Some Aspects of Lead and Silver Mining in the Aegean

I.     THE LEAD ISOTOPE TECHNIQUE

Lead is one of the few metallic elements which naturally has a significantly variable isotopic composition. Moreover ancient artefacts made of silver, gold, bronze and lead contain small or large amounts of lead, which is not surprising in view of the relatively simple extraction and refining methods used. In any case modern techniques for the isotopic analysis of lead are very sensitive, and easily permit the analysis of as little as 1 microgram of lead from an object containing as little as 0.01 % of lead; the isotopic composition can be established to better than ± 0.1%.

The potential of the study of the isotopic composition of lead in archaeological artefacts is as a "finger-print" technique, bearing on such matters as the provenance of the artefact and questions of authenticity. In particular terrestrial lead is one of the few elements which varies markedly in its isotopic composition in a manner dependent upon the geological and geochemical history leading to the formation of a particular lead ore (Doe 1970). Hence the isotopic composition of lead may vary from time to time and may be unique for each of a number of mines. It follows that the isotopic composition of lead in an artifact may in principle be matched with the isotopic composition of lead in lead ores from ancient mining areas so as to trace the mining region from which the lead in the antiquity was derived. The oasis of the lead isotope technique is discussed in more detail in reference (Gale 1978).

The technique has an important advantage over ordinary chemical analysis in that the isotopic composition of lead is not measurably affected by extraction, refining, or corrosion processes, so that whether the lead occurs as a major element in an artefact (lead sheet, lead pigments, kohl etc.) or as a minor element (bronzes, silver, gold, glasses and glazes etc.) the isotopic composition of that lead will be the same as the composition of the lead in the ore from which it derives.

The most serious problem in the application of lead isotope ratios to archaeology is the re-melting and mixing of metals originating from different geographical areas. Such mixing will tend to invalidate both lead isotope and chemical analyses. For example, at an early date many coins of the Roman Empire were struck in Rome, where silver from all parts of the Empire was mixed. With glazes, one does not have this problem of re-melting and mixing.

A further difficulty is that it is geologically probable that lead ores from some quite different geographical regions will prove to have lead isotopic compositions that are indistinguishable even when the most precise modem techniques are used. It may prove in some cases (silver cupelled from lead is an example) that chemical analyses will provide extra evidence to eliminate some isotopically possible ore sources, but in other cases an appeal also to archaeological plausibility may be essential.

Another difficulty is that there may be considerable isotopic variation within a single mining region, although no examples of this have yet been proved in archaeologically important areas. There are two well-documented modern studies - Laurion (Barnes et al. 1974) and Bleiberg (Koppel & Kostelka 1976) - where excellent isotopic homogeneity has been demonstrated throughout the mining district.

From our own studies we have data which suggests that there is isotopic homogeneity within the Laurion; La Union, Cartagena; Pontokerasia (Macedonia) and Siphnos mining regions.

II.     NATURAL SOURCES OF SILVER AND LEAD

The common primary igneous lead ore mineral is galena (PbS), and this mineral is undoubtedly the source of the great majority of lead from prehistoric times to the present day. Native lead metal does occur, but is extremely rare; it is most unlikely to have been used by early man. If used at all an artefact made from native lead would be very easy to identify by chemical analysis, since native lead is very pure - a typical analysis (Stos-Fertner & Gale) gives 5.2 ppm Ag, 0.01 ppm Au, 144.8 ppm Cu, 1.7 ppm Sb, 11.55 ppm As, 99.9998% Pb.

A typical lead/zinc or copper/lead/zinc ore vein is often capped by an iron-stained gossan (or iron-hat) of silicate clays and hydroxy carbonates; below this there is often a leached zone containing concentrated, re-precipitated oxidized metal in the form of sulphate, carbonate, chloride and sometimes native metal. In this zone lead can occur as cerussite (PbCO₃) and anglesite (PbSO₄), both free of silver unless (rarely) mixed mechanically with native silver. Cerussite is easily reduced to lead, and may be the mineral first mined and used for lead production at Thorikos.

Silver can occur (rarely) as the native metal in primary igneous deposits (chiefly in Norway, Bohemia, Canada) but the most common primary igneous minerals of silver are argentite (Ag₂S) and pyrargyrite (Ag₃SbS₃); both occur predominantly as exsolved minute specks in large host galena crystals. Primary native silver consitutes a minor proportion of primary igneous silver; most primary silver occurs in galena and most silver in ancient times must have been derived from it.

Below the leached zone in a lead/zinc deposit occurs an oxidized zone, just above the water table. This oxidized zone often contains cerargyrite (Ag Cl) and native silver mixed with some cerussite and anglesite. It is conceivable that early man may have made some minor use of both native metal and chloride ore (easily reduced to the metal) as a source of silver; use of native silver for an artefact would be easily proved by chemical analysis, since the native metal is exceptionally pure; >99.5% and usually >99.98% pure (Patterson 1971). It is in the oxidized zone of weathered galena veins that the small and extremely rare deposits of native lead are usually found.

It should not be overlooked that native gold, which occurs either as a primary lode deposit requiring mining or as alluvial nuggets or particles in placer deposits, can contain very large amounts of silver, certainly up to 40 - 45%. At some stage in Greek history the parting of gold from silver, probably by salt cementation, may have become known; indeed there is evidence that this process was used at Sardis in the time of Croesus (Hanfmann & Waldbaum 1970).

III.     LEAD AND SILVER MINES IN ANCIENT GREECE

The location of gold, silver and lead mines in Thrace and Macedonia was discussed at length by Casson 1926; his discussion of Macedonia was supplemented and corrected by a very thorough account by Davies (1932). Forbes (1964) gave a wide-ranging but ill-digested account with insufficient detail. Davies gave both a short general description (1932) and also the most extensive and accurate survey yet to appear (1935). Valuable information also appears in the general surveys of Fiedler (1840) and Ross (1840).

