Determining the Genesis and Cultural Significance of Deep Soil Features at Southeastern Connecticut’s Preston Plains Site more

Determining the Genesis and Cultural Significance of Deep Soil Features at Southeastern Connecticut’s Preston Plains Site Timothy Howlett Ives, Ph.D. University of Connecticut, 2010 Archaeologists excavating Archaic and Woodland Period sites on sandy, unconsolidated soils in the Northeastern U.S. have identified deep soil features (hereafter DSFs) that are challenging to interpret. Though hundreds of these basin-shaped features have been recorded, archaeologists do not agree as to whether or not they are anthropogenic. Competing hypotheses have suggested that DSFs constitute the remnants of semisubterranean pit houses, or, alternately, soil disturbances generated by naturally occurring tree throws. This dissertation presents a case study of a DSF complex at southeastern Connecticut’s Preston Plains Site. Its analytical design combines scholarship, empirically-based data assessments, and hypothesis testing to holistically inform an interpretation of the genesis and cultural significance of DSFs here. Its results discount the pit house hypothesis while supporting the tree throw hypothesis according to multiple lines of evidence. A simple and flexible model is proposed to explain how tree throws are modified through time to express the variety of forms and stratigraphies observed in DSFs. Furthermore, it is determined that the pit-and-mound microtopographies afforded by ancient tree throws at Preston Plains were targeted by small groups of Late Archaic Period (ca. 5000-3000 BP) foragers as elements of short-term residential sites. While archaeologists have already determined that Mesolithic and early Neolithic Europeans utilized such topographies as site elements, this study provides the most detailed set of supporting evidence of such behavior to date. Timothy Howlett Ives – University of Connecticut, 2010 Determining the Genesis and Cultural Significance of Deep Soil Features at Southeastern Connecticut’s Preston Plains Site Timothy Howlett Ives B.A., University of Connecticut, 1996 M.A., College of William and Mary, 2001 A Dissertation Submitted in Partial Fulfillment of the Requirements for the Degree of Doctor of Philosophy at the University of Connecticut 2010 Copyright by Timothy Howlett Ives 2010 APPROVAL PAGE Doctor of Philosophy Dissertation Determining the Genesis and Cultural Significance of Deep Soil Features at Southeastern Connecticut’s Preston Plains Site Presented by Timothy Howlett Ives, B.A., M.A. Major Advisor ______________________________ Kevin A. McBride Associate Advisor ______________________________ Natalie D. Munro Associate Advisor ______________________________ Daniel S. Adler University of Connecticut 2010 ii Acknowledgements I thank the Mashantucket Pequot Tribal Nation for their generous funding, interest, and encouragement regarding my investigation of the Preston Plains Site. Colleagues who provided moral and technical support include Kathleen Bouchee, Brian Jones, Dan Forrest, Julie Hartman-Brodeur, Sarah Holmes, Jason Mancini, and Zack Decker. Roberta Charpentier, the Museum’s Laboratory Director, provided excellent advice regarding methodology and data processing. I am grateful to all who worked at the site, including tribal members, contractors, university field school students, elementary school students, interns, and volunteers who are, regrettably, far too numerous to list. Mandy Ranslow generated the Surfer8-based artifact distributions used in this dissertation. Randy Nokes provided faunal identifications while Tonya Largy and David DeMello provided an analysis and report of botanical materials from Locus I. Robert Thorson shared several key observations that influenced my research approach and site interpretation. As academic advisors, Nicholas Bellantoni and Daniel Adler provided key critiques of my research design that galvanized this work product. Three student colleagues - Phillip Graham, Randy Nokes, and Sarah Sportman - worked along side me at Preston Plains in supervisory capacities. Natalie Munro provided guidance in my academic progress in a way that was clear, positive, and energizing. Kevin McBride, my primary advisor, afforded me the luxury of discovering my own path in my doctoral research, celebrating my dead ends as much as my victories. Consequently, my student experience has been empowering and enjoyable. Lastly, I thank my wife, Sonia, for her unwavering faith in my abilities to finish this work. She encouraged me to return to graduate school to complete this Ph.D. and helped me maintain focus throughout. iii Table of Contents Approval Page ……………………………………………………………………….…...ii Acknowledgements ………………………………………………………………….…..iii List of Figures and Tables …………………………………………………………..…..vii Chapter 1. Introduction ………………………………………………………………..….1 1.1. Problem: The Enigma of Deep Soil Features ………………………………........1 1.2. Purpose: Explaining DSF Formation and Cultural Significance ………………...5 1.3. Preston Plains Archaeology and the Author’s Involvement ………………….….7 1.4. Theoretical Outlook and Research Approach ……………………………...…...11 1.5. Analytical and Interpretive Design ……………………………………………..14 Chapter 2. Environmental and Cultural Context for New England ………………...…..16 2.1. The Late Pleistocene (ca. 18,000-10,000 BP) ………………………………..…17 2.2. The Early Holocene (ca. 10,000-8000 BP) ………………………………...…...22 2.3. The Middle Holocene (ca. 8000-5000 BP) ……………………………..………30 2.4. The Late Holocene (ca. 5000 BP-Present) …………………………………...…35 Chapter 3. Perceptions and Interpretations of DSFs in the Archaeological Record ……44 3.1. DSFs in Southern New England ………………………………………...…...…44 3.2. DSFs in the Middle Atlantic ……………………………………………..……..47 Chapter 4. Tree Throws: Formation, Character and Archaeological Relevance …...…..52 4.1. Formation and Character of Tree Throws ………………………………………52 4.2. Archaeological Relevance of Tree Throws ……..……………………………....59 Chapter 5. Methods of Data Collection and Processing ….…………………..…………65 iv 5.1. Field Methods ………………………………………………………………..…65 5.1.1. Test Pits ….……………………………………………………………..…65 5.1.2. Excavation Units …………………………………………………….……67 5.1.3. Machine Trenches ...………………………………………………………68 5.1.4. Test Units ...…………………………………………………………….…69 5.2. Laboratory Methods and Data Processing ...……………………………………70 5.2.1. Cleaning and Cataloging Artifacts ……………………………………..…70 5.2.2. Soil Flotation ….………………………………………………………..…71 5.2.3. Soil Water Capacity ………………………………………………………72 5.2.4. Radiocarbon Dating ………………………………………………………73 5.2.5. Long-Term Storage and Access ………………………………………..…73 5.2.6. Graphics Production ………………………………………………………74 Chapter 6. Setting and Geomorphology of Preston Plains …………………………..…76 6.1. Modern Site Setting ...…………………………………………………………..76 6.2. Geomorphology ...………………………………………………………………78 Chapter 7. Overview of Preston Plains Archaeology …..………………………………92 7.1. Younger Dryas Period Occupation ….……………………………………….…96 7.2. Early Holocene Occupation (ca. 10,000-8000 BP) …..……………………...…97 7.3. Middle Holocene Occupation ….………………………………………………99 7.4. Late Holocene Occupation (post-5000 BP) …..………………………………100 7.5. DSFs at the Preston Plains Site …..………………………………………...…104 Chapter 8. Locus I: A Case Study of Late Archaic Period Occupation and a Cluster of DSFs ………………………………………………….…………108 v 8.1. Artifacts and their General Distribution ...……………………………………111 8.2. DSFs …………………………………………………………………………..119 8.3. Discrete Features ………………………………………………………………127 Chapter 9. Testing a Hypothesis: Artifact Concentrations are Not Present on the Bottoms of DSFs at Preston Plains ...............................................................147 9.1. Testing Results for DSFs Identified in Machine Trenches in 2008 …..………148 9.2. Testing Results for DSFs Identified in Locus I Excavation Block …...………169 9.3. Summary of Observations …….………………………………………………174 Chapter 10. Local Tree Throw Examination ……………………………………..……177 10.1. Tree Throw 1: Avery Pond Site …..……………….……………………...…178 10.2. Tree Throws 2 and 3: Sandy Hill Site …..……………………….………..…181 10.3. Summary of Observations …………………………………….…………...…185 Chapter 11. Genesis and Cultural Significance of DSFs at Preston Plains ...…………187 11.1. DSFs at Preston Plains Began as Tree Throws………………………………187 11.2. DSF Variability Reflects Individual Formation Histories…………………...195 11.3. Cultural Significance of DSFs at Locus I: Hummocked Topography as an Element of Site Selection During theLate Archaic Period ......……...…200 Chapter 12. Significance of this Study to the Cultural Resource Management Industry and the Academy …….…………………………………………..…206 Sources Cited…………………………………………………………………...………214 vi List of Figures and Tables Figure 1-1. Figure 5-1. Location of Preston Plains Site in the State of Connecticut..……………..3 Map depicting all archaeological testing within the revised Preston Plains Energy Center Project Area …………………………………...…66 Figure 5-2. Plan map of Machine Trench 17, located outside of the revised Preston Plains Energy Center Project Area …………………………..…67 Table 5-1. Elevations of Machine Trench Datums Relative to Mean Sea Level …… ………………………………………………………………69 Table 5-2. Figure 6-1. All Radiocarbon Dates from the Preston Plains Site ………………....…74 View south of Avery Pond (foreground), Preston Plains openlands (farm tracts) and the Great Cedar Swamp (undeveloped forest, center-background), all elements of the same valley ..............................77 Figure 6-2. View southwest of house and garage at 455 Norwich Westerly Road in 2006 ……………………………………………………….....…78 Figure 6-3. Remnant glaciofluvial channel in south wall of MT-13. The darkest stratum (trowel-incised) is a buried A-Horizon ………………..80 Figure 6-4. Remnant glaciofluvial channel in south wall of MT-14. The darkest stratum (trowel-incised) is a buried A-Horizon ……………...…81 Figure 6-5. West wall of MT-9 (view south) showing continuous series of DSFs (darker basin-like forms) along edge of remnant glaciofluvial channel .................................................................................82 Figure 6-6. Tension fracture in sandy outwash (lower right) beneath a layer of gravely outwash in south wall of MT-13 (10-cm increment vii scale rod) ……………………………………………………………...…83 Figure 6-7. Yedoma-filled ice wedge cast in west wall of MT-15 (white tube markers at 2-m intervals) ……………….…………………………….…84 Figure 6-8. Yedoma-filled ice wedge cast in east wall of MT-16 (white tube markers at 2-m intervals; 10-cm increment scale rod). Note darker-colored krotovinas in the yedoma ......….……………………..…84 Figure 6-9. Figure 6-10. Figure 6-11. Painted turtle laying eggs near Locus 4 of Preston Plains, 2008 ………..86 Painted turtle hatchling found in MT-17, 2008 …………………………86 Woodchuck burrow complex in west wall of MT-11. Note the presence of a thin buried A-Horizon, visible as a thin, dark line immediately beneath the plowed A-Horizon ..………………………..…87 Figure 6-12. Plan view of probable small rodent burrow complex in glaciofluvial sands beneath DSF-5, Locus I (trowel points north) ...……………….…88 Figure 6-13. North wall of MT-17 (view west). Note root penetration through topsoil and feature soil (Fe.323), and absence of roots in C-Horizon sand (lightest soil at bottom) …………….………………………………89 Table 6-1. Soil water capacity measurements for representative Preston Plains soil types ………………………………………………….…………..…91 Table 7-1. Temporally diagnostic lithic points recovered from the Preston Plains Site …………………………………………………………….…93 Figure 7-1. Major Late Archaic Period site loci defined by Jones at Preston Plains (map produced by Jones 2003). The original Preston Plains Energy Center project area boundary is shown as an irregular red viii line.Green contours represent topography, while grey contours represent lithic artifact concentrations ..................………………………94 Figure 7-2. Distribution of lithic artifacts (≥1 cm) from 50-cm square test pits from the revised Preston Plains Energy Center Project Area. Updated as of 2009. (Surfer overlay by Mandy Ranslow) ……………..95 Figure 7-3. Barnes point (chert) recovered from Preston Plains (1-cm scale bar increments) ………………………………………………………………96 Figure 7-4. Early Holocene points recovered from Preston Plains Site. San Patrice on left (chert); Kanawha-like bifurcate on rightn(quartz) ………98 Figure 7-5. Late Archaic triangular points from the Preston Plains Site illustrating a gradient between Brewerton (4 on the left) and Squibnocket (4 on the right) types. All are made of Plainfield Formation quartzite …………………………………………………….101 Figure 7-6. Plan view of Fe.215, the only confirmed Woodland Period feature at the Preston Plains Site. Trowel points north ...............………………104 Table 7-2. Figure 8-1. DSFs identified in machine trenches at the Preston Plains Site ..… Plan of Locus I excavation block depicting all of the DSFs and most of the smaller, discrete features ……….........……………………109 Table 8-1. Table 8-2. Figure 8-2. Lithic tools and debitage, Locus 1 …………………………………..…112 Projectile points, Locus 1. Late Archaic types are italicized …………..113 Representative Late Archaic Period projectile points from Locus I, Preston Plains Site. Laurentian Tradition points are in the top row, while Narrow-Stemmed Tradition points are in the bottom row. ix …106 a. Otter Creek (argillite); b-c. Vosburg (quartzite); d-e. Brewerton (quartzite); f. Bare Island (quartzite); g. Narrow-Stemmed (argillite); h. Narrow-Stemmed (quartzite); i. Narrow-Stemmed (quartz); j-k. Squibnocket Triangle (quartzite); Squibnocket Triangle (quartz) ..........114 Figure 8-3. Lithic artifacts ≥1 cm recovered from beneath topsoil, Locus 1 excavation block (Surfer overlay by Mandy Ranslow). The counts depicted here may vary from those in profile drawings produced by the author in Chapter 9, which reflects minor differences in how Ranslow and the author sorted data …………….…116 Figure 8-4. Diagnostic points recovered from topsoil, Locus I excavation block (Surfer overlay by Mandy Ranslow)……………………………………117 Figure 8-5. Diagnostic points recovered from beneath topsoil, Locus I excavation block (Surfer overlay by Mandy Ranslow) .....………….…118 Figure 8-6. Calcined bone recovered from beneath topsoil, by weight (grams), Locus 1 excavation block ............................…………………...………119 Figure 8-7. Figure 8-8. Figure 8-9. Figure 8-10. Figure 8-11. Figure 8-12. Figure 8-13. Figure 8-14. Profile section drawing of DSF-1 and DSF-2 ……………….…………121 Profile section drawing of DSF-4 ………………………………….…..122 Profile section drawing of DSF-5 ……………………………...………123 Profile section drawing of DSF-6 ………………………………...……124 Profile section drawing of DSF-7 and DSF-8 ……………………….…125 Profile section of DSF-7 and DSF-9 ………………………………...…125 Profile section of DSF-9 …………………………………………….…126 Examples of primary hearth deposits at Locus 1, viewed in profile. x Fe.8 (top left, view west) is a pit hearth penetrating DSF-1 matrix. Fe.21 (top right, view south) is a pit hearth penetrating C-Horizon soils immediately adjacent to DSF-1. Fe.18 (bottom, view south) is a hearth deposit within, and conforming to, the shallow end of DSF-4………………………………………………………………...…129 Table 8-3. Table 8-4. Table 8-5. Table 8-6. Table 8-7. Table 8-8. Table 8-9. Table 8-10. Table 8-11. Table 8-12. Table 8-13. Table 8-14. Table 8-15. Table 8-16. Table 8-17. Table 8-18. Table 8-19. Table 8-20. Feature 2 summary ...…………………………………………………...130 Feature 3 summary ...………………………………………………...…131 Feature 4 summary ...………………………………………………...…131 Feature 5 summary ..……………………………………………………132 Feature 6 summary ..……………………………………………………132 Feature 7 summary ..……………………………………………………133 Feature 8 summary ..……………………………………………………134 Feature 9 summary ..……………………………………………………134 Feature 10 summary ..………………………………………………..…135 Feature 11 summary .………………………………………………...…136 Feature 12 summary .………………………………………………...…136 Feature 13 summary .………………………………………………...…137 Feature 14 summary .………………………………………………...…138 Feature 15 summary .………………………………………………...…138 Feature 16 summary .………………………………………………...…139 Feature 17 summary………………………………………………….…139 Feature 18 summary .………………………………………………...…140 Feature 19 summary .………………………………………………...…141 xi Table 8-21. Table 8-22. Table 8-23. Table 8-24. Table 8-25. Table 8-26. Table 9-1. Feature 20 summary .………………………………………………...…142 Feature 21 summary .……………………………………………...……143 Feature 22 summary .…………………………………………………...143 Feature 24 summary ……………………………………………………145 Feature 25 summary ……………………………………………………145 Feature 26 summary ……………………………………………………146 Summary of features identified in machine trenches at Preston Plains in 2008 (Class abbreviations. DSF=Deep Soil Feature; RC=Remnant Channel; U=unidentified) ……………….………...……149 Figure 9-1. Table 9-2. Table 9-3. Table 9-4. Table 9-5. Figure 9-2. Table 9-6. Table 9-7. Table 9-8. Table 9-9. Figure 9-3. Table 9-10. Table 9-11. Table 9-12. Profile of MT-13 showing location of TUs ……………………………151 TU-9 soil flotation summary …………………………………………...152 TU-10 soil flotation summary .…………………………………………152 TU-11 soil flotation summary …….……………………………………153 TU-12 soil flotation summary .…………………………………………154 Profile of MT-14 showing location of TUs .………………………...…154 TU-4 soil flotation summary ………………………………………...…155 TU-5 soil flotation summary ..……………………………………….…155 TU-6. soil flotation summary ..…………………………………………156 TU-7 soil flotation summary ………………………………………...…157 Profile of MT-15 showing location of TUs .………………………...…157 TU-1. soil flotation summary …………………………………………..158 TU-2 soil flotation summary ..……………………………………….…158 TU-3 soil flotation summary .……………………………………….…159 xii Figure 9-4. Table 9-13. Table 9-14. Table 9-15. Table 9-16. Table 9-17. Table 9-18. Table 9-19. Table 9-20. Table 9-21. Table 9-22. Table 9-23. Table 9-24. Table 9-25. Figure 9-5. Figure 9-6. Figure 9-7. Figure 9-8. Figure 9-9. Figure 9-10. Figure 10-1. Figure 10-2. Figure 10-3. Profile of MT-17 showing location of TUs ………………..………..…160 TU-13 soil flotation summary ..……………………………………...…161 TU-14 soil flotation summary ..……………………………………...…162 TU-15 soil flotation summary ..……………………………………...…162 TU-16 soil flotation summary .…………………………………………163 TU-17 soil flotation summary ...…………………………………..……164 TU-18 soil flotation summary …………………………………….……164 TU-19 soil flotation summary ……………………………………….…165 TU-20 soil flotation summary ..………………………………………...166 TU-21 soil flotation summary ..………………………………………...166 TU-22 soil flotation summary ..……………………………………...…167 TU-23 soil flotation summary ..……………………………………...…167 TU-24 soil flotation summary ..……………………………………...…168 TU-25 soil flotation summary ..……………………………………...…168 Lithic distribution in a cross-section of DSFs 1 and 2 ..……………..…170 Lithic distribution in a cross-section of DSF 4 …………………...……171 Lithic distribution in a cross-section of DSF-5 ……………………...…172 Lithic distribution in a cross-section of DSF-6 ……………………...…173 Lithic distribution in a cross-section of DSF-7 ……………………...…174 Lithic distribution in a cross-section of DSF-9 ……………………...…174 Profile section drawing of Tree Throw 1, Avery Pond Site ….......……180 Tree Throw 1, Avery Pond Site (10-cm increment scale rod) …………180 Profile section drawing of Tree Throw 2, Sandy Hill Site ………….…183 xiii Figure 10-4. Figure 10-5. Figure 10-6. Figure 11-1. Tree Throw 2, Sandy Hill Site (10-cm increment scale rod) ………..…183 Profile section drawing of Tree Throw 3, Sandy Hill Site …………….184 Tree Throw 3, Sandy Hill Site (10-cm increment scale rod) …..………184 Section of DSF-4, Locus 1, Preston Plains Site. Note the very steep and sharply defined boundary between feature matrix and bedded glaciofluvial deposits ………………………………………..…188 Figure 11-2. Longitudinal sections of Fe.304 (top left), Fe.300 (top right), and Fe.302 (bottom), Preston Plains Site. These relatively isolated DSFs retain patterns of inclined stratigraphy consistent with naturally occurring tree throws(10-cm increment scale rod) …..…190 Figure 11-3. Longitudinal section of modern tree throw, Hidden Creek Site. Note its striking similarity to Features 304 and 302 (DSFs) at the Preston Plains Site, as depicted in Figure 11-2 ……………………...…190 Figure 11-4. View west of a section of Fe.306 in MT-13, Preston Plains Site. Note the upthrust of fine sandy substrate that reflects a process of catastrophic soil rotation (10-cm increment scale rod) …………...……192 Figure 11-5. Section of Fe.318 in MT-17, Preston Plains Site. Note how this DSF penetrates fine-grained sands and bottoms out against a gravel substrate ……………………………………………………..…192 Figure 11-6. Hypothesized formation of “D-Shaped pit” morphology by a tree throw that involves deep soil rotation and shallow plate removal ….…194 Figure 11-7. Tree throw at Sandy Hill, with a new tree growing on the loam-rich mound (10-cm increment scale rod) …………………………...………197 xiv Figure 11-8. North wall of MT-17, Preston Plains Site. The clustered occurrence of DSFs (such as Fe.319, 320, 321) may eventually result in the development of extensive, unstratified soil masses (such as Fe.323) …………………………….……………..………...…197 Figure 11-9. Detail of Fe.319, MT-17, Preston Plains Site. Note the virtual absence of stratification ....…………………………………………..…199 Figure 11-10. West view of a section of DSF-2, Locus 1, Preston Plains Site. Note the lack of robust stratification within DSF matrix ……………………199 Table 11-1. Estimated tree throw dates and longevities of pit-and-mound pairs at Locus 1 ………………………………………….. ………………….