The relation between Papanicolau ("Pap") smear and the risk of cervical neoplasia was evaluated in a case-control study in which 145 women with cervical intra-epithelial neoplasia (CIN) were compared with 145 age-matched outpatient controls, and 191 cases of invasive cervical cancer were compared with 191 control subjects in hospital for acute conditions unrelated to any of the established or suspected risk factors for cervical cancer. Compared with women with no previous screening smear, the relative risk (RR) estimate for invasive cancer was 0.44 (with 95% confidence interval [CI] = 0.24-0.80) for those who had had one smear and 0.20 (95% CI = 0.13-0.32) for those who had had two or more smears. The corresponding estimates for CIN were 0.27 (95% CI = 0.10-0.71) and 0.12 (95% CI = 0.06-0.25), respectively. The risk estimates for intervals of more than 5, 3-5, and less than 3 years were 0.36, 0.18, and 0.10 for invasive cancer, and 0.45, 0.31, and 0.07 for CIN. These results were not materially modified by adjustment for the major risk factors for cervical cancer, such as indicators of socioeconomic status and sexual habits. According to the estimates of this study, 64% of invasive cervical cancers could be prevented by screening at intervals of more than 5 years, an additional 18% by reducing the interval to 3-5 years, and a further 8% to less than 3 years.