Skip to main content
National Science Review logoLink to National Science Review
. 2017 Feb 24;6(5):1024–1039. doi: 10.1093/nsr/nwx015

Tianshanbeilu and the Isotopic Millet Road: reviewing the late Neolithic/Bronze Age radiation of human millet consumption from north China to Europe

Tingting Wang 1,2,3,✉,4, Dong Wei 4,4, Xien Chang 5, Zhiyong Yu 5, Xinyu Zhang 1,2, Changsui Wang 1,2, Yaowu Hu 1,2,, Benjamin T Fuller 2,
PMCID: PMC8291513  PMID: 34691966

Abstract

The westward expansion of human millet consumption from north China has important implications for understanding early interactions between the East and West. However, few studies have focused on the Xinjiang Uyghur Autonomous Region, the vast geographical area directly linking the ancient cultures of the Eurasian Steppe and the Gansu Corridor of China. In this study, we present the largest isotopic investigation of Bronze Age China (n = 110) on material from the key site of Tianshanbeilu, in eastern Xinjiang. The large range of δ13C values (–17.6‰ to –7.2‰; –15.5 ± 1.2‰) provides direct evidence of unique dietary diversity and consumption of significant C4 resources (millets). The high δ15N results (10.3‰ to 16.7‰; 14.7 ± 0.8‰) likely reflect sheep/goat and wild game consumption and the arid climate of the Taklamakan Desert. Radiocarbon dates from four individuals indicate Tianshanbeilu was in use between 1940 and 1215 cal bc. The Tianshanbeilu results are then analysed with respect to 52 Bronze Age sites from across Eurasia, to investigate the spread and chronology of significant human millet consumption and human migration. This isotopic survey finds novel evidence that the second millennium bc was a dynamic period, with significant dietary interconnectivity occurring between north China, Central Asia and Siberia. Further, we argue that this ‘Isotopic Millet Road’ extended all the way to the Mediterranean and Central Europe, and conclude that these C4 dietary signatures of millet consumption reflect early links (migration and/or resource transfer) between the Bronze Age inhabitants of modern-day China and Europe.

Keywords: Inner Asian Mountain Corridor, Silk Road, Xinjiang, Old World crop globalization, Shang Dynasty

INTRODUCTION

Before the Silk Road introduced the luxuries and technologies of China to Central Asia/Europe, a collection of regional and non-formalized long-distance networks of trade and communication, maintained by nomadic pastoralists, connected the East and West (e.g. Trans-Eurasian exchange, Proto-Silk Road, Inner Asian Mountain Corridor, etc.) [1–8]. Located at the geographical confluence of eastern and western cultures, the modern-day Xinjiang Uyghur Autonomous Region (Fig. 1a and b) served as a bridge between the ancient tribes of the Eurasian Steppe and the peoples of the Gansu Corridor of modern-day China [9–13]. Recently, genetic studies found that the Bronze Age of the Eurasian continent was a dynamic period that witnessed frequent human interactions and migrations [14–17]. The eastward movement of Indo-European populations into China resulted in the introduction of numerous south-west Asian crops and domesticated animals [18–21] as well as technological advances (e.g. metalworking, horse riding, wheeled vehicles) and changes in social structure [1,7,22,23]. Archaeological evidence indicates that by ~2000 bc, Indo-Europeans were migrating to the Tarim Basin from Central Asia [1,5,24] and components of the Afanasyevo (3700–2500 bc), Okunevo (2500–1900 bc) and the Andronovo (1900–1500 bc) cultural complexes are found incorporated into some of the earliest known archaeological sites of Xinjiang such as Qiemu’erqieke (a.k.a. Ke’ermuqi) [25–27], Xiaohe [14,28,29] and Gumugou [30,31].

Figure 1.

Figure 1

(a) and (b) Maps showing the location of the Tianshanbeilu cemetery site (✦) in the Xinjiang Uyghur Autonomous Region of China; (c) Color-painted potteries (http://www.bangbenw.com/yx/tpk/2010-08/25/content_903355.htm); (d) Bronze mirror with image of a human face (http://www.cchmi.com/tabid/702/InfoID/11395/Default.aspx); (e) Gold earrings from tomb M325 (individual not studied) excavated at Tianshanbeilu (http://www.xjbs.com.cn/news/2015-04/01/cms1756084article.shtml).

However, this exchange was not unidirectional and crops, technologies and peoples also journeyed westward from the cultural centers of the Yellow River Valleys of China [12,19,32–35]. In terms of cultigens, foxtail millet (Setaria italica) and common millet (Panicum miliaceum) were vital resources for the ancient inhabitants of north China with a long history of cultivation and consumption [36–41]. At some point between the third to second millennium bc, but possibly earlier, millet was transported to the West and eventually adopted by populations outside of China [42]. At present, the earliest undisputed evidence for millet beyond the borders of modern China is the direct radiocarbon dates of millet grains (2460–2150 cal bc) at the site of Begash in southern Kazakhstan [12,43]—a steppe area linked to the Gansu Corridor by the Tianshan Mountains of China (Fig. 1b). Archaeobotanical studies have also discovered millet remains at numerous other Bronze Age archaeological sites along the Tianshan Mountains, and its nearby mountain ranges [35], which together form the Inner Asian Mountain Corridor [6]. This Inner Asian Mountain Corridor is believed to be a key artery for early East–West interactions [6,10,12] and was possibly bound together by the common cultures of the Srubnaya (1800–1200 bc) and Andronovo (1900–1500 bc) of the western and eastern Eurasian Steppes, respectively [23].

While archaeobotanical studies yield invaluable information about the presence of a particular cultigen at an archaeological site, they do not provide direct evidence that these crops were consumed in significant amounts by a population [13,34,44]. However, carbon (δ13C) and nitrogen (δ15N) stable isotope ratio analysis provides a direct method to reconstruct human and animal dietary patterns [45–47]. In particular, δ13C signatures are well suited as natural biomarkers of millet consumption in China and Eurasia due to the large isotopic differences between important C3 (e.g. wheat (Triticum aestivum), barley (Hordeum vulgare)) and C4 cultigens (e.g. foxtail and common millet). For example, C3 plants select against 13C when fixing CO2, and thus have more negative δ13C values (–30‰ to –22‰) than C4 plants (–16‰ to –10‰), with mean values of –26.5‰ and –12.5‰, respectively [48–51]. These different plant δ13C values are then incorporated into a consumer's body tissues with known corresponding offsets or fractionation factors [52,53]. To date, stable isotope studies in archaeology are most often carried out on bone collagen, and +5‰ is commonly suggested as the offset between bone collagen and diet for mammals [54–59]. For studies of bone collagen, it is also estimated that ~20% of the dietary protein must originate from C4 sources in order for the isotopic signatures to be distinguishable from a predominately C3 diet. Previous research has defined a δ13C cut-off value of approximately –18‰ as the difference between predominately C3 and mixed C3/C4 diets and a value of –12‰ as the approximate boundary between mixed C3/C4 diets and predominately C4 diets [50,60,61].

Here, results of the largest isotopic study of Bronze Age China (n = 110), as well as radiocarbon dates from four individuals, are presented to investigate human diet and age of habitation at the Tianshanbeilu (TB) cemetery site in the Xinjiang Uyghur Autonomous Region of China. TB is regarded as one of the most important archaeological sites in eastern Xinjiang, and displays a vast wealth of archaeological, anthropological and genetic evidence for significant links between the cultures of the Gansu Corridor and the Eurasian Steppe (Fig. 1a–e, Archaeological Background). Over 3000 artifacts (bronze, pottery, bone, stone, gold, silver, cowries, etc.) were discovered from the approximately 700 tomb burials [62] and the large regional diversity of these items indicate some of the earliest and most dynamic cultural interactions between the East and the West in China [24]. Only limited isotopic research has been published for Bronze Age sites from the Xinjiang region [63–65], including a pilot isotopic study of 10 individuals from TB [66] that revealed mixed C3/C4 diets. In addition to our new data, we conduct a detailed review of 52 Bronze Age archaeological sites, comprising >1000 individuals, from across Eurasia. Here, our aim is to use these human δ13C and δ15N results as natural isotopic tracers to construct the ‘Isotopic Millet Road’ (note this could equally be called the ‘Millet Road’ if focused primarily on archaeobotanical remains), which can be used to investigate the westward radiation of millet consumption from the Yellow River Valleys. These results are plotted to visualize sites from the second millennium bc where individuals who consumed significant amounts of millet have been recovered and to identify the many possible routes of interconnectivity between the Yellow River Valleys of north China, Central Asia, Siberia and Europe.

