PROCEEDINGS of the Society of Antiquaries of Scotland

Our full archive of freely accessible articles covering Scottish archaeology and history is available at http://archaeologydataservice.ac.uk/archives/view/psas/volumes.cfm

National Museums Scotland Chambers Street, Edinburgh

www.socantscot.org Charity No SC 010440 54 PROCEEDING E SOCIETYTH F O S , JANUAR , 1920Y12 .

I. REPORT OP THE EXCAVATION ON TRAPRAIN LAW IN THE SUMMER OF 1919. BY ALEXANDER O. CURLE, F.S.A. SCOT., F.S.A. To pick up the thread of the tale of our excavation on Traprain Law we must turn back to the report -which appeared in vol. ii. of the fifth r Proceedingsserieou e yeaf th o s rd 1916an . Afte o periodtw r f o s promising exploration and just as we had begun to adopt, after experi- ence methode th , s most "sitee suite were th w , o dt e reluctantly forceo dt abando r enterprisnou s beina a enatur f go e unsuite e seriouth r dfo s times through whic countrhthe passingywas , whe servicnthe evereof y able-bodied man was necessary to ensure victory. In course of time the only two of our staff who were fit for active service, our foreman Mr George Pringle and his stepson John Pitt, were absorbed in the army and fought in France. The war over, we rejoiced to learn of their safety, earln ana t dya e yea datth rn ei mad e preparatio resumptior nfo r ou f no work. As the use of the riddle is an indispensable step in our procedure, s uselesi t i o start s t operations unti e weatheth l s becomha r e fairly settled; we accordingly postponed the commencement until the end of April. Unfortunately Mr James E. Cree, who in the earlier years took so e mucburdeth f ho f supervisio o no t o s shouldersg hi o t n o n d ha , Americ businessn ao t beforbu , lef e assistee eh settl h o t t pla e e th ednm of campaign for the ensuing summer. In his absence Mr John Bruce assistancer camou o et gavd an , e ungrudgingly many day hourd san n so the hill. 191n I passee 5w d ove sligha r t ridge particularly noticeable th n o e lowest level consequencn ,i whicf eo e occupatiohth n ther onld eha y been partial. Accordingly realising this year that we had deeper ground in northwarde th o decidet s s fronu wa f t i ,o continu do tt than ei t direction as formerly. Takin northere gth t borderei s a n r flanfa d f ares unexploreko a aF d ground, to form a base of 60 feet due east and west, we pegged out an area measuring approximately 60 feet by 50 feet. To the east of this a small 25-foot section abutting on the upper part of area F had been explored in 1915, and north of this a farther 25-foot extension was delimited so as to produce a direct east and west front on the northern boundary of our new section. This area, including the small extension of 1915 namee ,w were d W aree . lateaG abl n greatee o rtako eth t n ei r part of another area, H, immediately to the northward of G, consisting e wester feeo 5 a th squarfeefn 0 2 feeo 5 n 0 t o y 5 tf b tno f e o end d ,an REPORT OF THE EXCAVATION ON TRAPRAIN LAW. 55 easterthe n tota halfThe l .are a explored durin pasgthe t season accordingly amount abouo st t 8500 square feet, or somewhat less than ^ acre. Before commencing work we were fortunate to find on the surface, coine th f smolea o turne fro e y mine b mon ,th p t dCrosraguelu a t , about which Mr George Macdonald enlightened us at our last meeting. This conceivably is referable to a fifteenth-century inhabitation by those in charge of the beacon, at that period to be used for spreading a warning of the coming of an English invasion from across the border. A portion handle oth flat a f eo mediaeval pitcher founcourse th n dexcavatiof i e o n probably owe presencs sit e same therth eo et cause. inchew fe f soia so f e removaOlo beneatnth ture d e th an f th f lo t hi highest floor level e las, th tha t f generao t l occupation t witme h s wa , abou foo1 t t beneat e surfacehth . Numerous large unshaped blockf o s the native rock lay scattered about, on many varied planes and angles, and the prospect of producing any reasonable plan out of them seemed hopeless; but by carefully considering each stone, and removing any that obviously owed their position to displacement, such for the most part lying loosely on the floor level without any hold, a rational scheme bega o evolvent , whic n beino h g plotted produce e plath d n given ni vers wa y t satisfactorI fig. 1 . fino yt d tha curvine th t g mas f stoneo s s laid bare in the portion of section G opened in 1915 formed the eastern elliptican a f o d len enclosure givin n interioga r measuremen fee0 4 tf o t by 20 feet. The letter G is placed in the centre of this enclosure on the e nortth plan n h t O i thersid. f o ee appea e foundationth r f threo s e small irregularly circular enclosures, and another lies to the south-west. Ther e periphera brea th s i ee mai n ki th f nyo enclosure towarde th s north-west, as if at this spot there had been a doorway, and some 17 feet distant in the same direction there is an indication of another enclosure of much the same form as the last, the west end of which has probably not yet been laid bare. As the stones on the plan are drawn to scale, t wili observee b l d thae averagth t e siz s veri e y considerable, several measuring about 5 feet in length by 2 feet to 3 feet in breadth. The removing of them took the efforts of three or four men at a time. There was nowhere any indication of actual building, nor did the amouny wa debriy suco t an hquantita n si wouls ya d point beini o t tg the ruin of stone walls. That the lines which form the enclosures were foundation obvious si s fro e contemplatiomth onle e th plan yth d f nan ,o reasonable hypothesis that suggests itself is that they were formed to carry wall f turfo s . On this level there were exposed the remains of four hearths. Of these three are so nearly complete as to make apparent the direction in which they were oriented, that t wili , observede lb , being practically east 6 5 PROCEEDING E SOCIETYTH F O S , JANUAR , 1920Y12 .

Scale of Feel

Fig. 1. Plan of Foundations on the highest level, Traprain Law. REPORT OF THE EXCAVATION ON TRAPRAIN LAW. 57 westrelatiod e an Th . f thesno e hearth e enclosureth o t st quit no es i s obvious. Ther hearto large n th s e i en hi elliptical enclosure liee son t bu , close besid t towardei south-easte sth closo s , ee presume thath f i t d wall of turf extende r beyondfa foundatioe limite th dth f o s t musni t have impinge fireplace e e ease areth f thisth th o n t ad o o n ,i T lai . d barn ei 1915, there is another which was associated with an area of paving, and may possibly have been in the centre of a large elliptical enclosure lying with its main axis north and south, which is suggested by what appeal-

ing. 2. Hearth on the highest level. to be foundations. In the enclosure shown towards the north-west corneplae th nf ro ther remaine hearthso ear tw f thesf o s O more . eth e westerly (fig. 2) is complete, but it will be observed from the illustration that it has been very rudely formed, not of carefully selected slips of edgn stono fort o f slabe t o kerb e se m sandstonf d th so an , e neatly fitted for o t e hearthf irregula mo th n i t bu , r block f stono sboulders d ean , evidentl readiese yth t that cam hando et . Similarly rud construction ei n was the hearth immediately to the south just outside the last-mentioned enclosure. It is perhaps worth noting that both these hearths are the same distance, 4 feet, from the walls adjacent to the north of them, to 8 5 PROCEEDINGS E SOCIETYOTH F , JANUARY 12, 1920. -

which the e alsar yo approximately parallel, whil e e heartth th e n o h southern margin of the plan shows a distinct trend from the true east and west direction towards the north-west, thus placing it more nearly paralle maie th ne adjaceno t lth axi f o s t enclosure almose Th . t total absence of charcoal, .or of discoloration of the soil, on this floor level was so remarkable as to make it evident that the period of occupation had been of the briefest. removaOe nth e remain th e lates f th lo f sto occupation t soma , o t e8 10 inches below, a second floor level was encountered. The plan (fig. 3) show e appearancth s e surfacth f eo e exposed. Ove l ther al lais r dewa bar ea well-define d floo pary t founf clayf an o no o rt t s a d, sucwa s ha e leveth l e aboveirregula e foundationsth s Th a . r r fa masse s a , f o s designatedo s e stonb n emuce ca ,ar h lightertheo d yr supplno , reay yan l indication of the lines of the enclosures that existed. Only in regard to the hearths is there any definite appearance of regularity. Of these there are shown the remains of seven laid bare last summer, and one in the south-east section expose n 1915i d f whico , h five orientear e d easd an t west. Moreover, when we place this year's plan alongside of that for 1915 we see that along the western side of our excavation we have a row of four hearths all practically in alignment and all open towards vere weste th yth n I centr. thif eo s year's plan fee0 5 , t fro westere mth n edge anothes i , r smaller hearth whic s orientehi n i ds i paralled an , to l alignment wit heartha hportioa kerbine th f n o whicf go h appears near centre 191e th th 5 f planeo . same Alsth en oi alignment t parallelno t bu , , ia thirs d hearte north-easth n i h t corne f thio r s year's plan. Lying nearly parallel to, and some 20 feet to the westward of the last, but out of alignment, is a small hearth closely resembling that towards systematicentre e plane th th Th f .e o c arrangemen thesf to e fireplaces thus disclosed seems to indicate that the dwellings with which they were connected were arranged in lines—or, may we say, streets? The marked difference in the general character of the remains on this and levep to le pointsth I ,think differena o t , t metho f constructiondo r fo , whereas the heavier foundations on the upper level, and the total absence of clay on the floor, strengthen the idea that the erections there were of turf, the slighter remains on this level, and the amount of clay over e flooth r suggests thae constructionath t l method employe n thii d s period was that of wattle and daub. Moreover, though we recovered exampleo n f thiso s materia lasn i l t summer's digging fros mwa t i ,thi s very level that we obtained so much in 1915. Where the floor levels are only separated by a few inches of soil on uneven ground it is inevitable that occasionally a mistake will occur, and there is no doubt that over a smal lsouth-ease areth n i a t par f sectioe secono t th , nH d level was REPORT OF THE EXCAVATION ON TRAPRAIN LAW. 59

1916

Scale of fill

Fig. 3. Plan of Foundations on the second level. 60 PROCEEDING E SOCIETYTH F O S , JANUAR , 1920Y12 .

o o

0(5 0

Explortd i

"O"

'O t> G H a. <3 'a 0 0

. • «

0 o

Soo/e

Fig. 4. Plan of Foundations on the third level. o REPORE EXCAVATIOTH F TO TRAPRAIN NO N LAW. 61 broken through, there bein remaint thao ga n t thiri te pointn th so d d an , exposed extene Th thif . o t s error was, however, slight. On the removal of this floor and. of some 5 or 6 inches of soil the third level was exposed. The remains found on it are shown on the plan (fig. 4). For the most part smallish stones are scattered about in a way that hardly suggests any structural form except as regards the hearths. Of these five were exposed set in a general direction north and south, but t uniformlwit no opee d hth nen y oriente eithen di r direction shows A . n e examplth y b e illustrated (fig, the5) . y were carefully fashioned an d laid thin i , s respect contrasting favourably witlevelp to h e . thosth n eo south-wese th n I t angl f areeo hearta s Gi r h toward e bac sth r soutko h end of which the remains of two parallel rows of stones, some 2 to 7 inches apart directede ar , f referencI . e third-levelmads i eth o et plan of 1915 similaa , r parallel arrangement wil observee b l north-wese th n di t corner of area F, the direction of which points exactly to the south end of the rows on area G, 12 feet away. At an identical distance to the south-eas e occurF row th n o sf a o shorst t singl whicw ero h possibly s beeha n connected wit arrangemente hth . Anothee founs th wa n r o d second level in 1915. The purpose which these lines were intended to serve is not clear. Their proximity to hearths would suggest the foundations of wind screens, but where, as in the instance now reported, the position e mai th e lin f ne th hearto e th axin i f o ss hi suc suggestioha n seems useless. Here, as on the level above it, the floor in places was formed othickla f y compacte claf o yd inche som5 dbe o t e4 s house e thickth f I s. were of wattle and daub, the disintegration of their walls under the influence of wind and weather after they were deserted would no doubt somn i e measur responsible eb e compacte th r this t efo bu , d appearance of the clay indicates, I think, constructed floors. f clay o d soi Witd an , l removae be wher hth e th ef o l claabsents ywa , to a depth of almost a foot in places, the foundations on the bottom were exposed (fig. 6). It will be remembered that in the 1915 excavation no remains whatever of foundations were found on this level. Last summer we were more fortunate, and two fairly regular segmental lines of stones a little to the north of the centres of Gr and H respectively show presumabl sitee ycirculath f so rhearth o hutsTw . s were laid bare, both well laid and circular, without any kerbstones, thus in respect both of shape and construction differing from the rectangular kerbed fireplaces of the higher levels. A record of the finds which was kept for area H shows that these, for the most part, are grouped near the centre of. the space enclosed by the segmental line of stones in that section, and also ove smala r e circulalnorte th areth f o ho at r hearth nea e westerth r n margin of the section. The general appearance of the soil on the bottom 62 PROCEEDINGS OF THE SOCIETY, JANUARY 12, 1920. stratu f moccupatioo n differs materially from l the al tha .n otherso t . s extremeli t I y blac discolouree placesn ki th d an , d layers hav mucea h

Fig . Heartthire .5 th dn ho level . greater thickness, tending to show that the site at the earlier period had been occupied for a longer time than during the later occupations. We have now considered the structural remains which were laid bare on each of the four identified levels that we have met with—taken in REPORT OF THE EXCAVATION ON TRAPRAIN LAW. 63

Explored a 1915

3°^i $ ^ (i4Qo& H V« >-f 00

Scale of Feel

Fig . Pla6 . Foundationf no lowese th n so t level. 64 PROCEEDINGS OF THE SOCIETY, JANUARY 12, 1920. the order in which they were revealed by the spade,—and I propose now to pass in review the relics found on-each. In order, however, to point t changeou developmentd an s s which occu s tima r e advances thud an ,s revea smala o lt l exten advance th t civilisation ei ninhabitantse oth f I , propose to start at the bottom level of all and work upwards. We shall also conside whole rth e relics from each direc o levet groupa s s a a l t o s , attention to the association of types or objects at more or less definite periods instead of treating of each class specifically, as was done in our previous reports.

