An Early Caddoan Period Cremation from the Boxed Springs Mound Site (41UR30) in Upshur County, Texas, and a Report on Previous Archaeological Investigations

Cite this Record Perttula, Timothy K.; Wilson, Diane E.; and Walters, Mark (2000) "An Early Caddoan Period Cremation from the Boxed Springs Mound Site (41UR30) in Upshur County, Texas, and a Report on Previous Archaeological Investigations," Index of Texas Archaeology: Open Access Gray Literature from the Lone Star State: Vol. 2000, Article 23. https://doi.org/10.21112/.ita.2000.1.23 ISSN: 2475-9333 Available at: https://scholarworks.sfasu.edu/ita/vol2000/iss1/23

this situation to a certain extent, and also spur renewed professional archaeological interest in this very significant prehistoric Caddoan mound center.

THE CREMA TEO REMAINS FROM THE BOXED SPRINGS SITE (41UR30). by Diane E. Wilson
The analysis of human remains from the Boxed Springs site (41 UR30) was limited to an isolated cremated burial in a large cemetery at the northern endl of the site. Ceramic vessels believed to be associated with the burial (see Perttula, below) suggest a date of ca. A.D. 900-1100 for the cremation. Although fairly common during the Woodland period (see discussion in Perttula 1999:387), cremated burials' are occasional occurrences at Caddoan sites, with the predominant form of burial being single extended and supine burials. During Early Caddoan times, however, crematiofil was apparently more comnno111 than in later times.
In the analysis of cremations, data collection is limited by preservation conditions, but they do provide valuable information on mortuary practices. The osteological examination reported here was conducted using standard techniques as outlined in the Texas A&M University, Physical Anthropology Laooratory Data Form for human remains, and those techniques presented in Buikstra and UbeJ,aker (1994). Observations were. made by hand with the aid of a 1 OX hand lens.
The bumed remains include isolated but relatively complete skeletal elements, but the majority were highly fragmented materials. When possible, bone identification and sex. and age information were recorded for that element. Isolated elements and larger fragments were also identified and inventoried. In addition, the degree and pattern of burning was recorded during the analysis.
From these remains it was determined that the cremation was that of a single individual. This individual was probably an adult male whose body was burned at a low temperature while the bones were articulated and in the flesh.
Most of the burned remains were reduced to splinters and extremely small fmgments of bone. The total weight of the cremation was 335.2 grams. The largest fragment was a nearly complete right talus, which measured approximately 55 mm in length. This suggests that the individual was male (Steele 1976), although it should be cautioned that. the talus was incomplete and had been burned and was visibly cracked longitudinally and transversely. Other identified fragments were: a left mandible fragment, a right patella fragment, four rib fragments, one vertebral neural arch fragment, three phalanx fragments of the feet, two metatarsal fragments, one fibula fragment, 28 long bone fragments, and four cancellous bone fragments. No teeth were present. The size and epiphyseal surfaces present suggest that the individua• was an adult.
None of the bone had a blue-gray calcined appearance. Instead, the bone was a pale tan color on the exterior surface with a black, charcoal-like appearance on the inner surfaces. This color patterlil indicates that the individual was burned at a very low temperature, likely as low as 200 degrees Centigrade (Buikstm and Ubelaker 1994). The l'ight tan color of the exterior surface of the bones also suggests that the bones were covered by nesh at the time of burning. Differential burning is evident, with little burning on the larger fragments, whereas most long bone fragments are more heavily charred. This again is typical of burning with flesh present of a primary articulated individual, mtller than after disarticulating.
Larger bones are predominantly cracked in a transverse plane, which is consistent with burning when the bone is green or covered with flesh. Cracking on bone surfaces was extensive. Very little warping was observed on the fragments, but most fragments were too small to observe this feature. Rootlets were present with the remains, but there was no evidence of root etching. No other biological alteration was observed on the larger fragments, and there was no evidence of pathological conditions, disarticulation, cut marks, or depressed fractures. It was not possible to take metric or non-metric observations.