Archaeological background

TB, also known as Linya, is located near the modern city of Hami in the Xinjiang Uyghur Autonomous Region of China in the eastern portion of the Inner Asian Mountain Corridor (Fig. 1a and b). TB is the largest Bronze Age cemetery site found in eastern Xinjiang to date [24,67]. Incorporating characteristics of the cultures from the Eurasian Steppes and the Gansu Corridor, the grave goods of the TB cemetery are remarkably diverse and indicate some of the earliest and most dynamic cultural interactions between the East and the West in China [24].

Large quantities of pottery discovered at TB were assigned to two different unique typologies (Fig. 1c). Group one has similar morphology, materials, and painted patterns with the pottery of the post-Machang and Siba cultures (~2000–1500 bc) from the Gansu Corridor [68]. Group two resembles pottery of the western Qiemu’erqieke (a.k.a. Ke`ermuqi) culture (~2000–1500 bc) from the Altai Mountains [4,25,26,69] and/or the Andronovo culture from the Eurasian Steppes [2,3,69]. Moreover, burial styles and bronze artifacts at TB display characteristics of the Eurasian Steppe cultures (Fig. 1d). The discovery of bronze drills and curved-backed bronze knives suggest these originated from the Okunevo culture (2500–1900 bc) of south Siberia, while the short bronze sword, sun-dried mud brick and the solid wooden wheel were possibly from the Sintashta-Petrovka culture (2100–1800 bc) of the South Ural region [4]. As presented in Fig. 1e, some of the earliest gold artifacts in Xinjiang, a pair of earrings, were also found in tomb M325 (individual not studied) and these are representative of the type recovered from the steppe areas of Eurasia. Physical anthropology and ancient DNA analysis support the genetic mixing of peoples from both the East and West at TB [67,69–71]. Examination of the morphological indexes from the cranium of 32 TB individuals found the population had some physical characteristics of both Mongolians and Europeans [62,67,70]. Mitochondrial DNA and Y-STR analysis was also conducted on 29 TB individuals [69] and 10 different haplogroups were detected, among which eight originate from East Eurasia, while the other two haplogroups (U and W) are considered West Eurasian lineages and are widely distributed in Siberia, Europe as well as in central and south-west Asia in the Neolithic.

MATERIALS AND METHODS

Collagen was isolated from 110 adult humans and one sheep/goat at the Key Laboratory of Vertebrate Evolution and Human Origins of Chinese Academy of Sciences, Institute of Vertebrate Palaeontology and Palaeoanthropology, Chinese Academy of Sciences using the protocol outlined in Richards and Hedges [72]. Collagen was measured at the Environmental Stable Isotope Laboratory (ESIL), Institute of Environment and Sustainable Development of Agriculture, Chinese Academy of Agricultural Sciences and at the Archaeological Stable Isotope Laboratory (ASIL), the Department of Archaeology and Anthropology at the University of the Chinese Academy of Sciences. The mass spectrometers were both IsoPrime 100 IRMS coupled with Elementar Vario. The isotopic results analyse the ratio of the heavier isotope to the lighter isotope (13C/12C or 15N/14N) and are reported as ‘δ’ in parts per 1000 or ‘per mil’ (‰) relative to the internationally defined standards for carbon (Vienna Pee Dee Belemnite, VPDB) and nitrogen (Ambient Inhalable Reservoir, AIR). The standards were Sulfanilamide, IAEA-600, IEAE-N-1, IAEA-N-2, IAEA-CH-6, USGS-24, USGS 40 and USGS 41 and, for every 10 samples, a collagen lab standard (δ13C value of 14.7 ± 0.2‰ and δ15N value of 6.9 ± 0.2‰) was also inserted in the run for isotopic calibration. The measurement errors were less than ±0.2‰ for both δ13C and δ15N values.

RESULTS

The sheep/goat and adult human sample information and isotopic results are presented in Table S1 (available as Supplementary Data at NSR online) and Fig. 2. All specimens had excellent preservation and produced good-quality collagen with C:N between 2.9 and 3.6 [73]. The sheep/goat has a δ13C value of –17.8‰ and a δ15N value of 8.7‰, which reflects a mainly C3 diet. The adult human δ13C results show a wide range of values (~10‰), from –17.6‰ to –7.2‰ (mean ± SD; –15.5 ± 1.2‰). In addition, the δ15N results are high and have a large range of values (~6‰) from 10.3‰ to 16.7‰ (mean ± SD; 14.7 ± 0.8‰). Radiocarbon dates on bone collagen from four individuals show that TB was in use from at least 1940 to 1215 cal bc.

Figure 2.

Figure 2

Human and sheep/goat δ13C and δ15N values from Tianshanbeilu. Ages of four individuals directly radiocarbon dated are also shown. See Table S1 (available as Supplementary Data at NSR online) for additional details.

DISCUSSION

Diet and radiocarbon dating at TB

In the northern latitudes of Eurasia, the two Asian millets (foxtail and common) are the only indigenous C4 crops likely to have been consumed in large quantities by ancient populations, and thus are responsible for the 13C-enriched isotopic signatures in humans and animals [9,34,74,75]. This makes δ13C values particularly sensitive markers with which to construct the Isotopic Millet Road, which can be used to detect the consumption and spread of millets from north China across the Eurasian Steppes and into Europe [34,39,47,76–80]. In this context, the isotopic results show that the TB humans consumed a significant amount of millets in their diets. Due to a lack of faunal isotope data, it is difficult to determine the specific type of animal protein consumed. However, the large number of sheep/goat remains excavated at TB [81] can be used to suggest these were a significant source of protein. In addition, wild game and freshwater fish were certainly eaten, as these foods were important components to human diets in Central Asia [9,82,83]. It is also noted that the high δ15N values of the TB humans (14.7‰) were also likely influenced by the arid environmental conditions of the Xinjiang region (see below).

Based on archaeological evidence, TB was occupied between 2000 and 1300 bc [24,67,70]. However, to determine a more precise chronology for the dietary patterns, four individuals with diverse isotopic values were directly radiocarbon dated (Table S1 (available as Supplementary Data at NSR online) and Fig. 2). The range of radiocarbon ages (1940–1215 cal bc) agrees with the archaeological evidence [84]. Interestingly, individual M599, with an exclusive C4 diet (millet), was also the oldest (1940–1765 cal bc). As this dietary pattern is characteristic of populations from the Gansu Corridor and the Yellow River Valleys, this discovery is possible evidence that there was an early movement of people from these regions to the TB area, which mirrors the archaeological and genetic evidence of extensive East–West interactions at this site (Archaeological Background).

The Isotopic Millet Road

Here, the stable isotope results of human remains (from studies with at least three individuals) from north China and bordering regions, Central Asia and Europe are examined (Table S2 (available as Supplementary Data at NSR online) and Figs 36). Our focus is mainly the second millennium bc, but, as many sites have no radiocarbon dates directly paired with isotopic results and since many sites also have large ranges of occupation, some late-third-millennium bc sites are also included. However, sites which mainly date to the first millennium bc or later were excluded. In addition, the δ15N results are plotted in relation to the modern mean annual precipitation (mm/yr) and the general climate zones of the individual sites (Table S3 (available as Supplementary Data at NSR online) and Fig. 4b) to visualize potential isotopic correlations with environmental conditions, such as aridity [85,86]. We acknowledge that the climate/precipitation changed considerably in the last 4000 years [87], but believe that these modern comparisons still provide valuable information with which to understand the human δ15N values, in addition to dietary intake. Here, the objectives are to examine the significance of the TB isotopic results in relation to other archaeological sites from the same general period. This permits the visualization of where and when there was significant dietary contact and exchange between the cultures of the Gansu Corridor/Yellow River Valleys and the Eurasian Steppe by using the δ13C signatures as proxies for human millet consumption.

Figure 3.

Figure 3

Map showing the spread of human millet consumption into Xinjiang and Central Asia along the Gansu and Inner Asian Mountain Corridors from the Yellow River Valleys during the second millennium bc.

Figure 6.

Figure 6

Map showing the Isotopic Millet Road for the radiation of human millet consumption from the Yellow River Valleys of China to Europe.

Figure 4.

Figure 4

Mean ± SD human δ13C (a) and δ15N (b) results from all Bronze Age sites surveyed along the Isotopic Millet Road. For detailed results and references, see Tables S2 and S3 (available as Supplementary Data at NSR online). Elevated δ15N values at western Gansu Corridor and Xinjiang sites are likely influenced by the increased aridity of the Taklamakan Desert and modern mean annual precipitation and general climate zones are shown in (b). Note: measurements > –12‰ reflect a predominately C4 diet, values between –12‰ and –18‰ indicate a mixed C3/C4 diet and values < –18‰ reflect a predominately C3 diet in China [39,76].

