Fire, furnaces and metallurgy

Woodcut: a man seated in front of a stone furnace holds a tool in the fire. — Furnaces used to process ore: woodcut from Georg Agricola’s book De re metallica (1556), page 176. Georg Agricola (woodcut from De re metallica); digitisation Wellcome Collection, CC BY 4.0, source

This whole page comes down to a single question: how hot can you get?

At a hundred degrees, water boils. At nine hundred, clay becomes as hard as stone. At eleven hundred, copper flows. At fifteen hundred, iron melts. Each threshold opens up a world. Each threshold took centuries.

Fire, and a date that cannot be fixed

When did humans begin to control fire? No one can say precisely.

The difficulty is not finding traces of fire: they are everywhere. It is proving that this particular fire was lit or kept going by someone, and not started by lightning or a bushfire.

Two sites come up again and again in the discussion:

Wonderwerk Cave (South Africa): burned bone and plant ash about a million years old have been found there. But the cave lies in a region where bushfires are frequent, and fire may have entered it naturally.
Gesher Benot Ya’aqov (Israel): burned seeds, wood and flint about 790,000 years old have been unearthed there. Clusters of burned flint are found throughout the sequence of occupation, which suggests an ability to rekindle fire at will. But neither burned bone nor a built hearth has been found there.¹

The argument that settles it — an elegant one

How can a fire lit by someone be told apart from a natural blaze? The excavators of this site found an answer, whose reasoning can be followed step by step.

A bushfire sweeps across a piece of land. So it burns more or less everything, everywhere.

Yet that is not what is observed. By recording the position of each burned flint fragment, the researchers find that the burning is concentrated in patches, in specific spots, while the rest of the site bears no traces.²

Localised fire, surrounded by intact material. A wildfire does not do that. A hearth does.

This is the reasoning used throughout this website: instead of asking “is this compatible with what I believe?”, ask “if the other explanation were true, what would be different?” Here, the answer is clear: everything would be burned.

Another detail points the same way. Six kinds of burned wood have been identified — and at least three of them bear edible plants: olive, wild barley and wild grape.³

From about 400,000 years ago, traces of fire become much more numerous in several regions of the world. Before that date, the question remains open.⁴

Even when a hearth is identified, its exact use often eludes researchers: heating, cooking or keeping predators away? The ashes do not tell. This is not a gap in archaeology, but the difficulty inherent in searching for a gesture rather than an object.

Earth that becomes stone

It was long taught that pottery appeared with agriculture: populations supposedly settled down, grew cereals and made vessels to store them.

Discoveries in East Asia have disproved this account. In Xianrendong Cave, in southern China, pottery sherds 20,000 years old have been discovered. At that time, the Earth was going through the Last Glacial Maximum. Agriculture did not exist: the makers of these pots were nomadic hunter-gatherers.⁵

The presence of soot on these fragments attests to their use in cooking. Pottery therefore precedes agriculture by at least 10,000 years. The old model — first the fields, then the pots — rested on data from the Near East and Europe alone.

Getting hotter

An open fire can reach 800 to 900 °C, but cools as soon as the supply of fuel stops.⁶

The invention of the kiln changes these constraints: it encloses the heat, concentrates it and maintains it. Adding a bellows forces in oxygen, raising the temperature further.

Each thermal threshold crossed allows new materials to be transformed:

Between 650 and 1,000 °C: clay turns into solid ceramic.
Between 1,050 and 1,200 °C: vitrification produces stoneware and glazed ceramics.
From 1,300 °C: the vitrification of kaolin produces porcelain, dense, resonant and translucent.⁷

In China, this progression can be observed over the long term. From 1500 BCE, glazed ceramics fired at around 1,200 °C appear.⁸

The exact date of the appearance of “true” porcelain is debated. This disagreement is not about the archaeological pieces, but about the very definition of the material: should one take the firing temperature, the proportion of kaolin, the translucency or the resonance of the piece? Depending on the criteria chosen, the chronology varies by several centuries.

