The perfect standard candle would have the following attributes. 1) It would be extremely bright so that it could be seen at high redshift. 2) It would always have exactly the same absolute magnitude (i.e. zero dispersion). 3) It would be found in great numbers in all types of galaxies. No single distance indicator has all these attributes, hence the proliferation of methods discussed at this conference.
One of the principle advantages of globular clusters (hereafter: GCs) is that they are very bright, overshadowed only by supernovae and entire galaxies as distance indicators. In addition, they are seen in both population I and II objects (e.g. spiral and elliptical galaxies), providing cross calibration between methods which are primarily available for one population type (e.g. Cepheid variables for population I). On the negative side they probably represent the extreme in terms of the dispersion in the value of the magnitude for a single object, which is 1.4 mag. However, this is largely irrelevant since a typical spiral galaxy has roughly a hundred globular clusters, an elliptical galaxy roughly a thousand, and a cD galaxy often over ten thousand globular clusters. Hence, in principle, the mean of the distribution can be estimated with uncertainties 0.14 mag (i.e. 1.4 - 100) for spirals, 0.04 mag for ellipticals, and 0.014 mag for cD galaxies. The intrinsic dispersion of M0V is 0.1 mag, making it competitive with the best of the distance indicators. At present the limiting factors are the local calibrators, the question of universality, and, for the estimate of the Hubble constant, the effects of peculiar velocities.
From an observational standpoint the GCLF has several advantages over many of the other distance indicators. For example, only a single exposure is needed to make the measurement rather than a series of perhaps a dozen separate observations spread over several weeks or months, as required for Cepheids, RR Lyrae variables, novae, and supernovae. Another advantage is that the most precise measurements can be made in the one galaxy where the question of whether the galaxy is a foreground or background object is not an issue, namely cD galaxies. This is one of the primary difficulties with population I methods. Finally, the number of high-quality GCLF's is rapidly increasing, thanks in part to the fact that the surface brightness fluctuation method of distance estimation requires identifying and removing globular clusters from their images (e.g. Tonry, Ajhar, Luppino 1990, Ajhar, Blakeslee, & Tonry 1994).
For these reasons it seems certain that the use of the GCLF to determine distances will become increasingly important in the coming decade. The plan for this paper is to review the current situation and provide an updated compilation of high quality GCLF measurements for bright elliptical galaxies (Section 2), examine the primary uncertainties and estimate the errors (Section 3), take a brief look at theoretical considerations (Section 4), make an estimate of the Hubble constant based on new observations in the Fornax cluster (Section 5), and summarize the discussion and suggest future directions (Section 6).
There are already excellent reviews on the general characteristics of globular clusters in galaxies beyond the local group (Harris 1991) and the use of globular clusters for measuring distance (Jacoby et al. 1992, van den Bergh 1992). These references can be used to backtrack earlier reviews (e.g. Harris & Racine 1979). An article by Harris (1996) at this conference provides more detail about GCLFs in spirals and dwarf ellipticals while a recent review by Larson (1996) covers the formation of globular clusters. The current review will focus on the past five years with special attention on GCLFs in bright ellipticals.