in

MARIE-HENRIETTE GATES

This year's report, on archaeological activities A?ikli, in now in their seventh season, and in the Turkey (fig. 1) during the 1994 excavation season and Plain at Qatalh6yiik and related Pinarbali, reopened the 1995 findings that have made their way into on pub-an ambitious scale in 1993. These multidiscipli- lic view, illustrates the extent to which long-term nary projects at lastjoin the long-term investigations projects can become invigorated by new directions, around Burdur. They will encourage others to enter and old issues can be revitalized by new discoveries.* this important field, where questions can as yet be While such developments could be seen as indepen- framed only in the most elementary terms (where, dent trajectories, in the case of Anatolian research for example, is the Cappadocian of the PPNA they are certainly encouraged by the annual meet-obsidian network?). On the Aegean coast, the sec- ings of the Turkish Ministry of Culture's General ond millenniumDi- B.C. is also receiving renewed atten- rectorate of Monuments and Museums. These schol- tion. Together with , Liman Tepe, and Panaztepe, arly forums serve not only as an opportunity for thethe remarkable Minoan finds from the new investi- directors of each project to present the season's gations at Miletos and the Mycenaean level en- findings. Far more significantly, they offer archae- countered in 1995 below the Ephesos Artemision ologists the chance to perceive and,assess their own promise a future reevaluation in perspectives on both research within the broader context of the region sides of the Aegean. and cultural period. Finally, the need to compress The neighboring Hittite kingdom also continues a season's results into a 15-minute presentation may to achieve a sharper focus. In a pattern set by the also be responsible for the precise articulation past of few years, two new excavations started up at Hit- the well-defined research goals that now character- tite sites in 1994: at Kinik, source of the Kastamonu ize the majority of these projects. Archaeology metal in hoard, and at Kilise Tepe in the G6ksu valley, Turkey is proceeding on all fronts toward the end on of the route between the plateau and the "Lower the century with an energy inspired and coordinated, Lands." At Ortak6y and Kugakli, researchers recov- to no small degree, by these meetings. Indeed, the ered more local color, together with their ancient large-scale projects that are celebrating centennials place-names, and Sarissa, over the course this decade-Ephesos in 1995, Miletos in 1999- of the year. These two projects seem fated to pro- illustrate these trends toward wider inquiries into ceed in tandem despite their geographic distance: the environmental and cultural background for their their Storm-Gods were invoked in immediate se- urban centers, and are providing excellent but by quence in, for instance, Suppiluliuma's treaty with no means unique models. Specialists in all periods Kurtiwaza (KBo 1.1- the first tablet in the Keilschrift- are proving themselves responsive to the challenge texte aus Boghazk6i series, published 80 years ago). To of expanding the scope of their investigations. these can be added another new city, Tikunani, men- In this light, several cultural areas are reemerging tioned in a recently published letter from the Great from a long sleep as dynamic centers for research. King Hattu'ili I to his vassal, its king Tunija/Tunip- central , for example, long iden- Tel'up (M. Salvini in SMEA 34 [1994] 61-80). This tified with a few isolated and independent sites, can letter would represent the earliest-known document be expected to take on a decidedly fuller personal- attributable to the first named Hittite king. Accord- ity in the coming years thanks to the excavations at ing to Salvini's reading, the tablet was written in

* This newsletter was written in large part from notes cially helped with Byzantine and bibliographical matters. taken at the 17th Annual Archaeological Symposium C.W.in Gates attended most of the survey sessions, and it is (29 May-2 June 1995), organized by the Turkish because of his careful notes that the survey projects can Ministry of Culture's General Directorate of Monuments be summarized here. It is a pleasure to express my warm- and Museums. I am very grateful to the speakers, to the est thanks to all of them, as well as to Tracey Cullen and many colleagues who sent me summaries and photographs, the AJA office for their encouragement and patient sup- and to Scott Redford, who wrote up the separate entries port. I am also grateful to Danielle Newland for revising for the Islamic sites (included here in their own section the map of Turkey (fig. 1), originally drawn by Liesbeth for the first time). Among my Bilkent colleagues, all of whom Wenzel. provided assistance, A. Ricci, N. Karg, and I. Ozgen espe-

American Journal of Archaeology 100 (1996) 277-335 277 - rSINOP Su PiAnar/A Demirc S Kanligegit* .KASTAMONU ikiztepeK . a EDIRNE aKaracakbyr Pontica Gazkman Te Silivri TRABZO ArsAraanep GUMOSHA AiTnos cus.B ao emBAYBURT ,, SnhoBGA .BURDURandir sKiRkLARELI Byktepe n SosHykKes..o L u BRSA Alaca. *ERZURUM *Troy Daskyleion Payalar ESKi$EHiR ANKARA .Bo?azk y DemircihajySk Sayosk .KARAKerkenes Da K SmintheioDn Seyda Ln merS. Bahggehisar. *rkeHY u anvurkalese?ordionaAliKar K me zano KAKHISAR AmoriumKItepe agndz Panaztepe *Ulubey Liman Te aAcemhyk Arsantepe eOrnma aciAe GAZANTEP KaURFA.. gBodrum. lOdemniz*AKSARAY . Su ie 'KOTAHra FAkhisar sHi .Kaman 0daALATYa .KeKuKraan/Ia LHallaem PaNysa nad KOenYA A riRANAKYAI SlaosDidymA.;?, LaginacePorsukAk 4Stratorla GAZANTEP URFAm*c Ti

Kndos FHiyek Karat-- Hacimusaar Y-SKENDERUN 0 2040/6 806 Tr s X a n th o s De re a z- U y r a .n e ..,

Fig. 1. Map of archaeological sites in Turkey discussed in the text 1996] ARCHAEOLOGY IN TURKEY 279

Tikunani while Hattu'ili saved engaged from in the a militaryclandes paign at the southeastern of Qanakkale's reaches ofmuseum his ki Unfortunately, the tablet excavated on which them it withwas w e together with several child's others, marble emerged chambe fro London antiquities market table, aand lyre not with in ivorythe of formal excavations. itemsInstead that of wouldrescuing not a hasit anonymity and presenting The secondit with isa historicala large m this significant text all has sides merely with complicat reliefs d difficult puzzle of Hittite on thegeography tomb of with Achilles yet query. quet. Whether this sa The number of antiquities to the tornearly fromfifth centurtheir texts and transformed, one within question a second, that will into n of time and place, increases context. Conservation at of allsuch the finds wasa carriedrate th subject has become familiar out by the Troia projectin thein an exemplary general spirit of pr cently Archaeology 48:2 cooperation. [1995] 44-56). Perhap alarming is the escalation Finally, of 1994 stealingappears to have been from the Year of theknow texts. Six directors reported Horse: Miocene hipparion during at Ozlfice and thein the Sinap 1995 posium that their sites, formation, among Neolithic horse the at Apikli, most late Middle visib sical ones, were the targets horses ofat Troy, clandestine Medieval horses at Kinet. dig The archaeological team The results at obtained Perge, from faunal alerted analysis, together to u cial activities in the with necropolis, archaeobotany, trace element found analysis, dendro- not smashed sarcophagus chronology, but and othershovels, techniques, have assumed hamme their picks abandoned in the place beside haste basic archaeological of dataflight to be reported by t The Turkish Ministry as ofa standard Culture component of ahas season's findings.just repa at great expense and after long negotiation BIBLIOGRAPHY large, inventoried sculptures from the gardens Erdek and Canakkale Museums, Conferences. The pieces 1992 and that 1993 annua ca have been removed with on excavations heavy given atequipment. the 15th and 16th one who participates inlogical this symposia illegal (May 1992 market and 1993) were be guilt: the scholars who as XV agree Kazz Sonuglarz to validate Toplantzsz I-II this(Ankara chandise; the dealers and XVI. Kazztheir Sonuglarz staff Toplantzsz who I-II conce(Ankara ugliness of their transactions breviated here with as KST 15:1elegant and 15:2 (1994),sho and high-society auctions; 16:2 (1995). and The 1993the reports buyers, on survey no and how small they imagine research, their XII. Arahtzrma purchases Sonuglarz Toplantz to It is against this backdrop 1995) and X. that Arkeometri the Sonuglarz resurfac Toplantz "Priam's Gold," after 1995),a disappearance appear here as Ara?ST lasting 12 (1995) an 50 years in the storerooms 10 (1995). Reports of forMoscow's the 1994 season, Pu t Museum and St. Petersburg's of this newsletter, Hermitage, were presented onshou 29 M considered (for bibliography, 1995 at the 17th see archaeological below). symposium The c stances of its discovery not appearin 1873, in print untilthe later photogra this year. Sophie Schliemann wearing Turkish itsmuseum finest personnel pieces, also convene th name given to what actually to present the representsresults of their work, at primari lea hoards excavated over and two conservation decades projects. - all Their belong 1994 re Romantic 19th-century Miize perspective.Kurtarma Kazzlarz Semineri Disputes (Ankara the ultimate ownership abbreviated of these here as objectsMKKS 5 (1995). shoul be understood as stemming The fromtook center this stage atout two approach to the past. ences. Better, Participants then, from Balkan to countriesdevo passion and energy toAnatolian more archaeologists pressing in Ankara oblig on 3-6J Another Homeric connection, to discuss "Thracians this andtime Phrygians: appro Pr came to light in 1994 Parallelism,"during asalvage session organized excavatio by Mid a necropolis near Biga, Technical ancient University Granikos. and the Bulgarian In the Ins of the Turkish press (Cumhuriyet, Thracology. A conference 18 held February in by two Late Archaic burials stitute offor Mycenaean"incomparable and Aegean-Anatolia his value," one of them illustrating (16-17 October 1995) a Trojanfocused on themthe Phr 280 MARIE-HENRIETTE GATES [AJA 100

their Anatolian milieu. A inbroader June 1995, an occasion range accompanied of by anAnatolian ex- issues was addressed in a hibitspecial entitled " session through the Ages." at The theTrojan 1995 ASOR meetings in metalwork that has(18 been housedNovember for the past 50 years 1995). The following 1996 conferences in Moscow's Pushkin Museum in was dueTurkey to go on dis- have been announced: "Ceramic Production Centers and play there on 1 March 1996, for one year. The cir- Exchange Networks in Hellenistic and Roman Ana- cumstances of its rediscovery are presented by D. tolia" (Istanbul, Institut frangais d'etudes anatoli- Easton in "Priam's Gold: The Full Story," AnatSt 44 ennes -late May); "Settlement and Housing in Ana- (1994) 221-43, and "The Troy Treasures in Russia," tolia through the Ages" (Istanbul, UNESCO-Habitat Antiquity 69 (1995) 11-14, and by M. Korfmann in II/Istanbul University-5-7 June); Third Interna- Studia Troica 5 (forthcoming). A full, detailed cata- tional Congress of Hittitology (Qorum - 16-20 Sep- logue to accompany the "Lydian Treasure," on view tember); "International Symposium on the Ottoman at the Museum of Anatolian Civilizations, Ankara, House" (Amasya, University of Warwick/British In- since its repatriation in 1993, is in press. stitute of Archaeology, Ankara- 24-27 September). Festschrifts and memorials. Halet (ambel's col- Journals and general publications. The widely cir- leagues and students have celebrated her many con- culated historical magazine Toplumsal Tarih included tributions to Anatolian archaeology with some works in its August 1995 edition a serious 10-page discus- of their own in Readings in Prehistory: Studies Presented sion by N. Asgari, U. Esin, and A. Ozyar on the state to Halet Cambel (Istanbul 1995), hereafter Readings of Turkish archaeology; it reflects an interest by the Cambel. (The editors--an anonymous collective- general public in issues beyond the more spectacular specify that this will be followed by a true Festschrift site presentations of Arkeoloji ve Sanat. More special- for her 80th birthday in 1996.) The colleagues, stu- ized publications have recently devoted entire issues dents, and admirers of Emily Vermeule have also to Anatolian topics: BiblArch 58:2 (1995) (but note marked her achievements in a volume edited byJ.B. that the maps should be used with caution), and Carter and S.P. Morris, The Ages of : A Tribute TelAviv 21:1 (1994) (on Urartu). Ankara Univer- to Emily Townsend Vermeule (Austin 1995), hereafter sity's Faculty of Language, History, and Festschrift Vermeule, with many articles on Troy and (AU-DTQF) announces the founding of an annual its neighbors. journal, Archivum Anatolicum, devoted to Anatolian A two-volume publication in the memory of Metin and philology (MBA-classical). The first Akyurt and Bahattin Devam, young archaeologists volume, edited by H. Ertem, appeared in 1995. tragically killed by a car bomb at Girnavaz in 1991, Since the Newsletter for Anatolian Studies compiles presents a rich collection of archaeological studies full and frequent lists of publications, the selection from prehistory to Medieval Anatolia: Studies for here, beyond items listed under individual sites, is Ancient Near Eastern Cultures: In Memoriam I. Metin deliberately kept brief. The first in the bibliographi- Akyurt - Bahattin Devam Anz Kitabz (Istanbul 1995). cal series for Bronze Age Anatolian excavations All archaeologists working in Turkey grieve at the has appeared, and can be welcomed as an essential sudden death of Osman Ozbek, head of the General reference: M. Korfmann, A. Baykal-Seeher, and S. Directorate's department overseeing excavations and Kil11 eds., Anatolien in derfriihen und mittleren Bronze- research, in June 1995 at the age of 43. Although zeit I: Bibliographie zur Friihbronzezeit (TAVO Beiheft he was responsible for the ultimate processing of B, 73.1, Wiesbaden 1994). For Hellenistic architectural all research paperwork, a task that required heroic sculpture, one can now consult E Rumscheid, Unter- patience at cyclical times during the year, he re- suchungen zur kleinasiatischen Bauornamentik des Hel- mained a master at welcoming each of us and show- lenismus (Beitriige zur Erschliessung hellenistischer ing genuine interest in our projects - even to thank- und kaiserzeitlicher Skulptur und Architektur 14, ing us personally for our efforts and discoveries. We Mainz 1994). And at the far end of the chronological miss him deeply. spectrum for this newsletter, recent discussions of Byzantine issues (many of them archaeological) have MIOCENE ERA been collected by C. Mango and G. Dagron eds., Payalar. The year 1994 marked the conclud Constantinople and Its Hinterland: 27th Spring Sympo- son for Berna Alpagut and her internati sium of Byzantine Studies, April 1993 (Society for the of investigators, who spent a full decade in Promotion of Byzantine Studies Publications 3, Cam- ing this Middle Miocene deposit southwest bridge 1995). It has proved to be Anatolia's richest site Museums, exhibits, and catalogues. The Byzantine mates, with specimens representing 7% wing of the Istanbul Archaeological Museum opened 19,000 fossil finds. The identification of st 1996] ARCHAEOLOGY IN TURKEY 281

may have been used as pounders, year-old Turolian phase and known a from study Mongolia toof pr mate diet based on teeth, Spain, areand previously bringing documented inthe Anatolia behavi of these 15 million-year-old around the Blackhominids Sea. Best represented into among sharpethe focus. faunal collections from 12 1 x 1 m soundings at Oz- Final publications of the excavations and special- liAce are hipparion (35%), followed by large mam- ized studies, including several dissertations, are in mals such as elephant and rhinoceros. The Ozliice progress, and will appear in a monograph series. All fossils are exceptionally complete, and include an fossil finds have already been recorded in a photo- elephant apparently trapped in a mudflow. Its west- graphic data base. Reports on the 1992 and 1993 ern extension, near Mytilene on the island of Samos, seasons appear in KST 15:1 (1994) 1-22 and 16:2 has produced similar taxa. (1995) 397-425. A general summary, coedited by PALAEOLITHIC Alpagut and P. Andrews, has been published in the Journal of Human Evolution 28 (1995) 301-405. Karain and 6kiizini. The concurrent pr Sinap Formation (Ankara). Survey and salvage ex- the Palaeolithic caves of Karain and ne cavations were continued for a sixth season in 1994 Okiizini, directed by Igln Yalglnkaya with a l in the Sinaptepe region 55 km northwest of Ankara, national team, pursued the programs of where hominid remains collected more than 30 years seasons during the 1994 campaign. At K ago had already alerted prehistorians to the region's cavations in the "K6kten Hall" continued to define potential as a bridge between Africa, the Near East, the main stratigraphy and its long span of Middle and . The current project, coordinated by Palaeolithic cultural components with Mousterian Berna Alpagut with Finnish and American colleagues and Levallois characteristics. Surveys in the imme- under the aegis of the Ankara Museum's director diate region have located an open-air site with typi- ilhan Temizsoy, has succeeded in locating geologi- cal Mousterian tools, probably a camp for the occu- cal deposits that form a continuous Middle Miocene pants of caves like Karain; and several more caves, to Early Pleistocene sequence spanning 13-14 mil- primarily Epipalaeolithic but often reused in the lion years. Over 100 sites have been mapped and in classical period. some cases tested with soundings. Of particular in- Results from the Okiizini excavations comple- terest are the hipparion fossils of both New World mented those of previous years, with further research and evolved Old World types recovered from several into the Upper Palaeolithic/Epipalaeolithic phases locations and different stratigraphic contexts. Future contemporary with the later occupation at Karain. seasons will begin to investigate the earlier, 15 The larger tools and triangular cores of the lower million-year-old tufa deposits that underlie the Sinap strata were gradually replaced, in the expected se- alluvial sedimentary formation. It is hoped that the quence, by Epipalaeolithic triangular microliths. Ex- region will eventually be linked to an intercontinen- ceptional 1994 finds that illustrate Okiizini's fine tal stratigraphic chronology. Preliminary reports with bone tool industry include a wing-tipped arrowhead, extensive bibliographical references appear in ArkST a harpoon, spatulas, and an object incised with de- 9 (1994) 237-56 and ArkST 10 (1995) 177-99. signs such as a fish. Continued use of the cave in Candir. The Middle Miocene deposits exposed 80 Neolithic times was again attested, in 1994, by a con- km northeast of Ankara, on the Ankara-Canklrl high- tracted burial accompanied by a large number of way, were investigated during a sixth season by Erksin stone, bone, and shell beads, a deliberately broken Giile? and an international team of collaborators. idol-like stone, and a ceramic cup. Excavations in the two localities already begun in Reports on the 1992-1993 seasons and associated 1993 produced fossil assemblages of bovids, probos- surveys appear in KST 15:1 (1994) 23-61 (Karain and cidae, and carnivores comparable to the previous Okiizini) and 16:1 (1995) 1-25 (Karain); for ESR dat- years' finds, and to the earlier fauna of the Sinaptepe ing of Karain tooth enamels and residue analysis of area. A lower jaw of Australopithecene type could stone tools, see ArkST 9 (1994) 33-37 (0. (etin and relate to Ethiopian hominids. For previous seasons, A.M. Ozer) and 257-66 (S. Demirci et al.). Caves found see KST 15:1 (1994) 63-75. by regional surveys are also reported by Yallnkaya 6zliice/Mugla. Berna Alpagut followed her brief in 1994 Anadolu Medeniyetleri Miizesi Konferanslarz survey of 1993 in the southern Aegean national park (Ankara 1995) 55-76, and by H. TaSkiran in AraSST near Mugla with a broader two-month investigation 11 (1994) 227-36. On art at Okiizini, see M. Otte et in 1994 of the region's Jurassic formations, in many al., Antiquity 69 (1995) 931-44. places threatened by industrial development and lig- Sehremuz-. The 1980s salvage excavations nite mines. The deposits belong to the 9 million- in the upper Euphrates near Samsat have appeared 282 MARIE-HENRIETTE GATES [AJA 100

Wm--

......

4:- ;-t&

4 7

......

44i

W Mull-,-

Fig. 2. Latmos-BeSparmak Dagi. Prehistoric cave painting. (Courtesy A. Peschlow-Bindokat)

in a final publication: G. Albrecht and H. Miuller- haps place it much earlier, even into the Upper Beck eds., Das Paldolithikum von Sehremuz bei Palaeolithic Samsat (10,000 B.C.). A preliminary publication, am Euphrat (Tilbinger Monographien zur with Urge- illustrations, appears in AntW 26 (1995) 114-17. schichte 10, Tilbingen 1994). Classical remains recorded by the Latmos area sur- U1agizh. The analysis of various pigments vey pro- are presented below under "Classical, Hellenis- cessed by the Upper PalaeolithiclAurignacian tic, occu- and Roman : Herakleia am Latmos." pants of this cave site in the coastal Hatay has been EARLY NEOLITHIC published by A. Minzoni-Deroche et al. in Antiquity 69 (1995) 153-58. These pigments, dating to 32,000 Hallan (emi. The 1994 excavations conducte b.p., would represent the earliest attested use Michaelof fer- Rosenberg, University of Delaware, at ruginous ore to produce the color red. largely aceramic site on a tributary of the Bat Latmos-Bepparmak Dagl. Although prehistoric River reinforced the results of the previous t settlement in the area is barely attested, the seasons, 1994 while extending its occupational phases Herakleia survey team directed by Anneliese two to three and perhaps four building le Peschlow-Bindokat discovered, in the northern Throughout these phases, circular structure reaches of the mountains, two small rock shelters closed the same open activity area, although t decorated in red and brown pigments with figural building materials and architectural details ch scenes. The better preserved of the two (fig. 2) rep- noticeably. The settlement was smaller (well u resents a number of upright figures, primarily 0.5 ha) than earlier estimates indicated, and pe women with exaggerated buttocks and breasts, and appropriate for several generations of a mode at least one clear scene of sexual intercourse. This tended family living in the eighth millennium type of parietal art is otherwise unknown in central They would relate to their contemporaries and Aegean Anatolia. It could date as late as the Chal- Jebel Sinjar and upper Tigris of northern Iraq colithic period; but context and subject would per- Rosenberg kindly reports: 1996] ARCHAEOLOGY IN TURKEY 283

