bs_bs_banner Archaeometry 56, 2 (2014) 279–295 doi: 10.1111/arcm.12015 MI D D LE B RONZ E AGE II B AT T LEA X ES FRO M R I S H O N L E Z I O N , I SRAE L : ARCHAE OL OG Y A N D META LLU RG Y * S. SHALEV,1,2 E. N. CASPI,3 S. SHILSTEIN,4† A. M. PARADOWSKA,5§ W. KOCKELMANN,5 T. KAN-CIPOR MERON1 and Y. LEVY6 1 Department of Archaeology University of Haifa, Mount Carmel, 31905 Haifa, Israel Charney School of Marine Sciences, University of Haifa, Mount Carmel, 31905 Haifa, Israel 3 Physics Department, Nuclear Research Centre, Negev, POB 9001, 84190 Beer-Sheva, Israel 4 Department of Physics, Weizmann Institute of Science, 76100 Rehovot, Israel 5 Rutherford Appleton Laboratory, ISIS Facility, Chilton Didcot, Oxfordshire OX11, UK 6 Israel Antiquities Authority, POB 586, Jerusalem 91004, Israel 2 Thirteen bronze battleaxes from Middle Bronze Age II graves at Rishon LeZion, Israel were analysed by ED XRF at multiple surface locations in order to determine their metallurgical composition. Six of these were further subjected to neutron diffraction using an ENGIN-X diffractometer in order to determine bulk phase composition. The results indicate that the previously established geographical and chronological variability in Sn–Cu (with occasional Pb) and As–Cu alloys found using the former method may be an artefact of preservation and conservation. In addition, the varying homogeneity determined by the latter reflects special treatment for improving on the metal cast. KEYWORDS: EASTERN MEDITERRANEAN, LEVANT, ISRAEL, ARCHAEOMETALLURGY, MIDDLE BRONZE AGE METALS, XRF, NEUTRON DIFFRACTION, ARSENICAL COPPER, TIN-BRONZE, LEADED-BRONZE INTRODUCTION Hundreds of copper-based objects dated to the Middle Bronze Age II (MB II; c. the first half of the second millennium bce) have been unearthed, largely in burials, all over the Levant. In this period, the development of more complex weapons (longer daggers, swords, complex battleaxes etc.) was made possible by alloying the copper, initially with arsenic (As) and, later, with tin (Sn) to produce, respectively, arsenical copper and tin-bronze (Philip 1991). These changes in the metal properties of weapons have been, up until now, better documented in the diachronic change in composition of smaller objects, such as toggle pins, which were probably made mainly from scrap re-melting (e.g., Shalev 2002, 309–10). This period also saw lead (Pb) beginning to play a greater role in alloying, in particular for thick and relatively large copper-based casts such as battleaxes. After a 2000-year history (fourth to third millennia bce) of mainly copper and arsenical copper metallurgy in Israel and Jordan during the Chalcolithic and Early Bronze Ages (e.g., Goldin et al. 2001), and before the predominance of tin-bronze and intensive Late Bronze Age Mediterranean trade in copper and tin ingots (c. second half of the second millennium bce; cf. Budd et al. 1995), the Middle Bronze Age constitutes a transitional period of metal alloying (e.g., Philip 1995a; Shalev 2009). It is therefore usually assumed that the MB copper-based objects alloyed with *Received 9 April 2012; accepted 22 November 2012 †Corresponding author: email sana.shilstein@weizmann.ac.il §Present address: Bragg Institute, ANSTO, Locked Bag 2001, Kirrawel DC NCW2232, Australia. © 2013 University of Oxford 280 S. Shalev et al. arsenic (As) are probably ‘older’ than similar objects that are alloyed with tin (Sn) and, often, lead (Pb), and that they are part of a ‘newly arrived fashion’. In a comprehensive study of the complexity of MB metal production and distribution, Philip (1995b, 523) stated correctly that ‘. . . the metalwork of the southern Levant and the eastern Nile Delta shows a remarkable degree of stylistic homogeneity during the MB II period’. By comparing the composition of the MBII metal objects from Jericho in Palestine and from Tell el-Dab`a in Egypt, Philip (1995b, 524) concluded that they are the products of ‘. . . two separate metal industries [that] were producing stylistically similar objects’. New analytical data of MBII artefacts from the past 12 years of archaeology in Israel—for example, finds from Aphek, Kabri, Gesher and Fasuta (Shalev 2000, 2002, 2007, 2010)—were published mainly within excavation reports and, therefore, are not yet included comprehensively in the above syntheses. In addition, the large volume of comparative data of similar objects and new analyses from Tell el-Dab`a (Philip 2006, 204–14), Byblos (El Morr and Pernot 2011) and Sidon (Doumet-Serhal 2003), offer an opportunity to consider some of the fundamental questions of MBII archaeometallurgy. Recently, a synthetic summary of the state of research dealing with MBII weapons (Shalev 2009) showed a current lack of correlation between metal composition and spatial and temporal distribution of identical types; that is, all possible alloys could be found region-wide and over time. However, within these alloys in identical objects, extreme quantitative variability is apparent, the cause of which has yet to be studied. In the present research, we wish to continue in this direction and try to better understand the possible causes for such a high variability in the quantities of the major elements detected in different objects of the same type and identical shape. To that end, a relatively large group of MBII battleaxes from all four major types (see Figs 1–4 and Table 1; see also Miron 1992) was chosen for analysis. They derive from a vast MBII cemetery in Rishon LeZion, south of Tel Aviv on the Mediterranean coast of Israel. The selection of a single site and a single cultural period was designed in order to better differentiate between the possible causes for such a large variation and to better distinguish between a real effect and an analytical artefact. A real effect means an actual difference in the quantities of major elements used for the production of these objects. In contrast, an analytical artefact could be caused by an inhomogeneity of the cast, corrosion or modern conservation treatment, as well as the use of different analytical methods and protocols. Each one of these factors, or a combination thereof, could cause the above quantitative variations. Therefore, in order to begin isolating the potential variables, the first stage entailed systematic XRF compositional analysis of different points on the cleaned surface of the same object with a single method and protocol. Following this initial stage, the results were further evaluated by subjecting selected previously analysed axes to neutron diffraction (ND) analysis in order to measure their bulk metal composition in different areas of the object. MATERIALS The 13 battleaxes found in different graves in the MBII cemetery at Rishon LeZion include unpublished axes from all four major known