Engineering Before Metals: The Prehistoric Technologies We Keep Underestimating

A newly identified tradition of heat-treating chert in northern Australia, at least 40,000–45,000 years old and probably older, belongs to a much deeper history of prehistoric materials engineering. Heated stone, manufactured adhesives, composite weapons, cordage and prepared pigments suggest that technological sophistication is poorly measured by the presence of metal, machines or permanent architecture.

At Nauwalabila I, a sandstone rockshelter in Arnhem Land, some stone flakes preserve an unusually informative sequence of scars. One surface is comparatively rough; another, cut across it later, is smoother. The difference is small enough to require careful measurement, but technologically it is consequential. The stone had been worked, heated, and then worked again.

Patrick Schmidt and Peter Hiscock examined 502 chert artefacts from the site's long archaeological sequence and classified 70, about 14 per cent, as diagnostic of heat treatment because they carried this contrast between pre-heating and post-heating fracture surfaces. On 15 particularly clear specimens, replica-tape measurements confirmed the visual assessment: the post-heating surfaces were, on average, substantially smoother than those produced before heating. Two of the diagnostic pieces came from deep in the deposit, more than 2.4 metres below the present surface.

Published in 2026, the study describes these finds as the oldest known systematic heat treatment of chert. That wording matters. They are not the earliest evidence that humans deliberately altered stone with fire. Heat treatment of silcrete in southern Africa has a much deeper history, with evidence reported from Middle Stone Age contexts extending well beyond 100,000 years ago and potentially to about 164,000 years ago at Pinnacle Point. What Nauwalabila adds is something different: early mastery of a stone that is considerably less tolerant of thermal mistakes.

The chronology also needs care. Optically stimulated luminescence dates sediment grains rather than the stone artefacts themselves, and Nauwalabila has been the subject of a long argument over movement within its deposits, including possible disturbance by termites. If artefacts and dated sediments remain in their original association, the deepest heat-treated pieces may be older than 50,000 years, perhaps around 60,000–53,000 years. Schmidt and Hiscock deliberately adopt a more conservative position: even under models allowing some vertical displacement, the evidence should be older than about 45,000–40,000 years. The claim is therefore impressive without requiring acceptance of the site's maximum age.

Making a better stone

To call heat treatment simply “heating a rock” obscures the technological problem.

Chert is a fine-grained silica-rich material, but not all chert behaves identically and not all heating improves it. Too much heat, too rapid a temperature rise, or inappropriate positioning in relation to a fire can fracture or destroy a useful piece. Chert also retains considerably more molecular and chemically bound water than many silcretes and has a tighter pore structure. During heating, that makes the escape of water more difficult and increases the danger of thermal failure. Later prehistoric chert-working traditions sometimes used carefully arranged, insulated heating structures precisely because gradual heating and cooling were necessary.

No Pleistocene “oven” has been identified at Nauwalabila, so it would be wrong to reconstruct a particular heating installation from the artefacts alone. Nor can archaeologists read an exact target temperature from each flake. What the artefacts demonstrate is process.

Somebody selected chert, worked it, exposed it to heat under conditions that altered its fracture properties without ruining it, and then returned to knap it again. Since many of the cherts brought to Nauwalabila probably travelled tens of kilometres, heat treatment may also have helped people obtain more useful cutting edge or more predictable flakes from a relatively costly imported raw material.

This is where “high technology” becomes a useful term, provided it is not confused with modern machinery. Technological sophistication can reside in control of transformation rather than in the apparent complexity of the finished object. A flake of stone may look simpler than a bronze sword. The operational knowledge needed to manufacture it need not be simple at all.

The same point was made much earlier in southern Africa. At Pinnacle Point, analysis of silcrete showed that Middle Stone Age people were deliberately using fire to improve stone for knapping. By around 72,000 years ago heat treatment was widespread in the assemblages studied by Kyle Brown and colleagues, with possible examples extending back to approximately 164,000 years ago. Heating was therefore not merely a domestic use of fire. Fire had become an engineering tool: a means of deliberately changing the properties of another material.

Nauwalabila makes this history more interesting because chert is less forgiving than silcrete. Schmidt and Hiscock consequently raise two possibilities. Chert heat treatment could have been independently developed after people reached Sahul. Alternatively, knowledge of thermal transformation may already have belonged to populations dispersing eastward through southern Asia, adapting established principles to different raw materials as they moved.

