THE PALEOAMERICANS:

Issues and Evidence Relating to
the Peopling of the New World

The question of human antiquity in the Americas remains one of the great unanswered research problems of science and one of the most contentious and debated issues in American archaeology. First entry is an important question. Important theoretical issues hinge on the location, origin and timing of migrations to the Americas. Despite seventy years of archaeological research and several apparent resolutions to the problem, the timing and mechanisms by which the New World was peopled continues to be debated by specialists. Currently, no definitive evidence exists to establish when the first Americans arrived, or who they were, to the satisfaction of the many researchers in diverse scientific fields focused on the problem. Additionally, researchers are proposing very different answers to this important question.

In this paper I will outline the areas of consensus and of contention regarding knowledge of Paleoamerican colonization. Bonnichsen and Turnmire (1999:1) proposed the term "Paleoamerican" for the late Pleistocene and early Holocene populations and archaeological complexes of the Americas to avoid an inference of biological continuity between the current Native American populations and the earliest population. Herein Paleoamerican refers to the first wave of migration and the founding population groups in the Americas. This distinction from all Native Americans, an easier semantic distinction than actual, will be further elucidated by the following discussions.

Regarding first peopling, I discuss evidence, understandings, results, and hypotheses which I characterize as within a general consensus, and those in contention. Researchers with slightly different approaches to the same problem with the same methods can arrive at distinct conclusions from seemingly the same evidence. To help clarify these differences, I begin my discussion with the ideas having the greatest consensus, thereby defining the common ground of thought. My discussions move to areas with distinct and opposing views, and include issues relating to the resolution of opposing views and theses. Finally, I move from the general to the specific by focusing on archaeological sites and the physical evidence at the center of the current debate.

THE CURRENT GENERAL CONSENSUS

Theoretically, because direct evidence cannot be presumed to be the oldest, dated sites are considered minimalist indications of actual antiquity. There is universal agreement that Homo sapiens sapiens occupied the Americas by 11,200 radiocarbon years ago, over 13,000 years ago (Toth 1991:53). This paradigm was a major shift in the history of American archaeology. In Folsom, New Mexico, the 1927 discovery of fluted projectile points embedded in the rib cage of an extinct bison provided unequivocal proof of human association with extinct megafauna and of late Pleistocene occupation in the Americas. Clovis is the most widespread cultural complex in North America. Clovis and Clovis-like points are found throughout North America. In the late 1950s, radiocarbon dating at the Lehner site in Arizona dated Clovis to 11,250 years B.P., (Haury, et. al., 1959). Haynes’ analysis dating Clovis culture to 11,000 to 11,500 years B.P., based on 40 radiocarbon determinations from seven sites, of which 31 were from the Lehner and Murray Springs sites, is considered noncontroversial (Rogers, et. al. 1992:286).

There is also universal acceptance of the route of origin of the prehistoric migrations, at least in general terms, from Asia via Siberia and Beringia. Beringia has been defined variously, as the continental shelf between Asia and North America inundated by the rising sea levels, more broadly to include existing land areas in northeast Siberia and western Alaska, even more expansively as the Pleistocene land mass between the Lena and MacKenzie Rivers, and as the land between the Laurentine Ice Sheet and the Verkhoyansik Range (Hoffecker, et. al., 1993:46). I accept the expansive use herein. The sea level was at its lowest, and Beringia most extensive in area, between 20,000 and 18,000 years ago (Hoffecker, et. al., 1993:46). The precise routes across the land bridge, along the shores, over sea ice, or combinations of the diverse possibilities, remain outside the current consensus.

Walking from Asia to America on dry land was minimally possible anytime between 30 to 14.5 kya (thousands of years ago) and maximally between 65 and 10 kya, and, before and after this time span, seasonally on sea ice (Wright 1991:138, Hoffecker, et. al., 1993:46, Meltzer 1995:37). Human capability to cross from Asia to America does not pose a fundamental problem. Boats were used along the Pacific rim at a very early date. Trace element sourced obsidian from Kozushima Island demonstrates use of boats to transport the obsidian to Japan at least 30 kya (Bonnichsen and Schneider 1999:509). Other water crossing evidence and their minimum dates include Okinawa by 18 kya, New Ireland by 34 kya, the Soloman Islands by 29 kya and Australia by 45 kya (if not far earlier) (Rogers, et. al. 1992:285, Fullager, et. al., 1996). However, archaeological data supporting a shoreline migration to the north and east is lacking and the Pleistocene shoreline is currently inundated.

The consensus about the route of migration extends to and overlaps with the geographic origin of the Paleoamericans, northeast Asia. Meltzer raises the possibility that migrations could move in both directions, with Paleoamerican populations influencing the genetic composition of Siberian populations and, consequently, later migrations back to the Americas (Meltzer 1993:165). At least from a theoretical perspective, migration models and bioarchaeological studies must account for the fact that gene exchange across Beringia has been possible in both direction during the entire occupation of the Americas.

BIOLOGICAL EVIDENCE

Perhaps the most important, persuasive, and broadly accepted evidence of northeast Asian and American population relationships is dental morphology. Teeth have very good preservation in comparison to other tissues, hence dental studies have the advantage of sampling prehistoric individuals. All Native American populations have teeth more like those of northeast Asians than any other world population (Turner 1994:131). Turner’s dental study of more than 15,000 crania indicate that Native Americans and Northeast Asians represent one of two major global dental clusters (1994:131). Turner concluded that "dental morphology indicates that the Americas were colonized by small bands of Siberians whose genetic ancestry was with the evolving late Pleistocene Mongoloid population" (1994:137). Greenberg, Turner, and Zegura (1986:480) reported that dental variation is greater in the north, and that there are three American dental (and parallel linguistic) clusters, Na-Dene, Aleut-Eskimo, and Macro-Indian. Of these, the Macro-Indian corresponds to the Paleoamericans.

