Limiting dilution analysis (LDA) was used to analyze the defect of production of IL-2 by synovial fluid cells in rheumatoid arthritis (RA). LDA is a relatively simple means of separating T-cell function from the potential effects of accessory-cells, suppressor factors and adsorption. The number of precursor cells for interleukin-2 (IL-2) secretion was determined among peripheral blood cells (PBL) from healthy control individuals, and in PBL and synovial fluid (SF) mononuclear cells from patients with rheumatoid arthritis (RA). Although the frequency of such cells in the PBL of the controls and RA patients was similar, that of the SF was significantly lower (about one sixth). This difference could not be accounted for by an inability of the SF cells to proliferate in culture, by the presence of suppressor cells, by absorption of IL-2, or by a relative lack of accessory-cell function in the SF fraction. Culturing synovial cells without stimulant for three days resulted in a two- to three-fold increase in the apparent frequency of secreting cells, but the same proportional increase followed similar manipulation of peripheral blood cells. Pre-incubation of normal lymphocytes in RA synovial fluid did not result in suppression of their ability to secrete IL-2. We conclude that the poor secretion of IL-2 after mitogen stimulation of SF cells is an intrinsic property of the T-cell, and not due to the presence of other factors. There was no evidence that this defect was selectively reversible by "resting" the cells prior to culture.