Serçe Liman1 11th Century Byzantine Shipwreck Excavation Excavation: 1977-1979 Project Directors: George F. Bass and Frederick H. van Doorninck, Jr.   Site Plans Artifact & Image Gallery Bibliography More

During the latter part of the 3rd decade of the 11th century, a Byzantine merchant ship made her way westward along the southern coast of Anatolia with cargoes she had taken on board somewhere along the coast of present-day southern Lebanon or northern Israel, then part of the Moslem caliphate of the Fatimids. Seeking shelter, she sailed before easterly winds into Serçe Liman1, a natural harbor on the southern coast of Turkey opposite Rhodes (seen right). When the ship reached a protected anchorage within the harbor close to the weather shore, the crew cast an anchor, but before a second anchor could be set, a powerful gust of contrary wind caused the vessel to swing around on her anchor cable toward the shore. The anchor held, but its shank broke under the strain. The ship was driven onto the rocks and sank, coming to rest at a depth of about 33 meters on an almost flat expanse of sand at the base of a steeply sloping rock face that descends beneath the water from the shore at the anchorage.

Shipwreck site showing the 2 meter metal grid system. (slide# GW-593) Photo: INA.

An expedition of the Institute of Nautical Archaeology and Texas A&M University excavated the Serçe Liman1 wreck during the summers of 1977-79. Popularly called the Glass Wreck, it had been chosen partly because of large quantities of broken glassware visible on the site, and also because almost nothing was known about how ships were built during the period to which the wreck belonged. Since the site was relatively flat, natural light levels sufficient to insure high-quality photographs, and most surviving hull remains were not significantly dislocated, a relatively simple mapping system in which a rigid metal grid of 2 m squares laid over the remains was combined with a daily routine of extensive photographic coverage proved adequate. Each 2-m-square area was subdivided into 16 smaller 50-cm-square areas for the purpose of recording where artifacts were found.

Simplified lines of the Serçe Liman1 ship. Photo: INA.

The ship had been a small two-masted vessel with lateen sails. The mainmast was stepped slightly aft of amidships, and the foremast, with a somewhat smaller sail, had probably raked forward over the bow. The ship had an overall length of perhaps only 50 Byzantine feet (15 m) and a breadth of 17 Byzantine feet ( 5.3 m). Despite these modest dimensions, her cargo capacity was some 30 metric tons, this due to her very full, box-like hold, with a virtually flat bottom amidships and an almost angular transition from the bottom to steep, flat sides. With her flat bottom and a keel that scarcely projected below the planking, the ship would have been a poor sailor when not well laden but had the advantage of being able to navigate in relatively shallow waters. She was probably similar to the qarib, a common ship type in Egypt at that time that sailed down the Nile from Cairo and then as far west as Tunisia and Sicily.

Serçe Liman1 exhibit in the Bodrum Museum of Underwater Archaeology. (slide # GW-3344) Photo: INA.

The surviving hull timbers were preserved with polyethylene glycol and reassembled at the Bodrum Museum of Underwater Archaeology in a building designed and built by the Turkish government for the display of the ship and its contents. Although only about 20 percent of the hull has survived, it is an important archaeological document that reveals to us at least one of the solutions that Mediterranean shipwrights had reached by the end of the first millennium to the problem of how to build a hull whose overall shape was determined primarily by frames erected before planking, rather than by planking erected before frames, the latter method being followed in the ancient Mediterranean world.

The ship was carrying nine iron anchors when she sailed into Serçe Liman1. Five anchors were stacked on the deck between the two masts with their shanks laying parallel to the hull's longitudinal axis. The other four were bower anchors ready for use with wooden stocks mounted in place. Two of them were on the bulwarks close by the anchors on the deck, one to either side. The other two bowers hung off either side of the ship's bow.

Forward starboard bower anchor following the removal of concretion. (slide# GW-2181) Photo: INA.

