No completely satisfactory experimental model exists to compare different techniques of preservation currently used in the distant procurement of a lung allograft. A canine model of left lung transplantation is described in which an inflatable cuff is placed around each pulmonary artery. Each cuff is connected to a subcutaneous reservoir, which allows alternate occlusion of either pulmonary artery. Functional assessment of lung function is made during ventilation of both lungs and after a 10-minute period of perfusion to the native lung alone and then to the transplanted lung alone. Systemic and pulmonary artery pressures are recorded continuously, and measurement of arterial blood gases and oxygen uptake are made immediately after the operation and again at 3 days. The animal is then put to death and the lungs are excised and weighed. Five dogs underwent transplantation of the donor lung immediately after excision (mean ischemic time = 55 +/- 7 minutes). Similar values for oxygen tension and oxygen uptake were obtained postoperatively for the right lung (oxygen tension = 420 mm Hg, oxygen uptake = 101 ml/min) and the left lung (oxygen tension = 368 mm Hg, oxygen uptake = 108 ml/min). However, carbon dioxide tension was elevated (right lung = 41 mm Hg, left lung = 52 mm Hg). Mean pulmonary artery pressure increased during allograft perfusion (right lung = 14 mm Hg, left lung = 24 mm Hg), although systemic blood pressure was unchanged. Similar results were observed at 3 days. The mean weight of the native lung was 101 +/- 2 gm and that of the transplanted lung, 128 +/- 6 gm. This model achieves consistent survival and allows serial observations of the functional adequacy of an allograft compared with a normal contralateral lung.