The geographic origin of Clovis technology: Insights from Clovis biface caches

Abstract

Multiple hypotheses have been advanced for the geographic origin of the Clovis technocomplex. Several competing hypotheses are considered in relation to the distribution of Clovis caches. Clovis caching behavior is interpreted as a strategy for maximizing exploration and migration rather than an embedded strategy associated with an annual foraging round. Based on this analysis, it is hypothesized that the Clovis technocomplex may have originated along the North Pacific coast or south of the Cordilleran ice sheet in the Puget Lowland before rapidly spreading across the continent.

KEYWORDS:

• Paleoindian
• fluting technology
• Clovis cache
• migration
• Western North America
• Clovis origins

Clovis is the label used for the earliest widespread, identifiable prehistoric technocomplex south of the Late Pleistocene Cordilleran and Laurentide ice sheets in North America. The Clovis tradition is short-lived and currently dated from 13,250–12,700 cal BP (Goebel and Keene 2014; Waters and Stafford 2007, 2013). A variety of site types are associated with the Clovis tradition, including encampments, quarry sites, megafauna kill or butchering sites, and cache sites.

The Clovis complex includes a well-developed bifacial lithic technology including projectile points, knives, gravers, and bifaces of various stages of manufacture (Ardelean 2014; Stanford and Bradley 2012:31–66). Clovis lithic assemblages may also include end and side scrapers manufactured from both flakes and blades. Bone implements in Clovis assemblages include shaft wrenches, beveled rods, and possible ivory points. A robust perishable industry involving fiber, wood, and hide is assumed to be a component of the Clovis complex, even though none of these kinds of perishable artifacts have yet been recovered from Clovis sites.¹

The most distinctive and diagnostic artifacts of the Clovis complex are the large unifacially or bifacially fluted points that are among the largest lanceolate projectile points in North America. The largest specimens can exceed 10 cm in length and are often found in caches and occasionally in kill sites (Buchanan et al. 2012).² Very large projectile points are a necessary component of a tool kit to hunt the large herbivores and megafauna that were the primary prey of Clovis groups (Haynes and Hutson 2014).

Clovis caches

A Clovis cache consists of a localized assemblage of artifacts that is intentionally set aside with an expectation that they would be recovered at a later date. Caches dating to the Clovis period are recognized by a combination of methods including chronometric dating, diagnostic artifacts, and distinctive toolstone reduction technology (Osborn 2016:162). Although Clovis caches are rare, they are widely distributed from the Columbia Plateau to the Great Plains (Huckell and Kilby 2014a; Kilby and Huckell 2014).

Clovis caching behavior was as short-lived as the tradition. Within several hundred years, Clovis technology gave rise to the Folsom complex. There is no evidence that extensive caching behavior was a part of the Folsom technological tradition (Collard et al. 2010:2513), although caching of finished artifacts occurred among late Paleoindian populations in western North America (e.g. Amick 2004; Davis et al. 2017). Consideration of the kinds of caches, their geographic distribution, and the source locations for some of the toolstone present in the caches can provide insights into the patterns of migration of Clovis groups more than 12,000 years ago and may, in turn, shed light on the geographic origins of Clovis technology.

The presence of fluted points among a Clovis cache assemblage or a radiocarbon date confirms the cache as Clovis in age. Other Clovis caches without fluted points have been identified as Clovis based on other attributes. Huckell and Kilby (2014b) enumerate 23 caches as Clovis. Since then, two additional Clovis biface caches have been reported, the Baller cache in Nebraska (Osborn 2016) and the Loa cache in Utah (Schroedl 2019).

Two other potential Clovis biface caches have been identified from southern Idaho near American Falls by Speer and colleagues (Speer et al. 2019). These caches were first reported in the 1960s. One cache, the Higgins cache, consisted of 3 large bifaces of quartzite, and the other is a cache of nine obsidian bifaces. These potential Clovis biface caches were never analyzed and their whereabouts are currently unknown. Also, Perlmutter (2015) describes the Bijou Creek biface cache found previously in northeastern Colorado. This cache consists of five large quartzite bifaces, some of the largest found in Colorado. This cache may be Clovis, but Perlmutter did not provide an assessment of cultural or temporal association. Neither the Bijou Creek cache nor the two bifaces caches from Idaho are considered confirmed Clovis caches and are not included in this analysis. Figure 1 depicts the geographic location of all of these caches, Table 1 enumerates them.

Figure 1 Map of Western North America showing the location of confirmed and unconfirmed Clovis biface caches and Clovis-attributed blade caches. The three cache regions discussed in the text are demarcated. The ice sheet boundaries, depicted in white, are approximate for the time span between 14,800–14,100 years ago (adapted from Potter et al. 2018 and Dyke et al. 2003). Key to cache locations listed in Table 1.

Table 1 List of Confirmed and Unconfirmed Clovis Biface Caches and Clovis-Attributed Blade Caches

Cache	Location	Cache Type	Reference	SiteNumber	Key to Map(Figure 1)
Anzick	Montana	Biface with projectile point(s)	Wilke et al. (1991)	24PA506	1
Baller	Nebraska	Biface	Osborn (2016)	-	2
Beach	North Dakota	Biface	Kilby and Huckell (2014)	-	3
Busse	Kansas	Biface	Kilby (2008)	-	4
Carlisle	Iowa	Biface	Hill et al. (2014)	13WA105	5
Crook County	Wyoming	Biface with projectile point(s)	Kilby (2008); Tankersley (1998, 2002)	-	6
CW	Colorado	Biface	Muniz (2014)	-	7
deGraffenreid	Texas	Biface with preform(s)	Collins et al. (2007)	-	8
Drake	Colorado	Biface with projectile point(s)	Stanford and Jodry (1988)	5LO24	9
East Wenatchee	Washington	Biface with projectile point(s)	Gramly (1993)	45DO482	10
Fenn	Wyoming, Utah, or Idaho	Biface with projectile point(s)	Frison and Bradley (1999)	-	11
Hogeye	Texas	Biface with preform(s)	Waters and Jennings (2015)	-	12
Loa	Utah	Biface with preform(s)	Schroedl (2019)	42WN1674	13
Mahaffy	Colorado	Biface	Bamforth (2014, 2015)	-	14
Rummells-Maske	Iowa	Biface with projectile point(s)	Morrow and Morrow (2002)	13CD15	15
Simon	Idaho	Biface with projectile point(s)	Butler (1963); Kohntopp (2001)	10CM7	16
Watts	Colorado	Biface	Patten (2015)	5LR10957	17
Anadarko	Oklahoma	Blade	Hammatt (1970); Kilby (2008)	-	18
Dickenson	New Mexico	Blade	Condon et al. (2014)	-	19
Franey	Nebraska	Blade	Grange (1964); Kilby (2008)	-	20
Green	New Mexico	Blade	Green (1963); Kilby (2008)	-	21
JS	Oklahoma	Blade	Bement (2014)	34BV180	22
Keven Davis	Texas	Blade	Collins (1999)	-	23
Pelland	Minnesota	Blade	Stoltman (1971); Kilby (2008)	-	24
Sailor-Helton	Kansas	Blade	Mallouf (1994); Kilby (2008)	-	25
Bijou Creek	Colorado	Potential Biface	Perlmutter (2015)	-	26
Higgins	Idaho	Potential Biface	Speer et al. (2019)	-	27
Unnamed	Idaho	Potential Biface	Speer et al. (2019)	-	28

