Saline hysterosonography was attempted as a routine, first-line screening test of uterine structure in 500 consecutive, unselected, infertile women. The procedure was completed in 96.8% (484/500) women and the observations were interpretable in 483 of these women. Intrauterine pathology was suspected in 67/499 (13.4%) women on plain ultrasound scan and 58/484 (12%) women with saline hysterosonography. Ultrasound alone had a superior specificity (96.3%) to sensitivity (81.8%) and better negative (97.6%) than positive (73.8%) predictive value for the detection of any intrauterine abnormality, using saline hysterosonography as the reference procedure. Suspected pathology at saline hysterosonography led to hysteroscopy in 20 women, after a median of 5.7 months (range, 1-14). The overall concordance rate between the two procedures was 65% with lesions suspicious of intrauterine polyps not present at subsequent hysteroscopy on six occasions. Criteria were established to help identify women with potentially self-limiting lesions, in whom a re-scan should be considered before resorting to hysteroscopy. The procedure was well tolerated with no significant complications. Saline hysterosonography appeared to be an acceptable first-line screening procedure for uterine structure which enhanced the predictive power of ultrasound alone for uterine anomalies and provided additional information which was potentially of value when planning operative hysteroscopy.