We have previously shown that in dogs with normal lungs, dependent placement of the puncture site reduced and/or stopped pneumothorax formation while breathing spontaneously. This experiment is now repeated in an acute injury model in dogs receiving either intermittent positive pressure ventilation (IPPV) or continuous positive pressure ventilation (CPPV). Pneumothorax was induced by percutaneous transthoracic insertion of a 20-gauge needle into the right lung. After a constant rate of pneumothorax formation was established (measured by evacuation of right pleural gas by chest tube), the dogs were alternated between 3 min with the puncture site up (left decubitus) or down (right decubitus). Only when the puncture was placed down did the rate of pneumothorax formation decrease (in every case), and this maneuver caused the leak to seal in 21 and 30% of the times with IPPV and CPPV, respectively. These results are consistent with the hypothesis that dependent placement of the puncture site allows reduction of both the alveolar size and alveolar to pleural pressure difference in the region surrounding the leak, thereby reducing and possibly stopping pneumothorax formation.