The occurrence of postoperative deep vein thrombosis (DVT) was studied in 209 elective surgery patients aged 40 years or over. Most of the operations were cholecystectomies or other major abdominal operations. In the detection of thrombosis the 125I-fibrinogen method was used. Of 209 patients, 51 (24.4%) developed postoperative DVT and of these 10 developed thrombosis in both legs. In 36% of the DVT cases the process started during the operation or on the first postoperative day. Clinical signs of DVT did not agree with the 125I-fibrinogen test very well, whereas the correlation of the 125I-fibrinogen test with phlebography was good: of the 17 patients with a postive 125I-fibrinogen test in whom a phlebography was done, the latter method revealed thrombosis in 14 patients (82.4%). The main purpose of the study was to determine whether the mode of mechanical ventilation used during anaesthesia has any effect on the occurrence of postoperative DVT. One hundred and nine patients received intermittent positive pressure ventilation (IPPV), whereas in 100 patients the intrathoracic pressure was decreased by applying intermittent positive-negative (--5 cmH2O) pressure ventilation (IPNPV). Ventilation was standardized in both groups by keeping the end-tidal CO2% at a constant level. Inspired oxygen concentration was the same in every patient. The decreasing effect of IPNPV on intrathoracic pressure and central venous pressure as compared with IPPV was confirmed in preliminary studies.