RELICS FRO LOWESE MTH T LEVEL. With few exceptions the relics from the lowest level are shown on e formeth , 9 rd figsillustratinan 7 . e smallegth r objectse mosth r t fo , part personal ornament bronzef so , shale lattee ,th articlee etc.d rth an ; s of iron. A selection of special objects are shown at natural size in fig. 8. Thoug t discovere havye e t hw ey no Trapraian n e sit do f th eo w nLa stage th settlemen f culturen o i n eme f wheo t n stone implements were in vogue, we have evidence that men of that time frequented the hill, for they left behind on the 'ground we excavated this year two stone axes bot, 2) h(figd Nos, wit 7 .an e butt-en1 .hth d broken off—No measurin1 . g 8f inches in length by 2f inches in breadth, and No. 2 measuring 2| inches lengtn i inche2 y h b breadth,—alsn si oa thir d stone objec) 3 . t (figNo , 7 .

measuring inche3 s in length, resembling an axe of a lighter form, ¥ T

and not completelS y fashioned. It is also imperfect, in this case one e facth side f o beine g broken e offobjecTh . s i carefullt y ground dow flatteneand n d alon sidthe g e cuttinedgesthe but g, edghas e never been finished. Probably belonging to this early period also is a leaf-shaped arrow-head of brown flint (fig. 7, No. 4), imperfect like the foregoing. have t witw recognisablo eme r Sh n ofa e Bronz potterye Ag e t bu , her relicew alsfe f tha o sa o t cultural epoch have been left behino t d sho whosn w ma tha ee weaponth t toold an ss wer f bronzeo e likewise frequented Traprain Law. In 1915 we found the point of a dagger: this year we have got about a couple of inches of the iipper part of the blade of another (fig. 7, No. 5). It seems significant of the manner in which they came there that they are both broken fragments. Beyond thes t seee no mhavo objectt o d ee founw s d this yeay an r relicearlien a f so r date than that perio whicIroe e th n Ag f dho apparently synchronises wit comine hth Agricola'f firse go th tf o d sen legion e th t sa century of our era. The association of first- and second-century Roman pottery with that of native manufacture, and the close similarity of the REPORE EXCAVATIOTH F TO TRAPRAIN NO N LAW5 6 . relics to such as have been found at Newstead and on other Roman sites, indicate this conclusion. Of bronze ornaments ther a pai f s bow-shapei o er d fibula? (fig, .7 ) wit1 . h figd No floriate , an .8 , 7 d Nosdan knop6 . d theian s r respective pins working on springs. They are identical and in style much I'esemble one found on the lowest level in 1914. It is interesting to find a pair, for,

though these brooches were known t1 o have been so worn connected by a chain, they are not frequently found together. Two other bronze fibulae, both imperfect, are shown in fig. 7, Nos. 8 and 9, and fig. 8, Nos. 2 and 3. These are of the purely Celtic type known as the " S " or dragonesque fibula, and were fairly generally distributed throughout Britain presene , thougth o largt o n tp hu e number have been found. Portion wero tw e f obtaineo s d her 1914n ei , also from the lowest level, and, in addition to the finds of specimens in Scotland recorded in r Reporou r thafo t t year, fina 2 e example fro e for f mCastlehillth o t , Dairy, Ayrshire, was reported to the Society in 1918, found associated with some pieces of Samian ware and other relics. To the list given

lay the late Professor Haverfield in the Report of 3 the Excavations at Corstopitu m190n i 8 shoul addee db dsecona d exampl t Cirencesterea , this one in the collection of Mrs Cripps. Of the two examples recovered last summer, that shown in fig. 8, No. 2, s almosi t complete, lacking e end onl e bronze heaon yTh th .t f a do e which it is composed is very rotten, and the enamelling, which was in rectangular panels on the body, has disappeared. The .second specimen (fig. 8, No. 3) is less perfect, but retains some faeds of green enamel on the remaining portion of the body. Of penannular fibulae there a.re two. The first (fig. 7, No. 10) is of a type common on Roman and Romano-British sites, the open ends of the ring terminating in small fluted bulbs. In our last year's (1915) excavatio e founw n d f whicthreeo l al h, came fro e thirmth d level, while in 1914 we had two —one from the lowest level, and one from that above. A second-century date may be assigned to these. e samth f eo e styls th i ) s 5 ea . fig d No e secon, an .8 , Th 11 d . (figNo , .7 first, but the brooch is larger and the terminals and general fashion heavier. The terminals are flattened at the end and are spirally fluted. Among the penannular brooches found at Newstead are two of similar design. Ther e threear e rings silverf o f s thesi o : . 12) e (figeNo on ; , anothe.7 r (fig. 7, No. 13) of bronze. These two are quite plain and call for no

remark thes a , y might have been mad servo e t purpose y ean thirde Th ., i Proc., xlix . . 1671 .p . , No fig , .23 2 Ibid., xlix . 169p . . ,2 fig Nos, d .24 an 1 . 3 Ibid., liii . . 1281 p . . , figNo , .4 5 VOL. LIV.

REPORE EXCAVATIOTH F O T TRAPRAIN NO N LAW7 6 . howeve . r14) (figNo s , undoubtedli 7 . ya finger-ring s fashionei t I . d of bronze wire int a ospira f threo l e coils measuring in diameter irif inch similaA . r ringe s lowesfounth wa n i dt levee th 1914n i l d an , occurrenc Scotlann ei f otheo d r example e Repors notei sth r n i dfo t that year. e havW 1e alse bezeoth l portio f anotheno r finger-rinf o g small size, whereof the gem has been of glass "or enamel. A personal ornament is also a beautifully patinated object of bronze wire spirally fluted, and imperfect (fig. 7, No. 15). The head has been formed into a loop by bending back the wire. It measures 2 inches in length. Of bronze clasps, dress-fasteners, or strap-mountings, as these objects may be, we found on this level four practically complete (fig. 7, Nos. 16-19) and the loop of a fifth. Three of these are of the style with a square head resemblin thire foune th gon dn d o leve 1915n i fourtle Th . h (fig. 7, No. 19) is of the petal and boss variety. A similar one came from the lowest level in the 1915 excavation. This design is a prevalent one in late Celtic ornament. With these last may be grouped objects of a somewhat similar nature als f bronzoo d usuallan e y designated harness-mountings fourte Th . h leve s produceha l d fivf theseo e , thre f whico e e showar h fign i n, 7 . e bacth kmetaa t a s l. ha Eac loop22 e o hon Nost , 0 presumabl2 . r yfo attachmen a leathe o t t f thero strapo m shoTw .petae boswd th an ls design doubled thire variatioa th ; s di f thano t theme whic hI hav t eno met with before, and consists of a somewhat larger boss in the centre of ovan a l leaende th f whic o s h have been slightly rolleo dtw backe Th . remaining harness-mountings illustratedt ,no , simplconsista f o e e boson : s wit hmouldea e othe dplaia th edgef o d rn an ,dis c f inc diametern hi . Buttons of the type of the former have been found associated with relics Bronze oth f e Age, therefors i t i 2 e possible that thirelia s si c fron ma earlier period. Various examples of these harness mountings were found at Newstead, while our excavations of the two previous years have also yielded a number. Wit e thirrelice th hth f do s leve l (fig) ther s bee11 . eha n accidentally illustrate dterrea t (No wit) 1 . h three bosses placed symmetricalls it n yo outline, which really belongs to this level. It measures 2^ inches by 2/j inche n extremi s e diameter s identicai t I . l with others founn o d the two lower levels in 1914. The find spot was on the eastern portion

of H, where no foundation3 s of any kind were met with. Its period therefore is a little open to doubt. A small elongated heart-shaped object (fig. 7, No. 23) with a ring,

1 Proc., xlix. 5 . 175. p . ,No , fig. Coffey26 e Bronze *Se e Ireland, n i Th ,. e 93 Ag . p 3 Proc., xlix. 1 . 180. p . ,No fig, .31 68 PROCEEDINGS OF THE SOCIETY, JANUARY 12, 1920.

broken suspensior fo , heade th a well-known t s na i , article toiletth f eo . e poin Th s imperfecti t t originallbu , t i ywoul e notchedb d . This i s presumed to have been a nail-cleaner. Such articles have been found associated with earpick tweezersd e rinan sth g d whic an ,s bee hha n n o this example suggests its having been a part of such a toilet set. A similar nail-cleane s foun wa Generay r b d l Pitt-River e Romanoth n si - British village at Rotherley. Fig. 7, No. 24, is a small disc of bronze with a pin projecting axially from eithe s beer ha facenmountina t I . somn go e object—evidentlf yo iron, fro e iromth n oxide that adhere s undeit o st r side. , 6 . No fign i , e 8 centr.e uppe d th th an n f , I eo r 25 par . f figo tNo , .1 s illustratei dthia n oval plat f bronzeeo , covered wit exceptionalln ha y beautiful dark green patina. It measures 3J inches by 1^ inch. Along one side is an ornament consisting of three circles of different diameter connecte doubly b d e line punctulation i s n don repoussy eb e workt I . seems not improbable that this device is intended to represent a trumpet, t a large o e mouthfe th th whic eon d s i an smale ,d hth en le circlon t ea the other the bell, the intermediate-sized circle being intended for some ornamented moulding round the centre. The single dots in the centres circlee oth f s have certainl placo yn sucn ei hrepresentationa musd an , t therefore be regarded as a trifling decorative licence. From end to end along the back extends a line marked by the material which has been employe fasteo dt n this objec s principalit o t t . e Readinth n I g Museum, amon e findth g s from Silchestere b y ma , seen a similar thin oval plate, bent, but measuring in its original state 5| inche lengthn i s . Alon s e ornamenteedgi th g t i e f o d w witro ha repousse dots towardd e sidean ,smala th centres e i ,t sa th l t circulabu , r perforation with similar dots aroun . it dAlon e bacth g k occurn a s identical mark of cement or material for attachment. Though the Silchester exampl a trifl s ei e larger than ours, doubthero n s ei t that thes objecto etw s have serve e samdth e purpose, evidentl ydecorativa e devicTrapraie e oneth Th n .specime eo w nLa clearls ni y late Celtic. Objects evidently intended for ornaments, but which never attained thao t tlarg o statustw ee imperfecar , t casting f bronze o smore Th e. perfec consist) 7 . t (figNo f littl , so 8 . e more tha a mouldee hal f non o f d knop, which might have decorated a spear-shaft. The other has apparently been intende e samth r e dfo purpose s beeha n t eve,bu n mora f eo failure in construction. have W e als opieca bronzf eo e wire about 4£ inches long, whicy hma have been parpina e heaf o th t; f wha havdy o ma te fn beeincpi nha lengtn i formed han d wit ha doubl e curv pieceo . 4)etw ;f No (fig o s, 8 . a large bronze link, 3£ inche lengthn si , very simila portiono t r f linko s s REPORT OF THE EXCAVATION ON TRAPRAIN LAW. 69

Fig. S. Particular Objects fro e lowesmth t level. 0 7 PROCEEDING E SOCIETYTH F O S , JANUAR , 1920Y12 .

found on the second level in 1915; three pieces of finely paginated bronze edging, triangula sectionin r , 2ii inches, 2j\ inchesinc1 h lonand , gre- spectivel plata y ; thif eo n bronz bend e witen t e intriveha on oloo a t ta p inc1 lengthn hi smala ; l stud wit biconicaha l head ^| inc diametern hi ; a small fragment, about f inch long, of bronze binding, semicircular in section, similar to that obtained in both previous years ; and a cruciform object of bronze, like the nave of a wheel, with four broken spokes projecting. fro8) . mt (figi No , .8 Related to the ornamental objects of bronze is one of silver (fig. 7, smala s i r wit applianc y t looplba I o han n . r ,o 9) . r efigfo d No , an . No8 , .26 attachmen backe bearind th fron e n ,an th o t tn g o withi nborderea d panel remaine th devica f so e concludin t eithega r extremit volutea e n yi Th . desig centre panee th th n f i elo appears havo t e been wilfully destroyed. Leaving the metal ornaments, let us turn to those of other material. Aformen i s r years have w , e recovered many piece f glasso s bracelets (fig. 7, No. 27) and a few glass beads. The fragments of glass bracelets number fourteen, of which five are opaque white, two opaque canary- yellow opaquone , e greenish yellow opaqu,two e whit differeneof t section with brown lines translucene on , t pale green translucene on , t pale green •wit hblua whitd an e e twisted cord alon e centragth l ridge transe on , - lucent pale green with lines of yellow trailed on the surface, and one piec f transluceno e t green glass coated with opaque yellow whicn i " h is set a small crimson lozenge, while a black and yellow twisted cord appears at the side. This last-mentioned fragment shows a narrow necking removae mad th oute e y t eitheth eb a f rd lo layeren f yelloo r w glass. This condition lase pointeth s t n Reporti wa t s ,a d ou , characterises all these pieces of glass bracelet which have a core covered with opaque glass of some other colour, and indicates that probably such pieces have been mounted with metal. Glass bead l througal s r course excavatiohth ou f eo n have occurred so seldo ms temptei thae on to believ t d e that they wer t woreno n i n quantities on a string, as is the fashion nowadays, but sparsely suspended from necklaces in a style known to have been followed in the early La Tene perio Gaul.n di 1 Last yea e lowesth r t level yielded four complete beade most a fifthd hal Th f an o fs . ) interestin(fig31 Nos, .o 7 t 8 .2 f go these is a bead of dark blue translucent glass, compressed in manufacture to a cubical form with its angles rounded (fig. 7, No. 28, and fig. 8, No. 10). O nfoue eacth rf hsideo s paralle e perforatioth o t l t nbeari sspiraa f lo polychrome opaque glass. The diameter is -fi- inch. Two of the other beads are of translucent glass of a pale green tint: one (fig. 7, No. 29) regularly formed on a mandril, the other (fig. 7, No. 30) apparently formed 1 Dechelette, Manuel, . 1322p . REPORT OF THE EXCAVATION ON TRAPRAIN LAW. 71 from a drop of molten glass allowed to drop and then perforated. A similar e thirbeafouns th dwa n o dleve noted 1915n i l an othern , o s s discovered elsewhere are given in the Report for that year. The fourth bead is discoid and of opaque yellow glass, measuring -^4 inch in diameter (fig. 7, No. 31). It is of a class of which a number have already been foun Trapraion d n Law cominall , g fro lowesmthe thiror t d levels. Man f theso y e beads have been foun Celtin o d c sites d thers an , wa e reported last year to the Society by Mr John Smith not only the discovery of beads but apparently also of rods of glass from which they were being produced fore f Castlehillth o t n i , , Dairy, Ayrshire, associated with other relic second-centura f o s y date.e lasTh t 1 example, representee on y db half a s somewhapari , f o t t thick discoid bead, measuring -*-§ incn hi diameter, of translucent lapis-lazuli blue. There were found a number of pieces of Roman glass, probably representing four different vessels e largeTh . r numbe f piecero s (eight), including the piece of a mouth of a bottle (fig. 7, No. 44), were of dark green, thick material. Four fragments were of a thinner glass of light green tint. One piece was of medium thickness and of similar light green colour, whil othee eon r fragmen thif o ns colourleswa t s glass. The only other objects of personal adornment which remain to be mentione e segmentth e dar f shalo s r jeeo t bracelet . 32)f so ,No (fig , 7 . which this leve yieldes ha l leso dn s than twenty-four mose th r t parfo , t triangular in section. At the extremity of one there has been cut a chevron witmediae e pointh h th n o tn e braceletridgth f o e . Smaller rings of this material are conspicuous by their absence. Of objects connected with industry there is a set of five small chisels r puncheo bronzn i s e o 15)(figt , Nos.8 1 . 1 .Thes e var lengtn yi h from m inch to 2^f inches. They are not all perfect, and the cutting edge of the largest example (No. 11) is slightly broken. B f bronzeo a thi r s i n ba , , 38 3-^ . f No Fig inche, 7 . lengtn i s Ty hFb inch in breadth, slightly curved in its length. One more object call r speciafo s l mention—a spear-but f bronzo t e a modifie f o ) 16 d . pear-shape No d fig , an 8 . , 41 . d(figNo form, 7 . inc1 , h in length, |^ inch in diameter at the mouth, and 1£ inch at widest part. It is quite plain except for a single incised line just below the mouth. Withi s stili smalt na li l fragmen f woodo t , indicating clearly that this objec bees heae butspear-shafe linch-pina a ha t nth f f th t do o no d . an t A butt approximatel f thio y s type t morbu , e ornate s founa wa ,n i d broch at Harray, Orkney; and what I believe to be another, from which a disc at the bottom has been removed, occurs among the Roman relics from Camelon preserved in the Museum. 1 Proc., liii. p. 128, fig. 3. 72 PROCEEDING F THO S E SOCIETY, JANUAR , 1920Y12 .

Besides the three whorls illustrated (fig. 7, Nos. 39 and 39a) there is another of the disc type made of sandstone, making four in all from e wholth f thieo s y possibl levele dis ma Th ca fift) . (fige 40 yhb . . 7,No coursn i f constructiono e t o indicatiothern bu ,y attemps i ean f o n t having been made to perforate it. Fig. 7, No. 42, shows a trapezoidal piece of yellow sandstone measuring 3| inches by 2f inches. On one face, which is dressed, is neatly scored a series of lines crossing at right angles, or running obliquely from a single line that bisects the centre of the surface. The arrangement of the lines on either side of the central line produces a resemblance to some syste Oghaf mo m writing onle Th y. other stone object illustrated (fig. 7, No. 43) is a hone, coffin-shaped in outline, with a perforation at broae th d end. Of objects of stone not illustrated there were found three stone balls of fro incm1 o Iht f inc n diameterhi a polishe; d dis f sandstono c e 2|-| inches in diameter, |4 inch thick, with the edge slightly bevelled; three stonesf theo e m on , oblong f light-coloureo , d sandstone, 2^ inches by 1-^f inch in breadth, imperfect; another imperfect, oblong, and slightly taperin ende on ,o gt wit n arriha s alon edgee thire gon th ; d oblong, 2J inches in length by 1£ inch in breadth, with the longer sides slightly curved a fla; t four-sided piec f sandstoneo e , measuring 3f| inchey b s 3£ inches, wit hcarefulla y fashioned arris alon edgee gon ; another somewhat similar piece of thinner sandstone, measuring 4fJ inches in length by 2j-f inches in breadth, with an arris along one side, the rest of the outline appearing to have been roughly rounded; about one half of a finely polished disc of dark micaceous stone, the diameter of which when complete having been 3^V inches. The iron objects from this level are, like iro e mosth n f thingo t s from sitee th rathen ,i r poor condition onle Th y. weapon a spear-hea, d (fig, .9 No. 1) is of a different form from any found hitherto, being leaf-shaped and slightly lop-sided, one side having more of a curve to it than the other. It shows also a slight opening at the base of the socket—a feature not hitherto met with in the spear-heads from the early levels, which, so far, have all been characterised by closed sockets. The opening, however shorts i lengte ,e spear-hea th Th f . ho 7^s di inches. Fig. 9, No. 2, is a heavy mortising chisel, well fashioned, with a solid head to receive the blow of a hammer. This is the second chisel of this class tha have tw e found e otheth , r referabl e second-leveth o t e l occupation having been recovered in 1915. The length of this example is 9J inches. The long pin (fig. 9, No. 3) measures lOf inches in length. Its purpose is not self-evident. REPORT OF THE EXCAVATION ON TRAPRAIN LAW.

