Journal of Worm Prehistory, Vol. 3, No. 3, 1989

The Later Prehistory of Mainland Southeast Asia Charles H i g h a m I

Until the recent expansion of research into the prehistory of mainland Southeast Asia, the development of civilizations employing Indian religious and political concepts had been seen as a major fulcrum in the area's history. The prehistoric sequence in the area is reviewed, with particular attention being paid to evidence for ritual, display, exchange, and ranking behavior within lowland sedentary communities. These date to the two or three millennia preceding Indian contact. It is argued that already by the end of the prehistoric period, there were complex centralized societies with a long tradition of recognizing status differentials among individuals, affiliated groups, and communities. The sequence incorporates the exploitation of rice, and bronze and iron working. The chronology of these innovations is compared with that in China. It is not possible at present to argue for single or multiple origins, but the area of southern China is suggested as a crucial one in further research to resolve these issues. KEY WORDS: mortuary behavior; ranking; exchange; domestication; craft specialization;

bronze working; copper mining; iron technology;SoutheastAsia; China.

INTRODUCTION This essay concentrates on the valleys of the Red, Mekong, and Chao Phraya rivers and the intervening uplands, an area corresponding in size to the Tigris-Euphates drainage system, the Danube valley from the delta to Vienna, or Mesoamerica from Teotihuacfin to Tikal (Fig. 1). During the past three decades, considerable progress has been made in delineating the principal chronological and cultural characteristics of this region, although it would be misleading to suggest that there is overall agreement, particularly ~Department of Anthropology,Universityof Otago, P.O. Box 56, Dunedin, New Zealand. 235 0892-7537/89/0900q)235506.00/0 © •989 Plenum Publishing Corporation

236

Higham

!!ii!iiiiiii!iiiiiiii!i!iiiiiiiiiii iiii!i!i!iiiiiii!!iii i~L~'/ "'::i:i:i:i:i:i:i:i:i:i:i:~i:!:!:!ii::.. : Dung Dau ~ ~" - Co Loa , Lung tlea -VietKite i::::' ~o

X/

Citau Can j • Dung Son

2

GULF OF B A C BO

\

7a River

Ban Don Ta Phet

¢

GULF

o

::~!"

98OE

!~)!

100°

104o

108°

Fig. 1. Map of mainland Southeast Asia showing the sites and places mentioned in the text. on the dating issue. From the early Holocene, the uplands of the area were occupied by small and probably transient groups of foragers usually described as Hoabinhian. It was in the Vietnamese provinces of Hoa Binh and Bac Son that Colani and Mansuy undertook pioneering fieldwork. The prehistoric material from the small rockshelters they excavated was ascribed to either the Hoabinhian or the Bacsonian cultures (Colani, 1927; Mansuy, 1924).

Later Prehistory of Southeast Asia

237

Biological remains from these sites indicate a catholic diet, ranging from the mammals of the canopied forest to fish and shellfish. In the upper levels, they encountered a trend toward the polishing of stone adzes where previously the stone technology had centered on the flaking of large river pebbles. Pottery, too, is often encountered in the later contexts. Similar remains from other rockshelters, particularly those examined by Gorman in northern Thailand, Pookajorn in central Thailand, and Mourer in Cambodia reveal that this upland adaptation was widespread (Gorman, 1971; Pookajorn, 1981; Mourer and Mourer, 1970). Spirit Cave is perhaps the best known of these rockshelters, because of claims for a very early development of agriculture there (Solheim, 1972). This is a small shelter located on the edge of a steep hillside a sharp 30-min climb from the nearest regular water source. The faunal remains are compatible with hunting and trapping within a closedcanopy forest. Yen's evaluation of the plant remains does not sustain the notion that the prehistoric occupants were concerned with food production, and it is argued that this rockshelter was one of many which gave shelter to mobile foragers. In the case of nearby Banyan Valley Cave, Gorman obtained radiocarbon dates which suggest occupation as late as the first millennium A.D. This dating, together with Yen's (1977) diagnosis of the rice remains from this site as probably coming from a wild species, emphasis the long-term stability of such mobile foragers adapted to the upland evergreen forest habitat. Considerable progress has recently been made in identifying the timing and location of early rice cultivation. This issue entails an appreciation of the series of sea-level changes which occurred after the end of the Pleistocene. The Southeast Asian mainland is surrounded today by extensive areas covered by a very shallow sea. The rapid rise in sea level, which was first manifested about 8000 B.C., submerged the former middle and lower reaches of the Red, Mekong, and Chao Phraya rivers and resulted in the loss of an immense area to human settlement. By 6000 to 4000 B.C., the sea rose higher than its present level, flooding what are now the lower plains of these three rivers. As a consequence, a mantle of marine clay blankets the Bangkok Plain up to a depth of 14m (Takaya, 1969). It is apparent that the rising sea drowned prehistoric settlements occupied during what may well have been a critical phase in the prehistory of Southeast Asia. The inundation of coastal settlements, which corresponds to part of the time span when upland rockshelters were occupied, is a major problem in any attempt to understand the area's prehistoric past. Some light, however, has been shed on this issue by recent research in the Bang Pakong valley, Central Thailand. The Bang Pakong is one of the four large rivers which flow into the upper Gulf of Siam. In its lower reaches, it cuts through clays laid down under marine or brackish conditions, and part of this area was chosen in 1984

238

Higham

for an intensive site survey linked with the excavation of a series of prehistoric and early historic sites. The choice of the area was strongly influenced by the presence, 22 km from the present coast, of the site of Khok Phanom Di. Trial soundings by Thai colleagues had shown that this mound, which covers 5 ha and rises to 12m above the surrounding plain, had been occupied when the sea was in the immediate vicinity (Suchitta, 1984; Noksakul, 1981; Pisnupong, 1984). The first stage in the research program was to undertake a site survey (Higham et al., 1987). This soon posed a problem. Although a number of low prehistoric mound sites were identified, our attention was also drawn to the presence of prehistoric sites stratified under deposits of marine clay. Clearly, early prehistoric settlements had been inundated as the sea level fluctuated. Identifying such submerged sites will be technically demanding. This situation encouraged a program of core boring in association with the excavation of Khok Phanom Di, which took place during the first 7 months of 1985. Bernard Maloney, who extracted the cores, has now completed his analyses of the sedimentological buildup in association with the pollen record. Eight AMS dates for the cores have been obtained. It is evident that the prehistoric mound built up over marine sediments, which reach a depth of about 5 m in its vicinity. The AMS dates suggest that these marine clays were being deposited from about 6000 B.C. Maloney has considered the pollen and fern spore frequencies and their state of preservation in conjunction with the concentrations of charcoal fragments in the cores. His conclusions indicate periodic episodes of burning, reflected in charcoal concentrations (Maloney et al., 1989). The earliest peak was associated with a rise in fern spores but not in grass pollen. It may have been the result of a natural conflagration or of human interference at such a distance from the coring location that grass pollen did not carry to it. This episode has been dated within the millennium 4800-5800 B.C. There was a second period of burning dated between 4500 and 5300 B.C., which again was associated with a rise'in pteridophyte spores. The third burning phase was matched by a sharp decline in the frequency of pollen from trees and a rise in grass pollen. Some of the grains fall within the known size range of pollen from rice. At the same time, there were increases in the pollen from plants which are known today as rice-field weeds. This phase has been dated to between 3960 and 4710 B.C. While all these episodes of burning could have been caused by one of several factors, the third was most probably the result of human interference. It resulted in an environment which favored the proliferation of plants such as rice, which are adapted to freshwater swamps. While it is dearly necessary to identify the prehistoric settlements associated with such an activity before final conclusions can be drawn, there remains clear primafacie evidence that the coastal tracts of the Bang Pakong survey area were settled by rice cultivators by at least the fifth and possibly the sixth millennium B.C.

Later Prehistory of Southeast Asia

239

While Khok Phanom Di itself was occupied when the site was situated on a major estuary, the third and second millennia B.C. saw a major expansion of settlement into the interior, low-lying river valleys of Southeast Asia. This expansionary phase of settlement involved communities which raised domestic animals, probably cultivated rice along swamp margins, and buried their dead in inhumation cemeteries in association with, inter alia, pottery vessels, jewelry, animal remains, and stone adze heads. We find such settlements near tributary streams up the course of the Mekong River, in northeastern Thailand, the Middle Country of the Red River valley above its confluence with the Black River, and surrounding the floodplain of the Chao Phraya River. It was within these communities that the knowledge of bronze-working spread. Two major copper mining, smelting, and casting locations are now known, which supplied ingots and finished artifacts for exchange. Ingots were cast using bivalve stone molds or clay molds into axheads, spearheads, bangles, arrowheads, and beads in lowland communities far removed from the sources of either copper or tin. Until recently, no metal-working involving copper rather than one of its alloys had been identified. Recent research in the Khao Wong Prachan valley, central Thailand, however, has recently disclosed a tradition involving copper casting (Bennett, 1988). The dating of the adoption of copper and bronze working is not yet settled. I have proposed that it occurred rapidly, with rivers and coasts as the channels of communication, from about 1400 B.C. (Higham, 1988b). Bayard (1988) and White (1986), however, prefer an initial date toward the end of the third millennium B.C. The resolution of this problem may well be achieved once we have a consistent series of radiocarbon dates from Non Pa Wai, a major smelting site in the Khao Wong Prachan valley. Pigott (personal communication) has recently obtained two radiocarbon determinations in the late third millennium B.C. from a clear copper-working context there. If further dates confirm this chronological context, then my more conservative estimate will be set aside. Again, the later cultural layers in a number of these sites reveal the spread of iron-working technology, using forged bloomery iron. The dating of this development is less controversial, with the period from about 500 to 600 B.C. onward being preferred. The latter half of the first millennium also witnessed two major developments from outside Southeast Asia, both of which had a considerable impact on the increasingly centralized chiefdoms which were developing in the lowland river valleys. The first involved the expansion of the Han Chinese to the south. This predatory policy saw the Red River Delta become increasingly sinicized to the point of incorporation as a series of commanderies (provinces) of the Han Empire (O'Harrow, 1979; Wheatley, 1983). The second saw the arrival of Indian traders off the coasts

240

Higham

of Southeast Asia, bringing with them a new range of exotic goods for exchange and political and religious concepts of statehood and government. As surviving Chinese accounts of Southeast Asia and inscriptions in Sanskrit add documentary sources to the archaeological record, so we can discern the development, from about 200 A.D., of larger, statelike polities in the coastal tracts best placed to control the burgeoning trade with India and China. Rather than reiterate well-rehearsed arguments on this chronological framework, I have chosen to explore a theme which, in my view, penetrates to the heart of human adaptation there, and in which it is argued that we can identify a long-term historic unity. The results of this discussion, and the sequence summarized above, are then set in a broader geographical frame of reference, which takes into account recent research into Chinese prehistory. The theme concerns domesticity, not necessarily of plants and animals but, rather, of people (Wilson, 1988; Higham, 1989). Within this context, a domestic community is sedentary, with an investment in permanent houses and concern with ownership of local resources often expressed in the development of specialized skills to exploit them. Life in such sedentary, domestic communities provided scope for competitive emulation in terms of rank not only between individual families, but between larger affiliated groups, including whole communities. The potential for inequality inherent in exchange transactions through gifting requires that particular attention be paid to the presence of exotic goods, particularly in mortuary contexts. Two further issues are exposed by this approach. The first is the degree to which long-term historic trends can be detected both within the prehistoric period and between it and the development of historic states. Thus, whereas in China, Chang (1986) has been at pains to stress the essential continuity in the development of historic states, in Southeast Asia this important development has been seen as a watershed, with Indian influence being the stimulus to change. An essential part of the discussion which follows is the view that Indianisation fueled an already complex cultural pattern with a long local ancestry.

THE ALLEGED FULCRUM: SOUTHEAST ASIA IN THE FIRST MILLENNIUM A.D. In addressing this theme, I commence not at the beginning of the sequence, or the end, but at the alleged fulcrum about two millennia ago, when Indian and Chinese expansion reached Southeast Asia. Illumination of

Later Prehistory of Southeast Asia

241

this period relies not only on the results of archaeology, but increasingly on historic documents and inscriptions carved usually in Sanskrit, but often too in a vernacular text. In about the middle of the third century A.D., the Wu emperor in Southeast China, shorn from access to the silk route by his northern rival, sent two emissaries to Southeast Asia probably to reconnoiter an alternative route to India and Rome. His servants, Kang Dai and Zhu Ying, reported that the area had walled cities and a king who taxed perfumes, gold, silver, and pearls. There was a writing system and rice was cultivated. These allusions, which survive in the seventh-century History of the Liang Dynasty, also describe an expansionary line of rulers, beginning with one Hun Panhuang, who subdued rivals on the borders of his domain and installed his sons and grandsons in their place. One grandson, a man of considerable prowess known as Fan Shiman, "used troops to attack and subdue neighbouring kingdoms, which all acknowledged themselves his vassals. He himself adopted the style of 'Great King of F u n a n ' " (Wheatley, 1961, p. 15). The two visiting Chinese encountered a representative of the Indian Murunda king, a visitation from India which probably already had a history of several centuries, for in the History of the Early Han, we read of a journey to Southeast Asia which probably took place about 100 B.C. Wheatley (1961, p. 8) has translated the relevant passage as follows: "Officials and volunteers put out to sea to buy lustrous pearls, glass, rare stones and strange products in exchange for gold and various silks. It is a profitable business for the barbarians, who also loot and kill." Most agree that these early references are concerned with the coastal regions from the deltaic plains of the Mekong and Bassac rivers across the Gulf of Siam to the Peninsula. Despite the problems involved in assessing Chinese histories written centuries after the event, it is hard to set aside the central issue, that there were centralized political structures by the beginning of the first millennium A.D., involving the principle of male succession to overlordship. Exchange of exotic items and centralized accumulation of resources within rice-growing communities took place. Archaeological validation of this finding is not new. It is almost half a century since Malleret (1959-1963) excavated within the enceinte of Oc Eo. This is a rectangular site covering 450 ha, set within four moats and five ramparts. Aerial photographs disclose that is was a node in an extensive canal network linking it with other centers and the sea. It is likely that smaller canals divided the interior of the site into wards. On the basis of both looted and excavated material, it is possible to provide a glimpse of the role played by the occupants of this site not only in a most extensive trade network, but also as a center for the manufacture of jewelry. Medallions of Antoninus Pius (A.D. 138-161) and his successor, Marcus Aurelius, evidence

