Survey Archaeology and the Historical Geography of Central Western Anatolia in the Second Millennium BC

Home , Beycesultan, Larisa (Ionia), Larisa (Lydia), Larisa (Troad), Lydia, Lydians

European Journal of Archaeology 20 (1) 2017, 120–147 This is an Open Access article, distributed under the terms of the Creative Commons Attribution licence (http://creativecommons.org/licenses/by/4.0/), which permits unrestricted re-use, distribution, and reproduction in any medium, provided the original work is properly cited.

The Story of a Forgotten Kingdom? Survey Archaeology and the Historical Geography of Central Western Anatolia in the Second Millennium BC

1,2,3 1,3 CHRISTOPHER H. ROOSEVELT AND CHRISTINA LUKE 1Department of Archaeology and History of Art, Koç University, I˙stanbul, Turkey 2Research Center for Anatolian Civilizations, Koç University, I˙stanbul, Turkey 3Department of Archaeology, Boston University, USA

This article presents previously unknown archaeological evidence of a mid-second-millennium BC kingdom located in central western Anatolia. Discovered during the work of the Central Lydia Archaeological Survey in the Marmara Lake basin of the Gediz Valley in western Turkey, the material evidence appears to correlate well with text-based reconstructions of Late Bronze Age historical geography drawn from Hittite archives. One site in particular—Kaymakçı—stands out as a regional capital and the results of the systematic archaeological survey allow for an understanding of local settlement patterns, moving beyond traditional correlations between historical geography and capital sites alone. Comparison with contemporary sites in central western Anatolia, furthermore, identifies material com- monalities in site forms that may indicate a regional architectural tradition if not just influence from Hittite hegemony.

Keywords: survey archaeology, Anatolia, Bronze Age, historical geography, Hittites, Seha River Land

INTRODUCTION correlates of historical territories and kingdoms have remained elusive. The historical geography of western This is hardly a novel situation. Long Anatolia in the second millennium BC has traditions of Biblical and Classical archae- become much clearer in recent decades. ology in the Mediterranean have roots in New discoveries and re-interpretations of attempts to connect material remains to both rupestral monuments in western textually attested peoples. In Anatolia Anatolia and archives from Hattusa, the itself, Hittite archaeology proper began in Hittite capital in north central Anatolia, the 1870s and 1880s with William have led to general consensus, if not unan- Wright’s(1874, 1884) and Archibald imity, on the broad outlines of many geo- Henry Sayce’s(1880, 1888) inspired con- political units in the area, from their first nections between previously unidentified appearances in Hittite texts to their inclu- hieroglyphic inscriptions, the biblical sion under imperial control. Yet while the ‘Hittites’, and the Egyptian ‘Kheta’, historical map appears more and more unveiling the Story of a Forgotten Empire, clearly delineated, the archaeological as Sayce called it. Similarly, Heinrich

© European Association of Archaeologists 2016 doi:10.1017/eaa.2016.2 Manuscript received 18 May 2015, accepted 10 February 2016, revised 29 January 2016 Downloaded from https://www.cambridge.org/core. IP address: 170.106.33.22, on 26 Sep 2021 at 16:00:59, subject to the Cambridge Core terms of use, available at https://www.cambridge.org/core/terms. https://doi.org/10.1017/eaa.2016.2 Roosevelt and Luke – The story of a Forgotten Kingdom? 121

Schliemann’s privileging of textual evidence for a singular capital site, but also for in his search for Homeric Troy is linked thriving communities in its hinterland. ontologically with Manfred Korfmann’sre- Our presentation of the second-millen- covery of empirical datasets at Troy that nium BC archaeology of the Central Lydia support arguments that the Hittites once Archaeological Survey (CLAS) study area knew the area as Wilusa (Starke, 1997; begins with a review of the historical geog- Korfmann, 2006). In such cases, however, raphy of central western Anatolia as illu- survey archaeology has played only a minor strated by Late Bronze Age Hittite role compared to excavation in helping to interests in and general information about flesh the bones of historical interpretation the area. We then discuss survey methods with material realities (see Matthews & and summarize the results analysed to Glatz, 2009). What is the potential for date. With the discovery of six fortified archaeological survey in previously unsys- citadels, five unfortified settlements, and tematically studied regions to recover data- at least twenty-three other sites of contem- sets that link well with and extend the porary activity in the region, this survey textual record in ways that excavation at project illustrates the presence of a well- single sites cannot? developed constellation of local communi- As of the late 1990s, the middle river ties in central western Anatolia that valley of the Gediz (the Hermos of greatly complements our understanding of Classical times) was one such unsystemat- regional settlement from texts alone. ically studied region. With few exceptions, Comparison with contemporary sites and historical and archaeological investigations materials in western Anatolia discovered had focused on the Iron Age kingdom of via survey and excavation also shows that Lydia in the first millennium BC, with this new evidence represents one of the visions of its capital Sardis commanding fullest pictures we have of a western vast hinterlands in Lydian and subsequent Anatolian settlement system of the second times. Yet in a 1998 article, Hittitologist millennium BC, from fortified citadel to J. David Hawkins suggested that Sardis, agricultural countryside. or a site in its environs, may have been an earlier capital of a forgotten Bronze Age kingdom too: the Seha River Land, an in- EMPIRE AND KINGDOM:HISTORICAL digenous realm and later Hittite vassal GEOGRAPHY IN LATE BRONZE AGE state recorded in Hittite texts of the Late WESTERN ANATOLIA Bronze Age (c. 1650–1200 BC). With the express purpose of examining the immedi- The story of the Hittites of north central ate hinterland of Sardis in all periods of Anatolia is well known to scholars of human history—not just the Late Bronze Mediterranean history and archaeology Age or Lydian periods—the Central Lydia (e.g. Bryce, 2005). By the mid-second mil- Archaeological Survey (CLAS) worked lennium BC,kingsrulingfromHattusa,a between 2005 and 2014 to explore an area capital city of monumental walls, temples, of c. 350 km2 around 10 km north of and palaces, interacted diplomatically and Sardis across the Gediz Valley. Among militarily with neighbouring peoples in all results from diverse periods, the work directions. Best known are Hittite interac- offers the best candidate yet known for a tions with Babylonian, Mittanian, Assyrian, capital of the Seha River Land. and Egyptian areas to the south and south- Furthermore, because of its systematic ap- east, where the conquest of previously inde- proach, the survey reveals evidence not just pendent states transformed the central

Downloaded from https://www.cambridge.org/core. IP address: 170.106.33.22, on 26 Sep 2021 at 16:00:59, subject to the Cambridge Core terms of use, available at https://www.cambridge.org/core/terms. https://doi.org/10.1017/eaa.2016.2 122 European Journal of Archaeology 20 (1) 2017

Anatolian kingdom into a hegemonic Kurunta over a combined group of Mira empire (Glatz, 2009). Of numerous king- and Kuwaliya, Manapa-Tarhunda over a doms affected by Hittite growth over the combined group of the Seha River Land courseoftheLateBronzeAge,agroupof and Appawiya, and, lastly, Alaksandu over politically aligned lands in western Anatolia Wilusa (Beckman, 1999:69–93). knowntotheHittitesasArzawa,orthe The precise locations of these vassal Arzawa Lands, remains one of the least kingdoms in western Anatolia have gained understood. resolution recently, after nearly a century Arzawa enters history fleetingly in of scholarship. Their existence, based on records of the seventeenth and sixteenth evidence provided by cuneiform tablets in century concerning the military campaigns archives at Hattusa, has been known since of Hattusili I and perhaps also those of his early in the twentieth century, yet only in grandson and successor Mursili I the last thirty years have nearly continuous (Heinhold-Krahmer, 1977:19–21; Bryce, epigraphical, archaeological, and topo- 2003:46–47, 2005; Yakubovich, 2008: graphical discoveries and syntheses 97). Little is revealed about the area at enabled their more confident placement that time, however, and it is not until on the map (Figure 1). Particularly influ- much later in the second millennium that ential in these geographical reconstructions more information is gained. were the masterful works of Frank Starke The annals of king Tudhaliya I/II de- (1997) and J. David Hawkins (1998) (see scribe campaigns against Arzawa, the Seha also Mountjoy, 1998, Sarı, 2013, and River Land, and Hapalla in the early to Alparslan & Doğan-Alparslan, 2015). mid fourteenth century (Heinhold-Krahmer, Wilusa, commonly associated with the 1977: 256–58; Bryce, 2003:48–49). By this Greek Ilios, has been tied to the Troad in time Arzawan kings had become powerful north-western Anatolia, with its capital at enough to threaten Hittite sovereignty and Troy, the only known fortified site in the negotiate marriage alliances directly with region (Pavúk & Schubert, 2014; Rose, Amenhotep III of Egypt (Moran, 1992: 2014:25–43). Mira and Kuwaliya prob- 101). According to later fourteenth- and ably coincided with the valleys of the thirteenth-century archives, Arzawan kings Küçük and Büyük Menderes (the Classical made their capital at Apasa, now identified Cayster and Meander) rivers, having a more confidently than ever as a predecessor capital at Apasa (classical Ephesos) and of later Ephesus (Büyükkolancı, 2007, sharing an inland border with Hapalla 2008). These kings repeatedly and rebelli- likely to be in the vicinity of modern Afyon ously associated with the representatives of (Hawkins, 1998:23–25, 31). The area Ahhiwaya, generally associated with a between these kingdoms, from the Classical Mycenaean Aegean realm that threatened Troad in the north to the Bozdağ (Classical the maintenance of Hittite hegemonic Tmolos) range in the south, and to unspeci- control (Beckman et al., 2011). Owing to fied points eastward, probably belonged to such associations with the enemy, the theSehaRiverLandandAppawiya. Hittite king Mursili II punished the coali- Appawiya can be connected to Classical tion of Arzawa Lands by dividing them into Mysia Abbaitis, located at the headwaters of at least four component parts, each kingdom the Simav (Classical Makestos) River, just to be ruled by a local lord whose power was north of the Simav (ancient Temnos) range. held in check by comprehensive treaties of The Seha River Land is most likely to have vassalage. Accordingly, acertainTargasnalli encompassed the Bakır (Classical Kaikos) came to rule over Hapalla, Kupanta- and broad Gediz (Classical Hermos) river

