The two-point correlation function has been the standard statistic for quantifying how galaxies are clustered. The statistic uses the positions of galaxies, but not their properties. Clustering as a function of galaxy property, be it type, luminosity, color, etc., is usually studied by analysing a subset of the full population, the galaxies in the subset chosen because they have a similar range of properties. We explore an alternative technique---marked correlations---in which one weights galaxies by some property or `mark when measuring clustering statistics. Marked correlations are particularly well-suited to quantifying how the properties of galaxies correlate with their environment. Therefore, measurements of marked statistics, with luminosity, stellar mass, color, star-formation rate, etc. as the mark, permit sensitive tests of galaxy formation models. We make measurements of such marked statistics in semi-analytic galaxy formation models to illustrate their utility. These measurements show that close pairs of galaxies are expected to be red, to have larger stellar masses, and to have smaller star formation rates. We also show that the simplest unbiased estimator of the particular marked statistic we use extensively is very simple to measure---it does not require construction of a random catalog---and provide an estimate of its variance. Large wide-field surveys of the sky are revolutionizing our view of galaxies and how they evolve. Our results indicate that application of marked statistics to this high quantity of high-quality data will provide a wealth of information about galaxy formation.
Download