Clinical, physiological, roentgenographic, and histological data concerning 502 patients who had open biopsy for chronic "interstitial" lung disease were reviewed. Mortality was 0.3%, the rate of complications was 2.5%, and the diagnostic yield was 92.2%. A modified Chamberlain approach in the second interspace is preferred for easy access to all lobes and mediastinum. Brief tube drainage is mandatory. Atelectasis and hemorrhage in the specimen are prevented by avoiding palpation and clamps, by delineating the wedge during full inflation, and by instant fixation. Customary biopsies of the tip of the lingula or middle lobe are avoided because these are common sites of inflammation, scarring, and passive congestion. Often, the most abnormal regions are biopsied apparently to aid the pathologist. Such selection has been the most important cause of meaningless histological findings and poor pathological, physiological, and roentgenographic correlations because these regions usually show end-stage disease of unrecognizable origin. Average lung is more likely to show an active and recognizable process.