242

Higham

ultimate contact with the Roman Empire, just as a Chinese mirror reflects exchange with the Empire of China. There is much Greco-Roman jewelry as well as Iranian coinage and a series of rings and seals bearing inscriptions in the Brfihmi script dated by de Casparis (1979) to the late first and early second centuries A.D. Tin amulets were locally cast in double stone molds, and glass beads were manufactured. The list of rare materials found at the site encompasses most of those available from Southeast Asia and beyond and indicates exchange contacts reaching far and wide. It includes diamond, amethyst, jet, amber, malachite, orpiment, sapphires, rubies, corindon, serpentine, feldspar, antigonite, topaz, beryl, zircon, jadeite, garnet, opal, jasper, onyx, agate, carnelian, gold, silver, and quartz. From the vantage point of Oc Eo and related centers, we can look forward to the establishment of a series of centralized polities, as well as back to the various prehistoric communities which preceded it. The former path, which traverses the very difficult territory of the so-called Indianized states of Southeast Asia, reveals several issues germane to this essay. Political and social life in the low-lying plains of the Mekong and Chao Phraya river systems and along the coastline of Vietnam increasingly revolved round a centralized court. The duration of these courts and their boundaries fluctuated, depending upon the capacity of the overlord to attract followers. Attracting followers turned on a number of factors. It seems that one of these was success in war against rivals. Another was the projection of an image through religious merit. Siva was particularly favored. Ascetic devotion to this deity was perceived as having the capacity to confer divine qualities. Wolters (1979) has stressed this process, leading to one seventh-century overlord, Jayavarman I, becoming "an incarnate portion of the god." A second leader, igfinavarman, was described in one of his inscriptions as being "like the sun in the sky, radiating an intolerable majesty, the issue of the revered kings of the earth. He was anointed with sacred water, provided blessings and was foremost among the virtuous." In order to express such qualities in a tangible form, it was the practice of such overlords to build impressive stone monuments to the gods and to endow the temples with gifts. The several monuments at Ig~napura, the court center of Tg~navarman, were ringed by walled enclosures, which, along with the temples themselves, were embellished with designs of Indian inspiration. Vickery (1986) has recently reexamined the inscriptions bearing texts in archaic Khmer, stressing that the provision of water was a further important issue in maintaining followers. These court centers were located in low-lying terrain which suffers a long dry season. The maintenance of dependents in permanent communities, particularly with the development of centers of population, would have necessitated the provision of water and food during the dry season. Thus, we can identify several important variables

Later Prehistory of Southeast Asia

243

in the development of the court societies: the magnetism of the overlord, his perceived divine qualities, and his strength measured in armed followers and ability to sustain adherents. These become strikingly evident with the foundation of an enduring and great court center at Angkor during the early ninth century A.D. Quite apart from the scale of this site, which basically recreated heaven on earth in stone, the overlords ensured water for agricultural purposes through the excavation of reservoirs, the largest of which had a capacity of 70 million m 3 of water. Groslier (1979) has suggested that the water was reticulated into downstream rice fields when natural precipitation fell below the minimum required for a successful harvest. One of the critical aspects of this phenomenon of a court society was the importance of display. Advertising status in the court of Angkor was as finely tuned as at the Versailles of Louis XIV. Indeed, the very name of the ruler who built Angkor Wat as his temple-mausoleum was Sfiryavarman, which in loose translation from the Sanskrit, means "Sun King." We can appreciate this propensity to courtly display through two contemporary accounts. The first describes Igfinavarman giving an audience in a great hall every third day. He wore a cap covered in gold and precious stones while reclining on a couch made of different kinds of aromatic wood. He was attended by five great officials and many lesser functionaries. This description refers to the seventh century A.D. We can pick up the same thread in 1297 A.D., through the eyes o f Z h o u Daguan, a Chinese visitor to the court of Indravarman II at Angkor, who described that when the king goes out, troops are at the head of the escort; then come flags, banners and music. Palace women,numberingfrom three to fivehundred, wearing flowered cloth, with flowers in their hair, hold candles in their hands, and form a troupe. Even in broad daylight, the candles are lighted. Then come other palace women, carrying lances and shields, the king's private guards.., carts drawn by goats and horses, all in gold, come next. Ministers and princes are mounted on elephants, and in front of then one can see from afar, their innumerable red umbrellas. After them come the wives and concubines of the king, in palanquins, carriages, on horseback and on elephants. They have more than a hundred parasols, flecked with gold. Behind them comes the sovereign, standing on an elephant, holding his sacred sword in his hands. The elephant's tusks are encased in gold. (Chandler, 1983, p. 75). The point of these quotations is simply this. Did such court societies, with their propensity for display, pageantry, and symbolism in architecture, emerge solely as a result of Indian "influence," or did Indian traders and brgthmans encounter local communities which already displayed a concern for status and rank? Until serious research into the prehistory of Southeast Asia began in the early 1960s, historic court societies had no local antecedents. This vacuum was filled, in the influential writings of Coed&, by emphasizing the seminal

Higham

244

nature of Indian influence on, essentially, local societies which were simply organized in almost every respect (Coed,s, 1968). As we survey the data upon which to reassess the nature of these prehistoric societies, we are still confronted by an uneven record when compared to areas endowed with a longer tradition of serious research into prehistory. We have at our disposal broad expanses of darknesss illuminated by the occasional beam of light representing a major excavation: Nevertheless, it remains possible to address the chosen theme in three successive contexts. The first concerns the nature of early settlement at the estuary of a major river. I then consider the expansion of prehistoric communities into the interior river valleys. The last context involves a number of instances where there is some evidence for the establishment of regional centers before or overlapping the onset of Chinese and Indian interest in the area.

COASTAL SETTLEMENT ON THE GULF OF SIAM We have seen that pollen cores taken from the vicinity of Khok Phanom Di suggest that burning and the encouragement of rice occurred in this region at least by the fifth millennium B.C. The pollen cores continued into the period when Khok Phanom Di was occupied, and high frequencies of grass and rice-field weed pollen continued to be associated with very high readings for charcoal particles. The 18 radiocarbon dates from cultural contexts in the site reveal occupation over a period of about six centuries, from 2000 to 1400 B.C. An assessment of the paleoenvironment of the site has been made on the basis of several complementary approaches. Maloney's pollen analyses, for example, have shown that it was occupied at a time when mangrove predominated, indicating a sheltered shore in the vicinity of the site. Among the foraminifera and ostracodes recovered from the cultural layers, McKenzie has identified specimens reflecting estuarine or delta backwater conditions, as well as weakly brackish conditions capable of supporting rice plants. In her analysis of the avifauna, West has recognized the remains of the Sarus crane (Grus antigone), stork (Ciconia sp.), and pelican (Peleeanus), all of which frequent swamp or river margins. Mason has also identified among the shellfish remains clear evidence for mangrove species, as well as those adapted to fresh or weakly brackish water and mudflats. Kijngam, who has worked on the remains offish, turtles, and crabs, has drawn attention to the presence of Lates calcarifer, a large anadromous species particularly valued today for eating. The biological remains from the site and its environs are unanimous in pointing to a settlement located on a sheltered, major estuary near a mangrove-fringed shore. An examination of the soil distribution allows us to recognize this estuary and the location of the site in relation to it (Fig. 2).

Later

Prehistory

of Southeast

Asia

245

. . ' : ~ ' ~ Bang Pakong River j ) I V,Y/'~"~ .............. 1 /./) . . . . . . . . . . . . . .

"~]'i}')

f .

)...'.'.'.".:':"'~.1~.)'

fJ f o r m e r tidal fiat / . ~ (brack3shdeposlt)

f

~/.:.:.....~~>~._[b,,..~qI!tKHOK • , ' . ' . . , . ' .

• ''+'.'.•

~.

.

.

.

.

.

.

P H A N O M DI

:

~

.

.

:

.I .

"...'..'.'t

. . . . . . . . . . . . .

+,

/ -.~

,'"

,,,

"'~

--

\

1"1

x

~,

,

2

-

.v . . - . . . . . - .

(

".

.. . . .-:.)

"% (,

. . . . . . =.' -.-., -..t:..i. -. -. -..,

"

' " "

. • ",

'/'~

:.,~'

.'...;" - .... ,'

• •

,.-" ..,/'%,

b j

~

"

\t

~

"'I"

7>

"-,,

"i t,' 'V

. ) ," . . . .

"-

•

.

.

-'-'n

.'. . . . . --'<

-

•

.,~

" ":"'~J'£__~

~,.~...Z z___z.e_

'-qX'

. . . . .

"i_ ~ i~.~i

,. -

~

;~.~,~_i;i.)

. . . . . .

" i

" i

2-.-

"-' " ' ~" \"m-a.477_.-~--*~

"~ . . . . . . . .

.-.t~,_

--"

/ ...... ,_~

t

.

" ~ . . . . .

~.\ . _ .- . - . . . % /

)

•

<.~

~i.~--- i

t ~

" I. "~,-"'. /

:.......... ,, \ I

.

........ .>'t.;.......... ~

-

.

~ ~ I 1 \ i :/ L.k-.~ - - - , ~ ~ , _ A " ~ ' . w T H . - " : " .~

. . k'~o

,

i

~ . _ ~ a

J ' " " -"/

II I r /

')

...s

. ' . " 77"i ",'.,'." > . ." . .;I ...... •; . . . . . It

.

¢~

..~.. ~ . , : . . . . , ' ~ . )

'.'.....

L

~,

f',

~

'.'<'.,','.'.,,,',',

. . . . '. . . . j. . .

.,

IP

t,. ilt

' .

or

,

\.:..::...\L..L~

,, ( .

4

"

~

-" ( m a r i n e d e p o s i t ) " . )

,

m

' .'.'.'.3

. . . . . . . . . . . . . . .

...............

t"--.../~JS. f k, / . , , , " ' t.)-.

~

I

.

....7.."

• .'..-..'.

•

.

f '

~

"'N"

• .:

•

::

r Z h i g h e r terrace a l l u v i u m

.~-~....,-

~.

''5,.')'~'"

i!i)! ! Fig, 2.

M a p o f t h e B a n g P a k o n g s u r v e y a r e a s h o w i n g p r e h i s t o r i c sites so f a r i d e n t i f i e d .

In addition to a diet including fish, shellfish, crabs, and turtles, there was also much evidence for the consumption of rice. This includes not only the remains of rice chaff in feces, but also its presence within the occupation layers and as a tempering agent in the pottery. In terms of material culture, over 400 shell knives have been recovered, on which the wear is most compatible with rice harvesting (T. F. G. Higham, 1988). Pisnupong (1988)

246

Higham

has also described granite hoes and sandstone grinding stones from the site. It is becoming increasingly clear that rice was harvested from the freshwater swamps in the site's vicinity. The excavation undertaken in 1985 involved one square measuring 10 x 10m. This configuration was chosen in the hope that it would reveal the spatial distribution of burials and, perhaps, structures. In the event, a cultural stratigraphy 6.8 m deep was uncovered. All but the top meter was laid down when the site was located on the estuary. Later occupation took place after the sea level fell, and the coastal environment was replaced by a wooded, inland habitat. Before turning to the mortuary data, it is necessary to consider the way in which the cultural deposits accumulated. The basal 1.2m comprises a series of ash and midden spreads, clusters of hearths, and numerous postholes. Among the material items, there are clay anvils for molding pottery vessels, fish hooks, net sinkers, and polished stone adze heads. It is felt that these remains indicate settlement including such domestic activities as shaping and firing pottery vessels. Two caches found near the very base of the cultural deposits contained well-used stone adze heads, anvils, and pebbles used to burnish pottery vessels. Thick accumulations of ash may have resulted from the open bonfire method of firing pots. Six human burials were found within these layers, setting a pattern for mortuary ritual which was to endure throughout cemetery use. Bodies were placed in a grave oriented so that the head pointed to the rising sun (Fig. 3). In most cases, the body was covered with powdered red ocher and wrapped in beaten bark winding sheets. Bodies were interred alongside and over others in enduring groups or clusters, each cluster including the remains of infants, children, and adults. A detailed analysis of these has made it possible to reconstruct several genealogies (Fig. 4). During the entire mortuary sequence, the occupants of the site were favored by access to a major river estuary with its abundant food resources. The shellfish of the mudflats and crabs of the mangrove could be collected by young and old, while anadromous fish came up the river. The river gave access to a hinterland which provided resources of stone, as did the then coastline, while maritime exchange would have brought shell jewelry to the site. Its occupants would then have had an advantageous position from which to control the distribution of such exotic jewelry, as well as their own pottery vessels, to the sites in the hinterland. Specified areas for closely related family members were established within two or three generations of initial settlement. The basic mortuary rite was maintained throughout the 20 or so generations represented. These clusters are interpreted as the collective burials of members of families related to each other and to common founding ancestors. Burials were accompanied

Later Prehistory of Southeast Asia

247

o c

.

.

.

. C~

.

.

.

.

.

.

.

.

Q) .

.

.

.

.

.

.

.

.

.

.

.

~

.

.

o

.

•

.

0

.

.

.

.

.

. . . . .

.

.

?

~

.

,.'-~a---a--~

: : : :o: : : :

::

.

kJ

\

'~

.

o

.

.

•

.o . . . . . .

~ o

u

.

.

q

.

.

.

"

9 o

::o:i:0 . . . . . .

0

o

~

o

o

0

ooo U

("~

~\\ Jk'~ l\ I 21~ I V ] " o

/I

o) i

< i

g

Eig. 3.

I

H

G

g

E

D

C

)i i: B

A

The burials of Mortuary Phase 3 at Khok Phanom Di showing pits and midden distributions relative to the collective burial areas.

by the digging of pits and probably feasting. Food remains were placed in these pits. Some families display a relatively high wealth expressed through grave goods; some contemporary families were poor. The wealth of the rich adults was matched by similar treatment for their infants. It is suggested that such wealth and status were not permanent but oscillated between families. This could reflect the aspirations and qualities of individuals. While the lines represented in clusters C and F continued over a long term, those in E, D, and B effectively terminated during MP 2 or early MP 3. This could reflect their extinction or fissioning. During MPs 3 and 4, women assumed increasing importance as the producers of pottery. Their predominance in clusters C and F is seen as evidence for the development of matrilocal residence, with husbands being introduced from other settlements or from other lineage clusters at Khok

248

Higham --

o]m

r

~,Iv-'

oln

-TA-~

'"A

~ 0 "'7 ,.'~

i0,,.

