Objectives: To ascertain the reasons for a low rate of response for breast screening., Design: All relevant aspects of the organisational process examined, including general practitioners' notes. Non-responders visited and interviewed., Setting: An inner city breast screening service working on the model advocated by the Forrest report., Subjects: 288 Women aged 50-64 registered with several general practices and invited for screening by post., Main Outcome: Determination of factors important for success of breast screening programmes., Results: After five women were excluded by their general practitioners the response rate was 129 out of 283 (46%), but 99 (35%) of the women did not receive their invitations because of inaccuracies in the family practitioner committee's database and general practitioners failing to check women's addresses completely., Conclusions: Increased rates of response will depend on enabling general practitioners to check addresses and on an increased awareness of the importance of information.