Analysis of the Prehistoric Artifacts from the Pace McDonald Site (41AN51), Anderson County, Texas

The Pace McDonald site (41AN51) is a poorly known prehistoric Caddo mound center on Mound Prairie Creek in Anderson County, Texas, in the upper Neches River Basin. With the permission of one of the landowners, Mr. Johnny Sanford, the Friends of Northeast Texas Archaeology are planning on initiating an archaeological research effort at the site in 2010. The ultimate purpose of this work is to learn more about the native history of this mound center-when it was occupied and used, and by which prehistoric Caddo group--its intra-site spatial organization, and ultimately obtain site-specific archaeological information that can help understand the site's place and role in the Caddo prehistory of this part of East Texas. It will be a long-term effort to accomplish these tasks. We intend to rely upon both archaeological (i.e., survey, surface collections, systematic shovel testing, and focused hand excavations) and archaeogeophysical disciplines (especially to complete a magnetometer survey of as much as the site as possible, as this has become an important aspect of Caddo archaeological investigations, to gather relevant archaeological information on the location and character of Caddo house features and outdoor activity areas, as well as the associated material culture remains and preserved plant and animal remains. One key aspect of our work is to understand the characteristics of the Caddo material culture from the Pace McDonald site, since this will have a large bearing on the age of the Caddo occupation, which has been a matter of dispute for some years. In this article 1 summarize the results and findings of a recent examination of the site's prehistoric artifacts (especially its prehistoric Caddo artifacts) in the collections of the Texas Archeological Research Laboratory at the University of Texas at Austin (TARL).


INTRODUCTION
The Pace McDonald site (41AN51) is a poorly known prehistoric Caddo mound center on Mound Prairie Creek in Anderson County, Texas, in the upper Neches River Basin (Pearce and Jackson 1933;Newell and Krieger 1949;Story 2000;Thurmond 1978). With the permission of one of the landowners, Mr. Johnny Sanford, the Friends of Northeast Texas Archaeology are planning on initiating an archaeological research ctTort at the. site in 20 I 0. The ultimate purpose of this work is to learn more about the native history of this mound center-when it was occupied and used, and by which prehistoric Caddo group--its intra-site spatial organization, and ultimately obtain site-specific archaeological information that can help understand the site's place and role in the Caddo prehistory of this part of East Texas.
lt will be a long-term effort to accomplish these tasks. We intend to rely upon both archaeological (i.e., survey, surface collections, systematic shovel testing, and focused hand excavations) and archaeogeophysical disciplines (especially to complete a magnetometer survey of as much as the site as possible, as this has become an important aspect of Caddo archaeological investigations, see Lockhart [2007], Walker and Perttula [200R], Walker [2009]. and McKinnon [20101), to gather relevant archaeological information on the location and character of Caddo house features and outdoor activity areas. as well as the associated material culture remains and preserved plant and animal remains.
One key aspect of our work is to understand the characteristics of the Caddo material culture from the Pace McDonald site, since this will have a large bearing on the age of the Caddo occupation, which has been a matter of dispute for some years.
In this article I summarize the results and findings of a recent examination of the site's prehistoric artifacts (especially its prehistoric Caddo artifacts) in the collections of the Texas Archeological Research Laboratory at the University of Texas at Austin (TARL).

SrTE SETTING
The Pace McDonald site is situated on a large and relatively Hat upland lanMorm (420-430 feet amsl) immediately north of Mound Prairie Creek. in central Anderson County, Texas, in the East Texas Pineywoods (Diggs et al. 2006). Mound Prairie Creek is a southward-and eastward-flowing tributary to the Neches River; the confluence of these two streams lies ahout 20 km to the east of the site. When the site was first visited and recorded in the 1930s, it was in a large cotton field (Pearce and Jackson 1933:2). In more recent years, it is in an improved pasture, and the site is apparently owned by several landowners, including the Texas Historical Commission (see below).
The site is on an expanse of Elrose fine sandy loam, 1-3 percent slopes (Coffee 1975:17 and Sheet 34 ). This is a relatively fertile upland soil that "formed under a pine-hardwood forest in stratified marine and alluvial sediment high in glauconitic sandstone" (Coffee 1975:17 and Table 2). A typical profile of the Elrosc fine sandy loam is a 25 em thick A-horizon that ranges from reddish-hrown to yellowish-red fine sandy loam developed atop a thick (ca. 165 em) Bt horizon composed of red, dark red, yellowish-red, or strong brown sandy clay loam and sandy loam. The underlying C horizon is a massive red loamy fine sand with an occasional fragment of glauconitic sandstone.

