Plant domestication,cultivation,and foraging by the first farmers in early Neolithic Northeast China: Evidence from microbotanical remains

Abstract

North China is regarded as a center of domestication for broomcorn millet (Panicum miliaceum) and foxtail millet (Setaria italica ssp. italica). The Neolithic Xinglonggou site (ca. 8000–7500 cal. BP) in the Liao River region has revealed the earliest macrobotanical evidence of domesticated millets in Northeast China, but controversy remains as to the importance of the millets in human diet. We employed an interdisciplinary approach involving analyses of starch grains, phytoliths, and usewear patterns to study a range of materials from Xinglonggou, including grinding stones, human dental calculus, and vegetative charcoal. The results demonstrate a broad spectrum of plant exploitation by the first farmers in Northeast China rather than dependence upon singular crops. Furthermore, three types of underground storage organs appear to be major staples, while millets were secondary to another early and important cultivated cereal, Job’s tears (Coix lacryma-jobi L.). Wild grasses and acorns also account for small portions of plants used. This study documents the northernmost and earliest occurrence of Job’s tears in temperate Northeast China, a species which may have originated in the subtropical regions. We argue that Job’s tears may have been one of the earliest domesticates in China along with millets.

Keywords

agricultural origins ancient starch grains Job’s tears millets phytoliths underground storage organs

Introduction

The investigation of the origins of agriculture is of enduring interest to archaeologists. China is regarded as one of the centers where plant domestication developed independently. The early Neolithic Xinglongwa culture (ca. 6200–5200 BC), which is characterized by sedentary village life, has revealed the earliest evidence of domesticated broomcorn millet (Panicum miliaceum) and foxtail millet (Setaria italica ssp. italica) in Northeast China (Zhao, 2011). Stable isotope data from human skeletons found at several sites indicate that C4 plants, assumed to be millets, were primary in the diet (Liu et al., 2012). Indeed, domesticated millets, dating to ca. 7670–7610 cal. BP, comprise 42.1% of preserved archaeological plant remains recovered by flotation from the Xinglonggou site of the Xinglongwa culture (Sun, 2014); however, this figure is much lower than the 99% of the seeds belonging to cultigens (primarily millets) in the flotation samples from the Bronze Age stratum at the same site (Zhao, 2004). We feel that the isotope and archaeological plant remain data are inconclusive and do not sufficiently support the assumption that subsistence of the Xinglongwa people was based predominantly on millet cultivation. We argue that additional indirect and direct indicators of food processing and consumption revealed from other materials can provide a more holistic understanding of early agriculture in Northeast China. Here, we report the results of an interdisciplinary study of a range of material remains from Xinglonggou. Methods used include analyses of starch grains, phytoliths, other plant microfossil, and usewear patterns.

Archaeological background

Xinglonggou is situated on a sloping area near the Mangniu River, a tributary of the Daling River, in Chifeng, central-east Inner Mongolia. During the mid-Holocene, the region belonged to the temperate forest zone (Winkler and Wang, 1993), and the annual average temperature was 3–4°C higher than at the present (Cui and Kong, 1992). The Chifeng area was characterized by warm and moist climate, and the landscape was covered with primarily grassland vegetation and small areas of broad-leaved forests during the Xinglongwa culture period (Xu et al., 2002).

The Xinglonggou site was discovered in 1982 and excavated from 2001 to 2003. The site comprises three separate localities, which belong to the middle phase of the Xinglongwa culture (ca. 8000–7500 cal. BP), the late Hongshan culture (ca. 5500–5000 cal. BP), and the Lower Xiajiadian culture (ca. 4000–3500 cal. BP), respectively. We discuss only Locality I (Xinglongwa) in this paper. The site (4.8 ha) of the Xinglongwa period comprised 145 semi-subterranean houses, from which indoor human burials and numerous artifacts were uncovered. A total of 37 houses, 28 indoor human burials, and 57 ash pits were excavated within an area of about 5600 m² in 2001–2003 (Figure 1). The artifacts found include pottery, stone tools, and ornaments made of jade, shell, bone, and stone. To date, only preliminary excavation reports have been published (Liu et al., 2004; Yang et al., 2000).

Figure 1.

Location of the Xinglonggou site and house remains under excavation.

Flotation methods were employed during the excavation. A total of 1082 flotation samples, 20–40 L each, were processed, and 2440 carbonized seeds were found, including 1026 millets (42.1%). Foxtail millet (n = 41) was outnumbered by broomcorn millet (n = 985). In addition to millets, the economically or medicinally important plants include Glycine soja (wild soybean, n = 8), Quercus sp. (acorn, n = 33), Astragalus membranaceus (n = 263), Chenopodium album (n = 479), and Artemisia annua L. (n = 309), among other taxa (Sun, 2014). It is notable that the great majority of the millets were recovered from only two houses (F10 and F31) (Liu et al., 2004; Zhao, 2004).

Grinding stones, including slabs (mopan) and elongate handstones (mobang), were commonly found on house floors at Xinglonggou. Similar tools from several Upper Paleolithic and early Neolithic sites in North China have been subjected to residue and usewear analyses in recent years, and the results suggest that they were primarily, although not exclusively, used as plant food-processing tools (see a summary in Liu, 2015). The presence of large quantities of grinding stones at Xinglonggou offers a great opportunity for analyzing possible plant residues. Human dental calculus provides the potential for direct evidence of food consumption. Small pieces of carbonized plant remains recovered by flotation, which were apparently not millets and identified as non-wood charcoal, were also subjected to microscopic analysis to understand the use of other plants.

We analyzed 22 grinding stones (14 slabs and 8 handstones from nine houses) for usewear traces and residue remains, dental calculus from 11 human individuals, and starch recovered from two pieces of non-wood charcoal from a house (F10).

Methods

We extracted residues from the grinding stones using sterile pipettes and distilled water. This simple protocol (Fullagar, 2006; Loy and Fullagar, 2006) has been tested and shown effective in our previous studies (Liu et al., 2011, 2013). Residue samples were processed for starch and phytolith extraction using the heavy liquid sodium polytungstate (in a specific gravity of 2.35). The process for extracting residues from dental calculus was based on published studies (Boyadjian et al., 2007; Henry and Piperno, 2008) with modification as follows: dental calculus from each tooth was scraped off with a clean blade and stored in a sealed plastic bag. The residue was then transferred to a test tube and mixed with sodium hexametaphosphate (10%) to disperse particles. Hydrochloric acid (7%) was then used to dissolve calcium. The charred archaeological plant samples were placed into tubes and dissolved in distilled water before being mounted directly on glass slides. All microfossil extractions were mounted in a solution of 25% glycerol on glass slides and scanned under a Zeiss Axio Scope A1 fitted with polarizing filters and DIC (Differential Interference Contrast) optics. Photographs were taken using a Zeiss Axiocam HRc3 digital camera and Zeiss Axiovision software version 4.8.

