America came late. Humanity had come from Africa, and spread out over Eurasia. This happened mainly during the ice age, when lower sea levels allowed walking migration where previously a significant swim or float had been required. They moved mainly along the coast line, quickly reaching southeast Asia, and from there even Australia, as early as 50,000 years ago. The precise dates may be in question.

Moving inland into Eurasia was more difficult, in part because of the mountain ranges left by the extinct Tethys ocean. River valleys provide some passable routes. The cold climate formed another barrier which limited the possibilities. Still, even the tundra could provide a good living for those who could adjust to the cold. Southern Europe and central Asia were occupied.

There was another continent waiting: America. The Bering Strait had become dry land and was now part of the Asian tundra. It is now known as name Beringia, but it no longer exists. People lived in Beringia probably from 25,000 years ago. At the time they were still part of the Siberian population, as shown by genetic evidence – of their dogs!

Alaska connected to Beringia and is sometimes referred to as East Beringia. It is not fully known when Alaska was settled. There are a few indications as old as 20,000 years (or perhaps even earlier) but the evidence that people were involved is not conclusive. The oldest securely dated human occupation in Alaska is at Swan Point, north of Delani; it has been dated to 14,200 BP. The site was occupied for perhaps only a few hundred years, although people would return at later times. During the brief early phase, microblades were used which are similar to those in use in eastern Siberia. Their diverse cuisine (from excavation of a hearth) included mammoth, horse and caribou! Only the latter would eventually escape extinction.

From Alaska into the plains of North America was no walk in the park. During the ice age, much of Alaska was ice-free: the area was too dry for ice growth. But the southern region was covered in ice, as was almost all of Canada: the glaciers blocked off all routes into the continent. 20,000 years ago was still the depth of the ice age: it was bad luck that the only traversable route into the Americas led to a region north of the impenetrable ice sheet!

Clovis First – or not?

An ice-free corridor opened up east of the Rocky Mountains, between the Cordilleran and Laurentide ice sheets. This may have happened as early as 15,000 year ago or as late as 13,000. By the later time, bison and elk had began to move across the corridor. The opening provided a route towards Montana. Only a limited number of people made it: the native populations have low genetic diversity, evidence of the ice age genetic bottleneck.

This may have been the origin of the Clovis culture. Clovis sites in North America are seen from 13,000 years ago, and cover a rather brief period of around 300 years. They are defined from a distinctive type of projectile point, which indicates a common origin. The sites are found across much of North America, but are absent from the Pacific coast.

But this ice-free corridor was not the only way south. From about 17,000 year ago, the islands along the coast of British Columbia began to become ice-free. Island hopping provided another route south, and this could have been done as early as 16,000 years ago. We do not know whether this route was used. Boats would have been needed. The oldest evidence of boat use is several thousand year later along the California coast. The oldest dates for confirmed human occupation along the route are less than 13,000 years old. But the coastal route was open earlier than the other corridor and could have given a limited number of people a head start.

Clovis was among the oldest population in the region. But it was not the only one. In the west (Oregon and Idaho), a different projectile style has been found, with an oldest date around 400 years before Clovis. At the southern point of South America, yet another style has been dated to 12,900 BP.

There are three conclusions from this. First, although the Clovis culture was early, it was not the first. People had already been present south of the ice sheet for perhaps 1000 to 2000 years. Second, people did not arrive any earlier than 17,000 years ago. This is supported by genetic studies of people, and by their dogs which genetically split from Siberian dogs no earlier than around 17,500 BP. Third, people arrived in South America within a few thousand years: genetic studies indicate that populations in Central America and South America separated around 13,000 years ago. After reaching South America, people moved south along both the Pacific and the Atlantic coast, over several hundred to several thousand years.

