Ban Non Wat

Ban Non Wat
Excavation at Ban Non wat
Location in Thailand
15°16′28″N 102°16′08″E
Type	Human settlement
Periods	Neolithic
Location	Non Sung district, Nakhon Ratchasima, Thailand
History
Formed	c. 17th century BCE
Abandoned	c. 5th–6th century CE (disputed)
Site notes
Archaeologists	Charles Higham

Ban Non Wat is an archaeological site in Non Sung district, Nakhon Ratchasima province, Thailand, near Phimai. It has been under excavation since 2002. The site covers about 20 hectares. The cultural sequence encompasses 11 prehistoric phases, which include 640 burials. Occupation runs from a Neolithic settlement placed in the 17th century BC through the Bronze and Iron Ages. The modern village of the same name lies nearer the Mun River. The site's chronology and the reading of its grave goods have both been questioned.

Excavations show that people were occupying the region during the Neolithic, Bronze, and Iron Ages. The sequence rests on 76 radiocarbon determinations treated with Bayesian analyses. Soil conditions in the area may affect that analysis. The determinations place the start of the Bronze Age in the late 11th century BC, and the transition to the Iron Age at about 420 BC.

Bronze, iron and lead were worked on the site. The wealth of the burials rises sharply in the early Bronze Age and falls away, read as the rise and decline of lineages of social aggrandizers. In 2011 the provincial governor moved to have it listed, with four neighbouring sites in the district, as a UNESCO World Heritage property.

Site and chronology

Earthworks and land use

Later in the Iron Age, banks and two moats surrounded the site, reticulating water from the adjacent river. The rice fields around the village are thought to have been irrigated thousands of years ago, on preliminary dating only. An area at the western edge of the site was used to corral domestic water buffalo and cattle.

Dating and phases

The late Neolithic is given as ending about 1280 BC, the Bronze Age as running from then to about 400 BC, and the Iron Age from 400 BC to about AD 300.

On 75 radiocarbon determinations, the merging of the final Bronze Age phase into the early Iron Age is placed in the 5th century BC. The four Iron Age phases of Ban Non Wat and of Noen U-Loke, two kilometres away, have been read as a single sequence of about eight centuries. It runs from the end of the Bronze Age to the formation of local states in the Mun valley.

No clear date separates the Bronze Age from the Iron Age here. The phasing of many of its burials rests on bronze grave goods in the absence of iron, and on the law of superposition. No sterile layer marks a break in occupation, and graves were cut down into earlier deposits, so the distribution of bronze, moulds, and crucibles by layer cannot be taken to reflect the original pattern.

The determinations were run on in-situ charcoal and on the freshwater shells placed as offerings with the dead. Burial 28, a man interred in a large lidded jar, gives the difficulty with unspeciated charcoal in one grave. Charcoal near his cranium and a bivalve shell near his hands are separated by an offset of six centuries. Burning of old wood by initial settlers clearing prime forest cannot be ruled out.

Small but significant amounts of early historic ceramics occur in the upper half-metre of sediment, on which evidence the abandonment of Iron Age sites in the 5th and 6th centuries may be more apparent than real.

Findings

Neolithic

(Left) Neolithic Phase 2 burials, typically containing only a handful of pottery vessels. (Right) A Bronze Age 1 infant burial adorned with shell bangles, signaling a shift toward more complex and elaborate mortuary practices.

The first Neolithic settlement has been placed in the mid-17th century BC. Stone for the adzes does not occur nearby, taken as evidence of exchange, and spindle whorls show that cloth was made on the site. Fish and shellfish from the Mun supplemented the rice and the domestic pigs.

The Neolithic phase shows occupation but no metallurgy. Fragments of bronze and pieces of mould and crucible did reach the Neolithic layers, but undisturbed middens of the period contained none of them. The Neolithic period is divided into an early and late phase. The earliest burials are a series of flexed burials thought to represent hunter-gatherers. These were partially contemporary with the initial Neolithic settlement by rice farmers who also raised pigs, hunted a wide range of animals, fished, and collected shellfish. Late Neolithic burials were often accompanied by simply decorated ceramic vessels and few other grave goods. Stone adzes, shells, and animal remains have also been found in the Neolithic burials. No red ochre was recovered from any of them.