Yellow River Valleys and the Gansu Corridor

The humans from the Yellow River Valleys have significantly elevated δ13C results (~–6‰ to –11‰; Fig. 4a). This illustrates how all diets at this period across north China were heavily dominated by millets and were relatively homogenous at the population level, although large differences in isotope values between individuals existed [40,44,75,88–101]. The δ15N values are between ~6‰ and 10‰, and again there is significant variation, which accounts for the high standard deviations (Fig. 4b). The modern climate is variable at these sites, and the modern mean precipitation ranges between ~350 and 700 mm/yr. In the Gansu Corridor, numerous populations also have 13C-enriched results similar to the Yellow River Valleys, but exceptions exist, with mixed C3/C4 diets occurring at Mogou, Lajigai, Shangsunjia, Ganguai and Huoshaogou. The δ15N values range between ~8‰ and 10‰, except for the far western populations of Ganguai and Huoshaogou, which show elevated results (~11‰ to 12‰). In addition, traversing from east to west along the Gansu Corridor, the modern mean precipitation steadily declines and reaches a value of 58 mm/yr near Huoshaogou.

As can be seen in Figs 5 and S1 (available as Supplementary Data at NSR online), the Yellow River Valley sites display a large isotopic range and have significant overlap with the Gansu Corridor sites: Qijiaping, Xiahaishi, Mozuizi and Wuba, but are different compared to Mogou, Lajigai, Shangsunjia, Ganguai and Huoshaogou. Further, there are many significant outliers suggesting the possibility that interconnectivity in terms of human migration was likely occurring between these regions. For example, at the most eastern site of Mogou, the δ13C results indicate that there was less millet consumption by the population. This site has a long period of occupation (1750–1100 cal bc) and the radiocarbon dates and mixed C3/C4 isotopic results are similar to TB [44]. The three outliers from Wadian plot within the main cluster of individuals from Mogou, while two individuals from Mogou have similar results to Xinzhai and the three outliers from Qijiaping (Fig. 5). At Liuzhuang, a significant outlier possibly originated from Huoshaogou or TB and, at Nancheng, a single individual is observed to plot near the Mogou group. The Wuba site also has an individual that plots with the Mogou data. Further, Mogou has seven individuals with elevated δ13C and δ15N values that directly cluster with data from Huoshaogou and a single individual outlier that plots with individuals from the sites of the Minusinsk Basin of Russia (Fig. 5). This wide diversity of isotopic results, in conjunction with the finding of starch grains of wheat, barley and millet in the dental calculus of three skeletons from Mogou [102], suggests that there was significant East–West contact occurring at the far eastern end of the Gansu Corridor during the Bronze Age [44]. Further, some individuals were buried with Andronovo-style bronze earrings [103], indicating that at least resource transfer and possible migration of individuals from the Eurasian Steppe (e.g. Sintashta-Petrovka culture) was occurring at Mogou.

Figure 5.

Figure 5

Scatter plotting the isotopic results of >1000 individuals from 53 Bronze Age sites from across Eurasia.

In contrast, Huoshaogou is the most western Bronze Age site of the Gansu Corridor and is located at the edge of the Taklamakan Desert. It shows some isotopic overlap with Mogou (~800 km to the south-east), as seven individuals found at Mogou plot with the Huoshaogou data. In addition, three individuals found at Huoshaogou have carbon and nitrogen isotopic similarities to individuals from the Yellow River Valley sites, suggesting that there was communication and migration between these locations (Fig. 5) [44]. Huoshaogou offers compelling evidence (in the form of radiocarbon dates directly paired with human δ13C values and wheat grains) for not only East–West contact, but also when this human interaction was initiated. Liu et al. [44] found that, before ~2000 cal bc, the diet at Huoshaogou was nearly exclusively based on millets, as seen at sites in the Yellow River Valleys. However, by ~1800 cal bc, the human δ13C signatures reflecting C4 plant consumption were diminished and, by ~1600–1300 cal bc, individuals consumed a mixed C3/C4 diet, which is in agreement with the TB results. Further, these isotopic findings at Huoshaogou are also supported by the direct radiocarbon dates of wheat grains, with the results demonstrating that wheat was present between 1880 and 1620 cal bc [104]. This collective evidence offers strong support that the cultures of what is now modern-day Gansu Province and the Eurasian Steppe (e.g. Sintashta-Petrovka, Andronovo) were in direct contact by at least 1800 bc (which coincides with the dates of the earliest Tarim mummies [1,30]). That these interactions intensified in terms of diet by approximately 1600 bc is relevant to Chinese Bronze Age archaeology, as this time coincides with the establishment and rise of the Shang Dynasty (~1600–1046 bc [1,36]).

Xinjiang

In Xinjiang, this trend of decreasing δ13C values with westerly longitude continues and there is no longer evidence that the populations were consuming diets predominately based on millet (Fig. 4a). Interestingly, the δ15N values are elevated (~12‰ to 15‰; Fig. 4b), and all sites are located in and around the Taklamakan Desert, which is one of the driest places in China, and is characterized by little rainfall and high evapotranspiration rates in excess of ~3000 mm/yr [105–107]. The hyper-arid climate and unique geology of the region result in high soil salinity, and these conditions are excellent for the preservation of organic material such as the Tarim mummies and their associated artifacts [1,30]. Past isotopic research has found correlations between elevated δ15N results and environmental factors such as aridity and salinity of the soil [85,86]. Thus, the 15N-enriched results of Ganguai, Huoshaogou and the Xinjiang sites are likely influenced by the arid environmental conditions of the Taklamakan Desert (Fig. 4b), in addition to the specific dietary habits of the populations at each site.

The Xinjiang data display robust isotopic evidence for Bronze Age human migration between the cultures of the Eurasian Steppes and the Gansu Corridor and Yellow River Valleys, and confirm this region acted as a bridge between the East and West. This is prominently illustrated at TB, where individual M599 (1940–1765 cal bc) is isotopically indistinguishable from some Gansu Corridor and the Yellow River Valley individuals (Fig. 5), and some of the TB humans plot directly with the south Kazakhstan individuals of Oli-Dzhailau. At TB, the wide range of δ13C values indicates that mixed C3/C4 diets from both the East and the West were consumed (Figs 4a and 5). This site is located ~400 km from Huoshaogou, and is the first major oasis encountered after crossing of the eastern edge of the Taklamakan Desert, which served as a formidable natural geographic barrier for exchanges between the East and West. However, based on the archaeological, anthropological, genetic and dietary evidence (Archaeological Background), a significant population of Di-qiang individuals were living and forming relationships with individuals from the Eurasian Steppe cultures in terms of family, trade and diet since at least 1900 bc, and possibly much earlier [24,69,71].

Yanghai has a late period of occupation compared to the other sites presented (~1200–800 bc), but was included in this survey as the early age range overlaps with the end of the second millennium bc. Like Ganguai, Huoshaogou and TB, the human δ13C results at Yanghai also indicate that the diet was a mix of C3/C4 resources (Fig. 4a) [65]. Yanghai has a wide range of δ13C and δ15N results that plot with the TB, Karasuk and north Kazakhstan sites (Fig. 5). Thus, the sites of Ganguai, Huoshaogou, TB and Yanghai were likely important incubators for cultural interactions and acted as gateways during the Bronze Age for the movement of people, goods and technologies, both into Central Asia via the Inner Asian Mountain Corridor and into the Yellow River Valleys via the Gansu Corridor.

One of the oldest Bronze Age sites in Xinjiang with substantial archaeological and anthropological evidence related to the Andronovo culture of the Eurasian Steppe is Gumugou (a.k.a. Qäwrighul) [63,64,108,109], and the isotopic results strongly support these links. Located in the Lop Nor region of the eastern part of the Tarim Basin (Fig. 3), the δ13C and δ15N results of Gumugou are identical to north Kazakhstan sites such as Akimbek, Kyzylkol and Tashik (Figs 5 and S1 (available as Supplementary Data at NSR online)). A distance of ~1500 km separates Gumugou from these north Kazakhstan sites, but there is a nearly direct water route of contact from the shores of Lake Balkhash by following the fertile Ili River Valley to the source of the Gongnaisi River in the Tianshan Mountains. From there, it is possible to connect with the Kaidu River and traverse down the southern side of the Tianshan Mountains to Lake Bosten, and then follow the Kongque River to the ancient Lop Nor Lake (dry since the 1960s) (Fig. 7). In support of this possible route, many inhabitants of Gumugou and the closely related site of Xiaohe were found buried with oars, fishing nets and in boat-shaped coffins, indicating that they were accustomed to life around bodies of water [29,31]. However, Andronovo individuals from north Kazakhstan sites may have also traveled to Gumugou and Xiaohe from the north-eastern shores of Lake Balkhash (Fig. 7). This route would have taken them through the Alataw pass and along the northern ridge of the Tianshan Mountains to the modern capital of Xinjiang (Urumqi) and into the Turpan Basin. From here, individuals could have either migrated east to TB or west to Lake Bosten and Lop Nor Lake, and eventually to Gumugou and Xiaohe, and more research (Sr and δ34S measurements) is planned to better understand these possible routes of human movement. Millet was not found at Gumugou, and this is supported by the predominately C3 human isotopic signatures of the site (Fig. 4a) [63,64]. However, wheat was recovered and a single grain was recently radiocarbon dated to 1890–1750 cal bc [109], which is similar to the results from Huoshaogou [104]. As well as small-scale farming of wheat, the population depended on animal husbandry, hunting and possibly fishing, which, in addition to the arid climate, would also account for the high δ15N values and this agrees with the isotopic results (Fig. 4b) [63,64,108].

