Archaeological perspectives on the origins of azuki ( Vigna angularis )

Abstract

Recent archaeological findings provide a hint on domestication of azuki (Vigna angularis var. angularis) in East Asia. This preliminary study investigates archaeological collections from various regions in Korea, China, and Japan, representing the long-term evolution (5300–1450 BP) of azuki. Changes in seed shapes show that cultural manipulation of azuki began as early as 5300–4800 cal. BP. Azuki beans from Neolithic contexts in Korea and Japan show a possible sign of morphological response to human management, at least 2000 years prior to the appearance of fully domesticated forms. Yabutsuru-azuki (V. angularis var. nipponensis), a conspecific progenitor of domesticated azuki, has been a supplementary diet and seed reserve for lean years in East Asia, and this tradition may have a long root to the mid Holocene. Archaeological evidence indicates the possibility of multiple origins of azuki, supporting recent phylogenetic data. A unique contribution of this research is its interdisciplinary attempt to examine both the biological and cultural processes of this staple legume domestication.

Keywords

azuki China domestication Holocene Japan Korea multiple origins Neolithic

Introduction

Azuki (Vigna angularis (Willd.) Ohwi et Ohashi var. angularis) has been an important pulse as a staple and ceremonial food in East Asia for thousands of years (Yamaguchi, 1992). Its conspecific wild form and presumed progenitor, Yabutsuru-azuki (V. angularis (Willd.) Ohwi et Ohashi var. nipponensis (Ohwi) Ohwi et Ohashi) grows widely in East Asia, which is generally regarded as the center of domestication (Ohashi, 1980; Smartt, 1990). Phylogenetic studies, however, do not provide a conclusive consensus on its specific origin(s) within East Asia (Kaga et al., 2008; Yang and Han, 1994; Zong et al., 2003). Historic texts describe relatively late occurrence of azuki compared with other crops, first appearing in the Eastern Zhou (1046–256 bce) (Chang, 1983). The later record should not be counted as evidence of absence of azuki prior to the Eastern Zhou period, considering that the texts usually described events later than reflected in archaeological evidence.

Archaeology has not been much incorporated into the study of azuki origins. A lack of archaeological analysis on azuki is partly due to the misidentification of azuki as soybean or vice versa (Lee et al., 2011; Sakamoto et al., 2006). The difficulty in identification of genus Vigna to species level (Yoshizaki, 1992) may make archaeological azuki even less visible. Preservation is another challenging issue. Changes in the pod structure, such as increasing size, reducing dehiscence and coloration, are the clearest domestication related traits in azuki (Kaga et al., 2008). Pods are rarely preserved in archaeological contexts, and often seeds are the only option available to study, without discernable color left in testae. Similar to other grain crops such as rice (Liu et al., 2007), seed size is a continuous variation which does not show a clear demarcation in complex azuki populations (Xu et al., 2000). Accordingly, a strict distinction of domesticated azuki from its conspecific non-cultigens may not be feasible through archaeological records. As seen in founder crops along the Euphrates River (Zeder, 2011), clear morphological changes including seed size increase, may have occurred quite late in the domestication process and can be no longer considered a leading-edge indicator of domestication. Throughout this paper, therefore, the term ‘azuki’ refers to Vigna angularis without specifying the subspecies because of the ambiguity in the archaeological record. The long-term goal of this study is not to pinpoint which archaeological sample represents a domesticated form based on a clear morphological sign, but to understand how domestication began by showing the general trend of human intervention of azuki evolution over an extensive time span. As a preliminary study, this paper documents morphological changes in azuki over 5000 years, mainly based on archaeological data from Korea.

