A model of increased microvascular permeability pulmonary edema was developed in chronically instrumented unanesthetized sheep using perilla ketone (PK). PK did not cause changes in pulmonary hemodynamics but did cause marked increases in the flow of protein-rich lung lymph. The changes in lung lymph flow were accompanied by radiographic evidence of both interstitial and alveolar pulmonary edema as well as hypoxemia. PK did not cause acute changes in lung mechanics. Dynamic compliance of the lungs and FRC decreased later, concomitant with the changes in lung lymph flow, radiographic evidence for pulmonary edema, and hypoxemia. Resistance to air flow across the lungs and specific conductance did not change significantly after PK infusion. The severity of the radiographic evidence for pulmonary edema observed after PK correlated with the severity of the concomitant hypoxemia and changes in dynamic compliance of the lungs. PK did not cause increases in the concentrations of cyclooxygenase products of arachidonic acid in lung lymph or plasma or changes in blood leukocyte counts. We conclude that PK causes increased lung microvascular permeability pulmonary edema without acute changes in pulmonary hemodynamics. This model permits study of the pathophysiologic aspects of increased lung microvascular permeability without the concomitant functional alterations that complicate most other experimental models of diffuse lung injury.