Based on assemblage contents, caches identified by Huckell and Kilby as Clovis can be divided into three categories, caches dominated by large bifaces and projectile points, caches dominated by large bifaces (some with preforms), and caches dominated by blades (Kilby 2015). Huckell and Kilby’s assignment of caches dominated by blades as Clovis is questionable as discussed below.

Putative Clovis blade caches

Huckell and Kilby (2014b) identify eight blade caches as Clovis: Anadarko, Dickenson, Franey, Green, JS, Keven Davis, Pelland, and Sailor Helton. Most of these blade caches are from the Southern Plains (Figure 1). The location of the Pelland cache in northern Minnesota appears to have been under the Laurentide Ice Sheet and may not have been accessible during the Clovis period, likely eliminating this cache as Clovis. A review of these blade caches indicates there is no associated chronometric dating or temporally diagnostic artifacts that would allow the unconditional assignment of these caches to the Clovis complex. Eren and others (2018) provide a general assessment of blade technology at an Archaic site in Oklahoma. They demonstrate that blades from Archaic caches in the region are statistically indistinguishable from blades within blade caches classified as Clovis (lacking radiocarbon dates or diagnostic artifacts) on the Southern Plains.

While there is no doubt there is a well-developed blade technology apparently dating to the Clovis period in southern Texas (Collins 1999), it does not directly follow that caches of blades from the Southern Plains must be Clovis, especially if they lack diagnostic implements or radiocarbon dates. Based on Eren and others’ (2018) analysis it is questionable whether blade-dominated caches from this region without associated dates or diagnostic Clovis artifacts should even be classified as Clovis caches. As discussed below, one hypothesis about the origin of the Clovis technology relies heavily on the presumption that these blade caches are Clovis in age and affiliation.

Clovis biface caches

There is a total of 17 confirmed Clovis biface caches from North America. There are seven Clovis bifaces caches with associated diagnostic projectile points and one of these, the Anzick cache (24PA506), has been radiocarbon dated to the Clovis period. There are ten other Clovis biface caches without diagnostic points (excluding the Bijou Creek cache and the two potential caches from Idaho). One of these, the Beach cache (Huckell and Kilby 2012), has been radiocarbon dated to the Clovis period. These ten biface caches can still be confidently identified as Clovis.

Schroedl (2019) compares the maximum length of bifaces in Clovis biface caches with Clovis points with the maximum length of bifaces in Clovis biface caches without diagnostic artifacts and determined that these two groups of large bifaces from the caches are statistically indistinguishable from each other along the dimension of maximum length. Three of these biface caches without diagnostic points contain preforms. The preforms had an average length 1 cm greater than projectile points in the caches with finished points. Thus, caches with exceptionally large bifaces and preforms, even without diagnostic projectile points, can still be identified as Clovis. Because Clovis points often exceed 10 cm in length, obviously creating such large points required an even larger piece of raw material, as reflected in these large bifaces.

Function of Clovis biface caches

Many researchers have attributed the deposition of Clovis biface caches to ritual behavior. Wilke et al. (1991:267, 268), noting that the Anzick cache contained human remains, posit that the known Clovis caches at the time were possibly “teaching kits” for use in the afterlife. Kilby (2008:216–219, 245; 2014:212) also argues that East Wenatchee, Simon, Fenn, and Anzick caches represent some kind of ritual afterlife cache. However, the lack of human remains within all the Clovis biface caches except Anzick undermines this hypothesis. We have little understanding of Clovis mortuary practices, but the lack of human remains among the caches suggest that Clovis caching behavior was not tied to human interments.

Gillespie (2007:183) believes that the caches involved the ritual burial of projectile points to assist Clovis groups with developing a “fixed sense of the landscape”. However, other artifact classes are more common in Clovis biface caches than projectile points. Another speculative and untestable function offered for the Simon cache is that it was a votive offering deposited to ward off natural disasters such as volcanic eruptions (Kohntopp 2010:38). Finally, the lack of commonality of the various ratios of implements, bifaces, and blades in these caches fails to support the hypothesis that Clovis caching is a result of ritualized behavior as proposed by a number of researchers (Kilby 2008; Lassen 2005:151; Morrow 2019:193). There is little evidence in the archaeological record that Clovis groups engaged in ritual behavior. Also, overlooked in these speculations is that flakes, biproducts of lithic reduction, are often incorporated into these caches, a highly unlikely form of ritual behavior.

A more parsimonious interpretation is that these caches had a secular utilitarian function of temporary storage of implements and toolstone in anticipation of future needs. According to Buchanan and others (2012) cached projectile points were intended for future use as hunting weapons, not ritual objects. In aggregate, these caches were utilitarian and were created to supply additional implements and toolstone in the form of bifaces and sometimes flakes for further reduction into tools. Cached at known locations, these assemblages served as stores of raw material and additional finished implements.

The distribution of toolstone in Clovis caches has been a cornerstone of discussions about the movement of Clovis groups across the landscape (Beck and Jones 2010; Kilby 2008, 2014; Pitblado 2017). Most of these discussions are based on an analogy to foragers who moved in a cyclical annual round returning to the same locations year after year (Steward 1938). But this historic analogy is not appropriate for Clovis subsistence practices which were first and foremost oriented toward hunting of large mobile herbivores and megafauna (Haynes and Hutson 2014). Unlike modern hunter gatherers tethered to home bases, Clovis groups, tracking and utilizing locally abundant resources, did not need to return to any fixed location (Haynes and Hutson 2014:303).

Hoppe and colleagues (1999) provide a method to reconstruct the migration patterns of extinct proboscideans and determined that these large animals may have migrated within a range varying from 120 km to 300 km in southeastern North America. Hoppe (2004) provides an analysis of a limited number of Clovis-age mammoth sites in the Plains region. She notes that mammoths traveled in family groups, although Clovis hunters only killed or scavenged mammoths on an individual basis. According to Hoppe, mammoth family groups ranged in this region more than 200 km. Clovis groups could have followed their prey over hundreds of kilometers without ever returning to a specific base camp.