"The 1994 excavations a general reportsucceeded in Anatolica 20 (1994) in121-40. coorFor the stratigraphic relationship 1994, see Anatolica 21 (1995) for1-12. three an four architectural levels to the north and south of A4skh Hoyiik. During the sixth excavation season the central area that remained open throughout the at this exceptional aceramic site on the banks of the site's occupation. The two upper levels, already known Melendiz, east of Aksaray and close to rich obsidian from earlier seasons, could be readily distinguished sources, Ufuk Esin deepened the sounding on the by their different construction techniques. In the mound's high north slope, and extended the hori- latest phase, the walls of the two semisubterranean zontal exposure of the upper settlement toward the circular structures were built of upright sandstone south and northeast. The northwest sounding illus- slabs and wooden posts. In contrast, the preceding, trated further the homogeneous and conservative second phase used pebblelcobble walls held together nature of this culture: house walls were rebuilt pre- with mortar. These are now represented by four struc- cisely on top of their predecessors (by 1994, as many tures, three of them circular and carefully paved with as eight successive rebuildings had been revealed), sandstone slabs, and the fourth smaller, unpaved, their single-room compartments at times indented and U-shaped. In 1994, a third and earlier level was to accommodate neighboring units so that no space also found to share all of the features of the second was wasted between the individual structures. On phase, including the two types of buildings. Evidence top of the mound, the large south building with case- for the presence of circular wattle-and-daub struc- mate walls revealed more of its atypical architecture. tures in levels 2 and 3 was provided by round plas- It now includes a court enclosed by a stone and mud- tered platforms, postholes, and one burnt wood brick wall and paved with large mudbricks (the superstructure. At the close of the season, traces of earliest attested in Anatolia), and would appear to a fourth level also began to appear. The two struc- have served a public, perhaps religious, function. tures of level 1 are still considered public because A pebbled street running along its northern side was of their furnishings. All of the level 2 and 3 build- followed to a fork and its northern and eastern ex- ings would now appear to be domestic. tensions. The houses lining the pebbled street in the "The local ground stone industry, now familiar northeastern sector were exceptionally constructed from previous seasons, included an unfinished pestle of clay slabs mixed with fieldstones, the floors and that was abandoned when the stone split, unfinished walls often painted red, and the units consisting of bowls and bowl blanks, as well as what may be a as many as three rooms. One of these was divided spindle whorl. Bone artifacts still consist primarily by narrow partitions into storage bins, whose con- of awls, in addition to toggles, polished boar tusks, tents may be retrievable with microsoil analysis. A and a second example of a decorative snake-shaped large refuse pit was also found in this district off the piece. pebbled street. As in the previously excavated areas, "The inhabitants of aceramic Hallan semi relied a few burials were located under the house floors. especially on pulses and nuts (such as wild almonds) The 1994 excavations again produced large quan- for their diet, not on cereals and wild grasses. Theytities of single and spacer beads in a variety of ma- were, however, experimenting with pig domestica- terials, some imported: stone, bone, solid heat-treated tion, again indicated by the 1994 data, as well as prac-copper and cold-hammered rolled sheets of natural ticing selective hunting of male sheep and goats. copper (from Ergani?), and seashell from the Medi- "The relationship between the aceramic settle- terranean. The obsidian toolkit includes triangular ment and the ceramic material found in 1993 on microlithic arrowheads retouched on one side, per- the south side of the mound was investigated in haps 1994 a link, like the architecture, to Cafer H6yiik. in a test sounding, down to a depth of 2 m without Botanical finds continue to favor a broad range of reaching either an underlying aceramic level wild or grasses with few cultivars, suggesting little in- sterile soil. Since the pottery and other finds in terest this in agriculture. The faunal assemblage, also deposit were sporadic and not associated with wild, any can now be expanded to include the horse. Thus architecture or features, they may reflect intermit- this broad and sophisticated site, consistently dated tent camping. The present evidence therefore sug-by radiocarbon and ESR to the eighth millennium gests a discontinuity along the mound's southern B.C., apparently relied on the region's natural bounty area, which should not be included when calculat- for food while bringing in exotic products from dis- ing the size of the earlier settlement." tant places. The 1992-1993 season reports are published in For the 1992 preliminary report, see KST 15:1 KST 15:1 (1994) 123-29 and 16:1 (1995) 79-94, and (1994) 77-95; specialist studies on human remains 284 MARIE-HENRIETTE GATES [AJA 100

and botanical specimens Thin gray(hackberries) burnished pottery, appearsome red ArkST 9 (1994) 23-31 (M. Ozbek) and 101-109 (S. painted wares, and infrequent chipped st Giillur), respectively. Esin considers the Ap~kli metal (with some obsidian) characterize this later Neo- finds and their evidence for early metallurgy in Read- lithic assemblage. The recutting and continuation ings jambel 61-77. of Garstang's stratigraphic sounding demonstrates Pinarbali/Konya Plain. Survey and several selec- that below levels XXIX-XXX lie another 7 m of oc- tive soundings carried out in 1993 and 1994 by cupational deposits down to sterile soil, with several Trevor Watkins and Douglas Baird as part of the new burnt architectural levels, and stone foundations at project at Qatalh6yilk have succeeded in providing the earliest stage of the settlement (radiocarbon- it with long-awaited relatives and antecedents in the dated to ca. 6000 B.C.). Lentils, emmer wheat, olives, Konya region. Several rock shelters and open-air sites almonds, and the standard range of domesticated were identified at Pinarbaip, near the small Sifley- animals, supplemented by some fishing and hunting, manhaci lake at the base of Kara Dag, 20 km from suggest a well-fed population. the main mound. Soundings in one of the shelters Excavations on top of the mound, which has been traced a long history of use (burials and hearths) from considerably modified by recent terracing, recorded Epipalaeolithic (pre-8000 B.C.) to aceramic and ce- three levels of large Medieval buildings that made ramic Neolithic times, the latest level being parallel liberal use of classical spolia including columns. with Qatalh6yuik. The Epipalaeolithic period was also Their stratigraphy requires modifications to the tested at the open-air site recorded in 1993 on the scheme published by Garstang. A preliminary re- lake edge. It is characterized by a chipped stone in- port on the entire 1993 project has been published dustry using an exceptional percentage of obsidian in KST 16:1 (1995) 27-41. (80%), and a wide range of fauna but- despite the ,atalh6yiik. During a second season preparatory lake- few fish. Such findings, together with the geo- to the resumption of excavations, Ian Hodder and morphological studies being undertaken in conjunc- an interdisciplinary team completed their topo- tion with the project, promise to clarify the environ- graphic survey of the 13.5-ha East Mound, and the mental and cultural setting in which the developed later 8.5-ha West Mound. Systematic surface collec- Qatalh6yilk Neolithic phase occurred. tion also delineated more clearly the mounds' For the 1993 survey, see Ara?ST 12 (1995) 421-27 chronological phasing. The latest Neolithic levels are and AnatSt 44 (1994) 13-15; the 1994 session is sum- located on the eastern rise of Qatal East. Classical marized in British Institute of Archaeology/Ankara and Byzantine settlements are found on the south- Research Reports 1994, 14-17. ern side of Qatal East, and on the eastern side of Qatal West; a Byzantine lower town also extends from LATER NEOLITHIC the western mound across the fields toward the east /Mersin. During the second over season a 10-ha of area. Cores into the deposits below Mel- the new Yumuktepe excavations codirected laart's by earliest Veli level (XII) indicate a further 4-5 m Sevin (historic levels) and Isabella Caneva of Neolithic (prehis- occupational deposit. toric levels), trenches were again opened Surfacein the scraping area and remote sensing (magnetom- of Garstang's northwestern Trench A to eter)correlate were especially the informative in a 40 x 40 m current project and the present state of square the onmound the high northern area of atal East, where with his discoveries of over half a century irregular ago. houses An were found clustered on either side upper 10 x 10 m trench uncovered portions of a narrowing of aalley. Scraping also determined that massive burnt mudbrick building (6000 B.P.)the Neolithic that can inhabitants, estimated at 5,000-10,000, be assigned to Garstang's Chalcolithic level were XVIdumping and refuse into the depressions on the is probably a western extension of his fortified site, while forbar- the most part building their new houses racks. Several rooms with doorways and onlytwo onwindows earlier wall foundations. Continued clean- appear to have been subdivided at a later ing of stage the 1960s' by excavation areas, and recording of flimsy partitions that suggest a long duration their sections, for this again included microsoil analysis of level. In situ pottery consisted mainly house of floors.red andResults suggest that the "shrines" origi- brown handmade bowls, and a small number nated as of domestic ves- buildings and assumed more spe- sels painted with yellow and brown zigzags. cialized functions Ubaid through time. New studies under- types were much rarer than indicated by taken Garstang. on the ceramic and flint finds from the earlier In a lower, Neolithic trench to the south, excavations a level some- are also attempting to assess whether what below Garstang's level XXVII produced the assemblages the from the "shrines" and houses differ stone foundations of a small buttressed significantly.structure whose walls, probably pise, were plastered The ancient white. landscape around (atalh6yuik is re- 1996] ARCHAEOLOGY IN TURKEY 285 emerging thanks to further "Porsuk geomorphological Culture." Houses were freestanding, study. bui The site originally stood of on mud oror mudbrick, near somethe on banksstone foundations. of the Qargamba River, whose Bricks alluvial are first attested deposits in level III. A one-room ensured struc- fertile fields for its farmers. ture of At mudbrick some on stone stage, foundations afrom channel level was dug to bring water between IV was apsidal. the two mounds. It remains to be discovered "Thewhen pottery ofthe the Porsuk river culture, changedwhich includes course away from the site, pre-Vin'a and elements,the effect is comprised this mainly might of bur- have had on the site's occupation. nished wares, slipped A wares,summary and painted wares. of Char- the 1994 season can be found acteristic in openBritish forms include Institute shallow bowls andof Archaeology/Ankara Research plates, Reportsdeeper bowls with 1994, everted rims, 11-14; and carinated for 1993, see AnatSt 44 (1994) bowls.10-13. Other typical vessels consist of jars with cy- For Neolithic and earlier lindricalsites necks, in andthe squat-bellied Qatalh6yfik pots with everted rims and vertical lug handles. The dark-burnished area, see above, under "Early Neolithic: Pinarbapl." Bademagaci H6yiik. Refik vessels Duru are often conducteddecorated with grooved a orsec- punc- ond excavation season in 1994 tate bands on on theirthis shoulders; large some have mound, surface well situated on the route rippling.leading Knobs areinland frequently attachedfrom to the the car- coast at up to the inations.Anatolian Some of the slippedplateau. wares have Alinear 6-m pat- deep sounding on the more terns 'painted' promising in the same reddish northeast- slip. Another ware ern side confirmed the previousclass, painted with redseason's and/or black spiraloid occupa- patterns tional phases: EB I-II, preceded on a white or creamby ground,a long recalls StarcevoNeolithic motifs. sequence of domestic architecture. "Figurines are also The characteristic pottery, of the Porsuk ofcul- local type distinct from the ture. Kurupay/H6yticek The earliest, from level I, with their culture,arms folded is unpainted, and includes andtall meeting bag-shaped mid-chest, can be compared jars towith tubular handles. Traces of I examples. a modest For the later andEarly more stylized Chalco- figurines, lithic settlement were again only foundthe large buttocks only are carefully in the modeled; south- other- ern part of the mound, aswise, theindicated figures are shown byarmless, red and the warestwo legs with festoons and chevrons are inseparated white only by apaint. shallow vertical The incision site's at reoccupation in the EB I theperiod front and back.is Similarmarked types occur by at Ilipinar pot- tery resembling the earliest VI (mid-fifth EB material millennium) and from in the Balkans. Beyce- Other sultan, large structures with finds includestone chipped foundations, stone tools, grinding stones and perhaps a fortified enclosure. and pestles, Cobbling marble bracelets, atgroundstone the celts,foot and of the northeastern slope maybone tools. be part of a rampart. For the 1993 season, see KST "Level VI,16:1 investigated (1995) in a very 69-77. limited area, should Kurucay. The first of the befinal linked toreports the succeeding on level this VII rather Neo- than to lithic and Early Chalcolithic the earlier site phases. near With level Burdur VII, the ceramic has as- been published by R. Duru, semblage Kurugay changes significantly I: 1978-1988. from the preceding Kazzlarzn Sonuglarz. Neolitik Porsuk ve cultural Erken phase, implying Kalkolitik a gap between (ak them. Yerlegmeleri (Ankara 1994). The later phase is characterized by black-burnished Orman Fidanligl. Turan Efe, wares withIstanbul generous straw University, temper and white-painted succeeded in his efforts to correlate the fifth- decoration; black-topped bowls with slightly incurved millennium contacts between Anatolia and the rims, or S-profiles; and horn-shaped handles on jars Balkans with a final season at this site southwest of and round-bellied pots. This pottery shows affinities Eskigehir, where quarrying and deep erosion layers with Yazir H6yfik and Late Chalcolithic made excavations exceptionally difficult, and strati- 1-2, and provides a terminus ante quem for the fied remains came as the reward of both patience Porsuk culture. It would ultimately link the stratig- and persistence. He kindly reports: raphy of contemporary sites in the region- "Salvage excavations were brought to a close inIlipinar, Yarimtepe, Toptepe - with western Anatolia 1994 at the end of the third campaign. Seven levels and Beycesultan. A copper pin with a double-spiral were distinguished (I-VII, with I the earliest) in an head (fig. 3) and a copper piercing tool from level excavated area of 320 m2. VII represent the earliest metal implements yet found "The finds from levels I through V form a single in northwestern Anatolia." continuum with clear internal development. Because Illplnar. Jacob Roodenberg resumed his research the pottery from these five levels reflects a new on Neolithicas- and Chalcolithic Balkan-Anatolian con- semblage, otherwise known only from survey in nections the with a seventh campaign in 1994 at this early upper Porsuk valley, we have chosen to call this farming the community east of the . Ex- 286 MARIE-HENRIETTE GATES [AJA 100

Fig. 3. Orman Fidanligi. Level VII: copper pin. Length ca. 7 cm.

cavations continued in the northern and western the cooperation of the (anakkale Mus squares, and confirmed the site's 40.5-m2 overall area phasing, of the large pre-Troy I which is complicated by spiral stratigraphy exposed immediately and the below the m slippage of upper levels into theManfred western Korfmann, gully. Tiibingen Unive Further study of the late sixth-millennium reports: wattle- and-daub burnt house in the large "Levels trench and W 12/13finds of the Kum Tep (level IX rather than X as in AJA the 98 new [1994] project 253 nowand refers to the Kum 255 fig. 3) suggests that it contained were anuncovered inner plat- below the IB deposit form, another link with the Balkans. This one-room trenches. Less expected was the disco house remained the standard type tlement at Ilipmar IA wasfor founded500 on top of a years. Below it, in level X, were foundflexed 15 burials burials of that pre-Karanovo typ also bear no resemblance to the Balkancontemporary Neolithic Ana- sequence (earlier fif tolian (intramural) tradition. The B.C.). children, This cemetery young represents the fi adults, and one 60-year-old woman an Earlywere placedNeolithic on occupation in the their left or right sides, flexed, withoutmust have fixed played orien- a key role in north tation, accompanying gifts, or aBalkan discernible interconnections." burial pit. This level marks the earliest Accountsoccupation of the 1993 at and the 1994 seasons appear as site, founded on virgin soil. part of the Troy preliminary reports in Studia Troica The level VI transition to the Early4 (1994) 1-50,Chalcolithic KST 16:1 (1995) 239-62, and Studia period, when mudbrick replaced Troicawattle-and-daub 5 (forthcoming). for the typical one-room house, was Kulaksizlar-Akhisar.investigated Rafet in Dingthe surveyed a marble western squares, at the edge of workshopthe ancient in the region settle- of quarries between Akhisar ment. A new cultural level, VB, wasand . defined Fifteen differenthere, to-types of blanks and gether with an unprecedented unfinished compartmentalized pieces include stone bowls and possible building fronted by a courtyard. figurines. A row The workshopof jars could and date to the period plastered reed containers, originally of Kum Tepe shaded IA, or as late by as the an Early Bronze Age. awning, was found in situ along the Apagh court's Plnar/Kirklareli. back Thejoint wall. Turkish-German Inside the doorway were two separate project led kitchen by Mehmet areas.()zdogan and Hermann Although the back of the building Parzinger was to destroyed investigate early Balkanby and northwest plowing, enough remained to indicate Anatolian that connections it was in Thrace built carried out a sec- on several terraces rising up the ond slope. season ofThe excavations building at A?agi Pinar in 1994. and its associated level would date The siteto maintainedthe mid-fifth its Balkan cultural affiliations millennium, a century or so later from than Neolithic level (viz. Chalcolithic VI. It in Anatolian chro- is discussed in Orient Express 1994:3, nology) to 69-71.Hellenistic times. Three stratified levels For the EB settlement at Hacilartepe, of domestic and structures its builtceme- of wood, with interior tery at Ilipinar, see below, under platforms "Chalcolithic and large circular and hearths, are associated Bronze Age: Western and Coastal with KaranovoAnatolia." III-IV pottery spanning all phases Kum Tepe. As part of the Troia of Projectthe Balkan Middleand Neolithicwith period (later fifth 1996] ARCHAEOLOGY IN TURKEY 287

Fig. 5. Arslantepe-Malatya. Building IV. Corridor wall with impressed and painted decoration. (Courtesy M. Frangipane)

its energies on the huge temple-palace complex known as Building IV, dated to the late fourth- millennium Uruk-affiliated level VIA (EB IA). The 35-m-long corridor that linked the units of this com- plex, and Temple B at its north end were excavated down to their floor levels, which had not been reached in 1993. Director Marcella Frangipane, Uni- versita di Roma, generously provided illustrations for the following summary. The long corridor (fig. 4) allowed communication Fig. 4. Arslantepe-Malatya. between theBuilding independent structures IV thatfrom made up the with gate in foreground and Temple B in upper righ the complex, and drainage via a water channel under figure for scale. (Courtesy M. Frangipane) its floor. At its lower south end, behind a monumental gate, it gave access to a series of storerooms. The millennium B.C.). Malachite corridor's inner sector,from the approach the to Templesecond B, suggests that metalworking was roofed, plastered, took and decorated place with impressed here. I earliest phase a flexed lozenges burial and geometric motifswithout in red paint (fig. grave 5), was found, a tradition resembling shared Uruk fashions. with It was later Ilipinar. replastered T was reoccupied, after aand longrevetted withhiatus, wood panels toduring secure a second the Age. Finally, the Hellenistic story. Burnt beams tumulus lying on the floor ofthat the corridor crow prehistoric mound, built give hope forup a dendrochronologicalfrom materials date. m from the earlier deposits, Temple B, 15was x 12 m, probablysurrounded two stories high by containing funerary giftsand bipartite inand plan (fig. sacrificed 6), proved also to be fur- anim Balkan funerary practice. nished with A twinned report cult items: twoon altars the or plat- firs son has been published forms, pairs in of clay KST basins, and 16:1two offering (1995) tables Ozdogan discusses northwest in the center of the cella. TheAnatolian entrance was on the Neo chronologies in Tiirk eastTarih side, away fromKurumu the corridor, throughKongresi a wing (1994) 69-79, and Readings of small rooms Cambel that contained 41-59. kitchen equipment For excavations at the such later as mortars Chalcolithic and pestles, but no pottery. Kanlig Large see below, under "Chalcolithic ceramic vessels in the cella and were concentrated Bronze in the Age: ern and Coastal Anatolia." northwestern corner of the room across from the entrance, and opposite the far altar. They included CHALCOLITHIC AND BRONZE AGE several unique types in local wares but Uruk shapes: Southeastern Anatolia and Cilicia two large jars in period VII fabrics (red with chaff Arslantepe-Malatya. The 1994 campaign focused temper), a red-burnished carinated jar, and a pair 288 MARIE-HENRIETTE GATES [AJA 100

Fig. 6. Arslantepe-Malatya. Building IV, Temple B M. Frangipane)

of wine bottles (fig. 7). Three tall-stemmed bowls were placed in front of the altar against the long wall. The vessels beside the other altar still contained food residues. A dozen more sealings (for a total of 50) were recovered from the building, especially from one of the side rooms where they appear to have fallen off a shelf near the window. Investigations in the area of Temple A, which lies to the west beyond Building IV, confirmed that it was a later construction (as indeed suggested by their different alignments) and could not be reached by the long corridor. In front of the temple, an enclo- sure with broad stone foundations formed a room reached by a flight of stone-paved steps. The room was burnt and evidently looted. The doorpost and its copper alloy door socket, held in place with nails, were found in situ. This extravagant use of metal, together with numerous other metal finds dating to this phase, give some measure of the resources and prosperity of the inhabitants of level VIA. The 1992 and 1993 preliminary reports appear in KST 15:1 (1994) 211-28 and 16:1 (1995) 165-76. Hacinebi Tepe. Further investigations into the interaction between fourth-millennium local entre- preneurs and Mesopotamian businessmen in the upper Euphrates valley were carried out during a third rescue season by Gil Stein, Northwestern Uni- Fig. 7. Arslantepe-Malatya. Wine jar from Temple B, Late versity, and codirector Adnan Misir, director of the Uruk period. (Courtesy M. Frangipane) 1996] ARCHAEOLOGY IN TURKEY 289

$anliurfa Museum. They ca. 8 x 9 m kindlyand 2.4 m high, thatprovide underwent a series the ing assessment: of repairs and remodelings. It was expanded at some "The 1994 excavations point bycontinued a lower terrace with ato connecting widen ramp and th vious operations in the staircase. northern, No trace of the superstructure southern, was pre- a ern areas of the 3.3-ha mound for an excavation area served, but the scale of the platform would have of 600 m2. The upper level of all three areas pro- suited a public building originally set in an open duced further evidence of the site's last major set- area. The first phase of the platform can be associ- tlement, dated to the Achaemenid/earlier Hellenistic ated with ash tips containing local ceramics and period. Large-scale public buildings in the western Anatolian-style stamp seals. Later, domestic rooms and southern trenches, and the stretch of defensive were built up against the platform's northern face wall uncovered in 1993, would suggest that the site's and eventually on top of it as well. A large pit cut strategic location on the major east-west river cross- into these later buildings still contained local pot- ing at Birecik was being fully exploited as a military tery and stamp seal impressions depicting mainly outpost. The buildings, whose mudbrick walls on cervids and lions, motifs common to this region as stone foundations were in places preserved to heights known from Arslantepe and Degirmentepe. A lime- of 1.5 m, had been carefully cleaned out and aban- stone stamp seal in this style, as well as two blanks, doned. They can be dated nonetheless to the late gives a firm local context to this recording system. fourth-early third century B.C. by a Hellenistic coin An influx of foreign goods and practices marks the found on a floor; and stratigraphically by intrusive end of the phase, and is best illustrated by the con- storage and trash pits, containing late third- and tents of a large pit: pottery exclusively of Uruk type, second-century B.C. pottery. Evidence for some pres- two cylinder seal-impressed jar stoppers, a tablet ence at the site during the century preceding Alex- blank, and a fragmentary clay tablet with a cylinder ander is indicated by the burial gifts in a tomb seal impression but no preserved notations (fig. 8). excavated in 1993, in particular a signet ring in In addition to these, a stamp sealing of local type Achaemenid style. suggests that, in the pit at least, the two recording "The Late Chalcolithic fourth-millennium settle- practices were contemporary. ments that represent Hacinebi's only other occupa- "Early phase B structures were also excavated in tional phases underlie the Hellenistic buildings in 1994 on the southern slope of the mound, where all areas of the site. The deposits are 5 m deep. They a stone platform built late in the pre-contact phase are clearly separable into an earlier and later phase to which the 1994 season was able to assign calendar dates backed by radiocarbon samples. Phase B, the "contact phase," is characterized by Uruk-affiliated artifacts such as clay cones, bullae, and tokens. These occur in contexts with local wares that roughly correspond to the Amuq G period, but belong in fact to the distinct ceramic horizon known from Arslantepe VIA [where it is classed with EB I] and Kurban VIA. Haclnebi's earlier, "pre-contact phase," or phase A, at present the first-known occupation of the site, would be equivalent to Amuq F and espe- cially Arslantepe VII, before the appearance of Meso- potamian traits. "During phase B, whose long duration is indicated by modifications to many of the structures belong- ing to several architectural levels, the northern area of the mound may have served as the residential dis- trict for a small enclave of foreign merchants. The concentration of Uruk ceramics, administrative arti- facts, and other characteristic Uruk items such as cruciform grooved stone weights and clay sickles ap- pears greater here. Below residential architecture dating to the middle and end of the phase, the 1994 Fig. 8. Hacinebi Tepe. Tablet with cylinder seal impression, season exposed most of a large stone/rubble platform, phase B. (Courtesy G. Stein) 290 MARIE-HENRIETTE GATES [AJA 100

continued to be used in duce the first completeensuing map of an period;EBA Near Eastern an near it, the stone foundations of a monumental wall city. with broad niches/buttresses on its inner face. It too "Soundings conducted in various parts of the was later subdivided by flimsy walls into narrow Outer Town to provide the map with archaeological rooms. Late in phase B, these were in turn cut into interpretations already invite a number of prelimi- by pits, one of them filled with beveled-rim bowls. nary observations. During the latest EB phase, that It would appear that the size of the settlement was recorded by remote-sensing, the eastern two-thirds decreasing over the course of phase B. of this district were densely built up without open "Phase A, defined as the pre-contact Late Chalco- spaces on a network of fairly straight roadways, lithic period of the earlier fourth millennium, was known from excavation to have been paved with thick exposed during the 1994 season in the northern and layers of cobbles and sherds. The houses flanking western areas of the mound. In the north, it consisted the streets were in some cases large-one building of terracing and a large wall set into the natural hill per block- and perhaps suited to extended families. on which the settlement was founded. This terracing In the central area, larger and more massive struc- was built over during phase B, in particular by a later tures may represent public buildings. Two intersect- terrace east of the one described above. The west ing thoroughfares on the map run on a rough area produced a well-stratified sequence of building north-south line across the western portion of the levels, all dating to phase A, and including two nar- Outer Town, and for hundreds of meters east-west row rooms, perhaps for storage, beside a courtyard along its northern edge. In contrast to this densely house excavated in 1993. A ceramic mold with copper planned residential area, the northern and western residues, and a copper chisel, both from the house, peripheries produced few magnetic anomalies; attest to metalworking in early Haclnebi; and a stamp whether their absence indicates gardens or activity seal impression once used to secure a woven basket areas, or results from non-archaeological factors, can shows an early interest in formal administration." only be resolved with future soundings. Finally, the The 1992 and 1993 seasons are presented in KST outer town was defended by a 6-m-wide fortification 15:1 (1994) 131-52 and 16:1 (1995) 121-40. See also wall on top of a massive glacis, as discovered in Stein et al., supra 205-60, in this issue ofAJA. A pre- trenches just inside the broad moat mapped and liminary analysis of the pottery has been published tested in earlier seasons. The layout of the Lower by S. Pollock and C. Coursey in Anatolica 21 (1995) Town at the foot of the mound proper appears to 101-42. share many of these urban features, including the Titri4 Hoyiik. Remote sensing and excavation con- system of streets. The suburbs, on the other hand, tinued for a fourth season at this large site (now may have followed different norms: limited surface estimated at 43 ha) in the upper Euphrates valley survey in an area outside and east of the Outer Town north of Urfa. As previous seasons had already produced significant numbers of blade cores, sug- demonstrated, Titri? conformed with the typical EB gesting that at least this suburb functioned as an III-IV Syro-Anatolian urban configuration of a lower industrial zone. town and suburbs extending across a broad area at "Broad trenches in the Outer and Lower Towns the foot of an mound. Since these lower again addressed the critical issue of the site's urban districts were not reoccupied after the close of the development in the Middle (Kurban IVC-B = EB III) period and their remains lie just below the surface and Late (Kurban IVA = EB III/IV) EB periods, with of open fields, the site is ideally suited for research further investigations into the stratigraphic sequence into the dynamics of a third-millennium B.C. resi- in these areas. In the eastern sector of the Outer dential center. Guillermo Algaze, University of Cali- Town, the Middle EB level, built on virgin soil, is char- fornia, San Diego, who codirects this project with acterized by the well-built stone foundations of urban Adnan Misir, Sanliurfa Museum, kindly reports: architecture, as yet exposed only in narrow soundings. "As in preceding seasons, the principal objective This level was followed by an apparent hiatus, dur- of the 1994 campaign was to clarify the urban struc- ing which the area was used as a cemetery with mul- ture of the Middle-Late EBA settlement at TitriS. tiple burials in cist graves and jars. During the Late To this end, the magnetic field gradient survey was EB reoccupation, when more modest housing fol- expanded to cover the entire Outer Town as well as lowed a different alignment, the location of these the previously unsurveyed areas of the Lower Town, graves seems to have been kept in mind when laying resulting already in one of the largest such maps for out the new constructions. The later houses are now a Near Eastern site. Ultimately, this survey will pro- illustrated by several broad exposures (fig. 9). They 1996] ARCHAEOLOGY IN TURKEY 291

