Cervical cancer incidence and mortality can be reduced by removal of precursor lesions detected at cytological screening. Organised screening, i.e. regular invitation of defined target groups, is generally considered more effective than opportunistic screening. The latter method however, is predominant in most settings. There is no scientific basis for advocating one type of screening or the other. Our aim was to compare the two types and to analyse their efficiency. We analysed 466,275 smears taken in an open cohort of 118,890 women during 1969-88. A computerised database permitted standardised classification of all smears and complete ascertainment of cancer in situ through record linkage. The number of in situ cancers detected per 1000 smears, the detection ratio, was used as an outcome measure both in univariate analyses and in multivariate logistic regression models. Cancer in situ was detected in 1076 women in the study cohort, with a detection ratio of 3.0 at organised and 2.1 at opportunistic screening, yielding an unadjusted odds ratio of 0.69 (95% CI 0.61-0.79). After adjustment for age and time period, the probability of detecting cancer in situ was around 25% higher with opportunistic than with organised screening (OR = 1.26; 95% CI 1.09-1.46). This difference in favour of opportunistic screening was most pronounced in the first 10 year period and disappeared during the last decade. The difference in efficiency between organised and opportunistic screening in the detection of cancer in situ was slight, if any. The dogma that organised screening is significantly more efficient than the opportunistic type needs reconsideration.