There is also a basic consensus regarding major phases of migration from Asia. In the 1980’s the synthesis of linguistic, dental and genetic data suggested three migrations from Asia to America, with each wave leading to a separate linguistic group (Greenberg, et. al., 1986, Gibbons 1996:31). This consensus does not extend to the probable dates of the migrations, or the particular details of when and where these groups branched.

Mitochondrial DNA (mtDNA) studies provide more conclusive information, perhaps, about migration direction than do dental morphology clusters. But, unlike dental studies, mtDNA studies have been based on contemporary populations and therefore reflect admixtures after migrations. Multiple migrations of closely related groups would not be evidenced by linguistics or genetics. The genetic studies lack sufficient prehistoric comparisons. Recovery of mtDNA from prehistoric remains is possible, but such opportunities represent a very limited sampling.

Other limitations of genetic studies must be kept in mind when their results are used in support of migration hypotheses. Molecular divergence can precede population divergence. When molecules in the mtDNA chain diverge reflects when a population’s genetic composition diverges, but does not necessarily coincide with when a population became genetically isolated. Mutations evidenced today may predate divergence, and statistical change may be due more to population dynamics than temporal depth. Lineages of mtDNA represent only the maternal line, and lineages can be lost in the original and derived gene pools. Also, use of present-day genetics is relative. Once discernible phylogenies can be altered, over thousands of years, by inbreeding, gene exchange, demographic alterations, natural selection, and random change in genetic frequency.

The genetic composition during the mutation of the virus may change and then the vaccine against coronovirus may not protect against infection, but if you use a Amoxicillin, it will be effective regardless of the mutation of the virus.

Given the limitations of mtDNA study, the genetic evidence nonetheless presents a clear enough picture of the genetic relationships of the Old World and Native American populations to derive inferences about Amerindian relationships. The three native American linguistic groups all carry only four mtDNA haplogroups, with each group characterized by a unique set of mutations, and the same variants are found in East Asian and Siberian populations, indicating a shared line of descent for all these groups (Gibbons 1996:31). Interpretation of this consensus dissolves into disparate conclusions, particularly so with regard to claims of temporal radiation and occupational depth (Meltzer 1995:29-30).

The ancestral Amerindian radiation has been estimated, using mtDNA mutation rate assumptions, at 21 to 42 kya and, more recently, at about 19 to 38 kya (Meltzer 1995:30). Different analytical techniques on different mtDNA sequences have produced different results, with radiation times as great as 41 to 78 kya, while another group using different assumptions put the Amerind lineage origin potentially within Clovis times (Meltzer 1995:31). Greenburg, et. al., (1986:480) concluded that the Amerind (Macro-Indian or Paleoamerican) language group differentiated over a "period probably greater than 11,000 years and beyond the limits of glottochronology." Unlike genes, languages are not biologically inseparable from populations. And given the baseline accepted time depth, a minimum of at least 11.5 kya, language is of limited utility in determination of first peopling. The greatest utility of linguistics in peopling of the Americas studies seems to be in distinguishing the Paleoamerican and later migrations, and in resolving issues with far less time depth than original peopling.

Schanfield, after comparing immunoglobulin allotypes of 28,000 Amerindians with Asian populations, concluded, based on the distribution of GM haplotypes, that at least four migrations of Asians entered North America, and estimated that the first migration occurred between 17 to 25 kya and the second between 10 to 13 kya (Schanfield 1992:381-397). Wallace and Torroni (1992:403), after comparing the mitochondrial DNA (mtDNA) in 167 individuals from six geographically disparate American tribes and in 153 individuals from seven east Asian populations, concluded that the American mtDNA populations descended from five Asian mtDNAs and cluster in four lineages. Wallace and Torroni consider distribution of the four lineages throughout North, Central, and South America as a demonstration that "Amerinds originated from a common ancestral genetic stock" with sequence frequencies indicating that "the Amerindian mtDNAs arose from one or maybe two Asian migrations that were distinct from the migration of the Nadene and ... about four times older than the Nadene" (1992:403). They placed the age of the four Amerindian lineages at 21 to 42 kya.

In 1993 Torroni, et. al. (1993:591), published results of mtDNA study of 411 individuals in 10 Siberian populations, which they report as supporting the hypothesis that "the initial migration occurred between 17 and 34 kya. They also report that one of the Amerindian mtDNA groups is not found in Siberia, raising the possibility of loss due to drift. or of a distinct migration to the Americas without impacting Siberian genetic composition (Torroni, et. al., 1993:603).

ADAPTATION

The environmental barriers to human migration to the Americas until the end of the Pleistocene were severe. Given migration via Beringia, the most significant limiting factor in New World peopling is adaptation to arctic or, at least, subarctic conditions. Adaptation is an area with considerable consensus about the broadest issue, that humans had to adapt to the climate in Beringia in order to successfully migrate to the Americas. Unresolved minor differences arise regarding specific adaptations, the degree of adaptation actually required, and when and where that adaptation arose.

Humans are tropical primates that require cultural adaptations to survive northerly latitudes, and hence, to have arrived in the Americas, capabilities such as control of fire, killing large mammals, hide working, tailoring clothing, sewing, and construction of shelters, all archaeologically demonstrable in the middle Pleistocene, are seen as components of arctic adaptation and prerequisite to peopling of Siberia and Beringia.

