Dental evidence for the peopling of the New World: some methodological considerations

Turner (1985b) and Greenberg et al. (1986) proposed that New World populations originated in northern Asia and entered the Americas in three migratory waves: Macro-Indian, Aleut-Eskimo, and Na-Dene. Biological support for this model comes from Turner's unweighted pair group (UPGMA) cluster analysis of discrete dental traits in world populations. Unfortunately, UPGMA analysis often creates suspect clusters and may not be valid as a method for displaying evolutionary relationships because it assumes that rates of evolution are equal among all populations. To test whether Turner's results are an artifact of his analytical method, I analyzed his published data (Turner 1985b) using two alternative techniques that do not assume homogeneous rates of change: a Wagner distance algorithm employing the Fitch-Margoliash criterion for goodness of fit and a maximum parsimony analysis using segment-coded dental trait frequencies. Both alternative methods produce trees that are similar to the UPGMA analysis results, supporting Turner's original results and basic conclusions. Comparisons of tree topology demonstrate that there is strong congruence between trees produced by all three methods, although the placement of certain populations, such as Na-Dene, depends on the method of analysis employed.