We conducted a large population-based survey of fragile X (FRAXA) syndrome in ethnically diverse metropolitan Atlanta. The eligible study population consisted of public school children, aged 7-10 years, in special education-needs (SEN) classes. The purpose of the study was to estimate the prevalence among whites and, for the first time, African Americans, among a non-clinically referred population. At present, 5 males with FRAXA syndrome (4 whites and 1 African American), among 1,979 tested males, and no females, among 872 tested females, were identified. All males with FRAXA syndrome were mentally retarded and had been diagnosed previously. The prevalence for FRAXA syndrome was estimated to be 1/3,460 (confidence interval [CI] 1/7,143-1/1,742) for the general white male population and 1/4, 048 (CI 1/16,260-1/1,244) for the general African American male population. We also compared the frequency of intermediate and premutation FRAXA alleles (41-199 repeats) and fragile XE syndrome alleles (31-199 repeats) in the SEN population with that in a control population, to determine if there was a possible phenotype consequence of such high-repeat alleles, as has been reported previously. No difference was observed between our case and control populations, and no difference was observed between populations when the probands were grouped by a rough estimate of IQ based on class placement. These results suggest that there is no phenotype consequence of larger alleles that would cause carriers to be placed in an SEN class.