Reconstruction
Zhoukoudian lies where limestone hills meet flat farmland, some forty-five kilometres southwest of central Beijing. Locality 1 is not a cave in the ordinary sense but the fill of a collapsed karst fissure — roughly forty metres of clay, breccia, silt and rubble, excavated in campaigns from 1921 onward. It produced the remains of about forty individuals of Homo erectus, more than a hundred thousand stone artefacts, and tens of thousands of animal bones.
Almost none of the original fossils survive. Fourteen partial crania, with jaws and teeth, were packed for shipment out of China in December 1941 and disappeared in the chaos of the occupation. Franz Weidenreich's monographs, casts and photographs, together with the specimens recovered after 1949, are what modern research works from — a fact that constrains every taxonomic and taphonomic study since.
Dating rests on cosmogenic nuclides. Aluminium-26 and beryllium-10 burial ages on quartz from the lower archaeological layers give 0.77 ± 0.08 million years, older than previous estimates and falling in a glacial interval. Homo erectus was living through winters at forty degrees north.
The food evidence is contested at nearly every point. The deposits hold cut-marked and percussion-broken bone, mostly deer — the sika-like Pseudaxis and the giant Sinomegaceros — with horse, boar and rhinoceros present. Carbonised hackberry stones from the occupation layers are the standard illustration of plant food. But taphonomic re-examination showed the assemblage is dominated by a carnivore, the short-faced hyena Pachycrocuta brevirostris. Sixty-seven per cent of the hominin bones themselves carry bite marks or fracture patterns consistent with hyena chewing, and the carbon coating the skulls of Locus G — read for decades as proof of roasting and cannibalism — is better explained as detrital carbon settled out of standing water, possibly in a hyena cache. Peking Man was food here as often as forager.
Fire is the other unresolved argument, and it matters because generations of textbooks used Zhoukoudian's thick grey 'ash layers' as the earliest proof of domestic hearths. Micromorphology and mineralogy of Layers 4 and 10 recovered no siliceous aggregates from wood ash and no hearth structures, and identified much of the dark sediment as water-borne organic matter; burnt bone was present but not demonstrably burnt where it lay. Excavation resumed in 2009 and reported burned sediment plus roughly a hundred burned bones from Layer 4, at least fifteen of them heated above 600 degrees Celsius, which is difficult to achieve without deliberate fire.
The defensible reading is intermittent fire use in a fissure that hominins shared, uneasily and probably seasonally, with hyenas — not a hearth burning continuously for hundreds of millennia.
Historical context
Northern China was a cold, strongly seasonal frontier for hominins. Loess was piling across the plateau during glacial phases, and the Zhoukoudian fissure opened onto grassland, patchy woodland and permanent water where short-faced hyenas, cave bears, wolves and leopards competed with people for shelter and carcasses. Homo erectus was by now spread from the Caucasus to Java. In Africa, Acheulean toolmakers were still working the Kenya Rift at Olorgesailie; in Europe, the flint knappers of Boxgrove lay a few hundred thousand years in the future. Fire was not yet an established human possession anywhere: the strongest early claim of comparable age is Gesher Benot Ya‘aqov in the Jordan Valley at roughly 790,000 years, and fire only becomes archaeologically routine after about 400,000 years ago — which is precisely why the Zhoukoudian ash layers carried so much argumentative weight, and why their reassessment mattered so much.
Evidence
Dating
Moderate26Al/10Be burial dating of quartz from the lower archaeological layers (Layers 7–10) gives 0.77 ± 0.08 Ma (Shen et al. 2009). The uncertainty is ±80,000 years, so '770,000' is a midpoint, not a precise horizon. The top of the sequence is far less well constrained.
Archaeological
StrongMore than 100,000 lithic artefacts across the Locality 1 sequence, predominantly quartz and chert flakes and cores, establish repeated hominin presence beyond dispute.
Zooarchaeological
ModerateCut-marked and percussion-broken cervid bone documents hominin carcass processing, but Boaz et al. (2004) showed the bulk of the bone accumulation — including 67% of the hominin remains — bears Pachycrocuta brevirostris damage. Hominin and hyena signals are mixed throughout.
Preparation method
LimitedCooking cannot be demonstrated. Weiner et al. (1998) and Goldberg et al. (2001) found no wood-ash minerals or hearth features in Layers 4 and 10; Gao et al. (2017) report post-2009 burned sediment and bone heated above 600°C in Layer 4. Intermittent fire is plausible, habitual cooking is not established.
Archaeobotanical
LimitedCarbonised Celtis (hackberry) endocarps from the occupation layers are routinely cited as plant-food evidence, but they were recovered under early excavation standards and their taphonomic association with hominin consumption, as opposed to natural deposition, was never established.
