Mapping a Novaculite Quarry in Hot Springs National Park

Novaculite quarries in the Ouachita Mountains of Arkansas and Oklahoma were created through largescale ex traction of lithic raw materials, used for stone tools by Caddos and other Native Americans over the past 11,000 years and in recent centuries by Euro-Americans for whetstones. Quarry sites are characterized by surface features like large pits, trenches, battered boulders, and debri s piles. This article summarizes the results of an Arkansas Archeological Survey research project that described and mapped surface features at one site (3GA22) to provide a better understanding of the problems and potential of documenting novaculite quarries.


Introduction
Novaculite, outcropping in the Ouachita Mountains of Arkansas and Oklahoma, ha s been valued as a raw material for stone tools for millennia. Ancient novaculite qua1Ties form some of the largest archeological sites in this area, but their investigation has been hampered by their large size and scale and their inaccessible mountain ridge-top location s. The Arkansas Archeological Survey has recently completed a project with the U.S. Forest Service to create an overarching research design to guide novaculite quarries research in the Ouachita Mountains (Trubitt et al. 2004 ). The 2002 mappin g project at 3GA22 was an offshoot of this effort, and was designed to explore techniques and methods for documenting quarry sites, as well as to periods of use fo r possible nomination to the National Register of Historic Places.

Some Background on Novaculite and Novaculite Quarries
the site boundaries an d The Arkansas Novaculite formation, made up of massive layers of novaculite and interbedded chert and shale, outcrops on ridges in the sou them Ouachita Mountain s alon g a 200 mile stretch from west of Little Rock, Arkansas, to Broken Bow, Oklahoma (Holbrook and Stone 1979). Like chert or flint, novaculite is a microcrystalline sedimentary rock formed by chemical precipitation of si liceous mi nerals, but Arkansas novaculite is thought to have been altered to some extent by diagenesis and metamorphosis during the formati on of the Ouachita Mountains (Holbrook and Stone 1979;Keller et al. 1985;Steuart et al. 1984 ). Novaculite is described as "a homogenou s, mostly whi te or light-col ored rock, translucent on thin edges. with a waxy to dull lu ster" (Holbrook and Stone 1979:2). However, there is some variation in color, texture, and luster. Colors from wh ite and gray to pink, red. tan, and black can be found (even at the same quarry). Texture and luster vary as well, from the hard fine-grained "Arkansas stone" to the more porous "Ouachita stone," to the weathered calcareous novaculite or tripoli (Griswold 1892:57-58 89-95. 103: Holbrook andStone 1979:4-5). The translucency of novaculite is often used as a sorting criterion (Jeter and Jackson 1994: 160) but trans lucency is rel ated to texture and varies in novaeulite (Luedtke 1992:69 and Appendix B). • Chipped stone tools made from Arkansas Novaculi te are ubiquitous on archeological sites in the Ouac hita Moun tains and adjacen t areas of the Gulf Coastal Plain in southwest Arkansas and southeast Oklahoma (e.g .. Early I 988; Schambach I 998 : Waddell et al. ! 995: Wyckoff I 968). As Holmes (1 89 1 :313 ) expansively described: ·' ... hundreds of square miles are strewn with flakes. fragments. failures and rejected pieces ! of novacu lite I, and the country around. from the mountains to the gulf. is dotted with the finished forms that have bee n used and lost.'' [n the mountains, novaculite could be procured directly fro m outcrops. from surfaces of talus slopes below ou tcrops, and in the form of cobbles from ri verbed gravels (Coleman et al. 2000;Marrin 1982;Waddell et al. I 995). Away from the Ouachita Mountains in southern Arkansas, eas tern Oklahoma. and northeastern Texas. novaculite could be obtained either from riverbed cobbles from the rivers flow ing out of the mountains or from grave l deposi ts in Pleistocene terraces. Novaculite could have a lso been procured from outcrops back in the Ouachitas an d transported directly or traded into these regions (Hemmings l 982: Perin o and Bennett 1978;Perttula 1984;Waddell and King 1990).
