The Archaeology of the Trobriand Islands, Milne Bay Province, Papua New Guinea


Trobriand Islands Milne Bay Province, Papua New Guinea.

In November 1998, Gotland University College, Visby, Sweden, started a new archaeological research project on the Trobriand Islands, Papua New Guinea, and during August-November 1999 sixteen Swedish archaeologists and osteologists carried out excavations and osteological analyses at new-found sites on the northern part of Kiriwina Island. The aim of the project is to study the introduction and subsequent cultural development of the Trobriand culture. Central questions at issue include the time of initial colonization of the area; the existence or non-existence of long-term cultural continuity in the islands as revealed by the archaeological record; the identification of possible hiatuses in the cultural development which may be associated with e.g. the influx of intrusive populations (as revealed by ongoing genetic studies of the skeletal material). In the initial stage, radiocarbon dates (AMS) and post mortem DNA-analyses on skeletal remains will form a crucial fundament for the planning of the forthcoming investigations. See full excavation report and the preliminary results of the investigations in Burenhult, G. (ed.) 2002. The Archaeology of the Trobriand Islands, Milne Bay Province, Papua New Guinea. Excavation Report 1999. B.A.R. International Series 1080.


The Trobriand Islands form part of the Melanesian archipelago, and politically they belong to Papua New Guinea. The unique and colourful culture of the Trobrianders, including their famous system of ceremonial exchange, the Kula ring, was studied extensively by Bronislaw Malinowski in the early 20th century. However, nothing is known about the earliest occupation of the islands. Nor do we know why the Trobriand Islanders differ markedly from other Melanesian populations with respect to their social organization, cultural set-up, physical characteristics, and genetics, instead bearing some resemblance to Polynesian populations further east. The Trobriand Islanders have been subject to intense anthropological studies, but apart from a partial excavation of one of the megalithic tombs on the island of Kiriwina in the 1940's, no stratigraphic archaeological excavations have been carried out on the islands. However, extensive surface collections of potsherds have been made and analysed in the 1970's , both on Kiriwina, Vakuta and Kuyawa, and some burial caves have been mapped. Also, surface collections of potsherds have been made on the nearby Amphletts and Goodenough Islands. Recently, archaeological investigations on Woodlark have produced important data on the early settlement on the island. We do not know if the megalithic tradition in the Trobriands (represented by e.g. the Otuyam tomb on central Kiriwina), was carried by ancestors of today's population or by an earlier and culturally different group of people. Present-day Trobrianders have no cultural relation to the megalithic structures on their islands, an unusual situation for Melanesia, where the megalithic traditions in most places are still very much part of the present cultures or, at least, well remembered and ethnographically documented. 



Trepanated human skull from Budou burial cave, Labai, during the investigations in 1999. The cave burials have been dated to c. AD 1570.


Much controversy still surrounds the arrival of the Austronesian language family, which presently dominates most parts of the Pacific. The traditional view holds that Austronesian-speaking peoples entered the archipelago from Southeast Asia around 3000 BC, but there is no evidence for such a large-scale migration into Melanesia in the archaeological record. The evidence for this scenario is mainly linguistic, but not even the linguists agree. It is possible, instead, that people have been moving in successively during a very long period of time, and many archaeologists believe that local evolution accounts for much of the cultural development in the area. The Lapita culture, for example, has traditionally been considered to be associated with migrating Austronesians from Southeast Asia, but, following intensified archaeological work in the area, it has also been suggested that this in fact developed in the Bismarcks, as a result of intense exchange networks, from where it spread into the Pacific. Several archaeologists today claim that the Polynesian culture, an extension of the Lapita, originated in eastern Melanesia, not East/Southeast Asia.



An area on northern Kiriwina around the villages of Mwatawa and Labai was selected for investigation. Numerous stray finds of polished stone adzes and axes of various types and sizes in the surrounding gardens, combined with phosphate surveys, promised to facilitate the localization and identification of earlier settlements in this area. This area also contains a number of caves with burial remains. The area around the villages of Mwatawa and Labai may also be of some interest with regard to the local oral tradition, which holds that the first Trobriand woman was born out of one of the nearby caves and that Labai is considered to be the oldest village in the Trobriand Islands. As a part of the 1999 investigations, also the oral traditions of Labai and Mwatawa, with respect to prehistoric sites and burial caves, were recorded in detail.

