Early pubertal boys (testicular volume, 4-6 mL) with constitutionally delayed growth and puberty were randomized to 3 months of treatment after a baseline 12-h overnight hormone profile: group 1 (n = 5), daily placebo; group 2 (n = 5), 2.5 mg oxandrolone daily; or group 3 (n = 6), 50-mg testosterone monthly im injections. LH and GH profiles (15-min samples) were analyzed by peak detection (Pulsar), Fourier transformation, and autocorrelation. FSH and testosterone levels were measured hourly, and insulin, sex hormone-binding globulin, insulin-like growth factor-I, and insulin-like growth factor-binding protein-3 levels were determined at 0800 h. Multiple regression was used to analyze the response to treatment (growth) with respect to baseline features. Endocrine variability was marked. Profiles ranged from unreactive to well established LH pulsatility and adult testosterone levels. The areas under the curve (AUC) for LH, FSH, and testosterone ranged 10-fold (4.4-46.3 IU/L.h), 8-fold (7.9-63.4 IU/L.h), and 45-fold (3.6-161.7 nmol/L.h), respectively. The growth response was individually varied, but significantly increased 0-6 months in the active treatment groups. Age, testicular volume, and LH AUC interacted significantly (r2 = 0.95; P < 0.05). Allowance for these produced a highly significant treatment effect (P = 0.006). Age, testicular volume, LH AUC, and testosterone AUC, but not treatment, significantly increased growth by 0-12 months (r2 = 0.88; P < 0.05). We demonstrate a spectrum of activation of the reproductive axis despite tight clinical staging. This, and not GH status at treatment commencement, influenced the growth response.