Substructure in galaxy clusters: A two-dimensional approach

We discuss a procedure for detecting and quantifying the presence of substructure in the projected positions of galaxies in clusters. New contour maps of the surface density of galaxies for the 56 clusters in Dressler's morphological sample are constructed by application of an adaptive-kernel technique. We employ the KMM algorithm to objectively partition the galaxies in each cluster into sub-populations, identify the best-fitting alternative to the null hypothesis of a single-population model, and estimate the statistical significance of the identified structures. The power of the KMM algorithm for this application is evaluated based on a study of simulated clusters. We estimate that 57% of the Dressier morphological-sample clusters have statistically-significant substructure. In all likelihood, this proportion represents a lower limit to the fraction of clusters with real substructure. We calculate the projected median distance of the D or cD galaxy (in the clusters which possess them) to the centroid of the nearest subcluster, and demonstrate that the dominant galaxies are located close to the peaks of the galaxy surface-density. (C) 1997 American Astronomical Society.