THE SMITH SITE: A SMALL HOPEWELL SITE OVERLOOKING THE STUBBS EARTHWORKS

Ted S. Sunderhaus

Cincinnati Museum Center

Rodney Riggs

Gray & Pape, Inc.

Frank L. Cowan

University of Cincinnati, Adjunct Assistant Professor

Ohio Archaeological Council © 2001

Extensive surveys, conducted in the late 1970s and early 1980s by the Miami Purchase Association for Historic Preservation (Genheimer 1984, 1996, 1997), documented 28 Hopewell sites along the Little Miami River presumed to have been employed by the builders and users of the nearby Stubbs Earthworks complex (Whittlesey 1851). The surveys found 17 discrete artifact concentrations on the south side of the Little Miami River in very close proximity to the former earthworks, while 11 additional artifact concentrations were mapped on more distant terraces on the north side of the river. In aggregate, these sites are referred to as the Stubbs Cluster.

Salvage excavations undertaken by the Cincinnati Museum Center in 1998 and 1999 (Cowan et al. 1998, 1999) focused on portions of four Stubbs Cluster sites (33Wa256, 33Wa257, 33Wa258, and 33Wa260), located within a 35 hectare property on which the Little Miami High School was then being built. Thirty-eight hundred square meters of excavation within and between previously designated artifact concentrations documented 21 Hopewell wooden structures, along with many other features, and provided strong evidence that the earthworks had extended onto that property. As a matter of practical convenience, we employed the generic designation of the Stubbs Earthworks site (33Wa1) to those excavations rather than the more specific site nomenclature used by Genheimer (1984, 1996, 1997).

Until 2001, none of the other Stubbs Cluster sites had been the subject of substantial subsurface investigation. That changed in June 2001 when construction for the Saddlebrook Subdivision began on a high, bluff-like ridge overlooking the Stubbs Earthworks. Like most salvage excavations, the Smith site investigation was undertaken with a minimum of resources under time-constrained circumstances. However, we benefited from lots of help and cooperation. The subdivision developer, John Adams of J. P. S. Development, LLC (and member of the American Institute of Archaeology) generously granted the Cincinnati Museum Center permission to conduct excavations amidst on-going construction work. We received cheerful assistance from Broshear Contractors of Hamilton, Ohio, which scheduled many of its massive excavations around our own small ones and expedited our work with the skillful removal of plow zone sediments from several areas of interest. Many local professional and avocational archaeologists toiled mightily throughout a very hot and humid month of on-and-off-again excavations.

The Smith Site (33Wa362)

Figure 1.  Location of the Smith Site Relative to the Stubbs Earthworks.
Figure 1. Location of the Smith Site Relative to the Stubbs Earthworks.
 


The Smith site (33Wa362) is located on a high, early-Wisconsinan Williamsburg terrace overlooking Big Foot Run, a small tributary of the Little Miami River in Hamilton Township, Warren County, Ohio. It also directly overlooks the Stubbs Earthworks, which was situated on a nearly level, late-Wisconsinan Fox outwash terrace some 12 to 14 meters below the Smith site (Figure 1). Genheimer's (1984, 1996, 1997) surface collections and artifact distribution maps provide an important background for these investigations, showing this to be a multi-component site with a predominance of Hopewellian artifacts and debris.

Three widely separated portions of the site were investigated in June and July 2001. All investigations were necessarily opportunistic in location, extent, and thoroughness, depending largely on the fortuitous exposure of subsurface features by heavy construction equipment and varied opportunities to conduct systematic excavation before the landform was forever changed and its archaeological deposits lost. All recovered artifacts and flotation samples and the field notes are curated at the Cincinnati Museum Center and will be the subject of more thorough analysis and reportage in the near future. This very preliminary report is intended to alert OAC members and other Hopewell scholars to the highlights of a new body of data concerning Ohio Hopewell.

Structure 1

Figure 2.  Smith Site Structure 1.
Figure 2. Smith Site Structure 1.
 


The most striking result of the 2001 excavations was the exposure of the complete floor plan of a Middle Woodland house-like structure. The structure was located at the southern extremity of the site; in fact it was actually just south of the site boundary as originally defined by surface artifact densities. A shallow road cut initially exposed several post molds; considerable plow zone removal with a tractor-drawn drag-scraper, shovels, and trowels was necessary to expose the complete structure, its interior, and its immediate surroundings.

