Ancient plant use at the site of Yuergou, Xinjiang, China: implications from desiccated and charred plant remains
- First Online:
- Cite this article as:
- Jiang, H., Wu, Y., Wang, H. et al. Veget Hist Archaeobot (2013) 22: 129. doi:10.1007/s00334-012-0365-z
- 382 Views
Archaeobotanical studies were undertaken at the Yuergou site, which is located in the Turpan basin in Xinjiang, China, and which has been dated to around 2300–2400 years b.p. Altogether 21 taxa were identified. Four cereal remains were identified, Triticum aestivum, Hordeum vulgare var. coeleste, Panicum miliaceum and Setaria italica. The first three were probably cultivated while the last one may not have been grown deliberately, but probably grew together with plants of P. miliaceum. A fruit stone of Ziziphus jujuba (Chinese date) was discovered, which showed that this may have been cultivated around the site during that time. Charcoal of Picea sp. was found, from wood which must have been used as fuel by the indigenous people. Fifteen taxa of wild plants were also identified, most of which can be considered as weeds, and which grew near the site. Burs of Xanthium strumarium were discovered. As nearly all of them were broken, the seeds may have been used by the ancient inhabitants. Since most of the cereal remains consisted of chaff, they must represent by-products. Furthermore, grains of Echinochloa crus-galli may also have been exploited as complementary food resources. All the above indicate that both cultivated and wild plants were used for cereals, fuel, or other purposes, and plant resources played important roles in the daily life of the ancient inhabitants of the Yuergou site.
KeywordsAncient agricultureCentral AsiaCerealsMilletWood charcoalCrop processing
The ancient civilization around the Turpan basin of Xinjiang can be traced almost as far back as 10,000 years ago. Chipped stone implements such as scrapers and choppers were discovered at the Gouxi site at Turpan. Furthermore, stone flakes, internal casts, stone spheres as well as microblades were unearthed at the Astana site, which is nearly 5,000 years old (Wang 1992). Later, the Turpan basin became a paradise for the ancient Gushi people from 3000 to 1900 b.p., as this and the adjacent areas fell under the influence of the Subeixi Culture (Lü and Zheng 2002). According to the sparse information to be found in the ancient written record, these fair-haired people had a nomadic lifestyle, but they also cultivated cereals as a food supplement.
The first systemic archaeobotanical study was conducted on the macroremains of seeds/fruits, leaves, wood, etc. which were excavated in a Subeixi Culture cemetery named Yanghai Tombs. It showed that the ancient people of Turpan knew how to take advantage of plant and animal resources, especially the former. Cultivated plants like Hordeum vulgare var. coeleste (syn. Hordeum vulgare var. nudum) (naked barley) and Panicum miliaceum (common millet) were used as cereals (Jiang et al. 2007b), while Vitis vinifera (grape) was cultivated for its fruit (Jiang et al. 2009). Some wild plants, such as Cannabis sativa (hemp) and Lithospermum officinale, were exploited for medicinal/psychoactive and decorative purposes respectively (Jiang et al. 2006, 2007a). Based on microremains of phytoliths and starch, a palaeodiet study of the Subeixi site indicates that noodles and cakes were made from P. miliaceum (Gong et al. 2011). However, in order to obtain more information and gain a clearer picture of the Subeixi Culture, it became necessary to study one more related site or tomb. Fortunately the newly excavated Yuergou site supplied more materials, and also provided an opportunity to improve our understanding of the ancient Gushi people of Turpan, including their agriculture, horticulture, crop processing and use of the indigenous wild plants.
The Yuergou District was once a crossroad in ancient Xinjiang. The Alagou valley was one of the main routes of the ancient Silk Road. During the 1970s, the Alagou–Yuergou tombs, belonging to the Subeixi Culture, were discovered. Nearly 80 tombs were excavated and many artefacts of great significance were unearthed (Wang 1981). In 2008 the Yuergou site was discovered during the construction of the Turpan–Kuqa multiple track railway. A salvage excavation was carried out by a working group from the Xinjiang Institute of Archaeology, under the direction of Professor Yong Wu, one of the authors of the present paper.
