To gain insight into the passage of androstenedione (A4) through the salivary gland, we measured concentrations of plasma total (TA4), free (FA4), and salivary (SA4) androstenedione during administration of dexamethasone and synthetic corticotropin to control subjects and hirsute women and compared these data with plasma total (TT), free (FT), and salivary testosterone (ST) concentrations. TA4 and FA4 were significantly lower in control subjects throughout (0, 15, 45, and 75 min) (P less than 0.05). SA4 in control subjects was significantly lower only in the follicular phase at 0 and 15 min. There was no significant difference between the increments in SA4 in response to corticotropin. The concentration of SA4 in the control women at 0 min was not significantly different from the concentration of FA4. At 45 and 75 min after corticotropin administration, however, SA4 was slightly higher than FA4 (P less than 0.01). In hirsute women, however, the concentration of SA4 was significantly lower than FA4 at all times (P less than 0.05). TT, FT, and ST concentrations were about twofold higher in the hirsute women than in control subjects throughout. In both groups, ST concentrations were three times as high as FT concentrations (P less than 0.001). The SA4:ST ratio was significantly lower than the FA4:FT ratio in both groups (P less than 0.001) because of higher ST than FT concentrations and similar or even lower SA4 concentrations in both groups. Both FA4:FT and SA4:ST ratios were lower in hirsute women, except for the FA4:FT ratio in control subjects in the luteal phase. Our data are compatible with 17-hydroxysteroid oxidoreductase activity in the salivary gland. If the difference in the ratios of A4 to T between hirsute women and control subjects is attributed to this hypothetical enzymatic activity, it would suggest a more rapid conversion of A4 to T in the hirsute group.