Category Archives: Projects & Research

WPA Highland Lakes Records Update by Annie Riegert

Annie Riegert is a Curatorial Technician at TARL. This article is part of the June 2018 TARL newsletter.

An ongoing digitization and records inventory project is underway at TARL. Under a 2017-2018 Texas Preservation Trust Fund (TPTF) grant, the project targets excavation and survey reports conducted in the late 1930’s as a subset of the larger Works Projects Administration (WPA) project. In the wake of the Great Depression and much needed employment, the program offered jobs on public works projects including archaeological survey and excavation throughout the country. Today, we are working on archiving the product of the WPA survey and excavation in Highland Lakes Area with a focus on Lakes Austin, Buchanan, and Travis. Further artifact analysis was conducted by the University of Texas in the 1960’s. Interest in the WPA project has produced many reports and analyses in various forms including records on microfilm. A database, which willhouse all excavation and survey data, will facilitate a more robust understanding of the extent of the WPA project in our local area. Additionally, digitization will enable greater access to records while also ensuring long term document preservation. We are looking forward to sharing the finished project for further research into the history of archaeology in central Texas.

Archive document depicting stone tools found at a WPA site in Travis County.

Goodbye to Flash! Changes Underway for Texas Beyond History by Susan Dial and Steve Black

As many of you know, we have been hard at work for the last two years with UT’s Liberal Arts Instructional Technology Services (LAITS) trying to bring our 17-year-old website, Texas Beyond History, into the 21st Century. We are happy to report that this complicated “makeover” is nearly complete and promises both esthetic and functional improvements. The project transforms our main entry portals as well as educational activities and interactive graphics originally programmed in Adobe Flash into a more modern technology accessible on many more platforms. This will allow viewers –whether in the classroom or in the field–to use tablets and, to a lesser extent, cell phones, to engage in TBH’s interactive learning activities, open interactive charts and maps, and fully utilize the resources of the website. TBH will have a fresh, new look but more importantly should function more smoothly.

Website technology has advanced exponentially since the founding of TBH in 2001, when Steve and I, along with student website developer Meg Kemp, unveiled the website and the first 20 site exhibits. At the time, we were excited to offer many interactive features for maps, graphics, and student learning activities using Flash technology. The “cat’s meow” for its time, this program provided exciting tools to incorporate animation and other interactive capabilities in maps and graphics (e.g. opening up stratigraphic layers in a profile map). Unfortunately, Flash is no longer being supported by many browsers and has been dropped altogether by Apple and some newer Android devices. Viewers who use Apple products, particularly iPads and Mac books, may have been encountering blank pages where our traditional TBH interactive maps and Kids Only revolving carousel should be.

As further complication, TBH was designed for “mousing” on a desktop or laptop, before touchpad navigation came into vogue. Many of our interactive scenes where users “mouse over and click” on segments of paintings to access more detailed information and site-specific photos of evidence (ie., Frank Weir’s remarkable painting of a prehistoric burial scene from Loma Sandia cemetery) cannot be utilized on these devices. As might be anticipated, this is a particularly critical problem in the classroom, and for K-12 teachers in particular, as schools increasingly are providing individual tablets for student use. For LAITS, the process has been especially challenging due to the volume of Flash content on TBH and markedly different formats in each of the Flash activities. There has been no “one size fits all” solution to reprogramming this content. Over the last year, however, LAITS web developers engineered a process to strip out content and imagery and then recreate the 40+ interactives using HTML5.

Along with the technical changes, there also will be a new look for TBH. Instead of the interactive Texas sites map, TBH will soon have a colorful and streamlined portal for accessing all of the website sections. (A revamped version of the familiar TBH map page will be accessible in a section called Site Explorer and made functional for all browsers.)

Rollout of our revamped website is slated for sometime this Fall (2018). This is particularly important because TBH is heavily used in university archeology classes as well as in 4th and 7th-grade classrooms. We continue to receive emails from Social Studies teachers who have been stymied by the non-working Flash activities, but are anxious to once again use educational interactives such as “Through the Eyes of the Explorer: Cabeza de Vaca on the South Texas Plains.” Older students (even university students, according to Texas A&M professor Alston Thoms,) have used the kids Flash activity “Stratification in Action!” to better understand complex stratigraphic processes such as that which occurred over thousands of years on the Medina River at the Richard Beene site, on which the activity is based.

This summer (with Steve back at TARL just in time for the TBH review process!) we will continue testing the updated website, checking new functions, and kicking the tires, so to speak. It is a painstaking process with numerous technological bugs lurking in the 60,000+ files that comprise TBH. Fortunately TBH Associate Editor Heather Smith and Education Advisor Carol Schlenk have been able to join in the effort. While change can be difficult (if not agonizing), we at TBH are determined to embrace the opportunity to usher this much loved and critically acclaimed public education website into the modern era. We are grateful for the time and dedicated efforts of the LAITS staff and student technical assistants.

And as for what lies beyond the website revamp process, Steve is already at work creating new plans and a vision of the future for TBH. Stay tuned!

Archeomalacology in Texas by Ken Brown

Ken Brown is a visiting researcher at TARL. This article is part of the June 2018 TARL newsletter.

Much of my research over the past couple of decades has focused on reconstructing Texas paleoenvironments by studying snails and sphaeriid clams. Snails are very moisture-sensitive, and most of the state (except east Texas) doesn’t have acid soils, so the shells usually preserve well, except where groundwater or shrink-swell processes have been active. If proper sampling is done, enough specimens can often be recovered to allow some rudimentary quantitative methods to be applied. Although snails are not especially temperature-sensitive, they make good paleo-moisture proxies, and because vegetation and moisture are correlated, snails can also hint at past vegetative cover. Aquatic snails are especially diagnostic of past hydrologic conditions. Some species can tolerate hot, stagnant, poorly oxygenated water with high solute levels, and others require cool, well-oxygenated, fresh flowing water, so an aquatic assemblage can tell us a lot about past stream conditions. Snails are sessile animals with short lifespans, which means that individual field samples capture paleoenvironmental profiles that are restricted in time and space.

