Re-evaluation of the age of Zhoukoudian, a prominent site of Homo erectus occupation in China, prompts a rethink of the species' distribution in both the temperate north and the equatorial south of east Asia.
The Zhoukoudian cave system, discovered in 1918 near the outskirts of modern-day Beijing, is one of the prize sites in palaeoanthropology (Fig. 1). It has produced a long sequence of well-preserved Homo erectus fossils and stone tools, unique in northeast Asia, that have been dated variously to between 0.4 million and 0.6 million years (Myr) ago. In their report on page 198 of this issue, Shen and colleagues1 push back the age of Zhoukoudian's lowest occupation level to 0.78 Myr ago. The new age estimate is consistent with the cold-climate fauna found within the locality's lowest levels, and it also correlates with revisions in the regional chronology of H. erectus sites2,3,4,5,6. Overall, the recent studies suggest that the onset of glacial cycles brought the open-land habitats in which H. erectus thrived.
The first fossils of H. erectus were found at Trinil in Java, Indonesia, in 1892 by Eugène Dubois. It was Dubois who coined the name 'erectus', based on a thigh bone that indicated upright posture7. Homo erectus had a heavily armoured and thickened skull, with an adult brain size in the 650–1,250-cm3 range8 — the equivalent value for modern humans (H. sapiens) is 1,100–1,800 cm3. Homo erectus stood 145–180 cm tall9, walked fully upright with a modern-like human footprint10, and used stone tools. The species is easily distinguished from H. sapiens by its distinctive torso, which was much more barrel-shaped and larger in volume11.
Fossil evidence now points to an origin of H. erectus about 2.0 Myr ago in equatorial Africa, followed by dispersal northwards and eastwards into Eurasia8. Incontestable evidence of that dispersal comes from the site of Dmanisi (Transcaucasia, Republic of Georgia) at 1.75 Myr ago and from southeast Asia (Java) by 1.6 Myr ago. Homo erectus survived to as late as 0.35 Myr ago in continental east Asia, and isolated populations persisted longer on the Indonesian archipelago. One group on Java, commonly known as 'Solo Man', may have survived until 50,000 years ago12.
As far as east Asia is concerned, Zhoukoudian has yielded 17,000 stone artefacts and fossil evidence for more than 50 H. erectus individuals8, and so provides a singular window on the life of H. erectus in the region's temperate north. Over the past several decades, discoveries in the surrounding basins and plains have broadened our knowledge, yet Zhoukoudian remained a temporal and environmental outlier. Surrounding Zhoukoudian are the open basins and plains comprising the H. erectus homeland at 40° N latitude. To the northwest, the lake-bed sediments of Hebei Province hold scattered open-air occupation sites (Maliang, Cenjiawan, Donggutuo, Xiaochangliang). To the southwest, Shanxi Province has yielded cranial remains of H. erectus at Gongwangling (Fig. 2, overleaf). Although these areas never saw glacial ice, glacial periods brought cooling and drying conditions. Open habitats, exemplified by grasslands and mixed steppes, expanded during glacial periods. Such environments favoured large grazing animals, with H. erectus being one of several large predators.
Magnetostratigraphy has been the primary dating technique for the lake and loess (wind-blown silt) deposits that harbour China's H. erectus record2,3,4,5,6. With this approach, one seeks evidence for Earth's known magnetic reversals preserved within a local sedimentary sequence; the age of a buried fossil or artefact horizon is determined by its location in the sequence. Early magnetic analyses produced impressively old ages of 1.7–1.4 Myr ago for Hebei's lake sites5,6. At nearly three times older than the previous age estimates for Zhoukoudian, it has been difficult to link the lake sites with the cave.
New work, however, is revealing the problems of open-air sedimentation rates, a crucial variable in magnetostratigraphy2. Revised ages have been provided for the Hebei sites of Maliang (0.78 Myr)4, Cenjiawan (1.1 Myr)3, Donggutuo (1.2 Myr)2 and Xiaochangliang (1.26 Myr)2, as well as for the Shanxi site of Gongwangling (1.22 Myr)2. With Zhoukoudian now estimated to be older, as reported by Shen et al.1 using an approach called cosmogenic 26Al/10Be burial dating, and the open-air sites to be younger, it seems that the early human occupation of northeast Asia began around 1.3 Myr ago. That occupation continued to at least 0.4 Myr ago based on the youngest occurrence of H. erectus at Zhoukoudian1.
Still at issue is how H. erectus dealt with the colder and drier glacial climates that prevailed at various times during the Pleistocene (1.8–0.01 Myr ago). The loess and palaeosol (ancient soil) sequence in Shanxi Province provides a record of at least ten glacial–interglacial cycles during the period of H. erectus occupation13,14. Comparing the ages of glacial (loess) and interglacial (palaeosol) periods with those of the H. erectus sites suggests that northeast China was occupied in both glacial and interglacial periods. During the early Pleistocene, before about 0.9 Myr ago, the cycles were shorter and less extreme. These conditions may therefore have enhanced the mammal-rich open environments on which H. erectus thrived, rather than bringing hardship, as is often thought. Once settled, H. erectus adjusted to the lengthening and deepening cycles.