Figure 1. Map of Late Bronze Age western Anatolia, with the location of significant kingdoms following Starke (1997) and Hawkins (1998).

valleys (Radt, 1992;Eastonetal.,2002: 98; periods, especially in Bin Tepe, or the Hertel, 2011). Until recently no place of ‘Thousand Mounds’—a72km2 tumulus sufficient archaeological significance—and cemetery commonly identified as the royal nothing that would suggest a regional cemetery of Lydian kings at Sardis, located capital—had been identified within this some 10 km to the south (Roosevelt, 2007, broad territory associated with the Seha 2009; Luke & Roosevelt, 2016). A few River Land, leading Hawkins (1998: 24) to intermittent investigations had also located prescribe constructively that, if not at Sardis, traces of Bronze Age remains dating to the the capital of Iron Age Lydia, ‘alarge third and second millennia BC (Hanfmann, BronzeAgesiteintheHermos[Gediz] 1968;Mitten&Yüğrüm, 1971, 1974; heartland should be sought’. Zimmerman et al., 2003). The interrelation between these sites, the presence of other sites, and the cultural systems they reflected SEEKING NEW SITES:REGIONAL SURVEY were all but unexplored aside from prelimin- IN CENTRAL WESTERN ANATOLIA ary presentations of material typologies intended primarily to assess chronology. Between 2005 and 2014, the CLAS A primary goal of CLAS was to fill out focused on the diachronic archaeology and this patchy understanding of the lake basin history of an area some 350 km2 in size, with a multipronged investigation of the surrounding Lake Marmara (Classical long-term record of human activity in the Lake Gygaia or Koloë) in the middle region and its relation to socio-political and Gediz Valley of central western Anatolia economic developments as well as chang- (Figure 2). Previous work in the area had ing environmental conditions. Accordingly, focused primarily on Iron Age and later CLAS adopted a variety of approaches

Figure 2. Map of the survey area of the Central Lydia Archaeological Survey showing the locations of sites mentioned in the text. Inset shows the location of the Marmara Lake basin in western Turkey.

including systematic, fieldwalking survey et al., 2014, 2015a, 2016; Luke et al., on a regional level, intensive gridded 2015;Luke&Cobb,2013). survey of selected sites, ground-truthing of Fieldwalking survey and bearing-based satellite imagery, geophysical prospection, survey followed standard research methods analysis of ancient and recent documen- with walkers spaced along set intervals of tary records, contemporary ethnography, 10–20 m, depending on the nature of the and paleoenvironmental investigations re- survey unit. Cultural material on the lating to paleoclimatic reconstructions and surface was counted by category (ceramic, the life history of Lake Marmara. Detailed tile, lithic, glass, etc.) and diagnostic discussions of methods and results from examples were collected for analysis in the each field season have been published in laboratory (except in 2013, when govern- the Arasţırma Sonuçları Toplantısı annual mental policies shifted, preventing material volumes and elsewhere (see Roosevelt, collection). These methods resulted in 10– 2007, 2009, 2010; Roosevelt & Luke, 20 per cent coverage of each survey unit. 2008, 2009, 2010, 2011, 2012, 2013; In addition, sites that had particularly Luke & Roosevelt, 2009, 2016, Roosevelt dense surface remains were selected for

more intensive survey with precise record- sub-ware varieties (e.g. Gold Wash, ing of surface remains and microtopogra- Slipped, Burnished, Painted) occurring phy (Roosevelt, 2014). Because no across a number of shapes (Figure 3). The regional chronology had been established commonest shapes are pedestalled bowls for the area, efforts focused on analyses of and cups and bead- and thickened-rim ceramic fabric and surface treatment to bowls and basins with ring-foot bases, create a working diachronic typology from some with horizontally attached upright Chalcolithic to Ottoman times (see Luke loop handles that are often deeply incised. et al., 2015). All shapes commonly feature horizontal CLAS results have greatly expanded our ridging on their exteriors. Gray Ware, understanding of prehistoric activities in another, rarer, medium- to fine-fabric the area, with stone tools dating to the ware, also has decorative sub-ware varieties Lower Palaeolithic (Roosevelt et al., (e.g. Silver Wash, Brown Wash, 2014: 340), several Chalcolithic sites of Burnished) and occurs in a similar variety the fifth and fourth millennia BC of shapes, though they are generally (Çilingiroğlu, 2015), and several more set- smaller in size (Figure 4). The coarse- tlements of the Early Bronze Age third fabric Red-Brown Coarse Ware occurs in millennium BC (Luke et al., 2015). The a wide variety of rounded and large jars, project attested also the vibrancy of com- the bodies of some of which feature munities settled around the lake continu- rounded to oval knobs (Figure 4). ing into later times, especially the Iron Parallels for these wares and shapes are Age Lydian, Achaemenid Persian (Luke well known from contemporary second- & Roosevelt, 2016), Middle and Late millennium sites in western Anatolia. To the Roman, Late Medieval, and Ottoman south and east, these include Beycesultan periods, up to the present day (Luke & (V–IVa and III/II; Lloyd & Mellaart, 1965: Cobb, 2013). 80–81, insert 1; Mellaart & Murray, 1995: Among the most surprising findings 21–22; Dedeoğlu & Abay, 2014)and was the identification of a group of Gordion (Gunter, 1991: 29). Elsewhere in roughly contemporary communities inha- inland western Anatolia, similar wares and biting both the peaks and valleys sur- shapes appear at Aphrodisias (Joukowsky, rounding the lake in the second 1986: 295, 323–27), Çine-Tepecik (Günel, millennium BC. These were dated primar- 2008: 136), BademgediğiTepe(Meriç& ily by ceramic material, drawing from Mountjoy, 2002;Meriç,2007:12),Ulucak western Anatolian datasets and the region- (Çilingiroğlu et al., 2004), and sites in the al typology developed from macroscopic, Akhisar plain (French, 1969). Further west, chemical, and petrographic analyses. excavationsonorneartheAegeancoast While survey results cannot provide have produced other comparanda from Troy precise dates in the way in situ excavations (VI; Blegen et al., 1953:38),Panaztepe may, we now have solid evidence for (Günel, 1999), Larisa and Bayraklı (10–14; second-millennium BC citadels and other Bayne, 2000:35,69,79),LimanTepe(III; settlements. The corresponding ceramic Erkanal & Şahoğlu, 2012:227),and repertoire includes two medium- to fine- Çesme-Ba̧ ğlararası and Kocabaş Tepe fabric wares and one coarse-fabric ware (Aykurt, 2008, 2010b). While comparison of (Luke et al., 2015: 433–36). ceramics from this broad range of sites is be- Red-Light Brown Ware, a medium- to ginning to allow the separation of Middle fine-fabric ware, is by far the most abun- from Late Bronze Age products (e.g. dant and includes a selection of decorative Mellaart & Murray, 1995:vol.2.2,21–22,

Figure 3. Selected Middle and Late Bronze Age Red-Light Brown Ware ceramics from lowland and citadel sites around the Marmara Lake basin.

57; Bayne, 2000: 33, 79), a significant survey. The division of the second millen- amount of continuity across these periods nium BC into Middle and Late Bronze Age still confounds precise dating (see Aykurt, phases according to the traditional dates 2013;Pavúk,2015). This is especially so for (c. 2000/1900–1650 BC and c. 1650–1200 ceramic material recovered from surface BC, respectively; Bryce, 2011; Yakar, 2011;

Figure 4. Selected Middle and Late Bronze Age Gray Ware and Red-Brown Coarse Ware ceramics from lowland and citadels sites around the Marmara Lake basin.

