THE VOID PHENOMENON EXPLAINED

We use high-resolution N-body simulations, combined with a halo occupation model of galaxy bias, to investigate voids in the galaxy distribution. Our goal is to address the "void phenomenon" of Peebles, which presents the observed dearth of faint galaxies in voids as a challenge to current cosmology. In our model, galaxy luminosity is determined only as a function of dark matter halo mass. With this simple assumption, we demonstrate that large, empty voids of similar to 15 h(-1) Mpc in diameter are expected even for galaxies seven magnitudes fainter than L-*. The predictions of our model are in excellent agreement with several statistical measures: (i) the luminosity function of galaxies in underdense regions, (ii) nearest-neighbor statistics of dwarf galaxies, and (iii) the void probability function of faint galaxies. In the transition between filaments and voids in the dark matter, the halo mass function changes abruptly, causing the maximum galaxy luminosity to decrease by similar to 5 magnitudes over a range of similar to 1 h(-1) Mpc. Thus, the boundary between filaments and voids in the galaxy distribution is nearly as sharp for dwarfs as for similar to L-* objects. These results support a picture in which galaxy formation is driven predominantly by the mass of the host dark matter halo, and is nearly independent of the larger-scale halo environment. Further, they demonstrate that Lambda CDM, combined with a straightforward bias model, naturally predicts the existence of the void phenomenon.