Additionally, the lack of stratified deposits at a majority of Clovis sites in western North America does not support a hypothesis of Clovis groups returning for seasonal reoccupations of campsites. Given the short duration of the Clovis period and wide geographic distribution of Clovis sites across the continent, there may not be any ethnographic analog of the behavioral adaptation Clovis people used to subsist, explore, and migrate across the continent in such a short period of time.

It is argued here that Clovis biface caches, excluding putative Clovis blade caches, were a behavioral strategy that allowed Clovis groups to rapidly move across the landscape tracking large game and exploring new territories. Without knowledge of toolstone sources ahead of them these caches represented insurance policies to provide access to finished tools and high quality toolstone for reduction into additional tools. The working hypothesis for this analysis follows Osborn (2016). Clovis biface caches were resupply depots that were not lost but were purposefully abandoned as Clovis groups moved across the landscape and discovered new toolstone sources. Osborn notes:

Clovis caches as signatures of colonization are viewed as resupply depots from which early human populations in North America replenished their lithic raw materials if they failed to locate new sources during movement(s) across the landscape. Insurance caches consist of cores, flakes, and bifacial implements placed on a landscape devoid of adequate toolstone. (Osborn 2016:160)

For the purposes of this analysis, the 17 confirmed Clovis biface caches are presumed to have been deposited and then abandoned as Clovis groups explored new regions. As these groups move from one region to the next, they cannot anticipate what new toolstone resources might occur there, so they may have transported and cached additional implements and toolstone as they traveled. If these caches were purposefully abandoned then the distribution of these caches relative to toolstone locations can provide insights into the patterns of Clovis movement and migration.

The rapidity of movement of Clovis groups across the landscape and the lack of established Clovis encampments diminishes the possibility that toolstone procurement was initially a result of active down-the-line trading among Clovis groups. The raw material was most likely obtained by direct procurement by individual Clovis groups rather than through trade or barter. Buchanan and others (2019) assume Clovis toolstone was procured through exchange within existing social networks among groups who had knowledge of local toolstone sources. Yet, as Clovis groups migrated and followed megafauna, toolstone sources would not have been known ahead of time. The proposed network by Buchanan and colleagues is only functional after all regions have been explored and all toolstone sources have been identified and therefore this is not an applicable model for initial colonizing populations.

Toolstone in Clovis biface caches

As noted above, toolstone used to manufacture artifacts in Clovis caches have been used to model movements of Clovis groups. Unfortunately, most of the toolstone identifications are based solely on macroscopic inspection. Highly distinctive siliceous material can sometimes be used to identify toolstone from specific geologic formations such as the banded Tiger chert derived from the Green River Formation found in several of the biface caches (Supplemental Table 1). However, much of the siliceous material in caches cannot be identified unless the toolstone has been subjected to petrographic or geochemical analyses.

Boulanger and colleagues (Boulanger et al. 2015:551) note that visual classifications of toolstone can be problematic because macroscopic toolstone identification can vary from observer to observer based on what they refer to as folk geological taxonomy. Ozbun (2015:2) notes that heated debates regarding the geological or mineralogical nomenclature for toolstone can only be resolved by analyses such as X-ray fluorescence (XRF), laser ablation inductively coupled plasma mass spectrometry (LA-ICP-MS), inductively coupled plasma mass spectrometry (ICP-MS), X-ray diffraction (XRD), neutron activation analysis (NAA), or near infra-red spectrometry (NIRS).

These analytical techniques are more likely to provide replicable scientific identification of Clovis toolstone than visual identification. However, toolstone from only two Clovis biface caches have been subjected to scientific analysis. Toolstone from the Beach cache has been identified to the Sentinel Butte source among the White River Group Silicates by INAA (Huckell et al. 2011). Using XRF, the toolstone from the Loa cache was sourced to the Wildhorse Canyon obsidian source (Janetski et al. 1988).

Kilby (2008) provided the first major discussion of toolstone identified in Clovis biface and blade caches. Kilby notes that his identifications are based on visual inspection however, his identifications continue to be repeated in the literature (Bradley et al. 2010) and are used as a basis for elaborate migration models (Beck and Jones 2010; Pitblado 2017). Some of Kilby’s toolstone identifications are challenged below.

Geographic origin of Clovis technology

Since the 1930s, spectacular fluted Clovis points have captured the interest of the public as well as the professionals. Clovis points are the earliest well documented, most widespread, and most easily distinguishable artifact type (Beck and Jones 2010, 82). Decades of research in North America have failed to identify any other archeological complex that is so unique, so widespread, and so old. It is not surprising that there has been an expectation that Clovis groups were the initial founding population in North America. The origin of Clovis technology and identity of the first colonizers of North America has often been conflated, giving rise to the Clovis First hypothesis (Bonnichsen and Schneider 1999; Haynes 1966; Martin 1973). While the Clovis complex is the earliest recognizable complex in North America it does not directly follow that its creators were the first colonizers south of the ice sheets. The following discussion focusses on where biface fluting technology might have originated, but it does not imply that Clovis populations were the first colonizers of North America. The controversy of whether earlier groups may or may not have inhabited portions of the continent prior to the arrival of Clovis groups is not a focus of this paper.

There are several competing hypotheses about where Clovis technology may have originated, each suggesting a different geographic region of origin. The distribution of Clovis biface caches and putative Clovis blade caches and the toolstone identified in these caches may help in evaluating some of these hypotheses.

The Ice-Free Corridor hypothesis (Fiedel 2002:421; Pitblado 2017) has been a common explanation for the initial colonization of the Americas south of the Late Wisconsin ice sheets from Beringia since the 1930s (Bryan 1968:163). Potter and others (2018) assert that the corridor was open beginning about 15,000 years ago. Dawe and Kornfeld (2017) argue that people migrating out of Beringia did not need a completely ice-free corridor to migrate south. Populations could travel among “countless nunataks, not just ice-free, but especially on south facing exposures, snow-free and perhaps vegetated. … Conditions conducive to colonization by plants and animals developed on these nunataks with time” (Dawe and Kornfeld 2017:68). There is no current chronological evidence to support the hypothesis that human populations were present in the corridor and migrating south before the corridor was completely open.

