One Thousand Years of Mediterranean Silver Trade to the Levant: A Review and Synthesis of Analytical Studies

Abstract

Silver exchanged by weight for its intrinsic value was the most important measure of value and means of payment in the southern Levant, starting from the Middle Bronze Age II–III through the Iron Age (~1700/1650‒600 BC). Since silver is not available locally in the Levant, its ongoing use as currency in the region triggered long-distance trade initiatives, and its availability or lack thereof had a direct impact on the economy. The continued use is evidenced in 40 silver hoards found in various sites across the region. A comprehensive study of lead isotopes and chemical analyses of samples obtained from 19 hoards enabled us to trace the origin of silver in the millennium during which it was extensively used as currency in the southern Levant and to identify constantly changing silver sources and concomitant trade routes. The results indicate that silver originated initially in Anatolia and Greece (~1700/1650–1600 BC) and shortly after from an unknown location in the Aegean/Carpathian/Anatolian sphere (~1600–1200 BC). After the collapse of Late Bronze Age Mediterranean trade routes, during Iron Age I (~1200–950 BC), there was a period of shortage. Silver trade was revived by the Phoenicians, who brought silver to the Levant from Sardinia and Anatolia (~950–900 BC), and later from Iberia (~900–630 BC). Further change occurred after the Assyrian retreat from the Levant, when silver was shipped from the Aegean (~630–600 BC). Following the devastation caused by the expanding Babylonian empire, silver consumption in the Levant practically ended for a century. Considering the isotopic results, combined with a detailed study of the context, chronology, and chemical composition, we demonstrate that all these factors are essential for the reconstruction of developments in the supply of silver in the southern Levant, and more generally. The changes in trade routes closely follow political and social transformations for over a millennium; exchange in this case was not only, not even mainly preconditioned by the environmental/geographic circumstances, as has often been argued for the Mediterranean. From an analytical point of view, we offer a protocol for the provenance of silver in general.

Introduction

In the ancient Near East and around the Mediterranean, silver was one of the most coveted commodities, of paramount importance both economically and symbolically. Potential silver sources in these parts of the Old World are few. Nevertheless, the precise acquisition and distribution networks of the metal and the ways in which these have been conditioned by historical circumstance, on the one hand, and generated infrastructures and interaction spheres, on the other, have never been outlined accurately and comprehensively enough to become historically significant. The main reason for this lacuna was the relative dearth of analytical studies of silver in the Mediterranean/ancient Near East to pinpoint its geographical source. We argue that other factors, such as the lack of chronological precision and the very methodologies used in the analytical protocols, also contributed to the rather blurred historical perspective regarding silver exchange and concomitant cultural phenomena.

In this paper we review and synthesize, from a long-term perspective (~1000 years), the results of a series of studies we have conducted over the last decade that focused on tracing the sources of silver that reached the southern Levant during the Bronze and Iron Ages. We concentrate on this region because this is where silver hoards are the most numerous, and where they are also accessible for sampling on a large scale. The many analyses, alongside precise chronology and careful methodology, enable us to reconstruct the changing sources of silver. In previous papers, we discussed separately each subperiod of the time span under investigation and also presented in detail the hoards, their contents and contexts, and the manner of their analysis (for a summary of these results, see Supplementary Material). Here, we concentrate on the wider emanating perspective. We start with generalities. We outline the significance of silver—used as currency—in the Near East, explaining why it was exploited and traded over long distances since the third millennium BC. We describe the relevant aspects of the production processes of silver from lead ores, and the main silver sources in the relevant regions. Then we explain the significance and uniqueness of the large number of silver hoards found in the southern Levant and the current state of debate regarding the provenance of the silver in them. Following a short review of the methods we employed, we broadly discuss the historical and archaeological implications of our main finds, contextualizing the results for each subperiod separately and then from a longue durée perspective. As well, since the analyses of silver items from the same hoards by other scholars yielded different interpretations, we explicate the differences in methodologies that, we believe, led to such diverse conclusions. We conclude by presenting the significance of synthesizing the analytical results, in a wide but also precise chronological framework, leading to new insights—both historically and regionally specific, but also more analytical and theoretical in general.

The Beginning: Silver Currency in Mesopotamia

Silver was used as a means of value, store of wealth, and means of payment and credit in the Near East starting in the mid-third millennium BC. It is mentioned in cuneiform sources as being exchanged according to local weight standards in standard forms such as rings, coils, and ax heads, as well as in cut silver ingots also known as hacksilber (e.g., Marti and Chambon 2019; Moorey 1999, p. 237; Peyronel 2010; Powell 1996; 1999; Radner 1999; Veenhof 1972, pp. 51–350). While other metals such as gold, copper, and probably tin were also used as measures of value, along with additional commodities such as barley or wool, the basic and most common metal of transaction was silver, which retained a relatively stable value in relation to other metals and commodities (Peyronel 2019; Powell 1990, pp. 79–80, 1996, pp. 227–228).

The persisting importance of silver in private and institutional economic transactions in Mesopotamia and Syria throughout the Bronze and Iron Ages is evident from cuneiform texts in several archives, including those of Mari (~18th century BC; Sasson, 2015), Emar (~14th–12th centuries BC), Ugarit (~12th century BC; Stieglitz 1979), and in Neo-Assyrian documents (mainly seventh–eighth centuries BC). Silver hoards unearthed in these regions, mostly containing cut silver in the form of rods, ingots, and jewelry, are amalgamations of wealth, hidden with the intent of future retrieval, therefore indicating their use as currency (Hudson 2020; Moorey 1999, p. 238; Peyronel 2010, 2019; Sherratt 2018).

Silver, as the main means of exchange, was a valuable commodity, essential for the Assyrian economy, which relied on its supply since the mid-third millennium BC (e.g., Sherratt 2018). Over 23,500 tablets written by Assyrian merchants at Kültepe-Kanesh in Anatolia in the 19th century BC reveal that Anatolia was the main source of silver to Mesopotamia during the Middle Bronze Age (MBA). Silver (and gold) was exchanged for tin, textiles, and other commodities using caravans that traveled back and forth between Anatolia and Assur (e.g., Barjamovic et al. 2012; Powell 1999; Radner 1999; Veenhof 1972; Yener 2015). Although not as well documented as for the Old Assyrian period, the quest for silver remained a driving force for long-distance trade during the following centuries, evident, e.g., in Kassite (mainly 15th–12th century BC) and Neo-Assyrian texts (mainly late ninth–seventh century BC; e.g., Bagg 2011; Gitin and Golani 2001; Kleber, 2016; Radner 1999).

Silver Production in Antiquity

Silver (Ag), which appears rarely in native form, was generally produced from silver-rich galena (PbS) and cerussite (PbCO₃) lead ores (Gale and Stos-Gale 1981b). The production process was two-stepped; smelting, involving the reduction of the lead ore into metallic Pb–Ag alloy, was followed by cupellation, namely, the oxidation of the alloy in a cupel to extract Ag (and gold) and separate it from litharge (lead oxide; e.g., Rehren and Klappauf 1995). In the Near East, silver was produced by cupellation since the fourth millennium BC (Nezafati and Pernicka 2012; Pernicka et al. 1998), while in the Aegean, this method was introduced no later than the second millennium BC (Bassiakos et al. 2018; Stos-Gale and Gale 1982). During the Iron Age (first millennium BC), there is evidence in Iberia for the production of silver from jarosite, which is not a lead ore but rather an iron-based mineral (KFe₃(SO₄)₂(OH)₆). The production process was similar to that of galena, yet required the addition of external lead for the cupellation process (e.g., Anguilano et al. 2010; Rovira and Hunt 2006). The introduction of this innovative process, as well as the complex organization required for mining the two ore types (galena and jarosite) and the smelting and mobilizing of the resultant lead across Iberia, is attributed to the Phoenicians (Hunt Ortiz 2003; Pérez Marcía 2013; Renzi et al. 2012; Rovira and Renzi 2013).

