Genetic estimates of the initial peopling of Polynesian islands actually reflect later inter-island contacts

Abstract

The timing of the initial peopling of the Polynesian islands remains highly debated. Suggested dates are primarily based on archaeological evidence and differ by several hundred years. Ioannidis et al. [2021] used genome-wide data from 430 modern individuals from 21 Pacific islands to obtain genetic estimates. Their results supported late settlement dates, e.g. approximately 1200 CE for Rapa Nui. However, when investigating the underlying model we found that the genetic estimator used by Ioannidis et al. [2021] is biased to be about 300 years too old. Correcting for this bias gives genetic settlement dates that are more recent than any dates consistent with archaeological records, as radiocarbon dating of human-modified artifacts shows settlement definitively earlier than the bias-corrected genetic estimates. These too-recent estimates can only be explained by substantial gene flow between islands after their initial settlements. Therefore, contacts attested by archaeological and linguistic evidence [Kirch, 2021] must have been accompanied also by demographically significant movement of people. This gene flow well after the initial settlements was not modelled by Ioannidis et al. [2021] and challenges their interpretation that carving anthropomorphic stone statues was spread during initial settlements of islands. Instead, the distribution of this cultural practice likely reflects later inter-island exchanges, as suggested earlier [Kirch, 2017].