PART ONE: MATERIALS 62 Although many arsenical copper artefacts of the Early Bronze Age appear to have been roughly cast to shape in open stone moulds and then forged to size, this practice died out soon after the 8 per cent tin bronze alloy came into general use. Most artefacts were cast almost directly to the finished size in fairly advanced closed moulds similar to that shown diagrammatically in Figure 1.4. Many of these moulds were made from clay into which internal cavities were precisely moulded before firing. During the early years of the third millennium foundry processes in the Land of Sumer appear to have developed rapidly. Before 3000 BC most of the copper artefacts from Ur were simply fabricated from sheet metal. By 2700 BC, however, very refined cast copper articles were being produced, probably by a version of shell moulding, using a prefired mould of thin clay. Long thin sections, such as the blades of swords and daggers, were also cast into thin shell moulds, which appear to have been heated to redness before the molten metal was introduced. By 2500 BC the Egyptians had developed considerable expertise in the production of hollow copper and bronze statuary. Many large Egyptian statues were cast with an internal sand core, which is still present in some of the figures which have been found. Smaller components, such as the spouts of copper water vessels, were undoubtedly made by the cire perdue (‘lost wax’) technique, which had obviously been mastered by Egyptian craftsmen before 2200 BC. Grecian bronze Daggers with bronze blades inlaid longitudinally with niello, a black compound of sulphur with copper or other metals which formed a background for lively and naturalistic pictures in gold and silver, figured very prominently in the grave goods found by Schliemann in 1876 when he opened the Royal Bronze Age tombs of Mycenae. Such weapons, and most of the other metalwork which was found, are now believed to be of Cretan origin. The shaft graves date from the sixteenth century BC and are therefore of the Late Bronze Age, the absolute chronology of which is now well established. By the beginning of the second millennium BC, Crete was beginning to obtain tin from Etruria and southern Spain, and their designs for spear heads and daggers were widely copied throughout the ancient world. The characteristic metal artefact of Crete was the double axe, an object of great ritual significance; the compositions of some of which range from pure copper to ductile 4 per cent tin bronzes, and finally to hard, brittle bronzes containing 18 per cent tin. The Dorians, who invaded the Peloponnese from the north-east across the Gulf of Corinth in the thirteenth century, had primitive metal working skills of their own, and from the amalgamation of Dorian vigour and Minoan NON-FERROUS METALS 63 sophistication eventually emerged the metallurgical virtuosity of classical Greece. Bronze statuary soon began to demonstrate a mastery of foundry techniques which remained unsurpassed until the European Renaissance. Such castings represent what is probably the high peak of Bronze Age technology, before the later domination by Rome of the classical world. Between prehistoric times and 500 BC, bronze technology had developed in a slow and irregular manner which does not relate in any metallurgical sense with the Early, Middle and Late Bronze Age classifications of the archaeologist. The major metallurgical discontinuity occurred around 1600 BC, in late Minoan times, when large copper ingots weighing more than 30kg (661b) first appeared, and the usage of bronze throughout Europe and the Mediterranean region began to increase very rapidly. The arsenical coppers and low tin bronzes hitherto employed had been gradually displaced by bronzes containing 8–10 per cent tin, similar in their general characteristics to those first utilized a thousand years previously at Ur. The emergence of the full Bronze Age, in the middle of the second millennium BC, is characterized by the almost universal employment of this type of alloy. However, after about 1600 BC the majority of the bronzes produced contained lead, which, in most cast objects of any intricacy, was present in concentrations between 5 and 10 per cent by weight, and which was added to improve casting fluidity. Chinese bronze The chronology of early Chinese bronze is more difficult to interpret, since there appears to have been no specific date before which no lead additions were made. The belief that China had little or no prehistorical metallurgical experience probably originated because of the massiveness and mature artistic competence of the Shang bronzes, which, having been looted from the Imperial Graves at Yin, were the first to make an impact on the artistic sensibilities of the West. It was assumed that the Chinese had waited until the middle of the second millennium BC, the culmination of bronze technology in the West, before it began, with no significant earlier experience, to produce large bronze artefacts of the highest artistic quality. However, more than 300 metallic artefacts dating from the period before 1600 BC have been discovered, and bronze was being produced in northern Thailand in 2000 BC (see p. 59). China obviously began to produce bronze and other copper alloys at an early date. The presence in Kuangtung Province of tin deposits, both alluvial and in lode form, might well have facilitated the process. The art of bronze founding, which is now believed to have originated independently in