The Cell-Sweep, a new cervical cytology sampling device, was prospectively evaluated. From September 1, 1992, to October 31, 1992, 548 cervical cytology smears were prepared using the Cell-Sweep. They were compared to 524 cervical smears performed between January 1, 1993, and February 28, 1993, using the Cytobrush and Ayre spatula. One smear in the Cell-Sweep group and nine in the Cytobrush/Ayre spatula group were unsatisfactory for evaluation secondary to drying artifact (P < .01). Seven (1.2%) of the Cell-Sweep smears and 13 (2.4%) of the Cytobrush/Ayre spatula smears were inadequate due to obscuring of the epithelial cells by blood (P = .11). Of the smears obtained with the Cell-Sweep, 66 (12%) contained no endocervical component versus 42 (8%) of the smears obtained with the Cytobrush/Ayre spatula (P = .03). In the Cell-Sweep group there were 130 (23.7%) abnormal smears versus 110 (20.9%) in the Cytobrush/Ayre spatula group (P = .3). However, high grade squamous intraepithelial lesions (SILs) were identified in 10 (1.8%) smears obtained with the Cell-Sweep and in 2 (0.4%) obtained with the Cytobrush/Ayre spatula (P = .02). The Cell-Sweep produced a small proportion of unsatisfactory smears due to drying artifact or blood. Even though the proportion of smears lacking endocervical cells was statistically greater with the Cell-Sweep than with the Cytobrush/Ayre spatula, there was no difference in the rate of detection of atypical squamous cells and low grade SIL. The Cell-Sweep appears to detect an increased number of high grade SILs. This device deserves further study with histopathologic correlation.