Content uploaded by Jiajing Wang
Author content
All content in this area was uploaded by Jiajing Wang on May 26, 2016
Content may be subject to copyright.
Revealing a 5,000-y-old beer recipe in China
Jiajing Wang
a,b,1
, Li Liu
a,b
, Terry Ball
c
, Linjie Yu
d
, Yuanqing Li
e
, and Fulai Xing
f
a
Stanford Archaeology Center, Stanford University, Stanford, CA 94305;
b
Department of East Asian Languages and Cultures, Stanford University, Stanford,
CA 94305;
c
Department of Ancient Scripture, Brigham Young University, Provo, UT 84602;
d
Zhejiang Research Institute of Chemical Industry, 310006
Hangzhou, China;
e
Department of Civil and Environmental Engineering, Stanford University, Stanford, CA 94305; and
f
Shaanxi Provincial Institute of
Archaeology, 710054 Xi’an, China
Edited by Dolores R. Piperno, Smithsonian Institution, Fairfax, VA, and approved April 26, 2016 (received for review January 27, 2016)
The pottery vessels from the Mijiaya site reveal, to our knowledge,
the first direct evidence of in situ beer making in China, based on
the analyses of starch, phytolith, and chemical residues. Our data
reveal a surprising beer recipe in which broomcorn millet (Panicum
miliaceum), barley (Hordeum vulgare), Job’s tears (Coix lacryma-
jobi), and tubers were fermented together. The results indicate
that people in China established advanced beer-brewing technol-
ogy by using specialized tools and creating favorable fermenta-
tion conditions around 5,000 y ago. Our findings imply that early
beer making may have motivated the initial translocation of bar-
ley from the Western Eurasia into the Central Plain of China before
the crop became a part of agricultural subsistence in the region
3,000 y later.
Yangshao period
|
alcohol
|
starch analysis
|
phytolith analysis
|
archaeological chemistry
In China, the earliest written record of beer appears in oracle bone
inscriptions from the late Shang dynasty (ca. 1250–1046 BC)
(1, 2). According to the inscriptions, the Shang people used malted
grains, including millets and barley/wheat (barley and wheat
are represented by the same Chinese character, lai)asthemain
brewing ingredients (1, 3). Scholars have hypothesized that the
Shang tradition of beer brewing has its origin in the Neolithic
Yangshao period (5000–2900 BC), when large-scale agricultural
villages were established in the Yellow River valley (4–6). The
hypothesis is possible, considering that China has an early tradition
of fermentation and evidence of rice-based fermented beverage has
been found from the 9,000-y-old Jiahu site (7). Certain types of
Yangshao vessels, including funnels and jiandiping (pointed-bottom
vessel) amphorae, show stylistic similarities to the brewing equip-
ment in the historical period and modern ethnographic records (6).
However, direct evidence for alcohol production from Yangshao
sitesislacking.
The Mijiaya site is located on a primary terrace northeast of
the Chan River, a tributary of the Wei River, in Shaanxi, North
China (Fig. 1 and SI Text). The excavations revealed two sub-
terranean pits with artifacts that appeared to resemble in situ
beer-brewing facilities (8) (Fig. 2 and Fig. S1). Both pits belong
to Banpo IV (or late Yangshao period) stratum, and an estab-
lished chronology based on ceramic typology and
14
Cdatesin
the region securely places the time period between 3400 and
2900 BC (9, 10). Pit H82 was 3.7-m deep with straight walls and
five steps leading down to the bottom (Fig. 2A); Pit H78 was
2.5–2.7 m deep with a flat bottom and a secondary platform on
one side of the walls (8). Three types of vessels were recovered in
both pits: wide-mouth pots, funnels, and jiandiping amphorae (Fig.
2B–D), all of which have yellowish residues on their interior
surface (Fig. S2). The shapes and styles of the vessels suggest three
distinctive stages in the beer-making process: brewing, filtration,
and storage. Interestingly, each pit also contained a pottery stove
(Fig. 2E). In brewing activity, heating equipment is often used to
maintain the optimal temperature for mashing. The stoves would
have been especially suitable for this operation. Our hypothesis is
that the artifact assemblages exclusively from the two pits repre-
sent a “beer-making toolkit.”To test our hypothesis, we conducted
starch, phytolith, and chemical analyses on the residues from two
complete funnels and pottery sherds from five jiandiping ampho-
rae and two wide-mouth pots.
Taxonomical Identification
The identification of starch grains and phytoliths was based on
morphological and morphometric analyses. The morphological
identification relied on a reference collection from over 1,000 Asian
and European economically important plant specimens, and by
consultation with published literature (11–16) (Table S1). For
morphometric analysis, we applied two computer-assisted methods.
First, a discriminant analysis model was used to separate the starch
grains of Job’stears(Coix lacryma-jobi ) from those of millets
(Setaria italica and Panicum miliaceum). Based on three variables,
including size, eccentricity of the hilum, and presence or absence of
curved arms on the extinction cross, this multivariate model has a
success rate of 82.4% for separating Job’s tears from millets (17).
Second, we conducted a morphometric analysis of the articulated
dendritic phytoliths. Dendritics are produced in the inflorescence
bracts of many common cereals, especially Triticeae species. Ar-
ticulated dendritics produce wave patterns of taxonomic signifi-
cance. Recent research has achieved good success at identifying
wheat and barley phytoliths and distinguishing them from the rel-
evant wild grasses (12, 13). A sample size as small as 30 dendritic
wave lobes can be adequate for a 90% confidence level in dis-
criminating among taxa based on differences in several wave lobe-
shaped morphometries. We measured the wave lobes of articulated
dendritic phytoliths from six Mijiaya residue samples following the
procedures outlined by Ball et al. (18). Each residue sample yielded
at least 30 measurable wave lobes in total, which allowed us to have
statistical confidence in the measurements of form factor, round-
ness, convexity, solidity, compactness, and aspect ratio (Table S2).
