Plate tectonics and continental drift

In a room full of instruments, a woman and a man stand in front of a machine that is printing a record. — Bruce Heezen and Marie Tharp watch a machine print out a depth record, aboard the oceanographic research ship USNS Kane, in 1968. Credit Line: AIP Emilio Segrè Visual Archives, Gift of Bill Woodward, USNS Kane Collection, CC0, source

In 1952, a woman who was not allowed on board ship found, drawing by hand, the proof that the continents move.

Her superior called her interpretation “girl talk”.

Then he spent weeks trying to refute it. He did not succeed.

The plates and their movements

The Earth’s surface is not a single, unbroken shell. It is cracked into some fifteen large pieces, called plates, along with many smaller plates. These pieces float on a hotter, softer layer, and they move. The science that studies this division and these movements is called plate tectonics — from the Greek tektôn, the carpenter, the one who assembles.

Their thickness, though, has to be pictured correctly. A plate is between 5 and 100 kilometres thick and thousands of kilometres wide. Relative to the Earth, that is thinner than an eggshell.

The plates do not move fast: from less than three centimetres to more than fifteen centimetres a year, depending on the place. The slowest advance at roughly the speed at which fingernails grow. The Pacific Plate is the fastest, the Antarctic Plate one of the slowest. But they never stop, and over millions of years that is enough to open an ocean.¹

Almost everything that moves on the Earth’s surface comes from this:

Where two plates move apart, magma rises and makes new ocean floor. This is how the Atlantic widens by a few centimetres every year.
Where two plates meet, one dives beneath the other or both crumple. This produces ocean trenches, volcanoes and mountain ranges: the Himalayas are the result of India pushing against Asia — it has travelled thousands of kilometres northwards, and the range is still rising by a few millimetres a year.
Where two plates slide past each other, they catch, the strain builds, then gives way all at once. That is an earthquake. The San Andreas Fault, in California, is of this type.²

One point is essential before going any further: the continents do not wander across the oceans. It is the whole plates — sea floor included — that move, and the continents sit on top of them. This is precisely the story the rest of this page tells: the idea took half a century to be accepted, and the proof came from a map drawn by hand.

What was believed before

In 1912, Alfred Wegener proposed that the continents move. The American geological establishment rejected the idea for forty years.³

There are two reasons for this rejection, and both must be given:

The first is scientific and legitimate: Wegener could not demonstrate by what mechanism continents could plough through the oceanic crust. He described a movement without being able to say what produced it.
The second is not: he was a meteorologist. He was accused of encroaching on a field that was not his own.⁴

The historian Naomi Oreskes, however, disputes this account for the United States: in her view, drift was rejected there not for lack of a mechanism, but because it seemed to conflict with the standards of practice of American geology.⁵

Marie Tharp’s working conditions

She joined the Lamont Geological Observatory at Columbia University in 1948, as a draughtswoman.⁶

Because she was a woman, she was not allowed to go aboard the research ships. She would not go to sea for the first time until 1965 or 1968, depending on the source.⁷

Her male colleagues went to sea to take the soundings: echo measurements giving the depth beneath the ship, column after column of figures.

She stayed ashore and converted these columns into topographic profiles. By hand.

What she saw in 1952

While aligning six profiles of the floor of the North Atlantic — each profile being a cross-section of the ocean floor along a ship’s track — she noticed a constant: a V-shaped notch ran down the centre of the ocean relief.⁸

The decisive methodological detail, which she herself emphasised, is this: the individual mountains didn’t match up from one profile to the next. Only the notch coincided — especially in the three northernmost profiles.⁹

This point is crucial. She did not reach her conclusion because everything looked alike. She reached it because one single thing recurred amid everything else that varied.

She recognised it as a rift valley, comparable to the one in East Africa — a region where the land is pulling apart. But this one lay at the bottom of the ocean.¹⁰

She drew the key conclusion from it: the ocean floor is spreading apart. New crust forms along this notch and pushes away what lies on either side.

