New evidence for long-distance trade in arsenical copper during the Early Bronze Age in the southern Levant: analysis of weapons from the Nesher-Ramla cemetery

Abstract

The paper presents new data regarding trade in metals during Early Bronze IB. Using chemical and lead isotope analysis of weapons from Early Bronze Age IB burials from the Nesher Ramla Quarry, located in the Shephelah (piedmont) bordering the Judean foothills, it is shown that complex metals were likely procured from eastern Anatolia. These data join similar analytic results regarding several artefacts from the Kfar Monash hoard and evidence from Tell es-Shuna, and it is suggested that metal trade might be considered as a possible conduit for the transfer of cultural ideas and modes of social organization on the eve of southern Levantine urbanism.

Keywords:

• Early Bronze Age
• southern Levant
• daggers
• lead isotope analysis
• urbanism

Introduction and background

Urbanization in the southern Levant during the late 4th millennium BCE is generally viewed as a diffusionist process (Greenberg 2002: 3). The evidence for possible influences from the two dominant urbanized civilizations of the period — Egypt and Mesopotamia — was recently summarized by Greenberg (2011) who pointed to the difficulties involved in correlating the urbanization of the southern Levant to the Egyptian sphere. While generally accepting Mesopotamia as the source, he notes a lack of ‘ … material evidence for direct contacts between the southern Levant and the Mesopotamian core: no imports or exports, nor any undisputed components of the Uruk toolkit, as exhibited in other sites of the Uruk expansion’. (Greenberg 2011: 236; see also Algaze 1993: 85–109). However, contacts between the southern Levant and the upper Euphrates region in the sphere of metallurgy have been recognized, emerging from the results of a metallurgical study of the EB IB Monash hoard (Hauptmann 2007; Hauptmann et al. 2011). These results pointed to the possibility of considering inter-regional urban transmission via the trade in metals. In this paper, additional evidence for such trade, as revealed via the chemical and isotopic analyses of weapons from Early Bronze IB (EB IB) burials from the Nesher-Ramla Quarry, is discussed, supporting the possibility that such contacts may have played a role in the transmission of cultural concepts at the dawn of urbanism in the southern Levant.

Context and finds

Nesher-Ramla Quarry EB I necropolis

The archaeological site at the Nesher-Ramla Quarry (henceforth NRQ) is located in the Lod Valley in the Shephelah (piedmont), bordering the Judean foothills, 5 km south-east of the modern cities of Lod and Ramla (Fig. 1).¹ The site extends over the slopes of two adjacent hillsides, which together form a low-lying, crescent-shaped area ranging between 110–125 m asl. Archaeological surveys and salvage excavations, conducted prior to the quarrying operations, yielded evidence of activity ranging from the Upper Paleolithic until the Mamluk period (Avrutis 2012: table 1.1).² The assemblage of copper-based weapons at the focus of the present paper was collected from nine natural, karstic cavities that were slightly modified and used as tombs, and 18 burial pits from the late phases of the EB I, dating to around the end of the 4th millennium BCE (Regev et al. 2012). In all but one case, the EB I burials reused earlier Late Chalcolithic (LC) caves for both interment and domestic purposes (Avrutis 2012: 7–32; 2018; 2020). Such cave tombs, hewn in the bedrock and serving as repositories for multiple burials over time are common to the southern Levant, and particularly, the central part of Israel (e.g., Amiran 1985; Ben-Tor 1975; Macalister 1912: pl. XIII).

Figure 1 Map of the southern Levant with the location of NRQ and additional sites mentioned in the text.

The interred individuals from the EB IB ne­cropolis at NRQ were buried in primary deposition and arranged on the floors of the caves, around the wall perimeters. Four adults, possibly of high social status, were laid out on stone-built pavements in three different burial caves. The deceased were accompa­nied by a variety of funerary offerings, including pottery, grinding stones, stone vessels, flint tools and metal weapons (Fig. 2). The quantity and variety of the offerings varied markedly from cave to cave and from interment to interment (Avrutis 2012: 121, table 4.4.1). Among the numerous local southern Levantine ceramic vessels were four of Egyptian origin (Avrutis 2020: fig. 4:1–4) and one came from the region of Syria-Anatolia, likely somewhere in the Euphrates Valley, comprising an extremely rare import (Avrutis 2020: fig. 7). Another type of grave in the NRQ EB I necropolis is the shallow, rock-cut tomb that is characterized by irregular forms, varying from rectangular to rounded, each containing a single individual adult. In contrast to the burial caves, very few offerings were found, usually a single dagger. This type of tomb is not common in the archaeological record of the EB I southern Levant. Several similar burials were unearthed in Tel Aviv (e.g., Barkan and Abu-Salah 2017: fig. 4; Kanias 2011: figs 1–3), where the interred were occasionally accompanied, inter alia, by a dagger (Barkan and Abu-Salah 2017: fig. 6). The mul­tiple burials in the caves most probably indicate that they were used by the local population to inter its dead for more than one generation. The ceramic and lithic deposits deriving from the NRQ burials all date to various phases with­in the late EB I. In terms of region­alizm, the material culture belongs to a late EB I southern sub-culture, with some notable influences from the northern part of the southern Levant (Avrutis 2012: 101–99, 213–20).

