Chapter 5. Lydian III: The Destruction Level (Last Quarter of the Eighth Century B.C.)

The first clearly defined occupation floor above the Iron Age levels in the Lydian Trench at HoB—and the lowest level that had large-scale exposure in the excavations—gives us a glimpse of the quickening life of Sardis as it began to grow into an international center. The elevation of the surface varies by about 0.6 m, from *97.6 to *97.0, according to the slope of the ground and accumulations of debris around specific features, such as the remains of walls, but in general it is remarkably uniform. The finds from this phase all over the excavated area bear out a characterization of growing prosperity and the renewal of foreign, especially western, contacts at the end of the long period of isolation and assumed poverty of the Iron Age.¹

The Lydian III occupation surface showed consistent evidence of damage from fire, as well as indications of a major battle. Among the many signs of violence were the remains of a young girl, another child, and at least fifteen adults.² The Destruction Level, as we call it, has provided the opportunity to observe the local life of ordinary Lydians at a particular moment in time, a moment that was interrupted by catastrophic burning. Through the pottery and other objects found here, one must try to envision their activities through the lens of the material remains, as if the destruction had been held back for another moment, the broken crockery restored, and the activities resumed. We had to concentrate on the inorganic remains, particularly the pottery, as almost all perishable materials had been consumed in the fire. Putting together the evidence for the fierce burning, the scatter of potsherds, and the skeletal remains, it is clear that there was a major conflict here.

The destruction was originally associated by George Hanfmann with the attack on Sardis in ca. 652 B.C. by the Kimmerians, a group of Central Asian nomadic tribes who are said to have entered Anatolia in the eighth and seventh centuries, and to have attacked or destroyed such cities as Gordion, Ephesus, and Sardis.³ This speculation was readily accepted and not really doubted until Judith Schaeffer noted that a critical Corinthian jug (HoB 355; Fig. 5.2), found in pieces scattered on the floor of the Destruction Level, was Late Geometric rather than Protocorinthian as had previously been thought. A reappraisal of the imported Greek pottery in this level at Sardis has led to the present conclusion that the destruction should be dated to the later part of the eighth century B.C., and not to the Kimmerian attack of ca. 652 B.C. Therefore, although earlier publications refer to this as the “Kimmerian Destruction,” we will refer to it simply as the Destruction Level.

The overall extent dating to Lydian III can be broken down into the East End, the Central Area, and the South Side (Fig. 5.1). A sub-area, which could be attached to the Central Area, is formed by the space around an oven or furnace (see Fig. 5.14); since this was clearly a place for a special activity, it is appropriate to deal with it separately.

The total area exposed in the Lydian Trench of HoB measured ca. 77 × 25 m, or 1,925 square meters,⁴ and thus the excavated area of Lydian III represents somewhat more than 40 percent of the total exposure of sector HoB.

From the southwest corner of the trench, the destruction surface slopes down to the north.⁵ In the middle of the trench it is more or less horizontal, and toward the east it rises a little over half a meter.⁶ This surface can be followed across the whole width of the trench from north to south,⁷ a distance of more than 30 meters. Traces of this surface can be found at the northeast and southeast corners of the trench too, but it is absent at the northwest. Over it was a layer of clay about 0.20–0.40 m thick that apparently had been laid down on purpose;⁸ above that, a layer of heavy gravel about one meter thick, deposited by floods, clearly divided it from the occupation level above (Lydian II; see Chapter 6), as seen elsewhere in the trench.

As we explored the 800 square meters of the Destruction Level, a consistent pattern of occupation emerged. Evidence for the unity of the level, in addition to the continuous layer of hard clayey earth across the trench, was the frequent occurrence of patches of heavy burning indicated by quantities of charcoal, ash, and reddened lumps of clay. In several instances, the heat was sufficient to alter the color of the pottery—in some cases to such an extent that joining pieces from the same pot have quite different colors, with no intermediate shades, showing that the vessel had been shattered into pieces, some of which landed in the conflagration, others not (HoB 326, HoB 338). The fire itself, destructive though it was, would not have compared to the heat and devastation observed at Gordion that arose from a major fire during the ninth century.⁹ The burning of the heavy wooden beams used in construction in that city had a much more dramatic effect on the pottery and other items, even causing pottery to melt,¹⁰ while the reed and straw buildings at Sardis could not have burned so hot as that.

A large quantity of occupational debris in the form of potsherds, disintegrated mudbrick, and miscellaneous artifacts was found in the layers of clayey earth separated by strata of mixed sand and gravel overlying the burned floor. Although the pottery recovered from two areas in the southern area of the trench was not taken directly off a floor, it was sealed between the Lydian III floor at *97.0 and the Lydian IV floor at *96.4, and therefore provides a useful general picture of ceramics and other finds in the early or middle of the eighth century (HoB 278).¹¹

Low, scattered remains of narrow fieldstone walls were found over the whole area.¹² They apparently supported mudbrick or wattle and daub superstructures with reed and clay roofs. In both the southwest and central areas of the trench, the surface is interrupted by a large number of storage pits (bothroi) that were dug into the Lydian III level from above during the period of Lydian II (see Chapter 6, pp. 85–86) (Fig. 5.3). The burning was particularly intense and was closely associated with the remains of structures in the form of stumps of walls, piles of stones, and lumps of mudbrick.

