Abstract: The first two volumes of the Catalogue of Galaxies and Clusters of Galaxies by Zwicky et al. list 3523 clusters of galaxies. Their distribution in space has been analyzed over both the entire region of the sky covered by the two volumes (about 5700 deg2) and also for that part of the region with galactic latitude greater than 60 (about 1500 deg2). The procedure is to divide the region of the sky surveyed into grid cells of various sizes, and for each sized cell to determine the frequency distribution f(t) of cells containing t clusters each. A ~2 test is used to estimate the probability that f ( t) would be obtained in a random sampling from a population whose frequency distribution is the binomial distribution B(t). The f(t) and comparison with B(t) are determined separately for clusters in each of Zwicky's five distance categories. It is found that f(t) approaches B (t) for very small and also for very large cells, but deviates from B (t) with extremely large statistical significance for cells of intermediate size, indicating a "clumpy" distribution of clusters. The cell size for which the maximum deviation of f(t) from B (t) is obtained is taken as the scale of the dumpiness. The angular scale found is in inverse proportion to the distance that corresponds to each distance class. Thus the clumps apparently represent real concentrations of matter in space. It is shown that interstellar and intergalactic absorption, and observational selection effects cannot account for the apparent clustering of the clusters. For a Hubble constant, H=75 km/sec x 10~ pc, the linear scale of dumpiness is about 5 X 10~' pc, in excellent agreement with the corresponding value found by Abell (Astrophys. J. Suppl. 3, 211, 1958) in his analysis of the clustering of the rich clusters catalogued from the Palomar Sky Survey. 3