A chance discovery of a 2,400-year-old shipwreck off the coast of Kibbutz Ma’agan Micha’el, 20 miles south of Haifa, has been yielding a storehouse of new insights into ancient seafaring and the shipwright’s art. Thanks to recent geomorphological changes—fluctuations in the sea level, sediment movements and shifts in local faults—along Israel’s seacoast, investigators are discovering more and more submerged antiquities.
We may have discovered, at last, a ship of Phoenician origin sunk along Israel’s Mediterranean coast almost 2,400 years ago. But the fact that after three seasons of excavation followed by intensive studies we still can’t be sure of the identity of the ship’s builders and sailors, illustrates just how difficult it is to interpret the evidence from this submerged vessel. What is certain, however, is that this is one of the best-preserved wooden-hulled ships ever discovered in the Mediterranean from such an early time.
The discovery was made by chance, along the shore of Kibbutz Ma’agan Micha’el, 20 miles south of Haifa, on the very spot where, more than 30 years earlier, marine archaeology in Israel had its beginning. In the fall of 1985, while diving along the coast in water less than 6 feet deep, a member of the kibbutz, Ami Eshel, noticed a pile of large stones not native to the Levantine coast. Could they possibly be the ballast from a sunken ship? Near the stones he saw some pieces of wood and broken pottery sherds. This was enough for him to alert representatives of the Israel Antiquities Authority and the Center for Maritime Studies at the University of Haifa, who confirmed that there was, indeed, an ancient shipwreck there worthy of further exploration.
Preliminary investigations were carried out soon after the discovery by Shelley Wachsmanna and Kurt Raveh of the Antiquities Authority. Pottery sherds and a broken ceramic oil lamp found on the site enabled them to assign a probable date of the fifth century B.C. My colleague Avner Raban of the Center for Maritime Studies and I explored the site in the spring of 1986. We were able to observe a great deal of wood preserved beneath a thick layer of ballast. To our amazement we saw hull planks in an excellent state of preservation thanks to their having been encased in an anaerobic environment; they were connected with pegged mortise-and-tenon joints, and the frames were secured to the hull with clenched iron nails.
The ship was obviously well worthy of excavation. To further determine the extent of the remains, Raban and I conducted another set of test soundings in the fall of 1987. We were able to define the northern boundaries of the ship and to gather further details of its size and condition. The University of Haifa then applied for and received a permit to excavate the site. To serve as field director of the excavation, we invited Jay Rosloff, a hull specialist who had studied under Professor J. Richard Steffy at the Institute of Nautical Archaeology, Texas A&M University.
In the spring of 1988, Rosloff carried out additional test soundings by cutting a narrow trench across the width of the vessel, roughly along the inshore end of the ballast pile. For recording purposes, datum poles, which define the plan of the site, were inserted in the sand by means of a water jet; they remained there throughout the excavation. By the same method, a “probe pipe” was used to define the ballast area and to determine the depth of sand over the site. The probe pipe disclosed a clay layer approximately 12 feet below sea level; this gave us the measure of the sand accumulation over the site.
Other test trenches helped to locate the vessel’s endposts, nearly 30 feet apart. One offshore trench produced an especially wide variety of finds, including a cooking pot whose outer surface was still coated with enough soot to stain our hands, a large plate known as a mortier, pottery sherds and a lead ingot. These finds suggested that the offshore end of the ship was the stern where the galley would normally be located. A thick layer of dunnage (loose material that served as padding) had been used to buffer the ship’s hull from possible damage by the heavy ballast stones. The dunnage consisted of cut twigs and branches of Mediterranean pine and pistachio, very well preserved.
Because the ship lay in shallow water, we had to deal with the problem of surge. This made working conditions difficult; sand continually drifted back into the excavation trench. The surge also affected the clarity of the water, making it murky and difficult to see through. To ameliorate these problems, we dredged a horseshoe-shaped trench around the site below the estimated depth of the ship’s keel, with the open end of the horseshoe facing shore. We then lined the outside of this trench with 1,350 sandbags to prevent more sand from migrating into the excavation area. Unfortunately, this man-made sandbar did little to eliminate the surge effect along the seafloor. Even waves as small as 1 to 1.5 feet made it impossible to work safely and avoid uncontrolled movements endangering the wood. In three seasons of work on the site (a total of 160 days), sea conditions allowed us to spend only 32 days actually excavating the ship and its contents. Because the site is small (only 270 square feet [75 square meters]), no more than three pairs of divers could work underwater at any one time, usually in shifts of about one-and-a-half hours.
