Hydroxyl radical production, detected by ethylene formation from methional, has been investigated in plasma, lymph and synovial fluid. In the presence of added iron--EDTA, addition of either H2O2 or xanthine and xanthine oxidase gave rise to hydroxyl radical formation that in most cases was not superoxide-dependent. The ascorbate already present in the fluid appeared to participate in the reaction. In the absence of added catalyst, the reaction was hardly detectable, the rate being less than 5% of that observed with 1 microM-iron--EDTA added. This implies that the fluids had little if any capacity to catalyse hydroxyl radical production via this mechanism.