The Role of Thera in the Bronze Age Trade in Metals

The geographical position of Thera and the rich and complex material excavated at Akrotiri suggests that this site was an important trading post for Late Bronze Age commerce.

Lead isotope analyses of metal artefacts from various Bronze Age sites in Greece and numerous ores from known metal deposits in the Mediterranean lead us to conclude, on present evidence, that only a few metal deposits were exploited in Bronze Age Greece. This locally produced metal may have been dominant within the Aegean, but some copper from Cyprus is also found on many Late Bronze Age sites in Greece.

The results of lead isotope and chemical analyses of Late Bronze Age lead and litharge suggest that at that time the predominant source of lead and silver in the Aegean was the polymetallic deposit at Lavrion in Attiki. However, the distribution of litharge, found predominantly on the Cycladic sites (including Thera), suggests that silver was not only extracted in Lavrion itself, but transported either as argentiferous lead ore, or silver rich in lead, to settlements for final silver production.

The scientific investigation of the metal trade in the Bronze Age Mediterranean is still in its infancy, but arleady the results of lead isotope analyses are suggesting new archaeological interpretations and stimulating a fresh look at the Bronze Age economy.

INTRODUCTION

From the beginning of the Bronze Age the Cyclades played an important role in the metal trade in the Eastern Mediterranean. The number of metal objects of silver, copper and lead excavated on the Cyclades does not leave any doubt that these materials were amongst the most important goods being distributed among the islands and along the coast of the Mainland. Lead isotope and archaeological evidence identified the island of Kythnos as the Early Cycladic source of copper, whilst Siphnos played the same role for the supplies of silver and lead. Later in the Bronze Age more economical ore deposits, on the Greek mainland, in Anatolia and in Cyprus were used, but the Cyclades retained their role of at least trading posts, if not actually active trading communities. One of the important recipients of all these metals throughout the Bronze Age was Crete. The geographical position of Thera and its apparent 'Minoanization' suggest that this island had an important role to play in Minoan trade. It also seems likely that the eruption of Thera may have upset the well-established trade routes and perhaps it might be possible to detect a change of metal supplies to Crete after that event.

It is well known that access to raw materials is of crucial importance to any society or culture. Access to necessary metal sources like copper, gold, silver and, in the Late Bronze Age, tin, was possibly as important for the development of wealth and power in Bronze Age society as is access to sources of energy in the second half of the 20th century. The politics of our world concentrate mainly on geographical areas and countries rich in crude oil or uranium deposits. Some five thousand years ago stone and metals may have played a similar role for economy and politics.

A good example of the importance of raw materials to the development of society in the Aegean is perhaps afforded by the Cycladic culture. It is well known that the island of Milos in the Cyclades was in the Neolithic and Early Bronze Age periods one of the most important sources of obsidian in the Eastern Mediterranean. It is not inconceivable that this fact was a spring-board for the comparative wealth of the Cyclades in the Early Bronze Age. This wealth was possibly extended by further exploitation of small Cycladic sources of copper and silver at the beginning of the Early Bronze Age.

Possibly in some cases in Bronze Age history the shift in the centres of wealth and power was directly connected with working out small copper, silver or gold deposits, or perhaps with a development of new, more economical (i.e. larger, more accessible or containing more easily smelted ores) ore sources, or changes in, or access to, a particular trade route. To follow such patterns of change it would be most desirable to be able to study directly the sources of origin of particular goods and raw materials.

In recent years great progress has been made in the development of scientific methods of examination of archaeological objects to identify the origin of raw materials. Particularly important advances have been made in the fields of pottery provenance studies (Jones 1986) and in provenance studies of metal artefacts (Gale and Stos-Gale 1989), though there have also been developments in provenance studies of obsidian (Gale 1981) and alabaster (Gale et al. 1988).

The most important scientific technique used for metal provenance studies is called thermal ionization mass spectrometry and is used to make comparative mass spectrometric analyses of the isotopes of lead present in ancient metal artefacts and metal ores (Gale 1989).

METHODOLOGY

To be able to discover the metal sources used in the Bronze Age, apart from numerous lead isotope analyses of ancient metals and by-products of metal smelting, an extensive study of the remains of ancient mining and metallurgy is necessary. At present we have such data on the ore deposits and slags from Timna in Sinai (Gale and Stos-Gale 1984), Cyprus (Stos-Gale et al. 1986), Cycladic lead-silver deposits (Gale and Stos-Gale 1981) and many copper deposits in Greece (Gale et al. 1985; Stos-Gale et al. 1986, 129; Gale 1986). Also published are some lead isotope data for the ores in Anatolia (Pernicka et al. 1984, Seeliger et al. 1985 and Wagner et al. 1987) and Sardinia (Gale and Stos-Gale 1987). Figures 1 and 2 give the localities of some of the important copper ore deposits in the Eastern Mediterranean.

To establish the characteristic fingerprint of a given ore deposit it is necessary to measure lead isotope ratios of at least 50 samples of all types of metal ores present in this deposit. When by-products of ancient smelting or refining operations, like slags or litharge, are available then it is possible also to establish a lead isotope fingerprint of a smelting or refining site (which is usually identical to that of the near-by ores). The lead isotope analyses give us three different ratios between the abundances of the four lead isotopes. The isotopes are called according to their mass; lead 204, 206, 207 and 208. These ratios describe a well-defined, usually egg-shaped, part of three-dimensional space which is often called a 'lead isotope field'. Without a sophisticated statistical treatment it is necessary to present such a three-dimensional 'field' in two diagrams. The data presented in this paper are on diagrams which show only ratios of three of the lead isotopes, expressed as the ratios of lead 208/206 and lead 207/206. Nevertheless, our conclusions are always based on the information obtained from all three ratios available.

