Platform mounds, once thought to have been constructed exclusively in the late prehistoric Mississippi period (A.D. 1100–1500) in much of the interior southeastern United States, are now known to have also been built during the preceding Middle to Late Woodland periods (ca. 200 B.C.–A.D. 1100) in regions previously believed to lack them (Anderson and Mainfort 2002:13, 16; Boudreaux and Johnson 2000; Brose 1988; Brown 1994; Jeffries 1994; Johnson et al. 2001; Knight 1990, 2001; Kwas and Mainfort 1986; Lindauer and Blitz 1997; Mainfort 1986a, 1988, 2013b; Mainfort and Walling 1992, 1996; Pluckhahn 1996, 2003; Rafferty 1983, 1987, 1990, 2002; Welch 1998). The Slate Springs Mound (22CA502, also known as the West Mound) in Calhoun County, Mississippi, is now recognized as such a Woodland site. In 1992, test excavation of the mound revealed fill stages and yielded charcoal samples that produced radiocarbon dates ranging from the early to middle first millennium A.D., within the Middle to early Late Woodland periods; in addition, chronologically diagnostic Middle to Late Woodland potsherds (sand-tempered fabric-impressed and cordmarked, grog-tempered cordmarked) were recovered from mound fill and from surrounding fields.
The Slate Springs Mound is unique in that it is currently the only confirmed Woodland-period platform mound in the extensive North Central Hills physiographic zone of Mississippi. The site is also anomalous in that only a solitary mound is present, whereas other Woodland platform mounds tend to occur in groups with other flat-topped mounds and/or with conical burial mounds; and, unlike at a number of such sites, no extraregional ceramics have been recovered. Overall, the Slate Springs site is enigmatic not so much for what is present, but for what is absent.
The intent of this report is to place this significant but little-known site in its context in regional prehistory by presenting the results of limited test excavation of the mound conducted in 1992, and of general surface collecting and shovel testing of the field surrounding the mound in 1991 and 1992. Prior to this report, information about the site was largely unpublished, consisting of the initial survey report (Haag 1952), a National Register of Historic Places nomination document (Baca 1992), and a Mid-South Archaeological Conference paper (Baca 1993). Brief mentions of the site appear in a few previous publications (Carleton 1999:152; Galaty 2008:267; Rafferty 2002:205, 222; Sims and Connaway 2000:221).
The Slate Springs Mound is a relatively small, flat-topped earthen structure (Figures 9.1 and 9.2) located approximately 3 km north of the village of Slate Springs, in southern Calhoun County, Mississippi, within the North Central Hills physiographic province (Figure 9.3). The mound is situated on a first terrace about 100 m east of the right bank of Shutispear Creek, a northwesterly flowing tributary of the Yalobusha River. The site is indicated as an “indian [sic] mound” on map sheet 87 of Soil Survey of Calhoun County, Mississippi (McMullen et al. 1965), in an area mapped as Freeland silt loam, 2 to 5 percent slopes, severely eroded (FrC3).
Figure 9.1. Slate Springs Mound (22CA502), February 27, 1992, looking north. Dewitt Spencer stands at the foot of the mound’s south corner. Photograph by Keith A. Baca.
Figure 9.2. Topographic map of Slate Springs Mound (22CA502).
Figure 9.3. Map of northern Mississippi showing locations of Woodland period platform mound sites, physiographic zones, and National Forest units discussed in text.
The site was initially recorded in the early 1950s by archaeologist William G. Haag, then of the University of Mississippi, during his pre-inundation survey of the Grenada Reservoir in the Yalobusha River valley and some of its tributaries. The results of Haag’s survey are presented in the unpublished report “Archaeological Survey of the Grenada Reservoir in Mississippi” (Haag 1952). Haag named the site the West Mound after Ruble West, the leaseholder of the property, but it has been renamed the Slate Springs Mound (see also Rafferty 2002:205, 222; Sims and Connaway 2000:221) to prevent its confusion with the previously recorded and published West Mounds (22TU520), a Late Mississippian site in the northern Yazoo Basin of Mississippi (Buchner 1996; Dye and Buchner 1988; Phillips et al. 1951:51, 321). Haag’s report briefly describes the site as follows:
Ca 1. The West Mound. A large, truncated-pyramidal mound with associated village. Located about two miles north of Slate Springs . . . U.S. [Army Corps of] Engineers Grenada Reservoir map F8. Although the site lies within the Skuna [actually Yalobusha] drainage on Shutispear Creek, its elevation is 262 feet . . . well above the flood pool. It is of some significance, though, to an understanding of the entire Reservoir and warrants inclusion in the survey. The mound is quite square, ninety feet on a side and rather uniform in height, twelve feet. For an area of perhaps 1000 feet radius about the mound sporadic artifacts are seen. At the time of the survey all the surrounding area was in pasture, so that but few potsherds and flints were revealed. Six sherds of Tishomingo Cord-marked and three sherds of Tishomingo Plain were found, and one side-notched, bevel-edged, projectile point [Haag 1952:1–2].
