We present a retrospective audit of all arterial embolectomies performed at the East Glamorgan General Hospital over an 11-year period (1980-1990). Eighty-seven patients (47M:40F), mean age 67 years (50-90 years) underwent 95 embolectomies, an incidence consistent with previous studies. There were 17 upper and 71 lower limb emboli with a mean delay before diagnosis of 29 h (range 1-264 h). In 66% of cases the cause was atrial fibrillation; 33% received immediate heparinization and 14% prophylactic antibiotics. Surgery was performed by a consultant in 12 and registrars in 75 cases, and under local anaesthesia in 80% and general anaesthesia in 20%. There was no anaesthetist present in 54% of cases. Few pre- or peroperative arteriograms were performed. The 30-day mortality was 45%, with an amputation rate of 15% and an overall postoperative complication rate of 62% with little improvement in these figures over the last 10 years. Factors increasing mortality were: delay before diagnosis, grade of surgeon performing the operation, and inadequate inflow or outflow at operation. Factors found to affect limb salvage rate adversely were a history of intermittent claudication, although such a history was not recorded in many cases, and lack of immediate preoperative heparinization. Although embolectomy is considered a 'registrar operation', reviewing our results it can be seen that it is an uncommon operation, in our series eight or nine being performed annually. Sometimes inappropriate surgery is performed upon patients in whom severe systemic illness may contraindicate any form of surgical intervention.(ABSTRACT TRUNCATED AT 250 WORDS)