The Pearson Site (41RA5) at Lake Tawakoni on the Sabine River, The Pearson Site (41RA5) at Lake Tawakoni on the Sabine River, Rains County, Texas Rains County, Texas

The Pearson site (41RA5) in the Blackland Prairie of East Texas is one of a number of aboriginal archaeological sites recorded during a 1957 archaeological survey of the flood pool of then proposed Lake Tawakoni on the Sabine River; the site is now inundated. The Pearson site was located on several low sandy rises across ca. 25 acres in the Hooker Creek-Sabine River floodplain, and these rises had both aboriginal and European artifacts on the surface. Johnson and Jelks, and Duffield and Jelks have argued that the Pearson site was the Tawakoni-Yscani village visited by a Spanish missionary in 1760 and part of a recently defined Norteno focus, a complex of sites apparently associated with the Wichita tribes. Schambach, by contrast, considers the Pearson site to be an 18th century Tunican entrepot.


Introduction
The Pearson site (41RA5) in the Blackland Prairie of East Texas is one of a number of aboriginal archaeological sites recorded during a 1957 archaeological survey of the flood pool of then proposed Lake Tawakoni on the Sabine River ( Figure 1) (Johnson 1957); the site is now inundated. The Pearson site was located on several low sandy rises across ca. 25 acres in the Hooker Creek-Sabine River floodplain, and these rises had both aboriginal and European artifacts on the surface. Johnson and Jelks (1958) and Duffield and Jelks (1961) have argued that the Pearson site was the Tawakoni-Yscani village visited by a Spanish missionary in 1760 and part of a recently defined Norteno focus, a complex of sites apparently associated with the Wichita tribes. Schambach (1996), by contrast, considers the Pearson site to be an 18 th century Tunican entrepot.

Ceramic Vessel Sherds
The ceramic vessel sherds in the Pearson site assemblage are from plain ware, utility ware, and fine ware vessels (Table 2). Fifty percent of the rim sherds and 74 percent of the sherds are from plain ware, and there are 54 decorated sherds from utility ware and fine ware vessels; 53.7 percent of the decorated sherds are from utility wares, and the remaining 46.3 percent are from fine ware vessels. The plain to decorated sherd ratio is 2.87. At the nearby Gilbert site, more than 61 percent of the decorated sherds (n=719) are from fine wares (Story et al. 1967:Tables 7 and 8; Perttula 2012: Table 3).  (Table 3). These sherds are quite similar to Edwards Plain as defined by Baugh (2014:7-12). At the nearby and roughly contemporaneous Gilbert site (41RA13), about 9 percent of a recently analyzed sample of ceramic vessel sherds have a sandy paste (Perttula 2012), and Story et al. (1967:Table 5) indicates that 5.6 percent of a larger sample of sherds from the Gilbert site are plain sandy paste wares. Baugh (2014) considers Edwards Plain to be a Kirikir'is or Wichita tribal pottery type that was made between ca. A.D. 1450-1725; its presence at both the Gilbert and Pearson sites suggests that this pottery continued to be manufactured until at least the 1740s-1760s, if not later.

___________________________________________________________________________
The other ceramic sherds are from vessels tempered with several different combinations of tempers, including burned bone, grog, hematite, and burned mussel shell (Table 3), the latter temper very rare in the assemblage. The most common temper is grog, comprising 38.2 percent of the assemblage, and sherds with grog and other temper inclusions account for 63.1 percent of the site sample; the highest percentage of grog-tempered sherds occur in the fine ware. Sherds with vessels that have hematite temper, either as the sole temper, or in combination with grog and/or bone, are present in 24.4 percent of the sherds. The highest proportion of sherds with hematite temper occurs in the plain wares (see Table 3). Sherds from vessels tempered with burned bone are most common in the utility wares (44.8 percent).
The utility wares from the Pearson site are most commonly represented by sherds with incised (51.7 percent of the utility wares) or brushed (31.0 percent) decorative elements, while 84 percent of the fine wares have engraved decorative elements (Table 4). The decorated sherds are uniformly from tempered wares.

