When did the continents start moving?

Sea floor is created at mid-ocean ridges and consumed at subduction zones, so almost none of it is older than about 200 million years. Continental crust is lighter and resists being dragged down, so it lasts far longer, though not intact: less than 7 per cent of what survives is older than three billion years, and much of the early crust has been destroyed.

That is why this question has to be asked of continental rocks and of the minerals inside them rather than of the ocean basins, which have been recycled many times over.

The signatures used include the chemistry of magmas, which carries a distinctive mark where fluid from a sinking plate triggers melting above it; rocks metamorphosed at the low temperatures and high pressures typical of subduction; and the compositions of tiny zircon crystals that survive from vanished rocks.

Each is indirect, and each can be produced in more than one way. Studies also differ over what counts as plate tectonics in the first place, and together those are why the dates disagree rather than converge.

Plate motion itself is the most direct test, and it has split too. In 2023 one team read near-constant magnetic field strengths recorded in zircons from 3.9 to 3.4 billion years ago as unchanging latitudes, and so as a stagnant lid rather than moving plates. In 2024 a second team reanalysed the same data and concluded that they give no evidence either way.

A 2013 review argues the question from the physics and judges the rock evidence from there. Whether plates ran in the past, it says, depends on what kept the surface heat flux roughly constant through Earth's likely thermal history.

On that reckoning plate tectonics could have run continuously through Earth history if, for example, the slow subduction of surface water altered the strength of convective stress in the mantle. Others argue from models that plate tectonics may need a mantle within about 150 degrees of its present temperature, which would leave the early surface as a stagnant lid.

Either way, the geological indicators carry assumptions about how the mantle behaved, and the review finds those assumptions either demonstrably wrong or untested.

A growing view is that the question needs restating. On that view plate tectonics is unlikely to have begun as a single global event at a distinct time: it began locally and spread, so an onset date is a date for when it became the dominant regime rather than for when it first happened anywhere.

That is also why the signatures give different ages. Many are minimum ages taken from the oldest surviving example, and local subduction can be triggered by a meteorite impact or a mantle plume as well as by plates.

Subduction needs crust and lithosphere rigid enough to be subducted, which depends on how hot they are, on water, which helps the mantle convect and the plates move, and on how much heat the interior had to shed.

Since Earth is the only example, everything about the necessary conditions is inferred from one case, which is a poor basis for a general law and is the best anyone has.

Plate tectonics recycles carbon, helps regulate climate over long periods and builds continents, so whether other rocky planets have it may bear on what life could become there. One geologist argues that a planet with oceans, continents and plates favours the evolution of a technological species, though plates may not be needed for life to begin.

A 2026 review expects imaging an exoplanet's surface well enough to tell a planet with plates from one without to stay difficult for the foreseeable future, though its atmosphere might carry indirect signs. For now, plate tectonics is known on one planet only.