The social context of early pottery in the Lingnan region of south China.

Pearson, Richard

Introduction

In the past few years archaeologists have confirmed that people in East Asia began to experiment with sedentary living, pottery making and plant and animal domestication as early as 14000 years ago, and very definitely by 12000 years ago. It appears that a number of early sites in China, Japan and the Russian Maritime province show evidence of a settled existence (sedentism), principally pottery, and pose many questions about their social, economic, and environmental contexts. Generally, pottery appears to have been invented in crude form about 15 000 years ago, during the end of the Palaeolithic. Keally et al. (2004: 349) conclude that 'the earliest pottery in East Asia and the whole OM Worm is now reliably dated to about 13 700-13300 b.p. (about 17200-14700 BP) in 3 regions: (1) Japan, (2) lower and middle parts of the Amur River Basin in the Russian Far East and (3) southern China'. At the end of this formative period, around 9000 years ago, it became more sophisticated and widespread (Cao in press).

What was the impetus for the invention of pottery? In this paper I summarise very briefly the information on early pottery sites of the Lingnan region (Guangdong and Guangxi provinces), and introduce the concept of prestige technologies as a possible explanation for the emergence of pottery making and domestication. Building on the substantial contributions of Brian Hayden (1995, 1998) and Prudence Rice (1999), I connect this explanation to changing trends in the use of agency theory by archaeologists (Robb 1999; Dobres & Robb 2000).

Social origins of pottery making

Brian Hayden has proposed that 'aggrandizing individuals seeking to promote their self-interest have been responsible for the development of prestige technologies including the use of metals, pottery and domesticated foods' (1998: 18). These 'aggrandisers' depended on favourable surplus economic conditions, and their prestige technology could be transformed to practical technology. In 1995 Hayden stated that since there is no record of pottery making before the development of complex hunter-gatherer and horticultural communities 'it is tempting to view the initial development of pottery as prestige technology' (1995: 260). He proposed that the ceramics were used as food containers for competitive display and consumption, but at the same time, he left open the possibility that aggrandisers might use other media for prestige food containers. If pottery were part of a prestige technology, Hayden postulated that it should initially occur as serving or feasting utensils--plates, bowls, liquid containers, or vessels for the preparation of prestige foods, which might involve boiling, brewing or straining. Finally he predicted that in such cases there would be a rapid development to specialised production of elaborately decorated forms involving large expenditures of labour (1995: 261).

Prudence Rice (1999) adopted the general theoretical position of Hayden, identifying unfired and low-fired clay to be part of early prestige technologies. She found the 'aggrandiser' theory approach of Hayden to be heuristic because it combined earlier culinary and symbolic explanations with more robust (fewer post hoc adaptationalist) implications for the origins and wider adoption of pottery. The culinary hypotheses were based on the idea of the discovery that sun-baked or fired clay for lining baskets or fireplaces could be used for rigid, relatively impermeable containers while the symbolic explanations focused on the early appearance of objects other than vessels, such as figurines, ornaments, beads, and spindle whorls. Rice found that the earliest pottery sites lack evidence for year-round sedentism and housing, and often seem to be based on a settlement subsistence system featuring semi-sedentary foraging and collecting, with seasonal movements from riparian to interior camp sites (1999: 21). She concluded that from the viewpoint of feasting and social models, it might be more appropriate to think in terms of pottery containers for short-term 'accumulation' rather than long-term storage. The adoption of pottery could have involved practical and prestige technologies in different areas and at different times.

Although Rice concluded that there was no single evolutionary path for the development or adoption of fired pottery, she seemed to favour the aggrandiser/competitive feasting model. From it she constructed some expectations as to the characteristics of early pottery containers used in feasting, and the sites where they are found (1999: 123). Early pottery should appear in the context of seasonal occupations rather than fully sedentary settlement. Early pottery vessels should appear (whether by invention or adoption) among complex hunter-gatherer groups as part of emerging rank distinction. They would be expected to consist of special-purpose vessels, associated with accumulating, storing, preparing or serving special foods. Such foods might be carbohydrates in protein-rich environments or fats and oils in areas with predominantly starchy diets. Vessel capacities (either size or number of vessels) should be large, i.e. sufficient for storage, serving, and consumption of contents. Lastly, vessels used for feast foods might be expected to be decorated, bearing stylistic information pertaining to the aggrandiser, his or her family and/or larger social group.

