Plant utilization at the Jiangxigou site during the middle Holocene

ARA-00018; No of Pages 9 Archaeological Research in Asia xxx (2015) xxx–xxx

Contents lists available at ScienceDirect

Archaeological Research in Asia journal homepage: www.elsevier.com/locate/ara

Plant utilization at the Jiangxigou site during the middle Holocene GuangLiang Hou a,⁎, Zhikun Ma b,⁎, Chongyi E. a, Weng Zhang c, Haicheng Wei d a

Key Laboratory of Physical Geography and Environmental Processes of Qinghai Province, School of Life and Geographic Science, Qinghai Normal University, XiNing 810008, China School of Archaeology and Museology, Peking University, Beijing 100871, China c Department of Geological Engineering, Qinghai Normal University, XiNing 810001, China d Qinghai Institute of Salt Lakes, Chinese Academy of Sciences, XiNing 810001, China b

a r t i c l e

i n f o

Article history: Received 9 January 2016 Accepted 13 January 2016 Available online xxxx Keywords: Northeastern margin of the Qinghai–Tibetan Plateau (QTP) Starch grain Pollen Millet Yangshao culture

a b s t r a c t Agricultural activities on the Qinghai–Tibetan Plateau (QTP) are difficult due to the hostile environment. Despite the difficulties, recently early agriculture has been documented by Holocene plant remains on the QTP. The details of the time, place, and mechanisms of agricultural origin as well as the evolution of agriculture on the QTP are still lacking. The Jiangxigou site (JXG2) contains artifacts dominated by microliths from the early to middle Holocene and provides important information on agricultural origins and evolution of prehistoric culture on the QTP. Here we report ancient starch grains and pollen extracted from ceramics and deposits excavated from each layer. The starch remains include 21 grains identical to starches from millets (Setaria spp. and Panicum spp.), probably including 24% domesticated foxtail millet (Setaria italica), indicating that humans at 3000 m above sea level (masl) in the northeastern margin of the Qinghai–Tibetan Plateau (QTP) began to use millets ca. 5600 cal BP. Considering archeological evidence, the reasons for millet utilization by plateau residents may be associated with the expansion and influence of the Yangshao culture via the Loess Plateau. © 2015 Elsevier Ltd. All rights reserved.

1. Introduction The Qinghai–Tibetan Plateau (QTP), known for its extreme cold and oxygen-limited environment, is the highest plateau in the world. Although undertaking agricultural activities on the plateau is difficult, current evidence suggests that agriculture began on the northeastern QTP from the middle to late Holocene (Chen et al., 2015). Current evidence particularly documents this process on the northeastern margin of Qinghai–Tibet Plateau. In this region, we find sites of the Yangshao culture (7000–5000 cal BP), which is the earliest Neolithic culture of the QTP. Prehistoric humans in the region cultivated millet along river valleys. Their agriculture has been documented by evidence of carbonized foxtail millet (Setaria italica) and broomcorn millet (Panicum miliaceum) (e.g. Xie, 2002) at sites of the Majiayao Culture (5300– 4000 cal BP). Furthermore, results of bone isotopic analysis of ancient humans associated with the Zongri Culture (5600–4000 cal BP) in the upper Yellow River valley indicate that local residents consumed C4 plants (presumably mainly foxtail millet and broomcorn millet) as staples (Cui et al., 2006), even though we still lack direct crop evidence (Chen et al., 1998). To date, many studies have shown that agriculture first appeared on the northeastern margin and east of the QTP, and then appeared in the ⁎ Corresponding authors. E-mail addresses: [email protected] (G. Hou), [email protected] (Z. Ma).

interior part of the plateau; meanwhile edible crops increased in diversity (Chen et al., 2015; Fu, 2001; D'Alpoim Guedes et al., 2014). Some scholars think that edible crops on the QTP first included foxtail millet and broomcorn millet, and later changed to a mixture of foxtail millet, naked barley, and wheat (Jia et al., 2012). Other researchers illustrate that wheat and barley rapidly replaced millets and became the main crops around 4000 cal BP. However, details on the time, place, and mechanism of agricultural origins as well as the evolution of agriculture on the QTP are still lacking. The Jiangxigou2 (JXG2) site contains continuous, relatively complete cultural strata from the Paleolithic to the Neolithic era and therefore is different from other sites, many of which have been destroyed by strong erosion on the QTP (Brantingham and Gao, 2006). Consequently, the JXG2 site is an important site for studying plant use of prehistoric humans from the Paleolithic to the Neolithic era on the QTP. A series of microliths, ceramics and other artifacts have been unearthed at the JXG2 site. Ceramics are closely associated with cooking, eating, and other activities (Piperno et al., 2000; Yuan, 2002; Yang and Jiang, 2010; Craig et al., 2013), and the interior walls of ceramic wares often contain lipid molecules, proteins, tartaric acid, starch grains, DNA, and other residues from prehistoric use (Craig et al., 2000; Zarrillo et al., 2008). In the present study, starch grains extracted from residues on unearthed ceramics were analyzed to reveal their function and the use of plants by ancient residents of JXG2 (the following methods described in Yang et al., 2012).

http://dx.doi.org/10.1016/j.ara.2016.01.003 2352-2267/© 2015 Elsevier Ltd. All rights reserved.