Silver Hoards in the Southern Levant

The use of silver as a standard of value and means of payment gradually spread from Mesopotamia to the Levant during the second millennium BC (Eshel et al. 2023), and from that point onward, played a significant role in the local economy, affecting long-distance trade strategies and political developments in the region (Eshel et al. 2019, 2021, 2022a, b, 2023). Forty silver hoards unearthed in various locations throughout the southern Levant comprise by far the largest corpus of pre-coinage silver currency known from the second and first millennia BC in any region around the Mediterranean. These hoards, dating ~1700/1650–600 BC (from the MBA III to the end of the Iron Age), indicate that silver was accumulated as wealth and means of payment in the region, at least for a millennium, until replaced by coins (e.g., Balmuth 2001; Eshel et al. 2018, 2019, 2021, 2022a, 2023; Heymans 2021; Kletter 2003; Thompson 2003).

There are no silver sources in the Levant, and there are several possibilities for its origin: Anatolia, which has silver-bearing ores in several regions including the Troad (Pernicka et al. 2003), the Black Sea region, and the Taurus Mountains, the latter being the nearest silver source to the Levant and extensively mined in antiquity (Sayre et al. 1992, 2001; Yener et al. 1991). Silver is also available in various regions in the Aegean, including the Laurion mines in Attica, mainland Greece, which was an important source of silver throughout antiquity (Gale and Stos-Gale 1981a; Stos-Gale and Gale 1982; Vaxevanopoulos et al. 2022). Additional ore sources include Romania (Baron et al. 2011), France (southern Massif Central; Baron et al. 2006; Orgeval et al. 2000), Sardinia (Begemann et al. 2001; Boni and Koeppel 1985; Stos-Gale et al. 1995; Valera et al. 2005), Iberia (Murillo-Barroso 2013; Stos-Gale 2001; Tornos and Chiarada 2004), Iran (Nezafati et al. 2008; Nezafati and Pernicka 2012; Pernicka et al. 2011, p. 638; Thornton 2014), and India (Craddock et al. 2013). However, in many of these regions, ores were probably not systematically exploited before the Roman period. The origin of silver can be traced based on the chemical and lead isotope compositions of the artifacts and by comparing the results to those obtained from ores; the latter were documented over the last 40 years and have been compiled into large databases used for provenancing archaeological artifacts (e.g., Oxford Archaeological Lead Isotope Database, also known as OXALID; also GlobalID, see Klein et al. 2022).

The many silver hoards in the southern Levant, therefore, offer a unique opportunity to trace the trade routes through which silver was supplied to the region and beyond for more than a millennium. The positive and research-oriented attitude of the Israel Antiquities Authority enabled the sampling of a large number of silver items. The combination of these two factors allowed us to conduct a series of provenance-based studies based on chemical and Pb-isotope compositions of 239 silver items obtained from 19 hoards excavated in the southern Levant (Eshel et al. 2019, 2021, 2022a, 2023). These results—reviewed and concluded for the first time in this paper—comprise the most comprehensive database of chemical and isotopic data of ancient silver ever compiled, in terms of numbers of sampled items and chronological scope. The results are also unique in the sense that each of the hoards was dated with high chronological resolution, enabling us to follow closely changes over time. In addition, we carefully address the question of mixed silver, both by identifying mixing lines on geological age models and by measuring detailed and accurate chemical compositions of items’ cores. Using this elaborate methodology, the longue durée perspective presented in this research allows us to demonstrate the close relationship between metal trade and political changes around the Mediterranean in antiquity.

Previous Studies of South Levantine Silver Hoards

Silver hoards from the southern Levant were first analyzed by Anna-Zofia Stos-Gale at the Isotrace Laboratory of the University of Oxford, for a project initiated by Miriam Balmuth. Sampling and partial publication of the results were performed by Christine Thompson (Stos-Gale 2001; Thompson 2007; Thompson and Skaggs 2013; OXALID). The study yielded 146 Pb-isotopic results from silver items of 14 of the southern Levantine silver hoards from ten sites: Tell el-ʿAjjul, Tell Keisan, Ashkelon, Beth Shean, ʿAkko, Tel Dor, ʿEin Hofez, Eshtemoa, Tel Miqne/ʿEkron, and ʿEin Gedi (see Figs. 1 and 2). These were obtained by Thermal-Ionization Mass Spectrometry (TIMS), which does not control analytical mass bias as efficiently as the Multiple-Collector Inductively Coupled Plasma Mass Spectrometry (MC-ICP-MS) and is therefore considered less accurate and prone to fractionation (Gentelli et al. 2021).

Fig. 1

Map of the southern Levant showing the location of the Bronze Age hoards mentioned in the text according to periods: Middle Bronze Age III (~1700/1650‒1600/1550 BC), Late Bronze Age I–II (~1600/1550–1200 BC), and Late Bronze age III (~1200–1150 BC). When the source of the silver is firmly established, it is mentioned in the figure. Map courtesy of Yoav Bornstein (illustrator)

Fig. 2

Map of the southern Levant showing the location of the Iron Age hoards mentioned in the text according to periods: Iron Age I (~1150–950 BC), Iron Age IIA (~950–800 BC), Iron Age IIB (~800–730 BC), and Iron Age IIC (~730–586/600 BC). When the source of the silver is firmly established, it is mentioned in the figure. Map courtesy of Yoav Bornstein (illustrator)

Recently, two other research groups, Wood et al. (2019, 2020) and Gentelli et al. (2021), suggested different interpretations for the isotopic results of some of the silver analyzed from the southern Levant. Below we point out the similarities and dissimilarities between these studies and ours and discuss them in light of the different methodologies applied by each group.

Chemical compositions of silver from the same southern Levantine Iron Age hoards were also obtained in the OXALID project, in additional studies (Kletter and De Groot 2007; Sass 2002; Shalev et al. 2014; Stos-Gale 2001), and in an initial study conducted by our research group (Eshel et al. 2018). In all of these studies, concentrations below 0.1 wt% were not reported, since analysis was performed using an Energy Dispersive X-ray Fluorescence (ED-XRF) analyzer, which measures surface compositions with a low detection limit (see Da Silva et al. 2023). Although only bulk chemical compositions of limited accuracy were reported in these studies, the results consistently showed that Iron Age hoarded silver typically contains Pb, Cu, and Au (Eshel et al. 2018; Kletter and De Groot 2007; OXALID; Sass 2002; Shalev et al. 2014; Stos-Gale 2001).

The Database and the Analytical Toolkit: Chemical and Pb-Isotopic Analysis of Silver

The 19 hoards sampled by our team (see Figs. 3 and 4) and their relative and absolute dates are listed in Table 1 (see also Figs. 1 and 2). The results also include ingots from the Iron Age IIC ʿUza fortress—though these were found in loose form, rather than as a hoard—and several silver items from an MBA III tomb in Megiddo (Eshel et al., in press). We carefully assigned most of the hoards to specific centuries and often even to subcentury time spans, as this was crucial for the identification of temporal trends in silver sources and trade networks during the Bronze and Iron Ages.

Fig. 3

A selection of Bronze Age and Iron Age I hoards analyzed in this study, ordered chronologically; the ceramic vessels are the original containers: a Abel Beth-Maacah jug and silver hoard as extracted, before cleaning and conservation (LB IIb). Image courtesy of the Tel Abel Beth Maacah Expedition by Gabi Laron (photographer); b–d three bundles from Beth Shean L. 88866 silver hoard (LB III). Images courtesy of the Beth Shean expedition; e Beth Shean L.1095 silver hoard (Iron Age I). Image courtesy of the Collection of Israel Antiquities Authority by Clara Amit; f Tell Keisan silver hoard before cleaning; (g) Tell Keisan silver hoard after cleaning (Iron Age I). Images courtesy of the Collection of Israel Antiquities Authority by Yael Yolovich (photographer).