The people buried here have been assigned, on the form and decoration of their pottery, to a different migratory population from that at Khok Phanom Di. That coastal cemetery lies 250 km away, and ochre was used there heavily and in powdered form.

Fourteen flexed graves were found, with shell beads of a form not found in any farming context. In the earlier of the two Neolithic periods the graves lie in the western part of the excavated area, with no nucleation into groups. Shell beads are almost absent in the earlier period and still rare in the later. Two adults were interred in large lidded jars, a form rarely found in the Southeast Asian Neolithic. The later period's pottery differs from the earlier, read as a possible break in occupation followed by the arrival of a different population.

A 2022 reassessment of chicken domestication found the oldest bones of a known domestic chicken here, dated to between 1650 and 1250 BC.

Bronze Age

Adults from the Bronze Age 3A period at Ban Non Wat typically wore multiple exotic shell and marble bangles

Dog burial from Ban Non Wat

Wealthy early Bronze Age burials appear with copper-base metallurgy, and grave goods increase sharply against the Neolithic, personal ornamentation becoming more common. The Bronze Age has been divided into five phases, the earliest beginning about 1050 BC and the latest ending about 420 BC. Bodies were commonly wrapped and placed in wooden coffins, with more variation in burial contexts in the late Bronze and early Iron Ages. Supine-extended burial is reported to be uncommon here, the usual form being a dug pit.

The seven burials of Phase 1 held shell bracelets and necklaces and a copper-base axe. One of them also held red ochre in powdered form, the earliest ochre at the site. Its source has not been established. Phase 1 is the first with the dead in wooden coffins in deeply cut graves, and the first in which they were grouped. One young woman of the phase lay in one shaped like a boat, with bivalve shells and a copper-base axe. Five of the seven graves held a socketed bronze axe or another implement.

Phase 2 graves are richer, with copper-base axes, anklets and rings and shell-bead necklaces, belts and earrings, and hold a wider variety of goods than any other mortuary phase at the site. Ochre is common in Phase 2, in most of the women's graves and about a third of the men's. In Phase 2 some of the dead were partially disinterred and the bones replaced in the grave, in one case with the skull set on a stack of limb bones. Infants of that phase were laid out beside adults in graves far larger than the body and with large numbers of pottery vessels, which has been taken to show that status could be inherited. Across the first three mortuary phases the rituals included placing fish, severed pig limbs, shellfish and chickens with the corpse. The partial exhumation and reinterment of the richest dead has been read as a possible indication of ancestor worship.

Phase 3 repeats the wealthy burials of the phases before it and is the last to do so. Its graves are oriented north-east or south-west, where other Bronze Age burials at the site mostly lie north or south. Phase 3 has been divided into two sub-phases, the earlier as wealthy as Phase 2 and the later with a sharp fall in mortuary wealth and no bronzes.

Phase 4 contains 162 burials in distinct groupings, and no evidence shows what the groupings represent. The phase divides into five clusters, and ochre is unevenly spread between them. That unevenness has more recently been read as evidence that the clusters stand for kin or social groups. The closeness of the graves and the rarity of disturbance between them have been taken to show that they were marked above ground.

Phase 4 marks a sharp decrease in grave goods against Phases 1 to 3, and Phase 5 continues the reduction. Ochre returns in Phase 5 in about half the graves, now in pellet rather than powdered form and often paired with bivalve shells. No juvenile of the phase was buried with it. The shift from powder to pellet has been read as a change in how the pigment was seen, and the solid form as a possible commodification, the pellets perhaps carrying a countable value. Through the Bronze Age the pairing of ochre with bivalve shells, and with burial jars whose interiors are painted blood-red, has been taken to show a growing emphasis on regeneration. Snake motifs and depictions of childbirth have been read as supporting that reading.

Clay was regularly placed in late Bronze Age and early Iron Age graves, read as an interest in pottery making or in dyeing fabric. The later rise in ochre, sometimes beside spindle whorls and grey clay, has likewise been linked to dyeing and to an artisan identity.

A dog was buried with two bowls, read as containers for food and water.