Perspectives of agency theory

Hayden's aggrandiser model needs revision on theoretical grounds. Its particular picture of the nature of agency in 'transegalitarian' societies affects the kinds of archaeological expectations and explanations sought by archaeologists such as Rice. There are two major approaches to power. The first focuses on the pursuit of power or prestige through stratagems that are rational in given circumstances (individual mastery model), whereas the second focuses on the creation of the actors' subjectivity through the process of agency. In the latter case, power is symbolically constituted and depersonalised, as a property of symbolic systems and institutionalised structure. The first type, as found in Hayden's aggrandiser theory, can be seen as reductionist and personalistic, a kind of methodological individualism. Saitta (1997: 263) states that it will be criticised for its strong reliance on certain ethnographic records and its gendering of power holders as male. A more nuanced view of agency assumes that power is depersonalised and symbolically constituted within a specific society. The emphasis is placed on social reproduction into which material substances are integrated. Decisions are made on the basis of many interlocking social and symbolic structures having many purposes and motivations beyond the individual acquisition of power.

Saitta (1999: 137) provides a powerful critique of prestige good models. In these models he notes that elites extract appropriate surplus and that the elite-subordinate relationship is exploitative. It is more likely that prestige goods are embedded in complex structural relationships. They could be seen as 'communal social entitlements required for reproduction rather than instruments of elite power' (1999: 137). Saitta notes that cases in which there is a poor match with available evidence and old theoretical assumptions about the natural acquisitiveness of the few impede our development of new perspectives (1999: 145).

Hegmon (2003: 220) notes that studies of the actions of leaders assume the universality of striving for aggrandisement, but this is not accepted by all writers on the subject. Such a picture of the individual aggrandiser is based on notions of the individual which come from later capitalism (Hodder 2000: 23, 25), and there is a strong possibility of legitimising modern social relationships by uncritically projecting them back in time. Johnson (2000: 214) states: 'the theoretical relationship between the individual, the social collective and agency will vary according to context. Consequently it follows that different methods will be appropriate for identifying agencies and developing convincing interpretations in different contexts.' Brumfiel notes a strong division between writers who believe that agents work towards goals that are in some way cross-culturally predictable, and those who believe that the goals were defined by unique culturally and historically specific logics and values (2000: 249). She concludes that 'the presence of aggrandizers in the archaeological record should be demonstrated rather than assumed' (2000: 253). Let us consider some of these ideas when we turn to the evidence of early south Chinese pottery and incipient cultivation.

Early sites and pottery from the Lingnan region

My discussion is limited to sites south of the Lingnan Mountains of south China (Figure 1). Other sites in the valley of the Changjiang (Yangzi) River are discussed by Zhao and Wu (2000), Wu and Zhao (2003) and Zhang (2002). In the Late Pleistocene there appears to have been a trend toward increased exploitation of local wild resources. In the final Pleistocene and early Holocene (c. 14 000-9000 years ago), Jiao (1994) noted the development of ground stone tools, perforated and chipped stone tools, shell and bone tools and the earliest pottery. He termed this assemblage Mesolithic, as did Zhang (2000). Zhang (2002) and Lu (in press) have also noted that pottery appears in a Mesolithic or 'pre-Neolithic' (my term) context. Changes occurred 9000 years ago with the development of completely polished stone tools and the widespread use of pottery. A few sites briefly described in English summaries are listed below. Debates on the dating of these and other sites can be found in a special issue of The Review of Archaeology (Kuzmin 2003).

Excavations at Dayan, Lingui County, Guangxi province, in 2000 yielded deposits dating from the Late Palaeolithic to the Middle Neolithic. Early pottery, which was incompletely fired, included a large cooking vessel (fu or bo) and a cylindrical fragment. The earlier portions of the site yielded shell midden containing human burials. Later layers yielded pottery resembling the pottery of Pengtoushan, Hunan Province, dating to around 9000 b.p. (Cao in press; Zhang 2002; Fu 2001). Dayan pottery is described as being thick and low-fired. Vessel form and function are unclear (Cao in press). Dingsishan, Nanning, Guangxi province, on the Bachijiang River, has yielded ceramics dating to earlier than 10 000 b.p. as well as ceramics from about 8000 b.p. The latter appear to be round-bottomed cooking vessels (Fu 2002). Cao (in press) states that the latter period of the site, dating from about 8000 b.p., displayed unique burial traits but does not give details. A rock shelter site at Liyuzui, Luzhou City, Guangxi province, has yielded two piles of ashes interpreted as the remains of hearths (Zhang 2000; Fu 2002). Chang (1986:102) reported that several flexed burials were recovered from the site but the precise details were not provided. Dating of shells from the lowest layer ranges from 18 000 to 23 000 b.p. (Onuki 2003: 78) but these samples came from below the cultural layer (Institute of Archaeology 1991: 217). Previously, Zhao (1998) estimated the age of the Neolithic component to be around 11 000 b.p., so there is considerable debate concerning the dating of this site. Liyuzui yielded soft, low-fired sand-tempered pottery. The vessel form is unclear.

