Lake Naconiche Archaeology And Caddo Origins Issues

Sometime around ca. A.D. 800, Lake Naconiche sites were no longer occupied by Woodland period groups of the Mossy Grove culture solely making sandy paste pottery or living as mobile huntinggathering foragers. At this time, from ca. A.D. 750-800 to around A.D. 900, colder and drier conditions began to dominate the local weather. After ca. A.D. 800, were the aboriginal groups Caddo peoples or acculturated Mossy Grove folks? Some findings from the Lake Naconiche archaeological investigations at the Boyette site (41NA285) are relevant to this issue of ethnic affiliations and local, but nevertheless regional momentous, cultural changes. Putting that in context, as best as can be discerned in the archaeological records of the Woodland period occupations at the Naconiche Creek (41NA236) and Boyette sites, if there is any evidence of increasing sedentism, it is only apparent after ca. A.D. 400 or perhaps even as late as ca. A.D. 650, during the latter part of the period. Even so, these occupations were not sedentary in the sense of them being year-round occupations (as with the Caddo settlement history at Lake Naconiche) or even multiseasonal occupations. The sites do not have accumulations of midden deposits, there is no evidence for the construction of sturdy wood structures, and there are only a very modest assortment of burned rock, pit, or post hole features at the Woodland period sites. It is hard to disagree with Story’s characterization of Woodland period settlements in the general area that they reflect “intermittent encampments by a relatively small group or groups over a considerable period of time.” Woodland period sites are widely distributed on many different kinds of landforms, implying the generalized use of a wide variety of habitats for settlements as well as foraging pursuits. Without a more fine-grained Woodland period chronology for Mossy Grove culture sites in East Texas, which we are a long way from achieving, it is not possible to evaluate suggestions by Corbin that there were subtle shifts on the landscape of peoples that may have been a response to changes in subsistence (i.e., the possible growing of cultivated plants). The absence of cultigens other than squash from Woodland contexts in the Lake Naconiche paleobotanical record casts some doubt on the assertion that horticultural economies were developed during this time locally, although the number of flotation and fine-screen samples from pre-A.D. 800 contexts is still miniscule. Thus, the virtual absence of cultigens from Woodland times does not yet constitute a robust evaluation of Corbin’s suggestion. The development of sedentary life along Naconiche Creek appears to have taken place after ca. A.D. 800 by successful hunter-gatherer foragers and pottery makers, specifically amongst the earliest Caddo residents of the valley. Neither the adoption of pottery or the adoption of horticultural subsistence strategies (i.e., the cultivation of maize) appear to have been triggering events that led to the ability of these people to maintain multi-seasonal residences in the same places.


INTRODUCTION
Sometime around ca. A.D. 800, Lake Naconiche sites were no longer occupied by Woodland period groups of the Mossy Grove culture ( Figure  1) solely making sandy paste pottery or living as mobile hunting-gathering foragers. At this time, from ca. A.D. 750-800 to around A.D. 900 (see Perttula and Nelson 2004:Figures 4 and 5), colder and drier conditions began to dominate the local weather. After ca. A.D. 800, were the aboriginal groups Caddo peoples or acculturated Mossy Grove folks? Some findings from the Lake Naconiche (Figure 2) archaeological investigations at the Boyette site (41NA285) are relevant to this issue of ethnic affiliations and local, but nevertheless regional momentous, cultural changes.
Putting that in context, as best as can be discerned in the archaeological records of the Woodland period occupations at the Naconiche Creek (41NA236) and Boyette sites (Perttula 2008:646-650, 663-668, 674-680), if there is any evidence of increasing sedentism, it is only apparent after ca. A.D. 400 or perhaps even as late as ca. A.D. 650, during the latter part of the period. Even so, these occupations were not sedentary in the sense of them being year-round occupations (as with the Caddo settlement history at Lake Naco niche) or even multiseasonal occupations. The sites do not have accumulations of midden deposits, there is no evidence for the construction of sturdy wood structures, and there are only a very modest assortment of burned rock, pit, or post hole features at the Woodland period sites. It is hard to disagree with Story's (1995:237) characterization of Woodland period settlements in the general area that they reflect "intermittent encampments by a relatively small group or groups over a considerable period of time." Woodland period sites are widely distributed on many different kinds of landforms, implying the generalized use of a wide variety of habitats for settlements as well as foraging pursuits. Without a more fine-grained Woodland period chronology for Mossy Grove culture sites in East Texas, which we are a long way from achieving, it is not possible to evaluate suggestions by Corbin (1998) that there were subtle shifts on the landscape of peoples that may have been a response to changes in subsistence (i.e., the possible growing of cultivated plants). The absence of cultigens other than squash from Woodland contexts in the Lake Naco niche paleobotanical record (see Dering 2008) casts some doubt on the assertion that horticultural economies were developed during this time locally, although the number of flotation and fine-screen samples from pre-A.D. 800 contexts is still miniscule. Thus, the virtual absence of cultigens from Woodland times does not yet constitute a robust evaluation of Corbin's suggestion.
