Quantitative pathology is a rapidly growing field of pathology. This field is developing methods allowing greater objectivity and better reproducibility in diagnostic decisions. The introduction of quantitative pathology as a teaching subject in medical education will help this field to become more widely and more correctly applied. Courses on quantitative pathology are organized yearly by the Department of Pathology at the University of Ancona. The 1986 course emphasized training of students in understanding basic problems such as variation caused by tissue processing, instrumentation, data handling and interpretation, and variation between observers. In this paper we describe experiments performed during the course on identification and quantitation of mitoses in malignant glial tumors. The participants, i.e. students of the postgraduate Pathology School, could identify mitoses with a 'substantial' reproducibility. However, when the observers were allowed freely to choose the fields to count, the reproducibility was 'slight'. The study suggests that sampling rules should be applied which secure reproducible application of morphometric method also on condition of free selection of fields. Because the application of such rules needs expertise in histopathology, it seems that morphometric method can best be applied by histopathologists themselves.