There are many reports of lead and silver mines in the classical literature. Within the Aegean area four main centres are attested by ancient authors: Laurion in Attica, the island of Siphnos, the island of Thasos and various districts in Macedonia and Thrace. Figures 1, 2 and 3 show the geographical location of these centres. A further important centre is that of Sardis in Lydia, the mint of the proverbially wealthy King Croesus.

Finds of silver-poor litharge and lead at Thorikos in Laurion, in an intact 9th century B.C. bed (Bingen 1964, 26 - 30) and in a bed dated to about 1500 B.C. (Servais 1965, 20 - 24) suggest that silver mining and the cupellation of argentiferous galena was practised at least from the end of the Middle Bronze Age; but the ancient authors record that the richest and deepest levels were exploited fully only in 483 B.C., (Ath. Pol. 27,7; Her. VII, 144). The question of the earliest working of Laurion will be discussed in more detail in section V.

The Aegean island of Siphnos was reported by Herodotus (III, 57) to have possessed in the late sixth century rich mines of gold and silver which enabled the Siphnians to build at Delphi a treasury of unusual magnificence. Pausanias (X, II, 2) records that at some unspecified date (probably in the 5th century B.C.) the mines were flooded by the sea. However, until recent field work by a Heidelberg/Oxford and then a Heidelberg/Bochum team there has been no certain evidence for the existence of lead/silver mines on Siphnos; there now is such evidence, including ancient shafts now below sea level, as reported by Wagner and Weisgerber (1978) and Gentner et al. (1978). In the N. Aegean there were various mining areas in Macedonia and Thrace. A mine near Lake Prasias (locality uncertain, see Davies 1932, 152) is said to have produced a talent of silver per day for Alexander I of Macedon (Her. V, 17), and the same ruler seems to have obtained at least partial control of mines around the river Strymon and Mount Pangaeus (see Fig. 8) which in the 6th century had been in the hands of various Macedonian tribes. 

Herodotus (VI, 46f) records that Thasos controlled mines both on its own island and on the opposite mainland, and that mining in Thasos itself was initiated by the Phoenicians in the early 1st millennium B.C. In 360 B.C. the Thasians are recorded (Diod XVI, 3, 7) as having established a settlement on the mainland at Krenides, "the Springs", and to have begun minting gold staters there; in 357 B.C. Philip II of Macedon took Krenides under his protection, renamed it Philippi (Diod. XVI, 8, 6) and issued gold and silver staters of excellent purity; both Philip's gold and silver made a tremendous impact on the ancient world.

Of the four main silver mining regions reported in the ancient literature the exact location of only one, Laurion, is now definitely known. On Siphnos the site of the ancient silver and gold mines is usually considered to be the Ayios Sostis peninsula; a more detailed account will be given in section VII. For Thasos Herodotus clearly states (VI, 47) that some of the mines lie on the eastern side of the island, facing Samothrace; however ancient silver mines are now known only on the western side of Thasos (Genther et al. 1978, 275; Davies 1932, 159 - 160). Little is known about the site of the ancient silver or gold mines in the Pangaeon region or on the mainland opposite Thasos. Pangaeon must have been a mining centre in 510 B.C. (Her. V, 23), probably in Pisistratus' day (Her. 1, 64) and perhaps much earlier; the best account of vestiges of ancient mining in this region is by Davies (1932, 155ff), who also discusses the region aroung Philippi.

IV.     LEAD ISOTOPE COMPOSITION OF ORES FROM THE AEGEAN REGION

In 1975 a collaborative programme began, financed by the Stiftung Volkswagenwerk, between the Max-Planck-Institut für Kernphysik, Heidelberg, and the Geological Age and Isotope Laboratory, Department of Geology, Oxford directed towards the attempt to discover the source of silver used in Archaic Greek silver coins, the attempt to locate ancient lead/silver mines and the investigation of ancient lead/silver metallurgy.

Part of the research involved fieldwork in the summers of 1975, 1976 and 1977 in the attempt to locate more exactly the silver mines mentioned by Herodotus; those chiefly involved have been W. Gentner, G.A. Wagner, O. Müller, J.D. Stavropodis, S. Manolarakis, H. Gropengiesser, H.G. Conrad, G. Weissgerber and N.H. Gale.

Many other sites mentioned in the archaeological, geographical and geological literature of the 19th and 20th centuries may well have served as lead/silver sources for ancient Greece. About 70 such sites have been visited and all sites at which traces of mining and/or smelting activities for lead were found and are shown on Figure 1. At this stage no claim is made that each of these sites is necessarily an ancient lead or silver mine; combined chemical, geological, physical and archaeological studies will be necessary to determine this. Further details of some of these sites are given in Gentner et al., Table 1.

In order to provide a basis of comparison for the work on lead isotopes in Aegean metallurgy and numismatics, analyses have been made of a number of galenas from mainland Greece and Turkey, especially Laurion, and also of galenas from Aegean islands. Samples collected on our own field trips are identified by "TG" numbers on Fig. 1 and in subsequent tables.

Table 1 and Figure 4 give the lead isotope ratios for galenas chiefly from the mainland; Tables 2 and Figure 5 give data for the Aegean islands (Fig. 5 gives also data for some lead artefacts). Figure 5 illustrates that ores form Laurion are clearly distinguishable from, for instance, Ayios Sostis on Siphnos, Thera, Lesbos and Samos. Taken together Figures 4 and 5 show that ores from places in Greece and Asia Minor mostly populate the field between Laurion and Siphnos. Some further data presented in Gale 1978, Fig. 2, shows that ores from Sardinia and Spanish ores from Los Linares and Rio Tinto fall in the uppermost part of the lead isotope diagram, are are clearly distinguishable from ores coming from around the Aegean.

The Laurion region has been studied in especial detail. The work of Barnes et al. (1974) first demonstrated that the isotopic composition of lead from modem Laurion galena and oxide samples, though essentially constant, had a small range of about 0.3%. We have extended the geographical and geological extent of the sampling of modem samples of Laurion galena, and our data (together with part of that reported by Barnes et al. 1974) is given in Figure 6 and in Table 3.

