Early colonization of Beringia and Northern North America: Chronology, routes, and adaptive strategies

Abstract

Recent archaeological and paleoecological work along both interior and coastal routes for early colonization of the New World has suggested that the interior route was impossible, leaving the coastal route as the only colonization route taken by Clovis ancestors. We review the geological, paleoecological, and archaeological record for Eastern Beringia and adjacent areas. Spatio-temporal patterning of known sites and evaluation of early interior and coastal route radiocarbon, luminescence, and cosmogenic dating, along with new analyses of obsidian distribution and adaptive strategies of early Beringians, indicate this assessment is premature and the interior route remains a viable hypothesis, available by at least 13,560 and possibly 14,900 years ago.

Introduction

The early colonization(s) of the Americas represent important avenues to explore human ecology, human-climate interactions, humans' role in extinctions, and adaptive capacities. Given current disagreements in archaeology, genetics, and interpretations of paleoenvironmental patterning, clarifying the nature, number, timing, and environmental contexts of these colonization events is critical. There are several points of widespread consensus that help guide this paper. The first Americans and/or the ancestors of the first widespread cultural traditions south of the Laurentide and Cordilleran Ice Sheets originated in Northeast Asia and expanded into Beringia before migrating farther south into North and South America. These populations were genetically related most closely to modern East Asians, but with significant admixture from a now-vanished Ancient North Eurasian population of interior Siberia (Raghavan et al., 2015). However, recent papers have asserted certainty for the coastal route of colonization and/or the impossibility of the interior route (Erlandson, 2013, Erlandson et al., 2015, Madsen, 2015), at least prior to 13,000 cal yr BP (Heintzman et al., 2016) or 12,600 cal yr BP¹ (Pedersen et al., 2016). Other studies of human genetics have made similar assertions, though providing no geological or archaeological evaluation of the data (Llamas et al., 2016, Skoglund and Reich, 2016). We contend these assertions of certainty are premature and are inconsistent with a comprehensive understanding of the extant data.

In this paper, we evaluate the Beringian archaeological record, including recent data generated by ongoing research programs directed by Potter, Holmes, and Reuther, in terms of chronological and spatial patterning of the earliest occupations and broad economic, technological, settlement systems and habitat use. We then provide a systematic evaluation of extant data from both hypothetical routes for entry into central North America, the Ice Free Corridor (interior route) and the Northwest Pacific Coast (NPC, coastal route). We test expectations for both routes for early (pre-Paleo-Eskimo) migrations and regional chronologies within Eastern Beringia and adjacent NPC areas. A clearer understanding of the ecological adaptations of early Beringians can provide us with avenues to link broad environmental change with human responses, including expansion, contraction, and land use shifts. We review the spatiotemporal distribution of archaeological sites in Beringia and adjacent alternative routes. We evaluate obsidian distributions in both coastal and interior regions with respect to expectations for coastal or interior colonization models. We also identify where we currently lack knowledge and propose that future research focus on these areas of ambiguity.

For this review, we must be clear about the colonization event(s) under consideration. Many previous colonization studies assume that Clovis complex-using populations were responsible for the initial peopling of the New World, the “Clovis First Model” (Meltzer, 2009:4–5). On one hand, there is support for this model from genetic evidence, which suggests the initial peopling of the New World by a single relatively small population (Raghavan et al., 2015) and a rapid colonization from north to south (Barton et al., 2004, Hamilton and Buchanan, 2007). Similarly, decreasing heterozygosity from north to south is consistent with a north-south movement (Wang et al., 2007), though it could be due to differential gene flow with Asians (Tamm et al., 2007). On the other hand, there are a growing number of sites proposed to pre-date the Clovis culture (Adovasio et al., 2013, Collins et al., 2013, Erlandson, 2013), although they are not without their critiques (Dincauze, 1984, Fiedel, 2013, Haynes and Huckell, 2016, Meltzer, 2009: 95–124; Morrow et al., 2012).

