Writing in Nature, Ardelean et al.1 and Becerra-Valdivia and Higham2 report evidence that the initial human settlement of the American continent happened earlier than is widely accepted, and some of this evidence suggests that expansion into the continent began at least 10,000 years earlier than was generally suspected. A study of radiocarbon dating of early archaeological sites by Becerra-Valdivia and Higham reveals that interior regions of Alaska, Yukon in Canada and the continental United States were already widely populated before 13,000 years ago. For decades, that time frame was widely considered to mark the earliest possible date of initial entry, until data from sites more than 13,000 years old in North and South America, first reported in the 1970s, raised the possibility of earlier arrivals3–5. Archaeological excavations in Chiquihuite Cave in northern Mexico by Ardelean and colleagues provide evidence of human occupation about 26,500 years ago. This Mexican site now joins half a dozen other documented archaeological sites in northeast and central Brazil that have yielded evidence suggesting dates for human occupation between 20,000 and 30,000 years ago6–12.
Following discoveries in the 1930s on the American Great Plains of distinctive, well-crafted stone spear points — of a type connected with the Clovis culture — alongside bones of mammoths, mastodons and a now-extinct bison species, archaeologists maintained, for many of the following decades, that the earliest people in the Americas were specialized big-game hunters who very rapidly expanded into North and South America, within 1,000 years of initial entry13. This model became known as the Clovis-first theory. It was later established that Clovis technology did not reach the southern continent. The time of their entry from Alaska into what is now the continental United States was thought to coincide with the opening of an ice-free corridor (Fig. 1) by around 13,000 years ago between the great northern continental ice sheet (called the Laurentide Ice Sheet) and the ice-covered northern Rocky Mountains (the Cordilleran Ice Sheet) in western Canada.
However, beginning in the mid-1970s, researchers identified archaeological sites in the Americas dated to earlier than 13,000 years ago, especially in South America. For example, the site of Monte Verde II in south-central Chile, initially dated to 14,500 years ago3, is a well-preserved open settlement with wooden structures and artefacts indicating a lifestyle based mainly around a diet of plants (subsequent discoveries revealed earlier occupations of this site14). Other early archaeological sites in South America on the Pacific coast, in the northern and central Andes, on the Caribbean coast, in the Brazilian uplands, in the Amazon basin, and on the Patagonian steppe in Argentina indicate that all major environmental zones of the region were occupied by people with diverse ecological adaptations and technologies before around 13,000 years ago4.
Becerra-Valdivia and Higham carried out a statistical analysis of radiocarbon dates from early archaeological sites widely distributed over the continent of North America and Beringia (the land that once joined Alaska and Siberia in the Bering Strait area). Their results now establish that, by 15,000 years ago, North America was also widely settled, with some data suggesting sparse occupation earlier than that; and several distinctive regional traditions in stone-tool technology had developed by 13,000 years ago. On the evidence of these early archaeological sites from more than 13,000 years ago, identified on both continents, the Clovis-first model must be discarded. Clearly, people were in the Americas long before the development of Clovis technology in North America.
Instead, the key issue becomes how much earlier the Americas were initially peopled than was previously thought. One aspect to consider is the route or routes that people took in expanding south of Alaska. This is the assumed entry point from northeast Asia through the Bering Strait area. However, for a long interval during the last major glacial advance (dated to between about 26,500 and 19,000 years ago15), the obvious route by land down through the lowlands east of the Rocky Mountains was blocked by the merger of the Cordilleran and Laurentide ice sheets. An alternative route down the Pacific coast by populations adapted to life at the shoreline has gained strength as a possibility, as a result of increasing archaeological research in coastal zones16. Another option to consider is an initial entry before the closure of the ice-free corridor during the last major glacial advance.
This is where the evidence from Chiquihuite Cave comes in. After an initial test excavation suggested that the site was of great antiquity, Ardelean and colleagues continued their research using a range of scientific techniques. They recovered stone artefacts of a distinctive technology located in layers with dates corresponding to around 27,000 years ago in the lowest parts of the cave’s sedimentary deposits, and the authors uncovered more artefacts in higher layers that dated to up to 13,000 years ago. The dating for the layer with the earliest artefacts indicates that there were people in northern Mexico at a time corresponding to the beginning of, or early during, the last major stage of glacial advance in North America.
Ardelean and colleagues’ suggestion that the initial entry date was as far back as 33,000 years ago, which is more than double the currently popular date of around 16,000 years ago, will be very hard for most archaeologists specializing in early America to accept. There will undoubtedly be challenges to this interpretation and close examination of the site data. The six Brazilian archaeological sites dated as older than 20,000 years ago, five in the centre of the state of Piauí6–11 and one in central Mato Grosso (the Santa Elina rock shelter)12, although expertly excavated and analysed, are commonly disputed or simply ignored by most archaeologists as being much too old to be real. The findings at Chiquihuite Cave will bring about fresh consideration of this issue.
One unanswered question is why no archaeological site of equivalent age to Chiquihuite Cave has been recognized in the continental United States, assuming that, with a Bering Straits entry point, the earliest people expanding south must have passed through that area. With the coastal-entry model, it might be presumed that the earliest archaeological sites are now submerged offshore by the rise in sea level at the end of the last ice age. For the continental interior, it might be a matter of identifying and carefully investigating geological or palaeontological localities of appropriate age, searching for traces of human presence, and re-examining previously discounted archaeological sites and collections for now-recognizable evidence of human behaviour. In the light of these new discoveries, archaeological research into this period should intensify.
Nature 584, 47-48 (2020)