…202 xv Chapter 1. Introduction 1.1 Problem: The Enigma of Deep Soil Features In the Northeastern U.S., archaeologists excavating Archaic and Woodland Period sites on deep, sandy, and unconsolidated soils have identified deep soil features (hereafter DSFs) that are challenging to interpret. Because the term “deep soil feature” could describe many phenomena beyond the context of this research, it is necessary to define the term as employed here. DSFs are a large (>2 m across) and deep (>50 cm deep), basin-shaped soil structures with matrix that is typically loamier than surrounding subsoil and not finely stratified. They may exhibit simple bowl-shaped forms in addition to “Dshaped pits” (sensu Egghart 2005) that often exhibit a crescentic or “D”-shaped plan. Profiles of the D-shaped pit variety often reveal a deepened section at one end and a shallower “shelf” at the other. DSFs may contain artifacts and anthrosols, or they may be entirely sterile. The author invented the semantically neutral term DSF to denote such phenomena in this work to avoid implied formative or functional interpretations that might obstruct an objective analysis. Though hundreds of DSFs have been excavated in New England and the Middle Atlantic, archaeologists do not even agree as to whether or not they are anthropogenic. This is a significant problem for the cultural resource management industry which uses publicly-derived funds to investigate DSFs under the assumption that they are cultural features, or at least culturally relevant phenomena. Because DSF complexes are expensive to investigate it is important that the archaeological community reach an 1 informed consensus as to what process(es) generated these features and what significance they had to past human populations. A primary goal of this dissertation is to bring us one step closer to that consensus. Numerous hypotheses have been developed to explain the presence of DSFs at archaeological sites. They have been hypothesized to be the remains of temporary living structures (Barnes 1972), as well as more substantial, seasonally occupied pit houses (Custer 1994). Alternately, DSFs have been hypothesized to be the subterranean elements of naturally occurring tree throws (Thomas and Payne 1981), some of which may have been modified or utilized by prehistoric populations (Mueller and Cavallo 1995; Mueller et al. 1997). The “D-shaped pit” variety of DSFs have been proposed to be anthropogenic tree throws generated to modify forest cover (Egghart 2005), with “grubbing pits” for root-burning forming their deep ends and subsequent tree throws forming their shallow ends. The presence of chipping debris in DSF matrix has also led to the speculation that DSFs represent “lithic disposal pits” (Hoffman 1983) on complex residential sites. In light of this polyphony, this dissertation‟s research outlook agrees that the nature of DSFs should be assessed through many lines of inquiry (Mueller et al. 1997) because these features are likely the result of complex formation processes where cultural and natural agents may overlap (Petraglia et al. 2005). Accordingly, this dissertation presents a case study of DSFs at southeastern Connecticut‟s Preston Plains Site (Figure 1-1) that is based on a multifaceted approach of empiricism and hypothesis testing. At Preston Plains both bowl-shaped and D-shaped soil structures have been identified, in addition to extensive conglomerates. During early 2 Figure 1-1. Location of Preston Plains Site in the State of Connecticut. 3 stages of field investigations, most of these features were tentatively interpreted as pit house remnants. Subsequent stages of field investigations were designed to test the pit house hypothesis by determining the presence or absence of artifact-strewn “floor layers” on their bottoms. The absence of floor layers in all (n=18) tested DSFs effectively discounts the pit house hypothesis. Through additional lines of inquiry, this research strongly supports the hypothesis that naturally occurring tree throws initiated DSFs. Background research, a geomorphological assessment, and examinations of contemporary tree throws collectively indicate that DSF morphology, and often stratigraphy, at Preston Plains closely reflects the subterranean signatures of naturally occurring tree throws. These tree throws do not appear be anthropogenic, as hypothesized by Egghart (2005), because they do not generally meet the expectation of containing charred wood concentrations in their deep ends - in fact, the opposite is true in some cases. This dissertation also seeks to explain the morphologic and stratigraphic variations that characterize Preston Plains DSFs. This is achieved by providing an overarching model that assumes each DSF has an individual formation history. This simple and flexible model explains how tree throws may become significantly transformed into the variety of expressions recognized as DSFs. Most importantly, this dissertation confidently determines that the pit-and-mound microtopographies afforded by tree throws at Locus I of Preston Plains were targeted by small groups of Late Archaic Period foragers as central elements of short-term residential 4 sites. Several archaeologists have already suggested that tree throws were used by prehistoric peoples (Bubel 2003; Kooi 1974; Mueller and Cavallo 1995; Pyddoke 1961; Whittaker et al. 2007), and European studies have confirmed the use of tree throw pits as site elements (e.g. Crombé 1993; Evans et al. 1999). However, this study provides the most detailed set of supporting evidence of such behavior to date. Tree throws occurred repeatedly in this locale throughout the Late Archaic Period, as did episodes of human occupancy. Their interplay is disentangled through a detailed spatial examination of artifacts and anthrosols within and surrounding DSFs here. This study demonstrates methods of sampling, analysis, and reporting that are applicable to future research of DSF complexes. 1.2. Purpose: Explaining DSF Formation and Cultural Significance This research addresses three interrelated questions: 1. How are DSFs formed at Preston Plains? In other words, what process(es) caused them to initially form, and how have they been influenced over time? It is expected that many factors are implicated in answering this question. This question is answered in Chapters 11.1 and 11.2. 2. What is the nature of Late Archaic human occupation at Preston Plains? While this site appears to have been intermittently occupied by prehistoric Native Americans during most of the Holocene, the Late Archaic Period represents an interregnum where the site was more heavily used. Archaeology is necessary to explain 5 why this place attracted Late Archaic foragers. This question is partially answered in Chapters 8, and more fully revealed in 11.3. 3. What is the cultural significance of DSFs at Preston Plains? In other words, did humans create, use, or modify elements of them? The answer to this question can only be ascertained after the previous two have been answered. answered in Chapter 11.3. Archaeological investigations of the Preston Plains site from 1999-2008 provided novel data sets useful to investigate the character of DSFs and Late Archaic land use on a glacial outwash landform. The site occupies at least 50 acres of MPTN property, and likely extends south and west on adjacent (un-surveyed) properties. Because additional sites containing DSFs have been recorded in New England and the Middle Atlantic, these regions present an immediately relevant geographic scope of application for the results of this study. Additionally, a broader relevance may emerge regarding any geographic zone with similar environmental and/or cultural histories to that of Preston Plains. The research questions are addressed through a multifaceted approach that involves background research, a geomorphological assessment, an examination of archaeological site data, hypothesis testing and excavations of contemporary tree throws. The outlook used to guide this research is adopted from Mueller (1995), who states that “our interpretations should not be viewed as an attempt to establish the false dichotomy that soil features of this kind must be pit houses or tree throws…researchers should be working with multiple modes of formation in mind.” This multivariate perspective calls for a synthesis of archaeological and geomorphological data, recognizing that the 6 This question is disposition of archaeological materials is never the result of one process only (Wandsnider 1987). 1.3. Preston Plains Archaeology and the Author’s Involvement The Preston Plains site, located on fee lands owned by the Mashantucket Pequot Tribal Nation (MPTN), has been surveyed by the Mashantucket Pequot Museum and Research Center (MPMRC) staff intermittently from 1999-2008 to provide the MPTN Planning Department with management recommendations regarding future development projects. The following provides an overview of the complex excavation history of this site. Preston Plains is a vernacular place name referring to the flat, glacial outwash terrain surrounding the intersection of Ledyard, Preston, and North Stonington‟s town boundaries. Though it may appear as little more than an oasis of farmland in the shadow of the rapidly expanding Mashantucket Pequot entertainment complex, Preston Plains resonates with the echoes of a rich cultural and environmental heritage Brian Jones, former archaeological field supervisor of the MPMRC, identified the Preston Plains archaeological site in 1999 while directing an archaeological sensitivity assessment (Phase I) of the Preston Plains Energy Center project area that was designed to determine the presence or absence of archaeological resources therein (Jones 1999). The project area, as originally proposed, encompassed an approximate 220-x-140 meter area of greater Preston Plains. Approximately 350 test pits, each measuring 50-x-50 cm square, were excavated at 10 meter intervals across a grid system oriented to magnetic north. According to distribution maps of the recovered artifacts, approximately 70 7 additional test pits were excavated at closer intervals as part of an initial evaluation (Phase II) of identified prehistoric artifact concentrations or isolated finds. Excavation of all test pits in 1999 resulted in the recognition of “approximately 25 clusters of prehistoric finds” represented by 580 artifacts consisting predominantly of stone chipping debris. A total of 394 historic period artifacts dating from the late 18th century to the present were recovered as well, though they were not determined to be elements of potentially significant cultural resources. Recovered projectile point types suggested occupation throughout the Archaic period, though most appeared to date specifically to the Late Archaic Period. Feature soils of “strong visibility” were also encountered, most often in association with prehistoric artifact concentrations. The University of Connecticut‟s Summer Archaeological Field School invested further efforts to evaluate Late Archaic Period cultural loci at Preston Plains from 20002004, in coordination with MPMRC staff. These efforts involved several site supervisors. Excavation blocks were used to explore artifact densities and record feature distributions within some of the more prominent artifact concentrations. This resulted in the designation of site loci. Large basin features were encountered in some areas, with the most robust examples at Loci 1, 2, and 4, where they formed overlapping complexes. These features were initially interpreted to be the subterranean remnants of pit houses, with some exhibiting a distinct “shelf” at one end and a deeper “storage pit” at the other. According to the results of these investigations, Preston Plains was thought “to contain evidence of multiple deep pit-house structures with stratified floor horizons” that reflect 8 “long-term reoccupation of this lakeside site” by Late Archaic Period forgers (Jones 2002:22). My involvement with the Preston Plains site began in 2006 when Kevin McBride (Research Director the MPMRC and Assistant Professor of Anthropology at the University of Connecticut) enlisted me as a teaching assistant for his University of Connecticut Summer Archaeology Field School. During one of our planning meetings, he suggested that the students excavate “Late Archaic pit houses” at Preston Plains. I enthusiastically agreed as I imagined myself recording elements of pit house architecture and collecting thousands of artifacts from floor layers. As I prepared by reviewing site documentation, I was particularly inspired by a block plan of Locus I drawn by Jones that depicted three “pit houses" in addition to several smaller carbon-rich features. I decided to have the 2006 field school expand this block using the same data recovery techniques as Jones. After field school ended, the MPM retained the students and me so that we could continue expanding the block until the summer‟s end. The expanded/completed block plan depicts 10 DSFs and over 20 smaller features that appear to have been deposited mostly during the Late Archaic Period (Ives 2006a). However, I saw no definitive evidence of pit houses, and neither did many site visitors. In fact, the supposed pit houses served as archaeological Rorschach tests, generating a wide variety of speculation among colleagues. By the summer‟s end, I felt so personally invested in the mystery of this “foster” site that, with the encouragement of Jones and McBride, I adopted it as the focus of my doctoral research. 9 I planned subsequent fieldwork projects at the Preston Plains Site and adjoining properties for the dual purpose of furnishing the MPTN Planning Department with management recommendations and collecting additional data sets useful towards evaluating DSFs. In the Fall of 2006, I planned the excavation of machine trenches in the revised Preston Plains Energy Center project area to investigate the spatial distribution of DSFs and assess their stratigraphic variability. Zachary Decker, MPM Crew Chief, served as on-site supervisor during this work effort (Decker 2006). Prior to excavating these trenches, the MPMRC crew excavated additional test pits to provide greater resolution of prehistoric artifact distributions. Important data was also collected relating to site geomorphology and formation processes during this work effort. With the help of the 2007 field school, I supervised an archaeological assessment survey of additional portions of the Preston Plains landform (Ives 2007a) on fee lands owned by the MPTN. We confirmed that the boundaries of the Preston Plains archaeological site extend far beyond those of the Preston Plains Energy Center project area. Additionally, we identified three additional sites (see Figure 1), two of which are small lithic scatters of unknown age (114-145, 114-146). At the third site, 114-106, I supervised a partial (budget-constrained) site examination that revealed the presence of two DSFs and substantial evidence of occupation by Late Archaic foragers (Ives 2007b), which is relevant to the interpretation of the adjacent Preston Plains Site. In 2008, I planned and supervised a second round of machine trenching in the Preston Plains Energy Center Project area to address a final set of research questions regarding the formation and cultural significance of DSFs. That was the last episode of field work that I conducted on 10 the Preston Plains landform prior to writing this dissertation. I have presented preliminary interpretations of Preston Plains archaeology to the Central Massachusetts Chapter of the Massachusetts Archaeological Society (2007c) and the Society for American Archaeology (2007d, 2009). To achieve the goals of this dissertation I draw upon some, but not all, of the data gathered during the complex excavation history of this site. Artifact distributions drawn from systematic testing of the Preston Plains Energy Center project area are reviewed. I extensively reflect on data from Locus I, partially because I am so familiar with it but also because it contained cultural deposits in spatial association with DSFs, which distinguishes it as an appropriate focus area for dissecting their interrelationship. I also draw heavily on data gathered during the two machine trenching episodes, which provide insight into site geomorphology, formation processes, and the distribution and variability of DSFs. With some regret, I concluded that providing detailed archaeological reconstructions of Late Archaic Period occupation at the other site loci required more time and effort than can I could practically invest for this dissertation. 1.4. Theoretical Outlook Research at Preston Plains has been guided by the assumption that every site and region is geologically unique (Waters 1992: xxi), and shares the perspective that every archaeological investigation begins as a problem in geoarchaeology (Renfrew 1976:2). This calls for an interdisciplinary approach grounded in geoarchaeology, which denotes the routine application of concepts and techniques drawn from the earth sciences to 11 archaeological research (Butzer 1982:35). Some of this work was accomplished through stratigraphic analysis and absolute dating techniques – the fundamental tools of geoarchaeology. Though the author has some formal training in geomorphology, outside specialists from the earth sciences were enlisted to generate additional observations, speculations, and interpretations useful for interpreting soil development and site formation processes at Preston Plains. Robert Thorson, professor of geology at the University of Connecticut, generously donated his time and expertise consulting at the site in 2008. Additionally, Kevin McBride and the author co-hosted a field tour of the Preston Plains site for the 2008 Northeast Regional Cooperative Soil Conference, based out of Rhode Island. Soil experts in attendance offered several useful observations regarding exposed soil profiles. Information gained through these collaborations has proven valuable for developing site interpretation. Most critically, this study illustrates that geomorphological conditions influence the position of archaeological material and that identifying these is prerequisite to developing an informed site interpretation (e.g. Holmes et al. 2008; Thorson 1990). Arguably, the enigmatic character of DSFs at Preston Plains seems to lie in their variability of size, morphology, stratigraphy, and orientation. Similar perceptions have been expressed by archaeologists investigating DSF complexes in the Middle Atlantic. For example, when visiting excavations at the Hickory Bluff Site in Delaware, Kevin Cunningham “remarked that the observed feature appeared rather “chaotic” in structure – that is, not one of our types looked exactly alike” (Petraglia et al. 2005:i). Assuming that DSFs at Preston Plains and Hickory Bluff are related phenomena, it seems likely that 12 many factors are implicated in the expression of such variability. Mueller, who has also investigated D-shaped pit complexes in the Middle Atlantic, advises researchers to assess each feature individually through detailed analysis using as many lines of inquiry as possible and cautions against classifying pit features according to the rigid dichotomy of pit house versus tree throw (Mueller et al. 1997). Considering multiple modes of formation is central to the investigation of features that are so variable in character (Muellerand Cavallo 1995), as understanding such modes is prerequisite to knowing the cultural past (Schiffer 1987:xx). Accordingly, investigations at Preston Plains have not only been carried out within the context of geoarchaeology, more importantly, under the broader assumptions of contextual archaeology (Butzer 1982; Schoenwetter 1981), “a systems approach in which the contextual components of the human ecosystem are reconstructed and the interactions between them are used to explain cultural stability and change (Waters 1992: 4).” The contextual components are viewed as interrelated within a broader system that includes climate, cultural, fauna, flora, and landscape subsystems. Appropriate specialists were enlisted to contribute to this broader contextualization, including zooarchaeologist Randy Nokes and archaeobotanist Tonya Largy. According to Schiffer (1987) meaningful reconstructions of human behavior at any site depend on reconciling two processes: cultural transformations and natural transformations. Thus, local and regional spatial patterning of archaeological resources should be interpreted in view of natural site formation processes occurring at similar scales of magnitude (Waters 1992: 11). These, in turn should be synthesized into a 13 landscape reconstruction that account for both cultural and natural components. Interpretive frameworks that are congruent with this outlook have been employed by some researchers studying DSFs in the Middle Atlantic (e.g. Mueller and Cavallo 1995; Mueller et al. 1997; Petraglia et al. 2005), and this study follows suit. 1.5. Analytical and Interpretive Design The analytical design of this dissertation combines scholarship, empirically-based data assessments, and hypothesis testing to holistically inform an interpretation of the genesis and cultural significance of DSFs at Preston Plains. The scholarship compiled for this purpose includes an overview of local environmental and cultural history (Chapter 2). This provides necessary context for interpreting the archaeological record of Preston Plains. Additionally, a summary of research relative to DSFs in the Northeast is provided that describes existing hypotheses that explain the nature of these features (Chapter 3). Because this dissertation supports the hypothesis that tree throws generated DSFs, a literature review on naturally occurring tree throws is provided in addition to a discussion of the archaeological significance of tree throws (Chapter 4). Empirically based elements include an assessment of site geomorphology (Chapter 6). This assessment demonstrates that beneath the seemingly mundane surface of Preston Plains lies a complex of soils that vary widely in their formation histories and physical character, and provides necessary context for the interpretation of DSFs therein. An overview of Preston Plains archaeology is presented that describes the chapters of 14 culture history represented here, as well as the relative intensity of occupations over time (Chapter 7). A case study of Locus 1 is also provided, which reports on an overlapping concentration of Late Archaic Period cultural materials and DSFs (Chapter 8). This data is reported in detail to reveal the nature of human occupation, the physical character of DSFs, and the distinct spatial patterning formed by the combination of the two. Furthermore, because this dissertation supports the hypothesis that tree throws generated DSFs, three contemporary tree throws in settings analogous to Preston Plains are excavated to assess their morphology and stratigraphy (Chapter 10). This actualistically collected data is used to bolster the final interpretation. The hypothesis that there are no artifact concentrations on the bottoms of most DSFs at Preston Plains is tested, the purpose of which is to discount the possibility that these features were created by Native Americans for use as pit houses (Chapter 9). It is assumed that if humans used the bottoms of DSFs as habitation surfaces this behavior would be reflected by remnant artifact concentrations. Quantitatively demonstrating the absence of such concentrations allows us to more confidently support the natural tree throw hypothesis for DSF genesis. The scholarship, observations, interpretations, and testing results presented in all of these chapters are drawn upon in Chapter 11 to holistically inform an interpretation of the genesis and cultural significance of DSFs at Preston Plains. The significance of this interpretation to the Cultural Resource Management industry and the academy is presented in the final chapter – Chapter 12. 