An archaeology of salt production in Fiji.

Burley, David V. ; Tache, Karine ; Purser, Margaret 等

[ILLUSTRATION OMITTED]

Introduction

Salt has played a pivotal role throughout world prehistory as an essential dietary supplement, as a commodity for specialised production and exchange and, for some societies, as the economic foundation upon which polities formed and wars were fought. Numerous researchers identify salt access as central to the development of complex societies (Connah 1991; McKillop 1995; Lovejoy 2003). Flad et al. (2005: 12618) claim its presence in all early states. Not surprisingly there is a profuse archaeological literature on salt in antiquity from diverse areas across the globe. While much of this focuses on archaeological and/or historical evidence for salt processing and its context in culture history, there has been some concern surrounding its broader role in economic and social processes. Indeed, treating salt production more generally as craft production positions it within a robust literature on craft specialisation, social aspects of production and the agency of producers, among other issues (see papers in Costin & Wright 1998; Flad & Hruby 2007a; Hirth 2009a).

The following paper provides and interprets data for solar evaporation salt extraction in Fiji, an industry previously undocumented within Oceanic prehistory. The short-lived salt-working site was excavated at the Sigatoka Sand Dunes on the island of Viti Levu and dates to the seventh century AD. We describe the site in its context, arguing for the employment of dedicated processing stations along the Sigatoka shoreline using large ceramic saltpans. These data allow us to infer specialised production, and they provide insights into its context, scale and organisation. We also explore the larger role of salt as a commodity in Fijian prehistory as well as its economic and social value as a product for exchange. The current shortage of archaeological data beyond Sigatoka makes these issues difficult to address, but scattered historical references to salt production and observations from a contemporary salt-making village at Lomawai, do provide insight into social and political processes at work, with potential implications for the archaeology of salt production in Fiji.

The archaeology of salt production at Sigatoka

The Sigatoka Sand Dunes at the mouth of the Sigatoka River on the Coral Coast of Viti Levu, Fiji, provide a unique and important archaeological complex for Fijian prehistory (Figure 1). Archaeological sites, buried episodically and rapidly by blowing sand over the past 2700 years, are today reappearing as sand erodes from the dune front slope (Marshall et al. 2000). Large-scale excavations here in 1965 by Birks (1973) provided insights into depositional processes, periods of dune stability, chronology and ceramic prehistory. From within a buried palaeosol labelled 'Level 2', Birks (1973: 44-5) recovered no fewer than 17 'rough finished' shallow flat-bottomed 'dishes' with diameters of 0.5m or more. He lists a range of potential functions for the vessels, one being for the evaporation of sea water to make salt (Figure 2).

More recent surveys at Sigatoka (Marshall et al. 2000; Burley 2005) report widespread and, at times, extremely dense concentrations of these dishes eroding from the palaeosol over a distance of 1 km along the sand dune shoreline. Birk's dishes are inferred now to be salt trays and their locations to indicate solar evaporation salt processing stations, logistically positioned along the shoreline to take advantage of sun and wind exposure (Burley 2003).

The salt trays are diagnostic of the Navatu phase at Sigatoka, a mid-sequence period in Fijian prehistory, generally dated between AD 200 and 1000 with regional variations (Clark 1999: 85; Burley 2005: 342). Navatu phase ceramics, other than trays, incorporate a variety of forms, importantly including a highly distinctive, well-fired, everted rim globular jar that frequently has carved paddle impression on the body and a decorative suite potentially including applied relief, end-tool or finger nail impressions and incised patterns (Frost 1979; Marshall et al. 2000; Burley 2005). These jars are found in tandem with the trays in the salt processing areas (Figure 3).

The Navatu phase settlement occurred in a small village close to the mouth of the Sigatoka River at the eastern end of the dune (Figure 1). Excavations in the village in 2000 and 2002 exposed or recovered a variety of habitation related features, ceramics, other artefacts and faunal materials (Burley 2005). Tray fragments are notably rare in the village assemblage, including only three in a collection of almost 11 000 ceramic sherds (Burley 2005: 327). A pooled mean age of radiocarbon dates (n = 3) for the Navatu village places it in the interval 610-675 cal AD at a 2[sigma] range.