Fig . Iro9 . n Objects fro e lowesmth t level. 74 PROCEEDINGS OF THE SOCIETY, JANUARY 12, 1920. , appear4 . No havo Figt , s 9 . e bee e spindlth n a Roma f o e n quern. It is almost the counterpart of one found on the second le el in 1915, except abous thai t i tt \ inch longer. f iroo nd ro measurin a s i , 5 . g No som , Fig9 .e 8^ inches n lengti h with a domical cap on the head which comes down well below the point of contact. It seems too long to have been a linch-pin, but what purpose it served is not apparent. Fig. 9, No. 6, is without doubt a Roman stylus of the well-known type, measuring 3J inches in length. It came from the very bottom e leveloth fmeano n . y Thib uniqua s s i s e fin an o dRomano-Britis h site, showin gnative thath f i te populatio t themselveno d nha s acquired f writino t ar g e theth d acquireyha d some interes instrumene th n i t t that produced it. Eight styli 'were found by General Pitb-Rivers in different s excavatioplacehi n i s t a Woodyates,n d numerouan 1 s specimens were also obtained at Woodcuts and Rotherley. More

recently Mr Donald Atkinson found one, and probably a second2 , on Lowbury Hill, Berkshire,3 and one was obtained by Mr and Mrs Cuiining- ton in Casterley Camp.4 I e nailar 9 s witd an h , Fig , Nos8 squar 9 , . 7 . e section. Several others, were found e largesth , t measuring 4£ inches n lengthi . / Fig. 9, No. 10 measures 2| inches in length, and is a thin pin of iron bent in the centre so that its two ends lie in parallel planes. The upper s imperfecti d en s purposIt t .apparent no s ei . ( stapla s i , Fige 11 2y . . V9 No inche, breadtn si 2^d lengthn h-i an . Fig. 9, No. 12, is a hook. \ Fig. 9, No. 13, shows three iron rings varying in external diameter from If inch to 2J inches. ' Fig..9, No. 14, is probably a small portion of the umbo of a shield. It is a thin convex fragment with a flange at the edge. '• cleata s i , . 15 . No Fig , 9 . ,' , illustrate1 . No . , indefinitFigan 10 s . e objec f irono t , 2^%- inchen i s length whico t , s attachehi d transversely, possibl corrosiony yb thia , n plate of bronze "with a button-like disc at one end. Fig. 10, Nos. 2 and 3, are small iron hooks purposely so made, and nailnot s the point whicof s h have been bent similaA . r hoowas k found in 1914,6 and another (noted infra, p. 84) came from the level immediately above this. The pottery from this level consists of fragments of native as well

Pitt-Bivers, Excavations, vol. iii., pi. clxxiv. 15-17. • " Ibid., vol. i., pi. xxix. 6-8. cv. 3. 1 Atkinson, Lowbury Hill, p. 51, pi. xv. 13, 14.

3 1 Wiltshire Archceological Magazine, No. cxix., vol. xxxviii., pi. ii., No. 10. 0 Proc., xlix. p. 198, fig. 45, No. 8. REPORT OF THE EXCAVATION ON TRAPRAIN LAW. 75

as of Roman wares, the former greatly outnumbering the latter. The native pottery is coarse, thick, and made without the aid of the potter's wheel, also usually much blackened by fire, indicating that the vessels have been chiefly used as cooking-pots. In previous reports details have been given illustrationd an , f specimeso n sherdsectionf o d san s supplied ns oa t bu , fresh data have transpired in the excavation of the past summer there is nothing previoue th o t d s statementsad o t . The Roman pottery has supplied fewer fragment f intereso s t tha previoun ni s pieceyearso n d ornamentef so ,an d wares that hav bee t witt eno nme h beforee Th . most noteworthy of the latter are shown in fig. 7, Nos. 45 and 46. The most characteristic fragment containee sar e th n di following list:— Portio moute amphoran th a f f no ho , round-bellied type ; second century. Fragmen f cooking-poto t , drab clay Fig-10- Three Object Irof o sn froe mth "Wit.,,h silver-gre., y sli,. p coating,. , scoredJi lightl- i ^1y lowest level. (£.) with a horizontal band of comb-markings and detached vertical lines; apparently of late first or early second century date. Fragment of decorated Samian bowl, form 37 (fig. 7, No. 45). Small fragment from the side of a Samian bowl, form 44 (fig. 7, No. 46), good hard clay and glaze (cf. Roman Frontier Post, pi. xl. fig. 20). Fragments of Castor ware cup Or beaker, form 55; second or third century. Smal fragmenm ri l f Samiao t n bowl 35/36, decorate recurvee th n do d rim with leaves en barbotine; Lezoux fabric late first century (cf. Roman Frontier Post, pi. xxxix. fig. 5). Small rim fragment, pale silver-grey fumed clay, hard, with well- smoothed inside surface; rare shape of open bowl with slightly inturned rim (cf. Kastell Wiesbaden, O.R.-L., xxxi. pi. xiii. 9, 13, 25, A.D. 83-121). Small fragment of a decorated Samian bowl, form 37; style of Germanus, A.D. 60-85. Remains of leafage and other ornamental detail indicate that this is a piece of the same bowl as a small sherd found 1915,n i 1 whereon appear e greateth s r charginpara f o t g lion. Some small fragment Belgif so c black ware. Rim fragment of a flagon with three receding cordons (cf. this piece 1915).n i t go 2wit e hon 1 Proc., 1. p. 96, fig. 21, No. 1. 2 Ibid., 1. p. 90, fig. 18, No. 7. 76 PROCEEDINGS OF THE SOCIETY, JANUARY 12, 1920. Fragments of a fumed grey cooking - pot, scored lightly with latticed lines. Fragment f Belgio s c black ware fragmentm ri ; f globulao s r pipkins r beakerso , fumed grey clay, coated with bitumen t polisheno , n o d outside surface. Fragments of Castor ware, white clay with brownish red slip coating. fragmenm Ri f Samiao t n bowl, form 18; late firs r earlo t y second century. Fragment of the side of a Samian bowl, form 31; second century. Small fragmen a Samiae sid th f eo f o nt bowl, form 37; portiof no a zon f "-shape"S o e d ornament. Fragment fro shouldee mth narrow-neckea f ro d cooking-po r storo t e vessel, coarse fumed grey clay with reddish core; • a cordon at junction of necshoulderd late kan b ey firsma ; t century. Two rim fragments of Upchurch or late Belgic black ware decorated round the outside with a lattice pattern, scored lightly with a blunt point (cf. Proc., xlix . 162. 4).p . ,No , fig18 . Lastly, this lowest level yielded two coins both from section H—one a denarius, the other a small brass, both of Faustina the Elder—which, along wit e potteryhth , serv indicato et e perioeth f occupationdo s A . both coins are of the types struck subsequent to the death of Faustina in 141 A.D., they clearly belong to the Antonine period of Roman occupation in Scotland. The foregoing notes on the Roman pottery show fragments of late firs r earlo t y second century wares s wela , s piecea l f vesselo e s th f o s Antonine period; thus pottery and coins supply evidence that this primary occupation continued well into the second century. Of the date of its commencement we cannot be certain; but, as we shall show e perios endurancit th late , f o don r e was much longer than thaf o t any one of the three occupations that succeeded it. RELICS FROM THE THIRD LEVEL. What we designated the third level as we reached it in succession from the latest refers, of course, to the second occupation, as we consider e relicth n chronologicai s l sequence, takin e levelth gp frou e s mth earliest. On level No. 3 we naturally do not find relics from the earlier Ston d Bronz e an lowese periods th e floo Ag eth f o ts r a ,occupatio n intervened between it and the ancient surface. One arrow-head of flint, however, of the lozenge-shaped variety, with a slight chip out of one thithery edge s founwa ha ,s it d, probabl actioe rabbita th f y no yb r o , the still more disturbing proceedings of someone bent on digging out

78 PROCEEDINGS OF THE SOCIETY, JANUARY 12, 1920. the rabbit. This is the fifth flint arrow-head that we have found, and s worthi t i f remaro y k tha e f lozenge-shapetheso th t f eo foue ar r d variety, whil e fift s th leaf-shapedi eh t .me e hav t w Thu no er fa s with a. single barbed specimen. Of bronze ornament e havw s e fewe e rlowe th tha n ro n level. There are two bow fibulas, both of quite distinct forms from those previously noted. One of these (fig. 11, No. 2, and fig. 12, No. 1) is a plaifibulw nbo a identica n styli l e with that w founbo n 1914.e i d Th 1 is rounded and plain; there is a concavity in the foot which has contained a boss, and there is a similar one at the head. The pin work simpla n o s e hinget uncommofore no Th ms i . Romanon no - British sitesd referencean , o othet s r occurrencee e giveth ar s n i n

Fig. 12. Fibulae, and Kings from the third level. Q.)

Keport for 1914. The second fibula (fig. 11, No. 3, and fig. 12, No. 2) is of a less common variety, and a beautiful example of its kind. The bow, which is quadrangular in section, \vith bevelled sides towards the foot, swells out to a trumpet-bell form at the head. On the a circula s i w r bo dis e cth cresfro f o mt which four small points projec t equaa t l distances apart, tendin produco t g squara e e appearance. This dis s outlinei c d wit ha thi n threa s coverei f gold o d dan d with enamel, which apparentl s beeha y n crimson spoA f .silveo t r form e centreth s . Three other t spotalon se f silveo se g ar reac h of the flat sides of the forward section, while a band of the same metal, flanke eacn o d h singla sid y eb e spot, runs alon e ridggth e from workn a spring pi n e e heado s th e disTh . o th .t c Ther s beeha e n a ring on the head which has been broken off. General Pitt-Rivers foun fibulda identicaf o a l form ornamented wit hstua f bludo e enamel 1 Proc., xlix. p. 169, fig. 2i, No. 3. REPORE EXCAVATIOTH F O T TRAPRAIN NO N LAW. 7& e Romano-Britisith n h villag t Woodcuta e s Common similaA .e on r was found at Wroxeter.1 ) measure3 . No d fig, Aan s12 . , rinft4 j. ginc No (fig n diamete, hi 11 . r and is formed of a plano-convex wire in such a way that the ends are turned back and slightly overlap. There is also a fragment amounting o 1t £ a coilsmal f o s l finger rin f , silverformeo g4) a . f o dNo (fig , 12 . tapered wire with a fine spiral rib, §|- inch in diameter. a penannula s i , 5 . No Fig, r 11 rin. g made fro mthia n plat f bronzeo e tapere eitheo dt rgreatese endTh . t diamete mors yha s ri f et i inch t bu , the appearance of a mounting than of a personal ornament. Of the so-called dress-fasteners there are three (fig. 11, Nos. 6-8). fashioneNos i 6 . d wit hsquara e roundea plat d ean d loop doed t an , sno differ essentially from those of the same pattern found on the lowest level. The same remark applies to fig. 11, No. 7, of the petal and boss variety. Conceivably it made a pair with fig.' 7, No. 19. The other specimen, however, of this class (fig. 11, No. 8) has all the appearance of incipient decadence in its lines. The boss has got flatter and broader, and the curves are less pleasing. Of the harness-mountings (Nos. 9-11 of fig. 11) No. 9, showing a double petal and boss design, equally with the last object shows a falling-off workmanshin i p e fro identicae stylth mth f o e l ornaments found on e e lowesth th trueo s d t an , e e levelcurvepetalno th Th e f .ar so s object is clumsily designed. The mounting next to it in the illustration, No. 10, is the exact counterpart of one found on the lowest level, and it is conceivable that the lowest level is its proper provenance, for both were e samfounth n ei d section (G)d wheran ,e th e dividing stratum is shallow and uneven an occasional confusion is, I fear, inevitable.

Another harness-mounting (fig. 11, No. 11) is a plate of bronze 1¥V inch square, furnished, like all the foregoing, with a loop at the back. Connected with harness was probably also the ring (fig. 11, No. 12). 7 t I measures 1T ¥ inc n diameteri h s peculia i e fashio t i Th n .f i ro n respect that the ring itself is not of the same thickness throughout, but swells out on two opposite segments in a style characteristic of Late- Celtic craftsmanship. One other ornamental relic of bronze was recovered from this level, a pin (fig. 11, No. 13, and fig. 13). It measures 6}| inches in length and, although bent, appears to be complete. It is similar to a pin obtained from the lowest level in 1914. The head is somewhat decayed, but it has been rounded, and on the front of the pin there is a slight projecting lip curving upwards at either end. It has some resemblance in the latter

Urinconium, Thomas Wright, p. 280, No. 5. 1 0 8 PROCEEDING E SOCIETYTH F O S , JANUARY 12, 1920.

respec a zoomorphic-heade o t t n founpi dn 1915 i d d an , describe d illustratean de Reporth r than i dfo t t year.1 The upper part of the head on this pin does not seem to have bee elaborateo s n t intervalA . s there occur zones of incised lines encircling the stem with the evident in- tention of keeping the pin in position. The only other object f bronzo s e recovered from this level were two in number. One appears to be the half of a disc, 1% inch in greatest diameter with a trefoil or quatre-

foil piercin7 g in the centre, and suggests the lockplate from tubulaa f o d ren padlock e th othee f o Th .r objecba a s i t bronze inche5 , lengthn si , J inc breadthn hi , quadrangular in section. Closely actuaalliee th o dt l bronze ornament moulde sar s for the manufacture of such articles. One complete mould sectioned and portions of three others are shown in fig. 11, Nos. 18-21sectionee Th . d moul founs dwa d s perfecinit t state as it was finished ready for use. It is of clay and bees ha n made exactl same th en yi metho thoss da e found in 1915 a metho, d fully describe e Reporth n di t for that s beeyear - s wilha e observenin A t b i .l , 1) . d No (fig, 14 . tende cascylindehala o e d t f on o ft r with five corrugations s beee outsideoha th n n d furnishean , d wit a hcor o t e preserve its hollow character. The three other moulds are each for pins. No. 19, and e heath , d2 . parNo t, figonly 14 hand-pia s bee. r ha , nfo n with six beads. We got a pin of this type from the lowest leve n 1914i l , whic d , figd fivan 14 .ha h, e 20 beads . No . No. 3, has been intended, and more probably used, to cast a pin of a type which, as far as my knowledge goes, is represented by no single existing specimen. It has, like e hand-pinth , bee nshoulderea heaa d dha s dha pind an , in the form of a cinquefoil. The full length of the pin as it left the mould has been only 1J inch, and, as will be notice diagrammatie th n di c representatio poine th , tit f no has been blunt. Probably a sharpened point was produced by hammering and the stem consequently lengthened before the object was finally finished. The remaining mould s apparentlha ) 21 . yNo , a (figsimplbee r 11 . nfo e pint bu , t presenti a peculias interioro featurtw s rthan ha i e t i t faces, one at right angles to the other. On that shown 1 Proc., 1, p. 39, %. 24, No. 3. REPORT OF THE EXCAVATION ON TRAPRAIN LAW. 81

