Previous studies of the erythropoietin response to anaemia in RA have yielded conflicting findings. Some have found the response to be impaired and others have found a normal response. We have compared erythropoietin (EPO) levels measured by radioimmunoassay, in 54 anaemic rheumatoid patients and 55 patients with iron deficiency anaemia but no inflammatory disease. The erythropoietin response in the rheumatoid patients was impaired compared with the control group (P < 0.025) but only seven rheumatoid patients showed a response which fell below the 95% confidence intervals predicted for the control group. Rheumatoid patients who fell within the highest quartile for serum ferritin concentrations (i.e. those most likely to have anaemia of chronic disease) had significantly lower EPO levels compared with the control group (P < 0.01). EPO levels in rheumatoid patients within the lowest quartile for ferritin (i.e. those with iron deficiency anaemia) were not significantly different from the control group (P = 0.670). The difference in EPO response between the RA patients in the upper and lower quartile for ferritin approached but did not achieve significance (P = 0.056). In a second study 15 anaemic RA patients were given a 5-day course of oral prednisolone 1.5 mgkg-1. Hemoglobin did not rise significantly until day 4 but EPO levels fell by day 1 (P < 0.005) and remained lower than pretreatment values throughout the study. Thus, in RA patients, anaemia of chronic disease is associated with inappropriately low EPO concentrations but this does not appear to be the major cause of the anaemia and Hb response to prednisolone does not depend upon an increase in EPO concentration.