Excavation of the Sculptor's Cave, Covesea. 177 I. The

EXCAVATIO E SCULPTOR'TH F NO S CAVE, COVESEA7 17 .

I. THE EXCAVATION OF THE SCULPTOR'S CAVE, COVESEA, MORAYSHIRE. BY Miss SYLVIA BENTON, M.A., F.S.A.ScoT. The cave is situated 1£ mile due north of Gordonstoun House, 6 miles north of Elgin, on the south shore of the Moray Firth. It takes its name from, the so-called Pictish sculpturings outlined on the walls and published in Sculptured Stones of Scotland and Early Christian Monuments Scotland.of When I obtained permission to see these in 1928, my companion, Miss Mollie Hair, calle attentioy m de fac th t o t nthae floos th twa r strewn with human bones. A knife stuck into the soil showed fat- looking occupation earth. The late Sir William Gordon-Gumming, Bart., was sympathetic to r projectou readild san y gav permissios eu mako nt furtheea r examina- tion. A week's work at trial trenches gave us objects of Bronze Age d Romaan n hop a date d f stratificationeo an , , thougo n thee d hw nha idea of the richness of the deposit. The following yea Williar s son e presenSi hi r ,th d r Alastairman Si t , with wonderful archaeological interest and a most unusual generosity, put at my disposal the tools of the Gordonstoun estate and a team of four men led by Mr Charles White. Whe nI starte d cave diggin I gsough vaitechnicain t nfor l advice. I should like to record here a few hints gained from my own experience. A dum always pi nuisancesa insidt bu , cavea t produceei s intolerable confusion of mind. Everything was put on a barrow and taken into the daylight. Here we soon learnt that the top soil was full of small bronzes which were quite invisible inside. Oncbarroa n eo e eart wth h migh t throug s pu welusen a te b lme d ha sieve e thiTh s. even when examining material whic he riddled b coul t no d , suc s claya h . Each receive s portioe siev hi dkned th ean n wo n whetheexamined ha e h r d notr o .t i Layer beed ha sn mad haphazarn ei stratup d to fashio e y mth b n ni burning and trampling on the various floors. Owing to their irregularity friable th d ean naturshinglee th f eo , these were useles stratificationr sfo . It was, however, found that a spade inserted above a floor went through easily and safely. If it "was an inch too high or too low considerable force had to be exerted, and there was great danger of breakage. In e invisible frailtvieth th f f wo o y e bronze this observatioe th f o s wa n greatest assistance to us. 2 1 VOL. LXV. 8 17 PROCEEDING E SOCIETYTH F O S , JANUARY 12, 1931. The same system was adopted when dealing with much of the clay. Layers of sand were interspersed among it, and the spade ran easily among these. Tearing lump f claso pieceo yt s hour after hou mosa s ri t severe test of human patience. In 1929 I had the advantage of the help of Mr David Wernham, who worked most manfully at these and other ungrateful tasks suc washins ha g bones. misse We verdhim y muc 1930hin , whe dealnwe t even more faith- fully with the clay. Sir Alastair again gave me the same team, and, as in 1929, we worked eight hours a day for five weeks.

THE CAVE. The cave faces north, looking across the Firth to the Caithness hills neven su e ra nort toucheTh d (fig. han 1) , .win it s d made sievint ga the entrance an unpleasant proceeding.

Fig. 1. Sculptor's Cave, Covesea, Morayshire. Even now it is difficult to find and difficult to reach. It is cut off at high tide in stormy weather; on two occasions I wondered whether I should climb round without being drowned. At the time of our earliest inhabitants, when the sea was little below the level of the passages, access could only have been possible at low tide t musI . t then have bee nmuca h less comfortable, thoug evenn ha more secluded, residence than it is to-day. The cliffs are red sandstone, and have been used for quarrying. The cave itself, as shown on the plan made by Mr W. Whittet, is abou fee0 fee9 t0 t 4 tlon d widgan e (fig . Sinc2) . e plas madeth nwa e ew have cleared anothe fee0 f 1 roofro t . Ther a passag s ei e behind going o

ffi W en O .±*-rT__j. ____ i ____ i ____ i ______i ____ I ____ i 3 o

= ROCKS. W= ROMANO- BRITISH HEARTH IN 3/TU. XYZ > BOTTOM OF BRONZE AGE. LAYER. RS- STRIP OF ORIGINAL SURFACE. LEFT ALONG THE. BASE LINE.. o Fig. 2. Plan of the Sculptor's Cave.

Fig. 3. Section of Floor of the Sculptor's Cave. 0 18 PROCEEDING E SOCIETYTH F O S , JANUAR , 1931Y12 . some 30 feet into the rock. Progress along this is by rolling, and it has not been shown on the plan. The two passages at the entrance are separated by a wall of rock. Falls of rock have occurred at various times in the south-west corner and outside. The surface of the rock is sculptured friablol e somd th ean muce f eo sar h decayed. originae stria Bth Ss f i p o l surface left alon base gth s i e lineZ ,XY the hard crust below the Bronze Age layer. W is a Romano-British hearth left -in situ above the Bronze Age layer. Strips of layer 2 have been left along the inside wall of both passages and in -CO. A good deal of excavation has been done outside the cave. The height above sea-leve s foun wa Majoly b d r Sleight, R.E. THE STRATIFICATION. There are three main layers of stratification (fig. 3). Layer 1, Occupation Earth.—This extends over the whole cave, varying deptn i h fro minche2 composes i feet 2 t I e blaco st . th f manf ko do y fires mixed with sand, gravel, and debris. As described in the intro- duction s dividei t i , y floorb d s hardene d tramplingfiry db an e e Th . gravel was so loose that coins were found in every layer, and pottery from different levels joined. Objects from the Bronze Age to a Viking rivet were foun thin i d s layer t objectbu , s fro e seconmth d centuro yt the fourth century A.D. predominated. Bronze Age objects were found in it chiefly at the back of the cave and along the east side. They were not found within 40 feet of the entrance (ABCD 4) to base line. Layer , ClayLayeredd 2 an Sand.—These mixee e uppear th n di r part. This dotted division belo ws composei f glutinouo d s clay t containI . s laminations from outsid e doos thi th eo ha ( t Asr d 2)characte,an n i r crevice e rocth kf o sbeyon d this point. Frofee0 4 fee5 tm 2 o frot t m base th e ful s lini f burn o t lei t black marks, charcoal Bronzd an ,e eAg objects. The clay dies out where the beach slopes up, about 45 feet from the door. The Laigh of Moray to the south has just such clay, more than 60 feet brick-yarde deeth t p a Elgin-Lossiemout e th n so h railway linke th s s A . between Lossiemouth and Covesea lighthouse are recently made land, s conceivabli t i e thae Lossieth t , flowing throug e Laighhth , entered 2a £ se mile e th ss presenwesit f o t t mouth. Probabl e wholyth e beach was muddy at this time when the sea had already started to retreat from the cave, and river and sea between them carried the mud into a sanse k e further th e cave s th A . , wind-blown sand covere e outdth - side clay, but while (A 3) - (D 3) was still moist the Bronze Age people came and lived in it and trampled their goods and chattels into it. The black line of their smoke can be traced from (A3)-(A1) outside, where we EXCAVATIO E SCULPTOR'TH F NO S CAVE, COVESEA1 18 . found separates potteri ) 3 D ( ypottery) - d3 ) lik 3 D fro( e A e - ( (m) Ath . 3 Roman layer by a hard burnt floor which has proved impenetrable to the smallest coin, although it is only about 1 foot below the surface. Near the door it is about 6 feet down. The intermediate sand was almost barren abov e clearlth e y defined Bronz e levelAg e . Opposite e rocth k wall, betwee e passageth n s outside a fal,f roc o l k seemo t s have crashed inhabitants e dowth n no , leavin t leas ga crushee on t d skull. Layer 3.—The shingl 25-fooe th f eo t beac s barrehwa n e excepth r fo t bonesblacw fe ka . lind ean Explanation of Marking.—Throughout, letters and numbers in brackets refer to the divisions of the cave. See plan (fig. 2). Objects stratified in the Bronze Age layer are marked 2nd. Objects not so marke froe e mixedar m th d layer. HEARTH (B3).—The only struc- tur were ew e abl identifo et s ywa an apsidal hearth, set with curbs roun dapse parth ef (figo t . Muc4) . h credit is due to Edwards, the workman, who found and cleared it on his own initiative. I believe it to be complete. It was at the bottom of the Roman period layer, and we assured ourselves that no coins lurked in the moist clayey Fig . Heart 4 . Sculptor'e th n hi s Cave. san whicn di e stonehth s were set, forming a barren layer above the Bronze Age clay. The photograph e heartoth f h foun t Trapraina d J seems very like r CurioursM e. found another at Grlenelg. There were traces of other such pavings. I

ought, perhaps mentioo t , largna e fla2 t stone evidently useheartha s da - stone befor e Romath e n s beeperiod ha moute e n t th th I left f . a ho t cave. BRONZE AGE. GOLD (C9).—Four pieces of False Ring-money (fig. 5, Nos. 1 to 4). These are small penannular rings, composed of a bronze core covered with gold. They have pointed terminals. They are in very bad condition e cor th ,s disintegrating i e d splittinan , e goldgth . Similar rings have been found in Scotland at Balmashanner, Forfar, and at Fuaraig

Grlen, Banffshire. They are common in Ireland. They 3 are too small

4 5 1 Proc. Soc. Ant. Scot., vol. liv. p. 62, in the second century. 2 Ibid., vol. Iv. p. 93. Ibid., vol. xxvi. pp. 182 ff. National Museum. 34 British Museum. Armstrong, Catalogue of Irish Gold Ornaments, plate xiv . 271pp ., etc. 5 182 PROCEEDINGS OF THE SOCIETY, JANUARY .12, 1931. to have been worn as rings, and some, when found, showed traces of bronze th string n i . e Suc7 s rin. ,a hgNo rings •were sometimes carried on bracelets.1 Measurements when found, '5 by '55 inch; weight, 3'5 dwts. (D2'/2nd).—A similar rin s gfoun wa (No ) d5 . e hiddeth n i n Bronz e clayAg e, beneat e rocks th a hshel f o . f (B4/2nd).—Another (No. 6) was stratified in the Bronze Age clay. (B3/2nd).—A stratified piece of leaf gold, perhaps from another ring, is not illustrated.

