Peopling the Americas — New evidence

Print Friendly
Peopling the Americas — New evidence

Artifacts from Buttermilk Creek

This slideshow requires JavaScript.

All images courtesy Michael Waters, Texas A&M University
Browse slideshow to see artifacts from Buttermilk Creek, Texas, date to about 15,500 years ago.

For decades, one name has dominated discussion of the ancient New World: Clovis. Tools representing the characteristic Clovis technology, first found in Clovis, New Mexico, in 1929, have long been considered the product of the first inhabitants of the Americas. With a tool style that’s been found across much of North America, Clovis was the best-selling brand in “the first Americans” competition.

Clovis technology is apparently a home-grown phenomenon, as it’s never been found in Northeast Asia, the source of migrants into the New World.

The oldest solid date for Clovis people is 13,100 years ago, says Michael Waters, an archeologist at Texas A&M University. Now, in an article in Science on March 25, Waters and colleagues argue that tools have been found near Austin, Texas, that date to 15,500 years ago.

The researchers found 15,528 artifacts at a site called Buttermilk Creek. Most of their finds were flakes busted off while making stone tools, but the site also yielded 56 stone choppers, points and scrapers.

Using a technique that calculated when an object was last in direct sunlight, “We took the most conservative route to estimate the age,” says Waters, who directs the Center for the Study of the First Americans at A&M. The stone tools and flakes were probably made by a band of hunter-gatherers who paused at the creekside site.

Seven stone arrows in a row, each with groove that starts at blunt end and goes to arrow's center

The Clovis tool style was marked by the lengthwise groove, a sophisticated bit of stone-work that probably helped secure arrowheads and spear points to shafts. Notice how this feature is absent from the pre-Clovis slide show, above?

Four men and one woman sitting in deep dirt pit, digging and recording with pen and paper

Courtesy Michael Waters, TAMU
Patience, please! Waters’s team of archaeologists comb the dirt to uncover more prehistoric treasures.

Looking for a date

Because no organic remains were available for carbon-dating, the scientists relied for dating on optically stimulated luminescence, or OSL. “OSL has been around for a long time, has been employed in geology for 30-plus years” for dating windblown sand and silt, says Waters. “It’s been compared to radiocarbon dates, toe-to-toe, and in all cases, OSL ages have been determined to be comparable.”

The OSL dating, which essentially figures how long something has been buried, took place at the University of Illinois, in Chicago, under the direction of Steven Forman.

The find at Buttermilk Creek is the latest — and one of the better documented — archeological sites to break the Clovis barrier. Others pre-Clovis finds have been made in Oregon, Wisconsin, Pennsylvania and even Chile.

The news got WhyFilers wondering:

If many other claims for pre-Clovis dates have failed to stick, is the new find really convincing?

What does the new confirmation of earlier occupation say about how people arrived from Northwest Asia?

Why have many archeologists resisted the possibility that the Clovis toolmakers were not the first inhabitants of the Americas?

How convincing?

To get the skinny on the Texas discovery, we phoned Steve Shackley, a professor of anthropology at the University of California at Berkeley. “Proof does not exist in science,” he told us, “but Mike [Waters] has made good, defensible arguments.”

Much of the discussion about the Buttermilk Creek site concerns the vertical position — the stratigraphy — of stone artifacts, and the Waters team went to great lengths to show that older material was under younger stuff, as expected in an undisturbed site. Undetected dislocations can confuse archeologists, who tend to think deeper is older and shallower is younger.

Buttermilk Creek actually offers a three-fer: Clovis artifacts are sandwiched above those now identified as pre-Clovis, but below artifacts are in a more modern style. “This site has all these time periods, superimposed, in the correct order,” says Shackley. Because Waters is “one of the foremost” experts in analyzing the geology of archeological sites, “I think it’s going to be difficult to defeat his stratigraphic work. He’s been very careful about it.”