In his summary analysis of the sites recorded during his survey, Haag (1952:27) comments that “[t]he truncated pyramidal mound is typical of those in the Mississippi Embayment except that it is not very large and occurs alon[e] rather than in a group.” Subsequently, he notes that “only in the extreme eastern end of the reservoir [area] does a truncated pyramidal mound appear (Ca 1). Because of its relatively small size and occurrence alone, this mound probably represents an early invasion of Temple Mound concepts” (Haag 1952:29). In his concluding recommendations for further archaeological work, Haag states that “[t]he pyramidal mound, Ca 1, need not be considered here for it is far outside the flood pool limits and will serve as a reserve site for future excavation by some interested agency. It may be said here, though, that this site would probably be most revealing as to the time and condition of the introduction of Temple Mound I ideas and practices” (Haag 1952:30).
Following Haag’s survey, the Slate Springs Mound was subjected to no further archaeological investigation for almost 40 years. Then in 1991, the Mississippi Department of Archives and History (MDAH) selected the site for fieldwork to obtain data for its nomination to the National Register of Historic Places. Accordingly, the author, then a staff archaeologist with MDAH, made two reconnaissance visits to the site in June 1991, and subsequently directed a test excavation and a topographic mapping survey of the mound in January 1992. The mound was listed on the National Register in December 1992. During the course of the fieldwork, artifacts were also collected from two newly discovered nearby sites (22CA552 and 22CA553). The artifacts and field records, plus reserved unanalyzed charcoal and sediment samples, are curated at MDAH’s Historic Preservation Division in Jackson, Mississippi.
At the time of the 1990s fieldwork, the Slate Springs Mound was covered with deciduous trees (Figure 9.1). At this writing, the mound’s tree cover remains undisturbed. In the early 1990s the area surrounding the mound was in pasture (Figure 9.1), but has since been planted with pine trees. The four corners of the nearly square mound point to the cardinal directions (Figure 9.2), a trait shared in common with several other regional Woodland-period platform mounds (Carleton 1999; Rafferty 2002:223). No earthen summit-access ramp, a common feature of platform mounds, was apparent during the author’s visits to the site. Possible vestiges of a ramp are suggested by the topographic map’s irregular, slightly projecting contour lines on the southwest side of the mound (Figure 9.2), but this is inconclusive and contrasts with the ramp locations on the northeast side of some other Woodland platform mounds (e.g., Rafferty 2002:223). A relic hunter’s pit measuring about 2 m in diameter and 1.2 m deep is located near the south corner of the mound’s platform summit (Figure 9.2).
Some discrepancy exists between the mound’s dimensions reported by Haag (1952:2) and those obtained during the 1990s fieldwork. Whereas the mound’s 1952 horizontal dimensions were stated as “ninety feet [30.5m] on a side,” the 1990s measurements for each side of the mound are considerably greater, averaging 125 ft/42 m per side at the base. Also, the 1990s maximum height measurement of 3 m (10 ft) is less than the height reported in 1952, “twelve feet” (3.7 m). It is not stated in the 1952 report whether the mound’s dimensions were measured or only estimated, but the reported height of 12 ft may be a result of “the common tendency to over-estimate the height of mounds,” as an early investigator of Mississippi’s Indian mounds noted (Brown 1926:26). However, that the 1990s mound measurements are wider and lower than the figures given in Haag’s report suggests possible vertical deflation and horizontal spreading of the mound, perhaps due to some disturbance and resulting erosion subsequent to Haag’s visit. As was observable during the 1990s fieldwork, and as evident in the topographic map (Figure 9.2), the northern half of the mound summit is not uniformly level, but slopes down toward the north, unlike the southern half of the summit, which is nearly flat. This unevenness of the summit is not mentioned in Haag’s 1952 report. An inconclusive suggestion of possible past disturbance of the mound’s original contours was provided by leaseholder Mickey West, who told the author in 1991 of his vague childhood memory (apparently dating to the mid-1950s) that the mound had been cleared and partially plowed by his father, Ruble West, in an attempt to put it into cultivation. This recollection is reinforced in that no trees are visible at the mound’s location indicated on the 1958 aerial photograph used for map sheet 87 of the Calhoun County soil survey book (McMullen et al. 1965). The reported cultivation activity, which evidently was soon abandoned, may explain the somewhat deflated appearance of the north half of the mound. However, since the south half of the mound retains a nearly level platform summit and the southwest and southeast sides of the mound are relatively steep, the essentially square, flat-topped form of the mound is still quite evident (Figure 9.2).
During the author’s first visit to the site in June 1991, two bore samples were taken from the mound with a 7-cm-diameter bucket auger in an initial attempt to gain an indication of its internal composition. Both sample columns were extended down to premound subsoil. The first auger column was placed at the top center of the mound and the second about 4.5 m north of the south summit corner (Figure 9.2). In general the auger testing revealed that the mound fill color and texture varies little between the top of the mound and the original premound subsoil. A conspicuous exception to the generally homogeneous character of the mound fill was first encountered in the form of a 2 cm-thick horizontal band of white (2.5Y 8/2) silty clay loam, which was revealed in the second auger column about 90 cm beneath the summit. This light-colored layer was subsequently found to extend to a point 2.5 m southwest of the second auger column, where it appeared as a 3- to 4-cm stratum in the cleaned-off north wall of the relic hunter’s pit, also at a depth of around 90 cm. This white stratum of fill contrasted sharply with the otherwise dark brown to yellowish brown silt loam and clay loam seen in the remainder of the relic hunter’s pit walls and in the auger samples.
