Abstract
Nodule analysis is designed to highlight the ways in which different lithic sources were utilized and incorporated into the stone tool industries of past societies. In 2008 and 2009, excavations in the Piceance Basin of Northwestern Colorado, an area with local chert and quartzite quarries, provided an opportunity to use a nodule analyses for a Section 106-driven project. The Aught-Six site lithic assemblage suggests that Archaic tool kits in Northwestern Colorado are likely to be heavily reliant on a variety of raw materials found across the region, though Bridger chert nodules were the primary objective pieces for the new tools during the basin house occupation. The results of the analyses not only highlight specific areas of projectile point production, but they also indicate that nodule analysis remains a powerful analytical method for understanding how raw materials are incorporated into the technological systems of semisedentary Archaic foragers.
Introduction
Minimum analytical nodule analysis (MANA) is a powerful method for understanding lithic technological organization (Andrefsky, 2009: 84; Larson and Kornfeld, 1997). MANA has been predominantly used in North America to understand lithic reduction strategies of mobile, typically Paleoindian, foragers (Bruce, 2001; Cooper and Meltzer, 2009; Hall, 1998; Hurst et al., 2010; Knell, 2004, 2007, 2012; Larson, 1990; Prasciunas, 2014; Sellet, 1999) and lithic industries among horticultural groups (Miller, 2016; White, 2012), though not among semisedentary Archaic foragers. MANA is most useful in regions where the geologic sources for toolstone are known and have a high degree of internal variability (e.g., color, texture, and inclusions) because it strikes a cost-effective balance between a reasonable level of analytical effort and an increased detail of acquired information while avoiding the time-intensive efforts of refitting (Knell, 2004, 2007, 2012; Larson, 1990, Larson and Kornfeld, 1997; Prasciunas, 2014). Recent Section 106-driven excavations in Northwestern Colorado offer the opportunity to use nodule analysis to better understand stone tool industries of Archaic foragers who utilized houses during their seasonal rounds.
Nodule analysis background
The origins of MANA lie in the development and application of refit studies (Frison, 1974; Kelly, 1985; Larson, 1990; Larson and Kornfeld, 1997). While refitting can provide detailed insights into technological organization, it is time-consuming (Bamforth and Becker, 2000; Larson and Ingbar, 1992; Stafford, 1985; Yerkes, 1989). Additionally, the prehistoric removal of the final tool kit, be it formal tools or large usable flakes, may result in large voids within the refitted nodule, causing the refit effort to be difficult, if not impossible (Larson and Kornfeld, 1997). Because it is time-consuming, and since refit lithic materials are typically sorted into similar raw material groups before refitting begins, “… the pieces in a nodule share a specific constellation of features that differentiate these pieces from others of the same raw material type” (Larson and Kornfeld, 1997: 4). This realization, at its core, is the basis for MANA (Kelly, 1985; Larson, 1990; Larson and Kornfeld, 1997; Miller, 2016).
The initial application of MANA is found in Kelly’s (1985) dissertation. The first step in MANA involves sorting the lithic debitage into raw material categories and then subdividing each raw material into more distinct groups based on variations in color and inclusions (Larson, 1990; Larson and Kornfeld, 1997, Miller, 2016). To increase the degree of coherency within nodule groups, the sorting is accomplished through a time-consuming process that involves the labeling of every artifact so that all of the debitage and tools from across the breadth of excavations can be spread out on a table in front of the analyst, intermixed, and visually compared without losing their provenience. Then, through comparison of color and inclusions, small groups of artifacts with similar color patterns, textures, and inclusion are grouped into minimum analytical nodules (MANs), which are the base units for MANA (Larson and Kornfeld, 1997: 7).
Knell (2004) introduced the concept of coarse-grained nodules, which he defined as nodules based on the geological origin of the raw material source and dominant colors (e.g., White River Group silicate). Knell dubs these coarse-grained MANs generalized nodules since they do not have the specificity of MANs, yet are more precise than basic raw material categories (i.e., general chert and general quartzite).
