by Nicholas Cahill
The Sardis Expedition conducted a ten-week season of excavation, conservation of monuments and artifacts, research, and publication during the summer of 2012 at the city of Sardis, followed by a fifteen-week season of excavation at Bin Tepe. We again offer thanks to the Ministry of Culture and Tourism for permission to carry out archaeological research, and especially General Director Abdullah Kocapınar, former Director General Murat Süslü, and Excavations Division Director Melik Ayaz, and Manisa Museum director Sevgi Soyaker. We particularly thank the representative of the Ministry of Culture, Melek Yıldızturan of the Ankara Museum, whose professional knowledge and commitment to archaeology and preservation were deeply appreciated, as was her sense of humor and good will.
Excavation at the city focused on a hill near the city center known as Field 49 and in the temple of Artemis (figs. 1, 2). Previous excavation on Field 49 had showed that it was occupied from the Lydian through the Late Roman periods, and was used as a cemetery thereafter. In the Lydian period this hill was enclosed with a monumental terrace wall, which was probably linked to the neighboring terrace, ByzFort, to form a single raised platform overlooking the lower city. We have previously proposed to identify this platform as the palatial quarter of Lydian Sardis.
The Lydian terrace wall had two phases: an earlier wall made with roughly worked polygonal boulders, and a later wall built of cut limestone ashlar masonry (figs. 3, 4, 5). One goal of excavation was to understand the early boulder terrace wall, one of the earliest examples of monumental construction at Sardis. On the north slope, the wall is about 3 meters thick, and has been exposed now to a total length of about 26 meters (fig. 3). On the west slope it is exposed for a length of about 11 meters, and its full width has not been exposed (figs. 4, 5).
One goal of excavation in this area was to recover deposits associated with the construction and use of the early terrace wall. The ceramic evidence in both north and west trenches was quite limited, and included only local ceramics, but suggests a date in the second half of the seventh century. This is a key period in the history of Sardis, when the Lydians were expanding their rule over western and northwestern Anatolia, concluding treaties with the Assyrians and Egyptians, and when gold from the Pactolus and elsewhere made the Lydians the richest people in the region, leading to the invention of coinage. The boulder wall is approximately contemporary with the earliest phase of the Lydian fortification, and may belong to a program to monumentalize the newly-wealthy capital.
The early boulder wall was replaced in the first half of the sixth century BC by a limestone ashlar wall (figs. 4, 5). It originally seems to have retained a packing of loosely set stones. This is preserved a couple meters higher than the preserved top of the terrace wall, and originally rose higher still.
Another objective of the summer was to locate Lydian buildings on top of this terrace, to learn more about the function of this area in the Archaic period. Three areas were opened, in the north, center, and southern parts of the hill. All these areas produced important remains of later periods. Hellenistic remains were found throughout the hill. In the southern trench, a sequence of massive foundations must have supported substantial buildings, but the superstructures of these buildings have now entirely disappeared (fig. 6). The foundations of one wall, sloping down towards the edge of the hill, indicates that the terrace had partly collapsed and eroded away by the Hellenistic era. The terrace must have been restored in the Hellenistic period, but this phase, like the superstructures of the Hellenistic buildings here, is now destroyed, perhaps in the earthquake that leveled Sardis in 17 AD.
Among the spolia built into one of these Hellenistic or early Roman (pre-17 AD) foundations is at least one fired brick, among the earlier examples of this technology in Asia Minor. Other bricks are found in slightly later Roman walls, which date to the first century AD.
Reconstruction after the great earthquake seems to have taken place relatively quickly, and includes a number of complicated early Roman phases. The terrace wall was rebuilt, reusing the same limestone blocks but set in a backing of mortared rubble (fig. 5). The most coherent building phase on top of the hill at least included two rooms, one with a well-preserved furnace (figs. 7, 8). The furnace has a domed chamber, and a bench with two clay-lined channels. When it was first discovered, we suspected that this furnace might have been used for the production of glass intaglios, since many unfinished intaglios have been found on the hill, many in early Roman contexts. However, during its excavation we discovered no intaglios in this room, and the furnace must have been used for some other purpose, perhaps for baking.
The walls of this and the adjacent room were almost entirely robbed out, but appear to have been made of plastered and painted stone and mudbrick. Thousands of fragments of painted wall plaster were found in the fill of the room (fig. 9). Designs include brecciated marble, floral bands, red panels with tendrils and scrolls, and other patterns (fig. 10). Large fragments of white ceiling plaster were found fallen in situ as well, preserving reed impressions on the back surfaces, with an elaborate cornice molding (fig. 11).
Other finds from the early Roman strata include 16 more unfinished glass intaglios from all trenches on the hill, making a total of 41 found here in the last four years (fig. 12). Among the figures represented are Jupiter, Ares, Hermes, Demeter, Athena, Hercules?, paired figures, a Capricorn, and a male head.