__________________________________________________________________________
The one appliqued body sherd has a single appliqued node (Figure 2a; see also Table 4), and may be from an Emory Punctated-Incised jar (see Story et al. 1967:144). The brushed and brushedincised sherds, comprising 41 percent of the utility wares, may also be from Emory Punctated-Incised vessels as Story et al. (1967:137) indicate that when brushed elements are present, they "may occur almost anywhere on the vessel, from the lip to the base." At the Pearson site, the one brushed rim has horizontal brushed marks, while the body sherds have opposed and parallel brushed marks.
The incised sherds (52 percent of the utility wares), all body sherds, have curvilinear or geometric lines (see Figure 2b and Table 4). These decorative elements are also consistent with Emory Punctated-Incised vessels, in that incised lines consist "of straight to slightly curved lines extending from below the rim to the base or to about the middle of the body" (Story et al. 1967:137). The one incised-punctated sherd, also from an Emory Punctated-Incised vessel (Story et al. 1967:137) in the assemblage, has a row of circular punctations, likely at or near the rimbody juncture of a jar, with vertical incised lines extending down the vessel body.
The 21 engraved fine ware sherds in the Pearson site collection (see Table 4) include a number from Womack Engraved vessels with cross-hatched (n=6) engraved zones, including one with a triangular zone  (Figures 3f and 4b) and one body sherd with diagonal-parallel engraved lines with tick marks may be from Variety B of Womack Engraved. Two other rim sherds, not identifiable to a type, have either a horizontal engraved line or horizontal-vertical-curvilinear engraved lines (Figures 3d and 4c). The other engraved sherds in the small assemblage of fine wares at the Pearson site can be grouped into a miscellaneous category: they include horizontal and diagonal opposed engraved lines; opposed engraved lines; parallel engraved lines; straight engraved line; and vertical-diagonal-curvilinear engraved lines.   Other fine wares at the Pearson site are either red-slipped (8 percent of the fine wares) on their exterior surfaces or have curvilinear or parallel trailed lines (8 percent of the fine wares, see Table 4). Slipped wares are described as "quite rare" at the Gilbert site (Story et al. 1967:112). The two trailed body sherds (Figure 5a

Selected Chipped Stone Tools
There are a number of dart points and arrow points in the Pearson site collection, reflecting a lengthy, if not intermittent, use of the site since ca. 5000 years B.P., during the Late Archaic, into the 18 th century A.D. The identified types in the collection follow Turner et al. (2011).

Dart Points
Of the more or less complete dart points in the Pearson site assemblage (n=51), 75 percent are identified to eight different defined types (Table 5), principally including Gary (n=10, 19.6 percent), Trinity (n=9, 17.6 percent), and Edgewood (n=7, 13.7 percent) (Figures 6 and 7). Thirteen of the dart points, all considered to be Late Archaic period in age, are categorized by the stem shape-i.e., expanding stem or straight stem-along with base shape and the presence of barbed shoulders-but have not been identified to a specific type.          There are two Woodland period point types dating from ca. 2500-1150 years B.P.) in the Pearson dart point assemblage: Edgewood (n=7) (see Figure 6d) and Gary (n=10) points (see Figure 7 and Table 5). The thickness and stem width data for the Gary points suggest that four are var. Camden (see Schambach 1982;Leith 2014), dating after ca. A.D. 400, and three others are var. LeFlore (ca. A.D. 1-400). The mean thickness of the var. LeFlore points is 7.7 mm and their mean stem width is 16.0 mm. Conversely, the mean thickness and stem width of the var. Camden points at the Pearson site is 6.73 mm and 13.13 mm, respectively.
In addition to the more or less complete dart points that can be identified to a defined type, there are 18 dart point fragments in the collection from the Pearson site. They are on quartzite (n=11, 61 percent), gray chert (n=3, 16.7 percent), coarse-grained grayish-white quartzite (n=2, 11.1 percent), reddish-gray chert (n=1, 5.6 percent), and very dark gray chert (n=1, 5.6 percent). The coarse-grained grayish-white quartzite originates at a quarry on Mill Creek in Van Zandt County in the upper Sabine River basin. According to Malone (1972:32), the quarry covers between 3-4 acres of an upland landform, manifested as outcrops of large boulders (Malone 1972: Figure 13) with "partially decorticated cores, as well as numerous flakes." This distinctive coarse-grained quartzite is present in the Pearson site assemblage, accounting for 4.8 percent of the Late Archaic dart points and 23.5 percent of the Woodland period dart points ( Table  6; see also Table 5); this material is represented on 15.4 percent of the unidentified expanding stem and straight stem dart points. The use of non-local cherts for manufactured dart points is much most common during Late Archaic period times, as is the case for local cherts, and novaculite, the latter likely obtained from Red River gravels well north of the Pearson site. Local fine-grained quartzite was primarily used for Woodland period dart points at the site, while local petrified wood was equally common in both the Woodland and Late Archaic period dart points. Almost 70 percent of the Gary dart points from a studied site (41VN63) in the upper Sabine River basin are made on the gray-white coarse-grained quartzite (see Perttula et al. 2017); 70 percent of the dart point tips are also made from this material. At the Yarbrough site (41VN6, see Johnson 1962) a few miles east of 41VN63, about 4 percent of the dart points are made from this distinctive quartzite, mostly having been heat-treated. This includes 5.2 percent of the Gary points (n=328), 2.1 percent of the Wells points (n=47), 5.3 percent of the Marshall points (n=19), 15.2 percent of the Ellis points (n=33), 6.0 percent of the Edgewood points (n=50), and 20.0 percent of the Wesley points (n=5); none of the arrow points (n=9) were made from this white/pink quartzite. The age of the dart points made from this quartzite certainly suggest that this material was primarily utilized in Late Archaic and Woodland period times, as is the case at 41VN63 and the Pearson site. At another site (unrecorded) on the Robert L. Richey property on the Sabine River a few miles west of 41VN63, the grayish-white quartzite raw material occurs on 61.5 percent of the small lithic debris sample (n=8). The few diagnostic artifacts from this site are decorated ceramic sherds that likely date from ca. A.D. 900-1200, indicating the continued use of this quartzite raw material in ancestral Caddo times (see Perttula and Richey 2017). None of the arrow points or arrow point preforms at the Pearson site are on the coarse-grained gray-white quartzite.