Once nodules are defined, whether true MANs or generalized nodules, analysts divide the nodules by the composition, frequency, and number of artifacts in each nodule group. Originally, this was done by dividing nodules into single- and multiple-item nodules (Larson and Kornfeld, 1997: 10–11). Single-item MANs are those groups with a single tool or piece of debitage that likely represent the presence of tools that were brought to the site in a completed form (Larson and Kornfeld, 1997). Multiple-item MANs are those that contain more than one artifact and likely represent on-site tool production from a larger objective piece (Larson and Kornfeld, 1997: 11). More recently, analysts have defined reduction scenarios by which to infer how each nodule was related to the movement of materials and tools through a site (Hall, 1998; Knell, 2004, 2007, 2012; Prasciunas, 2014; Sellet, 1999). That is, by understanding the form in which raw material was brought on-site and reduced into tools, it is possible to understand the differences between raw material acquisition and production (i.e., Scenarios 1, 2, and 3) versus the transport and use of previously manufactured items (i.e., Scenarios 4 and 5) (Knell, 2012: 331). The early stage reduction scenarios (e.g., Scenarios 1 and 2) are intended to represent the entire manufacturing process wherein raw or minimally tested nodules are brought to the site (or acquired on-site if it is a quarry) and reduced or modified. The functional difference between the two early stage scenarios is based on whether or not a tool was left at the site (Scenario 1) or transported away (Scenario 2). Scenario 3 represents a preform that is brought to the site and modified, though it does not distinguish between situations when tools are either discarded on-site or carried away. Scenarios 4 and 5 both include nodules that are brought on-site as a finished tool and then modified or maintained before being either carried away (Scenario 4) or discarded (Scenario 5). The five scenarios are then used as test hypotheses for understanding the role of each generalized nodule in the technological system of mobile Paleoindian foragers at the Hell Gap Site (48GO305) in Wyoming (Knell, 2012). If analysts use single- versus multiple-item MANs or more explicit scenarios, by dividing debitage collections into nodule groupings, archaeologists can better understand the interplay of raw material acquisition, stone tool production, tool use, and hunter-gatherer mobility.
The analysis presented here uses a middle ground of analysis that is more specific than generalized nodules (Knell, 2004, 2007, 2012) and less precise than true MANs (Larson and Kornfeld, 1997) to interpret a lithic assemblage from a Section 106-driven project. The following section provides an overview of the site and the excavations, as well as a description of our analytical approach that helped streamline the identification of nodules. The analysis provides insight into the raw material use and mobility strategies of hunters-gatherers (Andrefsky, 1991) while highlighting the way that a nodule-based approach can be incorporated into the routine artifact analysis of Section 106-driven projects.
Site background
Site 5MF3006, the Aught-Six site, is a 7-hectare prehistoric open campsite 16 km north of the Yampa River in Northwestern Colorado (Figure 1). The site is located on a gentle southeast-facing slope at an elevation of 1942 m along the West Prong of Spring Creek in the Spring Creek Valley, which is within the structural Sand Wash Basin of the Greater Green River Basin. The site crosses Bureau of Land Management (BLM) and private lands, though the work described herein occurred only on lands administered by the BLM-Little Snake Field Office.
Location of the Aught-Six site with reference to local toolstone sources.
The Aught-Six site was first recorded in the early 1990s as a surface expression of four artifact concentrations and a thermal feature surrounded by scattered lithic artifacts and fire-cracked rock (Pennefather-O’Brien et al., 1992). The site was further surveyed and excavated in 1992, 2005, 2006, 2007, and 2008 as part of Section 106 compliance work associated with the collocation of four pipelines. That work provided numerous radiocarbon dates and diagnostic projectile points that demonstrated multiple occupations at the site ranging in age from 6500 to 100 cal. BP (Alexander and Reed, 2009; Firor and Reed, 2009; Greubel et al., 2008; Landt, 2014; Landt and Prouty, 2017; Metcalf and Elkins, 2011; Metcalf and Slaughter, 2005; Redman and Chandler, 2004).
The main focus of excavations at the Aught-Six site since 1992 is a dark gray, charcoal-enriched horizon known as the Spring Creek Anthrosol that underlies an abrupt erosional contact (Landt, 2014; McFaul and Metcalf, 2009, 2011a, 2011b; Metcalf and McFaul, 2005). The Spring Creek Anthrosol was likely formed from 6360 to 4572 cal. BP during a time when several environmental proxies document a trend of increased seasonality and rainfall during the middle Holocene that created a stable aggrading landscape, which was followed by a shift to temperatures near modern levels that allowed for the development and preservation of the Spring Creek Anthrosol after 5172 cal. BP (Madsen et al., 2009). Holocene sediments at the site consist of weakly developed, stabilized yellow–brown alluvium with a sandy loam to medium coarse sand texture (Landt, 2014; McFaul and Metcalf, 2009). Despite being in a rain shadow in an arid environment with well-drained sediments, the sandy soils support a xeric-dominated floral community that includes sagebrush (Artemisia sp.), numerous varieties of prickly pear cactus (Opuntia sp.), greasewood (Sarcobatus sp.), and bunch grasses (Gramineae), with isolated stands of Utah juniper (Juniperus osteosperma) on surrounding hillsides.