We compared the means of these six morphometries for the Mijiaya
dendritics from each sample with the range of means observed in 20
Triticeae and other dendritic-producing species, including common
Significance
This research reveals a 5,000-y-old beer recipe in which broomcorn
millet, barley, Job’s tears, and tubers were fermented together. To
our knowledge, our data provide the earliest direct evidence of in
situ beer production in China, showing that an advanced beer-
brewing technique was established around 5,000 y ago. For the
first time, to our knowledge, we are able to identify the presence
of barley in archaeological materials from China by applying a
recently developed method based on phytolith morphometrics,
predating macrobotanical remains of barley by 1,000 y. Our
method successfully distinguishes the phytoliths of barley from
those of its relative species in China.
Author contributions: J.W. and L.L. designed research; J.W., L.L., T.B., and F.X. performed
research; J.W., L.L., T.B., L.Y., and Y.L. analyzed data; and J.W., L.L., and T.B. wrote
the paper.
The authors declare no conflict of interest.
This article is a PNAS Direct Submission.
1
To whom correspondence should be addressed. Email: jiajingw@stanford.edu.
This article contains supporting information online at www.pnas.org/lookup/suppl/doi:10.
1073/pnas.1601465113/-/DCSupplemental.
www.pnas.org/cgi/doi/10.1073/pnas.1601465113 PNAS Early Edition
|
1of5
ANTHROPOLOGY
wild species indigenous to China and domesticated species in-
troduced into China from Western Eurasia (Tables S3 and S4).
Results
A total of 541 starch grains were recovered from the funnels and
pottery sherds (Table 1). Of these grains, 488 starch grains (90.2%
of the total) were identifiable to various taxonomic levels com-
pared with our reference data. The starch assemblage mainly
consists of millet, Triticeae, and Job’stears,withasmallercontri-
bution from tubers that include snake gourd root (Trichosanthes
kirilowii), yam (Dioscorea sp.), and lily (Lilium sp.) (Fig. 3 A–F). A
high percentage of starch grains (n=166, 30.7%) exhibit signs of
damage, and among them two types of damage closely resemble
that produced by beer brewing. First, some grains show pits and
channels on their surface, ranging from being slightly pitted to
completely hollow (Fig. 4A). Second, a large number of starch
grains are swollen, folded, and distorted (Fig. 4B); some still retain
their individual boundaries, but many merge into one another.
These two damage patterns precisely match the morphological
changes developed during malting and mashing, as we observed in
our brewing experiments and that reported in published literature
(Fig. 4 Cand D,Fig. S3,andSI Materials and Methods)(19–22).
During malting, enzymes in sprouting cereals break down starch to
dextrins and simple sugars, creating typical surface pits and interior
channels in the grains (19). Furthermore, mashing involves heating
of the malt in water for a period, which causes gelatinization,
swelling, folding, and distortion of starch grains (19). Thus, the
damaged state of the starch grains in our archeological sample
provides strong evidence for the conclusion that those starch grains
are residues from the brewing process.
Phytolith data indicate the presence of cereal husks (Table 2).
Phytoliths identified to the Panicoideae grass subfamily pre-
dominate the assemblage. Seven vessels revealed η-shaped long
cells that are consistent with phytoliths formed in epidermal husk
tissues of broomcorn millet (Fig. 3J) (15). Cross-shaped phytoliths
showing a considerable variation in form and size, comparable with
the crosses produced by Job’s tears in our modern references, were
observed (Fig. 3H). Phytolith forms produced in Pooideae were
also identified. In particular, articulated dendritic phytoliths con-
sistent in pattern and shape with those produced in the husks of
Triticeae species were observed (Fig. 3K). Our morphometric
analysis of the articulated dendritics indicates their most probable
origin to be from the inflorescence bracts of barley (Hordeum vul-
gare). Although some of the mean measurements of the archaeo-
logical samples fall within the ranges of other taxa, all of the means
fall within those observed in H. vulgare (Tables S5 and S6). The
profile of phytoliths corroborates the starch grain assemblage, in-
dicating the presence of broomcorn millet, Job’s tears, and barley.
Our ion chromatographic (IC) analysis identified the presence of
oxalate, which develops during the steeping, mashing, and fer-
mentation of cereals (23). Calcium oxalate is a principle component
of “beerstone,”whichsettlesoutatthebeer fermentation and
storage containers, and has been used as a compound marker for
identifying barley beer fermentation in ancient vessels (24–27). We
conducted tests on the residues from Funnel 1, Pot 3, and Pot 5.
The results confirmed that high levels of oxalate are present in
Funnel 1 and Pot 5 (Fig. 5). Oxalate was not detected from Pot 3.
The oxalate concentration is 0.08% (80 mg/100 g) for Funnel 1 and
0.05% (50 mg/100 g) for Pot 5. Although oxalates occur naturally in
plants like spinach (Spinacia oleracea), rhubarb (Rheum rhabarba-
rum), and some Dioscorea and Lilium species, the vessel types in
this study are not suitable for storing any fresh plant.
To rule out the potential contamination from the enclosing soil
matrix and postexcavation conditions, we analyzed four samples as
control specimens. Three samples were analyzed for starch and
phytolith, including one from the sediment adhering to the exterior
surface of Pot 5 (control sample 1), one from a stone adze fragment
recovered from Pit H78 (control sample 2), and one from the
plaster material used for reconstructing Funnel 2 (control sample 3)
(Fig. S1). Compared with ancient samples, the control samples
contained a much smaller number of phytoliths and starch grains
(Tables 1 and 2) with no signs of damage. IC analysis of the exterior
scrapings from Pot 5 (control sample 4) detected no oxalate, a
result in clear contrast to a high level of oxalate found in the in-
terior surface of the same vessel. These observations suggest that
the high concentrations of starch, phytolith, and chemical residues
related to the beer-making process are exclusively present on the
interior surfaces of the Mijiaya vessels, which potentially have had
direct association with beer brewing.
Discussion and Conclusion
All three lines of evidence are consistent with the archaeological
data, indicating that the Yangshao people brewed a mixed beer with
specialized tools and knowledge of temperature control. Our data
show that the Yangshao people developed a complicated fermen-
tation method by malting and mashing different starchy plants.
Compared with millets, barley has much higher α-andβ-amylase
activities, which promotes the saccharification process (28). Tubers
Fig. 1. Geographical Location of the Mijiaya Site.