PREVIOUS INVESTIGATIONS AT THE PACE MCDONALD SITE
Lim.iteu ar<:haeological investigations at the Pace McDonald site since the 1930s indicates that the site covers ca. 11 acres (45.000 m 1 ) of the upland landfom1 adjacent to Mound Prairie Creek. The site has two deliberately construct<:u prehi storic Caddo earthen mounds, a large and associated habitation area (in- e luding several midden areas) and surface scatters of ceramic and lithic artifacts. and probably at least one associated cemetery area ( Figure I). There are also a numrer of small depressions. with an avcrag<: depth of 30-60 em. visible on the lam.lform that may represent borrow pits for sediments used by the Caddo to build the two eanhen mound:-.. In 1978, th<: depressions ranged from 6.1-25 min diameter ( Thum1ond 1978).
Jackson led the excavations of both mounds. In Mound No. 1, a circular shaft ca. 2.5 m in diameter was excavated by shovel to a depth of 3.28 m below the mound surface (bs). From Jackson's descriptions, there were several mound fill zones and perhaps two Caddo house archaeological deposits in the mound. From 0-147 em bs, there were red (0-102 <.:m bs) sandy clay loam and yellow (I 02-14 7 em hs) clay mound fill or capping zones, with few artifacts: these soils appear to be Bt-horizon sediments that were collected and used to finish the earthen mound. These redistributed sediments capped a 2.5 em thick ( 147.3-149.9 em bs) sandy red clay, that was described as "fairly hard. Seemed to be a floor level as it extended at the same thickness over the entire dug area. No midden material" (Pearce and Jackson 1933:6). If this was a lloor level, the structur<.: associated with it was not used for a lengthy period of time-given the absence of midden deposits-nor was it burned down-given the apparent absence of charcoal on the floor or in the yellow clay above it-before being dismantled and covered over with 147 em of mound deposits.
This apparent clay floor to a structure was capped above a reddish-brown sandy loam mound fill zone deposit that extended from 149.9-213 em bs. This fill zone capped a 49 em thick deposit (213-262 <.:m bs) or ash, "with <.:har<.:oal, lumps of red clay and some sand. Potsherd at top of the ashes; no other midden material" (Pearce and Jackson 1933:6). This ash deposit may represent a deliberate accumulation of ash within a confined space, as with the ash mound at the A. C. Saunders site ( 41 AN 19, sec Jackson 1936;Kleinschmidt 1982), or an accumulation or ash inside a structure that was burned down , as with the ash temple at the Crenshaw site (Schambach 1996). The available archaeological information from these early excavations is equivocal. but I suspect this deposit represents ash built up within a structure, along with the dismantled and burned remnants of that structure. mixed in with relatively clean earth.
The ashy deposits rested on a red sandy clay and red clay that extended from 262-328 em bs in the excavations. These deposits likely represent the buried Bt horizon of the Elrose fine sandy loam soil under Mound No. 1. lf that is the case. then apparently the immediately overlying ash accumulation rested on a prepared surface wherein the A-horizon was scraped away llrst (a common Caddo practice in and under locations where earthen mounds are to he built), as there arc no buried A-horizon (a fine sandy loam) underlying the ash deposit.
A " test hole" of unknown size was excavated by Jackson in Mound No. 2 (Pearce and Jackson 1933: 10). These explorations encountered a thick layer of hard-packed ash between 15-56 em bs. with gravels and lumps of red clay, but no apparent midden materials. The deposits above that (0-15 em hs) appear to represent the plowed portion of the ash deposit. As with the concentrated ash deposit encountered in Mound No. 1 between 213-262 em bs, the Mound No. 2 ash zone probably also represents a deliberate accumulation of ash within a confined space, such as a building, or an accumulation of ash inside a structure that was likely dismantled, abandoned, and destroyed. This deposit may represent ash built up within a structure, along with the dismantled (but not burned) remnants of that structure. mixed in with relatively clean earth.
The ash deposit rested atop a 43 em thick (56-99 em bs) zone of thin lenses ( 1-8 em) of red and yellow clay and sandy soil, probably mound fill zones. Immediately below the ash. however, Pearce and Jackson (1933:10) noted there was a 20 em thick (56-76 em bs) red clay zone; this may represent the initial mound platform of Mound No. 2, or the initial mound platform included both the red clay as well as the underlying red sandy lens (76-99 em bs) if the latter docs not represent a buried A-horizon underneath the mound. Excavations continued to 152 em bs. and the sediments encountered there were described by Pearce and Jackson ( 1933: 1 0) as a .. bluish mixture bearing streaks of yellow, red and gray. The bluish-gray clay-like composition is sticky and has a greasy feel." This deposit under the mound may represent undisturbed C-horizon deposits.