Also plotted is our measurement of the composition of the Thorikos ore sample C412, which is a mixture of cerussite and litharge from a large dump found by H. Mussche in the Thorikos excavations between an ancient mine and an ore washery in the fourth century B.C. level. The Oxford results plus those of Barnes et al., (1974) define very well the field of Laurion lead isotope composition, and it is satisfying that the 4th century B.C. ore sample from Thorikos falls within this field. Analyses of further lead samples (lead sheet, lead pottery repair rivets etc) from sixth to fourth century B.C. levels at Thorikos, five samples in all, are also given in Figure 6 and more are in hand. The data obtained so far establishes that in Laurion lead worked in earlier times all falls in the same isotopic field as that from modern ores at Laurion. Figure 6 shows also the distinctive composition of galenas now exposed near Akrotiri on the island of Thera.

The assumption by numismatists that Laurion was the source of the for most Athenian coins has now been tested by the analyses by Brill and Shields of seven Athenian bronze coins (fourth to first century B.C., see Brill and Shields 1972, 279 - 303 for descriptions) and at Oxford of eight Archaic silver Athenian tetradrachms (from the Asyut hoard, pre 475 B.C.) and three second century B.C. silver Athenian tetradrachms.

Figure 7 shows the data for these coins, all of which fall within the Laurion field within experimental uncertainty. The Athenian currency decree (450 - 420 B.C.) and the use of the Delian League treasure brought into Athens large amounts of silver of diverse geographical provenance. It would be interesting to investigate in some detail the isotopic composition of lead in Athenian silver coins of this period to see if these events are recorded in a change of isotopic composition from the Laurion composition so far observed. Müller and Gentner (1978) have shown that Athenian coins have also a distinctive chemical composition.

These measurements establish that the isotopic composition of lead from Laurion is well-defined and unlike that for any other ancient mining region so far sampled. To the extent that this is true any artefact which proves to have a lead isotope composition falling within the Laurion field may be assumed to contain lead deriving from Laurion.

Independent of the uniqueness of the Laurion composition, any lead or silver artefact containing lead of an isotopic composition outside the Laurion field cannot be derived from Laurion.

The credibility of these statements is vital to subsequent sections of this article, so that it is necessary to clear out of the way a misconception contained in the excellent review by Buchholz (1972) of lead in the Mycenaean culture. The misconception (Buchholz op. cit. 57 - 58) is that the characteristic of Laurion lead is that it has an isotopic composition lying anywhere within the field of compositions contained in 'Group L' as defined in the early paper by Brill and Wampler (1967) on isotope studies of ancient lead. The analytical techniques used in this early work give isotopic compositions of doubtful accuracy, but more important is the fact that 'Group L' was one of four quite arbitrary divisions made by Brill and Wampler merely as an aid to the discussion of the preliminary lead isotope data which they presented. As defined in Brill and Wampler (1967) 'Group L' is a field much larger than the field (the Laurion field) which we now know to include Laurion lead sensu stricto. The approximate relation between these two fields is shown in Figure 8; it is clear that an artefact can have a composition falling within 'Group L' as originally defined, but far outside the Laurion field characteristic of the Laurion mining region.

It is this quite undestandable confusion which led Buchholz to conclude, wrongly, that lead objects from the Cape Gelidonya ship wreck, from Cyprus and Sardis, together with lead ores from Turkey and Iran, all have a Laurion composition (1967, 55, 57 - 58); they do not. The same confusion led Buchholz to conclude that the lead isotope method had only a limited prospect of further clarifying sources of lead in the region Greece-Anatolia-Cyprus; this may yet turn out to be true, but at present the situation looks promising.

V.     THE DATE OF THE EARLIEST WORKING OF THE MINES AT LAURION

The history and some aspects of the geology of Laurion have been discussed by Ardaillon (1897) and Hopper (1968), who list most other important references. The geology is discussed in some detail by Marinos and Petraschek (1956).

The date at which the exploitation of the Laurion mines began has been a matter of controversy, with such scholars as Blümner affirming that significant expoitation began only in Archaic times (1887). Xenophon (De Vectigalibus 4, 2), expressing the view of the ancient Athenians, connected the beginning of silver production in Laurion with the name of Erichthonius and placed the beginning of silver production (which most probably postdates the beginning of lead production) in very early times.

Buchholz (1972, 53 - 55) has given perhaps the most recent discussion of the evidence that the exploitation of the Laurion mines began in the Bronze Age; the same topic has been discussed by Hopper (1968). Ardaillon (1897) suggested that the Laurion mines were discovered by the Phoenicians and Burrows (1907) surmised that they may have been worked in Minoan times, but neither author adduced any hard evidence to support these conjectures.

Gowland (1901), pointing to Stais' discovery of three Mycenaean tombs at Thorikos, thought it inconceivable that a people of their culture should occupy a district full of lead ores and not use them.

Stubbings (1947) drew attention to the archaeological evidence that Attica and Aegina were brought early into the sphere of Mycenaean domination and conjectured that the reasons may well have been commercial and economic. Silver and lead were both to be found at Laurion, and Aegina lies on the direct land and sea route, via Epidauros, between Attica and the Argolid.

These are, however, but reasonable conjectures; perhaps the first piece of slightly more tangible evidene was the discovery by Broneer (Broneer 1933, 352: 1939, 415 - 416) of large amounts of lead in a Mycenaean context on the Athenian Acropolis (4.1 kg of pig lead amongst the ruins of a Mycenaean house on the NE slope just below the Acropolis wall; large amounts - about 20 kg - of lead sheathing from the region of the Mycenaean fountain). Broneer concluded that it was likely that the mines at Laurion were worked already in the Mycenaean period and that the Mycenaean settlement at Thorikos was probably a mining town controlled by the Lords of Athens. We do not, however, know that the lead found by Broneer came from Laurion; it is just a reasonable conjecture which the lead isotope method can prove or disprove.