Azuki has been reported mostly from Korea and Japan (Crawford, 2006), although a few recent reports indicate presence of azuki in China. The recent starch analysis tentatively identifies the tribe Phaseoleae, possibly Vigna at the late Paleolithic context at Shizitan Locality 9 (c. 13,800–8500 cal. BP) in Shanxi province (Liu et al., 2011). More solid evidence comes from a much later context, the late Neolithic (locally known as the Late Longshan) Lianchengzhen site in Shandong (Crawford et al., 2005) (Figure 1). An AMS date on rice grains associated with azuki there confirms its association to the Late Longshan (TO10206, 3610±60 uncal. BP or 4060–3840 cal. BP. All dates given are in 1 σ calibrated ¹⁴C years) (Crawford et al., 2005). Another possible presence of azuki is reported from the Shang period (3600–3300 BP) at the Dunggasi site in Henan (Fuller and Zhang, 2007).

Figure 1.

Sites (in upper case letters) where both azuki and soybean were found. Other sites yielded only soybean.

Prehistoric Korea is regarded as a secondary region where domesticated crops, such as foxtail and broomcorn millets were introduced by the Middle Chulmun (or Neolithic) period (5500–5000 cal. BP) (Crawford and Lee, 2003). Recent data, however, indicates Korea as one of several areas where local domestication of soybean may have taken place (Kim et al., 2010; Lee et al., 2011). This paper raises the question of azuki domestication in prehistoric Korea, based on my study in the central and southern regions. One of the examples is the Middle Mumun (or Bronze, 2900–2400 BP) site at Daundong in Ulsan, South Gyeongsang province where huge amounts of charred azuki and soybean were recovered (Crawford and Lee, 2003). An AMS date on azuki (TO-8965, 2510±70 uncal. BP, 2740–2490 cal. BP) matches that on soybean (UCIAMS-60750, 2485±25 uncal. BP, 2710–2500 cal. BP) at Daundong. Findings from the Nam River valley added critical information on the even earlier presence of azuki in Korea (Lee et al., 2011). The Pyeonggeodong site in Jinju, South Gyeongsang reveals a large quantity of azuki and soybean at the Middle Chulmun context. The two direct dates on azuki there are 4960–4860 cal. BP (UCIAMS-60748, 4350±25 uncal. BP) and 4830–4650 cal. BP (UCIAMS-60749, 4175±25 uncal. BP). The Chulmun and Mumun sites are the main focus of this study.

Materials and methods

Comparative references of non-cultigen azuki are comprised of two groups. The first group (n=14) is from my collection of weedy azuki beans in the Nam River valley; they were collected from erect specimens near the archaeological site in 1998. The second group (n=56) is part of the Herbarium collection of the Crop Protection Division of the Rural Development of Agriculture of Korea (called RDA collection below). The specimens in this paper were collected from a vining specimen in North Jeolla (Accession no. 19/09/01). Domesticated azuki used here (n=30) is a market variety that I purchased in Korea in 2010.

Identification of archaeological Vigna relies on the pumule-hypocotyl (embryonic leaves-stems) shapes and sizes to cotyledons (Yoshizaki, 1992). All the archaeological azuki specimens (n=822) possessed charred cotyledons, with testa intact in some specimens. I measured only complete seeds in their full length and width (Figure 2). The earliest specimens were collected from the Middle Chulmun pits at Pyeonggeodong (Lee et al., 2011) (Figure 1). A majority of azuki in this study came from the Mumun-period Daundong site where legume concentration was particularly high in a house floor, counting 660 seeds per liter of soil floated. Azuki was also abundantly reported in vessels, floors, and pits at the Giheung-Gugal site, Gyeonggi province, dating to the Three-Kingdom and Joseon periods (c. 550 BP) (Lee, 2003). Charcoal from the Three-Kingdom pit was dated to 1840±40 uncal. BP (SNU-02037, 1820–1720 cal. BP) (Gijeon Archaeological Research Center, 2003).

Figure 2.