[FIGURE 1 OMITTED]

Excavation of a salt processing station

Excavations were undertaken at a salt processing station in 2006, located from sherds of salt trays about 1km west of the village site. The aim was to identify features or activity patterns associated with the production process and to gain a better understanding of the ceramic assemblages. The project included mapping and collection of surface ceramics as well as excavation of in situ material over a block area of 45.5[m.sup.2] (Figures 4 & 5). At the time of use, the salt processing station would have been situated on a flat or slightly rising terrace overlooking the Sigatoka beachfront. It was strategically positioned to take advantage of sun and wind exposure for the evaporation process but continued to be within easy reach of the water's edge. The former ground surface on which this activity took place is an inceptisol with well-defined A and B horizons (de Biran 2001: 51). Horizon structure indicates a stabilised land surface that, as found by Best (1989: 48) elsewhere on the dune, may have been consolidated by stands of Casuarina litoralis. A substantial cover of drift sand had buried the salt station, either causing abandonment or following it. Blowing sand, in fact, may have forced an exodus not only from this locale but the Navatu phase village as well (Burley 2005: 327).

[FIGURE 2 OMITTED]

[FIGURE 3 OMITTED]

[FIGURE 4 OMITTED]

Excavation of the processing station revealed minimal if any infrastructure. Ceramic fragments were spread across the area, from a central focus 1 m in diameter where there was a large segment of tray in situ. Faunal remains and artefacts other than ceramics were absent. The presence of only two small postholes suggests the trays were employed on the surface during use, rather than raised on platforms. A small hearth to the north-east is too limited in size or intensity for salt extraction from a boiling operation or for use in the firing of trays or jars. Diagnostic of the dedicated nature of the salt processing operation was the exclusive occurrence of only two types of ceramic vessel: trays (n = 7890 sherds) and the everted rim Navatu jars (n = 2090 sherds) (Figure 3). By comparison, equivalent sized excavations in the Navatu village in 2000 recovered nine other Navatu jar and bowl forms (Burley 2005: 327). The jars, we suggest, were for water transport and the trays were installed on an exposed and levelled surface as evaporation pans (Burley 2003: 313).

[FIGURE 5 OMITTED]

Salt tray manufacture

The salt trays are large flat-bottomed vessels with ovoid to circular platter-like forms. They are manufactured of tephra-based days with little care for aesthetic concern or consistency in stylistic elements. Coarse mineral sands of variable size were added as temper. Diameters range from 0.52-0.78m with estimated rim heights for 2006 specimens up to 65mm. Tray weights vary based on vessel shape and diameter. A 0.76m diameter circular vessel with almost complete rim course has a calculated weight of 12.5kg.

The size and weight of trays, and the limitations this creates in construction, firing and transport has been highlighted by Birks (1973: 45):

'Without sides high enough to give it additional strength, a vessel of this size, made of clay fired at a low temperature would be very fragile, and it seems unlikely that it could have been lifted even after firing without breaking under its own weight. [...] The conclusion u therefore reached that each [...] dish was probably dried, fuel placed on top of it, and the vessel fired and subsequently used on the spot where it was made.'

The problem of collapse during the formation of the tray was partially resolved by making it on a bed of leaves, twigs or mats. Impressions of these materials occur on the underside of virtually every base sherd. This matted layer could have provided support during sun-drying and would have facilitated transfer of the tray to a firing hearth. A probable firing hearth was excavated 70m east of the salt processing station. It took the form of a 1 x 1.5m shallow scoop containing charcoal, fire-broken volcanic stone, coral limestone pieces and large pottery sherds, including tray fragments (Figure 6). Pre-heated, these stones and sherds would act as a foundation that retained and distributed heat over the vessel surface during firing, aided by additional fuel heaped inside the tray as suggested by Birks. We believe this type of preparation was the only way a vessel of this size and form could be fired on an open hearth.

[FIGURE 6 OMITTED]

Appropriate potting clay does not occur at the Sigatoka Sand Dunes, nor do some other of the materials required in the production and firing process for tray manufacture. Clay, volcanic stone and limestone cobbles would need to be transported to the site, in most cases over a distance of 1km or more. A basket-sized quantity of unfired clay with degraded coral limestone cobbles found eroding from the dune surface in 2002 supports this idea (Burley 2005). Fuel for the firing process could be collected as driftwood along the beachfront, where large hardwood logs frequently wash ashore. Whether this was fully adequate for both tray production and domestic use in the village is unknown; additional fuels might also need to be imported.