Our first task was to remove most of the ship’s ballast. The stones, some of them weighing more than 100 pounds, were placed in boxes and pulled to shore on a sled tied to a van. The ballast was examined microscopically and geochemically. It consisted of three lithological groups- metamorphic, magmatic and sedimentary rock. The first group, comprising 65 percent of the total, was mostly blue schist. Initially, we thought the collective evidence of the rock types, when compared to the geology of the Mediterranean region, pointed to an origin somewhere on the Tyrrhenian seacoast, around Corsica or Calabria in southern Italy. After a thorough investigation by the expedition’s geologist, Dr. Arye Shimron, however, including visits to possible sites of origin, we now believe a more likely location is farther east, around the Greek islands or Crete or possibly Cyprus.
One by one, we excavated the ship’s 14 rooms (in nautical terminology a room is the space between one floor timber and the next). We started at the stern and worked toward the bow near the shore. Fearing damage by storm waves to exposed portions of the site, we initially adopted the strategy of excavating only a few feet of the hull at a time—then mapping, photographing and removing the contents.
The planking of the hull presented its own difficulties. To our surprise, we found that although the hull planks were soft, they were tough enough to resist breaking into short lengths. Some of the planks were several yards long and first had to be separated from adjacent planks by slicing the tenons with a knife. Then, by carefully breaking, and sometimes sawing them, we could place the planks in trays and transfer them to the freshwater tanks at the expedition base. This was the first phase of the conservation process.
We had no alternative to the use of such unorthodox methods, since all timbers had to be removed in sections that would fit our immediate transportation and conservation facilities. Leaving the wood under water for a prolonged period after exposure might have caused serious damage.
The preserved portion of the hull measures 37 feet (11.25 m) long and 13.1 feet (4 m) wide. Its displacement is estimated at 20 tons, over 12 tons of which was ballast. The hull was shaped like a wine glass. The slightly curved planks were attached to one another by a series of pegged mortise-and-tenon joints. Other structural components included full frames, a longitudinal stringer and part of a wale, vertical stanchions, floors, futtocks and a solidly built mast step. To our surprise, we found that the planks were not only joined by mortise-and-tenon joints, but the ends of the planks were lashed by cords to the endposts, knees and even an inch or so of the keel through triangular holes. The keel consisted of a single timber nearly 25 feet (8.5 m) long, 4.5 inches (11 cm) wide and 6.25 inches (16 cm) high, with a false keel, or shoe. All the structural components of the hull were of pine except the tenons and the false keel, which were of oak. The wood was in such an excellent state of preservation that it still retained the odor of resin. There was no sign of teredo (shipworm infestation) nor wear from recurrent sailing. Thus the ship may well have been on its maiden voyage when it sank!
A complete, one-armed anchor made of oak, with a lead-filled stock, was discovered off the ship’s starboard bow.1 The shank and arm were carved from a single timber and the arm’s tooth was copper coated. Remains of rope were observed around the crown and the lifting loop. The way the anchor was positioned makes us think that it was probably ready for use but had not yet been cast. This anchor is a unique find; it will be very important in the study of the evolution of anchors. This is the first time a complete, one-armed ancient wooden anchor has been discovered, although the existence of such anchors in antiquity had been known.
Three beautiful wooden artifacts were found. The first was an olive wood box in the shape of a leaf or heart with a swivel top. We assume it was used for cosmetics or jewelry. The other two were violin-shaped boxes; they too may have held cosmetics or jewelry.
Especially important to our understanding of naval architecture and the shipwright’s art is a collection of woodworking tools found in and around a basket made of plant fiber. With them were a large number of treenails, tenons and a wooden carpenter’s square. Only the wooden handles survived of some carving, cutting and boring tools. We also found a whetstone, which was probably used to sharpen the tools. Other wooden implements included toggles that were probably used to tie down the sail; in the bilge there were fresh chips of carpenter’s waste and some unused wooden stakes.