The wooden post structure (Figure 2) was 8-by-8 meters in its interior dimensions and nearly square with rounded corners. The structure was single-posted, except for the northwestern and northeastern corners, which were closely double-posted. No evidence for use of the wall-trench construction technique was detected, although that building method characterized all of the Stubbs Hopewell structures excavated on the lower terrace during the 1999 field season. Structure 1 had an unusually wide (3.16 meters) entryway centered on the south wall and what appears to have been a second, smaller (1.5 meters) entryway in the north wall at the northeast corner. The exterior post molds were quite regularly spaced with a mean center-to-center distance (excepting the two presumed doorways) of 84.7 +/- 8.3 cm. Interior post molds included a large, 45 cm diameter center post hole, two post molds along the southern interior, flanking the entryway, and four post molds along the north wall. The post molds of the exterior walls contained relatively large quantities of wood charcoal, although the interior post molds did not. It appears likely that the structure may have been destroyed by fire. In addition to its distinctively "Hopewellian" form, scraps of mica in the fills of several post molds confirm the Middle Woodland age of the structure.

A small, very shallow basin-shaped pit feature was found on the interior, very near the structure's west wall. Two pit features were located outside the structure. One small, very shallow pit was situated about a meter east of the southeastern corner of the structure. A much larger and somewhat deeper pit lay south of the southwestern corner of the structure. None of the excavated pit features yielded culturally diagnostic artifacts.

Although the "floor" of Structure 1 had been lost to plowing, there were no subsoil indications of an interior hearth. The apparent absence of such a feature, the unusually wide entryway of the structure, and the exposed ridge-top location of the structure all suggest that it was not designed or situated for cold weather use.

"Sandy Knoll" Excavations

Figure 3.  Smith Site Sandy Knoll Excavations.
Figure 3. Smith Site Sandy Knoll Excavations.
 


Approximately 30 meters north of Structure 1 was a slightly higher ridge-crest composed of a narrow band of reddish-brown very fine sand with clayey silt which abruptly changed to very fine white sand at a depth of approximately one-meter. A deeply graded road cut across the sand ridge provided a singular opportunity to clear an expedient profile, and two features were immediately exposed. Surface shovel-shaving along the edge of the road cut exposed more features. Eventually, with the assistance of the contractor's front-end loader and lots more shovel-shaving, an irregularly shaped surface of approximately 12-by-7 meters was fully exposed and found to contain nearly 100 features. Figure 3 illustrates the irregular excavation transect, the confirmed cultural features, and those features we did not have time to excavate. Absent from the map are many "features" that, upon excavation, turned out to be natural disturbances.

Although not all of the features in this area were excavated and confirmed to be cultural, it is clear that, in prehistory, this was a repetitively used area. Bounded on the east by a steep slope down to the Big Foot Run valley and on the north and south by lesser slopes, we suspect we exposed most of the cultural features lying along those margins of the sand ridge. However, the bulldozed truncation on the western side of the block excavation prevented any clear resolution to the chaotic array of post molds, and we did not have the opportunity to excavate systematic exposures on the sand ridge to the west of the road cut.

We estimate that the many post molds represent four or more overlapping post structures. Pit features, too, were numerous, and many of those also overlapped one another. Although fire-cracked rocks were relatively common, other artifact classes were notably sparse, and identifiable hearths were absent. Very few temporally diagnostic artifacts were recovered from the excavated features, but all of those artifacts (a few bladelets and sherds) are characteristic of the Middle Woodland period. It must be acknowledged, however, that Genheimer's early 1980s surface collections did recover projectile points from earlier periods from this as well as from other portions of the site.

Feature 37

Approximately 70 meters north of Structure 1 and some 30 or more meters north of the sandy knoll, the elevated Williamsburg terrace narrows to a point nearly surrounded by steep slopes that descend to the valley of Big Foot Run. On level ground atop this point, a very large, artifact-rich Middle Woodland pit was partially exposed by heavy earthmoving equipment. Two or three probable post molds were observed very nearby, but time-constraints and severe construction damage precluded the search for associated structures. Nonetheless, and despite the circumstances of its discovery and necessarily hasty excavation methods, Feature 37 merits special mention.

Feature 37 was approximately 1.5 meters in diameter, about 90 cm deep, and had nearly straight-sided walls and a rounded, basin-shaped bottom. The feature's fill was stratified with three primary cultural strata.

The uppermost stratum contained large quantities of Middle Woodland lithics, consisting primarily of several score kilograms of fire-cracked rocks and many chert artifacts, including 60 - 70 bladelets and fragments. The chipped stone assemblage included two obsidian bladelet fragments, as well as a small obsidian flake. Several fragments of very thin, heat-fractured bifacial tools were also recovered from this stratum. Flint Ridge (central Ohio) chert accounts for the vast majority of the bladelets, but Knox (eastern Tennessee) and Wyandotte/Harrison County (southern Indiana) cherts also occur among the other chert varieties. Tools and flakes of locally available gravel cherts were sparse to non-existent.

