A high frequency of mutations in the methyl CpG-binding protein 2 (MECP2) gene has recently been reported in males with nonspecific X-linked mental retardation. The results of this previous study suggested that the frequency of MECP2 mutations in the mentally retarded population was comparable to that of CGG expansions in FMR1. In view of these data, we performed MECP2 mutation analysis in a cohort of 475 mentally retarded males who were negative for FMR1 CGG repeat expansion. Five novel changes, detected in seven patients, were predicted to change the MECP2 coding sequence. Except for one, these changes were not found in a control population. While this result appeared to suggest a high mutation rate, this conclusion was not supported by segregation studies. Indeed, three of the five changes could be traced in unaffected male family members. For another change, segregation analysis in the family was not possible. Only one mutation, a frameshift created by a deletion of two bases, was found to be de novo. This study clearly shows the importance of segregation analysis for low frequency mutations, in order to distinguish them from rare polymorphisms. The true frequency of MECP2 mutations in the mentally retarded has probably been overestimated. Based on our data, the frequency of MECP2 mutations in mentally retarded males is 0.2% (1/475).