PENNSYLVANIA ARCHAEOLOGICAL DATA SYNTHESIS: The Raccoon Creek Watershed (Watershed D of the Ohio River Subbasin 20) Bridge Replacement Project T-319 Beaver County Bridge No. 36 (Links Bridge) Independence Township, Beaver County, PA Prepared for Pennsylvania Department of Transportation Engineering District 11-0 Submitted by GAI Consultants, Inc. 570 Beatty Road Monroeville, PA 15146-1300 GAI Project No. 2002-441-10 DECEMBER 2003
PENNSYLVANIA ARCHAEOLOGICAL DATA SYNTHESIS: The Raccoon Creek Watershed (Watershed D of the Ohio River Subbasin 20) Bridge Replacement Project T-319 Beaver County Bridge No. 36 (Links Bridge) Independence Township, Beaver County, PA Prepared for Pennsylvania Department of Transportation Engineering District 11-0 Written by Douglas H. MacDonald, Ph.D., RPA Lead Archaeologist With Contributions by Jonathan C. Lothrop, Ph.D, RPA David L. Cremeens, Ph.D., C.P.S.S. GAI Consultants, Inc. 570 Beatty Road Monroeville, PA 15146-1300 GAI Project No. 2002-441-10 DECEMBER 2003
TABLE OF CONTENTS List of Figures...ii List of Photographs... v Acknowledgements...vii CHAPTER I INTRODUCTION AND PROJECT SUMMARY... 1 A. INTRODUCTION...1 B. PROJECT SETTING...1 C. PREHISTORY OF THE RACCOON CREEK WATERSHED: AN INTRODUCTION...3 CHAPTER II PROJECT LOCATION AND ENVIRONMENTAL SETTING... 5 A. DRAINAGE...5 B. PHYSIOGRAPHY AND LANDFORMS (BY D.L. CREMEENS)...6 C. GEOLOGY (BY D.L. CREMEENS)...8 D. PREHISTORIC LITHIC SOURCES...9 E. PALEOENVIRONMENTS...12 CHAPTER III BACKGROUND AND KEY PROJECTS... 15 A. PASS FILES DATA...15 B. KEY PROJECTS...15 C. SUMMARY AND INTRODUCTION TO PREHISTORY CHAPTERS...19 CHAPTER IV PALEOINDIAN PERIOD... 21 A. PALEOINDIAN OVERVIEW...21 B. PALEOINDIAN MATERIAL CULTURE, CHRONOLOGY, AND SUBSISTENCE...21 C. PALEOINDIAN SITES...25 D. PALEOINDIAN SETTLEMENT PATTERNS AND LITHIC RAW MATERIAL USE...31 E. RESEARCH ISSUES...32 CHAPTER V EARLY ARCHAIC PERIOD... 35 A. EARLY ARCHAIC MATERIAL CULTURE AND CHRONOLOGY...35 B. EARLY ARCHAIC ARCHAEOLOGICAL SITES...36 C. EARLY ARCHAIC SETTLEMENT PATTERNS AND LITHIC RAW MATERIAL USE...39 D. EARLY ARCHAIC: CONCLUSION AND RESEARCH QUESTIONS...39 CHAPTER VI MIDDLE ARCHAIC PERIOD... 41 A. MIDDLE ARCHAIC OVERVIEW...41 B. MIDDLE ARCHAIC MATERIAL CULTURE AND CHRONOLOGY...41 C. MIDDLE ARCHAIC ARCHAEOLOGICAL SITE TYPES AND LOCATION TRENDS...42 D. MIDDLE ARCHAIC SETTLEMENT PATTERNS AND LITHIC RAW MATERIAL USE...45 E. MIDDLE ARCHAIC: CONCLUSION AND RESEARCH QUESTIONS...47 PA Prehistoric Data Synthesis, Raccoon Creek Watershed i
CHAPTER VII LATE ARCHAIC PERIOD... 49 A. LATE ARCHAIC OVERVIEW...49 B. LATE ARCHAIC MATERIAL CULTURE, SUBSISTENCE, AND CHRONOLOGY...49 C. LATE ARCHAIC SITE TYPES AND LOCATION TRENDS...53 D. LATE ARCHAIC SETTLEMENT PATTERNS AND LITHIC RAW MATERIAL USE...64 E. LATE ARCHAIC: CONCLUSION AND RESEARCH QUESTIONS...66 CHAPTER VIII EARLY WOODLAND PERIOD... 69 A. EARLY WOODLAND OVERVIEW...69 B. EARLY WOODLAND MATERIAL CULTURE AND CHRONOLOGY...69 C. EARLY WOODLAND, SUBSISTENCE, SITE TYPES, AND LOCATION TRENDS...73 D. EARLY WOODLAND SETTLEMENT PATTERNS AND LITHIC RAW MATERIAL USE...83 E. EARLY WOODLAND: CONCLUSION AND RESEARCH QUESTIONS...84 CHAPTER IX MIDDLE WOODLAND PERIOD... 87 A. MIDDLE WOODLAND OVERVIEW...87 B. MIDDLE WOODLAND MATERIAL CULTURE AND CHRONOLOGY...87 C. MIDDLE WOODLAND SITE TYPES, LOCATIONS, AND SETTLEMENT...89 D. MIDDLE WOODLAND LITHIC RAW MATERIAL USE...93 E. MIDDLE WOODLAND: OVERVIEW AND RESEARCH ISSUES...93 CHAPTER X LATE WOODLAND PERIOD... 95 A. INTRODUCTION...95 B. LATE WOODLAND MATERIAL CULTURE, CHRONOLOGY, AND KEY REGIONAL SITES...95 C. LATE WOODLAND SITES IN WATERSHED D...98 D. LATE WOODLAND LITHICS AND SETTLEMENT...100 E. LATE WOODLAND: SUMMARY AND RESEARCH QUESTIONS...102 CHAPTER XI SUMMARY AND CONCLUSION... 104 A. DEMOGRAPHY AND SETTLEMENT...104 B. LITHIC RAW MATERIAL USE AND SETTLEMENT PATTERNS...106 C. CONCLUSION: FUTURE AVENUES OF RESEARCH IN WATERSHED D, SUBBASIN 20...106 REFERENCES... 109 LIST OF FIGURES Figure 1. West-Central Pennsylvania (gray) and Raccoon Creek...1 Figure 2. Map of the Raccoon Creek Watershed and Vicinity...2 Figure 3. Geomorphological Cross-Section of Raccoon Creek at Links Bridge, Independence Township...7 Figure Figure 4. Lithic Raw Materials in Southwestern Pennsylvania...11 5. Map of Key Archaeological Sites & Projects in the Study Area and Vicinity...17 Figure 6. Miller Type Point from Stratum IIa. Point is Actual Size (from Boldurian 1985:284)....23 Figure Figure 7. Comparison of the Miller Type Point from Meadowcroft (far right; Boldurian 1985:284) with Steubenville Points from the East Steubenville Site (46Br31). All Points at Same Scale. Steubenville Point images courtesy of the West Virginia Division of Highways....23 8. Map of Paleoindian Sites...26
Figure 9. Possible Clovis Point Preform Fragment from Site 36Wh1312 (Scale: 1 in. equals 2.5 cm; 2000a)....29 Figure 10. Early Archaic Kirk Corner-Notched (left) and Stemmed (right) Points, Actual Size (from Michels and Smith 1967:683; East and Beckman 1992:46)...35 Figure 11. Changing Site Counts in the Early Holocene: Comparison of Site Counts (top) and Site Density per Decade (bottom), Watershed D (PASS files)...41 Figure 12. Middle Archaic Bifurcate Point, Actual Size (from Custer et al. 1996:31)...42 Figure 13. Early Woodland/Adena Artifacts (from Dragoo 1963:179)...70 Figure 14. Location of Early Woodland Sites Discussed in Text....72 Figure 15. Profiles of Cresap Mound (from Dragoo 1963:21)....76 Figure 16. Profile and Planview of McKees Rocks Mound (from McMichael 1956:148)...78 Figure 17. Comparison of Late Archaic and Early Woodland Artifact Counts, Sites in Cross Creek Drainage (Vento and Donohue 125-127)....83 Figure 18. Comparison of Late Archaic and Early Woodland Lithic Raw Material Use at Meadowcroft Rockshelter (top) and Mungai Farm (bottom) (after Vento and Donahue (1982:124-125)...86 Figure 19. Raccoon Notched Points (from Lantz 1989:9, 11)...87 Figure 20. Late Woodland Levanna Triangle Point (from Michels and Smith 1967:669)....96 Figure 21. Quantity of Late Woodland Radiocarbon Dates Over Time in the Upper Ohio Valley (data from Nass and Hart 2000:132-133)...97 Figure 22. Site Densities per Period from the End of the Late Archaic to the Late Woodland, Watershed D...98 Figure 23. Site Counts per Decade Per Period, Watershed D, Subbasin 20....105 Figure 24. Variation in Site Location, Watershed D, Subbasin 20....105 LIST OF TABLES Table 1. Archaeological Sites by Time Period in Watershed D, Subbasin 20....3 Table 2. Lithic Raw Materials in Southwestern Pennsylvania....12 Table 3. Comparison of Raw PASS File Data with Checked PASS File Data in Watershed D, Subbasin 20....15 Table 4. Paleoindian Sites, Watershed D (PASS Files)...25 Table 5. Paleoindian Site Location Data, Watershed D (PASS Files)...27 Table 6. Table 7. Table 8. Paleoindian Site Location Data, Watershed D (PASS files)...27 Site 36Wh1312: Projectile Point Data...30 Lithic Raw Material Use at Paleoindian Sites in the Raccoon Creek Watershed D and Vicinity....31 Table 9. Site 36Wh1312: Cross-Tabulation of Chert Type by Artifact Type....32 Table 10. Early Archaic Sites, Watershed D (PASS Files)...37 Table 11. Early Archaic Site Location Data, Raccoon Creek Watershed (PASS Files)...37 Table 12. Early Archaic Site Location Data, Raccoon Creek Watershed (PASS Files)....38 Table 13. Lithic Raw Material Use at Single Component Early Archaic Sites, Watershed D (PASS Files)...39 PA Prehistoric Data Synthesis, Raccoon Creek Watershed iii
Table 14. Cross-Tabulation of Cross Creek Site by Lithic Raw Material Type for Early Archaic Artifacts (Vento and Donohue 1982:124-127)....39 Table 15. Middle Archaic Sites, Watershed D (PASS Files)....43 Table 16. Middle Archaic Site Location Data, Watershed D (PASS Files)...44 Table 17. Middle Archaic Site Location Data, Watershed D (PASS Files)...46 Table 18. Table 19. Table 20. Cross-Tabulation of Single Component Middle Archaic Sites and Lithic Raw Materials, Watershed D (PASS Files)...47 Cross-Tabulation of Cross Creek Site by Lithic Raw Material Type for Middle Archaic Artifacts (Vento and Donohue 1982:124-127)....47 Late Archaic Sites, Watershed D (PASS Files)...54 Table 21. Cross-Tabulation of Site Setting by Nearest Water, Watershed D Sites (PASS data)....55 Table 22. Table 23. Table 24. Table 25. Table 26. Table 27. Cross-Tabulation of Site Setting by Diagnostic Artifact Type at Late Archaic Sites, Watershed D Sites (PASS data)...56 Cross-Tabulation of Nearest Water by Type of Diagnostic Artifact at Late Archaic Sites, Watershed D (PASS Files)...56 Late Archaic Site Location Data, Watershed D (PASS Files)...56 Late Archaic Site Location Data, Watershed D (PASS Files)...57 Cross-Tabulation of Site by Lithic Raw Material (Presence or Absence) at Single Component Late Archaic Sites (PASS Files; X=present; --=absent)....64 Cross-Tabulation of Cross Creek Site by Lithic Raw Material Type for Late Archaic Artifacts (Vento and Donohue 1982:124-127)....65 Table 28. Early Woodland Sites, Watershed D, Subbasin 20 (PASS Files)...81 Table 29. Early Woodland Site Location Data, Watershed D, Subbasin 20 (PASS Files)....81 Table 30. Table 31. Table 32. Table 33. Table 34. Early Woodland Site Location Data, Watershed D (PASS Files)...82 Cross-Tabulation of Cross Creek Site by Lithic Raw Material for Early Woodland Artifacts...84 Middle Woodland Sites, Watershed D (PASS Files)...89 Middle Woodland Site Location Data, Watershed D (PASS Files)...90 Middle Woodland Site Location Data, Watershed D (PASS Files)...91 Table 35. Cross-Tabulation of Middle Woodland Site by Lithic Raw Material (Percent), Watershed D....93 Table 36. Late Woodland Sites, Raccoon Creek Watershed (PASS Files)...99 Table 37. Late Woodland Site Location Data, Watershed D, Subbasin 20...101 Table 38. Late Woodland Site Location Data, PASS Files (Watershed D)....101 Table 39. Cross-Tabulation of Late Woodland Site by Lithic Raw Material (Percent), Watershed D...102 Table 40. Site Location Data, All Periods, Watershed D, Subbasin 20....105 Table 41. Lithic Raw Material Use over Time, Watershed D, Subbasin 20....106
LIST OF PHOTOGRAPHS Photograph 1. Traverse Creek within Raccoon Creek State Park. View East....1 Photograph 2. Photograph 4. Photograph 5. The Confluence of Raccoon Creek and the Ohio River. View South across the Ohio River from Route 68...5 Cross Creek near Site 36Wh298, Cross Creek Village, in the Southern Portion of Watershed D. View Southeast...6 Typical Upland Setting in Project Area. View South across Mungai Farm in the Southern Portion of Watershed D...6 Photograph 6. The Raccoon Creek Valley near Links Bridge, Independence Township. View South...6 Photograph 7. Photograph 8. Photograph 9. Photograph 10. Photograph 11. Photograph 12. Blades found in Association with the Miller Point, Stratum IIa, Meadowcroft Rockshelter (http://people.delphiforums.com/mcconaughy /meadowcroft/.htm)....24 View of Meadowcroft Rockshelter from above Cross Creek. View Northeast (photo courtesy of J. Herbstritt and J. Adovasio)....28 Setting of Site 36Wh1312 within the Wolf Fun Valley. View North...30 Brewerton Points (Left; from York County, Pa); and Steubenville Points (Right; from East Steubenville Site, 46Br31). Steubenville Point Photo courtesy of West Virginia Division of Highways...49 Aerial Photograph of the Leetsdale Site. View Northeast (www.lrp.usace.army.mil/.htm). Photograph courtesy of U.S. Army Corps of Engineers...61 Aerial Photograph of the East Steubenville Site. View North. Courtesy of West Virginia Division of Highways...62 Photograph 13. View of McKees Rocks Mound Prior to the 1896 Excavations (from Dragoo 1963:153)...78 Photograph 14. Photograph 15. Excavator holding Stone Bowl/Mortar from Leetsdale Site (www.lrp.usace.army.mil/.htm; Photo Courtesy of the U.S. Corps of Engineers....79 View Northwest toward the Georgetown Site (36Bv29) and the Ohio River...82 Photograph 16. View of Hill in the Former Location of Avella Mound. Excavations at the Site Removed Most of the Mound which was on top of the Hill. View Northwest...92 PA Prehistoric Data Synthesis, Raccoon Creek Watershed v
(This page intentionally blank)
ACKNOWLEDGEMENTS The Scope of Work for this data synthesis was coordinated by Patrick Roberts (Environmental Manager) and Joseph Verbka (Qualified Professional Archaeologist) of PennDOT District 11-0, Chan Funk of the Pennsylvania Historic and Museum Commission (PHMC) Bureau for Historic Preservation (BHP), and GAI s Ben Resnick, M.A., RPA, Jonathan C. Lothrop, Ph.D., RPA and Douglas H. MacDonald, Ph.D., RPA. Ben Resnick, M.A., RPA, was project manager, while Douglas H. MacDonald, Ph.D., RPA, was Principal Investigator and author of the technical report. Jonathan C. Lothrop contributed sections of this report as well. Lisa Dugas conducted background research at the State Museum in Harrisburg. David L. Cremeens, Ph.D., CPSS, conducted the geomorphological study for the Links Bridge project (MacDonald 2003b) which is included in Chapter II of this report. GAI would like to express its gratitude to the following individuals at the PHMC-BHP in Harrisburg who facilitated completion of the project in one way or another: Chan Funk, Pete Van Rossum, and Kurt Carr. Each of these individuals went out of their way to provide assistance during the course of this project. In addition, James Adovasio of Mercyhurst College and Conrad Weiser of the United States Army Corps of Engineers, Pittsburgh Division, provided access to artifacts, reports, and unpublished documents from the Meadowcroft and Leetsdale projects, respectively. PA Prehistoric Data Synthesis, Raccoon Creek Watershed vii
(This page intentionally blank)
CHAPTER I INTRODUCTION AND PROJECT SUMMARY A. INTRODUCTION This document is a prehistoric archaeological data synthesis for the Raccoon Creek Watershed (Watershed D) of the Ohio River Subbasin 20, prepared by GAI Consultants, Inc. (GAI) for PennDOT District 11-0 (Figures 1 and 2). For comparative purposes, the report also includes research report information from areas peripheral to Watershed D, including the Upper Ohio Valley and southwestern Pennsylvania, in general. The report should prove useful for archaeologists or laypersons interested in the Native American prehistory of western Pennsylvania and the Upper Ohio River Valley. This document also provides contextual information and research questions for reviewers of archaeological projects at PennDOT and the Pennsylvania Historical and Museum Commission s (PHMC) Bureau for Historic Preservation (BHP) for assessing the (National Register) significance of identified sites. Figure 1. West-Central Pennsylvania (gray) and Raccoon Creek (dot). (http://www.dcnr.state.pa.us/st ateparks/parks/racc.htm) B. PROJECT SETTING The Ohio River (PA Subbasin 20) encompasses 3,084 sq. miles of total drainage area in Pennsylvania and includes hundreds of watersheds along its 981 mile-long trajectory between Pittsburgh, Pennsylvania, and Cairo, Illinois. The Raccoon Creek watershed (Photograph 1) (Watershed D of Ohio River Subbasin 20) comprises 327 square miles within the Pittsburgh Low Plateau Section of the unglaciated Appalachian Plateaus Province, including northern Washington County, southern Beaver County, and western Allegheny County. Raccoon Creek begins near Burgettstown, in northern Washington County, and gently meanders northward to its confluence with the Ohio River northwest of Pittsburgh near the Beaver and Allegheny County line (Figure 2). Traverse Creek (see Photograph 1), Little Traverse Creek, and Service Creek are its major tributaries in the watershed. South of Raccoon Creek, the watersheds of Harmon Creek and Cross Creek drain westward into the Ohio River, but are included in Watershed D. Photograph 1. Traverse Creek within Raccoon Creek State Park. View East. As noted above, the PHMC-BHP and the Pennsylvania Department of Environmental Protection (2002) also include Cross Creek and Harmon Creek within the Raccoon Creek Watershed D, even though both of these creeks flow westward directly into the Ohio River; they are not feeder streams of Raccoon Creek. Nevertheless, to be consistent with the Pennsylvania Archaeological Site Survey (PASS) files (administered by the PHMC-BHP), we also include Cross Creek and Harmon Creek in Watershed D. The implications of this inclusion are great, as extensive research has been conducted within Cross Creek by the University of Pittsburgh at sites such as Meadowcroft Rockshelter (Carlisle and Adovasio 1982), providing a wealth of data by which to assess the history of the region. PA Prehistoric Data Synthesis, Raccoon Creek Watershed 1
MAP OF WATERSHEDS D, E, AND F SUBBASIN 20 Figure 2. Map of the Raccoon Creek Watershed and Vicinity gaic0nsultants, INC. DRAWN DHM APPROVED jcl DATE 7/10/03 DWG. NO 20 02-441-10--C-A2 A Shenango B Beaver D Raccoon E Buffalo/Wheeling Ohio River Subbasin 20 Ohio River Subbasin 20 C Slippery Rock G Ohio F Chartiers Shaded Area See Map at Right Ohio River Cross Creek Buffalo Creek WEST VIRGINIA To Ohio River D Beaver County Harmon Cr. PENNSYLVANIA Raccoon Creek E Wheeling Creek Allegheny County Washington County Pittsburgh Ohio River Chartiers Creek Scale F 0 5 10 miles Pennsylvania Figure 2 Map of the Raccoon Creek Watershed and Vicinity 2 PA Prehistoric Data Synthesis, Raccoon Creek Watershed
C. PREHISTORY OF THE RACCOON CREEK WATERSHED: AN INTRODUCTION The cultural historical chronology used in this report differs from the traditional Upper Ohio Valley sequence first defined by Mayer-Oakes (1955); namely, we do not include a Late Prehistoric period in the sequence, as is common in southwestern Pennsylvania and West Virginia. Instead, we use the traditional Pennsylvania chronology as established by Raber (1985) in the Comprehensive State Plan for the Conservation of Archaeological Resources. In this sequence, the Upper Ohio Valley Late Prehistoric Period is subsumed in the larger Late Woodland period. In so doing, the report maintains consistency with the chronology used in the Pennsylvania Archaeological Site Survey (PASS) files. For at least the last 12,000 years, Native Americans have occupied Watershed D of the Ohio River Subbasin. More than 200 prehistoric site components from seven different time periods have been identified at archaeological sites in Watershed D (Table 1). Prehistoric occupation was initiated during the Paleoindian period, at the end of the Pleistocene epoch, when small bands of Native Americans migrated into the previously uninhabited region. Meadowcroft Rockshelter (36Wh297), located in the Cross Creek drainage of Watershed D, is one of the most important Paleoindian archaeological sites in North America and may date to as early as 14,500 years ago. Table 1. Archaeological Sites by Time Period in Watershed D, Subbasin 20. PERIOD TIMEFRAME (YEARS BP) SITE COMPONENTS* IN WATERSHED D Paleoindian 16,000-10,000 8 Early Archaic 10,000-8000 18 Middle Archaic 8000-5300 45 Late Archaic 5300-3000 69 Early Woodland 3000-2100 22 Middle Woodland 2100-1200 23 Late Woodland 1200-400 30 TOTAL -- 215 *No. of Sites from PASS data with components for respective periods, based on type of diagnostic artifacts at sites. Subsequent to the occupations at Meadowcroft, Native Americans of the region used beautifully crafted fluted spear points and atlatls (spearthrowers) and traveled over a wide area. This mobile forager lifestyle continued through the end of the Early Archaic period until approximately 8,800 or so years ago. At this time, Native American populations steadily increased and there was a change in subsistence and settlement pattern. In contrast to the Paleoindian period, Native Americans of the Early and Middle Archaic (ca. 8,000 years ago) occasionally lived in base camps on terraces of rivers for a few months of the year. From these base camps, Native Americans traveled to uplands and low-mid-order tributaries to collect seasonally available resources. Travel in the local area increased at this time, as reflected by the use a wide variety of cherts from southwestern Pennsylvania, including Ten Mile, Uniontown and Monongahela cherts. The change in subsistence pattern led to a dramatic population increase during the initial portion of the Late Archaic period, approximately 5,200-4,000 years ago. Late Archaic Native Americans utilized Brewerton side- and corner-notched projectile points and traveled extensively within western Pennsylvania rivers and stream corridors. By the end of the Late Archaic period (ca. 4,000 to 3,000 years ago), evidence from Panhandle Archaic/Steubenville sites along the Ohio River and Raccoon Creek indicate increased nucleation of populations and increased riverine subsistence. In other parts of Pennsylvania, this transition to increased sedentism and riverine adaptations is called the Transitional or Terminal Archaic. However, typical artifacts of the Transitional Period, including steatite, pottery, horticultural remains, and broadspear projectile points, are not well represented at sites in western Pennsylvania. As such, this report does not include a chapter on the Transitional/Terminal Archaic Period. Instead, the Transitional is included in the Late Archaic. During the Early Woodland period (3,000 to 2,100 years ago), Native Americans of the Adena culture first produced pottery and incorporated small amounts of domesticated PA Prehistoric Data Synthesis, Raccoon Creek Watershed 3
foods into their diets, including corn, beans, squash, and sunflower, as well as a variety of other wild seed crops. Meadowcroft Rockshelter on Cross Creek has yielded some of the earliest evidence of pottery use and plant domestication in the Upper Ohio Valley. Increased sedentism and ceremonialism is a hallmark of the Early Woodland Adena culture, as reflected in the large burial mounds with human burials and wealth items along the Ohio River and its major tributaries. During the Middle Woodland (2100 to 1200 BP/800 A.D.), the Adena culture gave way to the Hopewell, another mound-building culture thought to be either a cultural continuation of Adena or a population replacement. During the Middle Woodland, mound building continued, as did the reliance on agriculture and sedentism. Raccoon Notched projectile points are well represented at sites in the Raccoon Creek drainage and vicinity (Lantz 1989). These points are typically recovered with Watson Farm pottery at sites along the main stem of Raccoon Creek, as well as much of central and western Pennsylvania. By approximately 800 A.D. (1,200 years ago), Late Woodland site counts increased in southwestern Pennsylvania, especially to the south and east of Raccoon Creek in areas occupied by the Late Woodland Monongahela culture (see Figure 1). Some researchers in the Upper Ohio River Valley call this period the Late Prehistoric Period. For all intent and purposes, the Late Woodland discussed in this report and the Late Prehistoric are the same. The Late Woodland population increase was likely spurned by the increased availability of resources due to agriculture and the increasing use of semi-sedentary villages. Populations in the Raccoon Creek watershed, as well as in the nearby Ohio River and Monongahela River watersheds, relied extensively on agriculture and sedentism during the Late Woodland. The Raccoon Creek watershed contains fewer Late Woodland sites than the Monongahela River watershed and the Upper Ohio River in West Virginia. While several sites in Cross Creek yielded Late Woodland components, only a few sites along the Ohio River are true Late Woodland villages within Watershed D. No Late Woodland villages have been excavated along Raccoon Creek, suggesting that it may have been used primarily for hunting and gathering. At villages along the Upper Ohio River, Chartiers Creek and the Monongahela River, Native Americans used domesticated crops and abundant wild resources to subsist, including fish, shellfish, animals, and wild plants. This report provides a comprehensive overview of the prehistory of the Raccoon Creek watershed and is organized into 12 chapters, including four background/summary chapters and seven chapters that summarize prehistory by time period. Chapter II describes the project setting and provides information on the landscape and resources available to prehistoric Native Americans. A comprehensive overview of paleoenvironments and lithic raw materials is also provided in Chapter II. Chapter III provides an overview of methods utilized during the collection of data for the report and summarizes the key archaeological projects and sites in Watershed D of Subbasin 20. Chapters IV through X provide overviews of the Paleoindian, Early Archaic, Middle Archaic, Late Archaic, Early Woodland, Middle Woodland, and Late Woodland periods. The final chapter provides a summary overview of major cultural and demographic trends over time in the study area. The ultimate goal of this report is to provide a context for future research in the region. By using the information gathered in this report, future archaeologists will hopefully be able to better understand the important research issues and cultural historical milestones of the last 12,000 years of Native American lifeways in the study area. 4 PA Prehistoric Data Synthesis, Raccoon Creek Watershed
CHAPTER II PROJECT LOCATION AND ENVIRONMENTAL SETTING This chapter provides an overview of the natural environment of the Raccoon Creek watershed and vicinity, including descriptions of the region s physiography, drainage, soils, bedrock, floral and faunal resources, as well as past and present environments. In addition, this chapter includes a detailed overview of lithic raw materials available to prehistoric Native Americans in the sub-basin. Analysis of the variety of cherts and other stones used by Native Americans provides insight into prehistoric use of the landscape. 10 km northwest of Canonsburg in northern Washington County. Raccoon Creek meanders northerly to its confluence with the westwardflowing Ohio River, 10 km northwest of Aliquippa in central Beaver County (see Photograph 2). A. DRAINAGE Ohio River Subbasin 20 begins in Pittsburgh, Pennsylvania, as a result of the confluence of the Monongahela and Allegheny Rivers; it subsequently flows northwesterly past its confluence with the Beaver River north of Aliquippa and past the mouth of Raccoon Creek just east of Shippingport. Subbasin 20 encompasses seven watersheds in Pennsylvania (Watersheds A-G; DEP 2002): A) Shenango River; B) Beaver River; C) Slippery Rock Creek; D) Raccoon Creek; E) Wheeling and Buffalo Creeks; F) Chartiers Creek; and G) Upper Ohio River. The current project area encompasses Watershed D Raccoon Creek and Cross Creeks, among others but will also discuss pertinent archaeological sites from peripheral areas as well. Watershed D of the Ohio River Subbasin 20 encompasses a total drainage area of 327 sq. miles. The northern border of the watershed is the Ohio River between the mouth of Raccoon Creek (Photograph 2.) and the Ohio border (see Figure 2). The Ohio and West Virginia state lines form the western watershed boundary. The Cross Creek watershed is the most southern in Watershed D, while Raccoon Creek forms the watershed s eastern border. Raccoon Creek is the main stream within Watershed D; it forms at the confluence of Burgetts Fork and Cherry Run in Burgettstown, Photograph 2. The Confluence of Raccoon Creek and the Ohio River. View South across the Ohio River from Route 68. Several large tributaries, including Wingfield Run, Traverse Creek, Little Traverse Creek and Service Creek make it a relatively high-order stream by the time it reaches the Ohio River, approximately half-way between Shippingport and Beaver, Pennsylvania. The majority of the land located south and west of the Ohio River in Beaver County drains into Raccoon Creek. Raccoon Creek State Park in the center of the watershed is one of the largest state parks in Pennsylvania, encompassing a total acreage of 7,323 acres, including the 100-acre Raccoon Creek Lake, formed by the damming of Traverse Creek (see Photograph 1). The other major drainage in Watershed D is Cross Creek (Photograph 3), which forms the southern border of the watershed and flows westerly for ca. 32 km to the Ohio River (Carlisle et al. 1982:6). Cross Creek s major tributaries include the North, Middle, and South Forks of Cross Creek. Each of these creeks has its headwaters at springheads in the uplands of PA Prehistoric Data Synthesis, Raccoon Creek Watershed 5
northern Washington County. Photograph 4 shows a typical upland setting at Mungai Farm on a ridge divide between the Cross Creek and Raccoon Creek watersheds southwest of Burgettstown. rolling interfluves separated by relatively narrow, steep-walled, moderately incised valleys (Figure 3; Briggs 1999). L a n d f o r m s As depicted in Figure 3, the general landform shape throughout the Raccoon Creek Valley is that of a maturely dissected plateau with relatively broad ridgetops (see Photograph 4) and narrow, deep valleys (Photograph 5). In general, slopes are relatively steep to very steep along the lower portions of the valley walls, and then become more moderate closer to the ridgetops (Ellyson et al. 1974). Photograph 3. Cross Creek near Site 36Wh298, Cross Creek Village, in the Southern Portion of Watershed D. View Southeast B. PHYSIOGRAPHY AND LANDFORMS (BY D.L. CREMEENS) Physiography Subbasin 20 occurs within the Appalachian Plateaus physiographic province, specifically within the Unglaciated Appalachian Plateaus section. The Appalachian Plateaus province is characterized by relatively flat-lying, predominantly clastic rocks that are higher in elevation, and younger in age than surrounding provinces (Thornbury 1965). The plateau is bounded on all sides by outward facing escarpments, subtle to the south and west, and prominent to the north and east (Allegheny Front). The boundary or division between the Glaciated Appalachian Plateau and the Unglaciated Appalachian Plateau sections is the Late Wisconsin glacial boundary, approximately 20 km north of the Ohio River. The study area is entirely within the unglaciated Pittsburgh Low Plateau section. Within Pennsylvania, the Pittsburgh Low Plateau section is the lowest in the Appalachian Plateau province, and has a relatively narrow range of depth of dissection. The upland surface of the prototypical area forms a true plateau of broad, Photograph 4. Typical Upland Setting in Project Area. View South across Mungai Farm in the Southern Portion of Watershed D. Photograph 5. The Raccoon Creek Valley near Links Bridge, Independence Township. View South. 6 PA Prehistoric Data Synthesis, Raccoon Creek Watershed
Scale 0 0.5 1.0 miles Figure 3. Geomorphological Cross-Section of Raccoon Creek at Links Bridge, Independence Township West WEST-EAST CROSS-SECTION, RACCOON CREEK Upland Residuum East Pennsylvanian Bedrock Early Pleistocene Terrace Strath Historic Floodplain/Low Terrace Pennsylvanian Bedrock Historic/Late Holocene Alluvial Fan Natural Levee Raccoon Creek Silty Late Holocene/Historic Alluvium Natural Levee Abandoned Channel Late Holocene Low Terrace Silty Clay Late Holocene Alluvium Not to Scale Figure 3 Geomorphological Cross-Section of Raccoon Creek at Links Bridge, Independence Township PA Prehistoric Data Synthesis, Raccoon Creek Watershed 7
With the exception of the Ohio River and a few of its large tributaries, the valley floors of most streams are very narrow with little to no flood plain development. The Raccoon Creek valley is approximately 150 to 300 meters wide with a relatively level valley floor easily distinguished from the valley wall (see Figure 3; see Photograph 5). Valley walls are near-vertical-to -vertical bedrock outcrops to the uplands or to terrace straths. Relief in the valley ranges from 91-107m from the upland ridgetops to the valley floor. The terrace straths are 9 to 27m above the valley floor, and approximately 60m below the ridgetops. Within the valley floor, Raccoon Creek meanders broadly across the valley floor with large cutbanks into the uplands and straths. Tributaries enter the Raccoon Creek valley floor in the form of alluvial fans. Many of these fans have pushed Raccoon Creek from one side of the valley to the other. The overall drainage pattern of Raccoon Creek with its tributaries is dendritic. Several large rectangular bends or meanders of the creek are present upstream and downstream of the project area (see Photograph 5). These large rectangular bends are associated with well-developed terrace straths. Independence Township in Beaver County has no documented glacial sediments, including outwash, associated with any of the Pleistocene glaciations. However, there are erosional landforms (terrace straths) associated with Pleistocene stream downcutting (see Figure 3). The uplands consist predominantly of residual Gilpin Series soils (Smith 1982). The terrace straths are mapped with the Monongahela, Tyler and Allegheny soils, all with well-developed profiles containing argillic Bt horizons and fragipan Bx horizons. Colluvial soils, represented by the Ernest Series, are mapped at the base of some slopes. The valley floor of Raccoon Creek is mapped with the weakly developed Pope, Philo and Atkins alluvial soils. C. GEOLOGY (by D.L. Cremeens) The rocks within the southern portion of Beaver County belong to the Middle Pennsylvanian Allegheny Group, as well as the Upper Pennsylvanian Conemaugh and Monongahela Groups. In general, the Middle Pennsylvanian represents upper delta plain facies, while the Upper Pennsylvanian Conemaugh Group and Monongahela Group represent a change back to lower delta facies (Wagner et al. 1970). The Monongahela Group, divided into the Pittsburgh and Uniontown Formations, is a sedimentary sequence that as a whole is dominated by limestones and dolomitic limestones, calcareous mudstones, shales, and thin-bedded siltstones and laminites (thin-bedded rocks). Rocks of the Monongahela Group are only present at the surface as isolated knobs or hills in adjacent south Hopewell Township, and in southern Hanover Township near Frankfort Springs. The Pittsburgh coal was stripmined out of the base of the Monongahela Group rocks in Hanover Township; the resultant strip-mined lands are contained in Raccoon Creek State Park and in Hillman State Park in Washington and Beaver Counties. The Conemaugh Group, which covers much of the study area, is divided into two formations, the stratigraphically higher Casselman Formation and the Glenshaw Formation. The Ames limestone separates the two formations. The Glenshaw Formation is distinguished by having several widespread marine units, both limestone and shale, in the stratigraphic succession (Edmunds et al. 1999). The Casselman Formation consists of shales, fluvial sandstones, marine limestones and thin coals formed in a lower delta plain environment (Vento 2001). The bedrock units at the bridge crossing have been ascribed to the Birmingham shale member of the Casselman Formation (Vento 2001). The only exposure of Allegheny Group rocks in Independence Township is in the northeast corner, north of the town of Independence, where the Allegheny Group is exposed in the Raccoon Creek valley. Structurally, the Pittsburgh Low Plateau prototypical area has a very gentle southwest dip into the Dunkard Basin, overprinted by gentle, northeast-trending folds that decrease in amplitude northwestward (Briggs 1999). In the pre-glacial (pre-pleistocene) landscape the major steams generally drained northwest to the Erie basin. Based on work done by Leverette 8 PA Prehistoric Data Synthesis, Raccoon Creek Watershed
(1934), most geologists believe that the master stream for the southwestern Pennsylvania region was the pre-glacial Monongahela River (Wagner et al. 1970). The pre-glacial Monongahela River headed in central West Virginia, just as the modern stream does, and flowed north to present day Pittsburgh (Harper 1997; Wagner et al. 1970; Leverette 1934). From Pittsburgh, the river flowed northwest to near Beaver, Pennsylvania, along a course now followed by the modern Ohio River downstream from Pittsburgh. At Beaver, the pre-glacial Monongahela River continued to flow northward toward New Castle, Pennsylvania. North of New Castle, the river turned westward and flowed into Ohio near Youngstown where it finally turned northward again and flowed into the Erie Basin. The north-flowing Monongahela system existed for a long enough period of time that broad meanders developed in the streams in its drainage basin, and the streams were widening rather than deepening their valleys (Wagner et al. 1970). Sometime in the early Pleistocene, continental glaciers moved into northwest Pennsylvania and effectively blocked the northflowing streams. The blockage resulted in ponding, and in the establishment of a new drainage outlet to the south. The modern Ohio River was formed taking with it the drainage basin of the pre-glacial Monongahela River at Beaver, Pennsylvania. Work by Jacobson et al. (1988) place the timeframe for the flow direction reversal sometime between the early Pleistocene magnetic reversal at 772,000 years BP and the Illinoisan Stage (302,000 to 132,000 years BP). The new south-flowing Ohio system rapidly downcut the valley floor. The pre- or early- Pleistocene drainage is preserved in the modern Monongahela, Allegheny, and Ohio basins in a series of abandoned stream channels and cutoff meander bends (Kaktins and Delano 1999). These geomorphic features are partially preserved as a series of relatively highelevation, flat-topped terraces, often as great as 90-95m above the present stream levels. These remnant valleys are known as the Parker Strath, from a typical terrace near Parker, Pennsylvania, and from the Scottish word strath meaning a wide flat valley (Wagner et al. 1970; Kaktins and Delano 1999). A terrace strath implies a bedrock-cored valley floor, elevated and dissected, that may be mantled by alluvium or other deposits. A map by Marine (1997) shows the extent of glacial Lake Monongahela, based on elevation, as extending upstream in the Raccoon Creek Valley. This suggests that some of the terrace straths may have a mantle of fine-grained lacustrine deposits. The new base level of the reversed, south-flowing Ohio drainage, and the increased drainage area for the basin resulted in a deep stage erosion event that isolated the Parker Strath as much as 15-20m above the riverbed. The erosion event is believed to have occurred in the late portions of the Illinoisan Stage or in the Sangamon Interglacial Stage (approximately 132,000 to 79,000 years ago). Erosion and downcutting continued throughout the Wisconsin Stage until the Parker Strath was isolated as much as 90m above modern stream levels. Vento (2001) indicated that the terrace straths, described as a broad structural bench at a nominal elevation of 900 feet above mean sea level and 20-30m above the active stream channel (Raccoon Creek), likely correspond to the Parker/ Carmichaels strath terrace. D. PREHISTORIC LITHIC SOURCES A variety of primary (bedrock) and secondary (stream cobble) lithic toolstone sources were available to prehistoric Native American groups inhabiting the Raccoon Creek watershed and vicinity (Figure 4). Most primary sources of lithic raw materials were available in the southern portion of Watershed D, while secondary sources were available in the north. As such, Native Americans in the Raccoon Creek Valley likely curated high-quality stone from other regions and used local glacial cobble chert for expedient daily tasks. Regional sources of chert are described below. Western Pennsylvania Uniontown chert or flint occurs in Late Paleozoic Uniontown formation limestone PA Prehistoric Data Synthesis, Raccoon Creek Watershed 9
deposits of southwestern Pennsylvania, distributed predominantly southwest of the project area near Canonsburg. Uniontown chert is the most prominent primary source chert in the Ohio River subbasin. Uniontown chert cobbles were recovered by Eisert (1974) in both Chartiers Creek and Little Chartiers Creek near Canonsburg in Watershed F of Subbasin 20. Uniontown is light-gray to light olive-gray (5Y6/1), very pale orange (10YR8/2), and yellowish-gray (5Y8/1) of medium to fine grain. The chert occurs in nodules ranging from 2.5 to 30.5cm in diameter (Eisert 1974). As with the Uniontown chert described above, Monongahela chert outcrops within the Uniontown member of the Pennsylvanian Monongahela Group. Vento and Donohue (1982:119) note its presence along Cross Creek within the current study area, while Eisert (1974) observed outcrops near Chartiers and Little Chartiers Creeks near Canonsburg. Monongahela chert is described as being dark to light gray with limonite staining on weathered surfaces and bedding planes (Vento and Donohue 1982: 120). The current author also observed Monongahela chert near the headwaters of Wolf Run, a tributary of Buffalo Creek, northwest of Washington, Pennsylvania in Watershed E. The chert was of a fairly low quality within limestone nodules eroding on the ground surface of a steep sideslope. Ten Mile chert occurs as thin lenses within the lower member of the Greene Formation and the upper member of the Washington Formation along the Ten Mile Creek drainage, Green County, southwestern Pennsylvania. Outcrops of Ten Mile chert have been observed along the Cross Creek drainage in the current study area. Ten Mile chert has dark, carbonaceous laminations within a dark grayish-brown, olivebrown or grayish-blue matrix. Weathering causes a patina rind of pale greenish-gray to grayish-white (Vento and Donahue 1982). the Racoon Creek study area, primary outcrops of Onondaga chert occur within the Lower- Middle Devonian Onondaga Formation across the glaciated region of New York, and discontinuously in the Ridge and Valley of Pennsylvania, Maryland, and West Virginia (Luedtke 1992:129). Secondary cobbles of Onondaga chert are distributed much more widely, occurring across northwest Pennsylvania within the boundaries of the terminal Pleistocene glaciation, and was locally available to Native Americans in the northern portions of the Raccoon Creek drainage (Holland 1999). Scattered cobbles of Onondaga chert were likely distributed as secondary deposits within the Raccoon Creek and Upper Ohio River valleys. Gull River (Yellow Onondaga/Huronian) Chert is most typically dusky yellow (5Y 6/4) to moderate yellow (5Y 7/6) with an opaque, coarse grain. Gull River is also known as Huronian, or Yellow Onondaga (Fogelman 1983), and derives from Ordovician deposits in Ontario, Canada. Gull River was distributed widely as secondary cobbles across western New York, northeastern Ohio, and western Pennsylvania, within the boundaries of the terminal Pleistocene glacial advance (Holland 1999). Gull River chert is common at archaeological sites in western Pennsylvania. Scattered cobbles of Gull River chert are likely present as secondary cobbles within the Raccoon Creek Valley. Sewickley Chert is found within the Pittsburgh Formation of the Pennsylvanian Monongahela Group in southwestern Pennsylvania. This chert is gray to very dark-gray moderate quality chert with occasional freshwater ostracod fossils. Cortex typically is gray to tan (Vento and Donahue 1982). Onondaga Chert--a high-quality chert of dark to bluish-gray with microfossil striations of lighter colors (Vento and Donahue 1982). The stone has a medium texture, a shiny luster, and is moderately translucent. To the north and east of 10 PA Prehistoric Data Synthesis, Raccoon Creek Watershed
Figure 4. Lithic Raw Materials in Southwestern Pennsylvania. PA Prehistoric Data Synthesis, Raccoon Creek Watershed 11
Table 2. Lithic Raw Materials in Southwestern Pennsylvania. MATERIAL AGE GENERAL LOCATION REFERENCE Uniontown Chert Pennsylvanian Canonsburg, SW PA Eisert 1974 Monongahela Chert Pennsylvanian Canonsburg, SW PA Eisert 1974; Vento and Donohue 1982 10-Mile Chert Permian Southwest PA Holland 1999 Onondaga Chert Pleistocene cobbles Northwest PA Vento and Donohue 1982 Gull River Chert Pleistocene cobbles Northwest PA Holland 1999 Sewickley Chert Pennsylvanian Southwest PA Vento and Donohue 1982 Brush Creek Chert Pennsylvanian SE Ohio/NE Kent/SW PA Gatus 1985;Olafson 1964 Flint Ridge Chert Mid-Pennsylvanian East-central Ohio Lepper et al. 2001 Upper Mercer Chert Lower Pennsylvanian East-central Ohio Luedtke 1992 Kanawha Chert Penns/Kanawha F. Central WV; SW Pa? Reger 1921; Reppert 1978 Northern West Virginia and Eastern Ohio Brush Creek Chert, or Hughes River chert, is a high-quality chert from the Brush Creek limestone member of the Pennsylvanian Conemaugh Group. Its primary source areas are southeastern Ohio, northeastern Kentucky, northwestern West Virginia, and far southwestern Pennsylvania (Gatus 1985; Olafson 1964; Vento and Donahue 1982). Brush Creek occurs within limestone as nodules and lenses up to 60cm (2 feet) in diameter. It is medium to fine grained with a translucency of 0.5mm (Luedtke 1992:118; Shott 1990; Vickery 1996: 74). Brush Creek is typically gray-brown (N3/ to N4/) to tan (10YR 5/1) with fine texture and moderate translucency. Kanawha chert is bluish-black to black to light gray and is of medium to fine grain (Reger 1921: 226-240; Reppert 1978). This chert is a member of the Pennsylvanian Kanawha Formation (see Figure 4) and occurs in a basin approximately 42 km (70 miles) by 64 km (40 miles) in parts of Boone, Kanawha, Clay, Nicholas, Webster, and Fayette Counties, West Virginia (Reger 1921:227; Reppert 1978:3). Reppert (1978:4) identifies three Kanawha chert facies in the primary source area, with the quality of stone declining within secondary alluvial settings and as one progresses eastward from the central primary source basin. Flint Ridge/Vanport Formation Chert occurs within the Vanport Formation of the Middle Pennsylvanian System of east-central Ohio, approximately 100 km (93 miles) to the west of the Links Bridge project area. Flint Ridge is variably homogenous, mottled, laminated, or brecciated and often has small veins of chalcedony or quartz crystals. Color ranges from white to dark-gray, with yellow, pink, red, and blue not uncommon. This stone was widely distributed during prehistory, occurring at sites greater than 200 kilometers (124 miles) distant (Tankersley 1989:269). Upper Mercer/Coshocton Chert derives from the Upper Mercer Limestone member of the Lower Pennsylvanian System of east-central Ohio. Upper Mercer frequently has inclusions of white or blue chalcedony and typically is black to dark-gray (Luedtke 1992:136; Tankersley 1989:289). This stone was distributed widely along the Ohio River and its tributaries, extending well into northwestern Pennsylvania (Lantz 1984; Tankersley 1989:270). Several surface-collected Late Woodland sites in the project vicinity possessed triangle points produced from Upper Mercer/Coshocton chert. E. PALEOENVIRONMENTS The climate of southwestern Pennsylvania is classified as humid continental (Trewartha 1967). Winters are cold and snowy at high elevations, but thaws typically prohibit permanent snow pack in low elevation river valleys. Summers are occasionally warm at high elevations, but are frequently warm in valleys. Mean summer temperature for Beaver County is 20 C (70 F), with a January mean of 1.6 C 12 PA Prehistoric Data Synthesis, Raccoon Creek Watershed
