uluburun pulak aja 1988

The Bronze Age Shipwreck at Ulu Burun, Turkey: 1985 CampaignAuthor(s): Cemal PulakSource: American Journal of Archaeology, Vol. 92, No. 1 (Jan., 1988), pp. 1-37Published by: Archaeological Institute of AmericaStable URL: http://www.jstor.org/stable/505868Accessed: 05/12/2009 13:42

Your use of the JSTOR archive indicates your acceptance of JSTOR's Terms and Conditions of Use, available athttp://www.jstor.org/page/info/about/policies/terms.jsp. JSTOR's Terms and Conditions of Use provides, in part, that unlessyou have obtained prior permission, you may not download an entire issue of a journal or multiple copies of articles, and youmay use content in the JSTOR archive only for your personal, non-commercial use.

Please contact the publisher regarding any further use of this work. Publisher contact information may be obtained athttp://www.jstor.org/action/showPublisher?publisherCode=aia.

Each copy of any part of a JSTOR transmission must contain the same copyright notice that appears on the screen or printedpage of such transmission.

JSTOR is a not-for-profit service that helps scholars, researchers, and students discover, use, and build upon a wide range ofcontent in a trusted digital archive. We use information technology and tools to increase productivity and facilitate new formsof scholarship. For more information about JSTOR, please contact [email protected].

Archaeological Institute of America is collaborating with JSTOR to digitize, preserve and extend access toAmerican Journal of Archaeology.

http://www.jstor.org

http://www.jstor.org/stable/505868?origin=JSTOR-pdf

http://www.jstor.org/page/info/about/policies/terms.jsp

http://www.jstor.org/action/showPublisher?publisherCode=aia

The Bronze Age Shipwreck at Ulu Burun, Turkey: 1985 Campaign*

CEMAL PULAK

Abstract Excavation of a Late Bronze Age shipwreck, tentative-

ly dated to the 14th century B.C., was continued by the Institute of Nautical Archaeology at Ulu Burun near Kas, Turkey, in 1985. New finds included more copper, tin, and glass ingots; Mycenaean, Cypriot, and Near Eastern pottery; bronze tools, including axes, adzes, chisels, drill bits, and tongs; bronze weapons, including swords, a dirk, a dagger, and arrowheads; balance-pan weights in a variety of materials and shapes; gold and silver jewelry, some as scrap; a scarab, a stone plaque, and a fragmentary gold ring inscribed with Egyptian hieroglyphs; beads of stone, faience, amber, and bone; fragments of faience rhyta; shell rings; and a globed pin of the type usually dated to the end of the Mycenaean period and later. The east-west route of the ship postulated on the basis of 1984 finds seems certain, but the nationality of the ship remains elusive.

During the summer of 1985, the Institute of Nauti- cal Archaeology (INA) continued excavation of a Late Bronze Age shipwreck off Ulu Burun, near Kay, in southern Turkey.' The first campaign, in 1984, had

already shown the site's extraordinary richness and diversification of finds. The bulk of the cargo, com-

prising primarily raw goods, included copper and tin

ingots in the so-called "ox-hide" and "bun" shapes, glass ingots, orpiment, Canaanite amphoras filled with resin, unworked elephant and hippopotamus ivory, fruits and probably other foodstuffs. Manufac- tured trade items were limited to Cypriot ceramics and beads of various materials. The portion of bronze tools and weapons representing commercial commodi-

ties, shipboard items, or personal possessions can be estimated only after the site has been completely exca-

* The excavation was financed by the Institute of Nauti- cal Archaeology (INA), the National Geographic Society, and the National Science Foundation. The excavators were Cemal Pulak, acting field director and assistant project director; Donald A. Frey, photographer; Robin C.M. Piercy, chief of operations; Tufan Turanli, captain of the Virazon; Murat Tilev, engineer; Yancey Mebane and Karl Ruppert, physicians. Staff also included archaeologists Douglas Hal- dane, Michael Halpern, and Faith Hentschel; Texas A&M graduate students William Lamb, Ralph Pedersen, and Ste- phen Vinson; ethnobotanist Cheryl Haldane; illustrator Netia Piercy; and conservators Jane Pannell and Robert Payton. Askin Cambazoglu represented the Turkish Gener- al Directorate of Antiquities. I would like to express my

vated. Based on Mycenaean ceramic evidence, the

shipwreck was dated broadly to the 14th century B.C.,2 and comparison of artifacts from the site with those found on land sites pointed to a sailing route from east to west. That the ship was carrying major resources of Asia and Cyprus seems certain; the na-

tionality of the vessel, however, remains elusive. With less than half of the site excavated, Ulu Burun

has already supplemented our knowledge of trade, es-

pecially in raw materials, gained from the approximately contemporaneous Amarna tablets, Egyptian art, and finds of similar materials discovered from sites on land. The cargo of the Ulu Burun ship, by far the largest single deposit of Bronze Age raw materials

discovered, has yielded the earliest known ingots of tin and glass. Complete excavation and study of the several hundred copper and tin ingots are expected to eluci- date greatly the routes and mechanisms of supply in the Eastern Mediterranean metals trade during the latter half of the second millennium B.C. The impres- sive collection of bronze weapons, tools, and other ob-

jects aboard is a welcome contribution to our know-

ledge of a period relatively meager in bronzes from datable contexts. The association of Cypriot, Canaan-

ite, Mycenaean, and Egyptian artifacts in a closed de-

posit not only provides new insight into problems of

chronology, but also facilitates cross-dating, and sup- plements our understanding of the relations between the Levant and the Aegean at the onset of a period of unrest and great changes.

The Ulu Burun ship is the earliest seagoing vessel ever found. Only a small section of the ship's hull has

gratitude to Prof. George F. Bass, project director, for his advice and constructive criticism, and to Ann Bass for edit- ing this manuscript.

The discovery of the site in 1982 is described by C. Pu- lak and D.A. Frey, "The Search for a Bronze Age Ship- wreck," Archaeology 38:4 (1985) 18-24. Artifacts raised for dating and identification purposes in INA's 1983 survey are reported in G.F. Bass, D.A. Frey, and C. Pulak, "A Late Bronze Age Shipwreck at Kay, Turkey," IJNA 13 (1984) 271-79. For the 1984 campaign, see G.F. Bass, "A Bronze Age Shipwreck at Ulu Burun (Kay): 1984 Campaign," AJA 90 (1986) 269-96.

2 Bass (supra n. 1) 285, 293.

American Journal of Archaeology 92 (1988) 1

CEMAL PULAK

Fig. 1. Upper half of wreck seen from east

been uncovered and that only briefly examined under water, but our knowledge of ship construction has been extended already by nearly a millennium. We now know that in the late second millennium B.C. there were ships in the Eastern Mediterranean which were built by an edge-joined technique similar to that used in later Greek and Roman hulls. Although the

stone weight anchors from Ulu Burun are similar to

examples found on land sites and in isolation under water, they are the first to be associated unequivocally with a Bronze Age ship.

The second excavation campaign began on 15 June 1985 and continued until the end of August. Half the

2 [AJA 92

THE BRONZE AGE SHIPWRECK AT ULU BURUN, TURKEY: 1985 CAMPAIGN

N 0 P

Fig. 2. Upper half of wreck. Objects mentioned in text are indicated by their K(as) W(reck) numbers. KW 200,201, and 237 are not shown.

staff again lived in a camp built onto the southeastern face of the rock promontory, while the other half lived aboard INA's 20-meter research vessel, Virazon, moored above the wreck.

Since the areas to be excavated in 1985 had been mapped in 1984 by photogrammetry and by triangu-

lation with meter tapes, we could begin excavating as soon as we had installed on the site our telephone booth (an air-filled plastic dome for diver safety) and air lifts (suction pipes for removing overburden). Newly uncovered objects were measured from the previously established datum points and plotted on the

3 1988]

CEMAL PULAK

plan. New datum points were fixed as needed, all carefully tied to the master points.

Diving twice daily, with nearly six hours between dives, six days a week, we compiled 91 dives to between 50 and 140 ft. (15 and 43 m.), 908 dives to between 140 and 150 ft. (43 and 46 m.), and 183 dives to between 150 and 160 ft. (46 and 49 m.), totalling 428.4 hours on the wreck in 1985.

THE WRECK

The sketch plan of the site made during the 1983 survey shows wreckage scattered over an area approximately 10 x 18 m.; the shallowest recorded objects are 41 m. deep, and the deepest about 51 m.3 Only about a third of this area, mainly its shallower reaches, was excavated in 1984, and only partially in most places. A primary objective of the 1985 campaign was to complete excavation of this uppermost region (fig. 1). The wealth of finds and heavy encrustation encasing the delicate material there, however, pre- vented the realization of our goal. Meticulous dissec- tion of the rock-hard matrix with chisels and hammers was extremely slow; small, hand-held pneumatic chisels were abandoned because they lacked the necessary control for extracting fragile items. The problem was exacerbated in 1985 when areas previously designated as bedrock proved to be tin ingots interspersed with other artifacts. The situation was especially complicated by the corrosion of the tin into amorphous mass- es indistinguishable from calcium carbonate deposits and bedrock. When tin was encountered by excavators, its immediate area was cleared and the borders of the ingot defined for excavation and removal. Work in this upper area by three divers for more than two months resulted in expanding the known boundaries of the wreck by an additional 2 m.2 (fig. 2).

An area about half a meter square (L-M intersec- tion at 11) was especially rich in finds. Once its crust of corroded tin and encrustation had been lifted, a small pocket of loose sand yielded artifacts including a mace-head (KW 278); bronze arrowheads; bronze tools, including three sickle blades, two chisels and a drill bit (KW 579: fig. 17); lead fishing-net weights; a large lead weight, perhaps the ship's sounding lead (KW 267: fig. 41); bronze balance-pan weights, including a lead-cored example decorated with shepherd and sheep figurines (KW 582: fig. 37); a gold flower (KW 361) and roundel (KW 551: fig. 32), both with granulated decoration; faience beads; a gold- framed bone or ivory scarab (KW 338: fig. 34) and a plaque of greenish stone (KW 481: fig. 35), each

carved with hieroglyphs; and a unique copper ingot (KW 388).

A larger deposit nearby in grid squares M-10, M- 11 and N-10 also yielded an abundance of objects: numerous lead fishing-net weights; terra-cotta lamps (KW 485: fig. 6, KW 502); a stone mace-head (KW 486: fig. 26); a bronze dirk (KW 296: fig. 23); arrowheads; a pair of tongs (KW 378: fig. 19); chisels (KW 536, 566); an adze blade (KW 576: fig. 12); part of a gold signet ring with Egyptian hieroglyphs on its bezel (KW 603: fig. 33); balance-pan weights of unidentified stone, hematite, and bronze or copper; and copper ingots in both bun and ox-hide forms. A large number of tin ingots in quarter ox-hide pieces, as well as half a tin ox-hide ingot (KW 644: fig. 5), were also recovered. Four other ingots, probably tin, appear to be cast in the bun shape (KW 401, 409, 519, 642).

Directly below the deposit was a relatively flat shelf thought in 1984 to be bedrock. Investigation in 1985, however, revealed this to be an encrusted layer of Ca- naanite amphoras. The positions of the amphoras indicate that they were arranged in neat rows before the ship broke up. Their downward slippage was inhib- ited by what initially appeared to be a terrace-like rock ledge. This proved to be a neatly stacked mass of at least 20 copper ox-hide ingots.

This ledge area, which yielded nearly all the glass ingots and many other finds during the previous campaign, now appears to be much deeper than previously assumed. It produced more glass ingots, lead fishing-net weights, a stone balance-pan weight, more metal ingots and Canaanite amphoras, a bronze sickle blade, a bronze arrowhead, and bronze pin KW 570 (fig. 36).

Removal of the intrusive Byzantine anchor in area N-11 revealed the ironic cause of its loss; its fluke had caught on several copper and tin ingots.

On either side of the heavily concreted ledge area are sand-filled gullies where work progressed more easily. The northern gully, virtually devoid of artifacts in its upper reaches, save for a few lead fishing-net weights, a bronze drill bit (KW 381: fig. 18), and a large disc-shaped stone balance-pan weight (KW 382), stretches downward along the sloping sea bed. The large rock outcrop in the middle of the site forms the gully's southern flank. Stone anchors and four- handled ox-hide ingots are visible at the gully's middle and lower reaches, while scattered ingots and a single stone anchor mark its deepest extension.

The excavation of the southern gully, at whose upper end rested small pithos KW 250, had yielded sev-

3 See sketch plan in Bass et al. (supra n. 1) 274 fig. 3, and Bass (supra n. 1) 271 ill. 2.

4 [AJA 92


eral of the more exciting finds of the 1984 campaign: elephant and hippopotamus ivory; gold, silver, faience, and amber jewelry; bronze weapons; Myce- naean, Cypriot, and Syrian pottery; a Mycenaean seal; and a gold chalice among them. Work in 1985 was concentrated on expanding and fully excavating the upper extension of this gully. The pithos, wedged between and concreted to bedrock, had to be chiseled free before it could be raised; the silt inside was saved for identification of possible organic contents, but already we know that it contained 21 lead fishing-net weights, a small pilgrim flask (KW 438: fig. 7), and an oil lamp (KW 437) identical to the lamps found in the pithos (KW 251) raised in 1984. Directly under

pithos KW 250, buried in deep sand, emerged a well-

preserved short sword (KW 301: fig. 21), apparently of Aegean origin. A second, equally well-preserved short sword (KW 275: fig. 20), seemingly Near East- ern in origin, surfaced next to a silver plate deliberately crumpled for scrap and 10 bronze arrowheads about a meter further upslope. The same area produced, among other objects, two centrally pierced lead discs (KW 298: fig. 39a, KW 459) and a long bronze spearhead (KW 261). Downslope of the pithos, a large deposit of several hundred fish bones was uncovered. Complete skeletons of small fish are represented, but whether they are the spilled contents of a container or the accumulated excrement of a large predator once living in the pithos is still under investigation.

A similar deposit of hundreds of murex shell opercula or "doors," identified as belonging to the species Phyllonotus trunculus L. by John Taylor of the Gas- tropod Section in the British Museum of Natural His- tory, was found trapped between several copper ox- hide ingots. The operculum, a horny plate which provides protection by closing off the open end of the shell

4 Personal communications with David S. Reese. 5 P. trunculus was one of the species utilized extensively

in the shell purple-dye industry of the ancient Mediterra- nean. Mounds of crushed purple-dye producing shells have been found at a number of sites. These production centers used species available locally; see D.S. Reese, "Industrial Exploitation of Murex Shells: Purple-Dye and Lime Pro- duction at Sidi Khrebish, Benghazi (Berenice)," Libyan Studies. 11th Annual Report of the Societyfor Libyan Stud- ies (1979-1980) 79-86, for a detailed account. A listing of purple shells from archaeological deposits in the Eastern Mediterranean is in D.S. Reese, "Palaikastro Shells and Bronze Age Purple-Dye Production in the Mediterranean Basin," BSA (in press). At Sarepta on the Lebanese coast, crushed shells of P. trunculus were deposited in a refuse pit of the 13th century B.C. (J.B. Pritchard, Recovering Sarep- ta, a Phoenician City [Princeton 1978] 126-27 with ills. 121-22). For a possible purple-dye installation comprising a series of pools cut in bedrock designed to receive both fresh

after the animal has withdrawn, seldom survives expo- sure. In this case, however, they undoubtedly were

preserved by copper impregnation. Apparently this is the first time such opercula have been reported from an archaeological context.4 At first, it was assumed that they came from murex snails which had died under the ingots, but the scarcity of shells in the immediate area, their presence between ingots in spaces too narrow to admit living animals, and the apparent stacking of three and even four opercula suggest that the deposit was the remains of a shipment carried in a bag. They probably represent a by-product of the purple-dye industry,5 for it is unlikely that these murex shells were collected for their opercula alone.

The medicinal uses of murex opercula have been noted by Pliny,6 but certain gastropod opercula also were used in antiquity as incense or as an incense in- gredient; under the name onycha (Greek onyx), opercula even constituted one of the ingredients of the Holy Incense.7 The purpose of the opercula recovered from the Ulu Burun shipwreck is unclear, but it may be that they represent a shipment of incense.

The lower mid-region of the gully yielded more spearheads, a crushed and poorly preserved one-handled tin cup, and several faience, amber, and stone beads.

Beads were found in far greater numbers in the area which fans out into a sloping field of sand directly below the gully. Here also were balance-pan weights of the standard sphendonoid and domed shapes, as well as three zoomorphic forms (fig. 37) comprising a duck (KW 350), a bull (KW 335), and a sphinx (KW 468). Close by, a Mycenaean stirrup jar (KW 308), ca. 0.20 m. in diameter, was found in fragmentary condition. A smaller stirrup jar (KW 305: fig. 8), equally fragmentary, lay about a meter upslope of the

and salt water, see A. Raban, "Some Archaeological Evi- dence for Ancient Maritime Activities at Dor," Sefunim 6 (1981) 20-21; A. Raban and E. Galili, "Recent Maritime Archaeological Research in Israel-A Preliminary Report," IJNA 14 (1985) 341-43. It recently has been affirmed that the blue-tinted variety of purple or hyacinthine purple (Latin purpura hyacinthea; Hebrew tekelet) was prepared exclusively from P. trunculus, and that this species was pro- cessed separately from other purple-producing species; this explains the enormous mounds of its shells near Sidon, separated from a second pile containing shells of other species (I. Ziderman, "First Identification of Authentic Tekelet," BASOR 265 [1987] 25-33). That the opercula from the Ulu Burun shipwreck represent only those of P. trunculus may, therefore, be more than a mere coincidence.

6 Pliny, HN 32.120; and Reese 1979-1980 (supra n. 5) 85, citing Pliny.

7 Exod. 30.34; I owe this reference and the suggestion of incense to Irving Ziderman.

5 1988]

CEMAL PULAK

first, next to the tin cup. Still farther downslope, in area L-15, were a poorly preserved bronze dagger (KW 621: fig. 24) and fragments of a faience rhyton in the form of a ram's head (KW 565); a unique am-

phora of exceptionally large capacity (KW 588) was found next to the rhyton and may have crushed it. Two pithoi, still in situ, lie almost against the western side of the rock outcrop.