When the finds from the many different places in the trench that go to make up the Destruction Level were combined, there were many matches (if not always joining) among pieces of pottery that had been found in different years. The Corinthian Late Geometric jug (HoB 355; Fig. 5.2), mentioned above, that was critical for dating the Destruction Level, was composed of many pieces found over the years and finally combined by Judith Schaeffer, who wrote, “The fragments were scattered, as though broken and trampled underfoot during the siege.”¹³ A striking example of the spatial displacement of fragments from the same pot occurred when a banded body fragment of an East Greek Geometric krater from a 1985 test trench was found to join a piece that had been found in 1966, nineteen years earlier and at least 13 meters away.¹⁴ Even so, sufficient numbers of whole pots or large pieces found close together put the integrity of the level beyond doubt and offer an important “fixed point” for the development of Lydian pottery. They also allow a glimpse of the penetration of Greek pottery in substantial quantities to the domestic and industrial levels of ordinary Lydian society in this period.

Since the surviving architectural fragments are scanty and do not yield sufficient evidence for an unassailable interpretation of either the inside or the outside of any structures, it is not possible to present the finds by building, as will be done for Lydian II and I. Whether the structures were houses or workshops is not clear because of the circumstances of their destruction and the removal of many stones. The bits of walls and the rich pottery and other finds that go with them suggest that the inhabitants in this period were better off than their predecessors, and help us to picture Lydians working with the artifacts and living in the buildings of which so little is preserved.

A substantial area of about 170 square meters was exposed in the northeast corner of the trench (at E5–15/S85–102, at ca. *97.60). A widespread floor¹⁵ produced heavy concentrations of charcoal, undoubtedly due to the violent attack that left evidence of fire over much of the trench at this level. There was little in the way of evidence of a building or shed here, except for baked mud with imprints of reeds (HoB 507A) and a single concentration of stones.¹⁶ The importance of the area lay in our learning that the whole level continued to the north, and in our retrieving the groups of pots and other artifacts, especially a bone toggle (HoB 324), associated with it (see HoB 310–HoB 322).

The succession of floors in this area is more complicated than elsewhere (Fig. 5.4). Seven distinct floors from Lydian I to Lydian IV can be distinguished here, but none of them is complete or offers a substantial exposure to indicate a major level; and only two of them are from Lydian III. The pottery proportions from several samples in this area were 30–50 per-cent Gray Ware; 30–40 percent coarse red ware; 10–30 percent cooking ware; and 5–10 percent painted and miscellaneous.

By far the most pottery from the Destruction Level was found in the Central Area, including several well-preserved Lydian and imported pieces. In fact, because of the number of recognizable imported pots, the Central Area is important both for the chronology of the occupation levels and for the pottery styles of the Lydians. Those imported pots allow us to establish a chronological horizon of the last quarter of the eighth century B.C.

The Central Area of the trench displayed considerable disturbance in Lydian III, but we were able to establish a floor level, at *97.4, on which had once stood several small buildings that were destroyed by a fierce fire. The pattern of patches of burning and of wall stubs was clear, as well as the remains of bothroi dug in from above during Lydian II, or hollows used for unclear purposes on the top of the destruction debris (Fig. 5.5). Finds from the floor level are primarily domestic and local, but there was also a large amount of metal debris and imported Greek pottery that can be dated to the second half of the eighth century.

Many of the pits are close together, and several were apparently intentionally connected by runnels in the clay surface of the ground (Fig. 5.6). This means that we must consider some sort of purposeful activity centered around these pits. About seven meters to the west, at W10–18/S100–105 *97.6–97.0, further evidence of the erosion of the surface of clay levels was found (Fig. 5.7). An irregular grooved surface nearby (Fig. 5.8) may have traces of the impressions of the sticks of wattle and daub. The runnels as well as the erosion of the clay surface would have to be placed directly after the destruction of Lydian III and before the deposition of the thick layer of gravel that rose to above *98.0 in some areas.

The Central Area gives us the most coherent view of the Lydian III Destruction Level and the densest and most varied evidence of human activity. In spite of their poor state of preservation, buildings had clearly been here. Two short wall fragments, one at the east at W0/S100 and one at the north at W3/S95 (see Fig. 5.4), are the nucleus of the Central Area. They may have formed the outline of a rectangular building ca. 8.5 × 5 m, perhaps with the south wall dependent on two large supports whose postholes were found directly to the south of the interrupted lines of wall stubs. A slightly shorter building with a porch is also a possibility, but at present no other Lydian building is known with this kind of design.