In contrast, Pedersen and colleagues (2016) note that even if the lobes of the Cordilleran and Laurentide ice sheets were partially or fully physically separated, human transit could not occur until the biodiversity in the corridor was sufficient to support migrating populations. Within an ice-free or icy corridor, Clovis groups needed more than just access to intermittent ice-free areas. Large prey would have to be present and access to toolstone and other resources would be necessary to replenish items that were lost or worn out as these groups traveled south. Froese and colleagues indicate the ice-free corridor was not passable “until well after 14,000 cal yr BP and likely until nearer 13,200 cal yr BP” (Froese et al. 2019:32) while others (Pedersen et al. 2016) do not believe it was passable until about 12,600 years ago – too late for a Clovis entrance below the ice sheets from Alaska.

When the ice-free or icy corridor finally opened to allow movement of human populations between Beringia and the northern Great Plains, it appears that Clovis fluting technology moved northward into the Arctic, not southward. Current research (Smith and Goebel 2018) indicates that as the ice-free corridor widened, bifacial fluting technology diffused northward into the subarctic and Arctic. Based on a statistical analysis of fluted projectile point styles, including projectile points from the northern Clovis caches of East Wenatchee, Anzick, and Drake, Smith and Goebel establish that the “ northern fluted complex morphology and technology was not independently invented in the north, but originated proximately in the Ice-Free Corridor, and ultimately from Clovis in temperate North America” (Smith and Goebel 2018:4120). They also note that the fluting technology observed in the north has the greatest affinity to Clovis points in western North America (Smith and Goebel 2018:4119).

Ives and others (2019) also infer that Clovis fluting technology was either carried northward to the subarctic and Arctic by migrating populations or the technology was transferred northward by diffusion. If fluting technology diffused northward or Clovis groups from the Northern Plains carried fluting technology north, then Clovis technology could not have originated in the Arctic nor could it have been carried southward through an ice-free or icy corridor.

Another hypothesis about the specific origin of Clovis technology is that an early maritime-adapted population along the southern coast of Baja California developed bifacial fluting technology and migrated north back along the sea of Cortez and then to the American Southwest via the Colorado River to eventually arrive onto the Great Plains (Anderson et al. 2014). In this Baja California/Sonora/Colorado River model, Anderson and others (2014:188, 189) hypothesize that Clovis groups colonized the Southwest by first reaching the mouth of the Colorado River. From there, some migrants followed the Colorado River upstream and then traversed over the Rocky Mountains to the Arkansas, Platte, and Missouri river systems. Other Clovis groups are presumed to have followed the Gila River eastward to migrate across modern-day Arizona, New Mexico, and southeast into Texas.

If biface caches are an integral part of the initial exploration of continent, there would be an expectation that such caches would be found along this route to the Northwest Plains where Clovis biface caches are common. With the exception of two Clovis biface caches in Texas, the Hogeye and deGraffenreid caches, all other bifaces caches in North America occur above the 38th parallel, far north of the proposed southern entrance into the interior of the continent.

Additionally, Haynes’ (2011) map of Clovis sites and isolated finds in Arizona indicates that Clovis locales are concentrated in the eastern part of the state. No Clovis encampments are noted along the lower Colorado River nor in the dissected topography of the Grand Canyon as would be expected if Clovis groups were moving along the river corridor. Further upstream, no Clovis encampments have been identified along the Colorado or Green rivers in Utah (Copeland and Fike 1988) that would indicate migration from south to north. A southern Baja California origin for Clovis is not supported by the distribution of sites along major river corridors in the southwestern portion of North America nor by the distribution of Clovis biface caches.

Another hypothesis, posited by Bradley and Stanford (2004), is that Clovis originated from a base of Solutrean culture in eastern Europe. This hypothesis has received little support among researchers because there is no substantial physical evidence to support it. Several researchers (O’Brien et al. 2014; Straus 2000) note that beside the several-thousand-year time differential between the end of the Solutrean and the origin of Clovis technology, there is no evidence that Solutrean populations were sufficiently adapted to maritime resources to make the journey to the North American continent or later deposit the Clovis biface caches.

Neither the ice-free corridor hypothesis, the Baja California hypothesis, nor the Solutrean hypothesis provides a satisfactory explanation of the arrival of Clovis technology in North America. There are, however, two other competing hypotheses about the possible origin of Clovis technology that need to be considered. Beck and Jones (2010, 2012) propose that Clovis technology originated in the Gulf Coastal Plains of modern-day Texas and then spread out to the north and northwest until Clovis groups arrived at the coast in the Pacific Northwest. In contrast, Croes and Kucera (2017) are the most recent proponents of the hypothesis that Clovis technology entered the interior of the continent from the Pacific Northwest just south of the receding Cordilleran ice sheet. Each of these hypotheses is discussed in detail below. The geographic distribution of Clovis biface caches, putative Clovis blade caches, and other Clovis sites in the West can only support one of these hypotheses.

Gulf Coastal Plains origin of Clovis technology

Beck and Jones are the primary proponents of the hypothesis that Clovis technology originated on the Gulf Coastal Plains in modern-day Texas near the Aubrey site (Beck and Jones 2010, 2014, 2015). Other researchers (Jennings and Waters 2014; Waters et al. 2018) also believe that Clovis technology originated on the Gulf Coastal Plains but from the Buttermilk Creek complex at the Gault site. The Beck and Jones’ model proposes that a blade-oriented Clovis technology originated in Texas and that bifacial reduction technology increased in importance as Clovis groups expanded north and northwest, migrating from Texas to Montana to ultimately arrive at the Pacific Northwest (Northern route), while other groups made their way westward across the Southwest (Southern route) (Beck and Jones 2010:86, 2012:85, 2015). Presumably, the great distance from Texas to the Pacific Northwest resulted in the late arrival of Clovis groups in the Intermountain West.

In order for Clovis technology to be a late arrival in the Pacific Northwest, Beck and Jones argue that Western Stemmed Tradition (WST) (Bryan 1980) groups had colonized the Pacific Northwest long before Clovis technology arose. They state “The point we want to emphasize here, however, is that people were in the Intermountain West before Clovis was present anywhere” (Beck and Jones 2010:106).

Goebel and Keene (2014) evaluate the radiocarbon record for WST components in the West and conclude that as a technocomplex, WST does not have any greater time depth than Clovis technology; at most, Western Stemmed points may be contemporary or coeval with Clovis points. Reid, reviewing 11 potential locales predating the Clovis time period in Idaho, determined that there is “no convincing support for the temporal priority of the Western Stemmed over the Clovis tradition” (Reid 2017:72). Another recent Bayesian analysis of dates associated with an early Western Stemmed point, the Windust point, confirms that at most, this point type is coeval with Clovis points (Brown et al. 2019:490), undermining the claim that WST predates Clovis occupation in the West.