Figure 7.

Figure 7

Map showing the possible routes that connected the Lop Nor region and Lake Balkhash. Note the carbon and nitrogen isotopic similarities of Gumugou and north Kazakhstan individuals in Figs 5 and S1 (available as Supplementary Data at NSR online).

Xiabandi is located at the far western end of the Tarim Basin and is situated along both sides of the Taxkorgan River (Fig. 3). Based on archaeological and radiocarbon evidence, the site dates to ~1600 bc [110]. The site has strong affiliations with the Andronovo culture, and silver and bronze earrings, identical in design to Eurasian Steppe burials, were found in some graves as well as the remains of goats. No archaeobotanical remains were recovered, but the isotopic results indicate that the population consumed a predominately C3 diet [111]. However, six individuals consumed some millet and were possible migrants to the community [111]. The majority of Xiabandi human isotope values overlap with the data of individuals found at sites from north Kazakhstan and the Minusinsk Basin, and the most extreme outlier is similar to the Karasuk population (Fig. 5). Additional research is needed combining isotopic results, direct human radiocarbon dates and ancient DNA to understand the chronology of these dietary patterns in this region, especially at sites along the southern and western portions of the Tarim Basin.

South Siberia, Russia

The Minusinsk Basin of Russia is located ~1300 km north-west of TB and has a long period of human occupation during the Bronze Age, with sites from the Okunevo (2500–1900 bc), Andronovo (1900–1500 bc) and Karasuk (1500–900 bc) periods. The mean human δ15N values from the Minusinsk Basin sites are ~11‰, and the climate is continental and warm/humid with a modern mean precipitation of 341 mm/yr (Fig. 4b). Svyatko et al. [9,112] used radiocarbon dates directly paired with δ13C values to determine the chronology of when millet first appeared in the human diets of this area. The Okunevo/Andronovo individuals showed no evidence that C4 plants were consumed in significant proportions (Fig. 4a). In addition, the Okunevo/Andronovo data of the Minusinsk Basin clusters with the north Kazakhstan sites, especially the western Novoilinovka, Lisakovsk and Bestamak (Figs 35 and S1 (available as Supplementary Data at NSR online)). This is not surprising given the temporal and cultural similarities between these regions. However, after ~1500 bc, with the onset of the Karasuk culture, there is a rapid change in the δ13C values—evidence for a sharp shift to a diet that was a mix of C3/C4 crops, wheat and millets.

Svyatko et al. [9,112] proposed that these millets spread to the Minusinsk Basin from contacts with northern China and, interestingly, the δ13C and δ15N results of the Karasuk individuals are nearly identical to those from Ganguai, and also isotopically similar to TB, Yanghai, Huoshaogou and Mogou individuals (Fig. 5). The TB results support the work of Svyatko et al. [9], since they are isotopically similar in δ13C and directly overlap with the same period when the mixed C3/C4 diets first start to appear at the Karasuk sites. Given the large amount of archaeological material from the Altai Mountains, as well as the anthropological and genetic evidence of admixture between individuals of Siberian and Di-qiang origin [69], it is likely the TB and the Karasuk sites were interacting with each other on a significant scale, despite the distance and barriers of geography (Fig. 1c–e, Archaeological Background). Thus, these results suggest the existence of a Bronze Age ‘Siberian Express’ or a long-distance network of migration and resource transfer that stretched over 2000 km from the Minusinsk Basin of Russia to at least the eastern end of the Gansu Corridor at Mogou 2000 years before the ‘Silk Road’ (Fig. 6 and S2 (available as Supplementary Data at NSR online)). That these previously established links rapidly strengthen in the Xinjiang region during the middle portion of the second millennium bc is likely linked to technological advances in horse riding, chariots and metal working, which increased the mobility and migration capabilities of the cultures of the Eurasian Steppe [16,23,113,114]. In China, this period (~1600 bc) is also highly significant, as it marks the establishment of the Shang Dynasty [115]. The observation that these events chronologically coincide has not escaped the notice of past researchers, and there is speculation that increased interaction and exchange with the cultures of the Steppe (e.g. Sintashta-Petrovka, Andronovo, Karasuk, etc.) helped to stimulate the rapid rise and technological advancements of the Shang Dynasty [1,7,23,36,116].

North Kazakhstan

In contrast, all individuals from Bronze Age sites from the plains of north Kazakhstan (Kyzylkol, Akimbek, Aschisu, Tashik, Dariya, Tegiszhol, Bestamak, Lisakovsk, Novoilinovka) have low δ13C values (–19.0‰ to –18.0‰), indicating little or no significant human millet consumption, with diets predominately based on C3 crops (Figs 3 and 4a) [13,117,118]. Thus, it appears that these sites were not significantly interacting with the populations of the Yellow River Valleys of China in terms of millet agriculture. However, exceptions exist and some individuals (e.g. Bestamak, B3540; Lisakovsk, L3137, L3155) had δ13C values > –18‰ and δ15N values > 13‰ and an outlier from Tegiszhol overlaps with three individuals from TB (Fig. 5). Ventresca Miller et al. [117] suggested that the Bestamak and Lisakovsk individuals might have consumed more fish or C4 plants in relation to the other members of these sites, so there is the possibility of some migrants from Xinjiang in these populations (note the isotopic similarities with Gumugou, Fig. 5 and S1 (available as Supplementary Data at NSR online)) and future genetic research is needed. In north Kazakhstan, the δ15N values are elevated between ~11‰ to 14‰ but, compared to south Siberia (~340–420 mm/yr), there is a general decrease in modern mean precipitation (~310–320 mm/yr), and the climate is warm, humid and continental. As with Xinjiang, these high human δ15N values were likely influenced by the environmental factors as well as the specific diets of these populations (Fig. 4b), and future isotopic analysis of archaeological plant remains will provide more information on the baseline isotopic values in this region.

South Kazakhstan and Central Asia

The south Kazakhstan sites of Kyzyl Bulak and Oi-Dzhailau display a large range of isotopic results and are located along the northern slope of the Tianshan Mountains (Fig. 3), approximately 2000 km from TB [13]. Like TB, the inhabitants at both of these sites had similar δ13C values, indicative of mixed C3/C4 diets and significant millet consumption (Fig. 4a) [13]. The human δ15N values at Kyzyl Bulak (10.7‰) and Oi-Dzhailau (13.8‰) show large differences even though they have similar climates (hot humid continental). While the modern mean rainfall is somewhat different between these sites (Fig. 4b), the isotopic difference (3.1‰) represents approximately one trophic level and was more likely influenced by dietary preferences than environmental conditions. Three Kyzyl Bulak individuals are similar to the isotopic values from the Minusinsk Basin, and the fourth plots near TB and Oi-Dzhailau (Fig. 5). In contrast, the Oi-Dzhailau data overlap with TB, but two individuals cluster with the north Kazakhstan populations, and this is possible evidence that migration and exchange occurred between the sites of north and south Kazakhstan as well as TB, but genetic analysis is necessary to confirm this. In addition, Kyzyl Bulak and Oi-Dzhailau date to the same general time period (1750–1400 cal bc) as TB (Fig. S2, available as Supplementary Data at NSR online). This is strong evidence in support of the Inner Asian Mountain Corridor as an early main conduit for the spread of millet (as well as cultural and technological exchanges and migration) from the Yellow River Valleys via Xinjiang to south Kazakhstan (Figs 6 and S2 (available as Supplementary Data at NSR online)) [6].