Perhaps the best archaeological evidence for the date of the first exploitation of Laurion for lead and silver comes from the excavations at Thorikos. In a general way the decision to establish a Bronze Age settlement at Thorikos was most probably made because of its excellent harbour, its favourable position for defence and because of its possibilities for controlling the mining region of south Attica. The excavations under the direction of H. Mussche have shown that there are at least five Mycenaean chambered tombs on the Acropolis at the top of the Velatouri; the oldest tomb dates at least to the first half of the 15th century B.C. Trial excavations of a densely built living quarter have shown that the Acropolis was inhabited from about 2900 B.C. onward.

On the lower slopes of the Velatouri, excavations have revealed a large densely populated living area dating from the seventh to third centuries B.C. This area is concentrated around workshops for lead and silver working starting from the end of the fifth century B.C. At least five conventional ore washeries have been found, dating from the fifth century B.C., and the ancient mines from which the ore was taken have also been found and partly explored.

For our purpose, however, the most important finds have been made in excavations of the necropoles and on the Acropolis. Bingen (1964, 26 - 30), excavating in necropolis West 4, which had been in use from Protogeometric times to the 4th century, found the remains of a Protogeometric house in the midst of the necropolis.

The largest room in this house had a floor consisting of a layer of clay and ashes in which a grinding stone and a Protogeometric oenochoe were found; in the same floor two basins were found together with fragments of pottery of which one (TC 46.498, Bingen 30: Fig. 18) had a decoration with swastikas separated by an undulating line. One of the basins contained some blocks of litharge which analysis shows to contain less than 0.0001 % Ag (Laurion galena contains typically 0.13 to 0.30% Ag) and must be the remains of cupellation for silver; the evidence suggests at least some exploitation of the Laurion mines in the 9th century B.C.

Servais (1965, 20 - 24), excavating on the summit shoulder of Velatouri, found the ruins of a Middle Helladic private house, the main room containing a typically MH child burial. On the second floor (between levels -40 cm and -50 cm) were found four flat cakes of lead (which probably fell as molten metal on the beaten earth), two fragments of litharge and some scoria. With the lead and litharge were found many sherds of MH Grey Minyan and matt painted ware, the foot of a cup, some sherds of Middle Cycladic aspect and some obsidian fragments (see Servais 1965, 23 - 24), collectively suggesting a context between 2000 and 1600 B.C. In the level below Servais found a sherd (TC 65.69f) dateable "towards 1550 - 1520 B.C."; he concluded that this gave a terminus post quem for the lead and litharge samples, and that one cannot be much mistaken in dating the soil with litharge and lead towards the end of the 16th century B.C. This seems to be the best archaeological evidence we have so far for the beginning of the working of Laurion, and the earliest date for cupellation in Attica.

Turning to the evidence from lead isotopes, Figure 9 shows the data for 8 early lead artefacts, one bronze and one silver artefact for which the lead isotope composition falls within the Laurion field. As already mentioned the Laurion field of isotopic composition is so far unique, so that we may at present identify any silver or lead artefact with this composition as having derived ultimately from Laurion.

Taking these objects in order we have first the two Late Minoan objects from the Patso Cave and Lyttos, Crete. The ram figurine we cannot fix better in time than between 1580 and 1150 B.C.; the female figurine from Lyttos probably falls within 1400 - 1200 B.C. The lead wire from Chamber Tomb 530, following Wace's discussion (Wace 1932, 110), is probably early LH III, say about 1400 - 1300 B.C., though it may be as early as 1500 - 1450 B.C.

The lead bar, and the lead filling of the duck weight, both from El Amarna, may be dated between 1360 and 1373 B.C., whilst the lead net sinker from Abydos can be dated as early XVIII dynasty (about 1500 - 1450 B.C.). The lead pottery rivet sample from Ayia Irini, Keos, is to be dated about 1600 - 1450 B.C. The two lead door socket samples and the bronze sample from Akrotiri, Thera, are all pre-1450 B.C. The nine lead samples, taken together, give us confidence that Laurion was being worked between about 1200 and say 1600 B.C. The single XI Dynasty Middle Kindom silver sample, from Dendera, suggests that at least some silver was being produced in Laurion at some time in the period 2133 - 2000 B.C. The lead isotope data seems, even at this early stage of research, to give confidence that Laurion was certainly being worked at the beginning of the Late Bronze Age, and perhaps in the Middle Bronze Age. The Egyptian artefacts with Laurion composition also suggest some sort of contact between the Mycenaean and Egyptian cultures at this time.

VI.     THE SOURCES OF LEAD AND SILVER IN THE AEGEAN IN THE BRONZE AGE

One may hope to derive some information about sources of lead and silver, and the intensity of metallurgy, in the Bronze Age by studying the frequency of finds of lead and silver artefacts as a function of time and location. The chief sources for such a study are Renfrews's paper on Cycladic Metallurgy (1967), Buchholz's paper on lead in Mycenaean metallurgy (1972) and Branigan's monograph on Aegean metalwork (1974). In the analysis which follows it must be borne in mind that there may be some bias introduced due to inadequate coverage in the chronological or geographical distribution of the excavations, but archaeology is never free from this difficulty.

Buchholz (1972, 21 - 36, 55) catalogued finds of lead objects in the Aegean and found the distribution to be 40 in the E.B.A., 8 in the M.B.A. and 199 in the L.B.A. One may similarly use the catalogue in Branigan's monograph on Early and Middle Bronze Age metalwork (Branigan 1974, 155 - 205) to demonstrate that there are 60 lead objects from the E.B.A., 12 from the M.B.A. (plus 15 of doubtful age); significantly silver shows a similar trend with 100 objects from the E.B.A. and 22 from the M.B.A. (plus 31 doubtful). Though one would expect lead metallurgy to precede silver metallurgy, unless both were introduced together from an older metallurgical culture, the Aegean peoples seem to have been practising a combined silver/lead metallurgy from the E.B.A.

