In 1975 Duke University Medical Center, a retrospective and prospective survey of respiratory cytopathologic specimens was undertaken for the ten-year period 1970 to 1979. The purpose of this study was to document the role of cytopathology in the diagnosis of lung cancer at this institution. This paper presents the results of the cytopathologic and histopathologic typing of cases of lung cancer seen at Duke University Medical Center from 1970 to 1974. During this period, 9,892 cytologic specimens from the lower respiratory tract were processed. Cytopathologic diagnoses of cancer with tissue confirmation were made on 483 specimens from 232 patients. Because original cytologic diagnoses, but not histopathologic diagnoses, had been made in conformity with a modified WHO classification of lung neoplasms, all histopathologic material was reviewed and reclassified when necessary. This was carried out by one of the authors (E.H.B.) as a blind review without benefit of knowledge of either preexisting cytopathologic or histopathologic diagnoses. Twenty-six patients were excluded from the current study because of lack of satisfactory histologic material. In 94 patients classified by histopathology as having squamous cell carcinoma, 76.4% of the positive cytologic specimens were also called squamous cell carcinoma; 18.6% were interpreted as large cell undifferentiated carcinoma. In 39 patients classified by tissue as having large cell undifferentiated carcinoma, the cytology agreed in 42.4% of the positive specimens. For the 29 patients thought histologically to have small cell undifferentiated carcinoma, the same diagnosis was rendered in 95.5% of the cytologically positive specimens from these patients. For the adenocarcinoma group of 43 patients, a cytopathologic diagnosis of adenocarcinoma was made in 67.8% of the positive specimens.