Figure 9.4. Stratigraphic profile of test excavation (west wall), Slate Springs Mound (22CA502).
In January 1992, a 1-×-1-meter test pit was excavated to a depth of 2 meters beneath the mound summit (Figures 9.2 and 9.4). The excavation could not be widened or extended to the base of the mound due to time limitations. The excavated mound fill was screened through quarter-inch (.64 cm) mesh. Following excavation, the pit was backfilled. The unit was designated by its northeast corner, 0N11E, as measured by tape 11 m east of a large aluminum stake which was driven into the approximate top center of the mound as a datum for the topographic mapping survey. (Upon completion of the fieldwork, the stake was driven completely into the ground as a permanent reference marker for future work at the site.) The pit was excavated by strata (fill-loading zones and features discernible by variability in color and texture of the soils), and within zones by arbitrary levels (10 cm or less). Stratigraphic zones (labeled A through N in Figure 9.4) were generally faint; the west wall of the test unit exhibited the most variability (Figure 9.4).
It is notable that high-contrast, white silty clay loam identical in color and texture to the deposits observed in the second auger column and in the relic hunter’s pit appeared at roughly the same depth in the 1-×-1-m test unit. In profile, however, the deposits of white fill as observed in the west wall of the 1 × 1 (labeled E in Figure 9.4) did not occur in a thin, even layer as was seen in the auger column and relic hunter’s pit, but instead resembled individual basket loads that were dumped and left unspread. However, the fact that this light-colored fill appeared at approximately the same depth in several subsurface exposures suggests that this conspicuously contrasting zone represents a horizontally extensive capping layer laid down between mantles of fill. The occurrence of strata with contrasting colors has been noted in other pre-Mississippian flat-topped mounds, an effect described by Knight (1990:171) as “a sandwiching of layers.”
Above and underneath the white-colored deposits, the mound fill throughout the test pit was of relatively uniform but subtly varying texture and coloration indicative of fill loading, consisting of irregularly alternating layers of loam, silty loam, and clay loam of slightly different shades of brown. Beginning in Zone I at approximately 100-cm depth and continuing down to 120 cm in the west half of the test unit, increasingly abundant charcoal, including chunks as large as 2 cm in diameter, was encountered in the fill matrix.
The base of the charcoal-containing fill of Zone I was underlain by Feature 1, a double lens-shaped layer (labeled F.1 in Figure 9.4) of silt loam with a hardened, brittle texture and a yellowish-brown color (10YR 5/4) resulting from in situ burning of wood, evidenced by charcoal fragments within the lenses of Feature 1 as well as in the directly overlying fill matrix of Zone I. Due to the restricted horizontal extent of the excavation, whether the charcoal and the underlying fire-hardened and -discolored soil represents the remains of a possible building which stood on the summit of an earlier mound stage is unknown, although no obvious preserved structural elements such as post-molds, charred thatch, charred wooden posts, wall trenches, or fired daub were found. The intact state of Feature 1 suggests that it may be a burned surface or hearth on a previous platform summit (e.g., Knight 1990:170–171; Lindauer and Blitz 1997:173) upon which a fire was burned shortly preceding the addition of more fill of a new construction stage to create a higher mound summit.
Feature 2 consisted of a discrete deposit of charcoal (labeled F.2 in Figure 9.4) within Zone M, located at a depth of ca. 170–180 cm. Unlike Feature 1, no fire-discolored or -hardened fill was associated.
Few artifacts were encountered during testing of the mound. All of the prehistoric items are interpreted as incidental inclusions in fill presumably brought in from a nearby occupation area.
Six Woodland-period potsherds (see below) were recovered from the 1-×-1-m test unit: one sand-tempered fabric impressed from the 10–20-cm level below the surface in Zone B, one sand-tempered cordmarked from the 110–120-cm level in Zone I, one sand-tempered plain and one grog-tempered plain from the 160–170-cm level in Zone N, and one grog-tempered plain from the 190–200-cm level in Zone O. One additional grog-tempered plain sherd was recovered from a depth of 125 cm in the bore hole placed about 3 m south of the 1 × 1.
Four samples of wood charcoal obtained from the test excavation were submitted to Beta Analytic, Inc., for radiocarbon dating (Table 9.1). The previously published results (Sims and Connaway 2000:221) have been recalibrated using CALIB 5.0.2 (Reimer et al. 2004) for presentation here.
