KRF look-a-likes in Tennessee

30 Days of Tennessee Archaeology, Day 22

Ryan M. Parish, PhD and Ellis Durham
University of Tennessee, Memphis

There often exist significant overlapping visual characteristics both within and between chert sources.  This fact coupled with our propensity to identify raw materials as exotic potentially causes misidentifications of prehistoric stone artifacts.  In turn, the identification or misidentification of chert artifacts leads to conclusions regarding social networks, mobility, and human behavior that may be unsound.  Knife River Flint (KRF) was a widely utilized and circulated tool stone in the northern Great Plains throughout the entire prehistoric record.  It is a common material type found in artifact assemblages throughout North Dakota and into Canada.  KRF is also a material type identified on sites in Maine and Florida and just about every state in between.  In Tennessee, KRF is rarely but occasionally identified usually as projectile points and other curated artifacts.  Are these artifacts really made from KRF, the source of which is located over 2,000 km (1,400 miles) away?


Figure 1.  KRF and Tennessee look-a-like flakes.  Can you guess which is which? *

Visually, KRF is said to be pretty distinctive, a translucent fine grained amber, root beer, beer bottle glass colored material sometimes with small white fossil inclusions.

However, researchers in the source region identify visually similar varieties of chert, chalcedony, petrified wood, and silicified peat.  The main prehistoric procurement sites of KRF occur in Dunn and Mercer County, ND as secondary gravel deposits thought to belong to the glaciated/eroded Golden Valley Formation. Glacial and alluvial deposits of KRF extend southeastward from Montana to Iowa.   The historically documented Hidatsa and Mandan tribes maintained a broad trade network within which KRF was exported although the true extent of which has yet to be confirmed using analytical techniques in most regions.


Figure 2. The Dodge Quarry site, ND. A typical undulating procurement site where KRF was dug out of the soil matrix. Sites such as this are common in Dunn and Mercer Counties. 

As Barbara Luedtke used to assert, ‘assume a local source first’, we must be critical of KRF claims in Tennessee.  What, if any chert sources in Tennessee or vicinity could be similar in description? First, before we speculate on possible local look-a-likes let’s briefly recap major toolstone sources in Tennessee.  The Central Basin (Nashville) contains older Silurian and Ordovician carbonate sources such as Brassfield (waxy, green/blue/gray/orange) and Bigby Cannon (fine grained, laminated brown/black).  Camden chert, Devonian in age, occurs just west of the Central Basin and provides a medium grained white chert with opaque inclusions.  The Western and Eastern Highland Rim contains abundant Mississippian epoch aged Fort Payne (coarse to fine grained white/tan/mottled brown/dark gray), Warsaw (fossiliferous blue/gray), St. Louis (waxy, green/blue/gray), and Ste. Genevieve/Monteagle (waxy, blue, gray) sources.  Cretaceous gravels; Horse Creek (medium grained tan/red/yellow/gray) exist along the Tennessee River at the edge of the Highland Rim and Coastal Plain of southern Tennessee, northeastern Mississippi, and northwestern Alabama.  The Coastal Plain in western Tennessee is dominated by Pliocene Upland Complex gravel (medium grained tan/brown) sources.  Finally, the Valley and Ridge of eastern Tennessee has Ordovician sources such as Knox chert (fine grained translucent black/blue) and quartzite (medium grained white/tan).

Though the authors do not claim to have conducted an exhaustive survey of tool stone sources, a fair amount of time visiting, surveying, and sampling various sources allows us to offer up a speculation of local KRF look-a-likes.  The extensive and diverse gravel deposits within and along the Mississippi River contain chert that certainly provided prehistoric people with a wide selection of materials.  However, we have never observed chert gravels that are fine grained/waxy or translucent.  The multi-colored Cretaceous gravels at the northern inflection point of the Tennessee River likewise do not contain material similar to KRF.  Neither do the Valley and Ridge sources appear to provide a visual match for KRF despite the range of white/opaque/blue/black translucent varieties present.  A match has not currently been observed in the Ordovician, Silurian, and Devonian chert sources of the Central Basin and vicinity.