The upper section of this sandy area (grid squares 1-13 and J-13 to J-15), which produced a Mycenaean stirrup jar and a gold roundel in 1984, now has been excavated completely, yielding several more amphoras, two broad-chisels (KW 423, 424), a Mycenaean cup (KW 334: fig. 9), several amber beads, and fragments of a large pithos.

The sandy area also extends to the north and east around the rock outcrop. The northern extension, containing several stone anchors and the pithos raised in 1984 (KW 251), merges with the northern gully. To the south, however, the sandy slope drops steeply into deeper water where a partly buried pithos is visible at a depth of 51 m.

Because both campaigns were devoted primarily to the excavation of the shallower part of the wreck, the true extent and nature of the scatter in deeper parts remain unknown. It would be premature, therefore, to

assign stern and bow areas to the site. When the com-

pleted excavation allows these determinations, and the final site plan reveals artifact distributions, we should be able to speculate with greater confidence about which items were cargo and which were the personal effects of those on board. Until then, however, it must suffice to note that the west-east orientation of the wreck is along the gradient of the sea bed, and that three rows of copper ox-hide ingots appear to be positioned athwartships with a row of anchors, possibly spares, arranged in groups between them. At each end of the ship and beyond the last row of ingots, there

appear to be shorter rows of copper and tin ox-hide

ingots.

8 The belief that these four-handled copper ingots were cast to imitate dried ox skins, and that each ingot was equal in value to the price of an ox, has been invalidated (Bass [supra n. 1] 275; J.D. Muhly, "The Copper Ox-Hide Ingots and the Bronze Age Metals Trade," Iraq 39 [1977] 81 with n. 52).

9 H.-G. Buchholz, "Keftiubarren und Erzhandel im zweiten vorchristlichen Jahrtausend," PZ 37 (1959) 4-8, fig. 2. Based on a poorly preserved, incomplete ingot raised in 1982, the Ulu Burun copper ingots had been initially identified as of Buchholz's Type 2 and perhaps Bass's Type 2c (Bass et al. [supra n. 1] 273 fig. 2). Most of the examples

THE FINDS

Ingots The bulk of surviving cargo on the Ulu Burun ship

consists of copper ingots of two major types: those in the so-called ox-hide shape;8 and plano-convex, discoid or bun ingots. A unique, small rectangular slab was also recovered.

By far the most numerous ingots are in the ox-hide shape of Buchholz's Types 2 and 3,9 flat, oblong ingots with protrusions, or handles, at each of their four corners. One surface of each ingot is always rough, while the other is much smoother. Seventeen such ingots were excavated and raised in 1984, in addition to one recovered in 1982 when the wreck was discovered.10 With nine recovered in 1985, the ox-hide ingots raised to date total 27. It remains impossible to estimate their number on the wreck. A preliminary count of some 84 visible ingots in 1983 suggested no less than 150 in ox-hide form, but in 1985 we realized that what in one case we had assumed to be a single row of seven ingots was in reality a pile stacked at least eight ingots deep in places.

None of the ox-hide ingots have yet been cleaned of marine encrustation to reveal possible primary or secondary marks. That such marks exist, however, is shown by ingot fragment KW 632, found near a stone anchor in area L-14. The unobscured mark II, not represented in the compendium of ingot marks in the Cape Gelidonya publication,1' appears on the ingot's rough, upper surface.

The weights of most ingots have not yet been determined because of heavy encrustation, but their dimensions reflect the same degree of variability seen during the previous campaign. Out of nine intact ingots only five allowed proper measurements. The largest (KW 404) is 0.82 m. long and 0.44 m. wide, with a maximum thickness of 0.06 m.; it weighs ca. 28.8 kg. The smallest (KW 628) is ca. 0.73 m. long and 0.37 to 0.40 m. wide, with a maximum thickness of 0.05 m.; it weighs 26.4 kg. The only other ingot that has been

raised in 1984 and 1985, on the other hand, approximate more closely Buchholz's Type 3. The wide variety of ingot shapes at Ulu Burun now demonstrates gradations between the two types, making distinctions between them most sub- jective. It seems that all three of Buchholz's types are present on the site, but we caution that most of the ingots remain to be raised, conserved, and studied.

10 Bass et al. (supra n. 1) 271, 273 fig. 2; Bass (supra n. 1) 276.

11 G.F. Bass, Cape Gelidonya: A Bronze Age Shipwreck, TAPS 57, pt. 8 (1967) 72 fig. 90.

[AJA 92 6


cleaned of concretion, raised in 1984, weighs 17.9 kg. (KW 184).

Some ingots suffered more from corrosion than others. In many instances they have been reduced, at their edges and especially on their handle-like protrusions, to a dark brown porous substance, devoid of metal, that yields at the slightest force. It is impossible to determine the original dimensions of ingots in this condition, and in many cases even to distinguish upper from lower surfaces.12

Until the discovery of the Ulu Burun ship, known full-sized ingots of Type 3 were limited to two intact and several fragmentary examples from Cyprus. Al- though none had been dated by stratigraphic excavation, some were attributed to about the 12th century B.C. on the basis of associated artifacts and the re- examination of their findspots.13 Consequently, the type is generally seen as the ox-hide ingot's fully de-

Fig. 3. Ox-hide ingot of Type lb. 1:4

12 For the poor state of preservation of the Ulu Burun ingots, see Bass (supra n. 1) 276.

13 For full-sized intact ingots of Type 3 and their proposed dates, see Buchholz (supra n. 9) 28-29 nos. 1-2, pl. 3.1-2; Bass (supra n. 11) 57, 61 nos. 3, 5; and H.W. Catling, Cypriot Bronzework in the Mycenaean World (Oxford 1964) 267-68 nos. 1, 3 and pl. 49.a-b, c, with two miniature ingots of Type 3 from Enkomi on p. 269. Another miniature ingot of Type 3, alleged to have been found at Makarska in Dalmatia (Buchholz [supra n. 9] 37 no. 57, 35 pl. 5.5; Bass [supra n. 11] 61 no. 75) is probably also from Cyprus (L. Vagnetti, "Osservazioni sul cosiddetto ripostiglio di Makar-

veloped form before its disappearance at the end of the Late Bronze Age.l4 Bass demonstrated, however, cit-

ing among others the Type 3 ingots depicted in two

Egyptian tomb paintings, that ingots cannot always be dated as precisely as Buchholz would indicate.15

Three stacked copper ingots of Buchholz's Type lb

(fig. 3), KW 390 the uppermost, were also recovered from the newly excavated upper part (N-10) of the wreck; a tin ingot fragment was found trapped between the lowest two. These Type lb ingots are similar in shape and size to the example found nearby in 1984.16 They are asymmetrical in that no two of their sides are equal (KW 389: 1. 0.32-0.33 m.; w. 0.24- 0.25 m.; max. th. 0.05 m. KW 390: 1. 0.30 m.; w. 0.24-0.25 m.; max. th. 0.05 m.).

Ingots of Type lb appear to be more common dur-

ing the 16th and 15th centuries B.C. and continue into the 14th century, but the problems of dating ingots by shape have already been mentioned. The Ulu Burun

examples provide the first direct evidence for the syn- chronous use of ingot Types 1, 2, and 3. It is possible, however, that we have at Ulu Burun a separate cate-

gory of smaller, perhaps fractional ingots which did not require handles for carrying, for the Type lb in-

gots on the wreck appear to be smaller than Type lb

ingots found elsewhere and depicted in Egyptian tomb

paintings.7 A trapeziform (i.e., no two sides are parallel) cop-

per slab (KW 388: 1. 0.29-0.32; w. 0.20; th. 0.04), with distinctly beveled edges, surfaced in area M-11

only about a meter from the Type lb ingots. Together with the 19 bun ingots excavated in 1984,

and the single example raised during the 1983 survey, 39 bun ingots have been recovered. As with the ox- hide type, bun ingots excavated in 1985 exhibit much

diversity in size; the largest (KW 514) measures 30 cm. in diameter and 3.2 cm. in thickness, and the smallest (KW 106) is 22 x 3.9 cm. If bun ingots should be divided into two basic categories by size, a reasonable dividing point would be about 27 cm. in diameter. In accordance with such a division, 70% of

ska," Studi ciprioti e rapporti di scavo 1 [Rome 1971] 210-11, 213-14, and 203 fig. 1, 209 fig. 7 with English summary on 216). Thus, all ingots of Type 3 known prior to the discovery of the Ulu Burun shipwreck appear to have been found on Cyprus.

14 Buchholz (supra n. 9) 6, 7 fig. 2; Catling (supra n. 13) 271-72.

15 Bass (supra n. 11) 69. 16 Bass (supra n. 1) pl. 17.1 lower right. 17 For representations of Type lb ingots on Egyptian tomb

paintings, see Bass (supra n. 11) 62-66 figs. 62-69, 71, 74-76, 78 and 80.

7 1988]

CEMAL PULAK

these ingots would be classified as "small," with an average diameter of about 23 cm.

One well-preserved bun ingot (KW 397) is inscribed with a mark resembling an open-sided rec- tangle on its relatively smooth convex bottom. Most of the bun ingots have yet to be fully cleaned and conserved, but conservator Robert Payton reports from Bodrum that of the 19 freed of encrustation, four are of the "large" size and are unmarked, whereas 10 of the 14 "small" ingots are incised on their convex surfaces with the mark T (mark 37B or 39B of Gelidonya ingots).18

Although metallurgical studies of the Ulu Burun ingots have not been completed, analyses have shown that the ox-hide ingots and at least some of the bun ingots on the Cape Gelidonya ship were cast from the same, nearly pure, copper. The two major types of ingots appear to differ only in size and, therefore, in the amount of copper in each ingot; the different shapes have been attributed to the possibility of different production techniques or areas, or to the need for smaller units of copper.19 Whether the ox-hide ingots them- selves resulted directly from primary smelting of copper-bearing ores or from remelting of raw copper metal is uncertain. The discovery on the Ulu Burun wreck of smaller Type lb ox-hide ingots makes it less likely that bun ingots were intentionally cast as fractional units of the ox-hide form. It has also been suggested that primary bun ingots, that is those which were not formed from remelted scrap metal, may in fact be the copper left at the bottoms of smelting fur- naces after slag tapping, while ox-hide ingots were cast outside the furnace, possibly even with remelted bun ingots.20 The Cape Gelidonya and Ulu Burun

18 Bass (supra n. 11) 72 fig. 90. 19 J.D. Muhly, T.S. Wheeler and R. Maddin, "The Cape

Gelidonya Shipwreck and the Bronze Age Metals Trade in the Eastern Mediterranean," JFA 4 (1977) 357, 358 table 1.

20 R.F. Tylecote, "The Late Bronze Age: Copper and Bronze Metallurgy at Enkomi and Kition," in J.D. Muhly, R. Maddin and V. Karageorghis eds., Early Metallurgy in Cyprus, 4000-500 B.C. (Nicosia 1982) 94; see also Muhly et al. (supra n. 19) 354.

21 Buchholz (supra n. 9) 28-39; Bass (supra n. 11) 52-69; G.F. Bass, "Cape Gelidonya and Bronze Age Maritime Trade," in H.A. Hoffner, Jr., ed., Orient and Occident, Festschrift Cyrus H. Gordon (Alter Orient und Altes Testa- ment 22, Neukirchen 1973) 29-31; and Catling (supra n. 13) 267-69. Also E. Galili, N. Shmueli and M. Artzy, "Bronze Age Ship's Cargo of Copper and Tin," IJNA 15 (1986) 32-34 with figs. 7-8; B. Dimitrov, "Underwater Re- search Along the South Bulgarian Black Sea Coast in 1976 and 1977," IJNA 8 (1979) 70, 73 with fig. 3; and F. Lo Schiavo, E. Macnamara, and L. Vagnetti, "Late Cypriot Imports to Italy and Their Influence on Local Bronze- work," BSR 53 (1985) 10-13; F. Lo Schiavo, Nuragic Sar-

wrecks, however, provide ample evidence of the con- current shipment of bun and ox-hide ingots during the Late Bronze Age, indicating that the bun shape did not serve as an intermediary product before being re- cast in the ox-hide shape; the shape of one was not preferred over the other. The widespread occurrence of the basic ox-hide ingot shape, from Sardinia to Mesopotamia and from Egypt (as representations) to the Black Sea,21 has suggested to some a central authority exercising control over the production of and the trade in this important commodity.22 The recur- rence of differing weights for the ox-hide ingots, even after taking into account the change in weight due to corrosion, probably indicates that these ingots were not intended for use as currency but rather represented a quantity of blister copper subject to weighing and evaluation during each commercial transaction. Their broadly common weight standards would sim- plify accounting procedures by allowing for a rough but quick reckoning of a given quantity of raw metal prior to weighing.23 The shape itself probably evolved merely for its ease of transportation over long dis- tances on pack animals. That the shape also developed because it was more practical for shipboard stowage than that of discoid ingots,24 however, has been ne- gated by the discovery of the latter on the Cape Geli- donya and Ulu Burun ships.25 Once all the Ulu Bu- run ingots are excavated, cleaned and studied, they will provide an important source of information for our understanding of the mechanism behind the ox- hide ingot trade.

More than 40 tin ingots and many ingot fragments reduced to unrecognizable lumps have been found on the wreck. Of these, 17 appear to be fragments of ox-

dinia in Its Mediterranean Setting: Some Recent Advances (University of Edinburgh Department of Archaeology Oc- casional Paper 12, 1985) 3-4, and figs. 2, 3 on 26-27; additional references are in J.D. Muhly, "The Role of Cyprus in the Economy of the Eastern Mediterranean during the Sec- ond Millennium B.C.," in V. Karageorghis ed., Acts of the International Archaeological Symposium "Cyprus between the Orient and the Occident" (Nicosia 1986) 55-56.

22 Views differ as to the identity of this centralized authority. Some favor a base in the Aegean (Catling [supra n. 13] 271), others in Syria (Bass [supra n. 1] 294-95) or Cyprus (R. Maddin, T.S. Wheeler, and J.D. Muhly, "Tin in the Ancient Near East: Old Questions and New Finds," Expe- dition 19:2 [1977] 46).

23 N.F. Parise, "I pani di rame del II millennio a.C. Consi- derazioni preliminari," in Atti e memorie del 1? Congresso Internazionale di Micenologia (Roma 1968) 128; C. Zac- cagnini, "Aspects of Copper Trade in the Eastern Mediter- ranean During the Late Bronze Age," in Traffici micenei nel Mediterraneo (Taranto 1986) 414-15.

24 Buchholz (supra n. 9) 2. 25 Bass (supra n. 11) 69 n. 78.

8 [AJA 92


hide ingots, four of bun or plano-convex ingots, and the remainder mostly too corroded for attribution to any specific type. The reversion of tin, especially in ingots trapped beneath copper ox-hide ingots, to a grayish powdery pulp that dissipates easily when dis- turbed under water may partly be due to the tin's galvanic reaction with copper.26 In other cases, however, the tin has fared better and it is possible to discern an ingot's general shape in spite of its rough and blistered surfaces. When these bubble-like blisters are broken off, a gray-white, hard crystalline structure can usually be observed beneath. In these cases, the transformation of tin from its white state to a gray, granular substance with few metallic properties is due to a totally different mechanism. This allotropic modification of tin into gray tin ("tin pest") is observed when tin of high purity is subjected to low temperatures (13.2? C) for substantial periods.27 That this reaction may be responsible for the condition of the Ulu Burun material is substantiated by the purity of the tin; analysis of a small fragment recovered during the 1983 survey revealed 99.5% tin.28 Robert Maddin, who sampled most of the copper and several of the tin ingots raised in 1984, reports that he has observed gray tin. This transformation occurs with a volume change of about 20 to nearly 25%, resulting in crumbling gray granules. Given favorable conditions and time, this transformation continues until all the white tin is con- sumed. Apparently, this is the first time gray tin has been uncovered in an archaeological context.29

Of the 17 tin ingots in ox-hide form, 16 represent quarter ingots, each preserving one handle (fig. 4). The seemingly intact quarter-ingots range in size from 28 cm. long, 15 cm. wide, and ca. 3.5 cm. thick (KW 641) to 25.8 cm. long, 11.7 cm. wide, and 3.6 cm. thick (KW 399). Whether the partial ingots represent exact quadrants of the ingots from which they were cut cannot be ascertained, but it seems reasonable to assume that at least some attempt was made to maintain the uniformity of the pieces. Payton reports that of the few ingots of tin which have been cleaned, one (KW 200: fig. 4), although poorly preserved, appears to be marked with the sign T, the same sign observed on most of the copper bun ingots. A second tin ingot (KW 201) is pierced and is marked on one of its surfaces

26 For behavior of galvanic couples in sea water, see N. North, "The Role of Galvanic Couples in the Corrosion of Shipwreck Metals," IJNA 13 (1984) 133-36, with bibliography.

27 J.A. Charles, "The Development of the Usage of Tin and Tin-Bronze: Some Problems," in A.D. Franklin et al. eds., The Search for Ancient Tin (Washington, D.C. 1977) 26; Maddin et al. (supra n. 22) 41.

Fig. 4. Quarter ox-hide ingot of tin (KW 200). 1:3

with a straight line; this may represent a third tin-ingot type, but its full shape has not been determined.

Ingot KW 644 (fig. 5; 1. 0.36; w. 0.30; max. th. 0.055), raised from area M-11 as part of an encrusted mass of tin ingots, is the only half ox-hide ingot of tin recovered to date. It is not known if it was cut deliberately or broken during the wrecking of the ship.

