Considerable attention has been paid to dating the earliest appearance of hominins outside Africa. The earliest skeletal and artefactual evidence for the genus Homo in Asia currently comes from Dmanisi, Georgia, and is dated to approximately 1.77–1.85 million years ago (Ma)1. Two incisors that may belong to Homo erectus come from Yuanmou, south China, and are dated to 1.7 Ma2; the next-oldest evidence is an H. erectus cranium from Lantian (Gongwangling)—which has recently been dated to 1.63 Ma3—and the earliest hominin fossils from the Sangiran dome in Java, which are dated to about 1.5–1.6 Ma4. Artefacts from Majuangou III5 and Shangshazui6 in the Nihewan basin, north China, have also been dated to 1.6–1.7 Ma. Here we report an Early Pleistocene and largely continuous artefact sequence from Shangchen, which is a newly discovered Palaeolithic locality of the southern Chinese Loess Plateau, near Gongwangling in Lantian county. The site contains 17 artefact layers that extend from palaeosol S15—dated to approximately 1.26 Ma—to loess L28, which we date to about 2.12 Ma. This discovery implies that hominins left Africa earlier than indicated by the evidence from Dmanisi.
Subscribe to Journal
Get full journal access for 1 year
only $3.90 per issue
All prices are NET prices.
VAT will be added later in the checkout.
Rent or Buy article
Get time limited or full article access on ReadCube.
All prices are NET prices.
Ferring, R. et al. Earliest human occupations at Dmanisi (Georgian Caucasus) dated to 1.85–1.78 Ma. Proc. Natl Acad. Sci. USA 108, 10432–10436 (2011).
Zhu, R. X. et al. Early evidence of the genus Homo in East Asia. J. Hum. Evol. 55, 1075–1085 (2008).
Zhu, Z. Y. et al. New dating of the Homo erectus cranium from Lantian (Gongwangling), China. J. Hum. Evol. 78, 144–157 (2015).
Zaim, Y. et al. New 1.5 million-year-old Homo erectus maxilla from Sangiran (Central Java, Indonesia). J. Hum. Evol. 61, 363–376 (2011).
Zhu, R. X. et al. New evidence on the earliest human presence at high northern latitudes in northeast Asia. Nature 431, 559–562 (2004).
Ao, H., Dekkers, M. J., Wei, Q., Qiang, X. & Xiao, G. New evidence for early presence of hominids in North China. Sci. Rep. 3, 2403 (2013).
Liu, T. S. et al. Loess and the Environment (China Ocean, Beijing, 1985).
Rutter, N., Ding, Z., Evans, M. E. & Liu, T. Baoji-type pedostratigraphic section, Loess Plateau, north-central China. Quat. Sci. Rev. 10, 1–22 (1991).
Lisiecki, L. E. & Raymo, M. E. A Pliocene-Pleistocene stack of 57 globally distributed benthic δ18O records. Paleoceanography 20, PA1003 (2005).
Ding, Z. L. et al. Stacked 2.6-Ma grain size record from the Chinese loess based on five sections and correlation with the deep-sea 18O record. Paleoceanography 17, 5-1–5-21 (2002).
Heller, F. & Liu, T.-S. Magnetostratigraphical dating of loess deposits in China. Nature 300, 431–433 (1982).
Kukla, G. & An, Z. Loess stratigraphy in central China. Palaeogeogr. Palaeoclimatol. Palaeoecol. 72, 203–225 (1989).
Zheng, H., An, Z. & Shaw, J. New contributions to Chinese Plio-Pleistocene magnetostratigraphy. Phys. Earth Planet. Inter. 70, 146–153 (1992).
Heslop, D., Langereis, C. G. & Dekkers, M. J. A new astronomical timescale for the loess deposits of Northern China. Earth Planet. Sci. Lett. 184, 125–139 (2000).
Guo, Z. T., Jiang, W. Y., Lu, H. Y., Wu, N. Q. & Yao, X. F. Pleistocene climate extremes in East Asia and their causes. Earth Sci. Frontiers 9, 113–120 (2002).