The development of sedentary life along Naconiche Creek appears to have taken place after ca. A.D. 800 by successful hunter-gatherer foragers and pottery makers, specifically amongst the earliest Caddo residents of the valley. Neither the adoption of pottery or the adoption of horticultural subsistence strategies (i.e. , the cultivation of maize) appear to have been triggering events that led to the ability of these people to maintain multi-seasonal residences in the same places.

THE CASE OF THE BOYETTE SITE
The Boyette site has archaeological deposits that are relevant to the discussion of Caddo origins. Our work here consisted of extensive block excavations (Block I and II) on an upland ridge toe slope, and small alluvial terrace above Telesco Creek (Figure 3a-b ); the site covers ca. 1.2 acres (Perttula 2008:181-209).
The relevant characteristics are as follows: first , there are radiocarbon-dated features and  archaeological materials at the site that fall in the general temporal interval of interest, ca. A.D. 800-850-as well as immediately before and after that time-when the Caddo cultural tradition is generally acknowledged to become recognizable in the archaeological record in East Texas (cf. Story 2000). primarily to be from a substantial Late Woodland period occupation. The plain sandy paste rim sherds from Block I are almost always from direct or vertical walled vessels (92%). There are a few rims with inverted (4%) or everted (4%) rim profiles. About 71% have rounded lips, 21% have flat lips, and two others are beveled (either towards the interior or exterior vessel wall surface). Another has a rounded, but exterior folded lip-commonly seen in Caddo pottery vessels from the site-and another has a rounded but pointed lip. Vessels range from 11-34 em in orifice diameter. The mean orifice diameter of these vessels-most likely cooking jars and bowls-is 18.4 ± 4.6 em, generally medium-sized on average. The plain sandy paste pottery vessels from the Boyette site have relatively thin walls and a rounded, thick base. Rim walls on average range from 6.75-7.16 mm; vessel walls are on average 6.92-7.18 mm in thickness, indicating the manufacture of smoothed and uniform vessel wall contours of medium thickness, well-suited to cooking use. Base sherds of sandy paste pottery range from 10-10.87 mm in thickness at the site, with the thickest bases among the plain sandy paste sherds from the Early Caddo component in Block I. About 16-18% of the vessel sherds have been smoothed or floated on interior and/or exterior vessel surfaces. The smoothing was done before the vessel was fired, and while the clay paste was malleable.
There is not much difference between Woodland and Early Caddo components in how the sandy paste pottery vessels were fired by aboriginal potters at the Boyette site (Table 3), indicating a technological continuity. Between 59.4-66.7% of the sherds are from vessels fired in a reducing environment, although in the earlier component more vessels were apparently left to cool in the fire rather than pulled from it to be cooled in the open air. Firing in an oxidizing or incompletely oxidizing environment was not the preferred firing method during either archaeological component.
Sherds that are from vessels that were smothered, sooted, or possibly reheated comprise between 11-14.0% in the two components. The relative frequency of these firing conditions in vessel sherds is comparable to that documented from the plain sandy paste sherds in Block II at Boyette.
There are 245 plain sandy paste sherds from Block II at the Boyette site (Table 4), including 23 plain rims and four base sherds, all from rounded base vessels. The highest proportions of sandy paste sherds occur below 50 em bs.
