Although the risk of Down's syndrome increases with maternal age, women under 35 bear about 80 percent of the infants born with this condition. We prospectively investigated the utility of measuring maternal serum alpha-fetoprotein during the second trimester in women under 35 in order to identify pregnancies in which the fetus was affected with Down's syndrome. Over a two-year period, 34,354 women in this age group were screened. Amniocentesis was offered when the risk of Down's syndrome, calculated as a function of maternal age and maternal serum alpha-fetoprotein concentration adjusted for maternal weight and race, was 1:270 or higher, the risk for a 35-year-old woman. This threshold was exceeded in 1451 women in whom gestational age was confirmed by ultrasound; 9 women in this group had a fetus with the syndrome. In three women whose fetuses had trisomy 18 and one whose fetus had trisomy 13, the calculated risk of Down's syndrome was 1:270 or higher. Thus, among women in whom the risk exceeded our cutoff point, 1 in 161 were found to have a pregnancy in which the fetus was affected with Down's syndrome; the figure was 1 in 112 for all autosomal trisomies. Eighteen pregnancies involving Down's syndrome, three involving trisomy 18, and two involving trisomy 13 were not associated with a calculated risk above the cutoff point. The available data indicate that in our population, using a cutoff for risk at which 5 percent of women under 35 are offered amniocentesis, we will detect one quarter to one third of pregnancies in which the fetus has Down's syndrome.