More than twenty years ago, Ashbaugh and coworkers,(1), in a now classic article, described. 12 patients whose striking but uniform clinical, physiologic, roentgenographic, and pathologic abnormalities distinguished them from among 272adult patients who had received respiratory support in the intensive care units of Colorado General Hospital and Denver General Hospital. The 12patients all had severedyspnea, tachypnea, cyanosis that was refractory to oxygen therapy, decreased respiratory system compliance, and diffuse alveolar infiltrations on their chest radiographs. Pathologic examination in seven patients who died revealed atelectasis, vascular congestion and hemorrhage, severe pulmonary edema and hyaline membranes. Shortly afterward, Petty and coworkers (2), called this constellation of findings the adult respiratory distress syndrome (ARDS). Since then, ARDS has been recognized as an entity throughout the world and has been the subject of numerous conferences, hundreds of articles, and several books. As a result of this activity, much descriptive information has been obtained about ARDS, and we have learned a lot about its clinical features, physiologic disturbances, prognosis, and pathologic findings. And yet formidable problems remain: there is disagreement about exactly what ARDS is and on what causes it; more importantly, available empiric treatment is inadequate, and mortality remains unacceptably high (600/0 or more)(3). This appears to be one of the few points of agreement among investigators, but even that statement is arguable (4). We believe that much of the controversy concerning ARDS is explained by the lack of a satisfactory definition of this elusivesyndrome. How can you collect, much less compare, epidemiologicdata and mortality figures when there is no uniformly accepted (and used) definition? How can you study basic pathophysiologic mechanisms, understand natural history, and above all, evaluate new therapeutic approaches in what appears now to be an amalgam of many different disorders? The purpose of this article, therefore, is to propose an expanded definition of ARDS that takes into account new knowledge about