Recently, archeologists have wrestled with reconstructing Folsom mobility and subsistence practices on the Southern Plains. Tool assemblages at kill and camp sites often include a variety of high quality lithic materials from sources separated by over 300 km. The various lithic materials manifest in any specific assemblage are interpreted as a direct indication of the past movements of the Folsom group. The area encompassed by the plotting of the various lithic sources equates to at least a partial mapping of the area traversed by a particular group. In addition, the size and amount of reworking exhibited on projectile points and the presence of non-fluted or pseudo-fluted points in the assemblage have been used to model a highly efficient and conservative lithic technology sustainable in a highly mobile system.

Likewise, analyses of bison kills and processing/camps have produced subsistence information consistent with the mobility reconstructions (Jodry 1987; Bamforth 1985; Bolderian 1987; Kelly and Todd 1988; Todd et. al 1990). Against the backdrop of more lush vegetation, higher carrying capacity for grazers, and low human population densities, Kelly and Todd (1988) have proposed a hunting and mobility model for early Paleoindian cultures. This model incorporates aspects of bison kill/processing strategies with mobility.

The model proposed by Kelly and Todd (1988:234-235) contains four basic parts:

Within this framework of Paleoindian mobility and subsistence, Kelly and Todd (1988:238) propose that Paleoindian groups, upon making a kill, would immediately begin to search for another herd. When a herd was located, the group would begin pursuit, taking with it any processed materials from the last kill. Pursuit took precedence over completion of the processing of bison remains for bulk storage.

Given the mobility aspects of the model proposed by Kelly and Todd (1988), what would the accompanying lithic technological system be like?

If, as Kelly and Todd suggest, finding bison for the next kill took precedence over complete processing of the last kill, then perhaps other factors such as lithic procurement would take a back seat to the search for more bison as well. Such a situation is reminiscent of Binford's (1979) concept of embeddedness. According to Binford (1979:273) in such a system, the procurement of lithic material is conducted during the course of some other subsistence task (e.g. game searching) and tool production/maintenance is performed during other work schedules (e.g. processing of animals). One of the tasks repeatedly identified at Folsom processing sites is the retooling of weapons (Jodry and Stanford 1992), suggesting that point production and hafting occurred at the same time as bison processing. The recovery of redundant artifact classes at Folsom camps near kills indicates the same tasks were performed at all camps (Kelly and Todd 1988). A similar situation is reported for the Hanson site, a camp site at a lithic source that may have been near a kill (Ingbar 1988). This is consistent with an embedded lithic procurement and production system.

That visits to a lithic quarry were embedded in the task of searching for bison is not a new idea about Folsom organization. However, when modeling the lithic technology and developing mobility systems based on lithics, the embeddedness of the lithic system is often lost. Re-embedding lithic technology into the broader subsistence/mobility system may reduce problems encountered in applying lithic-based models to the archeological record.

An Example

Recent characterizations of the Folsom mobility and subsistence systems include source mapping of the lithic types used in the manufacture of projectile points (Gramly 1980; Hofman 1991). The lithic materials represented by the projectile points are viewed as direct indicators of movement across the landscape between various lithic sources. In this scheme, the material comprising the highest percentage of the assemblage represents the last lithic source visited. The composition of the assemblage is modified by the number of kills conducted since a quarry visit. Such kills result in the replacement and reworking of broken projectile points.

There are some underlying assumptions to these reconstructions. First, it is assumed that a visit to a quarry area leads to an immediate replacement of small, heavily reworked points (Hofman 1991:351) and an increase in the use of a particular material--that of the quarry--in the projectile point assemblage. However, in a truly embedded system, the newly acquired stone would only be employed when tools broke or wore out. The newly acquired lithic material would be in the form of blanks, roughouts, and cores acquired from the source. If reconstructions of the mobility system are correct (Kelly and Todd 1988) the Folsom hunters would be in search of bison and would not disrupt this search for any length of time.