Sarı, 2013;Pavúk,2015) is only minimally sizes and densities in nearby locales. At each helpful with respect to the ceramic repertoire citadel and at selected valley sites, combina- of central western Anatolia. We are forced tions of intensive gridded collections of to generalize that the second-millennium BC movable material remains (e.g. ceramics sites of the Marmara Lake basin were occu- and lithics), resultant material analyses, and pied primarily in the middle part of the mil- study of immovable remains (e.g. architec- lennium, between the eighteenth and ture, bedrock modifications), via satellite- thirteenth centuries BC. image analysis and microtopographic surface Of the thirty-four sites of this period, mapping with both ground-based and aerial local communities fortified at least six ridges methods, provide rich datasets from which and hilltops as defensible citadels and occu- to interpret overall site extent, internal or- pied a similar number of places of varying ganisation, and relative chronology. Our

Figure 5. View of the occupation mound of Kılcanlar Höyük looking southeast.

working hypothesis is that these sites form settlements. Discovered at a place called thecoreofanextensivecommunityrepre- Boyalı Tepe and unfortunately not relo- senting an indigenous kingdom in the heart cated recently, this cemetery consisted of of the middle Gediz Valley. graves within large jars, or pithoi, many fragments of which were collected and tentatively dated to the Middle Bronze ASCATTERING OF LOWLAND Age (Hanfmann, 1968: 10). SETTLEMENTS Both the twenty-three low-density scatters and the five relatively higher-density set- Of the thirty-four sites discovered during tlements were found distributed across the thecourseofCLASworkanddatedtothe area, on or between the gently rolling foot- second millennium BC based on material hills surrounding Lake Marmara, if not on remains, twenty-three were characterized by the valley floor itself. Such places were well their small size and low ceramic densities situated to take advantage of fresh water (Figure 2). These twenty-three small sites, sources (e.g. local springs and streams) as or scatters—each represented by a handful well as fertile and well-drained lands for of sherds (typically amounting to 0.005– agriculture and more rugged uplands for 0.010 sherds/m2 spread over no more than grazing. All but one of the five settlement 0.05 ha)—are likely to represent the ephem- areas bore evidence for earlier occupation in eral traces of rural activities. Larger, prob- the third millennium BC,andallsawvarying ably more permanent, communities are intensities and durations of settlement after represented by five additional sites identified the second millennium BC, attesting to the by ceramic scatters of relatively higher long-term benefits of such locations. East, density, architectural remains, and/or occu- north, and southwest of the lake, respective- pation-mound form (Figure 5). These five ly, the settlements at P2012.12, Buğdaylık larger settlements ranged in area between and Kılcanlar Höyük, and Razoğlu Höyük ̇ less than 1 ha and c.4.5haandhadamean and Ikiz Tepe appear well situated also for density of 0.164 sherds/m2 (Table 1). monitoring communications in those A cemetery identified by members of general directions. While these settlements the Archaeological Exploration of Sardis demonstrate certain locational preferences, during their work in Bin Tepe in the late citadels in the Marmara Lake basin reveal 1960s can be added to this small group of even more strategic situations.

Table 1. Scatters, settlements, and citadels in The Marmara Lake basin. Site type & ID Size (ha) Sherds Density (#) (sherds/m2)

Scatters S2005.71; P2005.17, 43; S2006.39, 62, 141, 306; S2009.489; T2009.4, <0.05 1–7 0.005–0.010 8; S2010.137, 185, 558; P2010.2; S2011.20; S2012.982, 1438; P2012.3, 7; S2013.101, 102, 103, 650 (23 in total) Settlements Razoğlu Höyük (P2010.3) <1 19 0.1079 Kılcanlar Höyük (P2009.1) >2.5 67 0.2405 P2012.12 c. 4.1 31 0.01098* ̇ Ikiz Tepe (P2010.9) >4.5 141 0.1429 Buğdaylık (P2006.22) <10* >15 0.00038* Citadels Kızbacı Tepesi (P2007.4) 0.7 / 34 118 0.1483 Koca Dere (P2012.14) 1.1 / 2.5* 151 0.1372 Gedevre Tepesi (P2007.1) 1.2 / * 117 0.1104 Asar Tepe 1 (P2005.30) 2.1* / * * * Asar Tepe 2 (P2006.24) 3.4 / * 69 0.0088* Kaymakçı (P2006.1) 8.6 / >25 373 0.1200

Sites are presented in order of increasing size within each category. All sizes and densities are approximate. The two size values given for citadels correspond to the fortified area and total built environment, respectively. Sherd numbers refer to those datable to the second millennium BC. Asterisks mark site sizes and densities that are unreliable owing to lack of survey and/or non-systematic collection methods.

ASUITE OF CITADELS (DEMs) resulting from surveys of microtopography (Roosevelt, 2014). The exter- The six citadels in the Marmara Lake ior wall circuits of the citadels are easiest basin are defined here as settlements that to identify, as they tend to follow the are sited on naturally elevated hilltops or natural topography and are thus curvilinear ridges providing strategic views of sur- in form. In contrast, gates and interior fea- rounding landscapes and/or routes of com- tures (e.g. streets, houses, courtyards) are munication and that are fortified by at typically rectilinear. In addition to architec- least one circuit of walls enclosing archi- tural remains, settlement at citadels is indi- tectural complexes built with frequent re- cated by scatters of varied cultural material, course to the advantages of bedrock including flaked and ground stone tools outcrops (Figures 2 and 6). These features and especially ceramics, the latter with an are perceptible in both archival aerial average density of 0.127 sherds/m2. imagery from the 1940s to the 1990s and In contrast to the relative uniformity of QuickBird satellite imagery from the general topographical and surface material 2000s. Shadows cast by natural raking characteristics, the citadels of the Marmara sunlight can delineate these features in Lake basin vary in size, ranging between such images, but they are best visualized 0.7 ha and 8.6 ha in internal structure, as analytically via digital elevation models indicated on the surface by exposed wall

Figure 6. Panorama of the Marmara Lake basin taken from the northwest. Visible citadels are indicated by arrows; those out of view are in square brackets.

faces and slight undulations in microtopo- side. Cup-marks are found on adjacent stone graphy that reveal subsurface architectural outcrops. The natural bedrock topography remains, and in extent of extramural built and a series of terrace walls appear to form environments, including additional evi- stepped platforms rising to and defining a dence of settlement as well as conspicu- higher core of the citadel along its western ously modified bedrock outcrops. These fortification wall, surrounded on the north, latter features are defined by ‘cup-marks’ east, and south by relatively continuous ter- of roughly conical, concave form and races.Ontheflatpeakofthiscore,ac. 8× varying diameter and depth that appear 13 m rectangular building is oriented east- singly or in groups and are likely to have west and contains at least one interior cross- served ritual purposes common elsewhere wall. This and other architectural complexes in Late Bronze Age Anatolia (Neve, within the citadel appear to be oriented or- 1977/78, 1996; Luke & Roosevelt, in thogonally to their nearest exterior circuit press). These citadels are presented here in wall, itself defined by the trace of the natural order of increasing size of fortified space. topography, thus eschewing more comprehensive grid planning. The DEM resulting from microtopo- Kızbacı Tepesi (Figure 7) graphy survey reveals a c. 10 m-wide terrace that spirals clockwise around the Kızbacı Tepesi, also known as Kayalıtepe, is citadel, beginning c. 20 m below the the smallest, yet perhaps best defined, circuit wall on the east and perhaps pro- citadel in the basin, with a 0.73–1.08 m- viding a gently sloping route to the gate at wide circumference wall enclosing an area of the north. Extramural remains nestled 0.7 ha atop an elongated and roughly north- within these terraced features include nu- south oriented, steep-sided hilltop with merous structures, densest on the east commanding views to east, south, and west, slope, and clusters of cup-marks, especially 160 m above the plain. The site was first on the southeast and southwest ridges. identified in 2007 during regional survey; it Gridded survey within and outside the was mapped for microtopography in 2007 citadel produced ceramic material that, if and 2013 and documented with intensive not just the result of downslope erosion, gridded collections in 2012 (Roosevelt & demonstrates the contemporaneity of the Luke, 2009; Roosevelt et al., 2014, 2015a). terraced structures within and without the Primary access to the fortified area citadel, as well as later but not extensive appears to have been provided at its northern use of the eastern slope in Hellenistic and end by a c. 2 m-long passageway narrowing Roman times (Roosevelt et al., 2014). In to a width of 1.6 m between two square bast- sum, the survey of this citadel indicates a ions or towers measuring c. 5 m on one built environment spread over 34 ha.

Figure 7. Hillshaded digital elevation model (DEM) of Kızbacı Tepesi and view to the north.