Figure 2 Funerary offerings from Burial Cave F-55.

In the absence of any late EB I settlement re­mains at NRQ, a possible can­didate for the mortuary population of the ne­cropolis is the large, contemporary settlement at Tel Lod, situated 5 km to the north-west (van den Brink 2002; Yannai and Marder 2000). The presence of several Egyptian imports at the former site is consis­tent with the occurrence of Egyptian materials un­covered in the NRQ burial caves, including ceramic vessels and calcite (alabaster) mace heads (cf. van den Brink and Braun 2006).

Daggers (Fig. 3:1–3)

A dagger is a thrusting or stabbing haft weapon, used for close combat. Of the 19 daggers found in the NRQ tombs, 18³ belong to the most typical and widespread weapon of the entire Early Bronze Age, a type characterized by an elongated blade with a flattened, rhombic cross-section and a rectangular plain butt, containing two to six holes made for hafting by means of rivets. These are arranged in pairs or in a ‘V’-shape. Occasionally, in the course of tool repair, the butt was extended at the expense of the blade in order to strengthen the hafting, and the angles were blurred. Dimensions of these daggers vary in length between 134–470 mm, with a maximum width of 19–51 mm, and a maximal thickness of 4–9 mm. In various typologies, these daggers were denoted ‘Type 18’ (Maxwell-Hyslop 1946: 21, pl. II:18), ‘Narrow Daggers Type 2’ (Philip 1989: 103–04, fig. 76:b), ‘Type D’ (Shalev 2004: 133).

Figure 3 Selected dagger and spearhead from NRQ, sampled for this study: 1. Dagger L11230 W188; 2. Dagger L12055 Y768; 3. Dagger L11246 W405; 4. Spearhead L306 3096.

The earliest stratified recorded example of this dagger type was found in an early EB I layer at Beth Shean (Level XVI; FitzGerald 1935: 9, pl. III:25). In later phases of EB I, 39 such daggers were also recorded, e.g., at Azor (Ben-Tor 1975: 22, 26–27, fig. 12:4, pl. 22:1), Tell en-Naṣbeh (McCown 1947: 264, pl. 104:1), ‘Ein Assawir (Gorzalczany and Sharvit 2010: 106, fig. 13:1), Qiriyat Haroshet (Salmon et al. 2008: 13*, 18*, fig. 12), Giv‘atayim (Sussman and Ben Arieh 1966: fig. 10:1) and Ma‘abarot (Porath et al. 1985: 194, fig. 81:4). The type continues into the EB II–III, in much reduced quantities (n = 8), e.g., at Jericho (Kenyon 1960: fig. 66:3), Bab edh-Dhraʻ (Schaub and Rast 1989: 444–47, fig. 258:1–4) and Lachish (Tufnell 1958: pl. 22:6), and increases again during the Intermediate Bronze Age (IBA; second half of the 2nd millennium BCE, Regev et al. 2012) (n = 117), with a wide distribution, in the north from Hanita in the Upper Galilee (Yannai and Rochman-Halperin 2008: 3*, fig. 3:2) and Wadi Gaʻalan in Golan Heights (Dolmen no. 36, Epstein 1985: 40, pl. IX:19) to central Negev sites in the south (Cohen 1999: figs 53B:1; 80:5; 139B:1, 3). Examples were recorded, e.g., at Jericho (Kenyon 1965: figs 24:1–5, 73), Furedis (Hess 1980: fig. 1:8, 9), Kh. Ibreiktas (Porath et al. 1985: 121, fig. 17:5) and Nahal Taliya (Alexandre 2000: 120*, fig. 220). In total, there are 165 examples known during the course of EB I to IBA. Daggers of this type do not appear in the succeeding period of the Middle Bronze Age.