types (Shalev 2009): duckbill, narrow flat shafthole and rounded shafthole axes (Miron 1992, 71–80, Types 1–3). Their typological characteristics are detailed in Table 1. These include: a duckbill axe; 12 narrow-blade shafthole axes, of which 10 are Type 1 (Miron 1992, 71–4), one is a Type 2 (Miron 1992, 71, 74–5) and one is a Type 3 (Miron 1992, 71, 75). These 10 flat shafthole axes are the largest concentration of this type found in Israel; all were found in a mortuary context in what are known as ‘warrior burials’ (e.g., Hallote 1995, 2002; Ilan 1995; Philip 1995a,c). © 2013 University of Oxford, Archaeometry 56, 2 (2014) 279–295 Middle Bronze Age II battleaxes from Rishon LeZion, Israel 281 Figure 1 The main types of axe (after Miron 1992) from the Rishon LeZion Cemetery: (a) duckbill axe (BA8, 743-7327); (b) shafthole axe Type 1 (BA6, 654-6362); (c) shafthole axe Type 2 (BA15, 769-8131); (d) shafthole axe Type 3 (BA16, 755-7849). The XRF compositional analysis of the axes from Rishon LeZion may be compared to data from 39 similar axes from other sites (Shalev 2009, 72–3), including analyses of 23 duckbill axes, 12 flat axes (Type 1) and four rounded shafthole axes (Types 2 and 3), as well as 13 analyses of similar axes from Tell el-Dab`a (Philip 2006, 208–9, tables 17 and 18) including one duckbill, two flat and 10 rounded shafthole axes. The results of the compositional analyses of these finds are detailed along with their typological assignments in Table 2 and are discussed in greater detail below. However, it is worth noting the typological similarity and the overall compositional similarity of the metal in the Rishon LeZion axes to those of the previously published analyses of axes from other sites (Tables 3–5). © 2013 University of Oxford, Archaeometry 56, 2 (2014) 279–295 282 S. Shalev et al. Table 1 Typological characteristics of the MBII battleaxes from Rishon LeZion Type Laboratory and field nos. Duckbill Flat shaft Flat shaft Flat shaft Flat shaft Flat shaft Flat shaft Flat shaft Flat shaft Flat shaft Flat shaft Round shaft Round shaft BA8; 743-7327 A-2731 BA3; 506-5077 BA4; 1081-9232 BA5; 1076-9780 BA6; 654-6362 BA7; 656-6418 BA9; 1037-6854 BA10; 1085-9247 BA11; 1107-9370 BA12; 1118-9418 BA14; 1086-9293 BA15; 769-8131 BA16; 755-7849 Table 2 Weight (g) Length (mm) Width (mm) 104.41 124.00 95 123 103 115 132 90 108 102 103 108 113 155 165 42 17 15 15 19 14 15 16 15 13 15 35 17 92.34 103.01 148.87 152.94 XRF surface analysis (wt%) of axes from Rishon LeZion Type Laboratory and field nos. Duckbill Flat shafthole Flat shafthole Flat shafthole Flat shafthole Flat shafthole Flat shafthole Flat shafthole Flat shafthole Flat shafthole Flat shafthole Rounded shafthole Rounded shafthole BA8; 743-7327 A-2731 BA3; 506-5077 BA4; 1081-9232 BA5; 1076-9780 BA6; 654-6362 BA7; 656-6418 BA9; 1037-6854 BA10; 1085-9247 BA11; 1107-9370 BA12; 1118-9418 BA14; 1086-9293 BA15; 769-8131 BA16; 755-7849 % As (c1–c2) % Sn (c1–c2) % Pb (c1–c2) n.d. 0–0.5 n.d. n.d. tr 2.2–2.3 n.d. n.d. n.d. 1.5–9 tr 0.3–0.4 0.9–1.8 8–20 8–12 8–15 0.4–0.9 8–18 tr 5–16 9.5–24.5 11.5–13.5 1.3–2.9 5–20 12–13 tr 1–6 0.5–1.8 0–0.8 tr 0.5–8 0.1–0.2 1–3 0.6–7.5 tr 4–21 2–15 2.5–5 0.3–0.7 c1–c2, ranges of concentration measured at different points; n.d., not detected; tr, traces. The duckbill axe (Table 3) occurs mainly in inland northern Syria/southern Anatolia, and on the Syrian and Lebanese Coastal Plain (e.g., Gerstenblith 1983, 89–91; Yogev 1985; Philip 1989, 49–50, 282–91; Miron 1992, 58–71; El Morr and Pernot 2011, 2), although its distribution extends as far as Mari in the east, Kültepe in central Anatolia in the north, and as far south-west as Tell el-Dab`a in Egypt. In Palestine, this axe type is common primarily in the north, but now occurs as far south as Rishon LeZion, south of modern Tel Aviv. Apart from a few specimens without an archaeological context, such axes have only been found in tombs dated to the Middle Bronze Age IIa. The production mode of the duckbill axe has been discussed in detail by Philip (1989, 1991) and by Shalev (2009, 70–2). It was cast in a two-piece stone mould with a core, probably of clay, which was inserted for the production of the hollow socket. Three fragments of such moulds were © 2013 University of Oxford, Archaeometry 56, 2 (2014) 279–295 Table 3 Site Reg. no. 1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 24 Byblos Byblos Kadesh, Syria Baghouz, Syria Sidon Sidon Gesher Aphek Aphek Hama Kabri Tell el-Dab`a Ain es-Samiyeh Hama Rehov Kurdani Byblos Gesher Byblos Gesher Byblos Byblos Kafr el-Djarra Byblos 10646 (6271) 10794 AO 3226 AO 18201 S/1820 S/3003 89-556 47-2305 47-2040 SA-812 4220 6139 S2011 SA-813 4 5 10646 (6273) 89-588 7903 (6281) 89-590 10646 (6272) 940 (8349) AO 10958 10645 (6270) Cu As Sn Pb Fe 92.8 90.0 59.85 67.87 82.59 72.52 66.45 83.26 68.34 86.60 83.30 88.00 82.07 81.80 85.00 77.50 89.2 86.81 81.8 95.31 72.8 71.4 76.48 83.5 n.d. n.d. n.d. n.d. 0.02 0.14 0.20 0.22 0.24 0.30 0.42 0.40 0.49 0.56 <1.0 <1.0 1.2 2.04 3.0 3.23 3.5 4.8 5.77 7.8 4.0 10.0 8.16 14.08 0.56 0.42 15.8 4.27 5.90 5.30 12.25 11.0 14.77 8.30 7.00 11.0 8.0 0.13 2.8 0.02 3.4 n.d. 10.45 1.21 2.8 n.d. 30.54 17.35 0.73 26.23 17.1 0.27 25.5 5.10 1.50 <0.1 0.772 6.80 5.50 10.5 0.7 10.2 12.4 0.78 20.3 23.2 12.24 6.8 n.d. n.d. 0.38 0.21 10.88 0.52 0.18 0.21 0.32 0.05 0.15 0.131 0.07 n.d. 0.32 n.d. 0.32 n.d. n.d. 0.24 n.d. Reference El Morr and Pernot (2011, 5, table 4) El Morr and Pernot (2011, 5, table 4) Shalev Shalev Le Roux et al. (2003, 59, table 1) Le Roux et al. (2003, 59, table 1) Shalev (2007, 110, table 7.3) Shalev (2000, 279, table 13.2) Shalev (2000, 279, table 13.2) Philip (1991, 94, table 1) Shalev (2002, 309, table 8.3) Philip (2006, 210, table 19) Dever (1975, 31, fig. 2) Philip (1991, 94, table 1) Shenberg (1985) Shenberg (1985) El Morr and Pernot (2011, 5, table 4) Shalev (2007, 110, table 7.3) El Morr and Pernot (2011, 5, table 4) Shalev (2007, 110, tables 7.2 and 7.3) El Morr and Pernot (2011, 5, table 4) El Morr and Pernot (2011, 5, table 4) Shalev El Morr and Pernot (2011, 5, table 4) Middle Bronze Age II battleaxes from Rishon LeZion, Israel All objects are from burials dated to the MBIIa; n.d., not detected; all references to Shalev represent new data published here. 283 © 2013 University of Oxford, Archaeometry 56, 2 (2014) 279–295 No. The chemical composition (wt%) of the duckbill axes 284 S. Shalev et al. Table 4 No. Site Reg. no. 1 2 3 4 5 6 7 8 9 10 11 12 13 14 Megiddo tomb Megiddo tomb Sidon Kfar Vradim Yiftahel Tell el-Dab`a Yiftahel Tell el-Dab`a Hama Aphek Kafr el Djarra Kafr el Djarra Gesher Fasuta 385 389 S/1744 The chemical composition (wt%) of the flat shafthole axes P-46 3082 P-45 811 5E-802 KV 98-19 AO 10961 AO 10960 89-587 F-37 Cu As Sn Pb 97.24 88.42 85.46 85.00 93.59 85.00 86.60 