The medical records of all patients with rheumatoid arthritis diagnosed in Rochester, Minnesota, from 1950 to 1975 were examined to determine how many of them had malignant neoplasms. Follow-up averaged more than 14 years with outcome diagnoses complete in 98 percent of cases. All diagnoses of malignant neoplasm in this cohort were identified through the centralized record system based at the Mayo Clinic. Approximately 40 percent of these patients with rheumatoid arthritis were at least 60 years old at diagnosis. For comparison, the expected number of malignancies has been calculated using age-specific and site-specific rates from previous Rochester studies and the number of years of observation from date of diagnosis of rheumatoid arthritis to the date of last examination. Risk ratios have been calculated by dividing the observed number by the expected number of malignancies. Exact 95 percent confidence intervals around the risk ratios were calculated assuming that the observed number of cases has a Poisson distribution and that the expected number is fixed. Those patients with rheumatoid arthritis who had a malignancy before diagnosis have been analyzed separately, because they are at a higher risk. With the exception of multiple myeloma, no association was found between rheumatoid arthritis and subsequent cancer of any site.