We describe the genetic structure and interrelationships of nine south Indian tribal groups (seven from Andhra Pradesh and two from the adjoining states of Tamil Nadu and Kerala) using seven polymorphic loci (ABO, MN, RH, PGM, ACP, PGD, and LDH). R matrix analysis indicates that the Andhra Pradesh tribes are clustered and that the Kadar and Irula are genetically isolated from them. This dispersion of populations has been explained by the combination of relatively high frequencies of the alleles RH D and MN M in the Kadar and the relatively high proportions of the allele PGM*2 in the Irula. The Mahaboobnagar Chenchu subgroup is isolated from other Telugu-speaking groups because of high frequencies of the PGM*1 and ACP*A alleles. The regression of mean per locus heterozygosity (H) on distance from the gene frequency centroid (rii) reveals considerable levels of external gene flow among the Lambadi, the Yerukula, and the two Chenchu subgroups and more homogeneity in the Kolam, Koya, Yanadi, Irula, and Kadar. Mantel statistics were used to assess the relative effects of nonbiological processes (i.e., language and geography) on the morphological and genetic patterns of these subdivided populations. The significance of correlations was determined between different data sets (genetic, dermatoglyphic, anthropometric, geographic, and linguistic) at three levels involving nine, six, and five populations. Although multiple correlation analysis reveals significant combined effects of geography and language on genetics, anthropometrics, and dermatoglyphics, highly significant partial correlations suggest strong effects of geography on both anthropometry and genetics. Our analysis indicates that geographic factors have an overwhelming effect on the genetic differentiation of the south Indian tribal groups.