There seems to be little contention about glaciation and its impact on possible migration, or about the microclimatic variations along the Pacific coast of North America. During the late Pleistocene glacial maxima, about two-thirds of the present Alaskan land mass was ice free (Hamilton and Goebel 1999:156). Ocean moderated coastal refugia were interspersed with glaciers from Alaska to Puget Sound. The change in ocean level has inundated the Pleistocene coastline and continental shelf along possible early Pacific coastal migration routes. Inland routes, to the east and west of the Rockies in Canada, were glaciated from 20 to 14 kya minimally, and fossil vertebrates in the glaciated regions do not predate 11.3 kya (Meltzer 1995:38). There are no sites in the hypothetical Ice-Free Corridor, (between the Laurentine and the Cordilleran glaciers) to provide convincing evidence of human activity prior to the last major glaciation, nor are any known from the Alaska and British Columbia coasts (Wilson and Burns 1999:235). To have survived passage through this region, it is generally agreed, humans first adapted elsewhere, specifically in the northern latitudes of Asia.

Nearly all lower Paleolithic sites in Siberia have either problematic geological contexts, undateable contexts, or questionable artifacts, and no artifacts from radiometrically dated primary contexts are known (Goebel 1999:209-212). Excepting possibly the lowest layers at Denisova Cave, middle Paleolithic occupation of Siberia, based on associated fauna and stratigraphic contexts, appears to date to about 70 to 45 kya, is limited to southern Siberia, and is associated with the Neanderthal Mousterian industry (Goebel 1999:212).

Some 20 upper Paleolithic sites (ranging from 42 to 30 kya) in southern Siberia, below 55° latitude, demonstrate adaptive changes in hominid technology (Goebel 1999:213). Hearths, dwellings, and some artworks are indicated, and new tool industries are in evidence, centered on unifacial and bifacial blades, with retouched blades, scrapers, and burins. Generalized hunting, with ten or more prey species. is evidenced at the few sites with faunal analysis (Goebel 1999:214-215). Between 26 and 19 kya, evidence supports the spread of humans into the subarctic mammoth-steppe of central Siberia (as far as 60° latitude, and possibly further, latitudes comparable to Beringia). Substantial shelters with central hearths and storage pits were constructed, bone, antler and ivory artifacts abound, bone awls and needles appear, bone and ivory flaking is seen, and decorative and figurine arts are evidenced (Goebel 1999:216).

Dateable sites indicate a decline in Siberian population during the glacial maximum, dated to 19 to 18 kya, followed by the spread of microblade technologies from the Lake Baikal region to the Pacific coast and the Arctic coastal plain by 12 kya (Goebel 1999:218-223).

Two recent, possible Homo sapiens sites pose questions about when the earliest adaptations to latitudes comparable to those of the Beringia occurred. Traces of an approximately 36 kya occupation (stone artifacts, animal bones and a mammoth tusk with human-made marks) are reported from Mamontovaya Kurya, situated above 66.5° latitude in the Ural mountains, the oldest documented evidence for human presence at such a northerly latitude (Pavlov, et. al. 2001:64). The previously earliest dated arctic sites in Eurasia dated to 13 to 14 kya (Pavlov, et. al. 2001:65).

The Mamontovaya Kurya faunal assemblage, and radiocarbon dating of a series of mastodon teeth from Finland, suggest a dry and ice-free period, with a steppe-like, open grassland environment in arctic and subarctic Siberia at the time (Gowlett 2001:33). Study of Greenland ice cores indicate a colder climate record from 60 to 35 kya, but one with a series of temperature fluctuations of up to 15° C., implying movements of steppe environments from southeast Europe to the arctic (Gowlett 2001:34). An occupation site at Ust’Mil’ II near 62° north latitude was radiometrically dated to 35 kya and less securely dated sites suggest older occupation (Rogers, et. al. 1992:285).

Human intrusion near the Ural mountains, while indicating possible earlier arctic adaptation that previously known, is nowhere near Beringia. Another factor to consider regarding the climate in Beringia, particularly in comparisons to inland Siberia, is the proximal margin of the Pacific Ocean and the climate moderating effects of the oceanic mass. This difference may be manifested in Pacific Rim findings that support the view of earlier penetration of northern latitudes than previously thought. A series of more than 40 Middle Paleolithic sites (ranging from 150 to 200 kya) with simple core and flake technology are known from the northeast coast of Honshu Island, Japan (Bonnichsen and Schneider 1999:508). A series of 10 thermoluminescence dates date the cultural bearing layer of Diring Yuriakh (61 degrees north latitude on the Lena River) to greater than 250 kya and less than 320 kya (Bonnichsen and Schneider 1999:508). According to Bonnichsen and Schneider (1999:508), "cold adaptation to subarctic conditions no longer should be considered as a factor limiting human movement into the New World during middle and late Pleistocene times."

THE ARCHAEOLOGY

Linguistics and genetic studies provide inferences, albeit theory-driven, about temporal placement, but lack the all important aspect of direct dating. Ultimately, the issue of earliest migration is defined by archaeological evidence in the Americas. Herein I divide the discussion of archaeological evidence into geographical regions, Siberia (discussed above), Beringia/Alaska, North America, and South America. Before discussing the archaeological evidence in these regions I will continue presenting points of consensus and contention. In this case, that is with regard to the archaeological evidence, the points of contention clarify the issues better than the points of consensus.

