Across Atlantic Ice

by Dennis J. Stanford

  Although there is no definitive proof that the knife and the mastodon are part of the same archaeological assemblage, we contend that the clues from context in which they were deposited and the relatively undamaged condition of the knife overwhelmingly support their association. The edges of the blade are sharp, and there are micro-traces of knife use-wear on the distal surfaces and haft wear on the proximal end. The survival of these use-wear micro-traces indicates that the artifact remained in its primary context until disturbed by the dredge. Thus it was not tumbled, abraded, polished, or damaged by fluvial transport, wave action, or water transport of any kind until brought on board the Cinmar.22

  FIGURE 4.8.

  Location of the South Mountain rhyolite source in the Chesapeake Bay region relative to the Cinmar Site at the time of the Last Glacial Maximum.

  The excellent preservation of the Cinmar mastodon remains resulted from their rapid burial in a freshwater peat bog. Mastodons were apparently attracted to ponds and bogs.23 Freshwater springs, ponds, and peat bogs would have been common on the outer continental shelf during the LGM.24 We suggest that Paleolithic people killed and butchered the mastodon while she was in or next to a bog. The knife was lost during this activity. The anaerobic bog environment preserved the bone for ten millennia without physical changes. Human remains were likewise preserved for thousands of years in peat bogs in northern Europe.

  FIGURE 4.9.

  Post LGM sea level rise over the continental shelf showing mean sea level at the time of culturally associated diagnostic artifacts between 23,000 and 10,000 years ago.

  Near the end of the LGM, about 14,500 years ago, the collapse of the continental glaciers caused the sea levels to rise as much as 16 meters in a couple of hundred years.25 The elevated sea levels breached the James Peninsula Barrier Islands and barrier island lagoons, briefly transforming the Cinmar peat bog into a low-energy saltwater tidal marsh. The knife, tooth, and tusk became lightly stained by iron sulfide and rust-colored iron oxide from oxidation by aerobic bacteria of the marsh. The tidal marsh in turn was also inundated by the rising sea level, causing oxidation of the knife and bone to cease. If the sea level’s stabilization had left the knife in a salty environment for a significant period of time, it would have dissolved. But because of the fortunate location of the Cinmar Site on the west side of the barrier island lagoon, the knife was protected from ablation by tides and storms. With the isostatic subsidence—the pushing down of the earth’s mantle by the weight of glaciers on the crust—of the James Peninsula, the site was likely buried in shallow ooze until disturbed by the dredges of the Cinmar. Whether or not the knife was used to butcher the mastodon, or even whether the two go together as an archaeological assembledge, the presence of the knife at this depth would indicate that it dates to the time before sea level rose at the end of the LGM.

  The only other plausible scenario is that the biface was lost overboard by a later prehistoric sailor. But it seems highly unlikely that someone lost a knife overboard at that exact location and it came to rest in the relatively small area occupied by the mastodon remains. The odds against it are even greater given that this style of biface is extremely rare in the Chesapeake region. We asked a Smithsonian collection specialist to examine the museum’s extensive archaeological collections for a technological match from Maryland, Virginia, Delaware, Pennsylvania, or North Carolina. He found only two matches, and little is known about where they were found or with what they were associated.

  PRE-CLOVIS IN NORTH AMERICA?

  In the fall of 1975, we had an encounter with the thorny issue of the existence of pre-Clovis in North America. A team of archaeologists and students including Bruce and me were finishing up the excavation of an 11,500-year-old bison kill known as the Jones-Miller Site in Colorado. A local construction operator, Gary North, came by the archaeological field camp with some large bones he had unearthed on a nearby farm. We identified the bones as mammoth, which went extinct in North America at the end of the last ice age. The next morning our crew dashed off to examine the area where North had found the bones. We soon realized that there were a lot of bones in the excavation pit and perhaps an entire mammoth might be found. We were especially hopeful that human artifacts would be associated with the beast. The farm owner, Junior Selby, gave us a few days to uncover the rest of the bones, so North decided to work on another of his projects, at the Dutton farm a few miles south.