Proper sampling means fine-mesh wet-sieving of fairly large samples and picking of whatever shows up, including shell fragments and juveniles. I use a standard nested set of oversize brass geologic sieves (18 inch diameter; Fig. 1). Mesh sizes are #10 (2 mm), #18 (1 mm), and #35 (0.5 mm). The British were the earliest proponents of this kind of research, and they often use one-liter samples, perhaps modeling their sampling methods after those of pollen analysts. But England is a wet place replete with snails, and while one-liter samples may be adequate there, Texas is a lot drier, so my target sample size is usually about 12-15 liters. There are perhaps 185 or so species of terrestrial snails native to Texas, and recent research suggests there might be another 60 or so aquatic species. Many of these (perhaps as many as 60% for terrestrial species?) are too small to be captured on quarter-inch archeological field screens. Most or nearly all the juveniles will fall through standard field screens, even for the larger species, so fine mesh sampling is essential if the complete spectrum of species is to be recovered.

Figure 1. Nested geologic sieves used for sampling snails.

For paleoenvironmental studies, I’ll try to collect a few pilot samples first to find out if snails are preserved at all, what condition they’re in, and how abundant they are. If the site looks promising, I’ll try to collect a continuous column of samples from a representative profile, perhaps in 10 cm vertical increments. To get the required 15 liters from a 5 cm thick sample requires a column about 60-80 cm wide, which is much more extensive than the usual pollen sample column. I try to avoid sampling features, since by definition features are places with extensive human disturbance, but for some reason many of the snail samples reported in the literature (for example, Wilson-Leonard) have come from cultural features.

In the field, I measure out 15 liters of sediment, plus a little extra allowing for shrinkage, bag it in sandbags, and return it to TARL, where it is weighed and the volume re-measured after drying. Then the 15-liter sample is soaked in tapwater overnight (with dispersant if clay-rich) and wet-sieved through nested sieves. Small specimens are identified with a binocular microscope at low magnification (usually no more than 10X). Shell fragments are also saved and weighed, because recovering fragments is the only way to estimate how many fragments have been lost to breakage. Juveniles are tabulated separately and identified if possible. Snails are much like vertebrates, in that individuals often don’t develop species-diagnostic characters until adulthood is reached, so sometimes juveniles can only be identified to the family or genus level. On the other hand, if there is only one species in a genus present at a site, it is sometimes possible to assume that all the juveniles are representatives of the same species. For example if the only adult Gastrocopta found at a site are Gastrocopta pellucida, it is probably safe to assume that any juvenile Gastrocopta belong to that same species. If juveniles can be identified, adult/juvenile ratios can be computed. This will tell us something about juvenile mortality, which is linked to environmental conditions. High juvenile mortality probably means severe water stress, although fecundity can also be an evolutionary adaptation to living in stressful habitats.

Most of the formal archeomalacological studies in Texas (those involving extensive fine sieving of sediments) have been done east of the 98º longitude line, or in other words east of the 32-inch annual precipitation isohyet, in the better watered part of the state. I’ve done studies at the Berger Bluff, Vara Daniel, and Fish Creek Slough sites, and other researchers have done studies at the Aubrey, Wilson-Leonard, Richard Beene, and various other sites. There have been a few studies in the Texas Panhandle and High Plains, at Lubbock Lake, Lake Theo, and at Mackenzie Reservoir (species lists with no specimen counts or sample volume given are provided for the Rex Rodgers and Snail Bed sites). Species lists are available for the Plainview site and several other sites on the High Plains (Neck 1995), but again, specimen counts and sample volumes are not given. At least one site (41CC112; Treece 1992:Table 15) in west-central Texas has been reported.

The northeast and southwest quadrants of Texas, on the other hand, are a blank slate. The archeomalacology of East Texas remains completely unknown. The acidic, sandy soils and constant saturation with phreatic water are hostile to shell preservation. Indeed, I would speculate that there may be some calcium-poor areas in East Texas where even living populations are absent today.

The southwest quadrant of Texas is mostly unexplored. One of the earliest archeomalacological studies in the state was done by Cheatum (1966) as part of the Amistad paleoecological project at Eagle Cave, Bonfire Shelter, Devil’s Mouth, and Devil’s Rockshelter, but the number of samples, volumetric size of samples, and number of specimens are not disclosed; intrasite provenience is reported only by stratum. Thus, we are left only with a laundry list of species. More recent studies at Bonfire and Skyline Shelter provide quantitative data, but both of these sites are rockshelters, and the Lower Pecos remains understudied.

In the entire vast area of the Big Bend (more than 12,300 square miles, not counting parts of Culberson, Hudspeth, and Reeves counties), no formal archeomalacological studies have been done until recently, as far as I know. The Big Bend is truly a blank slate. The potential of the driest two-thirds of Texas for these studies is largely unknown. Pierce (1987) recovered large numbers of specimens at Lubbock Lake, but most of these came from bedload gravels or cross-bedded sand in Stratum 1. The research potential of alluvial terrace deposits in west Texas has been largely unknown. Cheatum’s faunal lists from Devil’s Mouth, and Devil’s Rockshelter offer promise, but without specimen counts, their productivity is unclear. We know that snails are moisture-seeking organisms, and we know that they are most abundant and diverse in the wettest parts of Texas, but how well do they function as paleoenvironmental proxies in the driest parts of the state?

