AIMS--To test the hypothesis that absence of squamous cells in cervical smears obtained by an endocervical sampling technique is more prominent in patients with a cervical ectropion. METHODS--Prospective study exploring the relation between the composition of cervical smears obtained using an endocervical cotton swab in patients with (n = 188) and without (n = 341) a cervical ectropion. Subjects were 529 consecutive patients from whom a cervical smear was prepared at a university gynaecological clinic. RESULTS--In 7% of the endocervical samples no squamous cells were found. There was no correlation, however, between the presence or the size of an ectropion and the absence of squamous cells in those samples. CONCLUSIONS--It was confirmed that endocervical sampling alone is insufficient to obtain good quality cervical smears. The presence of an ectropion proved to be an unreliable predictor of the absence of squamous cells.