15 Chapter 2. Environmental and Cultural Context for New England To interpret the relationship between ancient peoples and their environments, archaeologists draw upon a wide variety of data sets. These include archaeological data that directly reflect environmental character, such as charred plant remains and faunal material, in addition to the inferred functions or meanings of artifact types in regard to what resources they were used to exploit, influence, or represent. Natural science contributes additional data categories from the realms of bathymytry, pollen and plant macrofossils, geology, and ice cores. By synthesizing such information, this chapter presents a diachronic interpretation of human occupation in New England against a backdrop of environmental evolution as context for the archaeological analysis, geomorphological interpretation, and modeling of landscape evolution for the Preston Plains site. For the convenience of discussion, this section is organized according to four sequential climactic periods: the Late Pleistocene, Early Holocene, Middle Holocene, and Late Holocene. radiocarbon years. Bear in mind that local environments may change according to regional factors at different rates and times (Bryson et al. 1970), and that human adaptations to these local environments are probably rapid and nuanced. Therefore, the following discussion Unless otherwise stated, all dates are expressed in uncalibrated should be recognized as highly simplified, describing only the broadest of cultural and environmental trends. 16 2.1. The Late Pleistocene (ca. 18,000-10,000 BP) During the Wisconsin glaciation, most of Canada and much of the northern United States was blanketed by the Laurentide Ice Sheet, which reached its maximum extent from ca. 20,000-18,000 BP. New England was beneath this sheet, which deposited Long Island as a terminal moraine when it began retreating at ca. 18,000 BP (Uchupi et al. 2001). Connecticut was uncovered by about 16,000 BP. By 14,000 BP, all of southern New England was free of ice, and areas of higher elevation had emerged in the north (Davis and Jacobson 1985), with only lowlands and most of Maine still covered. By 12,000 BP the ice sheet was positioned along the north bank of the St. Lawrence Seaway. There is some disagreement regarding the timing of glacial recession in northern New England. Marine invasion of the Central St. Lawrence River lowlands has recently been dated at ca. 11,100 BP (Richard and Occietti 2005) which, in turn, suggests that previous regional chronologies for the retreat of the Laurentide Ice sheet are 4001000 years younger than previously thought. According to the delayed isostatic rebound of interior areas relative to coastal areas (Snow 1980:106), several proglacial lakes were formed, including Lake Winooski (VT), Lake Merrimack (NH) and Lake Ashuelot (NH). Lake Hitchcock was the most extensive glacial lake in New England, which, at its maximum, extended from Rocky Hill, CT to St. Johnsbury, VT (Rittenour n.d.) and had a maximum width of 20 miles (Patton and Kent 1992). This lake drained at ca. 13,500 BP (Stone et al. 1998), responding to erosion of the sill that impounded it and isostatic rebound to the north. The newly formed Connecticut River began to incise Hitchcock‟s lakebed deposits. 17 Sea level had been lowered during the last glacial maximum, exposing areas of the Continental Shelf (Gaudreau 1988). Southern New England‟s early post-glacial coastline extended far beyond its present limits, while northern New England‟s coastline was not vastly different (Uchupi et al. 2001). Between 12,000 and 8000 BP marine transgressions flooded much of the deglaciated coastline. The Hudson Shelf and Block Island valleys (submerged paleolandscape features) were likely formed when the natural moraine dams of pro-glacial lakes catastrophically broke, entraining large quantities of earth materials, in addition to late Pleistocene magafauna living along the lake shores. As the Laurentide Ice Sheet retreated, a continuum of tundra-woodland-forest zones followed without major hesitation or reversal (Davis and Jacobson 1985). Ice calving in the St. Lawrence Seaway cut off the ice supply to New England by ca. 13,000 BP, while ice still extended further south in areas west of New England. This resulted in New England appearing as a tundra and woodland “corridor” for at least a millennium (Davis and Jacobsen 1985). The southern New England landscape experienced periglacial conditions until ca. 12,500 BP, with postglacial winds depositing a ubiquitous blanket of aeolian sediment across the region. (McWeeney 1999:6). Retreating glacial ice was initially followed by tundra vegetation dominated by dwarf willow, sedges, herbs and shrubs that existed in southern New England from ca. 15,000-13,000 BP (Stone et al. 1998). By 14,000 BP, poplar and mixed forests had moved into southern New England, glacial ice had retreated into Maine and the northernmost reaches of New Hampshire, and between the two was an extensive zone of tundra. The onset of the Bolling warm period (ca. 13,000-11,000 BP) prompted further changes in New England‟s climate favorable for 18 boreal and temperate vegetation (McWeeney 1999). The approximate order of arrival for tree taxa during the Late Pleistocene is: “poplars (13,000-12,000 year B.P. in the south), spruces, paper birch, and jack pine, followed by balsam fir and larch, and possibly ironwood, ash, and elm, and somewhat later by oak, maple, white pine, and finally hemlock (10,000-9000 year B.P. in the south)” (Davis and Jacobsen 1985). By 12,000 BP zones of poplar, followed by mixed woodlands and forest, extended northeasterly into southern New Hampshire and Maine, reflecting accelerated climatic warming (Davis and Jacobson 1985). A mixed hardwood-conifer forest was well-established in southern New England (McWeeney 1999), tundra was confined mostly to northern New England (Davis and Jacobsen 1985), and a remnant glacial ice mass was present over some of northern Maine. After 12,000 BP trees spread broadly into northern New England, and as forests became more established their composition diversified until 9000 BP. The mixed conifer-northern hardwood forests of the Pleistocene-Holocene transition lacked good analogues with modern forests (Delcourt and Delcourt 1987). After the Bolling warm period, colder conditions of the Younger Dryas climactic reversal ( ca. 10,800-10,200 BP) prevailed across New England as glacial ice readvanced in Canada and northern Maine (Borns et al.2004). The Younger Dryas is characterized by a global temperature drop of approximately 4 degrees, which is reflected in southern New England by an increase in spruce, white birch, and alder pollen, along with a decrease in oak pollen (McWeeney 1999). During this time some wetland basins turned into grassy marshes. Major climactic transitions, such as the onset and termination of the Younger 19 Dryas, were not smooth, but chaotic events that required extremely rapid reorganizations in atmospheric circulation (White and Barlow 1993; Yu and Eicher 1998). The end of this interval is marked by the appearance of oak charcoal in Connecticut dated to approximately 10,200 BP and a decrease in boreal tree taxa (McWeeney 1999). Human population of New England was not possible until a biotic community had established itself on the post-glacial landscape (Boisvert 1999:2). The time lapse between deglaciation and biotic establishment may have been decades or centuries. The oldest known cultural signature is assigned to the Gainey-Bull Brook Phase (ca. 10,700 BP) of New England‟s Paleoindian Period (ca. 10,800-10,500 BP), which post-dates the Clovis culture horizon of the New World. Speiss et al. describe the Gainey-Bull Brook Phase colonization as coinciding with “the beginning of the end of the distinctive Paleoindian pattern” of greater North America (1998:249). Curiously, the Bull Brook Phase pioneers arrived at least two millennia after substantial communities of vegetation had become established in southern New England, and during the onset of the Younger Dryas cold interval. Why this region was not colonized earlier is a mystery, but Clovis populations may have preferred now submerged portions of the Atlantic Continental Shelf, and/or their sites may be insufficiently visible in New England‟s archaeological record. Lavin notes that oysters were available in abundance on exposed portions of the southern Atlantic Shelf during the Younger Dryas, though it is not known if Paleoindian populations exploited them (1988: 103). The pioneers associated with the Gainey-Bull Brook Phase continued to inhabit the region through the Younger Dryas, with their descendants leaving archaeological deposits associated with the Michaud-Neponset (ca. 20 10,200), Crowfield (unknown dates), and Nicholas (10,100-10,050) phases (Speiss et al 1998). The human population of New England was small during the Younger Dryas, generally estimated at less than 25,000 individuals (0.14 persons/km2) by Snow (1980:157), while Jones proposes a much lower cap of 2000 individuals (0.01 persons/km2) by 10,000 BP (2004). By the end of this period, the region hosted a variety of point types, some of which were probably contemporaneous, which likely emerged from an interplay of population growth, immigration, and emigration. The first arrivals developed a restricted wandering system after bands fissioned and/or new bands arrived from elsewhere (Snow 1980:150). The Paleoindian subsistence adaptation was probably similar to ethnographic examples from the subarctic that feature “caribou hunting, small mammal trapping, and seasonal plant use in a region which straddles the treeline” (Speiss et al. 1998: 227) A focal strategy of transhumant caribou hunting has been proposed (Pelletier and Robinson 2005), and though mammoth and mastodon are known to have inhabited the region in the late Pleistocene, as of yet, there is no evidence demonstrating that they were hunted by resident Paleoindians (Speiss et al. 1998:226). The use of high-quality, exotic cherts reflects high mobility, though it has not been resolved if lithic procurement was logistically organized or embedded in seasonal population movements (sensu Binford 1979). The settlement pattern is thought to reflect “short-term occupation by small groups which rarely returned to the same location.” Unusually large sites, such as Vail and Bull Brook, may represent “social aggregations of the kind observed among caribou hunters in 21 the arctic and subarctic” (Pelletier and Robinson 2005:165; Robinson et al. 2009). This cultural system lost its regional monopoly during the early Holocene. 2.2. The Early Holocene (ca. 10,000-8000 BP) Warmer atmospheric temperatures of the Holocene caused glacial retreat to continue, sea level to rise, and shorelines to reconfigure in the balance of eustatic and isostatic shifts. New England‟s environmental changes presented new parameters for the expression of human occupation. Warm and cold climate fluctuations characterized the first few hundred years of the Early Holocene, with the greatest occurring between 10,000 and 9600 BP (McWeeney 1999). Seasonal temperature fluctuations were more extreme than in the present, with hotter summers and colder winters. Warm-climate tree species began to flourish in southern New England, replacing the cold-adapted spruce, larch, and fir (Davis 1983). Charcoal from local sites reflect the establishment of diverse tree species (McWeeney 1999:9), including white pine, birch, beech, and oaks. By 9,000 BP a closed forest canopy was established across greater New England that included white pine, oak, elm, ash, birch, ironwood, and sugar maple (Davis 1983), though it was not homogenous in character. As the ameliorating climate caused differences between upland and lowland vegetation to develop, the regional gradient of vegetation composition across southern New England was likely established between 9500 and 8000 BP (Oswald et al. 2007). Additionally, the northernmost portions of New England hosted a different suite of vegetation. At about 9500 BP the Laurentide Ice sheet stood just inland from the north 22 shore of the St. Lawrence River (Dumais 2000:85), which probably contributed to a colder and drier local climate. Accordingly, vegetation likely favored cold-adapted trees (spruce, aspen and fir), and perhaps open areas featuring herb or shrub tundra. Palynological data from Moulton Pond, Maine indicate that the local forest closely resembled the pine-northern hardwood forests of modern central New England (Davis et al. 1975). The warmer and drier Hypsithermal Period, which began at approximately 9000 BP and continued through the Middle Holocene, is associated with the lowering of water tables and development of shrub swamps in southern New England. This climactic trend did not occur without interruption. From ca. 8400-8000 BP the climate system of the North Atlantic Basin experienced a cold shift of about half the magnitude of the Younger Dryas, where a cooler, drier, and windier climate prevailed (Alley et al. 1997). Archaeological perceptions of early Holocene New England have been heavily influenced by Ritchie (1965) and Fitting (1968), who envisioned the region as largely abandoned by humans from ca. 10,000-5000 BP due to the spread of a boreal forest that afforded few opportunities for foraging. This hypothesis has been largely abandoned in lieu of new archaeological discoveries and environmental data. In a drastic revision, Nicholas argues that New England was as an expanse of landscape mosaics during the early Holocene that would have been particularly attractive to foragers (1987:105). According to his Glacial Lake Basin Mosaic model, more recent evidence: 23 suggests a diverse set of landscape features developing within former lake basins, consisting of lakes, ponds, extensive wetlands, and emergent riverine systems.” The effect of these wetland mosaics was the creation of local areas of high resource diversity, productivity, and reliability within the northeastern interior during the early postglacial period. His model is supported by evidence that there were high water levels in southern New England during the 10th millenium BP. Nicholas emphasizes that the diversity of resources in wetland localities made them “primary locations” on the landscape. The greatest density of wetland environments probably existed in Maine, where 10 percent of the state‟s surface area is projected to have been open ponds during the early Holocene (Robinson and Petersen 1993:70). New England‟s early Holocene archaeological record constitutes a thin but surprisingly complex fabric. Discoveries of the past twenty years have populated the seemingly barren cultural landscape once envisioned by Ritchie (1965) and Fitting (1968) with diverse cultural manifestations: the Late Paleoindian tradition, the Piedmont-toAtlantic Slope Macrotradition, and the Gulf of Maine Archaic tradition. Though their chronological spans have yet to be precisely anchored, current data suggests they were coeval during most, if not all, of the ninth millennium BP (Ives 2006; Jones 2006). This leaves us to ponder the dynamics of human colonization and territoriality that played out across varying social and environmental contexts. 24 The notion of a temporal abutment of the Paleoindian and Archaic traditions, once thought to be definable by the disappearance of fluting and the appearance of notching (Griffin 1964, 1985), has been complicated by the recognition of the continued use of non-fluted lanceolate points. The temporal span of the region‟s Late Paleoindian period has been conservatively estimated at ca. 10,000-9000 BP (Doyle et al. 1985), and more liberally at ca. 10,200/10,000-8000 BP (Petersen et al. 2000:113). Mounting evidence from northern New England seems to support the latter more strongly (ibid.:118), with the Stones Throw site date as the newest addition (Ives 2005, 2006). Late Paleoindian projectile point styles are diverse, and the lack of a precise dating framework makes it difficult to determine whether these are sequential, overlapping, or coeval. The fluted point forms of the Paleoindian tradition gave way to unfluted, parallel flaked, lanceolate projectile points (Ritchie 1985:414). Some local forms appear to constitute an Eastern Lanceolate tradition that evolved, in situ, from local Paleoindian predecessors. While earlier Paleoindian assemblages often include diverse materials from a broad geographic area (e.g. Speiss et al. 1998), the Late Paleoindian period, as expressed in parts of northern New England, is characterized by more local, mixed nonchert lithic materials (Petersen et al 2000:122). This probably reflects a reduction in rates of mobility and the constriction of social networks (Jones 2004). Aside from continued stylistic drift of projectile types, their lithic toolkits do not evidence diversification. In the northeastern United States rare examples of points have been found that resemble “the broad family of Plano points of the western United States and of the upper 25 Great Lakes Region” (Ritchie 1985:414). Since this observation was first made, evidence has been mounting for the intrusion of a Plano-derived complex in northern New England. Excavations in southeastern Quebec have revealed substantial Plano-complex occupations that represent the approximate eastern limit of a hypothesized migration, evidenced by a linear distribution of related sites extending from the Western Plains (Dumais 2000). The recovery of Plano-like points from sites in Maine and New Hampshire support the hypothesis that this complex intruded into northern New England. The scarcity of Late Paleoindian sites across the region suggests a low population density, while continuities in lithic technology suggest the persistence of a subsistence strategy focused on caribou hunting. Most Late Paleoindian point finds are from northern New England, suggesting a correspondingly northern occupation. Dincauze (1971;1976) envisions “the entire Atlantic coastal area from North Carolina to New Hampshire” as a “single great culture area” by the eighth millennium B.P, based on widespread similarities of stemmed projectile point styles. This Atlantic Slope Macrotradition is defined, in part, “by a common bifacial chipped-stone technology” (Forrest 1999:81). It appears to represent the expansion of the Piedmont Tradition, which originated in the Carolinas. Northeastern examples of Piedmont Tradition point types, such as Hardaway, Kirk, and Palmer, are thought to date to between 9600 and 8500 BP in the Northeast (Funk 1996:13), though their occurrence is exceedingly rare and securely dated examples are scarce (Forrest 1999:81). The only known concentration of specimens in southern New England is from a cluster of sites in the Robbins Swamp Basin in northwestern Connecticut (Nicholas 1988:272-273). 