New insights into the Mesolithic use of Melos obsidian in Anatolia: a pXRF analysis from the Bozburun Peninsula (southwest Turkey)
Introduction
Southwest coast of Anatolia and its immediate hinterland have remained largely empty with respect to prehistoric investigations, even though the regions to the north and east have yielded a relatively rich archaeological record (Çilingiroğlu et al., 2020, Erdoğu et al., 2021, Otte et al., 1998, Otte et al., 2003, Özdoğan et al., 2012, Roosevelt et al., 2019, Taşkıran et al., 2021). One recent survey project was initiated in the Bozburun Peninsula to alleviate this situation (Atakuman et al., 2020). Bozburun Peninsula is located at the southwestern tip of Anatolia and lies to the north of the island of Rhodes (Fig. 1), positioning it at the juncture between Anatolia, the Aegean, and the Mediterranean. The survey project began in 2017, and one of the primary research objectives was to inquire into the role the peninsula might have played during the poorly understood Mesolithic-Neolithic transition in light of its strategical location. Finds from the survey indicate that the Bozburun Peninsula was repeatedly visited along a time frame between the Middle Palaeolithic and the Neolithic/Chalcolithic. The discovery of sites in Bozburun with flake-based lithic assemblages typical of the Aegean Mesolithic is especially noteworthy as forager presence is already known from nearby islands such as Ikaria, Fournoi, and Chalki in the Dodecanese directly across the sea from southwest Anatolia (in addition to the more distant islands such as Crete, Kythnos, Lemnos, Naxos, Sikinos, and Youra) (Sampson, 2019), hinting at the contemporary occupation of both the Aegean Islands and western Anatolia in the periods preceding the Neolithic. Such a situation has ramifications for how the Mesolithic-Neolithic transition in the Aegean can be framed with respect to migration and local adoption scenarios, a debate that is far from resolved within the archaeological research (Carter, 2019, Leppard, 2021, pp. 12-15; Reingruber, 2018; see also Atakuman et al., 2020 for a synthesis). What prevents the contribution of a substantial Anatolian perspective to the debate is the lack of excavated pre-Neolithic sites in western Anatolia, which is also the reason why ancient DNA research at the moment is inconclusive with respect to the nature of this transition (Kılınç et al., 2017). One exception is the site of Girmeler, roughly 110 km to the east of the Bozburun Peninsula, where ongoing excavations have revealed early layers that are well poised to provide insight into the transformations that took place within this timeframe (Erdoğu et al., 2021).
Another group of finds that provides clues about the connectivity in the prehistoric Aegean is the large number of obsidians recovered from the Bozburun Peninsula (Atakuman et al., 2020). Obsidian is a naturally occurring volcanic glass, often with a lustrous appearance, that can yield very sharp edges when fractured (Dewbury and Russell, 2007, Ericson et al., 1975). Such properties of obsidian were likely among the contributing factors that increased its prehistoric demand both in Bozburun and elsewhere (Ibáñez et al., 2016), together with its relative scarceness as a resource in the wider Mediterranean basin (Renfrew et al., 1966, Tykot, 2017b). Obsidian had to arrive in Bozburun from a long-distance, as neither the Bozburun Peninsula nor southwest Anatolia contains obsidian deposits. Closest sources lie in the Aegean, where obsidian outcrops in three islands in the southern part of the sea (Renfrew et al., 1965): Melos and Antiparos in the Cyclades, and Giali (alternatively transcribed as Yali or Gyali) in the Dodecanese. Another cluster of obsidian sources is known from Central Anatolia in the Cappadocia region, with exposures in Acıgöl, Hasan Dağ, Nenezi Dağ, and Göllüdağ (Poupeau et al., 2010). In between, there are reported outcrops in north-central and western Anatolia, but none of these sources appear to be extensively utilized (if at all) in prehistory (Chataigner et al., 1998, Düring and Gratuze, 2013).
Relying exclusively on visual identification in the face of such a plurality of sources brings problems: while some of the peralkaline obsidians can be recognised based on their greenish colours, most of the obsidian sources around the Mediterranean are alkaline or calc-alkaline in composition and exhibit overlapping visual characteristics (Williams-Thorpe, 1995; Table 2). Luckily, it is possible to distinguish different obsidian sources using their geochemical signatures obtained through a number of characterisation techniques (Shackley, 1998). Archaeological studies have taken advantage of these various techniques (see, for example, Section 3.2 for a selection of methods used in a number of previous studies in the Aegean) to track the prehistoric transportation of obsidian since the 1960s (Williams-Thorpe, 1995). More recent decades have witnessed a further upsurge in the investigation of ancient interaction and mobility patterns owing to the increasing popularity of portable X-ray fluorescence (pXRF) systems (Freund, 2013, Shackley, 2011, pp. 12-15). Likewise, the goal of this paper is to present the results of the pXRF analysis of obsidian artefacts recovered from the Bozburun Peninsula to shed more light into the procurement strategies employed by its prehistoric inhabitants.
The survey project in Bozburun started in 2017, and was continued in the 2018 and 2019 seasons (Atakuman et al., 2020). A total of 45 find spots were identified (Fig. 2), yielding mainly chipped stone artefacts but also rare pottery sherds, ground stone implements, and a few other small finds. 15 of these 45 find spots contained a total of 649 obsidian pieces, including cores, blade/bladelet and flake blanks, tools, and debris (Fig. 3). The overwhelming majority of the obsidian finds come from the sites of Sarnıç, Sobalak, Çakmak, Zeytinlik, and Hurma. The characteristics of the overall chipped stone assemblages of these five sites show parallels with those of the Aegean Mesolithic, while further diagnostic tools hint at visits or occupations during the Middle Palaeolithic, Upper Palaeolithic, Epipalaeolithic, and the Neolithic/Chalcolithic periods (Atakuman et al., 2020, Table 2). Apart from these five sites and the cave of Kayabaşı (which did not yield any obsidian objects), the remaining find spots in Bozburun are represented by small numbers of finds that do not allow (so far) confident deductions regarding their chronology. For the purposes of this article these latter localities are collected under the rubric of “minor sites”.