Several tin plano-convex, discoid or bun ingots were also recovered. Of these, KW 401 has a maximum diameter of 23.2 cm. and a thickness of 4 cm. Three incomplete ingots appear to have the same general shape; two can be reconstructed to approximately the same dimensions, while the third, if indeed it is a bun ingot, has a preserved diameter of 30.5 cm. and a thickness of 7.3 cm.30 Whether the manufacture of tin

28 Bass et al. (supra n. 1) 277. 29 According to James D. Muhly, quoted by Robert Mad-

din in a letter of 9 November 1984 to George F. Bass. 30 Five tin ingots, three marked apparently with syllabic

signs, have been reported from an undated context, possibly from a LBA shipwreck, off the coast of Israel (Galili et al. [supra n. 21] 25-32). They were probably cast in shallow pits, some still retaining the debris of the pits. The ingots are

1988] 9

CEMAL PULAK

Fig. 5. Half ox-hide ingot of tin (KW 644). 1:4

ingots in ox-hide and bun shapes-the same forms used in copper ingots-suggests a similar production technique rather than the control of metals production by a central authority remains to be investigated.

Pottery The 67 Canaanite amphoras raised so far represent

only about half of the total number estimated on the wreck. The Ulu Burun amphoras are still being cleaned and conserved and have not yet been studied in detail. Only one has been drawn, and comparisons with other amphoras are tentative. Similarities between the Ulu Burun amphoras and Canaanite amphoras from Mycenae, Menidi, Tell Abu Hawam, and the tombs near the Persian Garden at Akko have been noted,31 but mention may also be made of somewhat similar amphoras from Megiddo32 and Byblos,33

flat on top and convex on the bottom, are extremely irregular, and remotely resemble the Ulu Burun bun ingots in section; parts of them had been cut off in antiquity. Two rectangular tin ingots, found offshore near Haifa, were dated to the LBA on the basis of engraved signs resembling those of the Cypro-Minoan syllabary (Maddin et al. [supra n. 22] 45-47), but two 5th-century B.C. ingots, probably from the same offshore source, suggest a 5th-century date for all (M. Artzy, "Arethusa of the Tin Ingot," BASOR 250 [1983] 51-55).

31 Bass (supra n. 1) 277. 32G. Loud, Megiddo II (Chicago 1948) pl. 59.11, from

stratum VIII. 33 J.-F. Salles, La Necropole "K" de Byblos (Boulogne

and two examples from the sea off the coast of Israel;34 Raban, on the basis of chemical analyses of the amphoras from the sea, proposes that amphoras of this type originated along the Syro-Palestinian coast from Akko to Ugarit.35 The undated amphoras excavated by Tsountas in Tombs 58 and 59 at Mycenae remain the best parallels.36

A unique amphora of large proportions (KW 588), but incomplete from the neck up, resembles a large amphora from the Akko tombs.37 No dimensions are given for the Akko jar, but from the scale provided the Akko jar seems somewhat taller than the Ulu Bu- run example, although with about the same shoulder diameter.

It is now possible to divide the Ulu Burun amphoras tentatively into three basic sizes. Amphora KW 588 has a maximum preserved height of 49.5 cm. (ca. 4 cm. missing from the neck and rim), and maximum diameter of 39.5 cm. Only a few leaf fragments and several fig seeds were recovered by sieving the contents of this capacious jar. The second size comprises amphoras averaging 58-59 cm. in height and ca. 29 cm. in diameter, while yet a third size appears to average 50 cm. in height and 24 cm. in diameter; these last two sizes are not clearly distinct, and intermediate forms do exist. A seemingly unique amphora, mistakenly described in 1984 preliminary reports as

being 50 cm. tall and 33 cm. in diameter, due to a cata-

loging error, was in fact 24 cm. in diameter. Sieving of contents yielded a yellow resin from at

least 45 amphoras, or 67% of the total number excavated. Several of the amphoras produced a few fig seeds, but it is possible that these are intrusive since fig seeds were previously found in amphoras containing large amounts of resin.

Eight amphoras contained one potsherd each, while two others had several small, apparently joining sherds. Bass suggested, as one possibility, that these sherds were intentionally placed in amphoras, because

they appeared too large to have fit through the jars'

1980) 119 pl. XVII.4. 34 A. Raban, The Commercial Jar in the Ancient Near

East: Its Evidence for Interconnections amongst the Biblical Lands (Diss. Hebrew University 1980, in Hebrew) pl. 15- G.3 and 7.

35 Raban (supra n. 34) 6; see also R.E. Jones, Greek and Cypriot Pottery, A Review of Scientific Studies (British School at Athens Occasional Paper 1, 1986) 572-73.

36 For good illustrations of jars no. 2924 and 4569, see A. Xkerstrom, "More Canaanite Jars from Greece," OpAth 11 (1975) 190 figs. 11 and 13.

37 S. Ben-Arieh and G. Edelstein, Akko. Tombs Near the Persian Garden ('Atiqot [English Series] 12, Jerusalem 1977) pl. XII.2.

[AJA 92 10


necks without guidance.38 It is more probable, however, that these sherds supported wet sealing clay while it was shaped over the mouths of the jars; sherds placed deep inside resin-filled amphoras would more likely have survived intact. Jar and amphora sealings recovered from 18th-Dynasty Malkata clearly indicate that before sealing, a stopper of reed, mud, or pottery was placed over the container's mouth;39 reed or grass stoppers continued in use nearly 2,000 years later.40 Several stoppers, a potsherd over a reed plug for example, could be employed in combination.41 In rare instances leaves and bungs of chopped chaff mixed with an adhesive or mud were substituted. The stopper's main purpose apparently was to prevent the wet mud of the sealing from contaminating the amphora's contents. That a similar sealing practice may have been employed on the Ulu Burun amphoras is strengthened by the presence of copious organic detritus in the bottoms of some jars. This organic layer, sometimes found directly under a potsherd and comprising primarily grass blades, leaves, seeds, and resin chips, may represent the remains of a stopper or bung formed from plant material. When the ship sank, the unbaked clay sealing dissolved and water pressure may have forced the stopper, and in some instances the potsherd, into the jar. That an implosive force was felt by some jars, almost certainly after being structurally weakened by impact, is demonstrated by the rim and neck fragments found inside them. It must be remem- bered, however, that not a single amphora raised from the wreck gives any positive indication of the technique by which it was sealed, and that possible recon- structions remain speculative.

Cheryl Haldane, who is studying the contents of these amphoras, reports that nearly all the organic deposits recovered in substantial amounts contained many leaves and, especially, the fruits of a Pistacia tree (Pistacia cf. terebinthus).42 A few amphoras also contained ovine or caprine digital bones (phalanges),

38 Bass (supra n. 1) 278. 39 C. Hope, Excavations at Malkata and the Birket Habu

1971-1974, V. Jar Sealings and Amphorae of the 18th Dyn- asty: A Technological Study (Warminster 1978) 6, 14, 31 fig. 8, pls. I-II. A similar method for sealing jars appears to have been used in Mycenaean Greece (C.G. Koehler, "Han- dling of Greek Transport Amphoras," in J.-Y. Empereur and Y. Garlan eds., Recherches sur les amphores grecques [BCH Suppl. 13, 1986] 53),- although publication of these sealings does not specify the types of jars sealed (R.M. Dawkins, "Laconia. 1-Excavations at Sparta, 1910. 2-The Mycenaean City near the Menelaion," BSA 16 [1909-1910] 9-10, pl. III).

40 D. Colt ed., Excavations at Nessana I (London 1962) 309, pl. XXV.18-19.

41 Hope (supra n. 39) 43 table 2 lists stopper combina-

as well as shells of land snails. It is possible, therefore, that some of the organic detritus may be remnants of previous contents, in which case these amphoras were reused when filled with their final cargo of resin.

Preliminary research by Curt Beck, of Vassar Col- lege's Amber Research Laboratory, had indicated that samples of resin taken from the amphoras in 1984 probably represented the family Burseracaea, which includes frankincense (Boswellia spp.) and myrrh (Commiphora spp.).43 A few samples recently submit- ted to John S. Mills of the National Gallery of Lon- don, however, have been reported to be a Pistacia resin, most likely that of P. atlantica (also known as P. terebinthus var. atlantica),44 which yields the so- called Chian turpentine or terebinth resin. This identification certainly explains, in part, the presence in amphoras of the Pistacia remains mentioned above.

The sticky, semifluid nature of terebinth resin would necessitate its storage in containers imperme- able to the resin such as amphoras. On the other hand, mastic (obtained from P. lentiscus) and frankincense, both harvested in the form of dry tears,45 would be more efficiently transported in lighter containers such as baskets, as they are carried today. The possible presence of terebinth resin in Egyptian tombs may indicate the resin's ritual significance in antiquity,46 but in modern times it is substituted for chewing gum, and in the eastern Egyptian desert, it is used in the prepara- tion of perfumes.47

A few flecks of orpiment or yellow arsenic recovered from two amphoras in 1985 are probably intrusive. The quantity of loose and scattered orpiment in the area, as well as these traces, must have spilled from the amphora filled with the substance in the same location (M-12).48

Pithos KW 251, the largest of six storage jars visible on the wreck, was raised in 1984. As previously reported, it contained 18 pieces of seemingly Cypriot pottery, some still stacked inside one another, and thus

tions, and p. 41 pl. IV illustrates sealings with potsherds and grass stoppers.

42 C. Haldane, "Archaeobotanical Remains from Four Shipwrecks off Turkey's Southern Shore," presented in Istanbul on 12 September 1986 at the Fifth OPTIMA Conference.

43 Bass (supra n. 1) 277-78. 44 John S. Mills, letter of 16 December 1986 to E.T. Hall. 45 A. Lucas, Ancient Egyptian Materials and Industries4

(London 1962) 91, 321; Mills (supra n. 44). 46 Lucas (supra n. 45) 324. 47 M.E. Kislev, "Reference to the Pistachio Tree in Near

East Geographical Names," PEQ 117 (1985) 133. I thank John S. Mills for bringing this article to my attention.

48 Bass (supra n. 1) 278.

1988] 11

CEMAL PULAK

dark-brown fabric and grayish-brown slip they also contrast sharply with the typical Syro-Palestinian lamps on the wreck that have coarse red-orange fabric with reddish-brown surfaces. The latter differ from the lamps recovered from the pithoi not only in color, but also in shape and fabric; their nozzles are straight- sided and deeply pinched, while their fabric is more crumbly. One of the lamps belonging to the latter category has a fire-blackened nozzle, perhaps indicating its use aboard the ship, while those from the pithoi are pristine.

Before the pithoi are cleaned of encrustation and drawn, their profiles cannot be studied. One of the larger jars (KW 252) is decorated with a fine, raised ridge at the juncture of its neck and body, and three wider horizontal parallel ridges circumscribe the body around the shoulder. The smaller pithos (KW 250: h. ca. 0.80; diam. ca. 0.62) appears to be plain. Except for its shorter neck and squatter profile, it has the same general form as its larger counterpart. These large storage jars are similar to the ovoid pithoi with widened shoulders from Pyla-Kokkinokremos,51 Athi- enou52 and Minet-el-Beida.53 The general type is

Fig. 6. Lamp KW 485. 1:2

it served as a protective container for the material inside.49 Whether the pithoi on the ship constituted cargo as well as containers for cargo, however, remains unknown.

Concreted to the sides of the upper end of the southern gully, smaller pithos KW 250 contained little sediment and no visible artifacts. While transferring the sediment into plastic bags for later examination, however, we found 21 lead fishing-net weights, a perfectly preserved ceramic lamp (KW 437) and pilgrim flask KW 438 (fig. 7). The lamp resembles those found in 1984 in the pithos KW 251, but the pilgrim flask is a new addition to the ceramic repertory from these containers. The pithos was chiseled free from the bedrock and raised to the surface.

The saucer-shaped lamps from the pithoi do not appear to be common on Cyprus,50 but with their

49 Bass (supra n. 1) 279. 50 These lamps are treated in greater detail by Bass (supra

n. 1) 279, 281, 282 ill. 14. 51 V. Karageorghis and M. Demas, Pyla-Kokkinokremos.

A Late 13th-Century B.C. Fortified Settlement in Cyprus

/a: : : :

Fig. 7. Pilgrim flask KW 438. 1:2

(Nicosia 1984) 34 no. 18A, pls. XXI.18A, XLI.18A. 52 T. Dothan and A. Ben-Tor, Excavations at Athienou,

Cyprus, 1971-1972 (Qedem 16, Jerusalem 1983) 113, 114 fig. 52.2.

53 C.F.A. Schaeffer, Ugaritica II (Paris 1949) 208 no. 27,

12 [AJA 92


found in Late Cypriot II-III contexts,54 but compar- able material appears to date to the 13th-12th centuries B.C.55

The only fairly certain Cypriot pottery discovered in 1985 is an incomplete Base-ring II bowl (KW 277) excavated in grid square L-12. It is very similar to the examples found in pithos KW 251 the previous summer. A trail of sherds from other Cypriot vessels between this Base-ring II bowl and pithos KW 251 may point to the storage jar's original upslope position before it rolled downslope.

An intact wall-bracket (KW 304: 1. ca. 0.40; h. of bowl 0.17), of the plain type found during the 1983 survey and the 1984 excavation, was uncovered next to a stone anchor in area M-14.56 Wall-brackets of the same basic design are well known from both Cyprus and the Syro-Palestinian coast.57

Two Syrian lamps, KW 485 (fig. 6) and KW 262, found side by side, are of red-orange fabric with reddish-brown surface. As mentioned above, they differ both in shape and color from the examples found in the pithoi; both lamps appear to be of the same type as lamp KW 59 found in 1984.58 A unique concave bottom on lamp KW 485 (fig. 6), enabling it to sit flat on its base, is unlike the usual thickened, flat bottoms of

Fig. 8. Stirrup jar KW 305. 1:2

209 fig. 86.27. 54 Dothan and Ben-Tor (supra n. 52) 113; P. Astr6m,

SwCyprusExp IV. Pt.1C, 261 Type 1Ea. 55 For a brief discussion of similar jars and their dates, see

Dothan and Ben-Tor (supra n. 52) 113. 56 Bass (supra n. 1) 275, 292; illustrated example in Bass

et al. (supra n. 1) 273, 276 fig. 6. 57 Bass et al. (supra n. 1) 273, 276; Bass (supra n. 1) 292

n. 139. 58 Bass (supra n. 1) 285, 287 ill. 22.

Fig. 9. One-handled cup KW 334. 1:2

later types which, according to Amiran, made their first appearance in the Near East in LB IIB.59

Four Syrian pilgrim flasks in two distinct sizes were raised in 1985. A small flask (KW 438: fig. 7; h. 0.175; max. diam. 0.133) was in pithos KW 250, while an even smaller example (KW 604: h. 0.153; max. diam. 0.123) lay under a stone anchor in grid square L-14.

Lying partly beneath and damaged by the same stone anchor was one of the larger flasks (KW 600: h. ca. 0.30; max. diam. ca. 0.24); the southern gully yielded the fourth flask (KW 434: h. 0.312; max. diam. 0.254), also damaged and partially encrusted. All four pilgrim flasks are of the same general type as those recovered previously from the wreck.60 One of each size (KW 438: fig. 7, and KW 604), however, displays a slightly different rim. Unlike those with a triangular-sectioned, heavy lip, these flasks have a funnel-shaped mouth with a much smaller rim, appearing to sprout from the plastic V-like decoration formed by the join of the circular handles to the short neck.

Two of the pilgrim flasks contained fig seeds. Bass has cautioned, however, that rather than representing contents, the seeds may be the remains of dried-fig stoppers.61 Good parallels for the flasks have already been noted from Tell Abu Hawam and Hazor.62 Others are from the Akko tombs,63 although some of the smaller Akko vessels appear more rounded at the juncture of the separately formed halves of the body, and Lachish, where the funnel-shaped mouth is represented fairly closely by a flask from Level VI.64 The Ulu Burun pilgrim flasks appear to be undecorated, but further cleaning may reveal painted decoration,

59 R. Amiran, Ancient Pottery of the Holy Land (New Brunswick, N.J. 1970) 190.

60 Bass et al. (supra n. 1) 276, 277 fig. 7; Bass (supra n. 1) 284-85, 286 ill. 21.

61 Bass (supra n. 1) 285. 62 Bass (supra n. 1) 285. 63 Ben-Arieh and Edelstein (supra n. 37) 23 fig. 11.1-4,

pls. III.4, IX.3-4, XII.3. 64 0. Tufnell et al., Lachish IV. The Bronze Age (London

1958) 217, pl. 84.954.

m

1988] 13

. . w - -

.. .. .. . . . . Ii

CEMAL PULAK

as on other seemingly undecorated vessels from the wreck.

Mycenaean finds in 1985 include the two fragmentary stirrup-jars and the one-handled cup mentioned above. Globular stirrup-jar KW 305 (fig. 8; h. ca. 0.12; max. diam. ca. 0.12), recently restored, appears to be closest in profile to Furumark's FS 171, and is decorated on its body with groups of fine lines flanked by single wide bands, and on its shoulder zone with the multiple stem motif (FM 19:25).65 The larger stirrup-jar (KW 308) is also decorated with bands. Designs on its upper body will become clear after cleaning, but the jar appears to be globular and has a false neck with concave profile similar to that of KW 305. The disc of the false neck is flat on KW 308 and slightly convex on KW 305. The fabrics of both are cream buff and well refined. Their surfaces, covered with a slip of the same color, are decorated with red- brown paint.

The deep semi-globular cup (KW 334: fig. 9; h. 0.055; max. diam. 0.085) is well rounded below and the fusing of its slightly closing sides with its lip in a continuous curve forms a bell-shaped profile; the base is hollowed out, and the single handle rises a little above the rim. The linear decoration of the usual heavy red-brown bands on the cup's outer surface occurs around the rim, the lower body at the handle at- tachment, and the base. The intermediate zone between the two upper broad bands is undecorated while the region between the middle and base is filled with seven more or less equally spaced thin lines. The shape of the cup conforms to FS 214.66

Glass Six more discoid ("truncated cone") glass ingots

were recovered from M-12, the area which yielded all but two of the 15 intact ingots and other fragmentary pieces found in 1984.67 The ingots' lower (smaller) faces are rounded and smooth while their upper (larger) faces, with sharp and irregular edges, are rougher. In some examples the two faces are not parallel. The largest ingot of 1985 (KW 333) has a maximum diameter of 0.155 m., a minimum diameter of 0.125 m., and a thickness of 0.055 m.; and the smallest (KW 383) a maximum diameter of 0.138 m., minimum diameter of 0.122 m., and a thickness of 0.06 m. An unusually irregular ingot (KW 385) has a

65 A. Furumark, The Mycenaean Pottery I. Analysis and Classification (Stockholm 1972, repr. of 1941 ed.) 31 fig. 6.171, 299 fig. 47.25. See also P.A. Mountjoy, Mycenaean Decorated Pottery: A Guide to Identification (SIMA 73, Goteborg 1986) 77, 79.