Pig. U. Nos. 1-3, Clay Moulds, with diagram f relativo s e Castings , Groove4 ; d Dis Seemenf5 c• c diagrammatid an c completio f Potterno y Saucer l fro e thirAl mth . d level.' ('.) VOL. LIV. 2 8 PROCEEDING SE SOCIETYOFTH , JANUAR , 1920Y12 .

illustratioe matrixn inth pi e th , whils i n e otheth e s s plaini i r t I . conceivable that a simple pin matrix and a plain surface existed on the counterpart of the mould with their positions reversed. Of other ornaments not of bronze there were found, as on the level below, a number of segments of glass armlets (fig. 11, No. 14). Of these there were seven; of which two were of opaque white glass with no ornamentation ; two of opaque white, but with thin blue threads trailed surfacee o nf palth o eo translucentw ; t blue glass, each with three blue whitd an e intertwisted cord mouldings,e alonlaie on crese don d gth an t along each side. The last fragment was a small one of pale green translucent glass. observes A r slaswa ou t Reportn di e braceletth , f opaquso e whitr eo yellow glass wer f heavieo e r make than those mad f transluceno e t material. In addition to the foregoing, there was found a small wedge- shaped fragment, rathe thico rto hav o kt e bee narmlety paran f to , such as we have found so far (fig. 11, No. 15). It is of blue and white millefiore s beeha glass nsectioa d an , f somno e plano-convex rod r ring,o , which has had its surfaces ground or rubbed down to their present state of smoothness style Th suggestivs ei . f Romaeo n manufacture. Fig. 11, Nos. 16 and 17, show two objects of amber, one-half of an annular bead wit hdiametea smala inch 1 d f ro lan , discoid bead f inch diametern i . Since we left the lowest level a remarkable change has occurred in the prevailing fashion of shale or jet ornaments. Not a single example s com e heavha ligho th t e f o yt segment f armleto s numerouo s n o s that level, but we have in place four rings and a bead of this material (fig. 11, Nos. 22-26). Of these No. 22 is a slightly made ring, imperfect, wit hdiametea Iff ro f inch t alon I appearancn .i s ei e lik armletn ea t ,bu 'smalo to s li wrisfoe t i th r f anyono t vera t ye bu small child. d Nosan 3 2 . e flattene24ar d rings, probably worn suspende, 25 . s beadsda No d an ; with a diameter over all of 2 inches, has in all probability been put to a similar smala purpose s i , l 26 spherica . . No Fig , 11 .l bead, flattenet da perforatione th f eitheo d measurind en r an , gf inc diametern hi . Part oa verf y similar beas recoverewa d d fro e bottomth m leve 1915n i l . No. 27 of fig. 11 is one half of the head of a pin similar in character to one found on the upper level in 1914. The last article for personal use to be mentioned is a thin flat object fashione f clay-stono t ou d e inta e disfor th of cmo wita h tongue projecting in one direction, the disc surrounded by an incised lin . 28) e eithe n notc(figo No A . t , .e rjunctiohcu 11 th side t th ea f no tongue and disc has evidently been made to hold a cord used for the suspension of the object. The design, which is possibly phallic, is very REPORE EXCAVATIOTH F O T TRAPRAIN NO N LAW3 8 . muc samhthe tha eas t employe abovdecoratdto pin e heathe ethe dof described. e numbeTh f completo r e whorle thir ) founth 29 dn . o ds No (fig , .11 level amounts to eight, while seven different halves give a representative total of fifteen. Of these ten were fashioned from stone, one from the foot of a small Samian bowl, one each of Roman and native pot respectively, and two of lead. Among those of stone is a half-whorl made of clay-slate botn o , h face f whico s h numerous lines have been scratched with some sharp instrument. One of the objects of lead which I have include whora s da s onli l y f inc diametern hi . For the first time this season we meet with the small discs of stone or potter occasionalld yan f glasyo s which were evidently use playings da - me r countersno f thesO .have ew x (figd 31) esi , Nos an ., 11 varyin0 .3 g in diameter from ^ inch to If inch. Of stone objects ther founs irregulae ewa on d r oval dis f sandstoneco , measuring superficially 2f- inche 2y ^b s inches a coprolite; , 1| incn hi diameter, similafoune on 1914, n whichd i o t rinformed an m a I , , must

have come from carboniferou1 s shale beds discoiA . d objec f coarso t e pottery, If inc diameten hi f inc y depthn hb i r , wit hmarkea d hollow round the middle giving it a bobbin-like aspect (fig. 14, No. 4) is of unascertained use. A small bal f bakeo l d clay wit hdiametea f Ifo r inch doe t falsno l foregointhe of intgo any categories. Similar pellets have been foundin each of the previous years. A small number of pieces of Roman glass (fig. 11, Nos. 32-33) came from this level. The principal relics of iron recovered from the third level are shown n figo . .15 There are two spear-heads—one (No. 1) leaf-shaped, measuring 5f3^ inche lengthn si e otheth , r (No mor) 2 . e lanceolate, measuring 6| inches. Both have closed socket d well-definean s d midribs n thesi , e respects showing features which apparently distinguishe e purelth d y Celtic spear-heads of our two earlier occupations from those of the later periods. gooA d pai f shearro s (No measurin, 3) . g 8£ inchee th lengthn si f o e ,ar usual type—a type that has proved its effectiveness by its survival to e presenth t time wit a hsomewha t restricted use similaA . r pairt bu , smaller, was found in 1915 on the second level. There are three knives (Nos. 4, 5, and 6), two with triangular pointed blades (one lacking the point), the other with seemingly a rounded point. 1 Proc., xlix . 201p . . 84 PROCEEDING E SOCIETYTH F O S , JANUAR , 1920Y12 . A shouldered pin (No. 7) belongs to a type of which examples have been foun bronzn i d r previou ou eacn ei f ho s year's digging. Similar pins in iron were found respectively in the kitchen-midden at Gallanach, Lawse Th Obanfore t ,th a Monifietht n i ; , For Iroe fth arshire nn i d an ; Age burial at Moredun, near Edinburgh.1

Fig. 15. Iron Objects from the third level. (J.) smala s Noi l8 . hook Ij^ inc lengthn hi similaA . r hoo s founkwa d 1914.n i 2 No muc9 . h resemble sprina s g whic s beehha n hel position di a y nb pin throug hrinbrackea d gan t eithea t r end t measureI . s 6| inchen si length. 1 Proc., xxxviii., 19C6 . 433,p . 8 , flgs.-Fd an , 7 , 2 Ibid., . xlix.8 . ,No , fig45 . REPORT OF THE EXCAVATION ON TRAPRAIN LAW. 85 There is illustrated by fig. 16 a small mortising chisel of iron, rectangula section i r n throughou measurind an t incheg2 lengthn i s . Fig. 17 shows two aspects of a flattened oval object of iron wit hrectangulaa r transverse piercing, apparently for a strap to pass through. It measures superficially If inc |thicknesfn y i hb incd han s \ inch. Thouge hth rounded upper surface is now covered with iron, there is clearly discernible at the edges a thin plate of bronze, demonstrating that the iron which now overlies it is due to oxidisation. The shape which this plate of bronze must take at once recalls the ornamented plate found on the lowest level (fig. 8, No. 6), and, though of considerably larger size latte originallthe , may r y have been employed decorato t e uppeeth r surfac similaa f o e r though larger oval boss. Fig . Mortisin16 . g The pottery fragments from this level show the usual Chise f iroo ln mixtur f coarso e e native ware with Roman wheel-made from the third pottery. Amon e formeth g r ther s nothinei g that calls level. (}.) ^ J for any special remark except one small fragment that appear aboue b o st t one-thir a diminutiv f do e saucer wit hdiametea f ro risinm ri Igfa \incd hincan h abovbottoe levee th th f em o l (fig, .14 No. 5). A very similar saucer was found in a broch at Brabstermire, in the parish of Can- nisbay, Caithness, a hollowepebbl f o t e ou d 2f inche lengtn si inche2 y h b breadt n i s d han f inch in thickness.1 Such objects do not appear to have been crucibles — a stone is unsuit- able for such a purpose — but they may have been used for containing pigment of some sort. The Roman pottery shows a considerable variety of periods. There are a few pieces of Saniian ware — the following being those worth f remarkyo . 17 val W A rim fragmen6 t of a Samian ware bowl ^Bronze'J? from 5the^ thir* d ,level. (J.) (form 37) with an ovolo border, showing a beaded line beneath and at the lower edge what appears to be the upper part of a large scroll (fig. 11, No. 34). The style is eastern Gaulish or latest e secondLezouxs e datth it th f ed r o o probabl,an ,d en e yth beginning of the third century. A small rim fragment of a similar bowl with a thick bead moulding •probably of third-century date. Royal Commission Ancientn o Monuments (Scotland), Third Report, Caithness,. 37 . No 1 86 PROCEEDINGS OF THE SOCIETY, JANUARY 12, 1920. A fragmen smala f o t lsame bowth f eo lmateria l (for wit) m37 h panel ornament having beaded borders (fig. 11, No. 35). In the upper portion of the narrow panel on the left is a small caryatid (Dechelette, No. 656) use Cinnamusy db , Devixtus, etc.; late firs earlr o t y second century. A rim fragment of a bowl or saucer (form 18); first half of the second century. Two base-fragments of a cup (form 33) with the first three and last three letters ALB—ANI of a central potter's stamp, possibly part of the name ALBUCIANI—a second-century potter. Of the Roman wares other than Samian the most conspicuous are the piece f Rheniso s h pottery, part f bulbouo s s beaker f blacso k slip ware decorated with zones of scrolls and berries in white slip bordered by groov d roulettan e e markings (fig , d Nos37).11 an ;6 .3 period A.D. 150-250. Inoteworths ti y that non f thieo s war s beeeha n y founan n i d of the Roman sites in Scotland excavated by the Society. Its appearance here therefore must be dated subsequent to the withdrawal of the Romans, i.e. the very end of the second or the first half of the third century. Numerous pieces of this ware were found at Silchester.1 It f intereso s i noto t t efragmentw thafe a t s "were foun excavatinn i d g the broch of Keiss, parish of Wick, Caithness, and are preserved in the Museum. In 1915 pieces were recovered from each of the levels except the lowest. A small fragment of grey Upchurch ware, part of a globular beaker decorated wit hzona f rouletto e e hatching. Fragment f Upchurcho s r lato , e Belgic, ware r moro ,e eparton f o s pear-shaped pots decorated with vertical lines made with a rounded point f palo , e grey body with darker surface—smoothed abovd an e below the decorated zone; early second century. Fragments of this ware were found in the lowest level, probably its proper position, in 1915. fragmenm Ari a smal f o t l globular cooking-pot with slight upright bead rim; found also at Corbridge; Antonine period, A.D. 138-192. Several fragments of Castor ware of white body with black slip coating ; second to fourth century.

Three coins came from this level: from the upper or eastern half of Gr a smal, l bras f Probusso , A.D. 276-281 similaa d ,an r coi Allectusf no , 293-296; while from the centre of the western portion of H came a small brass of Carausius, A.D. 287-293. These third-century coins found at different places throughout the level may be taken, along with the fragment f Rhenisso h pottery shoo ,t w tha e occupatioth t f thino s level in the particular area excavated last summer extended well into, if not i May, The Pottery found at Silchester, p. 103, pi. xlii. A. REPORT OF THE EXCAVATION ON TRAPRAIN LAW. 87 throughout, the third century. This is a much later date than was attributed to it in the previous Reports from the evidence then obtained, the latest coin connected with it found up to 1915 being of the reign of Antoninus third-centurPiuse Th . y coins s yieldewhicha t hi d this last season are the first coins of that period that have hitherto been found on the hill. The significance of this is not clear. It may imply that during the period represented by each level the occupation was progressive over t simultaneoussurface no th d ean inhabitante ,th s movin freso gt h ground as occasion arose. Or it may sh&w that the duration of the occupations was much longer than we have hitherto estimated. As to the length of occupation of this particular level in last summer's excavation, I shall show- late than ro t relativ primare thao eth t f to y occupatio briefs wa ,t ni or else the site was used by a much less dense population.

RELICS FBOM THE SECOND LEVEL. We have now reached the period of the third consecutive occupation of the explored area—that relative to the second floor level reached in

Fig. 18. General illustration of Relics from the second level (other than of Iron). (Ca. J.)

our digging from the surface. It is that on which the hearths lay for mose th t part easwestd werd an t an alignment,n ei . Th ee relic bulth showe t wilf I ar sko observee figb . ln n i .18 d that 8 8 PROCEEDING E SOCIETYTH F O S , JANUAR , 1920Y18 . e numbeth f objecto r s foun s muci d h smaller than that obtained from either of the two preceding occupations. The only fibula recovered (fig. 18, No. 1, and fig. 19, No. 1) is of the penannular typediametee Th . Is i fr inch t originallbu , t ywouli d have been a trifle more, the brooch having been slightly crushed. The terminals are exactly of the style to be seen on one half of a penannular fibula found in 1915l on the third level, and believed to be zoomorphic. Though the style is the same respecte on n i , ther differa s ei - ence in the terminals—that of the 1915 brooch shows a barely perceptible expansion in thickness fro me wir th tha ef o t formin e maith g ne th par f o t brooch, while in the more recently recovered example th e ends broade t markedlyou n . As was related in our last Re- port, this typ f brooco e s ha h been r ReginalshowM y b n d prototype Smitth e e b th ho t f eo later' Celtic brooches wite hth richly wrought terminals, the best-known example of which from Scotland is the Hunterston Brooch broadenine Th . e th f go terminal examplr sees sou a n i e indicate importann sa t sten pi Fig . Pin Broocd 19 . san f Bronzho Disd f Iroean co n the development. It is prob- fro e seconmth d level. (}.) ably of fourth-century date. Two small slightly penannular rings of-bronze (fig. 18, Nos. 2 and 3) have more probably bee movable nth e head pinf so s than fibulaB. Of actual bronze pin e recoverew s d an , 4 d . twoe (figNo , On .. 18 fig. 19, No. 4) is a modified form of the hand-pin referred to above. t The numbe f beado r s reduceupow e heano nth s threedo i dt t measureI . s 2| inches in length. In an article tracing the development of the hand- pin in Great Britain and Ireland, Mr Reginald Smith gives to this form a fourth fifth-centuro t - y attribution. e relegatioTh 2 f thino s example e fourtth o t h century wils a show,e b l considerinn ni e dat gth f thi eo s occupation probabls i , y correct. ) measure3 . No figd , san .19 , 35 ^ . e othe inche(fign No Th pi r, .18 s 1 Proc., 1. p. 101, fig. 23, No. 5. 2 Opuscula Archceologica Oscari Montelio, p. 287, fig. 14. REPORT OF THE EXCAVATION ON TRAPRAIN LAW. 89 identicae th e lasfounf n th o t pi lengthn i s n leveli le o d typ d th f ,an ,e o shown in fig. 13. A pin with a similar head but larger was found in thire th d leve 1915.n i l 1 Closely connected a clawithale f pine yon o fh th moulds si , If incn hi length (fig. 18, No. 6), for casting a small hand-pin, which has apparently had five beads on the head.

Fig . Moul a Spear-butt20 . r dfo d diagrammatian , c drawin f Castinggo . (}-.)

