NINETEENTH-CENTURY ALIBAMU-KOASATI POTTERY

Aspects of southeastern Indian cultures were maintained through centuries of assimilation of European material culture. Unlike many tribes who discontinued the manufacture of traditional pottery, some Alibamu and Koasati groups produced traditional wares during their migrations from the Tennessee and Alabama region, through Mississippi, into Louisiana and Texas. Native ceramic technology and maize-oriented culinary traditions continued to be used by them and related tribes such as the Choctaws, Creeks, Seminoles, and Chickasaws into the twentieth century. Pottery vessel assemblages from selected household and village contexts are used to model gender-based aspects of subsistence, foodways, and culture change and continuity among the Alibamus and Koasatis.

Many Southeastern Native American groups continued to manufacture their ceramics well into the nineteenth century, despite long-distance migrations from original heartlands (Williams 1981). The maintenance of traditional ceramic technology varied among Indians: some tribes assimilated all types of European vessels, while other groups continued to make pottery that runs the technological and stylistic gamut from traditional wares to colono-wares for European household use (Henry 1980; Holmes 1903; Hunter 1994; Saville 1920; Speck 1907, 1925, 1928: 394-425; Vernon and Cordell 1993; Waselkov 1989: 60). Other Native American groups and many individuals completely assimilated European goods and became acculturated to European lifestyles.

Ethnological collections of Native American pottery vessels and Works Progress Administration interviews on file with the Oklahoma Historical Society indicate that traditional pottery technology and maize-based culinary traditions of various native and immigrant tribes lasted well into the nineteenth century (Perdue 1993; Schmitt and Bell 1954). The Jennie McElrod photograph collection shows the lifeways and pursuits of rural Native Americans in Oklahoma in the 1920s, and illustrates that native ceramic, wooden, and woven vessels were still employed in various traditional methods of preparing maize (e.g., hominy, sofkee, or osafki) among the Creeks, Choctaws, Chickasaws, and Seminoles; food residues (i.e., wood ash lye from soaking sakes) can be seen adhering to the interiors of some large vessels (Gettys and Watkins 1984:113-129; Wright 1958:156). Large flaring-rim jars and large carinated bowls were apparently the dominant vessel types and are those most frequently preserved in ethnographic collections (Schmitt and Bell 1954). Small native clay serving vessels, storage containers, and direct-fire cooking vessels were replaced by European ceramic, glass, and metal vessels among most Oklahoma tribes sometime during the nineteenth century.

The continuance of aboriginal ceramic technology and culinary traditions long after the introduction and ready availability of English- and American-made ceramics, in a milieu of rapid cultural change, is an important measure of ethnic stability in southeastern Native American societies. The material culture evidence of this stability was carried west with the migration of traditionalist tribal members. Among the Alibamus and Koasatis, female roles in the production of pottery and the preparation of maize perpetuated the maintenance of traditional pottery vessel forms and decorative motifs. Pottery vessel assemblages from selected Alibamu-Koasati archaeological sites can be used to measure these gender-based aspects of subsistence and foodways. Other types of material culture from the selected sites have been examined, but are not the focus of this paper.

In this paper we examine nineteenth-century culinary traditions of two tribes, the Alibamus and Koasatis, as evidenced by a single artifact class: pottery. The Koasatis were known as the "White Cane" people, and were closely allied with the Alibamus or "Medicine Gatherers," migrating separately to Louisiana and Texas, but yet remaining in close proximity to each other (see Flores 1977:55). The Alibamu-Koasatis were primarily allied with the French during the eighteenth century. Following the Treaty of Paris in 1763, they, some Creeks (e.g., Pakana Muscogee), Apalachees, and other southeastern tribal fragments, began a series of migrations into Mississippi, Louisiana, and Texas. Some Koasatis who remained in their heartland along the Tennessee River of northern Alabama were apparently assimilated among the Cherokees (Jacobson 1954, 1974; Jacobson et al. 1974). This mixing of ethnic groups created changes in pottery technology and style among some peoples, such as the Apalachee, and consequently the interpretation of ethnicity on the basis of ceramic material culture from the archaeological record must be critically examined.