The site at Miaoyan, Guilin City, Guangxi Province, has yielded deposits dating from 20 000 to 12 000 b.p. Zhang (2002) states that the small sample (5 sherds) of pottery from Miaoyan are similar to the plain pottery of Xianrendong, Jiangxi. Recalibrated AMS dates from potsherds from Zone 5 give dates of about 18 000 to 19 000 b.p. (Zhao & Wu 2000: 237). Miaoyan pottery is sandy, thick, hand-built and undecorated. It is light brown or black on the outside and mostly black on the inside. The midden in the cave site at Zengpiyan, Guilin City, Guangxi Province, has yielded 18 human burials. Chang (1986: 102) reports that they are mostly flexed, that 6 individuals had an artificial perforation on the top of their skulls and that some of the skeletons had been coloured with red ochre. The dating of the burials is not clear. Underhill (1997: 135) states that the stone tools included digging stick rings and elongated pebbles with flat ends believed to be pestles. The pig bones of Zengpiyan are of special significance. From a total of 67 individuals identified from jaw bones, 40 yielded age determinations, 65 per cent being between 1 and 2 years old. This indicates human harvesting of the pig population (Ren 1995:41). The pig canines seem to show size reduction typical of domesticated pigs (Underhill 1997:135). The Zengpiyan pig remains appear to be dated about 8000 years ago (Ren 1995). The earliest pottery was fired at less than 250[degrees]C, and tempered with quartz particles at least 1.5mm in diameter. Surfaces were cord marked. The only shape is a wide-mouth semi-hemispherical vessel. Pottery from 10 000 to 8000 b.p. was slab built; from 8000 to 7000 b.p. it was wheel made, in various shapes including stands (Zhongguo Shehui et al. 2003) (Figures 2-4). Following new excavations and interdisciplinary analysis at Zengpiyan in 2001, a new evaluation of the site has been published (Zhongguo Shehui et al. 2003). Five phases, each of roughly 1000 years' duration, from 12 000 to 7000 b.p., were distinguished. The authors concluded that there was no evidence for plant or animal domestication at the site.

However, remains of rice thought to be in an early stage of domestication, dating to 12000 to 14000 b.p. have been found at the Yuchanyan site, Hunan province, only 200km to the east of Zengpiyan, in the Yangzi Basin (Zhang & Yuan 1998). Phytoliths of rice thought to have been collected have been found in the Niulandong site in northern Guangdong province (Zhongguo Shehui et al. 2003: 699). Two burials of individuals in crouched squatting position were recovered in Phase IV (9000 to 8000 b.p.). One of these had two clam shells placed together to cover the skull. In the Guilin area the authors note that natural stone slabs have been found covering burials dating from 12 000 to 8500 years ago, while tools are found with burials dated 8000 years ago (2003: 693).

In general, the sites show evidence of sedentism in the form of extensive shell deposits and burials, and the burials show some kinds of ritual treatment, but status distinctions are not clear. None of the early burials possesses grave goods. Red ochre is a precious substance, but it was used for the collective. However, Hayden believes that few people may have received such a purposeful burial, and that they are probably elite (Hayden pers. comm. November 2003). The dating of the burials remains uncertain. While vessel shape is difficult to determine from the Lingnan sites the consensus seems to be that the earliest vessels are round-bottomed cooking vessels. Early pottery from Yuchanyan and Diaotonghuan, north of the Nanling Mountains, also seems to consist of round-bottomed cooking vessels. Zhao and Wu state that they could have been used for cooking rice or snails, clams and fish (2000: 237).

