Systematic Problems With the Original Eddington Experiment of 1919
Abstract
The observation that starlight passing near the Sun was bent as predicted by Einstein's General Relativity during the 1919 total solar eclipse was one of the great triumphs and turning points of all science. The headline at the top center of the New Times front page declared "Men of Science More or Less Agog Over Results of Eclipse Observations, Einstein Theory Triumphs". Still, there were substantial problems and uncertainties in the decade after 1919 as researchers sought to verify and duplicate Eddington's results. And there have been persistent whispering rumors and claims that Eddington somehow 'fudged' his data. These rumors are certainly wrong, as proven by modern measures and analysis of the original plates, as well as by detailed examinations of the documented analysis and discussions.With modern knowledge, we can recognize a variety of imperfections in the analysis of Eddington and Dyson. Most of these (e.g., differential aberration, color-dependent refraction) are too small to affect Eddington's basic claim with ~10% accuracy in the Einstein shift. The biggest systematic error, even in modern times, is a degeneracy between the plate scale and the Einstein shift. With only stars measured at similar distances from the Sun, we cannot distinguish between the case where there is the predicted Einstein shift outwards plus the correct plate scale and the case where there is no Einstein shift plus a very slightly mis-measured plate scale. Plate scales cannot be measured perfectly, and plate scales will change on many time scales, as the observer refocuses and as the changing ambient temperature changes the telescope tube length.For Eddington's 1919 plates, they used only a handful of stars, measured only in one direction, all at similar distances from the Sun, so they are in the case where the degeneracy can dominate the derived measure for the Einstein shift. Further, their plate scales were taken with plates exposed a continent away, months later, and at greatly different temperatures. So, how could Eddington have gotten what we now know to be the correct measure of the Einstein shift - despite not having solved the degeneracy problem?
- Publication:
-
American Astronomical Society Meeting Abstracts #230
- Pub Date:
- June 2017
- Bibcode:
- 2017AAS...23011903S