Arrow Points
There are 14 typologically identified arrow points in the Pearson site collections at TARL (Table 7), primarily Alba (Figure 9a-e) and Fresno (Figure 10a-c)   The Steiner arrow point is one of the earliest arrow point forms in East Texas (Shafer and Walters 2010), estimated to date to the Late Woodland period. The Pearson site specimen is made from a local red chert, as is the post-A.D. 800 Catahoula arrow point (see Table 7). More than 57 percent of the Alba arrow points are made on locally available quartzite and chert raw materials, with the remainder on non-local cherts that may have originated in Central Texas lithic sources. By contrast, 100 percent of the Fresno points and preforms are on non-locally available cherts (see Table 7), possibly procured from Red River gravels, indicating a considerable change in the aboriginal procurement of lithic raw materials from Early Caddo period times to the 18 th century, as well as the intensity of regional interaction/exchange.

___________________________________________________________________________
The glass beads from the Pearson site are dominated by blue colors (n=1074, 47.8 percent) and white (n=1154, 51.3 percent). Other colors include red (n=20, 0.9 percent), black (n=1), and emerald (n=1). The glass bead color pattern for the Pearson site is much different than roughly contemporaneous sites in the Nacogdoches County area of East Texas (see Avery 2017: Figure 4 and Table 1), as blue beads comprise at least 60-70 percent of the assemblages. In sites in the Red and Sabine river basins with glass beads-most especially the Gilbert (41RA13), Womack (41LR1), and Roseborough Lake (41BW5) sites-the assemblages are dominated by white and blue beads, with small percentages of red beads (Avery 2008: Figure 4); at the Gilbert site, however, about 13 percent of the glass beads are black (Jelks 1967; Avery 2008: Table 1), compared to less than 0.1 black beads at the Pearson site. It is possible that these color differences reflect tribal/ethnic associations, as well as temporal trends (Avery 2008:61). Walthall (2015:276-277) also suggests similarities in bead colors due to inter-tribal gift giving and exchange and shared trade sources, likely French traders.
A general sequence of bead types from late 17th century to 19th century sites in Texas and Northwest Louisiana can be proposed based on comparative data on sites with large bead assemblages where the beads were uniformly classified using the Harris and Harris (1967) classification system (Table 10), and as such provides an ideal data set to determine how the bead assemblage from the Pearson site compares with regional trends in bead use at Native American sites in the region. For this purpose, the more than 180 bead types in the Harris and Harris (1967:139-155) scheme were combined into eight broad groups based primarily on bead size, decoration, and method of manufacture (i.e., drawn or wire-wound beads). Groups VII and VIII (Table 10) represent several bead types that Harris and Harris (1967:157) suggest appeared in the bead trade during two specific temporal intervals: 1767-1780 and 1780-1820 respectively.    The sites can be readily separated into three chronologically distinctive bead assemblages: those that date before 1700, ca. 1700-1767, and ca. 1760-1850 (see Table 10). The first group is dominated by medium-sized to large white, blue, and black beads of simple construction, with less than 30 percent of the beads being small drawn beads of the same colors.
The 1700-1760 sites have more small drawn beads than do the pre-1700 sites, with one group ranging from 56 to 72 percent, along with significant numbers of medium to large drawn beads (Group I) and drawn and tubular-shaped Cornaline d'Aleppo beads (Group V). The other group, dating from ca. 1720-1780, including the Pearson site, has a much higher proportion of small drawn beads (95-98 percent), with Cornaline d'Aleppo beads (see Table 10). Cornaline d'Aleppo beads in particular seem characteristic of many eighteenth-century Texas Caddo and Wichita sites, more so than many other historic Native American sites in the southeastern United States.
Large striped (Group II) and wound beads make their first appearance in the 1700-1760 sites. This is also generally the case for colonial French Louisiana sites in the Southeast (Smith 2002), with the exception of earlier 17th century striped beads from a few sites that appear to be of Spanish origin (Smith 1983(Smith , 1987(Smith , 1990Ricklis 1994).
The bead assemblage data show a clear temporal trend in the frequency of the small drawn "garment" or "embroidery" beads (Group IV) replacing the larger and heavier "necklace" beads (Groups I-III) by ca. 1750. This trend or shift in bead size has been previously noted by Gregory (1973) and Hunter (1990) in Texas and Louisiana 18th century aboriginal sites. This overall bead size trend appears to culminate in the mid-19th century in Texas and Northwest Louisiana sites, along with the appearance after 1800 of large faceted beads (see Table 10). To reiterate, by the early 19th century, small drawn beads comprised more than 90-95 percent of the beads from this group of sites, and the larger beads were primarily faceted (see Table 10).