Figure 2. Site locations in Brewster and Val Verde Co.

In July, 2013, I visited the Genevieve Lykes Duncan site (41BS2615), a stratified Paleoindian site on the O2 Ranch, near the confluence of Terlingua Creek and Davenport Draw in Brewster County, and collected a continuous column spanning the entire Holocene. In May, 2016, I visited the Sayles Adobe site (41VV2239), a high terrace site in Eagle Nest Canyon near Langtry. Tori Pagano collected a discontinuous series of eight pilot samples for me, from sandy sediments representing about a 2500-year span of the Late Holocene. I have now processed and analyzed the samples from both of these sites (Fig. 2), and in future issues of this newsletter, I hope to summarize what I’ve learned.

References Cited

Cheatum, E. P.
1966 Report on Mollusk Shells Recovered From Four Archeological Sites in the Amistad Reservoir. Pages 227-243 in Dee Ann Story and Vaughn M. Bryant, Jr. (assemblers), A Preliminary Study of the Paleoecology of the Amistad Reservoir Area. Final Report of Research Under the Auspices of the National Science Foundation (GS-667).

Neck, Raymond W.
1995 Molluscan Remains. Pages 59-67 in Vance T. Holliday, Stratigraphy and Paleoenvironments of Late Quaternary Valley Fills on the Southern High Plains. Geological Society of America, Memoir 186.

Pierce, Harold G.
1987 The Gastropods, with Notes on Other Invertebrates. Chapter 6 (pages 41-48) in Eileen Johnson (ed.), Lubbock Lake. Late Quaternary Studies on the Southern High Plains. Texas A&M University Press.

Treece, Abby C.
1992 A Study of Five Annular Burned Rock Middens from the O.H. Ivie Reservoir, West Central Texas. Unpublished MA thesis, University of Texas at Austin.

Identification of an Unusual Mineral Telltale in a Leon Plain Sherd from the Edwards Plateau of Central Texas by David Glen Robinson

David Glen Robinson is a visiting researcher at TARL. This article is part of the June 2018 TARL Newsletter.   

Recent ceramic petrographic work at the Texas Archeological Research Laboratory (TARL) has identified an unusual tempering material in a sherd of Leon Plain pottery from the Albert and Bessie Kronkosky State Natural Area in Bandera County, Texas. In petrographic work performed on a sherd of Leon Plain for Texas Parks and Wildlife Department under the supervision of Aina Dodge, the author identified the mineral olivine. The mineral was a co-temper with the more usual bone temper common in Leon Plain ceramics.

Olivine is an igneous mineral forming deep within volcanic vents or channels. It is considered an intrusive mineral and a component of basic, granular volcanic rocks such as gabbros. At the earth’s surface, olivine and other intrusives erode relatively rapidly, thus they are somewhat rare in surface deposits and not widely distributed. In hand samples, or gem forms, olivine is olive green, hence the name, and has vitreous luster.

The microphotographs presented here are all from the subject sherd, BN266, designated from its site, 41BN266. All the photos are taken in cross-polarized light, or xpl. This was done to show the array of distinctive colors of various specimens of olivine in the sherd matrix.

Given the rarity of olivine, the likely production locale of the pottery was several score miles northeast of 41BN266 on the Llano Uplift, where intrusive rocks were brought to the surface during the uplift events. An alternative source for intrusives is in the Big Bend and Marfa volcanic regions hundreds of miles to the southwest of Bandera County. More specific estimates of the sherd’s resource/production locality are not feasible at this time.

Figures 1-4.Microscopic petrographic images of olivine mineral temper identified in Leon Plain pottery from 41BN266.


The ceramic petrography on the 41BN266 sherd was sponsored by TPWD and supervised by Aina Dodge. Timothy K. Perttula served asthe ceramic consultant to the project. Their guidance is gratefully acknowledged.

Marine Shell Gorgets from Ancestral Caddo Sites in the Southern Caddo Area by Timothy K. Perttula

Timothy K. Perttula is a visiting researcher at TARL. This article is part of the June 2018 TARL Newsletter.   

Plain and engraved gorgets are rare occurrences on ancestral Caddo sites in the southern Caddo area, as they have been found in burial features at only 13 sites in Southwest Arkansas, Northwest Louisiana, southeastern Oklahoma, and East Texas (Table 1). These include 12 plain and 28 engraved gorgets, and the engraved gorgets have a number of different styles as defined by Brain and Phillips (1996). A number of these gorgets are in the collections of the Texas Archeological Research Laboratory at The University of Texas at Austin, and I have documented them as part of the recent study of a 15th-16th century engraved gorget from the Pipe site (41AN67) in the holdings of the Gregg County Historical Museum in Longview, Texas.

Fifty-five percent of the gorgets (and 57 percent of the engraved gorgets) from southern Caddo sites are from Middle Caddo period Sanders phase burial features in the East Mound at the T. M. Sanders site (Figure 1), and no other site has more than three gorgets (see Table 1). The gorgets occur in Middle Caddo (ca. A.D. 1200-1400, n=22) (Figure 2), Late Caddo (ca. A.D. 1400-1680, n=16) (Figure 3), and Historic Caddo (ca. A.D. 1680-1800, n=2) burial features (see Table 1).

Figure 1. Engraved turkey gorget in Burial 20 at the T. M. Sanders site (41LR2) in the Red River basin.
Figure 2. Plain gorget from the Jess
Alford site (41HP5) in the South Sulphur River basin.