In west-central Arkansas, the use of novaculite spans the range of prehistoric periods (e.g., Schambach 1998). There is some evidence that the heaviest use was during the Archaic period (e.g., Baker 1974:28-29), perhaps as novaculite became part of the Late Archaic period Poverty Point exchange system (Jeter and Jackson 1994). Different regions show different temporal trends. however. In the Felsenthal region (Ouachita River valley in southern Arkan sas), novaculite use was higher during pre-Mississ ippian periods than during the Mississippian period (Hemmings 1982: 242-244;Kelley 1984). But in a study of Lake Fork Reservoi r sites in northeastern Texas. non-local lithics (including novaculite) had low freque ncies in Archaic period assemblages and hi gher frequencies in Caddo contex ts (Perttula 1984: 137-I 39). In Hot Springs. Arkansas, quarrying novaculite for whetstones became an important local industry earl y in the nineteenth century and continued during the twentieth century (Gri swold 1892; Whittington l 969). Novaculite and tripoli are still bei ng mined commerciall y today (Steuart et al. 1984 ).
Nineteenth-century descriptions of novaculite quarries (Featherstonhaugh 1968[ 1844 j: l l0-11 I; Griswo ld ! 892: 175-1 76; Holmes 189 l, 1974[ 19191: 196-200;Jenney 1891) li nked their use to the Indians previously living in the region and noted evi dence of novaculite tool manufacture at habitation sites in the area. W. H. Holmes ( 1974) used novaculite quarries as examples in his treatise on aborig inal stone quarrying in North America. Only brief descriptions appeared in the literature in the mid-twentieth century (Lemley l 942; Whittington l 969).
More recent archeological investigation of novaculite quarries and workshop sites has focused on recording new si tes and exami ning the spatial di stribution of novaculite tool-work ing acti vities. A l 973 reconnaissance of Hot Springs National Park by the Arkansas Archeological Survey recorded several new novaculite quarry and tool manufacturing sites (Baker 1974, l 982). As part of his thesis research, Baker ( 1974,1982) also conducted limited test excavations at a qu arry si te near Magnet Cove (3GA+8/3HSl58/3HS433). Recent surveys of U.S. Forest Service lands have recorded numerous quarry sites and examined the patterning of novaculite reduction activities across the landscape (e.g., Waddell and Waddell 1992;Waddell et al. I 995;Williams et al. 1993 ). Forest Service archeologists have been actively (focumen tin g novaculite quarry and workshop sites. Etchieson ( 1997) describes the types of large-scale features found at quarry sites. including pits and trenches, battered houlders and outcrops, artificial benches created by surface stripping to expose novaculite, shelters or caves enlarged quarrying novaculite. and trails to quarries. Coleman 's (2003} anal ysi s of the lit hic assemblage from test excavations at a Middle Archaic period novacu!ite workshop site investigated ideas about huntergatherer mobility patterns and technological organinttion. During the Arkansas Archeological Survey/Arkansas Archeo!ogical prog.ram at L,1ke in 993, a nov aculite quarry site (3 PLJ49 ) was mapped and tested (Coleman ct al. J 999: Hilliard l 995). A more detail ed analysis of the novac ulite de bitage from th ese excavations is c urrentl y underway al the Survey ·s Henderson Researc h Station .