That leaves the Mississippian aged sources of the Highland Rim, varieties of which overlap visually with KRF.  Specifically, the upper Fort Payne and lower Warsaw occasionally produces a root beer translucent chert with white fossil inclusions.  Also, both the Upper St. Louis and Ste. Genevieve/Monteagle sources are waxy and occasionally occur as tan/dark brown materials.  Some would refer to these varieties as a chalcedony. I would shy away from the use of the term chalcedony, as the KRF look-a-like materials do not appear to be dominated by a fibrous quartz crystalline structure.

The take away theme is this; that the many varieties of cherts that prehistoric people selected and used may sometimes have significant overlapping visual characteristics.  This fact makes identification of exotic material from long distances away appealing but potentially flawed.  The 240 lbs. of KRF collected from nine sources in North Dakota this summer may help us for the first time confirm or deny the presence of KRF in Tennessee using reflectance spectroscopy.


Figure 3.  Examples of KRF.  Note heavy blueish and white patina on some samples.

*Answer – KRF is the top sample.  The bottom two are probably a St. Louis chert variety.




Split-cane Technology from an Experimental Ethnoarchaeological Perspective

30 Days of Tennessee Archaeology, Day 23

Megan King
University of Tennessee, Knoxville

In the Southeast, split-cane technology is believed to represent an ancient tradition spanning thousands of years, and it remains a respected and valued art form among Southeastern Indians. When cane is found in the archaeological record it can take the form of formal artifacts, the remnants of basketry or matting, pieces of carbonized torches, or as impressions in clay or prepared surfaces. When cane baskets or mats are studied, the focus tends to be on recording and documenting stylistic treatments and construction techniques, and while this is important few scholars have attempted to investigate the tools, techniques, and by-products associated with split-technology. In the absence of direct evidence, stone tools might be one of the only means to identify the manufacture and production of this ancient technology in the archaeological record. As part of my dissertation research I wanted to know if stone tools would be efficient in the processing of river cane and I also wanted to document any wear that the processing left behind. I began by conducting a few cane processing experiments on my own based on descriptions in Sarah Hill’s Weaving New Worlds: Southeastern Cherokee Women and their BasketryThe mechanics sounded simple enough and I was optimistic, excited, and completely unprepared.

Hill described cane preparation as being quite demanding and notes that it is one of the most challenging of all raw materials used by contemporary Cherokee weavers. This I can attest to is an understatement. Cane processing is difficult, it is arduous, and it is dangerous. In the process of splitting, peeling, and scraping cane culms I was cut, scratched, punctured, and even needed eight stitches. It became evident after only just a day of working with cane that my inexperience was not only causing me bodily harm, but more importantly it was hindering my experiments. It became clear that the most effective way to replicate realistic use contexts of river cane and stone tools was to incorporate an ethnoarchaeological perspective, wherein the individuals using the tools were performing tasks in which they were practiced and proficient. As a result, I contacted Roger and Shawna Morton Cain, expert basket makers from the United Keetoowah Band of Cherokee Indians (UKB) in Oklahoma. Each have been awarded the honor of National Treasure within the Cherokee Nation for their contributions towards the preservation and revival of Cherokee art forms. Their knowledge of river cane and skill level as weavers is unparalleled, and while neither had experience processing cane with stone tools, they were given the opportunity to familiarize themselves with the technology and practice prior to our experiments.


Sample of Flake Tools Preferred for Cane Processing Experiments.


Experimental Cane Processing Experiments Conducted with Expert Basket Weavers and Cherokee Nation Treasures Roger and Shawna Morton Cane.

Over the course of several days I worked closely with Roger and Shawna. We visited a local canebrake where Roger carefully selected 10 culms for processing. When selecting culms for a basket Roger and Shawna follow the cycles of moon and tend to prefer those that were at least 3 to 5 years. They only harvest what they need and make a concerted effort to converse and manage canebrakes. Roger has even worked to establish the Cherokee Nation River Cane Initiative, which has been actively identifying and mapping existing canebrakes on tribal land in northeastern Oklahoma. While working together were able to process over a dozen river cane splits, and Roger processed an additional 120 splits after I left. Of the more than forty flake tools available for use, the Cain’s preferred large cortical backed flakes that could be used interchangeably to split, peel, and scrape the fibrous culms. They proved unequivocally that simple flake tools were efficient implements for processing river cane, and they even remarked that some of the flake tools were more efficient than modern steel blades.

