It has been reported that there is no relationship between serum testosterone level and erectile function and that testosterone mainly acts on the brain and not on the penis. However, we have found that erectile dysfunction in some patients with low serum testosterone improved after androgen replacement therapy. We therefore studied the effects of testosterone on the contraction and relaxation of the corpus cavernosum penis. Four groups of rabbits were studied over a similar time span. The testosterone group were non-operated and received testosterone propionate 10 mg/day for 14 days before sacrifice. Two groups (castration+testosterone group and castration group) were castrated 28 days before sacrifice. The castration group received sesame oil for 14 days, and the castration+testosterone group received testosterone 10 mg/day for 14 days. Entire penis were removed immediately after death by desanguination and dissected to obtain corpus cavernosum strips. These strips were suspended under 1 g tension in a 20 ml organ bath containing Tyrode's solution. After equilibration for 2 hours, the addition of drugs was started. Changes in tension were monitored with an isometric transducer and recorded. The vasoactive drugs used were norepinephrine bitartrate, 1-phenylephrine hydrochloride, clonidine hydrochloride, acetylcholine chloride and vasoactive intestinal polypeptide. The contractility of the corpus cavernosum strips increased significantly in the testosterone group and the castration + testosterone group, but decreased significantly in the castration group as compared with the control after adding norepinephrine, phenylephrine and clonidine.(ABSTRACT TRUNCATED AT 250 WORDS)