We posit that chance plays a major role in the occurrence of many common malformations that cluster in families but recur less frequently than expected for simple Mendelian traits. Once the role of random effects is accepted, the segregation of such malformations may be explained on the basis of Mendelian transmission of a single abnormal gene that predisposes to, but does not always result in, the abnormal phenotype. We apply a stochastic (probabilistic) single-gene model to the occurrence of malformations in mouse and man. The stochastic single-gene model suggests the feasibility of isolating individual genes that determine morphogenesis and sets limits on the precision with which the recurrence of malformations can be predicted.