SOURCES OF METALS IN THE EARLY BRONZE AGE

It has been suggested in the past, by Branigan (1974) and Renfrew (1967) amongst others, that in the Bronze Age the most likely sources of metals for the peoples in the Aegean were local, sometimes quite small, deposits of copper and lead/silver ore. On the other hand Schachermeyr (1954, 1350) and others (for instance, Hartmann and Sangmeister 1972) have suggested that metals came to the Aegean from the East.

To put these theories to the test we have explored many of the ore deposits in the Cyclades and compared their lead isotope compositions with numerous Early Bronze Age lead, silver and copper artefacts from the Cyclades. The results of this research have been published elsewhere (Gale and Stos-Gale 1981; 1985) and the conclusion drawn from the lead isotope analyses is that indeed small local sources of metals seem to be quite extensively used in the Early Bronze Age Cyclades. However, it also became quite clear that the metallogenetic map of Greece cannot be used uncritically as a guide to the metal sources used by particular settlements. In fact it seems that the smelting and extraction of metals, even in the Early Bronze Age, was a centralized operation, which is not really surprising since the processes involved require a certain expertise and appropriate installations. It seems that economics was governing the Early Bronze Age metal industry in very much the same way as is true in our times. Perhaps at one time only one or two deposits in the Aegean were exploited and then the less economical ones were gradually phased out when richer or more easily exploited ore deposits were found.

This point emerges clearly from the results of lead isotope analyses of Early Cycladic metal artefacts, as chiefly two ore deposits have been so far identified as sources of metal for the Early Cycladic people; Siphnos for silver and lead and Kythnos for copper. The multi-metallic ore deposits of Lavrion in Attica were already used in the Early Bronze Age both for copper (Stos-Gale 1988; 1989) and silver (Gale et al. 1984). This result seems quite reasonable if we take into account the geographical position of these deposits (coastal areas, near to old established obsidian trading routes) and the type of ores occurring there. All three deposits contain even now small quantities of very rich ores which are most suitable for primitive smelting; also perhaps it is not accidental that the rich copper ores exploited on Kythnos and in Lavrion often contain arsenic which allowed in the Early Bronze Age the easy, if probably accidental, production of superior copper metal - arsenical copper.

Metal production in the Early Bronze Age on Siphnos and Kythnos has also been proved through excavations on these islands (see Stos-Gale et al. 1988 for Kythnos, and Wagner and Weisgerber 1985 for Siphnos). Both these islands produced remains of Bronze Age mining and metal smelting activities and in both cases the archaeological and scientific (thermoluminescence and/or C14) dating points to exploitation in the Early Bronze Age.

In his book assessing the contacts between Egypt and the Aegean in the Bronze Age, Helck (1979, 38) has remarked that the oldest (Early Minoan II) ports on Crete are on the islands of Mochlos and Pseira in north-east Crete. He concluded that their position resulted from trade with Anatolia, and that the trade route ran along the western coast of Anatolia for silver. Many more analyses of Early Minoan lead and silver from Crete are needed to test this hypothesis, but our analyses so far of lead and silver from Mochlos suggest a different contact. It seems that these ports were getting at least some of their metal from the Cyclades. The results of lead isotope analyses of lead objects from Mochlos are presented in Figure 3; three of these objects are consistent with an origin from Siphnian ores. Throughout the Bronze Age lead was most likely not smelted for its own sake but only as a by-product of silver smelting (see Gale et al. 1984, 405 - 406); therefore it can be expected that lead and silver of the same origin might be found on one site. Likewise, the majority of copper objects from the Early Minoan I/II site of Ayia Photia are made of copper from the Cycladic island of Kythnos (Gale 1986). This is not altogether surprising because Seager mentioned typological parallels between Mochlos and the Cyclades, whilst Ayia Photia has Cycladic affinities strong enough to prompt Doumas (1976) to suggest that this site might have been a Cycladic colony on Crete.

However, the presence of metals of Cycladic origin on Crete does not necessarily mean that the users of the metals were Cycladic settlers: the lead isotope analyses of the Early Minoan objects from south-central Crete, from the Mesara tholos tombs, have proved that most of them also are made of Kythnian copper (Fig. 4) (Gale 1986). On the other hand the lead isotope data for the copper objects from the Mesara tholoi suggest also that already in Early Minoan times Crete had access to more than one source of copper.

Crete itself has only few, and rather poor occurrences of copper and no silver deposits worth exploiting (Fig. 5). It is possible that a very limited amount of copper from the southern coast of Crete (Miamou, Chrysostomos) was used in the earlier period of the Bronze Age (Gale 1986), but so far there is no evidence that these small occurrences were ever used on a large scale. It seems that the Minoan demand for copper was much greater than the local supplies of copper could have ever sustained. Therefore from the Early Bronze Age onwards the Minoans must have imported copper, as well as gold and silver.