Pottery remains consisted of 70 items, many of them complete. They included utensils for daily use by the sailors such as jugs, mortier plates, lamps, a cooking pot and a water jar (pithos); storage jars (a minimum of nine of the “basket handle” type) and some decorated amphorae; miniature juglets and black glazed vessels, including two salt cellars (which probably belonged to individual sailors, a kind of personal possession). Many of the ceramic vessels seem to have originated on Cyprus or the Palestinian coastline and date to the fifth to fourth centuries B.C. Some of the pottery has parallels in the Greek ware of the same general period. We are currently pursuing neutron activation analysis of the pottery, which we hope will give us a more conclusive indication of its origins.b Organic materials included a small woven basket, remains of a mat and a great amount of rope varying in width between .08 and 1.6 inches (2–40 mm). The rope consists of three-strands with a left-hand twist.
Remnants of food included grape, fig and olive seeds, as well as barley. A preliminary study of the pollen samples shows them to be of an eastern Mediterranean coastal variety that tends to grow during the summer. This indicates that the ship was sailing in the usual sailing season—from spring to late autumn. From the few food remnants that we were able to collect during the excavation, we are still trying to reach some conclusions about what the sailors ate.
The metal finds included what seems to be a small copper incense shovel, iron and copper nails and lead fishing weights. Surprisingly, iron nails were used to fasten the ship’s frames to the hull; we would have expected copper nails in ship construction at this early period. We also found one lead ingot. There could have been others that we did not find, or this single specimen could have served as raw material for use on board.
Because of geomorphological changes along the sea coast—minor sea-level fluctuations, sediment movement and changes in local faults—more and more antiquities buried in shallow waters are being discovered along Israel’s coastline. As a result, our coastline is providing a rapidly growing source of information about ancient maritime activities. The chance find of a late fifth- or early-fourth-century B.C. ship off the coast of Ma’agan Micha’el is an important link in the chain of development of ancient shipbuilding. Until now, there has been a gap between the fourth-century B.C. ship discovered near Kyrenia, Cyprus,2 and the more recent find of a sixth-century B.C. ship found off the coast of Gela, Sicily.
The Kyrenia ship was especially important to us in our comparative studies. That ship was completely excavated and, after conservation, reconstructed; it is now on permanent display at the Kyrenia Castle. An exact replica of it has been built and is today sailing the high seas.
There are many important similarities between the Kyrenia ship and ours. The Ma’agan Micha’el ship met its fate in about 400 B.C.; the Kyrenia ship went down about 75 years later. Ours was just over 44 feet (13.4 m) long in all, while the other measured 49 feet (14.9 m). Both were built by the shell-first construction method, in which the planks of the hull are joined first and only later are the frames added. The average plank size on both ships is almost identical- 19 to 23 feet (5.8 to 7 m) long, 8 to 10 inches (20 to 25 cm) wide and 1.5 to 2 inches (3.8 to 5.1 cm) thick. The distance between tenons on both ships is 4.75 inches (12.1 cm), with the width of the tenons being 1.5 to 2 inches (3.8 to 5.1 cm); the mortise depth is 3.1 inches (7.9 cm). Both ships also had a one-piece keel about 30 feet (9.1 m) long, and the mast-step structure, which holds the mast in place, is virtually the same on the two vessels.
There were nevertheless many important differences between the two ships. Ours was made of 14 full frames (often called ribs by nonprofessionals), while the Kyrenia vessel was composed of 41 half-frames. The average distance between frames was 29.5 inches (74.9 cm) on our ship and 9.8 inches (24.9 cm) on the Kyrenia ship. Ours were fastened to the planks by iron nails, in contrast to the copper nails of the Kyrenia vessel.
Our ship had no sheathing; the Kyrenia ship’s entire bottom was lead sheathed. Lashing was incorporated into our ship’s bow and stern structures; no trace of lashing was found on the other ship. Lastly, one more of the many differences between the ships should be mentioned here- the Ma’agan Micha’el ship’s ballast represented two-thirds of its displacement; in the Kyrenia vessel, part of the cargo, such as millstones, served as ballast.
Our ship plied the waters at a time of peak Phoenician maritime activity and trade in the Mediterranean. Could our ship be Phoenician? The mortise-and-tenon joints, along with the lashing, might provide a clue. The composition of the ballast is another key; if it was not picked up along the way in a port where another ship disposed of it to take on a new load of cargo, then we may have a hint as to where the ship’s voyage began.
