The vulnerability of neuroepithelial cells in the central nervous system (CNS) to neoplastic transformation results from the interaction of several factors: the existence of a reserve population of stem cells, the capability of differentiated cells to reenter the kinetic cycle, the number of replicating cells at risk at a particular time, the length of time during which a particular cell population remains in the cycle, the state of differentiation and the further differentiation potential of that population, and the steps of differentiation that are achieved in successive cell generations. This concept explains many aspects of CNS tumor incidence and the relationship of central neuroepithelial embryonal tumors to tumors of adult cell type. The incidence of different types of central neuroepithelial tumors can be correlated with the width of the window of neoplastic vulnerability. Examples illustrating the existence of only a narrow window include such rare tumors as medulloepitheliomas, cerebral neuroblastomas, gangliogliomas and ependymoblastomas. By contrast, cerebellar medulloblastomas, astrocytomas, mixed astrocytomas and oligodendrogliomas, and glioblastomas exemplify instances in which a relatively wider window of vulnerability exists in the light of cellular neuro-ontogeny and of the capacity of glial cells for postnatal replication. The relationship that may occasionally be established between the development of a glioma and the production of cellular gliosis such as may follow brain injury or accompany multiple sclerosis can also be viewed in the light of that concept. Increasing awareness is needed concerning the development of postradiation gliomas, in particular after the apparently successful treatment of acute lymphocytic leukemia.