Search

Anthemion stelae of the Archaic and Early Classical periods-tall gravestones crowred by volute-palmette designs-are found at many sites in Greece and western Asia Minor. This article presents five previously unpublished fragments from Sardis in Lydia and two gravestones discovered in 1912 but never thoroughly described, bringing to at least 10 the total number of such stelae found at Sardis. Most of these are only fragments, and none bears any trace of figural decoration. The anthemion stelae from Sardis nevertheless include some objects of superlative craftsmanship, and they help to illuminate relations between Lydia and neighboring regions in the sixth and fifth centuries B.C.* In 1912, the first American expedition to Sardis, directed by H.C. Butler of Princeton University, discovered two limestone anthemion stelae still standing in front of a rock-cut tomb (Tomb 813) in the Great Necropolis on the west bank of the Pactolus stream.' To date, these stelae are the only two such gravestones ever found in situ at Sardis. Since 1912, however, stray or reused fragments of the finials of at least eight more anthemion stelae have come to light, making Sardis an unusually rich source of these monuments.2 In addition to the limestone stelae from Tomb 813, the first Sardis Expedition also found a single fragment of another anthemion stele, made of marble. The provenience of this fragment is uncertain; it was published in the volume devoted to the Sanctuary of Artemis just across the Pactolus stream from the Great Necropolis.3 Between 1958 and 1975, the second and ongoing Sardis Expedition, sponsored by Harvard and Cornell Universities, recovered fragments of another four sixthor fifthcentury anthemion stelae, all of marble.4 One of these was found reused in a Roman wall in the Bath/Gymnasium complex on the edge of the Hermus River plain. The other three were stray finds from the valley of the Pactolus, two discovered near well-known clusters of Lydian rock-cut tombs, and a third found during the construction of the excavation house next to the Sanctuary of Artemis. Since 1975, fragments of three or four more marble anthemion stelae have been recovered, one, whose identification is uncertain, from a field east of the Bath/Gymnasium complex, two from the excavation of a group of Late Roman houses south of the same complex (across the Izmir-Ankara highway), and a fourth from new excavations in the compound of the Sardis Expedition's excavation house. The function-as gravestones-of the stelae discovered in the various cemeteries of Sardis is clear.5 It is conceivable that the fragments of anthemion finials found near the Sanctuary of Artemis (for example, at the site of the current excavation house) crowned votive stelae, but as this sanctuary lies just across the Pactolus stream from the Great Necropolis, and lesser cemeteries are located on the same (west) side of the stream both to the north and to the south, it is likely that these too are fragments of * This article is based on research conducted under the auspices of the Sardis Expedition, and I am grateful to the Expedition's field director, C.H. Greenewalt, jr., both for permission to work on this subject, and for his advice and encouragement. My thanks are also due to Catherine S. Alexander, who made the reconstruction drawings of the stelae published here, and to R.R.R. Smith and the anonymous AJA reviewers, all of whom made many helpful criticisms and suggestions. The following abbreviations are used in this article: SardisRep 2 G.M.A. Hanfmann and N.H. Ramage, Sculpture from Sardis (SardisRep 2, Cambridge, Mass. 1978). Samos XI B. Freyer-Schauenberg, Sanmos XI: Bildwerke der archaischen Zeit und des strengen Stils (Bonn 1974). Published anew as 3a and 3b below; see bibliographical references listed in catalogue entry. The Great Necropolis is a cemetery containing over 1,000 tombs. 2 I exclude from consideration three anthemion stelae of late fifthor fourth-century date, published in SardisRep 2 as nos. 240-42. For the intriguing but unsubstantiated suggestion that the Borgia stele, reportedly discovered in situ in front of a tomb, was found at Sardis, see SardisRep 2, 23, n. 86. H.C. Butler, Sardis II: Architecture, Pt. 1: The Temple of Artemis (Leiden 1925) 80, ill. 93. The fragment was published in this volume because of its resemblance to pieces of architectural ornament found in the Sanctuary of Artemis; its provenience is not specified, however, and it is quite possible it was found somewhere else. Published anew as SardisRep 2, no. 48. 4 SardisRep 2, 23-28, 73-76, nos. 45-46, 49. The fourth, unpublished until now, is our fragment 4. 