Fig. 9. Titri? Hoyfik. Late EBA housing were built as mudbrick evidence superstructuresfor local mortuary customs: the burials onwere t foundations, and included used for single families, several and included roomsviolin-shaped a cobbled courtyards with marble idols hearths in or outside the burial, and and metalovens. T of one house contained an oval basin and a drain; jewelry (fig. 10). The tradition of burial in cist tombs tartaric acid residues identify the room as an instal- is also attested for the pre-urban early third- lation for processing grapes. There appear to be two millennium settlement (see AJA 99 [1995] 216 and architectural phases here, both following similar fig. 7). parameters. "Results of the 1994 campaign support the pre- "The Lower Town, in contrast, was occupied with- vious assessment of the urbanization process at Titri? out interruption, with larger and more impressive as a swift and well-coordinated project that occurred buildings that signal a neighborhood of higher status, in the Middle EBA. Subsequent development in the especially on the western side. Portions of two large various sectors of the site proved more variable, with buildings with massive limestone foundation blocks the Outer Town undergoing significant transforma- were uncovered on either side of a cobbled street. tions while the Lower Town continued largely un- They were built in the Middle EB period, and re- changed. Palaeobotanical and environmental re- mained in use with few modifications throughout search indicates that the inhabitants practiced a the lifetime of the city. At a late stage, a stone hypo- mixed economy relying on grain and legumes, 'cash geum was installed in the southern building below a partitioned room, with access through a semi- circular dromos blocked by a limestone door. The tomb contained at least two (poorly preserved) in- dividuals, 42 Late EB vessels including Syrian bottles, and several bronze pins. Well-appointed stone-lined tombs were also characteristic of the Middle EB extra- mural cemetery 400 m northwest of Titri?, first in- vestigated in 1981 by H. Hauptmann and again by us in 1994. Although the cemetery has suffered from Fig. 10. Titri? Hoyiik. Silver bracelet from Middle EBA plowing and other disturbances, it produced good extramural cemetery. (Courtesy G. Algaze) 292 MARIE-HENRIETTE GATES [AJA 100

crops' such as grapes, and cuit wall wildthat was not nuts preserved. and These defenses fruits were i cluding acorns, in an environment built during two separate phases, that in the mid/laterwas moi and more wooded than third today. and late third/early Why second this millennia prosper B.C. The city contracted to less accumulatedthan aevidence 10th still confirms of its that at size least parts and treated to the mound at the close of the third mil- of the lower town continued to be occupied into the lennium remains a question for future seasons." Middle Bronze Age, although the outer town and The 1992-1993 seasons are reported in KST 15:1 its industrial quarters had by then been abandoned. (1994) 153-70 and 16:1 (1995) 107-20; a comprehen- For the 1992 season, see KST 15:1 (1994) 177-92. sive article on the 1994 season appears in Anatolica Oylum H6yiik. Under the direction of Engin 21 (1995) 13-64. Ozgen, the 1994 season at this major site, near Kilis Kurban Hoyiik. Patricia Wattenmaker discusses and the ancient (and modern) crossing into north- the organization of domestic craft production in western , focused on the 22-m-high eastern slope rural southeastern Anatolia with reference to Kurban and its stratigraphic sounding. Hellenistic, Iron Age, H6yiik's third-millennium phases in G.M. Schwartz and Middle Bronze buildings were traced, many of and S.E. Falconer eds., Archaeological Views from the them burnt with their contents in situ. Well-preserved Countryside: Village Communities in Early Complex Soci- mudbrick walling of MB II date promises to become eties (Washington, D.C. 1994) 109-20. a monumental structure in future seasons. A sound- Kazane Hoyiik. In their third salvage campaign ing on the southern slope, where the bronze figu- on the outskirts of Urfa at Kazane, whose extensive rine of a Hittite smiting god was allegedly found in lower terrace has been bisected by the GAP irriga- 1993, produced a gold pendant and three architec- tion channel, Patricia Wattenmaker and codirector tural levels that bore no visible connections to the Adnan Misir conducted stratigraphic soundings for figurine. For a photograph of the Hittite figurine, the early phases of the site, and opened broader areas and reports on the 1991 and 1993 seasons, see KST to investigate its EB III-MB urban phases. 16:1 (1995) 95-105. A sounding for Kazane's earliest-attested period, Tilbepar. A 1994 survey preliminary to full-scale on the southeastern edge of the outer town, produced excavations was undertaken by Christine Kepinski- Late Halaf levels characterized by pottery with Lecomte to the southeast of Gaziantep, at the 60-ha bucranium motifs, stamp seals, and a high percent- site crowned by the Crusader fortress of Turbessel, age of obsidian tools. If these belong to the same fief of the kingdom of . Systematic collection Halaf settlement noted from survey elsewhere at the in the northern part of the terrace or lower city pro- site, its size must be reevaluated at 15-20 ha, rivaling duced Halaf and Ubaid pottery, and the full span the largest sites known for this culture (and com- of later periods through the Middle Ages until the parable to southeast of Kahramanmara?: Mongol destruction in A.D. 1263. The Middle Bronze see below). On the west slope of the main mound, Age is especially well represented. A similar occu- a narrow step trench investigated Late Chalcolithic pational sequence is suggested by the ceramics col- and EB sequences without reaching the underlying lected on the 6-ha mound (180 x 320 m) on which Ubaid. Occupation appears to have evolved here with- the fortress was built. Entrances to the lower city, out any interruptions. which overlooks a tributary of the Euphrates to the Investigations into the EB III Syro-Anatolian ur- north, are visible on the southern, eastern, and west- ban explosion that transformed Kazane into a 100-ha ern sides. For earlier surveys of the site, see H.H. city, twice the size of contemporary Titri?, were von der Osten, Explorations in Hittite Minor, 1929 carried out on the terraces east and southwest of (OIC 8, Chicago 1929) 76-77; A. Archi, P.E. Pecorella, the mound. In the eastern lower town, 300 m2 of an and M. Salvini, Gaziantep e la sua regione (Rome 1971); EB III monumental building were uncovered: its par- and P. Sanlaville ed., Holocene Settlement in North Syria tial plan has mudbrick rooms with well-plastered (BAR-IS 238, Oxford 1985). Continued survey and floors and interior hearths inside a 5-m-wide wall soundings took place in 1995. with stone foundations. This architecture contrasts KahramanmaraS Survey. Elizabeth Carter, Univer- sharply with an industrial quarter excavated in the sity of California, Los Angeles, spent a second two- southwestern outer town, where flimsy structures month season surveying the valleys south of MaraS. were erected on either side of a cobbled street. The She kindly reports: city was protected by a fortification system also "The 1994 season increased the survey area to ca. examined in 1994 on the eastern ridge of the lower 1,100 km2 and the previous year's inventory by 100 town. It consisted of a gravel rampart, 8 m high and new sites, for a current total of 227 spanning the an estimated 40 m wide, to support a mudbrick cir- Palaeolithic to the Medieval periods. The new sites 1996] ARCHAEOLOGY IN TURKEY 293

are situated for the jacentmost to low hilly part country thatin would the have provideddrain the Aksu and Erkenez Su to the southwest and south- excellent grazing lands and a ready source of stone east of Mara?. They underscore the general settle- and flint. Exceptional size (18 ha, and 8-10 m high), ment trends that were formulated in 1993. pottery of "true" eastern Halaf type, and large num- "Regional cultural patterns were determined bybers of stone bowl fragments and lithics lying on the two contrasting features. On the one hand, the rocky surface make it an excellent prospect for research outcrops that divide the valley channeled internal into the dynamics of this widespread northern Meso- communications and encouraged the development potamian culture dating ca. 5000 B.C." of local cultural features, especially noticeable in Excavations began at Domuztepe (site KM-97) in ceramic traditions independent of those outside this 1995. For the 1993 survey results, see Ara?ST 12 (1995) area. Secondly, two communication routes of great 331-41. interest promoted contacts with the outside, and eco- Sirkeli. The third season directed by Barthel nomic importance to the sites that controlled them: Hrouda at this eastern Cilician mound on the pres- the first on the river branches east and west of Mara?, ent west bank of the Ceyhan combined excavation the route into the Anatolian highlands; and the sec- with magnetic remote sensing, in an effort to assess ond along the Aksu, leading south and east from both the stratigraphic history and the extent of the Mara? toward Gaziantep and the Euphrates valley. ancient site. Deep soundings on the eastern and west- Thus the cultural and settlement patterns in many ern sides of the mound determined that it was first instances retained a strong local character, while settled on bedrock during the Chalcolithic period, reflecting developments outside the region. For ex- 5 m below the present surface. The Chalcolithic occu- ample, a sharp change occurred in the Late Chal- pation was well defined in the western trench, colithic period with the appearance of a largely with domestic architecture beside a road. Handmade, undecorated and mass-produced ceramic tradi- chaff-tempered pottery has a red- or yellow-slipped tion, distinct from that of the contemporary Uruk- surface with a black core; a few sherds were painted. affiliated sites to the southeast, but sharing the new A cylinder seal of Amuq G-H type, showing a fox techniques of manufacture and distribution. At the standing on a lion, was the only sign of an EB phase, same time, the number of sites expanded, with many although EB material was found in previous seasons. small regional centers connected to specialized sites. Middle Bronze architectural levels with typical A typical center southwest of Mara?, Ozen H6yiik, Cilician Painted ware were uncovered in the central is a 6-ha mound with six or more related sites nearby; part of the mound. Here, as elsewhere in the sound- pottery may have been mass-produced at one, andings, the evidence suggests a gap from the end of flints knapped at several others. the Middle Bronze until the Iron Age, when a large "In the Early Bronze Age, settlements again in-fortified settlement was established and maintained creased in number and spread out over all of the into the . Houses of this later phase valley's subareas, with at least one major urban center have been excavated both in the central and eastern at Danigmend Tepe (perhaps 50 ha), on the route trenches, and on the northern side where magnetic to the east with its comparable settlement history. prospection also located a large structure surrounded After a decline from the Middle Bronze through the by other buildings. Iron Ages, perhaps to be explained by a polarizing If the Late Bronze Age has so far proved elusive center at Mara?, the number of sites soared in Hel- on the mound proper, nonetheless another Hittite lenistic and Roman times, and tapered off by theroyal presence at Sirkeli was discovered in 1994. Once Middle Ages. Although Iron Age Mara? is inacces- vegetation was cleared from the foot of the cliff where sible because of the modern city, the discovery of the famous Muwatalli relief overlooks the river and an unfinished gate lion in a stone quarry overlook- the site, a second, smaller Great King appeared in ing the Aksu served as a reminder of the region's mirror image 13 m behind and to the north of activities in the Neo-Hittite period. Muwatalli's. H. Ehringhaus has published excellent "A second project of the 1994 survey was system- photographs and a short description in AntW 26:2 atic collection at the Halafsite ofDomuztepe, already (1995) 118-19. (For another royal relief at Keben, on recorded during the first season in the southeastern the western border of the Muwatalli's "Lower Land," corner of the survey area. The site commands a key see his article in AntW 26:4 [1995] 215-19.) pass in the hills along the route linking the eastern Kinet Hoyiik. A third excavation season under and western sections of the central valley, and also my direction took place in 1994 at this eastern lies at the head of an alluvial fan forming some of Cilician harbor, generally identified with classical the richest agricultural land in the region. It is ad- Issos in the northeastern corner of the Mediterra- 294 MARIE-HENRIETTE GATES [AJA 100

LA

- - - - - ...... i:~ iii: :;-:----~:--_~ii_. i-i'~Liiiii M-00. MEN-::

-ii~~i-i~i :~:-i--Ii__i::l: - :M- k M K ::?1?:?: -~i-:--- -:-i-

Fig. 11. Kinet Hoyiik. Late Bronze II storeroom with in nean. On the west (seaside) slope cylinder of sealthe in 26-m-highMiddle Assyrian style, representing mound, two separate operations a god began strangling investigating an ostrich (fig. 13), was found on the site's Late and Early Bronze the floor phases. of the western The laterroom; although the seal trench produced three rooms ofseems a towell-planned antedate the building Late by at least two cen- Bronze II (13th century B.C. ) turies,building it would with confirm a small an Assyrian presence at central court, and a storeroom Kinet. containing The building overliesseven an Ca-equally burnt monu- naanite amphoras and other large mental jarsMiddle on Iron orAge besidestructure that was partially their potstands (fig. 11). This level'sexposed inpottery 1993,. and isexpanded iden- in 1994. An unex- tical to the and pected repertoire, bonus of the seasonwith was the the conclusive dat- "drab ware" plates and potmarks ing of that the glacis characterize and fortification system to ca. 400 the late Hittite Empire. The building B.C., thanks towas Attic destroyed black-glazed imports and a late by hostile forces: bronze arrowheads basket-handled were amphora strewn in the context of a tower about, and a bronze axehead was found embedded and gate. in the courtyard wall. For the earlier history of the For the 1992 season, see KST 15:1 (1994) 193-200. site, a small trench at the foot of the 1992-1993 step S. Steadman presents the Chalcolithic and EB pot- sounding exposed what appears to be casemate wall- tery from the regional 1991 survey in AnatSt 44 (1994) ing preserved over 2 m high, and built of alternating 85-103. fieldstone and brick courses on top of another, simi- lar casemate system. The later version would date Eastern, Northern, and Central Anatolia to the EB I period, and may mark the western for- Sos Hiyiik/Erzurum. As part of the continuing tified edge of the site. Sherds from the casemates project to document northeastern Anatolia's cultural included sixth-millennium Neolithic wares com- history, Antonio Sagona and his collaborators at the parable to Mersin's. They suggest an early settlement Erzurum Museum began new excavations at Sos in the immediate vicinity. H6yfik, a small mound 25 km east of Erzurum on Other operations on the upper west side of the the route to Pasinler. Soundings carried out at the mound expanded the Middle and Late Iron siteAge in 1987, and cuts made by the overlying modern architectural sequence to eight building levels village,span- had already suggested that it could fill some ning the ninth to fourth centuries B.C. A large,of the gaps encountered in the Bayburt region, par- violently burnt building (fig. 12) proved to be ticularlythe for the second millennium B.C. source for ash tips containing Neo-Assyrian pottery The 1994 excavations uncovered stratified deposits in a 1993 trench to the northwest. A half-preserved dating to the Medieval, Iron Age, and Early Bronze 1996] ARCHAEOLOGY IN TURKEY 295

Nv.4 ...... IBM son. 7?71

mill REM

QW,

...... P3

ME

MEN, i

W; S' ONAX oil OF At- ME Emil,

. 'x

10 'am ...... VOW OWN ME t

10?

WA 01 9ft WWI

P1 ?j?i ?3?";.? pm gg I t o ! ......

0. 2251 CK ?i? . '.51

K&A

Fig. 12. Kinet H6yiik. Middle Iron Age building with Neo-Assyrian pottery. periods. In an upper operation, three 13th-century and P. Brennan in Ara4ST 12 (1995) 305-16. The re- A.D. levels of domestic housing were preceded port includes by a geochemical study of obsidian from a large Hellenistic building, the latter destroyed the survey in and the Biiyiiktepe excavations. a violent fire that nonetheless preserved a woodenSivas Survey. For A. Tuba Okse's long-term survey column capital in situ. Excavations down the project mound in the Sivas area, see Ara4ST 11 (1994) 243-58 slope exposed two levels of partial buildings and dating 12 (1995) 317-29. to the EB-MB transition at the close of the third Ikiztepe. mil- Onder Bilgi's long-term research at this lennium B.C. The latest phase of glossy black early Early metallurgical center continued in Transcaucasian Karaz ware was well represented 1994 with by excavations in several areas, including the bowls, cups, jars with geometric relief decoration, large sounding at the highest part of the mound. and trays with elaborate incised fronts unique More EBto architectural the levels of wood structures with Erzurum region. The lithic industry used elaborateobsidian hearths (source of the fires that regularly almost exclusively, and includes the tanged ravaged projectile the settlement) were uncovered below the points that are the hallmark of the period. Ties with the Caucasus are indicated by a class of graphite- burnished pottery known from the Martkhopi

kurgans northeast of Tbilisi; and with the Trialeti -,:-: -----...... i---in~ cultures by gray wares with stippled bands, and red- slipped wares painted with black chevrons. : ;::::::::: ...... ::: :--- ::::: :::--:--::::: iiiiioiiiofi Survey in the area located cemeteries of pithos ...... burials set in the center of stone circles. They were found to contain either skeletons or tomb gifts, and cannot at present be securely dated. Biiyuiiktepe Hyiik. The 1994 study season and an overview of the 1990-1992 excavations are sum- marized in KST 16:1 (1995) 161-64. Bayburt Survey. The final season of a four-year Fig. 13. Kinet Hoyiuk. Middle Iron survey in the Bayburt area is published by A. Sagona seal and impression. (Photo T. (ak 296 MARIE-HENRIETTE GATES [AJA 100

Iron Age and Hellenistic systems) levels. embarking One on the suchlong and slow building, route be- thought to be a temple, tween contained Assur and Kanesh, aand largeall for personal number gain. of finds including a crucible There were and many suchterracotta cities in the Old Assyrian female com- figurines. The figurines have mercial world,large and Kiultepeears is withmerely the multiple Anatolian piercings, and look like simplified version, he would argue-terracotta but in archaeological versions terms of the bronze and silver ones known from Alaca and it must be recognized as exceptional. Horoztepe. They are also found at Ikiztepe in tombs. The variations in house types that have been re- Reports on the 1992 and 1993 seasons are pub- vealed by an increasingly large exposure of level II lished in KST 15:1 (1994) 235-44 and 16:1 (1995) are providing more evidence for social distinctions 141-60; for analysis of metal objects, see ArkST 9 within the commercial district. In one elegant neigh- (1994) 77-99. Suggested reconstructions for the typi- borhood with straight, parallel streets and stone-lined cal Ikiztepe EB architecture are illustrated in KST sidewalks, the houses were planned with large reg- 15:2 (1994) 583-96. ular rooms enclosing flagstone-paved courts. Else- Acemh6yiik. Further excavation of the service where, more modest versions of the standard three- quarters northwest of the Hatipler palace, where unit house (office with archive, sitting room, court) Aliye Oztan has been investigating its Middle Bronze were appropriate for smaller entrepreneurs. A stratigraphic history, produced more of the level 3 cramped building with irregular rooms and school (Assyrian Colony period) architecture whose mud- tablets shows that insufficient funding for educational brick walls were reinforced near the floor level with establishments is indeed universal. a foundation course of juniper timbers. These These architecturally based social hierarchies can timbers, numbering 24 by the close of the 1994 be confirmed by the written accounts recorded in season, were recycled from burnt buildings dated the house archives. In 1994, 1,563 more tablets were to the EBA on dendrochronological grounds by P.I. added to the total of over 17,000 from formal exca- Kuniholm, and further confirmed by radiocarbon vations; with them came a harvest of 1,310 seal im- analysis. They span the years 2671 to 2169 B.C. ? pressions, some on envelopes not yet opened when 10, and have nothing to do with the date of the ser- the houses were destroyed by fire. One envelope had vice buildings themselves (a warning against indis- been stamped with two round seals in the Anatolian criminate reliance on scientific dating!). tradition beside an Old Babylonian cylinder sealing. The service quarters in level 3 were extensive. Their Another had been impressed with an Early Dynastic rooms and courts were equipped with ovens, and III heirloom, the property of a correspondent with trash pits with discarded vessels such as the white- antiquarian tastes. Were these commercial accounts slipped wares with brown-painted decorations known not sufficient to demonstrate the Assyrian commu- from Killtepe-Karum II. A large court with a circu- nity's business flair, then the jewelry that accom- lar hearth, and its adjoining room with a platform panied their burials would give irrefutable proof of produced a number of bone handles, pierced shells, their success: a wealth of gold and silver ornaments and a marble macehead-like tool. A fragmentary such as rings and melon-headed pins, and a pin with obsidian rhyton implies elegant tableware. In the a lapis head. earlier level 4, a structure with an internal platform Although there have been few contexts from the also contained quantities of bone tools, over 200 karum that suggest an interest in formal, structured knucklebones, and a large inventory of smashed religion, one room excavated in 1994 may have served ceramic vessels. Reports for the 1992 and 1993 sea- a religious function. It was furnished with a hearth sons can be found in KST 15:1 (1994) 245-55 and or podium and benches along the walls, and con- 16:1 (1995) 189-92; for the ongoing resistivity survey tained a quantity of miniature white-slipped vessels project conducted by M. Drahor, see ArkST 9 (1994) of assorted shapes decorated with brown paint, a 1-11 and 10 (1995) 229-43. pitcher with four parallel spouts, several pitchers with Kiiltepe. Despite more than 45 years of research pinched rims that poured at right angles to the at ancient Kanesh, principal city of Nesha in the MBA, handles, and more of the anthropomorphic "bread- the 1994 season yielded still unfamiliar finds and boxes" with yawning mouths. The season also pro- new data about the activities of 18th-century B.C. duced an incantation text, the third of this category Assyrian merchants in the Anatolian highlands. Di- among the entire tablet inventory. These precautions rector Tahsin Ozgiig, who has delighted in rediscov- apparently did not stave off widespread disaster, ering all aspects of these merchants' daily lives, re- which the excavator would attribute to local polit- minds us that these excavations allow a perspective ical turmoil. of the period on a human scale, with individuals (not For a recent publication on the seals, see B. Teissier, 1996] ARCHAEOLOGY IN TURKEY 297

