The Bodrum Museum of Underwater Archaeology, located in the 15th-century Castle of St. Peter, will open a new exhibition hall in the autumn of 1995. Prepared in collaboration with the Institute of Nautical Archaeology, the new display will feature artifacts from the seventh-century Byzantine trading vessel excavated at Yassiada under the direction of George Bass. The exhibit will be housed in the castle's chapel, originally built by the Knights of St. John in the first half of the 15th century A.D. and restored by Spanish knights in 1519-20.
During the summer months, up to 1,800 visitors tour the Bodrum Museum every day. Tourists come from all walks of life, from archaeologists to fishermen, engineers to shepherds. The Yassiada exhibit has been designed to give this diverse audience a clear understanding of the meaning of the archaeological remains. To this end, it will portray the vessel anchored in a harbor, with the bottom part of the ship shown under water. Visitors should receive the impression that the ship did not wreck on the reef at Yassiada, but continued on to make port in Halikarnassos (the ancient name of Bodrum). Sound and light effects will be used to transport visitors into the atmosphere of a seventh-century harbor.
Immediately upon entering the hall, visitors will encounter a full-size reconstruction of the vessel's stern, designed by INA ship reconstruction specialist Fred Hocker and built by a team of Nautical Archaeology Program and Turkish students. To the left of the replica, a 1:10 scale model will allow visitors to see how the ship would have appeared during its working life. Cabinets along the wall will hold finds from the ship's galley and storage lockers, including the large steelyard inscribed "Georgios Presbyteros Naukleros," indicating the name of the vessel's captain. The artifacts in the showcases will be exhibited as goods are presented for sale in harbor shops.
Next, guests will climb a staircase to the ship's deck, from which they can view the harbor scene. A museum staff member dressed as Captain Georgios will meet the visitors in the galley and provide
information about shipboard life. He will also demonstrate the uses of various galley objects, such as drawing wine from an amphora with a wine thief. Visitors can then step down to the harbor and walk across the "sea" to a special glass platform. Under the platform, a sea-bottom scene with an arrangement of amphoras will recreate the wreck site. Showcases will display
additional artifacts from the vessel, including several large anchors. The whole experience is designed to appeal to the five senses and correspond to our concept of a living museum.
- Oguz Alpozen, Director, Bodrum Museum of Underwater Archaeology
The crowd that gathered along the edge of the excavation pit quieted suddenly as the western-most cover slab was removed from the tomb. Diffuse light coming in the chapel door filtered into the pit below and revealed five skulls stacked in a row against the end of a rectangular chamber. Three had collapsed, one lay on its left side, and one, robust with heavy brow ridges, stared straight up out of the pit. A slightly sour smell drifted up from the dark earth beneath the top layer of bones. As we took pictures and began the delicate cleaning process, we wondered how a group of nautical archaeologists had ended up excavating INA's first conventional land site.
Four of us (Nautical Archaeology Program students Taras Pevny and Tommi Makela along with the authors) came to Bodrum late in May to begin construction of the Yassiada replica in the Knights' Chapel. When we arrived, the Museum had already purchased the wood for the replica, many cubic meters of pine and elm, some of which had been delivered and stacked in the chapel. The rest arrived later in the summer, several tons at a time, and each piece had to be carried up the medieval stairways and narrow passages of the outer baileys to reach the courtyard. With the help of several Kurdish workmen, we carried pieces of wood weighing hundreds of pounds each up to the chapel at a run (it makes the unevenness of the stairs easier to take), much to the amusement of tourists using the same stairways. After resurrecting the same 40-year-old German bandsaw that had been used to build the Serge Limani sectional replica in 1986-87 (see INA Newsletter 15.3: 14-21), we began to cut large balks of elm and pine into frames and a sternpost. By this time, we had been joined by Cagdas Oralkan, a Turkish student from Bilkent University in Ankara, and Ozlem Buyuran, a student from Bosporus University.