Materials and methods
The plant remains were collected from the earthenware jar in F1a (Figs. 3c, 4). The jar is 60 cm in diameter, 30 cm in remaining height. The bottom of the jar was broken and replaced by muddy material (Fig. 3d). The plant remains, mixed with a little stone, earth, bones, etc., were collected by dry sieving with 0.5 mm mesh size. Subsequently plant remains of wood, stems and dung of sheep/goat were picked out by hand, while seeds/fruits and spikelets were extracted using finely pointed tweezers and later photographed under a stereomicroscope. Transverse, radial, and tangential sections of the charred wood were prepared with a razor, and then mounted on a stub for SEM studies. The plant remains are now deposited in the Department of Scientific History and Archaeometry, Graduate University of Chinese Academy of Sciences in Beijing.
Two specimens were selected for age determination. One was a piece of charred wood, selected from the plant remains inside the earthenware jar; the other was a stem of an annual herb, extracted from the plant remains close to the bottom of the pot. The specimens were radiocarbon dated with an accelerator mass spectrometer (AMS) at Peking University, and then calibrated using IntCal04 (Reimer et al. 2004) and OxCal v3.10 (Bronk Ramsey 2005).
Plant nomenclature follows the revised English edition of the Flora of China (http://www.efloras.org/flora_page.aspx?flora_id=2). The botanical terms used in the description of cereal remains and crop processing come from Jacomet (2006) and Van der Veen and Jones (2006). The botanical terms used for structural descriptions of the secondary xylem are based on the relevant references, Fahn et al. (1986), Schweingruber (1990), Lin and Hu (2000) and IAWA (2004).
Radiocarbon dating of the ancient plant remains
14C age (years b.p.)
Calibrated age, 1σ-range (years b.c.)
Calibrated age, 2σ-range (years b.c.)
Stem of an annual herb
2295 ± 50
410–350 (40.3 %)
290–230 (27.9 %)
510–460 (2.1 %)
420–200 (93.35 %)
2460 ± 35
750–680 (23.2 %)
600–500 (27.5 %)
760–680 (25.2 %)
670–410 (70.2 %)
Plant remains found at the Yuergou site
Hordeum vulgare var. coeleste
Description of the cereal, fruit, and wood remains
Setaria italica—11 spikelets were found inside the earthenware pot. Spikelet elliptical, 2.1 mm long, 1.2 mm wide and 1.1 mm thick on average. Husk (lemma and palea) brown in colour, lemma smooth with shallow grooves, palea papillate (Fig. 5h).
Triticum aestivum—9 rachis segments, 9 awns, 10 lemmas, 13 paleas, as well as 12 glumes, were unearthed from the earthenware pot. The plant remains of T. aestivum display its natural colour, and the rachis segments show typical bread wheat structure (Fig. 5b, e).
Hordeum vulgare var. coeleste—7 grains and 323 rachis segments were found inside the earthenware pot. Grains brown in colour, 5.8–6.1 mm long, 2.8–3.2 mm wide, and 2.1–2.2 mm thick; rachis segments compressed, triplet; glumes lanceolate (Fig. 5c, d).
Ziziphus jujuba—Only one stone was discovered inside the earthenware pot. The stone is black in colour, 10.5 mm long, 3.5 mm in diameter. Both the apex and the base are tapered. The stone is partly broken, but still displaying a fusiform shape. No seed was observed inside the locule (Fig. 5f).
Picea sp.—Transverse section: boundaries of growth rings distinct, transition from early to late wood gradual, vertical resin canals present (Fig. 6e). Tangential section: rays uniseriate to multiseriate (fusiform); uniseriate rays 1–17 cells high, while multiseriate rays with horizontal resin canals (Fig. 6c). Radial section: bordered pits usually in one row, rarely biseriate (Fig. 6e); cross-field pitting piceoid to taxodioid, 1–6 per cross-field (Fig. 6d).
Agricultural and horticultural remains
In all, four cereal taxa were discovered, Triticum aestivum, Hordeum vulgare var. coeleste, Panicum miliaceum and Setaria italica. The former two are typically of western derivation, while the latter two were first cultivated in Asia. According to previous studies, there have been agricultural activities in Xinjiang since ca. 4000 b.p. For instance, caryopses of T. aestivum were discovered in the Gumugou tombs (around 3,800 years b.p) (Wang 1983; Yan and Yang 1999), and later spikelets of both P. miliaceum and caryopses of T. aestivum were excavated in the coeval Xiaohe cemetery (Li et al. 2011). During the period of the Subeixi Culture (1000 b.c. to 100 a.d.), agriculture was apparently well-developed as many cereal remains have been excavated. For instance, plant remains of P. miliaceum, T. aestivum as well as H. vulgare var. coeleste were excavated from the Yanghai tombs (Jiang et al. 2007b), while cakes and noodles made from P. miliaceum were excavated at the Subeixi site (Gong et al. 2011). In addition, dental caries was a very common complaint among the deceased of the coeval Wulabo tombs. This pathological condition shows that starchy food may have played an important role in their daily diet (Wang 1992). In conclusion, agricultural activities became an indispensable part of daily life for the people belonging to the period of the Subeixi Culture, although most of them still led a nomadic lifestyle.