In this system, extensive refurbishing of the assemblage at a lithic source would occur only if a kill took place at or near the quarry. Perhaps the Hanson site is an example of this (Ingbar 1988). Such a kill would enable extended exploitation of a lithic source since the group would be processing the animals. Otherwise, only a quick stop at the quarry to obtain suitable materials (probably in the form of large bifaces) would be possible. These materials would be incorporated into the assemblage as the need arose. Knapping this material would occur as a general camp activity conducted while the group was searching for another bison herd or while processing the animals after a kill.

The lithic technology of Folsom groups has been characterized as a curative system (Bamforth 1986; Boldurian 1991; Brosowske 1996). A typical assemblage consists of highly curated lithic source material in the form of bifacial cores, large flakes that could be fashioned into a number of tools upon demand, formalized scraping tools, beaked tools, and projectile points. The projectile points often undergo numerous resharpening episodes. In addition to the curated assemblage, Folsom sites often yield numerous expediency tools made from flakes struck from the bifacial cores or from large flake blanks obtained at the quarry. These expediency tools usually consist of unifaces and unifaces with projections (often called gravers).

Modeling lithic acquisition and use during Folsom times will be differentially expressed in the curated assemblage and the expedient tool assemblage (Bamforth 1985). Because the archeological record is predominantly composed of discarded items, the archeologist is seeing the last decision regarding a curated item--that of discard. For a projectile point at a kill site, we assume 1) the projectile is no longer considered usable and is discarded or 2) the projectile was lost. Either way, the projectile point was part of a functioning curated system prior to its loss/discard at the kill site.

Utilized flake tools at the same kill site, however, are a component of the expediency tool assemblage and reflect the lithic material carried for the purpose of making such implements. Expediency tools will be made from bulk material acquired at the quarry and will reflect the last lithic source visited. Again, the decision to discard depends on the flake's usefulness in future retooling episodes. Many of the flakes recovered from kill sites are too small to have served as projectile point blanks and thus, were probably produced from a core specifically for the purpose of butchering the animals at the kill.

The Curated Assemblage

Since projectile points have been used to reconstruct Folsom mobility, a model is proposed based on the replacement of the curated projectile points in the site assemblage at kill/processing sites.

If a quarry visit is embedded in the act of searching for game, then it can be assumed that the hunters had already refurbished their weapons in anticipation of the hunt. In this instance the first infusion of a newly acquired lithic material into the curated assemblage might not occur until after the first kill following a quarry visit. At that time, broken or lost points would be replaced using material obtained from the lithic source. As a result, that particular knappable stone would see a percentage increase in the overall curated tool assemblage. Furthermore, that particular lithic material might not contribute significantly to the curated tool assemblage until several kill episodes had necessitated refurbishing the majority of the assemblage.

In this system, tools already hafted and determined usable would not be replaced simply because a new lithic source was available. The newly acquired lithic material would be employed when tools broke and needed replacement. In this case, archeologists might recover a single tool of a particular lithic source in an assemblage dominated by another source. In these instances, the usability of a point is the determining factor in whether the specimen is curated, not the material from which it is made. Once a point is determined usable and time and materials are invested in hafting it, it would require more than the timely availability of another lithic source for it to be summarily replaced.

The actual affect that an embedded lithic system would have on the curated point assemblage, then, is expressed as a direct relationship between the number of points of a particular lithic material and the number of retooling episodes since visiting that source. This relationship would change once another lithic source had been visited. At this stage, the first material would show a drop as it was being replaced by the new material type. As the group moves to yet a third source, the cycle begins again. Background noise includes possible infusion of groups from other areas of the Plains in response to variations in bison movements.