The anchors have straight arms that are inclined downward so as to form with the shank an inverted Y, a design that made it possible to reduce shank length and consequently the incidence of shank breakage. Y-shaped anchors were used in the Black Sea and in much, if not all, of the Mediterranean from the 10th to the 13th century. The arms of eight Glass Wreck anchors are set at an angle of about 110N with respect to the shank. The other anchor, among those stacked on the deck, had shorter arms set at an angle of only 98N.

The anchors fall into three weight-sizes: 150 Byzantine pounds (48 kg), 175 pounds (56 kg), and 200 pounds (64 kg). All of the shanks that had not been broken and repaired have an identical length of 6 Byzantine imperial spans (1.40 m). Anchor arms normally had a length of 2 Byzantine feet, and all stock apertures had an identical diameter of 3 Byzantine fingers. This dimensional standardization (not present in the anchors from the earlier 7th-century Byzantine ship at Yassiada) and the relatively small interval between weight sizes suggests that by the 11th century, the Byzantine state was controlling the weights and numbers of anchors on ships about as closely as we know was done at 13th-century Venice, where the number and weight of anchors used on ships of any particular size was regulated by law.

Personal gear carried onboard the ship included this wooden comb and scissors. (slide# GW-2720) Photo: INA.

Living areas were located within the bow and stern compartments, and apparently also in the midships area. A small cargo that could have fit into a single package and some personal possessions were found in the bow compartment and may have belonged to just one person. The cargo included over a dozen items of glassware and several glazed bowls; personal items included a toilet kit containing a comb, razor, scissors and some coins; a couple of weights for weighing coins; two cooking pots, one still containing goat, or possibly sheep, bones; a well-used glazed bowl; and perhaps one or more of three wine amphoras also found here.

Most food stores were kept either in the midships area or in the stern compartment. The shipboard diet included meat (pig, goat, and possibly sheep), fish (tunny, tub gurnard, bass and drum), almonds, assorted fruits, and olives; there may have been live goats on the ship. There were three large nets with floats, a casting net, and a multi-tined spear for catching fish. The recovery of charcoal only from the stern area suggests that most, if not all, shipboard cooking was done here. The consumption of pork was restricted to those living in the stern, and possibly the bow, compartment; perhaps this was true of fruit as well. Another possible indication of social stratification on the ship is the fact that a chess set had been kept in the stern compartment while a backgammon piece was recovered from the midships area.

Fatimid Islamic glass coin weights. (slide# GW-2536) Photo: INA.

Weighing equipment, kept almost entirely in the stern and well suited for a Byzantine-Fatimid trading venture, consisted of a Byzantine steelyard, 3 balances, two large sets of balance-pan weights, one Byzantine and the other Fatimid, and glass weights for weighing Fatimid gold and silver coins. Some of the glass weights bear legible dates, the latest being either 1024/25 or possibly 1021/22. The voyage must have occurred not long after the latest weights were issued, during a time of improving Fatimid-Byzantine relations, affirmed by a peace treaty in 1027.

Filigree gold earring. (slide# GW-2417) Photo: INA.

In keeping with what was then the normal practice, little coinage appears to have been present on the ship. Only three gold coins, all quarter dinars of the Fatimid caliph al-Hakim (996-1020), 15 clippings from other Fatimid gold coins , and some 40 Byzantine copper coins, later 'anonymous' issues of Basil II (976-1025), were recovered from the site. The clippings were used in place of silver coins, which were then in short supply and therefore reserved for local commerce. A small amount of jewelry, usually found in close association with coins and including a half-dozen silver rings and a gold earring, may have served as bullion, which merchants generally preferred to carry in the form of jewelry. The paucity of coins and bullion coupled with the presence of three Byzantine lead seals for documents suggests that merchants on board may have used letters of credit as a way of minimizing potential losses to pirates.

Sword with decorated bronze hilt. (slide# GW-2222) Photo: INA.

Piracy was an ever present danger, and the ship was well supplied with defensive arms. Iron spearheads belonging to 11 thrusting spears and 52 javelins, along with the remains of perhaps three swords, were recovered. One of the swords had a bronze hilt decorated on either side by an elaborately-feathered bird, the hamsa, a motif of Indian origin then popular in the Black Sea region.