Seed morphology of azuki, Vigna angularis. (a) Modern domesticated azuki, scale 1 mm; (b) azuki from the Chulmun Pyeonggeodong site, Korea; (c) azuki from the Mumun Daundong site, Korea; (d) azuki from the Joseon-period Giheung-Gugal site, Korea. The ‘Ph’ and ‘H’ indicate the pumule-hypocotyl (embryonic leaf and stem) and the hilum. The ‘L’, ‘W’, and ‘T’ stand for length, width, and thickness, respectively. All archaeological azuki specimens are in the same scale.

For comparison, this study includes size data published elsewhere: the Middle Satsumon period (1200–1150 BP) Sakushu-Kotoni River site in Hokkaido, Japan (Crawford, 1986; Crawford and Yoshizaki, 1987); and the Late Longshan (4600–3900 BP) Liangchengzhen site (Crawford et al., 2005).

Seed lengths, widths, and thickness were measured by using a stereo zoom microscope with an eyepiece reticule of 100 µm or measured with the assistance of digital imaging software (Nikon Element). The maximum dimension of the side where the plumule-hypocotyl is visible is labeled as width (Figure 2). The maximum dimension of the side where the hilum is in place is called thickness in this study.

Results

Botany of azuki

Yabutsuru-azuki (V. angularis var. nipponensis) is generally considered a progenitor of domesticated azuki (Smartt, 1990), although there are different hypotheses on its origin, as summarized by Lumpkin and McClary (1994). Azuki is a predominantly self-pollinating species, but heterozygous genotypes, indicating outcrossing, were identified when transgenic azuki is grown in areas where its wild and weedy relatives occur (Wang et al., 2004). These three forms of azuki can generate complex interbreeding populations, which are distributed widely in Japan. Domesticated azuki has a long pod-filling period and thus flowers earlier but matures later than wild azuki. Genes found related to germination and flowering time in cultivated azuki may confer a selective advantage to the hybrid derivatives under some ecological conditions, and explain why azuki has evolved as a complex population in Japan (Kaga et al., 2008). Such populations are also found in Korea (Lee, 1993). The sequences in the trnL intron and trnL-F intergenic spacer of cpDNA indicate that cultivated, weedy, and wild azuki share the same sequence with a specific deletion of 51 bp (Yano et al., 2005). This similarity indicates cross proximity of the three in their evolutionary pathway, suggesting that the Far Eastern population of weedy or wild azuki is the direct ancestor of the domesticated azuki.

The weedy form shows an intermediate phenotype between domesticated and wild azuki for several traits (Yamaguchi and Nikuma, 1996). For example, the weedy form has larger seeds and more variable color patterns on testa than the wild form. The commercial variety of domesticated azuki is red, but many landraces have various sizes and colors of testae and cotyledon. The RAPD analysis suggests that genetic variations are higher in the intra-populations of weedy and wild azuki than those of the cultivated azuki (Xu et al., 2000). Weedy populations, although having different morphology from the wild ones, are usually considered ecotypes of the wild that are adapted to different habitats.

Most genetic studies on azuki domestication have been conducted on modern materials except for the following cases. Based on a specific deletion of 51 bp, Yano et al. (2005) confirms the presence of domesticated azuki in the following sites: the Middle Yayoi Kanaba site in Kushu (2300 BP); the Satusumon Sakushu-Kotoni River (1200–1150 BP) and Ohkawa sites (650 BP) in Hokkaido; and the Late Edo Suo Kojubunji site (155 BP). These archaeological materials represent rather later varieties that developed long after the onset of azuki domestication.

Morphological comparisons

Domesticated references in this study represent one of the commercial varieties with red testa. Although some of them have relatively small seeds, they are still larger than non-cultigen counterparts (Table 1). Some wild specimens from the RDA collection are considerably smaller than modern weedy references and archaeological seeds. Each pair student t-test (α=0.05) confirms that seeds from the weedy and wild forms are different in overall shape and size, represented by the two ratios: length/width; and length × width × thickness (Figure 3a, b).

Table 1.

