Emergence of copper pyrotechnology in Western Asia

# Abstract The Chalcolithic Southern Levant (4500 to 3800 BCE) is especially well-known for its outstanding copper objects, such as the crowns and mace heads, found, among others, in the Nahal Mishmar Hoard as the largest and most prominent assemblage. They are made in the lost wax casting technique with polymetallic copper alloys, whose ore sources are located in the Anatolian or Southern Caucasian mountains. The combination of this metal type, exclusively used in the Chalcolithic Southern Levant, and the earliest evidence for this technologically complex casting process in West Asia attest to a unique metallurgical development in the Chalcolithic Southern Levant. Based on promising archaeological finds, Fazael was suggested as tentative production site. This metal working tradition was paralleled by an unalloyed copper metallurgy with production sites confined to the Nahal Beer Sheva, where copper ores predominantly from Faynan were smelted and the unalloyed copper was cast in open moulds to tool-shaped objects. The major aspects of the lost wax casting process and its mould design are understood since many years from the study of mould remains attached to the metal objects and of the polymetallic copper alloys itself. However, the absence of in-situ production remains (e.g., furnaces, mould fragments) made it impossible to gather more knowledge about the operational sequence. In addition, archaeological evidence for the production and processing of the polymetallic copper alloys remains to be found; the few ore pieces with compatible chemistry in the Southeastern Anatolian sites Arslantepe and Norşuntepe were apparently not used in smelting activities. Moreover, the contrast in the preservation of production remains from the two Southern Levantine metallurgical processes seems odd, even if the lost wax casting technology is generally somewhat ephemeral. In contrast to the lost wax casting process, the unalloyed copper process can be reconstructed in more detail. Furnace remains from Abu Matar and Shiqmim indicate the use of pit furnaces, in which the ore was smelted. In a second step, the copper prills were mechanically extracted, melted in crucibles and then cast in open moulds. Nevertheless, important details remain unclear, such as the draught technique. Additionally, some aspects of the current process reconstructions seem very impracticable and thus questionable. Beside these technological aspects of the Chalcolithic metallurgy in the Southern Levant, the origin and evolution of the innovation "lost wax casting" was not investigated in detail, yet. While many studies on the cultural developments and the metallurgies in the Chalcolithic Southern Levant and the other West Asian regions exist, an inter-regional perspective was rarely taken despite the clear connection between these regions by the polymetallic copper alloys and other objects. Part of the discussion about the metallurgy of the Chalcolithic Southern Levant is also the role of the metal objects in the society and their sudden disappearance at the end of the Chalcolithic. Although several studies tackled this topic already, it remains debated, not least because the general social organisation of the Chalcolithic Southern Levant remains debated. For these reasons, this project addresses three main aims: (a) Refining the current reconstructions of the metallurgical processes in the Chalcolithic Southern Levant by combining the (re)analysis of the metallurgical assemblages from Abu Matar and Fazael with an experimental approach; (b) Tracing the evolution of the innovation "lost wax casting" in an inter-regional perspective, and (c) providing new ideas about the cultural role of the metal objects for the social system of the Chalcolithic Southern Levant. Beside the experimental and analytical work for the first part, an extensive literature review is the base for all three aims. ## Archaeological background The Chalcolithic Southern Levant is a continuation of the Late Neolithic and covered today's Israel as far South as the Northern Negev, the Westbank, and the Jordan valley. It can be subdivided in two phases, termed here Early and Late Chalcolithic. The Early Chalcolithic (4500--4300 BCE) is characterised, among others, by cornets, ceramic vessel figurines, and large architectural structures ("temples") in Gilat, En Gedi, and Teleilat Ghassul. The Late Chalcolithic (4300--3800 BCE) is characterised, among others, by the vanishing of those traits and the emergence of metallurgy. However, many aspects in the material culture remain unchanged, such as secondary burials (often ossuaries placed in caves), the lack of individually assignable burial items, stone mace heads, V-shaped bowls, and basalt bowls. Many of the metal items seem to be skeuomorphs or imitations of non-metal objects, establishing a close link between this new and the longer-used materials. Reconstructing the social organisation of the Chalcolithic Southern Levant is challenging because neither the settlements nor the burials provide clear indicators for status differences. Therefore, models of hierarchical chiefdoms exists as well as heterarchical models with heads of households. At the end, the Chalcolithic ideology and with it the entire social system seems to collapse and most of the prestige items including the lost wax cast polymetallic copper alloys disappeared with the onset of the Early Bronze Age. The Chalcolithic in the Southern Levant is a period with many innovations. Beside horticulture and the full establishment of secondary products, the slow-turning potter's wheel and metallurgy are probably the most important ones. The innovations are paralleled by a significant increase in craft specialisation with dedicated workshops for, e.g., flint tools, basalt bowls, and metal items, as well as an increase in the standardisation of pottery, especially of V-shaped bowls. Exchange was organised in a two-tiered system. A network connecting the entire region and expressed, e.g., by a uniform _chaîne opératoire_ in pottery production was overlaid by networks in the northern and southern half, characterised by perforated flint discs and unalloyed copper items plus ivories, respectively. In addition, contacts to the neighbouring cultural entities such as the Timnian in the Negev desert and to regions beyond the