80.13 66.72 86.97 93.71 90.03 n.d. n.d. 0.03 0.03 0.12 0.13 0.19 0.20 0.30 0.35 0.37 0.39 3.47 4.34 9.17 10.78 0.11 10.06 12.27 6.64 6.00 6.70 5.30 9.76 14.26 11.49 0.03 0.02 2.21 0.93 2.28 1.18 1.90 4.14 0.10 0.61 5.10 0.45 18.31 0.69 0.04 4.35 Fe 0.23 0.05 0.01 0.20 0.04 0.03 0.05 0.07 0.04 0.12 2.27 0.73 Analysis Philip (1989, 51) Philip (1989, 51) Le Roux et al. (2003, 59, table 1) Shalev (2009, 72, table 2) Shalev and Braun (1997, 95, table 11.3) Philip (2006, 208, table 17) Shalev and Braun (1997, 95, table 11.3) Philip (2006, 209, table 18) Philip (1991, 94, table 1) Shalev (2009, 72, table 2) Shalev Shalev Shalev (2007, 110, tables 7.2 and 7.3) Shalev (2010, 44, table 3) All objects are from burials; all except no. 14 are dated to the MBIIa; no. 14 is dated to the transitional MBIIa–MBIIb; all references to Shalev represent new data published here. Table 5 The chemical composition (wt%) of the rounded shafthole axes No. Site Reg. no. Cu As Sn Pb Fe 1 2 3 4 5 6 7 8 9 10 11 12 13 14 Tell el-Dab`a Tell el-Dab`a Aphek Tell el-Dab`a Jericho Tell el-Dab`a Tell el-Dab`a Tell el-Dab`a Tell el-Dab`a Tell el-Dab`a Jericho Tell el-Dab`a Tell el-Dab`a Rumeida 133 339 10764/60 2187 29/63 359 1755 4148 349 403 29/62 3060 4716 T-34 88.00 94.00 71.56 94.00 86.07 96.00 91.00 99.00 95.00 96.00 86.33 95.00 95.00 93.78 0.04 0.06 0.19 0.30 0.45 0.50 0.80 1.20 1.40 1.40 1.63 1.86 3.10 3.42 8.10 <0.20 6.50 <0.20 6.68 <0.10 6.70 <0.10 <0.20 <0.20 6.72 0.30 <0.10 0.14 0.40 0.01 0.15 0.06 1.23 0.03 0.34 0.04 0.13 0.28 0.10 0.17 0.45 0.61 0.29 1.70 0.29 0.35 0.02 0.92 0.15 0.65 1.07 0.87 0.16 0.67 0.37 0.81 Analysis Philip (2006, 209, table 18) Philip (2006, 209, table 18) Shalev (2000, 279, table 13.2) Philip (2006, 209, table 18) Khalil (1980, 151) Philip (2006, 209, table 18) Philip (2006, 209, table 18) Philip (2006, 209, table 18) Philip (2006, 209, table 18) Philip (2006, 209, table 18) Khalil (1980, 151) Philip (2006, 208, table 17) Philip (2006, 209, table 18) Shalev (2009, 73, table 3) All objects are from burials and are dated to the MBIIb except for nos. 3 and 7, which are dated to the MBIIa. found at Byblos, two of which were found out of context (Dunand 1939, 20, pl. 108; Dunand 1954, 96, pl. 184) and one in strata 1–10 (Dunand 1939, 198, pl. 108) of the Middle Bronze Age II. The mould was made from two flat steatite slabs, joined by means of at least two drilled holes. One of the connecting holes is situated at the centre of the back part of the mould, while the other is on one of the front sides of the blade. A wide conical sprue was left as a runner at the front of the carved blade for pouring in the molten metal. No risers are visible on the mould, so it could be assumed that the runner was wide enough to let the gases escape during casting. The relatively slow cooling rate of the cast metal is visible in some of the preserved dendritic structures. These © 2013 University of Oxford, Archaeometry 56, 2 (2014) 279–295 Middle Bronze Age II battleaxes from Rishon LeZion, Israel 285 show that the rate was insufficient to allow total homogenization during solidification, for those metal areas that were less affected by further thermal and mechanical treatment. Thus, the probable use of a stone mould for the casting of these axes is indicated. After casting, the conical block of metal filling the sprue was cut and probably saved for re-melting. The cast axe then underwent further thermal and mechanical treatment by being at least partially homogenized mainly in the blade area, an effect that could very well be achieved by several cycles of annealing and hammering. In several metallographic samples taken from the socket or bulk areas of similar axes, remains of as-cast dendritic structures are preserved, showing that the cast metal in these particular areas was not totally homogenized during treatment after casting. This treatment is visible in the microstructure of the metal in the Gesher examples (Shalev 2007, 11–12), as well as in other duckbill axes (Branigan et al. 1976, 18). This process was probably carried out in order to sharpen and slightly harden the blade. The hammering process of the blade was not extensive considering the relatively low level of hardness (not exceeding 140Hv for the Gesher duckbill axes), and the fact that copper alloys with 3% arsenic, or more than 13% tin, could be hammered, as in the case of dagger blades (Shalev 1996, 13), to a hardness of more than 200Hv and 240Hv, respectively (for details of the metal properties of such alloys, see Northover 1989). The axe was then secured to a wooden handle inserted into the socket by means of nails hammered into the handle immediately above the top of the axe (e.g., Philip 1995a, fig.1-1; Kan-Cipor Meron 2003, fig. 17; Philip 2006: fig. 4-2). It is interesting to note here that although this type of axe was treated after casting in order to be suitable for use as a weapon, none of the known axes show any signs of wear or use on their blade. The duckbill axes, although all dated to the beginning of the period (MBIIa), are found to be made of two different metal alloying compositions, with no visible chronological or geographical distinction: (1) arsenical copper (2–3% As), with no tin; and (2) tin-bronze (5–16% Sn), with either no arsenic or traces (less than 1% As). The axe from Rishon LeZion (Tables 1 and 2) belongs to the latter group. Lead in alloying quantities (1.5–30.5 Pb) was added to most of the two alloy types. Only in seven cases (one for arsenical copper, five for tin-bronze and one for a low-As, -Sn and -Pb alloy) was less than 1% of lead detected. As lead is not soluble in copper and segregates during solidification, the total relative amount of lead could be affected by a small and not necessarily representative sampling area. Moreover, all counts of tin higher than 13–14% are probably analyses of areas enriched by corrosion or an inhomogeneous matrix. Hardness determinations of arsenical copper items show that the use of this alloy does not affect the quality of the object (Northover 1989). The presence of lead in appreciable quantities in some cases (10–30.5% measured Pb) is unique to this type of axe, with no known parallels in any other Middle Bronze Age metal objects from Palestine (e.g., Shalev 2000, 281, table 2). Like the duckbill axe, the Type 1 flat shafthole axes (Table 4) and Types 2 and 3 rounded shafthole axes (Table 5) belong to a large typological group of over 20 items (Gerstenblith 1983, 91; Miron 1992, 71–4). These types already replace the duckbill axe within the MBIIa period (Ziffer 1990, 71–2). The distribution of this group is certainly more limited than that of the duckbill axe: it extends throughout modern Israel, between Tell el-`Ajjul in the south and Safed in the north, and into southern Lebanon. Isolated