The site is in a large lithic-rich region that contains numerous sources of toolstone-quality chert, siltstone, orthoquartzite, and metaquartzite and past occupants of the site would rarely “have been more than a brisk one- or two-day walk from … outcrops of serviceable toolstone” (Gould and Saggers, 1985; Greubel et al., 2018: 161; Meltzer, 1989). Notable formations with recognizable toolstone materials include Bridger, Morgan–Madison, Uinta Mountain Group, and Green River formations, some of which can also be found as lag deposits and in valley-bottom alluvium with mixed cobbles of flakeable quartzite, chert, and other materials. Because the essence of nodule analyses lies in intramaterial similarities and difference in toolstone, the following sections describe some of the most identifiable sources of available toolstone in the region.
Morgan–Madison
The Morgan and Madison formations are geologically distinct, and the cherts they yield typically include orange–red hues and may include varieties that are locally identified as jaspers (Greubel et al., 2018). Morgan–Madison cherts can be found within 21 km of the Aught-Six site near Juniper Mountain (Greubel et al., 2018: Table 43), as well as further downriver near the confluence of the Little Snake and Yampa rivers (Figure 1).
Uinta
The Uinta Mountain Group yields a Precambrian metaquartzite with a red or reddish purple hue (Greubel et al., 2018; Miller, 1991). Uinta quartzites are widely distributed by alluvial processes and have been discovered many kilometers from the nearest mapped Uinta Mountain Group sources (Redman and Chandler, 2004; Stahl, 2009; Tweto, 1979). A documented exposure of the Uinta Mountain Group outcrops 24 km away on Juniper Mountain (Figure 1), which is a rough proxy for distance to acquire the toolstone (Greubel et al., 2018: Table 43).
Bridger
The Bridger formation mainly outcrops in the Sand Wash Basin, the heart of which is roughly 32 km west of the Aught-Six site (Figure 1). The Sand Wash Basin contains abundant chert beds, chert nodules, petrified wood, and silicified stromatolites (Buchheim et al., 2000; Ingalls and Park, 2010; Miller and Poulson, 2010; Murphey and Daitch, 2007; Stucky, 1977; Stucky et al., 1996). Some of the cherts in the Sand Wash Basin are referred to as tiger chert, as are some cherts from Southwestern Wyoming, because of the distinct and alternating, light and dark brown banding (Greubel et al., 2009; Loosle, 2000; Miller, 1991). Bedrock layers of chert were prehistorically mined in the heart of the Sand Wash Basin, while chert nodules and silicified organics are available as out washed lag deposits and isolated clusters in and around the southern perimeter of the basin (Landt and Watkins Morris, 2018). The amount of cortex on a nodule may dramatically differ if it was acquired from a bedrock mine or from lag deposits (Figure 2). If quarried from chert beds or lag deposits, all of the silicified lithics in the Sand Wash Basin were usable as raw material and all of which are typically identified as Bridger Formation cherts (Miller and Poulson, 2010; Stucky et al., 1996).
Examples of Bridger chert nodules.
Green River
The Green River Formation outcrops across a large area as surface exposures (Greubel et al., 2018; Love, 1977; Miller, 1991). Some of the best known outcrops of the Green River formation are along the Delaney Rim and near Wamsutter, Wyoming, both of which are 97 to 110 km from the site (Figure 1), though the Aught-Six site is mapped within roughly 2 km of a surface exposure of the Green River Formation (Tweto, 1979). The most easily recognizable variety of Green River cherts is of an oolitic variety (i.e., Wamsutter Oolitic) that contains densely packed spherical oolites (Miller, 1991: 591). However, because numerous types and colors of cherts can be recovered from the Green River Formation, it is likely “that a substantial portion of the materials classified as other chert originated in the Green River Formation” (Greubel et al., 2018: 163).