Continuing west along the Inner Asian Mountain Corridor, directly dated millet grains have also been found during the same timeframe at Ojakly (1740–1610 cal bc) in eastern Turkmenistan [6,11,35,119] (Fig. 6). This finding, as well as the discovery of small bundles of ephedra (Ephedra sp.) with the Tarim mummies at sites such as Gumugou and Xiaohe, suggests interactions with the Bactria–Margiana Archaeological Complex (2300–1700 bc) of the Iranian Plateau [1,23,30]. In addition, archaeological finds of millet have been reported from Shortughai in northern Afghanistan [6,120,121] and at many other sites in Iran, Iraq, Syria and Turkey during the second millennium bc, but these grains were not directly radiocarbon dated [8]. Thus, we hypothesize that future isotopic studies of Bronze Age human populations in Kyrgyzstan, Tajikistan, Uzbekistan, Turkmenistan, Afghanistan, Pakistan, India, Iran, Iraq and Turkey will uncover evidence of mixed C3/C4 human diets, and thus directly extend the Isotopic Millet Road from its current terminus in southern Kazakhstan (Fig. 6).

Europe

In Bronze Age Europe, human bone collagen δ13C values that could be interpreted as evidence of millet consumption are found at sites in Ukraine (Glubokoe Ozero II), northern Greece (Rymnio) and in northern (Olmo di Nogara) and central Italy (Grotta Misa, Felcetone) [122–126] (Fig. 6). Of these locations, the Italian site of Olmo di Nogara [107] currently displays the most convincing evidence that a European population consumed millet in similar amounts to the individuals at Oi-Dzhailau and TB (Figs 4a and 5). The Olmo di Nogara individuals are believed to date (no radiocarbon dates were performed) to the same general period (~1600–1200 bc) as the sites in south Kazakhstan, Siberia and China, and it is possible that these sites might have had links in terms of human connections such as migration and resource transfer. This could have been facilitated by a communication network maintained between the Srubnaya and Andronovo cultures of the western and eastern Eurasian Steppes, respectively [23], and future research (radiocarbon dating, ancient DNA and archaeobotanical analysis) at Olmo di Nogara is needed.

Millet was also recovered at many other Bronze Age sites in Europe (Fig. 6) [32,42,127–129] but, as few of these grains were directly radiocarbon dated, the inferred dates based on associated materials are less secure [42]. However, the direct radiocarbon dating of millet grains revealed ages (1610–1260 cal bc) at the sites of Fajsz18 (Hungry), Bruchenbrücken/Friedberg (Germany) and Măgura-Buduiasca (Romania) that overlap with the dates from the sites of south Kazakhstan, Minusinsk Basin and Xinjiang (Figs 6 and S2 (available as Supplementary Data at NSR online)) [42]. Further, a millet grain was directly dated at Assiros in Greece (1395–1285 cal bc) [130] and Valamoti [127] reported an earlier radiocarbon date of 2200–2030 cal bc for broomcorn millet grains from the Greek site of Skala Sotiros. In addition, millet was recently directly radiocarbon dated at Qasim Bagh (1495-1305 cal BC) in India [131] and Guamsky Grot (1110-910 cal BC) in Russia [132]. While these results are intriguing, additional isotopic, genetic and archaeological studies of Bronze Age sites (from eastern/southern Europe, Central Asia and Russia) are required to confirm that this Isotopic Millet Road stretched directly from the Yellow River Valleys of China to Europe and to understand what this form of communication/interaction may have meant to these distant populations.

CONCLUSIONS

As a period characterized by dynamic East–West migrations, the second millennium bc of Eurasia witnessed frequent exchanges of crops, animals, tools, as well as ideas among different populations. Accompanying these interactions, extensive changes in daily life and social structures occurred in the early civilizations across the Eurasian continent, including both ancient China and the Mediterranean. In particular, China experienced a rapid increase in human population and settlements from the late Neolithic to the Bronze Age [7,115], and this set the stage for the ‘revolutionary’ development of the Chinese civilization and the formation of early states in China. It is in this context and during this critical transitional period that millet agriculture/consumption started to significantly expand westward out of the Yellow River Valleys and appear in Xinjiang, south Kazakhstan, Minusinsk Basin and Europe by the Isotopic Millet Road. We suggest that increased East–West contact, in the form of resource transfer, warfare, marriage, migration and the rise of the Xia and Shang Dynasties, was a major driver for this Bronze Age radiation of millets. Thus, the isotopic results presented here, combined with the archaeology, physical anthropology and genetic evidence, indicate that sites in Xinjiang (such as TB) held unique positions between the Bronze Age cultures of the East and West with the direct mixing of many aspects of daily life, from family to foods.

Supplementary Material

nwx015_Supplement_File

ACKNOWLEDGMENTS

We are grateful to Xianglong Chen, Yating Qu, Ying Nie, Huiyun Rao and Qiuyu Wang for assistance with this work.

FUNDING

This work was supported by the National Basic Research Program of China (2015CB953803), the Chinese Academy of Sciences International Visiting Scholar Fellowship (2016VBC002), the National Natural Science Foundation of China (41373018) and the National Natural Science Foundation of China Research Fund for International Young Scientists (41550110224).