China, appears to have first been practised in the adjacent provinces of Hansu, Shensi and Honan, where Chinese culture first began. PART ONE: MATERIALS 64 The Shang bronzes, produced at Anyang between 1400 and 1027 BC, are now world famous for their artistic merit and technical virtuosity (see Figure 1.5). These ritual objects, which were produced entirely by casting, were notable for the richness of their relief patterns and also for their mass and solidity, although some smaller wine vessels were also produced. One vessel weighed over 1.6 tonnes and had obviously been cast in a multipart mould made from prefired clay segments. Most of these Shang artefacts had lead contents less than 1 per cent by weight, and some contain negligible quantities of lead. It would appear lead was an accidental impurity. The tin content of the alloys varies considerably, however, in what appears to be a random manner from about 1.8 per cent to more than 20 per cent. The founders of Anyang could hardly have regarded tin as a rare or expensive commodity, to be used as economically as possible. However, the most impressive, highly decorated and least utilitarian items of Shang bronze Figure 1.5: Bronze ritual vessel ‘Fang Ding’ Shang dynasty, eleventh century BC. Decorated with a pattern of snakes and roundels. Courtesy of the British Museum. NON-FERROUS METALS 65 tend to be those with the highest tin content, and rarely contain more than half a per cent of lead. Items of utility, however, such as weapons and accoutrements, contain smaller quantities of tin and higher concentrations of lead. It seems that the Shang foundrymen were well acquainted with the effects of composition on the characteristics of the bronzes they produced. Thus when the objective was to produce large, intricate and decorative castings which would not be subjected to mechanical ill treatment, they preferred to use a very high tin content, rather than a smaller quantity of lead to obtain the casting fluidity required. When weapons were being made, however, brittleness was a characteristic to be avoided. The tin content was then very considerably reduced and a minimal quantity of lead added to improve the castability of the alloy. The melting point of a binary 20 per cent tin bronze is approximately 880°C: a similar quantity of lead only reduces the melting point of copper to 1000°C. The high tin, lead-free alloys would therefore be easier to melt and cast in large quantities and this must have been a factor which encouraged their use in a situation where tin was plentiful and intricate castings weighing perhaps a tonne were being produced. Bronzes containing 20 per cent tin are invariably associated with the manufacture of bells, and Theophilus in the twelfth century AD, following what was obviously an age-old tradition, recommended this composition for its purity of tone. The bell is generally regarded as being a Chinese invention, and the cast artefacts of the Shang and Chou dynasties are associated with the rituals of ancestor worship where music had a magical significance. It seems possible, therefore, that the ritual vessels were expected to be sonorous as well as artistically impressive. The distinctive green patina of ancient Chinese bronzes has occurred by a process of slow corrosion during the long internment of the bronzes in deep loess of Central China. This yellow, porous, windblown, sandy clay contains a good deal of calcareous material, which, when taken into partial solution by circulating groundwaters, is believed to have played an essential role in the patination process. Greek metallurgical specifications At a very early period the Greeks attempted to standardize the composition and mechanical proportions of bronze. An inscribed stela found at Eleusis in 1893 provides what is probably the earliest known specification for bronze and cites a decree concerning the manufacture of empolia and poloi, the bronze fittings which were used for joining the drums of stone columns. The structure involved was the Philonian Stoa, a portico which was erected around 360 BC in front of the much older Telestrion at Eleusis. It is specified that the contractor PART ONE: MATERIALS 66 will use for the blocks copper from Marion, the alloy being made of twelve parts, eleven of copper to one of tin. He will deliver the blocks clean, rigid and four- square and will round off the dowels on a lathe as in the example provided; he will fix them into the blocks straight, and perfectly rounded, so that they can be rotated without deviation. Bids for the contract are to be made at so much per mina of bronze, and the contractor will weigh out the bronze while there is constantly present one of the building commission, either the public recorder or the site supervisor. It seems that the Greeks were aware of the weakening effect of lead on bronze and preferred the binary tin-copper alloy for structural items. The requirement in the specification that the poloi should be turned on a lathe is also a matter of unique importance: this Eleusinian inscription apparently provides the earliest evidence for the use of a metal-turning lathe in the Greek mainland. The requirement of the specification that an 8.3 per cent tin bronze should be used for these castings shows that the fourth-century Greeks were well aware of the effect of tin on the properties of the alloys they employed. From the weapons which were found in the Mycaenean tombs it is evident that metal workers, even in the sixteenth century BC, knew exactly