Archaeological Investigations at the Walnut Branch (41CE47), Ross I (41CE485), and Ross II (41CE486) Sites, Cherokee County, Texas

Cite this Record Perttula, Timothy K. and Stingley, Kevin (2017) "Archaeological Investigations at the Walnut Branch (41CE47), Ross I (41CE485), and Ross II (41CE486) Sites, Cherokee County, Texas," Index of Texas Archaeology: Open Access Gray Literature from the Lone Star State: Vol. 2017, Article 48. https://doi.org/ 10.21112/.ita.2017.1.48 ISSN: 2475-9333 Available at: https://scholarworks.sfasu.edu/ita/vol2017/iss1/48


___________________________________________________________________________
The majority of the sherds are from vessels tempered with grog and crushed hematite (47 percent) and grog (34 percent) ( Table 2). In total, more than 95 percent of the sherds have grog as a principal temper in vessel manufacture. Sherds from vessels tempered solely with burned bone comprise only 2.7 percent of the Kegley assemblage, but another 10.8 percent have burned bone aplastics in combination with hematite, grog, or grog and hematite. About 5.4 percent of the sherds are from vessels with a sandy paste. ___________________________________________________________________________ About 87 percent of the utility wares are from vessels with brushed or brushed-incised decorative elements (Table 3). These are likely from Bullard Brushed jars (Suhm and Jelks 1962:Plate 11); the thickness of some of the recovered sherds (> 10 mm) suggests that in some cases large jars were manufactured and used at the Walnut Branch site. The few incised and pinched sherds are from Maydelle Incised and Killough Pinched vessels (see Suhm and Jelks 1962:Plates 46 and 52). The typological ___________________________________________________________________________ Fine ware sherds are not common in the Kegley sherd assemblage from the Walnut Branch site, accounting for only 8.8 percent of the assemblage (see Table 3). Two of the engraved sherds have curvilinear engraved lines, and they may be from Poynor Engraved vessels (see Suhm and Jelks 1962:Plate 62). The trailed sherd with opposed curvilinear trailed lines is from a Keno Trailed bowl (see Suhm and Jelks 1962:Plate 44f). The Keno Trailed vessel sherd suggests that the Walnut Branch site was used to some extent after ca. A.D. 1680, but the other sherds perhaps point to a principal use of the site sometime during the Late Caddo period Frankston phase (ca. A.D. 1400-1680). We will return to the question of the estimated age of the Walnut Branch site, as well as the Ross I and Ross II sites, after discussing the character of the artifacts recovered from these sites in the 2017 archaeological investigations.

Surface Collection of the Walnut Branch stream bed from the Walnut Branch site and the Ross I site
The surface collection of the Walnut Branch stream bed was done in February 2017, before the Walnut Branch site to the Ross I site to reconnoiter for surface artifacts eroded from the cut bank (see Figure 2); at that time, it wasa thought that this was one site area, not two, so the surface collection artifacts were not separated by location along the stream bed. Visibility along the cut bank itself was poor because of vegetation growing along it, and no artifacts were noted in the cut bank.
The 2017 surface collection recovered 136 Caddo ceramic vessel sherds, 55 plain and 81 with lithic debris and part of the polished bit from a greenish-gray siliceous shale celt.

Walnut Branch Site (41CE47)
Intensive shovel testing was completed at the Walnut Branch site. This established that the site is approximately 110 x 100 m in size, or ca. 2.7 acres (Figure 3). A total of 72 shovel tests on the landform have cultural materials (Table 4) (Table 5). The sherdlet/sherd size index is 0.28. The mean density of sherds is 6.7 sherds per positive shovel test, or ca. 53.6 sherds per square meter of archaeological deposits. The range of sherds by shovel test is 1-26.     A range of other artifacts were recovered in the shovel tests at the Walnut Branch site, particularly burned clay pieces, wood charcoal, lithic debris, ground stone tools, and ceramic elbow pipe sherds (Table 6). Notably, the one lead ball found at the site came from a shovel test just north of Cluster B (see Figure 4). The three ceramic pipe sherds are from Clusters A and B, as are the two chipped stone tools; the ground stone tools are from Cluster B.