At present, neither model can be demonstrated. The obvious test lies between Africa and Australia, in the chert-rich archaeological sequences of South and Southeast Asia. Those sequences have rarely been examined systematically for the microscopic signatures of intentional heat treatment. The absence is therefore, for now, as much a research gap as an archaeological pattern.

Technology is often a sequence, not an object

Archaeologists traditionally classify technologies through surviving artefacts: handaxes, points, scrapers, blades. Yet many of the more demanding prehistoric technologies were not single-object technologies at all. They were sequences involving materials that were collected at different places, transformed separately and brought together only at the end.

Hafting is an obvious example.

A stone point attached to a wooden shaft is no longer simply a stone tool. It is a composite device whose performance depends on the interaction of the point, shaft, binding and often an adhesive. The dimensions of one component constrain another. A failed junction may destroy the usefulness of an otherwise well-made point.

Possible stone-tipped spears from Kathu Pan 1 in South Africa, dated to roughly half a million years ago, have been interpreted from impact damage, basal modification and experimental comparison as evidence of very early hafting, although that interpretation has subsequently been questioned in the specialist literature. Whatever the precise status of the Kathu Pan evidence, hafted technologies are securely established later and eventually become an important part of both Neanderthal and Homo sapiens technological repertoires.

What tends to disappear is the joining technology.

At southern African Middle Stone Age sites, microscopic residues and experimental studies have shown that adhesives could themselves be manufactured substances. Work at Sibudu and other sites has identified plant-derived tars and compound adhesives used in hafting. Recent chemical and experimental research is particularly revealing because Podocarpus, one of the plants used, does not conveniently exude a ready-made sticky resin. Adhesive must instead be produced from plant material through heat-driven transformation. Experimental work indicates that tar can be generated from leaves, including by allowing smoke-derived products to condense on cooler stone surfaces and then scraping off the resulting substance.

At Sibudu, analyses published in 2024 found evidence that Middle Stone Age inhabitants were not restricted to a single recipe. Tars from different plants were produced, and some were subsequently incorporated into ochre-containing compound adhesives. That implies technological choices based at least partly on desired material behaviour rather than merely the opportunistic use of whatever happened to be sticky.

There is a useful warning here. Archaeologists have sometimes moved too quickly from a complicated manufacturing recipe to claims about “complex cognition”. Experiments with birch bark tar, used by Neanderthals, have shown that apparently sophisticated substances can sometimes be discovered through surprisingly simple procedures. The presence of tar alone therefore does not tell us how difficult its manufacture was.

But the reverse is equally important. Chemical work on the approximately 200,000-year-old birch tar from Königsaue in Germany indicates that the material was probably produced under oxygen-restricted conditions rather than by the simplest experimentally demonstrated method. The argument has shifted accordingly. The interesting question is no longer simply whether Neanderthals could make glue. It is which production procedures they knew, how reliably they reproduced them, and how such knowledge was transmitted.

That is a much better archaeological definition of technological sophistication.

The missing technologies were made of plants

Stone dominates the Palaeolithic record partly because stone survives.

This creates a serious distortion. A toolkit containing stone flakes, wooden shafts, woven bags, fibre bindings, nets, baskets, bark containers and adhesives might leave archaeologists thousands of stone artefacts and almost none of everything else. What looks like a “stone technology” may actually have been a technology in which stone formed only the durable cutting component.

A tiny find from Abri du Maras in southeastern France makes the problem unusually clear.

On the underside of a Levallois flake from a Neanderthal occupation level dated approximately 52,000–41,000 years ago, researchers identified a fragment of cord only about 6.2 millimetres long and half a millimetre wide. Microscopy revealed three bundles of inner-bark fibres, each twisted in one direction and then plied together in the opposite direction to produce a stable three-ply cord. Raman spectroscopy supported the identification of the material as plant fibre.

Nothing shows that this particular fragment was rope from a haft. It could have belonged to a bag, net or some other object, or it may simply have entered the deposit with the flake. That uncertainty should remain.

