A critical review of radiocarbon dates clarifies the human settlement of Madagascar. (1st October 2019)

• Record Type: Journal Article
• Authors: Douglass, Kristina; Hixon, Sean; Wright, Henry T.; Godfrey, Laurie R.; Crowley, Brooke E.; Manjakahery, Barthélémy; Rasolondrainy, Tanambelo; Crossland, Zoë; Radimilahy, Chantal
• Is Part Of: Quaternary science reviews. Volume 221(2019)
• Publication Date: 2019-10-01
• Subjects: Quaternary -- Archaeology -- Paleogeography -- Madagascar -- Indian Ocean -- Radiogenic isotopes -- Island colonization -- Extinction; Geology, Stratigraphic -- Quaternary -- Periodicals; Stratigraphie -- Quaternaire -- Périodiques; 551.79
• DOI: 10.1016/j.quascirev.2019.105878
• Languages: English
• ISSNs: 0277-3791

Abstract

Abstract: The timing of the human settlement of Madagascar, one of the last large landmasses to be settled by people, remains a key topic of debate in archaeology. Despite decades of research, recent estimates for initial settlement are increasingly divergent and span ca. 9000 years: the widest colonization window for any island within the reliable range of radiocarbon ( 14 C) dating. 14 C dating of archaeological sites and remains of butchered animals provide important evidence of when the island was first settled, but the reliability of these dates requires critical evaluation. Applying principles of chronometric hygiene, we present the first systematic review of Madagascar's 14 C chronology to clarify the island's settlement. Our findings support human presence by at least 2000 cal BP and suggest that an Early Holocene arrival is possible. The nature of such an early presence on the island, however, remains elusive due to a lack of contextual information. Highlights: Estimates of Madagascar's settlement diverge by ca. 9 ka, the widest colonization window of any island within the range of 14 C dating. The distribution of 14C determinations is heavily skewed toward the Late Holocene. Outlying Early and Middle Holocene dates drive most of the variation in confidence intervals for human arrival estimates. Outliers include dates ranked highly for reliability, precision and clear association with human activity. People and now-extinct fauna coexisted for at least a millennium prior to major faunal population crashes.