Fig. 4

A selection of Iron Age II hoards analyzed in this study, ordered chronologically; the ceramic vessels are the original containers: a Dor silver hoard (Early Iron Age IIA). Image courtesy of the Collection of Israel Antiquities Authority and the Tel Dor Expedition; photo^© The Israel Museum by Ardon Bar-Hama (photographer); b ʿAkko silver hoard (Iron Age II). Image courtesy of the ʿAkko Expedition by Michael Eisenberg (photographer); c ʿEin Hofez silver hoard (Late Iron Age IIA). Image courtesy of the Collection of Israel Antiquities Authority by Warhaftig Venezian (photographer)

Table 1 List of hoards sampled for lead isotope analysis in previous studies (Eshel et al. 2019, 2021, 2022a, b, 2023), the relative chronology, absolute chronology, site, findspot, suggested source of metal, number of samples obtained from each hoard, and general chemical composition

Two-hundred thirty-nine silver items obtained from 19 hoards were sampled for chemical and Pb-isotopic compositions. The results were obtained through drilling, revealing the composition of the items’ core, rather than their surface, with high accuracy (the margin of error is within the range of a few ppm [parts per million]). Sampling permissions and current locations of the hoards are specified in Eshel (2020), appendix B. For a review of the methods, see Supplementary Material, with reference to our previous studies.

Chemical Composition

We have previously shown that trace element concentrations in silver items offer little help in identifying the source of the silver (see Supplementary Material). However, minor elements (Pb, Cu, Au, and Bi) that were not fully separated during the cupellation process can be used to identify and differentiate between idiosyncrasies resulting mainly from the production process and later biography. Significantly, Au and Cu are useful in identifying alloying, which may affect the isotopic results.

The general average and median of Cu, Au, Pb, and Bi through time, based on the chemical compositions of all 239 silver artifacts sampled, are displayed in Supplementary Material (Table SI, Fig. S1a–d). The results show a dramatic increase in Cu concentrations in silver items during the LBA III (~1200–1150 BC; Fig. S1a) and lower but still high concentrations during the Iron Age I (~1150–950 BC). Au average is uniquely high in the Early Iron Age IIA (~950 BC; Fig. S1b). Pb concentrations generally did not change with time, except for a significant increase during the Iron Age IIC (~680 BC; Fig. S1c). Bi concentrations are relatively high during the Iron Age IIB (~800–730 BC; Fig. S1d). Detailed chemical compositions are available in Eshel et al. (2019, 2021, 2022a, 2023), Additional explanations are presented in the Supplementary Material.

The results of the chemical analyses of the silver emphasize the importance of incorporating chemical data in the study of silver sources. Specifically, the deliberate addition of Cu, which may introduce lead into the alloy, affects the Pb-isotopic composition, and therefore identifying this addition is a perquisite for proper interpretation of Pb-isotope values (see more below).

Lead Isotope Analysis

Although lead isotope analysis has several limitations, we demonstrated that silver sources can often be positively traced (Eshel et al. 2019, 2021, 2022a, 2023). Correct provenancing depends, nonetheless, not only on high quality data but also on methodology, and therefore considering precise contexts and chronology, chemical composition, mixing and geological constraints, are all essential prerequisites to trace the sources of silver (see Supplementary Material). The results, presented for the eight subperiods examined, are available in the Supplementary Material. These show, first and foremost, that in each period the isotopic composition of the silver was different.

Discussion: Interpreting the Results from a Long-Term Perspective

The high-resolution chronological approach implemented in this study to a large number of silver hoards and items within them, combined with detailed chemical and isotopic analysis, affords a broad view of the development of silver supply to the southern Levant in the Bronze and Iron Ages. It also enabled us to contextualize our results historically—in some periods more clearly than in others.

Earliest Silver Hoards: Middle Bronze Age III

Contrary to the dominant paradigm, according to which silver hoarding in the southern Levant was essentially an Iron Age phenomenon (e.g., Heymans 2021; Thompson 2003, 2007, p. 161), there is unequivocal evidence for the use of silver as means of value and exchange and for the hoarding of silver, beginning in the Middle Bronze Age (Fig. 5a).

Fig. 5

Changes in silver sources over time: a Middle Bronze Age III–Late Bronze Age II (~1700/1650‒1200 BC, probably Anatolia and possibly also the Aegean); b Iron Age IIA early (~950–900 BC; Anatolia and Sardinia); c Iron Age IIA late–Iron Age IIB (~900–730 BC; Iberia); d Iron Age IIC (~730–600 BC; Greece). Maps courtesy of Svetlana Matskevich (illustrator)

The MBA III (~1700/1650–1600/1550 BC) in the southern Levant reveals direct cultural continuity from the MBA II (~1800/1750‒1700/1650 BC). Both periods are characterized by developed urbanism (inspired by earlier developments in Syria and Mesopotamia) and by a prosperous Canaanite culture (Greenberg 2019, pp. 180–182; Yasur-Landau 2019). Large fortifications, ramparts, and glacis, alongside intensifying use of bronze and cuneiform writing (mostly evident at Hazor), point to the existence of territorial units with centralized rule, with extensive economic and human resources at their disposal (e.g., Burke 2008; Finkelstein 1992; Horowitz et al. 2018). The use of silver as a means of exchange, a practice which originated in Mesopotamia, may have been introduced to the region from Anatolia and Syria. This is indicated by the typology of the items in the hoards and the source of the silver itself, which probably originated from Anatolia. Anatolian silver, as mentioned above, had been used as currency in Mesopotamia since the late third millennium BC.

Maritime trade between the Levant, Egypt, and Cyprus is well attested during the MBA II–III (e.g., Greenberg 2019, p. 219; Maguire 1995, p. 54; Marcus 2007; Samet 2017). Beyond this sphere, there are limited indications for Levantine maritime connections with Cilicia in Anatolia and Minoan Crete in the Aegean, throughout the Middle Bronze Age (Akar 2009; Cline et al. 2011; Cohen-Weinberger and Goren 2004; Kislev et al. 1993; Malamat 1998, pp. 34–35; Merrillees 2003; Niemeier and Niemeier 2000). The two silver items from the Shiloh hoard, consistent with the Laurion ores in Attica, provide therefore a rare attestation of commercial liaisons with the Aegean beyond Crete during the MBA.

Turning to the east, the prominent center of Mari on the Euphrates, which maintained terrestrial connections with Anatolia, also traded with Hazor, which was culturally part of the southern periphery of the Syrian (northern Amorite) realm (Maeir 2000; Malamat 1983, 1993; Marti and Chambon 2019; van Koppen 2007). However, other Middle Bronze Age cities in the southern Levant were mostly beyond Mari’s economic network, and south of Hazor there is no evidence for literate administrations (Samet 2017; Shai and Uziel 2010, pp. 72–74).

In Egypt, during the MBA II–III a Levantine population commonly known as the Hyksos gradually gained control over Lower Egypt and established their capital at Tell el-Dab‘a in the Nile’s Delta (ancient Avaris; 15th dynasty, ~1640–1532 BC; Ben-Tor 2009; Bietak 2010, 2013). The climax of this process is roughly parallel to the end of the Middle Bronze Age in the Levant (MBA III). Marcus (2006, 2019, 2022) repeatedly emphasized the significant involvement of the Hyksos in maritime trade and concomitant economic power, describing Tell el-Dab‘a as ‘Venice on the Nile’ (Marcus 2006) and suggesting that this was a major factor that enabled their political takeover. We cautiously suggest that the involvement and the rising power of the Canaanite Hyksos provide the context for the silver trade and its use as currency during the MBA II–III. However, the possibility that the trade in silver to the region during the Middle Bronze Age was part of the multifaceted Syrian influence on the southern Levant remains (Eshel n.d.).

Cypriot Involvement? Middle Bronze Age/Late Bronze Age Transition

Tell el-ʿAjjul on Canaan’s southern Mediterranean coast, a convenient anchorage, probably served as a port of entry into Canaan and a stopover on the way to Egypt during the MBA/Late Bronze Age (LBA) transition (~1600/1550–1500 BC) and LBA I, prior to its destruction (Bergoffen 2001; Fischer and Sadeq 2000; Steel 2002) (Fig. 5a). The Pb-isotope compositions of silver from Tell el-ʿAjjul differ from that of the MBA III items clearly demonstrating the introduction of silver from new sources to the Levant at this time (Eshel et al. 2023).