Iron Age burials

An infant burial from the Iron Age 1 period at Ban Non Wat features remarkable bronze anklets

With the Iron Age, a new range of exotic ornaments accompanied the dead, including carnelian, agate, and glass. Such jewellery is present in the earliest Iron Age cemetery from at least the 4th century BC, read as evidence of long indirect contact with India. Knowledge of iron smelting and forging has been attributed to the same contact, reaching the upper Mun by way of the peninsula. The site's glass shares its dominant composition with that of Non Ban Jak and Noen U-Loke, read with them as evidence of a South Indian or Sri Lankan network, perhaps emerging as late as about AD 200. Gold ornaments, among them rings, beads, and earrings, have also been reported from burials here and at Noen U-Loke and Non Ban Jak.

In contrast with the Bronze Age, the Iron Age provided significantly fewer burial goods. The number of bronzes in the graves rises markedly with the first Iron Age phase, dominated by bangles, anklets, and rings. The earliest Iron Age graves are described as carrying kits of small iron knives, chisels, and awls held in cloth bags, heavy socketed hoes, and iron bangles, with iron spearheads welded onto a socketed bronze haft in three cases. Bronze castings in the same phase include hollow anklets filled with bronze balls worn by a child and lost-wax armbands worn by an infant. In the last Iron Age phase infants were interred in lidded vessels of distinctive form with glass beads and agate ornaments, and cast bronze finger and toe rings were abundant.

Burials during the Iron Age at Ban Non Wat are divided into two phases. Phase 1 Iron Age burials are very similar to those of Phase 5 of the Bronze Age, the presence of iron often the only distinguishing factor. Phase 2 Iron Age burials are often marked by Phimai Black pottery besides iron work and exotic goods. The amount of grave goods remains far below that of some Bronze Age burials.

The earliest Iron Age cemetery holds 133 intact graves, among the largest samples of the period in Southeast Asia, and is contiguous with Bronze Age Phase 5 as the cemetery expanded eastward; the two have been read as one cemetery. Principal component analysis found a single compact group with a few outliers and no nucleus of wealthy individuals marking social distinctions.

Within the first Iron Age phase, the graves fall into two spatial groups, one with the head laid to the north and one to the south. Ochre occurs in both groups, around the head and feet in the northern one and at the right arm or hand in the southern. The difference has been read as two mortuary traditions in use at once. Ochre is absent from the second phase, which is dominated by infant jar burials. The new symbolism and the use of grave markers may have reduced the standing of ochre, its survival in pellet form being read as a gesture to tradition alongside newer techniques. Three further Iron Age phases are recognised in the settlement, with very few graves in them. Across the Bronze and Iron Ages ochre never reaches half the graves of any phase. It usually lies outside the body, on the extremities or in accompanying vessels, read as placing during or after burial.

Iron Age crafts, health and subsistence

Phimai Black from Ban Non Wat, Noen U-Loke, and Ban Suai was made in each community, with the same forms, decoration, and rice-husk temper. The technique itself has been traced to the middle Ganges and the comparable Northern Black Polished Ware.

Over a thousand spindle whorls were recovered, from Bronze Age contexts through to the late Iron Age. Thirty-five were found at Non Ban Jak, the difference read as more intense and more specialised manufacture of cloth here, unless it reflects the greater area excavated.

A study of the dentition found the highest Iron Age rates of carious lesions and of antemortem tooth loss among the three excavated Iron Age sites of the valley. The Neolithic sample carries the regional maximum for carious lesions and for advanced wear. The Iron Age dental sample is small. It was drawn from 244 of the 630 individuals excavated, most from the earliest Iron Age, chosen for complete and well preserved dentition, and treated as a single block of about a thousand years. A rate of natural population increase of 1.26 percent has been calculated for the Iron Age here, below replacement level and read as a population in decline.

No evidence of congenitally missing teeth has been reported from the prehistoric population here, where the trait is common at Noen U-Loke and Non Ban Jak. No graves filled with rice, of the kind found at Noen U-Loke, and no iron sickles have been found here, although the sequence stops short of the last Iron Age phase of the other two sites. The rice of the Bronze and early Iron Age is japonica, while the late Iron Age carries a markedly higher proportion of indica, read as the introduction of new wet rice varieties. The change from dry to wet cultivation has been contrasted with Non Ban Jak, which used wet rice from its foundation.

