Bones In A Canyon Wall<\/p>\n
The canyon wall revealed them in the summer of 1959 \u2014 two human femora, the large bones of the thigh, projecting from sediment nearly 37 feet below the modern surface of Arlington Canyon on Santa Rosa Island, California.<\/p>\n
Phil C.<\/p>\n
Orr had not come to this particular spot looking for human remains.<\/p>\n
The curator of anthropology and paleontology at the Santa Barbara Museum of Natural History was on Santa Rosa Island investigating the evidence for pygmy mammoths \u2014 the dwarf elephants whose fossils had turned up in the island’s ancient sediments \u2014 when erosion in the canyon wall exposed something that immediately changed the direction of his work.<\/p>\n
Two femora.<\/p>\n
Human.<\/p>\n
Buried impossibly deep.<\/p>\n
Thirty-seven feet is not the depth at which a burial becomes ancient \u2014 burial depth alone means little in archaeology.<\/p>\n
But Orr had spent years studying the layered sediments of Santa Rosa Island and its geological history.<\/p>\n
He recognized the stratum from which the bones protruded: water-laid sediments, deposited by ancient streams, that belonged to the terminal Pleistocene \u2014 the final chapter of the last Ice Age.<\/p>\n
The bones, in his assessment, came from a time roughly 10,000 years before the present.<\/p>\n
That assessment, later refined by modern methods to approximately 13,000 calendar years ago, would make the Arlington Springs individual among the oldest securely dated human remains ever found in North America.<\/p>\n
But there was something the age alone did not immediately announce about this find.<\/p>\n
It required a moment of geographic context to grasp.<\/p>\n
Santa Rosa Island is not California.<\/p>\n
It is separated from the California coastline by the Santa Barbara Channel, a body of open ocean.<\/p>\n
In the terminal Pleistocene, the Channel Islands offered less water to cross \u2014 sea level was substantially lower than today, reducing the gap between island and mainland \u2014 but the gap remained.<\/p>\n
<\/p>\n
Even at the lowest sea level of the late Pleistocene, Santa Rosa and its neighboring islands formed a single large landmass that researchers now call Santarosae, and Santarosae was still an island.<\/p>\n
It had never been connected to the mainland during the time that human beings were on Earth.<\/p>\n
Someone had crossed open water to reach it.<\/p>\n
That single deduction \u2014 that a human being was standing on an offshore California island roughly 13,000 years ago, and that reaching it required a water crossing \u2014 became one of the most significant individual data points in the debate over how and when the Americas were first settled.<\/p>\n
The bones could not tell Orr the person’s name, language, or exact life story.<\/p>\n
They could not directly answer the larger questions they raised.<\/p>\n
But their presence in that canyon wall 37 feet below the surface announced something to every researcher who later understood the geology of the site:<\/p>\n
By the time the last Ice Age ended, maritime technology was already part of the human story in the Americas.<\/p>\n
This is the story of those two bones, the man who found them, the science that eventually fixed their age, and what that age meant for the history of a continent.<\/p>\n
Phil Orr And The Museum That Sent Him To The Island<\/p>\n
Phil C.<\/p>\n
Orr was a scientist of the mid-twentieth century, operating in the era when the scientific professions were less fragmented than they became later, when a single curator at a regional museum of natural history might responsibly study archaeology, paleontology, and the physical geology that underpinned both.<\/p>\n
Orr had joined the Santa Barbara Museum of Natural History in the late 1930s and spent most of his professional life there.<\/p>\n
His interests centered on the California Channel Islands \u2014 the archipelago off the southern California coast that consisted, at that time, of eight named islands divided between northern and southern groups.<\/p>\n
<\/p>\n
The northern Channel Islands \u2014 San Miguel, Santa Rosa, Santa Cruz, and Anacapa \u2014 were his particular area of focus.<\/p>\n
Santa Rosa Island was, and remains, the second largest of the Channel Islands, a roughly 84-square-mile piece of terrain some 50 miles west of Santa Barbara.<\/p>\n
It is rugged country \u2014 canyons cut by streams through deposits of marine terrace and alluvium, ridges covered in shrubby vegetation, coastline alternating between sea cliffs and small beaches.<\/p>\n
The island belongs to the Channel Islands National Park, established in 1980, but in Orr’s time it was a working cattle ranch.<\/p>\n
He had been coming to Santa Rosa Island for years before 1959, building a picture of its geological and archaeological history.<\/p>\n
<\/p>\n
He had investigated its Pleistocene-era deposits, documented evidence of the pygmy mammoths whose remains appeared in the island’s ancient sediments, and compiled a growing record of the island’s human prehistory.<\/p>\n
The island had been occupied by Native people long before Europeans arrived, and Orr was assembling evidence of that long history.<\/p>\n
The canyon where he found the bones in 1959 was called Arlington Canyon \u2014 a deep arroyo cutting through the island’s sedimentary layers in its western interior.<\/p>\n
The canyon is fed by Arlington Springs, a water source that had supported life on the island across geological time.<\/p>\n
Streams running through the canyon had, over millennia, deposited layered sediments recording climate changes, vegetation changes, and the biological history of the island.<\/p>\n
<\/p>\n
It was while examining the canyon walls for mammalian fossils \u2014 the bones of the pygmy mammoths \u2014 that Orr saw something in the sediment that was distinctly not a mammoth.<\/p>\n
Two femoral fragments.<\/p>\n
Human.<\/p>\n
Eroding from a layer that, by every indication he could read in the stratigraphy of that canyon wall, dated to the end of the Pleistocene.<\/p>\n
Orr recognized what he had immediately.<\/p>\n
He also recognized the problem.<\/p>\n
Radiocarbon dating \u2014 the technique that had given archaeologists their first reliable clock for measuring the age of ancient organic material \u2014 was still a young science in 1959.<\/p>\n
It had been developed by Willard Libby in the late 1940s, and the methods available in 1959 were not capable of producing the precise age determination that fragmentary ancient bone required.<\/p>\n
<\/p>\n
The bones were too important to risk on an inadequate test.<\/p>\n
He preserved them for the future.<\/p>\n
What Orr Published And What He Could Not Know<\/p>\n
Orr published a brief account of the Arlington Springs find in the journal Science in 1962 \u2014 a paper that remains a foundational document in the archaeology of early California.<\/p>\n
The paper was direct: bones of a man, found at a depth of 37 feet in water-laid sediments on Santa Rosa Island, California, dated by radiocarbon at 10,000 years before the present.<\/p>\n
The dating at 10,000 years before present was based on associated charcoal from the same stratum \u2014 not directly on the bone itself, because the chemical preparation methods required for reliable radiocarbon dating of ancient bone protein were not yet available.<\/p>\n
<\/p>\n
Charcoal is well-suited for radiocarbon dating; it concentrates carbon in a chemically stable form that resists contamination reasonably well.<\/p>\n
But charcoal from the same layer as the bones is not the same as a date from the bones themselves, and the relationship between the charcoal and the human remains \u2014 whether they were truly deposited at the same moment \u2014 required geological interpretation.<\/p>\n
Orr’s interpretation, based on his reading of the stratigraphy, was that the bones and the charcoal were contemporaneous.<\/p>\n
The bones had come from the same water-laid sediments that contained the dated charcoal, from a depth that placed them at the end of the Pleistocene.<\/p>\n
<\/p>\n
Even if the 10,000-year figure was accepted as approximate \u2014 and even that was already an extraordinary claim in 1962 \u2014 the implication was significant.<\/p>\n
In 1962, the question of when the first humans reached North America was still organized around the Clovis paradigm: the idea that the first Americans were the makers of the distinctive fluted stone projectile points first found near Clovis, New Mexico, dated to roughly 11,500 years before present by conventional radiocarbon dating.<\/p>\n
A human being in California approximately 10,000 years ago was not outside the Clovis window, but it was on its trailing edge, and the island context made it geographically unusual.<\/p>\n
The bones went back into museum storage.<\/p>\n
<\/p>\n
They would wait there for three decades before the technology needed to test them properly would exist.<\/p>\n
What Orr could not have known in 1962 was that the bones were older than his estimate.<\/p>\n
Significantly older.<\/p>\n
The charcoal date of approximately 10,000 radiocarbon years was accurate for what it measured, but radiocarbon years and calendar years are not the same thing, and the calibration curve that converts one to the other involves complexities that were not fully worked out in 1962.<\/p>\n
The true age of the Arlington Springs individual, in calendar years, was something that later science would have to establish.<\/p>\n
<\/p>\n
He also could not have known how the island geography \u2014 the channel of open water separating Santarosae from the mainland \u2014 would eventually become the central argument of the entire find.<\/p>\n
The Island At The End Of The Ice Age<\/p>\n
Before continuing the dating story, it is necessary to understand the place.<\/p>\n
Santa Rosa Island today is an island.<\/p>\n
It sits roughly 50 miles west of Santa Barbara, surrounded by the Santa Barbara Channel on one side and the open Pacific on the other.<\/p>\n
The channel that separates it from the mainland is roughly 26 miles wide at its narrowest.<\/p>\n
In the summer, the channel can be relatively calm; in winter, it can be difficult.<\/p>\n
<\/p>\n
The island is accessible by boat or small aircraft; it is not accessible on foot.<\/p>\n
In the terminal Pleistocene, the geography was different.<\/p>\n
Sea level during the peak of the last glacial maximum \u2014 roughly 20,000 years ago \u2014 was approximately 350 feet lower than today.<\/p>\n
With sea level that far below its present position, vast areas of the continental shelf that are now submerged were dry land.<\/p>\n
The geography of coastal California looked fundamentally different from how it looks today.<\/p>\n
The four northern Channel Islands \u2014 San Miguel, Santa Rosa, Santa Cruz, and Anacapa \u2014 were connected.<\/p>\n
<\/p>\n
Instead of four separate islands, they formed a single large landmass.<\/p>\n
Researchers call this ancient superisland Santarosae.<\/p>\n
It was roughly 125 kilometers long and at its maximum extent substantially larger than any of today’s individual Channel Islands.<\/p>\n
It had peaks where today’s islands have peaks, canyons where today’s islands have canyons, and a coastline that extended across areas now submerged.<\/p>\n
By 13,000 years ago \u2014 the approximate date of the Arlington Springs individual \u2014 sea level had risen from its glacial-maximum low but had not yet reached the modern position.<\/p>\n
The channel separating Santarosae from the California mainland was narrower than today.<\/p>\n
<\/p>\n
But it existed.<\/p>\n
The channel has always existed.<\/p>\n
There is no period within the last two million years \u2014 no time when any human being was alive anywhere on Earth \u2014 during which Santarosae was connected to the mainland by dry land.<\/p>\n
The minimum width of that channel during the terminal Pleistocene, when sea level was lower, has been estimated at roughly 4.5 to 7 miles \u2014 a body of open ocean, even at its narrowest.<\/p>\n
More recent reconstructions using high-resolution bathymetric and sediment data have refined the estimates, but the basic conclusion is unchanged: reaching Santarosae required crossing open water.<\/p>\n
<\/p>\n
This is the geographic fact that makes the Arlington Springs individual significant beyond what mere antiquity would convey.<\/p>\n
A 13,000-year-old human being anywhere on the mainland of North America would be extraordinary.<\/p>\n
A 13,000-year-old human being on an island that required an ocean crossing to reach is something more: evidence that maritime technology was already in the human toolkit at the end of the Ice Age, on the Pacific coast of North America.<\/p>\n
The island was also, in 13,000 BCE, a different place ecologically from the scrubby, relatively dry terrain of today.<\/p>\n
The late Pleistocene climate was cooler and wetter.<\/p>\n
<\/p>\n
The island supported forests, wetlands, and the fresh water systems implied by the very name Arlington Springs.<\/p>\n