The actual arrival probably occurred shortly after 15,000 years ago which would have been almost immediately after the ice-free corridor began to appear. This timing comes from several well-dated remains of hunting of mammoth and camels (!) in Wyoming, Michigan and along the Gulf of Mexico, giving dates between 15,000 and 14,500 years ago. Surprisingly, there are also very old dates from Peru and Chile at 14,200 years ago which requires a rapid spread across every different climate zones. The tip of South America was reached by the time of Clovis.

So far, this forms a consistent picture where a single and isolated founder population spread from Alaska during the waning of the ice age, found a walking buffet on the American plains and rapidly spread out across the new land. The ages favour the coastal route over the in-land corridor, but we lack direct evidence for this.

There are some complications. One theory has a group from Europe reaching the Americas early on. This is based on skull appearances, but it fails against genetic studies: there is no trace of European ancestry among the original native population. The second complication comes from indications of human presence older than 20,000 years ago, at a time the ice sheets still blocked off any access and well before the date where genetic studies place the Beringia founder population. Footprints at White Sands have been dated to 21,000 to 23,000 years ago. But other groups find that these dates could be overestimated by as much as 7,000 years which would bring them into line with the arrival time mentioned above. It remains subject to on-going discussion. Finally, there is a claim that some tribes in the Amazon have a minor genetic link to native populations in PNG and Australia, indicating a common ancestral population. That claim is again disputed, but it is clear that ancient populations east and west of the Andes are not as closely linked to each other, probably related to the separate migration routes along the Pacific and Atlantic coast.

So far for the state of knowledge on the peopling of the Americas. Time to move on to the topic of this post.

Monte Verde

The location is in the Lake District of Southern Chile, along the Chinciuapi Creek which flows into the Maullin River. The landscape is relatively flat and the creek meanders. For the readers of VC, we may mention the (irrelevant) information that it is only 40 km west-southwest of Calbuco volcano – the one with the spectacular VEI-4 eruption in 2015.

The discovery of the site started with farmers finding a strange bone, subsequently identified as from the elephant family. Digging started in 1977, and quickly revealed signs of occupation – including butchery of the elephant’s relatives, as well as ancient forms of llamas. The site had been protected by an anaerobic peat bog, covering the remains and preventing further human interference. There are several separate sites, of which the one called Monte Verde II was found first and remains the best studied.

The site contained many hearths of varying sizes, and 12 huts about 10 square meter in size. Stone tools were found, some as cobbles split to make a workable edge, and in two cases as flaked points. Logs and branches had been worked. Animal remains were plentiful, and plant remains were varied and came from distant places ranging from the coast to the Andes. They included the oldest potato known, and many varieties of seaweed that must have come from at least 60-km away. (At the time, the sea coast was further away.) There was even a footprint left in the clay.

Carbon dating yielded an age of around 14,200 years. In 1989, when the results were first presented, such an old date was not easily accepted. It was older than Clovis, and ‘everyone knew’ that the Clovis culture had been the first. Pre-dating Clovis was bad enough, but doing so this far south in South America was worse. How had people gotten here so quickly? Had they bypassed North America? Was this the beginning of the end of the Clovis primacy? Or was the date wrong?

Later work found some indication for an even older layer of occupation, called Monte Verde I. This is strongly disputed, as there is no direct evidence for human involvement in the debris. The date can be several thousand years before Monte Verde II. However, the discussions have focussed on Monte Verde II, as there is extensive evidence for human activity, some of which requiring a trading network and thus an established (albeit thinly populated) society, together with C-14 dating. It is hard to argue with this!

In 1997, a group of independent specialists visited the site to judge the strength of the evidence. They arrived with doubts but by the end of the visit strongly supported the findings. After this, the date became generally accepted. From now on, the Clovis people were firmly relegated to a status as late-comers. Chile had the new oldest ancestor and the peopling of the Americas had to be re-thought. This is the origin of the current view as described above. Blame Monte Verde.