ARCHAEOLOGICAL CONTEXT OF THE CREMATION, by Timothy K. PerUula
Little specific information is available on the archaeological context of the cremation reported on by Diane E. Wilson. It is known that the cremation was a single burial feature dug in a large cemetery at the northern end of the site (north of a large metal farm building north of areas #4 and #5 on Figure 2); there is no evidence available that suggests that the cremated individual was burned in place; rather, it appears that the individual was cremated elsewhere (on the site?), and the burned remains were gathered up for burial, along with two (or three?) complete ceramic vessels placed in the burial pit as funerary objects with the cremation. Vessels apparently found with the cremation included a Spiro Engraved bowl and an East Incised bowl, along with a Red River style long-stemmed pipe.

PREVIOUS INVESTIGATIONS AT THE BOXED SPRINGS SITE, by Timothy K. Perttula and Mark Walters
The Boxed Springs site was originally identified by a Mr. Norman Bedgood of Tyler, Texas, while he was flying his airplane along the Sabine River. He was a member of the East Texas Archeological Society (Ef AS), and he informed Sam Whiteside, a fellow ETAS member, who undertook archaeological investigations in 1962. At the time, the site was owned by the Ledsinger family, from Kilgore, Texas, who maintained a fishing camp on the property.

Whiteside Investigations
In 1962, Sam Whiteside excavated in two mounds at the Boxed Springs site, and collected a limited amount of Caddoan artifacts from the surface in the vicinity of several of the mounds. In his 1962 notes. he referred to the mounds by numbers (i.e., #2, #3, #6, and #7), and his excavations took place in mounds #3 and #7. To be consistent with subsequent archaeological investigations at the Boxed Springs site by Dr. Dee Ann Story and the University of North Texas (see below), and with the provenience information on the collections housed at the Texas Archeological Research Laboratory, The University of Texas at Austin (T ARL), hereafter the mounds will be referred to by letters, as follows:     . The human remains were poorly preserved in Burial I , and the soil around the bone and artifacts showed no evidence of a burial pit, suggesting the burial may have been placed on the mound as it was being built.
The top of the mound had apparently been previously disturbed by pothunters, as there were numerous looter holes in it, and the human remains in Burial I appeared to have been disarranged. One piece of baling wire was found near Burial 1, and this was probably discarded by the pothunters.
Whiteside's notes indicate that few artifacts were present in the mound fill, other than pieces of lithic debris. The absence of sherds in the mound fill suggests that this part of the Boxed Springs site (including the borrow pit area just east of Mound A, see Figure 3) had not been previously occupied by Early Caddoan groups before the construction of Mound A.
Continued excavations in Mound A indicated that the mound was a maximum of 6 feet (1.83 m) in height and 35 x 43 feet (10.7 x 13.1 m) in length and width (see Figure 4).
With the removal of all the mound fill with the tractor blade, a large circular ring of clay was exposed near the base of the mound (see Figures 4 and 5). This ring of red sandy clay was 0.6 m in width and 16 em thick, and marked a berm around a large clay-and sandfilled pit that extended approximately 0.9-1.2 m below the original ground surface under the mound~ the top of the berm was between 0.6-1.2 m below the top of the mound ( Figure  7).
Excavations of the pit proceeded with shovel and trowel to about 30-40 em above the pit floor, and the remainder of the pit was carefully troweled until the pit floor was reached. The pit was 12.5 x 13.5 feet at the top (3.8 x 4.1 m), and narrowed to 10.5 x 11.5 feet (3.2 x 3.5 m) at the bottom (see Figure 7), and had rounded comers and slightly sloping walls. The upper and central portion of the pit had a white sand fill, with the remainder of the fill and berm grading from a sandy red clay to a very dense red clay just above the floor. The pit floor was covered with a 2.5-5 em thick lens of bluish-gray clay, probably obtained from Pleistocene alluvial deposits along the Sabine River.