Please cite this article as: Hou, G., et al., Plant utilization at the Jiangxigou site during the middle Holocene, Archaeological Research in Asia (2015), http://dx.doi.org/10.1016/j.ara.2016.01.003

2

G. Hou et al. / Archaeological Research in Asia xxx (2015) xxx–xxx

2. The Jiangxigou site The JXG2 site (36°35′25″N, 100°17′47″E) is located at 3312 masl on the southern shore of Qinghai Lake, 118 m above the lake level, and lies at the junction between the plains near Qinghai Lake and the hills leading to the Qinghai Nanshan mountains. The cultural deposits at the site are exposed on the edge of the associated terrace (Rhode et al., 2007). The profile of cultural layers at JXG2 is about 120 cm thick (Fig. 2) and contains a relatively large amount of microliths, animal bone fragments, as well as charcoal and ceramics. Stratigraphic subdivisions were conducted during fieldwork by close examination of the color, texture, and structure in the JXG2 profile. A combined pedological and stratigraphic description of the JXG2 profile is presented in Table 1. (See Fig. 1.) According to the color and particle size of soil, the profile is divided into four strata: Layer 1 (0–30 cm; pale yellow in color; with loose soil) consists of the modern soil, a relatively large amount of roots and sand, and a small number of microliths; the brightness value of the chrominance index is high and the degrees of redness and yellowness relative to the entire profile are high. Layer 2 (30–75 cm; dark gray; brown silty clay, relatively firm) contains a comparatively large number of microliths, bone fragments, and ceramics. The degrees of brightness, red and yellow in the chrominance index of this stratum are all low. The degree of redness and yellowness of sediments is related to the content of hematite and goethite respectively and are related to pedogenesis, which, in turn, is a product of climate. The brightness of soil is related to carbonate content, water content, roughness, and other factors, and reflects relative precipitation. Layer 3 (75–114 cm; gray; brown silty clay, hard and firm) contains a relatively large number of microliths and bone fragments. It is the lowest stratum in the profile that contains significant quantities of artifacts. All three aspects of chrominance have high values. Layer 4 (below 114 cm; typical loess) contains a very small number of microliths and bone fragments. Compared with the previous layers, the brightness value of the chrominance index is low, while the degrees of redness and yellowness are significantly higher.

3. Materials and methods 3.1. Materials Wearing clean disposable plastic gloves, the operator used trowels to excavate in 1.5 m × 0.5 m quadrats. Potsherds were excavated from

archeological strata at intervals of 10 cm and then washed. All necessary permits for the described field investigations were obtained by Qinghai Normal University. After wet-screening, microliths, animal bone fragments, ceramics, and charcoal were collected from the 11 layers (the 0–10 cm layer is the ground surface layer, which is disturbed by modern human activities and hence which was not studied) (Table 2). A total of 59 environmental samples (pollen, color) were collected along the profile at intervals of 2 cm, starting with 2 cm below the surface. Five pottery sherds were excavated from the profile (at depths of: 39 cm, 46 cm, 50 cm, 61 cm, and 75 cm), while another six sherds were collected from flotation samples (depths: 30–40 cm, 40–50 cm, 50–60 cm, and 60–70 cm). On the whole, ceramics were unearthed at 30–75 cm, with four pieces unearthed at 40–50 cm, hence, the ceramics are relatively concentrated in their associated depth. The deepest ceramic was an orange fragment unearthed at 75 cm. Two charcoal samples collected from the JXG2 were sent to the Accelerator Mass Spectrometry Laboratory of Peking University. Eleven sherds, whose serial numbers are JXG2:P1-JXG2:P11, were examined. Four sherds with charred residues were selected (Fig. 3 a–d). Charred residues were extracted by gently scraping the material from the interior and external surfaces onto aluminum foil squares. Then the charred residues were collected in the storeroom at the Qinghai Normal University. Eight sediment samples were analyzed as control specimens, including one from the dust of the storeroom where the artifacts were stored, and seven others from the seven cultural layers (42, 46, 48, 50, 52, 61, and 65 cm depths) where the sherds were collected. 3.2. Methods Pollen grains were extracted from the sporopollen samples collected from JXG2 by using the heavy liquid flotation method detailed elsewhere (Li et al., 2006). Detailed steps for analyzing starch grains from the residues found on the ceramics are as follows (after Yang et al., 2010, 2012): 1. Before analysis began, all experimental implements including test tubes and glass stirring rods were cleaned with an ultrasonic bath, and then boiled for 15 min to avoid any possible contamination of the archeological samples. 2. Organic matter was removed by adding H2O2 (concentration: 6%), then ultra-pure water was added. This was followed by centrifuging the sample (3500 r/m, 10 min). This process was repeated multiple times until the liquid became neutral.

Fig. 1. Outline map of Eastern Asia and the Qinghai lake Basin. a. Numbers refer to sites discussed in text 1. Dadiwan, 2. Majiayao, 3. Jiangxigou2, 4. Zongri, 5. Karuo, 6. Qugong, 7. Changguogou, 8. Nuomuhong. b. Main archeological sites at Qinghai Lake. 1. Heimahe, 2. 151 site, 3. Loula reservoir, 4. Yantaidong, 5. Bronze Wire, 6. Baifosi, 7. Shaliuhe.

Please cite this article as: Hou, G., et al., Plant utilization at the Jiangxigou site during the middle Holocene, Archaeological Research in Asia (2015), http://dx.doi.org/10.1016/j.ara.2016.01.003

G. Hou et al. / Archaeological Research in Asia xxx (2015) xxx–xxx

3

Fig. 2. JXG2 profile and the unearthed pottery.