The rim sherds (both plain and decorated) have direct (91%), inverted (4.5%), and everted (4.5%) rim profiles. One has a beveled lip, 50% have a rounded lip, another 42% have flat lips, and one other has a pointed lip. The mean orifice diameter of the sandy paste vessels from Block II at the Boyette site is 19.0 ± 3.63 mm, with a range of 13-29 em. Medium to large-sized vessels were used in Woodland period times at the Boyette site. About 55% of the vessel sherds have been smoothed on interior and/or exterior surfaces, smoothing occurring a bit more commonly on the vessel exterior (57%) compared to the vessel interior (54%). Less than 1% have been burnished, and 2% have remnants of organic residues preserved on them.
The majority of the plain sandy paste sherds from Block II are from vessels fired in a reducing environment, regardless of depth (62.7-70%); most of these were subsequently cooled in an open or oxidizing environment (Table 5). Incompletely oxidized vessel sherds are notably more abundant in the 0-50 em component, while sherds from thoroughly oxidized vessels are more abundant in the probable Woodland period deposits below 50 em bs ( Table  5). The distinctive smudged, sooted, or reheated vessel sherds account for 12.5-15% of all the vessel sherds from the various block contexts, comparable to the sandy paste vessel sherds from the Block I sandy paste sherds from the Boyette site, and the tempered wares from the deepest archaeological deposits in Block II.   notched rim sherds is telling with respect to the likely age of the Woodland period occupation in Block I, in that lip notched rims appear to be more abundant in pre-AD. 300 contexts at Lake Naconiche (Perttula 2008:433 ). Incised sandy paste vessel sherds are more common in the lower Block I archaeological deposits, while incised-punctated vessel sherds are more common in the upper Caddo component (see Perttula 2008: Table 7-4). The incised sherds are primarily from vessels decorated with a series of parallel-

Decorative Element
Incised parallel incised-V-shaped line parallel incised, broad line broad opposed incised lines diagonal incised broad curvilinear incised line horizontal and diagonal incised Subtotal Puncta ted small circular puncta ted rows curvilinear circular punctated rows tool punctated rows Subtotal Incised-Puncta ted straight incised line-circular punctated row broad incised line-triangular zone filled with large circular punctations incised line-circular puncta ted zone horizontal, circular, and panel incised zones filled with tool punctates broad incised line and tool punctations Subtotal

Rocker stamped
Totals probably horizontal-lines, although opposed, curvilinear, diagonal, and horizontal-diagonal elements are also present ( Figure 4). This is not much different than the incised sandy paste sherds in the overlying Caddo component.
Punctated sandy paste sherds include circular as well as tool punctated elements. These punctations are arranged in either straight or curvilinear rows (see Figure 4). The predominance of circular punctations is also characteristic of the punctated sandy paste sherds from the overlying Caddo component.
The incised-punctated sherds have either circular/semi-circular, triangular, or paneled incised zones filled with tool punctations or large circular punctations (see Figure 4). Sixty percent of the (see Brown 1998), dated from ca. A.D. 100-300 in the lower Mississippi valley; Girard (2008 personal communication) suggests that Troyville pottecy may date as late as the peliod of ca. A.D. 400-700 in the lower Mississippi valley, and thus the occurrence of this sherd at the Boyette site would not be out of place in a late Mossy Grove context. The Block II decorated sandy paste sherds (n=l6) at Boyette include incised (43.8%), incisedpunctated (37.5%), lip notched (12.5%), and incised-rocker stamped (6.3%). The decorated sherds complise only 6.1% of all the sandy paste sherds from the Block II excavations at Boyette.
The incised sandy paste Woodland sherds have straight-line elements. This includes single straight broad lines (n=l), single straight V-shaped lines (n=3), rim sherds with broad but shallow horizontal and vertical incised lines (n=2), and broad straight and diagonal incised lines (n= 1).
Among the incised-punctated sandy paste sherds from Block II, the designs consist of straight incised lines forming tiiangular zones filled with punctations of various sorts. Punctations used as filler include tool (n=3) and circular (n=2) punctations. One body sherd-from a sandy paste carinated bowl-has straight incised lines with one row of circular punctations alternating with a row of tool punctations.
The one incised-rocker stamped sherd (40-50 em bs, probably Marksville Stamped, var. Troyville) has a broad and shallow incised line, probably part of a curvilinear zone filled with rocker stamping. The lip notched rims (both from below 50 em bs) have shallow opposed notches along the lip.