Two generic lithic sources (Source A and Source B) can be used to illustrate this model (Fig. __). Let’s enter this cycle immediately following a major kill and processing episode in which stores of large bifaces and flake blanks are nearly spent and a visit to a nearby lithic source (Source B) is warranted. At the last kill and processing stop, the group’s hunting assemblage was retooled and consists of many heavily reworked points. The stores on hand were adequate to retool enough spears for another kill episode if the opportunity presented itself. But, because the lithic stores were low, the search for bison was directed into the vicinity of the Source B lithic material. While in the process of looking for bison, the group stopped at the Source B quarry long enough to restock with large bifaces and flake blanks. Since the hunting weapons had already been retooled with the remainder of the Source A stores, there was no need to retool at the quarry.

Following the quarry visit, the tool kit assemblage would consist of hafted projectile points made from Source A lithics and large bifaces and flake blanks made from the recently acquired Source B material. The first kill following the visit to Source B would leave at the kill site broken and lost projectile points made of Source A material. Flake knives and other expediency tools used in the initial butchering of the bison would be made of the recently acquired Source B material. Thus, any lost or broken butchering tools left at the kill site, along with any resharpening flakes, would be of Source B material. Retooling at the processing camp would include the manufacture of projectile points from the Source B lithic stores and use of expediency tools also made from Source B materials. Any points made from Source A material that were still hafted and deemed suitable for another kill would be maintained in the assemblage. Source B projectile points would only be introduced into the assemblage to replace points lost at the kill or broken beyond the point where resharpening was a viable option.

As the group leaves this kill/processing camp, its lithic assemblage would consist of projectile points made of a mixture of Source A and Source B materials, and cores and flake blanks made of Source B cherts.

As repeated kills are made, the projectile point assemblage will witness a drop in overall percentage of Source A material as these points are replaced by Source B stores. The non-projectile point assemblage would continue to be dominated by Source B lithics.

This scenario could be changed if a kill had been made near the Source B quarry. Such a circumstance would allow the group to visit the quarry while processing of the animals continued. Because a kill had just occurred, any lost or broken points could be retooled using the newly acquired Source B material. And, as processing of these animals continued, large flake knives and other expediency tools made from the Source B materials could be employed. Upon leaving this site, the group’s assemblage of projectile points will still be dominated by Source A cherts since the Source B chert was only used to replace broken or lost components. The group, having replenished its stores (large bifaces and flake blanks) with Source B material, continues in search of bison for the next kill.

In this model, lithic material would be collected in the form of bifaces or flake blanks, and the projectile point assemblage would not see an immediate use of this material. However, at the first kill following a quarry visit, the new material would be used to replace broken or lost projectile points. Since kill sites usually contain far fewer projectile points than the total number of animals killed (Cooper Upper Kill 13 pts/29 animals, Middle Kill 7 pts/29 animals, Lower Kill 7 pts/20 animals), it is assumed that the percentage of an assemblage lost during a kill is low. Hence, initial replacement of points with the new material would not constitute a major percentage increase of that lithic type after a single kill/retooling episode. However, the lithic type would see a steady increase in the assemblage after each kill, until at some point, nearly all projectile points would be made from the last visited quarry type.

The Non-Curated or Expediency Assemblage

A different scenario is anticipated for the quarry blanks and expediency tools in a given assemblage (Fig. __). The non-curated assemblage would almost immediately be dominated by the recently acquired lithic type. Especially, if, as Hofman (1991) predicts, raw material would be dwindling as a result of repeated retooling and processing activities after numerous kills. Hence, the amount of raw material retained by members of the group would dwindle. Thus, the quarry visit would target replenishing raw material in the form of bifaces (Bolderian 1990; Bolderian et al. 1987; Stanford and Broilo 1981) and flake blanks. The first kills following a quarry visit would already make use of the newly acquired materials for expediency tools since these tools are made at the kill. The non-curated tool assemblage would be dominated by recently acquired quarry lithic material while the curated assemblage would still be dominated by previously acquired lithic types.

The inverse relationship between curated and non-curated components of the assemblage would be visible only until the curated assemblage became dominated by the material obtained at the last quarry stop (i.e. after numerous retooling events). If enough retooling events took place, the entire assemblage would be dominated by the last source visited.

