Characteristic analysis is proposed as a multivariate technique for comparing regional unknown cell data with data for explored areas. Original data are first converted into binary arrays according to whether or not each measured variable at a given location is considered to be favorable (1) or of indeterminate favorability (0) in terms of mineralization. A characteristic vector is calculated for the model selected, and each region cell is compared with the model vector. Region cells are ranked according to the measure of association with the model, and the results to be interpreted are plotted on a map. An interactive computer program was written to perform the analysis.Twelve separate mineralized areas were selected as models in the Coeur d'Alene district, Idaho. The area comprising1 all twelve mineralized areas was selected as a thirteenth model. Geochemical data for Hg, Te, Cu, Pb, Zn, Ag, Cd, As, Sb, Na, and K in 4,000 rock samples were used as the variables. The map derived for each model shows the distribution of degrees of association of all cells in the district. A structural restoration of the aggregate mineralization model shows that nonmodel cells located between major models have high degrees of association with the aggregate model, implying continuity of mineralization. Because these cells would be the logical potential target areas considering the known geology, it is concluded that characteristic analysis is a powerful analytical tool in exploration.