More than 77 percent of the plain and engraved gorgets from the southern Caddo area are from burial features in Middle and Late Caddo period mound centers on the Red River, and approximately 83 percent of the southern Caddo area sample come from burial features on Red River valley sites. The remainder of the gorgets are from sites in the Ouachita (n=1), Little (n=1), Sulphur (n=1), Sabine (n=2), and Neches (n=2) river basins in southwest Arkansas, southeastern Oklahoma, and East Texas. The gorgets in sites in the Sulphur, Sabine, and Neches river basins in East Texas are from non-mound burial features.

Figure 3. Triskele style engraved gorget in Burial 19 at the Paul Mitchell site (41BW4) in the Red River basin.

Although rare, the clear association of marine shell gorgets with ancestral Caddo burial features (typically the burials of adult
males) found at mound centers is indicative of the display of marine shell jewelry as a prestige good (Deter-Wolf and Peres
2015:179). Such goods were restricted to the use (in life and at death) of the Caddo social elite living at these mound centers,
as they (a) signified and enabled access to supernatural powers and ritual knowledge by the wearer of the gorget, (b)
legitimized political power in communities both local and regional through the acquiring and display of symbolic materials
with high value social currency, and (c) sanctified claims to ancestral Caddo origins, landscapes, and landmarks (cf. Deter-Wolf
and Peres 2015:170; Marquardt and Kozuch 2016:23). The iconography present on marine shell gorgets on Caddo sites
warrants detailed consideration, therefore, because it reflects belief systems of different Caddo groups, as well as the social
relationships between different Caddo groups as well as more far-flung Mississippian groups (see Brain and Phillips 1996).

References Cited

Bell, R. E. and D. A. Baerreis
1951 A Survey of Oklahoma Archaeology. Bulletin of the Texas Archeological and Paleontological Society 22:7-100.

Brain, J. P. and P. Phillips
1996 Shell Gorgets: Styles of the Late Prehistoric and Protohistoric Southeast. Peabody Museum Press, Peabody Museum of Archaeology and Ethnology, Harvard University, Cambridge.

Deter-Wolf, A. and T. M. Peres
2015 Embedded: Five Thousand Years of Shell Symbolism in the Southeast. In Trends and Traditions in Southeastern Zooarchaeology, edited by T. M. Peres, pp. 161-185. University Press of Florida, Gainesville.

Etchieson, M.
1981 A Shell Gorget from the Kirkham Site. The Arkansas Archeologist 22:1-3.

Harris, R. K., I. M. Harris, J. C. Blaine, and J. Blaine
1965 A Preliminary Archeological and Documentary Study of the Womack Site, Lamar County, Texas. Bulletin of the Texas Archeological Society 36:287-365.

Jackson, A. T.
1934 Jess Alford Plantation on old channel of South Sulphur River 2 ½ miles North of Nelta and 18 ½ Miles Northeast of Sulphur Springs, Hopkins Co., Trenched July 16, 1934 to July 17, 1934. MS on file, Texas Archeological Research Laboratory, The University of Texas at Austin.

Jackson, A. T., M. S. Goldstein, and A. D. Krieger
2000 The 1931 Excavations at the Sanders Site, Lamar County, Texas: Notes on the Fieldwork, Human Osteology, and Ceramics. Archival Series 2. Texas Archeological Research Laboratory, The University of Texas at Austin.

Krieger, A. D.
1946 Culture Complexes and Chronology in Northern Texas, with Extensions of Puebloan Datings to the Mississippi Valley. Publication No. 4640. The University of Texas, Austin.

Marquardt, W. H. and L. Kozuch
2016 The Lightning Whelk: An Enduring Icon of Southeastern Native American Spirituality. Journal of Anthropological Archaeology 42:1-26.

Moore, C. B.
1912 Some Aboriginal Sites on Red River. Journal of the Academy of Natural Sciences of Philadelphia 14(4):526-636.

Perino, G.
1981 Archeological Investigations at the Roden Site (MC-215), McCurtain County, Oklahoma. Potsherd Press No. 1. Museum of the Red River, Idabel.

Perttula, T. K.
2011 The Pipe Site, a Late Caddo Site at Lake Palestine in Anderson County, Texas. Journal of Northeast Texas Archaeology 35:47-80.
2014 The Mitchell Site (41BW4): An Ancestral Caddo Settlement and Cemetery on McKinney Bayou, Bowie County, Texas. Special Publication No. 32. Friends of Northeast Texas Archaeology, Austin and Pittsburg.

Skinner, S. A., R. K. Harris, and K. M. Anderson (editors)
1969 Archaeological Investigations at the Sam Kaufman Site, Red River County, Texas. Contributions in Anthropology No. 5. Department of Anthropology, Southern Methodist University, Dallas.

Todd, J.
2011 An Unique Shell Gorget from Wood County, Texas. Journal of Northeast Texas Archaeology 35:1-4.

Webb, C. H.
1959 The Belcher Mound, a Stratified Caddoan Site in Caddo Parish, Louisiana. Memoirs No. 16. Society for American Archaeology, Salt Lake City.

Buried in Soil Samples from the Gault Site by Laura Vilsack

Laura Vilsack is a visiting researcher from the Gault School of Archeological Research. This article is part of the March 2018 TARL Newsletter.   

In 2001 and 2002, field crews collected 610 bulk soil samples from a single excavation block (Area 12) at the Gault Site in Bell County, Texas. Today, these soil samples are currently being stored at the Texas Archeological Research Laboratory (TARL). I recently contacted TARL and borrowed a portion of the soil samples for magnetic susceptibility analysis to expand upon my thesis research completed in 2016 (Figures 1 and 2). For my thesis work (Vilsack 2016), I focused mainly on the vertical stratigraphy and depositional processes within Area 12 specifically relating to a unique prehistoric cobble feature dating to approximately 14,900 cal B.P. Additional research, from the loaned soil samples, allowed me to extend sampling horizontally across the block instead of just a single vertical profile.