Fig. 2. The “beer-making toolkit”from Mijiaya Pit H82: (A) Pit H82 illustration
in top and cross-section views, (B) funnel 1, (C) pot 6 in reconstructed form,
(D) pot 3 in reconstructed form, and (E) pottery stove.
2of5
|
www.pnas.org/cgi/doi/10.1073/pnas.1601465113 Wang et al.
contribute starch and sugars for fermentation, and they also add a
sweeter flavor to the beer. The Yangshao people probably de-
veloped their recipe through repeated experiments.
The discovery of barley in the beer residues suggests a social
motivation in the initial stage of crop translocation (29). Barley
was first domesticated in Western Eurasia and later introduced
into China, presumably through the Central Asian steppe. The
timing and nature of the crop’s initial adoption in China is still
not well understood (29–31). In the Central Plain, macro-
botanical remains from the Yangshao sites are generally well
preserved, dominated by millets and very few other cereal
types, and no evidence of barley has been reported. The ear-
liest evidence of barley comes from some sporadic finds in
Bronze Age sites, all dated around or after 2000 BC (31, 32).
Not until the Han dynasty (206 BC–AD 220), three millennia
after Mijiaya, had this crop become an important part of
Table 1. Counts of starch grain types from vessels
Artifact no. Triticeae
Job’s tears,
C. lacryma-jobi
Millet,
P. miliaceum
Snake gourd
root, T. kirilowii
Lily,
Lilium sp.
Yam,
Dioscorea sp. Tuber* UNID Total Damaged
Funnel 1 24 49 78 0 0 0 0 7 158 47
Funnel 2 23 13 11 4 0 1 2 4 58 19
Pot 1 17 6 3 1 0 0 10 3 40 6
Pot 2 7 3 3 0 0 1 0 1 15 14
Pot 3 11 5 11 0 0 0 1 9 38 23
Pot 4 39 17 23 2 3 1 6 6 98 20
Pot 5 11 19 14 9 0 0 7 4 66 17
Pot 6 14 10 7 0 1 0 4 14 50 15
Pot 7 6 6 4 0 0 0 1 5 24 5
Total 152 128 154 16 4 3 31 53 541 166
Percent, % 28.61 23.7 28.5 3.0 0.7 0.6 5.7 9.8 100.0 30.7
Control sample 1 1 0 0 0 0 0 0 2 3 0
Control sample 2 3 0 0 0 0 0 0 3 6 0
Control sample 3 0 0 0 0 0 0 1 3 4 0
*A general category that includes snake gourd root, lily, and yam. UNID, unidentified.
Fig. 3. Starch and phytolith types from Mijiaya vessels (the starch types showing DIC and polarized views): (A) Broomcorn millet (P. miliaceum). (B) Triticeae. (C)Job’s
tears (C. lacryma-jobi). (D) Snake gourd root (T. kirilowii). (E)Yam(Dioscorea sp.). (F)Lily(Lilium sp.). (G) Bilobate. (H) Cross. (I) Rondel. (J)η-shaped phytoliths, consistent
with broomcorn millet. (K) Dendritic epidermal phytoliths, consistent with barley (H. vulgare). (Scale bars: 10 μminA,H,andI;20μmInB–G;50μminJand K.)
Wang et al. PNAS Early Edition
|
3of5
ANTHROPOLOGY
subsistence in the Central Plain (29, 33). The microbotanical
remains of barley at Mijiaya account for the earliest occur-
rence of this crop in China.
It is possible that the few rare finds of barley in the Central
Plain during the Bronze Age indicate their earlier introduction
as rare, exotic food. The Mijiaya farmers probably obtained small
quantities of barley grains through exchange or cultivated the
plant along with other cereals. We suggest that barley was ini-
tially introduced to the Central Plain as an ingredient for alcohol
production rather than for subsistence. Because to our knowl-
edge this is the first study that applies morphometric analysis to
the dendritic phytoliths from China, future research with more
comprehensive phytolith data from other Neolithic contexts is
needed to test our hypothesis.
The practice of beer brewing is likely to have been associated
with the increased social complexity in the Central Plain during
the fourth millennium BC The late Yangshao period in the Wei
River region was characterized by hierarchically organized
settlement patterns, interpolity competitions, construction of
large public architectures at regional centers, and ritual feasting
likely organized by elite individuals and involving alcohol
consumption (34) (SI Text). Like other alcoholic beverages,
beer is one of the most widely used and versatile drugs in the
world (35), and it has been used for negotiating different kinds
of social relationships. The coincidence of beer production with
other lines of material evidence suggests that competitive
feasting was actively developing. The production and con-
sumption of Yangshao beer may have contributed to the
emergence of hierarchical societies in the Central Plain, the
region known as “the cradle of Chinese civilization.”
Materials and Methods
Residue Extraction Methods Summary. Chemical samples and two control
samples (1, 3) were obtained by scraping off the sediments from the pottery
surfaces with clean blades. Other residue samples were extracted by using an
ultrasound bath or an ultrasound toothbrush. Starch and phytolith samples
were floated from the residues using the heavy liquid sodium polytungstate
at a specific gravity of 2.35. Extractions obtained from the residue samples
were mounted in 50% (vol/vol) glycerol and 50% (vol/vol) distilled water on
glass slides and scanned under a Zeiss Axio Scope A1 fitted with polarizing
filters and differential interference contrast (DIC) optics, at 200×and 400×
for both starches and phytoliths. Images were taken using a Zeiss Axiocam
Hrc3 digital camera and Zeiss Axiovision software v4.8. See SI Materials and
Methods for details.
Determination of Oxalate Using IC. Residue samples from Funnel 1, Pot 3, and
Pot 5 were analyzed by a Dionex ICS 5000 with a conductivity detector, in
Zhejiang Research Institute of Chemical Industry. The columns used were an
Ion Pac-AS 11-HC Analytical (250 ×4 mm I.D.) and an Ion Pac AG11-HC guard
column (40 ×4 mm I.D.) with an anion ASRS suppressor, operated at 20 °C.
The eluent was 20 mM KOH and the flow rate was 1.5 mL/min; the injection
loop was 25 μL. Each residue sample produced two experimental samples.
