The old star clusters found in the halos of all large galaxies have long been realized to have mean luminosities that are at least roughly the same everywhere we look. Typical globular clusters have absolute visual magnitudes MV -7 to -7.5, comparable to the brightest supergiant stars; and in giant E galaxies which may contain many thousands of globular clusters, the brightest individual clusters reach luminosities of MV -11, more luminous than any stellar indicators (except supernovae), and detectable at distances of more than 100 Mpc with today's ground-based CCD imaging capabilities. These characteristics are the basis of their attraction as standard candles (see Figure 6).
Figure 6. CCD image of a 1-arcmin field near the center of NGC 3311, the central cD galaxy in Hydra I at a redshift V0 = 3400 km s-1. This image is a 7 x 1000-s exposure in B taken at the CTIO 4-m telescope. Even under average (1") seeing conditions, large numbers of globular clusters are measurable.
The use of clusters for distance measurement is mentioned in passing by Hubble (1932, 1936b), though only in reference to large spiral galaxies (M31, M81, M101) for which they are intrinsically much less suitable. Their first published use in a giant elliptical, their most natural application, seems to be by Baum (1955), who simply compared the brightest globulars in M87 with those in M31. Other attempts were made to match the brightest clusters in the Virgo and Local Group galaxies (Racine 1968; Sandage 1968; van den Bergh 1969; de Vaucouleurs 1970; Hodge 1974), but these were eventually given up when it became clear that the numbers of globular clusters in these galaxies were very different, and that large statistical differences in the true luminosity of the brightest clusters could therefore be expected (but see Sec. 4.6 for a modern recasting of this approach).
The pioneering concepts developed by Racine in the early 1970's suggested that the luminosity function of the entire set of globular clusters in a galaxy had identifiable structure which must hold considerably more distance information than did just the brightest few clusters. Along these lines, the first fully modern use of globular clusters as standard candles is the work of Hanes (1977a, b, 1979), with deep photographic samples of clusters around several Virgo ellipticals. However, further major progress had to await the advent of CCD imaging, which permitted the key Virgo galaxies to be surveyed to much greater depth (van den Bergh et al. 1985; Cohen 1988; Harris et al. 1991), and brought the brightest clusters around far more distant galaxies immediately within reach (Harris 1987a; Thompson and Valdes 1987; Pritchet and Harris 1990).