The uptake of breast cancer screening in women aged 45-64 years, living in Aylesbury Vale district, was determined between 1984 and 1988. During this period, self-referral gave way to invitation as the method of gaining access to the service. In the initial period of open access the estimated uptake was 27.9 per cent. A pilot study of response to invitation (and reminder) in two general practices resulted in a compliance of 80.6 per cent. Extended to 20 general practices in Aylesbury Vale, overall compliance with invitation was 73.7 per cent. The response amongst previously screened volunteers was higher (91.5 per cent) than amongst unscreened women (67.6 per cent). There was no significant trend in compliance with age although the oldest women had the lowest uptake. Second round compliance in the two pilot practices reduced to 77 per cent, with 86.1 per cent of a cohort of screened women responding to a second offer of screening. The implications of these results for the national breast cancer screening programme are discussed together with the difficulties associated with the routine derivation of uptake. It is concluded that not only must the accuracy of population registers be improved, but also standard definitions of uptake measurements should be agreed and adopted.