We present the mass-metallicity relation foreseen by our self-enrichment model. At first glance, such a globular cluster formation scenario is disproven by the lack of any obvious correlation between the luminosity and the metallicity of globular clusters, as claimed by previous authors (e.g. Djorgovski & Meylan 1994; Ashman & Zepf 1998). However, we stress here that there are numerous sources of scatter between the theoretical (mass, [Fe/H]) relation, applying to the gaseous progenitors of globular clusters, and the observed (luminosity, [Fe/H]) plot, applying to globular clusters. These scatter sources are, for instance, the star formation efficiency with which the globular cluster stars form out of the proto-globular cluster cloud, the mass losses undergone by globular clusters with time (see the tidal tails exhibited by some clusters), the variations in the mass-to-light ratio from one globular cluster to another. We also caution that the search for a mass-metallicity correlation should be restricted to a given globular cluster subpopulation, namely the Old Halo group. Despite the numerous sources of scatter, the globulars of this group are characterized by a boundary in the mass-metallicity diagram and by a correlation in the sense expected by simple pressure-bounded self-enrichment models. Self-enrichment models remain a viable hypothesis for galactic halo globular cluster formation.
Acknowledgements
This research was supported partly by the European Community under grant HPMT-CT-2000-00132 and partly by contracts Pôle d'Attraction Interuniversitaire P4/05 (SSTC, Belgium) and FRFC F6/15-OL-F63 (FNRS, Belgium).
Copyright ESO 2001