Variability in the production and use of ceramics at Troy during the
Late Bronze Age
A macroscopic and microscopic fabric analysis of Troy VI and Troy VIIA pottery
Ayla Lisa Krijnen
Variability in the production and use of ceramics at Troy during the
Late Bronze Age
A macroscopic and microscopic fabric analysis of Troy VI and Troy VIIA pottery
Student, student number: Ayla Lisa Krijnen, 5872316 Research master thesis Supervisor: Dr. J. Hilditch Second reader: Dr. G.J. van Wijngaarden University of Amsterdam, faculty of Humanities, department of Archaeology
Amsterdam, August 1, 2014
Contents
List of abbreviations p.5
List of illustrations p. 6
Acknowledgements p. 9
Chapter one: Introduction p. 10
Chapter two: Troy in a Late Bronze Age context p. 18
2.1 The geographical setting of Troy p. 18
2.2 The geology of the Biga peninsula p. 19
2.3 Architectural and ceramic developments during the Late Bronze Age p. 21
2.3.1 Troy VI Early p. 21
2.3.2 Troy VI Middle p. 25
2.3.3 Troy VI Late p. 27
2.3.4 Troy VII A p. 29
Chapter three: Methodology p. 32
3.1 The data p. 33
3.2 Macroscopic fabric analysis p. 35
3.3 Microscopic fabric analysis p. 37
3.3.1 Inclusions p. 38
3.3.2 Clay matrix p. 40
3.3.3 Voids p. 41
Chapter four: Results p. 42
4.1 Results of the macroscopic fabric analysis p. 43
4.2 Results of the microscopic fabric analysis p. 65
Chapter five: Discussion p. 97
5.1 Variability within the local Trojan fabrics in the LBA p. 97
5.2 Neighbouring imports at Troy in the LBA p. 101
5.3 Distant imports at Troy in the LBA p. 104
Chapter six: Conclusion p. 107
References p. 109
Appendix 1: tables and figures p. 117
Appendix 2: microscopic fabric descriptions p. 126
Appendix 3: fabric pictures p. 146
List of abbreviations
AGW I Anatolian grey ware I AGW II Anatolian grey ware II
BA Bronze Age
EIA Early Iron Age
ESB Eastern Sigillata B
BPW I Burnished plain ware I BPW II Burnished plain ware II
CF Coarse fraction
CP Cooking pot
CPX Clinopyroxene
EBA Early Bronze Age
FF Fine fraction
GM Ground mass
INAA Instrumental neutron activation analysis
LBA Late Bronze Age
LCLO Large closed (vessel)
LH Late Helladic
n.a. Not available
NAA Neutron activation analysis OA Optically active/Optical activity
PBA Post Bronze Age
PPL Plain polarized light
RCW I Red coated ware I RCW II Red coated ware II
RW I Red ware I
RW II Red ware II
RWW I Red washed ware I
RWW II Red washed ware II
SEM Scanning electron microscope
XP Crossed polarized light
XRF X-ray fluorescence
List of illustrations
Tables
Table 1: Pottery of Troy VI (after: Pavúk 2005, 218, fig. 81) Table 2: Chronological grouping of the dataset
Table 3: Frequency label (Whitbread 1995, 379, table A3.1. After Kemp 1985) Table 4: contextual information, macroscopic fabric group F1
Table 5: contextual information, macroscopic fabric group F2 Table 6: contextual information, macroscopic fabric group F3 Table 7: contextual information, macroscopic fabric group F4 Table 8: contextual information, macroscopic fabric group F5 Table 9: contextual information, macroscopic fabric group F6 Table 10: contextual information, macroscopic fabric group F7 Table 11: contextual information, macroscopic fabric group F8 Table 12: contextual information, macroscopic fabric group C1 Table 13: contextual information, macroscopic fabric group C2 Table 14: contextual information, macroscopic fabric group C3 Table 15: contextual information, macroscopic fabric group C4 Table 16: contextual information, macroscopic fabric group C5 Table 17: contextual information, macroscopic fabric group C6 Table 18: contextual information, macroscopic fabric group C7 Table 19: contextual information, macroscopic fabric group C8 Table 20: contextual information, macroscopic fabric group C9 Table 21: contextual information, macroscopic fabric group C10 Table 22: contextual information, macroscopic fabric group C11 Table 23: contextual information, macroscopic fabric group C12 Table 24: contextual information, macroscopic fabric group C13 Table 25: contextual information, macroscopic fabric group C14 Table 26: contextual information, macroscopic fabric group L1 Table 27: contextual information, macroscopic fabric group L2 Table 28: contextual information, macroscopic fabric group L3 Table 29: contextual information, macroscopic fabric group L4 Table 30: contextual information, macroscopic fabric group L5 Table 31: contextual information, microscopic fabric group 1 Table 32: contextual information, microscopic fabric group 2 Table 33: contextual information, microscopic fabric group 3 Table 34: contextual information, microscopic fabric group 4 Table 35: contextual information, microscopic fabric group 5 Table 36: contextual information, microscopic fabric group 6 Table 37: contextual information, microscopic fabric group 7 Table 38: contextual information, microscopic fabric group 8 Table 39: contextual information, microscopic fabric group 9 Table 40: contextual information, microscopic fabric group 10 Table 41: contextual information, microscopic fabric group 11 Table 42: contextual information, microscopic fabric group 12 Table 43: contextual information, microscopic fabric group 13 Table 44: contextual information, microscopic fabric group 14 Table 45: contextual information, microscopic fabric group 15 Table 46: contextual information, microscopic fabric group 16 Table 47: contextual information, microscopic fabric group 17 Table 48: contextual information, microscopic fabric group 18 Table 49: contextual information, microscopic fabric group 19 Table 50: contextual information, microscopic fabric group 20 Table 51: contextual information, microscopic fabric group 21 Table 52: contextual information, microscopic fabric group 22 Table 53: contextual information, microscopic fabric group 23 Table 54: contextual information, microscopic fabric group 24 Table 55: contextual information, microscopic fabric group 25 Table 56: contextual information, microscopic fabric group 26 Table 57: contextual information, microscopic fabric group 27
Table 58: contextual information, microscopic fabric group 28 Table 59: contextual information, microscopic fabric group 29 Table 60: contextual information, microscopic fabric group 30 Table 61: contextual information, microscopic fabric 31 Table 62: contextual information, microscopic fabric 32 Table 63: contextual information, microscopic fabric 33 Table 64: contextual information, microscopic fabric 34 Table 65: contextual information, microscopic fabric 35 Table 66: contextual information, microscopic fabric 36 Table 67: contextual information, microscopic fabric 37 Table 68: contextual information, microscopic fabric 38 Table 69: contextual information, microscopic fabric 39 Table 70: contextual information, microscopic fabric 40 Table 71: contextual information, microscopic fabric 41 Table 72: contextual information, microscopic fabric 42 Table 73: contextual information, microscopic fabric 43 Table 74: contextual information, microscopic fabric 44 Table 75: contextual information, microscopic fabric 45 Table 76: contextual information, microscopic fabric 46 Table 77: contextual information, microscopic fabric 47 Table 78: contextual information, microscopic fabric 48
Table 79: Petrographic fabric groups in which typologically local ceramics occur Table 80: Compositional profile (CF) of the local Trojan groups
Figures (appendix 1)
Fig. 1: (a) Map of the Mediterranean; (b) detail of the Aegean region (Tartaron 2013, 2, fig.1.1) Fig. 2: Relative and absolute chronology of habitation periods at Troy (Korfmann 2006, 5, fig. 5) Fig. 3: New ceramic phases of Troy VI (Pavúk 2005, 332)
Fig. 4: Vessel shapes of AGW ceramic phase 1 (Pavúk 2005, 136, fig. 70a)
Fig. 5: Vessel shapes of AGW ceramic phase 2 (left) and ceramic phase 3 (right) (Pavúk 2005, 137, 138, fig. 70b, 70c)
Fig. 6: Vessel shapes of Island ware (Pavúk 2005, 145, fig. 74a)
Fig. 7: Vessel shapes of coarse wares Troy VI (Pavúk 2005, 142, fig. 73)