Outside the southern Levant, only three daggers belonging to this type are known, two from Lebanon, at Kamid el-Loz (Mansfeld 1970: 124, taf. 38:4) and below the Temple of the Obelisks at Byblos (Dunand 1954: 906, fig. 1062) and one from western Syria, at Qatna (Tomb IV; du Mensil du Buisson 1935: 155, fig. 55: left); all date to IBA. This spatial distribution indicates that the daggers under discussion belong to the indigenous metallurgical tradition of the southern Levant.

Spearhead (Fig. 3:4)

A spear is a thrusting and throwing shaft weapon, used for middle and far-range combat. The spearhead is of the type characterized by an elongated leaf-shaped blade with a pronounced midrib; the barbs of the blade can be either rounded or slightly pointed, and the tang, round in cross-section, tapering towards the hooked tip. The proportions between the blade and tang is 3:1 or 4:1. These spearheads are deemed ‘Tanged Spearheads Type 7’ (Philip 1989).

The type was introduced in late EB I, appearing at Kfar Monash (n = 4; Hestrin and Tadmor 1963: 279, figs 10: 3; 11:1–3, pl. 29), Bat Yam (Shalev 2004: 133) and NRQ (Avrutis 2012: 227–29, fig. 9.1; pl. 9:1). It was not recorded during EB II and EB III, yet re-appeared during the course of IBA, albeit in a smaller size and lacking the midrib. Examples have been recorded in the northern and central parts of Israel, e.g., at Hazor (Yadin et al. 1961: pls CCXLIV: 23; CCCXLII: 3), Beth Shean (n = 3; Oren 1973: 39, figs 20:3; 21:14, 20) and Jabel Qaʻaqir (Dever 2014: 279–89, fig. IE.3:5). Outside the southern Levant, one or two examples from Ras Shamra resemble the spearheads under discussion (Schaeffer 1962: 244, fig. 33:22, 335, fig. 4:9). It seems that this type of spearhead, like the daggers, belongs to the local metallurgical tradition.

Metallurgy in the southern Levant during the Early Bronze Age

Analysed metals

Several metal objects and metallurgical assemblages from the Early Bronze Age were previously analysed. The latter include production remains from Ashkelon-Afridar, dated to the beginning of EB I (Segal et al. 2004), and Tell esh-Shuna, in the Jordan Valley, dated to the Late EB I (Rehren et al. 1997). The former include a couple of axes from early EB I Yiftah’el (Shalev and Braun 1997), as well as objects from the above mentioned Kfar Monash hoard (Hauptmann et al. 2011), dated, based on typological considerations, to EB IB (Sebbane 2003), a small group of axes from a large building at Tel Beth Shean, dated to the same period (Segal and Yahalom-Mack 2012), several tools from EB IB–EB II at Pella (Philip et al. 2003), and a group of objects, mainly small tools, from Tel Arad, mostly dated to EB II (Hauptmann et al. 1999). Not all of the analysed objects from Bab edh-Dhra', Numeira and Jericho which appear in Hauptmann 2007 can be assigned with certainty to a particular time within the Early Bronze Age; thus, it appears that we lack metallurgical analyses securely dated to the later part of the Early Bronze Age (EB III).

Copper production and metalworking

Multiple lines of evidence, including fieldwork, radiocarbon dating and lead isotope analyses, point to Faynan as the main source of copper during the Early Bronze Age. Faynan appears to have been settled from the mid-4th millennium and most of the 3rd millennium BCE. During this time, major technological advances took place in the region, with a shift to sub-surface mining of the Dolomite Lime Shale unit (DLS) and wind-operated furnaces (Hauptmann 2007: 151). Secondary processing of copper is evidenced during EB II at both Barqa el-Hetiye and Khirbet Hamrat Ifdan at Faynan (Adams 2003; Fritz 1994; Levy et al. 2002). After a short recess, activity was renewed at the latter site during late EB III and early IBA (see Ben-Yosef et al. 2016 and references therein).

Outside Faynan, evidence for metalworking in the southern Levant, is derived mainly from late EB I villages which were located, for the most part, in south-western Canaan. These include Ashkelon Barnea, Tel Malhata, Halif Terrace, Site H (Nahal HaBesor), Tel Erani and Lod (Genz 2000; Milevski 2011: 123–24). Evidence for metalworking in the north of Canaan include the above mentioned remains from Tell esh-Shuna, dated to the late EB I, and isolated crucibles from Yiftah’el and Metzer, both from early EB I contexts. While metalworking in south-west Canaan was not studied systematically, its distribution and relative proximity to Faynan indicates the use of copper from this region. In contrast, analysis of the metallurgical remains from Tell esh-Shuna, including crucibles and moulds with remains of copper with arsenic and nickel, which could not have been derived from Faynan (or anywhere else in the Arabah and Sinai), indicates the procurement of copper from a northern source (Rehren et al. 1997). No metalworking remains were reported from EB II–III settlements (Genz 2000).

