Blood loss at surgery for pelvic floor repair was measured in 122 females anaesthetized by one of six standardized, orthodox methods. Two basic techniques of anaesthesia were used, one including spontaneous respiration and the other intermittent positive pressure ventilation (IPPV). Arterial hypotension was induced with trimetaphan in addition to the two basic techniques, and epidural anaesthesia was induced with the technique of spontaneous respiration. There was no difference in blood loss in relation to the mode of respiration employed. The addition of hypotension caused a fall of blood loss of about 50 per cent. Epidural analgesia also caused a significant fall of blood loss. At similar levels of arterial pressure blood loss in the epidural group was not significantly different from that in the hypotensive group. Patients over 65 years bled significantly less than those in a younger age group. The experience of the surgeon had no effect on blood loss, and operating time was unaffected by bleeding. The significance of these results is discussed. A pilot study in which epidural anaesthesia was combined with IPPV is described. The significance of the wide range of results is postulated to be that the causes of surgical bleeding are unknown.