The next archaeological investigations at the Pace McDonald site did not take place until the late 1970s when Pete Thurmond and Ulrich Kleinschmidt examined the site and took a surface collection, made a pace map of the surface distribution of prehistoric artifacts, the two earthen mounds, ami the locations of several possible small horrow pits (sec Thurmond 1978). They noted that the site appeared to he well-preserved, with little evidence of looting at that time. This work led to the placement of the Pace McDonald site on the National Register of Historic Places a few years later, and the purchase of one acre containing Mound No. 1 by the Texas Historic Commission in the early 1980s.
Despite the absence of professional archaeological work at the site bet ween 1933-1978, the Pace McDonald site had not been forgotten by East Texas Caddo archaeologists. Alex D. Krieger, University of Texas archaeologist. had examined the collections obtained by A. T. Jackson from the site as part of his renowned study of the early Caddo mound center at the George C. Davis site (41CEI9; Newell and Krieger 1949: 190). According to Story (2000:22), Krieger "believed that both the Alto and Frankston foci were represented at the site. hut also rceogniLed that it had not been adequately explored." Following the 197R surface investigations by Thurmond and Kleinschmidt at the Pace McDonald site, these archaeologists returned to the site in March 1981 to investigate the reported erosion on the back slope of Mound No. I. While they were there, they obtained a surface collcetivn of artifacts from Mound No. 1 and an area to the northwest of the mound.
ln March 1984, Klcin~chrnidt and Susan Lisk of The University of Texas at Austin inspected the site, noting that there were three pot holes (ca. 0.9 m in depth) in Mound No. 1, and also noting that a one lane oil top road had been built that crossed the site. This road had cut through a I 0 em thick midden area ca. 250 m north of Mound No. 1, in the vicinity of possible borrow pit depression 5 (see Figure  1). A second midden area had been encountered by local collectors and landowner who were building a fence just west of Mound No. I (sec Figure I). Finally, Klcinst.:hmidt and Lisk were told by local collectors that several prehistoric Caddo burials had been reportedly discovered and excavated at the site.
More information on the reported buriab was provided in 1985 and 1986 by another local amateur archaeologist who was familiar with the Pace McDonald site and other archaeological sites along Mound Prairie Creek (TARL files). This amateur archaeologist obtained information that three prehistoric Caddo burials (Burials 1-3) had hcen excavated not far to the west and south of Mound No. 1 (see Figure I). These burials were in grave pits approximately 75-90 em deep that were oriented east-west or north-south: the human remains were reported to be in a poor preservation condition. The Caddo burials had funerary offerings, including plain and decorated ceramic jars, howls, and bottles (n=9, 1 to 4 vessels per burial) and slone artifacts: a celt, a hammcrstone, a pitted stone, and nine round ~tone~ (possible polishing stones'!). Jan Guy ofTARJ. identified three of the vessels from Burials 1 and 2 as a Hickory Engraved bottle. a cf. Bowles Creek Plain howl, and a cf. Weches Fingernail Impressed, vw: Altn bowl (see Stokes and Woodring 1981: 185-18o and Figures 22m and 23b-c); there were no photographs availahle of the vessels from Burial 3. Jan Guy (TARL Pace McDonald files, 3/86) suggested that the vessels .. date to either the early or Middle Caddoan periods." None of these funerary objects have been properly documented, and it is currently unknown who has control of these artifacts from the Pace McDonald site.
Finally. the last mention in the records at TARL of a professional investigation of the Pace McDonald site was a visit by personnel from the Office of the State Archeologist at the Texas Historical Commission in March 1996. They noted that Mound No. I, owned by the Texas Historical Commission, wa~ badly overgrown.

COLLECTIONS AT TH~ TEXAS ARCHEOLOGICAL RESEARCH LABORATORY
The prehistoric artifacts from the Pace McDonald site in the Texas Archeological Research Laboratory collections consists of a variety of ceramic and lithic artifacts. The ceramics include plain and decorated ceramic vessels sherds (n=975) and a ceramic pipe. There are also chipped stone tools in the collection. including dart points (n=44), arrow points (n=31 ), two bifaces. and a single llake tool, as well as ground stone tools (n=6), and lithic debris (n=ll). The collection also contains animal bone (n=5) and a single unmodified marine conch shell.

Ceramic Vessel Shcrds
Combining the various TARL collections obtained from the Pal:C McDonald site from the early 1930s to the mid-1980s, including surface collections, the 1933 ex.cavations, and donated collections. the collections from the site amount to a total of 975 ceramil: vessel sherds. Approx.imately 77% of the vessel sherds are from undecorated vessels or the undecorated portions of decorated vessels; 22.6% of the sherds, including both fine wares (engraved and red-slipped sherds) and utility wares (wet paste decorations), are decorated (Table I).