The second major stratum consisted of a dense mixture of cultural debris, including impressive quantities of pottery, mica, fire-cracked rock, charcoal, and burnt soil, with lesser amounts of calcined bone. About a dozen large and small pieces of cut mica were recovered from this stratum, including one complete lozenge-shaped cut-mica artifact. Thousands of additional mica scrap fragments without shaped edges were also present.

The third stratum contained relatively few artifacts. These consisted of a few pottery sherds, a very few pieces of chert and fire-cracked rock, and only moderate amounts of wood charcoal.

The pottery recovered from Feature 37 is estimated to represent only about 30 percent of that originally present in the pit. The remainder were reduced to crumbs by the weight of heavy earthmoving equipment (earthworms at the very base of the pit were crushed flat by that weight). Nonetheless, at least 369 ceramic sherds, weighing more than 1.6 kilograms, were recovered. The ceramic content of this feature is particularly notable when one considers that no more than 174 sherds were recovered from all 28 Stubbs Cluster sites, combined, during the late 1970s and early 1980s surface collecting and test pitting.

The Feature 37 sherds represent a minimum of twenty-three vessels (Table 1). McGraw Plain and McGraw Cordmarked types predominate, accounting for about 57 percent of the identified vessels. Southeastern series vessels amount to 26 percent of the ceramic assemblage. Among the five Turner Simple Stamped vessels, two are tetrapod forms, and one of the tetrapodal vessels has a Brangenberg-like rim with a single row of upwardly projecting hemiconical punctates just below the sharply everted rim. Hopewell series vessels constitute 17 percent of the assemblage.

Table 1. Minimum Numbers and Percentages of Vessels by Type and Series.

Ceramic Series and Types Minimum Number of Vessels Ceramic Type Percentages Ceramic Series Percentages
Scioto Series     57
   McGraw Plain 8 35  
   McGraw Cordmarked 5 22  
Southeastern Series     26
   Turner Simple Stamped 5 22  
   untyped Simple Stamped 1 4  
Hopewell Series     17
   untyped Brushed (Chillicothe Brushed?) 2 9  
   untyped Hopewell Incised 1 4  
   untyped Roughened 1 4  
 

 
Total 23 100 100

Hopewell series and Southeastern series vessels are often interpreted as "ritual" fine-wares. These fine-wares account for 43 percent of the ceramic assemblage, which seems to be a very high proportion for a residential locality. In contrast, McGraw Plain and Cordmarked, generally considered to be the more prosaic "utilitarian" wares associated with the Ohio Middle Woodland period, account for only 57 percent of the total assemblage.

Grit-tempered ceramics predominate, amounting to about 71 percent of the assemblage by weight (Table 2). Sherds with combinations of both grit and grog account for another 21 percent of the sample. Limestone-tempered ceramics, sand-tempered ceramics, and sherds with combinations of tempering agents also occur, if only in very small quantities.

Table 2. Ceramic Weights by Temper.

Temper Type Weight in Grams Weight (%)
Grit 1,153.18 71
Grit/Grog 338.92 21
Limestone 95.59 6
Grit/Limestone 25.27 2
Grit/Grog/Limestone 3.51 <1
Limestone/Grit 1.00 <<1
Sand 0.95 <<1
     
Total 1,618.42 100

The quite varied mix of ceramic tempering materials and other production details within a contemporaneous assemblage may reflect design alternatives to facilitate the production of functionally differentiated vessel forms. On the other hand, varied production habits may also suggest that not all the pottery vessels were the products of a local, indigenous population. Instead, it might suggest that some vessels were brought to the site or were produced there by visitors of different pottery-making traditions who came from more distant regions.

Compared with the previously investigated Stubbs Cluster sites, Feature 37 surpasses all other known features in the abundance of its artifact content. In fact, Feature 37 has undoubtedly yielded more Middle Woodland ceramics than all previous Stubbs Cluster excavations combined, despite the excavation of nearly a thousand features on the lower terrace in 1998 and 1999.

Concluding Observations

The physical setting of the Smith site is quite striking. Situated high atop a steep-sided, ridge-like terrace above the Stubbs Earthworks complex, it must have offered a commanding view of the ceremonial facilities below as well as a good view of a stretch of the Little Miami River valley.

Today, the view from the Smith site is largely blocked by dense, second-growth forest growing along the steep terrace slopes. However, archaeobotanical evidence from previously excavated features at the Stubbs Earthworks site indicates that oaks and hickories dominate the wood charcoal assemblage. Understory species and tree species that favor mesic slopes, such as maples, are notably under-represented in the Stubbs flotation samples thus far analyzed. We infer, on the basis of the "Firewood Indifference Hypothesis" (Asch and Asch 1986:497-498), that the landscape surrounding the Stubbs Earthworks was dominated during Middle Woodland times by an open oak-hickory forest. It might be further suggested that Middle Woodland populations purposefully managed the local forests, probably by controlled burning, to thin undergrowth and economically unimportant tree species to favor the canopy-spread of select, fire-tolerant mast producers (e.g., Asch and Asch 1986:438-446). Such a strategy for promoting edible nut production would have been particularly useful in a location where large numbers of people would have congregated on a periodic basis. In addition, that sort of forest management would have ensured a very impressive overview of the huge geometric earthwork complex below; such an outcome just doesn't seem unintentional for Ohio Hopewell populations.