The sternpost, which consists of two large, curved pine pieces fastened together in what is often called a "thunderbolt of Jupiter" scarf, plus an inner post and elm false post, took over a week to fashion. Too large to fit easily through the bandsaw, it required several days of steady work with handsaw and adzes to shape (fig.2). The scarf itself, with faces that had to be fit together in several different planes, took nearly a day to finish. An ancient shipwright would no doubt have laughed at our slow progress.
The working plan was to rough out all of the primary internal timbers before beginning any assembly. Taras had carefully measured frame shapes off of the detailed plans and transferred these to balks of timber for sawing, so we were relatively confident that there would be a minimum of fairing after setting things up. Guiding four-meter frames through the bandsaw sometimes took four people and good teamwork, and often involved shouting in two languages (three, if Tommi decided to curse in his native Finnish). By the end of June, all of the floor timbers and most of the futtocks had been sawn out as well as the keel and sternpost, and it was time to prepare the chapel for the assembly of the ship.
Because the ship had to look as if it were floating in the floor, we needed a roughly trapezoidal hole about three meters long, five meters wide, and one and a half meters deep. We were a bit reluctant to cut into the floor of a 500- year-old building, but soon realized that the floor was concrete, poured during a restoration in 1964. Underneath 10 cm of concrete was a layer of rubble, also from the 1964 restoration. Beneath that lay plain brown dirt.
We had not dug far into the dirt when we discovered the first archaeological remains. The tops of two stone piers appeared near the east end of the pit, and the dirt was full of potsherds and broken glass. The potsherds dated from the 20th century back to Hellenistic times (c.330 to 31 B.C.), and were typical of fill dirt from anywhere in Bodrum. As we dug further, more walls came to light, and we decided that the removal of the fill needed to proceed more archaeologically. We had been concerned before digging that we would encounter remains of some sort in such a prominent location, and had actually discussed the possibility of finding graves associated with the chapel, but we were unprepared for the richness of the finds or the complexity of structures in the small hole we had opened. Danish excavations in the Museum had revealed remains of earlier structures, from Hellenistic walls to Byzantine mosaics, so our discoveries should not have presented such a surprise.
The building in which we were digging, the Chapel, was built when the Knights Hospitalers of St. John of Jerusalem (a chivalrous order of European knights founded during the Crusades) began their work on the Bodrum castle early in the 15th century. In the early 16th century, the chapel was rebuilt or improved by the Spanish knights of the order, but it is uncertain how much of the 15th-century structure was retained. Damage from French shelling during World War I has further confused the issue, and little remains now of the original floor. If the current chapel stands on the same foundations as She 15th-century building, then nothing we found in the pit can be conclusively associated with the period of occupation by the Knights, as none of the walls or pier foundations we discovered is exactly aligned with the current walls. The Zephyrion promontory was occupied by Selcuk Turks before the Knights arrived (much of the inner defense works of the castle dates from the medieval Turkish occupation), so some of the remains beneath the floor may date from that period, but the later Turkish occupants of the castle left the chapel standing and converted it to a mosque.
Many of the walls we found in the pit were built of mortared rubble, often of reused stone. Such masonry is a common feature of later Roman and Byzantine architecture in Asia Minor, and also can be seen in Turkish building. Many of the fragments incorporated into the standing walls of the castle, as well as those in the excavation, date to the Hellenistic period, and many of the later remains are built around large blocks of Hellenistic masonry. Although nothing conclusively dates these walls, the lowest layers of old fill around them contained Byzantine pottery. The mortared rubble walls seem to be approximately contemporaneous with each other, and may all belong to the same, larger structure, although this structure cannot be identified from our small excavation.
Two main foundation walls were exposed. Wall A, oriented north-south, is abutted by wall B in the southeast corner of the excavation. Wall B. while of the same general type of construction, is faced with squared stones and extends down to bedrock, 1.6 m below the current chapel floor. In places it stands 1.3 m high and is built over a Hellenistic wall foundation stone. Included in this wall is a faced ashlar stone that forms the west end of the wall and one edge of a drain floored with reused roof tiles.
