Transitions : Archaic and Early Woodland Research in the Ohio Country [1 ed.] 9780821442968, 9780821417966

Citation preview

TRANSITIONS TRANSITIONS

ARCHAIC AND

E A R LY

WOODLAND

RESEARCH IN THE

OHIO COUNTRY

Martha P. Otto & Brian G. Redmond

Otto FM:Layout 1

9/21/08

2:23 PM

Page i

Otto FM:Layout 1

9/21/08

2:23 PM

Page ii

Otto FM:Layout 1

9/21/08

2:23 PM

Page iii

ARCHAIC AND E A R LY WO O D L A N D RESEARCH IN THE OHIO COUNTRY

Edited by Martha P. Otto and Brian G. Redmond

OHIO UNIVERSITY PRESS Athens in association with the OHIO ARCHAEOLO GICAL COUNCIL

Otto FM:Layout 1

9/21/08

2:24 PM

Page iv

Ohio University Press, Athens, Ohio 45701 www.ohioswallow.com © 2008 by Ohio University Press All rights reserved

To obtain permission to quote, reprint, or otherwise reproduce or distribute material from Ohio University Press publications, please contact our rights and permissions department at (740) 593-1154 or (740) 593-4536 (fax). Printed in the United States of America Ohio University Press books are printed on acid-free paper ƒ ™ 15 14 13 12 11 10 09 08 5 4 3 2 1

Library of Congress Cataloging-in-Publication Data Transitions : archaic and early Woodland research in the Ohio country / edited by Martha P. Otto and Brian G. Redmond. p. cm. Includes bibliographical references. ISBN-13: 978-0-8214-1796-6 (hc : alk. paper) ISBN-13: 978-0-8214-1797-3 (pb : alk. paper) 1. Indians of North America—Ohio—Antiquities. 2. Woodland Indians—Ohio—Antiquities. 3. Paleo-Indians—Ohio—Antiquities. 4. Excavations (Archaeology)—Ohio. 5. Ohio— Anitiquties. I. Otto, Martha P. II. Redmond, Brian G. (Brian Gerald), 1958– III. Ohio Archaeological Council. E78.O3T73 2008 977.104’97—dc22 2008017893

Otto FM:Layout 1

9/21/08

2:24 PM

Page v

CONTENTS

Preface

vii

chapter 1: Archaic Manifestations in Southwestern Ohio and Vicinity Kent D. Vickery

1

chapter 2: Cultural Resource Management and the Analysis of Land Use Patterns of the Archaic in North-Central Ohio Craig S. Keener, Kolleen Butterworth, and Crystal L. Reustle

29

chapter 3: Down by the River: Late Archaic through Terminal Archaic Dynamics at the Davisson Farm Site (33LE619), Lawrence County, Ohio Matthew P. Purtill

41

chapter 4: Early Woodland Prehistory (1000–1 bc) in the Western Lake Erie Drainage Basin David M. Stothers and Timothy J. Abel

79

chapter 5: Three Early Woodland Occupation Loci in the Chartiers Creek Drainage, Allegheny County, Pennsylvania James A. Robertson, Douglas C. Kellogg, and Robert G. Kingsley

117

chapter 6: The Early Woodland Component at 33RO583, a Multicomponent Site in Jefferson Township, Ross County, Ohio: Preliminary Results Anne B. Lee, Andrew R. Sewell, M. Brooke Thompson, Steve Martin, and Tommy Y. Ng

143

chapter 7: Early Woodland Ceremonialism in Context: Results of LCALS Research at the Munson Springs Site (33LI251) Paul J. Pacheco and Jarrod Burks

159

chapter 8: Upland Settlement in the Adena Heartland: Preliminary Evidence and Interpretations from Two Early Woodland Nonmortuary Habitations in Perry County, Ohio John F. Schweikart

183

v

Otto FM:Layout 1

9/21/08

2:24 PM

Page vi

chapter 9: Political Complexity and Mound Construction among the Early and Late Adena of the Hocking Valley, Ohio Elliot M. Abrams and Mary F. Le Rouge

214

chapter 10: Observations on the Early Woodland Cultural Landscape in the Central Muskingum Valley of Eastern Ohio Jeff Carskadden

232

chapter 11: Smoking Pipes and Early Woodland Mortuary Ritual: Tubular Pipes in Relation to Adena Sean M. Rafferty

271

chapter 12: The Dominion Land Company Site: An Early Adena Mortuary Manifestation in Franklin County, Ohio Ann C. Cramer

284

chapter 13: The Adena Complex: Identity and Context in East-Central Indiana Beth K. McCord and Donald R. Cochran

References

vi

Contents

334

361

Otto FM:Layout 1

9/21/08

2:24 PM

Page vii

P R E FA C E

Archaeology in the Ohio area, especially relating to societies who were responsible for constructing earthen mounds, holds a significant place in the history of U.S. archaeology. Archaeological inquiry began with the study of those sacred earthen mounds initially built by indigenous societies some three thousand years ago. The American myth of an unknown, imaginary race of nonindigenous peoples responsible for building these dispersed conical mounds, particularly abundant in the middle Ohio Valley, spread with western expansion. It was not until Cyrus Thomas’s excavations of many of these mounds in the late nineteenth century that this myth was eventually dispelled, with credit rightfully given to indigenous peoples. A central question posed by the emergent profession of archaeology then became, who were these people? Excavation of the large conical Adena mound in Chillicothe, Ohio, by William Mills in 1901 yielded a richness of funerary artifacts that initiated Adena studies. Continued excavations of earthen conical mounds over the next five decades added to the inventory of material traits that defined the Adena culture, presumed at the time to represent a single, unified tribe analogous to contemporary Native American tribes. These trait lists continued to grow as archaeology refined its chronological placement of cultures, establishing the “Early Woodland” culture as distinct from that of the “Late Archaic” period, which roughly subsumed the transition from nomadic to sedentary archaeological cultures. By the 1960s, archaeologists accepted the Adena as a specific kind of Early Woodland society: a shamanistic, kin-based culture centered along the middle Ohio Valley but with influence— evidenced by their iconic burial mounds—elsewhere in the eastern portion of the continent. Since the 1960s, Adena research established the construction of burial mounds ca. 500 bc, placing this funerary practice more firmly in the chronology of the Early Woodland period. Posts located under the perimeters of these mounds were recognized not as the remnants of domestic structures, but rather as relating to ceremony. The idea that mounds identified the loci of hamlets was thus reconsidered. Small sites adjacent to mounds similarly were viewed not as residences but as mortuary camps—places where people temporarily stayed while preparing for and participating in the funeral of the deceased to be interred in that mound. Advances were also made in modeling horticultural origins. By the 1950s, research had established the species tended by the Early Woodland peoples, as archaeologists and botanists collaborated to identify chenopods, maygrass, marsh vii

Otto FM:Layout 1

9/21/08

2:24 PM

Page viii

elder, and a variety of other local species, rather than maize from Mexico, as a significant component of the economy of these increasingly sedentary groups. This transition from a hunting-gathering economy to one supplemented with gardened or managed plant species remains a key area of research today. Developments in the field of anthropology ensued. The notion of a single Adena culture was discarded as a classificatory artifact too easily created by virtue of the presence of mounds. In fact, the term “Adena” itself is seen as obstructive to analytic growth and is used now solely as a heuristic term. Anthropologists also achieved important insights in conceptualizing the historical process from hunter-gatherer to more sedentary societies. By analytically separating the creation of ceramics, horticulture, and sedentism, anthropologists emphasized the variability among these Early Woodland societies, as distinct from their Late Archaic ancestors. Further, agency rather than some inevitable, organic emergence of change is now the framework for a more humanistic scale of investigation. However, a mature understanding of the diversity within Early Woodland societies as opposed to that of the Late Archaic period, as well as of the relationships among these societies, has been slow in coming. The integration of data with social models from controlled historic comparisons—analogs—has languished. Domestic site excavation and settlement surveys remain limited. The lament of all archaeologists—the need for more refined regional chronologies —typifies Early Woodland research, precluding advancements in historical ecology. This volume seeks to rectify, to some degree, these limitations. The following collection of essays explores some of the significant cultural transitions experienced by Archaic and Early Woodland societies in the Ohio area between approximately 4000 bc and ad 100. The essays document evidence for a growing complexity of the social, political, and ceremonial lives of these early people. They combine presentations given at two symposia sponsored by the Ohio Archaeological Council, “Hunter-Gatherers to Horticulturists: The Archaic Prehistory of the Ohio Area” in 1995 and “The Early Woodland and Adena Prehistory of the Ohio Area” in 1997, along with several essays submitted specifically for this publication. The data are based on traditional archaeological field research, on analyses of museum collections, and on the results of a number of cultural resource management (CRM) projects. Their geographic coverage extends to all parts of Ohio and to adjoining states. While some focus on individual sites, others broaden their view to larger regions such as the western Lake Erie Basin and the Muskingum and Hocking river valleys. Besides providing new information on these topics, many of the chapters include previously unpublished radiocarbon dates that enhance our understanding of the chronology of Archaic and Early Woodland cultures.

viii

Preface

Otto FM:Layout 1

9/21/08

2:24 PM

Page ix

Kent D. Vickery’s contribution constitutes a major summary of the author’s significant investigations documenting Archaic settlement patterns and chronology. The essay focuses on his work in southwestern Ohio, particularly the Late Archaic Maple Creek phase, defining not only artifact types but also settlement patterns and subsistence. Craig S. Keener, Kolleen Butterworth, and Crystal L. Reustle combine large quantities of data from a number of small surface sites located in north-central Ohio as part of a lengthy US 30 highway survey project. Looking at these small sites from a regional viewpoint suggests varying Archaic settlement patterns in two different physiographic areas. Matthew P. Purtill’s essay, contributed specifically for this publication, chronicles a recent intensive CRM investigation of the Davisson Farm site, a major Archaic occupation on the Ohio River, and provides insights into Late/ Terminal Archaic settlement patterns and relationships with contemporary groups to the northeast. David M. Stothers and Timothy J. Abel discuss some of the ambiguities of the cultural concept of “Early Woodland” and provide an interpretation of Early Woodland mortuary programs, settlement-subsistence patterns, and exchange systems in northwestern Ohio and adjoining sections of Michigan and Ontario. The bases for their work are investigations sponsored through the Western Lake Erie Archaeological Research Program. James A. Robertson, Douglas C. Kellogg, and Robert G. Kingsley document three sites in the Chartiers Creek drainage south of Pittsburgh, Pennsylvania, which provide an uncommonly detailed view of settlement and subsistence during the transition from Late Archaic through Early and Middle Woodland in the Upper Ohio valley. The chapter by Anne B. Lee, Andrew R. Sewell, M. Brooke Thompson, Steve Martin, and Tommy Y. Ng on the Early Woodland component of 33RO583 in Ross County, Ohio, was specifically solicited for this publication. A CRM project to investigate a highway right of way in 2001 produced a paired-post structure of a type normally found underlying Adena/Early Woodland burial mounds, but in this case without a mound covering. This rare discovery suggests the site had special, likely ceremonial, functions. A public archaeology program in Licking County, Ohio, generated the investigation of the Munson Springs site as described by Paul J. Pacheco and Jarrod Burks. Although the site is multicomponent, the culminating event was the construction of an Early Woodland mound. In addition to describing the stratigraphy and cultural remains recovered from the mound, Pacheco and Burks expand their presentation to discuss the regional context of the site, focusing on the Late Archaic through Early Woodland period.

Preface

ix

Otto FM:Layout 1

9/21/08

2:24 PM

Page x

John F. Schweikart investigated two central Ohio examples of the normally elusive Early Woodland habitation sites. The two sites, located during an archaeological investigation of a gas line in Perry County, produced lithics, ceramics, and a posthole pattern suggesting a temporary shelter. These data help expand our understanding of Early Woodland people beyond their mounds and ceremonial sites. Elliot M. Abrams and Mary F. Le Rouge report on a multiyear research project examining the Early Woodland people of the Hocking River valley. In particular, they chart the contrasts between Early Adena and Late Adena mounds, suggesting that increases in mound size through time reflect increasing levels of political organization among these Early Woodland people. Jeff Carskadden presents a survey of Early and Late Adena sites, including habitations, mounds, and earthworks, in the Muskingum River Valley area of eastern Ohio. He provides a chronology and discusses the relationships between the various site types and the particular physiographic areas in that region. In addition to locational analysis, he also delves into variations in mortuary practices and identification of individual social groups. Beth K. McCord and Donald R. Cochran look at Adena manifestations in east-central Indiana. Radiocarbon dates from many of these sites are contemporary with Hopewell; indeed a number of the sites also produced diagnostic Hopewell artifacts. McCord and Cochran conclude that Adena and Hopewell are components of the same ceremonial system. Sean M. Rafferty’s study of tubular smoking pipes focuses on their geographic distribution and their contexts within particular Adena sites. These contexts, plus data derived from ethnohistoric accounts, highlight the likely ritualistic functions that these pipes served. Ann C. Cramer analyzes the artifacts and field notes of the salvage excavation of the Dominion Land Company site in Columbus in the 1950s. In particular, she defines the ceramic variety, Dominion Thick, as a dominant ceramic type associated with Early Adena in central Ohio. Martha P. Otto and Elliot M. Abrams

ACKN OWLEDGMENTS I wish to thank all of the authors for their contributions to this volume and their patience during the time it took to publish it. I particularly thank Brian Redmond for shepherding the papers through the printing process. His contribution to this volume is invaluable. Martha P. Otto

x

Preface

Otto FM:Layout 1

9/21/08

2:24 PM

Page xi

Otto FM:Layout 1

9/21/08

2:24 PM

Page xii

Otto CH1:Layout 1

1

9/21/08

2:17 PM

Page 1

Archaic Manifestations in Southwestern Ohio and Vicinity Kent D. Vickery

The southwestern corner of Ohio and adjacent areas are located in the western mesophytic forest zone of the Interior Low Plateau physiographic province in the central Ohio Valley. Within this region, various research endeavors that include distribution studies, survey data, and excavations offer glimpses into its Archaic occupational history. This chapter provides a brief overview of such history, which spans—in Griffin’s (1967) scheme—some 7,000 years from Early Archaic (ca. 8000–6000 bc) through Middle Archaic (ca. 6000–4000 bc) to Late Archaic (ca. 4000–1000 bc). Adoption of this temporal framework over those of Willey (1966) or Stoltman (1978), for example, is arbitrary. Subsequent to Griffin’s period scheme, deeply stratified sites have been excavated in various parts of the Midwest and Midsouth. Some of the betterknown sites include St. Albans in West Virginia (Broyles 1966; 1971; Brashler, Kite, and Freidlin 1994); Longworth-Gick (Collins 1979), Deep Shelter (Dorwin et al. 1970), and Morrisroe in Kentucky (Nance 1986); Swan’s Landing in Indiana (E. Smith 1995); and various sites in Tennessee reported by Chapman (1976), all of which have chronometrically dated Archaic sequences that often form the basis for internal chronologies for Early, Middle, and Late Archaic culture periods (e.g., Chapman 1985; Stothers and Abel 1991; DeRegnaucourt 1995; DeRegnaucourt and Georgiady 1998, 22). 1

Otto CH1:Layout 1

9/21/08

2:17 PM

Page 2

However, it seems doubtful that any one particular chronological scheme, and the often projectile point–based divisions and subdivisions thereof, will serve on a pan-Midwest scale. There are at least two reasons for this: (1) the specific projectile point types often used to define and recognize phases (or similar archaeological units) in one geographic area are often sparsely represented— if at all—in portions of the Midwest and Midsouth away from their “homeland” localities or regions, and (2) there is a lack of consistency in the use of the Early, Middle, and Late Archaic labels throughout this stretch of territory whereby culture periods (sensu Griffin) are confused with lifeways. As noted by Vickery (1999), much of this confusion derives from cultural adaptations that followed in the wake of the warm, dry Xerothermic/ Hypsithermal Interval as it slowly but inexorably made its way across the Midwest from west to east. Its peak, ca. 5000 bc, in portions of southern Illinois and adjacent Missouri at the southwestern corner of the Midwest (Brown and Vierra 1983) apparently prompted at least some western Midwest and northern Midsouth archaeologists to recognize the Middle Archaic as a new, emergent culture period beginning—in some schemes—around 4000 bc. Thus, the beginning of the Middle Archaic in the western Midwest coincides with its termination in the Midwest heartland, farther east—at least for those who strictly follow Griffin’s culture period scheme. The later appearance of the Xerothermic in the eastern Midwest may have contributed to the “demise” of the Late Archaic and paved the way for the transition to the Early Woodland. Logically, one may conclude from this “discrepancy” that cultural responses in the form of changed lifeways to the same climatic episode indeed occurred, but at different times and in geographically separated areas. Thus, it is inappropriate to apply the label Middle Archaic, for example, across the entire Midwest while relegating it everywhere to the same temporal span within the Archaic cultural continuum.