It is certain from our work that some of the metal supplies in the Aegean Early Bronze Age came from the Cyclades. We do not yet know how far the Cycladic metals were travelling, but it is very likely that at that time the Cycladic deposits together with those of Lavrion were able to meet most of the demands of the Early Bronze Age south Aegean. The presence on Crete of some quantities of metal of quite different, possibly Near Eastern or Anatolian origin, might indicate only a more complex trading network, specific to the Early Minoan culture, rather than a general shortage of metal in the Aegean.

It is very likely that Kythnos and Lavrion were the earliest copper deposits in this area which came to the attention of its inhabitants. Our lead isotope analyses of the copper artefacts from the Late Neolithic site of Kephala on Kea prove their origin from Kythnos and Lavrion, and the lead isotope analysis of the silver necklace from the Alepotrypa Cave is consistent with its origin from Lavrion (Fig. 6). Several years ago we saw in a private collection of a Lavrion miner native silver from the Lavrion mines; the Neolithic silver objects may perhaps have been made from the native metal, though at present there is no proof of this.

The analyses of copper objects from the site of Ayia Irini on Kea show (Fig. 7) that from the Middle Bronze Age until the Late Bronze Age there was a steady supply of Lavrion copper to this Cycladic island. It was at first very surprising therefore that lead isotope data show that the assemblage of metals from the EC III settlement of Kastri on the nearby Cycladic island of Syros represents the use of several copper sources and that none of them is either Kythnos or Lavrion (Stos-Gale et al. 1984). Excavation showed that pottery of the Kastri group constitutes almost the exclusive ceramic record of the short-lived hill-forts in the Cyclades (Doumas 1988). It would be most interesting now to analyse for their lead isotope composition copper-based objects from the other sites of this group. Not less interesting is the fact that the metals from Kastri itself include the use of an advanced technology of tin bronze whilst contemporary metals from Ayia Irini on Kea, Amorgos and other Cycladic sites are made still of arsenical copper.

Professor Doumas (1988) has argued that the northern Aegean 'refugee-pirates tried to expand their authority into the Cyclades' and finally they were 'expelled from the Cyclades as is suggested from the short life of the hill-forts and by their violent abandonment.' These conclusions agree very well with the results of the lead isotope analysis. In Figure 8 the results of lead isotope analyses of objects from Kastri on Syros are compared with the results of the analyses of objects from Troy II (Gale et al. 1985). From these results, it looks very likely that the sources of copper used for the production of objects from Kastri were the same as the sources supplying copper to Troy II. In the light of the tin bronzes excavated at Kastri on Syros it is also significant that Troy is the only Anatolian site yet known which used tin bronze plentifully in the Early Bronze Age.

OXIDE INGOTS AND SOURCES OF METALS IN THE LATE BRONZE AGE

As yet we do not have full chronological evidence for metal supplies in the Aegean Bronze Age. At present we have almost no analytical data for the Middle Bronze Age period. We hope that gradually, in collaboration with Greek archaeologists and scientists, we will be able to achieve a fully comprehensive picture of the Aegean metal trade in the Bronze Age. However, our research into the Late Bronze Age copper sources gives us already an interesting insight into this problem for later times. 

In the Late Bronze Age, Cyprus is traditionally thought to play a dominant role in copper production in the Eastern Mediterranean. There is no need to emphasize here the evidence from Cyprus itself that the Cypriot copper industry in the Late Bronze Age was booming. The occurrence of Mycenaean and Minoan pottery in Cyprus, especially in the period Late Cypriot II (that is the 14th and 13th century BC), but also to some extent a century earlier, led archaeologists to believe that Mycenaeans and Minoans were coming to Cyprus to procure copper. Therefore our next question is: how much Cypriot copper can be found amongst the Mycenaean and Minoan bronzes in the Late Bronze Age?

Particularly interesting here is the problem posed by Catling (1979): Cretan burials are rich in bronzes in the period from 1650 to 1450 BC, when there is little evidence for copper working on Cyprus and not much Minoan pottery on Cypriot sites of this date. However, in the following LC II period (1450 - 1200 BC) there is abundant pottery evidence for contact between the Aegean and Cyprus, but there seems to be much less bronze in the Aegean, even if this is the period when Cyprus is producing a great deal of copper.

One of the obvious answers to this paradox would be of course that until the end of LM IB copper metal was coming to Crete not from Cyprus but from elsewhere. Obviously, this is exactly a case where lead isotope analysis can lead to a better understanding of the Bronze Age metal trade.

A very good starting point here is provided by lead isotope analyses of the LM I Cretan oxhide ingots from Ayia Triada, Kato Zakro, Tylissos and Gournia. Anthony Harding in his recent book, perhaps thoroughly disappointed with the scientific examinations of the ingots for the purpose of establishing their provenance, concluded that: 'At present the most suitable means of assessing the likely points of origin and destination of the oxhide ingots are archaeological.' (Harding 1984, 50.)

Today we must disagree with this statement; lead isotope analyses, combined with trace elemental neutron activation analysis, of a whole range of oxhide ingots is providing more and more exciting evidence about their origin. It has been proved already by the lead isotope method that the fragments of oxhide ingots found on Cyprus are made of Cypriot copper (Stos-Gale et al. 1986, 129); this is true for ingots found at the widely separated sites of Skouriotissa, Mathiati, Enkomi and Ayios Dimitrios. The discovery of LM IA pottery at the northern Cypriot sites of Toumba tou Skourou and Ayia Irini led Catling to suggest that Minoans came then to Cyprus for copper, and therefore the Ayia Triada ingots dated to LM IA should be also made of Cypriot copper. Alternatively he also suggested that the LM IA pottery might instead be evidence for immigrants in Cyprus fleeing from the destruction of Thera (Catling 1980, 12).