Each experimental sample (0.03 g) was dissolved in 3 mL nitric acid and di-
luted with 50 mL distilled water. Standard solutions were made from
solutions of sodium oxalate in a concentration range between 0.0005 and
0.002 mg/mL. The calibration curve was established by linear regression
analysis of peak height vs. added concentration of oxalate. The detection
limit for oxalate under the given conditions is 0.036%. Control sample 4
was analyzed by a Dionex DX-500 IC with a conductivity detector, in the
Environmental Measurements Facility at Stanford University. The eluent
was 20 mM NAOH and the flow rate was 1 mL/min. Standardized solutions
were made from solutions of oxalic acid in a concentration range between
0.1 and 10 mg/mL. Other laboratory conditions were the same as those
used for the three residue samples.
ACKNOWLEDGMENTS. We thank Dr. Maureece Levin, Mike Bonomo, and
David Hazard for their comments on previous drafts of the paper;
Dr. Zhouyong Sun for making arrangement and providing access to the data;
Fig. 4. Damaged starch grains from Mijiaya vessels and brewing experiments.
(A) Mijiaya starch grains showing pitting and channeling. (B) A Mijiaya starch
grain showing swollen, folded, and distorted characteristics. (C)Fermented
broomcorn millet (P. miliaceum) starch grains from the brewing experiment
using broomcorn millet and barley (H. vulgare). (D) A fermented gelatinized
starch grain from the brewing experiment using broomcorn millet and barley.
(Scale bars: 10 μminAand C;20μminBand D.)
Table 2. Counts of phytolith types from vessels
Phytolith types
Taxonomic
association
Funnel
1
Funnel
2 Pot 1 Pot 2 Pot 3 Pot 4 Pot 5 Pot 6 Pot 7 Total
Control
sample 1
Control
sample 2
Control
sample 3
η-Type epidermal
sheet element
(P. miliaceum)4 3 0 313506 252 0 0
Cross (Panicoideae) 37 23 26 813263524535 394 0 2 0
Bilobate (Panicoideae) 49 18 16 622624502215 282 0 2 0
Polylobate (Panicoideae)3221421415144010
Articulated
dendritic
(Triticeae) 2 11 0 76332427 4 0 177 0 0 0
Undulate
trapezoid
(Pooideae) 4 4 0 2 7 4 3 3 0 27 0 0 0
Epidermal sheet
element
(Poaceae) 0 14 4 12121211 611 82 4 0 3
Bulliform (Poaceae) 4 3 5 013230 210 0 0
Rondel (Poaceae) 14 2 4 424613 400 1 0
Total 117 80 57 254 116 151 157 89 71 1,092 6 6 3
4of5
|
www.pnas.org/cgi/doi/10.1073/pnas.1601465113 Wang et al.
Hao Zhao for assisting in the collection of residue samples; Ganrong Wang
and Lijing Zheng for providing help in the ion chromatography analysis; and
two reviewers for their constructive comments. This research was supported
by the Min Kwaan Chinese Archaeology Fund from the Stanford Archaeol-
ogy Center, a summer research grant from Center for East Asian Studies, a
travel grant from Stanford Archaeology Center, and a private donor.
1. Wen S, Yuan T (1983) Yinxu Buci Yanjiu: Kexue Jishu Pian (A Study of Oracle Bones
from Yinxu: Science and Technology) (Sichuansheng Shehui Kexueyuan, Chengdu,
China), 1st Ed.
2. Zhang D (1994) Yinshang jiuwenhua chulun (A preliminary study of Shang alcohol
culture). Zhongyuan Wenwu 3:19–24.
3. Chen M (1956) Yinxu Buci Zongshu (A Comprehensive Study of the Oracle Bone
Inscriptions from Yinxu) (Kexue, Beijing).
4. Huang HT (2000) Science and Civilisation in China. Vol 6: Biology and Biological
Technology. Part V: Fermentations and Food Science. (Cambridge Univ Press, Cam-
bridge, UK).
5. Li Y (1962) Dui woguo niangjiu qiyuan de tantao (The origin of alcoholic beverages in
China). Kaogu 1:41–44.
6. Bao Q, Zhou J (2007) Zhongguo Chuantong Gongyi Quanji: Niangzao (Chinese
Traditional Technology: Brewing) (Daxiang, Zhengzhou, China), 1st Ed.
7. McGovern PE, et al. (2004) Fermented beverages of pre- and proto-historic China. Proc
Natl Acad Sci USA 101(51):17593–17598.
8. Shaanxi Provincial Institute of Archaeology (2012) Xi’an Mijiaya–Xinshiqi Shidai Yizhi
2004-2006 Kaogu Fajue Baogao (2004–2006 Excavation Report of the Neolithic Site
Mijiaya in Xi’an) (Kexue, Beijing).
9. Institute of Archaeology, Chinese Academy of Social Sciences (1991) Zhongguo Kaoguxue
Zhong Tanshisi Niandai Shu juji (Radiocarbon Dates in Chinese Ar chaeology 1965–1991)
(Cultural Relics Publishing House, Beijing).
10. Yang Y (2013) Shaanxi Yangguanzhai Yizhi Yangshao Wenhua Zhongwanqi Qihou
Huanjing Jilu Ji Yizhi Gurenlei Dui Quyu Huanjin Yicun (The Climatic Record and
Human-environmental Relations from Middle to Late Yangshao Period: A Case Study
at Yangguanzhai, Shaanxi). Master’s thesis (Xibei University, Xi’an, China).
11. Piperno DR (2006) Phytoliths: A Comprehensive Guide for Archaeologists and
Paleoecologists (AltaMira Press, Lanham, MD).
12. Ball T, Gardner JS, Brotherson JD (1996) Identifying phytoliths produced by the in-
florescence bracts of three species of wheat (Triticum monococcum L.,T. dicoccon
Schrank., and T. aestivum L.) using computer-assisted image and statistical analyses.
J Archaeol Sci 23(4):619–632.
13. Ball TB, Gardner JS, Anderson N (1999) Identifying inflorescence phytoliths from se-
lected species of wheat (Triticum monococcum,T. dicoccon,T. dicoccoides, and
T. aestivum) and barley (Hordeum vulgare and H. spontaneum) (Gramineae). Am J
Bot 86(11):1615–1623.