Several axes from the shipwreck's collection of repair tools. (slide# GW-2255) Photo: INA.

Ship stores kept in the stern compartment included spare parts and tools. The former included pulley parts, other wooden rigging elements, and lead fish-net weights. For foraging and camping on land, there were two axes, a billhook, a mattock and a pick; perhaps the ship had a shovel with an iron-sheathed blade that did not survive. Close parallels for some of these foraging tools have been found in Bulgaria. Among the carpentry tools recovered is everything needed to keep the ship in good repair (adzes, a bow drill and bits, chisels, claws for extracting nails, a hammer, a mallet, a plumb bob, a rasp, a saw, and assorted tacks and nails), except for a file for the saw. Of particular interest is a set of caulking tools apparently in use at the time of the sinking; perhaps the ship had sought shelter due to a leaking hull. The fact that the ship's carpentry tools and weighing equipment had been kept together in the same basket suggests the distinct possibility that the captain had been both a merchant and the ship's carpenter.

The ship's cargoes were diverse and in some cases quite small. This was a time of relatively free trade in the Mediterranean, and wholesale merchants, faced with ever-changing commodity prices, minimized risk by dealing in a diversity of goods. The cargoes included some three tons of glass cullet, some 80 or more items of glassware, several dozen cooking pots, several dozen glazed bowls, several half-dozen lots of jugs and gargoulettes (one-handled jugs with a built-in filter), raisins, sumac, wine, and a perishable cargo occupying the forward half of the hold that could not be identified.

Broken glassware constituted a substantial portion of the ship's cargo and would have been carried in wicker baskets similar to the one seen here. (slide# GW-1393) Photo: INA.

The glass cullet consisted of two tons of raw glass broken up into manageable pieces and one ton of broken glassware and glass-making waste from some Islamic glass factory on the Fatimid Syrian coast. Almost certainly transported in wickerwork baskets, it was being carried in place of stone ballast in the after quarter of the hold. Shipping glass cullet cheaply as ballast often made economic and/or technical good sense, since melting glass requires a much lower temperature than making glass does and a desired kind of glass can not always be made from locally available raw materials.

Some of the many types of glassware found on the Serce Liman1 shipwreck. (slide# GW-1183) Photo: INA.

On-going efforts to sort and mend between half a million and a million shards of broken glassware recovered from the wreck belonging to between 10,000 and 20,000 vessels has produced more than 200 distinctly different shapes, some unique, including various types of beakers, cups, bowls, bottles, jars, ewers, jugs, plates and lamps. These vessels illustrate for the first time the general nature of medieval Islamic glassware made in Syria and reveal a regional style characterized by an extensive use of simple, patterned molds to decorate vessels in a highly cost-effective manner. For example, vessels begun in the same patterned mold were sometimes then further blown into a variety of quite different shapes.

A sample of the different kinds of ceramic vessels carried on the Serce Liman1 ship. (slide# GW-2150) Photo: INA.

The ceramic cargoes were fragile wares packed in small groups, probably in wicker baskets or crates. A half dozen thin-walled gargoulettes of white fabric were found in the stern. They have filters with delicate fretwork and pseudo-Kufic designs and may be Syrian copies of similar Egyptian gargoulettes. Some three dozen thin-walled cooking pots, several baking dishes, a half dozen two-handled jugs and a half-dozen gargoulettes, all made of similar purplish-red fabrics, were stowed just aft of midships in the hold. The cooking pots and baking dishes have a dark, lead glaze on their interior surface and are examples of a type of ware widely used in Syria and Palestine during the period of the Crusades.

Islamic glazed Sgraffito ware and splash ware. (slide# GW-2041) Photo: INA.

Almost four dozen Islamic glazed bowls, which find their closest archaeological parallels at Caesarea on the northern coast of Israel, had been stowed, again in small groups, in both compartments and in the hold just aft of midships; only a few were in use. Most of the bowls were examples of either splash ware, so called because various colors of glaze were poured or splashed on the bowl's interior in such a way as to form a decorative pattern, or sgraffito ware, so called because decorative designs were carved into the clay. In this instance. the designs are rather deeply and broadly carved in what is known as the champlevé fashion. Although conventional chronology had assigned a 9th-century date to splash ware and a 12th-century date to the beginning of the champlevé style, the two techniques appear together on some of the Glass Wreck bowls, a fact that has lead to a reassessment of the chronology of Islamic glazed wares.