Are then the Cretan oxhide ingots made of Cypriot copper? The lead isotope data show that the Ayia Triada ingots are not made of copper from Cyprus (Gale and Stos-Gale 1986). These ingots originate from two quite different ore deposits of quite different geological model ages (640 Ma and 375 Ma). There are no known copper ore deposits of these ages in Cyprus, or Greece for that matter.

However, again the similarities with Trojan objects are striking (Gale and Stos-Gale 1986, 99). Figures 9 and 10 summarize the results of lead isotope analyses up-to-date of oxhide ingots. The oxhide ingot from Tylissos of LM IB period seems to originate from the same ore source as ingot B from Ayia Triada. However, the six ingots from Kato Zakro which are contemporary to those from Ayia Triada, originate mostly from different sources. Only five have been analysed so far; one of them is consistent with origin from the same ore deposit as the bulk of Ayia Triada ingots. The remaining four ingots do not originate from any copper source investigated so far and do not even form a coherent group. One of these ingots, which falls into the Ergani Maden field, is not consistent with this ore source in its third ratio. Another one falls near to a tight group of ingots from Gournia; those again do not come from Cyprus but originate from an at present unknown ore source. The remaining two ingots again have different lead isotope compositions which are not consistent with any ore deposit which we can identify geographically at present. We may note at this point that oxhide ingots from Emporio on Chios, Kea and Mycenae are made of Cypriot copper (Fig. 9, 10).

Additional interesting information comes from the silver and gold content in the copper of the ingots. It has been suggested by Tylecote et al. (1977), based on their copper smelting experiments, that gold and silver might be good trace element indicators of provenance. Our neutron activation analyses of these elements in samples of oxhide ingots from Cyprus show that their concentrations (normalized to 100 per cent copper) fall into a well-defined region of the gold-silver diagram (Fig. 11); also the gold-silver content in other Late Cypriot objects (not included on this diagram) falls into the same area. If we compare the gold-silver data of the oxhide ingots from Ayia Triada with this 'Cypriot' area, then the difference is quite clear (Fig. 12). The ingots from Kato Zakro and Gournia are also quite different in this respect, as is clear from Figure 13. On the other hand the ingots from Emporio on Chios and one from Mycenae, which have Cypriot lead isotope ratios, plot again inside the 'Cypriot' area. The two ingots from Kyme represented also on this diagram have not yet been analysed isotopically; most likely they will prove to be of non-Cypriot origin. This trace elemental information is helpful only to a certain extent: for example Cypriot copper of Middle Cypriot date does not conform to the 'Cypriot' area as defined by the Late Cypriot objects (Gale et al. 1989), although lead isotope data proves it to have been made from Cypriot ores. This difference in content of gold and silver between Middle Cypriot and Late Cypriot copper probably reflects a change of copper smelting technology, or a change from the use of oxidized to sulphidic ores, between the two periods.

The results of these analyses make it clear that the sources of copper for the LM I oxhide ingots excavated on Crete were not in Cyprus and not in Lavrion or Kythnos. Perhaps this result is not really surprising. There is a body of evidence of contacts between Crete and Syria and Mesopotamia in Early and Middle Minoan times. The studies of Near Eastern cylinder seals on Crete led Møller (1982) and Strøm (1982) to believe that the Minoans must have ceased to purchase Mesopotamian goods after 1650 BC. If the oxhide ingots came to Crete from that direction then it might mean that the contacts have ceased perhaps later (the dating of the ingots is not very exact, MM III - LM IA). Nevertheless the lead isotope analyses of the Late Minoan II objects from the Unexplored Mansion (see the next paragraph) suggest that at least after 1450 BC copper metal from these earlier sources was not reaching Crete at all. Other important evidence for strong contacts of Middle Minoan Crete with the Near East comes from the studies of ashlar masonry in the Aegean by Gunnel Hult (1983). Her studies revealed that the earliest ashlar buildings on Crete date to MM IB (beginning of the second millennium BC) more or less contemporary to similar buildings in Ugarit, whilst the technique of ashlar masonry was unknown on Cyprus, in Anatolia and Greece until much later (about the 12th century BC).

At present we have only limited evidence for sources of copper for Crete after the LM IB destruction. From Knossos we have analysed the bronze objects from the Late Minoan II Unexplored Mansion (Stos-Gale et al. 1986a, 134 and Fig. 14). The majority of analysed objects in this hoard (fifteen in total) is made of copper from Lavrion, three are made of Cypriot copper and another three come from an ore deposit of at present unknown geographical location and distinctly dissimilar to any sources of copper used on Crete in earlier times.

Another body of evidence for Cretan copper supplies comes from several Late Bronze Age sites in Western Crete (Fig. 15). Again, the dominant source of copper there from LM IA to LM IIIB was Lavrion. Four objects of LM III date are made of Cypriot copper, and one object of LM IIIB1 date from Armeni might be made of copper from Crete. Two objects of LM IIIB date are possibly of the same unknown origin as the three objects from the Unexplored Mansion mentioned previously. Access to samples of copper from Akrotiri on Thera has so far been rather limited. Figure 16 shows lead isotope analyses of ten LM IA bronzes, showing that six are apparently of Lavrion copper and four of copper from Cyprus.

