6.1 Globular Cluster Ages

There are also many excellent recent reviews covering in great detail the ages obtained for Galactic globular clusters (i.e., from a comparison of observed color magnitude diagrams and theoretical evolution models). At the moment, there is a fairly broad consensus that Galactic globular clusters are most likely at least 14-15 Gyr old (e.g. Chaboyer et al. 1996; VandenBerg et al. 1996; Shi 1995).

It is not widely appreciated that the largest uncertainty in the globular-cluster ages results from uncertainties in the distances to the globular clusters, which currently are based on statistical parallax measurements of Galactic RR Lyrae stars or on parallaxes for nearby subdwarfs(e.g. Renzini 1991; Chaboyer et al. 1996; VandenBerg et al. 1996). Although the ages of globular clusters are widely regarded as theoretically-determined quantities, in the process of determining ages, it is still necessary to interface theory with observation and transform the observed globular cluster magnitudes to bolometric luminosities (via an accurate distance scale). The subdwarf and RR Lyrae statistical parallax distance calibrations currently differ by about ~ 0.25-0.30 mag. Unfortunately, as emphasized by Renzini, small errors in distance modulus (0.25 mag or 13% in distance) correspond to 25% differences in age. Even with improved parallax measurements (for example, soon to be available from HIPPARCHOS), there are many subtle issues (e.g., reddening, metallicity, photometric zero-points) that combine to make it a very difficult problem to achieve distances to better than 5% accuracy.

As discussed previously in many contexts (e.g. Walker 1992; Freedman & Madore 1993; van den Bergh 1995, and most recently by Feast & Catchpole 1997), there is also currently a discrepancy in the Cepheid and RR Lyrae distances to nearby galaxies. If the Cepheid distances are correct, it would imply that the absolute magnitudes of RR Lyraes are brighter (by about 0.3 mag) than suggested by statistical parallax and Baade-Wesselink calibrations for Galactic RR Lyraes (e.g. see VandenBerg, Bolte & Stetson 1996 for a recent discussion). This brighter RR Lyrae calibration agrees well in zero point with that from Galactic subdwarfs. Based on the models of VandenBerg et al. 1997, applying this calibration (adopting M_V(RR) = 0.40 mag) to the metal-poor globular cluster M92, results in an age of 15.8 ± 2 Gyr. If the fainter RR Lyrae distance scale is correct, the age derived for M92 based on these same recent models increases to ~ 19 Gyr. Alternatively, if the Feast & Catchpole calibration of Galactic Cepheids based on HIPPARCHOS parallaxes is correct, then the resulting RR Lyrae calibration is even brighter (M_V(RR) = 0.25 at [Fe/H] = -1.9), and the corresponding age for M92 would be reduced to about 13 Gyr (based on the same Vandenberg models). A new calibration of Galactic metal-poor subdwarfs, also based on new HIPPARCHOS parallaxes, appears to confirm these younger ages (Reid, private communication). It is interesting to note that while the distances to nearby galaxies have converged to a level where they no longer have a factor-of-two impact on the Hubble constant, subtle differences of only a few tenths of a magnitude in distance modulus can still have very significant impact on cosmology, through the ages determined from stellar evolution.