The channel between Santarosae and the mainland, despite being narrower than today, was still a rich marine environment: kelp forests grew in the cold upwelling water, and the marine mammals, fish, and shellfish that depend on kelp-forest ecosystems were abundant.<\/p>\n
The person who reached the island 13,000 years ago arrived in a landscape of considerable ecological richness, separated from the mainland by a minimum of miles of open water, accessible only to someone who already knew how to cross open water.<\/p>\n
The Pygmy Mammoths And The Late Pleistocene Context<\/p>\n
<\/p>\n
A brief note on the other remarkable animals of Santarosae in the terminal Pleistocene \u2014 the pygmy mammoths \u2014 because they appear in the same general timeframe and the same general landscape as the Arlington Springs individual, and their presence illuminates the environment that person entered.<\/p>\n
Pygmy mammoths evolved on the Channel Islands from the Columbian mammoths that reached the islands during periods of lower sea level.<\/p>\n
Isolated on the island, cut off from mainland gene flow, the mammoths’ size decreased over generations \u2014 the island rule, the ecological tendency for large species to shrink when island resources are limited.<\/p>\n
The pygmy mammoths of the Channel Islands were substantially smaller than their mainland relatives: roughly the size of modern Asian elephants rather than the massive proportions of Columbian mammoths.<\/p>\n
These animals were alive on Santarosae during the terminal Pleistocene, overlapping in time with the earliest documented human presence.<\/p>\n
<\/p>\n
What interaction, if any, the person buried at Arlington Springs had with these animals is not known.<\/p>\n
The Orr 1962 paper specifically stated that no artifacts were directly associated with the oldest bones \u2014 no spear points, no evidence of hunting or butchering \u2014 and that the burial context appeared to be accidental, a depositional event associated with the sediments near a cienaga, a wetland or marshy area.<\/p>\n
The relationship between the first island humans and the pygmy mammoths is a question the evidence at Arlington Springs alone cannot answer.<\/p>\n
The mammoths eventually went extinct, as did the mainland megafauna, at the end of the Pleistocene.<\/p>\n
<\/p>\n
Whether hunting pressure from humans contributed to that extinction on the islands is debated and requires evidence beyond what the Arlington Springs site provides.<\/p>\n
The mammoths are mentioned here not because the Arlington Springs individual hunted them \u2014 that is undemonstrated \u2014 but because they share the same landscape and the same general period, and their presence tells us something about the biological richness of the island that the first humans entered.<\/p>\n
Why Early Radiocarbon Dating Had Limits<\/p>\n
The decision to preserve the Arlington Springs femora for future analysis rather than consuming them in tests that were not yet adequate was, in retrospect, one of the more consequential archival decisions in the history of American archaeology.<\/p>\n
To understand why, it is necessary to understand what radiocarbon dating can and cannot do, and what the science looked like in 1959 and 1962 compared to what it became.<\/p>\n
<\/p>\n
Radiocarbon dating \u2014 more formally, carbon-14 dating \u2014 works because all living organisms incorporate a specific radioactive isotope of carbon, carbon-14, from the atmosphere during their lifetimes.<\/p>\n
Carbon-14 is produced when cosmic rays strike nitrogen atoms in the upper atmosphere, and it circulates through the carbon cycle \u2014 from the atmosphere to plants through photosynthesis, from plants to animals through eating, and so through the entire food web.<\/p>\n
As long as an organism is alive, the ratio of carbon-14 to stable carbon-12 in its tissues reflects the atmospheric ratio of those isotopes.<\/p>\n
When the organism dies, it stops incorporating new carbon.<\/p>\n
<\/p>\n
The carbon-14 already present begins to decay at a known rate: the half-life of carbon-14 is approximately 5,730 years.<\/p>\n
By measuring the ratio of carbon-14 to carbon-12 in an ancient sample and comparing it to the ratio in a living organism, researchers can calculate how much time has elapsed since the sample’s carbon was last renewed \u2014 since the organism died.<\/p>\n
The elegant simplicity of the method masks several complications that became apparent as the science matured.<\/p>\n
The first complication is calibration.<\/p>\n
Carbon-14 production in the atmosphere is not perfectly constant over time \u2014 it varies with solar activity and other factors.<\/p>\n
<\/p>\n
This means that a specific carbon-14 age, expressed in radiocarbon years before present, does not correspond directly to a specific calendar age.<\/p>\n
The conversion from radiocarbon years to calendar years requires a calibration curve: a table or graph that accounts for atmospheric carbon-14 variation over time, built from samples of known age \u2014 tree rings being the most important \u2014 whose carbon-14 content can be measured directly.<\/p>\n
The calibration curves available in 1962 were primitive compared to what exists today.<\/p>\n
The conversion from radiocarbon years to calendar years could not be done with the precision that later science would achieve.<\/p>\n
A date expressed as “approximately 10,000 radiocarbon years BP” translated to a calendar age with substantial uncertainty.<\/p>\n
<\/p>\n
The second complication is contamination.<\/p>\n
Ancient bone is porous.<\/p>\n
Over thousands of years, it can absorb carbon from its surrounding environment \u2014 carbon from groundwater, from the decomposition of surrounding organic material, from mineral sources.<\/p>\n
This absorbed carbon, if it is younger than the bone itself, will make the bone appear younger than it actually is.<\/p>\n
If it is older, the bone will appear older.<\/p>\n
Good radiocarbon dating of bone requires chemical preparation \u2014 careful removal of the contaminant carbon to leave only the original biological material.<\/p>\n
In 1959 and 1962, the chemical preparation methods available for bone were not reliable enough for the kind of very ancient, fragmentary bone represented by the Arlington Springs femora.<\/p>\n
<\/p>\n
The bones were too important and too rare to subject to procedures that might produce an unreliable number.<\/p>\n
A third issue involved which part of the bone was being dated.<\/p>\n
Bone contains multiple types of organic and inorganic material.<\/p>\n
The most reliable for radiocarbon dating is collagen \u2014 the protein that forms the structural scaffold of bone.<\/p>\n
But extracting clean collagen from very old bone requires chemical processing that was beyond routine practice in the early 1960s.<\/p>\n
Collagen degrades over time, and ancient bone may contain only a fraction of its original collagen content, mixed with diagenetic products \u2014 materials produced by the chemical alteration of the bone in the burial environment.<\/p>\n
Separating what was originally there from what accumulated afterward requires methods that were developed later.<\/p>\n
<\/p>\n
Orr understood these limitations.<\/p>\n
He preserved the bones, published what he could responsibly claim, and waited.<\/p>\n
The Bones In Museum Drawers<\/p>\n
For nearly thirty years, the Arlington Springs femora sat in storage at the Santa Barbara Museum of Natural History.<\/p>\n
This was not neglect; it was the appropriate scientific practice of an era when the concept of leaving samples for future improved methods had become recognized as a strategy of genuine value.<\/p>\n
The idea that material should be held back for future technology had been gaining currency in archaeological circles through the 1960s and 1970s, as the speed of methodological improvement in the dating sciences became apparent.<\/p>\n
<\/p>\n
What accelerator mass spectrometry could do, and what improvements in bone chemistry could achieve, were not fully clear in the 1960s \u2014 but researchers were beginning to understand that the science was not static.<\/p>\n
AMS radiocarbon dating \u2014 accelerator mass spectrometry \u2014 was the key advance.<\/p>\n
Conventional radiocarbon dating measures the radioactive decay of carbon-14 atoms by counting the particles emitted as they decay.<\/p>\n
This requires a relatively large sample \u2014 typically several grams of carbon \u2014 and produces a measurement based on the statistical rate of decay over a counting period.<\/p>\n
AMS dating, by contrast, measures the carbon-14 atoms directly: it uses a particle accelerator to separate and count the actual atoms, rather than waiting for them to decay.<\/p>\n
This produces two major advantages.<\/p>\n
<\/p>\n
First, it requires a much smaller sample \u2014 a few milligrams of carbon rather than several grams.<\/p>\n
Second, it is substantially more precise for very old samples, because the uncertainty of direct counting is lower than the uncertainty of decay-rate measurement.<\/p>\n
For very old, very rare, fragmentary human bone \u2014 exactly the situation at Arlington Springs \u2014 AMS dating was transformative.<\/p>\n
It meant that the bones could be tested without consuming more than a tiny fraction of their material.<\/p>\n
The smaller sample requirement opened the possibility of dating material that would have been too small or too precious for conventional radiocarbon.<\/p>\n
<\/p>\n
By the late 1980s and early 1990s, AMS dating had become a practical tool in archaeological research.<\/p>\n
At the same time, bone-protein chemistry had advanced substantially.<\/p>\n
The methods for extracting, purifying, and characterizing ancient collagen \u2014 the most reliable organic fraction of bone for radiocarbon purposes \u2014 had improved to the point where researchers could produce more reliable dates from ancient bone than had been possible in the 1960s.<\/p>\n
The Arlington Springs bones were waiting for exactly these advances.<\/p>\n
What Bone Chemistry Had To Solve<\/p>\n
<\/p>\n
When John R.<\/p>\n
Johnson, by then the curator of anthropology at the Santa Barbara Museum of Natural History and Orr’s eventual successor in that role, and colleagues including the geochronologist Thomas Stafford Jr. turned their attention to the Arlington Springs femora in the late 1980s and early 1990s, they faced a chemical puzzle.<\/p>\n
Ancient bone is not simply frozen organic material.<\/p>\n
Over thousands of years in the burial environment, the original bone undergoes chemical changes.<\/p>\n
Proteins are lost \u2014 some more quickly than others, depending on their specific chemical properties and the conditions of the burial environment.<\/p>\n
<\/p>\n
Contaminants accumulate.<\/p>\n
The resulting material is a complex mixture of original biological compounds and diagenetic products that must be separated, characterized, and selectively dated.<\/p>\n
The two protein fractions that researchers used for the Arlington Springs bone were collagen and osteocalcin.<\/p>\n
Collagen is the most abundant protein in bone \u2014 the structural framework that, when mineralized by calcium phosphate crystals, gives bone its strength.<\/p>\n
It is a large, stable protein that, under favorable conditions, can survive in chemically recognizable form for tens of thousands of years.<\/p>\n
<\/p>\n
The key word is favorable: collagen preservation depends strongly on temperature, pH, moisture, and microbial activity.<\/p>\n
Cold, dry, neutral pH conditions preserve collagen well; warm, wet, acidic conditions degrade it rapidly.<\/p>\n
California island sediments are not an ideal collagen-preservation environment \u2014 the temperatures are moderate and the sediments are well-watered \u2014 but the specific conditions at Arlington Springs apparently allowed some collagen to survive.<\/p>\n
Osteocalcin is a different protein \u2014 a smaller molecule, specific to bone and dentin, involved in the mineral metabolism of living bone.<\/p>\n
Because of its small size and specific chemistry, osteocalcin tends to be more susceptible to degradation than collagen.<\/p>\n
<\/p>\n
It also has a different relationship with contaminating materials.<\/p>\n
The problem that emerged when both proteins were dated from the Arlington Springs bone was striking: the ages disagreed.<\/p>\n
By thousands of years.<\/p>\n
The osteocalcin date, extracted using formic acid, came in at 6,610 \u00b1 60 radiocarbon years BP.<\/p>\n
The collagen date, extracted using XAD purification \u2014 a more refined chemical method that removes contaminants more thoroughly \u2014 came in at 10,960 \u00b1 80 radiocarbon years BP.<\/p>\n
Two different proteins from the same bones, dated by the same method, producing results that differed by more than four thousand radiocarbon years.<\/p>\n
<\/p>\n
Which one was right?<\/p>\n
Why The Collagen Won \u2014 And What That Number Meant<\/p>\n
