The prehistory of a Friction Zone: first farmers and hunters-gatherers in Southeast Asia.

Higham, C.F.W. ; Guangmao, Xie ; Qiang, Lin 等

[ILLUSTRATION OMITTED]

Introduction

The Neolithic Revolution in East and Southeast Asia and its aftermath have received much attention over the past two decades. It is now established that the cultivation of rice and the domestication of pigs and cattle took place in the Yangtze Valley even if the timing of the various stages in this process remain to be fixed (Fuller et al. 2009). This transition was, according to Bellwood (2005), the prime stimulus for the expansion of farming groups first to Taiwan and thence south to the Philippines and Island Southeast Asia. He has traced this on the basis of archaeological remains that link the islands with the Asian mainland, and the linguistic evidence for a deep antiquity in the Austronesian (AN) languages spoken on Taiwan (Bellwood & Oxenham 2008; Gray et al 2009).

When Reid (1994) identified structural relationships between Austronesian languages and the Austroasiatic (AA) language Noncowry, spoken on the Nicobar Islands, he returned to a century-old proposal by Schmidt (1906), that the AN and AA languages share a common ancestry in the Austric phylum. This lead Blust (1996) to pose the possibility that there was a second demographic expansion of rice farmers on the mainland. This would have involved movement up the Yangtze to Yunnan, and then by river south and west into Southeast Asia and India. Intriguingly, there are cognate words from the Munda group in India, and several AA languages in Southeast Asia, for words associated with rice and its cultivation (Higham 2002).

This possibility, that there was a major diffusion of rice farmers into Southeast Asia ultimately from the Yangtze Valley, has received some support from archaeological evidence (Rispoli 2008; Zhang & Hung 2010). In the first instance, there is the establishment of village communities of a permanent nature, with cemeteries in which the dead were inhumed in an extended, supine position with mortuary offerings. One of the commonest offerings comprises pottery vessels which, from southern China into Vietnam, Thailand and Cambodia, were decorated with incised and impressed designs which bear an uncanny resemblance over considerable distances. Further parallels are found in the very presence of rice, along with the domestic dog, pigs and cattle. Spindle whorls link the Chinese and Southeast Asian Neolithic sites, and suggest that weaving was an established part of Neolithic life (Cameron 2002). A prudent interpretation of the chronological evidence suggests that the first Neolithic groups began to settle the inland plains of Southeast Asia around 2000 BC. Under these circumstances, mainland Southeast Asia falls into what Bellwood has termed a 'Friction Zone', where farming groups expanded into an area long settled by established hunter-gatherers (Bellwood 2005).

Documenting this model for demographic expansion has paid much attention to the relevant Neolithic sites (Sorensen & Hatting 1967; Hoang & Nguyen 1978; Higham & Kijngam 2009). However, there has been less research on a vital aspect of the process as a whole: was the Neolithic intrusion a deluge or a trickle, and what was the relationship between the intrusive rice farmers and indigenous hunter-gatherers?

Identifying who these hunter-gatherers were, and illuminating their adaptation in a region that, even at the height of the last glacial, would have been benignly warm, is not straightforward. Both archaeological evidence and the results of simulation studies concur that the initial spread of Anatomically Modern Humans out of Africa followed a coastal route involving India and Southeast Asia (Escoffier et al. 2008). The DNA of surviving hunter-gatherers in Southeast Asia, such as the Andamanese and the Semang, reveal close correspondence with African groups. However, nowhere else was so much land lost to the rising sea during the Holocene. With the sea level at its lowest, an area the size of India, known as Sundaland, was not only there to be settled, but also led on to the very doorstep of Australia (Figure 1). Sundaland would by all accounts have favoured human settlement. Centred in the equatorial regions, it presented the same warm conditions as the African homeland. The tropical estuary is one of the three richest habitats in terms of bioproductivity, and Sundaland contained the ancestral courses of some of the world s great rivers, not least the Mekong and the Chao Phraya.

[FIGURE 1 OMITTED]

The drowning of Sundaland between and 14 000 and 5000 BP rules out virtually any knowledge of the Anatomically Modern Humans who lived there over a period of at least 60 000 years (Soares et al. 2008). As a result, the hunter-gatherers of Southeast Asia have often been given rather a poor press. The first serious exposure began in northern Vietnam, when Colani (1927) and Mansuy (1924) explored rockshelters north and south of the Red River Delta and described flaked stone tool industries as Hoabinhian and Bacsonian. Compared with the Upper Palaeolithic of France, the quality of workmanship was unimpressive. As research extended into Cambodia, Laos and Thailand, so many more such rockshelters were examined. Again, a similar set of unifacially-worked river cobbles, known as sumatraliths, were encountered, together with points and short axes. The Hoabinhian achieved a brief notoriety in the 1960s following Gorman s excavation of Spirit Cave in northern Thailand. For the first time, he undertook fine screening of the cultural deposits, and recovered plant remains. These, in conjunction with small polished stone knives, some potsherds and a set of radiocarbon determinations, were interpreted by Solheim (1972) as early evidence for a Neolithic Revolution since some of the seeds encountered resembled peas or beans. This claim was given some support in 1972 with the recovery of rice chaff at the nearby site of Banyan Valley Cave, again in conjunction with potsherds.

