The origin and nature of central nervous system tumors primarily composed of primitive neuroepithelial cells have intrigued neuropathologists since 1925, when the prototype tumor in this group, the medulloblastoma, was defined as a specific type by Bailey and Cushing. Theories regarding this group of tumors have been widely divergent, ranging from classification as sarcomas, neuroblastomas, spongioblastomas, ie, poorly differentiated gliomas (including astrocytic, oligodendroglial, and ependymal), to primitive tumors with a capacity for multidirectional differentiation. Studies of both human and experimental primitive neuroectodermal tumors have yielded considerable information about biological features, although controversies remain regarding interpretation of the findings. Moreover, many investigators remain fixed upon the notion that discovery of the "cell of origin" of these tumors will somehow enhance understanding of their nature. Rationale for relevance of this view is examined against the current state of knowledge regarding carcinogenesis as a whole and studies of human and experimental primitive neuroectodermal tumors in particular, and it is concluded that the time has come to discard the archaic concept of cell of origin and devote resources to investigations of the phenotypic character of the tumor and its relationship to prognosis and treatment.