Our research focuses on the ceramic assemblages from two known Alibamu-Koasati archaeological sites representing the westward migration and attempted resettlement of these peoples. These sites were selected after examining the collections from many sites reputed to have been occupied by immigrant or native Indians in Louisiana, Arkansas, and Texas. There are problems with data quality; both sites were surface collected with minimal controlled excavation, spatial control is basically lacking, and the representativeness of each site's artifact sample as a whole is questionable. Nevertheless, we selected these two archaeological sites because of the quality of the historical identifications, the presence of large native pottery assemblages, and the relative cultural homogeneity of the collection (i.e., limited admixture with earlier prehistoric and later historic occupations) .

The Carolina Bluffs #3 site (16BO207) on the Red River in northern Louisiana contains a native pottery "behavioral" household assemblage, primarily an inuse domestic activity set (e.g., DeBoer 1983:21; Pauketat 1987:12) with several reconstructible vessels typical of Creek sites in Alabama (Knight 1987). The Carl Matthews site (41PK2), near the Trinity River in the coastal plain of Southeast Texas, seems to represent primarily a "discard" assemblage (see DeBoer 1983:21), although two whole vessels were found that may be from either a "behavioral" household assemblage or grave lot(s). Thus, these two sites represent a partial spectrum of the range of Alibamu-Koasati archaeological contexts in Arkansas, Louisiana, and Texas.

The continued maintenance of various traditional maize dishes and associated aboriginal ceramic vessel forms reflects the cultural resilience of the Alibamu and Koasati groups despite rapid cultural change, displacement, and intertribal amalgamation in the seventeenth and eighteenth centuries. The routes of Alibamu and Koasati migrations across the southern United States have been closely investigated and are important to claims for restitution (Jacobson 1954, 1974; Marsh 1974; Martin 1974). However, the determination of ethnicity and the level of acculturation is often difficult, if not impossible, to ascertain in the archaeological and ethnohistorical records. The present study provides preliminary evidence that may allow more refined identifications of immigrant Native American subsistence agriculture and foodways through key aspects of their ceramic technology.

Native American material culture, subsistence strategies, and polities changed through progressive acculturation. Some traditionalist groups, however, maintained vestiges of their past cultural lifeways in various hinterland and core cultural areas. According to Waselkov (1993:125), historic Creek culture was essentially "a synthetic one incorporating both European trade goods and traditional artifacts, values, and activities in a new, distinctive, stable cultural format." The Creek polity that formed after the Yamassee War in 1715-1716 was noted for its assimilation of disparate cultural and linguistic groups, which included the Alibamus and Koasatis.

The pressure of the Euro-American colonial frontier on Native Americans to become "civilized," and to be "removed," led to a class struggle (e.g., traditionalist versus immersionist) within tribal and pantribal alliances and the encroaching Euro-American frontiersmen. These political and economic conditions were the driving forces of the frontier between Native American and Euro-American populations in the Southeastern U.S. (see Braund 1993; Drinnon 1972, 1980; Henri 1986; Usner 1992). These conditions also impelled some southeastern Indians to favor cooperation with Euro-Americans, and for others to reject cooperation, with the latter tending to move ahead of the advancing frontier into the Trans-Mississippi South.

The parallel migrations of the Alibamu-Koasatis, and other groups such as the Pakana Muscogees (Jacobson et al. 1974), from Alabama to Texas is indicative of a close, shared kinship, and cultural point of view. Fairbanks (1962:64-68) was among the first to point out the gender-specific conservatism among the Creeks; this was later elaborated by Mason (1963:6873). The early Alibamu-Koasati immigrants to Louisiana and Texas likely then reflect more traditionalist segments of these tribes. The persistence of ceramic technology provides a spatially and temporally traceable measure of these same gender relationships, as well as the character of subsistence agriculture and foodways among the traditionalist Alibamu-Koasati immigrants.

This group was led by Echean, and their settlement consisted of loose, linear clusters of houses, termed "Upper" and "Lower" villages. Echean was reported by Francois Grappe or "Touline" in 1806, and by John Sibley in 1807, to be an Alibamu-Koasati chief (see Abel 1922:82-84; Flores 1984:149). In 1806, Echean, along with the Caddo chief Dehahuit, allowed the Freeman-Custis Expedition to pass into Spanish territory (Flores 1977:64). Since Francois Grappe, or Touline, was a prominent friend of the Caddos (see Drinnon 1972:178, 183, 227; Lee 1989) his identification of Echean as the recognized chief of the Alibamus and Koasatis is believed to be accurate.