items have also been found at both ends of the ‘Fertile Crescent’, with one example in Hama, Syria and more than 14 from Tell el-Dab`a in Egypt (Philip 2006, 138–9). These types are dated to the end of MBIIa and continue into the MBIIb period (Philip 2006, 139). The Type 2 axe was in use until the end of the MBII period. According to their shape, it was suggested that such axes also could have had some symbolic role in the society of the southern Levant (Philip 2006, 140). © 2013 University of Oxford, Archaeometry 56, 2 (2014) 279–295 286 S. Shalev et al. The production method of these types is similar to the one described above in detail for the duckbill axe. Fragments of a two-piece steatite mould for casting these types of axes have been found at Megiddo, Byblos and Tell el-Dab`a (Miron 1992, 56). As above, metallographic samples from the blade and socketed and bulk areas of typologically related axes show a similar metallographic picture as in the duckbill axe type; that is, the as-cast dendritic structure remains and massive annealing and hammering were discerned in the blade area. In several cases, as in the former type, metal nails for securing the wooden handle are still found on top of the socket (e.g., Philip 1995a, figs 1 and 2). As with duckbill axes, all chemical analyses of these types show similar groups of alloys, with differing alloying ratios. The flat shafthole axes were similarly made of either arsenical copper (3.4–4.3% As), with no tin but with some iron (0.7–2.3% Fe), or from a tin-bronze alloy (5–12% Sn), with less than 0.5% arsenic. Lead in much lower quantities than in the above (1–5% Pb) was detected in both alloy types, but not in all of them. All the tin-bronzes have similar amounts of tin (6.5–6.8% Sn), one with 3.4% arsenic and another with 1.2% lead, and the third with no significant traces. Two such axes from Tell el-Dab`a were made of copper with a small amount of iron and little or no arsenic (0.5% or n.d. As) and one was made just from unalloyed copper. The compositional data of those objects might reflect analysis of corrosion as well as an actual different original metal base. To date, the overall analytical results of the MBII metals clearly show the use of at least two different major alloys for the production of the aforementioned axe types. In addition, a comparison of the compositional data within different types shows variations in relative quantities of alloying elements. These reflect, possibly, the amount of scrap used in the production of each type, and are dependent on the scrap composition that was available at a particular time and place (Shalev 2009). Nevertheless, there is a striking difference in the quantities of the major alloying elements: a factor of 18 for the Sn content in duckbill axes (4–16% Sn); a factor of 10 for flat axes (6–12% Sn); and fairly constant compositions (6–7% Sn) for the rounded shafthole axes. The opposite trend is observed for the relative amount of arsenic, which seems to be more constant when it is the major alloy in duckbill axes (3–4% As), but that may be the result of the limited group analysed thus far. The high variability in the relative amount of lead is omitted from the discussion here due to the insolubility of Pb in the metal solid solution, and as the random results do not show any logical pattern in relation to the other alloying elements. What can be observed is a change in the frequency of use of lead in the various types of MBII weapons. Lead is frequently used in duckbill axes (up to 30% in 73% of the analysed objects); less in flat shafthole axes (up to 18% in 57% of the analysed objects); and even less in the rounded shafthole axes (only up to 1% in 14% of the analysed objects). This selective use of lead is more prominent when other types of weapons, such as daggers and spearheads, are considered (Shalev 2009, 74, 76–8). Whether or not the high variations in alloying quantities represent real variations in the original alloys or an ‘artefact’ caused by differing analytical methods, sampling spots, and the object’s state of preservation and conservation, deserve further investigation. METHODS In order to investigate the possible causes for the observed variability in analytical results, a research programme was designed wherein multiple locations were analysed in sequence on the surface of each of the axes from Rishon LeZion. First, all of the axes were treated by standard conservation procedures in the laboratories of the Israel Antiquity Authority. The metal composition of the axes was analysed by X-ray fluorescent (XRF) analysis of their surface, using a © 2013 University of Oxford, Archaeometry 56, 2 (2014) 279–295 Middle Bronze Age II battleaxes from Rishon LeZion, Israel 287 bench-top model EX-310LC energy-dispersive spectrometer produced by Jordan Valley Co. (for a detailed technical description, see Shalev et al. 2006). A voltage of 35 kV and a specially added filter made of pure Al on the detector window (0.24 mm in thickness) were used for the binary Cu-based alloys. A limit of detection of about 0.05–0.10 wt% was achieved for metals such as Sn, Pb and As. Standards were specially prepared for calibrating the measured quantities, mixtures of the metal oxides with defined concentration ratios. The relative accuracy of the concentration was determined to be 5%, on average. That accuracy level was cross-checked by analysing mixtures of thin powders of metals with defined ratios and by measuring several certified alloy standards of tin-lead bronze. For alloys with Pb and Sn in contents higher than 5%, a correction for mutual attenuation of the Pb and the Sn peak lines was introduced. The relative accuracy in measuring higher concentrations of up to 20% of Pb and Sn is about 10–15%; for high Pb content, the limit of detection for As is not higher than 0.3%. The measurements were carried out using a current of 0.2 mA with 300–600 s exposure times, on at least four or five different points on the surfaces of each axe, with exposure areas of 2 mm in diameter. Some of the flat shafthole axes were scanned along the total length from the blade to the socket (in these cases, the number of measured points could increase to upwards of 10 on each axe). The variation of concentration for all the analysed elements is presented in Table 2. As is typical for XRF analysis, the results represent only the surface layers with a depth penetration of no more than several tens of microns. Although the results of these measurements are discussed in detail below, it is already worth mentioning here that the results show that the possible explanation for the high variability being caused solely by different analytical methods may be rejected. Clearly, a similar variability can be found even when using a single method