Early Caddoan burials in mounds at the George C. Davis (F134 in Mound C, Story 1997) and Mounds Plantation sites (Burial Pit 5, Mound 5, Webb and McKinney 1975) also had clay berms around the centrally-placed burial pits, and they also contained multiple individuals. The burial pits were much larger at these two sites than Burial 2 at Boxed Springs: 7 x 5.5 m at George C. Davis (Story 1997: 18) and 5.8 x 5.2 m at Mounds Plantation (Webb and McKinney 1975:54). At both these sites, the burial pits were covered with a log tomb; no direct evidence for a log tomb was recovered from Burial 2 at Boxed Springs, but the depressed lens of white sand in the center of the pit (see Figure 7) suggests that the pit feature may have been left open for a time, or covered with a wood frame, and the sand may have washed into the top of the pit.
At George C. Davis, Fl34 had eight extended adult burials, but they were oriented northsouth, with their heads facing south (Story 1997: Figure 12). There were extensive burial offerings (Story 1997:21-23). The Mounds Plantation burials in Burial Pit 5 were placed in five groups of extended burials (totaling 21 individuals, including three fetuses), with four groups oriented east-west, with the head facing eac;t (Webb and McKinney 1975: Figure 5). A wide variety of funerary offerings were placed in Burial Pit 5 at Mounds Plantation, including ceramic vessels, ceramic pipes, arrow points, chipped stone lrnives, wood artifacts, cane matting, stone beads and pendants, as well as other items.
The pit feature at Boxed Springs, Burial 2, contained three sets of teeth enamel along the eastern side of the pit; no other human remains were p!ieserved itn the featUlie (Figure 8).
The broad spacing of the sets of teeth enamel suggest that Burial 2 contained the bodies of three individuals that were laid out east-west, with the head facing west. Most of the artifacts placed on the floor of the burial pit were along the north side of the feature, north of the third set of teeth enamel, although three of llhe clusters of arrow points were in the central and southern part of the feature, and one ceramic vessel (Vessel 5, a plain bottle) was near the eastern pit wall (see Figure 8). Funerary objects placed in Burial 2 included two large chipped stone bifaces, five groundstone celts, several polished stones, a ferruginous sandstone tool, four dusters of armw points, and seven cemmilc vessels (V. I through V. 7, see Figure 8).
The large and wcl1-made chipped stone bifaces were found together on the n0rtl'lem side of the pit, and immediately north of the five groundstone celts (see Figure 8). Bod11 appear to have been made from non-local cherts--probably from Central Texas--and one may have been heat-treated based on its lustrous appearance ( Figure 9). The larger biface (about 16 em in length) is a Gahagan Biface (Shafer 1973:224 and Figure 19), while the other compares well with the Group 2 bifaces at the George C. Davis site (Shafer J973:Figurc 19w).
The ferruginous sandstone tool is a tabular piece about 20 em in length and 4 em in width ( Figure I 0), with straight and smoothed edges. The Boxed Springs tool resembles tabular sandstone saws recovered in Early Caddoan contexts at the George C. Davis site (Shafer 1973:317 and Figure 25h-i). It was placed between Vessels 3 and 4, and near a pom ished pebble (see Figure 8).
The two polished stones are pebble-sized pieces of chert about 3-4 em in length and width ( Figure 11 ). They may have been used to polfslrl ceramic vessels after they had been successfully fired. One of the polished stones was near the femtginous sandstone tabufar tool, and the other was amidst arrow point cluster # 1 in the northern part of the burial pit (see Figure 8). The m.ano included in Figure 1 1 was apparently found in close association with Burial 1 (see above).
The five celts in Burial 2 were found cached together near the northernmost set of teeth enamel, and immediately south of the two large chipped bifaces and Vessel 1, a plaim bottle (see Figure 8). The celts alie small in size, rangilllg from 5-10 em ~n l1 engt.h, and are relatively narrow compared to the width of the bifad ally beveled bit and poll ( Figure 12). On one angular celt, the bit width is only 13 mm (see Figure 12, second from left), while the bit width on the other four specimens ranges between ca. 26-32 mm. The celts were made from non-local raw materials, probably of metamorphic or igneous origin and originating in the Ouachita Mountains of southeastern Oklahoma and southwestern Arkansas, and at least one may be diorite or Hatton tuff (see Figure 12, far left).