However, these claims are now set aside. The dating of Banyan Valley Cave revealed that the rice was as late as the first millennium AD. The rice itself came from a wild plant (Yen 1977), and resin on the surface of one of the Spirit Cave sherds provided a determination of about 1400 cal BC (Lampert et al. 2003). Spirit Cave itself is a tiny rockshelter perched on a steeply sloping hillside. It is well positioned for broad spectrum hunting, gathering and collecting, but not for any form of agriculture.

In 2002, Rasmi Shoocondej (2006) returned to the general area of Spirit Cave to excavate the cavern of Tham Lod, with spectacular results. The cave itself measures 30m across, and three areas were opened. Deep deposits of cultural remains were revealed, the oldest stretching back to at least 35 000 years ago. At that juncture, the climate was cooler and moister, and the excavations reveal how Anatomically Modern Humans adapted to the interior uplands of Thailand, for the cave lies in rugged limestone country 640m above sea level. As at Spirit and Banyan Valley caves, the hunter-gatherers lived near a stream, for the Nam Lang lies only 250m distant from the base, from which the inhabitants hunted the local wild cattle, pig and deer. Indeed, a very wide range of local species was identified among the animal bones many of which were smashed beyond recognition. They also fished, and collected large quantities of shellfish. Several different environmental zones were exploited, including the canopied evergreen forest and the river margins.

Each part of the cave excavated seems to have been used for a different purpose. In one, there were hearths and evidence for habitation, but in another, hearths were absent and the area was used for the manufacture of stone tools typical of the Hoabinhian, for there were sumatraliths and short axes among the assemblage, while hammerstones were abundant. Local stone was used for these artefacts, including sandstone, andesite and mudstone.

After about 15 000 BC, the cave was used for burying the dead. Four skeletons were uncovered (Pureepatpong 2006). One adult, dating to about 12 000 years ago, was found in an extended position, associated with shellfish. An adult female, who died when about 25-30 years of age, was interred in a flexed position and the radiocarbon determination suggests that she was interred about 13 500 years ago.

More recent research has shown that this hunting and gathering tradition in the forested uplands of Southeast Asia has a long ancestry. Increasingly early dates are being obtained in Vietnam, while in southern Thailand, the site of Lang Rongrien has revealed early occupation. This large cavern is located on a limestone tower that lies between two streams. Its airy shelter has attracted human settlement over a long period, and when excavated from 1983 by Douglas Anderson, the remains of their activities were found (Anderson 1990). These included charcoal from hearths dated from at least 38 000 until 27 000 years ago. This time-span does not mean that the cave was occupied permanently. Rather, bands of hunters and gatherers sheltered there briefly before moving elsewhere. This part of Thailand attracts much rain, four times that of the dry north-east, and even today, the hills round Lang Rongrien support a luxuriant rainforest. Thailand has seen many environmental changes over the years, and even during these three phases of occupation, the coast moved between about 30 and 100km away from the cave. Hence, we find an absence of any marine fish or shellfish and assume that the inhabitants collected and hunted locally available sources of food.

The millennia of occupation in such sites, and the likelihood that there was also a long-term occupation of the coastal tracts of Sundaland, is also seen in the deep cultural deposits at Niah Cave, where the sequence extends to over 40 000 years of occupation (Barker et al. 2007; see here p. 492).

Coastal adaptation

By about 4400 years ago, the sea level was slightly higher than at present, making it possible to trace former shorelines. Prehistoric settlements associated with these offer the only insight we have to the way of life that had developed in Sundaland over the previous millennia. Many such sites have been examined in Vietnam and southern China, and they reveal a marine adaptation, as well as the remains of pottery vessels, ground stone tools and human burials often in a flexed position. Rather confusingly, many Southeast Asian colleagues generically describe these sites as Neolithic, although there is no evidence for any form of agriculture or the raising of domestic animals.