(30 F). The growing season in Beaver County ranges between 119 and 177 days. Annual precipitation averages 98cm (38 inches) per year, and is heaviest between May and August (Smith 1982). Between 21,000 and 17,000 BP, the Laurentide ice sheets reached their maximum extent in eastern North America, with the Kent Moraine of the Erie Lobe extending as far south as central Ohio and the Beaver/Lawrence County line in western Pennsylvania (Crowl and Sevon 1999:226). The southern boundary of the 304- meter- (1,000-foot-) thick ice sheets was located approximately 20 km (12 miles) north of Raccoon Creek s confluence with the Ohio River. Northeastern North America experienced relatively rapid warming between 14,000 and 11,000 years ago (Gates 1993:84; Stingelin 1965). The northward retreat of the glaciers and a steady succession of vegetative types in Greater Appalachia reflect this dynamic transition (Gates 1989; Gaudreau 1988:218; Watts 1979). During the late Glacial period, tundra was widespread at areas adjacent to glaciers and at higher elevations in Appalachia (Larabee 1986; Maxwell and Davis 1972:506; Whitehead 1973:625). By 12,000 BP, paleoenvironmental sites in Pennsylvania revealed pollen assemblages that suggested widespread boreal forest of spruce and pine. Data from New Paris No. 4, Bedford County, Pennsylvania (Guilday et al. 1964) and Hartstown Bog, Crawford County, Pennsylvania (Walker and Hartman 1960), suggest ameliorating boreal forest conditions at approximately 11,000 BP The pollen diagram at Mt. Davis Marsh, near Meyersdale in Somerset County, Pennsylvania, was dominated by spruce, pine, and fir (Stingelin 1965:50). By 11,500 BP, in southwestern Pennsylvania, Meadowcroft Rockshelter (Washington County) vertebrate remains (Adovasio et al. 1998:11) revealed a temperate Carolinian fauna, as well as oak, hickory, and pine, suggesting the initial emergence of the Mixed Mesophytic forest. Subsequently, white ash invaded the area, along with beech and chestnut (Castanea) (Watts 1979:452). Upland sites in northcentral Pennsylvania, such as Tannersville Bog (Monroe County) and Longswamp (Berks County), revealed deciduous forest by 10,000 BP (Davis 1984:172). These data suggest that lowland sites likely have supported a Mixed Mesophytic forest suite for at least 11-12,000 years. As climates ameliorated during the hypsithermal interval, between 10,000 and 5000 BP, cooladapted boreal forest species declined in importance in the upland Appalachian Plateau, including such paleoenvironmental sites as Mt. Davis Marsh, Tannersville Bog, and Potts Mountain Pond (Davis 1984:178; Joyce 1988:197). Glaciers completed their retreat northward to the Arctic and many of the species of flora and fauna of central Appalachia spread rapidly northward to colonize the once-glaciated portions of the Appalachia Plateau (Gates 1989; Gaudreau 1988:218; Watts 1979). Approximately 5,000 years ago, climates began to cool slightly across eastern North America (Gajewski 1988:259), resulting in an increase in red spruce at upland sites such as Tannersville Bog and Cranberry Glades. This transition marked the end of the middle Holocene hypsithermal warm episode (Davis 1984:178). Increases in charcoal are noted in pollen diagrams across eastern North America during the late Holocene (Davis 1984; Fredlund 1989). Between 4,000 to 2,200 years ago at Gallipolis Locks and Dam on the mid-ohio River in Mason County, West Virginia, charcoal flecking increases in sediments, likely due to increased fire disturbance related to human modification of the landscape for agriculture. Between 2200 years ago and the present, the dominating pollen type was Ambrosia, or ragweed, associated with the fluorescence of regional agricultural practices. Pollen from corn and other domesticates begins to appear in the paleoenvironmental record, while other regional sites show dramatic increases in grasses with the rise in agriculture (Davis 1984:178). PA Prehistoric Data Synthesis, Raccoon Creek Watershed 13
(This page intentionally blank) 14 PA Prehistoric Data Synthesis, Raccoon Creek Watershed
CHAPTER III BACKGROUND AND KEY PROJECTS GAI reviewed archaeological and historical reports and publications, as well as PASS files data on recorded sites to develop overviews of the prehistory of the Raccoon Creek Watershed and vicinity. Pertinent archaeological sites in adjacent areas, including the Upper Ohio River Valley, are also included to provide a comparative context. These summaries provide an archeological and historical context for assessing potential site significance and for predicting the locations and types of archaeological sites that might be present in the sub-basin. A. PASS FILES DATA PASS files data were kindly provided by the Bureau for Historic Preservation (BHP) for all archaeological sites in Watershed D. These data proved invaluable in understanding the prehistory of the study area, as more than 450 prehistoric sites were recorded in Watershed D, including information on site type, location, age, lithic raw materials, and artifacts. As reflected in Table 3, 1,019 prehistoric site components were identified at the 492 archaeological sites in the Raccoon Creek watershed. Site forms for many of these sites did not include information regarding diagnostic artifacts to support the component designation. For the purposes of data integrity, GAI utilized only the sub-sample of sites that listed diagnostic artifacts from a specific time period on the archaeological site form. In so doing, the total number of useable components was reduced to 215 at 141 sites in the Raccoon Creek watershed. PASS files data regarding lithic raw material type were useful only for a small sub-sample of the 141 sites with diagnostic artifacts. Within PASS files, lithic raw material use is not segregated by component; thus, for sites with multiple components it is not possible to determine which raw materials were utilized during the respective site occupations. Thus, only the 61 single component sites with diagnostic artifacts are useful in assessing lithic raw material use. These 61 sites are comprised of 2 Paleoindian, 4 Early Archaic, 18 Middle Archaic, 20 Late Archaic, 5 Early Woodland, 6 Middle Woodland, and 6 Late Woodland components. This sub-sample of sites will be used to assess lithic raw material use within the PASS files data. Table 3. Comparison of Raw PASS File Data with Checked PASS File Data in Watershed D, Subbasin 20. PERIOD RAW PASS DATA 1 CHECKED DATA 2 Paleoindian 12 8 Early Archaic 15 18 Middle Archaic 56 45 Late Archaic 76 69 General Archaic 169 -- Transitional 13 -- Early Woodland 55 22 Middle Woodland 61 23 Late Woodland 36 30 General Woodland 96 -- General Prehistoric 430 -- Total Components 1019 215 Total Sites 492 141 1 =raw unchecked data as presented in PAS Chronol table within BHP PASS files; 2 =PASS data checked for diagnostic artifacts B. KEY PROJECTS To supplement PASS files data and collections analysis, GAI reviewed all available archaeological reports from Watershed D to better evaluate the types of sites found in the project area. Forty-three cultural resource management (CRM) projects (with available reports) have been completed within Watershed D, several of which identified archaeological sites with diagnostic artifacts. Nevertheless, these cultural resource management reports consist entirely of Phase I surveys with an occasional project conducting limited Phase II testing. Outside of the Meadowcroft and related Cross Creek studies, only one data recovery excavation (Dravo Site; Davis 1988) has been conducted within Watershed D. The CRM reports provide little if any data regarding specialized research issues, such as Native American subsistence, lithic raw material use, or technology. This lack of data is, in and of itself, PA Prehistoric Data Synthesis, Raccoon Creek Watershed 15
an important finding of this report. The total lack of data recovery investigations within the Raccoon Creek Valley proper is astounding, given the sheer number of previously recorded archaeological sites. Clearly, more work needs to be conducted in the watershed at the data recovery level. Information regarding important sites can be found in regional-scale studies, including: Adovasio s Cross Creek project (Carlisle and Adovasio 1982; Adovasio and Page 2002); Mayer-Oakes (1955) Prehistory of the Upper Ohio River Valley; Dragoo s studies of the Archaic Hunters of the Upper Ohio Valley (1959) and the Adena (Mounds for the Dead, 1963); Lantz s 1989 Raccoon Notched Point study; and Herbstritt s (1981b) archaeological site survey within southwest Pennsylvania. Given the dearth of archaeological research data within the Raccoon Creek Valley proper, research reports from peripheral areas, including the Upper Ohio Valley, Chartiers Creek, and the Monongahela River were incorporated into the study to provide meat to the bones of regional prehistory. Together with data provided in the Cross Creek study, these data provide a means to assess general trends in the prehistory of southwestern Pennsylvania. Key Studies and Sites in the Raccoon Creek Watershed (Watershed D) (Figure 5) Cross Creek Survey Between 1973 and 1982, the University of Pittsburgh s Archaeological Research Program, led by James Adovasio, conducted an in-depth study of archaeological sites within the Cross Creek drainage of the southern portion of Watershed D (see Photograph 3). Cross Creek flows westerly into the Ohio River in northern West Virginia from its headwaters in northern Washington County. Adovasio and his colleagues (Boldurian 1985; Carlisle and Adovasio 1982) conducted an intensive survey of the drainage, identifying 236 prehistoric sites. This catalog of sites provides an outstanding window into prehistoric settlement and site-use patterns in Watershed D. Among the more important sites identified during the Cross Creek Survey is Meadowcroft Rockshelter (36Wh297), which yielded evidence of the earliest human occupation in eastern North America, as well as some of the earliest recorded evidence of plant domestication and pottery in southwestern Pennsylvania. Other sites investigated by the Cross Creek project include the Krajacic Site (36Wh351), Cross Creek Village (36Wh298), Avella Mound (36Wh415), and Mungai Farm (36Wh106). These sites provide a wealth of data to assess the entire span of Native American occupation of the study area. Georgetown Site (36Bv29) Mayer-Oakes (1955:178-184) and Davis (1988, n.d.; Davis and Lantz 1987) report the Georgetown Site (36Bv29) as being located on the south side of the Ohio River in the northwesternmost corner of the Raccoon Creek watershed. The site yielded an array of pottery styles spanning the Woodland Period, including Early Woodland Half-Moon Cordmarked, Middle Woodland Watson Cordmarked, and Late Woodland Monongahela Plain, from discrete strata with radiocarbon dates. Mayer- Oakes (1955) used this ceramic assemblage to help refine Woodland pottery types in the Upper Ohio River Valley (Johnson 1977:61). The site is near a recently built, man-made lake and industrial development. Based on a recent field view, intact portions of the landscape are present, but much of the area nearby the site has been destroyed; it is uncertain whether intact portions of the site are present. Lower Field/Shippingport Site (36Bv4) Another Woodland village site investigated by Mayer-Oakes (1955), the Lower Field/ Shippingport site was located on the floodplain of the Ohio River. The Shippingport Nuclear Power Plant was built in the former site location. Davis (1988) reports that the skull of one of the Late Woodland burials at the site yielded a Fort Ancient projectile point. Additional details regarding this site are provided by Mayer-Oakes (1955). 16 PA Prehistoric Data Synthesis, Raccoon Creek Watershed
Figure 5. Map of Key Archaeological Sites & Projects in the Study Area and Vicinity. gaic0nsultants, INC. DRAWN DHM APPROVED jcl DATE 7/10/03 DWG. NO 20 02-441-10--C-A5 Key Sites in and Near Watersheds D, E, and F Map Key Sites outside of Watershed D Sites in Watershed D Watershed Boundary River/Stream not in Watershed D County Line River/Stream in Watershed D Ohio River Subbasin 20 Grave Creek Mound Cresap Mound Ohio River Pennsylvania Watson Farm Wheeling Creek Georgetown/Dravo Ohio River Buffalo Creek WEST VIRGINIA To Ohio River D Beaver County Harmon Cr. East Steubenville Avella Mound Meadowcroft Cross Creek Village PENNSYLVANIA Lower Field/Shippingport Crivallero Raccoon Creek Cross Creek Outdoor Theater Krajacic E Royal Tartan Allegheny County Washington County Mungai Farm Leetsdale Ohio River Chartiers Creek Scale F MayviewBend/ Mayview Depot Kelso Wylie 36Wh924 McKees Rocks Morganza Village Pittsburgh Monongahela River Crawford Grist #2 0 5 10 miles Peters Creek M o u n d Crall M o u n d Figure 5 Map of Key Archaeological Sites and Projects in the Study Area and Vicinity PA Prehistoric Data Synthesis, Raccoon Creek Watershed 17
Outdoor Theater Site (Raccoon Notched Point Study) Raccoon Notched points were first identified at the Outdoor Theater Site (36Bv24) on a terrace of Raccoon Creek near Aliquippa. Lantz (1989) provides a detailed description of the site and its lithic assemblage, as well as artifact studies from several other sites within the Raccoon Creek Valley. In addition to Mayer-Oakes (1955) seminal study, Lantz s synthesis on Raccoon Notched points provides the bulk of Middle Woodland data for the study area. The Outdoor Theater Site was destroyed by industrial and transportation development in Aliquippa. Crivallero Site (36Bv122) Excavated by Louis Berger and Associates, Inc. (1998), this site yielded evidence of Late Archaic and Early Woodland occupations in an alluvial setting near the confluence of Traverse Creek and Raccoon Creek. The site was interpreted to be a short-term lithic scatter and did not proceed beyond the Phase I/II level. These excavations for PennDOT yielded pertinent information regarding lithic raw material use during the mid-late Holocene. Pertinent Sites in Nearby Areas Leetsdale Site (36AL480) The Leetsdale archaeological project was funded by the United States Army Corps of Engineers (USACOE) as a mitigation of adverse effects to Site 36AL480, the construction area for the new Braddock Dam (Davis 2001; Fenicle 2003; Hardlines 2001; Schuldenrein et al. 2003; Vento et al. 2002). Site 36AL480 is located on the northern shore of the Ohio River, downstream approximately five miles west of the three rivers confluence in Pittsburgh. This site is located in Watershed G of Subbasin 20, less than 1 mile north and east of the Watershed D boundary. As the crow flies, the Leetsdale Site is less than five miles east of Raccoon Creek as it flows near Aliquippa. Excavations at Site 36AL480 were conducted between 2001 and 2003 by several archaeological consultants working for the USACOE. The Leetsdale Site yielded multiple stratified components that provide information regarding Middle Archaic, Late Archaic, Transitional Archaic, and Early Woodland lifeways in the Upper Ohio Valley. East Steubenville Site (46Br31) Located approximately five miles west of the Pennsylvania state line in the Northern Panhandle of West Virginia, the East Steubenville Site (46Br31) was excavated by GAI Consultants, Inc. on a bluff top overlooking the Ohio River (Lothrop 2001a; Mohney 2002). This site is just south of the confluence of Harmon Creek and the Ohio River; thus, it is only 5 miles removed from the boundaries of Watershed D. East Steubenville is the type site for the Panhandle Archaic complex of the Upper Ohio River Valley and yielded dozens of Steubenville Stemmed, lanceolate, and Brewerton points. Several well-dated, Late Archaic burials excavated at the site provide information regarding burial practices, subsistence, and other social issues of Late Archaic Native Americans in the region. McKees Rocks (36AL16) and McKees Rocks Mound Sites Located on a hilltop overlooking the confluence of Chartiers Creek and the Ohio River, the McKees Rocks Site (36AL16) contains an Adena Mound and a late-late Woodland village (Buker 1968). Located in Watershed G of Subbasin 20, the site was excavated first in 1896 and later by the Allegheny Chapter of the Society for Pennsylvania Archaeology in the early 1960s. Buker (1968) interprets the McKees Rocks site to be a late manifestation of the Monongahela culture with close ties to Fort Ancient peoples. The village site contained a stockade with a central plaza and multiple burials. An Adena mound the McKees Rocks Mound was excavated on the same landform by the Carnegie Museum in 1896, but was apparently destroyed prior to the 1960s excavations. Several regional syntheses (Dragoo 1963; Mayer-Oakes 1955:145-152; McMichael 18 PA Prehistoric Data Synthesis, Raccoon Creek Watershed
1971) have included discussions of the mound. Both McKees Rocks and McKees Rocks Mound yielded Steubenville stemmed points as well, indicative of prior Late Archaic occupations (Mayer-Oakes 1955:139-141). demography, settlement patterns, site types, stone tool manufacture, and lithic raw material use. Watson Farm (46Hk34) The Watson Farm Site (46Hk34) is a late Middle Woodland to early Late Woodland village site on the Upper Ohio River in Hancock County, West Virginia. First investigated by Dragoo (1956) and the Carnegie Museum, the type site for the Watson phase of the Middle Woodland period in the Upper Ohio Valley. The site yielded small stone and earth burial mounds associated with multiple occupations characterized by small villages and hamlets. Wylie Site (36Wh247; 36Wh283) The Wylie Site is a Late Woodland Monongahela village within the nearby Chartiers Creek Valley (Watershed F of Subbasin 20). George (1995) reports on three other similar villages of likely Monongahela cultural affiliation in the valley. These sites, along with McKees Rocks, Georgetown, and Watson (see above) are the closest sites with Monongahela affinities to Raccoon Creek. No Monongahela sites have been excavated within the Raccoon Creek Valley proper. C. SUMMARY AND INTRODUCTION TO PREHISTORY CHAPTERS The following chapters provide an overview of the prehistory of the study area, utilizing research reports and PASS files data. The chapters are organized by time period: Paleoindian, Early Archaic, Middle Archaic, Late Archaic, Early Woodland, Middle Woodland, and Late Woodland. As noted earlier, this cultural historical sequence follows that established by Raber (1985) to maintain consistency with PASS files. The review encompasses the last 14,000 years of prehistory in the Raccoon Creek Watershed, comparing and contrasting data from the nearby Upper Ohio Valley and vicinity to better understand prehistoric Native American subsistence, PA Prehistoric Data Synthesis, Raccoon Creek Watershed 19
(This page intentionally blank) 20 PA Prehistoric Data Synthesis, Raccoon Creek Watershed
CHAPTER IV PALEOINDIAN PERIOD 14,000 to 10,000 BP A. PALEOINDIAN OVERVIEW Until recently, the Clovis culture, with its famous fluted spear points, was thought to be the oldest Native American culture of the Americas at ca. 11,000 BP (Bonnichsen and Turnmire 1991). However, excavations at sites throughout the Americas have challenged the well-established Clovis-first paradigm and stimulated extensive debate regarding the nature of the original founding population of the Americas (Adovasio and Page 2002; Clark et al. 2004). Excavations at Monte Verde, Chile indicate that Native Americans occupied coastal South America at least 12,000 years ago (Adovasio and Pedler 1997; Dillehay 1997; Meltzer 1989, 1995), suggesting the occupation of North America several hundred years prior. Recent excavations at sites along the California coast (Jones et al. 2002:215) propose that Paleoindian populations were coastally adapted, suggesting a coastal migration with subsequent migrations to the interior and to eastern North America (Fladmark 1983). Early Paleoindian sites along the eastern seaboard including Meadowcroft (Adovasio et al. 1978; Carr and Adovasio 2002a), Cactus Hill (McAvoy and McAvoy 1997), and a handful of others (Adovasio and Page 2002:263-275) support the likelihood of a pre-clovis population in North America. The early dates from Monte Verde and these east coast sites, and the generally slow nature of prehistoric human population migrations (Housley et al. 1997; MacDonald 2004), suggest that the founding population of North America may have been present as early as 14,000 to 16,000 years ago, with populations reaching the eastern seaboard and western Pennsylvania soon thereafter (Carr and Adovasio 2002a). B. PALEOINDIAN MATERIAL CULTURE, CHRONOLOGY, AND SUBSISTENCE Pre- Clovis Chronology If humans occupied Chile 12,000 years ago and arrived via a northern migration route across the Bering Strait, pre-12,000-year-old archaeological sites should exist across North America, although controversy surrounds the issue (Jones et al. 2002; Turner 2003). Few sites have survived the scrutiny demanded of such early Paleoindian sites (Meltzer 1989); however, Meadowcroft Rockshelter (36Wh297) in the Cross Creek drainage of Watershed D yielded a Miller lanceolate point, blades, and features from Stratum IIa with associated uncalibrated dates of (Carr and Adovasio 2002:8; Stuckenrath et al. 1982:80): 9 3 5 0 ± 7 0 0 B C ( 1 0, 6 0 0-1 2, 0 0 0 B P ; SI- 2491), charcoal from firepit/fire floor middle 1/3 of F46 1 0, 8 5 0 ± 8 7 0 B C ( 1 1, 9 3 0-1 3, 6 7 0 B P ; SI- 2489), charcoal from firepits/ lower 1/3 of F46 These two dates provide a conservative estimate of the earliest Paleoindian occupation of Meadowcroft, between 10,600 and 13,670 years ago (Carr and Adovasio 2002:8). Stuckenrath et al. (1982:80) list five additional uncalibrated dates from stratum IIa that suggest even earlier occupations of the shelter; however, these dates are less well accepted and not typically discussed as being associated with human activity: 1 1, 2 9 0 ± 1 0 1 0 B C (14,250-12,230 BP; SI- 2065), charcoal from firepits/ lower 1/3 of F46 1 1, 3 2 0 ± 3 4 0 B C (13,590-12,910 BP; SI- 2480), charcoal from firepits/ lower 1/3 of F46 1 2, 9 7 5 ± 6 2 0 B C (15,545-14,305 BP; SI- 1872), charcoal from firepits/ lower 1/3 of F46 PA Prehistoric Data Synthesis, Raccoon Creek Watershed 21
1 3, 1 7 0 ± 1 6 5 B C (15,285-14,955 BP; SI- 1606), charcoal from firepits/ lower 1/3 of F46 1 4, 2 2 5 ± 9 7 5 B C (17,150-15,200 BP; SI- 2354), charcoal from firepits/ lower 1/3 of F46 Since the publication of these early dates (Adovasio et al. 1977), the site has experienced intensive scrutiny, with some not accepting the early dates (Haynes 1980; Mead 1980) and others supporting them (Adovasio et al. 1980, 1990). The dates have been questioned due to the presence of coal seams near the rockshelter which are hypothesized to have contaminated charcoal within features at the site (Haynes 1980; Mead 1980). Adovasio (Carr and Adovasio 2002; Adovasio and Page 2002), as well as an independent study by Goldberg and Arpin (1999), has provided evidence to sufficiently refute objections to the early dates, including studies of soils that found no evidence of coal contamination within the cultural deposits of the rockshelter. Based on these data, this report accepts the two bracketing dates of 10,600-13,670 BP for the Miller point and associated artifacts. Another objection to the Late Pleistocene age of Meadowcroft is that the faunal and botanical assemblages from the Pre-Clovis occupations of the shelter are essentially modern and may not reflect the types of environments thought to have been present in the terminal Pleistocene of southwestern Pennsylvania (Guilday and Parmalee 1982:172). Guilday and Parmalee (1982:172), in the Meadowcroft synthesis volume, state that: In view of the age of the dates associated with the lower Stratum IIa material and the location of Meadowcroft in relation to the Laurentide Ice Front, the absence of boreal species is surprising [as] the effects of glacial cooling extended far south of Pennsylvania in eastern and central North America. Using data compiled by Cushman (1982), Adovasio and Page (2002:178) refute this position, arguing that even at the glacial maximum, in many instances the forest composition was not unlike its modern composition [and that] some 11,000 years ago, oak, elm, and ash were growing amid the spruce a mere fifteen miles from the limit of the glacier. The controversy surrounding the environmental setting of the Cross Creek drainage during the Pre-Clovis occupations at Meadowcroft will likely continue until enough data are presented to support one side or the other. More data are necessary from additional paleoenvironmental sites before a final judgment can be made regarding the forest composition during the latest Pleistocene in southwestern Pennsylvania; however, the data collected to date tend to support the hypothesis that Cross Creek in the vicinity of Meadowcroft was comprised of a temperate forest quite similar to that existing in the region today. Pre- Clovis Technology The Pre-Clovis lithic assemblage from Meadowcroft is comprised of flaking debris, blades, a large flake knife (the Mungai Knife), and a lanceolate projectile point the Miller Point. The Miller point from Meadowcroft (see Figure 6) is small, lanceolate, and was produced from fine-grained local chert. While the geologic provenance for the chert is uncertain, it is found locally (pers. comm.. J. Adovasio, June 2003) and is opaque and light gray, with yellow and purple striations on both faces. Boldurian (1985:298) describes the Miller point as having a distinctive base with straight margins that articulate with the straight basal attribute margin at angles of 97 degrees. The point was produced via the removal of parallel, overlapping biface thinning flakes, most of which traverse the centerline to create a lenticular cross-section. Personal examination of the point revealed slight unifacial beveling 22 PA Prehistoric Data Synthesis, Raccoon Creek Watershed
along two lateral margins, suggesting possible retouching for use as a knife. Figure 6. Miller Type Point from Stratum IIa. (from Boldurian 1985:284). According to Boldurian (1985:298), thinning of the base is unique, including lateral thinning on one face, and lateral and end thinning on the opposite face. Slight grinding is also present on all portions of the base to prepare it for hafting. The Krajacic Site (36Wh351), located on a hilltop ca. 10 miles southeast of Meadowcroft, yielded fragments of possible Miller points in various stages of biface reduction, each of which contribute to the notion that the points fit a Paleoindian reduction pattern (Boldurian 1985:308-325). Probable Miller points and point preforms from the Krajacic Site resemble Late Paleoindian lanceolate point forms from the North American Plains, including Hell Gap and Agate Basin (see Carr and Adovasio 2002:12; Frison 1991:60-63). Other sites with possible Miller points include Pershina Farm (36Wh608; Raccoon Creek drainage; Boldurian 1985:303) and Mungai Farm (36Wh106; Burgetts Fork; Carlisle and Adovasio 1982), both of which were identified during the Cross Creek Survey. While Miller points are considered to be of Paleoindian age by the Meadowcroft and Krajacic Site researchers (Adovasio and Carr 2002a; Boldurian 1985), their morphological similarity to Late Archaic Steubenville Stemmed points is noteworthy, and highlights potential problems in their identification (see Figure 7). Figure 7 compares the Miller type point from Meadowcroft and four Steubenville Stemmed points from the East Steubenville Site (46Br31) in the Panhandle of West Virginia. The Steubenville points in this figure were collected by GAI during data recovery excavations for the West Virginia Division of Highways. Meadowcroft and East Steubenville are less than 10 miles from one another. Surface-collected Miller points were found with Steubenville points at the Krajacic, Pershina, and Mungai Farm Sites (Boldurian 1985:303). At Meadowcroft, a Steubenville Stemmed point was found in Stratum III, stratigraphically above the Miller point. The general dearth of Miller points from well-excavated contexts only Meadowcroft has yielded a Miller point from a stratified context and the abundance of Steubenville points at sites in the Upper Ohio Valley (Lothrop 2001a; Mohney 2002; Mayer- Oakes 1955) has raised doubts as to the cultural affinities of some of the Miller points from open sites, including Krajacic, Pershina, and Mungai Farm (Boldurian 1985; Gardner 2002). Boldurian (1985:292-308) confirms that the two point types are very similar and difficult to distinguish based on morphological traits (Adovasio 1983). Figure 7. Comparison of the Miller Type Point from Meadowcroft (far right; Boldurian 1985:284) with Steubenville Points from the East Steubenville Site (46Br31). All Points at Same Scale. Steubenville Point images courtesy of the West Virginia Division of Highways. Blades were also recovered in association with the Miller Point at Meadowcroft (Photograph 6) and at the Krajacic Site (Boldurian 1985). Prismatic blades from the Krajacic and Meadowcroft Sites are morphologically similar PA Prehistoric Data Synthesis, Raccoon Creek Watershed 23
and suggest standardization of blade production during the early Paleoindian occupations of the Upper Ohio Valley. According to Boldurian (1985:93), most specimens have either two or three parallel flake scars on the dorsal surface, with flat or multifaceted platforms. The prismatic blades appear to have been removed from small prepared blade cores (Boldurian 1985:94). The Miller point, a heavily retouched Kanawha chert flake tool (the Mungai Knife), and the blades comprise a Pre-Clovis toolkit with an Eurasiatic, Upper Paleolithic flavor, according to Carr and Adovasio (2002:8). Photograph 6. Blades found in Association with the Miller Point, Stratum IIa, Meadowcroft Rockshelter (http://people.delphiforums.com/mcconaughy /meadowcroft/.htm). Nevertheless, blades were also found at Meadowcroft in Stratum IIa (upper), Stratum III, Stratum IV, Stratum V, Stratum VII, and the near-surface Stratum XI (Fitzgibbons 1982:100-101). In addition, blades of similar morphology to the Paleoindian blades from Meadowcroft (Boldurian 1985:236) were found with a Steubenville point at Cross Creek Village, approximately four miles upstream from Meadowcroft on Cross Creek near Avella. Adovasio et al. (1998:19) state that the only blades recovered from any Archaic contexts are so dissimilar to the Paleoindian forms and so unstandardized as to suggest that most of them are accidents. Adovasio (pers. comm. 2003) also states that the blades from above Stratum IIa are morphologically different from those associated with the Miller point in Stratum IIa. While feature dates and stratigraphic context support the hypothesis that the artifacts in Stratum IIa including the Miller point, blades, debitage, and the Mungai Knife are Paleoindian, more data need to be provided to evaluate the Meadowcroft lithic assemblage. Specifically, blades in the various site levels need to be compared and contrasted to evaluate the hypothesis that the Stratum IIa toolkit is unique to the Paleoindian levels. Clovis Material Culture and Chronology Subsequent to the possible pre-clovis occupations in Watershed D, Native American hunters utilized a variety of fluted points, including Clovis, Gainey, and Barnes in the East and Midwest. The Clovis culture (ca. 11,800-11,000 BP) was widely spread across the Americas, including Pennsylvania, while the Gainey and Parkhill phases were focused in the Great Lakes (Lantz 1984; Shott 1993). The Clovis tool kit is characterized by fluted spear points (Figure 8), largely produced from exotic cherts indicative of long-distance mobility. According to Howard (1991:257), Clovis points are comparatively large lanceolate points with lenticular to oval cross-sections. The points have slightly convex edges, gradually tapering tips, and their greatest width is at or near midpoint. Moderate sharpening and reshaping is common, often blunting the tip. The flutes are the most characteristic trait of the Clovis point, but they are not as visually impressive as Folsom points from western North America. In contrast to the Folsom point, which entailed the removal of one broad, long flute, Clovis points include the removal of multiple channel flakes from the base (Mounier et al. 1993), presumably a strategy to reduce the chances of point breakage during final fluting (Ellis and Payne 1995). No Clovis sites with radiocarbon dates have been excavated in western Pennsylvania. The most proximate welldated site is Shawnee-Minisink on the Delaware River. This eastern Pennsylvania site dates to 10,940±90 BP (Dent 2002: 56). 24 PA Prehistoric Data Synthesis, Raccoon Creek Watershed
Fluted-point sites typically yield a variety of other lithic tools as well, including finely shaped endscrapers, sidescrapers, spokeshaves, limaces, blades (Collins 1999), piece esquillees (wedges), and an assortment of other cutting tools (Custer 1996:104). This tool kit suggests a subsistence pattern largely oriented toward the hunting of game. In formerly glaciated regions, including areas north of the Ohio River, subsistence was likely oriented toward the hunting of megafauna (Meltzer 1988:15), while in unglaciated areas (including the entire project area), a wider variety of game were likely procured, including woodland caribou, deer, bear, and other smaller game (Carlisle and Adovasio 1982). Paleoindian Subsistence Archaeological evidence indicates that eastern North American Paleoindians practiced a generalized hunting and gathering economy, including collection of a variety of wild plants and nuts and the hunting of a variety of game (Carlisle and Adovasio 1982; Dent and Kauffman 1985; Dent 1991; Lepper 1983; Meltzer 1988). In unglaciated portions of Pennsylvania, including the current study area, Late Pleistocene environments likely were comprised of mixed hardwood forests in lowlands and mixed conifer-spruce-hardwood forests in higher elevation uplands between 14,000 and 10,000 years ago (for more details, see Chapter II). These wooded settings favored a more universe diet breadth, such as that used by contemporary Native Americans in deciduous and boreal forest settings (Steegman 1983). Faunal and floral remains from Meadowcroft Rockshelter support the contention of a mixed diet breadth between 14,000 and 10,000 years ago. While botanical and faunal remains were limited, they indicate use of a variety of plants and animals adapted to a temperate climate during the Paleoindian occupation of the rockshelter (Guilday and Parmalee 1982:173). However, the temperate faunal assemblage has raised serious questions regarding the antiquity of the lower Paleoindian levels (Mead 1980:579), as most climatic reconstructions indicate that boreal forest species should have been dominant at the time of Paleoindian occupation (ca. 14,000-12,000 BP). C. PALEOINDIAN SITES Paleoindian Site Locations: Trends from PASS Fi les Data PASS files indicate the presence of 8 previously recorded Paleoindian sites within the Raccoon Creek Watershed (Table 4), including five sites with fluted points and three with possible pre- Clovis Miller Points (see Figure 8). Two additional Paleoindian sites (Royal Tartan and Krajicic) were identified nearby in Watershed E, both of which are shown in Figure 8. The locations and types of Paleoindian sites in the Raccoon Creek watershed and vicinity indicate that Paleoindians traveled frequently on ridge divides of river and stream corridors, while camping mainly on uplands overlooking stream valleys (see Figure 8; Table 5 and Table 6). Table 4. Paleoindian Sites, Watershed D (PASS Files) SITE NO. SITE NAME ARTIFACT TYPE 36Wh0106 36Wh0110 Mungai Farm Mungai Saddle Pre-Clovis /Miller Pt Fluted Points 36Wh0223 Cuprik Clovis Pre-Clovis 36Wh0297 Meadowcroft /Miller Pt 36Wh0375 MS #66 Clovis Marosi 36Wh0407 Farm Clovis Pershina 36Wh0608 Farm Miller 36Wh1032 C V Cowden Paleoindian COUNTY Washington Washington Washington Washington Washington Washington Washington Washington MUNICIPALITY Smith Twp. Smith Twp Cross Creek Twp. Jefferson Twp. Cross Creek Twp. Cross Creek Twp. Robinson Twp. Cross Creek Twp. PA Prehistoric Data Synthesis, Raccoon Creek Watershed 25
Figure 8. Map of Paleoindian Sites. gaic0nsultants, INC. DRAWN DHM APPROVED jcl DATE 7/10/03 DWG. NO 20 02-441-10--C-A8 Paleoindian Sites in and Adjacent to Watershed D Map Key Sites with Fluted Points Sites with Miller Points Watershed Boundary River/Stream in Watershed D River/Stream not in Watershed D County Line Ohio River Ohio River Subbasin 20 Ohio River D Beaver County Harmon Cr. Meadowcroft Buffalo Creek WEST VIRGINIA To Ohio River PENNSYLVANIA Raccoon Creek Marosi Farm Cross Creek Krajacic E Wheeling Creek Allegheny County Washington County Mungai Saddle C u p r i k Pershina Farm MS 66 Royal Tartan Mungai Farm Ohio River Chartiers Creek Scale F Pittsburgh Monongahela River 0 5 10 miles Pennsylvania Figure 8 Map of Paleoindian Sites 26 PA Prehistoric Data Synthesis, Raccoon Creek Watershed
Table 5. Paleoindian Site Location Data, Watershed D (PASS Files) SITE NO. SITE NAME SITE TYPE SETTING LANDFORM ELEV. NEAREST WATER 36Wh0106 Mungai Farm Open Slopes Mid-Slopes 1220 Trib Burgetts Fork 36Wh0110 Mungai Saddle Open Saddle Upp. Slopes 1240 Burgetts Fork 36Wh0223 Cuprik Village Upland Hilltop 1320 Raccoon Creek 36Wh0297 Meadowcroft Rockshelter T0/T1 Terrace 850 Cross Creek 36Wh0375 MS #66 Open Upland Ridgetop 1300 Trib Cross Creek 36Wh0407 Marosi Farm Isolated Find Slopes Mid-Slopes 1200 N Fk Cross Creek 36Wh0608 Pershina Farm Village T0/T1 Terrace 1160 Raccoon Run 36Wh1032 C V Cowden Open Upland Ridgetop 1380 Cross Creek SITE NAME Table 6. Paleoindian Site Location Data, Watershed D (PASS files). NEAREST WATER DIST. TO WATER DIRECTION OF WATER STREAM ORDER DIST. TO CONF. (M) DIR. OF CONF. (FT) Mungai Farm Trib Burgetts Fk. 50 North 1 710 East Mungai Saddle Burgetts Fk. 400 Southeast 1 600 East Cuprik Raccoon Creek 230 Northeast 1 1220 Northwest Meadowcroft Cross Creek 60 South 5 400 Southwest MS #66 Trib Cross Crk 360 Southeast 1 1230 Northwest Marosi Farm N Fk Cross Crk 280 Northwest 3 500 West Pershina Farm Raccoon Run 115 West 2 150 Southwest C V Cowden Raccoon Creek 200 Southwest 1 1500 Northeast The eight Paleoindian sites in Watershed D are all within Washington County; no previously identified Paleoindian sites (with confirmed diagnostic artifacts) are present within Beaver County s portion of the study area (see Table 4). Four of the sites are located within the Raccoon Creek watershed, while three sites are within the Cross Creek Watershed (see Table 5). Within Watershed D, five of the sites are on small upland tributaries, while three overlook Cross Creek or Raccoon Creek proper. An additional site 36Wh520 in the Cross Creek drainage near Avella is supposed to have yielded a late Paleoindian Dalton-like point (Carlisle and Adovasio 1982:258-259); however, PASS files data only describe Archaic artifacts at this site. No additional information is available to confirm the presence of a late Paleoindian artifact at the site and, thus, it is not discussed further in this study. Each of the eight Paleoindian sites in Watershed D is located within approximately 20 miles of one another, in or near the Cross Creek drainage. Even the sites within the Raccoon Creek watershed are within a few miles of Cross Creek. All but one of the sites (Meadowcroft) occur at elevations of between 1,000 and 1,380 ft. amsl (see Table 5). Meadowcroft is located at an elevation of 850 ft. amsl on Cross Creek, a high-order tributary of the Ohio River. Six sites are on upland ridgetop, slopes, or saddles overlooking creek valleys, while only two sites are located directly on floodplain/terrace settings. Each of the sites, regardless of landform, occurs within 400m of a water source, with a mean of 212m from nearest water. Distance to the nearest stream confluence varied from 150 to 1,500m (mean, 789m) and does not appear to have been a factor in Paleoindian site placement (see Table 6). Key Paleoindian Sites in Watershed D Meadowcroft Rockshelter (36Wh297) Meadowcroft Rockshelter (36Wh297; Photograph 7) is the most important archaeological site in Watershed D, as it contains stratified deposits possibly extending as far back as 14,000 years ago (see discussion of site chronology above). Meadowcroft is discussed above, as it represents the type site for the Miller point and its associated blades. The site is located 2.5 miles northwest of Avella on the north bank of Cross Creek. The rockshelter faces south overlooking the creek at an elevation PA Prehistoric Data Synthesis, Raccoon Creek Watershed 27
of 850 ft. amsl (see Photograph 7). The interior shelter measures approximately 700 sq. ft., while the overhang is ca. 43 ft. above the extant cave floor. Details of excavations can be found in Adovasio et al. (1977), Carlisle and Adovasio (1982), and Carr and Adovasio (2002a; see References for other sources of information as well). Photograph 7. View of Meadowcroft Rockshelter from above Cross Creek. View Northeast (photo courtesy of J. Herbstritt and J. Adovasio). As discussed above, the Meadowcroft toolkit from the middle portion of Stratum IIa (Boldurian 1985:131) is projected to date to between 11,300±700 and 12,800±870 BP (Carr and Adovasio 2002a:8) and minimally includes the above-described Miller projectile point, blades, the Mungai Knife and flaking debris (see Figure 6 and Photograph 7). A cursory review of the flaking debris assemblage by the author in June 2003, suggests an emphasis on late-stage biface and projectile point production at the site; most flakes were less than 0.25 inches in maximum dimension. Along with the Miller point, the blades, and the knife, the late-stage flaking debris suggests that Meadowcroft was the location of a short-term hunting and lithic reduction camp for Pre-Clovis hunters traveling through the Upper Ohio Valley and vicinity. Adovasio has stated that the stratigraphic context of the various occupations is sound and that each component, from Paleoindian to Late Woodland, is represented by clusters of artifacts and features on distinct living floors within the rockshelter (pers. comm., 2003; Adovasio et al. 1977, 1978; Carlisle and Adovasio 1982). Adovasio also notes the general lack of bioturbation, including rodent digging, in the Archaic and Paleoindian strata (pers. comm. J. Adovasio 2003). Preliminary data within the various published articles support these assertions regarding the stratigraphic integrity of the site. However, bioturbation has been noted in other levels (Adovasio et al. 1977:66) and its effects at the site need to be evaluated in the final report. The lack of refits between the various site components (pers. comm. J. Adovasio 2003) also suggests minimal vertical artifact displacement at the site; if this is the case, presentation of these refitting data in a formal publication would support the stratigraphic integrity of artifacts at the site. A technical report would be most beneficial in that it would provide details regarding: cultural materials associated with each living floor; stratigraphy and the effects of bioturbation; and features (including profiles, planviews, and associations with specific artifacts), although data regarding the latter are available in Diane Landers dissertation (Beynon 1981). Mungai Farm (36Wh106) Mungai Farm is on a ridge divide between the Burgetts Fork, a tributary of Raccoon Creek, and the South Fork of Cross Creek (see Photograph 4). The farm contains a group of site localities with extensive artifact assemblages near the headwaters of upland streams. Miller points are the only Paleoindian artifacts recorded at the site, although several varieties of Archaic and Woodland point forms were also collected by the landowner and Adovasio s team (Fitzgibbons 1982:108). Vento and Donohue (1982:125) provide an overview of the lithic assemblages at the site. 28 PA Prehistoric Data Synthesis, Raccoon Creek Watershed
Mungai Saddle (36Wh110) Lantz (1984:211) reports that the Mungai Saddle Site has yielded three fluted points and several rejects. This site is near Mungai Farm (above) on uplands above the headwaters of the Burgetts Fork, within the Raccoon Creek Watershed. No additional information is available regarding this surface-collected site. Pershina Farm Site (36Wh608) As with Mungai Farm, the only Paleoindian artifacts recovered at the Pershina Farm Site were Miller type points (Boldurian 1985:303). This site, along Raccoon Creek in Watershed D, contains an extensive Steubenville component as well (Boldurian 1985:303). The previous discussion regarding the similarity of Miller and Steubenville points is especially pertinent for Site 36Wh608. Paleoindian Sites in Nearby Areas Based on data from PASS files, an additional seven sites, including Sites 36Wh351 (Krajacic) and 36Wh1312 (Royal Tartan #2), were identified in Watersheds E and F of Subbasin 20. Lantz (1984) also reports on the presence of a cluster of some 20 Paleoindian sites along Chartiers Creek in Watershed F east of Cross Creek. Krajacic Site (36Wh351) Site 36Wh351 the Krajacic Site is on an upland hilltop that was identified by the University of Pittsburgh during the Cross Creek project (Carlisle and Adovasio 1982; Boldurian 1985). Miller points and blades from this site are discussed above. The hilltop is the drainage divide for the Buffalo and Cross Creek watersheds. Within PASS files, the site is placed within Watershed E, due to its proximity to Buffalo Creek; however, the site is less than a mile from the southern boundary of Watershed D. The Krajacic Site yielded a wide variety of artifacts from different time periods within a plowed field. No features were identified and no radiocarbon dates were obtained from the site. Artifacts include Miller points in various stages of reduction along with stemmed and lanceolate Steubenville points. As discussed above, lithic artifacts from the site were examined by Boldurian (1985) for his doctoral dissertation. The main focus of the study was a comparison of the Miller points, blades, and blade cores, with similar artifacts from Meadowcroft, located approximately 5-6 miles northwest. Royal Tartan #2 (36Wh1312) GAI s excavations at the Royal Tartan Site (36Wh1312) yielded a small Clovis-like point fragment (MacDonald 2000a; Figure 9). The site is located at an elevation of 1204-1208 feet amsl on the first terrace overlooking Wolf Run, approximately 5-10 miles southeast of Cross Creek (Photograph 8). Wolf Run s headwaters are a spring approximately 400 meters to the southeast of the site; a seasonal tributary of Wolf Run bisects the terrace 50 meters to the east of the site. A small spring, the source of this seasonal runoff stream, is approximately 150 meters to the northeast of the site. This site location is typical of Paleoindian sites in the region, given its proximity to water sources and its position within an upland setting overlooking a loworder stream valley. The site would have been an ideal hunting overlook (see Photograph 8). Figure 9. Possible Clovis Point Preform Fragment from Site 36Wh1312 (Scale: 1 in. equals 2.5 cm; MacDonald 2000a). PA Prehistoric Data Synthesis, Raccoon Creek Watershed 29
Photograph 8. Setting of Site 36Wh1312 within the Wolf Fun Valley. View North. Five positive shovel tests mark the boundaries of the 2,500 square foot (232 sq. meters) Royal Tartan Site. A natural gas pipeline marks the eastern boundary of the site, while double negative STPs mark the other directional boundaries. All STPs revealed an Ap-Bt horizon sequence and extended approximately 50cm below surface and 15cm into sterile subsoil (Figure 6). STP H5 revealed one lithic artifact, a probable Flint Ridge chert Clovis projectile point fragment (Table 7; see Figure 9). The current author interprets the point as Clovis, an interpretation confirmed by Tom East and Ken Mohney of Skelly and Loy, Inc., and Stan Lantz, a Field Associate with Carnegie Museum of Natural History; nevertheless, Jon Lothrop of GAI believes the point is too fragmentary to type. The projectile point has several traits that are typical of Clovis. The base is slightly concave, with extensive pressure flake retouching. Two flake scars on one face extend from the base parallel to the central axis of the point. The most lateral biface reduction flake was likely intended to guide the possible channel flake, which overlies the smaller biface thinning flake and measures approximately 22.8mm from the base. This method of setting up a channel flake removal is typical of most Eastern Clovis fluted points (Callahan 1979). The base of the point and the intact lateral edge possesses fine pressure flaking, as well as some minor grinding to facilitate hafting. Remnant biface thinning flake scars extend across the centerline on both faces of the biface. Fine, parallel pressure flaking extends along the slightly excurvate edge of the biface. In this interpretation, the original tip of the artifact is no longer intact and the broken edge was reworked via fine pressure flaking into an oblique cutting edge. This edge resembles a projectile point tip, and is a source of confusion regarding the artifact s orientation. Based on the schematic diagram in Figure 9, the width of the original point was approximately 33.6mm, with a thickness of 8.1 to 8.9mm, and a width-to-thickness (w:t) ratio of 3.73. A Clovis point from the Flint Run Site in Virginia had a comparable w:t ratio of 3.84; however, most complete Clovis points have w:t ratios of between 5 and 10 (Callahan 1979). The point from Site 36Wh1312 is likely a preform that broke during channel flaking and was subsequently resharpened to function as a knife (see Figure 9); thus, the w:t ratio calculation may not be pertinent for this point. Ten other lithic artifacts (all flakes) were recovered during shovel testing at the site. Results of lithic raw material analysis are presented in the next section. Table 7. Site 36Wh1312: Projectile Point Data. FS STP COND. MAT. TYPE STRAT WGT LTH WIDTH* THICK 1 H5 Base F.Ridge projectile point Bt 10.5 44.5 33.6 8.8 *Width is projected based on reconstruction of the point. 30 PA Prehistoric Data Synthesis, Raccoon Creek Watershed