Our modern plant collections include over 950 specimens. We specifically analyzed those starch-rich and economically important samples relevant to the research area, including more than 170 samples belonging to 84 species in 46 genera of 19 families.

For usewear analysis of the grinding stones, Polyvinyl siloxane (PVS or peel) was applied to the artifacts to provide portable and durable records for microscopic analysis. Previous research works, using PVS on usewear patterns from grinding stones in China (Fullagar et al., 2012; Liu et al., 2010a, 2010b, 2011, 2013), have established valuable reference data for the study of ancient tools. Our experimentally derived usewear comparative collection of tools includes usewear examples from the processing of stone, mineral, wood, shells, seeds, nuts, and tubers. These experimental studies provide the analytical variables of usewear to be examined in the Xinglonggou artifacts: stage of polish development (low, medium, high); polish reticulation pattern; polish topography; striations, such as furrow (V shape in intersection), sleek (U shape in intersection), and fine; pitting and pecking; and surface micro-topography.

Results

Starch types identification

A total of 800 starch grains were recovered from the tools and dental calculus. They can be classified into seven types based on their morphology and size (Figures 2 and 3; Table 1), and these types are identifiable as certain plant taxa when compared with our reference data (Figure S1, available online). Many of the starch grains show sign of damage (n = 198; 24.8% of the total), characterized by broken edges, rough surface, deep fissures, pronounced lamellae, or a dark area in the center of the extinction cross. These features are consistent with the morphological changes after grinding based on experimental studies (Ge et al., 2010; Henry et al., 2008). The morphology of the starch types is described in SI Text (available online). The significance of the plants identified is presented in the ‘Discussion’ section.

Figure 2.

Starch types from Xinglonggou (each grain shown in DIC and polarized views).

Figure 3.

Xinglonggou starch sizes compared with modern reference samples.

Table 1.

Starch counts of Xinglonggou grinding stones and dental calculus.

	Type I	Type II	Type III		Type IV	Type V		Type VI	Type VII
	Lilium spp.	Dioscorea polystachya	Trichosanthes kirilowii	USOs	Coix lacryma-jobi	Setaria italica	Panicum miliaceum	Triticeae grasses	Quercus spp.	UNID	Total
Slabs
GS1	19				3	2	4	2		2	32
GS2	3				4						7
GS3			2		4						6
GS4					15	6	10	2	1	1	35
GS5	2		4	1	2					2	11
GS6	41	2			40	22	9			9	123
GS7	5		2	2	8	3	4			1	25
GS9	3		4		7						14
GS10	1				26	2		4		3	36
GS14	24			10						4	38
Subtotal	98	2	12	13	109	35	27	8	1	22	327
%	30.0%	0.6%	3.7%	4.0%	33.3%	10.7%	8.3%	2.4%	0.3%	6.7%	100%
Mortar-slabs
GS8	5				43	18	3			1	70
GS11	15		3	6	3					4	31
GS12	2				1						3
GS13	3			4	2	2				1	12
Subtotal	25		3	10	49	20	3			6	116
%	21.6%		2.6%	8.6%	42.2%	17.2%	2.6%			5.2%	100%
Handstones
R1	2				10	2					14
R2					2	2				2	6
R3					11	16	1			3	31
R4	2				16	1		1		2	22
R5	1			1						3	5
R6	7				46	11	1	25		5	95
R7	1			5	50	20	8	11		14	109
R8		1		9	28	8	2	5	1	3	57
Subtotal	13	1		15	163	60	12	42	1	32	339
%	3.7%	0.3%		4.2%	46.2%	17.0%	3.4%	11.9%	0.3%	9.1%	96%
Dental calculus
M11				1							1
M13										2	2
M15										1	1
M22								1			1
M23				2				8			10
M24					1			1			2
M25								1			1
Subtotal				3	1			11		3	18
%				16.7%	5.6%			61.1%		16.7%	100%
Total	136	3	15	41	322	115	42	61	2	63	800
Total %	17.0%	0.4%	1.9%	5.1%	40.3%	14.4%	5.3%	7.6%	0.3%	7.9%	100%

Type I starch grains (n = 136; 17% of the total; from 17 tools) resemble those from Lilium sp. bulbs. At least eight lily species are widely distributed in Northeast China today (Ma, 1985a; Yang et al., 1996). When compared with our reference data, Type I starch grains best match to Lilium pumilum DC and Lilium tigrinum; both can be found in the region.

Type II starch grains (n = 3; 0.4% of the total; from two tools) resemble those from yams including Dioscorea polystachya. Wild yam is distributed widely in China, growing in forests, scrub forests, herb communities, mountain slopes, along rivers, and roadsides. It is also commonly cultivated (Wu and Raven, 2000). It is unclear whether wild yam grew in the Liao River area during the Neolithic period due to the lack of data. Only cultivated yam grows in the region today (Ma, 1985a).

Type III starch grains (n = 15; 1.9% of the total; from five tools) are most comparable with the root of snake gourd (Trichosanthes kirilowii in the Cucurbitaceae family). This particular root has been known as an edible plant since antiquity (see discussion below). Snake gourd is widely distributed in northern and southern China, growing in open forests, shrub lands, grasslands, and fields beside villages in many regions of China (Wu et al., 2011). It is unclear, however, whether or not snake gourd was native to the Liao River region. It is cultivated in Inner Mongolia today (Ma, 1980).

Type IV starch grains (n = 322; 40.3% of the total; from 20 tools and 1 calculus sample) best match to the seed of Job’s tears (Coix lacryma-jobi). Job’s tears grow near streams, in marshy valleys, moist fields, and by houses. The cultivated forms are widely distributed in China, but the wild ones grow mainly in tropical and subtropical areas (Wu et al., 2006). It is cultivated in Inner Mongolia today (Ma, 1983).

Type V starch grains (n = 157; 19.7% of the total; from 15 tools) best match to millets of the Panicoideae sub-family and may be further identified as broomcorn millet (V.1) and foxtail millet (V.2), based on the discriminant analysis of a statistical model. Because the success rates for separating the two millets in the model are rather low (below 70%) (Liu et al., 2014), we identify Type V starch as millets without attempting to separate them at the species level. Wild foxtail millet (Setaria viridis) produces starch smaller in size than that of domesticated foxtail millet, although the morphologies of the two species are similar. Flotation samples from Xinglonggou have revealed only three seeds of wild foxtail millet (0.12% of the total seeds) (Sun, 2014); therefore, it is possible that millet starch grains are primarily, if not all, derived from domesticated millets.