EA R LY ARC HAIC In a distributional study of Early Archaic projectile points similar to that of Luchterhand (1970) in the lower Illinois Valley, Litfin (1993) identified 883 specimens of demonstrably Early Archaic origin among more than 26,000 points curated at the Cincinnati Museum Center that had at least county provenience, then plotted their distribution in a 63-county area of southwestern Ohio,

2

K. D. Vickery

Otto CH1:Layout 1

9/21/08

2:17 PM

Page 3

southeastern Indiana, and northern Kentucky to yield the following sequential or partially overlapping phases: Charleston phase—(8000–7530 bc), the initial Early Archaic, accommodating Charleston Corner-Notched points; Thebes phase—(7530–7410 bc), the early Early Archaic, accommodating Thebes, Dovetail, and Lost Lake points; Kirk phase—(7410–6900 bc), the middle Early Archaic, accommodating Kirk, Palmer Corner-Notched, and Fractured Base points; Pine Tree phase—(7200–6870 bc), the late Early Archaic, accommodating Small Kirk, Pine Tree, and Pine Tree Corner-Notched points, and St. Albans phase—(6870–6000 bc), the terminal Early Archaic, accommodating St. Albans, Le Croy Bifurcated Base, Kanawha Stemmed, and Lake Erie Bifurcated Stem points.

Except for the monotypic Charleston phase, each of Litfin’s phases is represented by more than one projectile point–defined complex of shorter duration presumed to correspond with distinct, but likely related, bands or groups of bands. Although not universally held to be a valid measure of catchment extent, Litfin used the flint raw materials from which these points had been made as a basis for inferring the sizes and shapes of the exploitative territories for each of his five Early Archaic phases. Litfin suspects that catchment size decreased throughout this 2,000-year span of time as food resources became more abundant and more varied with the post-Pleistocene stabilization of interior waterways and the gradual encroachment of deciduous forest elements from refugia in the Midsouth and Southeast, allowing Early Archaic hunter-gatherer bands to progressively fill new ecological niches. Evidence for this view is an apparent reduction in the distances traveled to obtain flint for at least projectile point manufacture from presumed base camps represented by concentrations of points of a single type. Some ephemeral occupations on the part of bands not defined as complexes, but represented by at least 10 points, were inferred to have had encampments outside Litfin’s study area (for example, Kessell, Lost Lake, Fractured Base, Schoonover [MacCorkle]). Relatively widespread use of the area began in early Early Archaic with the establishment of base camps near the Ohio River, where deciduous forest elements had become established before expanding into the surrounding uplands. Thebes and Dovetail points made from Crescent chert (northeastern

Archaic Manifestations, Southwestern Ohio and Vicinity

3

Otto CH1:Layout 1

9/21/08

2:17 PM

Page 4

Missouri/southeastern Illinois) and one Dovetail of Kaolin chert (southern Illinois) in the Early Archaic levels of site 12SW89 in Mexico Bottom of Switzerland County, Indiana (Houston, McHugh, and Michael 1984), suggest a western origin, perhaps representing a “colonization” of the region from the west along the Ohio River. This component of 12SW89 also yielded one Kirk Corner-Notched (small variety) point of southern Illinois–derived Mill Creek chert, associated with late Early Archaic. During the terminal Early Archaic, characterized by bifurcated points, territories were much reduced in size. Individuals or small task-specific groups from base camps exploited predominately locally available chert. A slight expansion of catchments evident throughout the period perhaps reflects slight population increases. Excavated Early Archaic components are present at the Manning site (33CT476) in Clermont County, Ohio (Lepper 1994); the Leonard Haag site (12D29) in Dearborn County, Indiana (Tomak, Tomak, and Reidhead 1973; Reidhead and Limp 1974); and site 12SW89 in Switzerland County, Indiana. The Haag site is on a terrace adjacent to the flood plain of the Great Miami River, near its confluence with the Ohio River; Manning and 12SW89 are on the valley floor of the Ohio River. At Manning, a buried 7500–7000 bc occupational zone with Kirk Corner-Notched points overlay a Paleoindian component. A Kirk Corner-Notched occupation at the Haag site underlay a midden deposit; some of the points were associated with a living floor at least 25.58 m2 in extent and about 1 m below ground surface. The excavators relate unifacial knives, an expanded-base drill, and endscrapers to similar types found in the Early Archaic zones at the St. Albans site. The Early Archaic component at 12SW89 is represented by Amos Corner-Notched, small and large Kirk CornerNotched, Kirk/Palmer, Lake Erie Bifurcated-Stem, Le Croy Bifurcated-Base, and St. Albans points. The surface-exposed Ferris site (33CT31) on the valley floor of the Ohio River in Clermont County (Theler and Dalbey 1974) is virtually a singlecomponent Early Archaic base camp. All but seven of 57 Early Archaic points are referable to Palmer Corner-Notched, while two are Kirk Corner-Notched and three points are either Kirk or Palmer. The other Early Archaic points are one Thebes and one Hardaway Dalton, the latter of which is transitional between Paleoindian and Early Archaic (Vickery and Litfin 1994). Also present were three steep-edged, unifacial endscrapers, with and without graver spurs; large, percussion-detached blades, and five preforms—four of which were of Kentucky-derived Boyle chert that were plowed up together—plus one of Kanawha Black chert from West Virginia. Localized concentrations of spherical 4

K. D. Vickery

Otto CH1:Layout 1

9/21/08

2:17 PM

Page 5

hammers in one area and endscrapers in two others probably represent loci of chert working (hammers) and hide processing (endscrapers) or craft production (or both). Limited testing revealed a sub–plow zone cultural deposit to at least 54.86 cm below ground surface that was unaccompanied by organic staining or ecofacts. A Palmer Corner-Notched point was present at a depth of 51.82 cm. The Dallas Burton site (33CT58) is a small (ca. 0.2 ha) upland encampment that has yielded about 30 Thebes points, but no Kirk/Palmer points. Subsistence and settlement data are limited for the Early Archaic in southwestern Ohio. Wood charcoal from a post-Dalton, pre-Kirk feature at Manning was identified as juniper, spruce, walnut, and hickory (Lepper 1994, 146), indicating a mixed coniferous-deciduous forest composition. The presence of a pitted (nutting) stone this early apparently represents the technology to process nuts; four such implements were present in the slightly later Kirk Corner-Notched-bearing zone. As might be expected, the largest Early Archaic sites that likely were base camps tend to be near major waterways, with smaller (special purpose/extractive?) encampments or bivouacs scattered throughout the uplands. Ferris is unusual in having been exposed on a major river valley floor. In virtually every case where relatively deep testing has occurred in surface-exposed sites of later time periods along the Ohio River, Early Archaic occupation has been detected. Therefore, it is likely that Early Archaic sites in such settlement situations are underrepresented and this also may pertain to those on the valley floors of tributary waterways. Lepper (1994, 148–49) recognizes three Archaic settlement types: Type I settlements are “large sites with numerous bifaces”; Type II settlements are “small sites with few bifaces,” and Type III settlements are “ephemeral occupations with few artifacts.” He attributes Ferris and occupation 3 (the Kirk zone) at Manning to Type I, and occupation 2 at Manning (Paleoindian–Early Archaic) to Type II.

MI DDLE ARCHAIC The Middle Archaic period is as elusive in southwestern Ohio as it apparently is elsewhere in Ohio and vicinity. In southeastern Indiana, site 12SW89 is represented by a small Middle Archaic component stratigraphically between the Early Archaic and Late Archaic occupations. This component yielded Kirk Serrated, Kirk Stemmed, Morrow Mountain, Stanly, and White Springs projectile Archaic Manifestations, Southwestern Ohio and Vicinity

5

Otto CH1:Layout 1

9/21/08

2:17 PM

Page 6

points, but is unrepresented by faunal or botanical remains. In Clermont County, the Logan Locality 3 site (33CT72) is one of at least four bluff-base settlements with a Middle Archaic component. Because such sites would have been subject to burial by slope wash, more encampments with this distinctive settlement pattern may exist. However, in at least this locality, the distribution of Middle Archaic sites generally seems to have focused on uplands rather than on the Ohio River valley floor. If Chapman’s (1985) and Nance’s (1986) assessment of the temporal placement of Middle Archaic points is valid for southwestern Ohio and its environs, the first millennium of this culture period (6000–5000 bc) should be composed of at least three sequential phases: (1) Stanly, Kirk Serrated, and Kirk Stemmed, which may predate the beginning of Middle Archaic by 200 years or so and continue to at least 5800 bc; (2) Cypress Creek, around 5800–5500 bc; and (3) Eva and Morrow Mountain, around 5500–5000 bc. These perhaps would be followed by Guilford and Sykes (White Springs/Damron?) from 5000 to 4000 bc (Chapman 1976, table 7.1), with Big Sandy apparently transitional from Middle to Late Archaic. Guilford and Sykes/White Springs/Damron points are rare in southwestern Ohio and vicinity, suggesting interpretations such as: (1) a hiatus in occupation here from 5000 to at least 4000 bc; (2) the continuation or later reappearance of earlier Middle Archaic point styles throughout that millennium; (3) buried sites; and (4) the manufacture of point types currently unrecognized as belonging to the late Middle Archaic period.

L ATE ARC HAIC Late Archaic people intensively occupied the southwestern corner of Ohio, as evidenced by abundant sites and site components throughout the area. Based on extent of surface debris and artifact scatter, as well as sub–plow zone midden accumulations containing large quantities of material remains, they include some of the largest and most information-rich archaeological resources in this region.

Transitional Archaic An ephemeral Transitional Archaic (sensu Witthoft 1953) presence in southwestern Ohio likely represents the adoption of projectile point styles that

6

K. D. Vickery

Otto CH1:Layout 1

9/21/08

2:17 PM

Page 7

include Ashtabula-like, Lehigh, Perkiomen, Snook Kill, Susquehanna Broadspear, and Orient Fishtail. Sherds of imported soapstone vessels, another hallmark of the Transitional Archaic, were present at the Wilson Locality 1 (33CT48) and Maple Creek (33CW52) sites in Clermont County and at the Driving Range site (33HA586) in Hamilton County, Ohio (Kreinbrink 1996). The upland multicomponent East Fork site (33CT36) yielded an Ashtabula point, two Snook Kill points, and a Snook Kill hafted chisel.

Laurentian Archaic If a late Middle Archaic occupational hiatus indeed is represented in southwestern Ohio and vicinity, it also may have extended through the early Late Archaic, from around 4000 to 3000 or 2750 bc. Perhaps part of this gap is filled by a Laurentian Archaic (Ritchie 1969) presence in this area. Laurentian or Laurentian-related assemblages in Ohio are acknowledged in all synthetic treatments of the Ohio Archaic (Blank 1970; Britt 1967; Geistweit 1970; Prufer and Long 1986; Prufer, Pedde, and Meindl 2001). While artifacts referable to Brewerton and to other Laurentian phases co-occur in this area, projectile points and hafted scrapers of the former are present in greater quantities than other Laurentian types. The earliest Laurentian manifestations are dated ca. 3200–3000 bc in their New York “heartland,” and their appearance in the central Ohio Valley is presumed to have been coeval or slightly later. Serving to identify this rather distinctive cultural complex are projectile points and hafted scrapers of at least the Normanskill, Genesee, Otter Creek, and Vosburg types, in addition to Brewerton Side-Notched (abundant), Brewerton Corner-Notched (somewhat less abundant), Brewerton Eared-Notched (rather uncommon), and Brewerton Eared-Triangle (rare). Lamoka points are a minority type; nevertheless, they—along with Motley and side-notched points that may be related to Big Sandy or Big Sandy variants with a probable northern Midsouth origin— co-occur with these Laurentian types in local Late Archaic assemblages. Such projectile points and hafted scrapers probably were incorporated into the artifact inventories of resident populations as early as the middle Late Archaic (2750–1750 bc) or perhaps even earlier. They represent minority types in local assemblages of that millennium and tend to increase in representation in later Archaic sites as indicated by a greater proportional abundance and a wider variety of styles (types).

Archaic Manifestations, Southwestern Ohio and Vicinity

7

Otto CH1:Layout 1

9/21/08

2:17 PM

Page 8

“Pure” Laurentian assemblages in southwestern Ohio and vicinity have eluded recognition. Perhaps the best excavated context is 12SW89, where a Brewerton component occurring from 3.34 to 3.50 m below ground surface is represented by tools or weapons of Otter Creek, all the Brewerton types except Eared Triangle, and by specimens resembling Lamoka and Normanskill types. In Clermont County, the strongest representation of Laurentian artifacts—as surface collected assemblages only—are from the following five sites: (1) Retzler (33CT37) on the T-1 of the East Fork, Little Miami River, with Brewerton Corner-Notched, Brewerton Side-Notched, Genesee, and Vosburg points, along with four atlatl weights; (2) Jacobs (33CT40) on the same terrace, with Brewerton Side-Notched points and one Vosburg drill; (3) Renscke no. 1 (33CT105), with projectile points or tools of the Brewerton Side-Notched, Normanskill, and Lamoka types, as well as two three-quarter-grooved axes and a bell pestle; (4) site 33CT125 in an upland stream valley, which yielded one Brewerton SideNotched point, one Brewerton Side-Notched scraper, one Normanskill, and two Lamoka points; and (5) Jarman Locality 2 (33CT78), also in the uplands, with Brewerton Side-Notched, Corner-Notched, and Eared Triangle types. Gilbert/Commonwealth Inc. (1986, 141–45) reported one Otter Creek and three Brewerton series points from four sites in Clermont County: 33CT428, 33CT444, 33CT449, and 33CT451. Apart from Laurentian and Transitional Archaic, the only other reasonably well defined Late Archaic cultural complexes in southwestern Ohio and vicinity are the Central Ohio Valley Archaic and Maple Creek phases. In both—but particularly in the former—McWhinney Heavy-Stemmed projectile points (Geistweit 1970, 149–50; Heilman 1976, 13–15) are abundant in proportion to points of other types. The origin of the McWhinney point type is unknown, but is presumed to have been local. Its distribution is apparently restricted to the central Ohio Valley, where it is known to occur in northern Kentucky and at least the southwestern and southeastern portions of Ohio and Indiana. Generalizing from an examination of points from the Indian Knoll and Carlston Annis sites in western Kentucky, McWhinney points do not occur in Indian Knoll phase sites, although some thick-stemmed specimens may have inspired their (later?) manufacture. Many points identified as McWhinney Heavy-Stemmed in the literature are not of this type. They do not co-occur with known Middle Archaic points, but are commonly associated with Laurentian types; hence, it is likely that their inception dates between ca. 3000 bc and 2750 bc.

8

K. D. Vickery

Otto CH1:Layout 1

9/21/08

2:17 PM

Page 9

Central Ohio Valley Archaic The Central Ohio Valley Archaic is so named because of its similarity to what Winters (1969) called the “lower Ohio Valley Archaic” (or Wabash Valley Archaic) in east-central Illinois, having traits in common with Indian Knoll phase sites in western Kentucky. In the southwestern Ohio area, it is dated to the millennium from 2750 to 1750 bc (two-sigma calibration to 3632–1945 bc) (Reimer et al. 2004) on the basis of a consistent series of radiocarbon dates from excavated sites (see table 1.1). The temporal placement is thus middle Late Archaic, a chronological position relative to the succeeding Maple Creek phase that is duplicated for the Illinois manifestations at the stratified Koster site in the lower Illinois Valley (Houart 1971). Recognition of the Central Ohio Valley Archaic is based on excavations at the following sites in Hamilton and Clermont counties, Ohio (see fig. 1.1): DuPont Village (33HA11), a ca. 4.61-ha base camp on the low-lying T-1 of the Ohio River, near its confluence with the Great Miami River; Dravo Gravel (33HA377), a ca. 2.29-ha settlement at the edge of a high terrace overlooking the Great Miami River; Bullskin Creek (33CT29), a 1.01–1.21-ha base camp on a high terrace overlooking Bullskin Creek, near its confluence with the Ohio River, and Logan (33CT30), a 0.32-ha encampment adjacent to Maple Creek on the valley floor of the Ohio River, where the assemblage seemed to be transitional between the Central Ohio Valley Archaic and the Maple Creek phase.