D. PALEOINDIAN SETTLEMENT PATTERNS AND LITHIC RAW MATERIAL USE Based on the data presented above, Paleoindian populations in the Raccoon Creek Watershed and vicinity used upper terraces of small streams and upland flat areas near streams and springs as short-term camp locations. These locations typically provided excellent viewsheds for hunters looking for game in the nearby stream valleys (see Photograph 8). Data regarding lithic raw materials at these Paleoindian sites are useful in interpreting settlement patterns and trade/travel realms. PASS files provide only general information regarding lithic raw material use, rather than specific raw material types. Another limitation of PASS data is the fact that raw materials are not differentiated based on the type of projectile point; thus, only single component Paleoindian sites can be used to assess lithic raw material use at Paleoindian sites identified in PASS files. Within Watershed D, single component Paleoindian sites include 36Wh223 (the Cuprik Site) and 36Wh375 (MS #66). Site 36Wh375 identifies chert/flint as the only lithic raw material recovered at the site, while Site 46Wh223 does not have information regarding lithic raw materials. One other site 36Wh407 (the Marosi Farm Site) contained multiple components, but identifies Flint Ridge chert as the material for the single Clovis point found at the site. Of the more than 40 cultural resource management reports available for review at the BHP for Watershed D, none contained information regarding Paleoindian lithic raw material use. Only the Cross Creek studies (Boldurian 1985; Carr and Adovasio 2002a; Vento and Donohue 1982) provide lithic raw material data for Paleoindian sites in Watershed D. At Meadowcroft, the Paleoindian levels of the site yielded a variety of lithic raw materials, summarized in Table 8. (Vento and Donohue 1982:124). Carr and Adovasio (2002a:7-9) state that the flaked stone inventory in Paleoindian levels of the site was produced from local Monongahela chert and non-local stones, including Flint Ridge and Upper Mercer cherts from east-central Ohio and jasper from central Pennsylvania. Vento and Donahue (1982:124), however, fail to identify jasper or Monongahela chert in the Paleoindian assemblage. Nevertheless, Adovasio confirmed in a recent discussion (pers. comm., June 2003) that these two lithic raw materials were in fact recovered in the Paleoindian levels of the site. Vento and Donahue (1982:124) identify local Brush Creek chert (31.6%) as the dominant material in the Paleoindian levels at the site. Upper Mercer (20.4%) and Flint Ridge (17.3%) cherts comprise a high percentage of the lithic assemblage, suggesting the curation of exotic materials. Other local cherts, such as 10 Mile (2.0%), and Uniontown (1.0%) comprise only a small portion of the lithic assemblage. The high percentages of chert from the Ohio Valley (Brush Creek) and east-central Ohio (Flint Ridge and Upper Mercer) suggests travel to Meadowcroft from the west along the Ohio Valley. Table 8. Lithic Raw Material Use at Paleoindian Sites in the Raccoon Creek Watershed D and Vicinity. SITE WATER- SHED UNION- TOWN 10 MILE MONON. CHERT BRUSH CREEK UPPER MERCER FLINT RIDGE Mungai Farm D -- -- -- 100% -- -- -- Meadowcroft D 1.0% 2.0% -- 31.6% 20.4% 17.3% Onondaga, Kanawha MS #66 D -- -- -- -- -- -- Chert/flint Marosi Farm D -- -- -- -- -- 100% -- Royal Tartan E 18.2% 45.5% -- -- 27.3% 9.1% -- Krajacic E -- -- 100% -- -- -- -- OTHER PA Prehistoric Data Synthesis, Raccoon Creek Watershed 31
The lone Paleoindian artifact at the nearby Mungai Farm Site (36Wh106) was a reported Miller Point produced from Brush Creek chert (Vento and Donohue 1982:125). At Krajacic (36Wh351), the Miller point was produced from Monongahela chert. Several other Miller point preforms in various stages of reduction were also recovered at Krajacic; lithic raw materials for these bifaces were Monongahela chert (n=12) and 10 Mile chert (n=1; Boldurian 1985:309). Lantz (1984) provides an overview of Paleoindian projectile points in western Pennsylvania, including information on lithic raw material use. Upper Mercer, Flint Ridge and Onondaga cherts were used to produce most of the points in his sample. Outside of Watershed D, GAI s report (MacDonald 2000a) of the Royal Tartan #2 site on a tributary of Buffalo Creek (see Photograph 8) provided lithic raw material data for a small Clovis lithic scatter. Of the 11 artifacts from the site, seven were produced from local cherts, including five 10 Mile chert flakes and two Uniontown chert flakes (Table 9). Four artifacts were produced from non-local cherts, including three Upper Mercer flakes and the Flint Ridge Clovis point. Of the seven flakes which were identifiable to type, five were biface reduction or shaping flakes, indicative of the late-stages of biface reduction, while two Uniontown flakes were decortication flakes. The flaking debris and projectile point data show an emphasis on the use of non-local cherts in the finishing of late-stage bifaces (see Table 9). Data from these various sites indicate use of a variety of local cherts during the Paleoindian period in Watershed D and vicinity (see Table 8). Of the non-local cherts, Flint Ridge and Upper Mercer cherts were recovered at several sites, indicating the curation of high quality lithic raw material during travels from eastcentral Ohio. Site location data indicate that travel likely occurred along upland drainage divides and ridges overlooking tributaries of the Upper Ohio Valley (Lantz 1984). E. RESEARCH ISSUES This summary of data regarding Paleoindian sites in and near the Raccoon Creek Watershed D has generated several research issues which should be considered when conducting archaeological work in the area. Thirteen Paleoindian research questions are listed below. As stated by Meltzer (2002:162), most of these questions could be resolved by publication of a final report of investigations from Meadowcroft. The questions surrounding the Miller point and associated blades have tremendous implications for several of the sites discussed above, including Pershina, Mungai Farm, and Krajacic. Table 9. Site 36Wh1312: Cross-Tabulation of Chert Type by Artifact Type. RAW MATERIAL DECORT BIFACE SHAPING FLAKE INDET. POINT TOTAL FLAKE REDUCTION FLAKE FRAGMENT Uniontown 2 0 0 0 0 0 2 Upper Mercer 0 0 2 0 1 0 3 10 Mile 0 1 2 2 0 0 5 Flint Ridge 0 0 0 0 0 1 1 Total 2 1 4 2 1 1 11 32 PA Prehistoric Data Synthesis, Raccoon Creek Watershed
This list of research questions is by no means comprehensive, and should be used only as a starting point for generating additional research issues. Archaeological sites which can provide information pertaining to these research questions will likely meet National Register Criterion D; thus, unless they lack integrity, sites which address these research questions will be eligible for listing in the National Register for Historic Places: 13. What does the presence of fair amounts of Flint Ridge and Upper Mercer cherts at Paleoindian sites in the watershed tell us about settlement patterns and travel routes? 1. When did Native Americans first occupy Watershed D of Subbasin 20? 2. Are there any other Pre-Clovis Meadowcroftlike sites in the Raccoon Creek Valley? If so, where are they? 3. How can we explain the similarity between Miller and Steubenville points? 4. Is the Clovis toolkit a descendent of that used by Native Americans at sites like Meadowcroft? Or, is population replacement responsible for the technological changes that occurred between the pre-clovis and Clovis sub-periods? 5. What types of features were excavated in the Paleoindian levels of Meadowcroft? 6. How do the blades in the Paleoindian site levels compare to those in other strata of Meadowcroft? 7. What types of tools were produced and utilized by Paleoindians in the watershed and what do they indicate about a region of origin? 8. How can we explain the apparent Paleoindian preference for locating sites in upland settings, including hilltops, saddles, and upper terraces of small streams and springs? 9. Is this site location preference unique compared to later periods? If so, why? 10. What types of features were utilized by Paleoindians? 11. What types of foods were exploited by Paleoindians? 12. What do lithic raw materials tell us about Paleoindian travel and trade patterns? PA Prehistoric Data Synthesis, Raccoon Creek Watershed 33
(This page intentionally blank) 34 PA Prehistoric Data Synthesis, Raccoon Creek Watershed
CHAPTER V EARLY ARCHAIC PERIOD 10,000 to 8800 BP A. EARLY ARCHAIC MATERIAL CULTURE AND CHRONOLOGY Approximately 10,000 years ago, eastern North America experienced the final transition from a cool, boreal environment of the Pleistocene to one dominated by the modern-like Mixed Mesophytic forest (Carr 1998a), as described in Chapter II. At this time, Native Americans apparently began to schedule their activities and specialize in methods of seasonal resource extraction in response to the existence of a more diversified resource base. Although archeological research on the Early Archaic period in the region is limited, it is likely that patterns characterizing the Northeast in general were also typical of western Pennsylvania and the Upper Ohio Valley (Carr 1998a). As during the Paleoindian period, Early Archaic Native Americans likely continued their use of a forager subsistence system, gradually becoming collectors by the end of the period (Binford 1979, 1980). Carr (1998a:49, 60) and Stewart and Katzer (1989) suggest that the region likely sustained a slight population increase during this period, likely due to ever-increasing familiarity with the Holocene resource base. Early Archaic Material Culture During the Early Archaic, fluted and unfluted lanceolate projectile points were replaced in Native American toolkits by notched and stemmed points with deep lateral edge serrations (Figure 10). This transition in point types may reflect changes in hafting technologies due to a change in hunting strategies or prey. Alternatively, the switch from fluted points to notched points represents a purely stylistic change, in which the high costs of fluted point production were rejected in favor of a more efficient mode of production. Some estimates of eastern fluted point production indicate that 15-20 percent of attempts at fluting ended in failure (Ellis and Payne 1995: 471). The costs of fluting may have been deemed too expensive by Early Archaic Native Americans, resulting in the predominance of Early Archaic projectile point types, including Kirk stemmed and cornernotched points and Palmer corner-notched points (Figure 10). Figure 10. Early Archaic Kirk Corner-Notched (left) and Stemmed (right) Points, (from Michels and Smith 1967:683; East and Beckman 1992:46). Eighteen Early Archaic sites in the Raccoon Creek watershed were identified as such due to the presence of Kirk corner-notched or stemmed points or Palmer points (see Figure 10). Kirk corner-notched points have medium-large, serrated triangular blades with straight to concave bases with corner notches. According to Justice (1987:78), Palmer corner-notched points and Kirk corner-notched points are morphologically identical. Flaking patterns on Kirk/Palmer points are largely random, while edges were thinned by removal of long, narrow pressure flakes resulting in a flat face lacking a medial ridge. Edges have deep serrations, measuring 2-mm-wide and 3-mm-deep on average. Bases of Kirk/Palmer corner-notch points were thinned via pressure flaking, while tangs on notches are predominantly rounded (Broyles 1971:65). Kirk stemmed points are medium sized with slightly expanding stems. The blades are long, narrow and, as with Kirk corner-notched points, have very deep and pronounced serrations. Stems expand toward the base which ranges from flat to slightly concave. Sides to the stem are straight to slightly concave, with bases thinned by pressure flaking. In addition to projectile points, lithic assemblages from Early Archaic sites, such as St. Albans on the Kanawha River in West Virginia (Broyles 1971), include a variety of hafted drills, knives, endscrapers, and sidescrapers. Sites typically have high PA Prehistoric Data Synthesis, Raccoon Creek Watershed 35
proportions of bifaces in various stages of reduction, depending on proximity to stone sources. Bifurcate points are typically considered Middle Archaic diagnostics, but they have been dated to as early as 9400 BP from the Sandts Eddy Site on the upper Delaware River, eastern Pennsylvania (Bergman et al. 1998:72-73). These points typically date to circa 8000 BP throughout the east and, for the purposes of this report, are considered to be Middle Archaic artifacts and are discussed in the next chapter (Justice 1987:82-99). According to Adovasio et al. (1998:19), the Early Archaic tool kit at Meadowcroft included utilized flakes, limited numbers of other bifaces and unifaces, modified bone and wood, and basketry. If details regarding the association of these artifacts with features were unavailable, they would facilitate a better understanding of the Early Archaic culture. Early Archaic Chronology and Radiocarbon Dates Carr (1998a:62) and Gardner (1987) suggest that the Early Archaic persisted between 10,000 and 8800 years BP in the northeast. Broyles (1971) identified Kirk corner-notched points in Early Archaic levels at the St. Albans Site on the Kanawha River, West Virginia, with bracketing radiocarbon dates of between approximately 9850 and 8750 BP (6800 BC; Kinsey 1972), while Carr (1998a) reported a date of 9250 BP at the Fifty Site in Virginia. The only Early Archaic radiocarbon date in Watershed D is 7165±115 BC (9000-9230 BP; SI- 2491), obtained from charcoal from firepit/firefloor, middle 1/3 of unit within the upper portion of Stratum IIa at Meadowcroft Rockshelter (Stuckenrath et al. 1982:80). Two other dates from this stratum are 6060±110 BC (7900-8120 BP; SI-2064) and 9350±700 BC (10,600-12,000 BP; SI-2491). information regarding the context of the date was available in the PASS files. B. EARLY ARCHAIC ARCHAEOLOGICAL SITES PASS Files Data PASS files indicate that 18 sites within Watershed D contain Early Archaic site components with diagnostic artifacts (Table 10). Tables 10-12 provide detailed site location data on the sites. Kirk and Palmer corner-notched and Kirk stemmed points were the predominant diagnostic artifacts at each of the sites (Table 10). As these tables show, a majority of sites (n=11) are within fairly level (<5% slope) floodplains (T0) or terraces (T1) of creeks or rivers, with six sites located on upland hilltops, saddles, or slopes (Table 11). Seventeen of the sites are open sites (including one lithic reduction location and two isolated finds), while Meadowcroft is the only rockshelter with Early Archaic artifacts in the watershed. The prevalence of sites in river and creek floodplains/terraces is a contrast to Paleoindian sites, which are predominantly on uplands, slopes, or saddles overlooking stream valleys. Adovasio et al. (1998:1) report a similar Early Archaic site distribution pattern within the Cross Creek drainage of Watershed D. Perhaps this change in site placement is a reflection of travel patterns, with Early Archaic Native Americans preferring to travel along river terraces, while Paleoindians traveled predominantly along upland ridges. Another hypothesis would be that Early Archaic populations began to establish longer-term base camps in alluvial settings, with uplands used for seasonal food procurement. If this is the case, then the switch from a forager to a collector subsistence strategy occurred earlier than proposed in the region. Cowin (1991) suggests that this change in subsistence was fully entrenched by the Middle Archaic in western Pennsylvania. An additional date of 10,000 BP is listed in PASS files for Site 36Wh559 the Point Site in Hanover Township, Washington County. No 36 PA Prehistoric Data Synthesis, Raccoon Creek Watershed
Table 10. Early Archaic Sites, Watershed D (PASS Files) SITE NO. SITE NAME ARTIFACT TYPE COUNTY MUNICIPALITY 36Bv0026 Kochanioski Kirk Beaver Hopewell Township 36Bv0040 Crevallero Kirk Beaver -- 36Bv0050 St. Joe #2 Kirk Beaver Potter Township 36Bv0060 Philles Island Kirk Beaver Shippingport Borough 36Bv0241 Dravo #2 Kirk Beaver Greene Township 36Wh0106 Mungai Farm Palmer/Kirk Washington Smith Township 36Wh0297 Meadowcroft Kirk Washington Jefferson Twp. 36Wh0401 MS #96 Kirk Washington Independence Township 36Wh0475 Berrinski Camp Kirk Washington Mount Pleasant Township 36Wh0501 FB 50 Palmer Washington Cross Creek Township 36Wh0565 Beadling No. 1 Kirk Washington Hanover Township 36Wh1046 Kirk Site Kirk Washington Cross Creek Township 36Wh1088 Fort City Kirk Washington Mount Pleasant Township 36Wh1093 Davidson Site Kirk Washington Hopewell Township 36Wh1118 Lowry #9 Palmer/Kirk Washington Mount Pleasant Township 36Wh1152 Vanzin Kirk Washington Hopewell Township 36Wh1156 Luba #2 Kirk Washington Hopewell Township 36Wh1191 Lowry #12 Kirk Washington Mount Pleasant Township Table 11. Early Archaic Site Location Data, Raccoon Creek Watershed (PASS Files) SITE NO. SITE NAME SITE TYPE SETTING TOPO ELEV. NEAREST WATER 36Bv0026 Kochanioski Open T0/T1 Floodplain 780 Raccoon Creek 36Bv0040 Crevallero Open Upland Hill Ridge/Toe 920 Trib. Racc. Creek 36Bv0050 St. Joe #2 Open T0/T1 Floodplain 700 Raccoon Creek 36Bv0060 Philles Island Open T0/T1 Island 680 Ohio River 36Bv0241 Dravo #2 -- T0/T1 Floodplain 680 Ohio River 36Wh0106 Mungai Farm Open Slopes Middle Slopes 1220 Trib Burgetts Fork 36Wh0297 Meadowcroft Rockshelter T0/T1 Terrace 850 Cross Creek 36Wh0401 MS #96 Open Saddle Saddle 1180 Trib Cross Creek 36Wh0475 Berrinski Camp Open T0/T1 Terrace 1080 Raccoon Creek 36Wh0501 FB 50 Isolated Find Slopes Middle Slopes 1120 S Fk Cross Creek 36Wh0565 Beadling No. 1 Open T0/T1 Stream Bench 940 -- 36Wh1046 Kirk Site Open Slopes Middle Slopes 1250 Raccoon Creek 36Wh1088 Fort City Open T0/T1 Terrace 1220 Raccoon Creek 36Wh1093 Davidson Site Open Upland Ridgetop 1280 Raccoon Creek 36Wh1118 Lowry #9 Lith.Red. T0/T1 Terrace 1120 Trib Cross Creek 36Wh1152 Vanzin Open Slopes Hillslope 1200 Trib Cross Creek 36Wh1156 Luba #2 Open T0/T1 Terrace 1160 -- 36Wh1191 Lowry #12 Open T0/T1 Terrace 1100 -- For both the Paleoindian and Early Archaic, access to water was a key to site placement. Fifteen of the 18 Early Archaic sites are within 100m of a water source, while the remaining three sites are less than 310m from a water source. The mean distance to a water source for the sites is 102 meters for Early Archaic sites, compared to a mean of 212 meters for sites during the Paleoindian period. This decrease in mean distance to water likely reflects the placement of sites within stream valleys during the Early Archaic, as opposed to uplands during the Paleoindian. In contrast to the Paleoindian period, in which sites were mostly on uplands near low-order streams, Early Archaic sites are mostly located in the Raccoon Creek drainage (n=8), including five along Raccoon Creek itself and three along feeder streams (Table 12). Five of the Early Archaic sites are within the Cross Creek watershed, including four on feeder streams and one on Cross Creek. Two sites are on the main stem of the Ohio River between the mouth of Raccoon Creek and the West Virginia state line. PA Prehistoric Data Synthesis, Raccoon Creek Watershed 37
Table 12. Early Archaic Site Location Data, Raccoon Creek Watershed (PASS Files). SITE NO. NEAREST WATER DIST. TO WATER DIRECTION OF WATER STREAM ORDER DIST TO CONF. DIR. TO CONF. 36Bv0026 Raccoon Creek 60 South 1 570 -- 36Bv0040 Trib. Raccoon Creek 120 Northeast 2 180 -- 36Bv0050 Raccoon Creek 40 South 1 1120 Southeast 36Bv0060 Ohio River 90 North 9 880 Northeast 36Bv0241 Ohio River 10 Northwest 9 440 Southwest 36Wh0106 Trib Burgetts Fork 50 North 1 710 East 36Wh0297 Cross Creek 60 South 5 400 Southwest 36Wh0401 Trib Cross Creek 0 On-site -- 830 Southeast 36Wh0475 Raccoon Creek 70 West 3 120 Southeast 36Wh0501 S Fk Cross Creek 80 North 2 220 Northeast 36Wh0565 -- 100 West 2 220 Northeast 36Wh1046 Raccoon Creek 200 Southwest 1 1300 South 36Wh1088 Raccoon Creek 180 South 1 340 Northwest 36Wh1093 Raccoon Creek 290 Southeast 1 440 North 36Wh1118 Trib Cross Creek 100 Southeast 2 180 Southwest 36Wh1152 Trib Cross Creek 310 South 1 2120 Northeast 36Wh1156 -- 80 Northeast 1 340 North 36Wh1191 -- 10 North 1 180 Southwest As during the Paleoindian period, access to a stream confluence was not a determining factor in Early Archaic site placement in Watershed D, with sites ranging between 120 and 2,120m from a confluence (mean=588m). This mean distance to a confluence is 200m less than during the Paleoindian period, likely another reflection of the movement of sites from uplands to lowlands during the Early Archaic. Adovasio et al. (1998:18) state that Early Archaic site locations in Cross Creek reflect a preference for site placement near stream confluences. While Early Archaic sites were generally closer to stream confluences than Paleoindian sites, Adovasio et al s (1998:18) data for Cross Creek are not matched by PASS file data shown in Table 12. PASS files data for Early Archaic sites in the Cross Creek watershed indicate an average distance of 750 meters from a confluence. Key Early Archaic Sites Meadowcroft (36Wh297) As described above, the lone Early Archaic radiocarbon date in Watershed D is 7165±115 BC(9000-9230 BP; SI- 2491) from Meadowcroft Rockshelter (Stuckenrath et al. 1982:80). The date from the site was obtained from charcoal from firepit/firefloor, middle 1/3 of unit within the upper portion of stratum IIa. This portion of stratum IIa, unfortunately, did not yield any diagnostic artifacts; however, immediately above stratum IIa, stratum IIb yielded a wide range of projectile points and pottery spanning the entire Archaic and Woodland. The only Early Archaic artifacts in stratum IIb were a Kirk Stemmed point and a Kirk-like serrated blade fragment. Lithic raw materials used in production of Early Archaic points are described below. Mungai Farm (36Wh106) Mungai Farm (36Wh106) yielded a collection of 85 Early Archaic projectile points, according to Fitzgibbons (1982:109) and Vento and Donohue (1982:125). Limited information regarding this site is also available in Boldurian s (1985) doctoral dissertation on the nearby Krajacic Site. Site 36Wh1309 (Watershed F) Site 36Wh1309 (Watershed F, Subbasin 20) was identified along Robinson Run, a tributary of Chartiers Creek, during a sewerline survey south of the town of McDonald (Skinner 1999). Shovel test excavations at the site yielded a Palmer Corner-Notched point produced from an indeterminate chert type and 12 flakes produced 38 PA Prehistoric Data Synthesis, Raccoon Creek Watershed
mainly from Monongahela chert. The site was avoided during sewer line construction and no additional work was conducted at the small open site. C. EARLY ARCHAIC SETTLEMENT PATTERNS AND LITHIC RAW MATERIAL USE PASS files include data for four singlecomponent Early Archaic sites in Watershed D that are useful in the evaluation of lithic raw material use (Table 13)). In addition, Vento and Donohue (1982:124-127) include detailed data regarding lithic raw material use for projectile points in the Cross Creek drainage. The four single component Early Archaic Sites are listed in Table 13, with chert/flint being the only lithic raw material identified in PASS files data for Early Archaic artifacts. As described in Table 14, a variety of lithic raw materials were utilized for Early Archaic projectile point manufacture at the Cross Creek sites, including Upper Mercer, Brush Creek, Flint Ridge, Onondaga, Kanawha, 10 Mile, and Uniontown cherts. These data indicate a high degree of travel for Early Archaic Native Americans, especially along the Upper Ohio River Valley and east-central Ohio. Settlement patterns and trade realms were clearly similar for Early Archaic and Paleoindian Native Americans in the Upper Ohio Valley, including Watershed D. D. EARLY ARCHAIC: CONCLUSION AND RESEARCH QUESTIONS E a r l y A r c h a i c S u m m a r y The 18 sites with diagnostic Early Archaic artifacts in Watershed D are a significant increase compared to the earlier Paleoindian period (8 sites). This is especially true if the number of sites is averaged per decade of the period. When this is done, the Early Archaic (120 decades; 18 sites) has 0.13 sites per decade, compared to only 0.03 sites per decade during the Paleoindian (250 decades; 8 sites). Population densities during the Paleoindian period were in all likelihood extremely low across North America (MacDonald 2004), with increasing site counts during the Early Archaic period likely reflecting real population increases over time, rather than site preservation bias. Table 13. Lithic Raw Material Use at Single Component Early Archaic Sites, Watershed D (PASS Files). SITE NO. SITE NAME 36Bv0050 St. Joe #2 36Bv0060 Philles Island 36Bv0241 Dravo #2 36Wh1088 Fort City POINT TYPE Kirk Kirk Kirk Kirk LITHIC NEAREST WATER Chert/Flint Raccoon Creek Chert/Flint Ohio River -- Raccoon Creek -- Raccoon Creek Table 14. Cross-Tabulation of Cross Creek Site by Lithic Raw Material Type for Early Archaic Artifacts (Vento and Donohue 1982:124-127). LOCATION ONOND AGA KANAWHA FLINT RIDGE UPPER MERC BRUSH CRK 10 MILE UNIONTOWN OTHER TOTAL PTS. Meadowcroft* % 16.6 50.2 -- 16.6 16.6 -- -- -- 6 Mungai Farm % 9.4 7.1 18.8 28.2 24.7 7.1 3.5 1.2 85 Cross Cr. Sites % 9.1 -- 27.3 18.2 9.1 18.2 -- 18.2 11 Total Points (n) 10 9 19 27 23 8 3 3 102 Total % 9.8 8.8 18.6 26.5 22.6 7.8 2.9 2.9 100 *For Meadowcroft, data is for Early and Middle Archaic points PA Prehistoric Data Synthesis, Raccoon Creek Watershed 39
Another major difference between the Paleoindian and Early Archaic periods is the distribution of the previously recorded sites. Whereas the 8 Paleoindian sites in PASS files were in or near the Cross Creek drainage, only five of the 18 Early Archaic sites are within the Cross Creek watershed. Most of the Early Archaic sites are distributed fairly evenly across Watershed D, with the majority in the Raccoon Creek drainage. Mean distances to nearest water and nearest stream confluence were reduced during the Early Archaic period as well, confirming the movement of most sites from uplands to alluvial settings. Based on the data in Watershed D, thus, the Early Archaic appears to represent a distinct cultural/demographic break from the Paleoindian period, as reflected by: 1) increases in population; 2) movement of most sites into alluvial settings, as opposed to uplands; and 3) increased use of a wide variety of landforms. The two periods share similar patterns of lithic raw material use, however, as reflected in projectile point assemblages. Points of the Paleoindian and Early Archaic periods indicate wide-ranging travel patterns within southwestern Pennsylvania, northern West Virginia, and east-central Ohio, as reflected in the use of Kanawha chert, Flint Ridge chert, Upper Mercer chert, and Brush Creek chert. Early Archaic Research Questions This summary of Early Archaic archaeological data in and near Watershed D of Subbasin 20 has generated several research issues which should be considered when conducting archaeological work in the area. Eleven Early Archaic research questions are listed below; this list is by no means comprehensive and should be used only as a starting point for generating additional research issues. Archaeological sites which can provide information pertaining to these and other research questions will likely meet the National Register Criterion D; thus, unless they lack integrity, sites which address these research questions will be eligible for listing in the National Register for Historic Places: 1. Was there cultural/demographic continuity between the Paleoindian and Early Archaic periods or was this a time of cultural transition? 2. Was there a significant increase in population during the Early Archaic period, as suggested by increasing site counts? 3. Why did Early Archaic populations place sites mostly in alluvial settings, while Paleoindians apparently preferred upland settings? 4. Do the differences in projectile point types and site placements indicate a change in subsistence strategy (e.g., from forager to collector)? 5. What types of foods were procured during the Early Archaic? 6. What types of tools were produced and utilized by Early Archaic Native Americans? 7. Was the change in projectile point technology precipitated by a change in prey species and/or hunting strategies or are the changes stylistic? 8. Were bifurcate points used during the Early Archaic in Watershed D? 9. What types of lithic raw materials were used during the Early Archaic? 10. Why do lithic raw material use patterns remain consistent between the Paleoindian and Early Archaic periods? 11. What do lithic raw materials tell us about Early Archaic travel and trade patterns? How do they compare to the Paleoindian and Middle Archaic? 40 PA Prehistoric Data Synthesis, Raccoon Creek Watershed
CHAPTER VI MIDDLE ARCHAIC PERIOD 8800 to 5300 BP A. MIDDLE ARCHAIC OVERVIEW Middle Archaic sites are characterized by the presence of projectile points with bifurcate point bases, a distinctive stylistic change compared to the predominantly corner-notched bases of the Early Archaic. Bifurcate points appear to date to the latter portion of the Early Archaic to the Middle Archaic and may suggest technological continuity between the two periods (Bergman et al. 1998:70; Stewart and Cavallo 1991). However, others (Carr 1998a:60-64, 1998b:79; Gardner 1987) dispute the cultural continuity between the Early and Middle Archaic, as there appears to be a significant increase in population that corresponds to the switch to bifurcate points. These changes may reflect additional cultural transitions, including subsistence and settlement pattern shifts. Watershed D experienced a 2.5-fold increase in the number of recorded Middle Archaic sites compared to the Early Archaic period (Figure 11), supporting Carr s (1998a:61; 1998b:88) assessment of a significant increase in population. Nevertheless, if duration of the respective periods is considered (as per Fiedel 2001), site count densities are similar between the Early Archaic (0.15 sites per decade per period) and Middle Archaic (0.13 sites), with the Early Archaic actually having a slightly higher site density. No site count increase was observed within the Cross Creek drainage either, as Adovasio et al. (1998:1-2) note a decline in intensity of occupation during the Middle Archaic, similar to the site density data presented here. Cowin (1991:48) characterizes the Middle Archaic settlement system as consisting of base camps positioned on Holocene-age river terraces, smaller resource procurement stations for seasonal plant and animal exploitation in upland settings, and lithic-reduction stations near bedrock outcrops of stone exploited for tool manufacture. As noted in the previous chapter, this settlement system likely was initiated during the Early Archaic and became further entrenched during the Middle Archaic. Site Count 50 40 30 20 10 0 0.16 0.14 0.12 0.1 0.08 0.06 0.04 0.02 0 0.03 0.15 Figure 11. Changing Site Counts in the Early Holocene: Comparison of Site Counts (top) and Site Density per Decade (bottom), Watershed D (PASS files). B. MIDDLE ARCHAIC MATERIAL CULTURE AND CHRONOLOGY 8 18 0.13 Paleo EA MA Period 45 Paleo EA MA Period Middle Archaic Material Culture Bifurcate point production is the major technological change between the Early and Middle Archaic periods (Figure 12). Middle Archaic point forms in the Raccoon Creek drainage are predominantly Kanawha stemmed, Neville/Stanly and LeCroy, with fewer examples of MacCorkle and St. Albans (Broyles PA Prehistoric Data Synthesis, Raccoon Creek Watershed 41
1971; Kuhn 1985). LeCroy bifurcate points are the most common projectile points at Middle Archaic sites in the study area (Cowin 1991; Broyles 1971). According to Broyles (1971:69), LeCroy points are fairly small with trianguloid blades with straight edges. Blades are serrated on most points (but not all) and the bases are deeply notched by the removal of a large central flake with surrounding pressure flakes. Stems are straight to slightly flared with no basal grinding. Figure 12. Middle Archaic Bifurcate Point, Actual Size (from Custer et al. 1996:31). St. Albans Site dated to approximately 8300 BP (Broyles 1971). An extensive Middle Archaic component at the West Water Street Site on the West Branch of the Susquehanna River in northcentral Pennsylvania yielded a date of ca. 7400 BP (Custer et al. 1996:33), associated with Neville/Stanly and LeCroy bifurcate points (see Figure 12). Cowin (1991) reports a suite of radiocarbon dates of between approximately 5500 and 7400 BP for bifurcate points at the State Road Ripple Site (36CL52) in Clarion County, northwestern Pennsylvania. As with the preceding Paleoindian and Early Archaic periods, Meadowcroft Rockshelter provides the only radiocarbon dates for Middle Archaic sites in the study area. Stratum IIb at the site yielded two features with Middle Archaic radiocarbon dates: 1) 4 7 2 0 ± 1 4 0 B C ( 6 5 3 0-6 8 1 0 B P ; S I - 2 0 5 5 ) from charcoal from firefloor/lower 1/3 of unit (Stuckenrath et al. 1982:80). While some researchers place Otter Creek points in the Middle Archaic (East and Beckman 1992:69-70; East et al. 1999:4-6), others discuss them in the context of the Late Archaic (Carr 1998a, 1998b; Kinsey 1972). In addition, Cowin (1991) states that Brewerton points, considered by most to be of Late Archaic age, are diagnostic of the Middle Archaic. For the sake of consistency with PASS files data, this report discusses Brewerton and Otter Creek points within the contexts of the Laurentian tradition of the Late Archaic period. In addition to projectile points, lithic assemblages from Middle Archaic sites include a variety of hafted drills, knives, endscrapers, and sidescrapers, as well as cores and utilized flake tools (Custer et al. 1996). Sites typically have high proportions of bifaces in various stages of reduction, depending on proximity to stone sources. Middle Archaic Chronology Carr (1998b:79) cites radiocarbon dates of 8900 and 8888 BP for bifurcate point levels from the Fifty Site (Virginia), while LeCroy levels at the 2) 4 3 4 0 ± 3 5 5 ( 6 6 4 5-5 9 3 5 B P ; S I - 2 3 5 8 ) from charcoal from firepit, middle 1/3 of unit (Stuckenrath et al. 1982:80). Cowin (1991:47) includes two additional radiocarbon dates of 6630±70 BP (PITT-292) and 6315±280 BP (PITT-122) from Meadowcroft; however, these dates are not included in lists published in Stuckenrath et al. (1982) and Adovasio et al. (1998). Due to the general lack of specific feature data in Meadowcroft publications, the context and proveniences of the features (and associated radiocarbon dates) in relation to Middle Archaic artifacts is uncertain. As stated by Cowin (1991:48), Meadowcroft researchers, as of 1990, were not able to relate a specific point to a specific date at the site. C. MIDDLE ARCHAIC ARCHAEOLOGICAL SITE TYPES AND LOCATION TRENDS As with the preceding two chapters, only sites with diagnostic point types are included in the discussion of PASS files data. Forty-five sites in Watershed D contained Middle Archaic 42 PA Prehistoric Data Synthesis, Raccoon Creek Watershed
bifurcate projectile points (Table 15), a significant increase in site counts compared to the Paleoindian (n=8) and Early Archaic (n=18) periods (see Figure 11). As noted above, however, if site counts are averaged by total decades per period (as per Fiedel 2001:107), the Middle Archaic (8800-5300 BP; 350 decades; 45 sites) and Early Archaic (10,000-8800; 120 decades; 18 sites) have nearly identical site count densities, with 0.15 and 0.13 sites per decade, respectively. These data in Watershed D, thus, contrast that presented by Carr (1998a:61; 1998b:88) for Pennsylvania as a whole, which suggest significant population increases during the Middle Archaic. Table 15. Middle Archaic Sites, Watershed D (PASS Files). SITE NO. SITE NAME ARTIFACT TYPE COUNTY MUNICIPALITY 36Bv0003 Old Indian Fort Bifurcate Beaver Shippingport Borough 36Bv0013 Circle Rock Bifurcate Beaver -- 36Bv0014 Mc Micale Bifurcate Beaver -- 36Bv0016 -- Bifurcate Beaver -- 36Bv0017 McElhaney Hog Farm Bifurcate Beaver -- 36Bv0019 -- Bifurcate Beaver -- 36Bv0020 Wilson Triangle Bifurcate Beaver -- 36Bv0022 Boyscout Camp Bifurcate Beaver Hopewell Township 36Bv0025 Deadend Bifurcate Beaver Center Township 36Bv0026 Kochanioski Bifurcate Beaver Hopewell Township 36Bv0036 Wassler #2 Bifurcate Beaver -- 36Bv0037 -- Bifurcate Beaver -- 36Bv0039 New Pottery Bifurcate Beaver -- 36Bv0052 John Bush Farm Bifurcate Beaver -- 36Bv0107 Raccoon Ch. Gyd. Bifurcate Beaver Hopewell Township 36Bv0123 Big Travers Creek Bifurcate Beaver -- 36Bv0133 McHaffic Farm Bifurcate Beaver Greene Township 36Bv0162 Emil Alam Bifurcate Beaver Center Township 36Bv0178 George Frey Bifurcate Beaver -- 36Bv0210 Lower Circle on Rock Bifurcate Beaver Raccoon Township 36Bv0221 Hilltop Bifurcate Beaver Potter Township 36Wh0106 Mungai Farm Bifurcate Washington Smith Township 36Wh0110 Mungai Saddle Bifurcate Washington Smith Township 36Wh0187 Clair Cowden Bifurcate Washington Cross Creek Township 36Wh0293 Cross Creek Bifurcate Washington Cross Creek Township 36Wh0297 Meadowcroft Bifurcate Washington Jefferson Township 36Wh0392 -- Bifurcate Washington Cross Creek Township 36Wh0399 MS #94 Bifurcate Washington Cross Creek Township 36Wh0400 MS #95 Bifurcate Washington Burgettstown Borough 36Wh0409 MS #104 Bifurcate Washington Smith Township 36Wh0508 -- Middle Archaic Washington Jefferson Township 36Wh0512 -- Bifurcate Washington Cross Creek Township 36Wh0688 Lowry #4 Bifurcate Washington Mount Pleasant Township 36Wh0720 Hillside Site Bifurcate Washington Mount Pleasant Township 36Wh0986 Lutz #1 Bifurcate Washington Mount Pleasant Township 36Wh1002 Scott Site Bifurcate Washington Smith Township 36Wh1096 Lutz #7 Bifurcate Washington Mount Pleasant Township 36Wh1112 Cumer Site Bifurcate Washington Smith Township 36Wh1115 Murgel Site Bifurcate Washington Mount Pleasant Township 36Wh1118 Lowry #9 Bifurcate Washington Mount Pleasant Township 36Wh1148 Brezinski #5 Bifurcate Washington Hopewell Township 36Wh1151 Lowry #11 Bifurcate Washington Hopewell Township 36Wh1152 Vanzin Bifurcate Washington Hopewell Township 36Wh1161 Gregorski Bifurcate Washington Jefferson Township 36Wh1201 Vanzin #3 Bifurcate Washington Mount Pleasant Township PA Prehistoric Data Synthesis, Raccoon Creek Watershed 43
As with the Early Archaic, many Middle Archaic sites are open camps or lithic reduction sites on the floodplains or terraces of creeks and rivers (Table 16). Of the 45 Middle Archaic sites, 21 are in alluvial settings, with an additional 12 sites on upland flats, hilltops, or ridges. An additional 10 sites are on slopes, while two sites are in saddles between upland ridges. Compared to the Early Archaic (mean elevation=1,027 ft. amsl), the mean elevation of Middle Archaic sites (1,052 ft. amsl) suggests a slightly greater preference for upland site placement. Overall, PASS data indicate that site locations comprise the entire suite of landforms in Watershed D, suggesting wide ranging use of the landscape during the Middle Archaic. Table 16. Middle Archaic Site Location Data, Watershed D (PASS Files) SITE NO. NEAREST WATER SITE TYPE SETTING TOPOGRAPHY ELEV. 36Bv0003 Ohio River Open T0/T1 Floodplain 660 36Bv0013 Raccoon Creek Petroglyph T0/T1 Terrace 940 36Bv0014 Raccoon Creek Open T0/T1 Floodplain 820 36Bv0016 Raccoon Creek Open Upland Hill Ridge/Toe 900 36Bv0017 Trib Raccoon Creek Open Upland Hill Ridge/Toe 860 36Bv0019 Raccoon Creek Lithic Red. T0/T1 Terrace 880 36Bv0020 Raccoon Creek Open T0/T1 Floodplain 800 36Bv0022 Raccoon Creek Open T0/T1 Terrace 860 36Bv0025 Raccoon Creek Open T0/T1 Terrace 820 36Bv0026 Raccoon Creek Open T0/T1 Floodplain 780 36Bv0036 Raccoon Creek Open T0/T1 Terrace 920 36Bv0037 Raccoon Creek Open Upland Hill Ridge/Toe 940 36Bv0039 Raccoon Creek Open Upland Hill Ridge/Toe 880 36Bv0052 Trib Raccoon Creek Isolated Find Upland Hill Ridge/Toe 960 36Bv0107 Raccoon Creek Open T0/T1 Stream Bench 940 36Bv0123 Raccoon Creek Open T0/T1 Floodplain 840 36Bv0133 Mill Creek Open Upland Hilltop 1060 36Bv0162 Raccoon Creek Open T0/T1 Floodplain 780 36Bv0178 Trib Raccoon Creek Open T0/T1 Stream Bench 1160 36Bv0210 Raccoon Creek Open T0/T1 Terrace 960 36Bv0221 Trib Raccoon Creek Open Upland Upland Flat 1100 36Wh0106 Trib Burgetts Fork Open Slopes Middle Slopes 1220 36Wh0110 Burgetts Fork Open Slopes Upper Slopes 1240 36Wh0187 S Fk Cross Creek Open Slopes Lower Slopes 1340 36Wh0293 Raccoon Creek Lithic Red. T0/T1 Terrace 1000 36Wh0297 Cross Creek Rockshelter T0/T1 Terrace 1000 36Wh0392 S Fk Cross Creek Open T0/T1 Floodplain 960 36Wh0399 N Fk Cross Creek Open Slopes Middle Slopes 1200 36Wh0400 Trib Burgetts Fk Open T0/T1 Stream Bench 1140 36Wh0409 Raccoon Creek Open Slopes Middle Slopes 1230 36Wh0508 Raccoon Creek Open Upland Hill Ridge/Toe 1100 36Wh0512 N Fk Cross Creek Open Upland Hilltop 1240 36Wh0688 Raccoon Creek Open Upland Ridgetop 1300 36Wh0720 Chartiers Creek Open Slopes Upper Slopes 1340 36Wh0986 Raccoon Creek Open T0/T1 Terrace 1100 36Wh1002 Raccoon Creek Open Upland Ridgetop 1340 36Wh1096 Raccoon Creek Open Slopes Middle Slopes 1260 36Wh1112 Raccoon Creek Open Upland Upland Flat 1360 36Wh1115 Chartiers Creek Open Saddle Saddle 1300 36Wh1118 Trib Cross Creek Lithic Red. T0/T1 Terrace 1120 36Wh1148 -- Open Slopes Hillslope 1120 36Wh1151 -- Open T0/T1 Floodplain 1080 36Wh1152 Trib Cross Creek Open Slopes Hillslope 1200 36Wh1161 Trib Scott Run Open Saddle Saddle 1160 36Wh1201 -- Open Slopes Lower Slopes 1130 44 PA Prehistoric Data Synthesis, Raccoon Creek Watershed
As during the Early Archaic, Middle Archaic sites are located mostly within the Raccoon Creek drainage (n=29 sites), including 22 along Raccoon Creek itself and 7 along feeder streams. Eight sites are located within the Cross Creek watershed, including seven on feeder streams and one on Cross Creek. These data may indicate a contrasting settlement pattern between Cross Creek and Raccoon Creek, with Early Archaic sites being located on low-order tributaries in the former and on the main branch in the latter. Similar to the Early Archaic, access to water was crucial to site placement with most sites within 100m (mode/median) to 150m (mean) of a water source (Table 17). As during the preceding periods, access to stream confluences was not an important factor in Middle Archaic site placement. Most sites were located between 360m (mode) and 621m (mean) from a confluence (compared to 789m during the Paleoindian period). While Early and Middle Archaic sites were clearly more proximate to stream confluences than Paleoindian sites, this appears to be due to a general preference for locating sites near water. The Cross Creek survey report (Vento and Donohue 1982:124-127) provides the only other available information to understand Middle Archaic lithic raw material use in Watershed D. Lithic raw material data are provided for Middle Archaic points from Meadowcroft, Mungai Farm, as well as for sites in the Cross Creek drainage as a whole (Table 19). At Meadowcroft, Kanawha chert was preferred for Early and Middle Archaic points, while Upper Mercer chert was preferred at Mungai Farm. Seven different lithic raw materials were utilized for points at Mungai Farm, suggesting a wide-ranging settlement pattern, with travel and/or trade throughout southwestern Pennsylvania and east-central Ohio. Only three other Cross Creek sites contained diagnostic Middle Archaic points, with two produced from Flint Ridge chert and one from Uniontown chert. These patterns of lithic raw material use largely resemble those of the preceding Paleoindian and Early Archaic periods. At the time of this writing, investigations at the Leetsdale Site (Watershed G, Subbasin 20) had just been completed in Spring, 2003, with the identification of Middle Archaic occupations at the base of excavations (pers. comm., P.Funk, PHMC-BHP, 2003; Schuldenrein et al. 2003). However, no detailed information regarding this cultural component has been made available for dissemination due to the recent date of discovery. Results of excavations at Leetsdale will provide a rare window into early-mid Holocene lifeways in the Upper Ohio River Valley. D. MIDDLE ARCHAIC SETTLEMENT PATTERNS AND LITHIC RAW MATERIAL USE PASS files and research reports provide data on lithic raw material use during the Middle Archaic in Watershed D. Sixteen sites in PASS files are single-component Middle Archaic sites; thus, lithic raw material data from PASS files can be used for these sites without corruption by site occupation during other time periods. In addition, two sites Meadowcroft and Mungai Farm have specific lithic raw material data for Middle Archaic projectile points. Data from PASS files indicate a nearly complete reliance on chert/flint in stone tool production at the 16 single-component Middle Archaic sites. Onondaga chert, chalcedony, and slate were present at one site each, while two sites did not include information regarding lithic raw materials. None of the 16 single-component Middle Archaic sites in PASS files identified the use of Flint Ridge or Upper Mercer cherts (Table 18). While these data are limited, they suggest the use of chert, several varieties of which were locally available for Middle Archaic Native Americans. Onondaga chert was also available in secondary cobble form throughout regional drainages and should be considered a local lithic raw material. The presence of slate as a material used in tool production at Site 36Wh720 is the first evidence of its use in stone tool production. PA Prehistoric Data Synthesis, Raccoon Creek Watershed 45
SITE NO. Table 17. Middle Archaic Site Location Data, Watershed D (PASS Files) NEAREST WATER DIST. WATER DIRECT. WATER STREAM ORDER DIST. CONF. DIREC. CONF. 36Bv0003 Ohio River 120 Southwest 3 340 Northwest 36Bv0013 Raccoon Creek 150 Northwest 1 260 Southwest 36Bv0014 Raccoon Creek 60 Southeast 4 200 Northeast 36Bv0016 Raccoon Creek 100 Southwest 1 440 South 36Bv0017 Trib Raccoon Creek 0 On-site 1 360 Northeast 36Bv0019 Raccoon Creek 80 Northeast 4 170 Northwest 36Bv0020 Raccoon Creek 120 East 4 360 North 36Bv0022 Raccoon Creek 0 On-site 1 320 -- 36Bv0025 Raccoon Creek 160 Northeast 4 360 -- 36Bv0026 Raccoon Creek 60 South 1 570 -- 36Bv0036 Raccoon Creek 80 Southeast 1 1000 North 36Bv0037 Raccoon Creek 140 West 1 780 East 36Bv0039 Raccoon Creek 140 North 1 240 Southeast 36Bv0052 Trib Raccoon Creek 140 West 1 340 North 36Bv0107 Raccoon Creek 100 South 1 540 West 36Bv0123 Raccoon Creek 20 North 2 60 North 36Bv0133 Mill Creek 190 East 1 860 South 36Bv0162 Raccoon Creek 60 Northeast 4 240 Southeast 36Bv0178 Trib Raccoon Creek 160 South 1 1240 Southeast 36Bv0210 Raccoon Creek 100 West 2 140 Northwest 36Bv0221 Trib Raccoon Creek 100 East 1 880 Southeast 36Wh0106 Trib Burgetts Fork 50 North 1 710 East 36Wh0110 Burgetts Fork 400 Southeast 1 600 East 36Wh0187 S Fk Cross Creek 220 Northwest 1 1400 Northwest 36Wh0293 Raccoon Creek 0 On-site 3 50 Northwest 36Wh0297 Cross Creek 60 South 5 400 Southwest 36Wh0392 S Fk Cross Creek 20 Southeast 3 320 Northeast 36Wh0399 N Fk Cross Creek 280 Southeast 1 620 Southeast 36Wh0400 Trib Burgetts Fk 150 Southeast 1 700 Southeast 36Wh0409 Raccoon Creek 50 East 1 870 East 36Wh0508 Raccoon Creek 15 East 1 180 South 36Wh0512 N Fk Cross Creek 100 West 1 490 Northwest 36Wh0688 Raccoon Creek 380 Northwest 2 790 Southwest 36Wh0720 Chartiers Creek 260 East 1 1050 East 36Wh0986 Raccoon Creek 80 West 2 100 Southwest 36Wh1002 Raccoon Creek 375 West 1 1300 North 36Wh1096 Raccoon Creek 220 South 1 480 Northwest 36Wh1112 Raccoon Creek 60 North 1 1200 North 36Wh1115 Chartiers Creek 160 South 1 1140 Southwest 36Wh1118 Trib Cross Creek 100 Southeast 2 180 Southwest 36Wh1148 -- 310 West 1 390 West 36Wh1151 -- 260 Southwest 1 480 Northwest 36Wh1152 Trib Cross Creek 310 South 1 2120 Northeast 36Wh1161 Trib Scott Run 800 West 1 1980 North 36Wh1201 Other 10 East 1 700 Southwest 46 PA Prehistoric Data Synthesis, Raccoon Creek Watershed