For gold an analysis of Branigan's catalogue shows 301 E.B.A. artefacts as against 71 M.B.A. artefacts (plus 76 of doubtful chronology). Geographically the 301 E.B.A. gold artefacts come dominantly from Troy (139) and Crete (135), with 13 from Lesbos, a few from the Cyclades and only 9 from mainland Greece. In the M.B.A. the emphasis shifts to Crete (61) then the undisputed centre of power in the Aegean, with 1 from Troy, 3 from the Cyclades and 6 from mainland Greece. Unpublished analyses of Bronze Age gold from Crete, Mycenae and Cyprus (Stos-Fertner) show that unrefined native gold was used, so that for gold we are talking about a highly skilled use of a natural product in contrast to the extractive metallurgy then practised for lead and silver.

Provided that it does not contain a hidden (archaeological) source of bias the character and chronological distribution of finds suggests that, within the Aegean, gold working was highly developed and that there was considerable activity in silver and lead metallurgy in the Early Bronze Age followed by a recession in the Middle Bronze Age and then a vigorous resurgence in the Late Bronze Age, starting about the middle of the second millenium. This finding accords with Renfrew's view (Renfrew 1967, 1 - 20) that the Early Bronze 2 period (circa 2700 to 2300 B.C.) was one of remarkable cultural and metallurgical intensity and that there appears to be no comparable intensity of activity again until the L.B.A. empires of the Minoans and the Mycenaeans. We should perhaps temper this statement with the evidence of continuity, even though at a reduced level, of gold working in Minoan Crete through the M.B .A.

Buchholz (1972, 23) demonstrated also that in the E.B.A. the finds of lead were distributed chiefly (32) in Troy, Lesbos, the Cyclades and east and central Crete, with relatively few (8) (and technically backward) finds on the Greek mainland.

Branigan's monograph (1974, 155 - 205) can be used to show that in the E.B.A. 8 of his catalogued lead objects are from the mainland and 52 from Troy, Lesbos, the Cyclades and Crete. For silver Branigan's catalogue lists 10 E.B.A. objects from the mainland and 90 from Troy, Lesbos, Limnos, W. Anatolia, the Cyclades and Crete. In the L.B.A. the resurgence in silver and lead metallurgy is accompanied by a change in the geographical distribution, which now shifts westward towards the Peloponnese, the Argolid, Attica and Euboea (Buchholz 1972, 21 - 24, Fig. 4).

The change in number of finds and in their geographical distribution as between the Early and Late Bronze Age suggests that sources of lead/silver for the Aegean may originally have come from the East (E.B.A.) but that new sources in the West (perhaps Cycladic, Attic or Euboean) may have begun to be exploited at the beginning of the L.B.A. (about 1600 BC); one of these sources may have been Laurion, if the arguments given for its early exploitation (section V) are accepted. On this view the M.B.A. may perhaps be a period when one set of lead/silver sources had been exhausted, or had become no longer accessible to the Aegean peoples, the resurgence in the L.B.A. having to wait for the exploitation of new sources.

Buchholz (1973, 282) demonstrates further that in Cyprus there are no E.B.A. lead finds, contrasting with 9 M.B.A. and 74 L.B.A. finds and, on the basis of the character of the L.B.A. finds, argues for a connection then of Cyprus with Anatolia.

Arguments on such bases can never be more than plausible, but such questions as the sources of Aegean lead in the E.B.A. as compared with the L.B.A., the date of the first exploitation of Laurion or Cycladic lead/silver sources, or the source of Cypriot lead in the L.B.A. are all questions to which lead isotope studies can be expected to make a significant contribution.

Lack of samples for analysis has so far severely limited progress, but the generosity of Dr. R. Moorey and Dr. M. Vickers (Ashmolean Museum, Oxford), Dr. C. Doumas and Dr. R. Tylecote has made it possible to make a start. The analyses are presented in Table 4 and Figures 5 and 10. They comprise analyses of three EC boat models from Naxos, one EC vee-shaped fragment from Amorgos, six LM artefacts from Crete, three artefacts from Akrotiri, Thera, four MC finger rings from Cyprus and one piece of lead from a pottery repair strap from Kea.

These are far too few objects on which to base any but the most tentative conclusions, and much more work is needed. Two LM Cretan objects, one (AE 15) a female figurine from Lyttos, the other (AE 1565) a figurine of a ram from the Patso Cave, have the Laurion composition and suggest metallurgical contacts between the Mycenaean mainland and Crete at some time between 1500 and 1100 BC. The vee-shaped EC lead fragment (AE 243) from Amorgos also falls within error of the Laurion field. Of the other Cretan samples the votive axe-adze (1927 - 1260) from the Psychro Cave falls within error of the Siphnian field; unfortunately there is no evidence for its chronology and Boardman (1961, 54, Fig. 26) suggests, on typological grounds, a date later than Minoan. The remaining Cretan artefacts (AE 80, AE 179 and AE 180) also have no assigned date, and two (AE 179 and AE 180) have often been suspected as forgeries; however they all plot in the same region of the lead isotope diagram.

The four "silver-lead" spiral rings from Cyprus fall in a region of the lead isotope diagram for which we have so far no comparative lead ores; we can at least say that the lead does not derive from Laurion, Siphnos, Thasos or the mainland opposite Thasos.

Cyprus does not appear to have any indigenous galena deposits; with Buchholz we may guess at a source in Asia Minor generally which the lead isotope method may eventually locate. The Middle Cypriot ring C 466 has virtually the same composition as the Cretan AE 80 ferrule from the Psychro Cave.

The three artefacts from Thera all have Laurion lead isotope compositions, quite unlike the composition of lead from galena exposed on Thera today. The lead object AE 243 from Amorgos has an isotopic composition quite different from the Naxos boats, and within error of a Laurion composition; however no lead ores from Amorgos have yet been analysed.

Perhaps the most interesting samples are the Naxos lead boat models, of EC II date. Now that all three boats have been analysed we see that their compositions all overlap with the Siphnian field. There is a high probability that these boats derive their lead from Siphnos, which leads to the interesting conclusion that lead working on Siphnos may have a very early beginning indeed in EB II (2700 - 2300 BC) times, although the peak of the activity of the mines was probably in the 6th century B.C.