Fig . Gol .5 Bronz d dan e Ring-money from Sculptor's Cave, (y.)

BRONZE (C9).—Two bronze ring ) wer8 s (figd e , Nos5 an foun. 7 . d along wit e gold-platehth d ring f similaso r shape. Diameters6d , '7an

inch. Weight dwts3 , I hav. e only seee otheon n r example fro - m Torostan.2 - Bracelets (fig. 6 (D 5)).—Penannular bracelet (No. 1), rectangular in section terminale Th . beatee sar fro p inside nu mth e outward foro st m irregular knobs. There are also two cast, horizontal ridges on the outside besid e knobseth brokeA . n s founpai wa rt Auchtertyre a d 3 with three similar ridges. Otherwis ebelievI Covesee eth aunique b pai o t re -thn i e British type Isles s certainlTh i e . f foreigyo n origin, coming

originally from Upper Bavaria, but it went into Switzerland, and 5

was there modified into forms almos4 t identical with ours. (D 4).—The

British Museum, Bronze Age Guide, p. 54, from Ely, Cambs. 6 Nationae th n I 21 l Museum. 3 Proceedings, vol. ix. p. 435. Not illustrated.

4 Childe Bronzee Th , Age, . 119p . 5 Anzeiger fur Schweitzerisches Altertumskunde, xxix., Taf. xiii., No. 1. 6 Landes Museum, Zurich 1327—. Zuriche thres d No , ha c e1t I La froridge. ende e mth th , t sa alss ha o t fainbu t decoration all round . 1723—No . 6 from Morigene ridg-e ,on whicon n eo s hha side only. In shape it is identical with ours. EXCAVATION AT THE SCULPTOR'S CAVE, COVESEA. 183 other bracele e paith rf o t(No . (C3/2nd).—Fou2) . r fragments (No, 3) . probabl bracelet.e on f yo l These were stratifie tere claye th Th -.n di minal e similarlsar y beate frop nu m the- inside r analysi . 208Fo p . .e se s (A0/2nd).—One slightly swollen termina s founwa l d abov e Bronzth e e

Fig. . Bronz6 e Bracelets fro Sculptor'e mth s Cave. (f.)

Age layer fee,3 t belo surface wth e outsid claye eth . This stratification is uncertain I attac d o importancan n h, . it (No o t e t illustrated.) (D5).—Bracelet with very slightly beaten terminals (No. 4). Cf. Braes of Gight.2 I distinguish two types of penannular bracelets with swollen terminals. Both come originally fro e Continent,mth seeme on havo t st 3bu e been

1 Cfunpublishen a . d example from Heathery Burn wit brokeha n terminal. British Museum. 2 Proceedings, vol. xxv. p. 135. 3 A.S.A., vol. xxix., Taf. . 3 xiixiv. d . an . No , 184 PROCEEDINGS OF THE SOCIETY, JANUARY 12, 1931. developed in Ireland, and it is generally found in gold--the Irish material. In these the terminals have been beaten up regularly all round into a mushroom n ScotlanI . d thee mosar y t frequent dow e westh n t coast, though the e easfoune th gold.yn ar i t n o d 1 Tw bronzn oi e have been foun t Killin,a d 2 Perth- shire. The seconr typeou s ,d i wit h terminals beaten from the outside to form a knob. These are much closer to their foreign prototypes, and those in bronze are confined in Scotland to the counties of Angus,3 Aberdeen,4 and the shores of the Moray Firth.8 In England one occurs at Heathery Burn Cave.6 Thu e havw s a speciae l kin f Swiso d s bracelet, concentrated in a limited area, and show- sign0 littlg 11 f modificationin o sr eo . BONE.—It has been found convenient to take all the larger bone instruments together, though, as is suggested below, some may be of Roman age. (D7).—Instrumen nettina te b sai o gdt needle, probably made from the metatarsal of the red- e deea holnatura Th s i er (fig. l7) . nerve-hole which has been utilised, perhaps increased. The depressions down the centre are probably both natural, but they are increased artificially. There laterae ar l cuts down both depressioe sideth f so n and the point is worn. Several similar instruments t withoubu , t perforation d retaininan , g the outside of the bone and part of the process, 7 Inch. were found at Heathery Burn Cave, and one other is in the Pitt Rivers' Museum, Oxford, from Pig. 7. Netting Needle from Jordan's Farm r specimeOu e . s founth wa nn i d Sculptor'e th s Cave. mixed layer, buried below shela rockf fo , along with

From Billhea Nationae th n di l Museum r IrelanFo Armstronge . se d . cit.,op , plates xvi. 1 and xvii. 2 Anderson, Scotland in Pagan Times. Bronze Age, p. 149. The fragment from Traprain appearthif o se b type o st . Proceedings, vol. xlix . 165p . . Balmashanner, Proceedings, vol. xxvi. pp. 182 ff.

3 4 Braes of Gight, Proceedings, vol. xxv. pp. 135 ff. Behill, Proceedings, vol. i. pp. 138 ff. 5 Auchtertyre, Proceedings, vol. Ivii . 154.p . Anderson . cit., op ,. 147 p ; Wester Ord, Proceed- ings, vol. lix. p. 115, No. 6 ; Conage, Pettie, Proceedings, vol. i v. p. 299. A bracelet is also mentioned as coming from Conage, Banff, which I have not been able to trace. 6 In the British Museum Not illustrated. It has one broken terminal. This shape in gold occurs at Alloa. Proceedings, vol. Ivii. p. 164 (fig. 21), and at Patcham.

British Museum, Bronze Age Guide, p. 49. 7 EXCAVATIO E SCULPTOR'TH F O N S CAVE, COVESEA5 18 . the bone slice (fig. 9, No. 5). (C 3/2nd).—Similar, but much smaller (fig. 8, transverse Th No . .1) e cut mostle pointe sar th n y.o Stratifie claye th .n di We should, perhaps, notice that these tools have the shape of a leaf- shaped sword. Ther bona s ei e reproductio haftea f no d daggee th n i r

18 If 19 1/20 1/21 ,1/22

j inches.

Fig. 8. Bone Scoops and Awls from the Sculptor's Cave. Nos. 1 to 14 and 17 from the Bronze Age layer.

Landes Museum in Zurich, said to be late-Neolithic, but one is tempted suggeso t t tha typologicalls swordsf i o t ti e ag .e th f yo Fig. 8, Nos. 1 to 14 were found stratified in clay, No. 17 was found at the back of the cave. Nos. 2, 6, 10, 11, 12, 14, 17 are of a very simple type. The bone, generally a metatarsal or a metacarpal, is split up polishee pointes insidb e i th y d e middle d th een ma d dan e (Nos, on ,6 , 2 . 10, 11, 14), or may not (Nos. 12, 17). I could not find this type among the hundreds of bone instruments in the National Museum. Professor Childe latele showeon ye foundm t Skara d a Brae t saibu ,d thae th t 6 18 PROCEEDING E SOCIETYTH F O S , JANUARY 12, 1981.

type was unusual there. It resembled No. 2, and both may have been used for dressing skins. This type is found in the Swiss lakes. I believe

e differenthaar 4 t1 Nosd te saman endth .r simila6 f eo o s r spatulate 1 instruments. There is a close parallel in Cambridge from Robenhausen.

No. 9 is a strong needle with a close parallel from Heathery Burn,2 3 Beveridge t unlikbuth no n t i e thert ei on es e i collectio thn ni e National s founwa Museumd6 1 rathe. No . r doubtfully stratifie s itseli d f an d indeterminate. No. 15 has a freakish resemblance to a human face. Nos. 7 and 8 are common in any collection of bone instruments. The Broch people seem to have had two main types of instrument. (1) They e ar 9 1 d an too, y splinte16 an k , d pointe 13 f an rwhico , 5 , , it d4 h instances. Such instruments are liable to occur anywhere at any time. e casth f eo a bon n pointed i (2f f eitheo )s ean a of Thed , t dit en r cu y 18, 20, 22.4 It is notable that these did not occur in the Bronze layer and that the bisected type did not occur among the unmixed objects of the Roman period. List of Instruments in fig. 8.—1. (C3/2nd) Netting needle? Probably metatarsal of red - deer. 2. (— B 0/2nd) Scoop. Probably proximal or anterior portio f metapodiao n f red-deero 2/2ndl D ( . )3 Punchers. . Lateral metacarpa 3/2ndB ( . l5 ) equus d Prickersan 3/2nd4 C ( . . 6 ). Scoop . 7. (B3/2nd) Pin. 8. (B4/2nd) Pin. 9. (D2/2nd) Needle. 10. (-Bl/2nd) Pricker. Probably a fragment of the proximal end of the metacarpal a ored-deerf . (B4/2nd11 . ) Pricker. Anterior fragmen e proxith f o -t e metacarpath f o a red-deerd f o en l. (-B0/2ndl 12 ma . ) Scoop. Metacarpal of a red-deer. 13. (-Bl/2nd) Pricker. 14. (C3/2nd) Scoop. 1.5. (G9) Pricker.. 16. (-Al/2nd?) Pricker. 17. (C7) Pricker. 18. (06) Scoop: t ) stratifiedPrickershee1 No C ( r . goat o p. 20 19 . .. Pricker. 21. (CO)Pricker. 22. (B 10)Pricker. I am indebted to Miss D. M. A. Bate of the Natural History Museum for these identifications. I take this opportunity to thank her for this mand an y other kindnesses. Bone Needle.—(C 9). This is a long, finely polished needle pointed at e shape th a bronz . r botcf , Fo he . endneedl8) . s t NidauNo a e(fig , 9 . - Steinberg.5 Ther s nothini e g lik t i froe m Scottish sites. (D3/2nd) Similar needles (fig. 9, Nos. 6 and 7).

