Women who have not accepted an invitation by their own general practitioner to have a smear test, or for whom voluntary population screening projects are unavailable because of lack of public funds, may be at increased risk of malignant cervical changes. Such women may be offered self-testing by a cytopipette on the basis of the present analysis of abnormal cell counts and cell density on the cytological slides. As one link in a prophylactic study using the cytopipette, 107 women with abnormal cell findings were reinvited for simultaneous sampling by irrigation smear and by ordinary technique. The cytological findings indicate that the two methods are equal in recovery of abnormal epithelium, provided that the scanned quantity of cells is increased fourfold in the irrigation smear method, this quantity being distributed over two specimens each with double cell density and giving at least 1800 squamous epithelial cells per visual field (125:1) or a total of approximately 10(6) epithelial cells in each cytological specimen. Under these conditions the two sampling methods have the same predictive value.