Plain to decorated sherd ratios (P/DR) from numerous Caddo sites in East Texas appear to hold considerable promise as an independent means of establishing the age of Caddo ceramic-bearing components (provided samples of plain and decorated shcrds arc larger than about 200-300 sherds per site; the Pace McDonald site meets this data threshold). When P/DR ratios from different ceramic assemblages can be linked with absolute ages as established by radiocarbon dating from those assemblages, this should allow further refinements in how P/DR ratios change through time in East Texas CadJo sites. Looking at Early Caddo to Historic Caddo ceramic assemblages in the region through time, ceramic assemblages have lower proportions of undecorated sherds through time and thus a lower P/DR ratio (Pertlula 200Ba:9, 315-317). Analyzed pre-A.D. 1200 sites (n=3 assemblages) have plain/ decorated sherd ratios that range between 2.97-4.80. Middle Caddo sites (ca. A.D. 1200-1450, n= 7) have ratios that range between 1.30-2.65. In known Late Caddo sites (n= II) in the Neches, Angelina, and Sabine river basins, by contrast, the P/DR ranges from only 1.30-0.47. Finally, post-A.D. 1680 Caddo occupations in the Neches-Angelina river basin have P/DR ratios that range from 0.20-0.30. The plain to decorated shcrd ratio (P/DR) is a relatively high 3.43 at Pace McDonald, suggesting the ceramic assemblage may date from pre-A.D. 1200 times; the Jecorated sherd assemblage, however, suggests a post-A.D. 1200 age (see below).
There are 39 rim shcrds in the vessel sherd wllcction. More than 51% are from decorated utility ware vessels (n=20); another 38.5% are from plain vessels (n=l5), and only 10.3% arc from line ware vessels (n=4, from both engraved and red-slipped vessels). One bowl or carina~ed bowl with interior and exterior red-slipped surfaces has a distinctive Redwine mode rim treatment (see Walters 20 I 0:7R).
The line ware sherds from the Pace McDonald site have both red-slipped ( 40(}(of the line ware sherds), red-slipped and engraved (2%•), and engraved ( 58lf(l) decorative elements (Tahle 2). The proportion of red-slipped sherds (from bottles and bowls/carinated bowls, especially the latter) is considerable for an upper Neches River basin Caddo site (Perttula 2008b). In East Texas generally, the manufacture and use of red-slipped pottery unemhellished with engraved decorations is most commonly seen in Middle Caddo ceramic traditions, whether it be in Caddo sites on the Red River or in parts of the upper Sulphur. Big Cypress, and Sabine River basins. Seventy percent of the red-slipped sherds are slipped on both the exterior and interior surfaces, while 25% arc slipped only O!il the exterior surfaces; these laller sherds are from bottles. ShcrJs from vessels slipped only on the interior vessel surface (probably howls) are nm common (5%).
There are a wide variety of engraved decorative elements in the Pace McDonald fine ware cc.ramics (Table 3). Of those that have more than just straight or parallel lines of uncertain orientation, this includes cross-hatched engraved lines; sherds with various kinds of hatched (Figure 2b, d, f-g, i, k-1) or cross-hatched (Figure 2a, h) elements. One distinctive bottle sherd has excised pendant triangles and hatched zones and hatched triangles ( Figure  2e). Other distinctive engraved sherds in the assemblage include a body shcrd with panels filled with opposed diagonal engraved lines ( Figure 2c) and a rim with vertical and opposed diagonal engraved lines ( Figure 2j).
Notably absent in the engraved fine wares at the site are types such as Holly Fine Engraved, Spiro Engraved, or Hickory Engraved (Suhm and Jelks 1962). These fine ware engraved types are considered material culture hallmarks of the Early Caddo period (as well as various decorated utility wares, and the Alto phase (e.g., Story 2000: 14) in East Texas. Their absence at the Pace McDonald site certainly would be indicative of the fact that the Caddo occupation here. postdates the Early Caddo period.
Among the Pace McDonald utility ware sherds are several different kinds of decorative methods represented, as well as Jistinctive decorative elements within each of the larger decorative methods classes (Tahle 4 ). These principally include sherds from vessels decorated with incised Jines ( 43.5% of  Figure  3a-b, d). These utility wares arc likdy from Canton Incised and Pennington Punctated-Incised vessels. Other shcrds have a zone of pun<:tations adjacent to simple geometric the utility wares), sherds with punctated elements (42.9%), and sherds from vessels decorated with incised-punctatcd clements (7 .1% ). Minor decorative methods documented in the site's utility wares arc brushed (2.9%), pinched (2.4%), and appliqued ( 1.2%) categories (Tahle 4 ).
incised elements (see Figure 3c-f), with the punctated clements apparently limited to the body of the vessel and the incised elements restricted to the rim.
One sherd has a narrow im:iscd zone or band filled with cane punetations (see Figure 3c); this decorative clement has been documentcd on Pennington Punctated-Incised vessels (see Suhm and Jelks 1962:Plate 6ld-e).