5 On the cemeteries of Lydian Sardis in general, see B.K. McLauchlin, Lydian Graves and Burial Customs (Diss. Univ. of California, Berkeley 1985) 11-78. 593 American Journal of Archaeology 98 (1994) 593-607 This content downloaded from 207.46.13.149 on Wed, 28 Sep 2016 04:44:23 UTC All use subject to http://about.jstor.org/terms 594 CHRISTOPHER RATTE [AJA 98 Fig. 1. Anthemion stele fragments: a) no. 4 (left); b) no. 1, side A (upper right); and c) no. 3a, fragment of finial (lower right) gravestones, reused in one way or another in the sanctuary.6 Of the fragments found in the vicinity of the Roman Bath/Gymnasium complex, one bears a Lydian inscription that identifies it as a gravestone.7 The location of the cemetery where this gravestone originally stood is uncertain. There is evidence or at least reason for arguing for the presence of cemeteries on almost every side of the Bath/Gymnasium complex, and a small fragment such as this one could in any case have traveled a considerable distance.8 While the other fragments from this area are probably fragments of gravestones as well, an Archaic and Classical sanctuary of Cybele may have been located nearby, and the possibility that these are fragments of v tiv stelae cannot be entirely exclud d, although none of he identifiably votive stel e found at Sardis is of the anthemion type.9 Although the stel e from Tomb 813 are the only Lydian anthemion stelae ever found in context, it is likely that most or all of the other known fragments were also associat d with rock-cut tombs, as opposed to umuli.10 Most of these other fragments come from the Pactolus valley, an area where few tumuli are known, but which is in Butler's phrase "honeycombed" with rock-cut tombs; and the few grave markers that have been found in direct association with tumuli are of a different form, knob-like "phalloi" instead of stelae." 6 In western Asia Minor and East Greece, anthemion stelae are definitely attested, by provenience or by inscriptions, as gravestones both at Sardis (the stelae from Tomb 813 and the inscribed example discussed below) and elsewhere, e.g., Samos (Samos XI, nos. 89-90, with inscriptions; nos. 89-95, 99, from the "west necropolis"). SardisRep 2, no. 49; here fig. 14c. s To the south lies the acropolis of Sardis, whose west face contains many small cemeteries; to the east and west lies the modern Izmir-Ankara highway, which apparently follows the route of the ancient road through Sardis. Nearby is a monumental structure that appears to be the Archaic gate of the city, and a cemetery just outside it would not be surprising. At the time of writing, however, it is still uncertain whether the Archaic city lay on the east or west side of the gate. See C.H. Greenewalt, jr., et al., "The Sardis Campaigns of 1988 and 1989," AASOR 51 (1994) 14-26, for the most recent report on the excavation of this monument. 9 See G.M.A. Hanfmann, Sardis from Prehistoric to Roman Times (Cambridge, Mass. 1983) 74. 10 It is also possible that some of these st lae were associated with ore modest types of tombs, uch as cist graves or sarc phagus burials; on Lydian tomb types, see McLauchlin (supra n. 5) 11-13. 1 Se generally McLauchlin (supra n. 5) 126-39. Most famous is the huge marker still extant on the summit of the tomb of Alyattes; see C.H. Greenewalt, jr., et al., "The Sardis Campaign of 1983," BASOR Suppl. 24 (1986) 2022. In 1993, a "phallus marker" was disc vered (by Rose Lou and Ugur Bengisu, unpublished) on top of a tumulus tomb, probably of the Persian p riod, in the Cayster River valley. The context of the "stele and memorial of Tos" mentioned by Hipponax is unknown; see E. Degani ed., Hipponactis testimonia et fragmenta (Leipzig 1983) 30-32, fr. 7; J. Pedley, Ancient Literary Sources on Sardis (SardisMon 2, Cambridge, Mass. 1972) 77, no. 280; McLauchlin (supra n. 5) 45-48. This content downloaded from 207.46.13.149 on Wed, 28 Sep 2016 04:44:23 UTC All use subject to http://about.jstor.org/terms 1994] ANTHEMION STELAE FROM SARDIS 595

The distribution of mainland pottery in the Cyclades during LC I, together with information about the relative abundance and degree of variety in which it occurs at different Cycladic sites, suggests a distinct pattern of attenuation with distance as one progresses from N to S down the "Western String" of islands. Despite the limited evidence so far available, it is possible to narrow considerably the range of exchange mechanisms likely to have produced such a pattern; it is proposed here that it may have been middlemen traders who introduced nonlocal ceramic products into the Cyclades at this time. Goods can be exchanged in various ways and it seems clear that, under favorable circumstances, different kinds of exchange will leave distinctive patterns in