Materials and methods

Twelve daggers, eight rivets and one spearhead from the NRQ necropolis were sampled by drilling (Table 1). The drillings (15–60 mg) were dissolved with aqua regia (1:3 concentrated HCl:HNO₃). Metal concentrations were determined using a quadrupole Inductively Coupled Plasma-Mass Spectrometer (ICP-MS, Agilent 7500cx). The ICP-MS was calibrated with a series of multi-element standard solutions (Merck; ME VI), standards of major elements and a blank. Drift was corrected with the help of internal standards (originally, 750 μg/L Sc, 100 μg/L Re and 50 μg/L Rh). Standard reference samples (US Geological Survey standard reference samples T-201 and T-209) were examined after calibration for accuracy assessment. Estimated precision of the major and trace elements are 3% and 5%, respectively. Major and trace elemental accuracy was <5% for all elements. Following the separation of lead in columns (Erel et al. 2006), lead isotopic ratios were measured using Neptune plus multi-collector ICP-MS. Thallium was used for mass-bias correction. SRM-981 standard was run with the samples yielding the following values (n = 6): ²⁰⁶Pb/²⁰⁴Pb = 16.932 ± 0.001, ²⁰⁷Pb/²⁰⁴Pb = 15.485 ± 0.001, ²⁰⁸Pb/²⁰⁴Pb = 36.680 ± 0.004.⁴

Table 1 List of artefacts sampled for this study

No.	Type	Sample	Locus	Basket	Figure
1	Dagger	NES 7	11130	V925	—
2	Dagger	NES 8	11229	W187	—
3	Dagger	NES 9	11230	W188	3:1
4	Rivet	NES 10
5	Rivet	NES 1	11231	B102	—
6	Dagger	NES 14	11231	W192	—
7	Rivet	NES 15
8	Dagger	NES 11	11246	W405	3:3
9	Rivet	NES 12
10	Dagger	NES 19	11249	W207	—
11	Rivet	NES 20
12	Dagger	NES 5	11250	W209	—
13	Rivet	NES 6
14	Dagger	NES 17	11251	W211	—
15	Rivet	NES 18
16	Dagger	NES 4	11286	W271	—
17	Rivet	NES 2
18	Dagger	NES 16	12055	Y768	3:2
19	Spearhead	NES 3	326	B3096	3:4
20	Dagger	NES 13	10890	T993	—
21	Dagger	NES 21	16T11/4022	16T11/40052	—

Results

The chemical composition of the daggers and their rivets is similar (Table 2). The daggers (n = 12) are made of copper with 1.1%±0.5 As, 0.3%±0.4 Ni, 0.3%±0.2 Sb and 0.5%±0.3 Fe. The rivets (n = 8) are similarly made of copper with 1.1%±0.4 As, 0.4%±0.4 Ni, 0.3%±0.2 Sb and 0.7%±0.5 Fe. Unlike the daggers, they contain, on average, slightly more lead (0.1%±0.1 Pb), particularly NES 10, 12 and 20. The single spearhead that was sampled for this study (NES 3) differs considerably from the riveted daggers, being comprised of unalloyed copper and with trace-element concentrations (excluding Pb) that are lower at least by one order of magnitude (0.03% Fe, 0.03% Ni, 0.002% As, 0.005% Sb). The Pb content is in the same range (0.02% Pb) as the daggers.