The Red River branch of the Alibamu-Koasatis may have been part of the Alabama-Koasati migration of ca. 30 families led by Red Shoes, Pia Mingo, and Billy Ashe to Bayou Rapides, Louisiana, from Mississippi Territory in 1793. Part of this community was led by Red Shoes to the Sabine River in Texas ca. 1800. In 1804, John Sibley was petitioned by 30 AlabamaKoasati, apparently led by Echean, who did not leave the Bayou Rapides settlement with Red Shoes, to relocate approximately 200 miles upriver in Caddo Territory. The six to eight Alabama and Coushatta families noted by the Freeman-Custis Expedition in 1806 in the vicinity of Carolina Bluffs were reported by Francois Grappe to have moved from Bayou Rapides ca. 18031805 (Flores 1984:149). Chief Echean lived several miles down the river from the main Coushatta village (16BO173 [McCrocklin 1988, 1990]; Flores 1977:55, 1984:149).

Carolina Bluffs is one of eight identified archaeological sites in what is assumed to be the "Lower Coushatta Village" governed by Echean (McCrocklin 1988, 1990). Some of these sites appear to have been occupied by Alibamu-Koasatis until the 1830s, and several contain subsequent Euro-American occupations. As noted above, 16BO207 was selected because the archaeological collection is the most homogenous in terms of the age of historic Euro-American artifacts, and native ceramics, of all of the likely AlibamuKoasati Red River sites reported and collected by McCrocklin (1988, 1990).

Carolina Bluffs was discovered, recorded, and partially excavated in 1988 by members of the Kadohadacho chapter of the Arkansas Archeological Society and the Northwest chapter of the Louisiana Archaeological Society. This project focused on the relocation of Alibamu and Koasati sites reported by the FreemanCustis Expedition of 1806 (Flores 1977, 1984). Nine archaeological sites (16BO 172, 16BO 173, 16BO175, 16BO176, 16BO185, 16BO186, 16BO187, 16BO205, and 16BO207) recorded during this project contained distinctive early nineteenth-century Native American material culture remains along with silver ornaments, glass beads, iron tools, and flintlock ordinance items assumed to be derived from the reported Alibamu-Koasati occupations (McCrocklin 1988,1990). These habitation sites extend several miles along the bluffs of the Red River upriver from Shreveport, Louisiana. Horizontal-log houses were described for this group (Flores 1977, 1984; McCrocklin 1988, 1990 ), and archaeological evidence (e.g., fired clay and mud-dauber nests with stick and board impressions) for this architectural style was recovered from several of the sites.

Figure 2. Location of Carolina Bluffs #3 (16BO207), Bossier Parish, Louisiana.

A metal detector was used by McCrocklin to locate isolated and clustered metal artifacts and to define the possible locations of structures based on the recovery of square nails (McCrocklin 1988, 1990). A single Native American pottery concentration (Feature 2) was excavated at Carolina Bluffs. Shovel probing and steel rod probing were used to define other Native American pottery concentrations (Figure 2). Spatial provenience is based at best on feature locations; the location of general finds is not known.

Three features were identified at Carolina Bluffs by McCrocklin (1988, 1990), who postulated the presence of a former building some 5 m south of these features and another possible building about 61 m to the north (Figure 2). Feature 1, identified as a shallow midden area, received some shovel probing but no excavations, and yielded glass beads, silver ornaments, iron arrowheads, and flintlock ordinance. Feature 2, encountered with a steel probing rod, was exposed in a 1.5 x 1.83-m excavation. It consisted of a dense concentration of Native American pottery sherds with polishing stones, copper, and dark green bottle glass (McCrocklin 1988:37). Feature 3, a possible burial, yielded silver ornaments, a silver staff ornament, glass beads, carbonized oak and hickory nut fragments, and peach pits. Other artifact concentrations-presumably pottery sherd concentrations- were also located with a steel probing rod, but were not excavated.

