Archaeological Sites Along King Creek in Western Nacogdoches County, in East Texas

The King Creek area of western Nacogdoches County in East Texas is known to be a locality where Historic Caddo sites (of the Allen phase, ca. A.D. 1650-1800) are abundant, or at least abundant relative to many other parts of East Texas. In addition to there being at least two branches of the late 17th-early 19th century El Camino Real de los Tejas that bisect the area on their way to crossings on the nearby Angelina River, three important Historic Caddo sites have been identified not far apart in the valley: J. T. King (41NAI5), David King (41NA32l), and Wes Wisener (41NA336); the David King and Wes Wisener sites lie a short distance south of the J. T. King site, on the west side of the King Creek valley. Tom Middlebrook suggests these sites, and other Historic Caddo sites yet to be identified in the valley, are part of an early 18th century Hainai Caddo village that had farmstead compounds dispersed across at least a 3-4 km stretch of the valley. In 2009, we had the opportunity to conduct archaeological survey investigations on private land on King Creek and one of its tributaries. The survey area is about l-1.5 km northeast of the J. T. King site. The principal focus of the survey work was to identify other Historic Caddo sites in the King Creek valley. Our interest in this property had first been piqued because the landowner had reported that a iron Spanish sword had been found on the property some years earlier; the landowner also mentioned that there were preserved segments of El Camino Real de los Tejas on the property. This article summarizes the findings from a first round of archaeological survey investigations


INTRODUCTION
The King Creek area of western Nacogdoches County in East Texas is known to be a locality where Historic Caddo sites (of the Allen phase, ca. A.D. 1650-1800) are abundant (Middlebrook 2007;Perttula et al. 20 lOa: Walker and Perttu la 20 I 0). or at least abundant relative to many other parts of East Texas. Tn addition to there being at least two branches of the late 17th-early 19th century El Camino Real de los Tejas that bisect the area on their way to crossings on the nearby Angelina River, three important Historic Caddo sites have been identified not far apart in the valley: J. T. King King site, on the west side of the King Creek valley. Tom Middlebrook (2010 personal communication) suggests these sites, and other Historic Caddo sites yet to be identified in the valley, are part of an early 18th century Hainai Caddo village that had farmstead compounds dispersed across at least a 3-4 km stretch of the valley. ln 2009, we had the opportunity to conduct archaeological survey investigations on private land on King Creek and one of its tributaries. The survey area is about 1-1.5 km northeast of the J. T. King site (see Figure I). The principal focus of the survey work was to identify other Historic Caddo sites in the King Creek valley. Our interest in this property had first been piqued because the landowner had reported that a iron Spanish sword had been found on the property so me years earlier; the landowner also mentioned that there were preserved segments of El Camino Real de los Tejas on the property. This article summarizes the findings from a tirst round of archaeological survey investigations SITE.. 'I A total of six new archaeological sites (41NA323-328) have been identified to date along this particular stretch of King Creek. Three of the sites have prehistoric and/or Caddo occupations (41NA323, 41NA325-326), a fourth appears to be of Historic Caddo age (41NA327), a fifth has both ll11known prehistoric and late 19th-early 20th century artifacts (41NA324), and the last site was occupied in the early to mid-19th century (41NA328). During the course of the archaeological survey investigations, a total of 16 7 prehistoric and 46 historic artifacts were recovered from surface collection and/or shovel testing (Table I); the majority of the artifacts (n= 14R), all aboriginal, arc from 41 NA327.
Prehistoric and/or Caddo Sites 41 NA323 is on a rise ar the edge of a wooded toe slope (31 0-320 feet amsl) overlooking the floodplain of King Creek. The estimated site area is 1200 m 2 (0.3 acres). This is the site that the. landowner indicated a Spanish sword was found eroding out of a cut bank some. years ago.