The reasons behind the shift in ore sources (possibly from Anatolia to the Aegean), which is indicated by the compositions of silver from Tell el-ʿAjjul, are unknown. We can only guess that they were related to sociopolitical developments within Anatolia. Exports of metals from Anatolia were apparently rare under the control of the Hittite empire (Lehner 2015). Thus, it is possible, for example, that the temporal fall of Cilicia/Kizzuwatna under Hittite control in the days of Ḫattusili I who reigned in the late 17th or early 16th century BC (Trameri 2020, pp. 137–140) affected the availability of Anatolian silver in the Levant. But the results are too few and too inconclusive to offer a historical reconstruction. Finally, since both Tell el-ʿAjjul and Mycenae were connected with Cyprus in this period, and given the privileged position of Tell el-‘Ajjul in trade with Cyprus during MBA III/LB I and LBA I (Bergoffen 2001; Fischer and Sadeq 2000), we suggest the possibility that Cyprus played an active role in this silver trade.

Under Egyptian Domination: Late Bronze Age I–II

The Late Bronze Age I–IIa in the southern Levant (15–14th centuries BC) was characterized by population decline following a series of destructions and abandonments of Middle Bronze Age urban centers. During this time, Egypt gradually gained control over the Levant, stationing garrisons in key Canaanite sites such as Jaffa and Beth Shean (Higginbotham 2000; Koch 2019).

During the LBA II (1400–1200 BC), known as the ‘The International Age’ across the Mediterranean, intense economic and diplomatic interactions linked the coastal sites of the Levant with Egypt, Cyprus, and the Aegean (e.g., Greenberg 2019, pp. 329–336; Stockhammer 2019; van Wijngaarden 2002, 2014). Direct contacts, for example, between Egyptian elites and Mycenaean palatial centers, possibly involved, inter alia, the exchange of Egyptian gold for Aegean silver (Gale 1980; Gale and Stos-Gale 1981b; Gill 2010; Kelder 2016; Stos-Gale and Gale 1982). The wealth and the expanse of this trade are best illustrated through finds from shipwrecks off the coast of Turkey at Cape Geledoniyah and Uluburun. The latter wreck, dating to ca. 1300 BC, is the richest find ever to record Bronze Age maritime shipping. The ship’s cargo included some ten tons of copper ingots, one ton of tin ingots, and almost 150 Canaanite storage jars (amphorae), alongside other prestigious finds (e.g., Bass et al. 1989; Pulak 2001). Growing Cypriot involvement in maritime trade around the Mediterranean is indicated, eventually reaching as far as Sardinia (Russel and Knapp 2017; Stos-Gale 2009; Yahalom-Mack et al. 2022) and possibly even Iberia (Broodbank 2013, p. 481).

The role of Anatolia within this network is increasingly gaining attention in scholarly literature, partly due to the recently confirmed attribution of Red Lustrous Wheel-made Ware to Anatolian workshops (e.g., Kibaroğlu et al. 2019). This pottery is ubiquitous in the eastern Mediterranean, especially in Cyprus. There are further indications for Anatolian involvement in LBA trade from the metal perspective. The use of lead from the Taurus Mountains has been identified in northern Levant (Ras Ibn Hani and Ras Shamra-Ugarit), in southern Levant off the Carmel coast (Hahotrim), in southeastern Cyprus (Hala Sultan Tekke), and in Egypt’s Delta (Qantir) (OXALID Database; Yagel and Ben-Yosef 2022; Yahalom-Mack et al. 2022). Tin ingots found on the Uluburun Shipwreck have been suggested to originate from the Taurus (Powell et al. 2021, 2022; Yener 2021; Yener et al. 2015; contra Berger et al. 2023). Anatolian silver could have easily been distributed to Mediterranean markets through lower (so-called Flat) Cilicia, which was one of the most important regions of contact between the Mediterranean and inland Anatolian distribution networks during the Late Bronze Age (Powell et al. 2021). Despite the clear involvement of the southern Levant in the LBA maritime trade, silver hoards during this period are scarce in the region, and only one small silver hoard, with silver ingots and scrap, is known from this period in the southern Levant, from Tel Abel Beth-Maacah (Yahalom-Mack et al. 2019). This is unexpected in view of the intensive exchange of commodities between the Aegean and the Levant especially during the LBA IIB. The near lack of silver hoards may be due to a general decline of silver use, due to Egyptian cultural influence, since in Egypt gold was preferred (Eshel n.d.; Golani 2013).

One item from the Abel Beth-Maacah hoard is consistent with Laurion ores, suggesting that some of the silver in the hoard may have been of Greek origin. The isotopic compositions of the remaining items indicated mixing and a general origin in the Anatolian–Aegean sphere. The silver differs in isotopic composition from the Anatolian MBA III silver, but is similar to some of the silver items from the MBA/LBA transitional phase and to unalloyed silver from a later hoard dating to the LBA III (see below). This preliminarily observation suggests, albeit based on limited data, that the sources of silver in the Levant did not change much throughout the Late Bronze Age.

The end of the Bronze Age around the Mediterranean ~1200 BC, known as the ‘Bronze Age Collapse,’ was marked by massive destructions and abandonments of palatial centers and the termination of associated administrative and economic apparatus. This brought to a sharp decline in Mediterranean trade to the Levant, reflected, for example by the near cessation of pottery importation (Broodbank 2013, pp. 585–590; Cline 2014; Gilboa 2022a; Killebrew 2014; Maran, 2009; Stockhammer 2019) and the near disappearance of Cypriot copper from the southern Levant (e.g., Yahalom-Mack et al. 2014).

Shortage of Silver: Late Bronze Age III–Iron Age I

The upheavals around the Mediterranean were bound to obstruct silver production and exchange. It seems that the collapse of the Hittite empire in the years flanking ~1200 BC also entailed the disruption of connections across the Taurus (Matessi and Lovejoy, in press), and the direct contacts between Egypt and the Aegean ceased (Cline 2014). Our results indicate a growing availability of hoarded silver after ~1200 BC, but of very low quality (debased with copper), suggesting a long period of shortage in silver. We have shown that the likely addition of south Levantine (Arabah) Cu most likely contributed Pb to the alloy, preventing a straightforward provenance of the silver (Eshel et al. 2021). As for Egypt, silver seems to have been available, used for example to produce silver coffins in the days of Psusennes I in the 11th century BC (Jurman 2015, pp. 54–55), but its source is unknown.

The recovery of maritime trade routes to the Levant was slow, gradual, and region-specific (Gilboa 2022a). The disappearance of the old centers of power forced societies in the southern Levant into a gradual transformation that generated new perceptions and social hierarchies (Gilboa 2014). Only during the Late Iron Age I (~1050–950), did the south Levantine settlements reach a new urban apex (e.g., Arie 2011, pp. 367–374, 472; Gilboa 2014, 2022a; Harrison 2004, pp. 107–108; Münger et al. 2011). It was accompanied by a significant intensification of trade, driven mostly by small-scale entrepreneurship (Bourogiannis 2013; Gilboa 2015; Sherratt and Sherratt 1991). The main instigators of these new trading initiatives seem to have been non-institutional Cypriots and Phoenician skilled merchants and seafarers (Gilboa 2015, 2022a; Sherratt and Sherratt 1991; Yasur-Landau, 2014, p. 122). Increasing Levantine maritime ventures remained, however, limited in their geographic extent, involving mainly the Levantine coast, Cyprus, and Egypt. The Aegean and Anatolia were probably not yet included in Levantine exchange spheres (see Gilboa 2022a; Gilboa et al. 2008, pp. 143–145; Kourou 2019; Maeir et al. 2009; Matessi and Lovejoy in press; Mazar and Kourou 2019). Mixing silver with copper continued during the Iron Age I, though copper percentages were lower than before. Because of this continued substantial mixing, the origin of the silver in these periods is presently undetectable.