Azuki bean sizes (mm). Values in bold are based on modern weedy specimens only (n=14) from the Nam River basin.

	Length	Width	Thickness	L/W	L/T	L×W	L×W×T
Domesticated azuki (n=30)
Max	8.4	6.1	6.3	1.7	1.8	48.8	302.4
Min	5.6	4.3	3.9	1.2	1.1	24.6	97.0
Mean	6.8	5.1	5.2	1.4	1.3	34.8	184.0
Std Dev	0.8	0.5	0.8	0.1	0.1	7.1	59.5
Non-cultigen azuki (n=70)
Max	5.1	3.0	3.8	2.1	1.7	13.8	52.3
Min	1.6	1.3	2.1	1.0	0.6	2.1	4.5
Mean	3.4	2.5	2.6	1.4	1.3	8.8	22.8
	4.5	2.4	3.0	1.9	1.5	10.8	32.4
Std Dev	0.8	0.4	0.3	0.3	0.3	2.5	8.1
	0.3	0.2	0.4	0.2	0.1	1.4	8.0
Chulmun specimens (n=37)
Max	4.1	2.6	2.7	1.8	1.9	10.7	28.3
Min	2.5	1.7	1.8	1.3	1.3	4.6	8.6
Mean	3.2	2.1	2.1	1.5	1.5	6.7	14.5
Std Dev	0.3	0.3	0.2	0.1	0.1	1.4	4.5
Mumun specimens (n=653)
Max	7.0	4.7	4.4	2.6	3.5	27.5	115.3
Min	2.6	1.4	1.0	0.8	1.0	5.2	6.8
Mean	3.8	3.0	2.8	1.2	1.7	11.6	39.1
Std Dev	0.6	0.3	1.0	0.2	0.7	3.0	22.7
N	653	652	207	652	207	652	206
Three Kingdom specimens (n=110)
Max	6.8	3.7	3.6	2.3	18.3	24.8	89.2
Min	3.0	2.2	0.3	1.1	1.1	7.7	5.1
Mean	4.6	3.0	3.0	1.5	1.7	14.0	42.0
Std Dev	0.8	0.3	0.4	0.2	1.6	3.5	14.8
N	110	108	110	108	110	108	108
Joseon specimens (n=12)
Max	6.1	4.0	4.3	1.8	1.7	22.8	98.0
Min	4.5	2.7	2.7	1.4	1.3	12.6	34.3
Mean	5.4	3.3	3.4	1.6	1.6	18.1	63.5
Std Dev	0.5	0.4	0.4	0.1	0.1	3.2	18.0
Liangchenzhen specimens (n=9)
Max	4.2	3.1
Min	4.0	3.3
Mean
Std Dev
Sakushu-Kotoni River specimens (n=61)
Max	7.1	4.8	5.0
Min	4.8	3.0	2.3
Mean	6.0	3.9	3.8
Std Dev

Figure 3.

Box plots of seed measurements of modern and archaeological azuki beans: (a) comparison of length/width ratios of two modern non-cultigen azuki groups, including weedy specimens from the Nam River valley and wild specimens from the Rural Development of Agriculture of Korea (RDA) collection; (b) comparisons of length × width × thickness ratios of two non-cultigen azuki groups; (c) comparisons of length measurements of modern and archaeological specimens. The top, bottom, and line through the middle of the box correspond to the 75th percentile (top quartile), 25th percentile (bottom quartile) and 50th percentile (median), respectively. The whiskers on the bottom extend from the 10th percentile (bottom decile) and top 90th percentile (top decile). The lines connect the means. Only the ranges are published for the Liangchenzhen site (LCZ) in China and Sakuhsu-Kotoni River site (SKR) in Japan. Each pair student’s t-tests (∞=0.05) are charted in the far right boxes. In (c), non-cultigen and Chulmun specimens (underlined) are statistically indistinguishable; (d) correspondence analysis expressed in a two-dimensional plot using azuki seed sizes (length, width, thickness) and shape (l/w). Reference non-cultigen azuki collected from the Nam River valley (square); non-cultigen azuki from the RDA collection (square with ‘w’); domesticated azuki (triangle); and Chulmun specimens (open circle).