Project excavations
Radiocarbon dates for the basin house occupation.
Note: n/a: not applicable.
All dates were pooled and calibrated with Calib 6.1.0.
The excavations provided a large collection of lithic debitage and tools that was mainly made of materials from the Bridger formation (Figure 3), which are available within 30 km of the site in the Sand Wash Basin. Because materials from the Bridger formation contain a high degree of internal variability (Bradley, 1929; Keesling, 2006; Miller, 1991; Stucky, 1977; Stucky et al., 1996), lithic artifacts from the Aught-Six site provided an opportunity to use a nodule-based analysis to better understand the nature of raw material acquisition and use during the Archaic era of Northwestern Colorado.
Examples of Bridger chert flakes.
Methods
Basin house flaked stone assemblage.
Unlike traditional MANA studies, artifacts were assigned to nodule categories, as the technological analysis was being conducted. From a practical perspective, laboratory space is often at a premium, and it is difficult to leave a large number of artifacts out while defining nodule groupings. Additionally, the process of labeling all of the debitage and tools to maintain proper provenience, while they are left out is cost prohibitive. Initially, the debitage and tools were assigned one of the nine coarse geologic raw material types: Bridger chert, Morgan–Madison chert, oolitic chert, other chert, chalcedony, Uinta quartzite, other quartzite, petrified wood, and other materials. A comparative raw material collection housed at Alpine was used to classify lithic artifacts from the Aught-Six site. Debitage was sorted through a series of standard grading sieves (e.g., 2.5 cm, 1.3 cm, 0.64 cm), and nodule codes (NCs) were given to artifacts larger than 0.64 cm in size. The 512 smallest pieces of debitage (e.g., those that passed through a 0.64 cm sieve) can be included in a general material type analysis but are excluded from the nodule analysis (Larson and Kornfeld, 1997). Numerical NCs (e.g., NC 76) and a description of the nodule (e.g., transparent dark brown Bridger chert with medium brown mottling) were assigned to each material type when a new color or new type of inclusion was encountered. Thus, our analytical nodules were created one at a time during the traditional lithic analysis and are not as time intensive to create as a refit nodule. This had the advantage of significantly reducing the amount of time needed to conduct the analysis because the artifacts did not have to be individually labeled. Moreover, only a relatively small space was needed to conduct the analysis.
Having assigned NCs during the technological lithic analysis, we then reevaluated the nodule descriptions and specific artifacts in an effort to provide a more precise intrasite comparison. The primary goal of this reevaluation was to lump the nodule categories back together and minimize any intranodule splitting that may have occurred during analysis. That is, was Bridger chert NC 26 (e.g., a light gray to tan to white chert) the same as Bridger chert NC 74 (e.g., a gray and matte tan chert)? Alas, they are not the same. We consulted the descriptions, the actual artifacts, and their archaeological context to assign the artifacts into site-specific nodule categories.
Results
Geologically, Bridger chert was the most frequently recovered raw material (68%), followed by nonspecific other cherts (15%), Morgan–Madison cherts (9%), and two types of quartzite (6%). Various other materials are present, but when combined are less than 3% of the assemblage (Table 2). A significant difference exists between the raw material categorization of the 512 small flakes that passed through a 0.64 cm sieve, which are excluded from the nodule analysis, and the debitage that was assigned an NC (χ2 = 112, df = 5, p < 0.001 Cramer’s V = 0.11). The difference is due to an unexpected decline in Bridger chert among the small flakes and a commensurate rise of small debitage in the other chert category. The difficulty in assigning small flakes to geologically specific nodule categories, whether Bridger chert or not, is taken as support for the exclusion of the small debitage from further nodule analysis herein.
Basin house nodule distribution.