3. Calcium was removed by adding HCl (concentration: 10%), and then ultra-pure water was added. This was followed by centrifuging the sample (3500 r/m, 10 min). After shaking, this process was repeated multiple times until the liquid became clear. 4. A 5% sodium hexametaphosphate (Calgon) was added to each sample, the mixture was shaken for 10 min, and then set in place for 12 h. Finally, the mixture was rinsed by ultrapure water. 5. After centrifugation, 1.8 g/cm3 CsCl was added to the tube before the mixture was subjected to flotation. 6. The upper layer of the mixture was transferred to a new tube. A small amount of 10% glycerol was added, and the sample was mounted using nail polish. Observations and measurements were made by using a biological microscope. Eight soil samples collected from the dust in the storeroom where the artifacts were stored, and from 42, 46, 48, 50, 52, 61, and 65 cm depths of the JXG2 profile were used to control for possible contamination. Then 6% H2O2 was added to remove organic matter and release any starch grains. Next, a 5% Calgon was added to disperse the soil and remove clay. Finally, 1.8 g/cm3 CsCl was added to extract the starch grains. In addition, in order to identify the starch grains extracted from the residues on the surface of the sherds, analysis of starch grains of modern crops in the area surrounding the QTP was performed. Data on

morphology of over 200 species of starch grain from the Poaceae, Leguminosae, Fagaceae, and other families collected from different regions of China (Yang et al., 2010, 2012, Yang and Perry, 2013) were used as reference. Fig. 4A illustrates the morphology of some of these starch grains. 4. Results 4.1. Dates We sent two charcoal samples to Peking University Accelerator Mass Spectrometry Laboratory, the value of the half-life for 14C is 5568 years, but the 13C/12C ratio is not measured. Together with dates previously published for the site, the two charcoal samples reported here show that sediments from the JXG2 profile date from the early Holocene, between over 8100 cal BP and around 2000 cal BP (Table 3). A depth–date relationship of the strata was established by using these data (Fig. 5). 4.2. Starch grain analysis on pottery sherds A total of 22 starch grains were extracted from the residues on four pottery sherds (Nos. JXG2-P7, Fig. 3a; JXG2-P5, Fig. 3b; JXG2-P6, Fig. 3c; JXG2-P9, Fig. 3d) (Table 4). One grain is difficult to distinguish

Table 1 Pedological and stratigraphic descriptions of the soil and sediment in the Holocene loess–soil profile at the JXG2 site, northeast Qinghai–Tibetan Plateau. Depth (cm)

Stratigraphic subdivisions

Pedo-stratigraphic description

0–30

Top soil

30–75

Dark gray-brown silty clay

75–114

Gray-brown silty clay

114-

Transitional loess

Pale yellow (7.5YR4/3); coarse silt; abundant plant roots, wormholes, and excrement; friable; abundant well-rounded spherical pellets (1.0–2.0 mm). dark gray-brown (2.5YR 8/3); silt; very friable; few earthworm burrows and excrement; abundant fragments of pottery, animal bones, charcoal, and microliths. Gray-brown (10YR 3/3); weakly developed soil (Ab); silt; fine white secondary carbonate concentrations; massive-blocky structure; friable; few earthworm burrows and excrement; few well-rounded spherical pellets (1.0–2.0 mm); massive fragments of pottery, animal bones, charcoal, and microliths. Pale yellow (2.5YR 8/3), sand and silt, very loose, very friable. Loose structure of homogeneous portion, no bedding, silty, average particle size of 29.08–29.18 μm, particle size b2 μm components accounted for approximately 5.03–5.36%, the aeolian deposits which belong to the typical of the loess.

Chroma values Lightness

Redness

Yellowness

31.9–38.8 (34.9)

4.9–6.0 (5.5)

10.5–13.4 (11.9)

34.1–37.3 (35.2)

4.7–5.3 (5.0)

10–11.7 (10.9)

34.3–44.3 (38.7)

4–4.9 (4.5)

9.1–11.1 (10.1)

36.4–38.7 (37.6)

4.8–5.2 (5.1)

10.9–11.7 (11.4)

The figure between brackets is average of Chroma Values.

Please cite this article as: Hou, G., et al., Plant utilization at the Jiangxigou site during the middle Holocene, Archaeological Research in Asia (2015), http://dx.doi.org/10.1016/j.ara.2016.01.003

4

G. Hou et al. / Archaeological Research in Asia xxx (2015) xxx–xxx

Table 2 Ceramics, stone tools, and animal bones across strata in JXG2 site. Stratum (cm)

Ceramics no.

Description

Microlithic tools

10–20 20–30

/ / JXG2-P1a JXG2-P2 JXG2-P3a JXG2-P4 JXG2-P5 JXG2-P6 JXG2-P7a JXG2-P8 JXG2-P9 JXG2-P10a JXG2-P11a / / / /

/ / Clay terracotta with black attachments on the surface White stoneware, cord mark Orange coarse pottery, attachments on four walls Orange coarse pottery, attachments on four walls orange coarse pottery, with attachments Coarse black pottery, with attachments Orange clay fine pottery, attachments on the rims Orange clay fine pottery, with attachments Orange coarse pottery, attachments at the bottom Orange clay ceramics Orange clay ceramics, attachments on the exterior wall / / / /

5 11

22 9

11

18

15

36

26

100

35

403

20 50 9 10 6

416 310 195 140 44

30–40

40–50

50–60 60–70 70–80 80–90 90–100 100–110 110–120 a

Animal bones

Denotes samples unearthed in situ; the remainder were recovered from the sieve.

and thus could not be identified (Fig. 4B.e). The remaining 21 starch grains were mainly polyhedral in shape with centric hilum, and transverse or Y-shaped fissures through the hila (Fig. 4B.a, b). After rotation, the grains still appeared as polyhedrons. The maximum lengths were in the range of 10.1–27.7 μm (the number of grains measured 5–14 μm and more than 14 μm are 16 and 5, respectively) with a mean diameter of 18.1 ± 4.5 μm.

