2008ApJ...683..321F

We test the hypothesis that the host galaxies of long-duration gamma-ray bursts (GRBs) as well as quasar-selected damped Lyα (DLA) systems are drawn from the population of UV-selected star-forming, high-z galaxies (generally referred to as Lyman break galaxies). Specifically, we compare the metallicity distributions of the GRB and DLA populations against simple models where these galaxies are drawn randomly from the distribution of star-forming galaxies according to their star formation rate and H I cross section, respectively. We find that it is possible to match both observational distributions assuming very simple and constrained relations between luminosity, metallicity, and H I sizes. The simple model can be tested by observing the luminosity distribution of GRB host galaxies and by measuring the luminosity and impact parameters of DLA-selected galaxies as a function of metallicity. Our results support the expectation that GRB and DLA samples, in contrast with magnitude-limited surveys, provide an almost complete census of z~3 star-forming galaxies that are not heavily obscured.