66 Furumark (supra n. 65) 49-50, 48 fig. 13.214.

maximum diameter of 0.149 m., and a thickness of 0.035 m.

The deep blue coloring of glass ingots raised in 1984 could be observed in almost every case in spite of the extensive encrustation and patination coating most examples. The situation differed greatly in 1985. Of the six intact ingots found, only three are definitely blue. The others are either green with black and white var- iegation, or, as in certain large ingot fragments, light brown or amber. It is almost impossible to ascertain the original color of these last pieces since they have been completely hydrolized and no true glass remains.

Analyses of the Ulu Burun glass by Robert H. Brill of the Corning Museum of Glass have identified cobalt as the coloring agent. The extensively "weath- ered" green, black, and amber glass, on the other hand, is more difficult to interpret. According to Brill, at least the green coloring appears to result mainly from copper, perhaps as part of the original colorant or perhaps as contamination from nearby copper ingots; the latter explanation is plausible since the green ingots were found directly below copper ox-hide ingots. The core of at least one green ingot, on more recent testing, proved to be blue. Because the amber glass contains neither cobalt nor other metals such as nickel, zinc, and manganese that accompany cobalt in the blue ingots, its original color was probably amber or another color.

Tools A large assemblage of tools, found mainly in areas

L-11, M-11, and K-14, includes sickles, axes, adzes, chisels, drill bits and/or awls, and a pair of tongs.

The sickles, varying in length from 11-18 cm., have yet to be cleaned of encrustation to reveal the exact shapes of their blades and the manner in which the handles are attached. Of four blades found, none have long tangs or sockets for hafting, but it is possible that sockets have broken off; the apparently square-butt ends on two of the blades, when cleaned, may prove to be short tangs with rivet holes. A fifth blade (KW 344: fig. 10), curved like a sickle but with its cutting edge on the outer edge, is probably a razor.

The wreck has yielded so far four flat ax and adze blades of two distinct types, all now cleaned and conserved.68 KW 587 (fig. 11; 1. 0.195; w. 0.062; max. th. 0.009) from area L-15 and KW 218 (1. 0.195; w.

67 Bass (supra n. 1) 274, 281-82 with ills. 15-16. Photo- graphs of two glass ingots have been published in JGS 28 (1986) 118 fig. 2 (the glass beads with the ingots are from a Mycenaean tomb at Miisgebi, near Bodrum, Turkey).

68 The distinction between an ax and an adze blade, especially in earlier periods, is rather vague. W.M.F. Petrie,

14 [AJA 92


I.

I \

Fig. 10. Bronze razor KW 344. 1:2

0.052; max. th. 0.010) from the previous campaign are lugged or trunnion axes of the same approximate dimensions and shape,69 save for the slightly more slender proportions and somewhat rounded butt of the former. The latter blade, although chamfered on both faces, appears to have an asymmetrical cutting edge and may, in fact, be an adze blade. The face with the least chamfer is incised with what is probably a stylized fish.70

Lugged axes were widespread during the Bronze Age, but axes from the Ulu Burun site are of Max-

Tools and Weapons (Warminster 1974, repr. of 1917 ed.) 5, lists six distinguishing features for the separation of the two tools. The situation is complicated, however, by blades that are identical in form, differing only by the manner in which their cutting edges are sharpened. Here we adopt the commonly accepted view, without passing judgement on its validity, that if a non-socketed blade, regardless of its thickness and length, has beveling or chamfering on both of its faces, it is called an ax; and, if this beveling is confined to only one face, it is identified as an adze.

69 Bass (supra n. 1) 292 ill. 32. 70 Inscribed decoration on tools, especially on ax and adze

blades, is not uncommon in the LBA. For a fish on a different ax-blade type, see Petrie (supra n. 68) pl. IV.127.

71 R. Maxwell-Hyslop, "Bronze Lugged Axe- or Adze

Fig. 11. Bronze lugged ax KW 587. 1:2

well-Hyslop's Type II with flat blade and small shoulder protrusions for lugs.71 This is the most common type in Palestine, although it is also found in Syria, Egypt, Anatolia and Cyprus.72 Sites in Pales- tine and Cyprus have produced blades somewhat resembling ours,73 but the closest parallels for KW 218 are from Hazor, dated by its excavator to the 14th century B.C.;74 and for KW 587, from the treasury of the Great Priest at Ras Shamra-Ugarit, dated to the same century by its excavator,75 but lowered to the second half of the 12th century by Catling.76 The length of

Blades from Asia," Iraq 15 (1953) 72. 72 Maxwell-Hyslop (supra n. 71) 81-82; for additional

references, see J. Deshayes, Les outils de bronze, de l'Indus au Danube (Paris 1960) Vol. 1, 116-17; Vol. 2, 55-56 his Type B; Catling (supra n. 13) 87.

73 For a similar blade from Palestine, see for example, Maxwell-Hyslop (supra n. 71) 81 and 85 fig. 5.5, and for Cyprus, see Catling (supra n. 13) 87 no. 1, fig. 8.11, pl. 6g.

74 Y. Yadin et al., Hazor II (Jerusalem 1960) 159, pl. CXCVI.8-9.

75 C.F.A. Schaeffer, Ugaritica III (Paris 1956) 270 fig. 234.8, 272 fig. 236, lower right.

76 Catling (supra n. 13) 286, but for a discussion of the dating to the 13th century of these bronze hoards, including that of the Great Priest at Ugarit, see J.D. Muhly, Copper

Pi

??;?:

??.7; i: ?-I Lt-

:?II;? ?:

''' ':; :i

??

)?-? -?:

''' ..??:? .:???j

I

., .. .

r? -i: ?'?

:?i ??i-: i?

1988] 15

CEMAL PULAK

which may have been used to lash the blade more se- curely to its handle. The type appears to be confined mainly to Egypt, its evolution traced by Petrie from the Third Dynasty to the 19th Dynasty.78 The blade, distinctly splayed and widest at the cutting edge, becomes narrower toward the butt and forms angular shoulders just below the rounded, semicircular head. The head is separated from the rest of the blade by a pronounced undercutting (neck). No good parallels for this adze have been found, but there is some similarity between our blade and one from Gurob in Egypt, dated by Petrie to the 18th or 19th Dynasty.79 Although both blades splay out toward their cutting edges, the Ulu Burun piece is more slender and its head more rounded. A blade from the South Rooms at Amarna is of similar shape and proportions, with the exception of its flattened butt.80

Fig. 12. Bronze necked adze KW 576. 1:2

the Ras Shamra blade, the shape and relative position of its lugs and the blade's semi-rounded butt-end are features which approximate the Ulu Burun material most closely. The Ras Shamra blade, however, is a typical adze by our definition. Among the bronzes found on the Cape Gelidonya wreck of approximately 1200 B.C. were at least four lugged blades.77 Al- though smaller in size, blades B107 and especially B108 from the Gelidonya wreck are good parallels for the Ulu Burun examples. It is too early to ascertain whether the Ulu Burun lugged axes constituted part of the ship's cargo, or were for shipboard use only. Together with the Gelidonya wreck, the Ulu Burun ship may provide clues to the westward spread of this predominantly Near Eastern blade type.

Blade KW 576 (fig. 12; 1. 0.205; w. 0.063; th. 0.009) is of the type designated by Petrie as a "necked adze." This tool has a unique, well-formed butt-end

and Tin (Transactions of the Connecticut Academy of Arts and Sciences 43, Hamden, Connecticut 1973) 375 n. 202.

77 Bass (supra n. 11) 95, 97 fig. 109.B107-B110, 98 fig. 110.B107-B110.

78 Petrie (supra n. 68) 16-17, pl. XVII.

91 ii:, .??-?, r::-: :r;i

:; ?!

1:?;?11

:?1 ?.ci

..?.r.l '"?,-: .? ?

.1 :I

'' :??:?

??' `:?'

:-.?? ?c-: -? ??? .?..? ?.?- -i ?' ? v. : '1 ?? .: ?-?-?

r :L:??::

?r?. .: '?'?

I?:

-?? ?I I I::

Fig. 13. Bronze necked adze KW 141. 1:2

79 Petrie (supra n. 68) 17, pl. XVII.91. 80 J.D.S. Pendlebury, The City of Akhenaten, Pt. 3: The

Central City and the Official Quarters (London 1951) Vol. 1, 75; Vol. 2, pl. LXXII.10:25.

L :?::I?:

,.. ?-1

F-:?:: ;?'?-'I ??

.-.-?

: ? ???? ? ': c.;. I?.? -, r??: ?? ..; ?.

:??-??.??, : ?;?. '? ?j

?.?? .? II. ?: .'!Z

T''3 ; 1? :??:??!

i`??? ', "? ''

??-; .? '

i?' ?,r ? ?,.r r, ; L?:- :I

?:?? r'. s- . '. ?: ??.: C ''?-?r+ 6??`::

r ; c?? : '?

..r

1

[AJA 92 16


Blade KW 141 (fig. 13; 1. 0.215; w. 0.052; th. 0.008), found in 1984, proved after conservation to be a necked blade. It differs from KW 576 by its straight sides and narrower blade. The cutting edge, sharpened on both faces, suggests that this is, according to our classification, an ax. The head, although somewhat rounded, is smaller and is not undercut distinctly as is that of KW 576. If the necks of these blades were intended to facilitate lashing to hafts, then the handle type best suited for the purpose would probably be that of an adze mount. Hence, the blade may be classified more appropriately as an adze despite the profile of its cutting edge.

Some affinity, without specific parallels, may be observed between KW 141 and copper models of necked-adze blades from the foundation deposits of Tuthmosis III.81 All four models have straight sides and shallow neck indentations, but their heads have flat butts, unlike the slightly rounded head of our blade. Perhaps the best parallel for its head is found in an incomplete stone mold from the turquoise mines of Serabit el-Khadim in the Sinai, where a number of stone molds for tools were recovered, including a fragment containing the butt for another necked-adze blade. The site has been dated by its excavator to the New Kingdom, probably the 18th Dynasty.82

The next group of tools comprises a series of six broad-chisel blades, five recovered in 1985. In their encrusted state these tools were assumed to be flat adze blades, but when cleaned, all proved to have relatively pointed butts. Some of the blades have minor splaying near their cutting edges while others are straight-sided. Two blades, fused together by encrustation, were uncovered in the southern gully. In the gully's lower reaches, a second pair was discovered under a Canaan- ite amphora and next to a Syrian pilgrim flask. The lengths of these tools vary, KW 376 being the longest (fig. 14; 1.0.201; w. 0.043; th. 0.0065), and KW 423 the shortest (1. 0.169; w. 0.036; th. 0.004), although most are closer to 20 cm. in length.

This blade type apparently is rare or nonexistent in the Near East, but there are many examples from the Aegean.83 The tool is treated by Petrie as an adze, whereas Catling and Deshayes classify it as a chisel.84

81 W.M.F. Petrie, Abydos I (London 1902) 30, pl. LXI.5. 82 1. Beit-Arieh, "Serabit el-Khadim: New Metallurgical and Chronological Aspects," Levant 17 (1985) 102 fig. 9.3, and 95 pl. VII.12; the details of the poorly preserved molds are difficult to discern from the small illustrations. For dating of site, see p. 115.

83 Deshayes (supra n. 72) Vol. 2, 38-39, pl. IX; Catling (supra n. 13) 106.

r:

; ?r?

????. :::i

'-?:? :??i

'r.

rC Eti:S

r.-??1

:?" ??., r;n:q

i?' ?:i ;r-.- ?' a ?' r .?e ?:! ::-.Z ?;;

[:. ?- "' -? ? ..?1 .?? :12;

'I t: ``:?e

i. 5 ':i 1:?? ''' ''' ??? ?; ? ?;.

Fig. 14. Bronze broad chisel KW 376. 1:2

Deshayes suggests that his chisel Subtype C3, which the Ulu Burun examples resemble most closely, ap- proaches a true adze blade because of its broad cutting edge. Based on the shape of the tapering tang that forms the chisel's narrow butt, this subtype is divided further into two groups. With the exception of chisel KW 264, which best fits Deshayes's category C3a, all of the Ulu Burun finds have shouldered tangs and belong to his Subtype C3b.

The few dated examples of Deshayes's Type C3 chisels are from the 14th to 12th centuries B.C.85 Al- though the number and close association of the chisels on the Ulu Burun wreck might suggest that they were cargo, the discovery of large tool kits on wrecks is not unknown.86

Three deep-bar chisels of varying lengths form the

84 Petrie (supra n. 68) 16; Catling (supra n. 13) 106 heavy chisels; and Deshayes (supra n. 72) Vol. 1, 88-89.

85 Deshayes (supra n. 72) Vol. 2, 38-39. 86 The seventh-century A.C. Byzantine wreck at Yassi

Ada, Turkey, contained among many other tools five nearly identical adzes: G.F. Bass and F.H. van Doorninck, Jr., Yassz Ada I (College Station, Texas 1982) 240-42 with ills.

1988] 17

CEMAL PULAK

According to Petrie, this lance-shaped and peculi- arly Egyptian tool was used from the First to the 19th Dynasty to cut and lever out wood chips from narrow mortises, a task for which its rectangular-sectioned shank was especially well suited.88 That deep-bar chisels were also used outside Egypt, albeit rarely, is illustrated by two 12th-Dynasty examples, closely resembling KW 276 and KW 307, found in Byblos.89 Two earlier chisels from the beginning of the Middle Kingdom, also from Byblos, are of the same general shape.90 A small deep-bar chisel with a partly preserved wooden handle still adhering to its butt was recovered from the Cape Gelidonya wreck.91 Another chisel, from the Tresor de Bronzes at Enkomi, has the same general form as the Ulu Burun pieces but differs from them by its angular shoulder taper;92 the hoard has been dated to the final stages of Cypriot Iron I (1200-1050 B.C.) by its excavator and to the second

I

I

Fig. 15. Bronze deep-bar chisel KW 276. 1:2

second chisel series. The largest is KW 276 (fig. 15; 1. 0.255; w. 0.027; th. 0.010), while KW 307 is slightly shorter but of heavier construction (1. 0.205; max. w. 0.023; th. 0.009); KW 536 is the smallest (1. 0.12; max. w. 0.015; th. 0.006). With their uniform thickness and wide blades that rapidly converge to form the tools' cutting edges, which are perpendicular to the plane of the wider faces of the blade, the Ulu Burun deep-bar chisels belong to Deshayes's category Alb, the Egyptian type. Chisel KW 536 with its tapering butt, on the other hand, may be of Deshayes's Sub- type B1.87

87 Deshayes (supra n. 72) Vol. 1, 102-103. 88 Petrie (supra n. 68) 20. 89 M. Dunand, Fouilles de Byblos I, 7926-1932 (Paris

1939) 151, pl. LXIX.2190-2191. 90 M. Dunand, Fouilles de Byblos II, 1933-1938 (Paris

1950, 1954) 22, pl. CLXXVIII.6900; P. Montet, Byblos et

Fig. 16. Bronze chisel KW 566. 1:2

l'Egypte (Paris 1929) 104, pl. LVIII.340. 91 Bass (supra n. 11) 100 no. B131, fig. 112.B131; 101 fig.

113.B131. 92 Invariably referred to as a cold-chisel (C.F.A. Schaeffer,

Enkomi-Alasia I [Paris 1952] 43-44, 42 fig. 3.25) or a tanged chisel (Catling [supra n. 13] 95).

[AJA 92 18

i? ?-? I?? ?.?: i?, ?, ? ??r

?' C?. . ?

??1, I' r? -s r. ??, :i -? ??

t? :??? ' TL

rC:-?

?' c-?_

' ;i? '? ?:?

i.-

r ;

? r,_R ??? 5??1'-?-

?n

?-I C:?"-:L ''0'

C.r? i? i?.-?? cr' .c- ?;.

???.-e -i ?, .?,

--? ?" '```` ''`' :??`??C

.r -?:? : ?r -r

I? -? r :??:lf .c??

???:-: 1\ it ?:?jc i? : .t .. ?,-,?;? ?

,.?I . ill_ii

,?+-- - cL'tlt

.?e ?? ':' ':_?;? v.:-? i ?" -??-??' r?.-?r

3 ?n,r


half of the 12th century by Catling.93 Unlike true deep-bar chisels, chisel KW 536 lacks their flaring cutting edges designed specifically to prevent the chisel's blade from being seized by the wood during the cutting of deep mortices. KW 536 appears to be somewhat similar in shape to chisels attributed to metalworking, but its lightweight construction reveals its probable function as a wood-cutting tool.94 Another type of bronze chisel, KW 566 (fig. 16; 1. 0.203), has a circular shaft and a tapering square butt for hafting.

No less than a dozen long and slender bronze tools of square and circular sections have been found, most still covered by encrustation. One cleaned piece, KW 579 (fig. 17; 1. 0.13; 0.01 x 0.01), is probably a large drill bit. Its tip has been forged into a lozenge-shaped point placed diagonally to the shank, and its butt end tapers slightly for the fitting of a socket. Catling suggests that similar pointed objects were bone or wooden hafted hand tools, and not machine tools.95

These tools are found in the Near East,96 the Ae- gean, and Cyprus.97 The best parallel for the Ulu Bu-

.,.