Sealings and Seals on Texts of investigations from in the northeastern Kiiltepe corner of theKaru (Uitgaven van het Nederlands city at Biyiikkaya, the largest Historisch-A of Hattusha's citadels. gisch Instituut te Istanbul By the close of the season,70, almost Istanbul all of the Hittite 1 Karahoyiik-Konya. Anfortification illustrated circuit had been uncovered. It wassummar built 1992 campaign appears in the standard in casemate KST fashion with15:1 regular pro-(1994 together with a plan jecting of towers, trench and on the north, X's a monumental level gate 1 Colony period) residential (again the city's largest) andfacing Yazlllkaya. religious Inside the where Sedat Alp has citadel, been soundings excavating were first made in the area durin of a several seasons. Two largebeams rectangular from building noted, adjacent along with two ro have been ring-dated others, by in a resistivityP.I. survey.Kuniholm Finds consisted of sev- to 1782 vv B.C. ? 37. He eral imperial has seals also and sealings, dated a hammer-seal ajuni of from the earlier level 6/7 to 2181 B.C. ? 10. Killtepe Karum type, tablets, and many miniature Kaman-Kaleh6yiik. Sachihiro Omura's 1994 cam- vessels just inside the northern gate. When Biiyiik- paign at the impressive mound that commands the kaya was remodeled during the Phrygian period, a direct route from Ankara to Kayseri, and from Hat- system of water channels was installed in connection tusha to the great Salt Lake, brought increased evi- with the gate and a roadway leading up to a district dence for the site's participation in Hittite and of small workshops. The earliest Iron Age settlement Phrygian affairs. In conjunction with the northern at Bogazk6y seems to have been located here, as had stratigraphic sounding, a broader exposure of the already been suggested by excavations in the 1950s. Late Hittite (Kaman III) phase uncovered more build- The pottery, all of it handmade, includes a painted ings with excellent brick architecture, thick plastered ware that could be considered a prototype for the walls and floors, and cobbled courtyards, all violently Phrygian animal style with deer (see I. Bayburtluoglu burnt and filled with fallen timbers. In situ finds in- in AA 1994, 324-25). cluded superbly burnished Hittite pottery, relief A new computer-generated topographic plan of ceramics, bronze vessels and weapons, stamp seal- the entire site is underway-a project estimated to impressed jar stoppers, 20 Hittite hieroglyphic Old take two years - in order to produce a more detailed Kingdom sealings and one stone seal with hiero- map of Bogazk6y's irregular and challenging terrain. glyphs, a crystal stamp seal representing an eagle, Reports on the 1992 and 1993 seasons appear in KST and a faience scarab. This city was protected by a 15:1 (1994) 293-308 and 16:1 (1995) 199-217; for 1994, circular fortification constructed in standard Hittite see AA 1995 (forthcoming). Seeher has published casemate fashion. It did not withstand the assault a retrospective of his predecessor Peter Neve's direc- that brought this phase to an end, however, trapping torship in BiblArch 58:2 (1995) 63-67. The inscribed 21 residents (so far!) whose skeletons were found bronze sword discovered outside the Lion Gate is among the ruins. placed in a typological context by A. Miiller-Karpe, Concurrent excavations of the Early Iron Age levels "Anatolische Bronzeschwerter und Siidosteuropa," (Kaman II), again on the northern side, uncovered Festschriftfiir Otto-Herman Frey zum 65. Geburtstag (Mar- several rooms of what may be an Early Phrygian burger Studien zur Vor- und Friihgeschichte 16, 1994) official building constructed of wood on stone 431-44; see also O. Hansen, "A Mycenaean Sword foundations. Underlying this ninth-century B.C. archi- from Bogazk6y- found in 1991," BSA 89 (1994) tectural phase are the handmade black-burnished 213-15; M. Salvini and L. Vanetti, "Una spada di tipo ware and pit houses that characterize the arrival of egeo da Bogazk6y," PP 276 (1994) 215-36; and H.-G. the Phrygians at Kaman, , and elsewhere: Buchholz, "Eine hethitische Schwertweihung,"Journal sunken floors reinforced in the corners with stones, of Prehistoric Religion 8 (1994) 21-41. and walls marked by postholes. Alaca Hiiyiik. P. Neve reconsiders the chronolog- Preliminary reports for the 1992 and 1993 seasons ical setting of the Sphinx Gate in N. Cholidis et al. are published in KST 15:1 (1994) 273-92 and 16:1 eds., Beschreiben und Deuten in der Archiologie des Alten (1995) 313-30, and S. Omura ed., Kaman-Kaleh'yiik Orients: Festschrift fiir Ruth Mayer-Opificius (Miinster 4 (Japan: Middle Eastern Cultural Center 1995). For 1994) 213-26. For textual analogies relating to the the annual regional surveys that are placing the site reliefs, see A. UInal in AnatSt 44 (1994) 207-18. in its cultural and environmental setting, see AraSST Ortakoy-Corum. Excavations conducted by Aygiil 11 (1994) 311-36 and 12 (1995) 215-44. Sijel in 1994 at this 8.5-ha imperial Hittite center, Bogazk6y. With the 1994 season, Jfirgen Seeher 50 km northeast of Bogazk6y, uncovered more of assumed the direction of excavations at the Hittite the monumental complex now called Building A, capital, and continued the previous year's program and extended to the upper citadel and a neighbor- 298 MARIE-HENRIETTE GATES [AJA 100

ing field. The northern basepart of the ofhill's eastern,Building southern, and A western functio slopes as a kitchen and service (fig. wing. 14). The site Alongwas therefore muchthe bigger building and more west side, which was followed complex than previously for 55noted. m, a sequen of corridors and cobbled courts filled with brick col- "The Hittite contributions to the site should now lapse from an upper story produced 250 tablets and be reevaluated in several key aspects. The basalt one bulla. The violent fire that burned the building, Cyclopean structure behind and north of the reliefs and bronze weapons-a dagger and spearheads- on the outcrop's summit extended west beyond von found within the debris all speak of catastrophic der Osten's plan and down the slope, to enclose a hostilities. Wood samples without bark now give a much larger platform or terrace than in the 1930 dendrochronological terminus ante quem of 1304 version. Scattered basalt architectural elements such vv B.C. ? 37 for the building's construction. Research as a column or statue base and a possible lintel block on the (now forested) upper citadel, which was sepa- suggest that the platform supported a monumental rated from Building A by a series of terraces, suc- building of the Hittite period instead of function- ceeded in tracing parts of its Hittite fortification sys- ing only as a retaining structure for the corbeled tem as well as a cobbled road. In contrast, soundings chamber. Local reports of postern gates at the site in the outlying field produced rubble walls of a later would indicate further constructions around the hill. period, perhaps connected to the Roman cemetery On the other hand, the tower and ceremonial ramp that overlies the Hittite site. that von der Osten identified as a monumental Further study of the more than 2,000 tablets re- approach, on the southern and southeastern slope covered here confirm that it was part of a Hurrian below the reliefs, may well need to be redated to the geographical and cultural region, with the Hurrian Phrygian period: their assembly, and the use of lime- language used for most of the archival records and stone blocks, match the stonework of the Iron Age omen texts. They also suggest that Ortak6y can now fortification enclosure that crowns the summit. be identified as ancient Sapinuwa. "The Middle and Late Phrygian reoccupation at Aligar-Konaksu valley. Imminent flooding of the Ga^vurkalesi, previously judged to have been a small Konaksu valley because of dam construction led hillfort, can also be extended well beyond the sum- Ronald L. Gorny and his 1994 team to intensify their mit thanks to the coherent system of radiating ter- regional survey of the Ali?ar area, both on the ground race walls that were planned along the slopes in 1994. and from the air with balloon photography. Sound- The Phrygian site would thus have measured approxi- ings were conducted 20 km northwest of Aligar within mately 1,000 m east-west and 600 m north-south the flood zone at the large site of adir H6yilk, which to cover both the valley floor, its sloping borders, like its more famous neighbor was occupied from and the high hill proper. Access to the citadel in its the Chalcolithic to the Late Roman periods. Notable northeastern corner, as suggested by von der Osten, findings include EB I obsidian debitage and tools, can now be substantiated by massive walling appro- violently burnt walling that could represent an Iron priate for a defensive tower or fortified gate. There Age or Hellenistic fortification system, and a Roman may also have been another entrance on the north- structure on the top. The site appears to have ex- west, via a ramp leading up to a smaller gate. The panded significantly toward the south in the Hel- Phrygian settlement, like its Hittite predecessor, has lenistic period. For a summary of the project, see become far more sophisticated and monumental BiblArch 58:1 (1995) 52-54. The 1994 season's prelim- than could have been anticipated before the start inary report appears in Anatolica 21 (1995) 65-100. of this survey." Gavurkalesi. In 1994, Steven Lumsden, Bilkent For the 1993 season, see Ara?ST 12 (1995) 267-80. University/Copenhagen-Carsten Niebuhr Institute, Kinlk-Kastamonu. A hoard of over 30 Hittite continued his intensive survey of the hill dominated metal vessels, among them three bull rhyta and a by the Hittite reliefs and masonry that were first bowl with figural reliefs and a hieroglyphic inscrip- studied by von der Osten in 1930. He kindly reports: tion, was brought to the Kastamonu Museum in 1990, "The second season's primary goal was to plan all after being unearthed in a gravel quarry to the north visible surface architecture, a project that was facil- of Kastamonu on a direct route to the Black Sea. itated by the year's exceptionally dry climate and lack In 1994, Aykut (lnaroglu began a survey and exca- of vegetation. In addition to the architectural fea- vation project at the site, now named Kinik after its tures recorded by von der Osten - the Cyclopean nearest village. Despite bulldozing and large-scale structure, the summit's enclosure walls, and the towers earth removal, the season produced more metalwork and walls directly below the reliefs - it was also dis- as well as stratified deposits. covered that walls could be traced all the way to the Initial surface collection on the site's steep slope 1996] ARCHAEOLOGY IN TURKEY 299

(KY / .\I .9 N K ?/"Q/

/ Cyc / Ken os e ~ -, 7 10Reke.

I I I

0 100 200

Fig. 14. G

netted a dedsev in the who Kas and omphSham least ofthree bu at the Fests slop levels camp asso very andburn i shops. shoul Hig well totop a do The wellK. Em w horns, tite an V layers Ozgi of s sisted (1990 only jects, discu and trum Festsc spe spacer Kugakli. The second, 1994 seasonbea at the 18-ha vessel Hittite center located onin a high plateau, 60 km southt a warrior of Sivas near the village of Ba?6ren, again attested a gold to the success of remote sensingshi in directing the re- urine. search targets of AndreasThe Milller-Karpe and his team. in contrast Geoelectrical/resistivity maps have now produced characteri clear plans of a city gate to the north, the casemate contest fortification wall with regularly spacedwi towers enclos- The inginscri the site, and more large structures to the west 300 MARIE-HENRIETTE GATES [AJA 100 of and parallel to the North Terrace temple exca in 1993. Excavations in three separate areas pr equally remarkable. Continued work on the 54- North Terrace temple showed that the buildin seen occasional repairs. Behind its twin cellas, known to include the typical Bogazk6y temple n in their back walls, a small room contained 61 b impressed with Early-Middle Hittite-style belonging to, among others, a Suppiluliuma Samili. Of particular interest is a hieroglyphic ing naming a local king Malmi-ZiTi-ma (fig. 15) building seems otherwise to have been strippe all contents before it was burnt, although one produced a cache of 12 superb wing-tipped arr heads. Investigations outside the temple, again southeastern wall, revealed an annex with two r tangular stone-lined baths supplied (or drained a water channel made of fitted terracotta pipe stratigraphic probe against the temple's we corner indicates that it was built above an Old Hit- tite level of secular structures that included a work- Fig. 15. Kugakli. Sealing with hieroglyphic royal name. shop. After the temple was abandoned, a third phase (Courtesy A. Miiller-Karpe) of rubble walls was constructed over the immediate area, although not over the temple itself. On the western side of the acropolis, which domi- violent destruction dated by red-burnished wares and nates the central and eastern part of the city, exca- a hammer seal to the end of the Hittite Empire. Like vations exposed more of the residential district that so many other Hittite fortified hill sites, Porsuk was was continuously occupied from the Old Kingdom reoccupied, after a hiatus, in the Early Iron Age. The through the Hittite Empire, and-like the temple -new settlement established itself within the walled violently burnt. A new trench on the top of the acrop- enclosure, and apparently specialized in ironwork- olis uncovered the archive of another large Hittite ing. A faience figurine of an Egyptian seated god- building (24 x 35 m), unfortunately damaged by an dess from this period indicates that the site main- Iron Age reoccupation. The archive has so far yielded tained its importance as a link with the coast. 43 tablets, among them six texts that discuss the Investigations were also pursued on the top of the festivals of a city named Sarissa, and the comings mound, where a third-century A.D. cemetery over- and goings of its king. Kugakli can therefore be iden- lies a Roman town of the first-second centuries A.D. tified with this well-known Hittite cult center, whose The occasional cross among the burials signals the patron weather god is invoked in the Qadesh treaty presence of an early Christian community. A report between Hattusili III and Ramses II. by C. Abadie-Reynal on the 1992 cemetery excava- For a brief 1994 season report, see Orient-Express tions appears in KST 15:2 (1994) 211-23. 1995:1, 3-4; a preliminary report on the survey andKilise Tepe. A second route between the Hittite first two seasons (1992-1994) appears in MDOG 127heartland and the Mediterranean followed the dra- (1995) 5-36, with an epigraphical report by maticG. G6ksu valley, which links the Konya plain with Wilhelm. western Cilicia- the "Lower Lands" that were con- Porsuk. Excavations at the Hittite citadel that con- trolled in the 13th century B.C. by the elusive capital trolled the valley between Ulukila and the Cilician Tarhunta''a. Surveys byJ. Mellaart (Belleten 22 [1958] Gates, the main pass from the plateau down to 311-45), D. French (AnatSt 15 [1965] 177-201), and, Kizzuwatna and the sea, were reactivated in 1994 more recently, the Silifke Museum located Roman under the direction of Olivier Pelon. Work continued and especially Byzantine sites in the valley, and a on the site's massive fortification system, of which concentration of prehistoric ones in the region of a sloping 3-m-wide corridor has now been followed Mut (Roman ). The construction of a for a length of 31 m, where it opened onto a large major dam on the lower G6ksu will, when completed, cobbled courtyard. Quantities of burnt wood beams flood a number of these, including gingene Tepe, were strewn in the corridor debris, attesting to a Maltepe, and Kilise Tepe (misnamed Maltepe in the 19961 ARCHAEOLOGY IN TURKEY 301

surveys) on opposite carved bankson the cliff face at Keben,of in thethe lower G6ksu river c south of Mut. In 1994, valley and salvage25 km inland from Silifke, excavations see H. Ehring- b Kilise Tepe under the haus, AntWdirection 26:4 (1995) 215-19. ofJ. Nichol gate, Cambridge University, who kindly off following report: Western and Coastal Anatolia "Kilise Tepe sits on a Kanlhgegit. high Mehmet promontory Ozdogan and Hermann Par- of conglomerate to the zinger east expanded theirof Thracian and project overlook in the Goksu River, near a crossingKirklareli region to Kanligegit, point where confirmea few sound- presence of a second ings wereprehistoric carried out in 1994. Although mound the site has ( Maltepe) on the other suffered side from of plowing, the and road valley. and railway con- A go at the mound's northern struction, it nonethelessend preservesand key other stratified ma- spr the immediate region terial. ensured Troy IV pottery was that recovered here,the the first area able for settlement andattestation agriculture in the region for an EB III connectionin antiq indeed today. Pottery between collected the Balkan cultures systematicaand the . Just the surface of the below100 this level,x 100a deposit ofm pre-Cucuteni site and rang Chalcolithic to Early Karanovo Iron V Balkan Age material types, again underscored includ the tite burnished wares. fourth-millennium The site gap observed was so far also throughout occu the Hellenistic and Byzantine these sites; but it also brings periods, the onset of this inter- with tine material and Alahan monastic ware' on the north ruption down by a few centuries. side suggesting the location of the church that gave Hacilartepe. Excavations in the periphery of the mound its name. Ilipinar (see above, under "Later Neolithic: Ilipinar") "Two large step trenches were opened on the were carried out here for a second year byJ.J. Rooden- mound's eastern and northwestern sides. The east- berg in an effort to find the settlements that used ern trench exposed, for a length of 25 m, a massive Ilipinar as a cemetery during the Chalcolithic and stone wall running north-south along the edge EB of periods. The Hacilartepe soundings again un- the slope to form a high, fortified enclosure on coveredtop modest wattle-and-daub/half-timber housing, of the Bronze Age mound. An Early Iron Age date a revival of the local Neolithic tradition. Abundant is suggested by architectural levels inside, or west, black-burnished pottery is in the EB style of the Troad of the fortification wall, which pitting suggests hadand Demircihiiyiik, but moldmade in local clay rather gone out of use by classical times. The deep north- than coiled. Since the connection between this set- west trench produced a stratified sequence of second- tlement and the Ilipinar EB cemetery 300 m distant millennium B.C. occupations capped by Iron Age is now secure, there is good hope for a Chalcolithic clay-lined pits for grain storage, and finally Byzan- predecessor in reserve for a future season. For the tine walling. Its Bronze Age levels were all destroyed first (1992) campaign, see KST 15:1 (1994) 171. The in violent fires, and consequently contained excel- pottery is analyzed by A. van As et al., in Newsletter lent floor deposits. For the two LB architectural levels, of the Department of Pottery Technology (Leiden Univer- close ties with the Hittite Empire are indicated by sity) 11-12 (1993-1994) 54-73. fine red- and buff-burnished wares, lentoid flasks, Demircihiiyiik. The small finds from the libation arms, and a stone seal of bulla type with 1975-1978 excavations are published by A. Baykal- the name of the official Satuwali. They were preceded Seeher andJ. Obladen-Kauder in M. Korfmann ed., by two Middle Bronze phases: the later with walling, Demircihiiyiik IV: Die Kleinfunde (Mainz 1995). a courtyard, floors with pottery typical of the Ana- Troy. The seventh season of the new Troia project tolian plateau, stamped terracotta crescents, and maintained a its momentum with an average team of hearth containing at least five burnt cervid horn 70 working for over two months in an area of 1,905 cores; and the earlier a mudbrick building with inte- m2. One major focus of the 1994 campaign con- rior benches and vessels datable to the EB/MB tran- cerned the Bronze Age site's urban development, as sition. Although it is at present difficult to speculate kindly reported here by director Manfred Korfmann, on Kilise Tepe's political status during the Middle Tiibingen University: Bronze and Iron Ages, one can already conclude from "The postulate that Troy's fortification enclosure the 1994 excavations that during the later second expanded steadily from period to period through- millennium, the site functioned as an administrative out the Bronze Age was reexamined in 1994, and center and no doubt military garrison of the Hittite can now be confidently refuted. It has become clear kings." that the first major fortification systems of mid-late For a recent article discussing the Hittite relief Troy I and II (2700/2600-2450/2400 B.C.), which were 302 MARIE-HENRIETTE GATES [AJA 100

significantly rebuilt and Bronze restored Age excavations at at Troyleast are presentedeight below tim served as the basis for underthe "Classical, walls Hellenistic, of and TroyRoman: Troad." III, inde an observation already Miletos. made A research by program Schliemann on Bronze Age Dorpfeld. The short-lived Miletos, Troy with new excavationsIII citadel as well as study thus and me added modifications to publication its ofimmediate the finds from earlier excavations,predeces In contrast, the enclosures began in 1994. ofIts main Troy goal is to characterize IV and Minoan V m have been located well beyond and Mycenaean influencethe and/orperimeter presence in Bronze of T VI's fortress, which overlies domestic architecture Age coastal Ionia. Wolf-Dietrich Niemeier, Heidel- wherever soundings have been opened. The Troy IV berg University, generously reports on his first period (close of the third millennium) appears to campaign: have suffered from considerable unrest: it includes "At the start of the six weeks' season, excavation six burnt phases, and the faunal remains suggest beganan on the Stadium Hill, where P. Hommel claimed exceptional reliance on hunting, and perhaps an im- to have uncovered a Mycenaean megaron in the poverished economy. House types followed the'Ana- early 1970s. Hommel's trench was cleaned, and new tolian settlement system' noted at Demircihilyiuk, with trenches were opened to its south. In the walls of long two-room structures sharing party walls, and the supposed Mycenaean megaron we identified each containing a domed oven immediately inside reused Archaic building material together with the entranceway. Remains from Troy V cap these sherds in of Archaic pottery. Thus the building must all areas, and therefore would have matched the ex- be redated, probably after the Persian destruction panded plan established in IV. With Troy VI (1700- of 494 B.C. In the southern trenches, we first came 1250 B.C.) and the establishment of a 2-ha lower town down on the ruins of the village of Eski Balat, then extending 400 m south of the mound's South Gate, on settlement debris of the third-fourth centuries the fortifications around the citadel were built and A.D. overlying bedrock. Stadium Hill, therefore, was replanned at least three, and more probably four, not occupied at any point during the Bronze Age. times. Structures in the lower town and on the cita- "Investigations were then shifted to the area of del show a complete disregard for preexisting re- the Temple, where Bronze Age remains had mains (whether burials or architecture), perhaps be- been revealed during campaigns early in this cen- cause their builders were newcomers to the site. The tury, and again in 1938 and the 1950s-1960s. We succeeding Troy VII, whose minimum of four phases opened a series of trenches underneath and south span the close of the Late Bronze and the beginning of the east-west Roman road south of the sanctuary. of -the Iron Age (1250-1040/1000 B.C.), marks In a two trenches, remnants of the Roman road were dramatic contraction in the city's fortunes. There- sitting directly on the stone foundations of a Late fore, the 'Onion Model' for the development of Troy Geometric/Early Archaic house; it is hoped that a can now be discarded in favor of more diachronic stratigraphic sequence of Dark Age and Late Bronze fluctuations in its urban configuration. levels will be recovered below this house in a future "The 1994 program of restoration consolidated season. In most of the other trenches, fill from the more of the Troy VI citadel walls and Hellenistic earlier excavations accounted for over 2 m of de- additions west of the South Gate, and began work posit. At the bottom of three, however, we finally in the area of the Northwest Bastion. A new walkway reached an undisturbed Bronze Age destruction level was built above the Troy II ramp, to divert traffic awaywith masses of burnt mudbrick debris, charcoal, and from the ancient access to the citadel. Hope for greatthe quantities of local Minoan domestic pottery long-term protection of the site and its setting, how-such as conical cups, tripod cooking pots, and bridge- ever, rests with the creation at Troy of a national park,spoutedjars (fig. 16). Fine decorated pottery imported for which we continue to make every possible effort." from dates this destruction to the transition Interim reports on the 1993 season have appeared from LM IB to LM II (ca. 1500 B.C.). The deposits in Studia Troica 4 (1994) 1-50 and KST 16:1 (1995) also contained fragments of elegant wall paintings 239-62; for 1994 in Studia Troica 5 (forthcoming). (fig. 17). The water table encountered at this level Frank Calvert's pioneering work at is cham- prevented us from proceeding further. With a pump pioned by M. Robinson in AnatSt 44 (1994) 153-68; installation next year, we hope to excavate this phase and by S.H. Allen in AJA 99 (1995) 379-407 andmore extensively, and especially to recover architec- Archaeology 48:3 (1995) 50-57. tural remains." The Troy treasure in the Pushkin Museum is dis- The post-Bronze Age excavations at Miletos are cussed in my introductory commentary. For Kumpresented below under "Classical, Hellenistic, and Tepe, see above, under "Later Neolithic"; the post-Roman: Ionia." 1996] ARCHAEOLOGY IN TURKEY 303

Fig. 16. Miletos. Local Minoan domestic

Liman Tepe. Geomorphological results of Hayat Erkanal's continued research archaeological Liman Tepe area has findings now in the previousdefined seasons' trenches, the where theBro topographical setting 1994 that campaign focusedensured on the Middle the and Earlysite' tion as a harbor: it wasBronze occupations. originally The three parallel on apsidal a build-prom flanked by two deep ings bays, excavated inwhich 1993 as level 4,erosion now to be dated g filled to form the present early in the MB shallow period, were supplemented inlet. by Aerian- graphs also revealed otheran two enclosure identical structures again extending oriented paral- water beyond and parallel lel and east-west; to they suggestthe the EB presence bastion of many dication that the settlement more aligned in organized was rows. muchMatt-painted warelarg its current situation would indicate. of Lerna V type, although not from a secure context, This information adds a further dimension to the nonetheless indicates contacts beyond the Aegean. The underlying EB-MB transitional level 5, exposed for the first time in 1994, in contrast bore no sign of neat planning. Its houses were irregularly laid out, with cobbled floors, ovens, and surrounding pits. Finds include a charming terracotta bear. Elsewhere, broad exposure of an EB III level has revealed the massiverubble walls of large structures perhaps comparable to the Argolid "corridor house" plan, one of them approaching palatial scale. Pottery re- covered in large quantities from a storeroom includes one Troy II red-burnished vessel and sauceboats. Chalcolithic pottery was also recovered in excavations near the EB fortification wall at the southern perim- eter of the site, perhaps in conjunction with an earlier defensive system across the inland side of the promontory. Preliminary reports appear in KST 15:1 (1994) 361-73 and 16:1 (1995) 263-79. The Urla municipality has also produced an attractive bilingual English- Turkish brochure about the excavations. Panaztepe. The 1994 excavations were carried out by Armagan Erkanal in three separate areas at the Fig. 17. Miletos. Minoan fresco fragments: white lily on red eastern base of the acropolis, to investigate the ground. (Reconstruction B. Niemeier) sequence and nature of settlements for the cemeteries 304 MARIE-HENRIETTE GATES [AJA 100 on which previous campaigns concentrated. Two the trenches confirmed that the Bronze Age site reoccupied only in the Byzantine and Ottom periods, when the region revived its former econo prosperity. The third trench exposed an LBI IIIA-B ashlar building with several regularly plann rooms. Finds included , a g Minyan fenestrated fruitstand, and a stone mold.J outside it was buried a 35-year-old woman, laid o on her back and accompanied by modest gifts (ter cotta beads or spindle whorls). Skeletal analy indicates that she died from violently inflic wounds. The fourth period represented at Panazte a long Middle Bronze Age sequence, produced intriguing structure, perhaps a cult room: one of walls was interrupted by a niche faced with an upri slab or orthostat, in association with an overturned Fig. 18. Uluburun. Beads. (Photo L. Ray Martin) cup and a burnt vessel filled with animal bones. Pottery suggests an EB-MB transitional date for this phase, as well as antecedents for the gray ware that Red Sea shell rings inlaid with red glass; ivory later developed into the full-blown LB type. The scepters and their pomegranate-shaped finials (fig. 19); crescent pendants in what seems to be a tin-lead excavator proposes that this site may provide a key alloy; bronze Bes pendants; an amethyst scarab blank to the chronological and geographical distribution and a second possible example in a black stone; and, of Minyan ware. For brief reports on the 1992 and 1993 seasons, see KST 15:1 (1994) 461-66 and 16:1 as raw materials, a complete ostrich egg, sections of (1995) 281-85. elephant tusks, hippopotamus tusks and canines, and more glass ingots. Cypriot export ceramics were Ephesos. A Mycenaean level in the Artemision found spilled out in the fill as well as inside a trans- sounding is reported below under "Classical, Hel- lenistic, and Roman: Ionia." port jar (fig. 20), the 10th such container from the site. Although different from the others because Ka4-Uluburun. The 1994 season at the LB ship- of its two ribbed handles and short stature, this wreck off the coast of Ka? brought to completion, (Cypriot?) pithos illustrated the usual merchandise: after 11 years, the longest and deepest excavation two White Slip II milkbowls packed inside a cari- project yet to be undertaken by the Institute of nated bowl, fragments of White Shaved juglets, the Nautical Archaeology. Cemal Pulak kindly offers the following commentary on his concluding campaign: most popular Cypriot type on board, and samples "Recovery and mapping projects, requiring a of every organic substance already attested. More exceptional were two short lengths of rope and bits season of nearly three months, completed our under- water investigations of the Late Bronze merchant of twine, bone game pieces, and a beautifully carved vessel that ran aground against the rocky Uluburun promontory toward the close of the 14th century B.C. The first involved excavating the last remaining cargo in the central part of the site, where an exception- ally deep deposit had collected in a gully sealed by the lowest layer of copper ingots (raised in 1993) and elements of the ship's hull. The finds ran the entire spectrum recovered from elsewhere on the wreck. Organic materials sieved from the sand matrix in- cluded pomegranate, fig, grape, and olive; herbs and spices (coriander and sumac); pulses and wild grasses; and nuts, such as almond and pine. Beads in a rich variety of materials, shapes, and sizes (fig. 18: faience [ribbed], agate, ostrich eggshell, quartz, steatite, amber, and chalcedony) increased the previous col- lection by the thousands. Other goods included Fig. 19. Uluburun. Ivory finials. (Photo L. Ray Martin) 1996] ARCHAEOLOGY IN TURKEY 305