The discordance between the osteocalcin and collagen dates from Arlington Springs was not unique to this site.<\/p>\n
It was a problem that the broader field of ancient bone dating had been grappling with, and the resolution of that problem had implications far beyond any single site.<\/p>\n
The shorter osteocalcin age \u2014 roughly 6,600 radiocarbon years, meaning the bones would be less old than even Orr’s original estimate \u2014 was rejected on chemical grounds.<\/p>\n
Osteocalcin, because of its small molecular size and specific chemistry, is more susceptible than collagen to contamination by younger carbon.<\/p>\n
<\/p>\n
It interacts with humic acids and other organic compounds from the soil environment in ways that can introduce modern or near-modern carbon into the sample, making it appear younger than it actually is.<\/p>\n
Formic acid extraction, the method used for the osteocalcin fraction, does not remove this contamination effectively.<\/p>\n
The collagen fraction, extracted using the more demanding XAD purification method, was a different matter.<\/p>\n
XAD purification \u2014 named for the Amberlite XAD resins used in the process \u2014 removes a wider range of contaminants, including the humic acids that cause problems with osteocalcin.<\/p>\n
The collagen that survived this process was more likely to represent the original biological material.<\/p>\n
The researchers also examined the amino acid composition of the collagen \u2014 its molecular building blocks \u2014 and compared it to what was expected from well-preserved ancient collagen of known characteristics.<\/p>\n
<\/p>\n
The results suggested that the XAD-purified collagen retained characteristics consistent with genuine ancient collagen rather than a contaminated or extensively altered product.<\/p>\n
To further support the collagen date, the researchers looked at other materials from the same stratum that had yielded the human bones.<\/p>\n
Charcoal from the stratum produced AMS dates in the range of approximately 10,000 to 11,500 radiocarbon years BP.<\/p>\n
Rodent bones from the same stratum produced similar dates.<\/p>\n
The cluster of dates from multiple independent sources \u2014 human bone collagen, charcoal, rodent bone \u2014 all converging in the range of roughly 10,000 to 11,500 radiocarbon years BP was compelling.<\/p>\n
The internal consistency of the dataset supported the conclusion that the human bones belonged to this age range, not to the younger age implied by the osteocalcin fraction.<\/p>\n
<\/p>\n
The key date \u2014 10,960 \u00b1 80 radiocarbon years BP, from XAD-purified collagen \u2014 became the central age estimate for the Arlington Springs individual.<\/p>\n
But 10,960 radiocarbon years BP is not 10,960 calendar years ago.<\/p>\n
From Radiocarbon Years To Calendar Years \u2014 Why The Difference Matters<\/p>\n
The conversion from radiocarbon years to calendar years is one of the sources of apparent confusion in reporting on ancient sites, and the Arlington Springs case illustrates why it matters.<\/p>\n
The problem is that carbon-14 production in the atmosphere has not been constant over time.<\/p>\n
It varies with solar activity, with changes in Earth’s magnetic field, and with other factors.<\/p>\n
<\/p>\n
When carbon-14 production was higher than average, organisms incorporated more carbon-14 into their tissues, which means that when those organisms are dated, their carbon-14 content looks like that of a younger organism \u2014 one that died more recently.<\/p>\n
Conversely, when carbon-14 production was lower, organisms appear older than they are.<\/p>\n
The calibration curves that correct for this variation are built from samples of known calendar age.<\/p>\n
Tree rings are the most important source: trees produce one ring per year, and the carbon-14 content of each ring can be measured independently, creating a year-by-year record of atmospheric carbon-14 variation extending back tens of thousands of years.<\/p>\n
Corals, cave speleothems, and other materials with independent calendar-year chronologies supplement the tree-ring record for periods beyond the range of the longest tree-ring series.<\/p>\n
<\/p>\n
Around the terminal Pleistocene \u2014 the period relevant to Arlington Springs \u2014 the calibration curve contains a region of significant divergence between radiocarbon years and calendar years.<\/p>\n
Carbon-14 production was different during and immediately after the last glacial maximum, and the result is that radiocarbon ages from this period translate to calendar ages that are substantially older than the radiocarbon age in years.<\/p>\n
A radiocarbon age of 10,960 \u00b1 80 years BP, calibrated using the standard calibration curves, translates to a calendar age of approximately 12,710 to 13,010 calibrated years before present \u2014 roughly 13,000 calendar years ago.<\/p>\n
The National Park Service’s Channel Islands Park timeline records the Arlington Springs occupation at 12,900 BP.<\/p>\n
<\/p>\n
The difference is significant: roughly 2,000 calendar years separate the raw radiocarbon date from the calibrated calendar date.<\/p>\n
Thirteen thousand years ago, not eleven thousand.<\/p>\n
The terminal Pleistocene, not the early Holocene.<\/p>\n
The last Ice Age, not its aftermath.<\/p>\n
In the context of North American prehistory, those 2,000 years matter enormously.<\/p>\n
They push the Arlington Springs individual back before the traditional Clovis chronological horizon \u2014 the period to which the most dramatic early-American debate had been attached \u2014 and into a period when, under the old Clovis-first model, there should not have been anyone anywhere near the California coast.<\/p>\n
What Later Geology Confirmed<\/p>\n
<\/p>\n
The radiocarbon dating of the bone proteins and the associated materials produced a strong case for the terminal Pleistocene age of the Arlington Springs individual.<\/p>\n
But radiocarbon dates alone do not constitute the full evidentiary picture, particularly for a site as geologically complex as Arlington Canyon.<\/p>\n
Returning to the canyon to understand the original depositional context of the bones was essential.<\/p>\n
The bones themselves had been removed in 1959-60 and were in museum storage.<\/p>\n
The stratum from which they had eroded was what remained in the field, and that stratum needed to be precisely relocated, mapped, and dated independently.<\/p>\n
This work involved collaboration between Johnson and a team of geologists who conducted detailed chronostratigraphic work at Arlington Springs \u2014 re-identifying the specific sedimentary unit that contained the bones by its lithological characteristics and its relationship to overlying and underlying strata, and independently dating materials from that unit using both radiocarbon and other methods.<\/p>\n
<\/p>\n
The results of the geological work corroborated the bone dates.<\/p>\n
The stratum from which the femora had eroded \u2014 identifiable by its position and its sedimentological characteristics \u2014 consistently produced dates in the range established by the bone collagen: around 10,000 to 11,500 radiocarbon years BP, calibrating to the terminal Pleistocene.<\/p>\n
The geological and stratigraphic evidence confirmed that the bones had been deposited in the late Pleistocene, not in some later period from which they had somehow sunk down or been reworked.<\/p>\n
The convergence of independent lines of evidence \u2014 bone protein chemistry, charcoal radiocarbon, rodent bone dates, and independent geological chronostratigraphy \u2014 produced what researchers describe as a robust age determination.<\/p>\n
<\/p>\n
No single method, applied in isolation, could have achieved this confidence.<\/p>\n
Each line of evidence constrained and confirmed the others.<\/p>\n
The Arlington Springs individual was not a 10,000-year-old person whose bones had been reworked into older sediments.<\/p>\n
He was a person who had lived, died, and been incorporated into those sediments at the end of the Pleistocene, approximately 13,000 calendar years ago.<\/p>\n
And he was on an island.<\/p>\n
The Water That Had To Be Crossed<\/p>\n
The geographic fact that gives the Arlington Springs age its continental significance has been mentioned already, but it requires fuller treatment.<\/p>\n
<\/p>\n
Santarosae \u2014 the joined superisland of the northern Channel Islands \u2014 was separated from the California mainland during the terminal Pleistocene by the Santa Barbara Channel.<\/p>\n
The Channel is a genuine piece of ocean, not a sheltered harbor or a river crossing.<\/p>\n
Its waters are cold, driven by upwelling from deeper Pacific depths, and traversed by currents.<\/p>\n
Even at its minimum width during the terminal Pleistocene \u2014 when sea level was substantially lower and more of the continental shelf was exposed \u2014 the crossing from mainland to island required navigating several miles of open water.<\/p>\n
The specific minimum distance has been the subject of detailed research.<\/p>\n
<\/p>\n
Estimates using sea-level curves and bathymetric data have placed the minimum Pleistocene separation at roughly four to seven miles at its narrowest point.<\/p>\n
More recent reconstructions using higher-resolution acoustic data of the seafloor sediment stratigraphy have refined these estimates, but the basic conclusion is unchanged: there was always water.<\/p>\n
The crossing was always a genuine maritime challenge.<\/p>\n
No rafting event \u2014 no accidental crossing on a floating log or debris field \u2014 can reliably deposit a person on an island miles offshore in open Pacific waters.<\/p>\n
Accidental ocean crossings of this type do occur, but they are survivable only by chance in very specific circumstances, and they cannot be the basis for a population established enough to leave remains at documented sites.<\/p>\n
<\/p>\n
The Arlington Springs individual was not the only person ever to reach Santarosae.<\/p>\n
The island shows evidence of multiple early occupations and of resource use extending back to the same general terminal Pleistocene period.<\/p>\n
What this tells us is that people were crossing to Santarosae with sufficient regularity that some of them died there and were buried there.<\/p>\n
That regularity implies capability.<\/p>\n
It implies not accident but intention.<\/p>\n
It implies watercraft \u2014 some form of vessel adequate for open-water crossing, operated by people who understood the crossing and could make it survive.<\/p>\n
What kind of watercraft?<\/p>\n
<\/p>\n
This is where honesty about the limits of the evidence is essential.<\/p>\n
The Chumash people of the historical period \u2014 the Indigenous inhabitants of the Channel Islands and mainland Santa Barbara coast who were documented extensively by Spanish missionaries and explorers beginning in the late eighteenth century \u2014 built and used a remarkable vessel called the tomol.<\/p>\n
The tomol was a plank canoe, constructed from boards split from driftwood and stitched together with cord sealed with pine tar, capable of carrying several people and managing the channel crossing efficiently.<\/p>\n
The tomol was a technologically sophisticated craft that represented centuries or millennia of refinement in design and construction.<\/p>\n
<\/p>\n
It would be tempting to project the tomol backward 13,000 years, to imagine the Arlington Springs individual crossing the channel in something like it.<\/p>\n
This projection is not supported by any evidence.<\/p>\n
The historical tomol is documented from the eighteenth century onward; the archaeological evidence for plank canoes extends back into the prehistoric period in the Channel Islands region, but not to 13,000 years before present.<\/p>\n
The vessels used by the terminal Pleistocene inhabitants of the coast \u2014 if they were watercraft as opposed to simple rafts \u2014 are entirely unknown in their specific design.<\/p>\n
What can be said is that some form of water transport was required.<\/p>\n
<\/p>\n
What form it took \u2014 whether simple bundled-reed craft, dugout logs, skin boats, or something else entirely \u2014 is beyond the reach of the evidence.<\/p>\n
The ocean crossing is required by logic; the vessel that made it is invisible to archaeology.<\/p>\n
The Second Opening \u2014 What An Island Crossing Means For The History Of A Continent<\/p>\n
By approximately the late 1990s and early 2000s, when the revised Arlington Springs dates had been published and the geological work had confirmed the terminal Pleistocene context, the implications for the broader question of how the Americas were first populated had become a subject of intense discussion.<\/p>\n
The traditional model of the peopling of the Americas \u2014 Clovis First \u2014 had been the dominant framework for most of the twentieth century.<\/p>\n
<\/p>\n
In this model, the first Americans were the makers of the Clovis projectile points, who entered North America from northeast Asia across the Bering Land Bridge sometime around 13,000 to 14,000 years ago and spread rapidly through the interior of the continent, following the herbivore herds of the late Pleistocene.<\/p>\n