* - 1

- :, :.-

:.

i

Fig. 17. Bronze drill bit KW 579. 1:2

93 Schaeffer (supra n. 92) 38. Cypriot Iron I corresponds to LC III (C.F.A. Schaeffer, "Enkomi," AJA 52 [1948] 176-177). Catling (supra n. 13) 286, however, prefers to date this material to the second half of the 12th century. See, however, Muhly (supra n. 76) for a 13th-century dating of these hoards.

94 For a metalworking chisel from Ras Shamra, see Schaeffer (supra n. 75) 261-62, 268 fig, 213.22; and for the Aegean, see S. lakovidis, Excavations of the Necropolis at Perati (Institute of Archaeology Occasional Papers 8, Los Angeles 1980) 90 with fig. 106.M36.

95 Catling (supra n. 13) 97 n. 5. 96 Schaeffer (supra n. 75) 262, 268 fig. 233.11 and 13; 273

fig. 237.

::t ?. : - -? ?

.:?C:. fi!

ib:-?;r '`'?'? t ,;r iv? ?-?'?'. i.. Ir ?'- '.?` I` `-`

'' '" "' ..r.? ????? ? ?.

?.c??? : .?., ?.

.? ??e y?!,`I'l? I I: "' '' ' "'

" ?' ????.? :r? '-S

" ??'i\? `??` ' ?'??'

F :.?;?-.?1. ic ??? :?\1.?

fr ???:: r..? ? c, i .,??. ????

??:3

?t ??:\ ??? z.l :s

?I?,

\ :?

''?'

?-. ?-? ?\-?,?

ifr'z?

?i ... .......

tlj ;!

g.;. ? ?

?r:?r ?:s .C,:,?

i-i::nCI f?li ??? I

Fig. 18. Bronze drill bit KW 381. 1:2

run example is perhaps a drill bit from Enkomi having the same approximate dimensions and dated to Late Cypriot III (1220-1075 B.C.).98 KW 381 (fig. 18; 1. 0.149), similar to KW 579, but with a straight, untwisted point, resembles an example from Enkomi that has been dated to about the same period.99

Uncovered among copper and tin ingots, arrowheads, lamps, and a dagger at the western, uppermost part of the site was a pair of bronze tongs (KW 378: fig. 19; 1. 0.525; th. 0.006). Cast as one long, narrow piece, the tool was then bent into its present shape, with a central loop, broad shoulders, and parallel blades. Its purpose aboard the ship is not understood, but it is shaped differently than, for example, smaller tongs with diverging blades which spread out immedi- ately below the top loop. The latter tools can be mani- pulated conveniently with one hand, and are well represented in the Aegean, on Kos and Crete, and in the eastern Mediterranean in Cyprus.100 They are also

97 Catling (supra n. 13) 96; Deshayes (supra n. 72) Vol. 2, 10.

98 J.-C. Courtois, Alasia III (Paris 1984) 21 no. 166, 22, 174 fig. 4.48, with additional Cypriot references.

99 J. Lagarce, "La cachette de Fondeur aux Epees (En- komi 1967) et l'atelier voisin," in C.F.A. Schaeffer ed., Ala- sia I (Paris 1971) 408-409, with fig. 17.6; for dating of the hoard,425. 100 For Kos, see Catling (supra n. 13) 99. Examples from

Crete and Cyprus are conveniently listed in Lo Schiavo et al. (supra n. 21) 23, referring to L. Vagnetti, "Testimonianze di metallurgia minoica dalla zona di Nerokourou (Kydo- nias)," SMEA 25 (1985) 155-73.

1988] 19

CEMAL PULAK

Fig. 19. Bronze tongs KW 378. 1:5

commonly represented in Egypt, both in tomb paint- ingsl01 and as archaeological finds.'02

The broad shoulders of the Ulu Burun tongs probably required both hands for use. Similar tongs from the Late Bronze Age have been found in Cyprus,103 Syria,'04 and Palestine.105 Six tongs are known from Sardinia, all with broad, marked shoulders.106 De- shayes lists several from Iran.'07 Each of these has a collar between the open loop at the top and the broad shoulders below, with the exception of two Sardinian tongs and the Akko find. The Akko tongs are dated by

101 E.g., the Tomb of Hepu at Thebes, in H.H. Coghlan, Notes on the Prehistoric Metallurgy of Copper and Bronze in the Old World (Pitt Rivers Museum, Occasional Papers on Technology 4, Oxford 1951) 68-69, with fig. 10, also in Bass (supra n. 11) 65 fig. 76; and the Tomb of Rekh-mi-Re', in N. de G. Davies, The Tomb of Rekh-mi-Re' at Thebes (New York 1973, repr. of 1944 ed.) 52, pls. III, LII and LV. 102 Pendlebury (supra n. 80) Vol. 1, 141; Vol. 2, pl.

LXXIX.3:162. 103 Catling (supra n. 13) 99, fig. 11.4-5, pl. 10.a-b; V.

Karageorghis, "A Late Cypriote Hoard of Bronzes from Sinda," RDAC (1973) 75-76 with fig. 2.5, pl. VIII no. 5. 104 C.F.A. Schaeffer, "Nouvelles fouilles et decouvertes de

la mission archeologique de Ras-Shamra dans le palais

their excavator to the 14th century B.C. and, with their more gentle shoulder bend, resemble the Ulu Burun pair most closely. The Ulu Burun tongs show no evidence of a collar.

Weapons The 1985 campaign produced a wealth of weapons

including short swords, a dirk and dagger, spears, arrowheads, and stone maceheads.

-I

-I

Fig. 20. Bronze sword KW 275. 1:4

d'Ugarit (Campagne 1951)," Annales archeologiques de Sy- rie 2 (1952) pl. III.1. 105 For Megiddo, see P.L.O. Guy and R.M. Engberg, Me-

giddo Tombs (Chicago 1938) pl. 125.10; for Akko, see Ben- Arieh and Edelstein (supra n. 37) 31, 37 fig. 15.3, pl. XI.12. 106 Lo Schiavo et al. (supra n. 21) 23-25 with fig. 9; Lo

Schiavo (supra n. 21) 8, 30, fig. 6; L. Vagnetti, "Cypriot Elements Beyond the Aegean in the Bronze Age," in Acts of the International Archaeological Symposium "Cyprus Be- tween the Orient and the Occident" (Nicosia 1986) 206. I thank one of the anonymous reviewers of this paper for the last reference. 107 Deshayes (supra n. 72) Vol. 2, 163.

[AJA 92 20



Bronze short swords KW 301 and KW 275 are of unrelated types; their close proximity may point to their having been stored together. Nearly identical in length, both weapons are shorter than conventional swords of the period, and, although they should perhaps be classified as dirks, we will refer to them as short swords because of their borderline lengths.'08

Sword KW 275 (fig. 20; 1. 0.454) is cast in one piece. The excellently preserved pointed blade becomes widest at about mid-length after which it tapers inward slightly toward the hilt. At the base of the double-convex sectioned blade are two decorative bands running across the blade, each comprised of three finely incised longitudinal grooves. The widen- ing of the blade at its juncture with the hilt forms two protuberances which serve as a hand guard. The edges

108 Swords, dirks and daggers are separated arbitrarily by length. D.H. Gordon classifies weapons between 14 and 20 in. long as dirks; daggers are shorter; swords, called short or long on the basis of their lengths, are longer than dirks: D.H. Gordon, "Swords, Rapiers and Horse-riders," Antiq- uity 27 (1953) 67. V.R.d'A. Desborough, The Last Myce- naeans and Their Successors (Oxford 1964) 67, retains the

of the hilt are raised or flanged for the framing of hilt-

plates. These plate inlays, fully preserved by a protective layer of encrustation on one side of the sword and partially on the other, are a composite of wood, shaped to fit the contours of the hand guard and pommel, and a long bar of ivory for the grip. A short third piece of wood and another small piece of ivory placed between the grip and pommel inlays probably wedged the whole assembly between the flanges. In addition to the

wedging, a black substance around the inlay edges may have been used for affixing the pieces to the hilt. There are no rivets on the sword.

Daggers or swords usually cast in one piece and having simple flanged tangs are of Maxwell-Hyslop's Type 31.109 That flanges as a hafting device originated in the Near East rather than in the Aegean, as initially thought, is demonstrated by Sandars.10 The type is fairly common on the Late Bronze Age Syro-

I


same general divisions, but alters the classification somewhat by referring to the mid-range weapons as short swords. 109 R. Maxwell-Hyslop, "Daggers and Swords in Western

Asia," Iraq 8 (1946) 33-35. 10 N.K. Sandars, "The First Aegean Swords and Their

Ancestry," AJA 65 (1961) 22-24.

1988] 21

CEMAL PULAK

Palestinian coast as reported from a number of sites, including Ras Shamra, Megiddo, Beth-Shan, She- chem, Gezer, Bahan, Tel Mor, Tell Fara, and Tell el- Ajjul.11 Daggers from Cyprus with similar flanged grips and crescentic pommels are attributed to the Le- vant. 12 A good parallel, although shorter, comes from the 14th-century tombs near the Persian Garden, just north of the ancient tell of Akko.113

The second sword (KW 301: fig. 21; pres. 1. 0.455, max. blade th. 0.006) is a typical Aegean product. With its single-piece construction-tang and blade cast together-and ribbed blade, flanged grip and cru- ciform shoulders with rounded lobes, the sword seems to fit Sandars's Class Di of later Aegean types.114 Un- fortunately, the sword's pommel is missing, but close examination of the badly deteriorated end has shown the pommel-tang to be most probably of the unflanged type, typical of Di swords. The sword's hilt-plates have not survived, but one rivet hole in the grip and two others low in the blade affixed the plates to the hilt. Since the ivory inlays of "Levantine" sword KW

Fig. 23. Bronze dirk KW 296. 1:4

"l Maxwell-Hyslop (supra n. 109) 35-36; Ben-Arieh and Edelstein (supra n. 37) 33 n. 43. 112 Catling (supra n. 13) 128, pi. 15.i-1. 113 Ben-Arieh and Edelstein (supra n. 37) 33, 40 fig. 18.1,

pl. VII.1, dated on p. 36; also B.M. Gittlen, "The Murder of the Merchants near Akko," Biblical and Related Studies Presented to Samuel Iwry (Winona Lake, Ind. 1985) 64. 114 N.K. Sandars, "Later Aegean Bronze Swords," AJA 67

(1963) 123-25, 146-48; additional references to Di swords

Fig. 24. Bronze dagger KW 621. 1:4

275 have survived, and those made of semiprecious materials would be virtually indestructible, the hilt- plates of KW 301 probably were fashioned from a less durable material such as wood. The sword is shorter than average for its class and its blade has a broad flattish midrib with three very fine ribs outlined by fine grooves instead of the characteristic well-marked high midrib. Other short weapons of this class without midribs or with only a broad flattish midrib, each with or without central ribs and grooves, do exist., i5

A third, poorly preserved sword (KW 155: fig. 22), possibly of a third type, was discovered during the 1984 campaign. Now cleaned and conserved, the blade's preserved length is 0.37 m. and its maximum thickness 0.009 m., but its tip and most of its tang are missing. The position of the hilt is revealed by three rivets of the large size found on some typologically earlier swords of its class. 1 Cylindrical in shape, but slightly constricted at the middle where they pierce the blade, the rivets are forged into conical caps at their ends. Two of the rivets are placed at shoulder level and the third is centrally located above the other two, an invariable arrangement for Sandars's Class Ci swords.ll7 Unfortunately, almost nothing remains of

are in J. Driessen and C. Macdonald, "Some Military As- pects of the Aegean in the Late Fifteenth and Early Four- teenth Centuries B.C.," BSA 79 (1984) 70-71. 115 Driessen and Macdonald (supra n. 114) 73 nos. 6, 9, 10,

20, 22, 25, 35, 36, and (incomplete) 32, 33, 38, none originally exceeding 0.5 m. in length. 116 Sandars (supra n. 114) 120. 117 Sandars (supra n. 114) 119.

[AJA 92 22


our hilt, let alone the pommel, for positive identification. Furthermore, the blade is far too poorly preserved to reveal its original section, although a well- defined midrib is still in evidence. The total length of the sword may have been over 0.60 m., about the minimum size for Class Ci.

Both Sandars and Catling believe that the finest

specimens of Ci and contemporaneous Di swords were made at LM II Knossos, and that the swords continued to be produced until the fall of Knossos. 18

Macdonald suggests the possibility of other mainland centers"9 which may have continued to produce the same type of swords. Whatever the case, elaborately decorated examples were replaced by more efficient and utilitarian types during the 14th century.120 Al- though the variant of Di swords without high midribs is confined mainly to dirk-sized weapons not exceeding 50 cm. in length, the loss of the midrib is recognized as a development leading to the moderately thickened double-convex blades without true midribs. Typologically, therefore, the variants of the Di class which stand at the end of the ornamental sword-making of the 15th and early 14th centuries, to which we may add the Ulu Burun example, form the transition to later swords. 21

An incomplete but well-preserved dirk (KW 296: fig. 23; pres. 1. 0.335) was recovered with bronze tongs KW 378, terra-cotta lamps KW 262 and KW 485, and a number of small Type lb copper and fragmen-

-e

tary tin ox-hide ingots in areas M-10 and N-10. A few arrowheads were also reclaimed from among these ob-

jects. The dirk's hilt and blade are cast in one piece; the grip side flanges near the hilt and blade juncture have been extended and folded over the hilt-plates in the shape of wings. The wooden hilt-inlays are partially preserved, but the pommel is missing. This dagger type, Maxwell-Hyslop's Type 32, appears to be confined mainly to Syria and Persia.122 In its developed form it is found in Syria during the 15th to 13th centuries B.C. and later in Iran. In addition to good but late parallels from Iran, a nearly identical dirk, also incomplete, comes from a 14th-century hoard at Ras Shamra.123

Dagger or knife KW 621 (fig. 24) has a preserved length of 0.284 m. Incised with a five-pointed star near the juncture of the hilt and blade, this poorly preserved piece appears to be of the same general type as blades KW 1 and the slightly larger KW 189 discovered in 1984.124 Similar daggers are found in Egypt,'25 but the closest parallels from Palestine are found at Tell el-Ajjul where they are described as a Canaanite type based on Egyptian prototypes.'26

Thirty-two arrowheads have been recovered, mainly in two concentrations. Their poor preservation and extensive encrustation make their identification difficult, and the actual number of arrowheads may be slightly higher. The larger group, with 17 pieces, was in and around grid square M-10; the smaller concen-

-e:

4;

Fig. 25a. Bronze arrowhead KW 572. 1:2; b. Bronze javelin(?) point KW 286. 1:2; c. Bronze blunt arrowhead KW 429. 1:2

118 Sandars (supra n. 114) 126-27; E.A. Catling and H.W. Catling in M.R. Popham, E.A. Catling and H.W. Catling, "Sellopoulo Tombs 3 and 4, Two Late Minoan Graves near Knossos," BSA 69 (1974) 243, 252. 119 Driessen and Macdonald (supra n. 114) 64-65. 120 H. Matthaus, "Two Mycenaean Bronzes," BSA 74

(1979) 168. 121 Matthaus (supra n. 120) 169; Sandars (supra n. 114)

130-32. 122 Maxwell-Hyslop (supra n. 109) 36-38, pl. IV.32-32a. 123 Schaeffer (supra n. 75) 258-60 with figs. 223-224.17. 124 Bass (supra n. 1) 282, 283 ill. 17. 125 Pendlebury (supra n. 80) Vol. 1, 135; Vol. 2, pl.

LXXVI.10. 126 Bass (supra n. 1) 282 with ns. 57-58.

1988] 23

B-?

CEMAL PULAK

tration of about 11 arrowheads is mainly from J-10. Most of the few better preserved pieces appear to be narrow and long, with long tangs of square or rhom- boid section. The only seemingly complete arrowhead devoid of encrustation (KW 572: fig. 25a) is approximately 0.096 m. long and 0.01 m. wide at the blade. The general blade type is well known in Egypt, Syria- Palestine, and Cyprus,127 where Catling ascribes his sub-elliptical barbless category to the Near East.128 That these arrowheads were also manufactured on Cyprus, however, is demonstrated by a limestone mold found at Hala Sultan Tekke.'29

Several of the longer and more slender points may be for javelins instead of arrows.'30 One such blade, KW 286 (fig. 25b), has a preserved length of 0.107 m., but was originally at least 0.12 m. long.

Blunt-tipped arrowhead KW 429 (fig. 25c) is ba- sically a cone with a square-sectioned tang attached to the cone's apex. The blunt head, designed to strike rather than to penetrate, was probably used for fowl- ing.131 Blunt arrowheads are known from a number of sites in the Near East, including Tell el-Ajjul,132 Gezer,133 Megiddo,'34 Beth-Pelet,135 Akko,136 and Lachish.137

The southern sand gully yielded four more spearheads, in areas K-12, K-13 and J-12, with a fifth farther downslope in 1-14. Almost all are heavily encrusted and fragile, especially around their sockets. At least three types of spearheads are represented. All are

Fig. 26. Stone mace-head KW 486. 2:5

127 Ben-Arieh and Edelstein (supra n. 37) 35 with ns. 58-62, and Catling (supra n. 13) 130-31 with n. 5, include references to a number of Near Eastern sites. 128 Catling (supra n. 13) 130-31. 129 K. Nicolaou, "Archaeological News from Cyprus, 1975," AJA 81 (1977) 525 with fig. 4; P. Xstrom "The Bronzes of Hala Sultan Tekke," in Muhly, Maddin, and Karageorghis (supra n. 20) 178, pl. XVIII.1.

130 Ben-Arieh and Edelstein (supra n. 37) 34-35. 131 F. Petrie, Ancient Gaza II. Tell el Ajjil (BSAE 54,

London 1932) 8. 132 F. Petrie, Ancient Gaza I. Tell el Ajjul (BSAE 53, Lon-

don 1931) pl. XXI.106; Ancient Gaza II (supra n. 131) 8, pl. XVII.161-168; Ancient Gaza III. Tell el Ajjiul (BSAE 55, London 1933) 6, with pl. IX.34.

socketed and all have a central ridge or midribs. One spearhead, relatively clean of encrustation on its lower half, appears to have a closed or undivided socket.