Fig. 20. Uluburun. Large Fig. 21.pithos Uluburun. with Cyprio Ship's (Photo R. Piercy) wood lid for an oval hull remains;box. and onResting reconstructing, by against means of der of the pithos was a three-dimensional, one leaf computerized from topographic a mapdip tical in appearance of the to site, the the stages by whichcomplete the ship broke apart e covered in 1988, but and came missingto rest on the steeply itssloping oceanivory floor. second leaf. Sections of the hull at the deepest end of the wreck "Personal equipment for the crew still suggests preserveda an important construction feature not pre- Syro-Palestinian origin for the ship: the galley wares, viously recovered: perhaps the approach to the bow, for example, and most of the tools, weapons, and as indicated by a slight taper produced by a flat scarf items such as pan-balance weights (numbering over or drop strake. Here were also found a section of 130: two pans in a wooden case were found in 1993) the keel-plank and portions of six alternating wide and saucer lamps (the Cypriot lamps, in contrast, and narrow starboard strakes. In the area of the were unused stock for sale). Two more cylinder seals, withies excavated in 1993, several pieces of lighter- one of them in Mittani common style, also support weight wood may prove to be blades from the ship's this identification. There were, however, some non- sweeps, or oars from the ship's tender (fig. 21). Be- Levantine personal vessels and seals -among them cause the wreck's structural wood has warped into the shapes of the seabed, information from the topo- a second Mycenaean lentoid seal found in 1994-- that could point to an international legation (commer- graphical plan in conjunction with the remains them- cial attaches?) bearing royal gifts, rather than simply selves will ultimately make it possible to determine traveling salesmen of different nationalities shipping the ship's appearance and cargo placement when luxury goods as part of their inventory. it set sail on its last journey. "The second goal of the season focused on the ex- "Dating the ship's final voyage more precisely may cavation, planning, and complete recovery of all the depend in large part on dendrochronological analy- 306 MARIE-HENRIETTE GATES [AJA 100 sis. So far, the wood used be used withfor caution, its however); construction and in a Turkish trans- is gesting the last quarter lation of in Bilimthe ve 14thTeknik 336 century(1995) 26-29. B.C. it is hoped that additional samples collected in IRON AGE including the keel-plank, will refine the date fur This will provide cross-dating Eastern Anatolia of great significa for the Uluburun artifacts, Anzaf. Progresswhose in recovering conservation the plan of the a ninth-century B.C. Urartian military complex built study-- after this lengthy fieldwork-- now become project's focus." by King Ishpuini to protect a strategic crossroad east Preliminary reports for of Lake 1992-1993Van was assisted by aerial are photographs publish that in KST 15:1 (1994) 375-97 revealed additional and features 16:1 of the lower(1995) fortress's 219- for the 1994 season, and enclosure the circuit. 1993-1994 Oktay Belli was regional able to identify s wreck surveys, see INA and Quarterly excavate a large bastion 21:4 on (1994)the northern 8-16 side, article on LB organic where cargoes a monumental bygate was C. protected Haldane by an inte- appeared in WorldArch rior tower.24 Wall(1993) foundations 348-60. of dense rubble sup- ported well-preserved mudbrick superstructures, METALLURGY their faces carefully plastered. Just inside the gate, Goltepe and Kestel (Ni'de). K. Asllhan hundreds Yener's of animal bones were lying in heaps on progress report on the 1993 season at top the of EBtwo andiron axes. Interim reports on this site, Iron Age mining center of G6ltepe and thewhich neigh- must have served in a double capacity as tax boring Kestel mines appears in KST 16:1collection (1995) center, appear in KST 15:1 (1994) 417-44 177-88. No excavations were undertaken in 1994. B. and 16:1 (1995) 287-312. Recent installments in the Earl's most recent replication experiments for smelt- long-term documentation of the region's Urartian ing tin are presented in The Oriental Institute News and dams and sophisticated water supply system can be Notes 146 (1995) 1-5. For a general study on tin mining found in Ara?ST 12 (1995) 353-83, and TelAviv 21:1 in Anatolia, see E. Kaptan, "Tin and Ancient Mining (1994) 77-116. Fourteen dams, some of them marked in Turkey," Anatolica 21 (1995) 197-204. with royal inscriptions, have now been located to the north, northeast, east, and southeast of Van. DENDROCHRONOLOGY Ayanis Kalesi. Altan Qilingiroglu's 1994 campaign Peter Kuniholm and his team collected over 500 at the seventh-century B.C. fortress founded by Rusa samples in 1994 for analysis at the Carolyn and IIMal- continued work in the previous seasons' excava- colm Wiener Laboratory for Aegean and Near East- tion areas. The circuit wall's superb andesite masonry ern Dendrochronology, whose tree-ring sequences has been exposed along the entire eastern side and now span over 6,000 years from both sides of the Adri- much of the southeastern side, revealing details such atic to Syria. Newly calculated dates relevant as to vertical channels to help drainage at the corner specific Anatolian sites discussed in this newsletter joints of the blocks. Proceeding further inside the have been included in the site summaries. Of monumentalpar- southeastern gate, where Rusa II's build- ticular interest for second-millennium absolute dat- ing inscription was discovered in 1993, it was found ing is a sample from Porsuk with an anomalous ring that the back had, at a later stage, been blocked that Kuniholm attributes to the climatic disturbances with brickwork, and a much smaller door had been brought on by the eruption in 1628 B.C. This pierced on the side to control traffic more effectively. sample would obviate the Mesopotamian "High Chro- On the upper citadel, more of the great pillared nology," and perhaps favor the "Low Chronology." hall was cleared: there were three rows of massive He presents the 1992 and 1993 project results in orthostat-lined piers to support the high ceiling (the ArkST 9 (1994) 281-91, and a general assessment in mudbrick walls are preserved to a height of 4 m). AJA 99 (1995) 99-102. Quantities of collapsed pine roof beams, many with mortises that should allow the reconstruction of the ARCHAEOBOTANY, PALYNOLOGY, PALAEOZOOLOGY roofing system, still suggest a dendrochronological Discussions of archaeobotany (N. Miller), palynol- date in the middle of both the century and Rusa's ogy (S. Bottema), and palaeozoology (H. Hongo and reign. On the west side, the plan of the great ware- R.H. Meadow), with particular reference to Anatolian house ("pithos building") was expanded. A sophis- archaeology, can be found in P.E. McGovern, "Science ticated drainage system was discovered under the in Archaeology: A Review," AJA 99 (1995) 91-99. M. long cobbled corridor that separates the large pithos Nesbitt's excellent summary of archaeobotany, agri- storeroom from the narrow wing of magazines. It culture, and diet in ancient Anatolia appears in consists of tapering terracotta pipes fitted end to BiblArch 58:2 (1995) 68-81 (the map on p. 79 should end and sealed with lead; a T-shaped pipe also gave 1996] ARCHAEOLOGY IN TURKEY 307 access into the drain. Inside the storeroom, brick Cilicia steps gave access either to an upper floor or high and Domuztepe. In 1994 Halet Qambel shelving. To the numerous bullae found here can continued her program of restorations on the now be added an Urartian tablet, the 33rd known Karatepe citadel; and excavations at Domuztepe, so far. A second storage structure with 2-m-tall pithoi where the street leading inside the fortress gate was surrounded by small, standard-sized handlelessjars followed further into the Iron Age residential dis- is also being uncovered underlying a Medieval ceme- trict investigated in previous seasons. The very burnt tery near the pillared building. houses here contain a full selection of kitchen equip- A progress report on the 1992 season is presented ment. Evidence for corbeled vaulting to roof the gate in KST 15:1 (1994) 445-56. Qilingiroglu discusses dec- was also discovered. orated stone vessels from Ayanis in TelAviv 21:1 (1994) The 1991/1992 report (KST 15:1 [1994] 309-23) in- 68-76. For an assessment of Rusa II's building pro- cludes photographs of the restored monumental gram, see P. Zimansky, '"An Urartian Ozymandias," statue of a storm god that was reerected in 1991 be- BiblArch 58:2 (1995) 94-100. side Karatepe's southern gate. Karagiindiiz-Van. The salvage project of previous Kinet H6yiik. See above, under "Chalcolithic and seasons to document the Urartian necropolis on Bronze Age, Southeastern Anatolia and Cilicia." Ergek Lake's eastern shore, 35 km north-northeast of Van, was reoriented by Veli Sevin in 1994 to the Central and Western Anatolia mound proper, when a rise in water level isolated Gordion. The 1994 season, under the guidance it on a promontory and indeed threatens to sub- of G. Kenneth Sams, again involved a broad range merge it. Three soundings produced a long sequence of issues, from conservation to stratigraphy and set- of occupational deposits. Immediately below the Me- tlement variations, ethnoarchaeology, and publica- dieval cemetery that crowns the site, the stone foun- tion projects. After a decade of disappointing ex- dations of a large building were found together with periments on mudbrick preservation, it was finally red ceramics decorated in black paint- the first time decided to rebury Megaron 1, and replicate it on that this ware, known from Hasanlu IIIB, can be asso- a platform in the same location. The stone walls of ciated with an architectural context in this region. Terrace Building IV were rerestored by replacing the Deeper in the soundings, a mixed level with Cauca- earlier mortar capping with a rusticated mortar and sian Haftavan VIB painted wares signals a second- stone layer. These techniques will be applied to other millennium occupation. The lowest phase reached structures. during the season contained EB Transcaucasian Excavations directed by Mary M. Voigt were ca graphite wares like those from Sos Hoyilk, together out in three separate areas. Gordion's later resi with red- and black-burnished Karaz pottery, but no tial phases were investigated on the western sid associated architecture. the citadel mound, in the 1950 T Trenches. Late Excavations were also carried out at the newly Phrygian walling was found to have been constructed created shoreline a few kilometers to the east of the on top of the usual 3.5-m-thick clay seal; but the clay mound, where a low 20-m-long oval feature in a field had been laid down over rubble packing or terrac- of tumuli proved to be a massive rectangular brick ing rather than over burnt buildings, as in the center platform, its plastered floor painted with brown, blue, of the mound. This architectural level was followed and pink designs outlined in black. This structure, by two substantial phases of Hellenistic housing with although burnt, produced no finds beyond a terra- elegant ashlar foundations, the stones occasionally cotta spindle whorl and some iron nails. It recalls recycled by the less impressive Galatian occupation the temple at Altintepe, and may have functioned that followed. A Roman house marks the end of the as a ceremonial center for the surrounding ceme- sequence. During the latter periods, the saddle be- tery. The tombs date to the early Iron Agelearly tween the citadel mound and the "Kiiik H6yik" Urartian period, and illustrate fully a pottery reper- was used as a cemetery. The Roman cremation and toire distinct from the Urmian version of this culture. inhumation burials excavated there in 1994 included The 1992-1993 seasons at the Karagfinduiz necrop- an adult female wearing, around her neck, a large olis are summarized in KST 16:1 (1995) 331-50. For bronze signet ring wrapped with string (a memento C.A. Burney's soundings on the mound in the 1950s, from someone with much thicker fingers?). Its en- see AnatSt 8 (1958) 157-209. graved bezel is decorated with a fisherman bran- imikugail. Veli Sevin has published the first final dishing his catch (fig. 22). The Galatian settlers put report on the first-millennium phases of his salvage this area to a related but altogether more dramatic excavations in the Malatya area as Imikuya4t I (Ankara use for ritual sacrifices of animals (equids and 1995). bovids), women, and children. Skeletons- some de- 308 MARIE-HENRIETTE GATES [AJA 100

Fig. 22. Gordion. Bronze signet ring from Roman burial. (Courtesy M.M. Voigt)

capitated, some with broken necks-were found Fig. 23. Gordion. Lower Town. M strewn throughout the excavated plaque. (Courtesy area; they M.M. had Voigt) evi- dently been left exposed rather than buried in pits. On the western side of theIron Age fortifiedsite, site overlookingeast of , 1,400 masl in the present course of the Sakarya, the northern a Cappadocianstep highlands,trench with adown com- the side of the mound and two plete soundingsphotographic record taken near from both athe hot-air river uncovered stratified remains balloon andof a helium Middle blimp. Topographic and plans especially made Late Phrygian date, followed on the groundby are an being Achaemenidcoordinated with the photo- phase that includes imported graphic Iranian documentation pottery. to produce a computer- A non- domestic structure spanning assisted planimetric the map. The Middle layout of the entire to Late Phrygian periods was filled 375-ha site with(2.5 x 1.5 km) arrowheads can already be assessed and, among other finds, an incised for specific structural ivory features plaque and overall plan. showing He a running warrior with a kindly sword reports: and shield (fig. 23). The houses on this side were "Oblique large shots from and the blimp well (fig. 24), appointed which with the Attic and Lydian proved pottery more successful inthat highlighting characterizes architectural the city's taste for refined remains tableware. than the vertical pictures Imported needed for map- prod- ucts continued to flood in from the west in Hellenistic ping, defined with more precision an important com- times as well, now documented by research on the plex on the southern side of the site near one of its amphoras that transported wine and oil from Rhodes, many gates. Here, a network of streets led to a tri- Miletos, Mendes, Thasos, and Chios to Gordion. angular open area, perhaps a plaza, faced by a monu- For the 1992 and 1993 season reports, see KST 15:1 mental building set on a pair of huge rubble terraces (1994) 467-79 and 16:1 (1995) 369-92. R.C. Hendrick- and reached by a ramp or staircase. The streets were son discusses the Hittite ceramic industry at Gordion also flanked by large enclosures. Natural topography in BiblArch 58:2 (1995) 82-90; the Iron Age pottery determined the placement of the entire settlement's in D. French and A. Qilingiroglu eds., Anatolian Iron streets and buildings, which were not oriented ac- Ages 3 (London 1994) 95-129; and Gordion's Bronze cording to an imposed grid. Likewise, the massive and Iron Age ceramic traditions in W.D. Kingery ed., defensive system, and the location of its gates followed Social and Cultural Contexts of New Ceramic Technologies the site's natural contours. Nonetheless, the site ap- (Westerville 1993) 89-176. The 1988-1989 seasons' pears to have been created as a coherent building stratigraphic soundings are published by M.M. Voigt program that included elaborate provisions for sup- in Anatolian Iron Ages 3, 265-93. Volume II has ap- plying water. The survey suggests that construction peared in The Gordion Excavations (1950-1973), Final was not yet finished when an intense fire destroyed Reports series edited by G.K. Sams: E.L. Kohler, The the city, damaging far more buildings than previ- Lesser Phrygian Tumuli, Pt. 1: The Inhumations (Phila- ously realized. Tumuli were later erected both inside delphia 1995). the abandoned site and in the surrounding region, Da'. Geoffrey Summers, Middle East a burial tradition now recognized as spanning sev- Technical University, Ankara, concluded his second eral periods and a variety of types, all different from season of aerial and geophysical survey at this Late the Phrygian type with timber chambers. 1996] ARCHAEOLOGY IN TURKEY 309

Fig. 24. Kerkenes Dag. Center of city, looking

'An extramural temple, its associated structures 25on the centralm2 ridge, withfollowed f walls at times preserved a strict orthogonal 3 systemm withhigh, only minor wasconces- id m to the north of sionsthe to irregularities fortified in the terrain. enclosure temporary with it. The Closer investigation masonry of the fortification includes walls re- tresses bonded into vealed the the presence walls, of 80 small a towers, structura many con- also used for attaching nected with the minor entrances,fortification in counterpoint to the to city wall at the main major gatessite. located Theon all four temple'ssides. Soundings in in although obscured two at separate the areas of theback, central ridge alsocontain allowed three rooms, and seems a closer assessment to of itshave architectural featuresbeen than orien the east." aerial photography had previously afforded. A unit The 1993 interim report on the site, which the of eight double cells, typical of the architectural excavator identifies with the Median fortress , blocks on the southeastern side of the complex, was appears in Ara?ST 12 (1995) 567-82; for 1994, see completely cleared. The blocks were organized in AnatSt 44 (1994) 15 and British Institute ofArchaeology/ two rows of four cells sharing a common back wall, Ankara Research Reports 1994, 18-20. but without communicating doorways: each cell was Giilliida'. Wulf Schirmer carried out a third entered through a single door that opened onto the season of aerial photography, mapping, and sound- street. The second type of unit, with communicat- ings in 1994 at the second Cappadocian Iron Age ing cells arranged on two sides of a courtyard, was fortified hill site that is currently being studied. sampled to the northeast of the complex in the long Gllfidag also appears to have been laid out as a block parallel to the city wall. In both areas the rooms unified building program, perhaps under the were bare of any pottery or finds, but must have auspices of the kings of Tabal, in the late eighth cen- served as (temporary?) housing. Since the entrances tury B.C. Unlike at Kerkenes Dag, however, natural to the central complex were once decorated with lion topography determined only the placement of the sculptures (now in the Nikde Museum) and the main fortification walls, which enclose a 150-ha area on city gate was also flanked with lions (taken to the the ridges of a volcanic crater 2,000 masl. The city's Kayseri Museum by R.O. Arlk), the entire city with buildings, especially the monumental complex and its monumental architecture and associated housing 310 MARIE-HENRIETTE GATES [AJA 100 must have served a special, the tourism industryformal, and archaeological and non-m research. function. The 1993 season is discussed in KST 15:2 (1994) G ivurkalesi, Kaman-Kaleh6yiik. For the Iron Age 243-52. phases, see above, under "Chalcolithic and Bronze . The 1994 season in the Pisidian high- Age: Eastern, Northern, and Central Anatolia." lands successfully expanded the impressive scale of . For the Phrygian phase, see below, under the four previous campaigns thanks to an even larger "Classical, Hellenistic, and Roman: ." interdisciplinary crew engaged in excavation, con- Sarh6yiik-Dorylaion. A. Muhibbe Darga has sum- servation, survey, ethnography, and the study of marized the results of her first four seasons (1989- ancient diets, export networks, and the changing land- 1992) at the Bronze Age, Phrygian, and later site on scape. Marc Waelkens, Catholic University, Louvain, the outskirts of Eskigehir in KST 15:1 (1994) 481-501, kindly reports: and the 1993 campaign in KST 16:1 (1995) 351-67. "Excavations to the north of the Late Hellenistic fountain in the area of the Roman library were CLASSICAL, HELLENISTIC, AND ROMAN carried out with the dual purpose of investigating Pamphylia, Pisidia its still active water source and preparing the monu- Perge. Excavations under the direction of Haluk ment's restoration. A sixth-century A.D. water supply Abbasoglu continued for a third season to investi- system was discovered, as well as the Late Hellenis- gate the Late Roman and Byzantine residential dis- tic and Roman ones that it adapted. The Byzantine trict immediately inside the city gate, where insulae version consisted of a series of settling tanks feed- were laid out with four houses per block. To the ingwest three terracotta pipes cut through the fountain's of the earlier campaigns' trenches was found a well-back wall. The water was then transferred west into preserved two-storied house organized around a large a preexisting vaulted service tunnel 1.5 m high, and courtyard, with mosaic floors and an oil press. Finds from there into an enormous tank near the south- include quantities of glass, such as sixth/seventh cen-western corner of the fountain house. The tunnel tury A.D. lamps with suspension loops similar was to itself replacing a Late Hellenistic predecessor those shown hanging above shop fronts in illumi- with a corbeled roof built of fieldstones that we fol- nated Byzantine manuscripts. Since the houses were lowed for 37 m before reaching a collapsed section. built directly on bedrock, it is now clear that this A terracotta channel in its floor would have assisted district reflects a final expansion of the city in Latein directing the flow of water, ultimately to the Upper Roman times (third century A.D.), and continued . Access to the tunnel through a slab in the prosperity throughout the Early Byzantine period, roof was blocked in the early third century A.D. when until the Arab raids in the seventh and eighth the slopes around the fountainhouse were leveled centuries. and paved to form an esplanade, requiring a rework- Efforts to locate traces of pre-Classical settlements ing of the water distribution system. in the area prompted renewed survey on the acrop- "A second discovery stemming from preparations olis. Although surface clearing uncovered primarily for the restoration of the library occurred to its west, Hellenistic remains-walling and paved roads lead- where the previous season had uncovered a large ing to the lower city- sporadic finds give glimpses room, suggesting that the library originally formed of earlier periods: a flexed burial with chipped stone part of a building complex such as a gymnasium. arrowheads as grave goods; sherds of sixth-century This interpretation must now be reassessed. When Attic black-figure; and a fifth-century lion-headed this area was opened up in 1994, the large room was drainspout reused in a Late Roman wall. Other found to be one of a row of shops with mortared efforts focused on contemporary problems, in par- rubble walls and brick stringcourses dating to the ticular the protection of the necropolis from illegal second quarter of the second century A.D., probably digging. For a 1991 progress report, see KST 15:2 at the same time as the library proper. They had gone (1994) 597-611. out of use by the library's last restoration during the Side. The restoration program on the theater and reign of the emperorJulian, but (as described in last the adjacent agora entered its second decade under year's report) served briefly as residential housing the direction of Ulkii Izmirligil. Progress into theafter the fire that destroyed the library in the later diazoma of the theater has located internal staircases fourth century, and until the Isaurian raids of A.D. that will permit an evaluation of how spectators cir- 404-406. culated throughout the building. Other traffic flows -"The excavations this season in the Upper Agora notably of today's tourists through the ancient site moved - on to the sequence of structures that delim- are causing the archaeologists far more difficulties, ited its northern 30-m-wide edge, and cleared its west- and underscore the potential for conflict between ern half. The earliest building there, still standing 1996] ARCHAEOLOGY IN TURKEY 311

Fig. 25. Sagalassos. North side of Upper Agora

7 m high, was designed with a plain facad stories supporting an elaborately decor floor, most probably in the shape of a Dor Structural features and associated "proto- ware" would suggest a construction date in quarter of the first century B.C. The facad backed by a thick layer of Roman concre floor level raised considerably some time i ond century, perhaps in connection with struction in front of it of a large nymphaeu that became the focus of the Agora's north the Antonine period (A.D. 160-180). The n um's aediculated theater facade, one story set on a 1.50-m-tall podium in front of a basin with a capacity of 170 m3. The f articulated with six projecting podia on e of five niches; a waterfall would originally caded down the semicircular central on Monumental marble sculptures were recov the four niches and podia excavated in 1994 end toward the center, a Dionysos and sat Aphrodisian marble (figs. 27-28); a fragme torso; a complete (fig. 29) app brought in from elsewhere and for whom was modified; and the feet of an wi on the base for a bronze statuette. The ny was reworked on several occasions over the follow-

ing 400 years. By the first half of the sixth century, Fig. 26. Sagalassos. North side of Upper Agora, nymphaeum it had been divided over two terraces into small facade, northwestern half. (Courtesy M. Waelkens) 312 MARIE-HENRIETTE GATES [AJA 100

Fig. 28. Sagalassos. North side of Upper Agora, nym- phaeum: detail of Dionysos head. (Courtesy M. Waelkens)

Polykleitan tradition, and positioned to overlook the plain where he engaged the Pisidians in battle. "The third year of large-scale efforts in the Lower Agora included the initiation of work on the 840 m2 bath complex that was erected against its north- eastern corner in the second century A.D. Over 300 blocks that had collapsed into the agora were re- Fig. 27. Sagalassos. North moved side to revealof partUpper of the baths' Agora: western facade Diony (fig. and satyr group, from 32),the which nymphaeum. was revetted with large (Courtesyashlars against a Waelkens) mortar and rubble core with brick string courses. A sounding inside the wall indicates that it is 4 m rooms for storage or workshops thick, suitable for supportingthat canan enormous be assobrick ated with two street levels. Coins indicate that the vault that is still partly standing. Restoration in the rooms were not abandoned before A.D. 640. agora proper completed repositioning the pavement "Northwest of the Upper Agora, excavations began slabs and honorific monuments over the entire ex- anew northeast of the Kakasbos temple, in the place cavated area. A monumental gate dating to the reign where R. Fleischer discovered eight frieze slabs with of Tiberius was also discovered in the agora's south- a chain of nearly life-size dancing girls over 20 years eastern corner: a single row of Corinthian columns ago. The monument to which they belong proved with lateral podia supported an entablature carved to be a tall square socle, 7.5 m on a side and overwith garlands and theatrical masks, and a finely 2.5 m high (its base was not reached in 1994). Three carved cornice. It was modified ca. A.D. 400, when more slabs (fig. 30) representing another four danc- a broad stairway was built outside it leading to the ing girls were recovered, as well as evidence for city's re- main north-south street. The structure col- pairs to the frieze during the Roman period (fig. lapsed 31: in the earthquake of A.D. 518 or 528, trap- Antonine head for the east corner dancer). The mon- ping several victims whose skeletons were found lying ument itself, however, can be identified as a Middle in the debris. Hellenistic heroon, thanks to associated fragments "Survey of the district recorded 19 new sites, rang- from a nearly 4-m-tall marble sculpture of a Hel- ing from the late Palaeolithic and Neolithic periods lenistic ruler. The head allows us to recognize him to Byzantine times. These include two Hellenistic for- as a youthful , portrayed in tifiedthe cities (fig. 33: Kepez Kalesi, perhaps ancient 1996] ARCHAEOLOGY IN TURKEY 313

lassos ware is presente 15:2 (1994) 408-18. Balboura, . R 1993 results of these lo byJ.J. Coulton on the neighboring Pisidian t 429-36 and 12 (1995) 49 8-10, 129-48. L. Biers at Balboura in AnatSt research at Oenoanda, s stitute of Archaeology/ 20-22. A parallel invest with a focus on its p carried out in 1993 by 483-96). . A final repo published by S. Mitche Ancient City in Peace