The secret of making porcelain was long closely guarded. It was not until 1712 that Father d’Entrecolles, a Jesuit based in Jingdezhen, sent a detailed description of it to France. Ten years later, a second letter was accompanied by samples.⁹

Getting metal out of rock

Some metals occur in nature in native form. They can be cold-hammered: copper, gold, platinum and meteoritic iron (recognisable by its high nickel content).

Around the Great Lakes, in North America, native copper has been hammered for at least 8,500 years. That is the age of the oldest securely dated object, a projectile point from Wisconsin; the tradition itself probably goes back about 9,600 years. It peaks between 7,000 and 5,000 years before the present, then declines.¹⁰

All this without ever melting the metal or making an alloy: it is gathered in native form, hammered and annealed. It is one of the oldest metalworking traditions in the world, perhaps the oldest. By contrast, in South America, metallurgy developed from the reduction of ores and the making of complex alloys.

Most metals are trapped in the form of ores. Extracting them requires reduction by heat in the presence of carbon (charcoal).

This invention was long dated to about 8,500 years ago, on the strength of a fragment of slag found at Çatalhöyük, in Anatolia. A reanalysis published in 2017 overturned this interpretation: the fragment is only half-fired, and it results from a house fire that accidentally melted green copper minerals used as pigment. It was not a smelting slag. The oldest established traces of copper extracted from its ore now go back to about 5000 BCE, in South-East Europe and in Iran — and in several places at roughly the same time, which weakens the idea of a single centre from which the technique supposedly spread.¹¹

The making of alloys followed:

Arsenical copper: since pure copper is soft, adding arsenic hardens it. Naturally present in some ores, this alloy was exploited very early.
Bronze: an alloy of about 90% copper and 10% tin. It flows better in the liquid state and is harder in the solid state.

A historical problem remains: Anatolia, rich in copper ore, lacks significant tin deposits. The exact origin of the tin used to make Anatolian bronzes 4,000 years ago remains an open question.

Iron requires much higher temperatures (melting point 1,538 °C). At first, metalworkers do not reach melting point: they obtain a spongy iron bloom that has to be shingled and hot-hammered to squeeze out the slag.¹²

In China, the development of high-temperature furnaces, blown by bellows, made it possible to melt iron and pour cast iron from the fifth century BCE, favouring the mass production of farming tools. In Europe, the blast furnace and cast iron only became widespread from the thirteenth century.¹³

This chronology calls for several clarifications:

China did not get ahead of the other centres across metallurgy as a whole: ironworking by direct reduction began earlier in Western Asia.
The production of liquid iron is also documented in several African technical traditions.
European blast furnaces resulted from local innovations and not from a transfer of technology from Asia.

A lost recipe: wootz

In southern India and in Sri Lanka, a technique of crucible steelmaking developed. Reduced iron is sealed in a closed earthenware crucible with organic matter (wood, leaves) supplying carbon, then brought to a high temperature and cooled slowly.¹⁴

The material obtained, called wootz, is a high-carbon steel. The slow cooling favours the separation of phases (iron carbides), drawing a characteristic wavy pattern on the surface of the polished metal.

These ingots were exported to the Middle East, notably to Damascus, where they were forged into renowned blades, known in Europe as “Damascus steel”.

In the eighteenth century, this production came to a halt. The exhaustion of deposits of ores containing traces of specific elements (such as vanadium) essential to the formation of the patterns, combined with the breakdown of craft networks, led to the loss of the original process. Modern metallurgical analyses make it possible to understand the structure of wootz, but reconstructing the historical process exactly remains difficult.

Stories claiming that these blades could cut a silk veil in mid-air belong to myth. The quality of wootz weapons varied and depended on skill at the forge.

Iron metallurgy in sub-Saharan Africa

In sub-Saharan Africa, iron metallurgy is attested from at least the first millennium BCE. The techniques rely mainly on the bloomery furnace, producing a solid iron that was then worked by forging.