A summary of the lead isotope evidence which we have at present for the sources of copper for Crete throughout the Bronze Age is presented in graphic form in Figure 18. It seems that perhaps before LM II Cypriot copper was not at all coming to Crete. Even in the later part of the Bronze Age Cyprus does not seem to be a main supplier of copper to Crete. The majority of the Late Minoan objects from the Unexplored Mansion and from Western Crete originate from Lavrion. It is possible that after the destruction of Thera copper from Lavrion was coming to Crete via Kea, Kythera and West Cretan ports. The increase of importance of Western Crete for Minoan trade after the eruption of Thera is represented in the increase of quantity of pottery from Chania on Kythera and Kea. It is not impossible that not only were the trade routes after the disaster changed to avoid the neighbourhood of Thera, but also that Knossos itself suffered from decline due to this natural disaster.

Returning to Catling's paradox we are left now with a question: where was the bulk of Cypriot copper exported? There is compelling evidence that the development of Cyprus in LC II took place in the framework of continuing contacts mainly with Syria, including Ugarit. One of the examples of these contacts is the rich shaft grave of LC IIA date excavated in Hala Sultan Tekke, containing rare bronze items with parallels in Western Asia, especially Palestine (Åström 1983). There is good reason to expect that at a certain point in Late Cypriot II or III Cyprus was drawn into the Near Eastern metal trade and only then started to export its copper in the shape of oxhide ingots to comply with Near Eastern and perhaps Egyptian trade standards. It is also just possible that in earlier times Cypriot copper was not cast into oxhide ingots on Cyprus itself.

The excavators of Ras Ibn Hani in Syria, Drs J. and E. Lagarce, gave us for lead isotope analysis two droplets of metal found near the only mould for an oxhide ingot as yet excavated (Largace et al. 1983). The results are presented in Figure 10 and it can be seen that this metal is of undisputed Cypriot origin.

Undoubtedly Cypriot copper in the later periods of the Bronze Age was also travelling to the West. Our question now is - how far West? Figure 10 presents some lead isotope analyses bearing on this question: one oxhide ingot fragment from Emporio on Chios of LH IIIC date is made of Cypriot copper; the whole oxhide ingot excavated by Tsountas in Mycenae, at present in the Numismatic Museum in Athens, is also made of Cypriot copper; fragments of oxhide ingots from the island of Kea are also Cypriot. However, the largest number of oxhide ingots and fragments found in the Western Mediterranean is on Sardinia. Is it possible that those also are of Cypriot origin? Of course Sardinia itself has ample supplies of copper. It has often been suggested by various scholars that the presence of Mycenaean pottery on Sardinia represents an Eastern Mediterranean quest for copper. The obvious conclusion therefore seems to be that these ingots are there because they were produced on Sardinia, from Sardinian copper, perhaps for export to the Eastern Mediterranean.

Thanks to the kind collaboration of Dr Lo Schiavo we were able to collect copper and lead ores from numerous deposits on Sardinia, though it should be pointed out here that on Sardinia lead deposits far outnumber copper ore deposits. The geology of Sardinia is quite complex, and therefore different occurrences, even different mines, have quite characteristic lead isotope compositions, differing according to their geological age (Gale and Stos-Gale 1987). Most of the Sardinian copper deposits also contain lead ore and it is quite probable that in prehistoric times the same ore deposits may have been exploited for lead and copper. Compared with Crete, the Cyclades or even Cyprus, Sardinia has a great wealth of ore deposits of lead, silver, copper and also a little tin. At present the excavations on Sardinia provide a great amount of evidence for Nuraghic metallurgy of copper, lead and iron. Unfortunately the dating of the Nuraghic finds is not very precise because of a shortage of reliable stratigraphy. It is especially difficult to relate chronologically the Sardinian oxhide ingots to other Nuraghic metal finds, since most oxhide ingot finds on Sardinia have no stratigraphy at all. Lo Schiavo (1985, 258 - 260) concluded that the archaeological evidence on Sardinia favours the existence of relations between Cyprus and Sardinia in the 13th century, while the rich local production is dated mostly to the 12th and 11th century BC. Since the ingots in the Eastern Mediterranean can be fairly accurately dated to the 13th and 12th centuries BC, one cannot exclude altogether the possibility that the oxhide ingots came from outside Sardinia before the Nuraghic metallurgy began to flourish to any extent. In fact there is more and more evidence from the archaeological excavations on Sardinia that Eastern Aegean goods must have been reaching this island at least as early as the 13th century BC (Ferrarese-Ceruti 1985). There is also evidence for a 'special relationship' between Sardinia and Cyprus, expressed in the similarity of their metalwork (Lo Schiavo et al. 1985). Also, recent scientific examination of imported Aegean pottery on Sardinia (Vagnetti and Jones 1988), apart from numerous examples of Mycenaean and Minoan fine pottery, has identified some coarse pithoi fragments excavated in Nuraghe Antigori as originating from Cyprus (best parallel with Maroni) and Kommos on Crete.