Diana Tugea, a master potter from Nakabuta village on the Sigatoka River, was engaged in 2002 to replicate a Navatu phase tray for a salt processing evaporation experiment. She was given a descriptive template, and further shown archaeological sherds from the Sigatoka Sand Dunes. Diana employed local clay mixed with fine-grained iron sand temper from the Sigatoka delta. Her initial two attempts failed in their firing on an open hearth where the trays broke, either as a result of inadequate heat distribution, inappropriate temper, or both. A third effort produced a 0.40 x 0.35m tray weighing 9.5kg that, while cracked in firing, was usable after patching (Figure 7). The tray fabric was semi-porous, a characteristic that we again attribute to inadequate firing and temper. When filled with salt water, the fabric saturated and slowly drained requiring a relatively continuous addition of fluid. Despite this problem, and cloudy weather conditions limiting evaporation, an incipient salt crust formed on the rim and surface of the tray with the cumulative addition of 6.5 litres of saltwater. Substantial variation in manufacture between archaeological trays and the contemporary one create a poor analogy for inference of the salt production process during the Navatu phase. The experiment, however, emphasises the difficulty in tray manufacture, and the specialised construction and firing processes required.

[FIGURE 7 OMITTED]

Characteristics of the Sigatoka industry

The use of dedicated processing stations for evaporation of seawater and the manufacture of evaporation pans is paralleled in several other areas throughout the globe (Chiang 1976; Andrews 1983; Muller 1984). Craft specialisation in salt or other commodities provides surpluses for exchange, where a producer's essential needs are met in whole or in part through barter, or through the support of managerial elites for whom the commodity is produced (Stein 1996: 25). A series of correlates from which specialised production can reasonably be argued has been offered by Evans (1978:115), including specialised areas for craft activities, specialised tools, exploitation of particular resources, and integration of the product in trade and exchange systems. The Sigatoka complex, it can be suggested, meets such requirements. Flad and Hrubry (2007b: 6) nevertheless argue that meeting the 'standards' of specialisation should not be a focus in and of itself. Rather 'such studies should explore the various parameters that comprise the organization of production. 'Here we examine four of these - the context of production, the intensity of production, the scale of production and the organisation of production.

We infer that the context and economic raison d'etre for salt harvesting at Sigatoka, on any level of production, was for trade to non-coastal communities where salt was required as a dietary supplement. Coastal peoples themselves with access to seawater and at least a partial diet of marine foods have limited need to process salt for domestic use. The intensity of production depends on whether salt processing at Sigatoka was a year long or seasonal undertaking, and whether the producers were full- or part-time participants. Solar evaporation of seawater is conditioned by a range of factors including rainfall patterns, sunlight hours, sunlight intensity, wind and variations in salinity among others (Akridge 2008). At Sigatoka there is acure variation in these conditions associated with well-defined wet or dry seasons. The wet season occurs in the November to April period when 75 per cent of Fiji's annual precipitation occurs (Harris 1963). Greater numbers of rainy days correlates with fewer sunlight hours while wet season weather patterns are additionally defined and impacted by tropical cyclones and intensified storm activity (Parry 1997). Coastal and interior precipitation rates vary, with rain shadow effect of the mountainous interior enhancing accumulated totals in the upper Sigatoka Valley. This results in a substantially greater wet season discharge from the Sigatoka River that, concomitantly, reduces salinity along the coastal margin where salt was being processed. We argue, thus, that the efficiency of salt production is substantially lowered in wet season months, if salt can be produced with any regularity at all. The absence of infrastructure in the excavated salt processing station, as would be required for shelter from periodic rainfalls or storms, supports the inference of a dry season harvest by intermittent producers.

The scale of production should be indicated by the number of discarded trays. We estimate a minimum of nine to ten trays within the 2006 excavated assemblage based on a cumulative sherd weight of 91.2kg and an assumption that the average tray weighs 9.5-10kg. This number, it is argued, should be multiplied by a factor of ten to account for dense surface concentrations of trays previously documented (Marshall et al. 2000: 29) and those remaining buried in situ. Further assuming that an average tray has a 0.65m diameter, and that the typical salt crust in a 65mm deep vessel after full evaporation is 5mm, a single processing event results in 1.66 litres of salt. Akridge (2008: 1457) calculates that solar evaporation of a 5mm salt layer in Chinese wooden pans with a 25-60mm depth takes five days. Employing only a third of the trays potentially present (n = 30), upwards of 900 litres of salt could be produced in five-day processing shifts over a three month period at Sigatoka. Multiple processing stations, prolonged dry season production and/or greater numbers of trays in use at the same time not only enhance this volume but imply a sizeable scale of production for the industry.