Comparative early pottery sites

To assist the interpretation of the Lingnan pottery, comparisons can be drawn with two other examples of early pottery use. From Japan, a study of 30 sites giving 97 age measurements considered most reliable has found four phases of development of early pottery (Keally et aL 2003). Three of these pottery phases occur in the Incipient Jomon: Phase 1, plain pottery; Phase 2, linear relief ware; and Phase 3, mainly punctate, nail impressed, and cord marked. The fourth phase, in Earliest Jomon, is decorated with a cord-wrapped string. Calibrated dates for the beginning of Phase 1 are around 14 280 BC or 16 230 BP. Uncalibrated dates for the four periods are Phase 1, c.13 500-12700 b.p.; Phase 2, c. 12700-11 400b.p.; Phase 3, c. 11 400-9800 b.p.; and Phase 4, c. 9800-9000 b.p. Keally et al. (2003: 9) note that the quantity of pot sherds from each site remains very low throughout the 5000 years of the Incipient Jomon. Sites of Phase 1 had a median of 28 sherds, while those of Phase 3 had 500. In Phase 4, at the beginning of the Earliest Jomon, pottery suddenly becomes abundant. Vessels tend to have wide mouths, deep bodies, and various types of bases, including fiat, round, and pointed. Keally et al. (2003: 3) state that the vessel shapes are appropriate for boiling or cooking, or for storing liquids or solids. Comparable dates have been reported for Siberia (Keally et aL 2003) and from northern Vietnam (Obata 2003).

A similar situation seems to occur in Amazonia, where early pottery has been found in contexts similar to those of the Lingnan region. Roosevelt (1995) reports that early pottery sites at two locations appear to be those of specialised river foragers. Sites of the Mina Culture from the Amazon estuary zone of Salgado yielded shell- or sand-tempered plain pottery with simple bowl shapes. Uncalibrated dates determined from charcoal, shell, and pottery temper are in the range of 5570 to 3490 years ago (1995:118). Sites of the Alaka Culture, Guyana, are also early with dates of about 5000 years ago (1995: 117). The Taperhina site, in the coastal lower Amazon area, yielded 11 AMS dates in the range of 7080 to 6300 b.p., several millennia earlier than the Andean or Mesoamerican pottery (1995:124). Sherds come from fragile sand-tempered hemispherical or inturned-rim bowls, 150 to 300mm in diameter at the rim. Only 3 per cent of 383 sherds displayed decoration, consisting of geometric incisions on bowl rims and shoulders. Roosevelt thought that they could have been used for seafood stew or soup, or for storage or display.

Discussion

Hayden's postulates that pottery should first appear as serving or feasting utensils are difficult to confirm from the Lingnan evidence. With regard to Rice's expectations, it is not yet clear whether the Lingnan sites were occupied seasonally or year-round, or whether the sites show any signs of rank distinction. Zengpiyan's special burial treatment of red ochre appears be an expression of group identity rather than of individual status distinctions. The pottery vessels do not seem to have served special purposes, although the unusual forms from Dayan do raise some interesting questions, which need to be resolved. Vessel capacity does not seem to be large and the ceramics are undecorated.

In China and Japan, elaborately decorated pottery begins to occur in abundance only several millennia after the inception of pottery. In the Lingnan area, the Middle Neolithic (Li 2002:81) marks the appearance of abundant decorated pottery while in Japan it appears with the Early Jomon period (5300 b.c. to 3500 b.c. based on uncalibrated dates). The social dynamics of highly decorated ceramics in the Chinese Neolithic case, in which they are often used in great quantities as grave offerings, vary from region to region, and are very different from those of the Japanese Jomon, where they rarely appear in burials and ranking cannot be clearly seen from burials until the final two stages of the Jomon Period.