Fig. 8: Vessel shapes of Troy VII A, Terrace house, destruction phase (Rigter 2013, 320, fig. 46)
Fig. 9: Geological map of the Biga peninsula (Knacke-Loy 1995, 147, fig. 1After Bingöl et al.1973; Okay et al. 1991)
Fig.10: Simplified geological map of the area of Troy (Göbel et al. 2003, p. 357, fig. 5. After Kayan 2000) Fig. 11: Geological map of the Biga peninsula, (Kibaroğlu & Thumm-Doğrayan 2013, 7, fig. 3)
Fig. 12: Catchment areas of the Karamenderes and Dümrek river (after Knacke-Loy 1994, 36, fig. 15)
Fig. 13: Fig. 13: Indication of the find spots of Cretan (yellow dots) and Milesian (blue squares) pottery attested in the data (Modified after: Easton et al. 2002, 83, fig. 6)
Fig. 14: Plan of architectural remains of Troy VI Early, Middle and Late (Becks 2006, 157, fig. 2) Fig. 15: Plan of architectural remains of Troy VIIA (Becks 2006, 158, fig. 3)
Fig. 16: Comparison chart for the visual estimation of percentage area (Quinn 2013, 82, fig. 4.9) Fig. 17: Comparison chart for estimating the degree of sorting (Quinn 2013, 87, fig. 4.15)
(appendix 3) Fig. 18: 02.07.2x40x Fig. 19: 5.7x25x Fig. 20: 64x25x Fig. 21: 31x40x Fig. 22: 8x40x Fig. 23: 02.04.2x40x Fig. 24: 68x40x Fig. 25: 30x40p Fig. 26: 123x25x Fig. 27: 43x25x Fig. 28: 145x25x Fig. 29: 18x40x 7
Fig. 30: 5.19x40x Fig. 31: 167x40x Fig. 32: 94x40x Fig. 33: 166x25x Fig. 34: 117x25x Fig. 35: 154x25p Fig. 36: 161x25p Fig. 37: 12x25x Fig. 38: 137x40p Fig. 39: 27x40p Fig. 40: 65x40x Fig. 41: 79x40x Fig. 42: 66x25x Fig. 43: 163x25p Fig. 44: 155x25x Fig. 45: 103x40x Fig. 46: 130x25x Fig. 47: 33x40x Fig. 48: 22x40x Fig. 49: 128x40x Fig. 50: 44x25x Fig. 51: 70x25x Fig. 52: 92x40p Fig. 53: 17x40x 8
Acknowledgements
I would like to express my gratitude to Prof. Dr. Ernst Pernicka (Eberhard Karls Universität Tübingen), Peter Pavúk (Charles University) and Marta Guzowska for their permission to study the material presented in this thesis. The typological study of the selected samples has been carried out by Peter Pavúk and Marta Guzowska. Without this work, the present study would not have been accomplished and I am therefore very grateful to both. My warmest thanks go out to my supervisor dr. Jill Hilditch (University of Amsterdam) for her endless help and patience while teaching me petrography and her constructive feedback during the writing process of this thesis. I would also like to thank my second reader, dr. Gert Jan van Wijngaarden (University of Amsterdam) for his advice and many valuable suggestions to the chapters. I would like to thank my mother, Louise Krijnen, for her constant support during this study as well as her financial assistance to realize this thesis. Finally, I want to thank Štĕpán Rückl, who is always there for me, for his help, support and love.
Chapter one
Introduction
The settlement of Troy, located in the northwestern part of Anatolia, is one of the best-known archaeological sites in the world (fig.1). Troy’s high profile can perhaps best be explained by the connections of the site with the Homeric epics. In these written sources, Troy is the setting of the famous battle between the Greeks and the Trojans. According to the literary works by Homer, Troy was a widespread city with large streets and a spacious central square. A large palace with more than fifty rooms was situated at the highest point of the citadel and colossal walls with high towers surrounded the city.1 The Homeric descriptions of Troy differ to a great extent from the archaeological remains at the hill of Hisarlık.2 Besides the imposing fortification walls still visible today, monumental architecture dating to the Bronze Age is scarce.3 Whether or not the archaeological remains of Troy can be identified with the city described in Homer’s poems remains unsettled.4 However, the historical connotation of the site certainly has contributed to the fact that in the last 143 years, large-scale archaeological research has taken place at Troy.
Within the long history of archaeological research at Troy, three major campaigns can be distinguished. The first official excavations at Troy started in 1871 under the direction of Heinrich Schliemann.5 After his death in 1890, Wilhelm Dörpfeld continued the project for two more campaigns in 1893 and 1894.6 The second investigations at Troy were directed by Carl Blegen who carried out excavations from 1932 to 1938.7 The third campaign started in 1988 under the supervision on Manfred Korfmann and Brian Rose. After Korfmann’s death in 2005, the prehistoric part of the project has continued under the direction of Ernst Pernicka.8
1 Blegen 1963, 13.
2The Hisarlık mound is the modern name for the settlement hill of Troy.
3 When compared to contemporary sites in Syria, Mesopotamia and central Anatolia. Korfmann 1998, 377; See
also: Grethlein 2008; Hertel and Kolb 2003; Easton et al. 2002.
4 Several other places in the vicinity of Troy have been suggested as the city described in Homer. Nevertheless,
most scholars today agree that the settlement on the mound of Hisarlık can be associated with Homer’s Troy. See: Jablonka 2010, 847. 5 Schliemann 1881. 6 Dörpfeld 1902. 7 Blegen et al.1951, 1953, 1958. 8 Korfmann 2006 10
The settlement of Troy has a long and largely continuous occupation history which can be divided into ten main periods or ‘cities’ (Troy I to Troy X, fig.2).9 Each of these periods can be subdivided into several architectural and/or ceramic sub-phases.10 Habitation at Troy starts at the Early Bronze Age (Troy I, ca 2920 to 2550 BC) and continues into the Early Iron Age (Troy VIIB, ca. 1180 to 950 BC).11 During this period of almost 2000 years, the settlement was repeatedly destroyed by earthquakes and fire but always rebuilt immediately afterwards. The evidence for occupation of the site in the period that followed is sparse but in contrast to earlier interpretations, recent excavations have shown that Troy was never completely abandoned during the first millennium BC (Troy VIII and IX ca 950 BC to 500 AD).12 However, it was not until the Hellenistic and Roman period that the settlement became of considerable size again.13 During the Byzantine era, the settlement hill of Troy was occupied once more for approximately two centuries (Troy X, ca. 1300 to 1500 AD). 14 Troy’s favorable geographical location (see chapter two, p.) may be one of the reasons for the site’s long habitation history and may explain why Troy, during several phases of its occupational sequence, is considered to be of superregional importance.15
Among the material culture studies undertaken with Trojan artifacts, ceramic research has always had a prominent role.16 In the past, pottery studies largely concentrated on the cultural links of Troy with overseas regions whereby, in particular, the connections between Troy and the mainland of Greece during the Late Bronze Age (LBA) have been a focal point.17 Evidence for the cultural influence of the Mycenaean world at Troy during the LBA is visible by small amounts of Mycenaean imports and vast quantities of locally made Mycenaean-style vessels.18 Moreover, imported ceramics from other distant regions such as Cyprus, the Levant and Crete also point to the fact that Troy was engaged in an
9
The occupational sequence of Troy was established by Schliemann and Dörpfeld. The chronological division has been refined by Blegen and supported by absolute dating methods by Korfmann. See Dörpfeld 1902; Blegen 1953; Korfmann and Kromer 1993; Kromer et al. 2003.
10 Blegen divided the occupational sequence at Troy in various sub-phases. The ceramic periodization of the
Late Bronze Age was recently refined by Pavúk. See: Blegen 1953; Pavúk 2005.
11 Korfmann 2006, 5.
12 Blegen believed that after the destruction of Troy VII, the settlement was abandoned for a few centuries. This
idea has recently been revised although the extent of the Early Iron Age settlement and its exact location is still under debate. See Rose 2013; Chabot Aslan 2002.