Two of the burial caves were selected for a detailed study, Selai Cave and Budou Cave. The interior of the caves themselves has been provisionally mapped, and burial depositions registered. The archaeological investigation includes a documentation of the position of all bones within the depositions with the aim of making possible a reconstruction of e.g. deposition and reburial traditions. Documentation methodology includes drawing and photography, using both conventional and digital cameras, and the methodology used in the caves will be equal to conventional stratigraphic excavation during open air conditions. Bone samples were also collected from a third cave in Labai, Obuwaga Cave. As the bone depositions in this cave show clear signs of recent disturbances, and most bones obviously are displaced, no detailed analyses were carried out. A rock-shelter burial overlooking the sea was found and investigated at Labai Beach, Bwara Tudava.

The osteological analyses comprise the determination of death age, sex and length; the determination of the minimum number of individuals (MNI); and a study of possible diseases or other pathological changes including dental paleopathology, injuries, congenital abnormalities, discrete traits, stress or activity markers such as activity-induced pathology, as well as possible cut-marks or defleshing marks. Also, paleodemographic issues will be addressed. Most analyses have been conducted outside the entrances to the caves, and the bones have been put back in original position after the investigation. Samples for radiocarbon dating (AMS), stable isotope analyses (diet reconstruction), trace element analyses, and post mortem DNA analyses were collected, commonly a small part of a tooth is used for these purposes.

A total of 23 prehistoric sites have so far been located in the Labai and Mwatawa areas by means of stray finds of e.g. axes/adzes and pottery, and phosphate surveys were used in order to determine the extension of the defined activity areas. Test excavations were carried out in order to determine the depth and character of the stratification. All located sites have been mapped using both traditional mapping methods and GPS (Global Positioning System). The surveys include a topographic landscape analysis using aerial photography and ordnance survey maps, as well as the detailed mapping produced within the project. Local site grids and excavation plans have, together with the mapping data, been compiled in MapInfo GIS program for the final analyses and presentations. Two sites were chosen for larger excavations, Odubekoya in Labai and Oilobogwa in Mwatawa.

The extensive field surveys and preliminary mapping, including phosphate surveys, as well as a comprehensive registration of stray finds, was carried out as an MFS survey (Minor Field Study) in October-December 1998 by two students from Gotland University College. The MFS program was financed by the Swedish International Development Authority (SIDA).

The MFS field project consisted of three major parts. The first part of the project was a registration of the numerous stray finds in the area, e.g. the artefacts found by the villagers in the gardens and other areas surrounding the villages. After the field-survey of stray finds, maps showing the distribution of findings were produced. GPS was used for the mapping. Thirdly, phosphate surveys were conducted on located settlements in order to determine their extension. The phosphate surveys were carried out by using the spot-test method.

The investigations of August-November 1999

Based on the results from the MFS surveys, a team of sixteen archaeologist and osteologists from Gotland University College, Visby, the Archaeoosteological Research Laboratory, Stockholm University and the Department of Archaeology, Stockholm University, made a series of test excavations of stone age settlements identified around the villages of Labai and Mwatawa on northern Kiriwina under the direction of Maria Davidsson. Also, a preliminary osteological study of the human bone material found in two of the burial caves in the Labai area, Selai and Budou, was carried out. One of the located settlements in Labai, Odubekoya, revealed a well preserved burial ground, and so far five graves containing at least eighteen individuals have been excavated. The bone material is very well preserved, in the burial caves as well as in the settlements and burial ground, due to the calcareous environment, and a series of samples have been collected for C14 dating and also for post mortem DNA and stable isotope analyses. Also, rich quantities of charcoal from the excavated settlements will allow for additional radiocarbon dates. Rich quantites of pot-sherds and stone artefacts, including obsidian, have been recorded.

