The First Americans

by James Adovasio

  Far from limiting herself to climatic considerations in 1989, Dincauze raised yet again the great specter earlier brought to the fore by Vance Haynes.

  The specter was, of course, contamination. According to Dincauze, the samples we sent off to be dated had not been pure and good.

  Which is why I was so lucky to have such generous funding for the excavation of Meadowcroft Rockshelter. Throughout the excavation, I had been ferocious about our radiocarbon dates and the technical perfection on which they had to rest, lest I wind up the subject of ridicule and insult.Well, that happened anyway, but here is how the true believers, those who worshiped at the Shrine of the Sacred Fluted Point, dealt with my radiocarbon dates.

  CARBON - 14 AND MEADOWCROFT

  None other than Vance Haynes himself wrote not long ago that nowadays archaeology is, essentially, all radiocarbon dating. At the same time, Vance has had the worst time swallowing the radiocarbon dates from Meadow-croft. Vance read the earliest reports on Meadowcroft and was an early visitor to the site. At one point, he actually wrote me to tell me that the work was unexceptionable and that he had found no holes in our data. I could scarcely believe it, and in fact I have saved his letter to this very day. But then a chain of events occurred that have, like his original letter, lasted to the present.

  In a conversation with our radiocarbon specialist, the late Bob Stuckenrath, who was a friend of his, Vance thought he heard Bob say that all of the early radiocarbon samples had to be arrested during the pretreatment process lest they completely dissolve or, as the technical jargon has it, “go into solution.” This in turn suggested to Vance that all of the early dates were contaminated with dissolved materials from some unknown source(s). He immediately became suspicious and freely and aggressively voiced his doubts to all and sundry who would listen. The old Tule Springs specter of contamination had arisen again, this time almost 2,180 miles to the east.

  Never mind that what Bob had actually told him was that the pre-treatment of two very small samples had been arrested and that all the rest had behaved normally. Never mind that Bob clarified his remarks with Haynes on several occasions, and never mind that the distinguished Oxford Radiocarbon Dating Facility in England had explicitly stated that no indications of contamination could be found in an early Meadowcroft sample it dated.

  Once raised, the shadow of contamination has never died, at least in the minds of Haynes and some of his acolytes. This is not science, but, I suppose, neither was Bob Stuckenrath's reaction to all this. He said, “Ifthey don't believe the evidence, fuck 'em”—definitely not scientific discourse, but not ill considered, either!

  Since then (1974), Vance has employed every ounce of creativity in finding ways to blow out of the water any dates from Meadowcroft that place humanity on the scene before 12,000 B.P. or—essentially—before Clovis. As were others at the time (including some of us even before the rest of the pack), he was quick to point out the large discrepancy in millennia between the pre-Clovis Paleo-Indian dates and those of subsequent people (Archaic Indians), and he recommended that more samples be taken and many more dates be run—and at more than one radiocarbon-dating laboratory.

  No one then, by the way, and no one since has questioned any of the dates obtained first or in subsequent tests that are younger than the semi-mystical date of 11,500 years B.P. That is, no one with the exception of two Haynes disciples, Ken Tankersley and Cheryl Munson, who have intimated that all of the dates were wrong. Paradoxically, Haynes never even questioned the date of 11,300 B.P., which is of Clovis age. All the dates that are post-Clovis came, it seems, from perfectly obtained samples from perfectly determined stratigraphy and were perfectly tested in the lab. Whatever must have gone haywire to produce the pre-Clovis dates suddenly became magically inoperative after the magic moment of 11,300 B.P., and critics like Haynes have come up with some astoundingly creative, if unscientific, explanations for this astonishing timing.

  By the later 1970s, we had obligingly obtained nearly forty dates from two different laboratories, which had told us essentially the same story. Since then the total number of dates we have run is fifty-two, with thirteen from the pre-Clovis strata. What they show—and this is really the most important thing about Meadowcroft—is the longest continuous use of a single place in all of North American prehistory.