Among the incised utility wares. the most popular decorative elements include widely-spaced crosshatched lines (likely either from Canton Im.:ised or Dunkin Incised vessels). opposed incised lines (Figure 3h, also probably from Canton Im:ised or Dunkin Incised vessels). parallel or straight incised lines of uncertain orientation (possibly body decorative treatments), and diagonal or horizontal incised lines on the vessel rim. These incised sherds may he from Davis Incised. Dunkin Incised, or Canton Incised vessels, or from other Caddo ceramic types with incised elements that have not been identified to date in the region. Most of the incised rim sherds in the assemblage are from cross-hatched incised vessels (see Table 4). The punctated sherds from the site arc comprised of a mixture of tool (55% of the punL·tated sherds, including tool punctalions arranged in linear rows), fingernail (33% ), large and small cin.:ular (11% ), and cane ( 1%) punctated element:;; (see Tahle 4). The only punctated rim sherds (n=3) have rows of tool punctatiuns on them. The majority of punctated sherds are body sherds, indicating that the bodies of many utility ware vessels are decorated with punctations: the decoration on the rims of these vessels was probably not    1 981) in the assemblage. This particular sherd has crescent-shaped punctated elements above or adjacent to a single straight incised line. A distinctive characteristic of the Pace McDonald utility wares is the occurrence of sherds from brushed, pinched, and appliqucd jars, but only in low frequencies (see Table 4 ). These three categories of decorated utility wares together comprise only 6.5% of the assemblage. They apparently represent diiTerent means of rim ami/or body decoration on jars, with parallel (likely vertical) brushing on jar bodies; vertical appliqued fillets on jar bodies; and vertical pinched rows on both the rim and body of jars. The typological identification of the brushed and appliqued wares is currently uncertain. There arc also three perforated body and base shcrds in the Pace McDonald collection. These likely represent spindle whorls, arc disk-shaped sherds (usually base sherds) that have a central perforation or hole drilled in them. The spindle whorl would have been affixed on a spindle to help maintain its rotary motion during spinning activities. The  presence of spindle whorls at the Pace McDonald site suggests that Caddo women were processing fibers to produce textiles (cf. Alt 1999). Materials that eou ld have been used include animal hair and various vegetable fibers, among them hemp, slippery elm, mulberry. milkweed, and nettle, as well as the hark of trees. A detailed analysis of technological attributes of the Pace McDonald ceramic shcrds wa~ not comluctcd for this project, due primarily to time constraints and the inability to examine the sherd cores in any detail (i.e., hy removing a small bit of the sherd lO examine a freshly broken profile of the core). Nevertheless, it was possible to observe that the vast majority of ceramic vessel sherds from the site arc from vessels tempered with grog (i.e., fired and crushed clay). occasionally in association with other tempers. This is the principal prehistoric Caddo ceramic practice in the upper Neches Rjver valley (see Perttula 2008b: Figure 6-70). More than 14% of the sherds do have ~:rushed and burned bone temper added to the day pa~te (Table 5). Proportionally, hone temper is used most frequently in plain ware~ and utility wares.

Ceramic Pipl'
Ceramic pipes and pjpc shcrds are common artifacts found in upper Neches River basin Caddo sites. especially those sites occupied after ca.
A.D. 1400 (Gilmore 1974;Jackson 1933Jackson , 1936Kleinschmidt 1982). Not too surprisingly then. a complete, but undecorated , L-shapcd (i.e., L-shaped angle between the bowl and the stem) elhow pipe  (Figure 4). The pipe is grog-tempered, and has been burnished on the exterior surface. It is 34.6 mm in height, and has a 55.6 mm stem length. The bowl orifice diameter is 23.7 mm, and the bowl itself is 3.7 mm thick. Along the stem, the exterior orifice diameter is 13.9 mm; the interior orifice diameter is 7.7 mm: and the stem is 3.1 mm thick. Elhow pipes are a style of ceramic pipe manufacture that began to be popular after ca. A.D. 1350 in East Texas and elsewhere in the Caddo area (see Hoffman 1967: Rogers andPcrttula 2004;Pcrttula 2008b), but are virtually the exclusive form of clay pipe made by the Caddo from the 15th century A.D.
on. An examination of the clay elbow pipes from mortuary contexts in the upper Neches River basin (see Perttula 2008b), from cemeteries of known age. indicates that the earliest elbow pipes (Var. A) are plain L-shaped forms. Radiocarbon and thermoluminescence dates on sherds indicate that L-shaped pipes at the Lang Pasture site (41AN38) date in the 14th century A.D., from ca. A.D. 1320-1400. ln other upper Neches River basin sites of known age (i.e., dating to the Frankston phase, subphase 1-3,

Arrow Points
The arrow points in the TARL collection at the Pace McDonald site arc dominated by parallelstemmed Alba points (as well as a single possible Alba point preform). These account for 55% of the arrow points (Table 6). Relatively common are contracting stem Perdiz arrow points (23°/n) and expanding stem and flaring barb Catahoula points ( 10% ).