While they worked I paid close attention to the tools as they were maneuvered along the culms and even took notice of refuse that accumulated at their feet. When I returned home I examined the tools for microscopic usewear and was surprised in the variability of polish development and distribution on the tools. Cane polish develops slower than I had thought and there were observable differences in polish formation on artifacts used to split, peel, and scrape the culms. In general, the polish formed in small isolated patches and gradually developed into a bright, voluminous polish that looked more woody than plant.


River Cane Polish in Early Stages of Development on Experimental Flake Tools Used to Peel and Scrape Culms.


River Cane Polish in Late Stages of Development on Experimental Flake Tools Used to Peel and Scrape Culms

Differentiating between river cane polish and wood might be difficult archaeologically, however, the location of the polish and the macroscopic damage that resulted from use might help future archaeologists make such distinctions.

These experiments underscore the need for additional experimental work with indigenous plants in the Southeast. They also demonstrate how important it is to work with source communities and utilize the knowledge of craft specialists and artisans who continue to work with native plants and carry on traditions of their ancestors. I feel privileged to have been worked with Roger and Shawna, to have been welcomed into their community and their home, and to have had the opportunity to watch them work with a plant that they are so passionate about and have essentially built their life and livelihood around.

A Lumper’s Take on Paleoindian and Early Archaic Projectile Points from the Mid-South

30 Days of Tennessee Archaeology, Day 16

Shane Miller
Mississippi State University

A few minutes ago, one of my colleagues, Derek Anderson, walked into my office, handed me a projectile point in a bag, and asked, “What would you call that?” In other words, what projectile point category, or type, is it? In this particular case, I shrugged and replied “Sykes or White Springs?” and we then proceeded to google those types so we could see pictures of other examples. We both decided it was a pretty good match. From this projectile point, he could say that the artifacts he was analyzing contained an artifact that dated to the Middle Archaic period.

Even though it was quick assessment, it was based on a large literature of archaeological sites, radiocarbon dates, and various projectile point guides like Cambron and Hulse’s projectile point guide, which is the most commonly used guide for the Tennessee River valley. There are also more far-reaching attempts to make sense of variability in projectile point types, like Noel Justice’s guide for eastern North America or the online Projectile Point Identification Guide.

I’ll be honest, I study this stuff for a living and I find the myriad of projectile point types daunting (not to mention the vast literature in archaeology on classification and typology), and in order to make sense of it all, I readily admit… I am a lumper. I look for similarities (rather than focus on the differences), and then compare those similarities in projectile point shape and design to the stratigraphic and radiocarbon record.

I take this approach for two reasons:

  1. It’s really hard for even modern knappers to replicate the same point shape over and over again. There’s a lot of contingencies to consider, from skill levels to variation in raw material. These are not things that were made on an assembly line.
  2. If I do decide to differentiate two types based on some minute detail, I prefer that decision to be backed up by a quantitative approach using geometric morphometrics. The words may seem fancy, but the premise isn’t – it’s a statistical analysis of shape.

Last year, Thad Bissett and I combined our spreadsheets of radiocarbon dates and stratigraphic information to a Bayesian statistical analysis to generate date ranges for our “lumper” categories of projectile point types from the Tennessee River. Partially inspired by Chris Moore’s work on the Savannah River Valley, it occurred to me that I could distill this information down to a cheat sheet for others to use while we’re still working on tweaking our analysis for publication.Cheat Sheet

Here’s a link to the poster where you can find some more technical detail on how we conducted the analysis.

And here is the cheat sheet that I prepared based on this analysis with pictures of points from archaeological sites from Kentucky Lake from the McClung Museum at the University of Tennessee and the Tennessee Division of Archaeology in Nashville. The two Cumberland projectile point images were provided by Jesse Tune and come Alabama and Kentucky.