Selai Cave

The entrance to Selai Cave is located 10.62 metres above the sea level, and the lowest part 3.75 metres above the sea level. The total depth so far explored is about 50 metres. Eight halls have been investigated, named Hall A-H, and altogether twenty-two depositions of human bones have been recorded, containing males, females and children. A number of the crania were trepanated. No grave-goods accompanied the buried individuals.
A striking feature in Selai Cave is that most of the bones are covered in calcitic speleothems, in some cases in the form of stalagmites. Samples from this speleothem growth have been collected for uranium dating and stable isotope analyses. The fact that the human bones and crania are often completely "cemented" onto the rock, and consequently not possible to move without causing damage, is an important taphonomic indication that their position has not been disturbed in modern times. This makes the bone depositions in Selai Cave especially important in the forthcoming studies of the burial traditions.

Radiocarbon date from Selai Cave:
Ua-15487 (Sample 1/99, dens, ID 60001) 315±55 BP / c. AD 1570 CAL.

Budou Cave

The entrance to Budou Cave is located 12 metres above the sea level, and the lowest parts are situated below the present sea level. The total depth so far explored is about 60 metres. Five halls have been investigated, named Hall A-E. An assemblage of fresh-water at the bottom of the cave constitutes a separate section of the cave, Hall B. Altogether, twenty-one depositions of human bones have been recorded, containing males, females and children. A number of the crania were trepanated. Several of the depositions have been placed in large sea-shells (Tridacna gigas). No grave-goods accompanied the buried individuals. 

Obuwaga Caves 1 and 2

The Obuwaga cave is fairly easily accessible and situated close to Labai village, and frequent visits to the cave in recent times have displaced the bone depositions. Photographic documentation from 1984 shows that the depositions of human bones appeared then to be more or less intact. Also two large ornamented clay vessels, containing human bones, were almost intact. During the investigation in 1999, both pots were found broken into pieces.
Samples of pottery and bone were collected for analyses.

Radiocarbon date from Obuwaga Cave 1:
Ua-15985 (Sample 2/99, dens, ID 60002) 445±75 BP / c. AD 1450 CAL.
Bwara Tudava Rock Shelter, Labai
At Labai Beach, overlooking the sea, a rock shelter burial was located about six metres above the beach, containing bones from several individuals. Three pointed limesticks made of human bones (radius) were associated with the burials. Samples were taken for radiocarbon dating and DNA analyses.

Radiocarbon date from Bwara Tudava Rock Shelter:
Ua-15990 (Sample 10/99, dens, ID 60010) 200±85 BP / c. AD 1650-1850 CAL.
Oilobogwa Site, Mwatawa
The Oilobogwa site is situated north-west of Mwatawa village. Twenty-seven square metres were excavated. Most excavated sections proved to be very thin, providing a total depth from topsoil to limestone rock of about 20-30 centimetres. Nine post-holes were documented in the excavated areas. The find material consists of pottery, tools and flakes of lithics and charcoal. The excavations at Oilobogwa revealed approximately six kilograms of pottery, about 21% of these were decorated. Among the lithic material, a large number of obsidian flakes were recorded. Dated charcoal samples are most likely associated with recent gardening.

Radiocarbon dates from Oilobogwa:
Ua-15988 (Sample 8/99, charcoal, ID 60008) 134±1 BP / c. AD 1740-1930 CAL.
Ua-15989 (Sample 9/99, charcoal, ID 60009) 105±70 BP / c. AD 1740-1930 CAL.
Odubekoya Site, Labai

The Odubekoya site is situated on a hill north-west of Labai village. Thirty-six square metres were excavated. As at Oilobogwa, the total depth of the stratigraphy was 20-30 centimetres. Ten deep holes in the solid coral rock were documented, all with a diameter of 30-40 centimetres, and a depth of as much as 2,5 metres. The holes may have been dug for planting prestige yam (kuvi).