--().~

12l,,

137 r-L~ I

: Z~---i oI . . . 61.

MP 3

I

I

I

~

~o ....

')~

F - - q ~ 3~A = 039 4] ~

2g ~/..] 025

]0 47

I 35 03~ -¢Om

]o;~

Fig. 4.

Me2 11,5 --q

Mp3

= 056 so

i

"dP4

~I~ ]IJl

r---T

--

I

4~ MP4

Ti~27

=015 17 --TII i ~Z~ = O 13 14 I~

]o~ ~

MP5 --

7

Two putative genealogies based on clusters C and F from Khok Phanom Di. (Note that there were further generations before and after this sequence.)

Phanom Di. The dominant position of women potters reached a peak during MPs 5 and 6. First, the family of cluster F achieved prominence, but this was followed by a reversal, with the cluster C family dominating the descendants of cluster F. Much energy expended on individual graves is noted, but high status was not permanently retained by one lineage. Again, we find oscillating wealth and status. The provision of wealthy graves for infants and children continued into MP 7, but this was the final chapter in the history of this cemetery. The environmental change which occurred with the fall in the sea was accompanied by the end of the prehistoric mortuary activity. This interpretation calls on the question of the place of the community represented at Khok Phanom Di and contemporary settlements in its orbit. During the course of a site survey in 1984, a number of other prehistoric sites were identified (Fig. 2). The only site rivaling Khok Phanom Di in terms of size and height was Khok Karieng. This mound was probably also located in an estuarine situation, about 15 km southwest of Khok Phanom Di. Any points raised about this site are speculative, because no excavations have been undertaken there. To judge from its size, however, it is a site which might have rivaled Khok Phanom Di in terms of both its strategic, estuarine position and the wealth of its occupants. In addition, a number of small, low mounds were found in situations suggesting that they were located in the hinterland, away from the coast and favored estuarine conditions. It is considered likely that fine pottery vessels and shell jewelry were items of exchange between the inhabitants of Khok Phanom Di (Fig. 5) and other

Later Prehistory of Southeast Asia

249

ii

I

i !

i

__.1__ 0

1

10 cm

0

P

lOcm

[

~

/

/

i°

Fig. 5. Artifacts from Khok Phanom Di: top row, shell bracelets; middle row, turtle carapace ornaments; bottom row, bone harpoon and stone adze.

sites in the Bang Pakong Valley. Some insight into the extent of the exchange orbit of Khok Phanom Di has been provided by Pisnupong's analysis of the sources of the stone found at the site. As may be seen in Fig. 6, sandstone was obtained from a distance of 90 kin, andesite and other stone used for adzes from up to 60 kin, and the granite for hoes came from 45 km to the south (Pisnupong, 1988).

250

Higham ,-.- ,, . . . . . . . . . "J'l ,,

(

Earn rhea m o e n g , . . _ _ . i - ~

W--',

---.

Sa,~stone (abraders) •

( o .,.7~ ./.

"~

I

\f

t,,.-.--',L

..fl.z.._~,;,,:

,~,'-,

f

,

1

t

,.4

\

h

Lain Phraya Than Micaceous sandstone

t ,~,,~"j

(ahade~s~

Limitofbrac~kish waterdepos,ts]~'~ ~

I

I

- "'It

•

1

----

~

~

~

.... ~

t,~

//3

~..wd

A

I

n

d

,

~

s

i

t

e

(adz'es) ,-,4''a~

~,fl "~,

/ ,• j

"v,

ocanic sand and si!tstone _~.~.F (adzes)

"~

K--~"~ "~ ~ -- NN/~o~_

IDo IDo

e

"

'"~ KhokPha~n~mDi ~ ¢ " " " '

I

~ tF...--galcareous f,~ N ~ . ~ " , ~ I,, ~ sandstone p'~-~,vi. - .N~--Z.,~ "--~Q,(adzes)'

\•

"~

~ ' ' n'Ire( misc .) ~,.~.~-~__N_N_N_N_N_N_N~ Biotltegra

(burnishing stones, pounders)

rhyolitic tuff (misc.)

,,<4__

1

leucogranite (hoes)

Grey slaty shale (bangles)

0

Fig. 6,

12

24

36

48 Km

Map showing the sources of the stone found at Khok Phanom Di.

Later Prehistory of Southeast Asia

251

It is also possible that exogamy involved the movement of marriage partners between communities. A marriage partner leaving a community would have been compensated for by the payment of valuables from the recipient group. It is also assumed that some settlements were founded by emigrant groups from Khok Phanom Di and that they maintained relations with the parent site through the exchange of valuables. Under this interpretation, Khok Phanom Di began as one of a number of similarly sized settlements, status being a matter of competitive emulation between, first, the families and, then, perhaps, lineage groups. When a series of other communities increasingly participated in the exchange network, we find that the leaders of Khok Phanom Di held an advantage on at least two counts. First, being a specialist potting community, they could feed into the system a determined number of fine vessels. These, to judge from the Khok Phanom Di cemetery, were regularly employed in mortuary rituals. Being located on a large estuary, they occupied a nodal position in exchange involving exotic marine products and, in particular, shell jewelry. If we are correct in assuming that these items were used in marriage payments, those able to control their distribution would have found a predictable route to status and rank which extended their influence well beyond the confines of Khok Phanom Di. We could add a third advantage. While food for subsistence was probably not a concern at this estuarine community, food for prestige may well have been. The provision of food surpluses destined for ceremonial feasting, as in the death rituals, would have come much more easily to them than to those in more marginal inland communities. It is, therefore, suggested that the deep-seated changes observed with MPs 5-6 occurred as the leader or leaders of the community at Khok Phanom Di assumed increasing control over the regional distribution of prestige goods. The energy expenditure, exclusiveness, and wealth of Burial 15 would reflect that the acknowledged leadership of this highly ranked woman extended beyond her own community. The highly ranked echelon at Khok Phanom Di would, then, have exercised powerful influence in smaller, dependent communities. Paradoxically, this situation, which developed after about 15 generations, depended ultimately upon the control of a rare and highly advantageous location. Such centrality endured for only a brief span, and it is presumed to have disintegrated when the sea level fell. Certainly, the cemetery was abandoned at that juncture. Khok Phanom Di was a center of pottery making which resulted in the manufacture of masterpieces. These were used in mortuary rites. It is felt very likely that women potters were preeminent in the local provision of valuables. Ambitious women could, by feeding masterpieces into the exchange system, acquire further valuables such as shell jewelry, which gave them status and

Higham

252

power. We find an increasing stress on women. They came to predominate numerically over males, and by MPs 5-6, some were accorded lavish burials. It is very hard to imagine the society allowing highly skilled women to join other communities as marriage partners, and there are grounds, among the genetic markers in human skeletons, to suspect that the society was indeed uxorilocal. A concentration of valuables essential to appropriate mortuary ritual and perhaps marriage settlements would have conferred not only rank and prestige, but also power on the women of Khok Phanom Di.

THE EXPANSION OF HUMAN SETTLEMENT Several conclusions from the Bang Pakong research program are relevant in the wider context of Southeast Asian prehistory. The first is the evidence in favor of a long period of coastal settlement, involving burning the vegetation and favoring rice proliferation, beginning at least during the fifth millennium B.C. This long chronology is matched in the northern extremes of Southeast Asia in the valley of the Yangzi River and at Hemudu on the southern side of Hangzhou Bay (Fig. 9). Another conclusion is that, under the condition of sedentism, the number of people living in domestic communities in Southeast Asia can grow. An important point about Maloney's research is that the presence of coastal communities in the Bang Pakong area over several millennia before Khok Phanom Di was settled provides sufficient time for the population to have grown. Moreover, the rapid reduction in the area available for settlement due to the rising sea level would itself have fostered the expansion of human societies into hitherto sparsely occupied or unpopulated territory. If the end of some of our clusters at Khok Phanom Di was the result of fissioning of sections of the community to found new settlements, then we have some grounds for anticipating a proliferation of prehistoric communities. This in turn could have involved territorial expansion. The coastal environment buffers inhabitants from the extreme effects of the long dry season in Southeast Asia. This reflects the availability of marine food irrespective of season. It is notable today that, as one proceeds into the drier interior lowlands, increasing importance is attached to food storage. Rice is readily stored for dry season consumption, and fish are dried or fermented with salt. There are several inland areas which have been examined for prehistoric settlement. These include the Middle Country above the junction of the Red and Black rivers and three separate and carefully selected parts of northeastern Thailand (Chantaratiyakarn, 1984; Wichakana, 1984; Wilen, 1987). A feature of all four is the finding that settlements seldom

Later Prehistory of Southeast Asia

253

covered more than 2 or 3 ha and that site location away from major river floodplains but near the midcourses of small tributaries near low-lying soils was preferred. Despite several intensive site surveys, no settlement has been found which approaches the fifth millennium dates inferred from the Khok Phanom Di core data, nor is there any well-defined evidence that these extensive plains were occupied by hunter-gatherer groups. At present, it is acknowledged that initial settlement probably occurred sometime during the third millennium B.C., although the data are fugitive and contradictory. For the Red River valley sites, which are collectively known to us as the Phung Nguyen culture, there are only three radiocarbon dates. All are from late contexts and are, at 2o- range, 1885-1415 B.C. (Bin-830), 1330-885 B.C. (Bin-1409), and 19751545 B.C. (Bin-891). There are four sites from northeastern Thailand which have basal levels belonging to the earliest known occupation in the area: Non Nok Tha, Ban Chiang, Non Kao Noi, and Ban Phak Top. A large sample predating the initial occupation of the last-named site has been dated to 3970 B.P. +_ 90 years (2635-2185 B.C., P-2407). No datable charcoal was found in the limited area excavated at Non Kao Noi. The dates for the lower horizons at Ban Chiang and Non Nok Tha have been the subject of numerous papers. I prefer an initial occupation for both sites between 2000 and 1500 B.C., but for the purposes of the present argument, an extra millennium for the prebronze contexts at either site is not of importance. The initial settlement of the Middle Country of Bac Bo is represented by a dense cluster of sites ascribed by Vietnamese prehistorians to the Phung Nguyen culture. It has been subdivided, on the basis of pottery typology, into three successive groupings. Of the 52 sites excavated, only 11 have yielded bronze. Metal, in the form of corroded fragments of bronze, comes only from the latest contexts; no bronze artifacts have yet been recovered. Consequently, in terms of metallurgy at least, most Phung Nguyenm sites appear to equate with early Non Nok Tha and Ban Chiang in northeastern Thailand. There are relatively few radiocarbon dates for this sequence. The Phung Nguyen phase appears to have developed into early Dong Dau by about 1500 B.C. The date of the earliest Phung Nguyen contexts is not yet known but they probably belong to the third millennium B.C. Nearly all sites are located above the confluence of the Red and Black rivers. They cover !-3 ha and are found on slightly elevated terrain near small stream confluences. The principal excavated site is at Phung Nguyen itself (Hoang Xuan Chinh and Nguyen Ngoc Bich, 1978). It covers 3 ha, of which 3960m 2 has been excavated, revealing a large sample of most interesting material culture (Nguyen Ba Khoach, 1980). No trace of bronze was recovered, but there was a substantial sample of pottery and stone artifacts. The stone adzes take a variety of forms, but the rarest was the shouldered variety, of which only four

254

Higham

were found. This compares with 777 examples of quadrangular form, some of which had been sharpened so consistently that they are broader than they are long. When one adds stone adze fragments for which the shape is indeterminate, 1138 adzes or adze fragments were recovered. There are also 59 small stone chisels, some with cutting edges only 10 mm wide, and almost 200 grinding stones were found bearing grooves which result from sharpening the stone adzes and chisels. Stone projectile points were rarer. Three reach the dimensions of a spear-point, but the remainder are more likely to be tanged arrowheads. The inhabitants of Phung Nguyen also fashioned stone rings, mostly in nephrite. The sample of 540 specimens has also subdivided into eight types, based on the shape of the cross section. Most are rectangular, but some are much more complex, having a range of ribs and flanges. While some are large enough to rank as adult bracelets, others have small diameters and were designed either for children or perhaps as earrings. Stone beads, mostly tubular and measuring up to 1.3 cm in length, were also made. There was a vigorous tradition of working clay. Clay pellet-bow pellets matching those found so often in the General Periods A and B sites in Thailand were found, as were clay net weights, but most attention was accorded the manufacture of clay vessels. It is on the basis of changing decorative styles that the Vietnamese archaeologists have advanced their threefold subdivision of the Phung Nguyen sites (Ha Van Tan, 1980). One site, Lung Hoa, has yielded a sample of 12 burials in an excavated area of 365 m 2. The excavators felt that there was some evidence for differential wealth despite the small sample. There were, for example, two burials associated with stone bracelets, beads, and earrings, as well as pottery and polished adzes. The other burials were accompanied only by pots and adzes. Clearly, larger samples are necessary before this possibility can be tested. There are sufficient data from northeastern Thailand to encourage Bayard (1984) to propose a fourfold cultural sequence. His General Period A is defined as "the development of at least semi-sedentary agricultural c o m m u n i t i e s . . , with noticeable but fairly weak social ranking." General Period B saw "the appearance of bronze technology with an increase in social ranking, but still at a simple-ranked level." The last period of relevence to this paper, General Period C, involved "the introduction of iron technology with a more marked increase in ranking." Non Nok Tha is a small (ca. 1.1-ha), low mound located on the eastern edge of a monadnock known as Phu Wiang. It is a difficult site to interpret. Its prehistoric cultural deposits hardly exceed 1 m in depth, within which 217 burials or mortuary contexts were recovered. These burials have been subdivided into 11 phases, known as Early Period 1-3 and Middle Period 1-8. None of the 18 graves belonging to the first two phases was associated with