Douglas Bamforth, an archeologist at the University of Colorado, says the Waters team has avoided three errors that often destabilize ancient archeological claims:

Were the artifacts made by people? “The big question which has occupied the whole debate for stuff older than 11,500 years is whether the objects are really artifacts,” says Bamforth. “There is no question that these stone artifacts were made by people; it’s a total non-discussion.”

Were the artifacts moved after burial? “People don’t sink in the ground, so we think the ground is stable,” says Bamforth, “but objects can move around through freeze-thaw cycles, geologic activity or burrowing animals.”

Are the dates reliable? Even dates from ol’ reliable carbon-dating have been disproved in the past, Bamforth says, but the optical dating used at Buttermilk Creek (which contained no organic material for carbon dating) seems careful and sound.

“They have absolutely dated the site, they absolutely have artifacts, and the article talks in great detail about how intact the sediment was, they have really addressed whether the artifacts are in place,” says Bamforth. “They have refitted the [stone] flakes to the tools; I am totally convinced they have an intact site” and solid dates.

But that does not prove, to Bamforth, that the artifacts are pre-Clovis — they may be early Clovis. “The deep levels at the site are certainly older than the oldest carbon-14 date on Clovis-style projectile points, which Waters very emphatically argues is the beginning of the Clovis period. But the first problem with seeing the deep levels as different from Clovis is that there seems to be exactly nothing in those levels that differs from Clovis [as the site does not contain arrow- or spear-points that would prove or disprove the case]. … So I do not see why the site is not just early Clovis.”

Not so fast!

Aware that the latest find may be seen as final vindication for the “Clovis was not first” viewpoint, we phoned Thomas Dillehay, professor of anthropology at Vanderbilt University and the University of Southern Chile, who fought for decades to have Chile’s Monte Verde site recognized as pre-Clovis. Now that Monte Verde is finally accepted as one of the best-confirmed pre-Clovis sites, we figured the experience would make Dillehay receptive to the new find.


Man crouching and man standing and leaning over, both looking at grassy stream bank. Stream runs behind them.

The age of artifacts found at site in Monte Verde, Chile was long at the center of a heated debate, but the scientific consensus says they are up to 14,500 years old — long predating the first Clovis toolmakers.

We were wrong. “I have a mixed opinion,” Dillehay told us, proceeding to list some shortcomings in the study. “It would be most convincing if there was standard radiocarbon dating, and even better if those dates were taken from features like hearths and food stains. OSL dating has become more reliable, but it’s still not as reliable as carbon-14, although the sequences do line up very nicely with sediment dating.”

Dillehay has questions about the three-layer sandwich of pre-Clovis, Clovis and post-Clovis material. “I’m not saying the materials are mixed. Geologists, to identify the strata, applied these excellent, meticulous sediment and particle analyses, but there was no clear visible stratigraphy to distinguish Clovis from pre-Clovis, and again this does not meet standard archeological criteria.”

Dillehay also points to the lack of “diagnostic, complete projectile points in either the Clovis and pre-Clovis material. In a discipline that has placed incredibly heavy emphasis on formal projectile points as the primary criteria for acceptance of a site, along with C-14 [radioactive carbon] dating, and geologic stratigraphy, I find this sort of acceptance, which seems to be uncritical, to be a major shift in the discipline.”

Still, Dillehay says “the interdisciplinary work is first rate, and I admire the multidisciplinary approach. But had there been C-14 dating and diagnostic projectile points, all this extraneous analysis would probably not be needed.”

It certainly would be nice to find arrow- or spear-points, says Waters, but “You can’t dictate what you will find. You have to roll with the punches.” Further excavation may or may not reveal a “smoking gun projectile point,” Waters adds. “We don’t know what kind of weaponry they used. In Siberia and Alaska, people were using a lot of bone, ivory and antler weaponry, and it might be that early folks in North America were using this as well.”

But due to heat and humidity, such organic material would not be preserved in the Texas site, he says.