**Link to a high-resolution version of the cheat sheet chart**

Archaeology at Rock Creek Mortar Shelter (40Pt209), Pickett State Park and Forest

30 Days of Tennessee Archaeology, Day 11

Jay Franklin and Lauren Woelkers
East Tennessee State University

Jay Fig1

Figure 1. Rock Creek Mortar Shelter Summer 2017.

This past summer we continued archaeological excavations at Rock Creek Mortar Shelter (40Pt209) as part of an ETSU archaeological field school. New for this year was operating from the Pickett State Park Archaeology Museum and ETSU Archaeological Research Station. The museum opened to the public late this spring. Travis Bow, Park Ranger and Interpretive Specialist at Pickett, will have more to say about the museum in a separate blog later this month. However, public archaeology was an integral facet for the students this year, and among their responsibilities were leading hikes to the rock shelter and interacting with volunteers on site. It was a positive experience for the students and the public.

What helped make the public archaeology so positive is the fortuitous location of Rock Creek Mortar Shelter so close to the museum (a 15 minute walk). I have blogged about our work at the shelter previously (2014 and 2015), so I will simply summarize the background. Rock Creek Mortar Shelter contains intact deposits ranging from at least the end of the Pleistocene to about 1000 years ago toward the end of the Woodland Period in this region (Figure 1). We do have one AMS radiocarbon date that is Younger Dryas in age but no definitive Paleoindian artifacts yet. For now it appears that the first foragers to habitually occupy the Upper Cumberland Plateau (UCP) did so at the onset of the Holocene during ameliorating climates. While we excavated in Woodland and Late Archaic layers this summer, too, I highlight the early Holocene finds.

Jay Fig2
Figure 2. Blade and proximal blade fragment.

We continue to find blades in the early Holocene levels (Figure 2). This is significant because archaeologists disagree on whether true prismatic blades were purposefully made during the Early Archaic. We have recovered perhaps two dozen at Rock Creek but only a couple of blade core fragments. Our lithic flake debris analyses are ongoing, but we believe most of the cores were repurposed for the manufacture of bifaces. We have recovered numerous Early Archaic bifaces from the site along with significant numbers of biface reduction and thinning flakes in the early Holocene levels (Figure 3). Our current analyses are focused on determining if we can identify tablet flakes in the assemblage as early stage indicators of preparing cores for the removal of prismatic blades. Stay tuned. . .


Jay Fig3

Figure 3. Early Archaic tools (from left to right: Pine Tree Corner Notched biface, St. Albans biface, unifacial endscraper/graver.

One of the most interesting finds of the summer was a flaked hematite biface from the early Holocene levels (Figure 4). We recovered a flaked hematite adze in 2015 (Figure 5). What makes the hematite biface find so interesting is that it makes two flaked hematite artifacts recovered from the early Holocene levels and we identified the source of the material this year: the back wall of the shelter (Figure 6).

Jay Fig4
Figure 4. Flaked hematite biface recovered in July 2017.

While it is not uncommon to find pieces of hematite in the sandstone bedrock (and iron-rich sandstone conglomerates), it may be uncommon to find blocks that have silicified well enough to knap. We have also recovered pieces of the hematite flaking debris in the early levels. These silicified deposits are not visible in the shelter walls today because they are buried but were almost certainly 10-11,000+ years ago when the shelter floor was 2 meters lower than today. We think these pioneering foragers opportunistically knocked these blocks from the wall and fashioned them into stone tools. The flaked hematite artifacts may have been used for wood working. Maureen Hays at the College of Charleston will examine these artifacts for microscopic use wear later this year.

Jay Fig5
Figure 5. Flaked hematite adze recovered in January 2015.

In sum, Rock Creek Mortar Shelter is yielding evidence of and information about the region’s earliest inhabitants familiarizing themselves with a new environment and raw material availability and distribution on the landscape. In our ongoing lithic analyses, we also hypothesize that these first foragers separated from late Pleistocene populations while prismatic blade tool manufacture was somewhat common in the southeast. This may, in part, help to explain why early Holocene peoples on the UCP were still making and using prismatic blade technology. We can also see that it did not last long. Once these early folks realized the abundance of high quality chert resources in the region (there are at least a dozen raw material sources within a day’s walking distance of the shelter), they moved away from prismatic blade tool technology and simply began making blade-like blanks. Calculated preparation and curation of prismatic blade cores thus became unnecessary, and lithic technology became more expedient.