Fragmentary stone mace-head KW 278 was found near the ovoid mace-head excavated in 1984 in area L-11, and smaller mace-head KW 486 (fig. 26; max. diam. 0.067) was recovered only about a meter distant in grid square M-10. Both are of the same slightly flattened spherical form with a raised collar around one end of the haft socket. KW 278 appears to be marble, while KW 486 is of a polished blackish stone, perhaps diabase.

A collared mace-head was also found on the Cape Gelidonya shipwreck; other specimens are noted from Anatolia, Palestine, and Cyprus.138

Jewelry Amber, faience, and stone beads were found in

quantity. The 14 amber beads found in 1985 vary greatly in shape and size. Many appear to be in the form of pierced, naturally occurring "pebbles" with minimally worked surfaces.139 Such pebbles, when found and collected, often have a flat side and a dome, or an irregular, lopsided shape with flattened surfaces. A thoroughly worked piece with an artificially carinated edge can be described as a flattened biconical disc.

Nine of the amber beads were found in the lower part of the southern gully (grid square K-13), the upper reaches of which yielded the majority of exotica recovered in 1984. Three beads are from area K-14, directly below the gully, and two others, probably strays from the main cluster, are from adjacent areas. A small bead fragment was found inside Canaanite amphora KW 428 in area L-12. A rock ridge sep- arates the southern gully from this jar and other rich artifacts to the jar's north in L-11, M-11, and N-1. This fragmentary bead is one of the strongest clues associating the objects from these two regions as hav-

133 R.A.S. Macalister, The Excavation of Gezer (London 1912) pl. LXXV.19-20. 134 Loud (supra n. 32) pl. 175.30-34; Guy and Engberg

(supra n. 105) pl. 126.5-9. 135 E. Macdonald, J.L. Starkey and L. Harding, Beth-

Pelet II (BSAE 52, London 1932) pl. LV.266-267. 136 Ben-Arieh and Edelstein (supra n. 37) 35, 44 fig. 22.17. 137 Tufnell (supra n. 64) 79, pl. 25 nos. 27, 32, 47. 138 Bass (supra n. 11) 126-28 with figs. 134.ST2 and 136.ST2. A piriform mace-head with raised collar on the lower end of the haft socket was found at Tarsus: H. Gold- man, "Excavations at Gozlu Kule, Tarsus, 1937," AJA 42 (1938) 35 with fig. 13. 139 A. Harding and H. Hughes-Brock, "Amber in the

Mycenaean World," BSA 69 (1974) 154.

24 [AJA 92


ing been originally stored in a common area on the

ship before it broke apart. It is possible that all the amber beads were part of the same necklace, but whether the other bead types found nearby belonged to the same necklace is not known. Curt W. Beck has identified the amber bead from amphora KW 428 as Baltic amber.140

Fourteen pale, gray-blue faience beads are of two

types. The more common type, examples of which were also found in 1984, is lentoid or biconical with

impressed spokes radiating from the centers of both faces. In some examples one face is more worn than the other, and in others only one face appears to be decorated. In better preserved specimens, however, the impressed fluting is deep and well defined on both surfaces, and forms pronounced crenations around the bead edges. Biconical faience beads of the same general form, but usually with more closely spaced flutes, are found commonly in LBA contexts in Cyprus,'14 the Syro-Palestinian coast,142 and the Aegean. 43 The second type, of the same bluish color, is ovoid or olive-

Fig. 27. Stone beads. 2:1

140 Bass (supra n. 1) 286. For a brief account of amber analysis and the validity of infra-red spectroscopy, see A.F. Harding, The Mycenaeans and Europe (London 1984) 59-60. 141 L. Astr6m, SwCyprusExp IV. Pt. 1D, 521, category 2b;

Dothan and Ben-Tor (supra n. 52) 128, 129 fig. 59.10; Courtois (supra n. 98) 147 no. 1226, 215 fig. 45.3; M. Yon, Salamine de Chypre II. La tombe T.I du XIe s. av. J.-C. (Paris 1971) 21 no. 42, pls. 16.42 and 17.42. 142 Ben-Arieh and Edelstein (supra n. 37) 26, 28 fig. 14.14;

Tufnell et al. (supra n. 64) 88, pl. 29.41; O. Tufnell, C.H. Inge, and L. Harding, Lachish II. The Fosse Temple (Lon- don 1940) 74, pl. XXXV.55; Y. Yadin, Hazor III-IV (Je- rusalem 1961) pl. 294.2. Bass (supra n. 1) 287 cites a bead from Alalakh because of the similarity of its spokes. 143 C.W. Blegen, Prosymna (Cambridge 1937) 308-309;

Fig. 28. Quartz bead KW 379. 2:1

shaped with incised longitudinal grooves. Similar ex-

amples are again from Cyprus,'44 Syria-Palestine,145 and Greece.146 Many of the faience beads have been found on the wreck with delicate objects such as gold jewelry fragments, the bone scarab, and the stone

plaque, as well as with small balance-pan weights. Area K-13 also yielded nearly half of the 11 stone

cylindrical beads, with the remaining stone beads collected from adjacent grid squares. This limited disper- sal suggests a single source, perhaps a bag or necklace

string. The beads are cut from a white or cream-colored stone, perhaps agate, with each piece having its own pattern of white, brown, or black patches and striations (fig. 27). Most are irregular in section, varying from 6-11 mm. in diameter and 3-6 mm. in thickness. Others include a cylinder of rock crystal (quartz) (KW 379: fig. 28; 1. 0.023; diam. 0.012), a bead of unidentified white stone, perhaps meerschaum, and a

cylindrical bead of bone. Three silver bracelets or armlets of the types found

during the previous campaign were excavated in the lower reaches of the same gully. Open-ended bracelet KW 284 (fig. 29), decorated with a series of chevrons below incised threads at each end, resembles KW 92 found in 1984.147 This design is paralleled in gold at Tell el-Ajjul,'48 and in bronze at Byblos,'49 but simi-

W. Rudolph, "Die Nekropole am Prophitis Elias bei Tiryns," Tiryns VI (Mainz 1973) 118, pl. 25.33; S. Hood, G. Huxley, and N. Sandars, "A Minoan Cemetery on Upper Gypsades," BSA 53-54 (1958-1959) 245, 257 fig. 34.112. 144 Xstrom (supra n. 141) 521 category 3. 145C.F.A. Schaeffer, Ugaritica IV (Paris 1962) 119, 98

fig. 81; Tufnell, Inge, and Harding (supra n. 142) 74, pl. XXXV.63. 146 Rudolph (supra n. 143) 59 pl. 32.14 and 19. 147 Bass (supra n. 1) 288, 290 ill. 27. 148 Petrie (supra n. 131) 6, pls. I-II; F. Petrie, Ancient

Gaza IV. Tell el Ajjil (BSAE 56, London 1934) 7, pls. XV, XVI.68. 149 Dunand (supra n. 89) 174, pi. LXXII.2540 and 2542.

1988] 25

CEMAL PULAK

Fig. 29. Silver bracelet KW 284. 2:3

lar bracelets and armlets without chevrons are also found in gold and silver at Tell el-Ajjul;'50 and in silver at Megiddo and Gezer.'15 Silver bracelet KW 273 (fig. 30), paralleled on Cyprus,152 is decorated with crosshatching between a pair of incised lines, the in- nermost pair of which has a series of chevrons below. Fragmentary bracelets, fragments of silver bars, and a crumpled silver plate recovered from the wreck are clearly scrap silver intended for recycling, but whether the intact bracelets are part of this scrap hoard cannot be ascertained at present. 53

Nearby were two rings, most likely fashioned from top shells (fig. 31). A central groove inscribed completely around each ring preserves a black, bitumen- like substance probably intended for affixing pre-cut inlay parts to the ring. On at least one ring (KW 414: fig. 31b; diam. 0.02; th. 0.003), an impressed zigzag pattern in this black material may correspond to the triangular shapes of missing inlay pieces. Top-shell rings are found in LBA contexts in the Aegean, Cyp- rus, and Syria.'54

Excavators discovered a few small gold items mainly

'50 Petrie, Ancient Gaza IV (supra n. 148) 5, 8, pls. XII, XIX, XX.155-158; O. Negbi, The Hoards of Goldwork from Tell el-'Ajjul (SIMA 25, Goteborg 1970) 49.

151 Loud (supra n. 32) pl. 226.3; Macalister (supra n. 133) 99-100 with fig. 286.

152 F.H. Marshall, Catalogue of the Jewellery, Greek, Etruscan, and Roman in the Department of Antiquities, British Museum (Oxford 1911) 39 no. 607, pl. V.607. 153 Other hoards of precious metals include intact pieces

alongside fragments of folded and cut material: e.g., Petrie, Ancient Gaza IV (supra n. 148) 5, pl. XII; J.D.S. Pendle- bury, "Preliminary Report of Excavations at Tell el-'Amar-

in and around area L-1 1. A badly deformed gold flower (KW 361), perhaps the finial for a pin, is decorated with granulation on the inner surface of its petals. The unopened petals clustered inside also are granulated.

Scrap gold from a disc pendant (KW 551: fig. 32 diam. ca. 0.036) was found in the same general area. Its central design, comprising horizontal petals or leaves with two vertical lines through their centers, all worked in repousse and delineated with a single irregular row of gold granules, probably represents a highly stylized and crudely executed lotus palmette of McGovern's Type IV.F.4.155 The medial line on the left center petal or leaf was undoubtedly destroyed during cutting; two other lines, inserted between the petals, radiate outwards. A single rolled-over ribbon loop for suspension, worked with banded edges, is affixed to the top. The lotus palmette on the Ulu Bu- run pendant is similar to palmettes on eight pendants, probably worked in repousse, found together at Beth Shemesh in LB IIB context.156 Pendant KW 551 had been deliberately cut in two in antiquity, with each half then folded over and beaten flat, in one case with bits of gold foil and wire inside; the spherical granules had been reduced to irregular discs during this beat- ing. The pendant was originally formed of two sheet- gold discs with the edges of the slightly larger, undecorated back plate folded around and crimped over the front plate, as were the edges of gold pectoral KW 94, found in 1984.157 This technique was common in the LB IA hoard jewelry at Tell el-Ajjul, which includes

Fig. 30. Silver bracelet KW 273. 2:3

nah 1930-1," JEA 17 (1931) 236, pl. LXXIII; C.F.A. Schaeffer, "Les fouilles de Minet-el-Beida et de Ras-Sham- ra, troisieme campagne," Syria 13 (1932) 22, pl. XVI.1; Karageorghis and Demas (supra n. 51) 60-62, pl. XXVIII.

154 D.S. Reese, "Topshell Rings in the Aegean Bronze Age," BSA 79 (1984) 237-38, and personal communications with David S. Reese.

155 P.E. McGovern, Late Bronze Palestinian Pendants (JSOT/ASOR Monograph Series Sheffield 1985) 47. 156 McGovern (supra n. 155) 47 with fig. 41.139. 157 Bass (supra n. 1) 287, pl. 17 fig. 3.

[AJA 92 26


aM

Fig. 31. Shell rings. a = KW 357, b = KW 414. 4:3

more than one folded plaque-pendant "forming a mere dump for the melting pot."'58 A falcon or wryneck earring from Hoard 277 at Tell el-Ajjul appears from a photograph to be joined in a similar manner, the crimped edges around the periphery of the decorated front plates being clearly visible.159 The similarity of manufacturing technique between the Tell el- Ajjul falcon earrings and the Ulu Burun pectoral has been noted.160 Two other earrings of unknown pro- venience in Leiden, also suspected to be products of Tell el-Ajjul workshops,161 are even closer parallels to pectoral KW 94. It is unlikely that the Ulu Burun piece is contemporaneous with these earlier examples; the former was probably manufactured at a later date and at a different center, perhaps Ugarit, where it is suggested that the tradition of gold jewelry working eclipsed at Tell el-Ajjul may have prospered from the early 15th century until the end of the 14th century.162 The similarity in workmanship of pectoral KW 94 and pendant KW 551 may suggest a common school for both.

Similar circular pendants with single ribbon loops for suspension, but made from a single sheet of gold, are common to many western Asiatic sites.163 Their common theme is a star with varying number of points or rays, all in repousse. These roundels have been regarded as amulets representing divine symbols, perhaps combining the apotropaic qualities of the deities they represented.'64

Such may also have been the function of large cir-

158 Petrie, Ancient Gaza IV (supra n. 148) 5. 59 0. Tufnell, "Some Gold Bird Ornaments: Falcon or

Wryneck?" AnatSt 33 (1983) 62-63 no. 4, pl. XXI.4a. 160 Pectoral KW 94 is discussed in detail by Bass (supra

n. 1) 287-88. 161 K.R. Maxwell-Hyslop, Western Asiatic Jewellery c. 3000-

612 B.C. (London 1971) 117-18. 162 Negbi (supra n. 150) 37. 163 Negbi (supra n. 150) 34-35; T.L. McClellan, "A Syrian

Fortress of the Bronze Age: el-Qitar," National Geographic Research 2 (1986) 435 with fig. 15.B-C.

cular star pendant KW 138 (diam. ca. 0.075) recovered from the wreck in 1984.165 Circular pendants at the necks of Syrians in the merchant fleet depicted in the Tomb of Kenamon at Thebes probably are similar devices.166

The upper half of an inscribed gold signet ring (KW 603: fig. 33; h. of bezel 0.014), purposely cut from an intact ring, was uncovered less than a meter away from the pendant. Its color, palest of any gold object recovered from the wreck, may derive from the ring's alloy. That the ring suffered the same fate as pendant KW 551 is obvious from chisel scars along the cut; the ring clearly was part of a hoard of scrap gold aboard the ship. The ring may have been cut in half to prevent its unauthorized use, but it is not yet known if the other half was also on the ship, as were the two halves of the circular pendant. Three engraved figures on the oval bezel have been damaged by the cutting. From photo- graphs and a cast, James Weinstein has suggested that although the composition and general shape of the ring reflect an 18th-Dynasty date, the crudely incised signs are not indicative of the high-quality workmanship generally expected of New Kingdom artisans. A defin- itive interpretation of the signs cannot be made without the missing half.167

Fig. 32. Gold roundel KW 551. 1:1

164 Maxwell-Hyslop (supra n. 161) 149.

165 Bass (supra n. 1) 289-90, pl. 17 fig. 4. 166 N. de G. Davies and R.O. Faulkner, "A Syrian Trading

Venture to Egypt," JEA 33 (1947) 40-46; N. de G. Davies, Private Tombs at Thebes IV. Scenes from Some Theban Tombs (Oxford 1963) pl. XV. 167 1 thank James Weinstein for his comments on the Egyp-

tian finds. Weinstein notes that the incised signs include a maat-feather (oriented in the wrong direction) on the left of the ring, a standing ba-bird in the center, and on the right a badly cut seated figure, probably of a female as suggested

1988] 27

CEMAL PULAK

Fig. 33. Gold signet ring KW 603. 3:1

Nearby, a scarab framed in gold (KW 338: fig. 34), an empty frame for another scarab (KW 479), and a

rectangular stone plaque carved with hieroglyphs were uncovered (KW 481: fig. 35). The scarab, carved of bone or ivory, is perforated longitudinally. The

piece is set in an oval frame formed of sheet gold; at each end, surrounding the perforation, is a rolled cylindrical sleeve of gold. On the scarab's base is an ornamental assemblage of hieroglyphs, but the promi- nent gold frame around the scarab probably rendered it unsuitable for use as a seal. It was more likely an amulet. Weinstein informs me that the placement of

hieroglyphs in three columns is common to the second half of the Second Intermediate Period. Also typical of scarabs from this period are the apparently unread- able combination of signs, the smooth back, the small notches cut along the sides to indicate the division between the prothorax and elytra, and the schematically rendered trapezoidal head and clypeus.168

In Syria-Palestine, however, the scarabs of LB IA continue Middle Bronze Age traditions down to the

early 15th century when 18th-Dynasty scarab types become common. Hence, if the scarab is a Levantine rather than an Egyptian product, it could have been made at any time during this period or even slightly later. The Ulu Burun scarab is similar to an example from Palestine with a different combination of signs.169

The empty gold frame (KW 479), for a smaller scarab, is similar to the frame of KW 338. Slightly dis- torted, it possibly was part of the scrap gold on board.

by the long headdress. The missing lower half of the ring probably contained other signs which may have formed an epithet. 168 James Weinstein, letter of 31 October 1986. 169 A. Rowe, A Catalogue of Egyptian Scarabs, Scaraboids,

Seals and Amulets in the Palestine Archaeological Museum (Cairo 1936) pl. 11.59; I thank James Weinstein for this

Rectangular plaque KW 481 (fig. 35; 1. 0.016; w. 0.01; th. 0.006) was cut carefully from a soft gray- green stone, probably steatite. It is pierced longitudinally for suspension, and neatly executed hieroglyphs on both of its wider faces refer to the god Ptah with

accompanying standard phrases: Lord of Truth, and Perfect in Favors. According to Weinstein, most of the small New Kingdom plaques are inscribed for Amun or Amun-Re and the naming of Ptah makes this piece unusual. It probably does not belong to the

very beginning of the 18th Dynasty, when hiero-

glyphs were more sharply cut, thinner, and somewhat crudely executed; it is likely of 15th- or 14th-

century manufacture.170 Weinstein notes that the 18th-19th Dynasty scarab

and plaque from Lachish have the same general inscriptions.17 To these we may add the rectangular plaque, again with the name Ptah but accompanied by a different series of signs, from Tomb 982 of the late 19th Dynasty at Beth-Pelet,172 and a specimen also

reading "Ptah, Lord of Truth," dated by Petrie to the mid-18th Dynasty.173

Fig. 34. Scarab KW 338. 3:1

reference. 170 Weinstein (supra n. 168). 171 Tufnell (supra n. 64) pls. 39-40.376-377. 172 Macdonald et al. (supra n. 135) 26, pl. LVII.357. 173 F. Petrie, Buttons and Design Scarabs (BSAE 38, Lon-

don 1925) 19, pl. XI.624.

