We established a highly specific enzyme immunoassay (EIA) for 5 alpha-androstane-3 alpha, 17 beta-diol 17-glucuronide (androstanediol-17G). Rabbit antisera raised against 5 alpha-androstane-3 alpha, 11 alpha, 17 beta-triol 17-glucuronide 11-glutaryl bovine serum albumin and a heterologous tracer of androstanediol-17G conjugated with horseradish peroxidase at the glucuronic acid group were used. The EIA showed excellent specificity: there were no remarkable cross-reactivities with related androgens. The assay range for urine samples was 0.3-30 ng/ml. Recoveries of standards added to samples were 100-108%. Intra-assay and inter-assay coefficients of variation were 2.9-4.4% and 5.7-7.9%, respectively. The EIA was applied to urine samples of 407 males and 322 females to determine developmental patterns and normal ranges of androstanediol-17G excretions in 11 age groups (0 y, 1 y, 2-3 y, 4-5 y, 6-7 y, 8-9 y, 10-11 y, 12-13 y, 14-15 y, 16-17 y, and over 18 y). Urinary androstanediol-17G/creatinine (androstanediol-17G/Cre) ratios in both sexes were high in infancy, tended to decrease during childhood, and began to increase near adolescence. While androstanediol-17G/Cre ratio in girls increased at 8-9 y and reached a plateau during adolescence, that in boys increased at 10-11 y and continued to increase throughout adolescence. Androstanediol-17G/Cre ratios in girls were higher than those in boys at 6-7 y (P < 0.05) and at 8-9 y (P < 0.01). Androstanediol-17G/Cre ratios in boys were higher than those in girls at 12-13 y and at older ages (P < 0.01). These developmental patterns are parallel to age-related changes in androgenicity and serum androstanediol-17G, suggesting that urinary androstanediol-17G/Cre ratio could be a good marker for androgenicity in childhood.