Liu, Q., Roberts, A. P., Rohling, E. J., Zhu, R. & Sun, Y. Post-depositional remanent magnetization lock-in and the location of the Matuyama–Brunhes geomagnetic reversal boundary in marine and Chinese loess sequences. Earth Planet. Sci. Lett. 275, 102–110 (2008).
Gibbard, P. & Cohen, K. M. Global chronostratigraphical correlation table for the last 2.7 million years. Episodes 31, 243–247 (2008).
Cohen, K. M. & Gibbard, P. L. Global chronostratigraphical correlation table for the last 2.7 million years. http://www.stratigraphy.org/index.php/ics-chart-timescale (2016).
Liu, Q. et al. Magnetostratigraphy of Chinese loess–paleosol sequences. Earth Sci. Rev. 150, 139–167 (2015).
Sun, Y., Clemens, S. C., An, Z. & Yu, Z. Astronomical timescale and palaeoclimatic implication of stacked 3.6-Myr monsoon records from the Chinese Loess Plateau. Quat. Sci. Rev. 25, 33–48 (2006).
An, Z. & Ho, C. New magnetostratigraphic dates of Lantian Homo erectus. Quat. Res. 32, 213–221 (1989).
Dai, E.-J. & Chi, H.-G. Discovery of palaeoliths at Lantian, Shensi. Vertebrata PalAsiatica 8, 152–156, 161 (1964).
Tai, E. C. & Hsu, C. H. New finds of palaeoliths from Lantian. Acta Archaeologica Sinica 1, 1–12 (1973).
Wang, S. et al. Newly discovered Palaeolithic artefacts from loess deposits and their ages in Lantian, central China. Chin. Sci. Bull. 59, 651–661 (2014).
Zhuo, H. X. et al. Chronology of newly-discovered paleolithic artifact assemblages in Lantian (Shaanxi province), central China. Quat. Res. 86, 316–325 (2016).
Cande, S. C. & Kent, D. V. Revised calibration of the geomagnetic polarity timescale for the Late Cretaceous and Cenozoic. J. Geophys. Res. 100, 6093–6095 (1995).
Ogg, J. G. in The Geologic Time Scale 2012 (eds Gradstein, F. M. et al.) 85–114 (Elsevier, Oxford, 2012).
Roberts, A. P. Geomagnetic excursions: knowns and unknowns. Geophys. Res. Lett. 35, L17307 (2008).
Kappelman, J. et al. in Geology and Paleontology of the Miocene Sinap Formation, Turkey (eds Fortelius, M. et al.) 41–66 (Columbia Univ. Press, New York, 2003).
Ranov, V. The ‘loessic palaeolithic’ in South Tadjikistan, central Asia: its industries, chronology and correlation. Quat. Sci. Rev. 14, 731–745 (1995).
Sun, Y. B., Lu, H. Y. & An, Z. S. Grain size distribution of quartz isolated from Chinese loess/paleosol. Chin. Sci. Bull. 45, 2296–2298 (2000).
Liu, Q., Deng, C., Torrent, J. & Zhu, R. Review of recent developments in mineral magnetism of Chinese loess. Quat. Sci. Rev. 26, 368–385 (2007).
This research was supported by the National Basic Research Program of China (Grant 2010CB833400), projects of NSFC (grants 41102115 and 41662012) and Projects of Chinese Academy of Sciences (grants KZCX2-SW-133, KZCX3-SW-152, 2013TIZ0008, XDB26000000, SKLLQG1525, SKLLQG1502, SKLLQG1501 and SKLLQG1122). This is contribution number IS-2546 from GIGCAS and Key Deployment Projects of IVPPCAS. We thank Z. An, R. Zhu, Z. Ding, Z. Guo, Z. Qiu, W. Liu, Y. Pan, W. Dong, H. Tong, H. Zheng, X. Tan, X. Qiang, H. Lu, Y. Pan and C. Deng for guidance in loess–palaeosol stratigraphy, palaeomagnetism and palaeoanthropology; the government of Lantian County and Cultural Relic Management Department of Gongwangling for helping our fieldwork, as well as our colleagues, Y. Kuang, Y. Han, S. Qin, H. Huang, S. Peng, M. Li, Z. Ruan, R. Deng, Y. Hao, Y. Chen, W. Chen, F. Li and Z. Li; and L. Hurcombe for advice on the artefacts.
