How did people reach the Pacific islands?

Settlement ran from west to east, against the trade winds that prevail today, which makes accidental drift an unlikely mechanism and a planned upwind exploration a sensible one: sail out against the wind, and a failed search has a fast route home.

That strategy is a proposal, argued from simulations and from sailing, and it has a rival. A 2014 reconstruction of the winds of the time found that the routes to New Zealand and Easter Island lay downwind when those islands were first settled, so the ability to sail to windward may not have been essential.

Settlement meant carrying enough people to found a population. Historians put each founding party at a family group of 30 to 200, which is not what a fishing canoe blown off course contains.

Experimental voyages in reconstructed canoes, navigating by stars, swells and birds without instruments, have made the crossing between Hawaii and Tahiti repeatedly since 1976. The first sailing trials showed the canoes could hold a course about 75 degrees off the wind, which is what makes an upwind search possible at all.

A 2011 study redated the settlement of East Polynesia from radiocarbon dates on reliable, short-lived material, and set aside dates from materials with large sources of error.

The result compressed the chronology sharply: the Society Islands were settled around 1025 to 1120 AD, four centuries later than had been assumed, and the remaining archipelagos around 1190 to 1290.

Later work has pushed the first steps back. Lake sediments on Atiu, in the Cook Islands, record people or their pigs from about 900 AD, which their authors read as exploration and return voyages a century or two before settlement. A 2021 genomic study supports the short chronology for most islands, with earlier dates for the Cook Islands and the Marquesas, and rejects the older long chronology, which put settlement nearly a thousand years earlier.

On the short chronology, the last push, to every remaining island group in the largest ocean on the planet, took about a century.

A 2020 study analysed genomes from more than eight hundred individuals across Polynesia and the Pacific coast of the Americas, and found Native American ancestry in several eastern Polynesian populations.

It dates the mixing to around 1200 AD, and finds the closest match on the American side in Indigenous groups of present-day Colombia.

Easter Island had long been assumed to be the point of contact. The study’s authors conclude that the contact came before Easter Island was settled, on an island settled earlier that they could not identify.

A 2024 study of fifteen Rapanui who lived before the slave raids of the 1860s found about 10 per cent Native American ancestry in every one. It dates the mixing to about 1250 to 1430 AD, a century or two later than on the other islands, and finds the closest match in the Central Andes. Its authors read that as a sign that there may have been more than one contact.

The genetics shows contact. It does not show who sailed. Polynesians reaching South America and returning with Native Americans, or South Americans reaching Polynesia, are both consistent with the genomes, and whether it happened once or more than once is disputed.

The sweet potato, an American plant found in the Polynesian archaeological record, has long been cited as evidence of contact, and a 2021 study notes that most scholars of the subject put it down to people crossing the Pacific. Some argue it need not have been carried by people at all: a 2018 study concluded it reached Polynesia before people did, while the authors of the 2020 genetic study suggest their contact could explain its spread by people.

The last islands were reached around 1200 to 1300 AD. Voyaging between archipelagos went on long after that: stone adzes in the Cook Islands came from as far as the Marquesas, about 2,400 kilometres away, over some three centuries from the early 1300s, and an adze found in the Tuamotus shares its chemistry with stone from Kahoʻolawe, in Hawaii. In the Cook Islands the exchange may have outlasted that in other archipelagos by a century or more. When and why the long-distance traffic ended is less clear.

Climate is one candidate: wind reconstructions suggest conditions for sailing to New Zealand lasted until about 1260, with return voyages possible at intervals into the early 1600s. Whether changing winds ended the traffic is not established.