Global evolutionary trends in galaxy evolution over cosmic times (e.g., Madau et al. 1996) are best inferred from individual or combined galaxy surveys covering a wide redshift range. Marchesini et al. (2009) derive the redshift evolution of the global stellar mass density for galaxies with stellar masses in the range of 108 < M* / M < 1013. Galaxies with lower stellar masses do not appear to contribute significantly to the mass density budget. Marchesini et al. find that approximately 45% of the present-day stellar mass was generated from 3 > z > 1 (within 3.6 Gyr). From z ~ 1 to the present, i.e., during the last ~ 7.5 Gyr, the remaining 50% were produced.
Considering cosmic star formation histories, there is compelling evidence for "downsizing" (Cowie et al. 1996): The stars in more massive galaxies usually formed at earlier cosmic epochs and over a shorter time period. For a pedagogical illustration, see Fig. 9 in Thomas et al. (2010). High-redshift galaxies tend to have star formation rates higher than those found in the local Universe. Also, these galaxies are typically more massive than low-redshift star-forming galaxies (Seymour et al. 2008).
Juneau et al. (2005) show that the star formation rate density depends strongly on the stellar mass of galaxies. For massive galaxies with M*> 1010.8 M the star formation rate density was about six times higher at z = 2 than at the present, remaining approximately constant since z = 1. The star formation rate density in their "intermediate-mass" bin, 1010.2 M* / M 1010.8 M, peaks at z ~ 1.5, whereas since z < 1 most of the activity was in lower-mass galaxies (Juneau et al. 2005).
For quasar host galaxies observed with the Herschel satellite, Serjeant et al. (2010) find that high-luminosity quasars have their peak contribution to the star formation density at z ~ 3, while the maximum contribution of low-luminosity quasars peaks between 1 < z < 2. The authors suggest that this indicates a decrease in both the rate of major mergers and in the gas available for star formation and black hole accretion.
3.2. Mass-metallicity relations
From an analysis of ~ 160,000 galaxies in the local Universe observed by the Sloan Digital Sky Survey (SDSS), Gallazzi et al. (2008) conclude that approximately 40% of the total amount of metals contained in stars is located in bulge-dominated galaxies with predominantly old populations. Disk-dominated galaxies contain < 25%, while both types of galaxies contribute similarly to the total stellar mass density.
Panter et al. (2008) derived the mass-fraction-weighted galaxian metallicity as a function of present-day stellar mass based on the analysis of > 300,000 galaxies from the SDSS. Panter et al. find a flat relation with little scatter (~ 0.15 dex) around ~ 1.1 Z for galaxies with masses 1010.5 M. Below ~ 1010 M there is a clear trend of decreasing metallicity of about 0.5 dex per dex decline in mass, although the dispersion is relatively large (~ 0.5 dex). When considering only (the very few) galaxies in which more than 50% of the light comes from populations younger than 500 Myr, only systems with stellar masses < 1010 M contribute significantly - another indication of downsizing.
3.3. Environmental trends
Galaxies in high-density regions are generally found to have higher metallicities than those in low-density regions (see, e.g., Panter et al. 2008). Sheth et al. (2006) find that galaxies with above-average star formation rates and high metallicities at high redshifts are situated mainly in galaxy clusters in the present-day Universe. Interestingly, Poggianti et al. (2010) infer that high-redshift clusters were denser environments with respect to both galaxy number and mass than contemporary clusters, which might have fostered the intense star-formation activity and growth in their most massive galaxies. The clustering strength of star-forming galaxies decreases with decreasing redshift (Hartley et al. 2010). Galaxies that are passively evolving at the present time (typically galaxies with halo masses > 1013 M) are twice as strongly clustered than present-day star-forming galaxies (which are typically at least a factor of 10 less massive).
Sheth et al. (2006) point out that at lower redshifts, star formation (either in terms of mass or fraction) is anticorrelated with environment such that dense environments have shown lower star formation rates than low-density regions during the past ~ 5 Gyr. Interestingly, in the present-day Universe star-forming galaxies appear to evolve fairly independently of their environment, with intrinsic properties playing a determining role (Balogh et al. 2004; Poggianti et al. 2008). Overall, environment seems to have had little influence on the cosmic star formation history since z < 1 (Cooper et al. 2008).
The famous morphology-density relation in clusters and groups (Oemler 1974; Dressler 1980; Postman & Geller 1984) describes the increase in the fraction of ellipticals with galaxy density, while the fraction of spirals declines. Similarly, there is a pronounced correlation of morphological types with cluster-centric radius (Whitmore et al. 1993) such that in the innermost regions of a cluster the fraction of ellipticals shows a strong increase, while the fraction of spirals drops sharply. The S0 fraction rises less steeply with decreasing radius and also drops in the innermost regions. These relations are interpreted as suggestive of the growth of ellipticals (and S0s) in high-density regions at the expense of spirals. Goto et al. (2003) find three different regimes depending on galaxy density: For densities < 1 Mpc-2, the morphology-density and morphology-radius relations become rather weak. For 1 - 6 Mpc-2, the fraction of late-type disks decreases with cluster-centric radius, while early-type spiral and S0 fractions increase. For the highest-density clusters with > 6 Mpc-2, also these intermediate-type fractions decrease with cluster-centric radius, while the early-type fractions increase.
For dwarf galaxies, similar global morphology-density and morphology-radius trends are observed in the sense that gas-deficient early-type dwarfs tend to concentrate within ~ 300 kpc around massive galaxies in groups or are predominantly found in the inner regions of clusters, while gas-rich late-type dwarfs are found in the outskirts and in the field (e.g., Grebel 2000; Karachentsev et al. 2002a, Karachentsev et al. 2003a; Lisker et al. 2007). The poorest low-density groups are dominated by late-type galaxies both among their massive and their dwarf members (Karachentsev et al. 2003a, Karachentsev et al. 2003b), while in richer, more compact groups the early-type fractions increase also among the dwarfs, as does the morphological segregation (see, e.g., Karachentsev et al. 2002a, Karachentsev et al. 2002b). Various physical mechanisms including ram pressure and tidal stripping are discussed to explain apparent evolutionary connections of dwarf galaxies with environment (e.g., van den Bergh(1994); Vollmer et al. 2001; Grebel et al. 2003; Dong et al. 2003; Hensler et al. 2004; Kravtsov et al. 2004; Mieske et al. 2004; Lisker et al. 2006a; Mayer et al. 2006; D'Onghia et al. 2009).