We compared satisfactory for evaluation but limited by (limited by) and unsatisfactory gynecologic cytologic diagnoses for samples collected by conventional smearing with those generated with the AutoCyte Prep in a population with a historic squamous intraepithelial lesion (SIL) rate of less than 1%. Results from 18,819 AutoCyte Preps were compared with 53,835 conventional cervical smears. Furthermore, 23 women ages 18 to 65 years undergoing annual Papanicolaou tests underwent sequential sampling with the AutoCyte Prep and the Surgipath C-E brush. Comparison of the AutoCyte Prep with conventional cytologic diagnoses revealed the following: unsatisfactory rate, down 97%; limited by rate, down 67%; low-grade SIL rate, up 86%; cervical cancer rate, up 300%; and high-grade SIL rate, unchanged. Examination of unsatisfactory and limited by cases for the AutoCyte Prep showed that 88% were due to absence of endocervical cells (ECs). Dual sampling showed no improvement in EC recovery over the AutoCyte collection device. Compared with conventional Papanicolaou smears, the AutoCyte Prep significantly decreased the rate of unsatisfactory and limited by specimens while increasing low-grade SIL and cancer detection and EC recovery. The majority of limited by specimens with the AutoCyte Prep were due to absence of ECs, but use of a brush-type device for better endocervical sampling did not enhance EC recovery.