
The prototype of the classical Cepheids is Delta Cephei, discovered 1784. The changes in brightness were found in the 1890s to be accompanied by and principally caused by changes in stellar temperature and also by changes in radius (see illustration). This was later explained in terms of pulsations in the outer layers of the stars that appear at a late evolutionary state (see pulsating variables); according to theory, the pulsations die down as the star expands and its surface cools.
The relation between period of pulsation and average brightness of Cepheid variables was discovered during 1908–12 by Henrietta Leavitt. Leavitt was studying Cepheids in the Small Magellanic Cloud, which were thus all at about the same distance. She recognized that the relation between their period and their average apparent magnitude was equivalent to a relation between period and average absolute magnitude or luminosity (see distance modulus). An independent determination of the distance of a Cepheid of known period would lead to the graphical relation between period and luminosity. It was quickly realized by Shapley that this period-luminosity relation was an invaluable tool for measurements of distance out to the nearest galaxies and thus for studying the structure of our own galaxy and of the Universe (see distance determination). The refurbished Hubble Space Telescope can study classical Cepheids in galaxies as far away as the Virgo cluster (about 15 megaparsecs). Much work has been done to establish the graph of period versus absolute magnitude, mainly involving independent measurements of the distances or luminosities of Cepheids. Baade and Kukarkin were consequently able to demonstrate in the 1950s the existence of the two distinct categories of Cepheids, having separate but parallel period-luminosity relations.