The early pottery of eastern Asia, which occurs in Mesolithic contexts (Zhang 2000), does not seem to be part of a prestige technology. What, then, is the social context of the pottery? The evidence suggests that small semi-sedentary or sedentary communities made very small quantities of pottery vessels for food preparation, since pottery is found in the living debris and is plain or minimally decorated. Vessel shapes seem to indicate food preparation rather than food consumption, although further research is needed to confirm this. Pottery may have served as a kind of communal social entitlement, being used for preparing some special food for the collective (Saitta 1999: 137). It has been proposed that the Zengpiyan pottery was used for cooking gastropods (Zhongguo Shehui et al. 2003: 691). This interpretation comes close to the ideas of Haaland (1997: 354) that pottery is associated with activities which become important with the advent of sedentism, activities centred around women and children and the hearth, such as cooking, firewood collection, water carrying, food consumption, and habitation. I would favour a similar interpretation for the Amazonian data rather than one of a prestige technology. In at least two parts of the world, where pottery is very old, it appears that it was not initially a prestige technology, being associated with social competition and ranking only at a subsequent stage of development, such as the Barra Phase of the Soconusco region of Chiapas at around 1500 b.c. (Clark & Gosser 1995: 210). Clark and Gosser (1995: 212-9) found multiple lines of invention in Mesoamerica, plain utilitarian pottery being associated with small egalitarian groups as a practical technology, and well-made, highly decorated pottery as a prestige technology for ritual use in a competitive, evolving, rank society. Thus the world's oldest ceramics have a rather different social context than some of the later cases.

Acknowledgements

A draft of this paper was prepared for the International Conference, Prehistory in South China and Southeast Asia, Guilin, China, December 30-14, 2003. I thank the Institute of Archaeology, Chinese Academy of Social Science, the Guilin City Council, Guangxi Municipality, and the Cultural Bureau of the Guangxi Municipality, China, for the invitation to participate in the conference and for their hospitality. Thanks are offered to Simon Kaner, Sainsbury Institute for the Study of Japanese Arts and Cultures, Yaroslav Kuzmin, Pacific Institute of Geography, Russian Academy of Sciences, Charles Keally, Sophia University, and Min Li, Department of Anthropology, University of Michigan, for assistance and advice. Figures 2-4 are reprinted from the Institute of Archaeology, CASS, Archaeological Team of Guangxi Zhuang Municipality, Zengpiyan Museum, Archaeological Team of Guilin City (Editors) 2003. Zengpiyan--A Prehistoric Site in Guilin, Beijing. Cultural Relics Publishing House. Plates V and VIII. Thanks to Professor Xianguo Fu.

Received: 2 January 2004; Accepted: 12 October 2004; Revised: 22 November 2004

References

BRUMFIEL, E. 2000. On the archaeology of choice: agency studies as a research paradigm, in M. Dobres & J. Robb (ed.) Agency in archaeology: 249-63. London and New York: Routledge.

CAO, B. in press. The earliest ceramics in China: the findings, contents, and a preliminary interpretation. Journal o fEast Asian Archaeology.

CHANG, K.C. 1986. The archaeology of ancient China. New Haven: Yale University Press.

CLARK, J. & D. GOSSER. 1995. Reinventing Mesoamerica's first pottery, in W.K. Barnett & J.W. Hoopes (ed.) The emergence of pottery: technology and innovation: 209-22. Washington: Smithsonian Institution Press.

DOBRES, M. & J. ROBB. (ed.). 2000. Agency in archaeology. London and New York: Routledge.

Fu, X. 2001. Guilin diqu shiqian wenhua mianmao lunkuo chuxian/The prehistoric framework of prehistoric Guilin. Wenwu Bao April 4: 1.

--2002. Guangxi di ige shiqian kaoguxue wenhua--Huning Dingsishan yizhi, in Li, Wenru (ed.) Zhongguo shinian baida kaogu xinfaxian/Top 100 New Archaeological Discoveries of China 1990-1999: 236-42. Beijing: Wenwu Publishing House.

HAALAND, R. 1997. Emergence of sedentism: new ways of living, new ways of symbolizing. Antiquity 71 : 374-85.

HAYDEN, B. 1995. The emergence of prestige technologies and pottery, in W.K. Barnett & J.W. Hoopes (ed.) The emergence of pottery: technology and innovation: 257-66. Washington: Smithsonian Institution Press.

--1998. Practical and prestige technologies: the evolution of material systems. Journal of Archaeological Method and Theory 5: 1-55.

HEGMON, M. 2003. Setting theoretical egos aside: issues and theory in North American archaeology. American Antiquity 68: 213-44.

HODDER, I. 2000. Agency and individuals in long term process, in M. Dobres & J. Robb (ed.) Agency in archaeology: 21-33. London and New York: Routledge.

INSTITUTE OF ARCHAEOLOGY, CASS. 1991. Radiocarbon dates in Chinese archaeology. Beijing: Wenwu Chubanshe.

