Describes a geometrical technique, based on the Dirichlet tessellation, for quantitative characterization of inhomogeneities in the spatial distribution of second-phase particles. The Dirichlet tessellation representation allows a detailed, statistical description of particle distribution properties, including local particle density and vector nearest-neighbor distances. Extension of the tessellation technique by specifying a distance over which second-phase particles are assumed to interact mechanically (particle interaction distance) facilitates quantitative characterization of particle clustering by allowing evaluation of characteristics such as the fraction of particles clustered, and the number, size, and spatial distribution of clusters. The tessellation and clustering characteristics of several types of computer-generated particle arrays (random, clustered, and hexagonal cellular) are presented. The effects of plane-strain deformation on the tessellation and clustering characteristics of the particle arrays have been examined. Tessellation and clustering characteristics of the random particle arrays are insensitive to plane-strain deformation.