Lawrence E. Abbott, Jr.

Paper contributed to The North Carolina Lithics Conference, The Uwharries, 1999
February 24-27, 1999, Randolph Community College, Asheboro, North Carolina

The Three Hat Mountain lithic quarry, 31DV51, is located within the central Piedmont of North Carolina along the western edge of the Carolina Slate Belt. This quarry was used by Archaic groups as a source of raw materials for the production of stone tools. This paper presents the results of research to document the range of variation of lithic raw materials present at 31DV51 using expedient, easily replicated, field and analytical methods involving random sampling and macroscopic variables.

The Three Hat Mountain lithic quarry is located in Davidson County, North Carolina along the western edge of the Carolina Slate Belt (Sundelius 1970). Three Hat Mountain is composed of three peaks within the Flat Swamp Member of the Cid Formation (Stromquist and Sundelius 1969; Milton 1984; Harris and Glover 1988). A geologic map of the area shows the mountain within a complex zone of metasedimentary and metavolcanics which consist mainly of felsic (tuffs and rhyodacites ) rock, bordered to the south and west by mafic volcanic rock (Carpenter 1982). The deposits that make up Three Hat Mountain are Precambrian in age (> 570 mya).

Lithic raw materials outcrop on the ground surface across the mountain in sizes that range from hand-size nodules to boulders many cubic meters in size. Prehistoric cultural debris have been reported by local collectors (Gary J. Curry, personal communication to Lawrence Abbott, 1981) and recorded by professionals (Mountjoy and Abbott, 1982; Abbott 1987, 1993). Cultural debris occurs mainly as refuse resulting from primary and secondary lithic reduction acivities. The cultural debris is spread over approximately 250 hectares (625 acres) of the mountain surface (Mountjoy and Abbott, 1982; Abbott 1987). The debris is interspersed with primary and secondary core fragments (Bradley 1975) and aborted quarry blades (Coe 1964).

Archaeological excavations were conducted at the site by the University of North Carolina at Greensboro in 1975 (Mountjoy and Abbott 1982). A single two meter square excavation unit was dug on the west slope of the mountain within an area of high concentrations of surface artifacts. Five distinct strata were observed and over 16,000 artifacts were recovered to a depth of 1.3 meters below ground surface (Mountjoy and Abbott 1982). The range of projectile points suggests the site was most frequently used by Archaic groups. The greatest frequency of these temporal diagnostics are related to the Middle (Guilford and Morrow Mountain) and Late Archaic (Savannah River and Types II, III, and VII quarry blades) (Coe 1964).

In general, the information above suggested that raw material procurement, along with primary and secondary lithic reduction activities were carried out across the mountain. Nodules of raw material protruding from the ground surface were collected and converted into primary and secondary cores and quarry blades. Much of the raw material appeared to have been transported from the mountain in these forms.

Three Hat Mountain is probably not unique as a prehistoric quarry within the Carolina Slate Belt. Numerous quarry sites have been documented throughout this general area (Sellon 1980; Baker 1980; Hargrove 1989; Daniel and Butler 1991; Daniel 1994; Abbott and Harmon 1998). What is unique concerns the Slate Belt itself as a geological formation used over time as a source of lithic raw material to supply populations moving and living, not only within, but also outside its boundaries. Unfortunately, the widely variable and metamorphosed nature of the mineralogical and structural elements that compose these rocks make them difficult to analyze in terms of lithic sourcing. As Davis (1992) notes, the main problem is that the basic mineralogy and geochemistry of the lithic components are similar within a range of various elemental combinations, but highly differentiated as a result of successive metamorphic and depositional episodes. In addition, these lithic components are widely spread across the Slate Belt and intermixed within formations rather than localized to specific outcrops. As a result, while it is not particularly difficult to recognize lithic material from the Slate Belt, it is difficult to determine a specific source of a given material. Because of this difficulty, very little work regarding lithic sourcing has been attempted within the Slate Belt until recently (Novick 1978; Abbott 1987, 1993; Daniel and Butler 1991; Daniel 1994).

Beginning in 1982, baseline research was conducted to address this problem (Abbott 1987). The basic research question was very simple. It consisted of whether it was possible to document the range of variation of raw materials available at a known prehistoric quarry (Three Hat Mountain, 31DV51) and see if that range could be identified on other non-quarry sites located outside of the Slate Belt. The primary challenges facing this research were first, how to adequately sample the naturally occurring raw materials from a large prehistoric quarry; and second, how to analyze and describe the range of variation. These challenges were met using a random sampling technique to collect the data and a set of macroscopic and tactile variables to document the range of variation within the raw material. The field and analytical methods will be discussed in detail below.