Most archaeological azuki specimens are smaller than modern domesticated ones but still larger than modern non-cultigens (Figure 3c). Archaeological azuki beans show a clear increase in average lengths through time. This trend is also detected in their width and thickness dimensions. Mumun-period azuki reveals a wide range of seed sizes, overlapping with the early ones from the Chulmun and the later ones from the Three-Kingdom and Joseon, and domesticated references. t-test (α=0.05) shows that most archaeological samples are different from one another and from all modern references in their lengths (Figure 3c). In contrast, the Chulmun specimens are statistically indifferent from the non-cultigen references.

Chulmun azuki is indistinguishable from the modern non-cultigen references (Figure 3c). In order to check any clustering patterns of these two populations, correspondence analysis was applied to the seed size and shape dimensions. Clearly visible is the isolation of some non-cultigen RDA collection, which was harvested from a wild vining stem (squares in the upper part in Figure 3d). Domesticated specimens are also tightly clustered and separated from the rest. Interestingly, most non-cultigen seeds from the weedy, erect stem from the Nam River valley cluster together with the Chulmun azuki. The proximity between the two may indicate Chulmun azuki beans resemble those of the weedy form rather than those of the wild.

The late Neolithic azuki beans from Liangchengzhen also fit the general chronological pattern of increasing seed size. They are larger than the earlier ones of the Chulmun and within the wide size range of the later ones from the Mumun period. The original report emphasizes that azuki was present in Liangchengzhen anthropogenic habitats without distinguishing its domestication status (Crawford et al., 2005). Azuki specimens from the Sakushu-Kotoni River site in Japan are generally larger than those from the later historical specimens from Giheung-Gugal, Korea. These two populations may reflect different regional varieties, but this possibility should be examined with a larger sample size.

Conclusion

Debates on origins of azuki

Northeast Asia is generally regarded as the geographical origin of domesticated azuki (Verdcourt, 1970), but specific locales are much debated. For example, based on distribution of non-cultigen azuki and its fertile hybridization with domesticated azuki, Yang and Han (1994) suggested China as the primary origin. Earlier examples, however, are reported from Japan and Korea than from China (Crawford, 2006). Among 100 findings of beans in Japan, identification of early azuki is often ambiguous. Several Jomon and Yayoi mung beans were probably misidentifications of azuki (Obata et al., 2007), considering that early domesticated mung bean was found no earlier than 3400–3100 BP in its primary origin in northern India (Fuller, 2011). The Early Jomon Torihama shellmiddden may provide presence of azuki as early as 8000 cal. BP, but more solid evidence of Vigna came from the Middle Jomon Shimoyakebe site, Tokyo (Sasaki et al., 2007). The oldest association there has a direct date of 5300–5060 cal. BP (MTC-05837, 4515±45 uncal. BP), but its identification is labeled as either azuki or soybean (Kudo and Sasaki, 2010). Soybean was also dated to 4960–4850 cal. BP there (PLD-9088, 4515±15 uncal. BP). The original report distinguished two types of Vigna (Sasaki et al., 2007), but larger seeds were reclassified as soybean later. Vigna type A in the report represents azuki, ranging from approximately 3.3 to 5.8 mm long and 2.4 to 3.8 mm wide. Azuki from the Chulmun Pyeonggeodong site in Korea, which is approximately contemporaneous to the Shimoyakebe site, shows a smaller size range (Table 1). It is a subject of future study whether the differences between the contemporaneous Chulmun and Jomon beans reflect real differences between regional varieties or simply the sample size bias. Azuki became more prevalent in archaeological records through time from the Late Jomon, around 4000 cal. BP (Obata et al., 2007). Based on the distribution of various types of natural azuki populations, Japan is considered a center of genetic diversity of azuki and a possible origin of its domestication (Kaga et al., 2008).