14. Ball TB, Ehlers R, Standing MD (2009) Review of typologic and morphometric analysis
of phytoliths produced by wheat and barley. Breed Sci Jpn 59(4):55–102.
15. Lu H, et al. (2009) Phytoliths analysis for the discrimination of Foxtail millet (Setaria
italica) and Common millet (Panicum miliaceum). PLoS One 4(2):e4448.
16. Weisskopf AR, Lee G-A (2014) Phytolith identification cri teria for foxtail and broomcorn
millets: A new approach to calculating crop ratios. Archaeol Anthropol Sci 8(1):29–42.
17. Liu L, Ma S, Cui J (2014) Identification of starch granules using a two-step identifi-
cation method. J Archaeol Sci 52:421–427.
18. Ball TB, et al. (Dec ember 2015) A morphometr ic study of variance in art iculated
dendritic phytolith wa ve lobes within selected sp ecies of Triticeae and Avenea e. Veg
Hist Archaeobotany , 10.1007/s00334-015 -0551-x.
19. Samuel D (1996) Investigation of Ancient Egyptian baking and brewing methods by
correlative microscopy. Science 273(5274):488–490.
20. Dronzek BL, Hwang P, Bushuk W (1972) Scanning electron microscopy of starch from
sprouted wheat. Cereal Chem 49:232–239.
21. Sun Z, Henson CA (1990) Degradation of native starch granules by barley α-glucosi-
dases. Plant Physiol 94(1):320–327.
22. Henry AG, Hudson HF, Piperno DR (2009) Changes in starch grain morphologies from
cooking. J Archaeol Sci 36(3):915–922.
23. Briggs DE, Brookes PA, Stevens R, Boulton CA (2004) Brewing: Science and Practice
(Taylor & Francis, London).
24. McGovern PE (2009) Uncorking the Past: The Quest for Wine, Beer, and Other
Alcoholic Beverages (Univ of California Press, Berkeley).
25. Michel RH, McGovern PE, Badler VR (1992) Chemical evidence for ancient beer.
Nature 360(6399):24.
26. Michel RH, McGovern PE, Badler VR (1993) The first wine and beer: Chemical de-
tection of ancient fermented beverages. Anal Chem 65(8):408A–413A.
27. McGovern PE, et al. (2005) Chemical identification and cultural implications of a
mixed fermented beverage from Late Prehistoric China. Asian Perspect 44(2):249–275.
28. Delcour JA, Hoseney RC (2010) Principles of Cereal Science and Technology (American
Association of Cereal Chemists International, St. Paul, MN).
29. Boivin N, Fuller DQ, Crowther A (2012) Old World globalization and the Columbian
exchange: Comparison and contrast. World Archaeol 44(3):452–469.
30. Jones M, et al. (2011) Food globalization in prehistory. World Archaeol 43(4):665–675.
31. Chen FH, et al. (2015) Agriculture facilitated permanent human occupation of the
Tibetan Plateau after 3600 B.P. Science 347(6219):248–250.
32. Flad R, Li S, Wu X, Zhao Z (2010) Early wheat in China: Results from new studies at
Donghuishan in the Hexi Corridor. Holocene 20(6):955–965.
33. Yu Y (1997) Han. F ood in Chinese Culture: Anthropological and Historical Perspectives,
ed Chang KC (SMC Publishing, Taipei, Taiwan), pp 53–83.
34. Liu L (2007) The Chinese Neolithic: Trajectories to Early States (Cambridge Univ Press,
Cambridge, UK).
35. Jennings J, Bowser BJ (2009) Drink, power and society in the Andes: An introduction.
Drink, Power, and Society in the Andes, eds Jennings J, Bowser BJ (Univ Press of
Florida, Gainesville), pp 1–27.
36. LiangX (1987) Shilun Shaanxi Miaodegou II culture (A Discussion on Shaanxi Miaodigou
II Culture). Kaogu Xuebao 04 :397–412.
37. LiangX (1994) Shilun keshengzhuang Erqi wenhua (A discussion onthe Keshengzhuang
Culture). Kaogu Xuebao 04:397–424.
38. An C, Feng Z, Tang L (2004) Environmental change and cultural response between 8000
and 4000 cal. yr BP in the western Loess Plateau, northwest China. JQuatSci19(6):529–535.
39. GPICRA (Gansu Provincial Institute of Cultural Relics and Archaeology) (2006) Gansu
Qin’an Dadiwan Xinshiqi Shidai Yizhi Fajue Baogao (Qin’an Dadiwan Excavation
Report) (Cultural Relics Publishing House, Beijing).
Fig. 5. IC of residues from Funnel 1, showing the presence of oxalate.
Wang et al. PNAS Early Edition
|
5of5
ANTHROPOLOGY
Supporting Information
Wang et al. 10.1073/pnas.1601465113
SI Text
Further Details of the Mijiaya Site. The Mijiaya site was discovered in
1923 by J. G. Anderson and excavated from 2004 to 2006. The
excavations revealed three separate cultural strata belonging to
established chronology in the region based on ceramic typology with
associated radiocarbon dates: the Banpo IV phase (belonging to
the late Yangshao period, 3400–2900 BC), the Miaodigou II phase
(2800–2450 B.C) (36), and the Keshengzhuang phase (2400–2000
BC) (37). The entire site measures around 45 ha with cultural
deposits 1.5–4 m in depth. A total of 166 Banpo IV pits were found,
and the majority of them are regular in shape with flat bottoms,
suggesting their initial function as storage pits (8). The ceramic ar-
tifacts from the earliest stratum are Banpo IV styles that are com-
parable to the related sites in the region, including Xiehu in Lantian,
Xijing in Shanxian, and Quanhucun in Huaxian. Pollen analysis
indicates that the climate of the Wei River valley was semiarid
during the Banpo IV phase, characterized by steppe vegetation and
herbaceous plants dominating the pollen assemblage (38).
Social Complexity During the Late Yangshao Period. The settlement
pattern of the late Yangshao period is characterized by site nu-
cleation and a two- or three-tiered settlement hierarchy. Several
large regional centers emerged along the Wei River valley for the
first time, including Dadiwan (100 ha) and Gaositou in Gansu and
Anban (70 ha) in Shaanxi. Palace-like architectures are present in all
three sites, occupying the central locations of the settlements. At
Dadiwan, for example, a multiroomed structure (F901; 290 m
2
in
size) was found in the center of the site (39). Ceramic remains
found in the structure include large-sized pottery storage urns, pile-
up bowls, and vessels of regularly graduated sizes. A large hearth
was located in the center of the major room. The house may have
functioned as a central place for activities of regional communities.