REFERENCES

  • 1. Mallory JP and Mair VH. The Tarim Mummies: Ancient China and the Mystery of the Earliest Peoples from the West. London: Thames and Hudson, 2000. [Google Scholar]
  • 2. Li Y. The transport routes in Xinjiang before the building of the Silk Road. Caoyuan Wenwu 2001; 1: 57–64. [Google Scholar]
  • 3. Li S. Interaction between Northwest China and Central Asia during the second millennium bc: an archaeological perspective. In: Boyle K, Renfrew C and Levine M (eds). Ancient Interactions: East and West in Eurasian Steppe. Cambridge: McDonald Institute Monographs, University of Cambridge, 2002, 171–80. [Google Scholar]
  • 4. Lin M. Origin and migration of Tocharian. Xiyu Yanjiu 2003; 3: 9–23. [Google Scholar]
  • 5. Sherratt AG. The trans-Eurasian exchange: the prehistory of Chinese relations with the West. In: Mair V (ed.). Contact and Exchange in the Ancient World. Honolulu: Hawaii University Press, 2006, 32–53. [Google Scholar]
  • 6. Frachetti MD. Multiregional emergence of mobile pastoralism and nonuniform institutional complexity across eurasia. Curr Anthropol 2012; 53: 2–38. [Google Scholar]
  • 7. Jaang L. The landscape of China's participation in the Bronze Age Eurasian Network. J World Prehist 2015; 28: 179–213. [Google Scholar]
  • 8. Stevens CJ, Murphy C and Roberts Ret al. . Between China and South Asia: a middle Asian corridor of crop dispersal and agricultural innovation in the Bronze Age. Holocene 2016; 26: 1541–55. [DOI] [PMC free article] [PubMed] [Google Scholar]
  • 9. Svyatko SV, Schulting RJ and Mallory Jet al. . Stable isotope dietary analysis of prehistoric populations from the Minusinsk Basin, Southern Siberia, Russia: a new chronological framework for the introduction of millet to the eastern Eurasian steppe. J Archaeol Sci 2013; 40: 3936–45. [Google Scholar]
  • 10. Spengler RN III, Frachetti MD and Domani PN. Late Bronze Age agriculture at Tasbas in the Dzhungar Mountains of eastern Kazakhstan. Quatern Int 2014; 348: 147–57. [Google Scholar]
  • 11. Spengler RN III, Cerasetti B and Tengberg Met al. . Agriculturalists and pastoralists: Bronze Age economy of the Murghab alluvial fan, southern Central Asia. Veg Hist Archaeobot 2014; 23: 805–20. [Google Scholar]
  • 12. Spengler RN III, Frachetti M and Doumani Pet al. . Early agriculture and crop transmission among Bronze Age mobile pastoralists of Central Eurasia. Proc R Soc Biol Sci Ser B 2014; 281: 20133382. [DOI] [PMC free article] [PubMed] [Google Scholar]
  • 13. Motuzaite-Matuzeviciute G, Lightfoot E and O’Connell TCet al. . The extent of cereal cultivation among the Bronze Age to Turkic period societies of Kazakhstan determined using stable isotope analysis of bone collagen. J Archaeol Sci 2015; 59: 23–34. [Google Scholar]
  • 14. Li C, Li H and Cui Yet al. . Evidence that a West-East admixed population lived in the Tarim Basin as early as the early Bronze Age. BMC Biol 2010; 8: 15. [DOI] [PMC free article] [PubMed] [Google Scholar]
  • 15. Li C, Ning C and Hagelberg Eet al. . Analysis of ancient human mitochondrial DNA from the Xiaohe cemetery: insights into prehistoric population movements in the Tarim Basin, China. BMC Genet 2015; 16: 78. [DOI] [PMC free article] [PubMed] [Google Scholar]
  • 16. Allentoft ME, Sikora M and Sjogren K-Get al. . Population genomics of Bronze Age Eurasia. Nature 2015; 522: 167–72. [DOI] [PubMed] [Google Scholar]
  • 17. Haak W, Lazaridis I and Patterson Net al. . Massive migration from the steppe was a source for Indo-European languages in Europe. Nature 2015; 522: 207–11. [DOI] [PMC free article] [PubMed] [Google Scholar]
  • 18. Flad R, Li SC and Wu XHet al. . Early wheat in China: results from new studies at Donghuishan in the Hexi Corridor. Holocene 2010; 20: 955–65. [Google Scholar]
  • 19. Jones MK, Hunt H and Lightfoot Eet al. . Food globalization in prehistory. World Archaeol 2011; 43: 665–75. [Google Scholar]
  • 20. Betts A, Jia PW and Dodson J. The origins of wheat in China and potential pathways for its introduction: a review. Quatern Int 2014; 348: 158–68. [Google Scholar]
  • 21. Barton L and An C. An evaluation of competing hypotheses for the early adoption of wheat in East Asia. World Archaeol 2014; 46: 775–98. [Google Scholar]
  • 22. Mei J, Liu G and Chang X. Preliminary analysis and study on early Bronzes unearthed from Eastern Xinjiang. Xiyu Yanjiu 2002; 1–10. [Google Scholar]
  • 23. Anthony DW. The Horse, The Wheel, and Language: How Bronze-Age Riders from the Eurasian Steppe Shaped the Modern World. Princeton: Princeton University Press, 2007. [Google Scholar]
  • 24. Shao H. The Development of the Pre-Historic Cultures in Xinjiang and the Interaction with Neighbor Cultures. Changchun: Jilin University, 2007. [Google Scholar]
  • 25. Yi M. Excavation of the ancient cemetery site, Ke’ermuchi, Xinjiang. Wenwu 1981; 23–32. [Google Scholar]
  • 26. Jia PW and Betts A. A re-analysis of the Qiemu’erqieke (Shamirshak) cemeteries, Xinjiang, China. J Indo-European Stud 2010; 38: 1–43. [Google Scholar]
  • 27. Wang M. On chronology of the Ke’ermuchi culture and Ke’ermuchi cemetery. Xiyu Yanjiu 2013; 69–80+139. [Google Scholar]
  • 28. Zhang Y and Abuduresule Y. Beifang Cemetery: thousands year old mystery buried in the hinterland of desert. Xinjiang Renkou Dilixue 2009; 68–75. [Google Scholar]
  • 29. Yang R, Yang Y and Li Wet al. . Investigation of cereal remains at the Xiaohe Cemetery in Xinjiang, China. J Archaeol Sci 2014; 49: 42–7. [Google Scholar]
  • 30. Barber EW. The Mummies of ürümchi. London: Pan Books, 1999. [Google Scholar]
  • 31. Mair VH. The rediscovery and complete excavation of Ordek's Becropolos. J Indo-European Stud 2006; 34: 273–318. [Google Scholar]
  • 32. Hunt HV, Linden MV and Liu Xet al. . Millets across Eurasia: chronology and context of early records of the genera Panicum and Setaria from archaeological sites in the Old World. Veg Hist Archaeobot 2008; 17: S5–18. [DOI] [PMC free article] [PubMed] [Google Scholar]
  • 33. Boivin N, Fuller DQ and Crowther A. Old World globalization and the Columbian exchange: comparison and contrast. World Archaeol 2012; 44: 452–69. [Google Scholar]
  • 34. Lightfoot E, Liu X and Jones MK. Why move starchy cereals? A review of the isotopic evidence for prehistoric millet consumption across Eurasia. World Archaeol 2013; 45: 574–623. [Google Scholar]
  • 35. Spengler RN III. Agriculture in the Central Asian Bronze Age. J World Prehist 2015; 28: 215–53. [Google Scholar]
  • 36. An Z. Archaeological research on Neolithic China. Curr Anthropol 1988; 29: 753–9. [Google Scholar]
  • 37. Zhao Z. Palaeobotanical studies on the origin of agriculture and civilization. Shehui Kexue Guanli yu Pinglun 2005; 82–91. [Google Scholar]
  • 38. Lee GA, Crawford GW and Liu Let al. . Plants and people from the Early Neolithic to Shang periods in North China. Proc Natl Acad Sci USA 2007; 104: 1087–92. [DOI] [PMC free article] [PubMed] [Google Scholar]
  • 39. Barton L, Newsome SD and Chen Fet al. . Agricultural origins and the isotopic identity of domestication in northern China. Proc Natl Acad Sci USA 2009; 106: 5523–8. [DOI] [PMC free article] [PubMed] [Google Scholar]
  • 40. Liu X, Jones MK and Zhao Zet al. . The earliest evidence of millet as a staple crop: new light on neolithic foodways in North China. Am J Phys Anthropol 2012; 149: 283–90. [DOI] [PubMed] [Google Scholar]
  • 41. He H. A study on the origin and spread of foxtail millet and broomcorn millet in view of globalization. Zhongguo Nongshi 2014; 16–25. [Google Scholar]
  • 42. Motuzaite-Matuzeviciute G, Staff RA and Hunt HVet al. . The early chronology of broomcorn millet (Panicum miliaceum) in Europe. Antiquity 2013; 87: 1073–85. [Google Scholar]
  • 43. Frachetti MD, Spengler RN and Fritz GJet al. . Earliest direct evidence for broomcorn millet and wheat in the central Eurasian steppe region. Antiquity 2010; 84: 993–1010. [Google Scholar]
  • 44. Liu X, Lightfoot E and O’Connell TCet al. . From necessity to choice: dietary revolutions in west China in the second millennium BC. World Archaeol 2014; 46: 661–80. [Google Scholar]
  • 45. DeNiro MJ and Epstein S. Influence of diet on the distribution of nitrogen isotopes in animals. Geochim Cosmochim Acta 1981; 45: 341–51. [Google Scholar]
  • 46. Richards MP, Hedges REM and Molleson TIet al. . Stable isotope analysis reveals variations in human diet at the Poundbury Camp Cemetery site. J Archaeol Sci 1998; 25: 1247–52. [Google Scholar]