With all this information to hand perhaps we should not be too surprised to find that all of the oxhide ingots and fragments from Sardinia which we have analysed so far are fully consistent with the lead isotope composition of Cypriot copper ores. On the other hand the bun or plano-convex ingots found also in Nuraghic contexts on Sardinia are without doubt made of Sardinian copper and are quite different from the oxhide ingots. In addition the Nuraghic Santa Maria in Paulis metal hoard is wholly made of Sardinian copper; so that there is no doubt that the indigenous Nuraghic metal industry used Sardinian copper ores (Gale and Stos-Gale 1987).

Perhaps the most interesting result so far of our investigation of the origin of oxhide ingots is the fact that nearly all oxhide ingots outside Crete, all of them of course later in date than the Minoan examples, are made of Cypriot copper. In contrast none of the (earlier) Cretan ingots are Cypriot, and they most probably originate from the Near East. It is too early yet to suggest sweeping conclusions, but our results indicate that the oxhide ingot trade did not originate on Cyprus, but somewhere in the Near East. At about the end of the 15th century, perhaps, this trade ceased in the Eastern Mediterranean and then Cypriot copper filled the vacuum. However, even then Cypriot copper appears not to have been a major source of this metal in the Mycenaean world. At the moment our earliest evidence for the presence of Cypriot copper metal in the Aegean is some scraps of metal from Late Minoan I Thera (Stos-Gale 1988) where three objects out of nine come most probably from Cyprus. (We must be cautious at this point because some of the samples for analysis were badly corroded and might have been contaminated through the corrosion.) Further west, there is some Cypriot copper on Ayia Irini on Kea and in Mycenae, in Perati and Tiryns; in particular there are oxhide ingot fragments of Cypriot copper in the Poros Wall Hoard at Mycenae (Demakopoulou et al. 1987).

However, the production of copper from Lavrion copper ores in the Late Bronze Age must also have been quite substantial. Apart from the Lavrion copper finds on Crete, there is copper from Lavrion on Kea (see Fig. 7) (Stos-Gale 1988), in Perati, Menidi and Tiryns (Demakopoulou et al. 1987), and finally, a large find of double axes and tripods from a Late Helladic III excavation of Dr Onassoglou in Mycenae, is all a product of Lavrion copper (Fig. 17).

CUPELLATION OF SILVER AT AKROTIRI - A CYCLADIC MONOPOLY?

Very little so far has been published about the possible existence of silver workshops on Akrotiri. Marinatos mentioned a little workshop of metallurgy in the basement cellar of rooms 4 and 4a of the West House (S. Marinatos 1974, 29).

Further on the same page Marinatos remarks that: 'Lead and possibly silver were extracted from the keroussite and galene, which existed and still exist on the island.' On the floor in this room a broken pot was also found '...still filled with a whitish substance, much like pumice, but heavier. Near the pot, a bowl containing the same material was found (Pl. 64a) ... Apparently we have before us a method of cupellation for extracting possibly silver from the lead minerals.' (S. Marinatos 1974, 29 - 30, and Pl. 63b, 64a.) Nearby, stone hammers and grinding stones were also found.

In summer 1978, after the Second Thera Congress, we received a sample of litharge from the West House for lead isotope analysis from Dr Puchelt, who was given it by Marinatos (No. LWHF in Gale and Stos-Gale 1981). This was a sample of litharge (as was recognized both by us and by Dr Puchelt), not cerussite. Also, during the Second Thera Congress in 1978 we were allowed by Dr Doumas to sample lead from the apothiki on Akrotiri and there we were shown two pieces of litharge, which were labelled as found in room Delta 9 (Nos. 1915 and 1916 in Gale and Stos-Gale 1981). The context and dimensions of the litharge sampled by us are given in Table 1. The information about the context was copied at the time from labels attached to these pieces; the measurements were taken by us.

TABLE 1. Description of litharge pieces sampled by Gale and Stos-Gale on Akrotiri in August 1978

A heavy stone, possibly from room Delta 16, was examined by Konophagos (S. Marinatos 1972, 35), who concluded that it was cerussite. However, if, as is possible, the X-ray analysis performed by Konophagos was an X-ray diffraction analysis of the material scraped from the exterior surface of the stone, this could easily have been merely the exterior weathered surface of a mass of litharge, since litharge (PbO) does tend to acquire with time an external crust of cerussite (PbCO₃).

All this information is important, since as far as we are aware, litharge is not mentioned at all in Thera excavation reports; litharge was certainly excavated at Akrotiri, however, though not recognized as such.

Litharge (tetragonal lead monoxide) does not occur in nature and on archaeological sites is associated with cupellation, that is, extraction of silver from silver-rich lead. It is formed when lead is heated in air and therefore is a common by-product of the cupellation process. The colour of litharge can vary from grey, through yellow to buff red. The name itself was derived from 'lithargyros' - silver stone. Both Pliny (33.35) and Dioscorides (5.62) refer to the derivation of litharge from cupelling lead for silver. Pliny calls litharge 'spuma argenti'. In Classical times in Greece litharge was commonly used as a pigment and perhaps for medicinal purposes. As far as we know there is no evidence from Bronze Age sites in Greece that litharge was used at the time as a pigment. (This is confirmed by a personal communication from Ellen Davies to Malcolm Wiener.) The analyses of pigments on Theran frescoes (Filippakis 1978) do not show lead as a major component in any pigment. Nor have analyses of pigments on Aegean pottery ever suggested the use of lead pigments.

