Patterns of Biological Variation Among the 20 Endogamous Groups of Dhangar Caste-Cluster of Maharashtra, India

REDDY, B. MOHAN; DEMARCHI, DARIO ALFREDO; MALHOTRA, K. C.

COLLEGIUM ANTROPOLOGICUM

Croatian Antropological Society

Lugar: Zagreb; Año: 2001 vol. 25 p. 425 - 442

0350-6134

We study patterns of variation among the 20 endogamous groups of Dhangars, acaste-cluster from Maharashtra State of India, who are semi-nomadic shepherds andcattle herders. To understand patterns of variation, we subjected the data on fourteenanthropometric measurements of about 2,500 adult males and data on 6 genetic markers,published among 13 of the 20 Dhangar castes, to R-matrix analysis, Harpendingand Ward model of regression of heterozygosity on the distance from centroid of the populations,spatial autocorrelation analysis and Mantel statistics of matrix correspondenceof the distances ? geographic, anthropometric and genetic. Results of multiple regressionanalysis suggest a high degree of association between allele frequencies and thegeographic longitude and latitude; R2 value suggests that about 70% of the variance inRH7 and ACP can be assigned to geographic distribution of groups. In case of anthropometry,this association with body size is found to be even stronger. Results of spatialautocorrelation analysis, as suggested by Moran?s (I), are somewhat complementary tothose based on multiple regression analysis. Mantel test indicates significant associationbetween anthropometric distances and the geographic distances, not between geographicand genetic distances. The extent of differentiation of Dhangar sub-castes ismuch higher in anthropometric traits (FST = 0.068) when compared to the genetic markers(FST = 0.023). Yet, the FST value obtained for genetic markers is larger than the averagefor the Indian populations, based on similar class of markers. The positioning of thegroups in the multivariate space reflects primarily geographic proximity of the groupswith reference to anthropometric dimensions while no tangible pattern is evident for geneticmarkers. The plot of average heterozygosity of the groups versus their distancefrom the gene frequency centroid seems to reflect population size variation, rather thangroupvariation in external gene flow.