Early Caddo sandy paste sherds There are 103 decorated sandy paste sherds recovered from the Block I excavations at the   Boyette site. More than 75% of these distinctive sherds were found from 0-50 em bs in an Early Caddo occupation; three others were found on the surface in the West Block Extension, and one came from a feature ( Table 7). The proportion of decorated sandy paste sherds in the overall sandy paste sherd assemblage recovered in this component is 11.9%, a good bit higher than in earlier Woodland period ceramic assemblages from Block II at Boyette or at the Naconiche Creek site (41NA236) in pre-AD. 400 contexts. This proportion of decorated sherds among the sandy paste sherds in the upper 50 em of Block I is still twice as low as is documented in the tempered Caddo wares from the site (ca. 24%).
There are a few-and almost assuredly mixed or incorporated into the overlying Caddo component from the underlying Woodland period deposits-incisedrocker stamped sherds (5%), one r ocker stamped sherd (1.3%), and one lip notched sherd (1.3%) .
The incised sherds are dominated b y straight-line motifs, either parallel, horizontal, opposed, diagonal (see Figure 4g), cross-hatched in a few instance, or a combination of horizontal and diagonal lines, all probably on vessel rims and/or upper vessel bodies. The same range of incised sherd decorative elements have been documented in the tempered Caddo wares from the Boyette site, but not in the same proportions. Among the sandy paste incised sherds, there is a much lower proportion of cross-hatched decorations (although the relative frequency of cross-hatching is not much different than is documented among No. 2 1 1 4 1 1 80 the tempered Caddo wares from Block II) , as well as lower amounts of both horizontal and diagonal incised decorative elements in the sandy paste sherds from Block I. Most of the incised lines are narrow and V-shaped in profile, although about 17% have broad incised lines. Two incised sherds have curvilinear incised elements, roughly comparable in prop01tions (6.4%) to the tempered Caddo incised sherds from Block I. One incised body sherd has deep and narrow zigzag incised lines (not duplicated among the tempered incised sherds from the Boyette site); its interior thickened body suggests this sherd is part of a sandy paste bottle.
Among the incised-punctated sandy paste sherds, there is also a wide variety of decorative elements. Most consist of straight, diagonal, or opposed incised lines (occasionally broad-lined but mostly narrow and V-shaped) that have created triangular or circular zones filled with different sorts of punctations on vessel rims. Triangular punctatedfilled zones are most common (see Figure 4a-b), although there are circular punctated-filled zones on a few sherds; circular punctations typically filled these incised zones. Again, these characteristics of the sandy paste incised-punctated sherds from the Boyette site are basically the same seen on the tempered Caddo incised-punctated sherds, although the frequency of curvilinear-circular zonedincised-punctated sherds are much less common (6.7 %, compared to between 29-49% of all the incised-punctated tempered Caddo sherds) among these sandy paste decorated sherds. Furthermore, fingernail and linear punctated-filled zones are absent among the sandy paste incised-punctated sherds, and the frequency of tool punctated elements (60% of the incised-punctated sherds have tool punctations) is considerably higher in the sandy paste sherds from Block I at the Boyette site.
One distinctive sherd (with a suspension hole) has broad and deep incised lines above two or three curvilinear rows of large circular punctations (see Figure 4h). A large suspension hole has been drilled through one of the curvilinear punctated rows. Another sherd has a single straight incised line with closely nestled small tool punctations on either side of the line; it does not appear to be part of a punctated-filled triangular incised element.
The punctated sherds include tool (see Figure  4d), cane, and circular punctated (also probably made with a tool, but the circular punctated marks are sufficiently distinctive to warrant a different categorization) elements. Most of the Block I punctated sandy paste sherds have circular punctations (6 1.5% ), either small or large in execution; the Caddo punctated sherds are mostly made with a tool that was triangular-shaped on its end, although circular punctated elements are characteristic of the Early ceramic set defined from the Boyette Caddo decorated sherds. The large circular punctations are deeply tool impressed-probably with a cane tool-causing a raised ridge of clay inside the punctation itself. The small circular tool punctations are pin-prick-sized (preserving the impression of the small tool head) and usually occur in narrow rows; these latter punctations have no counterpart in the tempered Caddo punctated sherds. Another notable difference between the sandy paste punctated sherds and the tempered puncta ted sherds is the absence of both fingernail or linear punctations among the former.
