2006MNRAS.370..702C

Using the third data release of the Sloan Digital Sky Survey (SDSS), we have rigorously defined a volume-limited sample of early-type galaxies in the redshift range 0.005 < z ≤ 0.1. We have defined the density of the local environment for each galaxy using a method which takes account of the redshift bias introduced by survey boundaries if traditional methods are used. At luminosities greater than our absolute r-band magnitude cut-off of -20.45, the mean density of environment shows no trend with redshift. We calculate the Lick indices for the entire sample and correct for aperture effects and velocity dispersion in a model-independent way. Although we find no dependence of redshift or luminosity on environment, we do find that the mean velocity dispersion, σ, of early-type galaxies in dense environments tends to be higher than in low-density environments. Taking account of this effect, we find that several indices show small but very significant trends with environment that are not the result of the correlation between indices and velocity dispersion. The statistical significance of the data is sufficiently high to reveal that models accounting only for α-enhancement struggle to produce a consistent picture of age and metallicity of the sample galaxies, whereas a model that also includes carbon enhancement fares much better. We find that early-type galaxies in the field are younger than those in environments typical of clusters but that neither metallicity, α-enhancement nor carbon enhancement are influenced by the environment. The youngest early-type galaxies in both field and cluster environments are those with the lowest σ. However, there is some evidence that the objects with the largest σ are slightly younger, especially in denser environments.

Independent of environment both the metallicity and α-enhancement grow monotonically with σ. This suggests that the typical length of the star formation episodes which formed the stars of early-type galaxies decreases with σ. More massive galaxies were formed in faster bursts.

We argue that the timing of the process of formation of early-type galaxies is determined by the environment, while the details of the process of star formation, which has built up the stellar mass, are entirely regulated by the halo mass. These results suggest that the star formation took place after the mass assembly and favours an anti-hierarchical model. In such a model, the majority of the mergers must take place before the bulk of the stars form. This can only happen if there exists an efficient feedback mechanism which inhibits the star formation in low-mass haloes and is progressively reduced as mergers increase the mass.