A site survey undertaken in 1991 in an area now about 25km from the shore of the Gulf of Siam identified many prehistoric settlements on the basis of deposits of marine shells associated with stone artefacts and pottery. When the former environment was reconstructed on the basis of the geomorphology of the region, it was found that each of these settlements had been located on the shore of a large marine inlet (Boyd 1998). One such site was extensively excavated. The cultural deposits at Nong Nor comprised a shell midden of varying thickness, dated by five radiocarbon determinations to about 2450 cal BC. This midden contained many potsherds and polished stone tools, but no evidence for rice cultivation or domesticated animals. Rather, it was numerically dominated by bivalve shellfish adapted to a sandy shore. The inhabitants were clearly adept blue water sailors. Entry to the open sea lay 5km to the west of the site, and the faunal remains revealed that they brought in eagle rays and bull sharks (Mason 1998). One burial was found in a site which O'Reilly (1998) has suggested was occupied for at least one season of the year. The seated and flexed skeleton of a woman was found under a covering of pottery vessels.

Perhaps two or three centuries after the occupation of Nong Nor, the sea level had fallen slightly. The estuary of the ancestral Bang Pakong River lies 14km north of Nong Nor, and there, we find the site of Khok Phanom Di (Higham & Thosarat 2004). This large mound, which dominates the extensive river flood plain, was excavated in 1984-85, and like Nong Nor, provides detailed evidence of late coastal hunter-gatherers. Unlike the brief occupation period at Nong Nor, Khok Phanom Di was settled continuously for four to five centuries from about 2000 BC. This is a particularly important period in Southeast Asian prehistory, because it saw the establishment of the first rice farmers. However, the first three phases of occupation at Khok Phanom Di appear to have involved a community descended from the people of Nong Nor, because they made virtually identical pottery vessels, bone tools and polished stone adzes. The initial occupation layer, for example, revealed a pit containing several adzes, and another with the clay anvils used to form ceramic vessels. These were incorporated within a thick shell midden that contained numerous hearths and postholes, some with mineralised wood still in place, to support structures. The cultural deposits at Khok Phanom Di accumulated very rapidly, and they contained the superimposed burials of the dead that have been divided into eight phases.

There are six graves in the earliest phase. A newly-born infant came first, followed by two men, a woman, a second infant and a two-year old. The bone thickness in the adults suggests they were healthy with a good diet, but there are also hints of illness (Tayles 1999). The two-year-old child suffered from a blood disorder which would have resulted in anaemia, and this also afflicted two of the adults, both of whom died when relatively young. Tayles (1999) has suggested that this disorder, probably a type of thalassaemia or other genetic condition, would have provided for some resistance to malaria but with anaemia as a side effect. This finding is significant, for it suggests that the initial population of the site had developed such genetic resistance over the millennia of living in a coastal environment with, probably, a high population of malarial mosquitoes.

With the second mortuary phase, graves were set out in clusters. Tayles s examination of the human skeletons has suggested that people were well fed and active. Men had strongly-muscled upper bodies, probably as a result of an activity such as canoeing or sailing. The men and women exhibit different patterns of tooth pathology. Perhaps the men spent considerable time away from the site trading or fishing. We also find high infant mortality, much evidence for anaemia, and the pelvic evidence suggests virtually universal female fertility. A notable amount of energy was also expended on mortuary rites. Some bodies had been wrapped in an asbestos shroud, sprinkled with red ochre and laid on a wooden bier in individual graves. Pottery vessels were expertly made, brilliantly burnished and incised with complex designs (Hall 1993; Moore 1993). One man was found with about 39 000 shell disc beads, and shell beads in barrel and funnel forms were found with many burials. Cowrie shells included in one grave are almost certainly exotic, and bangles were fashioned from fish vertebrae (Pilditch 1993). Other grave offerings included the teeth of rhinoceros and deer, a stone adze, a fishhook and the stones used to burnish pottery vessels. The range of grave goods echoes some of the activities suggested by Nancy Tayles as possible reasons for stress on peoples' bones: making pots and paddling canoes. No clear differences have been detected in the mortuary rituals and grave goods found with men and women, although some individuals stand out on the basis of either their barrel beads, or association with shell beads and pottery vessels.

During the course of the third mortuary phase, which followed with no evident time lag from MP2, we find evidence for some significant changes. The first granite hoe was encountered, along with several shell knives with use striations indicating use for harvesting. Faeces from burial 67 included domesticated rice remains, while burial 56 provided food residue from the lower abdominal area which comprised fish bones, scales, and rice chaff. During this phase there was a fall in the number of shellfish adapted to clean subtidal and intertidal marine sand, offset by a rise in those from the landward edge of the mangroves. It seems that the site was being distanced from the mouth of the estuary and that siltation was affecting shellfish beds. The isotopes in the human teeth now indicate that some of the women were raised in a different environment and came to Khok Phanom Di from elsewhere. It is at this point in the sequence that there appears to have been interaction with rice farmers whose settlements, we know, were being established inland.