There were many fragments of individual pots, such as the Late Geometric Corinthian jug mentioned above (HoB 355; Fig. 5.2) and the large imported East Greek Geometric krater (HoB 351). Also found were more personal artifacts—like a bronze fibula (HoB 363), a pair of tweezers (HoB 362), or straight pins—than had been found elsewhere. Lying on the floor were a Greek Geometric cup rim with a geometric “tree” design (HoB 347) and a bone handle for a knife with part of an iron rivet (HoB 367). The fiery aspect of the destruction was more obvious in patches of reddened earth and pieces of pottery so burned or reduced that the color of their body or decoration is sometimes altered or obscured. Again, many fragments of burned wattle and daub with reed impressions (HoB 358–HoB 361) bear witness to the conflagration that destroyed any building remains at this level.

Human Remains

Among the grisly reminders of the effects of the catastrophe on the human occupants of HoB was the skeleton of a girl of about six or seven years discovered on the destruction surface among carbonized reeds or straw (Fig. 5.9).¹⁷ A recent reexamination of her remains confirmed that there was no sign of disease on the bones, and the breakages on the cranium were postmortem.¹⁸ As found, the skeleton was missing her left leg and foot—for what reason we cannot tell—and her cranium was fractured. Otherwise, the rest of the bones were in good condition and not much disturbed at the time of discovery.¹⁹ This would seem to make the action of dogs or jackals unlikely, but one would expect a stump if the unfortunate girl had only one leg. Perhaps she was asleep when fire overtook what seems to have been a light structure, and the reeds might have fallen on her, or she might have been asphyxiated. This individual may have been associated with the more substantial building represented by a considerable pile of stones to the south, some of which had been burned and cracked by the heat. The partial remains of another child and two young adults were found nearby.²⁰

Further evidence of the human toll of the disaster was found in the northwest corner of the Central Area where a stone-lined pit was filled with human and animal bones. The wall lining the pit, ranging in width from one to four stones, widened to a maximum thickness of about 0.2 m and a maximum height of 0.6 m (Fig. 5.10).²¹ The pit ran for two meters, west to east, with a rounded corner at each end. The north side was buried under the edge of the trench.²² The pit clearly belonged with the Destruction Level, as indicated by its level and contents. A large fragment of the rim of a Corinthian Geometric kotyle from the area (HoB 349) helps to define the chronology of the whole level.²³ The sherd must have been near the top of the pit, because the bone report states that there were no sherds in the matrix.

This stone-lined pit contained a mass of bones, 90–95 percent of which showed evidence of burning. Human and animal bones were charred and mixed together. Human cranial fragments indicated that both males and females, ranging in age from childhood to maturity (ca. 35 + years), were present. An analysis of the human vertebral remains gave a minimum estimate of ten individuals in the pit, the rest being the bones of sheep or goat, cow, and horse.²⁴ We may suppose that a number of domestic animals from yards or stalls were caught in the attack and fire. The disposition of the bones within the pit does not indicate any formal burial, as there was no directional orientation in the layout. No signs of disease or violent death were found, suggesting that the bodies were burned after death; but on the other hand, the fact that the pit lies not far below the burned layer suggests that it served as an emergency burial resulting from the disaster associated with that level.

Five more victims of the destruction were found nearby, again having been put in what was apparently a hastily dug shallow hole about one meter deep below the floor level (Figs. 5.11, 5.12, 5.13).²⁵ Unlike the burned and mixed bones in the other pit, these skeletons were more or less articulated. It is likely that the bodies were thrown in hurriedly, together with some rather fine pots. At least one of this group was a woman, one a child aged about 9, and one a man who seems to have been a soldier, as indicated by the fact that he had survived a serious cut to the head.

The peculiar, contorted positions of the bodies seem to indicate that they were thrown, already dead, into the pit. The man had suffered a broken neck, either before or on entry into the pit, and the front and sides of the skull had received heavy blows. It is impossible to say which of the many broken bones represented the cause of death and which occurred after the individual had already died. Various slash marks on the left parietal bone appear to be healed sword cuts that, apparently, have nothing to do with the cause of death. The damage to the skull was obviously inflicted, caused by a blunt object, a fall, or both.²⁶ The wounds to his head, face, and neck suggest that he died in an attack.

The sand and gravel beds above the remains of the five humans in the shallow pit were disturbed. The excavator remarked that all the bones had been found in an area with a diameter of 2 to 2.5 m. Other disiecta membra were found in the same area, but, since they were either loose on the surface or sticking out of the edge of a modern gully created by rains between excavation seasons, we cannot tell just how many individuals might have been killed in the same conflagration.