As is the case with other excavated portions of the Stubbs Cluster, the Smith site yielded abundant evidence for wooden architecture. The only clearly interpretable structure was located in an area so depauperate in artifacts and other cultural debris that it was previously considered to be outside the defined site limits. The extraordinary sparseness of artifacts and debris, both within the plow zone and within subsurface features, suggest that this was not a structure occupied on a regular basis for a long period of time.

The "sandy knoll" excavations also revealed an abundance of wooden architecture, although there were so many episodes of structure-building that we can't make sense of individual structures within the limited space we were able to excavate. Examination of Genheimer's artifact distribution maps suggest that there were many plow zone artifacts in this general area, and we noted many fire-cracked rocks (but few other artifacts) in the plow zone as we shoveled it away. Nonetheless, the evidence suggests many episodes of use rather than long-term use. Identifiable hearths were non-existent, and excavated feature and post mold fills contained very little charcoal or burnt soil and very few lithic or ceramic artifacts.

The only Smith site feature to yield large quantities of artifacts and cultural debris was Feature 37, a feature of such artifact-richness that it figures as something of an anomaly in our experiences with Ohio Hopewell. The large pit contained extraordinary numbers of bladelets, fire-fractured biface fragments, fire-cracked rocks, small chert flakes, sherds from many vessels of varied production technologies, and scraps from the production of shaped mica artifacts. It also contained unusually large quantities of redeposited wood charcoal (including many very large wood charcoal fragments), burnt soil, and calcined bone fragments. Feature 37 clearly was refilled with materials cleaned up from one or more surrounding areas, but it does not exhibit an assemblage that could remotely be considered characteristic of normal residential debris. Rather, the contents suggest the results of ritual cleaning of adjacent, special-purpose activity areas.

Dancey and Pacheco (1997:Table 1.1) include the Smith site in their tabulation of long-occupied residential "hamlets." On the basis of available evidence, that interpretation now seems very unlikely. Detailed analyses of all classes of artifacts and cultural debris from the Smith site are still in the offing, but initial impressions indicate that the site was used frequently during the Middle Woodland period, but only for relatively brief periods of time and only for a limited range of purposes. It seems likely that the Smith site was an earthwork-focused site with "ritual" overtones, rather than a place of long-term residence.

References Cited

Asch, N. B., and D. L. Asch
 1986  Woodland Period Archeobotany at the Napoleon Hollow Site. In Woodland Period Occupations of the Napoleon Hollow Site in the Lower Illinois Valley, edited by M. D. Wiant and C. R. McGimsey, pp. 427-512. Kampsville Archeological Center Research Series 6. Center for American Archeology, Kampsville, Illinois.

Cowan, F. L., T. S. Sunderhaus, and R. A. Genheimer
  1998  Notes from the Field: An Update from the Stubbs Earthworks Site. The Ohio Archaeological Council Newsletter 10(2):6-13.

 1999  Notes from the Field, 1999: More Hopewell "Houses" at the Stubbs Earthworks Site. The Ohio Archaeological Council Newsletter 11(2):11-16.

Dancey, W. S., and P. J. Pacheco
  1997  A Community Model of Ohio Hopewell Settlement. In Ohio Hopewell Community Organization, edited by W. S. Dancey and P. J. Pacheco, pp. 3-40. The Kent State University Press, Kent, Ohio.

Genheimer, R. A.
  1984  A Systematic Examination of Middle Woodland Settlement in Warren County, Ohio. Unpublished report on file, Cincinnati Museum Center, Cincinnati, and Ohio Historic Preservation Office, Columbus.

 1996  Bladelets are Tools Too: The Predominance of Bladelets Among Formal Tools at Ohio Hopewell Sites. In A View from the Core: A Synthesis of Ohio Hopewell Archaeology, edited by P. J. Pacheco, pp. 92-107. The Ohio Archaeological Council, Columbus.

 1997  Stubbs Cluster: Hopewellian Site Dynamics at a Forgotten Little Miami River Valley Settlement. In Ohio Hopewell Community Organization, edited by W. S. Dancey and P. J. Pacheco, pp. 283-309. The Kent State University Press, Kent, Ohio.

Whittlesey, C.
  1851  Descriptions of Ancient Works in Ohio. Smithsonian Contributions to Knowledge 3. Smithsonian Institute Press, Washington, D.C.