As previously mentioned, there were four cl ust.ers of arrow points in Burial 2. Cluster # 1, at the northern end of the pit (see Figure 8), had 55 specimens. They apparently were not on shafts or in a quiver when they were placed in the burial pit, because they were turned in various directions. Whiteside's notes comment that the cluster #l points we.lie found lyi ng on a ferruginous sandstone rock (presumably the tahu~ar ferruginous sandstone piece described above), but his plan map of Burial 2 (see Figure 8) indicate that the tabular ferruginous sandstone piece was about 0.60 m east of cluster# I. Arrow point cluster #2 had four specimens, while cluster #3 had 36 points and cluster #4 had five arrow points. All the cluster #3 points were pointing south and probably were shafted and in a quiver. The direction of the cluster #2 and #4 arrow points was not determined. These three arrow point clusters were placed in the central and southern parts of the pit, in apparent association with the central and southern sets of tooth enamel (see Figure 8).
Stylistically, the arrow points from Burial 2 at Boxed Springs are rather morphologically homogeneous (Figures 13-16). Most have square or parallel stems with squared barbs and flat bases, or have square/parallel stems with concave bases (Alba Several of the square stemmed points have more sweeping barbs, similar to the Catahoula type, and there are a few specimens with a diamond-shaped stem comparable to the Hayes type. A variety of raw materials were used in the manufacture of the Boxed Springs arrow points, including fine-grained and coarse-grained quartzite, local cherts of brown and yellow hue, and high quality cherts that probably originated in the Ouachita Mountains and/or Red Ri vcr gravel 1 s (see Banks 1990).
The seven ceramic vessels in Burial 2 cluster in the northern end of the pit (see Figure 8). Vessel I is a plain bottle, and Vessel 2 is a small (ca. lO em in height) Spiro Engraved beaker ( Figure 17). These vessels were placed between arrow point cluster # 1, the two large chipped bifaces, and the five celts. Vessel 3 is a 13 em tall everted rim jar with a pinched body and a zoned incised-cane punctated decoration Olil the rim ( Figure 18). It was placed immediately north of the tabular ferruginous sandstone tool and 0.60 m east of arrow point cluster #I (see Figure 8).
Vessels 4 and 5 lay along the eastern side of the burial pit, between the pit wall and the central and northernmost sets of tooth enamel (see Figure 8). Vessel 4 is a large plain carinated bowl, and Vessel 5 is a plain bottle with a straight neck and a globular-shaped body ( Figure 19). The bottle. stands ca. 23 em tall.
Vessels 6 and 7 were placed next to each other at the northern end of the Burial 2 pit, between 15-30 em north-northwest of arrow point cluster# 1 (see Figure 8). Both are plain jars.
Not much information is available on the other three mounds at Boxed Springs. Whiteside's notes described Mound B as a .. sand fill md.," but what it covered was not ascertained. There apparently was a borrow pit area immediately to the south of this mound (see Figure 3). Mound C was 35 x 44 feet in dimensions (10.7 x 13.4 m), and apparently covered a house with a prepared clay floor. A 2 x 12 foot trench had been dug by pothunters into the center of the mound sometime prior to I %2, but what was discovered is not known. Finally, Mound D also apparently covered a house with a prepared clay floor, and the mound fill was a 45 em thick zone of sand.
Most of the artifacts collected by Whiteside at the Boxed Springs site have been donated to T ARL. These include all the funerary objects recovered from Burial 2 in Mound A, as well as a few plain and decorated sherds from north of Mound A (Figure 20), near Mound D (Figure 21 ), and from unspecified contexts ( Figure 22). Many of the decorated sherds have multiple horizontal incised lines, and several have impressed triangles below the horizontal incised lines (see Figure 21, cf. Coles Creek Incised, var. Coles Creek). Punctated and pinched rim and body sherds are also present, along with a large cross-hatched incised and punctated sherd from general contexts (see Figure 22, upper left) and a Holly Fine Engraved rim sherd (see Figure 22, second from Iefton bottom row).    Walters also retains a small collection of materials from Mound A and from general contexts. This general collection includes the following items: six unmodified pebbles, one tested cobble, six cores and core fragments, 18 pieces of lithic debris, one bifacial tool fragment, three pieces of burned clay, one parallel brushed ceramic body sherd, one Coles Creek Incised, var. Coles Creek body sherd (see Phillips 1970), two plain rims, and 87 plain body sherds. The lithic artifacts are on local quartzites, petrified wood, and cherts. About 18 percent of the ceramics are bone-tempered, including the one brushed body sherd.