Of the four cereals, the former three were probably cultivated around the Yuergou site, as many spikelets of P. miliaceum, chaff of T. aestivum, H. vulgare var. coeleste and P. miliaceum, grains of H. vulgare var. coeleste, as well as leaf sheaths of P. miliaceum were discovered together. Compared to the other three taxa, which are known in considerable numbers, plant remains of S. italica seem negligible as only three spikelets (grains with husk) were found. P. miliaceum and S. italica were the two most important cereals in the ancient Eurasian steppe, especially in East Asia. Both plants have similar edaphic requirements and life cycles, and are usually cultivated in the same area even nowadays. In this connection it comes as no surprise that these two cereals are usually found together in archaeological sites (Renfrew 1973; Marinval 1992; Hunt et al. 2008). However, the archaeological sites in Xinjiang seem to be an exception. For instance, no S. italica was discovered in the Xiaohe cemeteries (around 3700–3500 years b.p.) (Li et al. 2011), the Gumugou cemeteries (around 3800 years b.p.) (Wang 1983), or at the Xintala site (around 3800 years b.p.) (Zhao et al. 2012). In the Wupu cemeteries (nearly 3000 years b.p.) of Hami District, some cake, thought to be made from caryopses of S. italica, was found (Wang 1983). However, it is impossible to tell the ingredients by the naked eye. Without starch and phytolith analyses, the result is at best equivocal.
Plants of P. miliaceum are more drought resistant than those of S. italica, their evaporation rate being lower than that of S. italica, Sorghum bicolor, H. vulgare and T. aestivum, etc. (Dong and Zheng 2006). It is more easily cultivated in Turpan, where the precipitation is only 15 mm, while evaporation is 2,500 mm (Jiang et al. 2009). To date, there is still no physical evidence of S. italica in Turpan before the period of the Subeixi Culture. With ancient trade, war and human migration, the value of S. italica was eventually recognized by the ancient people living in Turpan. In the Astana cemeteries (around 3rd to 9th century A.D.), large amounts of S. italica were found, together with P. miliaceum, T. aestivum, Cannabis sativa and H. vulgare var. coeleste, etc. As the majority of the deceased in the Astana cemeteries were not indigenous, but had migrated from the eastern part of ancient China, and since there was a tradition for cultivating S. italica among the people of eastern China for countless millennia, it seems plausible that S. italica cultivation was introduced into Turpan, along with the migration of people from eastern China.
Apart from the cereal remains, a fruit stone of Ziziphus jujuba Mill. (Chinese date) was also found. Z. jujuba is a small, deciduous fruit tree belonging to the Rhamnaceae, which has been cultivated since the Neolithic period in northern China (Qu and Wang 1993). It contains many vitamins, sugars, and mineral substances, etc. and most cultivars can be eaten both fresh and dehydrated. Following its natural distribution, plant remains of Z. jujuba were mainly discovered in eastern China, such as Hebei and Henan provinces, and it was seldom found in northwestern China, especially Xinjiang. However, dried fruit and kernels were occasionally discovered in the Astana tombs in the Turpan Basin (Wang 1983). The discovery of Z. jujuba in the present study shows that it may have been cultivated in Turpan as early as 2,500 years ago. Fruit of Z. jujuba not only provided delicious food for the indigenous people, but also enriched their daily life.
Remains of trees
Among the charcoal examined, only Picea was identified. Picea is now an important forest resource in China, as it has straight timber of good quality, and is usually selected as construction material, as well as for making paper, instruments, etc. In Turpan P. schrenkiana, a common gymnosperm, is widely distributed on the northern slopes of the Tianshan (Heavenly Mountains) (Cheng et al. 1992). Carbonized wood of Picea was once found in the pits of Qin Terracotta Warriors and Horses, and was considered to have been used as pillars, crossbeams, as well as for timber (Wang et al. 2009). Desiccated wood of Picea was found in the Yanghai tombs, having been used for cemetery construction (Jiang 2006). In the present study, all the charred wood was found together with the other plant remains which were desiccated, in the same earthenware pot, so one cannot help thinking that these woods were explicitly selected as firewood by the ancient Yuergou inhabitants instead of being charred unintentionally.