13 Rose 2013, 158, 217.
14 Jablonka 2010, 856; Korfmann 2006, 7-8.
15Korfmann, 2006. However, not all scholars agree on the role and importance of Troy within an Anatolian
and/or Aegean context. See Easton et al. 2002; Hertel and Kolb 2003; Jablonka and Rose 2004; Kolb 2004.
16 Mommsen et al. 2001, 169.
17 Mountjoy and Mommsen 2006; Knacke-Loy 1994; Blegen 1953; Felts 1942.
18 In Troy VI Late, one fourth of the ceramics are made in Mycenaean shape. Pavúk and Rigter 2006, 236; See
also Mommsen et al. 2001 and Mountjoy and Mommsen 2006.
11
‘international’ exchange network during this period.19 Only recently, ceramics at Troy deriving from neighbouring regions such as Samothrace, Lesbos and the hinterland of Troy have been given greater attention.20 Furthermore, studies of the Trojan wares have shed light on the varied character of the local pottery.21 As a result, a fairly complex picture emerges when looking at the range of ceramic fabrics produced and used at Troy during the LBA.
As it appears, ceramics coming from both distant and neighbouring regions as well as a large spectrum of local wares are present at the site of Troy. The diversity of ceramic fabrics occurring at Troy during the LBA is the starting point of this research. In order to understand the character of a ceramic assemblage on a compositional level, a key concept is to understand what constitutes local. Therefore, one of the main goals of this thesis is to establish the criteria which, on a compositional and technological level, identify the local Trojan ceramics of the LBA. The second major goal is to understand the range of foreign fabrics occurring at Troy during this period. The range of fabrics, both locally made and imported, will be correlated to three aspects: chronology, vessel shape and ware, and spatial distribution across the site. A petrographic analysis is an appropriate method to elucidate the degree of continuity in the use of ceramic clay sources over time, shedding light on the local potting practices in a diachronic way.22 Correlations between fabrics and vessel types/wares provide insights into the degree of specialization of the potters as well as the choice of the consumers. Variability of fabric distribution across a site illuminates consumption strategies, which are particularly interesting in the case of imported ceramics.
In this study, 230 samples dating to the Troy VI Early to Troy VIIA period will be examined macroscopically and petrographically.23 The majority of the samples derive from deposits excavated by the new excavations at Troy carried out between 1988 and 2005.24 The typological study of the material, as well as the selection of the samples, has not been done by the author of the present study as the material initially formed part of a different project. Therefore, certain questions related to the sampling strategies may remain partly unanswered
19 Kovál 2006, 253. 20
Kibaroğlu and Thumm-Doğrayan 2013; Schubert and Pernicka 2007; Pavúk and Rigter 2006.
21 Pavúk 2005; Rigter 2013. 22 Grave et al. 2013, 1760.
23 I am very grateful to Dr. E. Pernicka for granting me the permission to study the material. I also would like to
thank Peter Pavúk and Marta Guzowska for providing me with the material and for their help throughout the realization of this study.
24 The new excavations at Troy was an interdisciplinary and international project undertaken by the Department
of Classics, University of Cincinnati (USA) and the Institut für Ur- und Frühgeschichte und Archäologie des Mittelalters, Karls Universität Tübingen (Germany).
12
in this thesis.25 However, the primary goals of the former project are largely comparable to the aims set for the current research. Therefore, this issue should not be considered as an obstacle to perform a successful fabric analysis.
Compositional ceramic studies have for long played a significant part in the history of Trojan pottery research. In 1942, Wayne M. Felts conducted a pioneering mineralogical study with pottery dating to the Troy I to Troy VIII period with samples collected during the Blegen excavations.26 His study was one of the first systematic attempts using a combination of both sediment and ceramic samples to assess the general compositional characteristics of archaeological ceramics.27 Felts was able to distinguish between the local and the imported ceramics at Troy and discovered, moreover, that different clay recipes or ‘pastes’ were in use at different times.28
A few decades later, Felts’ study was followed by an optical emission-spectral analysis (OES) carried out by Mervyn Popham and Richard Jones at the university of Oxford.29 In this research, 25 samples from Troy I to VII deposits were examined. The study pointed out that three local Trojan pottery groups could be distinguished and that chronological patterning in the use of ceramic resources could, to a certain degree, be attested.30 In 1989, Susan Heuck Allen conducted a Neutron Activation Analysis (NAA) using Anatolian Grey Ware and Tan Ware sherds found in LBA deposits on Cyprus and in the Levant.31 This study was later embedded in a larger project directed by Elizabeth French aiming to determine the main production centers of the Grey Wares found in Anatolia.32 This research confirmed the existence of three chemical groups of local Trojan pottery in which certain chronological variation in its use could be detected as well.33
The most extensive provenance study of Trojan ceramics was carried out by Anno Knacke-Loy in 1994.34 This study was the first to use a combination of elemental, isotopic and petrographic methods in order to distinguish the local pottery from the Mycenaean
25 More details on (the creation of) the dataset can be found in chapter three, p. 29. 26 Felts 1942, 237. 27 Felts 1942; Quinn 2013, 10. 28 Felts 1942, 237. 29 Jones 1986. 30 Ibid., 303-304. 31 Allen 1990. 32 Pavúk 2005, 219. 33 Ibid., 220.
34 Knacke-Loy 1994. The somewhat expanded results of this study were published in 1995. Knacke-Loy et al.
1995.
13
imports and, moreover, to understand the (changing) character of the local Trojan wares.35 The study was undertaken with 117 sherds from Troy I to IX, 70 sedimentary samples from six local sediments (five from the Karamenderes River and one from the Dümrek River) and eight comparative pottery samples from Mycenae, Tiryns, Demirci Hüyük and Kamid-el-Loz.36 Due to the complexity of the geology of the Troad, the petrographic analysis did not give satisfactory answers regarding the provenance of the samples.37 However, the chemical analysis showed that four distinct compositional groups (A, B, C, D) of local pottery were present in the sampled assemblage. These groups could be related to two sedimentary sources from the nearby river valleys. The majority of the Troy VI samples belong to the chemical group C and group D mainly comprises of Troy VII and IX material.38 Moreover, Knacke-Loy was able to distinguish between the Mycenaean sherds deriving from the Argolid and the locally made Mycenaean wares.39
A follow-up study of Knacke-Loy’s work was conducted by Hans Mommsen, Dieter Hertel and Penelope Mountjoy in 2001.40 In this study, 111 Troy I to Troy VIII samples were analyzed using neutron activation analysis (NAA).41 The research aimed at distinguishing the imported Mycenaean vessels from the locally made pots, to shed light on the chemical composition of the Grey and Tan wares of the second millennium BC and to understand any possible compositional variation within the Grey and painted Archaic wares of the first millennium BC.42 Several compositional groups could be identified, most of which appeared to be local to the Trojan region.43 Moreover, chronological patterning, indicated by the use of different clay sources over time, was attested in the sampled assemblage.44 One of the most striking conclusions of their work was, in contrary to the earlier work carried out by Knacke-Loy, that not a single Mycenaean sherd deriving from the Argolid was discovered.45 In 2006, Mountjoy and Mommsen carried out a follow-up study in order to discover the previously missing Argolid imports of Troy.46 In this research, another NAA study was carried out with 192 sherds, of which 155 samples were Mycenaean sherds found during the new excavations
35 Knacke-Loy 1995, 146. 36 Ibid. 37 Ibid., 171. 38 Ibid., 161-162. 39 Knacke-Loy 1995, 162-163. 40 Mommsen et al. 2001. 41 Ibid., 174. 42 Ibid., 171. 43 Ibid., 173-203. 44 Ibid., 202. 45 Ibid, 176. 46
Mountjoy and Mommsen 2006, 97.