In summary, while incised, punctated, and incised-punctated decorative elements are present in both the sandy paste sherds (from 0-50 em bs) and the tempered Caddo sherds from the Block I ceramic assemblage at the Boyette site (see Table 7), the two different assemblages do not have the same proportion of specific elements or motifs. Although sample size differences may play a role in the fact that there are considerable proportional differences between the two assemblages-or they may be in fact stylistically different (and hence temporally different?)-nevertheless the same decorative decisions were made by the potters that decorated the two wares. That is, among the incised sherds, simple straight and geometric designs were preferred; the punctated sherds were decorated most commonly with straight rows of punctations executed with a tool; and incised-punctated sherds usually had triangular incised zones fi lled with punctations. Usually, the incised zoned were filled with triangular punctations. This suggests that both wares were made during the Early Caddo occupation in Block I at Boyette.
Among the decorated sandy paste sherds from the Boyette site are a few larger sherds where vessel forms could be determined. Most appear to be from straight-walled vessels-probably jars and bowls with rounded bases-but there is at least one bottle sherd and several sherds from carinated bowls (one with rows of small circular punctations and another with opposed incised lines). The same kinds of vessel forms were noted in the tempered Caddo decorated wares.

Early Caddo tempered ceramics
The Early Caddo ceramics at Lake Naconiche are distinguished by engraved fine wares (Figure Sab) and incised, punctated, and incised-punctated utility wares. Among the rims from bowls and carinated bowls, Holly Fine Engraved is only present in Block II at the Boyette site (Table 8), and is certainly the most distinctive engraved ware in the Early Caddo ceramic set. Also in Block II, other common rims have sets of horizontal lines, diagonal lines, vertical and horizontal lines, or broadly excised horizontal and vertical engraved lines. These latter rims are from a vessel with a non-tempered sandy paste, suggesting there is a temporal relationship between this early engraved element and the continued use of sandy paste pottery, which is otherwise being made and used for plain or simple decorated vessels in East Texas up until the 9th century A.D. or later.
Early set engraved rims from Block I include a wide variety of decorative elements, primarily geometric designs (i.e., diagonals and opposed lines) as well as sets of horizontal lines (see Table 8), but geometric and horizontal engraved decorations are characteristic of Lake Naconiche engraved wares from the earliest to the latest prehistoric Caddo ceramics. More distinctive engraved rim elements include cross-hatching and hatched zones (oriented in diagonal, curvilinear, and vertical directions on the rim), as well as cross-hatched and hatched pendant triangles and a circle and cross ( Figure 6). Although the low number of engraved rims fro m Block II precludes definitive conclusions, it is interesting to note the absence of hatched engraved rims in these archeological deposits, but their relative frequency in the Block I engraved sherd assemblage (see Table  centimeters a 0 1 2 3 4 centimeters b Figure 5. Engraved rim and body sherds from the Boyette site: a, Block I; b, Block II. One of the rims has deep! y excised horizontal and diagonal lines, all enclosed within a rectangle (see Figure 61). This particular rim is from a sandy paste non-tempered vessel. A larger rim of the same vessel was recovered from Unit 3 in the test excavations at the Boyette site (Perttula 2002: Figure 4.107a).
Early Caddo engraved body sherds from Block II include hatched pendant triangles (Figure 7c, f) and Holly Fine Engraved (Table 9). Engraved body sherds in Block I are dominated by cross-hatched, hatched zones, and sherds with curvilinear elements, as well as large pendant triangles. Holly Fine Engraved sherds comprise 6% of the engraved body sherds.
In addition to these distinctive Early Caddo engraved body sherds, a goodly number also have simple straight or geometric elements, including horizontal lines, parallel lines, opposed lines, diagonal lines, and vertical lines. These body sherds comprise 28.1% of the Block I engraved body sherds (see Table 9).
The The bottle sherds from Block I at the Boyette site, also part of the Early Caddo ceramics, primarily have sets of curvilinear engraved lines, but both Hickory Engraved and Holly Fine Engraved bottles comprise part of this distinctive engraved assemblage (see Table 1 0). Less common, but still apparently diagnostic of the Early Caddo ceramic set are semi-circles and panel and negative oval elements (see Table 10).