Ross I (41CE485)
At the Ross I site, shovel testing determined that the site covers a ca. 104 x 60 m area, or ca. 1.5 acres ( Figure 5). A total of 31 positive shovel tests were excavated on the site (Table 7), or ca. 20.7 80 cm in thickness overlying a clay B-horizon in the deeper shovel tests; the sediments were very moist. One shovel test (ST 179) encountered a probable feature between 50-60 cm bs: a 10 cm thick lens of charcoal and ash.      There are a few other kinds of artifacts recovered in the shovel testing at the Ross I site (Table 9), principally burned clay (possibly remnants of hearths or earth ovens) and animal bones. The one ceramic elbow pipe sherd is in Cluster A, and animal bone primarily occurs in Clusters A and B, or not far from Cluster A, as does nutshell. Most of the burned clay pieces also are present in one or another of the clusters, except for the burned clay from ST 154 and ST 158 at the northern end of the site (see Figure 5); wood charcoal pieces also are present in this part of the site.

Ross II (41CE486)
Shovel testing at the Ross II site determined that the site is ca. 106 x 76 m in size, or ca. 2.0 acres ( Figure 7); a small 14 x 16 m area at the southeastern end of the site is referred to as Area A. A total of 36 shovel tests at the site had cultural materials (Table 10), which amounts to 18.0 shovel tests per acre of from 34-100+ cm bs. Where encountered, the clay B-horizon is either a strong brown or reddish-yellow color. No apparent cultural features were encountered in any of the shovel tests at the Ross II site.

___________________________________________________________________________
The shovel tests at the Ross II site recovered 237 ceramic vessel sherds and 134 sherdlets (Table 11); the sherdlet/sherd index is 0.57, much higher than at the Ross I site. The mean sherd density is 6.6 sherds per positive shovel test, with a range of 1-17 sherds per shovel test, or ca 52.8 sherds per square meter of archaeological deposits at the site.

A wide range of archaeological materials were recovered in the surface collections and intensive shovel testing at the Walnut Branch (41CE47), John Ross I (41CE485), and John Ross II (41CE486) sites.
Not counting the several hundred sherdlets (<1.5 cm in diameter), this includes 1068 plain and decorated ceramic vessel sherds (representing 91 percent of the artifact assemblage), ceramic elbow pipe sherds (n=5), clay coils (n=2), burned clay pieces (n=30), chipped stone tools (n=2), lithic debris (n=21), ground stone tools (n=7), early 18 th century lead balls (n=2), wood charcoal (n=21), charred nutshells (n=5), and animal bone (n=13). Each of these categories of recovered archaeological materials are discussed below, beginning with the ancestral Caddo ceramic vessel sherds.

Ceramic vessel sherds
A total of 1068 ceramic vessel sherds have been collected in the 2017 archaeological investigations at the Walnut Branch (41CE47), John Ross I (41CE485), and John Ross II (41CE486) sites along Walnut Branch, including ceramic vessel sherds found on the surface of the Walnut Branch stream bed. This includes 477 plain rim, body, and base sherds, and 591 decorated rim and body sherds (Table 13). The composite plain to decorated sherd ratio (P/DR) for the assemblage is 0.81, but the P/DR ranges from 0.68 for the surface collection materials, 0.71 at the Walnut Branch site, 0.87 at the John Ross I site, and 1.06 at the John Ross II site. The P/DR values at these sites are considerably higher than they are at contemporaneous Allen phase sites in the Neche cluster on Bowles Creek (Perttula 2017a), where P/DR values are less than 0.40; we will return to this topic in the concluding section of the article.  (Table 14). About 80 percent of the decorated sherds in the surface collection are from utility wares, primarily from Bullard Brushed sherds; the incised and incised-punctated sherds are likely from Maydelle Incised vessels (see Suhm and Jelks 1962). Two notable sherds in the surface collection are a Lindsey Grooved body sherd (Figure 10a), and a Patton Engraved rim sherd (Figures 11 and 12a). The decorative 2011: Figure 6-66a-d), so it has been dubbed Patton Engraved, var. Walnut Branch. A Patton Engraved, body sherd, with a straight engraved line with tick marks, is also present in the surface collection. There is also a bottle sherd with opposed curvilinear engraved lines (Figure 12b).      Figure 13. Selected decorative elements on utility ware sherds from the Walnut Branch site. Table 15) as well as horizontal-diagonal and opposed incised lines (see Table 15). They are from the bodies of Maydelle Incised vessels. About 4 percent of the decorated sherds from the site have one or more tool punctated rows, and these sherds are from utility ware jars (see Suhm and Jelks 1962:Plate 79); these punctated jars have not been assigned a type name.

