This article addresses the classic enigma about schizophrenia (SZ). The disease occurs with a lifetime prevalence of 1%, 80% of which is attributable to genetic factors. Females with SZ produce 50% as many children as normals, and males with SZ produce 25%. Genetic factors responsible for SZ should behave like lethal genes. Yet the prevalence of SZ remains around 1% throughout the world. How can that be? Additionally, CATIE concluded that the response of each individual with SZ to treatment with antipsychotic agents (effectiveness, side-effect profile, or long-term prognosis) cannot be predicted. Every case seems to be unique. Several recent publications have reported increased frequencies of single-nucleotide polymorphisms (SNPs) and of copy-number variants (CNVs) containing large regions of DNA in patients with SZ. These genetic perturbations often include neurodevelopmental genes. The overwhelming majority of SNPs and CNVs are post-fertilization mutations, occurring in somatic tissue, not germinal tissue. These mutations are a normal aspect of somatic cell division but occur more frequently in patients with SZ. Somatic mutations are not passed on to subsequent generations and therefore cannot account for the inheritance of SZ. Our speculation is that the genetic platform for SZ is the gene or genes that increase the number of de novo mutations in patients with SZ. We argue that balanced polymorphism is the most plausible hypothesis to account for the preservation of non-adaptive genes in nature-and, in particular, in SZ. Maladaptive genes in different combinations can confer increased fitness to the entire population, thus insuring their preservation in the gene pool. Somatic mutations explain both the sporadic occurrence of SZ within families and the wide variations in phenotypic expression of SZ. Increased frequency of somatic mutations may confirm greater overall fitness via balanced polymorphism to explain the maintenance of the SZ gene or genes within the human population.