Except for the date from the context closest to the top of the mound, the dates conform to the order of stratigraphic superposition. However, the implied temporal span of more than 500 years seems anomalous as the dates are derived from samples vertically distributed in only about a meter of fill. Moreover, no indicators for the passage of a lengthy period of time (mottling due to weathering, development of an A horizon) were observed in the zones of fill separating the four charcoal deposits. That the three samples yielding the earliest dates (Beta-63788, -64735, and -62801) were derived from small, discrete concentrations of charcoal gives rise to the suspicion that these three charcoal samples may represent secondary deposits, inadvertently included in loads of mound fill taken from earlier, off-mound occupation areas. The fact that the excavation also recovered habitation debris (potsherds) from the fill reinforces this scenario. Possible evidence of mixing of off-mound materials of different ages in mound fill can be adduced from the fact that the deepest level of the excavation, although it yielded the sample producing the oldest Middle Woodland date (1800 ± 70 B.P.; cal. 2-sigma range A.D. 70–390, rounding to the nearest decade), also yielded a grog-tempered plain sherd, in this area a likely Late Woodland artifact (see below). In contrast, the bedded deposit of abundant charcoal yielding the age of 1310 ± 50 B.P. (cal. 2-sigma range A.D. 640–780, rounding to the nearest decade), together with the underlying lens of fire-hardened and discolored soil, is an unambiguous product of in situ burning, marking an actual event associated with the mound’s construction and/or use. This later date, when considered alone, would seem to place the Slate Springs Mound in the Late Woodland period. However, the three earlier dates, although their depositional context may seem questionable, nevertheless cannot be dismissed out of hand and must be taken into account when considering the age of the mound. Taken together, then, the absolute dates appear to bracket the construction/use span of the mound within the Middle Woodland to early Late Woodland periods.
According to Haag’s 1952 survey report, only a few artifacts were collected from the field surrounding the mound, consisting of grog-tempered potsherds of likely Late Woodland affiliation (“[s]ix sherds of Tishomingo Cord-marked and three sherds of Tishomingo Plain”) and a probable Early Archaic diagnostic, a “side-notched, bevel-edged projectile point” (Haag 1952:2). The low recovery of material was attributed to the fact that the surface of the field was obscured by pasture vegetation at the time of Haag’s site visit. Although the 1952 report implies that surface artifacts were found in proximity to the mound, the 1990s investigations revealed that the nearest concentrations of cultural debris are located at least 150 m away; the only artifacts found near the mound were two potsherds (one sand-tempered plain and the other sand-tempered eroded or possibly fabric-impressed) from the bare surface of the field about 10 m west of the western corner of the mound.
At the time of the 1991 and 1992 site visits, the field around the mound remained in pasture, with grass cover ranging from sparse to absent in some areas and dense in others. As had been the case during Haag’s visit, artifact occurrence was quite scant, and no darkened soil indicative of midden was observed anywhere. As described below, nearly all off-mound artifacts found in the 1990s investigations were found on two low terrace knolls; these two discrete artifact scatter loci, designated sites 22CA552 and 22CA553, were respectively located several hundred meters southeast and south of the mound.
In 1991, 18 shovel tests were dug in various areas where the grass cover was relatively dense, in the northeast, southeast, and southwest quadrants of the field surrounding the mound. The shovel-test holes were about 35 cm wide and 35 cm deep; the soil removed from each hole was screened through quarter-inch mesh. These tests yielded only five artifacts (1 grog-tempered plain and 2 sand-tempered plain sherds, a biface thinning flake of pink chert, and the distal portion of a Madison triangular point of red chert), found in a cluster of four positive tests located around 270 m southeast of the mound. A couple of additional artifacts were collected from the surface of this area during a visit in 1992 (a battered hammerstone of yellow quartzite and a pitted ferruginous sandstone cobble). The area from which this shovel-test and surface material was obtained, designated site 22CA552, measures about 140 m north-south by 50 m east-west.
Reinspection of the field around the mound in 1992 found the grass cover on the southwestern quadrant of the field much diminished from what had been present during the 1991 visit. As a result, a general surface collection was made there from an area measuring some 150 m north-south by 50 m east-west. This area, the northern limit of which is about 150 m south of the mound, has been designated archaeological site 22CA553 by MDAH. A Late Archaic or Early Woodland Flint Creek point of Citronelle or Tuscaloosa Gravel chert was found, while Woodland-period diagnostics include two sand-tempered plain and two grog-tempered plain sherds, along with two unfinished, narrow bifaces with straight stems (possibly Middle Woodland) of gravel chert. Also recovered were a few dozen pieces of local chert debitage, as well as a hammerstone of pink quartzite, three smoothed tabular (metate?) fragments of ferruginous sandstone, and three broken pitted ferruginous sandstone cobbles.
