Captación y selección de materias primas en la primera metalurgia del Sureste de la península ibérica ; Raw material procurement and selection in Southeast Iberia’s early metallurgy

The role of metallurgy in the Copper Age communities of the Iberian Southeast is a recurrent question of archaeological research in western Europe. Based on lead isotope and trace element analyses of archaeometallurgical remains, this paper addresses the territorial organisation of metallurgical production during the Copper Age (3100-2200 cal BC) in the Vera Basin (Almería, Spain), the region with the earliest metallurgical evidence in western Europe. This paper comprises the study of materials from the three main settlements with metallurgical activity in the area (Las Pilas, Santa Bárbara and Almizaraque), as well as some metal objects from these and other sites (La Encantada I, Loma de Belmonte and Las Churuletas 1). The results support a model of small-scale regional production whereby settlements exploited the resources of their nearby surroundings (up to 30 km as the crow flies). However, metallurgical exploitation prioritised mineralisations rich in arsenic and other elements, even when other sources were more readily accessible: for the case of Las Pilas, the exploitation of Pinar de Bédar sources instead of Sierra Cabrera, closer to the site; for the cases of Santa Bárbara and Almizaraque, the sources of Cerro Minado. The possibility that Almizaraque and Las Pilas also exploited the minerals of Herrerías, although to a lesser extent, remains open. Broader exchange networks are indicated by the data from finished objects, from which greater mobility can be inferred.


INTRODUCTION
The Vera Basin, in the Iberian Southeast, is a large tertiary basin spanning some 320 km 2 and traversed by three rivers: Aguas, Antas and Almanzora. It is framed by the Almagro and Almagrera mountain ranges in the North; Cabrera mountain range in the South; Bédar and Lisbona ranges in the West and the Mediterranean Sea in the East (Fig. 1).
Thanks to the extensive archaeological works that started in the late 19 th century by Henry and especially Louis Siret, the extensive occupation of the Vera Basin during the Late Prehistory is well documented, with early metallurgical stages constituting a decisive moment of occupation (Camalich Massieu and Martín Socas 1999). This is the area yielding the earliest metallurgical evidence in Iberia (Ruiz Taboada y Montero Ruiz 1999) and a key region for the study of long term prehistoric social change, as it has a rich archaeological record from the Neolithic to the Bronze Age whereby progressive social stratification processes with increasing fortification of the sites, individualisation of burials and other changes in settlement and burial patterns can be observed, culminating in the El Argar society (see e. g. Chapman 2008;Aranda et al. 2015). Based on field-work at Copper Age (c. 3100-2200 cal BC) sites such as Almizaraque (Cuevas del Almanzora), Las Pilas (Mojácar), Santa Bárbara (Huércal Overa), Zájara and Campos (Cuevas del Almanzora) as well as at Bronze Age (c. 2250-1550 cal BC) sites such as El Argar (Antas), Fuente Álamo (Cuevas del Almanzora) or Gatas (Turre), the Vera Basin has become a priority study area, with much archaeological research especially devoted to investigate the role of early metallurgy in the process of social stratification.
Some authors see metallurgy as a driving force of social change and a causal aspect of social stratification due to the socio-economic changes it could have brought about. This hypothesis is based on the premise that metalworking is a complex technology requiring full-time specialists and therefore encouraging social division of labour which, with the degree of specialisation and intensification reached during the Bronze Age, would have also entailed a territorial division of labour (e. g. Lull 1983, Lull et al. 2010Chapman 2003;Delgado and Risch 2008;Molina González and Cámara Serrano 2009). Other authors give metallurgy a secondary social value and consider that, although it could have favoured the consolidation of elite power, under no circumstances did it entail a structural economic transformation that would have resulted in a process of social stratification. For these authors, metallurgy is considered to have been a low scale, domestic activity that required no full-time specialisation and therefore would not have generated dependence. The stratification processes of the Iberian Southeast during the Copper and Bronze Ages would have been based more on the appropriation of agricultural surpluses, and therefore both coercion and the ownership of land, or, in the case of southeastern Iberia, the control of water systems would have been decisive (e. g. Gilman 1987Gilman , 2001Montero Ruiz 1994;Murillo-Barroso et al. 2015;Murillo-Barroso and Montero Ruiz 2017).
Archaeometallurgical studies at different sites of the Iberian Southeast have revealed a broadly similar technological tradition, which survived with relatively limited change from the Copper Age to the Bronze Age (e. g. Rovira 2002aRovira , 2004. Copper metallurgy is characterised by the preferential use of oxidic ores (although sometimes naturally associated to sulphidic minerals) smelted in open vessels (crucibles or the so-called "vase-furnaces") with air supplied by tuyeres or blowing pipes. Owing to poor control of the temperature and atmosphere, which oscillates from highly oxidising to reducing, this smelting technology generates inmature slag with high viscosity and high copper losses. It therefore is a quite inefficient process. These questions have been highlighted in all the archaeometallurgical studies in the area of study (Delibes et al. 1989;1991;Rovira 2002b;Müller et al. , 2006Murillo-Barroso et al. 2017). This technology, inter- Trab. Prehist., 77, N.º 1, enero-junio 2020, pp. 87-112, ISSN: 0082-5638 https://doi.org/10.3989/tp.2020.12248 preted as a low specialised and low productive one, persisted during the Argaric Bronze Age according to the archaeometallurgical analyses (Moreno et al. 2010;Rovira et al. 2015), although noting changes in object manufacturing technology (with the generalisation of annealing, a technique rarely used during the Copper Age) as well as in the spatial and regional organisation of production.This contribution aims to investigate the territorial organisation of metallurgical production in the Vera Basin during the Copper Age (3100-2200 cal BC) and to evaluate the relationship between settlements -where metallurgical activity was identifiedand burial sites, where the finished objects are found. Using scientific methods (trace element and lead isotope analyses), this work focuses on the provenance identification of metallurgical finds and objects from a broad selection of production sites and burials from the Vera Basin and surrounding areas.