Type VI starch grains (n = 61; 7.6% of the total; from seven tools and four calculus samples) resemble many taxa in the Triticeae tribe of the grass family. Several genera of Triticeae are indigenous to North China and still found in Inner Mongolia today, such as Agropyron, Elymus, Roegneria, Hordeum, and Leymus (Ma, 1983). The characteristics of Type VI grains are particularly consistent with those from Agropyron cristatum and Agropyron desertorum in our reference data.

Type VII starch grains (n = 2; 0.3% of the total; from two tools) are similar to several Quercus species in our reference data and other publications (Yang et al., 2009), such as Quercus viriabilis which is widely distributed in China (Wu and Raven, 1999). Inner Mongolia today has two species of Quercus, Quercus liaotungensis and Quercus mongolica (Ma, 1985b).

Some starch grains are only identifiable as having a general underground storage organ (USO) origin (n = 41; 5.1% of the total; from nine tools and two calculus samples). These grains can be found in the USOs of plants including lilies and snake gourd root. We also observe starch grains that lack diagnostic features comparable to taxa in our reference collection; these cannot be identified and are classified as UNID (n = 63; 7.9% of the total; from 17 tools and two calculus samples).

Phytolith identification

Phytoliths recovered from the artifact residues and dental calculus number significantly lower than starch grains in the artifact and dental calculus samples. Only 89 individual phytoliths were observed; however, additional plant-derived material, such as fibers and vascular tissue, was noted (Figure 4; Table 2). Most vegetal fibers and vascular tissue lack diagnostic taxonomic value (Piperno, 2006); however, silicified tracheary elements (n = 21) from USOs appear in residues of tools used for the processing of USOs. Tracheary elements are relatively abundant in these samples (n = 21), comprising 23.6% of the phytolith assemblage.

Figure 4.

Selected phytolith types and fiber from Xinglonggou.

Table 2.

Phytolith counts of Xinglonggou grinding stones and dental calculus.

	Asteraceae	Panicoideae	Panicoideae	Panicoideae	Panicoideae	Panicoideae	Pooideae	Poaceae	Poaceae	Poaceae	Hair cell	Hair cell	Hair cell	Hair cell	Vascular	Fibers	Total
	Sheets	Husk Type 1	Husk Type 2	Epidermal	Bilobate	Cross	Short cell	Bulliform	Long cell	Short cell	Type 1	Type 2	Type 3	Type 4
Slabs
GS1																P	0
GS2											17	2			12	A	19
GS3																	0
GS4														1			1
GS5																	0
GS6	4	1							2	1	3				4		11
GS7											1					P	1
GS8	1						2		1		1				1	P	5
GS9											2					P	2
GS10				1				1			4				2		6
GS14																	0
Subtotal	5	1	0	1	0	0	2	1	3	1	28	2	0	1	19		45
%	11.1%	2.2%	0.0%	2.2%	0.0%	0.0%	4.4%	2.2%	6.7%	2.2%	62.2%	4.4%	0.0%	2.2%	42.2%		100%
Mortar-slabs
GS8																	0
GS11	1				3			1			7					A	12
GS12						1											1
GS13																P	0
Subtotal	1	0	0	0	3	1	0	1	0	0	7	0	0	0	0		13
%	7.7%	0.0%	0.0%	0.0%	23.1%	7.7%	0.0%	7.7%	0.0%	0.0%	53.8%	0.0%	0.0%	0.0%	0.0%		100%
															(Continued)
Handstones
R1									1		1						2
R2						1											1
R3								5			1		1			P	7
R4											1						1
R5																	0
R6			2					4	2		4						12
R7									3							P	3
R8											1				1	P	1
Subtotal	0	0	2	0	0	1	0	9	6	0	8	0	1	0	1		27
%	0.0%	0.0%	7.4%	0.0%	0.0%	3.7%	0.0%	33.3%	22.2%	0.0%	29.6%	0.0%	3.7%	0.0%	3.7%		100%
Dental calculus
M11							1				1					P	2
M13																	0
M15											1						1
M22																	0
M23																	0
M24																	0
M25																	0
M19											1				1		2
Subtotal	0	0	0	0	0	0	1	0	0	0	3	0	0	0	1		5
%	0.0%	0.0%	0.0%	0.0%	0.0%	0.0%	20.0%	0.0%	0.0%	0.0%	60.0%	0.0%	0.0%	0.0%	20.0%		100%
Total	6	1	2	1	3	2	3	11	9	1	46	2	1	1	21		90
Total %	6.7%	1.1%	2.2%	1.1%	3.3%	2.2%	3.3%	12.2%	10.0%	1.1%	51.1%	2.2%	1.1%	1.1%	23.3%		100%

The largest percentage (55.5%) of phytoliths in this assemblage is composed of hair cells, silicified appendages from the epidermal surfaces of leaves produced in eudicot plant families (Kealhofer and Piperno, 1998; Piperno, 2006). Hairs were classified into four categories (Figure 4; Table 2). Type 1 (n = 46) hairs are simple, unicellular, lanceolate cells. Type 2 (n = 2) hairs are simple, unicellular, lanceolate hair cells that are armed, having spine-like surface projections. Type 2 hairs are consistent with Boraginaceae. Type 3 (n = 1) includes a single multicellular hair with blunt tip observed with a cistolith inclusion at its base. Finally, Type 4 (n = 1) hairs are multicellular with pointed tip lacking cistolith inclusion. Only Type 2 hairs are identified at this time to the family level, and these are consistent with hairs observed in the borage family, Boraginaceae. Other evidence of eudicot plants in the phytolith assemblage comes from distinctive, black, perforated epidermal sheets (n = 6) produced in the daisy family, Asteraceae (Pearsall, 2000). These darkened epidermal sheet fragments, possibly composed of silicon dioxide, are produced in abundance in this family (Piperno, 2006).

Poaceae phytoliths comprise the remainder of the phytoliths present (n = 33; 36.5%). In this assemblage, bulliform cells, epidermal long cells with indistinct borders, and redundant short cells are all forms that are produced across taxa in the Poaceae. Subfamily identification was possible with several phytoliths: simple short cells produced in Pooideae-type (Festucoid) (Twiss et al., 1969) and Panicoideae-type phytoliths, such as bilobate forms (n = 3), cross-shaped forms (n = 2), and epidermal sheets containing bilobates. Cross-shaped phytoliths are consistent with those observed in Coix lacryma-jobi (Lu and Liu, 2003) and other grasses of the Panicoideae. Interlocking epidermal sheets of husk cells (n = 3) from Panicoideae grasses were also observed and classified as two types. The Type 1 husk phytolith resembles husk cells produced in Setaria italica, foxtail millet (Lu et al., 2006; Zhang et al., 2011). However, much more comparative work is needed toward identifying and differentiating cultivated millets from related weedy taxa encountered in archaeological samples (Madella et al., 2013; Weisskopf and Lee, 2014).

