Eight castrate, estrogen-replaced women were given 200 mg spironolactone daily for 4 weeks. The response of plasma dehydroepiandrosterone sulfate (DHEAS), testosterone (T), and androstenedione (delta 4A), all indicators of adrenal C19-steroid production, varied greatly among individuals. Sixteen women with idiopathic hirsutism were given night-time dexamethasone (DEX) and then superimposed spironolactone for 4 weeks, followed by DEX without spironolactone for an additional 4 weeks. As expected, DHEAS, T, and delta 4A declined on DEX treatment. On addition of spironolactone, there was little further change in DHEAS, while plasma T declined in 7 of 16 women, including all those whose T level had remained elevated despite DEX treatment; most values rebounded when spironolactone was discontinued. The authors conclude from intact DEX-suppressed women that ovarian T, especially when increased, is frequently lowered by spironolactone. Thus, both adrenal and ovarian androgen production (as measured by prevailing plasma levels) may be diminished by this agent. These highly variable effects on androgen production are unlikely to account for the consistent antiandrogenic effects reported clinically.