One half of a large clay mould (imperfect) is shown in fig. 20 with a diagram of the object which was intended to be cast in it. This has evidently bee nsocketea d spear-butt wit flatteneha d spheroid terminal measuring 2| inche lengtn i s ^ d inchan diameten hi mouth e th t a r. Of beads there occurred one discoid bead of amber, ii inch in diameter (fig. 18, No. 7), very similar to one recovered from the level below, and a, small cylindrical bea f opaquo d e green glass ^ inc n diametei h r 1 Proc., 1.9u.d , Nos, figan 23 .A 9 . 90 PROCEEDINGS OF THE SOCIETY, JANUARY 12, 1920. similao Tw . r 8) bead . f greeNo o s , (fig18 n. glass were found in 1915 —one on this level, and the other in filling in the soil. The horizon is therefore probably e e typperio correcte fourtth th th ef d o d han , century. An object whic s apparentlha h y been worn suspende a bea s s a di another fish coprolite (fig. 18, No. 9), ground into discoid form and arti- ficially perforated. A groove radiates from the perforation in one directio e edgeth o ,nt probabl y produce frictioe th cora y f db o dn when being amuletn wora s na r beado , . Pieces of glass bracelets turned up on this level also; four of them are illustrated (fig. 18, NQS. 10 and 11). One is of the opaque yellow glass d wit an ,a h markedl y triangular section, which from accumulated experienc I einclin attributo t e e earlieth o t er occupationsf O . the three other pieces one is translucent pale green and ornamented with yellow streaks, and the others are opaque white. All three are lightee oth f r make shod an , w a semi-ova l section. Shale, or jet, is represented by a single fragment (fig. 18, No. 12), one ringa hal f o f. The playing-men or counters, which first carne to light on the third level, appear agai n thisf theso nO . e objects ther f e seveno ar e x si , which are shown in the illustration (fig. 18, No. 13). Two of these have been fashioned from fragment f Samiao s n ware ,f shale threo d e an ,ear two are pebbles. Wit e stonhth eincludy objectma e e w sanothe r white fish coprolite,. with an extreme diameter of yf inch, spherical in form, and in its natural state; two hones, one (fig. 18, No. 16) measuring 3^V .inches by 1 inch by ^ inch othere illustratedt th ; no , , rectangula sectionn i r , measuring 5 4T g- inche y 1^b s- inc fy fhb inch a thi; n dis f porphyryo c , measuring 2^f inches in diameter, highly polished on one face; a quadrangular flake of sandstone with one edge sharpened, measuring 3-^ inches by 3f inches j a large scrape f fino r e black flint, measuring If inc Iy fhb inch, which obviously has been brought up by some means from a lower level. Lastly, with these miscellaneous stone relics we may.mention an oval discoid objec hardenef o t d clay, measuring 2^ inche 2y ^sb inches. This article s evidentlha y been employe r polishingfo d face e surfacth on e s n a , o e towards each end has been worn to a convex outline by attrition. e perforateTh d disc r whorlo s s from this level number twelve, of which ten are illustrated (fig. 18, No. 14). In the case of two of these objects the perforation is so small, only -£% inch in diameter, that it is barely conceivable that they were used to put on a spindle; the perfora- tions of the remaining ten quite suit them for their presumed purpose. lattee Tw th mad e f orar o e from piece Samiaf so n ware. REPORT OF THE EXCAVATION ON TRAPRAIN LAW. 91 learne Ae r excavatiosw ou n di f 1915no , ther notabla s ei e increasn ei the pieces of Roman glass found on this level (fig. 18, Nos. 15-17), and especially noteworth e fragmentth e ar y f delicato s e f vesselo d an s coloured glass referable to a period of occupation of the site that must be dated long subsequen withdrawae th o t te Roma th f o l n legions from e territorth y nort f Hadrian'o h s Wall. Ther e probablar e y twelve different vessels represente e followinth y db g pieces:— piece1 , yellow bronze tint. 3 pieces, very thin, pale green. 1 piece, very thin, pale moss green, with a band formed of finely scratched lines. 1 piece, of a rim, thin, slightly darker in tint than the last. Several pieces of colourless glass, much cracked. 2 pieces, very thin, pale yellow. 3 pieces, very thin, pale blue. piec1 littlea e thicke darked ran r tha laste nth . 1 piece, colourless, showing an everted rim and moulded edge. 1 piece thin, colourless, with moulding. A segment of a hollow rim of green glass. Several pieces of thick green glass bottles of large dimensions. There remai o mentiot n n before considerin e w potterth fe g e th y following objects of metal. 5 A lead disc, plano-convex in section, with a hole T ¥ inch in diameter, centre flas sunth it t n n ki undersideeo , evidentl enablo yt e objeceth o t t be fixed on the end of a pin of some sort; three short fragments of bronze binding, semi-tubular s possibl wa d suc s an ,a h e y th use r fo d edging of sheaths;' a short coil of silver wire ^| inch in length and -£$ inc diametern hi ; lastly, abou a dise hal f f bronzo on co tf e (fig, .19 No. 2), Irj- inch in diameter, ornamented with a series of four incised concentric circles. The iron objects recovered from this leve e fewear l numben i r r than from either of the lower levels. We are fortunate in having among them spear-heaa dcontrasn whicca e hw t with those belongin e earlieth o gt r occupations. It is shown in fig. 21, No. 1, and measures 4 inches in length. s readili t I y recognisabl a differen f o s ea t type longeo n , r leaf-shaped, t narrobu rounded wan d wit hmarkedla y split socket. This split socket characterises ha e spear-headdth s hitherto found fro uppee mth r levels, and is a feature of such weapons in use by the Saxons. A spear-ferrule (fig. 21, No. 2), measuring 5 inches in length, shows the same class of socket verA . ys foun simila wa thin o de s on lever 1915.n i l Fig, .21

have 1W e found piece f thio s s eac habove yearrula on 2 thin e s e o a ,(Proc., th s n leveo r o l 1. p. 116). * Proc., 1. p. 114, fig. 29, No. 3. 92 PROCEEDINGS OF THE SOCIETY, JANUARY 12, 1920.

No. 3, resembles the flat tang of a sword-blade, but it is greatly oxidised and its purpose uncertain. What appears to have been part of a hinge, 4f inches in length, is shown by fig. 21, No. 4. The object No. 5 of the same fig. has probably been a small movable handle for a vessel of some sort. Fig. 21, No. 6, which measures 5J inches in length following the curve, bears a strong resemblance to a half horse-shoe; it is, however, very thilightd nan . Halve f horse-shoeo s s have been founn o d this and the highest level in both of our previous excavations. Fig. 21, knivese ar , 8 e poin.d Nosth an Fro s amissing7 i t7 . . mNo . s bee.Iha t n

Fig. 21. Iron Objects from the second level, (i.)

of|the same triangularly bladeo tw de have t witth forw eme n s hi m a previous levels ,, measurin 8 whil . eNo g 3^f inche n lengthi s , thougha trifle smaller, is of the same as that represented by fig. 15, No. 6, from the level below. No. 9 is a curious little imperfect spoon-like object, with the remains of a tang, or handle. Its dimensions are actually l^z inch by \ inch. It is difficult to conceive what process it could have been employes capacitit s vero a s , ys yi din slight. Agai e nativnth e potter yremarko n call r t showI fo s . s even less variation than that found in 1915 and no fresh features. With regard to the Roman ware, as in 1915 there is a surprisingly large REPORT OF THE EXCAVATION ON TRAPRAIN LAW. 93 amount, though amon t gtheri absencn a s ei f freseo h typenotablf o r so e pieces followine Th . g list show principae sth l fragments:— Side fragment of a bowl of Samian ware, form No. 31; late second or third century. Base fragmen anothef o t r Samian bowl, form 18/37; good glaze. Base fragmen a Samia f o t n cup, for , showinm33 e lettef o on g A r potter's stam incise n para d f pan o t d circle; glaze brownis fumingy hb ; second century. Rim fragmen f largo t e Samian bowl, for wit, m37 h thicd k an bea m dri sides latf o ; e date. A small fragment of grey Upchurch ware, part of a globular beaker decorated wit hzona f rouletto e e hatching simila pieca o t re founn o d thire th d level. Small side uprighth e f fragmeno ) t "-shape"19 S . No t , (figd18 .sid f eo well-washed a bowre n i l d clay coated with pink slip, much rubbed off. Thi s potteri s kina f yo d mad t Sandfordea , Oxon Ashled an , y Railn i s Forese fourtw th Ne n i hte centuryth . Fragment of Castor ware, late second century. Fragment of imitation Samian ware bowl, form 31, with heavy bead rim in ordinary red tile clay. Small fragmen f Castoo t r -ware, piecbulboua f o e s beaker originally rough cast ^vith small bits of clay; coated with a clay slip red-brown merging to black; clay well washed, hard, light red in colour. Side fragmen cooking-poa f o t t scored with vertical lines; clay pale, nearly white, hard. Base of beaker of Castor ware in pale buff clay, coated with clay slip, red, merging into brown; fourth century. Small fragmen a cooking-po f o t 20). , No t Upchurc , (fig18 . h ware, clay pale grey with darker surface, decorated with loops scored wit hbluna t point, resemblin gpieca e from lowest level, foun 1914.n di 1 Part of neck of narrow-mouthed or bottle-necked vessel with everted rim ; fumed grey Upchurch ware. Rim fragment of a cooking-pot in pale buff, nearly white clay with dark core. Base fragments of a pear-shaped cooking-pot with beaded foot of Upchurch ware; clay hard, pale daub wit hdarkea r surface. A smal fragmenm lri Belgia f o t c cooking-pot, coated externally with polishehot d bitumen. A small fragmen f Rheniso t h ware decorated with white slip scrolls (fig. 18, No. 21). Thera somewha s i e t unsatisfactory minglin f periodo e g th n i s 1 Proc., xlix. p. 162, fig. 18, No. 3. 94 PROCEEDINGS OF THE SOCIETY, JANUARY 12, 1920. assigned e potterdateth f o s y fragments e probablTh . e caus f thieo s was the inequality of the depth of the soil lying between the levels on certain e areapartth f ,s o owinn outcroa e cono gt th f rock po -d an , sequent difficulty of determining exactly to which level any chance fragment actually belonged.

No coins wer ee seconfounth n o dd level this last seasonn i t bu ; general the few relics recovered correspond in character to those found same onth e leve 1915n i l n thaI . t yea obtainee w r coinso tw d , brasses of Constantine Junior (A.D. 317-340), and of Magnentius (A.D. 350-353); we may, therefore still regard the date of this occupation as having occurred in the fourth century.

HIGHESE TH T LEVEL. Comparativel numben i w y fe wers ra relic e eth s fro secone mth d level, those fro highese mth notable tar y fewer whed take an ; nw e into account the movements of soil, and the objects within it due to the presence of rabbits, it is doubtful if the number credited to the latest occupation should not be less than it is. Of fibulae, either bow-shaped or penannular, none were found. Two pins of bronze were recovered (fig. 22, Nos. 2 and 3); one (No. 2, complet, als2) . o No fig, e .23 though bent, measurin f incheg5 lengthn i s ; the other (No. 3, and fig. 23, No. 1), also apparently perfect, measuring 1-Jnr inch. The former terminates in a head which was designated zoomorphic in our last Report, where the style was discussed. This

example differs slightly from the two found formerly, in tha1 t the terminal portion is not bisected by a shallow groove on the front surface. Groups of lines have .been incised on the upper part of the stem, obviously to prevent the pin slipping too readily from its hold. The second pin is somewhat puzzling. It is short, and does not taper to a very fine point. The head is of the same zoomorphic pattern as the last, but from the square top there project two semicircular flanges t verse y close together, their opposing faces being practicall contactyn i . The e notyar , however, pierce woul e riveta on r ds dfo ,a have expected. The flanges obviously suggest attachmen o somt t e other objectt bu ; what that object was, or how the attachment was accomplished, remains unascertained. o ringTher tw f bronz(Noe a o se , eOn ar 7) . . e 7) (fig d , Nos.an 22 6 . simple ring the ends of which are not quite closed, calls for no remark. The other (No. 6), however, measuring 1 inch in diameter, has a moulded collar movin s circumferenceit n o g , indicating rather clearly that this . . ! Proc., l..p. 102. REPORT OF THE EXCAVATION ON TRAPRAIN LAW. 95

object has been part of a ring-headed pin such as was in vogue among the VikingSaxonse th laterd a t an ,sa date. Three dress-fasteners o calleds , f bronzeo , e wortar , h noticee On . simpla s i ) 8 e . squarNo , (fige22 . plate wit a loweha loo n po r planea , form we have already met with from the lower levels. It may be a survival presencs it r o , fortuitouse eb hery ema , seein gfora s thai m t i t

Fig. 22. General illustration of. Relics from the highest level (other than of Iron). (Ca. J.) which has hitherto been confined to the earlier occupations. The two other examplee morar ) e3 , als. inter10 soNo Nos , (figd , fig-22 an 23 .. .9 esting in that they represent a form new to us, and are, a pair, found, moute onth rabbia t f e a ho e firs e otheth t th thole n flooo d r an , r level below e disth .c f typeo The e y,ar differing fro e ordinarmth y styln ei having e s perfecshowa ,th y b nt example a segmenta, l expansion o n one side, increasing in breadth and ending abruptly at the top of the spirala plat e f beed o e looseacTh f ha eo s d nth . t i objechha en e f i s a t 96 PROCEEDING E SOCIETYTH F O S , JANUAR , 1920Y12 . been inset with three triangular beds of coloured enamel, red and white. The occurrence of these two identical objects in close proximity to each other at once suggests that, whatever their purpose, they were worn or used in pairs. From the highest level in 1914 came another of these articles e wit on ,n annula ha rs i head , , als4 . 11 o. . No fig, No Fig , .23 .22 a staple-shaped object of bronze with an extreme length of 1^-f inch breadtd an f inc f ho h betwee e outepointss s divideit nth i n t ro I .p du surface into a series of short segments by transverse grooves. Such an

Fig. Pins23 . , Dress-fastener d othean , r Object f Bronzo s e fro e highesmth t level. ($-.) object might conceivably have serve a chap dagger-sheatha s d a f eo e th , form being a usual one in such articles of Teutonic origin. Of mould portionso s thertw e e e ar heae on pina , th eacf n dI o r . hfo case (fig. 22, No. 4) the matrix is too much damaged to show what the actua l e othefor als, th s been 5 me matri. ha on r I th fig) .(figNo , .24 x .22 e hanth f do ha type s eon bee witr nfo h three pelletn spi similae th o t r e seconfounth n o d level. Of Roman glass there occurred a few small fragments (fig. 22, No. 12). s representewa t o segmentShalje tw r y eo b d f ringo s s (fig , Nos22 . 3 .1 d 14)an . Ther s alse tinei on oy pieca glas f o es armlee lighth f to t translucent variety (tig. 22, No. 15). e playing-meTh n number thre . 16)e No (fig. , 22 . REPORE EXCAVATIOTH F TO TRAPRAIN NO N LAW. 97 The following articles fashioned from stone were found:—a disc of sandstone wit hdiametea f lyo rf inch, carefully polishe botn do h face) 17 s. (figNo , .22 a large flat object of sandstone, rounded in shape, measuring 4^- inches in diameter by 7p" inc n thicknesshi e greateth ; r para f o t pointed oval block of sandstone with a con- cave surface on the upper face, seemingly produced by sharpening large metal blades, suc s swordha somewhaa ( s t similar stone secone founs th wa n do d leve s 1915i n i l d ,an illustrate e Reporth thar n di fo t t year);a 1 Fig. 24. Mould for a pin from the •wedge-shaped objec f sandstono t e grooved highest leveld relativan , e diagram. ({.) o ne side boton , n hmeasurino face d an s g 6| inche n lengthi s , If inc thicknessn i h , 4d J an inches n greatesi t breadt h; lastly e uppeth , r stona quer f eo n 19 inche diametern i s , wit ha socke t righa t t angles to the edge for a side handle, and a groove cut on the under side for a mill-rind 7 inches in length. The iron objectproportionatels a e sar w yfe as thos othef eo r materials. Ther brokea s ei n spear-head (fig. 25) of the leaf-shaped type with a closed socket, which presumably has been brough frop tu mlowea r level knifea ; , imperfect, and in two parts, with a broad double- edged blade, symmetrically pointed, closely resembling one found on the third2 level in 1914 ;3 a portion of another knife, single-edged and furnished wit htana fea g ; w naillastlyd an s; , the iron link of a chain in form of a figure 8 (show fign no . 25). Similar links were founn di a Saxon barrow on Lowbury Hill, Berkshire.4

The fragments of pottery found on this level are so few in number as to be almost negligible. Fig. 25. Spear-head and Link of Iron from the highest level, (i.) In regard to the native pottery, the whole amount might almoshande e palon helth e f . mb to n di Amon g these 1 Proc., 1. p. 130, fig. 39, No. 1.

