A paternal-age effect and the exclusive paternal origin of mutations have been reported in Apert syndrome (AS). As the incidence of sporadic AS births increases exponentially with paternal age, we hypothesized that the frequency of AS mutations in sperm would also increase. To determine the frequency of two common FGFR2 mutations in AS, we developed allele-specific peptide nucleic acid-PCR assays. Analyzing sperm DNA from 148 men, age 21-80 years, we showed that the number of sperm with mutations increased in the oldest age groups among men who did not have a child with AS. These older men were also more likely to have both mutations in their sperm. However, this age-related increase in mutation frequency was not sufficient to explain the AS-birth frequency. In contrast, the mutation frequency observed in men who were younger and had children with AS was significantly greater. In addition, our data suggest selection for sperm with specific mutations. Therefore, contributing factors to the paternal-age effect may include selection and a higher number of mutant sperm in a subset of men ascertained because they had a child with AS. No age-related increase in the frequency of these mutations was observed in leukocytes. Selection and/or quality-control mechanisms, including DNA repair and apoptosis, may contribute to the cell-type differences in mutation frequency.