We will continue excavations at Rock Creek Mortar Shelter in 2018. Next summer we will target deposits underneath the lowest ledge in the shelter. These deposits may extend back under the ledge as much as 4 meters.


Jay Fig6

Figure 6. Silicified conglomeritic hematite block from back wall of shelter.


Archaeology and Environmental Change at the End of the Ice Age

30 Days of Tennessee Archaeology 2016, Day 11

Jesse W. Tune
Department of Anthropology, Fort Lewis College

Environmental change –this term appears everywhere these days. It shows up in news media, social media, political debates, bill boards, TV and radio commercials, and even in archaeology. Terminology matters, and particularly for archaeology it’s important to talk about “environmental change” rather than “climate change.” This reflects the fact that it archaeologists are not just interested in past climates, but rather in understanding entire past environments (climate, plants, animals, rivers, landscapes, etc.), and how changes in the environment impacted the lives of past peoples. To address these questions we study environmental shifts over very long periods of time – centuries or even millennia. Much of my own research deals with environmental changes and subsequent human responses at the end of the last Ice Age (or Pleistocene).

The end of the Pleistocene was arguably the most dramatic period of environmental change that humans have ever experienced. Environmental records like ice and sediment cores reveal that this was a very chaotic time, as much of the Northern Hemisphere underwent the death throes of the last Ice Age. The climate began to warm up sometime around 18,000-16,000 years ago. Then about 12,900 years ago the world experienced the Younger Dryas (YD), an extreme and abrupt cold period that lasted some 1,300 years. The extent to which the YD affected humans is unclear. Undoubtedly, changes to human behaviors were directly related to the local severity of the YD. To understand how this chaotic period during the YD effected people living in Tennessee, we need to look at what people were doing before, during, and after the YD began.

One way to understand at how people lived in the past is to look at how tools (projectile points in this case) were being used. While there are a number of things that can complicate this type of analysis, comparing the length and width of points gives us a general idea of what was going on. When this length-to-width correlation is analyzed for points used before, during, and after the YD, a surprising pattern emerges.

9.03 - F1

Comparison of length-to-width for Clovis, Cumberland, and Dalton points.

The length-to-width of Clovis points (used before the YD) and Cumberland points (used during the beginning of the YD) is very similar. This suggests that they were used and resharpened in very similar ways. However, the length-to-width of Dalton points (used at the end of the YD) is significantly different, and suggests that they were used and resharpened in very different ways.

 We can also learn about changes in behavior by examining how and when points were discarded. That is, did people discard them when they broke, or did they continually resharpen points until there was basically nothing left? Again, Clovis and Cumberland are nearly identical, while Dalton points are very different. Broken and resharpened Clovis and Cumberland points were discarded in about the same frequencies. Dalton points were discarded after being resharpened nearly eight times more frequently than when they were broken. This tells us that people using Clovis and Cumberland points made new points when they broke, while people using Dalton points held on to them and continued to resharpen them until there was basically nothing left.

9.13 - F4

Frequencies of discard broke versus resharpened Clovis, Cumberland, and Dalton points.

Another way to understand how environmental change may have influenced changes in human behaviors is to look at what resources were used and where those resources were coming from. In this case, we can study the actual stone used to make projectile points. Much of Tennessee and the Midsouth is characterized by an abundance of chert (flint) suitable for making stone tools. The main chert types in Tennessee are Fort Payne and St. Louis, and their distributions essentially create an oval surrounding Nashville and Murfreesboro.

9.13 - F5

Distribution of Fort Payne and St. Louis cherts in Tennessee.

Prior to the YD, Clovis points were being made in exactly the same frequencies from both chert types. During the beginning of the YD, two-thirds of Cumberland points were being made from Fort Payne chert, while only approximately one-third were being made from St. Louis. By the end of the YD, almost three-fourths of Dalton points were made from St. Louis, and only about one-fourth were made from Fort Payne.