  I should point out here that none of the hundreds of American or foreign archaeologists who actually visited the site between 1973 and 1978 questioned the rigor or precision of the excavation, the data-recovery techniques, the stratigraphy, the context, or the association of dates, artifacts, and ecofacts. In other words, the dig at Meadowcroft matched the criteria that Hrdlíčka and Holmes set forth more than a century ago and that have been only slightly updated in the context of modern technology.

  I repeat them yet again:

  Artifacts of indisputable human manufacture recovered in primary depositional contexts

  Clearly defined, unambiguous stratigraphy

  Multiple radiometric determinations showing indisputable internal consistency

  Practically no one has ever found these criteria to be too stringent or, as one commentator put it, “legalistic.” Nor are they so lax that they require amendment. However, there have been a few suggested amendments. In fact, Paul S. Martin (of overkill persuasion) and a colleague, Frederick Hadleigh West, have insisted on what they call “replicability.” Now, replic-ability is a very important word in the hard sciences. In a physics or chemistry experiment, replicability is the sine qua non. The way an experiment is set up and carried out must be able to be done by someone else exactly the way you did yours, so that your results can be checked. That, of course, cannot be done exactly in the case of an archaeological dig, and what Martin clearly implied by the word was that lots of pre-Clovis sites need to be found before any one of them can be considered real. And that is, of course, not science at all, however much comfort it brought to those who have lived or died by the Clovis Curtain. (It took the discovery of only one live coelacanth, a weird fish of the Indian Ocean then thought to have been long extinct, to prove that the species still lived.) And perhaps more to the point, even without some sort of replicability, the criteria above have long been successfully used to doom hundreds of alleged but flawed pre-Clovis sites to a deserved oblivion.

  In any event, my dates were internally consistent and few sites have ever yielded more. There was really only one way left to invalidate the dates that had come out of Meadowcroft, and that was, as noted, contamination.

  Something had to have dribbled or floated or drifted into the lower strata of the rockshelter that rendered all (and only) the older, improper, even impolite radiocarbon dates not just wrong but too old by thousands of years. We had numbered the various strata in the nearly sixteen feet ofmaterial that overlay the shale bottom of the rockshelter with roman numerals, with Stratum I being the lowest and therefore oldest. We had received dates on material from Stratum I as old as 30,000 years B.P., but there was no cultural (meaning human) material in that stratum. It was what we call sterile.

  Next up (and younger) was Stratum II, which was in fact separated into two smaller units, with IIa being the one with dates from 19,000 to 11,000 B.P. that were so offensive to the Clovis lobby. This level just had to have been contaminated by something. The obvious candidate—this being Pennsylvania anthracite country—was coal. We knew that the nearest actual outcrop of coal was about a half mile away from Meadowcroft, making transport of coal into the rockshelter pretty unlikely unless humans had actually carried it in. Nevertheless, we had searched for coal particles in every sample taken for carbon dating with both optical and scanning electron microscopy, finding no contamination. None of the dating laboratories or other independent researchers who had looked at the samples had found any contamination by coal particles. We had also checked all these samples for any paleobotanical evidence, chiefly spores, of the sort that typically occur in the coal common to the area (coal that dates back to the Pennsylvani
an age). None turned up.

  Another compromising candidate was vitrinized wood or vitrite, ancient Pennsylvanian-period wood fragments that had become fossilized. We found two highly localized occurrences of vitrite on the north wall of the shelter immediately below the interface or boundary of Stratum IIa and the sterile Stratum I. The vitrite was horizontally or spatially separated from the earliest occurrences of human activity by a distance of more than twenty feet and vertically separated from it by almost two feet of sterile sandy deposits. In order to get vitrite into the culture-bearing levels, it would have had to be physically transported to the early firepits by someone who had dug it up and then added it in exactly the right doses to render the pits older than they actually are. They would have had to add just enough vitrite, for some unknown and unimaginable reason, to sequentially skew the age of each of the pits in precise chronological order—one hell of a feat! This, of course, would also apply to any coal brought into the site from elsewhere.

