Vasodilatation elicited by topical application of methyl nicotinate was measured by photoplethysmography at various positions on human ventral forearm skin. The time-to-peak response, the magnitude of the peak response, the area under the response-time curve and the time for the response to decay to 75% of the maximum value, were recorded at six positions on the left and right forearms of eight subjects. There was no significant difference between response on the right and left forearm sites and no difference in response between the lateral and medial sites on the forearms. In a second experiment, the vasodilatation was measured at proximal and distal positions on the forearm. The magnitude of the peak response at the proximal position was significantly higher than that found distally (P less than 0.01) as was the area under the response-time curve (P less than 0.05). These results suggest a possible cause for the variation observed in vasoconstrictor assays of corticosteroids.