The classification of galaxies as star forming or active is generally done in the ([O III]/Hbeta, [N II]/Halpha) plane. The Sloan Digital Sky Survey (SDSS) has revealed that, in this plane, the distribution of galaxies looks like the two wings of a s
eagull. Galaxies in the right wing are referred to as Seyfert/LINERs, leading to the idea that non-stellar activity in galaxies is a very common phenomenon. Here, we argue that a large fraction of the systems in the right wing could actually be galaxies which stopped forming stars. The ionization in these retired galaxies would be produced by hot post-AGB stars and white dwarfs. Our argumentation is based on a stellar population analysis of the galaxies via our STARLIGHT code and on photoionization models using the Lyman continuum radiation predicted for this population. The proportion of LINER galaxies that can be explained in such a way is however uncertain. We further show how observational selection effects account for the shape of the right wing. Our study suggests that nuclear activity may not be as common as thought. If retired galaxies do explain a large part of the seagulls right wing, some of the work concerning nuclear activity in galaxies, as inferred from SDSS data, will have to be revised.
We study the evolution of 82302 star-forming (SF) galaxies from the SDSS. Our main goals are to explore new ways of handling star formation histories (SFH) obtained with our publicly available spectral synthesis code STARLIGHT, and apply them to inve
stigate how SFHs vary as a function of nebular metallicity (Zneb). Our main results are: (1) A conventional correlation analysis shows how global properties such as luminosity, mass, dust content, mean stellar metallicity and mean stellar age relate to Zneb. (2) We present a simple formalism which compresses the results of the synthesis into time-dependent star formation rates (SFR) and mass assembly histories. (3) The current SFR derived from the population synthesis and that from H-alpha are shown to agree within a factor of two. Thus we now have a way to estimate SFR in AGN hosts, where the H-alpha method cannot be applied. (4) Fully time-dependent SFHs are derived for all galaxies and averaged over six Zneb bins spanning the entire SF wing in the [OIII]/H-beta X [NII]/H-alpha diagram. (5) We find that SFHs vary systematically along the SF sequence, such that low-Zneb systems evolve slower and are currently forming stars at a higher relative rate. (6) At any given time, the distribution of specific SFRs for galaxies within a Zneb-bin is broad and roughly log-normal. (7) The same results are found grouping galaxies in stellar mass (M*) or surface mass density (S*) bins. (8) The overall pattern of SFHs as a function of Zneb, M* or S* is robust against changes in selection criteria, choice of evolutionary synthesis models for the spectral fits, and differential extinction effects. (Abridged)