Figure 1. Example of the soil sampling strategy and location of samples collected from the field.


Figure 2. Resulting soil samples compacted into paleomagnetic sample cubes.

Analyzing the soil samples provides a great opportunity for me to address various questions concerning cultural activities around the cobble feature without having to return to the site to excavate the feature again. This feature is approximately 2 by 2 meters wide and varies in thickness between 5 to 15 centimeters. While the surface of the feature yielded very few artifacts, there was an abundance of Clovis and Older-than-Clovis lithic materials surrounding the cobbles.

After acquiring a broad understanding of the depositional environment surrounding the feature, my next goal is to determine if there were any potential signatures of cultural or natural anomalies directly above, at the same elevation as, and within 10 to 15 centimeters below the cobble feature. While magnetic susceptibility by itself does not reveal what type of anomalies occurred, it does aid in the identification of disturbances in the sediments relating but not limited to buried paleosols, movement of soils, organic debris left by humans and areas exposed to heat (Crowther 2003; Tite and Mullins 1971), middens (Dalan and Bevan 2002), and/or pits. Any anomalies detected with a magnetic susceptibility reader are of course dependent on the environment, type of soil, extent of the occupation, and density of the occupation. However, I expect and hope the results will highlight areas to focus further analysis on and reaffirm potential activity areas and/or toss zones identified by artifact densities and distributions that once surrounded this unique prehistoric feature.

References Cited

Crowther, J.
2003 Potential Magnetic Susceptibility and Fractional Conversion Studies of Archaeological Soils and Sediments. Archaeometry 45(4):685-701.

Dalan, Rinita A. and Bruce W. Bevan
2002 Geophysical Indicators of Culturally Emplaced Soils and Sediments. Geoarchaeology: An International Journal 17(8):779-810.

Tite, M.S. and C. Mullins
1971 Enhancement of the Magnetic Susceptibility of Soils on Archaeological Sites. Archaeometry 13(2):209-219.

Vilsack, Laura
2016 Investigations of Area 12: Gault Site. Master’s Thesis, Department of Anthropology, Texas State University, San Marcos.

41AG22: An Historic Caddo Allen Phase Caddo Site in the Neches River Basin, Angelina County, Texas by Timothy K. Perttula

Timothy K. Perttula is an Affiliated Researcher at TARL. This article is part of the March 2018 TARL Newsletter. 

41AG22 is an Historic Caddo site on Jack Creek in the Neches River basin, about 10 km southwest of Lufkin, Texas (Figure 1). The site was located and recorded in November 1939 by Gus Arnold as ET-622 of the University of Texas during the WPA-sponsored archaeological survey of East Texas. This ancestral Caddo site covered about 3 acres in a plowed field, and was marked by three possible mounds about 18-23 m in diameter.

Figure 1. Location of 41AG22 in East Texas.

In 1924, the landowner, a Mr. Jumper, plowed the land where the site is located, and uncovered a human burial with a European trade and glass beads. About 10 years later, Jack Creek flooded and cut a channel through one of the possible mounds, exposing more human remains and ceramic sherds. When Arnold visited the site in November 1939, he collected ancestral Caddo ceramic sherds, lithic tools and debris, and one human molar from the plowed field. As part of long-term East Texas Caddo archaeological research, I recently analyzed the ceramic sherds from the site held by the Texas Archeological Research Laboratory at The University of Texas at Austin (TARL).

 Five of the 199 sherds in the 41AG22 ceramic assemblage are from sandy paste Goose Creek Plain, var. unspecified vessels. These are indicative of a Mossy Grove Culture occupation at the site during some part of the Woodland period (ca. 500 B.C. to A.D. 800). The remainder of the ceramic sherds from 41AG22 (n=194) are from ancestral Caddo vessels. When Dee Ann Story examined the 41AG22 sherds in 1985, she suggested that the site had “at least two components—Early Ceramic (sandy paste sherds & dart points) and Late Caddoan. Very high frequency of brushing suggests quite late Caddoan. Notable is the thickness of some brushed sherds and the incidence of brushed-incised.”

The ancestral Caddo sherds from the sherd are overwhelmingly (91 percent) from grog-tempered vessels. Eight percent of the sherds are from vessels with grog and/or bone or crushed hematite inclusions, and only 0.5 percent of the Caddo sherds are tempered solely with bone.

Only about 15 percent of the ceramic sherds are from plain vessels or the undecorated portions of decorated vessels. One of the plain body sherds has a 5.4 mm drilled perforation, and this sherd may have been used as a spindle whorl. The remainder are from utility ware vessels (n=158, or 95.8 percent of the decorated sherds) or fine ware vessels (n=7, 4.2 percent of the decorated sherds). The plain to decorated sherd ratio (P/DR) of the assemblage is 0.18, consistent with Allen phase sites in the Neches River basin to the north (Perttula and Stingley 2017:Table 22) as well as a few of the Allen phase sites on Bayou Loco in the Angelina River basin, east of 41AG22 (Perttula and Marceaux 2018:Tables 17 and 18). The brushed to plain sherd ratio of the 41AG22 assemblage is 5.17, and the ratio of brushed to other wet paste sherds is 3.75; both ratios are consistent with an East Texas Allen phase ceramic affiliation.

More than 90 percent of the decorated sherds from 41AG22 have brushing marks, either as the sole form of decoration, or in conjunction with appliqued, incised, or punctated decorative elements. The brushed sherds are from Bullard Brushed vessels that have horizontal or vertical-diagonal brushing marks on the rim and opposed, overlapping, and parallel (likely vertically oriented) brushing marks on the vessel body. A number of the brushed-incised (n=20) and the brushed-punctated (n=2) sherds are also from Bullard Brushed vessels. Many of these sherds are from large and thick (>12-15 mm) jars.