  Never mind our responses to the critics' objections. Vance Haynes repeatedly suggested that contamination must be present. Perhaps it had percolated up into the site via groundwater.

  We knew it could not have, based on our nuanced understanding of what had brought sediments and other materials into the rockshelter throughout its history. The eleven separate natural strata that exist above the rockshelter's shale floor were all the result of three sources: occasional rockfalls from the ceiling, providing fragments from fist-sized rocks to boulders; a steady rain of mostly quartz grains, also from the ceiling; and what is called sheetwash, meaning sediments coming down from above the rockshelter during rainstorms.

  Nothing was entering the site from below. Indeed, any water movement from a fluctuating water table would have erased the chemical signature that marks the dripline, which moves toward the back of the rockshelter with the ever-retreating rockshelter roof. Inside the dripline, there is a high concentration of calcium carbonate (CaCO3) from mussel shells or the sandstone, while outside, this calcium carbonate signature has been removed by rain percolating into the sediments. The continuing presence of CaCO3 inside the present and to a lesser extent former driplines demonstrated that no upward water movement that would have erased the CaCO3 signature has occurred. In other words, the trail and track of the dripline could still be plainly seen. Additionally, our excavations conclusively showed that after 22,000 years ago, groundwater levels fell, never again to rise within thirty feet of the deepest occupation surface.

  Never mind. Vance Haynes wanted more reassurance, more proof that the site was not contaminated, and continued to call for it in various scientific journals. Others chimed in, notably the aforementioned Haynes disciple Ken Tankersley, then an associate professor at Kent State University in Ohio. Ken's life trajectory has not, it seems, included much by way of notable archaeological findings; instead he has made a career mostly from being a professional naysayer. Tankersley offered some fairly tortured reasoning to come up with contamination at Meadowcroft, including an astonishing wrinkle on the notion of “replicability.” He noted that contamination from vitrite had occurred at two other alleged pre-Clovis sites, so therefore it probably had happened at Meadowcroft. He too pointed to groundwater as the culprit.

  I would not be honest if I said I did not find all of this irritating in the extreme. Here were a handful of Clovis First critics hypothesizing a lot of contorted, even fanciful, scenarios by which the radiocarbon dates at Meadowcroft could be wrong—scenarios we had already systematically ruled out by the most painstaking methods available to us. We responded with what I now look back on as a somewhat (no, very) smart-ass paper in American Antiquity. We called it the “Yes, Virginia” paper; in it we ridiculed Vance Haynes and his acolytes and what we knew were ill-informed criticisms. A colleague, upon reading it, told me that we would surely be vindicated, but it would take a long time and even longer now that we had insulted Haynes, who was by now widely considered the chief arbiter of such matters. Indeed, he was at the time a member of the august National Academy of Sciences. That he had cast doubt on our findings stuck to our conclusions and reports like a lamprey sticks to its host fish, and we came to feel that no matter what we did in defense of our work, it would always be referred to as a site about which “questions” existed.

  All this despite a long list of accolades about how precise, how careful, how technically astonishing the excavation had been, as well as the collection of the data and every other technical aspect of the dig. Numerous archaeologists had commented, calling our work “impeccable.” Brian Fagan, of the University of California at Santa Barbara, said that the “careful work at Meadowcroft has set a standard.” Yet here were a few people, including Vance, carrying on about groundwater's having brought whole or dissolved particles of coal or vitrinized wood into the dating samples, even though no one had ever actually seen such a thing in any of the samples.

  There is a very good reason why no one has ever seen coal particles or vitrinized wood brought up through the ground into the samples: neither substance is soluble in water. Even this elementary fact was ignored. As I noted, once raised, the specter of contamination did not die.

