A retrospective study analyzing the case notes of 49 hospitalized adults, either in intensive care (n = 26) or in thoracic medicine units (n = 23), for acute bacteriologically proven pneumococcal pneumonia based on samples obtained other than by sputum examination. The mortality was 54% in intensive care and 17% in the thoracic medicine unit. This significant difference may be explained in part by a respiratory distress syndrome in whom there were adequate criteria on admission for 7 patients in the intensive care group. Among these latter only one patient had had a splenectomy. The others did not have underlying disorders (three were chronic alcoholics); 7 patients were shocked on admission, four with a leukopenia less than 5,000/mm3 and six had a thrombocytopenia less than 100,000/mm3; finally 6 had a temperature of less than 38 degrees C. 7 patients died in less than four days (mean 2 days) in a clinical context of refractory hypoxemia. The significance of the respiratory distress syndrome is probably very different from the usual pneumonia; it seems rather to be an integration of the toxins induced by the pneumococcus. Its presentation can be particularly misleading as regards the diagnosis; the prescription of antibiotics once a diagnosis is obtained would seem insufficient by itself in this context to obtain a cure.