One brushed-appliqued body sherd has diagonal brushing marks and a straight appliqued fillet (Figure 2a). Thirteen brushed-incised body sherds are from Spradley Brushed-Incised jars. This utility ware is found on Historic Caddo Allen phase sites in the Neches-Angelina river basins in East Texas. It consists of parallel brushing elements with overlapping straight incised lines that are opposed or perpendicular to the brushing.

Figure 2. Decorative elements on selected utility ware sherds from 41AG22: a, brushed-appliqued; b, incised-punctated.

The two sherds with parallel grooves in the 41AG22 assemblage are from a Lindsey Grooved vessel. Lindsey Grooved is an Allen phase utility ware type comprised of large bowls or jars with direct or slightly everted rims. The rims have shallow horizontal grooves. Lindsey Grooved vessels also occur in conjunction with appliqued, brushed, incised, or punctated elements, typically placed either at the rim-body juncture or on the vessel body. Other utility ware sherds have incised, incised-punctated, and punctated elements on rim and body sherds from vessels of unknown types. One of the incised-punctated sherds has two closely-spaced horizontal incised lines and an associated row of triangular-shaped punctations (see Figure 2b).

The fine ware sherds in the 41AG22 assemblage include sherds from vessels with engraved (n=6) or trailed (n=1) lines. Three of the engraved sherds are from Patton Engraved vessels; such fine wares are one of the principal types in Historic Caddo Allen phase sites in East Texas. One of the Patton Engraved body sherds has upper and lower parallel rows of linear tick marks, and is from a Patton Engraved, var. Allen vessel (Perttula 2011:Figure 6-66a). The other two Patton Engraved sherds cannot be identified to a specific variety (Figure 3a-b). Another engraved rim sherd has both horizontal and vertical engraved lines (Figure 3c). Two other engraved sherds are from different carinated bowls. The first (Figure 3d) has horizontal engraved lines and a diagonal/cross-hatched zone between two of the horizontal engraved lines. The second carinated bowl sherd has opposed diagonal engraved lines and zones filled with diagonal hatched lines (Figure 3e). The parallel trailed sherd is likely from a post-A.D. 1680 Keno Trailed vessel, probably a bowl (see Suhm and Jelks 1962:Plate 44; Schambach and Miller 1984:123).


Figure 3. Selected decorative elements on fine ware sherds from 41AG22: a-b, Patton Engraved; c-e, other engraved rim and body sherds.

This analysis of the TARL collections from the site indicates that it was used first during the Woodland period, but that the principal use of the site was by Caddo peoples affiliated with Hasinai Caddo groups after ca. A.D. 1680 and up to ca. A.D. 1730, based on the recovery of primarily grog-tempered vessel sherds from known Allen phase ceramic types, including Patton Engraved, Lindsey Grooved, and Spradley Brushed-Incised, and one sherd of Keno Trailed. The specific cultural affiliation of the Caddo occupants of 41AG22 is not known, as almost all of the ethnographic and archival documents concern Caddo groups living near the Camino Real de los Tejas, not farther down the Neches River (see Perttula 1992:Figure 22). The Nacono are one possibility, as they lived below and to the south of the Camino Real, but they are only mentioned infrequently after ca. A.D. 1716 in archival documents, and may have “lost their separate ethnic and band identity” (Perttula 1992:220) by that time.


References Cited


Perttula, T. K.

1992    “The Caddo Nation”: Archaeological and Ethnohistoric Perspectives. University of Texas Press, Austin.


2011    The Ceramic Artifacts from the Lang Pasture Site (41AN38) and the Place of the Site within an Upper Neches River Basin Caddo Ceramic Tradition. In Archeological Investigations at the Lang Pasture Site (41AN38) in the Upper Neches River Basin of East Texas, assembled and edited by T. K. Perttula, D. B. Kelley, and R. A. Ricklis, pp. 145-320. Archeological Studies Program Report No. 129, Texas Department of Transportation, Environmental Affairs Division, Austin.


Perttula, T. K. and P. Marceaux

2018    The Lithic and Ceramic Artifacts from the Spradley Site (41NA206), Nacogdoches County, Texas. Special Publication No. 50. Friends of Northeast Texas Archaeology, Austin and Pittsburg.


Perttula, T. K. and K. Stingley

2017    Archaeological Investigations at the Walnut Branch (41CE47), Ross I (41CE485), and Ross II (41CE486) Sites, Cherokee County, Texas. Journal of Northeast Texas Archaeology 76:31-70.


Schambach, F. F. and J. E. Miller

1984    A Description and Analysis of the Ceramics. In Cedar Grove: An Interdisciplinary Investigation of a Late Caddo Farmstead in the Red River Valley, edited by N. L. Trubowitz, pp. 109-170. Research Series No. 23. Arkansas Archeological Survey, Fayetteville.


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.

Prehistoric Flintknapping Tool Research for Experimental Archeology by Christopher W. Ringstaff

Christopher W. Ringstaff is a visiting researcher from the Texas Department of Transportation (TxDOT). This article is part of the March 2018 TARL Newsletter.

For the past several months, I have been looking at prehistoric flint knapping implements in the collections at TARL. As an experimental archeologist, having the best understanding of prehistoric tool kits used in chipped-stone tool manufacturing improves the quality and resolution of my research. The following is a brief overview of this research discussing sites, tools, and replicated tools.