immediate neighbours existed. These contacts were most likely maintained by mobile parts of the population, either fully nomadic groups such as the Timnians or mobile herders from the sedentary groups. Moreover, archaeogenetic studies suggest two independent immigration waves before or during the Chalcolithic from the North. The Southeastern Anatolian and Northern Mesopotamian region splits into several groups at the end of the Ubaid (mid-5th millennium BCE), which again can be grouped by their pottery into an Eastern and a Western group, separated by the Euphrates river. At the end of the 4th millennium BCE, the Western group splits into a northern group and a southern group, and the latter orientates itself closer to the Eastern group. A unifying element of all groups is the production of chaff-faced ware and the serial production of so-called Coba bowls. While sites in the Eastern group such as Tepe Gawra have monumental architecture, a vertically stratified society, and indicators for early urbanism, similar developments are attested in the Western group with Arlsantepe as its most important site only from the mid-4th millennium BCE on. For the time under study here, no clear evidence for a stratified society can be provided in the Western group. The vertically stratified societies in the Eastern group and later in the Western group is based on staple finance, i. e. the control over the access to and distribution of resources, most importantly food. Compared to the other regions, metal items are rare in Northern Mesopotamia and Southeastern Anatolia. From the second half of the 5th millennium BCE on, apparently only small copper tools were produced and they were probably communal items. Except for three burials with small gold and silver items, burials from the period under study do not contain metal in this region. Cultural processes in the Southern Caucasus are still difficult to reconstruct. Most parts of the Sioni culture were assumingly mobile groups with too ephemeral remains to be readily recognisable in the archaeological record, and only the settlements of Mentesh Tepe and Ovçular Tepesi were investigated in detail, yet. In general, the Southern Caucasus appears to be strongly influenced by the regions to its South; chaff-faced ware is often found together with local pottery. Smelted copper is attested since the 6th millennium BCE but clear evidence for metallurgy dates only to the second half of the 5th millennium BCE. Some of the few excavated graves contained metal items. A reconstruction of the social organisation is not possible yet, but an organisation in large family groups was suggested. The situation changes markedly during the second quarter of the 4th millennium BCE. The Leilatepe-Berikldeebi culture features metal workshops and monumental architecture. In addition, the first kurgans are erected. They contain many metal and other prestige items, indicating a vertically stratified society at this time. Iran is characterised by small dispersed settlements without social stratification and minor socio-economic inequality at the end of the Neolithic. Processes during the Chalcolithic are similar in all regions but happen in a different pace. Northwestern Iran is part of the Southern Caucasus-Northern Mesopotamia cultural sphere and cultural developments follow the processes there. The North Central Iranian Plateau sees a strong increase in craft specialisation, including workshop quarters with dedicated spaces for, e.g., metallurgy and pottery. Central buildings indicate some kind of coordinating entity but there seems to be no vertical stratification of the society. Examples for long-distance exchange are moulds for shaft hole axes with the corresponding axes found in the cemetery of Susa in Southwest Iran, or lapis lazuli. Metallurgy, the two-chambered pottery kiln, and textile production in Iran is first evidenced in Southeastern Iran, from where it spreads to the other regions during the second half of the 5th millennium BCE. Apart from these innovations, this region adheres to the Neolithic settlement structure and social organisation. The Zagros Mountains seem to be populated at this time by mobile groups with large central cemeteries. Some of the burials contain metal items. In all these regions, items made of metal and exotic stones seem to be used to indicate individually assigned status. A strong contrast provides Khuzestan in Southwest Iran, where monumental architecture for public and sacral purposes in Choga Mish and Susa, and elite residences in Susa indicate a development towards a vertically stratified society. The cemeteries of Susa and Chega Sofla yielded large amounts of metal items. The finds in Chega Sofla indicate a high skill of metal working unparalleled in the other Iranian regions at this time. Southeastern Europe is included here due to the earliest evidence for copper smelting, dating to the Vinča culture (6th millennium BCE) in modern day Serbia. In the mid-5th millennium BCE this culture collapses. Settlement activities and metallurgy shifts to the Western Black Sea coast with the cemetery of Varna I being probably the most prominent site. Metallurgy in this region is characterised by a large number of heavy copper tools (more than 4300 items) and a high innovativeness with the earliest gold finds, a short episode of "natural" bronze production from stannite, evidence for the alloying of copper and gold, and lost wax casting with gold during the mid-5th millennium BCE. Similar to the Chalcolithic Southern Levant, there is no indication for social stratification in the settlements and burials, despite the large amounts of gold and copper in burials and hoards. All of these regions were in contact with each other. Archaeological evidence for contacts between the Southern Levant and its North can be traced back as early as the Natufian (13 to 9.6 ka BCE) by Anatolian obsidian in the Southern Levant. The material remains of the Wadi Rabah culture (6th millennium BCE) have many traits that can be related to the Halaf cultural sphere in the North. Similar, Tel Tsaf (first half of the 5th millennium BCE) yielded many indicators for contacts with the Ubaid cultural sphere. During the Chalcolithic, contacts are attested by, e.g., Anatolian obsidian in Southern Levantine sites. Fan scrapers indicate a technological horizon stretching from Egypt over the Southern