1 Keller, Lake Dwellers, plat , froe 20 miii. . Meilen,No s bee t shortha , cu nt otherwisbu , e resembles No. 2. Steckborn, No. 16693, 40, 41, in Zurich, resemble Nos. 11 and 12. I have not been able to trace any other examples from the British Isles. I -have seen one from Macedonia and some from Thrace (unpublished information by the kindness of Miss Lamb). 2 Ethnological Museum 66, . 724 A . No , 3 Loc. cit. 1 Wace and Thompson, Prehistoric Thessaly, p. 148, from Tsangli and Tsani. It is common t Glastonbura Zurice th n i h d Museumyan . 5 Keller, Lake Dwellers, plate xxxvi., No. 13. EXCAVATION OF THE SCULPTOR'S CAVE, COVESEA. 187 (D 7) Perforated Slice.—This has been attributed to the Bronze Age because of its connection with the netting needle. Nothing exactly

10'

Fig. 9. Bone and other Objects from the Sculptor's Cave. (^.) bees likha n t e i recorded, though ther e somear e coarser parallels from Beveridge th e collection. rectangulars i t I 1 thao s , t cannoti a whor e b t l From Tota Dunai Bealacd gan h Ban, bot Nortn hi h Uist. 1 8 18 PROCEEDING E SOCIETYTH F O S , JANUAR , 1931Y12 . (fig. 9, No. 5). I know of no other instance of this shape. Perhaps it is ornamentn a . Classified List of Bone Instruments. Scottis Swissr ho . Swiss. Indeterminate. Fig. 8. Fig. 7. Fig. 8. 18. (06) 5. (D7) 7. (B3/2nd) 20. (not stratified) . (B4/2nd8 ) . (BIO22 ) Fig. 8. 15. (09) 1. (C3/2nd) 16. (-Al/2nd) 2. (-B0/2nd) 21. (CO) 6. (C3/2nd) 9. (D2/2nd) 10. (-Bl/2nd) 19. (01) . (B4/2nd11 ) 3. (D2/2nd) 12. (-B0/2nd) 4. (B3/2nd) . (C3/2nd14 ) 5. (B3/2nd) 17. (07) 13. (-Bl/2nd) Fig. 9. 6. (09) 7. (D3/2nd) 8. (D3/2nd) Ou f twenty-sio t x instruments, twelv e certainlar e y foreiga d an n good many more may be foreign. (A 0/2nd).—The basal portio a she f no d antle a red-dee f o r r (Cervus elaphus), with brow tine missing. Miss Bate writes: opinio"Iy nm t ni undoubtedls ha y been artificially worked—show somy nb e sharp incisions t tha e dista e curiouth ly brokeb wels a s s a d excavationl nen s (worn smooth by thongs, possibly) at the proximal end of the antler." The excavations are three in number, and show an unexpected regularity. The whole bone is much worn, the usual vertical corrugations being missing. Mr Reginald Smith compares bone chopping-blocks or anvils from L'Aquina, and, in truth, this tine could be fixed rather comfortably into

grounde th . Whe 1 L'Aquine nth a bone beed ha snlittla e worked, there were cut fros sa mshara p instrument. After further usag e surfaceth e became not unlike our excavations. They do not, however, occur in threes or on raised blocks or in horn. Worst of all, I am loth to admit a Mousterian product among our Covesea finds, even although it was found 2 feet below the Bronze Age level, just inside the door. POTTERY : Bronze Age.—In quality the pottery resembles that of the 1 Henri-Martin, L'Evolution du Mousterien, plate lix. EXCAVATIO E SCULPTOR'TH F NO S CAVE, COVESEA9 18 . t unparalleleno s brochsi t i t pottern bu ,di f Neolithiyo Bronzd can e eAg Zurictimee th n si h Museum. It is coarse; most of it contains large stones, like garnets in appearance, e.g. pots K, X, t, tj. Some of it contains large white particles that seee shellb e outm /j.,o , t 0 ,Th i.- side surface rasps lik efila e when the hand is drawn over it, particularly pot v. The only pot that show attempn a s t levigatioa t n is f. The inside of these vessels is generally smooth, notably so l excepd eal ; an e cas8 th tf eo n i rj are more or less blackened, and scraps of food often appear. Almost the only shape seems a pail-lik e b o et cooking-pot, wit hrima . which projects slightly inside (fig. 10). The pot curves out slightly below the rim. All bases preserved are flat. For s beeha fig, n, 8 rim10 . fig. e s11 se . restored with slightly too wide a 3 Inches. . 5 . No base . fig, Cf 11 . . The pail is indigenous to Great Britain; it appearsin steatite urns from Orkney undecoraten i d 1an d transition pottery from Chastle- Fig Bronz. . 10 Cooking-Po e eAg t froe mth ton,2s i Oxfordshire m ri e Th . Sculptor's Cave. essentially like those of the cremation urns. I am not aware that the two have been found in conjunction in this country. They are to

b eZurice seeth n i h Museum from Neolithic time Hallstatto st , I. 3 1 Anderson, Scotland in the Iron Age, p. 73, fig. 57, from Fair Isle. 2 To be published. Neolithic, No. 6223, from (Berne) Neuville-Chavannes. It is unlevigated, but smoother than

our3 pottery. The ware is grey and decorated with nail marks. Bronze Zuriche d Age: same c Th La e. rim, shape sizd an e, continue unchanged t leasA . t half a dozen vases might be quoted, though most have some decoration. No. 852, No. 1419, from Wollishofen alse oSe Mitteilungen. der Antiquarischen Gesellschaft n Zurich,i xxix., part iv., plat. e29 vii.. No , Morigen, No. 6287, found with leaf-shaped swords. Grandson-Courcelettes, whee e 6624. b y No .l grey-blacmads ma Thii t n i seur ; polishedd kan . It is found with winged axes, antennae sword, and open-work razors. It has lines incised round traca platen d suco f et o an , hlin.a po it shows e i th n no Hallstatt, I., Wangen Grave 4. It is of greyish-yellow unlevigated clay. It is hand made and undecorated founs wa d t I wit. bronzha e girdle. 190 PROCEEDING E SOCIETYTH F SO , JANUAR , 193112 Y . 8. This pot was found mostly at the back of the cave (fig. 10), but two pieces, which may join, were found 6 feet down in (- B 0). At any rate their fabric is identical. The rim is flattened horizontally. Cf. the rim of € No. 6 on fig. 11. The colour is brick-red shading to black. Restoration by Mr Young. e. Rim 6.—Thi. s als, No o wa fig e cave , t bace founth . 11 th po sf ,kt o da and has a similar rim and colour. £. (D3/2nd).—This pot was found in woeful disintegration at the bottom of the Bronze Age layer. Some s i indistinguishabl t i of e froe th m last. i). Thi t waspo s foune clath f yo n di (D3/2nds i m d (D2/2nd)ri an ) e Th . , showinsimila7 . o No figt r a , g .11 distinct slope down toward insidee sth . brigha f fabrio d re s ti ct I y thama t attempbn ea reproduco t Hallstatea t pigment. 6. Rim . 7.—This ,No wa fig , t .11 po s found with a mutton-bone inside it and crushea d skull outsid (l/2ndn -i B t ei ) under falle slopem nri rockse s Th . more than the last. Though much of t bee t survivesi no ns possiblha t i , o et Fig.. Section11 f Kirno s d Basan sf o e complete it. It is reddish in colour, with Bronz Pottere Ag e y froe Sculpmth - tor's Cave. (^.) white bits and garnets. d basean , m fig , Nosd i.Ri .11 an .1 5.—Rim somewhat smaller tha e lastth n . Very rough. Yellowish grey. K and A. Rim, fig. 11, No. 2.—Found in (B3/2nd and C3/2nd). The ism surfac similae ri laste Th th . o t er varie muco s h thanumbero tw t s have been given in case the fragments represent two indistinguishable pots: grey with garnets. /x. Rim, fig. 11, No. 4.—This is the only incurving rim except for one rather amorphous fragment alss i e thickes ot th I . fabrin i t cond can - tain bitg f skulllighs bi si o t I t Traprain, .a pin e colourn ki on . Cf .

which may not be1 out of place in the sixth layer. v. Rim similar to K and A.. This is dark red, with the surface of a nutmeg grater (D 2/2nd). f. Rim 3.—Th. ,No fig , .e11 shap mors i f thieo m esri lik Iron e a nAg rim than those we have been considering, and the fragment drawn was 1 Proceedings, vol. Ivi. 224pp . , 226, fig. .23 EXCAVATIO E SCULPTORTH F NO S CAVE, COVESEA. 191 found in the first layer in (B5), but other fragments of undoubtedly the same pot were found well stratified in (B 4/2nd). STONE.—(C3/2nd) Hammer-shaped piecf o e bore d sandstonre d e wit hchannea e on n o l side (fig whae s hari se . 12) t o dI t .purpos e this object could have served. Perhapss wa t i hung round the neck as an ornament. In west Ireland they had a way of boring granite1 for an equally unknown purpose. Sandstone hammers are found in Scotland, but this stone, less tha ncubi2 c inche n sizei s , would have Fig-12. Perforated Stone from been quite useless. the Sculptor s Cave. •($.) (B9) Broke flint-knifed nre yelloA . w flint-scrape s founwa r(C3n di ) area, but the depth is uncertain. ROMAN PERIOD. Bronze Rings.—Fig. 13, No. 7 (B3), is a coiled ring in very bad condition. It is common for a long period. Nos. 8, 9, 10, 11, 13 are small rings. There are fragments of several others not illustrated. No. 12 (B3) is a piec f knotteeo d wire whic doubtless hi s parallea common s ha n d i l an , Saxon times in the Ashmolean Museum. No. 14 (C 0) may be a piece of chain armour.2 plaia Figs i (n) B4 7 1 . penannula belony r ma ring gt I . to the Bronze Age. Bracelets 2).—D ( A broken bracele rectangulaf to r section surface Th . e is corroded, but it appears to have had some pattern. One terminal is a hook othee ,th brokens ri . This fastenin commos gi Roman no n sites.) 3 B ( 3 Probabl yfragmena twistea f o t d bracelet. Cf. Traprain Colchested 4an r B. M. (B 5) Piece of a chain with spirals. Pendants (fig. 13, (D4)).—I7 Nos1 (D2)d 5 .1 an (, 6 As suggeste1 i ,t4) d tha tage t legionariesr b thes sfo y ema ' earlbeltn a f yso type, where eth straps are cut into small strips, each with a tag. Ritterling thinks that 150 B.C. is a terminus ante quern for these.5 They are in bad condition, and Mr Young informs me that No. 17 is stained with leather. Nos. 18 (A 4) and 19 (C5) are little pendant saucepans or bullse which date from sixth century Greece onwards.6 Mr Beck quotes an example from Harlech said to be Roman and some from Bologna.7 (06).—1 . No , FigA15 . hook-like object files uppee i d Th .down en r , 1 Private collection. 2 Curie, Newstead, . 161p , plate xxxviii., figs. Govesemese e 8,10th Th f h.o a exampl coarsers ei . 3 On Colchester. In the British Museum. 4 Proceedings, vol. Ivii. p. 198, No. 11, 3rd level. Ritterling, Hoffieim, vol. xv., No. 73; vol. xi., Nos. 56, 57, 66. Bushe-Fox, Richborouffh,