VII.     THE ANCIENT MINES ON SIPHNOS

Herodotus mentions the gold and silver mines on Siphnos more than once, says that amongst the islanders the Siphnians were the richest (III, 57) and reports that they were at the height of their power at the time of Polykrates of Samos (538 - 522 B.C.).

Objective proofs of their wealth lie in the Parian marble decoration of their assembly hall and market place and in their maintenance at Delphi (from one tenth of the output of their mines) of one of the richest treasuries at the shrine of Apollo. When invaded by the Samians Siphnos was forced to give them 100 talents (equivalent to 2620 kg. of silver), a sum far beyond the resources of other Cycladic islands (Her. III, 58).

Pausanias records (X.8, XI. 2) that at some unspecified time (perhaps in the 5th. century B.C.) the Siphnian mines were flooded, because the Siphnians neglected the levy to Delphi; he records also the existence of gold mines on Siphnos. By the 1st century B.C. Strabo (10.5.1) records that the island had been reduced to abject poverty. The literary evidence, then, is of an island with considerable resources of at least silver (perhaps gold) but which were not by any means on the scale of Laurion and which were exhausted (or rendered inaccessible by flooding) sometime in the 5th century B.C.

The best topographic map of Siphnos (1:40,000) is by J. Birkett-Smith, London, 1976. Figure 11 is based on this map and shows the location of mines mentioned subsequently. The most recent survey of the geology of Siphnos is by Davis 1966, which includes also a geological map (1:50,000); earlier accounts were given by Foullon and Goldschmidt (1887) and by Ktenas (1907). Briefly, the island is composed chiefly of metamorphic rocks dominant amongst which are gneiss, schist, amphibolite and marble, with some glaucophane bearing rocks in the N.W. The major mineral wealth consists of oxide ores, chiefly of iron and manganese, which mostly occur as lenses in the marbles and along contacts between marbles and schists. The iron ore was extensively mined in the late 19th century, principally at the now abandoned mines of Ayios Sostis, Ayios Silvestros, Vorini, Tsingouras, Kapsallos and Xero Xylo (for references see Philippson 1959); all of these mines are marked on Birkett-Smith's map and in Fig. 11.

The first modem report of the possible location of one of the principal ancient mines on Siphnos is probably that of Tournefort (1727). Tournefort visited Siphnos in 1700 and was shown the entrance to the Ayios Sostis mine, but he did not enter the mine because of the difficulty and darkness of the place. Its situation, beside the sea, was sufficient to remind him of Pausanias' account of the flooding of the mines. Tournefort also mentions that smooth, greyish lead 'ore' is abundant on the island, is revealed by the rains, is a sort of natural cerussite, easily vitrifies and makes the cooking pots of the island very good; he states further that the 'ore' yields a lead approaching pewter (in colour?) which is used by the Siphnians to make small shot.

Personal observation (together with observations made by Fiedler in 1840) suggests strongly that the lead 'ore' mentioned by Tournefort is not a natural ore but small lumps of litharge left behind by ancient cupellation activities; the litharge still to be found today on the island at Ayios Sostis, Kapsallos and Platia Gialos has all the characteristics mentioned by Tournefort.

A more comprehensive description of features of Siphnos, including the ancient mines, was published by Fiedler (1840); our visits to Siphnos in 1975, 1976, 1977 have shown him to be very accurate in his observations. The islander sindicated to him the small peninsula, Ayios Sostis, in the north-east of the island, as the site of the ancient gold/silver mines. They were the only mine workings near the sea, and thus the only mines which could fit Pausanias' account of flooded mines. At this time, 1840, the major exploitation of the iron ore had not begun, so that Fiedler found the mines in much the same state as they had been when abandoned in antiquity. He described them as narrow tunnels and galleries chiefly in argillaceous iron ochre situated between red iron ore and marble. He found no gold or other minerals containing silver or gold. On the other hand he found scattered, hazel-nut sized pieces of slag and (rarely) litharge on the slopes of the peninsula, and some obsidian and fragments of pottery including tuyeres. Fiedler also described an occurrence of similar slags with litharge and obsidian near Platia Gialos, southeast Siphnos, but with no associated nearby traces of ancient mining.

Ross (1840) also described his visit to Siphnos in 1837, and again mentioned the Ayios Sostis peninsula as the ancient mining area of Herodotus and Pausanias; he also mentioned the occurrence on the fields (locality not specified) of pieces of litharge used by the Siphnian potters to glaze cooking pots.

The next well-known descriptions of ancient mines on Siphnos are by Bent (1885, 42), who purports also to describe the condition of the mines before the extensive 19th century mining for iron began. Bent describes visits to two mines, the first being that of Ayios Sostis. This he describes as having a very small entrance which "leads to a perfect labyrinth inside" and that "the precious metal must have been in veins, which these multitudinous passages followed up; along the sides there were quantities of niches, where the workmen evidently put their lamps.

The appearance of this mine inside is as if sparkling with silver, and the stones we broke off from the side had the weight and colour of lead, ..." Bent goes on to say that numerous pointed and cone-shaped axes have been found inside, that the walls of the mines show tool marks, and that on the peninsula close to the shore are hollows surrounded by scoria. Further, he describes seeing more hollows and scoria on the sea bottom just off the peninsula.

The second mine described by Bent was at a place called "the fissures" on the slopes of Mount Profitis Elias, undoubtedly the Kapsalos mine; Bent describes it as having the same appearance inside as the Ayios Sostis mine.

There is some question whether Bent's observations were made at first hand. Carolos I Gion in his history of Siphnos published in 1876 mentions every single observation made by Bent, and Bent's descriptions seem to come almost verbatim from Gion. Bent graduated from Oxford in 1875 and apparently visited the Cyclades in 1883. Gion published his book in 1876; the reader must be left to make his own judgement. For our purpose the descriptions remain as the only account of the mines immediately prior to their complete disturbance by later 19th century iron mining.