Levant to Northern Mesopotamia. Diversity and possibly intensity of exchange increases with the Late Chalcolithic: the polymetallic copper alloys and Canaanean Blades are clearly an import from Southeastern Anatolia/Northern Mesopotamia or the Southern Caucasus, while the slow-turning potter's wheel is a Southern Levantine innovation that spread towards the North. Contacts with regions even further away are attested by, e.g., lapis lazuli beads in Southern Levantine cave sites. Moreover, strong similarities in the social organisation of Southeastern Europe and the Southern Levant were noted. Nilotic shells in Southern Levantine sites evidence contacts between the Southern Levant and Egypt. In addition, the presence of a Southern Levantine population in Egypt is evidenced by vessels manufactured from local clay in the Chalcolithic Southern Levantine pottery tradition in Buto and a subterranean house in Maadi. However, extensive exchange between Egypt and the Southern Levant is best attested in Tall Hujayrat al-Ghuzlan at the Gulf of Aqaba. This site yielded evidence for extensive copper production and shipment of copper ingots to Egypt. It is culturally unrelated to the Southern Levantine Chalcolithic. Contacts between the Southern Caucasus and the regions to its South are close since the Neolithic, as Caucasian obsidian in Anatolian sites and Halaf pottery in Caucasian sites indicate. During the period under study here, the Southern Caucasus, Northern Mesopotamia and Southeastern Anatolia are part of the chaff-faced ware technological horizon. It was suggested that differences in the material culture indicate differences between mobile highland communities with "Caucasian" material culture and lowland "Mesopotamian" settlers rather than between separate geographical regions. The emergence of the Kura-Araxes phenomenon during the first half of the 4th millennium BCE seems to be a major disruptor of the contacts between the three regions. So-called Dalma pottery attest contacts of the Southern Caucasus with Iran. In addition, shaft-hole axes were produced in both regions and their occurrence dates to roughly the same time. These axes appear already a bit earlier in Southeastern Europe and are therefore a strong indicator for contacts with this region. Southeastern Europe was also in close contact with Anatolia, albeit mostly with Western Anatolia. The distribution of ring-shaped idols indicate an exchange network spanning from Southeastern Europe to Northern Anatolia. However, the mountains between Eastern and Northern Anatolia seem to be a cultural border. Evidence for contacts with Southeastern Europe is missing in East Anatolia and, vice versa, there is no evidence for contacts with the Southern Caucasus in Northern Anatolia. ## Archaeometallurgical background The earliest known evidence for pyrometallurgy was found in Southeastern Europe, where unalloyed copper is smelted and manufactured to awls, axes, hammers and other objects since the early 5th millennium BCE. Already in the mid-5th millennium BCE, smelting activities were extensive with large amounts of heavy tools such as axes and hammers (around 4.7 t of copper in total). Compared to this huge amount of metal, the number of known smelting sites is astonishingly small. Mixtures of green and black or violet minerals were used as ore. The ores were smelted under relatively oxidising conditions in pit furnaces lined with pottery sherds, and the process yielded almost no slag. For a short period, stannite was smelted as well, resulting in tin bronze. The metal was then cast under oxidising conditions. The earliest known smelted copper in Anatolia is the large chisels, axes and needles of Mersin-Yumuktepe (around 5000 BCE), while the earliest smelting sites date only to the first half of the 5th millennium BCE. Throughout the period under study here, there is almost no change in the smelting process. Copper ore, usually weathered sulphide ore, is smelted in crucibles, which were often placed in a pit. Subsequently, the copper prills were melted in crucibles and cast in open moulds. Beside unalloyed copper, copper with elevated levels of arsenic and sometimes nickel was produced. In the Southern Caucasus, the earliest known metal finds are beads made of arsenic copper from the 6th millennium BCE. However, the earliest secure evidence for smelting activities dates only to the second half of the 5th millennium BCE. An ore pile found in Mentesh Tepe indicates the smelting of weathered copper sulphides from ophiolite-hosted ore deposits. Slag finds indicate relatively oxidising conditions. Tuyère remains from Ovçular Tepesi indicate the use of blowpipes. Furnaces are yet to be found while crucible fragments were uncovered, indicating that smelting and melting were carried out in crucibles. There is only minor change in the metallurgy of the Southern Caucasus during the period under study. The earliest evidence for copper smelting in Iran dates to the first half of the 5th millennium BCE. In Tal-i Iblis, weathered polymetallic ores were smelted in crucibles. The smelting process was more reducing than in all other regions under consideration. From early on, arsenic copper is produced alongside unalloyed copper. Larger copper items are found from the 5th millennium BCE onwards, especially in the Zagros Mountains and the neighbouring Iranian regions. Advanced casting techniques are attested by the moulds for shaft-hole axes in Tepe Ghabristan, which feature movable ceramic cores for the creation of the shaft holes. As mentioned above, the largest metal assemblages were found in Susa and Chega Sofla in Khuzestan. However, smelting sites are yet to be found in Khuzestan and it seems more probable that copper was imported from the regions further North, e.g., the North Iranian Plateau. Being one of the central topics in this study, the metallurgy of the Chalcolithic Southern Levant is extensively discussed. As already indicated, two metallurgical traditions existed: polymetallic copper alloys cast in the lost wax technique, and unalloyed copper metallurgy with open mould casting. Evidence for unalloyed copper metallurgy is confined to the Nahal Beer Sheva. The vast majority of ores was mined in Faynan and brought to the settlements for smelting. A few ore pieces were brought from Timna. Stone anvils and crushing stones with traces of copper minerals