Koca Dere (Figure 8) and was subsequently documented with in- tensivegriddedcollectionsandmicrotopo- Koca Dere is located along a narrow ridge graphic mapping in 2013 (Roosevelt et al., within the natural ravine of a perennial 2014, 2015a). stream flowing into the north-western The citadel’s circuit wall, here c. 2m marginofthelakebasin.Ratherthanbroad wide, encloses a 1.1 ha space elongated in views, its location c. 15–35 m above the the north-south direction. Numerous traces ravine floor provides visible control only of walls and rubble scatters are discernible over the communication route in which it is within the circuit and appear to form situated, as it opens from the foothills in the roughly defined terraces that step up from north to the plain below. The site was iden- west to east; dense maquis and bedrock tified first in 2012 during regional survey outcrops preclude clearer understanding of

Figure 8. Hillshaded DEM of Koca Dere and view to the north-northeast.

their organisation. Two large cup-marks opposite bank of the stream to the west. are found within the inner settlement, one Microtopography survey again reveals a at the centre of the site and one overlook- probable access to the site from the south, ing the southern ravine. where a c. 6 m-wide terrace provides Extramural remains identified during access to the site via a c. 3 m-wide path the course of gridded survey include only (Roosevelt et al., 2015a). sparse ceramics, likely to be the result of erosion from the citadel itself, and isolated cup-marks in the immediate vicinity of the Gedevre Tepesi (Figure 9) site that cover a total area of 2.5 ha. Slightly further away, more complex clus- Gedevre Tepesi, also known as Kale ters of small cup-marks are found on the Tepesi, consists of a curving circuit wall

Figure 9. Hillshaded DEM of Gedevre Tepesi and view to the north.

that encloses 1.2 ha of protected space A ramped approach at the southwest atop a circular bedrock knoll with unob- takes advantage of bedrock outcrops to structed views to the east, across the lake provide access to the site. Within the basin, as well as to the northeast and circuit wall, the buried width of which was south, at a height of c. 60 m above the impossible to determine, a continuous plain. The site was first identified in 2006 terrace c. 13–21 m in width surrounds an via QuickBird image analysis and was elevated inner space roughly a quarter-circle ground-truthed in 2007, mapped for in shape, within which a buried wall microtopography in 2008, and documen- defines a concentrically smaller, quarter- ted with intensive gridded collections in circle space (Roosevelt & Luke, 2010:6–7). 2012 (Roosevelt & Luke, 2009, 2010; As determined by gridded survey, extra- Roosevelt et al., 2014). mural remains include only dense collections

Figure 10. Hillshaded DEM of Asar Tepe 1.

of ceramics clustered just outside the circuit (Roosevelt, 2007, 2014; Roosevelt et al., wall, especially along the ramped approach, 2015a); yet it remains the least explored of that probably result from slight downslope the citadels of the Marmara Lake basin. erosionratherthanfromextramuralsettle- The site comprises the uppermost ment (Roosevelt et al., 2014). Cup-marks manmade terraces of 2.1 ha on a natural are found in isolation or in groups both hill surrounded by a circuit wall of dimen- within the circuit and along the ridge to the sions that are indeterminable because it is northwest of the site. buried. A narrow terrace just within the wall surrounds the steeply rising slopes of the upper terrace. Investigation of extra- Asar Tepe 1 (Figure 10) mural evidence was limited to the lower southern slopes of the hill, where the Located at the southern edge of Bin Tepe, presence of second-millennium ceramics Asar Tepe 1 overlooks the broad Gediz may have been caused by downslope River valley at c. 40 m above the river itself, erosion from above. The DEM resulting the modern course of which runs just to its from the microtopography survey suggests south. The site was first discovered in 2005 a possible access to the site at its north- and mapped for microtopography in 2013 eastern end.

Figure 11. Hillshaded DEM of Asar Tepe 2 and view to the northeast.

Asar Tepe 2 (Figure 11) The citadel’s circuit wall, the buried width of which was again impossible to de- Situated on a high and broad bedrock termine, encloses 3.4 ha of protected space, knoll north of the lake, Asar Tepe 2 beginning with a nearly continuous and 8– affords commanding views across the 15 m-wide circumferential terrace retained entire lake basin at c. 150 m above the by the wall itself. The interior edge of the plain, restricted only to the east by natural terrace is defined by rising bedrock, atop ridges descending to the lakeshore. The which numerous rubble cairns indicate the site was first discovered in 2006 during re- locations of ruined architectural complexes gional survey and was mapped for micro- whose organisation is undefined. topography in 2008 (Roosevelt & Luke, Remains outside the circuit include only 2008, 2010). a conspicuous stone cairn at its southern

Figure 12. Hillshaded DEM and resistance survey results of Kaymakçı, and view to the northeast.

end, perhaps an Iron Age tumulus, and descends to the western shore of Lake traces of possible access routes on all sides Marmara. At c. 110 m above the sur- just below the walls. These and the pres- rounding plains, the site has commanding ence of discontinuities in the circuit wall views in all directions and controls passage at the north, east, and south may suggest between the ridge itself and the lake. The the locations of access points to the site was first discovered in 2001 during a citadel. Additionally, the lowland settle- regional survey of tumuli (Roosevelt, 2006, ments of Kılcanlar Höyük and Buğdaylık 2009); its full documentation began only are located nearby. later with architectural survey and QuickBird image analysis in 2006, intensive gridded collections in 2007 and 2009, Kaymakçı (Figure 12) microtopographic and geophysical survey between 2007 and 2013, and, most recent- Kaymakçı, the largest and most complex ly (hence not reported here), excavations site in the area, is located on the lower, which started in 2014 (Roosevelt & Luke, south-eastern ridge of Gür Dağ,asit 2008, 2009, 2011, 2013; Luke &

Roosevelt, 2009;Roosevelt,2014; Roosevelt geophysical and excavation results. Breaks et al., 2014, 2015a, 2015b, 2016). in wall circuits are suggestive of accesses The site of Kaymakçı is defined by a into these central sectors as well as citadel as well as by extramural remains in- through the exterior circuit of the citadel, cluding evidence for settlement, a ceme- especially at the southeast, yet specific gate tery, and numerous bedrock outcrops configurations are indeterminable at bearing cup-marks. Cup-marks occur present. singly and in groups of varying sizes, and Gridded survey inside and outside the are located near access points inside the citadel revealed a high density of finds citadel as well as along what could be within the citadel, unsurprisingly, and a routes connecting the citadel with springs lower density outside the citadel, spreading and the cemetery (c.f. Neve, 1977/78, over the terraced lower slopes of the ridge 1996). All in all, the settlement area at and stretching to the plain on the south- Kaymakçı extends over a minimum of 25 west and to the modern shore of Lake ha, with the total built environment Marmara on the northeast (Roosevelt & spreading well over 1 km along the ridge. Luke, 2011: 57). A nearby cemetery of At the core of this area, the 8.6 ha citadel both pithos and cist graves of the second comprises several circuits of walls that millennium BC is identifiable owing to its enclose seemingly distinct and hierarchic- partial plundering. Isolated architectural ally organized sectors. complexes have also been identified along Within the sinuous contour-hugging the Kaymakçı ridge northwest of the outline of the lowest circuit of walls, a citadel, and cup-marks are found in abun- broad terrace at the southeast contained dance both inside and outside the citadel, neighbourhoods built on different orienta- especially on the edges of ravines that tions, divided by pebble-streets, and used incise both sides of the ridge. for mixed domestic and household-indus- try purposes, as suggested by ongoing geophysical survey and recently initiated excavations (Roosevelt et al., 2015b). At PERSPECTIVES FROM THE ARCHAEOLOGY the north-western edge of the citadel, AND GEOPOLITICS OF SECOND- massively constructed towers project MILLENNIUM BC WESTERN ANATOLIA outside the c. 2 m-wide wall where they protect the gentlest approach along the Without firm control of the chronologies ridge; from there the wall continues to the of each of these sites—excavations have east towards a triangular bastion before commenced at Kaymakçı alone, and returning to the southeast along the con- chronological refinements are still forth- tours of the ridge. Around the central, coming—interpreting their individual highest area of the citadel, a curving importance, their relative phasing and de- terrace wall defines an area roughly a velopment within the second millennium quarter-circle in shape divided by another BC, and the significance of the group as a curving terrace wall into two densely built whole, remains challenging. To be sure, spaces. On the slopes south of this central however, the dataset offers unprecedented area, yet still within the outer fortification detail in regional coverage that includes wall, additional widely separated building not only movable material culture, but also complexes appear to have been used pri- the immovable features of modified cul- marily for storage, if not also for other tural landscapes. As suggested above, cer- purposes, as suggested again by recent amics indicate that the sites were occupied

Figure 13. Map of central western Anatolia, showing contemporary sites mentioned in the text.