Three of the globular amphoras. (slide# GW-2085) Photo: INA.

The ship was carrying at least 104 Byzantine amphoras, 3 in the bow compartment, and the rest about equally divided between the hold just aft of midships and the stern compartment. The presence of so many Byzantine amphoras initially left us perplexed, since the ship was carrying cargoes from Fatimid Syria. We now realize that the owners of the amphoras had used the jars over and over again as transport containers, selling their contents but always keeping the amphoras for further use. Over time, rims underwent significant damage from the recurring act of prying out stoppers, and handles were also occasionally broken off. Three amphoras on the ship had lost both their handles and their entire neck as well. In an effort to reduce the chances of further damage, the broken edges of rims or necks and the stumps of broken handles were carefully carved down and rounded off.

Two of the amphoras had been made to carry olive oil; the rest, to carry wine. The great majority of the latter were undoubtedly serving as cargo amphoras, very probably carrying wine, on this their last voyage, but the contents of some of the amphoras were probably for shipboard consumption, while the three amphoras without handles or neck may have been being put to some special use within the stern compartment. A study of the dimensions and capacities of the amphoras has revealed a high level of dimensional standardization and a well-developed capacity system that together imply a considerable degree of central imperial control.

There are potter's marks and marks of over a half-dozen different owners carved on most of the Byzantine amphoras. The potter's marks as a group have their best parallels among contemporaneous Bulgarian potter's marks. Many of the piriform amphoras are in design and fabric like piriform jars made in kilns recently discovered on the north coast of the Sea of Marmara not far west of Constantinople. We are presently inclined to think that the ship and many, if not all, of the merchants and crew on board were from a community in that region that included Hellenized Bulgarians in its population. The ship's most likely destination would appear to have been Constantinople, an important production center for Byzantine glass.

(text adapted from van Doorninck, Jr., F.H., "Glasvraget-et byzantinsk skib fra 1000-tallet," Hvad Middelhavet gemmer [Århus 1997] 121-136.)

Bibliography

Bass, George F. "The Shipwreck at Serçe Liman1, Turkey." Archaeology 32 (1979): 36-43.

Bass, George F. "Archaeologists in Wet Suits." In Science Year 1982: The World Book Science Annual, pp. 96- 111. Chicago, 1981.

Bass, George F. "The Million Piece Glass Puzzle." Archaeology 37 (1984) :42-47

Bass, George F. "The Nature of the Serçe Liman1 Glass." Journal of Glass Studies 26 (1984): 64-69.

Bass, George F., J. R. Steffy and F. H. van Doorninck,Jr. "Excavation of an 1lth-Century Shipwreck at Serçe Limani, Turkey." National Geographic Society Research Reports 17 (1984): 161-182.

Jenkins, Marilyn. "Early Medieval Islamic Pottery: The Eleventh Century Reconsidered." Muqarnas 9 (1992): 56-66. Describes how the excavation revolutionized accepted views of early medieval Islamic art.

Pulak, Cemal, and Rhys F. Townsend. "The Hellenistic Shipwreck at Serçe Liman1 Turkey: Preliminary Report." With "Appendix. The Transport Amphoras: Description and Capacities," by Carolyn G. Koehler and Malcolm B. Wallace. American Journal of Archaeology 91 (1987): 31-57.

Steffy, J. Richard. "The Reconstruction of the 11th century Serçe Liman Vessel. A Preliminary Report." International Journal of Nautical Archaeology 11 (1982): 13-34.

van Doorninck, Jr., F.H., "Glasvraget-et byzantinsk skib fra 1000-tallet," Hvad Middelhavet gemmer (Arhus 1997) 121-136. The most recent analysis of the 11th-century shipwreck.