found in Abu Matar indicate the beneficiation of ores. The copper ore was directly smelted in a pit furnace of about 30 to 40 cm diameter and about 20 to 30 cm depth with a collar-shaped furnace wall of about 10 cm height. The opening in the furnace wall was about 10 cm wide. The ores were rarely fully liquefied in this furnace and the smelting products are best labelled as reacted ore rather than slag. The widespread occurrence of delafossite and cuprite indicate rather oxidising conditions, similar to the other West Asian regions except Iran. Due to the short and incomplete melting of the ore, metallic copper was present as copper prills inside the reacted ore pieces. In the next step, they were mechanically extracted, melted in crucibles and cast into objects. The crucibles are usually oval bowls with a diameter of about 10 cm and an inner depth of about 7 cm. Casting moulds are yet to be found, suggesting the use of sand moulds. Several tuyère fragments are mentioned but except one from Abu Matar no details about them are published and identification of said fragment as tyuère fragment received justified criticism. For the lost wax casting process with polymetallic copper alloys, no archaeological finds concerning the production of this metal type exists, yet. Based on the chemistry (up to 25 % Sb and 15 % As, sometimes several per cent of Pb, Ag, or Ni) antimony-rich fahl ores were suggested as ore source. Lead isotope analyses point towards the Southern Caucasus and Southeastern Anatolia as source regions. Technological investigations especially of the metal objects in the Nahal Mishmar Hoard revealed that some were cast over cores usually made of ceramic but at least in one case also stone. Further, they revealed a very heterogeneous casting quality of these objects and the occasional repair of casting errors in the cast-on technique, probably directly after the mould was removed. Mould remains adhering to the metal objects allowed to reconstruct a multi-layered mould design with at least two layers. The inner layer is often made of a paste conssiting of clay from the Moza formation, organic temper and carbonaceous sand. The outer layer is made of either ferruginous clay or lime plaster. The outcrops of suitable raw materials for the moulds suggests a location of the lost wax casting workshop(s) somewhere in the Lower Jordan valley. During the planning stage of this project, Fazael in the Middle Jordan valley was suggested as potential production site due to the co-occurrence of crucible fragments -- the first ones outside the Nahal Beer Shea sites -- and a large number of fragmented lost wax cast items in its sub-site Fazael 2. Tall Hujayrat al-Ghuzlan and Tall al-Magass at the Gulf of Aqaba revealed a proto-industrial scale of copper production with ores from Timna. These sites are not part of the Chalcolithic Southern Levant cultural sphere, as mentioned above. They differ in their copper metallurgy by the use of crucibles rather than furnaces for smelting. In addition, it seems that copper ingots were the main product at these sites while proper ingots (in contrast to metal lumps) are absent in the Chalcolithic Southern Levant. Similar to copper metallurgy, the earliest evidence for the use of gold is found in Southeastern Europe. Although not the oldest site, the cemetery of Varna I (mid-5th millennium BCE) is the most famous one because of the large number of gold items found here. Gold was manufactured into personal ornaments, ring-shaped idols, miniature tools, and staff/sceptres. Most of the objects were cast, a few in the lost wax casting technique. As a result, this site provides also the earliest evidence for this casting technology. In Western Asia, gold items occur slightly later as single small objects such a wires or beads, often in combination with items made of other exotic materials. Compared to the large spatial extent in the area, they are very rare. Important examples are the gold bead in a burial in Tepe Gawra and another one in a burial in Grai Resh, both dating towards the end of the 5th millennium BCE. About the same time date the few gold items found in Chega Sofla. A golden lost wax cast bead in Tepe Hissar dates to the first half of the 4th millennium BCE. The only gold items of the Chalcolithic Southern Levant are the two gold and six electrum rings from the Nahal Qanah burial cave. The earliest gold in the Southern Caucasus dates to the second quarter of the 4th millennium BCE, among them two massive lost wax cast bulls in the Maikop kurgan. With the gold mine in Sakdrisi and a crucible in the close-by settlement of Dzedzvebi (both mid-4th millennium BCE), this region yielded the earliest evidence for the mining and smelting of gold ore. Silver objects are reported from Southeastern European sites contemporaneous with the earliest gold finds. However, their contexts are unclear and the earliest securely datable silver items in this region date to the second half of the 4th millennium BCE. They are pre-dated by finds from Southeastern Anatolia, Northern Mesopotamia, and Iran. Silver earrings found in a burial in Hacınebi date to the first half of the 4th millennium BCE, as does the first evidence for cupellation, found in Southeastern Anatolia. From the mid-4th millennium BCE on, cupellation is attested in several sites in Northern Mesopotamia and Iran. The earliest silver finds of the Southern Caucasus were found in the kurgans, among them two silver equivalents of the golden lost wax cast bulls in the Maikop kurgan. In the Southern Levant, silver is not found before the Early Bronze Age. In addition to the archaeometallurgical background, key aspects of the smelting processes with unalloyed copper ore and fahl ores as well as the properties of polymetallic copper alloys are discussed. The draught technique, central topic in the re-assessment of the Chalcolithic unalloyed copper process in the Southern Levant, can be reconstructed from the tuyères' inner diameters. Experimental work and thermodynamic calculations showed that a diameter of 5 to 10 mm is optimal for blowpipes, while tuyères with 20 to 30 mm inner diameter are optimal for bellows. It is highlighted that with the technology and furnaces available at this time, a reducing atmosphere and a complete melting of the ore can rarely be achieved. Instead, a solid-state reaction of the copper minerals to