Lead Bead
A fragment of a lead ball in the Pearson site assemblage was modified into a rounded lead bead. The lead bead is 8.9 mm in diameter.

Gun Parts
Among the gun parts recovered in the archaeological investigations at the Pearson site is a fragment of a cast brass trigger guard (Figure 12a; see also Blaine and Harris 1967: Figure 39) that was broken at the rear tang. The piece is 45.3+ mm in length, 13.1 mm wide, and 2.5 mm thick. In Lot 11, there is an engraved cast brass finial from the butt plate (Figure 12b; see also Duffield and Jelks 1961: Figure 13w; Blaine and Harris 1967: Figures 26 and 37c). The finial is 37.0+ mm in length, 13.7 mm in width, and 2.2 mm in thickness. There is also an iron horizontal pivot gun lock sear in the gun parts assemblage (see Duffield and Jelks 1961: Figure 13p; see Blaine and Harris 1967: Figure 29e-f). The sear is 32.0 mm in length, 23.9 mm in height, and ranges from 2.6-4.9 mm in thickness. The hole for attachment on the trigger arm (see Blaine and Harris 1967: Figure 27) is 4.2 mm in diameter.
One iron gun barrel fragment (Lot 11) from an 18 th century French trade gun is in the Pearson site artifact collection at TARL. The barrel has been flattened at one end (see Duffield and Jelks 1961: Figure  13a'), and is at least 102.9 mm in length, 23.9 mm in width, and with an interior diameter of 15.4 mm.

Gunflints
There are four blade gunflints in the Pearson site collection at TARL made from grayish-brown chert (n=2) (Figure 13a) or black chert (n=2) (Figure 13b-c); based on the color, the former may be French gunflints while the latter may be English gunflints (see Durst 2017: Table 18.3). Blade gunflints were generally made after the mid-18 th century, while earlier gunflints were of the spall type (Kenmotsu 1992:152). The gunflints range from 23.0-33.6 mm in length along the sides (see Durst 2017: Figure 18.4), and are perhaps for use with carbines, 17.0-23.2 mm along the heel, and are 5.4-8.0 mm in thickness. Three of the four gunflints are unifacially flaked, and have damage on the striking edge as well as the heel and/or the sides; one of these gunflints has a broad ridge (see Durst 2017: Figure 18.4) on one face. The one bifacially chipped gunflint has edge damage on the striking edge as well as one side.  Hamilton 1976Hamilton :33, 1979, two with sprue scars (see Keith 2017: Figure 19.6). These range from .43 cal to .56 cal in size, referring "to the bore diameter as it relates to size of the lead ball it took" (Keith 2017:479). There is also one poorly cast lead ball fragment; it is 0.48 inches/12.3 mm in diameter.

Lead Sprue
There are two lead sprue masses in the assemblage (Lots 1-77 and 4-58) (see Duffield and Jelks 1961:61); these are the left-over residues from the on-site manufacture of lead balls for use in muskets or fusils. The lead masses range in size from 26 x 16 x 5 mm to 44 x 27 x 11 mm in length, width, and thickness.