The so-called second level of the 1914 Report was in reality the third. Proc., xlix. p. 187. 3

42 Atkinson, The Romano-British Site on Lowbury Hill, p. 51, pi. xv., No. 15. VOL. LIV. 7 98 PROCEEDINGS OF THE SOCIETY, JANUARY 12, 1920.

pieces there is none worthy of particular note. The Roman pottery is equally triflin quantityn gi e fragment bule th f Th ko . e ostensiblar s y from lowe re remaininlevelsth f O . gnoty e followinpieceema th e sw g as possibly being contemporaneous:— A rim fragment of a bowl or platter, form 18 (fig. 22, No. 19), with upright side, havin ggroova e \ inch belosuggeso t edge p wli th e te ofth a bead moulding; clay hard, brownish-red, late ware imitating Samian; third or fourth century. Two small fragment f blaco s k slip glazed Rhenish ware decorated with zones of scrolls and berries with raised centres, in white slip, similar thao t t thire founth dn d o f these)showleve 0 o 2 e . l son (fig.No , .22 e paucitTh f relicsyo , notabl e extremelyth y small quantit f potyo - sherds, indicates clearly that the occupation of this level was of the briefest duration. Traces of charcoal, moreover, were of the slightest. As to the date of this occupation the coin evidence of 1915 placed it somewhere in the neighbourhood of the year 400 A.D. Although no coins -wer eactuae founth n o ld floor leve 1919n i l e fou,th r small silver pieces which were buried with the silver hoard, to be dealt with hereafter, supply confirmatory evidence of that conclusion.

e havpassew W no erevie n i d e groupwth f relico s s foun eacn o d h profitable b y ma noto t et i level e d featurean , s peculia eacho t r . In the group from the lowest level it is natural that we should meet with relics fro e periodmth f culturo s e earlier than those generally represented ovee areth r a excavated e ston;th suce ar haxe s e fragmenth a bronzd f an o t e blade. Remarkable, howevere th s i , collectio f segmentno t braceletf shalje o s r eo s — twenty-fou alln i rt I . is evident that, as far as the inhabitants of Traprain Law were con- cerned, these objects passed out of fashion, probably towards the latte e seconr th hal f o dfa singl centuryt no e r examplfo , s comha e e e higheth f o r froy levelsman glancA . t tha e e r findrecorou sf o d of these armlet previouo thn i s tw e s years emphasises this pointr fo , •witexceptionw fe h s they have com levelo e tw f earlies froo se mth t occupation. It is difficult to supply any reason for this, other than that, as happens now, fashion decreed a change. There is not, however, y sucan h marked increas e numbeth n i ef piece o r f glaso s s bracelets from the higher levels as would suggest that a preference for glass was the cause. The numbe f whorlo r s froe lowesth m t leve s singularli l y small when contrasted with thos o leveletw frose th mabov , notwithit e - standing that all the evidence points to the lowest level having been under occupation, if not for a much longer time at least by a denser REPORT OF THE EXCAVATION ON TRAPRAIN LAW. 99 population e bul th f ;potter o k y fro e respectivmth e levels s showa , n hereafter, being evidence of this. The true inference to be drawn from thi t quits no fac s ei t obvious markeo s t increasn bu ;da e th n ei number of spinning whorls found in later strata must surely imply an increas e amounth n i ef spinnin o t g carriey meanb n f spindlo o ds e whord an l oveparticulae th r r area. When that increase coincidess a , it does in this case, with a diminution in the number of other relics, the fact becomes more significant. The cause may have been a freer supply of raw material, wool, or flax—and one probable result an increase in the amount of textile materials woven by the inhabitants. But the case of the whorls does not stand alone in suggesting some chang f fashioe o emateria th n i nf clothino l g having occurred after r e earliesperioou th f do t occupation e occurrencTh e .bronz th f o ee and iron pins such as would be used rather in textile fabrics than in hides appears to point to the same conclusion. In 1915 we recovered four pins of bronze and the head of a mould for casting a pin, all from e thirth d level; while this yea have w r e pins from each .level except e earliestth d portion an o les, n sf o stha clax nsi y mould r castinfo s g such objects from similar sources. The third-level relics sho e emergencwth f threo e e fresh kindf o s objects—clay moulds, t shalrings je smald r eo an , l playing-men 1915n I . ' fragmente th l al f mouldso s e complet wels th a , s a l e moulds, belonged thire th r highe do o t r s levelsevidenti t I . , thereforet ar e , th tha n i t of casting a great advance seemingly coincided with the increase in spinning. As well as moulds, pieces of crucibles are found in considerable number e samth t ringsen je o s levels e , Th eithe. r flar roundeo t d t n versectioni no ye numerousar , t thebu , y made their appearance last seaso n leveli n s above tha whicn i t e braceletth h s were found. The playing-men are not represented in the relics from the lowest level. Their first appearance is on the third, and they increase in numbe e secondth n quotatioi rA . n fro e 191mth 5 Repor applicabls i t e here e relic :e th bul "f sTh o k have com loweso e tw fro e t mth levels , e cas th f thes eo n i et playing-mebu e ratinth s reversed—fooi e th f o r twenty-one found o camtw , e e lates froe th leve mth f to l occupation, ten from the level below it, six from the third, and only two from e lowesth t stratum,e unaccounteon e th "r havinfo d g been found while filling in. Now possibly the presence and distribution of these playing-me affory nma n indicatio a d e sourcth f f otheno eo r influences. The findin muco s f o hg Roman potter d glasa sitwhic n yan n o e o s h generae th l aspec e relicessentiallo s th f s i so t y Celtic clearly shows that contact wit e Romath h n invaders, durin e perioth g f theio d r occupatio Scotlann ni subsequend dan t thereto s responsiblwa , introe th r -efo 100 PROCEEDINGS OF THE SOCIETY, JANUARY 12, 1920.

duction of many new commodities into the domestic economy of the native Celts. The discovery of a checker-board cut on stone at Corbridge, as well as the finding on sites of Roman occupation of numerous small glass discoid playing-pieces, indicate thae Romath t n soldiery wiled away their leisure moments with some game played on a chequered board, and seemingly they introduced it to the native inhabitants r lasou te presenc n Reporth I . w pointee f w ho o et t ou d late Roman potter d glasyan s relativ o periodt e s long subsequeno t t the withdrawal of the legions from the north was indisputable evidence activn oa f e fact tradeth sd broughan , pase ligho th t t t n i summet r further justify that conclusion.

. NativFig26 . e Pottery from each level discn foo1 o , f to s diameter.

I have referred above to the evidence of duration of occupation of varioue th s levelgatheree b o st d fro e amounmth f pottero t y recovered from e 191 themth 5n I Repor. n attempa t mads wa tdemonstrato et e thi statine numbery sb th weighte d native gth an piecef th so t f o sepo s represented of the Roman wares. In this Report another method is employed. Fig show6 2 . e totasth l amoun f nativo t e pottery recovered from each of the four levels respectively, piled within circles of one- foot diameter. It will be seen at a glance that the pile from the lowest level, shown on the extreme left, is preponderately greater than that from any other; in fact, it exceeds the total from all the others. Similarly also the other relics from this level are more numerous. From the third level, the pile next to the right, the amount is unaccountabl yproportioe smallth t bu , n whic largee t bearhi th o t sr vers i y simila similae a cas th n wha ei o s t r rwa t arrangement madf eo the pottery obtained only from area G, the first half of the ground excavated last summer e seconTh . d a pilleve s e ha lbarel y half the* height of the third; similarly the-illustration of the relics from this level REPORE EXCAVATIOTH F O T TRAPRAIN O N N LAW1 10 .

show proportiosa f onlno y about hal f thoso f e fro e thirdmth . When w e extrem looth trifline t th ka e eg se righ quantitd an t y obtainen do the level of the latest occupation, and make allowance for pieces of pot brought up from below by disturbing agencies, we realise either that e inhabitantsth y b e r us thao , n i t t thepotterno y s barelywa y lived for any length of time on the site. Let us now look at fig. 27, which shows the Roman pottery identically set up "within circles of one-foot diameter. From e lowesth e figureth te have lef f levelth w o ,t en o , a considerable pile, showing a ready disposition on the part of the Celtic population to acquire the Roman wares. The next pile on the right bears a higher proportion to the pile from the earlier level than

Fig . Roma27 . n Pottery from each level discn o , s fooo1 f t diameter. wae casth s e wit e nativ th ht greatl no e s i poty d les,an s thae nth corresponding pile of the latter, indicating the increase in Roman influence. The next pile to the right, that of the second level, actually exceeds in bulk the corresponding pile of native pot as well as that of all similar pottery from level number three. Bearing in mind the paucit f relicyo s from this level compared with those fro thirde mth , the increase of Roman ware is remarkable. As for the Roman pottery from the highest level, if we remove the pieces of Samian ware, which are those on the left side of the circle, and all of which probably owe their presence on this late level to disturbance, we are left with some six or seven fragments, for only one of which there can be clearly claimed a thir r fourtdo h century origin eved an ,n that might quite well have been deposited originall e leve th r thi fa n l so ybelow w methoHo . d of indicating the duration of occupation on the respective floors is accurate s difficule othei th t i , s rsayo a t tfactort ;bu e numbeth , f o r relics, shows somewhat similar proportions, it must be approximately correct e factOn , . however e los,b mus tt sigh no . t Thougof t e th h 102 PROCEEDING E SOCIETYTH F O S , JANUAR , 1920Y12 .

area explored on each level is necessarily identical, it does not follow that at each of the periods represented the occupation extended over whole th e area thar t eacno ,a t h equalls periowa t di y dense.

Thus far I have dealt with what I may call the normal results of the season's excavation. There were, however, as is now well known, results which were so remarkable as to be regarded as purely abnormal. I refer, of course, to the find of Roman silver plate. The circumstances of such an historical event deserve to be placed fully on record. On Saturday, 10th May, I paid my weekly visit to the excavation, and, finding the whole of the second level floor on area G fully exposed, plotted such remains of foundations as were there, in order to enable e workmeth proceeo nt e followin th n do g Monday \vit e removahth f o l the exposed surface. Though the whole floor came under observation in procese th planningf o s , nothing attracted attentio spoy causo an t t t na e a suspicion of its having been previously disturbed. On Monday, as the foreman was gently breaking the surface with his pick, preparatory to the soil being passed through the riddle, in the south-west corner of G, a spo t a t marked wite asteriskn th ha , e plan 2 th f leveld n o s o an , 1 s s toopoinhi lf o camt contacn i e t with some substance strange th o t e touch. Puzzled, he gently inserted the pick again, and to his amazement e pointth brough n -a metao . up tl bowl wit ha beade d border (fig. 40). Further exploration with a knife revealed the fact that the bowl was not the only relic, but that there was apparently a pit of unknown depth ful f remarkablo l e objectf theso w f emetal o sfe wer A e. taken out— other bowls with beaded borders, and the wine-cup (fig. 35) which lay near the surface. The day's work over, such objects as had been unearthed were carefully concealed in the hut; the pit, still containing an unknown amoun f treasureo t s covered lefwa an ,t p wit du littlo n h e anxiety for the night, while one of the staff was despatched to East Linton to telephone a message that would bring me out on the following day. Wisely desirin keeo t ge discover pth absolutn ya e s secretwa t i , impossible to be explicit over the telephone; accordingly the message f sufficieno that no t s reachet urgencwa e dm induco yo cancet t e em l all engagements and proceed at once to the hill. It was four o'clock on Tuesday befor ereacheI wore spote beed th Th k.ha n proceeded with unde r Pringle'M r s directio mannea n ni r deservin l praiseal e f gTh o . temptation to clear out the whole hoard had been resisted. The limits of the cache having been located, the soil for a reasonable radius all round had been removed, and a sufficient amount of treasure was left still in situ to show me exactly the condition in which it was exposed. The sight that met my gaze on reaching the ground was one that the REPORE EXCAVATIOTH F TO TRAPRAIN NO N LAW3 10 . most sanguine of excavators can hardly ever have dreamed of. Bowls, cups, spoons a miscellaneou d an , s collectio f pieceno f plateo s , tarnished and soiled, but obviously of silver, lay spread on the ground. The immediate realisatio e notablth f o n e characte e finth d f o madr e th e momen neveforgottene e b on t o rt silvere Th . , stil situ,n i lpartiall y la y d partiallan f o yp betweeto e olargo th n tw n e stones t I appeare. d to be imbedded in a purple paste, to such an extent was the soil dis- coloured by the decay of the metal with which it came in contact. In some caserotte silveso sthe dhad r away tha crumbleit t d when touched, though wit t i hardene hr ai exposur e dth somewhat.o t e Apparently e meta th d becomha l e affecte sulphuy e soilb dth , n i duer o doubt,n , to animal matter which had decayed in it in the course of occupation, and additional evidence of this chemical change was forthcoming in the offensive odour -which was emitted from the hoard for many week s sbrough aftewa t i r t indoors n immediatA . e examinatiof o n e circumstanceth e finth d f o sshowe d tha hola t d bee frog ha endu m some surface at a higher level than the second, for—as was shown by the facts of the actual discovery—the relics at the top of the cache were barely covered by the soil that formed the floor of that occupation. Reference levep to le plae madth f th n o t o cleaei t r that durin perioe e gth th f do latest occupation there was no foundation over the top of the spot -which would have show deposina t during that perio havo dt e been improbable. Fortunately when the treasure was unearthed two coins were found, and in the washing of the soil which came from the pieces when the rough dirt was removed from them, two more were discovered. These four coins, small silver pieces, belonged, one to the reign of the Emperor Valens (364-378), one to the reign of Valentinian II. (375-392), and two to that of Honorius (395-423). Their condition showed that thed beeyha n subjected to little -wear by circulation before the date of the deposit. To assign such a date to the early years of the fifth century was therefor t unreasonableno e o coinN . s were recovered definitely from the top level this last summer; but in 1915 three small brass pieces, believed to be of the reigns of Constantine Junior, Valentinian, and Arcadius, seeme s occupatio e datit poino dt th f eo o t t s havinna g been early in the fifth century. The date previously assigned to the highest level therefore coinciding so nearly with the indicated date of the cache, there is every probability thae latte mads th t wa r e durin e perioearln gth a f yo d fifth-century inhabitation of the site. One foot of soil at this spot lay between the present surface and the highest floor level, while 10 inches separated that fro secone mth d level f thiI . s assumptio s correctni , then whee nth deposit was made a hole about 1^ feet wide was dug just outside the 104 PROCEEDING E SOCIETYTH F O S , JANUAR , 1920Y12 . south-west end of the building which then stood on the surface, for a depth of 10 inches to the floor level below and a matter of \\ feet further down till the presence of two large stones, which actually lay on the third leve r juso l t abov , stoppeit e y furthedan r progress. Thereupoe th n silver somf o , t eithe esac a ou receptacle r n ki ro deposites wa , d with little inche1 1 t somr o e smasbu ceremon 0 th e1 y f sla o hole p th eto n y i tile lth belo surfacee wth onld an y,bara e beneat o incs r ho h wha beed e ha tnth surfac thire th df e o occupatio e seconth n no d level. traco n Ther s e ewa whatever of any sack or chest, but the manner in which the objects lay suggested rather that they had been hurriedly thrust anyhow into the

Fig. 28. General view of the Hoard of Silver as found.

hole sees conditione b It n.y fro ma generae ms th a , l illustratioe th f no hoar dlefdoub (figno t28) . t tha mas the t silve of sforme had r d parof t some great spoil. Bowls and dishes finely decorated were represented by fragments; flagons were crushed inte smallesth o t possible compass, handles were twisted and wrenched from spoons, and cups were torn from their stem basesd largan s A . e numbe piecef ro s shape werth n eei of packets which had been folded over several times and hammered flat. It was evident that the artistic qualities of the plate had been held accouno on f t whateve d handleha thosy o b rt previoui dwh e s it o t s disposal in the hole. The total weight of the treasure recovered is some 77. troy0oz , s mananda , y large vessel e representear s y onldb y small fragments, the original weight must have been several times as much. The first idea that presented itself was that here we had loot obtained from the sack of some religious establishment. To give weight to this, there are a number of pieces of plate bearing Christian emblems, though the mere fact of such emblems being placed on any vessel does not REPORE EXCAVATIOTH F O T TRAPRAIN O N N LAW5 10 .