The pottery remains, certainly one of the main clues by which we attempt to date our ship and follow its sailing route, presently point us to Cyprus and Greece. Yet so many questions remain. Why did the ship land as it did, perpendicular to the coastline? It was abandoned, but was still accessible so close to shore. The ship carried a unique anchor; where and by whom was it produced? There was very little cargo. Could the ship itself have been the merchandise, perhaps the special order of a local trader? These are just some of the subjects the expedition team is pursuing.
The Ma’agan Micha’el ship itself is now completely excavated. We are currently engaged in the second phase of this elaborate research project—the conservation of the hull and of the ship’s contents. The first stage of this long and tedious process occurred during the three seasons of excavation. Once retrieved from the sea, every single piece of hull timber—loose or intact, all artifacts of wood, metal or other perishable material, as well as the ceramic ware—was immediately placed in fresh-water tanks. The timber was initially put in plastic trays, stacked one over the other to avoid friction. Later, with every wooden item recorded and marked with identity numbers, the trays were transferred to the permanent conservation tanks at the university laboratories in Haifa. For the next two years at least, the timber will go through a process of impregnation by heated polyethylene glycol (PEG). This waxlike chemical will be added to the tanks in slowly increasing concentrations. Simultaneously, the temperature of the solution in the tanks will be gradually raised. In this way the PEG will penetrate the cellular cavities of the deteriorated wood, replacing the water in the cells with PEG. Only then, after the wood has been restructured and its sturdiness recovered, will the reconstruction of the ship begin. A 1-to-10 scale model of the ship’s hull has recently been built, accurately following the components of the hull’s timbers as disclosed underwater in their original position. This model will provide an indispensable guide when we reconstruct the ship itself.
The excavation was carried out under the auspices of the Center for Maritime Studies, University of Haifa, Israel, and was made possible through the generosity of Sir Anthony Jacobs of London. Kibbutz Ma’agan Micha’el provided full board for the permanent staff during the excavation periods.
Permanent staff- Elisha Linder, project director and senior investigator; Jay Rosloff, field director and hull specialist; Mike Udell, assistant field director; Jerry Lyon, senior assistant; Eve Black, academic secretary; Ya’acov Kahanov, quartermaster; Yossi Tur-Caspa, geotechnician; Steve Breitstein, diving and operations officer; Danny Syon, photographer; Neria Piercy, technical illustrator and artist; Yitzhak Dagan, administrator; Shelley Waschsmann, representative of the Israel Antiquities Authority; Itamar Grinberg, underwarer video photographer; Arik Bernstein, filmmaker. Specialists- Dr. Michal Artzy, pottery; engineer Shlomo Eisenberg, metals; Dr. Mina Weinstein-Evron, pollen; Dr. Mordechai Kislev, botany; Dr. Yossi Mart, geomorphology; Dr. Avner Raban, ship archaeology; Dr. Arye Shimron, geology; Dr. Ella Werker, timber analysis; Dr. Carmela Shimoni, fiber analysis; Ms. Orna Cohen and Ms. Marina Rosovsky, conservators. Volunteers- USA- Judy Scheuer, Michael Halpern, Sam Turner; New Zealand- Anna Nichols and Chris Campbell; Austria- Anthony Abry; Switzerland- Rachel Crausez; England- Lucy Blue; Tel Aviv- Tami Shabi, Na’ama Bahat; Haifa- Sigal Namer, David Zell, Amos Lavee; Kibbutz Ma’agan Micha’el- Avraham Chasidim, Oren Linder, Modi Bracha, Shimeon Gil.
b. Maureen F. Kaplan, “Using Neutron Activation Analysis to Establish the Provenance of Pottery,” BAR 02-01.
1. See Jay Rosloff, “A One-Armed Anchor of c. 400 B.C. from the Ma’agan Michael Vessel, Israel. A Preliminary Report,” International Journal of Nautical Archaeology, vol. 20, no. 3 (1991), pp. 223–226.
2. See J. Richard Steffy, “The Kyrenia Ship- An Interim Report on its Hull Construction,” American Journal of Archaeology, vol. 89, no. 1 (1985), pp. 71–101.