between the eighteenth and the thirteenth exceptions in central western Anatolia centuries BC, contemporary with levels VI– (Figure 13). VIIa at Troy and levels V/IVc–II at One such citadel is Gavurtepe Höyük, Beycesultan. The clays used in the primary located to the east in the foothills of the ceramic wares found across each site, Bozdağ range near Alasehiŗ (Classical where tested, are chemically uniform, Philadelphia). Surveyed and excavated in further suggesting their contemporaneity the late 1980s and early 1990s, contem- and centralized production (Luke et al., porary levels at the site are defined by a 2015). Beyond ceramics, the majority of circuit wall enclosing c. 1.1 ha of protected citadels share also locational patterns of space, at the peak of which was a large cup-marks (Luke & Roosevelt, in press), megaron-shaped building (Meriç, 1987, similar ranges of lithics and ground stone 1989b, 1990, 1992, 1993). In size and or- tools, and similar surface configurations, ganization, Gavurtepe Höyük most closely suggesting general commonality of func- resembles Kızbacı Tepesi and Koca Dere tion. Similar topographic situations in the Marmara Lake basin, with their result also in the intervisibility of four of wall circuits enclosing elongated forms the citadels, while only Kaymakçı has partly dictated by natural topography. views of all. A better parallel can be found to the With respect to the probably unforti- west at Bademgediğı Tepe, located in the fied, lowland settlements of the second lower Küçük Menderes Valley, near millennium BC, numerous comparable sites Torbalı. As mapped and excavated in the in central western Anatolia have been early 2000s, the site is defined by a circuit known since the 1950s (e.g. French, wall atop a bedrock knoll that encloses 1969; Foss, 1982; Meriç, 1989a: 361), and c. 2.8 ha (Meriç & Mountjoy, 2002; some have been excavated in recent Meriç, 2007). Flanking the wall around its decades, for example Baklatepe (Özkan & interior is a broad terrace, and at the core Erkanal, 1999) and Çesmȩ Bağlararası of the site is another monumental circuit (Erkanal & Karaturgut, 2004;Şahoğlu, enclosing an oval area at the site’s apex. 2007; Aykurt, 2010b). Rarer are systemat- These features best parallel the broad cir- ically surveyed, mapped, and/or excavated cumferential terraces and circuits around comparanda for the citadels in the bedrock knolls at Koca Dere, Asar Tepe Marmara Lake basin, with a few notable 2, and Gedevre Tepesi, and the size of the

citadel falls comfortably within the range 2005). Further inland and south of the of those in the Marmara Lake basin. Menderes Valley, only Çine-Tepecik Bademgediği Tepe too, with its oval inner (Günel, 2008, 2010) and Aphrodisias circuit, is reminiscent of Gedevre Tepesi (Joukowsky, 1986; Mac Sweeney, 2010) and Kaymakçı. Yet, at these sites, the con- offer contemporary evidence, and while evi- centric quarter-circle shapes of their inner dence for Çine-Tepecik is growing thanks circuits are so particular that they appear to ongoing excavations, still little can be to represent a previously unidentified local said about its overall site organization. building tradition. For additional compar- If one looks outside the central part of anda, especially for Kaymakçı, one must western Anatolia, the citadel of Troy VI/ look further afield. VIIa—the best known site in the region Only little is known about the config- and the presumed capital of Wilusa— urations of other fortified sites of the compares well with the central area of second millennium BC in central western Kaymakçı, although Troy’s circuit walls Anatolia. Beneath the remains of later enclose a space less than one-quarter the periods, limited exposures of a second- area, even if its extramural settlement pro- millennium BC fortification wall on the vides a closer parallel (Jablonka, 2011; acropolis of Pergamon show that it was Pavúk & Schubert, 2014; Rose, 2014). As defined by components similar to those at made clear above, the size and internal Kaymakçı and was likely to have served as complexity of Kaymakçı stand out not just the primary site of the middle Bakır from Troy and other western Anatolian River valley (Radt, 1992; Hertel, 2011). citadels, but from its local peers too, sug- Panaztepe too, in the lower Gediz Valley, gesting its primacy as a local and broader appears to have been defined by a fortified regional capital. Given the reading of his- citadel, extramural settlement, and cemetery torical geography in Late Bronze Age areas; yet post-Bronze Age activities at the western Anatolia described above, then, site obscure overall site organization (Günel, Kaymakçı is a strong candidate for the 1999; Erkanal-Öktü, 2008;Aykurt,2010a; capital of the Seha River Land. 2013). The same is true of nearby Larisa We know strikingly little about the late and Bayraklı (Bayne, 2000;Hertel,2007), fourteenth- and thirteenth-century history ̇ and even Liman Tepe, on the Bay of Izmir of the Seha River Land, and close to (Erkanal, 2008;Mangaloğlu-Votruba, 2011; nothing for the period before that time. 2015), where excavations of limited second- Fragmentary Hittite archives nevertheless millennium remains allow for little more make clear that its kings were notoriously than ceramic comparisons. Further south, flexible in their allegiances, pitting Hittite KocabaşTepe in the Menderes Valley of against Ahhiyawan odds for self-interested ̇ Izmir (Aykurt, 2010a, 2013), Ayasoluk in purposes at various times. It was presum- Selçuk (Büyükkolancı, 2007, 2008), and ably to situate the Seha River Land more Kadı Kalesi south of Kusadaş ı (Akdeniz, firmly on the side of Hittite causes that 2006) were each probably fortified in the some of its kings appear to have been second millennium, but none has been given broader significance, even inter- investigated thoroughly to date, rendering marrying with Hittite royalty in the mid- comparison nearly impossible. Even Miletus thirteenth century (Singer, 1983: 216; is estimated to have encompassed only 5 ha Hoffner, 2009: 314). intheLateBronzeAge;theretoomoreac- The state of ceramic chronologies (see curate insights are impeded by lack of ex- Aykurt, 2010a, 2013; Pavúk, 2015) and posure (Greaves, 2002, 2007;Niemeier, limited nature of excavations make it

impossible to determine the relative development that help explain the regional phasing and development of the citadels in data (e.g. Wright, 2004, 2008). Thus, sites the Marmara Lake basin, whether the around the Argive Plain that may originally smaller sites generally pre- or post-date have been of similar size and lesser signifi- Kaymakçı, or whether they all developed cance became subordinate to Mycenae in at the same time, with Kaymakçı eventual- later phases. Similar models of development ly surpassing its neighbours through the may explain the prominence of Kaymakçı prominence and success of its leaders, the among its peers in central western strategic importance of its location, or Anatolia. other factors. Given the settlement patterns and evidence for the centralization of ceramic production (Luke et al., 2015), CONCLUSIONS however, it seems likely that one centralized authority did indeed come to the The recently discovered suite of sites fore. Nevertheless, that multiple citadels described above is the only physical evi- (and unfortified settlements for that dence of a central western Anatolian matter) may have co-existed in close prox- kingdom that had been forgotten or was, imity is hardly difficult to accept. In the at least, previously unknown outside his- vicinity of capital cities in Mira, the torical archives. While the preponderance Hittite king Mursili II (re-)fortified of evidence currently suggests that this several such strongholds following his suc- kingdom held sway over the realm known cessful campaigns (Hawkins, 1998:14– to the Hittites as the Seha River Land in 15). With respect to the suite of citadels the Late Bronze Age, this identification in the Marmara Lake basin as a whole, remains a prevailing theory, subject to however, parallels in western Anatolia confirmation or refutation by new evi- remain unknown to us. dence. At present, the composition and lo- In northern Anatolia, an example can cation of the network of sites, along the be found in the Classical region of Gediz Valley’s natural corridor connecting Paphlagonia (Matthews, 2004; Glatz & Aegean and inland Anatolian worlds, Matthews, 2005), where survey located a matches well the roles known to have network of Late Bronze Age strongholds been played by Seha River Land kings in located along the river Devrez, or Hittite negotiating power relationships between Dahara. However, this network formed Arzawan, Ahhiyawan, and Wilusan inter- not the core of a Bronze Age territory, but ests. Nonetheless—be it the core of the a contested frontier zone between Hittite Seha River Land or not—the Marmara and Kaska territories, with citadels repre- Lake basin, as elucidated by its archaeo- senting the militarized build-up of frontier logical remains of the second millennium lands in a region defined by long-term BC, clearly hosted the local communities of conflict. a forgotten kingdom. To the west, in the core of the The archaeological discovery of these Mycenaean Greek world, a similar density communities resulted directly from the of contemporary citadels can be found in diachronic, systematic methodology of and around the Argive Plain, including CLAS, which gave us an understanding Mycenae, Tiryns, Midea, and Argos, of indigenous settlement structures and among other sites. Owing to the finer regional organization that is more holistic chronological control of such sites, one than achieved elsewhere without such can propose models of settlement-pattern methods. Thus, we are able not only to

