The production of superoxide by the peripheral blood neutrophils of 19 patients with active rheumatoid arthritis was measured during treatment with sulphasalazine (SASP). The response to drug treatment was determined by change in plasma viscosity, CRP, early morning stiffness and articular index over a 10-point scale. Of the 19 patients studied, eight were considered to have responded well to SASP and seven to have responded poorly or not at all. Over the treatment period, plateau levels of superoxide production fell in seven of the eight responders (P = 0.028) compared with a non-significant fall in 3/7 of the non-responder groups. The initial rate of superoxide production also fell in the responder group, but this was not statistically significant. Initial values in both the responder and non-responder groups were comparable with those seen for normal controls. Analysis of drug levels showed all patients to be compliant with drug treatment; however, drug levels and neutrophil activity were not correlated. Studies of the effect of SASP and sulphapyridine on superoxide production in vitro showed no difference between good and poor responders. These results suggest that there is no inherent difference between good and poor responders regarding the susceptibility of their neutrophils to SASP. SASP's action on neutrophils, therefore, appears not to be its main mechanism of disease-modifying activity in RA.