Although testosterone (T) has striking effects on mature skeletal size and structure, it is not clear whether this depends exclusively on adult circulating levels of T or whether additional early-life factors also play a role. We have compared the androgen-deficient hypogonadal (hpg) mutant mouse with intact, orchidectomized, and T-treated non-hpg mice to determine relative contributions of adult and perinatal T to bone growth and development. At 3 wk of age, although trabecular and cortical bone structure was normal, bone turnover was significantly altered in hpg male mice; osteoid volume (OV/BV) and osteoblast surface (ObS/BS) were significantly lower and osteoclast surface (OcS/BS) significantly higher in hpg mice compared with age-matched non-hpg mice, pointing to a role for the perinatal T surge in determining bone turnover levels before sexual maturity. At 9 wk of age, the hpg bone phenotype mimicked closely that of age-matched non-hpg mice that had been orchidectomized at 3 wk of age, including low trabecular bone mass and high bone turnover. These bone phenotypes of hpg and orchidectomized non-hpg mice were all prevented by replacement doses of T or dihydrotestosterone (DHT), suggesting that these are determined by adult sex steroid hormones. In contrast, a short bone phenotype that could not be prevented by T or DHT treatment was observed in 9-wk-old hpg mice yet not in intact or castrated non-hpg mice. These data suggest a role for the perinatal T surge in determining adult bone length and confirms that adult circulating T determines adult bone density.