Table 18. Cross-Tabulation of Single Component Middle Archaic Sites and Lithic Raw Materials, Watershed D (PASS Files) SITENO CHALCEDONY CHERT/FLINT ONONDAGA CHERT SLATE MAT. NOT IDENTIFIED TOTAL 36Bv0013 -- X -- -- -- 1 36Bv0020 -- X X -- -- 2 36Bv0039 -- X -- -- -- 1 36Bv0107 -- X -- -- -- 1 36Bv0178 -- X -- -- -- 1 36Wh0110 -- X -- -- -- 1 36Wh0392 -- -- -- -- X 1 36Wh0399 -- -- -- -- X 1 36Wh0400 -- X -- -- -- 1 36Wh0508 -- X -- -- -- 1 36Wh0512 -- X -- -- -- 1 36Wh0688 -- X -- -- -- 1 36Wh0720 X X -- X -- 3 36Wh0986 -- X -- -- -- 1 36Wh1002 -- X -- -- -- 1 36Wh1096 -- X -- -- -- 1 Total 1 14 1 1 2 30 Table 19. Cross-Tabulation of Cross Creek Site by Lithic Raw Material Type for Middle Archaic Artifacts (Vento and Donohue 1982:124-127). LOCATION ONONDAGA KANAWHA FLINT RIDGE UPPER M. BRUSH C. 10 MILE UNION- TOWN Meadowcroft*(%) 16.6 50.2 -- 16.6 16.6 -- -- -- 6 Mungai Farm(%) 8.3 16.7 8.3 20.8 12.5 12.5 12.5 8.3 24 Crs Cr. sites (%) -- -- 66.6 -- -- -- 33.3 -- 3 Total Points (n) 3 7 4 6 4 3 4 2 33 Total % 9.1 21.2 12.1 18.2 12.1 9.1 12.1 6.1 100 *For Meadowcroft, data is for Early and Middle Archaic points OTHER TOTAL PTS. E. MIDDLE ARCHAIC: CONCLUSION AND RESEARCH QUESTIONS M i d d l e A r c h a i c S u m m a r y The increase in site counts is the most notable difference between the Early Archaic and Middle Archaic; however, this increase disappears when the duration of each period is taken into consideration. Settlement patterns and site locations are also similar for the two periods, with Early and Middle Archaic Native Americans preferring use of a wide range of landscapes, including river and stream terraces, as opposed to uplands/slopes/saddles, which were preferred site locations during the Paleoindian period. Compared to the Paleoindian period, Early and Middle Archaic sites are located closer to water and closer to stream confluences, likely reflecting the movement of sites from uplands to lowlands near streams and rivers. Lithic raw material use was also consistent between the Early and Middle Archaic, with Native Americans of both periods using local cherts from southwestern Pennsylvania, as well as fair amounts of Kanawha chert, Flint Ridge chert, and Upper Mercer chert. Based on these data, the Middle Archaic continued trends initiated during the Early Archaic. Only the increase in site counts of the Middle Archaic separates it from the Early Archaic; however, as noted above, if site counts are averaged by total decades per period (Fiedel 2002), the Middle Archaic and Early Archaic have nearly identical site count densities, with 0.13 and 0.15 sites per decade, respectively, compared to.03 Paleoindian sites. These data in Watershed D, thus, contrast that presented by Carr (1998a:61; 1998b:88) who suggests a population increase between the Early and Middle Archaic periods. While this population PA Prehistoric Data Synthesis, Raccoon Creek Watershed 47
increase may have occurred in other portions of the state, such as the Upper Juniata sub-basin of south-central Pennsylvania (MacDonald 2003a), the Upper Ohio Valley appears to have sustained a population increase between the Paleoindian and Early Archaic periods, rather than between the Early and Middle Archaic. Middle Archaic Research Questions This summary of Middle Archaic archaeological data in and near Watershed D of the Ohio River Subbasin 20 has generated several research issues which should be considered when conducting archaeological work in the area. Twelve Middle Archaic research questions are listed below; this list is by no means comprehensive and should be used only as a starting point for generating additional research issues. Archaeological sites which can provide information pertaining to these and other research questions will likely meet the National Register Criterion D; thus, unless they lack integrity, sites which address these research questions will be eligible for listing in the National Register for Historic Places: 8. Are any other types of stone tools beyond bifurcate points diagnostic of Middle Archaic occupations? 9. Was the change in projectile point technology between the Early and Middle Archaic from notched and stemmed points to bifurcate points caused by a change in prey species and/or hunting strategies or are the changes simply stylistic? 10. What types of lithic raw materials were used during the Middle Archaic? 11. What do lithic raw materials tell us about Middle Archaic travel/trade patterns? 12. Why were patterns of lithic raw material use similar between the Paleoindian, Early Archaic, and Middle Archaic periods in Watershed D? Does this reflect the establishment and continued use of territories and trade networks between 12,000 and 5,300 years ago? 1. Does the Middle Archaic represent a distinct change from the Early Archaic or was there cultural continuity between the two periods? 2. Why were Paleoindian sites located further from water and stream confluences than Early and Middle Archaic sites? 3. Why did Early and Middle Archaic Native Americans move their sites from mostly in uplands (as during the Paleoindian period) to mostly in alluvial settings? 4. Was there a population increase or decrease (or neither) between the Early and Middle Archaic? 5. Was there a change from a forager to a collector subsistence strategy during the Middle Archaic or did it occur sooner (Early Archaic)? 6. What types of foods were procured during the Middle Archaic? 7. What types of tools were produced and utilized by Middle Archaic Native Americans? 48 PA Prehistoric Data Synthesis, Raccoon Creek Watershed
CHAPTER VII LATE ARCHAIC PERIOD 5300 to 3000 BP A. LATE ARCHAIC OVERVIEW The Late Archaic period witnessed several dynamic changes in population, culture, and environment in eastern North America. The Late Archaic is generally considered to persist between approximately 5,300 and 3,800 years ago (MacDonald 2002b). Subsequent to the Late Archaic, the Transitional, or Terminal, Archaic period persisted between 3,800 and 3,000 years ago (Raber 1985: 31). In western Pennsylvania, however, the Transitional Archaic is not welldefined, as several of the key diagnostic artifacts broadspears, pottery, steatite, and evidence of incipient agriculture are rare to virtually non-existent at sites in this region. As such, only two sites in PASS files include components dating to the Transitional Archaic. For the purposes of this report, thus, we include these sites with Transitional Period artifacts within the Late Archaic. During the Late Archaic, population increases are noted across the Middle Atlantic (Custer 1988) and Pennsylvania (Raber et al. 1998). The continued rise in sea levels due to environmental warming apparently increased available biomass (Turnbaugh 1977), including estuarine resources. According to Custer (1988; Custer and Wallace 1982) and Turnbaugh (1977), resulted in the increased availability of resources: 1) stimulated population growth; 2) increased use of logistic subsistence/settlement patterns; and 3) led to the widespread establishment of regional exchange networks. Late Archaic Native Americans developed a well-defined schedule of resource exploitation (Cowin 1991; Raber 1995, 1999, 2000) with the increased exploitation of riverine resources at sites such as East Steubenville on the Upper Ohio River (Lothrop 2001a; Lothrop and Mohney 2003; Mayer-Oakes 1955; Mohney 2002). The following chapter examines these Late Archaic issues within Watershed D, as reflected in PASS files data and regional research reports. B. LATE ARCHAIC MATERIAL CULTURE, SUBSISTENCE, AND CHRONOLOGY Late Archaic Material Culture Within western Pennsylvania, diagnostic artifacts of the Late Archaic period include Laurentian point types (Dragoo 1959; Kinsey 1972:403-408; Ritchie 1965), such as Otter Creek and Brewerton (Photograph 9, left), as well as Panhandle Archaic artifacts, such as Steubenville Stemmed and lanceolate points (Mayer-Oakes 1955; Mohney 2002) (see Photograph 9, right). Photograph 9. Brewerton Points (Left; from York County, Pa); and Steubenville Points (Right; from East Steubenville Site, 46Br31). Steubenville Point Photo Courtesy of West Virginia Division of Highways (Ritchie 1971/Mohney 2002) states that Steubenville stemmed and lanceolate points include a biconvex to lenticular cross-section with a medial maximum thickness. Lateral margins are ovate to excurvate, with some specimens having apparent parallel flaking. PA Prehistoric Data Synthesis, Raccoon Creek Watershed 49
Point bases are laterally, rather than basally, thinned. Following Boldurian (1985:288), Mohney states that Steubenville stemmed and lanceolate points are largely similar, with the differences due to secondary resharpening within the haft. This sharpening will eventually eliminate the stem, by reducing the point s blade width. Mohney (2002:24; Ritchie 1971) describes Brewerton points (Photograph 9, left)) as one of the most ubiquitous forms in the Upper Ohio Valley. The points are notched and produced via staged biface reduction. According to Justice (1987:115), Brewerton points have triangular blades with straight to convex bases. Two variants are known, side-notched and cornernotched, with the former perhaps representing a resharpened variant of the latter. Brewerton points likely served as the tips to atlatl darts (Mohney 2002:28). Within the Cross Creek drainage, Cross Creek Village (36Wh293) yielded blades in association with a Steubenville point (Applegarth and Cowin 1982). The coincidence of blades with the points suggests that blades were perhaps part of the Late Archaic technological repertoire. Seven blades were recovered at the site, predominantly of the prismatic type (Boldurian 1985:232). Transverse cross-sections of the blades are triangular and trapezoidal. Blade curvature is less than 5mm on the blades, which average 3.4 cm (length), 1.5cm (width), and 0.4cm (thickness) in dimension. Additional technological changes of the Late Archaic, especially the Panhandle Archaic, include increased use of groundstone tools, such as manos and metates and pitted cobbles, for food processing (Dragoo 1959; Mayer-Oakes 1955; Lothrop 2001a). These objects are often referred to as site furniture (Schiffer 1983), meaning that they were produced and used at a residence for an extended period or were curated at that location for future use. In consort with site furniture, large storage pits and food processing features, such as hearths and smudge pits, are also used more frequently during the Late Archaic. These artifacts and features, if recovered at archaeological sites, imply extended stays at one location by entire family groups. Sites with these features and artifacts are often interpreted as base camps from which Native Americans made daily forays for resource procurement, a typical characteristic of a logistic settlement pattern, as defined by Binford (1980). Late Archaic Chronology Late Archaic Radiocarbon Dates in Watershed D Within the Upper Ohio Valley and vicinity, Brewerton points have been recovered at sites dating between 6090 and 4150 BP, while Steubenville points have been found at sites dating between 4270 to and 3400 BP (Mohney 2002:47). These dates suggest very little overlap between Brewerton and Steubenville, with the former occurring in the early portion of the Late Archaic and the latter being transitional between Brewerton and the Early Woodland. In this sense, Steubenville may represent the lost Transitional period in the Upper Ohio Valley and vicinity. As discussed earlier, only two sites in Watershed D of Subbasin 20 yielded typical Transitional Period artifacts (e.g., steatite and broadspears). As such, researchers in the Upper Ohio Valley may want to consider including Steubenville/Panhandle Archaic sites within the Transitional Archaic rather than the Late Archaic. Few radiocarbon dates are available for Late Archaic occupations in Watershed D of Subbasin 20. The only dates from the watershed are from several features at Meadowcroft Rockshelter. Eleven uncalibrated radiocarbon dates from Stratum III and the upper portion of Stratum IIb indicate fairly extensive Brewerton and Transitional Period occupations at the site: 1 1 4 0 ± 1 1 5 B C ( 2 9 7 5-3 2 0 5 B P ; SI-2053); charcoal from firepits/upper 1/3 of Stratum III (F-18; Transitional/Terminal Archaic) 1 1 5 0 ± 9 0 B C ( 3 0 1 0-3 1 9 0 B P ; SI-3030); charcoal from firepits/upper 1/3 of Stratum III (F-18; Transitional/Terminal Archaic) 50 PA Prehistoric Data Synthesis, Raccoon Creek Watershed
1 1 6 5 ± 7 0 B C ( 3 0 4 5-3 1 8 5 B P ; SI-2044); charcoal from firepits/upper 1/3 of Stratum III (F-18; Transitional/Terminal Archaic) 1 3 0 5 ± 1 1 5 B C ( 3 1 5 4-3 4 7 0 B P ; SI-1679); charcoal from firepit/middle 1/3 of Stratum III (F-18) 1 2 6 0 ± 9 5 B C ( 3 1 1 5-3 3 0 5 B P ; SI-1681); charcoal from firepit/upper 1/3 of Stratum IIb (F-46 upper) 1 8 2 0 ± 9 0 B C ( 3 6 8 0-3 8 6 0 B P ; SI-1680); carbonized basketry fragment/upper 1/3 of Stratum IIb (F-46 upper) 2 0 0 0 ± 2 4 0 B C ( 3 7 1 0-4 1 9 0 B P ; SI-2063); charcoal from firepits/middle 1/3 of Stratum IIb (F-46 upper; Late Archaic) 2 0 2 0 ± 8 5 B C ( 3 8 8 5-4 0 5 5 B P ; SI-2058); charcoal from firepits/middle 1/3 of Stratum IIb (F-46 upper; Late Archaic) 2 0 5 5 ± 8 5 B C ( 3 9 2 0-4 0 9 0 B P ; SI-2054); charcoal from firepits/middle 1/3 of Stratum IIb (F-46 upper; Late Archaic) 2 4 3 0 ± 5 0 0 B C ( 3 8 8 0-4 8 8 0 B P ; SI-2354); charcoal from firepits/middle 1/3 of Stratum IIb (F-46 upper; Late Archaic) 2 8 7 0 ± 8 5 B C ( 4 7 3 5-4 9 0 5 B P ; SI-1685); charcoal from firepits/middle 1/3 of Stratum IIb (F-46 upper; Late Archaic) Late Archaic Radiocarbon Dates in Peripheral Areas While few sites within the study area have Late Archaic radiocarbon dates, two important sites adjacent to Watershed D, East Steubenville (46Br31) and Leetsdale (36AL480), were successful in providing dates. Mayer-Oakes (1955) reports dates of 2270 BC (4220 BP) on shell from Site 46Br31, indicating a Late Archaic age for the East Steubenville Site, only a few miles west of Watershed D in the northern Panhandle of West Virginia. relatively narrow window of time at the site (pers. comm., J. Lothrop, 2003): 3 8 0 0 ± 6 0 B P (Cal BP 4420 to 4080; Cal BC 2470-2130; Beta 163725) from Feature 13 (from deer bone collagen) 3 7 4 0 ± 6 0 B P (Cal BP 4400 to 3980; Cal BC 2450 to 2030; Beta 163727) from Feature 46 (from deer bone collagen) 3 6 8 0 ± 7 0 B P (Cal BP 4230 to 3840; Cal BC 2280 to 1890; Beta 163728) from Feature 60 (from charred nutshell) 3 3 5 0± 110 BP (Cal BP 3910 to 3390; Cal BC 1960 to 1440; Beta 163726) from Feature 49 (from deer bone collagen) 3 8 1 0 +4 0 B P (human bone, Feature 5) 3 7 8 0 +4 0 B P (human bone, Feature 5) 3 8 6 0 +4 0 BP (human bone, Feature 5) Eight additional dates are available from shell from East Steubenville. Conventional age ranges for the shell dates are 4630-5310 BP, while corrected age ranges are 3040-3720 BP. Finally, two dates are available from fish bone 4590+40 BP and 4240+40 BP indicating possible fish use during the Brewerton Site occupations. The Leetsdale site on the Ohio River just east of Watershed D also has yielded evidence of several Late Archaic Brewerton occupations. Recent excavations during the 2002-2003 field season yielded three features associated with Brewerton projectile points. Dates for the respective features are (pers. comm., P.Miller, 2003): 5 4 8 0± 50 BP (Feature 484/Brewerton levels) 5 6 0 0± 5 0 B P (Feature 488/Brewerton levels) 5 4 5 0± 4 0 B P (Feature 494/Brewerton levels) Conventional radiocarbon dates are available for Panhandle Archaic/Steubenville features excavated by GAI at 46Br31. The eight dates are listed below, and suggest a single occupation or several repeated occupations during a PA Prehistoric Data Synthesis, Raccoon Creek Watershed 51
Two additional Phase II dates are reported for a hearth feature in the Late Archaic component as well (Fenicle 2003). The dates of 3370±40 BP (cal BC 1760-1600; Beta 177514) and 2940±40 (cal BC 1280-1010; Beta 177992) from wood charcoal also suggest a substantial terminal Late Archaic occupation at the site, indicated also by the presence of steatite artifacts. Radiocarbon Dates in Other Areas of Western Pennsylvania In confirmation of these radiocarbon dates, which suggest temporal variation between Brewerton and Steubenville projectile points and associated occupations, several sites across Pennsylvania have similarly dated Late Archaic components. The Wiser Site (36Ce442) was excavated during expansion of S.R. 220 along North Bald Eagle Creek north of Altoona. This site yielded an extensive Late Archaic occupation with Brewerton and other Late Archaic points. Eight radiocarbon dates ranged between ca. 5200 and 3900 BP (East et al. 1999:7-44). One of the dates 5220± 70 (Beta-86174) was directly associated with a Brewerton side-notched point. The Memorial Park Site (36Cn164) in Lock Haven on the West Branch of the Susquehanna River yielded a variety of Late Archaic points from well-dated contexts (GAI 1995:191-202). Twenty features, including hearths and smudge pits, yielded radiocarbon dates of between ca. 5200 and 4900 BP for its late Laurentian component. The recovery of large numbers of groundstone tools confirmed that the site likely functioned as a base camp with a fairly permanent population. Excavations at the Canfield Island Site (36Ly37; Bressler 1989), approximately 15 miles downriver from Memorial Park on the West Branch, yielded an extensive Late Archaic component dated to ca. 5100 BP. Thirteen Brewerton points were recovered, as were features containing butternut and hickory nut shells. In addition, 12 netsinkers suggest that fishing was important to Late Archaic site occupants. Excavations at the Brown Site in Armstrong County, Pennsylvania, yielded Brewerton points radiocarbon dated to 4140 +240 BC (George and Davis 1986). Based on these radiocarbon dates and associated projectile points from sites in Pennsylvania and West Virginia, the following date ranges are likely for Late Archaic sites with Brewerton and Steubenville point associations in western Pennsylvania: Laurentian/Brewerton Sites: ca. 5300 to 4000 BP Panhandle Archaic/Steubenville sites: ca. 4000 to 3000 BP Late Archaic Subsistence Numerous researchers (Cowin 1991; Custer 1988; Raber 1995, 1999) propose generally similar Late Archaic models of land-use that entail an increase in logistic subsistence patterns within Pennsylvania. Rather than being foragers, thus, in which families moved frequently to obtain resources (as was the pattern during the preceding Paleoindian, Early Archaic, and Middle Archaic), Late Archaic Native Americans were collectors, in which families were seasonally sedentary at one location (base camps) for extended periods (Binford 1979, 1980). From these base camps (typically thought to be located along major waterways), individuals procured resources (plants, animals, lithics, etc.) during short trips to uplands and to low-mid-order tributaries. In western Pennsylvania, as discussed below, reliance on riverine resources increased during the latter portion of the Late Archaic (Lothrop 2001a). Terminal Archaic settlements in nearby regions also indicate the increased use of base camps for logistic subsistence (Custer 1996; Raber 1995). Limited Late Archaic ethnobotanical data are available for Subbasin 20. Cushman (1982: 217) provides evidence for fruit and seed use at Meadowcroft. Associated with dates of circa 52 PA Prehistoric Data Synthesis, Raccoon Creek Watershed
3700-4000 BP, Meadowcroft Stratum IIb yielded two Chenopodiaceae sp. (pigweed) fragments and two Rubus sp. (cloudberry) fragments, indicative of wild plant processing during the Steubenville portion of the Late Archaic. In Stratum III (associated with dates of circa 3200 BP, six Cheneopodiaceae sp., 41 Rubus sp., 11 Vaccinum sp. (cranberry), and one Cercis sp. (a deciduous shrub) fragment were recovered, along with basketry fragments. These data, while limited, suggest increasing plant use over time at Meadowcroft a pattern reflected elsewhere during the Archaic to Woodland transition (Smith 1987). While sites with subsistence remains are rare within the boundary of the study area, data from adjoining regions provides some clue as to potential food resources. Evidence from Panhandle Archaic/Steubenville point sites in the Upper Ohio Valley suggests a wide diet breadth, including hunting and gathering of shellfish, fish, turkey, and various small and large fauna. Mayer-Oakes (1955) reports that sites also yield a wide range of tool types, including bone and antler tools (fish hooks, bone points, and needles), as well as groundstone and cobble tools (netsinksers and hammerstones) and stone tools, including Steubenville points. The range of tool types also suggests a diverse array of site activities, likely related to food processing and procurement of a wide range of foods. More details regarding the East Steubenville Site are presented later in this chapter. Elsewhere in central and western Pennsylvania, excavations in Late Archaic levels at Memorial Park (GAI 1995:516) on the West Branch Susquehanna River yielded hickory, bitternut, hazelnut, butternut, walnut, acorn, grape, and elderberry. Faunal remains at sites in adjoining regions suggest the frequent use of fish and other aquatic resources. At Canfield Island in Williamsport, on the West Branch of the Susquehanna River, over 400 netsinkers were found in Terminal Archaic levels (Bressler 1989:46). Several caches of 16-30 netsinkers suggest the use of fairly large nets for largescale fish collection. Large platform hearth and rock pavements may be associated with the smoking, roasting, or drying of fish (Bressler 1989:72; Custer 1996; Wall et al. 1996). C. LATE ARCHAIC SITE TYPES AND LOCATION TRENDS A total of 69 sites in Watershed D have yielded diagnostic Late Archaic artifacts, an increase of 24 sites over the Middle Archaic. The increase in site counts is even more significant if the length of the respective periods is taken into consideration. The Middle Archaic (45 sites; 350 decades/3,500 years) yielded 0.13 sites per decade, while the Late Archaic (69 sites; 180 decades/1,800 years) yielded 0.38 sites per decade, an almost three-fold increase. Of the 68 Late Archaic sites, 46 yielded Brewerton points, 11 yielded Steubenville points, and 16 sites yielded unspecified Late Archaic diagnostic artifacts. Included in these totals are five sites that yielded both Brewerton and Steubenville points. One site also yielded a Savannah River Terminal Archaic point. As discussed above, Cowin (1991) and others (Custer 1988; Raber 1995, 1999), propose generally similar Late Archaic models of settlement that entail an increase in logistic subsistence patterns. In this subsistence system, Late Archaic Native Americans likely utilized a wider range of landforms. Base campsites should be in alluvial settings near large-order streams, while smaller resource extraction sites should be in upland settings or near small drainages. As shown in Table 24 and Table 25, and summarized in Tables 21-23, 29 sites in Watershed D are in alluvial settings, including stream and river terraces, floodplains, and benches. Slopes above streams (n=17), upland settings overlooking streams (n=15) and saddle settings (n=7) comprise the remainder. Access to water continued to be important during the Late Archaic (mean=126m to nearest water), while access to stream confluences (mean= 691) was of minimal importance. Mean elevation for Late Archaic sites was slightly higher than during the Middle Archaic (1,128 ft. amsl versus 1,052 ft.), suggesting a wider use of landforms. PA Prehistoric Data Synthesis, Raccoon Creek Watershed 53
Table 20. Late Archaic Sites, Watershed D (PASS Files) SITE NO. SITE NAME ARTIFACT COUNTY MUNICIPALITY 36Bv0003 Old Indian Fort Brewerton Beaver Shippingport Borough 36Bv0017 McElhaney Hog Farm Brewerton Beaver -- 36Bv0022 Boyscout Camp Steubenville Beaver Hopewell Township 36Bv0035 Wassler #1 Brewerton Beaver -- 36Bv0036 Wassler #2 Brewerton Beaver -- 36Bv0040 Crevallero Brewerton Beaver -- 36Bv0052 John Bush Farm Steubenville Beaver -- 36Bv0122 Crivallero Site Brewerton Beaver Independence Township 36Bv0171 Thompson #2 Brewerton Beaver Independence Township 36Bv0210 Lower Circle on Rock Late Archaic Beaver Raccoon Township 36Bv0221 Hilltop Brewerton Beaver Potter Township 36Bv0230 St. Joe #1 Brewerton Beaver Potter Township 36Bv0240 Dravo Brew/Steub Beaver Greene Township 36Bv0250 Thomas Pate Brewerton Beaver -- 36Wh0001 Moore Farm Late Archaic Washington Smith Township 36Wh0106 Mungai Farm Brew/Steub Washington Smith Township 36Wh0170 J Alrutz #2 Late Archaic Washington Mount Pleasant Township 36Wh0181 Reservoir Late Archaic Washington Mount Pleasant Township 36Wh0183 Nukon #1 Late Archaic Washington Smith Township 36Wh0187 Clair Cowden Late Archaic Washington Cross Creek Township 36Wh0293 Cross Creek Steubenville Washington Cross Creek Township 36Wh0297 Meadowcroft Brew/Steub Washington Jefferson Township 36Wh0313 Kaposey #2 Steubenville Washington Independence Township 36Wh0314 Kaposey #3 Steubenville Washington Independence Township 36Wh0348 MS #38 Late Archaic Washington Cross Creek Township 36Wh0349 MS #39 Late Archaic Washington Cross Creek Township 36Wh0355 MS #45 Late Archaic Washington Cross Creek Township 36Wh0357 MS #47 Late Archaic Washington Cross Creek Township 36Wh0365 Yee Farm Brew/Steub Washington Cross Creek Township 36Wh0374 Rex Smith Farm Steubenville Washington Cross Creek Township 36Wh0389 Carter Farm Brew/Steub Washington Jefferson Township 36Wh0390 Bertovich Farm Brewerton Washington Jefferson Township 36Wh0391 -- Brewerton Washington Jefferson Township 36Wh0398 MS #93 Late Archaic Washington Cross Creek Township 36Wh0401 MS #96 Brewerton Washington Independence Township 36Wh0422 FB #6 Late Archaic Washington Hopewell Township 36Wh0475 Berrinski Camp Brewerton Washington Mount Pleasant Township 36Wh0549 McNinch Site No. 1 Brewerton Washington Hanover Township 36Wh0565 Beadling No. 1 Brewerton Washington Hanover Township 36Wh0566 Beadling No. 2 Brewerton Washington Hanover Township 36Wh0608 Pershina Farm Late Archaic Washington Robinson Township 36Wh0757 Capuzzi #2 Late Archaic Washington Robinson Township 36Wh0821 Stacko Site Brewerton Washington Cross Creek Township 36Wh0992 Lowry #7 Brewerton Washington Mount Pleasant Township 36Wh1007 Carter #3 Brewerton Washington Mount Pleasant Township 36Wh1031 Pritts Brewerton Washington Mount Pleasant Township 36Wh1037 Cross Creek Park #1 Brewerton Washington Cross Creek Township 36Wh1093 Davidson Site Brewerton Washington Hopewell Township 36Wh1095 Scott Site Brewerton Washington Smith Township 36Wh1111 Hamilton Twins Site Brewerton Washington Hopewell Township 36Wh1112 Cumer Site Brewerton Washington Smith Township 36Wh1115 Murgel Site Brewerton Washington Mount Pleasant Township 36Wh1147 Brezinski #4 Brewerton Washington Hopewell Township 36Wh1148 Brezinski #5 Brewerton Washington Hopewell Township 36Wh1151 Lowry #11 Brewerton Washington Hopewell Township 36Wh1152 Vanzin Brewerton Washington Hopewell Township 36Wh1155 Luba #1 Brewerton Washington Hopewell Township 36Wh1156 Luba #2 Brewerton Washington Hopewell Township 54 PA Prehistoric Data Synthesis, Raccoon Creek Watershed
SITE NO. SITE NAME ARTIFACT COUNTY MUNICIPALITY 36Wh1161 Gregorski Brewerton Washington Jefferson Township 36Wh1191 Lowry #12 Brewerton Washington Mount Pleasant Township 36Wh1192 Lowry #13 Brewerton Washington Mount Pleasant Township 36Wh1193 Lowry #14 Brewerton Washington Mount Pleasant Township 36Wh1194 Lowry #15 Brewerton Washington Mount Pleasant Township 36Wh1195 Lowry #16 Sav. River Washington Mount Pleasant Township 36Wh1200 Vanzin #2 Brewerton Washington Mount Pleasant Township 36Wh1201 Vanzin #3 Brewerton Washington Mount Pleasant Township 36Wh1202 Martin #2 Brewerton Washington Mount Pleasant Township 36Wh1316 Herbst #1 Late Archaic Washington Mount Pleasant Township 36Wh1318 Cherry Run Terrace Late Archaic Washington Mount Pleasant Township Table 21. Cross-Tabulation of Site Setting by Nearest Water, Watershed D Sites (PASS data). NEAREST WATER ALLUVIAL SLOPES SADDLE OTHER UPLAND TOTAL Cross Creek 3 -- -- -- 3 S Fk Cross Creek -- 1 -- 1 2 N Fk Cross Creek -- 1 -- -- 1 Trib Cross Creek -- 6 2 1 9 Trib Scott Run -- -- 1 -- 1 TOTAL: CROSS CREEK WATERSHED 3 (18.8%) 8 (50.0%) 3 (18.8%) 2 (12.5%) 16 Raccoon Creek 12 4 1 6 23 Trib Raccoon Crk -- -- 1 4 5 Raccoon Run 1 -- -- -- 1 Little Traverse Creek 1 -- -- -- 1 Cherry Run 1 -- -- -- 1 Trib Burgetts Fork -- 1 -- 3 4 TOTAL, RACCOON CREEK WATERSHED 15 (42.9%) 5 (14.3%) 2 (5.7%) 13 (37.1%) 35 Chartiers Creek -- -- 2 -- 2 Ohio River 2 -- -- -- 2 Kings Crk (Ohio River Trib.) 1 -- -- -- 1 Other 9 4 -- -- 13 TOTAL, OTHER WATERSHED 12 (66.7%) 4 (22.2%) 2 (11.1%) 0 (0%) 18 GRAND TOTAL 30 (43.5%) 17 (24.6%) 7 (38.9%) 15 (21.7%) 69 Sites within Cross Creek and its tributaries are generally on slopes above creeks, likely due to the narrow floodplain and terrace settings along Cross Creek. In contrast, sites in the Raccoon Creek drainage are predominantly in alluvial settings or uplands, rather than slopes. The distribution of sites within the respective drainages of Watershed D likely reflects the topography of the drainages, rather than settlement pattern preferences. Within Raccoon Creek, wide valley bottoms and upland flats are common, but slopes between the two landforms are fairly steep and are generally inhospitable (see Figure 3 in Chapter II). As shown in Table 21, twice as many Late Archaic sites are located within the Raccoon Creek watershed than within Cross Creek, similar to the Early and Middle Archaic periods. Sites with Brewerton points are evenly distributed between alluvial settings (n=21) and uplands, saddles, and slopes (n=25; Table 22). Similarly, Steubenville sites are fairly evenly distributed between alluvial (n=6) and upland (n=4) settings. Sites with Steubenville points are evenly distributed between Cross Creek (n=4) and Raccoon Creek (n=6), while sites with Brewerton points are predominantly within Raccoon Creek (n=21) or other drainages (n=17), while only eight sites in Cross Creek yielded Brewerton points (Table 23). Overall, these data indicate a higher density of Brewerton occupation compared to Steubenville occupation during the Late Archaic in Watershed D of Subbasin 20. If the number of sites is averaged based on the duration of each sub-period, the results are similar, with Brewerton (5300-4000 BP; 130 decades; 46 sites) yielding 0.34 sites per decade compared to 0.16 for Steubenville (4000-3300 BP; 70 decades; 11 sites). PA Prehistoric Data Synthesis, Raccoon Creek Watershed 55
Table 22. Cross-Tabulation of Site Setting by Diagnostic Artifact Type at Late Archaic Sites, Watershed D Sites (PASS data) ARTIFACT ALLUVIAL SADDLE SLOPES UPLAND TOTAL Brew/Steub 2 -- 1 2 5 Brewerton 19 6 7 9 41 Late Archaic 5 1 8 2 16 Sav. River -- -- 1 -- 1 Steubenville 4 -- -- 2 6 Grand Total 30 7 17 15 69 Table 23. Cross-Tabulation of Nearest Water by Type of Diagnostic Artifact at Late Archaic Sites, Watershed D (PASS Files). NEAREST WATER BREW/STEUB BREWERTON STEUBENVILLE LATE ARCHAIC SAV. RIVER GRAND TOTAL Cross Creek Shed 1 7 3 5 -- 16 Raccoon Creek Shed 3 18 3 11 -- 35 Other 1 16 0 -- 1 18 Grand Total 5 41 6 16 1 69 Table 24. Late Archaic Site Location Data, Watershed D (PASS Files). SITE NO. SITE NAME SITE TYPE SETTING LANDFORM ELEV. NEAREST WATER 36Bv0003 Old Indian Fort Open T0/T1 Floodplain 660 Ohio River 36Bv0017 McElhaney Hog Farm Open Upland Hill Ridge/Toe 860 Trib Raccoon Crk 36Bv0022 Boyscout Camp Open T0/T1 Terrace 860 Raccoon Creek 36Bv0035 Wassler #1 Open T0/T1 Floodplain 800 Raccoon Creek 36Bv0036 Wassler #2 Open T0/T1 Terrace 920 Raccoon Creek 36Bv0040 Crevallero Open Upland Hill Ridge/Toe 920 Trib Raccoon Crk 36Bv0052 John Bush Farm Open Upland Hill Ridge/Toe 960 Trib Raccoon Crk 36Bv0122 Crivallero Open T0/T1 Terrace -- Little Trav. Cr. 36Bv0171 Thompson #2 Open T0/T1 Stream Bench 790 Raccoon Creek 36Bv0210 Lower Circle on Rock Open T0/T1 Terrace 960 Raccoon Creek 36Bv0221 Hilltop Open Upland Upland Flat 1100 Trib Raccoon Crk 36Bv0230 St. Joe #1 Open T0/T1 Terrace 710 Raccoon Creek 36Bv0240 Dravo Open T0/T1 Terrace 760 Ohio River 36Bv0250 Thomas Pate Open T0/T1 Floodplain 800 Raccoon Creek 36Wh0001 Moore Farm Open Slopes Middle Slopes 1020 Raccoon Creek 36Wh0106 Mungai Farm Open Slopes Middle Slopes 1220 Trib Burgetts Fork 36Wh0170 J Alrutz #2 Open Slopes Middle Slopes 1140 Trib Cross Creek 36Wh0181 Reservoir Open Upland Hill Ridge/Toe 1110 Raccoon Creek 36Wh0183 Nukon #1 Open T0/T1 Terrace 1050 Raccoon Creek 36Wh0187 Clair Cowden Open Slopes Lower Slopes 1340 S Fk Cross Creek 36Wh0293 Cross Creek Village Lithic Red. T0/T1 Terrace 1000 Raccoon Creek 36Wh0297 Meadowcroft Rockshelter T0/T1 Terrace 1000 Cross Creek 36Wh0313 Kaposey #2 Open T0/T1 Floodplain 780 Cross Creek 36Wh0314 Kaposey #3 Open T0/T1 Floodplain 780 Cross Creek 36Wh0348 MS #38 Open Upland Ridgetop 1120 Trib Cross Creek 36Wh0349 MS #39 Open Slopes Middle Slopes 1120 Raccoon Creek 36Wh0355 MS #45 Open Slopes Middle Slopes 1160 Raccoon Creek 36Wh0357 MS #47 Open Slopes Upper Slopes 1270 Trib Cross Creek 36Wh0365 Yee Farm Open Upland Ridgetop 1220 Trib Burgetts Fork 36Wh0374 Rex Smith Farm Lithic Red. Upland Ridgetop 1180 S Fk Cross Creek 36Wh0389 Carter Farm Open Upland Ridgetop 1260 Raccoon Creek 36Wh0390 Bertovich Farm Open Slopes Middle Slopes 1180 Trib Cross Creek 36Wh0391 -- Open Slopes Lower Slopes 1120 N Fk Cross Creek 36Wh0398 MS #93 Open Slopes Upper Slopes 1300 Raccoon Creek 56 PA Prehistoric Data Synthesis, Raccoon Creek Watershed
SITE NO. SITE NAME SITE TYPE SETTING LANDFORM ELEV. NEAREST WATER 36Wh0401 MS #96 Open Saddle Saddle 1180 Trib Cross Creek 36Wh0422 FB #6 Open Slopes Upper Slopes 1240 Trib Cross Creek 36Wh0475 Berrinski Camp Open T0/T1 Terrace 1080 Raccoon Creek 36Wh0549 McNinch Site No. 1 Open T0/T1 Stream Bench 960 Other 36Wh0565 Beadling No. 1 Open T0/T1 Stream Bench 940 Other 36Wh0566 Beadling No. 2 Open T0/T1 Terrace 920 Kings Crk 36Wh0608 Pershina Farm Village T0/T1 Terrace 1160 Raccoon Run 36Wh0757 Capuzzi #2 Open T0/T1 Terrace 940 Raccoon Creek 36Wh0821 Stacko Site Open Upland Ridgetop 1270 Raccoon Creek 36Wh0992 Lowry #7 Open Upland Ridgetop 1230 Raccoon Creek 36Wh1007 Carter #3 Open Saddle Saddle 1270 Trib Raccoon Crk 36Wh1031 Pritts Open Upland Ridgetop 1390 Trib Burgetts Fork 36Wh1037 Cross Creek Park #1 Open T0/T1 Terrace 1040 Raccoon Creek 36Wh1093 Davidson Site Open Upland Ridgetop 1280 Raccoon Creek 36Wh1095 Scott Site Open Upland Ridgetop 1280 Trib Burgetts Fork 36Wh1111 Hamilton Twins Site Open Slopes Middle Slopes 1120 Trib Cross Creek 36Wh1112 Cumer Site Open Upland Upland Flat 1360 Raccoon Creek 36Wh1115 Murgel Site Open Saddle Saddle 1300 Chartiers Creek 36Wh1147 Brezinski #4 Open T0/T1 Floodplain 1120 36Wh1148 Brezinski #5 Open Slopes Hillslope 1120 36Wh1151 Lowry #11 Open T0/T1 Floodplain 1080 36Wh1152 Vanzin Open Slopes Hillslope 1200 Trib Cross Creek 36Wh1155 Luba #1 Open T0/T1 Terrace 1140 Other 36Wh1156 Luba #2 Open T0/T1 Terrace 1160 Other 36Wh1161 Gregorski Open Saddle Saddle 1160 Trib Scott Run 36Wh1191 Lowry #12 Open T0/T1 Terrace 1100 Other 36Wh1192 Lowry #13 Open T0/T1 Floodplain 1080 Other 36Wh1193 Lowry #14 Open T0/T1 Stream Bench 1200 Other 36Wh1194 Lowry #15 Open Slopes Middle Slopes 1140 Other 36Wh1195 Lowry #16 Open Slopes Lower Slopes 1120 Other 36Wh1200 Vanzin #2 Open Saddle Saddle 1340 Trib Cross Creek 36Wh1201 Vanzin #3 Open Slopes Lower Slopes 1130 Other 36Wh1202 Martin #2 Open Saddle Saddle 1370 Chartiers Creek 36Wh1316 Herbst #1 Open Saddle Saddle 1290 Raccoon Creek 36Wh1318 Cherry Run Terrace Open T0/T1 Terrace 1260 Cherry Run Table 25. Late Archaic Site Location Data, Watershed D (PASS Files). SITE NO. NEAREST WATER DIST. WATER DIR. WATER DIST. CONF. DIR. CONF. 36Bv0003 Ohio River 120 Southwest 340 Northwest 36Bv0017 Trib Racc. Crk. 0 On-site 360 Northeast 36Bv0022 Raccoon Crk. 0 On-site 320 -- 36Bv0035 Raccoon Crk. 130 Northwest 740 Southeast 36Bv0036 Raccoon Crk. 80 Southeast 1000 North 36Bv0040 Trib Racc. Crk. 120 Northeast 180 -- 36Bv0052 Trib Racc. Crk. 140 West 340 North 36Bv122 Little Traverse Cr. -- -- -- -- 36Bv0171 Raccoon Crk. 0 East 900 West 36Bv0210 Raccoon Crk. 100 West 140 Northwest 36Bv0221 Trib Racc. Crk. 100 East 880 Southeast 36Bv0230 Raccoon Creek 0 On-site 960 Southeast 36Bv0240 Ohio River 140 Northwest 520 North 36Bv0250 Raccoon Crk. 20 Southeast 80 Southwest 36Wh0001 Raccoon Crk. 150 West 170 Southwest 36Wh0106 Trib Burgetts Fork 50 North 710 East 36Wh0170 Trib Cross Creek 140 Southeast 250 South 36Wh0181 Raccoon Crk. 100 Southwest 175 South 36Wh0183 Raccoon Crk. 160 Southwest 450 Northwest 36Wh0187 S Fk Cross Creek 220 Northwest 1400 Northwest PA Prehistoric Data Synthesis, Raccoon Creek Watershed 57
SITE NO. NEAREST WATER DIST. WATER DIR. WATER DIST. CONF. DIR. CONF. 36Wh0293 Raccoon Crk. 0 On-site 50 Northwest 36Wh0297 Cross Creek 60 South 400 Southwest 36Wh0313 Cross Creek 20 Northwest 60 Southwest 36Wh0314 Cross Creek 20 West 300 East 36Wh0348 Trib Cross Creek 140 Southeast 560 South 36Wh0349 Raccoon Crk. 80 Northwest 720 South 36Wh0355 Raccoon Crk. 140 East 810 South 36Wh0357 Trib Cross Creek 120 Southwest 1020 South 36Wh0365 Trib Burgetts Fork 330 Southeast 860 Northeast 36Wh0374 S Fk Cross Creek 5 North 490 West 36Wh0389 Raccoon Crk. 0 East 750 Northwest 36Wh0390 Trib Cross Creek 90 Southwest 840 Southwest 36Wh0391 N Fk Cross Creek 0 -- 150 East 36Wh0398 Raccoon Crk. 200 South 1020 Northeast 36Wh0401 Trib Cross Creek 0 On-site 830 Southeast 36Wh0422 Trib Cross Creek 150 Southeast 510 North 36Wh0475 Raccoon Crk. 70 West 120 Southeast 36Wh0549 Other 100 Northwest 120 West 36Wh0565 Other 100 West 220 Northeast 36Wh0566 Kings Cr. 20 West 120 Northwest 36Wh0608 Raccoon Run 115 West 150 Southwest 36Wh0757 Raccoon Crk. 120 South 255 Southwest 36Wh0821 Raccoon Crk. 280 Southwest 940 Northeast 36Wh0992 Raccoon Crk. 400 Southeast 700 Southwest 36Wh1007 Trib Racc. Crk. 180 South 675 West 36Wh1031 Trib Burgetts Fork 400 Northwest 970 Southwest 36Wh1037 Raccoon Crk. 0 On-site 130 South 36Wh1093 Raccoon Crk. 290 Southeast 440 North 36Wh1095 Trib Burgetts Fork 330 Northwest 1100 Southwest 36Wh1111 Trib Cross Creek 90 North 6880 Southwest 36Wh1112 Raccoon Crk. 60 North 1200 North 36Wh1115 Chartiers Creek 160 South 1140 Southwest 36Wh1147 -- 160 South 460 Northwest 36Wh1148 -- 310 West 390 West 36Wh1151 -- 260 Southwest 480 Northwest 36Wh1152 Trib Cross Creek 310 South 2120 Northeast 36Wh1155 Other 20 Northwest 320 Northeast 36Wh1156 Other 80 Northeast 340 North 36Wh1161 Trib Scott Run 800 West 1980 North 36Wh1191 Other 10 North 180 Southwest 36Wh1192 Other 20 South 300 Southeast 36Wh1193 Other 60 West 800 Southwest 36Wh1194 Other 10 West 700 Southwest 36Wh1195 Other 60 East 500 Southwest 36Wh1200 Trib Cross Creek 120 Southwest 540 Northwest 36Wh1201 Other 10 East 700 Southwest 36Wh1202 Chartiers Creek 20 Southwest 1400 Northwest 36Wh1316 Raccoon Crk. 100 -- -- -- 36Wh1318 Cherry Run 100 -- -- -- Key Late Archaic Sites in Watershed D Research reports provide additional data regarding Late Archaic settlement patterns and site types. Key Late Archaic sites within Watershed D include a series of Panhandle Archaic sites identified by Mayer-Oakes (1955:139) within Beaver County. 36Bv17 McElhaney Hog Farm Site 36Bv17, the McElhaney Hog Farm Site, yielded grooved adzes in association with large numbers of Steubenville stemmed and lanceolate points (Mayer-Oakes 1955:141, 144). The site is located at an elevation of 860 ft. amsl on an upland hill ridge/toe overlooking a tributary of Raccoon Creek near the town of 58 PA Prehistoric Data Synthesis, Raccoon Creek Watershed
Aliquippa, Pennsylvania location typical of many Steubenville sites, including East Steubenville. The site likely served as a camp during the Late Archaic, the remains of which were extensively collected by Emil Alam (Mayer-Oakes 1955). 36Bv52 John Bush Farm Site 36Bv52, the John Bush Farm Site, also yielded grooved adzes in association with large numbers of Steubenville stemmed and lanceolate points (Mayer-Oakes 1955:141, 144). The site is located only a few miles southeast of 36Bv17 (see above) and was also extensively collected by Emil Alam and studied by Mayer- Oakes (1955). As with 36Bv17, 36Bv52 is located on an upland hill ridge/toe overlooking a tributary of Raccoon Creek a few miles south of Aliquippa. This site likely served as a resource procurement camp during the Late Archaic. 36Bv122 Crivallero Site 36Bv122, the Crivallero Site, is located on Little Traverse Creek near its confluence with Raccoon Creek. Excavated by Berger (1998) for PennDOT along SR 30 near Raccoon Creek State Park, the site yielded a small lithic assemblage (n=242) during Phase I and II excavations. Four Brewerton points were recovered, as were an Adena stemmed and five untyped point fragments. No features were identified, but the prevalence of Brewerton points indicates that it was the most substantial occupation. Because of the low artifact density, the lack of features, and the presence of multiple unstratified components, the site was not deemed eligible to the National Register. 36Bv230 St. Joe #1 In 1986, NPW Consultants, Inc. (Cosgrove and Michael 1986) conducted Phase I/II testing at four sites for the St. Joe Resources Company fly-ash disposal project in Potter Township, near Monaca. Site 36Bv230 was identified on a first terrace of Raccoon Creek via the discovery of 22 lithic artifacts, including one Brewerton sidenotched point and one untyped side-notched point base. Lithic raw materials included Onondaga and Uniontown chert. 36Bv240 Dravo The Dravo Site (36Bv240; Davis 1988) is located on a high upper gravel terrace of the Ohio River just west of the town of Georgetown. The site yielded an extensive lithic assemblage, including more than 14,000 artifacts, much of which was related to the reduction of locally collected secondary chert cobbles during the Laurentian Late Archaic. More discussion of Late Archaic lithic raw material use at the site is discussed below. 36Wh106 Mungai Farm The Mungai Farm Site (36Wh106) is located at an elevation of 1,220 ft. amsl overlooking the Burgetts Fork south of Burgettstown. The upland site yielded the largest collection of Late Archaic projectile points of any site within Watershed D. The site was excavated by the University of Pittsburgh during their Cross Creek survey (Fitzgibbons 1982:108-109). Together with the surface collection by the landowner, the site has yielded 52 Brewerton Corner-Notched, 39 Brewerton Side-Notched, 53 Brewerton-like, one Buffalo Stemmed and four Steubenville Stemmed points. Interestingly, few debitage or other tools were recovered during excavations or surface collection of the site (Fitzgibbons 1982:108). 36Wh293 Cross Creek Village Cross Creek Village (36Wh293) yielded a Steubenville point in horizontal association with 11 possible house structures and evidence of blade production (Applegarth and Cowin 1982; Boldurian 1985). The site is located on a floodplain/terrace above Cross Creek, approximately 5 miles east of Avella (see Photograph 3). Morphologically, the blades from Cross Creek Village resemble those from the Paleoindian levels of Meadowcroft, suggesting similar technological origins (1985:237). One difference was in platform preparation, however, with Meadowcroft blades being prepared and generally multifaceted and Cross Creek Village blades being unprepared and flat. Raw material type was also distinct between the sites, with the Meadowcroft blades being produced from a variety of local cherts PA Prehistoric Data Synthesis, Raccoon Creek Watershed 59
(including Monongahela, Ten Mile, Upper Mercer, Brush Creek and Uniontown) and the Cross Creek Village blades being produced only from Monongahela chert (Boldurian 1985:231-236) which is available from an outcrop near the site (Mohney 2002:13). The differences in raw material type likely reflect its differential availability, rather than the technological preferences of individuals at the two sites. The presence of 11 possible house structures at Cross Creek Village and the lack of blades from other Steubenville sites in the region raise doubts that Cross Creek Village is a single component site. If this is the case, and multiple components are represented at the site, then the age of the blades is in doubt and may be related to another occupation of the site not represented by diagnostic artifacts. The coincidence of structures and blades at Cross Creek Village may reflect a Middle Woodland age of the site, with the Steubenville point possibly representing an earlier occupation disturbed during structure construction during the Middle Woodland occupation. 36Wh297 Meadowcroft No specific descriptions of the Late Archaic component at Meadowcroft have been published, although compilations of point types by period indicate that the Late Archaic occupations were among the most intensive of any of the site visits (Adovasio et al. 1977). Seven Brewerton, five Buffalo Stemmed, and four Steubenville Stemmed-like points were recovered at the site, predominantly concentrated in Strata IIb and III. Radiocarbon dates for the Late Archaic occupations span the entire Late Archaic period, including earlier Brewerton and later Steubenville point occupations (see radiocarbon dates above). Key Late Archaic Sites near Watershed D 36Wh351 Krajacic While not as well known as the possible Pre- Clovis assemblage from the site, the Late Archaic Steubenville assemblage was fairly extensive at the Krajacic Site (36Wh351). As discussed in the Paleoindian chapter, the site is on an upland hilltop and was identified by the University of Pittsburgh during the Cross Creek project (Carlisle and Adovasio 1982; Boldurian 1985). The Krajacic Site yielded a wide variety of artifacts from different time periods within a plowed field. No features were identified and no radiocarbon dates were obtained from the site. Artifacts include Miller points in various stages of reduction along with stemmed and lanceolate Steubenville points. At Krajacic, the surface association of blades with both Steubenville and Miller points raises significant issues regarding the validity of the supposed Miller point and associated blades as a distinct early Paleoindian toolkit (Gardner 2002). 36AL480 Leetsdale Site 36AL480 is on the northern shore of the Ohio River, within Watershed G of Subbasin 20 less than 1 mile northeast of Watershed D (Photograph 10). The site is approximately 13 miles upstream from the mouth of Raccoon Creek and less than 5 miles east of Raccoon Creek as it flows near Aliquippa. Excavations at Site 36AL480 were conducted between 2001-2003 by several archaeological consultants working for the USACOE; excavations yielded multiple stratified components dating to the Early Archaic, Middle Archaic, Late Archaic, Transitional Archaic, and Early Woodland. Based on data available to date, a Brewerton Late Archaic component was identified within Stratum III, including at least four Brewerton points and a possible open-faced hearth (Hardlines 2002:25). More recent excavations in Area 3 have identified three Late Archaic occupations, including one which contained shellfish remains. During much of the Late Archaic period, the site was an island in the Ohio River. The abundance of flaking debris in the Late Archaic component indicates that Native Americans were collecting and reducing chert cobbles at the site (www.lrp.usace.army.mil/ lmon/late_archaic.htm). Radiocarbon dates of 5480±50 BP, 5600±50 BP, and 5450±40 BP were assayed from three Late Archaic features 60 PA Prehistoric Data Synthesis, Raccoon Creek Watershed