THE VERA BASIN: ARCHAEOLOGICAL AND GEOLOGICAL BACKGROUND
There are three main settlements with evidence of the full chaîne opératoire of metallurgical production in the Vera Basin: Las Pilas, Santa Bárbara and Almizaraque; there is also partial evidence of some stages of metallurgical works in other settlements such as Zájara and Campos. Along with these sites, there is a series of collective burial necropolises that have traditionally been associated with the populations that lived in the Vera depression and the Almanzora River basin, such Fig. 1. Location of sites mentioned in the text. White stars: Copper Age settlements (3100-2200 cal BC); coloured stars: Copper Age tombs or necropolises (3100-2200 cal BC); Triangles: Bronze Age sites (2250-1150 cal BC). 1. Las Churuletas 2. Santa Bárbara; 3. Campos; 4. Zájara; 5. Almizaraque; 6. La Encantada; 7. Las Pilas; 8. Loma de Belmonte; 9. Gatas; 10. El Argar. Design M. Murillo-Barroso. Base map modified from CGIAR-CSI SRTM 90m 53 Digital Elevation Data (http://srtm.csi.cgiar.org/), with permission from SRTM (Jarvis et al. 2008;Reuter et al. 2007)  as La Encantada, Loma de Belmonte and the necropolis of the Purchena Group, among others ( Fig. 1).
In this work, the metallurgical remains from the three above-mentioned settlements (Las Pilas, Santa Bárbara and Almizaraque) with production remains as well as the objects of some of the graves are investigated.

Las Pilas (Mojácar, Almería)
The site of Las Pilas, of approximately 5 ha in extension, is located on top of a plateau 30 m above sea level, close to the estuary of the Aguas River. The site was discovered in 1989 and it was subjected to three archaeological campaigns in 1990, 1991and 1994(Alcaraz Hernández 1992Camalich Massieu and Martín Socas 1999;Rovira Buendía 2007). In the last campaign, an area of 40 m 2 was excavated. The occupation of the site was structured in a sequence of ten phases, based on documented stages of the restructuration and re-organisation of the inhabited area. This occupation took place during the 3 rd millennium BC.
In terms of constructive features, the settlement is characterised by round huts, some of them partially excavated in the ground, generally built on a stone and clay plinth, with rammed-earth walls. They usually have a central post to hold the conical roof made of vegetal lattices waterproofed with clay. Several negative structures related to grain and water storage are associated to these huts. Some large walls have been documented too, possibly related to the demarcation or defence of this sector, which clearly underwent frequent functional re-organisation of spaces where domestic and craft activities co-existed.
Ten archaeological phases have been recorded at the site, where the last excavation campaign yielded a substantial archaeometallurgical assemblage including evidence of the whole production process: stone hammers, copper ores, crucible fragments and a complete one, blowing pipes, slag fragments, metal drops and a few metal objects (see Murillo-Barroso et al. 2017 for a technological study) (Fig. 2). Except for one copper awl and one copper mineral fragment recovered in the earliest phases (1 and 2 respectively), metal-working debris (i.e., slag and a copper lump) appear from phase 5 (2905-2579 cal BC 2σ) and reached the highest frequency in phase 9 (2578-2276 cal BC 2σ) (Muri llo- Barroso et al. 2017).

Santa Bárbara (Huércal Olvera, Almería)
The site of Santa Bárbara is located on top of a plateau 221 m above sea level, next to the course of the Almanzora River. It has approximately 1.5 ha in exten-sion, and its location offers direct access to water resources, visibility and potential control to communication routes. The site was excavated in 1990 and 1991 in a rescue operation (González Quintero et al. 1993;. The excavations showed that the site was fenced by a wall with at least one semi-circular bastion. Other craft activities such as textile production have also been proposed in the site (González Quintero et al. 2018), although metallurgical production stands out due to the amount of metallurgical debris documented all over the site. The metallurgical assemblage currently under study by some of us (MMB, MDCM, DMS and MMT) includes a large amount of metallurgical debris: copper ores, complete and fragmented crucibles, slag fragments, copper droplets and few objects (Fig. 2).

Almizaraque (Cuevas de Almanzora, Almería)
The site of Almizaraque is known since the 19 th century. It was excavated by L. Siret between 1903Siret between -1906Siret between and 1932Siret between -1933. He documented some negative storage structures (silos), rounded huts with some significant objects such as idols, and archaeometallurgical production remains. The site was re-excavated between 1980-1984by Delibes and colleagues (1996 who structured its occupation in five phases and documented its maximum extension, c. 0.3 ha. Absolute dates indicate an occupation from c. 3000 to 2200 cal BC, covering the three phases in which Copper Age is usually divided (Fig. 3). Metallurgical remains are documented from the last period of the first phase up to the abandonment of the site (Delibes et al. 1996) and comprise evidence of the whole metallurgical process: copper ores, slag and crucible fragments, metallic copper lumps and objects (for technological studies see Delibes et al. 1989Delibes et al. , 1991Müller et al. , 2006 (Fig. 4).
Although metallurgical activity is well documented at the three sites, it must be noted that no moulds have been found at any of them, and blowing pipe nozzles have only been documented at the site of Las Pilas. Metal objects are also scarce in these metallurgical sites (with the exception of Almizaraque, where they have been scarcely documented in modern excavations but where Siret recovered more than 70 objects). Therefore, some objects from tombs in nearby sites were included in our provenance analyses with the aim to evaluate their possible link with the production remains at these main sites.

La Encantada necropolis (Cuevas de Almanzora, Almería)
The necropolis of La Encantada is composed of three megalithic tombs (tholoi) located on a mound Trab next to the Almizaraque settlement. They were excavated by Siret at the end of the 19 th century and information is also collected by the Leisners (Leisner and Leisner 1943). La Encantada I is a tholos with a circular chamber 3.5-3.6 m in diameter and a corridor 3.1 m long, divided into three sections of 0.8 m, 1.2 m and 1.1 m, respectively. About 50 individuals were documented with abundant grave goods (Leisner and Leisner 1943).
