Pulmonary-artery (PA) wedge pressure (Ppaw) is usually considered to be an accurate reflectiokn of left atrial pressure (Pla), Recent reports had demonstrated that during controlled positive-pressure ventilation (CV) in normal lungs, a progressive increase in positive end-expiratory pressure (PEEP) produced a progressive positive Ppaw-Pla discrepancy when the PA catheter tip was positioned vertically above the left atrium (LA). In seven anesthetized dogs the Ppaw-Pla relationship (as transmural pressures) was studied at PEEP = 0, 5, 10, 15, and 20 torr, in two ways: 1) in compliant lungs with the PA catheter tip above the LA, CV was compared with spontaneous ventilation (SV); 2) during CV with PA catheter tip both above (Cup) and below (Cdown) the LA compliant lungs were compared with noncompliant lungs. A 35 per cent decrease in static pulmonary compliance was induced by administering oleic acid, 0.06 ml/kg, intravenously, and 0.01N HCI, 5 ml, intratracheally. In compliant lungs with the PA catheter tip above the LA during CV, Ppaw was significantly greater than Pla at PEEP ≥ 5 torr and equalled PEEP when PEEP ≥ 10 torr, whereas during SV, Pla at any PEEP. During CV in compliant lung, Ppaw recorded from Cup and Cdown was significantly greater than Pla at PEEP ≥ 5 and PEEP ≥ 15 torr, respectively. During CV in noncomplicant lung, Ppaw recorded from Cup and Cdown was significantly greater than Ppaw-Pla difference from Cup at PEEP ≥ 10 torr was significantly less in noncompliant lung than in compliant lung. It is concluded that SV affords complete protection of the Ppaw-to-Pla relationship even at high PEEP, whereas a decrease in pulmonary compliance affords moderate protection of the Ppaw-to-Pla relationship. These results are important because they further our ability to assess the accuracy of Ppaw during varying clinical situations.