Incised sherds at the Walnut Branch site have parallel and straight line elements (see
Less common utility ware sherds from the Walnut Branch site are from Lindsey Grooved vessels and Killough Pinched jars (see Table 15; see Suhm and Jelks 1962:Plate 46); together sherds from these types comprise only 3.6 percent of the decorated sherds in the assemblage. Other rare utility wares from the site have appliqued, brushed-appliqued, and brushed-incised-appliqued decorative elements (see Table 15). Table 15). About 40 percent of these sherds are from Patton Engraved vessels that have engraved lines with either linear or excised tick marks (Figure 14). These are from Patton Engraved, var. Allen, var. Freeman, and var. Fair vessels (see Perttula 2011: Figure 6-66a, c-d). One sherd with rectilinear engraved lines is from a Mayhew Rectilinear vessel (see Figure 9, and see Jackson et al. Table 15). Finally, one sherd from a globular bowl has horizontaldiagonal engraved lines on the upper part of the vessel, while the body has vertical brushing marks. (Table 16) Table 16). 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 (Marceaux 2011:140 and Figure 5.2).

One brushed-incised sherd from the John Ross I site is from a Spradley Brushed-Incised vessel (see
The sherds with grooved decorative elements at the John Ross I site are from Lindsey Grooved vessels (see Figure 10b), while the incised and incised-punctated sherds are from Maydelle Incised jars (see Table 16). The one tool punctated body sherd comprises only 0.9 percent of the decorated sherds from the site. Figure 15b). This sherd may be from either a Patton Engraved, var. Allen, var. Freeman, or var. Fair vessel (see Perttula 2011: Figure 6-66a, c-d). (Table 17). Another 18 percent of the sherds have incised and incised-punctated decorative elements, and they are likely from Maydelle Incised vessels. Other utility ware sherds in the assemblage have appliqued ridges (n=2), including a rim sherd that may be from a Cass Appliqued vessel (see Suhm and Jelks 1962:Plate 13), or parallel or straight grooves (n=3); these latter sherds are from Lindsey Grooved jars.   Figure 6-66e, g).

Temper and Paste Comparisons between the Three sites on Walnut Branch
Three different temper inclusions are present in the ancestral Caddo ceramic vessel sherds from the sites along Walnut Branch: grog, burned bone, and crushed hematite, either as the sole temper or in combination with one or more of the temper classes (Table 18). These tempers were used in varying proportions by ceramic ware, as well as from one site to another. While many of the ceramic vessel vessels with a sandy paste, indicating the use of a natural sandy clay. This data makes clear that more than one source of clay was used by the potters living along Walnut Branch. Clays used for vessel manufacture were probably gathered from nearby alluvial settings, but almost certainly within a short (1-7 km away, at most) distance from the Caddo settlements (e.g., Arnold 2000:343;Arthur 2006:52), so that an inordinate amount of time and energy was not expended by potters in hauling clay back to the sites. Arthur (2006:52) points out that potters would be likely to select lower quality clays for vessel manufacture than high quality clays if the latter were farther away.     Table 18). The use of grog temper is highest in the plain wares at the Walnut Branch site, ware sherds at all three sites. Hematite temper is well represented in the coarser paste plain wares and utility wares (between 33.5-54.5 percent of all these sherds have hematite pieces in the temper), but is most prevalent in the plain ware sherds at the John Ross II site and the utility ware sherds at the John II sites.
The use of sandy paste clays is common in each of the three wares at the three Walnut Branch sites (see Table 18). Among the plain wares, sherds from sandy paste vessels are virtually equally common at the Walnut Branch and John Ross I sites (42.6-43.0 percent). The use of sandy clays in the utility wares at both sites is even higher: 46.5-58.3 percent; only 21.5-34.5 percent of the plain wares at the John Ross percent). Table 19 presents summary data (the combined totals for all three wares at each site) on temper use and sandy clay use in the paste of the ceramic vessel sherds at the three Walnut Branch sites. More than 94 percent of the sherds at each of the sites have grog temper, followed by between 39.4-48.5 percent hematite temper use, and then bone temper use from 13.1-19.0 percent by site. The differences between That is not the case with respect to the use of sandy clays to manufacture ceramic vessels at the Walnut Branch, John Ross I, and John Ross II sites. Between 45.4-50.2 percent of the sherds at the Walnut Branch and John Ross I sites are from vessels manufactured with a naturally occurring sandy clay. Only 29.2 percent of the John Ross II sherds are from vessels made with a sandy clay (see Table  19). This is particularly the case in the plain wares from the site (see Table 18).    Figure 18a). The John Ross II pipe has a sandy paste and is tempered with grog.