The material culture remains from Carolina Bluffs included hand-carved silver ornaments common in the late eighteenth and early nineteenth century, faceted blue glass beads (dating ca. 1820-1825 in Louisiana [Donald G. Hunter, personal communication 1994]), iron knives and kettles, axe heads, late eighteenth-early nineteenth-century flintlock gun parts, and fired clay from wattle-and-daub houses or stickand-clay chimney chinking. Charred hickory nuts, acorns, and peach pits also were recovered from the site, concentrated in Feature 2 with the native ceramics. This suggests that the rendering of oils from plants was a culinary mainstay. Instead of a specialized activity site, the material culture indicates a short-term household assemblage. As noted above, the assemblage was not systematically excavated; however, the ceramic sherds from the site were recovered from meaningful contexts and appear to be from a culturally homogenous deposit.

The Carl Matthews site was discovered by a local farmer in 1929 after two ceramic vessels eroded out of the Long King Creek bank following a flood. University of Texas archaeologist A. T. Jackson (1933) explored the site in 1933 and described it as an "extensive Indian campsite" on a hillside above the creek. He found quantities of pottery sherds on the plowed site surface in association with European goods such as hammered copper and brass, bottle glass, and lead musket balls (Perttula 1994). Jackson conducted limited excavations to find the source of the pottery vessels, uncovering an area of charcoal and square nails along the creek bank that may have been from a burned log building, but he abandoned the effort after encountering two "historic" burials near the "campsite" (Perttula 1993, 1994).

Long King's village was described in 1831 by Jose Francisco y Madero, a land commissioner for the Mexican province of Coahuila y Tejas (Bexar Archives 1831). It was known as the Upper Village of the Koasatis along the Trinity River and had more than 60 families. Madero described a "large house destined to religious cult" along with rectangular wood or log buildings and circular houses built with wood, cane, and clay chinking. These circular houses may represent traditional Creek house styles.

The native pottery was selected for analysis since it was more systematically collected than other artifact types by both McCrocklin (1988) and Jackson (1933). European artifacts and food remains assumed to be associated with the Alibamu-Koasati are also briefly summarized. The key aspects of early nineteenthcentury Alibamu-Koasati ceramics from the two sites are described below in terms of vessel form(s) and function(s), rim decorative treatment, and ceramic paste/temper. Metrical information is provided for native vessel form classes at Carolina Bluffs and for rim-sherd mode classes at Carl Matthews. These data characterize traditionalist Alibamu-Koasati ceramic household and village assemblages and provide an initial model for interregional site comparisons.

Vessel batches were defined for the Carolina Bluffs ceramic assemblage because it is thought to represent a "behavioral set" of vessel types related primarily to a single Alibamu-Koasati household. The ceramics from Carl Matthews are primarily from a surface collection and probably represent a "discard assemblage" from an entire village. The two complete vessels "washed from the creek bank" at Carl Matthews may have been from grave lots or from an unidentified "behavioral" context.

The dominant artifacts (n = 614) from Carolina Bluffs are items produced by the Alibamu-Koasati: ceramics (n = 317), and triangular iron arrow points (n = 8). Also common were EuroAmerican items (n = 206) used by the Alibamu-Koasati. Food remains (n = 50) and bald cypress wood fragments (n = 4) were also present. Finally, non-Alibamu-Koasati (i.e., Caddoan) remains (n = 29), such as cobbles, flakes, fire-cracked rock, and grog-tempered and incised ceramics round out the assemblage. Due to lack of flotation and systematic screening, the charred hickory nuts, acorns, and peach pits in the assemblage are eliminated from further statistical comparisons, leaving 531 artifacts in the analytical assemblage.

Euro-American artifacts include 214 items, dominated by firearms or weapons-related items (n = 65; including eight iron arrowheads, a flintlock cock, and two "British" [gray chert] gunflints); Alibamu-Koasati/southern Euro-American architecture items (n = 53; including 36 stick-impressed daub and 17 mud dauber nests); and personal items (n = 46; including 28 beads, 17 silver ornaments or fragments, and one button). Tools or tool fragments (n = 25), Euro-American architecture items (n = 18), and Euro-American household artifact fragments (n = 7) round out the assemblage.