Usewear patterns and summary plant microfossil residues from grinding stones

The grinding stones are classified into three general types: (1) slabs (n = 10), which are nearly rectangular or square in shape and have a relatively flat working surface; (2) mortar-slabs (n = 4), which are slabs with a depression in the center of the tool, presumably used as mortar as well as a slab; and (3) elongate handstones (n = 8), which often have multiple lateral surfaces and were used for abrading; the distal ends sometimes were used as a pestle for pounding (Figure S2, available online).

The usewear traces of each tool type (Figure 3) are compared with reference data from our experimental grinding (Figure S3, available online). The usewear patterns are discussed here together with residue remains from each tool type, in order to explore their functional differences.

Slabs

Usewear patterns on the slabs show many similarities and can be described as five broad forms: low-level polish, medium-level polish with fine striations, high-level polish, high-level polish with fine striations, and pitting. The striations are often multi-directional. When polish is well developed, crystal grains usually show rounded edges. The most common form is medium- to high-level polish with or without fine striations, which are comparable with processing tuber and Job’s tears, based on our experimental study. The low-level polish is similar to those from grinding cereals, such as millet, but possibly also resulted from low frequency of use.

Starch and phytolith residues recovered from these slabs indicate that a wide variety of plant tissues were processed with these tools. Seven types of starch are identified (n = 327); the most predominant starch grains are from three types of USOs (38.2%), followed by Job’s tears (33.3%), millets (19%), Triticeae grasses (2.4%), and acorn (0.3%). Phytoliths recovered from the slabs also show a greater variety of types (n = 45) relative to the other tools with Type 1 eudicot hair cells being most abundant (60%), followed by tracheary/vascular elements (42.2%), Asteraceae sheet fragments (11.1%), and various grass phytoliths including a Panicoideae husk fragment. In addition, Type 2 eudicot hair cells identified to Boraginaceae are present. Vegetal fibers were also observed on four of the slabs.

Mortar-slabs

The usewear traces on the flat surfaces are similar to those of slabs, characterized by different degrees of polish with/without fine striations. On the other hand, polished high spots with short sleeks are present on the depression of each tool, suggesting that it was used as mortar for pounding and grinding plant materials. The fine striations and sleeks on the polished areas particularly resemble those from tuber processing in our experimental collection.

Residue samples from these mortar-slabs yielded 116 starch grains, including Job’s tears (42.2%), three types of USOs (32.8%), and millets (19.8%). Starches from Triticeae grass and acorn are absent. A total of 13 phytoliths were recovered from these tools. Type 1 eudicot hair cells are most abundant (53.8%), followed by Panicoid bilobate phytoliths (23.1%), Asteraceae sheet fragments (7.7%), Panicoid crosses (7.7%), and Poaceae bulliforms (7.7%). Fibers were observed in abundance on one slab and present on another.

Elongate handstones

All the handstones appear to have been heavily used, showing polish and striations visible with naked eye. The usewear traces are in general much more pronounced than those on slabs (Figure 5). Most used surfaces are covered with reticulate, medium- or high-level polished areas, and sleek striations are common. Like other tool types, these traces are similar to plant processing.

Figure 5.

Xinglonggou grinding stone usewear traces.

A total of 339 starch grains of six types were recovered from handstones, including predominantly Job’s tears (46.2%), followed by millets (20.4%), Triticeae grasses (11.9%), lily and yam (8.2%), and acorn (0.3%). A total of 27 phytoliths were observed in the residues as well. These include Poaceae bulliform cells (33.3%), Type 1 eudicot hair cells (29.6%), Poaceae long cells (22.2%), Type 2 Panicoideae husk fragments (7.4%), and one cross-shaped phytolith consistent with Coix lacryma-jobi.

In summary, the usewear patterns exhibited on the grinding stones and microfossils recovered from their residues are complementary. Usewear traces on all tools are consistent, although not exclusively, with plant processing, including abrading and pounding. The most commonly represented patterns, medium-level polish with fine striations, match to abrading plants, particularly tubers with high fiber content. Starch and phytolith data indicate that slabs and handstones were used for processing many types of plants. The mortar-slabs appear to be more specialized in function, processing mainly tubers, Job’s tears, and millets, as indicated from the starches recovered in their residues. However, all three of the tool types show a preponderance of eudicot hair phytoliths, Asteraceae phytoliths, and abundant tracheary elements in the slab residues. This suggests that all tool types were used to process non-starch bearing plant parts, such as leaves and fibers as well. Sediment samples from contexts adjacent to the artifacts were not available for analysis and comparison; however, the integrity of phytolith and starch samples in artifact residues is not in question (Kealhofer et al., 1999).

Residues from dental calculus

Among the 11 human calculus samples analyzed (Figure S4, available online), 7 yielded a total of 17 starch grains, and 3 yielded five phytoliths. All of the starches show signs of damage, likely caused by chemical and physical processing and cooking. These starch grains are identifiable to USOs (n = 3; 16.7% of the total), Job’s tears (n = 1; 5.6%), and Triticeae grasses (n = 11; 61.1%). The starch granule identified as Job’s tears is rather small and may belong to foxtail millet, as about 20% of the starch grains from these two plants overlap in morphology based on the statistical analysis (Liu et al., 2014). Interestingly, the great majority in number and the most recurrent type belong to Triticeae grasses, most of which are recovered from one individual (M23) (Figure 6; Table 1). Calculus from one individual (M11) yielded a single Pooideae short cell phytolith. Triticeae is a grass tribe within the Pooideae; therefore, this phytolith is consistent with Triticeae starch in the calculus assemblage. The predominance of Triticeae starch from dental calculus differs markedly from its minor proportion revealed in grinding stones. Starch grains from calculus are likely to have undergone gelatinization during cooking, which would have affected starch’s survival. Many factors could have affected the survival rates of starch from different plants after cooking, as gelatinization involves a complex interplay between temperature, moisture content and the presence of solutes, lipids, and proteins, as well as species-specific starch physicochemical properties (Crowther, 2012). Therefore, the proportions of different starch types from calculus may not be used to reconstruct the ratio of starchy foods in human diet at Xinglonggou.

Figure 6.

Starch recovered from Xinglonggou human calculus (1–4), charred plant samples (5–8), and experimentally charred Job’s tears (9, 10) (shown are DIC and polarized views of each granule).

