Material Remains

LITHICS

The sample of lithic artifacts has been divided into the primary morphological classes of chipped stone and groundstone tools. These stone tool categories were further divided into the various functional categories listed in Appendix A. As expected, chipped stone debitage, in the form of flakes (n=11,272) and blocky fragments (n=4,799) comprised the vast majority of all chipped stone debris. A relatively narrow range of chipped stone artifact classes were defined, with projectile points being the most common tool type recovered. Groundstone tools were very rarely found.

Although no formal analysis of the lithic sources for chert debitage has been carried out, it was evident during cataloging that much of the utilized raw material was locally available, fossiliferous cherts which had been heat-treated to improve flaking qualities. Somewhat more distant sources of high quality material, such as Indian Creek and Wyandotte cherts, were also exploited for tool manufacture but were not as well represented in the debitage sample. As illustrated in Appendix A, flakes outnumbered blocky fragments by more than two to one and represented all stages of stone tool manufacture (i.e., initial reduction through pressure retouch and resharpening). Such evidence of the complete range of stone tool debris supports the interpretation that the village settlement was occupied for sufficiently long periods of the year (many months) to permit the initial manufacture, use, repair, and discard of chipped stone tools.

The non-projectile point stone tool assemblage contained a significant number of what appeared to be expediently produced flake tools which were grouped under the label of "modified or utilized flakes" (n=70). These tools most often consisted of no more than medium-to-large sized flakes that were pressure flaked along one or more margins to produce a working edge(s). These artifacts were likely to have been used for a wide range of purposes that would have included scraping, cutting, slicing, drilling, and perforating.

Eleven examples of "lamellar flakes" in the stone tool sample mimic the form of true lamellar bladelets known from Hopewellian contexts in the Midwest. Each specimen from the Cox's Woods site exhibited a distinctive medial ridge on its dorsal surface but lacked the parallel margins, prismatic cross-sections, and platform preparation scars which are typical of Hopewellian bladelets (Montet-White 1968; Greber et al. 1981). No examples of blade cores were recovered from the site, which indicates that these unusual flakes were not the result of a deliberate blade technology, in contrast to what has been suggested for some Fort Ancient assemblages from Ohio (Barber 1978).

Both side- and end-scrapers were recovered during the summer 1993 and 1994 field seasons (n=23). All but five were made from flakes and were chipped on one side only. One large side-scraper, recovered from Unit I, resembled specimens recovered from both Paleo-Indian and Early Archaic components, and, consequently, this tool may represent the oldest artifact recovered from the Cox's Woods site. One of the tools was a combination end- and side-scraper that was formed from an incomplete triangular point preform, and another was a combination scraper-graver (Noel Justice, personal communication, 1994).

Seven examples of humpback bifaces (Plate 3b) were identified. These were small, triangular, coarsely-shaped tools with thick cross-sections and distinctive nodes or "humps" on one or both faces. These bifaces vary greatly in appearance with some examples closely resembling thick triangular points and others more like small preforms or point blanks. All specimens in the Cox's Woods assemblage, however, show at least some evidence of retouch and undoubtedly represent finished artifacts. Similar specimens of this artifact class have been recovered from numerous late prehistoric sites in the riverine Midwest and Great Lakes areas. The true function(s) of these tools remains uncertain; however, the interpretation offered by Jeske (1992), which proposes that humpback bifaces served not one but a variety of purposes, appears to be the most parsimonious at present.

Whole and partial specimens of 18 chert drills were recovered during the summers of 1993 and 1994 (Plate 3b). All but one of the complete or nearly complete examples were manufactured from used triangular projectile points (Noel Justice, personal communication, 1994). All these drills have thin, expanded bases and bit ends with elliptical to trapezoidal cross-sections. No examples of bipointed or "spike" drills, which occur on some Fort Ancient sites (Railey 1992:144-146) were recovered. The one drill made from a non-triangular point (Unit QQ, level 7) was manufactured from heat treated Muldraugh chert (also called Knobs). This specimen far exceeded in size all of the other drills recovered from the Cox's Woods site. This specimen measured 84.8 mm in length and had a base that had been resharpened into an end-scraper measuring 20.7 mm in width. The long, relatively narrow bit was trapezoidal in cross-section. In all likelihood this item appears to be Archaic in origin and represents a mixing of the earlier site components with the Oliver Phase materials.

The lithic collection also contained 28 (complete and fragmented) bifacially flaked pieces of chert for which no certain functions could be determined. These bifaces were all made from large to medium-sized flakes and ranged in form from coarsely chipped, amorphous flakes to rather thin, tear-drop shaped specimens. Among many archaeologists, such bifaces are referred to simply as "preforms" or "blanks", which may describe the function of the more crude examples from the Cox's Woods site. The more refined bifaces, however, might also have served as knives or scraping tools although these functions are not readily apparent from the morphological attributes of the specimens. All of these bifaces were larger than, and readily distinguishable from, the humpback bifaces.

The final non-projectile point category of chipped stone artifact to be described here is the chert core. The complete cores from the excavations were divided into the following sub-categories on the basis of differences in overall form: plano-convex (n=9), biconvex (n=5), polyhedral (amorphous) (n=13), bipolar (n=3), and unclassified fragments (n=3). All cores exhibited multiple flake scars and included examples from both nodular and tabular sources. Several specimens exhibited evidence of battering on one or both polar ends, and may have been utilized as chert hammers (Noel Justice, personal communication 1994).

Most of the diagnostic projectile points (Plate 3b) recovered from the site were of the late prehistoric, Madison type (n=79) (Justice 1987). For purpose of analysis, sixty-one of the most complete specimens were measured for the attributes of basal width, length (complete specimens only), and thickness. In addition, the sample was sorted into the following three categories of basal form: straight, convex, and concave. The results of this analysis conducted by Timothy M. Wright are presented in Table 4 1993 and 1994.