Nature thanks J. Kappelman, M. Petraglia, A. Roberts and the other anonymous reviewer(s) for their contribution to the peer review of this work.
The authors declare no competing interests.
Publisher’s note: Springer Nature remains neutral with regard to jurisdictional claims in published maps and institutional affiliations.
Extended data figures and tables
a, b, Grain size distributions for loess (a) and palaeosol (b) samples from the Shangchen section. The features of sediments in Shangchen section are the same as those from the Lingtai section31, which is one of the typical Chinese loess sections. This indicates that sediments from the Shangchen section are aeolian deposits that are similar to those at Lingtai. c, Comparison of major elements between the loess sediments at Lantian and Luochuan7. The x and y coordinates represent logarithmic values of the major element content of loess from Luochuan and Lantian, respectively, and the plot shows that loess from the two sites is geochemically similar (data fall on or close to the 1:1 line). d, Chondrite-normalized rare-earth element distribution patterns for loess and palaeosol samples from the Shangchen section. Characteristics of partition modes of rare-earth elements from loess and palaeosol samples in the Shangchen section indicate they are the same as those from the Luochuan section.
a, Temperature-dependent magnetization variations for four representative samples from the Shangchen section. Arrows indicate heating or cooling runs. These curves were obtained in air using a field of 100 mT, and indicate the presence of maghemite, dominant magnetite and haematite. This type of magnetic mineral assemblage is typical of that previously recovered for the Chinese loess–palaeosol sequence19,32. b, Hysteresis loops for representative samples after correction for paramagnetic slope. Ms, saturation magnetization; Mrs, saturation remanence; Bc, coercive force; and Bcr, coercivity of remanence. c, Anisotropy of magnetic susceptibility for 694 specimens from loess L5 to loess L28 in the main section that we studied, and parallel sections at the Shangchen locality. Left, Flinn diagram. L, lineation (κmax/κint). F, foliation (κint/κmin). Right, stereographic projections of anisotropy-of-magnetic-susceptibility ellipsoids of specimens. Data for κmax and κmin are shown as black squares and red circles, respectively. The data are indicative of normal, undisturbed sedimentary fabrics and support the magnetostratigraphic interpretation presented in this study. d, Equal area projections for the natural remanent (left) and characteristic remanent (right) magnetization directions for all 694 studied specimens.
a–o, Demagnetization diagrams for representative specimens from the Shangchen section. Solid and open symbols refer to data projected onto the horizontal and vertical planes, respectively. Scales of demagnetization temperature and coordinate axes are in degrees Celsius and mA m−1, respectively. In m, n, the specimens from the magnetic horizon of the Réunion excursion are shown. See Source Data for detailed listings of palaeomagnetic data from the Shangchen section.
Extended Data Fig. 4 Landscape in which palaeomagnetic sampling and artefact collecting were carried out at the Shangchen locality.