Coins and pottery document the occupation of the hill during the second, third, and fourth centuries, but little architecture is preserved. In the late Roman period a major construction phase is identifiable in all parts of the hill. In the north, this includes a building incorporating spolia from a small building, perhaps a shrine, of earlier Roman date. The blocks include two or more cornice blocks (one of them a corner block with the corner of a pediment and the setting for an acroterion, another with a lion’s head spout), and a frieze block with bucrania and garlands, matching one discovered in 1982 on this same field. A massive drain leads from this late Roman building down the slope to the northeast.
Towards the center of the hill is another building, perhaps a villa with tile and opus sectile floors, and a series of pipes and drains associated with hydraulic features, now eroded away (fig. 13). In the southern part of the hill, late Roman remains were entirely eroded away on top of the hill, leaving only a sunken basin full of pottery of the late 4th or early 5th c AD. On the west slope of the hill, a basement was dug into earlier levels, accessed presumably by a ladder or wooden staircase. A hoard of 123 small bronze coins, found in 2011 under a poorly made “bench” on the east side of this cellar, was cleaned and studied this season. The latest coins date to the end of the 5th c AD (Zeno, 474-491). The hoard contained a significant number of 4th and earlier 5th c coins, however, providing a picture of the range of coins circulating at this time.
The hill was apparently abandoned sometime later: the latest coin from the hill was a pentanumium of Maurice, 582-602 AD. After the buildings collapsed and were buried, the hill was used as a cemetery. Nine more graves were excavated in 2012, but as in previous years, there were no finds associated with these graves, and their date remains uncertain (fig. 14).
These later buildings and strata prevented us from reaching undisturbed strata on top of the hill, except at the edge where erosion had removed all Lydian strata down to the terrace packing. One area of burned debris was found, consisting of burned earth, mudbrick, and vitrified mud plaster. This is similar to the burned levels found elsewhere at Sardis, associated with the destruction of the city in the mid-sixth century. The pottery from this area included an unusually large amount of fine Lydian ceramics dating to the 7th and earlier 6th centuries BC. In 2011 one of the significant finds from this stratum was a jasper sealstone on a bronze mount. This year, the same stratum produced a small fragment of a burned ivory furniture inlay. The reverse is incised to provide grip for adhesive; the obverse shows the leg and foot of a figure wearing a skirt, and the mane of a lion, perhaps a “Potnia Theron” holding lions by their tails (figs. 15, 16, 17).
Notable in excavations during this and previous years is the lack of any architecture from the Achaemenid period. For the most part, Hellenistic buildings were constructed directly on the eroded remains of the Lydian buildings dating 300-400 years earlier. A pit was found at the north end of the hill, containing pottery of the late 6th or 5th c BC. The almost complete lack of Persian period pottery reinforces the suggestion made previously, that urban Sardis was forcibly abandoned under Achaemenid rule, leaving it an administrative and military center but not a city in the usual sense.
Sanctuary of Artemis
Research in the Temple of Artemis was intended to help understand the chronology and phasing of the eastern porch and peristyle, and included a small probe in the east pteroma. The cella of the temple was built in the Hellenistic period, while the porch and peristyle columns were not built until the Roman era. Moreover, the peristyle was begun on the east side, the original back of the temple, and was never carried much beyond the east facade. This must be associated with the division of the temple into two chambers, one facing west and the other facing east, with the new door cut through the original back wall. The new east-facing chamber presumably belonged to the imperial cult, and probably housed the half-dozen or so colossal statues of the Antonine imperial family found in and around the temple; the columns monumentalized this entrance. The conversion has generally been dated to the second century AD.
The probe dug in the pteroma showed that the mortared rubble bedding in which the column foundations were set is thicker than 0.7 m, and probably represents structural support for the columns rather than simply bedding for a floor (fig. 18). Moreover, it seems continuous from the porch to the peristyle columns, and there is no reason to think that the peristyle and porch were built at different times.
Excavation in 2011 recovered a rich deposit of pottery, bricks, rooftiles, and other debris in the foundation trench of one of the porch columns, some still mortared into the foundation of the column. In 2012 this deposit was mended and studied. It consists of five more or less complete vessels: a large lamp of Broneer Type 21, a local krater, a local jug, a pseudo-Coan amphora, and a cooking pot; and about 13 kg of sherd material including stamped ESB, thin-walled ware, and other diagnostic forms (fig. 19). The deposit dates to the middle decades of the first century AD. Much of the standing architecture, however, seems to date to perhaps a century later. It may be that work was begun on the foundations of the peristyle a generation or so after the earthquake of 17 AD, but was discontinued for some reason. The project was then renewed in the second century AD, although even then, only a few of the columns of the temple were ever erected. This adds a new and previously unsuspected phase to the Temple of Artemis. A possible parallel for this division of the cella is the temple of Apollo at Corinth, whose interior dividing wall may be associated with a remodeling of the temple in the mid-first century AD.1
Site Conservation and Lydian Altar
The three-year project to conserve the Lydian altar was completed (figs. 20, 21). The sandstone foundations of the stairs across the front of the altar were consolidated, stabilized, and then protected by installing new travertine blocks to replace the original marble stairs, which had been robbed in antiquity. In addition to protecting the original fabric, this reconstruction makes the building more easily understandable to visitors, and is relatively easily reversible. The walls of the altar were restored to the condition in which they were preserved when first excavated in 1910, and capped with slate. The interior of the altar was protected with geotextile and gravel, and the monument explained in new signage.