From Mound A, and probably from Burial 1, are nine lithic artifacts; the association is not clear because Whiteside's notes do not indicate that other stone tools were found in association with Burial 1 besides a celt (see Figure 6) and a mano (see Figure II). Among them is a small quartzite polishing stone, two large bifacial tool fragments (probably from Gahagan-style bifaces), and six Alba or Homan arrow points. The bifacial tool fragment'\ are made from gray and dark gray cherts that are not locally available~ one specimen has well-ground tool edges. Four of the six arrow points are made on local red chert and coarse-grained quartzite, and they have expanding stems with convex rounded bases. The other two arrow points are made of a black chert (Big Fork chert) that originates in the Ouachita Mountains (see Banks 1990). They also have expanding stems with convex rounded bases and squared shoulders; one has been broken by an impact fracture.
Of unknown provenience--but probably from Mound A investigations--are 14 sherds from the plain body of a small, thin ( 4.9 mm), and well-made bottle. The sherds have a burnished exterior.  Figure 3). There was also evidence of pothunting digging activities south and southeast of Mound C, near the bluff edge overlooking the Sabine River. Her map also includes a metal farm building to the north of the mound and plaza area (see Figure 23), and the farm building is situated on a large knoll near the apparent northern end of the associated Early Caddoan village.

Sabine River Mound Sites Study
The senior author of this section first inspected the Boxed Springs site during the course of the archaeological survey investigations in the proposed Texas Big Sandy project on Big Sandy Creek in Wood and Upshur counties, Texas (Perttula et al. 1986). He also had an opportunity to visit with Mr. Whiteside and examine the collections and available notes from the site.
In 1988, he returned to the site to document it as part of a larger study of mound sites along the Sabine River in Northeast Texas and Northwest Louisiana (Perttula 1989(Perttula , 1994. At the time, the site was in an overgrazed pasture, with well-worn cattle trails and bare patches across the area. Archaeological materials were noted on the surface west of Mound A and in a midden deposit at the south end of the landform between Mounds C and D (see Figure  23). Pothunting activities were noted south of Mound C, and this work had apparently found numerous ceramic sherds but no obvious features. The landowner also informed us that a large novaculite bifacial tool had been found east of Mound D during the excavation of a septic tank.
A small amount of artifacts were collected from a surface collection west of Mound A and between Mounds C and D. In the first collection was a fragment of a large bifacial tool of gray chert (probably of non-local origin), two core fragments of Ogallala quartzite and tan chert, and two grog-grit-tempered plain body sherds; one of these thin ( 4.2 mm) sherds had been burnished on its exterior surface.
The second surface collection between Mounds C and D had 12 pieces of lithic debris and five ceramic sherds. The lithic raw materials represented among the debris included ferruginous sandstone, yellow chert, red chert, Ogallala quartzite, gray chert, novaculite, and jasper. The novaculite and jasper, and probably the gray chert, are definitely from nonlocal raw material sources (i.e., Red River gravels in northeastern Texas). One of the sherds has fine line horizontal engraving, and it was tempered with finely crushed grog. Two other sherds are flat bases (11-12 mm in thickness) tempered with grog-grit-bone inclusions. The remaining sherds included a 6.5 mm thick grog-tempered plain body sherd and a 5.5 mm thick grog-grit-bone-tempered sherd from a plain bottle.
A single shovel test was excavated in the midden deposits between Mounds C and D, and nine artifacts were recovered. These included five pieces of lithic debris, all of local petrified wood, tan chert, and quartzite), an Alba arrowpoint of Ogallala quartzite, and three grog-tempered ceramic sherds. Two of the sherds are plain, while the third has a small fingernail punctated decorative element.