The inland plains

Deforestation, sedimentation and modern land use have all conspired to destroy or cover the likely occupation sites of the hunter-gatherers who once were adapted to the inland, riverine plains. Yet these extensive areas would have been attractive to such settlement. The rivers and lakes would have provided a ready supply of fish and shellfish. There is a long list of indigenous species, ranging from the elephant to the mouse deer and including several species of cattle and deer, as well as their predators, the tiger and leopard.

Seeking prehistoric sites, for example on the extensive Khorat plateau of north-east Thailand, is straightforward for the Iron Age, for many were ringed with moats and banks and are readily visible. There are also mounds under modern villages that formed during the Neolithic and Bronze Age. Identifying the more ephemeral settlements of hunter-gatherers, however, relies as much on luck as judgement. Thus, at the base of the moated Iron Age site of Ban Non Wat, we encountered a series of human graves in which the body was interred in a flexed position (Higham & Kijngam 2009). This, as we have seen, was particularly favoured by the indigenous hunter-gatherers (Figure 2).

Moreover, the mortuary offerings placed with these individuals contrasts markedly with those found in Neolithic cemeteries. One of the flexed interments at Ban Non Wat contained a plain ceramic vessel, a woman was found wearing a unique form of large shell bead, and another woman wore roughly finished shell disc beads. These graves date from the eighteenth to the eleventh centuries BC. On excavating into the natural substrate at this site, we encountered a deposit of shellfish at the very edge of the excavated area, with a radiocarbon determination of Ox-A 18496: 15 045 [+ or -] 55 BP (16 081-16 557 cal BC). No artefacts were found in association and only further excavations will reveal if this was a natural deposit or reflects early hunter-gatherer occupation. As at Khok Phanom Di, however, there is a suggestion of contact between hunter-gatherers and early rice farmers.

For more complete evidence for the presence of hunter-gatherers away from the uplands and the coast in Southeast Asia, one has to turn to the recent intensive archaeological investigations in southern China. The occupation of caves in this area is well established. Foremost ?s the site of Zengpiyan in Guilin, where the hunter-gatherer sequence covers five phases beginning in 12 000 BP and ending in 7000 BP (Anon. 2003). A feature of this site is the early manufacture of extremely thick and poorly-finished pottery vessels. It is thought that they were used to heat shellfish in order to open them more easily. Numerous seated and flexed burials were found in this site, together with evidence for stone tool manufacture and local hunting and gathering. More recent fieldwork has also identified 15 sites ascribed to hunter-gatherer groups which settled along the banks of rivers. These comprise extensive workshops for the production of stone tools, but pottery sherds are relatively rare. The distribution covers an extensive area from Baise in the west to Guilin in the north-east, and south to Nanning and the Vietnamese border. In western Guangxi, stone workshops as well as burials were found at the sites of Gexinqiao (Xie et al. 2003), Baida and Kantun (Xie et al. 2003), all of which are located near the Youjiang River and its tributaries. These sites are dated from about 9000 BP to 5000 BP. In western Guangxi, Gexinqiao, excavated in 2002, is an extensive site, of which 1600m2 were excavated. It was occupied about 6000 BP. Stone tools include flaked choppers, points and scrapers, as well as polished adzes. All stages of manufacture are represented, so it was clearly a workshop site, in which the distribution of the stone flakes, hammerstones and anvils make it possible to identify the position taken by each artisan. Pottery sherds were rare, and in the main they were simply cord marked. Two tightly-flexed human burials were found, associated with large river cobbles.

[FIGURE 2 OMITTED]

[FIGURE 3 OMITTED]

In central Guangxi, along the Hongshuihe River, several sites have been excavated. Beidaling is the most important, for a large stone workshop of over 1600m and eight burials were found in the lower part of the deposits, which are dated to 8000 BP (Figure 3). More than 50 000 stone artefacts were recovered from the stone workshop, including hammerstones, anvils, whetstones, flaked choppers, points and scrapers, as well as polished adzes and axes (Figure 4). All stages of manufacture are represented, indicating a stone workshop site. Pottery sherds are rare (Lin & Xie 2005). It is very similar to Gexinqiao.

In eastern Guangxi, five sites have been found along the course of the Xunjiang River. Datangcheng was excavated in 2006 over an area of 2000 [m.sup.2]. Again, we find a stone workshop with two periods of occupation, dated respectively to 8000 and 5000 BP. There was a dense accumulation of stone flakes and artefacts in various stages of manufacture, including many adze roughouts, hammerstones and whetstones. Much pottery was also found at this site, including large vessels with everted rims (Lin et al. 2007). There are several known sites along the course of the Yongjiang River in southern Guangxi, and at Chongtang a group of 26 human graves have been uncovered, in which the dead were interred in a tightly flexed position (Figure 5). Cowrie shells also provide evidence for exchange with coastal communities (He & Chen 2008).