4.3. Sporopollen analysis Pollen analysis shows that the number of pollen grains in each sample is more than 200, generally 200–500 (Fig. 6). The sporopollen spectrum of the JXG2 profile reveals that, since the Holocene, this region has been dominated by herbaceous pollens. The vegetation was primarily alpine meadow-steppe, which indicates a dry-cold or wet-cold climate. From the bottom to the top, the sporopollen spectrum could be divided into three zones which are notably different in sporopollen

concentration and composition. This process was executed using the Cluster Analysis tool of Tilia software (Grimm, 2011). In the first sporopollen zone (120–101 cm; ca. 11,000–9400 cal BP), pollen taxa are fewer. Artemisia accounted for an average of 90%. The next most abundant pollen type belonged to the Poaceae, with an average percentage of less than 10%. The Poaceae include ll small grasses whose diameter is less than 35 μm. There were also small amounts of Ephedra and Asteraceae, suggesting a relatively dry climate. Sporopollen zone 2 can be further divided into zone 2-a (101– 67.5 cm; ca. 9,400–6100 cal BP) and zone 2-b (67.5–37 cm; ca. 6100– 3800 cal BP). Artemisia accounted for approximately 80% of plants in 2-a. The Poaceae content increased to an average of 10%. Grasses with diameters above 35 μm began to appear. A small amount of Ranunculaceae, Rosaceae, Cruciferae, and Chenopodiaceae also appeared. When the pollen concentration reaches the highest amount in the Holocene, it indicates that the climate was warm and more suitable for the growth of vegetation. In 2-b, Artemisia decreased and accounted for approximately 60% of plants. Poaceae content increased obviously, accounting for 19% on average. At ca. 5000 cal BP, Poaceae reached the peak of 25.7%. Polygonaceae, Ranunculaceae, Asteraceae, and Rosaceae content also increased slightly, indicating decreased humidity. In sporopollen zone 3 (37–0 cm; ca. 3800–0 cal BP), Artemisia accounted for 60% on average, Poaceae content declined slightly to an average content of 13%; Cyperaceae and Chenopodiaceae content increased, and Ephedraceae began to appear discontinuously, which indicated that the climate was drier than that of prediabetic stage.

5. Discussion 5.1. Starch grains from sherds

Fig. 3. Ceramic artifacts from Jiangxigou a JXG2-P7; b JXG2-P5; c JXG2-P6; d JXG2-P9. Sampling location: white arrows indicate the interior wall; white dots indicate the exterior wall. Scale bar: 1 cm.

No starch grains were recovered from the dust in the curating room, indicating that the residues on the four pottery sherds were not contaminated by this source. Using starch grain analysis to analyze the functions of artifacts, the ideal method would be to collect the surface residues from both the artifacts and nearby soil samples, and then to compare the results of the analyses to determine the possibilities of sample contamination (Barton, 2007). No starch grains were recovered from soil samples at depths of 42, 46, 48, 50, 52, 61, and 65 cm, indicating that the residues on the four sherds were not contaminated by this source. To evaluate the possibility of contamination, we also extracted samples from the inner and outer surfaces of each sherd. If the sherds were contaminated after burial, the starch grain assemblages extracted from

Please cite this article as: Hou, G., et al., Plant utilization at the Jiangxigou site during the middle Holocene, Archaeological Research in Asia (2015), http://dx.doi.org/10.1016/j.ara.2016.01.003

G. Hou et al. / Archaeological Research in Asia xxx (2015) xxx–xxx

5

Fig. 4. The comparison of morphology of starch grains from modern plants and surface residues of the artifacts. A. Morphology of starch grains from modern plants a. Setaria italica collected from Hebei province, mean diameter: 10.3 ± 2.2 μm (n = 100); b. Panicum miliaceum collected from Gansu province, mean diameter: 7.22 ± 1.2 μm (n = 116); c. Triticum aestivum collected from Qinghai province, mean diameter: 18.7 ± 8.8 μm (n = 112); d. Sorghum bicolor collected from Beijing city, mean diameter: 16.7 ± 5.5 μm (n = 160); e. Vigna radiata collected from Gansu province, mean diameter: 17.4 ± 6.3 μm (n = 111); f. Fritillaria cirrhosa collected from Sichuan province, mean diameter: 23.1 ± 11.9 μm (n = 118). Scale bar: 10 μm. B. Starch grains extracted from surface residues of the artifacts (a, b) Starch grain from millets; (c, d) starch grain with blurred edges; (e) unidentified starch grain. Scale bar: 20 μm.