14
at Troy.47 However, the identification of pottery from the Argolid could not be accomplished.48
In 2003, a NAA study was carried out by Marta Guzowska, Ivelin Kuleff, Ernst Pernicka and Muharrem Satır on a series samples dating to the Troy VII B period.49
The aim of this research was to investigate the local compatibility of the so-called handmade ‘Barbarian’ and Knobbed wares occurring at Troy during this period. The geochemical analysis showed that at least a part of the pottery was imported to the site although the provenance of these sherds is as yet unknown.50 The study of Farkas Pintér, published in 2005, also focused on the coarse wares of the Troy VII B period.51 He combined a petrographic, chemical and X-ray fluorescence (XRF) analysis and was able to show that the Knobbed ware from Troy, which previously was thought to be imported to the site, was, in fact, mostly local.52
In 2007, Cornelia Schubert and Ernst Pernicka undertook an instrumental neutron activation analysis (INAA) to examine the fine wares of Bronze Age Troy in more detail.53 203 samples dating to the Middle and Late Bronze Age (Troy VI Early to Troy VIIA) collected during the excavations at Troy and during surface surveys carried out in the surrounding region were used.54 The research resulted in the establishment of a range of local chemical groups and a small number of ‘foreign’ groups.55 A striking observation was the evidence for multiple production centers for the Anatolian Grey ware and Tan ware, probably all in the vicinity of Troy.56 With regards to previous studies, Schubert and Pernicka were able to identify at least four chemical groups compatible to the studies of Knacke-Loy (1994, 1995) and Mommsen (2001, 2006).57 Moreover, one of the subgroups strongly resembled the chemical composition of pottery from Samothrace, although conclusive evidence for the origin of the variant found at Troy needs to be investigated further.58
47 Mountjoy and Mommsen 2006, 97, 98. 48 Ibid., 97. 49 Guzowska et al. 2003. 50 Ibid., 233. 51 Pintér 2005. 52 Ibid., 142.
53 Schubert and Pernicka 2007, 133. 54
Ibid.
55 Schubert and Pernicka 2007. 56 Ibid., 136. 57 Ibid., 135, 136. 58 Ibid., 135, 137. 15
In the same year, Hans Mommsen and Peter Pavúk conducted a NAA study using Grey Wares found at Troy and various other sites on Cyprus and in the Levant.59 In their study, several chemical groups could be distinguished of which the largest group was locally compatible with the Trojan region.60
Billur Tekkök and Ernst Pernicka were the first to carry out a NAA which solely focused upon the post-Bronze Age pottery found at the site of Troy. In their 2012 analysis, 49 Eastern Sigillata B (ESB) samples were examined chemically and compared to the local groups established by previous compositional studies of Knacke-Loy (1994) and Mommsen (2001 and 2006). The study showed that the vast majority of the ESB samples belonged to a non-local source.61
Peter Grave, Lisa Kealhofer, Pavol Hnila, Ben Marsh, Carolyn Aslan, Diane Thumm-Dograyan and Wendy Rigter conducted a NAA and XRF study with 199 LBA and EIA pottery samples and a number of sediment samples in 2013.62 The aim of this research was to evaluate the potential local ceramic resources of Troy in further detail. Two previously known local clay sources could be identified in their dataset and a third local ceramic source, probably deriving from a residual resource in the Troy region, was discovered.63 The third local source mainly comprised of sherds dating to the Early Iron Age.
A compositional study of Trojan pithoi was undertaken by Mustafa Kibaroğlu and Diane Thumm-Doğrayan in 2013. The sample consisted of three EBA, 20 LBA and one chronologically unknown pithos. In the research, four petrographic groups, each corresponding to a partly overlapping region in the Troad could be distinguished. Every group originates ca. 10 to 20 km away from Troy.64
As shown above, a large number of compositional studies aiming to identify the local Trojan ceramics, as well as the provenance of imported pottery at Troy, have been undertaken in the last few decades. A number of important conclusions derive from these studies: firstly, a certain chronological variation in the use of ceramic resources at Troy was attested in almost every study. Secondly, it appears that compositional groups established by various
59 Mommsen and Pavúk 2007, 25. 60
Ibid.
61 Tekkök and Pernicka 2012, 353. 62 Grave et al. 2013, 1764. 63 Ibid., 1760.
64Kibaroğlu and Thumm-Doğrayan 2013, 1.
16
research projects are often difficult to correlate to one another.65 This may be explained by the fact that each project uses its own methodology and unique group nomenclature which is not directly comparable to other studies.66 Thirdly, the subject of interest within ceramic compositional studies at Troy appears to have slightly shifted from a strong focus on distant Greek imports to a more local perspective.
The current research distinguishes itself from previous work on a number of aspects and may therefore form a useful contribution to the discussion on the compositional character of the (LBA) fabrics occurring at Troy. First of all, whereas a large number of chemical analyses of Trojan ceramics have been conducted in the last two decades, the number of petrographic studies is significantly smaller. Secondly, the majority of previous studies have mainly concentrated on chronological variation within ceramic assemblages. In this study, correlations between the fabrics and the shape, ware, and spatial distribution will be taken into consideration as well. And last but not least, while a large number of the previous provenance studies have focused upon one particular pottery class (e.g. Mycenaean pottery), one particular chronological period (e.g. Troy VII B) or one particular vessel type (e.g. pithoi), this research includes a wide range of pottery shapes and wares, dating to a number of (sub-) phases.
In the second chapter of this thesis, the geographical and geological context of Troy will be discussed and an overview of the most significant architectural and ceramic developments of Troy VI and VII A will be given. In chapter three, the methodology for the macroscopic and microscopic analysis will be outlined and in chapter four, the results of the analyses will be presented. The implications of these results will be discussed in chapter five and past scholarship of compositional ceramics studies at Troy will be integrated here. In the last chapter, the most significant results and the potential for further research will be discussed.
65 See for example: Mommsen et al. 2001, 176; Mountjoy and Mommsen 2006, 97. In some cases, chemical
groups established in different studies could, at least partially, be compared. See Knacke-Loy, 1994; Mommsen, 2001; Schubert and Pernicka, 2007.
66 Grave et al. 2013, 1762.
17
Chapter two
Troy in a Late Bronze Age context
The aim of this chapter is to outline the geographical, geological and cultural context of the settlement of Troy in the Late Bronze Age. In the first part of the chapter, the geographical setting of Troy will be examined whereby particular emphasis will be paid to the differences in landscape between Troy in the Late Bronze Age and the present day. These differences can have implications for the character of the site and the availability of the raw materials used for the production of ceramics. The second part of the chapter outlines the geology of the area of Troy: the so called Biga peninsula or Troad.67 In order to carry out a successful provenance analysis, knowledge of the geology of a region is required and, therefore, an overview of the different types of rocks and minerals present on the Biga peninsula will be provided here. In order to contextualize the samples which are used in this research, the main architectural and ceramic developments in Late Bronze Age Troy will be discussed per chronological sub-phases (Troy VI Early, VI Middle, VI Late and Troy VIIA) in the third part of this chapter.
2.1 The geographical setting of Troy
The site of Troy is situated on the Biga peninsula, located in the northwestern part of Anatolia (modern Turkey). The settlement lies on a ridge on the western end of a plateau which rises ca. 30m higher than the surrounding alluvial plains. To the north and northwest, the plateau ends steeply, to the west and south of the settlement, the plateau slopes gently. Troy is located at the junction of two rivers: the Dümrek (ancient Simoeis) river flowing north of the settlement and the Karamenderes (ancient Scamander) river flowing south of Troy (fig.10).68 The citadel of Troy is located on a ca. 15 m high artificial hill (tell or hüyük) created by thousands of years of habitation at the same place. The elevated position of the settlement overlooking the surrounding alluvial plains, the Aegean Sea to the west and the Dardanelles to the north makes the geographical location of Troy very strategic (fig. 1). Land routes connecting Anatolia and southeast Europe as well as sea routes between the Mediterranean and the Black Sea would pass by Troy.69
67 The two names are interchangeable. Troad is normally used in historical sources, Biga peninsula is the
modern name of the region.
68 Jablonka 2006, 168. 69
Hockmann 2003; Korfmann 1986.
18
Geological research in the Trojan area has shown that the sea level was significantly higher during the Bronze Age than it is today.70 6000 to 7000 years ago, the sea level reached a maximum peak which created large bays on many Mediterranean coasts.71 In the period which followed, the sea level stabilized and many of such bays were gradually filled in with sediments.72 This pattern, typical for the coast of Asia Minor, can be seen in for example Miletus, Ephesus and Troy.73 Although the sea in the direct vicinity of Troy was already very shallow during the LBA and in the western parts transforming into a swamp, Troy can be considered to be a coastal city during this period.74 Today, the site is surrounded by plains and the coast is ca. 5 kilometers away (fig.1, 10).
