Prenatal diagnosis was carried out in 138 pregnancies at 1-in-4 risk for cystic fibrosis (CF) by using closely linked DNA markers, including XV-2c and KM-19. In fully informative families, 25 of 123 (20%) fetuses were predicted to be affected; 16 of these 25 pregnancies were terminated and 9 were continued. Postnatal sweat tests are completed in 42 cases; the diagnoses were confirmed in 4 of 4 infants predicted to be affected and in 37 of 38 infants predicted to be unaffected. One infant predicted to be a carrier had an abnormal sweat test after birth, but the mother also had an abnormal sweat test, and there was no evidence of an error in linkage analysis. The data indicate that prenatal diagnosis using linkage analysis is fully informative in most families and is highly reliable with either chorionic villus sampling or amniocentesis. Although outcome data are available on only 42 pregnancies, based on our experience, on general principles of linkage analysis, and on the tight linkage of the known DNA markers with CF, we recommend that DNA analysis replace microvillar intestinal enzyme analysis for 1-in-4 risk pregnancies when DNA is available from the propositus.