Across Atlantic Ice

by Dennis J. Stanford

  A discovery at the Gault Site in central Texas provides another example of a possible Clovis dwelling. This is a square cobble pavement with straight sides and distinct corners that measures almost 2 square meters. The sides of the pavement line up with the cardinal points (figure 2.2). As the Texas archaeologist Mike Collins has pointed out to us, the odds against this happening randomly are 90 to 1 (based on a 360 degree circle). The structure is in what would have been a wet valley bottom in fine clay deposits. Surrounding this pavement and resting at the same elevation is a scatter of blades, biface fragments, limestone clasts (possibly artifacts), and bone fragments, leaving no doubt about the pavement’s association with an occupation. Collins’s recent re-evaluation has questioned the Clovis association based on the lack of diagnostic Clovis artifacts as well as a series of dates indicating that this feature and its associated artifacts may be older than Clovis.5 Investigations into this issue continue.

  There is, of course, a lot of speculation about how the feature was used. A likely possibility is that it was the floor of a small structure, constructed to elevate the occupants above the damp valley soil. Blade segments with evidence of grass and reed cutting were associated with the feature, and Collins imagines that the pavement was laid down and then covered with grass or mats, which in turn may have been covered with hides, making a comfortable platform on which to sleep. It is also possible that reed mats were used as walls to enclose the structure.

  Another Clovis cobble pavement was discovered in 1948 in Texas at the Kincaid Rock-shelter.6 This pavement covers a significant portion of the Clovis occupation level (there are more uses of the site in archaeological layers above Clovis), and artifacts and bones from food refuse were found in the cracks between the cobbles. A water seep presently keeps the interior of the shelter wet, and if the soil was damp during the Clovis occupation the cobbles may have been used to provide a dry surface. Why Clovis people selected this wet cave among the many dry caves throughout Texas remains a mystery. Nevertheless, these two sites establish pavement as a method known and used by Clovis people.

  In order to understand the intensity and use of the Gault Site, intensive refitting analysis has been done there. Flaked stone artifacts were checked to see if knapping sequences could be reconstructed by fitting flakes into flake scars on flakes, cores, and tools. If pieces found near each other fit together, it is likely that they were made and used in that location, and if people used the site only temporarily, especially as a manufacturing locality, one would expect a lot of the remains to be clustered in refittable clusters. On the other hand, if there is a lack of fittable pieces it means there was substantial mixing of artifacts caused by long-term camping activities. The Gault study identified few refits. Since there is little indication in many of the deposits of natural processes moving the artifacts, the paucity of refits probably resulted from human activities that moved individual pieces away from where they were made. Moreover, tools of local chert that were used, worn, refurbished, worn out, and finally discarded are common in the collection and indicate that the Gault occupations were intense and long-term enough to result in this thorough mixing and reuse of artifacts.

  FIGURE 2.2.

  Limestone cobble pavement thought to compose the floor of a structure at Gault, Texas. Solid line indicates limit of excavation.

  Another example of a large Clovis site is the Carson-Conn-Short site on the ice age banks of the Tennessee River, which overflow waters from Kentucky Lake currently seasonally inundate.7 Although this site has been only preliminarily tested, it appears to be a major Clovis camping area associated with a chert source. A meter or more of midden (cultural refuse) is spread continuously along the terrace for up to a kilometer. The occupation debris probably accumulated from both long-duration encampments and repeated shorter visits during quarrying activities. Although the brief tests have recognized no evidence of dwellings, the presence of storage pits and possible heat-treating features indicate that this site has great potential for producing evidence of shelters if and when more detailed excavations are conducted.

  EVIDENCE OF SETTLEMENT DIRECTION

  Almost fifty years ago Ron Mason observed that there were significantly more fluted points reported from the southeast than from the rest of the continent and that there was more variation in fluted point forms and technology in the east than in the west.8 Following the age-area hypothesis that the greatest number and variations of traits will occur near their center of origin and decrease as they defuse outward, he suggested the southeastern United States and not the West as the homeland of the Clovis complex.

