The concept of an infectious aetiology of rheumatoid arthritis with special emphasis on viruses in discussed. Two main streams of current thought were considered. One was that persistent viral infection of rheumatoid synovial lining or cartilage cells could act as a target for immune destruction; the other was that viral infection of some effector cells of the immune system could lead to an autoimmune process disordered immunoregulation. Initially, the involvement of rubella virus in the chronic inflammatory process of rheumatoid arthritis was looked at. A collection of synovial fibroblasts from patients with rheumatoid arthritis and from patients with osteoarthoses or other non-RA disease were examined for rubella antigens by immunofluorescence and radioimmunoassay. Sections of synovial membranes and cartilages were also looked at using an immunoperoxidase technique. There was no evidence of rubella antigens in any of the cells or tissues (target hypothesis). Lymphocytes isolated from synovial fluid and peripheral blood were compared for the expression of rubella antigens and both found to be negative (disordered immunoregulation). Direct detection of rubella antigens and isolation of virus from the cells of synovial fluid was attempted in an unusual case of rubella polyarthritis. The result was negative although the persistence of high levels of rubella Ig1 in this patient's serum suggested a persisting viral stimulus. The overall results provided little support for the various hypotheses that persistent infection with rubella virus underlies the rheumatoid process. Evidence in favour of a possible viral aetiology of rheumatoid arthritis was further sought by studying the immunoglobulins produced in vitro by cultured rheumatoid synovial membranes. These were examined for their ability to form complexes with 3H-labelled viruses, namely rubella, measles, adenovirus and feline leukaemia virus. A technique was developed to detect specific in vitro 14G labelled antiviral antibodies. No antiviral activity was found in immunoglobulins associated with or produced by rheumatoid synovial membranes. The possible involvement of retroviruses in the aetiology of rhematoid arthritic was also investigated. Antibodies cross-reacting with primateretrovirus antigens were sought for in sera from patients with 'autoimmme' diseases by means of solid phase radioimmunoassay. There was no difference in the distribution of the immunoglobulins bound to retrovirus antigens in the three groups of patients studied: i.e. rheumatoid arthritis, systemic lupus erythematosus and a group of non-RA patients. Absorption of rheumatoid factor did not alter this conclusion. Retrovirus antigens were not expressed on rheumatoid synovial and peripheral blood lymphocytes as judged by membrane immunofluorescence, radioimmunoassay and complement-mediated cytotoxicity. The specific anti-retroviral sera used in this study was produced in rabbits immunized with viral antigens grown in homologous system (rabbit calls and medium supplemented with normal rabbit serum), avoiding non-specific immunofluorescence previously detected with donated anti-retroviral sera. Immune complexes lodged in the rheumatoid synovial membranes did not contain, and other cells within the membranes did not express, retroviral antigens. These results give a little support to the hypothesis that activation of endogenous human retroviruses or an infection with horizontally transmitted retroviruses is associated with the rheumatoid process.