highlight the size and complexity of ACKNOWLEDGMENTS Kaymakçı—undoubtedly the primary site in the area—but also to understand its The research presented here was supported situation within a network of at least five by the National Science Foundation (Awards smaller citadels and at least five other un- BCS-0649981 and BCS-1261363), National fortified settlements, enabling future in- Endowment for the Humanities (Award vestigation of the internal dynamics— RZ5155613), American Research Institute from citadel to village to farmstead—of in Turkey, Institute for Aegean Prehistory, western Anatolian kingdoms in the Vecchiotti Archaeology and Midas-Croesus second millennium BC and their local ne- Funds, and the Karis, Loeb Classical Library, gotiation of political and economic Marion and Jasper Whiting, and Merops fortunes. Foundations. For permissions and support, This work is beginning to shed light on we thank the Ministry of Culture and broader dynamics within second-millen- Tourism, Republic of Turkey, and the nium BC western Anatolia too, with the Manisa Museum of Ethnography and identification of ceramic and other mater- Archaeology, and its former director ial connections to other sites in the region. S. Soyaker. We are also grateful to the 2005– The citadels of the Marmara Lake basin 2014 field teams of the Central Lydia may have been influenced by Hittite con- Archaeological Survey, Peter Cobb for illus- quest, but they probably also represent tration assistance, the communities of local traditions of fortification. The mor- Tekelioğlu and Hacıveliler, and the anonym- phological similarities between local cita- ous reviewers of a previous draft of this dels and those in nearby areas appear to article. provide the basic vocabulary for common fortification principles and thus may provide the first evidence for second-millennium (Arzawan?) traditions of fortifica- REFERENCES tion spread throughout western Anatolia. Alparslan, M. & Doğan-Alparslan, M. 2015. Further evidence for such a tradition will The Hittites and their Geography: Problems have to be sought in ongoing and future of Hittite Historical Geography. European research. Journal of Archaeology,18(1):90–110. In such research, we affirm the import- Akdeniz, E. 2006. II. Mursili’nin Arzawa Seferi ve Tarihi Coğrafya Açısından Bazı ance of diachronic and systematic survey – to produce datasets that do not simply Öneriler. Arkeoloji Dergisi, 7(1): 17 40. Aykurt, A. 2008. Batı Anadolu’da Gri engage with but also extend historical evi- Seramiğin Kökeni ve Gelisimi.̧ In: A. dence, long-privileged in the art of recon- Erkanal Öktü, S. Günel & U. Deniz, eds. structing historical geography; this will Batı Anadolu Doğu Akdeniz Geç Tunç Çağı Kültürleri Üzerine Yeni Arasţırmalar. allow us to gain broader perspectives on ı ı the distribution of historical communities Ankara: Hacettepe Üniversitesi Yay nlar , pp. 114–22. ̇ across peopled landscapes, from citadels to Aykurt, A. 2010a. Izmir Bölgesi Orta Tunç ceramic scatters. We expect that only with Çağı Seramiği. Belleten, 74/269: 1–69. such combined analyses of data from Aykurt, A. 2010b. Late Bronze Age Pottery survey archaeology and historical geog- from Çesmȩ Bağlararası. OLBA, 18: 1–63. raphy will we continue to expand, if not Aykurt, A. 2013. An Updated Assessment on Western Anatolian Middle Bronze Age also modify, our understanding of historic- Chronology in Light of Excavations of the al communities of second-millennium BC Izmir Region. Colloquium Anatolicum, 12: western Anatolia. 37–77.

Bayne, N. 2000. The Grey Wares of North-West Dedeoğlu, F. & Abay, E. 2014. Beycesultan Anatolia in the Middle and Late Bronze Age Höyük Excavation Project: New and the Early Iron Age and their Relation to Archaeological Evidence from Late the Early Greek Settlements. Bonn: Habelt. Bronze Age Layers. Arkeoloji Dergisi, 19: Beckman, G. 1999. Hittite Diplomatic Texts. 1–39. 2nd ed. Society of Biblical Literature Easton, D.F., Hawkins, J.D., Sherratt, A.G. Writings from the Ancient World Series, & Sherratt, E.S. 2002. Troy in Recent 7. Atlanta: Scholars Press. Perspective. Anatolian Studies, 52: 75–109. Beckman, G., Bryce, T.B. & Cline, E. 2011. Erkanal, H. 2008. Geç Tunç Çağı’nda Liman The Ahhiyawa Texts. Atlanta: Society of Tepe. In: A. Erkanal-Öktü, S. Günel & Biblical Literature. U. Deniz, eds. Batı Anadolu ve Doğu Blegen, C.W., Caskey, J.L. & Rawson, M. Akdeniz Geç Tunç Çağı Kültürleri Üzerine 1953. Troy: The Sixth Settlement. Yeni Arasţırmalar. Ankara: Hacettepe Princeton: Princeton University Press. Üniversitesi Yayınları, pp. 91–100. Bryce, T.R. 2003. The Relations between Erkanal, H. & Karaturgut, E. 2004. 2002 Yılı Hatti and Ahhiyawa in the Last Decades Çesmȩ Bağlararası Kazıları. Kazı Sonuçları of the Bronze Age. In: G. Beckman, R. Toplantısı, 25(2): 153–64. Beal & G. McMahon, eds. Hittite Studies Erkanal, H. & Şahoğlu, V. 2012. Liman Tepe in Honor of Harry A. Hoffner, Jr., on the (1992–). In: O. Bingöl, A. Öztan & H. Occasion of his 65th Birthday. Indiana: Tasķıran, eds. DTCF Arkeoloji Bölümü Eisenbrauns, pp. 59–72. Tarihçesi ve Kazıları (1936–2011). Bryce, T.R. 2005. The Kingdom of the Hittites. Anadolu Dergisi Eki. Ankara: Ankara New edition. Oxford: Oxford University Üniversitesi Basımevi, pp. 219–30. Press. Erkanal-Öktü, A. 2008. Panaztepe Geç Tunç Bryce, T.R. 2011. The Late Bronze Age in Çağı Mezarlıkları. In: A. Erkanal-Öktü, S. the West and the Aegean. In: S.R. Günel & U. Deniz, eds. Batı Anadolu ve Steadman & G. McMahon, eds. The Doğu Akdeniz Geç Tunç Çağı Kültürleri Oxford Handbook of Ancient Anatolia Üzerine Yeni Arasţırmalar. Ankara: (10,000–323 BC). Oxford: Oxford Hacettepe Üniversitesi Yayınları,pp.70–90. University Press, pp. 363–75. Foss, C. 1982. A Neighbor of Sardis: The Büyükkolancı, M. 2007. Apasa, das alte City of Tmolus and its Successors. Ephesos und Ayasoluk. In: J. Cobet, V. Classical Antiquity, 1(2): 178–205. von Graeve, W.-D. Niemeier & K. French, D. 1969. Prehistoric Sites in Zimmermann, eds. Frühes Ionien: eine Northwest Anatolia II. The Balıkesir and Bestandsaufnahme. Panionion-Symposion Akhisar/Manisa Areas. Anatolian Studies, Güzelcamli, 26. September–1. Oktober 1999. 19: 41–98. Deutsches Archäologisches Institut. Glatz, C. 2009. Empire as Network: Spheres Milesische Forschungen, 5. Mainz am of Material Interaction in Late Bronze Rhein: Von Zabern, pp. 21–26. Age Anatolia. Journal of Anthropological Büyükkolancı, M. 2008. Selçuk Ayasuluk Archaeology, 28(2): 127–41. Tepesi (Eski Efes) ‘Appasas’ mı?In:A. Glatz, C. & Matthews, R. 2005. Anthropology Erkanal-Öktü, S. Günel & U. Deniz, eds. of a Frontier Zone: Hittite-Kaska Relations Batı Anadolu ve DoğuAkdenizGeçTunçÇağı in Late Bronze Age North-Central Kültürleri Üzerine Yeni Arasţırmalar.Ankara: Anatolia. Bulletin of the American Schools of Hacettepe Üniversitesi Yayınları, pp. 37–55. Oriental Research,399:47–65. Çilingiroğlu, Ç. 2015. Bozyer (Salihli, Manisa) Greaves, A.M. 2002. Miletos: A History. Prehistorik Çanak Çömleği Üzerine Bir New York: Routledge. Değerlendirme. Arasţırma Sonuçları Greaves, A.M. 2007. Trans-Anatolia: Examining Toplantısı, 32(2): 55–70. Turkey as a Bridge between East and West. Çilingiroğlu, A., Derin, Z., Abay, E., Anatolian Studies,57:1–15. ̇ Sağlamtimur, H. & Kayan, I. 2004. Günel, S. 1999. Panaztepe II: M.Ö. 2. bine Ulucak Höyük: Excavations Conducted tarihlendirilen Panaztepe seramiğinin Batı Between 1995 and 2002. Ancient Near Anadolu ve Ege arkeolojisindeki yeri ve Eastern Studies, Supplement 15. Paris: önemi/Die Keramik von Panaztepe und ihre Peeters. Bedeutung für Westkleinasien und die Ägäis