Subsequent to the start of the iron mining the mines were visited by Foullon and Goldscmidt in 1887 (1887) and by Hauttecoeur in 1898 (1899). Foullon and Goldschmidt do little more than repeat Fiedler's (1840) observations. Hauttecoeur mentions also the mine of Tsingoures (near the Kapsalos mine) on the slopes of Mt Profitis Elias; he says the mine was abandoned for about 2000 years when some Jews came in 1650 to work the lead (but see Tournefort (1727) who records that the opposition of the Siphnians caused the Jews to abandon their enterprise). Hauttecoeur goes on to say that silver is stated to be still visible in the walls of this mine today - i.e. in 1898.

Gowland (1901) mentions on Siphnos considerable amounts of debris both of mines and of smelting operations, together with pieces of ancient lead with portions of cupellation hearths attached to them. Gowland appears to rely heavily on Landerer (1876) for his information; Landerer, however, gives no convincing description of the occurrence or mineral species of silver-bearing lead ores on Siphnos and for gold mines his only evidence is the accounts of Herodotus and Pausanias.

Graindor (1903), writing in 1903, admits that up to that time no traces of gold had been found on Siphnos. In referring to silver mines he several times mentions argentiferous lead or argentiferous galena (e.g. on p. 467 he writes: "Au sortir de la mine, la galène était broyée, lavée, puis fondue dans de petits fours creusés en terre"...) without actually writing that he has personally seen galena in the mines or on Siphnos at all. The many comparisons which he makes with Laurion cause one to wonder whether his galena on Siphnos is an unfounded extension from galena in Laurion. On the other hand he gives a figure for the silver content of the "galerie" on Siphnos(~ 2 kg. Ag per tonne) which suggests that he had seen either an analysis of a Siphnian argentiferous mineral or the mineral itself - but the mineral species is not given here, unless "galerie" is a misprint for "galene".

That Graindor had personally observed the interior of at least some of the ancient mines on Siphnos seems probable from the accurate observations which he records about them. He mentions, for example, some of the varieties of tool marks on the gallery walls which can be seen to this day, the way in which the ore was cut out and the practice of sorting ore from gangue inside the mine, the gangue being left inside the mine. Though Graindor's account depends in part on statements to be found in Gion (1876) and in Bent (1885), his observations go beyond anything to be found in those authors.

Much later still Davies (1932) visited the mines after the abandonment of the 19th century iron mining. He identified Ayios Sostis as the site of the ancient gold/silver mines, described the existence there of iron ores containing veins of argentiferous lead, zinc and copper, and found on the peninsula a chip of obsidian, part of a coarse globular pot but no prehistoric sherds. At Kapsalos he saw similar grubbings for veins of galena and argentite, with steeply descending adits 3' high by 2' wide; he found here also a coarse splaying rim and a few pieces of weathered slag (containing 2.96% Pb, 3.86% Zn, 35.59% Fe, 0.05% Sb, traces of Ag but no Au). He found slag (containing about 0.2% Pb) also at Ayios Andreas and at Aspropyrghos.

Modern soil analyses in a geochemical investigation for lead were carried out in 1975 by the Vening-Meinesz Laboratory, Utrecht, under the direction of Professor V.M.L. Schuiling; the results gave strongly positive indications (1 - 5 % Pb) for lead at Ayios Sostis, Ayios Silvestros, Vorini and (0.1 - 1 % Pb) near the Tsingouras mine.

The Heidelberg-Oxford field trip to Siphnos in 1976 visited the Ayios Sostis, Xero Xylo, Kapsallos, Vorini and Vathi mines, and also the site at Platia Gialos (seaward slopes of the hill rising immediately behind the Xenia) at which Fiedler described the occurence of slag. At the Platia Gialos site abundant heavy lead-containing slag was found of hazelnut to walnut size, several lumps of litharge, abundant worked and waste obsidian and some sherds, but no trace of furnaces, mines or in situ ore. Lead isotope analyses of slag and litharge from Platia Gialos are given in Table 5.

The most obvious sign of mining activity on Siphnos is the 19th century exploitation of iron. There are six large iron mines (Xero Xylo, Kapsallos, Tsingouras, Vorini, Ayios Silvestros and Ayios Sostis (see Fig. 11) striking roughly NNE across Siphnos from Oras Profitis Elias to Ayios Sostis, and there are remains of old cable ways and mine railways. There is a smaller mine at Faros, and two much smaller modern iron mines in the hinterland of Vathi.

The intensive working for iron has obliterated much that was seen by Bent and Fiedler; further a large part of the Ayios Sostis mine collapsed in 1968 (taking most of the old church with it), leaving gaping holes at the surface and a dangerous situation under ground. Similarly, at Kapsallos the "entrance... only lately discovered, being hidden by the thickness of the brushwood all around" of Bent is now two large openings blasted out by dynamite; however the modern workings cut through low, narrow, ancient adits in the iron ore which worked some other ore no longer present except in small traces.

The Vorini mines similarly present the dominant impression of large open-cast exploitation for iron, but there are also some apparently old mine shafts here. The Xero Xylo mine (Erimoxylo of O. Davies) now shows chiefly evidence of recent iron ore mining, but there are some traces of ancient mining. The very small mines in the hinterland of Vathi are modern, as are probably those at Faros.

On this occasion we made a fairly superficial examination of the old workings below the old church of Ayios Sostis; we found these indeed to be of ancient appearance with occasional traces of pick marks, and to follow closely the veins of the ore of interest, ignoring the iron ore. Inspection failed to reveal the mineralogical features described by Bent and Gion, and no galena, cerussite or other lead, silver or gold ores were found on this occasion. The peninsula to the northeast of the new Ayios Sostis church had regions abundantly covered with heavy lead-bearing slag of nut size and rare pieces of litharge, and there were many coarse pottery sherds and some obsidian, including one well preserved arrow head. Lead isotope analyses of slag and litharge appear in Table 5.

