The antiperinuclear factor (APF) test raises two main problems: the unpredictability of the cells used as substrate and the difficulty in expressing the results. We propose that 10% of the cells have to be stained by a given serum in order for it to be considered positive. APF were found to be present in 76% of rheumatoid arthritis (RA) patients, 3% of healthy controls and occasionally in disease controls. The production of APF was significantly (p less than 0.01) related to the presence of rheumatoid factor in RA, and IgG antibody was predominant in the APF test.