This paper describes a fuzzy expert system for identification of minerals in thin section. Linguistic variables and hedges, as well as fuzzy numbers, are employed to perform approximate modes of reasoning in arriving at approximate answers from an input set of incomplete and imprecise optical data obtained from minerals in thin section. The output consists of a ranked list of possible minerals. The knowledge base of this expert system, XMIN-F, consists of the optical properties of 142 minerals. The list can be expanded easily by the user through the use of a built-in editor. The program is coded in Turbo Pascal, and a user friendly environment is provided by way of numerous windows and menus.