Along the north side of the excavation pit, three tombs were uncovered (figs. 3, 4). The first of these, tomb 1, was discovered when a small hole was broken through its south side early in the excavation. Clearing the edges of the hole, originally mistaken for a cistern, revealed bones. We entertained visions of armor clad skeletons clutching rusty longswords, although we knew that medieval knights were not usually buried in their armor. Removal of the five cover slabs, all reused stones (one is a marble roof tile, probably from a Greek or Roman temple), revealed instead a grave that had been in use for some time. Twelve adults were stacked one on top of another in the plastered chamber, as well as several infants. The chamber is only 1.96 m long, 0.60 m wide, and 1.00 m deep, so it was fairly crowded. All of the dead seem to have been originally buried with their heads to the west and their arms crossed over their abdomens, but several of the skeletons had been disturbed and some of the bones swept down to the eastern end of the chamber. In all likelihood, this tomb had been in use for several generations, with new burials being laid directly on top of the bones of earlier interments. The last burial was a single individual, but others were almost certainly groups; the first interment probably consisted of three adults buried simultaneously. The central individual lay on his or her back, and the other two lay on their sides facing the opposite walls. Aside from the bones, little in the way of grave goods was found: a pendant, made of bits of copper sheet, some gilt rings sewn to cloth backing, a fragment of gold thread, a bone die, and a few scraps of leather.
Tombs 2 and 3, directly to the west of tomb 1, were of similar construction-- mortar or plaster over mortared rubble walls-- and shared a common side wall, made from a reused Hellenistic ashlar block. Tomb 2 contains a similar thick deposit of burials, but it was decided not to excavate this tomb, as it was not in the way of the exhibit construction. The cover slabs, which were cracked in places and might not protect the site, were removed and the tomb filled with sand. Tomb 3 had been emptied at some time in the past, and contained only a few bone fragments mixed with fill dirt. Interestingly, one of the artifacts in the fill dirt was a roof tile of the 19th or 20th century, so this area must have been exposed relatively recently.
The style of these burials is fairly typical Byzantine Greek Christian, and similar tombs have been found elsewhere in Bodrum. Newton, the original excavator of the Mausoleion site in the 1860s, reported nearly identical burials in eastern Bodrum. These too were close to the surface of the ground and contained jewelry made from copper sheet. Due to the scarcity of grave goods and confusing stratigraphy in our excavation, it is difficult to date the burials. They may be as early as the fifth or sixth century A.D., or as late as the 12th century. Several bone samples are currently being radiocarbon dated by Odense University in Denmark, and it is hoped that these dates will help to clarify much of the stratigraphy in the excavation.
At present, it is difficult to say whether the tombs lay inside or outside of a structure, although a later building, which included the bases for two large columns or piers, was built adjacent to tomb 1. Byzantine Christians buried their dead inside churches on occasion, just as western Christians did throughout the Middle Ages, and it is possible that the adjacent walls belong to a church. It is also possible that the Knights of St. John chose the site of a ruined Byzantine church for the erection of their chapel; the general alignment of many of the medieval wall remains is close to that of the current building. If these tombs did lie inside a church, it suggests that the occupants were people of some importance.
The promontory is known to have been occupied in the Byzantine period, and visitors to the castle can already see an early Byzantine (fifth or sixth century) mosaic pavement to the northeast of the chapel. Although Halikarnassos was not near the Byzantine frontier until the Selcuk conquest of the 11th and 12th centuries, it still provided a large, sheltered, defensible harbor along the coastal sailing route from Constantinople to the southern shore of Asia Minor. Unfortunately, although there is frequent mention of Strobilos, a nearby port and naval station, in Byzantine sources dating from the 8th -13th centuries, there is little if any reference to Halikarnassos, suggesting that it did not hold the same prominence it had in Hellenistic times.
Much of the Byzantine-style rubble masonry is built over and around older structures composed of large, dressed stone. This type of masonry is typical of ancient Halikarnassos and can still be seen in She few surviving ancient structures in the town, notably the city wall adjoining the INA headquarters. In the Hellenistic period, Halikarnassos was the capital of the kingdom of Caria; before that it had been the seat of one of the Persian satraps governing Asia Minor. It was a large, prosperous city and home of one of the seven wonders of the ancient world, the Mausoleion, the elaborate tomb of King Mausolos. Mausolos is thought to have had his palace on the island where the castle of St. Peter now stands, and Danish archaeologists excavating in the castle and elsewhere in Bodrum believe it lay close to the present chapel. A small section of dressed ashlar wall is visible behind the Glass Wreck exhibit hall to the north of the chapel, and other Hellenistic remains have been found behind the chapel, to the northeast.