Due to the more than 20m thick alluvial deposits in the plains around Troy, the exact position of the riverbeds of the Dümrek and Karamenderes river in the Bronze Age is very difficult to determine. We do know that the river channels have shifted frequently by normal fluvial processes in the last millennia.75 This means that the sedimentary clay sources in the vicinity of Troy were most likely not located where they are today.Moreover, since the plains surrounding Troy were largely under water during the BA, the clay sources were probably much more restricted in the Bronze Age.76
2.2 The geology of the Biga peninsula
A wide range of metamorphic, sedimentary and igneous rocks are present on the Biga peninsula (fig.9, 11). Extensive studies carried out in the last few decades have shed light on the geological complexity of the area.77 Troy is located between two active fault zones; these zones have developed under the influence of the north Anatolian fault system and the Aegean north-south extension.78 The tectonic unit of which the Troad forms part is called the ‘Sakarya zone’; this zone can be subdivided into four regions: the Kazdag massive, the Ezine zone, the Ayvacik-Karabiga zone and the Bayramic magmatic complex (fig. 9,11). The mineralogical composition of the rocks in these four regions will be discussed here.
70 Kayan 1995, 212; Kacke-Loy 1995, 146. 71 Tartaron 2013, 45. 72 Ibid. 73 Ibid., 141. 74 Kayan et al. 2003, 379. 75 Kraft 2003, 362. 76 Knacke-Loy 1995, 146.
77 See for example: Bingöl 1973; Okay 1991; Okay 1996; Okay and Satír 2000; Sengör and Yilmaz 1981. 78
Yilmaz 2003, 55.
19
The oldest rocks in the Trojan area are part of the Kazdag massive and rise up to 1767m above sea level.79 The rocks are metamorphic Paleozoic continental basement rocks such as granitoids, gneisses and metamorphosed basement sediments which consists mostly of quartz, plagioclase, biotite and hornblende. Diopside, garnet and titanite occur in smaller quantities.80 The Ezine zone comprises of the Karadag unit, the Denizgören Ophiolite and Çamlica micaschists. The Karadag unit consists of slightly metamorphosed Late-Paleozoic to Triassic sedimentary sequences of metashale, fine to medium grained metasandstone, metaquartzite, calci-schist, recrystallized limestone, metamorphosed fine-grained shale, siltstone, calciturbidite, pelagic chert and limestone.81 The Denizgören Ophiolite consists mainly of ultramafic rocks and serpentine. The Camlica micaschist outcrops in a wide area and shows a repetitive lithology which consists mainly of medium grained metasedimentary rocks and micaceous quartz-micaschists with a common mineral assemblage of quartz, muscovite, carbonate, albite, chlorite, clinozosite and garnet.82
The rocks of the Ayvacik-Karabiga comprise of two main units, the Cetmi Ohiolitic melange and the Alakeci Mylonitic zone. The Cetmi Ohiolitic melange is a tectonic melange largely made up of volcanic and pyroclastic rocks including upper Jurrasic and Lower and Upper Cretaceous limestone blocks, shale, greywacke and some serpentinite, radiolarian chert and spilitic basalt.83 The Alakeci mylonitic zone comprises of mylonitized gneiss and metaserpentinite. The fourth zone, the Bayramic magmatic complex, comprises of Late Oligocene granodiortite, quartz monzonite, monzodiorite and quartz diorite as well as Oliogene to Lower Miocene volcanic rocks such as basalts, andesites, trachyandesites, dacites, rhyolites and associations of pyroclastic rocks and sedimentary rocks such as siltstones, sandstones, clay stones and shales.84
The development of clay sediments in close vicinity of Troy started during the Tertiary.85 The oldest Tertiary deposits are the limestones which were formed during the Middle Eocene. They are overlaid by Miocene turbiditic series with andesites and andesitic tuffs.86 These
79
The rocks have isotopically been dated to the Devonian and mid-Carboniferous. Okay et al. 1996; Okay and Satır 2000. 80 Okay et al. 1991, 198. 81 Ibid. 82 Ibid. 83 Ibid, 202. 84 Birkle 1994, 143. 85 Ibid. 86 Ibid. 20
rocks were tectonically uplifted and eroded during the Oligocene.87 During the Late Oligocene continuing into the Middle Miocene, isolated amounts of alkali olivine basalt on the Biga peninsula resulting from the Late Miocene volcanic activity were deposited. These rocks consist mainly of plagioclase, orthoclase, quartz, green hornblende and biotite.88 During the Early Miocene, a large area of the Biga peninsula was covered by andesitic-dacitic and riolitic volcanites and tuffs. Based on geochemical analysis, these volcanics mainly consist of latites and quartzlatites.89 The upper Miocene is characterized by deposits of fluvial and shallow marine sediments. During the Pliocene, fluvial and limnic lacustrine deposits were formed in some areas of the peninsula.90 During the Quaternary, alkali-basaltic volcanism occurred which is characterized by olivine crystals in a fine grained matrix which consists predominantly of feldspar-ridges and opaque minerals.91
The deposits of alluvial plains around Troy consist mostly of alluvial sediments of the two rivers which meet here. The Mesozoic and Paleozoic marbles and serpentinites as well as the Tertiary volcanic deposits of the hinterland have eroded here. Occasionally, colluvial and maritime deposits are included in the alluvial sediments around Troy.92 The low plateau on which Troy is located and the high plateau in the vicinity of Troy consist of shallow maritime limestone’s (fig.10).93
2.3 Architectural and ceramic developments during the Late Bronze Age
2.3.1 Troy VI Early
Troy VI Early marks the transition of the Middle Bronze Age to the Late Bronze Age at Troy. In absolute chronology this phase can be dated from ca. 1750 BC to ca. 1590 BC.94 This phase corresponds to Blegen’s architectural phase VI A, B and C and Pavúk’s ceramic phase 1 and 2.95
87
Pintér 2005, 50.
88 Ibid.
89Birkle and Satır 1994, 143. 90 Pintér 2005, 50. 91 Ibid. 92 Göbel 2003, 342. 93 Ibid. 94 Pavúk 2005. 95 Blegen 1953; Pavúk 2005. 21
In Troy VI Early, the citadel of Troy was surrounded by an 8m high stone fortification wall consisting of large limestone blocks.96 The wall was constructed with vertical offsets which divided the structure into sections; this so called saw-tooth technique was used to provide extra stability against natural disasters such as earthquakes.97 The fortification wall could be entered by four gates; the largest gate functioned as the main entrance to the citadel and was situated at the southern end. Only this gate provided access for wagons and chariots.98 The other three gates were significantly smaller and could only be accessed by pedestrians.99
No traces of a Troy VI palace are found until the present day.100 A possible explanation for the absence of this building at the citadel of Troy is that during the Hellenistic occupation of the site in the 3th century BC, many Bronze Age remains were removed to construct the Athena temple. 101 The temple is situated at the highest point of the citadel which would have been the most favorable location to build a LBA palace as well. Moreover, many LBA remains from inside the citadel were dug away during the excavations of Schliemann between 1870 and 1890.102
Only a few Troy VI Early houses have been excavated and even less complete house plans could be determined (fig. 14).103 The houses of the Troy VI Early period often have stone foundations but wooden houses made with posthole constructions are also known.104 The sporadic evidence of the houses from this period may be explained by the re-use of much of the building materials during later habitation phases (Troy VIII and IX).105
During the Troy VI Early period, habitation at Troy extended outside of the citadel.106 Although the previous excavators at Troy did not doubt the presence of a lower city,107 its
96 Jablonka 2010, 853. 97 Becks 2006, 155. 98 Ibid., 156. 99 Ibid.,156. 100 Ibid., 156. 101 Ibid., 156. 102 Jablonka 2010, 853. 103 Becks 2006, 158. 104 Korfmann 1998, 52. 105 Blegen et al. 1953, 119.