The shallow hole seems to have had grave offerings dumped in with the articulated bones, as opposed to the other bone pit nearby, with the bones thrown in helter-skelter, that had none. A number of unusually elegant ceramics and bits of bronze found near and within the shallow hole might have been deposited there for this purpose. Among them were some small pieces of bronze,²⁷ including what was probably a piece of a fibula; a fine Phrygian-shaped round-mouthed jug (HoB 369); other fine pottery, such as a tall base of a riotously decorated white Bichrome pot (HoB 370) that is unparalleled at Sardis, with two kinds of checkerboard patterns, circles, and zigzags; also, a Black on Red stand (HoB 371) and a Black on Red plate (HoB 368). A Bichrome fragment that was found two years earlier, but quite near the skeletons, has a design that is unique among Sardis finds (HoB 333). On the other hand, it is possible that the shallow hole was part of a tidying up operation after hostilities had ceased, and the pottery could have been just part of the debris.

Oven/Furnace and Surrounding Area

West of the possible building in the Central Area were the remains on the Destruction Level of what was apparently an industrial installation in the form of an oven or furnace (Fig. 5.14).²⁸ Close by, in several areas, were substantial patches of burned clay and ash.²⁹ The persistent reddening indicates a freer flow of air and a higher temperature (ca. 600°C?) than is usual in a domestic facility, where there is normally considerable blackening from soot. This oven or furnace was not so obviously fired, nor so substantial, as the late seventh- through mid-sixth-century furnace groups in the sector Pactolus North that have been associated with the parting of gold and silver.³⁰ It was, furthermore, smaller and of a different shape. The fired clay surface showed sufficient differentiation of color and texture to make it clear that the oven or furnace had a raised oval wall around it, certainly forming a retaining curb and possibly a domed roof. Fragments of “tile” mentioned in the fieldbook, described as reddish pieces of partly fired clay in flat, slightly curved forms of no particular shape, may be remnants of walls and roof. The pieces were 0.02 to 0.03 m thick and 0.10 to 0.15 m on a side.

A small three-sided rectangular foundation associated with the oven at W12–14/S98–99.26 was neatly made with stones that were one course high and two stones wide (see Fig. 5.14). The earth was packed hard within it, but showed no signs of the burning that was so obvious in the surrounding area. It measured 2.06 × 1.26 m on the outside although the north wall was preserved for only 1.34 m. The east end was open.

The finds from the oven include pottery compatible with other groups from the same level, useful for chronology but not specialized enough to shed light on activities in the area. The most reliable count of the proportions of the different pottery categories, from a total of about 5 and a half boxes, was: Gray Ware 70 percent; cooking ware 10 percent; painted 10 percent.³¹ The proportions fit well with those recorded in the Central Area and the East End at this level, and a fragment of an East Greek skyphos (HoB 348), dated to the fourth quarter of the eighth century or beginning of the seventh century by Michael Kerschner,³² helps to refine the dating of the Destruction Level. A Lydian Black on Red globular jug (HoB 331), decorated with bands of small concentric circles, has a Phrygian flavor and is remarkably like a piece in Gray Ware (HoB 376; Fig. 5.16) from a cache of pots in the Central Area.

The group of sherds from the oven area is in striking contrast to the pottery from a patchy sandy floor that was not burned.³³ In the one and a half boxes that came from this area, the sherds were more varied and in better condition than was usual at this level. In addition, the excavator reported that “compared to most samples [from around here] Gray Ware is low, painted and cooking are high.”³⁴ There were also seventeen pieces of pithos and two pieces of breadtray—but no mention of intermediate coarse ware. The painted ware included Lydian Geometric types, Bichrome, and Streaked ware, and some pieces over-painted with white bands. The Gray Ware consisted of the usual types: large jars with sloping necks, several different band handles, and a variety of bowls. The cooking pots were of the normal globular shape. Rather unusual was a considerable number of fragments of Buff Ware plates. Only one import was noted.³⁵

Since there were no structural remains here and the sandy floor became rather spotty, it is difficult to reconstruct the relationship between the oven, the rectangular foundation, the Destruction Level, and the sandy patchy area. There appears to be a distribution pattern in the proportions of pottery transitional between that frequent in Lydian III and II. How this came about is unclear, since there is no record of disturbances like pits. A line at ca. S105 was the limit of fierce burning and destruction, and the burning at W10/S113–114 was rather localized.

Three parallel north–south walls in the South Side, although scanty, offer an indication of coherent planning or, at any rate, consistent orientation.³⁶ At the western edge of the trench,³⁷ the top of the east face of a wall at least four meters long was exposed in the scarp (Fig. 5.17, at left edge of plan). Its thickness is unknown but the stones are of a scale comparable to those of the narrow parallel walls located further east at the same general level. This westernmost wall disappeared into the north scarp,³⁸ and it was apparently interrupted in the southwest corner of the trench. Its top can be associated with a level that slopes from *97.4 at the south end (ca. S119) to ca. *97.0 at the north (ca. S115), but we suppose that the interior of the building was to the west, under the scarp, since there are no hints of building debris nearby on its east side.

In two small test pits at the south side of the trench³⁹ we were able to establish the existence of the fiercely burned clay floor of Lydian III (see Fig. 5.1 and Fig. 5.17), thus extending the evidence for the conflagration that was so consistently recorded on the Destruction Level. Catalogued finds other than pottery included a straight bronze pin with a rolled head (HoB 389) and a bronze cosmetic spoon or spatula (HoB 388). Also in the southernmost test pit (near the bottom in Fig. 5.17) was not only a wall of Lydian III,⁴⁰ but also, at the lower level of *96.75, a surprising and rare wall spur of Lydian IV.