26 The Piedmont Tradition, after expanding across most of the Atlantic Slope macroregion, appears to have gained a foothold in southern New England during the 9th millennium BP, as manifested by slightly higher numbers of bifurcate-base points. According to Forrest (1999:81): slightly younger Piedmont Tradition sites are more common and more easily characterized. Sites dating between 8600 and 7800 BP typically contain relatively numerous expedient stone tools, including a variety of scrapers and other unifaces, “choppers” and small numbers of bifurcatebased points. The Robbins Swamp, which yielded substantial quantities of bifurcates, is a proposed core area for Piedmont Tradition occupation in southern New England (Nicholas 1988), in addition to the Taunton River and Titticut basins in Eastern Massachusetts (Taylor 1976; Johnson 1993). The adjacent Titicut and Seaver Farm sites on the Taunton River of southeastern Massachusetts are postulated to represent a single base camp, the only one recognized in New England‟s archaeological record to date (Dincauze and Mulholland 1977). Many Piedmont-to-Atlantic Slope Macrotradition points were heavily curated and reflect a complex use-life, which is consistent with expectations of the personal gear (sensu Binford 1979) of mobile foragers. There is less evidence for regional exchange of exotic cherts (Snow 1980:172) for bifurcate manufacture, and an increasing preference 27 for local quartz and quartzites. Bifurcates from the Dill Farm site in East Haddam, CT were made from local quartz and quartzite, in addition to more exotic materials such as Boston Basin rhyolites and Hudson Valley cherts (Pheiffer 1986), suggesting significant mobility. A substantial cache of quartz chunks found at the Dill Farm suggest long-term planning. Non-projectile elements of Atlantic Slope assemblages feature “heavy cutting tools, choppers, and steeply retouched scrapers” (Forrest 1999:81). Large choppers, or digging tools, likely reflect increased exploitation of roots and tubers (Snow 1980:170). The subsistence strategy of the Piedmont-to-Atlantic Slope Macrotradition is inferred from a thin database, as diagnostic artifacts are usually isolated or found on sites with later components (Snow 1980:169). Lavin proposes that large population of deer and wild turkey could have been supported by the mast forest of the Early Archaic (1988: 101), and that local foragers primarily hunted and gathered “interior food sources, particularly deer, nuts, and freshwater fish” (1988: 104). Evidence of root/tuber exploitation supports an orientation to a more diffuse subsistence strategy than Paleoindian foragers, as suggested by Snow. A low population density is estimated by the relative scarcity of diagnostic projectile points. The settlement system was likely one of restricted wandering, defined as “communities that wander about within a territory that they define as theirs and defend against trespass, or on which they have exclusive rights to food resources of certain kinds. Movement within the territory may be erratic or may follow a seasonal round, depending on the kind of wild food resources utilized” (Beardsley et al. 1956:136). 28 The third early Holocene New England cultural tradition is the Gulf of Maine Archaic, which persists through the middle Holocene in northern New England according to cultural chronology of the Sharrow Site in Maine (Petersen and Putnam 1992). If the Sandy Hill radiocarbon series is representative of Gulf of Maine Archaic occupation in southern New England, it has a more limited range of ca. 9500-8000 BP in this region (Forrest 1999; Jones and Forrest 2003). This tradition is distinguished by its microlithic industry, which is thought to be associated with the production of compound tools (Robinson and Peterson 1993). Assemblages from the Brigham and Sharrow sites in Milo (Petersen 1991; Petersen et al. 1986), the Blackman Stream site in Bradley (Sanger et al. 1992), and the Sandy Hill Site at Mashantucket (Forest 1999) reflect the selection of immediately available, coarsegrained stones as raw material for microlith production. Though the origin of this industry is indeterminate, we should consider the earlier, local Paleoindian population as possible progenitors, as their technological repertoire included quartz microcore technology. This is evidenced at the Adkins site in Maine‟s upper Magalloway valley, which features a quartz microlithic sub-assemblage (Gramley 1988:22; ibid [see Plates 21-23]:109, 111). Large choppers and hoe-like forms recovered at Sandy Hill likely functioned as digging implements. Larger artifact classes have been recovered from Gulf of Maine components at the Brigham and Sharrow sites as well, including adzes, celts, and gull-channeled gouges (Robinson and Petersen 1993:68); tools well-suited for dugout canoe manufacture. 29 The subsistence strategy of the Gulf of Maine Archaic tradition appears to be centered on exploiting wetland environments, which likely provided highly predictable resource patches. The deeply stratified Sandy Hill and Sharrow sites, with their overlapping lenses of “black sand” deposits, suggest intensive occupational episodes with repeated reoccupations. At Sandy Hill, these deposits form floor layers interpreted to represent the subterranean remnants of pit houses. These are argued to represent fall/winter occupations, though seasonality has yet to be confidently ascertained. Microscopic analysis of vegetative tissues from feature soils identified several wetland plant species with edible tubers (Forrest 1999: 94). 2.3. The Middle Holocene (ca. 8000-5000 BP) The Middle Holocene is characterized by a global climactic warming that altered the character of local environments and influenced adaptive strategies of people. This warming trend is referred to as the Hypthithermal Interval (ca. 9000-5000 BP), or Holocene Climatic Maximum (Deevey and Flint 1957). Though scholars debate the timing and duration of this phenomenon, there is a general consensus that its environmental effects are robustly evident by 8000 BP and persisted for at least three millennia. New England data indicates that the middle Holocene was characterized by fluctuating environments with a predominant warming trend (McWeeney 1999:9). The interval of maximum warmth, expressed as an average rise of 2 degrees in global temperature, appears to have started at ca. 9000 BP and lasted until at least 5000 BP 30 (Davis et al. 1980). This is based on the distributions of pollen and plant fossils from six sites at different elevations in the White Mountains. The extension of white pine and hemlock to higher altitudes here during this period is interpreted as evidence of Holocene warming. Additionally, the pollen record from Rogers Lake in southern Connecticut reflects a decline in pine at approximately 8000 BP, which thought to reflect local changes in vegetation that coincide with the onset of the “prairie period” of the Great Lakes region (Davis 1969). Overall, an increase in herbaceous plants was accompanied by a decline in forest trees, and the extension of forest tree ranges (McWeeney 1999). The drier climate favored oak trees in southern New England, which is reflected in the prevalence of oak fuelwood at Middle Archaic sites (McWeeney 1999:10). Other dominant mast trees included hickory and chestnut. By the opening of this period, the 20% oak isopoll had already moved to the northern boundary of Massachusetts, where it has remained until the present (Snow 1980:173). North of the isopoll, Maine‟s interior forests were dominated by pine and hemlock from approximately 7500 to 5000 BP (Speiss and Lewis 2001:160). Water levels dropped throughout southern New England, shrinking lakes and turning shallow ponds into swamps or meadows (McWeeney 1999). Aeolian sediments, reflecting an increased frequency of forest fires, have been interpreted as Middle Holocene markers at stratified archaeological sites in Connecticut (Thorson and McBride 1988). Swamp cores from across southern New England show an increase in charred organics during this period (McWeeney 1999). Additionally, sea level rise and the 31 inundation of the southern New England coastline reduced the gradients of stream and river drainages which likely fostered the development of floodplains. The southern New England coastline was far from its present location at 8000 BP, and many Middle Holocene coastal sites were probably destroyed or submerged as it receded (Snow1980:173). Throughout the Holocene, there has been little change in the geomorphological character of regions along the Maine coast (Speiss and Lewis 2001:159). Coastal transgression in Maine was relatively rapid before 5000 BP, but as coastal environments shifted inland and up estuaries they consistently featured beaches and mudflat settings. Ragweed pollen in southern New England sediment cores appears to be of local origin (Faison et al. 2006), which supports the interpretation that local forests consisted of mosaics of openland vegetation and forests. The modern Prairie Penninsula region of the upper Midwest is viewed as an analogue to vegetation cover across much of southern New England‟s terrain during the Hypsithermal. This does not mean the landscape was a monotonous patchwork, however. Spatial heterogeny of forest composition appears to have increased during the Middle Holocene (Oswald et al. 2007), so there may have been particularly productive biomes in certain locales. Traditional interpretations of New England‟s paleoenvironmental record for the Middle Holocene are currently under question. While higher temperatures and lower lake levels have been proposed for this time, Shuman et al. (2004) recognize raised lake levels by 8200 BP, note the abundance of mesic hemlock and beech over dry-tolerant pines at that time, and state that cooler-than modern temperatures persisted until 6000 32 BP. Foster et al. (2006) concur with this reading of the past, recognizing 5000-3000 calendar years BP as a marked period of dryness and lowered lake levels. Archaeology at southern New Hampshire‟s Neville Site indicates that the Atlantic Slope Macrotradition persisted through the middle Holocene (Dincauze 1976). Neville, Stark, and Merrimack points recovered there and now accepted as temporally diagnostic of the middle Holocene, are morphologically similar to contemporaneous projectiles from the Carolina Piedmont area, which suggests a broad cultural affinity among populations along the eastern coast of North America. Human populations associated with the Atlantic Slope Macrotradition followed generalized subsistence strategies and concentrated their settlement around waterfalls, river rapids, major river drainages, wetlands, and coastal settings (Bunker 1992; Dincauze 1976) establishing large base camps along extensive wetland systems (Doucette and Cross 1997). A pattern of seasonal rounds in fairly large territories has been proposed, with annual movements structured around seasonally abundant resources (Dincauze and Mulholland 1977). Netsinkers and plummets appear in the region‟s archaeological record for the first time, and elevated levels of mercury in the soil at the Neville Site suggest harvesting of migratory fish. The orientation toward aquatic resources during this period has been interpreted to indicate that forests were less productive than in the following late Holocene Period. Data from the Lower Hudson River indicates that local foragers exploited shell fish by 7000 BP (Brennan 1974), and it is possible that earlier middens exist on the submerged coastal plain. However, shellfish 33 exploitation probably contributed little to daily dietary requirements prior to the Late Holocene (Lavin 1988). Archaeological signatures of the Gulf of Maine Archaic Tradition continue during the middle Holocene until approximately 6000 BP, but only in the core territory of the Gulf of Maine, according to current knowledge. Unifacial core technology, the hallmark of the Gulf of Maine Archaic Tradition, was largely abandoned in New England by the Late Holocene, which may coincide with an abandonment of a unique brand of semisedentary wetland foraging. However, elements of burial ceremonialism associated with the Gulf of Maine Archaic Tradition, such as cremation and the interment of ground stone objects, appear to have endured within the subsequent Moorehead Tradition of the Northeast‟s Maritime Areas (Robinson 1992). Dincauze initially recognized continuity in the cultural sequence of the Neville Site, with Merrimack points (6th millennium) sharing traits with the later Wading River and Squibnocket Stemmed points (5th millenium), implying that a technological tradition of stemmed points continually evolved in southern New England through the Middle and Late Archaic periods. She later rescinded “there is no demonstration at this site for development or any close relationship between Merrimack and any small stemmed points” (Dincauze 1976:128) and proposed an occupational hiatus for the sixth millennium BP (ibid:135). Alternately it should be considered that Merrimack points are commonly misidentified as Wading River points by archaeologists. 34 2.4. The Late Holocene (ca. 5000 BP-Present) Though the late Holocene is generally thought to represent modern environmental conditions, the climate fluctuated significantly (McWeeney 1999:10). Repeated cooling episodes, or “Little Ice Ages,” occurred during this interval: at 4330, 3290, 2250, 1550, and 650 BP, while a warming period also occurred at ca. 1000 AD. The period between 5000 and 3000 calendar years BP is characterized by dryness, lowered water levels, and significant changes in vegetation across southern New England (Foster et al. 2006; Lavin 1988: 106; Yu et al. 1997). Hemlock pollen abruptly declines in New England‟s record at approximately 4750 BP and disappears entirely within 500 years, and is thought to have been caused by climate change (eg. Fuller 1998). A concurrent decline of other forest taxa, such as oak, and intervals of drought at regional to continental scales further suggest that there were abrupt dynamics in forest ecosystems. Reduced precipitation in the Great Lakes area during this time may have been caused by frequent eastward extension of dry Pacific air (Yu et al. 1997), which suggests that lowered water levels in New England may be more closely related to moisture source history than paleotemperature. Moisture availability rose to modern levels by approximately 3000 BP, which, in turn, caused moisture dependant taxa such as chestnut to increase (Shuman et al. 2004). Increased percentages of white pine in the pollen record of Cape Cod confirms a return to moister conditions in New England by about 3500 BP (Winkler 1985). The stabilization of the southern New England coastline at ca. 4000-2000 BP is thought to have allowed the development of extensive marshlands that were highly productive for human 35 exploitation (Lavin 1988). Pollen assemblages reflecting the modern environment have only been deposited for the past 2,000 years (Davis 1969:421). Hickory trees, an important element of the region‟s mast forests, became established in Connecticut at ca. 5500 BP (Davis 1969). In southern New England numerous additional deciduous and conifer trees became established during the Late Holocene, with a terrestrial environment that included oak, beech, elm, maple, and hickory (McWeeney 1999:11). Portions of the Boylson Street Fishweir in Boston, dating to ca. 4000 BP, suggest a locally diverse forest environment, with taxa including red oak, beech, sassafras, alder, sycamore, aspen, white oak, dogwood, bayberry, larch and hemlock. Northern New England‟s interior forests continued to change during the late Holocene as well. After the disappearance of hemlock trees in central Maine at ca. 5000 BP, forests were dominated by birch and beech trees sustained by the cooler climate (Speiss and Lewis 2001:160). Continued cooling after 2000 BP fostered the shift to a spruce-fir dominant forest. The spread of spruce forest cover appears to correlate with the increased ratio of moose to deer in the archaeological record (ibid:161). There is no overarching model for costal adaptations in New England, as different localities have very distinct patterns of exploitation that closely reflect changing shoreline environments and shellfish abundance (Braun 1974). Overall, sea level rise slowed greatly in southern New England during the late Holocene, resulting in more extensive floodplain development, shellfish beds and coastal marsh environments (Lavin 1988). Indians in northern New England did not develop a robust coastal adaptation until 36 approximately 5000 years ago, mainly because isostatic rise substantially outpaced eustatic rise (Snow 1972). Coastal shellfish exploitation in northern New England became intensive approximately two thousand years ago when sea level was at or near modern levels and flooded estuaries provided more stable beds (Snow 1972). Coastal transgression in Maine was slightly slower (than in the Middle Holocene) from 5000 to 2500 BP, and relatively slow after 2500 BP (Speiss and Lewis 2001:159). Tidal amplitude in the Gulf of Maine continued to increase during the Late Holocene, as did the upwelling of cold waters. The resulting tidal mixing would have eliminated the warm surface and inshore waters conducive to swordfish and quahog (ibid:161). The high density of Late Archaic sites in New England and the almost exclusive reliance on locally available lithic materials suggests higher populations during the Late Archaic Period (ca. 5000-3000 BP) (Dincauze 1975), when three archaeological traditions, the Laurentian, Small or Narrow Stemmed, and the Susquehanna are featured in southern New England‟s archaeological record. Ritchie defines the Laurentian tradition as “an extensive cultural continuum, widely spread throughout northeastern North America” (1965:79-80) associated with a hunting and fishing culture that spread southward and westward from a cultural hearth centered on the upper St. Lawrence Valley, manifesting local adaptations as it assimilated new traits from other Archaic cultures. In southern New England, the Laurentian tradition is commonly identified by Otter Creek, Brewerton, and Vosburg points. Laurentian manifestations in Connecticut are equated with the Golet phase of the Late Archaic, with an associated date range of ca. 37 4800-4200 BP (McBride 1984). Ritualized cremation burials from the Bliss Site suggest complex mortuary practices (Pfeiffer 1984). The Narrow Stemmed tradition of southern New England‟s Late Archaic appears to have temporally, and perhaps culturally, overlapped with the Laurentian tradition. The Narrow Stemmed tradition is characterized by a quartz cobble lithic industry employing bipolar reduction techniques (McBride 1984), which is proposed to be an intrusive manifestation (Dincauze 1976:128) originating in the Middle Atlantic (Dincauze 1968: 214, 219). The earliest Narrow Stemmed points in southern New England date from ca. 4600-4500 BP (Dincauze 1975:24; Hoffman 1985) after which the population became denser and more dispersed across the landscape (Hoffman 1985). Coastal sites, generally rich in shellfish remains, were used primarily for summer occupations (Snow 1980). The use of food-grinding instruments suggests exploitation of vegetable foods, including acorns and other Mast Forest nuts. Mortuary sites associated with this tradition are only well known in the Taunton River drainage, where individuals were cremated (ibid:232). Two local phases of the Narrow Stemmed Tradition have been proposed based on survey data from the lower Connecticut Valley; the Tinkham and Vibert phases (McBride 1984). The Tinkham phase is identified according to the presence of Narrow Stemmed varieties of projectile points among cultural deposits dating from ca. 4200-2900 B.P. Though quartz is the most common material in Narrow Stemmed assemblages, quartzites are commonly found in assemblages in eastern Connecticut (McBride 1984:59). Settlement patterns suggest seasonal aggregations in wetland areas, with task-specific camps in woodland areas (McBride 1984:67). The Vibert phase dates to ca. 4000 B.P. 38 and is identified according to the presence of Squibnocket points in association with a cobble lithic industry (ibid:60). Pitted stones, possibly used to split nutshells or as anvils for bipolar reduction, also occur in these assemblages. Vibert Phase sites suggest use of “non-riverine areas, particularly in terrace zones” (ibid:70), and a lack of large sites suggests small, mobile bands making frequent residential moves (ibid:71). McBride‟s Vibert phase is the local expression of Ritchie‟s Squibnocket complex which features Wading River, Squibnocket