A Preliminary Analysis of Radiocarbon from the East Texas Archaic Period (Selden)

This post presents a few preliminary findings of a temporal analysis of the East Texas Archaic based solely upon the examination of radiocarbon dates from sites that have deposits that date to the period. Text and figures are reprinted here with the permission of the Journal of Northeast Texas Archaeology, where these findings were published in 2013 (for a more exhaustive overview, see JNTA 42). All assays employed in this effort were collected from research and cultural resource management reports and publications, synthesized, then recalibrated in version 4.1.7 of OxCal  using IntCal09.

The date combination process is used herein to refine site-specific summed probability distributions, illustrating—for the first time—the temporal position of dated archaeological sites with radiocarbon assays. In all, 73 radiocarbon dates from 34 sites serve as the foundation for this analysis of the East Texas Archaic period (ca. 8000-500 B.C.). All dates used in this analysis come directly from the East Texas Radiocarbon Database (ETRD). Within the sample, there are 19 sites with a single radiocarbon sample, eight sites with two dated samples, one site with three dated samples, three sites with four dated samples, one site with five dated samples, and one site with 14 dated samples. Of the 73 radiocarbon dates from the ETRD used in this analysis, one dates to the Early Archaic period (ca. 8000-5000 B.C.), eight date to the Middle Archaic period (ca. 5000-3000 B.C.), and the remaining 64 date to the Late Archaic period (ca. 3000-500 B.C.).

Date Combination

Archaic sites with combined radiocarbon dates include: Shell Lens (41FN130), Winston (41HE245), Finley Fan (41HP159), J. Simms (41NA290), Herman Ballew (41RK222), Mockingbird (41TT550), and 41UR77. The number of dates garnered through research at each of these sites is biased by variable research designs, mitigation strategies, and access to funding.

ETX Archaic Figure10

 Example – All and combined summed probability distributions for Archaic period dates from 41UR77 with 1σ and 2σ ranges, median ages, and number of samples.
(Figure appears courtesy of the 
Journal of Northeast Texas Archaeology)


In every case where date combination was applied, the new combined age replaced the assays used to calculate it. Upon completion of the date combination process, the summed probability distributions for all East Texas sites with Archaic-era radiocarbon assays were plotted chronologically. This allows us—for the first time—to view all of the Archaic-era assays at the regional scale.

In the future, it would be useful to apply some manner of chronometric hygiene to the Archaic sample of dates, whether following a conventional method or by vetting each date to ensure that each assay represents an Archaic component associated with some manner of human occupation (i.e., artifact manufacture or feature use). At this point it is unknown how many of these dates can actually be attributed to the Archaic occupation of the East Texas landscape, but this preliminary analysis does illustrate a fairly remarkable increase in the number of dates during the Late Archaic (ca. 3000-500 B.C.) period following a sparse dated record for the Early and Middle Archaic. The fact that the number of assays from each period increase through time is a familiar trend, and one that is often attributed to an increase in population size (although there are some substantive challenges to this approach).

ETX Archaic Figure11

East Texas sites with Archaic-era radiocarbon assays in chronological order.
(Figure appears courtesy of the Journal of Northeast Texas Archaeology)

Concluding Remarks

Although biases likely exist in the radiocarbon sample from sites in the region, it is evident that the most extensive Archaic occupation (concentration of dates) of East Texas occurred during the Late Archaic period. Certainly more dates are needed from Early and Middle Archaic horizons, but given the often ill-formed stratigraphy in archaeological deposits that occur throughout East Texas, finding suitable samples can be a challenge. Also, some measure of chronometric hygiene needs to be applied to this sample of dates to increase their resolution and temporal accuracy, and should the focus be upon population-based studies, a taphonomic correction may also be warranted. While large steps have been taken to explore East Texas archaeology, the Archaic period remains very ill-defined with respect to our understanding of the material culture as well as the chronology. The fact that only 73 dates from the East Texas Radiocarbon Database—which is currently composed of 1248 radiocarbon dates from East Texas—speaks to the need for further research.

About the Author

Dr. Robert Z. Selden Jris a Research Associate in the Center for Regional Heritage Research at Stephen F. Austin State University.


SAA Meets in Texas April 2014 (Dore)

The Society for American Archaeology (SAA) meets in Texas during April 2014! SAA’s new digital journal, Advances in Archaeological Practice, welcomes manuscripts from Texas archaeologists for its upcoming issues.

This exciting new peer-reviewed journal is the premier place to publish scholarly work on how we learn about the past, convey our findings in the present, and manage resources for the future. Advances authors take advantage of the journal’s extremely quick publication time, full-color magazine format, global digital distribution, and the ability to use the format to include digital content.

The practice of archaeology includes many topics, and all are welcome in the journal. These may include, but aren’t limited to, innovations in approach, technique, method, technology, business models, collaboration, compliance, process, ethics, public engagement, stewardship, and training,

To obtain additional information and submit a manuscript, please see the Advances website. Specific questions may be sent to the journal’s editor. Come on Texas archaeologists, show your stuff!

About the Author

Dr. Christopher D. Dore is the Editor of Advances in Archaeological Practice, and is affiliated with the School of Anthropology at the University of Arizona.


Concerning the US Forest Service Collections in East Texas (Selden and Perttula)

The US Forest Service (Sabine NF, Angelina NF and Davy Crockett NF, specifically) collections from East Texas have long been in the top five collections that Tim and I have wanted to delve into the most. Lucky for us, the USFS–Juanita Garcia, USFS Archaeologist, in particular–has been very receptive to our requests to study the artifact from this ill-understood region in the Southern Caddo Area. Following the most recent Caddo Conference in Tyler, we began the task of analysis. As of today, that analysis has come to an end. What is left; however, is a large-scale/significant effort to round up all of the reports and field records to explore how the occupations at these sites might relate to sites that lie beyond the boundaries of USFS property, and what the larger regional patterns (now inclusive of the USFS data) might tell us about the use of these lands by our prehistoric predecessors.

USFS - Texas

Location of USFS properties mentioned in the text – from http://www.fs.usda.gov/main/texas/about-forest/about-area.

While the results of our analyses will be presented in a series of forthcoming reports and articles, we thought that a brief–and very preliminary–overview of our findings might be of interest to the broader archaeological community.

Figure 0

Selected decorated and plain (rim) sandy paste sherds from the USFS collections.

In short, lands owned by the USFS in East Texas harbor an extensive archaeological record of sites occupied by Native Americans as early as ca. 10,000 B.C.  to as late as the 1680s (or later). Our assessment of the artifact assemblages recovered from work since the mid-1980s at these prehistoric sites points to a very dynamic use of the prehistoric landscape.

Figure 1

Selected projectile points from the USFS collections.