necessarily impl ybeed thaha n t i tconsecrate servicee th r dreligionfo f so . But the shape and designs of certain of the pieces bear so markedly a

Fig. 29. Spoons. religious character that it is reasonable to assume they have played a part in church ritual. firse th t n placeI , fig. 29 e extremshowth 1 s01 e righ Romaa t n spoon

othee 1Th r spoonl s show thin no s plat describee ear . 115p n .do 106 PROCEEDINGS OF THE SOCIETY, JANUARY 12, 1920.

of a well-known type with a long handle terminating in a sharp point n ovaa ld bowle sharplan Th . y pointed handl originalls ewa y intended to assist in the eating of shell-fish in the manner in which we employ a fork at the present day. A pierced scroll connected the bowl to the handle, but unfortunately the two parts have been torn apart. In the centre of the bowl, so deeply cut as to have actually broken through the pointsacree e bac e th on s born s th a t i ,k d a n o monogramo e Rh i Ch e th , standar Constantinef do bace s Th decorateki . d wit hseriea f shalloso w flutes radiating from a central line. The total length of the spoon and handle when joine s bee ha df inches n8 e presenc e Th christh . f mo e in this case n i seem e o poinus t s o t t the a°dministration of the Holy Com- munion. Though to this day in the Eastern Church the administration of the Communion in both kinds is carried y mean b a spoon t f o sou ,t therno s i e e westdocumentary ith n an y evidence to show that here alse sacreoth d rite was so performed. While Cabrol in his Dictionary states that it is always possible that this eucharistic practice did exist exceptionall e west th suge h n ,i y - gests, from references associating the spoon wit e patenth h , tha e formeth t r havy ma e been employe placinr dfo e gth small eucharistic loaves on the paten. There can be no doubt, I think, that the object shown in fig. 30 is that termed StrainerA Fig . 30 . . in early liturgical literatur ecolatoriuma or coluni. It is a circular bowl, measuring 2 inches in diameter, which has originally had a handle, of which the base only remains. To serve its purpose, which was to purify by straining the wine to be consecrated for communion, it is perforated at the bottom and around the edge. e perforationTh o monograme bottoth Rh i n i ms Ch for e , mth whil e those around the edge furnish the legend IESUS CHRISTUS. Strainers for wine \ver e eGreek th use d Romany b dan s n ordinari s y lifed an , after the introduction of Christianity were adopted by the Church. One is mentioned in the inventory, dated 471, known as the Carta Cornutiana, and a few are enumerated in the Liber Pontificalis. 1 Equally unambiguous as to its use seems the small flagon, badly broken, two views of the body of which are shown in figs. 31 and 32. 1 Atchley, Ordo Bomanus Primus,. 25 . p REPORT OF THE EXCAVATION ON TRAPRAIN LAW. 107 The e heighvesseth f o lt when complet s beeha e n about 8£ inches. The mouth is circular, with an inverted rim, after the manner of many flagons and vases in late Roman times. Below the mouth, with a short intervening neck, is a bulbous expansion covered with leaf ornament. Betweee th n d thian s body of the vessel, on the portion that form shouldere th s s i , a zonal panel filled with animals and trees. Beneath this panel, bounded abova borde y b ef o r triangular bosses rising from a foliaceous device d beneatan , h b ywell-definea d rope moulding, a zon s i e 2£ inches deep, cover- greatee th g in rbode parth yf o t of the vessel. On this there are executed in high relief in repousse certain momentous inci- dents from Holy Writ. In graceful pose, and with delinea- tion and technique recalling the art of an earlier period, Adam, with thumb and forefinger, is shown plucking the forbidden fruit, while the serpent, wind- ing round the tree trunk, has s heahi d extended towarde th s figur f Eveo e . Wit s bachhi k to Adam is Moses, clad in flow- g robesin a ,dignifie d figure, with his right arm outstretched strikin e rocth g k from which 1" watee th r gushes into cups held up by two small figures repre- Fig. 31. Fragment of a small Flagon showing "The sentin e Childregth f Israelo n . Fall of Man" and "Moses striking the Rock." Wit a hgrou p intervening, probably intende r Lord'ou r sfo dbetraya l by Judas, we next have the Adoration of the Magi. The Virgin, seate a chai n o dr wit a hhig h curved back r feehe ,t restina n o g stool, hold e infanth s t r armse child'Jesuhe Th n i .ss arm e outar s - stretche o receivt d e giftth e s e profferinwhicar e wisn th h n me eo g round dishes. The Virgin appears to wear a head-dress. Her gar- men s loosha t e hanging sleevese edg th f whic o et a , a doublh e line 108 PROCEEDING E SOCIETYTH E O S , JANUAR , 1920Y12 . of dots apparently signifies embroidery. The chair on which she is seated is of the class described by Pliny as fashioned of plaited osiers. The sarcophagus of St. John the Baptist at Ravenna shows a similar representatio e Adorationth f o n , wit e Virgith h n seate a basket n do - chair, resting her feet on a stool, while the scene is also depicted on other Christian sarcophagi of the fourth century.1 A circle of heavy sphere r beadso s encircle e basth s e of the flagon, while around the space betwee e ropen th the -d man \ moulding twists a vine from which bunche f grapeo s s sprin altern gi - nate directions. The next piece which beara s Christia npeculiaf o symbo e on rs i l interest, as it may possibly yet afford a clue to the provenance of a smale hoards i th t l I flas. k (fig. 33), now flattened by the ill-usage to which it has been subjected, wit moute hth h everte parn i d t dan broken off. Bammed int nece oth k s whai t appeara strap-en e b o t s d or buckle showing a fragment, still remaining, of the leather strap to s beewhicha n t i hattached e Th . heighe vesseth f s By^i o tl - inches. At intervals around it occur zones of gilding. At the base of the neck, lettern i s abou inc- | t h high, formed y smalb l punctulations- in n a s i , scription, legible without any diffi- Fig. 32. Fragment of a small Flagon showin culty runnind an , g from either side "The Adoration of the Magi." of a Chi Rho monogram flanked by Alpha Omega'.d an The inscription reads PRYMIACOEISIAPI. There are no stops, no spaces left between the letters, and no indication of con- traction. Numerous attempts have been made to find the nature and meanin f thigo s inscription. Professor Haverfield suggested that some reference to the Abbey of Priim in the Western Eifel was implied in a word PRYMIACO, adjectives terminating in -iacus having been not un- commo e fourtth n ni h centur Celtin yi c lands, while Professor Sayce sees in it a reference to the corrupt worship of Isis and Apis which had

Cabrol, Diet. Arch. Chretienne, p. 26, s.v. Chaire cpiscopale. 1 REPORT OF THE EXCAVATION ON TRAPRAIN LAW. 109

crept from Egypt intearle oth y Christia uncomn a n r Churchfo -t Bu . promising E the Avord " IACO(B)I" could be read after PR YM, supplying- another solution t unfortunatelbu ; r thiyfo s suggestio s e "a E nth s "i carefull s clearla d yyan othey formean rs a dlette then i r ' inscription. A likely suggestion, emanating from another source, is PR.YMIACO EISIAPI, admitting of the translation " to Prymiacus the son of beo t = t ascertaineye s Eisiapus"ha t i d t whethebu ; e latteth r r name

. SmalFig33 . l inscribed Flask.

wa e fourt a possiblsth n Gau i n e i hl on ecentury . Various other- suggestions made need not be mentioned here, so we must leave the solution undetermined in the meantime. More doubtful in their Christian attribution are three portions of the ornamented border of a large circular plate, two pieces of which e showe edgar Th s enrichei en figi n. .34 d wit n ovad a dish an lc moulding within whic a finel s i hy designed border whic s beeha h n inlaid with niello. The design is not uniform, but consists of alternate lengths of different pattern—one made up of composite hexagonal figures aroun dsquara e e othercentreth d , an ,forme fouy b d r curving- r

110 PROCEEDING E SOCIETYTH F O S , JANUAR , 1920Y12 .

lines intersecting each other at regular intervals and tracing a series of ovals filled with conventional leaf ornament. Separating the designs on the larger fragment illustrated, is a circle containing an equal-armed cross with floriated endsprobabilite Th . s thai y t this crosa mer s i s e

^ *y&: /^%iPi§w'm • ^-^'

Fig . Portion 34 .e Edg larga th f f eo so e Platter. conventional ornament with no religious significance, but the piece of plate on which it appears has been grouped with those which show undoubted Christian emblems e possibilitth n e cono b , -y y ma tha t i t nected with them. The platter 'of which these formed parts was of large size segmente th , s indicatin gdiametea f 19o r | inches. There are at least four pieces which show designs as character- istically e foregoinpagath s na Christiane ar g . REPORT OF THE EXCAVATION ON TRAPRAIN LAW. Ill A large flat dish with onl ya sligh t curve upwards fro e bottomth m to the edge is, though folded up, complete. In the centre is a circular disc from which radiate to the edge deep flutes terminating in scallops and alternating with flat surfaces. The latter are engraved with conventional leaf designs e bottoe dishth th n f ,mO o similarl. y engraved, there appear a figurs f Amphitrito e e seate a panther-heade n o d d sea- moiister. Attache o semi-ovoie sidy soldetw on b e e o t ar dr d bosses sharply pointe e loweth o rdt end terminatind an , n e uppei th d t en gra e necd a heaswan-likth f an ko d e bird e nec brino th t , k s a beng o s t the head against the boss and form a loop. Through this loop there probably passed a ring or chain after the manner of other scutcheon handles know Celtin no d Saxo an c ne othe bowlsth n r O sid. f thio e s dish the remains of cement show where the two other handles have been attached whicf o e t leash,a on amons i t miscellaneoue gth s relics. Among the Roman relics in Mrs Cripps' Museum at Cirencester there are two handle thif so s f bronze typeo f the o t e m,bu on , still retaining metad lan cement similar in appearance to that in the Traprain specimens. A peculiar featur f theseo e Roman handle tha s si e neck th t , which projects fro uppee mth e rboss parth f , o tturn s outward formd an s looa s p with the boss itself, while in the somewhat similar Celtic scutcheons the loop is forme turniny db nece e vesselgth e edgkth th n f inwar I o eo . t n do Museue th Philosophicae th f mo l Societ t Yory a bronza s ki e bowl found t Castleyardsa , York 1829n i , Anglian,e b sai s furnishei o dt t I . d with three handles, much resemblin e for th f gthoso m e attached e tth o Traprain dish but with the loop formed against the edge, as in the Celtic examples examination A . materiae th f no l wTith which these bossee ar s filled shows it to be a combination of lead and tin which has been covere exposee th n do d surface wit hcemena whico t e soldehth r would be applied, as otherwise the soft metal within the boss would inevitably have melted when the soldering bolt was used. The lead and tin, or realitn pewtei t i serve, ys is a re purpos dth f solidifyino e e bossth g o t , surface th f whiceo t woulhi d adher manner'thaa n ei t lead alone would not. A folded-up fragment, of some object at present indeterminate, is ornamented in repousse work with a figure of Pan. On the ground betwee pipess hi fees e nhi ,li t while lightly resting agains e fingerth t f so his left hand, with the crook over his arm, is his pedum or shepherd's crook figure Th representes .ei d nud modelled ean d with great mastery. Dancing figures with flowing draperies accompany Pan. A portion of the bottom of a bowl, itself broken across, is chased with a figur f Venuo e s rising fro e waves mgoddese th Th . s representei s d front view, her arms upraised holding up her long lank locks, which hang fro mstrangela y conical head, possibly surmounted wit hpointea d cap. 112 PROCEEDINGS OF THE SOCIETY, JANUARY 12, 1920.

Roun r nec doubla dhe s ki e strin f beadsgo , whil weare esh s armlets sa wel bracelets a l s seemingl same th f e y o backgroune nature th n I . a s di d sea-shellsfisan h e medallioTh . n whic s formeha h e bottoth d s mi encircled by a conventionalised scroll border, probably derived from a vin resolves leavesd ha ean t I . d itself int oseriea rathef so r solid volutes branching fro ma curvin e th gf o steme surface sted th Th m an f . o e volute markes i s smaly b d l punctulations desige Th f thi. no s bordes i r to be seen on a number .of objects from the cemetery of Piispok-Szent in Hungary and other cemeteries in that region. The art displayed on this

piece present1 markesa d contraso t t that which appeare mosn th o s f o t other ornamented fragmentse Th . figure of Venus is stiff and formal,' recalling the conventionalised form of some Hindoo goddess. Whatever its provenance, it certainly never issued from the same atelier as the piece with Pan or the flagon with scenes from Scripture. e piece th e bul f Th o sk beao n r emblem r ornameno s o associatt t e them definitely with either Christian or pagan worship, but, being portions of dishes, platters, bowls, or flagons, might quite possibly have serve purposedthe eithersof don,or e Fig. 35. Wine-cup. duty as table plate in a purely secular establishment. ' The wine-cup (fig. 35) is shown as if complete, though the foot is detache s d doubleheighIt an d . s 4i ^tup dinches , diamete f bowo r l 3|4 inches, diameter of base 4J inches. The broad foot, a little exceed- ing in diameter the mouth of the bowl, gave the cup a stability which modern drinking-vessel o ofteto s n lack from disregar f theso d e proportions. In addition to this cup, there are the remains of at least three others: in the case of one the foot alone is lacking, of another there is a detached bowl and a stem, while a broken portion of a crushed bowl attached to another stem are portions of a fourth cup. Such cups might well have been used for administration of the. wine in the Communion, it.bein practice gth earle th yn ei Churc transfeo ht precioue th r s liquid from the larger chalice to such smaller vessels. On the other hand, they may have been simply wine-cups use somn di e Roman villa. They bear Hampele Se ' , Ungarische Alterthiimer, . 329p . . ii REPORT OF THE EXCAVATION ON TRAPRAIN LAW. 113

no emblems, and there is nothing in their form to restrict their use to sacred ceremonial. A large flagon is represented by three pieces, two of which are e piece t On illustratedshowno , . n figo n36 . , consist f tho s e mouth and the neck, on the lower part of which is a knop or bulb wreathed in laurels. The mouth is so crushed that its form is uncertain. The second fragmen figure)e righe th th f n o t ,(o t from lower dow e vesselnth , s enrichei d with a broad ban f ornamento d , formee intersectioth y b d n oa serief f circleso s e spaceth , formeo s d being filled with leaf orna-

Fragmento Tw . Fig Flagon—panea 36 .f so l gilt. ment. Certain portions have been gilt, while all the rest of the design has been filled in with niello. The third fragment, which is the largest, shows two broad panels of geometric design, with shallow gilded mouldings above and below each. The upper one is interrupted at intervals by circular spaces, each containing a representation, gilt, oa wingef d cherub holdin a gbaske f fruito t , whil e loweth e r panel include t shortea s r interval a sserie f smalo s l medallions with human heads, also gilt. Another such vessel is in a more perfect state. The bottom has been torn off and the body and mouth crushed. The height of the flagon in its present condition is 10 inches. The mouth has been too much crushed enablo t s exaceit te determineb for o mt e meantime th n i d . Around the neck occurs a broad band bordered at top and bottom by a line of scallops between whic a ric s hi ornament derived from some trailing vol.. LIV. 8 114. PROCEEDINGS OF THE SOCIETY, JANUARY 12, 1920. plant, probabl e e backgrounvineyth Th . d betwee e scallopth n d an s around the plant motive has been inlaid with niello. At the junction e necoth f k wit bode deepla hth s yi y projecting moulding e roundeth , d edg f whiceo engraves hi d with feather ornament. Below thibode sth y e vesseoth f l slopes outwar e bulgth o edt with flat panels, each alter- nately filled with an engraved leaf design formed by leaflets springing from either bd centraa sidy repetitiona f an eo b ri l smallea n ,o r scale, of the plant design around the neck, with, as in that case, a background of niello. Fro vessee bulge mth th f eo l wher panele eth s ceas bode eth y is continued plain to the base.

