Plate tectonics might even be responsible for another atmospheric ingredient, and arguably the most important: oxygen.

A full 2 billion years before the Cambrian explosion, back in the Archean eon, Earth had hardly any of the air we breathe now. Algae had begun to use photosynthesis to produce oxygen, but much of that oxygen was consumed by iron-rich rocks that used the oxygen to make rust.

According to research published in 2016, plate tectonics then initiated a two-step process that led to higher oxygen levels. In the first step, subduction causes the Earth's mantle to change and produce two types of crust — oceanic and continental. The continental version has fewer iron-rich rocks and more quartz-rich rocks that don't pull oxygen out of the atmosphere.

Then over the next billion years — from 2.5 billion years ago to 1.5 billion years ago — rocks weathered down and pumped carbon dioxide into the air and oceans. The extra carbon dioxide would have aided algae, which then could make even more oxygen — enough to eventually spark the Cambrian explosion.

Plate tectonics may also have given life an evolutionary boost. Robert Stern, a geologist at the University of Texas, Dallas, thinks plate tectonics arose sometime in the Neoproterozoic era, between 1 billion and 540 million years ago. This would have coincided with a period of unusual global cooling around 700 million years ago, which geologists and paleoclimate experts refer to as "snowball Earth." In April, Stern and Nathaniel Miller of the University of Texas, Austin, published research suggesting that plate tectonics would have catastrophically redistributed the continents, disturbing the oceans and the atmosphere. And, Stern argues, this would have had major consequences for life.

"You need isolation and competition for evolution to really get going. If there is no real change in the land-sea area, there is no competitive drive and speciation," Stern said. "That's the plate tectonics pump. Once you get life, you can really make it evolve fast by breaking up continents and continental shelves and moving them to different latitudes and recombining them."

Stern has also argued that plate tectonics might be necessary for the evolution of advanced species. He reasons that dry land on continents is necessary for species to evolve the limbs and hands that allow them to grasp and manipulate objects, and that a planet with oceans, continents and plate tectonics maximizes opportunities for speciation and natural selection.

"I think you can get life without plate tectonics. I think we did. I don't think you can get us without plate tectonics," he said.

Stern imagines a far future in which orbiting telescopes can determine which exoplanets are rocky, and which ones have plate tectonics. Emissaries to distant star systems should aim for the ones without plate tectonics first, he said, the better to avoid spoiling the evolution of complex life on another world.

Cracking Earth's Shell

But everything depends on when the process started, and that's a big open question.

Earth formed about 4.54 billion years ago and started out as an incandescent ball of molten rock. It probably did not have plate tectonics in any recognizable form for at least 1 billion years after its formation, mostly because the newborn planet was too hot, said Craig O'Neill, a planetary scientist at Macquarie University in Australia.

Back then, as now, convection within the planet's inner layers would have moved heat and rock around. Rock in the mantle is squeezed and heated in the crucible of Earth's innards and then rises toward the surface, where it cools and becomes denser, only to sink and start the process again. Picture a lava lamp.

Through convection, vertical motion was happening even on the early Earth. But the mantle at that time was relatively thin and "runny," O'Neill said, and unable to generate the force necessary to break the solid crust.

"Subduction wasn't happening. There was no horizontal motion," Klepeis said. "So there was a time before continents, before the first continent formed" — the time before land, if you will. Earth would have had a so-called "stagnant lid," without disparate plates.

O'Neill published research in 2016 showing that early Earth might have been more like Jupiter's volcanic moon Io, "where you have a volcanically active regime, and not a lot of lateral motion," O'Neill said. As the planet began to cool, plates could more readily couple with the mantle below, causing the planet to transition into an era of plate tectonics.

This raises the question of what cracked the lid and created those plates in the first place.

Some researchers think an intrusion might have gotten things moving. In the past two years, several teams of researchers have proposed that asteroids left over from the birth of the solar system might have cracked Earth's lid. Last fall, O'Neill and colleagues published research suggesting that a bombardment of asteroids, half a billion years after Earth formed, could have started subduction by suddenly shoving the cold outer crust into the hot upper mantle. In 2016, Maruyama and colleagues argued that asteroids would have delivered water along with their impact energy, weakening rocks and enabling plate movement to start.

But it's possible Earth didn't need a helping hand. Its own cooling process may have broken the lid into pieces, like a cake baked in a too-hot oven.

Three billion years ago, Earth may have had short-lived plate tectonic activity in some regions, but it was not widespread yet. Eventually, cooler areas of crust would have been pulled downward, weakening the surrounding crust. As this happened repeatedly, the weak areas would have gradually degraded into plate boundaries. Eventually, they would have formed full tectonic plates driven by subduction, according to a 2014 paper in Nature by David Bercovici of Yale University and Yanick Ricard of the University of Lyon in France.

Or the opposite might have happened: Instead of cold crust pushing down, hot mantle plumes — like the kind that are driving Hawaii's eruptions — could have risen to the surface, percolating through the crust and melting it, breaking the lid apart. Stern and Scott Whattam of Korea University in Seoul showed how this could work in a 2015 study.

According to these theories, plate tectonics may have started and stopped several times before picking up momentum about 3 billion years ago. "If you had to press everyone's buttons and make them take a number, there's a running ballpark in the community that around 3 billion years ago, plate tectonics started emerging," O'Neill said.

Yet it's hard to know for sure because the evidence is so fragmentary.

"Oceanic crust is only 200 million years old. We're just missing the evidence that we need," O'Neill said. "There's a lot of geochemistry that's come a long way since the 1980s, but the same fundamental questions are still there."

The oldest rocks on Earth suggest that some sort of proto-subduction was happening as far back as 4 billion years ago, but these rocks are hard to interpret, O'Neill said. Meanwhile, sometime between 3 billion and 2 billion years ago, Earth's mantle apparently underwent several chemical changes that can be attributed to cooling, changing its convection pattern. Some geologists take this as a recording of the gradual onset and spread of tectonic plates throughout the planet.

"The real answer is we don't know," said Brad Foley, a geophysicist at Pennsylvania State University. "We've got these rocks, but we can't figure out what's the smoking gun that would tell us there is plate tectonics or subduction at this time, or there definitely wasn't."

Plates on Other Planets

So are tectonics essential to life?

Ultimately, the problem is that we have one sample. We have one planet that looks like Earth, one place with water and a slipping and sliding outer crust, one place teeming with life. Other planets or moons may have activity resembling tectonics, but it's not anything close to what we see on Earth.

Take Enceladus, a frozen moon of Saturn that is venting material into space from strange-looking fractures in its global ice crust. Or Venus, a planet that seems to have been resurfaced 500 million years ago but has no plates that we can discern. Or Mars, which has the solar system's largest volcano in Olympus Mons, but whose tectonic history is mysterious. Olympus Mons is found in a great bulging province called Tharsis, which is so gigantic that it might have weighed down Mars' crust enough to cause its poles to wander.

O'Neill has published research showing that a Mars-size planet with abundant water could be pushed into a tectonically active state. And others have argued that some regions in Mars' southern hemisphere resemble seafloor spreading. But researchers agree it hasn't had any action for at least 4 billion years, which is roughly the age of its crust, according to data from orbiters and robots on the surface.

"There is some argument that maybe very, very early on, it could have had plate tectonics, but my view is it probably never did," Foley said.