35,000-year-old sites in the rainforests of West New Britain, Papua New Guinea.

Pavlides, Christina ; Gosden, Chris

The growing story of early settlement in the northwest Pacific islands is moving from coastal sites into the rainforest. Evidence of Pleistocene cultural layers have been discovered in open-site excavations at Yombon, an area containing shifting hamlets, in West New Britain's interior tropical rainforest. These sites, the oldest in New Britain, may presently stand as the oldest open sites discovered in rainforest anywhere in the world.

Introduction

This note presents new Pleistocene dates from the island of New Britain, Papua New Guinea. Recent research at Yombon has revealed cultural deposits dating from 35,000 b.p. Previous archaeological work in West New Britain succeeded in revealing only terminal Pleistocene occupation (Specht et al. 1983). The new dates from Yombon extend the human occupation of West New Britain by 23,000 years, making them equivalent in age to the earliest sites in neighbouring New Ireland (Allen et al. 1988). While the 40,000 year old dates from the Huon Terraces in Papua New Guinea remain the earliest evidence of human occupation in coastal forest in this region (Groube et al. 1986; Groube 1989) the sites at Yombon are important because they indicate the use of inland forest environments and resources, including sources of high-quality chert, from the Pleistocene to the present. These data also run counter to claims that rainforest environments were only occupied for the first time in the late Holocene.

Prior to the discoveries at Yombon, the earliest Pleistocene sites in the Bismarck Archipelago were coastal caves (Allen et al. 1988; Wickler and Spriggs 1988). These sites have revealed the early spread of humans into island Melanesia and the marine adaptations necessary. Similarly, evidence of early colonisation of the high altitude mid- and upper-montane forest of interior Papua New Guinea indicates the penetration and manipulation of these environments by at least 30,000 b.p. (White et al. 1970; Gillieson & Mountain 1983). No evidence exists, however, for the utilization of the lowland forests. The sites at Yombon provide new evidence that lowland forested areas were used during the late Pleistocene and early Holocene.

The fact that we have found only a small number of sites and artefacts dated to the Pleistocene is unremarkable. In these tropical regions, high rainfall and such activities as forests clearance and gardening can result in high rates of erosion and extensive surface alterations which greatly diminish the possibility of in situ archaeological deposits of such antiquity. The discovery of Pleistocene cultural deposits in trenches over half a kilometre apart, all yielding similar radiocarbon determinations and stratigraphic sequences, makes us confident that they are in situ Pleistocene assemblages. This confidence is increased by the overall structure of the sites which give a consistent stratigraphic picture over a wide area (Pavlides 1993). The local topography of this part of West New Britain with its limestone ridges and valleys, coupled with the deposition of airfall tephras, are undoubtedly the key factors in preserving small patches of Pleistocene activities. In every trench containing remnants of the old tephra there were underlying Pleistocene deposits, possibly demonstrating taphonomic processes at work rather than low levels of human activities, which originally may have been much more widespread in this area. Specht (et al. 1981) positioned his excavations at Yombon on the plateaux rather than the valley bottoms, where these new sites occur, and failed to locate any deposits older than 4000 b.p. These data must also alert us to the fact that rainforest occupation in New Britain may have been widespread during the late Pleistocene.

The sites at Yombon are composed of a series of chemically distinct tephras which can be traced over large areas of the forest. These tephra layers provide the opportunity to look not just at individual sites but at whole landscapes, e.g. White's work at Kosipe (White et al. 1970) and Torrence's at Garua (Torrence 1992). The sequence of stone artefacts as seen in successive layers indicates the production of different forms and numbers of stone tools at various periods. Yombon's role as a source of high-grade chert is evident in layers of freshly quarried stone and primary knapping debris within the Holocene and Pleistocene layers (Pavlides 1993; Pavlides & Gosden in press). Specifically, the sites at Yombon reveal that by 35,000 b.p. people had discovered and were utilizing local chert sources, continuing this exploitation periodically until the close of prehistory.

