In this study, a numerical algorithm was developed to decompose the planar mass structure of paper into a random array of grey disks with a discrete size distribution. The optimum size and the frequency of these disks were determined such that the second order statistics of the corresponding random disk structure resembled that of the paper sample. Using this method, eighty two (82) commercial and laboratory-made samples were analyzed. It was found that; independent of the forming conditions, the average disk size was proportional to the standard deviation of the disk size distribution. The utility of this new tool in analyzing the effect of papermaking conditions on paper formation is illustrated.