The majority of the 27 tool-bearing nodules are made from Bridger chert (n = 17), though tools were also made from other cherts (n = 5), Morgan–Madison cherts (n = 3), a nondescript quartzite, and an oolitic chert. Seven of the tool-bearing nodules consist of singular chert bifaces of Bridger, Morgan–Madison, and nondescript cherts (NCs 54, 314, 386, and 47, respectively), a Bridger chert projectile point (NC 351), an oolitic chert scraper (NC 803), and an informal Bridger chert flake tool (NC 392) that are not associated with debitage of a similar nature. Four additional tool-bearing nodules consist of a broken projectile point of Morgan–Madison (NC 10) or nondescript chert (NC 106), or a flake tool of Bridger (NC 84) or Morgan–Madison chert (NC 137), which is associated with two or fewer pieces of debitage. Those 12 small tool-bearing nodules were likely brought to the site in a nearly finished form and were discarded as part of a maintenance trajectory. A nondescript tool-bearing quartzite nodule (NC 704) consists of a chopper and six pieces of debitage, which likely indicates the production of an expedient tool out of material in the nearby Spring Creek drainage that was then discarded. The remaining 16 tool-bearing nodules are associated with at least 10 pieces of debitage and are likely part of a production trajectory wherein preforms or relatively raw nodules were brought to the site and reduced into tools. In four cases, a unique Bridger chert nodule was reduced into five or more different tools and comparable amounts of debitage: two bifaces and five informal flake tools (NC 144); a projectile point, a biface, and four flake tools (NC 76); two bifaces and four flake tools (NC 70); and two bifaces with three flake tools (NC 192). Thus, while 40% of the unique nodules were a variety of Bridger chert, 80% of the tools were made from Bridger chert nodules, which is likely indicative of a high degree of Bridger chert tool production.
Regardless of tools and tool-production nodules, the majority of the nodule categories consist solely of debitage (83%, Table 3), suggesting that distinctions in the debitage assemblage might reveal additional information about tool maintenance and production at the Aught-Six site. Of the debitage-only nodules in the maintenance trajectory (e.g., those 98 nodules with one or two pieces of debitage), 36% of the debitage are technologically indeterminate and 32% are the result of bifacial reduction. Roughly one-quarter of the maintenance trajectory debitage are core-reduction flakes, and the remaining 9% are pieces of angular shatter. These patterns are proportionally consistent even when the flake types are split by raw material categories. Of the debitage-only nodules in the production category (e.g., those 42 nodules with multiple pieces of debitage), 32% of the debitage are angular shatter, 30% are the result of core reduction, 25% were bifacial reduction, and the remaining 13% are technologically indeterminate. The distribution of flake types in the production trajectory is significantly different when divided by raw material types (χ2 = 151.6, df = 9, p < 0.001 Cramer’s V = 0.13). The difference is driven by a shift in flake types among Bridger nodules in the production trajectory, such that debitage-only Bridger nodules tend to have a greater than expected amount of angular shatter and core-reduction debris, while all other raw materials tend to have more bifacial or indeterminate flakes, like those in the maintenance trajectory. While Tomka (1989) argues that angular shatter is a product of core reduction, some flint knappers make more angular shatter during early stage biface reduction than others who are reducing cores (Williams and Andrefsky, 2011), which suggests that angular shatter is more likely created during the earliest stages of reduction and not necessarily just during core reduction. Thus, while debitage-only nodules associated with the maintenance trajectory tend to reflect bifacial reduction, the debitage-only nodules associated with a production trajectory tends to exemplify a focus on the early stage reduction of Bridger chert nodules and the bifacial reduction of all other materials.
Because we did not excavate the entire site, we recovered only one core, and the smallest pieces of debitage, which could be small fragments of shatter or the detritus from retouching tools, were excluded from the nodule analysis, we did not attempt to use specific scenarios (Knell, 2004, 2007, 2012: 341; Prasciunas, 2014). Rather, we recognize that Scenarios 1 and 2 represent the primary reduction of relatively unmodified raw material with the intent to produce a preform or tool. Scenario 3 represents the introduction of a preform that is reduced to a tool, though it can be difficult to determine whether Scenario 3 exists as an isolated reduction trajectory or it appears as such because of sampling (Prasciunas, 2014; Sellet, 1999). Distinguishing between early stage production (e.g., Scenario 1 or 2) and middle stage reduction (e.g., Scenario 3) debitage is inherently difficult from excavation samples, especially when mined Bridger chert nodules may lack cortex when acquired. Within the Aught-Six data set, the distinction between early and late production scenarios utilized cortex, as well as sizes, and technological characteristics (e.g., reduced amounts of shatter and degree of platform scaring). Tool maintenance and discard trajectories (e.g., Scenarios 4 and 5) represent the final stages of tool use and curation. The prevalence of maintenance trajectories (e.g., Scenario 4) and debitage-only reduction strategies (e.g., Scenario 3) in the Aught-Six assemblage suggests that many more tools were retouched and not recovered than were actually made on-site.