Among the excavated Central Ohio Valley Archaic sites, McWhinney Heavy-Stemmed dominates the projectile point and hafted scraper types, with Brewerton series, other Laurentian, and a few other weapons and tools (including Transitional Archaic and Motley) present as minority types. Other traits include atlatl parts (ground-stone and antler weights; antler hooks), hardstone bell and limestone or dolomite roller pestles, grooved axes, and ululike bifacial flint knives with one straight and one excurvate edge. These “Maple Creek knives” persisted through the succeeding Maple Creek phase, as did McWhinney points and scrapers. An industry in glacier-transported tillite featuring predominantly unifacial scraper planes is diagnostic and occurs in (or derives from) only the Great Miami Valley and its immediate environs. They were created by the direct percussion removal of a few flakes from a

Archaic Manifestations, Southwestern Ohio and Vicinity

9

Otto CH1:Layout 1

9/21/08

2:17 PM

Page 10

Table 1.1. Radiocarbon dates for Late Archaic manifestations in southwestern Ohio Site

Cultural period

Archaeological unit

Radiocarbon years BP

Calibrated date range*

late Late Archaic

Maple Creek phase

3260 bp +/-330

2458–800 bc

middle Late Archaic

?

4065 bp +/-150

3010–2144 bc

Archaic to Maple Creek phase (?)

4115 bp +/-455

3771–1505 bc

middle Late Archaic

Central Ohio Valley Archaic

4100 bp +/-65 4125 bp +/-65 4435 bp +/-70 4485 bp +/-75

2876–2491 bc 2885–2496 bc 3339–2918 bc 3365–2928 bc

middle Late Archaic

Central Ohio Valley Archaic

2340 bp +/-55†

745–209 bc

middle Late Archaic

Central Ohio Valley Archaic

4470 bp +/-75 4550 bp +/-355

3357–2928 bc 4051–2213 bc

Maple Creek I, II (33CT52) Maple Creek III (33CT52) Logan (33CT30)

mid-late transitional from Late Archaic (?) Central Ohio Valley

DuPont (33HA11)

Dravo Gravel (33HA377) Bullskin Creek (33CT29)

*The 2-sigma range, based on Reimer et al. 2004. †

Date is unacceptably late by ca. 2000 years.

single, flat striking platform to create a steep functional edge. Use wear on this edge and experiments using replicated specimens suggest that they were used by pushing forward with the flat face down at about a 45-degree angle. Although McWhinney points and scrapers persisted into the succeeding Maple Creek phase, their distribution is restricted to portions of the central Ohio Valley and they are absent in Green River Archaic assemblages in western Kentucky. When combined with the diagnostic tillite industry and other traits

10

K. D. Vickery

Otto CH1:Layout 1

9/21/08

2:17 PM

Page 11

Fig. 1.1. Location of excavated Late Archaic sites mentioned in text.

of the southwestern Ohio assemblages, the complex is sufficiently distinctive to distinguish the Central Ohio Valley Archaic from the Indian Knoll phase and to justify recognizing it as a separate phase (sensu Willey and Phillips 1958). The stratigraphy of all four sites was similar. Although encountered only in certain areas of each site below plow zone, midden deposits no thicker than 76.2 cm were consistently exposed. They were underlain by a cultural deposit lacking the organic staining and contents characteristic of overlying midden. Feature types common to all four include garbage/trash pits, roasting pits or earth ovens (or both), graves, and postmolds. The predominant mode of burial was flexed, with nearly half accompanied by grave goods of some kind; occasionally, burials were sprinkled or covered with red ocher. DuPont

A total of 30 burials, 73 features, and 27 postmolds were excavated, most of them occurring in an extensive midden accumulation no thicker than 45.72 cm. With the exception of one sitting burial, all burials had been interred in flexed

Archaic Manifestations, Southwestern Ohio and Vicinity

11

Otto CH1:Layout 1

9/21/08

2:17 PM

Page 12

or semiflexed positions. Two individuals showed evidence of traumatic death. One had two projectile points embedded. The other, a male about 25 years old, had four points embedded, all of which had entered from behind. He was accompanied by a shell-and-cannel-coal-bead necklace, an antler atlatl weight, a chipped-stone drill, and a box turtle carapace container with a lump of red ocher in it. Red ocher had been sprinkled on the face. Another adult male had red and yellow ocher sprinkled over most of the remains and had been interred with a bird bone flute. Also with him were several bone beads that had probably trimmed his clothing. Limestone or dolomite roller pestles were associated with two burials; another was accompanied by four scraper planes, an antler flaker, and two pieces of worked antler. Other artifacts in burial association included a marble atlatl weight, a perforated bone shuttle, a perforated raccoon canine, a three-quarter-grooved axe, a turtle carapace container, projectile points, a small mammal, and unmodified deer toe bones. Among the various other artifacts at DuPont were celts, bone fishhooks, notched sinkers, groundstone bell and conical pestles, and pendants of stone, cannel coal, and drilled mammal teeth. Unique to DuPont is a distinctive type of earth oven referred to as a Dakota oven. There were two varieties—one is a deep basin while the other is nearly spherical with a truncated, inverted cone above that constricts the middle, where the two join. Characteristic of both types is a subterranean tunnel projecting outward near the base that joins one or more other ovens. The general form of these distinctive compound features suggests a suctionlike ventilator principle to allow the passage of air from one oven to another through an interconnecting shaft for more efficient combustion. Quantities of mussel shell suggest that the steaming of mussels occurred within them, but it is likely that they served other cooking tasks as well. Domesticated dog remains encased in fire-hardened clay were found in the intact portion of the base of one Dakota oven, suggesting that the animal had been baked there. Other earth ovens were large, often basin-shaped pits filled with burned limestone and cobble fragments. One had seven or eight postmolds, arranged in a semicircle around the edge, that probably represent a windscreen. Also unique to DuPont were platform burials. The platform of one such burial adjoined an earth oven, into which one leg of the individual had flopped, thereby establishing the fact that the creation of the platform postdated the use of the oven. Because of their typically huge size, it is inferred that each oven was a locus of communal gathering, where at least one family

12

K. D. Vickery

Otto CH1:Layout 1

9/21/08

2:17 PM

Page 13

engaged in recreation, crafting, and other activities. Observing that the burial had been covered and the adjoining oven filled with refuse and debris at the same time, it may be speculated that the death of the head of the social unit using the facility prompted the “disposal” of both as a unified act of respect. The available radiocarbon dates are clustered in two groups representing either two discontinuous occupations or, more likely, the earliest and latest occupations of the site locus. The DuPont site is discussed by Starr (1960, 99–100), who includes photographs of artifacts from it. Reports by Dalbey (1976, 1977a, 1977b), Featherstone (1977), and Theler (1977) pertain to the salvage excavations and analyses. Dravo Gravel

Salvage excavations at Dravo were restricted to a small portion of the site undisturbed by gravel extraction operations (Vickery 1978). Bulldozer stripping exposed 100 features, and 25 of them were totally or partially excavated. In three distinct areas—unique to Dravo among the four excavated sites— features were arranged in semicircular fashion around an open area. Each of these arc patterns was comprised of at least six or seven features, including one or more cooking and refuse pits. Single families may have lived within these arcs or in close proximity to them, using all the pits in each for their daily needs. Specialized activities apparently occurred nearby at several discontinuous loci throughout the site. While it is likely that at least some of the refuse pits originally had served as storage facilities, one storage pit in which the organic contents had since decayed had been capped with a layer of clay from which a wedge-shaped section had been removed, presumably in order to retrieve some of its contents. Another pit had been lined with clay. The occupants had deposited sterile soil over refuse and debris near the tops of several other pits—presumably in an effort to lessen or eliminate the odor that must have emanated from them. Here and elsewhere, Late Archaic people commonly seem to have observed “sanitary” conditions during the earliest site occupations by depositing refuse and debris in pits. Later, they abandoned the effort and disposed of such waste material on the ground, thereby creating midden deposits that increased in size and depth with growing numbers of people and greater occupational intensity. Excavated burials included two dogs and ten humans, the former represented only at Dravo among the four excavated sites. At least three of the humans had been buried in a tightly flexed position, three were semiflexed, and

Archaic Manifestations, Southwestern Ohio and Vicinity

13

Otto CH1:Layout 1

9/21/08

2:17 PM

Page 14

one child aged 6–7 years was buried sitting. Three individuals were present in one grave, although one—a female aged 35–40 years—had been intruded into it at a later date. One male aged 25–30 years had been interred with three tillite scraper planes and a flake core of the same material. An adult female had been buried with one infant and one child in her arms. Associated with the infant was a drilled human molar, while a carnivore canine had been placed beneath the child. Two unmodified deer toe bones also accompanied this cluster of burials. The number of different kinds of tillite tools at Dravo exceeded that at DuPont; represented were one example each of chopper, scraper, graver, celt preform, flake core, full grooved axe, and notched axe. Other artifactual traits included limestone or dolomite roller pestles, grooved mauls, V-head flint drills, and bone and stone artifacts with two to six or more tallylike notches or incisions. The latter included two undrilled, polished-slate gorgets, a bone fish gorge, and one bone shuttle with a single perforation through the proximal end. Bone shuttles occur rather commonly in the excavated Late Archaic sites in southwestern Ohio. Atlatl parts (including loaf and butterfly weights of groundstone), grooved axes, celts, and both bell and conical pestles were present, along with antler pins, drilled human and mammal teeth, and notched sinkers. Bullskin Creek

Crop cover restricted excavations to a small portion of the Bullskin Creek site; however, a midden deposit no thicker than about 60.96 cm was encountered. Among 13 excavated features were three earth ovens, a roasting pit, several refuse pits, and artifact caches. The roasting pit was typical among those encountered at other excavated Late Archaic sites, being circular in plan and shallow in depth. Within the pit preheated limestone slabs had been laid flat, on which food apparently had been roasted while exposed to view. Although artifact caches were found at DuPont and Maple Creek, one of three at Bullskin Creek was unique among the four sites in apparently representing the ceremonial paraphernalia of a shaman. Occurring in a pit originating in submidden cultural deposit were a deer antler atlatl hook, a box turtle carapace cup, the articulated remains of the forearm and talons of a great horned owl, a hematite rubstone, a concretion, an unmodified amphibian (salamander?) bone, and deer antler tools that included a tine flaker, projectile point, and awl. At a lower level, but having been placed on the ground surface

14

K. D. Vickery

Otto CH1:Layout 1

9/21/08

2:17 PM

Page 15

rather than in a pit, were a piece of limonite (a probable source of yellow pigment), a large worked deer antler, an Early Archaic projectile point, and a McWhinney point whose haft element was coated with a resinous substance— likely naturally occurring— and used to secure the point in its haft. Another cache was comprised of a hardstone roller pestle preform and associated chert hammer near the edge of a roasting pit. Of three tightly flexed burials, one, which had a projectile point of probable Early Woodland affiliation embedded, had been intruded into an earth oven. That oven yielded material radiocarbon dated to 2600 bc +/-355—the earliest chronometric assay obtained to date among the five excavated Late Archaic sites (see table 1.1). None of the burials, including one in a submidden cultural deposit, had associated grave goods. Other artifactual material from Bullskin Creek included hardstone atlatl weights, notched sinkers, a bone fishhook and gorge, a bone bead and pin, a perforated fragment of turtle carapace, both perforated and unperforated bone shuttles, rodent incisor chisels, a bone needle, bell pestles, a limestone roller pestle, celts, and grooved axes. Lo gan

Midden blanketed a portion of the Logan site, varying in thickness from 24.38 cm near the center to 60.96 cm along one edge of the site, where a drop in elevation occurred. The average thickness of the midden deposit was 39.62 cm. Of eight excavated features, four or five were refuse pits, one was a roasting pit, one was probably a storage pit for plant matter, and one was questionably a cooking facility. Near an edge of the site opposite the midden deposit was a cluster of three human burials, including two probable adults in tightly flexed positions and a semiflexed female about 15 years old at the time of death placed in a grave with a limestone slab pillow under her head. Lying across the forearm of the latter was a bone shuttle, and a small mammal bone had been placed immediately below the elbow. Artifactual traits at Logan include a hemisphere and hemisphere preform, one birdstone, bone awls (with and without incised lines), notched and perforated bone shuttles (some with an X heavily incised in the articular end), grooved and notched sinkers, a grooved bone needle, rodent incisor chisels, turtle carapace containers (one fragment of which was notched in the edge), a hematite rubstone, gorgets of rectangular and quadriconcave forms, and O-head and V-head chipped-flint drills. Other artifacts relate either to the

Archaic Manifestations, Southwestern Ohio and Vicinity

15

Otto CH1:Layout 1

9/21/08

2:17 PM

Page 16

Central Ohio Valley Archaic (grooved axes, an antler atlatl hook, and a section of a drill core presumed to have been the waste material from the manufacture of a granite atlatl weight) or the Maple Creek phase (Merom Expanding Stemmed knives, manos, a cloudblower pipe preform, and perhaps flaked and pecked limestone “axes” [grubbing tools]). McWhinney points and scrapers occur in both cultural complexes, as do Brewerton Side-Notched points. Thus, the occupation at Logan is tentatively regarded as transitional between the two, with a few items (e.g., hemisphere, gorget forms) hinting at yet a later occupation—perhaps transitional between Late Archaic and Early Woodland or fully within the latter culture period. Although a surface find, the birdstone is unique among the southwestern Ohio Late Archaic site assemblages recovered through excavation. Other Comp onents

The eastern portion of the Twin Mounds Village (33HA24E) is a small camp situated on an upland ridge high above the valley floor where the Great Miami River joins the Ohio River. It is included by Starr (1960, 104–6) in his archaeological survey of Hamilton County and was excavated in 1968 by Fred W. Fischer (Fischer 1968; A. M. Lee 1972). A total of twelve pits containing cultural debris and four postmolds occurred within a midden deposit about 45.72 cm thick. One of two burials excavated was a child in a loosely flexed position, six to eight years old, accompanied by three bone fishhooks and a bone (garment?) pin in the pelvic area. Another lacked grave goods and had been interred in a tightly flexed position. Three other flexed burials were exposed but not excavated. Some of the artifacts from the Twin Mounds–East site are McWhinney points, notched sinkers, an atlatl weight preform, a bell pestle, a box turtle carapace container, a grooved axe, pebble pendants, an antler handle, a bone musical rasp, and bone needles, awls, and a perforated shuttle. The nature of the artifactual assemblage suggests a Central Ohio Valley Archaic cultural affiliation. Other sites of this cultural complex are locally abundant, occurring not only in southwestern Ohio but in southeastern Indiana and northern Kentucky as well. Perhaps among the better known in Ohio are the White Oak Creek site in Brown County (33BR19), the Raisch-Smith Village (33PR1) (Moffett 1949; R. Long 1962; Geistweit 1970), and the McWhinney Village (33PR9) (Geistweit 1970), both in Preble County.