The field phase of this research centered on the collection of representative samples from Three Hat Mountain and the sites used for comparison. The data collection phase at the quarry utilized a set of dogleash collection units. Each unit measured two meters in diameter. The placement of these units was randomly selected using a table of random numbers to generate azimuth and distance measurements away from established points along a transect across the site. The transect was located centrally across the mountain and extended from the southernmost peak northwest across the two smaller northern peaks. These three peaks were selected because they were the most visible landmarks in the area and because the distance between the landforms was great enough to provide the number of points necessary to adequately sample the quarry. Each dogleash unit was measured from a separate point (at a constant interval) along this transect. Eighty-five dogleash units were established. Naturally occurring lithic materials outcropping on the ground surface were collected within each dogleash unit. No cultural materials were collected, although the extent of cultural materials was noted on site maps. Hand specimens were collected where possible. Specimens were extracted using a geologic hammer in areas where large boulders comprised the dogleash unit. Sample units devoid of any lithic materials on the ground surface were designated as sterile units. Six of the 85 units (7.06% of the total) were sterile.

The same data collection methods were employed at five sites used for comparison to Three Hat Mountain with the exception that only culturally derived lithic materials, instead of unaltered specimens, were collected. The transect was e stablished through the approximate center of each site. Site centers were established based on the extent of surface debris and the distribution of artifacts in shovel tests. Prior to this the general area surrounding each site was inspected to insure that no outcrops of knappable raw materials were present (this included streambeds and gullies). This was done to support the assumption that all cultural materials present on the sites used for comparison were imported from elsewhere rather than collected from expedient sources nearby. Each site selected for comparison contained temporal diagnostics within the range documented at Three Hat Mountain (Coe 1964).

The classification and analysis of the lithic material was based on visual inspection of the specimens, supplemented with a hand-held comparator and a 10X to 20X stereoscope. Specimens from all sample units were classified according to a set of macroscopic variables. These variables included groundmass, texture, luster, fracture quality, presence or absence of inclusions, weathering, and any inferred anomalous features. The only quantitative category was density. The specific gravity of each specimen was measured using a Jolly scale. This information was coded onto a macroscopic analysis sheet originally developed by Alan N. Snavely in consultation with the late Dr. J. Robert Butler (at that time part of the Department of Geology, University of North Carolina at Chapel Hill).

The information was compiled and sorted into lithic groups using dBase II software. The range of variation of the representative lithic groups at the quarry and the outlying sites was thus established for comparative purposes. Representative hand specimens of each group were examined by Dr. J. Robert Butler and categorized according to basic raw material type.

Twenty-six lithic groups were identified at Three Hat Mountain. The variation in the naturally occurring materials ranged widely from heterolithological breccia (very difficult to knap) to a very fine grained rhyolite with excellent conchoidal fracturing qualities. Other groups within the range included flow-banded rhyolite, coarse grained dacite, andesite, and porphyritic rhyolite.

All groups contained aphanitic grains with iron pyrite, chalcopyrite, and arcenopyrite within the respective groundmass. Grain sizes ranged between microcrystalline and cryptocrystalline. Groundmass matrix ranged between homogeneous and heterogeneous. Translucency was exclusively opaque. Luster ranged from earthy (dull) to semi-glossy (a low exhibition of reflected light). The texture of unweathered surfaces ranged from that of 240 grit sandpaper to a smooth, even surface texture, with very little friction or resistance to movement. Fracture quality ranged from friable to conchoidal. Most of the inclusions consisted of feldspar (plagioclase) and quartz.

Four lithic groups (Groups B, F, N, and O) were identified in nearly two-thirds of the sample units at Three Hat Mountain. Group B was a heterolithological breccia with dark differentially devitrified minerals. Group F was a flow- banded rhyolite with chalcopyrite within the groundmass. Group N was a breccia with undifferentiated pyrite inclusions within the groundmass. Group O was a porphyritic rhyolite with feldspar inclusions.

A distinct pattern emerged in terms of the distribution of lithic groups across the mountain. The distribution of the lithic groups was divided between the two smaller northern peaks where Group O was almost exclusively located, and the large southern peak where the majority of the other lithic groups were located.

The range of variation of lithic materials on the sites chosen for comparison revealed a distance-decay pattern within approximately 29 km in regards to similarities to the range of variation documented at Three Hat Mountain. The data suggested a preference for two particular lithic groups, Group O and Group X. Lithic Group O was the porphyritic rhyolite that dominated the northern peaks at 31DV51. Group X was a very fine grained, cryptocrystalline rhyolite located on the largest, southernmost peak. These materials, particularly Group O, were located the greatest distance away from the quarry at Three Hat Mountain and probably remained within a given cultural system as cores, blades, finished tools, and/or trade goods.