Lycia . In 1994, excavations were continued by Jiirgen Borchhardt and his team on the ninth and tenth residential terraces located in the northern part of the site, where a system of plastered water channels was traced. The Roman phase was also in- vestigated with soundings in the area of the porti- coed street, and with a study of the Roman fort. The season produced six more fragments of the frieze with a chariot race and Persian war chariots that orna- mented the Ptolemaion monument (see AJA 98 [1994] 264 fig. 9), an honorific type to which re- fers. Restoration of the colossal female statue that once stood in this building has been completed. Con- servation efforts have now turned to the Byzantine episcopal church, and to the necropolis. A more dra- matic rescue involved the heroon, which was boxed up and transported from its high and isolated loca- tion down to the site museum by helicopter, re- enacting, in Borchhardt's words, the feats of "a mod- ern Pegasus for Bellerophon." Reports on the 1992 and 1993 seasons appear in KST 15:2 (1994) 257-77 and 16:2 (1995) 229-51. Fig. 29. Sagalassos. North side of Upper Agora Nemesis (ht. 2.10 m), from Letoon-. the No large-scalenymphaeum. excavations took (Co Waelkens) place during the 1994 season, which was devoted to soundings, survey, conservation, and study under the Paleopolis or Darsa); directiona Hellenistic of Christian Le Roy. In the fortress;Letoon, how- large settlements and ever, importantan isolated new evidence for the sanctuarylayout of the Hellenistic-Early Imperial temenos did emerge from date." soundings, demonstrating A new display of finds that its enclosurefrom wall extended the 70 m1990-1995 beyond the together with the restored propylon. The size ofNemesis the sanctuary can nowstatue, be set up in the Burdur doubled Museum. from previous estimates, A and thegeneral gate and r pears in Archaeology the three48:3 temples (1995) can be relocated 28-34,from its south- with graph of the Alexander ern side to itshead; center. Excavations for in the thestage and 1992 seasons, see KST 15:2 orchestra (1994) building of the373-407 Letoon theater began and in 16 175-90, and AnatSt 44 the summer (1994) of 1995. 167-86. A study 314 MARIE-HENRIETTE GATES [AJA 100

Fig. 30. Sagalassos. Hellenistic heroon: newly di (Courtesy M. Waelkens)

Renewed investigations bronze in the crosses area of among the necrop the olis, which was damaged of in gabled recent Lycian forest sarcopha fires, covered two large and elegant period, Late and Antiquewere able (fifth to rec tury A.D.) villas with peristyles,blocks from fine an mosaics,Early Imp a architectural sculpture. entirely In another recycled successful into a Byzsu vey venture, the Hellenistic complete and inscription, Roman aqued 5.5 which supplied the Xanthos 60 m long,bath complex,was carved was in fG lowed back from its terminus (only andfour across are missing).three brid to a mountain spring 8 cated km awayto the near emperor 0avdirk6y Clau Its small spring building first was known fed by military a long govertunn with an interior walkway mous and interest channel. is theThe list hint of land network of peak fortresses to the other linked towns, by an here ext sive wall system was also in the districtdocumented under Veranius's administration. further. The Ceramic research focused inscription is beingon published the bylate Sencer Sahin. ampho For trade, and on the Medieval the 1992 andhandmade 1993 interim reports pottery, with many illus- wh clays have been traced trations,to sources see KST 15:2 (1994) in 279-301 the and 16:2immedi (1995) vicinity and near . 253-82. The long-term restora project at Xanthos proper . has Excavations had directedto extend by Cevdet Bay- its tention to the more recent burtluoglu damageconcentrated in inflicted1994 on the late resi- by insensitive municipality's dential phases roadworkof the site. On the central around terrace, be- Antiochus Gate, and satellite side the honorific dishes funerary installedmonument opposite in the an above the Byzantine monastic library, was found complex a public latrine furnishedon the with up acropolis. An interim report a room for a fee-collector.on the To its1991 west, the and usual se- 19 seasons appears in KST quence 15:2 of Late Byzantine(1994) squatters' 303-23. housing suc- . Fahri I?lk and ceededhis earlier collaborators Byzantine levels with more continue prosperous the previous seasons' research residences; so tooand on the restorationsouthern terrace, where of Roman baths, the Byzantine the lower level basilica was represented (whose by an Early Byzan- plan sembles one at Perge), tineand villa withthe geometric ancient floor mosaics. city's Restorations ma street; they also excavated were carried outByzantine in various parts of thegraves site. Reports w 1996] ARCHAEOLOGY IN TURKEY 315

Fig. 31. Sagalassos. Hellenistic heroon: Roman repair (sec- ond century A.D.) for the eastern corner figure of the frieze. (Courtesy M. Waelkens)

Fig. 32. Sagalassos. Partial view of a Roman bath building are published in KST 15:2 (1994) 253-55 and 16:2 beside the Lower Agora, at close of the 1994 season. (1995) 217-27. Inscriptional (Courtesy materialM. Waelkens) appears in a volume edited by S. 5ahin: Die Inschriften von Arykanda (Inschriften griechischer Stfidte aus Kleinasien 48, Bonn 1994). Klzilbel and Karabel, and the Iron Age Bayindir Hacimusalar-Elmah. A new excavation project tumuli. A survey in 1993 and soundings in 1994 pro- under the direction of ilknur Ozgen was inaugu- duced evidence for its later Hellenistic, Roman, and rated in 1994 at the largest mound (300 x 300 m, Byzantine phases, together with earlier pottery from and 13 m high) in the central Lycian plateau, 20 km mixed fill. Reused marble architectural and sculp- tural fragments, numerous inscriptions, and high- southwest of Elmall near Ak:ay. The site was orig- inally located on the shores of a large lake (recently quality imported ceramics reflect a well-connected drained), where it could control the major passage and sophisticated community during the Iron Age from the plateau south to the sea, as well as east and and classical periods. Excavations continued in 1995. west across - indeed, the role today enjoyed Kyaneai. Reports giving the 1992 and 1993 results by nearby Elmall for a route that the surrounding of this long-term survey are presented by E Kolb in mountains have kept unchanged since antiquity. Ara4ST 12 (1995) 83-98. For Balboura and Oenoanda, Cedar forests, excellent hunting and fishing, and the see above, under "Pamphylia, Pisidia." agricultural richness of a well-watered plain ensured the site's prosperity from Late Neolithic/Early Chal- Caria colithic times to the Byzantine period, and its de- . In an effort to rediscover more of the velopment into the area's only major ancient city. It honorific monuments that once stood in the public has tentatively been identified with Byzantine . district near the harbor at Kaunos, Cengiz Isik turned This project represents the first attempt to inves- in 1994 to the round peristyle building excavated tigate ancient urbanism in this plain. It is thus in- in 1969 beside the harbor stoa. He suggests that this tended to supplement and follow upon previous elegant structure of fine marble may be connected archaeological research here - notably at Bronze Age with an inscribed entablature block, reused in the Karata?-Semayiik, the Archaic painted tombs at later stoa, and mentioning a heroon dedicated by 316 MARIE-HENRIETTE GATES [AJA 100

Fig. 33. Kepez Kalesi, Pisidia. Hellenistic city w

the Kaunians in thanks covered to from the a plastered people cistern could ofpush theirRome. A monolithic inscribed pillar construction datefound back into the behindfifth century. For the sto may also have been moved 1991-1993 from seasons, see KSTa platform15:2 (1994) 161-88 and adjoin- ing a building further north, 16:2 (1995) 297-314.where a series of niche and socles would have displayed Keramos. The exceptionally statues well-preserved of monu- promi nent citizens. Ipik would ments stillplace visible among a thebase modern housesfor at Oren,a bronz statue on top of the pillar. on the north For side ofhis the G6kova reconstruction bay, were recorded of the Protogenes monument, in 1993 by Marcello and Spanu, other whose preliminary excavations re- at the site, see KST 15:2 port(1994) appears in Ara?ST325-39 12 (1995) 133-44.and He 16:2 suggests (1995 319-25. The fountain house that the structure in known the locally Harbor as "Bakicak," and Agora is discussed in a final report: previously identifiedC. Iplk, as a temple, Kaunos should be under- Araltzr malarz 11.2: Liman Agorasz stood Celme as an elaborate Binasz nymphaeum (Ankara of perhaps 1994) Burgaz-Datga. The history Trajanic dateof and this western (Italian)ancient construction. harbor was extended back into the Bodrum-Halikarnassos. Late Geometric Poul Pedersen provides period by Numan Tuna's 1994 excavations interim reports on the 1992 in and 1993the seasons, southeast with ern sector of the site, beside commentary the on the Mausolosmodern palace soundings, shoreline. in The southwestern trenches KST 15:2 (1994) exposed 135-47 and 16:2 (1995)more 327-34. of late Greek and Roman architectural . In addition levels to continuing from salvage excava- the pre- vious season. A chamber tomb associated with the tions in the northeastern necropolis around its upper domestic architectural phase (second-third prominent Roman mausoleum ("Clock Tower"), and century A.D.) was discovered intact, its stone door restoring to its original elegance the site's most elab- still in place. Inside were found 30 individuals, orateand funerary monument ("Fish Market") to serve numerous funerary gifts (lamps, pitchers, smallas a museum, Fede Berti and her colleagues reopened amphoras) placed on a niched shelf that lined excavations the in the agora and on the acropolis. Sound- walls. Soundings on the small acropolis indicate ings in the agora below the Hadrianic northeastern severe erosional loss. The first season is reported portico in revealed an earlier Doric version with KST 16:2 (1995) 283-95. unfluted columns, and an associated pavement with . Ramazan Ozgan devoted the 1994 season Roman inscriptions and graffiti. To the north of this, to restoration of the frescoed Hellenistic/Early fine isodomic masonry from a Hellenistic structure Roman house from the terrace beside the Aphro- and an associated pavement suggested that the drain- dite temple; and to the excavation of more of the age system here was originally installed at the same small one-room shops or public facilities built along time, rather than in the Hadrianic period as previ- three terraced streets on the promontory overlook- ously proposed. An opus sectile floor with poly- ing the south harbor. Late black-figure pottery re- chrome geometric designs could belong to the Hel- 1996] ARCHAEOLOGY IN TURKEY 317

lenistic structure, since are suchnow reemerging floors from the midstwere of olive groves.attested by the second century These activities haveB.C.; alerted thievesbut to thethe site's in-area also much reworked, and remained in use into the terest. In 1994, theater reliefs were removed with Byzantine lth century. heavy equipment. It is fortunate that they were then Below the Hellenistic level were family tombs abandoned along the 6demi? highway, and eventu- whose burial gifts include late fifth-century B.C. ally recovered. A progress report for 1992 can be pottery. It is not yet clear whether the tombs should found in KST 15:2 (1994) 115-20. be linked symbolically with the agora, or belong to . The program of urban study, map- a period when the area enjoyed a different function, ping, inventory, restoration, and regional survey con- since Iron Age burials were also excavated here in ducted by R.R.R. Smith and C. Ratte in 1994 main- earlier campaigns. Research was also carried out on tained the energetic standards of the previous season, the acropolis temple. For the 1992 and 1993 seasons, which had already succeeded in transforming our see KST 15:2 (1994) 125-34 and 16:2 (1995) 335-39. perception of this city's configuration and its rela- Related regional surveys conducted by Eugenio La tion to the surrounding landscape. Rocca are summarized in Ara?ST 11 (1994) 169-89. Excavations again concentrated on establishing . The marble sculptures from the more clearly the city's grid. The area behind (north Hekatomnid sanctuary have been published by of) the was cleared to document the Ann C. Gunter, Labraunda. Swedish Excavations and shift in orientation from the temple to Researches 11:5 (Jonsered 1995). the new street plan of the Late Hellenistic (first Stratonikeia. Limited excavations were carried out century B.C.) period. It was found that the transition by Yusuf Boysal in 1994. Reports on the 1992 and was achieved at the northern row of chambers, whose 1993 campaigns appear in KST 15:2 (1994) 121-23 front and back facades run parallel with the two and 16:2 (1995) 315-17. different alignments. A sideline of this study was the . In a second season directed by Ahmet reexamination of the Late Antique sculptor's work- Tirpan, excavations on the southeastern side of the shop excavated in the 1960s, whose many practice site outside the propylon of the sanctuary of Hekate pieces have recently been inventoried. The north- uncovered a further stretch of the road leading to south axial street of the new urban layout was also Stratonikeia, and bordered with tombs from the sixth discovered in 1994, in a sounding 30 m north of the century B.C. to the Byzantine period. Numerous in- tetrapylon (later incorporated into the Byzantine scriptions on the orthostats lining the road as it nears triple-apsed or Triconch church), where the east- the gate read XOPOE AEONTOE. The Doric stoa west axial street was traced in 1993. Aphrodisias has that frames the eastern side of the sanctuary was also thus recovered the two major streets that determined partially cleared: its columns and entablature col- its alignment. A further segment of the northern lapsed in place when the sanctuary was violently street leading to the southern agora was found at burnt. In the area of the temple proper, enough the western end of the Great Basilica. A Hadrianic entablature fragments were recovered to show the inscription referring to taxes on nails was reused work of five separate craftsmen. The season also here as a paving stone, placed (deliberately?) face produced a number of small fragments to supple- up. A blue marble base for an equestrian monument ment the temple frieze now in the Istanbul Archae- can be associated with a blue marble horse found ological Museum. Other sculptural finds include part here in the 1970s. of a round altar with a relief showing three dancing Mapping extended in 1994 to the stadium, a plan women, one holding a lyre, reminiscent of the maiden of which was drawn for the first time. Inscriptions frieze from the Sagalassos heroon (supra, fig. 30). on the seats record the names of guild members, city Outside the sanctuary and to the east, the sacred officials, and visitors from nearby towns. The sta- pool or circular nymphaeum was excavated to its dium's eastern end was transformed into an amphi- depth of 1.5 m, still preserving some of its travertine theater in Late Antique times, when an arcade was revetments and three of the columns that once stood erected along the top. The inventory of spolia, statues on its perimeter. Water was piped in from a source and their bases, and sarcophagi added several to the south. The pool had been repaired on at least hundred more numbers to the collection, with sar- two occasions. A still unpublished inscription from cophagi accounting for over 500 of the total. These here, found one century ago and reused in a nearby include 12 newly excavated by the local museum; one house, is said to relate to the pool's construction. of the sarcophagi illustrates two men on either side Nysa. Vedat Idil has spent several seasons exca- of a furnace, perhaps glassblowers or smiths. All of vating and restoring the town where was edu- the sculpture previously kept in the excavation com- cated, so that many of its Roman public buildings pound was moved into a secure, newly designed depot 318 MARIE-HENRIETTE GATES [AJA 100

that will also function as aother, workshop more general shift for in focusmarble has led to con- in- servation. The need for creasedsuch efforts protection at placing the site within had its diachronic been confirmed by the recent regionaltheft and ofenvironmental a head context. from Volkmar the von Sebasteion reliefs; it was recoveredGraeve, Ruhr-Universitdit (from Bochum, a kindly dealer) reports: and restored to its proper "Interdisciplinary place at projects the took precedenceend of in 1994 the season. over excavation, which - beyond the Bronze Age pro- As part of the ongoing regional gram survey directed by project, W.-D. Niemeier-was limited to the circuit walls, inner redoubt, and stadium of salvage work in the Aphrodite sanctuary. Archaeo- -on-the-Meander were mapped. A full pre- botanical and archaeofaunal projects, begun in 1992, liminary report of the 1994 campaign appears in are investigating the resources available to Milesians AJA 100 (1996) 5-33; the 1993 season is published in the Archaic and later periods. This study will com- in AJA 99 (1995) 33-58, and KST 16:2 (1995) 191-206. bine forces with the archaeological survey, now in R.R.R. Smith presents the corrected reconstruction its fourth year, of the 200-km2 area around the site of the Zoilos monument in Aphrodisias I: The Monu- to discover the ecological setting that made it pros- ment of C. Julius Zoilos (Mainz 1993). per and the extent to which it was exploited. In the Herakleia am Latmos. In 1994 Anneliese Peschlow- same spirit of inquiry, the Roman water supply system Bindokat continued to conduct her long-term sur- is being retraced via its canals and reservoirs. vey project in the environs of Herakleia, in recent "Geophysical surveys of the site have also produced seasons tracing the routes that connected the citysubsurface information about the city enclosure to sanctuaries and forts in the BeSparmak/Mt. wall's extension toward and around the Archaic resi- Latmos range and ultimately across to the northeast dential district of Kalabaktepe. The results, suggest- and northwest. These sites, like Herakleia, owe their ing a larger circuit than previously thought, will be remarkable preservation to their remoteness; testedand with soundings in 1995. Similar efforts in the their fine masonry, fortifications, and temples toBay of Lions have identified another large structure Hekatomnid ambitions. They also succeeded in pre- near the harbor. serving local Carian cults alive under a Hellenizing "A major reassessment of the location of Milesian veneer, as at Bagarcik high above Herakleia to thecemeteries will now be required as a result of new northeast. For 1993, see Ara?ST 12 (1995) 123-31; aresearch in the Kazartepe necropolis, which was 1994 report is published in AA 1995 (forthcoming). always assumed to have served during the Archaic Prehistoric cave paintings newly discovered in the period. Clearing of the chamber tombs and associ- Latmos area are discussed above, under "Palaeolithic." ated structures in preparation for their publication . The long-term survey and epigraphical has shown that only one honorific monument can project conducted by Enver Varinlioglu and French be dated this early, and must have commemorated colleague P. Debord to document settlement patterns an exceptional person. Although Miletos has now in Caria focused its attention in 1994 on Harpasa acquired a fine new monument, the Archaic necrop- (modern Esenk6y), south of Nazilli on the Akpay olis will nevertheless need to be sought elsewhere. (Harpasus) River. Soundings demonstrate that the "Conservation continued in the theater, the Baths Hellenistic town was built on an earlier foundation, of Faustina, and the Hellenistic Heroon I. In addi- going back at least to the sixth century B.C. It pros- tion, the cooperative project with the State Water- pered until Byzantine times, when the site contracted works Department to control the site's annual flood- in size and functioned primarily as a fortress. The ing is making good progress. A more discouraging Ottoman period revived its economic importance challenge is how to prevent illegal excavations, such as a market center. For earlier survey results, see as occurred during the winter in the Aphrodite sanc- Ara?ST 11 (1994) 199-204 and 12 (1995) 25-27, and tuary. Such incidents underscore the difficulties in- Asia Minor Studien 8 (1992) 17-22. volved in protecting even the most visible sites from looting." Ionia The 1992 and 1993 seasons are presented in KST Miletos. The 1994 season was spent in evaluating 15:2 (1994) 83-99 and 16:1 (1995) 405-17; a detailed past results, and assessing the directions in which summary of the 1988-1993 results is published by new research interests will lead the excavation proj- V. von Graeve et al., AA 1995, 195-333. ect in the coming years. A major decision involved . Klaus Tuchelt and his team pursued the reviving the Bronze Age investigations, which are project's current survey of the Sacred Road leading summarized above (see under "Chalcolithic and from the sanctuary of to Miletos. A small Bronze Age: Western and Coastal Anatolia"). An- spring and nearby stream marked the location of 1996] ARCHAEOLOGY IN TURKEY 319 an Archaic nymphaeum, to underlie the an eighth/seventh source century for B.C. level a with-seven century B.C. headless out statue intervening deposits; of whethera woman the hiatus is in- hol bird. In mid-Imperial times, deed reflecting a broader bath pattern, building or merely restrict- wa structed there. The area was later used for burials. ed to this area, cannot be determined on the present Restoration work continued in the Apollo temple, evidence. In the agora, at least, the earliest occupa- and 80 m to the north, in the village mosque, where tion began in the seventh century B.C., as Ionian a polygonal stone pavement and a reused Hellenistic Geometric pottery from the base of deep soundings architrave were revealed. A platform for an altar has has repeatedly indicated. Stratigraphic research in been discovered between the temple and the mosque, the agora has now been concluded, and the ceramic beside the sacred road. Brief summaries of the 1992 sequence is being prepared for publication. and 1993 seasons appear in KST 15:2 (1994) 69-82 Other projects of previous seasons continued: on and 16:2 (1995) 341-43. A full report on the Sacred the Hellenistic-Late Roman cemetery outside the Road and its monuments is in preparation: K. Tuchelt Damianos stoa; on the second-century A.D. Olym- et al., Didyma III:1: Ein Kultbezirk an der Heiligen Strasse peion and South Stoa, which now appear to be con- von Milet nach Didyma. temporary; at the Church of St. Mary, which reused . Publication research on the Agora and stone architectural elements from the Olympeion the Lower Gymnasium, regional survey, and resto- in its apse and earliest flooring, now to be dated ration of the theater were again carried out under during the fifth century A.D. but before the Council Wolf Koenigs's direction in 1994. For a progress re- of A.D. 431; on the stadium, where a marble portico port, see Ara?ST 12 (1995) 145-66, andAA 1995 (forth- of Hellenistic date (part of Lysimachos's circuit wall?) coming). may have given access to a formal road leading to Magnesia ad Meandrum. Orhan Bingol's efforts the Artemision; and on the water supply to the villas to free the Artemis sanctuary and surrounding struc- on the southern slope and elsewhere, a system that tures from a century of alluvium turned in 1994 to functioned until the Arab invasions. Conservation the temple proper, where he suggests that the east and large-scale restoration projects at the lower end entablature is lying complete except for its frieze of the main street involved reerecting the portal and (now in the Louvre in Paris). The results of the 1992 colonnade of the Temple of ; and wrestling and 1993 seasons can be found in KST 15:2 (1994) with the fig tree that had grown around the Hel- 41-52 and 16:2 (1995) 43-56. lenistic heroon or Androcleon, a monument dedi- /Torbali. The three excavation areas cated to the founder's cult and already repaired on of previous seasons were continued in 1994 by Recep several occasions in antiquity. The theater and Merit. A second stratigraphic sounding on the acrop- selected Roman chamber tombs are also being re- olis, where the earliest, Iron Age settlement was stored. For a summary of the 1993 campaign, see located, has pushed occupation back to the Subgeo- KST 16:1 (1995) 419-24. The Mycenaean finds from metric period, as well as confirmed that the acrop- the Artemision sounding are discussed by A. Bla- olis was abandoned from the fifth century B.C. until kolmer in OJh 63 (1994) 28-39. the Early Hellenistic period. Excavation of the Hel- . Juliette de la Genibre's program of the past lenistic stoa at the base of the acropolis was com- six seasons to document the history of the oracular pleted; the initial building is still to be dated in the sanctuary of Apollo and the adjacent Artemis temple, early third century B.C., with Roman and Byzantine and to restore their final and most flamboyant Hel- modifications. In the exceptionally well preserved lenistic-Roman phase, proceeded in 1994 with more theater, the stage building, intact even to its marble deep soundings and a large team of conservators. floor, was freed of 6 m of erosional fill. That the Soundings to the southeast of the Apollo temple have theater was entirely buried by the fourth century A.D. followed the continuation of four parallel rows of can now be demonstrated by the products of a glass hecatombs (second century B.C.) that attest to the workshop, installed on top of the erosion level above scale of sacrificial activities there. The possibility that the left parados. they overlay earlier installations, as suggested by Ephesos. The 99th excavation season at Ephesos reused blocks of similar type, was confirmed by the again expanded its team and its ambitions, with strati- discovery of some of these predecessors in situ. Large graphic research, conservation, and publication proj- numbers of molded votive terracottas, representing ects under the direction of S. Karwiese. The ongoing and perhaps offered by children and young girls as sounding at the northwestern end of the Artemision at Brauron, give the associated hecatombs an Archaic came down on a Mycenaean building level with wall date. Despite the problems of excavating below the foundations built of unworked fieldstones. It seems water table and deep silting, the sequence of Artemis 320 MARIE-HENRIETTE GATES [AJA 100 temples has now been butdefined before the Persian as destruction four of 530 ArchaicB.C. The v sions and two Hellenistic ancient oliveones. oil industry The required soundings two mills, hot duced a fine head from water, an andArchaic lime, thus bothsculpture, processing and storagea f and Attic Protogeometric facilities. sherds All of its elements that can be coulddocumented here,her a foundation for the Artemis including a water cult supply as from early.as channels that the bor- century B.C. The sanctuary's dered the road abruptbeside the factory. collapse While the rooms in later Roman period can themselves be gauged were built with by stone the walls, thedrama special- tumble of honorary monuments ized installations were tocut into the bedrock: south these pits o Apollo temple, where excavationsand vats, in shapes appropriate continued to their functions, rem ing the especially thick would silt. have been used for crushing the fruit, extract- In conjunction with the ing the restoration oil, and collecting it (fig. of 34). Thethe large colo plas- sal sculptural group teredrepresenting pits with a central circular Claros's depression were sa family-Apollo, Artemis, housings andfor the mills,Leto- whose twoa partsreprodu were held has been made by Lafarge-Copp6e, together with a pivot. In the northernwhose extension pre of experience with the Khorsabad the complex, intact sculptures pithoi set into the floors, their them for the task of castingrims flush with sculpture the pavement, were on probably this used (the Artemis leg alone to storeweighs the oil. They 7were tons!). sealed off when The the floors g was originally placed in were the repaved cella in the ofcomplex's the second Apollo phase (fig. 35)te during the reign of the probably emperor not long after the initialCommodus setup. Finally, the 8-m-high honorific column amphoras foundof Sextus in significant Apolleius numbers in these i ing reerected in its findspot deposits must haveon served the as portablesacred containers road for the 1992-1993 interim the reports,finished product." see KST 15:2 ( 53-68 and 16:2 (1995) 39-42. Bayrakh-Old . Excavations directed in . In 1993 Numan Tuna resumed D. Mustafa 1994 by Meral Akurgal continued in the fourth cen- Uz's project with a survey, and continued efforts to tury B.C. western residential district, where two more save the site from encroaching constructions. For an streets faced with tightly packed buildings were un- assessment, see Ara?ST 12 (1995) 167-76. covered. Among the houses was a workshop filled Klazomenai. Local industries of the Archaic with molds for terracotta figurines. North of the period were further documented during the 1994Athena temple, the Cyclopean enclosure wall was fol- campaign at Klazomenai, as Gilven Bakir and Yagar lowed further toward the east. Outside it, a corbeled Ersoy, Ege University, Izmir, kindly report: fountain house dating to the late seventh century "The season's work concentrated on two different B.C. was excavated completely so that its structural areas: on the southern acropolis slope along the aspects and relation to a nearby spring could be Yildiztepe necropolis road, where an industrial dis- studied before beginning restoration work. The trict occupied the outskirts of the city; and in the 2.5-m-high fountain house was built of rectangular western sector of the city proper, in a complex de- blocks on basalt foundations, abutting the different voted to the production of olive oil. The industrial masonry of the fortifications. Behind it, a conduit district, consisting of two main phases and their sub- covered with large slabs led to the spring proper, levels, began its activities in the second quarter of which was channeled via a deep square shaft. Simi- the sixth century with ceramic production. Kilns asso- lar systems are known from Miletos and Syros. The ciated with quantities of Wild Goat-style pottery were 1993 season is presented in KST 15:2 (1994) 37-40. found directly on bedrock, although a few Late Geo- Attic black-figure pottery from Bayrakli and other metric and Subgeometric sherds in the same deposit western Aegean sites is discussed by Y. Tuna-Norling may point to an earlier presence in the area too. In in Ara?ST 11 (1994) 437-51. the second half of the century, when a series of work- shops was constructed here as a second building phase, the pottery industry diversified to include Sardis. The Archaic and Classical lower city's mas- bone-carving and iron metallurgy. Significant num- sive fortification wall, Late Roman and Byzantine bers of cooking vessels from both levels suggest that housing, conservation, publication research, and the artisans working here also lived in their shops. potential sources for Sardis's limestone and marble "The olive oil factory first investigated in 1993 was claimed the careful attention of the Harvard-Cornell excavated further toward the north, where the exis- Expedition during the 1994 season. C.H. Greene- tence of two phases became evident. Both are dated walt,jr., University of California, Berkeley, graciously by ceramics to the second half of the sixth century, reports: 1996] ARCHAEOLOGY IN TURKEY 321

