Before remediation of a site with contaminated soil or groundwater, the contaminant plume must first be characterized. This involves sampling the contaminant concentration at a set of locations in and around the contaminated area. To present the measured concentrations in a meaningful form, the concentrations are typically interpolated to the nodes of a three-dimensional grid, and the plume is visualized by constructing iso-surfaces from the gridded data. The critical step in this process is the interpolation stage. Improper application of an interpolation scheme can result in grossly misleading three-dimensional plume maps. There are a number of problems that often occur when interpolating contaminant plume data, including generation of negative concentrations, oscillation of interpolated values, improper estimation of maximum concentrations, and skewing of the results due to data clustering. These and other difficulties associated with plume characterization are discussed, along with a simple set of guidelines for detecting and overcoming these problems.