JIAO, T. 1994. Gengshinshi mozhi quanxinshi chu Lingnan diqu de shiqian wenhua/Prehistoric cultures of the Lingnan area from the Late Pleistocene to the beginning of the Holocene. Kaogu Xuebao 1: 1-24.

JOHNSON, M. 2000. Self made men and the strategy of agency, in M. Dobres & J. Robb (ed.) Agency in Archaeology: 213-231. London and New York: Routledge.

KEALLY, C., Y. TANIGUCHI & Y. KUZMIN. 2003. Understanding the beginnings of pottery technology in Japan and neighbouring East Asia. The Review of Archaeology 24 (2): 3-14.

KEALLY, C., Y. TANIGUCHI, Y. KUZMIN & I. SHEWKOMUD. 2004. Chronology of the beginning of pottery manufacture in East Asia. Radiocarbon 46: 345-51.

KUZMIN, Y. (ed.). 2003. The nature of the transition from the Paleolithic to the Neolithic in East Asia and the Pacific. The Review of Archaeology, Special Issue. 24: 1-3.

LI, G. 2002. Subsistence of Neolithic Pearl River Area, south China. MA Thesis, Department of Anthropology, University of British Columbia.

LU, T.L.D. in press. Early pottery in South China and its archaeological significance. Journal o fEast Asian Archaeology.

OBATA, H. 2003. Shiberia: Enkaishu/Siberia and the Russian Maritime Province. Kikan Kokogaku 83: 80-84.

ONUKI, S. 2003. Chugoku nanbu--Reinan sanmyaku no kita to minami/South China--north and south of the Lingnan mountains. Kikan Kokogaku 83: 75-9.

REN, S. 1995. Gongyuanqian wuqian nian Zhongguo xinshiqi wenhua de jixiang zhuyao chengjiu/Important results from the Neolithic cultures in China earlier than 5000 BC. Kaogu 1: 37-49.

RICE, P. 1999. On the origins of pottery. Journal of Anthropological Method and Theory 6: 1-54.

ROBB, J.E. (ed.). 1999. Material symbols: culture and economy in prehistory. Carbondale: Center for Archaeological Investigations, Occasional Paper No. 26.

ROOSEVELT, A. 1995. Early pottery in the Amazon: twenty years of scholarly obscurity, in W.K. Barnett & J.W. Hoopes (ed.) The emergence of pottery: technology and innovation: 115-31. Washington: Smithsonian Institution Press.

SAITTA, D. 1997. Review of Foundations of Social Inequality, by T.D. Price & G. Feinman (ed.) Plains Anthropologist 42: 263-5.

--1999. Prestige, agency, and change in middle range societies, in J. Robb (ed.) Material symbols: culture and economy in prehistory: 135-49. Carbondale: Southern Illinois University Press.

UNDERHILL, A. 1997. Current issues in Chinese Neolithic archaeology. Journal of World Prehistory 11: 103-60.

Wu, X. & C. ZHAO. 2003. Chronology of the Transition from Paleolithic to Neolithic in China. The Review of Archaeology 24:15-20.

ZHANG, C. 2002. The discovery of early pottery in China. Documenta Praehistorica XXIX: 29-35.

ZHANG, F. 2000. The mesolithic in south China. Documenta Praehistorica XXVII: 225-31.

ZHANG, W. & J. YUAN. 1998. A preliminary study of ancient excavated rice from Yuchanyan Site, Dao County, Hunan Province, PRC. Acta Agronomica Sinica 24-4: 416-20.

ZHAO, C. & X. Wu. 2000. The dating of Chinese early pottery and a discussion of some related problems. Documenta Praehistorica XXVII: 233-9.

ZHAO, Z. 1998. The Middle Yangtze Region in China is one place where rice was domesticated: phytolith evidence from the Diaotonghuan Cave, northern Jiangxi. Antiquity 72: 885-97.

ZHONGGUO SHEHUI KEXUEYUAN KAOGU YANJIUSUO, GWANGXI ZHUANG MINZUJIGU WENWU GONGZUOTUAN, GUILIN ZENGPIYAN IZHI BOWUGUAN, GUILIN SHI WENWU GONZUOTUAN (ed.). 2003. Guilin Zengpiyan (Zengpiyan, Guilin). Beijing: Wenwu Chubanshe.

Richard Pearson, 1890 West 17th Avenue, Vancouver V6J 2M9 BC Canada (Email:[email protected])