During the past three years, we have conducted fragile X DNA studies for carrier screening and prenatal diagnosis using a previously described PCR protocol that accurately resolves normal FMR1 alleles and premutations and detects most full mutations [Brown et al., JAMA 270:1569-1575, 1996]. A total of 344 pregnant women with a family history of mental retardation of unknown cause were screened and 6 fragile X carriers were identified: two had full mutations, and four had premutations. The mentally retarded relatives of two other women were found to be fragile X positive although the women themselves were not carriers. In all, 6 carriers and 8 fragile X families were identified by this screening. We have also screened 40 pregnant women who were members of previously identified fragile X families, but whose carrier status was unknown. Ten were found to be carriers and were offered prenatal diagnosis. Prospective prenatal testing of 84 carrier women correctly detected 31 fetal samples (19 females, 12 males) with full mutations and 6 with premutations (2 females, 4 males). No false positives but one false negative occurred early on due to undetected maternal cell contamination. In addition, screening of 806 males with developmental delays of unknown cause gave positive results in 33 (4.1%). Potential problems and pitfalls of direct DNA testing are discussed. Because of the proven success of fragile X screening with direct molecular analysis, screening of all undiagnosed individuals with mental retardation and at risk pregnant women should now be considered. The identification of fragile X carriers and prenatal diagnosis of their pregnancies should significantly reduce the prevalence of this syndrome.