and a uniform measuring protocol. Moreover, such similar high variability in alloy content could be measured in different points of a single axe surface. Therefore, as a next analytical step, based upon the above results, and to better understand whether this compositional variability is reflective of a ‘true’ inhomogeneity in the original metal or is caused by other ‘non-production’ factors, such as surface preservation, changes caused by corroded parts of the surface or by the conservation treatment of the object (each factor of a combination thereof), it was decided to check the level of homogeneity in the preserved bulk volume of metal in these objects. A suitable method for collecting data about the bulk contents in a non-destructive manner is neutron diffraction analysis. The time-of-flight (TOF) method of neutron diffraction (ND) was employed in the pulse source at the ISIS, Rutherford–Appleton Laboratory (UK). Attenuation of thermal neutrons is not significant for common metals such as copper alloys. Therefore, samples with a thickness of up to several centimetres, such as MBII battleaxes, are highly suitable for such experiments. In this method, the set of measured diffraction lines corresponds to each crystal phase of the irradiated sample. The TOF values of these lines give information about the lattice constants and their intensities, both reflecting the amount of each metal phase. Typical limits of detection for minor phases are lower than 1%. The expected basic phase of tin or arsenical bronze is a face-centred cubic solid solution of Cu–Sn or Cu–As, respectively. The mean concentration of the solid solution is connected directly with the value of the lattice constant; any inhomogeneity in the solid solution causes some widening of the diffraction lines and, therefore, could be measured as well. In this study, an ENGIN-X diffractometer was used (Zhang et al. 2009, 78), which is optimized for measurements of residual stresses in different defined parts inside an object (‘gauge volume’). Diffraction measurements on the ENGIN-X were made with gauge volumes of 4 ¥ 4 ¥ 10 mm3 along the length of the objects in their central part, and also on the blades. Each scan lasts 2 h; all © 2013 University of Oxford, Archaeometry 56, 2 (2014) 279–295 288 S. Shalev et al. measurements were carried out during two full days of beam time. The analyses established the contained phases (Cu–Sn or Cu–As solid solutions, intermetallic compounds of Cu with Sn or As, a Pb phase, impurity phases such as oxides, corrosion products etc.) and inhomogeneities (segregation of Cu–Sn or Cu–As solid solutions) and any change of their concentrations throughout the objects. The ENGIN-X diffractometer has detector banks at 190° scattering angles, which are the optimal conditions to determine the lattice constants in two orientations simultaneously (for a detailed description of the method and its application in archaeological artefacts analysis, see Kockelmann et al. 2006). Here, the observed diffraction patterns, as well as calculated ones for modelling (of the existing phases and their crystal structures), were used. Both full profiles of the pattern were constructed and compared (for a detailed description of the data treatment procedure—i.e., Rietveld refinement analysis—see Larson and von Dreele 2004). RESULTS The addition of As (up to 5–6%), Sn (in the range of 10–14%) or Pb (up to 3%) to Cu lowers the melting temperature of the copper alloy, improves the casting properties of the melt (liquidation), and enables the hardening of the cast object by hammering and annealing cycles to a much harder metal than unalloyed copper (cf., Northover 1989; Shalev 1996, 12–13). Therefore, there are good metallurgical reasons for ancient metal smiths to use such alloys. The XRF results show that the duckbill axe was indeed produced from leaded tin-bronze (Table 2), but with inhomogeneous Sn and Pb distributions. The areas closer to the blade seem to be richer in the doping metals. The very high Pb content in axe BA8 could be caused by severe corrosion, although it was measured on a conserved surface. Therefore, the observation of very high Sn contents is problematic in this case, as the optimal tin concentration for achievement of a greater hardness in Cu-based solid solutions is less than 14–15%. The high Pb concentration is also unexpected, as Pb exists with Cu alloys as inclusions and, thus, in high quantities could negatively affect the mechanical properties of the metal. The results of the XRF study of flat shaft axes are also shown in the Table 2. Among these, one axe (BA5) was cast from almost pure copper. Three axes (BA7, BA12 and BA16) were cast from Pb containing As bronze and a very small amount of Sn. The As concentration is fairly close to optimal (no more than 6% in the solid solution; see Subramanian and Laughlin 1988); the observation of high Pb content in axe BA12 is again problematic. All the other axes were produced from leaded tin-bronze, and the Sn concentration and low amount of Pb are more or less reasonable in axes BA3, BA4, BA9, BA11, BA13 and BA15. However, axes BA6, BA8, BA10 and BA14 have significantly higher concentrations than optimal of Pb and Sn. Two axes (BA15 and BA16) have very elegant shapes (Miron 1992, Types 2 and 3). The other chisel-shape axes have a simple shape with an almost flat flank and top faces (Miron 1992, Type 1). It is possible to measure a change of the content from one side to the other by scanning the axes along the central line of one of the faces without being too much affected by XRF radical changes in the shape of the measuring area and exposing the same maximal flat surface where the maximal visibility of the primary beam and the detector window (the ‘measurement plane’) is achieved. The TOF-ND analyses were carried out on three flat shafthole tin-bronze axes BA3, BA6 and BA10, two chisel-shaped arsenical bronze axes BA12 and BA16 and the duckbill tin-bronze axe BA8. The TOF-ND patterns of all studied axes show three strong lines associate with the {111}, {200} and {220} family of Bragg reflections (Fig. 2), belonging to a face-centred cubic type lattice. The lattice parameters estimated from these reflections are in the range of 3.66 Å to © 2013 University of Oxford, Archaeometry 56, 2 (2014) 279–295 Middle Bronze Age II battleaxes from Rishon LeZion, Israel 289 Figure 2 Rietveld refinement analysis of the neutron powder diffraction data taken at 15 mm from the blade edge of tin-bronze axe BA6, measured on the north bank of ENGIN-X. The open circles, the continuous red line and the bottom green line are the measured data, the refined model and the difference between them, respectively. The refined model includes two Cu–Sn solid solutions along with Cu2O, CuCl and Pb. (See online for a colour version of this figure.) 