Previous Research at 3GA22
Site 3GA22. a novacu lite quarry in Arka nsas' s Hot Springs N ational Park. has been known to the arc heological community for over 100 years. W. H. Holme s (1891:314) described the site as "a number of and excavations dug in and about th e crest of th e ridge. This rid ge is a solid for mation of the novaculite \veathering out in irregular grayish fli nty -loo ki ng masse~ which protrude from the cres t or projec t on th e slopes, formin g short broken cl iffs from ten to twenty fee t in height. " Two of the largest p its. one a bout 150 ft. (ca. 46 m) in di ame ter and 25 ft. (ca. 8 mi deep, th e o ther larger but shallower, were formed by quarrying a novac ulite o utcrop down ward s from the surface an d discarding debris from knapping bl anks arou nd th e pit edge s (Holmes ! 1974: Figure 77] marked these pits "IY ' and "C ' on hi s map, shown here as Figure I) . Holmes ( 1891 ) hypothesi zed that fire was used for qu arrying, based on his o bservance of " blacke ned patches" on somt~ undercuts (Holmes 189 i .315 ). Q uanti ties of knapping debris were see n around the pi ts and on level areas of the ridgetop (the ''Great Worksh op"), and Ho lmes (1891 :3 15 and Plate 111 ) also hypoth es ized that bifacial blanks and preforms were taken from the quarries to be fini shed into tool s at other sites . He notes some evi dence of recent di sturban ce at the site by people searching for old Span ish gold mines , but writes : '"That the recent work has not seriously changed the contour of the ancient quarries is evide m from the fact that the entire mass of ejected material, interior and exterior, is composed of the parti all y shaped fragm en ts derived from ancient flakin g" (Holmes I 891 :3 l 5 ). Following visits to the site by local amateur archeologist F~)rest Sargent ( l 966), 3GA22 was recorded in the state site files in 1969 by Sargent and James Scholtz. Collections made at that time included hammerstones, novaculite debris, and worked pieces of novaculite (AAS/HSU Accession 69-40, 70-655). The presence of nearby abandoned whetstone quarries was noted. Baker ( 1974,1982) revisited the site during his master's thesis research in !973. He relocated Holmes' "'Great Workshop" and the novaculite "ledge" and quarry (Baker 1974:Figure !J, and extended the site boundaries on the eastern side with the addition of seven quarry pits, abou t 2-4 m in diameter and l-2 m deep with ''chipping debris, tool blanks, and hammerstones'' around them, scattered over an approximately 3 acre area (Baker 1974: l l; JGA22 site fonn; this area was 01iginally designated as JGA46 but is now included in 3GA22).
The nov aculite source on this mountain has provided raw material samples for published novaculite heat treatment experimen ts. Flenniken and Garrison (I 975) used a large piece of white novaculite to make bifaces and then heated them to 200°C, 450°C, and 500°C (392°F, 842°F, and 932°F). No change was seen at 200°C, but the 450-500°C samples showed a change in microscopic texture that resulted in a glossy appearance and improvement in knapping characteristics. Sollberger (in Bennett 1986:69-73) found that pieces of white novaculite showed a change to a pink or red color with an increase in luster, along with an increase in workability, when heated to 620°F (327°C). It should be noted that a wide range of novaculite colors can be seen in debris at 3GA22, but it is not clear whether this range reflects natural variation of the outcrops or the use of heat treatment.
There are archeological sites further east of 3GA22 along the mountain ridge (3GA47, 3GA 135, 3GA 138, 3GA141, and 3GA832-840) that have evidence of aboriginal quarrying in the form of shallow pits and trenches and scatters of chipping debris. In addition, there are unrecorded sites that reflect the nineteenth and twentieth century mining of novaculite for whetstone raw material (Griswold 1892:308;Whittington 1969:228). One of the companies operating in this area was Arkansas Abrasives, Inc., managed by Frank Thompson, which marketed a line of whetstones beginning in the l 950s and expanded into industrial abrasives and novaculite finishing media, later doing business as Buffalo Stone Corporation (Ms. Mary Little, 2002 personal communication; Mrs. Frank Thompson, 2002 personal communication). The overlapping of ancient quarry features by more recent activity at novaculite outcrops is typical of the Hot Springs region; nineteenth and twentieth ce ntury quarrymen often found and collected hammerstones during their work, and they likely started their operations in or near ancient quarries.
Recently, Mark Blaeuer ( 1995) of the National Park Service prepared a draft National Register of Historic Places nomination for 3GA22. Two issues raised during this process were the need to clarify the boundaries of the site and the need for more infonnation on the periods of use of the site. The 2002 mapping project at 3GA22 was intended to contribute infonnation for the first of these issues by describing and mapping surface features.

The 2002 Mapping Project: Methods and Results
New research at 3GA22 was designed to map the cultural features associated with novaculite quarrying activities in an effort to refine the site boundaries and to test techn iques and methods for researching this kind of large-scale lirhic extraction site. Creating measured maps of quarry sites that indicate cultural features such as battered outcrops and boulders, quarry pits, and debris scatters, is the first step in assessing and comparing the scale and types of lithic extraction and procurement activities, as Hatch (l 994) has shown for jasper quarries in Pennsylvania.