This model implied that the Pacific coast of North America was not the route \u2014 that interior corridors through the retreating ice sheets were where the first Americans moved.<\/p>\n
By the time the Arlington Springs revision was published, the Clovis First model was already under pressure from multiple directions.<\/p>\n
Monte Verde, a site in southern Chile dated to approximately 14,500 to 14,800 years ago \u2014 older than any Clovis site \u2014 had been exhaustively verified after a long period of controversy and had forced the conclusion that people were in the southern tip of South America before Clovis appeared in North America.<\/p>\n
If people were in southern Chile by 14,500 years ago, they could not have arrived via an interior route through North America after Clovis appeared around 13,000 years ago \u2014 the dates did not permit enough time for the population to reach Chile.<\/p>\n
<\/p>\n
Some prior migration, by some route, had to be invoked.<\/p>\n
The alternative route that Monte Verde and other pre-Clovis evidence pointed toward was the coast.<\/p>\n
If people had moved south along the Pacific coast of North America, following marine resources, they would have been able to move quickly \u2014 marine resources are relatively uniform along a coastline, reducing the need for adaptation to dramatically new environments.<\/p>\n
They would also have left evidence that is now largely invisible: the coastline of 14,000 to 15,000 years ago is now drowned under the risen sea that followed the end of the Ice Age.<\/p>\n
The archaeological sites of the first coastal migrants, if any existed, are on the seafloor.<\/p>\n
<\/p>\n
The Kelp Highway hypothesis \u2014 formally proposed by archaeologist Jon Erlandson and colleagues \u2014 put ecological flesh on this geographical argument.<\/p>\n
The Pacific Rim kelp forests, which stretch from northeast Asia down through the Aleutians and Alaska and along the North American coast to Baja California, form a relatively continuous ecological zone of high biological productivity.<\/p>\n
Marine mammals, fish, shellfish, and seabirds concentrated in kelp forests provide a rich and predictable food base.<\/p>\n
A maritime people skilled in exploiting kelp-forest resources could, in principle, move along this coast from northeast Asia toward the Americas while living off the sea’s produce along the way.<\/p>\n
In this context, the Arlington Springs individual’s presence on an island off the California coast approximately 13,000 years ago was exactly what a Kelp Highway migration would predict: people already skilled in maritime movement, already comfortable crossing open water, already harvesting marine resources, already established along the California coast at the close of the Ice Age.<\/p>\n
<\/p>\n
The Arlington Springs find did not prove the Kelp Highway model.<\/p>\n
No single discovery can prove a migration theory.<\/p>\n
But it was consistent with the model in every relevant way, and it demonstrated \u2014 empirically, not theoretically \u2014 that maritime capability was present on the California coast at approximately the time that coastal migration models proposed.<\/p>\n
Before The Old Story Said They Should Be There<\/p>\n
To appreciate what the Arlington Springs date of approximately 13,000 calendar years meant in the history of the Clovis debate, it helps to understand what the old story actually said.<\/p>\n
The Clovis First model, at its most restrictive, argued not merely that Clovis people were the first Americans, but that no one could have reached North America significantly before Clovis because the ice-free corridor \u2014 the route through the Canadian interior between the retreating Cordilleran and Laurentide ice sheets \u2014 was not open until approximately 13,000 to 14,000 years ago.<\/p>\n
<\/p>\n
Before that corridor opened, the argument went, the Americas were effectively cut off from Asia by the continental ice.<\/p>\n
This argument had a specific implication for the Pacific coast: even if someone had wanted to travel down the coast, the coast was blocked by glacial ice extending to the shoreline along much of British Columbia and Alaska.<\/p>\n
There was no viable coastal route until the ice retreated enough to open a passage.<\/p>\n
As evidence for earlier occupation accumulated, defenders of the Clovis First framework offered various responses.<\/p>\n
Monte Verde’s dates were disputed, then accepted.<\/p>\n
<\/p>\n
The coastal ice-blockage argument was revisited as geological evidence suggested that ice-free refugia may have existed along parts of the coast earlier than previously thought \u2014 that the coastal route was open, or intermittently open, earlier than the interior corridor.<\/p>\n
Against this backdrop, the Arlington Springs revision placed a person on a California island approximately 13,000 years ago \u2014 near the boundary of the Clovis horizon, but on the island that required maritime capability, and in a location consistent with a coastal rather than an interior migration route.<\/p>\n
A person arriving in California by a coastal route from the north would, by necessity, have maritime technology.<\/p>\n
If that person or their immediate ancestors had traveled down the Pacific coast following marine resources, they would be exactly the kind of people who could cross a channel to an offshore island.<\/p>\n
<\/p>\n
The island presence was not just consistent with coastal migration \u2014 it was most economically explained by it.<\/p>\n
Collagen, AMS, And The Science Behind The Revision<\/p>\n
The technical breakthrough that made the revised dating possible \u2014 AMS radiocarbon dating combined with improved bone-protein chemistry \u2014 deserves a more detailed explanation than the overview provided earlier, because the method’s logic is the foundation of the entire revised age claim.<\/p>\n
The problem of dating very ancient bone is a problem of separating what is original from what is contamination.<\/p>\n
Over thousands of years, bone in the burial environment accumulates material from its surroundings.<\/p>\n
<\/p>\n
Groundwater carries dissolved organic and inorganic compounds that can bind to the bone matrix.<\/p>\n
Microbial activity produces organic acids and other compounds that become incorporated into the bone.<\/p>\n
Root action introduces plant-derived carbon.<\/p>\n
None of these additions is the original carbon of the person \u2014 they are the carbon of the burial environment, ranging in age from very old to very recent.<\/p>\n
If a batch of ancient bone is radiocarbon dated without careful removal of these contaminants, the result will reflect a mixture of ages \u2014 the original biological carbon of the individual plus the accumulated environmental carbon.<\/p>\n
Depending on the nature and age of the contaminants, this can make the sample appear either older or younger than it actually is.<\/p>\n
<\/p>\n
The XAD purification method used by Stafford and colleagues at Arlington Springs addressed this problem more thoroughly than earlier methods.<\/p>\n
The basic approach involves dissolving the mineral portion of the bone (demineralization), isolating the protein fraction (which includes collagen), and then running the protein solution through columns packed with XAD resin.<\/p>\n
The XAD resin preferentially binds humic acids and other contaminants that would affect the radiocarbon age, while the purified collagen passes through.<\/p>\n
The purified collagen was then characterized chemically \u2014 its amino acid composition was checked to verify that it represented genuine ancient collagen rather than a diagenetically altered product that might not reliably date the original bone.<\/p>\n
The amino acid profile of the Arlington Springs collagen was consistent with ancient collagen that had been preserved but not extensively altered.<\/p>\n
<\/p>\n
With the contamination problem addressed and the authenticity of the collagen confirmed, the AMS measurement itself was made on a tiny fraction of the purified material \u2014 just a few milligrams of carbon, converted to carbon dioxide and then to graphite, and measured in the particle accelerator.<\/p>\n
The result: 10,960 \u00b1 80 radiocarbon years before present.<\/p>\n
This date, calibrated to calendar years, places the Arlington Springs individual at approximately 12,710 to 13,010 calendar years before present \u2014 the terminal Pleistocene, as the geological context independently confirmed.<\/p>\n
The Sex Assessment \u2014 A Long And Careful Revision<\/p>\n
There is a chapter of the Arlington Springs story that has sometimes been reported in sensationalized terms \u2014 as a dramatic revelation that “he was actually a she” \u2014 but which is more accurately understood as a lesson in the difficulty of biological sex assessment from fragmentary remains, and in how science revises its conclusions as evidence accumulates.<\/p>\n
Phil Orr, when he found the femora in 1959 and published his account in 1962, called the remains those of a man.<\/p>\n
<\/p>\n
The paper that introduced these bones to the scientific world was titled “Arlington Springs Man.” The size of the femora, as measured by Orr, was consistent with a male individual.<\/p>\n
When John Johnson and his colleague Don Morris re-examined the remains in 1989, they reassessed the sex determination and concluded that the bones were more likely those of a woman.<\/p>\n
At approximately the same time that the radiocarbon revision was being developed, the name “Arlington Springs Woman” was adopted and used for the individual in several publications and presentations.<\/p>\n
The 1999 presentations at the Fifth California Islands Symposium used the “Woman” designation.<\/p>\n
The basis for this reassessment was a technical comparison: Johnson and Morris compared certain measurements of the Arlington Springs femora to a reference population of known sex from another Channel Islands site.<\/p>\n
<\/p>\n
The comparison suggested the bones were more consistent with a female individual than a male.<\/p>\n
Then, in 2006, Johnson reversed the sex determination again.<\/p>\n
Further analysis by physical anthropologist Patricia Lambert, using femoral measurements compared to a reference collection from Tecolote Point \u2014 another Santa Rosa Island site \u2014 produced results that Lambert described as suggesting the individual was more likely male than female.<\/p>\n
Lambert specifically acknowledged that the results were not definitive, but the balance of the femoral metric evidence leaned toward male.<\/p>\n
The reason for the uncertainty is intrinsic to the evidence.<\/p>\n
<\/p>\n
Sex determination from human skeletal remains is most reliable from the pelvis: the female pelvis has specific adaptations for childbirth that distinguish it from the male pelvis in measurable ways, and these distinctions can be assessed with relatively high accuracy from an intact pelvis.<\/p>\n
The skull offers secondary evidence: male skulls tend to be more robust, with more pronounced brow ridges and other features, though with significant population variation.<\/p>\n
The femur alone \u2014 particularly two femora without the femoral heads, which are the most diagnostic part of the bone for sex assessment \u2014 is one of the less reliable skeletal elements for biological sex determination.<\/p>\n
Femoral length and robusticity correlate with sex, but the correlation is statistical and probabilistic, not categorical.<\/p>\n
Populations vary in body size, and what is a large female femur in one population may overlap with what is a small male femur in another.<\/p>\n
<\/p>\n
The Arlington Springs femora, critically, were also incomplete.<\/p>\n
The femoral heads \u2014 the rounded upper ends of the thigh bones that fit into the hip socket and that carry important sex-diagnostic measurements \u2014 were apparently not present in the museum collection.<\/p>\n
Phil Orr had measured the maximum femoral head diameter and published it in his 1968 monograph, but the physical specimen was not available for direct remeasurement by later researchers.<\/p>\n
The result is a sex determination that has been revised twice and remains probabilistic rather than certain.<\/p>\n
<\/p>\n
“More likely male” is the current assessment, and “Arlington Springs Man” is the name that has returned to use.<\/p>\n
But this story of revision is not embarrassing or scandalous \u2014 it is the normal scientific process of updating conclusions when new analyses are made, particularly when the underlying evidence is inherently limited in what it can support.<\/p>\n
What the bones cannot tell us, directly and definitively, is whether the individual buried at Arlington Springs was male or female.<\/p>\n
The probability assessment leans one way, with appropriate uncertainty acknowledged.<\/p>\n
What The Bones Do Not Tell Us<\/p>\n
<\/p>\n
