Flexible fiberoptic bronchoscopic examination was performed 254 times in 228 patients in a prospective study to determine what specimens would give the greatest yield in the diagnosis of lung cancer. In addition, we questioned whether postbronchoscopic sputum specimens were still the most accurate method of diagnosing lung cancer, as they had been when only the rigid bronchoscope was available. Material for cytopathologic examination was obtained from bronchial brushings, bronchial biopsy, bronchial brushings in saline solution, bronchial washings, and three postbronchoscopie sputum specimens. In the latter part of the study, patients with peripheral tumors were studied with the aid of biplane fluoroscopic techniques. Bronchial brushings and bronchial biopsy each yielded the highest percentage of positive specimens (65 percent); postbronchoscopic sputum specimens were less frequently positive (40 percent). The combination of bronchial brushings and bronchial biopsies gave the optimum overall accuracy (79 percent). Bronchial washings and postbronchoscopic sputum specimens did not add significantly to diagnostic yield, and we conclude that they should no longer be part of the diagnostic procedures routinely ordered. In peripheral lesions, diagnostic accuracy was greatly enhanced in the cases where biplane fluoroscopic techniques were employed.