Objective: To find an appropriate sampling device for a liquid-based procedure in the population screening for cervical cancer, focusing on bleeding at sampling and the amount of cells smeared. Methods and Materials: 1,000 consecutive women who underwent primary screening were studied. The specimens were obtained with the cotton stick/Cytobrush® method in the first 500 cases or with the Cervex-Brush® in the following 500 subjects, and were processed using the Thinlayer Advanced Cytology Assay System (TACAS™) following the manufacturer’s instructions. Results: (1) Bleeding at cellular sampling using the cotton stick/Cytobrush and Cervex-Brush methods occurred in 1.2 and 8.8% of the cases, respectively (p < 0.0001). (2) The incidences of cells obtained with the two methods which covered the whole area, <1/2 and ≥1/4, and <1/4 of the observation fields were 55.4 versus 62.2% (p < 0.05), 14.6 versus 9.4% (p < 0.05), and 2.0 versus 4.0% (p < 0.05), respectively. (3) The incidences of endocervical or metaplastic cells obtained with ≥500 and <10 were 34.6 versus 20.0% (p < 0.01) and 9.4 versus 18.4% (p < 0.01), respectively. In cases of cells covering <1/4, incidences with <10 were 0 and 0.6% (n = 3), respectively. (4) Detection rates of abnormal cytology were 3.4 and 5.2% (n.s.), including atypical squamous cells of undetermined significance in 2.4 and 3.2%. Conclusions: The cotton stick/Cytobrush is superior to the Cervex-Brush as a cellular sampling device for the TACAS liquid-based procedure.