The way this story begins, it could be the setup of a math or physics problem. Two astronomers and their assistants depart England, one pair headed south toward the island of Príncipe, the other southwest, to Brazil. Meanwhile the moon, as it loops around the earth, will soon occupy a position between the earth and the sun. Once the observers reach their tropical outposts, they assemble their equipment and wait for the moon and the sun to line up exactly: a solar eclipse.
The eclipse in question took place a hundred years ago, in May 1919, drawing a swath of shadow over the South Atlantic Ocean. In Príncipe and in Sobral, Brazil, during the brief period when the sun was blocked overhead, the observing teams tried to capture images of the Hyades, a cluster of stars whose position in the sky was close to the sun. Morning rain nearly foiled the attempt in Príncipe, but just in time the sky cleared up enough that the astronomer there, Arthur Stanley Eddington, managed to make photographic plates during the eclipse and send a telegram to a colleague back in England: “Through cloud, hopeful. Eddington.”
Before I’d read about these trips, knowing only of their existence, I’d thought of them, romantically, as though they were Victorian expeditions. I pictured imperial ships setting sail, laden with equipment and provisions, for far-flung territory, while ladies back on shore waved their handkerchiefs and so on. But that’s wrong, anachronistic; this was just after World War I, and Eddington and Edwin Turner Cottingham, a clockmaker who maintained the instruments on the Príncipe trip, were lucky to find rides on commercial transport. They had to wait a month on the island of Madeira until they found a Príncipe-bound ship that could take them, like two guys hailing a taxi. Once there, Eddington was the only astronomer present, and they conscripted plantation workers to help them.
The Brazil mission involved a few more scientists, but all the same it was less grand and more fragile than I’d assumed—and so I find it all the more incredible, what was accomplished by the two voyages. The hope shared by Eddington and his collaborators was that the data they collected during the eclipse would provide evidence for or against the predictions of general relativity, which Einstein had proposed in 1915, and which asserted that light traveling close to the sun would be bent more strongly by gravity than was previously thought. (Normally the light from the sun itself made it impossible to measure the deflection of starlight, hence the need for an eclipse.) In the years since, the war had handicapped scientific operations, and the few other eclipse expeditions that could’ve tested the theory came up short, in one case because the astronomer was arrested as a spy.
War or no war, eclipse missions were difficult. First of all, scientists had to figure out where the path of totality would land, and only in the nineteenth century had improvements in science and in mapmaking made those calculations reliable. Then they had to locate a spot within the path of totality where it would be feasible to go (not in the middle of the ocean, say, or the jungle) and try to anticipate something about the weather there and get the money to go and then actually make the journey, while transporting sensitive lenses. And then they’d pray for clear skies during the very short time of totality—almost six minutes, in the case of the 1919 eclipse as seen from Príncipe, which was a longer window than most. Eddington would later tell his mother and his sister that he barely glanced up at the eclipse as it was happening, busy as he was with his photographic plates. The plates, in this case, presented another challenge: The predicted shift in position of the stars would be less than the width of the stars’ images on the plate, and so the astronomers had to work extremely carefully both during the eclipse and afterward as they made painstaking measurements. The calculations required to distinguish true deflection from optical effects caused by the telescope were long and tedious.
The official announcement of their findings came in November, at a joint meeting of the Royal Astronomical Society and the Royal Society in London. The hall was packed, the atmosphere tense. A portrait of Isaac Newton hung in the background, reinforcing for the audience the possibility that Newton’s laws, the foundation of modern physics, might now be radically modified. Frank Watson Dyson, who was England’s Astronomer Royal and the director of the Brazil mission (though he didn’t make the trip himself), and Eddington announced in turn that their findings agreed with the relativity model: The eclipse images of the Hyades stars indicated that their light had been deflected by roughly the amount that Einstein’s theory predicted. (In the Newtonian scheme, the light would’ve been deflected only half as much.) “LIGHTS ALL ASKEW IN THE HEAVENS,” announced the New York Times a few days later. “Men of Science More or Less Agog over Results of Eclipse Observations.”
It was a triumph, or maybe it wasn’t—skeptics would later question whether Eddington hadn’t cooked the books to agree with the theory. The excitement over the results reflected not only fascination with the science but the heartening fact that they’d been achieved through cooperation between English and German scientists, which resonated in the wake of the war. And Albert Einstein, before then unknown outside of his field, became a celebrity. He was surprised by this and assumed his fame wouldn’t last.
What does it take to confirm a theory? In practice, science does not advance as neatly as in the schoolroom version of the scientific method, which implies that the results of experiments all but speak for themselves, ruling thumbs-up or thumbs-down on a hypothesis. It’s often possible to interpret results more than one way: As much as science depends on experimental evidence, it also depends on the building of consensus about what counts as good evidence, and what an experiment means.
