A method is described for the estimation of dihydrotestosterone (5α-androstan- 17β-ol-3-one) in human peripheral plasma. An aliquot of 5–10 ml of plasma with 3H indicator is extracted with solvent and chromatographed on low residue, methanol washed paper in the Bush B-3 system for 2–3 hr. The sample is then assayed by a displacement technique using diluted plasma from an estrogen treated or pregnant patient. Separation of bound from free steroid is obtained with small Sephadex G-25 columns. The % bound of the sample is converted to mass by comparison with a standard curve and corrected for recovery. Testosterone and dihydrotestosterone can be assayed simultaneously. Duplicate samples can be measured with an error of 10% for male and 13% for female samples and there is no indication for the presence of systematic errors. The nonspecific blank is 0.15±0.05 (sd) ng/sample. Prepubertal plasma did not contain measurable amounts of dihydrotestosterone. Adult male values are 54±19 (sd) and female plasma contains 14.8±5.4 (sd) ng/100 ml. No significant difference was observed between follicular and luteal phase values. Testosterone and dihydrotestosterone values from a small group of hirsute patients indicated that testosterone, dihydrotestosterone, or both androgens, may be increased.