Two positive shovel tests, documenLing archaeological deposits from 0-60 em bs, antl a small surface collection, recovered artifacts from what appears to be primarily a Caddo site that was occupied sometime after ca. A.D. 1200. The. temporal estimate is based on the occurrence of brushed (n=2) and brushed-punctated (n=l) ceramic sherds from the shovel testing; a parallel incised body sherd was also recovered at the site. The brushing of utility wares by Caddo potters in the. Angelina River hasin postdates ca. A.D. 1200, and these decorative methods on ceramic vessels continued until at least the early 18th century. Thus, it is pos-  Creek. Artifacts were noted over a 4000 m 2 area (I acre) in surface and shovel test contexts. The one positive shovel test documented archaeological deposits that extend from 0-40 em hs. Artifacts found at 41 NA325 include a piece of gray chert lithic debris from the surface, along with a grog-tempered diagonal incised body sherd from a Caddo utility ware vessel of unknown age. The one shovel test recovered a plain sandy paste Goose Creek Plain, vur: unspecified body shcrd. This type of pottery is found in ca. 500 B.C. to A.D. 800 Woodland period contexts in the Angelina River basin in East Texas (Story 1990;Perllula 2008).
The last of the prehistoric sites in this survey area is 41 NA326, which is located on a lower upland slope (31 0 feet amsl) overlooking the floodplain of King Creek and a tributary stream 150 m to the north that flows west a short distance to a confluence with King Creek. Shovel testing and surt'ace artifacts indicate that the site covers ca. 1350 m 2 (0.3 acres).
Two positive shovel tests document archaeological deposits between 20-40 em in depth.
Only prehistoric lithic debris (n=4) of an unknown age were found at 41NA326. These include flakes on local raw materials-brown chert (n=l), quartzite (n=l), and petrified wood (n=l)-as well as a non-local piece of gray novaculite.

Possible Historic Caddo Site
This possible Historic Caddo site (41NA327) is on a wooded alluvial rise and alluvial fan (290 feet amsl) immediately adjacent to the King Creek floodplain. It covers an estimated 10,800 m 2 (2.7 acres) as defined by 14 positive shovel tests. Archaeological deposits extend from 0-50 em bs, based on the shovel tests and the excavation of Unit l, a l x 1 m unit. The mean artifact density in the shovel testing at 41NA327 is 5.5 per positive shovel test, or ca. 44 artifacts per m 2 . The one excavation unit has documented an area with a density of 71 artifacts per m 2 ( Table 2). Ceramic sherds comprise more than 90% of the artifacts from the site (sec Table 1 ).
Tahle 4 provides a more specific hreakdown of the decorative elements identified in the 4lNA327 utility wares and fine wares. Most of the jars are decorated with brushing om the rim and/or body (probably from Bullard Brushed jars), although two rims are brushed as well as punctated below their folded over rims, and a fifth rim has horizontal incised lines on it. Jar bodies had primarily parallel (likely vertical in orientation), vertical, overlapping, and opposed brushed marks on them, with small amounts of body she.rds with brushed-incised lines (Spradley Brushed-Incised?, an Historic Caddo type in the Angelina and Neches river basins), brushedpunctatcd body sherds, a few inc.iscd sherds (including one with a cross-hatched element), and a single Historic Caddo ceramic assemblages in the Angelina River basin (Shawn Marceaux. November 20())5 personal communication). Another fine ware sherd has an engraved design (horizontal lines) on the rim of a carinated bowl, and the vessel body has hori..:ontal brushing marks (see Tahle 4). The brushing of fine ware vessel bodies is a practice seen on both Late Caddo (ca. A.D. 1400-1650) und Historic Caddo vessels in the Angelina-Neches river basins, so its occurrc nee at 41 NA327 would not be unexpected if the site dates to the HistOtic Caddo period. Other engraved shcrds arc currently unidentifiable to type, hut include examples with geometric elements, as well as two with panels filled with horizontal or parallel engraved lines (see Table 4).  n * this rim sherd has a possible Redwine mode rim (Walters 2010), with a folded over and flat lip: CB=earinmed bowl: +=King Engraved Overall, the ceramic sherd assemblage rrom 41 NA327 is quite similar in character to the three Historic Caddo sites situated l-2 km down~trcam on King Creek (Table 5). Lt is true that no Patton Engraved sherds-the quintessential Historic Caddo ccramic type in the Angelina-Neches river basinhave been found at 41 NA327, but then again only 13 small engraved sherds have been recovered at the site during our limited archaeological investigations. King Engraved, another Historic Caddo line ware type, is present at 41NA327. More telling is the amo unt of brushed sherds in the 41 NA327 assemblage, as the proportion of hrushed sherds in an assemblage as a whole. or its proportions in the decorated sherd assemblage, arc key measures in the identification of Historic Caddo sites in the Angelina River hasin in the absence of European trade goods (see Middlebrook 2007;Perttula et al. 201 Oa, 201 Ob;Walker and Perttula 2010).
