To evaluate a new liquid-based cytology technique, Papspin (Thermo Shandon, Pittsburgh, Pennsylvania, U.S.A.).Three thousand cervical samples were examined. Each cervix was sampled with a Cervex Brush (Roche, Oss, the Netherlands), used first for a Pap smear and afterwards for a Papspin. One cytospin was prepared from each vial.An identical rate of epithelial cell abnormalities (3.8%) was detected with the two methods. Diagnostic concordance was observed in 86% of the 114 cases. Differences in diagnoses occurred in 168 of 3,000 cases (5.6%) concerning fungal infection (22 cases), epithelial cell abnormalities (24 cases) and minimal differences within the nonneoplastic Bethesda category (122 cases). Endocervical cells were absent from 158 Papspins (5.3%) and 66 Pap smears (2.2%), while they were present in the respective Pap smear or Papspin. Seven Papspins were considered "satisfactory, but limited by ..." (SBLB) as compared to 33 Pap smears given the absence of endocervical cells.Discordances concerning epithelial cell abnormalities were observed in 24 of 3,000 cases (0.8%). Fungal infections were more easily diagnosed on Papspin. The absence of endocervical cells in 5.3% of Papspins is due to a bias of methodology. Quality improvement was evident on Papspin for SBLB specimens. HPV testing could be performed with good results.