Consequently, there has been a renewed interest in examining Clovis lithic technology across North America to re-visit some of the central tenets of the “Clovis First Model.” For example, one of the key aspects of the Clovis First Model is the similarity in lithic technology across North America, seen particularly in the production of prismatic blades and fluted bifaces (Bradley et al., 2010, Collins, 1999, Eren and Buchanan, 2016). However, while fluted bifaces are remarkably consistent across areas of North America, in some cases analysts have detected subtle variations in their production and final shape that have been interpreted as “cultural drift” derived from a founding population (Eren et al., 2015, Smith et al., 2014). In the Great Lakes and northeastern United States, people using Clovis technology and later post-Clovis groups were clearly the first people to move into areas that had been recently deglaciated (Eren, 2013, Lothrop et al., 2016).

However, what is less clear is the origin of Clovis technology and its relationship, if any, with purported pre-Clovis sites. Pre-Clovis sites vary in their integrity and chronological controls, but there is as yet little consensus as to their technological, typological, or derived historical connections with later widespread cultural traditions (Collins et al., 2013, Fiedel, 2013, Waters et al., 2013; but see also Adovasio et al., 2013). These sites may represent earlier migrations that left little or no genetic contribution to recent/modern Native Americans, or in other words “failed migrations” (Meltzer, 1989:472).

For our purposes, we do not assume that Clovis represents the earliest colonization. Instead, we contend that the Clovis complex and related Paleoindian materials are the first widespread cultural manifestation south of the Ice Sheets. Based on a sample of 13 sites, Waters and Stafford, 2007, Waters and Stafford, 2014:544) contend that the Clovis complex spans “11,080 ± 40 ¹⁴C yr BP (Lange-Ferguson, SD) to 10,705 ± 35 ¹⁴C yr BP (Anzick, Montana).” They question the accuracy of the radiocarbon dates and/or their association with artifacts at four sites (Aubrey, Texas; Casper, Wyoming; East Wenatchee, Washington; and Sheaman, Wyoming). Haynes et al. (2007), however, argue that the dates from Aubrey should not be summarily dismissed. Since then an additional site, El Fin del Mundo, a proboscidean kill site in Sonora with associated Clovis points, has been argued to pre-date this age range (Sanchez et al., 2014). Moreover, given the vagaries in the radiocarbon calibration curve associated with the onset of the Younger Dryas (Fiedel, 2014), and the statistical uncertainty associated with using a small number of sites to extrapolate an age range for Clovis (Prasciunas et al., 2015), there is basically a short chronology (12,940–12,680 cal yr BP [11,050 to 10,800 ¹⁴C yr BP]) and a long chronology (∼13,400–12,680 cal yr BP [11,600–10,800 ¹⁴C yr BP]) for the duration of the Clovis cultural complex (Miller et al., 2013).

Technological analyses indicate technological links between the Siberian Upper Paleolithic and Clovis complex (Straus et al., 2005), and we concur. There is a similarly robust record of human occupation in Eastern Beringia dating from 14,000 cal yr BP (Potter et al., 2013), with clear technological linkages to the Siberian Late Upper Paleolithic (e.g., Yi et al., 1985, Holmes, 2011). Similarly, we concur with studies that argue that Fishtail points from South America post-date the Clovis complex (Waters et al., 2015) and are likely derived from it as well (Morrow and Morrow, 1999; Pearson, 2004:91). Our goal in this article is to evaluate the colonization of Eastern Beringia and the route(s) used by Clovis and Fishtail complex progenitors.

Recent genetic research focused on both mtDNA (Llamas et al., 2016; this volume; Achilli et al., 2013, Perego et al., 2009) and nuclear genomic data (Raghavan et al., 2015, Skoglund et al., 2015) has transformed our understanding of population divergence, admixture, and migration into the Americas. Currently, data suggest that all modern Native Americans (excluding Eskimo populations) draw the vast majority of their ancestry from a single genetically diverse founding population (Hey, 2005, Raghavan et al., 2015, Reich et al., 2012, Zegura et al., 2004). Coalescence estimates for modern Native American haplogroups range from 25 to 15 kya, assuming different mutation rates (Llamas et al., 2016, Perego et al., 2010).