16

K. D. Vickery

Otto CH1:Layout 1

9/21/08

2:17 PM

Page 17

Maple Creek Phase The definition of the Maple Creek phase is based on horizons 1 and 2 of the Maple Creek site (33CT52) (see fig. 1.1), the type site for the phase (Vickery 1976; see also Anslinger 1986; Boisvert 1986; Ledbetter and O’Steen 1992; Duerksen and Doershuk 1998). Based on excavations, three distinct cultural horizons were recognized at the Maple Creek site. Horizon 1 was a hard-packed living floor—unique among the excavated Late Archaic sites in southwestern Ohio—with superficial burning and pit openings under thin, but undisturbed, cultural deposit and overlying plow zone. Horizon 2 was a discontinuous layer of fire-cracked rock on a surface with occasional burning, apparently representing the debris from above-ground cooking. Such loci and the inferred cooking practice of building a fire on a formerly stabilized surface, then placing food within or on piles of heated rocks, are unique to the Maple Creek site and specifically to horizon 2 therein. The observed distribution pattern of rocks suggests that they were scattered over the surface after the cooking features had served their purpose. Horizon 2 immediately underlay the living floor, but only in a restricted area near the center of the site. Horizon 3 occurred in a buried natural levee underlying either horizon 1 or horizon 2. Sheet midden no thicker than about 76.2 cm was localized elsewhere near the center of the site but occurred as a thinner accumulation in discontinuous spots along its longer axis. The main domiciliary area for horizons 1 and 2 apparently was the living floor and its environs in the center of the site. Specialized activities including butchering, flint working, and minor craft production took place near the peripheries of this area while refuse was deposited in or along the edge of the Ohio River during at least one episode of occupation when the river apparently flowed adjacent to the site (it now flows over 2 km away from it). Five or six features originating in horizon 1 were refuse pits, while three were earth ovens. One thermal feature consisted of a layer of large limestone slabs that had to have been laboriously transported to the site from outcrops 0.47 km distant. Although limestone was common in other Late Archaic sites closer to bluffs and other outcrops, river cobbles were the only at-hand lithic materials at Maple Creek that could be used for cooking and perhaps other tasks. It seems possible that the thermal conduction properties of limestone

Archaic Manifestations, Southwestern Ohio and Vicinity

17

Otto CH1:Layout 1

9/21/08

2:17 PM

Page 18

were deemed to offer an advantage not provided by river cobbles, leading one to infer a specialized function for this feature such as mussel steaming or a noncooking function—to simply provide heat and light. Also originating in the horizon 1 living floor was an intact, undismantled earth oven that offered a unique opportunity to reconstruct the specific cooking practice involved in its use. A cylindrical hole had been excavated inside a larger, shallow depression. Preheated rocks were then stacked inside the cylinder, leaving room around the column for, presumably, food to be baked, with soil packed throughout to serve as insulation. Rocks were then heaped on top of the cylinder and a fire built around it, inside the shallow depression. The column of preheated rocks diffused heat laterally through the food and dirt insulation; the fire above diffused heat downward, and the rock pile on top sealed in the heat so generated. Ill-defined postmolds occurred in isolation and as arc patterns in two areas. One of the latter bordered a shallow pit with evidence of intense burning and likely represents the remains of a windscreen, such as that encountered at DuPont. The other was not associated with any feature and was of a size that suggested a small windbreaklike structure. A human burial in horizon 1 had been interred in a flexed position and subsequently damaged by prehistoric pit construction. As represented by horizons 1 and 2, the Maple Creek phase is believed to have existed from ca. 1750 bc to 1000 bc, temporally corresponding with the late Late Archaic. The earlier date, which pertains to the horizon 2/3 interface, was interpolated from an annual average sedimentation rate calculated from radiocarbon dates in horizons 1 and 3. Applying the annual average sedimentation rate to the vertical distance from the radiocarbon dated contexts to the present ground surface, however, suggests the possibility that the site was occupied for 400 years or so after 1000 bc. In horizons 1 and 2 were perforated, notched, and grooved bone and antler weaving shuttles, turtle carapace containers, siltstone/sandstone abraders and files, sandstone and cannel coal discs, and projectile points or hafted scrapers of all four Brewerton types, as well as Vosburg, Big Sandy, Motley, Lamoka, Normanskill, Snook Kill, Cypress Creek II, Ashtabula, and other types. Also present were sherds of a thick, coarsely tempered ware in a context suggesting contemporaneity with the main, Late Archaic, occupation of the site (there is also a sizable sherd of an Adena Plain vessel whose isolated occurrence

18

K. D. Vickery

Otto CH1:Layout 1

9/21/08

2:17 PM

Page 19

just below plow zone leads one to believe that it may be affiliated with a localized Early Woodland component that is otherwise lacking in recognizable cultural deposits). The Maple Creek phase is related to the Riverton Culture as defined by Winters (1967, 1969) in the Wabash Valley of east-central Illinois. Traits shared in common by both cultural complexes extend to “points” of the Merom Expanding Stemmed and Trimble Side-Notched types, a chipped flint microtool industry in which microperforators, microdrills, and microgravers are represented, but on flakes rather than blades; manos; a paucity of groundstone tools, but with celts more common than grooved axes; sandstone tubular pipes of the cloudblower variety, and the absence or scarcity of atlatl parts and bell pestles. Also present at Maple Creek are limestone grubbing tools—one notched and one grooved—that may be a central Ohio Valley counterpart to limonite “hoes” described by Winters (1969) in Riverton Culture contexts. While similarities between the Riverton Culture and the Maple Creek phase are acknowledged to exist, they cannot be construed to be the same for several reasons: • The time spans do not correspond, with Winters acknowledging the beginning of the Riverton Culture in the Wabash Valley ca. 1400 bc— some 250 years after its appearance in the central Ohio Valley. • Artifactual traits such as a soapstone bowl, a stone tablet and stone tablet preform, stone and cannel coal discs, fired clay pottery, bone beamers (which also occur at DuPont), and—especially—McWhinney points and scrapers are not represented in Riverton contexts. • If the Maple Creek and Riverton people were one and the same, migrating from one locality to the other, it would not be unreasonable to expect the transportation of at least some indigenous raw materials or finished artifacts thereof as such treks were made, yet no such items have been reported (or recognized) in the assemblages of one as exotica in those of the other (see Early Archaic above for an example of such behavior). • The ability of prehistoric populations to navigate great expanses of territory is acknowledged. However, the geographic distance separating the two homelands (Illinois and Ohio), combined with the longevity apparently represented by their respective assemblages and continuity of occupation suggesting a degree of residential stability, creates doubt that they were the same group of people migrating seasonally or on some other basis from one region to the other.

Archaic Manifestations, Southwestern Ohio and Vicinity

19

Otto CH1:Layout 1

9/21/08

2:17 PM

Page 20

Merom and Trimble types are not projectile points. They are tools. Patrick Munson (pers. comm. 1974) suggested that careful examination of such specimens likely would lead to this conclusion. No such specimens exhibiting evidence of impact damage to the tip have been observed to date, although some exhibit small flake scars originating at the tip that rather obviously resulted from sustained pressure; that is, from use as a chisel rather than as a projectile, which should have caused massive damage to specimens so small. Furthermore, edge damage consistent with cutting and scraping occurs on most specimens. Only a few of the Merom and Trimble examples illustrated by Winters (1969, plates 13, 14) have broken tips and most of them appear to have been lateral snap fractures probably resulting from too much pressure having been applied on an edge while cutting or scraping or, less likely, from prying. Neither do they appear simply to have been resharpened specimens that broke from impact because the size of their blades is consistent with their haft elements, which are not likely to have been modified in any putative resharpening process if they remained hafted when their edges were rejuvenated or retooled. If the query then becomes one of what the Maple Creek people were using for projectile points, the answer is a variety of Laurentian, Transitional Archaic, and other types—some of which suggest a Midsouth stylistic origin—as well as McWhinney points. McWhinney points not uncommonly show evidence of impact damage, occur embedded in the DuPont male, and co-occur with Merom or Trimble tools (or both) not only at Maple Creek, but at Logan and Driving Range as well. At the latter site, one specimen of each occurred in the same feature where they were associated with a burial (Kreinbrink 1996, 66). For the reasons that McWhinney is one of the point types of Maple Creek phase people, that McWhinney points have a geographically restricted distribution that does not include Illinois, and that the definition of a phase (sensu Willey and Phillips 1958) specifies a small localized area for its occurrence, it is concluded that the Maple Creek phase complies with such a definition and that it is distinct from the Riverton Culture. Multicomponent sites with Maple Creek phase artifacts are numerous throughout southwestern Ohio and vicinity, particularly in proximity to sizable base camps (such as 33CT52 and 33CT34) close to the Ohio River. Minor quantities of such items in upland locations may represent food processing or gathering stations, or other extractive/special purpose encampments. Candidates for single-component Maple Creek phase sites include Sandy Ridge (33CT34) on the valley floor of the Ohio River in Clermont County and the Butler site

20

K. D. Vickery

Otto CH1:Layout 1

9/21/08

2:17 PM

Page 21

(33HA249) (G. Black 1934, 192; Starr 1960, 121) on a terrace of the Ohio River in Hamilton County. Both are sizable base camps with midden accumulations and abundant cultural materials. Horizon 3 at the Maple Creek site is not sufficiently well known from its traits to assign it to any defined phase or other cultural manifestation. Apart from a few manos and a microperforator, Maple Creek phase artifacts are rare or absent. The chipped stone tool assemblage features McWhinney points and scrapers, a Normanskill point, and Maple Creek knives. Other artifacts include simple bone shuttles, rodent incisor chisels, and a chert shredder. Among the ten excavated features were four roasting pits, two earth ovens, one possible storage pit, and an artifact cache of four bifaces. One of the earth ovens was small and cylindrical. A fire had been built on pieces of limestone laid flat on the bottom, probably in an effort to heat the pit walls. Remnants of the fire were then removed and cooking apparently took place in the heated interior without the aid of preheated rocks or direct fire. Another earth oven using this cooking principle was encountered at Bullskin Creek, both representing the only such examples among the excavated Late Archaic sites. The cooking principle is reminiscent of that of the Horno for baking bread that is still in use among various pueblo-dwelling peoples of the American Southwest. The extended burial of a male over 25 years old was immediately overlain by the thermal feature in horizon 3 containing limestone slabs. Having been buried in the horizon 3 natural levee, it would not have been exposed when the thermal feature was in use. Nevertheless, the bones of this individual showed signs of thermal stress from the fire feature above it.

S UB S I STENCE Faunal remains from the five excavated Late Archaic sites indicate the exploitation of a variety of resources, including 83 vertebrate and invertebrate species in which deer or deer and elk combined account for 70 to 87% of the total animal protein intake by weight at any one site. Nevertheless, some river-valley-to-rivervalley, and even site-to-site, variation is evident among the assemblages. For example, DuPont has a greater representation of aquatic species—including fish, mussels, and migratory birds—than do the other excavated settlements, likely due to its proximity to the broad flood plain with backwater lakes and swales that existed near the confluence of the Great Miami and Ohio rivers. Among

Archaic Manifestations, Southwestern Ohio and Vicinity

21

Otto CH1:Layout 1

9/21/08

2:17 PM

Page 22

the five sites, elk was identified only in the two near the Great Miami River in Hamilton County (DuPont and Dravo), indicating the exploitation of forested uplands in addition to aquatic habitats at lower elevations. At Bullskin Creek in Clermont County, the variety and relative abundance of raptorial birds is noteworthy, including red-tailed or rough-legged hawk and both great horned and barred owl. Likely, they were used in ritual/ceremonial contexts rather than as food items. This site is situated on a high terrace adjacent to bluffs where such raptors likely roosted. Edible plant remains from the sites and site components of both Central Ohio Valley Archaic and the Maple Creek phases indicate a heavy reliance on nuts, with the shells or meats of hickory, walnut, butternut, acorn, and hazelnut typically accounting for at least 97%, by weight, of the total edible plant remains. The seeds of fleshy fruits are a far distant second, while grain seeds and legumes (wild beans) rank third. No definite cultigens are reported and the cultivated or wild status of a cucurbit seed from the Late Archaic component of 12SW99, dated 2800–1650 bc, could not be determined. It is represented as either pepo squash or gourd (Houston, McHugh, and Michael 1984, v–8).

S ETTLEMEN T In the uplands of Clermont County, many sites are known with Laurentian projectile point and hafted scraper types occurring in larger numbers than McWhinney, Merom, or Trimble types. This proportional representation contrasts with the Central Ohio Valley Archaic and Maple Creek phase base camps and other sizable settlements on the valley floor near the Ohio River. Furthermore, these sites tend to be distributed in upland river valleys, while sites with McWhinney points and scrapers are predominantly on high ground above these valleys. If the former represent Laurentian Archaic manifestations, it is possible that the settlement pattern was one of base camps in the bottoms of upland stream valleys, with hunting-and-gathering stations in the surrounding hills. The same hills may have served as hunting-and-gathering stations for the Central Ohio Valley Archaic and Maple Creek phase encampments along the Ohio River, with occasional forays into upland stream valleys. Such a situation might have resulted in some mixture of assemblages, as was observed. Unfortunately, it is not known whether these components were contemporaneous or allochronic.

22

K. D. Vickery

Otto CH1:Layout 1

9/21/08

2:17 PM

Page 23

Central Ohio Valley Archaic sites occur more abundantly than sites of other Late Archaic complexes, at least in Hamilton County and probably Clermont County as well, both as large base camps predominantly on the valley floors of major waterways and as smaller encampments and components of larger, multicomponent sites in upland locations. The latter probably represent stations where various extractive activities were performed in support of the base camps. Regional and local base camps are recognized among Central Ohio Valley Archaic sites. The former is uniquely represented by the DuPont site, which, judging from its large size and intensity of occupation, likely served as the locus of amalgamation of several bands for a variety of activities that probably included trading, gift giving, feasting, mate selection, storytelling, crafting, and recreation. While such band aggregation was likely temporary (seasonal or semisedentary), the site apparently was occupied by at least one band or band segment year-round and over a span of some 400 years. On average, an estimated 75 to 100 people lived there. Local base camps were smaller and less intensively occupied—perhaps seasonally. Such sites are known to occur on at least the valley floor of the Ohio River, with smaller sites that were probably hunting and gathering stations in upland tributary river valleys and hills. These settlement types seem to characterize both Central Ohio Valley Archaic and Maple Creek phase sites. With respect to Late Archaic sites in a portion of the central Ohio Valley, Boisvert (1986) has expressed disenchantment with the ethnographic analogue and Riverton-inspired “cyclical fission/fusion” settlement model of “a seasonal round of movement by family bands who periodically fuse and split” (Boisvert 1986, 115) and what he refers to as the “‘Big Site’ mentality” (122) that overemphasizes the importance of large base camps that may lie buried along major waterways, but which may have been “effectively the same in terms of the number of people present on the site at any given time” (119) and in terms of task performance as “satellite or dependent campsites” (119). Instead, he favors the Central Based Wandering (Meggers 1956), or Hub and Spoke (Janzen 1977), model of semisedentary residence at certain loci or along narrow zones with concentrated resources, from which forays to satellite camps were relatively short in duration and limited in purpose (Boisvert 1986, 115). If such clustered resources are in relatively close proximity to one another, seasonal movement is minimized and a residual population may have remained at the same location throughout most of the year (Janzen 1977, 140). While adequate excavation data

Archaic Manifestations, Southwestern Ohio and Vicinity

23

Otto CH1:Layout 1

9/21/08

2:17 PM

Page 24

are lacking, the latter model seems to more closely fit Late Archaic settlement patterns in southwestern Ohio than does the former.