Later Prehistory of Southeast Asia

255

bronze artifacts. The basic problem with this site remains the controversy surrounding its chronology (Bayard, 1988; Higham, 1988b). Bayard prefers a duration of at least two millennia, while I have suggested a time span of a few centuries within the second millennium B.C., with the earliest bronze artifacts appearing about 1400 B.C. It is not easy to define the nature of social organization on the basis of the 18 burials from early prebronze contexts. The sample includes 2 adult males, 3 adult females, and 11 children, with 2 graves being indeterminate. Bayard (1988), however, has noted the wealth of Burial 14, which contained the remains of a 6-year-old child. The body was accompanied by six pottery vessels, two strings of shell disk beads around the waist, a bone tool, and cattle and pig bones. Pottery vessels abound at Non Nok Tha, and other early child graves contain three to eight pots each. Apart from these, the most common artifact associated with children was the small stone adze blade. One of the adult males was buried with nine pots, two strings of shell disk beads, two adze blades, and a grinding stone, while the second had only two pots, the bones of a pig's leg, and two grinding stones. One female had no grave goods, but the second had eight pots, and the last, five pots and an adze blade. One of the problems involved in the interpretation of these data is that the origin of the pottery vessels is not known. As Vincent (1987) has shown at Khok Phanom Di, the identification of centers of production is an important key to understanding the probable role of a community within an exchange system. In this context, the question of exotic versus local production of pottery vessels at Non Nok Tha, some of which display technical skills of a very high order; needs clarification. The stone adze blades may have been exchanged from quarry sources in the uplands to the west of the site, while the shell beads could have been fashioned from local freshwater shells or from exotic marine specimens. Despite these interpretative difficulties, Bayard (1988) has concluded that "some degree of ascribed ranking existed at Non Nok Tha from its first use as a burial place." If it is difficult to assess whether the jewelry and ceramics at Non Nok Tha are exotic, there is no doubting the status of bronze. Neither copper nor tin, which were employed in the earliest metal artifacts from the site, is found in the vicinity. A bronze axe head was found in one grave belonging to the third mortuary phase (Early Period 3). The following phase, Middle Period 1, yielded two graves containing crucibles and one with a sandstone mold. Thereafter, bronzes, and the crucibles and molds which attest local casting, are found, but never in quantity. Bayard has noted a rise in the quantity of grave goods during the period of early bronze and has suggested that an increase occurred in the degree of ascriptive ranking following the introduction of bronze working. Noting the presence of two forms of pots which

256

Higham

rarely occur together in Middle Period graves, he has further suggested that two different affiliative groups interred their members in the same cemetery over a period which, following his chronology, would have lasted for well over a millennium. While I have questioned this overall interpretation (Higham, 1988b), I agree that the mortuary data from the site are compatible with a ranked social structure. Bayard's suggestion that this may have been connected with tenure of the best, low-lying land suited to rice is entirely reasonable. Again, some measure of control over access to rarities, such as metal ingots, could have been a contributory factor. Pigott and Natapintu have added greatly to our appreciation of the nature of early copper metallurgy in Southeast Asia through their investigations at Phu Lon, near the southern bank of the Mekong River, about 160 km north of Non Nok Tha, and in the vicinity of Khao Wong Prachan, in the central plain of Thailand just to the northeast of Lopburi (Fig. 1). It is evident that considerable energy was devoted to the mining, dressing, and crucible smelting of a malachite ore at Phu Lon, activity so far dated as commencing between 1500 and 1000 B.C. (Pigott, 1984). Several sites have been investigated in the Khao Wong Prachan area (Pigott and Natapintu, 1988). Nilkhamhaeng, which has been dated from the late second millennium B.C., comprises a 2 to 3-ha mound reaching a depth of 4-5 m. Much of this accumulation is of thin deposits of host rock and crushed slag. Both the slag and the associated crucible fragments attest to smelting activity. Just as clay cylinders and anvils were found with burials at Khok Phanom Di, so at Nilkhamhaeng, one of the two inhumation burials was associated with a quantity of copper mineral placed near the cranium. A second important site, Non Mak La, was examined in 1985 by Surapol Natapintu (Natapintu, 1988; Bennett, 1988). Bennett has reported that an excavation square measuring only 5 x 4 m yielded 3500 kg of copper slag. She has further characterized the slag as being derived from an incompletely weathered sulfide ore. The prills were found to contain up to 5.5% arsenic. The hardening properties of arsenic, even when present as 1-2% in the finished artifact, are well known and should be taken into account when it is considered that an arrowhead from Non Mak La was made from unalloyed copper containing up to 0.8% arsenic. Non Pa Wai covers about 5 ha, and virtually the whole mound was built up of the residues of copper smelting. The earliest cultural context comprises a cemetery of at least 12 inhumation graves. There were few grave goods, but one included a pair of bivalve molds and a second individual was interred with a socketed copper-base ax head. It is this critically important context which has provided two radiocarbon dates in the late third millennium B.C. If these are confirmed by a larger series, then it will pose most interesting issues on the nature of exchange, since Khok Phanom Di, an estuarine center

Later Prehistory of Southeast Asia

257

which engaged in much exchange for stone, was occupied between 2000 and 1400 B.C. Yet none of the 154 graves yielded any traces of metal. According to Natapintu (1988), the pottery from this first context at Non Pal Wai matches that from a mortuary site of Huai Yai, burials from which have yielded shell H-shaped beads similar to those found in the latest mortuary context at Khok Phanom Di. At Huai Yai, however, these H-shaped shell beads were found with evidence not only for stone bangle and adze manufacture, but also for copper-working. We are thus edging closer to an understanding of the cultural and chronological framework for early copper-working in central Thailand. Metal-working was under way from 1500 B.C. or so, along with a vigorous exchange in exotic objects: shell, stone, pottery, and copper. This situation may well have been present already by the closing centuries of the third millennium B.C. If so, then exchange involving copper artifacts did not incorporate Khok Phanom Di or, at least, the areas excavated there. Since the occupants of this strategic estuarine site were so clearly given to the exchange of exotic goods, the absence of metal would pose interesting practical and theoretical issues. After a period of abandonment, the second phase at Non Pa Wai saw the accumulation of the remains of metal processing to a depth of about 3 m. Crucible fragments abounded, as did slag cakes which probably originated in smelts from individual crucibles. It seems likely that the copper thus obtained was poured into individual cup-shaped ceramic molds in order to produce ingots. Evidently ingots were locally cast in ceramic molds, although it is likely that many were exchanged in their raw form. The excavators have suggested that the mortuary evidence from these sites indicates that there were specialist metal workers. This upper smelting context at Non Pa Wai probably belongs to the later second and the first millennia B.C. It is possible to follow the course of bronze-working away from the immediate vicinity of the mines as a result of excavations undertaken at two other northeast Thai sites, Ban Chiang and Ban Na Di. While bivalve clay molds were preferred at the numerous sites in the Chao Phraya valley, in northeastern Thailand bivalve sandstone molds were preferred. During their excavations at Ban Chiang in 1975, Gorman and Charoenwongsa uncovered the remains of a clay furnace associated with crucibles and bronze. Similar clay-lined pit furnaces were recognized during the excavation of Ban Na Di (Higham and Kijngam, 1984; Higham, 1988a). It seems that ingots of copper and tin were secured through exchange from sources remote from the site. The vast majority of the metal castings was of a tin bronze, although one fish hook was made from copper (Rajpitak and Seeley, 1984). These were then mixed in small ceramic crucibles, raised to melting point in charcoal-charged furnaces, and then cast into bivalve stone molds. Ceramic molds at Ban Na

258

Higham

Di differ from those found in central Thailand, since they were used to cast bracelets using the cire perdue technique. Axes, spearheads, bracelets, fishhooks, and arrowheads predominated among the objects cast. Further sites involved in this activity are found in the valley of the Mekong and its tributaries, in Bac Bo, and along the coast of Vietnam. Through the associated mortuary remains at Ban Na Di we can obtain some insight into the social organization associated with a community which possessed metal-working skills. This site was first occupied when the properties of bronze were known, probably between 1400 and 1000 B.C. (Higham and Kijngam, 1984). Excavations took place in two parts of the site, about 30 m apart. A tight cluster of inhumation burials was found in each area. Each contained adult males, females, children, and infants and accumulated over the same time span. Twenty-three burials were found in one area (A) and 37 in area B. It is interesting to note that the incidence of certain grave goods was more frequent among those buried in area B. Thus, of the 12,438 shell disk beads, only 5% were found in area A graves. All five stone bracelets were found in area B. These were fashioned from exotic stone and were evidently so valued that, when broken, they were repaired by boring two holes and joining the pieces with a bronze tie. Evidently, the wirelike bronze ties were cast in place through the holes (Maddin and Weng, 1984). All the clay cattle figurines were found with area B interments, and of the 30 bronze artifacts found, 25 were associated with area B. One area B male was buried with 13 bracelets fashioned from marine trochus shells. When iron became available at the very end of the use of this cemetery, all the iron implements were found in area B burials. The excavators have concluded that the community responsible for this cemetery probably recognized at least two different social groups, of which one was richer and more highly ranked than the other (Higham and Kijngam, 1984). The subsistence activities of these people included the exploitation of rice. According to the analysis of the grains recovered from Ban Na Di, this rice was probably cultivated (Chang and Loresto, 1984). The faunal remains indicate a catholic diet, ranging from the maintenance of domestic cattle, pigs, and dogs to the hunting of large and small deer and the collection of shellfish. By far the most common food source represented was fish. We are reminded that the site lay within the floodplain of a substantial lake, which even within living memory was given to extensive flooding during the rainy season. The basal levels at Ban Na Di incorporate many sandy lenses laid down by floodwater. This diet provided a good plane of nutrition. According to the analysis of the human bones from Ban Na Di, bone development matched that for modern inhabitants of Ohio and a sample drawn from Finnish farmers (Houghton and Wiriyaromp, 1984).

Later Prehistory of Southeast Asia

259

I have argued before that the small village communities of this period when bronze was circulating widely, but before the development of iron technology, represent autonomous communities within which there was more than one unequally ranked social group (Higham, 1983; Higham and Kijngam, 1984). The exchange of exotic goods between such communities is seen within this framework as the surviving evidence for interlinked exchange networks within which there moved such commodities as copper, tin, or bronze ingots, high-quality stone used for adze blades and ornaments, sandstone for bronze molds, grinding and sharpening stones, shell jewelry from coastal sources, and high-quality pottery vessels. Less tangible valuables which equally could have circulated were silk cloth, fragments of which adhered to the bronze bracelets from Ban Na Di (Pilditch, 1984), beeswax used in bronze casting, salt, rice, and feathers. No information forthcoming during the last 5 years has made it necessary to modify this interpretative framework. On the contrary, the excavations at Khok Phanom Di, Non Pa Wai, Phu Lon, and Ban Tha Kae (Natapintu, 1984) have reinforced it by identifying production centers for pottery, copper, stone jewelry, and shell bracelets. Within such a system, as Dalton (1977) has stressed, the exchange of valuables involves embedded economic, technological, and social behavior. The measure of the success of an individual, an affiliated social group, or a community lies in competitive emulation. It is argued that such competition is an important factor in appreciating the sequel, the trend toward a dichotomy between large, wealthy, and powerful centralized communities and the smaller, dependent villages within their orbit.

A TREND TOWARD CENTRALIZATION: 500-1 B.C.

Most workers agree that during the latter half of the first millennium B.C., and perhaps for a century or two earlier, there were profound cultural changes in the riverine and lacustrine floodplains of Southeast Asia. These areas, now the foci of modern states and concentrations of population, witnessed the transition from autonomous to centralized societies and, with it, a marked intensification in the fields of production, agriculture, and ceremonial. The area was exposed to the expansion of the Chinese and Indian civilizations, which brought in their wake new products and novel ideas. The clearest manifestation of this change occurred in the Red River delta and the valleys of the Ma and Ca rivers, an area known to the Vietnamese as Bac Bo. Here, there is a less extreme climate than in most of low-lying mainland Southeast Asia because the dry season is tempered by moist winds which move across the Gulf of Bac Bo. On reaching land, they form a low cloud cover often associated with drizzle. As Gourou (1955) has shown, this

260

Higham

moist climate permits two crops of rice a year on favourable soils. The Red River is given to sudden rises in level, a characteristic compensated for in the last few centuries by the construction of earth banks. Its load of silt is so great that the delta is still advancing rapidly. During the prehistoric period, the delta area would have been subjected to widespread flooding, and indeed, the Phung Nguyen sites are located in the more elevated middle country above the delta proper. Thereafter, there was a continuing process of expansion involving occupation of the upper and middle delta itself. The period in question, named after the cemetery and occupation site of Dong Son, saw an increasing exposure to Chinese military expansion. Our appreciation of the Dong Son people is, indeed, colored by literary allusions to the area which have survived in Chinese documents (Wheatley, 1983). Viewed from the Han capital, the Dong Son people were the most distant of several groups known as the southern barbarians, and there is no doubt that the Chinese encountered a society controlled through paramount chiefs of high status. Although most of our information on Dong Son comes from burial sites, Co Loa is one major settlement which probably dates to the third century B.C., although it was much added to. It is located 15 km northwest of Ha Noi on the floodplain of the Red River. There are three sets of ramparts, the two outer being moated and oval in plan and the innermost being rectangular. There is no doubt that it was added to over several centuries, but there is some literary evidence that Co Loa was established as a center during the third century B.C. This has been confirmed by recent archaeological excavations which revealed Dong Son-style pottery under the middle rampart. In 1982, a complete Dong Son-style bronze drum containing over 100 socketed bronze plowshares was located near the central part of the site. One of the most important points about Co Loa is its size. The outermost ramparts cover about 600 ha, 200 times the area of Phung Nguyen. No concentrated survey of settlement sites has been undertaken for the Dong Son phase, and most of our information is drawn from cemeteries. This source of information is also restricted in the sense that many cemeteries are found in acidic soils and human bone has not survived. At Dong Son itself, for example, burials were recognized more from the disposition of artifacts than the presence of human remains. At Lang Ca, 314 burials were found but hardly any human bone. Even so, the excavators noted that a small group of graves was differentially rich and contained axes, daggers, situlae (large bucket-shaped vessels), and spearheads. Further evidence for rich burials has been found at Viet Khe and Chau Can (Luu Tran Tieu, 1977). Both have yielded opulent boat burials. The richest at Viet Khe was interred in a hollowed tree trunk about 4.5 m in length. It contained over 100 artifacts. No human remains were found, but there is little doubt that the bronze weapons,

Later Prehistory of Southeast Asia

261

receptacles, and utensils were grave goods. This particular coffin also contained bronze bells, relatively small bronze drums, and even a painted wooden box. Some of the metal goods in the coffin are said to be of Chinese derivation (Pham Mirth Huyen, personal communication). Interment in such impressive boat-coffins surely signaled the burial of a particularly important person. Wooden artifacts, as well as the human remains, were found inside some of the eight boat burials found at Chau Can. These are smaller than the Viet Khe example and contain fewer grave goods: bronze axes and spearheads with their wooden hafts in place, earrings made of a tin-lead alloy, and a bamboo ladle. The suggested local origin of the Dong Son bronze industry has been supported by recent research at Dong Son itself. The first use of the cemetery involved burials accompanied by pottery vessels but only a few bronze axes, spearheads, and knives. These are said to equate with the Go Mun phase in the Red River valley dating to 1000-500 B.C. (Ha Van Tan, 1980). The second phase saw a proliferation of bronze artifacts and extension of types to include daggers, swords, situtae, and drums. It is dated to the period 500-t B.C. The third and last prehistoric phase comprises burials containing objects of Chinese origin such as seals, coins, mirrors, and halberds. Ha Van Tan (1980) has suggested a date within the first century A.D. for these burials. Indeed, the site was later used for interments in the Han style after the incorporation of the delta region into the Han Empire during the first century A.D. We can learn much about the people of Dong Son from these surviving bronzes. As has been demonstrated at the Dong Son cemetery, both the quantity and the range of bronze artifacts increased greatly from about 500 B.C. This intensification of production can be illustrated in two ways. The Co Los bronze drum, for example, weighs 72 kg and would have entailed the smelting of between 1 and 7 tons of copper ore. One burial from the Lang Ca cemetery contained the remains of a crucible and four clay molds for casting an axe, a spearhead, a dagger handle, and a bell. The mold for the bronze dagger handle suggests that it may have been intended as a bimetallic weapon, with the blade itself being made of iron. Bimetallic spears are known from both Dong Son and Ban Chiang. The crucible, while retaining the shape of the earlier ones found in Dong Dau contexts, is much larger. Indeed, it could have held 12 kg of molten bronze (Vu Thi Ngoc Thu and Nguyen Duy Ty, 1978). Many novel artifact forms were developed, many revealing a high degree of decoration. The drums, situlae, and rectangular ornamental plaques suggest an interest in ritual and ceremony, while the daggers, swords, and halberds reflect concern with personal weaponry. Nor was agriculture overlooked: the socketed bronze plowshares represent agricultural intensification.