Archeological Support of the Model

Although no direct link can be established between specific sites in the Southern Plains suggesting the round of a single group, if the model is applicable, then there should be redundancy in the archeological record. Thus, the archeological record should contain sites that can be plugged into the various positions. Turning to the archeological record and the data provided by Hofman (1991), one trend that could support some aspects of this model is found in site assemblages containing Edwards and Alibates lithic materials. The Lipscomb bison kill site contains roughly 70% Edwards and 30% Alibates points (Hofman 1991:342). This kill occurred during the fall of the year. Moving further to the east, the Cooper site, also a fall bison kill has an assemblage from the Upper Kill containing 75% Alibates and 20% Edwards (5% is Niobrara jasper). This pattern appears to fit the model in that the percentage of Alibates projectile points increases as the distance and assumed number of retooling episodes from the Alibates quarry also increases. Hence, as the pre-quarry Edwards chert assemblage undergoes attrition, it is replaced by Alibates material acquired at the quarry.

The non-curated assemblage at Lipscomb contains numerous pieces of Alibates, including several large flakes suitable for projectile point preforms (Hofman 1991; Hofman et al., 1991) even though the curated assemblage (projectile points) contains only 30% Alibates and is predominantly (70%) Edwards Plateau chert (Hofman 1991). Further east, the non-curated assemblage in the Upper Kill at the Cooper site is entirely Alibates, whereas the curated is comprised of 75% Alibates. Thus, the model presented for the non-curated and curated assemblages appears to reflect the archeological samples from Lipscomb and the Upper Kill at Cooper. This model predicts that within several more retooling events after the Cooper kill, the curated and non-curated assemblage for this group will be nearly identical in composition with a preponderance of both curated and non-curated tools made of Alibates. The ability to determine the direction of seasonal movements will have an enormous affect on the closeness of fit of this model when applied to archeological materials. On the Southern Plains, we do not have sufficient information at this time to evaluate this model. The Middle Kill and Lower Kill at Cooper have significantly different assemblages from that seen in the Upper Kill, indicating they were at different stages in the stone-use cycle.

Effect of Embeddedness on the Reworked Assemblage

The presence of non-fluted or pseudo-fluted points and heavily reworked points in the Folsom curated assemblage provides another problem in modeling mobility patterns (Hofman 1991, 1992; Hofman et al. 1990). The manufacture and use of non-fluted points has been used to illustrate the conservative nature of Folsom lithic technology (Agogino 1969; Amick 1995; Judge 1973). The risk of projectile point breakage during manufacture is greatly increased by the fluting process (Hofman et al. 1990; Judge 1973; Kelly and Todd 1988). If lithic material is scarce, non-fluted points would appear to be an adaptive alternative to fluted points for bison hunting. Likewise, the rejuvenation or reworking of projectile points would conserve lithic raw material. Under this system, the number of non-fluted and heavily reworked fluted points should increase with distance from a source. But, as Hofman (1991, 1992) points out, it is the number of retooling events rather than the linear distance from a source that determines the amount of lithic material in the group. Even with this, though, the archeological record contains sites with high percentages of non-fluted, heavily reworked points near major lithic sources (Hofman 1991:345). The presence of non-fluted and heavily reworked fluted points at sites near lithic sources are consistent with an embedded lithic technology. In an embedded system, usable points would not be replaced until they were broken, lost, or otherwise determined unusable regardless of the proximity of a replacement source of material. The decision to discard a large number of non-fluted points most likely would be made at a kill that took place after the visit to a major quarry area. Thus, the assemblage at a kill just after a quarry visit would contain discarded non-fluted points and any broken or lost specimens. Fluted points made from the newly acquired lithic material would be incorporated into the assemblage during subsequent retooling events.

Again the Folsom points from the Upper Kill at the Cooper Site help illustrate the concept of embeddedness of Folsom lithic procurement systems. The recovery of small points illustrates that small Folsom points were not only perceived as still effective, but which in fact were because they came from among the skeletons of 29 bison at the Cooper site.