It is notable that no European ceramics were recovered from Carolina Bluffs, although they were recovered from all other Alibamu-Koasati sites along the Red River (McCrocklin 1990). There is no evidence that the Arkansas Archeological Society/ Louisiana Archaeological Society investigations at Carolina Bluffs were conducted any differently than at the other sites examined along the Red River. Since the range of European metal items recovered included machine cut nails, firearms parts, knives, kettles, silver ornaments, and some vessel glass fragments indicative of a typical frontier household material culture assemblage, the lack of European ceramic storage or tableware vessels may not be solely due to recovery or sampling problems. Instead, native clay vessels probably still functioned for storage and serving needs in the household, to the exclusion of imported ceramics.

Alibamu-Koasati Ceramics. Native Alibamu-Koasati ceramics comprise the dominant artifact class recovered from the site as a whole, as well as 59.6 percent (n = 317) of the analytical assemblage. These sherds are primarily temperless, although some sherds and one vessel have a fugitive temper thought to be shell, based on the platy structure of the voids. Pastes include very little sand or grit. There are no evident coil breaks.

Table 1. Rim, Neck, and Body Diameters (cm) and Thicknesses (mm) and Physical Attributes for Large Flaring-Rim Jars, Chattahoochee Roughened var. Chattahoochee from 16BO207.

Vessel Number	Rim Diam.	Rim Thickness	Rim Treatment1	Neck Diam.	Neck Thickness	Body Diam.	Body Thickness	Surface Decoration2	Interior Color	Exterior Color
1	22	6.2	r, af	19	9.2	36	7.5	br	7.5YR5/6-4/6	7.5YR5/6-4/6
2	18	6.2	r, p, af	20	6.7	-	8.0	br	10YR7/3-3/2	10YR7/3-3/2
13	26	7.5	r, af	21	8.2	-	9.0	br	10YR5/4	10YR5/4
14	30	9.2	r, af	22	8.1	-	6.0	br	10YR5/2	10YR5/2
15	24	9.2	r, f, af	22	9.5	-	-	-	10YR5/4	10YR5/4
37	-	-	-	18	8.5	36	8.0	br	5YR3/3	5YR3/3
16	26	-	r, f	20	-	36	-	-	10YR5/4	10YR5/4

¹ r = rolled; p = pinched; f = flattened; af = appliqué fillet.
² br = brushed.

All ceramic sherds were sorted according to color and vessel portion. Rim and body sherds then were sorted by surface treatment and closely examined for conjoins or identical decorative or surface attributes. This allowed the determination of individual vessels of specified form, paste, temper, and decoration. Approximately 20 percent of one vessel, and substantial rim, shoulder, and upper body segments of other vessels were reconstructed from the sherd assemblage.

Figure 3. Alibamu-Koasati ceramics from Carolina Bluffs: (a) deep flaring-rim jar, Chattahoochee var. Chattahoochie; (b) rim sherd, deep flaring-rim jar, Chattahoochee var. Chattahoochie; (c) hemispherical bowl; (d, e) carinated bowls (drawn by Mitch Stokley).

Four forms were identified in the assemblage of 37 minimum vessels: deep flaring-rim jars (Figure 3a, b); hemispherical bowls (Figure 3c); carinated bowls (Figure 3d, e); and flaring-rim bowls. By sherd count, deep flaring-rim vessels dominate the assemblage (n = 277; 87.4 percent), followed by hemispherical bowls (n = 23; 7.3 percent), carinated (non-incised) bowls (n = 12; 3.8 percent), and flaring-rim bowls (n = 5; 1.5 percent. The deep flaring-rim jars resemble the Chattahoochee Roughened, var. Chattahoochee (brushed) ceramic type (Knight 1987:201). These vessels have rolled rims, pinched and flattened lips, smoothed necks, and lightly (multidirectionally) brushed bodies.

Table 2. Rim, Neck, and Body Diameters (cm) and Thicknesses (mm) and Physical Attributes for Hemispherical Bowls from 16BO207.

