Timothy K. Perttula and Kevin Stingley are visiting researchers at TARL. This article is part of the September 2017 TARL newsletter.

During the summer of 1969 while doing fieldwork at the George C. Davis site (41CE19), Dr. Dee Ann Story sent out two of her students, Dan Witter and George Kegley, to survey sites in areas to the north of the Davis site (Story 1997). One of the sites recorded by Kegley and Witter was 41CE47. In the spring of 2017 Stingley revisited the site and recorded two new adjacent sites. This article will describe the work conducted at the sites and the range of artifacts found there.

During Stingley’s initial survey of the site area the landowner was able to point out the location of two shovel tests excavated by Kegley when he recorded the site; four ceramic sherds were found in these two shovel tests. Also pointed out was a section of Walnut Branch, a small tributary stream of Box’s Creek, where Kegley found 70 Caddo ceramic sherds; these sherds are now housed at the Texas Archeological Research Laboratory. Kegley noted in his site report that 41CE47 was in danger of eroding into the creek bed. However, around 1990 the creek changed course bypassing an oxbow that was dangerously close to that site.

The landowner noted several areas in the Walnut Branch floodplain where he had found a few surface artifacts, but said he did not know of any shovel testing ever having been done there. Before starting any shovel testing Stingley walked a 400-meter section of this east-west flowing creek immediately south of 41CE47. In the gravel bed he found 136 Caddo ceramic sherds; 55 plain and 81 decorated vessel sherds. A few lithic artifacts included pieces of petrified wood lithic debris and part of the polished bit from a greenish-gray siliceous shale celt.

Stingley began shovel testing at the westernmost end of a large field and floodplain that stretched 700 meters east-west and between 60-250 meters north-south. The western area was recorded as 41CE485. A total of 31 positive shovel tests were completed there that had 217 ceramic sherds, split almost evenly between plain and decorated sherds. Seven of the shovel tests in three spatial clusters contained between 10-30 sherds and one had a lens of ash and charcoal indicating a possible pit/hearth feature. Other artifacts recovered from this area included 13 pieces of burned clay, a pipe sherd, four charred nutshells, animal bone, and wood charcoal. A high water table limited the depth of shovel tests with most only reaching 40-50 cmbs.

Next, intensive shovel testing was done at the Walnut Branch site (41CE47), the original location recorded by Kegley. This work determined that the site was ca. 110 x 100 m in size. Seventy-two shovel tests returned cultural materials in the fine sandy loam of the Walnut Branch floodplain; again, a high water table limited the depth of shovel tests and the clay B-horizon was never reached in this area. A total of 480 ceramic sherds were found in the shovel tests, ranging from 1-26 sherds. Two different spatial clusters at the Walnut Branch site contained high sherd concentrations. These spatial artifact clusters likely represent at least 2-3 Caddo household compounds. Found in these areas were burned clay, wood charcoal, nutshell, animal bone, a chipped stone tool, a grinding stone, red ochre, a polished pebble, and a small lead ball (Figure 1a).

Immediately northeast of the Walnut Branch site was a slight elevated area of ca. 2.0 acres where site 41CE486 was identified and recorded. A total of 36 shovel tests contained cultural materials. Because of its slight elevation above the floodplain several shovel tests reached the clay B-horizon. The A-horizon sediments were strong brown to dark brown fine sandy loam. The shovel tests at 41CE486 recovered 237 ceramic sherds, again almost evenly split between plain and decorated sherds. The density of sherds from these 36 shovel tests point to two areas within the site that likely represent one or two household compounds and a plaza between them. Also found in these areas were an elbow pipe sherd, burned clay, wood charcoal, animal bone, lithic debris, a ground stone tool, and another small lead ball (see Figure 1b) similar in size to the lead ball found approximately 125 meters away at the Walnut Branch site.

A wide range of archeological material was found at the three sites in the shovel testing. Of the 177 total shovel tests excavated by Stingley, 79% were positive. More than 205 ceramic vessel sherds were found in the gravel bed of Walnut Branch by Kegley and Stingley combined. In total 1068 ceramic vessel sherds, five ceramic elbow pipe sherds, two clay coils, 30 pieces of burned clay, two chipped stone tools (including a Turney arrow point), 21 pieces of lithic debris, seven ground stone tools, two early 18th century lead balls, 21 pieces of wood charcoal, five charred nutshells, and 13 small pieces of animal bone were recovered. The ceramic sherds were from vessels that were predominately grog-tempered followed by sherds from vessels with hematite temper, and 19% or less had bone temper. Wood charcoal, nutshell, and animal bone are not abundant at any of the three sites indicating the poor preservation of organic remains in the moist fine sandy loam of the sites. Caddo ceramic vessel sherds from the following types were found at one or more of the sites: Patton Engraved (Figure 2) is the most common fine ware; along with only a very few Poynor Engraved and Mayhew Rectilinear sherds, Bullard Brushed and Maydelle Incised jars and sherds of Lindsey Grooved and Killough Pinched are in the assemblages at each site.

The artifacts from the sites indicate they were occupied mainly during the Historic Caddo Allen phase (post A.D.1680), with some limited use during the Late Caddo Frankston phase particularly at 41CE485. The archeological evidence also suggests that these sites were part of a Caddo Neche cluster that includes sites from the
Bowles Creek area, along the middle Neches River, and from a late occupation at the George C. Davis site. The two 18th century lead balls found at 41CE47 and 41CE486 suggest that Caddo peoples were in contact with Europeans in the area. Further work at the sites is planned, including remote sensing and the excavation of 1 x 1 m units in the areas of the different artifact clusters, hoping to identify cultural features from Caddo houses and pit features in outdoor activity areas.

References Cited:

Story, D.A.
1997 1968-1970 Archeological Investigations at the George C. Davis Site, Cherokee County, Texas. Bulletin of the Texas Archeological Society 68:1-113.