The routine clinical use of pulsatility index (PI) in the quantitative assessment of Doppler blood velocity/time waveforms, was assessed in normals and in patients with leg ischaemia. Two thousand and five hundred common femoral artery waveforms were studied in 57 subjects (38 males, 19 females). PI values showed a mean variability of 17% in subjects in normal sinus rhythm (NSR) (range 4.1–13.2). In 17 patients in atrial fibrillation (AF) but with no clinical arterial disease, PI varied 26%. Variability was reduced, using a ten beat average, to 7.5% in NSR and 9% in AF.Rise in PI correlated well with increase in peripheral resistance (R = 0.82) but PI was poorly related to arterial stiffness (R = 0.50). Although PI was able to identify patients with proximal arterial narrowing it was unable to distinguish the severity of disease as assessed arteriographically by minimum diameters (R = 0.30).From our observations we conclude that while PI has an established place in the assessment of peripheral arterial disease it must be applied with caution and averaging from ten heart beats is an essential manoeuvre. PI may prove to be a more reliable index for assessing distal rather than proximal disease.