In contrast to such a claim for the single origin, Zong et al. (2003) report seven distinct populations of azuki, including the ‘Japanese complex-Korean cultivated’, ‘Chinese wild’, ‘China Taiwan wild’, ‘Nepal-Bhutan cultivated’, and ‘Himalayan wild’ groups. Nucleotide diversity with geographical isolation of each group indicates that azuki has been domesticated from at least four progenitors with three or more geographical origins.

Given that weedy azuki is an ecotype of the wild form (Xu et al., 2000), this study proposes that expansion of wild azuki populations to anthropogenic habitats may have played a crucial role in domestication. That is, wild azuki populations relocated to human settlements experienced some morphological, physiological changes to adapt to a new habitat. Human intervention, such as harvesting and encouraging wild populations near their settlements, may have further prompted changes in wild populations, and eventually transformed them to domesticated azuki.

Cultural implications

Current studies of archaeological azuki enlighten our understanding of human involvement of azuki evolution, similar to the implications from my co-authored paper on soybean (Lee et al., 2011). First, a high density of azuki and its morphology similar to modern weedy azuki at Pyeonggeodong (4950–4650 cal. BP) support my hypothesis that human intervention initiated a seed size increase during the Middle Chulmun, well before azuki was fully domesticated. Middle Jomon data also indicate a sign of morphological response to human management as early as 5300–4850 cal. BP. Weedy azuki is a common ingredient for desserts and beanbags in Japan even today. In lean years farmers used to collect weedy azuki to consume and to secure a complement of azuki cultivars for the following years (Yamaguchi, 1992). The Neolithic inhabitants in Japan and Korea probably took advantage of such economic values of weedy azuki near their dwellings. Middle Chulmun people already cultivated two introduced domesticates, foxtail and broomcorn millets by 5500–5000 cal. BP (Lee, 2003), and thus managing annuals was not a novel idea for them. Weedy azuki was grown in their millet fields and became a part of agroecosystem like today. Cultigen-size azuki by the Middle Mumun (2740–2500 cal. BP) represents a product of a prolonged domestication process. Neolithic azuki should be understood in the context of niche construction efforts of small-scale human societies (sensu Smith, 2009). A narrow distinction of wild and domesticated species is not a useful concept in study of early human–plant interactions.

Second, azuki became widely adaptive to anthropogenic habitats in East Asia during the mid Holocene, producing larger seeds than those in natural habitats. Several sites prove this hypothesis, including the Middle Jomon (5300–4850 BP) (Kudo and Sasaki, 2010; Sasaki et al., 2007), the Middle Chulmun (4950–4650 cal. BP), and the Late Longshan (4060–3840 cal. BP) sites. In light of this archaeological evidence, azuki domestication may not be explained in terms of a single, narrowly defined origin. In later historical periods, azuki dispersed as far north as central Hokkaido and became one of the main crops by the Satsumon period (1400–1100 BP). The Sakushu-Kotoni River site (Crawford, 1986) may indicate a development of a large-seeded landrace in Hokkaido. This archaeological data support the recent phylogenetic studies that propose recurrent domestications of azuki in multiple localities (Zong et al., 2003).

Third, despite the limited preservation in archaeological settings, morphology of charred seeds, if characterized by multivariate analyses, is still a useful indicator to document a general trend of azuki domestication. It is, however, implausible to pinpoint the domestication status of individual archaeological specimens, solely based on the single variable of seed size, considering the wide variation in the complex interbreeding population of azuki (Xu et al., 2000). Larger sample sizes of both archaeological and modern populations from broad geographic areas will be a key factor for fruitful future research.

