The institute’s second excavation of 1975 was conducted near Şeytan Deresi (Devil’s Creek), on the north coast of Turkey’s Kerme Bay. AINA (now INA) had surveyed the wreck in 1973 and raised two huge pottery vessels, along with a number of pot sherds. The site seemed untouched since then.
The Site & Cargo
The jars had been found at the base of a sloping field of rock outcrops and boulders. No traces of wood had been found, to obvious disappointment, nor were there any non-ceramic objects other than a fishing weight, which was not necessarily antique.
This led to the belief the ancient ship capsized, for the sand was deep enough to have preserved traces of wood had there been any below the cargo. No effort was spared to locate hull remains and other concentrations of pottery.
The date of the loss of the cargo is circa 1600 B.C. based on parallels to ceramics from other sites such as Beyçesultan, Troy, and Minoan sites in Crete.
The shape of the two-handled pithoi from Şeytan Deresi resembles a middle Minoan spouted shape that continues, less similarly, into the late Minoan I period. The evidence points to a date in the late Middle Bronze or early Late Bronze Age.
The craft that carried this cargo need not have been large, and may have been no more than a coaster transporting newly finished jars from one village to another. This would explain why no cooking implements, lamps, and other ships wares were found.
Attributing his initial impression that the Şeytan Deresi pots belonged to the Bronze Age to wishful thinking, George Bass concurred with the view of the specialists he consulted after the survey and dated the wreck to the Iron Age or Archaic Period in his first reports of the discovery. The general shape and distinctive handles of the krater (SD 9) raised during the 1973 survey seemed to justify this late date for the specific artifact and the wreck as a whole, but convincing parallels for this and other Şeytan Deresi finds, from either the Iron Age or the Archaic Period, were lacking.
Even as a Geometric or Archaic assemblage, the discovery seemed important and the site worthy of excavation. INA scheduled a project for 1974 but military conflict in the region thwarted plans. Excavation finally took place the following year and lasted for almost six weeks. The original team comprised 18 members, yet on October 14, the last day of excavation, only six of the crew remained as the others, mostly student volunteers, had to return to their academic pursuits. It was the first excavation undertaken by the nascent Institute of Nautical Archaeology (INA), and the first official collaboration of the Institute with O?uz Alpozen, then a commissioner overseeing the project on behalf of the Turkish Archaeological authorities and now director of the Bodrum Museum of Underwater Archaeology.
The morphology of the site and complicated layout of the artifact scatter emerges from the site plan and the excavators’ descriptions. The finds spread over a 42-square-meter area but the main concentration, including the two complete vessels raised during the survey, lay at the northern end of the site in a sandy bottom at a depth of 32 to 33 m. Above a rocky slope located east of this sandy area, at shallower depths (27 m), another stretch of sand held a pithos (SD 13) not seen during the 1973 survey. The pithos contained large rocks, an amphora neck (SD 18), an amphora base (SD 17), a lead fishing weight (SD 22), and, significantly for the understanding of wreck formation processes, two sherds of SD 6 and SD 8, both amphoras of the same type. Not visible on the site plan are smooth stones, “presumably ballast” according to the excavator, of which the divers removed several basketfuls in the sandy area of the site. Also not indicated on the plan are additional finds from shallower waters. Bass mentions “large sherds and a handle identical to those of pithos SD 10” lying in 2 or 3 m of water and heavily concreted to the rocky bottom between the site and the shore.
Having moored a 50-ft barge over the site, the team laid down a PVC grid of four 2-m squares around the findspots of krater SD 9 and pithos SD 10. A metal grid soon replaced the PVC piping, and squares were added as the excavation progressed. Using 30 cm steel probes, they located concentrations of pottery buried in the sand. Since most of the sherds recovered from this area joined to form complete or nearly complete vessels, Bass was satisfied with the thoroughness of excavation.
In spite of combing of the area, the rocky slope, where the divers had seen and collected random sherds during the 1973 survey, did not yield any further artifacts in 1975, except for the base of amphora SD 6. Bass points out that in their hasty efforts in 1973, the divers are not likely to have collected all sherds present among the rocks and, adding that few of the sherds from the rocks joined, speculates that other parties may have visited the site before the archaeologists returned. A further possibility, not necessarily exclusive of Bass’s interpretation, is that sherds covered with concretions escaped notice both times; Bass himself notes that the base of SD 6 “was nearly invisible under concretion.” Furthermore, the relatively short duration of the project coupled with the dwindling crew towards the end allow for the possibility that, despite meticulous efforts, the divers missed valuable scraps of evidence scattered throughout the site.
After conservation and mending, the finds of the main group were estimated to represent 17 individual vessels, of which 10 were restorable to their original shape. A perusal of the site plan reveals that, in general, pieces from the sandy area, especially those belonging to vessels restored to complete or near-complete profiles, were lying in discrete groups. Despite exceptions to this pattern, it seems that most of the ceramic vessels lying in the sandy area had reached the seabed intact; four of them (SD 9, SD 10, SD 12 and SD 13) remained whole, and were slowly buried in situ. Others broke upon impact and a deep layer of sand gradually covered their remains. Some fragments, however, were found among sherds of different vessels. Currents and marine creatures may be responsible for such movements, which obviously occurred before substantial deposition of sand over the site. In one extreme case, the fragments of amphora SD 6 were dispersed over a relatively wide area, primarily in the deeper sandy section; the base was concreted to the rocky slope and a body sherd was found inside pithos SD 13 in shallower water.