OCCUPATIONAL HISTORY Southwestern Ohio and adjacent regions in Indiana and Kentucky were first populated by Paleoindian bands, whose presence is represented by low site density and very ephemeral occupations. This level of occupation continued through the initial Early Archaic, as represented by the Charleston complex and by Kessell point makers. Greater numbers of bands are evident during the later Early Archaic, beginning about 7500 bc. From this time on, sites are more numerous. Although still rather widely distributed over the landscape, they are consistently encountered whenever archaeological surveys are undertaken. By the middle Early Archaic Kirk phase (7410–6900 bc), at least one base camp had been established at the Ferris site. Ephemeral occupations again characterize this region during the early Middle Archaic, with possible clusters along stretches of major river bottoms or in the valleys of upland tributaries. During the late Middle Archaic and early Late Archaic, from ca. 5000 to 3000 bc, the region was either abandoned or sparsely occupied by bands whose artifactual assemblages have escaped recognition. It is possible that Laurentian bands, presumably from a homeland to the north and east, established base camps and other settlements in the area over the next 250 years or so. An influx of people into the central Ohio Valley is clearly evident during the middle Late Archaic, when this region was intensively occupied. This dense occupation is reflected in part by data compiled by Cinadr (1980) for Clermont County, where the largest culture period representation among 431 components is the Late Archaic, accounting for 20% (85 components) of the total. In Hamilton County, Late Archaic and “undifferentiated” Archaic combine to comprise 14% of the total site components recorded at the time. Likely, these groups moved into the area from the west, “fleeing” the Xerothermic Interval, which was very gradually making its way across the Midwest from west to east. The following Maple Creek phase is a “site-unit intrusion” into the central Ohio Valley. The slightly earlier dating of this manifestation here than in the Wabash Valley, where it was defined, suggests that its movement across the

24

K. D. Vickery

Otto CH1:Layout 1

9/21/08

2:17 PM

Page 25

Midwest was from east to west. It is interesting to note the presence of Merom Expanding Stemmed “points” (Holland 1970, 89, plate 16h–k) and a “Riverton Stemmed” point identified as such by Winters himself that “belongs to the Riverton Culture” (Holland 1970, 91) in southwestern Virginia, to the south and east of the central Ohio Valley. An inquiry to Charles S. Bartlett of the Archeological Society of Virginia as to whether such points are represented by chronometric dates in this area met with negative results. Bartlett (pers. comm. 1995) confirms that current use of the Merom label accommodates Trimble points as well, and that both are ubiquitous in at least southwestern Virginia. It may not be too far-fetched to suggest at least a morphological, and perhaps size, similarity between Merom/Trimble and Brewerton Corner- and Side-Notched points. For example, one would be hard pressed to distinguish two of Ritchie’s (1969, plate 29, nos. 15, 16) illustrated Brewerton Side-Notched points from Merom and two others (10, 11) are quite small in size. None appear to have been resharpened from larger specimens. Furthermore, all four of Ritchie’s Brewerton types have diminutive counterparts in Maple Creek phase assemblages. When and if the Merom points in Virginia are chronometrically dated, it will be interesting to learn whether the dates are even earlier than those in the central Ohio Valley. If they are, and if Brewerton points are indeed the early prototypes for Merom and Trimble tools, then an east-to-west movement for this complex must be seriously entertained.

I NT ERPRETATION S Howard Winters (1974) attributes the success of Midwestern Late Archaic cultures such as Indian Knoll and Riverton to a highly efficient, “narrowspectrum” economy that geared its subsistence efforts mainly to the exploitation of deer, freshwater mussels, and nuts—an idea with roots in Caldwell’s (1958, 1962, 1965) concept of “primary forest efficiency” that he believed is first manifested in the Late Archaic period. The idea of a narrow-spectrum economy for the Late Archaic in the central Ohio Valley is bolstered somewhat by niche width calculations undertaken by Emanuel Breitburg, who analyzed the five excavated faunal assemblages (Vickery and Breitburg 1991). The niche width value for total variety is 1.90 and for spatial variety, 1.30, indicating fewer different species exploited and the exploitation of more restricted environmental

Archaic Manifestations, Southwestern Ohio and Vicinity

25

Otto CH1:Layout 1

9/21/08

2:17 PM

Page 26

zones for food than among later Fort Ancient populations (with niche width values of 2.90 and 1.50, respectively) (Breitburg, pers. comm. 1995). Winters, however, also notes that areas where relative abundances of resources supporting such an economic orientation were rare in the Midwest, and suggests that there may have been competition for the right to occupy and exploit them. Such competition may have escalated toward the end of this period due to “subsistence crises” as environmental conditions contributed to “natural fluctuations” in the availability of such staple food resources, combined with their overexploitation during periods of declining productivity. This hypothesis may be investigated along several lines. As noted earlier, the Xerothermic Interval—an episode of generally warmer and drier conditions—was time transgressive across the Midwest, traversing the region from west to east, from Illinois ca. 5000 bc to northwestern Pennsylvania between 1650 and 1000 bc (Walker and Hartman 1960). The result was natural deforestation as uplands in some areas changed from forests to prairies, followed by the gradual drying of river valley bottomlands. In their Stages Pond pollen diagram in south-central Ohio, Shane, Snyder, and Anderson (2001, 31, 34–35) note high nonarboreal pollen, indicating greater openings in forest canopies, beginning ca. 2000 bc, an occurrence they relate to warm, dry conditions and prairie extensions from the west. These findings are consistent with those of Ogden (1966) for his Silver Lake pollen diagram in Logan County, Ohio. Both the Silver Lake and Medway Bog (Williams 1957) pollen diagrams in southwestern Ohio confirm that not only did hickory, walnut, and butternut experience “fluctuations” in their relative proportions during the closing centuries of the Late Archaic, but that these important nut-bearing genera also show a trend of declining abundance throughout this span of time. Acorn, which was exploited by Late Archaic bands, as well as having been a major food source for white-tailed deer, shows a similar trend. The findings of Schoenwetter (1974) in Kentucky tend to support these data, which suggest increasingly more extensive openings in the forest canopy. The ensuing declines in at least nuts and deer would have created population pressure that is manifested by conflict. There is evidence of armed aggression among Late Archaic—and even earlier—bands, likely having taken the form of small-scale raids for other bands’ food or access to portions of their exploitative territories. Violent deaths have been reported at Riverton, Indian Knoll phase sites, and sites in Tennessee (Mensforth 2001);

26

K. D. Vickery

Otto CH1:Layout 1

9/21/08

2:17 PM

Page 27

southwestern Ohio; and at the Lamoka Lake and Frontenac Island sites in New York. While some Late Archaic populations made adjustments in their diets in accordance with localized food availability—as represented by analyses of prehistoric faunal and floral assemblages (e.g., Featherstone 1977)—others responded to this ongoing problem of declining food resources, restricted mobility, and population pressure by outmigrating from at least some of the Late Archaic “strongholds,” as reported in the Wabash Valley, the Falls of the Ohio area near Louisville, southwestern Ohio, and portions of the Midsouth and southern New England. Near Louisville, Janzen (1977) reports a very intense occupation early in the Late Archaic, but after about 3200–3000 bc, his search for evidence that hunter-gatherer bands stayed in the area and evolved into later cultures was in vain. They apparently moved out of the area and headed upriver, eventually ending up in the central Ohio Valley. In 1974, Wendland and Bryson published the results of a worldwide study of more than 800 radiocarbon dates associated with pollen maxima and minima, sea level maxima and minima, and peat beds, from which compilation they identified 13 climatic “discontinuities.” These they represented as globally synchronous episodes of climatic change, although the specific way in which each was manifested varied in different parts of the world. Their goal was to correlate these climatic discontinuities with culture changes, allowing some 150 years after the onset of each climatic episode for people to “respond” in ways that might be detectable in the archaeological record. Minor discontinuities were recognized by Wendland and Bryson at 2290 bc and 1620 bc, while a major one occurred at 810 bc. It is not inconceivable that their minor discontinuities presented a double whammy to certain Midwestern Late Archaic peoples while the major one tolled their death knell. In 1968, Charles Erasmus (1968) coined the term upward collapse and applied the concept to the Maya. The localized disappearance of deciduous forests and their replacement by grasslands brought to a close the distinctive and, in many senses, impressive lifeways of various Midwestern Late Archaic cultures. However, the ecological conditions that existed then favored the growth of certain weedy pioneer annuals whose starchy and oily seeds became a welcome source of food and were soon domesticated as an indigenous Midwestern development to be grown in the gardens of Late Archaic and Early Woodland peoples.

Archaic Manifestations, Southwestern Ohio and Vicinity

27

Otto CH1:Layout 1

9/21/08

2:17 PM

Page 28

Thus it was that the Late Archaic collapsed, in the sense that formerly successful lifeways no longer could be sustained. But perhaps it was an “upward collapse,” for out of the ashes of the Late Archaic arose the phoenix of the Early Woodland and Adena.

ACKN OWLEDGMENTS The dedicated contributions of many students and volunteers involved in site surveys, field school and salvage excavations, and lab processing and analyses are gratefully acknowledged. Playing particularly important roles in these efforts are Jim Theler and Tim Dalbey for their field and lab work on the DuPont site and its contents, Bryan Featherstone for analysis of the DuPont floral remains, and Manny Breitburg for analysis of all excavated faunal assemblages. Special thanks are due A. J. Hill-Ariens for her preparation of figure 1.1.

28

K. D. Vickery

Otto CH2:Layout 1

2

9/21/08

2:26 PM

Page 29

Cultural Resource Management and the Analysis of Land Use Patterns of the Archaic in North-Central Ohio Craig S. Keener, Kolleen Butterworth, and Crystal L. Reustle

The CRM surveys of the SR 30 project identified 2,949 prehistoric sites (329 of which have an Archaic component) (Butterworth and Keener 1995), which can be used to address land use patterns. These data should be considered a sample representation of a portion of north-central Ohio. Mathematical calculations concerning landforms and site locations will be presented. Two different physiographic regions, Till Plains and Glaciated Plateau, are contrasted using landform data compiled in the database. The results of this analysis are presented and CRM data and methodology are evaluated in terms of their reliability and applicability for regional study.

B ACKGROUN D OF STUDY AR E A In accord with the Ohio Department of Transportation’s plans to relocate and widen sections of State Route 30 within a 100-mile-long corridor in northcentral Ohio (fig. 2.1), three to four alternate alignments, varying in width from 500 to 2,000 feet, were surveyed for archaeological resources. This survey was broken into six spatial segments, which were investigated by three consulting firms. ASC Group Inc. of Columbus, Ohio, was assigned segments II, IV,

29

Otto CH2:Layout 1

9/21/08

2:26 PM

Page 30

Fig. 2.1. Investigated segments of the SR 30 corridor in Ohio.

VII, XII (Dobson-Brown, Hunter, and Gibbs 1993; Gibbs, Frye, and DobsonBrown 1994; Grimes, Prosser, and Dobson-Brown 1991, 1992; Hillen and DobsonBrown 1991; Mustain and Dobson-Brown 1992, 1993; Sprague, Dodson-Brown et al. 1995; Whitman et al. 1995); the Center for Cultural Resource Research at the University of Pittsburgh investigated segment IX (Bush et al. 1995); and segment XIII was awarded to Gray & Pape Inc. of Cincinnati (Weed and Tuttle 1993). Geographically, the setting for this study is divided into two basic areas: the Till Plains and the Glaciated Plateau (fig. 2.2). The Till Plains have a some-

30

C. S. Keener, K. Butterworth, and C. L. Reustle

Otto CH2:Layout 1

9/21/08

2:26 PM

Page 31

what flat, well-drained topography. Flat ground moraines are predominant and are punctuated by end moraines that are higher than the adjacent landscape. Soils in the Till Plains formed in the high-lime glacial drift and glacial lake sediments of Wisconsinan age. The Glaciated Plateau has a more rolling topography and includes ground and end moraines as well as kames and outwash deposits. Soils in the Glaciated Plateau are formed in the low-lime glacial drift of Wisconsinan age (Goldthwait, White, and Forsyth 1961; ODNR 1973; Lafferty 1979). Of the six segments, three are in the Till Plains and three are in the Glaciated Plateau. The project area includes the upper drainages of the Lake Erie basin and the Ohio River, with the following secondary river systems: the Maumee, Muskingum, and Scioto (ODNR 1984). Topographically, the upper drainages of these secondary systems are typified by more heavily dissected stream patterns and by shallow valleys. The Archaic period in Ohio represents an approximately 7,000 year span split into three periods: Early Archaic (8000–4500 BC), Middle Archaic (4500– 3000 BC), and Late Archaic (3000–1000 BC). These periods are typically associated with typological changes in artifacts. The Middle Archaic is a distinct period during which many of the principle features of Late Archaic subsistence and settlement first appear. Additionally, the Late Archaic is characterized by an increase in site size, a change in diet (more plant use), increased evidence of semipermanent settlements, and an increased use of slate artifacts (Pratt 1979; Stothers and Abel 1991; Vickery and Litfin 1994). A transition from a nomadic foraging pattern to a more regionally focused subsistence strategy is the general trend characterizing the Archaic period.

T H E DATABASE The SR 30 investigations produced a CRM-derived archaeological database consisting of 2,949 prehistoric sites. Frequencies were tabulated for temporally affiliated and unassigned prehistoric sites (table 2.1). The database program Paradox 4.1 was used to store recorded information about the sites, and to tabulate counts and make summary statistics of site categories and characteristics. The following descriptive categories express the environmental and locational attributes of each site: site type, cultural/temporal period, local landform, soil type, distance to the closest water source, elevation, and site size.

CRM and the Analysis of Land Use Patterns

31

Otto CH2:Layout 1

9/21/08

2:26 PM

Page 32

Fig. 2.2. Physiographic sections of Ohio (modified from Goldthwait, White, Forsyth 1961).

Unassigned prehistoric sites (n = 2,478)—sites that did not yield any culturally diagnostic artifacts—composed 84% of the total number of sites within the six segments of SR 30. Of the remaining 16%, the Archaic period was best represented, with a total of 329 components, representing 70% of assemblages possessing diagnostic material. When the Archaic sites are temporally divided into Early, Middle, and Late periods, the Late Archaic has the highest number of components with a total of 189 (57%), while the Early Archaic is represented by 105 (32%) components (Butterworth and Keener 1995). When site totals

32

C. S. Keener, K. Butterworth, and C. L. Reustle

Otto CH2:Layout 1

9/21/08

2:26 PM

Page 33

Table 2.1. Cultural component totals for SR 30 investigations Cultural component

Count

Paleoindian

15

Early Archaic

105

Early to Middle Archaic Middle Archaic Middle to Late Archaic Late Archaic Unspecified Archaic

1 25 2 189 7

Early Woodland

32

Middle Woodland

25

Middle to Late Woodland Late Woodland Unspecified Woodland Late Prehistoric Unassigned Prehistoric Protohistoric

1 39 4 24 2,478 2

were broken down by topographical region, 1,814 sites, of which 208 (11%) have Archaic components, were in the Till Plains. The Glaciated Plateau contains 1,135 sites, 121 (11%) of which have Archaic components.

MET H O DOLOGY The methodology and goals of this study were based on principles of regional archaeology. Regional archaeology has been defined by several researchers, with the basic premise that a regional approach advocates an understanding of the interrelationships of sites in a given area in terms of spatial or temporal properties (Read 1974). This approach advocates a distributional study of material culture with a vigorous control of space that not only entails documentation of sites in a region, but also focuses on the intersite spacing of sites to expose land use patterns. This approach contrasts with type site descriptions of areas in which one large stratified site is typically used to define prehistoric populations in a uniform manner across broad geographical areas and over a

CRM and the Analysis of Land Use Patterns

33

Otto CH2:Layout 1

9/21/08

2:26 PM

Page 34

long time. This traditional methodology led to the development of phases and standardized descriptions of cultural land use patterns that often leave little room for considering variation in human populations and change through time. While type site descriptions rely on a single site as representative of human settlement patterns, a regional approach accounts for the many variables that may affect those patterns, including topography, distance to water, geomorphology, vegetation, and biotic communities. Examination of all or any combination of these variables will enhance our understanding of the variability of prehistoric settlement patterns. We hoped that application of this regional approach to the SR 30 database could establish tentative land use patterns. However, this method proved to be impracticable because of the limitations imposed by the variations of recovery methods employed during the surveys as well as the use of outdated soil books and traditional field techniques. A comprehensive regional archaeological approach was not possible without returning to the surveyor’s original field notes and perhaps even returning to the field. The only variable independently useful in this circumstance is the local landform in relation to its geographical setting. Other regional analyses of landforms and site locations have already been conducted in the Midwest. In Ohio, Flora Church (1988) incorporated CRM data in her analysis of Late Woodland and Late Prehistoric landform use patterns along the central Scioto River Valley. Robert Warren (1992) conducted another regional study in Illinois, where he used CRM data to determine land use patterns along the I-39 highway project. Warren’s study used topographical variables similar to those employed in the SR 30 investigations. The current study compared local landforms in the Till Plains and the Glaciated Plateau, concentrating specifically on the Archaic period. In addition, Early and Late Archaic components were compared to see if land use patterns changed through time. We presumed that there would be differences in site locations over time as a response to changes in the climate or drainage of the soil. Unlike Church’s study (1988), we used a larger number of topographical variables to describe site locations—Upland, Valley Floor, Ridge, Upland Flat, Glacial Lake Margin—instead of only two (upland and valley floor). The increase in landform categories provides more variation in the physiographic description of a region, and also accounts for the differential topographical descriptions and definitions of landforms among and between CRM companies. Another aspect to consider while reviewing this information is that the results are a reflection of research strategies that do not necessarily

34

C. S. Keener, K. Butterworth, and C. L. Reustle

Otto CH2:Layout 1

9/21/08

2:26 PM

Page 35

consider alluvial and colluvial depositional factors (see Brown and Vierra 1983). These factors will contribute to the underrepresentation or overrepresentation of sites in the sample region.