262

Higham

The Dong Son metal-worker was a master in the difficult field of bronze casting, but the status of local iron-working is not clear. During the Dong Son phase, Chinese iron technology reached a high level of proficiency. Their methods involved the demanding system of iron casting as well as forging. It is known from documentary sources that iron objects were exported from China to the south during this period, and the presence of some bimetallic spears in Dong Son contexts, with iron blades with bronze hilts that had been cast on, shows that iron, if by now locally worked, remained rare. It is also instructive to note that the rich burials of Viet Khe and Chau Can do not contain any iron grave goods. The status of iron during the Dong Son phase requires detailed analyses to determine the ratio of cast to forged artifacts. Evidence for local iron smelting is also basic to an appreciation of iron and its importance to the local people. At present, the direct Chinese contact with Dong Son people is the most likely means whereby knowledge of iron casting reached Bac Bo. The specialized bronze-workers there would doubtless have been interested in its properties. Iron did not threaten the central role of bronze in agriculture, war, or ceremonial. The casting of bronze plowshares is a direct application of metallurgical skills to the intensification of agriculture, which took place in conjunction with the increased repertoire and skill of the local bronze ateliers. Large bronze drums are one of the outstanding artifacts of the Dong Son workshops. The bronze drums themselves are not confined to Bac Bo but have a widespread distribution in southern China (Wu et aI., 1986). The ownership of drums is still a hallmark of high status among the ethnic minorities of Southern China, and their abundance in Dong Son contexts indicates a society given to ritual and display. It is fortunate that the Dong Son bronze-smiths decorated their drums with scenes drawn from the world around them: these allow a glimpse into the activities of the very lords described in Chinese documents. We can, for example, recognize the importance attached to elegant boats equipped with cabins and fighting platforms. They were crewed by paddlers and carried plumed warriors. The spears, halberds, and arrows found in aristocratic graves are seen in action, either being fired or, in one case, chastising a captive. The Dong Son drums themselves are represented mounted in sets of two or four, with the drummers on a raised platform. Houses were raised above ground level on piles, gable ends being supported by posts and decorated with bird-head carvings identical to those seen on the canoe prows (Fig. 7). These, and other scenes of musical instruments being played, bring home to us the importance of music in ritual activities at the Dong Son aristocratic centers. Although the culture of Dong Son is the best-known expression of the transition to the centralized chiefdom in mainland Southeast Asia, it was not alone. Indeed, the trend to centralization was widespread. Again, the

Later Prehistory of Southeast Asia

263

m Fig. 7. Scenesderived from the decoration on Dong Son drums: top row, two plumed warriors; middle row, a group of drummers sitting on a raised platform over their drums; bottom row, a musical quartet. foremost symbol of the new aristocracy, the D o n g Son drum, is by no means confined to the valleys of the Red and M a rivers: the distribution of such drums covers much of southern China where rich chiefdoms contemporaneous with the D o n g Son phase are well-known. F a r less research has been undertaken in the Chao Phraya plains than in Bac Bo, but our appreciation of the importance of this area is rapidly growing with, first, the identification of a vigorous exchange between coastal and inland communities in the second millennium B.C. and, second, the long tradition of copper-working at sites in the K h a o Wong Prachan valley. Indeed, the surge in copper production is clearly attested there by at least 1000 B.C., and it is within this context that we can, for the first time, identify a plausible context for the local origins of iron working. While analyzing the copper slag from N o n M a k La, Bennett recognized the use of a rich hematite ore as a flux in copper smelting. This hematite would "have been necessary to prevent the reduction of the highly silicious

264

Higham

ore to a porous and brittle sinter product" (Bennett, 1988, p. 130). The hematite deposits are locally available. In this situation, it is perhaps not surprising to find iron artifacts at Non Mak La, one of which was found in a mortuary context dated to about 600-500 B.C. (White, personal communication). Iron, the ores of which are widely distributed in lowland Southeast Asia, was rapidly adopted during the second half of the first millennium B.C. The Chao Phraya plains command the eastern edge of the Three Pagodas pass, one of the avenues of Indian contact. The Gulf of Siam also permitted the movement of people and exchange of goods by sea, which allowed the occupants of the coastal tracts to participate in the expansion of exchange which took place during the late first and early second millennia. That this involved contact with the Dong Son people is documented at one of the most intriguing sites, Ongbah. This great cavern, located in the upper reaches of the Khwae Yai River valley, has yielded the remains of three sets of paired drums of clear Dong Son affinities (Sorensen, 1973, 1979, 1988). The burial technique was extended inhumation in wooden boat-shaped coffins. A radiocarbon determination of one such coffin gave an estimate of between 400 and 1 B.C. There were also several intact burials within the cave, of which 10 were excavated. These lacked the wooden boat-coffins, but both groups contained a similar assemblage of iron implements. Sorensen has suggested inferior social status, rather than chronological change, to account for such a disparity in relative wealth. The 10 poorer burials have yielded several well-preserved iron implements, thereby affording insight into the uses of this metal, presumably during the last century or two of the first millennium B.C. Burial 5 incorporated an iron hoe on the chest. Burial 6 included a tanged knife, possibly a spearhead, and five beads. Other burials yielded chisels and arrowheads. One contained seven iron objects, all placed near the ankles. The coffin burials included strings of beads around the waist and neck; they also formerly included bronze and iron artifacts including earrings, bracelets, and at least one bronze vessel made of a very high tin-bronze, an alloy increasingly seen as a major innovation of General Period C. As has been mentioned, a set of drums was found near, and probably associated with, the coffins. These elaborate objects were decorated with zones of different motifs, including flying birds, human beings, and geometric designs. The wealth of this assemblage may well be due to the control of the local deposits of lead-ore. Contact of one sort or another between the occupants of the Chao Phraya and the Red River valleys is also indicated by a more recent find at Doembang Nangbuat. Suchitta (1985) has reported a series of artifacts of clear Dong Son affinities, the remains of ornamented drums and bowls. Whether this find reflects exchange or an actual movement of people from Bac Bo remains to be seen. Pisnupong (personal communication) has also

Later Prehistory of Southeast Asia

265

recently identified a bronze Dong Son-style drum on the eastern margin of the Gulf of Siam. Ban Don Ta Phet is located on the western margins of the Chao Phraya lowlands. Its position on a terrace commanding low-tying terrain is similar to the location of settlements on the Khorat plateau. This cemetery site has been the object of several excavation seasons since its discovery in I975 (You-di, 1978; Glover, 1980, 1983; Glover et al., 1984) and it already plays a central role in our understanding of the late prehistoric period in central Thailand. Unfortunately, bone does not survive in the acidic soils, but excavation has, as at Dong Son, revealed dispositions of artifacts, with the occasional scrap of bone, which are clearly the remains of human interments. These graves are richer than any yet documented from earlier sites in the Chao Phraya area. The number of complete or fragmentary pots averaged 20 per burial, a far higher figure than in the General Period B burial assemblages. The burials were also associated with tools and weapons made of iron, ornaments and bowls made from bronze, and both stone and glass beads. The blades of socketed iron spearheads were often bent back to break, or ritually kill them, prior to burial. One iron implement looks very like the modern sickle blade, while the application of iron technology is revealed by the socketed tips for either digging sticks or hoes, harpoons, and knives. Bennett's (1982) analysis of the 13 iron artifacts has shown that they were forged probably from the local limonite ores and then edge-hardened by hammering to produce what Bennett describes as a "fairly homogeneous low-carbon steel." This would have allowed the iron to retain a tough and easily sharpened cutting edge. The beads were manufactured from agate, carnelian, and glass. These are all exotic and recall specimens dated to the second and third centuries B.C. in India. There is a very strong probability that the beads are Indian exports. The burials at Ban Don Ta Phet have also furnished a series of thinwalled bronze bowls which exhibit interesting features (Rajpitak and Seeley, 1979). They are decorated with incised motifs and, occasionally, representations of people. The alloy is a high (19-21%) tin-bronze which is very brittle and hard to work yet imparts a gold-like color to the finished article. The number of such bowls at the site and the long tradition of established bronze-working point alike to local and, indeed, specialized manufacture. A few similar bowls have been found in India. Two come from the Bhir Mound at Taxila dating to the late first millennium B.C. (Marshall, 1951). The inferences are clear: exchange linked Indian and Southeast Asian societies toward the end of the first millennium B.C., and some individuals at Ban Don Ta Phet were interred with imported beads made of glass and semi precious

266

Higham

stone. It is unfortunate that the acidic soils at the site have ruled out the preservation of large enough samples of charcoal for radiocarbon dating, but one determination made on the basis of the organic content of pottery indicated use of the site at least during the second and third centuries A.D. (Glover et al., 1984). An important step toward a more complete understanding of this period has recently been taken by Ho (1984). Within the context of her analysis of the material from the excavations of Khok Charoen, she undertook a series of site surveys. These suggested strongly that, after an initial occupation during what she calls "the early metal age" (equivalent to our General Period B), there was a move toward centralization wherein one site grew differentially large and, it is assumed, exercised political dominance over others in its area. Indeed, Ho was able to identify three distinct areas each dominated by its large, and eventually moated, site. The centers are about 30 km apart from each other. She has ascribed these sites to a "high metal age," which saw a marked increase in bronze-working and the initial use of iron. This equates culturally and chronologically with General Period C. It is important to note how her results from the Chao Phraya valley match those proposed by Welch (1985) and Moore (1986) in the Mun valley and Higham and Kijngam (1984) in the Chi valley. There are, then, some grounds for suggesting that there was a trend toward centralization during the mid to late first millennium B.C. The chronological relationship between this trend and direct or indirect contact with Indian coastal traders remains to be determined. Evidence for the exchange of goods between coastal communities in Southeast Asia itself are apparent in the Dong Son-style drums from Ongbah, the few bronzes in the Dong Son tradition, and a two-headed ear ornament characteristic of the Sa Huynh culture at Ban Don Ta Phet (Suchitta, 1985). Events on the Khorat Plateau toward the end of the first millennium B.C. are best illustrated by the site of Non Chai (Bayard et al., 1986). This most important site is located in the upper reaches of the Chi catchment. It is on a small surviving tract of the old middle terrace, in such a location as to command low-lying alluvial soils moderately suited to rice cultivation. Its area cannot be stated with precision because the site has been removed for road-fill. According to plans made after removal had commenced, it covered at least 18 ha, and the excavators estimated an area during its prehistoric occupation of 38.5 ha. It was, therefore, considerably larger than any known General Period B site. Proximity to flooded areas is reflected in the aquatic resources identified by Kijngam (1979). He described a considerable number of shellfish, fish, crabs, frogs, and water turtle from the middens there. At the same time, the early settlers brought with them domestic water buffalo, cattle, dog, and pig. Like the occupants of Ba Na Di and Ban Chiang, they hunted

Later Prehistory of Southeast Asia

267

extensively. The bones of deer, crocodile, rhinoceros, and many small mammals occur in the faunal spectrum. The pottery from Non Chai is dominated by red-slipped and painted wares which echo later Ban Chiang styles (Rutnin, 1979). In this context, the radiocarbon dates confirm a relatively late prehistoric settlement and a rapid buildup of cultural material. The excavators have suggested that the site was occupied between about 400 B.C. and 200 A.D. The material culture from Non Chai includes small amounts of iron slag, but in abundance only from phase IV. There are also four glass beads from contexts earlier than phase III at the site and over 200 belong to phases IV and V. The surge in the number of beads probably dates to about 200-1 B.C. Clay molds for casting bronze bracelets and bells are likewise found in phases III-V contexts, although fragments of bronze and crucible fragments were identified in layers attributed to phases II-V. The critical point about this well-dated site is that it covered at least 18 ha at some point during the period from 400 B.C. until about A.D. 250, when it was abandoned. At present, we do not know during which part of the sequence, if any, the site actually attained that area under continuous occupation. If it did, and a population figure of 50 people per ha is adopted, then the site at its maximum extent would have harbored about 1000 people. Further, if estimates of nearly 40 ha for the site's area are employed, then the population could have been twice that figure. Further down the Chi Valley, Higham and Kijngam have undertaken an intensive site survey. It was found that prehistoric sites concentrate near the low terraces of the tributary streams and fringe the extensive tract comprising the Chi floodplain (Chantaratiyakarn, 1984). One site in this area, Ban Chiang Hian, incorporates a double set of moats, a reservoir, and possibly the remains of ramparts. Three sites have been excavated including Ban Chiang Hian itself. All three yielded a distinctive red-on-buff painted ware in the lowest occupation layers. Non Noi had only a horizon of early red-onbuff painted pottery, but at Ban Kho Noi and Ban Chiang Hian, this was superceded, in the mid-first millennium B.C., by a plainer ware. At that juncture, the excavators found the first evidence for iron and the water buffalo. The moated area of Ban Chiang Hian covers about 38 ha. It significantly exceeds the size of all other prehistoric sites in the surveyed area and was clearly a special central site. As with all other moated sites in the Khorat plateau, we cannot yet answer the question of its construction date. This is a critical issue, because the existence of so large a site with such extensive earthworks is precisely the sort of evidence which is held to reflect the existence of centralized chiefdoms. There is, however, some evidence which suggests that such social groupings were present in at least the Mun and Chi valleys during the period 400 B.C. to A.D. 300.

