The antigenic heterogeneity of human neuroectodermal tumors defined by both murine and human monoclonal antibodies (MAs) is reported; no patterns of reactivity defining degree of anaplasia, in vitro morphology, or immunogen used were apparent. We investigated the reactivity of 20 distinct murine MAs defining markers of glioma-associated or predominantly lymphoid distribution for 13 human glioma-derived (HGL) cell lines and frozen sections of 19 human glioblastoma multiforme (GBM) and six astrocytomas (AST). Methods included radioimmunoassay, immunofluorescence, immunohistochemistry, and absorption analysis. Two markers, HLA-A,B and human Thy-1, exhibited no deviation; all HGL cell lines tested bound high levels of specific MA. Individual HGL cell line reactivity with the MA panel ranged from 30 to 70%. HGL cell lines (7/13) which reacted with ≥50% of the anti-glioma MAs had the highest (30–70%) positive reactivity rates with the anti-lymphoid marker MA panel; complex antigenicity in one system correlated with multiple antigens in the other. Within the anti-lymphoid marker MA panel, subpopulations of 4/13 HGL cell lines were clearly positive for the HLA-DR (Ia) antigens; another 3/13 HGL cell lines were strongly positive for common acute lymphocytic leukemia antigen (CALLA). With the exception of Thymocyte 1 antigen (Thy-1), reactivity for early and mature T-cell markers was infrequent and sporadic. Lymphoid marker expression by HGL cell lines is highly heterogeneous, ranging from few (Thy-1 and HLA-A,B) to complex expression of Ia, T-cell, and lymphoid tumor markers. GBM and AST tissues were antigenically less complex; for each of 6/8 anti-glioma MA, 70–100% of GBM and 66–100% of AST were positive. Two MAs were highly reactive (7/10, 8/9) with GBM sections and minimally so (1/6) with AST. Antigenic expression in gliomas is complex and heterogeneous; however, clear differences in lymphoid marker expression, the identification of widely and rarely expressed glioma-associated antigens, and the potential of immunologic differentiation between GBM and AST by large panels of MAs will serve to reduce the complexity and may be of potential diagnostic or prognostic significance