the archaeological record. The purpose of this article is to consider the distributional pattern within the Cyclades of pottery manufactured on the Greek mainland during the Late Helladic I period (contemporary with Late Cycladic I in the islands and with Late Minoan IA in Crete). Although our evidence is still far from complete, we feel it is at present worthwhile to attempt to narrow the range of mechanisms or modes of exchange to those which seem most likely to have been responsible for the geographical distribution and variety of mainland pottery discovered thus far in the Cyclades. The realistic assessment of prehistoric exchange relations between the mainland and the islands is a necessary preliminary to any reconstruction of the political relationships between these two areas. THE CYCLADES AND THE GREEK MAINLAND IN LC I Much mainland pottery which is clearly of LH I date has been found at Ayia Irini on Keos, Phylakopi on Melos, and Akrotiri on Thera (ill. 1); all three sites on present evidence were apparently not only the dominant settlements on their respective islands, but also the principal links in a major trade route between Crete and the mainland.' There is no evidence, however, that settlements elsewhere in the Cyclades had ready access to mainland products in LC I. There are, moreover, few indications that in LC I (or, for that matter, in LC II) the central and eastern Cyclades were in contact with Crete, Ayia Irini, Phylakopi or Akrotiri.2 One might immediately object-and with some justification-that the islands of the central and eastern Cyclades (with the exception, perhaps, of Pa* This article is based on a paper delivered at the 82nd Annual Meeting of the Archaeological Institute of America, New Orleans, December 1980; the authors are grateful for funds from Cambridge University and from the University of Illinois at Chicago Circle, which made possible their attendance at this meeting. Davis's research on pottery of LC I date at Ayia Irini on Keos and Phylakopi on Melos has been made possible, in part, by a grant from the excavation fund of the Department of Classics, University of Cincinnati, by a Summer Stipend from the National Endowment for the Humanities, by a Grant-in-Aid from the American Philosophical Society, and by a Faculty Summer Fellowship from the Research Board of the University of Illinois at Chicago Circle; Cherry's, by a Small Grant in the Humanities from the British Academy. Special thanks are due to Mr. C.B.H. Macfarlane for his part in the study of pottery from Phylakopi and to Dr. Elizabeth Schofield and Miss Mariza Marthari for information and comments. I For LH I style pottery at Ayia Irini, Phylakopi and Akrotiri, see O.T.P.K. Dickinson, "The Definition of Late Helladic I," BSA 69 (1974) 8 and n. 56; for other imported pottery of LH I date at these three sites see infra. Cretan goods and influences on Keos, Melos, and Thera are the subject of two recent papers: E. Schofield, "The Western Cyclades and Crete: 'A Special Relationship,'" (Paper presented to the Oxford Aegean Seminar, 4th February 1981; forthcoming in the Oxford Journal of Archaeology) and J.L. Davis, "Minos and Dexithea: Crete and the Cyclades in the Later Bronze Age," in J.L. Davis and J.F. Cherry eds., Papers in Cycladic Prehistory (Monograph 14, Institute of Archaeology, UCLA, Los Angeles 1979) 143-57. 2 F r LM I types from Delos and LH II from Grotta on Naxos, see references in R. Hope Simpson and O.T.P.K. Dickinson, A Gazetteer of Aegean Civilisation (SIMA 52, G6teborg 1979). A base from a panelled cup found on Delos (BCH 71-72 [1947-1948] pl. 46a) is likely to be a Melian import and could be as late as LC I. Sherds from Paroikia on Paros (AthMitt 42 [1917] 66 figs. 72, 73) seem to be in the Polychrome Mattpainted "Mainland" style (as K. Scholes notes, BSA 51 [1956] 22) and should be LH I imports (see Davis [supra n. 1] 256-58). From the same site comes a sherd from an LM IB bell-cup (AthMitt 42 [1917] 71 fig. 79, no. 2; identified by E.M. Schofield), and possibly sherds in the Melian Naturalistic style (mentioned by Scholes, BSA 51 [1956] 20). A jug from Amorgos (BSA 51 [1956] 20) could also be as late as LC I. A sherd from the Northwest Promontory on Antiparos may be LC I/LM IA in date (see Hope Simpson and Dickinson, Gazetteer 324). The only other Minoan material in the Cyclades is earlier in date: MM sherds on the surface at Vigla on Naxos (BSA 69 [1974] 50) and cups of MM III types from a cist at A'la, also on Naxos (Hope Simpson and Dickinson, Gazetteer 331). For Cretan stone vases from Delos and Naxos, see P.M. Warren, "Minoan Stone Vases as Evidence for Minoan Foreign Connexions in the Aegean Late Bronze Age," PPS 33 (1967) 37-56.