Documentation of the Red McFarland Collection from the Boxed Springs Site
Red McFarland, from Whitehouse, Texas, dug approximately 150 burials at the Boxed Springs site in 1990. These burials were in a single large cemetery at the northern end of the site, away from the mounds, and the burials were reported to be almost exclusively single, extended, supine burials, with the notable exception of the one cremation reported on here by Wilson.
Very little information is available on any of the burials, either their orientation, depth, or condition of the human remains, and associational data are also sparse. The associational information will be summarized below in the discussion of the funerary objects documented from the Boxed Springs site. The original photographs and documentation notes are on file at the Archeology Division, Texas Historical Commission. It is our understanding that much of the Boxed Springs collection has been sold since it was documented in 1990.
A total of 78 lithic artifacts are in the collection, including large chipped bifaces, dart points, numerous arrow points, a scraper, celts, and a bipolar core. The three large bifaces are Gahagan bifaces; two are chipped from Central Texas chert, and the third is on an unidentified non-local chert. One of the large bifaces occurred in burial association with a plain bowl.
There are three large dart points or knives, all three with contracting stems; one of the tools has been extensively resharpened or reworked, and may have seen use as a knive. This tool is on novaculite, a non-local raw material. One of the other dart points is on a non-local raw material (Red River chert), and the third is from a local raw material.
The bipolar core is on a local chert, probably a small pebble available in upland gravels in the Sabine River basin. The single bifacial scraper is on a tabular piece of locally available petrified wood.
There are five groundstone celts in the McFarland collection. One is on a quartzitic sandstone from the Ouachita Mountains, and a second is on chert; the three others appear to be on locally available hematite.
McFarland found 65 arrow points in the cemetery. In one burial, the arrow points included Alba, Alba or Catahoula, and Agee types. The arrow points include seven Alba or Catahoula, 40 Alba, eight Homan, two Friley, one Steiner, two Agee, one comer-notched (Scallorn?), three contracting stem forms, and one unidentifiable fragment. The arrow points are predominately manufactured from locally available chert and quartzite raw materials (Table 1 ). Only 4.6 percent of the arrow points are on non-local lithics (see Table  1), and these (Ouachita Mountains cherts and siltstone) appear to be from Red River gravel sources (see Banks 1990).  Two of the burials apparently had copper-covered shell earspools, two earspools per burial (for a total of four shell earspools), but other associational information is not available.
There arc a number of copper-covered earspools at the George C. Davis site from F118, F119, F155, and F161, representing shaft tomb interments from Stage I-IV in Mound C.
Copper plated ear ornaments have been reported at Mounds Plantation (Webb and McKinney 1975:103-104 and Figure 14), but in both George C. Davis and Mounds Plantation contexts, the copper covered either stone or wood, not shell.
Vessels, vessel sections, sherds, a long-stemmed Red River pipe, and clay earspools comprise the ceramic artifacts in this Boxed Springs collection. As previously mentioned, the one long-stemmed Red River pipe was apparently found in association with F..ast Incised and Spiro Engraved vessels. The six clay earspools were found in six different burials, one earspool per burial, and one of these burials had a Crenshaw Fluted jar (cf. Durham and Davis 1975:36 and Figure 19) and an incised jar. In another case, a plain carinated bowl and a dish were noted to have been found together in a burial, and in a final example, one burial had six vessels placed with it, including a Hickory Engraved bottle, a plain bowl with a peaked rim, a plain bottle, a plain jar, a horiwntally incised jar, and a small tray. Table 2 lists the 160 ceramic vessels/vessel sections and 23 sherds in this Boxed Springs collection. Plain vessels include 36 percent of the 160 vessels, engraved finewares comprise 44 percent of the vessels, and the incised, punctated, punctated-incised, and pinched wares represent the remaining 20 percent of the assemblage (see Table 2). By comparison, 80 percent of the five vessels in Early Caddoan burial pits at the Mounds Plantation site were engraved (three Holly Fine Engraved and one Hickory Engraved), and the other vessel was an elaborate Crockett-Pennington hybrid (Webb and McKinney 1975).
At the George C. Davis site, I 00 percent of the tO vessels foi!Jnd in Mound C bullmal pits were engraved (Story I 997). Clearly emgraved vessels were particuli arly important ceramic funerary objects in Early Caddoan burial pits in northwestern Louisiana and East Texas.