The important point about the Guangxi sites, when considered in conjunction with the coastal settlements, is that Southeast Asia was well populated with different hunter-gatherer groups when the first rice farmers arrived from the north. This finding has a significant bearing on the question posed above, was this Neolithic expansion a deluge or a trickle, and what interaction might have taken place between the intrusive farmers and indigenous hunter-gatherers?

[FIGURE 4 OMITTED]

Interaction in the Friction Zone

There is little room for doubt that sedentary farming was established in Southeast Asia through the expansion into the region ultimately from the Yangtze Valley, where the transition from hunting and gathering to settled agriculture has been documented. There is no corresponding transition in Southeast Asia: early Neolithic sites were settled by established practitioners who raised cattle and pigs, brought their dogs, and cultivated rice and millet. This is seen clearly at Ban Non Wat, where initial settlement took place in the mid seventeenth century BC (Higham & Higham 2009; Higham & Kijngam 2009). The biological remains include domestic pigs, cattle and dogs, and rice remains have been identified. However, there was still much food gathering. The middens contain shellfish, fish and wild species with a dominance of deer. The dead were interred in an extended, supine position in association with fine and large ceramic vessels decorated with incised and painted designs. Some of these designs are paralleled in early Neolithic sites in central Thailand, Vietnam, Cambodia and Malaysia (Higham 2004; Wiriyaromp 2007). At least two early Neolithic individuals wore cowrie shell ornaments, indicating exchange ultimately with coastal communities. According to the radiocarbon dating of shells placed with the dead, the site was used as a cemetery by hunter-gatherers and rice farmers over the same period of time.

[FIGURE 5 OMITTED]

At least one of the motifs found on Ban Non Wat Neolithic pots is also found in central Thai sites, as well as the final mortuary phase at Khok Phanom Di. In a similar vein, H-shaped shell beads and shell discs of the forms worn at Khok Phanom Di recur in central Thai Neolithic sites. Khok Phanom Di is arguably the only site available where it is possible to explore, in more than a cursory way, a pattern of interaction between the two groups under discussion. As we have seen, during the third mortuary phase at this site there is evidence for significant changes in behaviour. Bivalve shells were converted into harvesting knives. We find the use of large, granite hoes. Rice was consumed. Human faeces contained mouse hairs and remains of the beetle Oryzophilus surinamensis, both of which are drawn to live in rice stores. There were subtle changes ?n the form of ceramic vessels and the introduction of novel design motifs (Vincent 2003). These changes coincided with the presence in the site of some women who were raised elsewhere (Bentley et al. 2007). Is it possible that they came into the community from Neolithic settlements which we know were then being established in the hinterland? Their presence coincided with the first hoes and reaping knives. We might have to await new and refined techniques for extracting their DNA, for hitherto none has been found. There are, however, some intriguing lines of evidence. Until the third mortuary phase, infant mortality was very high indeed. Tayles (1999) has suggested that the population experienced a high incidence of a gene for thalassaemia or other blood disorder that provides resistance to malaria. Selection in favour of this gene over centuries or millennia would reflect the fact that the ravages of malaria would be minimised. However, this would have been at the expense of a high incidence of anaemia. Moreover, if both parents carry particular variants of this gene, their offspring has little chance of survival. This would help us to understand why infant mortality was so high during mortuary phases 2 to 3A at Khok Phanom Di. Following the arrival into the community of the women from putative inland communities during phase 3B, however, infant mortality fell quite sharply. This might have been due to the newcomers lacking the blood disorder gene.

[FIGURE 6 OMITTED]

From the third mortuary phase into and including the sixth, we find that women rose in social prominence due to their expertise in making pottery vessels. However, towards the end of mortuary phase 4, saline conditions returned after a brief period of low sea level that permitted the local cultivation of rice. The shell knives and granite hoes were no longer found, and marine hunting and gathering dominated once again. This was also a period when some women were interred with remarkable wealth--in one case, with over 120 000 shell beads (Figure 6). It is thus a tenable hypothesis that at this site, between about 1700 and 1500 BC, individuals descended from Neolithic women were absorbed into a vibrant coastal hunter-gatherer community.

