We evaluated the feasibility, reliability, and acceptability of prenatal diagnosis of haemophilia A by DNA analysis of chorionic villi. Twenty-two women at risk to transmit the abnormal gene were referred for prenatal diagnosis, two of them twice. Two of the 22 women appeared to be non-carriers by DNA analysis. In one of these women, the results were known only after chorionic villus sampling had been carried out. Thirteen of the twenty carriers were heterozygous for an intragenic (Bc1I or XbaI) marker; six women were only heterozygous for the extragenic DXS52 (St14) locus. One of the women was homozygous for all the presently known DNA markers within or closely linked with the factor VIII locus. Twelve of the 22 fetuses at risk were male, ten were female. Seven of the 12 male fetuses were shown to be affected and were subsequently aborted. Four male fetuses appeared to be not affected. In one case, the diagnosis was made by use of an extragenic marker. The woman rejected fetal blood sampling to confirm the diagnosis. After birth, a normal factor VIII level was found in three of the four cases. The fourth pregnancy is still continuing. In one of the 12 male fetuses, no diagnosis at the gene level was possible. DNA analysis is expected to provide maximum certainty as to the phenotype of the fetus for approximately 60 per cent of the women; for another 37 per cent a rate of misdiagnosis of 4-5 per cent applies. In only 3 per cent of the cases will no diagnosis at the gene level be possible as yet. The new possibility of a prenatal diagnosis in the first trimester of pregnancy enabled some of these women to have a family of their own and was appreciated in particular by the women who underwent fetoscopy in an earlier pregnancy.