The current models for the first peopling are broadly divided into the Late-Entry and the Early-Entry models. The currently dominant Clovis-First model, one variant of the Late-Entry model, proposes that a small migration of hunters, equipped with thrusting spears and possible atlatls with fluted points, entered the Americas from Siberia around 11,500 years ago (Bonnichsen and Turnmire 1999:1). Late-Entry models generally envision that big-game hunters, after passing the glaciated northern latitudes of North America, entered a vast continental expanse populated by megafauna, then rapidly dispersed over the extent of the Americas. The various Early-Entry models all agree on peopling well before 11,500 years ago, however the precise timing and possible routes remain speculative (Bonnichsen and Turnmire 1999:1).

The main obstacle "to general acceptance of the Early-Entry model continues to be questions over dating and whether the artifacts reported from many sites are human or natural in origin" (Bonnichsen and Schneider 1999:510). The history of proposed pre-Clovis sites reads like a history of shipwrecks. Numerous proposed pre-Clovis site are no longer considered viable candidates. Meltzer (1995:22) pointed to 50 sites on a 1964 list, of which none survived scrutiny as of 1988, and, in 1997, stated, "So many have failed that the archaeological community has grown highly skeptical of any and all pre-Clovis claims." Rogers, et. al., (1992:286) point to 87 dates at 33 pre-Clovis sites with lithic artifacts that had buried components whose radiocarbon dates were greater than 11.5 kya. Taylor states that of the more than 100 major pre-Clovis sites published by professional archaeologists in major journals or monographs "only a small percentage remain under active review" (Taylor 1991:102). The most important of these are discussed below.

The burden of proof certainly lies squarely with the supporters of the Early-Entry models. Arguments for the Late-Entry model cannot be based on the lack of evidence, which is equivalent to proposing a null hypothesis. Nonetheless, it is important to discuss areas where evidence is lacking. Stylistic and technological continuity in early projectile point types can generally be traced on a regional level (Dillehay 1999:207). There is no current, widely-evidenced pre-Clovis technology associated with the earliest Paleoamericans. The scarcity of Paleoamerican physical remains is also problematic. The oldest human skeletons in the Americas, from the Wilsall site in Montana and the Mostin site in California, are dated to 10,600 � 300 and 10,470 � 490 years B.P. respectively (Taylor 1991:102). No more than 37 skeletal remains are older than 9,000 years, and, of these, only 11 are complete with fully measurable crania (Chatters 2000:292). While studies of Paleoamerican skeletal material in recent decades indicate unexpected diversity and physical distinction from modern counterparts (Chatters 2000:291), larger sample size is needed before conclusive inferences can be drawn.

There is consensus on the current archaeological standards and methods. If the current standards had always been in place, many sites would not now have a place in the pre-Clovis graveyard. The standards of proof have never been more refined. Evaluation of evidence from proposed pre-Clovis sites has eliminated all but a very few. Before discussing the specific sites and regions, I want to review the issues that have spoiled past claims. These issues should be kept in mind when considering some of the pre-Clovis candidates discussed below.

Discrimination of human and nonhuman fracture patterns in stone is essential to site identification. Human activity will produce an assemblage of flaked stone, patterned flaking of cores, and identifiable use-wear patterns. Natural processes can chip rocks, produce apparent cores, cause use wear, and can carry rocks over great distances. Manuports without artifactual associations are now suspect.

Spiral fractures and polish on redeposited bone, originally attributed to human agency, have been demonstrated for other taphonnomic processes, such as carnivore action and ice movement. Cutmarks on bone should exhibit predictable anatomical placement related to butchery. Carnivores and humans can produce similar bone fractures, and fracture patterns produced by humans can also be produced by other natural forces, hence, inferences of human activity based entirely on bone fracturing, without associated artifactual materials or specific patterning, are difficult to support.

Geological context must be assessed. Sites situated in high-energy locations, where cliffs, steep gradients, moving water or glaciation may have influenced deposits need to be assessed for geological alteration of deposits. Sites with moment of water must be assessed for contamination.

Several samples of material need to be radiocarbon dated to provide adequate statistical support of any early date at a site. Charcoal and megafauna remains associated with flaked rocks have yielded pre-Clovis dates, even though bones and wood can be used at a much later date than when the animal or tree lived. Even in caves charcoal can form from natural processes and does not always represent human activity. Carbon samples may be contaminated by limestone, coal or petroleum. For example, the Lewisville, Texas, site produced a radiocarbon date of 36,000 years B.P. from a definite hearth, then reinvestigation established that lignite had been burned in the hearth around 11,000 years B. P. (Wyckoff 1999:348).

The score sheet is not entirely imbalanced with only bad archaeology on the Early-Entry side. Well-accepted archaeological evidence has been used to question the Clovis-First model. A problematic issue for the Clovis-First model is that of regional co-traditions contemporaneous with Clovis. The Goshen cultural complex at the Hell Gap site in southeastern Wyoming is an earlier manifestation than Clovis, with dates as early as 11,400 years B.P. (Adovasio and Pedler 1997:574). Early-Entry proponents point to Clovis co-traditions, such as Nanena in Alaska, Chesrow in Wisconsin, Goshen in the northern Great Plains and the stemmed lanceolate point of the North American West. Early stemmed points in the Basin and Range physiographic province probably represent a tradition that evolved independently of, or was coeval with, Clovis (Bonnichsen and Turnmire 1999:12). Stemmed points were in use in the Great Basin before fluted points, before 11,500 years B.P. (Byran and Tuohy 1999:251, 253). Fluted points, while common surface finds, have been excavated at only five sites in the greater Great Basin, and classic Clovis points have never been recovered from radiocarbon-dated contexts in the Great Basin (Byran and Tuohy 1999:254-255). Available dates indicate that fluted points were in use in the Great Basin between about 11,000 and 8500 years ago (Byran and Tuohy 1999:255). A spectacular cache of classic Clovis points from Wenachee, Washington, and a stemmed point from northeastern Oregon are associated with the Glacier Peak ash, and therefore date to just after 11,250 years B.P. (Byran and Tuohy 1999:255).