Compared with the sites in the middle Yangshao period, public
buildings in the late Yangshao period became bigger in size, and
they are likely to have served for ritual ceremonies and feasting, at
both local and regional levels (34). The construction of large public
buildings during the late Yangshao period implies an increased
level of social hierarchy and complexity. Competitive feasting is
likely to have been conducted by the regional elite for obtaining
high social status.
SI Materials and Methods
Field Sampling. The field sampling took place in the Jingwei Ar-
chaeological Station in Xi’an, China. All artifacts for this study were
curated in storage. Only new and sterile plastic bags, test tubes,
pipettes, toothbrushes, and razor blades were used during the
sampling process. The procedure used is as follows.
First, three residue samples for IC analysis were taken from the
interior surfaces of Funnel 1, Pot 3, and Pot 5. The residues were
yellowish in color and firmly adhering to the interior vessel walls.
Residues from each vessel were scrapped off by using a clean razor
blade. By using the same method, two control samples were obtained
from the soil adhering to the exterior surface of Pot 5 (one for IC and
one for microbotanical analysis), and one control sample was taken from
the conservation plaster material on the interior surface of Funnel 2.
Second, all artifacts were subjected to sonication. An ultrasound
bath was used to extract residues from Funnel 1, Funnel 2, Pot 1, and
a control sample (stone adze). Each artifact was placed in a new,
polyvinyl bag with ∼15 mL of distilled water. For Funnel 1, Funnel 2,
and Pot 1, only the mouth part (up to 2.5–3 cm from the opening)
was immersed in the distilled water. For the stone adze, the entire
artifact was immersed. The bag containing the artifact then was
placed into the bath for 3 min. After 3 min, the artifact was re-
moved from the bag. For artifacts that were too big to be placed in
the ultrasound bath (Pot 2, Pot 3, Pot 4, Pot 5, Pot 6, Pot7), ul-
trasonic toothbrushes were used. First, each artifact was placed in a
new polyvinyl bag. The interior surface of the artifact was then
brushed gently with an ultrasonic toothbrush and at the same time a
pipette was used to add distilled water to rinse the surface. Only a
new ultrasonic toothbrush and a new pipette were used for each
artifact. The water and all sediment from the bag were transferred
into a new 15-mL test tube. Each tube was stored in a sealed plastic
bag before laboratory analysis.
Laboratory Techniques for Starch and Phytolith Analyses. The pro-
tocol for starch and phytolith extraction is as follows:
Sample concentration. Each 15-mL tube containing sediment and
water was topped off with distilled water and placed in a centrifuge
(Eppendorf 5804, Hamburg, Germany) for 5 min at 1,500 rpm to
concentrate the sample at the bottom of the tube. The supernatant
was then decanted.
Dispersion. Fourmicrolitersof0.1%EDTA(Na
2
EDTA•2H
2
O)
solution was added to each tube. Then the capped tubes were
placed in an automatic shaker for 2 h to disperse the sediments.
After being removed from the shaker, the tubes were filled to 15 mL
with distilled water and centrifuged for 5 min at 1,500 rpm, and the
supernatant was decanted.
Heavy liquid separation. Four microliters of heavy liquid sodium
polytungstate at a specific gravity of 2.35 was added to the tubes. The
tubes were then centrifuged for 15 min at 1,000 rpm. The top 1- to
2-mm layer of organics was carefully removed from each test tube by
a new pipette and then transferred into a new 15-mL tube. The
samples were topped off with distilled water and centrifuged for 5 min
at 1,500 rpm to concentrate the starchandphytolithatthebottomof
the tube, and the supernatant was decanted. The rinse was repeated
two more times to remove any remaining sodium polytungstate.
Slide mounts and microscope scanning. An aliquot of residue sample
was extracted with a pipette to a microscope slide and allowed to
dry. Then the residue was resuspended in 30–40 μL of 50% (vol/vol)
glycerol and 50% (vol/vol) distilled water. A coverslip was placed
on top, and the edges were sealed with nail polish. All slides were
scanned under a Zeiss Axio Scope A1 fitted with polarizing filters
and DIC optics, at 200×and 400×for both starch and phytoliths.
Beer-Brewing Experiments. The brewing experiments were based on
four sets of cereal grains. The four experimental sets consisted of
broomcorn millet (40 g), foxtail millet (40 g), a mixture of broomcorn
millet (30 g) and barley (10 g), and a mixture of foxtail millet (30 g)
and barley (10 g), respectively. Each set went through three brewing
steps, including malting, mashing and fermentation. The procedure
is as follows:
First, grains were immersed in water until they began to germinate.
The room temperature was around 20–28 °C. Most grains germi-
nated after 8 d, and they were drained and dried. Next, the malted
grains were crushed and mixed with heated water to achieve a final
temperature of 65 °C. The temperature was maintained for 2 h.
Finally, the mash was cooled in room temperature and allowed to
ferment in a brewing container for 2 d. During fermentation each
container was covered with a lid to create an anaerobic condition.
To obtain reference starch samples to compare with the ancient
starches, we took starch samples during the experiment. Two to
three malted seeds of broomcorn millet, foxtail millet, and barley
were taken for microscopic observation, and two patterns were
observed. First, all three types of malted grains had starches that
Wang et al. www.pnas.org/cgi/content/short/1601465113 1of8
showed pittings, channelings, or fissures radiating from their cen-
ters. The centers of the grains appeared hollowed but the outer
edges appeared undamaged (Fig. S3 A–C). Second, this type of
damage was common and appeared in around 90% of broomcorn
millet and foxtail millet starches; it was rare and appeared in about
1% of barley starches. A second batch of starch samples was taken
from each set when mashing was done. Various levels of swelling
and distortion were present in ∼5% of broomcorn millet, and 10–
15% of the other three samples (Fig. S3 D,G,J,andM). When
fermentation was finished, a third batch of starch samples was
taken from each set. Three patterns were observed from all four
brewing sets. First, pitted starch grains were still present, and their
outer edges also appeared damaged (see arrows in Fig. S3 E,H,
and K). Second, abundant starch grains exhibited swelling, distor-
tion, and loss of extinction cross. Many starch grains merged into
one another completely (Fig. S3 Fand O). Compared with mashed
samples, the fermented starch grains exhibited a higher level of
swelling and distortion; extinction cross was not observed in many
large starches. Third, some small-sized starch grains remained in-
tact (see arrow in Fig. S3L). However, we were unable to produce
quantified data from the fermented samples because a high pro-
portion of the starch grains was gelatinized and merged together.