The limitations of the Arlington Springs evidence deserve explicit treatment, because the find has sometimes been described in ways that exceed what the physical record supports.<\/p>\n
There is no skull.<\/p>\n
No face, no cranial structure, no features that would allow any reconstruction of the individual’s appearance.<\/p>\n
Any artistic rendering of “Arlington Springs Man” is creative interpretation, not scientific recovery.<\/p>\n
There is no full skeleton.<\/p>\n
Two femora \u2014 and those not entirely complete \u2014 are the physical remains from which every conclusion must be drawn.<\/p>\n
The total of bone material is small.<\/p>\n
The biological profile that can be constructed from it \u2014 probable sex, approximate stature, general robusticity \u2014 is limited by the fragmentary nature of the evidence.<\/p>\n
<\/p>\n
There is no associated artifact assemblage from the original bones.<\/p>\n
Orr’s 1962 paper was explicit: no artifacts were directly associated with the oldest bones.<\/p>\n
The individual was not buried with tools, food, or ornaments that were preserved alongside the femora.<\/p>\n
The context was not a deliberate grave with burial goods; Orr interpreted it as an accidental burial in sediments near a cienaga \u2014 a marshy area.<\/p>\n
Small chert flakes have been recovered from the relevant archaeological stratum during later work at the site, but these are from the cultural layer generally, not from direct association with the femoral remains themselves.<\/p>\n
There is no ancient DNA from this individual.<\/p>\n
<\/p>\n
The story of Arlington Springs Man is explicitly not a DNA story.<\/p>\n
DNA preservation in warm, seasonally wet California sediments is poor, and no genetic material has been recovered from or associated with the Arlington Springs femora.<\/p>\n
This means there is no genomic information about the biological population to which this individual belonged, no assessment of genetic connections to later peoples, no molecular ancestry data of any kind.<\/p>\n
The science of the site is stratigraphic, geochronological, and bone-protein chemistry \u2014 not genetics.<\/p>\n
There is no evidence of what this person ate, beyond the general inference that people on a resource-rich island with access to both marine and terrestrial foods would have used those resources.<\/p>\n
<\/p>\n
There is no cause of death, no evidence of trauma or disease preserved in the available remains.<\/p>\n
There is no name, no language group affiliation, no social role, no personal history beyond the single fact of presence \u2014 of being on that island at that moment in time.<\/p>\n
What the bones do tell us is precise and significant: approximately when this person lived, approximately where, and that the geographic location required maritime capability to reach.<\/p>\n
Those facts are genuine, well-supported, and consequential.<\/p>\n
But they are also narrow facts.<\/p>\n
<\/p>\n
The temptation to expand them into a narrative biography of “the first Californian” or “a pioneering navigator” should be resisted.<\/p>\n
The individual was a person whose full life, in all its human specificity, is as invisible to us as the lives of almost everyone who has ever lived.<\/p>\n
The Kelp Highway \u2014 A Model Explained Carefully<\/p>\n
The coastal migration model \u2014 sometimes called the Kelp Highway hypothesis after one of its most discussed ecological formulations \u2014 deserves explanation on its own terms, because it is often conflated with the specific Arlington Springs evidence in ways that overstate what the bones can demonstrate.<\/p>\n
The Kelp Highway hypothesis, as developed primarily by Jon Erlandson and colleagues, proposes that maritime peoples migrated from northeast Asia to the Americas along the North Pacific coast during the late Pleistocene, following the productive kelp-forest ecosystems that form a nearly continuous chain from Japan and the Russian far east, along the Aleutian Islands, down the northwest coast of North America, and south to Baja California.<\/p>\n
<\/p>\n
The appeal of this model is ecological: kelp forests are among the most biologically productive marine ecosystems on Earth.<\/p>\n
They support enormous concentrations of fish, marine mammals, shellfish, and seabirds.<\/p>\n
A people skilled in harvesting these resources would have had a reliable food supply along the entire coast, reducing the need for the kind of radical ecological adaptation required when crossing from a coastal to an interior environment.<\/p>\n
The model predicts rapid southward movement once the coast was entered.<\/p>\n
The model also predicts that much of the archaeological evidence for this migration is now invisible.<\/p>\n
<\/p>\n
The coastlines of 13,000 to 15,000 years ago are now under water \u2014 drowned by the sea-level rise that followed the end of the glacial maximum.<\/p>\n
Any sites that early coastal migrants built at or near the ancient shoreline are on the continental shelf, covered by ocean.<\/p>\n
Without underwater archaeological investigation, that evidence is inaccessible.<\/p>\n
This is a critical point for understanding why the Arlington Springs individual is so significant within the Kelp Highway framework.<\/p>\n
The bones were found not at the ancient coastline \u2014 which is now submerged \u2014 but inland, in canyon sediments.<\/p>\n
They survived because they were deposited in a geological context that preserved them above the water table.<\/p>\n
<\/p>\n
Most of the evidence for terminal Pleistocene coastal occupation is gone.<\/p>\n
The Arlington Springs bones are among the survivors.<\/p>\n
The Kelp Highway model is supported by multiple lines of evidence beyond Arlington Springs: the early radiocarbon dates from Channel Islands sites, the early dates from coastal California, and the presence of humans at Monte Verde in southern Chile \u2014 which requires that people had reached the tip of South America before the Clovis horizon, arguing for an earlier, and plausibly coastal, migration.<\/p>\n
But the model remains a model \u2014 a well-supported hypothesis rather than a settled account.<\/p>\n
<\/p>\n
Its specific mechanisms, timing, and the exact nature of the watercraft and subsistence strategies involved remain subjects of active research and debate.<\/p>\n
Arlington Springs contributes one data point: by approximately 13,000 years ago, someone was on a California island that required crossing open water.<\/p>\n
This is consistent with \u2014 and most easily explained by \u2014 a maritime tradition already established on the Pacific coast.<\/p>\n
It does not, by itself, prove the Kelp Highway hypothesis.<\/p>\n
The Chronostratigraphy That Sealed The Case<\/p>\n
The geological fieldwork that returned to Arlington Canyon in the years after the initial bone-chemistry dating was essential for converting a good radiocarbon date into a fully defensible scientific conclusion.<\/p>\n
A radiocarbon date from a bone sample tells you when the bone protein last exchanged carbon with the atmosphere \u2014 when the animal died.<\/p>\n
<\/p>\n
But it does not, by itself, tell you whether the bone was deposited in the stratum where it was found at the time the animal died or whether it was reworked \u2014 moved by water, soil creep, or other processes into a stratum of different age.<\/p>\n
The possibility of reworking is a standard concern in archaeology and geology, and for a find as significant as Arlington Springs, it had to be addressed.<\/p>\n
The geological and chronostratigraphic work \u2014 led by Johnson in collaboration with specialists from San Diego State University and other institutions \u2014 involved relocating the specific sedimentary unit at Arlington Canyon from which the femora had eroded, mapping its lithological characteristics, and independently dating multiple materials from it and from overlying and underlying units.<\/p>\n
The character of the sediments was itself informative.<\/p>\n
<\/p>\n
The water-laid alluvial deposits in which the bones were found were consistent with deposition by the stream that drained through Arlington Canyon \u2014 not a disturbed or reworked context, but a primary depositional unit laid down in a specific geological event or sustained by ongoing stream activity.<\/p>\n
The depth below the modern surface \u2014 11 meters, or approximately 37 feet \u2014 reflected the accumulated sediment of subsequent millennia rather than unusual depth at the time of deposition.<\/p>\n
Multiple AMS radiocarbon dates from charcoal, plant material, and rodent bone from the relevant stratum and from adjacent strata produced a coherent chronological sequence.<\/p>\n
The stratum containing the human remains dated consistently to the range of 10,000 to 11,500 radiocarbon years BP \u2014 the terminal Pleistocene, calibrating to approximately 12,000 to 13,500 calendar years ago.<\/p>\n
Strata above it produced dates consistent with an early Holocene age.<\/p>\n
<\/p>\n
Strata below it, where dated, were older still.<\/p>\n
The sequence made geological sense.<\/p>\n
The bones had been deposited in terminal Pleistocene sediments at a time when Arlington Canyon’s stream system was actively transporting and depositing material.<\/p>\n
The individual had died near the stream \u2014 perhaps at the marshy area that Orr described as a cienaga \u2014 and had been incorporated into the accumulating sediment.<\/p>\n
Nothing about the physical context argued for reworking.<\/p>\n
The bones were in primary depositional context, in a stratigraphic unit whose age was independently confirmed by multiple dating methods.<\/p>\n
<\/p>\n
The case was as complete as fragmented terminal Pleistocene remains can reasonably be made.<\/p>\n
A Deeper Indigenous History Of The Channel Islands<\/p>\n
The Arlington Springs individual lived approximately 13,000 years ago.<\/p>\n
What happened on the Channel Islands in the following thirteen millennia is a separate, complex, and only partially known story \u2014 but it is the context within which the modern meaning of the Arlington Springs find must be understood.<\/p>\n
The Channel Islands were continuously inhabited after the terminal Pleistocene, with an archaeological record extending from the time of Arlington Springs through the Holocene and into the historical period.<\/p>\n
Sites on Santa Rosa Island, San Miguel Island, and Santa Cruz Island document increasing human presence and complexity over time: evidence for fishing technology, shellfish harvesting, seed grinding, bead manufacture, and eventually the emergence of the social complexity and material culture that characterized the Chumash of the historical period.<\/p>\n
<\/p>\n
The Chumash people, documented extensively from the late eighteenth century onward through Spanish mission records, explorers’ accounts, and the ethnographic work of researchers like John Peabody Harrington, inhabited the northern Channel Islands and the adjacent mainland coast.<\/p>\n
They spoke a family of related languages and maintained extensive maritime commerce and cultural exchange.<\/p>\n
The tomol, their remarkable plank canoe, connected the island communities to each other and to the mainland.<\/p>\n
Whether the Arlington Springs individual was a direct ancestor of the historical Chumash is not something the available evidence can determine.<\/p>\n
Thirteen thousand years of population history \u2014 of migrations, population changes, cultural transformations, genetic drift, and the catastrophic disruptions of European contact \u2014 separate the terminal Pleistocene from the historical period.<\/p>\n
The genetic and cultural connections between the earliest Channel Island inhabitants and the historical Chumash are not documented in any straightforward way.<\/p>\n
<\/p>\n
John Johnson, who spent his career at the Santa Barbara Museum of Natural History studying both the prehistoric and historical Chumash, has addressed this question with appropriate care.<\/p>\n
When asked whether Arlington Springs Man was an ancestor of the Chumash, his response was honest: “We don’t know for sure.<\/p>\n
Maybe future studies will determine this.” The answer acknowledged both the deep Indigenous history of the Channel Islands and the genuine uncertainty about specific ancestral connections across such an enormous time span.<\/p>\n
What can be said without reservation is that the Channel Islands have been home to Indigenous people for at least 13,000 years \u2014 from the time of the Arlington Springs individual to the Chumash of the historical period and beyond.<\/p>\n
The island has never been uninhabited since the first humans reached it.<\/p>\n
<\/p>\n
That continuity of presence is part of the meaning of the Arlington Springs find.<\/p>\n
Repatriation And Modern Stewardship<\/p>\n
The modern chapter of the Arlington Springs story involves the return of the remains to the community with established legal and cultural connections to the Channel Islands.<\/p>\n