in Area 3 (pers. comm., P. Miller, KCI Technologies, Inc.). Photograph 10. Aerial Photograph of the Leetsdale Site. View Northeast (www.lrp.usace.army.mil/.htm). Photograph courtesy of U.S. Army Corps of Engineers. The site yielded one of the few assemblages of steatite artifacts in western Pennsylvania, and several projectile points more commonly found at sites in the lower Ohio River Valley and the Wabash River Valley of Illinois (Frye and Weiser 2003: 4-5). The presence of these artifacts in the Late Archaic/Transitional Archaic component indicates widespread regional interaction networks within eastern North America during the terminal portion of the Late Archaic period, approximately 4,000-3,000 years ago. West Virginia Panhandle Sites Lothrop (2001) reports that at least eight Panhandle Archaic sites, some with associated shell middens, have been recorded a few miles west of the current study area in the panhandle of West Virginia. Settings are variable, including three sites situated on first terraces above the Ohio River, one site located on a high T2/T3 Pleistocene terrace, and four sites situated on intermediate or higher Pleistocene terrace straths above the valley bottom. Mayer-Oakes (1955:132) reports on four of these Steubenville sites in the West Virginia panhandle along the Upper Ohio River. Three of the sites East Steubenville (46Br31), New Cumberland Heights (46Hk1), and Globe Hill (46Hk34) were upland ridgetop sites overlooking the Ohio River, while the Half- Moon Site (46Br29) was a large site with Steubenville points along the Ohio River bottomland south of Weirton. McKees Rocks mound (36Al16), overlooking the mouth of Chartiers Creek and the Ohio River, also yielded grooved adzes and Steubenville points. This distribution of sites as presented in Mayer- Oakes (1955) indicates a preference for Panhandle Archaic sites in upland settings, at least along the Ohio River. Nevertheless, one site 46Br29 was in an alluvial setting, confirming the data presented in Lothrop (2001) that indicate a wider use of landforms, including bottomlands, upper terraces, and ridgetops. This cluster of Steubenville sites confirms PASS file data which indicate a wide range of site locations during the Steubenville or Panhandle Archaic portion of the Late Archaic period. Recent excavations at the East Steubenville Site (46Br31), excerpted below from Lothrop s (2001a) management summary, reinforce this interpretation and indicate a wide diet breadth for Late Archaic Native Americans. 46Br31 East Steubenville Excerpted from Lothrop (2001a) The East Steubenville Site (46Br31) is the type site for the Panhandle Archaic, as described above. The site is located on a south-trending ridgespur overlooking the Ohio River near Weirton, West Virginia (Photograph 11). Ohio River pool elevation at this location is 644 feet (196 meters), while Site 46Br31 occupies the south end of the ridgespur at an elevation of 900 feet (274 meters) amsl. Mahan Run, a low-order stream, empties into the Ohio River immediately south of the ridgespur. The site is located only five miles west of the Pennsylvania state line and the current study area. As first reported by Mayer-Oakes (1955), avocational archaeologists discovered the East Steubenville Site in 1938, and intermittently conducted informal excavations thereafter; Mayer-Oakes report (1955) was based on their field notes and artifact collections. These PA Prehistoric Data Synthesis, Raccoon Creek Watershed 61
excavations identified a shell midden used by Late Archaic Native Americans. The site also yielded other food refuse, including bones of deer and other mammal, turkey, fish, birds and turtle. Excavations in the shell midden produced tools of perishable materials rarely preserved on other prehistoric sites in the region, including bone awls and harpoons of antler and bone. Ohio River 46Br31 Photograph 11. Aerial Photograph of the East Steubenville Site. View North. Courtesy of West Virginia Division of Highways. Stone tools from the site included Steubenville spearpoints and other flaked stone tools of Onondaga and Flint Ridge chert, as well as groundstone tools consisting of grooved adzes and axes, and bannerstone spear-thrower weights. A human burial and two dog burials were reportedly found at the site. Based on these discoveries at Site 46Br31 and similar finds at three other nearby sites, Mayer-Oakes defined the Panhandle Archaic Complex as a distinctive riverine adaptation of prehistoric Native Americans in the Upper Ohio Valley, dating sometime between 5000 and 3000 years ago. Three years after Mayer-Oakes report on the East Steubenville Site, Crane and Griffin (1958) reported a radiocarbon date of 2270 BC (4220 BP) on shell from the site. In 1999-2000, GAI conducted data recovery excavations at Site 46Br31 (East Steubenville) for the West Virginia Division of Highways. During Phase III work, GAI excavated 86 STPs, excavated 316 1x1-m test units, and stripped the entire ridgetop landform to identify features. These excavations resulted in the identification of 74 cultural features, including 5 human burials and 50 pit features. Of particular note among the 50 pit features found at Site 46Br31 are three features tentatively interpreted as pits for steaming freshwater mussel. Other cultural feature types include 10 FCR clusters, one postmold, two lenses of preserved shell midden along the west flank of the ridge, six miscellaneous features (including a tabular anvil stone), and the five human burials. The 10 FCR clusters consist of rock concentrations in which many of the rocks appear to been reddened or cracked by heat; in some cases, these FCR clusters also contained concentrations of charcoal. These features are tentatively interpreted as small hearth platforms. Four feature dates place site occupation between 3,400 to 3,840 years ago (see dates above). GAI archaeologists excavated the five human burials at Site 46Br31 between November 6 and November 17, 2000. Disarticulated human remains were also recovered from feature 92, a disturbed deposit discovered during stripping on the west flank of Site 46Br31 (see below). These burials include four pit interments located on the eastern flank of the ridge (Features 1, 37, 46, and 62), and a single pit burial, Feature 5, positioned along the southwestern margin of the ridge. In addition, all five burials are located along or within a few meters of where slopes on the ridge margins increase from approximately 20 to 35 percent. As one final observation on trends in burial location, Features 5, 37, 46 and 62 are all situated within the approximate former boundary of shell midden at the site; only Feature 1 lay outside the original distribution of shell discarded the site, situated further to the north on the eastern flank. Based on field counts, data recovery excavations at Site 46Br31 produced 52,419 lithic artifacts, 14,698 shell, and 3464 faunal remains (see Table 2). These artifact totals include large numbers of stone tools (383 bifaces, 37 62 PA Prehistoric Data Synthesis, Raccoon Creek Watershed
unifaces) and approximately 241 cobble and groundstone tools. The biface assemblage includes more than 200 projectile points; of these, most were Steubenville projectile points, while perhaps one-quarter represent Brewerton notched points. The recovery of several Brewerton points in pit features containing shell at Site 46Br31 suggest that shellfish procurement may have begun during the presumed earlier Brewerton component. According to Jon Lothrop of GAI (pers. comm.. 2003), artifact analysis (Dr. Kenneth Mohney, Renee Sobota, and Lisa Dugas) has revealed a diverse technology consisting of flaked stone, ground stone, cobble, bone and shell implements. As noted, hunting implements include Steubenville stemmed and lanceolate, and Brewerton projectile points. The nondiagnostic, flaked stone, assemblage is dominated by bifaces, including "unfinished" bifaces, drills, and an unusual biface tool tentatively termed square-bit bifaces. Preliminary microwear analysis of this last tool class by Dr. Melody Pope indicates scraper use on animal hides and other materials. Bifaces classified as unfinished early-, middle-, and latestage bifaces, at first assumed to represent point manufacturing rejects, also bear microwear traces of use as processing tools. Early-stage bifaces were employed for plant processing or woodworking, while middle- and late-stage bifaces show evidence of hide working. The inventory of ground stone tools includes threequarter-grooved adzes and celts, suggesting both heavy and light woodworking tasks. Cobble tools, consisting of hammerstones, pitted stones, manos and metates suggest plant and lithic processing activities. Bone and shell tools are also represented--these appear to be used for scraping activities and other tasks; their presence as an alternative technology may explain the scarcity of unifacial and expedient stone tools in this unusual assemblage. Studies of freshwater mussel shell (Dr. Harold Rollins, Lisa Dugas) show that Panhandle Archaic Americans collected 26 species. Habitat analysis shows that most species were harvested from large river, shallow water settings (likely from shoals in the Ohio River) 300 feet below the ridgetop setting of the East Steubenville site. Shellfish were probably harvested during the summer, when the river was at its lowest ebb. Faunal analysis by Marie-Lorraine Pipes and Steven Thomas reveals a diverse food base at East Steubenville, including 16 species of mammal (dominated by deer), 8 species of fish, 5 species of bird, 5 of reptile and 2 of amphibia. Deer were likely taken in fall and early winter. Among fish remains, channel catfish were most common, caught during middle and late spring. Fish comprise 38 percent of the faunal assemblage, emphasizing, along with shellfish, the importance of riverine resources in the Panhandle Archaic diet. Studies of botanical remains by Justine McKnight revealed that mast (including hickory, black walnut, and acorn) was commonly collected during the fall for consumption. Seeds of fleshy fruits and small grains are rare. Notably, there is no evidence of cultigens at the East Steubenville site, a striking contrast to sites of the contemporaneous Shell Mound Archaic in Kentucky; this absence of cultigens may be a diagnostic trait of plant use in the Panhandle Archaic. Osteological studies of human remains at the site by Dr. Paul Sciulli recognized six individuals, including one from a looted burial. All were adult, and analysis identified two males and two females, with ages at death ranging from 20-30 to 45-55. Common pathologies included severe tooth wear, presumably from a gritty diet, and degenerative joint disease, reflecting the rigors of the Panhandle Archaic lifeway. Situated on a ridgetop, the East Steubenville site is of moderate size, occupying 6400 sq. meters. Across this landform, shell midden and most features were distributed along the east and west flanks, areas for food resource processing; lithic artifacts and debris on the ridgecrest likely define general habitation and stone working areas. The site appears to represent a composite of many reoccupations by groups of modest PA Prehistoric Data Synthesis, Raccoon Creek Watershed 63
size. Seasonality data suggests Native American encampments during the spring, summer and fall/early winter; however, moderate artifact densities and lack of evidence for structures at East Steubenville suggests these occupations may have been typically intermittent rather than continuous in any given year. This scenario may signal increased redundancy of site use at resource-rich locations along the Ohio River during the latter portion of the Late Archaic. Evidence suggests that Native Americans at this time also occupied short-term, seasonal encampments outside the Ohio Valley proper as part of their settlement round. D. LATE ARCHAIC SETTLEMENT PATTERNS AND LITHIC RAW MATERIAL USE PASS files and research reports provide data on lithic raw material use during the Late Archaic in Watershed D of Subbasin 20. Twenty sites in PASS files are single-component Late Archaic sites, 17 of which provide data regarding lithic raw material use. In addition, several research reports provide more specific information useful in reconstructing Late Archaic lithic raw material use. Data from PASS files indicate that 13 of the 17 single-component Late Archaic sites (with lithic raw material data) yielded chert/flint artifacts. Flint Ridge and Onondaga cherts were used at three and two sites respectively, while chalcedony and sandstone were also recovered at one Late Archaic site each. None of the single-component Late Archaic sites in PASS files identified Upper Mercer chert or steatite, the latter of which is a hallmark of the terminal portion of the Late Archaic (Transitional Archaic). While these data are limited, they suggest the use of chert, several varieties of which were locally available in and near Watershed D. Onondaga chert was also available in secondary cobble form throughout regional drainages and should be considered a local lithic raw material. Late Archaic lithic raw material use data are also available in research reports from Meadowcroft, Mungai Farm, Dravo, St. Joe #1, and Crivallero in Watershed D, as well as the East Steubenville Site within a few miles of the watershed. All types of Late Archaic projectile points are lumped together for the Cross Creek studies, thus there is no means to distinguish differences in lithic raw material use and settlement patterns between Brewerton and Steubenville site occupations. Table 26. Cross-Tabulation of Site by Lithic Raw Material (Presence or Absence) at Single Component Late Archaic Sites (PASS Files; X=present; --=absent). SITENO Onondaga Chert/Flint Flint Ridge Chert Chalcedony Sandstone Not Identified 36Bv0035 X -- -- -- -- -- 36Bv0171 X -- -- -- -- -- 36Wh0170 -- -- -- -- -- X 36Wh0183 -- -- -- -- -- X 36Wh0348 X -- -- -- X -- 36Wh0355 X -- -- -- -- -- 36Wh0357 X -- -- -- -- -- 36Wh0398 -- -- -- -- -- X 36Wh0422 X -- -- X -- X 36Wh0549 X -- -- -- -- -- 36Wh0566 X -- -- -- -- -- 36Wh1192 X -- -- -- -- -- 36Wh1193 -- -- X -- -- -- 36Wh1194 X X -- -- -- -- 36Wh1195 X X X -- -- -- 36Wh1200 X X -- -- -- -- 36Wh1202 X -- -- -- -- Total Sites 13 3 2 1 1 4 64 PA Prehistoric Data Synthesis, Raccoon Creek Watershed
Table 27. Cross-Tabulation of Cross Creek Site by Lithic Raw Material Type for Late Archaic Artifacts (Vento and Donohue 1982:124-127). LOCATION ONON- DAGA KANA- WHA FLINT RIDGE UPPER M. BRUSH C. 10 MILE UNION- TOWN GLACIAL PEBBLE Meadowcroft (%) 6.1 29.0 12.6 18.4 15.1 7.7 1.2 2.4 7.6 245 Mungai Farm (%) 15.9 7.2 11.1 15.9 15.9 15.9 12.6 2.4 2.5 207 Cross Cr Sites (%) 19.4 6.5 6.5 6.5 13.0 32.2 3.2 6.5 6.5 31 Total Points (n) 54 88 56 80 74 62 30 13 26 483 Total % 11.2 18.2 11.6 16.6 15.3 12.8 6.2 2.7 5.4 100 OTHER TOTAL PTS Nevertheless, the Cross Creek studies, as shown above in Table 27, indicate the use of a variety of cherts, including (in order of preference) Kanawha, Upper Mercer, Brush Creek, 10 Mile, Flint Ridge, Onondaga, Uniontown, and glacial pebble. Local cherts and exotic cherts were used, indicating curation and replacement of broken exotic chert points with points produced from local stones. The high incidence of Flint Ridge and Upper Mercer cherts indicates travel and trade to east-central Ohio, while use of Kanawha chert indicates travel to north-central West Virginia. Recent studies of Kanawha chert artifacts by Brian Fritz (pers. comm. 2003; Mohney and Lothrop 2003) have indicated possibly local sources of Kanawha-like chert in southwestern Pennsylvania; given the high incidence of the material at the Cross Creek sites, a local source for the material is likely. Excavations at the Dravo site (36Bv240) revealed more than 14,000 lithic artifacts from an upper terrace of the Ohio River near Georgetown (Davis 1988, n.d.; Davis and Lantz 1987). Phase I/II excavations at the site yielded five Steubenville points and three Brewerton points (Davis and Lantz 1987), while Phase III studies yielded 27 Laurentian points (Brewerton and Otter Creek), but no Steubenville points. Fourteen of the Laurentian points were produced from local Onondaga cobble chert, with the remainder produced from Delaware, Upper Mercer/Coshocton, Flint Ridge and untyped gray chert (Davis 1988:65). At the St. Joe #1 Site (36Bv230) on Raccoon Creek near Monaca, Cosgrove and Michael (1986) identified a small lithic scatter with a Brewerton side-notched point. Lithic raw materials included 7 flakes of Onondaga pebble chert and one Uniontown chert flake, suggesting use of locally collected lithic raw materials. Berger (1988) conducted Phase I and II excavations at the Crivallero Site (36Bv122) near the confluence of Little Traverse Creek and Raccoon Creek. This site yielded 242 lithic artifacts, including 10 projectile points. Diagnostic artifacts were restricted to four probable Brewerton points and one Early Woodland Adena stemmed point. While several chert types, including Flint Ridge, Onondaga, Uniontown, Upper Mercer, Shriver, and jasper were identified in the assemblage, no tables presenting the counts and percentages of each type were presented in the report. Isolated Onondaga chert Brewerton projectile points were recovered during two Phase I surveys in Watershed D. One of the points was recovered on an upland near the mouth of Raccoon Creek and the Ohio River (Baker 1996), while the other was recovered during STP excavation along a small creek in Raccoon Creek state park (Eddins 1982:52). Excavations in the Late Archaic components of the East Steubenville Site, a few miles west of Watershed D, indicated differential use of lithic raw materials during the Brewerton and Steubenville occupations (Mohney 2002; Mohney and Lothrop 2003). A total of 119 Steubenville and 82 Brewerton points were recovered at the East Steubenville site. For the Brewerton assemblage, Onondaga chert (55%) and other local cobble cherts (20%) comprise 75 percent of the total projectile point assemblage, while Upper Mercer chert from east-central Ohio was also popular. In order of importance, Brewerton points were produced from PA Prehistoric Data Synthesis, Raccoon Creek Watershed 65
Onondaga, Gull River, Upper Mercer, and untyped black and gray cherts. In comparison, a wider variety of cherts were used to produce the Steubenville points at the site, including a predominance of bedrock cherts, such as Ten Mile chert. In order of importance, Steubenville points were produced from Ten Mile, untyped black, Upper Mercer, Brush Creek, and Uniontown cherts. Both Brewerton and Steubenville points were frequently produced from Flint Ridge and Upper Mercer cherts from east-central Ohio. The predominance of bedrock cherts for Steubenville points contrasts with the heavier use of local cobble cherts for Brewerton points. Mohney and Lothrop (2003) argue that the reliance on primary source cherts for Steubenville points was likely due to its higher quality and larger raw size. Steubenville points are larger than Brewerton points and it would have been difficult to produce Steubenville points using small chert cobbles. Thus, the morphology of the collected raw material was important in determining which raw materials were used during the Late Archaic period. The sources of lithic raw materials also suggest more widespread travel patterns for individuals using Steubenville points, compared to those that used Brewerton points (Mohney and Lothrop 2003). E. LATE ARCHAIC: CONCLUSION AND RESEARCH QUESTIONS Late Archaic Overview Data provided in PASS files and research reports provide a means to compare sites with Brewerton and Steubenville points within Watershed D of Subbasin 20. Generally, PASS data indicate that Panhandle Archaic/ Steubenville settlement was less dense (fewer sites), but encompassed a wider range of landforms and drainages than Brewerton sites. Sites with Steubenville points are within both uplands and lowlands and within the Raccoon Creek and Cross Creek drainages, while Brewerton sites tend to be more in lowlands of Raccoon Creek. Most importantly, Steubenville point sites are generally richer in artifacts, yielding abundant faunal and botanical remains, and wide varieties of tools and features (including human and dog burials). This appears to be true both in the Upper Ohio Valley at sites like East Steubenville and within the Raccoon Creek Valley, at sites like 36Bv17 and 36Bv52, discussed above. While these latter two sites did not yield shell middens or burials, they were comparatively more dense and diverse in terms of artifacts than Brewerton point sites (Mayer- Oakes 1955:141,144). Sites with Steubenville points were either occupied: 1) on multiple occasions during a restricted timeframe, as suggested by Lothrop (pers. comm., 2003); or 2) were occupied once by numerous individuals. If the former is true, then the sites suggest multiple reoccupations of the same landform for shell processing, an intensive subsistence strategy not commonly seen at Brewerton point sites. If the latter is true, then these Steubenville point sites could be interpreted as incipient villages. In contrast to these sites with Steubenville points, Brewerton sites in Subbasin 20 and the Upper Ohio Valley are more numerous, but generally consist of low-density lithic reduction stations and/or small campsites with limited artifact diversity. These differences in settlement patterns, site types, and chert use may reflect fundamental differences in land-use and lithic technologies of the respective Late Archaic sub-periods. Brewerton point sites are reminiscent of Early and Middle Archaic sites in Watershed D of Subbasin 20. However, by the end of the Late Archaic, during what some have called the Panhandle Archaic (Mayer-Oakes 1955; Mohney 2002), Steubenville point sites mostly resemble Transitional Archaic sites in other parts of Pennsylvania (compare with the Transitional Period of the Upper Juniata Subbasin; MacDonald 2003a). Sites with Steubenville points, thus, appear to mark the beginning of the transition from the mobile forager lifestyle of the Archaic to the more sedentary, horticultural lifestyle of the subsequent Woodland. While agricultural remains, steatite, pottery, and broadspears are not recovered at Steubenville 66 PA Prehistoric Data Synthesis, Raccoon Creek Watershed
/Panhandle Archaic sites, many other site traits are similar to Transitional Archaic sites, including similar timeframes, subsistence patterns, site sizes, and intensity of site occupations. As such, researchers in the Upper Ohio Valley may want to consider the Panhandle Archaic/Steubenville sites as representative of a distinct regional sub-culture within the Transitional/Terminal Archaic Period in the Upper Ohio Valley. Late Archaic Research Questions This summary of Late Archaic archaeological data in and near Watershed D of Subbasin 20 has generated several research issues which should be considered when conducting archaeological work in the area. Twelve Late Archaic research questions are listed below; this list is by no means comprehensive and should be used only as a starting point for generating additional research issues. Archaeological sites which can provide information pertaining to these and other research questions will likely meet National Register Criterion D; thus, unless they lack integrity, sites which address these research questions will be eligible for listing in the National Register for Historic Places: 7. If so, is it best to include Panhandle Archaic/Steubenville sites within the Transitional/Terminal Archaic period? 8. How can we explain the lack of Steubenville points at the Leetsdale site, given its proximity and similar age to Steubenville sites in the Upper Ohio and Raccoon Creek Valleys? 9. Why don t these Transitional-era sites in the Upper Ohio Valley yield early evidence of agriculture, as do sites in central and eastern Pennsylvania? 10. Why wasn t any steatite recovered at Steubenville, which dates to approximately the same time period as the Transitional Period occupations of Leetsdale? 11. Does Leetsdale represent the western-most extent of steatite use in Pennsylvania? 12. Are the birdpoints at Leetsdale the easternmost expression of that point type? 1. Does the Brewerton portion of the Late Archaic represent a continuation of the Middle and Early Archaic? 2. Does the increase in recorded Brewerton sites imply an increase in populations compared to the Early and Middle Archaic? 3. Why are there more Brewerton sites than Steubenville Sites? 4. What accounts for the differences in lithic raw material use between Brewerton and Steubenville populations? 5. Does the transition between Brewerton and Steubenville represent a population replacement or an in-situ evolution of Late Archaic cultures? 6. Does the increasing site size and density of Steubenville sites suggest that Native Americans were transitioning into a Woodland-type culture? PA Prehistoric Data Synthesis, Raccoon Creek Watershed 67
(This page intentionally blank) 68 PA Prehistoric Data Synthesis, Raccoon Creek Watershed
CHAPTER VIII EARLY WOODLAND PERIOD 3000 to 2100 BP A. EARLY WOODLAND OVERVIEW The Early Woodland period witnessed the continuation of trends that emerged during the end of the Late Archaic period at Steubenville/Panhandle Archaic sites. The emergence of pottery, mound building, and agriculture are hallmarks of the Early Woodland period. Although the subsistence base continued to be based on hunted and gathered resources, horticulture gradually assumed greater importance during the Early Woodland (Clay 1991). The emergence of the Adena cultural complex in the central Ohio Valley directly involved populations in Watershed D. Site densities are low for the Early Woodland in the study area, with only 22 Early Woodland sites compared to 68 during the preceding Late Archaic. Whether this reduction in site counts is due to population decrease/dispersement or nucleation of populations into fewer sites is an important issue. Beginning in the latter portion of the Early Woodland, Native Americans of the Adena culture built burial mounds and other ceremonial facilities along the Ohio River, as well as along the mid-atlantic coast (Adena) and in New York (Meadowood). Important mounds near the study area include Grave Creek Mound (Hemmings 1977), Cresap Mound (Dragoo 1963), and McKees Rocks Mound (McMichael 1956), all of which are on the Upper Ohio River. The structure and size of Adena ceremonial complexes was highly variable, ranging from small incipient Adena ceremonial sites to large mound complexes. In uplands above major drainages, Adena peoples occupied small camps and engaged in short-term extractive activities, such as lithic and food processing (Schweikart 1998:17). This site distribution pattern of mounds in larger river valleys and camps in uplands resembles that used by Adena cultures in the Middle Atlantic (Custer 1996:242) and Meadowood cultures in New York (Ritchie 1980:180-200; Snow 1980:267). Additional descriptions of chronology, material culture, and site types are described below for Adena sites in the Upper Ohio Valley and Watershed D. B. EARLY WOODLAND MATERIAL CULTURE AND CHRONOLOGY Early Woodland Material Culture As introduced above, the prominent settlement pattern for Early Woodland cultures in the Upper Ohio Valley consisted of mortuary/ritual sites on alluvial terraces of major streams and rivers and resource extraction camps in uplands and along smaller streams. As such, material culture between the two site types varies, with the ceremonial sites containing typical Adena burial goods, while the resource extraction camps contain a limited range of food production and processing artifacts, typical of Archaic forager camps. One change from the Archaic camps is the presence of pottery and occasional domesticates at Early Woodland campsites. Material culture for Early Woodland sites is discussed below. Figure 14 shows the locations of important Early Woodland sites mentioned in the text. McMichael (1971:89) and Mayer-Oakes (1955:153), based on data from the McKees Rocks Mound near Pittsburgh, and several other mounds along the Upper Ohio Valley, indicate that Upper Ohio Valley Adena burial mound sites typically contain one or more of the following diagnostic characteristics: Adena/Cresap Stemmed (Figure 13-1A) and Robbins (Figure 13-1C) points Adena Plain and Half-Moon grit-tempered cordmarked and plain pottery Groundstone (gorgets, tubular pipes, celts and tablets) and bone tools (awls, needles) PA Prehistoric Data Synthesis, Raccoon Creek Watershed 69
Rolled copper beads and bracelets and cut mica. Copper gorgets (Figure 13-3A; McMichael 1956:142) conical mounds with prepared soil base or floor upon high-order alluvial terrace underlying primary mound built of a midden sacred circle earthwork with interior ditches underlying mound central sub-floor pit clay cap for primary mound cremations log-tomb burials body extended in flesh red ochre deposits double-posted circular houses Some of the traits listed above, including mound type and location, burial practices, and ceremonial items are also common at Middle Woodland Hopewell mounds; thus, diagnostic artifacts and features with radiocarbon dates aid in distinguishing Early and Middle Woodland sites. Several mound locations in the Upper Ohio Valley were used repeatedly and, thus, may contain evidence of stratified pre-mound component Adena, and Hopewell occupations, as at McKees Rocks Mound near Pittsburgh (McMichael 1956:144-145). The pottery and point types mentioned above are also common at non-mound Early Woodland sites, including food processing camps and lithic-reduction stations. The diagnostic Early Woodland projectile points (see Figure 13) in western Pennsylvania include Cresap stemmed (2950 to 2400 BP), Adena stemmed/ovate (2750-2250 BP), Meadowood points (2500-2790 BP; Ritchie 1980:181), and Robbins stemmed points (2400 BP to 1750 BP) (Dragoo 1963; Justice 1987:191-196). Early Woodland points are also commonly produced from Flint Ridge chert, indicative of the widespread Adena ties to Ohio (George 1975). Figure 13. Early Woodland/Adena Artifacts (from Dragoo 1963:179). Pottery, comprised mainly of grit-tempered plain wares, such as Half-Moon Plain and Cordmarked and Adena Plain, was introduced during the Early Woodland. Grit-tempered Adena Plain was recovered at McKees Rocks Mound near Pittsburgh (McMichael 1956: 139) and is the most common pottery type at Adena sites in the lower and middle Ohio Valley (Lothrop 2001b). Mayer-Oakes (1955:214) states that Half-Moon Cordmarked ware is definitely associated with Adena mounds and is the earliest unit in our pottery sequence. Vessels have lugs and flat bases, similar to earlier steatite and groundstone vessels, another indication that it represents an early, transitional pottery type. Half-Moon Cordmarked was produced using coils malleated 70 PA Prehistoric Data Synthesis, Raccoon Creek Watershed
with a cordwrapped paddle (Mayer-Oakes 1955:184-185). Temper was crushed grit, predominantly granite, but occasionally chert or limestone. Interior surfaces of the vessels were smoothed, while the cordmarking on the exterior is vertical near the rim, but in multiple directions on the vessel body. Vessel thickness ranges from 10-20mm, with a mean of 14mm (Lothrop 2001b). Mayer-Oakes (1955:189) states that evidence from the Georgetown site [in Watershed D on the Ohio River] places this type at the bottom of the pottery sequence and it may well be the earliest pottery in the area. Confirmation of the early age of Half-moon cordmarked ware was found at Meadowcroft Rockshelter, which yielded two vessels dated to 2815 BP and 3065 BP, respectively (Johnson 1982:142), the earliest occurrence of this pottery type in the Upper Ohio Valley (see radiocarbon dates below). McKees Rocks Mound also yielded significant quantities (79 sherds) of chert-tempered Half-Moon plain pottery. All of the sherds are likely from one vessel, which was plain-surfaced, thick, crude, and flatbottomed (McMichael 1956:139). McMichael further comments that Half-Moonware at McKees Rocks generally resembles Fayette Thick, the most common type at Adena sites further downstream on the Ohio River to the south and west (Figure 14). Early Woodland Chronology Early Woodland Radiocarbon Dates in Watershed D Within the study area, Early Woodland sites with radiocarbon dates are infrequent, including only Meadowcroft Rockshelter and the Georgetown Site. At Meadowcroft, nine features yielded Early Woodland radiocarbon dates, suggesting frequent use of the site at that time (Stuckenrath et al. 1982:80): 1 2 5 ± 1 2 5 B C ( 1 9 5 0-2 2 0 0 B P ; SI- 2362); charcoal from firepits/upper 1/3 of Stratum V (F-14). 2 0 5 ± 65 BC ( 2 0 9 0-2 2 2 0 B P ; SI- 2487); charcoal from firepits/upper 1/3 of Stratum V (F-14; associated with 10-, 12-, and 14-row corn fragments). 3 4 0 ± 9 0 B C ( 2 2 0 0-2 3 8 0 B P ; SI- 2051); charcoal from firepits/upper 1/3 of Stratum IV (F-16; associated with 16-row corn fragments). 3 7 5 ± 7 5 B C ( 1 9 5 0-2 2 0 0 B P ; SI- 1674); charcoal from firepits/upper 1/3 of Stratum IV (F-16; associated with 16-row corn fragments). 5 3 5 ± 3 5 0 B C ( 2 1 3 5-2 8 3 5 B P ; SI- 2359); charcoal from firepits/upper 1/3 of Stratum IV (F-16). 7 0 5 ± 1 2 0 B C ( 2 5 3 5-2 7 7 5 B P ; SI- 3031); charcoal from firepits/upper 1/3 of Stratum IV (F-16). 8 6 5 ± 8 0 B C ( 2 7 3 5-2 8 9 5 B P ; SI- 1665); charcoal from firefloor/middle 1/3 of Stratum IV (F-16; Feature 60B; Half-Moon Cordmarked vessel; associated with squash remains). 8 7 0 ± 7 5 B C ( 2 7 4 5-2 8 9 5 B P ; SI- 1665); charcoal from firepit/middle 1/3 of Stratum IV (F-16; associated with squash remains). 9 1 0 ± 8 0 B C ( 2 7 8 0-2 9 4 0 B P ; SI- 1660); charcoal from firepits/firefloors/lower 1/3 of Stratum IV (F-16). 1 1 0 0 ± 8 5 B C ( 2 9 6 5-3 1 3 5 B P ; SI- 2049); charcoal from firepits/firefloors/lower 1/3 of Stratum IV (F-16). 1 1 1 5 ± 8 0 B C ( 2 9 8 5-3 1 4 5 B P ; SI- 1664); charcoal from firepits/upper 1/3 of Stratum III (F-18; Feature 60; associated with Half-Moon Cordmarked vessel and squash remains). PA Prehistoric Data Synthesis, Raccoon Creek Watershed 71
Ohioview Georgetown Figure 14. Location of Early Woodland Sites Discussed in Text. gaic0nsultants, INC. DRAWN DHM APPROVED jcl DATE 7/10/03 DWG. NO 20 02-441-10--C-A14 Early Woodland/Adena Sites in and Near Watershed D Map Key Sites outside of Watershed D Sites in Watershed D Watershed Boundary River/Stream in Watershed D River/Stream not in Watershed D County Line Ohio River Ohio River Subbasin 20 Grave Creek Mound Cresap Mound Pennsylvania Ohio River Buffalo Creek WEST VIRGINIA To Ohio River D Beaver County Harmon Cr. Meadowcroft PENNSYLVANIA Raccoon Creek Cross Creek Avella Mound E Wheeling Creek Allegheny County Washington County Mungai Farm Chartiers Creek Leetsdale Ohio River Scale F MayviewBend/ Mayview Depot McKees Rocks Mound Pittsburgh Monongahela River 0 5 10 miles Peters Creek Mound Crall Mound Figure 14 Map of Early Woodland Sites 72 PA Prehistoric Data Synthesis, Raccoon Creek Watershed
These dates suggest multiple occupations of the rockshelter between ca. 3,100 and 2,000 years ago, squarely within the Early Woodland. An Early Woodland Adena Ovate-based point was recovered in Stratum V, while the Early Woodland Feature 60 yielded the earliest Half- Moon cordmarked pottery in the Upper Ohio Valley. Corn and squash remains were also recovered in well-dated features (see above) in the Early Woodland levels of Meadowcroft. At the Georgetown Site on a first terrace overlooking the Ohio River (near the Ohio state line), Mayer-Oakes (Davis n.d.) obtained a date of ca. 2300 BP (173±80 BC) from a hearth feature with fragments of Half-Moonware ceramics. With the date of ca. 1100 BC from Meadowcroft, the two dates represent the earliest and latest occurrences of this pottery type in Watershed D. Early Woodland Radiocarbon Dates in Peripheral Areas While only Meadowcroft and Georgetown have yielded Early Woodland radiocarbon dates within the study area, several sites further south on the Upper Ohio River have yielded dated Early Woodland components. The Leetsdale Site (Fenicle 2003) has yielded two terminal Late Archaic/Early Woodland dates from charcoal in hearth features: 3370±40 BP (BC Beta-177514; Feature 1, Block 1) and 2940±40 BP (Beta-177992; Feature 3). With the Half- Moon ceramics, Leetsdale contains a substantial terminal Late Archaic/Early Woodland component. The Crawford-Grist #2 Site (36FA262), an Early Woodland hamlet on the Monongahela River, yielded two features with radiocarbon dates of 2370±150 BP (DIC- 3061A; Feature 46) and 2430±55 BP (DIC- 3105; Feature 53) (Grantz 1986:18). A radiocarbon date from a pooled charcoal sample obtained during coring at Grave Creek Mound, in Moundsville, West Virginia (only 25 miles southwest of the study area) yielded a date of 2150 BP ( 200±225 BC; Uga1324; Fowler et al. 1976:119; Hemmings 1984; McConaughy 1990:3). Additional dates of 2020±150 BP (M- 974), 2190±200 BP (M-975), 2240±150 BP (M- 976), and 2506±175 BP (Gulf) have also been recorded at Cresap Mound, a few miles south of Grave Creek Mound (McConaughy 1990:3). Cotiga Mound in the middle Ohio River Valley of Mingo County, West Virginia, yielded 20 Early Woodland radiocarbon dates between ca. 3000 BP and 2000 BP, with moundbuilding focused between 205 BC and 75 AD (Frankenburg and Henning 1994:179). C. EARLY WOODLAND, SUBSISTENCE, SITE TYPES, AND LOCATION TRENDS Early Woodland Subsistence Across Eastern North America, the Early Woodland subsistence base continued to be based on hunted and gathered resources, with horticulture gradually assuming greater importance during the Early Woodland, including the domestication of squash, chenopod, maygrass, sumpweed, and sunflower (Hart 1995b; Smith 1987). Ethnobotanical remains from various Early Woodland sites suggest that, while domesticates were introduced, they were dominated by the use of widely available wild plant foods (Adovasio and Johnson 1981; Ballweber 1989; Ritchie 1980). Meadowcroft Rockshelter yielded the oldest radiocarbon dates for squash (Cucurbita sp.) in the Upper Ohio Valley, including 1115±80, 870±75, and 865±80 BC (Cushman 1982: 207; Adovasio and Johnson 1981; Lothrop 2001b). Meadowcroft documents the first maize cultigens in the Ohio Valley as well, having yielded specimens of 16-row corn associated with dates of 375 and 340 BC, and 10-, 12-, and 14-row corn dated to 205 BC (Adovasio and Johnson 1981: 67, 72, Table 2). While acceptable as the earliest dates for maize in the Ohio Valley, the lack of similar evidence on any other sites in the region suggests that corn horticulture represented only a minor constituent of subsistence adaptations at this time. Meadowcroft has also produced evidence for limited use of seed-bearing plants and fruits PA Prehistoric Data Synthesis, Raccoon Creek Watershed 73
during the Early Woodland. Seeds of amaranth, goosefoot and knotweed are associated with a date of 980±75 BC (although apparently none show evidence of domestication); fleshy fruits, including remains of raspberry/blackberry, cherry and grape have been dated to 980 and 865 BC (Cushman 1982: table 4). Evidence of nut exploitation during the Archaic at Meadowcroft is also reflected for Early Woodland components by small amounts of walnut, hickory, and acorn associated with ninth and tenth-century BC radiocarbon dates. By 340 BC, greater quantities of preserved nutshell, especially walnut, may indicate increasing use of this resource (Cushman 1982: Table 3; Lothrop 2001b). Evidence of wild resource exploitation was also identified at the Leetsdale Site. Phase II excavations identified a Half-Moon Cordmarked pottery fragment that tested positive for rabbit during immunological analysis, while a stone bowl/mortar tested positive for mustard (Fennicle 2003). Finally, botanical remains from Early Woodland occupations at Crawford-Grist #2 in Fayette County, Pennsylvania, included freshwater mussel shellfish remains, hickory and acorn nut, black/raspberry, and starchy seed plants (Grantz 1986: 16-17). Adena Site Types in the Ohio River Valley Along with resource procurement camps of the Early Woodland (such as Meadowcroft Rockshelter), researchers currently recognize four basic Adena mortuary ritual site types in the Ohio River Valley: burial mounds, circular post structures, ceremonial circles, and large ditched villages (Clay 1998; Lothrop and Munford 2001). Only burial mounds and ceremonial circles are known to occur in the Upper Ohio River Valley. The site types and their respective roles in mortuary ritual activities are reviewed below in an excerpt from Lothrop and Munford (2001: 170-175). Burial Mounds represent the most obtrusive and most heavily investigated Adena mortuary ritual site type (Clay 1998). Adena mounds minimally represent a single construction episode capping one burial event; in many cases, these mounds were accretional and grew in size over time as additional burials and mound-capping episodes were added. Evidence for mortuary ritual associated with mound burials is highly variable in terms of post-mortem treatment of the body, choice of crypt or grave pit construction for interment of the body, types of grave goods included in the interment, and positioning of the burial within the overall mound structure, to name a few. Evidence of Circular Post Structures has been identified at many Adena mound sites in the lower and middle Ohio River Valley, most often below the mound itself (Clay 1998: 6-9; Clay and Niquette 1992). None of these site types has been identified in the Upper Ohio River Valley. This evidence consists of distinctive paired-postmold patterns arranged in symmetric circles up to 30m in diameter that represent the remains of former freestanding wooden post structures; the structures were originally interpreted as residential in function, perhaps representing the dwelling of a high status individual that was later buried within the mound (Webb 1941b). These pairedpostmold circles are now viewed as evidence of sometimes-roofless structures that predated burial mound construction and represent the locations of "mortuary camps" (Seeman 1977, 1986). Activities likely involved preparation of the deceased for burial, with the paired-post structure serving as a permeable screen arena for these mortuary rituals before burial elsewhere (Clay 1998:6-8). 74 PA Prehistoric Data Synthesis, Raccoon Creek Watershed
Ceremonial Circles represent a relatively rare Adena mortuary ritual facility (GAI 2002). In their simplest form, these small earthworks consist of a symmetric circular ditch with the excavated spoil placed on the outside of the ditch to form an exterior berm; in all cases, the ditch-and-berm construction is broken by one or more entryways (Clay 1998: 9-10; Otto 1979). In part, variability in ceremonial circles reflects decisions by corporate groups at one or more times to modify the site, thereby altering the ritual use of locations marked by these ceremonial circles (Clay 1998). At Mt. Horeb, a post circle was built within the ditch and berm construction (Webb 1941a). At other ceremonial circles, burial mounds were built within the limits of the interior ditch (e.g., the Biggs Circle- Hardesty 1964) or on top of the entire ceremonial Circle (Gay Mound; Clay 1998: 10). In other cases, ceremonial circles encompassed both mound and paired-post constructions (e.g., Dominion Land Company site) (Otto 1979). At Grave Creek, a ditch surrounding the 240-foot diameter mound was built either prior to, or during one of the stages of, mound construction (Hemmings 1978). Unlike other ceremonial circles, the circular ditch at Grave Creek lacked an adjoining exterior berm, a characteristic interpreted by Hemmings (1978: 24) as evidence that the Adena excavators of the ditch contributed the resulting spoil to mound construction (Hemmings 1978). Large Ditched Villages, the fourth Adena ritual site type, consist of very large, irregular subcircular enclosures the best example, Peter Village, involved a massive ditch and stockade construction (Clay 1988). Originally viewed as evidence of Adena "villages" within defensive constructions (Webb and Snow 1945), these sites are now interpreted as having been used by the Adena for ritual purposes and/or for processing and distribution of exotic ritual goods such as barite artifacts (Clay 1988). This site type is not known in the Upper Ohio River Valley. For the Upper Ohio Valley region, archeologists have recorded two of the four Adena mortuary site types defined above, including accretional burial mounds and ceremonial circle earthworks. Ditched village sites have not been identified with certainty, and paired-post circles are conspicuously absent (pers. comm., J.Lothrop, 2003; Lothrop 2001b). Key Adena/Early Woodland Sites near Watershed D Mayer-Oakes (1955:142-153) reports on several Adena mound sites in the Upper Ohio River Valley near Watershed D, which McMichael (1971:89) groups into three mound clusters, two of which are pertinent to the current study area. The Grave Creek mound group is centered on the Upper Ohio Valley in the northern panhandle of West Virginia, while the Forks of the Ohio group is centered near the confluence of the three rivers in Pittsburgh. Several Early Woodland habitation sites are also present near Watershed D, including Leetsdale, Mayview Bend, and Crawford Grist #2. Figure 14, above, shows the locations of these and other important Adena sites in and near the study area. The two mound clusters discussed here effectively surround Watershed D of Subbasin 20 and indicates a significant Adena presence in the study area. Grave Creek Mound Group Grave Creek Mound (Moundsville, West Virginia; Hemmings 1977; McMichael 1971) is ca. 25 miles southwest of Watershed D on the Ohio River and is the largest earthwork in PA Prehistoric Data Synthesis, Raccoon Creek Watershed 75
McMichael s Grave Creek Mound Group. Measuring 19 meters (62 feet) high and 73 meters (240 feet) in diameter, Grave Creek dwarfs all other recorded mounds in the Upper Ohio Valley and dates to ca. 2100 BP (Hemmings 1984:3; McConaughy 1990:3). Grave Creek mound's dimensions indicate that 1.2 million cubic feet of soil was required for construction, work that was probably carried out in several stages involving soil borrowing from locations on Grave Creek Flats and, in conjunction with associated mortuary rituals and grave construction, basket-by-basket placement of spoil at this location (Hemmings 1977, 1984). Fowler et al. s (1976:112) report on the 1975-1976 investigations at Grave Creek Mound indicated that the ditch underlying the mound was ca. 40 ft. wide, 4-5 ft. deep, and symmetrical with an opening to the south. Cores drilled within the mound itself indicated heterogeneous bands of fill likely denoting basket loads, a pattern observed at Cotiga Mound in Mingo County, West Virginia as well (Frankenberg and Henning 1994). Underneath the mound was a surrounding moat a possible example of a ceremonial circle, described above. Dragoo (1963:141) reports that more than 2,000 disk shell beads, hundreds of marginella shell beads, as well as numerous gorgets, copper bracelets, mica, and an engraved stone tablet were recovered in burials at Grave Creek Mound. Tombs at Grave Creek consisted of a heavy crib-like log structure, rather than simple pits with overlying logs as at Cresap Mound further downstream. The increased complexity of the Grave Creek burials may indicate a later occupation (Dragoo 1963; McConaughy 1990). Cresap Mound (Dragoo 1963; McConaughy 1990; McMichael 1971) is also within the Grave Creek Mound group of sites on the Upper Ohio River. Dating to ca. 2000-2500 BP (see dates above), Cresap Mound ( is located 6.5 miles south of Moundsville (ca. 30 miles south of the current study area). Excavated by the Carnegie Museum in 1958 under the direction of Don Dragoo, this stratified Adena mound was underlain by a prepared clay floor, a series of burials and other features, and an encircling backfilled trench. Dragoo's (1963: 33-35) sequence of feature and mound construction at the Cresap Mound documents changes over time in mortuary ritual at the site and provides a basis for evaluating the function and origin of Adena mounds. Figure 15. Profiles of Cresap Mound (from Dragoo 1963:21). 76 PA Prehistoric Data Synthesis, Raccoon Creek Watershed