Finally, the organisational context might be suggested as household-based rather than centralized elite-control, arguing from the dispersed processing nodes. There exists no recognisable advantage or functional explanation for separate production areas otherwise. Muller (1984) and de Leon (2009) convincingly argue for household-based organisation in their respective reassessments of Mississippian and Aztec salt industries; Hirth (2009b: 4) reports its widespread applicability to craft production generally. If the salt was not produced for local use (above) it becomes an economic supplement to household subsistence. Archaeological evidence in the Navatu village suggests food stress for the salt processing community (Burley 2005: 332). Salt harvesting in this light may have been a primary basis upon which the household economy was centred. In a comprehensive study documenting a millennium of salt-making procedures and equipment between the Middle Iron Age and the end of the Roman period in the Fenland of eastern England, Morris (2001: 397) notes cases where salt probably represents the major source of household income, even though full-time production was impossible due to a variety of weather conditions.

Discussion: context and analogies

Initial settlement of the upper Sigatoka Valley and interior highlands began 2000 years ago with a small and dispersed population. Field (2003: 258-9) suggests a food exchange system was in place almost immediately, given the presence of non-local riparian clams (Batissa violacea) in one of the earliest sites. Between 1500 and 1000 years ago archaeological site distributions indicate a population expansion that filled the valley bottom and foothills. It seems not coincidental that the beginning of this expansion period coincides with the sudden appearance of salt processing activities at Sigatoka. We suggest the procurement of salt facilitates, if not critically underpins, this expansion. The expanding interior settlement resulted in deforestation, slope erosion and increased sediment load in the Sigatoka River, the latter being a factor for enhanced sand dune growth at the mouth of the river (Dickinson et al. 1998). Ironically, the same inland population that created a demand for salt may have been an agent for the demise of its production at Sigatoka.

The Sigatoka River continued to provide a principal corridor into interior highland communities, and is later well-documented as a route for coastal/interior trade. Historical references by Tonganivalu (1917: 9) and Williams (1858: 94) specifically highlight salt as a component of this exchange, leading Tanner (1996: 234) to claim salt as a resource both prized and essential. There are few descriptions of indigenous salt production in Fiji. The most detailed of these by Williams (1858: 71) for the island of Vanua Levu notes only that it is 'good salt, but of a sandy colour', and that it is "procured by evaporation, and preserved near the fire in baskets made for the purpose" Rather than solar evaporation, this reference suggests a brine boiling operation.

Indigenous Fijian salt processing continues to be practiced in a single village, Lomawai, on the western coast of Viti Levu to the north-west of Sigatoka. Merewairite Butani, a Lomawai elder who was instructed in traditional practices of salt-making in her youth, is credited with a resurgence of this industry in the 1990s. In a 2006 interview in Lomawai, she explained and demonstrated her salt-extraction process, one based on the boiling of brine. Saltwater is acquired at low tide from two excavated collection ponds in the adjacent mangrove swamp. Brine reduction is then accomplished through boiling in a large diameter heavy gauge aluminium boiler over an open fire in a specially constructed 'cooking house' on shore. This produces approximately 150-200mm of salt in the pot after a full day of boiling. The salt is packed into 1 litre milk cartons for shaping into blocks and the block then finished with a dogo dina, a woven basket-like fringe made from the inner bark of the mangrove.

The Lomawai chief, Ratu Kini Vosailagi, described Lomawai as one of several former salt-making villages along the Coral Coast of Viti Levu, stretching from Malomalo on the south to Yako on the north but importantly including Tau, Nabila and Nakorokula. The antiquity of this group of villages in the production of Fijian salt is attested to in Thomson's (1908: 203) statement that, historically, 'salt came only from the salt-pans of the mangrove swamps' Salt was produced largely for trade, for presentation as prestige goods to visiting chiefs, or for presentation at traditional events. In keeping with this statement, the salt being produced during the 2006 visit was to be given to members of the Nadroga/Navosa Provincial Council who were about to meet at Lomawai. Ratu Vosailagi believed traditional salt was an instrumental component of Lomawai economy and prestige, and its production forged political and social alliances for the coastal villages of western Viti Levu.