The speculative nature of this suggestion ?s clear, but it makes a significant point. The expansion into Southeast Asia of rice farmers did not lead to the extinction of the long-established hunter-gatherers. Indeed, some of the latter continued to thrive, albeit in relatively remote and inaccessible places. It also leads to recent findings from the analysis of modern human DNA sequences. A particularly clear testing of this situation comes from the different ethnic groups that occupy the Malay Peninsula (Hill et al. 2006). On the basis of mtDNA sequences, the Semang, who are adapted to rainforest foraging and have a distinct Negrito appearance, reveal in haplogroups M21 and R21, a deep ancestry in the region that stretched back at least 60 000 years to the original Anatomically Modern Human expansion into Southeast Asia. Long-term continuity is further indicated by the recovery of ancient DNA from prehistoric hunter-gatherer skeletons in Moh Khiew Cave in Krabi province, dating to about 25 000 years ago, which suggests that they were ancestral to the modern Semang hunters of the same region (Oota et al. 2001). The Senoi differ from the Semang in being slash-and-burn agriculturalists. Their mtDNA is consistent with an indigenous component, as with the Semang. However, they also carry a strong admixture of Haplogroup Flala, which Hill et al. (2007) ascribe to an origin on the Southeast Asian mainland within the past 7000 years. This, they suggest, could originate in the expansion into the area of Neolithic farming groups who intermarried with the indigenous inhabitants.

On a broader canvas, the genetic evidence is now identifying major demographic events coinciding with the sea level rises that drowned Sundaland between 14 000 and 5000 years ago. Environmentally, this would have greatly expanded the extent of the coastline as a great archipelago was created, and reduced the extent of dry land. For coastal groups long adapted to ocean voyaging, this would have provided an opportunity for an increase in population and expansion of settlement. This is precisely what has been identified on the basis of the modern genetic evidence. At least for Island Southeast Asia, there is a growing body of genetic evidence for a demographic expansion out of Sundaland, and less emphasis on the later dominance of Austronesian speaking agriculturalists. As archaeological research gathers momentum on the mainland, one of the key issues to arise will be documenting the interaction between the indigenous hunter-gatherers and the Neolithic wave of advance from the north. It is predicted that the significance and contribution of the former will be considerable.

Acknowledgements

We are grateful to Dr Maru Mormina and Dr Nancy Tayles for commenting on this paper prior to submission and to Graeme Barker and Peter Bellwood for helpful reviews.

References

ANDERSON, D.D. 1990. Lang Rongrien rocksbelter: a Pleistocene, early Holocene archaeological site from Krabi, southwestern Thailand (University Museum Monograph 71). Philadelphia (PA): University Museum, University of Pennsylvania.

ANON. 2003. Zengpiyan--a prehistoric site in Guilin. Beijing: Cultural Relics Publishing House (in Chinese with English summary).

BARKER, G.W.W., H. BARTON, M. BIRD, P. DALY, L DATAN, A. DYKES, L. FARR, D. GILBERTSON, B. HARRISSON, C. HUNT, T. HIGHAM, L. KEALHOFER, J. KRIGBAUM, H. LEWIS, S. MCLAREN, V. PAZ, A. PIKE, P. PIPER, B. PYATT, R. RABETT, T. REYNOLDS, J. ROSE, G. RUSHWORTH, M. STEPHENS, C. STRINGER, J. THOMPSON & C. TURNEY. 2007. The 'human revolution in lowland tropical Southeast Asia: the antiquity and behavior of anatomically modern humans at Niah Cave (Sarawak, Borneo). Journal of Human Evolution 52(3): 243-61.

BELLWOOD, P. 2005. First farmers: the origins of agricultural societies. Oxford: Blackwell.

BELLWOOD, P. & M. OXENHAM. 2008. The expansions of farming societies and the role of the Neolithic demographic transition, in J.-P. Bocquet-Appel & O. Bar-Yosef (ed.) The Neolithic demographic transition and its consequences: 13-34. Dordrecht: Springer.

BENTLEY, A., N. TAYLES, C.F.W. HIGHAM, C. Macpherson & T.C. Atkinson. 2007. Shifting gender relations at Khok Phanom Di, Thailand: isotopic evidence from the skeletons. Current Anthropology 48(2): 301-314.

BOYD, W.E. 1998. The palaeoenvironment of the archaeological site at Nong Nor, in C.F.W. Higham & R. Thosarat (ed.) The excavation of Nong Nor: a prehistoric site in Central Thailand (University of Otago studies in prehistoric anthropology 18): 27-86. Dunedin: University of Otago; Oxford: Oxbow.

BLUST, R. 1996. Beyond the Austronesian homeland: the Austric hypothesis and its implications for archaeology, in WH. Goodenough (ed.) Prehistoric settlement of the Pacific (Transactions of the American Philosophical Society 86): 1 17-40. Philadelphia (PA): American Philosophical Society.

CAMERON, J. 2002. Textile technology in the prehistory of Southeast Asia. Unpublished PhD dissertation, The Australian National University.

