Serum testosterone and especially free testosterone is one of the parameters commonly used to evaluate androgen excess or deficiency. The authors equilibrated serum samples with 14C-labeled testosterone followed by an ammonium sulfate precipitation to compare the "apparent free testosterone concentration" with "total" serum testosterone concentration in the following populations: normal males and females; females presenting with gynecologic problems, particularly hirsutism and/or virilization; and males and females on maintenance hemodialysis. Total serum testosterone for each specimen was assayed with five different commercially available RIA kits encompassing a variety of technics: direct assay technics, assays utilizing extraction procedures prior to RIA; tritium-labeled tracer as well as iodine-labeled tracers. Clinical correlations improve strikingly when apparent free testosterone concentrations rather than total serum testosterone concentrations are used.