Ratu Vosailagi's comments on the role of salt as both a necessity and a prestige item, and on its importance in the establishment of social and political relations, relates precisely to ongoing concerns for craft production studies in general. Appropriately summed up by Flad and Hruby (2007b: 3), we must be concerned with the social aspect of production and the creation and perpetuation of social ties as a consequence. Salt producers and their product are given agency, rather than being passive components within an economic exchange network (Costin 1998; Clark 2007). When salt moved beyond a purely domestic need into the prestige system of chiefly elites in Fiji is difficult to determine. That it occurred prior to European contact is anticipated in the historic observations of Williams (1858: 94) who succinctly reports that the "Inland tribes of the Great Fiji [Viti Levu] take yaqona [kava] to the coast, receiving in exchanges mats, masi [bark cloth], and fine salt." The equivalence of fine salt with mats and masi is significant, for the latter are well-integrated prestige goods in traditional Fijian society. Salt, as both necessity and a prestige item, blurs a classificatory distinction central to many archaeological studies concerned with value, accessibility and the role of prestige goods in systems of emergent complexity (Flad & Hruby 2007b: 9).

Fijian society is foremost organised at the village level with groups of associated villages traditionally forming vanua. Inter-village salt production, as highlighted by Ratu Vosailagi, seems a catalyst in this type of political consolidation for western Viti Levu, as well as a mechanism through which social and political ties were maintained. With one exception, the distribution of sites having Navatu phase salt tray fragments is geographically restricted, most within the immediate vicinity of Sigatoka (Frost 1979; Burley & Tache 2008). Hypothetically at least, this cluster of sites may have formed a social and political unit of the type later found at Lomawai.

The exception is Karobo, 70km east of Sigatoka where several salt trays were excavated in the 1960s (Anderson & Clark 2009). The Karobo site parallels the salt processing station at Sigatoka in most respects. It was positioned on an exposed sand dune/spit of land on Korobo Beach. Trays and jars are the only two ceramic vessel forms present. The Navatu everted rito globular jar is absent but Clark (1999:159) describes a very high degree of consistency in jar form. Faunal remains were not recovered and non-ceramic artefacts include only two specimens (Anderson & Clark 2009: 108). Most important, Korobo is 6.5km east of the Navua River, another of the major waterways extending into the interior highlands of Viti Levu. We suggest Karobo represents either an extension of the Sigatoka salt processing group of sites or, potentially, it is part of an independent unit focused on inland trade along the Navua River.

Conclusion

The short-lived seventh-century AD solar-based extraction of sea-salt that took place in the Sigatoka Sand Dunes provides a first snapshot of a prehistoric island industry that must have been widely practised in the Pacific. Associated with the Navatu phase in Fijian prehistory, the industry employed large and heavy evaporation trays that in themselves were a specialised contingent craft. The trays were used in discrete salt processing nodes along the Sigatoka shoreline, intentionally positioned away flora the Navatu phase village.

We feel secure in our assertion that salt was being processed at Sigatoka for trade with non-coastal peoples. We also believe the evidence supports an interpretation that salt processing was seasonal, capable of production at a reasonably high scale, and was probably being undertaken by individual households acting as intermittent producers. The salt extraction industry documented at Lomawai 1300 years later relied on boiling brine collected in mangrove swamps, bur offers analogies to the earlier craft: the salt was supplied to users upriver as both a dietary supplement and prestige item. Later salt processing was also a catalyst for the development of social and political relations among salt producing communities. This latter observation has potential implications for interpreting the distribution of salt processing sites during the Navatu phase.

The production of salt through solar evaporation of sea water was not long lived in Fijian prehistory. We can only speculate as to why, given the importance of salt within the coastal/interior exchange network in the historic era. The development of a mangrove saltwater pond and boiling process might have proven more efficient. Alternatively mangrove salt may have produced a different and more attractive type of product. It may also be that the costs of solar evaporation, including the manufacture of large ceramic salt trays, simply outweighed the value of its return.

Acknowledgements

Each of the field projects at Sigatoka described in the text was undertaken as part of Simon Fraser University field schools in archaeological field methods between 2000 and 2006. We are grateful to the numerous students who participated in these field programs and thank them for their efforts. We also are grateful to Savaneka Dau and George Trail from Kulukulu village, Sepeti Matararaba, Fiji Museum, and Andrew Barton for their outstanding contributions to our programs over the years. Fieldwork at Sigatoka was conducted with permission or permit from the National Trust for Fiji, Fiji Museum, Fiji Ministry of Immigration and Mr Chris Work, Kulukulu village, who owns a large segment of the eastern dune end. Simon Fraser University International Programs and the Social Sciences Humanities and Research Council of Canada provided funding support to Burley. To each of the preceding we are indebted.

