Serum levels of testosterone (T), dehydroepiandrosterone (DHEA), and DHEA sulfate (DHEA-S) were measured in a group of 285 women with clinical signs of hyperandrogenism (hirsutism and/or acne). Levels of T were elevated in 75.8% of the patients, DHEA in 57.5%, and DHEA-S in 20%. Normal levels of all 3 androgens were found in 13.3%. Significant correlations were demonstrated among levels of all androgens. Only 7.7% of the patients had elevated levels of DHEA in the presence of normal T, and 3.2% had elevated DHEA-S and normal T levels. Subdivision of the study population on the basis of presence of acne, hirsutism, or both demonstrated no significant differences in androgen levels. Amenorrheic women had higher serum T levels than those with menstrual cycles. Women with laparoscopically demonstrated polycystic ovaries had significantly higher serum androgen levels than hyperandrogenic women with no laparoscopic evidence for polycystic ovarian disease. All 285 patients were treated with chronic low dose prednisone therapy. Overall suppression of all 3 androgens occurred in a large proportion of the patients. The pretreatment levels of DHEA or DHEA-S did not predict the responsiveness of T to chronic glucocorticoid therapy. On the other hand, a 2-day dexamethasone (DEX) suppression test quantitatively predicted the degree of T suppression achieved by chronic therapy. An overnight DEX suppression test was considerably less precise for this purpose. In conclusion, chronic low dose prednisone therapy resulted in suppression of serum T levels in a large proportion of women with hyperandrogenism of undetermined cause. The response to therapy could not be predicted on the basis of pretreatment serum DHEA or DHEA-S levels, but was predicted with a 2-day DEX suppression test.