106 Habitation outside the citadel already existed during the Troy II period and the Troy V period although the
evidence for the Troy V period is very sparse. During the Troy VI period, the evidence for habitation outside the wall is much more abundant. Blegen et al. 1958, 6-13; Korfmann 1992, 123-146; Becks 2003, 42-44; Easton et al. 2002, 82-94; Jablonka and Rose 2004, 616-620; For an alternative view see: Hertel and Kolb 2003, 77-86; Kolb 2004, 596-597.
107
Blegen 1953, 351; Dörpfeld 1902, 238; Schliemann 1884, 62-63.
22
existence could only be properly confirmed when the new excavations at Troy began.108 During the new excavations which started in 1988, a series of houses and other finds dating to the Troy VI and VII period were uncovered south and southwest of the citadel. Moreover, during a systematic surface survey carried out in 2002 and 2003 on the surrounding slopes of Troy, large quantities of Bronze Age pottery and, in particular, sherds dating the Troy VI and VIIA period were collected.109
The pottery from Troy VI Early can be subdivided into two ceramic phases: ceramic phase 1 and ceramic phase 2(fig.3, 4, and 5).110 Ceramic phase 1 dates from ca. 1750 to 1680 BC and ceramic phase 2 dates from ca. 1680 to 1590 BC. The ceramic repertoire of phase 1 stands largely in the tradition of the preceding Troy V period as can be seen by characteristics such as the wheel-made bowls and cups in red coated ware (RCW) and the beige plates in plain ware. These wares are fine, slightly porous and high fired. The coarse wares of ceramic phase 1 such as gritty ware and quartz coarse ware are similar to the coarse wares of Troy V although several new shapes are introduced as well (fig. 7).111
Nevertheless, several important changes in the pottery are visible as well: the most important development is the introduction of the Anatolian Grey Ware (AGW) at Troy.112 Only a small percentage of the ceramic repertoire of phase 1 consists of AGW but over the next few centuries this ware would become one of the most dominant pottery wares at Troy.113 AGW usually has a grey surface, is wheel-made, has a fine clay and is high fired.114 The grey surface is in huge contrast with the pottery from the previous periods since this pottery was always brown, red or beige.115 The ware has a soapy feeling and, especially in the early phases (AGW I), a shiny grey surface.116 It has been suggested that the shiny grey surface of AGW was meant to imitate metal vessels.117 The shapes of AGW in phase 1 show
108 Korfmann 1992, 123-146. The new excavations of Troy is a joint project by the Institut für Ur- und
Frühgeschichte und Archäologie des Mittelalters by the Eberhard Karls Universität Tübingen (Germany) and the Department of Classics of the University of Cincinnati (USA) and were carried out under the direction of M. Korfmann and B. Rose. Korfmann 1992, 123-146.
109 The surface survey was carried out by the Eberhard Karls Universität Tübingen under the supervision of
Peter Jablonka. See Korfmann 2004, 9-12.
110 The division of the ceramic phases of Troy VI Early is proposed by P. Pavúk. (2005) In this thesis, this
system will be followed.
111 Pavúk and Rigter 2006, 232-233.
112 Pavúk 2002b, 57; Pavúk and Rigter 2006, 233. 113 Pavúk and Rigter 2006, 233.
114
Ibid.
115 Ibid.
116 Pavúk 2005, 192.
117 Becks 2006, 159 Similar suggestions have been made for the Grey Minyan Ware occurring in Mainland
Greece during the Early Helladic Age III and Middle Helladic period See e.g. Rutter 1983.
23
strong resemblances with Aegean ceramic shapes such as the so-called Lianokladi-goblet and the rounded cup.118 However, the shapes of this period are not only marked by Aegean influences: the beard rim bowl which occurs at Troy from the first ceramic phase onwards is a typical Anatolian vessel shape (fig. 4).119
Imports make up a small percentage of the total ceramic repertoire of ceramic phase 1. Ceramic imports can be divided into pottery coming from neighbouring and distant regions.120 Neighbouring imports are ceramics produced in the close vicinity of Troy including the islands off the coast of Troy. Distant imports are the ‘exotic’ imports produced at for example the mainland of Greece or the south-west coast of Anatolia.121 The relative percentage of the neighbouring imports in the first ceramic phase is ca. 2-3%, distant imports only make up ca. 1% of the total ceramic assemblage (table 1).122 One Minoan jug was found in a cist grave of a child in area A7, situated directly southwest of the fortification wall.123 Approximately ten medium to large Cretan transport vessels were found in ceramic phase 1 strata as well.124 From the first ceramic phase onwards, the so called ‘Mica ware’ appears at Troy in deposits within and directly outside of the citadel.125 The highly micaceous ware usually occurs in thick-walled vessels such as pithoi or cooking pots. A total of 43 fragments dating to various sub-phases of Troy VI have been recorded thus far.126 The provenance of these vessels is still unknown, although Kythera or Northern Greece are suggested as possible production centers.127 In addition, distant imports from the Cyclades, possibly reaching the site from the neighbouring islands of Troy and a small amount of sherds, perhaps produced in northern Syria, were attested in ceramic phase 1 of Troy VI.128
In the second ceramic phase, the Troy V vessel shapes as well as the Aegean goblets and the red coated ware gradually disappear.129 AGW becomes more widespread and several new beige and red fine wares such as Burnished Plain ware II and Red Coated ware III are
118 Pavúk and Rigter 2006, 233. 119
Bayne 2000, 23.
120 This division has been introduced by P. Pavúk and will be used throughout this chapter as well. Pavúk 2005,
216-218
121 Pavúk 2005, 216. 122
Ibid., 218.
123 Recent reassessment of the chronology of the grave suggests that it can be dated to the end of Troy V (early
ceramic phase I) Blum 2006, 152-153; Girella 2005, 339.
124 Pavúk and Rigter 2006, 233. 125 Pavúk 2005, 206. 126 Ibid. 127 Ibid, 207. 128 Pavúk 2005, 215. 129
Pavúk and Rigter 2006, 233.
24
introduced.130 Two handled cups, kantharoi, oinochoai and two handled bowls with ripped shoulders are popular among the new shapes while several shapes from the first ceramic phase such as the bear-rim bowl are still in use as well (fig. 5).131 Small cups and basins, often decorated with white paint, are found in Plain ware.132
Large transport vessels, which strongly resemble vessels known from Samothrace and other neighbouring islands such as Lemnos and Lesbos, are found in ceramic phase 2 deposits at Troy (fig. 6). The so called Island wares comprise ca. 10% of the ceramic assemblage of Troy in the second ceramic phase, a significantly higher number than in the previous or the succeeding periods (table 1).133 The appearance of the Island wares in ceramic phase 2 points to intensive contacts between Troy and its neighbouring maritime regions during this period. Imports from distant regions comprise only ca. 0.5% of the pottery assemblage of ceramic phase two: a few large closed vessels from Miletus and several small closed vessels from the South Aegean are found in ceramic phase two deposits. 134
2.3.2 Troy VI Middle
The Troy VI Middle period roughly dates from 1590 to 1430 BC.135 This period corresponds to Blegen’s architectural phase VI D, E and F and Pavúk’s ceramic phase 3.136 The Troy VI Middle period is marked by a number of important architectural and ceramic innovations. At the beginning of Troy VI Middle, a new fortification wall was constructed which exceeded the previous wall in height and width.137 The wall of Troy VI Middle was slightly sloping which provided more stability against possible natural disasters. Apart from the fortification wall, the architectural evidence of Troy VI Middle is very scattered: only a few Troy VI Middle houses inside and around the citadel have been uncovered. We do know that, compared to the houses of the previous period, the houses of Troy VI Middle were larger.138
130
Pavúk 2005, 181, 187; Pavúk and Rigter 2006, 233.
131 Pavúk and Rigter 2006, 233. 132 Ibid.
133 Pavúk 2005, 218. 134
These vessels most likely come from Crete but a Cycladic origin cannot be excluded. Pavúk 2005, 211, 218.