  But say, for the sake of argument, that somehow the samples were contaminated by some magical vehicle. For a sample that was, say, truly 10,000 years old to appear to be 16,000 years old, it would have had to be contaminated by coal or vitrite in an amount equal to 35 percent of its weight. More than a third of the volume in each of the thirteen samples taken in Stratum IIa would have to be coal or vitrite. But why would the groundwater have stopped percolating upward exactly and convenientlyat the end of the period represented by Stratum IIa? Why wouldn't it have gone upward to contaminate more recent samples? And mathematically, the addition of 35 percent by weight of coal or vitrite to create a date of about A.D. 1200 would mean that the sample, if “uncontaminated,” would have had to be deposited in the ground some four thousand years in the future.

  It is remotely possible that the personnel in four different radiocarbon laboratories could have, upon the rigorous examination of each sample, failed to notice the admixture of 5 percent by weight of one or the other contaminant—but that 5 percent would create an error of as little as five or six hundred years—not six thousand. And to suggest that laboratory personnel could fail to note that a third or more of the sample was a contaminant implies a witless incompetence of astronomical, even unimaginable, dimensions. Only professional naysayers could assert that the samples were contaminated sufficiently to render the early dates so wholly inaccurate. But that is what I was up against. In this regard, it is perhaps worth noting that Tankersley and a colleague, in their zeal to discredit the early Meadowcroft dates, actually contrived a laboratory experiment in which bituminous coal appeared to partially dissolve but only after two hundred hours of immersion in water maintained at 100 degrees Centigrade! Interesting, I suppose—at least to a coal chemist—but most agree that such exercises have no bearing whatsoever on the contamination issue at Meadowcroft—much to Ken's chagrin.

  Finally, in 1994, after Haynes's many reiterations of his complaints and the gnat attacks from Dena and Ken and some others, Paul Goldberg and Trina L. Arpin of Boston University stepped in and collected sediment samples from Meadowcroft. Goldberg is the nation's leading scholar and practitioner in the arcane realm of microscopic analysis of sediments, and he made an exhaustive study of undisturbed blocks of sediments at Mead-owcroft. This involved cataloguing and analyzing the relationships between the particles of solid materials and the spaces between them, which are called “voids.” Indeed, there are many different forms of voids that are characteristic of particular sediments; some of these are irregularly shaped ones called “vughs,” some are large, rounded ones called “chambers,” and others are long, narrow ones called “channels.” Anything that contaminated the sediments at Meadowcroft would have had to pass through thenetwork of voids and would have left unmistakable signs, such as filling the voids with particles of
very fine material.

  Goldberg and Arpin concluded in a report in the Journal of Field Archaeology in 1999, “Most significantly in terms of the Meadowcroft dating controversy, we see no evidence of the effects of groundwater saturation of the sediments, nor was particulate coal material visible in any sample… nor do we see evidence of any other mechanisms by which particulate or non-particulate contamination could have been introduced into the sediments in general and into the charcoal samples in particular.”

  The Goldberg report effectively closed the book on the groundwater-contamination question—after twenty years—though not surprisingly, Haynes said it proved nothing. There has never been any particulate contamination found in any Meadowcroft radiocarbon samples, never any coal seam in the shelter. Furthermore, the earliest inhabitants of the rock-shelter were operating in a mosaic environment that has, beginning in 1975, proved to be wholly consistent with new information about life at the glacier's edge. Even if you average the six deepest dates associated with human artifacts—a procedure that makes no statistical sense at all though it is practiced by some archaeologists—people were in western Pennsylvania at least as early as sometime between 13,955 and 14,555 radiocarbon years ago. Not only that, but the human artifacts from those earliest times no longer exist and, in fact, never existed in isolation. Two other sites in the Cross Creek drainage, the Mungai Farm and the Krajacic locality, have both produced distinctive Meadowcroft-style implements, including cores, blades, and a Miller point. So also has Cactus Hill in Virginia. Replication!

  So, for now at least, the humans who stopped off at Meadowcroft represent the pioneer population in the upper Ohio River Valley and perhaps the entire Northeast. From their varied tool kit, we know they were hunter-gatherers who opportunistically foraged for a variety of plant and animal foods. They were not specialized hunters of big game, and there are no clear or unambiguous connections between them and the people who produced the Clovis artifacts. Those are now the generally accepted facts. I believe I have made the case.