Fig. 34. Klazomenai. Olive oil complex, with in

"Excavation again focused tury B.C.on the Two area major at th p western foot of the Acropolisthe 150-m (sectors stretch MMS/ that MMS/S), where a fortified includes, line at of the defense northe ex about 300 years, from dation the seventh trench tocontaine the fo

Fig. 35. Klazomenai. Olive oil complex: pithos visible. (Courtesy G. Bakir and Y. Ersoy) 322 MARIE-HENRIETTE GATES [AJA 100

Style and Corinthian unfinished or Orientalizingcrepis wall inside Karniyarik Tepe. East Re- wares) confirming its searchconstruction projects on the EB cemeteries in the excavated thir 30 ter of the seventh century years ago beside B.C. the Gygean The lake, Hellenisticgate reliefitse built over two cobbled roads that antedate the for- ware, Lydian pottery, and Roman terracotta water tifications. After the Persian destruction in the mid- conduits will lead to publications and doctoral dis- sixth century and the city's conversion into a satrapal sertations. Fikret K. Yegiil continued his architectural capital, the gate was blocked and traffic diverted from recording of the Artemis temple. this perennial east-west thoroughfare to some other "A survey for the stone sources that supplied route. The second phase, following shortly upon Archaic Sardis's contractors identified several lime- Sardis's rehabilitation, reused the earlier fortifica- stone quarries on the Bin Tepe ridge, and marble tions as a rampart to support and elevate a 5-m-thick quarries (one inscribed with XOPOE, or "boundary") stone wall built on top, while leaving exposed at least to the south and southeast of Sardis. Sampling of one part of the earlier version's facade. A mid-sixth the marble used in the Tomb of Alyattes and Kar- century occupation surface at the foot of this facade nlyarik Tepe, however, suggests in preliminary analy- produced more iron objects, bringing to over 300 sis that the best match is not with local types, but items the Archaic iron assemblage recovered from with marble from the Ku?ini quarries near Ephesos." this area since 1992. The defensive system continued Reports on the 1992 and 1993 seasons have been to enclose and protect the lower city into the fourth published in KST 15:2 (1994) 101-13 and 16:1 (1995) century B.C. 393-404. C. Ratte discusses the Archaic architectural "This area was occupied during the following cen- terracottas in Hesperia 63 (1994) 361-90; on this topic, turies, but architectural features remained modest see too E. Hostetter, Lydian Architectural Terracottas until the Late Roman period, when the thorough- (Atlanta 1994). fare took on a monumental character as an 18-m- wide avenue flanked with colonnaded porticoes. The Aeolis- sector excavated in 1994 (MMS/N) contained a paved . The four-month 1994 season at Per- walkway perpendicular to the south portico and run- gamon concentrated its research and conservation ning south, perhaps to one of the neighboring in- efforts on the acropolis's two most salient monu- sulae. Portico and walkway were substantially built ments: the Altar of Zeus and the Trajaneum, with of spolia, such as the torso of a draped figure whose outright excavation limited to soundings tracing the unfinished back destined this statue for a more dig- city's system of streets. Wolfgang Radt, Deutsches nified initial setting in a niche or against a wall. And Archiiologisches Institut, Istanbul, kindly discusses although the residents of Sardis during this period and illustrates the results: appear to have respected their predecessors' con- "It seemed timely to investigate once again the Per- structions only as a ready source of building material, gamon Altar, which has recently been the subject their own houses were carefully preserved and re- of renewed discussions concerning its date and de- stored well into Middle Byzantine times. When, in tails of reconstruction. To this end, the still intact the course of the season, we dismantled these houses foundations were cleaned in order to test the accu- for access to the underlying Archaic strata, we found racy of their plan; several of their compartments were that their walls consisted of stratified courses whose excavated for chronological indicators, and to reex- mortar contained pottery dating, in the uppermost amine the apsidal building that underlies them; and layers, to no earlier than the eighth or ninth century clues to the original placement of the marble super- A.D. The homogeneous external appearance of these structure were carefully studied. walls would have suggested a single phase of construc- "The altar's foundations consisted of a 4.5-m-high tion, instead of repairs actually spanning half a structural core designed as a casemate system, with millennium. rectangular compartments all of uniform size except "Conservation efforts focused on the pottery and in the narrower outer row (fig. 36). This core was metal finds from the Lydian residential unit exca- then entirely concealed with marble revetments; it vated in 1993, a chronologically precious subassem- should be noted, however, that the long-standing re- blage of conventional Lydian and East Greek vessels construction of the northeastern corner (see below, that predate the Persian destruction; painted funeral fig. 38) was put together from blocks that were not klinai supports recovered by the Museum found in situ. Previous excavations by Humann, von from an Archaic tumulus chamber at Bin Tepe; Conze, and, in 1961, by J. SchSifer had focused on painted wall plaster from the Late Roman housing the central portion of the foundations, and emptied in MMS; sections of the earlier fortification; and the three of the compartments. This year, in addition 1996] ARCHAEOLOGY IN TURKEY 323

Fig. 36. Pergamon. Altar of Zeus: found to cleaning the top andrelated to it. theThey must haveperimeter, belonged to the Early 10 m partments were cleared Hellenistic houses thatdown extended to the to south ofbedroc the other three reexamined. Altar, and either precededThe or painstakinwere coterminous with used in their construction can now be understood. the apsidal building itself. No closer date within the Their walls were built of large tufa blocks, which Early were Hellenistic period could be assigned to the laid down only one or two courses at a time apsidal until building as a result of this new study. the same level inside the compartments could be "Aob- considerable amount of pottery was recovered tained, tightly packing them with andesite rubble from the soil layers in the compartment fill, and there and a covering layer of earth (fig. 37). The packing stones included squared and polygonal blocks from dismantled Hellenistic buildings (such as the under- lying apsidal one), as well as small and very large rough fragments. This technique ensured that the monument would resist settling and even earthquake damage. "Efforts to learn more about the plan and nature of the underlying apsidal structure were not fully rewarded, in large part because the preparations for the later altar involved a meticulous clearing of the area. No trace of it could be discovered beyond its eastern apse, despite a careful search at the base of the compartments bordering its presumed north- ern, southern, and western walls. To Schiffer's pre- vious study, however, a few details can now be added. The outer face of the apse preserved some of its orig- inal red plaster coating, while the interior face was Fig. 37. Pergamon. Altar of Zeus: compartment D2, with plastered in white. Shallow drainage channels were Hellenistic rubble packing still in place in its western half. cut into the bedrock near the wall, but appear un- (Photo E. Steiner) 324 MARIE-HENRIETTE GATES [AJA 100

Fig. 38. Pergamon. Altar of Zeus, from the no was no difficulty in distinguishingand the Altar. Here a stepped street, the bordered Hellen by deposits from those previously deep drainage channels, wasexcavated, newly discovered and or o disturbed by the Byzantine followed for somehousing length. The carefully that cut stone overlay slabs area. The original deposits that paved its provide steps and covered thean drains excellent were often view of the entire range incised withof the local Greek letter ceramic X (fig. 39), an interest- produ in both fine and coarse ing contrastwares, to the Efrom found on the a western context streets. s in the first part of the The streetsecond was laid outcentury on a sharp diagonal B.C. course Alt this ceramic assemblage (northwest-southeast) can be at odds considered with the previously a able general indicator for recorded the orthogonal Altar's axes, and datethus requires of a re-cons tion, unfortunately neither assessment ofit the nor criteria directingthe few the Hellenistic identi coins from the fill can narrow down its date more planners. closely within the 185-160 B.C. range. The season's "The 15-year project headed by K. Nohlen to re- finds and those of Schaifer's investigations will be store the Trajaneum was successfully completed presented in a forthcoming publication of the en- with two concluding rituals: the placement of tri- tire monument, together with a new plan of the foun- lingual explanatory panels outside the entrance; and dations, and a detailed study of the marble super- a celebratory banquet in the precinct for everyone structure- for which the season recovered a number who participated in this undertaking. Conservation of small fragments. The compartments were filled of the painted stuccoes from the Masked Mosaic in at the end of the season, and the foundations en- Room in Building Z also made good progress." closed in a fence to end 'tourist erosion' (fig. 38). Reports relating to the 1992-1993 seasons can now "Soundings to determine Pergamon's urban lay- be found in KST 15:1 (1994) 563-90 and 16:2 (1995) out by locating its network of streets continued this 1-26. For 1994, see AA 1995, 575-96, with a full year in the unexcavated northeastern residential dis- bibliography of recent Pergamene studies. trict on the steep slope southeast of the Upper Agora . Omer Ozyigit conducted salvage inves- 1996] ARCHAEOLOGY IN TURKEY 325

Interim reports appear in KST 15:2 (1994) 1-9 and 16:2 (1995) 27-37.

Troad Troy. The project of 1994 focusing on the post- Bronze Age monumentalized the scale of the Hel- lenistic sanctuary, discovered Archaic architectural remains for the first time, and began to close the gap between the Bronze and Iron Age phases. C. Brian Rose, University of Cincinnati, kindly reports on his team's progress: "The excavation area north of the Upper Sanctu- ary was continued for a third season and expanded, in order to uncover broader plans for the Hellenis- tic South Building and the North Building's prede- Fig. 39. Pergamon. Northeastern residential distr ment slab with incised cessors Greek (fig. 40); and letter extended into X. Blegen's (Photo Upper E and Lower Sanctuaries to recover earlier phases in this precinct. The South Building, with its 3-m-deep tigations both above foundations ground of regular andmarl ashlars, under was an enor- wa efforts to record Foga's mous building past that would before have dwarfed every it other is e by the city's uncontrolled structure in the area.growth. The 20-m length Researcof exterior poorly preserved Cybele walling uncovered sanctuary, by the close of the season whose repre- tions were cut into bedrock, sents only a fraction uncoveredof the original size. Although Arc ing of the sanctuary's the superstructurefirst phase,of limestone blocks and has been thre es- ings (the most recent, sentially robbed a out,Genoese even its few remaining reuse). elements terracottas and tufa highlight plaques the importance ofshowing this second-century B.C.th in an oikos resemble architectural representations commission, now to be understood ofas Cyb in Marseilles, a Phocaean an enclosed plan colony rather than a stoa[and or portico. at It lat see below under "Troad"]. coexisted with the The nearby North Hellenistic Building, which was t on Maltepe, and the nearby set on a higher terraceArchaic formed by the fortificatremains of excavated in previous two earlier seasons, Hellenistic structures. were An iron double also st restored. The rubble used to construct the tumulus ax and an Archaizing gilded bronze sphinx found contains a wide range of Hellenistic pottery and figu- in 1993 in association with the North Building can rine wasters, clearly from a workshop's dump. They now be linked with the symbols of and would suggest an active local industry manufactur- that appear on their coinage. They may rep- ing even such types as West Slope ware. resent gifts from members of the Troad koinon to their The municipality's callousness toward ancient capital's sanctuary. Finally, traces of another large monuments is well illustrated by the opening inEarly Hellenistic structure with pebble mosaics and the Aphrodite sanctuary of a tea garden, which in- fragmentary wall paintings showing architectonic de- volved damaging the Ottoman tower that overlooks signs were recovered in a new trench between this it; and by an incident involving a marble replica ofarea and the Upper Sanctuary- the first evidence a Cybele plaque, set up by 6zyigit beside the linking the two areas architecturally during this time. sanctuary's spring, which was vandalized within 10 "For the earlier stages in this sacred precinct, the days of installation. Classical phase is so far too badly disturbed by stone For the 1992 and 1993 reports with poignant photo- robbing and pitting to allow any suggestions. But graphs of these tensions, see KST 15:2 (1994) 11-36 monuments of the Archaic period are appearing for and 16:1 (1995) 425-54; a study of the fortifications the first time. Under the porch of the North Build- has been published in REA 96 (1994) 1-34. ing, the corner of a building and a column base still Kyme. Research projects of previous seasons were in situ were exposed, certainly part of a structure pursued in 1994 by Sebastiana Lagona and her col- with an interior colonnade. An Aeolic capital in later laborators in the harbor, the theater, the Hellenistic fill probably belonged to the same building, whose Isis sanctuary on the north hill, and in the central stratigraphic situation would date it to the sixth cen- Byzantine and Roman residential district. Neolithic tury B.C. It appears to have had a short duration, and EB pottery has been recovered in mixed contexts. since it was sealed off by Early Classical debris. This 326 MARIE-HENRIETTE GATES [AJA 100

Fig. 40. Troy. The North and South Buildings ( looking north. (Courtesy C.B. Rose)

monument, a boundary dependent wall to its andsouth, were and equ an potential structure beyond The newit, were finds probably still do noco porary with the two though Archaic female altars deities excavat cou Blegen in the Upper Figurinesand Lower from Sanctuaries the Low "Previously unexcavated Cybele, deposits but identification in the Lo Sanctuary were investigated "Seventh-century in 1994, with B.C.result p both confirmed and mixedenlarged contexts Blegen's by assess Bleg Quantities of Orientalizing Lower andSanctuaries. Archaic Inceram 1994 ports were again recovered recovered here, fromas well stratif as a va of votive offerings (notably below a thefaience Archaic scarab, surfa two tacle" fibulae, and a terracotta together withfigurine). the gray The a remained in use until ofthe Troy Early VIII. Hellenistic Close study per when it was replaced byin a1994 new underlined one at the thesame f as the Hellenistic mosaic indistinguishable building was from fou s Throughout the sanctuaries' an indication history, that the thetwo kept separate from each chronological other by enclosure range than w and distinguished by thedating different proposed orientatio too by D the altars. Even in the A Hellenistic sounding besideperiod, the whe N altars were aligned, the a Protogeometrictwo enclosures remainestratum 1996] ARCHAEOLOGY IN TURKEY 327

VII:2 deposit, a sequence that would sugges lar sequencing of the two phases. A cach bronze fibulae dating to ca. 700 B.C., althou in mixed fill, promises more evidence for Iron Age in this sector in future seasons. "An ancillary benefit of the Lower City Br project to examine the Troy VI ditch was t tigation of this area's later phases. An E lenistic stone structure, although much d by later constructions, was found to be ali the street grid. This urban plan was thu already in place centuries before the Rom to which it has been attributed. A nearby emptied of nearly 17 m of fill dating from second century B.C. to the mid-first centu another area a limestone quarry was di which appears to have supplied the Hellen with building material. The area above it m become residential in later years, as indicat other well in use during the third centur the evidence of pottery and especially coi it had been tossed a gilded terracotta Cy a marble version of similar size and type whose Archaic roots have recently been il at Phocaea. The southwestern corner of the LB Lower City, in contrast, does not seem to have been re- occupied until the Late Byzantine period, when it functioned first as a cemetery, later as a storage area with large pits cut into bedrock, and finally as a field Fig. 41. Troy. Lower City. Marble statuette of Cybele. (Courtesy C.B. Rose) in Early Ottoman times." Interim reports on the 1993 season have appeared in Studia Troica 4 (1994) 75-104; for 1994, see Studia the 1994 season inventorying more of the colossal Troica 5 (forthcoming). Temple of Hadrian's disiecta membra, which include Assos. Umit Serdaroglu devoted the 1994 cam- unfinished drums for the columns that once stood paign to excavating sectors in the necropolis outside to a height of 17 m. Continuing survey of the asso- the city gate, and in the residential district on the ciated city extended to the theater and the Metro6n, terrace overlooking the gymnasium. He also carried where Roman female statuary was found. An illus- out a regional survey. trated report on the 1993 results appears in KST 16:2 Smintheon. The program of cleaning, restoration, (1995) 107-30. and soundings in the Apollo temple continued the Iznik. In a 15th season of excavations in the pace of previous seasons under the direction of theater, whose construction had apparently already Cogkun Ozgfinel, in whose honor an appreciative begun during Trajan's reign, Bedri Yalman freed the Gillpinar municipal council has renamed the street entire width of the stage building from its massive leading to the sanctuary. The stereobate has been overburden of reuse. The stage building housed completely cleared of overlying deposits, which stable in the Medieval period, with tethering holes contained, among other architectural debris, two drilled into many of the marble facade blocks. The fragmentary entablature blocks that appear to be splendid collection of glazed vessels excavated in the window frames. Soundings around the stereobate stratified sequence of ceramic workshops that occu- revealed a stretch of fourth-century B.C. header-and- pied this area in Byzantine and Ottoman times wil stretcher masonry that should belong to an earlier be presented in a forthcoming thesis. The 199 version of the building. season is discussed in KST 15:2 (1994) 425-54. Ainos-. Sait Bagaran, who has inherited the Propontis, Bithynia, Thrace, Pontus directorship of the Ainos project, widened and deep- . Abdullah Yaylall and his assistants spent ened the stratigraphic sounding on the acropolis 328 MARIE-HENRIETTE GATES [AJA 100 down to bedrock. The alongoccupational the site edge that was threatening sequence to collapse beg in the sixth century B.C. and with be destroyed. many It proved diagnostic to be a dense area ofAegea and East Greek wares, and was maintained without stratified kilns and related installations such as wells interruption into the Hellenistic period. The acrop- and storage areas. Unlike the Zeytinlik workshop, olis was then abandoned until the Byzantine era. On where production was limited to heavy potting, this the eastern side of the modern city, excavations in center made a few varieties of tableware and lamps the classical necropolis followed immediately on the along with cooking pots, tiles, and amphoras. The heels of clandestine diggers, whose approach -while kilns were adapted to this industry: the earliest (sec- systematic-requires less time and thus outstrips ond century B.C.) and the best preserved, an oval formal research at an alarming pace. The project 2.4 x 2 m with two central supports, contained over also progressed in its long-term program of restor- 20 tubes to diffuse the heat and support the more ing Enez's Byzantine and Early Ottoman monuments. fragile vessels during firing. It was abandoned and The 1992 report appears in KST 15:2 (1994) 455-73. replaced by a shallow well at the end of the Helle- M. Tunay's research on the Byzantine frescoes in the nistic period, until a second kiln was assembled here Haghia Sophia church is published in Ara?ST 11 (second century A.D.) and enclosed in a stone plat- (1994) 521-25. form to retain heat. This kiln was again replaced Zeytinlik/Sinop, Demirci. A comprehensive proj- two or three centuries later by a well. The amphora ect to investigate the Black Sea amphora industry industry sampled in the preceding year's survey is and its center in the Sinop region followed up on now confirmed through stratified finds. A few new the 1993 survey (see AJA 99 [1995] 246-47) with ex- types were previously known from the northern shore cavations in 1994 at two workshops: Zeytinlik, in an of the Black Sea, at the sites to which they were ex- eastern suburb of modern Sinop, and Demirci, 18 ported. Finally, although the 1994 excavations fo- km to the south. Dominique Kassab Tezg6r, French cused on Demirci's later ceramic industry, the site's Institute of Anatolian Studies, Istanbul, and her col- long involvement in this speciality is attested 300 laborators Yvon Garlan and Ismail Tatlican (direc- m to the north, where a fourth-century B.C. work- tor of the Sinop Museum) kindly provide the follow- shop has been identified." ing report: For the 1993 survey, see Ara?ST 12 (1995) 177-90; "The kiln site at Zeytinlik represents the earlier Y. Garlan discusses Sinop amphora stamps in CRAI of the two excavated workshops, which between them 1990, 492. The neighboring Byzantine site of Qiftlik span the Hellenistic and Roman periods to Early is summarized below, under Byzantine Period: Byzantine times. It was exposed and damaged by the Qiftlik/Sinop. foundation cut for a housing project in 1992 and Samlar necropolis/Amasya. Further examples of 1993, and excavation was thus restricted to salvage the barrel-vaulted communal tombs used by modest investigations on a single kiln. The workshop was Late Roman inhabitants of Amasya were excavated established on bedrock that had been cut back by by Metin Ozbek in a second salvage season in 1994. quarrying. Shortly thereafter, a large piriform kiln Samples taken from construction timber may place constructed with reused tiles and clay began to manu- the cemetery in an absolute dendrochronological facture primarily amphoras and tiles: stamps found framework. Previous findings are reported in KST in the kiln and bearing the name of the producer 16:2 (1995) 95-105. Sosias date it to 300/290 B.C. Associated pottery, and many wasters, indicate that the kiln remained active Phrygia for one century. Stamped examples numbered 169 . Klaus Rheidt and his team devoted the (all from Sinop with the exception of three Rhodian 1994 season to restoration projects in the porticoed types), including a few tiles impressed with the same street and at the Roman baths; to organizing the seals used on the amphoras. Of particular signifi- excavation's large collection of reliefs and inscriptions cance is the large proportion of stamps relative to from the site and its vicinity; and to pre-publication the number of amphoras recovered: roughly 80%, study. Progress reports for 1992 and 1993 can be in contrast to the 17-20% proposed in earlier studies. found in KST 15:1 (1994) 515-37 and 16:2 (1995) After the workshop ceased production, no doubt in 57-73; for 1994, see AA 1995 (forthcoming). the upheaval caused by Pharnaces's conquest of Sinop Pessinus. The 1994 season directed by John Dev- in 183 B.C., the area remained essentially unoccupied reker progressed on the combined fronts of exca- until the Late Ottoman period. vation, geophysical prospecting, and intensive "At Demirci, excavations were carried out in one regional survey that have characterized the past sev- of several anomalies recorded by geomagnetic survey, eral campaigns. Excavations again focused on the 1996] ARCHAEOLOGY IN TURKEY 329

city center. Trenches square north in the center ofof Tarsus, the where bulldozingtemple, for built under Tiberius a parkingfor lotthe had uncovered imperial the paved street cult, of a to the east, confirmed much olderthe municipal existence project. Levent Zoroglu of and a M Phrygian (seventh-mid-sixth his associates devoted the 1994century campaign to strati- B.C level as the earliest graphicoccupation. soundings around the 65-mBeyond stretch exposed th to the north, a possible the previous defenseyear, in order to determine wall its chrono- of E lenistic date and a massive sunken foundation were logical parameters. The 1.5-m-deep road foundations, overlaid in the second/first century B.C. by a large carefully set on a bed of sand with a sunken drain, building with black-and-white mosaics, already partly were found to overlie a Middle Hellenistic building encountered by Lambrechts in earlier excavations. level. It can now be concluded that the street was Because this area was badly disturbed by Byzantine laid out as part of an Early Imperial Roman quarrying for building materials, there remains un-urban reorganization in the first century B.C./early fortunately little of its Roman phase. first century A.D. The street was stabilized again with Survey and GIS analysis traced 7 km of the north- limestone paving blocks in Late Roman times. It was ern aqueduct's terracotta pipes, and found possible probably no longer functioning by the Medieval evidence for a Roman dam in a high valley at theperiod, when the portico columns were retrieved for far end. A second water supply was tapped at a source private housing. The archaeologists hope to persuade 8 km to the northeast. Four more Roman cemeteries the local council to establish an open-air museum were located, in the vicinity of their villages and farm- here. steads. A report on the 1993 season appears in KST Donukta4-Tarsus. Nezahat Baydur assesses her 16:2 (1995) 75-94. 10th and final excavation season in 1992 at the co- Ankara. The demolition of a large commercial lossal Imperial Roman temple on the outskirts of building in the city's old Ulus district gave unprece- the modern city in KST 15:2 (1994) 225-41. dented access to Ankara's pre-Republican history. Ex- Kelenderis. After a decade of excavations in cavations carried out in 1995 by the Ankara Archaeo- Rough Cilicia's best harbor and Anatolia's closest logical Museum and Cevdet Bayburtluoglu reviewed maritime link with , Levent Zoroglu spent the here the entire sequence of periods from Ottoman 1994 season conserving the 15-m-long Late Antique to Phrygian times, with especially substantial Roman mosaic for in situ display (see AJA 99 [1995] 249, fig. and Hellenistic phases. The most impressive remains 33); and restoring several of the town's Ottoman were a well-paved Roman street oriented n-s and monuments. For the 1992 season, see KST 15:2 (1994) exposed for a length of 69 m, neighboring structures 189-209. The first in the final report series has with marble floors, and a Roman female statue in appeared: L. Zoroglu, Kelenderis I: Kaynaklar, Kalzntzlar, marble. Classical, Archaic, and Phrygian pottery, Buluntular (Ankara 1994). without associated architecture, dates the earliest Karasis. Mustafa Sayar discusses and illustrates occupations founded on sterile soil. An illustrated this well-preserved Seleucid fortress, perched high report has been published in the Museum of Ana- above Anavarza/Anazarbos in northeastern Cilicia, tolian Civilization's Museum News 6 (1995) 5-6. in AntW 26:4 (1995) 279-82. Recent installments of Hierapolis. The previous years' vigorous pace of his annual epigraphic surveys in this region can be research and restoration was maintained by Daria found in Ara?ST 11 (1994) 137-60 and 12 (1995) 39-60. De Bernardi Ferrero in 1994. Excavations in the agora Kiiqiik Burnaz. This anonymous ancient harbor and on Frontinus Street uncovered and reerected in the northeasternmost tip of the Mediterranean more of their columned porticoes. Restoration con-remained invisibly buried in coastal dunes until 1987, tinued in the Apollo temple, the theater, and when the a local company began to bulldoze the site for Medieval fortress. The project begun in 1993 in sandthe and revealed buildings standing as high as 10 m. northern necropolis added four new tombs to A thethorough survey, carried out in 1994 by Jennifer over 100 already inventoried: these masonry hypogea Tobin, recorded some 20 identifiable structures and were conceived as houses to contain sarcophagi, placedin- their chronological range between the first cluding one Sidamara type. For the 1992 and century1993 B.C. and the 13th century A.D. The site was seasons, see KST 15:2 (1994) 341-50 and 16:2 (1995) laid out on a roughly 500 x 500 m cardinal grid, 345-60. perhaps as the port for inland Epiphaneia/Giizelhan when Pompey resettled Cilician pirates there in 67 Cilicia B.C. Much of the architecture, however, suggests that Cumhuriyet Square-Thrsus. Salvage investigations the site saw its greatest development in Late An- continued for a second season in the municipal tiquity; it may have profited from its location on the 330 MARIE-HENRIETTE GATES [AJA 100