The excavations and tephrochronology

Yombon is situated approximately 35 km northeast of Kandrian on West New Britain's south coast, approximately 6 [degrees] south of the equator. The site is deep in the lowland tropical rainforest of West New Britain, at an altitude of approximately 500 m above sea level. Rainfall in the area is greater than 6350 mm per year (Chowning 1980). Today the area of Yombon is inhabited by people who live in dispersed hamlets subsisting on a combination of wild and cultivated foods. Archaeological research in the region was first carried out between 1979 and 1981 at both Yombon and near-by Misisil Cave, the latter producing a date of 11,400 b.p. (Specht et al. 1983).

Pavlides has excavated 10 test trenches around Yombon with five more trenches dug in two neighbouring villages, Asiu and Sisisel. These excavations revealed stratified sites composed of alternate layers of in situ tephra and artefact-rich soils. The tephras seal cultural layers within these sites and allow the comparison of equivalent chronological units over wide areas of West New Britain. The artefactual sequences from Yombon are composed of both formal and amorphous stone tool assemblages knapped primarily from local chert and obsidian imported from the Talasea and Mopir areas (Pavlides 1993).

The combined results of the archaeological and geological investigations indicate the deposition of at least four tephras in the Holocene and many others during the Pleistocene in this area of West New Britain. Geological research in northern New Britain has identified Mount Witori, approximately 95 km away, as the source of these ancient tephras (Machida pers. comm.).

The depth of unweathered tephras varies within and between sites excavated in the area, as does the depth of the overlying soil layers. Some indication of the original thickness of tephras deposited on the area can be gauged by combining the thickness of each individual tephra and the overlying soil layer, which is the weathering product of the tephra. In the case of the WK2 eruption, excavated tephra and soil layers indicate a depth of between 20 and 50 cm. As well as covering huge areas of central West New Britain, the deposition of such quantities of tephra would have caused considerable problems for the inhabitants of the forest, including severe damage to forest trees, food plants and animals (see Blong 1982: 162). For the archaeologist these large volcanic events preserve a series of landscapes from the Pleistocene to the present. The unique chemical composition of each tephra allows chronological correlations to be made across the whole of northern and central West New Britain.

Sites containing remnants of the four major Holocene tephras, together with the Pleistocene evidence, have been revealed in trenches FIF/2, 3 and 4 on the Yombon airstrip. In all 10 stratigraphic units have been identified in this 3 x 1 m trench. The sites FYV/1 and 2, excavated at an area in Yombon called Eliva, are approximately 500 m east of the airstrip sites. These squares, approximately 12 m apart, contain most of the sequence identified in the airstrip sites with the exception of the latest tephras and soils which appear to have been totally weathered and incorporated into the topsoil layer. FIGURE 2 illustrates the tephra correlations and associated soil layers between the two areas.

Radiocarbon dates

Fourteen radiocarbon determinations are now available from the sites at Yombon and Asiu. These dates coupled with the known dates of Holocene volcanic eruptions provide an initial chronological sequence for this area of West New Britain. The integrity of the sites is indicated by the consistency of the radiocarbon dates for equivalent layers between squares for both the Holocene and Pleistocene layers.

The late Pleistocene soil layers are sealed beneath 30 to 50 cm of sterile clay and an old tephra of indeterminate age (labelled 'consolidated tephra' in FIGURE 2) in both the FIF and FYV sites. This earliest tephra was extremely compacted and intact across most of the excavated squares. On the basis of the tephra's stratigraphic position and available dates from the sites it is likely to have been deposited some time after 14,000 b.p.

At the airstrip sites FIF/2 and 3 the dates of 32,630[+ or -]400 and 33,600[+ or -]670 come from scattered lumps of charcoal surrounding the artefacts in a carbon-rich clay layer which lies directly below the old tephra at a depth of 160 and 170 cm below ground surface. These radiocarbon determinations date the upper and lower part of this stratigraphic unit across two squares. The two trenches excavated at Eliva, FYV/1 and 2, also revealed a carbon-rich clay layer directly below the old tephra. The date of 35,570[+ or -]480 from FYV/2 and 14,310[+ or -]100 from FYV/1 date the upper section of this layer at a depth of 155 and 180 cm below the surface. Chert artefacts found between 10 and 20 cm below the level of these dates in both squares indicate perhaps deposition of artefacts before 35,570 in FYV/2 and 14,310 in FYV/1.