What is not uncertain is the technological principle. Cordage converts many weak fibres into a stronger continuous material. Counter-twisting prevents the finished product from unravelling. Before twisting can even begin, suitable plant material has to be recognised, collected and processed.

The discovery therefore hints at an entire technological domain that is almost archaeologically invisible. Once cord exists, so do the prerequisites for nets, bindings, handles, traps, containers, mats and more elaborate ropes. This does not prove that every such object existed at Abri du Maras. It demonstrates why reconstructing Palaeolithic technology solely from stone produces an artificially impoverished picture.

Recipes, containers and prepared materials

A similar problem occurs with pigments.

Ochre is common at prehistoric sites, but a lump of red mineral tells us little by itself about what people did with it. Ochre can be used for colouring, adhesives, hide processing and other practical purposes. Its presence should not automatically be translated into “art” or symbolism.

At Blombos Cave on the southern Cape coast of South Africa, however, the archaeological context is much richer. In deposits dating to about 100,000 years ago, excavators found two Haliotis midae abalone shells associated with ochre, bone, charcoal, grindstones and hammerstones. Analysis showed that an ochre-rich liquid mixture had been prepared and stored in the shells. Its final use remains uncertain.

The unresolved function is almost beside the technological point.

Here are raw materials collected from different sources, size reduction by grinding, deliberate mixing and a container in which the finished preparation could be kept. The archaeological evidence catches not merely possession of pigment but a production operation.

The same distinction applies to heat-treated chert and manufactured tar. What distinguishes these technologies is not simply that prehistoric people knew useful materials when they saw them. They altered materials into states that do not ordinarily occur in usable form around them.

Stone was made easier to fracture. Leaves became adhesive. Fibres became cord. Mineral powders became prepared mixtures. Separate components became composite tools.

In each case, the finished product depends on steps whose results appear only later. A chert blank can be destroyed by poor heating. An adhesive must be prepared before the point is secured. Fibre has to be processed before the cord can be twisted. Ochre has to be ground before it can be incorporated into a mixture.

This is planning in a very concrete archaeological sense. It need not imply that prehistoric craftspeople understood mineralogy, polymer chemistry or thermodynamics as abstract sciences. Technical knowledge does not require formal theory. A craftsperson can understand that a material must be heated slowly, that one plant makes better tar than another, or that alternating twists produce stable cord without being able to describe the molecular mechanism involved.

What “high-tech” should mean in prehistory

There is a temptation to turn these finds into another story about supposedly primitive people being “more advanced than we thought”. That formulation retains the problem it claims to correct.

The Stone Age was never technologically uniform. Nor does innovation form a single ladder leading from stone to bronze, iron, steam engines and electronics. Technologies appear, disappear, spread, remain local, change function and sometimes have to be reinvented. A society can possess extremely refined knowledge in one material domain while having no need for another technology that later archaeologists regard as diagnostic of progress.

Nor should every multi-stage process be used as an intelligence test. The current debate over Palaeolithic adhesives has been valuable precisely because experiments demonstrate how difficult it is to infer cognition directly from the number of steps in a manufacturing sequence. Some apparently complicated procedures can emerge from repeated practical engagement with materials; some deceptively simple artefacts require extensive learned expertise.

What archaeology can identify more securely is technological organisation.

Nauwalabila shows stone worked on both sides of a controlled material transformation. Southern African sites preserve traditions in which fire altered stone and plants into materials with different mechanical properties. Hafting joined materials with very different characteristics into functioning devices. Abri du Maras preserves a few millimetres of a fibre industry that should otherwise have vanished. Blombos preserves the preparation and storage of a manufactured mixture.

Taken together, these cases make the conventional image of “Stone Age technology” increasingly unhelpful. Stone was only one element in technological systems incorporating heat, wood, plant fibres, adhesives, mineral additives and accumulated procedural knowledge.

The strongest claim that can presently be made is therefore not that Palaeolithic people possessed some unexpected modernity. They did not need to be modern engineers in disguise. Their technologies were adapted to different economies, materials and scales of production.

Rather, the evidence shows that controlled transformation of materials has a deep history. By well over 100,000 years ago, people in Africa were manipulating the physical properties of stone with fire. By the later Pleistocene, manufactured adhesives, composite tools and fibre technologies occur in several populations and regions. The Nauwalabila discovery now places deliberate chert heat treatment near the beginning of the known human occupation of Sahul, though whether the technique was invented there or carried eastward with migrating populations remains unresolved.