Previous studies showed that two main firewood collection strategies can generally be distinguished: collection based on either availability or selection (Out 2010). Wood of Picea contains resin, and is fit for burning. It also has a good combustion value like other gymnosperms. Although the possibility of natural distribution of Picea was rather limited around the Yuergou site during that period, the wood may have been collected during hunting or grazing activities. Selection was involved, as only Picea was identified among the charcoal. The ancient people may have recognized the good combustion value of Picea, which provided them with light, warmth, as well as safety.
Wild herb/shrub remains
Phragmites australis and Typha sp. are two hygrophilous plants. As the Yuergou site is located on a floodplain, the ancient people could have obtained both of them near the Aiweiergou River. Both plants have long, flexible fibres, making them good candidates for weaving. The ancient people living in Xinjiang have known their uses for quite a long time. In the Gumugou tombs, nearly 3800 years b.p., thread made of Typha sp. was discovered (Jiang Hongen, unpublished data). In the Yanghai tombs, both P. australis and Typha sp. were used for mat making, while the former was also used for thread making (Jiang 2006). As only florets of P. australis and leaf fragments of Typha sp. have been discovered in the present study, while no products of the above two plants were found, their exact use is still an open question.
Ephedra is a little shrub adapted to dry environments. In the present study, as only two stems but no reproductive organs were discovered, it could only be identified to the generic level. Three species, namely E. sinica Stapf, E. intermedia Schrenk ex C. A. Meyer and E. equisetina Bunge, containing more ephedrine, are considered to have medicinal value. Ephedra is a traditional medicine in China. It was used for curing asthma, coughs as well as being used as a diaphoretic (Anonymous 2005). Plants of Ephedra have been excavated in the Xiaohe cemeteries (Bergman 1935) and Gumugou cemeteries (around 3800 b.p.) in Xinjiang (Wang 2009). Some scholars are of the opinion that the plants were connected with primitive religions, and thus were accredited supernatural status and revered by the indigenous people (Bergman et al. 1939; Xia 1997). However, as there is little proof that Ephedra was exploited in the present study, its use by the ancient Yuergou people remains uncertain.
Of the wild plant remains, Xanthium strumarium (syn. X. sibiricum) (cocklebur) is of particular interest. As an annual herb X. strumarium is widely distributed in the plain, on roadsides, in wild areas, etc. With its hooked thorns, it is easily disseminated by animals and mankind and has become a synanthropic plant. X. strumarium is poisonous (List et al. 1979), but its fruit is used in traditional Chinese medicine for curing nasal congestion, wheals, spasms, etc. (Han 2005). Although seeds of X. strumarium are poisonous they can be used for diuresis and diaphoresis. To date, burs of X. strumarium were discovered in two Chinese Neolithic sites, namely Zaojiaoshu site in Henan (LWA 2002), and Chengtoushan site in Hunan (Liu and Gu 2007). In both cases they were considered to be weeds taken to the site by the ancient indigenous people. The species has also been discovered in an early Iron Age well at Hajndl, Slovenia. However, only a few involucra were collected and the fruits were still intact (Sostarić et al. 2009). In the present study the fruits were found in great numbers (138 pieces), and all of them were broken and the seeds missing. We believe that the seeds of X. strumarium may have been extracted and used by the ancient Yuergou people, although they could not be eaten raw. However, the exact purpose for which they were collected and broken requires further investigation.
Cereal processing and plant storage
Aside from ancient agriculture, the cereal remains discovered in the earthenware pot can also provide us with information on crop processing and plant storage during that period. Among the cereal remains, a total of 330 plant parts (323 rachises and seven grains) of H. vulgare var. coeleste (naked barley) were discovered. The well-developed triplets and grains indicated that they were harvested at maturity, instead of being stored in the ear. The triplets with the usually broken palea and lemma clearly indicate threshing. Of the 323 rachises, 289 were single triplets. The other 34 multiple rachis segments, which consisted of 2–4 triplets, contain 87 triplets. In total 376 triplets were discovered. Considering the fact that one triplet contains three grains, ideally 1,128 grains should have been harvested. However, only seven grains were found in the container. Similarly the nine multiple rachis segments of T. aestivum, containing 45 rachis segments, should have produced 135 wheat grains (rachis segment: grains ratio is 1:3, according to Alonso et al. 2008). However, no grains of T. aestivum were discovered. In these cases the clean grains of both the naked barley and the wheat should have been collected by winnowing or hand-sorting, and only the waste by-products would have been placed in the present container. According to Van der Veen and Jones (2006), the high ratio of free-threshing rachis internodes/grains indicates that the sample of both the naked barley as well as the wheat in the container should be by-products from an early processing stage. The seven grains of naked barley may represent accidental spillage during crop cleaning, which also occurred at the settlement mound of Tell Brak, northeast Syria (Hald and Charles 2008).