9.13 - F6

Comparison of chert types used for Clovis, Cumberland, and Dalton points.

At first glance it looks like this is a simple pattern to interpret. People making Clovis points had no preference for chert types, while Cumberland point makers preferred Fort Payne and Dalton makers preferred St. Louis. However, it may be more complicated (and interesting) than that…

If we consider where these two chert types are predominately found, then we can start to understand how people may have been using the landscape around them. Clovis and Cumberland points are found in similar distributions throughout the state, yet the types of chert used to make them are different. This may indicate that people making Clovis point were relatively unfamiliar with where certain types of chert occur, and were content using either raw material type. Later, people making Cumberland points would have been slightly more familiar with were certain chert types occur, and preferred to make their points from St. Louis material. Something different, however, appears to have been going on by Dalton times at the end of the YD. The distribution of Dalton points is much more limited that Clovis or Cumberland. In fact, in Tennessee, Dalton points are most frequently found in the same area where St. Louis chert is most prevalent. So rather than preferring to make Dalton points from St. Louis chert, people were simply using local stone resources to make points.

9.13 - F7

Distributions of Clovis, Cumberland, and Dalton points in Tennessee.

So what does all of this tell us about the relationship between human behavior and environmental changes at the end of the last Ice Age? Basically, environmental changes at the beginning of the YD do not appear to have led to major changes in human behaviors in Tennessee – this may not be the case in other areas. Rather, it appears that significant changes occurred near the end of the YD when the climate began to warm up. Projectile point technologies trend toward being resharpened more extensively and used longer before being discarded. The territories where points were discarded appear to become smaller from Clovis-to-Cumberland-to-Dalton, while technologies became more focused on locally available resources.

Editor’s note: for a more in-depth discussion of this topic, see the recent journal article article “The Clovis-Cumberland-Dalton Succession: Settling into the Midsouth United States during the Pleistocene-to-Holocene Transition.”

The Paleoindian and Early Archaic Record in Tennessee

30 Days of Tennessee Archaeology 2016, Day 3

Jesse W. Tune
Department of Anthropology, Fort Lewis College

The Tennessee Fluted Point Survey (TFPS) is one of the most comprehensive statewide archaeological surveys in North America. Not only does this survey differentiate between Paleoindian point types, but it also includes Early Archaic point types. To say that Tennessee possesses some of the densest concentrations of Paleoindian and Early Archaic artifacts in North America is somewhat of an understatement. To hopefully get this point across and try to provide a general summary of what it might mean, I’m going to throw a lot of numbers around and share some maps of where artifacts have been found. Then I’ll try to explain why Tennessee appears to be a unique place for Paleoindian and Early Archaic archaeology.

9.3 - F1

Image from: Morse, Dan, Phyllis Morse, and John Waggonee 1964, Fluted Points from Smith County, Tennessee. Tennessee Archaeologist 20:1(17).

First, a little background about Paleoindian and Early Archaic research in Tennessee. While there are exceptional numbers of Paleoindian and Early Archaic artifacts recorded across the state, nearly all have been recovered from plowed fields and eroded shorelines of lakes and streams. Artifacts recovered in these types of settings are typically not in their original location, and any datable materials that were buried with them are now long gone. Consequently, while individual counties in Tennessee having more Clovis points than Arizona and New Mexico combined (where Clovis research began), the majority of Paleoindian and Early Archaic research has been focused on other regions with more datable sites.

