On the Isle of Pines in New Caledonia, low mounds on a mineral-rich plateau have prompted explanations ranging from human construction to natural processes and the nesting behaviour of an extinct giant bird. The mounds are real. Their interpretation is harder: similar-looking structures elsewhere in the archipelago do not necessarily share the same age, material or purpose.
Approximate area reference: the map uses rounded coordinates for the Isle of Pines, not a surveyed mound or excavation. Coastal burial mounds and the central plateau structures occupy different settings. No ordinary public access to particular mounds has been verified for this article; the island reference is not visitor permission or a route to an archaeological site.
What needs explaining
The central plateau mounds include accumulations of chromite and iron-rich gravel, some with hardened calcareous cores. Archaeologists have debated whether the cores were deliberately made or formed naturally. Calling them concrete can be misleading if it is understood to prove that people mixed and poured a modern-style material. That is one of the questions, rather than an established description of their manufacture.
A useful recent synthesis is Christophe Sand, Jacques Bolé and David Baret’s 2024 archaeological chapter, especially its regional comparison on printed pages 331–335. It distinguishes the plateau examples from mounds of sand and coral material along the coast. Human burials have been found in some coastal mounds; that does not mean that every plateau mound has yielded a skeleton.
How birds entered the argument
Roger Green’s 1988 paper, Those mysterious mounds are for the birds, records the intellectual history of the bird hypothesis. He traced an informal suggestion to discussions following Jack Golson’s 1959–1960 excavations: perhaps a bird related to living mound-building megapodes had created the structures. At first, researchers lacked a suitable identified bird from New Caledonia.
Green explained how later discoveries changed that position. Bones of Sylviornis neocaledoniae were first described as those of a ratite, then reassigned by researchers to the megapodes. A giant local bird therefore appeared to supply the missing candidate. In 1988 Green presented the proposal seriously, while retaining difficulties such as the cores and features that excavators had interpreted as postholes.
The analogy had a concrete biological basis. Some living megapodes incubate eggs using heat from large heaps of organic material. Green outlined an argument in which accumulated vegetation, microbial activity and repeated use could account for mound growth and hardened internal features. His paper records that argument; it is not a direct observation of an extinct bird making an Isle of Pines mound.
The anatomical test
A later study tested a central assumption. Trevor Worthy and colleagues’ 2016 PLOS ONE investigation of Sylviornis examined roughly 600 fossil bones. It placed the bird outside the true megapodes and found that its hindlimbs lacked the specialised digging adaptations expected in a giant mound-building species. Its claws were closer in form to those of an ordinary ground bird than to the tools required for such extensive excavation.
The authors concluded that Sylviornis was not a plausible builder of the tumuli. Their discussion also noted the absence of supporting associations such as eggshells within the mounds. This substantially weakens the once-popular giant-bird explanation. It does not identify the origin of every mound by elimination. Other extinct birds and non-bird processes are separate propositions that need their own evidence.
The PLOS discussion favoured a vegetation-and-erosion explanation for the tumuli. The anatomical analysis and that broader landscape interpretation should be kept distinct: the fossil evidence directly addresses the animal’s capabilities, while mound formation requires archaeological and geological investigation as well.
The archaeological alternative
Sand, Bolé and Baret favour a human origin for an important group of mounds and suggest that burial practices may link the plateau structures with the better-understood coastal examples. They discuss confirmed burials at Tü’ü and a dated human bone from a coastal mound at Gadji. These finds provide firmer evidence of funerary use at those sites than the shape of a mound alone could supply.
Their extension of that interpretation to the plateau remains a hypothesis. They propose that acidic iron-rich material and drainage could account for the absence of surviving human bone in upland structures. An explanation for missing remains is not the same as finding those remains. The distinction matters when assessing how far the coastal evidence can be carried into a different soil environment.
The chapter also discusses mapped paths or stone alignments around some plateau mounds. Such arrangements are relevant to a cultural interpretation, but the authors still call for a focused multidisciplinary research programme. Their conclusion is a reasoned archaeological proposal rather than a declaration that every structure has been excavated and explained.
Why a regional comparison helps
Much of the 2024 chapter reports the authors’ own fieldwork at Tivoli on Lifou. There they recorded several forms of mound and partially excavated four. Two structures were interpreted as ovens for cooking Cordyline, with local names and relatively recent dates; other, larger mounds prompted a cemetery interpretation. Lifou is a comparison, not another name for the Isle of Pines or direct excavation evidence from its plateau.
The same review notes other uses and origins across New Caledonia: commemorative structures associated with exchange or alliances, and modern heaps produced by mechanical vegetation clearance. Those examples caution against a universal answer. A mound can be natural, funerary, commemorative or practical without all neighbouring or superficially similar structures having the same explanation.
Dates and the limits of the source trail
Early radiocarbon results have sometimes been used to suggest extraordinarily ancient builders. The 2024 review questions that reading because the origin of the dated core material is uncertain and land-snail carbon can complicate interpretation. A measurement on material inside a mound is not automatically a reliable date for the moment someone built it. This record therefore assigns no invented construction year.
The sources reviewed here include Green’s complete 1988 argument, the relevant abstract and discussion of the 2016 anatomical study, and the 2024 chapter’s fieldwork and regional comparison. Golson’s original excavation report and Louis Lagarde’s 2020 reassessment have not been independently read in full for this article. Their findings are described only where explicitly attributed through the later archaeological synthesis.
Assessment
The outcome is Disputed. The well-known Sylviornis nesting explanation has been strongly challenged by anatomical evidence. Published natural-process and human-burial interpretations still differ, and the evidence is not uniform across coastal and plateau sites. The most useful remaining question is which explanation fits each class of mound, supported by secure excavation, dating and environmental analysis.
The geographical reference for the Isle of Pines places this case in New Caledonia’s South Province. Its archaeological setting should be respected as part of a living Kanak landscape. A gap in present knowledge does not establish an unknown vanished civilisation or any paranormal construction method.