135 Pavúk 2005. 136 Blegen 1953; Pavúk 2005. 137 Becks 2006, 155. 138 Ibid., 158. 25
400m south of the citadel, a ditch dating to the Troy VI Middle period was discovered during a magnetic prospection survey in 1992.139 The ca. 120m long anomaly was initially thought to be a burned mud-brick wall, however, the excavation pointed out that the feature was, in fact, a ditch.140 The ditch is 2,5m deep, 6m wide and periodically bridged by approximately 10m long causeways.141 The function as well as the chronology of the ditch is under debate. 142 In particular since a defense ditch is usually accompanied by an actual fortification wall which, in the case of the ditch found at Troy, has not been discovered yet. However, if the ditch marks the border of the outer city of Troy during the LBA, this would suggest that the settlement area of Troy was much larger than the citadel alone (ca. 30 ha) and could inhabit ca. 6000 people.143
Compared to the architectural evidence of Troy VI Middle, the period is significantly better documented in terms of pottery. AGW becomes the dominant ceramic fine ware in phase 3 and other local red and beige wares stay popular. The most important development in the local ceramic production of Troy is the introduction of the so called Tan ware. Tan ware is a beige variant of AGW: the oxidizing atmosphere in which Tan ware is fired results in a light fabric while the reducing firing conditions of AGW produces its distinct grey colour.144 In phase 3, only a small percentage of the ceramic assemblage of Troy consists of Tan ware but over the next centuries the ware would become one of the most widespread wares at Troy.145 Several pottery shapes of the previous period, such as oinochoai and kantharoi, are still in use while other shapes such as the rounded bowl disappear.146 Grey carinated cups, two-handled deep bowls with plastic knobs (‘beads’) and two handled jars with horizontal incisions are among the most popular new shapes of this period (fig.5).147
At the start of the third ceramic phase, the Island wares still occur at Troy but towards the end of the period they have disappeared completely. Neighbouring imports comprise ca. 4-5% of the total pottery assemblage in ceramic phase 3. Distant imports make up ca. 2-3%
139 Becker et al. 1993, 122. The ditch was dated to the Troy VI Middle/Early VI Late period on the basis of
ceramic fragments, including several Mycenaean sherds found inside the ditch as well as a series of C-14 samples. Jablonka 1995, 61-76; Jablonka 1996, 73.
140
Blindow et al. 2000, fig 1.; Korfmann 2001a, fig. 1.
141 Becker et al. 1993, 117.
142 See: Easton et al. 2002; Hertel and Kolb 2003; Jablonka and Rose 2004; Kolb 2004
143 Tentative estimation by Korfmann ( Korfmann 1992, 138; Korfmann 1996, 91-92). Contra: Hertel and Kolb
2003, 74-76
144 Pavúk 2005, 197.
145 Pavúk and Rigter 2006, 235. 146 Ibid.
147
Ibid.
26
of the assemblage (table 1). 148 In the third ceramic phase, Mycenaean pottery and local imitations of Mycenaean-type vessels occur for the first time at Troy. The earliest Mycenaean imports can be dated to the LH IIA (the beginning of ceramic phase 3).149 Apart from the Mycenaean vessels, 63 fragments, mostly belonging to medium to large transport vessels, are considered to be Milesian imports.150 5 sherds belonging to Cypriot pottery were found in Troy VI Middle deposits.151 A small quantity of polychrome matt painted vessels were found in Troy VI Middle deposits as well. These vessels are handmade, thin walled and always appear in closed shapes. The vessels in question are found inside and in close proximity of the fortification wall.152 The provenance of these vessels has not been determined yet although Rhodes has been suggested as a possible production center.153 Two Minoan stirrup jars were found in graves and can be dated to the end of the third ceramic phase.154
Table 1: Pottery of Troy VI: absolute and relative proportion of local pottery, neighbouring imports (‘nahimporte’) and distant imports (‘fernimporte’) of ceramic phase one, two and three (Pavúk 2005, 218,
fig. 81)
2.3.3 Troy VI Late
The Troy VI Late period dates from ca. 1430 to 1300 BC. This period corresponds to Blegen’s architectural phases VI G and H and Pavúk’s ceramic phase 4. The architectural evidence of Troy VI Late is much more numerous than that of the preceding phases. At the start of this period, the fortification wall of Troy VI Late is enlarged again: the wall is now 4
148 Pavúk 2005, 218. 149 Mountjoy 2006, 242-243. 150 Pavúk 2005, 206. 151 Kozal 2006, 253. 152 Pavúk 2005, 210.
153 Pavúk 2005; Pavúk and Rigter 2006. 154
Mountjoy 1997, 283-285.
27
to 5 meters wide, 530m long and surrounds an area of 1,8 ha.155 The houses of Troy VI Late are larger than the houses of the previous periods and are often built on concentric terraces.156 Various types of large freestanding houses such as the megaron house, the L-shape houses with multiple rooms and two-floored houses with pillars are known from within the citadel (fig. 14).157 Outside the citadel, the houses usually had a rectangular shape.158 At the end of Troy VI Late, the fortification wall and most of the buildings inside and around the citadel were most probably destroyed by a violent earthquake.159
Very little is known about the funerary behaviors of inhabitants of Troy during the LBA. The only evidence comes from a small cemetery located ca. 550m south of the citadel. The south necropolis was partly excavated by Blegen in 1934 and can be dated to the end of the Troy VI Middle/Troy VI Late period. The cemetery was heavily disturbed: only 20 out of the ca. 200 graves were still more or less in its original state.160 The corpses were usually cremated and buried in urns although some examples of inhumation of infants in dual burials are also present.161 Since the grave goods accompanying the human remains in the south-necropolis were fairly poor, it is believed that this cemetery was used by the inhabitants of the lower city as opposed to the residents of the citadel.162 The other necropolis at Troy has, until the present day, not yet been discovered. The only other burial evidence in the vicinity of Troy is the LBA Beşik-tepe cemetery, located ca. 8km west of Troy.163
This cemetery is better preserved then the south-necropolis and is much richer: the death were often accompanied with rich grave goods such as jewelry, seals, weapons and Mycenaean vessels.164
Part of the pottery repertoire of ceramic phase 3 is still in use in ceramic phase 4 although new shapes such as plates (with or without a foot), shallow bowls, two-handled carinated cups and kylikes occur as well.165 Many of the vessel shapes of this period derive from the Mycenaean pottery repertoire although Anatolian vessel types, such as the pilgrim
155 Becks 2006, 155. 156 Bayne 2000, 38. 157 Becks 2006, 156. 158 Ibid., 158. 159 Blegen 1953, 15-20.
160 The reason for the heavy disturbance of the cemetery is that the burials were covered with relatively little
soil, a fortification wall dating to the Hellenistic period was built on top of the cemetery, military trenches were built here in 1915 and many of the stones were removed in later occupation phases of Troy. Blegen 1963, 115.
161 Becks 2006, 162. 162 Blegen 1953, 379; Becks 2006, 163. 163 Basedow 2000. 164 Becks 2006, 163; Basedow 2000, 53-144. 165
Rigter 2013, 8; Pavúk and Rigter 2006, 236.
28
flask, also appear at Troy.166 AGW and Tan ware are the two most common fine wares in this period. Both wares are produced in a very standardized manner indicating a high degree of organization of the ceramic production.167 In Troy VI Late, one fourth of the fine wares are produced in a Mycenaean style. However, only 2 to 3% of the pottery is actually imported from the Greek mainland.168 Mycenaean-style pottery appears in a variety of open vessel shapes, such as cups, kraters, and kylikes and closed shapes, such as amphorae, amphoroid kraters, alabastra, stirrup jars, rythons, flasks, and jars.169 Compared to ceramic phase 3, a somewhat larger amount of Cypriot pottery is known.170 With the exception of five sherds, the Cypriot pottery has all been found in and around the citadel.171
2.3.4 Troy VII A
Troy VIIA spans from ca. 1300-1190 BC. This period corresponds to Blegen’s architectural phase Vii and in terms of pottery, the period can be defined as the fifth ceramic phase.172 After the destruction of Troy VI, the settlement was largely rebuilt. Several architectural features of the previous period, such as the circular terraces and the fortification wall, were reused and new architectural structures, such as the two towers situated outside the southern and western part of the citadel, were constructed in Troy VII A. 173 On the foundations of the fortification wall, a new stone wall was placed and several of the old entrances were blocked.174 The road leading to the main entrance of the citadel was paved and provided with a water channel.175
The houses of Troy VIIA within the citadel are significantly smaller than the houses of Troy VI Late (fig. 15). The houses in this period were often built against the fortification wall and contained a large amount of storage vessels.176 The fact that some of the entrances to the citadel were blocked and the discovery of large amounts of storage vessels in the houses,
166 Pavúk and Rigter 2006, 236. 167
Ibid., 239.
