Sir Arthur Eddington (1882-1944) was quite a lucky man. He was a Quaker, which helped keep him out of World War I. He also had a talent for being at the right place at the right time. A mathematically inclined English astronomer, he happened to be working as a secretary at the Royal Astronomical Society when Dutch astronomer Willem de Sitter (1872-1934)’s letters and papers on Einstein’s new theory of general relativity began arriving in England. Quick to comprehend the implications of Einstein’s ideas, already in 1916 Eddington became a supporter of Einstein. This despite the fact that the theory had been formulated by a German at a time when anti-German sentiments were at an all-time high in England due to the war. Eddington, a pacifist, was among the first to teach the theory in the West (both at Cambridge University and at a meeting of the British Association). In 1918, he also produced a report on the theory for the Physical Society1. One issue however remained. The theory had not been empirically verified, and thus, largely remained a conjecture whose implications rendered many scientists skeptical.
Convinced of its validity, within three years of receiving de Sitter’s letters, Eddington had organized two expeditions, to the island of Principe in West Africa and to Sobral in Brazil. His goal: prove the validity of Einstein’s theory by measuring the deflection of light from distance stars (by the sun’s gravity) during a total solar eclipse. His team’s findings, announced at a meeting of the Royal Society in November 1919 and published in its proceedings the following year, indeed confirmed Einstein’s prediction. The deflection of starlight due to the gravitational field of the sun was twice that predicted by Newton’s theory—a new paradigm was established.
LIGHTS ALL ASKEW, IN THE HEAVENS
Men of Science More Or Less Agog Over Results of Eclipse Observations.
EINSTEIN THEORY TRIUMPHS
"Stars not where they seemed or were calculated to be, but nobody need worry."
— Headline in The New York Times, Nov. 10th 1919
This week’s essay is about Sir Arthur Stanley Eddington and his grand expedition to prove Einstein’s theory of relativity.
Einstein’s Prediction
Shortly after publishing his general theory of relativity in correct form in 1915, Einstein proposed three empirical predictions that, if any should prove to be wrong, would discredit his proposed theory. As he later stated:
"The chief attraction of the theory lies in its logical completeness. If a single one of the conclusions drawn from it proves wrong, it must be given up; to modify it without destroying the whole structure seems to be impossible."
Einstein’s three empirical tests of the validity of his theory (now called the ‘classical tests’) were:
An explanation of the perihelion precession of Mercury’s orbit.
An expectation of the angle of deflection of starlight by the Sun.
An expectation of the gravitational redshift of light.
The first prediction had been an outstanding problem in celestial mechanics dating back to Urbain Le Verrier (1811-77) and his 1859 observation that the planet Mercury deviates from the precession2 predicted by Newtonian physics. Various unsuccessful attempts at explaining Mercury’s precession had been made over the years (including the suggestion of a yet-unidentified hypothetical planet named ‘Vulcan’).
Einstein’s new theory settled the debate by showing how gravitation is mediated by the curvature of spacetime, explaining the amount of perihelion shift of Mercury previously observed. His calculations were published in 1916:
Einstein, A. 1916. "The Foundation of the General Theory of Relativity" (PDF). Annalen der Physik. 49 (7): 769–822
Experimental verification of the gravitational redshift of light (the third prediction) would in 1916 still be 38 years away. At the time, the sensitivity of astronomical equipment was insufficient to be able to detect such an affect, which according to Einstein would be easiest to detect near a white dwarf (“dying star”).
As such, with one prediction found true and another seemingly untestable, Einstein and his followers in 1916 placed their hopes on the second: the prediction that light passing through a gravitational field will deflect at an angular distance twice that which Newton’s classical theory assumes.
History
Sir Isaac Newton (1643-1727) mentions the bending of light due to gravity as early as in his 1704 book Opticks, but fails to provide any calculations for the expected angle of deflection. After Newton, the first to consider the interaction of gravity and light was perhaps Henry Cavendish (1731-1810) who in an unpublished 1784 “scrap” described “the bending of a ray of light by gravitation” (Lotze & Simionato, 2022). About twenty years later, Johann Georg von Soldner (1776-1833) published a paper pointing out that, in the Newtonian paradigm, starlight will be deflected when it passes near a massive object at a distance that is proportional to the mass of the object. Soldner’s work, based on Newton’s corpuscular theory of light predicted that light would be diverted by a factor of 0.84 seconds of arc. In the paper, written in 1801 but not published until 1804, Soldner pointed out that if it were possible to observe stars in close distance to the sun, it might be important to take this effect into consideration. At the time, however, such observations were impossible. His (translated) paper is available via Wikisource:
Soldner, J. G. v. (1801–1804). "On the deflection of a light ray from its rectilinear motion, by the attraction of a celestial body at which it nearly passes by". Berliner Astronomisches Jahrbuch: 161–172.
