The objectives of this study were to examine aspects of organization of a proposed national screening programme based in general practice. The target population of women aged 25-59 years and their general practitioners (GPs), in a defined inner city area, was identified from a population register of persons eligible for free medical services; a computerized system was developed for invitations and record linkage of cytology results. Smears were examined in one laboratory and follow up of women with abnormal smears was undertaken by one gynaecologist. A random sample of non-responders was surveyed by questionnaire. Response following two invitations was only 20%. Practices with male doctors only had significantly lower response rates (P < 0.001) than those with a female doctor/nurse. A survey of non-responders showed that over 20% of addresses were incorrect and 16% of those interviewed were ineligible for smear tests. A preference for a female to undertake smears was expressed by 67%, and 77% believed that the purpose of the cervical smear was to detect cancer. An accurate population register, health promotion, support for GP practices, provision of alternative venues for smear tests, development of computer systems, accurate data entry and fail-safe follow up are aspects of a cervical screening service which must be addressed prior to setting up a national service.