One law now explains both the heavens and the Earth. The English physicist and mathematician Isaac Newton publishes his Philosophiæ Naturalis Principia Mathematica. For the first time, a single law of gravity explains how the planets and the Moon move, why there are tides, and why things fall to the ground.

Mercury isn’t quite where Newton’s law says it should be. Its orbit is an oval, and the point where it comes closest to the Sun is called its perihelion. Over the centuries, that point slowly moves around the Sun. The pull of the other planets explains most of this movement, but the French astronomer Urbain Le Verrier finds that some is left over: about 38 arcseconds per century. An arcsecond is a tiny angle, 1/3,600 of a degree.

Le Verrier has seen something like this before: in 1846, the odd motion of Uranus led him to predict an unknown planet, and the German astronomer Johann Gottfried Galle found Neptune right where he said. So Le Verrier suspects unseen matter inside Mercury’s orbit: a planet, or a ring of small bodies.

The planet Le Verrier predicted is nowhere to be found. He named it Vulcan in 1860, after an amateur astronomer reported seeing it cross the Sun, but no sighting held up. A planet that close to the Sun is lost in its glare, except during a total solar eclipse, when the Moon blocks the Sun. So the American astronomer Charles Dillon Perrine photographed the sky around the Sun during the eclipses of 1901, 1905, and 1908. He found nothing big enough. Now the director of Lick Observatory, the American astronomer W. W. Campbell, declares the search over: there is no such planet. The German astronomer Hugo von Seeliger has suggested something fainter instead: a thin cloud of matter around the Sun, like the one that glows as the zodiacal light, too dim to photograph.

Can gravity bend light? If light is made of tiny particles, as Newton thought, it should fall like anything else. In 1801, the German astronomer Johann Georg von Soldner used Newton’s law to work out how much: starlight passing the edge of the Sun would be bent by 0.84 arcseconds. Now the German-born physicist Albert Einstein, a professor in Prague, works it out from an early version of a new theory of gravity, which he is still developing. He gets almost the same: 0.83 arcseconds. He asks astronomers to test it during a total solar eclipse, when stars near the Sun can be photographed.

A world war gets in the way of the test. Astronomers from several countries travel to the Russian Empire for the total solar eclipse of August 21, 1914, and three of the teams mean to test Einstein’s prediction. But the First World War has begun weeks earlier. Russia arrests the German team in Crimea, led by the German astronomer Erwin Freundlich, and takes their instruments. Freundlich is sent home weeks later in a prisoner exchange.

The other two teams are free to work, but clouds hide the Sun. One is Perrine’s team from Argentina, also in Crimea. The other is Campbell’s team from Lick Observatory near Kiev, with the American astronomer Heber Curtis. Einstein’s prediction stays untested. That gives him time to find a mistake in it: until late 1915, his calculation leaves out how gravity curves the space around the Sun.

Einstein solves the mystery of Mercury, with no hidden planet. At the Prussian Academy of Sciences in Berlin, he applies his new theory of gravity, general relativity, to Mercury. In it, gravity isn’t a force but the bending of space and time by matter. The theory makes Mercury’s perihelion move 43 arcseconds per century more than Newton’s law does. Astronomers by now put the leftover at 45 arcseconds, give or take 5, so that fits.

The same paper fixes a mistake in Einstein’s 1911 prediction. His old calculation left out how gravity bends space. Counting it doubles the effect: the Sun should bend starlight by 1.7 arcseconds. That’s twice his old number, and twice what Newton’s law gives. A week later, Einstein presents the final equations of his theory.

The war goes on, but in England, someone is planning ahead. The English astronomer Frank Dyson, the Astronomer Royal, points out that the total solar eclipse of May 29, 1919, is a rare chance to test Einstein’s prediction. The darkened Sun will stand in front of the Hyades, a cluster of bright stars. No other part of the Sun’s path through the sky has so many. With so many stars close to the Sun, photographs can show whether their light is bent, and by how much: by Einstein’s 1.7 arcseconds, by the half of that Newton’s law gives, or not at all.

The eclipse photos are in, and they side with Einstein. During the total solar eclipse of May 29, 1919, two teams of British astronomers photographed the stars around the darkened Sun. Arthur Eddington went to the island of Príncipe, off West Africa. Andrew Crommelin and Charles Davidson went to Sobral, in Brazil. Dyson organized both trips. Now, at a joint meeting of the Royal Society and the Royal Astronomical Society in London, they report the results.

The starlight was bent by 1.98 arcseconds, as measured with one of the two telescopes at Sobral, and 1.61 at Príncipe. Measurements this fine are never exact, but both are close to Einstein’s 1.7, and far from the half of it that Newton’s law gives. The main Sobral telescope gave a smaller value, but its images were distorted, the astronomers judged, by the Sun’s heat on its mirror, and they set it aside. Historians still argue about that choice.

Weeks earlier, when the first news reached Einstein in a telegram, he showed it to a philosophy student, Ilse Schneider. She asked what he would have said if the eclipse had disagreed with him. As she recalled decades later, he answered: “I would have had to pity our dear God. The theory is correct all the same.”

But a deeper problem arises: gravity may not fit with quantum physics. Quantum theory describes the world of atoms and light, where energy comes in tiny packets. In 1928, the English physicist Paul Dirac found an equation for the electron that obeys both quantum theory and Einstein’s special relativity, the theory of motion at high speeds. Dirac’s theory predicted a new particle, the positron, which the American physicist Carl Anderson found in 1932.

Now the Soviet physicist Matvei Bronstein tries to do the same for gravity, and he finds a problem. Unlike electric and magnetic fields, gravity can’t be measured precisely in a small enough region of space. He concludes that a quantum theory of gravity may call for “the rejection of our ordinary concepts of space and time.”

To this day, physicists have proposed theories that unite quantum physics and general relativity, like string theory, but none has yet passed a test that could have refuted it.

Einstein passes a test 27,000 light-years away. At the center of our galaxy, the Milky Way, sits a huge black hole called Sagittarius A*. A star named S2 orbits it every 16 years. Astronomers have followed S2 for 27 years, lately with the GRAVITY instrument on the Very Large Telescope in Chile. The GRAVITY collaboration, an international team of astronomers, reports that its orbit turns a little each time around, by about 12 arcminutes. An arcminute is 60 arcseconds. That’s just what general relativity predicts. Under Newton’s law, a lone star orbiting a lone mass would follow the same oval forever.