This grave has two absolute dates, one on carbon from a sample collected by Siret in 1906, unpublished, which could be associated with the oldest moments (although we are aware of the old wood problems), and another on bone that would correspond to the Late Bronze Age and could be correlated with the later materials found in the tomb (Lorrio and Montero Ruiz 2004) (Tab. 1).
Among the metal objects preserved in the Museo Arqueológico Nacional (MAN, Madrid), there is a small chisel, two fragments of a thicker chisel, a punch, a fragment of an axe, a Palmela point with a fragmented tip and a second point fragmented into two pieces that the Leisners describe as a spatula (Tab. 2, Fig. 5). Typologically, the whole assemblage is consistent with 3 rd millennium BC dates. Other metal objects corresponding to later reuses of these tombs such as a hoop or hook fragment, a hook, a low-lead bronze bracelet, a button and a possible fragmented nail were also recovered (Lorrio 2008). This paper presents the lead isotope analyses of the two Palmela points from La Encantada I.

Loma de Belmonte (Mojácar, Almería)
This collective burial is close to the site of Las Pilas and was excavated by L. Siret and his foreman P. Flores (Siret 2001(Siret [1891(Siret ], 1906 and published by the Leisner (Leisner and Leisner 1943). It is a tholos with a corridor of 4 x 1.4 m, divided into two sections, which gives access to the main chamber of 5 m in diameter. In the north side of the chamber, there was a polygonal niche, approximately 1.1 x 1.2 m, formed by five slate slabs. The Leisners, according to the information gathered from L. Siret and P. Flores' field notebooks, noted that 100 individuals were documented (Leisner and Leisner 1943)  queológico Nacional (Madrid) are as follows: three punches, the tip of a forth punch and a fragment of another possible mineralised punch, a fragment of a possible chisel, a folded sheet and a Palmela point (Tab. 2) (Fig. 5).
Lead isotope analyses results of metallic copper from the site of Terrera Ventura and the necropolis of El Barranquete (Stos Gale et al. 1999) will be also considered in this study as well as the only metallic object found in the Churuletas necropolis (Fig. 5). The awl from El Barranquete was recovered in the layer IV, i.e. the earliest one, of the chamber of tomb 9 (Almagro Gorbea 1973) which would be earlier than layer III dated between 2210-1940cal BC 2σ (Aranda et al. 2018. The awl from the Churuletas necropolis was recovered at Churuletas 1, dated between 2900-2670 2σ cal BC (Aranda et al. 2017). These are the only Chalcolithic objects analysed from the Almería province and would thus be contemporary with the metallurgical sites aforementioned (Las Pilas, Santa Bárbara and Almizaraque) and potentially related to them.

Mineralisations from the Vera Basin
The Vera Basin is renowned for its metallurgical wealth. Copper mineralisations are well-known in the area surrounding Las Pilas (Pinar de Bédar and Sierra Cabrera), Santa Bárbara (Cerro Minado) and Almizaraque (Sierra Almagrera and Herrerías) (Fig. 1). Prehistoric mining activity has only been well documented at Cerro Minado, with an absolute date of 2469-2336 cal BC 2σ (Delgado et al. 2014), although L. Siret recovered a couple of mining hammers in Herrerías as well, which could also be indicative of prehistoric mining (Montero Ruiz 2010) (Fig. 6). However, the vast intensity of Roman and especially modern mining activities, particularly in Pinar de Bédar and Herrerías, may have completely erased any prehistoric mining evidence. A geological survey was conducted in the area with the aim to better characterise these mineralisations both on their trace element patterns and their lead isotopic signatures (Murillo-Barroso et al. 2019). The mining districts sampled were Pinar de Bédar (Sierra de los Filabres) in the Mulhacén Complex; Lower Sierra Cabrera in the Alpujárride Complex; Herrerías  (Sierra Almagrera), also at the Alpujárride Complex but with a hydrothermal mineralisation; and Cerro Minado (Sierra de Almagro) in the Alpujárride Complex (see Murillo-Barroso et al. 2019 for further geological details).
The mines of Bédar are found in the Sierra de los Filabres. They were intensely worked in the 19 th and 20 th centuries as sources of iron and lead ores as well as for copper ores in its southeastern zone at Sierra de Alcornia from at least the 16 th century (Montero Ruiz 1994; Soler Jódar 2014; Soler Jódar and Hansen 2016). No indications of prehistoric works were documented; however the later intensive mining exploitation could have eliminated such evidence. In the present, the housing development executed on this terrain between 2006 and 2008 has destroyed or covered the majority of the copper mines and altered the conditions for future investigation.
Sierra de los Filabres belongs almost entirely to the Mulhacén Complex. Ore deposits in this location are mainly associated to marbles, are stratiform, and show a mineral association of galena-chalcopyrite-malachite-azurite (Montero Ruiz 1994). In addition the presence of cerussite, zinc-olivenite, olivenite, barite and smithsonite has been documented in smaller quantities. X-ray fluorescence (XRF) (Montero Ruiz 1994) as well inductively coupled plasma mass spectrometry (ICP-MS) (Murillo-Barroso et al. 2019) ore analysis revealed the high levels of zinc, arsenic, lead and uranium in these copper mineralisations.
The chain of Sierra Cabrera is the continuation of the Sierra Alhamilla towards the east, and separates the Vera Basin of the Aguas River to the north from that of the Carboneras River to the south. The Alpujárride Complex emerges in a band at the foot of the northern slopes of the range, and is followed by the Mulhacén Complex towards the south. From the Triassic, there is a carbonate formation composed of black and greyish-brown rocks with very dense stratification and with a few interspersions of blue phyllite layers. It is in these phyllites layers where small outcrops of copper ores are found. Copper minerals from Sierra Cabrera display lower levels of impurities (arsenic, zinc or lead) when compared to other mineralisations at the Vera Basin (Montero Ruiz 1994; Murillo-Barroso et al. 2019).