The earliest arrow point form at the Pace Mc-Donald site is probably the Catahoula type. This point is considered diagnostic of late Woodland (ca. A.D. 700-800+) components in the East Texas and Southeast Texas regions (Shafer and Walters 2010). Alba points, on the other hand, arc generally considered Formative to Early Caddo period (ca. A.D. 800-1200) arrow points, although because they are diagnostic of the Alto phase (Story 2000). they may date from as long a period of manufacture and nse as ca. A.D. 850-1300. They arc particularly well-represented at the George C. Davis site on the Neches River (Newell and Krieger 1949:161 and Figure 56a-h), where they are considered the only "resident type." The later (post-ca. A.D. 1200-1300) arrow points at the Pace McDonald site are dominated by Perdiz points. Perdiz points have been found in a number of East Texas Caddo sites that date from the 13th to the 17th century A.D., but as of yet, unfortunately, no temporally distinctive varieties have been defined within this hroad span of time that would permit a more definitive conclusion as to the age of this prehistoric occupation at the site.
The arrow points from the Pace McDonald site are predominantly manufactured from non-local cherts that are apparently from the Edwards Plateau

Dart Points
There are 44 dart points in the TARL collections from the Pace McDo nald site ( Table 7 ). The most common type is the contracting stem Gary, which accounts for 36% of all the dart points from the site. Based on the shape of the dart points, particularly their basal and stem forms, and the idcntifiL:ation of specific dart point types with known or estimated temporal limits (e.g., Story 1990 The dart points from the Pace McDonald site can be grouped into temporal periods, including Woodland. Late Archaic, and Middle Archaic ( Table  8). The usefulness of these groupings should be evaluated with the proviso that the majority of dart point types that occur in East Texas are not yet welldated hy secure archaeological association with a series of calibrated radiocarbon dates from features or single component archaeological deposits, hut the estimated temporal periods to which the dart points from the Pace McDonald site are assigned is the product of a few calibrated dates a~ well as extrapolations with better dated temporal sequences in the western Gulf Coastal Plain, Central Texas, the Ouachita Mountains. and the Ozark Highlands (cf. Schamhach 1 Y82; Story 1990;Trubitt 2009;Turner and Hester 1 Y9Y). Nevertheless, the groupings follow rather closely the artifact sequences for stone tools postulated hy Story (19YO: Figures 32 and 33) in her synthesis of the archaeology of the East Texas portion of the Gulf Coastal Plain.
A simple comparison of the number of dart points from the site that fall into each of these periods make evident that Woodland period dart points are by far the most abundant (52.3%), particularly the Gary (n=l6), Kent (n=3), and Godley (n=2) types. Late Archaic dart points comprise another 40.9% of the recovered points; these include primarily parallel stemmed and nat-based projectile point forms. expanding stem points with a narrow stem and a flat base, as well as one Williams point. Middle Archaic points include one Dawson and one possible Wells specimens, along with an expanding stem form with a concave hase (see Table 8).  Several different kinds of lithic raw materials were used in the manufacture of the dart points that ended up discarded at the Pace McDonald site, including a variety of cherts (most of non-local origin); petrified wood; and yuartzite. ll is clear that non-local chert raw materials were principally used in dart point manufacture during all temporal periods of occupation at the site, especially during the Middle and Late Archaic periods (see Table 8). Local cherts, quartzite, and petrified wood gradually became more important for dart point manufacture during and after the Late Archaic period, while the use of cherts diminished from 100-77 .8<1'(' in the Middle and Late Archaic periods to 60.91h· in the Woodland period.
These trends in the use of lithic raw materials, particularly the non-local chctts that most likely were from gravels that originated from source ar-ea~ to the west in Central Texas and the Edwards Plateau. suggest that the aboriginal populations that utilized the Mound Prairie Creek area prior to ca. 1200 years ago had access to a wide range of non-local lithic raw materials. This is probably because they were relatively mobile foraging populations that ranged west into the Trinity and Brazos River valleys-where high quality cherts could be obtained from gravel sources-and where they collected these high-quality lithic resources during the course of their settlement and fomging forays. Locally available lithic raw materials became more important as sources of chipped stone tools during the Woodland period. These later groups that used the Pace McDonald site likely had a more territorially-conJined settlement/foraging area in the East Texas Pineywoods, although certainly there were contacts between Pineywoods Woodland groups and peoples living in areas with high quality chert raw matcriab that led to the continued and extensive acquisition of non-local chipped stone for tool manufacture and use.