Migration routes

The timing of human occupation of North America bears heavily on their migration route from Northeast Asia, which is accepted, for geographic and genetic reasons, as the source of the first Americans. The melting of the last ice age during the Clovis period, starting roughly 11,000 years ago, producing an ice-free corridor through Northwest Canada that would have allowed transit into the North American interior.

Possible Migration Routes

Migrants from northeast Eurasia moved into the Americas through the ice-free corridor in Canada, or along the Alaska coast

A confirmed pre-Clovis date means the first Americans must have migrated by boat along the West Coast, as the ice-free corridor was ice-full around 15,000 years ago.

But the region was clogged with glaciers a few thousand years earlier, meaning that any early immigrants would have moved along the coast, either on foot, or via short hops in boats.

The possibility of coastal movement got a boost in a study1 published March 4, which reported the discovery of stone tools dating from 11,400 to 12,200 years ago on the Channel Islands west of Los Angeles.

According to study leader Jon Erlandson, an archeologist at the University of Oregon, the ancient residents of these offshore islands made delicate stone tools to hunt in the ocean. “The points we are finding are extraordinary, the workmanship amazing. They are ultra thin, serrated and have incredible barbs on them. It’s a very sophisticated chipped-stone technology.”

The stone artifacts are quite different from the fluted points left throughout North America by Clovis and the later Folsom peoples, who hunted big game on land, said Erlandson. “This is among the earliest evidence of seafaring and maritime adaptations in the Americas, and another extension of the diversity of Paleoindian economies.”

The find is yet another reason to doubt that Clovis was first, says Shackley. “When you get dates to 11,000 or 12,000 years ago, out on islands, that makes it tough for the Clovis-firsters, who reject maritime entry. On the Channel Islands, they had get out there by boat,” and if they were already using boats, that means they could also have boated down the West Coast, he adds. “A lot of people accept that now.”


Two stone tools rest in open hand, one half-moon-shaped blade and one sharp arrow point

The recent discovery of delicate stone weapons on California’s Channel Islands boosted the theory that the first Americans could travel by boat while entering the Americas.

But even if people did move south along the coast rather than inland, Dillehay says they probably needed a long time to reach Chile. “There are hundreds if not thousands of rivers that descend the western slope of the mountain chain from Alaska to Tierra del Fuego, and every river, whether major or secondary, is a temptation to head upriver,” slowing the overall southward movement.

And if Monte Verde was occupied by 14,500 years ago, this logic suggests that people reached North America much earlier than even the 15,500 pre-Clovis date in Texas.

Should we trademark the “pre-pre-Clovis” brand?

Opening of cave, three people sitting and writing, one person standing and writing, two people digging

Photo: University of Oregon Northern Great Basin Field School
In Paisley caves in south-central Oregon, researchers uncovered pre-Clovis artifacts and the oldest human DNA discovered in the Americas. Radiocarbon dates show that people lived in the caves between 12,000 and 14,340 years ago.

At any rate, the increasing number of solid pre-Clovis finds answers a riddle: How did Clovis artifacts appear in so many places at roughly the same time? According to the Waters report, “These data are evidence that by 15.5 ka [thousand years ago], human populations occupied the continental United States… . The sites of Cactus Hill, Virginia, and Miles Point, Maryland, hint that these [pre-Clovis] technologies may have been present a few millennia earlier. This early occupation of North America provides ample time for people to settle into the environments of North America, colonize South America by at least about 14.1 to 14.6 ka (Monte Verde, Chile), develop the Clovis tool kit, and create a base population through which Clovis technology could spread.” 2

When science gets ossified

Although we’ve covered the Texas discovery as a bit of gee-whiz archeology, it’s more accurate to say that the discipline proceeds by stacking study atop study, says Sissel Schroeder, a professor of anthropology at the University of Wisconsin-Madison and expert on ancient peoples of the Americas. Although most of the signs left by people who lived in North America will never be found, if they even still exist, “We work with the best information we have. The very small samples of data can make some of our interpretations less robust. Archeology is a cumulative science, so future finds can potentially add confirmatory evidence, or can disconfirm earlier conclusions; you just have to be open to recognizing that your interpretations could change.”