As genetic studies are reliant on modern populations and a few ancient samples, they can directly provide only limited information on the geographic locations of these ancient populations, how they changed in time, and routes of migration. The distributions among modern peoples of three mtDNA haplogroup, X2a and C4c centered at the southern end of the former IFC and D4h3a concentrated along the Pacific coast, have been hypothesized to reflect separate migration pulses (Perego et al., 2009, Kashani et al., 2012). As coalescence ages for these three haplogroups are similar, they may be evidence of two coterminous dispersal routes from Beringia to central North America. We note, however, that while D4h3a has been argued to reflect a remnant of an early coastal migration (Perego et al., 2009), the oldest occurrence of D4h3a is found within a Clovis individual (Anzick), located on the eastern side of the Rocky Mountains (Rasmussen et al., 2014). Based on the ages of Anzick and the younger Shuká Kaa individual, and the modern coastal distribution, these data are more consistent with a later (post-12,600 cal yr BP) expansion from the North American interior to the Pacific coast.

mtDNA analyses indicate divergence of Native American ancestors from East Asian ancestors and a long period of genetic differentiation (∼40–16 kya) prior to rapid colonization of/expansion in the New World around 16–12 kya (termed the ‘Beringian incubation or standstill model’) (Tamm et al., 2007, Mulligan et al., 2008; Faught, this volume). Later studies have narrowed this period of isolation to between 24.9 and 18.4 kya and 16 kya (Llamas et al., 2016 and this volume). The location of this isolation is unknown, but as (1) there is no evidence for human occupation in Chukotka, Kamchatka, or Eastern Beringia prior to ∼14,000 cal yr BP and (2) there is substantial evidence for widespread human occupation in Northeast Asia throughout the period after 40,000 cal yr BP, particularly in Southern Siberia (Goebel, 2002), we suspect that this isolation occurred farther west in Asia (for similar arguments, see Madsen, 2015, Buvit and Terry, 2016). Hoffecker et al. (2016) speculate this isolation may have occurred in the currently inundated part of Central Beringia.

Nuclear genomic analyses indicate a single ancestral Native American population, though with a deep divergence that predates Anzick, separating branches termed Northern Native Americans (Algonkians and the majority ancestry of Athabaskan groups) and Southern Native Americans (all other peoples in North, Central and South America) (Rasmussen et al., 2014). Reich et al. (2012) found that Native Americans descended from a single source population, while Na-Dene (e.g., Chipewyan) had evidence for limited admixture (∼10%) from an Asian source, different from that in Eskimo populations. Genomic studies also indicated that Native American ancestors entered no earlier than 23 kya after no more than an 8000 year period of isolation from East Asian populations (Raghavan et al., 2015, Llamas et al., 2016). However, some Amazonian Native Americans have been found to carry minor additional ancestry from a source most closely related to modern day Australasians (Reich et al., 2012, Raghavan et al., 2015, Skoglund et al., 2015). The spatial distribution and intensity of this ancestry component suggest two alternative hypotheses: (1) a recent migration from an Australasian-related source that only reached certain Amazonian groups, or (2) that the Native American ancestral population was structured and carried differential relatedness to separate Asian sources (Skoglund and Reich, 2016); these Amazonian groups, in turn, derive part of their ancestry from a distinct component of such ancestral population. At present, the timing and full nature of this Australasian signal is unknown (see Fiedel, this volume).

In sum, genetic studies are consistent with a scenario of an ancient separation (>20 kya) between ancestral Native Americans and East Asians, a population bottleneck between ∼18 and 15 kya (Skoglund and Reich, 2016) which took place either in Beringia or northeast Asia, followed by a rapid expansion consistent with a relatively late migration to/within the New World. Clovis (Anzick) has been linked to the southerly of the two basal branches (Rasmussen et al., 2014), Southern Native Americans. The other basal branch, Northern Native Americans, appears restricted to populations in Northern North America. Most genetic data suggest a single population pulse (Raghavan et al., 2015), lending support for the hypothesis that the earliest widespread cultural manifestation south of the ice sheets, Clovis and related complexes, represents the material culture used by this group after the time they migrated south of the ice sheets.