South America, where this argument is applied to an entire continent, does not exhibit a continent-wide pattern of distribution of a tool type, such as the Clovis manifestation of North America (Dillehay 1999:207). Gruhn reports a "now-solid body of evidence" supporting occupation of, and local adaptations to, all major environmental zones of South America during Clovis times in North America (Gruhn 1997:29). Sites with evidence of cultural materials from before 11 kya include Taima-Taima, Venezuela and several rockshelters in Brazil and Tierra del Fuego (Dillehay 1999:210-211). Caverna da Pedra Pintada, Monte Alegre, Brazil, contains rock exotic to the cave, rock painting, finely chipped bifacial and unifacial lithic tools, establishing human occupation (Roosevelt, et. al. 1996:361). The about 50 AMS and conventional radiocarbon dates spanning from 10,000 to 11,145 years B.P. clearly establish Pleistocene occupation and adaptation in the Amazon Basin and supports the view of a more complex Paleoamerican radiation than the Clovis-First model (Roosevelt, et. al. 1996:361). Amazonian Paleoamerican cultures, in contrast to the contemporaneous Clovis culture, evidenced generalized foraging subsistence practices, art styles, and different formal tools.

Broad spectrum economies documented by 11,000 years BP in South America are not seen in North America until about 10,000 years BP (Dillehay 1999 214-215). The regionally distinct unifacial tool traditions of South America, and the parallel development of unifacial and bifacial traditions, contrast with the North American uniformity and fluted traditions and supports rejection of the intrusive-Clovis culture model to explain fluting in some areas of South America (Dillehay 1999:208). Some archaeologists think the fluted point was developed in South America and diffused to the north (Dillehay 1999:208). The varied adaptations early in South America are not easily explained within a hypothetical framework that has Paleoamericans arriving in Alaska only a few centuries earlier. While not conclusive, these chronological comparisons, based on accepted archaeological evidence, are thought provoking.

The remainder of my archaeology discussion is focused on specific regions and sites.

Beringia and Alaska.

Archaeological sites in Beringia shed little light on the Clovis diaspora (Hoffecker, et. al., 1993:46). The Clovis-First model is based on the distinctive fluted-point assemblages and on mammoth and megafauna kills in the Great Plains and in the Southwest. However, there are no Clovis sites in Alaska, nor are mammoth kills in evidence (Meltzer 1995:24, Yesner 1996a:248).

The earliest archaeological sites in far northeast Asia date from 14 to 11 kya (Goebel 1999:224) The Berelekh site contained artifacts in relation to bones dated to 13,420 to 12,240 years B.P. (Hoffecker, et. al., 1993:50). The lowest occupation horizon of Ushki, in central Kamchatka, yielded an artifact assemblage and other items dated to 14,300 to 13,600 years B.P. (Hoffecker, et. al., 1993:50). In Beringia there is a difference between premicroblade and microblade complexes and an obvious chronological dichotomy between them (Goebel 1999:224). Clovis and other Paleoamerican complexes lack microblade technologies.

In Alaska the Nenama complex, dates to 11.8 to 11 kya (Goebel 1999:224). The earliest firm evidence of human occupation in Alaska dates to 11,800 years B.P. (Hamilton and Goebel 1999:156). The stratified sites of Dry Creek, Walker Road, Moose Creek and Owl Ridge, all in the Nenana valley region, produced a blade-and-biface assemblage with a probable age of 11,300 to 11,000 years B.P. (Hamilton and Goebel 1999:157). In the nearby upper Tanana valley, the Broken Mammoth, Swan Point, Mead and Healy Lake sites contain cultural occupation evidence dated from 11,800 to 11,000 years B.P. (Hamilton and Goebel 1999:157).

About fifty fluted points have been found in northern Alaska, most from undated surfaces, shallow sites or Holocene contexts (Hamilton and Goebel 1999:180). All the credible radiocarbon dates from associated charcoal are younger than the 11,300 to 10,900 years B.P. assigned to the Clovis horizon, and the points have technological differences, supporting the hypothesis that fluted point technologies in Alaska originated to the southeast in North America (Hamilton and Goebel 1999:181). No Clovis fluted points are found in the reputed Siberian homeland (Bonnichsen and Turnmire 1999:2).

Hoffecker, et. al., (1993:51-52) maintain that a quantitative comparison of selected Nenama and Clovis assemblages indicates that the two are similar and share diagnostic elements, that there is reason to include the Nenana Complex in the Paleoindian Tradition, that Nenana and Clovis probably represent the same population, and that revised dating of key Clovis sites suggests that Nenana predates Clovis by a century.

The Nenama complex is a firmly established cultural manifestation with numerous stratified sites. There is only one proposed Early-Entry site in eastern Beringia that still has proponents, Bluefish Caves in the Porcupine River Basin, Yukon Territory, Canada. Claims of great antiquity in the area have a convoluted history. A caribou tibia flesher, the most diagnostic human implement from the Old Crow basin, had been dated at near 28 kya. When redated using accelerator mass spectrometry (AMS) of small amounts of remnant collagen the bone produced a 1.8 kya date (Yesner 1996b:255).