In all, there are more than 150 sites with temporally diagnostic artifacts (projectile points and ceramic sherds) on Forest lands, and a number of the sites have evidence of multiple temporal components. Approximately 3.8% of the components yielded evidence of Paleoindian habitation demonstrated by the presence of Dalton and San Patrice points. Archaic sites (ca. 10,000 – 2500 B.P.) are more common with 10.8% of the components having a variety of stemmed and notched Archaic dart points and other chipped stone artifacts. Woodland period (ca. 2500 – 1150 B.P.) sites are widely dispersed across Forest lands, as almost 56% of the components on the three National Forests appear to date to this period. These sites also have sandy past pottery (Goose Creek Plain, var. unspecified) as well as Kent, Gary, and Godley dart points and/or early Friley (ca. A.D. 700) and Steiner arrow points. Three Woodland period sites on the Sabine National Forest also have Marksville Incised or Marksville Stamped vessel sherds.

Figure 2

 Selected decorated sherds from the USFS collections.

The ancestral Caddo sites on East Texas National Forest lands comprise approximately 30% of the known components. Those sites are marked by later arrow point styles (Alba, Bassett, Bonham and Perdiz types) as well as plain and decorated grog or bone-tempered ceramic vessel sherds. About 15% of the sites have Caddo arrow points and almost 38% of the USFS sites contain Caddo vessel sherds. These artifact classes indicate that Caddo peoples continuously used these parts of East Texas from as early as ca. A.D. 800 to at least the 17th century. One site on the Sabine National Forest (41SY280) also has a very thin piece of cut and shaped metal that may be part of a European trade good.

Figure 3

Examples of lithic raw material from the USFS collections.

Where do we go from here?

Data from this analysis will be used to complement the larger regional datasets that we have been working with of late. A number of the assemblages analyzed will be added to the ever-growing sherd database. Additionally, we pulled 221 sherds from 62 sites for INAA/petrography, and numerous flora (radiocarbon and plant identifications) and fauna (paleoenvironmental reconstruction and potential food resources) samples that can be added to the existing INAA,  petrography, and radiocarbon databases for the region.

More importantly, we plan to use these data to make some substantive recommendations to the USFS regarding the further evaluation of these important cultural resources, and help them to prioritize–and address–their current condition, archaeological research potential, and the management of archaeological resources on USFS lands in East Texas.

About the authors

Dr. Robert Z. Selden Jr. is a Research Associate in the Center for Regional Heritage Research at Stephen F. Austin State University

Dr. Timothy K. Perttula is a Research Affiliate in the Center for Regional Heritage Research at Stephen F. Austin State University, and Manager at Archeological & Environmental Consultants, LLC


A special thanks to the Chair of the SFASU School of Art, Chris Talbot, for providing the space and equipment needed to photograph the USFS artifacts.



Regarding Caddo INAA (Selden, Perttula, Eckert, and Carlson)

Instrumental Neutron Activation Analysis (INAA) is a technique pioneered by the archaeological chemistry community and represents one of a very few examples where the earth sciences adopted a technique from archaeology. Using small samples (archaeologically-recovered ceramics in this instance), INAA can—with high precision—measure both very low and very high concentrations of a wide range of geochemical elements.

The analysis of Caddo INAA began with 22 sherds analyzed at the National Institute of Standards and Technology (also known as NIST) in 1996. Since then, all analyses have been processed through the University of Missouri Research Reactor (MURR), where Caddo archaeologists have amassed the third-largest sample in their database (only surpassed by the Mimbres region of the American Southwest, and the Valley of Mexico).

Background – The Formative Years

During the formative years of Caddo INAA (late-1990s), many interesting trends began to emerge; however, the bulk of our efforts remained focused upon defining the extent of trade and exchange with groups located outside of the ancestral Caddo region.


Map illustrating the extent of the Northern and Southern Caddo Areas, and archaeological sites (stars) outside of the ancestral Caddo area where Caddo ceramics have been recovered.

While our gaze remained focused on exotic trade and exchange, the number of local sherds within the database continued to increase.

 BackgroundRenewed focus on intra-regional interaction

While intra-regional interaction had always been a topic of considerable interest, it was a 2007 study that pointed toward upcoming and significant challenges to the interpretation of currently defined statistical groups. In that 2007 study (from East Texas), 100 Caddo sherds from generally contemporaneous sites in the upper Neches River basin were submitted for analysis, but only 16 (16%) were able to be assigned to the existing statistical groups by MURR.


INAA statistical groups as defined in 2010.

“Overwhelmingly Homogenous” — an aside

During the course of reading through a pile of technical reports, I (Selden) received an invitation from some colleagues in our geology department to visit a site in the Angelina River basin. I had recently become interested in this area, and was exploring whether it may be possible to employ a study of ceramic petrofacies in this same stream basin. Upon my arrival, I was met with a substantial surprise. Within a one-acre study area, they had identified seven unique clays. Admittedly, some were much deeper in the profile than others (up to 1m), but all were accessible from a cutbank less than 500 meters away.

After reading the technical literature, there was one trend in particular that was troubling, and it had appeared in almost all of the Caddo INAA literature from MURR for the past six years–that the geochemistry of East Texas (and the larger Caddo region as a whole)–was “overwhelmingly homogenous.” After my recent visit with my colleagues in the Angelina River basin, this left me questioning how this could be. How could the region be characterized as geochemically homogenous when so much variation exists in local clay sources?

Honing In

In early 2012, we began working toward a solution to interpreting these complex data. In doing so, three research questions were developed to guide our efforts:

  1.  Is the ancestral Caddo region geochemically homogenous, and if not, what manner of geochemical variation exists?
  2. If the geochemistry is variable, is it possible to demarcate between the various geologic groups and–potentially–geologic formations?
  3. If geochemical variation can be found to exist between geologic groups, is it possible to then demarcate locally-manufactured ceramic assemblages based upon those same elements that segregate the geology?

Question 1: Geochemical Diversity

Among the first problems with the current interpretations was that of the consistent application of the calcium correction to the entirety of the Caddo INAA dataset. In a review of the submitted samples, we found that only 4% of the sherds in the INAA dataset were shell-tempered, and we made the decision to proceed with applying the calcium correction only to those sherds identified as shell or bone-tempered.


Number and concentration of calcium (Ca) in shell, bone and other (grog)-tempered sherds in the Caddo INAA database.

It should be noted that a number of outliers exist in the Other category that possess higher measures of calcium, and likely represent grog-tempered sherds in which the grog used by Caddo potters may have been grog-shell or grog-bone tempered sherds. This is an issue that can be further clarified with the incorporation of ceramic petrographic analyses.

A Solution Emerges

It took time, but a solution to our current conundrum eventually presented itself. In its current form, the dataset is put through the standard suite of statistics with R (r-project.org), then imported into GIS for further analysis.

In GIS, a z-score was calculated for each element, yielding a site-specific value (average). This provided us with an X and a Y coordinate for each site, and 33 Z coordinates (one for each element).

