The genetic structure of a tribal population, the Yanomama Indians V. Comparisons of a series of genetic networks*,†

SUMMARY The model of Cavalli‐Sforza and Edwards was used to investigate the gentic relationships of thirty‐seven Yanomama villages. Using gene frequency data from eleven genetic systems (MNS, P, Rh, Duffy, Kidd, Diego , Lewis, haptoglobin, Ge, PGM, and acid phosphatase), three hierarchical levels of analysis were carried out. The first level involved in the construction of a gentic network, based on genetic distance between villages, for each of three historically defined clusters of villages. Upon comparing the results with ethnohistory there was shown to be a good agreement between the historical development of a cluster and its genetic network. At the second level of analysis a genetic network was constructed in similar fashion for nineteen villages of the Western Yanomama subgroup, including the three village clusters previously analysed. At this level the integrity of the village clusters was largely maintaine, exceptions being due to specific socio‐political events that resulted in changes of the composition of the village gene pool. The highest level resulted in the construction of a network for all thirty‐seven villages which showed three of the five geographically defined tribal subgroups as distinct entities. The result of these analyses indicate that the model of Cavalli‐Sforza and Edwards can under certain circimstances be used when the requirements of the midel arenot met, and furthermore that the rate of divergence of human populations may approach a steady state at the level of major tribal subgroupings.