We studied the pharmacokinetics and pharmacodynamics of sc implanted pellets of fused crystalline testosterone. Three different regimens (6 x 100 mg, 6 x 200 mg; and 3 x 200 mg) were compared in a prospective, cross-over clinical trial in which androgen deficient men were administered the three-dose combinations in a randomized starting order at intervals of at least 6 months. Plasma free and total testosterone, sex hormone-binding globulin, LH, and FSH were measured before and at monthly intervals for at least 6 months after 111 pellet implantation in 43 men with hypergonadotropic (n = 22) or hypogonadotropic (n = 21) hypogonadism. Total and free testosterone levels peaked at the first month and were maintained at physiological levels for 4 to 5 (600 mg doses) or 6 (1200 mg dose) months after a single implantation. Absorption of testosterone from 100 mg and 200 mg pellets closely approximated zero-order throughout the effective life of the pellets and exhibited a half-duration of 2.5 months. The estimated rate of release of testosterone was 1.5 (95% confidence limits 1.3-1.6) mg/day.200 mg pellet as determined from direct measurement of residue in pellets recovered after extrusion and confirmed independently from percent absorbed-time plots. The bioavailability of testosterone was virtually complete and the time course was predictable from the total implant dose and, to a lesser extent, total initial surface areas of pellets. Despite wide fluctuations in testosterone, SHBG levels were not changed during 6 months. In men with hypergonadotropic hypogonadism, both LH and FSH levels were uniformly and markedly suppressed by increased testosterone after pellet implants. LH and FSH were highly correlated with each other (r = 0.87) and inversely with total (r = 0.47 and 0.45, respectively) and free (r = 0.46 and 0.47) testosterone levels. Nadir LH levels occurred at 1-3 months (600 mg) and 1-4 months (1200 mg) reaching levels comparable with eugonadal controls. In contrast nadir FSH levels occurred at similar times but remained elevated compared with eugonadal controls. We conclude that fused pellets of crystalline testosterone provides very satisfactory depot androgen replacement exhibiting many desirable features for androgen replacement.