The validity and reproducibility of measurements of the luminal diameter of the radial artery in man were investigated with Dermascan A, a 20 MHz ultrasound scanner. The luminal diameter of a reference object, a plastic tube, was measured with ultrasound to be 99.0% of stereomicroscopic measurements. By comparing the diameter of the reference object either filled with water or blood at 37 degrees C, the ultrasound velocity in human blood at 37 degrees C was calculated to be 1605 m/s. The intraobserver repeatability coefficients of in-vivo measurements of the radial artery in man were in the same range, whether measurements were repeated after 30 min (14%) or from day to day (12%). The interobserver repeatability coefficient was acceptable (15%) when the site of measurements was marked, whereas measurements without a mark resulted in a repeatability coefficient of 24%. The intravenous administration of 0.5 mg ergotamine tartrate to 5 subjects caused a decrease in the mean luminal diameter of the radial artery from 2.94 mm to 2.42 mm 1 h after ergotamine (p less than 0.05). This ultrasound method for measurements of the luminal diameter should thus be suitable for investigating the effects of physiological and pharmacological stimuli on the arteries per se.