New Horizons: Early Phoenician Silver Trade: Early Iron Age IIA

Long-distance maritime trade was resumed during this time, mostly through Levantine initiatives (e.g., Gilboa 2022a; Sherratt and Sherratt 1993) (Fig. 5b). Euboean pottery found in Levantine sites reflects ties with the Aegean (e.g., Gimatzidis 2020; Mazar and Kourou 2019; Stampolidis 2019, p. 500). Copper from the Arabah was widely distributed, found in the Levant, Egypt, and the Aegean (Ben-Dor Evian et al. 2021; Kiderlen et al. 2016; Vaelske and Bode 2019; Vaelske et al. 2019a, b; Yahalom-Mack et al. 2023). As for trade with the western Mediterranean, though Phoenician material remains found at Huelva in southern Iberia have been suggested to date to the mid-10th century BC (González de Canales et al. 2006, 2008; Mederos Martín 2006), these dates were questioned. The reasons are the problematic deposition of the finds, the unclear association between them and the radiocarbon-dated organics, and the lack of similar contemporaneous finds at any other Phoenician site in Iberia (e.g., Aubet-Semmler 2019, p. 75; Gilboa 2013, 2022b). Thus, the earliest unequivocal evidence for the buds of Phoenician commercial diasporas in several sites—both in Iberia and in other Mediterranean regions—does not antedate the late ninth century BC. This is the case not only for Iberia (Aubet 2008; Aubet-Semmler 2019; Gilboa 2013; Pappa 2013; Torres Ortiz 2008) but also for North Africa (Docter et al. 2008) and possibly for Sardinia (Rendeli 2018) and Kition in Cyprus (Iacovou 2014).

The Pb-isotope compositions of the silver dating to the Early Iron Age IIA (~950–900 BC) are consistent with galena from the Taurus Mountains, Anatolia, and with Pb ores in Iglesiente, Sardinia (Eshel et al. 2019, 2022a). This is the earliest clear evidence for the acquisition of silver beyond the Anatolian–Aegean sphere. Significantly, in this period, the quality of silver improved substantially, and its quantity in the southern Levant hoards increased (Eshel n.d.; Eshel et al. 2018, 2019).

The Sardinian silver found in the Southern Levant in Early Iron Age IIA contexts provides, therefore, the first evidence of silver trade with the western Mediterranean since the Late Bronze Age collapse. It originates from two sites on the northern coast of the southern Levant (ʿAkko and Tel Dor, 40 km apart), which in this period was part of the Phoenician economic sphere (Gilboa 2022a; Killebrew 2019; Lehmann 2021). We have suggested that the results embody Phoenician so called pre-colonization activities in Sardinia and outlined a geographic-temporal trajectory for Phoenician far-flung quest for silver. According to this reconstruction, Phoenicians brought silver to the Levant from southwest Sardinia ∼100 years before any de-facto Phoenician permanent holding in the western Mediterranean, and at least ~50 years before Phoenicians reached Iberia (see below). The results also suggest that Phoenicians were involved in silver production in Anatolia (Eshel et al. 2019).

The chronology that we have secured for the beginning of the Phoenician quest for silver in the west lays to rest once and for all previous suggestions to understanding this quest as a response to the Neo-Assyrian takeover of the Levant (Frankenstein 1979), as clearly the Phoenician endeavors predate the Assyrian control of the Levant, which started in the late ninth/eighth centuries BC (and see further below). It is also evident that other political entities that were suggested to have driven the Phoenicians westward, such as Aram Damascus in the ninth century BC (Fantalkin 2006), cannot be considered as initial catalizators.

Phoenicians in Iberia: Late Iron Age IIA and Iron Age IIB

Our results show that in the course of the Late Iron Age IIA (~900–800 BC), a shift in the sources that provided silver to the southern Levant occurred, from Sardinia and Anatolia to Iberia (Linares, Gador, Almeria), alongside small quantities of silver, which is consistent with the Taurus region in Anatolia (Eshel et al. 2022a) (Fig. 5c). The large hoards from ʿEin Hofez (~1.2 kg) and Eshtemoa (~23 kg) are evidence of the abundance of silver that continued to characterize the Levant during these periods.

Silver production in Iberia required the addition of lead to the Pb-deficient argentiferous jarosite ores in Pyrite Belt near Río Tinto, from Iberian lead ores (e.g., Kassianidou et al. 1995; Murillo-Barroso et al. 2016). Consequently, the isotopic compositions of Iberian silver points mainly to the source of the added lead. Silver attested in the southern Levant was produced primarily with lead from Linares in the Pyrite Belt—in the ʿEin Hofez and Eshtemoa hoards, and from Gador in Almeria in the ʿArad hoard (Eshel et al. 2019, 2022a). The Phoenicians were the first to apply this process in Iberia, which required the mobilization of Pb ores through riverine and maritime routes. It remains unclear whether adding external Pb for silver production was a Phoenician innovation or whether it was practiced earlier and if so, where.

It is generally accepted that the quest for metals, especially silver, was a crucial instigator of Phoenician endeavors to the west (Aubet-Semmler 2019; López-Ruiz 2022; Markoe 2005; Sherratt and Sherratt 1993). Scholars, however, debated whether this was the sole or first stimulus (for agricultural land and produce, see, e.g., Aubet 2001; Beitzel 2010; Dietler 2009; Wagner and Alvar 2003; for a total dismissal of metals as motivators, see Fletcher 2012). Our results show that silver production in Iberia started already in the ninth century BC and thus was a major goal from the outset of Phoenician settlement in Iberia (Eshel et al. 2019) and in the western Mediterranean in general (e.g., Docter et al. 2008; Gilboa 2013; Pappa 2013, pp. 6–8; Torres Ortiz 2008). The earliest acquaintance of Phoenicians with Iberian metal sources probably occurred in Sardinia, which held close commercial connections with Iberia. The excavations of the late ninth century BC market square of St'Imbenia in northwest Sardinia revealed Iberian amphorae alongside lead (Pb) from the Iberian Peninsula and material evidence of interconnection with Aegean and Levantine seafarers (Clemenza et al. 2021; Rendeli 2018) suggesting Sardinia as an early meeting point between these cultures.

Notwithstanding the newly discovered Iberian silver sources, the all but disappearance of Anatolian silver from the south Levantine markets is intriguing. Mines in the Anatolian highlands supplied the Levant for millennia and remained the closest and arguably the most accessible sources. It is difficult to understand this cessation since we do not know when exactly in the course of the ninth century BC this change occurred, but we propose that it was likely a result of Assyrian activities in Cilicia. Silver reaching the Levant from Anatolia was perforce marketed thorough Flat Cilicia. Assyrian aggressive intervention in this region began with Shalmaneser III, who during several campaigns, starting ca. 839 BC, took control of the plain of Cilicia, possibly also part of the Taurus Mountain ridge, conquered and destroyed cities, and subjugated and taxed local rulers. Famously, inscriptions record that the king climbed Mt. Tunni, “The Mountain of silver,” unanimously identified in the Bolkardağ mining ridge (Weeden 2010, pp. 39, 44; Yamada 2000, pp. 199–220, n. 469 on p. 213). Shalmaneser III’s tribute lists mention silver very frequently, in many inscriptions as the first item before gold, all this attesting to its value in Assyrian eyes. It is highly likely that the Assyrians channeled the Bolkardağ silver to suite their interests, mainly to heartland Assyria, and thus it disappeared from the south Levantine market.

In tandem, the incomparable yield of the Iberian ores, once the outstandingly efficient Iberian exploitation systems were established, rendered other sources redundant. The significantly poorer sources in Iglesiente, Sardinia (Atzeni et al. 2005; Pearce 2018), were abandoned as well.