Vessel Number	Rim Diam.	Rim Thickness	Rim Treatment1	Neck Diam.	Neck Thickness	Body Diam.	Body Thickness	Surface Decoration2	Interior Color	Exterior Color
31	18.5	7.0	p, f	22.0	8.0	22.0	10.0	sm	5YR4/4	5YR4/4
32	20.5	7.0	p	22.0	8.5	-	-	-	5YR5/4	5YR5/4
22	18.0	5.0	p	18.5	6.0	-	6.5	sm	5YR6/8	5YR4/1
26	16.0	6.2	f	-	7.0	-	-	-	5YR5/8	5YR5/8
27	16.0	6.2	p	-	-	-	-	-	5YR5/6	5YR4/6
28	23.0	6.0	f	-	-	-	-	-	10YR3/2	10YR3/3
29	22.0	7.0	f	-	-	-	-	-	10YR5/8	10YR5/8
30	21.0	6.2	p, f	-	-	-	-	-	5YR5/8	10YR5/8
33	23.0	5.5	p	-	-	-	-	-	5YR5/6	10YR6/6
34	23.0	5.5	p	-	-	-	-	-	10YR6/8	5YR6/8

¹p = pinched; f = flattened.
² sm = smoothed.

The orifice diameter and rim thickness, neck diameter and thickness, body diameter and thickness, rim and surface treatment, decoration, and interior and exterior colors were determined for seven deep flaring-rim jars (Table 1). These same attributes were measured for four hemispherical bowls that possessed suitable curvature for these determinations. The interior and exterior colors were recorded for an additional six vessels (Table 2).

Among the deep flaring-rim jars (Figure 3a), the orifice diameter ranged from 18 to 30 cm (mean = 24.3 cm); the rim thickness ranged from 6.2 to 9.2 mm (mean = 7.7 mm); the neck diameter ranged from 18 to 22 cm (mean = 20.3 cm); the neck thickness ranged from 6.7 to 9.5 mm (mean = 8.4 mm); and for three measurable vessels the body diameters were 36 cm, and the body thickness ranged from 7.5 to 9.0 mm (mean = 7.7 mm). Rim treatments included six rolled rims with applique fillets below the rim with fingeror dowel-impressed notches, one pinched rim, and two flattened rims.

The hemispherical bowl sherds were more fragmentary, and fewer dimensions could be recorded. None were brushed. The orifice diameters measured for 10 vessels ranged from 16 to 23 cm (mean = 20.1 cm), and the rim thickness ranged from 5.0 mm to 7.0 mm (mean = 6.2 mm). Rim treatments included five pinched, two pinched and flattened, and three flattened.

No measurements were calculated for carinated bowls or flaring-rim bowls (Figure 3d, e) since all sherds were small. None of these were brushed.

Summary and Comparisons. The large sherds from Feature 2 indicate that large ( > 24 cm orifice diameter) and small (<22 cm orifice diameter) flared rim jars (i.e., Chattahoochee Roughened, var. Chattahoochee [brushed] and similar forms without applique´ fillets and brushed bodies) were employed in the household, suggesting use of the pots for both storage of liquids (smaller) as well as communal (larger) serving (see Hally 1983, 1984; Pauketat 1987). Bowls were relatively uniform, with 16-23 cm orifice diameters, likely indicating serving functions. Orifice diameter and vessel size estimates were not possible for flaringrim bowls and carinated bowls.

There are no known pottery assemblages from the Tennessee and Alabama region that have been firmly associated with traditionalist Alibamu-Koasati groups. It is notable that the four basic vessel forms found at Carolina Bluffs are also found on Creek sites in A1abama. The Tukabatchee site in Alabama has received extensive investigations, defining "discard" assemblages from Creek households with pottery technology and use similar to those of the Alibamu-Koasati (e.g., Knight 1987). Many of the attributes of the Carolina Bluffs ceramics-the large flaring-rim jar vessel form with decorated strip fillets below the rim, sandy pastes and occasional crushed bone or shell temper; carinated bowls (cazuelas); and flared rim bowls-are also known from Alibamu-Koasati sites in Texas, including Carl Matthews (Perttula 1992, 1994).

A Tallapoosa phase (1750-1800) assemblage of 62 vessels from Tukabatchee consisted of 64 percent deep flaring-rim jars,20 percent carinated bowls (including incised varieties), 2 percent hemispherical bowls, and 3 percent Mississippian jars (Knight 1987:109, 121123). By contrast, a Late Tallapoosa phase (1780-1836) assemblage of 89 vessels from Tukabatchee consisted of 44 percent deep flaring-rim jars, 37 percent carinated bowls (with a significant decrease in incising), 9 percent hemispherical bowls, and 7 percent flaringrim bowls (Knight 1987:141, 156-164). In vessel form frequency, the Carolina Bluffs ceramic assemblage falls between the Tallapoosa phase (1750-1800) and the Late Tallapoosa phase (1780-1836) ceramic assemblage from Tukabatchee.