The original, mistaken dating of Old Crow bone material led to exploration and survey of the surrounding area. Bluefish Caves was found to contain a depositional context with cultural materials, including lithic artifacts, bone alteration from butchering, bone tools and examples of bone reduction by flaking (Cinq-Mars and Morlan 1999:203). A mammoth bone flake and its core were studied, and Cinq-Mars and Morlan (1999:205-206) concluded that the bone evidences butcher marks and fresh-state fracture, and that the flake was reduced bifacially and diagonally in a step-by-step ordered sequence. The bone collagen from the bone and flake were AMS dated to an average age of 23,500 years B.P. (Cinq-Mars and Morlan 1999:205).

Serious concerns still exist about the possibility that natural processes of bone breakage created the artifacts (Yesner 1996b:256, Wilson and Burns 1999:217), and about the depositional complexities (Hoffecker, et. al., 1993:50). The use of modified bone as an indicator of human activity without collaborating evidence has been shown to be ill-advised by Binford and by Bonnichsen and Sorg (Stanford 1999:286). Although radiocarbon dates from Bluefish Caves are considered reliable, they are from bone and the relation between the age of the bones and the time of their use as artifacts remains less than clear (Hoffecker, et. al., 1993:50). Unlike the Folsom bison find, this case lacks a hunting tool embedded in bone. Diverse scenarios are possible regarding when the bone was employed.

North America

The one site in North America outside of ancient Beringia commonly cited as having the best-dated evidence for human occupation in Pleistocene America is Meadowcroft Rockshelter in Pennsylvania. The available data here suggest that small groups of generalized foragers occupied midcontinental North America before 14,000 years ago (Lepper 1999:366-367). Adovasio, et. al. (1990:348), contend that Meadowcroft represents human presence by 14,000-14,500 years ago.

Meadowcroft, a deeply stratified multicomponent site, has produced the longest intermittent occupational sequence in the New World, and has so far yielded some 20,000 artifacts, 150 fire pits, 33 fire floors, 52 ash and charcoal lenses, a million faunal remains, and about 1.4 million plant remains (Adovasio, et. al. 1999:417-418). Of the one hundred and four charcoal samples submitted to four laboratories, fifty two have produced dates (Adovasio, et. al. 1999:420). All but four dates, in the Middle Archaic or younger, are in chronological order and internally consistent (Adovasio, et. al. 1999:420).

The Meadowcroft archaeological study was challenged by assertions that the dates older than 11,000 years suffered coal contamination (Adovasio, et. al. 1990:348). There are no anomalous later Paleoindian or Archaic artifacts associated with the pre-Clovis strata (Lepper 1999:366). Only the thirteen dates older than 12,800 �870 years B.P. have been questioned. According to the site’s researchers, critics of the Meadowcroft dates have selectively questioned only those dates that contradict the Clovis-First paradigm (Adovasio, et. al., 1990:349). Numerous chronologically sensitive artifact types have supported the radiocarbon dating in the post-Pleistocene range (Adovasio, et. al. 1999:422).

The dates were consequently reassessed using AMS. Of these the six that have "clear, undeniable and extensive artifact associations" average at 13,955 to 14,555 radiocarbon years ago. (Adovasio, et. al., 1999:422-427). Adovasio, et. al. (1999:420), contend that if the radiocarbon dates in this range are inaccurate, so are "all the myriad sites that have produced diagnostics presumed to be of similar age". Adovasio, et. al. (1990:349), also contend that the integrity of the deepest occupation surfaces is supported by a definitive separation of superimposed living floors on both qualitative and quantitative geological, geochemical, and sedimentalogical grounds, by lithic refitting and by raw material studies. They point out that there is no coal vein in or near the rockshelter and that every sample from Pleistocene-age deposits was examined for coal particles and coal-associated microflora using both optical and scanning-electron microscopy and no coal particle was ever detected despite the fact that contamination of the magnitude suggested would require a 35 percent fraction of the sample to be coal (Adovasio, et. al. 1990:351). Additionally, abundant artifactual associations proving that the post-12.8 kya dates are valid renders the argument for particulate contamination unconvincing and intrinsically unlikely (Adovasio, et. al. 1990:351-352). Adovasio, et. al. (1999:427), remain convinced that the Meadowcroft Rockshelter "represents the earliest bona fide evidence of human occupation south of the glacial ice in North America".

Meadowcroft is not the only North American site for which excavators continue to claim per-Clovis antiquity. However, none of the following sites, all of which have provided pre-Clovis dates, compare with the quantity of evidence offered at Meadowcroft. Pendejo Cave, New Mexico, is a stratified site with flaked stone tools and 55 radiocarbon dates extending back to more than 55 years B.P. (Bonnichsen and Turnmire 1999:19). MacNeish argues that the site contains 186 exotic lithic artifacts, logs, internal fires not attributable to natural causes, fingerprints on fired clay, and human hair in pre-12 kya levels (Bonnichsen and Turnmire 1999:13).

The basal occupation of Fort Rock Cave, Oregon, yielded one date of 13,200 � 720 years B.P. (Byran and Tuohy 1999:256). Near Tule Lake, California, a small fire pit horizontally proximate to chips of obsidian and bifacially worked material in a rockshelter yielded a radiocarbon date of 11,450 �340 years B.P. (Beaton 1991:5-7). This material overlays an older deposit with bone tools, bifacially worked pieces and flaking debitage associated with fragmentary faunal remains (Beaton 1991:7).