The excavation of a Classical cemetery at Roccagloriosa in Lucania has provided evidence for some unusual burials with funerary pyres. The practice of cremation is discussed in connection with representations on vases, and heroic overtones are suggested. The available archaeological evidence indicates that cremation was not a major burial practice among the Italic populations of Southern Italy,' and it certainly was of secondary importance in the Greek colonies along the coast of Magna Graecia in the Archaic and Classical periods.2 The presence, however, of ustrina (the charred remains of pyres, sometimes with burial in situ or, more often, with secondary cremation burials) in such areas as Samnitic Capua,3 Lucanian Paestum4 and the hinterland of Metaponto5 points to the growing popularity of cremation during the fourth century B.C. It is, therefore, of special interest to examine the evidence for a peculiar kind of cremation adopted in the course of the second half of the fourth century B.C. at Roccagloriosa in western Lucania (ca. 30 miles SE of Paestum).6 The La Scala cemetery at Roccagloriosa, at least the part which has so far been explored, was situated in a natural saddle of the Capitenali crest and limited to the east by a precipitous drop (pl. 59, fig. 1). The scattered trench graves covered with tile roofs occupied a wide, flat area. At the north and south ends of the central area, near the rock defining the "saddle," two walls of limestone blocks separated the two extremities of the "saddle" from the rest of the necropolis (ill. 1). The enclosure at the northern end, formed by a line of squared blocks set in two courses, contained the remains of two large pyres, T. 23 and T. 25 (pl. 59, fig. 2); the central part of each was covered by a cist of limestone blocks (ill. 2). The cremation in T. 25 (set just in front of chamber tomb 24, with some blocks of the latter still bearing traces of burning) may have included a funeral bed or wooden coffin decorated with gilded terracottas of Tarentine type (pl. 59, figs. 3-4).' The offerings, set on the pyre or in the area around it, included a variety of vases and a group of metal objects among which are six iron spits and an iron fire-dog (pl. 59, fig. 5). A lekane and black glaze salt cellars date the burial to the third quarter of the fourth century B.C. About a meter west of T. 25, remains of a similar, slightly larger burial (T. 23; ca. 3.5 x 4.0 m.) may be * A version of this article was presented at the 82nd General Meeting of the Archaeological Institute of America in New Orleans, December 1980. The following abbreviations are used: Antiche Civilta Lucane Atti del Convegno di Studi, Oppido Lucano, 5-8 Aprile 1970 (Galatina 1975). Atti Taranto (followed by Atti del Convegno di Studi sulla serial number and year) Magna Grecia (Taranto). EVP J.D. Beazley, Etruscan Vase Painting (Oxford 1946). SFor a general discussion of the problem, see G. Davies, "Burials in Italy up to Augustus," in R. Reece ed., Burial in the Roman World (CBA Report 22, London 1977) 13-15. Also, DarSag s.v. Funus, 1366-1409. 2 Good discussions of the typology of cremation burials found in the cemetery of a Greek colony in South Italy are in P. Orsi, "Locri Epizephirii-Necropoli Lucifero," NSc 1911, Suppl., 3-5; NSc 1913, Suppl., 47; NSc 1917, 160; "Megara Hyblaea-La Necropoli," MonAnt 1 (1889) 773-74. Also instructive in showing a general increase in cremation burials and ustrina during the late fourth-third century is the preliminary report on the Hellenistic necropolis at Reggio Calabria: E. Galli, NSc 1942, 166-234 (see his fig. 3 for a plan of the excavated burials). The growing use of cremation in the third century is well documented in various areas of Central and South Italy: for the Esquiline cemetery in Rome and some burials at Palestrina and Tuscolo, see Roma Medio Repubblicana Aspetti Culturali di Roma e del Lazio nei secoli IV e III a.C. (catalogue of the exhibition, Rome 1973) 188-96 (Esquiline), 270-72 (Palestrina), 305-307 (Tuscolo). 3 I am indebted to W. Johannowsky for this information. 4 A. Greco Pontrandolfo, DialAr n.s. 2 (1979) 46. 