As discussed by Ford (1977, 1980, 1981, 1989), Johnson (1984, 1988), and Peacock (1997, 2003, this volume, Chapter 8), archaeological investigation of the interior uplands of Mississippi has lagged behind that of better known adjacent regions to the west (the Yazoo Basin/Lower Mississippi Valley) and to the east (northeast Mississippi and Tombigbee River valley). In addition, because of the perceived presence in the North Central Hills of a “mixture of ceramic types which have been recognized and defined separately in the two regions” (Ford 1980:26), researchers in the North Central Hills have tended to borrow existing ceramic typologies from those surrounding areas. This approach was first taken in the analysis of North Central Hills ceramics by Haag, who stated that the Grenada Reservoir area “lies in a zone that is transitional between clay-grit [grog] tempering to the west in the Mississippi Delta and sand-tempering to the north and east in North Mississippi and Alabama” (Haag 1952:18). As a result, Haag (1952) classified his Grenada Reservoir survey ceramics using types that had been defined by Phillips et al. (1951) for the Lower Mississippi Valley and by Jennings (1941, 1944), Cotter (1950), and Cotter and Corbett (1951) for northeast Mississippi.
The latter four works established the Middle and Late Woodland Miller ceramic sequence for northeastern Mississippi. Jenkins (1981) modified and refined the Miller sequence for the central Tombigbee River valley, dividing the Miller sequence into the Middle Woodland-period Miller I (ca.100 B.C.–A.D. 300) and Miller II (ca. A.D. 300–600) phases, followed by the Late Woodland Miller III phase (ca. A.D. 600–1100). The date ranges for Millerstyle ceramics are a couple of centuries earlier north of the central Tombigbee Valley (Ford 1989, Rafferty 1987, 1990; Walling et al. 1991). The Miller ceramic sequence is characterized by gradual shifts in tempering agents and surface finishes through time: Miller I is dominated by sand-tempered fabric impressed (Saltillo Fabric Marked), Miller II by sand-tempered cordmarked (Furrs Cord Marked), and the Miller III phase by grog-tempered wares, both cordmarked and plain (Tishomingo Cord Marked and Tishomingo Plain, respectively). Peacock (1997) performed a seriation of ceramic assemblages from the Ackerman Unit of the Tombigbee National Forest, well within the eastern section of the North Central Hills (Figure 9.3), which revealed that the Miller sequence is applicable there. This finding necessitates a modification of Johnson’s (1988) model positing that the Miller ceramic tradition does not extend into the North Central Hills from the east.
In striking contrast is the pattern revealed by seriation of assemblages from the Holly Springs unit of the Holly Springs National Forest in the extreme northern portion of the North Central Hills (Peacock 1997; Figure 9.3): although the temporal shift in surface finishes from fabric impressed to cord-marked is the same as in the Miller region to the east and southeast, the change in temper through time is reversed, from grog to sand. Moreover, because surface-finished sherds in the Holly Springs assemblages are dominated by grog-tempered fabric impressed, with some grog-tempered sherds bearing individual cord impressions, Peacock (1997:245–246, this volume, Chapter 8) assigns this material to the Early Woodland (or early Middle Woodland) Tchula period (ca. 400 B.C.–100 B.C.), originally defined for northwest Mississippi by Phillips et al. (1951:432) and Phillips (1970:878–880). Similar Tchula assemblages have been identified from sites in the North Central Hills south of the Holly Springs National Forest (Ford 1989, 1990), to the west in the Loess Hills at the Batesville Mounds (Johnson et al. 2002), and in western Tennessee (Mainfort and Chapman 1994). Assemblages containing grog-tempered fabric-impressed materials representative of Tchula occupations in northern and northwestern Mississippi may somewhat predate northeastern and eastern Mississippi Miller I occupations with their assemblages dominated by sand-tempered fabric-impressed, although this is uncertain due to the shortage of radiocarbon dates from Tchula contexts (Ford 1996; Peacock 1996c, 1997, this volume, Chapter 8).
The Slate Springs Mound and nearby sites 22CA552 and 22CA553 are located in the heart of the North Central Hills, about equidistant between the Holly Springs Unit of the Holly Springs National Forest and the Ackerman Unit of the Tombigbee National Forest (Figure 10.3). How do these three sites fit, chronologically and spatially, in view of Woodland-period ceramic traits documented in other areas of the North Central Hills? Haag (1952: 2) classified his small collection of sherds from the field around the mound as “Tishomingo Cord-marked” and “Tishomingo Plain.” These types were named and defined by Jennings (1941:200–201) to denote pottery containing clay or grog temper in a sandy-textured paste. This is in contrast to clay/grog-tempered sherds from other sites in the Grenada Lake survey which are described by Haag (1952:20) as having a “smooth” and “chalky” textured paste, and classified by him as “Mulberry Creek Cord-marked” and “Baytown Plain,” following the Phillips et al. (1951) Lower Mississippi Valley typology. As previously discussed, the 1990s fieldwork recovered an additional six grog-tempered sherds, all plain: three from the mound and three from the nearby occupation areas (one from 22CA552, two from 22CA553). In this report, no paste texture distinctions are attempted for these sherds, as no known chronological, cultural, or functional inferences can be drawn from such an exercise (Peacock 1997:241–242, this volume, Chapter 8).
In addition to the grog-tempered sherds, 12 sand-tempered specimens recovered in the 1990s fieldwork include 1 sand-tempered fabric impressed from mound fill; 1 sand-tempered eroded (possibly fabric impressed) from the field just west of the mound; 1 sand-tempered cord marked from mound fill; and 9 sand-tempered plain (1 from mound fill and the rest from off-mound contexts). Sand-tempered fabric-impressed and cord-marked pottery in northeast Mississippi is commonly identified by the type names Saltillo Fabric Marked and Furrs Cord Marked, respectively (e.g., Jenkins 1981), after Jennings’s original designations (Jennings 1941:199–201, 1944:411–412).
