Age-specific incidence curves for clinical cancer of the cervix in England and Wales show progressive changes over the period 1963-1978; in particular, a large reduction in incidence is seen in the age group 35-54. Since screening on any scale began in the early 1960s, we have investigated how much of this reduction in incidence in the middle age range can be attributed to detection of pre-invasive disease. Data on registrations of in-situ cancer have been used to estimate the patterns that might have been observed in the absence of screening. The results indicate clear cohort effects on incidence, with rising rates in the generations born 1906-1921 and since 1931, with a fall in the decade between. In addition to this, screening has probably led to a substantial reduction in the number of cases of clinical cancer in women aged 35-54, but has had little effect over the age of 60 where virtually no screening has been performed. Below age 35 the observed increase in incidence may be considerably less than it would have been in the absence of screening.