A NEW TEST OF THE STATISTICAL NATURE OF THE BRIGHTEST CLUSTER GALAXIES

A novel statistic is proposed to examine the hypothesis that all cluster galaxies are drawn from the same luminosity distribution (LD). In such a "statistical model" of galaxy LD, the brightest cluster galaxies (BCGs) are simply the statistical extreme of the galaxy population. Using a large sample of nearby clusters, we show that BCGs in high luminosity clusters (e. g., L(tot) greater than or similar to 4 x 10(11)h(70)(-2) L(circle dot)) are unlikely (probability <= 3 x 10(-4)) to be drawn from the LD defined by all red cluster galaxies more luminous than M(r) = -20. On the other hand, BCGs in less luminous clusters are consistent with being the statistical extreme. Applying our method to the second brightest galaxies, we show that they are consistent with being the statistical extreme, which implies that the BCGs are also distinct from non-BCG luminous, red, cluster galaxies. We point out some issues with the interpretation of the classical tests proposed by Tremaine & Richstone (TR) that are designed to examine the statistical nature of BCGs, investigate the robustness of both our statistical test and those of TR against difficulties in photometry of galaxies of large angular size, and discuss the implication of our findings on surveys that use the luminous red galaxies to measure the baryon acoustic oscillation features in the galaxy power spectrum.