Further along the peninsula towards the sea we found more traces of ancient mines, which follow veins closely and ignore the adjacent iron ore. At the point of the peninsula, adjacent to the sea, are numerous vent-like ore bodies. Close examination revealed no hollows or scoria on the sea bottom.

In 1977 a joint Heidelberg-Bochum-Demokritos Institute team carried out a thorough survey of the ancient mines on the Ayios Sostis peninsula. A preliminary account appears in Gentner et al. and a much fuller account in Wagner et al. Figure 12 gives a topographic map of the peninsula whilst Figure 13 shows the sites of the ancient galleries which were investigated and the surface occurence of scattered ores fragments, slag and litharge. The peninsula consists largely of marbles intersected by three types of ore bodies - large, vein-like and vent-shaped. The ore bodies are chiefly Fe-Mn oxides containing haematite, limonite and various manganese oxides. Secondary minerals include malachite, azurite, galena, cerussite, anglesite and other secondary lead minerals; the highly oxidised state of the ores suggests that part of the ore-deposit above the present sea-level is the gossan of a deeper sulphide deposit. Chemical analyses of some of the secondary lead ores by Müller and Gentner (1978) show up to 0.5% Ag and below 1 ppm Au; mineralogical, electron probe and ore microscopical studies have been reported by Klein (1978).

Detailed descriptions of the mines and their interior appearance, of their flooding at sea level, of the tool marks to be found in the gallery walls, of the proof of the use of fire-setting and of crushing and handsorting of the ores inside the mines and other matters are to be found in Wagner et al. Observations in 1976 and 1977 show that the ancient miners were not interested in the iron ore, but in more precious metals; the slag and litharge scattered over the peninsula and the silver content of the ores (up to 0.5%) make it very likely that in antiquity Ayios Sostis was a silver mining and refining site. On the other hand Ayios Sostis ores found so far do not contain gold in economic quantities; gold, if present, must occur elsewhere on Siphnos.

The isotopic compositions of lead extracted from slag, litharge and ores from Ayios Sostis are given in Table 5 and plotted on Figure 5. From Table 5 it will be seen that the slags and litharge from Platia Gialos (perhaps another Siphnian silver extraction and refining centre?) have the same isotopic compositions - either the ores came from Ayios Sostis or Siphnian ores in general have the same isotopic composition.

In September 1978 further exploration of the Ayios Sostis mines revealed very many tool marks in the ancient galleries and a number of primitive, handleless stone hammers were found in the mines, made chiefly of hard glaucophane-bearing metamorphic rocks probably from the Cherronesos region in the NW of the island. Proof of the hand-sorting of ore inside the mines is afforded by the many walls of waste material to be found built wholly or partly across the small, ancient galleries.

These walls are chiefly composed of material a few centimetres across and are often sintered together by calcareous material deposited out of water percolating through the mines, which has also led to widespread stalactite and stalagmite formation inside the galleries. The waste material walls occasionally contain accidental pieces of ore, pottery and charcoal which may eventually allow direct dating of the mining activity. On the slopes (see Fig. 13) where slag and litharge occur there were found also many fragments of broken tuyeres.

Also in 1978 a fairly extensive exploration was made of the Frase and Bolioni mines in the Kapsallos region, whilst a rapid survey of the two Tsingouras mines revealed that the modern iron working had blasted out most of the ancient workings.

Though both the Frase and the Bolioni mines show abundant evidence of modem exploitation using dynamite to blast out the iron ore, both also retain abundant remains of a labyrinth of ancient drifts and galleries (which traverse but ignore the iron ore) full of tool marks and waste walls together with lamp niches, stalactites and stalagmites and the occasional piece of pottery and charcoal. The tool marks reveal the use of at least three sorts of tool: chisels ending in a wedge shape, chisels with a foursided prismatic point, and a larger gad having a flattish wedge-shaped end which seems to have been used in softer rock; all of these were probably struck with stone hammers.

Some practically vertical shafts were found; they were probably for ventilation or removal of ore - one had rough steps cut into the side. As at Ayios Sostis the ancient miners had removed practically all traces of the ore which they had mined, but some samples of apparently secondary lead ores were taken from inside the mines from places in the walls where there were abundant tool marks.

The ridge above the Frase mine was littered chiefly with marble and iron ore but, especially where it was exposed to the wind, there were roughly circular depressions (furnace sites?) and some pieces of slag and about fifteen pieces of litharge were found. These finds suggest that there was a third centre of lead/silver refining here. Together with the slag and litharge were found about a dozen pieces of worked obsidian, about ten pieces of black glaze sherds and many rough unpainted /unglazed sherds, some very thick as though from a furnace.

The most significant analytical finding is that the EC II lead boat models found on Naxos have the Siphnian lead isotope composition. This suggests that lead production on Siphnos may be very ancient indeed, and that the Siphnian mines may have been exploited (perhaps to only a minor extent) as far back as 2700 - 2400 B.C. The association of slag and litharge with obsidian, both at Ayios Sostis and Platia Gialos, does not weaken this suggestion, nor does the existence of Early Cycladic silver bracelets, diadems and bowls from Amorgos, Antiparos, Naxos and Syros. Although we do not yet know certainly the source of this silver it probably came from within the Cycladic culture; so far there is evidence for ancient mines and silver refining only on Siphnos amongst the Cycladic islands.

In later times the evidence now established for ancient mining in five places on the island (Ayios Sostis, Ayios Silvestros, Xero Xylo, Kapsallos and Vorini) makes it clear that inundation of the Ayios Sostis mines cannot have been the only cause of the collapse of the Siphnian economy. It is much more probable that all the mines on Siphnos had been exhausted of silver ore, at least to a level and at a time which made it impossible for the Siphnians to compete any longer with the rapidly increasing exploitation of Laurion. It is possible that the inundation at Ayios Sostis was but the final blow to the Siphnians who, incapable of competing with Laurion and with practically no revenue left, neglected at last their tithe to Dephi. The priests of Delphi, disappointed of their title, may well have sought to turn to their advantage the natural flooding of the Siphnian mines in promoting a legend very favourable to the power and reputation of Apollo.

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

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