Bifaces
Both of the two large bifaces in the collection arc made from a non-local chert. These bifaccs arc probably discarded fragments from the attempted manufacture of bifacially chipped dart points.

Flake Tool
A single flake tool with unilateral use wear is in the Pace McDonald collection. This expedient tool was made from a non-local (Edwards Plateau?) dark gray chert.
The ground stone tools in the collection are a disparate lot. with a grooved axe. a polished stone. a mano. and three celts (Table 9); the mano and polished stone are made from locally available lithic raw materials. The grooved axe probably represent s a Middle to Late Archaic usc of the site, as those are the time periods when this tool type was predominantly used by East Texas prehistoric peoples (cf. Turner 2006). The axe is made from a grayish-hlack diorite or tutT from a Ouachita Mountains soun:c area. and has been broken above the groove. lt is 60.4 mm wide at the groove. 69.7 mm from the groove to the bit, 38.2 mm wide, and the bit width is 65.7 mm.
The cobble-sized mano has been ground smooth on both surfaces from usc on a mctate or grinding slab. The quartzite polished stone was probably used to polish and finish ceramic vessels made at the site; it is 46.2 mm in length, 32.3 mm in width. and 12.2  1 0) indicate that the knapping of stone tools took place at the site, hut was not apparently a common activity. Both local (45%) and non-lo~al (55%) raw materials arc represented in the lithic debris.
The non-local lithic debris includes several different colors of chert, including several with a stream-rolled cortex, that likely originated in the Edwards Plateau area of Central Texas, and ~obbles and pebbles of this material can be found in stream gravels well to the cast of the Plateau. including the Brazos and Trinity River drainages. The local lithic raw materials arc also avai lable Marine Shell A small. pitted/etched, and unmodilied marine conch shell is in the collection from the Pace Mc-Donald site, but its provenience within the site is unknown. The conch is 73.3 mm in length and 49.0 mm in width.

Animal Bone
There arc five unidcntillablc pieces of animal hone in the collections. One of these has been burned.

SUMMARY AND CONCLUSIONS
The Pace McDonald site ( 41 AN5 I ) is a prehistoric Caddo mound center in the upper Neches River basin of East Texas; it also was occupied during the Archaic and Woodland periods. Its temporal and cultural relationships to other important mound centers in the region ( Texas Archeological Research Laboratory at The University of Texas at Austin represents the first step in a likely long-term effort to learn more about the nature of the Caddo occupation at this mount! center. What is known about the prehistoric Caddo occupation of the Pace McDonald site:: is that the site is on an upland landform overlooking Mound Prairie Creek, a tributary to the Neches River, has evidence of a settlement that covers at least ll acres , and two earthen mounds were constructed there by the Caddo. Both mounds appear to have been built to cover special purpose structures where significant deposits of ash was accumulated. similar to the main mound at the nearby A. C. Saunders site (see Jackson 1936) (see Figure 5) or the ash temple (albeit. without a covering mound) at the Crenshaw site ( Schambach 1996 ).
There also is a reasonably large sample of ceramic and lithic artifacts from the site that have been collected over the years, mainly from Jackson's 1933 excavations in the two mounds and a surface collection. and various surface collections gathered between 1978 and ca. 19R5. With respect to the prehistoric Caddo component, the material culture remains that can he associated with it include: (I) a significant percentage of plain ware vessel sherds (mainly grog-tempered), primarily incised, punctated, and incised-punctated utility wares. and both engraved (i.e., geometri<..: elements, significant proportion of hatched and cross-clements, mainly triangles, and carinated bowls and bottles) and (many) red-slipped fine wares; (2) an L-shaped elbow pipe, a 14th century ceramic innovation; and (3) Alba and Perdi7. arrow points and a few expedient flake tools.
Temporally, the overall impression obtained from the examination of the decorated sherds. the ceramic pipe, and arrow points is that the Pace Mc-Donald site was first occupied sometime before the 13th century A.D.-based on the predominance of Alba arrow points and some of the decorated utility wares--<:ontemporaneous with some part or the:: lengthy Caddo O(;cupation at the George C. Davis site (Story 2000). Without radiocarbon dates from the site, it is difficult to establish the early beginnings of this occupation with any precision, largely because the apparently lengthy use (ca. A.D. 850-1300) of many kinds of distinctive Alto phase or Early Caddo ceramic and lithic artifacts, as well as the inability to identify temporal varieties of these artifacts that have more discrete temporal periods of use.
