The 1919 eclipse expeditions: measuring bent starlight

General relativity predicted that starlight passing close to the Sun would change direction under its gravity. Ordinarily the Sun’s brightness hides these stars. A total eclipse briefly blocks the bright solar surface and permits photographs.

The eclipse of May 29, 1919, occurred against a field of bright stars. Frank Dyson organised observations, with Eddington and colleagues preparing to compare eclipse positions with photographs taken when the Sun was elsewhere.

References: [1]

Crommelin and Davidson observed from Sobral in Brazil, while Eddington and Cottingham travelled to Príncipe off Africa’s west coast. The two sites supplied separate sets of observations of the same eclipse.

Instruments had to be shipped, assembled and calibrated. The displacement was tiny on the photographic plates; weather, temperature and focus affected the star images, and comparison plates were also needed.

References: [1]

Cloud affected Príncipe, though usable star images were obtained. The two instruments at Sobral produced images of differing quality. Back in Britain, the team measured the plates and accounted for changes in instrumental scale.

The usable observations supported the magnitude predicted by Einstein, about 1.75 arcseconds at the Sun’s limb. Results were announced at a scientific meeting in November 1919 and published in detail in 1920.

References: [1]

The expedition turned a theoretical prediction into an observable quantity: select a field, record positions, compare photographs and evaluate gravitational models. Its paper documented equipment, weather, plate reduction and calculations for subsequent examination.

Light deflection was measured again in later eclipses and more precise observations, while gravitational lensing became a tool for studying distant galaxies. The 1919 expeditions preserved an early observational test of general relativity.

References: [1]