Four years prior to Einstein’s completing his theory in 1915, without crediting Soldner3, Einstein published a calculation (based on the equivalence principle of his special theory of relativity) which predicts that light passing near a body such as the sun will deflect at 0.83 seconds of arc, essentially the same prediction as Soldner but using a different theoretical framework. In the same paper, he encourages astronomers to test this prediction by taking advantage of the shielding of sun by the moon during solar eclipses and using astrometry to measure the positions of stars around the sun. In 1911 still four years away from completing the theory (see essay below), Einstein’s (and Soldner’s) prediction was wrong by more than a factor of two.
1912 Expedition to Brazil
Responding to Einstein’s request, German astronomer Erwin Finley-Freundlich (1885-1964) first investigated whether existing solar eclipse photographs would be suitable for verifying Einstein’s prediction. Director of the Argentine National Observatory Charles D. Perrine (1867-1951) (who himself had participated in four solar eclipse expeditions in 1900, 1901, 1905 and 1908) responded that he did not believe existing photographs would be useful for such testing. Freundlich next pushed Perrine to include observation of light deflection as part of the Argentine National Observatory’s program for the coming eclipse of 1912. Perrine agreed and lead a team of astronomers from the Cordoba Observatory to document the October 12th solar eclipse in Cristina, Brazil. Unfortunately, as is often the case in astronomy, weather on the day of the eclipse prevented the team from succeeding in their test of Einstein’s theory, the first of its kind.
Interestingly, Eddington had taken part in a British expedition to the same location to observe the same eclipse, but for different reasons. While there, Perrine and Eddington spent several days together and thus may have discussed Einstein’s theory and the basis of his 1911 (incorrect) prediction (Gates & Pelletier, 2019*).
1914 Expeditions to Crimea
Two years after the unsuccessful expedition in Brazil, Perrine, Freundlich and U.S. astronomer William Wallace Campbell (1862-1938) again set out to test Einstein’s prediction via the observation of light deflection near the sun during a total solar eclipse. However, this time, due to the location of the umbra4 of the eclipse in Crimea (then in the Russian Empire) only the expeditions lead by Perrine and Campbell were able to reach the location. As Germany had declared war on Russia three weeks prior to the eclipse, Freundlich and his team of German astronomers were forced to return home. The U.S. and Argentine teams successfully made it to the eclipse station but again, due to clouds, Einstein’s prediction of the deflection of light would remain untested.
The following year, in 1915, after a decade of revisions (see essay above), Einstein finally published his general theory of relativity, adjusting his prediction of the deflection of light from 0.83 seconds of arc to 1.75 seconds of arc. As was pointed out in Gates & Pelletier (2019)*, in hindsight, Einstein was indeed fortunate that weather and war prevented Perrine, Freundlich and Campbell from testing his 1911 prediction, as successful such tests would have shown his prediction to be off by a factor of more than two — questioning the validity of his new theory and, potentially, Einstein’s motivation for completing it.
1918 Expedition to Washington
Four years after the failed expeditions to Crimea, Campbell and his team again set out to verify Einstein’s theory by traveling to Goldendale in Washington, where another solar eclipse was set to occur5. Unfortunately, the Lick Observatory’s equipment, which Campbell’s team relied on, at the time was stored in Petrograd at the Poulkovo Observatory. Despite being shipped from Russia in August of 1917, due to the war, the equipment did not reach the U.S. in time for the June 8th eclipse. Campbell and his team thus had to take their photographs with poorly suited equipment borrowed from the Oakland Observatory, and so, again did not succeed in testing Einstein’s theory.
Eddington’s 1919 Expedition to Brazil
Although Eddington’s interest in Einstein’s theory may have begun as early as when he met Perrine in Brazil in 1912, his active involvement in the testing of the theory (is generally agreed) started in 1916. As the story goes, Eddington was working as a secretary of the Royal Astronomical Society at a time during World War I when German scientific journals were generally unavailable to British scientists. Indeed, a fractioning of the scientific community was afoot at the time, wherein German scientists (including Arnold Sommerfeld and Max Planck) opposed the use of the English language by German scientists, editors of books and translators. Fortunately, Einstein (a German opposed to the fractioning) maintained close ties with Dutch scientists and visited the Netherlands often.