Table 2 Chemical composition measured using ICP-MS (wt.%)

Sample	Fe	Co	Ni	Cu	Zn	As	Cd	Sn	Sb	Pb	Bi	Total
NES 1	0.47	0.000	0.105	89.6	0.000	0.97	0.000	0.04	0.244	0.02	0.001	92
NES 2	1.40	0.001	0.808	88.5	0.025	0.52	0.000	0.00	0.567	0.00	0.000	92
NES 3	0.03	0.000	0.031	92.9	0.000	0.00	0.000	0.00	0.005	0.02	0.000	93
NES 4	0.62	0.001	0.699	91.2	0.001	0.41	0.000	0.00	0.662	0.00	0.001	94
NES 5	0.23	0.003	1.180	94.0	0.002	0.68	0.000	0.01	0.228	0.02	0.000	96
NES 6	0.75	0.001	1.110	89.5	0.000	0.92	0.000	0.01	0.134	0.04	0.000	92
NES 7	1.06	0.001	0.102	95.4	0.003	1.40	0.000	0.00	0.656	0.02	0.000	99
NES 8	0.23	0.000	0.011	95.2	0.005	1.41	0.000	0.01	0.378	0.02	0.000	97
NES 9	0.14	0.000	0.276	95.8	0.000	1.02	0.000	0.00	0.290	0.02	0.000	98
NES 10	0.06	0.000	0.187	76.6	0.000	1.87	0.000	0.01	0.582	0.13	0.001	79
NES 11	0.91	0.001	0.439	92.9	0.003	1.27	0.000	0.00	0.431	0.01	0.000	96
NES 12	1.39	0.003	0.613	77.1	0.001	1.11	0.000	0.00	0.295	0.19	0.000	81
NES 13	0.43	0.001	0.043	91.7	0.004	1.16	0.000	0.01	0.149	0.01	0.000	94
NES 14	0.23	0.000	0.015	89.8	0.001	0.70	0.000	0.01	0.181	0.00	0.000	91
NES 15	0.45	0.000	0.104	91.2	0.000	0.84	0.000	0.00	0.157	0.02	0.000	93
NES 16	0.47	0.000	0.079	91.2	0.000	1.78	0.000	0.00	0.760	0.00	0.000	94
NES 17	0.38	0.000	0.017	89.1	0.001	1.87	0.000	0.02	0.231	0.00	0.000	92
NES 18	1.10	0.000	0.060	90.0	0.003	1.31	0.000	0.01	0.163	0.01	0.000	93
NES 19	0.35	0.001	0.581	92.3	0.000	0.24	0.000	0.00	0.073	0.00	0.000	94
NES 20	0.23	0.001	0.316	89.9	0.000	1.24	0.000	0.01	0.320	0.12	0.001	92
NES 21	1.08	0.004	0.065	90.7	0.001	1.05	0.000	0.02	0.027	0.04	0.000	93

The lead isotopic ratios range between ²⁰⁶Pb/ ²⁰⁴Pb = 18.9, ²⁰⁷Pb/²⁰⁴Pb = 15.69 and ²⁰⁸Pb/²⁰⁴Pb = 39.1 to ²⁰⁶Pb/²⁰⁴Pb = 18.0, ²⁰⁷Pb/²⁰⁴Pb = 15.64 and ²⁰⁸Pb/²⁰⁴Pb = 38.2. Figure 4 shows the correlation between the daggers and their rivets. Four of the six daggers that were sampled along with their rivets share similar isotopic values, suggesting that the same metal was used in the production of both the daggers and their rivets.

Figure 4 Lead isotopic ratios of weapons from NRQ. Correlation between the daggers and their rivets is indicated (see data in Tables 1 and 3).

Table 3 Lead Isotope composition measured using MC-ICP-MS

Sample	^206Pb/^204Pb	σ	^207Pb/^204Pb	Σ	^208Pb/^204Pb	Σ
NES 1	18.899	0.002	15.689	0.001	39.047	0.004
NES 2	18.655	0.003	15.666	0.001	38.754	0.004
NES 3	18.036	0.000	15.640	0.001	38.159	0.001
NES 4	18.632	0.002	15.666	0.002	38.744	0.005
NES 5	18.887	0.001	15.688	0.001	39.060	0.003
NES 6	18.883	0.001	15.689	0.001	39.074	0.002
NES 7	18.076	0.002	15.637	0.002	38.203	0.005
NES 8	18.434	0.001	15.654	0.000	38.529	0.001
NES 9	18.840	0.002	15.685	0.001	39.024	0.004
NES 10	18.906	0.000	15.691	0.000	39.099	0.001
NES 11	18.773	0.002	15.684	0.002	38.958	0.004
NES 12	18.346	0.001	15.657	0.001	38.512	0.002
NES 13	18.898	0.002	15.685	0.002	38.897	0.004
NES 14	18.788	0.003	15.659	0.002	38.898	0.005
NES 15	18.913	0.000	15.685	0.001	39.057	0.001
NES 16	18.436	0.003	15.657	0.002	38.570	0.005
NES 17	18.753	0.004	15.686	0.003	38.863	0.008
NES 18	18.806	0.003	15.681	0.002	38.911	0.005
NES 20	18.913	0.002	15.690	0.001	39.100	0.004
NES 21	18.553	0.001	15.658	0.001	38.626	0.002