The mapping project at 3GA22 was done under an Archaeol ogical Resources Protection Act permit (MWR-02-1) obtained throu gh the Hot Springs National Park and the Nation al Park Service ' s Midwest Archeological Center. The fie ld work took place over six da ys between February 15th and March 8th, 2002. Field personnel included Mary Beth Trubitt and Kate Wri gh t (Arka nsas Archeological Survey), assisted by Harry Hammond, Mary Little, Mildred Grissom, and Patrici a Heacock (Arkansas Archeological Society, Ouachita Chapter), and Michael Head and Jeffrey Gaskin (stude nts from Henderson State Uni versity). Mapping was done with a combination of techniques and equipment, including an electronic total station (Topcon GTS -2 l 1 D) with stadia/target, a hand-held global positioning system unit (Garmin eTrex Vista), and standard metric tape measures. After field maps were produced using Golden Software's Surf er program and Map tech' s TeJTain Navigator software. A fin al report was prepared and submitted to the National Park Service (Trubitt 2003).
Several logistical problems were enco untered during field work. Visibility for topographic mapping in a wooded environment and identifi cation of surface feature s was maximized by scheduling the field work for the late winter/early spring before trees and vines fully leafed out. Site accessibility was a factor, as equipment had to be caJTied from the vehicle up about 300 ft. of elevation to the site. A backpack for the total station and multiple crew members made it possible to carry equipment up the mountain. Another factor was the size of the site. While a total station has a much greater useful range than a transit in open areas, trees and rough terrain made multiple setups or mapping stations necessary. Topographic mapping only included the area of the site covered by the Holmes map . Outcrops further west, and areas of modern quaJTying to the south and east, were plotted on a digital quad sheet using the GPS unit. In addition, the novaculite outcrops on the site's northwest side presented mapping hurdles because the irregular and steep topography was difficult to access with equipment in the field and difficult to portray using the mapping software. Finally, differentiating between the older pits from Native American toolstone procurement activities and newer pits from Euro-American whetstone mining activities can be problematic. In the absence of archival research, detectable smface artifacts such as hammerstones or metal quarry tool s, or excavations to identify characteristic debitage, I used rectangularity and steepness of pit sides as an indicator of modern quarry features , or features modified by nineteenth and twentieth century whetstone quarrying.
Points taken with the GPS unit indicate the more ancient and more recent quaJTy features of 3GA22 extend about 1000 m along two ridges on the mountain between about 800-1100 feet amsl. A total of 18 quarry pits or trenches were plotted along the southern ridge; based on steepness of pit sides, square or rectangul ar outlines. and proximity to an old access road, these are identified as nineteenth and/or twentieth century features. Additional pits may be identified with more field survey. Examining the associated debris would give a more conclusive interpretation of period of use (some have flake s in the debris piles and may have overlapping ancient and modern use).
An area of about 250 m north-south x 200 m east-west c01Tesponds to the main area of 3GA22 used for novaculite toolstone quarrying, as indicated by surface fl aki ng debris, shallow to deep circular quarry pits, and worked/battered novaculite outcrops. In addition to GPS point plotting, this area was mapped topographically the total stati on (Figure 2), and corresponds to the area mapped previously by Holmes (see Figure l ). A total of 11 quarry features was identified in this area (Table 1 ), Most are pits or trenches (F-2, 3, 4, 5, 6, 8. 9. 