Plant remains of Panicum miliaceum (common millet) consisted of only 15 sheaths, 118 intact spikelets, as well as chaff (897/678 pieces of palea/lemma and 224 pieces of broken spikelets). This strongly suggests that the spikelets had been broken and the grains released after threshing. The grindstones excavated from the site may have been used for dehusking. As no clean grains (caryopses) were found, but a number of weed seeds were discovered, remains of P. miliaceum should be interpreted as by-products from a late processing stage according to the concept of Van der Veen and Jones (2006). The minor components of spikelets (118), which were not threshed, may represent unprocessed grain contaminants, or accidental spillage during crop processing, like those of the naked barley in the present study.
Apart from the cereal remains, there was some Echinochloa crus-galli in the mixture. Although 156 spikelets were still intact, it is significant that there were 178 broken spikelets as well as 27/47 lemmas/paleas. As with T. aestivum, no clean grains of E. crus-galli were discovered. That means plants of E. crus-galli should not be considered as weeds, but as complementary food. Plants of E. crus-galli and P. miliaceum ripen in the same season in Turpan. One possibility is that plants of E. crus-galli were present as weeds in a field of P. miliaceum, and spikelets of the former were harvested and then threshed together with those of the latter. Since spikelets of E. crus-galli can easily be separated by winnowing, it is also possible that spikelets of E. crus-galli were collected deliberately and then threshed for their clean grains. Previous studies showed that the protein, calcium and iron content of E. crus-galli is comparable to or greater than that of other major cereals like wheat or rice (Mandelbaum et al. 1995). During famines, spikelets of E. crus-galli were collected and then threshed for their grains, even as late as the 1960s in China (Ms. Jinrong Chen, personal communication 2012). In this connection the grains of E. crus-galli might also have been consumed consciously or unconsciously as a component of their diet by the ancient inhabitants of Yuergou.
Except for the wood charcoal and the 138 pieces of Xanthium strumarium, which should have no connection with crop processing, most of the plant remains consisted of small chaff and weed seeds. In this case, the cache of plant remains could be interpreted as by-products from day-to-day processing according to Van der Veen and Jones (2006). Although H. vulgare var. coeleste and T. aestivum ripen simultaneously in Turpan, they cannot be harvested at the same time as plants of P. miliaceum and Ziziphus jujuba. It seems unlikely that these plant residues represent a single storage event, but instead several storage episodes in which the pot was not completely cleaned. The scenario of sequential contamination of reused stores also took place in the tomb of Tutankhamun (De Vartavan 1990). Furthermore, the earthenware jar was not a container for clean grains, but for waste by-products, which may have been used for fuel and/or fodder.
Comparisons of tomb offerings and the plant remains from the Yuergou site
Many plant remains, well preserved due to the dry environment, have been excavated in different archaeological sites in Xinjiang. In the past several years, archaeologists have paid more attention to the plant remains which were preserved in the containers. From their descriptions the tomb offerings usually represented storage products. For example, only grains (caryopses) of T. aestivum were preserved in the grass basket of the Gumugou cemetery (Wang 1981). A cache of florets (caryopses with husk) of P. miliaceum was discovered in a jar from the Yanghai tombs, which could represent storage samples, as only a few glumes and husks of H. vulgare var. coeleste were found in the same vessel (Jiang et al. 2007b). In these cases, the tomb offerings were specifically chosen for further use in another world. Unfortunately, the number of spikelets of P. miliaceum as well as the chaff of H. vulgare var. coeleste was not quantified in the 2007 study. In order to obtain a better scientific interpretation of the cereal remains, the concept of crop processing should be adopted and applied in future work by Chinese archaeobotanists.
We are greatly indebted to the editor and the two anonymous reviewers for their critical revision of an earlier version of this manuscript. We would like to thank Enguo Lü for his help during field work in Turpan. Thanks are also due to Chengzhong Han for his many constructive suggestions. This study was supported by the program of National Natural Science Foundation of China (No. 41102114), as well as the “Strategic Priority Research Program–Climate Change: Carbon Budget and Relevant Issues’’ of the Chinese Academy of Sciences (Grant No. XDA05130501).