Many plant fibers are present in the calculus samples, an observation that is consistent with the residues from grinding stones. Interestingly, Type 1 eudicot hair cell phytoliths found in three of the calculus samples suggest that the origin of these phytoliths, also abundant in the grinding stone residues, represents a presently unknown processed plant.

Starch from charred plant samples

Two small pieces of charred archaeobotanical remains from Xinglonggou of unknown botanical origin were examined for starch content. A significant number of starch grains of similar morphology were visible in both samples, but also exhibit damage due to heating, that is, gelatinization and morphological change. The starch grains can be grouped into three types: (1) two grains that were round or polygonal in shape with a faint extinction cross; (2) mostly polygonal or round grains with damaged surfaces bearing deep fissures and no visible extinction cross although the grains are florescent under the polarized light; (3) grains with a granulate surface, some having fused to large pieces (Figure 6). These grains resemble Job’s tears starch in general morphology, although their sizes are rather small (5.4–18.1 µm), falling into the lower part of the size range of Job’s tears starch from modern reference and Xinglonggou grinding stones (Figure 3). In order to verify the identification of the charred plant remains, we conducted an experimental study of charring Job’s tears seeds.

Charring Job’s tears experiment

Starch grains can survive in charred plants (Hather, 1993), and experimental studies have shown that controlled temperatures and durations in burning seeds can produce results comparable with ancient charred seeds (Yang et al., 2011). In order to understand the morphological changes of Job’s tears starch during gelatinization process caused by heating, a process which can shed light on identifying Xinglonggou charred plant samples, we conducted an experimental study using furnace to char dry Job’s tears seeds.

We tested three temperature settings (200, 250, and 300°C) in order to establish the ideal range of the heating level. It appears that 250°C is the best temperature for our charring purpose, which is to char the seed without melting it. Four time intervals at 10, 20, 25, and 30 min were used to char the samples. To analyze the starch remains, a part of each charred seed was placed in a plastic bag with a small amount of distilled water. After a few hours, we gently compressed the sample inside the bag to release the remaining starch. The solution was then mounted with 50% glycerol on a glass slide and sealed with nail polish for analysis. The general trend of morphological change of starch during the gelatinization process can be summarized as follows:

Starch grains showed increased damages, with visible damage ranging from deep fissures or micro-pitting to a granulate surface, in the 10 and 20 min of heating.

More grains lost the extinction crosses with prolonged heating; after 25 min of heating, only a few small grains displayed extinction cross.

Starch grains tended to fuse and become carbonized at 25 min of heating, except for a few granules which displayed a damaged surface.

Small grains exhibiting a damaged surface and distorted cross rarely survived 30 min of heating.

The charred Job’s tears starch shows several features which match to those from Xinglonggou charred samples: surviving grains tend to be small in size with damaged surface (particularly deep fissures and granulate surface); some grains are fused; and grains are florescent, but extinction crosses are mostly distorted with only a few still clear in shape. The starch grains found in Job’s tears heated at 250°C between 20 and 30 min best match to the ancient charred samples (Figure 6), suggesting that the ancient Job’s tears seeds were probably heated under similar conditions. These results conform to the Xinglonggou charred plant samples and support their identification as charred Job’s tears.

Discussion

Based on the evidence from usewear traces and microfossil residues from multiple sets of material remains analyzed, Xinglonggou people consumed a considerable variety of plant foods. Three categories of plants are discussed below: USOs, cereals, and acorns.

USOs

Tuber starch grains belong to at least three taxa: lily, yam, and snake gourd roots (n = 209). They account for 25.7% of the total starch and occur on most tools examined (19 tools, 86.4% of the total). Likewise, vascular tissue and fibers likely originating from USO tissue are ubiquitous and appear in the residues of nine, 39%, of the tools analyzed.

It has long been suggested that consumption of USOs is closely related to human evolution (O’Connell et al., 1999), and the earliest evidence for processing USO starch with stone tools can be traced back to the Middle Stone Age in Africa (Mercader et al., 2008), 40,000 years ago in Highland New Guinea (Summerhayes et al., 2010), and 30,000 years ago in Europe (Revedin et al., 2010). In China, starch grains from USOs, such as lily, yam, snake gourd root, and cattail rhizome, have been recovered from Upper Paleolithic sites, dating to 30,000–11,600 cal. BP, at Shuidonggou in Ningxia (Guan et al., 2014) and Shizitan in Shanxi (Liu et al., 2011, 2013). These taxa appear more frequently on grinding stones from early Neolithic sites in North China, such as the Peiligang and Cishan cultures (9000–7000 cal. BP) (Liu, 2015). Yam, lily, and snake gourd root evidently were commonly consumed during the Upper Paleolithic and early Neolithic times in North China, although it is currently unknown whether or not they were also native to the Liao River region prior to the Neolithic. These plants also have been used in traditional medicine (Li, 1981) and as famine food (Zhu, 1406) during the historical period. They are widely cultivated in China today, but the domestication process is unclear. Given that these tubers/roots have been used as foods for millennia prior to the Neolithic period, it is likely that the Xinglonggou people, who were knowledgeable about crop cultivation, had already developed some techniques for managing or even cultivating these plants. Such strategies for the management of tubers and roots may be comparable to those documented in the ethnographic and archaeological record from Australia (Denham, 2011), North and South America (Anderson, 2005; Sheets et al., 2011), and Japan (Crawford, 2011).

Cereals

Starches from cereals are most abundant in the samples analyzed. Among them, Job’s tears accounts for 40% of the starch assemblage, a proportion that is considerably higher than that of millets (20%). The identification of charred Job’s tears starch from macrobotanical samples further confirms the exploitation of this plant by the Xinglonggou population. Phytoliths from Panicoideae grasses were observed in the residues of six tools, 26%, while only two of the phytoliths are consistent with foxtail millet and Job’s tears, respectively.