metallic copper happened. The use of weathered sulphide ores is beneficial for the process because the exothermic reaction of sulphur with oxygen increases the temperature and the reducing atmosphere in the furnace at least locally. Experimental studies showed that smelting of (weathered) fahl ores without prior roasting can result in the production of unalloyed copper alongside speiss of a composition similar to the polymetallic copper alloys found in the Chalcolithic Southern Levant. Examples from other geographical areas showed that polymetallic alloys with such high levels of antimony and arsenic are exclusive to the Chalcolithic Southern Levant. An exception are ingots from the Bronze Age Eastern Alps. They have a comparable chemical composition but were alloyed with unalloyed copper and melted under oxidising conditions before casting the metal, significantly reducing the amount of arsenic and antimony in the cast items. All other finds with a comparable chemical composition have either less than 10 % Sb or very low arsenic levels. Arsenic, and to some extent antimony, is extremely volatile under oxidising conditions. This makes the production of arsenic copper and polymetallic alloys not straightforward. Several theoretical and experimental approaches are presented but the best method with the technological knowledge of this time seems to be the production of arsenic copper through the direct smelting of heavily weathered fahl ores or the alloying of unalloyed copper with speiss. Arsenic and antimony decrease the smelting point of copper significantly, increasing its castability. Arsenic also increases the hardness and plasticity of copper significantly, while high concentrations of antimony make copper brittle to an extent that it cannot be mechanically worked any more. Consequently, the polymetallic copper alloys used in the Chalcolithic Southern Levant could only be used for casting decorative and status-communicating items. Depending on the different levels of the alloying elements in a metal, its colour differs. For example, depending on the levels of antimony and arsenic in the copper, the colours of polymetallic copper alloys range from copper to haematite-like, golden, and silver-like colours. The different colours of metal alloys was not only used by past metallurgists as indicator for the composition of the metals but was also deliberately exploited for aesthetic purposes, e.g., by the arrangement of differently coloured gold beads in a necklace in the Varna I cemetery. C. S. Smith concluded that aesthetic properties of new materials were always exploited before their usability for tools. Metal is not different from other materials in this regard. Consequently, the role of metal objects in past societies should be investigated as one material under many and not as superior to e.g. pottery or stone. The early metal items in Southeastern Europe are good examples for such a focus on the sensory properties of metals. They were embedded in an aesthetic concept of shininess together with graphitised pottery. It seems that the occurrence of many different metal types with different shiny colours (e.g. silver) over the 4th millennium BCE in Western Asia was motivated by the same pursuit for shiny materials. The example of the South American Muisca show that it is not necessarily the metal items that are of importance, but it can also be the production process (in this case lost wax casting) as the focal point of a ritual. ## Lost wax casting ### Visibility in the archaeological record Despite the large number of Chalcolithic sites in the Southern Levant, all mould remains so far were found attached to metal objects in protected places such as burial caves. As a result, it remains unclear whether mould remains at the production sites are not preserved, were found but not recognised as such or are yet to be found. Although previous studies identified a multi-layered design, the use of chaff temper, and the occasional use of plaster as key characteristics of the moulds, most of these features can only be recognised with petrographical methods and not readily in the field. Therefore, an archaeological experiment was carried out based on knowledge from previous studies and complemented by ethnographical records to investigate the recognisability of lost wax casting moulds in the archaeological record. Because layers made of different clays can be easily identified in thin sections, the recognisability of layers with the same clay but with different amounts and proportions of temper was investigated. The archaeological experiment was a multi-purpose experiment. Raw materials comparable to the materials used in the Chalcolithic were used to re-create the lost wax casting moulds. The moulds were then heated to remove the wax. At the same time, crucibles and a furnace were built according to the reconstructed furnaces of the unalloyed copper process to test their operability with bellows (see below) and to melt the copper and antimony for casting. Four runs were carried out, in none of them was enough metal melted for a successful casting. The moulds were fragile right after casting and their fragility increased significantly in the next couple of days. Complete cross-sections of two moulds as well as partial sections of the other moulds were prepared for petrographic examination. Recorded trampling and submerging experiments with mould pieces and other metallurgical ceramics were carried out to simulate mechanical stress and their interaction with water. Because none of the archaeological moulds showed interaction with the molten metal and the archaeological and experimental ceramic pastes are overall comparable, the experimental moulds are viable analogues of the archaeological ones. The post-experiment increase in the fragility of the moulds is best explained by the rehydration of the carbonaceous sand after heating and the associated volume increase. Crushing the moulds and sectioning them for petrography revealed that the inner and outer layer can be easily separated, masking their original multi-layered design. Petrographic analyses showed that the different layers can only be recognised by the different temper mixtures of the clay pastes. The alteration tests revealed that the mould fragments are stable when submerged for several hours but that one step is often enough to crush them into