Individual genetic admixture estimates (IA) from European Americans (EAs) were computed in 7,996 members of the Gila River Indian Community (Arizona). Parental populations for the analysis were European Americans and full-heritage Pima Indians. A logistic regression was performed on 7,796 persons, to assess association of IA with type 2 diabetes. The odds ratio, comparing diabetes risk in full-heritage EAs with full-heritage Pima Indians, was 0.329 (95% confidence interval [CI] 0.225-0.482). Proportional-hazards analysis was performed on 5,482 persons who were nondiabetic at their first examination and 1,215 subjects who developed diabetes during the study. The hazard risk ratio for IA was 0.455 (95% CI 0.301-0.688). Nondiabetic persons had significantly more European IA. In nondiabetic Pimans, multivariate linear regressions of quantitative predictors of type 2 diabetes mellitus, including fasting plasma glucose, 2-h post-load plasma glucose, and body-mass index, showed significant inverse relations with IA when controlled for sex and age. These results illustrate the ongoing evolution of populations by the mechanism of gene flow and its effect on disease risk in the groups with admixture. When the two parental populations differ in disease prevalence, higher or lower risk is associated with admixture, depending on the origin of the admixed alleles and the relative magnitude of the disease prevalence in the parental populations. These data also illustrate the strong genetic components in type 2 diabetes and are consistent with one susceptibility locus common to obesity and diabetes.