  As work progressed on the Selby farm, the geologist John Albanese determined that the bones were in a geologic level that dated older than 14,000 RCYBP (radiocarbon years before the present), at least 2,500 years before scientists thought the first humans appeared on the Colorado plains. However, some of the leg bones of the large animal appeared to have been intentionally broken to retrieve the fat-rich marrow. The absence of tooth marks suggested that someone had smashed these bones with a large, heavy object. If this was so, someone had broken these bones several millennia before the date of the earliest evidence of humans in North America. We therefore reasoned that some natural event would better explain how the bones were broken.

  North soon returned, excited to report finding more mammoth bones, in his new excavation on the Dutton farm. The next morning found a small excavation team twenty miles south looking at more mammoth bones. A jawbone slightly exposed by North’s large motor scraper stood in the way of construction progress. In an effort to salvage the specimen, the crew immediately began to excavate it from a geological deposit similar to that of the bones at Selby’s. The context suggested that both finds were from the same pre-human period.

  By mid-morning we discovered another bone lying next to the jaw. This too was mammoth bone, a short section of a rib that had been split by a series of blows along its margin. At one end it was beveled to a sharp point, which appeared to have become rounded and polished, perhaps through use as a tool.

  Joined that evening by the teams from the bison kill and the Selby farm, we gathered over a warm stew to discuss the day’s discoveries. Long into the stormy night we talked about the significance of the rib bone: Was it a tool made and used by a human more than 16,000 years ago? Or was there another explanation for its tool-like character?

  These issues were all the more intriguing since all of us were well versed in the accepted scientific explanations of New World origins. In all our textbooks and classes there was very little doubt that Clovis culture represented the first Americans, who came from Asia some 14,000 years ago. If the bone we found at the Dutton farm was a tool, and if the relative date for it was correct, it would be an artifact that was older than Clovis, and our textbooks and professors would be wrong. This was our first personal brush with heresy. Dennis

  We have identified only three other laurel leafs in private collections in the region. One of these was found in sand that was dredged offshore to rebuild a beach after a hurricane, one came from a near-shore clam bed, and another was found eroding from a Pleistocene terrace of the Susquehanna River in Dauphin County, Pennsylvania.

  Taken as separate collections, the Cinmar knife and the types of artifacts found in the lowest cultural levels at Meadowcroft, Cactus Hill, Miles Point, and Oyster Cove do not constitute large enough samples to characterize a cultural assemblage or flaking technology. However, if all five site assemblages are combined and treated as one complex, we begin to see a pattern emerge that includes laurel leaf–shaped bifaces, thinned biface projectile points, blades, burins, bladelets, polyhedral cores, and a host of non-diagnostic flake tools. We consider these tools to be a partial representation of a flaked stone tradition that is similar in many ways to Clovis, a variety of the Franco-Cantabrian Solutrean, and as it turns out, resembles the younger eastern Beringian Nenana Complex (see chapter 6).

  CONTINUITY FROM CACTUS HILL TO CLOVIS

  The main differences between Clovis point technology and the technologies of the earlier Mid-Atlantic assemblages is that the latter were made smaller and their bases thinned by the removal of a series of pres
sure flakes, not the distinctive fluting of the Clovis points. Thus the key issues for pinpointing the origins of Clovis are when and where purposeful fluting as a basal thinning technique came into use. We contend that people living in eastern North America perfected Clovis technology, so we examined both excavated and surface collections from the Mid-Atlantic and southeastern states. We found a wide variety of typological and technological treatments, ranging from straight through indented bases, with various amounts of pressure thinning, including unsuccessful attempts at fluting. If the unfluted Clovis-like specimens from the surface sites are proto-Clovis technology, their abundance suggests an early robust population in the southeast. That being the case, it is not surprising that a growing number of sites in the southeast have radiocarbon dates or stratigraphic contexts suggesting that they are proto-Clovis in age. Among these sites, Jefferson Island in Maryland, Page-Ladson in Florida, and Johnson and Coats-Hines in Tennessee are the most informative.