  Mason’s estimates of the number of fluted points were based on impressions, not on empirical data. But as the years passed, surveys reporting eastern fluted point finds and sites increased at a rapid rate, while western finds remained relatively rare. Twenty years later Louis Brennan, the editor of Archaeology of Eastern North America (AENA), in an effort to revive Mason’s hypothesis, organized and published a compilation of eastern fluted points.9 The AENA compilation covered the cis-Appalachian East from Quebec to Florida and recorded more than 5,800 points. Although the leading experts on Paleo-Indian archaeology were impressed by the data, they all agreed that the abundance of specimens alone was not proof of the antiquity or ultimate origin of Clovis. Even if the compilation demonstrated that there were more Clovis points recorded from the East, there were likely appropriate explanations—for example, the carrying capacity of the eastern habitats may have been more favorable to population growth, or there may simply have been more artifact collectors and archaeologists working in the East, where decades of agricultural activities likely plowed up more artifacts. Conversely, if there were fewer Clovis points recorded from the West, it was likely that greater erosion there had destroyed many Clovis sites and re-deposited fluted points. A study by the archaeologist Brad Lepper found a positive correlation between the numbers of fluted points reported from Ohio and the amount of acreage under cultivation and associated with contemporary population centers.10

  Can we evaluate whether or not a greater density of eastern Clovis points reflects better artifact visibility due to erosion and cultivation, plus more intense scrutiny and focused inventories? Certainly there are places where surface exposure is limited in the West, but this is not characteristic of most of the area. Virtually all of the grassland prairies have fallen to the plow, and surface erosion has been so extreme that a large percentage of the topsoil has been stripped off. We only have to look into the history of the U.S. Soil Conservation Service (now the Natural Resources Conservation Service) to see how widespread this has been and continues to be. Erosion has often reached below late Pleistocene levels, exposing massive tracts of Clovis and earlier deposits, leaving artifacts exposed along with gravels. This process was especially intense during the dust bowl days of the 1930s, and to a lesser degree during the 1950s. These severe droughts exposed an enormous area of what was once prime large-game habitat, especially for now-extinct Pleistocene species.

  Some of the most fruitful artifact collecting episodes in North American archaeology took place on the Plains and in the Southwest during the dust bowl. Large private collections of Paleo-Indian (and later) points were assembled from the extensive exposures of late Pleistocene deposits in both regions. Although many Folsom (the fluted point culture that followed Clovis in the High Plains) and later Paleo-Indian sites were documented, Clovis artifacts were rarely found, even though numerous mammoth remains were reported.

  Many western collections have been inventoried, and there are regional compilations similar to those in the east. For example, John Cotter and students from Denver University working under the direction of E. B. Renaud traveled throughout eastern Colorado, western Kansas, southwestern Nebraska, and northeastern New Mexico during the 1930s to record Paleo-Indian artifacts found by local collectors.11 Even though the study was conducted in an area where there was excellent exposure and visibility, as well as extremely active artifact coll
ecting, Clovis points were poorly represented in the collections and accounted for less than 0.04 percent of the total Paleo-Indian point count. It seems reasonable to us to compare these Clovis artifact density results to those in the east, and this pattern has continued to the present, with a large difference in the number of Clovis points and sites found in the western and eastern states.

  There is also a far greater diversity of Clovis descendant point forms and technologies in the east. These variations likely reflect local adaptations to varied ecological settings, but links among form, technology, use, and environment have not been established. In the east there are styles such as Debert (figure 2.3a), Cumberland (figure 2.3b), Redstone (figure 2.3c), Barnes (figure 2.3d), and Crowfield (figure 2.3e), to name just a few. Only a single variant, Folsom, is present throughout the western Plains (figure 1.1a), and there is a fluted variety in Alaska (figure 2.3f).12 Folsom likely developed as part of a bison-hunting specialization, and this point type expands along with the expansion of the grasslands that became the bison range. Consequently, we contend that the greater site density and diversity in point styles in the east support the conclusion that the Clovis tradition has a greater time depth there.