tiny crumbs, which are not recognisable in the archaeological record any more. Consequently, mould remains in archaeological sites must be expected to be rare. If preserved, they are friable and potentially rounded pieces of low-fired ceramics and, therefore, might be easily mistaken for heated soil crumbs. An additional universally applicable characteristic of the lost wax casting moulds is suggested: the mixture of vegetal and mineral temper. The validity of this criterion is underpinned by general technological considerations. Pottery in the Late Chalcolithic Southern Levant is made exclusively with mineral temper, and metallurgical ceramics exclusively with chaff temper. In case of pottery, this temper choice maximises mechanical strength and reduces porosity of the vessels. For metallurgical ceramics, a high porosity is desirable to increase heat insulation and to prevent spreading of cracks. Lost wax casting moulds had to combine both properties: The comparably thick moulds had to be stable enough to be heated from the outside to sufficient temperatures inside. At the same time, the high porosity especially in the inner layer prevents the spreading of cracks while allowing the air in the mould to escape during casting. With this criterion at hand, mould fragments can be reliably identified regardless of their multi-layered design, as long as archaeologists are aware to collect seemingly inconspicuous pieces of heated sediment or low-fired ceramics. ### The meallurgical assemblage of Fazael Fazael is located along the riverbank of the Wadi Fazael and consists of several Chalcolithic broad room houses and other archaeological features. Excavations in the broad room house Fazael 2 yielded many fragments of lost-wax cast and unalloyed copper items together with several crucible fragments and heated sediment nodules and, thus, was suggested to having been a lost wax casting site. Metallurgical installations such as furnaces are yet to be found, though. The broad room Fazael 5 yielded a small hoard, consisting of a head-shape standard into which a chisel, an awl and a spiral-shaped object were shoved. The broad house Fazael 7 has walls preserved to a height of more than 1 m and contained many metal fragments as well as complete metal items. Fazael 2 was radiocarbon dated to 4000 to 3900 BCE, i.e. to the very end of the Southern Levantine Chalcolithic. This date is supported by an incomplete Chalcolithic material culture and the appearance of elements typical for the Early Bronze Age, such as Canaanean Blades. Due to the already fragmented state of the many metal objects, a representative selection of them could be sectioned for metallographical investigation including SEM-EDX. In addition, several crucible fragments and heated sediment nodules were sectioned for petrography. The aim was to investigate whether the remains allow an identification of Fazael 2 as lost wax casting site and to gather new insights into the lost wax casting process. The chemical composition of the sampled metal objects is in agreement with previous analyses from other sites, although the polymetallic copper alloys have an overall slightly depleted arsenic concentration and are slightly enriched in the other alloying elements. The items can be subdivided into three groups based on their metallographic structure: unalloyed copper objects, objects with low levels of alloying elements and objects with high levels of alloying elements. These groups correspond to the chemical composition of the objects. They confirm existing notions of the Chalcolithic metallurgy but add two important observations: the metal of some items contain large amounts of silt-sized angular quartz, and some of the polymetallic copper alloys contain inclusions of unalloyed copper and, in one case, of a multi-phase copper alloy. The copper inclusions show that the polymetallic copper alloys were not completely melted upon casting. While a reliable estimate of the actual melting temperature is impossible due to the complex interplay of the alloying elements, all of them decrease the melting point of the alloys compared to unalloyed copper. Consequently, the metal could be liquid enough for casting well below the melting point of unalloyed copper. At the same time, the presence of unalloyed copper inclusions in the polymetallic copper alloys evidence some sort of mixing of the two metal types. In addition, the inclusions of multi-phase copper alloys in one sample indicate that such mixing must not necessarily involve only unalloyed copper. This is the first evidence for mixing of polymetallic copper alloys. The motivation for and location of the mixing action remains unknown. It could be related to, e.g., the alloying of speiss as some sort of master alloy with unalloyed copper or recycling. The whereabouts of the quartz inclusions must remain unknown as well. In contrast to a single sample from Bir es-Safadi, they occur in the entire section and not only at the surface. This excludes their origin as contamination from, e.g., the casting mould. Options are presented but none of them could provide a convincing hypothesis why silt-sized quartz is dispersed throughout the object. The crucible fragments can be separated into three petrographic groups. The first group does not consist of crucibles but vessel fragments made of Moza clay. The second group consists of the low-fired object F225a with a calcareous foraminifera-rich clay and carbonaceous sand. Only the third group contains actual crucibles as is indicated by the vitreous and bloated ceramic paste with vegetal matter, in good agreement with the metallurgical ceramics from the Nahal Beer Sheva sites. The heated sediment nodules seem to be made from the same clay as the crucibles but were heated to much lower temperatures. While some of them contain no temper at all, other contain vegetal temper, carbonaceous sand, or both. One nodule features two layers of different size fractions of the same clay. F2-Y55 is distinct from all these nodules because it is entirely blackened and contains rounded inclusions of vitreous copper-free material, together with vegetal and mineral temper. Comparison with sediments from the Wadi Fazael showed that crucibles and heated sediment nodules were made of this clay. According to the criteria established above, most of the heated sediment nodules can be reliably interpreted