Controversial issues relating to the pathobiology and classification of central nervous system primitive neuroectodermal tumors (PNETs) have plagued neuropathologists for more than 70 years. Hypotheses advanced in the mid-1920's have remained as fixed concepts in contemporary literature, largely consequent to repetitious support by a small number of neuropathologists despite a growing body of information discrediting these ideas from neuroembryologists, oncologists, neuroscientists and pathologists. Attention has largely focused upon PNETs arising in the cerebellum (commonly known as medulloblastomas ([MBs]), because about 80% of central nervous system (CNS) PNETs originate in this site. It has been asserted that the 20% which do not are biologically different, although most individuals agree that the histological features of PNETs that occur in different sites throughout the CNS are indistinguishable from those growing in the cerebellum. The historical aspects of this controversy are examined in the face of evidence that there is, in fact, a unique class of CNS tumors which should appropriately be regarded as primitive neuroectodermal in nature. Specifically, a number of different approaches to the problem have yielded data supporting this hypothesis. These approaches include the identification of patterns of expression among a variety of cellular antigens (demonstrated by the use of immunopathological techniques), molecular analyses of cell lines derived from these tumors, experimental production of PNETs and molecular genetic analyses. Differences of opinion among surgeons, oncologists and radiotherapists are typically resolved by conducting cooperative studies of patients with these tumors who are diagnosed and treated at multiple centers.