As summarized by Lothrop and Munford (2001:176), a condensed synopsis of these events at Cresap Mound includes: 1. Removal of topsoil from an area approximating the subsequent mound footprint; 2. Creation of central fire pit; 3. Spreading of circular, prepared clay floor measuring 12.2 meters (40 feet) across; 4. Creation of a circular trench (Feature 14) around the clay floor perimeter; 5. Activity in the southwest quadrant of Feature 14 interior space: interment of Burials 11 and 28 and features 10 and 19 burial pits, subsequently covered by construction of small burial mound ("West Primary Mound") containing additional burials; 6. Activity in southern sector of Feature 14 interior space: additional burials and cremation basins (including burial 21 pit superimposed on feature 14), covered by construction of second small mound; 7. Activity in eastern sector of Feature 14 interior space: Feature 28 burial, with an associated "clay bench" superimposed on the inner edge of the Feature 14 trench, covered by construction of third small mound containing additional burials; 8. Feature 15, cremation pit placed in depression between southern and eastern mounds, followed by construction of single conical mound cap covering all three small mounds subsequent intermediate stages of mound capping further expanded the mound footprint, covering all of Feature 14. The placement of premound and initial small mound features within and on the margins of Cresap Mound Feature 14 provides the obvious interpretation suggested by Clay (1998:18) for Adena ceremonial earthwork circles: that the Feature 14 trench served to define premound ritual space at the site and structure the performance of initial mortuary rites. Stratigraphic information indicates that the Feature 14 trench was filled prior to construction of the Feature 28 burial and the overlying small eastern mound, as well as subsequent capping episodes covering the three small initial mounds. Clay and Niquette (1992) have suggested that the submound prepared-clay surface at the Kirk Mound was periodically swept clean following ritual activities that preceded mound construction. At Cresap, if the same kinds of site maintenance were conducted on the prepared clay floor (the "stage" for premound mortuary rituals [e.g., ritual food preparation, feasting, kindling of cremation fires]), the Feature 14 trench could have served as a fortuitous trap or intentional receptacle for debris from ritual activities that was later swept off the prepared surface. Forks of the Ohio Mound Group Among the Forks of the Ohio Group of mounds (McMichael 1971), the most pertinent to this study is McKees Rocks Mound near Pittsburgh (McMichael 1956). Peters Creek Mound and Crall Mound, further south on the Monongahela River, are the eastern-most Adena mounds in southwestern Pennsylvania. McKees Rocks Mound was originally excavated in the late 1890s by the Carnegie Museum, with subsequent mound interpretations by McMichael (1956) and Mayer-Oakes (1955; Photograph 12). Based on these studies (Figure 16), the mound appears to have been built in three phases, including a prepared clay floor with a central burial of an Adena female. This burial was subsequently covered with a layer of charcoal and ash interpreted as the remains of a burned burial structure. Atop this layer of ash were placed two additional episodes of fill, effectively forming the shape of the conical mound. The mound contained mostly extended burials, with lesser amounts of bundle burials and cremations. The primary mound contained only the single burial, while the two later mounds contained five and 29, respectively, suggesting that the final occupation was the most intensive. Artifacts recovered from the burials include numerous columnella and marginella shell beads, antler and bone tools, a reel-shaped gorget, and a piece of plain twined fabric enclosing a copper sheath (Dragoo 1963:151-159). PA Prehistoric Data Synthesis, Raccoon Creek Watershed 77
Photograph 12. View of McKees Rocks Mound Prior to the 1896 Excavations (from Dragoo 1963:153) Figure 16. Profile and Planview of McKees Rocks Mound (from McMichael 1956:148). Peters Creek Mound on the Monongahela is the eastern-most Adena mound in the Upper Ohio Valley (see Figure 14). In 1890, the mound measured 80 ft. in diameter and was ca. six ft. tall, diminished by multiple plowings. The mound contained a central burial area with human bone and other features, but the early excavations were not controlled, so records are scant (Dragoo 1963:159). Artifacts from Peters Creek include celts, a grooved axe, shell beads, columnella beads, two copper gorgets, three trips of folded copper, two copper sheaths (one with a bear tooth), and abundant red ocher (Dragoo 1963:159-160). Crall Mound in Monongahela City, Washington County, Pennsylvania, is also within the Forks of the Ohio Group, measuring some 60 ft. wide and nine ft. tall. Several smaller mounds indicated the likely presence of mound complex within the current city limits, most of which have been destroyed by modern development. As with the other Adena mounds in the region, the base of the mound was comprised of a clay floor with a central burial area with skeletal remains of at least five individuals buried with numerous ceremonial items (including a copper gorget, tubular pipe, and a copper strip) and covered with wood or bark. The primary mound was then enclosed by another mound layer including ash possibly related to a burial structure burn event. Early Woodland Non-Mound Sites outside the Study Area In addition to Meadowcroft (discussed above), a handful of Early Woodland habitation sites have been excavated in the Upper Ohio Valley, including Crawford-Grist #2, Leetsdale, and Mayview Depot/Mayview Bend (Lothrop 2001b). The Crawford-Grist site #2 (Grantz 1986) is located in Fayette County, Pennsylvania, occupying an upland bench and saddle, 125 meters above the adjacent Monongahela River. Although no Adena points were recovered, flat-bottomed Adena Plain pottery, a fireclay tubular pipe fragment, and accepted radiocarbon dates of 480 and 420 BC, collectively indicate contemporaneity with some 78 PA Prehistoric Data Synthesis, Raccoon Creek Watershed
of the earlier Adena mortuary ritual sites in the region (although Half-Moon Cordmarked pottery suggests Early Woodland occupation as well). Plowzone stripping revealed a moderate density of small cultural features, including hearth basins; U-shaped, conical and flatbottomed pits; and several postmolds, some of which may represent temporary structures or windbreaks (Grantz 1986: 3-8). As noted earlier, feature flotation produced a small sample of subsistence remains, including freshwater mussel shell, fish scales and calcined bone, and carbonized remains of hickory and acorn, blackberry/raspberry, and possible carbonized starchy seed plants (Grantz 1986: 16-17). These data, in combination with a simple artifact assemblage of pottery and stone tools (biface fragments, expedient unifacial implements, hammerstones and anvils, and a rounded-end celt) suggests a suite of campsite activities consisting of tool manufacture, woodworking, expedient tool use, and food processing and consumption. Grantz (1986: 18) interprets the site as a short-term encampment by "one or more extended families" during the late summer through late autumn or early winter. The presence of mostly local lithic raw materials in the stone tool assemblage, and minor amounts of Flint Ridge and Upper Mercer cherts (Grantz 1986: 8-9), suggests either seasonal travels to eastern Ohio lithic sources or interaction with groups living in those areas. Located along the middle reach of Chartiers Creek in Allegheny County, Pennsylvania, the Mayview Depot and Mayview Bend sites (36AL124 and 36AL125) are important for their glimpses into Early Woodland settlement in valley bottom settings of the Upper Ohio Valley (Kellogg et al. 1998; Lothrop 2001b). Mayview Depot straddles a dissected T-2 terrace, while the nearby Mayview Bend site occupies a T-1 terrace. Excavations at Mayview Depot encountered 75 hearth and refuse pit features within a total site area of 3300 sq. meters; however, most of the features probably reflect post 5/400 BC occupation (Kellogg et al. 1998). At Mayview Bend (Robertson et al. 1998), the large number features radiocarbon dated to 1040-570 BC indicated that site occupation was heaviest during this Initial Early Woodland timeframe. Investigators interpret Mayview Bend as a location for limited or specialized activities during both Initial Early Woodland and later Woodland occupations and believe the site was contemporaneous with the nearby Mayview Depot habitation site (Lothrop 2001b). The pre-500 BC Woodland occupations at Mayview Depot and Mayview Bend may be viewed as "clusters of interrelated activities." Further, the observations of low artifact density at the Mayview Depot and Mayview Bend sites, and the near absence of stone tools that the latter site, together suggest that the paucity of recorded Early Woodland (1100-5/400 BC) sites in the Upper Ohio Valley may reflect the low archaeological visibility of many of these sites and raises questions about the criteria used for site definition. Finally, another habitation site occupied during the Early Woodland is the Leetsdale Site (46AL480) on the north bank of the Ohio River, less than a mile from Watershed D. As noted above (see radiocarbon dates), one of the most substantial components of the site was deposited ca. 3000 BP Seventeen Half-Moon Cordmarked sherds were recovered (Fenicle 2003) along with 380 lithics and a stone bowl/mortar (Photograph 13). Photograph 13. Excavator holding Stone Bowl/Mortar from Leetsdale Site (www.lrp.usace.army.mil/.htm; Photo Courtesy of the U.S. Corps of Engineers. PA Prehistoric Data Synthesis, Raccoon Creek Watershed 79
Flaking debris from Phase I-II excavations was produced largely from local pebble cherts. Features associated with the Early Woodland occupation include several postmolds, a refuse pit, and several hearths (Schuldenrein et al. 2003). No domesticates were recovered from the site, but several nut hull fragments were recovered in the Early Woodland component, indicating nut processing and/or consumption at the site. Early Woodland Sites in Watershed D, Subbasin 20 PASS file data indicate a reduction in site counts compared to the preceding Late Archaic, with 22 sites compared to 69, respectively. When duration of period is taken into account, the Early Woodland yields a site density per decade of 0.18, compared to 0.38 for the Late Archaic. However, Early Woodland site density is similar to that of the Panhandle Archaic/ Steubenville phase (0.16 sites per decade). Rather than a population reduction, thus, perhaps the reduced site counts are due to nucleation of populations into villages, which began during the latter portion of the Late Archaic at these Steubenville sites (see previous chapter for discussion). Alternatively, a mobile forager lifestyle was dominant during the Early Woodland, and site visibility is low. Based on the locations of the mounds and the reduced site counts, one might infer the presence of mounds/villages within Watershed D of Subbasin 20. While such mounds might be present, no Early Woodland Adena mound sites have been excavated or previously identified in Watershed D of Subbasin 20. As discussed above, the lack of documented mound sites along the main stem of the Ohio River in the study area may be due to their destruction or lack of recognition to date. The valley of Raccoon Creek near its confluence with the Ohio River likely also contained Early Woodland mounds, although areas further upstream and along the smaller streams, including Cross Creek, were most likely used as hunting and gathering areas, with site types restricted to small camps (e.g., Meadowcroft Rockshelter), a pattern reflected in PASS files data (Table 28). Alternatively, the lack of mounds may actually reflect a prehistoric behavior pattern. Clay (1991) suggests that locations of mound clusters were also locations of intersecting social group territories. As such, mounds/mound clusters were locations for ceremonial congregation. If this is true, then perhaps Watershed D of Subbasin 20 lacks mounds because of an absence of social territory intersection. While no Early Woodland villages are reported in site files in Watershed D, the Ohioview Site (36Bv9) is an example of the type of Early Woodland site one would expect to find in the study area (see Figure 14). This site on the north bank of the Ohio River is less than one mile west of its confluence with Raccoon Creek (Alam 1961; Faingnaert and Doyle 1977). Its location on the north bank places it within Watershed B of Subbasin 20. While much of the site was destroyed by development (Alam 1961), thousands of artifacts were collected over the years by the landowner, many of which indicate a substantial Adena component. Artifacts include a bird stone, gorgets, celts, an Adena stone tablet, and a cache of bifaces. No mound is known to have existed at this location, but the upper 18 inches of the site were stripped for topsoil in the 1950s. As Alam (1961:61) states: about three-quarters of a miles upstream from [the Ohioview Site, near the mouth of Raccoon Creek] a number of rock mounds have been destroyed by the continual erosion of the river bank A number of Hopewell and Adena type artifacts have been recovered. 80 PA Prehistoric Data Synthesis, Raccoon Creek Watershed
Table 28. Early Woodland Sites, Watershed D, Subbasin 20 (PASS Files) SITE NO. SITE NAME ARTIFACT TYPE COUNTY MUNICIPALITY 36Bv0029 Georgetown Half-Moon pott. Beaver Green Twp. 36Bv0036 Wassler #2 Adena (Stemmed) Beaver -- 36Bv0078 Parish Farm Adena (Stemmed) Beaver Hopewell Township 36Bv0210 Lower Circle on Rock Adena (Stemmed) Beaver Raccoon Township 36Wh0106 Mungai Farm Adena (Stemmed) Washington Smith Township 36Wh0124 Paul Adena (Stemmed) Washington Hopewell Township 36Wh0181 Reservoir Adena (Stemmed) Washington Mount Pleasant Township 36Wh0344 -- Early Woodland Washington Cross Creek Township 36Wh0365 Yee Farm Early Woodland Washington Cross Creek Township 36Wh0368 MS #59 Early Woodland Washington Independence Township 36Wh0374 Rex Smith Farm Early Woodland Washington Cross Creek Township 36Wh0389 Carter Farm Early Woodland Washington Jefferson Township 36Wh0421 FB #2 Adena (Stemmed) Washington Mount Pleasant Township 36Wh0475 Berrinski Camp Adena (Stemmed) Washington Mount Pleasant Township 36Wh0757 Capuzzi #2 Early Woodland Washington Robinson Township 36Wh1115 Murgel Site Adena (Stemmed) Washington Mount Pleasant Township 36Wh1161 Gregorski Adena (Stemmed) Washington Jefferson Township 36Wh1191 Lowry #12 Adena (Stemmed) Washington Mount Pleasant Township 36Wh1286 Vanzin #4 Adena (Stemmed) Washington Mount Pleasant Township 36Wh1317 Herbst #2 Early Woodland Washington Mount Pleasant Township 36Wh1318 Cherry Run Terrace Early Woodland Washington Mount Pleasant Township 36Wh0297 Meadowcroft Half-Moon pottery Washington Jefferson Township Table 29. Early Woodland Site Location Data, Watershed D, Subbasin 20 (PASS Files). SITE SITE NAME SITE TYPE SETTING TOPO ELEV (FT.) NEAREST WATER 36Bv0029 Georgetown Open T0/T1 Alluvial 680 Ohio River 36Bv0036 Wassler #2 Open T0/T1 Alluvial 920 Raccoon Creek 36Bv0078 Parish Farm Village T0/T1 Alluvial 780 Raccoon Creek 36Bv0210 Lower Circle on Rock Open T0/T1 Alluvial 960 Raccoon Creek 36Wh0106 Mungai Farm Open Slopes Slopes 1220 Trib Burgetts Fork 36Wh0124 Paul Open Upland Upland 1280 Raccoon Creek 36Wh0181 Reservoir Open Upland Upland 1110 Raccoon Creek 36Wh0344 -- Open T0/T1 Alluvial 1000 Mid Fk Cross Creek 36Wh0365 Yee Farm Open Upland Upland 1220 Trib Burgetts Fork 36Wh0368 MS #59 Open Upland Upland 1200 Trib Cross Creek 36Wh0374 Rex Smith Farm Lithic Red. Upland Upland 1180 S Fk Cross Creek 36Wh0389 Carter Farm Open Upland Upland 1260 Raccoon Creek 36Wh0421 FB #2 Open T0/T1 Alluvial 1100 Raccoon Creek 36Wh0475 Berrinski Camp Open T0/T1 Alluvial 1080 Raccoon Creek 36Wh0757 Capuzzi #2 Open T0/T1 Alluvial 940 Raccoon Creek 36Wh1115 Murgel Site Open Saddle Saddle 1300 Chartiers Creek 36Wh1161 Gregorski Open Saddle Saddle 1160 Trib Scott Run 36Wh1191 Lowry #12 Open T0/T1 Alluvial 1100 -- 36Wh1286 Vanzin #4 Open T0/T1 Alluvial 1240 -- 36Wh1317 Herbst #2 Open Saddle Saddle 1250 Trib Raccoon Crk 36Wh1318 Cherry Run Alluvial Open T0/T1 Alluvial 1260 Cherry Run 36Wh0297 Meadowcroft Rockshelter T0/T1 Alluvial 1000 Cross Creek PA Prehistoric Data Synthesis, Raccoon Creek Watershed 81
SITE Table 30. Early Woodland Site Location Data, Watershed D (PASS Files). NEAREST WATER DIST. TO WATER (M) DIREC. OF WATER ORDER OF WATER DIST. TO CONF. DIRECTION OF CONF. 36Bv0029 Ohio River 50 Northwest 9 380 No Conf. 9 36Bv0036 Raccoon Creek 80 Southeast 1 1000 North 4 36Bv0078 Raccoon Creek 100 West 4 260 Southwest 4 36Bv0210 Raccoon Creek 100 West 2 140 Northwest 2 36Wh0106 Trib Burgetts Fork 50 North 1 710 East 2 36Wh0124 Raccoon Creek 300 South 1 540 Southeast 2 36Wh0181 Raccoon Creek 100 Southwest 3 175 South 3 36Wh0344 Mid Fk Cross Creek 0 On-site 1 10 South 3 36Wh0365 Trib Burgetts Fork 330 Southeast 1 860 Northeast 2 36Wh0368 Trib Cross Creek 150 Southeast 1 1550 East 2 36Wh0374 S Fk Cross Creek 5 North -- 490 West 3 36Wh0389 Raccoon Creek 0 East 1 750 Northwest 3 36Wh0421 Raccoon Creek 0 Southeast 2 420 Southwest 3 36Wh0475 Raccoon Creek 70 West 3 120 Southeast 3 36Wh0757 Raccoon Creek 120 South 5 255 Southwest 5 36Wh1115 Chartiers Creek 160 South 1 1140 Southwest 3 36Wh1161 Trib Scott Run 800 West 1 1980 North 4 36Wh1191 -- 10 North 1 180 Southwest 2 36Wh1286 -- -- Southeast -- -- No Conf. 0 36Wh1317 Trib Raccoon Crk 150 -- -- -- No Conf. -- 36Wh1318 Cherry Run 100 -- -- -- No Conf. -- 36Wh0297 Cross Creek 60 South 5 400 Southwest 5 ORDER OF CONF. Twenty-one of the 22 sites in Watershed D with Early Woodland artifacts are open sites classified on the site form as either camps or villages (Table 29). Sites are also distributed fairly evenly between uplands/ saddles/slopes (n=10) and alluvial (n=12) settings, although access to water remained a key to site placement (Table 30). Mean distance to water for Early Woodland sites 130 meters is similar to other time periods, as is mean distance to the nearest stream confluence (597m). the artifact inventories. This restricted range of artifacts at Early Woodland sites in Watershed D may indicate use of the area as a resource procurement zone, rather than a ceremonial space for the Early Woodland Adena. The separation of ritual and residential sites is a hallmark of the Adena settlement pattern along the middle and lower Ohio River Valley (Clay 1998). Mean elevation of Early Woodland sites is 1,101 ft. amsl, compared to 1128 ft. for Late Archaic sites, indicating a slight preference for locating sites in alluvial settings. As during the Archaic, artifacts from Early Woodland habitation sites are largely restricted to lithics, with only Meadowcroft Rockshelter and Georgetown (36Bv29; Mayer-Oakes 1955:153) yielding Half-Moon ceramics (Photograph 14).PASS forms for Early Woodland sites do not list ceremonial objects in Photograph 14. View Northwest toward the Georgetown Site (36Bv29) and the Ohio River. 82 PA Prehistoric Data Synthesis, Raccoon Creek Watershed
Given the locations of most mounds upon largeorder rivers, the lack of mounds in Watershed D is not altogether surprising, as nearly the entire watershed is comprised of low-order stream valleys and uplands. The only known Early Woodland village site in Watershed D is the Georgetown Site, which lacks a mound (Davis and Lantz 1987; Mayer-Oakes 1955:153). The high density of artifacts, the presence of postmolds, and the site s location on an upper terrace of a major river suggest a village function, although excavations were limited (Davis and Lantz 1987). As noted above, a radiocarbon date from the site places its occupation at 173±200 BC. Along with the Ohioview Site on the north side of the Ohio River (discussed above), these two sites indicate the possible presence of incipient Early Woodland villages within and immediately adjacent to Watershed D. D. EARLY WOODLAND SETTLEMENT PATTERNS AND LITHIC RAW MATERIAL USE Overall, compared with the Late Archaic occupations of the same sites, occupation intensity was reduced during the Early Woodland, as suggested by significantly fewer lithic artifacts at Meadowcroft, Mungai Farm, and the other Cross Creek sites (Figure 17). On one hand, this pattern is not expected, given the hypothesis that populations merged into fewer (but larger) village sites during the Early Woodland. Given this hypothesis, Early Woodland sites should be more artifact-rich, not less than Late Archaic sites. On the other hand, this reduced intensity of site use may reflect the nature of the recorded Early Woodland sites in the study area. As noted, most sites are small hunting and gathering camps. Thus, the lower density of artifacts at sites in Cross Creek likely indicates less intensive use of sites in more remote uplands and smaller stream drainages a pattern seen elsewhere during the Early Woodland (Clay 1991: 3). These data support Clay s (1991) assertion that a mobile, forager lifestyle was dominant in the Early Woodland. Lithic raw material use patterns support the contention that Native Americans continued to incorporate much of the landscape into their travel patterns during the Early Woodland. At Meadowcroft Rockshelter in the Cross Creek watershed, dominant Early Woodland toolstones are Kanawha chert, Upper Mercer chert, Flint Ridge chert, and Onondaga chert, in percentages that largely mirror the Late Archaic occupation (Figure 18); also compare Table 27 and Table 31). Meadowcroft likely functioned as a shortterm resource procurement and processing camp for populations living along the Upper Ohio Valley in nearby West Virginia. Local lithic raw material use pattern was also emphasized at the Leetsdale Site on the Ohio River near Watershed D. Phase I-II occupations yielded 95% local cobble chert artifacts, as well as a variety of other lithic materials (Fenicle 2003:61). Artifact Counts 300 250 200 150 100 50 0 245 94 c 207 Figure 17. Comparison of Late Archaic and Early Woodland Artifact Counts, Sites in Cross Creek Drainage (Vento and Donohue 125-127). 31 Late Archaic Early Woodland 31 Meadowcroft Mungai Farm Cross Creek Sites 5 PA Prehistoric Data Synthesis, Raccoon Creek Watershed 83
Table 31. Cross-Tabulation of Cross Creek Site by Lithic Raw Material for Early Woodland Artifacts. LOCATION ONONDAGA KANAWHA FLINT RIDGE UPPER M. BRUSH C. 10 MILE UNIONTOWN GLACIAL PEBBLE Meadowcroft (%) 8.5 26.6 16.0 21.3 8.5 6.4 1.1 2.1 8.7 94 Mungai Farm (%) -- 3.2 45.2 35.5 9.7 -- -- -- 6.4 31 Cr. Crk. Sites (%) -- 20.0 40.0 40.0 -- -- -- -- -- 5 Avella Mnd (%) -- -- 25.0 50.0 25.0 -- -- -- -- 4 Total pts (n) 8 27 32 35 12 6 1 2 11 134 Total % 6.0 20.2 23.9 26.1 9.0 4.5 0.8 1.5 8.2 100 OTHER TOTAL PTS In contrast to Meadowcroft and Leetsdale, Early Woodland component sites in Cross Creek reflect the increased role of Adena ceremonialism. At Mungai Farm (see Figure 18), exotic cherts dominate the assemblages, with Flint Ridge (45.2%) and Upper Mercer (35.5%) accounting for 80 percent of the artifact assemblage. At Avella Mound (36Wh415) in the Cross Creek drainage, the Early Woodland occupation was brief, represented by only four projectile points, three of which were produced from exotic chert (Flint Ridge and Upper Mercer). At other Cross Creek sites, four of five Early Woodland points were produced from Upper Mercer and Flint Ridge cherts (Vento and Donohue 1982:126-127). Flint Ridge chert, likely due to its high quality and aesthetic appeal, was popular in ceremonial blade and point production at Early and Middle Woodland ceremonial sites (Dragoo 1963; Mayer-Oakes 1955:154). Mound sites typically reveal high percentages of Flint Ridge chert; thus, the prevalence of the chert at a site like Mungai Farm, a short-term occupation within uplands, is conspicuous. This is especially true when compared to the Early Woodland occupation at nearby Meadowcroft Rockshelter, which displays an Archaic pattern of lithic raw material use (Figure 18). While the Early Woodland occupation at Avella Mound was minimal (Applegarth and Cowin 1982), the presence of exotic lithics may indicate incipient ceremonialism that fluoresced during the subsequent, more intensive Middle Woodland occupation. At Mungai Farm, the presence of exotic cherts may indicate that Early Woodland Native Americans camped at the site while traveling to or from a ceremonial site. While Wallace (1971) does not identify any Native American paths near Mungai Farm, the site is on a ridge divide between Cross Creek and Burgetts Fork, a viable route between the Upper Ohio Valley in the West Virginia panhandle and the forks of the Ohio near Pittsburgh. E. EARLY WOODLAND: CONCLUSION AND RESEARCH QUESTIONS Early Woodland Overview The Early Woodland represents one of the most dynamic periods within the entire prehistory of the Upper Ohio Valley. During the Early Woodland, Native American foragers continued to diversify their resource base, including the first use of domesticated plants. A hallmark of many of communities was the presence of burial mounds at sites like McKees Rocks, Cresap, Grave Creek, Peters Creek, and Crall along the Upper Ohio Valley and vicinity. While no mounds are known within Watershed D, the area is bordered to the east and west by two major mound groups the Grave Creek Mound Group near Moundsville and the Forks of the Ohio Group near Pittsburgh. The lack of mounds in the current study area is, thus, fairly conspicuous, and is likely due to one of two reasons: 1) a lack of social group intersections at which mounds were often built (Clay 1991) or 2) their destruction or lack of identification. While mounds might be expected along the main stem of the Ohio in Watershed D, the remainder of the study area is comprised of loworder tributary streams and uplands, areas typically used as hunting and gathering areas during the Early Woodland period. As such, most Early Woodland sites in Watershed D 84 PA Prehistoric Data Synthesis, Raccoon Creek Watershed
appear to be camps used for resource procurement. Early Woodland lithic raw material use patterns varied from the preceeding Late Archaic in the use of Flint Ridge and Upper Mercer cherts for ceremonial tool production. While some campsites like Meadowcroft continue patterns of lithic raw material use of the preceding Late Archaic, other sites like Mungai Farm and Avella Mound contained high percentages of Flint Ridge and Upper Mercer chert artifacts, reflecting the Adena ceremonial influence. These site file and research report data support Clay s (1991) assertion that the Early Woodland was marked by hunting and gathering by dispersed populations, with mounds located at the intersections of social territories. The presence of exotic cherts at some sites in the study area reflects the participation of Early Woodland cultures at Watershed D within the Adena Inaction Sphere. Early Woodland Research Questions This summary of Early Woodland archaeological data has generated several research issues which should be considered when conducting archaeological work in the area. Thirteen Early Woodland research questions are listed below; this list is by no means comprehensive, and should be used only as a starting point for generating additional research issues. Archaeological sites which can provide information pertaining to these and other research questions will likely meet the National Register Criterion D; thus, unless they lack integrity, sites which address these research questions will be eligible for listing in the National Register for Historic Places: 1. What relationship, if any, was there between terminal Late Archaic Steubenville populations and their Adena descendents/replacements in the Upper Ohio Valley? heavily populated sites, or settlement trends resulting in reduced site visibility? 3. Does the reduction in occupation intensity of Early Woodland campsites reflect shorter stays and/or less frequent visits? If so, what accounts for this changing pattern of site use? 4. What was the influence of the Adena culture in Watershed D? 5. Did Adena individuals build mounds in Watershed D? If so, where are they? 6. Is it possible that Watershed D, given tbe widespread prevalence of small streams and uplands, simply did not experience widespread Adena ceremonialism; thus, the lack of recorded mounds? 7. How extensive was horticulture during the Early Woodland in Watershed D? 8. Was there an increase or decrease in the use of semi-sedentary base camps/villages during the Early Woodland? 9. Was hunting and gathering the predominant subsistence method? 10. What is the cause for the reduced site counts compared to the Late Archaic? Were populations nucleating within fewer sites or did populations actually decrease for some reason? 11. How can we explain the origins of the Adena culture along the Upper Ohio River Valley? 12. How extensive was trade/travel between Adena in the Upper Ohio Valley and other parts of the east (Middle Atlantic) and Midwest (lower and middle Ohio Valley)? 13. What accounts for the increased use of Flint Ridge and Upper Mercer cherts at the Early Woodland occupations in Cross Creek? 2. Does the reduction in site numbers reflect decreasing Early Woodland populations, population nucleation into fewer, more PA Prehistoric Data Synthesis, Raccoon Creek Watershed 85
% Artifacts 35 30 25 20 15 10 5 0 8.5 6.1 29 26.6 Meadowcroft 21.3 18.4 16 12.6 15.1 8.5 7.7 6.4 Late Archaic Early Woodland 7.6 1.2 2.4 1.1 2.1 8.7 Onondaga Kanawha Flint Ridge Upper Mercer Brush Creek 10 Mile Uniontown Glacial Pebble Other % Artifactsl 50 45 40 35 30 25 20 15 10 5 0 45.2 Mungai Farm 35.5 Late Archaic Early Woodland 15.9 15.9 15.9 15.9 12.6 11.1 7.2 9.7 6.4 3.2 2.4 0 0 0 0 2.5 Onondaga Kanawha Flint Ridge Upper Mercer Brush Creek 10 Mile Uniontown Glacial Pebble Other Figure 18. Comparison of Late Archaic and Early Woodland Lithic Raw Material Use at Meadowcroft Rockshelter (top) and Mungai Farm (bottom) (after Vento and Donahue (1982:124-125). 86 PA Prehistoric Data Synthesis, Raccoon Creek Watershed
CHAPTER IX MIDDLE WOODLAND PERIOD 2100 to 1100 BP (150 BC to A.D. 850) A. MIDDLE WOODLAND OVERVIEW In western Pennsylvania, the Middle Woodland period is characterized by a continuation of trends initiated during the Early Woodland, including burial ceremonialism, interregional exchange, and the increased importance of indigenous cultigens and maize. In this report, we follow the chronological model of Raber (1985); however, many researchers in the Upper Ohio Valley follow a model established by Mayer-Oakes (1955) which defines the Middle Woodland as occurring between In the lower and middle Ohio River Valley, the Hopewell Culture was dominant and it influenced Native Americans in the Upper Ohio Valley; however, while the current study area was within the Hopewell Interaction Sphere, it was on its eastern periphery and sites with Hopewell influence are rare in the study area. As during the Early Woodland, most habitation sites are small lithic scatters, with few recorded villages and ceremonial sites. Avella Mound on Cross Creek is the only known Middle Woodland mortuary ritual site recorded in the study area. This mound, and others in the surrounding region, confirm that mortuary ritual remains an organized focus of Native American life, but with noteworthy changes (J. Lothrop, pers. comm. 2003). Lantz s (1989) study of Raccoon Creek points indicates a concentration of Middle Woodland habitation sites along the Raccoon Creek Valley. The Middle Woodland appears to represent a period of continuity from the preceding Early Woodland. Agriculture likely increased in importance during the Middle Woodland, although only a few regional sites have yielded ethnobotanical evidence of domestication at this time. B. MIDDLE WOODLAND MATERIAL CULTURE AND CHRONOLOGY Middle Woodland Material Culture Middle Woodland technology represents a continuation of styles utilized during the Early Woodland period. Ceramics of the Middle Woodland increase in diversity, decorations become more common, and vessel thickness generally decreases with increasing technological sophistication (McConaughy 2000). Projectile point types include Hopewellian artifacts, such as Snyder s and Raccoon corner-notched points (Figure 19). Groundstone tools, including manos, metates, pitted cobbles, adzes, and tool sharpeners are recovered at sites as well (Mayer-Oakes 1955; Lantz 1989). Figure 19. Raccoon Notched Points (from Lantz 1989:9, 11). PA Prehistoric Data Synthesis, Raccoon Creek Watershed 87
The most prevalent Middle Woodland point type in the study area is the Raccoon Notched type (see Figure 19), as exemplified by the numerous points of this type from the Outdoor Theater Site (36Bv24) along Raccoon Creek near Aliquippa. Raccoon Notched points are small, and thin, with excurvate blades. Side-notched and cornernotched varieties are noted by Lantz (1989:10-12). Point locations are most common along tributary valleys, rather than the main stem of the Ohio. The Raccoon Creek Valley is identified as a Regional Center of these points (Lantz 1989:5). Other Middle Woodland projectile point types identified in southwestern Pennsylvania include Backstrum Side-notched (a regional variant of the Chesser point) (George 1992), Kiski Notched (George 1982), and Murphys Stemmed (George 1982: 208-209). As described by Lothrop (2001b) and Johnson (2003), limestone tempered Watson Cordmarked ware (Watsonware) is the major Middle Woodland pottery type in the Upper Ohio Valley (Dragoo 1956; Hemmings 1984; Maslowski 1973; Mayer-Oakes 1955; McConaughy 2000). Watson Cordmarked and Watson Plain pottery are typically tempered with crushed limestones. Vertical cordmarking is common, with smoothing on interior surfaces. Vessels are globular, with decorations produced by using a cord-wrapped paddle to make edge impressions on the rim (Dragoo 1956: 64; Hemmings 1984: 29). Body sherds for Watsonware average 6 to 7mm thick. Watsonware was largely a utilitarian pottery (Mayer-Oakes 1955: 216) and displays a broad distribution through the Upper Ohio Valley. Watson pottery has been found at the Outdoor Theater Site in association with Raccoon Notched points (Lantz 1989). Radiocarbon dates for Watsonware pottery span the entire Middle Woodland period at sites in the Upper Ohio Valley (Maslowski 1983:53). Several archaeologists have commented on variation in cordage twist direction on Watson Cordmarked sherds for Middle Woodland components in the Upper Ohio Valley (Lothrop 2001b). More northerly sites, including Watson and Meadowcroft (Johnson 1982; Maslowski 1973, 1980; McConaughy 2000), have yielded sherds with predominantly Z-twist cordage impressions, while.further south, Watsonware displays mostly S-twist cordmarking (Hemmings 1984). These differences may signify distinct populations occupying northern and southern portions of the Upper Ohio Valley drainage during the Middle Woodland (Johnson 1982; Lothrop 2001b; Maslowski 1973). Middle Woodland Chronology Radiocarbon dates are available from only two Middle Woodland sites in the Raccoon Creek Watershed, Meadowcroft and Avella Mound. Dates from these sites are listed below: Meadowcroft Rockshelter (36Wh297) A. D. 7 0 ± 6 5 (SI-3022); Stratum V, charcoal from firepits/upper 1/3 of unit. A. D. 1 6 0 ± 6 0 (SI-3027); Stratum V, charcoal from firepits/upper 1/3 of unit. A. D. 2 8 5 ± 6 5 (SI-3024); Stratum V, charcoal from firepits/upper 1/3 of unit. A. D. 6 6 0 ± 6 0 (SI-3026); Stratum VII, charcoal from firepits/middle 1/3 of unit. Associated with maize remains. Avella Mound (36Wh415) A. D. 8 5 0 ± 9 0 (SI-3498); on fill within Burial 7 A. D. 5 3 0 ± 6 5 (SI-3499); Burial 7; date from charcoal fused to skulls A. D. 7 9 0 ± 1 1 0 (SI-2943); Burial 4/Cremation; Flint Ridge Bennington corner-notched point with scraper/flakes Additional radiocarbon dates are available from sites within the nearby Upper Ohio Valley, many of which are associated with diagnostic points, pottery, or maize. Maslowski (1983:53-55) provides several radiocarbon dates associated with Watson pottery, including dates of A.D. 890±45 (Dic-1499) and A.D. 930±55 (Dic-1500) from the Watson type site on the Upper Ohio Valley just west of Watershed D. 88 PA Prehistoric Data Synthesis, Raccoon Creek Watershed
Additional dates for Watson pottery place its use well into A.D. 1200 (Maslowski 1983:53). McConaughy (2000: 7) suggests that these dates are too recent and were likely contaminated during processing; he provides several additional dates for Watsonware spanning the entire Middle Woodland (McConaughy 2000: 8). M i d d l e Woodland Subsistence At Meadowcroft Rockshelter, 10-and 12-row maize fragments were recovered in Stratum VII, indicating continued cultivation of this crop circa A.D. 660 (Adovasio and Johnson 1981:67-72). Various nut, fruit, and seed remains were also found in these levels, suggesting that a mixed subsistence pattern persisted in the study area during the Middle Woodland (Cushman 1982: Tables 3 and 4). In Watershed C of Subbasin 20, excavations of Early-Middle Woodland features (dating to A.D. 500-1000) at the Connoquenessing Site recovered carbonized seeds (goosefoot, persimmon and grape) and nuts (hickory, butternut and walnut) (Knepper and Petraglia 1996: 38). Mayview Bend on Chartiers Creek revealed a decline in nut use during the Middle Woodland (Lothrop 2001b). C. MIDDLE WOODLAND SITE TYPES, LOCATIONS, AND SETTLEMENT PASS Files Data: Middle Woodland Sites in Watershed D Twenty-three Middle Woodland sites are recorded in PASS files for Watershed D, Subbasin 20 (Table 32). When the length of the periods is taken into account, Watershed D contains 0.23 Middle Woodland sites per decade, compared to 0.18 sites during the Early Woodland. Most sites in the subbasin were identified based on the presence of Raccoon Notched points, with two sites Meadowcroft and Outdoor Theater also yielding Watson Farm pottery. Table 32. Middle Woodland Sites, Watershed D (PASS Files). SITE NO. SITE NAME ARTIFACT TYPE AGE COUNTY MUNICIPALITY 36Bv0021 Biscan Farm MW MW Beaver Independence Twp. 36Bv0024 Outdoor Theatre Raccoon N. MW Beaver Hopewell Township 36Bv0036 Wassler #2 Raccoon N. MW Beaver -- 36Bv0040 Crevallero MW MW Beaver -- 36Bv0114 Jarecek MW MW Beaver Independence Twp. 36Bv0179 Biscan Farm #2 MW MW Beaver Independence Twp. 36Bv0240 Dravo Raccoon N. MW Beaver Greene Township 36Bv0248 Crucible Steel Raccoon N. MW Beaver Greene Township 36Bv0253 South Side High School Snyders MW Beaver Greene Township 36Wh0106 Mungai Farm Raccoon N. MW Washington Smith Township 36Wh0133 J Alrutz Farm #1 MW MW Washington Mt. Pleasant Twp. 36Wh0349 MS #39 Raccoon N. MW Washington Cross Creek Township 36Wh0393 MS #87 MW MW Washington Mount Pleasant Twp. 36Wh0409 MS #104 MW MW Washington Smith Township 36Wh0410 MS #105 MW MW Washington Cross Creek Township 36Wh0415 Avella Mound C-14 MW Washington Cross Creek Township 36Wh0654 Huffman Hilltop MW MW Washington Mt. Pleasant Twp. 36Wh0984 Lutz #3 MW MW Washington Mt. Pleasant Twp. 36Wh1032 C V Cowden MW MW Washington Cross Creek Township 36Wh1147 Brezinski #4 MW MW Washington Hopewell Township 36Wh1158 Lutz #9 MW MW Washington Mt. Pleasant Twp. 36Wh1318 Cherry Run Terrace MW MW Washington Mt. Pleasant Twp. 36Wh0297 Meadowcroft Watson CM MW Washington Jefferson Township PA Prehistoric Data Synthesis, Raccoon Creek Watershed 89
The majority of Middle Woodland sites in the study area (n=14 of 23; Table 33) are within the Raccoon Creek watershed, with an additional five sites in the Cross Creek watershed and two sites on the Ohio River. Most sites are open habitations (n=20 of 23) near the larger rivers, including Raccoon Creek, Cross Creek or the Ohio River, with comparatively little use of loworder drainages (n=5), slopes (n=4), and uplands (n=6 sites). Of the 23 sites, 12 are in alluvial settings, while 11 are on uplands, saddles, or sideslopes. Access to water was a key to site placement (Table 34), with sites averaging 122m from nearest water, similar to the preceding Early Woodland (130m). As with other time periods, proximity to stream confluences was apparently not important to site placement, with sites averaging 720 meters from confluences. Mean elevation for Middle Woodland sites was 1,073 ft. amsl, slightly less than during the Early Woodland (1,101 ft.), continuing a trend toward lower elevations begun during the Late Archaic (1,128 ft.). Research Reports: Sites in Watershed D Only a handful of research reports discuss Middle Woodland occupations in the Raccoon Creek watershed. Lantz (1989) study of Raccoon Notched projectile point distributions is the most comprehensive overview of the Middle Woodland in the area. The Cross Creek study (Applegarth and Cowin 1982) also provides data regarding Middle Woodland occupations at Avella Mound and Meadowcroft (Adovasio et al. 1977). Mayer-Oakes (1955) also includes information regarding Middle Woodland ceramic types at sites in the study area (Johnson 2003). Table 33. Middle Woodland Site Location Data, Watershed D (PASS Files). SITE SITE NAME SITE TYPE SETTING TOPO ELEV (FT) NEAREST WATER 36Bv0021 Biscan Farm Open T0/T1 Alluvial 830 Raccoon Creek 36Bv0024 Outdoor Theatre Open T0/T1 Alluvial 760 Raccoon Creek 36Bv0036 Wassler #2 Open T0/T1 Alluvial 920 Raccoon Creek 36Bv0040 Crevallero Open Upland Upland 920 Trib Raccoon Crk 36Bv0114 Jarecek Open T0/T1 Alluvial 820 Raccoon Creek 36Bv0179 Biscan Farm #2 Open T0/T1 Alluvial 830 Raccoon Creek 36Bv0240 Dravo Open T0/T1 Alluvial 760 Ohio River 36Bv0248 Crucible Steel Open T0/T1 Alluvial 700 Ohio River 36Bv0253 South Side H.S. Open Upland Hilltop 1320 Little Traverse Creek 36Wh0106 Mungai Farm Open Slopes Slopes 1220 Trib Burgetts Fork 36Wh0133 J Alrutz Farm #1 Open Slopes Slopes 1200 Trib Cross Creek 36Wh0297 Meadowcroft Rockshelter T0/T1 Alluvial 1000 Cross Creek 36Wh0349 MS #39 Open Slopes Slopes 1120 Raccoon Creek 36Wh0393 MS #87 Open Upland Ridgetop 1320 S Fk Cross Creek 36Wh0409 MS #104 Open Slopes Slopes 1230 Raccoon Creek 36Wh0410 MS #105 Open T0/T1 Alluvial 1230 Raccoon Creek 36Wh0415 Avella Mound Burial Mound Upland Hilltop 960 Cross Creek 36Wh0654 Huffman Hilltop Rockshelter Upland Ridgetop 1420 Trib Cross Creek 36Wh0984 Lutz #3 Open T0/T1 Alluvial 1180 Raccoon Creek 36Wh1032 C V Cowden Open Upland Ridgetop 1380 Raccoon Creek 36Wh1147 Brezinski #4 Open T0/T1 Alluvial 1120 -- 36Wh1158 Lutz #9 Open T0/T1 Alluvial 1180 -- 36Wh1318 Cherry Run Open T0/T1 Alluvial 1260 Cherry Run 90 PA Prehistoric Data Synthesis, Raccoon Creek Watershed
SITE Table 34. Middle Woodland Site Location Data, Watershed D (PASS Files). NEAREST WATER DIST. TO WATER (M) DIRECTION OF WATER ORDER OF WATER DIST. TO CONF (M) DIR. OF CONF. 36Bv0021 Raccoon Creek 130 East 6 340 North 36Bv0024 Raccoon Creek 140 East 4 380 -- 36Bv0036 Raccoon Creek 80 SE 1 1000 North 36Bv0040 Trib Raccoon Crk 120 NE 2 180 -- 36Bv0114 Raccoon Creek 80 South 6 380 Southwest 36Bv0179 Raccoon Creek 100 North 3 200 Northeast 36Bv0240 Ohio River 140 NW 9 520 North 36Bv0248 Ohio River 0 On-Site 1 1940 South 36Bv0253 Little Traverse Creek 220 SE 1 1380 Northeast 36Wh0106 Trib Burgetts Fork 50 North 1 710 East 36Wh0133 Trib Cross Creek 260 SE 2 560 Southwest 36Wh0297 Cross Creek 60 South 5 400 Southwest 36Wh0349 Raccoon Creek 80 NW 1 720 South 36Wh0393 S Fk Cross Creek 180 West 1 760 South 36Wh0409 Raccoon Creek 50 East 1 870 East 36Wh0410 Raccoon Creek 10 West 1 1150 South 36Wh0415 Cross Creek 90 SE 5 200 Southwest 36Wh0654 Trib Cross Creek 440 NE 1 1240 Northeast 36Wh0984 Raccoon Creek 60 Northeast 1 320 Southwest 36Wh1032 Raccoon Creek 200 SW 1 1500 Northeast 36Wh1147 -- 160 South 1 460 Northwest 36Wh1158 -- 60 West 1 620 Southeast 36Wh1318 Cherry Run 100 -- -- -- -- Mayer-Oakes (1955) Mayer-Oakes (1955) and Alam (1961) discuss the presence of Watson cordmarked and Mahoning cordmarked ceramics at the Ohioview (36Bv9; Alam 1961), Georgetown (36Bv29), and Industry (36Bv128) sites. Raccoon Notched Study (Lantz 1989) In his 1989 study of the regional distribution of Raccoon Notched projectile points, Lantz reviews a Regional Center of the point type within the Raccoon Creek Valley. As shown in Table 32, several sites with Raccoon Notched points are present in the study area. Lantz (1989:45) identified 14 sites along Raccoon Creek and 33 in Beaver County. The Raccoon Notched point type was first recognized by Emil Alam at the Outdoor Theater Site (36Bv24) near Aliquippa on the first terrace of Raccoon Creek. The 20,000 sq. meter site was completely destroyed by industrial development west of Aliquippa along Raccoon Creek. Along with numerous Raccoon Notched points, the Outdoor Theater site yielded Flint Ridge cache blades (Mayer-Oakes 1955:154-155), drills, a scraper, triangular points, pendants, celts and Middle and Late Woodland pottery, including Watsonware (Lantz 1989:27). Avella Mound (36Wh415) As reviewed by Lothrop (2001b), Avella Mound on Cross Creek is the only investigated Middle Woodland burial mound in Watershed D of Subbasin 20 (Photograph 15; Applegarth and Cowin 1982:249-255). Located on a low knoll overlooking Cross Creek, this mound measured approximately 1 meter high and 18 meters wide. Mortuary activity involved construction of four stone burial structures, after which a prepared floor of white clay was laid down. Subsequent burials included a large stone cyst and several stone heaps. Burials included extended and inplace and redeposited cremations. Chronological markers include a Manker point from a cache, a Bennington point from a burial radiocarbon dated A.D. 790, and dates of A.D. 530 and 850 from another burial, indicating site occupation between 500 and 800 A.D. Grave goods are rare, being limited to a single cache containing a Manker point, flakes and cores, and a beaver PA Prehistoric Data Synthesis, Raccoon Creek Watershed 91
incisor. Lithic raw material use at the site is discussed in detail below, but, in general, indicates use of local cherts in addition to Flint Ridge and Upper Mercer cherts. represent accretional stone and earth structures in which most burial placement appears to have been haphazard (Lothrop 2001b). In contrast to Adena mounds, which were separate from habitation sites, Hopewell mounds were directly associated with them, suggesting a changing role for ceremonial mounds. At Watson Farm, mound matrix consisted of earth, quarried flagstone and river cobbles, with eight burials, including four extended (Dragoo 1956; Lothrop 2001b). Submound refuse pits likely represent premound habitation features associated with the village. Photograph 15. View of Hill in the Former Location of Avella Mound. Excavations at the site removed most of the mound, which was on top of the hill. View Northwest. Meadowcroft Rockshelter (36Wh297) In contrast to Avella Mound, Middle Woodland occupations at the nearby Meadowcroft Rockshelter indicate no ceremonial activity and abundant food processing tasks. Limestonetempered Watsonware pottery dominates the ceramic assemblage, with 354 sherds in 36 vessel clusters (Adovasio et al. 1977:61). As described below, a variety of lithic raw materials were used at the site, whereas at ceremonial sites (such as Avella Mound), most lithics were from non-local sources (Flint Ridge and Upper Mercer). Relevant Middle Woodland Sites Outside of Watershed The largest post-adena Middle Woodland mounds, such as Watson Farm and Fairchance, are found on the Ohio River in the northern panhandle of West Virginia (e.g., Dragoo 1959; Hemmings 1984). Watson Farm and Fairchance mounds exhibited dimensions similar to the larger mounds in the upper Allegheny Valley (Dragoo 1956; Hemmings 1984 a). Both Fairchance Mound revealed a prepared floor sand lens, laid down prior to construction of the 15x20-m mound (Hemmings 1984:10). As defined by Lothrop (2001b), the main burial (Burial 17) was a sub-floor rectangular slab-lined pit roofed with logs, stone slabs and clay (Hemmings 1984: 21) containing the remains of a young adult male and large numbers of grave goods, including over 600 marine shell disc beads, greenstone celts, gorgets, hematite and mica (Hemmings 1984: 21-24). Excavations of moundfill uncovered 49 burials, most of them extended, with several displaying stone slab grave pit construction. Based on these findings, Hemmings (1984: 20) concluded that the "Fairchance community was evidently organized on non-egalitarian lines" cited in Lothrop 2001b. The Billy #1 Mound (Westmoreland County) is located on the lower reaches of the Monongahela River (George 1992:17-21). Although badly disturbed, excavations at this mound in 1984 revealed remnants of a probable stone crypt constructed of flagstones and river cobbles (Lothrop 2001b). George (1992) states that mounds are found in the downstream portions of the Monongahela drainage in Pennsylvania (Lothrop 2001b). 92 PA Prehistoric Data Synthesis, Raccoon Creek Watershed
