The usefulness of an endocervical brush for cytologic sampling was studied in 288 consecutive women attending a colposcopy clinic. One hundred sixty had initial colposcopic examinations with directed biopsies and endocervical curettage (ECC). One hundred twenty-eight had follow-up examinations that included Papanicolaou smears and ECC some time after evaluation and/or treatment for cervical dysplasia. Within an established protocol for the evaluation and management of abnormal Papanicolaou smears, the endocervical brush was compared to ECC. When the combination of colposcopy and cervical conization showed that dysplasia was confined to the endocervix, the endocervical brush was significantly more sensitive than ECC in detecting this endocervical disease (P less than .05). In follow-up evaluations and in patients with unsatisfactory colposcopic examinations, endocervical brush cytology could replace ECC without affecting clinical management. In patients with a satisfactory colposcopic examination, ECC may be required only when endocervical brush cytology is abnormal. Endocervical brush cytology is less costly and painful and could prove to be a reasonable alternative to ECC for the initial evaluation of the endocervix. Endocervical brush cytology shows promise as a safe adjunct to the colposcopic evaluation of pregnant women, whereas ECC is contraindicated in such women.