4.6 Future Needs and Directions

Aside from its presently incomplete calibration, the principal observational problem associated with the GCLF method is that it can be applied effectively only to giant ellipticals. This feature is perhaps best regarded as a limitation rather than a weakness, since the gE's are most often found in the rich clusters that are the main landmarks in Hubble velocity space. But for any types of irregular and late-type spiral galaxies, it is unlikely that GCLFs will ever turn out to be useful standard candles.

On the theoretical side, prescriptions for future work are simple to state but quite challenging to execute. We need a theory of cluster formation specific enough to predict their complete mass spectrum and not just a mean globular cluster size. Following this, the dynamical interaction of a system of clusters with its parent galaxy environment needs to be well enough modeled to describe how the initial cluster mass distribution evolved into what we see today. Matching the models to the observations for galaxies of widely different types can then begin in earnest.

On the observational side, the most pressing need is to calibrate GCLFs for selected large disk and elliptical galaxies which are close enough for the turnover to be clearly measured. The distances to these must, of course, be established independently. Requirements for the spirals are particularly stringent (they must be almost perfectly edge-on so that the halo clusters are clearly visible). For nearby giant Sa/Sb disk galaxies, especially important systems for which new GCLF data have been produced include NGC 4594 (Bridges and Hanes 1992), NGC 4565 (Harris et al. 1992), and M31 itself (Reed et al. 1992); other potentially valuable candidates are listed by Harris et al. (1988a). The edge-on S0 NGC 3115 will also be of interest, though it has a rather sparse cluster system. Well situated, nearby large E galaxies are similarly scarce; the most interesting ones for extensive GCLF measurement may be NGC 3377 and 3379 in Leo (Pritchet and van den Bergh 1985b; Harris 1990), and NGC 4278 and 4494 in the Coma I group (Gregory and Thompson 1977). Adding these systems to the calibration should eventually lead to a fully self-consistent GCLF determination of the distances to Virgo and Fornax, the nearest rich galaxy clusters.

Once Virgo and Fornax are secured, the absolute distances to giant ellipticals in other systems at V₀ 4000 km s^-1 including Hydra I, Centaurus, Pegasus, and Perseus will then follow: these are close enough that the bright half of the GCLF is fully measurable (Harris 1988a), and thus can be fitted using the E-galaxy GCLF as a fiducial marker as described above. Finally, HST imaging with the WF/PCII camera has the capability to extend the reach of the GCLF technique by about 2 magnitudes more. With it, photometry to the turnover level can be obtained out to the Coma Cluster members (at V₀ ~ 7000 km s^-1) and slightly beyond.

There are also prospects for using the bright end of the GCLF as a secondary standard candle to continue much further out. The basis of the idea has been pointed out by Hanes and Whittaker (1987) and can be seen from Figure 7. If the GCLF is already well populated, as it is in a typical giant elliptical, then (m) falls off so steeply at the bright end that further increases in the cluster population do not change the level of the brightest few clusters very much. When one images a distant galaxy, the effect is very much like that of the planetary-nebula LF technique: the globular clusters ``switch on'' rather suddenly as a critical magnitude level is reached. Near the top end (M_V -10), it may be shown easily that M_V (n) 0.4 M_V^T where M_V (n) is the mean luminosity of the brightest few clusters (n ~ 10-20) and M_V^T is the galaxy luminosity. The brightest ellipticals in clusters of galaxies have an intrinsic luminosity dispersion (M_V^T) ± 0.4 mag (e.g. Sandage 1973b). Then, even if the cluster specific frequencies of giant E galaxies vary by factors of ~ 3 (Pritchet and Harris 1990), the parameter M_V(n) may act as a standard candle to the ± 0.5-mag level of precision. With ground-based photometry and CCD imaging at half-arcsec quality now conventional at some sites, the brightest clusters will be detectable in giant E galaxies at V₀(lim) 12,000 km s^-1. With the HST, the ultimate limits of the brightest-cluster technique should extend to about 7 magnitudes more distant than the Virgo Cluster, or a redshift V₀ 35,000 km s^-1. The method may then provide an interesting complement to the Tully-Fisher and D_n- techniques, which function over the same distance ranges.