However, there is a long history of Paleoindian and Early Archaic research in Tennessee. This research began in earnest in 1945 with Thomas M. N. Lewis’s study of fluted points in Tennessee. In an attempt to locate fluted points in buried contexts, in 1958 Lewis and Madeline Kneberg investigated the Nuckolls site, along the Lower Tennessee River, where they found an extensive surface collection of Paleoindian and Early Archaic artifacts along the shoreline. In 1964 Dan Morse and colleagues published the first map of the statewide fluted point distribution in Tennessee (see above), which included 278 points. By 1983, Alfred Guthe reported that 389 fluted points had been recovered in Tennessee. Like others before him, Guthe made the prediction that this number represented only a small portion of the fluted points that would eventually be documented from Tennessee. Turns out he was right…

9.3 - F2

Distribution of all fluted points recorded in North America. Image from PIDBA (

In 1988 John Broster and colleagues at the Tennessee Division of Archaeology began intensively expanding the TFPS. As of the most recent update (2013) a total of 5,497 Paleoindian and Early Archaic points are recorded in the TFPS, which are made accessible through the Paleoindian Database of the Americas. As David Anderson and colleagues discussed in their post yesterday, compiling large-scale datasets and distributions of artifacts is critical to asking and answering the big picture questions that often fascinate archaeologists.


Frequency of points in Tennessee by time period and region.

Paleoindian and Early Archaic points are concentrated toward the center of Tennessee, with an overwhelming majority of all points documented from the Highland Rim, Central Basin, and Coastal Plain. The vast majority of all points occurs in the Highland Rim (63.1%). The Central Basin (14.5%) and Coastal Plain (13.9%) have roughly equal frequencies of points. So points from these three regions alone make up nearly 95% of all Paleoindian and Early Archaic points in the entire state!

Another way to look at this data is to scale the densities of points to account for the different sizes of each physiographic region. The density of all Paleoindian and Early Archaic points throughout the state is 84 points per 1,000 km2.

So what does this really mean in terms of the early prehistory of Tennessee and archaeological research? Why is it such a big deal that there are so many artifacts found in Tennessee?


Densities of point types per 1,000 square km.

Well, basically there are far more Paleoindian and Early Archaic points in Tennessee than in most other states. Also, the distributions of points throughout the state reveal distinct patterns. Regardless of point type (or time period) there are more points toward the center of the state. While this is a complex issue to unwrap, it undoubtedly relates to the distribution of resources, like stone to make tools and riverine resources that would have attracted animals.

More importantly though, studying large-scale distributions of artifacts helps archaeologists interpret how people were using and interacting with their environments on the macroscale. These type of studies allow us to ask questions about settlement strategies, territorial ranges, migration routes, and trade and interaction networks (just to list a few).

Basically, this all tells us something about changes in how people organized themselves across the landscape, and how that organization system may have changed over time. Paleoindian points in Tennessee occur in a larger area than Early Archaic points do. Again, while this is a complex issue to understand, there appears to be a reduction in territories over time. Paleoindian groups were likely more using larger territories than subsequent Early Archaic groups.

A Tale of Two Archaic Sites: Identifying Correlations Between Intensity of Tool Use and Land-Use Patterns

30 Days of Tennessee Archaeology 2015, Day 24

Annie Melton
Texas A&M University

Archaeological sites litter the Tennessean landscape…or so I hear, not actually having stepped foot in the state! Not having stood on Tennessean soil, though, has not prevented me from experiencing the rich archaeological record or from taking part in research projects endeavoring to answer questions regarding prehistoric populations. For the past two years, Jesse Tune and I have analyzed hafted bifaces from the Fernvale (Williamson County) and Magnolia Valley (Rutherford County) sites, attempting to understand any possible correlations between the environments where these sites are situated and how intensely the tools were used.

Bifaces from the Fernvale site.

Bifaces from the Fernvale site.

Both the Magnolia Valley and Fernvale sites are within the Harpeth River watershed of Middle Tennessee, have extensive Late Archaic occupations, and have produced sizable samples of hafted bifaces from that period (53 for Fernvale and 33 for Magnolia Valley). Magnolia Valley is located in an upland setting near the headwaters of the Harpeth River, while Fernvale is situated on an alluvial terrace along the South Harpeth River. The seasonal occupations of occupation for these two sites during the Archaic was most likely was impacted by their differing upland/lowland settings. While Magnolia Valley was most likely occupied during the fall and winter, Fernvale would probably have been occupied during the spring and early summer seasons.