3.68 Å in the Sn-bronze axes, and in the range of 3.62 Å to 3.63 Å in the As-bronze axes. These values are in good agreement with cell parameters values of the a-phases, or solid solutions, Cu–Sn (ª 8–11% Sn) and Cu–As (ª 1–3% As), respectively (Grazzi et al. 2010). In most of the diffraction patterns, additional small reflections lines are observed associated with a Pb phase and corrosion products. These lines are strongest on both ends of the studied axes (close to the edge of the blade, and in the handle). The intensities and positions of the extra lines are in agreement with the CuCl, Cu2O, SnO2 and Pb phases for the tin-bronze axes, and with CuCl, Cu2O and Pb for the arsenical copper axes. In neither the tin-bronze axes nor the arsenical bronze axes were secondary intermetallic phases found (e.g., the d-phase in the Cu–Sn phase diagram; or Cu8As or Cu3As in the Cu–As phase diagram). The ICDD database1 was used for the identification of minor phases. Close inspection of the diffraction patterns taken from all tin-bronze axes shows significant broadening of the diffraction lines of the solid solution Cu–Sn phase, collected at a more than ~20–30 mm distance from the edges blades (henceforth, ‘bulk’) compared to lines in patterns obtained right on the edges. This broadening either increases or remains the same as a function of the gauge volume distance from the blade edge of each of the three flat shafthole axes (BA3, BA6 and BA10). In contrast, the socket of the duckbill axe BA8 (approximately 63 mm from the blade edge, close to ‘eyes’) has an observed line width similar to the one observed on the edge of the axe. A detailed Rietveld analysis of all measured diffraction patterns was undertaken. The refined model used consisted of two main phases (solid solutions with different cell constants), corrosion 1 Available from the International Centre for Diffraction Data, 12 Campus Blvd, Newtown Square, PA 19073-3273, USA. © 2013 University of Oxford, Archaeometry 56, 2 (2014) 279–295 290 S. Shalev et al. Figure 3 The diffraction peak profile for {111} reflection of solid solution Cu–Sn as measured by neutron diffraction on axe BA6 at x = 15 (solid circles, one homogeneous solution) and x = 81 mm (open squares, two solutions of different cell parameters). phases (i.e., Cu2O, CuCl for all tin-bronze and arsenical copper axes, and an additional SnO2 phase for the tin-bronze axes), and a metallic Pb phase. A good agreement between the refined model and observed data is achieved for all the above diffraction patterns. It is worth mentioning here that the model of two solid solutions does not reflect the real metallographic situation of one solid solution with a continuous wide range of Sn concentrations and a correspondingly wide range of cell constants of the face-centric cubic lattice. The model used here allows us only to evaluate quantitatively the above differences of Sn concentration (and lattice constants) in the cast alloy, although it does not necessarily reflect the real metallographic state that could well be of a dendritic segregation of continuous range, and not two well-separated solid solutions as modelled. For example, the cell constant values of the two main tin-bronze a-phase solid solutions as a function of the spatial position, x, along the BA6 axe, are depicted in Figure 4. The Rietveld refinement of patterns collected at distances of up to ~30 mm from the blade edge of BA6 gives a good fit, with only one copper-tin solid solution in the model. At a distance from the blade edge—that is, in the bulk of the axe—a good refinement requires two different solid solutions (about 60% and 40%), in agreement with the observed line shapes. Moreover, the difference in the cell constants of the two solid solutions increases as a function of distance (Fig. 4). Similar behaviour is also observed in BA3 and BA10, which are both chisel-shaped tin-bronze axes. In the duckbill axe BA8, the refinement indicates one solid solution phase close to the blade, with two different solid solutions in the bulk part, and, again, with one solid solution close to the ‘eyes’. Following the calibration curve published in Grazzi et al. (2010), and using the refined values of the cell constants associated with the Cu–Sn solid solutions, we are able to estimate the tin content in the different parts of the studied axes. The average tin content for all tin-bronze axes is in the range of 7.5–12 wt% for the chisel-shaped axes, and 10.5 wt% for the duckbill-shaped axe (Table 6). In all tin-bronze axes, the average tin content is constant along the length of the axe. However, the width of the tin concentration distribution, as evidenced by the use of two Cu–Sn solid solutions, increases from edge to bulk. For example, the width of tin distribution of © 2013 University of Oxford, Archaeometry 56, 2 (2014) 279–295 Middle Bronze Age II battleaxes from Rishon LeZion, Israel 291 Figure 4 The tin-bronze refined cell constant values for axe BA6 as a function of the distance from the blade edge: the open circles and solid squares (solid diamonds and open triangles) represent data measured by the two (north and south) detectors of ENGIN-X. Table 6 Neutron diffraction results for solid solution concentrations and minor phase amounts in bronze axes (Sn in wt%; minor phases in vol%) Object % Sn (c1–c2) BA3 BA6 BA10 BA8 BA12 BA16 5–10 5–11 9.5–13.5 8.5–12.5 % As (c1–c2) 1–2.6 % Cu2O % Pb 1–3 0–6 – 0–5 – 0–1.5 <0.5 <0.5 – <1 2 <0.5 c1 and c2 are the concentrations of two solid solutions (see text). axe BA6 (Figs 3 and 4) is 5–11 wt% in the bulk, indicating a considerable degree of tin segregation. On the other hand, the sharper tin distribution on the edge clearly indicates that the axe had undergone special heat treatment during annealing and hammering of the blade area, as described above in the reconstruction of the production process. For the arsenical copper axes, the diffraction line broadening is much less pronounced, since the changes of lattice parameters for low concentrations of the doping element (here As) are also smaller. However, some segregation behaviour similar to that observed in tin-bronze axes could be detected at least in axe BA12. Together with other impurities and corrosion phases, the diffraction line of Pb was observed in TOF diffraction patterns for all the studied axes (the estimated amounts for these phases are presented in Table 6). A comparison of the neutron diffraction and the XRF measurements shows some clear differences: according to the ND results, all the axes contain no more than 2% Pb, which is an amount high enough for facilitating the wetting during casting without © 2013 University of Oxford, Archaeometry 56, 2 (2014) 279–295 292 S. Shalev et al. damaging the mechanical