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