It was not the end of the discussions, and as can happen with the overthrow of a long established fact of supremacy, some of the discussions became vitriolic. The C-14 dates were questioned. This is not uncommon: an independent test is always useful, as is a comparison with for instance luminescence data (which can have its own uncertainty). Often the carbon is taken not from the objects themselves but from associated material. For instance, C-14 ages can not be measured from a footprint! Instead organic material found in the footprint may be used. Neither can a stone be directly dated. (This is also a problem for volcanic eruptions: lava does not contains useable carbon, but vegetation covered and perhaps burned by the flow may be used.) If the carbon is from water (e.g. fish or coral), it may gives dates that are centuries too old because it measures the time when the CO2 was last in the atmosphere. So these are normal discussions, but carrying more than usual emotions.

Eventually the discussions quieted down and the Monte Verde II dates became accepted. As more pre-Clovis sites were found, the ‘Clovis First’ hypothesis was weakened and this removed some of the emotional weight of Monte Verde. Still, the site remained exceptional. There is only one comparable date in all of South America, and that is in Peru, on the far side of the inhospitable and possibly impassable Atacama desert. All other dates are younger than 13,000 years. Genetic studies also favour that Central and South America population only separated around that time. Thus, Monte Verde still retains some of its contentiousness.

Re-dating Monte Verde

The discussion has now been re-opened by a team which carried out an independent investigation at the site. Their paper and especially their conclusion was published this month and it attracted a fair amount of attention. The following derives from that paper.

The Monte Verde site lies in a small valley which contains the current creek. The surrounding plain is a few meters higher. The different surface levels are indicated on the drawing below as T1 to T4. They correspond to three different layers of the soil, called SU1 to 3 – ‘SU’ apparently representing the first letters of the author. (This is not recommended in scientific publications, by the way!)

The SU1 layer is common around the wider area. It has formed over a long time. Carbon dated old wood from the site shows that the base of SU1 is 16,000 years old. The SU1b layer is a wood layer: measurements give it an age of 14,000 years, consistent with the reported dating for Monte Verde II. SU1c is a layer of organic clay, dated to 10,000 to 11,000 years but with some uncertainty.

Above the clay is a 5-cm thick tephra layer, indicated as SU1d in the figure. The team associates this with the Lepué Tephra, a widespread layer dated to 11,000 years ago. The layer above, SU1e, is considered a young layer of sediments and peat. (The paper assigns dates for SU1d for two pieces of wood of 3,000 and 100 years, but in one case this represents an intrusion (e.g., a plant root) and in another case the piece was mislabeled and belongs to SU1e – private communication.)

The T2 surface is where the debris from the Monte Verde II settlement is located. It formed from river bed erosion into the SU1 layer, followed by sediment deposition. This sediment is named SU2 in the figure. The Monte Verde II debris is just above the interface, so near the bottom of SU2. One piece of wood from near the bottom of this layer, albeit not collected in this study but in the 1997 visit, gives an age of 13,000 year. Measurements higher up give ages of 8,000 to 2,000 years. This indicates that SU2 is mostly considerably younger than SU1. The older piece of wood may have come from layer SU1b, and fallen into the ditch at the time of the erosion. However, this is speculative.

The authors propose that the erosion that formed the ditch or creek happened in the mid-holocene. Between 15,000 and 11,000 years, this region was a temperate rainforest and marshland. After 11,000 years the region dried out during a much warmer phase, allowing the creek to form and cut into the drier landscape. The proposed scenario is shown in the figure below, taken from their paper.

In this scenario, Monte Verde II dates to the middle holocene and is younger than 8,000 years. It is quite a change from the original 14,200 age! The C-14 dating was correct, but the association of the dated material with the settlement was not.

It is a strong case, though perhaps not yet fully conclusive. It explains much of the appearance of the site, and also puts Monte Verde II among a more developed culture where for instance the trading of food over long distance is plausible. It also avoids the need for a very fast expansion deep into South America. However, it requires that all the debris used in the previous C-14 dating from material associated with the Monte Verde II excavation came from material which had originally come from elsewhere, and had fallen in: the hamlet was built on this material. One might have expected a wider range of dates from this scenario.