Pig. 37. (1) Toilet Box, and (2) Handle. T Fig. 37, No. 1, is a cylindrical box with a convex moulding T F inch below whicn resteds 3s o i | lowee ha inchep t hth p I to . ca re n spari a th f o t height larga n I e. silver casket e Esquilin, parth f o t e Treasure th n ei British Museum e fouar , r almost identical boxe r poto s s placed symmetrically aroun dcentraa l bottle. This caske regardes i t lady'a s da s dressing-case, and the boxes are supposed to have been used for holding pomades and cosmetics. An object, obviously a handle, is also illustrated in fig. 37. It is 5J inches in length, and has a knob at the foot and another betwee o hollotw n w moulding e stemth n , o swit a hsimila r mouldin e uppeth t ra g end l thesAl . e mouldings have been giltA . p indicateto e flangth s t a etha e socke th ts hel ha td some spreading arrangement, and it has been suggested that this may be the handle of a flabarum, or fan. REPORE EXCAVATIOTH F O T TRAPRAIN NO N LAW. 115 A crushed-in conical object terminating in a fluted knob has been a cover or lid. Within it was caught the bowl of a spoon which was found to have engraved on it the Chi Rho monogram. Three other spoons (fig. 38) are of a different type from those mentioned above. Two have pointed oval bowls. The handle of each, which is somewhat slende d roundedan r , terminate a bird' n i s s e heade spooth Th n .o n left of the illustration has a swan's head with, apparently, a fruit in bills it . Originall turnebirs e heae th d wa yth f do d rounds thao wa s , t ti

Fig. 38. Spoons. parallel in direction to the central line of the spoon and the stem had one bend in it. A spoon of similar form, but of bronze, with a debased bir da doubl head an de steme th loo s illustratei n , o p Pilloy,y db s a

t founQuentinS t a d t I bear. s alon e bow e th glegen th l d PON1 E CURIOSE, i.e. "Handle carefully. similaA " r spoon t witbene bu , hon d steme th s relatee n sami ,i th ey db authorit havo yt e been founa n i d fourth-century cemetery at Spontin, and to have been in the Museum at Namur. A silver spoon of the same class bearing along the blade VOT. FECIT LETU s founSwa d a fourth-centurals n i o y cemetert ya Vermand. 1 Pilloy, Etudes, vol. ii, p 209. 116 PROCEEDINGS OF THE SOCIErY, JANUARY 12, 1920. The spoon on the right is altogether ruder in execution. The bird's head seems so debased as to be barely recognisable; the stem, which is of the same slender character as that of the other, is longer, and has, l probabilityinal doubla d ,ha e twistcentrae Th . l illustrae spooth n ni - t anotheye f tioo s rni type lengts It . 5s hi ^ inches bowle Th ., witha diameter of 2^V inches, is circular and has been spun on a lathe. In the concentrio e centrbottotw th f e o emar c circlese handle th n Th i . s ei fora dolphi f mo n holdin e spoos mouthe edgit th gth f n ni eo . There is a curve in its body, admirably contrived for the thumb if held in the right hand, while the flat tail would rest on the lower joint of the first e treasur fingerth e Britis th n I n ei . h Museum e ,Hil founth f o l n o d St Louis, Carthage, and belonging to the fourth or fifth century, there are a number of spoons with deep circular bowls similar to that above describe t witdbu h different handles. Amon smallee gth r objects found wer ea pai f handle ro fore th m n si of leopard , fro e or absencs mth f definiteo e spottin onen go leopara , d and a panther (fig. 39). These objects are represented as supporting themselves in an upright position with their fore paws gripping a segmental bar, which has evidently been attached to the curving side or edge of some vessel, while their feet have pressed against the base. Though both animal e beautifullar s y modellede th e leff , th o that n o t illustration is more highly finished than the other. The panther, if such e rightth n ,o doe, t appeabe no s t i havo t r e bee ne sammadth ey b e e otherhanth s da ; the modellin e animalth f especialld go an ,e th f yo hinde goodo s wherd t partface an , no nth lookes i es i , t frontwisda t i e will be observed that the lower jaw is not in the centre. The best-known sucof h use animaanalog the l yfor figure handlebowlas s the sin are s the treasure found at Petrossa in Roumania early in the last century. Each of these bowls or baskets, as they might well be called, had leopard handles at either side, the animals fastened to arms which project from the edge of the actual vessel. The art is later and the modelling incomparably poorer tha r nspecimens ou e thaCair f th o t on I . Museum there are two tankards of bronze, each with a single handle, e forlioa th f nm o n i with turned-iir outstretchetaid an l dr headfa o S . t beeno ther ns identifieha e d amon e Trapraigth n treasur vessely ean , r vesselpara o f o twhic o t , h these handles see havmo t e been attached. cannoWe t tell, therefore, whether they have belonge singla to d or e o separatt e vessels n whice curveo Th r h. ba d their respective fore paw f identicaso rese ar t l curve, but e illustras wil a th e ,see b ln i n - tion, e blockth s against which theire hinth t da pawe li t s presno o d s same angles. The smaller object on the upper part of fig. 39 appears to have been REPORE EXCAVATIOTH F TO TRAPRAIN NO N LAW7 11 .

projectine a foot Th whicn .o e supporter hth gba d article restes dha been bent downwards. finesthe One tof piecetreasurthe of ssomewhaa eis t shallow bowl,

Pig. 39. Foot of Stand and two Handles of Vessels. mutilate crushedd dan , wit hdrawa n wir efooneaa m edge d ri tth r ean outer circumference of the bottom. Around the side, between a species of ovolo border above and a border of dentelles below, is a broad frieze filled with figure f animalo s d occasionaan s l human mask n reliefi s . 118 PROCEEDINGS OF THE SOCIETY, JANUARY 12, 1920.

The animals, which are depicted with great vigour, are shown in savage pursuit of one another. Amon e fragmentgth f plato s e thersurprisina e ear g number which

Fig. 40. Broad-rimmed Bowl.

Fig. 41. Triangular Bowl. e decoratear d alon e edg f larggth o e w ewit ro bead ha s hammerep du from beneath, and we are fortunate in possessing six small bowls with this character of decoration in an almost complete state. That shown REPORE EXCAVATIOTH F TO TRAPRAIN NO N LAW. 119 n figi measure0 .4 inche6 s n diametei s f inche2 d n heighti san r s A . previously stated, thi firse s th bow t s objecwa l t recovered fro cache mth e and shows on its lower side the indentation caused by the pick. Fig. 41 is a small triangular dish measuring 4^ inches along each side 2j-d \an inche heightn i s . Lik e circulaeth r bowls above referret i , to d heava s ha y bead edgingbeade th thit n i s bu ,s case have been cast solid objectw fe e t shammere th no recoveref o d e an . a Thi on dup n d i s s i perfect sta,te. A variation from the simple flat rim of the beaded bowls is shown by e fragmenth s ornamentei t figm . ri Her .42 e eth d wit hhumaa n head between a grazing horse and a hound with its head to the ground. The modelling and execution of these animal figures are very vigorous, and

1" Fig. 12. f FragmenShalloo m Ri w f o Disht . bowe th •whicf o l he onl thith ys si surviving fragmen s bee ha f higt no h artistic value. The style is what is commonly known as Alexandrine, and somewhaa t close paralleseebowe a b n ni y l froma l m Carthage th n ei British Museum. The illustrate e lasb to t fragmen m s beeri d a ha (fign) f o tpar 43 .a f o t bowl wit hturned-dowa n edgflang the plac ein eat beading eof . Likethe last, it is an isolated fragment, while the design upon it and the method of executio bote nar h peculiar rowing-boaa n I . t with curved prod wan a high stern, stand two nude fishermen drawing in a net with a large fish enmeshed in it. In advance of the boat a long-legged bird is holding in snaker o l ee bils , whiln it a l e behin cuttle-fisa t di h floats upright wits hit tentacles outstretched charactee Th desige . th f thin ro no s piece belongs clasa o .t s generally recognise s Nilotida originn ci occasionallyd an , s a , affording a certain indication of the provenance of the scene depicted, a crocodile or hippopotamus is associated with the boat and with other aquatic creatures plaqueo Tw .Italn i s ornamenteo ys illustratee dar d 120 PROCEEDINGS OF THE SOCIETY, JANUARY 12, 1920. by M. Salomon Reinach,1 and their close resemblance in style to the design of this silver fragment is obvious. Let us now briefly consider some of the questions which this find of treasure invokes. In the first place, where did it come from ? So little plate of this character has been found in Britain that we may safely assume it was not raided from any local source. The difference in the character of the pieces—Christian vessels for church use, vessels with pagan devices, others that may have been simply secular plate—suggest l looteal t dno fro s establishmente thamwa on t ti . The presence also of certain Teutonic personal ornaments, probably

r f Dish. FragmenFigo 43 . m , Ri wit f o th Nilotic scene.

Gothic, considerably strengthens the case against any British source, d pointan o Gault s . These ornaments (fig o bucklesconsis) tw .44 f o t , o stratw p terminals (one onl f theyo m illustrated) e mountingth , a f o s belt, and a small fibula of Visigothic type. The numerous bowls and platters represented, which have been embellished with heavily beaded edges, indicate even more clearla y continental origin. The beaded bowls have been found over a large extent of the Roman Empire, yet the greater number, though chiefly bronzen i , have been recovere Francen di fourth-centure Th . y cemeteries of Abbeville and Vermand have yielded them, and an example identical with those from discovereTrapraiwas a paga Law nin d n gravin e Mecklenburg. Such are the factors which point to Gaul as the region from whic treasure hth s raidedewa . Who brought it to Traprain Law? This is an even more difficult question to answer than the foregoing, for if, as I believe, it was buried Reinaeh's Repertoire des 271d Reliefs,an . 5 21 . pp vol . ii . 1 REPOR TE EXCAVATIOOTH F TRAPRAIN O N N LAW. 121 occupante byth laso t wh sreare d their grou f dwellingpo e spotth n , o s then these people have left no relics behind that we have yet found to prove that they were other than native Celts. The extreme scarcity of pottery fragments f otheo s a r, relics, indicate bries theiw ho wa sf r sojourn e rine havw gOn .e found a forwhic f mo s hi thas bee ha tn classe s Saxonda t equallbu , have yw e found other things s whica e har

Fig. 44. Fragments of Leather Strap with silver studs, two Buckles, Fibula, and Belt Tab. certainly Celtic f positivO . e evidenc have w thu er fa snone muse w , t therefore rel t presenya conjecturen o t Celte Th s . sea-faring wera t eno , piratic folk. On the other hand, the Saxons, who at this period were engaged in harrying the sea coasts of Southern Britain and Gaul, un- doubtedly were seriouo S . menacsa e wer lattee eth r tha Romane th t s actually appointe officern da e " Cometh , s littoris Saxonici, organiso "t e resistanc theio et r attacks probable Th . e solutio e difficultth f no y seems to be that we have here part of a great booty acquired by one of these Teutonic a roverband se a serie f n o si f raid o s s into Gaul. differene Th t condition whicn si fine e plathw th d e perhaps indicate 122 PROCEEDINGS OF THE SOCIETY, JANUARY 12. 1920. .

that this has been an accumulated booty. Three beaded bowls except for natural decay triangulaa , r bowl severad an , l spoons were practically perfect when found n anotheI . r grou f objectso p a shallow, , fluted dish, a pair of small beaded bowls, a cover, a shallow bowl, and three spoons completee ar , t crushe,bu twistedr do . Several articles, only single halvee hoardth f whicn i o s, have har e been systematicall t througycu h e middleth largA . e numbe f bowlso r , platters, etc. e representear , d by small portions only. Various pieces somn i , e cases amounting to one-half of the originals, had been folded up into small packets hammered an d flat. Insido etw of these there were discovered piece f pewterso , presumablt ypu there fraudulently to°ade th o dt weight. Ther e eighar e t small packets approximately about 1 inch square, several times folded t intcu oover d shapean , , ready e cruciblefoth r lastld e an , yon drop of run silver (fig. 45). If we are to regard this as the fruia singl f o t e raidcomw ho , e about these difference n coni s - dition ? There is no ostensible e beadeth reaso y d wh nbowl s should have been kept intact, unless s possiblei s a , , they were Fig. 45. Fragments of Dishes-folded for melting, actuall previoue us n yi beino t s g and a Drop of run Silver. t i e s cacheplacei th r n i no d; apparent why certain, pieces should have been selected merely to be crushed, while others have either been cut in halves and separated or reduced to small fragments, and as regards the greater number of these, parted wit melter ho d down packete Th . s loaded with pewter, evidently so treated for purpose of direct trade, may quite well have been so prepare othey db r hands than those which were engage convertinn di g the plate into bullion. In fine, we have no certain evidence yet that e individualth placeo e wh silves th d r e cachinth e wer e originath e l raiders who obtained it in Gaul or elsewhere. questioe th A origins o it st s evidenf i n o t i , t fro e variemth d stylef so ornamen executiod an t n t comtha whole no tth es fro eworkshopha e mon . The general character of the art displayed on the finer pieces, such as REPORE EXCAVATIOTH F O T TRAPRAIN O N N LAW. the small flagon with the scenes from Scripture and the fragment with the representation of Pan, is classical, as are also the border designs on e remainth e thre th f eac o sf e o h larger flagons with designs which resemble those on objects in the Cairo Museum. There is little sign of that decadence which begins to make its appearance in the early Middle Ages whicd s an sometime, ha h s been associated with early Christian artcomparisoA . f fourth-centuro n y finds shows thae prevailinth t g style generalle ar s y thos n voguei e throughou Romae th t n Empirt ea timee th ; tha bowle range th tth f seo with beaded border fros si m Cairo Corbridgo t n Northumberlandi e d thae closesan th ;t t analogies with the rarer designs employed are to be found preserved in Egypt. How the treasure came to be hidden in a hole on Traprain Law, we valuabl cannoo importanto to to s d twa etellan t I fancy.I , havo t , e been private th individuale ehardls on i wealt y t i an d y f hcredibl,o an e thas it t existence was not known to many. Its concealment in a comparatively shallow hole indicates hasty actio presencn ni f someo e imminent peril. The most we can be sure of is that the anticipated danger materialised, and that none of those cognisant of the place of concealment ever returne recoveo dt r their wealth. Further exploratio reveay nma l fresh e presen th r knowledg r e lighou fo th factstf t o t bu , e doe t sufficno s e to dispel the darkness that enshrouds the history of this great hoard previous to its being buried on the shoulder of Traprain Law. This Report as regards the silver is necessarily incomplete, as the numerous piececoursin beinsare e of g opened out, examined repaire,and d by Messrs Brook & Sons. It was thought best, however, to give a preliminary report to be printed in the Proceedings, reserving the final accoun monograp a fine r th d fo f to issueSocietye e b th o hy t d b . Before closing, I desire to advert to the great benefit conferred on . BalfourJ . A e Societ t r onlth M no ,continuin n yi y yb countenanco gt e r excavationou t als generouslbu ,n o i y handine gfinde th th oveo t sl al r National Museum of Antiquities. With a less public-spirited landowner, awkward questions arisin f Treasuro w g la fro e em th Trov e might easily have produced trouble. To Mr John Bruce, F.S. A. Scot., we are peculiarly indebted generosits hi r fo , y mad excae carro t et possibli th s n -u yo r efo vation ; and similarly we are under obligation to the Carnegie Trustees for a valuable and much-appreciated grant from their Research Fund. We are also obliged to Mr Mark, on whose farm the Law is situated, for permitting us to carry on pur exploration. My personal thanks mano t e ydu friends e r Georgar M o t :e Macdonal r mucdfo h scholarly help; to Mr Thomas May for examining the Roman pottery and supply- wite m hg copiouin s notes thereon r MathiesoM o t ; e r beinth nfo t a g trouble to redraw the plans; to Dr M'Lintock, of the Royal Scottish 124. PROCEEDINGS OF THE SOCIETY, JANUARY 12, 1920. Museum r makinfo , g analysi f metalso s , etc. Lastlya y I ,desir pa o et tribute to the staff under Mr Pringle, whose intelligence and zealous discharge of their duties have produced the notable results which I have here the honour to relate.