Pilgrim's Road as an inviting been cleared out). stop Building forprojects wearyon the city out- tra ers. The best preserved skirts of damaged the elements buildings of three Byzantine exposed sum- the city center was connected mer residence complexes: to in anthe Aretas aqueduct of Romanus displays the necessary IVfurnishings Diogenes on the European ofside; anda onpublic the Ana- b chambers lined with waterproof tolian side, a bath near cement, the Damatris complex hypoc at tiles, and elaborate piping. Samandra, and Other walling west oflarge the main Bryasstruct com- are still contained within intact dunes. The site is plex at Kiigilkyali. The year's most spectacular acci- now protected by a fenced enclosure generously con-dental discovery occurred in the Kocamustafapaga tributed by a nearby industrial center. For mention district, inside the city walls on the Sea of Marmara. of the site in previous surveys, see Ara?ST 10 (1993) A foundation cut there came down on two intact 391 and ArkST 8 (1993) 357-67. sarcophagi, and below them, on a large figural mosa Porsuk. For the Roman period, see above under (fig. 42) from the floor of what must have been a "Chalcolithic and Bronze Age: Eastern, Northern, wealthy villa"' and Central Anatolia." E. Yficel's restoration work at Haghia Sophia i

Southeast Anatolia discussed in Ara4ST 11 (1994) 1-8 and MKKS 5 (1995 171-74. The Silivrikapi Mausoleum (erroneously . Reports by David Kennedy on his 1993 identified as such in AJA 99 [1995] 250, fig. 34) ha preliminary season at BelkislZeugma appear in KST been published by J.G. Deckers and tU. Serdaroglu 16:2 (1995) 207-15, Archaeology 48:2 (1995) 54-55, in JAC 36 (1993) 140-63. and AnatSt 44 (1994) 18-20. No formal excavations Great Imperial Palace, Istanbul. In 1994 E. Bolo were carried out in 1994. gnesi Recchi-Franceschini concentrated her thir survey season on the juncture between the uppe BYZANTINE Constantinian palace and the lowerJustinianic one, Istanbul. The struggle to record, preserve, in the area and of the great peristyle hall (site of th celebrate the monuments of this city's pastMosaic in Museum:the see below). She now dates the con- face of unprecedented urban growth achieved struction sev- of Mamboury DB - perhaps a church - to eral major victories in the past year, to offset the same the period by as the peristyle, whereas Mamboury now familiar chronicle of destruction. The follow- DC would represent a later phase in the palatia ing report was kindly prepared by Mehmet Tunay, complex's history. A report on the first (1992) seaso Istanbul University: appears in Ara?ST 11 (1994) 19-34. "This year's most celebrated moment for Byzantine Great Imperial Palace/Mosaic Museum, Istanbu studies was the long-awaited opening of the Archaeo- The restoration of the Mosaic Museum, in the logical Museum's new Byzantine wing in June 1995, Justinianic palace's peristyle court and its northeast- together with an exhibit entitled "Istanbul through ern hall, was revived and is nearing completion the Ages," where finds from excavations at the citythanks to a joint Austrian-Turkish project directed walls, the Hormisdas Palace, and the churches of St. by W.Jobst and B. Erdal. The museum project, con- Polyeuktos and Constantine Lips take pride of place. servation, and study of the mosaics (excavated in the It is also significant that an independent national 1930s and 1950s) are discussed in Ara?ST 11 (1994) commission has declared the entire historic district 9-18. New soundings, carried out in three sectors of Istanbul a protected zone. All building projects of the northeastern wing to determine more closely will be required to obtain clearance from designated the stratigraphic setting of the mosaics, confirm that authorities, and will proceed under close supervision they were installed in the first decades of the sixth in the case of archaeological findings. These regu- century, when the palace was restructured and lations should prevent the repetition of disasters restored.like those that affected the Bodrum Camii, the Hippo- Anastasian Wall Project, Thrace. A survey to map drome, the Forum of Constantine, and-more and analyze the 50-km circuit of long walls, con- recently- the Magnaura. structed in the early sixth century by Anastasius I "The pattern of recent years urgently needs inter- (A.D. 491-518) between Karacak6y on the Black Sea rupting. As a result of the 1994 municipal elections, and Silivri on the Sea of Marmara, was carried out all restoration work on the city walls has ceased, as in 1994 byJames Crow. Although the northern end have related projects on the Hormisdas and Mangana has eroded into the sea, the other parts of the de- Palaces (fortunately for the cisterns of the St. George fensive system - ditch, walls, gates, rectangular towers, monastery at Mangana, only after the debris accumu- and possibly pentagonal towers - have been well pro- lating there since the excavations of the 1920s had tected and preserved by vegetation. It is suggested 1996] ARCHAEOLOGY IN TURKEY 331

Fig. 42. Istanbul. Kocamustafapaga dist M.I. Tunay) that the walls may have been constructed to safeguard in C. Mango and G. Dagron eds., Constantinople and the preexisting aqueduct system, of which many sec- Its Hinterland (Cambridge 1995) 207-18. The final tions are still visible in this area. The long walls func- report is planned for 1996-1997. tioned for 150 years, after which Constantinople fell Ere'li-Marmara. The salvage excavations at the back on its city fortifications for protection. A sum- Middle Byzantine Basilica at Perinthos, with well- mary appears in British Institute ofArchaeology/Ankara preserved mosaics, are reported by M. Akif Iln in Research Reports 1994, 18. For the aqueduct survey MKKS 5 (1995) 27-37. conducted in the 1950s by F. Dirimtekin, see CahArch Gazik6y-Tekirdag. Nergis Giinsenin shifted her 10 (1959) 215-43.J. Crow has published a pre-survey investigations in 1994 from the Byzantine amphora study of the walls in C. Mango and G. Dagron eds., manufacturing center of ancient Ganos to shipwrecks Constantinople and Its Hinterland (Cambridge 1995) off the Marmara coast, and an amphora workshop 109-24. on the island of Marmara. She discusses her 1993 Ainos-Enez. See above, under "Classical: Pro- results in ArkST 10 (1995) 201-20. pontis, Bithynia, Thrace, Pontus." Kurqunlu/Gemlik-Bursa. In 1995 Mehmet Tunay Cyzicus-Iznik-Geyve Survey. The fifth and final began the cleaning and restoration of the 12th- season of a multidisciplinary project to record the century monastic Church of St. Abercius. human and natural geography of this province Ulubey/Ugak Survey. In the third installment of (Roman Bithynia) in Roman, Byzantine, and Otto- a five-year project to record Byzantine castles along man times extended its survey from Gemlik east to the Aegean coast, Mark Whittow turned south of the Sakarya River. Project directors Veronique Usak to the fertile Banaz plain, whose agricultural Frangois, Bernard Geyer, and Jacques Lefort com- prosperity in Roman and Byzantine times was well bined their fields of archaeology, geography, and his- illustrated by N. Firatll's excavations of the church tory to assess the complexities of the Iznik lake, whose at Sivasll. The 1994 survey focused on the large castle level has fluctuated over 4.5 m in historic times be- that dominates a canyon 2 km south of Ulubey. The cause of climate and human intervention; followed castle's relatively poor preservation, mainly at foun- the regional network of Byzantine routes and asso- dation level, in fact proved useful in revealing its ciated settlements; and evaluated the diachronic many phases of construction alternating with periods effects of heavy erosion. A report on the Bursa area of neglect. It may originally date to the seventh/eighth (1992 survey) has been published in Ara?ST 11 (1994) centuries, then fell out of use during the two cen- 65-71.J. Lefort discusses the Byzantine road systems turies of Byzantine economic expansion that pre- 332 MARIE-HENRIETTE GATES [AJA 100 ceded the defeat at Manzikert in A.D. 1071. The castle basilica, followed by extensive remodeling, perhaps was then reactivated in a second building stage, which in the second half of the ninth century, when the made use of Roman spolia for the repair and ex- building was transformed into a cross-domed basili- pansion of the towered enclosure, before being aban- can church. This restructuring involved the addition doned and revived yet again in the 13th century. Since of massive internal pillars to support a domed roof, written sources are few for this region, the archaeo- and a radical alteration of the church's furnishings. logical evidence takes on primary importance- Thus, two distinct levels of the opus sectile pave- particularly for the poorly understood transition ment in the bema can now be understood in con- from Byzantine to Turkish control. Whittow's report nection with this phasing. The location of the high for the 1993 Odemi? survey can be found in AnatSt altar in the middle of the bema floor was also iden- 44 (1994) 187-206 (with H. Barnes). tified thanks to fragments of the altar base still in . For inscriptional material, see situ. Precise evidence for secular reuse of the church L. Jonnes ed., The Inscriptions of Heraclea Pontica in the 13th century, dated by three coins issued by (Inschriften griechischer Stidte aus Kleinasien 47, the Seljuks of Rum, was concentrated in the eastern Bonn 1994). end where the apse and its superstructure must have Ciftlik/Sinop. Urgent salvage work to record the better survived the church's destruction. Archaeo- Byzantine and earlier site at Qiftlik, south of Sinop, botanical samples and carbonized sheep droppings before it erodes further into the sea, was carried out indicate that the structure was used as a farmyard in 1994 by Stephen Hill in cooperation with the for agricultural produce and animal husbandry. Sinop Museum. Excavation efforts concentrated on "Work on the Lower City fortifications focused on the Early Byzantine church with a square nave and the gateway west of the triangular tower, which - late fine mosaic floors, which show several phases of re- in its history-had been blocked by a crude stone pair. It may have been associated with an agricultural wall at its northern, inner end. The wall was as- estate and a monastic community. Earlier settlements sembled from reused blocks, including a keystone are indicated by Bronze Age artifacts. The mosaics decorated with a raised cross medallion, and prob- have been lifted and taken to the Sinop Museum. ably dates to the 11th century like other rubble dwell- A preliminary report on the 1993 survey appears ings constructed inside the gate. At the outer end in British Institute ofArchaeology/Ankara Research Reports of the gateway, the horizontal lintel blocks were still 1994, 3-5; and AnatSt 44 (1994) 15-16. lying in a row across the entrance (fig. 43). A coin For the nearby amphora factories at Demirci (1 km from the surface immediately below the collapse in- distant), see above, under "Propontis, Bithynia, dicates that the gateway was still standing until the Thrace, Pontus: Zeytinlik/Sinop, Demirci." last quarter of the 10th century. The lintel's interlock- Trabzon-Giimiighane Survey. James Crow ex- ing blocks recall W.J. Hamilton's description of an tended his survey area in 1994 from the Arakli- arch still visible in 1836, when he conclusively iden- Trabzon-Macka region of the two previous seasons tified this site as . toward the south and Gimii?hane, in order to dis- "In the Upper City, a second season in the north- cover the Byzantine communications network link- ern slope step trench did much to clarify the chro- ing communities, monasteries, and fortresses across nology and stratigraphy of the latest defensive sys- a highly irregular and difficult terrain. For the 1993 tem, which involved cutting back the mound to create survey, see AnatSt 44 (1994) 16-18. a steeper rampart and the construction of a substan- Amorium. The seventh (1994) excavation season tial fortification wall. A new 30-m trench inside the progressed in the Lower City's two ongoing areas of fortifications exposed a range of Late Seljuk and investigation, the church and the southwestern gate, Early Ottoman features and surfaces, including a as well as on the northern side of the Upper City room with fallen roof beams. These deposits, which mound. Survey in the northern foothills of the represent the final stages of permanent occupation Emirdag Mountains also documented the intensive in the 14th and 15th centuries, reflect the decline Roman and Byzantine exploitation of the hinterland of the site under the Seljuks, and Amorium's trans- connected to this regional urban center. C.S. Light- formation from a major Byzantine fortress city to foot kindly provides this assessment: a rustic backwater. The archaeological findings from "Excavations in the church concentrated on un- this part of the Upper City, together with historical covering the floor in the bema and naos, with muchand demographic documents, and environmental new evidence for the occupational history of analysis, the should provide a unique glimpse into the building. Two major construction phases are lifenow of a rural community in the Late Medieval clearly attested, rather than three as previously period." thought: the primary phase of the Late Antique aisled Reports for 1994 can be found in AnatSt 45 (1995), 1996] ARCHAEOLOGY IN TURKEY 333

Fig. 43. Amorium. Lower City gate: collapse and British Institute of Archaeology/Ankara monasteries, three small churches, one of several Research ports 1994, 8-11; for 1993, cemeteries see at the site,AnatSt and two chamber 44 complexes (1994) 10 and KST 16:2 (1995) 131-51. arranged in rows. The season's surprise came with Catalh6yiik. For the Byzantine remains, see above, the discovery of two 'underground cities' connecting under "Later Neolithic." to one of the rock-cut churches. The entrances were Canh Kilise-Akhisar. Although the Canll Kilise protected by an elaborate series of rolling stone is usually cited in handbooks as an example of Con- doors; and, although they appear to follow a village stantinopolitan style imported to provincial Cappa- scale, some of the rooms are exceptionally large (one docia, it has received little architectural study since measuring 35 m long) in comparison with the more Gertrude Bell's brief visit in 1906. Robert Ouster- famous examples in eastern Cappadocia. They cer- hout, University of Illinois at Urbana-Champaign, tainly served as refuges, since they are linked to build- kindly offers a summary of his first three-week ingssur- above ground. It is particularly significant that vey season to document more fully the church theyand would date, like the village, to the 10th-12th its surrounding village: centuries, well after the time of the Arab invasions "Much of our effort in 1994 was directed to the usually cited for other underground cities. preparation of a new plan, elevations, and sections "The nature of this settlement (and others like it of the elegant brick and stone church (fig. 44), usu- in the region) must also be reevaluated. Rather than ally dated because of its frescoes to the 11th century being strictly monastic, it may well represent a typi- with a few later additions. Its naos was constructed cal Byzantine town, with houses of varying sizes, with exceptional care, the dome's diameter measur- barns, and stables, coexisting with a few monasteries. ing nearly 15 Byzantine feet and half the width ofThis issue will require a fuller documentation of the the naos, in contrast with less exacting attention paid site in future seasons." to other sections of the building. Of particular in- Cappadocian Underground Cities Survey. A typo- terest is the use of brick almost exclusively as facing logical study of the multilevel underground refuges material, thus a veneer imitating techniques in the and villages that characterize the Cappadocian high- capital without affecting the local tradition of mor- lands was carried out for a fourth season in 1994 tared rubble and especially stone. by Roberto Bixio with a team of speleologists. They "We also wanted to situate the church within its have located 172 such settlements in the broad area architectural landscape- a settlement extending to between Yozgat, Nigde, and Aksaray, and recorded the east and west over perhaps 2 km. We were able a wide variety of architectonic systems, some of them to survey a 100 x 250 m area around and east of previously unknown. Particular attention is being the church, and draw a plan of its rock-cut building given to the hydraulic and drainage tunnels that remains and topographic features. Four distinct areas tapped underground streams for the settlements and were located here, comprising two so-called courtyard their fields, as well as to converse techniques for di- 334 MARIE-HENRIETTE GATES [AJA 100

Fig. 44. Canhl Kilise-Akhisar. Church from verting and storing excess second building rainwater. phase, and carved arcades A withreport Con- o first two seasons is publishedstantinopolitan features. in The exteriorAra?ST arcade and 11 43-56. its neighboring street are now completely exposed. Ak6ren (northern Cilicia). This sixth-century A.D. Pottery, including Cypriot types, spans the sixth to site 50 km north of Adana in the upper Seyhan val-the lth century. Restoration work on the church's ley recalls the "dead cities" of northwestern Syria bothfrescoes continued. For the 1992 and 1993 progress in its architecture and remarkable state of preser- reports, see KST 15:2 (1994) 475-86 and 16:2 (1995) vation. Gabriele Mietke, during a first 1994 survey, 361-75. focused her attention on two of its churches (one ISLAMIC with figural sculpture) and a mausoleum. For inscrip- tions, preliminary photographs, and comparable sites The following site summaries were generously p in the region, see M.H. Sayar's 1993 survey notes pared in by Scott Redford, Georgetown University Ara?ST 11 (1994) 137-60. Citadel. In the course of the 1994 ex Kilise Tepe. For the Byzantine period, see above, vations, Olu? Arik uncovered an eyvan, thought under "Chalcolithic and Bronze Age: Eastern, North- be the royal audience hall of the Seljuk palace. T ern, and Central Anatolia." open hall, facing the innermost courtyard of Gemiler Island/Fethiye Survey. The extensive palace, was flanked by two rooms whose entranc Medieval colony on the Lycian coast and offshore have yet to be identified. The remains of piers in islands was investigated for a fifth season in 1994 cate that it was also fronted by a portico. This arc by Shigebumi Tsuji with Kazuo Asano. The project's tectural schema recalls the larger palace at Kuba painstaking documentation of pilgrimage centers, bad, as do underglaze painted tiles (eight-poin necropolises, and churches, many with well-preserved stars, cruciform ones, and tile mosaics). Two tiles wall paintings, has appeared in a first preliminary mained in situ on the courtyard facade. The 1 report edited by S. Tsuji: The Survey of Early Byzantine season is presented in KST 15:2 (1994) 579-82. Sites in the Oliideniz Area (Memoirs of the Faculty ofLetters, Haci Baba Kiosk, Alanya. Scott Redford's 19 Osaka University 25, Osaka 1995). The 1993 season survey resulted in a site plan of this two-story pavi is summarized by K. Asano in Ara?ST 12 (1995) and its enclosure wall, 6 km from Alanya, as wel 407-19. plans, sections, and elevations of the building its Demre. In 1994, Yildiz Otiiken's excavations inside The complex can be dated to the second quarter the church of St. Nicholas proceeded to its interior the 13th century by its masonry, and by its inte arcade and podium sector. Architectural finds in- and exterior geometric fresco decoration. clude opus sectile flooring from the eighth-century, Kubadabad. Ruilhan Arik's 1994 season concen- 1996] ARCHAEOLOGY IN TURKEY 335

trated on the area surrounding thought to be a granary, and the what appears smaller to be a of this Seljuk-period site. military Thecanteen with road large ovens, between a well, and a water the and smaller palaces was channel. found The 1992 report appearsto runin KST 15:2 paralle(1994) the water system. The 513-32. enclosure wall and a po fountain were also investigated. Aksaray. Bekir Deniz beganGlaze excavations slag in 1994 reco in these areas underscores at the building the locally probability known as the "Darphane" orthat were being produced Mint,locally. which is in fact Soundings the zaviye, or dervish lodge, unco of extensive Early Bronze the Melik remains Mahmoud khan. Inhere, the process butof clear- no of the Byzantine period ing the easternattested and southeastern in sides otherof the struc- se Dendrochronological samples ture, a fragmentary from inscription thewas found norther that may of the larger palace give date partthe of the reassuring building to A.H. 711/ A.D. confirm 1310/11. that all itsjuniper pilings Tiles associatedwere with cut this structure, in 1231, however, and well its the reign of the builder, typological Alaeddin similarity to other Keykubad. zaviyes, would suggest Fo 1992 findings, see KST an initial 15:2 foundation (1994) in the 13th century.533-46. Erci? Fortress (Celebibagi). Iznik kilns. The long-term Salvage investigations operati of the against the rising waters famous ofOttoman-period Lake ceramic Van industry were continued con with increased urgency in 1994 under by the directionAbdiisselam of Oktay Aslanapa and Ul The entire site is now Araunder Altun. In the water,area south of the exceptHaci Hamza baths, for fortress's necropolis (whose workshops for low-lyingtiles were discovered reinstalled areas in- ar threatened). The 1994 effortsside a structure datableconcentrated to the fourth century A.D. on t ferring the late 12th-early Water supply via pipes15th was shared century with the nearby Is sarcophagi, tombstones, bathhouse. and The team burials-also restored several manypreviously to the Karakoyunlu dynasty- excavated kilns, as partto of thethe rapidly top developing of but the Despite weathering, they underfunded are Iznik often preservation zone.well The 1992 preser sea- and of a high artistic son isquality discussed in KST that 15:2 (1994) will 547-63. espe interest students of calligraphy. Ottoman kilns, Istanbul. Soundings When budgetary con- als firmed the presence of straints earlier in 1994 prevented cremation Filiz Yenigehirlioglu burial from the 1992 season, see excavatingKST soundings 15:2 in the(1994) Ottoman potters' 487-51 district . The major investigation around the Tekfur Sarayi, ofshe conducted Beyhan an ethno- K magarall's 1994 campaign archaeological surveywas of thea fewlarge traditional work-build perhaps a palace; its stratigraphic shops still surviving in Eyfip context and other districts andof mentary carved portal Istanbul. elements Their kilns closely suggest resemble those that from it to the lth century, after Iznik, and the the two industriesarrival were indeedof linkedthe in Tu This structure is multistoried, the later Ottoman period. with Clay sources two used by stairc the The lower floor, at basement current workshops level,will serve for mustcomparative analy-have lit from above since no sis windowswith the historical ceramicswere and found. tiles collected It sisted of a room with duringa raised earlier surveys. sleeping A detailed report onplatfor the pre- granary, and other viousrooms campaign appears with in Ara?ST lamp 12 (1995) 535-66. niche hearths that may have served for storage or as s It is suggested that the DEPARTMENTmain OF living ARCHAEOLOGY ANDareas were l on the upper story or stories. HISTORY OF ART . Nurettin Yardimci and his team focused FACULTY OF HUMANITIES AND LETTERS their 1994 excavations around the Aleppo Gate, clear- BILKENT UNIVERSITY ing the city wall on either side of it, and restoring 06533 BILKENT part of the defensive wall there. They also worked ANKARA, TURKEY on two buildings inside the gate: a vaulted structure, [email protected]