To bolster support for the Texas origin hypothesis and the late arrival of Clovis in the West, Beck and Jones (2010:95–96; Beck et al. 2019) argue that morphological variation in fluted points from western North America demonstrates that these points are chronologically younger than classic Clovis points east of the Rocky Mountains (Ellis 2013:128; Miller et al. 2014). Neither Pinson (2011) at the Dietz Clovis site (35LK1529) in south central Oregon nor Reid and colleagues (2015:61) analyzing Clovis points in Idaho accept that fluted points in these western regions are separate and distinct from Clovis fluted points. Pinson states the fluted points at the Dietz site “are indeed Clovis points, rather than Sunshine points or any other type of point” (Pinson 2011:290). In an analysis of the continent-wide variability of fluted points, Morrow and Morrow (1999) suggest that technological drift rather than chronological differentiation causes stylistic variations in large fluted points across the North American continent.

Standardized Clovis knapping techniques were likely transmitted through direct interaction between knappers from different groups (Sholts et al. 2012). As highly mobile Clovis groups became more separated in time and space across the West, they may have developed slightly different learning trajectories for lithic reduction and tool production (Eren and Buchanan 2016:6). Technological drift between groups could have occurred over large regions, all during the same time period. In contrast to Beck and Jones’ (2010) assumption, regional morphometric differences of fluted points do not automatically signal a temporal or cultural differentiation for highly mobile populations such as Clovis.

A key component of the Texas origin hypothesis is the distribution of Clovis biface caches on the Northern Plains and putative Clovis blade caches on the Southern Plains (Beck and Jones 2010:88, Figure 3). Beck and Jones (2015:96) argue bifacial reduction technology supplanted core and blade technology as Clovis groups migrated northward. Using a biological analogy, they argue that blade technology eventually became extinct as Clovis groups were forced to rely on bifacial core reduction technology as they moved north due to the decreasing frequency and intermittent availability of toolstone sources.

Northern route

Beck and Jones argue some Clovis groups subsequently migrated westward from the Northern Plains. They claim that “the Rocky Mountains acted as a significant barrier to the west (except to the far south), which explains the distribution of sites [Clovis biface caches] along the eastern front. Clovis technology did, however, eventually penetrate this barrier, as evidenced by the Fenn Clovis cache and … the Simon and East Wenatchee Clovis caches … ” (Beck and Jones 2010:86). Claiming that the geographic location of the Fenn, Simon, and East Wenatchee caches demonstrates Clovis breaching of the Rocky Mountain barrier is example of confirmation bias. It is an untestable assumption that the Rocky Mountains constituted a barrier to Clovis migration and there are no data that demonstrate these caches are chronologically later than those on the Northern Plains. However, the directionality of the toolstone sources represented in these three caches can be used to assess the westward migration hypothesis.

Fenn cache

The Fenn cache (Frison and Bradley 1999) consists of 56 specimens including bifaces, projectile points, a blade, and a chipped stone crescent. The original location of cache is uncertain and can only be assigned to somewhere within the tri-state area where Idaho, Utah, and Wyoming meet. Frison and Bradley (1999) identify the toolstone in this cache as Green River Formation chert (Tiger chert), jasper, Floy Wash agate (Bradley et al. 2010:8), quartz crystal, and obsidian. Green River Formation chert, referred to as Tiger chert, zebra flint, or zebra chert, is available throughout the Green River Basin as is jasper (Hausel and Sutherland 2000:203). Floy Wash agate, sometimes referred to as Utah agate or pigeon blood agate, is found throughout east central Utah. Bradley’s identification of this toolstone source is based only on visual inspection, so some of these specimens could have been derived from local siliceous sources found throughout the Green River Basin (Sutherland 1990; Hausel and Sutherland 2000:201–205).

The three smoky quartz crystal projectile points appear to have been manufactured from smoky quartz crystals from the Sawtooth Range to the west of the cache location (Kilby 2008:155). Nine of the specimens in the cache, including the largest biface, were manufactured from obsidian attributed to a southeastern Idaho source (Frison and Bradley 1999:78), which is at least 100 km west of the presumed cache location. While some of the toolstone represented in the Fenn cache is of local origin, other toolstone in the cache suggests that these Clovis groups had visited sources to the west before depositing the cache.

A single chipped stone crescent was included in the Fenn cache assemblage. Meighan and Haynes (1970:1220) note that crescents may be associated with fluted points in Clovis-age sites in the West, although recent researchers believe this artifact class only dates to between 12,000 and 8,000 years ago (Sanchez et al. 2017). The geographic distribution of chipped stone crescents generally coincides with the distribution of WST sites and rarely occurs elsewhere (Sanchez et al. 2017). The Fenn cache crescent, however, is manufactured from local Green River Formation chert which indicates the knapper had direct, first-hand knowledge of this tool form, knowledge that would not likely have been gained if the knapper came from the Northern Plains. The crescent, the quartz crystal projectile points, and the obsidian artifacts all suggest that the Clovis group who deposited the Fenn cache came from the west rather than the Northern Plains.

Simon cache

The Simon cache (Woods and Titmus 1985) is located in the plains below the foothills of the southern Sawtooth Mountains on the Snake River Plain in Idaho. This cache consists of 33 specimens (Kohntopp 2001:26) including bifaces, points, a convex end scraper, and a flake fragment (Kohntopp 2001:23–33). The Simon assemblage is dominated by a large variety of siliceous materials; Kilby (2008:111) notes as many as 14 different types of siliceous rock including chert, agate, and chalcedony are present in the assemblage. Kilby suggests that the some of the toolstone represented in the Simon cache was collected more than 600 km away on the Northwest Plains (Kilby 2008:158).

Within the Snake River Plain region, Reid and others (2015:57) note the distance that obsidian Clovis points are found from their toolstone source locations averages only 60 km. They question why Clovis groups would travel much farther to obtain non-obsidian toolstone. Specifically, why would they need long distance procurement trips to the Northwest Plains to obtain siliceous rock for caching when there are a number of sources for agate, chert, and chalcedony known from within the region (Idaho Department of Lands 2019) as well as sources in southwestern Wyoming like Green River Formation chert?

Kilby’s visual identifications for the toolstone in the Simon cache are likely incorrect (Santarone 2014:16–19). As noted above, visual characteristics, without petrographic analysis or other analyses, are insufficient to identify toolstone source locations when attempting to model long distance Clovis movement and migration. The sources for the siliceous materials in the Simon cache are unknown, but cannot be assigned to toolstone sources on the Northern Plains without further quantitative analysis.

The Simon cache also includes four large bifaces made from quartz crystals (Kohntopp 2001:27). Quartz crystals of a size necessary to manufacture these large bifaces are only known to have come from the Sawtooth Range, northwest of the cache location. Very large smoky quartz crystals, in excess of 30 cm, were found and photographed by climbers in the higher elevations of the Sawtooth Range in the 1970s (Fritz 2010). It is possible that the Clovis group who deposited the quartz crystal bifaces in the Simon cache were associated with the group who traveled east and abandoned the Fenn cache, which included finished quartz crystal projectile points.

