Total eclipse — Eddington proves Einstein
The eclipse that made Einstein famous overnight — starlight bending past the darkened Sun confirmed general relativity.
📍 Visible from where you are right now — computed in your browser
bearing · altitude · rise time · best time tonight
The moment of Total eclipse — Eddington proves Einstein: https://nowsky.org/e/eclipse-1919-eddington — opens at the right place and time, replayable, no login.
Available views: Orbit ✓ · Sky ✓
The story
On 29 May 1919 Arthur Eddington photographed a total eclipse through gaps in the cloud from Príncipe, a small island in the Gulf of Guinea. Totality ran up to 6 minutes 51 seconds, and the Sun happened to stand in front of the Hyades, a cluster rich in bright stars — a better arrangement for measuring the bending of starlight comes only once in centuries.
The target was starlight grazing the Sun. If Einstein's general relativity was right, the Sun's gravity would curve spacetime and shift the apparent positions of the stars by 1.75 arcseconds; Newtonian mechanics predicted half that, 0.87. The same eclipse was observed from Sobral in Brazil by Crommelin and Davidson, and from Príncipe by Eddington and Cottingham.
More than six months of plate analysis gave results matching Einstein: 1.98 ± 0.12 arcseconds from Sobral's 4-inch telescope, 1.61 ± 0.30 from Príncipe. They were announced at a joint meeting of the Royal Society and the Royal Astronomical Society on 6 November 1919, and the next day "Lights all askew in the heavens" ran across the world's newspapers. A little-known physicist became the most famous scientist alive overnight.
Newton's universe was replaced by Einstein's on the strength of one eclipse — which is why this is called the most important eclipse in the history of science. The orbit replay shows that day's Sun–Moon–Earth alignment; move the sky view to Príncipe and the Hyades stand beside the darkened Sun.
Why it matters
General relativity was complete in 1915 but almost untestable. It explained Mercury's perihelion shift, yet that was an explanation rather than a prediction, and the bending of light could only be measured with the Sun covered. A 1914 expedition to the Crimea was lost to the outbreak of war; 1919 was the first chance after it.
The measurement itself was at the limit. 1.75 arcseconds is the width of a coin a kilometre away, and thermal distortion of the plates or a shift of focus produces errors of the same size. Sobral's main-telescope plates were discarded for heat distortion; the backup 4-inch provided the decisive data.
There is a historiographical coda: later critics argued that the data were selected to fit a chosen conclusion, but a 1979 re-analysis of the original plates at Greenwich upheld the result. The same deflection has since been confirmed thousands of times more precisely, from Gaia in 2017 to the Event Horizon Telescope in 2019.
How to watch
The path ran from northern South America (Sobral, Brazil, in the morning) across the Atlantic to Príncipe in the Gulf of Guinea (afternoon) and on into central Africa. Sobral had 5 minutes 13 seconds around 9 a.m.; Príncipe about five minutes around 2 p.m., through gaps in cloud.
The 1919 work used photographic plates and mechanical micrometers. To repeat it today you photograph the stars near the Sun during totality and the same field at night six months later, then compare positions — amateurs reproduced the measurement with digital cameras and precise software at the 2017 and 2024 eclipses.
Set the sky view to Príncipe (São Tomé and Príncipe) near 14:00 UT on 29 May 1919 and Aldebaran and the Hyades appear beside the darkened Sun as they did that day; the orbit replay shows the alignment from space.
See this for yourself — open that sky at that moment · replay the moment in the orbit view
Background
The eclipse belongs to Saros 136, the series that produced the longest totalities of the twentieth century — 8 June 1937 (7 min 4 s), 20 June 1955 (7 min 8 s, the century's longest) and 30 June 1973 (7 min 4 s) followed it.
Einstein first computed 0.87 arcseconds in 1911 and doubled the value with the completed theory of 1915. Because expeditions to Argentina in 1912 and the Crimea in 1914 failed through weather and war, the earlier, wrong figure was never put to the test.
Eddington was a Quaker who refused military service on religious grounds; sending him on the expedition instead of to the front was the Royal Astronomical Society's compromise. A British astronomer confirming a German theory months after the armistice was also read as a symbol of science's internationalism.
FAQ
What did they measure?
How far the stars beside the darkened Sun were shifted from their usual positions — 1.75 arcseconds by Einstein, 0.87 by Newtonian mechanics.
What were the results?
1.98 ± 0.12 arcseconds at Sobral and 1.61 ± 0.30 at Príncipe — Einstein's side.
What about the data-selection controversy?
A 1979 re-analysis of the plates at Greenwich upheld the original conclusion, and Gaia, the EHT and others have since confirmed the effect far more precisely.
Why did the Hyades matter?
Measuring the shift needs several bright stars right beside the Sun, and in May 1919 the Sun happened to stand in front of the Hyades cluster.
Related
- The 1999 European total eclipse
- The 2027 total eclipse — six minutes of darkness
- The 2009 total eclipse — longest of the century
- The Sun — object index
- Aldebaran — object index
Times and figures are approximate — verify with official sources when planning travel. Simulation uses J2000 mean elements.