False Cougar Cave, Montana, produced human hair believed to be from a level dated at 14,500 years B.P. (Frison and Bonnichsen 1996:311). The Lovewell Mammoth site in north-central Kansas evidences high-velocity impact points and bone flaking of limb-bone fragments according to Holen (1996:69). Mammoth bone from the site, not the proposed bone tool, dated to 18,250 �90 years B.P. (Holen 1996:70).

Several other megafaunal bone sites have produced pre-Clovis dates. The La Sena mammoth kill site in Nebraska has produced 18,000 �190 and 18,440 �145 years B.P. dates from bone collagen (Bonnichsen and Turnmire 1999:12). The mammoth bones at the site are disarticulated and the long bones exhibit green bone fractures. At the Burnham site, in Oklahoma, chipped stone flakes were found in sediments with extinct bison faunal remains and carbon dating to between 26,000 and 40,000 years B.P. (Wyckoff, et. al., 1990:60-62, Wyckoff 1999:356). Nearly 60 "artifacts," mostly retouched flakes, were recovered from a deposit between 28,000 and 32,000 years old (Wyckoff 1999:357). Three field seasons consistently yielded human artifacts from deposits more than 25,000 years old, including 28 flakes from 20 different tools (Wyckoff, et. al 1990:62). Levi Shelter, Texas, has yielded bone tools, chipped stone tools and numerous flakes along with a 12,800 year B.P. radiocarbon date (Wyckoff 1999:349). Bonfire Shelter, Texas, has produced a 12,460 � 490 year B. P. date from charcoal flecks in a bone bed with circumstantial evidence of human activity (Wyckoff 1999:349).

The Cactus Hill site in Virginia appears to have "a radiocarbon dated archaeological manifestation temporally and stratigraphically below Clovis" (Goodyear 1999:435). In 1993, below a Clovis floor with two fluted points and seven quartzsite flakes, a charcoal concentration with quartzsite flakes and quartz core blades was AMS dated to 15,070 � 70 years B.P. (Bonnichsen and Turnmire 1999:16, Goodyear 1999:435). Later excavations produced a blade cluster associated with a soil sample dated to 16,670 � 730 years B.P. (Goodyear 1999:435).

Two Mexican sites continue, years after their excavation, to be proposed at valid pre-Clovis sites. At Tlapacoya, Mexico, presumed hearths associated with animal bone middens were dated to 24,00 � 4000 and 21,700 � 500 years B. P. (Lorenzo and Mirambell 1999:488-489). Lorenzo and Mirambell (1999:489) recently concluded that "these dates provide clear evidence of human activity at the site about 22,000 years ago." At the site a prismatic obsidian blade found under a tree trunk dated to 23,950 � 950 years B. P. was obsidian hydration dated to between 21,250 and 25,000 years B. P. (Lorenzo and Mirambell 1999:489).

Between 1977 and 1984, at El Cedral, San Luis Potosi, Mexico, excavation of a spring with an abundance of faunal remains produced, from a stratum dated to 33,300 � 2700 years B.P., a circular scraper manufactured from microcrystalline quartz by direct percussion (Lorenzo and Mirambell 1999:491). A limestone core was recovered from a stratum dated at 15,000 years B.P., and the nearest identified limestone source is 5 km distant (Lorenzo and Mirambell 1999:491). A hearth containing a charcoal lens dated to 31,850 � 1600 years B.P. and ringed with proboscibian tarsi "leave no room for doubt that it is the product of human activity," according to Lorenzo and Mirambell (1999:492). Three other roughly contemporaneous hearths dated to between 26,000 and 28,000 years B.P., while the two hearths producing the sites limiting dates dated to 37,694 � 1963 and 21,468 � 458 years B.P. (Lorenzo and Mirambell 1999:492). The seven superimposed hearths varied in size from 60 to 170 cm in diameter and contained many small fragments of burned bone (Lorenzo and Mirambell 1999:493).

South America

Monte Verde, Chile, is to South America as Meadowcroft is to North America. Meltzer wrote (1995:21), "The Monte Verde site in Chile is the most viable pre-Clovis candidate, although for now neither it nor any other site resolves when or by which route humans first came to the Americas." According to Adovasio (the excavator of Meadowcroft) and Pedler (1997:573), Monte Verde, with its exceptionally well preserved organic material and artifacts radiocarbon dated to between 12,500 and 13,000 years B.P. "may prove to be the seminal archaeological site that will finally prevail over the Clovis-first model."

Monte Verde was excavated from 1977 to 1985 and analyzed by more than 70 collaborating researchers. Monte Verde contains four discrete zones of buried cultural material and two components. The first, or basal, component, Monte Verde I, produced a radiocarbon determination of 33,370 �530 years B.P., associated with stones not yet shown to be artifactual (Roosevelt, et. al. 1996:363). The second component, Monte Verde II, has produced more than 30 radiocarbon determinations averaging 12,500 to 13,000 years B.P. (Adovasio and Pedler 1997:574).

Monte Verde II contains a wide variety of well preserved perishable materials, including wood, plants, bone, and hide (Dillehay 1999:210). Remains include grinding stones, modified stone, unifacial and bifacial flakes, and bola stones. More than 70 species of plants were recovered from the site. The site remains were transported from a variety of coastal and mountain habitats adjacent to the river basin (Dillehay 1999:210). The Monte Verde evidence also includes a human footprint from a living floor, cordage and knotted fiber.