Both mold and coil techniques were used by Creek women in the production of pottery (Bartram 1789[1987]:32). Brushing of the vessel body was common, and red filming was present in Creek ceramic assemblages in Alabama (Knight 1987), near the supposed homeland for the occupants of Carolina Bluffs. The Tuckabatchee ceramics are dominantly temperless or claytempered wares, having applique fillets below the rims with fingernail or pinched notching. The incised sherds from Carolina Bluffs do not resemble the incised motifs usually found on carinated bowls on either Atasi phase (1630-1650), Tallapoosa phase (1750-1800), or Late Tallapoosa phase (17801836) sites (Vernon J. Knight, personal communication 1990), and may instead be Red River Caddoan types.

As previously mentioned, two whole native ceramic vessels were recovered from the creek bank, whereas the sherds were collected by A. T. Jackson from a cultivated field "not far" from where the vessels were located. Since the ceramic collection was probably from the entire village area, and in a surface context, the analysis focused on the recognition of vessel forms through attributes of rim morphology, surface treatments, decoration, and tempering agents.

Vessels. The whole vessels are a hemispherical bowl (or cazuela) and a flaring-rim jar, common late eighteenth-early nineteenth-century Alibamu-Koasati vessel forms. The plain hemispherical bowl (Figure 4a) appears to have been constructed from a single, large lump of clay, rather than coil-constructed; the rounded to flat base is undifferentiated from the remainder of the vessel. It has a fine sandy paste, but no apparent temper. It appears to have been fired in an oxidizing environment, as indicated by its orange to light orange interior and exterior color. This vessel also has numerous exterior fire clouds from direct exposure to fire, probably from cooking use (see Hally 1983, 1984).

Figure 4. Vessels recovered from the Carl Matthews site (41PK2): (a) hemispherical bowl; (b) deep flaring-rim jar (drawn by Mitch Stokley).

Both interior and exterior walls of the bowl have been burnished, particularly near the rim of the vessel. It stands 13.2 cm in height, with a 5 mm thick inverted rim that is 1.5 cm in height. The lip is flat and 4 mm in width. The orifice diameter is 15.4 cm, somewhat smaller than the hemispherical bowls from Carolina Bluffs. Although the vessel is generally well constructed and finished, it does not fully stand on its own; rather, it tilts when set on a flat surface.

The flaring-rim jar (Figure 4b) is quite heavy and durable, and also appears to have been constructed from a large clay lump. Scraping and possible paddling marks are visible on the exterior surface near the base, and these may represent evidence for vessel shaping before firing. The vessel surface is roughened near the base, but otherwise undecorated. Horizontal scraping marks are visible on the interior along the upper vessel walls, possibly from use or manufacture.

The vessel is brown (lOYR5/3) to reddish yellow (7.5YR6/8) in color, with numerous fire clouds on the exterior body and rim. Such fire clouding is probably indicative of the vessel's use as a cooking jar (see Hally 1983, 1984; Skibo 1992). It has no obvious temper, but does have a medium sandy paste. The jar stands 15.6 cm high, the rim is 2.0 cm in height, and the vessel thickness is 6 mm at the lip. The lip itself is rounded and 2 mm in width. The orifice diameter is 14.2 cm.

This flaring-rim jar clearly resembles in form and paste the Chattahoochee Roughened var. Chattahoochee (brushed) type from late Tallapoosa phase Creek sites in Alabama (Knight 1987:201), but lacks the applique fillets commonly seen on these types of vessels. Two of the seven large flaring-rim jars from Carolina Bluffs also lacked applique fillets, however.

Sherds The sherd assemblage from Carl Matthews is comprised primarily of plain and decorated sandy paste sherds from well made jars and bowls, and possibly from one bottle. Visible tempering aplastics in the sample include grog, bone, and shell (Table 3). Temper was present in 13 percent of the sherds, with crushed bone and shell about equally represented. The untempered sherds (n = 58) listed in Table 3 are separated from the broader sandy paste group because of their distinctive black core (fired in a reduced environment) with very fine sandy paste. These sherds are from hemispherical bowls.