Clay coils
The two clay coil pieces found in the shovel testing at the Walnut Branch site are direct evidence for bs) is 20.0+ mm in length and 8.8 mm in thickness, while the second coil piece (ST 105, 20-40 cm bs) is 8.0 mm in diameter.

Burned clay pieces
The few small burned clay pieces (n=30 from the three sites) are likely remnants of clay hearths or earth ovens that were in use during the ancestral Caddo occupations of the sites, or the product of the from either the John Ross I site (n=13, 43 percent) or the Walnut Branch site (n=10, 33 percent).

Chipped stone tools
Only two chipped stone tools were recovered in the shovel testing at the sites along Walnut Branch,

Lithic Debris
Nineteen pieces of lithic debris were recovered in the shovel testing from the Walnut Branch (n=9) and John Ross II (n=10) sites (Table 20); no evidence for the manufacture or maintenance of chipped stone tools were found at the John Ross site I site (41CE485). The majority of the lithic debris are on chert raw materials from non-local sources-67 percent at the Walnut Branch site and 70 percent at the John Ross II site-likely from Central Texas Edwards Plateau sources. Cobbles and pebbles of these cherts can be found in outwash gravels in Neches River gravel sources (see Girard 1995:66;Shafer 1973). The translucent honey-colored chert is also called "beeswax" chert (Miller 2008:27). The only local raw material in the from the Walnut Branch stream bed during the February 2017 surface collection.  Table 20), either stream-rolled or roughened. It is probable that cortex-covered pebbles were minimally reduced at

Ground Stone tools and Red Ochre
Branch and John Ross II sites. From the Walnut Branch site, one of these tools is a polished stream-rolled pebble from ST 53 (0-20 cm bs); this pebble was likely used to burnish and polish the surface of ceramic vessels made at the site. There also is a ferruginous sandstone mano fragment (41+ x 34+ mm in length and width) from ST 97 (20-40 cm bs), and a pitted stone from ST 89 (10 cm bs). This ground stone tool is also made from local ferruginous sandstone. It is 70 x 56 x 37 mm in length, width, and thickness, with a 16 mm diameter pit or depression on one surface of the rock; the depression is 2 mm in depth. The on it, suggesting it was used as a source of pigment.
One of the ground stone tools at the John Ross II site is a ferruginous sandstone pitted stone (ST 163, 17 cm bs). One surface of the cobble has a 5.8 mm diameter pit or depression; the pit is 8.2 mm in depth. mano fragment was recovered in ST 190 (0-20 cm bs) at the site.
The ground stone tools from the three sites also include a greenstone celt polished bit fragment from the surface collection along the Walnut Branch stream bed (Figure 19). The greenstone or siliceous shale raw material was likely obtained from Red River gravels, although its source was the Ouachita Mountains in southeastern Oklahoma.

Lead Balls
Two small lead balls of likely early 18 th century age and French manufacture (see Jackson

Wood Charcoal and Nutshells
Pieces of wood charcoal and nutshell are not abundant at any of the three sites on Walnut Branch, sites. There are 21 pieces of wood charcoal, 71 percent of those pieces being recovered from the Walnut are from the John Ross I site (41CE485). One piece of charred nutshell from the Walnut Branch site was submitted for radiocarbon dating (see below).

