Annu. Rev. Astron. Astrophys. 1996. 34: 511-550 Copyright © 1996 by Annual Reviews. All rights reserved

2.3. Age of the Oldest Clusters

Walker (1992b, Figure 9) and Da Costa (1993, Figure 2) have shown that LMC clusters have a redder HB morphology than expected for their metallicities. Zinn (1993), van den Bergh (1993) have similarly shown the division of Galactic halo clusters into two major families in this metallicity-HB distribution plane. This division appears to be another manifestation of the "second parameter" effect in Galactic clusters and dSph galaxies. It is now common to associate this effect with an age difference of 2-3 Gyr (Lee et al 1994) between the two families of halo clusters. In our Galaxy, this effect is also accompanied by a kinematical signature (Rodgers & Paltoglou 1984, Chaboyer et al 1992, van den Bergh 1993, Zinn 1993, Da Costa & Armandroff 1995). If age is truly the second parameter determining the HB morphology, after metallicity, then we expect that the ancient LMC clusters will prove to be some 2 Gyr younger in the mean than old halo Milky Way clusters.

Suntzeff et al (1992) have used the data in Table 1 to argue that the old LMC cluster population is very similar to the Galactic cluster population outside the solar circle in terms of metallicity, cluster luminosity, number of RR Lyraes per unit cluster luminosity, and the ratio of luminous cluster mass to field star mass. The number of clusters in the LMC also is consistent with the luminosity difference between the LMC and the Galaxy. This similarity is somewhat surprising given that the dynamical mass of the Milky Way is 10 times that of the LMC, although the dynamical mass of the LMC is likely to be an underestimate (Suntzeff et al 1992).

If we scale the LMC to the SMC luminosity, we would expect the SMC to have about two old clusters: In Table 1, we find one old cluster with RR Lyrae and two substantially younger clusters. Little can be said of the general properties of the old clusters in the SMC due to small number statistics.

It is tempting to try to use the data in Table 1 (and the HB magnitudes given in Suntzeff et al 1992) to determine a M_V-[Fe/H] relationship for RR Lyraes. This relationship is an important and controversial one. We discuss it further in Section 2.5. The published cluster metallicities, however, have more scatter than can be tolerated for this problem - they are typically based on only one or two spectra or on the color of giant branches in the CMD. Even with better metallicities the relationship cannot be determined due to the small range in metallicities in these particular clusters and the unknown position of any cluster in front or behind the galaxy. It had been thought that the HB magnitude as a function of age was constant for stars older than a few Gyr (e.g. Olszewski et al 1987), allowing the metallicity effects to be derived by studying all LMC and SMC clusters older than a few Gyr. Recent results on the Carina dwarf galaxy (Smecker-Hane et al 1994; PB Stetson 1996, private communication) show, however, that 7- and 14-Gyr populations at [Fe/H] = -2 produce HB luminosities that differ by 0.25 mag, implying that age effects are as important as metallicity effects.

A more promising approach would be to search the large number of LMC RR Lyraes discovered by the MACHO project for large-amplitude, short-period RRab variables, which will tend to be metal rich. A few such stars are known in the MCs (Hazen & Nemec 1992), though they are clearly quite rare. Such a metal-rich sample, coupled with the much more numerous metal-poor RR Lyraes, should yield a definitive M_V - [Fe/H] relationship.