Nodule categories at the Aught-Six site.
Middle stage reduction trajectory nodules reflect the introduction of a preform to the site, which is then reduced to a tool (e.g., Scenario 3) (Knell, 2012). Debitage produced in this way typically lacks evidence of primary reduction, be that flakes with a large percentage of cortex or high proportions of angular shatter amid the debitage. We identified 48 middle stage reduction nodules at the Aught-Six site, and while there are anywhere from 2 to 41 pieces of debitage per nodule, the majority of the aggregated debitage is either technologically indeterminate or the product of bifacial reduction. Of these nodules, 38% are varieties of Bridger chert (Table 4). Equally, 4 of the 10 middle reduction tools are Bridger chert and 42% of the debitage are Bridger chert. The equivalent and relatively low percentage of Bridger chert among the reduction nodules indicates a shift from the previous tool-production scenarios to a consistent focus on the reduction of a bifacial preforms from a diversity of materials.
Nodules in a maintenance trajectory (e.g., Scenario 4) indicate the resharpening or retouching of a tool that was transported to the site in a finished state, though the tool remains absent (Knell, 2012). That is, the maintenance debitage should reflect the final stages of tool use and typically consist of a few small flakes with multifaceted platforms and no cortex (Knell, 2004: Table 8.2). Even with the exclusion of the smallest 512 flakes, we identified 96 maintenance nodules at the Aught-Six site, 35% of which are Bridger chert (Table 4). Each maintenance nodule contains either one or two flakes and the aggregated debitage at the Aught-Six site is nearly evenly divided between biface and core-reduction strategies, though the simple majority of flakes are technologically indeterminate. We included a nodule with an informal flake tool (NC 84) in the maintenance trajectories because it is only associated with one piece of shatter, which suggests a core was maintained as an expedient tool source. More than anything, the high number of maintenance nodules and the diversity of raw materials indicate that the majority of a flint knapper’s tool kit at the Aught-Six site consisted of a variety of materials.
Discarded nodules (e.g., Scenario 5) consist entirely of tools that were transported to the site and left or lost behind. There is little to no expectation for associated debitage in this scenario (Knell, 2012). We might expect one or two pieces of maintenance debris before the tool is discarded, though some of those may be included in the smallest debitage that was not assigned NCs. Of the eight discarded tools, which are all different nodules, four are Bridger chert. Only one of the eight tools is an informal flake tool and the rest are formal tools (i.e., four bifaces, two projectile points, and a scraper). The low number of discarded tools in the assemblage likely indicates that the Archaic tool kits retained a high degree of utility or that the tools were discarded in activity areas beyond the extent of our excavations.
At a coarse level, the Aught-Six assemblage suggests that an Archaic tool kit in Northwestern Colorado is likely to be heavily reliant on a variety of raw materials found across the region, though Bridger chert nodules were the primary objective pieces for the creation of new tools during the basin house occupation. The combination of early stage production nodules and middle stage reduction nodules suggests that the Aught-Six site occupants found it worthwhile to carry large nodules of Bridger chert. However, because 90% of the debitage assemblage lack cortex, the sources are far enough away that foragers found it inefficient to carry extra “waste” rock or undesirable material (Beck et al., 2002; Metcalfe and Barlow, 1992; Surovell, 2011). This may also suggest that regardless of the proximity of bedrock exposures, any artifacts made from the Green River Formation were not sourced locally (Beck, 2008; Elston 1990; Jeske, 1989; Jones et al., 2003). Additionally, the large proportion of maintenance nodules suggests that the tool kit of the middle Archaic foragers was not in need of being replaced. Rather, the tool kits were being supplemented with locally sourced materials. The Aught-Six assemblage indicates that prehistoric tool kits contained a large variety of nodules from numerous sources (i.e., maintenance trajectory), though the bulk of the on-site debitage was produced from locally available Bridger cherts that were easy to acquire and could be further reduced into tools and preforms at the Aught-Six site (i.e., early stage production).
Spatial patterning of nodules
Nodule categories can be used to identify spatial clusters that equate to singular flint-knapping events. In an effort to identify meaningful clusters, four early stage production nodules that consisted of more than 100 total flakes (e.g., greater than 5% of the assemblage) were individually assessed by technomorphological organization. While the flakes tended to be distributed across the site, three discrete NC clusters were identified, each of which likely indicates the production of tools around the basin house.