how much tin could be put into bronze before the alloy became too brittle for hot or cold forging. They improved the hardness and elastic limit of their bronze sword blades by judicious cold working, and knew that ductility could be restored, when required, by careful annealing. Pure copper rather than bronze was the preferred material where high ductility was required for the manufacture of rivets and other fixing components. METALLURGY IN THE ROMAN WORLD Although few metallurgical ideas or processes appear to have been truly Roman in origin, the improved standards of living associated with the rapid dissemination of the Roman way of life stimulated metallurgical demand, and encouraged the rapid diffusion of improved processes and techniques throughout the Empire. The primary metallurgical requirement of the army was iron (see Chapter 2). Bronze, in quite formidable quantities, was needed, however, both for military and non-warlike purposes, and coinage utilized vast quantities of gold, silver and copper. Roman civilization also brought with it the rapid expansion of highly organized urban life which required large quantities of lead for plumbing. The main metallurgical innovation of Roman times, however, was the general introduction of brass which was first used for coinage when tin became expensive. NON-FERROUS METALS 67 Roman copper and bronze Well before Imperial times the extraordinary mineral wealth of the remoter regions such as Spain, Germany and Britain became evident. The mines of Spain provided gold, silver, copper, iron and zinc in great abundance, the richer deposits being situated in a region known as Tartessus, twenty-five miles north of Huelva on the Gulf of Cadiz. At Rio Tinto (see p. 55) copper was extensively extracted in the Bronze Age, and the Phoenicians appear to have worked the deposits for silver between 900 and 700 BC. The Carthaginian empire was also dependent upon silver from Rio Tinto, and the Romans worked the Tartessian mines continuously from about 200 BC until AD 400. In Republican times the mines were worked on a limited scale, largely for silver. Under Augustus, however, the scale of operations increased very rapidly and deep tunnels were driven into the main chalcopyrite body to exploit secondary veins of rich copper ore. The mines appear to have reached their maximum output between the second and third centuries AD. The Roman troops left Spain in AD 410, and although the Visigoths took over the Peninsula in 475 they did nothing to revive the mining industry, which was also largely ignored by the Moors. The ancient Roman workings at Rio Tinto were not reopened until the middle of the sixteenth century, during the reign of Philip II. Rome did not begin to exploit the copper mines of Timna in the Negev (see p. 52) until the third century AD, although working continued until the middle of the seventh century when the Arabs invaded. Mining activity at Timna appears to have been controlled largely by the availability of fuel, charcoal obtained from the desert acacia. The horses of San Marco Most of the statuary of the Graeco-Roman world was cast in bronze. In fourth- century Greece leaded tin bronzes were generally used although zinc found its way into many of the later Roman bronzes, which by the middle of the second century AD tended to contain comparable quantities of zinc and tin. The characteristics of these alloys, which resembled those of the later gunmetals were, after all, well known, and utilized by all practical foundrymen. Thus, their low melting points and wide freezing ranges facilitated mould filling and helped to compensate for any incidental deficiencies in feeding and venting arrangements which must have been difficult to avoid when large, thin-walled castings of a complex nature were being produced in a primitive foundry. Zinc would also have deoxidized the alloys and assisted greatly in the reduction of casting porosity. However, some artists of this period cast their statues from what we would now regard as most unsuitable material. The four horses of San Marco in PART ONE: MATERIALS 68 Venice, when accurately analysed in 1975, were found to have been cast from virtually pure copper, containing only about 1 per cent of t1n and lead. Because of its high melting point and narrow freezing range, this alloy would have been difficult to melt and cast properly. The San Marco horses were cast by the indirect lost wax process, the technique used some 1500 years later by Benvenuto Cellini when he cast his celebrated Perseus. In this approach, the wax layer applied to the surface of the core is equal in thickness to the wall thickness of the casting to be produced. The wall thickness of the San Marco horses varies between 7.5 and 10.5mm, and originally they were gilded. From the duplex structure of the surviving areas of the gilded layer, it can be inferred that before the application of the gold leaf, the copper surface had been treated with an aqueous solution of gold mercury amalgam. When this had reacted chemically on the surface of the copper to produce the amalgamated layer required, the work would have been heated gently and the gold leaf applied. Theophilus, writing in the twelfth century AD, warned that this technique could not be used to gild a lead-containing substrate, and the information would probably have been available in Roman times. The artist who produced the horses of San Marco might have decided to produce his castings from copper because it was easier to gild than a bronze