Androgens play an important role in erectile function. However, the dose-response relationship between plasma testosterone levels and penile erection remains unclear. Intact (sham operated) or bilaterally orchiectomized mature male Sprague-Dawley rats were used. Two weeks after surgery, rats were infused continuously with either vehicle (polyethyleneglycol) or varying doses of testosterone (44, 88, 220 or 440 µg/day) for 14 days using subcutaneous osmotic infusion pumps (Study 1). In a separate study, 4 weeks after surgery, rats were infused with a lower range of testosterone doses (11, 22 or 44 µg/day) for 14 days (Study 2). In the first study, intact rats had a mean plasma testosterone concentration of 0.56 ± 0.12 ng/ml (1.9 nM), as determined by standard radioimmunoassay. In the second study, a more sensitive enzyme-linked immunoassay was used to measure the lower testosterone levels. Using this assay, intact rats had a mean plasma testosterone concentration of 2.02 ± 0.59 ng/ml. Intracavernosal pressure measurements indicated that orchiectomy resulted in a significant reduction in erectile function, when compared to intact animals, while testosterone infusion restored erectile function to varying degrees. Erectile function was maintained by a wide range of systemic testosterone levels as low as 10-12% of normal physiological plasma concentrations. Below these concentrations, erectile function was significantly and positively correlated with testosterone plasma levels in a dose-dependent manner. Interestingly, prostate tissue mass was positively correlated to plasma testosterone levels across all concentrations examined. Protein expression of neural nitric oxide synthase (nNOS) and phosphodiesterase type 5 (PDE 5) was reduced in penile tissue from orchiectomized animals and increased in testosterone-infused animals, as assessed by Western blot analyses. We suggest that testosterone at levels approaching one-tenth of the normal physiological plasma concentration may represent a threshold value, below which erectile function declines in a dose-dependent fashion. However, different androgen-dependent tissues may exhibit varying sensitivities to circulating testosterone with regard to growth and function.