Footnotes

Acknowledgements

The following individuals and institutions in South Korea helped me in collecting archaeological azuki: Hopil Yoon, Minjung Ko, Chunyoung Kim, Miyoung Kim, and Bongjoo Lee (Gyeonggnam Development Institute); Sung-Joo Lee (Gangneung-Wonju National University); Hyunjung Cho (Samgang Institute of Cultural Properties); the late Sang-kil Lee (Kyungnam University Museum); and Yonghwa Kim (Gyeonggi Cultural Foundation). The Rural Development of Agriculture in Korea provided modern reference samples. Gary W Crawford (University of Toronto) and Rory C Walsh (University of Oregon) provided useful comments on an earlier draft of this paper. Brendan J Culleton prepared charred samples (UCIAMS-60748 to 60750) for AMS dating at the Archaeometry Facility at the University of Oregon, which are dated at the WM Keck Carbon Cycle Accelerator Mass Spectrometry Facility at University of California Irvine. I also appreciate the insight provided by anonymous reviewers.

Funding

This research was partially supported by the Academy of Korean Studies Fellowship Program in 2012.

References

Chang

T-T

(1983) The origins and early cultures of the cereal grains and food legumes. In: Keightley

(ed.) The Origins of Chinese Civilization. Berkeley: University of California Press, pp. 65–94.

Crawford

(1986) Sakushu-Kotoni River plant remains. In: Yoshizaki

(ed.) Sakushu-Kotoni River Site: The Ezo-Haji Component Plant Remains. Sapporo: Hokkaido University, pp. 4–21.

Crawford

(2006) East Asian plant domestication. In: Stark

(ed.) Archaeology of Asia. New York: Blackwell Publishing, pp. 77–95.

Crawford

Lee

G-A

(2003) Agricultural origins in Korea. Antiquity 77: 87–95.

Crawford

Yoshizaki

(1987) Ainu ancestors and Asian agriculture. Journal of Archaeological Science 14: 201–213.

Crawford

Underhill

Zhao

. (2005) Late Neolithic plant remains from northern China: Preliminary results from Liangchengzhen, Shangdong. Current Anthropology 46: 309–317.

Fuller

(2011) Finding plant domestication in the Indian Subcontinent. Current Anthropology 52: S347–S362.

Fuller

Zhang

(2007) A preliminary report of the survey archaeobotany of the upper Ying valley (Henan Province). In: School of Archaeology and Museology in Peking University and Henan Provincial Institute of Cultural Relics and Archaeology (eds) Archaeological Discovery and Research at the Wangchenggang Site in Dengfeng (2002–2005). Zhengzhou: Elephant Press, pp. 916–958.

Gijeon Archaeological Research Center (2003) (ed.) The Yongin Gugalri Site. Suwon: Gijeon Archaeological Research Center Press (in Korean).

10.

Kaga

Isemura

Tomooka

. (2008) The genetics of domestication of azuki bean (Vigna angularis). Genetics 178: 1013–1016.

11.

Kim

Lee

Van

. (2010) Whole-genome sequencing and intensive analysis of the undomesticated soybean (Glycine soja Sieb. and Zucc.) genome. Proceedings of the National Academy of Sciences USA 107: 22,032–22,037.

12.

Kudo

Sasaki

(2010) Characterization of plant remains on Jomon pottery excavated from the Shimoyakebe site, Tokyo, Japan. Bulletin of National Museum of Japanese History 158: 1–26 (in Japanese).

13.

Lee

G-A

(2003) Changes in subsistence systems in southern Korea from the Chulmun to Mumun periods: Paleoethnobotanical investigation. Ph.D. dissertation, University of Toronto.

14.

Lee

G-A

Crawford

Liu

. (2011) Archaeological soybean (Glycine max) in East Asia: Does size matter? PLoSONE 6: e26720: http://dx.plos.org/10.1371/journal.pone.0026720.

15.

Lee

(1993) Illustrated Flora of Korea. 5th edition Seoul: Hyangmun-Sa (in Korean).