D. MIDDLE WOODLAND LITHIC RAW MATERIAL USE Data for Middle Woodland sites in the study area suggest continued use of a wide range of lithic raw materials in stone tool production, with an emphasis on local lithics at campsites and exotic lithics at sites with possible links to the Hopewell Interaction Sphere. Lithic raw material data are available from the Outdoor Theater Site, Avella Mound, Meadowcroft, and Mungai Farm, as well as from sites in the Cross Creek watershed (Table 35). At the Outdoor Theater Site (36Bv24; Lantz 1989:45) on Raccoon Creek near Aliquippa, several lithic raw materials were utilized in the production of Raccoon Notched points (see Table 35), including Onondaga chert (38%), Upper Mercer chert (38%), Flint Ridge chert 21%), and Plum Run chert from Ohio (2%). Mayer-Oakes (1955:154) reports cache blades of Flint Ridge chert at the site as well. These data compare favorably with sites in the Cross Creek drainage, including Meadowcroft, Mungai Farm, and Avella Mound (Vento and Donohue 1982:124-127). Cherts from eastcentral Ohio dominate the assemblages, with Flint Ridge comprising 23.2% of artifacts at the sites and Upper Mercer comprising 30.3 percent. The presence of high percentages of these cherts implies a link to the Hopewell Interaction Sphere, even at sites without ceremonial functions, such as Meadowcroft, Mungai Farm, and Outdoor Theater. Only Avella Mound was a mortuary site, and it yielded the least number of Flint Ridge and Upper Mercer chert artifacts of the group (see Table 35). In addition to these research report data, PASS files provide limited lithic raw material data for eight single-component Middle Woodland sites in Watershed D. Chert/flint was the predominant lithic raw material at the sites, although Site 36Bv114 in Raccoon Creek yielded jasper, with its most proximate sources in central Pennsylvania. Flint Ridge and Upper Mercer were not identified in PASS files for Middle Woodland sites. George (1982) provides lithic data from the Middle Woodland Blawnox Site (36AL19) on the lower Allegheny River as well; namely, evidence from this site indicates the decline of the Hopewell influence circa 500-800 AD. Most lithics were produced from local Onondaga chert, with only minor amounts of Flint Ridge, Upper Mercer, and eastern Pennsylvania Jasper. E. MIDDLE WOODLAND: OVERVIEW AND RESEARCH ISSUES The Middle Woodland in the Raccoon Creek watershed is best characterized as a peripheral zone to the Hopewell Interaction Sphere of the lower and middle Ohio River Valley. Sites like Meadowcroft Rockshelter and the Outdoor Theater Site indicate use of the study area largely for hunting and gathering, rather than ceremonial purposes. The prevalence of Raccoon Notched points at several Middle Woodland open sites confirms the importance of hunting and gathering in the Raccoon Creek Valley and vicinity. The lone exception to this pattern in the study area is Avella Mound in Cross Creek which contained evidence of ceremonialism and several human burials dated to ca. 500-800 A.D. Table 35. Cross-Tabulation of Middle Woodland Site by Lithic Raw Material (Percent), Watershed D. ONON- DAGA KANAWHA FLINT RIDGE UPPER M. BRUSH C. 10 MILE UNION- TOWN GLACIAL PEBBLE OTHER TOTAL POINTS SITE* (%) Meadowcroft 0 21.9 9.4 37.5 6.2 9.4 3.1 0 12.4 32 Mungai Farm 3.7 1.9 39.6 30.2 9.4 9.4 1.9 0.9 2.8 106 Cr. Cr Sites 13.3 6.6 33.3 26.6 6.6 13.3 0 0 0 15 Avella Mnd 19.4 3.2 12.9 19.4 12.9 3.2 0 19.4 9.7 31 Outdoor Theater 38.0 0 21.0 38.0 0 0 0 0 3 53 Mean Percent 14.9 6.7 23.2 30.3 7.0 9.1 1.0 4.1 5.6 -- *Data from Lantz (1989:45) for Outdoor Theater and Vento and Donohue (1982:124-127) for Cross Creek Sites. PA Prehistoric Data Synthesis, Raccoon Creek Watershed 93
Despite the low density of Hopewell ceremonial sites/mounds in the study area, cherts from eastcentral Ohio dominate some lithic assemblages, suggesting ample travel and trade within the Hopewell Interaction Sphere. Cultural stratigraphy and radiocarbon dates suggest that these connections were chronologically staggered, extending from approximately the first through 5th or 6th centuries A.D. in the Upper Ohio Valley, relative to the 200 BC-A.D. 400 duration for the Hopewell Interaction Sphere. (pers. comm., J.Lothrop, 2003; 2001b). The proximity of burial mounds and habitation sites at some locations represents the most striking settlement shift during the post-adena Middle Woodland in the Upper Ohio Valley proper (Clay 1998; Lothrop 2001b). The separation noted between Adena ceremonial sites and habitations is not present at Middle Woodland sites. Subsistence remains at post- Adena sites, exemplified by Meadowcroft, indicate reliance on a broad spectrum of upland and riverine hunted and gathered foods with little evidence of cultigens. The diversity of diagnostic and nondiagnostic artifact classes at sites like Outdoor Theater suggests the possible presence of Middle Woodland villages in the Raccoon Creek Valley. At sites like Meadowcroft, the more restricted subsistence remains may reflect preservation differences or shorter-term occupations (Lothrop 2001b). The Raccoon Creek and Cross Creek drainages, thus, were likely peripheral areas to the Hopewell heartland with continuity in site types and settlement patterns between the Early and Middle Woodland. Hunting and gathering remained the most important subsistence pattern, with agricultural foods forming only a minor portion of the diet. Despite this apparent separation from the Hopewell Interaction Sphere, the presence of high percentages of Flint Ridge and Upper Mercer cherts indicates some ties to east-central Ohio, while sites like Outdoor Theater and Avella Mound yielded ceremonial items, such as cache blades and burials, respectively, which also indicate the influence of Hopewellian ceremonialism in Watershed D. Middle Woodland Research Questions This summary of Middle Woodland archaeological data in and near Watershed D of Subbasin 20 (Raccoon Creek) has generated several research issues which should be considered when conducting archaeological work in the area. Nine Middle Woodland research questions are listed below; this list is by no means comprehensive and should be used only as a starting point for generating additional research issues. Archaeological sites which can provide information pertaining to these and other research questions will likely meet National Register Criterion D; thus, unless they lack integrity, sites which address these research questions will be eligible for listing in the National Register for Historic Places: 1. How does the Middle Woodland differ from the preceding Early Woodland and subsequent Late Woodland? 2. Do the similar site counts for the Early and Middle Woodland indicate demographic continuity in the first millennium A.D.? 3. What was the influence of the Hopewell Interaction Sphere in Watershed D of Subbasin 20? 4. What is the significance of the high percentages of Upper Mercer and Flint Ridge cherts at Middle Woodland sites? Does this imply links to the Hopewell Interaction Sphere? 5. What is the significance of the increased use of Raccoon Notched points in the study area? Is this a Hopewell-associated phenomenon? 6. Did populations become nucleated during the Middle Woodland? If so, where are the base camp/village sites? 7. What role did agriculture play during the Middle Woodland period? 8. How did settlement patterns change, as reflected in lithic raw material use at Middle Woodland sites? 9. Was there an increase in ceremonial/ritual activity in the study area during the Middle Woodland? 94 PA Prehistoric Data Synthesis, Raccoon Creek Watershed
CHAPTER X LATE WOODLAND PERIOD 1100 to 400 BP (AD 800-1550) A. INTRODUCTION During the Late Woodland, Native American populations increased, as did conflict associated with competition for resources. In the Upper Ohio Valley, many researchers label this period the Late Prehistoric, following Mayer-Oakes (1955; Johnson 2003; Sciulli et al. 2003). For all intent and purposes, the Late Prehistoric and Late Woodland discussed here refer to the same cultural events. Numerous researchers have provided regional syntheses or excavated Late Woodland/Late Prehistoric sites in western Pennsylvania, including the valleys of the Upper Ohio River (Davis 1988; Mayer-Oakes 1953, 1955; Rue 1990), the Upper Allegheny River (Burkett 1999; George 1978, 1998; Lantz 1982, 1989; Lounsberry 1997; Johnson et al. 1979), the Monongahela River (George 1974; Johnson and Athens 1988), among other lesser order streams and rivers (Herbstritt 1981a; Knepper and Petraglia 1996; MacDonald and Cremeens 2002; Skirboll and Hanson 1996). These studies provide an outstanding view of Late Woodland culture within western Pennsylvania in general, but lack specific data regarding the Raccoon Creek drainage in particular. The Monongahela culture was dominant during the Late Woodland, focusing their settlement on the Monongahela, Youghiogheny, and Upper Ohio River Valleys of southwestern Pennsylvania and northern West Virginia. In the Raccoon Creek drainage, the Monongahela influence is uncertain, although several sites have yielded Monongahela ceramics. PASS files include data regarding 30 Late Woodland sites in Watershed D. Raccoon Creek and vicinity is home to some important Late Woodland sites, including Georgetown and Shippingport, but most well-studied Late Woodland sites are located on the periphery of the current study area. Late Woodland village sites, such as McKees Rocks, Watson Farm, and hundreds of others, are all located outside of Watershed D. Figure 5 in Chapter III shows the locations of sites discussed in this chapter. Compared to other areas in southwestern Pennsylvania, Watershed D is characterized by a low density of villages and a comparatively high frequency of open habitations/camps. Thus, along with the Mahoning River to the north (MacDonald and Cremeens 2002), Raccoon Creek may represent a peripheral area during the Late Woodland with only marginal affiliations to the Monongahela. B. LATE WOODLAND MATERIAL CULTURE, CHRONOLOGY, AND KEY REGIONAL SITES Hallmarks of the Late Woodland period are shell-tempered pottery and triangular projectile points. Mayer-Oakes (1955:158) indicates that plain pipes, bird-bone beads, and cannel coal pendants are also diagnostic of Late Woodland occupations in the Upper Ohio Valley. The following section describes ceramic and projectile point types common to regional Late Woodland sites. W a t s o n P l a i n a n d W a t s o n C o r d m a r k e d P o t t e r y Watson Plain and Cordmarked wares continued to be used during the Late Woodland period. Watsonwares were coiled and subsequently malleated with a paddle and anvil. Temper is coarse limestone, with both interior and exteriors smoothed on the Plain variety and exteriors cordmarked on the other variety. Lips are square to rounded, with straight to slightly outflaring rims (Mayer-Oakes 1955:194-196). Maslowski (1983:53) reports the presence of Watson pottery at Meadowcroft Rockshelter associated with a radiocarbon date of A.D. 1025±65 (SI-2047), confirming its use into the Late Woodland. Mahoning Plain and Cordm a r k e d P o t t e r y Mayer-Oakes (1955:192-195) type description of Mahoning Ware remains the standard and includes three main types that date to the Middle-Late Woodland: Mahoning cordmarked, PA Prehistoric Data Synthesis, Raccoon Creek Watershed 95
Mahoning plain, and Mahoning incised (MacDonald 2000b). Mahoning cordmarked and plain types are regionally concentrated in the Beaver River drainage and vicinity and are considered typologically related to the shelltempered Monongahela cordmarked, while Mahoning incised is rarer and is most common in the Ohio Valley proper. Each variation was produced via the coil method, using crushed igneous rock (e.g., granite) or quartz temper. Mahoning cordmarked vessel decoration is largely confined to the lip area, which may also be notched or impressed. Coils were malleated with a cordwrapped paddle and cords were oriented vertically near the rim, but in varying directions on the remainder of the vessel. A minor number of sherds show evidence of fabric impressions. Decoration is confined to the rim, which was slightly flared and occasionally folded. Vessel necks were slightly constricted and vessel bodies were likely globular, although no known complete vessels have been identified. Sherd thickness is 3-10mm, with a mean of 7mm.Temper fragments comprise approximately 30-60 percent of the paste. Texture of sherds is regular and compact, with regular breaks along the coil line (Mayer-Oakes 1955:191-192). Mahoning plain sherds typically possess a smoothed interior and exterior, while the rarer Mahoning is typically incised with parallel straight lines on a smoothed exterior surface. Mahoning plain vessel rims are also slightly flared, with necks slightly constricted. Vessel lip is flattened and often L-shaped with interior and exterior overhang. Thickness of Mahoning plain sherds varies from 4-10mm, with a mean of 7mm (Mayer-Oakes 1955:192-193). Mahoning ware seems to occur on sites which are Middle Woodland or Hopewellian (Mayer- Oakes 1955:193), but Mahoning River Valley sites clearly indicate a Late Woodland period of use. The Coverts Crossing Site (36Lr75) yielded numerous Mahoning ware fragments with early Late Woodland dates (MacDonald 2000b:167). In addition, the Chambers Site (36Lr11) is one of six Late Woodland sites on the Mahoning and Beaver Rivers near New Castle that yielded Mahoning ware pottery and triangle points. Phase III excavations in the mid-1990s at the Chambers Site by Algonkian Archaeological Consultants, Inc. recovered a nearly complete Mahoning Cordmarked pot within a feature dated to between A.D. 779 and 1250 (MacDonald 2001). A small number of corn and gourd fragments were also recovered from a Late Woodland feature at Chambers. M o n o n g a h e l a C o r d m a r k e d, P l a i n, a n d I n c i s e d Monongahela shell-tempered wares are by far the most diagnostic ceramic type for Late Woodland sites in southwestern Pennsylvania. As defined by Mayer-Oakes (1955:196-198), Monongahela Cordmarked and Plain wares were typically tempered with crushed shell and produced using the coiling technique with malleation by paddle and anvil. Cordmarkings were produced using a cord-wrapped paddle. Exteriors are plain or cordmarked with some smoothing, while interiors are smoothed. Decorations are restricted to the lip and adjacent lower rim, with some lips showing cordmarking or punctation. Monongahela Incised wares were produced similarly, with the addition of incised parallel or rectilinear lines. Late Woodland Lithic Technology Generally, Late Woodland lithic technology in western Pennsylvania is characterized by the production of expedient tools for daily tasks using locally available cherts (Hart 1995a:522; Hatch 1980; MacDonald 2000, 2001; Stewart 1990). Late Woodland sites yield a variety of small triangle points, including Levanna (Figure 20) and Madison varieties. Typically produced from locally available cobble cherts, these small triangle points were used as arrow tips. Chesser notched, Jack's Reef Corner Notched (Justice 1987) and Backstrum Sidenotched (George 1992) also occur at Late Woodland sites in western Pennsylvania. Figure 20. Late Woodland Levanna Triangle Point (from Michels and Smith 1967:669). Actual Size. 96 PA Prehistoric Data Synthesis, Raccoon Creek Watershed
Late Woodland Radiocarbon Dates The radiocarbon date inventory for the Late Woodland in Watershed D is comprised of three dates from Meadowcroft Rockshelter: A. D. 1 0 2 5 ± 6 5 (SI-2047); Stratum VII, Floor 13; charcoal from firepits/mid 1/3 of unit; associated with Watson pottery (Maslowski 1983:53). A. D. 1 3 2 0 ± 1 0 0 (SI-3023); Stratum VIII, Floor 12; charcoal from firepit A. D. 1 2 6 5 ± 9 0 (SI-2363); Stratum IX, Floor 9; charcoal from firepit/upper 1/3 of unit Numerous radiocarbon dates are available from Upper Ohio Valley Late Woodland sites, including several near Watershed D. Nass and Hart (2000:132-133) provide a list of 72 radiocarbon dates for Late Woodland sites dating to between A.D. 1000-1600 in the Upper Ohio Valley. The quantity of dates increases over the course of the Late Woodland period, as shown in Figure 21, suggestive of increasing site counts and populations over time. No. of Radiocarbon Dates 35 30 25 20 15 10 5 0 18 Figure 21. Quantity of Late Woodland Radiocarbon Dates Over Time in the Upper Ohio Valley (data from Nass and Hart 2000:132-133). Near Watershed D, George (1995:28) reports on four sites in the Chartiers Creek Valley with radiocarbon dates: the Kelso Site, A.D. 1350; Morganza Village, A.D. 1300; Wylie, A.D. 1310 and 1350; and McKees Rocks Village, A.D. 1330. Church (1995) also reports on a well-dated Monongahela village in the northern West Virginia panhandle, the Saddle Site (46Mr95) with dates ranging from A.D. 1100-25 1000-1200 1200-1400 1400-1600 Years A.D. 29 1400, with increasing use of maize over the course of the various occupations (Church 1995:69). As noted above, the extensive research at Late Woodland village sites in southwestern Pennsylvania has resulted in hundreds of radiocarbon dates (Nass and Hart 2000), only a handful of which are from Watershed D of Subbasin 20. Pertinent Late Woodland Sites near Watershed D Numerous publications have provided details of Late Woodland cultures in southwest Pennsylvania, with several references provided at the beginning of this chapter. Given previous in-depth study and the low occurrence of Monongahela sites in the current study area, this report only touches on some of the key attributes of Monongahela settlements, as a basis for comparing them to sites in Raccoon Creek and vicinity. As of 1989, 332 Late Woodland Monongahela sites were recorded in western Pennsylvania, with a core of settlement centered near the Monongahela and Youghiogheny rivers (Johnson et. al. 1989). Several of the sites discussed in this section are shown on Figure 5 in Chapter I. Mass spectrometer analyses of human bone from Monongahela sites in West Virginia indicate that maize contributed up to 70 percent of the diet (Farrow 1986). Site types include villages in river bottomlands and along trail networks in uplands. Smaller camps are generally restricted to uplands and terraces of smaller streams and rivers. The Monongahela constructed large, multiseasonal, palisaded villages of one to five acres in size and probably contained from 100 to 150 inhabitants. These villages were generally circular and were often fortified with a round or oval stockade. Circular houses around a central circular plaza were abutted by large circular storage pits (Bennett and Porter 1986:17; Hart and Nass 1995; Nass and Hart 2000). A distinctive Monongahela storage feature was post-lined and pear-shaped and was often attached to a circular house (Mayer-Oakes 1955:161). PA Prehistoric Data Synthesis, Raccoon Creek Watershed 97
As described in more detail below, villages of this composition are not known in Watershed D, although three villages in Washington County are in uplands. The lack of excavations at these sites precludes further comparison. George (1995) describes four upland village sites a few miles south and east of Watershed D along Chartiers Creek in Watershed F. These four sites are the most proximate Late Woodland villages to the current study area and are shown on Figure 5 in Chapter III. Located on a hilltop overlooking the confluence of Chartiers Creek and the Ohio River, the McKees Rocks Site (36AL16) contains an Adena Mound (described above) and a late-late Woodland to Protohistoric village site (Buker 1968). Located in Watershed G of Subbasin 20, the site was excavated first in 1896 and later by the Allegheny Chapter of the Society for Pennsylvania Archaeology in the early 1960s. The village site contained a stockade with multiple burials and numerous storage features; however, Buker (1968:45) notes the marked absence of dwelling shapes or placement with the village confines. He suggests that a plaza was present within the stockade and that houses were outside the stockade s walls. Pottery and burial types (typically extended or semi-flexed) from the site indicate affinities with Monongahela are also common at Monongahela sites. The lack of distinct circular houses within the confines of the stockade led Buker (1968) to suggest possible Fort Ancient contacts, a suggestion not contradicted by George (1995:37). Subsistence at the site involved a mixed farming and foraging pattern, including maize agriculture, shellfish collection, and deer hunting. George (1995) comments on three additional sites in the Chartiers Creek Valley between Pittsburgh and Washington, Pennsylvania, within 20 miles of Watershed D. Radiocarbon dates for these sites all suggest fourteenth century A.D. occupations. The Kelso Site (36Wh23) and Morganza Village (36Wh48) are located on hilltops, while Wylie #3 (36Wh283) is located on a sideslope/bench near Kelso. These four sites (including McKees Rocks) indicate a preference for late-late Woodland village locations to be in upland settings that were easily defendable (George 1995). Finally, GAI (1986) identified a Late Woodland site (Site 36Wh924; see Figure 5) with likely Monongahela affiliations buried under a meter of alluvium in the floodplain of Chartiers Creek near Canonsburg. Only one flake was identified during initial shovel testing, but subsequent deep testing recovered 90 shell-tempered Monongahela pottery fragments and numerous flakes between 150 and 200cm below the ground surface. The location of this site has obvious implications for archaeological survey within similar settings in Watershed D. C. LATE WOODLAND SITES IN WATERSHED D Compared to the Early Woodland (n=22 sites; 0.18 sites/decade per period) and Middle Woodland (n=23 sites; 0.24 sites), the Late Woodland (n=30 sites; 0.43) experienced a significant population increase in Watershed D (Figure 22). This increase is similar to other regions of the state, such as the Upper Juniata River basin of south-central Pennsylvania (MacDonald 2003a). In Watershed D of Subbasin 20, site counts appear to have increased gradually from the low of 0.16 sites per decade of the terminal Late Archaic /Steubenville sites to the high of 0.43 per decade during the Late Woodland (see Figure 22). Site Count/Decades per Period 0.5 0.4 0.3 0.2 0.1 0 Steubenville 0.16 0.18 0.23 0.43 Figure 22. Site Densities per Period from the End of the Late Archaic to the Late Woodland, Watershed D. EW MW LW 98 PA Prehistoric Data Synthesis, Raccoon Creek Watershed
Within Raccoon Creek and vicinity, PASS files also indicate the continued prevalence of small open sites (n=22 of 30) during the Late Woodland, but with an increase in village (n=7) sites as well (Table 36). At Meadowcroft Rockshelter (Adovasio et al. 1977:61), the Late Woodland component yielded Jack s Reef and triangle projectile points, Mahoning and Monongahela ceramics, and wild food remains, but only a small number of maize fragments. Contrast these data with those from village sites in the Upper Ohio Valley which indicate a greater than 70 percent reliance on maize for some Monongahela populations (Farrow 1986). The high density of small, Late Woodland camps in Raccoon Creek resembles the pattern observed in the Mahoning River drainage to the north in Beaver and Lawrence Counties (Watersheds A and B, Subbasin 20). At the Coverts Crossing and Coverts Bridge Sites, on either side of the Mahoning River (ca. 50 miles north of the Ohio River), multiple Late Woodland camp occupations were recorded with evidence of extensive nut processing, but comparatively little maize agriculture (MacDonald 2000b, 2001; MacDonald and Cremeens 2002). As suggested by Church (1995:69-70), while maize agriculture certainly increased over time during the Late Woodland, hunting and gathering continued to be important, especially in areas peripheral to the Monongahela heartland. Table 36. Late Woodland Sites, Raccoon Creek Watershed (PASS Files). SITE NO. SITE NAME ARTIFACT TYPE MUNICIPALITY 36Bv0003 Old Indian Fort Triangles Shippingport Borough 36Bv0004 Lower Field/Shippingport Triangles/pottery Shippingport Borough 36Bv0029 Georgetown Watson CM Greene Township 36Bv0036 Wassler #2 Triangles -- 36Bv0040 Crevallero Triangles -- 36Bv0045 Poorhouse Run Late Woodland Potter Township 36Bv0076 Mack No. 2 Triangles Hanover Township 36Bv0078 Parish Farm Triangles Hopewell Township 36Bv0086 Route 30 Triangles -- 36Bv0169 Whriner No. 2 Triangles Hopewell Township 36Bv0180 Belich Farm Triangles -- 36Bv0202 Williams Farm Triangles -- 36Bv0213 Wildflower Reserve #3 Triangles -- 36Bv0248 Crucible Steel Triangles Greene Township 36Bv0250 Thomas Pate Triangles -- 36Wh0001 Moore Farm Late Woodland Smith Township 36Wh0022 Scarem & Foder Farm Triangles Hanover Township 36Wh0106 Mungai Farm Triangles/Jacks Reef Smith Township 36Wh0123 Shaffer Inc Monongahela CM Mount Pleasant Township 36Wh0297 Meadowcroft Triangles Jefferson Township 36Wh0338 MS #28 Triangles Independence Township 36Wh0404 Talkovich Farm Triangles Cross Creek Township 36Wh0409 MS #104 Triangles Smith Township 36Wh0410 MS #105 Triangles Cross Creek Township 36Wh0550 McNinch Site No. 2 Triangles Hanover Township 36Wh0758 Capuzzi #3 Late Woodland Robinson Township 36Wh1007 Carter #3 Triangles Mount Pleasant Township 36Wh1104 Cross Creek Log House Triangles Cross Creek Township 36Wh1118 Lowry #9 Triangles Mount Pleasant Township 36Wh1161 Gregorski Late Woodland Jefferson Township PA Prehistoric Data Synthesis, Raccoon Creek Watershed 99
As noted above, while PASS files indicate that most Late Woodland sites in Watershed D are open camps, seven sites in the study area are recorded as villages (Table 37). Johnson (2003) notes three main Late Woodland Monongahela villages along Raccoon Creek, including Murdocksville (36Bv88), Scarem (36Wh22), and Moore-Olah (36Bv38). These village sites are all in lowland terrace settings, a contrast to the upland Monongahela village pattern in the Mononghela drainage to the east. Johnson (2003) also notes the presence of three additional Monongahela village sites on the Ohio River in or adjacent to the current study area: Shippingport (36Bv4), Ohioview (36Bv9; Alam 1961); and Rochester (36Bv80). The four village sites in Beaver County are in floodplain/terrace settings of the Ohio River or Raccoon Creek, while the three village sites in more southerly Washington County are within uplands, saddles, or slopes (see Table 37). As described above, this upland location trend is typical of Late Woodland villages in the nearby Chartiers Creek and Monongahela River valleys, especially during the terminal Late Woodland/Late Prehistoric (George 1995). The upland setting of villages likely facilitated defense due to increasing intercultural conflict in the Upper Ohio River Valley and vicinity (George 1995). The Lower Field/Shippingport Site (36Bv4) on the Ohio River in Watershed D yielded multiple burials (Mayer-Oakes 1955), one of which contained a Ft. Ancient-type projectile point embedded in its skull (Davis 1988:34), supporting the hypothesis of increased conflict during the Late Woodland. This site, along with the three upland villages in Washington County and sites discussed above by Johnson (2003), indicates cultural links to the Monongahela for Late Woodland populations in Watershed D. At the Scarem Site (36Wh22), Johnson (2003) notes the presence of late Neutral Iroquois influence in the Raccoon Creek Valley, based on the presence of ceramic traits. The mean elevation of Late Woodland sites is 961 ft. amsl, a decrease of more than 100 ft. from the Middle Woodland (1,073 ft.) and a decrease of 140 ft. from the Early Woodland. This decrease over time indicates an everincreasing preference for sites to be placed at lower elevations, likely due to the increased role of agriculture and sedentism in lowland settings during the course of the Woodland. Mean distance to water for Late Woodland sites (98.3m) also decreased compared to the Middle Woodland (122m), as did mean distance to nearest stream confluence (522m versus 720m). Nineteen of the 30 Late Woodland sites are located in alluvial settings, a higher percentage than the Middle Woodland (12 of 23 sites) as well. These data support the elevation data in showing a Late Woodland preference for site placement in lowland, alluvial settings in Watershed D. D. LATE WOODLAND LITHICS AND SETTLEMENT As during other periods, the Cross Creek study provides some of the only data regarding Late Woodland lithic raw material use in Watershed D (Table 38). In contrast to the Early and Middle Woodland periods, Late Woodland sites generally contain a broader range of lithic raw materials, with an emphasis on locally collected cobble cherts. While Flint Ridge and Upper Mercer cherts were utilized extensively by regional Early and Middle Woodland groups, these exotic cherts were not as important during the Late Woodland, suggesting the severing of close ties to east-central Ohio. Lithic raw material compositions of artifacts from the Dravo Site (36Bv240) are similar to Meadowcroft and Mungai Farm, with locally collected cobble cherts dominating the assemblages (see Table 40). 100 PA Prehistoric Data Synthesis, Raccoon Creek Watershed
Table 37. Late Woodland Site Location Data, Watershed D, Subbasin 20. SITE NO. SITE TYPE SETTING LANDFORM ELEVATION NEAREST WATER 36Bv0003 Open T0/T1 Floodplain 660 Ohio River 36Bv0004 Village T0/T1 Terrace 680 Ohio River 36Bv0029 Village? T0/T1 Floodplain 680 Ohio River 36Bv0036 Open T0/T1 Terrace 920 Raccoon Creek 36Bv0040 Open Upland Hill Ridge/Toe 920 Trib Raccoon Crk 36Bv0045 Open T0/T1 Floodplain 680 Raccoon Creek 36Bv0076 Open T0/T1 Terrace 1020 N Fk Kings Crk 36Bv0078 Village T0/T1 Floodplain 780 Raccoon Creek 36Bv0086 Open T0/T1 Floodplain 840 Raccoon Creek 36Bv0169 Open T0/T1 Floodplain 780 Raccoon Creek 36Bv0180 Open T0/T1 Floodplain 820 Raccoon Creek 36Bv0202 Open T0/T1 Floodplain 820 Raccoon Creek 36Bv0213 Village T0/T1 Terrace 860 Raccoon Creek 36Bv0248 Open T0/T1 Terrace 700 Ohio River 36Bv0250 Open T0/T1 Floodplain 800 Raccoon Creek 36Wh0001 Open Slopes Middle Slopes 1020 Raccoon Creek 36Wh0022 Village Upland Hill Ridge/Toe 980 Raccoon Creek 36Wh0106 Open Slopes Middle Slopes 1220 Trib Burgetts Fork 36Wh0123 Village Slopes Upper Slopes 1170 Raccoon Creek 36Wh0297 Rockshelter T0/T1 Terrace 1000 Cross Creek 36Wh0338 Isolated Find Slopes Middle Slopes 1160 Raccoon Creek 36Wh0404 Village Upland Ridgetop 1200 Mid Fk Cross Creek 36Wh0409 Open Slopes Middle Slopes 1230 Raccoon Creek 36Wh0410 Open T0/T1 Stream Bench 1230 Raccoon Creek 36Wh0550 Open T0/T1 Terrace 970 Aunt Clara's Branch 36Wh0758 Open T0/T1 Stream Bench 980 Raccoon Creek 36Wh1007 Open Saddle Saddle 1270 Trib Raccoon Crk 36Wh1104 Open Slopes Middle Slopes 1160 Trib Cross Creek 36Wh1118 Lithic Red. T0/T1 Terrace 1120 Trib Cross Creek 36Wh1161 Open Saddle Saddle 1160 Trib Scott Run Table 38. Late Woodland Site Location Data, PASS Files (Watershed D). SITE NO. NEAREST WATER DIST. WATER (M) DIRECTION OF WATER ORDER WATER DIST. TO CONF (M) DIRECTION OF CON. 36Bv0003 Ohio River 120 Southwest 3 340 Northwest 36Bv0004 Ohio River 0 On-site 1 170 West 36Bv0029 Ohio River 50 Northwest 9 380 -- 36Bv0036 Raccoon Creek 80 Southeast 1 1000 North 36Bv0040 Trib Raccoon Crk 120 Northeast 2 180 -- 36Bv0045 Raccoon Creek 50 West 9 120 Southwest 36Bv0076 N Fk Kings Crk 40 South 1 320 North 36Bv0078 Raccoon Creek 100 West 4 260 Southwest 36Bv0086 Raccoon Creek 60 East 4 540 Northwest 36Bv0169 Raccoon Creek 60 West 4 220 South 36Bv0180 Raccoon Creek 40 East 4 600 North 36Bv0202 Raccoon Creek 40 North 1 180 West 36Bv0213 Raccoon Creek 60 Southeast 4 910 Southwest 36Bv0248 Ohio River 0 On-site 1 1940 South 36Bv0250 Raccoon Creek 20 Southeast 4 80 Southwest 36Wh0001 Raccoon Creek 150 West 1 170 Southwest 36Wh0022 Raccoon Creek 110 East 1 560 Northeast 36Wh0106 Trib Burgetts Fork 50 North 1 710 East 36Wh0123 Raccoon Creek 90 Southeast 1 300 Northwest 36Wh0297 Cross Creek 60 South 5 400 Southwest 36Wh0338 Raccoon Creek 30 North 1 180 East 36Wh0404 Mid Fk Cross Creek 190 East 2 300 Northeast PA Prehistoric Data Synthesis, Raccoon Creek Watershed 101
SITE NO. NEAREST WATER DIST. WATER (M) DIRECTION OF WATER ORDER WATER DIST. TO CONF (M) DIRECTION OF CON. 36Wh0409 Raccoon Creek 50 East 1 870 East 36Wh0410 Raccoon Creek 10 West 1 1150 South 36Wh0550 Aunt Clara's Branch 20 East 1 100 Southeast 36Wh0758 Raccoon Creek 100 South 5 350 Southeast 36Wh1007 Trib Raccoon Crk 180 South 1 675 West 36Wh1104 Trib Cross Creek 170 Southwest 1 480 South 36Wh1118 Trib Cross Creek 100 Southeast 2 180 Southwest 36Wh1161 Trib Scott Run 800 West 1 1980 North SITE Table 39. Cross-Tabulation of Late Woodland Site by Lithic Raw Material (Percent), Watershed D. ONON- DAGA KANA- WHA FLINT RIDGE UPPER M. BRUSH C. 10 MILE UNION- TOWN GLACIAL PEBBLE Meadowcroft % 5.5 18.2 7.3 36.4 16.4 7.3 5.5 3.6 -- 55 Mungai Farm % 5.5 -- 11.1 27.7 44.4 -- 5.5 -- -- 18 Cross Cr Sites % -- -- -- -- 50.0 -- -- -- 50.0 2 Dravo % 42.9 -- -- 50.0 -- -- -- -- 56.7 30 Total Pts (n) 10 10 6 32 18 4 4 2 19* 105 Total % 9.5 9.5 5.7 60.5 17.1 3.8 3.8 1.9 18.1 100 *includes 14 Delaware chert points (Davis 1988:73) OTHER TOTAL POINTS PASS files provide only limited additional data regarding lithic raw material use at Late Woodland sites in Watershed D. Six singlecomponent Late Woodland sites are present in the watershed, five of which list chert/flint as the only lithic raw material types. One site (36Bv169) also records the presence of Onondaga chert in the lithic assemblage. No sites record Flint Ridge or Upper Mercer cherts on PASS forms. While limited, these data confirm the emphasis on local lithic raw material use during the Late Woodland. E. LATE WOODLAND: SUMMARY AND RESEARCH QUESTIONS Late Woodland Overview PASS files data indicate a preference for Late Woodland site locations in lowland settings, especially compared to previous periods. This settlement tendency likely is related to the everincreasing role of agriculture and village life, both of which occurred largely in river bottoms in eastern North America (Smith 1987). Nevertheless, the low density of true Late Woodland village sites in Watershed D of Subbasin 20 is conspicuous. A comprehensive map of Late Woodland sites in the Upper Ohio Valley by Hart and Nass (1995:25) does not depict a single excavated site in Watershed D. As with the Mahoning River in Watershed A of Subbasin 20 (MacDonald and Cremeens 2002), approximately 50 miles north of the study area, Raccoon and Cross Creeks may have been peripheral to the Monongahela heartland. Nevertheless, PASS data indicate that Late Woodland sites are closer to water, lower in elevation, and more often in alluvial settings than Early and Middle Woodland sites, confirming that agriculture and sedentism increased in river bottoms during the Late Woodland. During the last few hundred years of the Late Woodland, villages in the southern portion of Watershed D and within nearby Chartiers Creek were placed on hilltops and sideslopes. Within Chartiers Creek and the Monongahela River valleys, these villages contained fortified palisades for defensive measures (George 1995). One of the only Late Woodland villages in Watershed D, the Lower Field/Shippingport Site (36Bv4; see Figure 5) on the Ohio River, yielded a burial with a projectile point embedded in its skull, also suggestive of population conflict. The increasing populations of the Late Woodland, as reflected in the increasing site counts in PASS files, likely inspired competition for resources. Climate changes approximately 800 years ago (ca. 1200 A.D.) may have resulted in reduced resource availability as well (Nass and Hart 2000). This combination of factors increasing populations 102 PA Prehistoric Data Synthesis, Raccoon Creek Watershed
and reduced resources may have spurred the reliance on maize agriculture and the increased settlement in alluvial settings. As noted above, analysis of human bone from Late Woodland sites in southwestern Pennsylvania indicates that nearly 70 percent of the Monongahela diet was composed of maize in southwestern Pennsylvania and northern West Virginia (Farrow 1986; Mayer-Oakes 1955). The limited nutritional intake of the maize-base diet apparently led to higher incidences of disease and skeletal deformities for Monongahela populations (Nass and Hart 2000:147). In support of this notion, Davis (1988) reports that one of the Late Woodland burials studied by Mayer-Oakes at Shippingport in Watershed D contained a skeleton with vitamin deficiency and congenital abnormalities. The lack of excavated Late Woodland villages precludes our ability to clearly define the role of the Monongahela in the study area. Given the proximity of the Monongahela heartland, its influence is presumed to have been great. The presence of seven Late Woodland village sites in Watershed D PASS files suggests an increasing role for sedentism and crop use. PASS data also indicate that Late Woodland sites were mostly in lowland settings, possibly to facilitate agriculture. Given the alluvial setting of many Late Woodland sites, burial of sites under late Holocene alluvium is likely within many floodplain and terrace settings of Raccoon Creek (see Geomorphology and Figure 3 above), as occurred at Site 36Wh924 along Chartiers Creek (GAI 1986). As discussed above, the Late Woodland component of this site was buried under 1.5-2.0 meters of alluvium, suggesting the possibility of deeply buried Late Woodland sites within Watershed D as well. area. Ten Late Woodland research questions are listed below; this list is by no means comprehensive and should be used only as a starting point for generating additional research issues. Archaeological sites which can provide information pertaining to these and other research questions will likely meet National Register Criterion D; thus, unless they lack integrity, sites which address these research questions will be eligible for listing in the National Register for Historic Places: 1. How did Late Woodland cultures (e.g., the Monongahela) differ from their predecessors of the Middle and Early Woodland? 2. Was the Monongahela a continuation of Middle Woodland trends or a population replacement within Watershed D? 3. How can we explain the apparent population explosion during the Late Woodland? 4. How prevalent was sedentism and agriculture in Watershed D? 5. Why were Late Woodland sites placed at lower elevations and closer to streams and to stream confluences? 6. Does the increased tendency for Late Woodland sites in alluvial settings reflect increasing agriculture in valley bottoms? 7. What best explains the increase in conflict during the Late Woodland? 8. Did increasing populations deplete resources enough to cause dietary stress? 9. What was the influence of the Monongahela in Watershed D? 10. Why haven t any Monongahela villages been identified in the watershed? Late Woodland Research Questions This summary of Late Woodland archaeological data in Watershed D has generated several research issues which should be considered when conducting archaeological work in the PA Prehistoric Data Synthesis, Raccoon Creek Watershed 103
CHAPTER XI SUMMARY AND CONCLUSIONS This final chapter provides an overview of the major chronological and demographic changes during the last 12,000 years in Watershed D, Subbasin 20, including Raccoon Creek and Cross Creek. PASS data provide a means to assess demographic change over time in this study area. Site counts, site locations, and site types can be assessed for each prehistoric period and compared to evaluate changing patterns during the last 14,000 years of Native American use of the area. A. DEMOGRAPHY AND SETTLEMENT As shown in Figure 23, increases in site counts per decade are concluded to indicate a Native American population increase between the Paleoindian and Early Archaic, continuing through the Brewerton Late Archaic in Watershed D. A second population increase is also suggested by site counts between the Steubenville Late Archaic and the Late Woodland. This report suggests that the two population changes are related to changing subsistence and settlement patterns. Based on the PASS data, the Early Archaic appears to represent a distinct cultural demographic break from the Paleoindian period, as reflected by: 1) increases in population; 2) movement of most sites into alluvial settings, as opposed to uplands; and 3) increased use of a wide variety of landforms. The two periods share similar patterns of lithic raw material use, including wide-ranging travel patterns within southwestern Pennsylvania, northern West Virginia, and east-central Ohio, as reflected in the use of Kanawha chert, Flint Ridge chert, Upper Mercer chert, and Brush Creek chert. The increasing availability of riverine and other resources between the Paleoindian and Late Archaic allowed the exploitation of a wider range of foods, with a culmination during the Brewerton Late Archaic. While Carr (1998a, 1998b) and MacDonald (2003a) have suggested that the major demographic change occurred between the Early and Middle Archaic in other parts of Pennsylvania, the PASS data for Watershed D of Subbasin 20 indicate population continuity between these periods, with the most significant increases between the Paleoindian and Early Archaic and the Middle Archaic and Brewerton Late Archaic, as depicted in Figure 23. This report hypothesizes that the reduced site counts between the Brewerton Late Archaic and the Steubenville Late Archaic may best be explained by the increased tendency for collective processing of foods, resulting in nucleation of populations into fewer sites. Thus, in this scenario, Brewerton populations were foragers moving to and from resource procurement areas, establishing short-term camps in a variety of settings, and thereby yielding Brewerton points from many small sites. While still incorporating a wide use of the landscape, Steubenville populations began to congregate to exploit concentrated resources; thus, a few large sites yield Steubenville points. The large shell heaps at Steubenville sites in the West Virginia panhandle support this hypothesis, as does the ubiquity of Brewerton points at small lithic scatters in the Upper Ohio Valley (Mayer-Oakes 1955; Mohney 2002). As such, the reduced Steubenville site counts likely reflect population congregation at sites, rather than population reduction. If increasing resource availability stimulated population growth between the Paleoindian and Brewerton Late Archaic (as reflected in Figure 23), we should expect the corresponding use of a wider variety of landscapes over the early and middle Holocene. Site location data from PASS files appear to confirm this scenario (Figure 24), as Middle Archaic and Brewerton Late Archaic sites tend to be more in uplands (ca. 54%) than lowlands (ca. 46%), compared to the Early Archaic in which most sites are in lowlands 104 PA Prehistoric Data Synthesis, Raccoon Creek Watershed
(61%). This increasing use of uplands between the Early and Late Archaic is mirrored in mean elevation of sites, as the elevation increased by 100 ft. between the Early Archaic (1,027 ft. amsl) and Late Archaic (1,128 ft. amsl). Mean distance to water and mean distance to confluence also increased between the Early Archaic (102m and 588m) and Late Archaic (126m and 691m), another indication that populations were moving across the landscape and more frequently into uplands. Sites/Decades per Period Percent Figure 23. Site Counts per Decade Per Period, Watershed D, Subbasin 20. 80 70 60 50 40 30 20 10 0 0.5 0.45 0.4 0.35 0.3 0.25 0.2 0.15 0.1 0.05 0 Paleo 0.02 0.03 EA 0.15 0.13 0.34 Early Paleo Paleo EA MA Brewerton Steubenville MA Brewerton Period Steubenville 0.16 0.18 0.23 EW MW LW Figure 24. Variation in Site Location, Watershed D, Subbasin 20. 0.43 Upland/Saddle/Slope Alluvial EW Period MW LW These trends of increased use of uplands and use of a wider variety of settings between the Early and Late Archaic were reversed between the Late Archaic and Late Woodland. Given the increasing role of sedentism, agriculture, and village life of the later Woodland, one might expect an increased tendency for sites in alluvial settings over the course of the Woodland (see Figure 24). This prediction is confirmed by site location data in Figure 24, which indicate an increase in use of alluvial settings (63% of sites) during the Late Woodland. Accordingly, the mean elevation of sites also decreases to 961 ft. amsl during the Late Woodland, compared to 1,128 ft. amsl during the Late Archaic (Table 40). Mean distance to water and mean distance to stream confluence is also lowest during the Late Woodland, confirming the tendency to move to valley bottoms conducive to agriculture, a trend that occurred throughout the Woodland in the eastern United States (Smith 1987). Table 40. Site Location Data, All Periods, Watershed D, Subbasin 20. PERIOD MEAN DIST TO WATER (M) MEAN DIST. TO CONF. (M) MEAN ELEVATION (FT. AMSL) Paleoindian 212 789 1208 E.Archaic 102 588 1027 M.Archaic 150 621 1052 L.Archaic 126 691 1128 E.Woodland 130 597 1101 M.Woodland 122 720 1073 L.Woodland 98.3 522 961 Thus, while botanical evidence for agriculture is minimal for Late Woodland sites in Watershed D (see Chapter X), site location data indicate that farming may have been an important subsistence activity. Populations clearly moved to river bottoms, most likely to grow crops. As noted in Chapter X, studies of Monongahela diet indicate a 70 percent reliance on maize agriculture. Additional excavations at Late Woodland sites along valley bottoms of Raccoon Creek are necessary to better understand the role of farming for Late Woodland populations in Raccoon Creek and vicinity. PA Prehistoric Data Synthesis, Raccoon Creek Watershed 105
B. LITHIC RAW MATERIAL USE AND SETTLEMENT PATTERNS Lithic raw material data provide a means to evaluate travel and trade patterns and technological choices over the last 14,000 years in Watershed D. Fortunately, the extensive excavations at sites in Cross Creek yielded excellent lithic raw material data, as described in the previous chapters. This section synthesizes those data to compare land-use patterns over time. For each period, projectile point data from the Meadowcroft, Mungai Farm, Cross Creek Village, and Avella Mound Sites are combined with the general survey results presented in Vento and Donohue (1982:124-127). More detailed analyses of lithic raw materials are presented for each period in the chapters above. To better understand patterns of toolstone use, the lithic raw materials were grouped into three categories in Table 41: local, non-local, and Kanawha chert. The local cherts include Onondaga, Brush Creek, Ten Mile, Uniontown, and cobble chert, while non-local cherts include Flint Ridge and Upper Mercer. Kanawha chert is separated due to the uncertain provenience of the material, with some suggesting that the most proximate sources are in central West Virginia (Vento and Donohue 1982) and others (Fritz, pers. comm., 2003; Mohney 2002) indicating possible local sources for Kanawha-like chert in southwestern Pennsylvania; thus, its sources could be either local or non-local. Unknown cherts were present as well, accounting for the remainder in each period. Table 41. Lithic Raw Material Use over Time, Watershed D, Subbasin 20. PERIOD LOCAL NON-LOCAL KANAWHA Paleoindian 34.6* 37.7 0.0 Early Archaic 43.1 45.1 8.8 Middle Archaic 42.4 30.3 21.2 Late Archaic 48.2 28.2 18.2 Early Woodland 21.8 50.0 20.2 Middle Woodland 36.1 53.5 6.7 Late Woodland 42.6 41.3 13.3 Total, Mean % 38.4 40.9 12.6 *Data synthesized from Vento and Donohue (1982:124-127); only Meadowcroft data used for Paleoindian period. The most obvious point to make about the data in Table 41 is its general uniformity over time. On average, local cherts were used to produce 38.4 percent of artifacts at the Cross Creek sites, while non-local cherts were used to produce 40.9% of artifacts. Variation from this pattern occurs during the Early Woodland, Middle Woodland, Middle Archaic, and Late Archaic. During the Middle and Late Archaic, local cherts dominate the assemblages of the sites, with significant quantities of Kanawha chert as well. This use of a variety of locally available stones likely reflects the ever-increasing use of a variety of landforms, discussed in more detail above and in Chapters VI and VII. Rather than traveling long distances therefore, Middle and Late Archaic Native Americans traveled extensively within a well-defined local territory. The Early and Middle Woodland reliance on non-local Flint Ridge and Upper Mercer cherts represents another significant variation from the normal toolstone-use pattern (see Table 41). This peak in non-local chert use reflects increased ceremonial ties to east-central Ohio Adena and Hopewell populations. As discussed in more detail in Chapters IX and X, the Adena and Hopewell influences were not great in Watershed D, but were apparent in the lithic raw material assemblages at sites like Mungai Farm and Avella Mound, both of which yielded very high proportions of Flint Ridge and Upper Mercer cherts. C. CONCLUSION: FUTURE AVENUES OF RESEARCH IN WATERSHED D, SUBBASIN 20 One of the main results of this study is the realization that very few in-depth archaeological studies (beyond the Phase I survey level) have occurred within the Raccoon Creek Valley proper. No research reports at all are available for several smaller streams in the northwestern portion of the study area, including Harmon Creek, King s Creek, and the Aunta Clara s Fork, all of which flow westerly into the Ohio River. With the exception of the outstanding work conducted within Cross Creek at sites like Meadowcroft Rockshelter by Adovasio and his colleagues, only one data recovery investigation 106 PA Prehistoric Data Synthesis, Raccoon Creek Watershed
