COUNTER-EVOLUTION OF FAINT FIELD GALAXIES

We adopt a new approach to explore the puzzling nature of faint blue field galaxies. Instead of assuming that the local luminosity function is well defined, we first determine whether any nonevolving set of luminosity functions for different spectral types of galaxies is compatible with the observed marginal distributions in optical and near-infrared counts, B-R colors, and redshifts. Exploiting a nonnegative least-squares method, we derive a new no-evolution model that is found to fit all the observations surprisingly well, from B less-than-or-equal-to 15 to the limit of B is similar to 24 for redshifts and B is similar to 26 for colors and counts. Contrary to previous conclusions, the faint galaxies in excess of our new no-evolution model are red rather than blue. Although our fits are far better than previous no-evolution models, the remaining deviations from the observations still suggest the need for some, but slight, evolutionary component or model revisions. We conclude that models more exotic than mild luminosity evolution, such as those requiring rapid evolution in star formation rates, disappearing dwarf galaxy populations, high values of the cosmological constant, rapid mergers, or substantial nonconservation of galaxy numbers with time, are no longer as compelling. Our results resurrect field galaxies as promising probes of the curvature of space.