The major function of a classification of neoplasms is to correlate with prognosis. During the 1980s two trends developed in the classification of gliomas: 1) the presence of necrosis in astrocytomas defines a subset with the worst prognosis (glioblastoma multiforme) and 2) no comparable subset of oligodendroglioma appears to exist, either histologically or prognostically. Even the worst of the oligodendrogliomas has a better prognosis than glioblastomas. These asymmetries must be kept in mind when one is considering the diagnosis of glioblastoma. The absence of a specific stain for oligodendrogliomas, the difficulty in differentiating small anaplastic cells from oligodendroglial cells and the common occurrence of mixtures of normal and/or reactive oligodendroglia and astrocytes in gliomas combine to create potentially confusing misclassification of many gliomas.