Between July 20 and August 2, 2024, a team from the University of the South Pacific and The Fiji Museum took part in archaeological research on Vanua Levu (Fiji) in the Provinces of Bua and Macuata. The project is a collaboration funded by the University of the South Pacific Research Office and European Commission via OCSEAN. The project was designed to address questions related to two shell midden deposits and to train local students in archaeological methods during the field school.
For more than 120 years after it was first described by a naturalist, questions concerning the origin and formation processes of the ca 80 m-long shell deposit along the Bua River, locally known as Rokodavutu, remained largely unanswered. In 2016, researchers from USP, Fiji Museum, and The University of the Sunshine Coast obtained the first radiocarbon dates from the site. The results were anomalous, in that shell samples taken in close proximity to one another showed vastly different age ranges, from about four thousand years to five hundred years Before Present. The early dates pre-dated any previously known evidence of human occupation in Fiji and other parts of Remote Oceania, which is commonly attributed to the arrival of migrants bearing elements of the Lapita Culture about three thousand years Before Present.
Map of sites of the Archaeological Field school and excavations
It was decided to return to the site in July 2024 to obtain additional radiocarbon dates and possibly uncover artifacts or other evidence of human occupation, as well as provide a more detailed stratigraphic description of the sediments above and surrounding the midden exposed at low tide to assess the degree of environmental disturbance in an active fluvial setting by ‘slicing’ through the currently exposed riverbank instead of the usual vertical excavations.
Team members from USP, Fiji Museum, and the local community
This approach suited the project’s objectives and also saved time to reach the ca 40 cm-thick shell deposit overlaid by about 4 m of clayey sediments with no visible artifacts. The research conducted in 2024 was successful in obtaining additional samples for dating (to be submitted to a dedicated laboratory at the University of Waikato in New Zealand) and a record of the stratigraphic profile.
Shell midden at Rokodavutu at low tide
No artifacts were identified. Because of their absence, the possibility remained that the midden was the result of a natural deposit when sea-level was appreciably higher than at present, about 1.5 m between 4000 and 3000 years Before Present. However, the roughly even distribution of only three edible shellfish species is suggestive of likely selection by people rather than an ancient intertidal environment where many more species would be expected to be identified. It is hoped that sediment samples collected at the site will yield micro-artifacts and plant microfossils indicative of human impact. No specific oral traditions connected to the site were recorded, other than the area being close to a former SDA mission.
Stratigraphic profile at Rokodavutu (Note: layer 5 is where the shell midden is identified)
Another midden site was also investigated on the ‘shell island’ of Culasawani in Macuata, measuring about 8000 m². Like Rokodavutu, no specific oral traditions were collected. The midden was briefly investigated in 2023 by a team from the Fiji Museum and The University of the Sunshine Coast. Radiocarbon dates clustered neatly around 1670 Before Present, about one thousand years after the end of Lapita. Some scattered plain pottery sherds were recovered from three test pits during the most recent research at the site.
The ‘shell island’ at Culasawani
There is a consensus among team members that the extensive shell deposit is likely the result of a tsunami (or another big wave) that might have disturbed part of an anthropogenic midden deposit (as suggested by the clustering of the dates), together with the dominance of the same three shellfish species previously identified at the Bua River midden. It is hoped that the dates obtained on shells will be matched with known records of tsunami or big wave events along the coast. Again, sediments samples from the excavation pits were collected for micro-artifact and microfossil analyses.
Quite unexpectedly, the research team was able to find the time to look for the presence of a pottery-bearing site inland from Otubelu Bay on the western side of Bua Province, which was suspected to exist when a survey was conducted eight years previously. In 2016, pottery sherds, some decorated, were found on the reef flat at low tide. A single test pit a few meters inland from the current shoreline did not yield any pottery or other artifacts. Because of fluctuations in sea-level, it was believed that any relatively intact cultural deposit would be found further inland. During the 2024 research, large numbers of pottery sherds were again located at low tide. A surface survey from the shoreline to the base of nearby cliffs led to the discovery of several pottery sherds of various sizes, including decorated rim and other fragments with distinctive Lapita designs.
Decorate Lapita pottery from Nakorovatu
Cleaning pottery in the field
On cursory examination, the designs did not display any dentate-stamped decoration (a hallmark of Early and Middle Lapita), and instead had decoration reminiscent of Late Lapita. Three test pits were excavated at Nakorovatu, yielding more pottery. One pit yielded the base of a pot associated with charcoal suitable for radiocarbon dating. It could not be determined if it was of Lapita or post-Lapita age.
Uncovering the base of a pot at Nakorovatu
These sherds were carefully handled so as not to potentially contaminate them for subsequent analysis of food residue to be conducted at ANU, possibly adding to the corpus of information on ancient pottery use in Oceania so far limited to the analysis of sherds from the Early Lapita site of Teouma in Vanuatu.
Comments