COLANI, M. 1927. Lage de la pierre dans la province de Hoa Binh. Memoires du Service Geologique de rindochine 13: 1

ESCOFFIER, L., N. RAY & M. CURRAT. 2008. Where was paradise? A simulation study of the spread of early modern humans in heterogeneous environments, in S. Matsumura, P. Forster & C. Renfrew (ed.) Simulations, genetics and human prehistory: 9-18. Cambridge: McDonald Institute for Archaeological Research.

FULLER, D.Q., L. QIN, Y. ZHENC;, Z. ZHAO, X. CHEN, L.A. HOSOYA &C G.-E SUN. 2009. The domestication process and domestication rate in rice: spikelet bases from the Lower Yangtze. Science 323(5921): 1607-1610. doi: 10.1126/science. 1166605.

GRAY, R.D., A.J. DRUMMOND & S.J. GREENHILL. 2009. Language phylogenies reveal expansion pulses and pauses in Pacific settlement. Science 323(5913): 479-83. doi: 10.1126/science. 1166858.

HALL, D.M. 1993. Aspects of the pottery decoration, in C.F.W. Higham & R. Thosarat (ed.) The excavation of Khok Phanom Di: a prehistoric site in Central Thailand. Volume 3, material culture part 1 (Reports of the Research committee of the Society of Antiquaries of London 50): 239-74. London: Society of Antiquaries of London.

HE, A. & X. CHEN. 2008. New discovery from the excavation of Chongtang site in Chongzuo, Guangxi. China Cultural Relics News 1621 (in Chinese).

HIGHAM, C.F.W. 2002. Languages and farming dispersals: Austroasiatic languages and rice cultivation, in P. Bellwood & C. Renfrew (ed.) Examining the farming/language dispersal hypothesis: 223-32. Cambridge: McDonald Institute for Archaeological Research.

-- 2004. The opposed human figure at Khok Phanom Di. Journal of the Siam Society 92: 15-36.

HIGHAM, C.F.W. & T.F.G. HICHAM. 2009. A new chronological framework for prehistoric Southeast Asia based on a Bayesian model from Ban Non Wat. Antiquity 83: 125-44.

HIGHAM, C.F.W. & A. KIJNGAM. 2009. The origins of the civilization of Angkor. Volume 3, the excavation of Ban Non Wat: introduction. Bangkok: The Fine Arts Department of Thailand.

HIGHAM, C.F.W. & R. THOSARAT. 2004. The excavation of Khok Phanom Di: a prehistoric site in Central Thailand. Volume 7, summary and conclusions (Reports of the Research committee of the Society of Antiquaries of London 62): 239-74. London: Society of Antiquaries of London.

HILL, C., P. SOARES, M. MORMINA, V. MACAULAY, W MEEHAN,J. BLACKBURN, D. CLARKE, J. MARIPA RAJA, P. ISMAIL, D. BULBECK, S. OPPENHEIMER & M. RICHARDS. 2006. Phylogeography and ethnogenesis of Aboriginal Southeast Asians. Molecular Biology and Evolution 23: 2480-91.

HILL, C., P. SOARES, M. MORMINA, V. MACAULAY, D. CLARKE, PB. BLUMBACH, M. VIZUETE-FORSTER, P. FORSTER, D. BULBECK, S. OPPENHEIMER & M. RICHARDS. 2007. A mitochondrial stratigraphy for Island Southeast Asia. American Journal of Human Genetics 80(1): 29-43.

HOANG, X.C. & N.B. NGUYEN. 1978. The excavation of the archaeological site of Phung Nguyen. Ha Noi: Nha Xuat Ban Khoa Hoc Xa Hoi (in Vietnamese).

LAMPERT, CD., I.C. GLOVER, R.E.M. HEDCES, CP. HERON, T.F.G. HICHAM, B. STERN, R. SHOOCONDEJ & G.B. THOMPSON. 2003. Dating resin coating on pottery: the Spirit Cave early ceramic dates revised. Antiquity 77: 126-66.

LIN, Q. & G. XIE. 2005. An important discovery from the excavation of Beidaling site in Duan, Guangxi. China Cultural Relics News 1374 (in Chinese).

LIN, Q., G. XIE & L. ZHU. 2007. An important discovery from the excavation of Dangtangcheng site in Guiping, Guangxi. China Cultural Relics News: 1534 (in Chinese).

MANSUY, H. 1924. Stations prehistoriques dans les cavernes du massif calcaire de Bac-Son (Tonkin). Memoires du Service Geologique de llndochine 11 (2).

MASON, G.M. 1998. The shellfish, fish and crab remains, in C.F.W Higham & R. Thosarat (ed.) The excavation of Nong Nor: a prehistoric site in Central Thailand (University of Otago studies in prehistoric anthropology 18): 173-224. Dunedin: University of Otago; Oxford: Oxbow.