Once the site-specific z-scores were rendered, it became a question of how best to display these data. Interpolation proved to be the best option, but it was a struggle to find a representation that would best illustrate what was becoming evident in the numbers.


Geographic distribution of the 33 elements produced from Caddo INAA data (ceramics). White represents low values, and red represents high values.

These results illustrate the high degree of geochemical variability that occurs across the ancestral Caddo region. However, it still left us wondering whether it was possible–using the criterion of abundance–to link our findings with variations found in local geologic groups and formations.

Question 2: Geology

Early last fall, we set out to test whether it was possible to demarcate between geologic groups (and possibly formations) using the geochemical maps. By overlaying the digital Geologic Atlas of Texas, we concluded that it is possible to segregate between geologic groups using the Caddo INAA data. In all, we found 16+ areas where geologic groups or geologic formations can be said to vary substantially enough (based upon our sample) to warrant further exploration.

Blog Fig 2

Geographic distributions of Ce, Dy, Eu, Hf, La, Lu, Nd, Sm, Ta, Tb, Yb, and Zn within 25 miles of the Kitchen Branch site, found to successfully discriminate between the Claiborne and Wilcox Groups.

The division of these data based upon measures of geologic variability makes it possible to demarcate on a broad scale between sites and ceramics from Caddo vessels manufactured from Claiborne Group clays, and those manufactured from clays in the Wilcox Group. Further separation can be achieved with elements associated with the Weches and Sparta Sand Formations in the Claiborne GroupSimilarly, the undivided Wilcox Group can be contrasted with geochemical data from sites atop the Carizzo Sand Formation.

Blog Fig 3

Results of cluster analysis using geochemical elements that appear to correlate with local surficial geology: a) INAA clusters representing the Wilcox (blue) and the Claiborne (green) Groups; b) subdivisions of the Claiborne cluster illustrating a geographic increase in Ce that may correlate with the inclusion of tempering agents from the Weches and/or Sparta Sand formations; and c) subdivisions of the Wilcox cluster illustrating a geographic increase in Dy that may correlate with a gradual decrease in tempering agents from the Carrizo Sand through space. 

The results of this study illustrate that the chemical composition of ceramics associated with ancestral Caddo populations is apparently diverse and highly variable, hinting of further successes in ceramic provenance identifications for more robust samples of sherds from sites in this region.

Blog Fig 4

Distribution of lanthanum atop geologic groups, noting contributing sites. 

Question 3: Cultural Implications

Once we realized that the geologic groups and formations were capable of being segregated, we decided to take a run at a type-specific query. Within the database is a sizeable sample of Ripley Engraved sherds, and we wanted to see if it was possible to use those data to ask a more pointed question.


Ripley Engraved vessel (O NAGPRA 2012.1.510) from the Tuck Carpenter site – for more information about this vessel, click here

Within the Big Cypress Creek drainage, five possible political communities have been identified based upon key sites, burial offerings, and artifact assemblages. We wanted to know if we could segregate the ceramics using the geochemistry of the principal fine ware (Ripley Engraved).


Location of Titus-phase Caddo political communities in the Big Cypress Creek drainage.

Using arsenic (As), iron (Fe) and vanadium (V), we were able to successfully demarcate between three of the possible political communities using elements identified with the geochemical maps. Results tentatively point to the local manufacture of ceramics by Titus phase political communities, and illustrates potential movement of ceramic vessels between these communities.

Figure 5

Three Titus phase political communities (PC) can be segregated based upon evidence gathered from the geochemical maps. In the 3D scatterplot, PC 2 is represented by Group 1, PC3 by Group 2, and PC4 by Group 3 – refer to map above for PC locations.

While these results were very exciting, we also found several other spatial trends in the geochemical data worth noting. In the case of chromium (Cr), scandium (Sc) and vanadium (V), the increase in geochemical values from northwest to southeast is apparent in the Big Cypress, Little Cypress, and Sabine River drainages. In a contrasting pattern, hafnium (Hf), rubidium (Rb), and zirconium (Zr) values share similar geospatial patterns, although the pattern associated with Rb represents the inverse of Hf and Zr. This latter pattern may hold the key to discriminating between clay compositions in ceramics from these three drainages as more samples from a greater number of sites become available.


Geochemical distributions of Cr, Sc, V, Hf, Rb and Zr. Dotted line represents the Titus phase boundary. Red represents high values, and white represents low values.

This represents the first ceramic type-specific discussion of INAA results from the ancestral Caddo region, and highlights the successful application of GIS to the analysis and interpretation of the Ripley Engraved dataset. While much remains to be learned with regard to the chemical constituents of ceramic pastes in the Caddo region, this analysis marks a substantial step toward furthering our understanding of the Caddo INAA dataset.

A New Approach to the Analysis of INAA Data

We are currently developing a new approach to the analysis of INAA data using an SQL query that will designate a spatial location for sherd-specific assignments based upon the geochemical maps. The current iteration allows for the assignment of sherds to a specific area.

This new approach gives us a way to ask the same question of all 33 geochemical maps – where is it likely that a vessel was manufactured? Our colleagues have been very gracious in allowing access to their data, and we have amassed a sizable dataset from Central Texas that we will use to test this method; it will likely take another year of tests before the final version is ready for use.

Final Thoughts

The geochemical maps discussed herein can be used to identify potential areas that warrant further exploration (and more samples). That knowledge may then produce the clues necessary for us to begin furthering our hypotheses concerning the cultural transmission of Caddo ceramic technological practices, and the exchange of material goods. Progress on this front should help us to better understand Caddo economies, craft specialization, ideology, and possibly even Caddo origins. While the method of analysis continues to evolve, these preliminary results point to an area of Caddo research where we can begin to make significant and meaningful analytical gains.

About the Authors

Dr. Robert Z. Selden Jr. is a Research Associate at the Center for Regional Heritage Research at Stephen F. Austin State University.

Dr. Timothy K. Perttula is a Research Affiliate at the Center for Regional Heritage Research at Stephen F. Austin State University, and Manager at Archeological & Environmental Consultants, LLC.

Dr. Suzanne L. Eckert is an Associate Professor in the Department of Anthropology at Texas A&M University.

Dr. David L. Carlson is an Associate Professor in the Department of Anthropology at Texas A&M University.


A New View of our 3D Scans! (Selden)

Many of you know that we have been working to refine the way that our 3D scans appear in our repository (CRHR:ARCHAEOLOGY). While we were unable to identify a method for including the imagery in the current (CONTENTdm) platform, we have augmented it with a parallel (ScholarWorks) platform. This allows us to share a low-resolution image that is viewable by those using a slow internet connection, and offers a more expedient method of viewing/manipulating the scans.

In this new platform (below), users can be connected to the metadata through links in the Abstract, as well as a button at the top right (in red below). The stable URL for each vessel–in this case FIN-S18–is listed under the Recommended Citation. You can share this page with friends on Facebook, Twitter, Google Plus, Email, or click on the plus sign to see a variety of others! And don’t forget to take a look at, rotate, and otherwise manipulate the 3D imagery! To see the 3D model in ScholarWorks, click here.


