First results from reviewing the diagnostic biopsies of Hodgkin's lymphomas (HL) are reported. Biopsies from 1,140 patients were evaluated by consensus diagnoses according to an extended classification of the British National Lymphoma Investigation. 95 of the recruited cases (8.3%) were omitted and not approved as Hodgkin's lymphoma. The remaining 1,045 biopsies were classified as follows: Lymphocytic predominance 31 (2.7%); Nodular sclerosis (NS) 660 (57.9%), Mixed cellularity (MC) 159 (13.9%); Lymphocytic depletion 8 (0.7%); unclassifiable Hodgkin's 148 (13.0%). The unproved Hodgkin's cases [95 (8.3%)] were divided into non-Hodgkin's lymphomas 32 (1.9%), uncertain due to inadequate techniques 32 (2.8%), borderline between Hodgkin's- and Non-Hodgkin's lymphoma 23 (2.0%). Major dissent on this question involved 17 cases (1.5%) and 1 case which was non-malignant. All unclassifiable, borderline or dissent cases were reassessed after the histologic techniques were improved, and immunophenotyping and clinical data reevaluated. The rate of agreement among the observers was about 81.6%, varying between 23.8% in grade 2 NS to 85.0% in both NS groups. Only 62.8% of all primary diagnoses were approved by the final diagnoses of the panel. Important differences in the classification of the British National Lymphoma Investigation concerns the NS-group and mainly its grade 2 subtype. MC was identical in both classifications. Clinico-pathologic correlation of actuarial survival times revealed a significantly worse outcome of MC vs NS, < 20% after 80 months observation. Only slightly significant better survivals were found in grade 1 vs grade 2 NS. Significant differences in unclear compared to all Hodgkin's, were found and the worst survival was in the NHL group.