TABULAR DATA OMITTED

TABULAR DATA OMITTED

Stone tools

The artefacts found within the Pleistocene layers consist of flakes and angular fragments struck from locally obtainable chart and fire-cracked chart pieces. One of the artefacts recovered from the Pleistocene context in FYV/2 is a large retouched flake weighing 142 g and measuring 106 mm at its longest point.

Chert is the only lithic resource deposited within the Pleistocene layers at these sites. Later there are chert flakes and retouched tools, and small amounts of obsidian, imported from the Talasea and Mopir sources on New Britain's north coast. The production and deposition of bifacially and unifacially flaked tools occurs between 4000 and 3500 b.p. These artefacts are made from local chert in a variety of stemmed, waisted and oval forms. At approximately 3300 b.p. the interior areas of West New Britain were smothered by tephra from the greatest Holocene eruption of Mount Witori, WK2. After this eruption, a major change occurs in the nature of lithic reduction and the amount of archaeological material at the sites decreases (Pavlides 1993).

Yombon in the rainforest debate

In debate regarding modern humans' colonizing capabilities and the environments that are hard to colonize, rainforests are seen as a particularly difficult environment for pre-agricultural humans (Bailey et al. 1989; Headland 1987; Headland & Reid 1989; Hart & Hart 1986; Eggert 1992; Bellwood 1990). Although rain-forests have extremely high biomass, the plants and animals necessary to sustain human life are highly diverse and dispersed, making this environment difficult for hunter-gatherers.

The Yombon region is today covered by tropical rainforest. Although there is no pollen data from this area or from other lowland forest in Papua New Guinea to show evidence of the Pleistocene vegetation, there is every reason to believe that the area was covered by rainforest during the late Pleistocene. Its equatorial position, low altitude and high rainfall all favour the establishment of a rainforest zone. In the Papua New Guinea Highlands, the treeline was above 2000 m throughout the last glacial maximum; there is little evidence for a drop in either temperature or rainfall (Swadling & Hope 1992; Enright & Gosden 1992) sufficient to disrupt totally the lowland forest habitats. Furthermore, palynological evidence from South East Asia indicates little forest fragmentation during times of higher glacial aridity (Bellwood 1990; Walker 1982; Whitmore 1984).

Equatorial rainforests are among the most stable, as well as most complex environments in the world. Having been least affected by Quaternary glacial cycles (Flenley 1979), rainforests can provide insights into the changing human use of a constant set of resources, impossible from any other ecosystem. At Yambon and other sites discovered in Melanesia, we see people existing in both coastal and inland environments without the benefits of agriculture by at least 35,000 b.p. The discovery of Pleistocene sites in the rain-forest regions of West New Britain are therefore not totally unexpected given the geographical spread of modern humans into almost every other environmental niche at this time.

Acknowledgements. This paper concerns discoveries made as part of Pavlides' ongoing doctoral research. We are grateful to the Australian Research Council Small Grants Scheme who provided the funding for the research carried out in 1993. The National Museum of PNG offered their support and encouragement as did the New Tribes Mission, Hoskins. Jim Allen and Brendan Marshall offered helpful comments on various drafts of this paper. Margaret Wade and Wei Ming drew the figures. Hiroshi Machida of the Tokyo Metropolitan University did the initial tephra sourcing of the Yambon material, and Peter Jackson has completed subsequent analyses; we thank them both for their continued support and efforts. Finally we would like to thank the people of Yambon, Asiu and Sisisel villages and our team of excavators, Eve Flaim, Lincoln Hayes and Robert Mondol.

References

ALLEN, J., C. GOSDEN & J.P. WHITE. 1989. Human Pleistocene adaptations in the tropical island Pacific: recent evidence from New Ireland, a greater Australian outlier, Antiquity 63: 548-61.

BAILEY, R.C., G. HEAD, M. JENIKE, B. OWEN, R. RECHTMAN & E. ZECHENTER. 1989. Hunting and gathering in the tropical rain forest: is it possible?, American Anthropologist 91: 59-82.