During the second half of the ninth and especially the eighth century BC (Late Iron Age IIA and Iron Age IIB), several kings of Assyria also initiated campaigns to the Levant and gradually annexed significant parts of it (e.g., Bagg 2011; Berlejung 2012; Faust 2021, pp. 60–72; Radner at al. 2023). In the southern Levant, the Assyrians conquered and destroyed the Kingdom of Israel in the late eight century BC, while other political entities succumbed to vassalage, such as the kingdom of Judah and the Philistine cities (e.g., Faust 2021). The Phoenician cities in Lebanon (the northern Levant) were usually spared from destruction, even when they rebelled. Scholars agree that among the reasons for this leniency was the Assyrian reliance on the fleets of the Phoenician city-states and their supraregional naval contacts and skills. The empire’s great kings and other elite benefited from the precious commodities brought by the Phoenicians from Egypt, Cilicia, and the western Mediterranean, which they taxed heavily (e.g., Aubet 2001, pp. 85‒95; Bagg 2011, p. 390, table 3E; Elayi 2018, pp. 129‒154; Fales 2017, p. 247; Fantalkin 2018; Parpóla 2003, p. 102; Radner 2004, pp. 155‒157; Sader 2019, pp. 131‒135, 252‒255; Sherratt and Sherratt 1993; Sommer 2004; Younger 2015, pp. 189‒198). The results presented here confirm this historical reconstruction, showing that the Phoenician market was dominated by Iberian silver for more than a century, and that Iberia was also the main silver ore source for Judah and Philistia where silver hoards dating to the Iron Age IIB were found.

East Greek Trade during the Iron Age IIC

The political situation in the southern Levant changed dramatically again when the Assyrians retreated from Western Asia ca. 640–620 BC to deal with threats in the east, leaving behind a political and administrative void (Faust 2012, 2018, pp. 48–49; Gilboa and Sharon 2016, p. 248). For a short time, the Saitic Egyptians (the 26th Egyptian dynasty) took advantage of the situation, attempting to regain power in the Levant. They filled the military and administrative vacuum and established their rule over part of the former Assyrian territories, especially along the Mediterranean coast of the Levant (e.g., Agut-Labordère 2013; Čapek and Lipschits 2019, pp. 29‒30, n. 56; Schipper 2011). In the southern Levant, new commercial circuits are in evidence, involving several coastal and near-coastal sites that survived the Assyrians, mainly in Philistia and the ʿAkko Plain in southern Phoenicia (Fig. 5d). These commercial spheres are indexed by conspicuous quantities of so-called East Greek (mainly south Ionian) pottery and to a lesser extent Corinthian ceramics and Cypriot pottery. The establishment of the Greek, including East Greek trading ‘colonies’ in Egypt during this period, most famously Naukratis on the Canopic (west) branch of the Nile (Fantalkin 2014; Schlotzhauer and Villing 2006; Villing and Schlotzhauer 2006), is the best-known aspect of this new Greek-Egyptian-Levantine (and most probably also Cypriot) exchange sphere. It enabled, inter alia, the arrival of Greek silver to the southern Levant for the first time since the Bronze Age (Eshel et al. 2022a).

The Pb-isotope ratios of Iron Age IIC (~630–600 BC) silver are mostly consistent with the Laurion mines in Attica and the island of Siphnos in the Cyclades, alongside a few items of silver that show consistency with Iberian lead ores (Eshel et al. 2022a). The shift from Iberia to Laurion, which we have shown to have occurred specifically in the second half of the seventh century BC, seems to be directly related to the Assyrian retreat (see above), following which the Phoenicians seem to have lost their privileged position as providers of silver to the Neo-Assyrian empire. The market opened then to new agents, especially East Greek traders, who imported silver from Laurion. This shift probably had a deep effect on Phoenician activities in the western Mediterranean and may have been one of the factors that eventually, in the sixth century BC, contributed to its detachment from the homeland in the Levant (Eshel et al. 2022a).

All this ended with the Babylonian takeover of the Levant, involving large-scale destructions in the southern Levant between ~604 and 586 BC (Kahn 2006; Schipper 2011), marking the end of the Iron Age. No data pertaining to silver exists for the economically and demographically depleted southern Levant for most of the sixth century BC. When eastern Mediterranean trade revived in the early Persian period (the early fifth century BC), extensive contacts were resumed between the southern Levant with various parts of the Aegean and Anatolia (Gilboa et al. 2017; Lehmann et al. 2019; Martin and Shalev 2022; Stewart and Martin 2005). This period, however, is outside our scope here.

Same Pb-Isotope Results, Different Interpretations

The unique assemblage of south Levantine silver hoards drew much scholarly attention in recent years, yet the analyses of items from the same hoards by other scholars yielded different interpretations. While we suggest that the sources of silver that reached the Levant constantly changed over time, Wood et al. (2019) claimed that Anatolia and Iberia (or possibly Sardinia, which they could not differentiate isotopically) provided silver to the Levant throughout the Iron Age, from ~1200 BC and continuously until the seventh century BC. Gentelli et al. (2021) claimed that only a very small portion of the silver is consistent with Anatolian or Iberian ores, and that only one hoard (Beth Shean 1095, 11th century BC) contained a significant portion of silver consistent with Sardinia. They suggested that the Aegean and central European silver sources dominated the trade of silver to the southern Levant from the end of the Late Bronze Age throughout the Iron Age. As such a variety of opinions stems from the analysis of the same hoards, we are concerned that despite all the intense research, an important phenomenon in the history of the Mediterranean remains vague. Therefore, we outline below the differences in methodologies, which, we believe, led to such diverse conclusions.

Wood et al. (2019) based their study on the OXALID results from samples obtained by Thompson (2007) and analyzed by the Oxford Archaeological Lead Isotope Laboratory. Wood et al.’s (2019, fig. 6) main methodology involved a calculation of the Pb-model crustal age of the metal versus the compositional Au/Ag ratios. Where the gold to silver ratio did not match the crustal age, they proposed that mixing of silver from different sources occurred, identifying linear mixing lines. We, however, take issue in this method (see also Gentelli et al. 2021), for several reasons.

First, since gold (Au) was occasionally added to silver through remelting and alloying (Eshel et al. 2018), it cannot be used automatically to identify silver sources, as opposed to Wood et al.’s basic assumption. In addition, both the lead isotope ratios and chemical compositions available on OXALID, on which Wood et al. (2019) based their study, are of lower quality than the results obtained in the present study (see above). Specifically, Au concentrations were measured by XRF, which has a low detection limit. Moreover, comparing Au concentrations and Pb-isotope ratios is expected to produce curved rather than linear mixing lines (e.g., Albarède et al. 2020; Eshel et al. 2021, fig. 8), since each ore contributes different Au concentrations.

In addition, the Pb-crustal age model in Wood et al. 2019 (presented as age histograms) records the decay of ²³⁵U to ²⁰⁷Pb, and ²³⁸U to ²⁰⁶Pb, and thus uses only two isotopic ratios (²⁰⁶Pb/²⁰⁴Pb, and ²⁰⁷Pb/²⁰⁴Pb), while ignoring the Th-Pb system represented by ²⁰⁸Pb/²⁰⁴Pb. The latter ratios are indeed presented (Wood et al. 2019, fig. 5), yet they are not incorporated in the Pb crustal age plots, which, as a result, do not capture the full LIA picture. Had the authors used all three ratios, they would be able to distinguish between some Anatolian and Greek lead ores, since despite the overlap of these ores on the ²⁰⁷Pb/²⁰⁴Pb versus ²⁰⁶Pb/²⁰⁴Pb plot, they still differ isotopically. This is well evidenced by the low ²⁰⁸Pb/²⁰⁴Pb ratios of Laurion (Greece) ores relative to Anatolian ones (e.g., Stos-Gale and Gale 1982; Yener et al. 1991). Wood et al. (2019) also grouped together different ores within large geographical areas, e.g., Sardinia and Iberia. They treated all of the ores as one entity, despite the significant variability in their isotopic compositions (e.g., Iglesiente versus Sulcis in Sardinia). Due to this generalization, they could not distinguish isotopically between individual ores in Sardinia and Iberia (Wood et al. 2019, fig. 8), while in fact, the Iglesiente ores in Sardinia clearly have distinguishable isotopic values (²⁰⁶Pb/²⁰⁴Pb, ²⁰⁷Pb/²⁰⁴Pb, and ²⁰⁸Pb/²⁰⁴Pb) in relation to, e.g., the Linares ores in Iberia.