I examined collections from sites spanning three regions of Texas: Central Texas, the Lower Pecos, and the Coastal Plain. Sites studied included Gault (41BL323), Fate Bell (41VV74), Eagle Cave (41VV163), Morhiss Mound (41VT1), and the Crestmont Site (41WH39). The sites range in age from Early to Late Archaic. The items from Gault may be older but are from the Pearce Collection, an impressive but poorly provenienced collection made by one of the earliest investigators of the site. The site types from which these knapping implements were recovered are varied and include open campsites, rock shelters, and burial sites.

The knapping implements reviewed included antler billets for direct percussion, antler tines for pressure flaking, and antler segments commonly referred to as punches or drifts for indirect percussion. Although many of the specimens were utilitarian and were apparently discarded, in the case of Crestmont and Morhiss Mound, some items were found together as knapping kits interred with deceased individuals as grave goods. Notable are the knapping kits found with burials associated with Features 4 and 9 at the Crestmont Site, which included an array of well manufactured punches. Also remarkable was the knapping kit from Burial 119 at Morhiss Mound that included a wide range of knapping tools including billets, tines, and possible punches. All of the Burial 119 artifacts remain covered in ochre.

Figure 1. White-tail antler billet for soft-hammer direct percussion from Fate Bell Shelter (41VV74).
Figure 2. Experimental white-tail antler billet for soft-hammer direct percussion.

The attributes of these artifacts were used to model experimental knapping tools for replication and comparative use-wear. As an example, the size and weight of the antler billet from Fate Bell, among the larger ones thus far observed, was used to model an experimental billet as shown in Figures 1 and 2. Of particular interest during the course of this research were the punches used for indirect percussion. From the attribute data collected, replica punches were made and used in extensive experimentation. Initial results show consistency in use damage between the actual and experimental tools as shown in Figures 3 and 4. Results from this research should be completed in time to be presented at the 2018 TAS annual meeting.

Figure 3. Probable punch for indirect percussion from the Gault Site (41BL323). Note impact spall.
Figure 4. Replicated punch used for indirect percussion experiements. Note impact spall from use.

The Perkin Pike: A Unique Lithic Artifact from Southeast Texas by Wilson W. Crook, III

Wilson W. Crook, III is an Affiliated Researcher with TARL. This article is part of the March 2018 TARL Newsletter. 

During our recent work with the Andy Kyle Archeological Collection on behalf of the Sam Houston Regional Library and Research Center in Liberty, Texas, a number of unique bifacially flaked artifacts were observed from sites near the Trinity River in Liberty County (Crook et al. 2017). These artifacts were constructed from small, oblong cobbles that retained their original shape on the proximal end but had been bifacially flaked on the proximal end to form a sharp, tapered point. Similar tools have been recovered from sites within the McGee Bend (now Sam Rayburn) Reservoir in the mid-1960’s (Jelks 1965, 2017). Jelks named the artifacts “Perkin Pikes” but noted that this was a temporary designation as their true function was unknown. As a result, a study was undertaken using high power digital microscopy in an effort to observe use wear patterns which might help determine the tool’s true function.

The Perkin Pike

In a detailed study of 13 sites prior to the construction of the McGee Bend (Sam Rayburn) Reservoir, Ed Jelks noted the occurrence of 121 artifacts which he termed “Perkin Pikes”. The artifacts were constructed from relatively small (<100 mm) oblong river cobbles. The most common lithic material was chert but cobbles of other materials (quartzite, silicified wood) were also used. Many of these cobbles were rectangular in cross-section. In each case, the proximal end of the cobble was left unmodified while the distal end was bifacially flaked to form a tapered point. Dimensions for artifacts made from chert ranged from 28-72 mm in length, 19-41 mm in width, and 10-23 mm in thickness. Jelks did not make an attempt to determine the artifacts’ function, stating that the term “Pike” was simply used to describe the tools’ tapered point (Jelks 1965, 2017). Perkin Pike-like tools were also described by Tunnell (1961) from McGee Bend, by Wheat (1953) from Addicks Reservoir in Harris County, and by Jelks (1962) from Texarkana Reservoir.

Table 1. Physical Measurements of the Perkin Pikes from the Andy KyleArcheological Collection from sites in Liberty County, Texas. Average length:width ratio = 2.09. Average length:thickness ratio = 3.04.

Perkin Pikes from Liberty County

In our study of the 30,000+ piece Andy Kyle Archeological Collection, a total of 15 tools were identified which seemed to be identical to Jelks’ Perkin Pikes. The artifacts were found in collections from four sites in Liberty County including Wood Springs (41LB15), Savoy (41LB27), Knight’s Bayou (41LB61), and Moss Hill (41LB65). Eleven of the artifacts were constructed from chert cobbles, two from quartzite and two from silicified (petrified) wood (Table 1).

Size of the artifacts was fairly consistent with lengths ranging from 55.5 – 77.6 mm; widths from 28.8 – 38.0 mm; and thicknesses from 14.2 – 25.7 mm. Average tool dimension was 67.7 x 32.4 x 22.2 mm (see Table 1). Average length-to-width ratio is close to 2:1 and average length-to-thickness ratio is near 3:1. All of the artifacts had a rectangular cross-section with the cobble’s original cortex remaining on the proximal end to about one-third the length of the cobble. The distal end was bifacially flaked into a crude point. An example of a typical Perkin Pike from the Moss Hill site is shown in Figure 1.

Figure 1. Typical Perkin Pike typeartifact from the Moss Hill (41LB65)
site in Liberty County. Note the presence of cortex on the proximal end of the tool.