BELLWOOD, P. 1990. From late Pleistocene to early Holocene in Sundaland, in C. Gamble & O. Soffer (ed.), The world at 18,000 BP 2: 255-63. London: Unwin Hyman.

BLONG, R.J. 1982. The Time of Darkness: local legends and volcanic reality in Papua New Guinea. Canberra: Australian National University Press.

CHOWNING, A. 1980. Culture and biology among the Sengseng of New Britain, Journal of the Polynesian Society 89: 7-31.

DODSON, J. (ed.). 1992. The Naive Lands: prehistory and environmental change in Australia and the Southwest Pacific. Melbourne: Longman Cheshire.

EGGERT, M. 1992. The central African rain forest: historical speculation and the archaeological facts, World Archaeology 24(1): 1-24.

ENRIGHT, N.J. & C. GOSDEN. 1992. Unstable archipelagos -- southwest Pacific environment and prehistory since 30,000 BP, in Dodson (1992): 160-98.

FLENLEY, J.R. 1979. The equatorial rain forest: a geological history. London: Butterworth.

GILLIESON, D. & M.-J. MOUNTAIN. 1983. Environmental history of Nombe rockshelter, Papua New Guinea Highlands, Archaeology in Oceania 18: 45-53.

GROUBE, L. 1989. The taming of the rain forests: a model for late Pleistocene forest exploitation in New Guinea, in D. Harris & G.C. Hillman (ed.), Foraging and farming: the evolution of plant exploitation: 292-304. London: Unwin Hyman.

GROUBE, L., J. CHAPPELL, J. MUKE & D. PRICE. 1986. A 40,000 year-old human occupation site at Huon Peninsula, Papua New Guinea, Nature 324: 453-5.

HART, T.B. & J.A. HART. 1986. The ecological basis of hunter-gatherer subsistence in African rain forest: the Mbuti of Eastern Zaire, Human Ecology 14(1): 29-55.

HEADLAND, T.N. 1987. The wild yam question: How well could independent hunter-gatherers live in a tropical rain forest ecosystem?, Human Ecology 15(4): 463-91.

HEADLAND, T.N. & L.A. REID. 1989. Hunter-gatherers and their neighbours from prehistory to the present, Current Anthropology 30(1): 43-66.

PAVLIDES, C. 1993. New archaeologicel research at Yombon, West New Britain, Papua New Guinea, Archaeology in Oceania 28: 55-9.

PAVLIDES, C. & C. GOSDEN. In press. The mysteries of West New Britain, Paradise.

SPECHT, J., I. LILLEY & J. NORMU. 1981. Radiocarbon dates from West New Britain, Papua New Guinea, Australian Archaeology 12:13-15.

SPECHT, J., I. LILLEY & J. NORMU. 1983. More on radiocarbon dates from West New Britain, Papua New Guinea, Australian Archaeology 16: 92-5.

SWADLING, P & G. HOPE. 1992. Environmental change in New Guinea since human settlement, in Dodson (1992): 13-40.

TORRENCE, R. 1992. What is Lapita about obsidian? a view from the Talasea sources, in J.C. Galipaud (ed.), Poterie Lapita et peuplement: actes du colloque Lapita, Noumea, Nouvelle Caledonie Janvier 1992: 111-26. Noumea: Orstom.

WALKER, J. 1982. Speculations on the origins and evolution of the Sunda-Sahul rain forests, in G.T. Prance (ed.), Biological diversification in the tropics: proceedings of the fifth international symposium of the Association for Tropical Biology: 554-75. New York (NY): Columbia University Press.

WHITMORE, T.C. 1984. Tropical rain forests of the Far East. 2nd edition. Oxford: Clarendon Press.

WICKLER, S. & M. SPRIGGS. 1988. Pleistocene occupation of the Solomon Islands, Melanesia, Antiquity 62: 703-6.

WHITE, J.P., K.A.W. CROOK & B.P. RUXTON. 1970. Kosipe: a late Pleistocene site in the Papuan Highlands, Proceedings of the Prehistoric Society 36: 152-70.