10, 11 ), one is a concentrated area of debris on the surface (F-1, Holmes' "Grear Workshop"). and one is an outcrop with evidence of working (F-7). The quaJTy features are concentrated on the northern ridge. Features on the southern ridge include a shallow ov al quarry pi t (F-4) that is thought to be from toolstone quarrying, two depressions (F-10, 11 ) that appear to be from more recent whetstone quarrying (or older pits with more recent disturbances), and a small depression (F-9 ) th at may be either a quarry feature or a large tree fall. In a flat saddle between the two ridges is an extensi ve area covered with chippi ng debris (F-1 ). CAnDoAs ARCHEOLOGY Jot:R"1AI. •, --A 'spring' is shown on the Holmes map and was mapped in 2002 as we ll. There was some water found in the bottom of this gully during our field work in February-March, but none at the tim e of a July site vis it, so thi s may be seasonal water drainage rather than a perennial spring. Ancient quarry workers may have preferred novaculite outcrops with nearby springs or seeps, since working on a mountaintop for any length of ti me would require water (carried up or acce ssible nearby). Sprin gs have been noted near ot her novacu lite quarri es visited in the Ouachita Mountains, and I suspect that this is not coincidental. Feature I, Holmes ' "Great Workshop," is a large sprawl nf novaculite debris in the saddle of flatter land between the ridges and gull ies. This area is literally covered with novac ulite debitage. highly visible because of its density and a rece ntly re-cut right-of-way access road. The F-1 boundaries approximate the area of heaviest exposed debris (novacu lire flakes and shatter. as well as chipped novaculite bifaces/fragments and sandstone hammerstones/fragments), and novaculite continues beyond this on the surface. During our field work. we mapped and described (but did not collect) a total of 17 sandstone hammerstones or fragments and one novaculite dart. point ( Table 2). The ham merstones are made of a raw material that is not fo und on the ridgetop forma tio n bu t was rran sported from lower elevations. The dait point, the on ly Jiagnostic identified during the tie Id work. is typed as a Fourche Maline period Gary. mr. Le Flore/Bodcaw point (Schambach 1998:58-59. l 18-121 ).

Great Workshop
A B E C, D G The main area of quarry in g on the northern ridge ( Figure 3) has se veral distinct pits (F-2. 3. and 5. Fi gure 4). a large pit fi lled with piles of debris (F-6. Fi gure 5), and adj acenr exposed outcrop (F-7, Figure 6). The adjacent oval pits labeled F-2 and F-3 (Figures 7 and 8) are separated by a li ne of large novaculite boulders that appears to be the surface of a novaculite outcrop left in pl ace whi le adjacent seams were quarried out. A scatter of novaeulite boulders lie on the surface of the ridge j ust east of these two pits. The F-6 pit and F-7 outcrop area (see Figures 3, 5-6) is a large feature complex created by quarrying to ex pose a seam of high -qual ity toolstone visi ble at the base of the outcrop (Baker I 974: 10-1 l ). Some debris from this quarry ing was left in pi les within F-6 wh ile other debris forms an extensive talus slope on the south side of the ridge.
It is clear that the northern ridge was extensively remodeled by the quarrying and knapping activities that took place here in the past. A three-dimensional view of the northern ridge (Figure 9) shows that so much novaculite was removed that over 50 m of the ridge crest has been oblitera ted. Presumably this acti vi ty took place over the span of thousand s of years. The one diagnostic recorded during this fi eld work suggests use at least during the Fourc he Mali ne peri od (ca. 650 B.C. -A.D. 950). Dating the use of this complex of quarry features remain:-as a major research question for this site. and may he answerable through exi:avation.

Conclusions
Ncwaculite quarries are a major site type in the Ouachita Mountain:-.. and one that pre~enb numerou~ logistic and methodological problems a:-well as great potential for inwstigating past human activitie:-, in this region. There are certainly logistical problems to deal with when researching these large mountaintop quarry sites. Tree foliage limits ,i-,ibility of surface features. so surveying during late winter/early spring is optimal for visibility and weather conditions.
The size and scale and rugged terrain of novaculite quarries makes mapping these sites more difficult. Three mapping methods were used during this project. Plotting GPS points taken at quarry features provided an indication of the extent of the site and its location on the geological quadrangle map. but not at sufficiently large scale to show details of the site or its features. Higher resolution global positioning system units could be used for more detailed site mapping. A topographic map was created of the main quarry feature area using a total station. In conjunction with this, more detailed sketch maps of individual quarry pits were made using standard tape measures. These techniques showed the relationships between features and between features and the landscape. However, novaculite outcrops with extreme topographic variation proved difficult to show on the maps produced. and the more detailed mapping was time-consuming.
