The Saccomanno technique of sputum preparation is widely used. This study evaluates the role of this preparation in conjunction with fresh smears in the diagnosis of lung cancer. All sputum samples from September 1973 to July 1975 showing atypia were randomized and negative controls added. The Saccomanno and fresh smears were evaluated independently and blindly and classified as negative, atypical, suspicious, or cancer. When each preparation was compared with the original diagnosis, the diagnostic accuracy for 55 squamous carcinomas was similar (fresh 95%, Saccomanno 86%) but significantly less in the Saccomanno preparations of 22 small cell carcinomas (fresh 100%, Saccomanno 24%) and 26 adenocarcinomas (fresh 96%, Saccomanno 52%). Four cases negative on fresh smears were suspicious or diagnostic of cancer on the Saccomanno slides. There were no cases of small cell carcinoma in which the Saccomanno preparations added information not available on the fresh smears. The authors conclude that in conjunction with fresh smears, the Saccomanno preparation may contribute to the diagnosis of non-small cell carcinomas but does not appear to aid in the diagnosis of small cell carcinoma.