The deepest hole ever drilled, on the Kola Peninsula in Russia, reached a little over 12 kilometers before the heat stopped the drill. That is less than one percent of the way to the center of the Earth, about 6,370 kilometers down. Everything we know about the deep interior comes from indirect evidence, above all from earthquakes.

Seeing with earthquake waves

Earthquakes send waves through the whole planet, and seismometers around the world record when they arrive. Two kinds matter most. P waves push and pull the rock as they go and can travel through solids and liquids. S waves shake the rock from side to side and can travel only through solids. Early in the twentieth century seismologists noticed a large shadow on the far side of the Earth from any big quake where S waves never arrived. Because S waves cannot pass through the outer core, scientists concluded that the outer core is liquid. In 1936 the Danish seismologist Inge Lehmann studied faint P waves and showed that inside the liquid core sits a solid inner core.

The picture that emerged has four main layers. A thin rocky crust; a thick mantle of hot rock that is solid but can flow very slowly over millions of years; a liquid outer core of iron and nickel, whose churning produces Earth's magnetic field; and a solid inner core. For plate tectonics, a different division matters. The lithosphere is the rigid outer shell, made of the crust and the uppermost mantle, and it is broken into plates. Below it lies the asthenosphere, a hotter, softer layer of mantle on which the plates move. Continental crust is thick and relatively light; oceanic crust is thinner and made of denser basalt.

Wegener's bold idea

In 1912 the German meteorologist Alfred Wegener argued that the continents had once been joined in a single landmass and had since drifted apart. He laid out his case in The Origin of Continents and Oceans (1915). The coastlines of South America and Africa fit together, and Wegener argued that the rocks and fossils matched too. Comparing the continents, he wrote, was "just as if we were to refit the torn pieces of a newspaper by matching their edges and then check whether the lines of print run smoothly across." His evidence included fossils of Mesosaurus, a small freshwater reptile, found in both South America and southern Africa, an animal that could hardly have swum an ocean. He also pointed to fossils of the fern-like plant Glossopteris across the southern continents, and to scratches left by ancient glaciers in places that are now tropical. The idea was widely rejected for decades, partly because Wegener could not explain what force moved the continents. Wegener died on an expedition across the Greenland ice sheet in 1930.

Mapping the hidden ocean floor

The breakthrough came from the sea. After the Second World War, ships used echo sounders to measure the depth of the ocean. At Columbia University's Lamont Geological Observatory, the geologist and cartographer Marie Tharp turned thousands of these soundings into maps, working with Bruce Heezen. Tharp found a rift valley running along the crest of the Mid-Atlantic Ridge, a sign that the seafloor there was being pulled apart. Heezen at first dismissed her finding as idle talk; within a few years it was central to his own work. Their maps revealed a system of ridges about 65,000 kilometers long winding through every ocean, the longest mountain chain on Earth.

Seafloor spreading

In 1962 the American geologist Harry Hess proposed that new seafloor forms at the ridges, where hot mantle rises and melts, and then moves away to either side like two conveyor belts. He modestly described his paper as something closer to poetry about the Earth than proven fact. The evidence came quickly, including magnetic patterns in the seafloor rocks. Dating confirmed it: the seafloor is youngest at the mid-ocean ridges and grows older with distance from them. It also explained a puzzle. Continents hold rocks about four billion years old, but almost no ocean floor is older than about 200 million years, because old seafloor sinks back into the mantle at deep-sea trenches.

Where plates meet

By the late 1960s these ideas had merged into the theory of plate tectonics. Plates meet in three ways. At divergent boundaries they pull apart, as at mid-ocean ridges and the East African Rift. At convergent boundaries they come together. Where oceanic crust meets a continent, the denser ocean plate sinks, or subducts, forming a trench and a chain of volcanoes such as the Andes. Where two continents meet, neither sinks easily, and the crust crumples upward. The Himalaya formed where India collided with the rest of Asia, and is still rising. At transform boundaries, plates slide past each other sideways, as along the San Andreas Fault in California. Plates move a few centimeters a year, roughly as fast as fingernails grow. What drives them? Heat flowing out of the mantle plays a part, but slab pull, the weight of old, cold seafloor sinking at trenches and dragging the rest of its plate behind it, is thought to be the strongest force.

Worked example: Iceland

Iceland is one of the few places where a mid-ocean ridge rises above the sea. At Thingvellir, the site of Iceland's ancient parliament, visitors can walk along a rift valley between cliffs that mark the edges of the North American and Eurasian plates. The valley widens by about two centimeters a year. Fresh lava erupts along cracks parallel to the rift, so rocks are youngest near the center of the island and older toward the east and west coasts, just as on the seafloor. Iceland also sits above an unusually hot plume of mantle, which is why it stands above the waves at all.