While foxtail and broomcorn millets are known to be the first cereals cultivated in North China, as early as 10,000–8000 years ago (Yang et al., 2012; Zhao, 2014), Job’s tears is a much less understood grain. Its domestication process is currently unclear. This plant is believed to be native to tropical and subtropical regions of Asia (Arora, 1977; Lim, 2013). Genetic study suggests that southwest China may have been the original center from which several taxonomical variations of Coix evolved (Jiang et al., 2013). It is widely distributed in Asian subtropical and temperate regions today, often cultivated, and used as food, medicine, and decoration (Chen and Phillips, 2006). From Neolithic China, seeds of Job’s tears have been discovered primarily at water-logged sites at a small number of locations along the Yangzi River, dating to ca. 5000–2000 BC (Guedes et al., 2013; Liu and Gu, 2007; Zhejiang Institute of Archaeology, 2003). Starch grains identifiable as Job’s tears, however, have been recovered more frequently from grinding stones and pottery vessels in many Neolithic sites. As evident from starch analysis, Job’s tears first appeared by ca. 6500 BC at Shunshanji in the lower Huai River region (Zhang et al., in press) and became widespread during the 6th millennium BC in Yangzi, Huai, and Yellow River regions (Liu, 2015; Yang and Jiang, 2010). The current study suggests that its distribution had extended to the Liao River valley by the mid-6th millennium BC. The northward movement of Job’s tears coincides with the increased reliance on rice- and/or millet-based agriculture during the mid-Holocene climatic optimum in Neolithic North China. Job’s tears produces much larger seeds than millets do; therefore, it is entirely possible that it was preferred by Neolithic peoples. Xinglonggou is located outside the natural distribution of wild Job’s tears. Most likely, Job’s tears was brought as a cultigen to the region. Situated near a moist river valley, Xinglonggou would have been an ideal location for cultivating Job’s tears. Our identification of Job’s tears here records the northernmost distribution of this plant in the 6th millennium BC and the earliest evidence for Job’s tears cultivation in Northeast China.

Starch grains from millets are lower in proportion (20%) than those from Job’s tears and USOs, but occur on most tools (68%). These patterns suggest that millets were commonly processed for food, but in smaller quantities than other plants, particularly Job’s tears. Millets may be cooked as whole grains for consumption without being ground into flour. Likewise, other starchy foods, such as Job’s tears, yam, and lily, may also be cooked without grinding. These foods occur in greater abundance than millets in the grinding stone residue. Therefore, we cannot assume that millets were much more significant in human diet based on their overall low abundance in the micro-remains.

Both millets and Job’s tears utilize the C4 carbon fixation mechanism, and together they account for 60% of the Xinglonggou starch assemblage. On the other hand, the three tubers, Triticeae grasses, and acorn are species that utilize the C3 carbon fixation mechanism, and they account for 32% of the total starch. These results are consistent with previous stable isotope analysis of human skeletons indicating that C4 foods were predominant in diet (>50%). Rather than assuming millets were the primary source of C4, our research indicates that C4 foods were primarily derived from Job’s tears.

Triticeae starch grains have been found on stone tools (flaked tools and grinding stones) from several Upper Paleolithic and Neolithic sites in North China (Liu, 2015). The indigenous Triticeae grasses, such as Agropyron, Elymus, and Leymus, have never been domesticated, but are commonly used as fodder today (Jiang, 2007). Domesticated wheat and barley, also belonging to the Triticeae tribe, are believed to have been introduced to China from West Asia during the 3rd millennium BC (Barton and An, 2014; Zhao, 2009) and are not considered here. The indigenous wild Triticeae grasses were evidently used for food as early as the Upper Paleolithic in China, and the Xinglonggou people also adapted this tradition.

Acorn

Starch grains from Quercus sp. acorns have been found on grinding stones dating to the Pleistocene–Holocene transition period (Liu et al., 2011) and early Neolithic (Liu, 2015; Tao et al., 2011) in northern China. This nut seems to have been commonly used as food in prehistory. Acorns contain a considerable amount of tannic acid, which needs to be removed by grinding and leaching before consumption (Mason, 1996). However, acorn is unlikely to have been consumed regularly by the Xinglonggou people, given that only 0.3% of starch grains are identifiable as Quercus. The labor-intensive and time-consuming process of acorn may have made this nut undesirable as a staple food when other higher ranked plants (e.g. USOs) were readily available.

In addition, phytoliths from Asteraceae, Boraginaceae, and other eudicot plants may represent additional plants processed with these stone tools. While more specific taxonomic identification of these plants is not possible with the current evidence, many species of Asteraceae and Boraginaceae are reportedly consumed as food or otherwise utilized for medicinal purposes in the region today (Hu, 2005). It is likely that species within these plant families were utilized in the past as well.

Conclusion

It is understood that the transition from hunting and gathering to intensive farming was an extensive and complicated process in human history. However, we feel that there has not yet been sufficient study in China to document the details of plant use and regional variation during this transition. By employing an interdisciplinary approach, the current study demonstrates a complex picture of plant exploitation by the Xinglonggou people who were among the first farmers in Neolithic China.

The consumption of roots and tubers was a Paleolithic tradition and continued for millennia during the early Neolithic. The three types of USOs (lily, yam, and snake gourd root) revealed in Xinglonggou were the most widely distributed. These are commonly utilized as food and medicine in China, pointing to a shared food cultural tradition with a deep history. In addition to domesticated millets, Job’s tears evidently was a major staple food at Xinglonggou. Evidence suggests it was cultivated by early Neolithic peoples. We argue that in addition to millets, Job’s tears may have been one of the earliest domesticated plants utilized in China and that it contributed to the early agricultural systems which would later form the economic backbone of Chinese civilization. More research should be directed to its domestication process, and to distinguish its wild and domesticated forms in macro- and micro-remains.

This study for the first time provides a picture with fine-grained resolution for a wide range of plants in the diet of the early Neolithic people in Northeast China. It will help to develop further research strategies for more studies on long-term changes in plant use during the transition from low-level food production to intensive agriculture in China.

Footnotes

Acknowledgements

We are grateful to Ruichang Wang (Institute of Archaeology, Chinese Academy of Social Sciences) and Hanlong Sun (Zhejinag Provincial Institute of Cultural Relics and Archaeology) who assisted the collection of residue and usewear samples in Chifeng, Inner Mongolia. Zhijun Zhao provided the charred plant remains from Xinglonggou. We thank two anonymous reviewers for their very constructive comments. This research is supported by the Min Kwaan Chinese Archaeology Fund in the Stanford Archaeology Center, Stanford University.

Funding

The authors received no financial support for the research, authorship, and/or publication of this article.

References

Anderson

(2005) Tending the Wild: Native American Knowledge and Management of California’s Natural Resources. Berkeley, CA: University of California Press.

Arora

(1977) Job’s-tears (Coix lacryma-jobi): A minor food and fodder crop of northeastern India. Economic Botany 31: 358–366.

Barton

C-B

(2014) An evaluation of competing hypotheses for the early adoption of wheat in East Asia. World Archaeology 46: 775–798.

Boyadjian

CHC

Eggers

Reinhard

(2007) Dental wash: A problematic method for extracting microfossils from teeth. Journal of Archaeological Science 34: 1622–1628.

Chen

Phillips

(2006) Coix Linnaeus. In: Wu

Raven

Hong

(eds) Flora of China, Volume 22: Poaceae. St Louis, MO: Missouri Botanical Garden, pp. 648–649.

Crawford

(2011) Advances in understanding early agriculture in Japan. Current Anthropology 52: S331–S345.

