This paper reviews the characteristics and the results of 15 years of experience with a preventive program, based on carrier screening and prenatal diagnosis, designed to control thalassemia major in the Sardinian population. The education of the population about thalassemia and the modalities for its prevention was accomplished via the mass media. Carrier screening was carried out voluntarily on couples of child-bearing age. Prenatal diagnosis was initially carried out by fetal blood analysis; since 1983, it has been done by DNA analysis on non-amplified or amplified DNA. Different chorionic villous sampling procedures have been used. Nowadays, we have adopted the transabdominal approach because, in our experience, it seems to be associated with a low risk (2%) of fetal mortality. At the present time, the beta-thalassemia mutations are detected directly by dot-blot analysis of amplified DNA with 32P- or horseradish peroxidase-labeled allele-specific oligonucleotide probes. Two oligonucleotide probes, one complementary to the codon-39 nonsense mutation, which accounts for 95.7% of the beta-thalassemia chromosomes in the Sardinian population, and the other complementary to the frameshift at codon 6, which is the second most common mutation in our population (2.1%), allow us to make prenatal diagnosis in the large majority of cases. Notwithstanding a careful dissection of maternal decidua from chorionic villi, co-amplification of maternal sequence was detected in 4 out of 425 cases tested by this procedure. In order to avoid this pitfall, the simultaneous amplification of highly polymorphic VNTR (variable number of tandem repeats) segments could be used. On the whole we have so far carried out 2711 prenatal tests: 1130 by fetal blood analysis, 1156 by oligonucleotide hybridization on electrophoretically separated DNA fragments, and 425 by dot-blot analysis on amplified DNA with allele-specific oligonucleotide probes. Two errors occurred by fetal blood analysis and none by DNA analysis. The incidence of thalassemia major declined from 1:250 live births in the absence of prevention to 1:1000 after the establishment of this program, indicating that carrier screening and prenatal diagnosis are effective means for preventing thalassemia major at the population level.