During the excavation of Odubekoya five burial depositions were found, containing at least seventeen individuals, males, females and children. None of the crania were trepanated. Various grave-goods accompanied the skeletons in most of the graves, including obsidian (memetu), magical stones (bina-bina), axes/adzes (utukema) and pottery. The excavations at Odubekoya revealed approximately fifteen kilograms of pottery, about 28% of these were decorated. Test pits outside the main excavation area were excavated with the aim to demarcate the site, and the settlement and burial ground can be provisionally estimated at sixty-one by forty-five metres, or c. 2.745 square metres.

Radiocarbon dates from the Odubekoya burial ground:
Ua-15467 (Grave 1, Sample 4/99, dens, ID 60004) 930±80 BP / c. AD 1100 CAL.
Ua-15468 (Grave 2 / Individual 5, Sample 5/99, dens, ID 60005) 1100±70 BP / c. AD 950 CAL.
Ua-15986 (Grave 3, Sample 6/99, dens, ID 60006) 755±70 BP / c. AD 1250 CAL.
Ua-15987 (Grave 5, Sample 7/99, dens, ID 60007) 1045±80 BP / c. AD 1000 CAL.

Preliminary conclusions

The archaeological and osteological investigations in 1999 have revealed a series of data that may indicate significant cultural change on the Trobriand Islands between AD 1250 and AD 1450. This applies to both bone deposition traditions, the manipulation of the bones, ceremonial features, and grave goods. The radiocarbon dates so far available from the Odubekoya burial ground show that the individuals were interred between c. AD 950 and AD 1250, while dated samples from individuals in two of the burial caves center around c. AD 1500. Also, the inhumations at Odubekoya are final interments with skeletons in anatomical positions, as opposed to the cave burials, the bone depositions of which can be shown to be secondary burials, e.g displaying defleshing marks, with the bones in non-anatomical positions. Furthermore, a large number of the crania in the caves have been trepanated, while none of the individuals so far excavated at Odubekoya display this feature. The dates from the Odubekoya burial ground corresponds well to Bickler's first phase on Woodlark, which has been dated to c. AD 800-1200.

Looking at the pottery associated with the burials at Odubekoya, and the sherds documented in the Obuwaga burial cave, again distinct differences can be shown. A detailed presentation and analysis of the ceramics will be presented elsewhere. It is reasonable to believe that the megalithic tradition developed on the Trobriand Islands during this intermediate period (Austen 1940). Such a date would correspond well to the appearance and subsequent development of most other megalithic traditions in the Pacific area.

If the cultural change suggested here between AD 1250 and AD 1450 is the result of a continuous local development, or an appearance of new people, cannot be archaeologically determined without further excavations. However, the results from recent mitochondrial DNA, HLA and Y-chromosome polymorphism analyses on today's Trobriand population have indicated strong Polynesian markers. The mtDNA data are highly suggestive of a recent migration of Polynesian maternal lineages to the Trobriand islands, and the results argue for a considerable back migration from Polynesia to island Melanesia and coastal New Guinea in recent times. The forthcoming DNA-analyses of the human bones from the Odubekoya burial ground and the burial caves may prove to be of great importance in this context.


Burenhult, G. 2000a. The Trobriand Islanders – Original Settlers or Later Migrants? I: Migrations and Exchange in a Historical Perspective. No Barriers Seminar Papers, Vol. 3, 2000. The Kon-Tiki Museum, Institute for Pacific Archaeology and Cultural History. Oslo.

Burenhult, G. (ed.) 2002b. The Archaeology of the Trobriand Islands, Milne Bay Province, Papua New Guinea. Excavation Report 1999. B.A.R International Series 1080.

Burenhult, G. 2011. Begravningsgrottorna på Trobrianderna. I: Traveller's Club Malmö Jubileumsbok 1936-2011. Malmö.

Burenhult, G., Davidsson, M., Larsson, A.-C., Svensson, S. & Venturi, G. 2000. Bland grottor och gravar i Söderhavet. Populär Arkeologi Nr 1, 2000.

©2014 Burenhult Photography | Powered by Arkeomedia