268

Higham

The valley of the Lam Siao Yai is about 60 km southeast of Ban Chiang Hian. Here again, we encounter a large moated site. Non Dua commands an extensive deposit of rock salt as well as low-terrace soils suited to rice cultivation. The salt exposure, known as Bo Phan Khan, is surrounded by evidence of industrial activity in the form of mounds and quantities of thick-walled pottery. Excavations in one such mound exposed evidence for salt extraction to a depth of about 6 m, beginning, according to the radiocarbon date, in the first-second century A.D. Some examples of the crudely fashioned industrial wares were found during excavations within the moated site itself, which suggests that its occupants extracted the salt. The extent of the activity, measured in terms of the huge mounds which have accumulated all around Bo Phan Khan, points to extensive production beyond that necessary to satisfy local demand. Non Dua, the moated site, also yielded a d e e p stratigraphic sequence, and the initial phase of occupation has been assigned to the period 500-1 B.C. Some of the distinctively decorated rims and body sherds have been noted in phase 2 at Ban Chiang Hian, but otherwise the pottery there was not matched in the Middle Chi valley. Further information on the Khorat plateau settlement during General Phase C has been obtained by Welch (1985) and Moore (1986). Welch worked in the region surrounding the great Khmer center of Phimai (Fig. 8). His site surveys, undertaken to document and explain the beginnings of

Mun

ir a

ve)

y

/ T a i l 77m15 u e 15L~15 '

I

°

o(.~/~Lvi.,-x

tlo

.

R.

;;- --1

"

I -It

'

1

/.

-.',,

//"

,<-

'

,I "-J~ " " "l'J. (

J

/ . . . . . . .

"

1,J' .

-

-

.

_ ~ .

./',t

II

~

, . . . . . . . . . .

,~

~

.

. . . . . . . . . .

,

~

+

.

.

.

.

.

.

.

l •

,( ....

La Lot R.

(,

/:,

i

./ !--

I. 1

.

i

.

f ~ . ] . . . . . .

.

.

~ '~ "~ [

,

Land

above

175 m

,...~,o ........ -\

k

,

. t\ ~

•

•

t-'~ u- j ~,

/"

/ "')

'J

/

-X_ / ~

J

J

k kC-~ . . . . . . .

. . . .

~ ~, " . . . . 102 3 0 ' " . . . .

.( ,~

"--'J L

Fig. 8. Map of the Phimaisurveyarea, after Welch(1985).

)

h

/

\ l O k"m m - J

Later Prehistory of Southeast Asia

269

centralization in the area, revealed 15 prehistoric sites. He found a concentration of them on recent terraces elevated above the floodplain. Excavations at two sites, the most productive of which was Ban Tamyae, allowed him to establish a chronological and cultural framework for the upper Mun valley. The five test pits excavated at Ban Tamyae covered an area of 24 m 2 and revealed a cultural stratigraphy about 2.5m deep and comprising eight layers. Bronze was found in the lowest, with iron appearing for the first time in level 7. Among the well-preserved faunal material, Welch found the remains of domestic pigs and cattle, with the first positively identified water buffalo bone coming from a level 4 context. It seems that the excavations encountered occupation areas, so burial data are not available. On the basis of the very large sample of pottery, Welch subdivided his Formative Period (600 B.C.-A.C. 600) into four phases, dated principally on the basis of seven radiocarbon dates from Ban Tamyae. The sequence began with the Tamyae phase, which corresponds best, it seems, with General Period B. There followed the Prasat phase (600-200 B.C.), Classic Phimai phase (200 B.C.-A.D. 300) and the Late Phimai phase (A.D. 300-600). It was during the Prasat phase that iron-working was established. Welch was able to observe a marked trend toward centralization during the Prasat and Classic Phimai phases as some settlements, such as Phimai itself, grew larger. We return to his stimulating interpretation of this situation when reviewing possible reasons for the widespread trend toward centralization during General Period C. Moore (1986) has recently completed an innovative and important study of a sample of the Mun valley moated sites identified on the basis of the World War II photographic archive, the Williams-Hunt collection. She has analyzed 91 of these sites on the basis of their size, architecture, and distribution. Unlike the Chi valley survey described above, Moore examined the moated settlements as a technological group, rather than as a constituent of an overall settlement pattern which includes both the larger moated and the smaller unmoated sites. The Mun valley moated sites are distinctive in their circular layout and common occurrence of multiple moats and banks. Moore has calculated the area of each. The resulting data are not directly comparable with the areas of Ban Chiang Hian and Non Chai because she included the area of the moats and ramparts in her figures, whereas the area of 38.5 ha for Ban Chiang Hian was taken within the earthworks. Nevertheless, the figures are of considerable interest. In her small group, the area varies up to 20 ha. Sites of the intermediate group have areas of between 21 and 40 ha, while the large sites extend up to 68 ha. These are far greater than those estimated for General Period B villages. The distribution of these large enclosed sites includes the floodplain of the Mun River, the lower terrace, the middle terrace, and on three occasions, the high terrace. Of 91 recorded sites, 78 were located on the floodplain

270

Higham

(N = 15), tow terrace (N = 50), or low to middle terrace (N = 13). Again, compared with the General Period B villages, there seems to have been a significant expansion of settlement to include the middle and high terraces. When Moore turned to the architecture and degree to which moats and reservoirs encircled sites, she found that the elevated settlements incorporated more water-control systems and commonly exhibited more stages in their expansion. This is hardly surprising where the water supply was likely to be a particular problem in the drier middle- and high-terrace areas. In terms of chronology, Moore has suggested that the initial phase in the provision of water-control measures took place during General Period B, perhaps during the first half of the first millennium B.C. The major expansion, however, is ascribed to the period 500 B.C.-A.C. 500, that is, the currency of General Period C. One possible reason cited for this expansion of settlement away from slightly elevated terrain on the floodplains to the knolls of middle- and high-terrace land commanding tributary stream margins is the pull of such resources as timber, salt, and lateritic iron ore. Certainly, a high proportion of these sites, even on the basis of their modern names, indicates local iron smelting. The large moated sites in northeastern Thiland are very much a feature of the Mun and Chi valleys. They are significantly larger than any other site in their surrounding territory. Non Chai, Phimai, Ban Chiang Hian, and Non Dua have produced distinctive local styles of pottery, and there is some suggestion that vessels were exchanged between centers or, at least, regions. Iron-working and bronze casting were undertaken, and glass beads are included among imported items. No Dong Son import has been found in a stratigraphic context in one of the moated sites, but excavations have been minimal. Objects of Dong Son origin, however, are known in northeastern Thailand from surface of looted sites, and two axe halberds were found in a burial context at the site of Ban That (Kethutat, 1976). Given the contemporaneity of these sites with Dong Son, it is interesting that no bronze plowshares have yet been found in the Khorat plateau. If we accept a figure of 50 people per ha as a reasonable population estimate for these prehistoric sites, Non Chai would have accommodated at least 1000 people, and Ban Chiang Hian, twice that number. The earthworks at Ban Chiang Hian are substantial and would have entailed much energy in their construction. Even if the moats and reservoir were a mere 1 m deep, 100,000 m 3 of soil would have been moved in their excavation. It has been estimated that it would have taken 500 well-fed adults a year to complete the task (Chantaratiyakan, 1984). Again, the very presence of a reservoir argues for the need to supply water to a large number of people during the dry season. The point is clear: the moated sites represent a signal departure from

Later Prehistory of Southeast Asia

271

the earlier system of village autonomy and involved a far higher investment in buildings. Dating moat construction is a major priority.

THE TRANSITION FROM AUTONOMY TO CENTRALITY Research into the later prehistory of Southeast Asia has now shown that there was a tradition of living in sedentary communities for at least two millennia before the first Chinese and Indians entered the region. The occupants of Khok Phanom Di over a span of 20 generations seem to have competed for high status, if such status be indicated by the ritual of an individual's burial. Some of the women at Khok Phanom Di fashioned pottery vessels with exemplary skill, and it has been argued that their importance as the makers of the exclusive and beautiful enhanced their status positions and encouraged an uxorilocal residence pattern. Occupation of the interior plains was probably under way by the third or early second millennium B.C. by domestic groups. Bronze-working was subsequently adopted within these small village communities, which already had an established practice of exchange in exotic artifacts. These have been described as primitive valuables, and they underwrote the maintenance of economic, social, and political relationships between participating communities. It is argued that the exchange of valuables and the establishment of affinal ties promoted alliance. Imbalance, or fluctuations in access to valuables, involved oscillations in rank. Within such a system, much devolves upon the leaders of dominant lineages. They had superior access to prestige valuables, could foster affinal relationships, and could play an entrepreneurial role in the transfer of goods. The distribution of stone, marine shell, and metals, the principal surviving materials documenting such exchange, reflects goods crossing the landscape along a variety of routes and in different directions. Further research will surely illuminate the former existence of interlinked exchange networks with major concentrations in, for example, the Chao Phraya catchment, the Mekong valley, and the maritime tracts of Viet Nam. It was, it is held, exactly along such nodes, particularly in the last two networks, that knowledge of bronze traveled. The system was also internally flexible with regard to permanence of rank. It is worth pausing to review such flexibility, as it will sharpen our perception of the transition to centrality. During General Period B, sites were located in small stream valleys and along the margins of the main river floodplains. The technique of rice cultivation probably involved the exploitation of enclaves of land subjected to only a limited degree of flooding following the onset of the monsoon. However, climatic unpredictability involving drought in some years and

272

Higham

major flooding in others would doubtless have prejudiced predictable success in the rice harvest. The broad-ranging subsistence activities would, accordingly, have been a valuable buffer against the effects of climatic extremes. At the same time, however, intensifying production to permit participation in a prestige-good exchange system would have come more readily to those commanding the most extensive tracts of cultivable land. Indeed, for many river valleys, good land and communications were their principal resource: there were no local deposits of ore, stone, or marine shell. There are, then, conditions under which certain communities might prosper unduly. Among these may be numbered access to a circumscribed resource, be it salt, copper, clay, lead, tin, or good land; control over a strategic position such as a mountain pass or river crossing; or a monopoly, through the fortunes of geography, over access to prestige exotic goods such as glass beads, iron, or agate jewelry. The transition to General Period C, with its establishment of highly ranked groups in centers underwritten by intensified production and exchange, represents the growth to prominence of one settlement and its occupants over others within its orbit. In Bac Bo, we can isolate several variables which were intensified, although this area is a special case given its direct contact with Chinese expansion. In terms of agriculture, there was the advent of plowing and double cropping. As G o o d y has shown, the application of animal traction to soil preparation greatly magnifies production and thereby makes it more feasible to produce surplus food (Goody, 1976). A rice surplus can be employed to attract and maintain followers, thereby concentrating people. It may also be converted, through provisioning specialists, into visible status objects, not least the great ceremonial drums, situlae, body plaques, and, in terms of prowess, daggers, swords, and ax halberds. On the domestic scene, we can also note the elaborate decoration applied to houses, a decoration matched on the impressive boats. A further and most significant application of a food surplus lies in the provision of feasts. This aspect of ceremonial and ritual behavior is often overlooked by prehistorians, due, perhaps, to the difficulty of identifying its presence. In the case o f the Dong Son material culture, however, it is clear that the large, ornamented, bronze vessels were intended for display. It is a widespread phenomenon that a means of attaining high rank and prestige lies in providing sumptuous displays of food not so much for eating as for admiring (Wilson, 1988). This relationship has been succinctly described by Evans-Pritchard in the context of the Azande: The more the wivesthe more the labour and the more the food; the more the food the greater the hospitality the greater the following;the greater the followingthe greater the prestige and authority. (Evans-Pritchard, 1971, p. 223) This sequence of events is no less applicable to the situation in Bac Bo during General Period C, except that it is necessary to substitute water control and the plow in place of women in agricultural production.

Later Prehistory of Southeast Asia

273

There was also a great proliferation of skill on the part of the Dong Son smiths, and much more metal was mined, moved, and shaped. It is selfevident that the organization and purpose of all this activity, involving as it did the maintenance of permanent ateliers and control over far-flung ore sources, involved intensification. In this manner, permanent ranking of an elite, craft specialization, and intensified agriculture may well have interacted with each other. Nor must it be forgotten that the Dong Son aristocrats controlled the maritime and riverine routes which exposed Bac Bo first to the importation of Chinese goods, as seen in some of the offerings contained in the Viet Khe boat burial, and then to the admission of Chinese armies. There is a consistent body of evidence that the use of iron and importation of glass and exotic stone beads took place along the coasts of central Viet Nam, in the Chao Phraya area, and even on the inland Khorat plateau, during the last few centuries of the first millennium B.C. In the latter area, although shielded from Chinese contact by the Truong Son range and remote from immediate, coastal contact with Indian expansion, Non Chai, on the Khorat plateau, still grew to cover at least 18 ha, and during its occupancy (ca. 400 B.C.-200 A.D.), witnessed the use of iron and importation of exotic glass beads. At Ban Na Di, after the initial presence of a few iron objects in two graves, local iron smelting was initiated during the buildup of level 5. Higham and Kijngam (1984) have suggested that the major changes noted at Ban Na Di and Ban Chiang, signaled by the advent of red-painted pottery and associated artifacts, reflect the expansion of groups from the fringes of the emergent Chi valley chiefdoms. The same conclusion has been reached by Moore in her consideration of the Mun valley moated sites, where, for the first time, expansion into the higher country away from the low-lying marshes is documented. In the Mun valley context, the moated sites grew considerably larger than the villages of their autonomous predecessors and were, in all likelihood, centers where petty chieftains prospered on the basis of ironworking, the salt trade, and increased agricultural production to sustain their followers. The control of water through substantial moats made this move a feasible proposition. It has been suggested that increased agricultural production was, as in Bac Bo, based upon the buffalo-drawn plow and fixed fields demarcated by bunds (Higham and Kijngam, 1979; Higham et al., 1981). In more recent publications, the presence of plowing has been questioned because of the continuing lack of any artifactual evidence corresponding to that available in Bac Bo (Higham and Kijngam, 1984). This last publication advocated an agricultural system where the naturally flooded margins of the major rivers were brought into production, but not necessarily through the medium of plowing or fixed field creation. Of course, this situation would again alter were actual plows to be recovered from General Period C contexts there.