The first raw material cluster contains 38 pieces of debitage from matte tan to gray Bridger chert (NC 70) that surround a roasting pit (F3 in Table 1) located roughly 5 m behind and uphill of the basin house (Landt, 2014: Figure 56). The debitage includes 19 indeterminate flakes, 4 core-reduction flakes, and 15 early stage biface flakes. None of the debitage retains cortex, suggesting preliminary reduction elsewhere. No NC 70 tools were recovered from that area, though non-NC 70 tools were recovered. The NC 70 cluster is suggestive of early to middle stage biface production where the tool was utilized and carried elsewhere.
NC 76, which is a transparent dark brown Bridger chert with medium brown mottling, clusters in two areas of the site (Landt, 2014: Figure 57). The first cluster contains 40 pieces of debitage, most of which are indeterminate fragments that surround a hearth located level with and 6 m to the left as a person would have exited the basin house. Nine of the flakes are early stage biface flakes found in adjoining units that all measure 6 to 13 mm in size. The second cluster consists of five early stage biface flakes, two pieces of angular shatter, and three technomorphologically indeterminate flakes that were found in four adjoining units next to a hearth located 7 m behind and uphill of the basin house, though 7 m west of the NC 70 cluster above. Although the biface flakes range from 13 to 28 mm, the rest of the debitage are 6 to 13 mm in size, all of which was recovered in a 0.64 cm mesh. The only bifacial tool of NC 76 is a complete Elko Corner-notched projectile point that is located in the second cluster. The bifacial reduction event in the second cluster may have produced the projectile point, which was then abandoned or lost in the sandy sediments of the site, though the product of the bifacial reduction event in the first cluster was carried outside the extent of our excavations and likely away from the site.
The final clusters are made from a semitransparent to opaque brown/gray Bridger chert (NC 144). NC 144 was a popular material, as evidenced by scattered densities across the excavation (Landt, 2014). Most of the NC 144 debitage are indeterminate flakes or are a mix of biface and core-reduction flakes with no clear reduction trajectory. There are, however, two clusters of NC 144 that may indicate individual reduction events. The debitage in the first cluster, which is next to the same roasting pit (F3 in Table 1) with the NC 70 cluster above, was mostly bifacial reduction, though the cluster contains an expedient flaked tool and a few core-reduction flakes that may suggest that a core and biface of NC 144 were reduced in the same location. The NC 144 material in the second cluster, which overlaps with the aforementioned projectile point cluster of NC 76, consists of 10 early stage biface flakes that are associated with five indeterminate flakes. As with NC 76 above, the second NC 144 flake cluster is also associated with an NC 144 Elko Corner-notched projectile point. The presence of multiple discrete and identifiable MAN clusters with projectile points in the same location suggests a patterned behavior. The fact that all of the identified clusters are related to biface reduction and two of those events are also related to a projectile point suggests that the nodule analysis is sensitive enough to isolate discrete locations of tool production. In this way, and if each nodule can be viewed in terms of reduction scenarios and as a limited set of events deriving from singular objective pieces (Knell, 2004, 2007, 2012; Larson and Kornfeld, 1997), then the consequent nodule counts reflect a conservative number of flint-knapping events at a site or in a component.
Technological organization
The patterning of nodules and technological reduction trajectories can also offer insight into the mobility patterns of the site’s occupants; core-reduction-focused tool discard events are similar to lithic schemes ascribed to sedentary populations, whereas bifacial tools are often indicative of mobile populations (Parry and Kelly, 1987). Excluding a few relatively intact projectile points, the abandoned tools were typically expedient tools, bifaces broken during manufacture, or abandoned early stage bifaces (e.g., Stage 2), all of which maintain minimal utility. Thus, the limited nature of high-utility tools within the component suggests that the tools simply represent a normal discard rate at the Aught-Six site and not a specific technological strategy, per se. That is, previously worked nodules, whether preforms or finished tools, were brought to the site and worked or maintained. A few tools were lost or discarded in the sandy sediment of the site, and the bulk of the tools/objective pieces were carried away from the site. A group that foresees travel away from raw material sources or that needs diverse tool kits is expected to conserve and carry a large number of nodules. The nodule analysis shows that while the bulk of the on-site production was dedicated to making Bridger chert tools, the majority of the tool kit nodules (e.g., objective pieces) were removed from the area, which is taken as further evidence that the site occupants were highly mobile groups or individuals who conserved their tool kits by making tools from local materials during the basin house occupation.
