1 Using an in vivo model of erectile activity, the eects of sildena®l were studied in mice lacking neuronal or endothelial nitric oxide synthase (nNOS and eNOS, respectively). 2 Under pentobarbitone anaesthesia, intracavernous pressure (ICP) and mean arterial pressure (MAP) were monitored continuously in wild-type, nNOS7/7 and eNOS7/7 mice. The magnitude of erectile activity was quanti®ed as the ratio of ICP to MAP. 3 No dierences in basal ICP or MAP were observed amongst wild-type, eNOS7/7 and nNOS7/7 mice. Electrical stimulation of the cavernous nerve (ESCN; 4.0 V, 16 Hz, 1 ms, 30 s) evoked increases in ICP and ICP/MAP as well as penile tumescence. Responses to ESCN were reduced in nNOS7/7, but not in eNOS7/7 mice. 4 L-NAME (50 mg kg 71 , i.v.) signi®cantly increased MAP and attenuated erectile responses in both wild-type and eNOS7/7 mice. 5 Sildena®l (1 mg kg 71 , i.v.) augmented electrically-evoked erectile activity in a voltage-dependent manner in wild-type mice and facilitated erectile responses in eNOS7/7 mice. By contrast, sildena®l failed to augment the diminished erectile responses in mice lacking the nNOS isoform. 6 These data reveal the relative importance of nNOS, compared to eNOS, as the critical NOS isoform in the control of erectile function and illustrate that the nNOS isoform is required for sildena®l-induced facilitation of erectile responses in vivo in mice.