Several daggers and rivets cluster around higher values, representing a geologically younger ore source, while at the lowermost values, representing a geologically older source, are the unalloyed spearhead (NES 3) and a single riveted dagger (NES 7), which is chemically similar to the other daggers (Fig. 5). The samples in between appear to be a mix of these two sources. However, when the isotopic distribution of the samples is compared with the Pb contents, additional information is revealed (for the use of such methodology in archaeology, see Eshel et al. 2021; Pollard and Bray 2015). It becomes evident that at least three sources are needed in order to account for all the NRQ samples: (1) a geologically young copper source, relatively rich in Pb (as well as in As and Ni, ‘Source A’); (2) a slightly older, Pb-poor source (poor also in As and Ni, ‘Source B’); and (3) a geologically older source (also poor in As and Ni, ‘Source C’) with intermediate Pb concentrations (Fig. 6, Table 2). The majority of the NRQ samples are comprised of mixtures of the two younger sources (Sources A and B), while only a few contain mixtures of Source A and Source C.

Figure 5 Lead isotopic ratios obtained from the NRQ weapons plotted against a two-stage Pb-Pb age model (Stacey and Kramers 1975), showing that the samples, as a group, cross several isochrons. For the results see Table 3. For ²⁰⁸Pb/²⁰⁴Pb see Figure 7.

Figure 6 Lead isotopic ratios of weapons from NRQ plotted against 1/[Pb (%)] (see data in Tables 2 and 3). Additional Early Bronze copper-based objects are plotted for comparison: ‘EB objects’ — securely dated to EBI–III from Numeira, Jericho and Bab edh-Dhra' (Hauptmann 2007: tables 8.5–8.7), Tel Beth Shean (Segal and Yahalom-Mack 2012), Pella (Philip et al. 2003) and Arad (Hauptmann et al. 1999); ‘Monash’ — EBI objects from the Monash hoard (excluding Cu sheets, after Hauptmann et al. 2011); ‘Arslantepe’ — objects from Levels VII–VIb1 from Arslantepe (Hauptmann et al. 2002).

The source of the unalloyed copper (Source C)

The unalloyed-copper spearhead from NRQ (NES 3) is consistent with the Arabah ores both chemically, being very low in trace elements (although Pb contents is on the lower end of Pb concentrations in copper from this region, see Hauptmann 2007: 73–79), and isotopically (see Fig. 7). As evidence from Faynan points to this region as the main copper source during this time (Hauptmann 2007: 211–15, 275), and since the spearhead is highly consistent with the DLS ores from Faynan, this region is more likely than Timna to have been its source. This also appears to be the case with additional unalloyed copper objects dated throughout the EB period (Fig. 7), although all the sites from which EB copper objects were sampled are located, as Hauptmann indicated, either in the proximity of Faynan (i.e., Arad, see Hauptmann et al. 1999 and Bab edh-Dhra'), or further along the Jordan valley (i.e., Numeira, Pella and Tel Beth Shean, see Hauptmann 2007: 272–88; Philip et al. 2003; Segal and Yahalom-Mack 2009). It thus appears that our Source C (above, Fig. 6) may be identified with Faynan DLS ores.

Figure 7 Lead isotope ratios of Early Bronze objects as presented in Fig. 4 (see caption there for references), plotted against a two-stage Pb-Pb age model (Stacey and Kramers 1975), as well as selected copper ores from the Arabah (for Amir-Avrona in Timna see data in Gale et al. 1990; Hauptmann 2007, for Faynan DLS see Hauptmann et al. 1992), copper and lead ores from Levels VII–VIA in Arslantepe (Hauptmann et al. 2002), and copper ores from Anatolia (Begemann et al. 2003; Hirao et al. 1995; Seeliger et al. 1985; Wagner et al. 1986; 1989; 2003).