In Classicar times in Lavrion litharge was used as a constituent of hydraulic cement used for lining water cisterns and ore washing tables, but no use of litharge for any purpose has yet been demonstrated by excavations of Bronze Age sites.

By heating litharge in reducing conditions it is possible to convert it again to lead. However, our chemical analyses of Bronze Age Aegean lead and litharge show that the lead metal usually has 10 to 100 times higher silver content than litharge; this suggests that litharge was not converted to lead after the silver was extracted but was thrown away, as was possibly the case on Kea, where Caskey mentions in his excavation report (Caskey 1962, 273) '...a great many lumps of litharge (PbO) at Aghia lrini.' In a footnote on the same page he acknowledges R.H. Brill, who identified the pieces of litharge for him.

Cerussite, in contrast with litharge, is a natural lead carbonate formed commonly by the action of carbonated waters on lead ores, or as a corrosion product on lead artefacts and also on the surface of pieces of litharge if buried underground. As suggested earlier, it is very likely that pieces of litharge excavated by Marinatos have been repeatedly mis-identified by scientists as cerussite if actual samples for analyses were taken only by scraping the surface of these heavy 'pebbles', which did consist of a cerussite alteration product on the surface of a mass of litharge. In 1978 we were allowed to break two of the objects which we sampled (1915 and 1916) which revealed the interior to contain bands of pink and purple litharge with characteristic crystalline structure. On the whole it is quite easy to distinguish with a naked eye between litharge, which has a well-defined crystal structure, and white, powdery cerussite. We are quite sure that our samples are of litharge. Therefore it seems reasonable to conclude that silver smelting was conducted on Akrotiri, as was also the case on Kea and Milos.

The presence of litharge on an archaeological site should play an important role in any assessment of the economy and trade of its inhabitants. Using the lead isotope technique we can trace the origin of the lead-silver ore and, as will be seen later, sometimes rather unexpected conclusions can be drawn.

The Aegean is rather well-endowed with lead/silver deposits. Therefore it is not difficult, just on the basis of the map of mineral deposits of Greece, to find for nearly each larger Bronze Age site a suitable, nearby lead/silver ore deposit, from which the inhabitants of the settlement could in theory smelt their own metals. Several archaeologists have suggested this possibility in the past, most of them, it seems mainly as a counter to Schachermeyr's diffusionist theory and the suggestions by Hartmann and Sangmeister that metals came to the Aegean largely from the East.

The theory of metal smelting, be it copper or lead, as a 'cottage industry' on each site, however, does not bear scrutiny. Firstly, not all of the metal deposits on the map are suitable mineralogically for ancient technologies. In the case of lead ores, the silver content is also in many mineralizations far too low to be of any interest to ancient metallurgists. Secondly, metal smelting is a skilled job and requires specialization of a similar type to that needed for decorating pottery, or the manufacture of stone vessels. To produce any metal on a larger scale a proper social organization is necessary just as much as access to the source of ore, which itself has to be recognized for what it is. It is possible that at the beginning of the Late Bronze Age the inhabitants of the Cyclades played in this respect an important role in silver production in the Aegean. The proof of this occupation lies in the finds of litharge on the sites of Phylakopi, Ayia Irini and Akrotiri. With the presence of some lead ore on most Cycladic islands the occurrence of lead and silver objects and litharge at Cycladic sites is not necessarily surprising. However, our lead isotope and chemical analyses of galenas from the Cyclades and lead and litharge from Cycladic sites, as well as lead from Crete, show a quite different picture from that of local exploitation (Gale and Stos-Gale 1981). Rather surprisingly all of this LBA lead and litharge (including the lead weights from Thera and Crete) is consistent with their origin from the lead/silver deposits at Lavrion. Additionally the LBA lead metal has a rather high silver content, which indicates that very rich lead/silver ores were extracted. So far, on the bases of lead isotope and chemical analyses most of the Cycladic lead ore occurrences can be discounted as sources of metal for these objects. We have analysed most known lead/silver ores from the Cyclades.

The lead isotope compositions for the different ore deposits do not spread greatly in most cases and there is very little overlap between the isotopic fingerprints of different islands. The only galenas which plot close to the field established for the Lavrion deposit are those from Kea, Makronissos, Seriphos and Anaphi. We collected samples of the ores and examined these occurrences ourselves (only the one sample from Makronissos was given to us). On all these islands occurrences of galena are very small and there are no traces of any ancient exploitation of the ore. Moreover the silver content in contemporary ore is not higher than 300 ppm. On Kea galena occurs in discordant veins of marble on the southern tip of the island; the quantity is not great but it was profitable enough for short and limited exploitation at the end of the previous century. There are definitely no traces of any old mining or metal smelting anywhere in the vicinity of these few outcrops. Finally, the overall low silver content in this ore precludes its smelting for primitive silver recovery.

In fact our lead isotope analyses of copper, lead and silver from Greek archaeological sites of different periods of the Bronze Age suggest at present that there was no great number of metal sources used in the Aegean at that time. It seems that the principles of economics were just as important then as now: the metal deposits exploited in the Bronze Age were very few and limited to the richest ones (which guaranteed mineralization suitable for quick and efficient extraction), of easy access and with a good supply of wood for smelting. It seems that, throughout the Bronze Age, Lavrion in Attica fulfilled all these conditions, both for copper and lead. Lavrion is certainly one of the most spectacular deposits of rich silver ores of all kinds. This must have been discovered quite early, since our lead isotope analyses of silver from the Late Neolithic Alepotrypa cave site suggest that it originated from the island of Siphnos, though some does come from Lavrion.