  Although most fluted points have been found out of context and on modern ground surfaces in the east, there are major Clovis deposits buried in early terraces of eastern rivers. The densest accumulations of Clovis artifacts have been encountered at Shawnee Minisink on the Susquehanna River in Pennsylvania, Thunderbird on the Shenandoah River in Virginia, Carson-Conn-Short on the Tennessee River in Tennessee, Williamson in Virginia, Adams in Kentucky, Sinclair Site in Tennessee, and the Pine Tree and Topper Sites on the Savannah River in South Carolina.13 The Gault Site in central Texas, possibly the largest accumulation of Clovis artifacts yet identified, is situated on a small stream associated with springs. Many of these dense surface accumulations of Clovis artifacts are near or even at chert sources and have artifact assemblages that indicate an extensive range of domestic activities beyond simply obtaining tool stone. No comparable sites are known in the west.

  FIGURE 2.3.

  Post-Clovis fluted point types: (a) Debert; (b) Cumberland; (c) Redstone; (d) Barnes; (e) Crowfield; (f) Alaska fluted.

  The few western Clovis campsites, such as Murray Springs in Arizona and Sheaman in Wyoming, are much smaller than the eastern sites, and most are at or near kill sites with associated processing areas, such as Lehner and Murray Springs in Arizona and Lange-Ferguson in South Dakota. The Colby site, in Wyoming, has been interpreted as a meat cache.14 Even though there are excellent quarry locations with high-quality tool stone used by Clovis flintknappers, associated campsites have not been found. We suggest that the large eastern sites represent a complete cross section of Clovis population that camped on or near stone sources for extended times or for shorter terms but revisited the sites many times, whereas the paucity of evidence found at the western quarry locations, kill/processing sites, and campsites suggests that these localities were for short-term use or specific activities performed by a small subset of the Clovis population.

  Large blades are rare in western Clovis assemblages but extremely common at many southeastern sites. The scarcity of blade technology in the west is not related to the availability of high-quality stone. Excellent sources of flaking material (e.g., Alibates dolomite from the panhandle of Texas, quartzite from the Spanish Diggings in Wyoming, and jasper from southeastern Wyoming) are available throughout the west and were used by Clovis people for projectile points but seldom for blades. The differential distribution of large blade technology may be related to chronology; that is, blades are also absent from late Clovis assemblages such as Thunderbird in the east. Or they may have been used in specific tasks that might have been gender related. Use-wear analyses of blade tools suggest that they were used for hide working and for the construction of matting, both frequently female tasks in historically documented foraging societies. The dearth of blades may signify that hide procurement and processing and the establishment of more durable living structures were not part of western activities. Along with the minimal occupation debris left at campsites, the lack of blades may point to small, mobile exploration parties rather than larger bands. However, it can be argued that blades occur in the same proportions to projectile points in the west as in the east, since Clovis artifacts are relatively rare in the west. Unfortunately, it is difficult to evaluate the circumstances of the occurrence of blades as few reports include relevant chronological or use-wear data.

  Even when the various indications of greater Clovis population in the east are accepted, they are usually still interpreted as evidence of a west-to-east settlement pattern. For example, David Anderson and Jack Gillam produced geographic distribution maps based on a systematic database compiled from fluted point reports in official government files.15 When plotted at the county level across the continental United States, the density of fluted points is clearly greater in the east (figure 2.4). On the basis of this information Anderson and Gillam proposed several colonization models—all based on the assumption that Clovis arrived in the west and expanded to the east. Accordingly, they explain the high density of eastern points by supposing that the first migrants were few in number and that their population grew as their territories expanded southeastward along major rivers such as the Missouri and Mississippi to eventually end up on the eastern seaboard. The eastern environment was much more productive than the western, resulting in a large human population concentrated in several geographic centers.

  FIGURE 2.4.

  Density of Clovis finds and raw material movement, 13,000 BP. (Based on Anderson and Gillam 2000.)

