A new intraoperative method for blood flow measurement, ultrasonic transit-time flowmetry, was validated regarding its variation and error of measurement. Both common iliac arteries of five piglets were dissected under inhalation anaesthesia. Three probes, calibrated by the manufacturer, were used. Variability was evaluated by repeated measurements from the native artery and under exsanguination controlled with a withdrawal pump. The errors of measurement were calculated by comparing the flow with the exsanguinated volume by time. For each vessel 34-36 separate exsanguinations were carried out, with flow ranging from 10 to 500 ml/min. The coefficients of variation for the two 5 mm probes measured from the native artery were 3.8% and 4.2%, and 4.5% for the 3 mm probe. With the use of the withdrawal pump the coefficients were 2.5%, 2.2% and 2.1%, respectively. The two 5 mm probes showed a mean error of measurement of 0.7% (95% confidence interval (CI) was -10.4 to 13.25) and -11.4% (95% CI -20.6 to -1.1), while the error for the 3 mm probe was 3.4% (95% CI -18.1 to 14.0). The small variation of all probes indicated a good reproducibility. With accurate calibration of the probes, ultrasonic transit-time flowmetry appears to be a valid and adequate method for clinical use.