Seven of the artifacts were examined under high power (20-220x) using a Dino-Lite AM-4111-T digital microscope. Several of the specimens showed a high degree of polish on the proximal end of the artifact on the cobble cortex. This sheen is consistent with the tool having been hafted while in use (Keeley 1980). On the distal end, five of the artifacts showed some polish on the tip of the point. This polish was present only on the distal end of the point and not on any of the lateral edges extending back from the point toward the unmodified proximal end. Detailed examination showed the polish to extend around the circumference of the distal end, consistent with wear from a rotary motion on a material such as leather (Keeley 1980). It is therefore suggested that the Perkin Pikes examined from the Liberty County sites were used as a perforator type tool.

Conclusions and Discussion

The Perkin Pike is a minor but consistent component of the Late Prehistoric occupation in sites from Liberty County. The 15 artifacts observed from the Andy Kyle Archeological Collection are consistent in size and composition to those described by Jelks (1965, 2017) from sites in the Angelina River drainage now submerged beneath Sam Rayburn Reservoir. As noted by Jelks (2017), chert cobbles were the preferred lithic material for manufacturing Perkin Pikes although examples are known both from Sam Rayburn Reservoir and from the Kyle Collection from both silicified wood and quartzite. Workmanship is limited with only the distal end of the cobble being flaked to a bifacial point. Microscopic examination shows polish developed on the unmodified dorsal and ventral surfaces of the proximal end which is indicative of the tool having been hafted during use. Rotary polish was observed on the distal tip on several specimens from the Kyle Collection suggesting the tool was used on material such as leather. As such, the Perkin “Pike” might be better defined as the “Perkin Perforator” and used by the aboriginal inhabitants of Southeast Texas as a specialized tool for the manufacture of leather (probably deer skin)  goods.

References Cited

Crook, Wilson W., III, Robert J. Sewell, Linda C. Gorski and Louis F. Aulbach
2017 The Andy Kyle Archeological Collection. Report of the Houston Archeological Society, No. 29:13-56. Houston.

Jelks, Edward B.
1962 The Kyle Site: A Stratified Central Texas Aspect Site in Hill County, Texas. University of Texas Archeological Series No. 5.

1965 The Archeology of McGee Bend Reservoir, Texas. Unpublished Ph.D. Dissertation. Department of Anthropology, The University of Texas, Austin.

2017 The Archeology of Sam Rayburn Reservoir. CRHR Research Reports: Vol. 3, Article 1.

Keeley, Lawrence H.
1980 Experimental Determination of Stone tools Uses: A Microwear Analysis. University of Chicago Press, Chicago.

Tunnell. Curtis D.
1961 Evidence of a Late Archaic Horizon at Three Sites in the McGee Bend Reservoir, San Augustine County, Texas. Bulletin of the Texas Archeological Society 32:123-158.

Wheat, Joe Ben
1953 River Basin Survey Papers No. 4: Archeological Survey of the Addicks Dam Basin, Southeast Texas. Bureau of American Ethnology Bulletin 154. Washington, D. C.

Hunting Bannerstones at TARL by Gus Costa

Gus Costa is a visiting researcher from Moore Archeological Consulting, Inc. This article is a part of the March 2018 TARL Newsletter. 

Last year our team unearthed several curious bits of groundstone items in advance of road construction on Houston’s west side. One of these objects had been drilled all the way through; a defining characteristic of problematic, perforated, groundstone objects commonly known as “bannerstones.” As the leading “rock geek” of our office, I had the chance to help better determine what these broken bits represent, what they might mean, and why they’re important.

Descriptions of bannerstones go back to the Civil War era when they were regularly labeled as “problematicals”. Some authors called them whale-tails, thunder-bird emblems, ceremonial axes, maces, and butterfly stones. Charles Conrad Abbott coined the term “bannerstone” in 1876. Most early observers were stuck by the elaborateness of these soft and fragile stone items. For nearly half a century the academic consensus was that these items represented non-utilitarian, ceremonial items. In the 1920s bannerstones found alongside antler hooks and handles led to the popular inference that bannerstones were used as spear-thrower or atlatl weights. The debate continues today as to what function(s) these enigmatic artifacts might have served.

Figure 1. Fragmentary bannerstones from TARL’s collections (Photos B and C courtesy of Marybeth Tomka and Jason Barrett, scale is 3 cm). All are broken along the “barrel” or drilled perforation which is oriented perpendicular to the lateral wing-like elements shown here in plan view. Specimen A, is half of a butterfly type bannerstone from Flournoy Farm in Wood County. Butterfly bannerstones forms are uncommon in Texas but are more commonly reported in the northeastern region. Specimen B is half of a winged semi-lunar bannerstone from Morhiss (41VT1) in Victoria County. This specimen is notable as it represents the only published example of a broken bannerstone with a secondary perforation in Texas. Such extra-perforations are occasionally encountered elsewhere where they might represent mending activities or ornamental uses. Specimen C is a semi-lunar bannerstone from 41WA7 in Walker County.

Bannerstones appear during the Archaic Period between 6000-5000 B.C. in the Ohio River Valley and parts of the Mid-Atlantic. They proliferate during the Middle Archaic and are replaced in the Late Archaic (at approximately 1800 B.C.) by boatstones (another type of groundstone artifact) and gorgets. Bannerstones are scarcely known in the archeological record of Texas. Only a handful of bannerstones are present within the Texas Archeological Research Lab (TARL). Last year I visited TARL to examine some of these artifacts as part of a larger effort to characterize the Archaic bannerstones of Texas. At present I’ve documented nearly 40 bannerstones from all over Texas, primarily in the southeastern areas along the Gulf Coastal Plain. Bannerstones are fascinating, uniquely American artifacts that are rare, neglected and difficult to make. They can also be useful indicators of long distance exchange of materials and ideas. Petrographic analysis of the bannerstones found by the Moore Archeological team show that an exotic source of granite, hundreds of miles away from Houston, was exploited by Archaic people.