A British naval surgeon named J. Linton Palmer came ashore on Easter Island in 1868 and counted roughly a hundred people living among hundreds of toppled stone giants. He wrote down what he saw. Statues "thrown down in all directions and were all more or less mutilated," an island that had clearly held something larger and no longer did.
That image did a century and a half of work. It became the ecocide parable: a people who cut down their last tree, starved, turned on each other, and left their monuments face down in the dirt as a warning to the rest of us. Jared Diamond gave it a wide audience in Collapse in 2005. It is a very good story. It has a moral, a mechanism and a body count, and it fits on a single slide.
Michael Marshall's piece in the September issue of Scientific American walks through two decades of evidence that the story was assembled backwards. Palmer was not looking at the aftermath of a society that destroyed itself. He was looking at the aftermath of smallpox, tuberculosis, slave raiding, annexation and sheep.
The collapse narrative is not one claim. It is a chain of them, and they do not all fail together. Some of it is solid. Most of the load bearing parts are not.
This part is not in dispute and never was. Pollen cores and paleoecology confirm that Rapa Nui lost its forest and that the endemic palm disappeared. Everything that follows is an argument about what that meant, not whether it happened.
Carl Lipo and Terry Hunt have spent years building a case, synthesized in 2025, that the Polynesian rat did a great deal of the work. Rats arrived with the settlers (nobody is sure whether as stowaways or as provisions), they live in trees, and they eat palm nuts. Lipo's phrase is "rat candy." The oldest deposits at Anakena hold hundreds of rat bones per square meter, and preserved palm nuts are chewed through to the seed.
What the story inflatedA palm that takes 70 years to fruit, a rat population with no predators eating the seeds, and people clearing plots for food. The forest had no path back. Framing that as a civilization too foolish to notice its last tree requires ignoring most of the mechanism.
The opposite, more or less. The Rapanui responded to thin soil and unreliable rain by engineering their way through it. They practiced lithic mulching, working stones into the soil so that phosphorus and potassium leached out, erosion slowed, and evaporation dropped. Mara Mulrooney and colleagues measured the result: rock garden soils carry more nutrients than the soil around them. Stone walled enclosures called manavai sheltered crops from the wind.
Annette Kühlem's team excavated Ava Ranga Uka a Toroke Hau between 2008 and 2024 and found dams, basins and underground aqueducts moving water from the Terevaka crater toward the southern coast. She calls it advanced hydraulic engineering. A prolonged drought starting around 1550 is exactly the sort of thing such a system exists for. They also found a planting pit cut into bedrock, filled with garden soil, holding a palm, next to an ahu. These people were planting trees, not just felling them.
What the story inflatedKühlem puts it plainly. The ecocide hypothesis is "a story of people being too stupid to realize what they were doing." The sites show mindful planning under bad conditions.
Statues left at the quarry or abandoned on roads lean noticeably forward, which is a strange design choice until you try to move one. Lipo and Hunt walked a 4.35 ton replica 100 meters in 40 minutes using ropes and 18 people, roughly the way you shuffle a refrigerator across a kitchen. The upright base was carved later, at the destination. "The statues were carved to move," Lipo says. Around 600 of them made it out of the quarry.
What the story inflatedThe timber budget. If the statues walked, the rollers and sledges that made statue building look like an act of ecological self harm were never needed. That said, the 2025 paper is subtitled as a response to critics, which tells you the transport question is argued rather than closed.
J. Víctor Moreno-Mayar and colleagues sequenced ancient DNA from 15 individuals held in Paris, radiocarbon dated between 1670 and 1950, and published in Nature in 2024. A crash severe enough to define the island's history should leave a scar in genetic diversity. There is no scar. Their models show slow growth over centuries and a population that never got very large.
The 2024 satellite and machine learning survey of the rock gardens points the same direction: far less cultivated land than earlier estimates, enough for a few thousand people, not the tens of thousands the collapse math requires.
As for the violence, Moreno-Mayar notes there is no archaeological evidence of fortification, which is conspicuous on an island whose people were demonstrably very good at stonework. No mass graves. No pile of skull fractures or blade marks. Not much weaponry. Nobody claims Rapa Nui was uniquely peaceful (Kühlem doubts it was), but intertribal conflict is not the same thing as societal breakdown.