Raw material availability also serves as a distinguishing factor between these two sites, and is potentially related to the intensity of tool usage as well. While relatively decent-quality nodules of Fort Payne chert were easily accessible along gravel bars in the South Harpeth River near Fernvale, the area surrounding Magnolia Valley did not have this fortune, and exhibits only poor quality, fossiliferous Fort Payne Chert found along a nearby ridgeline.

Excavations at Magnolia Valley, summer 2014. Photo c/o the Rutherford County Archaeological Project.

Excavations at Magnolia Valley, summer 2014. Photo c/o the Rutherford County Archaeological Research Program.

Therefore, we had two sites, occupied relatively contemporaneously but in distinctly different local environments. Our question was: would these differences also be reflected in how the hafted bifaces from each site were resharpened and reused over time? In order to understand how intensely points were used, we used Andrefsky’s Hafted Biface Retouch Index (HRI), which examines the frequency of resharpening flakes removed along the lateral edges of a biface’s blade. Higher amounts of retouch are represented by higher HRI values (maximum value of 1.0), while low HRI values represent less retouch (minimum value of 0.0). Bifaces which have a significant amount of retouch and high HRI index are also expected to be much less standardized in terms of length, width, and thickness than is seen in an assemblage with lower HRI values. This is because through extensive retouch a tool’s width, length, and thickness will be modified, thereby producing a high amount of variability in an assemblage with high HRI values.

Overall, we analyzed 33 bifaces from Magnolia Valley, whose HRI value averaged 0.83. We found that Magnolia Valley’s bifaces have a relatively high rate of retouch.  For Fernvale, we analyzed 53 bifaces and found that the average HRI value was 0.78, just slightly less than Magnolia Valley’s. We also noticed that more than half (51%) of the Fernvale bifaces had an HRI value of 0.8 or less.

Pic_9_Fernvale_MV_HRI_ComparisonNow to compare! With the results for both sites, interesting differences began to emerge. Using Mean Values, Magnolia Valley’s assemblage has slightly higher HRI than Fernvale. This met our expectations that, given the lack of decent-quality raw materials in the vicinity of the Magnolia Valley, the bifaces would be used more intensively. However, both sites exhibit relatively high HRIs, suggesting intensive resharpening and reuse occurred at both locations.

We expected to see more standardization in the length, width, and thickness of bifaces in the Fernvale assemblage, given its lower HRI value (the assumption being that more intensive use results in greater variation and vice versa). However, we noticed that the opposite was true: Late Archaic bifaces from Fernvale exhibited less standardization than those from Magnolia Valley.

Overall, we are looking at two sites where stone tools were intensively used and resharpened as opposed to site occupants manufacturing new tools. This behavior may reflect raw material availability: In areas with limited resources such as Magnolia Valley, people are expected to exhaust more relative utility than in areas with abundant resources. However, decent-quality lithic raw material is relatively abundant in the South Harpeth River Valley. Another possible explanation for the high HRI values from Fernvale may lie in the inter-group relationships of the Late Archaic. In her analysis of the skeletal remains from Fernvale, Dr. Shannon Hodge identified evidence of interpersonal conflict, including signs of warfare and trophy taking. While lithic resources near Fernvale are regionally abundant, increased population pressures and warfare may have influenced the degree to which stone tools were being used. These conflicts may have imposed increasing constricted territorial boundaries, limiting people’s access to specific raw material source locations. Therefore, site inhabitants engaged in more intensive biface reduction in order to increased the use-life of their tools.

Panorama of excavations at the Fernvale site in 1985. Photo c/o the Tennessee Division of Archaeology.

Panorama of excavations at the Fernvale site in 1985. Photo c/o the Tennessee Division of Archaeology.

At the same time, Fernvale’s easy access to a lithic material may explain the greater variability seen within the assemblage. Occupants of Magnolia Valley would have needed to travel further in order to acquire necessary raw materials, resulting in more standardized assemblages. For the less standardized Fernvale assemblage, this may be a result of the high amount raw material variability found near the site.

*Editor’s Note: The 2013 report on excavations at the Fernvale site was published as part of the Tennessee Division of Archaeology’s Research Series, and is available for free download via the TDOA web page. You can read more about work at Magnolia Valley in the 2014 blog posts by Jesse Tune and Tim de Smet.