[AJA 92 28


An unexpected find is bronze pin KW 570 (fig. 36), accidentally broken during excavation. Two fragments join cleanly, while the third, globed section does not. Either the break has suffered subsequent damage, or a smaller fourth piece, now lost, is responsible for this discrepancy. The shank and its slightly prolate globe were cast as one piece; the pin's head and lower shank were lost in antiquity. The original length of the pin cannot be reconstructed, but the existing pieces, when joined end-to-end, are about 10.7 cm. long, the maximum shaft diameter being 2 mm.

Long pins seem to have been popular in Central Europe during the 14th and 13th centuries B.C.,174 but appeared in Greece only at the end of the LH IIIC period and beginning of Submycenaean,175 becoming more common during the Dark Ages. Most scholars now agree in deriving the general shapes of the Greek pins from some region to the north of Greece, although pins of these two regions are not identical.176 Whether the appearance of long pins (and fibulae) in Greece heralds, as has been suggested, a change in clothing fashion-possibly due to a change in climate and/or a change of population-remains speculative.177 Of long pins, the type with a small nail-like

Fig. 35. Stone plaque KW 481. 3:1

174 J. Bouzek, The Aegean, Anatolia and Europe: Cultural Interrelations in the Second Millennium B.C. (SIMA 29, Goteborg 1985) 166. 175 Bouzek (supra n. 174) 167. 176 P. Jacobsthal, Greek Pins and Their Connexions with

Europe and Asia (Oxford 1956) 181; Desborough (supra n. 108) 54; V.R.d'A. Desborough, The Greek Dark Ages (London 1972) 298-99; Bouzek (supra n. 174) 165-67. 177 V. Milojcic, "Die dorische Wanderung im Lichte der

vorgeschichtlichen Funde," AA 63/64 (1948/1949) 12-36; Desborough (supra n. 108) 56; Bouzek (supra n. 174) 167; M.S.F. Hood, "A Note on Long Bronze Pins," in J.N. Cold- stream and M.S.F. Hood, "A Late Minoan Tomb at Ayios Ioannis near Knossos," BSA 63 (1968) 214-216, 218. I. Ki- lian-Dirlmeier ("Der dorische Peplos: ein archaologisches Zeugnis der dorischen Wanderung?" ArchKorrBl 14 [1984] 288-89), however, observes a gradual evolution in the pro-

r

?

Fig. 36. Bronze pin KW 570. 1:2

head and a globe toward the top of its shank,178 to which the Ulu Burun example conforms most closely, appears to be confined to central Greece during the early Submycenaean period. The Greek pins differ from similar pins from Anatolia, Cyprus, the Balkans, Italy, and Central and Eastern Europe by the placement of their globes farther down their plain shanks.179 Thus, Greek pins appear to be indepen- dently produced even though their general form is probably of foreign inspiration.180 Some authors, however, suggest a connection between this form and the much earlier composite pins from Shaft-Grave Circle B at Mycenae.181 The earliest examples in Greece of this globed type, apparently preceding other types of long pins, come from several LH IIIC contexts.182 One earlier exception may be a pair of large bronze pins from Chamber Tomb 61 at Mycenae. Akerstr6m, however, rejects the general LH II-IIIA dating for this series of tombs. He assigns the pins and associated fibulae to later, secondary burials within the same tomb, partly on the basis of the Submyce-

duction of pins, burial customs, and clothing fashion during the transition from Mycenaean to Submycenaean times, rather than a sudden change. I owe the last reference to one of the anonymous reviewers of this paper. 178 This pin belongs to Jacobsthal's classical Submyce-

naean Type ([supra n. 176] 1-2); Desborough's Type A ([supra n. 176] 296 fig. 33.E, 297), and A.M. Snodgrass's Type I (The Dark Age of Greece [Edinburgh 1971] 226,227 fig. 81.1). 179 Bouzek (supra n. 174) 166, 164 fig. 84; Harding (supra

n. 140) 136; Catling (supra n. 13) 239, pl. 41j. 180 Harding (supra n. 140) 136-37; Bouzek (supra n. 174) 167. 181 J. Deshayes, Argos, Les fouilles de la Deiras (Etudes

peloponnesiennes 4, Paris 1966) 204-207; Snodgrass (supra n. 178) 226. 182 Bouzek (supra n. 174) 167.

1988] 29

CEMAL PULAK

naean character of the long pins.'83 Bouzek, on the other hand, emphasizes the reversed globe-to-head proportions of these pins and sets them apart from the typical globed type. Based on the associated fibulae, he dates the material to ca. 1200 B.C.184

The fragmentary dress pin from Ulu Burun, if found elsewhere, would have been dated to the 12th century at the earliest.

Miscellaneous Finds The 22 balance-pan weights of hematite, uniden-

tified stone, and bronze were found mostly in grid squares K-14 and M- 11. Most are of common sphendonoid and domed types, but "sugar-loaf" and discoid examples are also represented. Three solid-cast zoomorphic bronze weights (fig. 37), comprising a recumbent bovine (KW 335), a duck (KW 350), and a couchant sphinx (KW 468), were found close together. These complement a pair of bronze frog weights excavated some distance away in 1984, but not recognized before laboratory conservation (KW

237, fig. 38, weighs 6.6 gr.; lead-filled and badly corroded KW 220 has a surviving weight of 19.5 gr.). A fourth bronze weight (KW 582: fig. 37), found further upslope, was adorned with human and animal figurines on its upper surface.

The bovine, most probably a bull-calf, lies with its head positioned on the same axis as its body. The animal has the typical humped shoulder and heavy build of a bull, yet its horns are not fully developed. The legs are drawn up under the body and the tail appears in relief, curled over the hindquarters. Despite its small size (1. 0.032), the form is executed with naturalism. With a preserved weight of 16.1 gr., it may correspond to two shekels of the Babylonian standard.

The less detailed and smaller bronze duck (1. 0.022), weighing 8.3 gr., is perhaps a Babylonian shekel. Its head faces forward and its bill is pressed tightly against its neck in a simple but graceful composition.

The sphinx, recovered with a hematite sphendonoid weight concreted to its head, has a typically feline posture and tufted tail. Measuring 6.8 cm. in length, it

Fig. 37. Balance-pan weights (left to right: KW 468, KW 350 and KW 335, with KW 582 above). Approx. 1:1

183 X. Akerstrom, "Mycenaean Problems," OpAth 12 184 Bouzek (supra n. 174) 161. (1978) 77-79.

[AJA 92 30

i: fE:

:: f:S:i:

k :

, 11

i:

--


Fig. 38. Bronze balance-pan weight KW 237. 2:1

weighs about 80.6 gr., perhaps corresponding to 10 Babylonian shekels.

The remarkable bronze weight with figurines, best described as a reversed truncated cone with domed top, is hollow cast and filled with lead. Expansion of the corroding lead has extensively damaged the piece, which now weighs 410 gr. On the upper surface of the domed top are skillfully rendered bronze figurines of a man, probably a shepherd dressed in a cloak and cap, and two sheep (a lacuna next to the sheep probably held a third animal). The shepherd sits on his left leg with the left hand resting on his knee. The rather short right leg is tucked up against his chest and the right hand is spread, palm down, on the ground. Al-

though the nose has suffered some damage, other detailed facial features are still clear. The facing heads of the recumbent sheep are turned almost at right an- gles to the axes of their bodies.

Zoomorphic weights in use are recorded in tomb

paintings and representations of New Kingdom Egypt.185 That recumbent bovines are the most widespread of the zoomorphic shapes is attested by many finds from the eastern Mediterranean. In Ras Shamra (Ugarit) alone, more than a half dozen pieces, some filled with lead, have been found.'86 Cyprus has produced similar weights. A remarkable collection of 14 weights recently discovered in Kalavassos comprises three hematite sphendonoids; a cast-bronze calf and two bull heads; a reclining bull, a boar, and what ap-

185 Ben-Arieh and Edelstein (supra n. 37) 57 n. 33 for some references; F. Petrie, Ancient Weights and Measures (War- minster 1974; repr. of BSAE 39, 1926 ed.) 6. 186 J.-C. Courtois, "Le trisor de poids de Kalavassos-Ayios Dhimitrios 1982," RDAC (1983) 120-21. 187 Courtois (supra n. 186) 117-30, pl. XVII, with other

bovine Cypriot weights listed on 120-21; A.K. South and I.A. Todd, "In Quest of Cypriote Copper Traders: Excava- tions at Ayios Dhimitrios," Archaeology 38:5 (1985) 42. 188 J.B. Pritchard, Sarepta. A Preliminary Report on the

Iron Age (Philadelphia 1975) 69, fig. 62.6; Ben-Arieh and Edelstein (supra n. 37) 57-58 with n. 34, 59 fig. 25.24, 62, and pl. XX.38; Pendlebury (supra n. 80) 109, 125, pl. LXXVII.1-2.280; Petrie (supra n. 185) pl. XVI; Cour- tois (supra n. 186) 121 n. 36.

pears to be a lioness; a duck; a negroid head; and three

cylindrical bronze weights, two of which have crossed animal figures in relief on their upper surfaces.'87 Bo- vine weights, however, are extant all along the Syro- Palestinian coast and in Egypt.188 Most of those found in stratified contexts appear to date mainly from the 14th century, but some may be slightly later.

Duck-shaped weights are commonly found in the Near East, although few are of bronze.189 A small bronze duck-weight attributed to the Babylonian weight system and composed with the bird's head turned backward in the sleeping position, typical of this system, was found at Enkomi.190 A second bronze duck-weight, much heavier and with a forward-facing head on an extended neck, was in the Kalavassos cache.'91 Frog-shaped weights appear to have a similar distribution, having been found in Alalakh, Larsa, Enkomi, and Egypt.'92

Sphinx figurines, mainly pendants, small beads, and seals are found commonly in Egypt, on the Syro- Palestinian coast, and to a lesser extent on Cyprus.193 To the best of my knowledge, however, no bronze fig- urine in the form of a sphinx has been previously identified as a weight; some weight-like bronze examples appear to have been used as votive figurines.194

Although the cylindrical body of the weight with a shepherd bears some resemblance to the lead-filled bronze weights with crossed animal figures on their

upper surfaces from Kalavassos,195 the Ulu Burun weight remains unparalleled.

Three of the Ulu Burun zoomorphic weights were used frequently enough to warrant their storage with the more common types; even the smallest of these forms, we may now be sure, were used for weighing and were not amulets or the like as has been suggested.'96 Because the weights have not been fully conserved, and because we expect to find many others on the site, they have not yet been studied in detail.

Pierced lead discs KW 298 and KW 459, not yet

189 Courtois (supra n. 186) 123 with n. 49. 190 Courtois (supra n. 186) 123; Schaeffer (supra n. 92)

441, pl. A facing p. 128. 191 Courtois (supra n. 186) 123, pl. XVII.5. 192 Courtois (supra n. 98) 43-44, 185 fig. 15.35; Petrie (su-

pra n. 185) pl. IX. 193 Courtois (supra n. 98) 148 no. 1239, 215 fig. 45.25,

pl. XVI.9. 194 E.g., Dunand (supra n. 90) pi. CXVI.14499. 195 Courtois (supra n. 186) 123-25, pi. XVII.6-7. A weight

from Enkomi, similar in shape to that from Ulu Burun, but with a loop handle, bears a series of animals in low relief on its sides (Schaeffer [supra n. 99] 411, 413, 418 fig. 22c-d, 420 fig. 23a-b, 422 fig. 24.2. 196 Ben-Arieh and Edelstein (supra n. 37) 58.

1988] 31

CEMAL PULAK

with crimping impressions still visible on some, they must have been assembled on nets and were not spares or trade items. Whether as cargo or for shipboard use, one net seems to have been inside the pithos. Whatever the case, if all 21 weights are from a single net, we

may be able to reconstruct its length. Similar net

weights have been found on Cyprus at Enkomi and Athienou.200 The Governor's Tomb at Tell el-Ajjul also yielded a large quantity of net weights in groups mostly based on multiples of 17, leading Petrie to con- clude that 17 or 18 lead weights were normal for a

Fig. 39a. Lead disc KW 298. 1:2; b. Pierced hematite disc KW 325. 1:2

cleaned of concretion, are similar to weights of Aegean origin;197 the former (fig. 39a) has a crudely scratched

spiraling spoke design on its surface. Another possible weight is hematite disc KW 325 (fig. 39b). Pierced

longitudinally, it is reminiscent of lentoid seals, but its thickness, weight, and unmarked surfaces make it also suitable for use as a balance-pan weight.

Other finds from 1985 include fragments of faience vessels, mainly from the sandy area below the southern gully, where similar fragments ,rere recovered in 1984.198 The precise shapes of the vessels will not be known until more pieces are found, but a series of fragments comes from a rhyton in the form of a ram's head (KW 565). That a second rhyton was aboard the ship, suspected in 1984, is now confirmed by the dis-

covery of a second muzzle and a large section of the

rhyton itself. The two rhyta appear to be similar in

shape and probably in size. Other fragments, now being assembled in the Bodrum Museum of Under- water Archaeology by Jane Pannell, represent a female head (fig. 40). Both forms are closely paralleled in faience at Enkomi and Tell Abu Hawam.199

One hundred seven folded lead strip fishing-net weights or sinkers were recovered from the site in three clusters: 35 from area L-11, 21 from M-10, and 21 inside pithos KW 250; the remaining pieces were found mostly downslope of the first two concentrations. Since all the lead strips are folded and crimped,

197 K.M. Petruso, Systems of Weight in the Bronze Age Ae- gean (Diss. Indiana University 1978) 204 no. 274 from Kommos is similar but smaller. 198 Bass (supra n. 1) 290-91 pl. 17.5-6. 199 A.S. Murray, A.H. Smith, and H.B. Walters, Excava-

tions in Cyprus (London 1900) 33, pl. III; R.W. Hamilton, "Excavations at Tell Abu Hawam," QDAP4 (1935) 65, pls. XXVIII, XXX. I do not know of Aegean examples of these types, but E.J. Peltenberg, in a letter to G.F. Bass, pointed

Fig. 40. Faience goblet. The left eye of a woman's face is visible in the lower center; her hair shows under the cylindrical headdress. 1:1

out H. Frankfort's unverified reference to a female-headed faience goblet from Rhodes in The Art and Architecture of the Ancient Orient4 (Harmondsworth 1977) 274; I have not yet seen Peltenberg's discussion of ram's head rhyta in Ki- tion V. 200 Courtois (supra n. 98) 47, 185 figs. 15.15A-15B, 236

pl. XXI.19; Dothan and Ben-Tor (supra n. 52) 135, pl. 47.4, 126 fig. 57.18-19, and esp. 20 with crimp-marks still evident.

ciD

32 [AJA 92

r?r


light net.201 A collection of 18 net weights concreted together, presumably from a single net, was recovered from the Cape Gelidonya shipwreck.202

A large, pyramidal lead weight (KW 267: fig. 41; h. 0.10), pierced at its narrow upper end by a hole for

suspension, was found with the net weights in area L- 11. Although rectangular in section, its basal corners have been deformed and rounded by heavy use. The weight may have served either as the heavy foremost sinker of a fishing net or perhaps as the ship's sound-

ing lead. Apart from the single stone weight-anchor raised in

1984, no anchors have been moved. Seven of the eight anchors uncovered to date are grouped between the two uppermost rows of copper ingots. Although some of the anchors are still partly buried, they seem to come in three sizes: three large, two intermediate, and one very small. The last, probably of marble or other

light-colored limestone, is too small to be an effective

ship's anchor, and perhaps served as a hawser weight or as a spare for the ship's boat. The eighth anchor, of the largest size, lies amidst copper ox-hide ingots at the deeper end of the wreck.

Removal of anchor KW 145 in 1984 exposed bal- last stones lying directly over parts of the ship's hull. No further work was done in this area in 1985, but after removing several ox-hide ingots and amphoras

Fig. 41. Large lead weight KW 267. 1:2

201 Petrie, Ancient Gaza III (supra n. 132) 6, pl. IX.35. 202 Bass (supra n. 11) 131, 132 fig. 139.L1. For similar net weights in the Aegean, see Iakovidis (supra n. 94) 96; I owe this reference to Mark Rose.

203 G.F. Bass and C. Pulak, "The Late Bronze Age Ship- wreck at Ulu Burun: 1986," AJA 91 (1987) 321. 204 Mountjoy (supra n. 65) 77.

slightly upslope of this area, in square P-14, we ex-

posed a large, curved timber. Because of a stone anchor lying partly above it, we could not determine if the piece was a large frame, the endpost, or the curved end of the ship's keel.

After this report was written, another campaign was conducted at Ulu Burun. Among the finds being studied for the 1986 report are a gold scarab of Queen Nefertiti, two Near Eastern cylinder seals, intact and

scrap Canaanite jewelry of gold and silver, three more

hippopotamus teeth, an ostrich eggshell, logs of "Afri- can ebony" (Dalbergia melanoxylon), Minoan and/or Mycenaean stirrup-jars and pitcher, more zoomorphic weights, beads, a ram's head rhyton of faience, an

ivory-hinged wooden writing tablet (diptych), and

many more objects similar to those found previously; the stone anchors now total 12.203

DISCUSSION

The wealth of artifacts recovered in 1985, mostly chiseled out of rock-hard concretion, reveals the years of excavation ahead. Thus, it remains early to attempt an accurate dating of the ship or a resolution of its course. A 14th-century B.C. date for the wreck has been suggested previously on the basis of ceramic evidence. The most precisely dated object, kylix KW 57, has been ascribed to the early phase of LH IIIA:2, but it could have been in use for years before its loss. Stir-

rup-jar KW 137 had been broadly dated to LH IIIA:2; the shape, however, becomes popular early in the latter part of LH IIIA:2,204 and the floral design on the jar's handle zone is a late feature,205 which, in a

slightly more stylized shape, becomes the common form of the unvoluted flower in LH IIIB.