as remains of lost wax casting moulds. Being a single find, the function of F225a must remain unclear. In conclusion, the study of the metallurgical assemblage of Fazael allowed identifying Fazael as production site for lost wax cast objects by the identification of in-situ lost wax casting mould fragments for the first time. In addition, it provided for the first time direct evidence for the mixing of polymetallic alloys with unalloyed copper and tentatively also with other polymetallic alloys. ## Re-assessing the metallurgical assemblage of Abu Matar Based on the experience from past smelting experiments, some aspects of the current reconstruction of the Chalcolithic unalloyed copper process in the Southern Levant seemed impractical, particularly the use of blowpipes through the collar opening, and covering the furnace with a large ceramic lid, as suggested by J. M. Golden. In an attempt to refine this reconstruction, the metallurgical ceramics and slag of the Abu Matar excavations in the 1990s were re-assessed and led to the identification of several important but hitherto overlooked pieces. The probably most important ones are two pieces with a clear channel-like feature of about 2.5 to 3.0 cm diameter. While one is made of chaff-tempered ceramic and only slightly vitrified with some greenish spots, the other is completely vitrified. In addition, a large slag fragment features a similar channel with a larger diameter. Further, several ceramic objects are particularly thick, rounded on one side and have slagged sides that extent over the edge to another side. Two other ceramic rim fragments deviate visibly in their paste from the other metallurgical ceramics by the presence of carbonaceous sand and the absence of chaff temper. The slags and some of the ceramics were sampled for microscopic and SEM-EDX analyses. For the remaining objects, the macroscopic features were of interest and sampling did not promised any additional insights. Results of the analyses show that the sand-tempered ceramics are secondarily used fragments of a V-shaped bowl and a hole-mouth jar. While secondary use of V-shaped bowls as crucibles was already previously encountered, this is the first evidence for the secondary use of other vessel types in the metallurgical process. The slagged fracture on one side of the hole-mouth jar fragment suggest the use of hole-mouth jar sherds as furnace cover. Such a use would also allow to cover the furnace with objects that are much easier to handle than one large ceramic lid. The completely vitrified fragment with the channel-like feature turned out to be of ceramic origin rather than slag. Together with the other ceramic fragment with a channel-like feature, they are interpreted as tuyère fragments due to the lack of other objects in the metallurgical process with such a narrow channel. Their diameters suggest the use of bellows. Operating the furnace with bellows instead of blowpipes through the collar opening appears much more sensible because it allows heating of a larger area in the furnace while operating the furnace more efficiently with less effort. In addition, the archaeological experiment showed that a bellows-operated furnace can produce sufficiently high temperatures for copper smelting. A reconstruction for the use of the rounded ceramic fragments was not entirely possible. They do not appear to be part of the furnace wall nor do other sites report any comparable objects. It is suggested here that they may have been some kind of mobile platform for the crucibles to melt the copper prills next to the mould, and fragments of the "clay cakes" J. Perrot mentioned but described only poorly. Based on these results and interpretations, the reconstruction of the unalloyed copper process can be refined: Smelting of copper ore in a furnace covered with hole-mouth jar fragments to keep the heat and to increase the reducing conditions. Bellows were used for draught. The furnace was operated in a rotating manner, placing fresh fuel and ore on the side opposite the tuyère and pushing it closer to the tuyère when the material there reacted. The copper prills were then mechanically extracted from the slag as previously reconstructed and then melted in crucibles. Because handling of the crucibles inside the pit furnace is difficult, they were most likely placed on small mobile platforms next to the moulds, where they were covered in small heaps of fuel. Probably, blowpipes provided draught for this step because they allow much more control over the intensity and direction of the air stream. ## Evolution of the innovation "lost wax casting" Based on the extensive literature review, a large West Asian interaction sphere consisting of Southeastern Anatolia, Northern Mesopotamia, the Southern Caucasus and Iran can be reconstructed. Within this interaction sphere, materials but also ideas could spread quite easily but were not necessarily adopted everywhere. Moreover, two fundamentally different status-communicating systems could be identified. The Southern Levant, parts of Iran, and Southeastern Europe communicated status through metal objects and other so-called prestige items. A characteristic of this system is the use of metal items and other exclusive goods as burial goods. In contrast, Northern Mesopotamia and other parts of Iran used monumental architecture and staple finance to communicate status. The status-communicating system of the Southern Caucasus is unclear but the absence of monumental architecture and the presence of metal items in some of the burials suggest it belongs to the former one. On another strand, except for the Southern Levant, lost wax cast items in West Asia at this time are always made of gold and massively cast. It is therefore suggested that a technological package "lost wax casting with gold" spread from Southeastern Europe through the Caucasus into the West Asian interaction sphere, probably together with the shaft hole axes. Depending on the compatibility with the status-communicating system, this technological package was modified. While the shaft hole axes were readily adopted in the Southern Caucasus and Iran, the technological package becomes only visible in the archaeological record after some delay, seemingly unchanged but in the very different ideological setting of a vertically stratified society. In Northern Mesopotamia and Southeastern Anatolia, only