Human cultural diversity in prehistoric Fiji

Human cultural diversity in prehistoric Fiji Ethan E. Cochrane Remote islands and their human, animal and plant populations have long fascinated archaeologists, biologists and geographers. In this article, the chronology, diversity and interactions of human cultures in some small islands of the Fiji archipelago are explored, particularly through the application of sophisticated chemical analyses of the composition of prehistoric pottery.

Remote islands and their human, animal and plant populations have long fascinated archaeologists, biologists and geographers.In this article, the chronology, diversity and interactions of human cultures in some small islands of the Fiji archipelago are explored, particularly through the application of sophisticated chemical analyses of the composition of prehistoric pottery.
T he Fiji Islands (Fig. 1) were first inhabited in approximately 900 by colonists sailing eastwards from island Melanesia.1 Like all the founding populations of western Remote Oceania -from Vanuatu and New Caledonia, to Fiji, Tonga and Samoa -the first Fijians were part of a related group of colonizing peoples sharing aspects of language, biology and also material culture, including the well known Lapita pottery.Many archaeolo gists, anthropologists and other scholars suggest that, over the past three millennia, these once-similar populations diverged fr om their common origins.2Our current research in Fiji investigates the generation of cultural difference over 3000 years of human occupation.Specifically, how do we explain present-day cultural diversity across Fiji and Remote Oceania?Is cultural divergence the most appropriate model?3

Culture history of Fiji
To measure changes in human cultural diversity we must first generate a repre sentative and precise description of the archaeological record in a region.This has been a primary goal of our research in the western islands of the Fiji archipelago over the past few years.In the Yasawa Islands, the focus of our current research (Fig. 2), chronological variation in the material culture record shares broad similarities with the rest of Fiji.
The Y asawa Islands were first inhabited

AUSTRALIA
in approximately 700 and have been home to human populations since then to the present.The initial colonization and sus tained occupation of the Yasawas probably occurred several hundred years after the initial habitation of sites in other parts of Fiji, particularly in the east of the archipel ago.The earliest identified occupations of the Yasawa Islands, at sites such as Olo (Fig. 3) and Qaranicagi, is indicated by pottery with vessel forms and decorative attributes similar to the so-called "ter minal Lapita"4 pottery found at other sites in Fiji and Remote Oceania.The earliest inhabitants of the Yasawas lived in small communities on the coasts, but probably also spent time in the uplands, where occupation sites have been found in caves and on ridges.Unfortunately, the limited research on early settlement patterns and mobility means that we can only speculate about this aspect of life in the islands.
The earliest inhabitants also left a record ofboth artefacts and fo od remains, indicat ing a heavy reliance on marine resources, but the earliest inhabitants of the Yasawas consumed chickens in addition to plant resources that were probably grown in gar dens.Modified shells found at Olo may have been used as root peelers, and ham mer stones may have been used to extract the kernels fr om nuts.Evidence of a reli ance on marine resources occurs through out the Yasawa Islands sequence, with fishbone and marine shell occurring in deposits of all ages.Other animal resources Pottery is the most abundant category of artefact recovered in the Y asawa Islands and it displays a range of decorative and for mal variation that is used to divide Fiji-an prehistory into phases or periods of relative cultural homogeneity.Like all archaeologi cal phases, those used in Fiji chop a record of continuous temporal and spatial vari ation into convenient packages.7Yasawa Islands pottery, as with pottery throughout Fiji, is classified into the following phases:

PA CIFIC
• Sigatoka phase (900-500) pottery assem blages comprise sherds that are often intricately decorated and known by the archaeological term Lapita.
• These assemblages contain a diverse array of vessel shapes, including cari nated bowls, pot stands, water jugs, and various types of cooking pots.• Ra phase ( AD 1800-1900 ) pottery ass em blages are characterized by increasingly complex incised and applique patterns.
There is little evidence fo r large-scale craft specialization associated with pottery manufacture in Fiji.Most pottery appears to have been made by households that were not integrated into larger specialized pro duction systems, but no systematic research has investigated the possibility of special ized ceramic production.
Archaeologists have long argued that similarities in pottery decoration and ves sel shape across a region may reflect inter action between producers of the pottery.10Assemblages of pottery to be found in the Yasawa Islands are similar to those found throughout Fiji, as described by the phases above, and it therefore appears that people in the Yasawas were interacting with pop ulations throughout Fiji, sharing informa tion about pottery at similar intensities throughout prehistory.However, the simi larities in decorative features and vessel shape between pottery in the Yasawas and the rest of Fiji may not precisely measure interaction between different groups.Other potential measures of interaction, such as language , 11 • 12 show great differences across the archipelago and suggest that Fijian populations may have interacted less fre quently than implied by these pottery data.The resolution of these apparent contra dictions is one of the principal objectives of my research in the Fijian islands.