Historical interpretation
ModerateThe site is best read as a palimpsest of episodic hominin visits interleaved with long intervals of carnivore denning, rather than as a continuously occupied home base with a permanent hearth.
Geographic attribution
StrongLocality 1 is a mapped, monumented locality with a published stratigraphic column; layer attributions for the fossil and artefact finds are documented.
Sources
- 1.Shen, G.; Gao, X.; Gao, et al. (2009). Age of Zhoukoudian Homo erectus determined with 26Al/10Be burial dating. Nature 458(7235): 198–200. doi:10.1038/nature07741Scientific literature
- 2.Weiner, S.; Xu, Q.; Goldberg, et al. (1998). Evidence for the Use of Fire at Zhoukoudian, China. Science 281(5374): 251–253. doi:10.1126/science.281.5374.251Scientific literature
- 3.Goldberg, P.; Weiner, S.; Bar-Yosef, et al. (2001). Site formation processes at Zhoukoudian, China. Journal of Human Evolution 41(5): 483–530. doi:10.1006/jhev.2001.0498Scientific literature
- 4.Gao, X.; Zhang, S.; Zhang, et al. (2017). Evidence of Hominin Use and Maintenance of Fire at Zhoukoudian. Current Anthropology 58(S16): S267–S277. doi:10.1086/692501Scientific literature
- 5.Boaz, N.T.; Ciochon, R.L.; Xu, et al. (2004). Mapping and taphonomic analysis of the Homo erectus loci at Locality 1 Zhoukoudian, China. Journal of Human Evolution 46(5): 519–549. doi:10.1016/j.jhevol.2004.01.007Scientific literature
- 6.Roebroeks, W.; Villa, P. (2011). On the earliest evidence for habitual use of fire in Europe. Proceedings of the National Academy of Sciences 108(13): 5209–5214. doi:10.1073/pnas.1018116108Scientific literature
- 7.Goren-Inbar, N.; Alperson, N.; Kislev, et al. (2004). Evidence of Hominin Control of Fire at Gesher Benot Ya‘aqov, Israel. Science 304(5671): 725–727. doi:10.1126/science.1095443Scientific literature
Limitations
- majorthe framing omits that Locality 1 is substantially a carnivore accumulation
Calling Zhoukoudian an 'iconic deep-time site' of Homo erectus without qualification reproduces the home-base reading that taphonomy has dismantled. Boaz et al. (2004, doi:10.1016/j.jhevol.2004.01.007) found 67% of the hominin skeletal sample bears bite marks or fragmentation attributable to large carnivores, principally the short-faced hyena Pachycrocuta brevirostris, and reinterpreted the Locus G carbon — long cited as evidence of roasting and cannibalism — as detrital carbon deposited under water. Homo erectus was prey at this site as well as predator.
- majorthe epoch label asserts fire that this site cannot securely deliver
Zhoukoudian is the canonical early-fire site, and its 'ash layers' were the standard textbook evidence. Weiner et al. (1998, doi:10.1126/science.281.5374.251) and Goldberg et al. (2001, doi:10.1006/jhev.2001.0498) found no wood-ash siliceous aggregates and no hearth features in Layers 4 and 10, and attributed the dark sediment largely to water-deposited organic matter. Gao et al. (2017, doi:10.1086/692501) report burned sediment and bone heated above 600°C from renewed Layer 4 excavation. Fire here is a live dispute; a reconstruction showing a long-maintained communal hearth states as fact what is still argued.
- moderatedate start is falsely precise
The 770,000 BCE figure comes from Shen et al. (2009, doi:10.1038/nature07741), whose 26Al/10Be burial age is 0.77 ± 0.08 Ma and applies specifically to Layers 7–10, not to the whole forty-metre fill. The ±80,000-year uncertainty should be visible in any text that quotes the number.
- moderatedate end is not securely dated
Hominin fossils at Locality 1 span roughly Layers 3–10 and the uppermost archaeological layers are usually placed somewhere between about 400,000 and 300,000 years ago on biostratigraphic and U-series grounds. A hard 400,000 BCE endpoint is a convention rather than a measured horizon.
- moderatethe single most consequential fact about the material is missing
The original Peking Man fossils — fourteen partial crania plus mandibles and teeth — were lost in December 1941 while being evacuated, and every subsequent study rests on Weidenreich's casts and descriptions plus post-1949 finds. Any account of the site's evidential strength that omits this overstates what can be re-examined.
- minorthe hackberry plant-food evidence is weaker than usually presented
Carbonised Celtis endocarps from the occupation layers are the near-universal citation for Peking Man's plant diet, but they were recovered under early excavation standards from a fissure fill with demonstrable water transport, and their association with hominin consumption rather than natural accumulation has not been established.