East Wenatchee cache

The East Wenatchee cache, located in western Washington state along the Columbia River, consists of an assemblage of bifaces, projectile points, bone implements, flakes, and blades. The toolstone identified in the East Wenatchee cache includes chert, chalcedonies, and Ephrata agate (Kilby 2008:Table 20a). Waitt (2016:441) notes that the Columbia River Basalt Group has many potential sources of chalcedony and agate. A list by Pattie (1983) indicates that a number of sources of siliceous materials occur in the foothills and mountains above the Kittitas Valley, 30 km south of the East Wenatchee location. The toolstone sources for the East Wenatchee cache appear to be restricted to locally available sources pending further analysis. While the cache does demonstrate Clovis presence in the far Pacific Northwest there is no basis to assume that the Clovis group who deposited the cache arrived from east of the Rocky Mountains as hypothesized by Beck and Jones.

Loa cache

The Loa cache was not available to Beck and Jones for consideration because it was not recognized as a Clovis biface cache until recently (Schroedl 2019). The cache consists of 10 large obsidian bifaces and flakes, each of which was sourced to the Wildhorse Canyon obsidian source. The cache was recovered about 100 km east of the source location, supporting an interpretation of eastward migration rather than westward.

Discussion

The four Clovis biface caches west of the Rocky Mountains do not support a Clovis migration westward from the Northern Plains nor the assertion that the Rocky Mountains were a barrier to Clovis transits. Tiger chert from the Green River Basin is found at the Cook County and Mahaffy caches east of the Rocky Mountains. It strains any logistical provisioning model that westward traveling Clovis groups procured Tiger chert and then traveled back more than 400 km over the Rocky Mountains to deposit these caches before journeying further west. Also, the East Wenatchee, Simon, Fenn, and the Anzick caches, undermine Beck and Jones’ hypothesis that the shift from blade and blade core technology to bifacial core reduction was due in part to lack of access to toolstone sources. The Clovis groups who deposited these caches had access to a large number of different toolstone sources as evidenced by the wide variety of materials visually identified in these caches.

Southern route

Beck and Jones rely on the distribution of putative Clovis blade caches, prismatic blades, and blade cores in the West to support their migration hypothesis (2015:89, Figure 5.3). They identify their southern westward Clovis migration route as starting in north central Texas, moving across southern New Mexico, crossing into eastern Arizona, and entering the edge of the Great Basin in southwestern Utah (2010:90, Figure 5). Beck and Jones (2010:93) accept the southern blade caches as Clovis, although none of these caches have associated radiocarbon dates and none contain temporally diagnostic fluted points. Beck and Jones fail to explain why the southern blade-dominated caches lack projectile points and exceptionally large bifaces which are diagnostic Clovis elements of caches on the Northern Plains.

It is questionable if any of these southern blade caches are Clovis in age. There is little evidence that separate caching of blades was a primary component of Clovis caching behavior if, as hypothesized here, that in addition to caching finished tools, the primary purpose of caching large bifaces was to provide material for later manufacture of large bifacially fluted points as back-up hunting equipment (Buchanan et al. 2012). Even if the blade-dominated caches from the Southern Plains are eventually dated to the Clovis period their function appears to be fundamentally different than that of Clovis biface caches.

The distributional evidence of blade caches does not support southern Clovis movement westward beyond the Southern High Plains of New Mexico because no blade caches are found further west than the Green and Dickenson caches recovered near Blackwater Draw in eastern New Mexico. Also, toolstones that dominate these southern blade caches, Edwards Plateau chert and Alibates agatized dolomite, are not represented in assemblages further west than the Southern High Plains.

Moving beyond blade caches and considering reported Clovis locales in the Greater Southwest, Clovis toolstone use demonstrates a more likely west to southeast movement rather than the westward directionality proposed by Beck and Jones. The Loa Clovis biface cache in central Utah was found 100 km to the east of the Wildhorse Canyon source location in the eastern Great Basin (Janetski et al. 1988). A Clovis point of obsidian from the Wildhorse Canyon source, one of three known obsidian sources in Utah utilized by Clovis groups, was identified at the Blackwater Draw site in New Mexico, more than 900 km to the southeast of the source (Copeland and Fike 1988:7–16). Vance (2011:15) notes that this projectile point contradicts Beck and Jones’ hypothesis of northwest migration of Clovis groups.

Vance (2011) reanalyzed of the toolstone from the Lime Ridge site, 42SA16857 (Davis 1989), a Clovis encampment in southeastern Utah. Beside the presence of local siliceous materials, the analysis noted the presence of wonderstone rhyolite, a material obtained from sources 300 km to the north or northwest. Vance (2011, 142) states “the presence of wonderstone rhyolite in the Lime Ridge assemblage indicates the same north-to-south movement apparent from the composition of other Clovis sites in both Arizona and New Mexico”, referring to Murray Springs (AZ EE:8:25) in Arizona and the Mockingbird Gap site (LA 26748) in New Mexico. At the Murray Springs Clovis site in Arizona, Haynes and Huckell (2007) note the presence of artifacts manufactured from toolstone sources several hundred kilometers to the north, leading Haynes (2011) to suggest that Clovis groups in Arizona and New Mexico arrived from the north via the Lime Ridge site in southern Utah.

Hamilton and colleagues (2013) discuss the toolstone used at the Mockingbird Gap site, a large Clovis encampment representing multiple occupation events in the Rio Grande rift region in central New Mexico. A variety of local raw materials were used in the manufacture of implements at the Mockingbird Gap site. Given the location of this site within Beck and Jones’ westward path of Clovis migration, toolstone sources from the Southern Plains, such as Alibates agatized dolomite and Tecovas jasper should be well-represented in the assemblage. No Clovis points in the assemblage were manufactured from these toolstone sources and Alibates agatized dolomite was only represented in “very low frequencies” in the assemblage (Hamilton et al. 2013:256). Chuska chert, Perdernal chert, Correo “China” chert, and obsidian, occurring in lesser quantities than local materials in the assemblage, are all derived from lithic sources outcropping up to several hundred kilometers north and west of the site on the Colorado Plateau (Hamilton et al. 2013:256).

These data indicate that at each of these sites in Utah, Arizona, and New Mexico, Clovis groups were using toolstone sources to the north and northwest from the Colorado Plateau rather than sources to the east or southeast. There is no evidence of toolstone transport from prominent southern sources such as Alibates agatized dolomite or Edwards Plateau chert beyond the Southern High Plains as would be expected by the Beck and Jones model. Beck and Jones’ hypothesis that Clovis technology originated in the Gulf Coastal Plains and was carried westward to the Pacific Coast is not supported by the archaeological record in the Intermountain West.

