How did a solar eclipse prove Einstein right?
The short answer
In 1919, expeditions led by Arthur Eddington photographed stars near the Sun during a total solar eclipse. The stars appeared shifted from their normal positions, showing that the Sun's gravity bends starlight. The measured shift matched Einstein's prediction of about 1.75 arcseconds, roughly double what Newtonian physics predicted.
Einstein's idea was that space itself bends
General relativity, published in 1915, proposed that space is not empty nothing but something that can be distorted. Place a massive object in it, such as a star, and the space around it sags. Light always travels the straightest available path, but near a massive object the straightest available path is itself curved. So starlight passing close to the Sun should be deflected slightly, making the star appear shifted from where it ought to be.
Testing it required switching off the Sun
To measure the shift you must photograph stars sitting right next to the Sun. On a normal day that is impossible: the glare drowns out every star in the sky. They are up there, but completely invisible. The only natural circumstance that removes the Sun's light while leaving the stars visible is a total solar eclipse, which provides a few minutes of darkness in the middle of the day.
Two expeditions, one eclipse, terrible weather
In May 1919 two teams sailed from England. Arthur Eddington went to the island of Príncipe off West Africa; a second team set up at Sobral in Brazil. Two sites, because cloud at one might still leave clear sky at the other. On Príncipe it rained on the morning of the eclipse. A gap in the clouds opened with minutes to spare. Of sixteen photographic plates exposed there, cloud ruined most. Two were usable.
The measurement was absurdly small
Months later came the comparison. Plates taken during the eclipse were laid against plates of the same patch of sky photographed at night, when the Sun was nowhere near. The stars near the Sun's edge had shifted outward by an angle comparable to spotting a coin from a few kilometres away. Einstein had published his predicted figure in advance, about 1.75 arcseconds. Newtonian physics predicted roughly half that. The plates landed near Einstein's number.
Announced in London, and Einstein became famous overnight
The results were presented in November 1919 and the press reaction was extraordinary. Newspapers declared a revolution in science and the overthrow of Newtonian ideas. A relatively obscure physicist in Berlin became, within days, the most famous scientist alive. It is one of the clearest examples in history of a single carefully planned measurement changing how we understand the universe.
You rely on this every time you use a map
This is not merely historical. GPS satellites carry extremely precise clocks, and because they sit further out of Earth's gravitational field and move quickly, their clocks tick at a measurably different rate from clocks on the ground. Without relativistic corrections applied, satellite navigation positions would drift badly, accumulating substantial error each day. Your phone quietly applies the physics that eclipse confirmed, every time it locates you.
Sources
Full transcript: The Eclipse That Changed Physics Forever
The eclipse that proved Einstein right. 1919, the sun goes out right on schedule and two stars sitting beside it move. Except stars don't move. Their light was bent on its way to us by the sun itself and that tiny shift caught on a glass photograph in the middle of an eclipse turned a quiet physicist in Berlin into the most famous scientist on Earth overnight.
Today, the eclipse that proved Einstein right. Welcome back to What and Why, 11 days to our eclipse. This is the story of the most important few minutes of darkness in the history of science. Einstein's idea from 1915 sounds impossible.
Space isn't just empty nothing, it's a thing like a stretched sheet. Put something heavy on it, a star, a planet, our sun, and space sags around it. Now, light always takes the straightest path it can find, but near the sun, the straightest path is curved. So, starlight that grazes past the sun gets bent just a whisker on its way to your eye, which means a star sitting next to the sun should look shifted, out of place, moved.
Lovely idea, completely wild. Now, prove it. Here's the problem, to catch the shift, you have to photograph stars sitting right next to the sun. Try that on a normal day, glare.
The sun drowns out every star in the sky. They're up there right now, this second, and you can't see a single one. You need the sun's light off and the stars on at the same moment. There is exactly one switch in nature that does that, a total eclipse.
A few minutes of night in the middle of the day, just long enough to point a telescope and shoot. May 1919, two expeditions sail from England chasing one eclipse. Arthur Eddington takes the island of Principe off the coast of West Africa. A second team sets up across the ocean in Sobral, Brazil.
Two sites, one reason. If clouds win at one, the other might still get sky. Wooden telescopes, glass photographic plates, months of travel for minutes of eclipse, and on the morning of eclipse day on Príncipe, it rains. Then, with minutes to spare, a gap in the clouds.
Totality, one of the longest of the whole century, six plus minutes of night. The shutters fly. 16 plates on Príncipe. Clouds spoil most of them.
Two survive. Months later comes the moment of truth. Take the eclipse plates, take plates of the same patch of sky shot at night when the sun was nowhere near. Lay one over the other, and the stars moved.
Every star near the sun's edge nudged outward by an angle so tiny, it's like spotting a coin from 3 km away. Einstein had named his number in advance, 1.75 arc seconds. The old physics said half that. The plates from Príncipe and Sobral landed right around Einstein's number, and nowhere near the old one.
The sheet really sags. Light really bends. November 6th, 1919. The results are announced in London, and the world loses its mind.
Revolution in science, Newtonian ideas overthrown, lights all askew in the heavens. That one's real. Einstein wakes up the most famous physicist alive because the moon switched off the sun for a few minutes, and somebody was ready with a camera. And this isn't just history.
You used bent space-time today. GPS satellites carry ferociously precise clocks, and up where they orbit, space-time genuinely runs differently. Their clocks tick a tiny bit faster than yours. Skip Einstein's correction, and your map pin drifts kilometers off every single day.
Turn it on, and the pin sits exactly where you're standing. Your phone quietly runs the physics that eclipse proved 1919 in your pocket. And now it's your turn. 11 days, glasses sorted, that's episode two.
Know what the afternoon will look like minute by minute, episode three. August 12th, eyes up. But one question keeps nagging at me. If eclipses are this useful, this spectacular, why are they so rare?
New moon happens every month, why doesn't the sky go dark every month? That's episode six, right after the big day. So remember, heavy things bend light's path. That was the wild idea, and a 1919 eclipse caught it happening, that was the proof, and your phone corrects for it every day, that's the receipt.
Big science hiding inside a few minutes of dark. 11 days, August 12th, see you under the shadow. And the day after, right here, what we all just saw. See you on What and Why.
Transcript of the episode, lightly cleaned for readability.
Sample question
Question 1 of 10
What did Einstein's general relativity predict about light passing near a massive object?
- Its path would be bent, because space itself is curved by mass correct
- It would slow down and stop
- It would change colour permanently
- It would be absorbed entirely
Light follows the straightest available path, but near mass that path is curved.
That is one of 10. The rest are not shown here, and the answer order changes every time you take the quiz.
Written by Watt & Why. Published 1 August 2026. Last reviewed 17 October 2026.