Seventeen individuals from a pedigree with complete androgen insensitivity, [testicular feminization (TF)] are presented. Their hormonal evaluation was compared with those of normal males and male pseudohermaphrodites with primary 5a-reductase deficiency. The mean plasma testosterone to dihydrotestosterone ratio was 12 ± 3 in normals, 24 ± 8 in TF subjects (P < 0.001), and 41 ± 14 (P < 0.001) in 5a-reductasedeficient subjects. In 4 TF subjects the MCRs for testosterone and dihydrotestosterone were normal. The dihydrotestosterone blood production rate averaged 383 jug/day in normals, 162 jig/ day in TF subjects, and 86 /ig/day in 5a-reductase-deficient subjects. The conversion ratio of testosterone to dihydrotestosterone averaged 2.53 in normals, 1.8 in TF subjects, and 0.63 in 5a-reductase-deficient subjects. The mean plasma estradiol level was 2.8 ± 1.0 ng/100 ml in normal males, 4.8 ± 1.3 ng/100 ml (P < 0.001) in TF subjects, and 3.1 ± 1.3 ng/100 ml (P < 0.5) in 5a-reductase-deficient subjects. The fractional plasma protein binding of testosterone in TF subjects and 5a-reductase-deficient subjects was similar to that in normal males. The mean urinary etiocholanolone to androsterone ratio was 0.87 ± 0.34 in normals, 1.28 ± 0.46 (P < 0.001) in TF subjects, and 4.90 ± 2.15 (P < 0.001) in 5a-reductase-deficient subjects. The mean urinary ratio of 5/?-tetrahydrocorticosterone to 5a-tetrahydrocorticosterone was 0.53 ± 0.22 in normal males, 0.76 ± 0.21 in TF subjects (P < 0.02), and 4.59 ± 4.5 (P < 0.001) in 5a-reductase-deficient subjects. The mean urinary 5/?-tetrahydrocortisol to 5a-tetrahydrocortisol ratio was in the normal male range in the TF subjects, but was markedly elevated in the 5a-reductase-deficient subjects. The data suggest that in the TF subjects, there is a decrease in peripheral 5a-reductase activity related to C-19 androgen 5a-metabolism, which is a secondary manifestation of androgen resistance. This differs from the situation in the male pseudohermaphrodites with 5a-reductase deficiency, where the defect affects hepatic and peripheral 5a-reduction with a marked decrease in both 5a C-19 and C-21 metabolites. (J Clin EndocrinolMetab 54: 931, 1982)