Fig. S1. Analyzed Mijiaya artifacts not included in Fig. 1. Artifacts and their discovery contexts from the upper row to the lower row: (A) funnel 2 (H78), red
circle indicating the sampling location of control sample 3; (B) pot 1 (H82); (C) pot 2 (H82); (D) pot 4 (H82); (E) pot 5 (H78), red circle indicating the sampling
location of control sample 1 and control sample 4; (F) pot 7 (H78); (G) stone adze (H78).
Fig. S2. Residues from the interior surface of funnel 1.
Wang et al. www.pnas.org/cgi/content/short/1601465113 2of8
Fig. S3. Starch grains from brewing experiments: (A) Malted broomcorn millet (P. miliaceum); (B) malted foxtail millet (S. italica); (C)maltedbarley(H. vulgare);
(D) mashed bro omcorn millet; (E) fermented broomcorn millet, showing pittings and damaged outeredges; (F) fermented broomcorn millet, showing pitting (Left)
and gelatinization (Right); (G) mashed foxtail millet; (H) fermented foxtail millet, showing pitting and damaged outer edge; (I) fermented foxtail millet, showing
gelatinization, merging, and loss of extinction cross; (J) mashed broomcorn millet and barley; (K) fermented broomcorn millet and barley, showing big hollows in
the centers and damaged outer edge s; (L) fermented broomcorn millet and barley, showing one gelatinized grain (Left) and one undamaged grain (Right);
(M) mashed foxtail millet and barley; (N) a starch grain from fermented foxtail millet and barley, showing channeling and distortion; (O) a cluster of completely
merged starch grains and loss of extinction cross from fermented foxtail millet and barley.
Table S1. Morphology of starch grains
ID Grain shape Size range (mean), μm Hilum Fissures Lamellae Extinction cross
Broomcorn millet,
P. miliaceum
Polygonal and subround,
facetted
3.77–13.87 (9.06) Mostly
centric
“Y”,“V”, or linear
forms
Absent “+”Shape with
straight arms
Triticeae Round or oval,
flat surface
6.85–38.38(21.30) Centric Rare Visible on
large grains
Mostly “+”shape
Job’s tears,
C. lacryma-jobi
Polygonal and subround,
facetted
5.1–28.28(16.3) Centric or
eccentric
Common, “Y”,“V”,
and linear forms,
or many fine lines
radiating to the
edge
Visible on
some
Mostly straight,
sometimes
with bent or
Z-shaped arms
Snake gourd root,
T. kirilowii
Spherical or regular oval,
bell-shape, semispherical,
and nearly semispherical
with facets
8.61–30.61(18.85) Centric or
eccentric
Some with short
linear fissure
Visible on
large grains
Mostly with bent
arms, sometimes
straight
Yam, Dioscorea sp. Irregular triangular
or oval shape
12.34–23.11(18.09) Eccentric Some with short
linear fissure
Visible on
most grains
Bent arms
Lily, Lilium sp. Irregular triangular 16.88–34.71(24.41) Eccentric Some with short
linear fissure
Visible on
most grains
Bent arms
Table S2. Description of phytolith morphometries analyzed
Name Description
Form factor Equals 4 ×Area ×π/Perimeter, it is 1.0 for a perfect circle and diminishes for irregular shapes.
Roundness Equals 4 ×Area/π×Length
2
, it is 1.0 for perfect circle and diminishes with elongation of the feature.
Solidity Ratio of area to convex area; It is 1.0 for a perfectly convex shape, diminishes if there are surface indentations.
Compactness Ratio of the equivalent diameter to the length.
Convexity Ratio of convex perimeter to perimeter; It is 1.0 for a perfectly convex shape, diminishes if there are surface indentations.
Aspect ratio Equals length/width.