Although few in number, the sherds with textured surface finishes are not inconsistent with the Miller ceramic tradition of the Middle and Late Woodland periods. This impression is reinforced by the absence of any pottery that might suggest affinity with the Early or Middle Woodland Tchula tradition documented farther north in the North Central Hills and in the Yazoo Basin, i.e., grog-tempered fabric-impressed or grog-tempered individual cord impressed (Peacock 1997, this volume, Chapter 8). Although Haag (1952) found clay/grog-tempered fabric-impressed sherds at 5 of the 48 sites he recorded in the Grenada Lake area (classified in his report as “Withers Fabric-impressed” following the Phillips et al. [1951:73] Lower Mississippi Valley typology), those sites are located in the western portion of his survey area, in Grenada County. No such material has been found at the Slate Springs Mound, or, according to the MDAH state archaeological site inventory, at any of the 32 additional recorded prehistoric sites within a 10-mile (15-km) radius of the mound.
The presence of a partial Madison triangular point along with grog-tempered pottery in the assemblage from off-mound site 22CA552 reinforces the probable presence of a Late Woodland occupation, as the Madison point was introduced to Mississippi during Late Woodland times (McGahey 2000:200–201). The recovery of three grog-tempered sherds from the mound excavation, including one found in the deepest level (190–200 cm), implies that the mound was built in the Late Woodland period, an impression reinforced by the radiocarbon date most likely representative of an in situ context: cal. A.D. 640–780, rounding to the nearest decade (see Table 9.1).
A particularly revealing category of negative evidence is signified by the absence of Mississippi-period (post-A.D. 1100) shell-tempered ceramics in any of the collections from the mound, from nearby sites 22CA552 and 22CA553, and from all but one of the additional 30 recorded prehistoric sites within a 10 mile (15 km) radius of the mound. (The sole exception is site 22CA514, located some 5 miles [8 km] north of the mound, from which 2 shell-tempered sherds are recorded.) The extremely scant presence of Mississippian occupations in the region in general is evident in the results of Haag’s (1952) Grenada Lake survey: only one shell-tempered sherd (Barton Incised) was found out of a total of 1,287 sherds collected from 48 sites. A later survey of several reservoir areas in the North Central Hills likewise recovered very few shell-tempered ceramics (Broyles et al. 1982:149). Further confirming the scarcity of Mississippian occupation of the region are the results of a more recent survey that included wide-ranging areas of the North Central Hills and adjacent Loess Hills (Johnson 2001:137–138): of a total of 919 sites recorded in a 157,214-acre survey universe, only 22 sites yielded shell-tempered sherds. Rafferty (2002:222) notes that “[m]any . . . Woodland flat-topped mounds . . . had been recorded or discussed in earlier literature under the assumption that they were constructed during Mississippian times.” In the case of the Slate Springs Mound and its environs, the lack of shell-tempered ceramics in any context strongly suggests that the mound is not the product of a Mississippian occupation. This is conclusively confirmed by the mound’s suite of exclusively Woodland-period radiocarbon dates discussed above.
Peacock (1997:238, 244, 253) has noted that most of the Tchula and Miller sites recorded in the National Forests of the North Central Hills tend to be located in two separate major drainage basins. The streams in the Tchula area (the Holly Springs Unit of the Holly Springs National Forest and adjacent areas of Lafayette County) (Figure 9.3) flow westward to the Yazoo Basin in the Mississippi River drainage, where Tchula stylistic traits are most common. In contrast, the Miller-affiliated Woodland-period sites of the Ackerman Unit of the Tombigbee National Forest (Figure 9.3) are located on land drained by the Noxubee River, a tributary of the Tombigbee River. Peacock (1997:244) has posited that “the Miller ceramic tradition [may be] bounded [on the west] by . . . the drainage divide between the Tombigbee and Yazoo River basins,” although he cautions that “there is no simple, overreaching correlation between drainage basins and ceramic styles” (Peacock 1997:253). Indeed, the Slate Springs Mound and other Middle to Late Woodland-period sites in the area with Miller-like ceramics do not conform to this dichotomous geographic pattern, as these sites are located not in the Tombigbee drainage, where the Miller tradition is centered, but on tributaries of the Yalobusha River, which flows west into the Yazoo Basin where the Tchula and later Marksville, Baytown, and Coles Creek traditions are dominant (Kidder 2002). It should be noted, however, that these sites, while west of the Tombigbee-Yazoo Basin drainage divide, are not far from it: the Slate Springs Mound is in the upper reaches of the Yalobusha River drainage, only about 15 miles (24 km) west of the divide (Figure 9.3). In this respect, the Slate Springs site shares something in common with the much larger, multi-mound sites of Ingomar (22UN500) (Rafferty 1987, 1990) and Pinson (40MD1) (Mainfort 1986a, 1988, 1996, 2013b). As noted by Rafferty (1987:148) and Johnson (1988), Ingomar and Pinson are also Woodland-period sites with platform mounds located near the divides of major drainage systems. In view of the Miller-like ceramic assemblages of the Slate Springs locality, the Tombigbee-Yazoo divide in the North Central Hills might be regarded not as a discrete borderline between the Miller tradition on the east and the Tchula, Marksville, Baytown, and Coles Creek traditions to the west, but as a fairly broad transitional zone. (See also Rafferty’s [1994] seriation of assemblages containing Miller tradition ceramics from sites on both sides of the Yazoo Basin-Tombigbee Basin divide in Union, Lee, and Pontotoc counties, Mississippi.)
