Production rates of testosterone were determined in healthy men and women during the follicular phase of their menstrual cycle using the stable isotope dilution technique and analysis by gas chromatography/mass spectrometry. In an initial series of experiments, 0.07 mg/h (men) or 0.01 mg/h (women) 1,2-d-testosterone was infused for 36 h. After an equilibration period of 12 h, blood samples were obtained at 20-min intervals throughout 24 h. In men, no diurnal rhythmicity of testosterone production was observed, whereas in women, testosterone production rates were largest from 0400-1200 h. In a second series of experiments, the infused dose of 1,2-d-testosterone was reduced to 0.015 mg/h (men) and 0.0001 mg/h (women), respectively. Blood samples were obtained at 20-min intervals during the last 12 h (0800-2000 h) of the observation period. In accordance with results obtained by others using radioactive tracers, estimated production rates were 147 +/- 31 microg/h (3.7 +/- 2.2 mg/day in men) and 1.8 +/- 0.6 microg/h (0.04 +/- 0.01 [corrected] mg/day in women). To reduce both the duration of the experiment and the number of samples to be processed, we subsequently demonstrated that similar production rates may be obtained when the equilibration period before blood sampling is reduced to 6 h and the period of blood sampling is reduced to 4 h. This protocol is suitable for clinical use in a routine setting to obtain analytically correct estimates of testosterone production in vivo.