1 In the present study, we have investigated the e ect of berberine in rabbit isolated corpus cavernosum and measured the intracavernous pressure (ICP) change after intracavernosal injection of berberine in rabbit. 2 Berberine alone suppressed the basal tone and induced a concentration (0.1 ± 100 mM)-dependent relaxation in phenylephrine (PE)-precontracted corpus cavernosum. 3 Tetrodotoxin (0.1 and 1 mM) treatment had no signi®cant e ect on the berberine-induced relaxation. Phentolamine (1 and 10 mM), propranolol (1 and 3 mM) and atropine (1 and 3 mM) were also without e ect. These results suggest that berberine might cause relaxation of the cavernosal strip by direct action on the corpus cavernosum, not by a neuronal e ect. Furthermore, muscarinicand b-adrenoceptors were not involved. 4 Berberine-induced relaxations were signi®cantly reduced by endothelium removal and by exposure to L-N G -nitro arginine methyl ester (0.1 and 0.3 mM), but not indomethacin (30 mM). 5 In endothelium-deprived corpus cavernosal tissues, berberine-induced relaxations were signi®cantly reduced in high K + medium (KCl=60 mM), by charybdotoxin (ChTX) and 4-aminopyridine (4-AP) but not by glibenclamide and apamin. 6 After intracavernous injection of berberine (1, 2, 3 and 5 mg kg 71 ), the ICP rose from 12.7+3.6 to 13.2+5.4, 25.3+6.1, 46.5+8.2, and 63.4+10.2 mmHg, respectively. The duration of tumescence ranged from 11.5 ± 43.7 min. 7 The results show that berberine possesses a relaxant e ect on rabbit corpus cavernosal tissues which is attributable to both endothelium-dependent and-independent properties. While the former component is apparently due to the release of NO from sinusoidal endothelium, the endotheliumindependent mechanism involved in berberine relaxation is probably linked to ChTX-and 4-APsensitive K + channel activation in the cavernosal vasculature.