Moreover, Wood et al. (2019) dated some of the hoards erroneously (see Eshel et al. 2022a, footnote 3) and generally did not consider the chronology of the south Levantine hoards. Instead, they plotted the results from all the hoards together. As a result, they considered together in one graph, hoards that span a millennium, from the 16th to the sixth century BC, during which major cultural, economic, and geopolitical changes occurred, as described above. They also overlooked the possibility that LI compositions of most of the Early Iron Age silver in the southern Levant were affected by the addition of copper from local Arabah mines (Timna and/or Faynan). Disregarding this extraneous lead contribution led to incorrect interpretations, for example, suggesting that the silver in the 11th century BC Tell Keisan hoard originated from Iberia. In fact, nearly 50 items from four Iron I Age hoards (Beth Shean 1095, Tell Keisan, Megiddo 2012, and Ashkelon) contained silver mixed with copper from the Arabah (Eshel et al. 2021). Only one item from the late Iron Age I Ashkelon hoard might have been produced from Sardinian silver, and three additional items from the same hoard may have been mixed with western silver. Therefore, even if ‘western silver’ was indeed present in these hoards as suspected by Wood and his colleagues, it cannot be identified using lead isotope analysis.

Moreover, Wood et al. (2019) also claimed to be able to distinguish between native silver and silver produced by cupellation, based on the former’s low Au and Pb concentrations. Indeed, native silver is poor in Pb (<0.05 wt% and often undetected; Boellinghaus et al. 2018; Murillo-Barroso et al. 2014, table 1; Pernicka 2013, table 6), however, their claim could not be validated, since it is based on XRF surface analysis of Au, Pb, and Cu, which has a poor detection limit (0.2 wt%). In fact, Pb concentrations in south Levantine silver consistently range between 0.05 and 3 wt% (Eshel et al. 2019, 2021, 2022a, 2023), indicating that all the silver was most likely the product of cupellation and not of native silver.

Finally, Wood et al. (2020) further applied their proposed methodology to a reanalysis of lead isotope data from OXALID to determine the source(s) of the Tel Dor hacksilber hoard. The authors concluded that silver previously identified by them as having originated from Thera or Kythnos in the Aegean (Wood et al. 2019) was in fact mined at Kalavasos in south Cyprus. Therefore, considering the manifold ties between Phoenician Dor and Cyprus, they claimed that Phoenicians learned silver production in Cyprus. These conclusions are unsupported, since the presence of both Ag and Pb ores in Cyprus is highly unlikely (Gentelli et al. 2021; Kassianidou 2013, p. 50).

Gentelli et al. (2021) presented a new statistical approach to interpret Pb-isotope analysis of silver. They analyzed 45 additional pieces of hacksilber from five hoards (Megiddo Area H, Eshtemoa, Tel Dor, ʿEin Gedi, Tel Miqne-Ekron) measured by MC-ICP-MS, which handles fractionation successfully, and combined their data with the results obtained by our group (Eshel et al. 2019).

Gentelli et al. (2021) relied only on high-quality ore databases and refrained from using isotopic data of ores suspected to be imprecise. They also considered each hoard and subperiod separately. Therefore, they have a solid basis for a high-resolution chronological study. However, within each hoard, the Pb-isotopic results were treated using a clustering method to identify statistically distinct groups of data named ‘convex hulls,’ and to identify and constrain ore sources consistent with the isotopic signature of each group. This method has been applied for each hoard separately. According to the authors, they grouped the samples of each hoard into several ‘convex hull’ groups, separated by minimal variance within isotopic clusters and maximal variance between isotopic clusters. A ‘hit’ was defined when a result obtained from an ore was consistent with a ‘convex hull’ representing a statistically distinct group of results within a sampled hoard.

The clustering of each entity, however, was based on assigning the same significance to all the results within each hoard, and thus it does not differentiate between end members and mixing lines. Therefore, the more mixed silver is being analyzed, the more the analysis will be prone to errors. In addition, since Gentelli et al.’s (2021) analysis is based on statistics, the larger the ore database, the more ‘hits’ will be found between ‘convex hulls’ and ores. The size of the ore datasets therefore has a direct effect on the identification of provenance. For example, Gentelli and colleagues proposed that the Tel Dor hoard contains mostly silver originating from the southern Massif Central ores in France (Cévennes). Isotopically, in fact, these ores fall on the mixing line between Iglesiente in Sardinia and Taurus in Anatolia, and therefore the results that plot with the Cévennes ores are more likely a result of mixing silver from these two origins.

Finally, Gentelli et al. (2021) did not measure chemical compositions; they also disregarded the chemical data we published. They do not accept the possibility that when silver was mixed with copper, the copper contributed Pb to the alloy. They argue that Pb is not soluble in copper, yet overlook the fact that copper artifacts and ingots from the Arabah contain 1–5% Pb (Hauptmann 2007, p. 201). This can be explained by the fact that Pb, which indeed is insoluble in copper, resides along grain boundaries or in inclusions (Ben-Yosef et al. 2016; Yahalom-Mack et al. 2014, fig. 9c), yet once remelted it is expected to contribute Pb to the Ag–Cu alloy. Thus, their provenance determinations of Ag–Cu alloys (common in the Iron Age I, see above) should not be accepted.

The main conclusion of Gentelli et al. (2021) was that Aegean sources, including Laurion, Macedonia, and Thrace, alongside Cévennes in France, were major sources of silver to the southern Levant for a long time, since the end of the Late Bronze Age through the Iron Age. Based on their conclusions—that Anatolia and Iberia did not supply silver to the Levant in the Iron Age (and Sardinia only minimally), we suspect that some of the ores from these regions were underrepresented in their ore database vis-à-vis Aegean and central European ores. In all, we do not accept the interpretation of the results by Gentelli and colleagues and conclude, as outlined above, that in the Iron Age, the role of Laurion in silver supply to the Levant was significant only during the Late Iron Age IIC, in the late seventh century BC. Aegean ores may have contributed silver to the Levant during the Bronze Age, but this cannot be presently confirmed based on the data at hand.

We reiterate that out results are still restricted by the limitations of Pb-isotopic analysis, some of which we attempted to overcome by using a large dataset and by incorporating geological considerations into the interpretation, as described above. It is possible, for example, that unknown ores will be revealed in the future and, consequently, alter some of the conclusions presented here.

Conclusions: Analytical and Historical Implications

In a series of studies, we followed with high chronological resolution the sources that supplied silver to the Levant over a period of more than 1000 years. The results reflect rapid fluctuations in silver ore sources, which often changed completely rather abruptly.

Silver sources of the Bronze Age can be traced to Anatolia and the Aegean in general, regions that are closer to the Levant relative to the Iron Age sources. However, the specific sources are unknown. Only three Bronze Age silver items (two from Shiloh and one from Abel Beth-Maacah; Fig. S2b) are consistent with Laurion. During the Late Bronze Age III and Iron Age I, extensive mixing with Levantine copper hinders the provenancing of the silver. The sources of silver from the mid-10th century (Early Iron Age IIA) and onward are more clearly identified. Large-scale Phoenician silver exploitation can be demonstrated at this time in the Taurus Mountains in Anatolia and in south Sardinia, and from the ninth century BC (Late Iron Age IIA and Iron Age IIB) in Iberia. Chemical analysis suggests that the Phoenicians introduced innovative silver production methods, but the precise process and the way these innovations were introduced, are unknown. Yet the results emphasize that western Mediterranean silver production (in Sardinia and Iberia) is a clear Iron Age phenomenon, unknown in the Bronze Age. Finally, during the seventh century BC (Iron Age IIC), Laurion became the main source of silver.

