During a 7-year period (1988-94), we diagnosed 102 fetuses with trisomy 18, and 54 with trisomy 13; in 6.9% of the trisomy 18 and in 40.7% of the trisomy 13 fetuses, there was a facial cleft. On the basis of (1) these frequencies of facial cleft in trisomic fetuses; (2) the reported prevalence of facial cleft in mid-trimester fetuses; and (3) estimates of the prevalence of trisomies 18 and 13 at 20 weeks of gestation in a population with the maternal age distribution of all deliveries in England and Wales, it was calculated that 6.5% of fetuses with a facial cleft would have trisomy 18 or 13. This estimated frequency of trisomies was significantly lower than the 26% observed in 111 fetuses with a facial cleft that were referred to our unit for fetal karyotyping. These findings suggest that the patients with a facial cleft examined in a referral center are preselected in favor of those with multiple abnormalities, and therefore a higher frequency of associated chromosomal defects. In the future, with improving quality of ultrasound equipment and standards of scanning, it is likely that more cases of isolated facial cleft will be identified and, consequently, the observed frequency of chromosomal defects should decrease.