This paper traces the work history of patients with rheumatoid arthritis (RA) from the year of diagnosis to 1985. The paper also describes the risk factors for work loss among patients with RA. It uses data from a panel of 698 RA patients, observed for 4 years, from the practices of a random sample of northern California rheumatologists. Of these 698, 353 had worked for pay at some point in their lives. Three hundred six of the 353 had worked when diagnosed as having RA. Of these 306, 157 (51%) were no longer working in 1985. Forty-seven individuals started working after the onset of illness, but of these, approximately one-third had stopped working by 1985. In all, 50% of RA patients with some work experience stopped working within a decade of diagnosis, 60% within 15 years, and 90% within 30 years. We found that the probability of work loss is lessened among persons in jobs that have few physical requirements, among those with high levels of discretion over the pace and activities of work, and among those who were able to stay on the job held when the diagnosis was made. The probability of work loss is increased among service workers. The findings of this longitudinal study, showing that work characteristics profoundly alter the probability of work loss among persons with RA, are consistent with the findings of our earlier cross-sectional studies of work outcome and RA.