In 2022, following a request under the Native American Graves Protection and Repatriation Act (NAGPRA), the Arlington Springs femora were repatriated to the Santa Ynez Band of Chumash Mission Indians.<\/p>\n
The remains were returned for reburial according to the tribe’s customs and traditions.<\/p>\n
<\/p>\n
The repatriation represented the appropriate application of the principles embedded in NAGPRA: that Native American human remains and associated cultural items held by federal agencies and federally funded institutions must be inventoried, that affiliated tribes must be notified and consulted, and that the remains must be returned upon tribal request.<\/p>\n
The cultural affiliation question \u2014 which is central to any NAGPRA repatriation \u2014 was determined in favor of the Chumash connection to the Channel Islands.<\/p>\n
While the specific biological ancestry of the Arlington Springs individual cannot be established by the available science, the long-established connection of Chumash people to the Northern Channel Islands and the Santa Barbara Channel region, documented archaeologically and historically over thousands of years, was sufficient to establish the cultural affiliation required by law.<\/p>\n
The repatriation was not a scientific loss.<\/p>\n
<\/p>\n
The scientific work that could responsibly be done with the Arlington Springs femora \u2014 the radiocarbon dating, the bone-protein chemistry, the morphological analysis, the chronostratigraphic context \u2014 had been completed.<\/p>\n
The bones had been measured, sampled, analyzed, and documented in the peer-reviewed literature.<\/p>\n
What remained was physical custody of remains whose scientific work was substantially done, and whose appropriate custodians were the Indigenous community with the deepest historical connection to the land where they were found.<\/p>\n
The Santa Ynez Band of Chumash Mission Indians received the remains in 2022 and proceeded with reburial according to their traditions.<\/p>\n
The scientific record of the site \u2014 the published papers, the data, the museum collections of associated materials \u2014 remains accessible to researchers.<\/p>\n
<\/p>\n
The person who died on that island 13,000 years ago was returned, in some sense, to the landscape to which they had come so long before.<\/p>\n
What A Single Find Cannot And Cannot Prove<\/p>\n
The Arlington Springs individual occupies an unusual position in North American prehistory: a single fragmentary find whose significance is disproportionate to the physical evidence it consists of.<\/p>\n
Two femoral fragments are a thin basis for continental claims.<\/p>\n
They cannot support identification of a specific cultural affiliation, a genetic lineage, a migration route, or a subsistence strategy.<\/p>\n
They tell us that one person was present on one island approximately 13,000 years ago and that the island required a water crossing to reach.<\/p>\n
<\/p>\n
That one fact \u2014 one person, one island, one crossing, one date \u2014 is enough to establish that maritime capability existed on the terminal Pleistocene California coast.<\/p>\n
It is enough to demonstrate that the Americas were being settled at the margins and on the islands, not only in the interior.<\/p>\n
It is consistent with, and supportive of, the coastal migration models that have largely replaced the interior-corridor-focused Clovis First model.<\/p>\n
But it is not proof of any complete migration theory.<\/p>\n
A single site cannot prove a continental-scale model.<\/p>\n
Theories of how the Americas were populated rest on accumulating evidence from many sites \u2014 Monte Verde, Paisley Caves in Oregon, the broader Channel Islands archaeological record, the Daisy Cave assemblage on San Miguel Island, sites in coastal Alaska and British Columbia, the genetic evidence from ancient and modern Indigenous populations.<\/p>\n
<\/p>\n
Arlington Springs is one piece of a much larger puzzle, and its role is to contribute evidence about a specific place and time, not to settle all debates.<\/p>\n
The debates, in fact, continue.<\/p>\n
The timing of the initial peopling of the Americas, the number of separate migration waves, the routes taken, and the relationships between the earliest known skeletal remains and later populations are all subjects of active research and genuine scientific disagreement.<\/p>\n
Ancient DNA science, which has transformed what is known about human population history elsewhere in the world, is producing new insights about the genetics of early Americans as more ancient individuals are analyzed.<\/p>\n
The specific place of terminal Pleistocene California in the broader story of American origins is still being worked out.<\/p>\n
<\/p>\n
In this context, the Arlington Springs individual remains what it always was: fragmentary, significant, and irreplaceable evidence of human presence at a specific point in space and time.<\/p>\n
The science built around it has been careful, rigorous, and appropriately hedged.<\/p>\n
The conclusions it supports are real but limited.<\/p>\n
The questions it raises remain partly open.<\/p>\n
The Migration Map Changes \u2014 And Keeps Changing<\/p>\n
The decades since the Arlington Springs revision have seen the old Clovis First model give way to a more complex picture of initial American settlement.<\/p>\n
The basic conclusion that the Americas were settled earlier than Clovis, and that the Pacific coast was likely a significant route of entry rather than merely a theoretical alternative to the interior, has been substantially supported by accumulating evidence.<\/p>\n
Monte Verde in Chile remains the most dramatic single site pointing to pre-Clovis occupation of the Americas.<\/p>\n
<\/p>\n
Paisley Caves in Oregon has produced coprolites \u2014 preserved feces \u2014 containing human mitochondrial DNA dated to approximately 14,300 calendar years before present: before Clovis, on the Pacific side of the continent.<\/p>\n
The genetics of ancient and modern Indigenous peoples show patterns consistent with multiple waves of migration, with the deepest ancestral lineages tracing to northeast Asian populations moving into the Americas during the late Pleistocene.<\/p>\n
The Channel Islands, in this evolving picture, are a significant node.<\/p>\n
The Arlington Springs individual represents the oldest well-dated human remains from the island context.<\/p>\n
But the broader Channel Islands archaeological record extends the story: Daisy Cave on San Miguel Island has produced evidence of human occupation from approximately 11,600 to 8,000 years ago, with shellfish exploitation, fishing, and early shell bead manufacture represented.<\/p>\n
Other sites on Santa Rosa Island document human presence from the early Holocene onward, with the record deepening and becoming richer as sea level stabilized and the islands took their modern form.<\/p>\n
<\/p>\n
The picture that emerges from the Channel Islands as a whole \u2014 from the terminal Pleistocene Arlington Springs individual through the early Holocene occupations at Daisy Cave and other sites to the documented Chumash of the historical period \u2014 is of an island landscape continuously inhabited, continuously exploited for marine resources, and continuously connected by maritime technology to the mainland coast.<\/p>\n
This is a picture of deep continuity: people on the California coast for at least 13,000 years, maintaining the maritime skills that made island living possible across the radical environmental changes of the post-glacial period.<\/p>\n
The Canyon And The Channel<\/p>\n
Return, finally, to the physical place.<\/p>\n
Arlington Canyon cuts through the interior of Santa Rosa Island from a spring source down to the coast.<\/p>\n
In the late Pleistocene, the canyon was a different world from today: wetter, possibly wooded in places, with the stream that built the alluvial deposits still active, the cienaga or marshy area at the canyon bottom providing fresh water and the biological richness that comes with permanent water in a semi-arid landscape.<\/p>\n
<\/p>\n
The person whose femora were found 37 feet below the modern surface was somewhere in this canyon when they died, or was washed into the depositional context by the stream after dying nearby.<\/p>\n
The bones were incorporated into the accumulating alluvium, covered by subsequent deposition, buried deeper and deeper as millennia passed, until erosion in the modern canyon wall brought them to the surface where Phil Orr could see them.<\/p>\n
From the canyon, if you climbed to the ridge above, you would look south and east and see water.<\/p>\n
The Santa Barbara Channel \u2014 the crossing that was required to get to this island \u2014 would be visible on a clear day as a band of blue between the island’s cliffs and the California mainland.<\/p>\n
The distance from here to there is navigable today by small boat in a few hours.<\/p>\n
<\/p>\n
Thirteen thousand years ago, in whatever vessel the first visitors used, it was navigated too.<\/p>\n
Someone looked at that water, understood it as a route rather than a barrier, and crossed it.<\/p>\n
They carried with them \u2014 in their knowledge, their technology, their community practice, their understanding of the sea \u2014 the maritime capability that the island required.<\/p>\n
That capability is what the Arlington Springs femora document, in the end.<\/p>\n
Not specifically this individual’s crossing \u2014 the bones cannot tell us the story of a particular day or a particular voyage.<\/p>\n
What they document is the consequence: that by the time those bones were deposited in Arlington Canyon’s accumulating sediment, human beings were already on this island.<\/p>\n
<\/p>\n
They had already crossed the channel.<\/p>\n
The crossing was already within the range of human capability and human practice.<\/p>\n
What the bones do not tell us is everything else.<\/p>\n
The name of the person.<\/p>\n
The names of the people who may have brought them here, or met them here, or buried them here \u2014 if burial was even what happened, rather than the accidental incorporation of remains into the cienaga’s sediments.<\/p>\n
The language they spoke.<\/p>\n
<\/p>\n
The boats they used.<\/p>\n
The specific route they traveled to reach this island.<\/p>\n
These absences are permanent.<\/p>\n
Thirteen thousand years of sediment, sea-level rise, and time have erased them.<\/p>\n
What survives is two femora, a suite of radiocarbon dates, a geological context, and the geographic fact that the island existed and the channel existed and the channel required crossing.<\/p>\n
A road was eventually built across the land where the Arlington Springs femora spent their long burial \u2014 but it is a dirt track used by the ranchers who worked the island before it became a national park, not a highway of continental significance.<\/p>\n
The significance lies underground, in the sediment, in the stratigraphy that preserves the record of what happened here at the end of the Ice Age.<\/p>\n
<\/p>\n
The canyon wall still stands.<\/p>\n
The spring still flows.<\/p>\n
The channel is still there, still blue, still separating the island from the mainland.<\/p>\n
And somewhere in that long geological record, between 11 meters of accumulated sediment and the present surface, is the stratum where two femoral bones were deposited 13,000 years ago \u2014 before the world we inhabit existed, before the sea rose to its present level, before the four modern Channel Islands separated from the single landmass of Santarosae, before any written record existed anywhere on Earth.<\/p>\n
Someone was there.<\/p>\n
<\/p>\n
That is what Phil Orr saw in 1959, in the form of protruding bone.<\/p>\n
That is what Thomas Stafford and John Johnson confirmed in chemical analysis.<\/p>\n
That is what the geological record corroborated in stratigraphy.<\/p>\n
Someone was there, on an island that required crossing the sea.<\/p>\n
The bones cannot tell us who, in the full human sense.<\/p>\n
But they tell us, with reasonable scientific certainty, when \u2014 and the when turns out to be one of the most important facts in the story of how this continent came to be inhabited.<\/p>\n
The first Californians may have been islanders.<\/p>\n
<\/p>\n
And the first islanders had to have been sailors.<\/p>\n
The Clovis Paradigm And Why It Fell<\/p>\n
To fully appreciate what the Arlington Springs revision meant for the history of American archaeology, it is necessary to understand the intellectual weight of the Clovis First paradigm \u2014 and why challenging it was so difficult for so long.<\/p>\n
The distinctive fluted stone projectile points first identified near Clovis, New Mexico, in the late 1920s and early 1930s were a revelation.<\/p>\n
They appeared, consistently, in association with the bones of extinct Pleistocene megafauna \u2014 mammoths, giant bison, horses.<\/p>\n
<\/p>\n
The association demonstrated not just that humans had been in North America during the Pleistocene, but that they had been hunting its largest animals.<\/p>\n