RES ULTS The five identified landforms were quantitatively analyzed on the basis of temporal affiliation. Percentages of each landform type for each of these variables were calculated within all the temporal periods represented in the database. Variable percentages associated with the Archaic period were visually compared between the two geographical settings (Till Plains and Glaciated Plateau) and within and between the Archaic periods within the regions. Four charts or graphical presentations were developed as a result of the following calculations: the Archaic as represented throughout the project area, comparison of the Archaic in the Till Plains with the Archaic in the Glaciated Plateau, comparison of the Early Archaic versus the Late Archaic in the Till Plains, comparison of the Archaic versus the unassigned prehistoric in the Till Plains.

The Archaic Representation in Both the Till Plains and the Glaciated Plateau Figure 2.3 shows the percentages of Archaic sites found on five landforms (Upland, Valley Floor, Ridge, Upland Flat, Glacial Lake Margin) along the entire SR 30 corridor. Archaic sites identified across the entire corridor were typically located in upland situations. Upland sites comprised 67% of the Archaic, while 14% of the sites were on the valley floors. These numbers may be a reflection of the percentage of landforms surveyed (undetermined at this time).

The Archaic in the Till Plains versus Archaic in the Glaciated Plateau A comparison of Archaic sites in the Till Plains versus those in the Glaciated Plateau produced differences between the locations of sites within the two physiographic areas. As shown in figure 2.4, 79% of the Till Plains sites were in upland situations and only 3% were on the valley floor. Those on the Glaciated Plateau are in the uplands and on ridges 60% of the time, whereas

CRM and the Analysis of Land Use Patterns

35

Otto CH2:Layout 1

9/21/08

2:26 PM

Page 36

Fig. 2.3. Landform distribution Archaic sites along the entire SR 30 corridor.

Fig. 2.4. Landform distribution for Archaic sites of the Till Plains versus Archaic sites of the Glaciated Plateau.

Otto CH2:Layout 1

9/21/08

2:26 PM

Page 37

Fig. 2.5. Landform distribution of the Early Archaic sites versus Late Archaic sites of the Till Plains.

40% of the sites are on valley floors. This apparent dissimilarity of site location between the two physiographic regions would not have been revealed had no distinction been made between the two geographical regions (compare fig. 2.3 with fig. 2.4).

Early Archaic versus Late Archaic in the Till Plains Since more sites were discovered in the Till Plains, we were able to conduct a more specific analysis for the Early and Late Archaic phases in this region (fig. 2.5). A large percentage (79%) of both Early and Late Archaic sites are in upland situations (fig. 2.5). The only differences between the two time periods were Late Archaic increases in the percentages of sites on glacial lake margins (6.4% increase), ridges (1.6% increase), and valley floors (1.6% increase). A decrease in the number of sites on the upland flat (8.5% decrease) was also noted.

CRM and the Analysis of Land Use Patterns

37

Otto CH2:Layout 1

9/21/08

2:26 PM

Page 38

Fig. 2.6. Landform distribution of Unassigned Prehistoric sites versus Archaic sites of the Till Plains.

The Archaic versus the Unassigned Prehistoric in the Till Plains Figure 2.6 compares the Archaic sites and unassigned prehistoric sites in the Till Plains. Unassigned prehistoric sites, often derided for lacking diagnostic artifacts, do provide invaluable information on prehistoric land use. If seen as a general representation for human use of the landscape from Paleoindian to Protohistoric, the information that unassigned sites express for a region can be used as a baseline for contrasts with specific temporal periods and regions. The comparison of the Archaic and the unassigned prehistoric revealed that the upland landform was the predominant location of sites in both categories (unassigned prehistoric, 72%; Archaic, 79%). In summarizing the charts, the greatest contrast of variables appears when comparing sites from the two main geographical settings. There is an apparent difference among Archaic site locations between the two physiographic regions, which may indicate different adaptive responses to two environments. However, site location within either region did not appear to be significantly different.

38

C. S. Keener, K. Butterworth, and C. L. Reustle

Otto CH2:Layout 1

9/21/08

2:26 PM

Page 39

These general observations are preliminary and strictly based on the data provided in the CRM reports. Different methodologies and recovery techniques used by each of the CRM firms may have caused underrepresentation or overrepresentation of temporally diagnostic sites and their locations, thus skewing the results of our calculations. Additionally, the percentage of the types of landforms surveyed needs to be accounted for in order to adjust the percentages gleaned from the CRM data provided. The graphs presented here can and should be tested by future site information and other environmental variables before any solid regional conclusions can be drawn about landform use in north-central Ohio. Although this regional study of the SR 30 corridor is introductory in nature, it holds much potential for the development of land use patterns based on physiographic regions. This preliminary study indicates that the Archaic site locations differ when geographic settings are compared. The patterns revealed by this study are further testable by discovering new sites or by adding sites not used in the original formation of this database. Ideally, if other environmental and locational variables of north-central Ohio are studied according to principles of regional archaeology, then perhaps a more comprehensive land use pattern can be developed for this specific area. Because of its size and general cohesiveness, the SR 30 project affords a unique opportunity to analyze site landform use data from a regional perspective. However, problems did surface during the compilation of data for this paper. These questions are a reflection of problems with procedures for data recovery used by CRM surveys, which is also a problem with the field of archaeology in general. Do we as professional archaeologists, as scientists, use the same methods, measurements, and descriptive techniques when we define a site? The answer is no. For example, two critical pieces of information, site size and artifact density, could not be used in this analysis because of the interpretative and biased manner in which sites are recorded. Site size is based on the researcher’s own skills and ideas, not on a consistent scientific formula. As a result, two different archaeologists could record two different site sizes for the same location. Obviously, this situation produces real obstacles for regional archaeology that must be addressed if a comprehensive study of a region is to be undertaken. Once this problem is overcome, regional interrelationships can be elaborated. For example, it would be possible to look at artifact and cluster densities in upland situations versus those on the valley floor in order to address variation and changes in land use.

CRM and the Analysis of Land Use Patterns

39

Otto CH2:Layout 1

9/21/08

2:26 PM

Page 40

Other discrepancies in the database caused problems for regional interpretations. For example, the terminology used to describe local landforms on Ohio Archaeological Inventory (OAI) forms were inconsistent and left much room for interpretation. Landform descriptions varied among the CRM firms, and even within individual firms. A definitive landform description needs to be developed and used consistently. Discrepancies occur in comparisons of soil types between counties. In several instances, only a soil association name or a specific soil type name was listed in the site data rather than both names as requested on the OAI form. This situation is complicated by out-dated information from older county soil surveys that sometimes do not identify soil type. As a result, we were unable to make a reliable soil analysis of the study area. It is unknown what effect recovery methodologies used by the CRM companies doing the SR 30 surveys had on the number and location of sites. Although the project has produced the largest amount of archaeological information, to date, for several north-central Ohio counties, the question remains whether or not the sample was large enough to conduct regional interpretations. Using data collected from CRM surveys for regional research-oriented investigations can be problematic, but it is not unworkable. There currently exists a vast body of invaluable data recovered by CRM surveys that remains relatively unused. Every year, CRM investigations lead to the accumulation of thousands of newly recorded sites in Ohio. Once these sites are located, they are recorded in the Ohio Archaeological Inventory (OAI), housed in the Ohio Historical Society’s State Historic Preservation Office. Archaeologists interested in regional land use patterns have already begun to use the data; this interest should continue to grow, especially with the greater ability to analyze large databases on computers. Appropriate research strategies could and should be developed to use this important source of information.

ACKN OWLEDGMENTS We are much indebted to the following people, who helped point us to relevant data or supplied us with pertinent information and critiques: Jarrod Burks, Erica Keener, and the Ohio Department of Transportation archaeological team, specifically Jim Addington and Dr. Bruce Aument.

40

C. S. Keener, K. Butterworth, and C. L. Reustle

Otto CH3:Layout 1

3

9/21/08

2:28 PM

Page 41

Down by the River LATE ARCHAIC THROUGH TERMINAL ARCHAIC DY N A M I C S A T T H E D AV I S S O N FA R M S I T E ( 3 3 L E 6 1 9 ) , L A W R E N C E C O U N T Y, O H I O

Matthew P. Purtill

The time between ca. 4000 and 700 bc has long been identified as a period of cultural transition and significant technological change (e.g., Stoltman 1978). Within the middle Ohio Valley, the stimulus for the development of many of the cultural traits typically associated with prehistoric life for the next 2,000 years—increased sedentism and parochialism, horticulture/agriculture adoption, elaborate ceremonialism, interregional trade, technological innovations—can be traced to this period. Regional archaeological research has begun to relate such changes to cultural developments in a broader geographic expanse. Although research in neighboring states such as West Virginia and Kentucky has contributed to our understanding of these trends (e.g., Granger 1988; Hughes, Kerr, and Pecora 1991; Ledbetter and O’Steen 1992; Pedde and Prufer 2001), well-documented excavation data from south-central Ohio has been either limited in scope (e.g., localized surface surveys with minor subsurface excavation), not easily assessable (e.g., archaeological contract reports), or even misleading (e.g., the poorly defined Dunlap phase, see below). Recent cultural resources management (CRM) excavations at the multicomponent Davisson Farm site (33LE619), in Lawrence County, Ohio (Purtill 2001, 2002c; Purtill et al. 2001), have provided critical data regarding the Late Archaic through Terminal Archaic periods in south-central Ohio. For the

41

Otto CH3:Layout 1

9/21/08

2:28 PM

Page 42

purposes of this chapter, Late Archaic refers to the period between 4000 and 1200/1000 bc, and the Terminal Archaic between 1200/1000 and 700 bc. As the term is here used, Archaic refers to a developmental stage (sensu Willey and Phillips 1958) while both Late Archaic and Terminal Archaic denote discrete sequential temporal spans (i.e., periods within this stage).1 Davisson Farm first was identified professionally during Phase I CRM surveys in the winter of 2001 by Gray & Pape Inc. of Cincinnati, Ohio. Phase I investigations consisted of controlled surface collection and limited shovel testing of a series of plowed agricultural fields (fig. 3.1). The Phase I work indicated that the site retained high potential for inclusion in the National Register of Historic Places (NRHP), and Phase II eligibility testing was completed in the spring of 2001 (fig. 3.2). These investigations involved 100% surface collection within 10 x 10 m quadrants, followed by mechanical stripping in search of features. A total of 1,750 m2, representing about 1% of the site, was excavated and revealed 32 prehistoric features. Four samples submitted for 14C dates provided a cluster of dates between 2000 and 1000 bc. Surface collections produced almost 7,000 artifacts, including a large assemblage of formal tool classes. Based on the abundance of subsurface features and areas of high artifact densities, the site was determined eligible for inclusion in the NRHP in 2001. Because the site could not be avoided by the project, mitigation excavations (Phase III) were conducted. Mitigation focused on extensive mechanical striping and feature excavation. Toward that end, a total of 3,200 m2, representing approximately 4% of the site, was excavated during the summer of 2001. Results of these investigations indicated that, although the site was multicomponent, the primary occupation occurred between the Late Archaic and Terminal Archaic periods.

T H E SITE AN D ITS EN VIRO N M E N TA L SE T T I N G Davisson Farm is in Hamilton Township, Lawrence County, Ohio, along a third terrace (T3) of the Ohio River (figs. 3.1, 3.2). Lawrence County is situated within the unglaciated Appalachian Plateau physiographic province (Noble 1975). Although located in a periglacial setting, the area undoubtedly was influenced by Illinoian and Wisconsinan glaciation events that occurred 112 km north of the project area. Processes such as nivation, frost-wedging, hydrofrac-

42

M. P. Purtill

Otto CH3:Layout 1

9/21/08

2:28 PM

Page 43

turing, and gelifluction were likely active during such glacial advances. Glacial outwash discharge into the Ohio River also played a role in terrace formation (McCleary and Hamilton 1998). This portion of the valley is approximately 1.6 km wide and is characterized by three discontinuous terrace levels (T1, T2, T3) in addition to a narrow floodplain (T0). Geomorphological investigations in Lawrence County suggest that the T1 terrace dates to the Holocene; whereas, the T2 and T3 terraces were formed earlier—by the end of the Pleistocene (Purtill et al. 2001; Vento 2001). Importantly, both the T2 and T3 terraces represent geologically stable landforms, not often inundated by Holocene flooding. The site is in the Mixed Mesophytic Forest region (Braun 1950, 39–121) characterized by a combination of oak, hickory, tulip poplar, walnut, maple, and other hardwoods (Sears 1941; Gordon 1969). Exact forest composition, however, is highly variable and dependent on microenvironmental conditions such as topography, sunlight exposure, soil type, and moisture. Vegetational surveys in nearby Scioto County, Ohio (Gordon 1969, 86), and neighboring counties in Kentucky (Braun 1950, 95), demonstrate the former occurrence of oak, hickory, and pine on elevated hillsides and hilltop environs (especially above 240 m above mean sea level [amsl]). Beech, maple, and walnut, by contrast, were more prevalent along valley floors. The Davisson Farm site proper is on the southern T3 terrace edge. The site overlooks a small, unnamed tributary that bisects the lower T2 terrace to the south. This tributary also bounds the site to the east. To the north, the site extends, discontinuously, along the T3 terrace tread and terminates 300 m south of steep bedrock bluffs. Within the project area, the site is approximately 8.9 ha. However, project constraints precluded determining its actual dimensions, which are undoubtedly larger. Site soil types, as mapped by the United States Department of Agriculture (McCleary and Hamilton 1998) include members of the Elkinsville-SciotovilleNolin soil group. Deep trenches (trenches G1 and G2) excavated during eligibility testing (Phase II) and bucket auguring conducted during the survey (Phase I) exposed soil profiles consisting of Ap-Bw-stacked C soil horizons. Both trench profiles revealed increasingly sandier (i.e., more and more coarse) soils with increased depth. River gravels associated with an abandoned Ohio River channel of Pleistocene age were noted in several locations, most prominently within trench G2 at a depth between 2.5 and 3 m. Importantly, no deeply buried soil horizons or artifact zones were encountered.

Late Archaic–Terminal Archaic Dynamics, Davisson Farm Site

43

Otto CH3:Layout 1

9/21/08

2:28 PM

Page 44

Fig. 3.1. Aerial photograph showing terrace formations and location of the Davisson Farm site, Lawrence County, Ohio.

Otto CH3:Layout 1

9/21/08

2:28 PM

Page 45

Fig. 3.2. Plan view of Davisson Farm showing combined Phase II and III investigations.

FEAT URES Investigations revealed 80 prehistoric features that date to the Late Archaic and Terminal Archaic components. Temporal designation of features was based on the presence of diagnostic artifacts, 14C dates (table 3.1, fig. 3.3), and proximity to other dated features. To characterize the assemblage, a typology based on morphological characteristics was developed. Morphology provides a more accurate significatum for feature definition and classification than feature content. The validity of this approach at Davisson Farm is strengthened by evidence suggesting that site features were not single-use facilities, but instead often represented “mixed” assemblages from multiple-reuse and in-filling episodes (see below). The potential danger of equating feature content with feature function has been noted by other researchers as well. For example, at the Late Archaic Missouri Pacific no. 2 site in west-central Illinois (McElrath and Fortier 1983), a negative correlation between pit volume and material density led the authors to conclude that “the variation in cluster densities of material, as

Late Archaic–Terminal Archaic Dynamics, Davisson Farm Site

45

Otto CH3:Layout 1

9/21/08

2:28 PM

Page 46

Table 3.1. Late Archaic through Terminal Archaic radiocarbon determinations from the Davisson Farm site Feature

Lab no.