Boyette site vessel
The one vessel from a funerary context at the Boyette site is a Holly Fine Engraved globular bowl from Feature 13A in Block II (Figure 8a). The engraved decoration is confined to the rim, and consists of sets of 12 large triangular panels around the rim filled with diagonal engraved lines that are pitched in opposite and alternating directions from one triangular panel to the next (Figure 8b

Utility Wares at Boyette
Among the incised utility wares, the decorative elements that differentiate the Early Caddo from later Caddo ceramic assemblages are the more common use of cross-hatching (including cross-hatched and horizontal sets of lines), diagonal incising on vessel bodies, and the occurrence of various Dunkin Incised motifs (Table 11) on both rim and body sherds (Figure 9d, h and Figure 1 Ob ). Cross-hatched rims comprise between 16.7-32.1 % of the rims from both blocks at the Boyette site, and 2.3-2.4% of the incised body sherds have diagonal incised lines. In later Caddo ceramic assemblages at Lake Naconiche, only 3.7-6.7% of the incised rims are cross-hatched; diagonal incised rims are much more common (Perttula 2008: Table 7-10).
The punctated rim and body sherds at the Boyette site are characterized primarily by a wide variety of decorative elements (Table 12), among them being the ubiquitous tool punctated row element. However, the most distinctive punctated elements a in the Early Caddo ceramics compared with later punctated utility wares is the more common use of rows of fingernail ( 40% of the punctated sherds, but less than 20% at each of the other Lake Naconiche sites), linear, and circular and small circular punctations as decorative elements (Table 13). Another distinctive punctated decorative element is the use of free or randomly spaced tool punctates on the vessel body.
In the case of the incised-punctated decorative elements, the Early ceramic set includes as diagnostic Weches Fingernail Impressed, var. Weches (see Figure 9b and Figure lOa), incised triangles filled with tool punctations, and in Block IT at Boyette, rims with horizontal incised lines above rows of tool punctates; the latter two incised-punctated decorative elements are also present in later Lake Naconiche assemblages (Table 14). In Block I, there also are a considerable proportion of rims with curvilinear or circular incised zones filled with linear or tool punctates (see Figure 9c  and semi-circles filled with punctations (see Figure  9a); these resemble designs seen on Crockett Curvilinear Incised vessels. One Early Caddo rim has cross-hatched incised lines forming diamonds filled with punctations (see Figure 9f). The incised-punctated body sherds in the Early Caddo ceramic assemblages from the Boyette site share one decorative stylistic tendency: the use of circular or curvilinear incised zones filled with punctates (Table 15): between 28.6-49% of the body sherds from the Boyette site blocks have this distinctive decorative element on utility wares. Both blocks at the Boyette site also have many triangular incised sherds filled with punctates: in the case of Block I, most of them are filled with tool punctates (as is also the case with the later Lake Naconiche incised-punctated ceramic assemblages), while fingernail punctates and cane punctates were more often employed to fill these incised zones in the Block II ceramics.
Incised-punctated sherds with incised lines either above or below rows of tool or fingernail punctates are particularly common in Block II at the Boyette site. As with the decorated rims, Weches Perttula-Lake Naco niche Archaeology And Caddo Origins Issues 83 Fingernail Impressed, var. Weches sherds are present in the incised-punctated body sherds characteristic of the Early ceramic set.