By the late 17th-early 18th centuries. for example, in Caddo sites in the Angelina River basin such as Deshazo (41NA27), Mayhew (41NA21), Steven Spradley (41 NA2011), and the Mission San Jose de la Nasoni sites (41RK191,198,200), brushed shcrds account for hetween 50-90% of all the decorated sherds. All of the King Cn:ck sites, including 41 NA327, discussed here have assemblages where between 69-78.5% of the decorated sherds are hru shed. Furthermore, more than 51% of all the sherds (plain and decorated) from these generally contemporaneous sites have brushed decorations, and the highest proportion (58.8%) is documented from 41 NA327 (see Table 5).
Another measure of the age of the Caddo oc-(;Upation at 41 NA327 is the plain/decorated sherd ratio (P/DR) (see Table 5). This ratio appears to hold considerable promise as an independent means of establishing the age of Caddo ccramk-bearing components on Historil: Caddo sites in the Angelina-Neches River hasin. For example. Histori(; Caddo sites in thi!> region have P/DR values that arc less than 0.60. but Caddo sites that arc older have P/DR values that range from 0. 77 to greater than 4.00. For example, the P/DR ratio from the ca. A.D. 1675-1715 occupation at the DeshaLo site (41 NA27) is only 0.30 (Fields 1995 ). The Allen phase component at the Kah-hah-ko-wha site (41CE354) has a P/DR of0.20 (Petttula and Nelson 2007a:72, 74), while Historic Caddo (Nahcdache Caddo) ceramic assemblage on San Pedro Creek (Perttula andNelson 2006. 2007b: Pcrttula et al. 20 I Oa) have P/DR ratios that range from 0.07-0.60. The P/OR ratios from the four Caddo sites on King Creek listed in Table 5 range only between 0.33-0.51, and it seems n:asonable to conclude that all four sites are generally contempo raneous Historic Caddo settlements. Further investigations at 41 NA327 will be geared to acquiring a larger sample of both plain and decorated shcrds to refine the select ceramic comparisons. us well as determining if European trade goods are present at the site: these goods arc rare (a few beads, lead balls, kettle fragments) at the other King Creek Historic sites.
In addition to the possible I listoric Caddo ceramic sherds from 41 NA327, there are three Goose Creek Plain. \'W: unspecified body sherds from ST 11 (0-20 em bs), ST 12 (0-20 em bs), and ST 19 (20-40 em bs). The recovery of these shcrds is evidence that 41NA327 was also OC(;upied by Mossy Grove Culture people (see Story 1990: Perttula 2008) some time during the lengthy Woodland period. A dart point fragment from ST 7 (40-50 em hs) suggests that 41NA327 was also occupied during the Archai~.: period, perhaps ca. 4000 13.P. or earlier.
The dart point tip and blade, made from a light gray chert. has a distinctive hcvclcd blade, an allrihule more commonly seen on points made and used prior to the Late Archaic in East Texas.
A single pie~.:e of burned animal hone was re- The prehistoric artifacts on the surface consist of three pieces of non-corti~.:al lithic debris of an unknown age. The lithic debris are dark gray chert (n=2), probably from a non-lo~:al (i.e., Central Texas) raw material sour~.:e, and a white quartzite (n=l), possibly Glover quartzite from a Neches River source (see Perttula and Nelson 2006).