North Pacific coast origin of Clovis technology

If Clovis fluting technology did not originate in eastern Europe, the Arctic, Baja California, or the Gulf Coastal Plains, the only remaining hypothesis to consider for the origin of Clovis technology is somewhere along the North Pacific coast. A Pacific coastal migration entrance into North America has been hypothesized for more than 40 years (Fladmark 1979) and has been gaining wider acceptance (Erlandson et al. 2008; Lesnek et al. 2018). Lesnek and others (2018) indicate that a coastal migration route along the North Pacific coast was passable by about 16,000 years ago and Potter and others (2018) indicate that the biotic environment along the Pacific Northwest coast could have sustained populations that migrated south along the Pacific coast. Froese and colleagues (2019) believe that about 15,000 years ago lowland landscapes along the Pacific coastal route were accessible and within 500 years there was a diversity and abundance of terrestrial animals.

Morrow (2019:193) sees parallels in Clovis tool assemblages with assemblages from the Shaw Creek Flats area of central Alaska. The Shaw Creek Flats area includes one of the densest known concentrations of buried, in situ Pleistocene archaeological sites in the Americas (Lanoë et al. 2018). Sites in the area evidence both biface and bladelet (microblade) technology along with the presence of ivory points, needles, and ivory rods. The earliest component of the Swan Point site (CZ4b) in the Shaw Creek Flats area is dated is from 14,150–13,870 cal BP (Lanoë et al. 2018) and predates Clovis sites south of the ice sheets.

The distance from the Swan Point site in central Alaska to south of the receding Cordilleran ice sheet along a North Pacific coastal corridor is about 2400 km. Morrow (2019) notes that Clovis groups traversed the continent, a distance of almost 4000 km, within perhaps as few as ∼20 generations, so it is reasonable to assume that populations in central Alaska could find their way south along the North Pacific coast in the same or lesser amount of time to arrive below the Cordilleran ice sheet and start migrating eastward across the continent.

Croes and Kucera (2017) hypothesize that populations migrating south along the North Pacific coast encountered nothing but the Cordilleran ice sheet to the east for almost 2000 km until arriving below the ice sheet in the Puget Lowland in modern-day northwest Washington state. Until 14,500 years ago the Puget Lowland was covered by the Puget Lobe of the Cordilleran ice sheet (Booth et al. 2003). About that time the ice sheet began disintegrating into floating ice when marine water entered the Strait of San Juan de Fuca and lifted the ice sheet. This caused the wholesale collapse of the Puget Lobe from the Puget Sound to as far north as the Strait of Georgia in a short period of time (Easterbrook 2016). After the recession and disintegration of the Puget Lobe, the Puget Lowland was essentially unglaciated.

The occurrence of scattered fluted Clovis points in the Puget Lowland area (Croes et al. 2008) demonstrates that this area represents the first deglaciated interior continental setting south of the Cordilleran ice sheet accessible to human populations. Importantly, two potential Clovis sites are located in the Puget Lowland region. At Ayer Pond (45SJ454), an assortment of Bison antiquus bones was discovered on Orcas Island with evidence of possible human butchering dating to about 13,800 years ago (11,990 ± 25 ¹⁴C yr BP) (Kenady et al. 2011). Mastodon (Mammut americanum) skeletal remains were discovered in a kettle pond on the Olympic Peninsula with a putative bone projectile point buried in a rib element at the Manis Mastodon site (Waters et al. 2011 [no site number reported by the authors]). This site, also dating to about 13,800 years ago based on an average of four radiocarbon dates (11,960 ± 17 ¹⁴C yr BP), was identified as “pre-Clovis” by Waters and others (2011). As a label, pre-Clovis is an inappropriate, if not subtly pejorative, term because the true time span of the Clovis period is unknown. Even though Waters and Stafford (2007, 2013) provide a tight time range for dated Clovis sites, the sample size of radiocarbon dates associated with Clovis occupations is small and, most importantly, as Prasciunas and Surovell (2015) note, there is no way to determine whether currently dated Clovis sites include either the earliest or latest sites along the Clovis temporal continuum.

If, in fact, the Puget Lowland was the gateway for Clovis exploration of the interior of the continent, it would be expected that Clovis sites in this area would date earlier than Clovis sites elsewhere in North America. Although these two sites are questionable, they do date a few hundred years earlier than the earliest Clovis sites in the interior of the continent, as would be expected. Assuming that the embedded fragment in the Manis mastodon rib and the butchering marks on the Ayer Pond bison are anthropogenic, these two sites could be the earliest radiocarbon-dated Clovis sites south of the ice sheets.

Because the Cascade Mountains were never glaciated, Clovis groups from Puget Lowland could explore eastward unimpeded. The location of the East Wenatchee cache is less than 200 km directly east of the Puget Lowland. Although the direction of toolstone transport from the artifacts in the East Wenatchee cache is uninformative, Fiedel and Morrow (2012:381) suggest that the relatively large Clovis points in the cache indicate they may be some of the oldest known Clovis artifacts, because Clovis points appear to diminish in size over time.

The presence of bifaces and bladelets in the cache suggests to Morrow (2019:193) that the Clovis technology identifiable in the East Wenatchee cache is a derivative lithic technology originating from assemblages occurring in the Shaw Creek Flats area of central Alaska. Morrow (2019:193) also suggests that the cache could be the earliest known Clovis cache based on stratigraphic evidence that indicates the cache was deposited shortly after the Glacier Peak eruption dating between 13,700–13,400 years ago.

In addition to exploring from the Puget Lowland, colonizing populations could have used the Columbia River as a corridor for migrating into the interior of North America (Anderson et al. 2014:188). The Columbia River would have provided an unimpeded passageway east through the Cascade Range onto the Columbia Plateau. There is little evidence of Clovis along the Columbia River itself, although the Willamette River drainage, the first major inland tributary of the Columbia River, has extensive evidence of megafauna remains (Gilmour et al. 2015) and isolated Clovis points (Connolly 1994).

From the Pacific Northwest, Clovis groups could have quickly moved south and southeast to spread over the Columbia Plateau, the northern Great Basin (e.g. Dietz site [Pinson 2011]) and the Snake River Plain demonstrated by the Simon cache, Heil Pond (100E1548) (Reid et al. 2015), and scattered Clovis points in Idaho (Speer et al. 2019). According to Speer and others (2019) migrating populations, such as Clovis, “drifting eastward from the Pacific coast would soon discover that the corridor running from Farewell Bend on the Oregon – Idaho border to South Pass in the Wyoming Basin was the easiest route to the plains” (Speer et al. 2019:148).