This is the first of many more 3D scans to come–we will be adding the remainder of our 3D collection to the ScholarWorks platform over the coming weeks–and we hope that you take time to enjoy the scans. That said, please venture over to ScholarWorks, and dig in – click on the underlined text below to take a look.


O NAGPRA 2012.1.510

Caddo vessel from the Turner Collection (O NAGPRA 2012.1.510)
Scan appears courtesy of the Anthropology and Archaeology Laboratory
Animation courtesy of the Virtual Curation Laboratory

About the Author

Dr. Robert Z. Selden Jr. is a Research Associate in the Center for Regional Heritage Research at Stephen F. Austin State University (seldenjrz@sfasu.edu).


An Unusual Case of Good Charcoal Preservation in Central Texas (Bush and Kibler)


Pecan (Carya illoinensis) hulls still clinging to the branches in late winter. The nuts dropped in fall and early winter.

Archeologists in central Texas have long been frustrated by poor plant preservation on open-air sites. Even wood charcoal is often sparse due to the alkaline environments on much of the Edwards Plateau (Braadbaart et al. 2009). Conditions in the more acidic Llano Uplift can be slightly better, but wet/dry and freeze/thaw cycles still take their toll on charcoal.

The Jayroe Site (41HM51) provides an example of unusually good charcoal preservation in Central Texas. While located in the Western Crosstimbers region, the Jayroe Site’s riparian locale is similar to alluvial environments found across the more limestone-dominated landscapes of Central Texas, where alluvial deposits and landforms primarily consist of dark loamy to clayey calcareous sediments. The Jayroe artifact and feature assemblages rest on top of a buried cumulic soil formed on such sediments that comprise the first terrace of the Leon River.

The buried soil is similar to the Leon River paleosol identified downstream at Fort Hood (see Mehalchick et al. 1999) in its pedogenic character and geomorphic context. The Leon River paleosol, yielding radiocarbon ages on humates of 1160 ± 40 BP and 1010 ± 70 BP, contains discrete Late Archaic and Austin phase assemblages (Mehalchick et al. 1999:215, 268; Nordt 1992). The stratigraphic position of the Jayroe Site cultural materials on the surface of this buried soil are an indication, along with accompanying radiocarbon ages and temporally diagnostic artifacts (e.g., Perdiz arrow points and Late Caddo ceramics), of the site’s post-Austin phase or Late Prehistoric II age.


A basin-shaped hearth (Feature 9) atop the paleosol at the Jayroe Site.

What makes the Jayroe Site so noteworthy though is its high degree of contextual integrity and preservation of organic remains (e.g., vertebrate faunal and charred botanical materials). This remarkable degree of integrity and preservation is due in large part to the fact that the site’s surface, the surface of the paleosol, was buried by 60 to 200 cm of interbedded sand and mud deposits that are slightly altered pedogenically. The deposition of this alluvial mantle contributed to the high degree of preservation in three ways.

First, the apparent rapid and deep burial of the occupation surface at the Jayroe Site preserved the spatial integrity of the site’s features and artifacts, largely sealing them off from various biological agents that tend to churned and mix sediments, Secondly, the sand component of the alluvial drape, which consists of fine quartz sands derived from outcrops of lower Cretaceous sands in the surrounding landscape, provide a less alkaline environment that facilitated in the preservation of organic materials, particularly charcoal. And third, the encapsulating sandy deposits are less susceptible to shrinking and swelling under alternating dry and wet conditions, conditions that would have accelerated the mechanical break down of charred botanical remains into smaller, largely unidentifiable particles.  


Concentration of bison bones and ash lens (upper left) atop the paleosol and below the interbedded sands and muds at the Jayroe Site.

Even the better preservation conditions at Jayroe did not result in the preservation of uncarbonized plant parts. A few uncarbonized seeds were recovered, but these are consisted with modern seed rain expected on the site and are interpreted as such (except possibly the hackberry, whose possible preservation from ancient times on this and other sites deserves a blog post all its own). In terms of botanical preservation, the main difference between Jayroe and less quickly-buried sites lies in the recovery of seeds and pecan nutshell.


Hawthorn (Crataegus cf. viridis) seed from the Jayroe site. Scale in mm.

The seeds consisted of a sedge and some hawthorn seeds, probably a component of ancient seed rain burned by accident, and hawthorn seeds. Hawthorn trees were part of the riparian forest understory near Jayroe, and hawthorn wood charcoal was recovered in some of the flotation samples. Hawthorns are a crabapple relative and have similar dry, edible fruits.


Pecan nutshell (Carya illinoinensis) from the Jayroe site. Scale in mm.

Although pecans grow throughout central Texas, pecan nutshell is not terribly common on prehistoric sites. Prior to Jayroe, one of us (Bush) had encountered a total of eleven archaeological pecan shell fragments in the last five years spread over three sites in central Texas (Siren, Herbert Ranch, and Site 41CV947). Pecans are a species of hickory, and they belong to the hickory-walnut botanical family. Not surprisingly, pecans are most similar to hickory nuts in nutritional content (although they have somewhat more fat and less protein), and both are more similar to walnut than to acorn.

Proximate Analysis of Four Edible Tree Nuts and Corn Meal per 100 g Dry Weight (USDA, ARS 2013)







Fat (g)






Protein (g)






Carbohydrate (g)






Water (g)






Energy (kcal)






Grant Hall points out that the nutrients in pecan and other hickories, especially the linolenic fatty acids, would have been particularly important for hunter-gatherers who relied on lean meat for a portion of the year (Hall 2000).

Traditional hickory processing methods used by Iroquois, Choctaws, Cherokees, and many other tribes involve pounding hickory nuts into small pieces and then heating them in water, where the oil can be skimmed off, the nutmeat retrieved from suspension, and the shells allowed to sink to the bottom (Fritz et al. 2001; Moerman 1998). Experiments by archeologists show that this process yields a much larger number of calories per labor invested than does cracking and picking (Talalay et al. 1984:353). Other common nuts cannot be processed in this manner because either their meats float (acorn, hazelnut) or the nuts become bitter (walnut). These methods also work for pecan nuts.

In Texas today, pecans and other hickories are typically available for collection from mid-October through the first week in December (McEachern et al. 1977). In good years, harvest can begin as early as late September. The earliest possible harvest is best because it minimizes loss to predators such as squirrels.

References Cited

About the Authors

Dr. Leslie L. Bush is a CRHR Research Affiliate, and the proprietor of Macrobotanical Analysis.

Mr. Karl Kibler is an Archeologist/Geomorphologist at Prewitt and Associates, Inc.