. 48 . volp silvee . th ii 5 . . r Cf sstraine t Traprainra , Proceedings, vol. . 9 Ivi. . ,No , fig20 . 7 Del Picino and Montegiragio. 192 PROCEEDINGS OF THE SOCIETY, JANUARY 12, 1931. the lower broken. It is of stouter make than most of the objects of this period. Nos. 4 (A 5) and 3 (D4). Probably the bolt and key of a barrel- a hollo s i 3 w . squarNo f o e locke ben d whicTh en t. hp u coul n ru d

15 19 17

. BronzFig13 . e Object Romae th f so n Period fro Sculptor'e mth s Cave. ({.) EXCAVATIO E SCULPTOR'TH F NO S CAVE, COVESEA3 19 . No. 4, so that the springs on both sides could be compressed and then withdrawn similaA . r bolt withou e sprin th ts foun g wa t Traprain.da

No. 2 (AO) seems to be a sword guard or scabbar1 d top. A straight-faced bronze guard with a Celtic frill is found on an iron sword at New- stead,2 and plain bronze bands are found on iron sworde Britisth n i sh Museu t Oxforda d man . A smaller mounting occurs in the second century at Traprain.3 Mr Curie thought it out of place and attributed it to the Bronze Age. Covesea supports his stratification as (A 0) is at least 5 feet abov e Bronz5 th e e . levelAg No e . unexplainen (Ala s i ) d object whic s perwa h- Fig. 14. Bronze Penannular haps inscribed. "TBBS a ver s i "y uncertain King from the Sculptor's Cave, (f.) reading. Bronze Wire Beads (D 3).—One complete bead and several fragments were found r BecM .k writes: " These, although foun Roman i d d nan Saxon period, in this country were rare at that date."4 From its position

. BronzFig15 . e Object Romae th f so n Period fro Sculptor'e mth s Cave, (j.) t oughi belon o t Romae t th o gt n period. Ther Nationae foue th ear n i r l Museum—two from Traprain,9 one from Jarlshof,6 and one from Dun Beag."

1 Proceedings, vol. Ivii. p. 222, fig. 28. 2 Curie, A Roman Frontier Post, p. 185, plate xxxiv., No. 10. 3 Proceedings, vol. Ivi. p. 212, 1, p. 213. Ashmolean Museum, from Wheatley Grav Proceedings, e27 Birminghame oth f Philological Society,4 vol. iv., part i., p. 179. 5 Proceedings, vol. Ivi . 213. 212p p . , . ,1 National Museum. 6 National Museum. VOL. LXV. 13 194 PROCEEDING E SOCIETYTH F O S , JANUAR , 1931Y12 .

Fragments.—Small bronze hook (C8d variouan ) s fragments from different places, including broken pin steins. Bronze Wry-necked Pins with Circular Heads.—These are among our most attractive finds. Such pins are not uncommon in Scotland and elsewhere t theibu , r evolutio chronologd nan t presena s yi t somewhat obscure. I cannot follow Mr Reginald Smith in his claim that the Keady pin Lingroe ,th we e Bowermaddeth mouldth l d al an e , ar n npi same. They threseee b mo et different pins. The Keady pin illustrateds a , , appear havo st e small corrugationl al s round (not two-third way)e th f ,so whic hI believ e represen threae th t d wound round to keep the pin in position. The Lingrow mould has some

decoration at the base, possibly like ou1 r No. 7, and the corrugations

are half way to becoming pellets, as in Traprain, and Tents Muir, Fife. 3

The BoAvermadden pin is becoming a hand pin 2 below, like our No. 5, and has corrugations intermediate between the Keady pin and the Fifeshir aboven pi e chronologA . y base thin do s misconceptiot no s i n very promising, and when we apply it to .Traprain we find the ring- headed pin, supposedly dated before them,4 present in all layers from earle th y seconfifte th h o dcentury.t 5 It seems probable that there are more elements present in the evolutio f thithan o Smitr spi nM h -woul dpita allows leavo yi t t I e. out the sunflower pin, and I would suggest that the rolled pin with a crook in the neck, which the Swiss used to attach their sword girdles,

cannot be neglected, especially as one of these girdles reached Braes o0 f Gight.7 Leavin earliee gth r histor futura r yfo e occasion s confinu t le , e our- e exampleselveth o t s s actually , seem 1 befor . e . th sFigNo us e , .16 simples d thereforean t , perhaps e earliesth ,e th f t o typ d een wite on h cloth wound ove e basemuse th rW .t remember, however, that simple f persistino y wa a gtype d alonha s g with their more elaborate deriva- tives, e.g. the lap-over pin at Traprain, already quoted.8 No. 7 would come next, with three swathes of cloth at the base still looking like cloth. Nos. 8 and 9 are different stylisations of this motive. No. 2, with three definite pellets, develop e rose-pin.s th fro , 3 mint. 8 oNo Nos, .4

Proceedings, vol. xx. p. 348, where see references9 .

21 Ibid., vol. Ivii. p. 210, No. 3. Nationae th n I 3 l Museum. Smith Kit.,. . 347op , p ,first e "Enth f , do beginnin secone th f go d century A.D. thinI " k that i nsubsequena 1 t pape pinsn ro , "Opuscul Montelium,d aa Smitr "M h dated these earlier. 5 Proceedings, vol. Iviii. p. 262, "Analysis." Neuchatel Museum, from Auvernier. See also illustration in Keller, ed. Lee, plate xlviii. No.6 8, from Estavayer. The ribbing below is unusual. 7 Loc. cit. 8 Smith, loc. cit., Early Ness pin. " Proceedings, vol. Iviii. p. 259, Nos. 5 and 6. EXCAVATIO E SCULPTOR'TH F NO S CAVE, COVESEA5 19 . a stylisatio 6inte 9 , . ar ano 0 No dbar1 f no s betweens i pellets5 . No . the hand pin, which seems to go back to No. 7. Thes datelase o saib e eb ttw t o Traprai ddt a piny fourte sma th o nt h centur r later. yo e rose-pi Th 1 n only occurs there. Pellet bossed an s s

678 9 10

Fig. 16. Bronze Pins from the Sculptor's Cave. (*.) see o havmt e I believhad d an , e still have n attractioa ,e th r fo n inhabitants of East Scotland. It is the underlying principle of the decoration of the stone balls. The hand pin must have had four pellets when unbroken. The two at Traprain have five and three pellets respectively.1 All these pins exemplify a love of severe symmetry of ornament which is now at any rate a national characteristic. Contrast the pin r SmitfigureM . hby d from sandy beds e , stilth or , ln i more e on ,

Ashmolean Museum from Woodperry.2

2 Se1 e Analysis, loc. cit. Archaeological Journal, vol. Hi. p. 120, fig. 19, No. 10, and p. 122. 196 PROCEEDING F THO S E SOCIETY, JANUARY 12, 1931.

e parallelear pino 0 1 tw s d y frodb an , mNos6 , Nort4 . h Berwicn ki the British Museum and in the National Museum. Alpine th l s were founmixee th n di d followinlayee th n ri g divisions :— No. 1 (07), No. 2 (C8), No. 3 (D5), No. 4 not stratified, No. 5 (A 5), No. 6 (C8), No. 7 (06), No. 8 not stratified, No. 9 (08), No. 10 (04). I should like to date Nos. 3 and 5 in the fourth century and the others provisionall secone th n yi d century. Zoomorphic Pins, fig. 9, No. 9 (0 7).—Large pin found inside a bird bone. The type is common at Traprain, where it is present in all layers. It was also found in earliest Newstead, c. 80 B.C. No development of type is apparent at Traprain.1 Fig. 13, No. 6 (D4) is a similar but smaller

Fig. 17. Bronze Toilet Instruments from the Sculptor's Cave. (ca. f.) example, except that it has ducks' -heads on two sides. The form of attachmen e r sam coi, Romath ou s . 43 than alsa es r i n tCf o oou t. twisted grattoir. 7 . , No fig , .17 Toilet Instruments, 3).—PincerB ( 1 . d No figsen , wite .17 on h broken .e sam th This e ha attachmens e fourth-centurth s a t y coin, . (Wit03 5) . hNo goo. 43 d. surfacsprine No , figth 23 d . gean still working. lesn I s . goo5) NoC ( d 5 .verconditionn I conditiond y. (ba 6 0 5) . No . . No. 2 (D 1) seems to be a nail-cleaner. When it was shown on the screen a pattern of incised lozenges became visible. No. 4 (B 1) may also be a nail-cleaner. It is sharpened at both ends. A similar instrument was found in the early Saxon cemetery at Girton College.2 For the pattern, cf. Traprain.3 No. 7 (B3). Grattoir of twisted wire. The head is zoomorphic, a bill being faintly visible below the double atachment rings.