In 1912, Einstein’s friend Paul Ehrenfest (1880-1933) moved to Leiden to succeed Hendrik Lorentz (1853-1928) as professor of theoretical physics at the University of Leiden. A close friend, Einstein visited him there often (even during the war when obtaining permissions to travel were difficult to come by) (Kennefick, 2019*). On one such visit, in 1916, Einstein also met with astronomer Willem de Sitter (1872-1934), with whom he discussed his theory and its astronomical predictions. De Sitter, who was a mathematician by training and later became the Director of the Leiden Observatory, understood Einstein’s work and communicated it to scientists in England. Eddington, a benefactor of De Sitter’s communiqués, later commissioned De Sitter to write articles specifically on the astronomical implications of the theory for publication in the Monthly Notices of the Royal Astronomical Society. These papers were published in the years 1916-17 and are available via the links below:
De Sitter, W. (1916-07-14). "On Einstein's Theory of Gravitation and its Astronomical Consequences. First Paper". Monthly Notices of the Royal Astronomical Society 76(9): 699–728.
De Sitter, W. (1916-12-08). "On Einstein's Theory of Gravitation and its Astronomical Consequences. Second Paper". Monthly Notices of the Royal Astronomical Society 77(2): 155–184.
De Sitter, W. (1917-11-09). "On Einstein's Theory of Gravitation and its Astronomical Consequences. Third Paper". Monthly Notices of the Royal Astronomical Society 78(1): 3–28.
Enthusiastic about the contents of De Sitter’s papers, Eddington took it upon himself to be the first to lecture on Einstein’s theory to an English-speaking audience. He did so first at Cambridge and later at meeting of the British Association.
1919 Expeditions to Africa and Brazil
At 36 years old in 1918, Eddington was eligible to be drafted into the military following the introduction of wartime conscription in Britain in 1916. Raised as a Quaker, Eddington applied for an exemption from this duty, first on the grounds of national interest and later on the grounds that he was a conscientious objector. Astronomer Royal Sir Frank Watson Dyson (1868-1939) supported Eddington’s request, writing in his favor that he had an essential role in a coming expedition to Principe to observe a total eclipse, leading the tribunal handling the case to grant Eddington a twelve month exemption. The war ended before the exemption did.

Under the direction of Dyson, two expeditions set out to photograph the 1919 solar eclipse in the specific pursuit of measuring the deflection of starlight due to the gravitational field of the sun. The expeditions were funded by a government grant of 100 pounds for adapting telescopes and 1,000 pounds for “various expenses”. To account for weather, one team set out to the West African Island of Principe and the other to the Brazilian town of Sobral. The conditions for obtaining results would be especially favorable, as the sun would be passing across a large cluster of starts in the constellation Taurus—giving many opportunities to photograph bright starts near the sun (at the specific time when it was blocked by the moon).


Sobral Location
The Royal Greenwich Observatory (lead by Dyson) sent a team consisting of Charles Davidson (1875-1970) and Andrew Crommelin (1865-1939) to take photographs from Sobral, a small town in northeast Brazil. Although Dyson coordinated the expedition, he himself did not travel, for unknown reasons. Davidson and Crommelin arrived on April 30th with a 13-inch aperture main telescope and a backup 4-inch telescope borrowed from astronomer Aloysius Cortie (1859-1925). On the day of the eclipse they were lucky to have clear skies and an eclipse which lasted for 5 minutes and 13 seconds starting at 9 am. Their main telescope recorded twelve stars, but the photographs were blurred and so ultimately discarded by Dyson from the final analysis. The official observations from the Sobral location, thus, amounted to photographs of seven stars recorded by the smaller “backup” telescope (seen in photograph below, the right-most square “tube”).
Principe Location
Eddington himself lead the team traveling to Principe, an island in the Gulf of Guinea, West Africa. Traveling with him was Edwin Turner Cottingham (1869-1940), a fellow of the Royal Astronomical Society whose expertise was astronomical timekeeping. The two arrived on April 23rd with a telescope from the Cambridge Observatory similar to the main telescope sent to Sobral. The weather was poor on the day of the eclipse, and so an overcast sky impacted the quality of their photographs of five stars in the Taurus constellation.

The Analysis
To test Einstein’s prediction that the sun’s gravity bends starlight, Dyson and Eddington set out to compare the positions of the stars they had photographed at Sobral and Principe with photographs of the same constellation at night (when the gravitational field of the sun is not significantly affecting the path of the light from the stars as it reaches earth). Although such comparison photographs should ideally be taken in the same location as the eclipse photos, Eddington chose to instead photograph the Oxford sky on a clear night.