The origin of the CuAsNi(Sb) alloys (Sources A and B)

Different variations of CuAsNi alloys were in use in the southern Levant during the Chalcolithic period, as reflected by the prestigious objects of the Nahal Mishmar hoard (Shalev and Northover 1987; 1993; Tadmor et al. 1995). Originally, this composition was considered a natural alloy, but while the Arabah (and Sinai) could certainly be excluded, based on the composition of the ores (Hauptmann 2007: 280–81), a compatible ore source was never identified (Golden 2014 and bibliography therein). In addition, despite the fact that stylistically, the Nahal Mishmar hoard was deeply rooted in the Ghassulian Chalcolithic tradition of the southern Levant, suggesting that the objects were produced locally, the actual place of production remains unknown (but see Goren 2008). It has been suggested that the difficulty in identifying the source of the metals is related to the fact that the alloys were artificially made by alloying or co-smelting complex metals and minerals, and possibly base-metal speiss (the by-product of smelting complex arsenic) or antimony-rich copper or lead ores (Rehren et al. 2012; Thornton et al. 2009).

Our results point to Sources A and B as being of a possible east Anatolian origin. As seen in Figs 4–7, a large group of daggers and rivets from NRQ sampled for this study cluster around ²⁰⁶Pb/²⁰⁴Pb = 18.9, ²⁰⁷Pb/²⁰⁴Pb = 15.69 and ²⁰⁸Pb/²⁰⁴Pb = 39.1, denoted here Source A, together with the alloyed objects (mainly adzes) from the contemporary Kfar Monash hoard. Both these groups are consistent with complex ores and CuAsNi objects from Arslantepe (Hauptmann et al. 2011), as well as with additional 4th and 3rd millennia BCE sites in the upper Euphrates area, such as Hassek Höyük, Norsuntepe and Tülintepe (see Hauptmann 2007: 297 and Hauptmann et al. 2011: 60 for references). Objects from Arslantepe Str. VII–VIb fall on the same mixing line as the objects from NRQ, between Sources A and B (Fig. 6). Palmieri et al. (1999) note a continuity in metallurgical practices throughout Levels VII (3700–3400 BCE), VIA (3400–3000 BCE) and VIB (3000–2900 BCE), although the use of polymetallic ores is mainly ascribed to Str. VIA (Uruk period).

Hauptmann et al. (2011: 75–76) noted a ‘conspicuous geographic congruity of the find sites of arsenic-nickel rich copper artefacts from the 4th and 3rd millennia BCE with the occurrence of Mesozoic ophiolithic rocks of the Tethyan Eurasian Metallogenic Belt (TEMB)’ and particularly with ‘isotopic composition of lead in ores from the ophiolite-hosted “giant” copper deposit of Ergani Maden and other deposits in comparable geological contexts close by’. Indeed, substantial evidence of metal production, including the presence of ores, was recorded in these east Anatolian sites which appear to have specialized in metal production (e.g., Palmieri et al. 1999; Yakar 2002). Lehner and Yener (2014) suggested that ore was smelted elsewhere and the metal was brought to Arslantepe and neighbouring sites for further processing, however, this does not explain the presence of ores there, unless these were required for alloying practices. Thus, as noted above, our data shows that the NRQ daggers are made by mixing two different Anatolian ores (Sources A and B), similarly to the Arslantepe objects.

Discussion

Use of alloyed vs unalloyed copper

The present assemblage consists only of weapons, and mainly of a single type, the riveted dagger. Significantly, the one weapon which differs typologically, the spearhead, is also very different in composition, being made of unalloyed copper. This may indicate a specialized production of riveted daggers, a case which may be supported by the use of the same metal to produce both the daggers and their rivets. Based on the objects from the EB sites mentioned above (Beth Shean, Pella, Arad, Bab edh-Dhra', Numeira and Jericho), it appears that unalloyed copper was used mainly for the production of tools. A more complex picture emerges, however, from the study of the Kfar Monash hoard wherein the adzes, albeit of an Egyptian type (Hestrin and Tadmor 1963; Sebbane 2003), were produced using a CuAsNi alloy, while one of the two analysed spearheads and two riveted daggers were made of unalloyed copper (Hauptmann 2007: 275–85; Hauptmann et al. 2011). The source of the unalloyed copper was likely Faynan, while the complex metals used for the production of the alloys may have originated in the upper Euphrates region, as detailed above. The daggers (with their rivets) and the spearhead from the NRQ burials analysed in the framework of this study thus complement and support previous understanding of the metallurgical technology during EB IB–EB II, when unalloyed and alloyed copper were used contemporaneously for weapons of different types.