Sierra Almagrera, adjacent to Herrerías, is geologically uniform, mainly composed by materials of the Paleozoic bedrocks of the Alpujárride Complex. Mineralisations from Sierra Almagrera are rich in galena, silver and copper sulphides with a host rock of graphite-rich phyllite and quartzite. Chalcopyrite has been documented in the mine Ramo de Flores and native copper, alongside with azurite and malachite in the Jaroso Ravine. Arsenic is not detected in copper ores analysed from Sierra Almagrera although lead is usually associated to copper minerals (e. g. La Estrella Mine 1 , Montero Ruiz 1994).
Herrerías, adjacent to the seams of sulphites and sulphosalts of Pb-Sb-Ag in the Sierra Almagrera, constitutes a very well defined mineral zone inside the Vera-Garrucha basin (Martínez Frías 1991;López Gutiérrez et al. 1993) where copper ores have also been identified. In Herrerías a deposit of oxides and hydroxides of Fe-Mn, sulphides of base metals, and native silver can be found. Both Herrerías and Sierra Almagrera, dike type deposits, are considered epigenetic, with a similar temperature of ore deposition (200-300 °C) and very similar 34 S isotope values in the barites which accompany the mineralisations in both deposits. Due to this, a similar hydrothermal genesis for both, probably in a submarine environment, has been proposed (Martínez Frías 1992). This similitude in the genesis of the two deposits is also reflected in similarity and proximity in their lead isotopic fields (Murillo-Barroso et al. 2019). The intensive mining of Herrerías makes the identification of copper samples difficult and would have erased any evidence of prehistoric mining works.
The last copper mineralisation in the Vera Basin is Cerro Minado in the Sierra de Almagro, mostly occupied by the Alpujárride Complex. Two major tectonic units have been distinguished, a lower and an upper. The most attractive copper and cobalt ores are found in the upper unit, composed basically of schists and quartzites, phyllites and dolomites (IGME 1980). The primary deposits of copper, cobalt, silver and mercury are found filling a dense network of fissures which is a consequence of, amongst other factors, the brecciation of the dolomites, what has allowed the primary paragenesis to suffer an intense oxidizing and carbonating alteration. More than 100 types of minerals have been identified at Cerro Minado (Favreau et al. 2013) including native copper, malachite, cuprite, olivenite, etc. Elemental composition analysis revealed the high levels of arsenic, cobalt and nickel, due to the coexistence of erythrite and annabergite with copper minerals and arsenates (Murillo-Barroso et al. 2019).
Recent studies have documented stone peaks and hammers from Cerro Minado, most of them on the surface and two associated to a gallery (Delgado et al. 2014). Absolute dates confirm its exploitation during the Copper Age, contemporaneous with the occupation of several sites in the area such as Almizaraque, Las Trab Pilas or Santa Bárbara. Its exploitation has been also proposed in the neighbouring region of the Guadalentin Valley, Murcia (Escanilla 2016) but based only on elemental analyses and lacking isotopic confirmation.
The lead isotopic fields of these mining districts were defined and three different isotopic fields were identified: Cerro Minado, Herrerías/Sierra Almagrera and Pinar de Bédar/Sierra Cabrera. Probably due to a similar genesis, Herrerías and Sierra Almagrera cannot be isotopically discriminated, and the same applies to Pinar de Bédar and Sierra Cabrera. However, discrimination is possible on the basis of their different trace element patterns (for detailed discussion see Muri llo-Barroso et al. 2019).

MATERIALS AND METHODS
More than 1000 archaeometallurgical finds were recovered during the archaeological excavations of the three metallurgical sites (Las Pilas, Santa Bárbara and Almizaraque). The collections include ore fragments (maximum fragment weight of 260 g), slag fragments, slagged crucible sherds and complete crucibles, stone hammers and grinding stones, copper droplets or lumps, and only in the case of Las Pilas, fragments of ceramic blowpipe nozzles (MNI five) as well as few metallic objects (Figs. 2, 3 and 5).
Selected materials were sampled for further analyses at the UCL Institute of Archaeology's Wolfson Archaeological Science Laboratories. We employed a stratigraphic sampling frame that considered archaeological and typological information, as well as the results of 591 screening analyses by portable x-ray fluorescence (pXRF) using an Olympus Innov-X Systems Delta Premium. Samples were selected from across the compositional groups qualitatively identified by pXRF. The sample set amounts to a total of 48 samples, including 21 samples of copper ores, 17 slag fragments and 10 copper objects or lumps from the three metallurgical sites, which were selected for lead isotope analyses, seeking to cover the whole production process in each site.
The sample is completed with four copper objects from the nearby necropolis of La Encantada I, Loma de Belmonte and Las Churuletas 1. Data from the literature from Almizaraque, Terrera Ventura and El Barranquete are also used for comparative purposes. In addition, trace element analysis was conducted for 18 archaeological ore samples from Almizaraque, Las Pilas and Santa Bárbara in order to compare them with the trace element patterns of the geological samples in the area (Tab. 3).
Trace element analyses of ore samples were conducted by inductively coupled plasma mass spectrom-etry (ICP-MS) at the SgIker-Geochronology and Isotope Geochemistry Facility of the University of the Basque Country UPV/EHU (Spain).
For each sample, 80 mg were digested in concentrated HNO 3 from Merck Pro-Analysis in closed Savillex PFA vessels set on a hot plate at 120ºC during 24 hours. Digestion was conducted two consecutive times due to the low dissolution level in some samples. This is probably due to the high silicate portions of the mineral gangue, although our interest is on the soluble portions of metalliferous minerals. Subsequently, the solution obtained was brought to dryness and taken up in diluted HNO 3 . The solution obtained was centrifuged and small amounts of undissolved residues were found in most samples. Only the fraction soluble in nitric acid, i.e., the dissolved fraction, was analysed. Dilution after dissolution was done gravimetrically using an electronic balance with precision of 0.1 mg in order to prevent errors induced by volumetric dilutions. All reactives used in sample preparation were distilled in the laboratory by sub-boiling distillation using an Acidest quartz distiller in order to avoid contamination. Deionised water was obtained using a Millipore Elix device and polished to obtain a resistivity ≥ 18 M Ohm-cm with a Barnstead EasyPure system.