Dating of other painted stirrup jars is less precise, for small globular jars KW 171 and KW 305 (both FS

171) fall into the general LH IIIA:2 period. Large globular stirrup jar KW 308, on the other hand, probably represents a shape which may have already become rare by the latter phase of LH IIIA:2.206

The bell-shaped profile of semi-globular cup KW 334 is characteristic of LH IIIA:21-IIIB, and a fragmentary large coarse-ware stirrup jar, not yet studied in detail, appears to be broadly dated to the LH IIIA- IIIB periods.

205 The final drawing of KW 137, Bass (supra n. 1) 291 ill. 29, the entire decoration cleaned, was added to that article as it went to press and thus was not considered in the published discussion of the jar. 206 Stirrup jar KW 308 is incomplete and unrestored, but its shape appears to be closest to FS 170 (Mountjoy [supra n. 65] 77).

33 1988]

CEMAL PULAK

The Mycenaean ceramic assemblage so far recovered, therefore, would not be out of place in the latter half of the LH IIIA:2 period. Kylix KW 57, on the other hand, is most likely of an earlier style-presumably a treasured possession-and not of contemporary manufacture.

Hankey and Warren, in their study of Aegean chronology based on Mycenaean pottery, have determined that the latter phase of the LH IIIA:2 period corresponds to the reign of Akhenaten, and that pottery of this period may not be dated any later than the time of Tutankhamun.207 Thus, it would appear that the Ulu Burun ship sank sometime during, and most likely at the end of, the Amarna period, or slightly later, and there really is no substantial evidence for a refinement of this date.

If our typological attributions are correct, the Ae- gean swords, however, appear to corroborate the dating suggested by the kylix. Although not all are se- curely dated, the contexts in which similar swords have been found are chronologically confined to LM/LH II-IIIA:2; but the smaller flat-bladed weapons, perhaps precursors to later swords of Sandars's Type Dii of LH IIIA:2 and IIIB periods,208 may have been manufactured for a longer time than previously believed. Furthermore, more valuable and durable than pottery, bronzes, especially swords, would have been treasured heirlooms. The Near Eastern sword also fits comfortably in the LH IIIA:2 range.

Most of the tools are of types with long histories, but a few, namely the deep-bar chisels, heavy chisels, and necked adzes, become relatively scarce after the 14th and 13th centuries B.C., although later, rare examples do exist. A useful chronology for dating tools, however, has not yet been developed, and doubts are even raised about using typology to establish ethnic origins of tools.209

The Mycenaean lentoid sealstone KW 134 appears to date to the second half of the 14th century, but the excessive wear on the sealstone from extended use suggests a later date for the sinking of the ship.210 With some finds tentatively dated to LH IIIA:2, and more likely to the late phase of this period, the dis-

207 V. Hankey and P. Warren, "The Absolute Chronology of the Aegean Late Bronze Age," London University Insti- tute of Classical Studies Bulletin 21 (November 1974) 142-52. 208 Sandars (supra n. 114) 130. 209 J.D. Muhly, "The Nature of Trade in the LBA Eastern Mediterranean," in Muhly, Maddin, and Karageorghis (supra n. 20) 256. 210 Bass (supra n. 1) 284. 211 Y. Portugali and A.B. Knapp, "Cyprus and the Aegean:

A Spatial Analysis of the Interaction in the Seventeenth to

covery of globed pin KW 570 was unexpected, as pins of this type do not otherwise appear until well into the 12th century. Present evidence is too scanty, the finds being only partially studied, to date accurately the de- mise of the ship, but based on evidence available so far, a date at the end of LH IIIA:2, or probably just after the Amarna period but still within the 18th Dyn- asty, would not be unrealistic.

Even if a post-Amarna date is preferred for the Ulu Burun ship, it is still close enough in time to this well- documented segment of history to assume that the general economic structure of the Levant had not changed appreciably,211 and that inferences from the Amarna letters as well as the evidence from tomb paintings could aid us in understanding the context of the Ulu Burun ship.

The only extant depiction of a Mediterranean merchant venture from the 14th century B.C. is the scene from the tomb of Kenamun at Thebes illustrating the arrival of a Syrian merchant fleet at an Egyptian port.212 Porters unload cargo including Canaanite amphoras and a pilgrim flask similar to those found on the Ulu Burun wreck. The roundels on the necks of some of the crew may represent star-disc pendants of the type already recovered during excavation. Pithos KW 251, recovered with its contents of mostly intact Cypriot pottery, allows us to speculate about the pithoi shown on ships' decks, perhaps also filled with pottery as well as other goods. Other tomb paintings depict tribute bearers carrying not only elephant tusks, Canaanite amphoras, and other vessels of the types found on the wreck, but also copper ox-hide ingots.213

The exact amount of copper on the ship will not be known until all is excavated, but a total of 200 ox-hide ingots is not an unreasonable estimate. If we assume that each ingot has an average weight of 25 kg., this would correspond to approximately 183 talents of copper in the shape of ox-hide ingots alone, reminding us of the Amarna letters mentioning shipments of 100 (EA 34) and 200 (EA 33) talents of copper from Alasia to Egypt.214 If the 500 talents of copper mentioned in one Amarna letter (EA 35), ironically ac-

Fourteenth Centuries B.C.," in A.B. Knapp and T. Stech eds., Prehistoric Production and Exchange (Institute of Ar- chaeology Monograph 25, Los Angeles 1985) 65. 212 See supra n. 166; Bass (supra n. 1) 293-94. 213 Bass (supra n. 1) 294. 214 Bass (supra n. 1) 293. Zaccagnini ([supra n. 23] 414)

noting the variations in phraseology of references to copper in the Amarna letters, translates the passages according to their formulation as "number of copper ingots" or "talents of copper."

34 [AJA 92


companied with an apology for the small quantity of copper sent, represents instead, as Zaccagnini be- lieves, only 500 shekels of copper,215 then the Ulu Bu- run ship consignment corresponds to the largest quantity of copper shipment ever recorded in ancient texts! The Amarna letters also mention possible glass ingots, elephant tusks, gold jewelry and silver, and weapons among other royal gifts, all of which are again matched on our vessel. Could this cargo, then, represent a royal shipment of the type exchanged between the Syro-Palestinian coast, Egypt, and a land called Alashia, all of which is vividly described in the Amarna letters? This certainly seems plausible when we consider that the single largest hoard of ox-hide ingots prior to the discovery of the Ulu Burun ship came from the Gelidonya wreck, which yielded only 34 ingots. The artifact assemblage from Cape Geli- donya led Bass to assign that wreck to itinerant entre- preneur-smiths sailing westward along the coast in search of goods and markets wherever available.216 The nature of trade in the Bronze Age, and of ancient economy in general, is not fully understood, but the Ulu Burun ship, carrying some of the major resources and luxury items of Asia and Cyprus, contrary to the Gelidonya venture, may represent a single shipment destined primarily for a specific port, and may in fact be our first direct evidence for state-administered trade based on gift exchange.

The bulk of the cargo, comprising copper and tin ingots, glass ingots, Cypriot pottery, ivory and Can- aanite amphoras filled with resin, all certainly taken on at a Syro-Palestinian or Cypriot port, points to the ship's east to west sailing route. The Canaanite amphoras containing terebinth resin could have been taken on at a port such as that at Ugarit.

The large pithos containing Cypriot export pottery could be of Cypriot origin. The saucer-shaped lamps stored in this pithos, however, appear to be of the type commonly found on Syro-Palestinian sites, although they are different still from the fire-blackened examples also recovered from the ship; the latter lamps, again of Syro-Palestinian type, were obviously for shipboard use and not cargo. If the unused lamps inside the pithos were Syro-Palestinian, then the pithos may have been packed at an Asiatic port such as that

215 Zaccagnini (supra n. 23) 414. 216 Bass (supra n. 11) 163-64. 217 Bass (supra n. 1) 295 n. 163. 218 G. Dossin, "La route de l'etain en M6sopotamie au temps de Zimri-Lim," RAssyr 64 (1970) 101-103; Bass (supra n. 1) 294; Muhly (supra n. 21) 60. 219 Bass (supra n. 11) 64. 220 Bass (supra n. 1) 294; J.D. Muhly, "Sources of Tin and

at Ugarit where large quantities of Cypriot pottery and Canaanite amphoras have been found in storage,217 perhaps for export. If the lamps are Cypriot, on the other hand, then it is more likely that the pithoi were taken on at Cyprus.

The pilgrim flasks are probably of Syrian origin, but a pilgrim flask found with a lamp in another pithos complicates the matter again. Glass ingots almost certainly originated in Syria-Palestine, as did the gold and silver jewelry, and probably most of the tools and weapons. Near Eastern and Cypriot bronze tools of this general period are similar. The majority of the Ulu Burun tools, and to some extent the weapons, may therefore just as likely have originated in Cyprus. Ivory must have come from Syria, although Amarna tablets also mention its shipment from Alashia to Egypt, most probably in transshipment from Syria in this case.

The copper ingots probably originated on Cyprus as the island's abundant copper supplies have been recognized as a major source throughout antiquity. The tin may have been taken on at Syrian ports, most likely at Ugarit, which had at an earlier date received tin from Mari, some for Caphtorite merchants perhaps to be later shipped to the Aegean.218

Newly excavated half ox-hide and bun ingots verify that tin was cast and traded in the same forms as copper ingots. Most tin ingots on the Ulu Burun ship were cast originally in the ox-hide shape, and most are fragmentary. These are the earliest tin ingots known, and their shapes strengthen Bass's identifications of various gray and white ingots in Egyptian tomb paintings as tin.219 They probably represent the form in which tin was shipped from primary smelting areas near mines or from other processing centers reason- ably close to tin sources. Since the description of tin trade in ancient texts from western Asia hints at a source of tin located somewhere to the east, in Iran and even beyond in Afghanistan or Central Asia,220 these ox-hide ingots presumably represent the preferred form in which tin was transported overland by donkey caravans,221 lending weight to the supposition that the type was designed for ease of handling and transportation by animals.222 Hence, it is unlikely that tin ingots were broken down into smaller sizes for

the Beginnings of Bronze Metallurgy," AJA 89 (1985) 281. 221 For a study of the caravan trade based on the relative prices of donkeys at various Near Eastern centers during the Bronze Age, see M. Heltzer, "The Metal Trade of Ugarit and the Problem of Transportation of Commercial Goods," Iraq 39 (1977) 206-208. 222 The later transport across Gaul of knucklebone-shaped

tin ingots on pack animals is related by Diodorus Siculus

1988] 35

CEMAL PULAK

convenience of handling. It also seems unlikely that the Ulu Burun tin ingots had been cut into pieces during the voyage, as trade goods at ports along the way, for that would not explain the absence of intact tin ingots among the dozens of cut fragments on the wreck. It has been suggested that fragmentary copper ox-hide ingots, such as those at Gournia and on the Cape Gelidonya wreck, were broken into pieces at the ateliers where they were forged or cast into copper and bronze implements.223 Similarly, tin ingots may have been broken down at their point of receipt not only for being melted down, but also perhaps for use in commercial transactions.224 If this is indeed the case, then we may assume that the Ulu Burun tin ingots do not represent ingots of a shipment procured directly from a single source, but are rather a collection gathered by barter, levies or gifts.

One of the quarter ox-hide ingots of tin is apparently incised with the same mark that is seen on most of the copper bun ingots. Copper and tin are usually mined in different geographical regions, so the presence of the same mark on ingots of these different metals suggests that these marks were incised at the point of receipt and/or export to the west rather than at the production end of the metals. That the marks were more or less centered on the tin ox-hide ingot fragments may also indicate that they were incised after the ingots were broken up. Whether the tin ingots were loaded on the Syro-Palestinian coast or on Cyprus has yet to be determined. What is known, however, is that the tin was destined for a port somewhere to the west of Cyprus. It has been suggested that from the Middle Helladic period onward the Ae- gean world was using tin from northwest European sources, especially those in southwest England and Brittany.225 The presence of blue glass ingots chemi- cally identical to glass amulets from Mycenae aboard a ship laden with tin and copper ingots may indicate that the Aegean tin market was at least shared, in part, by imports from the east.

(V.22), as pointed out by Buchholz (supra n. 9) 2, and Muhly (supra n. 76) 262; Muhly (supra n. 220) 288, 289 ill. 5. 223 J.D. Muhly, "Cypriote Copper: Some Geological and Metallurgical Problems," in Acts of the International Ar- chaeological Symposium, "The Relations between Cyprus and Crete, ca. 2000-500 B.C." (Nicosia 1979) 95; Bass (supra n. 11) 71-72. 224 Although of an earlier period, i.e., mid-second millen-

nium B.C., Middle Assyrian texts show that the price of a field was frequently quoted in terms of minas and shekels of tin (Muhly [supra n. 209] 257). 225 Muhly (supra n. 220) 287.

Almost all of the items on board the ship, cargo or otherwise, could have been taken on either at a Cypri- ot or Syro-Palestinian port, with some of the goods probably in transshipment, making it impossible to determine the ship's point of departure. It is possible that port calls in both areas were made before the ship sailed westward along the Anatolian coast.

The Ulu Burun ship was undoubtedly sailing to a region west of Cyprus, but her ultimate destination can be surmised only from the distribution of objects matching the types carried on board. It has been suggested that the Dodecanese played an important role as a commercial entrepot for Aegean and Levantine goods,226 and it may not be mere coincidence that several swords similar to KW 301 come from this region, and that the Mycenaean pottery assemblage from the wreck finds some of its closest parallels on Rhodes, the largest island in the group.

Ancient texts as well as archaeological evidence reveal that Mycenaean settlements imported goods from the Near East in quantity, but it is noteworthy that Cypriot pottery is relatively scarce at Mycenaean sites.227 Since the Ulu Burun ship demonstrates that Cypriot wares of various forms did reach the Aegean, the simplest explanation of their near absence at these sites is that the cargo of pottery was not originally destined for the mainland. Could it be that after unload- ing its primary cargo of copper, tin, resin, and glass on the islands, or the mainland, or even at one of the settlements on the Anatolian coast, the voyage was to continue? The recent discovery of Cypriot and Ca- naanite pottery at Kommos nominates Crete as another possible port of call.228 The ship's primary car- goes could have been destined originally for Crete as well. We already know that at about this time Ugari- tic ships visited Crete, for we are reminded of one such venture returning home with a cargo of grain, a fer- mented beverage, and oil.229 After delivering most of her raw material in the Aegean, would the Ulu Burun ship then have sailed south to Egypt, perhaps to the

226 Sandars (supra n. 114) 128; Portugali and Knapp (supra n. 211) 52-53. 227 Bass (supra n. 1) 295. 228 J.W. Shaw, "Excavations at Kommos (Crete) during 1980," Hesperia 50 (1981) 246-47, pl. 60a, d; "Excavations at Kommos (Crete) during 1981," Hesperia 51 (1982) 170, pl. 50d; "Excavations at Kommos (Crete) during 1982- 1983," Hesperia 53 (1984) pl. 50d; "Excavations at Kom- mos (Crete) during 1984-1985," Hesperia 55 (1986) 239, pl. 58a-b. 229 G. Nougayrol, Le palais royal d'Ugarit III (Paris 1955) 107-108; also cited by Portugali and Knapp (supra n. 211) 66.

36 [AJA 92


natural harbor at Mersa Matruh, where excavations on a small island east of the town have uncovered

quantities of Cypriot, Minoan, and Mycenaean wares?230 The investigators of that site believe the island served as a victualizing station for ships sailing from Crete toward the Nile Delta and the coast of Pal- estine. Could the Cypriot pottery have been intended as trade goods for final provisioning of the ship before sailing home, or for a later venture to these regions by way of the desolate North African coast? The limited quantity of pottery so far uncovered, comprising 18 assorted vessels from the pithos, and perhaps a little more than a dozen other pieces scattered about the site, certainly does not represent a cargo of any signifi- cant commercial value, but may have been sufficient for the acquisition of the necessary supplies.

The nationality of the vessel remains uncertain. We have yet to uncover and identify the pottery used aboard the ship, but seven pieces of Mycenaean pottery raised to date, all different with the exception of two stirrup jars, suggest shipboard items. Furthermore, kylix KW 57 and cup KW 334, unlike the Cypriot open containers comprising part of the ship's ceramic cargo, must have been personal drinking cups, for the forms are hardly suitable for use as containers. Along with a few other coarse-ware bowls not yet studied, it is almost certain that these Mycenaean ceramics constituted shipboard items; were they being reused by a Near Eastern crew or did they belong to a Mycenaean crew aboard the ship? The purpose of this pottery is

230 D. White, "Excavations at Bates Island, A Late Bronze Age Egyptian Trading Station," AJA 90 (1986) 205-206. 231 Bass (supra n. 1) 296. 232 Mostly associated with women, these long pins have

uncertain, but the Mycenaean merchant's seal suggested to Bass the presence of a Mycenaean on board.231 The recently discovered globed pin (KW 570), so far unparalleled in the Near East, is of the type worn by Mycenaeans as part of their clothing.232 Cast of bronze, hence of negligible scrap value, the pin probably belonged to a Mycenaean, perhaps even the person who owned the seal. Was this person an official on a royal mission, a wealthy merchant returning from a successful trading venture, or was he only a passenger merchant of moderate means, perhaps with a small share of the cargo? The Ulu Burun excavation raises many questions which we may never be able to answer.

Whether or not the presence of a Mycenaean on the Ulu Burun ship suggests a like origin for the vessel itself remains unknown, but for the time being, available evidence may slightly favor a Mycenaean home port for the Ulu Burun ship.

It will take several years to excavate completely the site. Results of new laboratory analyses and information uncovered daily during each campaign have already reshaped some of our interpretations of the finds. The material presented above should be reviewed with caution, and the conclusions, for the most part, should be regarded only as an enthusiastic exercise.

INSTITUTE OF NAUTICAL ARCHAEOLOGY

AT TEXAS A&M UNIVERSITY

DRAWER AU

COLLEGE STATION, TEXAS 77840

often been found in pairs, but single examples as well as those associated with male burials also exist (Desborough 1972 [supra n. 176] 295).

1988] 37

uluburun pulak aja 1988

Documents