Fijian cultural diversity: new questions and methods
Our current research in Fiji is based on a framework of evolutionary and ecological theory, within which human diversity is explained as the result of a variety of We generated geochemical provenance data for 260 sherds from sites throughout the Yasawa Islands using an analytical technique known, rather dauntingly, as laser-ablation inductively coupled plasma mass-spectrometry ( LA-ICP-MS, for shortlY The chemical data were subsequently ana lyzed using multivariate statistical tech niques to define groups of pot-sherds with similar compositions and thus probably from pots made from the same, or similar, clay sources.By studying the geology of the Yasawa Islands, we were able to link sherd compositional groups to general geographical areas having particular types of clay resources.Finally, after examining the varying proportions of sherd composi tional groups present in Yasawa Islands assemblages over time, we were able to track the changing geographical scale of geological clay resources used by the island inhabitants. The results of the LA-ICP-MS analyses indicate the shifts in clay resources used by Yasawa Islands populations and, there fore, the likely changes in the interaction between different groups (with interaction measured by the use of clay resources from different localities).For the first several hundred years of occupation in the Yasa was , people used vessels manufactured from clay deposits distributed throughout the Yasawa Islands and probably also the Mamanuca Islands to the south.This sug gests that the earliest groups in the Yasawa Islands interacted within a broader inclu sive population fram ework that was spread across the Yasawa and Mamanuca Islands, although the proportions of different clay types in these early pottery assemblages indicate that people usually used vessels manufactured fr om clay sources nearest them (Fig. 6).For the earliest populations, the frequency of interaction across space appears to decrease with increasing dis tance.
Some 1000 years later, at about AD 500, the pottery assemblage in levels 17-14 of the southern Yasawas site of Qaranicagi, on Waya Island, contains approximately equal proportions of sherds made fr om clays derived fr om the two principal geo logical sources (Fig. 6 ) .This suggests that, by AD 500, the inhabitants of Waya Island made equal use of pottery produced from clay deposits across the Yasawa Islands and Mamanuca Islands and were, there fore, part of an interaction system that stretched across the Yasawa-Mamanuca island arc.
This widespread interaction ends at about AD 1000, as indicated by pottery

Figure 1
Figure 1 Th e southwest Pacific, showing major island groups and the boundary between Near and Remote Oceania.The box on Fiji shows the area covered by Fig ure 2.

Figure 2 Figure 4 Figure 3
Figure 2 Western Fiji, showing the Ya sawa and Mamanuca Island groups, and some principal archaeological sites in the Yasawa Islands.-I

Figure 5
Figure 5 A view of central Nacula island, looking northeast.The site of Druidrui is located on top of the rocky hilltop in the middle ground.processes, including the interaction and transmission of ideas between individu als , variation and change in environmental and ecological factors, adaptation, and convergence.13If contemporary cultural diversity in Fiji is at least partly a product of variation in the spatial scale and fre quency of human interaction in prehis tory, what aspects of pottery variation might help us track any such changes in interaction?Provenance analyses of arte facts are one method for tracking inter action in prehistoric populations.Artefact typically refers to the geo graphical location where an artefact was made.By comparing artefact provenance with the location of the artefact's final deposition in the archaeological record, we can estimate the geographical range covered by the people involved in the manufacture, distribution and use of the artefact.Changes in the geographical ranges of populations over time may indicate changes in the spatial extent of interaction.Cochrane & Neff14 recently conducted geo chemical provenance analyses of archaeo logical pottery from sites in the Yasawa Islands to compare the likely location of pottery manufacture with the location of its ultimate deposition.We generated geochemical provenance data for 260 sherds from sites throughout the Yasawa Islands using an analytical technique known, rather dauntingly, as laser-ablation inductively coupled plasma mass-spectrometry ( LA-ICP-MS, for shortlY The chemical data were subsequently ana lyzed using multivariate statistical tech niques to define groups of pot-sherds with similar compositions and thus probably

Figure 6
Figure 6 Grouped bar chart showing pro portions (with 95% confidence intervals) of sherd compositional gro ups in Yasa wa Islands pottery assemblages.Assemblages are shown in chronological order, with the oldest at the fo ot and the youngest at the top of the chart.Q = Qaranicagi site.