  • 47. Hu Y, Wang S and Luan Fet al. . Stable isotope analysis of humans from Xiaojingshan site: implications for understanding the origin of millet agriculture in China. J Archaeol Sci 2008; 35: 2960–5. [Google Scholar]
  • 48. Farquhar GD, Ehleringer JR and Hubick KT. Carbon isotope discrimination and photosynthesis. Annu Rev Plant Physiol Plant Mol Biol 1989; 40: 503–37. [Google Scholar]
  • 49. Tieszen LL. Natural variations in the carbon isotope values of plants: implications for archaeology, ecology, and paleoecology. J Archaeol Sci 1991; 18: 227–48. [Google Scholar]
  • 50. Lee-Thorp JA. On isotopes and old bones. Archaeometry 2008; 50: 925–50. [Google Scholar]
  • 51. An C, Dong W and Li Het al. . Variability of the stable carbon isotope ratio in modern and archaeological millets: evidence from northern China. J Archaeol Sci 2015; 53: 316–22. [Google Scholar]
  • 52. Schwarcz HP. Some theoretical aspects of isotope paleodiet studies. J Archaeol Sci 1991; 18: 261–75. [Google Scholar]
  • 53. O’Leary MH. Carbon isotope fractionation in plants. Phytochemistry 1981; 20: 553–67. [Google Scholar]
  • 54. DeNiro MJ and Epstein S. Influence of diet on the distribution of carbon isotopes in animals. Geochim Cosmochim Acta 1978; 42: 495–506. [Google Scholar]
  • 55. van der Merwe NJ and Vogel JC. 13C content of human collagen as a measure of prehistoric diet in woodland North America. Nature 1978; 276: 815–6. [DOI] [PubMed] [Google Scholar]
  • 56. Lee-Thorp JA, Sealy JC and van der Merwe NJ. Stable carbon isotope ratio differences between bone collagen and bone apatite, and their relationship to diet. J Archaeol Sci 1989; 16: 585–99. [Google Scholar]
  • 57. Ambrose SH and Norr L. Isotopic composition of dietary protein and energy versus bone collagen and apatite: purified diet growth experiments. In: Lambert JB and Grupe G (eds). Prehistoric Human Bone. Berlin: Springer, 1993, 1–37. [Google Scholar]
  • 58. Tieszen L and Fagre T. Effect of diet quality and composition on the isotopic composition of respiratory CO2, bone collagen, bioapatite, and soft tissues. In: Lambert J and Grupe G (eds). Prehistoric Human Bone. Berlin Heidelberg: Springer, 1993, 121–55. [Google Scholar]
  • 59. Jim S, Jones V and Ambrose SHet al. . Quantifying dietary macronutrient sources of carbon for bone collagen biosynthesis using natural abundance stable carbon isotope analysis. Br J Nutr 2006; 95: 1055–62. [DOI] [PubMed] [Google Scholar]
  • 60. Hedges REM. On bone collagen-apatite-carbonate isotopic relationships. Int J Osteoarchaeol 2003; 13: 66–79. [Google Scholar]
  • 61. Pearson JA, Buitenhuis H and Hedges REMet al. . New light on early caprine herding strategies from isotope analysis: a case study from Neolithic Anatolia. J Archaeol Sci 2007; 34: 2170–9. [Google Scholar]
  • 62. Wei D and Shao H. Ancient demography study on Hami Tianshan Beilu cemetery. Bianjiang Kaogu Yanjiu 2012; 463–70. [Google Scholar]
  • 63. Zhang Q and Zhu H. Carbon and nitrogen stable isotope analysis of the human bones from the Gumugou cemetery: a preliminary exploration of the diet of the eary population in Lop Nur. Xiyu Yanjiu 2011; 91–6 .+  142. [Google Scholar]
  • 64. Qu Y, Yang Y and Hu Yet al. . The extraction and the C, N isotope analysis of hair keratin from Gumugou cemetery in Xinjiang. Geochimica 2013; 447–53. [Google Scholar]
  • 65. Si Y, Lu E and Li Xet al. . Exploration of human diets and populations from the Yanghai Tombs, Xinjiang. Chin Sci Bull (Chinese Version) 2013; 1422–9. [Google Scholar]
  • 66. Zhang Q, Chang X and Liu G. Stable isotopic analysis of human bones unearthed from Tianshan Beilu Cemetery in Hami, Xinjiang. Xiyu Yanjiu 2010; 38–43 .+  122-3. [Google Scholar]
  • 67. Wei D. Human Biological Variation and Population Affinity during Bronze-Iron Age in Xinjiang Hami Region. Changchun: Jilin University, 2009. [Google Scholar]
  • 68. Li S. Research of Siba culture. In: Binqi S (ed.). Kaoguxue Wenhua Lunji, Vol. 3. Beijing: Cultural Relics Press, 1993, 80–121. [Google Scholar]
  • 69. Gao S, Zhang Y and Wei Det al. . Ancient DNA reveals a migration of the ancient Di-Qiang populations into Xinjiang as early as the Early Bronze Age. Am J Phys Anthropol 2015; 157: 71–80. [DOI] [PubMed] [Google Scholar]
  • 70. Wei D, Zhao Y and Chang Xet al. . Craniometric variation of ancient skulls from the Hami Tianshan North Road Cemetery. Acta Anthropol Sin 2012; 395–406. [Google Scholar]
  • 71. Li H. Y-Chromosome Genetic Diversity of the Ancient Northern Chinese populations. Changchun: Jilin University, 2012. [Google Scholar]
  • 72. Richards MP and Hedges REM. Stable isotope evidence for similarities in the types of marine foods used by late mesolithic humans at sites along the Atlantic Coast of Europe. J Archaeol Sci 1999; 26: 717–22. [Google Scholar]
  • 73. DeNiro MJ. Postmortem preservation and alteration of in vivo bone collagen isotope ratios in relation to palaeodietary reconstruction. Nature 1985; 317: 806–9. [Google Scholar]
  • 74. Wang R. Photosynthetic pathway and morphological functional types in the steppe vegetation from Inner Mongolia, North China. Photosynthetica 2003; 41: 143–50. [Google Scholar]
  • 75. Ma M, Dong G and Lightfoot Eet al. . Stable isotope analysis of human and faunal remains in the western Loess plateau, approximately 2000 cal BC. Archaeometry 2014; 56: 237–55. [Google Scholar]
  • 76. Pechenkina EA, Ambrose SH and Ma Xet al. . Reconstructing northern Chinese Neolithic subsistence practices by isotopic analysis. J Archaeol Sci 2005; 32: 1176–89. [Google Scholar]
  • 77. Fu Q, Jin S and Hu Yet al. . Agricultural development and human diets in Gouwan site, Xichuan, Henan. Chin Sci Bull 2010; 55: 614–20. [Google Scholar]
  • 78. Lanehart RE, Tykot RH and Underhill APet al. . Dietary adaptation during the Longshan period in China: stable isotope analyses at Liangchengzhen (southeastern Shandong). J Archaeol Sci 2011; 38: 2171–81. [Google Scholar]
  • 79. Atahan P, Dodson J and Li Xet al. . Subsistence and the isotopic signature of herding in the Bronze Age Hexi Corridor, NW Gansu, China. J Archaeol Sci 2011; 38: 1747–53. [Google Scholar]
  • 80. Guo Y, Hu Y and Gao Qet al. . Stable carbon and nitrogen isotope evidence in human diets based on evidence from the Jiangzhai Site, China. Acta Anthropol Sin 2011; 30: 149–57. [Google Scholar]
  • 81. Xue W. Research of Animal Burial from Prehistoric Xinjiang. Beijing: Central University of Nationalities, 2011. [Google Scholar]
  • 82. O’Connell TC, Levine MA and Hedges REM. The importance of fish in the diet of central Eurasian peoples from the Mesolithic to the Early Iron Age. In: Levine MA, Renfrew AC and Boyle K (eds). Prehistoric Steppe Adaptation and the Horse. Cambridge: McDonald Institute for Archaeological Research, 2003, 253–68. [Google Scholar]
  • 83. Murphy EM, Schulting R and Beer Net al. . Iron Age pastoral nomadism and agriculture in the eastern Eurasian steppe: implications from dental palaeopathology and stable carbon and nitrogen isotopes. J Archaeol Sci 2013; 40: 2547–60. [Google Scholar]
  • 84. Reimer PJ, Bard E and Bayliss Aet al. . IntCal13 and Marine13 radiocarbon age calibration curves 0–50,000 years cal BP. Radiocarbon 2013; 55: 1869–87. [Google Scholar]
  • 85. Schwarcz HP, Dupras TL and Fairgrieve SI. 15N enrichment in the Sahara: in search of a global relationship. J Archaeol Sci 1999; 26: 629–36. [Google Scholar]
  • 86. Szpak P. Complexities of nitrogen isotope biogeochemistry in plant-soil systems: implications for the study of ancient agricultural and animal management practices. Front Plant Sci 2014; 5: 288. [DOI] [PMC free article] [PubMed] [Google Scholar]
  • 87. Tang Z, Chen D and Wu Xet al. . Redistribution of prehistoric Tarim people in response to climate change. Quatern Int 2013; 308: 36–41. [Google Scholar]
  • 88. Zhang X, Wang J and Xian Zet al. . Dietary research on ancient humans. Kaogu 2003; 2: 158–71. [Google Scholar]
  • 89. Wu X, Xiao H and Wei Cet al. . Isotopic evidence on the dietary structure, angricultural stratygies and animal domestication of humans and pigs from Xinzhai Site, Henan Province. Science for Archaeology, 2nd edn. Beijing: Science Press, 2007. [Google Scholar]
  • 90. Zhang X, Qiu S and Bu Get al. . Stable carbon and nitrogen isotope analysis on human bones from Erlitou site and Taosi site. In: Technology, Mosa (ed.). Origin of the Chinese Civilization Project. Beijing: Science Press, 2009, 286–95. [Google Scholar]