Metalworking

Lead isotopes in the slag and cast metal have been matched to ores of the Khao Wong Prachan valley in central Thailand. The copper was therefore drawn from within mainland Southeast Asia by about 1000 BC, not imported. The socketed axe from the Phase 1 grave of a two-year-old is 98.1 percent copper. None of the Phase 1 copper analysed matches the lead isotope signature of any known Southeast Asian copper mine. Metal prills on the crucibles all carry copper with substantial tin and so cannot have cast near-pure copper axes. The earliest axes have been read as imports from a distant source, possibly in southern China, rather than the work of resident founders.

Lead isotope analysis of 79 artefacts from Non Ban Jak places all the Ban Non Wat samples analysed with them inside its range. Three early Iron Age samples, a lead fragment and two leaded bronze 'cordiforms' from two burials, are consistent with a silver ring from Non Ban Jak. The ring's silver has been read as cupellated from an argentiferous lead source worked for lead as much as eight centuries earlier. The cordiform shape is known in the Khao Wong Prachan valley as a form of copper ingot, but these two are leaded alloys and not ingots of that kind. The same study holds it inadvisable to assume that unanalysed rusty objects are iron rather than unrefined copper or bronze, and gives a Bronze Age phase 4 artefact here as a possible case. Two cordiform bronzes from one early Iron Age grave match moulds and castings from Nil Kham Haeng in the same valley. Some of the copper at the site is compatible with lead isotope signatures from Vilabouly Complex in upland Laos.

Graves cut through clay floors, one preserving a charred timber beam from a probable wall foundation, and the occupation layers yielded evidence of weaving, bronze casting, iron forging, pottery making, and the casting of lead. Several occupation surfaces were littered with the broken bones of cattle, water buffalo, and deer together with iron knives and large stones read as butchering anvils. About a dozen working floors of the Iron Age have been uncovered. Also found were crucibles and fragments of bronze and iron. Bronze axes and chisels from the site are reported to have been cast in sandstone and fired-clay moulds. A Bronze Age Phase 4 grave held 29 clay moulds for casting socketed axes and bangles, its occupant otherwise poorly endowed. The bangle moulds are of a concertina form and were used in sets, one set at the occupant's shoulder able to cast twenty bangles in a single pour.

Casting evidence peaks in Phases 4 and 5. The number of crucibles rises through the Bronze Age to 117 in the latest layer, then falls away sharply in the first Iron Age context. The crucibles were tempered with rice chaff and then lagged with a silica coating, which spared the vessel wall and let one serve more than a single melt. The same two treatments appear at Phu Lon, Ban Chiang, and Ban Na Di. Of the 359 casting moulds, those for ornaments are rectangular in section and those for axes, chisels, and spears semicircular, each implement cast singly. Crucibles of this form, small spouted bowls found at many sites in northeastern Thailand, have been grouped as a common Southeast Asian type.

About 600 fragments of bronze come from occupation contexts, most without surviving form and many lying near the furnaces as casting debris. Hardly any tools or weapons occur in those contexts despite the number of axe moulds, which has been attributed to the recycling of large broken pieces. The number of stone adzes fell sharply once bronze axes were available.

In the early Iron Age the numbers of crucibles, moulds, and bronzes fall, while the clay-lined furnaces remain in use. Caches of clay socket plugs for bivalve axe moulds were made. Some are still sheathed in lead, which would have melted out and made the alloy more fluid. No bronze axe was found in any early Iron Age burial or occupation context.

