The diagnostic accuracy and complication rate of transbronchial forceps lung biopsy combined with bronchial brushing were analyzed in 164 consecutive patients, of whom 95 had nonmalignant disease and 51 had malignant disease. Eighteen patients with abnormal chest radiographs had normal transbronchial forceps lung biopsies and no established diagnosis. Satisfactory specimens were obtained in 153 patients (93%), with diagnostic accuracies of 62, 64 and 67%, respectively, in infectious, interstitial and malignant lung disease. The over-all diagnostic accuracy was 57%. Seven patients were diagnosed as having metastatic carinoma to the lung by transbronchial forceps lung biopsy but brush biopsy was positive in only 3. Of the 23 patients with primary lung carcinoma diagnosed by transbronchial forceps lung biopsy, 21 had positive brush biopsies. An additional 3 patients with bronchogenic carcinoma were diagnosed only by brush biopsy. Fluoroscopy was essential for accurate positioning of the biopsy instrument. Increased risk factors were present in 83 patients (51%), among whom the immunosuppressed group presented special problems. Significant bleeding in 15 patients (9%) was controlled by conservative management. Pneumothorax occurred in 7 patients (4%). There were no deaths. Transbronchial forceps lung biopsy in experienced hands is safe and well tolerated. The physician, nevertheless, must be prepared to handle major complications, especially bleeding. The procedure has a reasonable diagnostic yield in high-risk patients who have a variety of lung lesions.