The establishment of magnetostratigraphy in the Shangchen locality was based on the traditional and effective-linking methods (using marker layers and palaeomagnetic reversal boundaries of offset sections29). The comprehensive main section and timescale of the Shangchen locality have been established from five subsections, the spatial distribution of which is shown below. a, b, The main sections are exposed continuously along the same gully. Because of the steep terrain and multistep gully bottom (see c, d), we divided the sampling into four subsections (offset sections) in this gully. Several marker layers—such as L9, L15 and L25—that are white and thick are obvious in the section at short range and can easily be used to link the subsections. The subsections I, II, III and IV contain the following layers: I (L5–L15), II (L15–L25), III (L25–L27), and IV (L27-L28, on the other side of the slope). c, The steep subsection II with fresh outcrops of original loess and palaeosols. Many artefacts were found within more than 10 horizons in subsection II or on both its sides. The first artefact was found here in S22. Note the two individuals on the upper part of the section as an approximate scale. d, Subsection V, which is a short parallel section opposite subsection II (c). e, Many artefacts were found in situ in palaeosol S22 during our sampling and excavation. Palaeomagnetic analysis confirms that the artefacts were collected from undisturbed loess or palaeosol (see Supplementary Information). The steepness of slope (shown in c, d) prevents large-scale excavation. More details can be seen in Extended Data Fig. 6. f, g, The section of the exploratory trench (named subsection KW) contains layers L25–L28. It is part of the same small hill as the main composite section (subsections I–IV), and is located about 500 m northeast of palaeomagnetic section IV (see Extended Data Fig. 5a). Based on the outcrops of marker layers L9, L15 and L25 in the loess–palaeosol sequence, these sections can easily be linked (see Extended Data Fig. 5 for further details).
a, Schematic plan of the subsections of palaeomagnetic sampling. The numbers 800 and 900 refer to the height in metres above sea level. The contour interval is 20 m. Subsection KW refers to the additional parallel section from L25 to L28 with the excavation into S27 and L28. b, Schematic of the method used to link the five subsections in the Shangchen locality. This figure and Extended Data Fig. 4 show the loess–palaeosol stratigraphy, with key marker layers, sampled subsections and the locations where artefacts were found. The details of the method used to link the five subsections are given below. Subsections I and II are linked on the basis of the L15/S15 boundary. Measured horizons of subsection I start from the mid-to-lower part of L5 and run downward to the base of marker layer L15 (to the boundary of L15/S15). Measured horizons of subsection II start from the lowest part of L15 and run downward to the lower part of L25. Because the measured thickness of marker layer L15 is the same in both subsections I and II, we cut out the overlapped L15 in subsection II. Thus, subsection I ends at the base of L15, and subsection II begins at the top of S15. Subsections II and III are linked on the basis of the top boundary of the Olduvai subchron. A small normal-polarity segment of the Olduvai subchron was detected in the marker layer L25 at the base of II. Measured horizons of subsection III start from the lower part of L25 to the top of L27, and record the entire Olduvai subchron with a small reversed-polarity segment on the top (in L25) and at the base (top of L27) of the subsection. Thus, the overlapped normal-polarity segment of the Olduvai subchron at the base of subsection II and a small reverse-polarity segment on top of subsection III were cut out. The reverse-polarity segment at the base of subsection II is linked to the normal-polarity top segment of subsection III, and the whole normal-polarity segment of the Olduvai subchron is contained within subsection III. Subsections III and IV were linked on the basis of the base of the Olduvai subchron. Measured horizons of subsection IV start from the top of L27 to the mid-to-upper part of L28, with a small normal-polarity segment of the base of the Olduvai subchron at the top. After cutting out both the small reverse-polarity segment at the base of subsection III and the small normal-polarity segment on top of subsection IV, the reverse-polarity segment just below the Olduvai subchron in subsection IV is linked to the normal-polarity base of the Olduvai segment in subsection III. Subsection KW contains horizons of L25 (marker layer) to L28, including units L25, S25, L26, S26, L27, S27 and L28. Against the general background of reversed magnetostratigraphy, two normal magnetozones—the Olduvai subchron (1.78–1.95 Ma) and the Réunion excursion (2.13–2.15 Ma)—are recorded. Therefore, section IV and section KW are linked by the base of S27 (that is, S27 at the base of section IV is linked to L28 at the top of subsection KW). Subsection V is opposite of subsection II, which is on the other side of a narrow gully and has two distinct marker layers, L15 in the upper half and L25 in the lower half. There is a palaeosol layer with many artefacts in the middle part of subsection V. The stratigraphic correlation and horizon number demonstrate that this layer belongs to S22. Palaeomagnetic measurements on samples collected from this parallel section all showed reversed polarity.
