KentuckyFC (1144503) writes "In the early hours of the morning on 24 February 1987, a neutrino detector deep beneath Mont Blanc in northern Italy picked up a sudden burst of neutrinos. Three hours later, neutrino detectors at two other locations picked up a second burst. These turned out to have been produced by the collapse of the core of a star in the Large Magellanic Cloud that orbits our galaxy. And sure enough, some 4.7 hours after this, astronomers noticed the tell-tale brightening of a blue supergiant in that region, as it became a supernova, now known as SN1987a. But why the delay of 7.7 hours from the first burst of neutrinos to the arrival of the photons? Astrophysicists soon realised that since neutrinos rarely interact with ordinary matter, they can escape from the star's core immediately. By contrast, photons have to diffuse through the star, a process that would have delayed them by about 3 hours. That accounts for some of the delay but what of the rest? Now one physicist has the answer--the speed of light through space requires a correction. As a photon travels through space, there is a finite chance that it will form an electron-positron pair. This pair exists for only a brief period of time and then goes on to recombine creating another photon which continues along the same path. This is a well-known process called vacuum polarisation. The new idea is that the gravitational potential of the Milky Way must influence the electron-positron pair because they have mass. This changes the energy of the virtual electron-positron pair, which in turn produces a small change in the energy and speed of the photon. And since the analogous effect on neutrinos is negligible, light will travel more slowly than them through a gravitational potential. According to the new calculations which combine quantum electrodynamics with general relativity, the change in speed accounts more or less exactly for the mysterious time difference. Voila!"