Quantitative determination of analytes of interest was done by means of a Thermo XSeries 2 inductively coupled plasma mass spectrometer (ICP-MS) equipped with collision cell (CCT), an interphace specific for elevated total dissolved solids (Xt cones) and shielded torch. A concentric nebuliser and quartz expansion Peltier-cooled chamber were employed. Rh solution, used as internal standard, and multielemental solutions for the initial tuning and calibration of the mass spectrometer, and for quality control (QC) of the results were prepared from 1000 ppm high purity Merck multi-element standard solutions for ICP, stabilised in diluted HNO 3 2 to 6%.
The internal standard was added by means of an automatic online addition kit in order to prevent random errors (further details on the instrumental method are given in García de Madinabeitia et al. 2008). Error estimation for each element is established using the error propagation equation of Miller and Miller 2004. Uncertainty of the results corresponds to a 95% confidence level (see analytical errors 2SE in Supplementary File, SF.1).
Most of the lead isotope analyses were conducted via multi-collector inductively-coupled plasma -mass spectrometry (MC-ICP-MS) at the Geochronology and Geochemistry Service (SGIker) at the University of the Basque Country (Spain).
Aliquots of 0.05 g of the samples were prepared, which were dissolved with 0.5 mL of 7N HNO 3 overnight in a 70°C heating plate. After digestion the liquid has evaporated at 70°C. The residues were redis- solved in 0.5 N HBr and processed by conventional liquid chromatography with AG1-X8 ion exchange resin (Manhès et al. 1984), obtaining final solutions of 2 mL of 6N HCl with the purified Pb. These solutions were taken to dryness. The lead samples were dissolved in 1.5 mL of 0.32N HNO 3 and if necessary they were diluted to a final concentration of 150-200 ng Pb/g solution. The samples were introduced as wet aerosols in a Neptune (Thermo Fisher Scientific) MC-ICP-MS using a PFA micronebuliser with nominal suction of 100 µL/min (Elemental Scientific) and double pass dual cyclone-Scott expansion chamber. Actual baselines (electronics + chemistry) have been subtracted from the measurements of a chemical blank for 60 s prior to each sample. The spectrometric measurements were carried out in 105 cycles with an integration time of 8 s per cycle.
A Thallium reference material NBS997 with a normalised ratio of 205 Tl/ 203 Tl = 2.3889 was used for the internal mass correction (see Chernyshev at al. 2007 and references therein for further methodological detail). The reliability and reproducibility of the method were verified by regular measurements of the certified reference material NBS981 interspersed between the measurements of the samples, and in the same conditions. The average uncertainties (2SE) for the NBS981 during this study are 0.0004 for the 206 Pb/ 204 Pb ratio, 0.0005 for the 207 Pb/ 204 Pb, 0.0014 for the 208 Pb/ 204 Pb, 0.0001 for the 208 Pb/ 206 Pb and 0.00003 for the 207 Pb/ 206 Pb (NBS981 values used are those proposed by Baker et al. 2004). Uncertainties of measurements are smaller than symbols used in all graphs.
The lead isotope analyses of the three Palmela points from the necropolis of La Encantada and Loma de Belmonte was conducted at the Department of Earth Sciences (University of Geneva, Switzerland) in collaboration with Massimo Chiaradia. After dissolution in a mixture of 1 mL 7M HCl and 1 mL 14M HNO 3 , samples were dried, conditioned in a few drops of 4M HBr and Pb was separated from the matrix using micro-columns loaded with AG-MP1-M resin in a hydrobromic medium. The purified Pb was redissolved in a 2% HNO 3 solution spiked with a Tl standard. Isotope ratios were measured using a Thermo Neptune Plus MC-ICP-MS  Pb/ 204 Pb. The in-run corrected Pb isotope ratios were further corrected for external fractionation (due to a systematic difference between measured and accepted ratios of NBS981) by a value of +0.5‰ amu using the NBS981 values proposed by Baker et al. 2004.
Procedural blanks are <200 pg and are insignificant compared to the amount of sample Pb.

Trace element composition
The ore samples recovered in archaeological contexts display similar compositional characteristics. In all cases the samples are of polymetallic copper ores, most of them with elevated levels of arsenic (Tab. 4). Nonetheless some significant differences are observed. In order to better visualise these differences, the concentration of the most significant trace elements in the copper ores was normalised to the recommended composition of the upper continental crust in (Rudnick and Gao 2014). This normalisation aims to highlight the deviations in the chemical composition of copper ores from mean continental crust values. Normalised element patterns of ores from Las Pilas, Almizaraque and Santa Bárbara are plotted in Fig. 7, and compared with element patterns of copper mineralisations of the Vera Basin.
At the three sites, the ore samples show elevated levels of cobalt and nickel, especially in the samples from Santa Bárbara (with median values of 1136 ppm Co and 1391 ppm Ni, as opposed to 63 ppm Co and 299 ppm Ni at Las Pilas and 537 ppm Co and 170 ppm Ni at Almizaraque), which is consistent with the mineralisations of the region, especially for Cerro Minado, with high Co and Ni values and where erythrite and annabergite have been identified in copper ore samples (Murillo-Barroso et al. 2019).
The levels of zinc and lead are relatively high in the ores from Las Pilas (median values of 2.1% Zn and 0.7% Pb), as opposed to lower levels in the samples from Almizaraque (0.3% Zn and 110 ppm Pb) and especially from Santa Bárbara (0.1% Zn and 51 ppm Pb). Those high levels of Zn and Pb are consistent with Pinar de Bédar copper mineralisations. In contrast, the levels of arsenic and silver are very similar in the samples from the three sites: median values of 4.2% As and 40 ppm Ag at Las Pilas, 3.5% As and 105 ppm Ag at Almizaraque and 3.4% As and 119 ppm Ag at Santa Bárbara. One should also highlight the elevated levels of uranium in the samples from Las Pilas (median 95 ppm U), significantly higher than those from Santa Bárbara (5 ppm U) or Almizaraque (2 ppm U), although this high U content has been also documented in the copper ores of Pinar de Bédar (Murillo-Barroso et al. 2019).
