Testosterone (T) is required for spermatogenesis, particularly in the conversion of round spermatids between stages VII and VIII of spermatogenesis. T is generally thought to be the major androgen involved in adult spermatogenesis due to the high local concentration within the testis, whereas its more potent 5α-reduced metabolite dihydrotestosterone (DHT) is thought to be the active androgen in peripheral tissues. The current study investigated whether 5α-reduction of T to DHT is involved in the restoration of spermiogenesis in vivo in a setting in which testicular T levels were markedly reduced. Adult male rats were given 3-cm T plus 0.4-cm estradiol implants for 9 weeks to suppress serum LH and testicular T levels and thereby inhibit spermatogenesis. Increasing doses of T (3-, 6-, 10-, and 24-cm implants) were then given for 4 days to restore spermatogenesis in the presence or absence of a 5α-reductase inhibitor (L685,273). The hourly production rates of round spermatids in stages I-III, IV-VI, VII, and VIII were assessed using stereological techniques, and the conversion of round spermatids between stages VII and VIII was then assessed as an index of androgen action on spermiogenesis. Testicular androgen levels were measured by HPLC and RIA. The 5α-reductase inhibitor significantly suppressed (P <0.05) the hourly production rate of round spermatids at the 3- and 6-cm T doses, but not at the 10- and 24-cm doses. The conversion of round spermatids between stages VII and VIII was suppressed (P <0.05) by the inhibitor only at the 3- and 6-cm doses. The 5α-reductase inhibitor had no effect on testicular T levels, but suppressed (P <0.05) DHT levels at the 6-, 10-, and 24-cm doses. We conclude that the 5α-reduction of T is involved in the restoration of spermiogenesis at the lower administered doses of T and that these data are the first description of a role for 5α-reduced androgens in adult spermatogenesis.