  • 91. Ling X, Chen L and Xue Xet al. . Stable isotopic analysis of human bones from the Qingliang Temple graveyard, Ruicheng Country, Shanxi Province. Disiji Yanjiu 2010; 30: 415–21. [Google Scholar]
  • 92. Zhang X, Qiu S and Zhong Jet al. . Stable carbon and nitrogen isotope analysis of human bones from Qianzhangda cemetery from Tengzhou, Shandong. Kaogu 2012; 9: 83–96. [Google Scholar]
  • 93. Hou L, Hu Y and Zhao Xet al. . Human subsistence strategy at Liuzhuang site, Henan, China during the proto-Shang culture (~2000–1600 BC) by stable isotopic analysis. J Archaeol Sci 2013; 40: 2344–51. [Google Scholar]
  • 94. Atahan P, Dodson J and Li Xet al. . Temporal trends in millet consumption in northern China. J Archaeol Sci 2014; 50: 171–7. [Google Scholar]
  • 95. Wang Y, Nan P and Wang Xet al. . Dietary differences in humans with similar social hierarchies: example from the Niedian Site, Shanxi. Acta Anthropol Sin 2014; 1: 82–9. [Google Scholar]
  • 96. Ma M, Dong G and Liu Xet al. . Stable isotope analysis of human and animal remains at the Qijiaping Site in Middle Gansu, China. Int J Osteoarchaeol 2015; 25: 923–34. [Google Scholar]
  • 97. Chen X, Guo X and Hu Yet al. . Dietary recontruction on human bones from Muzhuzhuliang site, shenmu, Shanxi. Kaogu yu Wenwu 2015; 5: 112–7. [Google Scholar]
  • 98. Cheung C, Jing Z and Tang Jet al. . Examining social and cultural differentiation in early Bronze Age China using stable isotope analysis and mortuary patterning of human remains at Xin’anzhuang, Yinxu. Archaeol Anthropol Sci 2017; 9: 799–816. [Google Scholar]
  • 99. Chen X, Fang Y and Hu Yet al. . Isotopic reconstruction of the Late Longshan Period (ca. 4200–3900 BP) dietary complexity before the onset of state-level societies at the Wadian Site in the Ying River Valley, Central Plains, China. Int J Osteoarchaeol 2016; 26: 808–17. [Google Scholar]
  • 100. Ma Y, Fuller BT and Wei Det al. . Isotopic perspectives (δ13C, δ15N, δ34S) of diet, social complexity, and animal husbandry during the proto-shang period (ca. 2000--1600 BC) of China. Am J Phys Anthropol 2016; 160: 433–45. [DOI] [PubMed] [Google Scholar]
  • 101. Ma M, Dong G and Jia Xet al. . Dietary shift after 3600 cal yr BP and its influencing factors in northwestern China: evidence from stable isotopes. Quat Sci Rev 2016; 145: 57–70. [Google Scholar]
  • 102. Li M, Yang X and Wang Het al. . Starch grains from dental calculus reveal ancient plant foodstuffs at Chenqimogou site, Gansu Province. Sci China Earth Sci 2010; 53: 694–9. [Google Scholar]
  • 103. Xie Y, Qian Y and Mao Ret al. . Mougou Cemetery of Lintan, Gansu. Kaogu 2009; 7: 10–7. [Google Scholar]
  • 104. Dodson JR, Li X and Zhou Xet al. . Origin and spread of wheat in China. Quat Sci Rev 2013; 72: 108–11. [Google Scholar]
  • 105. Ren M, Yang R and Bao H. An Outline of China's Physical Geography. Beijing: Foreign Languages Press, 1985. [Google Scholar]
  • 106. Liu D, Li Z and Jiang Set al. . Physical Geology of Asia. Beijing: Commercial Press, 1996. [Google Scholar]
  • 107. Luo C, Peng Z and Yang Det al. . Research on the environmental evolution of Lop Nur in Xinjiang, China. Ziran Zazhi 2006; 28: 37–41. [Google Scholar]
  • 108. Wang B. An general introduction to the agricultural archaeolgical study of Xinjiang. Nongye Kaogu 1983; 1: 102–7 +  2. [Google Scholar]
  • 109. Zhang G, Wang S and Ferguson DKet al. . Ancient plant use and plaeoenvironmental analysis at the Gumugou Cemetery, Xinjiang, China: implications for dessicated plant remains. Archaeol Anthropol Sci 2017; 9: 145–52. [Google Scholar]
  • 110. Xinjiang Institute of Cultural Relics and Archaeology. Xiabandi Cemetery in Xinjiang. Beijing: Cultural Relics Press, 2012. [Google Scholar]
  • 111. Zhang X, Wei D and Wu Yet al. . Carbon and nitrogen stable isotope analysis of human bones from the Xiabandi Cemetery in Xinjiang, China: implications to the cultural interaction between the east and the west. Chin Sci Bull 2016; 61: 3509–19. [Google Scholar]
  • 112. Svyatko SV, Mallory JP and Murphy EMet al. . New radiocarbon dates and a review of the chronology of the prehistoric populations from the Minusinsk Basin, Southern Siberia, Russia. Radiocarbon 2009; 51: 243–73. [Google Scholar]
  • 113. Kuzmina E. On the origin of the Indo-Iranians. Curr Anthropol 2002; 43: 303–4. [Google Scholar]
  • 114. Legrand S. The emergence of the Karasuk culture. Antiquity 2006; 80: 843–59. [Google Scholar]
  • 115. Thorp RL. China in the Early Bronze Age: Shang Civilization. Philadelphia: University of Pennsylvania Press, 2005. [Google Scholar]
  • 116. Motuzaite-Matuzeviciute G, Kiryushin YF and Rakhimzhanova SZet al. . Climatic or dietary change? Stable isotope analysis of Neolithic–Bronze Age populations from the Upper Ob and Tobol River basins. The Holocene 2016; 26: 1711–21. [Google Scholar]
  • 117. Ventresca Miller A, Usmanova E and Logvin Vet al. . Subsistence and social change in central Eurasia: stable isotope analysis of populations spanning the Bronze Age transition. J Archaeol Sci 2014; 42: 525–38. [Google Scholar]
  • 118. Lightfoot E, Motuzaite-Matuzeviciute G and O’Connell TCet al. . How ‘pastoral’ is pastoralism? Dietary diversity in Bronze Age communities in the central Kazakhstan Steppes. Archaeometry 2014; 57: 232–49. [Google Scholar]
  • 119. Rouse LM and Cerasetti B. Ojakly: a Late Bronze Age mobile pastoralist site in the Murghab Region, Turkmenistan. J Field Archaeol 2014; 39: 32–50. [Google Scholar]
  • 120. Askarov A. La Bactriane á l’aube de la civilization. Les dossiers d’Archéologie 1993; 185: 60–9. [Google Scholar]
  • 121. Spengler RN III and Willcox G. Archaeobotanical results from Sarazm, Tajikistan, an Early Bronze Age settlement on the edge: agriculture and exchange. Environ Archaeol 2013; 18: 211–21. [Google Scholar]
  • 122. Triantaphyllou S. A Bioarchaeological Approach to Prehistoric Cemetery Populations from Central and Western Greek Macedonia. British Archaeological Reports, 2001. [Google Scholar]
  • 123. Privat KL. Palaeoeconomy of the Eurasian Steppe: Biomolecular Studies. Oxford: University of Oxford, 2004. [Google Scholar]
  • 124. Tafuri MA, Craig OE and Canci A. Stable isotope evidence for the consumption of millet and other plants in Bronze Age Italy. Am J Phys Anthropol 2009; 139: 146–53. [DOI] [PubMed] [Google Scholar]
  • 125. Triantaphyllou S. A stable isotope analysis of skeletal assemblages from Greek Macedonia. In: Papathanasiou ARM and Fox S (eds). Were They What They Ate? Dietary Reconstruction in Greece from Stable Isotope Analyses. Athens: Occasional Wiener Laboratory Series, 2014, 113–53. [Google Scholar]
  • 126. Varalli A, Moggi-Cecchi J and Moroni Aet al. . Dietary variability during Bronze Age in central Italy: first results. Int J Osteoarchaeol 2016; 26: 431–46. [Google Scholar]
  • 127. Valamoti SM. Millet, the late comer: on the tracks of Panicum miliaceum in prehistoric Greece. Archaeol Anthropol Sci 2016; 8: 51–63. [Google Scholar]
  • 128. Gyulai F. The history of broomcorn millet (Panicum miliaceum L.) in the Carpathian-Basin in the mirror of archaeobotanical remains. I. From the beginnings until the Roman Age, Columella. J Agr Environ Sci 2014; 1: 29–38. [Google Scholar]
  • 129. Moreno Larrazabal A, Teira Brión A and Sopelana Salcedo Iet al. . Ethnobotany of millet cultivation in the north of the Iberian Peninsula. Veg Hist Archaeobot 2015; 24: 541–54. [Google Scholar]
  • 130. Wardle K, Higham T and Kromer B. Dating the end of the Greek Bronze Age: a robust radiocarbon-based chronology from Assiros Toumba. PLoS One 2014; 9: e106672. [DOI] [PMC free article] [PubMed] [Google Scholar]
  • 131. Spate M, Zhang G and Yatoo Met al. . New Evidence for early 4th millennium BP agriculture in the Western Himalays: Qasim Bagh, Kashmir. J Archaeol Sci: Rep 2017; 11: 568–77. [Google Scholar]
  • 132. Trifonov VA, Shishlina NI and Lebedeva EYet al. . Directly dated broomcorn millet from the northwestern Caucasus: tracing the Late Bronze Age route into the Russian steppe. J Archaeol Sci: Rep 2017; 12: 288–94. [Google Scholar]

Associated Data

This section collects any data citations, data availability statements, or supplementary materials included in this article.

Supplementary Materials

nwx015_Supplement_File

Articles from National Science Review are provided here courtesy of Oxford University Press

RESOURCES