Table 1 lists the number of obsidian finds from different sites and their proportion in each assemblage. The site of Hurma is especially notable as it yielded 426 of all 649 obsidians encountered in the Peninsula. Hurma, together with Zeytinlik, is also characterised by an intensive use of obsidian compared to other resources (>70%). But one note of caution for these proportions is that all the BPS sites are known only from surface investigations at this moment and lack a stratified sequence, and the palimpsest nature of these assemblages might be amplifying or masking any chronological patterns that originally existed. A second note concerns the survey collection strategy regarding different raw materials: since the mere presence of obsidian in the Bozburun Peninsula indicates human activity, all encountered obsidian objects were recorded. For other raw materials, the survey opted for those objects that have a clear place in the operational sequence of lithic manufacture and use, i.e. pieces that exhibit evidence of human involvement (as opposed to local geological materials scattered and deposited as colluvium without any hint of direct interference).
Section snippets
Materials and methods
Preliminary observations of the Bozburun assemblages had determined at least two distinct obsidian groups based on macroscopic qualities (Atakuman et al., 2020, p. 16, Table 1): tentatively, those obsidians exhibiting a glossy black surface with white spherulites were accepted as originating from Giali, while the duller, dark grey semi-translucent pieces were considered to be Melos obsidians (Carter et al., 2016, Fig. 13; Milić, 2014, Fig. 9). Yet, as mentioned in the previous section, visual
Results
Our results indicate that the Bozburun obsidians can be divided into four groups based on the relative concentrations of the trace elements Rb, Sr, and Zr (Fig. 5, left). A comparison of these results with those from the published literature reveals that the Melos, Giali, and Göllüdağ sources are represented in the Bozburun Peninsula (Fig. 5, right). The following 3.1 Sources not represented in Bozburun, 3.2 Melos and Giali, 3.3 Göllüdağ (and other Central Anatolian obsidians) provide an
Obsidian distribution around the Aegean: implications from Bozburun
Ultimately, our pXRF results indicate that the island of Melos, including its two distinct outcrops, was the primary source of obsidian used in the Bozburun Peninsula, followed by the obsidians that originated in the island of Giali (Table 4). These islands lie to the west of the peninsula, between 90 (Giali) and 330 (Melos) km in distance. In addition to these two sources, obsidians from Göllüdağ East (in Central Anatolia) are also represented by six pieces in Bozburun, deriving from outcrops
Conclusion
Our pXRF analysis of the 649 obsidian artefacts from the 2017–2019 surveys of the Bozburun Peninsula revealed the long-distance and long-term connections of the inhabitants of the peninsula in both directions. Obsidians from the Melos A and Melos D outcrops, at the other side of the Aegean, appear to have reached Bozburun before the Neolithic - the earliest recorded landfall of these materials in the Anatolian mainland. Pre-Neolithic forager networks in the Aegean therefore incorporated at
CRediT authorship contribution statement
Hasan Can Gemici: Conceptualization, Formal analysis, Investigation, Methodology, Validation, Writing – original draft, Writing – review & editing. Murat Dirican: Investigation, Methodology, Validation, Writing – review & editing. Çiğdem Atakuman: Conceptualization, Investigation, Project administration, Supervision, Writing – review & editing.
Declaration of Competing Interest
The authors declare that they do not have any competing financial or personal relationships which could have inappropriately influenced this work.
Acknowledgements
This work was supported by the European Research Council (ERC) project ‘‘NEOGENE: Archaeogenomic Analysis of Genetic and Cultural Interactions in Neolithic Anatolian Societies’’ (grant number 772390). Geological samples from Eastern Anatolia were kindly provided by İsmail Baykara from Gaziantep University. We are also grateful to Ellery Frahm and one anonymous reviewer for their comments and constructive criticism.
References (74)
- et al.
Provenance of obsidian artifacts from the Chalcolithic site of Dava Göz in NW Iran using portable XRF
J. Archaeol. Sci.: Rep.
(2018) - et al.
Chemical characterization of obsidians from different Mediterranean sources by non-destructive SEM-EDS analytical method
J. Archaeol. Sci.
(1999) - et al.
Détermination de provenance d’une sélection d’obsidiennes du palais minoen de Malia (Crète)
C.R. Palevol.
(2008) - et al.
New investigations of the Göllüdağ obsidian lava flows system: a multi-disciplinary approach
J. Archaeol. Sci.
(2011) - et al.
Multiple sources: the pXRF analysis of obsidian from Kenan Tepe, S.E. Turkey
J. Archaeol. Sci.: Rep.
(2016) - et al.
The character and use of the Soros Hill obsidian source, Antiparos (Greece)
C.R. Palevol
(2012) - et al.
Networks and Neolithisation: sourcing obsidian from Körtik Tepe (SE Anatolia)
J. Archaeol. Sci.
(2013) - et al.
A new programme of obsidian characterization at Çatalhöyük, Turkey
J. Archaeol. Sci.
(2006) - et al.
Obsidian circulation in the Early Holocene Aegean: a case study from Mesolithic Damnoni (SW Crete)
J. Archaeol. Sci.: Rep.
(2018) - et al.
Turkish occurrences of obsidian and use by prehistoric peoples in the Near East from 14,000 to 6000 BP
J. Volcanol. Geoth. Res.
(1998)