168 The provenance of Mycenaean style pottery from Troy has been examined chemically by various scholars.
See: Knacke-Loy 1994; Mommsen (eds.) 2001; Mountjoy and Mommsen 2006.
169 Mountjoy 2006, 242-245. 170
In total, 61 fragments of Cypriot pottery were found at Troy out of which 13 can be securely dated to a Troy VI Late deposit. Kozal 2006, 253.
171 Kozal 2006, 253.
172 Although the Troy VIIA period is not included in P. Pavúk’s pottery study, the period has been defined by
him as the fifth ceramic phase, since a continuation in pottery traditions is visible. Pavúk 2005, 332.
173 Becks 2006, 156 174 Ibid. 175 Blegen 1958, 48. 176 Ibid., 6-7. 29
could be an indication that the inhabitants of Troy experienced a sense of insecurity during this time.177 Outside the citadel, several freestanding houses are known from the Troy VIIA period but only a single house could be fully excavated. This house, situated west of the citadel in area Z7/8, contained two floors and had a central room with various smaller rooms connected to it (fig. 15).178 The house provides us with important insights in the living conditions of inhabitants of Troy during the Troy VIIA period. Similar to the houses found within the citadel, a large number of storage vessels were placed in the house. In various rooms, spear points were found and on the street outside of the house, a sling stone was discovered.179 These finds are interpreted as the remains of a lost war.180 The house was eventually destroyed by a fire.181
In terms of pottery shapes and wares, Troy VI Late and Troy VIIA show many resemblances. The fact that no clear cultural break is visible in the pottery may suggest that Troy was (largely) inhabited by the same people.182 Nevertheless, based on the relative proportions of the wares and the appearance of several new wares and shapes, the two periods can be distinguished: Tan ware becomes the most popular ware and a new brown variant of the Tan ware is introduced as well.183 Plates with thickened rims, a series of small jugs, a specific type of carinated cups and a deep two-handled bowl are among the new shapes of Troy VIIA (fig. 8).184
Mycenaean style pottery still appeared frequently at the site and the local imitations were often produced in Tan ware and AGW during Troy VIIA.185 Mycenaean pottery occurs in open shapes such as cups, kylikes, skyphoi and bowls, and closed vessels such as amphorae and alabastra. Besides the Aegean influence on the pottery at Troy, a mixture of traditions is illustrated by an Anatolian vessel type containing Mycenaean decoration.186 Small quantities of Cypriot imports occur in various wares such as White slip II, Base Ring II, White Shaved ware and Plain White wheel-made in Troy VII A deposits.187 Two
177 This evidence has led scholars to believe that this phase can be appointed to the Homeric Troy. Blegen 1963;
Becks 2006; Jablonka 2010. 178 Becks 2006, 159. 179 Ibid. 180 See note 177. 181 Becks 2006, 159.
182 Rigter 2013, 8; Pavúk and Rigter 2006, 237. 183
Pavúk and Rigter 2006, 237; Blegen et al. 1958, 22.
184 Rigter 2013, 9. 185 Mountjoy 2006, 247.
186 Mountjoy 1998, 41; Mommsen et al. 2001, 189; Mountjoy 2006, 248. 187
5 samples from a VIIA and 17 samples from a VI Late/VIIA deposit are known. Kozal 2006, 251-253.
30
fragments of Canaanite jars which probably originate from the Levant, were discovered as well.188
188 Pavúk and Rigter 2006, 237.
31
Chapter three
Methodology
One of the main aims of ceramic compositional studies is to establish the compatibility of the pottery with locally available sediments and to reconstruct the range of ceramic resources that were in use at a specific site or area at a specific time. If corresponding rocks, minerals or association of minerals appear in the pottery and the nearby geological sources, we say that a fabric is locally compatible. When this is not the case, we may suggest that the ceramic in question was produced elsewhere and transported to the place of deposition. There are a number of reasons to transport a ceramic vessel such as trade, exchange or the migration of people. Therefore, both the local production of ceramics and the pottery produced elsewhere (imports), can shed light on social, cultural and economic aspects of a society.
In this research, the ceramic composition or the fabric of 230 samples from LBA Troy will be studied. Fabric refers to the arrangement, size, shape, frequency and the composition of components of a ceramic.189 An integrated approach using both macroscopic and microscopic research will be performed to study the mineralogical composition of the samples. In this study, a qualitative visual approach will be used, which means that the sherds and the thin sections will be sorted manually, after which they will be described. Grouping is a crucial step in the analysis of ceramic thin section, which can lead to an archaeologically meaningful compositional patterning within ancient ceramic assemblages.190 The basic concept of a fabric group is that the combination of the inclusions, the clay matrix and the voids differ in a significant way from other samples within the studied assemblage. Differences in the firing process or secondary calcite can create visual differences although they are not related to the compositional character of the sherd and are therefore insignificant to the grouping process.191 The presence or absence of rare inclusions or other features is normally also not a valid characteristic to split an otherwise coherent group because small variations in the abundance or size of the inclusions do not have to be archaeologically significant. 192 The visual grouping of sherds and thin sections will be done without taking the other characteristics of the ceramics, such as morphology, stylistic characteristics, location of excavation and chronology, into account.
189 Whitbread 1995, 368. 190 Quinn 2013, 71. 191 Ibid., 76. 192 Ibid, 76. 32
Supporting compositional ceramic analyses, such as neutron activation analysis, scanning electron microscopy or X-ray fluorescence, will not be conducted in this study. To explore the local compatibility of the sherds, various sources will be consulted. No geological samples could be taken for this study, which limits the potentials of the analytical results. Nevertheless, geological maps, geological reports and former compositional studies of Trojan ceramics are available and will be used in comparison to the established macroscopic and petrographic fabric groups. 193
Macroscopic fabric analysis and petrographic research are, to some extent, subjective methods of analyses. The grouping of an assemblage partly depends on the experience and the perception of the analyst on a particular moment.194 Moreover, in order to be able to identify the distinguishing properties of a mineral, large or well-formed examples are usually required.195 While such examples are often available when studying rocks, ceramics generally contain small and isolated inclusions and are therefore far more complex to recognize. As a result, the process of grouping is subjective on multiple levels.196 Besides the complexity of the identification process, a great diversity exists in the techniques and terminology used to carry out macroscopic and petrographic analyses. In order to minimize any confusion on terminology and, more importantly, to optimize the comparability of this study to other future studies, the methodology will be clarified in greatest detail as possible in this chapter.
The two scales of compositional studies, carried out in this research, are based on the system developed by Ian Whitbread.197 By using a system developed by the same person for both the macroscopic and microscopic analysis, it is hoped that the comparison between the two scales of analysis will be less complicated since the variables largely overlap (see below). Before outlining the methodology of the compositional ceramic analyses, the sampling strategies that led to the creation of the current dataset, will first be discussed.
3.1 The Data
The dataset used in this thesis consists of 230 ceramic sherds from Troy (table 2). The majority of the samples (77%) comes from Troy VI Early to Troy VIIA deposits (ca.
1750-193 Knacke-Loy 1995; Mommsen et al. 2001; Guzowska et al. 2003; Mountjoy and Mommsen 2006; Mommsen
and Pavúk 2007; Schubert and Pernicka 2007; Tekkök and Pernicka 2012; Grave et al. 2013; Kibaroğlu and Thumm-Doğrayan 2013. 194 Quinn 2013, 76. 195 Whitbread 1995, 367. 196 Ibid. 197 Whitbread 1989. 33
