Several speakers at this symposium have alluded to the zeroth-order agreement between the Type II supernovae (SNe) stellar yields, as predicted by the models of those most responsible for driving progress in the field - i.e. Arnett (1991,1996); Maeder (1992); Woosley & Weaver (1995); Langer & Henkel (1995); Thielemann et al. (1996), hereafter referred to as A91, A96, M92, WW95, LH95, and TNH96, respectively. It is important though for those entering (or indeed, already involved in!) the chemical evolution field to be cognizant of the fact that there are important first- and second-order differences between the yield compilations.

Within collapsing protogalaxies, thermal instability leads to the formation of a population of cool fragments which are confined by the pressure of a residual hot background medium. The critical mass required for the cold clouds to become gravitationally unstable and to form stars is determined by both their internal temperature and external pressure. Massive first generation stars form in primordial clouds with sufficient column density to shield themselves from external UV photons emitted by nearby massive stars or AGNs. Less massive photoionized clouds gain mass as they undergo cohesive collisions with each other, and lose mass due to ram pressure stripping by the residual halo gas. Collisions may also trigger thermal instability and fragmentation into cloudlets. While most cloudlets have substellar masses, the largest become self-gravitating and collapse to form protostellar cores without further fragmentation. The initial stellar mass function is established as these cores capture additional residual cloudlets. Energy dissipation from the mergers ensures that the cluster remains bound in the limit of low star formation efficiency. Dissipation also promotes the formation and retention of the most massive stars in the cluster center. On the scale of the protogalactic clouds, the formation of massive stars generates intense UV radiation which photoionizes gas and quenches star formation in nearby regions. As gas density accumulates in the center of the the galactic potential, the self regulated star formation rate increases. At the location where most of the residual gas can be converted into stars on its internal dynamical timescale, a galaxy attains its asymptotic kinematic structure such as exponential profiles, Tully-Fisher, and Faber-Jackson laws.

The merger model for elliptical galaxy formation has received increasing attention since it was first suggested by Toomre & Toomre (1972). Van den Bergh (1984) pointed out a problem with the idea that elliptical galaxies were formed by simply combining two, or more, spiral galaxies. He noted that the specific frequency (SN, number of globular clusters per unit galaxy light) is systematically lower for spirals than for ellipticals. Schweizer (1987) suggested that globular clusters might be expected to form in the merger process, thereby alleviating or possibly eliminating the SN problem. Ashman & Zepf (1992) developed this idea into a merger model for globular cluster formation with testable predictions.

I would like to focus on one aspect regarding the evolution of Galactic stellar populations that is particularly relevant to discussions at this symposium: Where were the sites of early star formation in the Galaxy? The large scatter in abundance ratios for metal poor stars suggests multiple early settings of star formation in the Milky Way. In this and other ways, interpretation of detailed stellar chemical abundance analyses are converging with those of spatial-kinematical analyses of field stars, star clusters and satellite galaxies.

Chemical evolution studies of the Clouds based on detailed abundance analyses are essentially limited nowadays to young objects like HII regions (Dufour 1984; Garnett et al. 1995), planetary nebulae (e.g. de Freitas Pacheco et al. 1993), supernova remnants (Russell & Dopita 1990) and supergiant stars either of F-K type (Hill et al. 1995) or B ones (e.g. Rolleston et al. 1996 and references therein). One exception to this was presented by Dopita et al. (1997) where HST data permitted to date planetary nebulae nuclei and to present an age vs. metallicity model.

We outline a method to infer the global history of star formation in galaxies from absorption-line observations of quasars. The application to existing data leads to the conclusion that most stars formed at relatively low redshifts (z ≲ 2). The results obtained by this absorption-based method are consistent with those obtained subsequently by emission-based methods.