Where did immigrants to the Americas come from more than 10,000 years ago? Why is this region considered the most likely source?

Were all claims for pre-Clovis inhabitation rejected based on poor scientific evidence, or were some rejected for other reasons?

How does the increasing acceptance of pre-Clovis inhabitation change our understanding of the ancient world?

But scientists, like other people, can get stuck, she adds. “It seems easy for certain interpretive frameworks to become quite entrenched, and repeated over and over again. Into the 1920s, it was hugely debated that there were even people in the Americas” at the end of the last ice age. “There were a number of very provocative finds that led scholars to suggest that people had been here at the end of the Pleistocene [about 12,000 years ago], but wasn’t until the find at Folsom, New Mexico [in 1926] that scholarly acceptance began to develop.”

The Folsom find, soon followed by the discovery of those distinctive fluted points near Clovis, New Mexico, sparked “a transformative intellectual step for archeologists,” says Schroeder. “This was a radical shift in thinking.”

“You build a reputation based on a particular perspective,” says Bamforth, “and it’s hard to see evidence that is in opposition; we all believe we are really good at what we do.” Those who gain fame for overturning the conventional wisdom can wind up in the opposite corner, defending their own views long after contradictory evidence arises.

Some early claims for pre-Clovis sites were based on faulty excavation or inaccurate dating, which left a tradition of doubt, Bamforth says. For example, erroneous radiocarbon dates arose after dig sites were contaminated with groundwater. And European-style artifacts unearthed in the Hudson River valley, once interpreted as evidence for ancient European immigration, actually came from ship’s ballast that was dumped into the river, Bamforth told us.

Once archeologists got used to refuting claims, that skeptical attitude itself became entrenched, says Bamforth. “Because people were making such poor claims, very powerful people in the field clamped down on any claims for antiquity, and often the rejected claims turned out to be correct. People at the Smithsonian famously had nothing to do with Folsom until finally the evidence carried the day. There’s a famous photo showing a Folsom spearpoint between the ribs of an extinct bison. That’s proof you can’t argue with.”

Clovis-first is dead, at long last!

After 40 years of assault, Clovis-first seems dead at last. The Texas find “anchors the fact that people were here in the 14,000 or 15,000 year range, there is no longer an argument with that,” says Bamforth.

As the technology of archeology improves, Waters expects some of the most interesting finds to emerge from South America. “We have this North American bias. I’ve heard a lot about early sites in South America of the same age [as the Texas site] or older that nobody hears about. If you think about the immensity of South America, there is no way Clovis was first. There are going to be some amazing finds in the next 10 years, given the South American evidence, the work with genetics and DNA. The story of the first Americans is going to stay exciting.”

– David Tenenbaum

Terry Devitt, editor; S.V. Medaris, designer/illustrator; Jenny Seifert, project assistant; David J. Tenenbaum, feature writer; Amy Toburen, content development executive


  1. Paleoindian Seafaring, Maritime Technologies, and Coastal Foraging on California’s Channel Islands, Jon M. Erlandson et al, Science, 4 March 2011.
  2. The Buttermilk Creek Complex and the Origins of Clovis at the Debra L. Friedkin Site, Texas, Michael R. Waters, et al, Science, 25 March 2011.
  3. Spear points found in TX.
  4. Center for First Americans.
  5. Clovis not first people.
  6. Oldest radiocarbon remains in Oregon.
  7. Prehistoric Beringia.
  8. Gene flow of early Americans.
  9. Emergence of people in North America.
  10. Settlement of the Americas.
  11. Radiocarbon dating.
  12. Meadowcroft rock shelter.
  13. Interactive map of pre-Clovis sites.