1. This study examined the potential role of impaired nitric oxide production and response in the development of endoneurial ischaemia in experimental diabetes. Rats were anaesthetized (Na pentobarbitone 45 mg kg-1, diazepam 2 mg kg-1) for measurement of sciatic nerve laser Doppler flux and systemic arterial pressure. Drugs were administered into the sciatic endoneurium via a microinjector attached to a glass micropipette. 2. In two separate studies comparing diabetic rats (streptozotocin-induced; 8-10 wk duration) with controls, nerve Doppler flux in diabetic rats (Study 1, 116.6 +/- 40.4 and Study 2, 90.1 +/- 34.7 (s.d.) in arbitrary units) was about half that measured in controls (219.6 +/- 52.4 and 212.8 +/- 95.5 respectively; P < 0.005 for both). There were no significant differences between the two in systemic arterial pressure. 3. Inhibition of nitric oxide production by microinjection of 1 nmol L-NAME into the endoneurium halved flux in controls (to 126.3 +/- 41.3 in Study 1 and 102.1 +/- 38.9 in Study 2; both P < 0.001), with no significant effect in diabetic rats, indicating markedly diminished tonic nitric oxide production in the latter. D-NAME was without effect on nerve Doppler flux. 4. L-Arginine (100 nmol), injected after L-NAME, markedly increased flux in controls (by 65.8% (P < 0.03) and 97.8% (P < 0.01) in the two studies) and by proportionally similar amounts in diabetic rats [75.8% (P < 0.001) and 60.2% (P < 0.02)]. The nitro-donor, sodium nitroprusside (SNP; 10 nmol) had similar effects to L-arginine in both groups (increases of 66.0% in controls and 77.5% in diabetics; both P < 0.002). 5. A second diabetic group, treated with evening primrose oil performed exactly like control rats in respect of responses to L-NAME, L-arginine and SNP. 6. These findings implicate deficient nitric oxide in nerve ischaemia of diabetes and suggest correction thereof as a mechanism of action of evening primrose oil.