im 2. Jahrtausend. Ankara: Türk Tarih Lloyd, S. & Mellaart, J. 1965. Beycesultan Kurumu Basımevi. Volume II: Middle Bronze Age Architecture Günel, S. 2008. Çine-Tepecik kazıları ve bölge and Pottery. London: British Institute of arkeolojisine katkıları. In: A. Erkanal- Archaeology at Ankara. Öktü, S. Günel & U. Deniz, eds. Batı Luke, C. & Cobb, E. 2013. Dwelling in Anadolu ve Doğu Akdeniz Geç Tunç Çağı Hacıveliler: Social Engineering Policies in Kültürleri Üzerine Yeni Arasţırmalar. the Context of Space, Place, and Ankara: Hacettepe Üniversitesi Yayınları, Landscape in Rural, Western Turkey. pp. 114–22. Anatolian Studies, 63: 155–73. Günel, S. 2010. Mycenaean Cultural Impact Luke, C. & Roosevelt, C.H. 2009. The on the Çine (Marsyas) Plain, Southwest Central Lydia Archaeological Survey: Anatolia: The Evidence from Çine- Documenting the Prehistoric through Tepecik. Anatolian Studies, 60: 25–49. Iron Age Periods. In: S.W. Manning & Gunter, A.C. 1991. The Gordion Excavations M.J. Bruce, eds. Tree-Rings, Kings, and Final Reports Volume III: The Bronze Age. Old World Archaeology and Environment: Philadelphia: The University Museum. Papers Presented in Honor of Peter Ian Hanfmann, G.M.A. 1968. The Tenth Kuniholm. Oxford: Oxbow Books, pp. Campaign at Sardis (1967). Bulletin of the 199–218. American Schools of Oriental Research, 191: Luke, C. & Roosevelt, C.H. 2016. Memory 2–41. and Meaning in Bin Tepe, the Lydian Hawkins, J.D. 1998. Tarkasnawa, King of Cemetery of a ‘Thousand Mounds’. In: O. Mira: ‘Tarkondemos’,Boğazköy Sealings Henry & U. Kelp, eds. Tumulus as Sema: and Karabel. Anatolian Studies, 48: 1–31. Proceedings of an International Conference Heinhold-Krahmer, S. 1977. Untersuchungen on Space, Politics, Culture, and Religion in zu seiner Geschichte nach den hethitischen the First Millennium BC. Berlin: De Quellen. Texte der Hethiter, 8. Heidelberg: Gruyter, pp. 407–28. Carl Winter Universitätsverlag. Luke, C. & Roosevelt, C.H. In press. Cup- Hertel, D. 2007. Der äolische Siedlungsraum Marks and Citadels: Evidence for Libation (Aiolis) am Übergang von der Bronze- zur in Second-Millennium B.C.E. Western Eisenzeit. In: J. Cobet, V. von Graeve, Anatolia. Bulletin of the American Schools of W.-D. Niemeier & K. Zimmerman, eds. Oriental Research. Frühes Ionien: eine Bestandsaufnahme. Luke, C., Roosevelt, C.H., Cobb, P. & Panionion-Symposion Güzalçamlı 26. Çilingiroğlu, Ç. 2015. Composing September–1. Oktober 1999. Milesische Communities: Chalcolithic through Iron Forschungen, 5. Mainz am Rhein: Verlag Age Survey Ceramics in the Marmara Philipp von Zabern, pp. 97–122. Lake Basin, Western Turkey. Journal of Hertel, D. 2011. Das vorklassische Pergamon Field Archaeology, 40(4): 428–49. und sein Siedlungsprofil. Istanbuler Mac Sweeney, N. 2010. Hittites and Mitteilungen, 61: 21–84. Arzawans: A View from Western Hoffner, H.A. Jr. 2009. Letters from the Anatolia. Anatolian Studies, 60: 7–24. Hittite Kingdom. Writings from the Ancient Mangaloğlu-Votruba, S. 2011. Liman Tepe’de World, 15. Atlanta: Society of Biblical Geç Hellas IIIC dönemi. Anadolu/ Literature. Anatolia, 37: 43–73. Jablonka, P. 2011. Troy in Regional and Mangaloğlu-Votruba, S. 2015. Liman Tepe International Context. In: S.R. Steadman & during the Late Bronze Age. In: N.C. G. McMahon, eds. The Oxford Handbook of Stampolidis, Ç. Maner & K. Kopanias, Ancient Anatolia (10,000–323 BC).Oxford: eds. NOSTOI. Indigenous Culture, Oxford University Press, pp. 717–33. Migration and Integration in the Aegean Joukowsky, M.S. 1986. Prehistoric Aphrodisias. Islands and Western Anatolia during the An Account of the Excavations and Artifact Late Bronze and Early Iron Age. Studies. 2 vols. Providence: Brown Proceedings of an International Symposium University. held at the Koç Research Center for Korfmann, M.O. ed. 2006. Troia – Archäologie Anatolian Civilizations, Istanbul, March ̇ eines Siedlungshügels und seiner Landschaft. 31–April 3 2011.Istanbul: Koç University Mainz am Rhein: Philipp von Zabern. Press, pp. 647–68.

Matthews, R. 2004. Landscapes of Terror and Neve, P. 1996. Schalensteine und Schalenfelsen Control: Imperial Impacts in Paphlagonia. in Boğazköy-Hattuša(2.Teil). Istanbuler Near Eastern Archaeology, 67(4): 200–11. Mitteilungen,46:41–56. Matthews, R. & Glatz, C. 2009. The Historical Niemeier, W.-D. 2005. Minoans, Mycenaeans, Geography of North-Central Anatolia in Hittites, and Ionians in Western Asia the Hittite Period: Texts and Archaeology Minor. New Excavations at Bronze Age in Concert. Anatolian Studies,59:51–72. Miletus-Millawanda. In: A. Villing, ed. Mellaart, J. & Murray, A. 1995. Beycesultan The Greeks in the East. London: British Volume III, Part II: Late Bronze Age and Museum, pp. 1–36. Phrygian Pottery and Middle and Late Özkan, T. & Erkanal, H. 1999. Tahtalı Barajı ̇ ̇ Bronze Age Small Objects. London: British Kurtarma Kazısı Projesi.Izmir: Izmir Institute of Archaeology at Ankara. Arkeoloji Müzesi Müdürlüğü. Meriç, R. 1987. 1985 yılı Alasehiŗ kazısı. Kazı Pavúk, P. 2015. Between the Aegean and the Sonuçları Toplantısı, 8(2): 259–71. Hittites: The Western Anatolia in Second ̇ Meriç, R. 1989a. 1988 yılı Izmir, Manisa illeri Millennium BC. In: N. Stampolidis, Ç. Arkeolojik yüzey arasţırması. Arasţırma Maner & K. Kopanias, eds. NOSTOI. Sonuçları Toplantısı, 7: 361–66. Indigenous Culture, Migration and Integration Meriç, R. 1989b. 1987 yılı Alasehiŗ kazısı. in the Aegean Islands and Western Anatolia Kazı Sonuçları Toplantısı, 10(1): 157–70. during the Late Bronze and Early Iron Age. ̇ Meriç, R. 1990. 1988 yılı Alasehiŗ kazısı. Kazı Istanbul: Koç University Press, pp. 81–114. Sonuçları Toplantısı, 11(1): 179–90. Pavúk, P. & Schubert, C. 2014. Die Troas in Meriç, R. 1992. 1990 yılı Alasehiŗ kazısı. Kazı der Mittel- und Spätbronzezeit. In: E. Sonuçları Toplantısı, 13(1): 227–35. Pernicka, C.B. Rose & P. Jablonka, eds. Meriç, R. 1993. 1991 yılı Alasehiŗ kazısı. Kazı Troia 1987–2012: Grabungen und Sonuçları Toplantısı, 14(2): 355–63. Forschungen I. Forschungsgeschichte, Methoden Meriç, R. 2007. Ein Vorbericht über eine und Landschaft, Teil 2. Studia Troica spätbronzezeitliche befestigte Höhensiedlung Monograph, 5. Bonn: Rudolf Habelt, pp. bei Metropolis in Ionien: Die Arzawa-Stadt 864–907. Puranda?In:J.Cobet,V.vonGraeve,W.-D. Radt, W. 1992. Die frühesten Wehrmauern Niemeier & K. Zimmermann, eds. Frühes von Pergamon und die Zugehörigen Ionien: eine Bestandsaufnahme. Panionion- Keramikfunde. Istanbuler Mitteilungen, 42: Symposion Güzalçamlı 26. September–1. 163‒234. Oktober 1999. Milesische Forschungen, Roosevelt, C.H. 2006. Tumulus Survey and 5. Mainz am Rhein: Verlag Philipp von Museum Research in Lydia, Western Zabern, pp. 27–36. Turkey: Determining Lydian- and Meriç, R. & Mountjoy, P. 2002. Mycenaean Persian-Period Settlement Patterns. Pottery from Bademgediği Tepe (Puranda) Journal of Field Archaeology, 31(1): 61–76. Ionia: A Preliminary Report. Istanbuler Roosevelt, C.H. 2007. Central Lydia Mitteilungen, 52: 79–98. Archaeological Survey: 2005 Results. Mitten, D.G. & Yüğrüm, G. 1971. The Arasţırma Sonuçları Toplantısı, 24(2): 135–54. Gygean Lake, 1969: Eski Balikhane, Roosevelt, C.H. 2009. The Archaeology of Preliminary Report. Harvard Studies in Lydia, from Gyges to Alexander. Classical Philology, 75: 191–95. Cambridge: Cambridge University Press. Mitten, D.G. & Yüğrüm, G. 1974. Ahlatlı Roosevelt, C.H. 2010. Lydia before the Tepecik beside the Gygean Lake. Lydians. In: N.D. Cahill, ed. The Lydians Archaeology, 27: 22–29. and Their World. Catalogue of an Exhibit at Moran, W.K. 1992. The Amarna Letters. the Yapı Kredi Vedat Nedim Tör Museum, ̇ ̇ Baltimore: Johns Hopkins University Press. Istanbul.Istanbul: Yapı Kredi Culture, Mountjoy, P.A. 1998. The East Aegean-West Art, and Publishing, pp. 37–73. Anatolian Interface in the Late Bronze Roosevelt, C.H. 2014. Mapping Site-Level Age: Mycenaeans and the Kingdom of Microtopography with Real-Time Ahhiyawa. Anatolian Studies, 48: 33–67. Kinematic (RTK) Global Navigation Neve, P. 1977/78. Schalensteine und Satellite Systems (GNSS) and Unmanned Schalenfelsen in Boğazköy-Hattuša. Istanbuler Aerial Vehicle Photogrammetry (UAVP). Mitteilungen, 27/28: 61–72. Open Archaeology,1:29–53.