Four CRHR Research Posters Presented at the Texas Academy of Science (Selden, Perttula and O’Brien)

Over the weekend, four CRHR research posters were presented at the annual meeting of the Texas Academy of Science in Galveston, Texas on the topics of INAA and 3D Morphometrics. The posters can be seen below – click on each one to see the full-screen version. Many thanks to my co-authors on these posters – Timothy K. Perttula and Michael J. O’Brien; these projects would not have been possible without their help and guidance. The posters–along with a few others–will also be making their way to the Caddo Conference later this month.



Slide 1

Slide 1


Ancestral Caddo Ceramics in East Texas (Perttula and Selden)

The most distinctive material culture item of the ancestral Caddo groups that lived in East Texas (Figure 1) from ca. A.D. 900 to the 1830s were the ceramics they made primarily for cooking, storage, and serving needs. The decorative styles and vessel forms of ceramics found on sites in the region hint at the variety, temporal span, and geographic extent of a number of ancestral Caddo groups that lived in this area. The diversity in decoration and shape in Caddo ceramics is considerable, both in the utility ware jars and bowls, as well as in the fine ware bottles, carinated bowls, and compound vessels. Ceramics are quite common in domestic contexts on habitation sites across the region, and also occur as grave goods in mortuary contexts.

 Figure 1

Figure 1. The Southern and Northern Caddo Areas in Arkansas, Louisiana, Oklahoma, and Texas. Figure prepared by Sandra Hannum.

The Caddo made ceramics in a wide variety of vessel shapes, and with an abundance of well-crafted and executed body and rim designs and surface treatments. From the archeological contexts in which Caddo ceramics have been found, as well as through inferences about their manufacture and use, it is evident that ceramics were important to the ancestral Caddo in: the cooking and serving of foods and beverages, for the storage of foodstuffs, as personal possessions, as incense burners, as beautiful works of art and craftsmanship (i.e., some vessels were clearly made to never be used in domestic contexts), and as social identifiers. In the case of the later, certain shared and distinctive stylistic motifs and decorative patterns on ceramic vessels marked closely related communities and constituent groups.

The stylistic analysis of Caddo ceramics from sites in East Texas has focused on the definition of recognizable decorative elements, patterns, and motifs on the rim and/or body of the quite diverse fine wares (i.e., the engraved and red-slipped vessels, including carinated bowls and bottles) and utility wares, usually cooking or storage jars and simple bowls. These decorative distinctions have both temporal and geographical distributions across East Texas, and in some cases, across the broader Caddo area, and the recognition and unraveling of those distributions  has been key to the reconstruction of settlement and regional histories of different Caddo communities as well as their socio-cultural character.

The stylistic distinctions that have been recognized in East Texas Caddo ceramics are based primarily on the pioneering typological research done by Alex D. Krieger, Clarence Webb, Dee Ann Suhm (Story) and Edward B. Jelks in the 1950s and early 1960s. In 1962, Suhm and Jelks presented descriptions of 60 Caddo ceramic types that had been identified in Caddo sites in East Texas and the Caddo archeological area up to that time. According to Suhm and Jelks (2009:3), since 1962:

the Caddoan [sic] types, at least those found in Texas, have changed surprisingly little, more tweaked than substantially altered. Elsewhere in the Caddoan [sic] area, a relatively modest number (considering the amount of pottery usually found at the sites) of new types have been defined, although many varieties of existing types have been introduced and design motifs, even design element categories, have been recognized, especially by archeologists working in Arkansas.

While the ceramic types defined by Suhm et al. and Suhm and Jelks in 1954 and 1962, respectively, are still useful classificatory constructs for Caddo archeological research, a number of new Caddo ceramic types have been recognized in East Texas archeological sites since the mid-1960s—some better defined than others. Most of them are poorly known among archeologists that work on Caddo sites in the region. Some new varieties have also been identified among several of the well-known types defined in the 1950s, including Poynor Engraved, Hume Engraved, Ripley Engraved, and Wilder Engraved; these varieties may have more discrete temporal and geographic boundaries than when first defined.

Many archeologists working in the East Texas Caddo area continue to rely, erroneously, on the estimated ages of types offered by Suhm and Jelks (1962). However, with the advent of relatively extensive radiocarbon dating of Caddo sites in the region—and the seriation of burials in cemeteries of different ages—as well as many new archeological research investigations, much more accurate temporal estimates for the manufacture and use of ancestral Caddo pottery types are apparent. Distinctive sets of ceramic vessels and assemblages of different ages and areas within the region.

Table 1 represents our efforts to partition the known ancestral Caddo ceramic sets in East Texas. The stylistic diversity in the decorated wares on East Texas Caddo sites has led to the recognition of distinctive stylistic motifs and types that have unique spatial and temporal distributions (although these are still being refined) (Figures 2-5), and the distribution of these ceramic sets can be linked with the identification of culturally specific Caddo groups, phases, and vessel assemblages in the East Texas archeological record.

 Table 1. East Texas and mid-Red River Caddo Ceramic Sets.

Early Caddo set, ca. A.D. 900-1300

 Bowles Creek Plain

Canton Incised

Coles Creek Incised

Crenshaw Fluted

Crenshaw Lobed

Crockett Curvilinear Incised

Davis Incised

Dunkin Incised

Duren Neck Banded

Hickory Engraved

Hickory Engraved, var. Chapman

Holly Engraved

Kiam Incised

Pennington Punctated-Incised

Spiro Engraved

Weches Fingernail Impressed

Williams Plain

 Middle Caddo set, ca. A.D. 1100/1200-1300/1400, upper Red River, cf. Sanders phase, and in parts of East Texas

 Canton Incised

Leaning Rock Engraved

Maxey Noded Redware

Monkstown Fingernail Impressed

Paris Plain

Sanders Engraved

Sanders Plain

Spoonbill Engraved

Broaddus Brushed

Nacogdoches Engraved

Pineland Punctated-Incised

Reavely Brushed-Incised

Tyson Engraved

Washington Square Paneled

 Middle Caddo set, lower Red River, ca. A.D. 1200-1400

 Dunkin Incised

East Incised

Friendship Engraved

Haley Complicated Incised

Haley Engraved

Handy Engraved

Hempstead Engraved

Pease Brushed-Incised

Allen phase, ca. post-A.D. 1650

Bullard Brushed

Constricted Neck Punctated

Deshazo Brushed-Appliqued

Hood Engraved (effigy bowls)

Hume Engraved

Killough Pinched

King Engraved

La Rue Neck Banded

Lindsey Grooved

Mayhew Engraved

Patton Engraved

Spradley Brushed-Incised

 Belcher phase set, ca. A.D. 1500-1680

 Avery Engraved

Belcher Engraved

Belcher Ridged

Cowhide Stamped

Foster Trailed-Incised

Glassell Engraved

Hodges Engraved

Karnack Brushed-Incised

Moore Noded

Taylor Engraved

Latest Belcher phase set, 1680+ (and other post-1680 contexts)