  But there are lines of evidence that seem to effectively challenge the eastward-expansion-of-Clovis hypothesis. One argument involves the nature of caching behavior. Caches are places where large numbers of artifacts or small groups of projectile points were buried. Clovis people left three varieties of caches: groups of blades, groups of projectile points, and mixes of projectile points, biface preforms, and tools.16 Currently, the best-known Clovis caches are west of the Mississippi River, but a growing number are being identified farther east. The caches may have been made for various purposes, ranging from logistical caching of tools for later retrieval to ritual offerings never intended to be reclaimed. In a thought-provoking paper, the Southern Methodist University archaeologist David Meltzer hypothesized that when Clovis people colonized a new, uninhabited landscape, they anticipated the possibility of difficulty in finding raw materials for tool replacement.17 To prepare for such an eventuality, they created insurance caches of artifacts across the landscape as they moved away from known resources. Thus, caches served as resupply depots; if explorers didn’t find stone sources in the new territory, they could resupply from a stash strategically near the margins of known territory. This strategy would be essential for exploring and learning new lands, making it possible to venture farther and more widely across otherwise unknown terrain. Presumably, if new stone sources were found the explorers would not need to return to a cache, which may account for the number of these features in the archaeological record.

  We agree that the caching behavior pattern has great practical benefits during exploration ventures, and we contend that the source areas for the raw materials found in a cache clearly indicate the general areas and directions from which explorers came. To test the directionality of Clovis exploration, we assembled information on the long-distance transport to the western caches of stone identifiable to specific sources. In six out of seven caches (the seventh contained only local material), all of the stone obtained from sources more than 150 kilometers away originated from sources south or east of the cache location. Only the Green River chert in the Crook County, Wyoming, cache traveled from west to east, but it also went north. The single Clovis point found in this cache probably originated in the Spanish Diggings area to the south. Although this pattern is not conclusive, it does not support the notion that Clovis folks
were exploring from the west or northwest; instead, it supports just the opposite direction of travel.18

  If Clovis culture originated in the east and spread westward, the eastern radiocarbon dates should be earlier than the western, and there are some indications that this is the case. Conversely, if Clovis originated in the northwest and moved toward the southeast, the earliest dates should be in the northwest. So far, they are not (figure 2.5).19 The oldest accepted radiocarbon assays for Clovis are 11,540±100 and 11,590±90 RCYBP (radiocarbon years before the present; see the introduction for information about this dating method), both from the Sloth Hole Site, Florida, a long way southeast from any northwestern entry point imagined in a Beringian origin hypothesis. The youngest dates associated with “Clovis-like” fluted points are 10,400 RCYBP, from Charlie Lake Cave in British Columbia, Canada, and circa 10,300 RCYBP from northwest Alaska.20

  If correct, these radiocarbon dates indicate at least a 500-year period during which eastern Clovis people honed their fluted point technology before expanding westward to the prairie margins. This extended period would account for the abundance and diversity of artifacts found in the east, along with large occupation sites with deep refuse middens. The younger western radiocarbon dates, minimal usage of the western landscape and stone resources, and presence of insurance caches of raw materials originating from distant southeastern sources suggest that Clovis bands were expanding their territories out of the east and southeast.

  FIGURE 2.5.

  Clovis sites with calibrated radiocarbon dates.

  It is our opinion that the rapid expansion of Clovis peoples into the Southwest and the Plains would have taken place at the onset of the Younger Dryas, a period of rapid climatic change when the weather turned cold and dry. It is clear that water tables were dropping in the western Plains and Southwest, as evidenced by hand-excavated wells at Blackwater Draw and Murray Springs, and this situation may have been more widespread than we currently understand.21 Plummeting temperatures and decreased moisture would have had a major impact on the climate and ecology, causing a significant reduction in plant and animal resources. If eastern Clovis relied on coastal estuarian resources as a primary subsistence element, a drop in sea surface temperature would have caused food shortages. Consequently there may not have been sufficient resources to support a large southeastern human population. Moreover, during this cold, dry period many of the larger animals, such as the mammoth, mastodon, sloth, camel, and horse, became extinct in North America.