inner wall and external surface of each sherd should be similar. Therefore, the degree of contamination can be judged by comparing the results of the starch grains extracted from both sides of each sherd. A total of 20 starch grains were extracted from the inner sherd walls, an amount far greater than grains collected from the external surfaces (only 2 starch grains). These differences suggest that starch grains collected from the sherd inner surfaces were related to ceramic function. Furthermore, scorching, smoke trails and blurred effects on some grains (Fig. 4B.c, d) may also imply that the four sherds were used for cooking (Crowther, 2012). Starch grain morphology of 38 species covering 28 genera and 13 tribes within 4 subfamilies of Poaceae examined in previous studies were examined for comparative purposes (Liu et al., 2010; Yang et al., 2012). Also we collected common plants such as wheat and barley that grow in the QTP now for comparative purposes (Fig. 4A.a–e). Starch grains from the genera Oryza, Eriochloa, Echinochloa, Panicum, Setaria, Digitaria, Coix, and Eleusine are generally polygon shaped. One-to-one comparisons with modern reference collections led us to identify 21 starch grains as derived from Poaceae taxa. The mean diameter of starch grains from genera Oryza, Eriochloa, Echinochloa, Digitaria and Eleusine are shorter than those of the starch grains recovered from our four pottery sherds. Meanwhile, the surface of starch grains from the genera Coix are smoother than those of the starch grains in our sherd sample. In contrast, the morphological features of the 21 starch grains are identical to starches from millets (Setaria spp. and Panicum spp.). For more detailed understanding, Yang et al. (2010, 2012) also examined 31 modern samples of both wild and domesticated millets derived from the

seeds of 7 species within the genus Setaria and 2 species within the genus Panicum, and determined diagnostic morphological characteristics. Since 23.8% starch grains with smooth surface are larger than 14 μm from samples of JXG2-P7, JXG2-P6 and JXG2-P9, and that is the limit of wild millets in size (Yang et al., 2012), we think that some granules were likely from domesticated foxtail millet (S. italica). Meanwhile, the mean diameter of the grains is a little larger than their modern counterparts, a factor which may be related to cooking (Crowther, 2012). In the present study, starch grains interpreted as Setaria spp. and Panicum spp. were found on four pottery sherds. Their age was approximately 5600–3600 cal BP. One artifact (no. JXG2-9) was found at a stratigraphic depth of 60–70 cm, and stratum 65 cm was carbon dated to 4850 ± 40 rcybp (5616 ± 40 cal BP). Another artifact (no. JXG2-7) was unearthed from 50 cm, and the remaining two ceramic artifacts were unearthed from depths of 40–50 cm. Based on the depth date relationship, the 50 cm stratum corresponds to approximately 4600 cal BP, whereas the 40–50 cm stratum corresponds to 4600–3600 cal BP. In other words, prehistoric inhabitants in this area may have started to use millet plants as early as 5500 cal BP, and continued to use these plant resources in the subsequent millennia until 3600 cal BP. 5.2. Pollen analysis from the JXG2 site There is a high content of Artemisia pollen in the JXG2 profile. As for the source of Artemisia pollen, the Qinghai Lake basin has been mainly

Table 3 Dates of the strata in the JXG2 site. Stratigraphic depth (cm)

Material

25 54 65 75 81 95 108

Ceramic Ceramic Charcoal Charcoal Charcoal Charcoal Charcoal

Lab numbers

Dating method

Date

Calibrated age range (95.4% C.I.)

BA111965

OSL OSL AMS14C AMS14C AMS14C AMS14C AMS14C

1970 ± 90 years ago 4973 ± 254 years ago 4850 ± 40 rcybp 5925 ± 35 rcybpa 7330 ± 50 rcybp 8170 ± 50 rcybp 7325 ± 35 rcybpa

5576–5656 cal BP 6666–6802 cal BP 8014–8215 cal BP 9010–9267 cal BP 8027–8192 cal BP

BA111966

Radiocarbon dates were calibrated using the Calib v.7.1 online software package (Stuiver and Reimer, 1986; Stuiver et al., 2015), using the XXX Calibration curve (REF). a Denotes data measured in the present study; the rest of the data are from Rhode et al. (2007).

Please cite this article as: Hou, G., et al., Plant utilization at the Jiangxigou site during the middle Holocene, Archaeological Research in Asia (2015), http://dx.doi.org/10.1016/j.ara.2016.01.003

6

G. Hou et al. / Archaeological Research in Asia xxx (2015) xxx–xxx

Fig. 5. Diagrams showing pedo-stratigraphy, lightness, redness, yellowness, and strata depth-date relations of JXG2.

characterized by alpine meadows/alpine grassland vegetation during the Holocene. Therefore, Artemisia pollen was likely produced in situ, from somewhere in the Qinghai Lake basin. Picea belongs to the alpine cold warm evergreen coniferous forest, which in the northeast edge of the QTP are mainly distributed at 2500–3950 masl, on the shady slope of Qilian mountain. JXG2, however, is located in the transition area between the Qinghai Lake basin and the Qinghai Nanshan, where Picea forest do not grow now. Picea forests may have grown in the area at certain altitudes in the Qinghai Nanshan during the Holocene Megathermal when there were more suitable climatic conditions, but this area is far away from JXG2. The low representation of Picea may reflect the difficulty of sporopollen transmission, which resulted in low Picea content. Salix belongs to temperate deciduous broad-leaved forest, which are mainly distributed in the Yellow River and HuangShui valley at 1700– 2600 masl on the northeastern edge of the QTP. In that zone, hydrothermal conditions are amenable to the growth of Salix, which has an upper limit of distribution lower than that of Picea. There is no distribution of Salix in the study area now, although Salix oritrepha lives in a few

