Thirty-six consecutive unselected patients, who had apparently previously undergone saphenofemoral ligation for primary uncomplicated long saphenous varicosities and who had then re-presented with recurrent thigh varices emanating from the groin, underwent preoperative clinical assessment, hand-held Doppler and duplex ultrasonographic examination and varicography to establish the presence or absence of saphenofemoral incompetence as the cause of recurrence. All patients underwent reexploration of the saphenofemoral junction (SFJ) via a lateral approach. Twenty-six patients had an intact SFJ (type I recurrence) and ten had varices arising from either a thigh perforator, or from abdominal or perineal veins (type II recurrence). Clinical examination alone was poor at distinguishing type I from type II recurrence. Doppler ultrasonography was sensitive (88 per cent) but non-specific (40 per cent). In contrast, duplex scanning was insensitive (42 per cent) but extremely specific (100 per cent) and accurate, with a positive predictive value of 100 per cent. Varicography also had a specificity and positive predictive value of 100 per cent, a sensitivity of 73 per cent and in addition provided a precise anatomical 'road-map'. A combination of clinical examination and hand-held Doppler ultrasonography seems to be the most appropriate first-line method of preoperative assessment in these patients. Duplex ultrasonography, if available, will provide additional useful information about both the SFJ and the presence of thigh perforators. Contrast examination may be reserved for patients who have equivocal results on non-invasive investigations, who have had more than one previous groin operation or who have, in addition, deep venous disease.