The long narrow trench of South Side was filled with the bottoms of bothroi from Lydian II, as well as a number of short scraps of walls of Lydian III (Fig. 5.18). The nearest structural remains are substantial fragments of two parallel walls 2.5 m apart, running on the same alignment as the westernmost wall, both about 1.5 m long (Fig. 5.19).⁴¹ Near both these walls were piles of stones that were probably fallen debris but might have formed a patch of pavement or a working platform such as we see later in Building H and Building J. A line of single stones⁴² that forms a right-angled corner might be the remains of one of the small rectangular storage bins or foundations that become a standard feature in the better-preserved buildings of Lydian II and I (e.g., Buildings G and F). Just to the east of the more southerly wall were a small pit and a hearth with a broken quern set on end (marked on the plan, Fig. 5.17). The floor continues to the east beyond these walls, running northwest to southeast, and rising higher toward the south scarp.⁴³

Another group of three parallel wall fragments (or four if one counts the small scrap at the south scarp) was found nearby to the east,⁴⁴ but the remains had been disturbed and dug into. In one place the circular outline of the base of a bothros that had been dug in from above, from a Lydian II level, showed clearly. These walls are taken to be the long sides of the original structures rather than crosswalls, by extrapolating from probable lines of the buildings to the north and the general lines of the later Lydian II structures, which may have followed traditional practice in determining their layout.

At the southern end of the wall at W8/S115–116, there is a gap of 0.20 m after which the wall continues southward. This gap is surely too close for adjoining houses, even by Lydian standards, but we have not been able to explain why it is there. A group of small stones all on one level (*97.58) makes a neat corner at the west side of the wall. The stones are carefully set, suggesting that they made a small platform that continued uninterrupted even where there was the gap in the wall. This same north–south wall is interrupted on the north side by the base of another bothros of about 0.60 m in diameter, after which the wall appears to continue to the north, but with a narrower width. The alignment of north and south fragments of the wall is such that there can be no doubt of their unity even if there were a doorway or opening to account for some of the gap. As with many of the other walls, there was one fine face and one less carefully made, even when the wall is only one stone thick, as is the northern portion of this one.

About 0.70 m to the east of the north–south wall just described is another narrow wall only one stone wide, and this one is made of smaller stones. It too was interrupted by the east edge of the bothros.⁴⁵ This group of walls in the South Side represents the most substantial of the wall fragments of Lydian III, but even so, most of them are hardly more than one course high and two stones wide. Still further to the east in the South Side is a narrow wall ca. 0.25 m wide that makes a corner extending about one meter west and one meter south.⁴⁶ Its function is unknown.

Overall, the floor level of the walls on the South Side ranged from *97.4 to *97.0, where the pottery from this area was found. It consisted of rather more Gray Ware than in some other parts, but it is not characterized by the increase in cooking or pithos wares that is the mark of the earlier levels. Among the group of standard Lydian wares, including various simple red slip and Lydian Geometric, such as the Black on Red stemmed dish (HoB 379), were two imported wares, including fragments of two East Greek Geometric cups (HoB 382 and HoB 383). The second of these is a predecessor of the later bird bowls. Insofar as it is a recognizable import from a known series, this piece is important for the dating of the area. A wide span has been suggested: Hanfmann, at the time of discovery, thought it should be placed in the first half of the seventh century. Boardman puts similar pieces from Emporio in the range of 690–630,⁴⁷ and Coldstream puts the beginning of the type in Rhodes at 745.⁴⁸ But Michael Kerschner and Nezih Aytaçlar, who examined HoB 383, suggest a date in the second half of the eighth century.⁴⁹ The presence of the Corinthian jug (HoB 355; Fig. 5.2) in the same horizon makes a stronger case for an earlier date, ca. 725 B.C., although the piece does not have to be at the beginning of the series and suggests that Boardman’s range is on the low side.

Among the several metal finds were fragments of a bronze fibula (HoB 387). The circular bosses that would have decorated the bow are preserved, as well as part of the catch plate. This personal item accompanies other bronze pieces found in the Lydian III level of the South Side (HoB 388–HoB 389).

A change occurs among the pottery shapes in Lydian III: suddenly there are more cups, and of a different kind, because of the appropriation of the Corinthian skyphos: for example, a Lydian skyphos with two handles, a rounded body, and a high ring foot was found on the Destruction Level (HoB 329). This shape (or rather a less rounded version of it) eventually replaces the one-handled cup seen so widely in the Gray Ware of the previous period (see on cups and skyphoi in Chapter 1, p. 6). Lydian dishes become more common, and have a different profile than earlier examples. Compare, for instance, HoB 325,⁵⁰ with an earlier example such as HoB 278,⁵¹ which has a much shallower bowl and flat rim (again, see Chapter 1, p. 5). Considerable differences in the storage vessels also may be seen in the Lydian III level. Pithoi, found in such abundance in earlier levels, showed that the locals in those periods had very high storage requirements; there seems to have been less reliance on storage in large jars in Lydian III than in Lydian IV and the Late Bronze Age. Perhaps in earlier periods the Lydians were more dependent on individual household storage, while the people living during the Lydian III era could procure supplies as they needed them, rather than having to keep large stocks in big vessels.