Radiocarbon dating, as it developed from the late 1940s onward, placed Clovis consistently in the range of approximately 11,000 to 13,000 radiocarbon years before present \u2014 which calibrated to approximately 13,000 to 14,000 calendar years.<\/p>\n
Clovis points were found across a vast geographic range: from the southeastern United States to the Pacific coast, from Alaska to Central America.<\/p>\n
Their distribution suggested rapid spread \u2014 the archaeological signature of a founding population moving quickly through an uninhabited continent with abundant game and no human competitors.<\/p>\n
<\/p>\n
The model was coherent, dramatic, and well-supported by the evidence available through the 1970s.<\/p>\n
The model also had powerful institutional defenders.<\/p>\n
Several of the most respected researchers in American archaeology built careers on it and were not inclined to see it overturned without extraordinarily compelling evidence.<\/p>\n
The standard for acceptance of pre-Clovis claims was accordingly high \u2014 as it should be for extraordinary claims \u2014 and many proposed pre-Clovis sites were found, on close examination, to have chronological problems or contextual ambiguities that undermined their claims.<\/p>\n
This created a situation that critics called a self-fulfilling standard: the threshold for accepting pre-Clovis evidence was set so high that sites with genuine but imperfect evidence were dismissed, while sites with perfect evidence remained elusive because perfect preservation is rare.<\/p>\n
<\/p>\n
Monte Verde eventually met even the most demanding standards.<\/p>\n
Intensive scrutiny by a team that included former Clovis First defenders produced a conclusion that the site was genuine and its dates were valid.<\/p>\n
The paradigm had to change.<\/p>\n
But changing a paradigm is not the same as establishing an alternative one.<\/p>\n
The fall of Clovis First opened the question of what came before and by what route, without immediately providing a clear answer.<\/p>\n
The alternatives \u2014 coastal migration, multiple migration waves, earlier interior entry \u2014 each had supporters and evidence, but none achieved the clean consensus that Clovis First had once provided.<\/p>\n
<\/p>\n
Arlington Springs, with its terminal Pleistocene island bones, was a contribution to this post-Clovis environment: evidence from a specific place and time that coastal presence was a reality by 13,000 years ago, even if it could not specify what route was taken to get there.<\/p>\n
The Radiocarbon Calibration \u2014 Understanding The Curve<\/p>\n
The conversion of the Arlington Springs radiocarbon age to calendar years is worth examining in more detail, because the calibration curve for the terminal Pleistocene has its own interesting history and its uncertainties have sometimes been underappreciated.<\/p>\n
The IntCal calibration curves \u2014 the internationally recognized standards for converting radiocarbon ages to calendar ages \u2014 are regularly updated as new data are incorporated.<\/p>\n
The curves available when the Arlington Springs revision was published in the early 2000s were less precise in the terminal Pleistocene than later versions, because the tree-ring record that provides the most reliable calibration data did not extend back 13,000 years, and calibration for that period relied more heavily on other materials with larger age uncertainties.<\/p>\n
<\/p>\n
The terminal Pleistocene is also a period when the curve has a distinctive shape \u2014 a relatively steep slope in some regions, meaning that a small change in radiocarbon age corresponds to a relatively small change in calendar age.<\/p>\n
This reduces the calendar-year uncertainty in some respects: even with radiocarbon measurement uncertainty of \u00b180 years, the corresponding calendar-year range is constrained by the curve’s shape.<\/p>\n
The more recent IntCal calibration curves \u2014 versions released through 2020 \u2014 incorporate additional data from marine sediments, cave speleothems, and extended floating tree-ring chronologies, and are more precise in the terminal Pleistocene than earlier versions.<\/p>\n
The Arlington Springs date, recalibrated with the most recent curves, produces a calendar range centered on approximately 12,900 calendar years ago \u2014 consistent with the range that had been published and consistent with the NPS timeline designation of 12,900 BP for the site.<\/p>\n
<\/p>\n
The precision of the calibration matters because it affects the specific claim about where in North American prehistory the Arlington Springs individual falls.<\/p>\n
A date of 13,000 calendar years places the individual near but slightly before the main Clovis horizon; a date of 12,000 calendar years would place the individual after it.<\/p>\n
The calibrated range for Arlington Springs \u2014 approximately 12,710 to 13,010 calendar years, using the calibration available from the published work \u2014 places the site squarely in the Clovis horizon or slightly preceding it.<\/p>\n
This chronological positioning is significant: the Arlington Springs individual was alive at approximately the same time that Clovis hunters were using their distinctive fluted points elsewhere in North America.<\/p>\n
<\/p>\n
This simultaneity \u2014 maritime-capable people on a California island at the same time that other people were hunting mammoths in the interior with distinctive stone tools \u2014 was one of the arguments for multiple migration routes and multiple cultural traditions already present in the Americas by the end of the Pleistocene.<\/p>\n
The Drowned Coast And What It Hides<\/p>\n
One of the most important points in the coastal migration discussion concerns evidence that will almost certainly never be found: the archaeological record of the terminal Pleistocene coastline.<\/p>\n
The coastline of California 13,000 years ago was substantially farther from its modern position.<\/p>\n
With sea level roughly 80 to 85 meters lower than today, the continental shelf was exposed.<\/p>\n
The modern shoreline of the Pacific coast is backed by mountains and sea cliffs in many places, but the shelf extends seaward for many miles before dropping off.<\/p>\n
<\/p>\n
In the terminal Pleistocene, all that shelf was land \u2014 coastal lowland with rivers, wetlands, beaches, and the kind of ecologically rich transitional zone between land and sea that human populations around the world have historically preferred.<\/p>\n
The people who may have traveled south along this coast, following marine resources, would have lived at and near the shoreline.<\/p>\n
Their camps, their middens, their shelters \u2014 if any existed \u2014 would have been built on ground that is now 80 to 85 meters underwater, far offshore.<\/p>\n
Without extensive underwater archaeological investigation, these sites are not accessible.<\/p>\n
This has a specific implication for how evidence of coastal migration is assessed: the absence of sites from the terminal Pleistocene coast does not mean people were not there.<\/p>\n
It means their sites are submerged.<\/p>\n
<\/p>\n
The available record reflects the survival of exceptional cases \u2014 sites that happen to be preserved in contexts above modern sea level, such as the inland depositional context at Arlington Canyon, or the elevated terrace sites on the Channel Islands that stood above the maximum late Pleistocene sea level.<\/p>\n
The Arlington Springs individual survived the post-glacial sea-level rise because they died inland, in a canyon where stream sediments buried their remains well above the water level at any subsequent period.<\/p>\n
Had they died closer to the shore \u2014 had their remains been deposited in a coastal camp or midden \u2014 nothing would remain.<\/p>\n
The sediment that covered them would now be at the bottom of the Pacific Ocean.<\/p>\n
This taphonomic fact \u2014 the systematic destruction by sea-level rise of all evidence from the most probable locations of early coastal occupation \u2014 is why even a fragmentary inland find like Arlington Springs is of disproportionate scientific value.<\/p>\n
<\/p>\n
It is evidence that survives from a period whose coastal record is entirely lost.<\/p>\n
Phil Orr’s Broader Research \u2014 The Man Behind The Discovery<\/p>\n
Phil Orr did not arrive at Arlington Canyon in 1959 as a naive discoverer stumbling onto an unexpected find.<\/p>\n
He was a seasoned researcher with decades of experience studying exactly the kinds of geological and archaeological contexts that the Arlington Springs site represented.<\/p>\n
Orr had joined the Santa Barbara Museum of Natural History in the late 1930s and built his career around the Channel Islands.<\/p>\n
He was trained in archaeology, paleontology, and the physical geology that underpinned both \u2014 the kind of broad naturalist preparation that characterized mid-century scholars who might today be split into multiple narrower specialties.<\/p>\n
<\/p>\n
His work on Santa Rosa Island extended back many years before 1959.<\/p>\n
He had been investigating the island’s Pleistocene fauna, its archaeological sites from various periods, and the geological history recorded in its sediments and terraces.<\/p>\n
When he saw the bones in the canyon wall, he had the geological context to immediately understand what he was looking at: the stratum, its depth, its depositional character, its relationship to the island’s climatic and geological history.<\/p>\n
He also had the scientific judgment to understand the limitations of the available dating methods and to make the decision \u2014 crucial in retrospect \u2014 to preserve the bones for future analysis rather than consuming them in procedures inadequate to the task.<\/p>\n
<\/p>\n
Orr left the Santa Barbara Museum of Natural History in 1969, after more than thirty years.<\/p>\n
His 1968 monograph, “Prehistory of Santa Rosa Island,” published by the museum, remains a foundational document in Channel Islands archaeology.<\/p>\n
His legacy was carried forward by John Johnson and others, who applied the improved methods of later decades to the materials Orr had documented and preserved.<\/p>\n
The Early Holocene After Arlington Springs \u2014 What Came Next On The Island<\/p>\n
The Arlington Springs individual did not live in isolation on the island.<\/p>\n
The broader archaeological record of the Channel Islands documents a sustained human presence extending from the terminal Pleistocene through the early Holocene and beyond.<\/p>\n
<\/p>\n
Daisy Cave, a rockshelter site on San Miguel Island \u2014 the westernmost of the Channel Islands, adjacent to the western end of Santarosae \u2014 has produced one of the most informative assemblages of early Holocene material from the California coast.<\/p>\n
The site yielded evidence of occupation from approximately 11,600 to 8,000 calendar years ago, with a remarkably rich organic preservation.<\/p>\n
Shell middens document intensive shellfish harvesting.<\/p>\n
Bone tools, fishhooks made from shell, and cordage made from sea grass are among the preserved organic materials.<\/p>\n
The fishhooks from Daisy Cave are among the oldest in the New World \u2014 evidence that early Channel Islands people were engaged not just in shellfish gathering but in active fishing.<\/p>\n
The cordage documents sophisticated fiber technology.<\/p>\n
<\/p>\n
The overall assemblage from Daisy Cave supports a picture of early Holocene Channel Islands people who were skilled maritime foragers, exploiting the rich marine environments of the kelp-forest ecosystem.<\/p>\n
Other early Holocene sites on Santa Rosa Island itself document the expansion of human presence as the terminal Pleistocene gave way to the Holocene.<\/p>\n
Evidence from approximately 11,800 to 11,100 years ago shows human use of estuarine shellfish and coastal marsh resources \u2014 extending the known use of such productive transitional environments well back before the Middle Holocene.<\/p>\n
As sea level continued to rise through the Holocene, Santarosae began to break apart.<\/p>\n
Anacapa was the first island to separate, approximately 10,300 to 10,900 years ago, as rising water engulfed the low-lying land that had connected it to the larger island.<\/p>\n
<\/p>\n
By roughly 10,000 years ago, the four modern northern Channel Islands had become distinct \u2014 and the crossing from island to island, or from any island to the mainland, required more sophisticated watercraft than the initial crossing to the terminal Pleistocene superisland may have required.<\/p>\n
The technology rose to meet the challenge.<\/p>\n
Shell beads, fishing technology, and sea-grass cordage appear in early Holocene sites at dates around 10,500 years before present.<\/p>\n
The evidence of more elaborate maritime technology becoming more prominent as the islands became more isolated is consistent with the expectation that watercraft sophistication would increase in response to increasing water gaps.<\/p>\n
