Bulk chemical analysis is shown to be insufficient for the archaeometric interpretation of coarse ware pottery. Accompanied by quantitative petrographic analysis of the temper, much more conclusive information can be obtained at little more expense. Three small groups of prehistoric pottery from Switzerland have been examined with both techniques, showing that expensive chemical bulk analyses may be obsolete or even misleading, and in any case need further confirmation and interpretation by petrographical means. Assuming a rigorous sampling strategy, the number of analyses necessary for valid statements can be substantially reduced relative to those necessary for the statistical interpretation of bulk analyses only.