What the story inflatedCannibalism, which was the detail everyone remembered, rests on the thinnest evidence of anything in the chain.
Jacob Roggeveen's fleet sighted the island on Easter Sunday, 1722, and his sailors shot 10 to 12 Rapanui within days of landing. His journal describes good soil, decent weather and cultivated food. Cook arrived 52 years later and found an island with almost nothing to offer and few people on it.
Then it got worse. Smallpox and tuberculosis. Peruvian blackbirding raids in 1862 and 1863 that carried off, as Kühlem notes, precisely the people who went out to greet visitors, meaning the knowledge holders. Chilean annexation in 1888, then a sheep ranching lease that ran until 1953 and confined the Rapanui to Hanga Roa, which they could leave only with a permission slip. "The island was totally scoured by sheep."
That is the landscape twentieth century archaeologists walked into and read as evidence of indigenous self destruction.
The revision is not unanimous, and the piece says so. The 2024 rock garden mapping is contested: critics argue satellite methods can miss upland cultivation or gardens that left fainter traces, and the population ceiling it implies depends on that. The walking moai hypothesis has published opponents. And the aDNA result rests on 15 genomes, all of them dating to after the collapse was supposed to have happened, which makes the absence of a bottleneck an inference from surviving diversity rather than a direct look at the period in question.
None of that rehabilitates the ecocide story. It does mean the replacement is a working consensus rather than a settled one, and the honest version of this piece says so out loud.
What I keep circling is not the archaeology. It is the twenty year lag.
Mulrooney was in the audience at Auckland in 2006 when Diamond wondered aloud what went through the mind of the Rapanui who cut down the last tree. She went and looked at the evidence behind the line, and the more she read, the less it held. That was two decades ago. The parable kept getting taught anyway, because it was a superb parable and because the alternative (a small society that managed a hard island competently for six centuries and was then destroyed by outsiders) does not deliver a moral to a general audience nearly as cleanly.
The failure mode here is not that anyone lied. It is that a story became so useful for making a point that nobody with an incentive to keep making the point went back to audit it. Palmer's ruined island was real. The explanation was supplied later, by people who needed one, and it stuck because it was satisfying.
I recognize this pattern from my own field. Medicine runs on teaching stories, and the good ones outlive their evidence by decades. The tell is always the same: a claim that everyone repeats, nobody cites, and which arrives with a moral attached. Morals are load bearing in a way that data is not. They are also much harder to retract.
The Rapanui did not destroy their island. They engineered their way through 600 years on the least forgiving inhabited rock in the Pacific, and then people arrived by ship and took it apart in a century.
The version we all learned was not a finding. It was a moral looking for a case study, and Easter Island was the right shape.
Sources
Primary: Michael Marshall, "What really happened on Easter Island?" Scientific American, September 2026 issue, published August 18, 2026. scientificamerican.com
Ancient DNA: Moreno-Mayar et al., "Ancient Rapanui genomes reveal resilience and pre-European contact with the Americas," Nature, 2024. doi.org/10.1038/s41586-024-07881-4
Rock gardens and population: DiNapoli et al., "Island-wide characterization of agricultural production challenges the demographic collapse hypothesis for Rapa Nui," Science Advances, 2024. doi.org/10.1126/sciadv.ado1459
Rock garden soil nutrients: Mulrooney et al., Archaeology in Oceania. doi.org/10.1002/j.1834-4453.2010.tb00082.x
Rats and deforestation: Lipo and Hunt, Journal of Archaeological Science, 2025. doi.org/10.1016/j.jas.2025.106388
Walking moai: Lipo and Hunt, "The walking moai hypothesis: archaeological evidence, experimental validation, and response to critics," Journal of Archaeological Science, 2025. sciencedirect.com
Hydraulic engineering at Ava Ranga Uka a Toroke Hau: Kühlem et al. doi.org/10.1007/978-3-030-91127-0_19
Deforestation record: Quaternary Science Reviews, 2020. doi.org/10.1016/j.quascirev.2020.106751
Drought beginning c. 1550: Quaternary Science Reviews, 2016. doi.org/10.1016/j.quascirev.2016.08.015
Settlement chronology: Hunt and Lipo, Science, 2006. doi.org/10.1126/science.1121879