MOORE, M. 1993. The burnishing stones, in C.F.W Higham & R. Thosarat (ed.) The excavation of Khok Phanom Di: a prehistoric site in Central Thailand. Volume 3, the material culture part 1 (Reports of the Research Committee of the Society of Antiquaries of London 50): 105-118. London: Society of Antiquaries of London.

OOTA, H., K. KUROSAKI, S. POOKAJORN, T. ISHIDA & S. UEDA. 2001. Genetic study of the Paleolithic and Neolithic Southeast Asians. Human Biology 73: 225-31.

O'REILLY, D.J.W. 1998. Nong Nor phase one in a regional context, in C.F.W. Higham & R. Thosarat (ed.) The excavation of Nong Nor: a prehistoric site in Central Thailand (University of Otago studies in prehistoric anthropology 18): 509-522. Dunedin: University of Otago; Oxford: Oxbow.

PILDITCH, J.M. 1993. The personal ornaments, in C.F.W. Higham & R. Thosarat (ed.) The excavation of Khok Phanom Di: a prehistoric site in Central Thailand. Volume 3, the material culture part 1 (Reports of the Research Committee of the Society of Antiquaries of London 50): 119-76. London: Society of Antiquaries of London.

PUREEPATPONG, N. 2006. Recent investigations of early people (late Pleistocene to early Holocene) from Ban Rai and Tham Lod rock shelter sites, Pang Mapha district, Mae Hongson province, northwestern Thailand, in E.A Bacus, I.C. Glover & V.C. Pigott (ed.) Uncovering Southeast Asia's past. 38-45. Singapore: NUS Press.

REID, L.A. 1994. Morphological evidence for Austric. Oceanic Linguistics 33: 323-44.

RISPOLI, F. 2008. The incised and impressed pottery style of mainland Southeast Asia: following the paths of Neolithization. East & West 57(1-4): 235-304.

SCHMIDT, W 1906. Die Mon-Khmer Volker: ein Bindeglied Zwischen Volkern Zentralasiens und Austronesiens. Btaunschweig: F. Vieweg und sohn.

SHOOCONDEJ, R. 2006. Late Pleistocene activities at the Tham Lod rockshelter in highland Bang Mapha, Mae Hongson province, northwestern Thailand, in E.A. Bacus, I.C. Glover & V.C. Pigott (ed.) Uncovering Southeast Asia's past. 22-37. Singapore: NUS Press.

SOARES, P., J.-A. TREJAUT, J. LOO, C. HILL, M. MORMINA, C.-L. LEE, Y.-M. CHEN, G. HUDJASHOV, P. FORSTER, V. MACAULAY, D. BULBECK, S. OPPENHEIMER, M. LIN & M.B. RICHARDS. 2008. Climate change and postglacial human dispersals in Southeast Asia. Molecular Biology and Evolution 25(6): 1209-1218.

SOLHEIM, WG. II 1972. An earlier agricultural revolution. Scientific American 226(4): 34-41.

SORENSEN, P. & T. HATTING, 1967. Archaeological investigations in Thailand. Volume 2, Ban Kao. Part 1: the archaeological materials from the burials. Copenhagen: Munksgard.

TAYLES, N.G. 1999. The excavation of Khok Phanom Di: a prehistoric site in Central Thailand. Volume 5, the people (Research Reports of the Society of Antiquaries of London 61). London: Society of Antiquaries of London.

VINCENT, B.A. 2003. The excavation of Khok Phanom Di: a prehistoric site in Central Thailand. Volume 6, the pottery (Research Reports of the Society of Antiquaries of London 70). London: Society of Antiquaries of London.

WIRIYAROMP, W. 2007. The Neolithic period in Thailand. Unpublished PhD dissertation, University of Otago.

XIE G., Q. LIN & C. PENG. 2003. Gexinqiao Neolithic site in Baise city, Guangxi. Archaeology 12: 3-6 (with English abstract).

YEN, D.E. 1977. Hoabinhian horticulture: the evidence and the questions from northwest Thailand, in J. Allen, J. Golson & R. Jones (ed.) Sunda and Sahul. Prehistoric studies in Southeast Asia, Melanesia and Australia: 567-99. London: Academic Press.

ZHANG, C. & H.-C. HUNG. 2010. The emergence of agriculture in southern China. Antiquity 84: 11-25.

C.F.W. Higham (1), Xie Guangmao (2) & Lin Qiang (2)

(1) Department of Anthropology, University of Otago, P.O. Box 56, Dunedin, New Zealand (Email: charles.higham@otago.ac.nz)

(2) Guangxi Institute of Cultural Relics and Archaeology, 34 Minzu Ave., Nanning 530022, Guangxi, China

Received: 3 March 2010; Accepted: 4 July 2010; Revised: 7 July 2010