In addition to pottery fragments, iron objects from the area included a small “spear point” with a tang that had been purposely bent over and a large piece of iron that may have been part of a hammer or pick (HoB 364, HoB 365). An iron hook (HoB 366) was found on the floor at *97.4 together with large amounts of pottery.⁵² Several iron arrowheads or points, some corroded together in a cluster,⁵³ were also found in the debris. In excavations elsewhere at Sardis, as opposed to areas by the city wall, arrowheads are rare, and it is significant that so many more were found all in one civic or residential area and associated with a widely fatal conflagration. Traces of reeds or straw could also be observed in the iron corrosion product. Another object from the same “lump” was a riveted furniture fitting of uncertain use but conventionally called a bracket.⁵⁴ There are also traces of wood grain on its interior and clear imprints of straw and grains on its exterior. The same area also produced pieces of two large knives⁵⁵ and another piece of iron, best described as a spear butt that tapers and has a hollow end.⁵⁶

In the Central Area, around and just under the five skeletons that had been thrown in a shallow hole (Figs. 5.11, 5.12, 5.13), the pottery included much reddish pithos and Gray Ware. The painted ware is mostly local, and there is more Bichrome than was observed further south in the trench. We would take this to be accidental, perhaps a result of digging from a later level into an earlier one, since Bichrome is more common once again in the southwest corner, where there was also considerable interruption from the digging of bothroi. The distribution holds up whether the sherds are tabulated by estimated proportion or, as was the case with small lots, actually counted. There were no striking imported pieces of pottery except for a small body sherd from a Greek Geometric krater (HoB 353), found in sand a bit to the south of the stone-lined burial pit (Fig. 5.10).⁵⁷ This find indicates that the revival of Greek imports goes back well into the eighth century. Black on Red predominates among the painted pots. Bowls in this ware and in Gray Ware were particularly noted, and the large white Bichrome stand (HoB 370) and the Black on Red stand with flaring base (HoB 371) are noteworthy.

Several good examples of pottery typical of this phase were retrieved from the heavy gravel that overlay the Destruction Level. They can be confidently associated with the level in general, although their original context is lost. Here and there in Lydian III a special function might be assigned to a particular area, such as where the proportion of cooking pots and animal bones rises far beyond the usual numbers, and a kitchen might be inferred. But in general the interrupted nature of the level defies anything like certainty. A typical box at this level contains much plain ware and Gray Ware, but also a consistent mixture of utilitarian ware, finer decorated ware, and imports. Since we are dealing with areas of gravel with mixed contents as well as areas of disintegrated mudbrick or floors maintaining their original (i.e., destroyed) definition, we must approximate the succession of pottery wares. It is clear, however, that this Destruction Level coincides with the highest frequency of painted Lydian pottery in the Black on Red and Brown on Buff styles. The association with Greek imports of Geometric style can give us some help with the chronology.

The Bichrome style, which relies upon different combinations of geometric motifs for its decoration, has not yet reached its peak. This is clear for the plates and stemmed dishes that grow in popularity rather later, at the same time as the general introduction of the East Greek Wild Goat style. No piece of Wild Goat or any imitation of it has been identified in the pottery found in the Destruction Level.

In the past, as already explained, this level had been associated with the second (and successful) attack of the Kimmerians in ca. 652 B.C. A reappraisal of the imported Greek pottery in this level at Sardis, however, has led to the present conclusion that the destruction should be dated to the last part of the eighth century B.C. Among the imported ceramics belonging to this level is the Late Geometric Corinthian jug (HoB 355; Fig. 5.2),⁵⁸ found over several years in many pieces. The combination of shape and decoration on this pot has a range between 750 and 700 B.C. We cannot just pick an arbitrary date within that range because it must fit with other Greek imports whose ranges do not precisely match that of the Late Geometric jug. For example, the East Greek paneled krater on a high foot (HoB 351) might be described as sub-Geometric and its range can spread from the later eighth, continuing into the beginning of the seventh century.⁵⁹ The same is true for the cup fragments that belong to the bird skyphos class (HoB 347).⁶⁰ By balancing the various ranges in which any of the pots might have been made, and the understanding that they were all broken at one specific time, we believe that the last quarter of the eighth century is the best estimate for the date of this destruction.