That unresolved question may be the most productive one. If early chert heat treatment is eventually identified along the route through South and Southeast Asia, the Australian evidence could become part of a technological tradition transmitted across enormous distances during the dispersal of Homo sapiens. If it is not, Nauwalabila may instead document an independent Australian solution to a difficult materials problem.

Either result would tell us more about prehistoric technology than simply moving another invention a few thousand years further back.

Selected Sources and Further Reading

1. Schmidt, Patrick & Peter Hiscock. 2026. “Earliest Lithic Heat Treatment in Australia is the World's Oldest Known Treatment of Chert.” Journal of Paleolithic Archaeology 9, 18. https://doi.org/10.1007/s41982-026-00262-5
2. Brown, Kyle S., Curtis W. Marean, Andy I. R. Herries, Zenobia Jacobs, Chantal Tribolo, David Braun, David L. Roberts, Michael C. Meyer & Jocelyn Bernatchez. 2009. “Fire As an Engineering Tool of Early Modern Humans.” Science 325: 859–862. https://doi.org/10.1126/science.1175028
3. Wadley, Lyn, Tamaryn Hodgskiss & Michael Grant. 2009. “Implications for Complex Cognition from the Hafting of Tools with Compound Adhesives in the Middle Stone Age, South Africa.” Proceedings of the National Academy of Sciences 106: 9590–9594. https://doi.org/10.1073/pnas.0900957106
4. Henshilwood, Christopher S., Francesco d’Errico, Karen L. van Niekerk, Yvan Coquinot, Zenobia Jacobs, Stein-Erik Lauritzen, Michel Menu & Renata García-Moreno. 2011. “A 100,000-Year-Old Ochre-Processing Workshop at Blombos Cave, South Africa.” Science 334: 219–222. https://doi.org/10.1126/science.1211535
5. Wilkins, Jayne, Benjamin J. Schoville, Kyle S. Brown & Michael Chazan. 2012. “Evidence for Early Hafted Hunting Technology.” Science 338: 942–946. https://doi.org/10.1126/science.1227608
6. Hardy, B. L., M.-H. Moncel, C. Kerfant, M. Lebon, L. Bellot-Gurlet & N. Mélard. 2020. “Direct Evidence of Neanderthal Fibre Technology and Its Cognitive and Behavioral Implications.” Scientific Reports 10, 4889. https://doi.org/10.1038/s41598-020-61839-w
7. Schmidt, Patrick, Tabea J. Koch, Matthias A. Blessing, F. Alexandros Karakostis, Katerina Harvati, Veit Dresely & Armelle Charrié-Duhaut. 2023. “Production Method of the Königsaue Birch Tar Documents Cumulative Culture in Neanderthals.” Archaeological and Anthropological Sciences 15, 84. https://doi.org/10.1007/s12520-023-01789-2
8. Schmidt, Patrick, Armelle Charrié-Duhaut, Edmund February & Lyn Wadley. 2024. “Adhesive Technology Based on Biomass Tar Documents Engineering Capabilities in the African Middle Stone Age.” Journal of Human Evolution 194: 103578. https://doi.org/10.1016/j.jhevol.2024.103578
9. Schmidt, Patrick, Armelle Charrié-Duhaut, Edmund February & John Parkington. 2026. “Biomass Tar, Podocarpus Leaves and the Diversity of Southern African Stone Age Adhesives: New Data from Elands Bay Cave.” Archaeological and Anthropological Sciences 18, 82. https://doi.org/10.1007/s12520-026-02451-3

Suggested keywords

prehistoric technology; lithic heat treatment; Nauwalabila I; Arnhem Land; chert; pyrotechnology; prehistoric adhesives; hafting; composite tools; fibre technology; cordage; materials engineering

Editorial summary

The 2026 identification of very early chert heat treatment at Nauwalabila provides a timely reason to reconsider prehistoric technology not as a succession of increasingly elaborate objects, but as a deep history of materials knowledge, controlled transformation and multi-stage production whose most sophisticated components have often disappeared from the archaeological record.