1 Analysis, loc. cit. 2 Rollings wort O'Reillyd han Anglo-Saxone Th , Cemetery t Girtona College, (B)40 2 . . p Fox, Archaeology of the Cambridge Region, plate xxvii., 2, dated late sixth century, p. 248. : Proceedings, vol. Ivii., fig. 23, No. 1, fourth century, by stratification. No. 2, fifth century, by stratification. Ibid., vol. lv., fig. 21, No. 15; also late. EXCAVATION OF THE SCULPTOR'S CAVE, COVESEA. 197 These toilet instrumentn i e d commo ar e sBronzan th e n i nAg e Roma Saxod nan n times. Needles, fig , Nos .13 (5 D (D4.1 3).—Bronzd an ) e needle a usua f o s l type which is not found in Scotland. No. 1 with ring for attachment. No , fro5 .s positionmit e Roma, th shoul f o ne db period . Nos .(052 ) and 3 (D5) fragments of another type. Penannular Brooch, fig , d 4NosAan 13 . (B6).—Thi4 . a simpl s i s e

j Inches

1 2 4 38 11 78 9 10 Fig . Bon18 . e Pins fro Sculptor'e mth s Cave. example of a pleasing type. Slightly zoomorphic terminals have been hammeret incorporateye t s rectno i e ring e t th dI - ar .n back d i d an , angula n sectioni r d lightlan , y punched s impossibli t I . o hazart e d a date.1 Bone wer, 2 Pins, d e founan ) fig Nos, 6 A d.18 ( .besid1 d ehana re f ko a pierce s i 1 d. hair four-leafeNo . d clover, beautifully polished with punch marks between lightly incised diamond steme th I knon . so f wo no exact parallel, but the style is not uncommon in the brochs.2 Knife mark e clearlar s y visibl . Barrel-shape6) e A abou( 2 e head. th t No . d pin. This specimen is also finely polished and in excellent condition. The shape is fairly common in Roman sites in England. A knife seems to have been used on this head too. No. 3 (C 6). Ring-headed pin. Well 1 Cf. Girton . cit.jop , plate iii., als orina t Traprainga , Proceedings, vol fig, .. 78 liv3 . . p . 2 Crai Boynef go , Banff, Proceedings, vol. viii . 290.p . 198 PROCEEDINGS OF THE SOCIETY, JANUARY 12, 1931. f rougheo polisheside e t on bu , n r o dmake . Knife marke seeb no t s round the head. No. 4 (B 6).—Spoon. There is very little polish about this head. The side of the head suggests little slices with a knife, but there are ^10 actual cuts, as in the last three examples. Cf. Borness Cave.1 It has . Thi7) C s( los s 5 pointspecime it . t No . n differs very little from fig, .8 ) 8 a littl B s ( i et 8 I natteNo e 4)pointd . D th 8 .,( an t 7 ra ., 5) NosA ( .6 e brokear n point bees f ha pinso ns t I suggeste. d that such casualties frequently occurre n sheli d l mounde extractioth n i s f shell-fisho n . There were enough limpets in Covesea to support this theory.2 No. 9 (B 7) is a type of needle very common in Scotland. An irregular holbees eha n bore slightla n di y polished e splinterBeveridgth e Se . e Collection from Nort he Nationa Uisth n i t l Museum.3 No. 10 (B 7), cf. fig. 8, No. 7. 8).—BonC ( 0 1 . eFig No s founSheath, 9 .wa e , pind9 Th . ., figNo , .9 inside it. Such sheaths are well known,4 and pins have sometimes been found inside.5 Fig. 9, No. 1 (D 4).—Fragment of a rectangular comb with chevrons engrave . Therit n e knifdo ar e e cuts betwee e teethth n straight A . - sided rectangular com s unusuali b semicirculaA . r one, also broken, and with a geometric pattern, comes from Kettleburn. It is probably

datee b o dt fairly late wit e Trapraihth Girtod nan n toilet instruments. 6 . Fla8) C t( bon3 . eFigNo ring, knifo 9 . N .e mark e visiblsar thisn eo , so it may belong to the earlier period. It has clear polishing marks.7 Fig. 9, No. 2 (C 5). Bone bead. This is well made, but not highly polished as most of those of this size from Skara Brae. Fig. 9, No. 4 (B 7).

Right upper incisor of a beaver. These are 8 reported from Loch Nevis in the fifteenth century, but their remains are not common in Scotland.9 Beads and Glass, fig. 19, Nos. 1 to 7.—Amber, Nos. 1 and 2 (07).—Mr Beck writes of No. 2 : " The beautifully made annular bead appears to me pre-Romane b o t e sam th probabld f e o an dar, o date b tw k e o e yat sth blue beads referred to later. The other amber beads are not sufficiently definite to date." No. 3 (D 3). Irregular, with a tiny hole. No. 4 (B 6). The intermediatf holo s ei e size. Ther e side,s alsei s on oa hola n o e 1 Proceedings, ff5 . ff.30 0 . ; 49 vol pp . . volpp xi . . x . 2 T. C. Lethbridge, "Shell Mounds and Winkle Pins," Archceologia Cambrensis, June 1928. 3 Also Wester Broc Keisd han s Broch, Caithness Nationan I . l Museum. 4 Broc Burrayf ho , Orkney, Proceedings,. 56 d an 5 . volpp . .xi Keller, op. cit., plate xlii., 6 and 7, p. 290. Thigh of a stork used as needle-case. 56 Catalogue, p. 237. No illustration. Cf. Comb from Skail. Proceedings, vol. xxii. p. 284. 7 Cf. a rougher ring from Archerfield. Proceedings, vol. xliii. pp. 265 ff. publishede b o T 8 alse Se o .bead s from Jarlsho Westerbrocd an f Nationan ki l Museum. 0 Identification by Miss D. M. A. Bate. EXCAVATIO E SCULPTOR'TH F NO S CAVE, COVESEA. 199 if some other ornament had" rested against it. No. 5 (D 3). Very small d irregularan (D4)conditioe 6 . Th .No . f thino s bea s rathedi r poor. Th . Thie3) holD s( larg a beasmall7 s ei . s deroughls ha No i hol . d ean y rectangular. Glass.—The report kindly sent me by Mr Beck is here given in inverted commas I hav. ee Scottis noteaddeth w n fe o sd a h materiale th l Al . weights are given by him.

14 /•**"/ ..Si V /

15 16 17

. FigBeads19 . , etc. Glassf o , , Ambert fro Sculptor'Je e md th an , s Cave. ({.) 7).—C ( 8 Lon . gNo , FigSquare19 . Cylinder.—Weight, 1'216 gramme; sp2%37, .g. . " This bea mad s di pala f eo blue glass whicn i , hI believ e colourine th g materia coppere b o t l . Bead f thio s s glass t usuallbu , y spherica r oblateo l foune ar , t Lakenheatha d , Salisbury Plain, etcI . cannot date this e beadeithe b t expec o bu t ,r t i Romat r Saxono n . The form is not common, but I have met it in pre-Roman, Roman, and Saxon times. So far as I can remember, I have never met it in this glass." Most of the square beads I have seen in England are larger. There ibrokesa n bea f thido s size (its length canno determinede b t darn i ) k blue glass from Traprain.1 Proceedings, vol. Iviii, fro level5 d . . 269mp 2n . No ,. , fig17 . 1 200 PROCEEDINGS OF THE SOCIETY, JANUARY 12, 1931. Small oblate blue5).—WeightA ( 5 bead,1 . No , '567 gramme . aboug . sp ;t 2'2. "This bead, which is of practically the same glass as the last, is similar to beads from Salisbury Plain arid other places in this country. The specific gravity, however, is lower, which may be due to hidden bubbles. Unfortunately we cannot exactly date them, but it seems probable that the e Romayar r Saxonno , when foun thin i d s country. ContinentOe nth Bohemian ,i , they dat abouo et B.C. 0 t"10 Small annular green bead f clearo (C6)9 glass, . .No Weight, '833 gramme . g.sp ,; 2'54. "A good example of ordinary bottle glass found both in Roman d Saxoan n sites. . Traprain.Gf " 1 Pentagonal cylinder n opaquei green . Weight3) D ( 0 glass,1 . , '82No 0 gramme g.. ,sp ; 2'28 . "commoA n type of bead t infrequenno , n Romai t n times , e mucwoulb t h no dearlie r but might continue to Saxon times." In appearance this is very like a bluish-green hexagonal bead foun t Traprain.a d 2 Annular beadf o clear green (06)6 glass,1 . . No Thi s madi s e fro e hollomth w rint ga havd e basvessel a en d smoothedth e sidef e o d an bee Th t an s. ncu . It is from a base,3 not a rim, as it expands horizontally in both directions; cf. a base from Traprain,4 which is called a rim in the text, but quite clearly expands almost at right angles on both sides. No (-AO)7 1 . chiA f opaqu.p o glassd ere . "Thi a glas s si s coloured with copper, whic s include i he for th f minutm o n i d e particlef o s metallic copper." Single wave bead, No. 18 (D 6). " White wave with four crests on a cobalt blue base." Weight, 2'375 grammes; sp. g., 2'419. "This bead I consider is probably an early bead. The slightly higher specific gravity (compareo t e du e b below9 d 1 y wit. ma )No h the white wave, some of the white glasses being very heavy. On the other hand, the bead is lighter in colour, and there seems to be a tendency for these cobalt beads to have a higher specific gravity when the colour is lighter. In any case it has a lower specific gravity than any Saxon cobalt bead that I have tested. But it is not on account of the specific gravity that I say it is early. The whole appearance suggests a pre-Roman date to me, although several such beads have been found with Roman remains in parts of England." Similar though not identical beads have been found at Newstead and Traprain. Large dark blue oblate bead, No. 19 (C 7). Weight, 4'233 grammes ; sp. g., 2'386. " This bead, whic mads hi e from cobalt glass I believe , is , , pre-Roman, probably

Tenewhola e L Th .e appearanc coloue th d r ean agre e much more closely ; wit e earlhth y beads than the wito ylatee d hth r Saxo o r n varietf yo same th e glass. Also, although thert sufficienno s ei t divergenc specifin ei c 1 Proceedings, Secon30. and d centurvol29, .66, livy pp. .A.D . 2 Ibid., vol. Ivii. p. 212, fig. 26, in the centre from 2nd level. 3 I owe this identification to Mr Harden. * Proceedings, vol , " .segmena liv91 . p .hollo a f o t w rim." EXCAVATIO E SCULPTOR'TH F NO S CAVE, COVESEA1 20 .