Platform mounds, whatever their age or cultural affiliation, are rare in the North Central Hills of Mississippi. Other than the Slate Springs Mound, the only known examples are Nanih Waiya (22WI500) in Winston County (Carleton 1999), Hurricane Landing (22LA516) in Lafayette County (Thorne 1981), and the Old Hoover Place (22HO502) in Holmes County (Lorenz 1990, 1996). Of these three mounds, Hurricane and Old Hoover are known to date to the post-A.D. 1000 Mississippi period. The mound at the multi-occupation Nanih Waiya site in Winston County has been proposed as a Middle Woodland-period structure by Carleton (1999:125), but this is uncertain, as it has not been excavated and directly dated by absolute means. Consequently, the Slate Springs Mound is currently the sole confirmed Woodland-period platform mound in the entirety of the North Central Hills (Figure 9.3).
To cite other known Woodland-period mounds of flat-topped configuration in northern Mississippi, it is necessary to widen the geographic scope beyond the North Central Hills to encompass adjacent physiographic regions (Figure 9.3). To the west in the Loess Hills, the construction of the Batesville Mounds (22PA500) was initiated the Early Woodland period and may have extended into the Middle Woodland; both Early to Middle Woodland Tchula tradition ceramics and Middle Woodland Marksville pottery were recovered (Johnson et al. 2002). On the Pontotoc Ridge to the east, a radiocarbon date obtained from the one remaining mound at the Thelma site (22CS501) suggests that it was built in the Late Woodland period, but the ceramics from the site include shell-tempered sherds as well as grog-tempered material, interpreted by Johnson and Atkinson (1987:69) as indicative of a later, transitional Late Woodland-Mississippian occupation (see also Rafferty, this volume). In the Flatwoods, the platform mound at Ingomar has been dated definitively to the Middle Woodland period (Rafferty 1987, 1990).
Looking beyond northern Mississippi to adjacent regions, the Pinson Mounds site in western Tennessee (Mainfort 1986a, 1988, 1996, 2013b) exhibits a number of similarities to Ingomar, and the two sites are apparently roughly contemporaneous (Mainfort and McNutt 2004). Also in western Tennessee are Middle Woodland platform mounds at the Johnston site (40MD3) (Kwas and Mainfort 1986) and the Savannah site (40HR29) (Welch 1998). Along the Tennessee River in northern Alabama, the Walling Mound (see Knight 1990:1–4 for the various site numbers employed) and the Florence Mound (1LU10) (Boudreaux and Johnson 2000) have been identified as Middle Woodland-period platform mounds of the Copena tradition. A few Baytown- (early Late Woodland-) period platform mounds may be present in the Yazoo Basin of western Mississippi (Kidder 2002:82).
Aside from the fairly superficial parallel of the mere presence of a platform mound, it is strikingly apparent that the Slate Springs site shares little else in common with any of the other sites mentioned above in terms of size and complexity. In contrast to most (if not all) of the other platform mounds mentioned above, the Slate Springs Mound is rather small. As mentioned above, it is also notable that the Slate Springs site consists of a solitary mound, whereas most other Woodland-period platform mounds occur in groups with other flat-topped mounds and/or with conical burial mounds. There is no evidence for additional mounds anywhere in the vicinity of the Slate Springs site; Haag’s (1952) report does not mention any mounds nearby, and no topographical features suggestive of mound remnants were observed during the 1990s investigations. On some Woodland platform mounds, midden deposits containing food remains (animal bone, charred plant material, etc.) have been found on the summits or dumped down the sides; this has been interpreted as evidence of ritual feasting (e.g., Blitz and Mann 2000:38; Knight 1990:158–164, 2001:323–325; Lindauer and Blitz 1997; Milanich et al. 1984:102). However, no bone or non-wood floral material of any kind has been found at or around the Slate Springs Mound.
Haag (1952:29) interpreted the Slate Springs Mound’s relatively small size and solitary occurrence as evidence of an undeveloped, early phase of a platform mound-building tradition in the region. Accordingly, Haag (1952:29–30) assigned the mound to Ford and Willey’s (1941:328–330, 344) Temple Mound I stage, which was equated with an “early Middle Mississippian period” in western Mississippi (Ford and Willey 1941:348). In Ford and Willey’s pan-regional culture-historical scheme, a relatively continuous and brief developmental trajectory from Temple Mound I to Temple Mound II or fully developed Mississippian was postulated. Phillips (1970:7) realigned Temple Mound I with Coles Creek, a Late Woodland culture of the Lower Mississippi Valley spanning ca. A.D. 700–1200 (Kidder 2002:69). However, it is likely not the Coles Creek period but the preceding early Late Woodland Baytown period (A.D. 500–700) and/or Middle Woodland Marksville period (200 B.C.–A.D. 500) of the LMV (Kidder 2002) that coincide with the age range of the Slate Springs Mound.