A linear discriminant analysis (LDA) was conducted over the geological sample set to optimise separation between known groups, with a 96.5% of the variance explained (Fig. 8). Comparison of the trace element patterns of the archaeological ore to the geological samples shows a resemblance between the archaeological ores from Las Pilas and the geological deposits at Pinar de Bédar (particularly the levels of zinc, lead and uranium), which is consistent with the closer proximity between the Las Pilas site and these mines (Figs. 7 and 8).
The cluster formed by the majority of the samples from Almizaraque and Santa Bárbara most closely resembles the trace element pattern of Cerro Minado, although it is not possible, with the data available, to conclusively distinguish the deposits of Herrerías and Cerro Minado on the basis of their trace element compositions. The most useful element to discriminate between these two deposits appears to be lead: while the average lead level is 34 ppm at Cerro Minado, the content of this element is significantly higher at Herrerías, with an average of 530 ppm (see Murillo-Barroso et al. 2019). Previous analyses would seem to corroborate this trend: specifically, those conducted by Siret on ores from Herrerías (cf. Montero Ruiz 1994) show an average of 15% Pb, while in those from Cerro Minado analysed using XRF by I. Montero no lead was detected (with a detection limit of 0.1%, Montero Ruiz 1994). As such, it could be considered that lead is a differentiating element between the deposits of Herrerías and Cerro Minado.
In this regard, one may highlight that, in our set of samples from Almizaraque and Santa Bárbara, the samples from Santa Bárbara show low median lead values (50.9 ppm) while the samples from Almizaraque show higher median values (110 ppm).
The deposits of Sierra Cabrera are well differentiated on the basis of their pattern of trace elements, with none of the archaeological samples falling in this cluster.

Lead isotope analysis
Generally, lead isotope analyses of archaeometallurgical remains of these archaeological sites do not show very high radiogenic values, with 206 Pb/ 204 Pb ranging from 18.44 to 18.95 (Tab. 5). However, relevant observations can be presented. Regarding the production debris (ore and slag fragments), a tight cluster can be identified at the site of Las Pilas, where most samples plot c. 18.8 on the 206 Pb/ 204 Pb axis. All these samples are from the same phase (Phase 9 dated from 2578 to 2276 cal BC 2σ, Murillo-Barroso et al. 2017) and thus could correspond to a single event. The situation is different for the other sites: production debris from Almiza-raque plot mostly between 18.55-18.75 on the 206 Pb/ 204 Pb axis, and those from Santa Bárbara mostly between 18.6-18.8 in 206 Pb/ 204 Pb (Fig. 9).
Conversely, the three metal objects from Las Pilas do not match the tight cluster of ores and slag fragments, plotting at 18.56, 18.66 and 18.70 on the 206 Pb/ 204 Pb axis, respectively. This could indicate that these three objects were produced during a different metallurgical operation or maybe elsewhere, although two of them plot close to one slag fragment (18.55) and one ore sample (18.65) from the same site. Interestingly none of the metal objects analysed matches the production debris cluster from Las Pilas; nothwistanding the small sample size, this could indicate that the metal related to this production event was not consumed at any of these three sites of the area. Conversely, two of the four objects from Almizaraque and the three objects from Santa Bárbara can be related to the production debris of each of these sites (although two of the objects from Santa Bárbara being in the limit) indicating an on-site consumption of the metal produced (Fig. 9). Using thermal ionization mass spectrometry (TIMS) data from the literature and MC-ICP-MS results from our own sample set, three isotopic fields were distinguished: Cerro Minado, Pinar de Bédar/Sierra Cabrera and Herrerías/Almagrera. When plotting archaeological samples against these isotopic fields, several observations can be made: the majority of the metal-production waste from Las Pilas, whose values group into a cluster around 18.8 ( 206 Pb/ 204 Pb), coincide in all bivariate plots with the isotopic field of Pinar de Bédar/Sierra Cabrera (Fig. 10). Although isotopically it is not possible to differentiate between the copper deposits of Pinar de Bédar and Sierra Cabrera given their similar genesis, it is possible, as already mentioned above, to distinguish them by their trace element patterns (Figs. 8 and 9): the copper ores at Pinar de Bédar have much higher mean levels of arsenic (7.6%), zinc (3.4%) and lead (3.4%) than those from Sierra Cabrera (455 ppm As, 276 ppm Zn and 31 ppm Pb on average) (Murillo-Barroso et al. 2019). These compositional differences allow us to posit that the origin of the archaeological metalworking remains from Las Pilas was Pinar de Bédar given that the average content of arsenic, zinc and lead from the ores and fragments of slag recovered at this site also show relatively high average values for these three elements (3.8% As, 2.7% Zn and 1.4% Pb on average for the ores [Tab. 4], and 9.7% As 2 O 3 , 19.7% ZnO and 6.5% PbO on average in the fragments of slag [Murillo-Barroso et al. 2017]). Nevertheless, although the majority of the metalworking remains from Las Pilas can be associated with the deposits of Pinar de Bédar, we also note the minor presence of ores from other mining areas. This applies to sample LP8479 which can be associated with the isotopic field of Cerro Minado (with lower levels of lead and zinc and relatively elevated levels of Co and Ni), and the sample of copper LP6972, which falls in the overlap zone between Cerro Minado and Herrerías/ Almagrera. The sample appears to correspond rather more with Herrerías/Almagrera in the lead isotope plots (Fig. 10), although its composition resembles more closely the ores from Cerro Minado, with high levels of Co and Ni. However, the samples from Las Pilas show a trend along a lead growth curve like the ores from Cerro Minado while the ores from the Bédar range seem to form a secondary isochron (Fig. 10). As such, it is plausible that the isotopic field of Cerro Minado may expand in this direction if further samples are analysed, although the composition of both ores and slags from Las Pilas are more consistent with the mineralisations of Pinar the Bedar than to the ones at Cerro Minado regarding their high levels of Zn, As, Pb and U. Trab. Prehist., 77, N.º 1, enero-

DISCUSSION
The site of Las Pilas is located in the foothills of the Sierra Cabrera at barely 1 km from areas where copper deposits have been identified, while the mines of Pinar de Bédar are found at a distance of 10 km as the crow flies. There are several reasons why the metalworkers from Las Pilas would choose to make use of the ores from Pinar de Bédar despite having available copper resources in their immediate surroundings. It is possible that this was a voluntary and intentional choice, given the greater richness in arsenic of the Pinar de Bédar ores as compared to those of Sierra Cabrera, allowing the direct smelting of arsenical copper. Nonetheless, it is difficult to evaluate whether the visibility of these resources in prehistory was significantly greater than those of Sierra Cabrera. Nowadays the potential of the copper deposits at Pinar de Bédar is greater than that of the veins of copper which appear spread amongst the phyllites of the Sierra Cabrera, but as has been mentioned, the intense modern exploitation  of the zone makes it impossible to evaluate the state of the deposits in prehistoric times.