Early Caddoan burial pits (pits D-F, I-K, M, N, Q, and R) in Mound C at the Crenshaw site (Durham and Davis 1975), on the other hand, had a much lower proportion of engraved wares among the ceramic funerary offerings ( 46 percent, majnly Spiro Engraved, Hickory Engraved, and Holly Fine Engraved), amd a correspondingly higher percemtage of plain ( 16 percent) and incised, incised-punctated, punctated, and p£nched (27 percent) vessels, much like the vessel assemblage in the Boxed Springs non-mound cemetery. The most significant difference between the Mound C vessel assemblage at Crenshaw and the Boxed Springs assemblage is the high proportion of plain bottles and bowl1 s at Boxed Springs (sec Table 2).
However, Early Caddoan village area pottery at the George C. Davis s~te is not domm nateo by engraved sherds. In a large sample described by Stokes and Woodring (1981), engraved sherds represent about 26 percent of several thousand decoliatcd sherds, with the remainder of the decorated sherds comprised of incised, incised-punctated, fingernail impressed, pinched, and neck banded specimens. Thus, engraved vessels weli e used, broken, and discarded in village contexts during Early Caddoan times at this mound center, but the village assemblage is dominated by utility ware jars and bowl 1 s. T fue Boxed Springs vessel assemblage from the non-mound cemetery is also composed of many utility ware jars and bowls, suggesting it was derived from residential or every-day use at the site.  Figure 24a). There are also distinctive Coles Creek Incised vessels with horizontal incised hnes on the rim and! triangular impressions/punctations at the rim/body juncture (see Figure 24b-c and Figure 27f), a Kiam Incised bowl with multiple incised lines on the rim and fingernail punctations on the body (see Figure 27i), and Weches Fingernail Impressed jars and bowls with puuctated bodies and rims (see Figure 24f and Figure 26c). A punctated rim jar and a pinchedpunctated jar are also present in the collection (see Figure 24d-e). One incised-punctated bowl has rim peaks and a design of multiple horizontal imcised lines with small punctations between the incised lines (see Figure 27e).
The decorated bottles are dominated by horizontal engraved bodies and necks from small and large Hickory Engraved specimens (sec Figure 25a-b, d). One Hickory Engraved bottle has multiple curvilinear and concentric engraved lines on the globular body (see Figure 25e), and an unidentified engraved bottle has four horizontal engraved lines below the bottle neck, with sets of semi-circular engraved lines pendant from the lowermost horizontal engraved line (see Figure 25c).
The plain and decorated bowls (primarily with engraved decorations of Spiro Engraved, Holly Fine Engraved, and Hickory Engraved) from the non-mound cemetery have a diversity of forms, including simple bowls with straight walls and direct rims (see Figure  26c, Figure Texas Archaeology No. 12 (2000) 8 A c Figure 28. Decorated vessels from cemetery: a, Incised jar; b, Crockett Curvilinear Incised; c, unidentified engraved carinated bowl. Vessel a is 12.2 em in beigbt. g), globular bowls (see Figure 27d), peaked bowls (see Figure 26b and Figure 27h), and shallow but wide bowls (see Figure 27j and Figure 28c). One fonn resembles a gourd or seed jar (see Figure 26d), and it has an engraved motif along the rim, and deep vertical flutes on the body, resembling Crenshaw Fluted fonns; another fluted fonn from Boxed Springs has a Spiro Engraved motif with fluted channels cut through the engraved motif.

Journal of Northeast
One of the carinated bowls has a series of large engraved pendant triangles (see Figure 26a) along the rim, reminiscent of engraved motifs more commonly seen in Middle Caddoan contexts in the upper and middle reaches of the Sabine River basin. Another, probably a Crockett Curvilinear Incised vessel, has horizontal incised lines on the rim, and vertical and concentric circles on the body between panels filled with small punctations (see Figure  28b). One of the shallow but wide bowls has a unique engraved motif of horizontaJ lines on the rim and a series of panels filled with short interlocking scrolls and smaJI excised circles or dots (see Figure 28c).