After the field work was done, several kind" of aerial photographs of Hot Springs National Park were examined . While exposed novaculite in the modern quarries shows up strongl y against the wooded vegetation of adj acent areas. the smaller exposures. such as the F-2. 3. 5. 6. and 7 complex mapped at 3GA22. were not visible . In the future it may be worthwhil e lO use a combination of high resolution aeri al photograph-, and more accurate GPS technology to generate large-s,:ale map~ that focus on the locations and situation of cu ltural features. Since quarry feature~ are often scattered over large areas. detailed topographic mapping using a total station might focus on configurations of the feature,, themsehes (quarry pits. talus slopes of debris. worbhop area~ 1.  • Mapping of novaculite quarry sites is an important first step in unde rs tandin g the size and scale o f these s ites and the types of ex tractio n tec hniques used that resulted in °features of d iffe re nt types observable on th e surfaces. Presumably the objecti ve was simi lar from quarry to quarry -to expose high quality no vacu lite and extract it from the surroundin g rock or soi l -but the situation differed within and between mountains and left be hind differing surface alterati ons . It is clear that novaculite quarrying for toolstone was done by Caddos and their ancestors, and perhaps other Indian groups as well. over a period of thou sands of years. One of the o ngoing research problems is dating the period(s) of quarry use. Diagnostic artifac ts found elsewhere indicate that novaculite was the favored raw material for chipped stone tools for most of the long prehistoric sequence in th is region. It is not c lear, howe ve r, whether quarrying was actively done at ou tcrops throug hout this time span, or whether riverbed cobble sources or talus or surface pieces were used during specific time peri ods. If no vac ulite from outcrop was e xchanged beyond the local area, there remains the question of whether local people obtained and then traded it or whether outside groups came to get their own supplies directly.
Surface diagnostics at quarry sites are few, especially after collecting by vis itors over the last two hundred years. Excavations of quarry features and works hop areas may produce diagnostic artifacts and/or sam ples suitabl e for datin g, but sampling strategies for excavati on and analysis must be developed to deal with the large amounts of artifacts uncovered (e.g .. Torre nce 1984). In general, few finished tools are expected at most quarry si tes , compared wi th the quantities of quarrying and chi pping debris.
Unlike most other North American che,ts or flints , novacu lite has continued to be quarried or mined into the modern era. Qua rrying novaculite for w hetstones began in the early 1800s around Hot Springs, and has continued, along with other commercial uses, through the twentieth ce ntury. Quarry sites (includin g 3GA22) show e vidence of use by both Euro-Americans and Indians in different time periods. and the more recent rock quarrying may have oblite rated older portions of this and other sites . Distinguishing o lder quarry features from the more rece nt ones could be done by examining debris, since distinctive.tools and technologies were used by roolstone quarrying versus whetstone mining. While angular pieces of no vacu li te debris characteri ze the bulk of the natural talus and the cultural waste piles in and around quarry pits and outcrops. close examination may expose hammerstones and identifiable fl akes versus metal tools, drill marks for black powder charges, or di stincti ve whetstone debris (although this may not be so straightforward; see Martin [ l 982: l 20] for a cautionary tale of " flakes" resulting from modern tools used to test stone quality pri or to whetstone q uarryi ng).
Nume ro us no vaculite quarries a re preserved on U.S. Forest Service and Nationa l Park Service lands. Using both old and new field techniques to document these sites has benefits both for researching novaculite tool production and exchange systems and for inte rpretation of these historical resources for the public. While there are many logis tical probl e ms, the potential fo r learning about past lifeways is significa nt. Specific research areas inc lude sourcing studi es to tie novaculite artifacts to their source quarries: investigatio n of quarry ing tools and techniques ; analysis of the byproducts of novac ulite tool manufacture at quarries, workshops, and habitation sites to understand the spatial and social organization of production; analyzing the distribution of novaculite tools to interpret the mechanisms for novaculite exchange; and investigating the people involved in these activities us both producers' and consumers ' (Trubitt et al. 2004). Novaculite quarries are but one part of the tool production and exchange systems. When novaculite qwmying is conceptualized as part of these larger cultural systems, we can shift focus from artifacts to the people who made and used these tools. The spaci al distributions of novaculite artifacts show in a materi a l for m the relationships between groups of interactions between people on a regional and inter-regional scale.