The study of stellar populations in galaxies is entering a new era with the availability of large and high‐quality data bases of both observed galactic spectra and state‐of‐the‐art evolutionary synthesis models. In this paper we investigate the power of spectral synthesis as a means to estimate the physical properties of galaxies. Spectral synthesis is nothing more than the decomposition of an observed spectrum in terms of a superposition of a base of simple stellar populations of various ages and metallicities, producing as output the star formation and chemical histories of a galaxy, its extinction and velocity dispersion. Our implementation of this method uses the recent models of Bruzual & Charlot and observed spectra in the 3650–8000 Å  range. The reliability of this approach is studied by three different means: (1) simulations, (2) comparison with previous work based on a different technique, and (3) analysis of the consistency of results obtained for a sample of galaxies from the Sloan Digital Sky Survey (SDSS). We find that spectral synthesis provides reliable physical parameters as long as one does not attempt a very detailed description of the star formation and chemical histories. Robust and physically interesting parameters are obtained by combining the (individually uncertain) strengths of each simple stellar population in the base. In particular, we show that, besides providing excellent fits to observed galaxy spectra, this method is able to recover useful information on the distributions of stellar ages and, more importantly, stellar metallicities. Stellar masses, velocity dispersion and extinction are also found to be accurately retrieved for realistic signal‐to‐noise ratios. We apply this synthesis method to a volume‐limited sample of 50 362 galaxies from the SDSS Data Release 2, producing a catalogue of stellar population properties. Emission lines are also studied, their measurement being performed after subtracting the computed starlight spectrum from the observed one. A comparison with recent estimates of both observed and physical properties of these galaxies obtained by other groups shows good qualitative and quantitative agreement, despite substantial differences in the methods of analysis. The confidence in the present method is further strengthened by several empirical and astrophysically reasonable correlations between synthesis results and independent quantities. For instance, we report the existence of strong correlations between stellar and nebular metallicities, stellar and nebular extinctions, mean stellar age and equivalent width of Hα and 4000‐Å  break, and between stellar mass and velocity dispersion.

We have re-evaluated empirical expressions for the abundance determination of N, O, Ne, S, Cl, Ar and Fe taking into account the latest atomic data and constructing an appropriate grid of photoionization models with state-of-the art model atmospheres. Using these expressions we have derived heavy element abundances in the ∼310 emission-line galaxies from the Data Release 3 of the Sloan Digital Sky Survey (SDSS) with an observed Hβ flux F(Hβ) > 10 −14 erg s −1 cm −2 and for which the [O iii] λ4363 emission line was detected at least at a 2σ level, allowing abundance determination by direct methods. The oxygen abundance 12 + log O/H of the SDSS galaxies lies in the range from ∼7.1 (Z /30) to ∼8.5 (0.7 Z ). The SDSS sample is merged with a sample of 109 blue compact dwarf (BCD) galaxies with high quality spectra, which contains extremely low-metallicity objects. We use the merged sample to study the abundance patterns of low-metallicity emission-line galaxies. We find that extremely metal-poor galaxies (12 + log O/H < 7.6, i.e. Z < Z /12) are rare in the SDSS sample. The α element-to-oxygen abundance ratios do not show any significant trends with oxygen abundance, in agreement with previous studies, except for a slight increase of Ne/O with increasing metallicity, which we interpret as due to a moderate depletion of O onto grains in the most metal-rich galaxies. The Fe/O abundance ratio is smaller than the solar value, by up to 1 dex at the high metallicity end. We also find that Fe/O increases with decreasing Hβ equivalent width EW(Hβ). We interpret this as a sign of strong depletion onto dust grains, and gradual destruction of those grains on a time scale of a few Myr. All the galaxies are found to have log N/O > -1.6, implying that they have a different nature than the subsample of high-redshift damped Lyα systems with log N/O of ∼-2.3 and that their ages are larger than 100-300 Myr. We confirm the apparent increase in N/O with decreasing EW(Hβ), already shown in previous studies, and explain it as the signature of gradual nitrogen ejection by massive stars from the most recent starburst.

We use the WHα versus [N ii]/Hα (WHAN) diagram introduced by us in previous work to provide a comprehensive emission‐line classification of Sloan Digital Sky Survey galaxies. This classification is able to cope with the large population of weak line galaxies that do not appear in traditional diagrams due to a lack of some of the diagnostic lines. A further advantage of the WHAN diagram is to allow the differentiation between two very distinct classes that overlap in the low‐ionization nuclear emission‐line region (LINER) region of traditional diagnostic diagrams. These are galaxies hosting a weakly active galactic nucleus (wAGN) and ‘retired galaxies’ (RGs), i.e. galaxies that have stopped forming stars and are ionized by their hot low‐mass evolved stars. A useful criterion to distinguish true from fake AGN (i.e. the RGs) is the value of ξ, which measures the ratio of the extinction‐corrected Hα luminosity with respect to the Hα luminosity expected from photoionization by stellar populations older than 108 yr. We find that ξ follows a markedly bimodal distribution, with a ξ≫ 1 population composed by systems undergoing star formation and/or nuclear activity, and a peak at ξ∼ 1 corresponding to the prediction of the RG model. We base our classification scheme not on ξ but on a more readily available and model‐independent quantity which provides an excellent observational proxy for ξ: the equivalent width of Hα. Based on the bimodal distribution of WHα, we set the practical division between wAGN and RGs at WHα= 3 Å. Five classes of galaxies are identified within the WHAN diagram: pure star‐forming galaxies: and WHα > 3 Å; strong AGN (i.e. Seyferts): and WHα > 6 Å; weak AGN: and WHα between 3 and 6 Å; RGs (i.e. fake AGN): WHα < 3 Å; passive galaxies (actually, lineless galaxies): WHα and W[N ii] < 0.5 Å. A comparative analysis of star formation histories and of other physical and observational properties in these different classes of galaxies corroborates our proposed differentiation between RGs and wAGN in the LINER‐like family. This analysis also shows similarities between strong and weak AGN on the one hand, and retired and passive galaxies on the other.