CONCLUSIONS
The Early Caddo occupations at the Boyette site (41 NA285) appear to be contemporaneous with the earliest Alto phase component at the George C. Davis site on the Neches River, dating as the latter does from the mid-9th century A.D. However, the fine wares and the utility wares found there do not suggest that the Boyette site is a component of the Alto phase, although such sites have been identified in the Angelina River basin (see Story 2000: Figure  5). Story (2000:20) has previously pointed out that "components of this phase are no where common even though some of the diagnostics, such as Weches Fingernail Punctated and Holly Fine Engraved, have wide distributions." Such appears to be the case here, because while there are a few sherds of Holly Fine Engraved and Weches Fingernail Impressed in the Boyette site decorated sherds, they do not dominate the decorated sherd assemblages-along with Davis Incised, Dunkin Incised, Crockett Curvilinear Incised, Pennington Puncta ted-Incised, Hickory Engraved, and Duren Neck Banded-as they do as the George C. Davis site (Stokes and Woodring 198l: Table 24). For example, Stokes and Woodring (198l:Table 24) note that Holly Fine Engraved vessel sherds and Weches Fingernail Punctated sherds comprise both between 16-41% of the more than 14,000 decorated sherds from mound and domestic contexts across the site, and incised-punctated Crockett Curvilinear Incised and Pennington Punctated Incised sherds are also fairly well-represented (2-19% by excavation areas) at this mound center. Only a handful of sherds from the Boyette site were identified as coming from either Holly Fine Engraved or Weches Fingernail Impressed/Punctated vessels. Less than 13% of the sherds at the Boyette site have incised-punctated decorative elements, although between 30-50% of these have curvilinear zoned incised and punctated elements, few of which remotely resemble in execution Crockett Curvilinear Incised vessels. At best, then, the few similarities in vessel decorations in both fine wares and utility wares between the Boyette site and the well-known George C. Davis site are indicative of contemporaneous Caddo occupations-and perhaps even a modicum of contact/interaction-but they do not belong to the same Caddo groups. Instead, the Boyette site is apparently a component of a local and culturally separate Caddo community in the upper Angelina river basin, one that is currently taxonomically unidentified.
One question that languishes unanswered is the cultural relationship between the latest M ossy Grove sites in East Texas and the earliest Caddo sites in the region. Concerning the historical traditions of the Alto phase Caddo peoples that lived in this general area, Story (2000 :25) has commented that "there are no earlier archeological remains in the middle and upper Neches River basin that can plausibly be identified as an antecedent complex to the Alto phase component [at the George C. Davis site]." She goes on to speculate that "earlier Caddoan developments [earlier than the late A.D. 800s] must have taken place elsewhere, probably to the northeast in either the Sabine or Red River basins." These suggestions go hand in hand with the notion that the George C. Davis Caddo mound center represents a founding colony in a prut of East Texas that was not previously within the territory occupied by Caddo peoples. Corbin (1989:121) also subscribes to the notion that the Caddo occupation of East Texas, or at least those areas south and west of the Sabine River, originates outside of East Texas and that the Caddo were newcomers to the region. He also proposed that the Caddo populations who had come into the area in the A. D. 800s lived coevally with the East Texas Woodland peoples (i.e., the Mossy Grove Culture peoples) who were already there, and that these Woodland peoples continued "their dispersed lifestyle, only slightly displaced on the landscape, with the additions of maize, better pottery and the bow and arrow." Under this scenario, the: indigenous Woodland population was acculturating and modifying some of what was early Caddo culture into their own lifestyle to create a post-early Caddoan culture we call Late Caddo on a cultural base that was already in place and never disappeared. The only place where early Caddo blinked into almost instantaneous existence in this area was at a few specific sites (Corbin 1989:124).
Probably the only means to fully evaluate the relationships between, and cultural affi liations of, the Mossy Grove Woodland period groups and the earliest Caddo archaeological sites will be exten-sive bioarchaeological and genetic studies of DNA, oxygen/strontium isotopes, and genetic markers preserved on human skeletal remains, as these provide the highest probability of establishing cultural and biological relationships between various groups of people recognized in the archaeological record. It is doubtful, however, that this can ever be fully achieved because to date not a single Woodland period burial has ever been found in a domestic site in East Texas, and 9th century Caddo sites are almost as rare, whether with burials or not. Furthermore, it is an open question whether DNA or traces of genetic markers left on human skeletal remains are even preserved in any such sites occupied on the cusp of the Woodland to earliest Caddo time periods.
That being said, the prehistoric occupations at the Boyette site are nevertheless relevant to the questions and scenarios posed by Story (2000) and Corbin (1989). From radiocarbon dates obtained in the excavations here, as discussed above, there are two occupations of interest, one that dates (with a 95% probability) from cal A.D. 667-847 (with a mean calibrated intercept from three dates of AD 743) and the other that dates (with a 95% probability) between cal AD 873-1075, with a mean calibrated intercept (from four dates) of AD 960. Two of the four dates have a mean calibrated age range of AD 750-990 (calibrated intercepts of AD 880 and 900), while the other two have a mean calibrated age range of AD 995-1160 (calibrated intercepts of AD 1020 and 1040). These radiocarbon ages suggest that the second occupation may be represented by two different episodes of settlement.