Animal Bone
Animal bone is also not well preserved at the three sites on Walnut Branch. Of the 13 pieces recovered in the shovel testing, 61.5 percent are from the John Ross I site (41CE485) archaeological deposits. In addition to the abundant ceramic vessel sherds, the shovel testing recovered ceramic elbow pipe sherds, chipped stone tools, a single ground stone celt and other ground stone tools, lithic debris, and 18 th century lead balls at the Walnut Branch and John Ross II sites. The latter are testimony to Caddo contact with Europeans in the area. Other recovered artifacts include a few pieces of burned clay, wood charcoal, nutshells, and burned and unburned animal bone.

Summary and Conclusions
The sites range in size from ca. 1.5 acres (John Ross I), 2.0 acres (John Ross II) to 2.7 acres (Walnut Branch site).The densities of ceramic vessel sherds at the three sites, in combination with other artifact artifact clusters at each of the three sites: two clusters at the Walnut Branch site, three clusters at the John Ross I site, and two artifact clusters at the John Ross II site. These clusters cover between ca. 900-980 square meters at the three sites, and are likely evidence for 2-3 ancestral Caddo household compounds at each site, along with open areas (informal plazas), and extramural work areas. Remote sensing work is planned in the spring of 2018 around these artifact clusters/household compounds at the three sites along Walnut Branch, along with controlled metal detector work to recover European metal artifacts. ethnographic and archaeological information on the accumulation of utility ware sherds (i.e., what Varien [1999] calls cooking pot sherds) in domestic contexts suggests that 4000-8000 grams of cooking pot sherds would be accumulated through breakage and use per year on a residential site occupied by Native meter in those artifact clusters (>80 sherds per square meter), and the proportion of utility ware sherds in the assemblages (46-51 percent of the recovered sherds), we estimate that there are approximately 180,000 grams (36,000 sherds, weighing on average 5.0 grams each) of utility ware sherds at the John Ross I site, 184,000 grams (ca. 36,800 sherds) at the John Ross II site, and ca. 196,000 grams (ca. ca. 39,200 sherds) at the Walnut Branch site. These values suggest the utility ware sherds at the three sites would have accumulated in about 22.5-45 years (John Ross I), 23-46 years (John Ross II), and 24.5-49 years (Walnut Branch).
Given that Caddo wood structures would probably only last about 20 years or so before they began to deteriorate and needed to be replaced (see Good 1982:69), then the utility ware accumulation data from the Walnut Branch sites suggests that the artifact clusters were created from ceramic debris deposited either by: (1) one house per cluster or farm compound that may have been used and rebuilt; (2) one or two houses that were in use contemporaneously, each farm compound contributing to the accumulation of ceramic sherd debris; or (3) one or two houses in a farm compound that were occupied at separate times, but close together in time, each contributing at different times to the trash deposits. or stylistic difference in the ceramic assemblages, either in the use of grog or bone temper, in the relative proportion of plain to decorated sherds, brushed to plain sherds, or brushed to other wet paste sherds (Table 21). As such, the three sites constitute a related group of similar ceramic assemblages made by Caddo potters at about the same time that shared a common tradition of ceramic practice. The Walnut Branch sites appear to be related to the Allen phase components previously recognized as belonging to the Neche cluster (Perttula 2017a). The Neche cluster of ceramic vessel sherd assemblages includes several Allen phase Historic Caddo sites on Bowles Creek and the Neches River (41CE291) as well as one Late Frankston phase (ca. A.D. 1560-1680) investigated component of the George C. Davis site (41CE19) (Fields and Thurmond 1980) on the northern part of the terrace east of the Neches River, and a Middle Caddo component at 41CE289 (Figure 21). The Allen phase components in Groups I and II have high proportions of brushed sherds and ratios of brushed to other wet paste sherds (Table 22). These assemblages are almost exclusively comprised of grog-tempered vessels, but differences between the sites in the proportion of bone-tempered vessels (either as the sole temper or in combination with grog) suggest that two contemporaneous groups of Allen phase sites are present in the Neche cluster. These two groups (I and II) also are notably different in brushed to plain sherd ratios (Table 22).