Whether this destruction should still be regarded as the result of an unrecorded attack of the Kimmerians, or the result of some local disturbance arising from traditional jealousies within the Lydian aristocracy, or even an attack by another foe altogether, we cannot tell from the archaeological record. The Kimmerians are leading contenders as the enemy, since they are well attested as ranging over Asia Minor and attacking established cities with considerable success in the late eighth and early seventh centuries B.C. What might also be inferred here to lend weight to the theory of outside attack is the fact that only 100 meters to the east of HoB, a huge fortification wall was erected sometime in the seventh century; this was built on the foundation of a yet earlier wall.⁶¹

A civil war among the Lydians cannot be ruled out entirely, however, since the circumstances of Gyges’ coming to power are obscured in romantic details.⁶² One interpretation of the story might be a series of “palace coups” in an unstable era, in the generations before Gyges, perhaps centering around a passionate intrigue as at Rome during imperial times or at the Byzantine court. If this were the case, it would probably have been settled within the palace circle, with minimal effect upon the commercial and domestic affairs of ordinary Lydians. If, however, this is only the excuse or final act of substantial and traditional hostility between factions of the Lydian nobility, then it is possible that an army marched from elsewhere in Lydian domains (Daskyleion or Adramyttion?) and forced the issue.

To summarize: The Destruction Level of Lydian III dates to the last quarter of the eighth century B.C. The traditional interpretation of the historical sources puts the successful Kimmerian attack on Sardis that resulted in the capture of the citadel in the mid-seventh century, at a period that does not fit the chronological horizon proposed for the Destruction Level as recorded in HoB. As will become clear in the next chapter, there is no sign of general fiery destruction within the next level above it. We conclude, as other excavations have, that if the Kimmerians captured Sardis, we have not yet discovered any archaeological traces. The Lydian III Destruction Level seems to have included significant human death by violence, but the identification of the aggressor(s) remains uncertain.