Roosevelt, C.H. & Luke, C. 2008. Central Sarı, D. 2013: The Cultural Development of Lydia Archaeological Survey: 2006 Western Anatolia in the Third and Results. Arasţırma Sonuçları Toplantısı,25 Second Millennia BC and its Relationship (3): 305–26. with Migration Theories. In: A. Mouton, Roosevelt, C.H. & Luke, C. 2009. Central I. Rutherford & I. Yakubovich, eds. Lydia Archaeological Survey: 2007 Luwian Identities: Culture, Language, and Results. Arasţırma Sonuçları Toplantısı,26 Religion between Anatolia and the Aegean. (2): 433–50. Culture and History of the Ancient Roosevelt, C.H. & Luke, C. 2010. Central Near East 64. Leiden-Boston: Brill, pp. Lydia Archaeological Survey: 2008 Results. 305–27. Arasţırma Sonuçları Toplantısı,27(2):1–24. Sayce, A.H. 1880. The Hittite Monuments. Roosevelt, C.H. & Luke, C. 2011. Central Proceedings of the Society of Biblical Lydia Archaeological Survey: 2009 Results. Archaeology, 10(9): 76–78. Arasţırma Sonuçları Toplantısı,28(3):55–74. Sayce, A.H. 1888. The Hittites: The Story of a Roosevelt, C.H. & Luke, C. 2012. Central Forgotten Empire. London: The Religious Lydia Archaeological Survey: 2010 Tract Society. Results. Arasţırma Sonuçları Toplantısı,29 Singer, I. 1983. Western Anatolia in the (1): 383–400. Thirteenth Century BC According to the Roosevelt, C.H. & Luke, C. 2013. The Hittite Sources. Anatolian Studies, 33: Central Lydia Archaeological Survey: 2011 205–17. Work at Kaymakçı and in the Marmara Starke, F. 1997. Troia im Kontext des histor- Lake Basin. Arasţırma Sonuçları Toplantısı, isch-politischen und sprachlichen 30(1): 237–54. Umfeldes Kleinasiens im 2. Jahrtausend. Roosevelt, C.H., Luke, C., Cobb, P., Studia Troica, 7: 446–87. O’Grady, C. & Sekedat, B. 2014. The Wright, J.C. 2004. Comparative Settlement Central Lydia Archaeological Survey: 2012 Patterns during the Bronze Age in the Work at Kaymakçı and in the Marmara Northeastern Peloponnesos, Greece. In: S. Lake Basin. Arasţırma Sonuçları Toplantısı, Alcock & J. Cherry, eds. Side-by-Dive 31(1): 333–55. Survey. Comparative Regional Studies in the Roosevelt, C.H., Luke, C., Sekedat, B. & Mediterranean World. Oxford: Oxbow Cobb, P. 2015a. The Central Lydia Books, pp. 114–31. Archaeological Survey: 2013 Work at Wright, J.C. 2008. Early Mycenaean Greece. Kaymakçı and in the Marmara Lake Basin. In: C.W. Shelmerdine, ed. The Cambridge Arasţırma Sonuçları Toplantısı, 32(2): 239–58. Companion to the Aegean Bronze Age. Roosevelt, C.H., Cobb, P., Moss, E., Cambridge and New York: Cambridge Olson, B.R. & Ünlüsoy, S. 2015b. University Press, pp. 230–57. Excavation is Destruction Digitization: Wright, W. 1874. The Hamah Inscriptions: Advances in Archaeological Practice. Hittite Remains. British and Foreign Journal of Field Archaeology, 40(3): 325–46. Evangelical Review, 23(87): 90–99. Roosevelt, C.H., Luke, C. & Sekedat, B. Wright, W. 1884. The Empire of the Hittites. 2016. The Central Lydia Archaeological London: Nisbet. Survey: 2014 Work at Kaymakçı and in Yakar, J. 2011. Anatolian Chronology and the Marmara Lake Basin. Arasţırma Terminology. In: S.R. Steadman & G. Sonuçları Toplantısı, 33(2): 251–62. McMahon, eds. The Oxford Handbook Rose, C.B. 2014. The Archaeology of Greek and of Ancient Anatolia (10,000–323 BC). Roman Troy. Cambridge: Cambridge Oxford: Oxford University Press, pp. 56– University Press. 93. Şahoğlu, V. 2007. Çesme-Ba̧ ğlararası: A New Yakubovich, I.S. 2008. Sociolinguistics of the Excavation in Western Anatolia. In: F. Luvian Language (unpublished PhD dis- Felten, W. Gauss & R. Smetana, eds. sertation, University of Chicago). Middle Helladic Pottery and Synchronisms, Zimmermann, T., Banerjee, A. & Huth, J. Proceedings of the International Workshop 2003. Frühe Steinwerkzeuge aus Anatolien: held at Salzburg October 31–November 2, Archäologische und Mineralogische Unter- 2004. Vienna: Austrian Academy of suchungen. Archäologisches Korrespondenz- Sciences Press, pp. 309–22. blatt,33:57–74.

̇ BIOGRAPHICAL NOTES 34433, Istanbul, Turkey. [email: [email protected]] Christopher H. Roosevelt (PhD 2003, Cornell University) is an Associate Professor at Boston University and Koç Christina Luke (PhD 2002, Cornell University, Director of the Research University) is an Associate Professor at Center for Anatolian Civilizations and the Koç University, Co-Director of Gygaia Kaymakçı Archaeological Project, and Co- Projects, and Editor of the Journal of Field Director of Gygaia Projects. His research Archaeology. She has published widely on a interests include archaeological and spatial variety of topics, including Anatolian technologies and the archaeology of archaeology, ethnography, and heritage. Bronze and Iron Age Anatolia and the eastern Mediterranean. Address: Koç University, Department of Archaeology and History of Art, Address: Koç University Research Center Rumelifeneri Yolu, Sarıyer, 34450, ̇ ̇ for Anatolian Civilizations, Istiklal Istanbul, Turkey. [email: christinaluke@ Caddesi No: 181, Merkez Han, Beyoğlu, ku.edu.tr]

L’histoire d’un royaume oublié ? Prospections archéologiques et géographie historique dans le centre-ouest de l’Anatolie au second millénaire av. J.-C.

Dans cet article nous présentons des indices archéologiques jusqu’à présent inédits sur l’existence d’un royaume datant du milieu du second millénaire av. J.-C. situé dans le centre-ouest de l’Anatolie. Ces données, relevées par la Central Lydia Archaeological Survey dans le bassin du Lac Marmara dans la vallée du Gediz en Turquie occidentale, sont apparemment en accord avec les reconstructions basées sur les sources écrites relatives à la géographie historique de l’âge du Bronze conservées dans les archives hittites. Le site de Kaymakçı en ressort en particulier comme capitale régionale et les prospections archéologiques systématiques nous permettent d’élucider les dynamiques de l’habitat local, allant au-delà d’une corrélation de type traditionnel entre la géographie historique et les chefs-lieux. De plus, une com- paraison entre divers sites contemporains de l’Anatolie occidentale nous permet d’identifier des traits communs aux types de sites, ce qui pourrait indiquer une tradition dans l’architecture de la région et non pas seulement l’influence de l’hégémonie hittite. Translation by Madeleine Hummler

Mots-clés: prospection archéologique, Anatolie, âge du Bronze, géographie historique, Hittites, Pays de la rivière Seha

Die Geschichte eines vergessenen Königreichs? Archäologische Geländeaufnahmen und historische Geografie im zentral-westlichen Anatolien im zweiten Jahrtausend v. Chr.

In diesem Artikel werden bisher unbekannte archäologische Hinweise auf ein Königreich des mittleren zweiten Jahrtausends v. Chr. im zentral-westlichen Teil Anatoliens vorgelegt. Diese Angaben wurden von der Central Lydia Archaeological Survey im Becken des Marmarasees im Tal des Flusses Gediz im Westen der Türkei aufgenommen. Die materiellen Belege scheinen mit der Auswertung der historischen Geografie der späten Bronzezeit — die sich auf schriftliche Quellen, die in hethitischen Archiven erhal- ten sind, stützt — gut zu korrelieren. Vor allem zeichnet sich die Siedlung von Kaymakçı als regionaler Zentralort aus und die Ergebnisse einer systematischen archäologischen Feldaufnahme bieten neue Einsichten in die lokale Siedlungsstruktur, die über traditionelle Zusammenhänge zwischen historischer

Geografie und Hauptorte hinausgehen. Außerdem zeigt ein Vergleich mit zeitgenössischen Siedlungen im zentral-westlichen Teil Anatoliens, dass es gemeinsame Kennzeichen in der Form der Siedlungen gibt, was auf eine regionale architektonische Tradition weist, und nicht nur auf einen Einfluss der hethitischen Hegemonie. Translation by Madeleine Hummler

Stichworte: archäologische Prospektion, Anatolien, Bronzezeit, historische geografie, Hethiter, Land am Seha-Fluss