Minor cytological abnormalities--low-grade squamous intraepithelial lesions (LGSIL) and atypical squamous cells of undetermined significance (ASCUS)--are the most frequent abnormalities observed in smears performed in screening for cervical cancer. Surveillance through repeated smears has been proposed as an alternative to immediate colposcopy to reduce cost and avoid overloading colposcopy clinics. The aim of this review is to discuss the advantages and inconveniences of cytological surveillance of women with ASCUS or LGSIL, as reported in the literature. Cervical smears are not sufficient to establish the precise diagnosis of a lesion. In fact, the mean prevalence of normal cervices, high-grade CINs and cancers is estimated to be 58%, 7% and 0.5% for ASCUS and 34%, 27% and 0.2% for LGSIL. Cytological surveillance shows that minor cytological abnormalities regress in 28-69% of cases. In 6-9%, cytological regression masks a high-grade CIN, and a cancer may develop sooner or later in 0.30-1.12% of cases. Cancers often occur in patients temporarily lost to surveillance or in those with normal smears alternating with smears showing minor abnormalities. Strict modalities of surveillance are therefore required. The surveillance has to be carried out with smears repeated every 4-6 months for 2 years. If all smears are normal the regular screening at 3-year intervals may be resumed. On the other hand, an immediate colposcopy is recommended if a new abnormality is observed in these consecutive smears, regardless of its severity and of the time elapsed since the initial abnormality was discovered. With these recommendations in mind, cytological surveillance may slightly reduce the number of colposcopies but without reducing significantly the cost. At present, no arguments based on cost or efficacy clearly favour cytological surveillance over immediate colposcopy.