Crowther

(2012) The differential survival of native starch during cooking and implications for archaeological analyses: A review. Archaeological and Anthropological Sciences 4: 221–235.

Cui

Kong

(1992) Neimenggu zhong dong bu diqu quanxinshi gaowenqi qihou bianhua de chubu fenxi. Zhongguo Quanxinshi Danuanqi. Beijing: Kexue Press, pp. 72–79 (in Chinese).

Denham

(2011) Early agriculture and plant domestication in New Guinea and Island Southeast Asia. Current Anthropology 52: S379–S395.

10.

Fullagar

(2006) Starch on artifacts. In: Torrence

Barton

(eds) Ancient Starch Research. Walnut Creek, CA: Left Coast Press, pp. 177–203.

11.

Fullagar

Liu

Bestel

. (2012) Stone tool-use experiments to determine the function of grinding stones and denticulate sickles. Bulletin of the Indo-Pacific Prehistory Association 32: 29–44.

12.

Liu

Chen

. (2010) Shiwu jiagong guochengzhong dianfenli sunshang ji zai kaoguxue zhong de yingyong [Experimental study of starch damages during food processing and its application in archaeology]. Kaogu 7: 77–86 (in Chinese).

13.

Guan

Pearsall

Gao

. (2014) Plant use activities during the Upper Paleolithic in East Eurasia: Evidence from the Shuidonggou Site, Northwest China. Quaternary International 347: 74–83.

14.

Guedes

Jiang

. (2013) Site of Baodun yields earliest evidence for the spread of rice and foxtail millet agriculture to south-west China. Antiquity 87: 758–771.

15.

Hather

(1993) An Archaeobotanical Guide to Root and Tuber Identification (Oxbow Monograph 28). Oxford: Oxbow Books.

16.

Henry

Piperno

(2008) Using plant microfossils from dental calculus to recover human diet: A case study from Tell al-Raqa’i, Syria. Journal of Archaeological Science 35: 1943–1950.

17.

Henry

Hudson

Piperno

(2008) Changes in starch grain morphologies from cooking. Journal of Archaeological Science 36: 915–922.

18.

S-y

(2005) Food Plants of China. Hong Kong: The Chinese University Press.

19.

Jiang

(ed.) (2007) Zhongguo Zuowu jiqi Yesheng Jinyuan Zhiwu: Siyong ji Lufei Zuowujuan (Crops and Their Wild Relatives in China: Vol. Forage and Manure Crops). Beijing: China Agriculture Press (in Chinese).

20.

Jiang

Z-d

Guo

J-h

Chen

Q-f

(2013) Yiyi shu zhiwu DNA duoyangxing fenxi (DNA diversity analysis of Coix L.). Guangdong Nongye Kexue 2: 124–142 (in Chinese).

21.

Kealhofer

Piperno

(1998) Opal Phytoliths in Southeast Asian Flora (Smithsonian Contributions to Botany, No. 88). Washington, DC: Smithsonian Institution Press.

22.

Kealhofer

Torrence

Fullagar

(1999) Integrating phytoliths within use-wear/residue studies of stone tools. Journal of Archaeological Science 26: 527–546.

23.

(1981) Bencao Gangmu (Materia Medica, Arranged according to Drug Descriptions and Technical Aspects). Beijing: Public Health Press.

24.

Lim

(2013) Coix lachryma-jobi. In: Lim

(ed.) Edible Medicinal and Non-Medicinal Plants: Volume 5, Fruits. New York: Springer Science + Business Media Dordrecht, pp. 243–261.

25.

Liu

(2007) Chengtoushan yizhi de zhiwu yicun. In: Hunan Institute of Archaeology and Cultural Relics and International Research Center of Japanese Culture (eds) Chengtoushan in Lixian. Beijing: Wenwu Press, pp. 98–106 (in Chinese).

26.

Liu

Jia

Zhao

. (2004) Excavation of the Xinglonggou settlement site in Chifeng city, Inner Mongolia, in 2002-2003. Kaogu 7: 3–8 (in Chinese).

27.

Liu

(2015) A long process towards agriculture in the Middle Yellow River valley, China: Evidence from macro- and micro-botanical remains. Journal of Indo-Pacific Archaeology 35: 3–14.

28.

Liu

Bestel

Shi

. (2013) Paleolithic human exploitation of plant foods during the last glacial maximum in North China. Proceedings of the National Academy of Sciences 110: 5380–5385.

29.

Liu

Field

Fullagar

. (2010a) A functional analysis of grinding stones from Donghulin, north China. Journal of Archaeological Science 37: 2630–2639.

30.

Liu

Field

Fullagar

. (2010b) What did grinding stones grind? New light on Early Neolithic subsistence economy in the Middle Yellow River Valley, China. Antiquity 84: 816–833.

31.

Liu

Bestel

. (2011) Plant exploitation of the last foragers at Shizitan in the Middle Yellow River Valley China: Evidence from grinding stones. Journal of Archaeological Science 38: 3524–3532.

32.

Liu

Cui

(2014) Identification of starch granules using a two-step identification method. Journal of Archaeological Science 52: 421–427.

33.

Liu

Jones

Zhao

. (2012) The earliest evidence of millet as a staple crop: New light on Neolithic foodways in north China. American Journal of Physical Anthropology 149: 283–290.

34.

Loy

Fullagar

(2006) Residue extraction. In: Torrence

Barton

(eds) Ancient Starch Research. Walnut Creek, CA: Left Coast Press, pp. 197–198.

35.

Liu

K-B

(2003) Morphological variations of lobate phytoliths from grasses in China and the south-eastern United States. Diversity and Distributions 9: 73–87.

36.

H-Y

N-Q

Yang

X-D

. (2006) Phytoliths as quantitative indicators for the reconstruction of past environmental conditions in China I: Phytolith-based transfer functions. Quaternary Science Reviews 25: 945–959.

37.

(ed.) (1980) Neimenggu Zhiwuzhi: Di Wu Juan (Flora of Inner Mongolia: Volume 5). Hohhot: Neimenggu Renmin Press (in Chinese).

38.

(ed.) (1983) Neimenggu Zhiwuzhi: Di Qi Juan (Flora of Inner Mongolia: Volume 7). Hohhot: Neimenggu Renmin Press (in Chinese).

39.

(ed.) (1985a) Neimenggu Zhiwuzhi: Di Ba Juan (Flora of Inner Mongolia: Volume 8). Hohhot: Neimenggu Renmin Press (in Chinese).

40.

(ed.) (1985b) Neimenggu Zhiwuzhi: Di Yi Juan (Flora of Inner Mongolia: Volume 1). Hohhot: Neimenggu Renmin Press (in Chinese).