Despite the absence of ship timbers, it seems that the assemblage at Şeytan Deresi represents a single cargo that sank in a shipwreck, as opposed to items jettisoned from aboard a ship, or debris from shore. The location of the site at considerable depth off the eastern point of Kerme bay precludes the possibility of refuse from land, especially since no settlement site is reported in the vicinity. The uniform fabric of the ceramics ascribe the assemblage to a single source and a single shipment while the relatively concentrated layout of the finds and the presence of ballast stones among them contradict the scenario of cargo jettisoned from a ship in distress.
Comparison & Interpretations
In his study of stratification and contamination in ancient Mediterranean shipwrecks, A.J. Parker provides a simple classification of shipwrecks, based on their state of preservation. He groups shipwrecks into three broad classes: well preserved, relatively coherent, and scattered. In well-preserved shipwrecks, “coherent timber structure and cargo survive, with a dense concentration of material.” In relatively coherent shipwrecks, “part of the disturbed cargo, and, possibly, some structural timbers, survive more or less in situ, though the rest of the wreck is dispersed or lost,” while in scattered shipwrecks only fragments of the cargo survive, widely spread, and small finds or fragments of the hull are only occasionally preserved.
Individual shipwrecks are likely to fall between categories, yet Parker’s classification is useful in defining the degree of disturbance each site has suffered. The relatively dense main concentration of pottery, which appears to have survived more or less in situ, the presence of ballast stones, and additional ceramic finds scattered over a wide area including much shallower waters, combined with the absence of any hull remains or small finds, place the Şeytan Deresi assemblage between Parker’s classes 2 and 3.
Despite its nature as an open deposit, the Şeytan Deresi site offers sufficient evidence for the reconstruction of the main shipwreck event and speculation on the ship’s original contents. The intrusive material at Şeytan Deresi is easily detectable; comprising distinctly random pieces it does not detract from the uniformity of the main assemblage. Its presence indicates that, in addition to the main event of wreck formation at Şeytan Deresi, at least three other incidents, probably minor and chronologically far removed from one another, contributed to the final morphology of the site.
The circumstances that caused the shipwreck at Şeytan Deresi pertain to local geography and weather conditions. The coast of the Gulf of Gokova features a series of parallel mountainous ridges and deep valleys that run in a north-south direction perpendicular to the shore. The excavation team had pitched their tents on the beach of the small bay at the issue of a deep valley known as Şeytan Deresi. The wreck site itself lies off the bay’s easterly point, labelled “Gatal Burun” (Cape Fork) in nautical charts. These ominous names warn the traveler about the prevailing conditions in the area. Mehmet Turguttekin, then captain of the Kardeler, told Bass that when he was a young man he had seen a waterspout right inside the bay. Bass describes the eerie atmosphere at camp in the evenings, which he ascribes to the unpredictable gusts of wind sweeping down the valley and menacing the camp only to die down as suddenly as they had appeared. On one occasion, the violence of such a blast caused the anchors of the expedition’s barge to drag, while at another time it snapped a cable. Is it likely, as Bass postulates, that a similar incident caused the ancient wreck? The location of the underwater site approximately 100 m off the easterly point of the bay suggests that the ship was rounding this cape when she came to grief. Since Kerme bay does not seem to hide the navigational hazards, such as reefs or shallows, that beset the neighboring bay of Maz?, a violent gust of wind that hit the vessel unexpectedly as she was entering the bay and caused her to list dangerously and take in water or capsize may explain the disaster.
Having rejected the idea that the cargo might have been tossed overboard to lighten a ship in distress, Bass wondered initially whether the ship had tipped and spilled its cargo when already on its way to the bottom or had capsized while still at the surface, spilled its cargo, and then floated away to be smashed on the rocky shore. Later, based on the group’s failure to locate even the slightest pieces of hull timbers he leaned decisively towards the idea of capsizing. He states that the sand in the area of the main concentration was deep enough to have preserved at least some of the timber, had the hull gone straight to the bottom with the cargo. In very shallow water (2 or 3 m deep) near shore, findings of pithos fragments, including a basket handle similar to those of pithos SD 10, provide a relevant, yet ambiguous, additional clue. Bass offers two possible explanations for the presence of these sherds so close to shore. If the entire cargo spilled as soon as the vessel capsized, then some of the pithoi might have floated away to be smashed onto rocks near the shore. It must be kept in mind that one of the intact pithoi (SD 13) had certainly floated away and come to rest in a sandy area about 30 m away from the main concentration in shallower water (27 m) above the rocky slope.
Presumably, the hull hit the rocks nearby, or was carried out to sea by the surge. If, on the other hand, the vessel spilled most but not all of her cargo upon capsizing, she may have floated towards shore with a couple of pithoi still in her hold. According to this scenario, when she eventually hit the rocky shore, her broken timbers floated away, leaving pithos fragments to indicate the area of the final impact.