The apparent popularity of red-slipped pottery at the Pace McDonald site is consistent with a Caddo occupation during the 13th and 14th centuries A.D., which was when a tradition of manufacturing redslipped carinated bowls and bottles was established across much of the western part of East Texas. At the Middle Caddo period Jamestown Mound site ( 41 SM54 ), for example, more than 26% of the decorated sherds (n=84) arc from red-slipped vessels, including one. Maxey Noded Redware bottle sherd (Perttula and Walker 2008:7).
The end of the Caddo occupation at the Pace McDonald site can be more precisely established as ca. A.D. 1350 to A.D. 1400, based on the presence of the early L-shaped elbow pipe in the assemblage, and perhaps also by the one red-slipped Redwine mode rim (see Walters 20 10). Other indications that the Caddo occupation must have lasted into and through the 14th century include Perdiz arrow points. the brushed and pinched utility ware sherds, and much of the distinctive engraved fine wares found at Pace McDonald. These line wares from the site may be part of a Middle Caddo period (ca. A.D. 1200-1400) East Texas style zone or ceramic tradition in the Angelina and Neches river basins. as well as parts of the middle and upper Sabine River basin. at other better known and studied sites such as Washington Square, Oak Hill, Redwine, and Jamestown (see Figure 5). The engraved fine wares at these sites have hat<.:hed or cross-hatched curvilinear and vertical ladders or narrow panels, as well as hatched and cross-hatched triangles, pendant triangles, or rectangular panels with engraved triangles (see Rogers and Pcrttula 2004); in some instances, there are engraved vessels with vertical and triangular panels filled with wncentric circles (Hart and Perttula 20 I 0). A number of the engraved fine ware vessels from these sites, but not yet identified in the assemblage of Pace McDonald fine wares, have horizontal interlocking, slanting, anti vertical scrolls-including negative S-shaped scrolls-as their principal motif. There are also rayed circles/ sun clements and the swastika cross-in-eircle, but these are absent in the small sample of fine wares in the site collections.
It is interesting that the George C. Davis site may also has a component that dates to the:: Middle Caddo period, and that has a <.:eramic assemblage that may be part of this East Texas ceramic tradition, although red-slipped sherds are not at all common in the ceramic assemblage (Stokes and Woodring 1981:222-223). Story (2000: 13-14) suggests there are ''late Alto phase" ceramics (i.e., dating after ca. A.D. 1200) in certain areas of the site that include "brushed utility wares, a hallmark of the Frankston phase, as well as Maddux Engraved, Pease Brushed-Incised, and other pottery usually associated with the Middk: Caddoan Bossier phase." Stokes and Woodring ( 19R I :206) indicate that this collection of ceramics also include those with incised or punctated panels as well as appliqucd strips. At the George C. Davis site, interestingly, the Maddox Engraved shen1s from "late Alto phase" contexts arc described as having narrow bands (either straight or curvilinear) filled with cross-hatched lines , very similar to several of the decorative elements in the Pace McDonald fine wares (sec Figure 2c-g). Story (2000:23) examined, some years ago, the Pace McDonald collection at the Texas Archeological Research Laboratory at The University of Texas at Austin, and her comments are apt, as she "found little evidence for either an Alto phase or Frankston phase connection. The age and affiliation of this, one of the few other mound sites near Davis I George C. Davis], an: yet to be cstahlished." Thurmond ( 1978:26), some years before, concluded that the Pace McDonald site was "transitional between the two foci [Alto and Frankston] .'' Essentially, what both Story (2000) and Thurmond (197R) arc suggesting is that the Pace McDonald Caddo occupation dates prior to the onset of the Late Caddo Frankston phase. which occurred around A.D. 1400, which is a reasonable suggestion, and hegan sometime hefore the end of the Alto phase. If the ca. 450 years of the "Alto phase" could he suhdivided (and there is not necessarily any good reason to think that they can) into two sequent early and late suh-phases of ca. A.D. 850-1075 and ca. A.D. 1075-1300, l think it would be fair to say that the Pace McDonald site-while not an Alto phase component-probably began to be occupied and used at a time contemporaneous with this hypothetical late subphase (ca. A.D. 1075-1300), and its occupation ~:ontinued into the 14th century A.D. Temporally speaking then, the Pace McDonald site is best viewed as an Early to Middle Caddo period mound site whose occupation overlapped with that of the longoccupied George C. Davis mound center (most likely the component there associated with the construction of the Mound B platform around ca. A.D. 1200, see Story 1 997:65), but apparently continued afrer the George C. Davis mound site was abandoned in the early A.D. 1300s. In cultural terms, then, and relying heavily on the character of the decorated ceramic sherds in the assemblage, the Pace McDonald site was an imrx.1rtant mound center huilt by a prehistoric Caddo group that was apparently related to others of similar socio-political character in the upper Neches. middle and upper Sabine, and the Angelina River basin (see Figure 5). Whether it was a suhsidiary or satellite to any other contemporaneous Caddo mound center~ is not known.