The lack of silver sources in the Levant meant that silver supply depended heavily on long-distance trade. Indeed, first and foremost, the chronological trajectory reveals that each period stands alone in terms of the silver sources and the concomitant trade routes. Thus, for example, while large quantities of silver from Anatolia are evident in the southern Levant during the Early Iron Age IIA (in the hoards from Dor, Beth Shean, and ʿAkko), a hoard from the Late Iron Age IIA (ʿEin Hofez) contains almost exclusively silver from Iberia (n=27/29) and scarcely any from the Taurus, Anatolia (n=2/29; Eshel et al. 2019). The two artifacts that are consistent with the Taurus may suggest that silver continued to arrive from this region during the Late Iron Age IIA on a limited scale, or, alternatively, that small quantities of Anatolian Early Iron Age IIA silver continued to circulate. Therefore, the results strongly suggest that in contrast to other metals and items such as gold artifacts or glyptics, silver was not accumulated over long periods, a global phenomenon (see ‘silver losses’ in Patterson 1972). This created a constant need for fresh silver supply. Therefore, when high-resolution chronology is employed for investigating sources of silver, the results become a good index for the silver circulating in the Levantine market during each period, though in reality, changes in ore sources were probably gradual.

This constant need for silver was a push factor for endeavors to search for new metal sources and an instigator of long-distance trade. Thus, we recognize the shortage of silver in the Levant during the Iron Age I as the instigator for the Phoenician voyages to Anatolia and the western Mediterranean during the Iron Age IIA, around the mid-10th century BC (Eshel et al. 2019, 2021; see above). Later on, during the eighth century BC, silver procurement is probably one of the main reasons for the Assyrians granting the Phoenicians permission to sail freely, while taking assertive measures to block Greek mercantile activities (Fantalkin 2006, p. 201). This is evidenced both from Assyrian texts and pottery distribution in the Levant: Assyrian kings, starting with Tiglath-pileser III in the late eight century BC, were fighting Ionian ‘pirates’ along the shores of the Levant and Cilicia, and regulated maritime traffic along the Levantine shores (Radner and Vacek 2020). As well, in the eighth century BC, Greek pottery in large parts of the Levant reached a very low ebb (Fantalkin 2006; for the exception of al-Mina in Syria, see Radner and Vacek 2020). When Assyria lost its hold over the southern Levant ca. 630 BC, there was nothing anymore to hinder the establishment of Ionian (East Greek) commercial and other contacts with the Levant, enabling silver trade from Laurion (Eshel et al. 2022a) and East Greek activity in Naukratis and other sites in Egypt at this time (Fantalkin 2014). These changes in trade patterns probably affected the entire balance of trade, particularly with the western Mediterranean (Iberia).

Looking at the broader implications of our results, we start with the analytical ones. The method of lead isotope analysis for the provenance of metals has several limitations and thus was accepted by some metallurgists as reliable only when used to exclude ores rather than to correlate artifacts with specific ones (e.g., Knapp 2000; Pollard 2009). We have, however, demonstrated in the abovementioned publications that silver sources can often be positively traced (Eshel et al. 2019, 2021, 2022a, 2023). Since different scholars reached contradicting conclusions regarding the sources of Levantine silver, we have demonstrated above that, for correct provenancing, high-quality data are not enough. The methodology is important, as is the consideration of precise contexts and chronology, chemical composition, mixing, and geological constraints. All these are essential factors that are prerequisites for the identification of ore sources and developments in the supply of silver, in our case to the southern Levant.

Our results are germane for a plethora of phenomena in the study of the history and archaeology of the Mediterranean in the timeframe investigated, even though we do not claim that the developments observed in the southern Levant can be projected to other eastern Mediterranean regions at this point (see below). Some of these issues have already been alluded to above, including trade circuits in the eastern Mediterranean and beyond during the Middle Bronze Age, especially towards the end of the period (Akar 2009; Marcus 2006; Yener 2007); the way(s) the LBA collapse affected Mediterranean trade with the Levant (Gilboa 2022a as opposed to Bell 2016); the vicissitudes of south Levantine contacts with the Aegean including the phenomenon of the relative ubiquity of Euboean ceramics in the Levant during Iron Age IIA, which now seems unconnected to any silver supply from the Aegean; the nature and chronology of Phoenician activity in the silver regions around the Mediterranean; Phoenician ‘expansion’ to the west and its effects on several economic, demographic, and cultural processes in these lands, especially in relation to Cilicia (Lehmann 2008), Sardinia (Bernardini 2008, 2018–2019; Terpstra 2021; Tronchetti 2015), and Iberia (Botto 2018; Campos and Alvar 2013; Murillo-Barroso et al. 2016; Pappa 2013); the meaning of the ubiquitous Ionian ceramics in the Levant and in other Mediterranean regions in the seventh century BC (for Iberia, see Aubet 2007); and—a related issue—the reshaping of Mediterranean connectivity after the fall of Assyria (Vacek 2017; Villing and Schlotzhauer 2006). Ironically, specifically for the Late Bronze ‘International’ Age, we hardly contribute new insights, as explained above. Beyond the phenomena listed above, we submit that our results will probably be relevant for any discussion of pan-Mediterranean developments in the time stretch investigated, especially for the Iron Age (Hodos 2020). The results also contribute to the discussion regarding the identification of the still elusive biblical ‘Tarshish,’ which was the destiny of joint ventures of Hiram King of Tyre and King Solomon bringing back silver among other commodities (e.g., Beitzel 2010).

From a more theoretical perspective, our results are significant for ongoing debates regarding the very nature of Mediterranean connectivity. On the one hand, nodes of contact related to silver trade were of course primarily dictated by the not-too-many locations of silver ores around the Mediterranean. On the other hand, we clearly show that these geographical circumstances, including factors such as distance from the consumers or ease of sailing (structure, as formulated by the Annales School of historical thought) cannot be considered the only, even not the main aspects to be considered. Such a deterministic approach has most pointedly been formulated for the Mediterranean in Peregrine Horden and Nicholas Purcell’s Braudelian and extremely influential hypothesis (2000, e.g., pp. 348–350, cf. Broodbank 2013). Our results are not compatible with such views. Rather, silver sources, trade routes, distribution nodes on the receiving end, and agents constantly and sometimes rapidly changed, to a large extent following demographic and political transformations (elements of conjoncture and histoire événementielle; see in the same vain Concannon and Mazurek 2016; Gilboa 2022a; Panagiotopoulos 2015; cf. Whittow 2001). Consequently, more often than not, all-encompassing models for commercial connectivity in the Mediterranean in the historical periods are doomed to fail (see Murray 2023).

Future Prospects

Our research concentrated on a single subarea of the Levant. As we hope to have demonstrated above, the question of silver trade should be investigated from a nuanced chronological, as well as regional frame of reference. Naturally, additional systematic sampling in adjunct regions as the northern Levant, Cyprus, and Egypt is needed to form a more comprehensive view of silver trade in the Mediterranean during the Bronze and Iron Ages.

The use of silver as currency, which is closely linked to that of the sources of silver, is discussed in a forthcoming paper (Eshel n.d.), but a few words are in order here. Textual evidence for the earliest use of silver as currency in the southern Levant is limited to three MBA Akkadian tablets from Hazor in northern Canaan, mentioning payments in silver shekels (numbered Hazor 5, Hazor 7, and Hazor 18 in Horowitz et al. 2018). However, all three texts were found out of context and therefore could not be accurately dated within the MBA (Eshel et al. 2023). Therefore, the silver hoards in the southern Levant currently provide the best dating proxy for tracing the beginning of the use of silver as currency there, to the MB II–III. Investigating the use of silver requires attention to issues such as the weight and content of the silver hoards, the related weighing systems (e.g., Kletter 2003; Rahmstorf 2007), the depositional nature of the hoards, the forms of the ingots, and more (see Eshel et al. 2018, 2019, 2021, 2022a, 2023). These are discussed in the abovementioned paper (Eshel n.d.), suggesting that the use of silver as currency, invented in mid-third millennium Mesopotamia, gradually spread throughout the ancient Near East and Europe. Silver money is suggested to have developed as a mercantile phenomenon rather than initiated by city-state institutions. Merchants relied on silver to acquire and distribute commodities over long distances, as demonstrated during the late third and early second millennia in the exchange between Anatolia and Assyria (e.g., Barjamovic et al. 2012; Yener 2015 and above). The use of silver as currency in the Near East is one of the reasons for the growing need for silver and the constant development of trade routes in search for it.