properties of the cast. Thus, the ND results show that the alloy contents in all the bulk parts of the analysed axes are favourable in the relative amounts of all the major doping elements (Sn or As), as well as in the relatively low Pb concentration. These results indicate that all the XRF measurements showing higher concentrations of Sn and Pb in axes BA6, BA8, BA10 and BA14 and of As and Pb in axe BA12 (Table 2) are related solely to surface layers and do not represent the true bulk composition. These measured effects of surface enrichments by all the doping elements, including lead, could reflect a change caused by surface corrosion (Ingo et al. 2006), as well as alteration in original metal ratios caused by conservation treatments. CONCLUSIONS The ND analyses of the MBII battleaxes from Rishon LeZion may be closely linked to the surface compositional analysis, as well as to the metallurgical production reconstruction based upon the archaeological remains (stone mould fragments) and metallography of different parts of similar artefact types. They all show a highly controlled alloying in both tin-bronze and arsenical copper production in more or less optimal amounts of tin (up to 14 wt% Sn) or arsenic (up to 5 wt% As). The limited amounts of lead in all analysed axes (up to 2 wt% Pb) is optimal for wetting the cast and, thus, for facilitating the pouring of metal into the mould and the proper filling of the surface details and decorations. In addition, the amounts of Pb, Sn and As are most advantageous for achieving the desired mechanical properties, as is visible mainly in the blade area. Annealing and hammering of the blade could result in greater hardness, making the axe a more efficient weapon with a hard blade for cutting and softer inner bulk that could better act as a shock absorber, reducing brittleness and, consequently, reducing the likelihood of the cracking and fracture of the axe. Another important aspect revealed by the ND results is that in all analysed tin-bronze axes, the average tin content in the solid solution Cu–Sn is constant along the length of the axe. However, the change in the width of the diffraction peaks: thinner in the blade and wider in the bulk, shows that the Cu–Sn solid solution in the blade is more homogeneous than in the bulk. This is also made visible by the occurrence of two different Cu–Sn solid solutions at the same positions in the bulk of the axes. For example, the difference in the peak shapes of axe BA6 (Fig. 3) corresponds to the difference in tin segregation (5–10 wt% in the bulk), although the average Sn values are the same. On the other hand, the lower difference in tin content of the two a-phase components on the edge indicates that the axe had undergone special heat treatment for homogenization, which is necessary for subsequent hardening of the edge by hammering. Neither in the tin-bronze axes nor in the arsenical copper axes were secondary intermetallic phases found (e.g., the d-phase in Cu–Sn; and Cu8As or Cu3As in Cu–As). The ND results also show that the bulk amount of lead in all analysed axes has not exceeded 2% Pb, which is exactly high enough for better casting and low enough not to damage the mechanical properties of the cast. Thus, the concentration of the main doping element (Sn or As), and the reasonably small Pb concentration in the bulk of all measured axes, are optimal for the casting and further annealing and hammering of the blade. The results of this neutron diffraction study afford an opportunity to establish clearly the high level of ancient technological control in the production of these axes. This sophisticated control by ancient smiths in the production of these axes stands in stark contrast to the inhomogeneity in the use of two alloys for the same axe types. This situation is well known from other sites in the southern Levant—Aphek, Kabri, Kurdani, Byblos and Tell © 2013 University of Oxford, Archaeometry 56, 2 (2014) 279–295 Middle Bronze Age II battleaxes from Rishon LeZion, Israel 293 el-Dab`a (Shenberg 1985; Philip 1995a, 2006; Shalev 2000, 2002; El Morr and Pernot 2011)— and strengthens the assumption that the ancient metal smiths who produced the artefacts from Rishon LeZion used metals that were available at a particular time and place (Shalev 2009). These special well-defined axe shapes (Types 1–3) are maintained despite the use of different compositions, a fact that may imply their important role for southern Levantine society in this period (Philip 1995a, 77; Philip 2006, 140; El Morr and Pernot 2011, 10–11). This significance is correlated with their occurrence in mortuary contexts and usage as grave offerings and/or equipment for warrior burials (Hallote 1995, 2002; Ilan 1995; Philip 1995a,c; Doumet-Serhal 2003; El Morr and Pernot 2011). The comparison of the ND data with surface XRF results of the same axes shows a clear difference in the measured ratios of alloys (Sn, As and Pb). This discrepancy is probably caused by the measuring of partially corroded surfaces. If this is indeed the case, then the conservation procedures of cleaning and conserving ancient metals in the laboratory could well affect the surface composition, a fact that has to be considered when analysing the surface of a ‘cleaned’ and conserved ancient metal. This effect is particularly important when the quantitative results of surface analysis data are being evaluated for interpretation. The metallurgical results presented here, deriving from totally non-destructive ND analysis, show a close agreement with metallographic and compositional analyses of similar objects. Thus, the potential exists for drawing more and more metallurgical and metallographic data without the need to drill or cut samples. ACKNOWLEDGEMENTS The authors wish to thank the IAA for allowing us to take these axes from Israel to the UK to be analysed at ISIS. We wish to thank ISIS for enabling us to use their ND state-of-the-art facilities and experts. We wish to convey our special thanks to the anonymous referee whose comments showed the high credit he or she gave to our deep knowledge in metallurgy, probably higher than what we really have, and the comments that forced us to stress, better and to a slightly deeper extent, the relations between the ND data and the overall metallurgical and specific archaeometallurgical knowledge. This referee clearly deserves to be acknowledged as a crucial component of the research carried out in this paper, and we sincerely hope for future co-operation with him or her in studying this aspect further. REFERENCES Branigan, K., McKerrell, H., and Tylecote, R. F., 1976, An examination of some Palestinian bronzes, Journal of the Historical Metallurgy Society, 10(1), 15–23. Budd, P., Pollard, A. M., Scaife, B., and Thomas, R. G., 1995, Oxhide