A small island off the coast of Fiji’s Vanua Levu appears to be almost entirely composed of discarded shellfish remains, accumulated over roughly 1,200 years of human activity. A peer-reviewed study published in the journal Geoarchaeology presents evidence that the site, located near the village of Culasawani, is a massive shell midden rather than a natural geological formation. The finding challenges assumptions about how early Pacific communities interacted with their coastal environments and raises questions about whether such waste accumulation was incidental or served a functional purpose.
What the excavation revealed
The research team, led by geoscientist Patrick D. Nunn, excavated the island using four 1-by-1-meter test pits. What they found was striking: a deposit roughly 20 to 40 centimeters thick, composed of an estimated 70 to 90 percent shell. The island itself spans approximately 3,000 square meters and sits just 20 to 60 centimeters above mean high tide, making it barely visible above the waterline. The deposit was largely unstratified, meaning the layers of shell showed little vertical differentiation, a detail that carries weight for interpreting how the material accumulated.
Radiocarbon dating places the shell deposit at approximately 1,200 years old, situating the island’s formation well within the period of established human settlement across the Fijian archipelago. Pottery fragments were also recovered from the site, adding another line of evidence for sustained human presence rather than a single catastrophic natural event. Together, the physical setting and the artifacts suggest a long-lived coastal activity zone rather than a transient campsite or a one-off disaster deposit.
Why the midden explanation holds up
The paper’s title poses the question directly: “Midden or Muddle?” Two competing explanations exist for how a small island could end up packed with shells. One possibility is a natural process, such as a tsunami or storm surge, that deposited marine debris in a concentrated area. The other is that generations of people discarded shells from their meals in the same location over centuries, gradually building up the landform through everyday behavior.
The evidence tilts decisively toward the midden explanation. The shell assemblage consists almost exclusively of edible species, according to the study’s analysis. A tsunami or storm would be expected to sweep up shells indiscriminately, mixing edible and inedible species, coral fragments, and other marine debris. The selective composition of the Culasawani deposit points instead to deliberate human discard. The absence of typical high-energy storm markers, such as coarse mixed gravels or broken coral rubble, further weakens the case for a purely natural origin.
Combined with the pottery fragments and the unstratified nature of the deposit, the picture that emerges is one of long-term, repeated food processing at a single location. If people were regularly gathering shellfish from nearby reefs and mudflats, then cleaning, cooking, and discarding them in one favored spot, the slow accumulation of waste could have created a raised platform that only later came to be recognized as an “island.” In this interpretation, the landform is not a passive stage for human activity but a direct product of it.
This distinction matters beyond academic taxonomy. If the island is indeed a midden, it represents one of the more dramatic examples of how routine human activity can reshape coastal geography. People eating shellfish and tossing the remains in the same spot, year after year, effectively built new land. That process underscores how even small-scale subsistence practices can leave large, enduring imprints on shorelines, with implications for how archaeologists read other low-lying islands and shoals across the Pacific.
What is verified so far
The core findings rest on solid institutional ground. The study was authored by researchers affiliated with the University of the Sunshine Coast and partner institutions including the University of the South Pacific, Fiji National University, the Fiji Museum, and University College Cork. The paper is published in Geoarchaeology, a peer-reviewed Wiley journal, and is available as an open-access PDF through the University of the Sunshine Coast’s institutional repository, allowing other researchers to scrutinize its methods and conclusions.
The physical measurements are internally consistent across the paper and secondary reporting. The island’s area, elevation, deposit thickness, and shell concentration are all quantified with specific ranges rather than vague estimates. The excavation design, while modest at four test pits, is appropriate for a preliminary geoarchaeological survey of a site this size and offers a representative cross-section of the deposit. The radiocarbon dating provides a temporal anchor, though the study’s exact calibration methods and lab protocols are not detailed in the available summaries and would need to be consulted directly in the full article for technical assessment.
Patrick D. Nunn, the lead author, has prior published work on related sites in the same region. An earlier paper in Geosciences documents his fieldwork on Vanua Levu and discusses a rediscovered historical formation known as the “Mbua Shell-Bed,” which has its own radiocarbon dates and interpretive debates. In that work, Nunn argued that some dates may predate conventional models of when humans first settled Fiji, a claim that remains contested but provides important context for the Culasawani research by highlighting how shell deposits can complicate regional settlement chronologies.
In the Culasawani case, however, the estimated age of around 1,200 years falls comfortably within widely accepted timelines for human occupation of Fiji. That makes the island less controversial as evidence for early settlement and more significant as a window into how already-established communities organized and intensified their coastal resource use. The study’s cautious framing (raising alternative explanations and then weighing them against the data) also lends credibility to its main conclusion that the island is best understood as a shell midden.
What remains uncertain
Several significant gaps exist in the available evidence. The radiocarbon age of approximately 1,200 years comes from secondary summaries rather than a detailed breakdown of lab methods, sample selection, or calibration curves in the public reporting. Without knowing which specific shells were dated, whether marine reservoir corrections were applied, and what calibration software was used, the precision of that age estimate cannot be independently evaluated from summaries alone. Marine shells can yield deceptively old dates if local reservoir effects are not carefully accounted for.
No direct statements from Fijian Indigenous communities or the Fiji Museum regarding the cultural significance of the site have been published alongside the study. The Fiji Museum is listed among the institutional affiliations, which suggests some level of local engagement, but the nature and depth of that collaboration is not described in the reporting. For a site that may reflect centuries of community food practices and possibly ancestral territories, the absence of Indigenous voices in the public discussion is a notable gap and limits how confidently outsiders can interpret the island’s social meaning.
The study also lacks comparative primary data on similar shell islands elsewhere in Fiji or the broader Pacific. Nunn’s earlier work on the Mbua Shell-Bed provides some regional context, but it addresses a different type of formation on land rather than a low-lying offshore feature. Whether other shell-dense islands exist in Fiji’s coastal waters and whether they share the Culasawani site’s characteristics are questions the current paper does not appear to answer. Without a wider sample, it is difficult to know if Culasawani is exceptional or part of a broader, underrecognized pattern of shell-built landforms.
The unstratified nature of the deposit introduces its own ambiguity. While the authors interpret it as consistent with a midden that built up steadily over time, an unstratified shell layer could also result from post-depositional mixing by wave action, bioturbation from burrowing organisms, or other disturbance processes that blur original layers. The paper apparently considers and rejects these alternatives based on the edible-species composition and the absence of mixed sediment types, but the strength of that argument depends on how thoroughly non-food shells and non-shell materials were identified and quantified.
Another open question concerns the island’s exact function within past coastal lifeways. A dense shell midden might reflect routine household refuse, specialized processing for trade or feasting, or even deliberate construction to stabilize a landing place or fish-processing station. The limited excavation area and the lack of reported structural remains or tools make it hard to distinguish among these possibilities. Further fieldwork (expanding the excavation grid, conducting detailed faunal analysis, and integrating oral histories) would be needed to clarify whether Culasawani’s shell island was simply a convenient dump or a carefully maintained feature of a broader coastal landscape.
For now, the Culasawani midden stands as a compelling case study in how accumulated everyday actions can literally build islands. It underscores the value of combining geoarchaeology, radiocarbon dating, and regional historical work to re-examine seemingly ordinary landforms. At the same time, the unresolved questions about dating precision, comparative sites, disturbance processes, and cultural interpretation highlight how much remains to be learned from this shell-built island and others like it that may still be hiding in plain sight along Pacific shores.
More from Morning Overview
*This article was researched with the help of AI, with human editors creating the final content.