As with most baseline research, this study generated as many questions as it answered. Future work concerning sourcing analysis within the Slate Belt should concentrate on locating the quarry sites and developing viable methods to document the range of variation at these sites. Certainly the usefulness of macroscopic analysis should be further tested at other quarries to determine its viability as a method to describe the range of raw materials. A successful documentation of the ranges of variation for known quarries may provide information regarding source-specific use of certain locations by prehistoric groups.

Sourcing analysis within the Carolina Slate Belt cannot be fully addressed until specific sources can be identified with specific raw materials. The methods presented in this paper provide a cost-effective and expedient means to begin addressing this mammoth task. At this point it is not clear if source-specific analysis, at a comprehensive level, is possible given the complexity of the Slate Belt. However, if attempts are not made to understand what was available to prehistoric people, and how it was collected and distributed over time, then we will never know if it is possible.

For a full discussion of this research, along with the basic assumptions regarding the results, please refer to Abbott (1987). For information (to include photos) regarding the lithic groups documented at Three Hat Mountain, please contact Lawrence Abbott at the following email address: fals1@netpath.net.

Abbott, L. E., Jr.
1987 An Investigation of Lithic Resources Within Certain Sites in Davidson County, North Carolina. Master’s thesis, Department of Anthropology, Wake Forest University.

1993 The Production, Use, and Distribution of Metavolcanic Raw Material From the Slate Belt: An Example From Central North Carolina. Paper presented to the 58^th annual meeting of the Society for American Archaeology, St. Louis.

Abbott, L. E., Jr. and M. A. Harmon
1998 The Acquisition of Siliceous Raw Materials From the Carolina Slate Belt: Intentional or Fortuitous Lithic Procurement Within the South Central Piedmont of North Carolina? Paper presented at the 63^rd annual meeting of the Society for American Archaeology, Seattle.

Baker, M. C.
1980 Archaeological Investigation of Two Prehistoric Lithic Sites in Chatham County, North Carolina. Report submitted to the North Carolina Department of Transportation by Archaeological Research Consultants, Inc., Chapel Hill.

Bradley, B. A.
1975 Lithic Reduction Sequences: A Glossary and Discussion. In Lithic Technology, edited by E. Swanson, Mouton Publishers, The Hague.

Carpenter, P. A.
1980 Geologic Map of Region G, North Carolina. North Carolina Department of Natural Resources and Community Development, Raleigh.

Coe, J. L.
1964 The Formative Cultures of the Carolina Piedmont. Transactions of the American Philosophical Society No. 54.

Daniel, I. R., Jr. and J. R. Butler
1991 Rhyolite Sources in the Carolina Slate Belt, Central North Carolina. Current Research in the Pleistocene 8:64-66.

Daniel, I. R.
1994 Hardaway Revisited: Early Archaic Settlement in the Southeast. Ph.D. dissertation, Department of Anthropology, University of North Carolina at Chapel Hill.

Davis, J. D.
1992 Prehistoric Woodland Lithic Exchange in the Great Bend Region of The Yadkin River Valley. Master’s thesis, Department of Anthropology, Wake Forest University.

Hargrove, T.
1989 An Archaeological Survey of Morrow Mountain State Park, Stanly County, North Carolina. Report submitted to the North Carolina Division of Parks and Recreation by Archaeological Research Consultants, Inc., Raleigh.

Harris, C. W. and L. Glover, III
1988 The Regional Extent of the ca. 600 Ma Virgilina Deformation: Implications for Stratigraphic Correlation in the Carolina Terrane. Geological Society of America Bulletin 100:200-217.

Milton, D. J.
1984 Revision of the Albemarle Group, North Carolina. U. S. Geological Survey Bulletin 1537-A:69-72.

Mountjoy, J. B. and L. E. Abbott, Jr.
1982 An Archaic Quarry and Stone Knapping Location on Three Hat Mountain, North Carolina. In Collected Papers on the Archaeology of North Carolina, Archaeological Council Publication Number 19, Raleigh.

Novick, A. L.
1978 Prehistoric Lithic Material Sources and Types in South Carolina: A Preliminary Statement. South Carolina Antiquities 10:422-437.

Sellon, M. R.
1980 Preliminary Archeological and Historical Investigations at Three Alternative Airport Areas in Davidson County, North Carolina. Ms. on file, Office of State Archaeology, Raleigh.

Stromquist, A. A. and H. W. Sundelius
1969 Stratigraphy of the Albemarle Group of the Carolina Slate Belt in Central North Carolina. Geological Survey Bulletin 1274-B.

Sundelius, H. W.
1970 The Carolina Slate Belt. In Studies of Appalachian Geology: Central And Southern, edited by G. W. Fisher. Interscience Publishers, New York.