Received: 5 August 2009; Revised: 28 January 2010; Accepted: 5 March 2010

References

AKRIDGE, G.D. 2008. Methods for calculating brine evaporation rated during salt production. Journal of Archaeological Science 35: 1453-62.

ANDERSON, A. & G. CLARK. 2009. Fieldwork in southern Vit Levu and Beqa Island, in G. Clark & A. Anderson (ed.) The early prehistory of Fiji (Terra Australia 31): 87-120. Canberra: ANU E-Press.

ANDREWS, A.E 1983. Maya salt production and trade. Tucson (AZ): University of Arizona Press.

BEST, S. 1989. The Sigatoka dune burials. Unpublished manuscript on file, Fiji Museum, Suva.

DE BIRAN, A. 2001. The Holocene gemorphic evolution of the Sigatoka delta, Viti Levu, Fiji. Unpublished PhD dissertation, University of the South Pacific.

BIRKS, L. 1973. Archaeological excavations at Sigatoka Sand Dune site, Fiji (Fiji Museum Bulletin 1). Suva: Fiji Museum.

BURLEY, D.V. 2003. Dynamic landscapes and episodic occupations: archaeological interpretations and implications in the prehistory of the Sigatoka Sand Dunes, in C. Sand (ed.) Pacific archaeology: assessments and prospects (Cahiers de l'Archeologie en Nouvelle Caledonie 11): 327-35. Noumea: Departement Archeologie, Service des Musees et du Patrimoine.

--2005. Mid-sequence archaeology at the Sigatoka Sand Dunes with interpretive implications for Fijian and Oceanic culture history. Asian Perspectives 44: 320-48.

BURLEY, D.V. & K. TACHL 2008. Archaeology of the Sigatoka Sand Dunes National Park and adjacent area, report on the 2006 field season. Unpublished manuscript on file, Department of Archaeology, Simon Fraser University.

CHIANG, T.C. 1976. The production of salt in China, 1644-1911. Annals of the Association of American Geographers 66: 516-30.

CLARK, G.R. 1999. Post-Lapita Fiji: cultural transformation in the mid-sequence. Unpublished PhD dissertation, The Australian National University.

CIO, K, J.E. 2007. Craft specialization's penumbra: things, persons, action, value and surplus, in R.K. Flad & Z.X. Hruby (ed.) Rethinking craft specialization in complex societies: archaeological analysis of the social meaning of production (Archaeological papers of the American Anthropological Association 17): 20-35. Arlington (VA): American Anthropological Association.

CONNAH, G. 1991. The salt of Bunyoro: seeking the origins of an African kingdom. Antiquity 65: 479-94.

COSTIN, C.L. 1998. Introduction: craft and social identity, in C.L. Costin & R.P. Wright (ed.) Craft and social identity (Archaeological papers of the American Anthropological Association 8): 3-16. Arlington (VA): American Anthropological Association.

COSTIN, C.L. & R.P. WRIGHT (ed.) 1998. Craft and social identity (Archaeological papers of the American Anthropological Association 8). Arlington (VA): American Anthropological Association.

DICKINSON, W.R., D.V. BURLEY, P.D. NUNN, A. ANDERSON, G. HOPE, A. DE BIRAN & S. MATARARABA. 1998. Geomorphic and archaeological landscapes of the Sigatoka Sand Dune site, Viti Levu, Fiji. Asian Perspectives 37: 1-31.

EVANS, R.K. 1978. Early craft specialization: an example from the Balkan Chalcolithic, in C.L. Redman, W.T. Langhome Jr., M.J. Berman, N.M. Versaggi, E.V. Curtin & K.J.C. Wanser (ed.) Social archaeology: beyond subsistence and dating:. 113-29. New York: Academic Press.

FIELD, J.S. 2003. The evolution of competition and cooperation in Fijian prehistory: archaeological research in the Sigatoka valley, Fiji. Unpublished PhD dissertation, University of Hawaii.

FLAD, R.K., J. ZHU, C. WANG, P. CHEN, L. VON FALKKENAUSEN, Z. SUN & S. LEE. 2005. Archaeological and chemical evidence for early salt production in China. Proceedings of the National Academy of Sciences 102: 12618-22.