  Jefferson Island in the Chesapeake Bay has produced unfluted Clovis-like projectile points along with an assemblage of other tools exposed in and eroding out of the Tilghman Paleosol.26 This stratigraphic placement dates the site occupation to sometime between the LGM and YD climatic intervals (circa 13,000–25,000 years ago). Because of sea level rise the site is now on an island, but at the time of occupation it was situated on a terrace above the Susquehanna River. The projectile points (figure 4.10a–c) and blade tools (figure 4.10f–g) from Jefferson Island overlap the range of technological variation noted in both Cactus Hill and Clovis assemblages. Although the points are not fluted, they are basally thinned. Evidence of overshot flaking reduction is apparent from a tool made on an overshot flake (figure 4.10d). Among the blades and tools made on blades are two with abrupt steep edge retouch, one of which has flat flaking on its ventral surface and a possible burin scar (figure 4.10g), and two retouched blades (figure 4.10i–j). Two oblong, crescent unifacial scrapers (figure 4.10k–l), likely made on blades, were found along with a micro–end scraper (figure 4.10h), and an adze-like tool (figure 4.10e).

  The relatively large size of these artifacts and the excellent raw materials used in their manufacture indicate that the Jefferson Island flintknappers collected their raw stone farther up the Susquehanna River, closer to or at bedrock deposits, than did the Miles Point knappers. Perhaps this difference indicates that exploration for and discovery of new resources over a wider geographic region took place in the years after the occupation of Miles Point.

  Most of the Jefferson Island artifacts were collected from the intertidal zone, where they were eroded out of context onto the beach. However, a projectile point was found in situ but partially exposed in the Tilghman Paleosol. Since no other archaeological occupations occurred at this location, we suggest that all of these artifacts were once buried in the paleosol deposit and represent a single assemblage.

  FIGURE 4.10.

  Jefferson Island artifacts (a–l) and Florida pre-Clovis points (m–p): (a–c) projectile points; (d) overshot flake tool; (e) adze-like tool with outline of obverse side and reverse side showing bit and cross section of bit end; (f) tool with abrupt retouch made on the snapped proximal end of a blade; (g) snapped blade with abrupt edge retouch on the right dorsal edge, flat flaking on the ventral side of the blade, and arrow designating a possible burin scar; (h) microscraper; (i–j) retouched blades; (k–l) unifacial crescent scrapers/knives; (m–o) Page-Ladson points; (p) Suwannee point. Outlines show cross sections.

  At the Page-Ladson Site in Florida, three unfluted, indented base projectile points were found (figure 4.10m–o), but unfortunately they were displaced from their original context. Still in situ, however, mastodon ivory and bones with cut marks were found associated with chert tools. Seven radiocarbon assays run on these bones average 12,420 RCYBP. The flake tools found with the mastodon and the weapon tips were made of a chert from the Flint River Basin in Georgia. And perhaps not coincidently, strontium-90 studies of the mastodon bone indicate that these animals also came from Georgia. This co-occurrence may be serendipitous, but the investigators think it highly probable that the points were originally associated with the mastodon.27 Furthermore, the projectile points from Jefferson Island are technologically very close to the Page-Ladson points, and the former’s stratigraphic placement in the pre-YD Tilghman Paleosol supports the assumption that the unfluted points found at Page-Ladson were, indeed, used to kill mastodons 14,600 years ago. It appears that all of these unfluted points are pre-Clovis and can provide information about the later stages of the development of Clovis technology.

