The effect of isosorbide dinitrate (ISDN), 5 mg sublingually, on the diameters of coronary artery stenoses (n = 27) was examined in 20 patients. Another 18 patients with angiographically normal coronary arteries received the same amount of ISDN and were used as controls. Prestenotic and stenotic diameters were measured with a vernier calliper having an accuracy of 0.05 mm. The degree of stenosis was expressed as percentage of cross-sectional area reduction. ISDN caused uniform dilatation of every normal epicardial artery; mean increase in diameter was 21 per cent (range: 17 to 26%). In 18 stenoses (28 to 95% obstruction) there was very little change after ISDN. The mean prestenotic diameter increased from 2.82 +/- 0.48 mm to 3.05 +/- 0.43 mm and the mean stenotic diameter from 1.45 +/- 0.49 mm to 1.59 +/- 0.51 mm. However, in the nine other stenoses (35 to 89% obstruction) the mean degree of obstruction decreased significantly from 68 +/- 15.6 per cent to 47 +/- 15.6 per cent after ISDN. This improvement was a result of a significant increase of the mean stenotic diameter from 1.71 +/- 0.47 mm to 2.41 +/- 0.55 mm, whereas the prestenotic diameter showed only an insignificant increase from 3.17 +/- 0.63 mm to 3.31 +/- 0.58 mm after ISDN. In four patients with two obstructions in different coronary branches ISDN dilated one without significantly affecting the other lesion. From the data we conclude that ISDN can dilate some coronary artery stenoses but that this response may vary from one site to another even in the same patient.