The Fit

Look at a world map and the coasts of South America and Africa look like two pieces of a torn page. People had been saying so for three hundred years.

Wegener did the thing that turns a resemblance into an argument. If the page was torn, then the lines of print should run across the tear.

So he checked. Fossil species, rock formations, mountain belts, coal seams, glacial scratches — four independent kinds of writing, and they all line up.

And he was dismissed for thirty years, because he could not say what moves a continent.

The textbook picture in 1910 was a cooling Earth. As the interior shrinks, the crust wrinkles, and the wrinkles are mountain ranges.

Continents stayed where they were made. The problem was that identical fossils kept turning up on continents separated by whole oceans.

The accepted answer was land bridges: enormous isthmuses that once connected the continents and then conveniently sank without trace.

That is the intellectual situation Wegener walked into: a theory that explained mountains and a patch that explained fossils, with nothing holding them together.

Wegener was not a geologist, which is part of why the geologists would not listen.

His doctorate was in astronomy, his working field was meteorology, and in 1906 he and his brother set a world record by staying up in a balloon for fifty-two and a half hours.

He spent four expeditions on the Greenland ice cap studying polar air, and he was genuinely good at it.

In 1911 he read a paper listing fossil species that appear on both sides of the Atlantic, went looking for more, and found that the pattern was everywhere.

On the sixth of January 1912 Wegener stood up in Frankfurt and told a room of geologists that the continents had once been a single mass and had since moved apart.

The coastline fit is suggestive but sloppy, because a coastline is just where the sea happens to stand today. The real edge of a continent is the outer edge of its shallow shelf, often a hundred kilometres further out.

Fit those edges instead and South America and Africa close to within about a degree, which is as good as the data allows.

Press the button and close the Atlantic.

Wegener’s picture was a single supercontinent he called Pangaea, surrounded by one ocean, which broke apart and whose fragments drifted to where we find them.

The timescale is the part people find hard. Pangaea assembled about three hundred and thirty-five million years ago and started to split about a hundred and seventy-five million years ago.

The Atlantic is still opening, at around two and a half centimetres a year. That is roughly how fast your fingernails grow.

Slow enough to be invisible in a human life, and fast enough to move a continent five thousand kilometres in a couple of hundred million years.

Here is why the fit is not a coincidence. Switch on the layers and watch what crosses the join.

Mesosaurus was a small freshwater reptile. Its fossils are in Brazil and in South Africa, and it could not swim an ocean.

Glossopteris was a seed fern whose seeds are far too heavy to blow across water, and it is found in South America, Africa, India, Australia and Antarctica.

The Appalachian mountains stop at the Atlantic and start again in Scotland and Norway, same age, same rocks.

And coal, which only forms in warm swamps, is mined in Spitsbergen, well inside the Arctic Circle.

The strongest evidence is the strangest. In India, central Africa, Brazil and Australia there are rocks scoured by a continental ice sheet about three hundred million years ago.

Those places are tropical. An ice sheet thick enough to grind bedrock could not have sat on them where they are now.

And the scratches have directions. On today’s map they point in incoherent directions, outward from nothing.

Put the continents back into Pangaea and all the scratches radiate neatly from a single ice cap over what is now southern Africa and Antarctica.

Wegener knew the weakness and said so. He had assembled the evidence that continents move, and he could not say what moves them.

The forces he suggested — the Earth’s rotation flinging continents towards the equator, and tidal drag from the Moon — are real, and they are far too feeble. Harold Jeffreys showed they were off by orders of magnitude.

That was fatal. Geologists concluded that if the proposed mechanism is impossible, the conclusion must be wrong too — which does not follow, but is how people reason.

At a 1928 symposium in New York the theory was taken apart. One senior figure called it utter, damned rot.

The mechanism existed. One year after Wegener died, the British geologist Arthur Holmes proposed that the mantle convects — slow currents of hot rock rising, spreading sideways, and sinking.

The crust rides on top of those currents like scum on a simmering pan. Continents are not ploughing through the ocean floor; the ocean floor is moving too.

It was essentially right. It was also untestable with the instruments of 1931, and it was treated as speculation.

The evidence that settled it did not come from the land at all.

Wegener spent his last years leading a scientific expedition to the Greenland ice cap, measuring its thickness and the weather above it.

In 1930 the team at the inland station, four hundred kilometres from the coast, had not received their supplies and would not survive the winter without them.

Wegener led a relief party through worsening weather and reached them on his fiftieth birthday. There was not enough food for everyone to stay, so he turned round with one companion and started back.

He died somewhere on the way, probably of heart failure. His companion buried him carefully in the snow and then also died. The body was found the following May, marked by a pair of skis set upright in the ice.

After the Second World War, ships with sonar mapped the sea floor properly for the first time, and found a mountain chain running down the middle of the Atlantic with a rift along its crest.

They also found something that should not be possible. The ocean floor is young. Nowhere on Earth is the sea bed older than about two hundred million years, while continental rocks go back four billion.

In 1962 Harry Hess drew the conclusion: new sea floor is being made at the ridges and pushed outwards, and old sea floor is being destroyed down the deep trenches.

The continents are not ploughing through the ocean. They are passengers.

The clinching evidence was magnetic. The Earth’s field reverses at irregular intervals, and lava freezing at a ridge locks in whichever direction the field had that day.

So the sea floor is a tape recording. Fred Vine and Drummond Matthews realised in 1963 that if the floor spreads, the recording must be symmetrical: the same sequence of stripes on both sides of the ridge, mirrored.

They went and looked, and it is.

There is no way to produce a mirrored magnetic barcode except by making new floor in the middle and carrying it outwards in both directions. Within five years plate tectonics was standard geology.

What Wegener could only infer, we can now measure. Permanent satellite stations track the continents in real time.

The Atlantic is widening by about two and a half centimetres a year. The fastest plates move around ten.

He was wrong about the speed — his estimates were far too fast, because he was using inaccurate longitude measurements from Greenland. He was right about everything else.

And the mechanism he lacked turned out not to be in the continents at all. It is in the ocean floor, which he had no way to see.

Wegener assembled four independent lines of evidence pointing to one conclusion, and offered a mechanism that was wrong.

His critics attacked the mechanism, which was fair, and then dismissed the evidence, which was not. The pattern in the rocks did not stop being there because nobody could explain it.

Thirty years were lost, and he did not live to see it fixed.

The chapter stills in this film are AI-generated. The photograph of Wegener, his own reconstruction maps and the fossil leaf are real. Sources are listed below.