Both legs of 29 patients with venous disease and those of 15 controls without venous disease were assessed by duplex ultrasonography. The duration of reverse flow after release of manual calf compression was measured in the common femoral, long saphenous, popliteal and short saphenous veins. Before undertaking the study, the reproducibility of the technique was evaluated in six subjects by repeating the examination over 3 consecutive days; the coefficient of variation of the test was 7·3 per cent. The 95 per cent confidence interval (c.i.) of the median (0·16 s) of all measurements in the normal limbs was 0·12-0·18 s. The 95 per cent c.i. for the 95th percentile of all measurements in normal limbs was 0·32-0·52 s. In limbs with clinical evidence of venous disease at least one of the sites examined was found to have reverse flow lasting longer than 0·5 s. These data suggest that the measurement of reverse flow after release of manual calf compression is a reproducible technique. While the method records some reverse flow in normal veins, its duration is unlikely to exceed 0·5 s; significant reflux is therefore defined as reverse flow exceeding 0·5 s.