41.

Madella

Lancelotti

García-Granero

(2013) Millet microremains – An alternative approach to understand cultivation and use of critical crops in Prehistory. Archaeological and Anthropological Sciences. Epub ahead of print 27 February. DOI: 10.1007/s12520-013-0130-y.

42.

Mason

SLR

(1996) Acornutopia? Determining the role of acorns in past human subsistence. In: Wilkins

Harvey

Dobson

(eds) Food in Antiquity. Exeter: University of Exeter Press, pp. 12–24.

43.

Mercader

Bennett

Raja

(2008) Middle Stone Age starch acquisition in the Niassa Rift, Mozambique. Quaternary Research 70: 283–300.

44.

O’Connell

Hawkes

Jones

NGB

(1999) Grandmothering and the evolution of Homo erectus. Journal of Human Evolution 36: 461–485.

45.

Pearsall

(2000) Paleoethnobotany: A Handbook of Procedures. San Diego, CA: Academic Press.

46.

Piperno

(2006) Phytoliths: A Comprehensive Guide for Archaeologists and Paleoecologists. Oxford: AltaMira Press.

47.

Revedin

Aranguren

Becattini

. (2010) Thirty thousand-year-old evidence of plant food processing. Proceedings of the National Academy of Sciences 107: 18815–18819.

48.

Sheets

Dixon

Guerra

. (2011) Manioc cultivation at Ceren, El Salvador: Occasional kitchen garden plant or staple crop? Ancient Mesoamerica 22: 1–11.

49.

Summerhayes

Leavesley

Fairbairn

. (2010) Human adaptation and plant use in Highland New Guinea 49,000 to 44,000 years ago. Science 330: 78–81.

50.

Sun

(2014) Research on plant remains from Neolithic to early bronze age in Upper West Liao River Region. PhD Dissertation, Inner Mongolia Normal University.

51.

Tao

Guo

. (2011) Starch grain analysis for groundstone tools from Neolithic Baiyinchanghan site: Implications for their function in Northeast China. Journal of Archaeological Science 38: 3577–3583.

52.

Twiss

Suess

Smith

(1969) Morphological classification of grass phytoliths. Proceedings – Soil Science Society of America 33: 109–115.

53.

Weisskopf

Lee

G-a

(2014) Phytolith identification criteria for foxtail and broomcorn millets: A new approach to calculating crop ratios. Archaeological and Anthropological Sciences. Epub ahead of print 16 March. DOI: 10.1007/s12520-014-0190-7.

54.

Winkler

Wang

(1993) The Late-Quaternary vegetation and climate of China. In: Wright

Jr Kutzbach

Webb

III . (eds) Global Climates since the Last Glacial Maximum. Minnesota, MN: University of Minnesota Press, pp. 221–264.

55.

Raven

(eds) (1999) Flora of China, Volume 4: Cycadaceae through Fagaceae. Beijing and St. Louis, MO: Science Press and Missouri Botanical Garden Press.

56.

Raven

(eds) (2000) Flora of China, Volume 24: Flagellariaceae through Marantaceae. Beijing and St. Louis, MO: Science Press and Missouri Botanical Garden Press.

57.

Raven

Hong

(eds) (2006) Flora of China, Volume 22: Poaceae. Beijing and St. Louis, MO: Science Press and Missouri Botanical Garden Press.

58.

Raven

Hong

(eds) (2011) Flora of China, Volume 19: Cucurbitaceae through Valerianaceae, with Annonaceae and Berberidaceae. Beijing and St. Louis, MO: Science Press and Missouri Botanical Garden Press.

59.

Q-h

Yang

Z-j

Cui

Z-ij

. (2002) Pollen analysis of Qiguoshan Section and the living environment of ancient peoples in the Chifeng Area, Inner Mongolia. Scientia Geographica Sinica 22: 453–457.

60.

Yang

Liu

Shao

(2000) Neimenggou Aohanqi Xinglonggou xinshiqi shidai yizhi diaocha [Survey of the Xinglonggou Neolithic site in Aohan banner, Inner Mongolia]. Kaogu 9: 30–48 (in Chinese).

61.

Yang

Liang

. (1996) Dongbei Baiheshu zhiwu jiqi zaipei liyong [Cultivation and use of Lilium plants in Northeast]. Guotu yu Ziran Ziyuan Yanjiu 2: 60–62 (in Chinese).

62.

Yang

Zhou

. (2011) Investigation of the ultrastructural characteristics of foxtail and broomcorn millet during carbonization and its application in archaeobotany. Chinese Science Bulletin 56: 1495–1502.

63.

Yang

Jiang

(2010) Starch grain analysis reveals ancient diet at Kuahuqiao site, Zhejiang Province. Chinese Science Bulletin 55: 1150–1156.

64.

Yang

Kong

Liu

(2009) Zhongguo beifang zhuyao jianguolei dianfenli xingtai duibi [Morphological comparison of starch grains of major types of nuts from north China]. Quaternary Sciences 29: 153–158 (in Chinese).

65.

Yang

Wan

Perry

. (2012) Early millet use in northern China. Proceedings of the National Academy of Sciences 109: 3726–3730.

66.

Zhang

Yin

. (in press) Plant resource utilization at Shunshanji site in Jiangsu Province based on plant remains analysis. Dongfang Kaogu (in Chinese).

67.

Zhang

. (2011) Phytolith analysis for differentiating between foxtail millet (Setaria italica) and green foxtail (Setaria viridis). PLoS ONE 6: e19726.

68.

Zhao

(2004) Cong Xinglonggou yizhi fuxuan jieguo tan Zhongguo beifang hanzuo nongye qiyuan wenti [The origins of dry-land agriculture in north China based on the flotation results from the Xinglonggou site]. In: Nanjing Normal University (ed.) Dongya Kaogu. Beijing: Wenwu Press, pp. 188–199 (in Chinese).

69.

Zhao

(2009) Eastward spread of wheat into China – New data and new issues. Chinese Archaeology 9: 1–9.

70.

Zhao

(2011) New archaeobotanic data for the study of the origins of agriculture in China. Current Anthropology 52: S295–S306.

71.

Zhao

(2014) Zhongguo gudai nongye de xingcheng guocheng – Fuxuan chutu zhiwu yicun zhengju [The process of origin of agriculture in China: Archaeological evidence from flotation results]. Quaternary Sciences 34: 73–84 (in Chinese).

72.

Zhejiang Institute of Archaeology (ed.) (2003) Hemudu. Beijing: Wenwu Press (in Chinese).

73.

Zhu

(ed.) (1406) Jiuhuang Bencao (Herbal for Relief of Famines). Taipei: Shangwu Press (in Chinese).