Endothelial dysfunction might be an important and early event in the pathogenesis of major cardiovascular diseases. Therefore, the evaluation of endothelial function in humans may be of great clinical relevance. Usual methods for that purpose are either invasive and/or technically demanding. In the dermal microcirculation, endothelial function may be assessed noninvasively from the laser Doppler measurement of increases in blood flow after either the transdermal application of acetylcholine by iontophoresis, or the release of transient arterial occlusion (reactive hyperemia). An endothelium-independent response may be provided by the iontophoresis of sodium nitroprusside. This approach is notable for technical simplicity, but of uncertain reproducibility. Sixteen young, healthy, nonsmoking males were examined in the fasting state. Changes in skin blood flow were measured with a laser Doppler imager during the iontophoresis of acetylcholine and sodium nitroprusside, as well as during reactive hyperemia, on two different days, at each of two different sites on the volar face of the forearm. Nonspecific effects related to the stimulation of terminal nerve fibers by the iontophoretic current were suppressed by prior surface anesthesia. The iontophoresis of acetylcholine and sodium nitroprusside induced a seven- to eightfold increase in dermal blood flow. The corresponding figure for peak reactive hyperemia was approximately fourfold. The mean coefficients of variation of responses recorded on different days, on the same site, in the same individual were <10% for iontophoresis of acetylcholine and for peak reactive hyperemia, and between 10 and 20% for iontophoresis of sodium nitroprusside. This day-to-day variation was significantly smaller than the site-to-site variation (p < 0.01 for all three responses). Endothelium-dependent and -independent responses of dermal blood flow evaluated with laser Doppler imaging are highly reproducible from day to day, at least in healthy nonsmoking young male subjects, and provided some simple precautions are observed, foremost among which is the strict standardization of the recording site. These observations may have implications for the testing of endothelial function in clinical studies.