ingots, recycling and the Mediterranean metal trade, Journal of Mediterranean Archaeology, 8(1), 1–3. Doumet-Serhal, C., 2003, Weapons from the Middle Bronze Age burials at Sidon, Archaeology & History in the Lebanon, 18, 38–57. Dunand, M., 1939, Fouilles de Byblos I, 1926–32, Paris. Dunand, M., 1954, Fouilles de Byblos II, Part I, Paris. El Morr, Z., and Pernot, M., 2011, Middle Bronze Age metallurgy in the Levant: evidence from the weapons of Byblos, Journal of Archaeological Science, 30, 1–20. Gerstenblith, P., 1983, The Levant at the beginning of the Middle Bronze Age, Dissertation Series No. 5, American Schools of Oriental Research, Winona Lake, IN. Goldin, J., Levy, T. E., and Hauptmann, A., 2001, Recent discoveries concerning Chalcolithic metallurgy, Journal of Archaeological Science, 28, 951–63. Grazzi, F., Bartoli, L., Siano, S., and Zoppi, M., 2010, Characterization of copper alloys of archaeometallurgical interest using neutron diffraction: a systematic calibration study, Analytical and Bioanalytical Chemistry, 397, 2501–11. © 2013 University of Oxford, Archaeometry 56, 2 (2014) 279–295 294 S. Shalev et al. Hallote, R. S., 1995, Mortuary archaeology and the Middle Bronze Age southern Levant, Journal of Mediterranean Archaeology, 8(1), 93–122. Hallote, R. S., 2002, Real and ideal in Middle Bronze Age tombs, Near Eastern Archaeology, 65(2), 105–11. Ilan, D., 1995, Mortuary practices at Tel Dan in the Middle Bronze Age: a reflection of Canaanite society and ideology, in The archaeology of death in the ancient Near East (eds. S. Campbell and A. Green), 117–39, Oxbow Books, Oxford. Ingo, G. M., DeCaro, T., Riccucci, C., Angelini, E., Grassini, S., Balbi, S., Bernardini, P., Salvi, D., Bousselmi, L., Cilingiroglu, A., Gener, M., Gouda, V. K., Al Jarrah, O., Khosroff, S., Mahdjoub, Z., Al Saad, Z., El-Saddik, W., and Vassiliou, P., 2006, Large scale investigation of chemical composition, structure and corrosion mechanism of bronze archaeological artefacts from Mediterranean basin, Applied Physics A, 83, 513–20. Kan-Cipor Meron, T., 2003, Middle Bronze Age metal objects from the Rishon Le-Zion Cemetery, Unpublished M.A. thesis, University of Haifa. Khalil, L. 1980, The composition and technology of ancient copper alloy artifacts from Jericho and related sites, Unpublished Ph.D. thesis, Institute of Archaeology, London. Kockelmann, W., Siano, S., Bartoli, L., Visser, D., Hallebeek, P., Traum, R., Linke, R., Schreiner, M., and Kirfel, A., 2006, Applications of TOF diffraction in archaeometry, Applied Physics A, 83, 175–82. Larson, A. C., and von Dreele, R. B., 2004, General Structure Analysis System (GSAS), Los Alamos National Laboratory Report LAUR, 86–748, Los Alamos, NM. Le Roux, G., Véron, A., Scholz, C., and Doumet-Serhal, C., 2003, Chemical and isotopical analyses on weapons from the Middle Bronze Age in Sidon, Archaeology & History in the Lebanon, 18, 58–61. Miron, E., 1992, Axes and adzes from Canaan, Praehistorische Bronzefunde IX, 19, Stuttgart. Northover, P., 1989, Properties and use of arsenic–copper alloys, in Old World archaeometallurgy (eds. A. Hauptmann, E. Pernicka and G. A. Wagner), 113–18, Selbstverlag des Deutschen Bergbau-Museums, Bochum. Philip, G., 1989, Metal weapons of the Early and Middle Bronze Ages in Syria–Palestine, BAR International Series 526, British Archaeological Reports, Oxford. Philip, G., 1991, Tin, arsenic, lead: alloying practices in Syria–Palestine around 2000 B.C., Levant, 23, 93–104. Philip, G., 1995a, Tell el-Dab`a metalwork patterns and purpose, in Egypt, the Aegean and the Levant (eds. W. V. Davies and L. Schofield), 66–83, British Museum Press, London. Philip, G., 1995b, The same but different: a comparison of Middle Bronze Age metalwork from Jericho and Tall ad-Dab`a, Studies in the History and Archaeology of Jordan, V, 523–30. Philip, G., 1995c, Warrior burials in the ancient Near-Eastern Bronze Age: the evidence from Mesopotamia, western Iran and Syria–Palestine, in The archaeology of death in the ancient Near East (eds. S. Campbell and A. Green), 140–54, Oxbow Books, Oxford. Philip, G., 2006, Tell el-Dab`a XV: metalwork and metalworking evidence of the Late Middle Kingdom and the Second Intermediate Period, Verlag der Österreichischen Akademie der Wissenschaften, Wien. Shalev, S., 1996, Archaeometallurgy in Israel: the impact of the material on the choice of shape, size and color of ancient products, in Archaeometry 94: the proceedings of the 29th International Symposium on Archaeometry (eds. Ş. Demirci, A. M. Özer and G. D. Summers), 11–15, Tübitak, Ankara. Shalev, S., 2000, Metal artifacts: archaeometallurgy, in Aphek—Antipatris I (eds. M. Kochavi, P. Beck and E. Yadin), 278–87, Tel Aviv University, Tel Aviv. Shalev, S., 2002, Metal artifacts, in Tel Kabri—the 1986–1993 excavation seasons (ed. A. Kempinski), 307–18, Tel Aviv University, Tel Aviv. Shalev, S. 2007, Metallurgical analysis, in The Middle Bronze Age IIa cemetery at Gesher: final report (eds. Y. Garfinkel and S. Cohen), 109–14, American School of Oriental Research, Boston, MA. Shalev, S., 2009, Metals and society: production and distribution of metal weapons in the Levant during the Middle Bronze Age II, in Techniques and people: anthropological perspectives on technology in the archaeology of the proto-historic and early historic periods in the southern Levant (eds. S. Rosen and V. Roux), 69–80, De Boccard, Paris. Shalev, S., 2010, Metal objects from Fassuta, `Atiqot, 62, 43–9. Shalev, S., and Braun, E., 1997, The metal objects from Yftah’el II, in Yiftah’el (ed. E. Braun), 92–6, Israel Antiquities Authority Reports, No. 2, Jerusalem. Shalev, S., Shilstein, S. Sh., and Yekutieli, Y., 2006, XRF study of archaeological and metallurgical material from an ancient copper-smelting site near Ein-Yahav, Talanta, 70, 909–13. Shenberg, C., 1985, Analysis of the weapons by X-ray fluorescence method, `Atiqot, 17, 112. Subramanian, P. R., and Laughlin, D. E., 1988, The As–Cu (arsenic–copper) system, Bulletin of Alloy Phase Diagrams, 9, 605–17. © 2013 University of Oxford, Archaeometry 56, 2 (2014) 279–295 Middle Bronze Age II battleaxes from Rishon LeZion, Israel 295 Yogev, O., 1985, A Middle Bronze Age cemetery south of Tel Rehov, `Atiqot, 17, 90–113. Zhang, S. Y., Godfrey, E., Kockelman, W., Paradowska, A., Bull, M. J., Korsunsky, A. M., Abbey, B., Xu, P., Tomota, Y., Liljedahl, D., Zanellato, O., Fitzpatrick, M., Daymond, M. R., Toda, R. M., Holt, R. A., Kelleher, J., Siano, S., and Sansitisteban, J., 2009, Review: high-tech composites to ancient metals, Materials Today, 12, 78–84. Ziffer, I., 1990, At that time the Canaanites were in the land: daily life in Canaan in the Middle Bronze Age 2, 2000–1550 B.C.E., Eretz Israel Museum, Tel Aviv. © 2013 University of Oxford, Archaeometry 56, 2 (2014) 279–295