Stemmed, and Squibnocket Triangle points (Ritchie 1969: 214-5). Squibnocket components from Martha‟s Vineyard evidence an emphasis on hunting of both terrestrial and marine mammals. Shellfish were heavily utilized, while fishing was of less importance. Bear in mind that archaeologists with theoretical perspectives based in cultural ecology have encouraged us to move beyond discussing the Late Archaic record in terms of cultural-historical traditions (ie. Laurentian versus Narrow Stemmed) (Tuck 1978; Snow 1980), and some have made concerted efforts to do so (e.g. Lavin 1988; Pfeiffer 1984). However, the temporal framework and lexicon of these traditions still provide a common basis for discussion of the Late Archaic in New England, though associated meanings have become more varied and flexible over time (Funk 1988). Northern New England hosted another cultural manifestation during the Late Archaic Period known as the Maritime Archaic Adaptation, or Tradition, which “occurs in the drainages of Maine and New Brunswick, as well as in the other Maritime provinces, Newfoundland, and around the shores of the Gulf of St. Lawrence” (Snow 1980:188). People associated with this tradition probably alternated their activities 39 between summer coastal locations and the winter interior locations (ibid: 200). Interior hunting likely focused on woodland caribou, while coastal hunting was more diversified. Archaeology at the Turner Farm in Penobscot, Maine revealed year-round occupation of a coastal site that is interpreted to represent a population‟s home base (Speiss and Lewis 2001). Significant resources included swordfish, cod, deer, and shellfish. The material culture of the Maritime Archaic is distinguished by “bayonets” of bone, swordfish sword, and slate, in addition to adze and gouge heads that were probably used to create large dugouts for sea hunting as well as traveling inland via tributaries (Snow 1980). Foragers associated with the Moorehead phase (ca 4500-3800 BP) (Bourque 2001:51-61) of South-Central Maine used a Narrow Stemmed technology and participated in a mortuary subsystem known as the Moorehead Burial Tradition (ca. 5000-3800 BP) (Sanger 1973). This highly ritualized subsystem is geographically bounded (Robinson 2003) and predates the appearance of the Moorehead phase culture unit. Utilitarian furnishings were interred in red-ochre graves, in addition to objects of striking workmanship, most notably, ground-slate bayonets. During the Terminal Archaic Period (ca. 3600-2800 BP), which bridges the Archaic and Woodland periods, more intensive exploitation of riverine and coastal settings occurred in New England. Increased development of coastal marsh environments coincided with an increased frequency and size of shell middens (Lavin 1988: 108). Susquehanna tradition sites are markers of the Transitional Archaic Period. The best known are cremation cemetery complexes that echo those of the preceding Moorehead Burial Tradition. Susquehanna cemetery complexes are consistently marked by the 40 presence of dark, greasy pits containing calcined bone fragments, grave goods, broken or “killed” blades, ground-stone tools and steatite bowls, and red ochre (Dincauze 1968; Leveillee 2002; Robinson 1996). The use of heavy steatite vessels suggests increased sedentism by resident populations. Diagnostic tool forms of this tradition include Broadpoints of various types (Atlantic, Coburn, Perkiomen, Snook Kill, Susquehanna Broad) and Orient Fishtail points. Though Turnbaugh emphasizes that Broadpoint assemblages “cap Narrow Stemmed point levels at several sites in the Northeast” (1975:56), some stratigraphic contexts indicate that the Susquehanna and Narrow Stemmed Traditions were coeval during the fourth millennium BP (Dincauze 1968:86). Dincauze suggests that after an extended period of separate co-existence the two traditions merged to constitute the Orient Phase of the Terminal Archaic. At the Turner Farm Site in Maine Susquehanna tradition occupants were less maritime-oriented than their Moorehead phase predecessors, which is explained by a hypothesized collapse of the Moorehead summer fishery by the introduction of cold water into Maine (Speiss and Lewis 2001:161-162). This hypothesis is further supported by an increased reliance on resources available in the intertidal and near-shore zone, such as fish, birds, and seals, by coastal populations of the Susquehanna Tradition. During the Woodland Period (ca. 3000-450 BP) New England‟s population further diversified their subsistence base, increasingly relying on shellfish and horticulture and eventually establishing year-round coastal and riverine settlements. This period is divided into the Early, Middle, and Late periods. Early Woodland Period ( ca. 3000-1600 BP) sites are typically identified by the presence of Meadowood, Lagoon, and 41 Rossville points, as well as grit-tempered, cord-marked Vinette I ceramics. Exotic lithics such as jasper, chert, and chalcedony are often present. Several indigenous plants appear to have been cultured including goosefoot, sumpweed, sunflower, pigweed, and knotweed (McBride 1978; Streuver and Vickery 1973). Because Early Woodland Period Sites are generally underrepresented in New England‟s archaeological record, a population decline has been proposed (Dincauze 1974; Lavin 1988). Climactic or environmental changes, sociocultural change, or epidemics may have contributed to the so-called “Early Woodland collapse” (Fiedel 2001). Conversely, the apparent paucity of Early Woodland sites may simply reflect the biases of site recognition strategies (Juli and McBride 1984). Direct association of Narrow Stemmed projectile points with Woodland Period radiocarbon dated contexts (Herbster 2004; Herbster and Chereau 1999, 2001, 2003), as well as stratigraphic association of Narrow Stemmed points with Woodland types (Cuzzone and Hartenberger 2009; Lavin 1985), alerts us to the probability that some Woodland Period assemblages are misidentified as Late Archaic. Diagnostic artifacts of New England‟s Middle Woodland Period (ca. 1650-1000 BP) include Jack‟s Reef Pentagonal and Corner-Notched points, Fox Creek points, and rocker and dentate-stamped ceramics. Middle Woodland Period assemblages commonly feature exotic lithic materials such as Pennsylvania jasper, which was commonly used to manufacture Jack‟s Reef points from A.D. 500 to 800 (Goodby 1988; Luedtke 1987). This trend suggests the operation of long-distance exchange networks extending from Labrador to Pennsylvania and beyond (Dragoo 1976; Fitting 1978; Snow 1980). Furthermore, Midwestern cultural influence is suggested by the fact that Jack‟s Reef 42 projectile points are virtually identical to points found in contemporary Hopewell burial contexts in Ohio (Strauss 1992). Late Woodland Period (ca. 1000-450 BP) sites are most often located in coastal environments, around interior freshwater ponds and wetlands, and adjacent to large tributary streams. Diagnostic artifacts include Madison and Levanna points and cordwrapped, stick-impressed, and incised ceramics. During the Late Woodland Period, southern New England‟s islands appear to have hosted significantly larger populations than the mainland, which may be explained in part by higher ratios of productive shallow water habitats to land (Nixon 2004:16). Isotopic analysis of human skeletal remains from Nantucket suggests that people subsisted largely on marine fish, mammals, and shellfish (ibid:14). Human density estimates based on historical sources suggests that the carrying capacity of the offshore islands of Nantucket, Martha‟s Vineyard, and Block Island was 3.5 to 10 times greater that that of the mainland shore (Nixon 2004). Abundant marine foods included oceanic fish and seals as well as shallow water fin and shellfish. Some mainland coastal settings, such as Greenwich Cove in Rhode Island, appear to have had year-round settlement during the final centuries of the pre-contact period (Bernstein 1990). Corn cultivation may have played a relatively minor role in the prehistoric diet of southern New Englanders prior to European contact (Nixon 2004), with year-round village life and a heavier dependence on corn arising in response to historic economic activities (Ceci 1980:80). 43 Chapter 3. Perceptions and Interpretations of DSFs in the Archaeological Record The current state of research indicates that DSFs are most commonly encountered on Archaic and Woodland period sites that contain deep, sandy and unconsolidated soils. Unfortunately, there is no consensus among archaeologists as to how these features formed or what they meant to prehistoric populations. Because excavation strategies and reporting standards have varied from one investigation to the next, comparing site data to construct a synthetic interpretation of DSF genesis and cultural significance remains challenging. While DSF morphologies are consistently reported, their stratigraphy and associated artifact distributions are often underreported. DSFs have been discovered at eastern North American sites in varying abundances, and researchers have generated several hypotheses regarding their genesis and cultural significance. Relevant literature from New England and the Middle Atlantic are briefly reviewed. 3.1. DSFs in Southern New England In the early 1970s the Massachusetts Archaeological Society conducted salvage excavations at the Bear Swamp II Site in Berkley, MA. Like Preston Plains, this site is situated on glaciofluvial deposits that cap glaciolacustrine sediments (Barnes 1972: 26-7). This site is best known for providing an early radiocarbon dated association for NarrowStemmed point use (ca. 4640 BP), and perhaps less remembered for two DSFs (Features 12 and 72). In her doctoral dissertation, Barnes classified them as “deep pits” and 44 described them as elliptical in plan with “a red-orange lining covered by mixed layers of ashy-grey and red-orange fill” (1972:81). Feature 72 was 3 m wide and 4.5 m long, while the smaller one exhibited similar morphology but was approximately one half as large. Each had a shallower extension on the western side containing a charcoal concentration. Few artifacts were recovered from feature fill though materials recovered from the vicinity were interpreted to be Late Archaic in age. Feature 72 had a small hearth in its upper portion that appeared to be intrusive. Barnes interpreted these DSFs to be anthropogenic and offered two functional interpretations (1972: 81-2). First, she suggested they were used for “working or sleeping quarters in cold weather,” though the smaller one seemed “uncomfortably small” for this purpose. She cited the lack of postmolds as supporting this temporary-shelter hypothesis, adding that the charcoal deposits may have been generated by a series of small fires used for heating. However, in view of the fact that these DSFs contained few artifacts, she alternately hypothesized that they were covered pits used to dry and smoke fish, meat, or skins. In North Reading, Massachusetts four DSFs (Features 1,2,3, and 11) were investigated by Doucette and Flynn (2008) at Locus 4 of the J.T. Berry Site. The J.T. Berry site is situated on a relatively level, excessively drained glacial outwash formation. Found in close spatial association with Middle and Late Archaic period deposits, the DSFs were ovate (ranging from approx. 150 – 250 cm long) and deep (ranging from approx. 120-150 cm deep), contained charcoal, and exhibited multilayered, reddened soils. Cultural material was most abundant in the topsoil above these features, while the 45 matrices of Features 1, 2, and 3 contained very few artifacts. Charcoal collected from the upper portion of Feature 1 yielded a radiocarbon date of 1890 ± 60 BP, which falls within the Middle Woodland Period, while a sample from its lower portion provided a Late Archaic date of 4760 ± 60 BP. These DSFs were tentatively identified as ceremonial features, possibly burials; however, their reddened sediments were thought to reflect fireinduced oxidation as opposed to the introduction of red ochre, which is frequently recognized as a hallmark of Archaic Period burials. Multi-year investigations of the Charlestown Meadows Site in Westborough, Massachusetts revealed a large number and variety of features, some of which could be classified as DSFs. This site was situated on a silty outwash formation at the edge of an ancient lakebed, and contained cultural components dating from 9100-3500 B.P (Hoffman 1993). Of particular interest, Area I contained a feature complex that included “nine large, deep pits filled with lithic waste” that were recovered that were interpreted as “lithic disposal pits” (Hoffman 1983). These occurred in close spatial association with Middle, Late, and Terminal Archaic Period occupational episodes, with Late Archaic Period artifacts predominating. Two DSFs were discovered at the Avery Pond Site, located on a level portion of an excessively drained outwash formation (kame mound) adjacent to the Preston Plains Site. (Ives 2007b). One (Feature 1) was ovate (approx. 2-x-1.5 m.), fairly shallow at the bottom (52 cmbs), and contained two layers of charcoal-rich, oxidized soil separated by a layer that contained less charcoal. Squibnocket triangle points and associated Late Archaic Period chipping debris was concentrated in the uppermost portions of this 46 feature, and generally centered upon it. A radiocarbon date from charcoal in the uppermost portion of this feature, in direct association with Squibnocket triangle points, yielded an uncalibrated radiocarbon date of 270 ± 40 BP, which is considered too young to be associated with the lithic materials. Wood charcoal from the bottom of this feature was dated at 3690 ± 80 BP. The other DSF (Feature 2) was circular (approx. 1.5 m in diameter) and slightly deeper (approx. 75 cmbs) than Feature 1. Narrow-Stemmed projectile points and associated Late Archaic Period chipping debris were recovered from its upper portions, but its lower portions were largely devoid of artifacts. A wood charcoal sample from the bottom of this feature yielded a radiocarbon date of 4400 ± 40 BP, while the upper portion yielded a date of 940 ± 40 BP. Curiously, both of these DSFs were slightly deeper at their northern ends, had shelf-like sections at their southern ends, and had lithic artifact distributions concentrated in their upper portions. While DSFs are clearly a southern New England feature type associated with outwash settings, researchers have not generated much dialogue regarding their interpretation and cultural significance. This has not been the case in the Middle Atlantic region of the U.S., where researchers have long debated the nature of DSFs. 3.2. DSFs in the Middle Atlantic Hundreds of DSFs have been reported on the mostly sandy and unconsolidated soils of the Delaware Coastal Plain (Egghart 2005) where local archaeologists have cultivated a more active dialogue than their New England colleagues regarding the genesis and cultural significance of these enigmatic features. This section provides a 47 brief review of archaeological investigations of key sites over the past four decades and prominent explanatory models developed by their researchers. Archaeological features interpreted to be semi-subterranean shelters were first reported in Delaware in the 1970s (Petraglia et al. 2005: 10[2]), at sites including Poplar Thicket (7S-G-22), Island Field (7K-F-17), Mispillion (7S-A-1) and Warrington (7S-G14). Additional remains of suspected pit houses were discovered at the Delaware Park Site (Thomas 1981) and Clyde Farm Site (7NC-E-6A) (Custer 1989). The proposed pit house features at all of the aforementioned sites contained some evidence interpreted to be internal hearth remnants, and post molds were frequently reported. Large numbers of subsurface anomalies at the Snapp (Custer and Silber 1994), Carey Farm (Custer et al. 1995), and Leipsic (Custer et al. 1994) sites were recognized by Custer (1994) as forming household clusters of semi-subterranean dwellings that had been truncated by historic plowing. According to his Degraded Pit House Model (sensu Egghart 2005), Custer proposed that most D-shaped pits are semi-subterranean pit house basements truncated by historic period plowing. Additionally, he recognized deepened sections of D-shaped pits as sub-basement storage pits, citing their lack of internal stratification as evidence of rapid filling prior to abandonment. Custer applied this interpretive model broadly in interring site function during the 1990s. For example, he interpreted 197 soil anomalies at the Leipsic site to be house remains, which implies that the site was subject to substantial earthmoving activities by ancient Native American residents (Custer et al. 1994). Custer‟s model has generated skepticism among some colleagues. For example, 48 LeeDecker cites the lack of postmolds and other feature types expected with house remains as evidence against the Degraded Pit House Model (Egghart 2005). Other researchers in the Middle Atlantic have proposed a tree throw hypothesis for DSF genesis (Petraglia et al. 2005). At the Hollingsworth Farm site in Elkton, Maryland, a complex of large pit features with irregularly shaped bottoms, organic-rich soils and Native American artifact content were interpreted to be tree throws (Thomas and Payne 1981). The discovery of additional D- and kidney-shaped features at the Charles Robinson Plantation in Delaware compelled Thomas to suggest that tree throws played a central role in the genesis of similar features throughout Delaware. Mueller and Cavallo‟s research (1995) supports the tree throw hypothesis, while considering the possibility that that tree throws could be used or modified by prehistoric people. Excavations at the Gabor site in Delaware identified several D-shaped pits that they interpreted to be tree throws based on the paucity of associated artifacts and lack of organic soils, internal stratigraphy, and post molds. Furthermore, they compared signatures of archaeologically confirmed pit house remains, natural tree throws, and Dshaped pits from archaeological sites. In view of this comparative approach, they concluded that the D-shaped pit forms were most likely initiated by naturally occurring tree throws, recognizing the possibility that these were utilized or modified by humans. Mueller et al. (1997) advised future researchers to assess each feature individually through detailed analysis using as many lines of inquiry as possible, and cautioned against classifying pit features according to the rigid dichotomy of pit house versus tree throw. 49 This advice was well-taken by Petraglia et al. (2005) whose research and interpretation of DSFs acknowledges that “since cultural and natural agents often overlap and interact in the context of basin formation and use, it is worthwhile to consider these archaeological manifestations as the result of complex processes” (ibid: Section 10-1). Their analysis and interpretation of pit features at Hickory Bluff in Delaware was informed by archaeological examination of extant tree throws from the site and nearby areas. Additionally, they conducted experiments designed to investigate aspects of basin formation over time. In the balance of their interpretation, they note that so many factors influence the distribution of soil and archaeological materials within pit feature contexts that it is difficult to isolate natural from anthropogenic dispositions (Petraglia et al. 2005: Section 9-29). Additionally, collaboration with local Nanticoke tribal consultants during investigations of the Hickory Bluff site alerts archaeologists to the possibility that sacred or ceremonial features, such as sweat lodges, might generate a palimpsest of pit and hearth-like features over time that constitute, resemble or contribute to the formation of DSF complexes (Petraglia and Cunningham 2006). Egghart (2005) contributed another model to explain the formation of D-shaped pits of the Middle Atlantic. His Culturally-Induced Tree Throw Model postulates that these features are anthropogenic and primarily reflect prehistoric forest management strategies. According to his model, humans began the formation process by excavating a deep hole next to the base of a tree (which constitutes the deep end of a D-shaped pit), partially exposing its root system. Next, they built a fire in the pit to burn through exposed roots, rendering the tree structurally weak. 50 It would eventually be felled naturally by wind (leaving root plate negative that constitutes the shallow end of a Dshaped pit), likely in a storm event. By selectively culling trees in this manner, prehistoric people may have achieved several goals such as increasing mast production and supplies of limb firewood, in addition to clearing areas well-suited for nascent horticulture or the promotion of edible wild plant foods. His model is compelling because it accounts for the presence of a deepened section at one end of the feature and any charcoal it may contain. 