 Natchitoches Engraved

Hodges Engraved

Glassell Engraved

Keno Trailed

Ebarb Incised

Foster Trailed-Incised

 Frankston phase set, ca. A.D. 1400-1650

 Bullard Brushed

Fair Plain

Hood Engraved (effigy bowls)

Hume Engraved

Hume Plain

Killough Pinched

La Rue Neck Banded

Maydelle Incised

Poynor Brushed

Poynor Engraved, multiple varieties

 Kinsloe phase, post A.D. 1680-1830

 Darco Engraved

Emory Punctated-Incised

Henderson Plain

Keno Trailed

Natchitoches Engraved

Patton Engraved

Simms Engraved 

McCurtain phase set, ca. A.D. 1300/1400-1700

 Avery Engraved

Clark Engraved

Emory Punctated-Incised

Hudson Engraved

McKinney Plain

Nash Neck Banded (shell)

Simms Engraved

 Texarkana phase set, ca. A.D. 1400/1450-late 17th century

 Avery Engraved

Barkman Engraved

Bowie Engraved

Foster Trailed-Incised

Hatchel Engraved

Karnack Brushed-Incised

Keno Trailed (latest part of phase)

McKinney Plain

Moore Noded

Nash Neck Banded

Pease Brushed-Incised

Simms Engraved

Titus phase set, ca. A.D. 1430-1680

Anglin Corn Cob Impressed

Bailey Engraved

Bullard Brushed

Cass Appliqued

Gilmer Engraved

Harleton Appliqued

Johns Engraved

Karnack Brushed-Incised

Killough Pinched

La Rue Neck Banded

Maydelle Incised

Pease Brushed-Incised

Ripley Engraved, multiple varieties

Taylor Engraved

Turner Engraved, multiple varieties

Wilder Engraved, multiple varieties

Latest set in Titus phase area, ca. A.D. 1680+ (best known at the Clements site (41CS25)

 Clements Brushed

Darco Engraved

Hatinu Engraved

Keno Trailed

Simms Engraved

Taylor Engraved

Post-A.D. 1680, mid-Red River and upper Sabine River basin

Emory Punctated-Incised

Natchitoches Engraved

Simms Engraved

Womack Engraved

Womack Plain


Figure 2. Important Early Caddo sites: 1, George C. Davis; 2, Fasken; 3, Roitsch; 4, Taddlock; 5, Hudnall-Pirtle; 6, Grace Creek; 7, Bison A; 8, Hale; 9, Boxed Springs; 10, Pace; 11, Boyette; 12, Joe Meyers; 13, Crenshaw; 14, Mounds Plantation; 15, Gahagan; 16, Jaggers; 17, Henry Chapman; 18, Bowman; 19, Bentsen-Clark. Figure prepared by Sandra Hannum.

Middle Caddo Sites_Final

Figure 3. Important Middle Caddo sites, major Red River Caddo centers occupied during the Middle Caddo period, and defined Middle Caddo period phases. 1, Harling; 2, Sanders; 3, Fasken; 4, Roitsch; 5, Holdeman; 6, Hatchel; 7, Hurricane Hill; 8, 41RR181 and Little Mustang Creek; 9, 41TT670; 10, 41CS150; 11, Coker (41CS1); 12, 41TT372; 13, 41FK70; 14, Benson’s Crossing; 15, Crabb (41TT650); 16, Harold Williams; 17, 41UR21; 18, Big Oaks; 19, Griffin Mound; 20, 41UR133; 21, 41UR8; 22, McKenzie; 23, Spoonbill; 24, 41RA65; 25, T. M. Moody; 26, 41WD518; 27, Yarbrough; 28, Charlie Crews; 29, Jamestown; 30, Carlisle; 31, Langford; 32, Bryan Hardy; 33, 41HS74; 34, Old Brown Place; 35, Oak Hill Village; 36, 41PN14; 37, Musgano (41RK19); 38, Pace McDonald; 39, 41CE42; 40, 41CE289; 41, George C. Davis; 42, 41CE290; 43, 41NA20; 44, Washington Square; 45, Tyson; 46, 41SA123; 47, 41SA89; 48, Knight’s Bluff; 49, 41FK7; 50, Hudnall-Pirtle; 51, Gray’s Pasture; 52. Redwine.

Late Caddo Phases - FINAL

Figure 4. Late Caddo period phases in East Texas and immediately surrounding areas.


Figure 5. Clusters of Historic Caddo sites and defined phases. Figure prepared by Sandra Hannum.

Refining and further bracketing the age and intra-site chronological relationships of the ceramics in ancestral Caddo sites in East Texas remains to be fully accomplished, but work is underway through intensive radiocarbon dating efforts (including the dating of organic residues preserved on ceramic vessels and sherds). It is also important that the old and new ceramic types used in the region (including ceramic types yet to be recognized) be fully defined, and differences and similarities in ceramic decoration and manufacture be established. Such analyses can be employed then to answer questions of the social and cultural affiliation of ancestral Caddo groups, and the place of  particular ceramic assemblages within specific communities of Caddo people.

Suggested Readings

Early, A. M.
2012    Form and Structure in Prehistoric Caddo Pottery Design. In The Archaeology of the Caddo, edited by T. K. Perttula and C. P. Walker, pp. 26-46. University of Nebraska Press, Lincoln.

Perttula, T. K.
2013    Caddo Ceramics in East Texas. Bulletin of the Texas Archeological Society 84:181-212.

Story, D. A. and E. B. Jelks
2009    Foreword. In Handbook of Texas Archeology: Type Descriptions, edited by D. A. Suhm and E. B. Jelks, pp. 1-5. Reprint Edition, Gustav’s Library, Davenport, Iowa.

Suhm, D. A., and E. B. Jelks (editors)
1962    Handbook of Texas Archeology: Type Descriptions. Special Publication No. 1, Texas Archeological Society, and Bulletin No. 4, Texas Memorial Museum, Austin.

Suhm, D. A., A. D. Krieger, and E. B. Jelks
1954    An Introductory Handbook of Texas Archeology. Bulletin of the Texas Archeological Society 25:1-562.

About the Authors

Dr. Timothy K. Perttula is a Research Affiliate at the CRHR, and Manager at Archeological & Environmental Consultants, LLC

Dr. Robert Z. Selden Jr. is a Research Associate at the CRHR


Turner Collection, O NAGPRA 2012.1.504
Vessel image appears courtesy of the Anthropology and Archaeology Laboratory
Vessel animation appears courtesy of the Virtual Curation Laboratory


CRHR:ARCHAEOLOGY – A Note from the Architects (Ellis and Wackerman)

In the Fall 2013 semester, Dr. Selden approached us with an idea for a digital collection that would document and showcase his work, and that of others, on Caddo ceramics and aspects of Texas archaeology and history. At the time, we were members of the East Texas Research Center (now with the Center for Digital Scholarship), and this collection would reside in the department’s digital archives. Work on the collection, eventually named CRHR:ARCHAEOLOGY, began in October 2013, developed slowly, and is now one of our primary priorities.

