
The sky over Sobral was overcast that morning, which was terrible news for men who had crossed an ocean to photograph stars in the daytime. On 29 May 1919, Charles Davidson and Andrew Crommelin of the Royal Greenwich Observatory had exactly one chance - five minutes and thirteen seconds of totality - to help settle an argument about the shape of the universe. Then the clouds pulled apart. The plates came out. And a cattle-and-cotton town in the Ceará backlands became, for a few weeks that November, the most consequential dot on the map of physics.
Einstein's general theory of relativity, published in 1915, made a prediction that could be checked: starlight grazing the Sun should bend by about 1.75 arcseconds, roughly twice the 0.87 arcseconds that Newtonian gravity allowed. Checking it required photographing stars whose light passed close to the Sun, which is possible only during a total eclipse. The eclipse of 29 May 1919 was extraordinarily favourable. It was long - nearly seven minutes at maximum along its track - and the darkened Sun would sit against the Hyades, a dense, bright cluster in Taurus. The Astronomer Royal, Frank Watson Dyson, sent two expeditions to catch it. Arthur Eddington went to Príncipe, off the West African coast. Davidson and Crommelin went to Sobral, where a Brazilian team under Henrique Morize - Lélio Gama, Domingos Fernandes da Costa, Allyrio Hugueney de Mattos, Teófilo Lee - set up alongside them. On Príncipe that morning it stormed.
Sobral had two instruments, both fed by coelostats - driven mirrors that hold the moving sky steady on a fixed lens. The main rig paired a 13-inch astrographic telescope with a 16-inch coelostat shipped from Greenwich. The backup was a modest 4-inch lens borrowed from the Jesuit astronomer Aloysius Cortie. The big instrument caught twelve stars, but its images came out blurred, and the deflection they implied fell closer to the Newtonian figure than to Einstein's. The small one caught seven stars, sharply, and gave a deflection slightly above Einstein's prediction. The blurred astrographic plates were set aside. The 4-inch plates - clean, unambiguous, and taken with the humblest telescope on the site - were judged the trustworthy ones. Eddington's cloud-hampered plates from Príncipe leaned Einstein's way too, but they were thin evidence on their own.
At a joint meeting of the Royal Society and the Royal Astronomical Society, Dyson delivered the verdict: "After a careful study of the plates, I am prepared to say that they confirm Einstein's prediction." The newspapers took it from there. Within days a German theorist barely known outside physics departments had become the most famous scientist alive, and the transformation is usually traced to that meeting - which is to say, in large part, to seven star images recorded above a Brazilian town most of the reporters could not place on a map. The story has never fully settled. Critics have argued for decades that discarding the astrographic plates was convenient rather than principled. But Daniel Kennefick has made the case that without the Sobral photographs the 1919 result would have been inconclusive altogether, and a 1979 reanalysis of plates from both Sobral instruments, using modern measuring equipment, supported the original conclusion. Later work suggests Eddington's handling of the data was less biased than the accusation assumes. The historiography is genuinely contested; the plates themselves are not.
Sobral was already old by 1919. The Areriús people lived along the Rio Acaraú before the settlement of Caiçara grew up around a pair of churches on a cattle ranch, and on 5 July 1773 that settlement was raised to Vila Distinta e Real de Sobral. Its wealth came from charque - beef salted and dried for export - and by the late eighteenth century the charqueadas had made Sobral richer than Fortaleza. The nineteenth century brought harder arithmetic. The town opened its first cemetery, São José, in 1861, once burials beside the churches were prohibited. The catastrophic droughts of 1877-1879 emptied the sertão and, in their aftermath, brought the Estrada de Ferro de Sobral: a railway whose station opened in 1882 to move highland produce down to the coast at Camocim. Cotton mills followed in the twentieth century.
The town has not let the eclipse go. A monument in Praça do Patrocínio marks where the instruments stood; the Museu do Eclipse was built beside it in 1999, and a planetarium was inaugurated on the same square in 2015. Sobral counted 203,023 residents at the 2022 census, making it the fifth-largest municipality in Ceará after Fortaleza, and it now carries two public universities and more than a hundred schools. Its exports run to comedians and singers as well as data - Renato Aragão and Belchior both grew up here. The setting is unforgiving: 70 metres above sea level on a plain that bakes toward 37°C, taking around 888 millimetres of rain a year and taking almost all of it in March and April. The rest of the year, the sky is very reliably clear. Which, in 1919, turned out to matter enormously.
Sobral sits at 3.686°S, 40.344°W on the Ceará plain, 70 m above sea level, with the wall of the Serra da Ibiapaba rising to around 950 m about 70 km to the west - the most useful visual reference in the area. Cel. Virgílio Távora Airport (SNOB / QBX) lies 2 km from the centre at 64 m elevation with a single 1,033 m asphalt runway, 10/28; it ceased commercial service on 26 February 2023 when flights moved to the newer Sobral Regional Airport. Fortaleza's Pinto Martins International (SBFZ / FOR, 25 m) is the nearest major field, roughly 200 km east on the coast. Recommended sightseeing altitude 3,000-5,000 ft AGL for the historic core and Praça do Patrocínio. Visibility is generally excellent: annual rainfall is only about 888 mm and is concentrated in March and April, so expect haze and convective buildups in those two months and clear, hot, high-visibility conditions for most of the rest of the year.