which might well have contained substantial quantities of lead. They were popularly attributed to Lysippus of Sicyon, the official sculptor of Alexander the Great, who worked in the fourth century BC. It is possible, however, that they were cast eight hundred years later, during the reign of Constantine, although a more probable date would be some time during the second century AD. Lead and silver Lead is comparable to copper in its antiquity and was probably obtained in its metallic form by reduction from its ore at an earlier date. The earliest known metallic artefacts, small beads and pins of copper dating from the period between 9000 and 7000 BC, were hammered from crystals of native, unmelted copper (see p. 48). However, at 6500 BC horizons in the ancient Anatolian mound of Catal Hüyük beads of fused lead have been found in close association with hammered beads of native copper. It appears that lead must have been extracted from its ores at an earlier date than copper, especially in view of the low melting point of lead and the ease with which its oxide can be reduced by carbon. Native lead is rarely found. Its sulphide, however, galena, upon which most lead ores are based, is fairly abundant and would have been instantly recognizable by primitive man because of its lustre, high density, and by the structural perfection of its dark grey cubic crystals. When exposed at an outcrop, galena weathers and oxidizes. Anglesite, the sulphate of lead, is a mineral formed during NON-FERROUS METALS 69 the intermediate stages of this reaction process which generally concludes with the formation of cerussite, a lead carbonate. Galena usually contains between 0.03 and 0.1 per cent by weight of silver, and the economics of lead extraction are generally determined by the value of the silver present in the ore. Metallic lead has been found in Iraq, in a 6000 BC context at Yarim Tepe, and at the fifth-millennium site of Arpachiyeh. Lead artefacts have also been obtained from the fourth-millennium sites of Anan and Hissar III in Iran and Naqada in Egypt. The earliest evidence of lead being used for aesthetic rather than practical purposes is provided by an Egyptian figurine 5cm high, dating from about 4000 BC, which is now in the British Museum. It is unusual in having been carved from a solid block rather than being cast to shape; the lead is of such high purity and contains so little silver that it could apparently have been obtained only by reducing the oxide residues from a silver cupellation process. Few leaden artefacts of comparable age have been found, although a mid- fourth millennium cemetery at Byblos in Lebanon recently yielded more than 200 silver artefacts. Silver objects of the late fourth millennium have also been unearthed in Palestine, in Ur and Warka in Mesopotamia, and also from various sites in Asia Minor. It seems probable that they emerged as a by- product of lead refining operations. Lead artefacts of the third and second millennia BC include spindle whorls, weights for fishing, and wire which appears to have been used for repairing pottery. Lead weights, based on a unit module of 61g (2.202), have also been found throughout the Aegean in Middle and Late Bronze Age contexts. Shapeless lumps of lead were found in that layer of the Hissarlik ruins corresponding to the period between 3000 and 2500 BC, and many fine silver artefacts dating from about 2700 BC were found by Sir Leonard Woolley in the Royal Graves at Ur. Numerous deposits of silver-rich lead are known in the Aegean area, several of which were described by classical authors such as Aeschylus, Herodotus and Strabo. In the fifth century BC, silver from the mines at Laurion provided the wealth needed to support the Athenian Empire. Other famous silver mines were those in the Pangaean region of Macedonia, and the Cycladic island of Siphnos. Mining at Siphnos began some time between 3150 and 2790 BC. The silver recovered from the Early Cycladic tombs appears never to have been melted down for recycling, and this has made it possible to utilize isotope abundance analysis to identify the lead mines from which the silver was obtained. Specific identifications can now be made by measuring the relative abundance of the natural isotopes in each element present in the metallic artefact. This approach allows the origin of the major alloying constituents to be identified, not from the impurities they contain, but from their own specific internal atomic characteristics. The metals extracted from every ore deposit each have an isotope distribution spectrum which is specific to, and characteristic of that deposit. Isotope abundance ratios are not influenced in PART ONE: MATERIALS 70 any way by smelting or refining processes, or by the presence or absence of impurities introduced during alloying. The silver artefacts in the Cycladian tombs, therefore, which appear to have been made from virgin metal and almost immediately interred, have isotope abundance ratios which make it possible to identify very precisely the lead mines from which the silver was obtained. Of sixteen artefacts, six were found to have come from the Laurion mine near Athens and eight from the Siphnos field. A number of lead and silver artefacts from the shaft graves at Mycenae (c.1550 BC) were made from metal coming from the Laurion mine. Egyptian silver artefacts of the 10th and 11th Dynasties (2175–1991 BC) also appear to have been made from Laurion metal. Isotope ratio measurements on Cretan lead and silver artefacts