• 1The open siting of the buildings of Lydian III is somewhat at odds with the close urban pattern of the eighth century revealed in the excavations at Old Smyrna, Zagora on the island of Andros, and Lathuresa in Attica, where defense seems to have been an important consideration. See on Old Smyrna: Cook 1958; Akurgal 1983, pp. 22–35, fig. 13f. Zagora: Cambitoglu, Coulton et al. 1971, pp. 13–36. Lathuresa: Lauter 1985. For a general overview, cf. Stampolidis, Maner, and Kopanias 2015.
• 2Hanfmann, “Sardis 1966,” p. 33, n. 5.
• 3First proposed in Hanfmann, “Sardis 1960,” p. 12, n. 8, with bibliography; Hanfmann and Mierse, SPRT, pp. 28–29; see Sardis M10, p. 7. See also Kristensen 1988, with review by P. Zimansky (1994); Sauter 2000; Ivantchik 2001a; Bouzek 2007. On Gyges’ correspondence with Assurbanipal of Assyria for aid against these raiders, see Cogan and Tadmor 1977 and Spalinger 1978. At the time, the Early Phrygian destruction level at Gordion was also attributed to the Kimmerian attack; it has now been redated to a century earlier (Rose and Darbyshire 2011).
• 4Or perhaps a little more, because we have not included the area between W20–30/S120–125 that was not taken much deeper than the Hellenistic level in this part, at *100.0.
• 5At level *97.4.
• 6Going to the east from E0, it rises almost to *98.0 around E3/S104.
• 7At W5.
• 8G. F. Swift, Jr., in Hanfmann, “Sardis 1965,” pp. 10–11.
• 9Rose 2013b, p. 6.
• 10Sams 1994, pp. 2–7, esp. p. 3.
• 11Also HoB 279, HoB 280, HoB 281.
• 12See Ramage, Sardis M5, pp. 4–5 for additional details of Lydian walls.
• 13Sardis M10, pp. 17, 19, cat. Cor 1, pl. 4. Hanfmann, “Sardis 1965,” p. 10; Hanfmann and Mierse, SPRT, p. 29, fig. 38.
• 14HoB 353A. A piece from E4–6/S109.5–111 *97.74–97.64 was found by A. Ramage to join another from W9/S104–109 *97.40–97.00.
• 15The floor was found below another patchy floor that was associated with some building remains from a later floor above it, at *98.2; see Chapter 6 on Lydian II.
• 16At E5/S95.
• 17Joel S. Savishinsky, “1966 Grave and Human Skeletal Remains” (Sardis Expedition field report, 1966), p. 3.
• 18Melis Koruyucu, “2016 Anthropological Report of Sardis Bones from Ankara DTCF” (Sardis Expedition field report, 2016).
• 19According to Savishinsky, “1966 Grave and Human Skeletal Remains.” See Hanfmann, “Sardis 1966,” p. 33, fig. 5, and Hanfmann and Mierse, SPRT, p. 27, fig. 34.
• 20Those remains included the mandible of a male, aged 19–20, of a female, aged about 8, and the radius of a male perhaps in his 20s, according to the account and bone report by anthropologist Joel Savishinsky. Hanfmann, “Sardis 1966,” p. 33, n. 5, and Hanfmann and Mierse, SPRT, p. 27, fig. 34.
• 21G. F. Swift, “House of Bronzes: Lydian Trench: Midseason 1966” (Sardis Expedition field report, 1966), p. 2. It was located at W4–6/S86–87 *96.9–96.3.
• 22Fieldbook HoB 1966.i:119, 143, 125, 153, 159; Fieldbook Bones 1966.ii:17–39. Hanfmann, “Sardis 1966,” p. 33, n. 5.
• 23Sardis M10, cat. Cor 2, pp. 19–20, pl. 4 (“from earth with bones”). This piece could have been in use at the same time as the Late Geometric jug that came from the destruction floor, HoB 355.
• 24Joel S. Savishinsky, “Human and Non-Human Bones, Mid-Campaign Report” (Sardis Expedition field report, 1966), pp. 5–9; Joel S. Savishinsky, “1966 Grave and Human Skeletal Remains” (Sardis Expedition field report, 1966), p. 3.
• 25W6.3–8.4/S90–93 *96.2.
• 26David Finkel, bone report 1968 (Sardis Expedition field report, 1968).
• 27Fieldbook HoB 1968.ii:13: “Bronze – 5 bits, incl. 1 pc. of fibula? 1 pc. Sheet.”
• 28W12–13/S87–88 *97.0.
• 29Sidney M. Goldstein, archaeologist and conservator, wrote a report on several trial trenches and tests conducted on the recovered material. He thought it was impossible to draw firm conclusions about the purpose of these structures, but favored the idea of an oven beside a storage bin in the courtyard of a house.
• 30Hanfmann and Mierse, SPRT, pp. 34–41; Ramage and Craddock, Sardis M11, pp. 83–87.
• 31Pithos fragments and coarseware sherds are sometimes taken together, and cooking and coarse ware share some characteristics, so pithos and coarseware sherds have not been included in the specific proportions here.
• 32Personal communication at Sardis, July 1–2, 2016. Cf. Coldstream, GGP, p. 277, pl. 61d, his Bird-Kotyle Group.
• 33Nearby at W14–18/S103–108 *97.00–96.80.
• 34Gray 40 percent, painted 30 percent, plain buff 15 percent, cooking 15 percent.
• 35Fieldbook HoB 1968.iii:17: “1 pc Greek amphora or hydria; pale gray-buff surface, black bands somewhat streaky” and iii:19: “1 pc Greek pale buff jar.” Not catalogued.
• 36W5–33/S110–119.
• 37At level *97.0.
• 38The wall runs from W33–34/S114.5–118.4.
• 39At W21–25/S118–119 and at W16–20/S110–111 *96.40. Only the more southerly one is shown on the plan in Fig. 5.1.
• 40Top of wall at *97.70.
• 41One at W22.5/S112.5–114, and the other at W20.5/S115–116.5.
• 42At W22/S111.5 *97.13.
• 43To *97.5.
• 44Between W5 and W12.
• 45It ran from S112.5–116.5 at W7.5.
• 46At ca. W5/S114. Its top is at *97.71, and the floor in the area is at *97.55.
• 47Boardman 1967, p. 192 and closed shapes 547, 551, 554. These are said to be current in his periods II and III, which includes most of the seventh century.
• 48Coldstream, GGP, pp. 277–79, the Bird-Kotyle Workshop; another example, from the Central Area, is HoB 348.
• 49Personal communication, Sardis 2016.
• 50Also HoB 374 and HoB 379.
• 51Also PC 26 and PC 27.
• 52HoB 328, HoB 336, part of HoB 338, HoB 342, HoB 345, HoB 351, HoB 356.
• 53HoB 366A. Compare Cahill 2010c, p. 352. Also Greenewalt and Rautman, “Sardis 1994 and 1995,” pp. 490–93.
• 54Several of these “brackets” were found with many other pieces of iron in the debris of Lydian houses side of the Lydian city wall: Greenewalt, Rautman, and Cahill, “Sardis 1985,” p. 64.
• 55The first with dimensions of 0.095 × 0.065 m, a piece now lost; max. W. 0.02. The second, M17.5, found at E5/S95–100 *97.6, was ca. 0.20 m long, 0.015 wide, and more than 0.003 thick.
• 56It measures 0.13 × 0.035 m and the diameter of the socket is ca. 0.022.
• 57Sherd found at W4/S94 *96.4.
• 58Sardis M10, pp. 5, 7, and cat. Cor 1, p. 19, pl. 4; Hanfmann and Mierse, SPRT, p. 29 and fig. 38.
• 59Boardman 1967, p. 106, fig. 62.
• 60See Coldstream, GGP; Boardman 1967, p. 134, no. 440.
• 61Attempts to clarify this point in the excavation seasons of 2009 and 2010 produced no answer. See Cahill and Kroll 2005; Cahill, “Sardis 2009”; Cahill, “Sardis 2010.”
• 62Herodotus 1.8ff. and a different but more political version in Nicolas of Damascus, FGrHist 90 F 45 (Sardis M2, no. 35 with comments).