The effect of preventive intermittent calf compression on the incidence, distribution and extent of venous thrombosis after elective hip replacement was examined by randomized trial in 90 patients who were screened for postoperative thrombosis with 125I‐fibrinogen leg scanning and impedance plethysmography, followed by routine venography on the seventh postoperative day. Venography showed that leg compression reduced the incidence of calf vein thrombosis from 45 per cent (21/47) in untreated patients to 16 per cent (7/43) (P<0·005), but not that of proximal (i.e. popliteal or femoral) vein thrombosis, which occurred in 23 per cent of treated and 26 per cent of untreated patients. However, proximal vein thrombosis appeared to be less extensive in treated patients. Proximal vein thrombosis was found in 40 per cent of patients who had hip replacement by a modified Charnley technique (17/43 patients), and 9 per cent of patients in whom a posterior surgical approach was used (4/43 patients) (P<0·005), strongly suggesting that surgical technique may influence the proximal vein thrombosis rate after elective hip replacement.