FLAD, R.K. & Z.X. HRUBY (ed.). 2007a. Rethinking crafi specialization in complex societies: archaeological analysis of the social meaning of production (Archaeological papers of the American Anthropological Association 17). Arlington (VA): American Anthropological Association.

FLAD, R.K. & Z.X. HRUBY. 2007b. 'Specialized' production in archaeological contexts: rethinking specialization, the social value of products, and the practice of production, in R.K. Flad & Z.X. Hruby (ed.) Rethinking craft specialization in complex societies: archaeological analysis of the social meaning of production (Archaeological papers of the American Anthropological Association 17): 1-19. Arlington (VA): American Anthropological Association.

FROST, E.L. 1979. Fiji, in J. Jennings (ed.) The prehistory of Polynesia: 61-81. Cambridge (MA): Harvard University Press.

HARRIS, N.V. 1963. Rainfall patterns in Fiji. New Zealand Geographer 19: 25-45.

HIRTH, K. (ed.) 2009a. Housework, craft production and domestic economy in ancient Mesoamerica (Archaeological papers of the American Anthropological Association 19). Arlington (VA): American Anthropological Association.

--2009b. Housework and domestic craft production: an introduction, in K. Hirth (ed.) Housework, craft production and domestic economy in ancient Mesoamerica (Archaeological papers of the American Anthropological Association 19): 1-12. Arlington (VA): American Anthropological Association.

DE LEON, J.P.. 2009. Rethinking the organization of Aztec salt production: a domestic perspective, in K. Hirth (ed.) Housework, craft production and domestic economy in ancient Mesoamerica (Archaeological papers of the American Anthropological Association 19): 45-57. Arlington (VA): American Anthropological Association.

LOVEJOY, P.E. 2003. Salt of the desert sun: a history of salt production and trade in the Central Sudan (African studies series 46). Cambridge: Cambridge University Press.

MARSHALL, Y., A. CROSBY, S. MATARARABA & S. WOOD. 2000. Sigatoka: the shifting sands of Fijian prehistory (University of Southampton Department of Archaeology Monograph 1). Oxford: Oxbow.

McKILLOP, H. 1995. Underwater archaeology, salt production and coastal Maya trade at Stingray Lagoon, Belize. Latin American Antiquity 6: 214-28.

MORRIS, E. 2001. Salt production and distribution, in T. Lane & E. Morris (ed.) A millennium of saltmaking: prehistoric and Romano-British Salt production in the Fenland (Lincolnshire archaeology and heritage report series 4): 389-404. Heckington: Heritage Trust of Lincolnshire, Fenland Management Project.

MULLER, J. 1984. Mississippian specialization and salt. American Antiquity 49: 489-507.

PARRY, J.T. 1997. The north coast of Viti Levu Ba to Ra. Air photo archaeology and ethnohistory (Bulletin of the Fiji Museum 10). Suva: Fiji Museum.

STEIN, G.J. 1996. Producers, patrons, and prestige: craft specialists and emergent elites in Mesopotamia from 5500-3100 BC, in B. Wailes (ed.) Craft specialization and social evolution: in memory of V. Gordon Childe:. 25-38. Philadelphia (PA): University of Pennsylvania, Museum of Archaeology and Anthropology.

TANNER, A. 1996. Colo Navosa: local history and the construction of region in the western interior of Vitilevu, Fiji. Oceania 66:230-51.

THOMSON, B.H. 1908. The Fijians: a study of the decay of custom. London: Heinemann.

TONGANIVALU, D. 1917. Fijian property and gear. Transactions of the Fijian Society 1: 1-18.

WILLIAMS, T. 1858. Fiji and the Fijians, Volume 1. London: Alexander Heyland.

David V. Burley (1), Karine Tache (2), Margaret Purser (3) & Ratu Jone Balenaivalu (4)

(1) Department of Archaeology, Simon Fraser University, Burnaby, BC, V5A 1S6, Canada (Email: burley@sfu.ca)

(2) Department of Anthropology, Universite de Montreal, C.P. 6128, Centre-ville, Montreal, QC H3C3J7, Canada

(3) Department of Anthropology, Sonoma State University, 1801 East Cotati Ave, Rohnert Park, CA 94928-3609, USA

(4) Fiji Museum, P.O. Box 2023, Government Buildings, Suva, Fiji Islands