This was the mechanism Wegener had been missing for forty years.

The reaction and the cross-checking of data

She took her interpretation to Bruce Heezen, her superior. She reported his reaction as follows:

“When I showed what I found to Bruce, he groaned and said, ‘It cannot be. It looks too much like continental drift.’”¹¹

She adds an explanation that is worth reading: at the time, believing in continental drift was almost a form of scientific heresy. Almost everyone in the United States thought it impossible.¹²

Then: “Bruce initially dismissed my interpretation of the profiles as ‘girl talk.’” He ordered her to recalculate everything and redraw everything.¹³

She started again. The valley was still there.¹⁴

This quotation is taken from a text written by Marie Tharp herself in 1999.

What tipped the balance was an independent cross-check. In the same laboratory, the epicentres of undersea earthquakes were being mapped for an entirely different purpose: the routing of transatlantic communications cables. When the two maps were overlaid at the same scale, the epicentres fell within the V-shaped notch.¹⁵

Two data sets, gathered for different reasons, that coincided. It took about six months for Heezen to be convinced.¹⁶

The sceptic who supplied the proof

What followed is even better.

At first, Jacques Cousteau did not believe in this valley. He set out to sea in the Atlantic, equipped with an underwater camera, intending to show that it did not exist.¹⁷

The images showed the cliffs of the valley.

He wanted to refute; he confirmed. That is the best thing that can happen to a hypothesis: someone who wants it to fail goes to check for himself, and reports the opposite.

The driving force and plate boundaries

How the plates interact

The Earth’s lithosphere is divided into rigid plates that interact at their boundaries in three ways:

Type of boundary	Movement	Geological result
Divergent (e.g. ridge)	The plates move apart	Creation of oceanic crust (rift)
Convergent (e.g. subduction)	The plates move towards each other	One plate dives beneath the other or forms mountains
Transform (e.g. fault)	The plates slide sideways	Frequent earthquakes without creation/destruction of crust

What makes the plates move?

For a long time, the dominant theory assumed that the plates were pushed by convection currents in the mantle, like rafts on water.

That explanation has changed. Today we know that the main force comes from the edge of the plates themselves: where a cold, dense oceanic plate sinks into the mantle (subduction), its own weight pulls it downwards and drags the rest of the plate along with it.¹⁸

It is not a push from below: it is a pull from the edge.

How fast do the continents move?

The order of magnitude was given above — a few centimetres a year. Accumulated over time, it adds up as follows.

Over 80 years (the span of a lifetime): a few metres — at five centimetres a year, about four metres.
Over 1 million years: a few tens of kilometres.
Over 100 million years: the opening or closing of an ocean.¹⁹

Interactive

At the speed of the plates

The plates do not move fast: a few centimetres a year. But they never stop.

The question

A few centimetres a year: enough to open an ocean?

An Icelandic landscape of plain and rocks. — Þingvellir, Iceland: a rift valley that marks the crest of the Mid-Atlantic Ridge, between the North American and Eurasian plates. Bryan Ledgard, CC BY 2.0, source

Without JavaScript, here are the animation’s data.

Three speeds
Where	Speed	Source
Mid-Atlantic Ridge (average)	2.5 cm a year	1
Pacific Plate against North American Plate (average)	5 cm a year	2
The fastest	more than 15 cm a year	3

Two benchmarks from the article (calculated values)
Time span	Speed	Distance
80 years (a lifetime)	5 cm a year	4 m
1 million years	2.5 cm a year	25 km

The distances are calculated: speed multiplied by time, as if the plate always kept the same speed. The drawing of the plates is not to scale.

The takeaway A few centimetres a year is invisible over a lifetime. Over millions of years, it is enough to open an ocean.