The people of the early Holocene Channel Islands were building on a maritime tradition that the Arlington Springs individual represents at its terminal Pleistocene beginning.<\/p>\n
What The Site Means For Other Early American Discoveries<\/p>\n
<\/p>\n
Arlington Springs does not exist in isolation in the record of early Americans.<\/p>\n
It must be understood alongside a set of other discoveries that together form the current picture of terminal Pleistocene North America.<\/p>\n
The Anzick Child, a male infant buried approximately 12,600 calibrated years ago in Montana with Clovis artifacts and ochre, provided the first ancient genome from a Clovis-associated individual.<\/p>\n
The genetic analysis placed him on the lineage ancestral to most modern Indigenous people of the Americas \u2014 demonstrating that at least some Clovis people were the biological ancestors of later Native American populations.<\/p>\n
The Anzick burial was associated with genuine Clovis technology: the fluted points and other distinctive Clovis artifacts that define the complex.<\/p>\n
<\/p>\n
The Spirit Cave mummy from Nevada, dated to approximately 10,600 calibrated years ago, represents an individual from the Great Basin with a genome showing both connections to modern Native Americans and some distinctive features of early coastal populations.<\/p>\n
This individual’s genome has been analyzed and contributes to the genetic picture of early post-Clovis western North American populations.<\/p>\n
The La Brea Tar Pits woman (Pit 10), found in a late-19th century excavation of the famous Los Angeles tar pits and recently re-analyzed with ancient DNA, has been dated to approximately 10,000 to 13,000 years ago \u2014 the terminal Pleistocene to early Holocene.<\/p>\n
Her ancient genome has been analyzed and shows connections to the coastal populations of the period.<\/p>\n
<\/p>\n
Each of these individuals contributes a data point to an evolving genetic and archaeological picture of how the Americas were populated.<\/p>\n
None of them, individually, resolves all questions.<\/p>\n
Together, they are building a picture that is more complex than the simple Clovis First model and more complex than any single coastal-migration model.<\/p>\n
Arlington Springs, without ancient DNA, cannot contribute to the genetic picture.<\/p>\n
It contributes instead to the geographical and chronological picture: coastal California, terminal Pleistocene, island context requiring maritime capability.<\/p>\n
The bones locate a person, not a genome, but location is sometimes the most important information.<\/p>\n
<\/p>\n
The Science Of Bone \u2014 Why Two Thigh Bones Can Date A Lifetime<\/p>\n
A brief technical note on why the femur \u2014 the thigh bone \u2014 is so significant for radiocarbon dating, and why the specific chemical fractions within it require careful separation.<\/p>\n
The femur is the longest and largest bone in the human body.<\/p>\n
It contains substantial amounts of compact cortical bone \u2014 the dense outer layer of bone \u2014 and cancellous bone in the interior, surrounding the medullary cavity that produces blood cells.<\/p>\n
Cortical bone is the best source of ancient collagen for radiocarbon dating: its dense structure reduces the infiltration of environmental contaminants compared to more porous cancellous bone or to flat bones like the skull.<\/p>\n
<\/p>\n
Collagen in cortical bone is the protein scaffold that, when mineralized with calcium hydroxyapatite crystals, gives bone its distinctive mechanical properties.<\/p>\n
In living bone, collagen makes up roughly 20 to 30 percent of the wet mass of cortical bone.<\/p>\n
Over thousands of years, collagen degrades \u2014 but under favorable preservation conditions, recognizable collagen can survive for tens of thousands of years.<\/p>\n
The XAD purification method used at Arlington Springs addresses the most significant chemical challenge in ancient bone dating: the contamination of preserved collagen with humic acids.<\/p>\n
Humic acids are the complex organic molecules produced by the decomposition of plant and animal matter in soil.<\/p>\n
<\/p>\n
They are abundant in any soil or sediment that has supported biological activity, and they can bind to ancient bone, introducing carbon from the burial environment that does not reflect the age of the bone itself.<\/p>\n
The XAD resins used in the purification process selectively bind humic acids while allowing purified collagen to pass through.<\/p>\n
The result is a collagen fraction that is substantially freer of humic contamination than what earlier extraction methods could achieve.<\/p>\n
For the Arlington Springs femora, which had spent approximately 13,000 years in alluvial sediments through which water had been moving \u2014 water that carried dissolved organic matter from the surrounding vegetation and soils \u2014 this purification was critical.<\/p>\n
The contamination problem was severe enough that the osteocalcin fraction, extracted without XAD purification, gave a date thousands of years younger than the purified collagen.<\/p>\n
<\/p>\n
Only the XAD-purified collagen gave a result consistent with the geological age of the stratum.<\/p>\n
The Distance That Mattered<\/p>\n
One of the most frequently discussed technical points in the Arlington Springs literature is the question of how far the mainland was from Santarosae during the terminal Pleistocene \u2014 the minimum distance that had to be crossed by whoever reached the island.<\/p>\n
This question has been addressed by multiple researchers using different methods, and the answers have varied somewhat depending on the sea-level reconstruction used and the specific methods applied.<\/p>\n
The general consensus, expressed across the literature on Channel Islands paleogeography, is that the minimum distance from Santarosae to the California mainland \u2014 at the narrowest point of the Santa Barbara Channel at maximum glacial low sea level \u2014 was roughly four to seven miles.<\/p>\n
<\/p>\n
As sea level rose from the glacial maximum toward its present position over the period from approximately 20,000 to 7,000 years ago, this minimum distance increased.<\/p>\n
By the time of the Arlington Springs individual, approximately 13,000 years ago, sea level had risen from its lowest glacial position but had not yet reached the modern level.<\/p>\n
The precise channel width at that date depends on the specific sea-level reconstruction used.<\/p>\n
What all reconstructions agree on is that there was water.<\/p>\n
Even at the most favorable sea-level position, crossing from mainland California to Santarosae required navigating several miles of open ocean.<\/p>\n
The channel was not a calm, protected body \u2014 it is the Santa Barbara Channel, an exposed coastal waterway subject to wind, waves, and currents.<\/p>\n
<\/p>\n
Crossing it required more than wading or jumping: it required a vessel capable of sustaining a crew through those conditions.<\/p>\n
A minimum crossing of four to seven miles of open Pacific water represents a genuine maritime challenge.<\/p>\n
For reference, the distance from Manhattan to Staten Island via the Verrazano-Narrows is about two miles; the English Channel at its narrowest is about 20 miles.<\/p>\n
The Channel Islands crossing was between these extremes \u2014 substantial but not extreme by the standards of documented historical ocean crossings in traditional watercraft.<\/p>\n
People crossed it.<\/p>\n
<\/p>\n
The presence of the Arlington Springs individual demonstrates this.<\/p>\n
The Island Record Beyond Arlington Springs<\/p>\n
The Arlington Springs site is the earliest securely dated human remains from the Channel Islands, but it is not the only early evidence of human presence.<\/p>\n
The broader record of the islands helps place the Arlington Springs individual in context.<\/p>\n
Early occupation sites on Santa Rosa Island have been documented through systematic archaeological survey and excavation.<\/p>\n
Sites associated with the early Holocene show evidence of shellfish harvesting, particularly of the abundant intertidal and subtidal species of the rocky Channel Islands coast: mussels, abalones, sea urchins, and others.<\/p>\n
<\/p>\n
The middens \u2014 accumulations of shell, bone, and other debris that mark habitation sites \u2014 allow reconstruction of the marine environments that early islanders were exploiting.<\/p>\n
One important line of evidence from the broader early Channel Islands record is the association of human sites with the remains of the pygmy mammoths.<\/p>\n
The relationship between early islanders and the mammoths is a topic of ongoing research.<\/p>\n
Some sites show temporal overlap between early human occupation and the presence of mammoth remains in the geological record, suggesting the possibility that early islanders and pygmy mammoths were contemporaries on Santarosae.<\/p>\n
Whether humans hunted the mammoths, or exploited carcasses opportunistically, or had no direct relationship with them at all, is not established.<\/p>\n
<\/p>\n
The evidence for active mammoth hunting on the Channel Islands specifically is limited.<\/p>\n
The broader Channel Islands record also includes evidence of early Holocene shellfish harvesting that extends the implications of Arlington Springs: if people were using marine resources on the islands from the terminal Pleistocene through the early Holocene, they were adapting to changing conditions as the islands separated and the marine environment evolved.<\/p>\n
The kelp forests that characterized the Channel Islands coast \u2014 and that are central to the Kelp Highway model \u2014 were available to these early peoples, providing the rich marine resource base that sustained island living across thousands of years.<\/p>\n
The Sea That Rose And What It Erased<\/p>\n
<\/p>\n
One of the most geologically dramatic events in the history of the Channel Islands was the sea-level rise that followed the end of the last Ice Age.<\/p>\n
As global temperatures increased after approximately 18,000 years ago, the great ice sheets of the Northern Hemisphere began to melt.<\/p>\n
The enormous quantities of water they had sequestered \u2014 water that had been removed from the ocean through evaporation and precipitated as snow and ice over thousands of years \u2014 returned to the oceans.<\/p>\n
Sea level rose, globally, over the subsequent millennia.<\/p>\n
The rise was not uniform in rate: there were periods of rapid rise, periods of slower rise, and complex local effects from the removal of ice mass that had been depressing the land beneath it.<\/p>\n
<\/p>\n
For the Channel Islands, the consequences of sea-level rise were transformative.<\/p>\n
The vast lowlands of Santarosae \u2014 the wide, flat areas that would have been most suitable for human habitation, closest to the water, richest in the transitional habitats that human communities prefer \u2014 were progressively inundated.<\/p>\n
As the sea rose, the habitable area of Santarosae shrank.<\/p>\n
People living near the old shoreline would have had to move upslope as the water advanced.<\/p>\n
Eventually, the water level reached the low points that connected the four modern islands, and Santarosae began to break apart.<\/p>\n
<\/p>\n
Anacapa separated first, roughly 10,300 to 10,900 years ago.<\/p>\n
The other islands followed as the low saddles connecting them were overtopped by the rising sea.<\/p>\n
By roughly 10,000 years ago, the four modern northern Channel Islands were distinct, separated by water.<\/p>\n
This process was, in geological terms, rapid.<\/p>\n
In human terms, it unfolded over many generations \u2014 slowly enough that no individual experienced it as a cataclysm, but fast enough that it was visible within a lifetime.<\/p>\n
The people who lived through this transformation watched their island shrink and divide.<\/p>\n
<\/p>\n
The maritime knowledge that had brought them to the island in the first place \u2014 the ability to cross water, to navigate the channel \u2014 became progressively more essential as the island’s own internal landscape became more fragmented.<\/p>\n
The sea-level rise also erased the evidence of the earliest occupations.<\/p>\n
The terminal Pleistocene campsites and middens that would have been on the lowland margins of Santarosae \u2014 the places where early inhabitants built their fires and processed their food and buried their dead \u2014 are now underwater.<\/p>\n
The Arlington Springs canyon, cut into higher ground, survived.<\/p>\n
Most of what else happened on the ancient shore of Santarosae between 13,000 and 10,000 years ago is on the continental shelf, under water.<\/p>\n
A Note On The Historical Chumash \u2014 And The Careful Bridge<\/p>\n
<\/p>\n
The Chumash of the historical period \u2014 the people who were living on the Channel Islands and the adjacent California mainland when Spanish missionaries arrived in the late eighteenth century \u2014 represent the best-documented Indigenous culture of the Channel Islands region.<\/p>\n