Almost every aspect of the work that we have done with CRHR:ARCHAEOLOGY has been unique for us, professionals who prior to working with Dr. Selden had created numerous digital collections ranging over many subject areas. The work on CRHR:ARCHAEOLOGY would be, in our opinion, likely unique to any other digital archives and archivists as well and there are several reasons supporting this. And, of course, if the work is unique then the product will probably be unique as well. For us, being in a service-oriented profession, it is this latter aspect of the project which is most significant. We would then like to describe some of our work on this project, in the process highlighting some of the unique aspects of the collection.


In the context of any digital archives, one of the more unique aspects of CRHR:ARCHAEOLOGY is the descriptive information supplied in CONTENTdm. Referred to as metadata, this information in its most basic form is a set comprised of a field, for example Site Name, and a value, for example Vanderpool. On average, 10 to 15 fields are used by most collections. Prior to CRHR:ARCHAEOLOGY, the highest number of fields used by one of our collections was around thirty. CRHR:ARCHAEOLOGY uses fifty-four fields, of which Dr. Selden created more than thirty-five.

Although impressive, the number of fields alone isn’t the only determiner of uniqueness. Generally, the fields used for digital collections of cultural heritage objects are similar: Title, Subject, Description, Date(s), Format and a few others. CRHR:ARCHAEOLOGY has these fields, but it also has discipline-specific fields that up to this point have probably been little used in other digital collections. Some of these fields include Interior and Exterior Surface Treatment, Orifice Diameter (cm.), Pigment Use/Location and Base Shape.

On one hand, the metadata used in CRHR:ARCHAEOLOGY is significant in that it presents to users a novel set of resources. Yet, considering that these users are most likely professionals – let’s just say archaeologists in general – the significance can be more measured. Through CONTENTdm’s Advanced Search features, users are allowed to sort and group items within a collection based on a number of selected metadata fields: a user could limit a search result to only display items that possess metadata values “A” and “B”, “A” or “B”, etc. For example, within the CRHR:ARCHAEOLOGY database a user could limit the results to only display those items that are 5 cm. in height and that were also found at the Vanderpool site. For a researcher, such a search method can allow patterns in the information to more easily develop, perhaps providing a combination of values previously unrecognized.


In planning and developing what would become the webpage for CRHR:ARCHAEOLOGY, there were several examples – existing within the department’s digital archives and on external sites – which were referenced. The external examples were essentially CSS templates (openly distributed CSS files accompanied for display via HTML files), with the examples found in the department being modifications of similar templates.

Eventually, an external template was chosen from templatemo (www.templatemo.com). Using a pre-designed template did allow for the application of a rather complex design which would have otherwise either been difficult or required quite a lot of time to create. But, as alluded to above, modifications were necessary. With any such template, there are various features that must be edited, namely text and images. The amount of time taken to complete these changes was, though, relatively small, as all that was needed was a rudimentary knowledge of HTML and information provided by Dr. Selden. Yet, for the CRHR:ARCHAEOLOGY webpage, there were several other changes that required much more effort and time. One of these surrounded an image carousel that was included in the original design template. This carousel displayed one image at a time, cycling through five or more. To move and control the cycle of the images, the template included a section of javascript, which also regulated the speed and interval of each image cycle.

During the process of creating the web page, we noticed that certain areas of the CONTENTdm database were not functioning properly. The issues were only present on the database associated with CRHR:ARCHAEOLOGY – each collection having a singular relation to the database – so it was apparent that the source of these would be found somewhere in the template code. After a week spent reviewing the HTML and CSS, with minor edits and tweaks during that time, the problem that was causing these issues could not be found. As a last resort, we dismantled the entire site structure – at that point, the main page and four secondary pages were essentially complete, with text, images and active links throughout. When we did this and reloaded the database, the issues were corrected.

We then reconstructed the site, piece by piece, refreshing the database at each step and checking for any problems. This strategy worked, as we found, when embedding the code for the image carousel, that the database issues reappeared. The source of the problem was the javascript used for the carousel. With that section of code removed, and with the accompanying HTML and CSS left intact, the database issues were not present. Yet, without this javascript the carousel could not work. There were many other carousel options which could have been chosen at that point, but it was decided to use a simple, static header: simplicity in design has been a focus for CRHR:ARCHAEOLOGY since the beginning.

3D Models

One aspect of CRHR:ARCHAEOLOGY as a project for which we have not achieved the desired level of success is the capture and display of 3D models of Caddo pottery. As any reader of this blog knows, such models occupy a significant part of Dr. Selden’s work, and the display was, in fact – if our memory is correct – one of the initial discussion points during the introductory period of this project. At the time, while we could not find a comparable content management system, such as CONTENTdm, that successfully displayed 3D models of any type (although, several institutions did claim to have 3D galleries or collection, but in reality these were 2D images of 3D representations), we were confident that we could find a solution of some type.

There were at least three ways in which we could have displayed the 3D models: within the CONTENTdm database; within a CONTENTdm web page; within a third party viewer. The listing follows order of preference, as the second and third methods of display would not provide immediate access: for both, the user would have to navigate through a set of links to arrive at a model. As these models would be accompanied by 2D representations, metadata and other relevant information and files, separating access was thought of as not being ideal.

The first method, displaying models within the CONTENTdm database, did not work. There were several reasons for this but the main was that the image/object viewer used within the database would not support the common – or any at all – 3D file formats. Within the second method there are two sub-methods: displaying a model through HTML or through a third-party viewer. The first sub-method was somewhat of an attempt at a work-around of the first method. This was possible, at least in theory, as although the viewer used by CONTENTdm does not support 3D file formats, the database will accept and allow the upload of any format. The process, then, would be to embed the 3D file in a CONTENTdm web page by means of HTML tags. For the second sub-method, everything would be the same except for the use of a third-party viewer. Ultimately, these sub-methods did not work, and while the reason is not entirely clear it likely had something to do with the manner in which the file was accessed by the code or the viewer. The third method, displaying a model via a third-party viewer, with the file residing in that site’s database, also has two sub-methods: link from CONTENTdm to the site on which the third-party viewer resides, or embed the third-party viewer in a CONTENTdm page. There are some similarities between this and the second method, with one main difference: the files are also hosted by the third-party. These methods did technically work, but there were issues concerning file integrity and sharing and so neither were used.

So, there were many attempts but not one met everyone’s expectations. These did not though completely rule out the inclusion of 3D models, just the display: for some of the objects within the CRHR:ARCHAEOLOGY database a 3D model has been included. These models were created by Dr. Selden in the usual manner, and the files do reside in CONTENTdm, but the eventual output format, instead of OBJ, ICF or any other, is PDF. In the CONTENTdm database, these PDF 3D models appear as 2D representations – the 3D functions, such as zoom and rotate are not accessible. Yet, the user can download these files and, after loading them into a 3D image viewer, interact with a functional 3D model. Although this method does provide some functionality, other ways in which to display 3D models for CRHR:ARCHAEOLOGY are still being explored.

