What happened on Rapa Nui?

Pollen records show the island was covered in a palm forest and that palm pollen gives way to grasses after settlement. The invasive Pacific rat, carried by the settlers, is implicated alongside clearance for cultivation, and a review of the pollen record argues that drought played a part too.

That much is not in dispute. What is disputed is what followed.

The chronology was revised before the collapse story was. A 2006 study dated the earliest layers at Anakena, reassessed the older radiocarbon dates, and placed settlement around 1200 AD. Earlier accounts had put it as early as 800.

That shortens the whole sequence: the statues, the forest loss and the adaptation to a treeless island all fit into about five hundred years rather than as many as nine hundred.

It does not make the forest loss abrupt. The early pollen work read as a sudden island-wide replacement of palm by grass, and the continuous sediment cores taken since show clearance happening at different places at different times and different rates. That part of the original story is the part that has gone.

A 2021 study modelled radiocarbon dates from across the island together with palaeoenvironmental records, using a method that estimates population trajectories and their uncertainty rather than reading them off a curve.

It found no evidence of a pre-contact collapse. Population appears to have grown steadily until European contact, with land use shifting in response to the changing environment rather than failing with it.

Statue construction continued into the period when the collapse is supposed to have occurred, which is difficult to reconcile with a society in crisis.

That study and the redating both come from the same research group, which is worth saying. The independent check arrived in 2024, from ancient genomes: a separate team sequenced fifteen Rapanui individuals who lived after European contact and before the slave raids, then reconstructed effective population size, a genetic measure rather than a head count, backwards through the centuries the collapse is placed in. They reject a severe bottleneck in the 1600s. Two methods with nothing in common now point the same way.

One model of what the island could feed gives two figures for the peak population before contact: about 3,500 and about 17,500. What separates them is an assumption about how much nitrogen the soils made available.

Since the collapse story depends on a large population falling to a small one, the disagreement about the peak is not a detail. Europeans in 1722 found a few thousand people, so the high figure requires a crash and the low one does not.

The 2021 study reads its results as fitting the low figure. A 2024 satellite survey from the same group found about a fifth of the rock gardens the model had assumed, and its authors put what the island could then feed, fish and other food included, at about three thousand people.

Showing that the island did not empty itself is not the same as showing that nothing went wrong. Palaeoecological work continues to find environmental stress that the adaptation account has to absorb, including a drought during the Little Ice Age that one palaeoecologist proposes as a driver of changes in settlement and cultivation, without ruling out small falls in population. The 2021 radiocarbon study, drawn mostly from coastal sites, finds no decline in activity to match.

So the live argument is no longer collapse against no collapse. It is about how much strain a small island society absorbed, by what means, and which of the changes visible in the record are responses to climate rather than to what the settlers did to the forest.

It is a complete story with a moral, and it has been used as a small model of the whole planet overusing what it has.

The revised account is less useful rhetorically: a society adapting successfully to an environment it had altered, and then being devastated by outsiders. That it is less useful is not an argument against it.