Although adult respiratory distress syndrome (ARDS) has been a recognized entity for over 20 years, estimates of its incidence have been very controversial. The most quoted figure is from a 1972 National Heart and Lung Institute Task Force, which estimated 150,000 cases/year in the United States, an incidence of about 75 cases/100,000 population. No experimental study, however, has adequately addressed this issue. We were in a unique position to answer this controversy because the hospital of one of the investigators is located on an island (Las Palmas, Canary Islands, population 700,000). All patients who required mechanical ventilation (other than during anesthesia and immediate postoperative care) were admitted to the same adult intensive care unit. A prospective study to determine the incidence of ARDS was carried out over a 3-yr period. Inclusion criteria were: (1) a known predisposing illness for ARDS; (2) PaO2 less than or equal to 55 mm Hg with FIO2 greater than 0.5 with 5 cm H2O PEEP, without improvement in 24 h; (3) bilateral pulmonary infiltrates; (4) no evidence of left ventricular failure. An average of ten patients per year, representing an incidence of 1.5 cases/100,000 population, were diagnosed as having ARDS and the mortality rate was 70%. Using a more liberal clinical criterion of PaO2 less than or equal to 75 mm Hg with FIO2 greater than or equal to 0.5, 44 more patients with ARDS, representing a total incidence of 3.5 cases/100,000 population, were identified. In conclusion, the overall incidence of ARDS was 1.5 to 3.5 cases/100,000 population, an incidence that is much lower than most previously published estimates.