gravity to lay much stress upon it, it is interesting to note that all the cobalt blu Tena eL e beads from Englan Francd dan e tha I havt e tested have specific gravity between 2'22'34d severad 8an an , l fro Eastere mth n Mediterranea Italr n o 2'32 e yar , whils Saxoe th t n one I shav e testee dar 2Covesee -485Th . a specime t 2'38na decidedls 6i y neare earle th r y beads. I therefore think that it was made before the Roman times, though it may quite possibly have been used by Romans." Jet annular bead, fig. 19, No. 12 (D 4). This is broken, but is finely polished, with a bevel on the inside. Ther similaa s ei rNationae beath n di l Museum fro e Culbimth n Sands. " ThiBronza e beadse b appearo t e Ag .e Botm workmane o st h th - conditioe th shimateriae d th p an f no l mak placee t thaa em t i t date." Annular bead with some unknown substance (D4)inside,3 1 . . No "This bea s exceptionalldi y roug manufacturen hi heave Th .y scratches over the surface are very unusual. I have seen them on jet and stone beads attributed to the Bronze Age, but I do not see why such a bead as this madunskillemighe n b a t y eb tno d workma period.y an t Scotlann a I " d such beads are not uncommon. The present specimen has parallels in e Nationath l Museum from Culbin, Luce Bay Trapraind (D7),an 4 1 . . No . Fragmen braceleta f o t pierce r hangindfo thed gan n again brokent I . is beautifully worked. This habi re-usinf o t g bit braceletf so pendants sa s seem havo t s e been particularly commo Morayshiren i n s thera , e ar e several examples fro Culbie mth n Sands. chiA f orpimen p. o 6) B ( 1 1 t . (AsNo Fig 2, S3)19 .. This identification s suggeste wa r Bec confirmed M kan y r Balsilli db M y b d e after experi- ment. It is a bright yellow glass-like substance, which is inclined to flake away Balsillir M . e inform thae sm tnatura s thiit s si l condition and that t hithertno s ha o t i been reported from e Scotland th uses y wa db t I . Romans bot pigmenr poisonr h fo t woul fo i t d rase bu ,dan t b hazar o ht d a, guess as to its significance on the present occasion. (B4). A smaller chi orpimentf po . Suggested date of some beads:—Bronze Age.—Nos. 1, 2, 12, 18, 19. Second Century A.D.—No . 9 .Fourth Century A.D.—Nos. 15 , 10 , 8 . Lead.—Three pierced lead discs have been found suitabl r netefo - sinkers. " One has approximately I'l inch diameter and sp. g., 5'65. It s muci h corroded, which probabl specifiyw accountlo s cit r gravityfo s . It is an unusual type and difficult to date." Probably it is of the Roman period. Counters.—Numerous flat discs have been found suitabl r countersefo , of many materials—sandstone, gneiss, shell, clay. Their use has not been •determined, but they are common on many sites.1 Stone Files.—-(D 4), a broken rectangle, 4 inches by '8 inch by -5 inch. It 1 Traprain, passim. 202 PROCEEDING E SOCIETYTH F O S , JANUAR , 1931Y12 . is well made and tapers at the ends. (D 4). Flat rectangle, 5'45 by 1*5 by -35. (B 4). Broken at both ends: about 1 inch square. Iron.—(A 6). A Viking rivet. This appears to be the latest find in the cave. Cf. Rivets from Carnan More, Otternish. Samian Pottery (fig. 20).—All these pieces e mucar , , excep2 h . No t worn. Nos. 1, 4, and 5 look as if this were deliberate. The rims 4 and 5

.11

Fig. 20. Samian Ware from the Sculptor's Cave. (J.).t

see mhavo t e been ground dow polishingy nb . Pieces worn like thie sar to be seen in the National Museum from Traprain, Mumrills, and Newstead. The quality is rough and coarse. No. 1, Dragendorf 37. This shows a rather late pattern. No. 2, Dragendorf 37. The only fragment in good condition. It is a rim with a small pattern of divided . circles3 Ornamented . No . t indeterminatebu , , Dragen4 . No . - dorf 37. The rim of an even thicker pot than No. 1. It seems to have e beginninth inscriptionn a f g, o Dragendor 5 . No . . Thi33 f anothes si r worn rim. Native Pottery of the Roman Period.—a. Fig. 21, No. 2. Rim. This EXCAVATION OF THE SCULPTOR'S CAVE, COVESEA. 203 polishea s i d wheel-made oll f second-centurao y typeclae s blacyTh i . k weld an l levigated. Scattered over many. division7) C ( o t s) fro4 B m( Flat bottomed. /8. No. 3. Black, hand-made, soft ware, with indication s evertei f polishingm o s ri d e Th . and decorated with finger-marks. It produces false rim describeds sa Skarn i J a Brae. Like eth softs olli t ai , crumbly weld an , l levigated. This pot seems to have been widely scattered. The height was probably about 8 inches. y. No. 1. This is yellowish in colour. The rim is everte d brokendan , the' shoulders sloping. . FigSection21 . f Rimo s s It is soft and inclined to split off in layers like of Native Pottery from the Sculptor's Cave. (J.) a. Like a and /3 it is widely scattered. It was probably abou inche0 1 t heightn si .

CONCLUSIONS. Bronze Age.—Several archaeologists have been quite shockee th t a d suggestion of a landing in Moray Firth. " Do not you know," said they, x ha"thasFo r saiM td that ther absorptios t impositiowa e no d nan n in the north and east ? "2 Other e indignanar s behaln o t f Irelando f . "The Scottish Celte ar s always supposed to have come from Ireland." Have these supposers counte e swordEase th th p tn du Coaso s considered an t moulde dth t sa Traprain, including even a sword mould? However this may be, I have lease th t t doubno t tha landine t on ther s f foreignersgo e wa , probably many, in the Moray Firth. More excavation is required before we can estimate their extent and importance. Certain modification havy ma se taken e culturplacth n ei f thes eo e people on their way from Switzerland, but there is no evidence that these modifications took Britise placth n ei h Isles. The pottery is probably made of local clay; it has grown a little coarser, and may have lost a few nail-marks. In shape it is still com- pletely Swiss. The bracelets have still the distinguishing habit of being beaten3 from inside th e out. Several example Ziirice sth shown i h e Museunm m were almost identical. From Balmashanner to Wester Ord they show little

1 Childe, Proceedings, vol. Ixiii . 269p . . 2 Ant. Journ., vol. vii . 126 p .I hav . ideo en a what kin burial-urf do n these people usedt A . presen onle ar ye concernetw d with their cooking-pots. Thos Covesean ei e th n t leasti a , e ar , pre-absorption state. As I read Mr Fox's theory it pre-supposes the existence ol such a state. 3 C/. Grandson Courcelette 2495. sNo 6 (cast), Zurich. 204 PROCEEDINGS OF THE SOCIETY, JANUARY 12, 1931.

variet f typeyo . Not e otheth e r exotic r articlEasterou n ei n county hoards whic s arrivehha d without modification e sworth , d girdle from Brae f Gight.o s 1 The Ring-money is the only purely western product in this settlement. A curious feature is that Switzerland seem havo t s e abandoned bone instrument Middle t leassth a y b te Bronz Voug. e M Age d inclines ai an , d to think that the type represented by our Nos. 11- and 12 was confined at Bevaix at least to the earliest Neolithic. Either our people must have left Switzerland earl d acquirean y e braceletth d s later r theo , y may have bee nbackwara d tribretaineo ewh d bone instruments until the Bronze Age. There are great difficulties in the way of the inquirer after bone instruments. t appeaTheno o graveyn d i r r hoardso s , and, even when they are present in settlements, archaeologists may dislike them and throw them away. The bone instruments show a wide range of types, and they seem to have arrived unmodified. Perhaps I may be allowed to mention the moulds for two socketed e Elgiaxeth n i s Museum suggeso t , t tha f thesi t e Late Bronze Ag e people were traders they had come to stay. I should also like to mention twelve flat axes in the same Museum

from seven different sites, to suggest that the cave-dweller2 s were not the first invaders of the Firth, and Professor Childe has cited the Gallic For t Burgheaa t provo dt e that the laste y th wer .t eno worts i t I h noticing that bone instrument apparentln a f o s y foreign

type flourish at Heathery Burn, a canonical part of the invasion. 3 E ROMATH N PERIOD. Owing partlfriable th o yt e naturstilsoid e th an ll f moreo e probably to the efforts of former explorers, it has not been possible to find any stratificatio thin i n s period havee W . , however dateo tw , s established by the pottery, dated in the first half of the second century and by e coinsth 354o t fro 7 . m33 With the aid of Traprain it has been possible to group certain articles with one or the other. Second Century. Five piece Samiaf sollao n a d . nan . 2 . No fig, d .21 an , Fig20 . Tag r beltssfo . . 17 d an , FigNos, 16 , 13 .15 . 1 Proceedings, vol. xxv . 135.p . 8 Mr Callander has already noted that all the Scottish moulds for these objects come from our North-eastern Counties. Proceedings, vol. xxxviiiff7 . 48 . .pp 3 Crawford, Ant. Journal, vol. ii. p. 33. EXCAVATIO E SCULPTOR'TH F O N S CAVE, COVESEA5 20 . Bight wry-necked pins. Fig. 16, Nos.l,2,4,and6tolO. Bronze sword guard. Fig. 15, No. 2. Two glass beads. Fig. 19, Nos. 9 and 16. The hearth at the bottom of the layer. Fig. 4.

Fourth Century. The coins. Figs. 22 and 23. Two bronze pins. Fig. 16. 5 , Nosd an .3 Bronze nail-cleaners with geometric patterns. . 4 Fig d Nos, an . 17 2 . Bronze pin and grattoir. Fig. 13,No.6,andfig.l7,No.7. Bone wit e pinon h s: geometric pattern. . 2 Fig Nosd , an 18 . 1 . Broken comb. . 1 . FigNo , 9 . Blue beads. . 15 d Figan Nos, , 19 .10 , .8

feeo d l I rather strongly that non f theso e e article e reallar s y suitable for cave-dwellers. Perhaps I am prejudiced by the extreme detectinn i difficultd ha e bronzese gth w ykepe W t. losing excavation knives and entrenching tools. How could those people have kept nail- cleaner tweezersd woulo an s wh dd hav an ,e hear eth cleao t t n nailn si the Sculptor's Cave? Scottis t ric hn no littl i h site e ear s bronzes, perhaps because th e dwellings were so dark. Half a dozen is a large number for a broch, and Borness1 heads the list of caves with five articles. On the other judgn ca e froon recordse s ma th r hand fa laco ,s n f a ther, ko s ewa pottery t TraprainA .ratey an ,t a ther , massess peoplr ewa d Ou . eha only three, or at most six, cooking-pots in their two periods. Contrast this record with tha f Wookeo t y Hole,2 where there were quantitie f Celtio s c potter d seveyan n piece f bronzo s a resul s a ef o t firse th t campaign. There t leasta , cleaa s ,i r cas f occupatioeo n ovea r long period. The pottery records of the Victoria Cave, Settle, are less satisfactory,

e bronzeth t bu e mucs therar finr o cav w ho to fo e eto no mens i t I . 3 accepted that they cannot have been deposite y Celtdb s fleeing from Saxons, but there may have been other refugees in Yorkshire in the second century with Brigantine wars and Pictish raids and may be some of our Covesea men reiving southwards. Perhaps they found tagwhad t whesan no t n Newstea Birrend dan s were hastily evacuated. s alsi ot I possible thae expeditioth t f Severuo causen0 21 e th n di s ladies of the Laigh bo seek sanctuary, and that they died at Covesea 1 Corrie: Hunt, Clark, Proceedings, vol x. p. 490; vol. xi. p. 305. Wookey Hole, Archceologia, vol. Ixii. p. 565.