Returning to the lead isotope analysis results, the bulk of the samples of metal production remains from Santa Bárbara and Almizaraque resemble the isotopic range of Cerro Minado (Fig. 10). In the case of Santa Bárbara, all of the slag samples and four out of the six ore samples analysed can be associated with Ce rro Minado. Of the other ore samples, HOSB8/30 falls in a peripheral zone of Pinar de Bédar/Sierra Cabrera isotopic field, although its trace element pattern does not match either of these deposits. The second ore sample, HOSB11/48, would fall in a zone which cannot currently be associated with any known isotopic field in Iberia.
In the case of Almizaraque, six out of the nine ore samples analysed can be associated with Cerro Minado although two of them (AA1100C, with 1.4% As [Delibes et al. 1989] and ALMZ_1096, with 0.9% As [Tab. 4]) fall in the limit of Herrerías/Almagrera and Cerro Minado. Of the three other samples, AA1099B, with relatively elevated As, Pb and (to an extent) Zn contents (Delibes et al. 1989), resembles the isotopic field of Pinar de Bédar/Sierra Cabrera although falling in a peripheral area in some plots. Nonetheless, the samples ALMZ_M16 and PA0320 fall close to Cerro Minado, without fitting with any known isotopic field, which opens the possibility that ores from other deposits were also being exploited, although to a lesser extent. In the case of the slag samples from Almizaraque, four of them can be associated with the isotopic field of Cerro Minado: PA1443F and AA1104E match with Cerro Minado, while AA1106D and AA1080B fall partially in the space between the geological samples analysed using MC-ICP-MS and TIMS (Fig. 10). However, taking into account the greater analytical error in the older TIMS analyses, a link to Cerro Minado cannot be completely ruled out. Finally, the last slag sample (ALMZ_AA1094) falls in a zone which cannot be associated with any known isotopic field in the region.
In a study of Egyptian samples it has been proposed that the contribution of lead from ceramics could distort the isotopic signature of the slag samples, especially when the mineral resources used have a low lead content (e. g. Rademakers et al. 2017), something which is true in the cases of the ores from Cerro Minado and from Sierra Cabrera (c. 30 ppm on average). However, in our case, a similar spread is observed for both ore and slag samples from Almizaraque and Santa Bárbara, thus indicating that the explanation derives rather from the use of diverse mineral resources.
Overall, it appears that Cerro Minado is the main source of ores employed at Almizaraque and Santa Bárbara, although ores from other deposits could be arriving sporadically. Santa Bárbara is located barely 1 km, and Almizaraque around 20 km as the crow flies, from Cerro Minado. In the case of Santa Bárbara it is clear that they were exploiting the nearest copper mineral resources to the site. However, in the case of Almizaraque the situation is similar to that of Las Pilas, since the Herrerías mines and the deposits of Sierra Almagrera are barely 500 m from Almizaraque. In spite of this, it appears that the majority of the metal production remains from this site match better the isotopic field of Cerro Minado. The preferential choice of some ore sources in this case does not appear justified by the arsenic levels in Cerro Minado (2.5% on average (Murillo-Barroso et al. 2019). Although the As levels of Sierra Almagrera ores are indeed typically low (<0.1% As on average in the XRF analyses by Montero Ruiz 1994 and <0.01% As in the recent XRF analyses from the mine La Estrella 2 ), the deposits at Herrerías, adjacent to Almizaraque, also show moderate levels of arsenic (1.5% on average, Murillo-Ba rroso et al. 2019). Again the importance of variables such as the visibility, accessibility and potential of the deposits in prehistoric times must be highlighted, although these variables are difficult to evaluate in present times.
We have to bear in mind as well the difficulty in defining the isotopic field of Herrerías, which cannot be differentiated from Sierra Almagrera, and from which two of the analysed samples are displaced significantly towards the zone which covers the isotopic field of Cerro Minado. In figure 10 the two 'outliers' from Herrerías are also represented, which could be indicators of the direction in which the isotopic field might expand should further analyses be conducted. Were that so, the overlap with the isotopic field of Cerro Minado could turn out to be considerable; based on the limited dataset available, the differences in the trace element patterns are limited too (with the apparent exception of the mentioned lead levels). It would be necessary to expand the ore sampling from Herrerías to achieve greater definition, although given the extensive modern exploitation of the area it will be difficult to obtain new samples for analysis.
Overall, at the three sites studied, copper sources with naturally elevated levels of other elements, especially arsenic, zinc or lead, were prioritised ahead of other sources with lower levels (especially of arsenic), even when the latter were at a closer distance. This model differs from that proposed for the Valle del Guadalentín (Murcia): here, Escanilla (2016) has proposed the deliberate mixing of arsenic-rich ores from the distant sources of Cerro Minado (beyond around 50 Trab. Prehist., 77, N.º 1, enero-junio 2020, pp. 87-112, ISSN: 0082-5638 https://doi.org/10.3989/tp.2020.12248 km) with local, arsenic-poor copper carbonates, yielding metal with lower arsenic levels than if Cerro Minado ores were smelted directly. Nonetheless, this constitutes a model based exclusively on elemental analyses, which should be verified isotopically.