  The Page-Ladson points were originally grouped together with similar points under the name Suwannee (figure 4.10p). Suwannee points have now been divided into three types that may have temporal significance. The Page-Ladson type is thought to be the oldest.28 The overall point morphology and manufacturing techniques, including overshot flaking, are similar to Clovis, but without the flutes. These artifacts are well known to archaeologists working in Florida, especially in the northern part of the state, where they are common, but they have a broad distribution, occasionally showing up along the eastern seaboard as far north as Delaware and as far west as central Texas, were they were found below a Folsom occupation at Horn Shelter near Waco.29

  An undated collection from Harney Flats, Florida, based on its similarity to the Page-Ladson materials and dissimilarity to other materials in the region, helps to define the artifact assemblage content of this early Florida technology and reveals that the Page-Ladson-Suwannee lithic repertoire incorporated a well-developed polyhedral core and blade technology.30 At the Ryan/Harley Site and other localities found by divers seeking relics in underwater sinkholes and channels along the Wacissa River, Suwannee points have been found associated with a wide range of extinct fauna, including Pleistocene horse, tapir, and mastodon.31 A time span of 13,900–14,600 has been attained from worked ivory and butchered bone found at these locations, consistent with the dates from Page-Ladson. The excellent preservation conditions provided by the limestone sinkholes have expanded the variety of known ivory tools to include decorated beveled rods (figure 2.13e), a barbed point (figure 2.13f), and hooks for atlatl spear throwers (figure 2.13i–k).32 Many of these items are also associated with chipped stone tools, but no diagnostic projectile points have been found. It is unclear whether these were part of the proto-Clovis or Clovis assemblages, but the radiocarbon dates from Page-Ladson indicate that hunters killed and processed mastodons in Florida prior to Clovis times.

  Evidence of the final phase of proto-Clovis development comes from the Johnson Site, on the Cumberland River in downtown Nashville, Tennessee. This site was originally reported as a Clovis occupation.33 Twenty-six late stage preforms, all broken during final fluting (figure 4.11), were recovered from an occupation level, along with seventeen blades and blade tools, end scrapers, and miscellaneous flake tools, while other artifacts were found redeposited on the beach below the exposure. Three charcoal features from the occupation level dated to 11,700±980, 11,980±110, and 12,660±970 RCYBP, respectively. Though the standard deviations are large for two of the dates, we think the deviation for the third feature is within an acceptable range, and all three dates average about 14,000 years old. As these dates are at least 500 to 1,000 years earlier than currently accepted Clovis dates, some scholars dismissed the site as incorrectly dated or not a Clovis occupation.34

  FIGURE 4.11.

  Late-stage preforms from the Johnson Site (b, d, f) and Carson-Conn-Short Site (a, c, e) in Tennessee: (b, d, f) Proto-Clovis late-stage biface failures resulting from not preparing a beveled fluting platform; (a, c, e) late-stage bifaces showing typical Clovis beveled fluting platforms; (e1) platform bevel; (e2) face intended to be fluted.

  Because we accept the dates, we re-analyzed the assemblage and came to an unexpected conclusion: these are not Clovis artifacts, but Southeast Early Paleo-American (proto-Clovis) artifacts! The assemblage includes a plane face point; two end scrapers, bo
th retouched across their dorsal surfaces; and retouched blade fragments and bladelets. In addition, there are fourteen biface basal fragments, all of which were broken during late-stage fluting attempts gone awry (figure 4.11b, d, and f). This same breakage is also seen in some bifaces from the Carson-Conn-Short Site in Tennessee (figure 4.11a, c, and e). Unlike Clovis artisans, the flintknappers who ruined the Johnson Site bifaces did not prepare proper fluting platforms, instead striking the base of the biface no matter what its morphology.

  BEFORE THE BEGINNING?

  In the 1970s, an archaeological excavation of a seventeenth-century colonial homestead on Eppes Island, Virginia, turned up an unusual biface. The investigators of the site recognized it as a Solutrean laurel leaf. They concluded that the biface, which they thought was made of French flint, was a relic brought to the New World by an early colonist.