To find anything analogous to Zhoukoudian and its surrounding basins, one must travel 5,000 km south to central Java. At 7° S latitude, the early Pleistocene lake and river deposits at Sangiran have yielded fossils of nearly 80 H. erectus individuals and several occupation sites. The H. erectus presence at Sangiran has a firm basal age of 1.6–1.5 Myr ago, and begins during an early Pleistocene glacial cycle15. A plausible view of events is that low sea level during this glacial period linked some of today's Indonesian archipelago with mainland southeast Asia. A previously mainland fauna, including H. erectus, made its way southwards, and thrived in the open woodland and savanna during the relatively dry glacial periods16. Interglacials brought higher sea levels, which submerged lowlands and increased moisture, shrinking open habitats. In Java, H. erectus persisted through these cycles until the Late Pleistocene12.
The temperate and equatorial populations were probably not connected. One factor may have been the existence of the subtropical primal forest in southern China and Indochina inhabited by the Stegodon–Ailuropoda fauna, which included the giant panda, gibbons, orangutans and the giant extinct ape Gigantopithecus17. The forest and its fauna may have formed an impenetrable ecological barrier between the two H. erectus populations dwelling in open habitats. The picture emerges of separate H. erectus populations existing outside a vast and ecologically unattractive forest. Parallel foraging niches existed in open and edge environments in both the temperate and equatorial environments.
Palaeoanthropologists have assumed that east Asian H. erectus originated with one group that dispersed just once from a single source in Africa. It may be time to rethink. As evidence for the ability of H. erectus to penetrate the subtropical primal forest falls away, so does the connection between Zhoukoudian and earlier Sangiran. If Zhoukoudian and the northern open-air sites represent a population, that population seems to have arrived later than originally thought.
Africa and southwest Eurasia are equally likely as the areas from which east Asian H. erectus originated. The two Asian groups may always have been independent and derived from separate dispersals. One group could have found a coastal route across the Arabian Peninsula to the southern flanks of the Himalaya and into coastal southeast Asia8; Sangiran represents a product of this dispersal. Another group could have passed through central Asia and southern Mongolia to enter the temperate basins and plains of northeast Asia, Zhoukoudian being a result.
Finally, 26Al/10Be burial dating — the technique applied by Shen et al.1 — is of considerable interest in its own right. It capitalizes on the formation of unstable isotopes, 26Al and 10Be, in quartz grains exposed to cosmic radiation at the land surface, and their differential decay in deeply buried sediments shielded from cosmic radiation. In caves, where surface sediment rapidly becomes deeply buried, burial dating provides a cross-check on age estimates derived by magnetostratigraphy. The approach also serves as a dating option in situations where magnetostratigraphy is problematic.
Early Pleistocene lake sediments in southwest China have been recently dated using the new technique18. This study used burial ages to constrain the age of fault movements that influenced the evolution of the middle Yangtze River. Likewise, 26Al/10Be burial dating will advance our understanding of H. erectus throughout its range by providing a new way of dating the Hubei Basin lake sites, as well as other H. erectus sites throughout east Asia.
Shen, G., Gao, X., Gao, B. & Granger, D. E. Nature 458, 198–200 (2009).
Li, H. et al. Quat. Res. 69, 250–262 (2008).
Wang, H. et al. Sci. China (D) 49, 295–303 (2006).
Wang, H. et al. Quat. Res. 64, 1–11 (2005).
Zhu, R. X. et al. Nature 431, 559–562 (2004).
Zhu, R. X. et al. Nature 413, 413–417 (2001).
Shipman, P. & Storm, P. Evol. Anthropol. 11, 108–116 (2002).
Boaz, N. T. & Ciochon, R. L. Dragon Bone Hill: An Ice Age Saga of Homo erectus (Oxford Univ. Press, 2004).
Lordkipanidze, D. et al. Nature 449, 305–310 (2007).
Bennett, M. R. et al. Science 323, 1197–1201 (2009).
Franciscus, R. G. & Churchill, S. E. J. Hum. Evol. 42, 303–356 (2002).
Swisher, C. C. III et al. Science 274, 1870–1874 (1996).
Zhou, L. & Shackleton, N. Earth Planet. Sci. Lett. 168, 117–130 (1999).
Liu, T. & Ding, Z. Annu. Rev. Earth Planet. Sci. 26, 111–145 (1998).
Larick, R. et al. Proc. Natl Acad. Sci. USA 98, 4866–4871 (2001).
Bettis, E. A. III et al. J. Hum. Evol. 56, 11–24 (2009).
Rink, W. J. et al. Quat. Res. 69, 377–387 (2008).
Kong, P. et al. Earth Planet. Sci. Lett. 278, 131–141 (2009).
About this article
Comptes Rendus Palevol (2018)
Quaternary Science Reviews (2013)
Comptes Rendus Palevol (2012)
High-resolution record of the Matuyama-Brunhes transition constrains the age of Javanese Homo erectus in the Sangiran dome, Indonesia
Proceedings of the National Academy of Sciences (2011)