Stargas Models of the Universe and the Rise of Statistical Astronomy

At the turn of the 20th century, William Thomson (Lord Kelvin), inspired by kinetic gas theory, calculated the dimensions of the universe based on stellar velocities in the vicinity of the Solar System, giving rise to "stargas" models of star clusters -- and of the universe -- pursued from 1904 to the early 1920s by J. C. Kapteyn, Henri Poincaré, Arthur Eddington, Karl Schwarzschild, James Jeans, C. V. L. Charlier and Albert Einstein. The attraction of stargas models, and subsequent formation of statistical astronomy as a subfield of astrophysics, is clarified by the correspondence of Kapteyn and Schwarzschild, in particular. Stargas models of the universe, including Kapteyn's island universe, did not stand up well against observations afforded by the big new North-American telescopes, as E. R. Paul pointed out in 1981. However, the demise of stargas cosmological models in the 1920s did not spell the end of stargas models of star and galaxy clusters. On the contrary, the theorems and methods introduced in this context served as the foundation for stellar dynamics in later decades.