Independent research libraryEvidence first · sources openHow claims are graded
SCRIBEThe research behind the world’s biggest questions
SCR-076 · Human Origins · HUMAN EVOLUTION · FIRE · TECHNOLOGY

When Did Humans First Control Fire — and When Could They Make It?

Share this investigation

The first burned deposit, the first managed flame and the first ability to ignite fire are not the same milestone. New evidence pushes opportunistic fire use deep into the Early Pleistocene, while direct evidence for fire-making appears much later.

Last reviewed 15 August 20269 sourcesVerdict version 1
When Did Humans First Control Fire — and When Could They Make It? — SCRIBE editorial image
SCRIBE editorial image · visual context, not direct evidence unless a separate source credit says otherwise
Evidence rule: Fire has at least four archaeological thresholds: encountering a natural blaze, transporting or maintaining it, repeatedly organising activity around it, and deliberately producing a new flame. Burned bone or sediment can establish heating without proving every later step.

One discovery question is really four

Popular accounts ask when humans ‘discovered fire’, but archaeology cannot reduce the transition to one moment. A hominin could exploit a landscape after wildfire, carry embers to a camp, feed and preserve an existing flame, organise repeated hearth-like activity, or generate sparks on demand. Those behaviours have different evidential signatures. Natural burning, sediment movement, mineral staining and later disturbance can also imitate parts of the record, so the strongest cases combine context, temperature-sensitive analysis, spatial patterning, artefacts and repeated use.

Wonderwerk now reaches farther back

A 2026 PLOS ONE study re-examined small-mammal bones from Strata 10 and 11 inside South Africa’s Wonderwerk Cave. The team used a rapid bone-luminescence protocol and checked it against Fourier-transform infrared spectroscopy. Stratum 11 lies within a dated interval of about 1.79–1.07 million years ago, probably towards its older end. Burned remains occurred within an in-situ archaeological sequence deep inside the cave, where wildfire is a much poorer explanation than at an open-air site. The authors therefore interpret the pattern as fire repeatedly brought into the cave by hominins. That is strong evidence of use; it is not an ignition kit and does not show how the flame began.

The one-million-year layer remains the contextual anchor

Wonderwerk Stratum 10, dated to roughly one million years ago, had already produced microscopic wood ash, burned bone and heated sediment together with stone artefacts showing heat-related fractures about 30 metres inside the cave. The combination is stronger than an isolated blackened object because the material is associated and preserved in place. No constructed hearth or fire-lighting implement was identified, so the most defensible conclusion is that hominins managed fire at the site, possibly after acquiring it from a natural source.

Open-air claims are important but harder to close

Earlier or broadly contemporary burning has been reported at African sites including Koobi Fora and Chesowanja, while new spectroscopy has identified heat exposure among artefacts and animal remains at Israel’s Evron Quarry between roughly 1.0 and 0.8 million years ago. These finds widen the search, but open landscapes introduce wildfire, redeposition and mixed-context alternatives. An early date alone does not make a case stronger: the crucial question is whether human activity, burning and secure stratigraphy converge in the same place.

Gesher shows repeated control by about 790,000 years ago

At Gesher Benot Ya‘aqov in Israel, dense clusters of burned flint microartefacts recur across several archaeological horizons rather than being evenly scattered. Later work linked some of those combustion areas to fish remains heated at low temperatures. The pattern supports repeated, spatially organised use of fire and probably cooking. Researchers have inferred that the occupants could make fire at will, but the site has not produced a preserved strike-a-light set. Repeated control is directly supported; deliberate ignition remains an inference from that regularity.

Barnham is the current fire-making threshold

At Barnham in Suffolk, a 400,000-year-old buried land surface contains repeatedly heated sediment, fire-cracked flint handaxes and two fragments of iron pyrite. The pyrite is locally rare and can produce sparks when struck with flint, making deliberate transport for ignition the explanation proposed by the Nature research team. Geochemical work indicates localised, repeated high-temperature burning rather than a passing wildfire. This is the earliest current archaeological package specifically directed at fire-making. It is still contextual evidence: no action was observed, no hominin fossil identifies the fire-maker, and the published case depends on the convergence of pyrite, heat and site formation rather than one diagnostic object.

Fire use does not automatically date routine cooking

Fire could provide light, warmth, protection, landscape access and occasional cooked food before it became a dependable part of daily subsistence. The cooking hypothesis proposes a much earlier evolutionary dependence on cooked food, but anatomy cannot by itself supply a date for habitual ignition and archaeological preservation is uneven. A decisive earlier fire-making case would need a securely dated context combining repeated combustion with a transported ignition material or tool, diagnostic use-wear or residues, and independent exclusion of wildfire and redeposition.

Research record

Marin-Monfort et al. (2026) — New evidence for Early Pleistocene use of fire at Wonderwerk CavePeer-reviewed · open access · open sourceBerna et al. (2012) — Microstratigraphic evidence of in-situ fire at Wonderwerk CavePeer-reviewed · open access · open sourceStepka et al. (2022) — Hidden signatures of early fire at Evron QuarryPeer-reviewed · open access · open sourceAlperson-Afil (2008) — Continual fire-making at Gesher Benot Ya‘aqovPeer-reviewed · open sourceZohar et al. (2022) — Evidence for cooking fish 780,000 years ago at Gesher Benot Ya‘aqovPeer-reviewed · open sourceRoebroeks & Villa (2011) — Earliest evidence for habitual fire use in EuropePeer-reviewed review · open access · open sourceDavis et al. (2026; online 2025) — Earliest evidence of making fire at BarnhamPeer-reviewed · open sourceGoldberg, Miller & Mentzer (2017) — Recognizing fire in the Paleolithic archaeological recordPeer-reviewed review · open sourceWrangham (2017) — Control of fire in the Paleolithic: evaluating the cooking hypothesisPeer-reviewed review · open source
SCRIBE VERDICT

When Did Humans First Control Fire — and When Could They Make It?

Direct answer: Hominins were probably transporting and managing naturally acquired fire at Wonderwerk Cave between about 1.79 and 1.07 million years ago, with especially strong in-situ evidence by roughly one million years ago. Repeated control is well supported at Gesher Benot Ya‘aqov by about 790,000 years ago. The earliest current evidence specifically for making fire is Barnham at about 400,000 years ago. No single date captures the whole transition.

BEST SUPPORTED: opportunistic fire use and maintenance extend into the Early Pleistocene; secure contextual evidence is present at Wonderwerk by about one million years ago, repeated spatial control is established by about 790,000 years ago, and Barnham currently provides the earliest converging evidence for deliberate ignition at about 400,000 years ago. UNRESOLVED: exactly which hominin groups crossed each threshold, how often early fire was used, whether Gesher’s occupants could always ignite it, and when cooking became obligatory. NOT DEMONSTRATED: one universal ‘discovery of fire’, an ignition date inferred from burned material alone, or a requirement that the earliest Homo erectus routinely cooked every meal.

Share this investigation
Share the evidence

Share this investigation

Send the sourced investigation rather than a headline without its evidence trail.

Open research · SCR-076

Add to this investigation

Found a stronger source, correction, counter-evidence or new lead? Send it to SCRIBE. Contributions are reviewed against the evidence policy and do not change a published verdict automatically.

The investigation ID and page address are prefilled. Public credit is optional.