As outlined above, other than the exceptions of the Hurricane Landing Mound in Lafayette County and the Old Hoover Mound in Holmes County, no Mississippi-period mounds, and few Mississippian sites of any kind, have been discovered in the North Central Hills. The preponderance of evidence accumulated from throughout the North Central Hills supports the view that, for presently unknown reasons, most of the region was largely abandoned by the beginning of the Mississippi period (Peacock 1997:252, 2003:50–51; Peacock et al. 2008). Because the Slate Springs Mound is now known to date to the Middle or early Late Woodland period, the site does not represent an Emergent Mississippian platform mound-building trend in the region, as believed by Haag. On the contrary, it seems to be a temporally and spatially isolated anomaly.
Rafferty (2002:224) has stated that “[c]rucial to understanding the relationships among [platform] mounds and to Woodland habitation sites is information about occupation duration, continuity of mound building, kinds of features and artifacts found in mounds, and contemporaneity.” The chronological and functional relationships, if any, between the Slate Springs Mound and the two nearby sites 22CA552 and 22CA553 are uncertain, but both the ceramic and chronometric data suggest that they are partially contemporaneous in the Middle and/or early Late Woodland periods. The occupation debris at both 22CA552 and 553 is not plentiful, conforming to the pattern noted by Rafferty (2002:224): “Habitation evidence is either unknown or seems to be relatively light at most of the large Middle Woodland mound sites in the north-south strip from Tennessee to Mississippi. . . . Most Middle Woodland habitation sites in the central region are hamlet-sized.”
Knight (1990:162–163) has postulated that the widespread but scant occurrence of Woodland-period platform mounds across the southeastern United States suggests sporadic, relatively short-term, non-hereditary concentrations of political power held by a few individuals in otherwise non-stratified societies (see also Milanich et al. 1984). The isolation and small size of the Slate Springs Mound conforms to this hypothesized pattern. How this mound was integrated into a settlement system context is unknown due to the lack of intensive and extensive survey in the locality. Assuming that the mound represents some sort of public sociopolitical center, it does appear that no site hierarchy exists in the surrounding region, given the small size of the mound and the lack of superordinate (i.e., multimound) sites anywhere in the area. As a result, the influence exerted by the site most probably was limited to a relatively small territory. This impression is reinforced by the lack in the (admittedly small) artifact samples from the mound and from nearby sites of extraregional ceramics or other exotic trade materials, which have been found at larger Woodland mound centers like Ingomar (Rafferty 1990:100–101) and Pinson (Mainfort 1986a:35–46).
Although modest in scope, the investigations accomplished at the Slate Springs Mound so far have been fairly productive, but more detailed insights into this site’s functional role in the regional Woodland-period settlement system must await further excavation of the mound and intensive survey of the surrounding area. Fortunately, the mound’s long-term protection is secure due to its formal designation as a Mississippi Landmark. Under provisions of the Mississippi Antiquities Law, destruction or alteration of Mississippi Landmarks is prohibited, ensuring that this ancient monument will be preserved as a valuable repository of archaeological data.
I wish to express my appreciation of the late David Fant, Cliff Jenkins, Evan Peacock, Janet Rafferty, and Julie Baca, all of whom generously volunteered their assistance in the fieldwork. Evan also deserves much credit for finally prevailing upon me to revise and update my two-decades-old conference paper for publication—praestat sero quam nunquam. Mrs. Ruble West and Mickey West, leaseholders of the land on which the mound is located, provided gracious hospitality during the fieldwork. Because the Slate Springs Mound is on sixteenth-section public school land, thanks are due to Dewitt Spencer, the Calhoun County superintendent of education at the time of the 1990s site investigations, for his enthusiastic support during the process of designating the mound a Mississippi Landmark. Jeffrey Alvey, my colleague at the Cobb Institute of Archaeology, efficiently prepared the digital figures from my smudged old field drawings.
Finally, although he was not directly involved in this project, Sam Brookes nonetheless certainly deserves recognition for doing so much in his leadership role with the National Forests in Mississippi to bring to light the prehistory of Mississippi’s North Central Hills (in addition to many other regions of the state). Specifically, Sam saw to it that Zone Archaeologists were assigned to National Forest units in that region, and these individuals included the above-named David Fant and Evan Peacock for the Holly Springs National Forest and the Tombigbee National Forest, respectively. Without the copious amounts of Woodland-period data that Sam’s dedicated archaeologists recovered from these two forests, the Slate Springs Mound would have been much harder to place in perspective. It is through such networks of colleagues and friends that archaeological knowledge of this once little-known region has advanced.