Radiocarbon years BP

Calibrated date range*

42

161296

2670 +/-70

1007–594 bc

80

161297

2790 +/-70

1127–807 bc

111

161299

2810 +/-70

1192–814 bc

95

161298

3070 +/-60

1487–1130 bc

9

155077

3190 +/-40

1600–1392 bc

8

155079

3190 +/-40

1600–1392 bc

7

155078

3250 +/-90

1746–1316 bc

26

155076

3300 +/-130

1921–1296 bc

45

161294

5120 +/-70

4048–3712 bc

*The 2-sigma range, based on Reimer et al. 2004.

measured from pit contents, is, therefore, probably due to differences in the intensity of activities conducted around pits rather than in them. It would also appear that the function of the pits at this site cannot be detected by the presence of specific cultural materials found in pits” (McElrath and Fortier 1983, 87, emphasis added). Although feature content was not used directly in the classification of feature types, feature content represents cultural activities at a sitewide, and temporal, level. Measurements of each feature included length₁, length₂, depth, and volume (for the results, see table 3.2). Length₁ represents the longest feature dimension whereas length₂ is the dimension perpendicular to length₁. Depth was determined from profile drawings and represents the maximum feature depth. The volume of basin-shaped pit features was estimated using standards developed during the FAI-270 Highway Project in Illinois (e.g., McElrath and Fortier 1983, 31–32). There are six feature types defined for the Davisson Farm assemblage: large basin-shaped pits, small basin-shaped pits, fire-cracked rock (FCR) clusters, postmolds, small basin-shaped pits or postmolds, and surface hearths or burned areas. In general, these types are similar to ones recorded at contemporary middle Ohio Valley sites, such as the Maple Creek site in southwestern Ohio (Vickery 1976), the Grayson site in northern Kentucky (Ledbetter

46

M. P. Purtill

Otto CH3:Layout 1

9/21/08

2:29 PM

Page 47

Fig. 3.3. Late Archaic through Terminal Archaic 14C date ranges (uncalibrated and 2-sigma calibrated) from Davisson Farm.

Otto CH3:Layout 1

9/21/08

2:29 PM

Page 48

Table 3.2. Feature measurements by feature type Descriptive statistics (cm) Feature and measurement

N

Min

Max

Mean

SD

Large basin-shaped pits Length₁ Length₂ Depth

6

100

200

145.00

34.28

6

90

135

106.50

16.22

6

10

Volume (dm3)

6

43.84

54 456.16

36.66

16.08

228.93

142.25

65.41

18.14

Small basin-shaped pits Length₁ Length₂

44

35

44

34

84

56.70

14.52

Depth

44

8

52

20.75

8.56

44

4.87

114.68

31.43

21.78

11

13.00

35.00

21.54

6.35

11

12.00

35.00

18.63

6.86

11

8.00

40.00

19.36

9.90

13

15.00

68.00

43.61

17.51

13

15.00

60.00

37.07

14.96

2

28.00

108.00

68.00

56.57

2

26.00

99.00

62.50

51.62

3

Volume (dm )

96

Postmolds Length₁ Length₂ Depth FCR clusters Length₁ Length₂ Surface hearth/burned area Length₁ Length₂

and O’Steen 1991, 1992), the Corey site in northwestern West Virginia (Hughes, Kerr, and Pecora 1991), and the Bluestone site in central Kentucky (Brooks, Brooks, and Collins 1979) (for extended discussion, see Purtill 2002c, 122–23).

Large and Small Basin-Shaped Pits Large basin-shaped pits (hereafter, large pits) (n = 6) were shallow pits with oval to round plans and slightly sloping side walls (fig. 3.4). By definition, the length₁ of large pits was at least 1 m. In contrast, small basin-shaped pits (small

48

M. P. Purtill

Otto CH3:Layout 1

9/21/08

2:29 PM

Page 49

pits) (n = 45) had generally circular plan views and steeply sloping side walls. Length₁ of small pits was less than 1 m. Charcoal occurred in nearly all features, whereas burned earth was noted with much less frequency (25%). Typically, internal feature stratigraphy, or zoning, was not present in either feature type; feature soil was homogenous throughout. Both large and small pits appear to have functioned, almost exclusively, as cooking/roasting pits rather than as storage facilities. As represented in figures 3.4 and 3.5, large pits were morphologically variable and do not appear to have been constructed following a rigorously preconceived “design.” The construction of small pits, however, appears to have been more standardized, especially for length₁, length₂, and volume variables (see fig. 3.5A, D). By contrast, when depth is plotted against other variables (volume and length₂) little patterning resulted (fig. 3.5A, B). The difference between depth and volume plots suggests that prehistoric populations were more concerned with constructing small pits of a standard volume rather than a standard depth. Within the large-pit category, several features (6, 20, 42) contained internal, smaller, subpits originating at the base of each feature and extending downward several centimeters (fig. 3.4). Such subpits were generally circular and steep sided. Feature 6 also deserves additional note because of a shelflike outer perimeter on which numerous unburned river cobbles were cached, presumably for later use. This design type was not identified elsewhere at the site. Finally, feature 8 (small pit) deserves special note because it was the only feature to contain an internal, FCR-lined circular “hearth” at its base. Similar feature designs occur at contemporaneous sites in the region (e.g., Grayson site [Ledbetter and O’Steen 1991]). Both large and small pits contained minimal amounts of artifacts, arguing against their secondary use as trash receptacles in most cases. Large and small pits yielded between zero and 3.168 g of flotation-derived microdebitage per liter (mean = 0.280 g/liter for large pits, 0.294 g/liter for small pits). Features 15 (small pit), 17 (small pit), and 20 (large pit) were the only features that yielded high artifact counts during screening (n = 58, 107, 272, respectively), suggesting a later, secondary refuse function. Densities (by weight) of fire-cracked rock varied greatly among pit features, suggesting differential cleaning or dumping behavior during use and abandonment. Pit features contained FCR densities between 10.25 g/dm3 and 952.52 g/dm3 (mean = 93.33 g/dm3 for large pits and 157.19 g/dm3 for small

Late Archaic–Terminal Archaic Dynamics, Davisson Farm Site

49

9/21/08 2:29 PM

Fig. 3.4. Representative plan views and profiles of selected features at Davisson Farm.

Otto CH3:Layout 1 Page 50

Otto CH3:Layout 1

9/21/08

2:29 PM

Page 51

Fig. 3.5. Scatter plot diagrams showing feature volume, depth, and length measurements.

Otto CH3:Layout 1

9/21/08

2:29 PM

Page 52

pits). In other words, several features were densely packed with FCR while others contained almost none. Initially, the investigators believed that standardized measures of feature content held promise for providing clues as to pit feature function. Basinshaped pits (large and small) containing high FCR densities may represent intact cooking features (features whose cooked contents were never retrieved by the site occupants). If true, there would be a strong positive correlation between FCR weight and food-item density within features. A nonparametric correlation test (Kendall’s Tau-b) was calculated for paired variables, including FCR versus nutshell weights ( b = 0.199; p = .138) and FCR weight versus seed count densities ( b = 0.136; p = .352) (see Purtill 2002c, 118–21). The resulting weak correlation between FCR weight and food-item density fails to support the hypothesis that FCR-filled features represented intact facilities. These tests do indicate, however, that the features were used, abandoned, and reused frequently. Importantly, with the possible exception of feature 111— which contained an unusually high seed count, principally purslane (Portulaca spp.) (n = 240)—none of the pit features appear to have retained their cooked contents in toto, as has been described at other Late Archaic sites (e.g., Vickery 1976). There was some variation in the richness of recovered botanical species between pit feature types (see Subsistence discussion below for more detail). Typically, small pits contained a greater variety of plant species and higher seed counts than large pits (0.128 versus 0.047 g/liter, respectively). Both feature types contained hickory, black walnut, and acorn nutshell. Seed remains from large pits included bean family (Fabaceae) seeds and grape family (Vitaceae) seeds. Small pits contained a wider variety of seed types including Cheno-Am seeds,2 chenopod (Chenopodium spp.) seeds, bedstraw (Galium spp.) seeds, grass family (Poaceae) seeds, knotweed (Polygonum spp.) seeds, mustard (Brassica spp.) seeds, pokeweed (Phytolacca americana) seeds, purslane (Portulaca spp.) seeds, grape family seeds, prickly sida (Sida spinosa) seeds, squash rind (Cucurbita cf. pepo), and several unidentified seed fragments. Interestingly, subpits identified within large pits contained higher densities of seeds than large pits without subpits, suggesting different functions for various feature components. Finally, several features contained small, calcined bone fragments, indicating some animal processing (e.g., cooking, roasting), or perhaps refuse dumping. The lack of additional faunal remains is likely the result of poor preservation instead of an initial absence.

52

M. P. Purtill

Otto CH3:Layout 1

9/21/08

2:29 PM

Page 53

Based on a range of temporal indicators from Davisson Farm and other regional sites (e.g., Purtill 2003), both large and small pits were constructed for long spans of time between Late Archaic and Late Prehistoric times. Late Archaic and Terminal Archaic 14C dates were obtained from features 7, 8, 9, 26, 80, 95, and 111 (table 3.1, fig. 3.3). In addition, features 15, 16a, 20, 111, and 112 contained diagnostic Late Archaic or Terminal Archaic projectile points.

FCR Clusters FCR clusters (n = 13) are defined as a tight clustering of FCR pieces at the Ap/Bw soil interface that were not associated with a “visible” pit. Such clusters are highly variable in size and shape. An abundance of botanical remains occurred in two soil samples taken from FCR clusters. Included within these samples were acorn, walnut, and hickory nutshell, squash rind (Cucurbita cf. pepo), a knotweed seed, a chenopod seed, a Cheno-Am seed, and a grape family seed. Although FCR clusters may represent surface roasting hearths, it seems more likely that they resulted from pit cleaning, that is, that their dumped contents were redeposited from other features. No diagnostic artifacts—and only minimal amounts of debitage—were recovered from these features.

Postmolds Only 12 postmolds were identified during excavations. Most were directly associated with pit features, suggesting a related use such as a drying/cooking rack, spit, and firewall that may also have served as a windbreak. Although no unambiguous house patterns were noted, an alignment of three charcoal-rich posts (features 30, 105, 106) provides the site’s best structural candidate (fig. 3.2). The post pattern describes a slight arc approximately 10 m long. There were no additional pits in the interior of the arc. If the arc of postmolds does represent a structure, it was apparently an open-walled shelter or domicile similar to ones described at other Late Archaic sites such as Kentucky’s Grayson site (Ledbetter and O’Steen 1991, 192), the Maple Creek and Dravo sites near Cincinnati, Ohio (Vickery 1980, 31, 38), and the Mabel Hall site (33LE97), also in Lawrence County, Ohio (OAI on file at Ohio Historic Preservation Office, Columbus). Assuming the site was used as a residential base camp, the absence of further evidence for habitation structures is puzzling. All identified postmolds

Late Archaic–Terminal Archaic Dynamics, Davisson Farm Site

53

Otto CH3:Layout 1

9/21/08

2:29 PM

Page 54

contained substantial charcoal staining or FCR (or both), thereby creating a highly visible archaeological phenomenon. It seems likely that additional posts were present at the site, but their presence went undetected due to poor preservation and absence of observable artifact classes (e.g., FCR). Absence of deep postmolds, wall trenches, shallow semisubterranean house basins, and similar evidence for substantial structures designed for year-round or year-to-year habitations (or both) strengthens the interpretation that the site was seasonally reoccupied. There is an attendant likelihood that site occupants erected ephemeral domestic structures that have escaped archaeological detection.

Small Basin-Shaped Pit/Postmold This is a provisional category used to accommodate two features that could not be classified as either small pits or postmolds. The volume of both features was small (4.62, 2.43 dm3), arguing against their use as cooking pits. However, length₁ and length₂ measurements were nearly double those of postmolds. Only sparse artifactual debris (FCR and one piece of debitage) and no plant remains were recovered during excavation.

Surface Hearths/Burned Area Surface hearths or burned areas (n = 2) are burned areas with irregular outlines, no underlying depressions, and only traces of artifactual debris. The lack of intensive burning suggests short-term use. Similar feature types have been identified as surface hearths or surface fires in previous literature (e.g., Broyles 1971, 18–24).

A RTIFAC TS A total of 14,989 prehistoric artifacts, including debitage, tools, and FCR, was recovered from Davisson Farm during investigations. Excluding diagnostic artifacts from other time periods reduces the total to 13,774. Although multicomponent, the abundance of Late Archaic through Terminal Archaic diagnostic artifacts, as well as 14C assays, suggest that the main occupation of the site dates to this span of time. Tables 3.3 and 3.4 provide a summary of the assemblage; for representative artifact photographs, see figures 3.6 through 3.9.

54

M. P. Purtill

Otto CH3:Layout 1

9/21/08

2:29 PM

Page 55

Table 3.3. Davisson Farm chipped-stone assemblage Chert raw material type*

BC PA KB TIC LPC OFR UM

Z

WY OC UNID TOTAL

Projectile points Merom cluster Brewerton Eared Triangle Brewerton Eared-Notched Brewerton Side-Notched Trimble Side-Notched Late Archaic Stemmed cluster Matanzas Side-Notched Matanzas cluster Little Bear Creek Feeheley Gary Contracting Stemmed Unidentified point fragments

1 2 3 — — — — — — — — 4

— 1 3 1 1 — 1 — 1 — — 6

— — — — — — — — — — — 1

— — 1 — — — — — — — — 2

— — — — — — — — — — — —

— — 1 — — — — — — — — 2

— — — — — — — — — — — —

— — — — — — — — — — — —

— — — — — — — — — — — —

— — — — — — — — — — — 1

— — 8 1 2 2 6 1 — 1 1 30

1 3 16 2 3 2 7 1 1 1 1 46

Bifaces Preform (stage 1–2) Preform (stage 3–4) Drills (tapering bit/unspecified) Hafted scrapers

11 7 — 1

8 9 1 2

10 — — —

6 4 1 3

8 2 — —

— — — —

1 — 1 —

— — — —

— — — —

1 — — —

24 16 2 4

69 38 5 10

2 1 2 1

3 — 3 2

2 — — —

— — — —

5 — — —

1 — — —

— — — —

— — — —

— — — —

— — — —

6 — 3 —

19 1 8 3

7 2 2 20 — — 1 13 62 120 204 124 — — — 5

3 — 12 —

— — 2 —

— — — —

— — — —

1 — 1 —

4 8 870 —

47 24 1,464 6

34 82 36 4 1 20

— 7 14 29 1 17

1 4 7 14 1 6

— — — — — 1

— — 2 — 1 2

— 2 3 2 2 3

6 138 284 212 31 269

60 502 878 573 98 742

—

—

—

—

—

—

2

7

502 547 460 698 354

87

37

1

5

16

1,931

4,638

Unifaces End scrapers Gravers Side scrapers Steep-edge end scraper Debris Cores Checked pebbles Shatter/blocky fragments Primary decortication flakes Secondary decortication Flakes Primary flakes Secondary flakes Shaping/finishing flakes Thinning flakes Flake fragments Miscellaneous debitage TOTAL

8 2 69 1

5 2 87 42 140 126 54 125 12 22 89 117 —

2

2 10 43 97 97 169 65 68 20 7 95 123 3

—

*BC = Brush Creek; PA = Paoli; KB = Kanawha Black; TIC = type I chert; LPC = local pebble chert; OFR = Ohio Flint Ridge; UM = Upper Mercer; Z = Zaleski; WY = Wyandotte; OC = other chert; UNID = unidentified chert

Otto CH3:Layout 1

9/21/08

2:29 PM

Page 56

Thirty-nine projectile points (79% of the total) are attributed to the Late Archaic or Terminal Archaic periods. It is likely that most of the 46 others that could not be typed also represent Late Archaic or Terminal Archaic forms. Davisson Farm further is characterized by a high frequency of Brewerton series points, especially the Brewerton Eared-Notched type (Ritchie 1971, 17). Several other researchers have noted the occurrence of Brewerton series points in the middle Ohio Valley (e.g., McKenzie 1967; Blank 1970; Geistweit 1970). Their presence has led many archaeologists to characterize much of Late Archaic manifestation in south-central and southeastern Ohio as Laurentian-like, although this term is not widely accepted (e.g., Murphy 1975, 79–96). Review of the literature for several area Late Archaic and Terminal Archaic sites reveals considerable intersite variability of specific point types, or combinations thereof, within the region (Purtill 2001). For example, unlike Davisson Farm, the contemporary Grayson site, 40 km south of Davisson Farm in Kentucky, has been defined as a Maple Creek phase site based on the dominance of Merom Expanding Stemmed and Trimble Side-Notched points (n = 33; 57%) and an almost total absence of Brewerton series points (Ledbetter and O’Steen 1992, 87–102). Ongoing academic and contract research in the region is demonstrating that the dominance of Brewerton Eared-Notched points seen at Davisson Farm, and at immediately adjacent sites, is not a trend witnessed elsewhere in the area. Neither several CRM projects throughout Lawrence and Scioto counties in Ohio (e.g., Purtill 2002a; 2002b; 2002d; 2003) nor my inspection of amateur artifact collections from the area have revealed high frequencies of Brewerton Ear-Notched points outside the Davisson Farm area. Although it is tempting to suggest that such trends reflect the presence of multiple ethnic groups operating within the Ohio Valley at this time, the true nature of this phenomenon is currently unknown. At the very least, it seems obvious that multidirectional, perhaps even competing, spheres of cultural influence were acting on resident Late Archaic and Terminal Archaic populations. The remainder of the assemblage consisted of a wide range of unifaces, bifaces, debitage, FCR, and groundstone tools. This assemblage suggests a wide range of tasks performed by the site occupants, including generalized domestic, fabricating/processing, and hunting activities. The bifaces appear to represent more preforms than finished implements. Bifaces tended to have thick cross sections, sinuous lateral margins, and no macroscopic signs of use. In general, the majority of the bifaces correspond with Callahan’s (1979) production stages 1 through 3. Importantly, none of them appear to represent projectile point preforms for large, Late Archaic stemmed varieties such as the Buffalo 56

M. P. Purtill

Otto CH3:Layout 1

9/21/08

2:29 PM

Page 57

Fig. 3.6. Selected Brewerton Eared-Notched and Eared Triangle projectile points recovered from Davisson Farm: (a–h, j–k) Brewerton Eared-Notched; (i) Brewerton Eared Triangle. (Photo courtesy of Gray & Pape, Inc.)