appropriate areas such as the ShaLiu River valley that flows into the North Bank of Qinghai lake. In addition, Salix are zonal plants, with pollen that is difficult to transmit long distances, resulting in the fact that Salix does not appear in the sporopollen spectrum of JXG2. The results of analysis of the JXG2 pollen spectrum of grasses was roughly similar to that of the lake sediments and pollen from Qinghai Lake (Shen et al., 2005). The content of Poaceae of the lake core during ca. 6000–0 cal BP is 13%, with the maximum above 25%. The content of Poaceae of the lake core during ca. 11,000–6000 cal BP is only 10%, with a maximum surpassing only 15%; The content of Poaceae at JXG2 shows the same change. The content of Poaceae of JXG2 during 6000–0 cal BP is 15.7%, with a maximum also surpassing 25%. However, the content of Poaceae of JXG2 during 11,000–6000 cal BP is 8%, with a maximum surpassing only 15%. The two Poaceae curves reach maximum value around 5000 cal BP. Grasslands containing Poaceae in the semi-arid areas of northern China generally have a low percentage of Poaceae plants, typically not exceeding 10% (Zhou et al., 2011; Zhao et al., 2011). In contrast, the percentage of Poaceae pollens in JXG2 reached a peak value of 25.7% around

Table 4 Starch grain data on the surface of the ceramics. Name of the artifacts

No. of the artifacts

Sampling location

Starch grain from millets

Starch grain of unknown type

Total

Ceramics

JXG2-P7

Interior wall Exterior wall Interior wall Exterior wall Interior wall Exterior wall Interior wall Exterior wall

8 1 1 0 7 1 3 0 21

1 0 0 0 0 0 0 0 1

9 1 1 0 7 1 3 0 22

JXG2-P5 JXG2-P6 JXG2-P9 Total

Please cite this article as: Hou, G., et al., Plant utilization at the Jiangxigou site during the middle Holocene, Archaeological Research in Asia (2015), http://dx.doi.org/10.1016/j.ara.2016.01.003

G. Hou et al. / Archaeological Research in Asia xxx (2015) xxx–xxx

Fig. 6. Sporopollen composition at different dates in JXG2 (unit of sporopollen concentration: grain/g), Poaceae* from Shen et al. (2005).

Please cite this article as: Hou, G., et al., Plant utilization at the Jiangxigou site during the middle Holocene, Archaeological Research in Asia (2015), http://dx.doi.org/10.1016/j.ara.2016.01.003

7

8

G. Hou et al. / Archaeological Research in Asia xxx (2015) xxx–xxx

5000 cal BP. This percentage is far greater than the average for natural Poaceae. The reason for the high content of Poaceae at about 5000 cal BP is climate fluctuation. More humid climatic conditions are reflected in the higher Poaceae content. 5.3. Comparing plant use at nearby sites Prehistoric foxtail millet and broomcorn millet remains were distributed broadly across the northeastern margin of the QTP and surrounding areas (Fig. 7). At the Dadiwan site (7900 cal BP), the earliest millet remains in northwest China have been found (Xie, 2002). At the Hulijia site (5450 ± 135 cal BP) of the Miaodigou phase of the Yangshao Culture, flotation revealed millet remains, providing evidence of the earliest use of millet on the edge of the Qinghai–Tibet Plateau (IA, CASS and Institute of Archaeology, Qinghai Province, 2001; Dong et al., 2013). These crops are also found to exist widely in Majiayao Culture sites (5300–4300 cal BP) (Jia et al., 2012). On the northeastern margin of the plateau, only the Hulijia site has dates for S. italica and P. miliaceum that are close to that of the JXG2 site. The time of millet remains at all other sites are later; therefore, JXG2 has some of the earliest S. italica and P. miliaceum remains on the northeastern margin of the QTP. The remains of Setaria spp. and Panicum spp. found in our study are currently the first definite evidence for the earliest use of millets at an altitude of over 3000 masl in the QTP region. The discovery of the earliest use of millets (Setaria spp. and Panicum spp.), very likely including domesticated foxtail millet at this site was not expected and may be related to the expansion of the Yangshao culture in the Loess Plateau and the special geological constitution of the adjacent Loess Plateau. Neolithic culture in northern China reached its peak during the Yangshao Culture period (7000–5000 cal BP). It then rapidly expanded during the Miaodigou phase of the Yangshao Culture (6000–5000 cal BP) (Barton et al., 2009). Over 5000 Yangshao sites that date to this time are known. The core sites are distributed in the basins of the Weihe and Jing Rivers. A large-scale expansion of the Yangshao Culture continued eastwards into the lower reaches of the Fen River