The majority of points from the Upper Kill were made of Alibates (Fig.__). Following Hofman (1991:349), this would indicate that the Folsom group had most recently been at the Alibates quarry in the Texas panhandle. Expedient large flake tools recovered from the kill are also made of Alibates, further strengthening the interpretation that these people had recently visited the Alibates source area. At least one of these flakes was of sufficient size to serve as a Folsom point blank. But, instead of replacing the several small, reworked Folsom points with new, larger points immediately following the quarry visit, this group employed the smaller points until they were no longer effective, or perhaps were lost in the kill. The use of large Alibates flakes for expedient tools in processing the animals at Cooper indicates this Folsom group had the raw materials available to replace the small points still contained in the weapon assemblage prior to the Cooper kill. It is assumed that following the kill at Cooper, the lost or discarded small points were replaced by those made from Alibates procured at the quarry sometime earlier. Because the discarded assemblage at Cooper is dominated by Alibates, it is assumed that the overall assemblage was dominated by points made of this material. The number of kills prior to Cooper or when Alibates became the dominant lithic type is unknown. What is known, however, is that other lithic types, including Edwards Plateau chert and Niobrara were still in the curated tool assemblage even given the diminutive size of the points.

Discussion

This model, once again, is primarily concerned with the lithic assemblage and, in so doing, is elevating lithics from an embedded position in the system. However, the principal theme espoused here remains consistent with the re-embedding of the lithics back into the subsistence system. For example, if the Folsom group is primarily following/searching for bison, then their trek across the landscape could be quite erratic. Rather than a cyclical movement between two or three lithic source areas, the search path could continually place the group near the same lithic source time and time again. A group that consistently found itself near the Edwards Plateau source would have lithic assemblages dominated by that type regardless of the distance they eventually wandered from the source. This could explain the dominance of Edwards Plateau chert in the assemblages throughout west Texas. Only on occasion might the group travel into the area of Alibates while following/searching for bison. On very rare occasions this same group might continue northward into the vicinity of Niobrara material. The primary goal was to find bison; finding suitable lithic material was secondary.

Even though archeologists emphasize the type of lithic material, the Paleoindians might have emphasized the tools. The deciding factor was not what type of stone the tool was made of but whether or not the implement was still perceived as capable of performing its task. This is not to say that Paleoindians did not have a preference for one lithic source over another. Instead, even given the availability of a suitable lithic source, individuals did not immediately replace every tool in their tool kit with the newly acquired material. Tools were replaced with specimens of the most recently acquired material when they were lost or broken. The decision is based on the assemblage, not the lithic type. The determining factor was the suitability of the assemblage to meet the perceived needs of the group. New lithic material was only employed when the assemblage was found lacking in some way.

Because of the difference in the use-life between expediency and curated tool forms, curated tools would reflect a greater number of lithic types in a system that routinely exploits different lithic source areas. On the other hand, expediency tools at a given site would be dominated by one lithic type, that of the last source visited.

If lithic procurement is embedded, then Folsom groups could move about the landscape following bison herds in any direction and not be concerned if they were near a lithic source until replenishment of the raw material dictated a quarry visit. Then, while in pursuit of bison, the group could stop by the nearest quarry. A highly mobile circuitous movement across the Southern Plains landscape would ensue. Only the scarcity of bison (and maybe extreme seasonal conditions) would force groups into other areas containing alternative lithic sources. Until seasonality studies are completed in more sites across the Southern Plains the exact nature or direction of movements cannot be ascertained. The model of lithic procurement, use, and discard outlined above explains all patterns of stone uses seen on the southern Plains except one: The lack of Alibates points and tools on sites near the Edwards Plateau. If Alibates is being picked up by the same groups using Edwards Plateau chert, then Alibates tools and points should be transported back to the Edwards Plateau in use-assemblages.