DETECTION AND SELECTION EFFECTS IN OBSERVATIONS OF FAINT GALAXIES

Cosmological predictions of number count and redshift distribution of faint galaxies are made by incorporating detection and selection effects in the theoretical framework. It is shown that the detection rate for galaxies at given apparent magnitude strongly depends on the detection criterion and magnitude measurement algorithm, and that a large number of faint galaxies with high redshift could remain undetected in recent galaxy surveys. By using a current model of galaxy luminosity evolution with the assumption that galaxies do not evolve in diameter or number, the predictions are compared directly with the uncorrected B(J)R(F)IK-band galaxy count data reaching B(J) = 27 mag or the equivalent depth and also with the Durham redshift distribution to B(J) = 23 mag. It is concluded that a flat and low-density OMEGA0 is similar to 0.1 universe with nonzero cosmological constant is compatible with the data presented, but a flat and high-density OMEGA0 = 1 universe shows a marked discrepancy with all the data. A hypothesis of galaxy evolution in diameter and/or number is tested against faint observations. Central surface brightness versus apparent magnitude relation is found to be sensitive solely to a change of intrinsic size of galaxies so that it is used as a probe of their diameter evolution. Comparison between the predicted and observed relations indicates that the bulk of galaxies have undergone little diameter evolution over a considerable fraction of their age. A merger-driven number evolution model is critically reconsidered from a viewpoint of detectability of high-redshift premergers. Contrary to usual expectations, the observed high count at B(J) = 27 mag may not be explained in the context of a high-density OMEGA0 = 1 universe because many high-redshift premergers could not be counted according to the criterion adopted in recent galaxy surveys.