The Earth's seven continents may seem like terra firma, but geologists know the "firma" part doesn't prove out. The buoyant shelves of granite that tower miles above the seafloor have a protean past more like a slow-motion movie than a still life. The heat of the planet's molten innards, for example, propels movement of crustal plates over its surface. This process, called plate tectonics, has fractured landmasses over the eons and reassembled them in haphazard continental mosaics. Hence the territory that one geologist dubbed the United Plates of America is made up of no fewer than six separate geologic provinces.
"If you're standing in Idaho, you might be on a piece of crust that actually formed near Australia, and now it's been juxtaposed next to a piece of crust that formed somewhere near Russia," says Paul Sylvester of the Memorial University of Newfoundland.
There's also evidence that the continental crust can be recycled through Earth's interior. Geologists have come to wonder whether such recycling obliterated the planet's earliest landmasses. As they try to reconstruct the early architecture of Earth, a question looms: How much ancient earth was there on ancient Earth?
The short answer: not much. Although the planet congealed out of interstellar dust and gas some 4.5 billion years ago, the oldest known rock formations on Earth go back just 4 billion years. Infant Earth spun on its axis every 14 hours, and its atmosphere was a mix of carbon dioxide, nitrogen, methane, and sulfur. It might have had three layers: a dense core, a lighter mantle, and a thin, hardened crust. But much of that crust would have been destroyed by the roiling heat of the planet's interior, as well as by incoming asteroids that regularly vaporized whole oceans and raised temperatures as high as 1700 degrees Celsius for thousands of years.
It's no surprise, then, that geologists have found only a few sites more than 3.5 billion years old, and these areas are often less than 300 miles across.
"This period is full of lots of transitions and very little geologic record," says Samuel Bowring of the Massachusetts Institute of Technology, who found the oldest known formations in northwest Canada several years ago.
As Earth cooled and asteroid impacts became less frequent, the continental crust had a better chance of surviving. The crust forms when repeated melting of rock in the planet's mantle creates a lightweight magma that rises to Earth's surface. This lightweight rock "floats" atop the heavier crust in ocean basins. Where it ends up depends on the movement of the oceanic plates, which are shifted laterally by the subterranean churning of the mantle. As the oceanic crust dives into trenches, the continental crust collects above them. Eventually it piles up above sea level, becoming dry land.
Continental crust is on average five times thicker than oceanic crust, and it tends to be far more ancient. Whereas the oldest known oceanic plates are around 200 million years old, more than half of the continental crust that exists today dates to about 2.7 billion years ago, when the Earth experienced a burst of magmatic enthusiasm that's been unequaled since. A third of the crust was produced by lesser events at 1.9 billion and 1.2 billion years ago. The rest is even younger.
With this chronology quite literally etched in stone, geologists have come to believe that the continents formed slowly over time, adding a bit here, a bit there, in growth spurts of granite that waited nearly 2 billion years after the planet's conception to begin in earnest. But Bowring, Sylvester, and others have come to question that reasoning. They've taken their cue from the late geologist Richard Armstrong, who proposed in the 1960s that Earth produced plenty of lightweight crust in its infancy. Armstrong was inspired by the discovery that other terrestrial planets, such as Mars, had differentiated into core, mantle, and crust early in their history.
"The idea is that the planet is hottest right after it forms," says Sylvester. "And if you consider the planet as a sort of factory extracting this light material out of a big ball of dense material, you'd expect that to occur early on. That's when you'd expect this stuff to sweat out of the mantle. It never made sense that [the crust] would slowly dribble out over billions of years."
But if Earth was cranking out continental crust from the get-go, where has all the really old stuff gone? Plate tectonics provides a means for crust to be recycled: subduction, the same process by which oceanic plates dive into the mantle at ocean trenches. "But nobody had any idea of how you would get continents to be subducted down those trenches," says Sylvester. So Armstrong's theory was never taken seriously.
Several mechanisms for the subduction of continental crust have since emerged. Studies suggest that sediments eroding from the edges of continents onto oceanic plates get carried into the mantle in small doses. Subducting oceanic crust can also act like sandpaper, tearing off bits of overlying continental crust as it sinks. And when two continental plates collide, the undersides of the continents might detach and sink down into the mantle. Most geologists now agree that continental crust gets recycled they just don't agree on how much recycling occurred in the first 2.5 billion years of Earth's history.
In the past decade, geologists have traced the concentrations of certain elements in primitive samples of continental crust and mantle to determine how far back in time recycling may have occurred. When the crust melts out of the mantle, these elements get concentrated in it, leaving the mantle depleted. But if crust is recycled, the elements should be redistributed, leaving more complex chemical signatures in both mantle rocks and recycled crust. Based on such signatures in the oldest rock formations, some geologists claim that robust recycling was taking place at least 4 billion years ago.
That possibility is bolstered by the recent dating of hardy minerals called zircons from Western Australia. The minerals, lodged in newer sediments like crumbs in a carpet, are up to 4.4 billion years old.
"That suggests that we had some continental crust very early on," says Kent Condie, a geologist at the New Mexico Institute of Mining and Technology, "but it doesn't tell us how much we had. That's the problem."
The problem is complicated by recent findings that subducted plates may go much deeper in the mantle than previously suspected. That may prevent geologists from establishing reliable comparisons between primitive crust and mantle. But even as the debate over ancient continental crust heats up, the terrain itself is cooling. By now, 4.5 billion years after its birth, Earth is fueling far fewer major tectonic events. In island arcs like Japan and the Aleutians, new continental crust is still piling up, but at a geriatric pace. Like Venus and Mars, Earth is on its way to becoming a dead planet; the heyday of its continents is long gone.
"You just don't get as much production of crust anymore," Sylvester laments. "It's kind of sad."