Basin of the Yellow River, northwards to the Hetao Plain in Inner Mongolia, and westwards to the HuangShui watershed (Bureau of National Cultural Relics, 1991, 1996, 1998, 2003, 2007, 2010). In the Yellow River valley on the northeastern margin of the QTP, the Andaqiha site (5870 ± 97 cal BP) and Hulijia site (Dong et al., 2013) which belong to the Miaodigou phase of the Yangshao Culture have recently been discovered. These sites were likely created during a largescale expansion of the Miaodigou phase of the Yangshao Culture. The Yangshao Culture is known for cultivation of millet and the exquisite technique of making ceramics (Xie, 2002). Our findings at the JXG2 site clearly show that before the arrival of the Yangshao Culture, there already were native hunters who were capable of making microliths, who hunted goral, deer, antelope, and other wild animals (Table 2). Sites belonging to the same culture of the native hunters also include the Layihai site (7652 ± 68 cal BP) (Gai and Wang, 1983) and the Shalongka site (8320–8060 cal BP) (Dong et al., 2013). However, neither pottery, polished stone tools, nor millet were found in any of these sites. At the depth of 61 cm of the JXG2 profile, painted pottery sherds were found with drawings composed of dots and curves that are representative of the Yangshao Culture or early Majiayao phase (JXG2-P10). Such cultural exchanges became frequent with the rapid expansion of the Miaodigou phase of the Yangshao Culture. As a result of this exchange, the social life of JXG2 microlithic hunters underwent profound changes that coincided with the adoption of pottery and the utilization of millets. The northeastern margin of the QTP thus entered the Neolithic era. 6. Conclusion From 10,000 to 6900 cal BP, JXG2 was a site characterized by microliths, and it contained no ceramics. Subsequently, ceramics were found in remains dating from 6900–3600 cal BP. Millets (Setaria spp. and Panicum spp.) likely including 23.8% domesticated foxtail millet (based on plant microremains) were found in residues from four pottery sherds, indicating that prehistoric residents at JXG2 site used millets. The earliest pottery sherds containing millets may date to about 5500 cal BP, and

Fig. 7. Main millet remains from the northeastern margin of the QTP and the expansion of the Yangshao culture. a. Main millet remains: 1. Dadiwan, 2. Buziping, 3. Baidaoping, 4. Qinggangcha, 5. Majiawan, 6. Linjia, 7. Hulijia, 8. Hetaozhuang, 9. Liuwan, 10. Yuanyangchi, 11. Huishan. other sites mentioned in the text: 12. Andaqiha, 13. Layihai. b. Painted ceramics unearthed at stratum 61 cm in JXG2.

Please cite this article as: Hou, G., et al., Plant utilization at the Jiangxigou site during the middle Holocene, Archaeological Research in Asia (2015), http://dx.doi.org/10.1016/j.ara.2016.01.003

G. Hou et al. / Archaeological Research in Asia xxx (2015) xxx–xxx

this is currently the earliest evidence of millet use at over 3000 masl on the QTP. The use of these early millets found at JXG2 on the northeast margin of the QTP was a result of the rapid expansion of the Miaodigou style of the Yangshao Culture to the QTP along the Yellow River. The Miaodigou style expanded rapidly to the QTP between 6000– 5000 cal BP, leading to a transition of lifestyle for plateau natives who primarily lived on hunting with microliths. The use of ceramics and millet increased the agricultural component of the economic life and this helped the indigenous culture on the QTP enter the Neolithic era. Acknowledgments This work was supported by the China NSF (grant 41550001, 41361047,41301045). Meanwhile, we thank David Rhode (University of Washington), David B. Madsen (University of Texas) and P. Jeffrey Brantingham (University of California, Los Angeles) for completing field work with us and other help. We were responsible for the related work on the pottery sherds. References Barton, H., 2007. Starch residues on museum artefacts: implications for determining tool use. J. Archaeol. Sci. 34, 1752–1762. Barton, L., Newsome, S.D., Chen, F.H., et al., 2009. Agricultural origins and the isotopic identity of domestication in northern China. Proc. Natl. Acad. Sci. 106 (14), 5523–5528. Brantingham, P.J., Gao, X., 2006. Peopling of the northern Tibetan plateau. World Archaeol. 38, 387–414. Bureau of National Cultural Relics, 1991. Atlas of Chinese Cultural Relics — Fascicule of Henan Province. China Cartographic Press, Beijing, pp. 13–20 (in Chinese). Bureau of National Cultural Relics, 1996. Atlas of Chinese Cultural Relics — Fascicule of Qinghai Province. China Cartographic Press, Beijing, pp. 15–23 (in Chinese). Bureau of National Cultural Relics, 1998. Atlas of Chinese cultural relics — fascicule of Shaanxi province. Xi'an Cartographic Press, Xi'an, pp. 17–28 (in Chinese). Bureau of National Cultural Relics, 2003. Atlas of Chinese cultural relics — fascicule of Inner Mongolia province. Xi'an Cartographic Press, Xi'an, pp. 19–26 (in Chinese). Bureau of National Cultural Relics, 2007. Atlas of Chinese Cultural Relics — Fascicule of Shanxi Province. China Cartographic Press, Beijing, pp. 16–26 (in Chinese). Bureau of National Cultural Relics, 2010. Atlas of Chinese Cultural Relics — Fascicule of Gansu Province. Surveying and Mapping Press, Beijing, pp. 22–26 (in Chinese). Chen, H.H., Ge, S.B., Li, G.L., 1998. Discuss to the nature of Zongri relic. Kaogu (Archaeol.) 5, 23–27. Chen, F.H., Dong, G.H., Zhang, D.J., et al., 2015. Agriculture facilitated permanent human occupation of the Tibetan Plateau after 3600 BP. Science 347 (6219), 248–250. Craig, O., Mulville, J., Pearson, M.P., 2000. Archaeology: detecting milk proteins in ancient pots. Nature 408, 312. Craig, O., Saul, H., Lucquin, A., 2013. Earliest evidence for the use of pottery. Nature 496, 351–354. Crowther, A., 2012. The differential survival of native starch during cooking and implications for archaeological analyses: a review. Archaeol. Anthropol. Sci. 4, 221–235.