Work still continues on CRHR:ARCHAEOLOGY within the Center for Digital Scholarship. At this point, a major task is quality control: identifying and removing inconsistencies or inaccuracies found in the database and the web page. Yet, at the same time, we are continually looking to further develop the project to more fully realize the goals of Dr. Selden.

About the Authors

Ann Ellis is the Metadata Librarian in the Center for Digital Scholarship at SFASU (aellis@sfasu.edu)

Dillon Wackerman is the Digital Archivist in the Center for Digital Scholarship at SFASU (wackermad@sfasu.edu)

**I would like to take a moment to thank these two creative individuals and the Center for Digital Scholarship at SFASU. Without their hard work and patience (and tolerance of my constant prodding), the idea of CRHR:ARCHAEOLOGY would have remained just that – an idea.**


Ripley Engraved Bowl Motifs and Titus Phase Communities (Fields)

This post is a preview of part of a paper I’ll be giving at the upcoming 2014 Caddo Conference and East Texas Archeological Conference, in which I’ll talk about ways in which we tried to address the issue of Titus phase communities based on excavations we did in 2010 at sites on the U.S. Highway 271 Mount Pleasant relief route project. One part of that study, done by my colleague Eloise Gadus, involved looking at the distributions of motifs on Ripley Engraved bowls to assess whether they might be informative about local community extent, not just for our immediate project area (the Tankersley Creek valley) but for the broader Titus phase area as well.

The study included 599 vessels from 17 sites: the cemeteries at the Thomas B. Caldwell, A. P. Williams, and Duncan Anderson sites in the middle Tankersley Creek valley; three other excavated cemeteries close to the project area at the Mockingbird, Alex Justiss, and Pilgrim’s Pride sites; the Tuck Carpenter and Johns cemeteries farther south in the Big Cypress basin; several cemeteries in the middle part of the basin at the Lone Star Lake, Rumsey, and Keeling sites; and four cemeteries at the Henry Williams, Enis Smith, Henry Spencer, and Frank Smith sites in the Little Cypress Creek basin. Also added were two points of comparison in the Sabine River drainage. These are the two family cemeteries at the Pine Tree Mound site and graves from several sites on Caney Creek southwest of our project area.


This map shows the locations of Titus phase cemeteries used in Ripley Engraved bowl variety comparisons.

We recognized 18 varieties of Ripley Engraved bowls in our study based on structural differences in the engraved motifs. Nine of the varieties were newly defined, and the others had been defined by Tim Perttula and colleagues based on motifs originally illustrated in Pete Thurmond’s 1981 M.A. thesis.

Analysis of the distributions of these varieties found a great deal of overlap spatially, providing little support for the idea that variation in Ripley bowl motifs is a productive way to consistently see group identity and local community boundaries, at least not variation as captured by this analytical scheme. While it is true that that some of the ceramic vessels in the Tankersley Creek mortuary assemblages do look different than those from the Alex Justiss and Pilgrim’s Pride sites not far to the east and south, and this could be seen as evidence of boundaries between local communities, the Tankersley Creek ceramics are decidedly similar to those from the Tuck Carpenter and Johns sites even farther south, as well as those from the Mockingbird site to the north, and they are consistent with assemblages across a large area covering much of the upper part of the Big Cypress basin.

Even if potters and groups of potters chose motifs and variable expressions of those motifs in part to reflect social identify, they did that within the context of a common widespread ideology. Further, Caddo potters could and did innovate in motif construction, while still using a set number of basic structures and elements, and these motifs were understood and accepted by communities across and beyond the Titus heartland. Thus, vessel trade, shifting community boundaries and centers of political power through time, and group coalescence and splitting could make it extremely difficult to see community associations in the ceramics.

The one pattern we noted in our study relates to Titus phase communities viewed broadly, but not to communities narrowly defined, and probably is part of the same pattern that led Tim Perttula to see upstream and downstream ceramic subtraditions in the Titus phase heartland and to relate them to what he calls core communities. In our data, this pattern can be seen in the following: (1) we could place 12 of the 15 collections into two groups, albeit ones that are not very homogeneous; and (2) most members of these groups have distinct spatial distributions (see map above).

One group consists of collections from the following seven sites: Thomas B. Caldwell, A. P. Williams, Duncan Anderson, Mockingbird, Tuck Carpenter, Johns, and Henry Spencer. In all seven, Carpenter is the predominant Ripley Engraved variety, with variety Spencer being equally dominant at the Henry Spencer site alone. Beyond this, no two assemblages look exactly alike.



These are drawings of motifs representing the Carpenter (top) and Spencer (bottom) varieties of Ripley Engraved bowls. Variety Carpenter is especially common at cemeteries in the upstream part of the Cypress Creek basin.

The second group consists of the following five collections: Frank Smith, Enis Smith, Henry Williams, the Middle Cypress sites, and Alex Justiss. What unites them is that each is dominated by variety McKinney and/or variety McKinney-Enis Smith and that variety Gandy ranks second or third.


These are drawings of motifs representing the McKinney (top), McKinney-Enis Smith (middle), and Gandy (bottom) varieties of Ripley Engraved bowls. All three varieties are relatively common at cemeteries in the downstream part of the Cypress Creek basin.

Three collections do not fit into either group and also are unlike each other. Each of these three outliers—Pilgrim’s Pride, Pine Tree Mound, and the Caney Creek sites—has its own constellation of predominant varieties.

All of the sites in the first group but one, Henry Spencer, are in the northwestern part of the basin, and all of those in the second group, except Alex Justiss, are in the central part. Not surprisingly, two of the outliers, Pine Tree Mound and the Caney Creek sites, are geographically separated from the other sites. The third, Pilgrim’s Pride, is not.

These distributions suggest five main conclusions. First, they support the contention that different but related core communities with distinct ceramic subtraditions occupied the two parts of the Titus phase heartland. Second, they suggest that there were ties between the southeastern heartland core community and whatever was going on in the Little Cypress basin to the south. Third, with one site in each group being out of place spatially, they suggest that there was movement of potters and people between core communities. Fourth, parts of the Titus phase area outside the Cypress Creek basin supported their own core communities. And fifth, sites like Pilgrim’s Pride, which is an outlier ceramically but not spatially, imply that at certain times and places within the heartland there were small local communities that chose to decorate their pottery differently than their neighbors, presumably reflecting different ideas about connections between engraved motifs and a widely held Caddo belief system.

About the Author

Mr. Ross C. Fields is the President of Prewitt and Associates, Inc.(rfields@paiarch.com)

**Next week we’ll discuss some preliminary findings from a number of East Texas Cemeteries (including a pet cemetery) that the CRHR is investigating with our new GPR**