Several distinct arthritic syndromes now have been recognized in HIV-infected persons. These comprise seronegative spondarthritis, including classic Reiter's syndrome and psoriatic arthritis associated with HLA-B27, and undifferentiated arthritis usually confined to the lower limbs, unassociated with other lesions, and unrelated to any known genetic marker. In such cases great care should be taken to exclude infection. In addition, a syndrome of short-lived but sometimes severe arthralgias also occurs. Spinal pain is a major problem in some patients but ankylosing spondylitis appears to be rare among this group. Psoriasis probably occurs more often in the HIV-infected group than in the population in general and may be especially severe in those patients with arthritis. Arthritis has been reported in the United States, Europe, and Africa among persons considered to be at high and low risk for HIV infection. Arthritis can occur at any stage of HIV infection, but the true prevalence of arthritic syndromes and the nature of their association with HIV infection remains unclear. In view of the development of Reiter's syndrome in some patients, precipitating bacterial infections have been sought as the culprits. In a minority of cases, shigella, yersinia, and campylobacter infections have been implicated, but in the majority of cases, no specific infection has been identified. In most patients depletion of circulating CD4-positive lymphocytes is present by the time that arthritis is detected, but only limited data on synovial immunopathology are available. In some patients changes of nonspecific chronic synovial inflammation are present and synovial fluid cell counts are high. In other patients evidence of inflammatory changes is minimal. Human immunodeficiency virus has been isolated from joint fluid and identified in large mononuclear, probably dendritic, cells and lymphocytes. Synovium from patients dying with AIDS but with apparently normal joints also shows significant abnormalities that could lead to joint disease in long-term survivors. The possibility of a viral etiology of arthritis in some cases is suggested by the induction of arthritis in animals by lentivirus infection; it also is possible, however, that HIV enhances the effect of mechanisms that can operate in the absence of HIV infection. Conventional treatments of rheumatic lesions, including intraarticular steroids, appear to be safe and reasonably effective. Anecdotal evidence suggests that treatment with methotrexate and azathioprine leads to exacerbation of HIV disease and should be avoided.