Wang et al. www.pnas.org/cgi/content/short/1601465113 3of8
Table S3. Range of mean morphometries of articulated dendritic wave lobes observed in all bract types from all inflorescence locations for all accessions of selected species from
modern comparative species: Triticum,Avena,Secale,Agropyron, and Bromus
Genus Triticum Avena Secale Agropyron Bromus
Species T. aestivum T. dicoccoides T. dicoccon T. durum T. monococcum A. sativa S. cereale A. cristatum A. mongolicum B. japonica
Form factor 0.607–0.725 0.646–0.711 0.643–0.745 0.623–0.707 0.668–0.817 0.635–0.721 0.635–0.709 0.691–0.744 0.672–0.711 0.680–0.739
Roundness 0.500–0.614 0.517–0.608 0.531–0.619 0.552–0.603 0.532–0.617 0.526–0.616 0.504–0.626 0.526–0.603 0.512–0.594 0.537–0.614
Solidity 0.936–0.973 0.940–0.970 0.918–0.979 0.922–0.957 0.954–0.984 0.904–0.978 0.928–0.968 0.966–0.979 0.952–0.972 0.963–0.974
Compactness 0.691–0.781 0.714–0.777 0.724–0.785 0.738–0.769 0.725–0.783 0.718–0.783 0.706–0.774 0.738–0.774 0.712–0.769 0.730–0.782
Convexity 0.921–0.944 0.912–0.941 0.905–0.948 0.889–0.940 0.935–0.949 0.907–0.946 0.922–0.942 0.937–0.948 0.937–0.944 0.941–0.947
Aspect ratio 1.463–1.638 1.469–1.710 1.500–1.667 1.463–1.611 1.490–1.710 1.423–1.775 1.503–1.690 1.482–1.826 1.458–1.758 1.435–1.641
Wang et al. www.pnas.org/cgi/content/short/1601465113 4of8
Table S4. Range of mean morphometries of articulated dendritic wave lobes observed in all bract types from all inflorescence locations for all accessions of selected species from
modern comparative species: Elytrigia,Leymus,Roegneria,Hordeum
Genus Elytrigia Leymus Roegneria Hordeum
Species E. elongata L. secalinus R. mayebarana R. ciliaris R. pendulina H. vulgare H. bulbosum H. comosum H. secalinum H. distichon
Form factor 0.719–0.742 0.706–0.740 0.705–0.741 0.719–0.737 0.621–0.746 0.644–0.736 0.708–0.757 0.717–0.744 0.693–0.760 0.642–0.748
Roundness 0.582–0.628 0.573–0.625 0.552–0.603 0.588–0.613 0.621–0.629 0.448–0.617 0.563–0.632 0.564–0.634 0.525–0.644 0.458–0.626
Solidity 0.965–0.983 0.960–0.973 0.967–0.978 0.972–0.980 0.969–0.976 0.954–0.981 0.977–0.988 0.971–0.977 0.978–0.983 0.979–0.985
Compactness 0.760–0.791 0.754–0.789 0.741–0.774 0.765–0.782 0.786–0.792 0.664–0.784 0.748–0.794 0.748–0.796 0.721–0.802 0.671–0.790
Convexity 0.941–0.948 0.939–0.946 0.943–0.949 0.942–0.947 0.945–0.946 0.934–0.952 0.943–0.950 0.943–0.945 0.945–0.950 0.945–0.948
Aspect ratio 1.412–1.580 1.392–1.562 1.489–1.641 1.463–1.544 1.414–1.449 1.497–2.090 1.458–1.657 1.449–1.707 1.410–1.829 1.482–2.123
Wang et al. www.pnas.org/cgi/content/short/1601465113 5of8
Table S5. Mean morphometries of articulated dendritic wave lobes
observed in Mijiaya residue samples
Morphometry
Sample
Funnel 1 Pot 2 Pot 3 Pot 4 Pot 5
ND 252241027
NL 33 307 131 97 208
Form factor 0.729788 0.696355 0.668985 0.657271 0.715702
Roundness 0.547273 0.488202 0.455939 0.474625 0.533909
Convexity 0.949152 0.948322 0.947763 0.935229 0.942255
Solidity 0.980485 0.971811 0.971519 0.95774 0.975072
Compactness 0.736576 0.693264 0.669 0.684094 0.726654
Aspect ratio 1.666182 1.948433 2.07529 2.03824 1.737486
ND, the number of articulated dendritics that formed the wave lobes measured; NL,
the number of wave lobes measured in each sample.
Wang et al. www.pnas.org/cgi/content/short/1601465113 6of8
Table S6. Comparison of mean morphometries of articulated dendritic waves in
Mijiaya vessels with the range of mean wave morphometries from selected
species from modern comparative species
Reference taxon Sample Funnel 1 Pot 2 Pot 3 Pot 4 Pot 5
ND 2 52 24 10 27
NL 33 307 131 97 208
Triticum aestivum Formfactor xxxx
Roundness x x
Solidity x x x
Compactness x x x
Convexity x x
Aspect ratio
Triticum dicoccoides Form factor x x x x
Roundness x x
Solidity x
Compactness x x
Convexity x
Aspect ratio x
Triticum dicoccon Formfactor x xxxx
Roundness x x
Solidity xxxx
Compactness x x
Convexity xxxx
Aspect ratio x
Triticum durum Form factor x x x
Roundness
Solidity
Compactness
Convexity x
Aspect ratio
Triticum monococcum Form factor x x x x
Roundness x x
Solidity x xxxx
Compactness x x
Convexity x xxxx
Aspect ratio x
Avena sativa Formfactor xxxx
Roundness x x
Solidity xxxx
Compactness x x
Convexity x x
Aspect ratio x x
Secale cereale Form factor x x x
Roundness x x
Solidity x
Compactness x x
Convexity x x
Aspect ratio x
Agropyron cristatum Form factor x x x
Roundness x x
Solidity x x x
Compactness
Convexity x x x
Aspect ratio x x
Agropyron mongolicum Form factor x
Roundness x x
Solidity x x x
Compactness x x
Convexity x
Aspect ratio x x
Elytrigia elongata Form factor x
Roundness
Solidity x x x x
Compactness
Convexity x x x
Aspect ratio
Wang et al. www.pnas.org/cgi/content/short/1601465113 7of8
Table S6. Cont.
Reference taxon Sample Funnel 1 Pot 2 Pot 3 Pot 4 Pot 5
Leymus secalinus Form factor x x
Roundness
Solidity x x
Compactness
Convexity x
Aspect ratio
Roegneria mayebarana Form factor x x
Roundness
Solidity x x x
Compactness
Convexity x x x
Aspect ratio
Roegneria ciliaris Form factor x
Roundness
Solidity x x
Compactness
Convexity x
Aspect ratio
Roegneria pendulina Formfactor x xxxx
Roundness
Solidity x x x
Compactness
Convexity
Aspect ratio
Elongation
Bromus japonica Form factor x x x
Roundness x
Solidity x x
Compactness x
Convexity x
Aspect ratio
Hordeum vulgare Formfactor x xxxx
Roundness x xxxx
Solidity x xxxx
Compactness x xxxx
Convexity x xxxx
Aspect ratio x xxxx
Hordeum bulbosum Form factor x x
Roundness
Solidity x
Compactness
Convexity x x x
Aspect ratio
Hordeum comosum Form factor x
Roundness
Solidity x x x
Compactness
Convexity
Aspect ratio x
Hordeum secalinum Form factor x x x
Roundness x x
Solidity x
Compactness x x
Convexity x x x
Aspect ratio x x
Hordeum distichon Formfactor x xxxx
Roundness x x x x
Solidity x
Compactness x x x x
Convexity x x
Aspect ratio x xxxx
ND, the number of articulated dendritics that formed the wave lobes measured; NL, the
number of wave lobes measured in each sample; x, mean for the sample falls within the range
of means for the reference taxon.
Wang et al. www.pnas.org/cgi/content/short/1601465113 8of8