Comparisons between the Oxford plate and the three sets of photographic plates obtained at Sobral and Principe all showed deviations. Analysis of the blurry images from the main telescope at Sobral showed a deflection of 0.86 seconds of arc, consistent with predictions based on Newton’s law of gravity. Due to the poor quality of the images, though, these calculations were ultimately discarded from the final analysis. Eddington’s photographs of the overcast sky at Principe were also of poor quality, as many stars were either obscured by the halo from Sun’s light or covered by the moon. These limitations taken into account, the calculated deflection at Principe was 1.61 arc seconds (+/- 0.3 arc seconds). Finally, analysis of the sharper images produced by the smaller telescope at Sobral showed a deflection of 1.98 arc seconds (+/- 0.12 arc seconds), slightly larger than Einstein’s prediction of 1.75 arc seconds (and more than twice that predicted by Newton’s theory of gravity).
As Dyson, Eddington and Davidson’s report to the Royal Astronomical Society later concluded:
“The greatest weight must be attached to those obtained with the 4-inch lens at Sobral. From the superiority of the images and the larger scale of the photographs, it was recognized that these [results] would prove to be the most trustworthy”.
As Einstein later stated to the Brazilian press "The problem conceived by my brain was solved by the bright Brazilian sky".

A Final Test
With two of Einstein’s three proposed tests of his relativity theory confirmed and accepted—in 1916 and 1919—by 1920 only the observation of a gravitational redshift remained unaddressed.
Einstein’s prediction was that light will be redshifted when it passes close to a strong gravitational field. Einstein first proposed the idea in 1907, at the time inspired by the equivalence principle of theory of special relativity. The prediction was that electromagnetic waves traveling away from a gravitational field would (due to gravity) lose energy. Such a loss would correspond to a decrease in the wave frequency of the light and thus an increase in its wavelength (a redshift).
To test the prediction, Einstein proposed measuring the spectral lines of a white dwarf star, which has a strong gravitational field. A first attempts was made in 1925 by Walter Sydney Adams (1876-1956), but the results (measurements of the redshift of Sirius-B) were criticized for having been contaminated by light from a nearby star (Sirius). Indeed, the first accurate measurement of a gravitational redshift did not occur until 1954, when astronomer Daniel M. Popper () measured a 21 km/s gravitational redshift of 40 Eridani B, a white dwarf.
If you are interested in reading more about Eddington’s expedition, I recommend Daniel Kennefick’s 2019 book No Shadow of a Doubt: The 1919 Eclipse that Confirmed Einstein's Theory of Relativity*.
I hope you enjoyed this essay,
Best
Jørgen
Related Privatdozent Essays
About
At the time of the writing of this essay, the Privatdozent newsletter goes out to 11,585 subscribers via Substack.
References
Earman, J. & Glymour, C. 1980.. "Relativity and eclipses: the British eclipse expeditions of 1919 and their predecessors". Historical Studies in the Physical Sciences. 11 (1): 49–85.
Einstein, A. 1911. "On the Influence of Gravitation on the Propagation of Light". Annalen der Physik. 35 (10): 898–908.
Gates Jr, S.J. & Pelletier, C., 2019. Proving Einstein right: the Daring expeditions that changed how we look at the universe. Hachette UK.
Kennefick, D.J., 2019. No shadow of a doubt: the 1919 eclipse that confirmed Einstein's theory of relativity*. Princeton University Press.
Lotze, K.H. and Simionato, S., 2022. Henry Cavendish on Gravitational Deflection of Light. Annalen der Physik, 534(7), p.2200102.
Pivetta, M. & de Oliveira Andrade, R. 2019. When light bent. Pesquisa. Available at: <https://revistapesquisa.fapesp.br/en/when-light-bent-2/>
Entitled Report on the Relativity Theory of Gravitation.
Precession is slow movement of the axis of a spinning body around another axis due to a torque acting to change the direction of the first axis. It can be seen in the circle slowly traced out by the pole of a spinning gyroscope.
Noted anti-Semite and Nazi sympathizer Philipp Lenard () went on to accuse Einstein of plagiarism in a 1921 Annalen der Physik paper. Einstein’s lack of a credit to Soldner is however generally accepted to have been an oversight. There is no indication that Einstein (or anyone else) at the time were aware of Soldner’s 100 year old paper, as corpuscular theories of light by that time were out of fashion.
The umbra is the innermost and darkest part of a shadow, which during an eclipse is where the total eclipse occurs.
According to Earman & Glymour (1980), the force behind Campbell’s team’s efforts were due to astronomer H. D. Curtis (), who both brought Einstein’s prediction to Cambell’s attention, set up the instruments during the eclipses, took the photographs in 1918 and did the final measurements (p. 63)
Beautiful piece. It's intriguing to see how earlier scientists like Cavendish and Soldner had touched upon concepts that Einstein would later formalize in his general theory of relativity. This underscores the evolutionary nature of scientific ideas, where new theories build upon the groundwork laid by predecessors, sometimes unknowingly.