51 Chapter 4. Tree Throws: Formation, Character and Archaeological Relevance 4.1. Formation and Character of Tree Throws Because so many DSFs at Preston Plains resemble naturally occurring tree throws in terms of morphology, and sometimes even stratigraphy, tree throws merit special consideration as a hypothesis for DSF genesis. Scholars have identified a wide variety of factors that influence whether or not a tree throw will occur in a given setting, and have proposed many explanations that account for the regularity and variability in their size, morphology, stratigraphy, and longevity. This section synthesizes key points on these topics based on a review of pertinent literature, chiefly from the disciplines of forestry, ecology, and archaeology. A tree throw (aka. tree-fall, tree-tip, windfall, windthrow, uprooting) is the toppling of a tree and coincident uprooting of its root mass, which may be caused by any combination of internal or external factors. Focal vocabulary surrounding the discussion of tree throws commonly includes the following terms, the definitions of which may vary somewhat from one source to another: Ball: an uplift with a thick cross-section, supported by deep root system Blowdown: the uprooting of many trees in a single, climactically-driven event Deadfall: trees that fall after being dead or weakened Hinge: the margin of contact between the ground surface and an uplift Lee, Leeward: upwind; direction opposite from which a tree fell 52 Mound (aka. Knoll, Pillow): topographically convex tree throw remnant Pit (aka. Cradle): topographically concave tree throw remnant Uplift: a recently uprooted root mass and its adhering soil matrix Plate: an uplift with a thin cross-section, supported by shallow root system Treefall: synonym for tree throw, or the direction in which a tree has fallen Windward: downwind; direction in which a tree fell When a tree throw occurs, a pit is formed on the windward side of the trunk as soil is torn by the rotating root mass to form an uplift. Pits are rarely circular, most often appearing ovoid or D-shaped in appearance, though crescentic forms occur as well (Landhor 1993). Degradation of the uplift is contributed to by several processes including rain splash, frost heave, soil creep, faunal activity and root decay (Beatty and Stone 1986) which cause the formation of a mound. After a fallen tree has entirely decayed, topographic evidence of the tree throw persists for some time as a paired pit and mound, alternately termed cradle and knoll (Malde 1964). External factors that have been implicated in causing tree throws include wind, weight of snow or ice, the falling of adjacent trees, and human action (Langohr 1993; Petraglia et al. 2005: 10-5). Modern weather patterns in southern New England have been shown to correlate with blowdowns which tend to topple trees in one direction (Wessels 1997). The strongest winds of spring and summer are generated by easterly traveling thunderstorm microbursts that hit the ground and radiate outwards, though the majority of trees they topple point due east. Southeasterly to southerly treefalls are 53 usually the product of fall and winter arctic air masses that bring high-pressure gales. Strong cyclonic storms with counterclockwise rotations tend to throw trees in westerly directions. Northwesterly treefalls can be confidently attributed to hurricanes, while southwesterly treefalls are usually caused by hurricanes or winter northeasters. In addition to blowdowns are tornadoes, though these are uncommon occurrences in Connecticut due to the hilly terrain. Tornadoes down trees in all directions, and are prone to snapping pines at midtrunk. Deadfalls occur in any direction and can be caused by less dramatic weather conditions. It is critical to remember that different weather patterns existed in New England‟s ancient past and may not correspond with contemporary treefall patterns. Additionally, it has been argued that climactic shifts, which have occurred repeatedly during the Holocene, may increase wind speeds and, in turn, the frequency of tree throws (Gribbin 1990). The aftermath of a 2008 ice storm in western Massachusetts indicates that the weight of ice frequently causes trees to fall in succession, in both clustered and linear patterns (Ouimet et al. 2010). Local slope may have little influence on the distribution of trees uprooted by ice. Fraser (1962) argues that the depth and extent of tree rooting and the physical conditions of soil are the two most important factors determining whether or not a tree throw will occur, with exposure and elevation as secondary determinants. A tree‟s stability is based largely on how effectively it is anchored by its root system, though the mechanical strength of soil matrix is also a factor (Bennie 1996). Fritzsche (1933) views horizontally running roots as struts that support trees during wind storms, though heart roots (aka. tap roots) and small roots are also vital in tree anchoring (Day 1950). The 54 greatest source of stability is likely provided by the extensive network of tiny roots that continually spread out from the larger, more centrally located, roots (Farb 1964:94; Petraglia et al. 2005: 10-5). Relatively small increases in rooting depth can increase a tree‟s resistance to throwing significantly, and well-drained soils tend to promote deeper rooting. Healthy trees tend to cause more soil disturbance when thrown than do dead or dying trees because decaying root systems break more easily (Swanson et al. 1982). As a tree become older it is more likely to have dead or dying roots, which tend to snap during tree falls, creating more expansive but shallower pits and knolls than would a younger tree (Bubel 2003:65). Generally speaking, rooting strength declines as moisture levels increase (Rowlands 1939). Day (1950) explains that soils that tend to become saturated for extended periods of time cannot be held together by tree root systems as effectively, and are places where tree throws are more likely to occur. Constant saturation of soil also tends to encourage rotting of deep roots. However, tree throws also occur on freely draining soils that permit deep rooting, though not necessarily as often. Rotting of deeper roots can occur in large trees due to lack of water, eventually resulting in butt-rot. Some New England species, such as beech, appear to be more resistant to throwing (Day 1950), while others, such as white pine, are far more prone to blowdowns (Rowlands 1939). This is probably due, in part, to the fact that root structure varies across species (Farb 1961), but the wind-resistance of the crown is an important factor as well (Brewer and Merritt 1978). Forest heterogeny and the distance between individual trees must also be accounted for when considering the likelihood of tree throws. Stands of even age are more wind resistant than stands of mixed age which experience more 55 turbulence (Rowlands 1939; Foster and Aber 2004: 238). Trees growing in densely stocked stands tend to be less resistant to toppling than open-grown trees. Furthermore, trees growing on the edges of forests are more resistant to toppling than interior trees because they tend to develop stronger rooting and stem structures (Mergen 1954). Trees may develop an unequal distribution of root types in response to frequent, unidirectional winds, with greater lateral root growth occurring on the lee and windward sides of a tree parallel to the dominant wind direction (Stokes et al. 1995 [in Bubel:58]), though this may not equip the tree to resist high winds coming from unusual directions. Descriptions of 19th century tree stump removal techniques (Hays 1910:123-5) suggest that the size and configuration of a tree throw‟s subterranean disturbance reflects, to some degree, the species of tree involved. Poplar stumps are described as small and easy to remove while white pine stumps grow large and have very extensive, but not deeply penetrating, root systems that are difficult to dislodge. Hickories and oaks also develop large stumps and have very strong tap roots. At least three classification schemes have been proposed for tree throws. Beatty and Stone (1986) type tree throws as being hinge or rotational. Hinge tree throws are formed when a tree pivots on the edge of the uplift away from the tree base, while rotational tree throws form when the uplift is sheared along the edge. Hinge tree throws are subtyped as being complete, ≤ 90°, or > 90°, all of which are further subtyped by the size of their uplift. Rotational tree throws are subtyped as incomplete, ball and socket, and thrust. Langohr (1993) classifies tree throws according to the following aspects, or combinations thereof: 1) isolated falls, 2) serial/domino falls, 3) complete falls, 4), 56 incomplete falls, 5) rotation of uplift during fall, 6) complete uplift extraction out of pit, 7) uplift remains in pit, 8) uplift falls back into pit after being cut. Based on fieldwork conducted in 1999, Bubel developed another typology for tree throws (2003), which enlists many of the aspects defined by Langohr (1993) but translates even more expressions tree throws may assume (for a full description see Bubel 2003). Profiles of tree throw bisections have demonstrated the complex and variable nature of stratigraphies that may be anticipated (Petraglia et al. 2005). Over time, the pit produced by a tree throw becomes filled with soil from the mound. Factors affecting the rate and character of pit infilling include soil character, regional climate, faunalturbation, and ground slope (Petraglia et al. 2005). The stratigraphic character of soils (re)deposited in the pit may depend greatly on the rate of decay of the tree‟s root system (Schaetzl et al. 1990). Hardwoods tend to decay quickly (Beatty and Stone 1986), which may cause soil to fall from upturned root masses in clumps so quickly that they have no opportunity to be mixed through surface pedoturbation (Hall 1988; Troedsson and Lyford 1973) and may retain a heterogenous character. Conversely, tree roots that decay slowly, like pine, may release their soil matrix through a slumping and washing process that is so gradual that no traces of the original soil horizons are preserved. Tree throw stratigraphy may also feature redeposited gravel, rocks, and large soil clasts (Lutz 1960). While tree throw topographies lose distinctiveness over time (Denny and Goodlett 1956), they remain discernable for varying periods that probably reflect the tree throw‟s initial magnitude, local sediment characteristics, and regional differences in climate (Petraglia at al. 2005; Schaetzl and Follmer 1990). 57 Tree throw topography in tropical/humid climates have been observed to become indistinct within a decade (Putz 1983), which is not generally the case in cooler climates. For example, Schaetzl and Follmer (1990) report tree thrown topography in Michigan that is estimated to be up to 2500 radiocarbon years old, while another in the same state estimates sandy mounds to have a mean age of up to 500 years (Beatty and Stone 1986; Swanson et al. 1982). Pit and mound pairs of substantial magnitude are projected to last for hundreds of years in New England (Wesssels 1997:123) Further complicating factors shaping tree-thrown landscapes include the growth of subsequent trees in pit and mound topography. When a tree falls it creates a gap in the canopy that is usually capitalized on quickly by new growth (Wessels 1997:120). Mounds provide good germination sites for subsequent tree growth in some settings, while in others the pits provide more suitable locations. The author has noted this phenomenon in forested portions of Mashantucket‟s Sandy Hill Site, where a much of the new tree growth occurs on mounds (relics of The New England Hurricane of 1938). If some of these trees were later thrown, they could (hypothetically) contribute to the formation of expanding or overlapping complexes of tree throw modified soils. Tree throws influence the ever-changing composition of forests. Each creates an ecological microsite that affords different levels of light, water, and soil nutrients; thus, post-blowdown forest areas tend to exhibit greater environmental heterogeny (Foster and Aber 2004: 248-51). Soil formation appears to be significantly faster in pits versus mounds (Bubel 2003: 68). Studies of pit-and-mound topography in western Massachusetts and Northern Michigan both found that the accumulation of organic debris 58 accelerated pedogenesis in pits, while soils on mounds were slower to develop (Schaetzl 1990; Veneman et al. 1984). The slopes on the north and south facing portions of mounds and pits may experience significant temperature differences, and water infiltration rates may be significantly greater on mounds than undisturbed soil (Lutz 1940). 4.2. Archaeological Relevance of Tree Throws Tree throws are archaeologically relevant because they are primary agents of floralturbation (Waters 1992:307) that relocate artifacts and soil materials (Schiffer 1987:212). Tree throws are generally responsible for the mixed or interrupted horizons in forest soils (Johnson and Watson-Stegner 1990:544). Brewer and Merritt (1959) studied tree thrown topography in Michigan and estimated that the total soil surface of a forest would be disturbed by tree throws within a period of 3000 to 5000 years. In a study of tree throws in New York State, Muller and Cline (1959) described tree throws as having a long-term plowing effect on forest soils. Denny and Goodlett‟s study in Pennsylvania showed that most soils shallower than one meter are no more than a few hundred years old (1956). Massive amounts of sediment may be displaced by tree throws over time (Wood and Johnson 1978), and the process of uprooting may bring large objects to the surface (Denny and Goodlett 1956). Hypothetically speaking, larger artifacts and stones can repeatedly cycle up and down through soil profiles as a long-term consequence of existing in a tree thrown landscape (Schiffer 1987:212). Horizontal displacement of soil 59 occurs as well. Over time, tree throws can severely compromise the spatial patterning of archaeological materials at a site. For example, at the Paleoindian Period Debert Site in Nova Scotia prolonged tree throw turbation has significantly blurred artifact associations and feature boundaries in the uppermost portions of the soil column (MacDonald 1968). Tree throws are also cited as significantly compromising the spatial integrity of shallowly buried Woodland Period sites in forested regions of the Northeast (Strauss 1978). Artifacts can also be transported into much deeper contexts than they were originally deposited. This was first emphasized by Holmes (1893) who showed that tree throws transported artifacts at the Babbit Site in Minnesota to depths exceeding 1.25 m in glacial till (Waters 1992:307). A tree throw may also generate a depositional or postdepositional pattern that is predictably organized and, thus, archeologically meaningful, provided that the tree throw can be typologically identified and subsequent turbation is not dramatic. Bubel (2003) models anticipated patterns as they would appear in a tree throw‟s longitudinal crosssection and plan (2003: 113-18). These are summarized below in reference to her typology: Case 1: Artifacts are beneath the surface when the tree is thrown. Type C or I windthrow: rotational (ball/socket) tree throw causes strata to become vertically inclined in center of feature; artifacts are concentrated in center of feature, and not in loamy deposits in windward and leeward margins. 60 Type B or D windthrow: hinge (plate) tree throw causes artifacts to become broadly dispersed in a mound deposit that collapses into the feature and onto windward ground surface; loamy deposits deeply buried in leeward side and beneath the mound on the windward ground surface do not contain artifacts. Case 2: Artifacts were on the surface at the time the tree was thrown. Type C or I windthrow: rotational (ball/socket) tree throw causes artifactbearing humic deposits to become concentrated deeply in the windward and leeward ends of the feature. Type B or D windthrow: hinge (plate) tree throw causes artifacts in humic deposits to become concentrated deeply in the leeward end of the feature, and beneath the mound on the windward ground surface. Case 3: Tree throw was used by people immediately after occurring. Type B or D windthrow: Hinge (plate) tree throw exposes uplift that is used as a windbreak for a shelter built on the windward side; artifacts will be deposited only in loamy soil that becomes capped by collapsing mound. (* Bubel does not think the leeward side of a tree throw would make a suitable habitation site due to spatial limitations and assumed wind direction, so she models no such scenario) More scenarios are certainly possible, as well as multiple combinations thereof. Nonetheless, Bubel brilliantly alerts us to the possibility of translating the interplay 61 between tree throws and archaeological signatures, and her models will be referenced later in this dissertation. Tree throws are also archaeologically relevant because they may contribute to the formation of spatially distinctive deposits that appear anthropogenic (Waters 1992:307). For example, regarding the archaeological record of interior Alaska, Thorson suspects that some features interpreted to be hearths are merely the products of colluvial mixing of surface artifacts and forest-fire charcoal into tree throw pits (1990:416). If his suspicions are correct, they may, in part, account for instances of older artifact types being found in apparent association with younger charcoal. Features originally proposed to be PreClovis Period hearths on California‟s Santa Rosa Island have been reconsidered to be tree throw remnants where the burning of associated root masses generated charcoal (Wendorf 1982). Horseshoe-like features in the Netherlands, once thought to be the remains of habitation structures, have been shown to be morphologically and stratigraphically analogous to the subterranean remnants of tree throws (Kooi 1974). According to a synthetic review of archaeological data from central and northern Europe, Newell (1981) convincingly argues that most, if not all, of the Mesolithic pit dwelling structures reported in twentieth century literature are merely tree throws that disturbed pre-existing Mesolithic living surfaces (Newell 1981). However, some archaeologists have argued that tree throws are not only postdepositional agents, but topographic features that humans exploited in certain instances. Crombé (1993) reports five Final Paleolithic-to-Mesolithic sites in northern Belgium‟s East Flanders Province where tree throw features were found to contain very 62 distinct flint concentrations. Of particular interest, tree throw features from the Mesolithic Kruishoutem and Oeudeghein sites contained concentrations of flints in what would have been their shallow (pit) ends (see plans [Fig.5] and profiles [Fig.6] in Crombé 1993). This pattern demonstrates a correlation between tree throw-induced pit-andmound topography and cultural activity, though the precise use of the tree throw pits is not determined. Evans, Pollard and Knight (1999) report a similar pattern in Cambridgeshire, England where early Neolithic midden deposits were found within what once constituted the shallow (pit) ends of tree throws. This pattern was found among tree throw features on the Cam at Hinxton and at Barleycroft Farm along the River Great Ouse. The recovered assemblages appeared to be domestic in nature, containing both flints and pottery sherds, and are comparable to middens recovered from contemporary domestic structures. This pattern is argued to reflect middening that resulted, at least in part, from in situ activity. Such use of tree throw features may reflect occupations, and perhaps reoccupations, of forest clearings by early Neolithic populations of the British Isles who retained a “forest identity” after the introduction of shifting cultivation. It has also been speculated that uplifts presented convenient windbreaks for expedient tool manufacture or resource processing activities (Pyddoke 1961:85). Investigators of the Edgewater Park Site, a Late Archaic Period campsite in Iowa, speculated that a “deep pit feature” (fitting the author‟s definition of a DSF) there may have been a tree throw subsequently used as a flintknapping locale (Whittaker et al. 63 2007:31). Alternately they proposed that debitage may have been deposited at the base of a tree that uprooted, causing artifacts to become redeposited into the root ball cavity. Another important point is that tree throws can transport archaeological deposits beneath the reach of historic plowing, thus serving as preservation mechanisms. The author realized this while excavating Native American cultural deposits at the Stones Throw Site in New Hampshire (Ives 2006b). This site is located on a sandy outwash terrace where most lithic artifacts were recovered from plowed topsoil. However, at the epicenter of a small Late Paleoindian lithic scatter a probable hearth remnant was discovered beneath the plowzone in the matrix of an ancient tree throw. This feature, and its valuable radiocarbon data, would probably not have been evident if a tree throw had not fortuitously transported feature matrix beneath the plow‟s destructive reach. 64 Chapter 5. Methods of Data Collection and Processing This chapter describes the methods of data collection, laboratory processing, and synthesis that contribute to the interpretive model of DSF formation and cultural significance presented in Chapter 11. 