The quality of the cervical smear, a decisive factor in the efficacy of population screening, can depend on the sampling method utilized. An analysis was made of the performance of five sample takers in a screening program, each of whom had made approximately 5,000 smears, and of the five sampling methods each had used: spatula alone (method A), Cytobrush plus spatula (method B), Cytopick (method C), cotton swab plus spatula (method D) and Cervex brush (method E). The differences between the sample takers and the sampling methods were significant in both the detection of grade III cervical intraepithelial neoplasia (CIN III) (P less than .01) and in the production of smears containing endocervical cells (EC+) (P less than .018). The data obtained firmly establish the importance of the presence of endocervical cells for smear adequacy. The results of this study indicate that (1) method B (Cytobrush plus spatula) and method C (Cytopick) give superior results in the preparation of EC+ smears and in the detection of CIN III and thus should be used in population screening programs, and (2) methods A and D should not be used for cervical cytologic sampling in such programs.