The historic artifacts found at 41 NA324 are likely from a tenant farm that was known to be in this general location (George Perry Campbell, May 2009 personal communication). The artifacts include four pieces of plain whiteware (n=2) and porcelain (n=2),lale 19th to early 20th ~.:enlury stone ware sherds (one shcrd with a Bristol glaze with a cobalt tint: one sherd with an Albany glaze: three shcrds with a brown kad glaze; and one salt-glazed sherd) (see Greer 198 I;Lcho 1987), and a clear tableware glass shcrd.
EARLY TO MID-19TH CENTURY SITE 41 NA328 is an early to mid-19th century farmstead on an upland landform ( 350 feet amsl). In addition to ar~.:haeologi~.:al deposits that extend from 0-40 em bs over a 2500 m~ area (0.6 acres), there is a visible well depression al the site, and there are ruts from a section of El Camino Real de los Tejas that run just south of the farmstead.
Seven positive shovel tests at 41 NA32R contain early to mid-19th century plain (n=7) and decorated (n=5) whileware ceramks, cut nails (n= I R. I g2o-l89l, see Wells 2000). brown beer/snuff bottle glass (n=l), aqua-colored window glass (n=l). dauh or burned clay (n=2). and hand-made brick fragments (n=l ); the dauh and brick fragments are likely from a chimney associated with a wood-framed structure that stood on the site. The density of artifacts in the shovel testing is 5.0 artifacts per positive shovel test. or ca. 40 artifacts per m 2 • One of the de~.:orated whitewarc sherds from 41 NAJ2g is a hlue painted shell-edged rim with a non-scalloped rim hut impressed lines. Blue shelledged plates and planers with unscalloped rims and impressed lines were being made by the 1840s, while the earlier symmetrical scalloped shell-edged ware continued to he made into the 1S30s (Hunter andMiller 1994. 2009:13 ); this earlier form is absent in the 41 NA32R artifact sample. The other four decorated sherds are transfer-printed. either brown (n=I). light hlue (n=l), red (n=l). or black (n=l). The date ranges of production of the ditlcrcnt colors of transfer-printed <:eramics found at the site are: hrown (1818-18n9). light hlue (1818-1867), red (1818-1880), and black (1785-1864) (Samford 2000:Tahle 5). Because these transfer-printed sherds are whitewares. and based strictly on the production date ranges, the occupation at 41 NA328 could have ranged from ca. 1830s-l880. Mean beginning and end production dates for the transfer printed wares suggest these sherds are from vessels that were most likely manufactured between ca. 1830-1843 (Samford 2000:Tahle 5).
The one piece of aqua-colored window glass suggests that at some time. the wood-framed strw.:turc at 41NA328 had a window installed in one or more of the walls. The pane thickness (1.9 mm) of the one window glass shcrd suggests the window pane was manufactured and installed in the early 1870s (Moir 1987).

CONCLUSIONS
The recent archaeological survey of a parcel of private properly along King Creek in western Nacogdoches County in East Texas has documented sporadic usc of the land since at least 4000 years ago by ancestors of the modern day Caddo Nation of Oklahoma. The existence of a branch of El Camino Real <.It: los Tejas that cuts across the property. and the possibility that a Spanish sword was found at 41 NA323, suggests use of the land during the Spanish and/or Mexican colonial periods (ca. 1690 to 1836). The land wa~ also settled by Anglo-Americans in the 1830s-probably after the Caddo peoples had been removed from their traditional East Texas homelands and after Texas became an independent Republic-and used as a farm since that time.
With respect to the aboriginal use of the land, it was during occupations by Caddo peoples (after ca. A.D. 1200 ami as late as the early 1 Xth century), that the land along King Creek was apparently most intensively used for scttkmenl. The Caddo sites (41NA323, 41NA325. and 41NA327), although poorly known due to limited archaeological investigations, wen:. likely farmsteads or small hamlets (in the case of 41 NA327, the possible Historic Caddo site). Agricultural fields would also have heen situated around each of the habitation sites, a~ well as foot trails connecting the sites to others in the King Creek valley, including the Historic Caddo sites not