5.1. Field Methods Data collection at the Preston Plains site involved several field methods, including the excavation of test pits, excavation units, machine trenches, and test units. Because most of the data discussed in this dissertation was collected using these methods, it is appropriate to describe them so that their strengths and limitations are apparent. Most of the archaeological testing discussed in this dissertation is depicted on Figures 5.1 and 5.2. 5.1.1. Test Pits Test pits were excavated across the original Preston Plains Energy Center Project area at 10-m interval grid nodes. When excavated at this relatively large interval, they were intended to determine the presence or absence of cultural material. Additional test pits excavated at 5-m grid nodes were used to further assess spatial distributions of prehistoric cultural material and gain a larger sample of artifacts. Test pits measured 50-cm square, were excavated by hand, and are identified by the grid coordinates of their southwest corners. These were usually terminated after 65 Figure 5-1. Map depicting all archaeological testing within the revised Preston Plains Energy Center Project Area. 66 Figure 5-2. Plan map of Machine Trench 17, located outside of the revised Preston Plains Energy Center Project Area. penetrating C-horizon soil or when the practical limits of hand excavation had been reached, which was most often at 100 cmbs. Soil was screened through a ¼ inch hardware mesh, and artifacts recovered from each 10-cm level were bagged and labeled according to their provenience. Within these levels, whenever possible, artifacts from different soil strata were collected separately. A soil profile from every test pit was measured relative to the ground surface and recorded, including details on stratigraphy, soil texture, and soil color. Screened backdirt was shoveled back into each test pit, restoring the ground surface to its original contour. 5.1.2. Excavation Units Excavation units were used primarily to gain larger samples of prehistoric cultural material and/or investigate cultural features and other soil anomalies. When a series of excavation units was excavated adjacent to one another, they were collectively referred to 67 as an excavation block. Relatively large excavation blocks were excavated at Locus 1, 2/2B, 4, and 5. Excavation units typically measured 1-m square, were excavated by hand in 10cm vertical levels, and were labeled according to the grid coordinates of their southwest corners. Within each level, each 50-cm square quadrant was excavated separately and, whenever possible, artifacts from different soil strata were collected separately. Soil was screened through a ¼-inch hardware mesh, and recovered artifacts were bagged and labeled according to their provenience. Excavation units were usually terminated after Chorizon soil had been encountered or when the practical limits of hand excavation had been reached. A soil profile from every excavation unit was drawn and recorded stratigraphy, soil texture, and soil color. 5.1.3 Machine Trenches Machine trenches (MTs) were excavated with a backhoe equipped with a 5-foot wide digging bucket, and assigned sequential designations (MT-1, MT-etc.) To comply with Occupational Safety and Health Administration Standards regarding trenching and excavation, one end of every trench had a sloped egress and most were excavated to maximum depths of approximately 4 feet. If a trench was excavated to a depth greater than 4 feet, one side was graded back to reduce the risk of a cave-in. All exposed soil profiles were trowel-scraped and digitally photographed, and at least one wall from each trench was drawn according to scale. In 2006, trench profiles were drawn in reference to line levels that were not tied into known elevations (relative to mean sea level). In 2008, 68 an arbitrary datum was established for each machine trench that was measured relative to mean sea level, so that profile drawings and archaeological sampling could be tied into known elevations. Datum levels are marked on scaled profile drawings provided in this dissertation, and their real elevations are listed in the table 5-1: Table 5-1. Elevations of Machine Trench Datums Relative to Mean Sea Level (msl). Machine Trench MT-13 MT-14 MT-15 MT-16 MT-17 Datum Elevation 40.76 meters above msl 40.46 meters above msl 39.36 meters above msl 39.36 meters above msl 40.17 meters above msl In certain cases, machine trench laterals were excavated to gain further information on feature morphology and stratigraphy immediately prior to backfilling. These consisted of short trenches (3-10 feet long) extending perpendicularly from the walls of primary machine trenches. Because these were excavated opportunistically during the process of backfilling primary trenches, their sidewalls were only cleaned and photographed. These machine trench laterals are not depicted on any testing maps included in this dissertation, and will only be referred to in instances where they have yielded critical information. 5.1.4. Test Units Test Units (TUs) are 50-cm square test pits excavated along the edge of machine trenches dug in 2008, but unlike regular test pits all soil from the TUs was collected. These were excavated in judgmentally selected locations to investigate features and 69 collect control data from natural soil profiles. Each was excavated in 10 cm arbitrary levels measured relative to its associated machine trench datum. Strata collected from strata within each level were collected separately when possible. Five soil samples were collected from each level of a TU: Pollen Sample: 400 ml. (approx ½ of a 5-x-8 inch bag) Phytolith Sample: 400 ml. (approx ½ of a 5-x-8 inch bag) 1-liter Soil Samples: Two of these were taken from each level, for anticipated analytical purposes. One of each was subsequently used for soil water retention tests. Flotation Sample: All remaining soil after the removal of the previous samples was floated in one sample. 5.2. Laboratory Methods and Data Processing Several laboratory and data processing methods were used to translate the recovered artifacts and soil samples from Preston Plains into archaeologically meaningful data. This translation included washing and cataloging artifacts, floating soil samples, measuring soil water capacity, and producing analytical graphics. The long-term curation of the recovered artifact assemblage is also described. 5.2.1. Cleaning and Cataloging Artifacts Artifacts collected in the field were organized according to provenience and transported to the MPMRC Laboratory. Physically and chemically stable items, such as 70 lithics, were subsequently cleaned by soaking in tap water and brushing as necessary to remove any adhering soil clasts. Friable items, such as charred wood, were dried and lightly brushed. All materials were cataloged into the MPMRC Database - a custom computer program designed using Microsoft Office Access 2002/3 database software. Materials that displayed similar attributes such as materials type, functional and typologial classes, size range, etc. were grouped and cataloged by lots. These lots are stored in polyethylene resealable bags containing provenience information. These bags, in turn, are organized within archival-quality cardboard trays that are stacked in boxes for long-term storage. Several artifact analysts invested significant labor in this project. Botanical specimens (seeds, nuts) and wood from feature contexts at Locus I were identified by Tonya Largy and David DeMello while Randy Nokes identified calcined bone fragments from Locus I feature contexts as well as TU contexts. I provided final identifications of lithic tools from Locus I, following up on initial identifications made by Brian Jones. David Perry provided soft plant tissue identifications for Features 8 and 12 from Locus 1. 5.2.2. Soil Flotation Soil samples were processed through flotation – a water-sorting method – to extract any microscopic cultural material contained therein. This process removes silt and clay from a submerged soil sample through agitation, which yields two “fractions” of material for collection. The light fraction consists of buoyant materials, chiefly organics, while the heavy fraction consists of materials that sink, chiefly inorganics. Most of these 71 fractions were sorted under low-power magnification by staff of the MPMRC for preliminary identification and cataloging. Some light fractions were set aside for Tonya Largy and David DeMello (consulting botanical analysts) to sort as they provided final identifications of botanical materials from Locus I. During the winter season soil samples were floted using the Model A Flote-Tech Machine in the Laboratory of the MPMRC. During the warmer months flotation was performed outside at a reservoir behind the headquarters of the Mashantucket Pequot Tribe‟s Department of Natural Resources. 5.2.3. Soil Water Capacity A laboratory test was devised to estimate water capacities of representative soil types. It did not determine “field capacity,” which can only be measured from soil that is in situ; rather, it merely measured the percentage of weight gained by an air-dried soil sample immediately after being saturated. For the purposes of discussion, this value is referred to as soil water capacity (hereafter SWC). SWC samples were culled from the 1-liter soil samples collected during TU excavation and spread out on aluminum foil to dry for over 30 days. A device was assembled that would hold a dried-and-weighed soil sample (of approximately 200 ml each) while water (400 ml) was poured through it without significant loss of soil material. This device was made from a plastic, Penzoil 1-Pint funnel. The mouth of this conical form measured 12.9 cm while the bottom was trimmed with a saw to leave a 2.4 cm circular drain hole. Across this hole, a very fine screen mesh (355 micrometers) was 72 glued that allowed water to drain from the soil without significant loss of silt or clay. The mechanism worked surprisingly well. Each sample was allowed to drain for 30 minutes (by which time every sample had ceased to drip) before its weight gain was measured. 5.2.4. Radiocarbon Dating All radiocarbon samples from the Preston Plains Site were processed by Beta Analytic in Miami, Florida. Several samples were dated using the radiometric technique, while others were dated using accelerator mass spectronomy (AMS). Each radiocarbon date from the Preston Plains Site presented in this dissertation is expressed in conventional (ie. not calibrated to calendar years) radiocarbon years and is followed by its unique Beta Analytic sample identification number. All radiocarbon dates from the Preston Plains Site are listed in Table 5-2. 5.2.5. Long-term Storage and Access The MPMRC has a permanent on-site collections storage facility that is climate controlled. Temperature ranges from 68-72 degrees, with a typical 24 hour variation that is no more than 5 degrees. Humidity ranges from 50-60% with a typical 24 hour variation of no more than 3%. Archaeological materials are stored in archival quality compactor shelving. The collections storage facility is monitored 24 hours day by security personnel, and the greater building (the MPMRC) is protected by fire, smoke detection, and alarm systems. Researchers interested in studying materials from the Preston Plains Site may contact the Laboratory Director of the MPMRC, who can 73 Table 5-2. All Radiocarbon Dates from the Preston Plains Site. Beta Analytic Sample # 130446 131054 219137 219138 219139 219140 222824 222825 222826 226147 227750 227751 131943 132294 132293 136708 Conventional 14C date BP 4230 ± 60 5290 ± 80 4120 ± 60 4270 ± 50 4710 ± 50 4070 ± 50 5030 ± 50 7620 ± 60 4090 ± 40 2230 ± 40 4550 ± 40 4130 ± 40 4650 ± 70 4490 ± 70 4310 ± 70 3990 ± 80 Locus Feature 8 12 18 19 Material wood charcoal wood charcoal wood charcoal wood charcoal wood charcoal wood charcoal carb. hickory (carya sp.) wood charcoal wood charcoal wood charcoal wood charcoal wood charcoal wood charcoal wood charcoal wood charcoal wood charcoal Technique radiometric radiometric radiometric AMS AMS AMS AMS AMS radiometric AMS AMS AMS radiometric radiometric radiometric radiometric 1 20 21 22 24 26 215 216 217 1 4 2 6 not assigned 2 4 5 temporarily transfer requested materials from the collections storage facility to the laboratory for examination. 5.2.6. Graphics Production Final versions of site maps and profile walls were rendered using Adobe Photoshop CS4 Extended image editing software. Hand-drawings were initially produced in the field on 1/10 inch graph paper, and scanned copies of these were imported into the image editing software. Each scanned copy served as the background layer for one graphic, which was generated by adding layers. Each additional layer depicted a key element, such as stratigraphic boundaries, provenience references, and soil 74 strata labels. The final versions of graphics were yielded simply by dropping the background layer and exporting the remaining collection of layers as a single PDF file. Mandy Ranslow produced all maps that depict lithic distributions. She started by importing the Preston Plains artifact catalog into Microsoft Office Excel 2003 spreadsheet software and eliminating all material other than lithic tools and debitage with a maximum dimension of at least 1-cm in size. Lithic artifacts with a maximum dimension less than 1-cm in size (such as “shatter” and “microflakes”) were excluded from these analyses because they were not collected as consistently, having been effectively recovered in flotation samples but frequently lost through soil screening in the field. Next, she filtered this data to produce a series of series of spreadsheet files, with each one containing data from a specific set of contexts and/or artifact categories for analysis. She, in turn, imported each of these files into Surfer8 contouring and surface plotting software, which she used to generate graphic representations of lithic distributions. Lastly, she took each of these representations and superimposed them on background maps furnished by the author. 75 Chapter 6. Setting and Geomorphology of Preston Plains 6.1. Modern Site Setting Preston Plains is a vernacular place name referring to the flat openlands near the intersection of Ledyard, Preston, and North Stonington‟s town boundaries in southeastern Connecticut. This area generally lies within a physiographic transition from Connecticut‟s Eastern Uplands to Coastal Lowlands (Connecticut Development Commission 1963). Preston Plains occupies a north-south oriented valley and are used primarily for corn and sod farming today (Figure 6-1). Though the terrain here is relatively level compared to the surrounding uplands, the topography is subtly distinguished by low-relief outwash plateau and bar-like morphologies with meandering borders. Its surface gently plunges southward at an average 0.5% downslope, eventually becoming submerged beneath the forested mire of Mashantucket‟s Great Cedar Swamp. Avery Pond (a glacial kettle) and a complex of forested kame formations bound Preston Plains on the north while forested uplands lie to the east and west, and are variably developed for residential and commercial use. Preston Plains is drained primarily by Indian Town Brook, which also gathers water from nearby reservoirs including Avery Pond, Lake of Isles, and the Great Cedar Swamp. This brook begins its clockwise course as an eastern outlet of Avery Pond. From there, it meanders southward along the eastern edge of Preston Plains, and veers west when it reaches Mashantucket‟s Great Cedar Swamp. Indian Town Brook joins Shewville Brook which flows west and eventually drains into Poquetanuck Cove, the Thames River, and lastly Long Island Sound. 76 Figure 6-1. View south of Avery Pond (foreground), Preston Plains openlands (farm tracts) and the Great Cedar Swamp (undeveloped forest, center-background), all elements of the same valley. The USDA assigns most soil at Preston Plains to the Hinckley-MerrimacAgawam soil unit, with Hinckley and Agawam soil types predominating. Hinckley soils are deep and excessively drained, and exhibit rapid permeability. They usually occur on kames, eskers, and outwash plains and terraces. Agawam soils are deep, well drained, and loamy, and exhibit rapid permeability in their substrata. They usually occur on outwash plains and stream terraces. Both of these soil types are suited to farming and woodlands. Archaeologically investigated portions of the Preston Plains Site occupy an area of Agawam fine sandy loam on 3 to 8 percent slopes (Crouch 1983). This gently sloping, well drained soil is medium to strongly acid, and warms up and dries rapidly in the spring. 77 Archaeologically investigated portions of the Preston Plains site are located at 451 and 455 Norwich Westerly Road (CT Route 2), both fee lands owned by the MPTN. A 1920‟s Craftsman style house and associated outbuildings stood on a portion of the 451 parcel until they were demolished in 2006 (Figure 6-2). This house formerly served as the headquarters of the MPTN Department of Natural Resources, and a manicured lawn was maintained here until 2006. A local farmer continues to cultivate much of the 455 parcel, planting corn and rye grass in recent years. Figure 6-2. View southwest of house and garage at 455 Norwich Westerly Road in 2006. 6.2. Geomorphology Preston Plains formed through a series of depositional and erosional events, the most dramatic of which occurred during the late Pleistocene. The southward grind of the Laurentide Ice Sheet carved the bedrock valley that contains Preston Plains. At, or shortly after, 17,000 BP, this glacier retreated from the area, leaving a proglacial lake that 78 was initially ice-dammed and caused the deposition of glaciolacustrine mud in its southern end (Thorson and Webb 1991:18). The clearing of a natural outlet to the west probably lowered the level of this lake, though impoundment by the Ledyard Moraine on the south prevented its complete drainage. Glaciofluvial deposition followed, partially filling this basin to establish the surface recognized as Preston Plains. Landforms such as this are valley sandurs, which form “when meltwater is forced to follow a valley” (Krigström 1962: 328). On average, Preston Plains outwash facies prograde to the south slightly, as does the overall topography here. Elevations at the Preston Plains Site range from approximately 120-130 feet above mean sea level. Clast-supported gravel and bar accretion structures are abundant, and the way in which sands are plane bedded and crossbedded suggests that a slowly rising or stationary lake level existed immediately downgradient from the Preston Plains Site during glaciofluvial deposition. In an attempt to determine whether or not glaciolacustrine mud is buried beneath the glaciofluvial deposits in this portion of the valley, a test hole was mechanically excavated at the southern end of MT-15. The results were inconclusive. The test hole deeply penetrated glaciofluvial deposits to a depth of 4 meters before it was terminated because the machine could not reach any deeper. Portions of a remnant glaciofluvial channel were recognized during subsurface investigations. It probably originated north of the Preston Plains Energy Center project area, beyond the limits of archaeological testing. Its trajectory within the project area can be estimated by tracing a line from north-to-south through the following grid points: N130 E30, N110 E35, N90 E40, N70 E40. South of N70 E40, the channel becomes 79 shallower and transitions into a broader fan that deposited medium-to-fine grained sands, as encountered in MT-12, 15, and 16. Deep sections of this channel cross-cut by MT-13 and 14 (Figures 6-3 and 6-4) show lower fill dominated by clast-supported material, which is probably a combination of channel lag and initial erosion of channel walls. The upper/central fill is dominated by matrix-supported material that appears eolian in origin. Stormbursts and frozen-ground (ie. periglacial) runoff likely reworked eolian fines from the surrounding terrain into the remnant channel, mixing them into matrix-supported gravel before depositing them on top. This channel would have attracted the first vegetation at Preston Plains because its concentration of fines retains more water than the surrounding glaciofluvial deposits. The arrival of pioneer plants (e.g. alder, cottonwood, dwarf birch, grass, heath, sedge, willow) further promoted entrapment of eolian fines. Figure 6-3. Remnant glaciofluvial channel in south wall of MT-13. The darkest stratum (trowel-incised) is a buried A-Horizon. 