Photo credits 3

The relief of the South Atlantic floor, between South America and Africa (NOAA). NOAA, Public domain, source
The breakup of Pangaea: the continents, at first joined, move apart to their present positions (United States Geological Survey, ’This Dynamic Earth’). Jacquelyne Kious and Robert I. Tilling (USGS), Public domain, source
Blue Marble: Next Generation, with land topography and sea-floor bathymetry, in equirectangular projection: map produced by Reto Stöckli, NASA Earth Observatory (NASA Goddard Space Flight Center). Reto Stöckli, NASA Earth Observatory, modified by NordNordWest, Public domain, source

Sources 4

United States Geological Survey, This Dynamic Earth: The Story of Plate Tectonics, “Understanding plate motions” page (https://pubs.usgs.gov/gip/dynamic/understanding.html).
The rate of spreading along the Mid-Atlantic Ridge averages about 2.5 centimeters per year (cm/yr), or 25 km in a million years.
Open access open the source · checked on 24 October 2026
United States Geological Survey, This Dynamic Earth: The Story of Plate Tectonics, “Understanding plate motions” page, passage on the San Andreas Fault (https://pubs.usgs.gov/gip/dynamic/understanding.html).
Along it, the Pacific Plate has been grinding horizontally past the North American Plate for 10 million years, at an average rate of about 5 cm/yr.
Open access open the source · checked on 24 October 2026
United States Geological Survey, This Dynamic Earth: The Story of Plate Tectonics, “Historical perspective” page, 1st paragraph (https://pubs.usgs.gov/gip/dynamic/historical.html) […] “Understanding plate motions” page, part on divergent boundaries (https://pubs.usgs.gov/gip/dynamic/understanding.html).
The theory of plate tectonics states that the Earth's outermost layer is fragmented into a dozen or more large and small plates that are moving relative to one another as they ride atop hotter, more mobile material. […] The rate of spreading along the Mid-Atlantic Ridge averages about 2.5 centimeters per year (cm/yr), or 25 km in a million years. […] The Arctic Ridge has the slowest rate (less than 2.5 cm/yr), and the East Pacific Rise near Easter Island, in the South Pacific about 3,400 km west of Chile, has the fastest rate (more than 15 cm/yr).
Open access open the source · checked on 24 October 2026
United States Geological Survey, This Dynamic Earth: The Story of Plate Tectonics, “Understanding plate motions” page (https://pubs.usgs.gov/gip/dynamic/understanding.html), paragraphs on the types of boundary.
Divergent boundaries occur along spreading centers where plates are moving apart and new crust is created by magma pushing up from the mantle. […] Such destruction (recycling) of crust takes place along convergent boundaries where plates are moving toward each other, and sometimes one plate sinks (is subducted) under another. […] The zone between two plates sliding horizontally past one another is called a transform-fault boundary, or simply a transform boundary.
Open access open the source · checked on 24 October 2026

Open full page · All interactives

The word that remains

We still use the expression “continental drift”. This vocabulary dates from the time when it was believed that the continental masses slid directly over the ocean floor. The explanation has changed (it is whole plates that move), but the term has stayed.

The historical credit and what followed

The first results were presented in the mid-1950s. In keeping with the customs of the time, Bruce Heezen published the work as principal author and received a disproportionate share of the initial credit: the first published abstract does not mention Marie Tharp.²⁰

In 1977, Tharp and Heezen published the first composite world map of the ocean floor, revealing an undersea mountain chain more than 60,000 km long.²¹

Marie Tharp’s role would be fully recognised later: she received the Hubbard Medal in 1978, and in 1997 the Library of Congress ranked her among the most important cartographers of the twentieth century.²²

Note 1

The theory of plate tectonics states that the Earth's outermost layer is fragmented into a dozen or more large and small plates that are moving relative to one another as they ride atop hotter, more mobile material. […] The rate of spreading along the Mid-Atlantic Ridge averages about 2.5 centimeters per year (cm/yr), or 25 km in a million years. […] The Arctic Ridge has the slowest rate (less than 2.5 cm/yr), and the East Pacific Rise near Easter Island, in the South Pacific about 3,400 km west of Chile, has the fastest rate (more than 15 cm/yr).

United States Geological Survey, This Dynamic Earth: The Story of Plate Tectonics