In his theories of special and general relativity, Einstein proposed that space and time weren’t the absolutes we think of them as being—if a person on earth could observe what was happening inside a spaceship traveling close to the speed of light, yardsticks would appear to shrink, clocks would seem to slow down. But if you considered three-dimensional space and one-dimensional time together, as four-dimensional space-time, he theorized, there were certain measurable quantities that remained the same for all observers. (Had it been up to him, Einstein would’ve preferred to call it the theory of invariants, rather than the theory of relativity: The point was not, as the layman might mistakenly assume, that everything is relative, truth always dependent on the observer, but that there were in fact certain measurements that remained consistent, no matter the frame of reference.)
Space-time, as it turned out, was not flat and uniform. It buckled and sagged, its four-dimensional topography distorted by massive objects, and its curvature was equivalent to the force of gravity—indeed this was another way of understanding gravity, as the curvature of space-time. Near a massive object like the sun, in this model, the geometry of space-time changes, and light rays travel a different path than they would in empty space. It’s these radical, mind-boggling ideas that Eddington and Dyson found some evidence in favor of—the curvature of space-time!—although Eddington himself remained circumspect in announcing the eclipse findings. He was careful to say that they confirmed “Einstein’s law rather than his theory.” In other words, they agreed with the prediction that light near the sun would be deflected, more so than in the Newtonian model, but didn’t prove that space-time is curved. (And so, were he alive today, Eddington might take issue with the subtitle of Daniel J. Kennefick’s No Shadow of a Doubt: The 1919 Eclipse That Confirmed Einstein’s Theory of Relativity.)
Einstein’s work had plenty of detractors, before and after the eclipse findings were publicized. It was already a transformative period for astronomy, which had formerly been taken up by wealthy amateurs who could erect their own observatories, but was becoming increasingly professionalized and restricted to large institutions—and then here came this counterintuitive theory that required advanced mathematics to fully understand. The old guard didn’t like this, and anti-Semites didn’t like Einstein: In 1920, opponents went so far as to hold an antirelativity rally in the auditorium of the Berlin Philharmonic, and Nazis would condemn “Jewish physics.”
The meaning of the eclipse results would be contested not only by contemporaries, but by those skeptics who looked back on the experiment and decided that it was more or less a case of confirmation bias. Some of the controversy stemmed from the fact that in the course of analyzing the eclipse photographs, Dyson decided to throw out some of the Brazil data. The team there had made plates using two different instruments, and the pictures from one of them were out of focus, likely due to the distorting effect of the sun’s heat on a mirror used in the imaging. But, critics asked, had the data been tossed because it was flawed or because it didn’t yield the answer that general relativity predicted, an answer that (they speculated) Eddington was already biased in favor of? Eddington, a man of great mathematical intelligence and an entirely mysterious personal life, had been such a dedicated pacifist during the war that he’d nearly lost his job, and people would later say that his desire for international cooperation and solidarity influenced him to confirm Einstein’s work.
The idea that there was something dubious about the analysis drifted into conventional wisdom—so much so that Stephen Hawking would write, in his 1988 book A Brief History of Time, that “[Eddington and Dyson’s] errors were as great as the effect they were trying to measure. Their measurement had been sheer luck, or a case of knowing the result they wanted to get, not an uncommon occurrence in science.”
But Hawking, argues Kennefick, had it wrong. Kennefick is a physics professor at the University of Arkansas, Fayetteville, and his book’s purpose is to very thoroughly rebut the skeptics, which he accomplishes in part through a careful and technical review of the instruments, the data, and an astronomer’s 1978 reanalysis of the data using a computer, but also by standing up for Frank Dyson. Eddington would become well known for his public lectures and radio broadcasts about relativity, leaving Dyson in the shadows, but it was Dyson, who’d been unsure about relativity before the eclipse and so probably not biased to find in its favor, who made the decision about which data was valid.
Astronomers no longer need wait for an eclipse to find evidence for light-bending. By the middle of the twentieth century they were able to repeat the same type of experiment using radio telescopes, while now they can collect data from space-based satellites. The detection of gravitational waves—vibrations in space-time generated by accelerating masses, such as a pair of black holes colliding—provided further landmark confirmations of general relativity. In Gravity’s Century, science journalist Ron Cowen gives a brief, accessible account of the 1919 eclipse and subsequent advances in cosmology, touching upon dark matter, dark energy, quantum gravity, and black holes. It’s a very quick and readable introduction to some of the exotic findings that came in Einstein’s wake.
The announcement, in April, that scientists had managed to capture an image of the black hole at the center of another galaxy offered yet another corroboration of the power of Einstein’s theory. That accomplishment relied on data from an international consortium of radio telescopes and involved more than two hundred researchers around the globe. By comparison, the 1919 missions seem poignant and tiny—I now picture those efforts as though from above, a handful of people trapping starlight like kids catching rain in a cup—but no less magnificent.
Karen Olsson’s The Weil Conjectures: On Math and the Pursuit of the Unknown will be published by Farrar, Straus and Giroux in July.
