Women using the medical clinic of a public hospital were interviewed about their Pap smear histories to assess the accuracy of self-reported smears and to identify groups in need of further screening. Interview data from 263 women were compared with cytology files and hospital records. In spite of considerable agreement between patient report and record, patients reported significantly more recent smears than were documented. Accuracy of recall was not dependent on age or birthplace, but on the length of time since the last smear. About half the women had been screened within the past two years, whereas one tenth had never been screened. Women aged 65 years or older had fewer recent smears than younger women, while foreign born women were more likely never to have had a Pap smear than were United States born women. We conclude that self-reported information is useful in assessing Pap smear histories, but the screening rates that result should be treated as high estimates. Significant opportunities for screening exist in ambulatory care sites in low-income communities.