To assess the influence of compromised pulmonary function on the risk of thoracotomy, two groups of patients over 40 years old were compared. The 85 patients with impaired function (group 1) had less than 60% of predicted value for at least two of four standard spirometric tests. Fifty-seven patients comprised group 2 (control), and a majority in both groups had pulmonary resection. Fifty percent of group 1 patients who received tracheostomy or mechanical ventilation failed to survive. Postoperative respiratory insufficiency occurred in 17 patients, with nine deaths. Hospital mortality for group 1 was 18.8%, with bronchopleural fistula and pneumonia with respiratory insufficiency as the principal causes of death. A mortality of 10.5% in group 2 suggested that operative risk is approximately doubled by compromised pulmonary function and that technical or infectious complications exert greater influence than reduced pulmonary function per se.