Extended Data Fig. 6 Stone artefacts found during the sampling of S22 (1.54–1.57 Ma) and S24 (1.71–1.73 Ma) at the Shangchen section.
a, First discovery of a stone artefact in S22 (‘place I’). b, Schematic profile of the section-cutting of place II. c, The place II section was excavated into the original palaeosol layer to a depth of about 150 cm. Note the presence of stone artefacts on a ledge in the original palaeosol, below the weathered crust. d, View of the excavation in subsection V. e, Profile of the original slope surface with a weathered crust of only 3–6-cm thickness, and no slope wash. Palaeomagnetic analysis of the samples taken around the stone artefacts confirmed that they were in undisturbed deposit. f, View of the excavation in S24. Note the weathered surface crust upslope, the slope wash to the left of the cutting and section cleaning debris below the excavation. g, Schematic profile of excavation into S24. h, Side view of place I of the section cut by the excavation, showing two artefacts within fresh original palaeosol. Note that there is a clear difference between the thin weathered crust and the original loess. i–k, Discovery and exposure of a large quartzite core in place II of the section (i, j), which led to the removal of the core shown in k. Details of the large core (SC080710-1) are given in Supplementary Table 6. Analysis of the palaeomagnetic samples taken alongside the stone artefacts showed that the loess and palaeosol was undisturbed, and did not represent slope wash.
Extended Data Fig. 7 Details of artefacts excavated from the roadside section at the Shangchen locality from S27, which occurs below the Olduvai subchron and is dated to approximately 2.1 Ma.
The hill slope is covered by a dark brown slope wash of 10–30-cm thickness that is easily distinguishable from the in situ loess in the roadside section. After cleaning the section, stone artefacts were exposed between 1.5 and 3 m below the top of the section, and extracted from in situ loess. a, b, This shows the obvious difference between slope wash and the original loess–palaeosol layer. b, First finding of a stone standing out of the surface crust which is situated about 1.5 m below the top of slope wash. c–f, The finding and collecting process for the artefact is the same in Extended Data Fig. 6. Details of the artefact SC 20120507-3—which was found after digging >30 cm into the fresh original palaeosol layer—are given in Fig. 4a, Supplementary Tables 6, 7 and Supplementary Video 1. g–i, A stone is marked with its specimen number before removal (g), and the artefact is disclosed step by step (h, i). j, The stone artefact after removal. k, View of core (SC20120507-1). Clear flake scars are evident. This piece is shown in Fig. 4d and details are given in Supplementary Tables 6, 7.
Extended Data Fig. 8 The stratigraphic partition, grid layout and distribution of artefacts and fossils in the exploratory trench (subsection KW) in S27 and L28.
a, b, The vertical section (a) and the horizontal plane (b) of the exploratory trench, which is composed of a primary and a secondary trench. The primary exploratory trench is 6 m long, 1.4–1.78 m deep and 2 m wide, and the secondary trench is 1.7 m long, 2 m deep and 2 m wide. The excavated area (approximately 15.4 m2) was divided into a grid of 1-m × 1-m squares. I, II, the numbers of grid lines on the vertical section. A1–A6 and B1–B6, numbers of the grid squares on the horizontal plane in the primary trench. –A1, –B1, the numbers of grid lines on the horizontal plane in the secondary trench. In the south and the north of the primary trench, 108 palaeomagnetic samples were continuously collected in two sampling trenches. The north sampling trench is 0.8–1 m wide and 2.3 m deep, and the south sampling trench is 0.4 m wide and 1.2 m deep. Some stone artefacts and fossil remains were excavated in situ in S27 and L28 from our exploratory trench. The features of eight artefacts can be seen in Fig. 4, Extended Data Fig. 10 and Supplementary Table 6. The fossils from the excavation were primarily tooth and bone fragments (Extended Data Fig. 10) of different animals, including a cervid, a bovid and a suid. No bones or jaws were complete. The fossil remains and stone artefacts occurred on the same horizon and were close to one another. Note that the artefacts and fossils also occur in the deepest part of the trench and not near the modern slope surface.