1 J. Coleman Carter, personal communication. To be considered, although strictly speaking outside the area of Magna Graecia, is the evidence from the hinterland of Gela reported by P. Orlandini, ArchCl 20 (1968) 152-54, with statistical table showing the relative frequency of the two rituals. 6 A preliminary report on the 1976-77 excavations is in NSc 22 (1978) 383-421; see also the preliminary report of the 1978 season, including a general description and plan of T. 23 and T. 25, NSc 24 (1980) 103-109. For some general comments on the cremation burials, see Expedition 22 (Spring 1980) 41-42. For a general description of the site, see D. Ridgway, JHS-AR 1981-82 (forthcoming). 7 R. Lullies, Vergoldete Terrakotta Appliken aus Tarent (RbmMitt EH 17, 1962) pl. 3; considering their quality, the Roccagloriosa pieces may well be imports directly from Taras. For finds of similar terracottas in funerary contexts, see R. Bartoccini, NSc 1936, 131 (from Taranto); Greco Pontrandolfo (supra n. 4) 48-49 (from Paestum).

thickness from 0.I2 m. at the top edge to o0.4 m. at the bottom. The cist was composed of two pieces very closely fitted together and measuring I.i8 m. and 0.43 m. in length respectively. The grave lay, at an angle approximately NWSE, at a depth less than half a meter below the modern surface and completely sunk into the soft greenish rock called "pasparos" which is native to the site. A trench wider than the sarcophagus had been cut to receive it and the space remaining when it was in position had been filled with small fragments of the rock which had been dug out. A very well preserved skeleton of an adult was found in the tomb together with ten small pouring vessels and a small amphora, all of the geometric period. There were also some metal objects. These finds will be described and discussed in this article. For the first time the existence of a larnax tomb of the geometric period cannot be disputed. The only other known instance is that of a tomb at Corinth2 (AJA 9 [19051 413, fig. i) where attribution to the geometric period is highly probable but not absolutely certain since early geometric vases were found very near the grave but not actually in it (ibid. pls. XI-XVI; Saul S. Weinberg, Corinth VII, pt. I, Cambridge 1943, i6, nos. 54-66). The other contemporary parallel is the "aschenkiste" found in a rock tomb on Thera. Note especially a vase illustrated by H. Dragendorff (Thera II, Berlin 1903, 28, fig. 74) which came from a tomb with geometric and protoarchaic burials. See also ibid. 56, fig. 190. Two more geometric vases found very near the tomb and two Corinthian vases found in a field not far from the village were given to me when I visited the site. These will be described and discussed together with the rest of the finds from Clenia.

PERIOD V. The transition from Lerna IV to Lerna V occurred without a serious break in continuity. Apsidal house plans persisted, often appearing directly above the place of an older building. Gray Minyan ware was made in canonical Middle Helladic shapes, the kantharos and the ring-stemmed goblet. Some stone implements are like those of Period IV. Innovations include dark Argive and light yellow Minyan wares, Matt-painted pots, and a class of vessels with decoration in lustrous paint (not well known before Lerna was excavated). Intramural burials in cist graves became very common suddenly. Bones of horses and domestic fowl are now found for the first time. Much wider foreign contacts are attested by the appearance of small handmade flasks, possibly from the Balkans, and imported Middle Minoan IA pottery in the first phase of Lerna V. As the period progresses, with many successive building levels, there are examples of MM II and MM III ceramic styles and corresponding types of Cycladic pottery. The latest Middle Helladic strata have been lost by erosion in most parts of the site, but there is no reason to doubt that the transition to the Mycenaean period was gradual, as elsewhere in the Argolid. Two large shaft graves, much like those of Circle B at Mycenae, were found to have been emptied and refilled in antiquity. In the filling were great quantities of potsherds reflecting late MH styles and those of LH I. We call this period Lerna VI.