Evolutionary population history of early Paleoamerican cranial morphology

Abstract

The nature and timing of the peopling of the Americas is a subject of intense debate. In particular, it is unclear whether high levels of between-group craniometric diversity in South America result from multiple migrations or from local diversification processes. Previous attempts to explain this diversity have largely focused on testing alternative dispersal or gene flow models, reaching conflicting or inconclusive results. Here, a novel analytical framework is applied to three-dimensional geometric morphometric data to partition the effects of population divergence from geographically mediated gene flow to understand the ancestry of the early South Americans in the context of global human history. The results show that Paleoamericans share a last common ancestor with contemporary Native American groups outside, rather than inside, the Americas. Therefore, and in accordance with some recent genomic studies, craniometric data suggest that the New World was populated by multiple waves of dispersion from northeast Asia throughout the late Pleistocene and early Holocene.

Keywords: cranial shape; geometric morphometrics; peopling of Americas; population history.

Fig. 1. Geographic and historical relationship between populations.

(A) Map showing the geographic position of each contemporary population and the Paleoamerican sample from Lagoa Santa (Brazil). Stars denote the waypoints used to calculate more realistic geographic distances between populations. (B) Tree topology representing the hierarchical model of historical divergence among populations. The null history model places the Paleoamericans as a sister group to their nearest geographic neighbor (Chubut, Patagonia).

Fig. 2. Population affinities based on craniometric data.

2D nonmetric MDS plots of population morphological affinities based on average Procrustes distance. (A) Cranium. (B) Vault. (C) Basicranium. (D) Face. Mantel tests were used to assess the congruence between affinity patterns based on different cranial data sets. The cranium was strongly correlated with the vault (r = 0.873), face (r = 0.786), and the basicranium (r = 0.704). The vault was also relatively strongly correlated with the face (r = 0.624) and the basicranium (r = 0.609), and the face and basicranium were moderately correlated (r = 0.327). All Mantel tests were significant at the α = 0.0083 level following Bonferroni adjustment.

Fig. 3. Results of multiple-effects model comparisons.

History models found to be significantly better than the null model show the position of the Paleoamerican population (black diamond) in green (α = 0.05) and red (α = 0.01), with the absolute best model in black for each of the four cranial modules.