Silver extraction at the three Cycladic sites on Kea, Milos and Thera was carried out inside the workshops, in inner rooms (Kea) or cellars (Thera). It seems that such a pattern would fit a theory where lead sulphide (galena) would be roasted, perhaps near the mines, to remove the sulphur as sulphur dioxide, followed by smelting to produce raw lead metal. This product would then be transported to a much more secluded site, where an assay of the silver content by cupellation could be made. Following that, lead with sufficiently high silver content could be cupelled on a larger scale and silver extracted; the lead with insufficient silver content could be used for lead weights, etc.

If this theory is right, then some rather interesting social and economic models might emerge for the organization of the mining operations and export of metal from Lavrion. It is possible that the production of copper metal was done at Lavrion (though some copper smelting slags were excavated at Ayia Irini - lead isotope data on some of these slags are consistent with Lavrion again), but the presence of litharge produced from lead metal derived from Lavrion ore on many archaeological sites suggests that silver extraction from raw material originating from Lavrion was often carried out far from Lavrion.

Perhaps silver had not only special quality as an item of luxury and social stratification. It is known that Minoans delighted in noble metal vessels and cups, some of which might have had special ritual or religious meaning. For example, Dr Karageorghis, in his paper for this conference, describes a Near Eastern ritual of offering cut hair to the gods in gold and silver vessels. If Minoan silver vessels had similar use and if the religious and state authorities were inseparable, then it is possible that the production of silver was a specially controlled activity. Further, the excavators in Mallia and Kato Zakro mention that in both of these palaces there were workshops of 'precious objects for sacred use.' (Chryssoulaki and Platon 1987.) Perhaps objects of silver were also counted among them.

On the other hand, of course, silver was perhaps a very useful commodity in the trade for tin, which was not readily available in the Aegean and had to be imported. Malcolm Wiener (1987) concludes that: 'In interstate trade the value of a coin in antiquity was simply its approximate bullion value. In Babylonian and Assyrian texts silver often serves as a unit of measure...' and further: 'Although copper vessels were used regularly in the Ramesside period as currency their value was usually stated in terms of silver.'

If in the Minoan state silver played effectively the role of currency, that again, would be a good reason to keep the silver production under state control in some way. Perhaps the authorities licensed craftsmen to extract silver, and in return provided them with a supply of excellent quality raw material (silver-rich lead metal).

So far we do not know of any finds of litharge dated to LH I on the Mainland, save for some litharge in Middle Helladic strata at Thorikos. On Crete, as far as we know, only some litharge was found in the LM II Unexplored Mansion. It would be interesting to study the distribution of this material on other Aegean sites and then to analyse their lead isotope composition for their origin. Silver extraction is in some ways a unique process of metal production in that it can be carried out with comparatively little equipment on a small scale and bring a large profit, providing that the technology is well known and good raw material is cheaply supplied.

CONCLUSIONS

There is no doubt at present that Cypriot copper was coming to the Aegean in some quantity in the Late Bronze Age (possibly from the end of the 15th century BC and certainly in the 13th century BC). But there is no evidence that the trade in copper from Cyprus was vital to the Mycenaean civilization. It looks almost as if the copper from Cyprus was reaching the Aegean as a sideline trade. On the other hand it is possible that Lavrion copper did not travel beyond Crete and the Cyclades. There is no evidence as yet that copper from Lavrion was cast into oxhide ingots, whilst the majority of objects counted as Cypriot copper in the Mediterranean outside Cyprus are oxhide ingots or their fragments, dating mostly to Late Bronze Age III times. Figure 19 summarizes our present evidence about the sources of copper for particular sites at particular periods of the Late Bronze Age. Our hypothesis based on the lead isotope evidence so far is that the Minoan and Mycenaean long distance trade in Late Bronze Age II - III times may not have been chiefly concerned to obtain copper; it may rather have been concerned more with the import of tin and perhaps gold. Would it be from the East via Ugarit and Cyprus, or perhaps from the north or the west Mediterranean?

The Lavrion mines provided throughout the Bronze Age relatively large amounts of silver. We have been able to analyse only a few silver objects from the Shaft Graves in Mycenae. Most of these seem to have come from Lavrion. Large amounts of lead from Tiryns and Mycenae originate wholly from Lavrion; this seems to be true also for Crete, Kea, Milos and Thera (see Fig. 20). It is not likely that this lead was smelted for its own sake -almost certainly it provides a proof of extensive silver production. This silver would provide an excellent commodity for trade abroad for tin, and together with largely sufficient resources of copper for local needs, it might have been a sound basis for the Mycenaean economic power. It is possible that the Theran economy benefited not only from participating in the exchange of ready made goods, but that the island was also involved in a certain amount of processing of raw materials into desirable objects of trade. The presence of large cakes of litharge (of Lavrion origin) at Akrotiri proves that some production of silver, either from silver-bearing galena, or silver-rich lead metal, was carried out on that site. There is no doubt that trade and commerce blossomed in the Aegean in the Late Bronze Age. The crucial geographical position of Thera, on the crossroads between the western coast of Anatolia, Crete, Cyprus and mainland Greece may have given it a prominent position in the development of the metal trade between the eastern and western Aegean.

----------------------------------------

----------------------------------------