A model to explain the mutational process and population dynamics of the fragile X syndrome is presented. The mutational mechanism was assumed to be a multipathway, multistep process. Expansion of CGG repeats was based on an underlying biological process and was assumed to occur at two time points: meiosis and early embryonic development (mitosis). Meiotic expansion was assumed to occur equally in oogenesis and spermatogenesis, while mitotic expansion was restricted to somatic, or constitutional, alleles of maternal origin. Testable hypotheses were predicted by this meiotic/mitotic model. First, parental origin of mutation is predicted to be associated with the risk of a woman to have a full mutation child. Second, "contractions" seen in premutation male transmissions are predicted not to be true contractions in repeat size, but a consequence of the lack of mitotic expansion in paternally derived alleles. Third, a portion of full-mutation males should have full-mutation alleles in their sperm, due to the lack of complete selection against the full-mutation female. Fourth, a specific premutation-allele frequency distribution is predicted and differs from that based on models assuming only meiotic expansion. Last, it is predicted that approximately 65 generations are required to achieve equilibrium, but this depends greatly on the expansion probabilities.