The Hellenistic remains discovered in the chapel consist of the base of a fortification wall, a level stone platform extending to the east beyond the excavation area, and several ashlar stones incorporated into the Byzantine masonry, including the large stone dividing tombs 2 and 3 (fig. 4). The wall itself is actually two fitted stone walls with rubble fill between, although most of the western side of the wall is missing. Surviving elements indicate chat the western side was finished, but the eastern face of the wall was quarry faced, or only roughly dressed, and was probably above ground level. Two courses of large stones survive, and She absence of "pry holes" (which serve to key the stones of two courses together) in the top of the upper course suggests that this is the top of the wall. Normally, such keying of stones is used in terrace or retaining walls to resist the outward pressure of the earth, but there seems to be no evidence of an associated terrace here. Although it may have been obliterated by later construction (in this case, walls A and B), the low height of the stones relative to the bedrock and surrounding terrain suggests that these stones are probably not a true wall, but a socle (or foundation) for a brick wall. The leveled upper surface of the rubble fill also suggests a base for a brick wall. The upper surface is covered by a thin layer of hard, pink granular mortar, which is not usually found before the Roman period, but may have been used in Hellenistic foundation work. Vitruvius wrote in the first century B.C. that Halikarnassos was famous for its brick masonry covered in stucco, and mudbrick walls on the stone socles are a common form of ancient Greek construction in Asia Minor. Stone socles were used to protect mudbrick from groundwater, and a deep drain on the east side of our socle is further evidence of this construction.
Ancient Authors report that the mouth of the main harbor of Halikarnassos was guarded by two fortifications, one on the Zephyrion promontory and another across the bay, and that the promontory guarded the "secret harbor" used by Carian warships (parts of this harbor were investigated in the early 1980s by INA divers during waterfront renovation work carried out by the town of Bodrum). The wall beneath the chapel may be part of one of these forts, but it is difficult to say how it fits in with the other bits of Hellenistic walls found to the north and northeast, or if it relates to the palace of Mausolos.
Although the remains excavated in the summer of 1994 may seem more enigmatic than informative, they do suggest the wealth of valuable information buried beneath the walls and courtyards of the castle of St. Peter. This stronghold is merely the successor to a series of fortifications built on the commanding Zephyrion promontory since ancient times. The tombs point to the proximity of a Byzantine church, and the discovery of the Hellenistic walls may contribute to the eventual location of the palace of Mausolos. Part of the dificulty in interpreting these remains lies in the small size of the area investigated. In many respects, our little excavation is more of a sounding or test trench, providing a glimpse of the many people who have used the rocky island at the mouth of Bodrum harbour over the last two and a half millennia and a guidepost for future work in the castle.
Further excavations in the castle and its environs are being conducted by Odense University, as an outgrowth of their work at the site of the Mausoleion. In the coming year, we hope that detailed study of the skeletal remains will begin; physical anthropologists at Texas A&M University are excited by the prospect of comparing the evidence from so large a sample, especially when there is the possibility that the individuals may be related. We also hope to make some headway in cataloguing the large variety of postsherds and roof tile fragments, as these may be important indicators of the date and function of the structures. The results of this research will be published in a new journal, Halikarnassian Studies, and will be an important aspect of the broader picture of ancient and medieval Bodrum/Halikarnassos.
Acknowledgements: We would like to thank, first of all, the Director, Oguz Alpozen, and the staff of the Bodrum Museum of Underwater Archaeology, who were more than generous in providing equipment and materials for the excavation. We also offer thanks to Dr. Poul Pedersen and Odense University for friendly advice, assistance with radiocarbon dating, and the opportunity to publish in their journal. Conservation help was provided by Dr. Claire Dean.