Iron Axes
The first of the iron trade axes in the Pearson site collection has a broad blade and curved bit and a wide eye hole (42.2 mm in height and between 29.0-62.0 mm in width) (see Duffield and Jelks 1961: Figure 13a; see also Harris et al. 1967: Figure 23). The axe blade is 71.0 mm in width, a maximum of 82.0 mm in height, between 14-28 mm in width from near the hole to the edge of the blade, and 14.28 mm in thickness. The eye hole has a triangular shape, with 44.0 and 47.0 mm edges and a 47.0 mm diameter; the iron of the eye hole is 7.0 mm thick.
The second iron axe (Lot 19-92-198) has a blade composed of two layers of iron that have split apart slightly at the curved blade edge. The blade is 84.0 mm in length, only 5-10 mm in width, thinning towards the bit, and between 27-58 mm in height. The eye hole on this axe is 42 mm in height, with a 31.0 mm inner diameter; the eye hole is only 4.0 mm thick.
Blaine (1988, 1992:189) suggests that the iron axes or hatchets at the nearby Gilbert site (41RA13) "were used primarily for working metal rather than wood." Blaine (1992:189) suggests this is part of the aboriginal experimentation with metals that took place on aboriginal historic contact sites where metal objects were apparently readily obtained in trade with the French.

Tinkler
The one sheet brass tinkler in the assemblage is a pointed base form (see Harper et al. 1967: Figure 43a-e) with an opening at the top ( Figure  14; see also Duffield and Pearson 1961: Figure 13i). It is 27.0+ mm in length and 6.2 mm in width.

Sheet Brass Bail Ear and Sheet Brass Kettle Fragments
The assemblage has two pieces of folded sheet brass from kettles as well as one kettle bail ear (see

Summary and Conclusions
The Pearson site (41RA5) at Lake Tawakoni on the upper Sabine River in the Blackland Prairie in modern-day Rains County, Texas, is a multi-component aboriginal site that was excavated by the Texas Archeological Salvage Project in 1960 (Duffield and Jelks 1961), prior to construction of the lake. The work documented a wide but shallowly dispersed archaeological deposit on natural sandy alluvial rises in the Sabine River floodplain, and these deposits include temporally diagnostic artifacts from Middle Archaic, Late Archaic, Woodland, Early Caddo, and 18th century contexts. This last component, the most substantial of the different components at the site, was of particular interest to Duffield and Jelks (1961) because they argued that the Pearson site was part of a Tawakoni-Yscani village occupied in the 1760s. During their work, a diverse range of historic European trade goods were documented as the site's artifact assemblage, primarily including gun parts from French fusils, glass beads, and various iron or brass tools or containers. along with triangular arrow points of the Fresno type, and 209 ceramic vessel sherds from two different ceramic traditions from an unknown number of vessels.
Although there are no radiocarbon dates from the Pearson site, the recovered glass beads, gun parts, and engraved and tempered Caddo pottery found in the work suggest that the site was occupied in the mid-18 th century. The work done by Johnson and Jelks (1958) and Duffield and Jelks (1961) suggests that the Pearson site was part of a Tawakoni-Yscani settlement visited by the Spanish in 1760. The ceramic vessel sherds found there are from both Wichita (15.3 percent) and Caddo (84.7 percent) traditions, but the overall low density of ceramic sherds suggests that the production of vessels was less important by ca. 1750 to the Tawakoni-Yscani than was their reliance on Caddo manufactured vessels obtained through French trade networks as well as copper kettles to cook and store foods (see Beach 2015:86). Consequently, I view the existence of the two ceramic traditions at the Pearson site not to be evidence that Tawakoni/Yscani and ancestral Caddo peoples lived together on the site, as Baugh (2014:44) did in his analysis of the ceramic vessel sherds from the Lowrance site (34MR10) in south-central Oklahoma, but as evidence of a Tawakoni/Yscani settlement whose occupants obtained an unknown number of Caddo vessels in trade and exchange, along with European trade goods. Such ceramic vessels may have been obtained, for instance, from the Caddo occupants of the Gilbert site (41RA13) in the upper Sabine River basin or ancestral Caddo groups on the Red River (Story et al. 1967;Perttula 2012). The occurrence of several thousand sherds from Caddo fine ware, utility ware, and plain ware vessel sherds in the various midden features at the mid-18 th century Gilbert site does indicate that the site was occupied by Caddo peoples that regularly made, used, and broke ceramic vessels during the course of the occupation, and likely traded or exchanged vessels with the Tawakoni/Yscani. Instrumental neutron activation analysis of three sherds from the Gilbert site (Perttula 2012:13) also indicate that they came from vessels made with local sub-region 5 clays in the upper Sabine River basin.