Interpretation

Grave wealth and social change

The wealth of the burials has been read as a cycle, not a rise. It peaked between about 1000 and 800 BC and then fell away, and the site did not recover its earlier standing until about 200 BC, in the early Iron Age. The explanation offered for the accumulation, an emerging social hierarchy, does not account for the loss. A further suggestion, that the reversal followed changes in trade and climate beyond the region, is offered as speculation, since no temperature series or tree-ring study covers the period. A later account reads the peak as the rise of one or more lineages of social aggrandizers, who dominated the community for about four generations before the burial pattern gave way to rows of graves with little display of wealth. It finds no wealthy subset in the later Bronze Age phases or the earliest Iron Age despite repeated statistical analysis, and reads the first bronze as having produced aggrandizers where iron did not. The Phase 2 graves have been named as a possible exception to a general lack of association between bronze and hierarchical status elsewhere in the region. The same evidence has been read as supporting a regional pattern in which the social use and consumption of bronze varied site by site and shifted over time and space.

The wealth of Phase 2 is confined to the central part of the site. A second exposure of graves of the phase, about 40 metres to the north-east, held no metal and few exotic ornaments. The nucleated graves of the same date at Ban Lum Khao were virtually without bronze. At Ban Prasat, on the same exchange route, the centrally placed dead of the phase were as rich as those at the centre here.

Social elites are not attested at Ban Chiang, Non Nok Tha, or Ban Na Di. The 47 burials of Bronze Age 2 and 3A here held over 150,000 shell beads, with hundreds of shell and marble bangles. Ban Chiang, in more graves, yielded almost none.

The site's position has been read as the mechanism behind that wealth. The dead of the first three mortuary phases were interred over at least eight generations, with wealth described as unparalleled at any other Southeast Asian site. The site stands at the eastern entrance to a pass over the Phetchabun Range, the logical point of entry for prestige goods from the Chao Phraya basin. The Khao Wong Prachan copper mines lie 150 km to the west and the range itself yields marble. Eastward the Mun meets the Mekong, beyond which are the copper mines of Sepon, and the upper Mun has salt of its own. Such a natural exchange bottleneck is held to give aggrandizers preferential access to prestige goods.

Disputed dating and attribution

The site's dates bear on the origin of Southeast Asian bronze. One model derives the region's bronze-working tradition from metalworkers associated with the Seima-Turbino culture, who are said to have reached the region about 2000 BC. Against this, the burials here are dated to between 1259 and 1056 cal BC, while the regional sequence dates to between about 1600 and 540 BC. The older determinations have been questioned because they come from disturbed deposits, while Seima-Turbino material is scarce in the region. Both positions are presented as models that must remain in flux.

The site's dates have been questioned, on the ground that precautions were not taken in collecting them and that the results are therefore unreliable. The Bayesian analysis built on those dates has been questioned on the same ground, and so have the significance and meaning of the artefacts and burial goods. A further objection concerns the reach of the chronology rather than its reliability. It argues primarily from a programme of shell dating at this site alone and was not cross-dated against the sites with the earliest metal. On that reading it therefore applies to the florescence of bronze in the immediate area of the upper Mun valley and not to the first appearance of the technology elsewhere in Thailand. The chronology's authors described it as the "rejection of all previous attempts to date Southeast Asian prehistory radiometrically". The importance of the bronze grave goods is also disputed, on the ground that the introduction of metal carried a different social meaning here than in other cultures. The soil in the area is more acidic, which can interfere with both excavation and analysis.

Occupation was continuous from the Bronze Age onwards, while in the Neolithic the site may have been less densely populated and used seasonally.

Pottery comparisons here and at Non Ban Jak across the late Iron Age and the proto-historic period have been read as showing that the resident populations were not replaced by incoming Dvaravati people. The pottery changed gradually, the technology being adopted and adapted locally and not carried in by new arrivals.

Fieldwork and heritage listing

The excavations are partially funded by the Earthwatch institute. Nakhon Ratchasima Rajabhat University has worked at the site with the University of Otago, and the local Tambon Administrative Organization funded a learning centre there. In 2011 the provincial governor moved to have the archaeological sites of the upper Mun basin in Non Sung district listed as a UNESCO World Heritage property. The property was called The Upper Mun River Basin Archaeological Sites, Non Sung, Nakhon Ratchasima (แหล่งโบราณคดีลุ่มแม่น้ำมูลตอนต้น โนนสูง นครราชสีมา). The area was given as five locations in the district, Ban Non Wat, Ban Lum Khao, Ban Prasat, Non Muang Kao, and Noen U-Loke.