of the period 1700–1450 BC indicate that about 80 per cent of the metal came from Laurion and only 10 per cent from Siphnos. Many lead and silver artefacts were preserved under the volcanic ashes which submerged Akrotiri, on the island of Thera, in 1500 BC. About 96 per cent of this metal came from Laurion and only 4 per cent from Siphnos. Rio Tinto At this period, Rio Tinto was being worked only for copper; silver was first exploited in this area between the twelfth and eleventh centuries BC. The chalcopyrite ore which appears to have been worked for silver around 1200 BC at Rio Tinto was taken from a thin, concentrated cementation layer, only a few centimetres thick, between the oxidized gossan and the principal mass of copper ore (see p. 55). It could have contained around 40g of silver per tonne. Up to this time the bulk of Mediterranean silver had been derived from lead by cupellation. Since Rio Tinto copper contained little lead, it is evident that lead oxide, or a lead ore must have been added to the smelting charge to take up the silver. The chalcopyrite ore might have been fused directly in a small shaft furnace to form a copper sulphide matte, from which the silver was absorbed by molten lead, and the alloy thus obtained would then have been cupelled to extract the silver. The silver deposits at Rio Tinto were controlled successively by the Phoenicians, Carthaginians and, from about 200 BC the Romans. By this time the wealth of Tartessus, the biblical land of Tarshish, had become legendary throughout the Graeco-Roman world. Silver production at Laurion Silver was extracted at Laurion from a geological formation of three layers of whitish marble and limestones separated by two layers of micaceous schist. NON-FERROUS METALS 71 Contact deposits of silver-bearing lead concentrated below each layer of schist. The galena-based ore from the lower contact deposit was particularly rich and is reputed to have yielded between 1200 and 1400g of silver per tonne in the time of Themistocles. This was not exploited until the beginning of the fifth century BC. Earlier mining was confined to the upper contact layer, which was closer to the surface, and yielded ore based largely upon the mineral cerussite, a carbonate of lead. The lower, richer contact layer in the mines was discovered in 484 BC and 100 talents of the silver it produced were used by Themistocles to build the fleet which destroyed the Persians at Salamis in 480 BC. Subsequent references to the mines in plays by Aristophanes and Pherecrates suggest that they were then operated under state control. During the fourth century the mines and mining rights at Laurion still belonged to the Athenian state, who auctioned off operating franchises to private contractors and speculators. The industry declined very rapidly towards the end of the fourth century, when the silver coinage standard of the Athenian state was superseded by the gold standard introduced by the Macedonian kings. The last Athenian silver tetra-drachmas were minted in the second half of the first century BC, by which time the mines were largely exhausted, and once prosperous mining towns such as Thorikos had been virtually abandoned. Attempts made in the mid-nineteenth century AD to revive mining at Laurion were unsuccessful. The ore extracted at Laurion in the fourth and fifth centuries BC was crushed with iron mallets and then ground to a sandy consistency in rotating stone hourglass mills or in hopper querns. The crushed ore emerged as a fine powder from which the lighter, earthy material was removed by water washing in some very ingenious helicoidal washers. A continuous current of water, during its passage around the spiral, carried the suspended particles of ore from cup to cup and classified them. The rich ore was the first to be deposited, then the poorer ore and finally the sand and silts. When the water reached the lower tank all suspended solids had been deposited and it could be returned to the stand tank, so completing the circulation process. The argentiferous lead ore, concentrated in this way, was then taken from Laurion to smelting complexes at Thorikos, Puntazeza and Megola Pevka, where vertical shaft furnaces, using locally produced charcoal, were used to reduce the ore to metal. The silver content of the lead alloy obtained from this smelting process was subsequently recovered by cupellation, and taken to the Athenian mint, at the south-east corner of the ancient Agora, where the famous silver drachmas, the ‘owls of Athene’ would have been struck. Some of the lead produced was used to a limited extent for sealing and for holding in place the iron clamps used to prevent relative movement between large marble building blocks. From the extensive deposits of litharge in the Laurion, it must be concluded that the value of lead at this time did not always justify the cost of recovery from cupellation waste. . towards the end of the fourth century, when the silver coinage standard of the Athenian state was superseded by the gold standard introduced by the Macedonian kings. The last Athenian silver tetra-drachmas. suspended particles of ore from cup to cup and classified them. The rich ore was the first to be deposited, then the poorer ore and finally the sand and silts. When the water reached the lower tank. silver from the mines at Laurion provided the wealth needed to support the Athenian Empire. Other famous silver mines were those in the Pangaean region of Macedonia, and the Cycladic island of Siphnos.