16.

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.

17.

Liu

Lee

G-A

Jiang

. (2007) Evidence for the early beginning (c. 9000 cal. BP) of rice domestication in China: A response to Fuller et al. The Holocene 17: 1059–1068.

18.

Lumpkin

McClary

(1994) Azuki Bean: Botany, Production and Uses. CAB International.

19.

Obata

Sasaki

Senba

(2007) Impressions on pottery revealed cultivation of Glycine max subsp. max (soybean) in the late to latest Jomon preriods in Kyushu Island. Japanese Journal of Historical Botany 15: 97–114 (in Japanese).

20.

Ohashi

(1980) Systematic position of ‘Azuki’ (Vigna angularis (Willd.) Ohwi et Ohashi). In: Gakkai

Nippon Ikushu

(ed.) Ikushugaku Saikin no Shinpo. Tokyo: Keigaku Shuppan, pp. 73–76 (in Japanese).

21.

Sakamoto

Ishikawa

Nakamura

. (2006) Species identification of 6,000-year-old beans from Sannai-Maruyama site, Aomori, Japan. Japanese Fossil Research 39: 1–6.

22.

Sasaki

Kudo

Momohara

(2007) Utilization of plant resources reconstructed from plant macrofossils during the latter half of the Jomon period at the Shimo-yakebe site, Tokyo. Japanese Association of Historical Botany 15: 35–50 (in Japanese).

23.

Smartt

(1990) Grain Legumes. Cambridge: Cambridge University Press.

24.

Smith

(2009) Resource resilience, human niche construction, and the long-term sustainability of Pre-Columbian subsistence economies in the Mississippi River Valley corridor. Ethnobiology 29: 167–183.

25.

Verdcourt

(1970) Studies in the Leguminosae-Papilionoideae for the Flora of Tropical East Africa: IV. Kew Bulletin 24: 507–569.

26.

Wang

Kaga

Tomooka

. (2004) The development of SSR markers by a new method in plants and their application to gene flow studies in azuki bean [Vigna angularis (Willd.) Ohwi & Ohashi]. Theoretical and Applied Genetics 109: 352–360.

27.

R-Q

Tomooka

Vaughan

. (2000) The Vigna angularis complex: Genetic variation and relationships revealed by RAPD analysis, and their implications for in situ conservation and domestication. Genetic Resources and Crop Evolution 47: 123–134.

28.

Yamaguchi

(1992) Wild and weed azuki beans in Japan. Economic Botany 46: 384–394.

29.

Yamaguchi

Nikuma

(1996) Biometric analysis on classicficaiton of weed, wild and cultivated azuki beans. Journal of Weed Sicence and Technology 41: 55–62 (in Japanese).

30.

Yang

R-j

Han

Y-g

(1994) The geographical distribution of Vigna angularis var. nipponensis in Liaoning Prov. of China and the artificial crossing between it and V. angularis var. angurais. Acta Agronomica Sinica 20: 607–613 (in Chinese).

31.

Yano

Yasuda

Yamaguchi

(2005) A test for molecular indentifcation of Japanese archaeological beans and phylogenetic relationships of wild and cultivated species of subgenus Ceratotropis (Genus Vigna, Papilionaceae) using sequence variation in two non-coding regions of the TRNL and TRNF genes. Economic Botany 58: 135–146.

32.

Yoshizaki

(1992) Ancient cultigen distribution. Kojogaku Janaru 355: 2–13 (in Japanese).

33.

Zeder

(2011) The origins of agriculture in the Near East. Current Anthropology 52: S221–S235.

34.

Zong

X-x

Vaughan

Tomooka

. (2003) Preliminary study on geographical distribution and evolutionary relationships between cultivated and wild adzuki bean (Vigna angularis var. angularis and var. nipponensis) by AFLP analysis. Plant Genetic Resources: Charaterization and Utilization 1: 175–183.