has taken place in Watershed D, the work at the Dravo Site by Davis (1988) on the south bank of the Ohio River near the Ohio state line. While Mayer-Oakes (1955) conducted work at several sites in this area, most of it entailed analysis of existing collections. When excavations were conducted by Mayer-Oakes, the lack of available technical data limited their ability to address specific research questions. Hundreds of sites are recorded in PASS files in the Raccoon Creek Valley, thus the minimal extent of archaeological work is astonishing, especially given the extensive studies that have taken place nearby in the Ohio River, Cross Creek, and Chartiers Creek. While the prehistory of these areas is very well known, especially for certain time periods, the prehistory of the Raccoon Creek Valley is a relative unknown and demands additional research. ability to answer research questions discussed in this report. Even the smallest of sites may provide pieces of information regarding landuse, technology, and subsistence that would contribute to a better understanding of prehistoric change within Raccoon Creek, Cross Creek, Traverse Creek, Little Traverse Creek, Service Creek, King s Creek and the many other drainages of Watershed D, Subbasin 20. While many upland sites are likely to reveal their presence by surface scatters of artifacts in plowed fields (e.g., Mungai Farm), sites in alluvial settings may be buried under more than a meter of late Holocene alluvium (see Geomorphology and Figure 3 above; MacDonald 2003b). For example, Site 36Wh924 along Chartiers Creek contained a Late Woodland Monongahela component with features and shell-tempered ceramics between 1.5 and 2.0 meters below ground surface (GAI 1986). As currently required by the BHP, deep testing is recommended in floodplain/terrace settings of larger streams and rivers within Watershed D. More work clearly needs to be conducted to better evaluate the prehistory of Watershed D, especially within the Raccoon Creek Valley and its smaller tributaries. Within Cross Creek, the outstanding work by Adovasio and his colleagues provided a wealth of information that should be synthesized in a comprehensive technical report. While PASS data are also an excellent means to evaluate changing lifeways and demographics over time in Watershed D, more in-depth archaeological projects are required to address the research issues. Archaeological sites within the Raccoon Creek watershed should be evaluated to assess their PA Prehistoric Data Synthesis, Raccoon Creek Watershed 107
(This page intentionally blank.) 108 PA Prehistoric Data Synthesis, Raccoon Creek Watershed
REFERENCES Adovasio, J.M., R. Fryman, A.G. Quinn, D.C. Dirkmaat, and D.R. Pedler 1998 The Archaic West of the Allegheny Mountains: A View from the Cross Creek Drainage, Washington County, Pennsylvania. In The Archaic Period in Pennsylvania, pp. 1-28, edited by P.A. Raber et al. Pennsylvania Historical and Museum Commission, Harrisburg. Adovasio, J.M., J. Gunn, J. Donahue, and R. Stuckenrath 1977 Meadowcroft Rockshelter: Retrospect 1976. Pennsylvania Archaeologist 47(2-3):1-93. 1978 Meadowcroft Rockshelter, 1977; An Overview. American Antiquity 43 (4): 632-651. 1982 The Meadowcroft Rockshelter/Cross Creek Archaeological Project: Retrospect 1982. In Meadowcroft: Collected Papers on the Archaeology of Meadowcroft Rockshelter and the Cross Creek Drainage, pp. 257-270, edited by R.Carlisle and J.Adovasio. Papers Prepared for 1982 Society for American Archaeology Meeting, Minneapolis, MN. 1983 The AENA Compilation of Fluted Points in Eastern North America: A Perspective from Meadowcroft Rockshelter. Archaeology of Eastern North America 11: 6-11. 1990 The Meadowcroft Rockshelter Radiocarbon Chronology 1975-1990. American Antiquity 55(2):348-354. Adovasio, J.M., J. Gunn, J. Donahue, R. Stuckenrath, J.E. Guilday, and K.Volman 1980 Yes Virginia, It Really is that Old: A Reply to Haynes and Mead. American Antiquity 45(3): 588-596. Adovasio, J.M. and W.C. Johnson 1981 The Appearance of Cultigens in the Upper Ohio Valley: A View from Meadowcroft Rockshelter. Pennsylvania Archaeologist 51(1-2):63-80. Adovasio, J.M. and J. Page 2002 The First Americans. Random House, New York. Adovasio, J.M. and D.R. Pedler 1997 Monte Verde and the Antiquity of Humankind in the Americas. Antiquity 71:573-580. 2000 A Long View of Deep Time at Meadowcroft Rockshelter. Paper presented at the 65 th Society for American Archaeology Meeting, Philadelphia. Alam, E.A. 1961 A Preliminary Report on a Stratified Site at Ohioview. Pennsylvania Archaeologist 31(2):61-77. PA Prehistoric Data Synthesis, Raccoon Creek Watershed 109
Applegarth, J.D. and V.L. Cowin 1982 Excavations at Cross Creek Village (36Wh293) and the Avella Mound (36Wh415), Washington County, Southwestern Pennsylvania. In Meadowcroft: Collected Papers on the Archaeology of Meadowcroft Rockshelter and the Cross Creek Drainage, pp. 241-256, edited by R.Carlisle and J.Adovasio. Papers Prepared for 1982 Society for American Archaeology Meeting, Minneapolis, MN. Baker, P.H. 1996 Phase I Cultural Resource Survey of the Proposed 55±Acre Beaver Valley Retail Project, Center Township, Beaver County, Pennsylvania, ER 96-0753-007. Submitted to JDN Development Company, Inc. By Garrow and Associates, Inc., Latrobe, Pennsylvania. Ballweber, H.L. 1989 A Study of Six Early/Middle Woodland Sites in Westmoreland County, Pennsylvania. Pennsylvania Archaeologist 59(2):46-99. Bennett, R.N., and W.D. Porter 1986 Cultural Resources Survey of a Proposed 11.7 Mile Pipeline Corridor in Kanawha County, West Virginia. Prepared by WAPORA, Inc. For Columbia Gas Transmission Corporation, Charleston, West Virginia. Berg, Thomas M., John H. Barnes, W. D. Sevon, Viktoras W. Skema, I. Peter Wilshusen, and Dawna S. Yannacci 1989 Physiographic Provinces of Pennsylvania. Prepared for the Commonwealth of Pennsylvania, Department of Environmental Resources, Office of Resources Management, Bureau of Topographic and Geologic Survey. Berger, Louis and Associates, Inc. (Berger) 1998 Phase II Archaeological Investigation, Site 36Bv122, Proposed Little Traverse Creek Bridge Replacement, SR 0030, Section B05, Independence Township, Beaver County, Pennsylvania, ER 92-3454-007. Prepared for PennDOT District 11-0 by Louis Berger and Associates, Inc., East Orange, NJ. Bergman, C., J. Doershuk, R.Moeller, P.LaPorta, and J.Schuldenrein 1998 An Introduction to the Early and Middle Archaic Occupations at Sandts Eddy. In The Archaic Period in Pennsylvania: Hunter-Gatherers of the Early-Middle Holocene, pp. 45-76, edited by. P.Raber et al. Pennsylvania Historical and Museum Commission, Harrisburg. Beynon, D.E. 1981 The Geoarchaeology of Meadowcroft Rockshelter. Unpublished Ph.D. Dissertation, Department of Anthropology, University of Pittsburgh. Binford, L. R. 1979 Organization and Formation Processes: Looking at Curated Technologies. Journal of Anthropological Research 35:255-273. 1980 Willow Smoke and Dogs Tails: Hunter-Gatherer Settlement Systems and Archaeological Site Formation. American Antiquity 45(1):4-21. 110 PA Prehistoric Data Synthesis, Raccoon Creek Watershed
Boldurian, A.T. 1985 Variability in Flintworking Technology at the Krajacic Site: Possible Relationships to the Pre-Clovis Paleoindian Occupation of the Cross Creek Drainage in Southwestern Pennsylvania. Unpublished doctoral dissertation, University of Pittsburgh. Bonnichsen, Robson, and Karen L. Turnmire, eds. 1991 Clovis: Origins and Adaptations. Center for the Study of the First Americans, Corvallis, Oregon. Braun, E.L. 1950 Deciduous Forests of Eastern North America. The Blakiston Company, Philadelphia. Bressler, J.P. 1989 Prehistoric Man on Canfield Island 36Ly37, Lycoming County, Pennsylvania. North Central Chapter No. 8, Society for Pennsylvania Archaeology and the Lycoming County Historical Society. Briggs, R.P. 1999 Appalachian Plateaus Province and the Eastern Lake Section of the Central Lowland Province. In The Geology of Pennsylvania, edited by C.H. Shultz, pp. 362-377. Pennsylvania Geological Society, Harrisburg, and Pittsburgh Geological Society. Broyles, B.J. 1971 Second Preliminary Report: The St. Albans Site, Kanawha County, West Virginia, 1964-1968. West Virginia Geological and Economic Survey, Report of Archaeological Investigations 3, Morgantown. Buker, W.E. 1968 The Archaeology of McKees Rocks Late Prehistoric Village Site. Pennsylvania Archaeologist 38(1-4): 3-49. Burkett, Kenneth 1999 Prehistoric Occupations at Fishbasket. Pennsylvania Archaeologist 69(1):1-100. Butts, C.F. 1945 Geologic Atlas of the United States, Hollidaysburg-Huntingdon Folio Pennsylvania. Pennsylvania Geologic Survey, Harrisburg. Callahan, E. 1979 The Basics of Biface Knapping in the Eastern Fluted Point Tradition: A Manual for Flintknappers and Lithic Analysts. Archaeology of Eastern North America 7(1):1-180. Carlisle, R.C. and J.M. Adovasio 1982 Meadowcroft: Collected Papers on the Archaeology of Meadowcroft Rockshelter and the Cross Creek Drainage. Prepared for the 47 th Annual Society for American Archaeology Meeting, Minneapolis. PA Prehistoric Data Synthesis, Raccoon Creek Watershed 111
Carlisle, R.C., J.M. Adovasio, J. Donahue, P. Wiegman, and J.E. Guilday 1982 An Introduction to the Meadowcroft/Cross Creek Archaeological Project: 1973-1982. In Meadowcroft: Collected Papers on the Archaeology of Meadowcroft Rockshelter and the Cross Creek Drainage, pp. 1-30, edited by R.Carlisle and J.Adovasio. Papers Prepared for 1982 Society for American Archaeology Meeting, Minneapolis, MN. Carr, K.W. 1989 The Shoop Site: Thirty-Five Years After. In New Approaches to the Past, edited by W.F. Kinsey and R.W. Moeller, pp. 5-28. Archaeological Services, Bethlehem, Ct. 1998a The Early Archaic Period in Pennsylvania. Pennsylvania Archaeologist 68(2):42-69. 1998b Archaeological Site Distribution and Patterns of Lithic Utilization during the Middle Archaic in Pennsylvania. In The Archaic Period in Pennsylvania: Hunter-Gatherers of the Early and Middle Holocene Period, edited by P.Raber, P.Miller, and S.Neusius, pp. 77-90. Pennsylvania Historical and Museum Commission, Harrisburg. Carr, K.W. and Adovasio, J.M. 2002a Paleoindians in Pennsylvania. In Ice Age Peoples of Pennsylvania, edited by K.W. Carr and J.M. Adovasio, pp. 1-50. Recent Research in Pennsylvania Archaeology No. 2, Pennsylvania Historical and Museum Commission, Harrisburg. 2002b Ice Age Peoples of Pennsylvania. Recent Research in Pennsylvania Archaeology No. 2, Pennsylvania Historical and Museum Commission, Harrisburg. Chatters, J.C. 2000 The Recovery and First Analysis of an Early Holocene Human Skeleton from Kennewick, Washington. American Antiquity 65(2):291-316. Church, F. 1995 Monongahela Subsistence and Settlement in the Northern West Virginia Panhandle: the Saddle Site. Archaeology of Eastern North America 23:57-72. Clarke, G. and C.M. Barton 2004 The Settlment of the American Continents. University of Arizona Press, Tucson. In Press. Clay, R. B. 1980 The Cultural Historical Placement of Fayette Thick Ceramics in Central Kentucky, Tennessee Anthropologist V(2): 166-178. 1986 Adena Ritual Spaces. In Early Woodland Archeology, edited by Kenneth B. Farnsworth and Thomas E. Emerson, pp. 581-595. Center for American Archeology Press, Kampsville, Illinois. 1987 Circles and Ovals: Two Types of Adena Space. Southeastern Archaeology 6(1): 46-56. 112 PA Prehistoric Data Synthesis, Raccoon Creek Watershed
1991 Adena Ritual Development: An Organizational Type in a Temporal Perspective. In The Human Landscape in Kentucky s Past: Site Structure and Settlement Patterns, edited by C. Stout and C.K. Hensley, pp. 30-39. Kentucky Heritage Council. 1998 The Essential Features of Adena Ritual and Their Implications. Southeastern Archaeology 17(1):1-21. Clay, R. B. and C. M. Niquette 1992 Middle Woodland Mortuary Ritual in the Gallipolis Locks and Dam Vicinity, Mason County, West Virginia. West Virginia Archeologist 44(1&2):1-25. Coe, J.L. 1964 The Formative Cultures of the Carolina Piedmont. Transactions of the American Philosophical Society 54(5), Philadelphia, Pa. Collins, M.B. 1999 Clovis Blade Technology. University of Texas Press, Austin. Cosgrove, K.L. and R.L. Michael 1986 Phase II Archaeological Testing of Sites 2A, 3A, 3B, and 3C, St. Joe Resources Company, Fly-Ash Disposal Project, Potter Township, Beaver County, Pennsylvania, ER 86-0679-007. Submitted to Michael Baker, Jr., Inc. By NPW Consultants, Inc., Uniontown, Pennsylvania. Cowin, Verna L. 1985 The Woodland Period. In A Comprehensive State Plan for the Conservation of Archaeological Resources, Volume II, edited by Paul A. Raber, pp. 185-193. Pennsylvania Historical and Museum Commission, Harrisburg. 1991 The Middle Archaic in the Upper Ohio Valley. In Journal of Middle Atlantic Archaeology. Vol. 7, pp.43-53. Crowl, G.H. and W.D. Sevon 1999 Quaternary. In The Geology of Pennsylvania, edited by C.H. Shultz, pp. 225-231. Pennsylvania Geological Society, Harrisburg, and Pittsburgh Geological Society. Cushman, K.A. 1982 Floral Remains from Meadowcroft Rockshelter, Washington County, Southwestern Pennsylvania. In Meadowcroft: Collected Papers on the Archaeology of Meadowcroft Rockshelter and the Cross Creek Drainage, pp. 207-220, edited by R.Carlisle and J.Adovasio. Papers Prepared for 1982 Society for American Archaeology Meeting, Minneapolis, MN. Custer, J.F. 1988 Late Archaic Cultural Dynamics in the Central Middle Atlantic Region. Journal of Middle Atlantic Archaeology 4: 39-60. 1996 Prehistoric Cultures of Eastern Pennsylvania. Pennsylvania Historical and Museum Commission, Harrisburg. PA Prehistoric Data Synthesis, Raccoon Creek Watershed 113
2001 Classification Guide for Arrowheads and Spearpoints of Eastern Pennsylvania and the Central Mid-Atlantic. Pennsylvania Historic and Museum Commission, Harrisburg. Custer, J.F., and E.B. Wallace 1982 Patterns of Resource Distribution and Archaeological Settlement Patterns in the Piedmont Uplands of the Middle Atlantic Region. North American Archaeologist 3(2): 139-172. Custer, J.F., Scott C. Watson, and Daniel N. Bailey 1996 A Summary of Phase III Data Recovery Excavations at the West Water Street Site (36Cn175), Lock Haven, Clinton County, Pennsylvania. Pennsylvania Archaeologist 66(1): 1-53. Davis, C. E. n.d. Georgetown Site (36Bv29). National Register of Historic Places Registration Form. Submitted to the United States Dept. of the Interior, National Park Service by Christine Davis. 1988 Phase III Data Recovery, the Dravo #1 Site, 36Bv240, ER 87-0879-007. Prepared for Dravo Corporation by the Carnegie, Section of Anthropology. 2001 Phase I Archaeological Survey and Limited Phase II Testing for the Leetsdale Concrete Casting Facility, Allegheny County, Pennsylvania, ER 87-0469-042. Submitted to the PHMC-BHP by Christine Davis Consultants, Inc., Verona, Pennsylvania. Davis, C.E. and S.W. Lantz 1987 Phase I/II Cultural Resource Survey of the Georgetown Project, Beaver County, Pennsylvania. Report submitted by the Carnegie Museum, Pittsburgh, to Dravo Corporation (ER No. 97-0879-007). Davis, M.B. 1984 Holocene Vegetational History of the Eastern United States. In Late Quaternary Environments of the United States, edited by H.E. Wright, Jr., pp. 166-181. Longman Press, London. Dent. R.J., Jr. and B.E. Kauffman 1985 Aboriginal Subsistence and Site Ecology as Interpreted from Microfloral and Faunal Remains. In Shawnee Minisink: A Stratified Paleoindian Site in the Upper Delaware Valley of Pennsylvania, pp. 55-79, edited by C.W. McNett. Academic Press, New York. Dillehay, T.D. 1997 Monte Verde: A Late Pleistocene Settlement in Chile: Paleoenvironment and Site Context. Smithsonian Institution Press, Washington, D.C. Dincauze, D.F. 1976 The Neville Site: 8,000 Years at Amoskeag, Manchester, New Hampshire. Peabody Museum Monographs No. 4. Harvard University, Boston. 114 PA Prehistoric Data Synthesis, Raccoon Creek Watershed
Dragoo, D.W. 1956 Excavations at the Watson Site, 46Hk34, Hancock County, West Virginia. Pennsylvania Archaeologist 26(2):59-88. 1959 Archaic Hunters of the Upper Ohio Valley. Annals of Carnegie Museum 35: 139-245. 1963 Mounds for the Dead: An Analysis of the Adena Culture. Annals of Carnegie Museum 37. Edmunds, W.E., Skema, V.W., and N.K. Flint 1999 Pennsylvanian. In Shultz, C.H. (ed.) The Geology of Pennsylvania, pp.148-170. Harrisburg, Pennsylvania Geological Survey and Pittsburgh, Pittsburgh Geological Society. East, Thomas C. and K.A. Beckman 1992 Blair County S.R.6220 Wetland Replacement. Phase I and Phase II Cultural Resources Report Prepared for Pennsylvania Department of Transportation District 9-0 by Skelly and Loy, Inc. Engineers-Consultants, Monroeville and Harrisburg, PA. East, T.C., R.B. Duncan, M.G. Sams, C.T. Espenshade, P.H. Baker, and A.T. Vish 1999 Phase I/II Archaeological Testing and Phase III Data Recovery, U.S. Route 220 Improvement Project, S.R. 0220, Section C10. Report submitted to PennDOT District 2-0 by Skelly and Loy, Inc., Monroeville, Pa. Eddins, J.T. 1982 Archaeological Site Survey of Defined Areas within Raccoon Creek State Park, Beaver County, Pennsylvania. Submitted to the Dept. of Environmental Resources by the Carnegie Museum of Natural History, Pittsburgh. Eisert, Ronald W. 1974 Monongahela Flint and Associated Workshops in the Chartiers Valley, Washington County, Pennsylvania. Pennsylvania Archaeologist 44(3):33-39. Ellis, C. and J.H. Payne 1995 Estimating Failure Rates in Fluting Based on Archaeological Data: Examples from NE North America. Journal of Field Archaeology 22:459-474. Ellyson, W.J., Kunkle, M., Ruffner, J.D., and J. Webb 1974 Soil Survey of Brooke, Hancock, and Ohio Counties, West Virginia. USDA- Soil Conservation Service, Washington D.C., Govt. Printing Office Faingnaert, D. and W. Doyle 1977 A Report on the Skeletal Remains of the Ohioview Site, A Late Prehistoric Village. Pennsylvania Archaeologist 47(4):8-26. Farrow, D.C. 1986 A Study of Monongahela Subsistence Patterns Based on Mass Spectrometric Analysis. Midcontinental Journal of Archaeology 11(2):153-180. PA Prehistoric Data Synthesis, Raccoon Creek Watershed 115
Fenicle, D. 2003 Draft Report: Phase II Archaeological Testing: 36AL480, Leetsdale Industrial Park, Leetsdale, Allegheny County, Pennsylvania, ER 87-0469-042. Submitted to USACOE Pittsburgh District by the ASC Group, Inc., Columbus, Ohio. Fenneman, N.M. 1938 Physiography of Eastern United States. McGraw-Hill Book Company, New York. Fiedel, S.J. 2001 What Happened in the Early Woodland? Archaeology of Eastern North America 29: 101-142. Fitzgibbons, P.T. 1982 Lithic Artifacts from Meadowcroft Rockshelter and the Cross Creek Drainage. In Meadowcroft: Collected Papers on the Archaeology of Meadowcroft Rockshelter and the Cross Creek Drainage, pp. 91-111, edited by R.Carlisle and J.Adovasio. Papers Prepared for 1982 Society for American Archaeology Meeting, Minneapolis, MN. Fladmark, K.R. 1983 Times and Places: Environmental Correlations of Mid-to-Late Wisconsinan Human Population Expansion in North America. In Early Man in the New World, pp. 13-42, edited by R. Shutler. Sage Publications, Beverly Hills, Ca. Fogelman, G.L. 1983 The Pennsylvania Artifact Series Lithics Book. Paulhamus Litho, Inc. Fowler, D. B., E. T. Hemmings, and G. R. Wilkins 1976 Some Recent Additions to Adena Archaeology in West Virginia. Archaeology of Eastern North America 4:110-121. Frankenberg, S. R. and G. E. Henning 1994 Phase III Data Recovery Excavations of the Cotiga Mound (46MO1), Mingo County, West Virginia. GAI Consultants, Inc. Project Number 90-509-12. Fredlund, Glen G. 1989 Holocene Vegetational History of the Gallipolis Locks and Dam Project Area, Mason County, West Virginia. Cultural Resource Analysts, Inc. Contract Publication Series 89-01, Lexington, Kentucky. Frison, G.C. 1991 Prehistoric Hunters of the High Plains. Academic Press, New York. Frye, L. and C. Weiser 2003 The Leetsdale Project. PAC Newsletter 24:3-6. GAI Consultants, Inc. 1986 Phase II Archaeological Investigations of Seven Prehistoric Sites in Three Sewerage System Subdistricts, Washington County, Pennsylvania. Report submitted to North Strabane Township Municipal Authority by GAI Consultants, Inc., Monroeville, Pa. 116 PA Prehistoric Data Synthesis, Raccoon Creek Watershed
1995 Archaeological Investigations at the Memorial Park Site (36Cn164), Clinton County, Pennsylvvania. Final Report Submitted to the U.S.A.C.O.E., Baltimore District by GAI, Monroeville, Pennsylvania. Gajewski, K. 1988 Late Holocene Climate Changes in Eastern North America estimated from Pollen Data. Quaternary Research 29:255-262. Gardner, W. M. 1987 Paleoindian and Archaic in West Virginia: An Overview for the State Historic Preservation Plan. West Virginia State Historic Preservation Office, Charleston. 2002 The Paleoindian Problem Revisited: Observations on Paleoindian in Pennsylvania (A Slightly Shorter Slant). In Ice Age Peoples of Pennsylvania, pp. 97-104, edited by K. Carr and J. Adovasio. Pennsylvania Historical and Museum Commission, Harrisburg. Gates, D.M. 1993 Climate Change and its Biological Consequences. Sinauer Associates, Inc., Sunderland, Massachusetts. Gatus, T. W. 1985 Chert Resources and Patterns of Prehistoric Acquisition and Utilization in the Yatesville Reservoir Area of Extreme Northeastern Kentucky. In Prehistoric and Historic Sites Archeology in the Proposed Yatesville Reservoir, Lawrence County, Kentucky, edited by Charles M. Niquette and Theresa K. Donham, pp. 421-432. Cultural Resource Analysts, Inc. Contract Publication Series 85-13, Lexington, Kentucky. Gaudreau, Denise C. 1988 The Distribution of Late Quaternary Forest Regions in the Northeast. In Holocene Human Ecology in Northeastern North America, edited by G.P. Nicholas, pp. 215-256. Plenum Press, New York. George, R.L. 1971 The Archaic of the Upper Ohio Valley: A View in 1970. Pennsylvania Archaeologist 41(1-2):9-15. 1974 Monongahela Settlement Patterns and the Ryan Site. Pennsylvania Archaeologist 44(1-2):1-22. 1975 Some Woodland Point Types in the Upper Ohio Valley. The West Virginia Archaeologist 24:22-31. 1978 The McJunkin Site, A Preliminary Report. Pennsylvania Archaeologist 48(4):33-47. 1985 The Archaic Period. In A Comprehensive State Plan for the Conservation of Archaeological Resources, edited by P.Raber, pp. 181-184. Pennsylvania Historical and Museum Commission, Harrisburg. PA Prehistoric Data Synthesis, Raccoon Creek Watershed 117
1992 Early Late Woodland in Western Pennsylvania; the Backstrum Side Notched Point Evidence. Pennsylvania Archaeologist 62(1):62-72. 1995 Getting High: Chartiers Valley Monongahela and the Troublesome 14 th Century. Archaeology of Eastern North America 23:27-40. 1998 McFate Artifacts in a Monongahela Context: McJunkin, Johnston, and Squirrel Hill. Pennsylvania Archaeologist 67(1):35-44. George, R.L. and C.E. Davis 1986 A Dated Brewerton Component in Armstrong County, Pennsylvania. Pennsylvania Archaeologist 56(1-2):12-20. Gleach, F.W. 1985 A Compilation of Radiocarbon Dates with Applicability to Central Virginia. Quarterly Bulletin of the Archeological Society of Virginia: 40(4):180-200. Goldberg, P. and T.L. Arpin 1999 Micromorphological Analysis of Sediments from Meadowcroft Rockshelter, Pennsylvania: Implications for Radiocarbon Dating. Journal of Field Archaeology 26(3):325-342. Grantz, D.L. 1986 Archaeological Investigations of the Crawford-Grist Site #2 (36FA262): An Early Woodland Hamlet. Pennsylvania Archaeologist 56(3-4):2-21. Griffin, J.B. 1974 The Adena People. The University of Tennessee Press, Knoxville. Guilday, J.E., P.S. Martin, and A.D. McCrady 1964 New Paris No. 4: A Pleistocene Cave Deposit in Bedford County, Pennsylvania. Bulletin of the National Speleological Society 26:121-194. Guilday, J.E. and P.W. Parmalee 1982 Vertebrate Faunal Remains from Meadowcroft Rockshelter, Washington County, Pennsylvania: Summary and Interpretation. In Meadowcroft: Collected Papers on the Archaeology of Meadowcroft Rockshelter and the Cross Creek Drainage, pp. 163-174, edited by R.Carlisle and J.Adovasio. Papers Prepared for 1982 Society for American Archaeology Meeting, Minneapolis, MN. 1965 Animal Remains from Sheep Rock Shelter (36Hu1), Huntingdon County, Pennsylvania. Pennsylvania Archaeologist 35:34-49. Hardesty, D. 1964 The Biggs Site: A Hopewellian Complex in Greenup County, Kentucky. Probes 14-21. 118 PA Prehistoric Data Synthesis, Raccoon Creek Watershed
Hardlines Design Company 2001 Interim Report: Prehistoric Deposits at 36AL480, Leetsdale Industrial Park, Leetsdale, Allegheny County, Pennsylvania. Submitted to the USACOE Pittsburgh District by Hardlines Design Company, Columbus, Ohio. Harper, J.A. 1997 Of ice and waters flowing: the formation of Pittsburgh s three rivers. Pennsylvania Geology 28: 2-8. Hart, J.P. 1995a Summary and Conclusions. In Final Report, Archaeological Investigations at the Memorial Park Site (36CN164), Clinton County, Pennsylvania, pp. 507-548. Final Report Submitted to the U.S. Army Corps of Engineers, Baltimore. GAI Consultants, Inc., Monroeville, PA. 1995b Storage and Monongahela Subsistence-Settlement Change. Archaeology of Eastern North America 23:41-56. Hart, John P. and J.P. Nass, Jr. 1995 Modeling Monongahela Subsistence-Settlement Change: Introduction. Archaeology of Eastern North America 23:23-26. Haynes, C.V. 1980 Paleoindian Charcoal from Meadowcroft Rockshelter: Is Contamination a Problem? American Antiquity 45(3):582-588. Hemmings, E. T. 1977 Investigations at Grave Creek Mound 1975-76: A Sequence for Mound and Moat Construction. The West Virginia Archaeologist 36(2):59-68. 1978 Exploration of an Early Adena Mound at Willow Island, West Virginia. Report of Archaeological Investigation No. 7. West Virginia Geological and Economic Survey, Morgantown, West Virginia. 1984 Investigations at Grave Creek Mound 1975-1976: a Sequence for Mound and Moat Construction. West Virginia Archaeologist 38(2):3-49. 1985 West Virginia Radiocarbon Dates and Prehistory. West Virginia Archaeologist 37(2):35-44. Herbstritt, James T. 1981a Bonnie Brook: A Multicomponent Aboriginal Locus in West-Central Pennsylvania. Pennsylvania Archaeologist 51(3):1-59. 1981b Final Report, Prehistoric Archaeological Site Survey in Pennsylvania Region II, Southwestern Pennsylvania 1980. Prepared for the PHMC by the Center for Prehistoric and Historic Site Archaeology, California, Pennsylvania. Holland, John D. 1999 The Cherts of Northwestern Pennsylvania: Indigenous and Otherwise. Paper presented at the 70 th Annual Meeting of the Society for Pennsylvania Archaeology, Brookville, PA. PA Prehistoric Data Synthesis, Raccoon Creek Watershed 119
Housley, R.A., C.S. Gamble, M.Street, and P. Pettitt 1997 Radiocarbon Evidence for the Lateglacial Human Decolonization of Northern Europe. Proceedings of the Prehistoric Society 63:25-54. Howard, C.D. 1991 The Clovis Point: Characteristics and Type Description. Plains Anthropologist 35(129): 255-262 Jacobson, R.B., Elston, D.P., and Heaton, J.W. 1988 Stratigraphy and magnetic polarity of the high terrace remnants in the Upper Ohio and Monongahela Rivers in West Virginia, Pennsylvania, and Ohio. Quaternary Research 29: 216-232. Johnson, W.C. 1977 Ceramics. In Meadowcroft Rockshelter: Retrospect 1976, edited by J.M. Adovasio et al. Pennsylvania Archaeologist 47(2-3):59-64. 1982 Ceramics from Meadowcroft Rockshelter: A Re-Evaluation and Interpretation. In Meadowcroft: Collected Papers on the Archaeology of Meadowcroft Rockshelter and the Cross Creek Drainage, pp. 142-162, edited by R.Carlisle and J.Adovasio. Papers Prepared for 1982 Society for American Archaeology Meeting, Minneapolis, MN. 2001 Who were those Gals? Cordage Twist Direction and Ethnicity in the Potomac River Basin, Preliminary Evidence suggesting Population Continuity and Replacement during the Late Woodland Period. Paper presented at the Middle Atlantic Archaeological Conference, Ocean City, MD. 2003 Prehistoric Culture History of the Upper Ohio Valley. Unpublished Manuscript. Johnson, W.C. and W.P. Athens 1988 Observations on Late Prehistoric Period Monongahela Culture Settlement Patterns within the Allegheny Plateau Section of Fayette, Westmoreland, and Adjacent Allegheny, Armstrong, and Indiana Counties, Southwestern Pennsylvania. Paper presented at the 55 th Annual Meeting of the Eastern States Archaeological Conference Federation, Toronto. Johnson, W.C., W.P. Athens, M.T. Fuess, L.G. Jaramillo, K.R. Bastianini, and E. Ramos 1989 Late Prehistoric Period Monongahela Culture Site and Cultural Resource Inventory. Report prepared by the University of Pittsburgh Cultural Resource Management Program. Pennsylvania Historical and Museum Commission, Harrisburg. Jones, T.L., R.T. Fitzgerald, D.J. Kennett, C.H. Miksicek, J.L. Fagan, J. Sharp and J.M. Erlandson 2002 The Cross Creek Site (CA-SLO-1797) and its Implications for New World Colonization. American Antiquity 67(2):213-230. 120 PA Prehistoric Data Synthesis, Raccoon Creek Watershed
Joyce, Arthur A. 1988 Early/Middle Holocene Environments in the Middle Atlantic Region: A Revised Reconstruction. In Holocene Human Ecology in Northeastern North America, edited by George P. Nicholas, pp. 185-214. Plenum Press, New York. Justice, N.D. 1987 Stone Age Spear and Arrow Points. Indiana University Press, Bloomington. Kaktins, U. and Delano, H.L. 1999 Drainage basins. In Shultz, C.H. (ed.) The Geology of Pennsylvania, pp.378-390. Harrisburg, Pennsylvania Geological Survey and Pittsburgh, Pittsburgh Geological Society. Kellogg, D.C., R. Kingsley, R. Varisco 1998 Archaeological Data Recovery at the Mayview and Mayview Bend Sites, Allegheny County, Pennsylvania. Report prepared for PennDOT District 11-0 by John Milner Associates, Inc. Kinsey, W.F., III 1972 Archaeology in the Upper Delaware Valley. Pennsylvania Historical and Museum Commission, Harrisburg. Knepper, Dennis A. and Michael D. Petraglia 1996 Prehistoric Occupation at the Conoquenessing Site (36Bv292), An Upland Setting in the Upper Ohio River Valley. Archaeology of Eastern North America 24:29-58. Kuhn, R.D. 1985 A Morphological Analysis of Bifurcated-Base projectile Points in the Northeastern United States. Man in the Northeast 29:55-69. Lantz, S.W. 1982 A Cultural Resource Survey of L.R. 93 (T.R. 62) Irvine Bridge Located near the Town of Irvine, Warren County, Pennsylvania. Report prepared by Carnegie Museum of Natural History, Pittsburgh. 1984 Distribution of Paleo-Indian Projectile Points and Tools from Western Pennsylvania: Implications for Regional Differences. Archaeology of Eastern North America 12: 210-230. 1989 Age, Distribution and Cultural Affiliation of Raccoon Notched Point Varieties in Western Pennsylvania and Western New York. Bulletin of the Carnegie Museum of Natural History No. 28, Pittsburgh. Larabee, P.A. 1986 Late Quaternary Vegetational and Geomorphic History of the Allegheny Plateau at Big Run Bog, Tucker County, West Virginia. Unpublished Masters of Science Thesis. The University of Tennessee, Knoxville. PA Prehistoric Data Synthesis, Raccoon Creek Watershed 121
Lepper, Bradley T. 1983 Fluted Point Distributional Patterns in the Eastern United States: A Contemporary Phenomenon. Midcontinental Journal of Archaeology 8(2):269-286. Lepper, B.T., R.W. Yerkes, and W. H. Pickard 2001 Prehistoric Flint Procurement Strategies at Flint Ridge, Licking County, Ohio. Midcontinental Journal of Archaeology 26(1):53-78. Leverette, F. 1934 Glacial deposits outside the Wisconsin terminal moraine in Pennsylvania. Pennsylvania geological survey, 4 th series, General geology Report 7, 123 p. Lothrop, Jon In Prep Panhandle Archaic Americans in the Upper Ohio Valley: Archaeological Investigations at the East Steubenville (46Br31) and Highland Hills (46Br60) Sites. WV Route 2 Follansbee-Weirton Road Upgrade Project, Brooke County, West Virginia. Report prepared for Whitney, Bailey, Cox & Magnani, LLP. n.d. Panhandle Archaic in the Upper Ohio River Valley: Data Recovery at the East Steubenville (46BR31) and Highland Hills (46BR60) Sites. Report in Preparation for West Virginia Division of Highways by GAI Consultants, Monroeville, PA. 1989 The Organization of Paleoindian Lithic Technology at the Potts Site. In Eastern Paleoindian Lithic Resource Use, edited by C.Ellis and J.Lothrop, pp. 99-137. Westview Press, Boulder. 2001a 2001b Management Summary, Phase III Data Recovery at Prehistoric Sites 46Br31 and 46Br60, U.S. Route 2 Follansbee-Weirton Road Upgrade Project, Follansbee, Brooke County, West Virginia. Prepared by GAI Consultants, Inc. for Whitney, Bailey, Cox & Magnani, Bridgeport, West Virginia. Chapter 2. Study Unit 1: Upper Ohio Valley, Woodland Prehistoric Context for West Virginia. Unpublished Manuscript prepared for the West Virginia State Historic Preservation Office, Charleston, W.V. Lothrop, J.C. and B. Munford 2001 Ritual or Special Purpose Use of Site 46Ni267. In Phase III Data Recovery Investigations of Site 46Ni267, Nicholas County, West Virginia, pp. 167-176, prepared by D.H. MacDonald, B.A. Munford, and J.C. Lothrop. Submitted to the West Virginia Division of Highways by GAI Consultants, Inc., Charleston, W.V. Lounsburry, Kelly M. 1997 The Smith Site: The Chautauqua-McFate Culture in the Upper Allegheny River Valley in Southwestern New York. Pennsylvania Archaeologist 67(1):21-34. Luedtke, Barbara E. 1992 An Archaeologist s Guide to Chert and Flint. Archaeological Research Tools 7. Institute of Archaeology, University of California, Los Angeles. 122 PA Prehistoric Data Synthesis, Raccoon Creek Watershed
MacDonald, D.H. 2000a 2000b Phase I Archaeological Survey, Royal Tartan Golf Course, Buffalo and Canton Townships, Washington County, Pennsylvania, ER 00-0010-125D. Report prepared for W.B. Stewart by GAI Consultants, Inc., Monroeville, Pa. Phase III Data Recovery Investigations at the Coverts Crossing Site (36Lr75), Lawrence County, Pennsylvania. Report Prepared for PennDOT District 11-0 and Taylor Engineers, Inc., New Castle, Pa. by GAI Consultants, Inc., Monroeville, Pa. 2001 Phase III Data Recovery Investigations at the Coverts Bridge Site (36Lr22812), Lawrence County, Pennsylvania. Report Prepared for PennDOT District 11-0 and Taylor Engineers, Inc., New Castle, Pa. by GAI Consultants, Inc., Monroeville, Pa. 2003a Pennsylvania Archaeological Data Synthesis: The Upper Juniata River Subbasin 11 (Watersheds A-D), Walter Industrial Park: Mitigation of Adverse Effects, USDCEDA, Greenfield Township, Blair County, PA. ER 00-2888-013. Prepared for Keller Engineers, Inc., by GAI Consultants, Inc., Monroeville, Pa. 2003b Phase I/II Archaeological Investigations, Bridge Replacement Project T-319, Beaver County Bridge No. 36 (Links Bridge), Independence Township, Beaver County, Pennsylvania. Prepared for PennDOT District 11-0 by GAI Consultants, Inc., Monroeville, Pa. 2004 Paleoindians as Estate Settlers: Archaeological, Ethnographic, and Evolutionary Insights into the Peopling of the New World. In The Settlement of the American Continents, in press, edited by G.Clarke and M.Barton. University of Arizona Press, Tucson. MacDonald, D.H. and D.L. Cremeens 2002 Archaeology and Geomorphology of the Coverts Crossing (36Lr75) and Coverts Bridge (36Lr228) Sites, Lawrence County, Pennsylvania. Pennsylvania Archaeologist 72(1):17-50. Marine, J.T. 1997 Terrace deposits associated with ancient Lake Monongahela in the lower Allegheny drainage. Unpublished Master s thesis. University of Pittsburgh. Maslowski, R.F. 1973 An Analysis of Cordmarked Watsonware. Pennsylvania Archaeologist 43(2):1-12. 1983 The significance of Cordage Attributes in the Analysis of Woodland Pottery. Pennsylvania Archaeologist: 51-60. Maxwell, Jean A. and Margaret Bryan Davis 1972 Pollen Evidence of Pleistocene and Holocene Vegetation on the Allegheny Plateau, Maryland. Quaternary Research 2:506-530. Mayer-Oakes, W.J. 1953 An Archaeological Survey of the Proposed Shenango River Reservoir Area in Ohio and Pennsylvania. Anthropological Series, No.1. Carnegie Museum, Pittsburgh. PA Prehistoric Data Synthesis, Raccoon Creek Watershed 123
1955 Prehistory of the Upper Ohio Valley: An Introductory Archaeological Study. Anthropological Series, No. 2. Annals of the Carnegie Museum, Pittsburgh. McAvoy, J.M. and L.D. McAvoy 1997 Archaeological Investigations at Site 44Sx202, Cactus Hill, Sussex County, Virginia. Research Report Series No. 8. Commonwealth of Virginia Department of Historic Resources, Richmond. McConaughy, M.A. 1990 Early Woodland Mortuary Practices in Western Pennsylvania. West Virginia Archaeologist 42(2):1-10. 2000 The Watson Farm Site (46Hk34): a Possible Late Middle Woodland Village in the Panhandle of West Virginia. Paper Presented at the Society for American Archaeology Meeting, Philadelphia. McMichael, E. V. 1956 An Analysis of McKees Rocks Mound, Allegheny County, Pennsylvania. Pennsylvania Archaeologist 36:128-151. 1971 Adena-East, an Appraisal of the More Easterly Extensions of the Spread of the Adena Phenomenon. In Adena: The Seeking of an Identity, edited by B.K. Swartz, Jr., pp. 88-97. Ball State University, Muncie, Indiana. Mead, J.I. 1980 Is it Really that Old? A Comment about the Meadowcroft Rockshelter Overview. American Antiquity 45(3):579-582. Meltzer, D.J. 1988 Late Pleistocene Human Adaptation in Eastern North America. Journal of World Prehistory 2: 1-52. 1989 Why Don t we Know When the First People Came to North America? American Antiquity 54:471-490. 1995 Clocking the First Americans. Annual Review of Anthropology 24:21-45. 2002 Thomas Wolfe and Eastern North American Paleoindians. In Ice Age Peoples of Pennsylvania, pp. 159-166, edited by K. Carr and J. Adovasio. Pennsylvania Historical and Museum Commission, Harrisburg. Mohney, K.W. 2002 Steubenville Projectile Points: Social and Technological Functions in the Panhandle Archaic of the Upper Ohio Valley. Unpublished doctoral dissertation, University of Pittsburgh. 124 PA Prehistoric Data Synthesis, Raccoon Creek Watershed
Mohney, K.W. and J.C. Lothrop 2003 Raw Material Availability and Use During the Late Archaic in the Upper Ohio Valley: A View from the East Steubenville Site. Paper presented at the Pennsylvania Archaeological Council Symposium, Recent Advances in Lithic Sourcing Studies, State College, Pennsylvania. Mounier, R.A., J. Cresson, and J.W. Martin 1993 New Evidence of Paleoindian Biface Fluting from the Outer Coastal Plain of New Jersey at 28-OC-100. Archaeology of Eastern North America 21:1-23. Nass, J.P., Jr. and J.P. Hart 2000 Subsistence-Settlement Change during the Late Prehistoric Period in the Upper Ohio River Valley: New Models and Old Constructs. In Cultures before Contact: the Late Prehistory of Ohio and Surrounding Regions, edited by R.A. Genheimer, pp. 124-155. The Ohio Archaeological Council, Columbus. Olafson, Sigfus 1964 West Virginia Flints Used in Ohio. Ohio Archaeologist 14(2):37-39. Otto, M.P. 1979 Hopewell Antecedents in the Adena Heartland. In Hopewell Archaeology: The Chillicothe Conference, edited by D.S. Brose and N omi Greber, pp. 9-14. The Kent State University Press, Kent, Ohio. Pennsylvania Bureau of Historic Preservation 1991 Cultural Resource Management in Pennsylvania: Guidelines for Archaeological Investigations. Bureau of Historic Preservation and Pennsylvania Historical and Museum Commission, Harrisburg. Pennsylvania Department of Environmental Protection 2002 Pennsylvania DEP Watershed Notebook. In http: www.dep.state.pa.us/dep/ deputate/watermgt/. Pennsylvania Department of Environmental Protection Home Internet Site. Raber, P. A. 1985 A Comprehensive State Plan for the Conservation of Archaeological Resources. Pennsylvania Historical and Museum Commission, Harrisburg, Pennsylania. 1995 Prehistoric Settlement and Resource Use in the Aughwick Creek Valley and Adjoining Areas of Central Pennsylvania. Pennsylvania Archaeologist 65(1):1-18. 1999 Late Woodland Settlement in the Uplands of Central Pennsylvania: the View from the Mykut Rockshelter, 36Hu143. Paper presented at the 65 th SPA Meeting, Philadelphia, April 6, 2000. Raber, P.A, P.E. Miller, and S.M. Neusius, eds. 1998 The Archaic Period in Pennsylvania: Hunter-Gatherers of the Early and Middle Holocene Period. Pennsylvania Historical and Museum Commission, Harrisburg, PA. PA Prehistoric Data Synthesis, Raccoon Creek Watershed 125
Reger, D. B. 1921 West Virginia Geologic Survey. Charleston, W.V. Reppert, R. S. 1978 Kanawha chert: Its Occurrence and Extent in West Virginia. Open File Report 96. West Virginia Geological and Economic Survey, Charleston. Ritchie, W.A. 1965 The Archaeology of New York State. The Natural History Press, Garden City. 1969 The Archaeology of New York State (revised edition). The Natural History Press, Garden City. 1980 The Archaeology of New York State (revised edition). Harbor Hill Books, Harrison, New York. Rue, David J. 1990 Cultural Resources Report for Proposed Construction of Columbia Gas Transmission Corporation s 672-Foot Pipeline Replacement in Lawrence County, Pennsylvania. Project No. C7019, 3D/Environmental Services, Inc., Cincinnati, Ohio. Schiffer, M.B. 1983 Toward the Identification of Formation Processes. American Antiquity 48(4): 675-706. Schuldenrein, J., F. Vento, and S. Selby 2003 Draft Report: Interim Geomorphology Report for Archaeological Site 36AL480, Area 3 South, Leetsdale Industrial Park, Leetsdale, Allegheny County, Pennsylvania. Submitted to the USACOE Pittsburgh District by Geoarcheology Research Associates, Inc., Riverdale, New York. Schweikart, J. F. 1998 An Upland Settlement in the Adena Heartland: Preliminary Evidence and Interpretations from Two Early Woodland Non-Mortuary Habitations in Perry County, Ohio. Unpublished Manuscript on File at ASC Group, Inc. Columbus, Ohio. Sciulli, P.W., L.R. Lease, B.P. Brown, and N.E. Tatarek 2003 A Preliminary Investigation of Population Dynamics of the Upper Ohio Valley: AD 1200-1650. Archaeology of Eastern North America 31:1-14. Seeman, M. F. 1977 The Hopewell Interaction Sphere: The Evidence for Interregional Trade and Structural Complexity. Ph.D. Dissertation, Indiana University. University Microfilms International, Ann Arbor. 1986 Adena Houses and Their Implications for Early Woodland Settlement Models in the Ohio Valley. In Early Woodland Archaeology, edited by Kenneth B. Farnsworth and Thomas E. Emerson, pp. 564-580. Center for American Archeology Press, Kampsville, Illinois. 126 PA Prehistoric Data Synthesis, Raccoon Creek Watershed
Shelford, V.E. 1963 The Ecology of North America. University of Illinois Press, Urbana. Shott, M.J. 1990 Childers and Woods: Two Late Woodland Sites in the Upper Ohio Valley, Mason County, West Virginia. Program for Cultural Resource Assessment Archaeological Report 20. University of Kentucky, Lexington. 1993 The Leavitt Site: A Parkhill Phase Paleoindian Occupation in Central Michigan. Memoirs of the Museum of Anthropology No. 25. University of Michigan, Ann Arbor. Skinner, S.M. 1999 Phase I Literature Review and Archaeological Survey for a Proposed Sewer Pipeline in Robinson, Cecil, and Mount Pleasant Townships, Midway and McDonald Boroughs, Washington County, Pennsylvania. Report submitted to Gannett Fleming, Inc. By ASC Group, Inc., Columbus, Ohio. Skirboll, Esther R. and Roger W. Hanson 1996 Preliminary Findings at the Wolf Creek Site (36BT82), Butler County, Pennsylvania. Pennsylvania Archaeologist 66(1):54-67. Smith, B.D. 1987 The Independent Domestication of Indigenous Seed-Bearing Plants in Eastern North America. In Emergent Horticultural Economies of the Eastern Woodlands, edited by W.F. Keegan, pp. 3-48. Center for Archaeological Investigations, Southern Illinois University at Carbondale, Occasional Paper No. 7. Smith, R.V. 1982 Soil Survey of Beaver and Lawrence Counties, Pennsylvania. USDA- Soil Conservation Service. Government Printing Office, Washington, D.C. Snow, D.R. 1980 The Archaeology of New England. Academic Press, New York. Steegmann, A. Theodore 1983 Boreal Forest Adaptations: The Northern Algonkians. Plenum Press, New York. Stewart, M. and J.Cavallo 1991 Delaware Valley Middle Archaic. Journal of Middle Atlantic Archaeology 7:19-42. Stewart, M. and J. Kratzer 1989 Prehistoric Site Locations on the Unglaciated Appalachian Plateau. Pennsylvania Archaeologist 59(1):19-36. Stingelin, R.W. 1965 Late-Glacial and Post-Glacial Vegetational History in the North Central Appalachian Region. Unpublished Doctoral Dissertation, Department of Geology and Geophysics. Pennsylvania State University, State College. PA Prehistoric Data Synthesis, Raccoon Creek Watershed 127
Straus, L.G. 2000 Solutrean Settlement of North America? A Review of Reality. American Antiquity 65(2):219-226. Stuckenrath, R., J.M. Adovasio, J. Donahue, and R. C. Carlisle 1982 The Stratigraphy, Cultural Features and Chronology at Meadowcroft Rockshelter, Washington County, Southwestern Pennsylvania. In Meadowcroft: Collected Papers on the Archaeology of Meadowcroft Rockshelter and the Cross Creek Drainage, pp. 69-90, edited by R.Carlisle and J.Adovasio. Papers Prepared for 1982 Society for American Archaeology Meeting, Minneapolis, MN. Tankersley, Kenneth B. 1989 A Close Look at the Big Picture: Early Paleoindian Lithic Resource Procurement in the Midwestern United States. In Eastern Paleoindian Lithic Resource Use, edited by B. Lepper and J. Lothrop, pp. 259-292. Westview Press, Boulder. Thornbury, W.D. 1965 Regional Geomorphology of the United States. John Wiley and Sons, London. Trewartha, G.T. 1967 An Introduction to Climate. McGraw-Hill, New York. Turnbaugh, W.H. 1977 Man, Land, and Time: The Cultural Prehistory and Demographic Patterns of North Central Pennsylvania. Lycoming County Historical Society, Williamsport, Pennsylvania. Turner, C.G., II 2003 Three Ounces of Sea Shells and One Fish Bone Do Not a Coastal Migration Make. American Antiquity 68(2):391-396. Vento, F.J. 2001 Geomorphology Study. In Phase I Summary/ Phase II Work Plan, Penndot district 11-0 Open End, Link s Bridge #36, T-319, Independence Township, Beaver County, Pennsylvania. Prepared by Christine Davis Consultants, Inc. Vento, F.J., and J. Donahue 1982 Lithic Raw Material Utilization at Meadowcroft Rockshelter and in the Cross Creek Drainage. In Meadowcroft: Collected Papers on the Archaeology of Meadowcroft Rockshelter and the Cross Creek Drainage. Symposium at 47th Annual Meeting of the Society for American Archaeology, Minneapolis, Minnesota. Vento, F.J., J. Schuldenrein, M.P. Purtill 2002 Final Report, Geomorphology of Archaeological Site 36AL480 at Leetsdale Industrial Park, Leetsdale, Allegheny County, Pennsylvania, ER 87-0469-042. Prepared for USACOE Pittsburgh District. 128 PA Prehistoric Data Synthesis, Raccoon Creek Watershed
Vickery, K. D. 1996 The Flint Sources. In Recent Excavations at the Edwin Harness Mound, Liberty Works, Ross County, Ohio, edited by N omi Greber, pp. 73-85. MCJA Special Paper No. 5. Kent State University Press. Wagner, W.R., Heyman, L., and Gray, R.E. 1970 Geology of the Pittsburgh area. Pennsylvania Geological Survey, 4 th Series, General Geology Report 59, 145 p. Wall, Robert D., R. Michael Stewart, John Cavallo, Douglas McLearen, Robert Foss, Phillip Perazio, and John Dumont 1996 Prehistoric Archaeological Synthesis. Trenton Complex Archaeology: Report 15. The Cultural Resource Group, Louis Berger and Associates, Inc., East Orange, Jew Jersey. Prepared for the Federal Highway Administration and the New Jersey Department of Transportation, Bureau of Environmental Analysis, Trenton. Wallace, P.A.W. 1971 Indian Paths of Pennsylvania. The Pennsylvania Historical and Museum Commission, Harrisburg. Walker, P.C. and Hartman, R.T. 1960 The Forest Sequence of the Hartstown Bog Area in Western Pennsylvania. Ecology 41 (3): 461-474. Watts, W.A. 1979 Late Quaternary Vegetation of Central Appalachia and the New Jersey Coastal Plain. Ecological Monographs 49(4):427-469. Webb, W.S. 1941a 1941b Mt. Horeb Earthworks, Site 1, and the Drake Mound, Site 11, Fayette County, Kentucky. Reports in Anthropology and Archaeology 5(2). University of Kentucky, Lexington. The Morgan Stone Mound, Site 15, Bath County, Kentucky. Reports in Anthropology and Archaeology 5(2). University of Kentucky, Lexington. Whitehead, Donald R. 1973 Late-Wisconsin Vegetational Changes in Unglaciated Eastern North America. Quaternary Research 3:621-631. PA Prehistoric Data Synthesis, Raccoon Creek Watershed 129
130 PA Prehistoric Data Synthesis, Raccoon Creek Watershed