Fig. 3.7. Merom Cluster projectile points from Davisson Farm: (a–c) Trimble SideNotched; (d) Merom Expanding Stemmed. (Photo courtesy of Gray & Pape, Inc.)

Otto CH3:Layout 1

9/21/08

2:29 PM

Page 58

Fig. 3.8. Hematite plummets from Davisson Farm. (Photo courtesy of Gray & Pape, Inc.)

Fig. 3.9. Selected ground- and pecked-stone artifacts from Davisson Farm: (a, c) pitted stones; (b) celt; (d) groundstone fragment; (e) bit end of groundstone adze. (Photo courtesy of Gray & Pape, Inc.)

Otto CH3:Layout 1

9/21/08

2:29 PM

Page 59

Straight Stemmed type noted at several nearby sites (e.g., Buffalo site [Broyles 1976]). The presence of debitage, cores, and biface preforms indicate that all stages of chipped-stone tool reduction occurred on-site. Hafted scrapers also were identified at the site. Although several researchers contend that, in many cases, these artifacts represent purposely manufactured tool types (e.g., Blank 1970; Kent Vickery, pers. comm.), the majority, if not all, of the Davisson Farm specimens likely represent reworked projectile point basal fragments. Based on basal characteristics, several hafted scrapers were made from Big Sandy (n = 1), Raddatz (n = 1), Saratoga Cluster (after Justice 1987) (n = 1), Brewerton Eared-Notched (n = 3), and Matanzas Side-Notched (n = 2) varieties. Only Early, Middle, and Late Archaic point types are represented among the hafted scrapers, suggesting that such reworking did not extend much, if at all, beyond Late Archaic times at this site and its vicinity. Use of raw materials focused primarily on locally available resources such as Brush Creek, Type I chert,3 and local pebble chert4 (table 3.3). Kanawha Black and Paoli chert varieties, both of which crop out south of the site, also were recovered in high frequencies. Although cobbles of both Kanawha Black and Paoli chert could have been transported to the local area by river currents, significant spatial variation between these types and locally available pebble chert (see below), in combination with the presence of soft (i.e., unweathered) cortex on several pieces of Paoli, suggest procurement from primary outcrops. Finally, only a small percentage of Kentucky Flint Ridge, Wyandotte, Ohio Flint Ridge, Upper Mercer, Zaleski, and Delaware chert was recovered, suggesting limited exploitation of these nonlocal materials. Although present, ground- and pecked-stone tools (n = 21) were not abundant (table 3.4). Two hematite plummets were recovered during surface collections (fig. 3.8). One complete specimen that was highly polished and “egg-shaped” measured 39.4 mm by 24.3 mm and weighed 50 g (fig. 3.8b). A single shallow groove had been incised around the midsection. Several potlidlike scars on this specimen possibly represent fire damage or natural exfoliation (e.g., freeze-thaw). The second plummet is a fragmented midsection of cruder workmanship or perhaps was unfinished (fig. 3.8a). Although similar in design to the complete one, the sides of this plummet are flat instead of rounded and display striations suggesting that the artifact had been worked with a coarse material (e.g., sandstone). A poorly executed groove encircles the artifact’s midsection. Due to the plummet’s fragmentary nature, only the width, 30.4 mm, could be measured.

Late Archaic–Terminal Archaic Dynamics, Davisson Farm Site

59

Otto CH3:Layout 1

9/21/08

2:29 PM

Page 60

Table 3.4. Davisson Farm groundstone assemblage Raw material type*

HEM

GR

Q/QZ

GS

SA

CH

UNID

Total

Celts

1

—

—

—

—

—

2

3

Hammerstones

—

—

1

—

1

3

3

8

Pitted stone

—

—

1

—

3

—

1

5

Plummets

2

—

—

—

—

—

—

2

Tapered adze

—

1

—

—

—

—

—

1

Misc. groundstone

1

—

—

—

—

1

—

2

TOTAL

4

1

2

—

4

4

6

21

*HEM = hematite; GR = granite; Q/QZ = quartz/quartzite; GS = greenstone; SA = sandstone; CH = chert; UNID = unidentified material.

The remainder of the ground- and pecked-stone tool assemblage consists of hammerstones (n = 7), pitted stones (n = 5), celts (n = 4), a tapered adze, and two miscellaneous groundstone artifact fragments (fig. 3.9). Three among this assemblage, two celts and one pitted stone, were derive from suspected Late Archaic or Terminal Archaic features. The groundstone and pecked-stone tools suggest a range of activities, including woodworking/cutting (celts, adze), chipped-stone reduction (hammerstones, possible pitted stones), and plant processing (pitted stones, possibly hammerstones). The scarcity of groundstone nut-processing implements is not commensurate with the presumed importance of this resource in the diet of site inhabitants. Perhaps some, or most, nut processing occurred off-site at extractive loci near nut-bearing trees. The scarcity of such implements has been noted at other Late Archaic sites in the middle Ohio Valley (e.g., Vickery 1980). Also occurring in less-thanexpected numbers were bone tools, more likely due to poor preservation rather than their original absence.

Microscopic Use-Wear Analysis Twenty-three chipped-stone tools were selected for microscopic use-wear analysis by Richard Yerkes of the Ohio State University (Yerkes 2002) to determine tool use patterns, site function, occupational duration, and to provide a better understanding of resource procurement strategies. The methodology employed by Yerkes is largely based on one outlined by Semenov (1964) and

60

M. P. Purtill

Otto CH3:Layout 1

9/21/08

2:29 PM

Page 61

modified by Keeley (1980) (see Purtill 2002c for a more detailed discussion on methodological protocol). Yerkes used both low-power (6x and 50x) and highpower (50x and 1,500x) stereoscopic approaches. Of the 23 artifacts studied, nine (39%) presented no evidence of use. However, the remaining artifacts show evidence of a range of tasks mostly associated with butchering and hide working. Materials worked include hide/meat (five occurrences), bone/meat (one), dry hide (one), hide? (one), bone/hide (one), and bone/antler (one). Yerkes notes that tool use appears relatively unintensive, suggesting that artifacts were used expediently, with a low incidence of curation. No evidence for woodworking or plant processing was indicated. All analyzed point types have fractures or use consistent with performance as projectiles. These artifacts do not appear to have been used as knives, drills, or perforators, as suggested for Merom Expanding Stemmed and Trimble Side-Notched points at other Late Archaic sites (Vickery 1980, 27). Two projectile points examined by Yerkes had been reworked to serve as hide scrapers, one of which is a fragmented Matanzas Side-Notched point. For this artifact, Yerkes indicates that the absence of basal use-wear (e.g., grinding, microchipping, etc.) suggests a secure haft. He further points out that the hafted scrapers in this assemblage were “broken or damaged projectile points that were recycled, rather than scrapers that were manufactured from biface blanks (2002, 2).” Yerkes indicates that at least one artifact—a recycled Brewerton Side-Notched point—lacked microscopic evidence consistent with hafting, suggesting that the tool was used by hand. Also of note is a Brewerton Eared-Notched point that exhibited traces of use-wear along its base resulting from contact with bone, perhaps suggesting attachment of the point to a bone shaft or foreshaft. No other points analyzed exhibited similar traces.

Perishable Artifacts Surprisingly, flotation of soil from feature 80, which yielded a radiocarbon age of 2790 +/-70 bp (840 bc) (table 3.1, fig. 3.3), contained three fragments of charred twine—an extremely rare find outside dry rockshelter/cave or submerged environs. Two of the fragments are rather small and represent only a single strand of double-ply twine. The third fragment is 1 cm long with four Z-twists per cm. The twist angle was approximately 25 to 45 degrees, consistent with category 3 (see Drooker 1992, fig. 11). The diameter of this third fragment of twine was 3 mm. Organic material used for the twine was not identified,

Late Archaic–Terminal Archaic Dynamics, Davisson Farm Site

61

Otto CH3:Layout 1

9/21/08

2:29 PM

Page 62

although its preserved condition suggests that it could be. It is uncertain if these pieces derive from a larger item, such as a basket or bag, or if they represented twine used to secure some item.

S PATIAL PATTERN IN G OF A R T I FACT S Surface collecting revealed high densities of all categories of artifacts and debris along the southernmost T2-T3 terrace edge (see Purtill 2002c). Tools were especially concentrated east of E4680, where a high density of tools (0.31/10m2) was noted, in contrast to areas west of E4680 (0.18). Importantly, little evidence exists for the spatial segregation of function-specific activity loci (e.g., chippedstone tool production, butchering/hide working, etc.), as one would expect for a planned settlement like a village. Instead, observed distributions more likely resulted from a series of discrete, but slightly overlapping, occupations through time. In order to better investigate the nature of intrasite patterning, a subsample of 1,000 pieces of debitage surface collected from five discrete areas (groups 1 through 5) was examined (see Figure 3.2 for group location). Each group covered 2,500 m2 and was distributed to meet three basic criteria: 1) to provide an equal representation of areas along the intensively occupied T2-T3 terrace edge; 2) to test areas containing dense feature clustering that may represent individual occupations; and, 3) to provide a large enough subsample of debitage artifacts (> 100) from each group for comparative statistical analyses. Originally a sixth group was defined to test a feature cluster centered on feature 26 (between N0930–0970 and E4590–4630; see fig. 3.2), however, low debitage densities (n = 18) made direct intergroup comparisons unwarranted and the group was removed from these analyses. First, a Pearson’s chi-square test (SPSS Base 8.0) was calculated in order to characterize chert type composition by group and to investigate possible variations in the distribution of different chert varieties across the site. Chert selected for this test included Brush Creek, Paoli, local pebble chert, Type I, and Kanawha Black. Because of low frequencies, Wyandotte, Ohio Flint Ridge, and Upper Mercer chert were excluded from this analysis in an effort to avoid skewing the results (D. Thomas 1986, 298). In general, the analysis revealed significant intergroup variation in chert use (χ 2 = 94.475, df = 24, p < .001).

62

M. P. Purtill

Otto CH3:Layout 1

9/21/08

2:29 PM

Page 63

Cross-tabulation results between group and chert type provide further insight into the nature of the intergroup variation. Low residual scores5 for Brush Creek, local pebble chert, and Type I chert demonstrate no statistical differences in the utilization of these materials for each group. In contrast, high residual scores for Kanawha Black and Paoli varieties suggest that these materials account for the majority of the intergroup variance noted in the initial chi-square test. Specifically, group 1 is characterized by high relative variability in both Kanawha Black, which is greatly overrepresented (residual = 7.7), and Paoli chert, which is underrepresented (residual = -3.6). In contrast, only group 4 is characterized by a heavy underrepresentation of Kanawha Black (residual = -5.0). To further test the impact of Kanawha Black and Paoli chert on overall group variance, the initial chi-square test was recalculated, this time excluding Kanawha Black and Paoli. The second test (χ 2 = 15.225, df = 12, p < .229) indicates that when Kanawha Black and Paoli chert types are removed from the sample, the remaining chert types are equally distributed across all groups. This trend suggests that all site occupants exploited locally available chert (local pebble chert, Type I, Brush Creek) in similar ways through time. As noted initially, high variance of Kanawha Black and Paoli chert appears to be restricted to groups 1 and 4 (see fig. 3.2). Low residual scores (between +2.0 and -2.0) for these varieties in the remaining areas (groups 2, 3, 5) suggest similar use patterns. A third chi-square test was calculated that excluded groups 1 and 4 to see if significant differences existed between the remaining groups when all chert varieties were included. The results demonstrate that even with Kanawha Black and Paoli included, no significant differences between groups 2, 3, and 5 regarding chert type usage are evident (χ 2 = 10.924, df = 10, p < .363). These tests indicate that site occupants used locally available chert sources in similar ways across the site. In site areas represented by groups 2, 3, and 5, nonlocal (but easily accessible) Kanawha Black and Paoli cherts were similarly used. Variation in the presence of Kanawha Black and Paoli in groups 1 and 4, however, indicates that patterns of exploitation of raw chert were not universal across the site. An overrepresentation of Kanawha Black in western site areas reflects heavier use of this low-grade material than in other site areas. Although a substantial Middle Woodland occupation also existed in the western area (Purtill 2002c), Hopewell populations’ well-documented penchant for highquality chert varieties (e.g., Ohio Flint Ridge) argues against associating

Late Archaic–Terminal Archaic Dynamics, Davisson Farm Site

63

Otto CH3:Layout 1

9/21/08

2:29 PM

Page 64

Kanawha Black’s substantial presence with such a late occupation. Instead, I interpret the difference in raw materials as resulting from diachronic changes in chert exploitation patterns within the Late Archaic and Terminal Archaic periods. Because other data sets suggest that Davisson Farm represents a repeatedly occupied seasonal camp for several thousand years during Late Archaic through Terminal Archaic times, variation in chert resource procurement areas (e.g., a shift from north to south), the breakdown or establishment of regional intergroup relations, new technological needs, or perhaps other explanations could account for such shifts in chert exploitation strategies.

S UBSISTENCE Subsistence data come almost exclusively from feature-derived paleoethnobotanical remains. Poor preservation, presumably due to highly acidic soils (McCleary and Hamilton 1998), restricted recovered bone to “bits” too fragmented to identify. A total of 205 liters of soil was extracted for flotation separation and analysis from suspected Late Archaic (111 liters) and Terminal Archaic (94 liters) features. A Flote-Tech flotation machine was used for soil separation and analysis of the recovered botanical remains was undertaken at Gray & Pape’s Cincinnati laboratory under the direction of Michele Williams (for a detailed discussion of methodology, see Purtill 2002c). The results of this analysis are summarized in tables 3.5 through 3.8. The high level of nutshell ubiquity in both Late Archaic (100%) and Terminal Archaic (90%) features attests to the dietary importance of nuts (table 3.5). The recovery of black walnut (11.47 g), hickory (6.23 g), and acorn (0.001 g) nutshell6 suggests that site occupants exploited both upland forests and valleys (Gordon 1969; also see below). When considered by temporal period, Late Archaic features contained nearly equal amounts of hickory (an average of 0.39 g per 10 liters floated) and black walnut (0.36 g/10 l), with only minor amounts of acorn (