On the Northern Plains, Clovis groups could have fissioned, with some groups carrying Clovis biface technology north and northwest by following a periglacial zone along the Rocky Mountain foothills into the ever-expanding ice-free corridor (Ives et al. 2019; Smith and Goebel 2018; Waters et al. 2015). Other Clovis groups could have continued eastward over the Continental Divide and then radiated outward to the northeast, east, and southeast resulting in the direction of toolstone transport of Tiger chert apparent in some of the Clovis biface caches (e.g. Crook County and Mahaffy) on the Northern Plains (Supplemental Table 1).

Other Clovis groups migrated south from the Snake River Plain. Heil Pond is an example of a Clovis encampment on the southern Snake River Plain from which Clovis groups could have moved south along Lake Bonneville. The Hell’n Moriah Site, 42MD1067, (Davis et al. 1996) at the southern end of Lake Bonneville is one of several Clovis encampments in that area that has been investigated. From this area Clovis groups could have continued southeast to Lime Ridge and then on Murray Springs and Mockingbird Gap as discussed above.

Based on this interpretation, the most parsimonious hypothesis for the origin of Clovis fluting technology is that it developed somewhere along the North Pacific coastal migration route from Alaska or possibly in the Puget Lowland of Washington. Figure 2 depicts the proposed Clovis exploration eastward from the Pacific Northwest.

Figure 2 Hypothesized spread of Clovis groups from the Pacific Northwest to the Northern and Southern Plains.

This hypothesis, although not presently confirmed, lends itself to additional independent testing and GIS modeling (e.g. Buchanan et al. 2016; Hamilton and Buchanan 2007). Although current eco-regions are not good proxy measures for environmental regions at the end of the Pleistocene, at least eight of the Clovis biface caches were deposited at or near the edge of the modern-day Northwestern Forested Mountains ecozone (Omernik and Griffith 2014) (Figure 3). Additional geolocational, quantitative, and chronological analyses of Clovis sites, isolated Clovis points, as well as Clovis biface caches are necessary to illuminate migration patterns across the western landscape by Clovis groups.

Figure 3 The location of Clovis biface caches in relation to the Northwestern Forested Mountains ecozone, highlighted in yellow (Omernik and Griffith 2014).

Conclusion

As presented here, people practicing a Clovis lifeway and utilizing a unique fluting technology most likely entered North America below the ice sheets from the Puget Lowland in northwest modern-day Washington state. There is no substantive evidence that Clovis fluting technology originated on the Gulf Coastal Plains of modern-day Texas, along the Ice-Free Corridor, Baja California, or eastern Europe.

Given the contemporaneity of the Clovis technology and the WST in the West, it is not likely that WST is the progenitor of Clovis fluting technology especially given the divergence between the two lithic technologies (Davis et al. 2012). Additionally, Smith and others (2020) indicate that Clovis and WST groups utilized different adaptative strategies. Western Stemmed Tradition groups were often tethered to littoral resource patches around Pleistocene pluvial lakes and frequently returned to caves and rock shelters resulting in stratified deposits. In contrast, highly mobile Clovis groups rarely reoccupied the same campsites.

Compared to the mobile strategy of the Clovis tradition, the tethered strategy of the WST was very successful and long-lived, lasting for thousands of years. Even if the WST is determined to predate the Clovis tradition in the West, WST groups did not colonize the interior of North America. There is no evidence that WST groups ever explored or traveled eastward beyond the Columbia Plateau, the Great Basin, or the Snake River Plain. The current evidence suggests that vast areas of the interior of western North America including the Great Plains, the Rocky Mountains, the Colorado Plateau, and portions of the Basin and Range Province were uninhabited and first explored by Clovis groups.

Clovis biface and tool caching provided a strategy for maximizing Clovis hunting, exploration, and migration, allowing for rapid expansion of Clovis groups across the landscape. This analysis suggests Clovis fluting technology developed along the North Pacific coast or the Puget Lowland before spreading across the continent. These founding Clovis groups focused on terrestrial resources and were initially adapted to a cold weather, periglacial environment (Dawe and Kornfeld 2017; Kornfeld 2013:42) before migrating into the interior.

The concept of a single founding population with a single technological tradition entering the Pacific Northwest or anywhere else in North America appears to be an over-simplification of the colonization process in this hemisphere. It is possible that different groups with different technological traditions and cultural values migrated inland from the coastal regions about the same time. These groups could have entered the continent synchronically or successively in such a short time period as to be chronologically indistinguishable in the archeological record (Reid et al. 2015:73; Smith et al. 2015:362).

For example, Erlandson and others (Erlandson et al. 2007:62) posit that the first migrants along the Pacific coast were paleocoastal populations exploiting littoral resources. However, other separate groups focused on terrestrial resources, some with Clovis bifacial fluting technology and some with WST technology, could have migrated into the interior while paleocoastal populations continued migrating south.

Clovis groups may not have been the first colonizers of North America, but these groups did carry the first widespread identifiable lithic technology into the unoccupied interior of the continent. There is an important distinction between the first exploratory groups entering a previously unexplored region and the populations who occupied the area later and benefited from the generational knowledge of their predecessors. These first individuals and groups had to explore the region before they could construct mental maps of the geography and resource patches such as toolstone sources. Once discovered, the locations of toolstone sources and resource patches could be shared with other group members and later among descendant populations. Caching of implements and toolstone was an important component of this process as each new area was explored. While Clovis groups may not represent the first wave of human migration into the North American continent south of the ice sheets, the distinctive and unique nature of Clovis technology represent an important anchor point in the prehistoric timeline of North America.

Acknowledgements

I thank Robert Hoard, J. David Kilby, Marcel Kornfeld, and several unnamed and anonymous reviewers for assistance and helpful comments on numerous previous drafts of this paper.

Disclosure statement

No potential conflict of interest was reported by the author(s).

Supplemental material

Supplemental material for this article can be accessed here: https://doi.org/10.1080/00320447.2021.1888188.

Additional information

Notes on contributors

Alan R. Schroedl

Alan Schroedl has been active in teaching, research, and archeological administration for more than 40 years. He continues to be interested in the culture history of western North America. Email: [email protected]

Notes

1 Osborn (2014) presents a strong case that eyed needles and spurred flake gravers were used to sew garments for weather protection during Clovis and later Paleoindian periods.

2 The largest reported Clovis point is an obsidian point measuring 24.7 cm recovered from the surface near the East Wenatchee cache in Washington state (Gramly 1993:51).