Because of the debate over the dating of the Monte Verde site, a group of archaeological specialists visited the site to assess its validity. They had to examine the artifacts, study the site’s stratigraphy and evaluate the reliability and internal consistency of the radiocarbon determinations. Their on-site visit focused on depositional circumstances, confirmation of stratigraphy, validation of depositional context, anthropogenic reality of alleged living structures, and potential sources of carbon contamination at or near the site (Adovasio and Pedler 1997:577). All the specialists agreed that there were indisputable human artifacts in a primary depositional context mantled by a peat layer and without indication of materials deriving from later depositional horizons at the site (Adovasio and Pedler 1997:578). The specialists reviewed the Monte Verde collections, geology, stratigraphy and chronology before reaching a consensus that Monte Verde "is a bona fide site" (Bonnichsen and Schneider 1999:497).

No other site in South America exhibits the preservation and number and kind of artifacts at Monte Verde, so it is not surprising that the other sites claiming pre-Clovis antiquity are also questioned. Several open sites and rock art sites in Brazil with flaked stones, painted rock spalls, and possible hearths have produced numerous, consistent radiometric date extending to 50 kya, but human presence this early, and the evidence offered to support it, continues to be questioned (Roosevelt 1996:374). The following sites are discussed in recent publications.

In 1979 Laming-Emperaire reported results of excavations at Lapa Vermelha IV, Minas Gerais, Brazil, with quartz and limestone artifacts in a fill dated to 15,000 �400 years B.P., and in a sedimentary zone dated at 22,410 and greater than 25,000 years B.P. (Gruhn 1997:29). Also in Minas Gerais, Lapo de Boquete rockshelter has produced a basal assemblage with unifacial flake tools with four charcoal radiocarbon dates between 12,070 and 11,000 years B.P. (Gruhn 1997:29). The Abrigo de Santana do Riacho produced a hearth dated to 11,960 �250 years B.P. in association with quartz flakes (Gruhn 1997:29).

In coastal Venezuela the Taima-Taima site indicates use of points in hunting mastodon by 13,000 years B.P. (Gruhn 1997:29). The evidence includes butcher marks on the mastodon bone. The 19 dates from wood, bone and mastodon bone collagen range from 14,400 to 11,860 years B.P. (Roosevelt, et. al. 1996:363) are questioned because of possible contamination from lignite and limestone.

Other very ancient dates from South America include 11,740 �110 at Tibitó, Columbia, 11,800 �930 at Pachamachay, Peru, 11,600 �190 at Quereo, Chile, and 12,600 �600 at Los Toldos, Patagonian Chile. Vialou, et. al., reported finding stone artifacts and modified wood with radiocarbon dates of 22,500 �500 and 23,320 �1000 years B.P. from the Santa Elina rockshelter in Mato Grosso, Brazil (Gruhn 1997:29).

Many pre-Clovis dates are outliers, including Abrigo do Sol, Brazil with three dates between 14.7 and 12.3 kya distributed among 24 dates after 11.6 kya, Ibiqui sites with two dates on plant remains with flakes around 12.7 kya, a single charcoal date of 12,560 in Guiterrero Cave, Peru, and charcoal and megafauna dated to 12,795 and 12,360 years B.P. at a Peruvian coast open site with 21 other dates after 10,430 years B.P. (Roosevelt, et. al. 1996:363).

At Pedra Furada, Brazil, modified stones hint at a possible human presence. Pedra Furada illustrates the scenario of natural forces creating alleged stone tools. Pleistocene strata at the site, consistently dated to 50 to 17 kya (42 dates), allegedly contained quartzite cobble tools. Nearby are similarly flaked geofacts recently fallen from 100 m above.

SHIFTING PARADIGMS?

While most Clovis-First proponents remain unpursuaded that a paradigm shift is about to occur, Goebel (1999:208) writes:

Goebel now presents two alternative hypotheses for the peopling of Beringia and the Americas, one via the mammoth steppe before the last glacial maximum 25-20 kya, and colonization during the last glacial maximum 14-12 kya (Goebel 1999:224). Bonnichsen and Turnmire (1999:2-3) perceive a paradigm shift towards acceptance of the Early-Entry model predicated on the acceptance of Monte Verde, Chile, as a 12,500 year old site, identification of regional co-traditions contemporaneous with Clovis and several other pre-11,500 sites in the Americas. They also point out that the Early-Entry model does not nullify the possibility that a Late-Entry migration is correct (Bonnichsen and Turnmire 1999:3).

CONCLUSION

I would like to think that what is needed to establish a pre-Clovis occupation is a site with complex material remains in undisturbed geological context with radiocarbon dated material assuredly associated with human activity. Meadowcroft and Monte Verde certainly offer that and more, while many readily accepted Clovis sites could not pass this test. The number of claimed pre-Clovis sites that have failed this test certainly influences the perception of new claims. Meltzer (1995:33) wrote: "The first site to break the Clovis barrier will have to be utterly unimpeachable in all respects." In my view there will not be a paradigm shift, a new general consensus, until at least several unimpeachable, new and comparable sites join Meadowcroft and Monte Verde.

Even if new evidence surfaces and is widely accepted as the oldest dated site or cultural manifestation, it will be just another minimal indication of actual human antiquity in the Americas. The mystery and excitement surrounding the questions about the first Paleoamericans will outlive the next paradigm. Our vision of the past is necessarily incomplete, hence an expectation of agreement about the past seems unrealistic. The timing and mechanisms by which the New World was peopled will probably always be debated.

Extinct fauna mural image: Permission courtesy Museo Nacional
de Antropología, Mexico City. Download the deskpicture file.