2 3 Victoria Cav e: Boy d Dawkins, Cave Hunting,. 81 . p 206 PROCEEDING E SOCIETYTH F O S , JANUAR , 1931Y12 .

with their trinkets upon them. t emergeI 1 d from Professor Bryce's report that there was a scarcity of adult males in the bones he examined. With regarlatee th ro dt period e cointh , s requir leso n e s explanation. Trade will hardly meet the case. The coins were ornaments rather than currency, and why here and nowhere else in Scotland? It s temptini conneco gt t them with Picts sout d ove e Bordean hth r n i r 367. Claudius tells of famous massacres in the far North two years later.2 l thiIal f s seem rememberee fancifulo b to st i t le , d that thera e ear prodigious number of human bones to be explained and that nine of them show beheadings.8 This questio s bouni n e drawingp wit th e u ddate th f hth o e n o s •walls I hope. o republist d h those already publishedd somad ed an

otherspostpono t havI t d ebu ha , e this till later. 4 It is worth noticing that Samian pottery is the last dated find on e Cav e flooth f Constantine,th f eo o r whose sign e connectear s d with

e Wemyssth wito 6s sign hd 5 oursan s . Lastly, thar DiacM t k dates the Auquhollie7 Stone 350 to 500, and one of our unpublished signs seems identical wit s (twhi h o circles joined). Professor Childe pointet ou d that it was awkward to make unconnected strangers put up memorials apotropair o c signs years afterwards. overwhelmem a y obligations I m e lisf th o t y b d . Besides those already acknowledged, I owe much to the kindness of Museum Directors d staff an t Elgina s , Aberdeen, Edinburgh, London, Cambridge, Oxford, Neuchatel, Zurich, besides counse d inspiratioan l n from archaeologists and scientists too numerous to mention.

HUMAN BONES. Professor Bryce kindly examined e humasomth f eo n bone 1928n i s . made H e sameth e observation whic s subsequentlhwa y made indepen- dently by Professor Low that there was a preponderance of children's bones. There were human bone n boti s h layers t manbu , ye morth n i e mixed laye rbronze thath n i e layer. 1 Professor Myres' pleasing suggestion. 2 Claudian, Panegyricus de Quarto Consulatu Honorii Augusti, lines 31-33. See p. 17. A local myth of cremation in the cave has arisen. It is without foundation. 3Romilly Alien, Early Christian Monuments, vol. iii. p. 130; J. Stuart, Sculptured Stones of

Scotland.4 5 Wac Jehud ean , Proceedings, vol. xlix . 237ffpp . . 6 Wemyss, ibid., . 107 p vol . . xi . 7 Diack, ibid., vol. lix. p. 257. EXCAVATIO E SCULPTOR'TH F NO S CAVE, COVESEA7 20 .

UNIVERSIT ABERDEENF YO , ANATOMY DEPARTMENT, MARISCHAL COLLEGE, ABERDEEN.

Comments on the Human Bones from the Sculptor's Cave, Covesea. In this collection the human bones are so fragmentary and mixed that it is not possible to observe any characteristics of racial significance or differ- ences between the bones of the respective layers. The large proportion of bone f younso g individual noteworthys si . ALEX. Low. 18th December 1930.

EXTRACT FROM PROFESSOR Low's DETAILED REPORT ON THE HUMAN BONES. (07).—Small wisp of human hair—reddish—made up of strands eight inches plait o long humaf Tw so . n hair—reddish. (B4).—Axis showing upper borde y f spinb o r d lamins f an of e t cu e shara p instrument. (04).—Atlas showing upper articular processes and anterior border shara y b pf instrumentof t cu . (07).—Atlas showing lower left articular surface cut off by a sharp instrument. •» (D4).—Axis showing upper border of spine and laminse cut off by a sharp instrument. (D7).—Axis with upper border of spine and upper border of right righlaminf of t t forwaracu d into odontoid process. Axis where a complete cut has gone through the laminse, pedicles, and body of the vertebra. Thes vertebrx si e a indicate thae individualth t s must have been beheaded.1

REPORT ON THE ANIMAL BONES FROM THE SCULPTOR'S CAVE, COVESEA. By R. M. NEILL, M.C., M.A., Lecturer in Zoology, Universit f Aberdeenyo . Wit e kinth h d hel f Miso p s Isobel Dean, B.Sc. I ,hav e examined this collection of animal remains. There are 536 items in the collection, which is largely the customary debris of a midden or dump.2 Dodgsor D 1 n discovered three more vertebrae similarly out. 2 This doe t includsno enumbea r of bones examine Professoy db r Ritchie. 208 PROCEEDING E SOCIETYTH F O S , JANUAR , 1931Y12 . The bone e mostlar s y much broken d includan , e remainx o f o s (Celtic ox), red-deer, roe, sheep, pig, dog d fox,an , among mammals; e boneth f severao s l gull-like bird severad an s l bones fro e heamth d l d familyco e ,th o a togethefisf o h r wit a hsingl e valv f oystero e - shell. Three bones were found a definitelshar d y heavb an p- t im ycu y plement such as an axe. In one group were five split pieces of long bones (red-deer), with remains of the marrow adhering. The bones are not of great age and on the whole are in good preservation; a few fragments are calcined and a number of others discoloured and blackened by fire. None were mineralised.

REPORT ON TWO SPECIMENS OF BRONZE FROM SCOTLAND. By MR O. DAVIES. The analyse e Imperiath st a wer e l m eColleg y carrieb f t o eou d Science, owing to the kindness of Professor Baker in putting Laboratory facilities at my disposal there. Both specimens were badly rusted thereford an , I ehav t bote epu hth figures as found and the results when scaled to 100 per cent., in order that the character of the bronze may appear. The results, therefore, wilt entirellno y represen originae th t l composition coppes a , somd an r e other thing slightle sar y soluble. (C 3 2nd) Section of Bracelet.

Original Figures, per cent. Scaled. u 41-4C 8 74-79 n S 4-42 7-97 b P 8-74 15-76 . tr . tr Zn b S -13 -23 As -30 -54 e F -39 -70 . tr . tr Ni . Bi , Co , NoAg This bronza typ f o e s thaei vers i t y commo Wesn ni t Europe before Roman times, and goes back into Bronze Age III. or IV. There should be especially noticed the high lead percentage.

To this list should be added the horse and the beaver identified by Miss Bate. Mr Neill also calle attentio1y dm workee th o nt d deer tine.—S.B. EXCAVATION OF THE SCULPTOR'S CAVE, COVESEA. 209

(C 5) Needle.

Original Figures centr pe , . Scaled. Cu 70-57 90-35 Sn 6-78 8-68 b S -29 -37 b P -14 -18 e F -32 -41 Co -01 -01 , Zn. Bi Ni , , As , NoAg An ordinar n bronzeyti . There shoul e noticeb d e fairlth dw lo y percentag tineof , wel l cent.beloper ,w10 whic h woul normadbe for l pins and such objects that did not require to be very hard but which t i woul surprisine db fino t g d r afte e Earlfo th e r yus Bronzn i e Ag e DAVIES. O . swords, etc. Mr Davies notes that the percentage of lead at Covesea is less than socketee thath f o t d axe t Heathera s y Burn Cave, e 24'r cent.h 3pe d an , does not think that the two have the same source.

. ' ' ' COINS;' There are nine Roman coins dated nearly to the middle of the fourth o 354t century f thes7 O . e limit33 e Th e coin. ar s s eight show no sign weaf s teao d thre d rpiercedan e ran ear . Roman Occupation.—These two facts go to show that the coins were not the ordinary small change of the district. Therefore they were probabl t somya e distance fro n occupiema d Roman settlement. This conclusion is supported by the presence of five scraps of Samian pottery from four different bowls which have also been put to a new use.1 The olla alone would sugges nativa t e potter Roman yo n lines withie nth travelling radiu f sucso h pots uniques i , butt i s ,t cannoa i , takee b t o nt prov eRomaa n occupatio Morayshiref no ,secone eveth n ni d century. e barbariaOth f n coins t seemi , s possibl discuso et s seventy-nint eou of some two hundre d twentyan d . Their condition when founs wa d deplorable rathes i t i a triumprd an , h r EdwardfoM r s tha mano s t y legiblee ar . Those determinee whicb n hca d were confine e followinth o t d g five types:—Gloria Exercitus, 3; Felicitas Temporum, 54; Victoria Augusti; ,3 Roma, 2. , As at Lydney and Richborough, the Felicitas Temporum type is predominant. 1 See p. 202. 4 1 VOL. LXV.

212 PROCEEDINGS OF THE SOCIETY, JANUARY 12, 1931., Mrs Wheeler, besides allowin r o set collectiohe ee m g n before publication s mosha , t kindly alloweo studt r manuscripte he ym d . Our coin e largerar s , less fragmentary, less degenerate madd an ,n o e differena t principle smallesr Ou .bigge s i , t coin59 . r tha,No Lydnee nth y Minimissirni, and only about three of them -were found at Covesea. I am only certain of clipping in one instance, No. 55. t LydneA ycompleto n ther s ewa e coin foundgooa d dan , manf yo our coins have nothing missing. We have some variety of types. Our coins are degenerate, but their idiosyncrasies have a certain attraction. Very few are mere nonsense, like so many of the Lydney n exceptioa s i 8 4 coins . n (NosNo . . 44-7 t published)t Elgin9a ye t no , . Mr Casey explained to me that a large number of the Lydney coins are made by chopping up a rod of bronze. This causes impurities to concentrat e middleth n ei , wher e coith e n begin degenerateo t s r Ou . coins all split up the sides, showing that they were stamped out of a plate of metal. Consequently there seems to be a good deal of difference in date between the two finds. Mrs Wheeler places her coins early in the fifth century. Mr Pierce thinks that our coins do not differ much in date from their prototypes, the last of which is dated 354. He also considers that they hav a ehomogeneit f theiyo r own—that they for mgroupa . The knowledge I have gained from describing them inclines me to think that he is right—that I should be able to distinguish one of the Covesea group from a Bichborough coin of the period if I found it in discovey da barrowa e on hopry f thiI y .trues ma esi ma e tha ,e w w t from what English county our Caledonians filched them. My thanks are especially due to Mr Pierce for his help with the barbarous coins. The Roman coins were identified by Sir George Macdonald.