In 2018, near the Falémé valley (Senegal), excavations uncovered an iron-smelting workshop used from the fourth century BCE to the fourth century CE. The site revealed that the bottoms of the furnaces were wedged with palm-nut shells.¹⁵

The origin of ferrous metallurgy in sub-Saharan Africa is the subject of two hypotheses: an independent invention or a spread from North Africa or the Nile valley. Uncertainties linked to radiocarbon calibration for certain periods call for caution, and archaeometric research is continuing.

Technical choices shaped by culture

To say of a civilisation that it had not mastered iron assumes a single pattern of technical evolution. The metalworkers of the Andes developed complex skills directed not towards the pursuit of hardness, but towards working with colours and surface finishes.

They created alloys of copper and gold (tumbaga). Through successive thermal and chemical treatments (depletion gilding), the copper was removed from the surface, revealing a layer of pure gold.

According to the archaeologist Heather Lechtman, this technique probably reflected not so much a wish to save precious metal as a conception in which the essence of the material had to emerge at its surface.¹⁶

The environmental and human cost

Reducing and working metals requires massive quantities of fuel.

Before mineral coal became widespread, metallurgy relied exclusively on charcoal. In Europe, the location of forges was frequently dictated by proximity to forests rather than to ore deposits, as transporting wood was more costly.

The impact of metallurgy on forests varies by region: while some areas suffered local deforestation, others were subject to rigorous forest management (coppice-with-standards) to guarantee a lasting supply of charcoal.

Mining also relied on forced labour. In the Andes, the mita system — a pre-Columbian rotating communal labour — was redirected by the Spanish colonial administration in the 1570s to supply the silver mines of Potosí. Sixteen provinces had to provide thousands of workers (the mitayos) every year. The working conditions in the tunnels and during the processing of ore by mercury amalgamation led to significant excess mortality.¹⁷

Summary of thermal thresholds

Temperature	Materials & Processes	Equipment required
600 – 700 °C	Simple firing of clay, open fire	Hearth
900 – 1,000 °C	Fired ceramic, dense terracotta	Simple kiln
1,100 °C	Melting of copper, reduction of simple ores	Furnace with air supply
1,200 – 1,300 °C	Stoneware, vitrification, porcelain	High-performance kiln
1,500 °C	Melting of iron, production of cast iron	Blast furnace / Powerful blowing apparatus

🔍 Archaeological methods and historiographical revisions

Metallographic analysis (examination of polished sections of metal under the optical or electron microscope) makes it possible to reconstruct the thermal and mechanical treatments an object has undergone: hammering, quenching, annealing.

Chronological advances also rest on the re-examination of known sites. The reassessment of the dating of the Xianrendong pottery by Xiaohong Wu’s team was achieved by taking samples of charcoal and bone from strata already identified, then analysing them by accelerator mass spectrometry (AMS).

Note 1

Here we show that micromorphological and Fourier transform infrared microspectroscopy (mFTIR) analyses of intact sediments at the site of Wonderwerk Cave, Northern Cape province, South Africa, provide unambiguous evidence--in the form of burned bone and ashed plant remains--that burning took place in the cave during the early Acheulean occupation, approximately 1.0 Ma.

Berna, Francesco et al., “Microstratigraphic Evidence of In Situ Fire in the Acheulean Strata of Wonderwerk Cave, Northern Cape Province, South Africa”, Proceedings of the National Academy of Sciences, vol. 109, no. 20, 2012, p. E1215-E1220

The presence of burned seeds, wood, and flint at the Acheulian site of Gesher Benot Ya'aqov in Israel is suggestive of the control of fire by humans nearly 790,000 years ago.

Goren-Inbar, Naama, Alperson, Nira, Kislev, Mordechai E. et al., “Evidence of Hominin Control of Fire at Gesher Benot Ya’aqov, Israel”, Science, vol. 304, no. 5671, 2004, p. 725-727