274

Higham

In approaching the issue of possible agricultural intensification through plowing and water control, Welch (1985) has offered the interesting alternative view that it began not so much in response to population pressure in enclaves of suitable marshy terrain but, rather, to cope with the marked unpredictability of rainfall which characterizes northeastern Thailand. Indeed, he is skeptical that there was ever an initial problem with population pressure during the prehistoric or historic periods in Southeast Asia. Since some aspects of intensified cultivation, such as transplanting, require increased labor at peak seasons of the year, Welch further suggests that it was intensification itself which encouraged an expansion of population. Welch's hypothesis might well be sustained by future research, but it is also stressed that agricultural intensification, particularly the increasing importance of rice production, could also have been stimulated by a desire to attract followers, and therefore prestige and power, through lavish feasting behavior. At present, we can point to various possible contributing factors. It does appear that centers developed, that iron tools and weapons proliferated, that populations grew and expanded, and that new sources of prestige goods became available. At the same time, more attention was given to the production of beautiful and prestigious possessions. We are confronted by a complex series of changes, and the reasons are more likely to lie in several interacting shifts in behavior. The growth of certain settlements and employment of iron and imported jewelry also occurred in the Chao Phraya valley. These changes are seen here as consistent with the breakdown of the long-standing affinal alliance and exchange system between independent communities. Occupants of sites such as Non Chai and Phimai had easy access to iron ore and commanded nodal positions in the upper reaches of the major Khorat plateau rivers. The Sa Huynh sites, likewise, were able to control coastal traffic. Any movement up the Chao Phraya River would involve passage through the territory of such growing centers as Ban Tha Kae. Several interacting variables can therefore be identified. Population growth and agricultural intensification are noted and were doubtless interactive despite uncertainty over which came first. But the social change was critical in that it involved the growth of centers as foci of population growth and leadership. There is some archaeological evidence that these centers controlled the large-scale production of salt, exchange of exotic artifacts, and smelting of iron ore. It is proposed that these variables and their interactions reflect the flexibility and opportunism inherent in the system characterizing General Period B, wherein those controlling the best tracts of agricultural land and access to a new range of exotic prestige goods enjoyed considerable advantages. In every instance of increased centralization, there is a recurrent thread of evidence in favor of increasing ritual and display activity. The most

Later Prehistory of Southeast Asia

275

impressive are the Dong Son bronze drums, weaponry, boat burials, and personal ornamentation. But we also see the same trend in the Ban Don Ta Phet bowls and imported semiprecious stone jewelry. In upland Laos, there is clear evidence for particular individuals being given opulent burials in huge stone jars. Along the coast of Viet Nam, jar burials included iron weapons, imported jewelry, and beautifully worked burial urns. Even in the remote Songkhram valley, the potters of Ban Chiang made outstandingly beautiful painted vessels for inclusion in human graves. Intertwined in this model is the developing exposure to Indian and Chinese expansive forces. The sequel to this early contact was alluded to above, for we have reached the alleged fulcrum, the turning point when Southeast Asian communities became increasingly exposed to Indian contact. It is suggested that exotic artifacts and ideas were accepted and adopted by people with a long history of behavior centered on life in domestic, sedentary communities. Over a period of at least two millennia, there is evidence for widespread exchange of exotic goods, rice exploitation, display of rank in mortuary rituals, and craft specialization. These new finding emphasize more of a continuity, less of a dislocation, between the prehistoric and the early historic groups described-in Chinese histories.

PREHISTORIC

SOUTHEAST

ASIA IN ITS WIDER CONTEXT

An essential factor in this appraisal of domesticity and sedentism in Southeast Asia is that the exchange of specialized products expedited the transfer of knowledge. In an area where uplands and dense forest cover are an impediment and waterways are an encouragement to exchange, we might reasonably consider the broader geographic framework within which Southeast Asian communities operated. What follows is essentially a series of questions pinpointing avenues of future enquiry, rather than arguments for or against particular propositions. At the commencement of the Holocene cultural sequence, the southern provinces of China sustained groups who intermittently occupied rockshelters and flaked implements from pebbles. Heijinglong in Yunnan, for example, is a rockshelter which has yielded ash layers in association with deer, bear, bovid, and macaque bones (Chang, 1986). An important series of open sites characterized by a flaked stone technology, but no evidence of pottery or polished stone implements, has been identified at Xijiaoshan in Guangdong Province. The so-called Hoabinhian pattern of broad-spectrum foraging evidently recurred to the north. If it is confirmed that the evidence of the pollen cores from the Bang Pakong valley reflects rice cultivation during the fifth millennium B.C., then it is intriguing to note that this is of a similar order of antiquity to the

276

Higham

outstanding evidence for rice cultivation which has been obtained from Hemudu, on the northern bank of the Yaojiang river in Zhejiang Province. Here, the lowest of four cultural layers has been dated to 5000-4600 B.C. The question posed, and one which can be answered only by considerably more archaeological research, concerns the initial development of rice cultivation. Was there one focus, from which expansion of people or the spread of the idea took place, or was there a series of foci, in which coastal or riverine sedentary communities intensified and modified rice independently of each other? We can be sure that in China, the development of both millet and rice resulted in population growth which saw the expansion of settlement over an increasingly greater area. Recognizing this, Chang (1986) has proposed the term "Chinese Interaction Sphere." These cultures, he notes, "became closely linked, and they share common archaeological elements that bring them into a vast network within which the cultural similarities are qualitatively greater than without" (Chang, 1986, p. 234). By 4000-3000 B.C., this Interaction Sphere stretched from Jilin in the north to Guangdong in the south and from western Gansu to Shandong in the east (Fig. 9). This area, measuring 2000 x 2500 km, incorporated virtually the entire coastline of modern China. Linkages between the component "cultures" of this interaction sphere center upon artifacts, including pottery forms, polished stone adze types, and items of antler and shell. In the southern part of the area, south of the Yangzi River, rice was an important component of the diet. Pearson and Underhill (1987) and Keightley (1985) have emphasized that mortuary practices among these groups also display a number of common characteristics. These include communal cemeteries, often with distinct clusters of individuals, the provision of a range of grave goods including pottery vessels and exotic ornaments, a regular but regionally variable orientation of the grave, and with time, increasing attention to status differential expressed in the location and size of the grave and the quantity of funerary offerings placed with the dead. Thus, by the third millennium B.C., the coastal and major riverine areas of modern China were occupied by agricultural communities linked through the exchange of exotic items and increasingly given to signaling status differentials in mortuary rituals. This situation poses a fundamental series of questions for the student of prehistory in Southeast Asia. Chang's (1986) work of synthesis is confined to the modern state of China, but we must ask whether the coastal communication network, part of his Interaction Sphere, continued in a southerly direction to include the Red, the Mekong, and even the Chao Phraya valleys as well as the Vietnamese coast. This could, for example, have involved the flow of goods and ideas between essentially localized exchange networks such as that reconstructed for Khok Phanom Di on the basis of the exotic stone found there. What we find in the Phung Nguyen sites of the

Later Prehistory of S o u t h e a s t Asia

Landabove 1 8 2 m

i i i l;i ii'i

277

'.i i ' . ; ; i . ; ;

;~,5<%

~ > i ;< .'#'~;->.--..yr,2~i ' ...F>. . .. .. .. .. .. .. .. .. .. .. .. . . . . . . ?<.-.~' - . .'5"7-~., ,~ev,. 7%~.172;2117222212f'~'>~.2 2 2 i 2 i 2 i 2 2.)"

;-:Land above 2473

m::

':;7:

;:

- "-

i iiii)i}

'i

•

~ • 5;':7i,;;'-.

.

¸ . i"

.

.

.

.

.

.

.

.

.

.....

,'.@ :,/, . . . . . . . .

.2.~ j ~ . Z 2 j / 3/ . . . .. .. .. .. . . . . . ........

f)

./ .

.

.

.

.

.

.

.

.

.

.

.

.

. . . J. . . . . . . . . . . . . . . . . . .

)

• .,:2',

/o 0

km

500

Fig. 9. The Chinese Interaction Sphere o f C h a n g (1986), 4000-3000/2000 B.C., set out in relation to possible cultural areas in S o u t h e a s t Asia which could have participated.

Red River valley mirrors the essential features of sites further north. There is the same extensive use of incised and cord-marked pottery. Although the rarity of graves means that our knowledge of pot shapes is not great, if one adopted the same broadly comparative stance as Chang, then one could point to forms of vessel with parallels in the southern sites of his Interaction Sphere. Stone adze blades were numerous and bracelets of

278

Higham

nephrite, amphibolite, quartzite, and jasper were manufactured (Nguyen Ba Khoach, 1980). Again, in very general terms, the mortuary behavior involving extended inhumation with a range of grave goods matches that found so extensively in China. This does raise not the arid and pointless issue of whether innovations first occurred in China or in Southeast Asia but, rather, addresses the possibility that exchange of exotic goods and ideas was a mechanism which linked communities distributed over much of the mainland of East Asia. This issue, to which no resolution is in sight, assumes particular interest when we turn to the question of bronze-working. According to Sun and Han (1981), the earliest evidence for bronze casting in China is found at the Majiayao culture site of Linjia, Gansu. A bronze knife 12.5 cm long was found, cast in a bivalve mold from an alloy comprising a 6-10% tin bronze. This find dates in the region of 3000 B.C. They note that by 2400-2000 B.C., both tin, lead, and tin-lead bronzes were being cast in the Huanghe Basin from Gansu to Shandong. Han and others (1986) have described the proliferation of bronzes particularly in Gansu, where copper ores are abundant, during the early second millennium B.C. The sequel was the supreme bronze industry of the Shang. Within the framework of a third millennium inception of bronze-working in northern China and the existence of the Chinese Interaction Sphere of exchange, the student interested in metallurgy now faces a fascinating question. Did the knowledge of the properties of copper, tin, and lead ores pass between the communities subscribing to the Interaction Sphere from an ultimate origin in Gansu? Alternatively, was the Southeast Asian bronzeworking tradition an independent phenomenon owing nothing to developments further north? It is too early to offer a judicious review of these alternatives, but several relevant points can be made. Both China and Southeast Asia were occupied, during the centuries or even millennia preceding the first metal working, by societies participating in extensive exchange in stone adze blades, for which the raw material was often located in ore-bearing uplands. There were also specialist potting communities in both areas with established pyrotechnic skills. Again, there was a widespread interest in jewelry fashioned from exotic materials by specialists. In order to consider the way in which copper smelting and casting began and spread, we need much more knowledge. We need to know more about the presence, dating, and technology of bronzeworking in the coastal area between the Yangzi and the Red rivers. It is also necessary to expand our data on the chronology of copper- and copper alloy-working in Southeast Asia. At present, choosing between a single and a multiple origin for East and Southeast Asian metal working is not possible. It must, however, be observed that the recent identification of smelting sulfide copper ores high in arsenic in the Khao Wong Prachan Valley, perhaps as early as the late third millennium B.C., tends to favor a multiple origin.

Later Prehistory of Southeast Asia

279

The issue of iron-working is slightly less cloudy. According to Han and others (1986) and Li (1975), knowledge of iron technology in China originated in the midseventh to sixth centuries B.C., if one discounts the Shang working of meteoric iron. It is well-known that, from the sixth century B.C., the casting of iron artifacts was developed in conjunction with the continued hot-working of bloomery iron, particularly in the manufacture of swords. In Southeast Asia, at least outside the immediate area of Han colonization, iron casting was not adopted, but the forging of iron spears, knives, bracelets, arrowheads, and agricultural implements became widespread during the latter half of the first millennium B.C. Bennett (1988) has suggested that knowledge of the properties of iron could well have been recognized first when hematite ore was used as a flux in copper-working. At Non Mak La, the forging of iron has been dated recently to about 600-500 B.C. (Joyce White, personal communication). From this consideration of Southeast Asian prehistory in a wider context, it can clearly be seen that we are still in a very early stage of enquiry. In my own very limited experience, I have invariably found that what I anticipate from an excavation invariably differs, and often radically, from what is eventually concluded. This is inevitable when working in an area where there is still so much to be learned. But if this essay has shown that some progress is being made and that there are fascinating issues to explore, then it will have achieved its objective. REFERENCES Bayard, D. T. (1984). A regional phase chronology for Northeast Thailand. In Bayard, D. T. (ed.), Southeast Asian Studies at the X V Pacific Science Congress, Otago University Monographs in Prehistoric Anthropology 16, Dunedin, pp. 161-168. Bayard, D. T. (1988). Bones of contention: the Non Nok Tha burials and the chronology and social context of early South East Asian bronze. Paper read to the conference on ancient Chinese and Southeast Asian bronze age cultures, Kioloa, Australia. Bayard, D. T., Charoenwongsa, P., and Rutnin, S. (1986). Excavations at Non Chai, Northeastern Thailand. Asian Perspectives 25(1): 13-62. Bennett, A. (1982). Metallurgical analysis of iron artifacts from Ban Don Ta Phet, Thailand. B.Sc. report, Institute of Archaeology, University of London. Bennett, A. (1988). Prehistoric copper smelting in Central Thailand. In Charoenwongsa, P., and Bronson, B. (eds.), Prehistoric Studies: The Stone and Metal Ages in Thailand, Thai Antiquity Working Group, Bangkok, pp. 125-135. Chandler, D. (1983). A History of Cambodia, Westview Press, Boulder, Colo. Chang, K. (1986). The Archaeology of Ancient China, Yale University Press, New Haven, Conn. Chang, T. T,, and Loresto, E. (1984). The rice remains. In Higham, C. F. W., and Kijngam, A. (eds.), Prehistoric Investigations in Northeast Thailand, British Archaeological Reports (International Series), Oxford, pp. 384-385. Chantaratiyakarn, P. (1984). The research programme in the Middle Chi. In Higham, C. F. W., and Kijngam, A. (eds.), Prehistorie Investigations in Northeast Thailand, British Archaeological Reports (International Series), Oxford, pp. 565-643. Coed,s, G. (t968). The Indianised States of Southeast Asia, East-West Centre Press, Honolulu. Colani, M, (1927). L'~ge de la pierre darts la province de Hoa Binh. M~moires d~ Service Geologique de l'indochina XIII: 1.