Source and trade of alloyed copper

The distribution of the NRQ dagger type favours its local production in the southern Levant. This leads to the conclusion that rather than the daggers themselves, raw material for their production was transferred to the southern Levant from eastern Anatolia or its environs. The question is, in what form? Were foreign complex minerals or speiss co-smelted in the Levant together with local Arabah copper minerals or, alternatively, were ready-made complex alloys traded in metallic form (as ingots)? Were these in turn alloyed with Arabah copper? The pattern presented above shows a clear mixing between the two younger sources (A and B), indicating that the alloys were artificially produced through mixing and thus were likely traded in ingot form. A mixing between Source A and Source C, which we identify as the DLS ores from Faynan, is suggested here based on the comparison between 1/Pb and ²⁰⁶Pb/²⁰⁴Pb, but appears to have been the exception rather than the rule in the assemblage under study; sample NES 7 appears to be the best example of such a mixture (Fig. 6).

Another question pertains to the place of production. The majority of metalworking sites during the late EB I were limited to south-western Canaan and were more likely working unalloyed copper from Faynan, rather than alloyed copper from far northern sources (as detailed above). Evidence from Tell esh-Shuna in the Jordan Valley, however, is a good indication that workshops using complex metals did exist in the southern Levant at this time and should, perhaps, be sought in the more northern parts of the region (Rehren et al. 1997). Notably, Rehren et al. (1997) viewed this technology and the metal sources as representing continuity from the Chalcolithic period. However, there is a considerable gap between the Ghassulian Chalcolithic and the EB IB, so that in the present state of research, we might point to similarity rather than continuity.

Cultural transfer?

Interaction between the southern Levant and the upper Euphrates region is reflected mainly by the trade in metals. Very little additional evidence exists to date, although a single pottery vessel from the NRQ necropolis which was likely imported from this region should be noted (Avrutis 2020). The metallurgical data presented above raise the possibility that during this period, the metal trade, via direct or indirect contacts, served as a conduit for cultural transference, possibly including concepts related to urbanism that were widespread in Mesopotamia. The village-based rural region of Arslantepe and its environs demonstrated rather sophisticated metallurgical skills. Yet, it was not at the core of the Urukian urban phenomenon and the level of interaction between this region and urbanized southern Mesopotamia remains under debate (e.g., Algaze 2005; Frangipane 2001: 325–40). A possible scenario entails merchants/agents from urban centres in southern Mesopotamia or northern Syria who were active in bringing the metals to the southern Levant from the Arslantepe region. Another possibility is that Canaanite agents actively sought out this raw material and, in doing so, encountered at close range the urban reality on their way to the source area and its environs. They then brought these concepts back to a society and economy that was primed to accept the actions required to implement such far-reaching developments and changes (e.g., Greenberg 2011; 2019). Additional analytical studies such as those presented in the present article can contribute to a better understanding of such interregional contacts and processes of transmission.

Analyses of well-dated metal objects from the urbanized EB II–III are also required in order to ascertain the continuity of such contacts into the latter part of the Early Bronze Age. Hauptmann (2007: 296), for example, suggested that the ongoing metal-trade connections with Anatolia were instrumental in the arrival of the Khirbet Kerak culture to the southern Levant. Several prestige items, or their conceptualization, such as stone and ivory bull figurines (Beck 1995: 23; Ben-Tor 1972) and decorated bone tubes (Zarzecki-Peleg 1993), were considered to have arrived from Mesopotamia and Anatolia, as did certain cultic ideas such as the ‘Sacred Marriage’ motif (de Miroschedji 2011). As mentioned above, a sharp decrease occurred in the number of riveted daggers during EB II–III, accompanied by a considerable intensification of copper production at Faynan with the establishment of secondary metalworking in Khirbet Hamrat Ifdan in the EB II and again in the late EB III (see above). Whether these phenomena are inter-related, and what their far-reaching implications for inter-regional connections are, remains to be determined.

Acknowledgements

We thank Ruhama Bonfil for the graphics and Nava Panitz-Cohen for her editing and valuable comments. The study was funded with the generous support of the Philip and Muriel Berman Center for Biblical Archaeology at the Institute of Archaeology, The Hebrew University of Jerusalem.

Notes

1 The site is also known as el-Khirbeh, the Arabic term for ruin.

2 Ar­chaeological exploration has been conducted at the site from 2006 until the present on behalf of the Zinman Institute of Archaeology of the Uni­versity of Haifa. Excavations are directed by S. Kol-Yaʻakov, and since 2014, co-directed with V. W. Avrutis.

3 To date, 20 copper-based EB I weapons have been found in the NRQ burials; the present paper discusses 13 of them, including 12 daggers and a single spearhead.

4 Both analyses were performed by Ofir Tirosh in the Institute of Earth Sciences, The Hebrew University of Jerusalem.