Extended Data Fig. 9 The excavation of the exploratory trench (the subsection KW) in layers S27 and L28 at Shangchen locality.
a–c, The primary (a, c) and secondary (b) exploratory trenches. d, e, Palaeomagnetic sampling trench in the north of the primary exploratory trench. The boundary between the cultivated horizon (cropland soil) and original loess horizon (S27) is very clear. The top layer is a grass-covered cropland soil of a depth of 20–40 cm (without any crops planted at this time). Underlying the cultivated layer are clear, homogeneous, solid and undisturbed loess–palaeosol strata. Several stones and fossils were found in situ in the original palaeosol–loess strata (S27–L28). These stones and mammalian fossils were kept in the reserved pillars of palaeosol and loess during the exploration process (a–c). The Réunion excursion is situated in the lower part (L28) of the palaeomagnetic sampling trench (e).
Extended Data Fig. 10 Selected artefacts and fossils from the exploratory trench (subsection KW) in S27 and L28.
These are the oldest artefacts discovered at the Shangchen locality during this investigation. a, Quartzite core (SC-K4) from the boundary between S27 and L28 at a depth of 60–70 cm below the boundary between cultivated soil and fresh original palaeosol, and 164 cm above the end of the Réunion excursion. See Fig. 4i and Supplementary Tables 6, 7 for details. b, c, The oldest artefacts SC-K30 and SC-K55 within L28 from the exploratory trench, from 69–114 cm above the end of the Réunion excursion. See Supplementary Table 6 for details. d, Quartzite flake tool (SC-K5) from the exploratory trench, from the boundary between S27 and L28 at a depth of about 75 cm below the boundary between cultivated soil and fresh original palaeosol and 149 cm above the end of the Réunion excursion. See Fig. 4h and Supplementary Tables 6, 7 for details. e, Flake tool SC-K54 from within the L28 from the exploratory trench. f, Mandibular fragment of a small bovid (SC-K60) from the same horizon of S27 as the stone artefacts found during the excavation of S27 and L28. Top image shows the occlusal view of the mandible. g, A left mandibular fragment of a large cervid (SC-K10), which was found near and in the same horizon of S27–L28 as the stone artefacts SC-K4 and SC-K5. In the scales, each division represents 1 cm.
This file contains Supplementary Text, Tables and References: I. The loess-palaeosol timescale, including (1) The loess-palaeosol timescale and (2) Dating by sedimentation rates. II. Main methods and results, including (1) Grain size, (2) Mineralogy, (3) Geochemistry, (4) Palaeomagnetism. III. Archaeological field procedures, and IV. Stone artefacts from the Shangchen locality.
The 3D animation feature of the artefact (SC2012-0507-3) from layer S27.
The 3D animation feature of the artefact (SC2012-0507-2) from layer S27.
The 3D animation feature of the artefact (SC 20120502-6) from layer S23.
About this article
Cite this article
Zhu, Z., Dennell, R., Huang, W. et al. Hominin occupation of the Chinese Loess Plateau since about 2.1 million years ago. Nature 559, 608–612 (2018). https://doi.org/10.1038/s41586-018-0299-4
Magnetostratigraphy and Paleoenvironments of the Kuntila Lake Sediments, Southern Israel: Implications for Late Cenozoic Climate Variability at the Northern Fringe of the Saharo-Arabian Desert Belt
Frontiers in Earth Science (2020)
American Journal of Primatology (2020)
Magnetic characteristics of Guangshan loess from northern piedmont of Dabie Mountains, east-central China
Geophysical Journal International (2020)
Quaternary International (2020)
The Paleolithic in the Nihewan Basin, China: Evolutionary history of an Early to Late Pleistocene record in Eastern Asia
Evolutionary Anthropology: Issues, News, and Reviews (2020)