the new material gold seems to be adopted. An explanation might be that the ideological aspects of the technological package -- communicating status through metal objects -- was incompatible with the communication of status through staple finance and monumental architecture while gold as new material was compatible to the newly discovered silver. The Southern Levant adopted and heavily modified the technological package. Instead of gold, polymetallic copper alloys were used, the produced objects are much larger than anything else produced in this technique before, and ceramic cores were used instead of massive casts. The reasons for these modifications seem to be technological and aesthetic. Gold is not readily available in the Southern Levant and the large amounts necessary for the produced objects probably exceeded by far, what was available elsewhere -- if it was available there. In addition, polymetallic copper alloys have a much lower melting point and it seems likely that it was only because of this that such large objects could be produced in the lost wax casting technique. The aesthetic component was probably equally important: Depending on the composition of the polymetallic copper alloys, not only the new materials gold and silver could be imitated but also haematite. Haematite mace heads were produced long before the advent of metallurgy in the Chalcolithic Southern Levant and vessel figurines, ossuaries and other ceramic objects often decorated with red bands. Ceramic cores were likely used to save on metal. A key factor for this quick adoption in the Chalcolithic Southern Levant is the compatibility of the technological package with the existing social system. It is argued here that metal items in the Chalcolithic Southern Levant are representations of objects made of other materials (e.g., stone, pottery, wood). The arrival of metallurgy in general and the lost wax casting technology in particular did not only fill a need in the local population for such shiny representations. At the same time, the strong modification of the technological packages makes it likely that it was stripped from its ideological aspects except for the most basic concepts, status-communication through special items and use of shiny materials, and fully included into the local ideology. ## Role of metal in the Chalcolithic Southern Levant Previous research already indicated that both, the unalloyed copper objects and the lost wax cast objects were status-communicating items. Therefore, their role in the Chalcolithic Southern Levant must be discussed together. Rather than a full-fledged hypothesis, some ideas are presented. It is suggested that all copper-based metal objects in the Chalcolithic Southern Levant are representations (skeuomorphs, imitations) of objects made of other materials. While this is rather obvious for tool-shaped objects and mace heads, other objects were likely made of perishable materials. For example, previous studies suggested that the decorated mace heads might represent seeds or bulbs and some objects in the Nahal Mishmar Hoard have similarity with reed canes. If the ideology of the Chalcolithic Southern Levant is so heavily ritualised as suggested by some researchers, the metal objects are likely to be ritual representations of mundane objects. However, some iconic objects of the Chalcolithic Southern Levant do not have metallic representations, such as V-shaped bowls and basalt bowls. A likely explanation is that they were already ritual objects and, therefore, did not need metallic representations. The turn of the Early to the Late Chalcolithic and the associated change from architecture to metal items for status display might indicate a shift in the ritual system of the Chalcolithic Southern Levant. It is suggested that this shift might be related to a decline of the Chalcolithic ideology due to a decreased power in explaining the changing environment. Metal objects might have been the opportunity to reinforce this power. It is further suggested that one of the key elements to understand the role of metals in the Chalcolithic Southern Levant is their colour. We know from other cultures that gold is often associated with the sun, while silver is associated with the moon. Both were important divine entities in early farming societies. In the Chalcolithic ideology, haematite-like colours or red in general might be related to the ancestors. Thus, metallurgical practices might have been a way to connect with the divine sphere and transform objects from the mundane sphere into the ritual sphere. Smelting and melting can be easily turned into audio-visually impressive rituals, through which this communication could be performed. Previous studies suggested that many aspects of the Chalcolithic ideology are related to a cycle of death and rebirth. Metallurgy was probably firmly embedded in it. Smelting haematite-rich copper ore to metal might symbolise the rebirth of ancestors. Transforming items symbolising them into their metallic equivalents might have been part of rituals connected to the passage of deceased into the nether world. If they were indeed symbolic representations of the deceased, the central aspect of metallurgy might have been the ritual transformation of the mundane item into its ritualised representation. This must not necessarily be connected to metallurgical processes, i.e. their production. This could explain why metal items were part of exchange networks together with other ritual objects. At the end of the Chalcolithic, the ideology collapses entirely and most of the prestige-items disappear. Environmental conditions probably changed to an extent that the Chalcolithic ideology could not provide suitable explanations for them anymore, leading to ritual failure. In addition, internal conflicts might have originated from the important ritual role of metal and might have contributed to the decline of the ritual framework of the Chalcolithic ideology. With the collapse of the Chalcolithic ideology, the lost wax cast items and the polymetallic alloys lose their purpose, whereas the utilitarian value of unalloyed copper was recognised, perhaps inspired from the Aqaba sites. Consequently, the lost wax casting metallurgy was abandoned while the unalloyed copper metallurgy was incorporated into the efficiency-focussed and goal-oriented mindset of the Early Bronze Age.