Pure luminosity evolution models: too few massive galaxies at intermediate and high redshift

We compare recent galaxy data at low and high redshift to pure luminosity evolution (PLE) models, which assume that massive galaxies were assembled and formed most of their stars at high redshift (z > 3) and have evolved without merging or substantial dust obscuration since then. Previous studies, typically comparing to only one or a few different PLE parametrizations, painted a diverse picture of whether or not the evolution of bright early-type galaxies is consistent with such models. Here we attempt to gain further insight by exploring a wider parameter space. Our models span the full range of plausible metallicities, initial mass functions (IMFs) and star formation histories. We require them to reproduce the abundance of galaxies by colour and luminosity in the Sloan Digital Sky Survey, and we investigate whether they can simultaneously fit (i) the observed galaxy counts as a function of redshift in magnitude-limited surveys with K < 20, and (ii) the colour and M/L ratio evolution of red-sequence galaxies in clusters. All models that are consistent with (ii) predict galaxy counts at 1.5 < z < 3 that lie above the observations. This finding does not change with the incorporation of moderate dust extinction, confirming previous studies, which concluded that, for an IMF slope similar to the Salpeter value, such models lie far above the data. The progenitors of most present-day massive galaxies must be much more heavily extincted than currently known galaxies at z >= 1.5 to match the observed counts at these redshifts. Alternatively, the majority of massive galaxies may have assembled at later redshifts as suggested by some hierarchical formation models.