A volcanic connection

The crucial point of the argument lies in the Lepué Tephra. This layer was not found at the Monte Verde II village itself, but only at the T1 area. This suggests that this tephra layer was removed by the ditch erosion, and thus that the ditch or creek formed after this time. (The alternative, not mentioned in the paper, would be that this layer was removed long after the village had been there. A second erosion surface was not observed, but a single flood around the incised river bed could perhaps have done this.) It is now up to other teams, most likely the original one, to make these checks and to confirm (or otherwise) this new result.

What do we know about this tephra? The Lepué Tephra is found across northwest Patagonia, and is assigned to Volcán Michimahuida, near Chaitén. It lies 160 km southeast of Monte Verde. Chaitén had a major rhyolitic eruption in 2008 (VEI 4), after millennia of dormancy. In contrast, Michimahuida is andesitic, so its eruptions are easy to distinguish from Chaitén. Michimahuida is also much taller than Chaitén, at 2400 meters.

As an aside, this region of southern Chile may be under-watched: there have been three major eruptions here already in the past 20 years: Chaitén in 2008, Puyehue-Cordon Caulle in 2011 and Calbuco in 2015.

The volcano is glaciated which can make it hard to study. It had a phreatomagmatic VEI-6 (!) eruption dated to 10,500 years ago, and a VEI-5 3,000 years later. There was also an eruption in 1834, observed by Charles Darwin, when lava reached the sea. The summit glacier fills in a caldera which may have formed in the 10,500 year old Lepué eruption. It is one of the largest eruptions in this region since the ice age, based on the extent of the tephra. It also formed a pyroclastic flow which filled in a valley and is known as the Amarillo Ignimbrite.

If Monte Verde II is as old as it used to be, it will not be the only such site in the area. In that case, there must be other finds waiting to be found underneath the Lepué Tephra. People were already present before this eruption, but the question is whether they show similarity to Monte Verde II. If not, this would further strengthen the case for the younger date.

Final points

Monte Verde II started out being controversial. It pulled the rug from underneath the Clovis First hypothesis, and people took this personally. A site visit by independent experts was needed before it became accepted. But the disputes did not die down. Every aspect was challenged. One wonders, how much of that was down to personal opinions rather than the weight of evidence? This recent paper returns to Clovis First and links the re-dating of Monte Verde II to the case for the primacy of the ice-free corridor for populating North America, against that of the coastal route. It is a bit of a jump.

Regardless of the final date of Monte Verde II, the prior research had already shifted the paradigm. It opened up the study of the pre-Clovis settlement of the Americas, allowing an unbiassed look at the timing of the opening corridor versus the dates of pre-Clovis sites in North America. Clovis First now seems rooted in the past: the question is what society gave rise to Clovis.

I’ll end with Tom Dillehay, the original investigator of the site, who in 2019 responded to another paper questioning Monte Verde II. He wrote: “it is healthy for the discipline to criticize new and old discoveries, but it should be done in a constructive, judicious, balanced, and empirically justifiable manner.” This new paper has done just that. I am looking forward to the inevitable response with great interest.

Albert, March 2026

Further reading (but may not be open access)

Waters, 2019, Late Pleistocene exploration and settlement of the Americas by modern humans. Science, Vol 365, https://www.science.org/doi/10.1126/science.aat5447

Dillehay, 2008, Monte Verde: Seaweed, Food, Medicine, and the Peopling of South America, Science, 320, https://www.science.org/doi/full/10.1126/science.1156533

Surovell et al., 2026, A mid-Holocene age for Monte Verde challenges the timeline of human colonization of South America, Science, Vol. 391, https://www.science.org/doi/epdf/10.1126/science.adw9217