Turning to the lumps of metallic copper and finished objects from the three sites, it is significant that none of the metallic remains coincide with the isotopic fields of Pinar de Bédar/Sierra Cabrera or Herrerías/ Almagrera, especially given that the use of these resources has indeed been verified, at least in the case of the site of Las Pilas. Instead, metallic samples from all three sites can be linked to Cerro Minado. At least one sample from each site can be associated with this isotopic field in all bivariate plots, with a second from each site falling in peripheral zones in some diagrams. Further, it is significant that three objects (one from each site) have very similar isotopic values, clustering around 18.55 ( 206 Pb/ 204 Pb), a zone around which metal production remains from the three sites also group, and which does not coincide with any known isotopic field in the close area, unless new analyses from Ce rro Minado or Herrerías expand their isotopic fields in this direction. This could indicate the existence of other deposits exploited, at least occasionally, by the three sites and which are either completely exhausted or yet to be characterised.
With regards to the isotopic signature of the remaining objects analysed, it is worth highlighting that the punch recovered in the grave at Churuletas 1 does not resemble any of the pieces from the Southeast nor the metal production remains from the Vera Basin, having an isotopic signature more similar to the objects from southwestern Iberia which could be associated with deposits of an older geological age. Figure 11 shows how objects from the southwest, in this case represented by samples from the provinces of Seville and Badajoz identified with circles (Hunt 2003;Hunt et al. 2009;Hunt et al. 2012), tend to cluster mostly in the range 18.   For the remaining artefacts from the area, only the Palmela-type arrowhead from Loma de Belmonte and the sample from Terrera Ventura can be related to the deposits in the Vera Basin, which in both cases coincide with the isotopic field of Cerro Minado. In the case of La Encantada I, the two arrowheads resemble three objects recovered in La Pijotilla (Badajoz), one punch, one saw, and a possible ingot, for which Ossa Morena has been proposed as their ore source (Hunt 2003).

CONCLUSIONS
The aim of this paper was to evaluate the provision and management of copper ore resources at three metallurgical sites in the region where the earliest metallurgy in Iberia is documented. Having archaeological remains related with the whole productive process as well as a few final metal objects, it has been possible to both identify the copper sources exploited and to evaluate if the objects found at each site had been produced on-site or obtained by other means.
Based on the combination of lead isotope analyses and trace element compositions we can establish the following: 1. As suggested by prehistoric mining evidence at Cerro Minado, its exploitation during the Copper Age by the sites of Santa Bárbara and Almizaraque is well confirmed.
2. The exploitation of the copper resources of Pinar de Bédar by the site of Las Pilas is also confirmed despite the absence of prehistoric mining evidence in this district. Intense modern mining works starting in the 16 th century and developing in the area especially during the 19 th and the 20 th centuries might have erased any prehistoric mining evidence in the field, which adds value to lead isotope analysis in combination with trace elements composition as an indirect proof of prehistoric copper mining.
3. There is no isotopic concordance between metallurgical remains and finished objects at every site. While most of the production debris from Las Pilas clusters quite tightly together matching the isotopic field of Pinar de Bédar, none of the objects can be associated to this cluster or isotopic field. That would suggest a metallurgical production focused mainly on the exploitation of specific resources and a movement of objects within a network probably including Santa Bárbara and Almizaraque. These two sites shared the resources of Cerro Minado, as confirmed in objects and production debris, therefore suggesting self-sufficient production.
4. Copper resources with higher levels of zinc, lead and/or (especially) arsenic appear to have been prioritised over other copper resources with lower levels, despite being geographically at further distance from the sites. This is the case for the exploitation of copper ores of Pinar de Bédar instead of Sierra Cabrera at Las Pilas, or Cerro Minado instead of Herrerías or Almagrera at Almizaraque. 5. Regional 'provinces' of metal exploitation and circulation can be inferred when comparing southeastern and southwestern metal objects, although some sporadic long-distance exchange would have happened, as exemplified by the awl from Churuletas 1 and the Palmela points from La Encantada I.
All in all, the analyses of Copper Age production remains and artefacts from the Southeast of the Iberian Peninsula allow us to propose a model of local production using mainly, but not only, copper resources from the nearby surroundings. Despite the prioritisation of sources rich in arsenic and other elements, even though found at a greater distance from the settlements, the distances are no more than 30 km as the crow flies. In the case of Las Pilas, this preference for resources rich in arsenic, zinc and lead from Pinar de Bédar over those from Sierra Cabrera, might also be due to the greater metal-producing potential of these copper resources as compared to the low-grade copper ores from Sierra Cabrera, which would have a lower yield.
In the case of Almizaraque it seems unlikely that the metal workers at this site were unaware of the existence of the copper ores at Herrerías, at barely 500 m from the settlement and with arsenic levels also notably elevated (1.5% on average), indicating that neither arsenic content nor proximity were the determining factors at the point of choosing a resource. Other issues must also have had an important role to play in the choice of these ores, such as a greater tradition of mining at Cerro Minado, social relations with other communities such as that of Santa Bárbara (perhaps involving collaboration in mine-working), or factors relating to the visibility or accessibility of the resources.
As concerns the finished objects, although some of those from Almizaraque and Santa Bárbara could be linked to metal production remains at their respective sites, the same cannot be said in the case of Las Pilas, where it seems likely that metal produced in other settlements may be arriving. This evidence points to the circulation of goods at the regional scale and may-Trab. Prehist., 77, N.º 1, enero-junio 2020, pp. 87-112, ISSN: 0082-5638 https://doi.org/10.3989/tp.2020.12248 be even further. The intensity of this process cannot be for now determined; however, given the relative abundance of both ore resources and production sites in the same area, metal exchange is unlikely to have been driven by hierarchical relations.
From a wider perspective, this study confirms that the metal production model during the Iberian Copper Age is neither one of dependence on a single mineral source nor one of controlled access of resources given the abundance and widespread distribution of copper ore resources available. The evidence thus appears to support a model of domestic, not specialised and not monopolised, metallurgical production. Against this background, high-resolution local studies, such as the one presented here, are mandatory to progressively understand the development of metal production and consumption at the peninsular scale. In this sense, this study is only a starting point that should be expanded to address the regional and supra-regional levels.