Paleoindians are typically perceived as a highly mobile group and nodule analyses at Paleoindian sites in eastern Wyoming (e.g., the Cody-complex at Hell Gap site locality I and V or Clovis occupations at the Sheaman site) can be used as a standard by which to assess the degree of mobility among Archaic foragers who seasonally lived in semisubterranean structures (Knell, 2012; Prasciunas, 2014). Within the Cody I assemblage at the Hell Gap site, which represents a campsite near abundant raw material (Knell, 2012; Prasciunas, 2014), many more tools were discarded than were produced (e.g., emphasis on a discard trajectory or Scenario 5). Alternatively, at the campsite/workshop Cody V assemblage (Knell, 2012), numerous preforms were reduced to tools as middle stage reduction events (e.g., Scenario 3). The Sheaman site assemblage, which has also been interpreted as a campsite, indicates a focus on the continued reduction of nonlocal bifacial tools that were then removed from the site (e.g., emphasis on Scenario 3 or 4, though further distinction between the two scenarios were not made) (Prasciunas, 2014). By comparison, the Aught-Six site assemblage is more akin to that of the Sheaman site, although at the Aught-Six site, the majority of the recovered tools were expediently manufactured of local material during early stages of production (e.g., Scenario 1) and the majority of the tool kit during the basin house assemblage is focused on maintaining tools that retain high utility (e.g., Scenario 4). If the scenarios have any relationship to the mobility of the groups, then the occupants of the Aught-Six site took the opportunity to manufacture some new Bridger chert tools, though the bulk of their tool kit remained focused on the conservation of previously constructed tools, which might further suggest that they were frequently near areas of ubiquitous raw material during their seasonal round or that they had just moved from a Bridger chert source area.
Conclusions
Nodule analysis revealed that the emphasis of the lithic industries at the site was the maintenance of tools that retained utility (e.g., Scenario 4), with a secondary emphasis on the production of Bridger chert tools which were then likely transported off-site (e.g., Scenario 1) sometime between 6500 and 6300 cal. BP. The various lines of evidence all seem to reflect a brief, seasonal occupation in support of local foraging or resource collection, and the presence of the basin house suggests that the site is a residential base at the center of all subsistence and cultural activities (Binford, 1980; Landt, 2014). The core reduction that took place was likely conducted to make flake blanks or produce expedient tools, which were used during their occupancy at the site. While the majority of the debitage was produced during the manufacture of Bridger chert tools, the small proportion of middle stage reduction tools suggests that numerous bifaces were transported off-site. This is in stark contrast to the high number of tools that were maintained at the site, which indicates that the site occupants conserved the raw materials, as would be expected of highly mobile people. The use of analytical nodules suggests that the site occupants were utilizing a wide variety of nearby raw material sources while camping and took the opportunity to replenish their tool kit before moving to a new camp area.
Nodule analysis is a powerful analytical tool that can greatly enhance the interpretation of lithic assemblages and identify individual activities in a 6400-year-old archaeological site. Our analysis also highlights the way that lithic assemblages with a high degree of raw material variability (e.g., Bridger chert in Northwestern Colorado) are managed and utilized by semisedentary Archaic foragers who live in an area rich with toolstone sources. Practically, we demonstrated that using nodule analysis can produce meaningful results on the scale of Section 106-driven projects.
Footnotes
Acknowledgments
The authors would like to thank Anthony T. Boldurian, Editor, North American Archaeologist, Rand A. Greubel, Christopher T. Hall, Edward J. Knell, Justin D. McKeel, and Michael J. Prouty, for their comments, guidance, and advice on earlier versions of this paper. Any errors of facts or misinterpretations are ours alone. Thanks also goes out to the Bureau of Land Management-Little Snake Field Office who continues to advance science in a rural corner of Colorado All artifacts from the excavations are curated at the Museum of Northwest Colorado, Craig, CO.
Declaration of Conflicting Interests
The author(s) declared no potential conflicts of interest with respect to the research, authorship, and/or publication of this article.
Funding
The author(s) disclosed receipt of the following financial support for the research, authorship, and/or publication of this article: The financial support during the excavation and analyses of the assemblage as well as during the writing of this paper was provided by Alpine Archaeological Consultants, Inc.