280

Higham

Dalton, G. (1977). Aboriginal economics in stateless societies. In Earle, T., and Ericson, J. (eds.), Exchange Systems in Prehistory, Academic Press, London, pp. 191-212. de Casparis, J. (1979). Palaeography as an auxiliary discipline in research on early Southeast Asia. In Smith, R. B., and Watson, W., (eds.), Early South East Asia, Oxford University Press, OxforJ, pp. 380-394. Evans-Pritchard, E. E. (1971). The Azande: History and Political Institutions, Oxford University Press, Oxford. Glover, I. C. (1980). Ban Don Ta Phet and its relevence to problems in the pre- and protohistory of Thailand. Bulletin of the Indo-Pacific Prehistory Association 2: 16-30. Glover, I. C. (1983). Excavations at Ban Don Ta Phet, Kanchanaburi Province, Thailand, 1980-81. South-East Asian Studies Newsletter 10: I-4. Glover, I., Charoenwongsa, P., Alvey, P., and Kamnounket, N. (1984). The cemetery of Ban Don Ta Phet, Thailand, results from the 1980-1 season. In Allchin, B., and Sidell, M. (eds.), South Asian Archaeology, Cambridge University Press, Cambridge, pp. 319-330. Goody, J. (1976). Production and Reproduction, Cambridge University Press, Cambridge. Gorman, C. F. (1971). The Hoabinhian and after: subsistence patterns in Southeast Asia during the late Pleistocene and early recent periods. World Archaeology 2(3): 300-320. Gourou, P. (1955). The Peasants of the Tonkin Delta, Human Relations Area Files, New Haven, Conn. Groslier, B. P. (1979). La cit6 hydraulique angkorienne. Exploitation ou surexploitation du sol? Bulletin de l'EcoIe Franfaise d'Extr~me Orient 66: 161-202. Ha Van Tan (1980). Nouvelles recherches pr6historiques et protohistoriques au Viet Nam. Bulletin de l'Ecole Franfaise d'Extr~me Orient 68:113-154. Han, R., Watts, G., and Kennon, N. (1986). The development of Iron and Steel technology in ancient China. Materials Australia 18(9): 12-16. Higham, C. F. W. (1983). The Ban Chiang culture in wider perspective. Proceedings of the British Academy LXIX: 229-261. Higham, C. F. W. (1984). The social structure of the Ban Na Di prehistoric population. In Bayard, D. T. (ed.), Southeast Asian Studies at the XV Pacific Science Congress, OUSPA 16, Bunedin, pp. 72-86. Higham, C. F. W. (1988a). Prehistoric metallurgy in Southeast Asia: Some new information from the excavation of Ban Na Di. In Maddin, R. (ed.), The Beginning of the Use of Metals and Alloys, Massachusetts Institute of Technology Press, Cambridge, Mass., pp. 130-155. Higham, C. F. W. (1988b). The social and chronological contexts of early bronze working in Southeast Asia. Paper read to the conference on ancient Chinese and Southeast Asian bronze age cultures, Kioloa, Australia. Higham, C. F. W. (1989). The Archaeology of Mainland Southeast Asia, Cambridge University Press, Cambridge. Higham, C. F. W., and Kijngam, A. (1979). Ban Chiang and Northeast Thailand; The palaeoenvironment and economy. Journal of Archaeological Science 6(3): 211-234. Higham, C. F. W., and Kijngam, A. (1984). Prehistoric Investigations in Northeast Thailand, British Archaeological Reports (International Series) 231(-3), Oxford. Higham, C. F. W., Kijngam, A., Manly, B. F. J., and Moore, S. J. E. (1981). The third bovid phalanx and prehistoric ploughing. Journal of Archaeological Science 8(4): 353-365. Higham, C. F.W., Bannanurag, R., Maloney, B.K., and Vincent, B. A. (1987). Khok Phanom Di: The results of the 1984-5 excavation. Bulletin of the Indo-Pacific Prehistory Association 7: 148-178. Higham, T. F. G. (1988). A Microwear Analysis of the Shell Knives from Khok Phanom Di, Unpublished B.A. Honours dissertation. Department of Anthropology, University of Otago. Ho, C. M. (1984). The Pottery of Kok Charoen and lts Farther Context, Ph.D. thesis, University of London, London. Hoang Xuan Chinh and Nguyen Ngoc Bich (1978). Di Chi Khao Co Hoc Phung Nguyen, Nha Xuat Ban Khoa Hoc Xa, Ha Noi (Vietnamese). Houghton, P., and Wiriyaromp, W. (1984). The people of Ban Na Di. In Higham, C. F. W., and Kijngam, A. (eds.), Prehistoric Investigations in Northeast Thailand, British Archaeological Reports (International Series), Oxford, pp. 391-412.

Later Prehistory of Southeast Asia

281

Keightley, D. (1985), Dead but not gone: The role of mortuary practices in the formation of the Neolithic and Bronze Age culture, ca. 8000 to 1000 b.c. Paper read at the annual meeting of the American Historical Association, New York. Kethutat, P. (1976). Preliminary report on the test excavation at Ban That, Tambon Ban Ya, Amphoe Nong Han. Journal of Anthropology, Thamasat University, Bangkok 1976:27-38 (Thai). Kijngam, A. (1979). The faunal spectrum from Non Chai. Silpakon 23(5): 102-109 (Thai). Li, C. (1975). The development of iron and steel technology in ancient China. Kaogu JFuebao 2: 1-22 (Chinese). Luu Tran Tieu (1977). Khu ?do Co Chau Can, Archaeology Institute, Ha Noi (Vietnamese). Maddin, R., and Weng, Y. (1984). The analysis of bronze wire. In Higham, C. F. W., and Kijngam, A. (eds.), Prehistoric Investigations in Northeast Thailand, British Archaeological Reports (International Series), Oxford, pp. 112-116. Malleret, L. (1959-1963). L'Arch~ologie du Delta du Mekong, Ecole Francaise d'Extrdme Orient, Paris. Maloney, B. K. (1989). Palaeoenvironments of Khok Phanom Di: The pollen, pteridophyte spore and microscopic charcoal record. Unpublished section in Volume 1 of the Khok Phanom Di excavation report. Maloney, B. K., Higham, C. F. W., and Bannanurag, R. (1989). Early rice in Southeast Asia, palynological and archaeological evidence from Thailand. Antiquity 63: 363-370. Mansuy, H. (1924). Stations pr6historiques dans les cavernes du massif calcaire de Bac-Son. Bulletin de la Service Geologique d'lndochina 11: 1. Marshall, J. (1951). Taxila, Cambridge University Press, Cambridge. Moore, E. (1986). The Moated Mu'ang of the Mun River Basin, Ph.D. thesis, University of London, London. Mourer, C., and Mourer, R. (1970). The prehistoric industry of Laang Spean, Province Battambang, Cambodia. APAO 5: 128-46. Natapintu, S. (1984). Ancient settlement at Ban Tha Khae in Lopburi. Muang Boran 10(4): 1-4. Natapintu, S. (1988). Current research on ancient copper-base metallurgy in Thailand, In Charoenwongsa, P., and Bronson, B. (eds.), Prehistoric Studies." The Stone and Metal Ages in Thailand, Thai Antiquity Working Group, Bangkok, pp. 107-124. Nguyen Ba Khoach (1980). Phung Nguyen. Asian Perspectives 23(I): 23-54. Noksakul, D. (1981). A Study of the Culture and the Environment of an Ancient Community at Khok Phanom Di, Changwat Chonburi, Master's thesis, Silapakon University, Thailand. O'Harrow, S. (1979). From Col-loa to the Trung sisters' revolt: Viet-Nam as the Chinese found it. Asian Perspectives 22(2): 140-164. Pearson, R., and Underhill, A. (1987). The Chinese Neolithic: Recent trends in research. American Anthropologist 89(4): 807-822. Pigott, V. C. (1984). The Thailand archaeometallurgy project 1984: Survey of base metal resource exploitation in Loci Province, Northeast Thailand. Southeast Asian Studies Newsletter 17: 1-3. Pigott, V. C., and Natapintu, S. (1988). The Thailand archaeometallurgy project: 1984-7. Investigating the origins of metal use in Southeast Asia. Paper read to the conference on ancient Chinese and Southeast Asian bronze age cultures, Kioloa, Australia. Pilditch, J. (1984). Sections on the material culture. In Higham, C. F. W., and Kijngam, A. (eds.), Prehistoric Investigations in Northeast Thailand~ British Archaeological Reports (International Series), Oxford, pp. 57-222. Pisnupong, P. (1984). A preliminary report on the 1982 excavation at Khok Phanom Di, Fine Arts Department, Bangkok, Thailand. Pisnupong, P. (1988). Stone Working at Khok Phanom Di, Master's thesis, University of Otago, Dunedin. Pookajorn, S. (1981). The Hoabinhian of Mainland Southeast Asia: New Data from the Recent Thai Excavations in the Ban Kao Area, Master's thesis, University of Pennsylvania, Philadelphia. Rajpitak, W., and Seeley, N. J. (1979). The bronze bowls from Ban Don Ta Phet: An enigma of prehistoric metallurgy. World Archaeology ll(1): 26-31.

282

Higham

Rajpitak, W., and Seeley, N. J. (1984). The bronze metallurgy. In Higham, C. F. W., and Kijngam, A. (eds.), Prehistoric Investigations in Northeast Thailand, British Archaeological Reports (International Series), Oxford, pp. 102-112. Rutnin, S. (19779). The Pottery from Non Chai, Northeast Thailand, Master's thesis, University of Otag~. Dunedin. Solheim, W. G. II. (1972). An earlier agricultural revolution. Scientific American CCVI(4): 34-41. Sorensen, P. (1973). Prehistoric iron implements from Thailand. Asian Perspectives 16(2): 134-73. Sorensen, P. (1979). The Ongbah cave and its fifth drum. In Smith, R., and Watson, W. (eds.), Early South East Asia, Oxford University Press, Oxford, pp. 78-97. Sorensen, P. (1988). The kettledrums from Ongbah cave, Kanchanaburi Province. In Sorensen, P. (ed.), Archaeological Excavations in Thailand. Surface Finds and Minor Excavations, Scandinavian Institute of Asian Studies Occasional Paper No. 1, Copenhagen, pp. 95-156. Suchitta, P. (1984). Khok Phanom Di: A prehistoric shell mound on the eastern coast of Thailand. In Bayard, D. T. (ed.), Southeast Asian Studies at the X V Pacific Science Congress, Otago University Monographs in Prehistoric Anthropology 16, Dunedin, pp. 1-13. Suchitta, P. (1985). Early iron smelting technology in Thailand and its implications. Paper read at the Research conference on early Southeast Asia, Silapakon University, Bangkok. Sun, S., and Hart, R. (1981). A preliminary study of early Chinese copper and bronze artefacts. Kaogu Xuebao 3:287-302 (Chinese). Takaya, Y. (1969). Topographical analysis of the Southern Basin of the central plain of Thailand. Tonan Ajia Kenkyu 7(3): 293-300. Vickery, M. (1986). Some remarks on early state formation in Cambodia. In Mart, D. G., and Milner, A. C., (eds.), Southeast Asia in the 9th and 14th Centuries, Research School of Pacific Studies, Australian National University, and Institute of Southeast Asian Studies, Singapore, Canberra, pp. 95-115. Vincent, B. A. (1987). The ceramics. In Higham, C. F. W., Maloney, B., Bannanurag, R., and Vincent, B. A. Khok Phanom Di: The results of the 1984-5 season. Bulletin of the IndoPacific Prehistory Association 7: 148-178. Vu Thi Ngoc Thu and Nguyen Duy Ty. (1978). A tool set for casting bronze from Lang Ca (Vinh Phu). Khao Co Hoe 26:36-39 (Vietnamese). Welch, D. (1985). Adaptation to Environmental Unpredictability: Intensive Agriculture and Regional Exchange at Late Prehistoric Centers in the Phimai Region, PhD. thesis, University of Hawaii, Honolulu. Wheatley, P. (1961). The Golden Khersonese: Studies in the Historical Geography of the Malay Peninsula to A.D. 1500, University of Malaya, Kuala Lumpur. Wheatley, P. (1983). Nglgara and Commandery, University of Chicago Department of Geography Research Paper Nos. 207-208, Chicago. White, J. C. (1986). A Revision of the Chronology of Ban Chiang and Its Implications for the Prehistory of Northeast Thailand, Ph.D. thesis, University of Pennsylvania, Philadelphia. Wichakana, M. (1984). The analysis of Upper Songkhram settlement patterns. In Higham, C. F. W., and Kijngam, A. (eds.), Prehistoric Investigations in Northeast Thailand, British Archaeological Reports (International Series), Oxford, pp. 546-564. Wilen, R. (1987). Excavation and site survey in the Huay Sai Khao basin, Northeastern Thailand. Bulletin of the Indo-Paeific Prehistory Association 7:94-117. Wilson, P. (1988). The Domestication of the Human Species, Yale University Press, New Haven, Conn. Wolters, O. W. (1979). Khmer hinduism in the seventh century. In Smith, R., and Watson, W. (eds.), Early South East Asia, Oxford University Press, Oxford, pp. 427-442. Wu, K., Sun, S., Zhang, S., and Wang, D. (1986). Bronze drum making techniques in ancient Southwestern China. Paper read at the second International Symposium on the beginning of the use of metals and alloys, Zhenghou. Yen, D. (1977). Hoabinhian horticulture: The evidence and the questions from Northwest Thailand. In Allen, J., Golson, J., and Jones, R. (eds.), Sunda and Sahul, Prehistoric Studies in Southeast Asia, Academic Press, London pp. 567-599. You-di, C. (1978). Nothing is new. Muang Boran 4(4): 15-16.