Roman Coins:— Figs. 22 and 23. 1. (BO) IM CAE MAGNENTIVS AVG. Magnentius. FELICITAS [REIPVBLICE]. Magnentius standing. In ex., TRP. Coll., viii. p. 9, No. 5. This coin has had some wear, but not much. 2. (C8 CONSTAN[TIVN D ) . ConstantiuC] S B IVNNO s Gallus. PEL TEMP REPA[RATIO], In ex., CSL[C]. . 25 Coh.. No , , viii32 . .p In excellent condition t seemI . s almost new. Someon starteds eha boro t eholea t abandonebu , efforte dth . 3. (C7) [CONSTA]N[TIVS] P F AVG Constantius II. VICTORIA n ex.I E . [DDJAVG, SARINN Q C . GC Cf. Coh., vii. p. 484, No. 293. Pierced. Thi somd s ha stil s ecoii s usaget li nha quitt bu , e sharp. EXCAVATIO E SCULPTOR'TH F NO S CAVE, COVESEA. 213

4. ) (ROM4B A Perio f Oonstantindo . eI Wolf and twins. Coh., vii. p. 330, No. 17. Very much rubbed, but the type is quite clear. Pierced. 5. (A5) [FL IVL CONSTANTI]VS NOB CAES. Constantius II. GL[ORIA EXERCITVS]. Coh., vii. p. 455, No. 92. This is broken round the edges, but it seems to have been quite new when lost. ) [CONSTJAN3 C ( . 8 AVGF SP[ ] Constan. sI VIC[T]0[RIAE DD] AVGG Q NN. In ex., TRP. Coh., vii. p. 431, No. 176. This is splitting up the middle, but the lettering is quite new. . Not illustrated. 9. (C3) CONSTANTINVS IVN NOB C Constantine II. GLORIA EXE[RCITVSJ. Coh., vii. p. 377, No. 114. Good condition. MAGNENTIVN D . ( A) 10 6 F AVG P S . Magnentius. VICTORIAE DD NN AVG ET CAES. In wreath, VOT V MVL X. In ex., AMB. . 68 Coh.. No , , viii19 . p . This is another new coin. 43. (B 4) Helmeted head of Roma and possibly ROM visible. Like the other Roma coin, thi mucs i s h wornpierces i o t hunI d tw . dan n go wire rings, the ends of which overlap. Barbarous Coins— 6. (C 5) Gloria exercitus type. Diademed head imitates the Constantius type; cf. No. 2 or 5. * Ther a poo s ei r attemp produco t t elegenda reliee Th s i .f lowdesige th , n clear though weakly executed. Splitting. Sam) 4 B 7( e. type. Diademed head mucA . h robuster version tha laste nth . reverse . 9 WornTh . s nearerei No , o brokent d splitan ,o n ; lettering visible. 11. (C 5) Same type; reverse nearer to No. 5. It has achieved CO in the right place, but the end of the inscription, though visible s i illegible, e heaTh d. seems Constantinian. reversee Th , though lumpy vigorouss i , . 12. (A 5) Fel. Temp. The diademed hea vers di y prototypes closit o et . Reverse. — Sign f lettero s se correctb whic y ma h. Design vigorous. Condition good. The coin was all there and only slightly rubbed. Hea) 5 13 f D Constantind( o . I N . veiledeI . . . . . Constantine in a chariot. Doubtful letters. (Cf. Cohen, vii . 318 .p . 760No , . Mattingly, Roman Coins, PI. Ixiii., No. 6, p. 299.) 14. Fel. Temp. A neat little head like No. 11 V S P E Reverse.—An excellent charge LE r TEM Ther jaggea s ei dmiddlee holth n ei . 15. (A 6) Fel. Temp. Head seems bearded. STA. TIVS. Even the apices are clear. In ex., CPLG 214 PROCEEDING E SOCIETYTH F O S , JANUAR , 193112 Y .

16. (B 3) Fel. Temp. Thin Coustantinian head. The reverse is faint but spirited. 17. (B 4) Fel. Temp. reverse Th e an weldo g obvers t l togetherno o ed obverse Th . e lack sfats neci chid d .kan nan The revers nearls ei y complet figuree th neate d sar e an thint , ino f . Fel) 4 B .( Temp. 18 . This woul e complet r ath ds fa e sees b a mo 7 t e1 versio. No f no concerns the obverse. The diademed Constantinian head has become marvellously hairy. 19. (B4) Fel. Temp. Face going off right. Reverse.—The legionar s becomweayw ha fe k a estraggle n so s e horsrume left hi s s losTh th head s ha hi t .i ep t bu , magnificent. Fel) 4 20B .( . Temp. Obverse broken. Reverse.—Lumpy. Spear and falling horse visible Ex. U V 21. (B4) Fel. Temp. Diademed head. This fac n intermediatea e seemb o t s e type between the last three and No. 22, which is recognisably Constantinian. Reverse.—Lumpy, horse going off left. Letters ? 22. (C3) Fel. Temp. Constantinian gravita becoms sha regulaea r photograph stare. Reverse:—The legionary is reduced to legs. 23. (B 4) Fel. Temp. This head is similar to the last, but there is less of it. All that remains e revers wa e horseth e th f h t o ss I seem ei f i . s a s consciously being use separata s da e design coinee Th .r meant to make a horse's head framed in dots; at the last minute he squeezed in the rest of the animal. Vie) 4 24B .( .Aug . Hardly shows -any signs of degeneration. Diademed head. AVG in good lettering. Reverse.—The right-hand Victory has been left out by careless striking. Pel) 4 25A .( . Temp. REP obverse Th goods ei . Reverse.—The legionar vanishins yi g heavenwards. Pierced. Pel) 6 26.A ( Temp. . This head is barbarous, but belongs to the family. Reverse.—Much blurred.S AV V I GT " 27. (B4) Fel. Temp. This is a very bad head. An extra nose has slipped under the chin. I V Reverse.—The e horsrum th somd f an epo e straggling legsI . believe tha tant a bee s gha n deliberately hammered ou thif to s coin, so that it might be attached as an ornament. Cf. Du Chaillu, vol. ii. p. 313. The Viking Age. Similar bronze wire beads to those there shown have been found in the cave. See p. 193. EXCAVATION OF THE SCULPTOR'S CAVE, COVESEA. 215

28. (3 A) Pel. Temp. Diademed head, same type as 23. Reverse.—Lumpy treatment; most of legionary lost. 29. (Fel) B5 . Temp. This seems to aim at 24, taut only the back of the toga remains. Reverse.—The Briton is falling off the best horse of the collection. He is a Celtic (?) triumph F E L E in the field. 30. (B3) Fel. Temp. Diademed and civilised head. Reverse.—The legionare legth f so agaie yar n give place nth f eo honour s tunic skire rume hi s spearth f Th f o thi ,.d p o an , horse th alse ear o visible. 31. (Not stratified) Fel. Temp. Diademed head with the hair stylised. Reverse.—The treatmen weas i tskeletald kan e t mosth bu ,f o t desig stils ni l lineheree Th s .\mde horse'e rth s taihare lar d to explain. 32. (A3) Fel. Temp. Constantinian head. The reverse design seems good, but the coin is much decayed. REP ) 6 BarbarouA ( . 33 same th s f eheado genera t . Onl,bu 38 . yl No type . Cf . coine th n .face o th s ei Reverse.—Lost. 34. (Not stratified) Roma. The wolf and twins are just visible. Pierced and in very bad condition. 35. (A 4) Fel. Temp. Constantinian head, all present. R.S. T. Skeletal, Minoan-like reverse coie nTh . seem havo st e been e hammereon t a t dou end as if to facilitate attachment. 36. (B3) Fel. Temp. Obverse much better than reverse. Hatchet face. Thick heavy jaw. He has the air of a philosopher. Reverse.—The shield on the right has grown enormous. The legionary's lege weedye th e horssar rume th d Th f ean p.o legs of his rider are visible. 37. (03) Fel. Temp. The details of this head, especially of the chin, are very poor, straggling. I V In the reverse the design entirely fails to distinguish between the horse and the falling horseman. 38. (05) Fel. Temp. Rather hairy Constantinian head. The reverse is much more sensible in the photograph than in real life. Split. 39. (4B) Fel. Temp. Low relie left e ; heaf th , o dt nearl y worn away. Reverse.—Severe t thos disintegratiobu , e in slantin t se s gha n lines probably represen legionarye tunie th tth f co therd an , e may be the legs of the horse to the left. 40. (B5) Fel. Temp. Fairly neat diademed head to left, but the nose appears to be snub. C O N C AN 6 21 PROCEEDING E SOCIETYTH F O S , JANUAR , 193112 Y . Reverse.—The design appear havo st e been complete coie th n t ,bu is worn. Pierced. 41. (B3) Fel. Temp. Diademed head wit hlona g neck . RlchborougCf . . hII Plat f eo coins. A V G Goo completd dan e reverse. 42. (B 3) Pel. Temp. Good diademed head. Reverse.—Gigantic legionary on the left; unfinished ; hole in the middl horsee th f eo .

NOTE. e coin e havTh w s e been dealing with have been fairly complete, and what remains is extremely decadent. The later examples (not yet Lydnee publishedth f o yy coinsgoine wa tha n ar )I e .g th t collectiof no o e coinsn ove th d thousana r an , fo g bi d o e flangecoinsto th l e al ar s, single desig t seriecompletes ni fa r collectio e ou s th firs e n f th I o t. n ni (Nos. 17-20), the gentleman has lost his chin. Degeneration, like old age s recal,u e leane t seemth lth Le r . so bount fa inclino e t d th o t e pillow-like horses of Late Mycensean seals. . These published horses specialise in enormous rumps. There is another breed of horse in the collection which fades away into a curve, e.g. No. 57, and ends as a straight line, as in No. 48.