9

Cui, Y.P., Hu, X.W., Chen, H.H., Dong, Y., Guan, L., Weng, Y., Wang, C.S., 2006. Stable isotopic analysis on human bones from Zongri site.Di Si Ji Yan Jiu. Quat. Sci. 26, 604–611 (in Chinese). D'Alpoim Guedes, J., Lu, H.L., Li, Y.X., 2014. Moving agriculture onto the Tibetan Plateau: the archaeobotanical evidence. Archaeol. Anthropol. Sci. 6, 255–269. Dong, G.H., Jia, X., Elston, R., 2013. Spatial and temporal variety of prehistoric human settlement and its influencing factors in the upper Yellow River valley, Qinghai Province, China. J. Archaeol. Sci. 40, 2538–2546. Fu, D.X., 2001. The discover, identification and research of plant crop remains from Neolithic in Changguogou site in Tibet. Kaogu (Archaeol.) 3, 66–74 (in Chinese). Gai, P., Wang, G.D., 1983. Excavation report on a Mesolithic site at Layihai, Upper Yellow River. Renlei Xuebao (Acta Anthropol. Sin. 2, 49–59 (in Chinese). Grimm, E.C., 2011. Tilia software v. 1.7. 16. Illinois State Museum. Research and Collection Center, Springfield. IA, CASS (Institute of Archaeology, Chinese Academy of Social Sciences), Institute of Archaeology, Qinghai Province, 2001i. Excavation of the Hulijia Site in Minhe, Qinghai province. Kaogu(Archaeol.) 12, 12–25 (in Chinese). Jia, X., Dong, G.H., Li, H., 2012. The development of agriculture and its impact on cultural expansion during the late Neolithic in the Western Loess Plateau, China. The Holocene 23, 85–92. Li, X.Q., Shang, X., Zhou, X.Y., 2006. Integrative method of sieving and heavy liquid in pollen analysis of loess-data analysis and processing. Ganhanqu Dili (Arid Land Geogr.) 29, 663–667 (in Chinese). Liu, L., Field, J., Fullagar, R., 2010. A functional analysis of grinding stones from an early Holocene site at Donghulin, North China. J. Archaeol. Sci. 37, 2630–2639. Piperno, D.R., Ranere, A.J., Holst, I., 2000. Starch grains reveal early root crop horticulture in the Panamanian tropical forest. Nature 407, 894–897. Rhode, D., Zhang, H.Z., Madsen, D.B., 2007. Epipaleolithic/early Neolithic settlements at Qinghai Lake, western China. J. Archaeol. Sci. 34, 600–612. Shen, J., Liu, X.Q., Wang, S.M., 2005. Palaeoclimatic changes in the Qinghai Lake area during the last 18,000 years. Quat. Int. 136, 131–140. Stuiver, M., Reimer, P.J., 1986. A computer program for radiocarbon age calibration. Radiocarbon 28, 1022–1030. Stuiver, M., Reimer, P.J., Reimer, R.W., 2015. CALIB 7.1. http://www.calib.qub.ac.uk/calib/. Xie, D.J., 2002. Prehistoric Archaeology of Gansu Province and Qinghai Province. Cultural Relics Press, Bejing (in Chinese). Yang, X.Y., Jiang, L.P., 2010. Starch grain analysis reveals ancient diet at Kuahuqiao site, Zhejiang Province. Chin. Sci. Bull. 55, 1150–1156. Yang, X.Y., Perry, L., 2013. Identification of ancient starch grains from the tribe triticeae in the north China plain. J. Archaeol. Sci. 40, 3170–3177. Yang, X.Y., Kong, Z.C., Liu, C.J., 2010. Morphological characteristics of starch grains of millets and their wild relatives in north China. Di Si J. Yan Jiu (Quat. Sci.) 30, 364–371 (in Chinese). Yang, X.Y., Zhang, J., Perry, L., 2012. From the modern to the archaeological: starch grains from millets and their wild relatives in China. J. Archaeol. Sci. 39, 247–254. Yuan, J., 2002. The origin of rice and pottery in Yuchanyan, Dao County, Hunan Province. In: Yasuda, Y. (Ed.), The Origins of Pottery and Agriculture. Roli Books, New Delhi, pp. 157–166. Zarrillo, S., Pearsall, D.M., Raymond, J.S., 2008. Directly dated starch residues document early formative maize (Zea mays L.) in tropical Ecuador. Proc. Natl. Acad. Sci. 105, 5006–5011. Zhao, Y., Yu, Z.C., Zhao, W.W., 2011. Holocene vegetation and climate histories in the Eastern Tibetan Plateau: controls by insulation driven temperature or monsoon derived precipitation. Quat. Sci. Rev. 30, 1173–1184. Zhou, X.Y., Li, X.Q., Zhao, K.L., 2011. Early agricultural development and environment effects in the Neoloithic Longdong basin(East Gansu). Chin. Sci. Bull. 2011, 56. http:// dx.doi.org/10.1007/s11434~010~4268~x.

Please cite this article as: Hou, G., et al., Plant utilization at the Jiangxigou site during the middle Holocene, Archaeological Research in Asia (2015), http://dx.doi.org/10.1016/j.ara.2016.01.003