The metric analysis of the triangular points showed a remarkable consistency in form and size. Basal widths and maximum lengths exhibited standard deviations of 2.4 and 3.4 mm with respective means of 15.8 mm and 25.3 mm. In a similar fashion, thickness measurements were extremely uniform with a standard deviation of only 0.8 mm around a mean value of 4.0 mm. These points exhibited straight to slightly incurvate lateral edges with no evidence of serration.

The analysis of the basal forms of triangular points has been used as a means to seriate triangular projectile points from Fort Ancient sites in Kentucky (Railey 1992), Ohio (Litfin et al. 1993), and West Virginia (Graybill 1981). These studies appear to support the interpretation that the modal basal shape of Fort Ancient Tradition points changed over time from convex, through straight, to concave in form. The frequency distribution of basal shapes for triangular points from the Cox's Woods site revealed the popularity of convex based points (n=28) and points with straight bases (n=25) to be roughly equivalent, and both far outnumber the concave forms (n=4) (Table 5). If these results are compared with existing seriations of Fort Ancient points, then the parity of convex to straight based forms in the Cox's Woods sample would appear to place the site occupation roughly contemporary with the middle Fort Ancient components (ca. AD 1250-1400) in the central Ohio River Valley. This interpretation is supported by the stylistic attributes of pottery from the site as well as the two fourteenth- century radiocarbon determinations.

Finally, a study of the lithic raw material used in the manufacture of triangular projectile points from the Cox's Woods was undertaken as part of a broader investigation of chert utilization during the Oliver Phase. The sample of triangular points was examined by Donald Schmidt, an archaeology student, with the assistance of Noel D. Justice (Curator of Collections, GBL) (Schmidt 1993). Each projectile point was examined macroscopically in an effort to identify the kind of cherts used in its manufacture. Identifications were made using an extensive chert reference collection housed at the Glenn A. Black Laboratory as well as published descriptions of southern Indiana lithic resources contained in the archaeological literature (Munson et al. 1983; Munson and Munson 1984; Cantin 1989). The source areas for each chert type were identified using provenience information archived with the Laboratory reference collection, and these data were supplemented with information derived from a 1989 study by Cantin (Schmidt 1993). The resulting identifications of chert types are included in Table 3.

In summary, the triangular points from the Cox's Woods site were manufactured from a selection of raw materials that ranged from medium to high quality for flaking. These cherts were obtained from source areas located between three and thirty-five kilometers from the village. The analyzed sample was dominated by high quality materials such as Wyandotte (n=10), Holland (n=10), St. Genevieve (n=8), and Indian Creek (n=8) with fewer numbers of points made from lesser quality, fossiliferous cherts such as Haney (n=5), Harrodsburg (n=5), and others (n=7) (Table 5). This apparent predominance of cherts such as Wyandotte or Holland may indicate a decided preference on the part of the village flint-knappers for high quality raw materials. However, these results may equally be the result of the relative proximity of these excellent chert sources to the Cox's Woods site. In fact the Cox's Woods site is closer to these southern Indiana high quality source areas than are most other Oliver Phase villages (such as the Clampitt site) where local and semi-local lithic resources are of lesser flaking quality (Schmidt 1993; Redmond 1994).

A total of 60 non-triangular projectile points and fragments were recovered during the investigation (Plate 4a). Of this sample, 17 points were identified as to type and temporal affiliation (Appendix A). These points reflect a sequence of non-permanent habitation which began in the Early Archaic period (as represented by a St. Albans Side-Notched point, a Lake Erie Bifurcate, and a Fox Valley point (Justice 1987)) and ended sometime in the Late Woodland period (as documented by one Racoon Side-Notched point (Justice 1987) (ca. 8000 B.C. to A.D. 900).

Nine of these non-Oliver Phase points were recovered in the northern portion of the site, in proximity to Lick Creek, and therefore may testify to the occupation of small seasonal campsites which focused on the exploitation of aquatic resources in the immediate vicinity. On the other hand, all of these points were recovered from secondary contexts (i.e., village midden deposits) and, therefore, could have originated anywhere on the site. In any case, the lack of evidence for extensive pre-Oliver Phase cultural deposits within the site boundaries would seem to suggest that all earlier occupations were short-term and rather ephemeral.

The groundstone tool assemblage from the Cox's Woods site contained nine sandstone abraders that exhibited deep grooves resulting from the sharpening of bone, antler, or stone implements (Plate 4b). All sandstone abraders originated as amorphous chunks of local sandstone rock and appeared to have been expediently manufactured (but definitely curated) tools. A fragment of a drilled sandstone discoidal (Plate 4b) and a small fragment of a sandstone mortar or metate were recovered as well. Two small spherical objects, one made of sandstone the other of an unidentified stone, were identified as hammerstones based on the presence of small pits and abrasion marks on the surfaces of both. Three anvil or pitted stone fragments manufactured from sandstone, possibly quartzite and granitic rock were recovered from the site. In addition, four igneous-type hammerstones (one ground surface) were identified. These hammerstones and granitic anvil fragment represent the only examples of artifacts made from crystalline rock found at the site. Finally, a large fragment of cannel coal that showed the effects of grinding and polishing along one margin was recovered (Plate 4b). The purpose of this artifact remains unknown; however, the presence of an object manufactured from such a non-local raw material makes it worthy of note. Cannel coal objects of more recognizable form (including trianguloid-, diamond-, or claw-shaped pendants) have been recovered from Fort Ancient sites in Ohio, West Virginia, and Kentucky (Griffin 1966; Graybill 1981; Turnbow 1992).
[return to part 3 menu] [return to main table of contents] [continue to bone]