Since Bright's day the study of glomerulonephritis has largely consisted in the observation of the clinical, biochemical, and morphological characteristics of patients. Linked to this has been the attempt to name and classify these data in a fashion which improved prediction of outcome and response to treatment. This has proved difficult, since there is no absolute correspondence between clinical syndromes and the morphological findings in the kidney. The histological data have, however, allowed good prediction of prognosis and response. We now have enough information from experimental models of nephritis and from clinical observation to outline the pathogenesis of glomerulonephritis. Immune mechanisms are certainly concerned in some patients, the role of chronic soluble complex deposition in the kidney being particulary important. Non-immunological origins and mediation of injury, however, are at least as important. Injury to the glomerulus is mediated through the interaction of complement and coagulation mechanisms, leading through the release of soluble factors to inflammation. Repair is probably central in producing irreversible glomerular damage. Treatment can now be directed towards the events concerned in the pathogenesis of nephritis, the most accessible at the moment being the immediate mediators of injury. At a more fundamental level better identification in individual patients of the events leading to the injury is needed, such as the isolation, avoidance, or elimination of the antigens involved in chronic soluble complex disease. Laboratory techniques to begin these investigations in man are now available and need to be applied more widely in the clinical field to patients with glomerulonephritis.