Two columns of the temple have stood to their full height since antiquity. Their capitals, however, are split by natural flaws in the marble. These were cleaned and consolidated to prevent further damage (fig. 22, 23).
In the century since the temple was excavated, its walls have become covered by black cyanobacteria and gray-green lichens, which erode and damage the stone. Conservators developed a new protocol for killing these microorganisms without harsh chemicals or pressurized water. This will be pursued in future years.
A three-year project to consolidate and stabilize the mosaics of the Synagogue was also finished in 2012. Conservators treated the mosaics of the Forecourt, stabilizing areas around the mosaic panels, filling voids under the mosaics, replacing infills, repairing cracks and losses, and removing biological growth. The original mosaic inscriptions had been removed to the Manisa Museum and replaced with painted casts. These were repainted with Silin paint.
Work also proceeded on the “Touristic Enhancement Project” to roof the Synagogue and Lydian Fortification, and develop for visitors the area between them, including the Roman marble road, Lydian gate, late Roman houses, and other remains.
Torrential rains during the winter of 2011 caused part of Building A, an unexcavated late Roman building in central Sardis, to collapse. The surviving segments of the north face of this building had spalled off since antiquity, leaving the upper parts dangerously overhanging and in danger of further collapse. We therefore stabilized the remaining portions of the wall to prevent further collapse; these will be completed during coming years.
Research and publication included the completion of a catalog of all coins discovered since 1971, to be published in print and on line, study of Hellenistic and Roman ceramics, including analysis by XRF, study of figural terracottas, archaeoseismology, and sarcophagi. Publication of sectors HoB and PC proceeded with further study of Lydian ceramics and stratigraphy, while other ongoing publication efforts include churches, the Synagogue, and the temple of Artemis.
An important focus of excavation and research at Sardis over the past 50 years is the great burial mound in the necropolis of Sardis, Bin Tepe, known as Karnıyarık Tepe (figs. 24, 25). This tumulus, 230 m in diameter — as wide as the Great Pyramid in Egypt— and 53 m high was tunneled between 1963-1966 by Crawford H. Greenewalt, jr. (figs. 26, 27). Early excavations discovered the crepis wall of an earlier tumulus, about 90 m in diameter, which was buried, still unfinished, when the tumulus was expanded to its present diameter. The tunnels continued to the center of the mound, and more than 300 m were dug before the project was brought to a halt. The Expedition returned to this tumulus over the years to do geophysical survey, coring, study the masonry and other aspects of this important monument, as well as maintaining the excavation houses. Christopher Ratté showed that this was not the tomb of Gyges, as Prof. Hanfmann had believed, but must date to the sixth century BC.2
In 2011 a third campaign of geophysical survey was completed. The mound is so big that ERT and magnetometer survey from the top of the mound failed to locate even the tunnels we knew were there. But survey within the tunnels revealed a strong anomaly near the mound’s center. This was detected in three different techniques, ERT, magnetometer, and ground-penetrating radar. ERT showed an anomaly measuring about 4 x 6 m, indicating an area of high electrical resistance such as a void in the tumulus fill. The magnetometer showed a strong dipole at the same location, indicating the presence of a large iron object. Ground-penetrating radar showed a reflective anomaly here, too. In 2012, therefore, we began a joint excavation with the Manisa Museum, after the regular Sardis season had ended. The tunnel floors were cleared of tons of fallen earth, and a new railroad, electrical and ventilation systems, and steel shoring were installed (fig. 28).
Two tunnels were excavated through the anomalies down to bedrock, but without finding any trace of a chamber or other feature. The explanation for the anomalies is difficult to understand, but it is clear that there is no burial chamber at this location.
Nonetheless, we continued excavation for almost four months, digging about 220 meters of tunnels, covering about half the area within the crepis, and showing that there is no chamber in those parts of the mound. I wish to thank Nur Soyer, Nilufer Parlıti, a hard-working team of about 20 workmen, and particularly Teoman Yalçınkaya, for their constant good cheer and almost superhuman efforts in very difficult and uncomfortable circumstances. Among the interesting but unexplained results were the discovery of three roughly worked limestone cylinders, set upright in the fill (fig. 29). A similar cylinder was discovered in 1965 near the crepis wall; their purpose remains a mystery. But in the end, one can only quote Prof. Hanfmann as he too admitted defeat in 1966: “in the battle of man against mound, the mound won.”
In the spring of 2013 Sardis and Bin Tepe were inscribed to the UNESCO World Heritage Tentative List. We look forward to helping to bring this application through to completion, and to contributing to the preservation of the ancient remains and modern beauty of this magnificent site.
(Adapted from Kazı Sonuçları Toplantısı 35: 119-135)