The earlier of the two occupations (identified in the deepest archaeological deposits in Blocks I and II) has sandy paste Goose Creek Plain and decorated Mossy Grove ceramics, Gary and Kent dart points (and probably some early arrow point forms), a few features, but no evidence of structures, middens, burials, or use of cultigens. In most respects, this early occupation at the Boyette is a fairly typical Mossy Grove period occupation. There is one (to some) troubling aspect in the material culture of this component: decorated sandy paste vessel sherds are apparently atypically abundant (more so than any other known Mossy Grove component, unless all of them have moved by bioturbation from overlying Caddo archaeological deposits, which is unlikely), and the incised, incised-punctated, and incised decorative elements almost eerily presage the same ceramic vessel decorations noted in the later ca. AD 985 component. In the latter occupation, these styles of vessel decoration are common on both sandy paste and tempered pottery wares. There are at least a few examples of non-traditional vessel forms in the ca. AD 743 component, including carinated bowls and a bottle. Such vessel forms are well represented in the later ca. AD 985 component, as they are in post-AD. 1100 Naconiche Caddo ceramic vessel assemblages.
What about the ca. AD 985 component: is it affiliated with the Caddo or is it an acculturated Mossy Grove site? First, it can be noted that cultigens are absent in the archaeological deposits associated with this occupation, although the numbers and arrangements of features suggest that this occupation was a relatively sedentary one as there is evidence of sequential central hearths from two different houses that date to the earlier of the two later occupational episodes (i.e., ca. A.D. 750-990 from radiocarbon, but centering around A.D. 880-900). There are stemmed arrow points in the assemblage, including those of the Alba type (the dominant type in the Alto phase) and an abundance of ceramic vessel sherds, including many from carinated bowls and bottles that have engraved designs (i.e., Holly Fine Engraved and Hickory Engraved) much like those noted from other early Caddo contexts in the region. Much of the pottery is tempered-primarily with grog-but sandy paste pottery remains an important part of the ceramic vessel assemblage in this later component. More importantly, the sandy paste pottery in this late 9th-early lOth century occupation is commonly decorated with the same decorative elements common in the tempered wares, even including some amount of engraved sandy paste pottery.
From the evidence at hand from the Boyette site at Lake Naconiche, it appears that there were changes in material culture-the use of temper in the manufacture of pottery vessels, subtle changes in vessel form, and innovations in pottery vessel decoration-that were either underway by the mid-8th century and/or had been adopted by the aboriginal peoples living along Naco niche Creek by the late 9th century, a period of some 150 years (or at least six generations). The choice and inspiration to decorate sandy paste pottery cannot be laid at the feet of any Caddo colonizers from the George C. Davis site or others of its ilk because this was taking place at least one century before the appearance of that site on the Neches River. The same may be said for the appearance of carinated bowl and bottle forms in late Mossy Grove contexts at the Boyette site. Such innovations as these appear to have devel-oped amongst Mossy Grove groups without having to invoke a Caddo "influence," if that is relevant in the context posed by Story (2000) and Corbin (1989). Perhaps it is plausible that further ceramic innovations such as adding temper to the paste of vessels, or even choosing to decorate a vessel after it was fired rather than while the vessel still had a wet paste, were not beyond the creative reach of the people that lived along Naconiche Creek or in other areas of East Texas.
Thus, in the end, and based on admittedly very sketchy archaeological information, I do not view the 9th century occupation at the Boyette site as acculturated Mossy Grove groups influenced by the superior culture of the Caddo who were expanding into the area. Rather, I view the 7th and 8th century population at Lake Naconiche as directly antecedent to the 9th century population that lived at the Boyette site. In most particulars, the preponderance of archaeological evidence from this later occupation indicates that the population that lived there was Caddo, or at least one of many different groups living in East Texas that can be considered Caddo in an ethnic sense. The 7th and 8th century A.D. Woodland population is considered ancestral to the Caddo. This does not mean that the Woodland or earliest Caddo populations in the Attoyac Bayou basin had any ethnic or underlying genetic relationship with the founding population at George C. Davis-that issue still remains to be teased out. It does mean that the George C. Davis Caddo population was not the only one in East Texas in the 9th century A.D.