Résumé : The object of this work was PURPOSE: to develop a methodology that enables net tumour growth, a balance between actual tumour growth and tumour cell loss, to be approximately evaluated. METHODS: The methodology proposed relies on detecting the growth fraction immunohistochemically by means of MIB-1 antibody labelling combined with cell density determination, carried out on 5-microm-thick Feulgen-stained histological sections with computer-assisted microscopy. The series investigated included 25 oligodendrogliomas (OLG-II), 9 anaplastic oligodendrogliomas (OLG-III). 13 astrocytomas (AST), 14 anaplastic astrocytomas (ANA) and 8 mixed oligoastrocytomas (OLG-AST). RESULTS: The results show that the biological characteristics of some cases were in total accordance with their histopathological diagnoses. This was the case for the "weakly proliferating weakly dense" OLG-II and AST-II tumours, and for the "highly proliferating highly dense" OLG-III and AST-III ones. In contrast, the biological characteristics of some cases seemed to contradict the histopathological case labels. This was the case for the "highly proliferating highly dense" OLG-II and AST-II tumours, the biological aggressiveness of which would be undervalued on the basis of the morphology-based grading system alone, and also for the "weakly proliferating weakly dense" OLG-III and AST-III tumours, the aggressiveness of which would be overvalued. CONCLUSIONS: Combining the determinations of the MIB-1 and the cell density variables appears to be satisfactory in terms of the cell kinetic characterization of glial tumours as a complement to the prognostic information given by a morphology-based grading system alone.