80 Figure 6-4. Remnant glaciofluvial channel in south wall of MT-14. The darkest stratum (trowel-incised) is a buried A-Horizon. The rate of infilling of this remnant channel would have been greatest early on, and expressed the advantage that sedimentation rates held over compaction/subsidence. Interestingly, buried A and B-horizon soils were recorded beneath plowed topsoil in the central portions of this channel (see Figures 6-3 and 6-4). Historic plowing across the glaciofluvial plateaus of Preston Plains truncated what must have been very thinly developed natural A and B-horizon soils. However, pedogenesis was far greater in the remnant channel, resulting in more thickly developed soils, most of which escaped the reach of the plow. Interestingly, this channel contains the greatest concentration of DSFs found at Preston Plains. A longitudinal cross section of this channel exposed in MT-9 revealed that its stratigraphy is almost continuously interrupted by overlapping DSFs (Figure 6-5). This assessment deliberately avoids presenting the interpretation that DSFs were initiated by tree throws so that it may more objectively inform such an interpretation later in this dissertation. Lengthy review of primary data from DSFs is presented in the 81 following chapters (particularly Chapters 8, and 9), addressing their distribution and character, in addition to their spatial relationships to archaeological deposits. Figure 6-5. West wall of MT-9 (view south) showing continuous series of DSFs (darker basinlike forms) along edge of remnant glaciofluvial channel. Preston Plains also appears to have been modified by earthquakes induced by isostatic uploading. Seismicity may have had several effects, one of which was the production of tension fractures in glaciofluvial sediments. Some of these episodes occurred during glaciofluvial deposition, as indicated by a tension fracture in a sandy outwash that is capped by a coarse-grained layer without such a tension fracture (Figure 6-6). 82 Figure 6-6. Tension fracture in sandy outwash (lower right) beneath a layer of gravely outwash in south wall of MT-13 (10-cm increment scale rod). Depressions within glaciofluvial strata (recorded in MT-9, 12, 15, 16) filled with homogenous, fine-grained material with no evidence of bedding were discovered (Figures 6-7 and 6-8). These basins (especially the one depicted in Figure 6-8) exhibit morphology typical of ice wedge casts. The fine-grained fill they contain is typical of yedoma – a loess deposit that can occur in massive beds and is commonly associated with permafrost environments (Tomirdiaro 1978; Walter et al. 2007). The soil material tentatively identified as yedoma was only observed in this area of Preston Plains, which has a slightly lower surface than surrounding areas. This area is also at the southern end of the remnant glaciofluvial channel where it transitions into a fan, so the accumulation of fines here is not surprising. 83 Figure 6-7. Yedoma-filled ice wedge cast in west wall of MT-15 (white tube markers at 2-m intervals). Figure 6-8. Yedoma-filled ice wedge cast in east wall of MT-16 (white tube markers at 2-m intervals; 10-cm increment scale rod). Note darker-colored krotovinas in the yedoma. The establishment of plant communities across the periglacial landscape of Preston Plains facilitated the accumulation of an eolian mantle, the remains of which is blended into plowed topsoils that are typically more fine-grained than their substrates. Deflation of soils near plateau edges and inflation of soils immediately downslope 84 obscured the angular postglacial topography throughout the Holocene, a process that was likely augmented by historic agricultural activities. Excavations at Locus I confirmed this process - the vast majority of soil profiles from the excavation block (located on a plateau edge in the N45 W30 vicinity) transitioned from a shallow, plowed A-Horizon to a C-Horizon of glaciofluvial gravel at 30 cmbs. Thicker topsoil deposition was recorded immediately below this plateau (in the vicinity of N34 W40), with stratigraphy featuring a deeper plowed A-Horizon (0-40 cmbs) over a buried A1-Horizon (40-50 cmbs). Many other natural processes modified Preston Pains during the Holocene. It may be assumed that a multitude of faunaturbative agents were involved, but reviewing all possibilities is neither practical nor necessary for the purposes of this dissertation. For brevity, a few interesting examples will be shared that lend character to Preston Plains. In the spring, painted turtles (Chrysemys picta) dig nests in the topsoil here to depths of about 10 centimeters (Figure 6-9). The sandy soils of Preston Plains are ideal for turtle nests, and most hatchlings appear to be destined for the shore of Avery Pond (Figure 610). While a single nesting turtle disturbs little more than a handful of soil, the cumulative actions of thousands of generations of turtles here must have resulted in lateral topsoil displacement, especially near the pond edge. In June 2008, the author noted abandoned turtle nests, recently vacated by hatchlings, along the northern edge of Preston Plains at average intervals of approximately 20 m. 85 Figure 6-9. Painted turtle laying eggs near Locus 4 of Preston Plains, 2008. Figure 6-10. Painted turtle hatchling found in MT-17, 2008. Woodchucks (Marmota monax) caused particularly deep soil disturbances at Preston Plains. While suspected burrow chambers were occasionally recorded in DSFs and the remnant glaciofluvial channel, one massive woodchuck burrow complex was documented that demonstrates the substantial earthmoving potential of these animals. In the west wall of MT-11, an active burrow chamber (complete with bedding), was revealed amidst a stratified palimpsest of collapsed chambers that extended to a depth of 86 270 cmbs (Figure 6-11). This complex appears to have been formed within, and obscured the edge of, the eastern margin of the remnant glaciofluvial channel. One of the distinguishing features of this complex was the tendency for some of its strata, suspected burrow chamber remnants, to undercut bedded glaciofluvial sands. The impact of large rodents on the archaeological record should not be underestimated. Because woodchucks transport excavated soil to the ground surface, it may be reasonably assumed that as abandoned chambers fill with overlying sediment they may contribute to the development of deeper A-and B-Horizon soils. The presence of a thin buried A-Horizon (see Figure 611) beneath the plowed A-Horizon in MT-11 may reflect such a process. Burrowing by pocket gophers on archaeological sites in California has been shown to size-sort artifacts in soil columns, with larger items concentrated on the bottom (Boeck 1986; Johnson 1989), while the displacement of artifacts by armadillos is not so systematic (Araujo and Marcelino 2003). At the least, it may be inferred that where woodchuck borrowing has been documented, vertical displacement of artifacts and mixing of soil strata occurred. Figure 6-11. Woodchuck burrow complex in west wall of MT-11. Note the presence of a thin buried A-Horizon, visible as a thin, dark line immediately beneath the plowed A-Horizon. 87 Additionally, smaller rodents and other burrowing animals (vertebrate and invertebrate) relocated soil material over time. At Locus I, a burrow complex suspected to have been generated by small rodents was detected in glaciofluvial sands immediately underlying DSF-5 (Figure 6-12). While root casts may leave similar branching patterns (Schiffer 1987: 210), this complex featured expanded chamber sections that are more typical of rodent activity. Figure 6-12. Plan view of probable small rodent burrow complex in glaciofluvial sands beneath DSF-5, Locus I (trowel points north). Floraturbation also modified soils at Preston Plains throughout the Holocene. Topsoils were continually penetrated by root systems, though the character and abundance of vegetal cover varied over time. Field investigations revealed that virtually all contemporary root systems avoid penetrating glacially bedded sand and gravel, while 88 most other soil types contain roots when tree or bushes are nearby (Figure 6-13). Roots turbate soil and may displace artifacts in any direction as they enlarge in diameter, and after they die they leave casts that may fill with redeposited soil and artifacts (Waters 1992: 309). Soil that contains tree roots may also become displaced when trees sway in the wind. The most significant mechanism of Holocene soil displacement at Preston Plains is probably tree throws, a topic that will be more thoroughly explored in Chapter 10. Figure 6-13. North wall of MT-17 (view west). Note root penetration through topsoil and feature soil (Fe.323), and absence of roots in C-Horizon sand (lightest soil at bottom). A particularly striking aspect of Preston Plains geomorphology is the variable rates at which different soil types dried when exposed to air. Understanding that the ability to retain water dramatically affects the degree of bioturbation in a given soil, soil water capacity (hereafter SWC) was measured for several representative soil types. The 89 methods for measuring SWC are presented in Chapter 4. Sixteen soil samples were chosen to represent the diversity of soil types encountered at Preston Plains, and the results are presented below in Table 6-1. To summarize some of these results, samples of glaciofluvial sand (Sample 2) and gravel (Sample 1) yielded relatively low SWC values of approximately 13-17% and 32% respectively. Samples of yedoma (Sample 3) yielded dramatically higher values (55-57%), as did plowed topsoil (Samples 3, 9, 11, and 15) (39-46%) and most matrix-supported feature fill from the glaciofluvial channel (Samples 8, 10, and 14) (39-62%). The implications of these results inform the model of DSF formation presented in Chapter 11. In sum, this geomorphological assessment demonstrates that beneath the seemingly mundane surface of Preston Plains lies a complex of soils that vary widely in their formation histories and physical character. This assessment provides an informed basis of landscape formation that contextualizes the interpretation of Preston Plains DSFs in their formation and relationship to cultural deposits. It also allows us to generate two critical expectations regarding the disposition of archaeological material across the site – expectations that are confirmed by data in the following chapters: 1. Artifacts deposited on glacioflucial plateaus should have a tendency to remain at shallow depths, unless turbated through a secondary process. 2. Soil structures with high water capacities should contain artifacts at greater depths than in surrounding areas. Once an area of subsoil is disturbed, through natural or cultural agency, its water capacity should be increased by the introduction of fine-grained 90 fractions from the surface. This, in turn, should attract subsequent turbation and mixing of surface artifacts to greater depths. Table 6-1. Soil water capacity measurements for representative Preston Plains soil types. Sample Weight (grams) Dry 324.1 326.9 317.6 323.1 208.9 204.4 243.8 236.3 228.6 217.3 222.9 214.6 359.0 333.0 230.5 263.4 245.2 231.9 258.5 252.5 262.8 278.7 296.8 300.3 219.0 217.3 197.6 193.0 268.4 269.5 263.3 263.2 Wet 380.4 370.8 421.2 432.2 325.4 322.6 357.4 345.1 360.8 343.3 350.9 340.8 431.8 409.1 337.1 372.0 358.5 341.4 360.0 362.0 362.6 382.6 393.8 399.7 337.4 334.0 321.1 313.0 375.6 378.0 380.8 378.5 Soil Water Cap. (% wt. gain) 17.4 13.4 32.6 33.7 55.7 57.8 46.6 46.0 57.8 58.0 57.4 58.8 20.3 22.9 46.2 41.2 46.2 47.2 39.3 43.4 38.0 37.4 32.7 33.1 54.1 53.7 62.5 62.2 39.9 40.3 44.6 43.8 Sample # Soil Type Provenience Color and Texture Test Run 1 2 1 2 1 2 1 2 1 2 1 2 1 2 1 2 1 2 1 2 1 2 1 2 1 2 1 2 1 2 1 2 1 2 3 4 5 6 Glaciofluvial Gravel (C-Horizon) Glaciofluvial Sand (C-Horizon) Yedoma (C-Horizon) Plowed A-Horizon (over Fe. 307) Buried A-Horizon (Fe.307) Buried B-Horizon (Fe.307) Clast-Supported Feature Fill (Fe.307) Matrix-Supported Feature Fill (Fe.307) Plowed A-Horizon (over Fe.306) Matrix-Supported Feature Fill (Fe.306) Plowed A-Horizon (over Fe.319) Matrix-Supported Feature Fill (Fe.319) Plowed A-Horizon (over Fe.304) Matrix-Supported Feature Fill (Fe.304) Plowed A-Horizon (over Fe.316) Matrix-Supported Feature Fill (Fe.316) MT-14, TU-4, 110-120 cmbd MT-17, TU-15, 129-140 cmbd MT-15, TU-1, 90-100 cmbd MT-13, TU-11, 60-70 cmbd MT-13, TU-11, 75-85 cmbd MT-13, TU-11, 85-90 cmbd MT-13, TU-11, 120-130 cmbd MT-13, TU-11, 150-160 cmbd MT-13, TU-9, 20-30 cmbd MT-13, TU-9, 60-70 cmbd MT-17, TU-14, 100-110 cmbd MT-17, TU-14, 150-160 cmbd MT-14, TU-5, 60-70 cmbd MT-14, TU-5, 110-120 cmbd MT-17, TU-17, 130-140 cmbd MT-17, TU-17, 170-180 cmbd 7 8 9 10 11 12 13 14 15 16 10YR 5/6 YW BN gravel and coarse sand, bedded 2.5Y 6/4 light YW BN fine sand, bedded 2.5Y 5/4 light OL BNclayey silt 10YR 3/2 very dark GY BN fine sandy loam with trace of gravel 10YR 2/2 very dark BN fine sandy loam 10YR 4/6 dark YW BN fine sandy loam with trace of gravel 10YR 4/4 dark YW BN gravel with loamy fine-tocoarse sand 10YR 4/4 dark YW BN loamy fine sand with gravel 10YR 3/2 very dark GY BN fine sandy loam with trace of gravel 10YR 4/6 dark YW BN loamy fine sand with gravel 10YR 3/3 dark BN fine sandy loam with trace of gravel 7.5YR 4/6 strong BN loamy medium fine sand with trace of gravel 10YR 3/2 very dark GY BN fine to medium sandy loam with trace of gravel 10YR 3/3 dark BN fine sandy loam with trace of gravel 10YR 3/3 dark BN fine sandy loam with trace of gravel 10YR 4/6 dark YW BN loamy fine sand with trace of gravel 91 Chapter 7. Overview of Preston Plains Archaeology The majority of this chapter presents a general overview of Preston Plains archaeology that is based largely on diagnostic point types, feature concentrations, and radiocarbon data. Because a synthesis of all data collected from this site has not been produced, this overview provides the reader with a sense of what chapters of cultural history are represented here, at least according to archaeologically robust evidence. The last section of this chapter describes the locations of DSFs at the site and presents some generalizing observations on their character and distribution. More detailed analysis and interpretation of these features will be presented in following chapters. Table 7-1 lists all diagnostic points (total=166) recovered from the Preston Plains Site to date. The most frequently recovered diagnostic points consist of types attributed to the Late Holocene climactic period, and more specifically to the Late Archaic Period of culture history (84%). Not reported in Table 7-1 are 75 untyped lithic points, which account for 31% of all recovered specimens. Most of these are stemmed and triangular forms consistent with Late Archaic Period templates; thus, most untyped points probably date to that period. As of 2003, Brian Jones had defined four primary concentrations of Late Archaic Period deposits in and near the original Preston Plains Energy Center project area, Locus 1, 2, 4, and 5 (Figure 7-1). The abundance of lithic materials and the presence of hearth deposits in these locales distinguished them as the most archaeologically significant portions of the site according to their potential to contribute new information to regional 92 Table 7-1. Temporally diagnostic lithic points recovered from the Preston Plains Site. Period Younger Dryas Early Holocene Middle Holocene Span (years BP) 11,000-10,000 10,000-8000 8000-5000 Point Type Barnes Bifurcate (MacCorckle) Bifurcate (Kanawha-like) San Patrice Neville Bare Island Beekman Triangle Material chert rhyolite quartz chert chert quartzite quartzite quartzite argillite chalcedony hornfels quartz quartzite rhyolite quartz argillite quartz quartzite chert argillite quartz quartzite argillite quartz quartzite quartz argillite quartzite unid. argillite quartz quartzite chert argillite rhyolite rhyolite argillite quartz quartzite Count 1 1 1 1 2 3 2 4 6 1 1 3 18 1 2 3 3 5 1 1 14 25 2 5 3 2 1 17 1 3 22 2 2 1 1 1 1 2 1 Total 1 3 5 Brewerton (all variants) Lamoka Narrow-Stemmed (untyped) 5000-3000 Late Archaic Normanskill Otter Creek Squibnocket Triangle Late Holocene Squibnocket Stemmed Sylvan Side-notched Vosburg 148 Wading River Broadspear (untyped) Snook Kill Wayland Notched Meadowood Rossville 3600-2800 Terminal Archaic 3000-1600 Early Woodland 4 5 archaeology. Of particular interest, clusters of DSFs were discovered at Locus 1, 2, and 4. These were originally interpreted to be the remnants of pit houses occupied by Late Archaic Period foragers (Jones 2002:22). 93 Figure 7-1. Major Late Archaic Period site loci defined by Jones at Preston Plains (map produced by Jones 2003). The original Preston Plains Energy Center project area boundary is shown as an irregular red line. Green contours represent topography, while grey contours represent lithic artifact concentrations. Supplemental test pit excavation directed by the author within the revised (smaller) project area affords a slightly more refined distribution of lithic materials (Figure 7-2). On the whole, this distribution reflects the accumulation of Late Archaic Period activity areas across the landscape, though it likely also reflects other, less robust components. 94 The following sections provide brief descriptions of prehistoric occupation at Preston Plains. They are organized broadly by climactic period, and more specifically by cultural historical periods within each climactic period. Figure 7-2. Distribution of lithic artifacts (≥1 cm) from 50-cm square test pits from the revised Preston Plains Energy Center Project Area. Updated as of 2009. (Surfer overlay by Mandy Ranslow). 95 7.1. Younger Dryas Period Occupation Younger Dryas Period occupation at Preston Plains is indicated by the recovery of one artifact, a fluted point of the Barnes type (Figure 7-3). Barnes points are attributed to the Middle Neponset Phase (ca. 10,300 BP) of New England‟s Paleoindian Period (Speiss et al. 1999), and their distinguishing features include long channel flakes, flared basal ears and a remnant fluting nipple. The recovered specimen is a complete tool though it is short compared to others reported from New England. It appears to represent a heavily reworked point or the recycled base of a point that snapped during use. It is made of reddish-brown Munsungen chert, a material from northern Maine. Figure 7-3. Barnes point (chert) recovered from Preston Plains (1-cm scale bar increments). This point was recovered at Locus 5, from the southwest quadrant of N170 W8, in B1-subsoil. No associated Paleoindian artifacts or features were identified in the vicinity. 96 The recovery of this isolated point indicates that Preston Plains was traversable by ca. 10,300 BP and hosted at least one episode of Paleoindian hunting activity. It may be associated with a short-term Paleoindian encampment on a portion of Mashantucket‟s Sandy Hill Site, where a graver of identical material was recovered in 2009. 7.2. Early Holocene Occupation (ca. 10,000-8000 BP) Early Holocene occupation of Preston Plains is indicated by the recovery of three points associated with the Piedmont Tradition, the precursor to the Atlantic Slope Macrotradition. The first closely fits the San-Patrice type (Figure 7-4), which is distinguished, in part, by the bifacial or unifacial removal of short flute-like flakes from the base. It is made of dark grey chert likely derived from New York State, possibly from an outcropping of the Normanskill Formation in the Hudson River Valley drainage, and closely resembles a number of other specimens recovered from the Lower Hudson (cf. Figure 4 in Brennan 1977). This specimen was recovered from the Locus I excavation block, specifically from the northeast quadrant of N50 W37 from within DSF matrix (110-120 cmbs), where the boundaries of DSF-1 and DSF-2 overlapped. No associated Early Holocene features were identified at Locus I. While there was a relatively thin distribution of chert debitage at Locus I, also likely derived from a source(s) in New York State, part or all of this material may be associated with the Late Archaic components. 97 Figure 7-4. Early Holocene points recovered from Preston Plains Site: San Patrice on left (chert); Kanawha-like bifurcate on right (quartz). Two bifurcate-base points were recovered as well, one of which is an unprovenienced surface find made of rhyolite. It closely resembles the MacCorckle type, which is commonly attributed to the Early Holocene and distinguished by its relatively large size and wide-eared base. This point‟s lack of context precludes any recognition of closely associated Early Holocene deposits. A quartz bifurcate of exceptionally crude workmanship was recovered from the Locus I excavation block (see Figure 7-4), specifically from the northwest quadrant of N51 W40 in DSF-3 matrix (40-50 cmbs). It resembles the Kanawha type, which is commonly attributed to the Early Holocene and distinguished by a narrower, stem-like bifurcated base. While this point was recovered less than 5-meters from the San Patrice point, an association cannot be confidently determined. San-Patrice points are generally 98 thought to predate bifurcate base points in the Northeast, though their concurrent use cannot be ruled out. In sum, occupation of Preston Plains by early Holocene foragers was relatively sparse if determined by the paucity of associated diagnostic points recovered. associated features or artifact assemblages were recognized. No 7.3. Middle Holocene Occupation Middle Holocene occupation of Preston Plains is indicated by the recovery of five Neville-like points attributed to the Atlantic Slope Macrotradition. These were all recovered from disparate locations. One quartzite specimen was recovered at Locus 5, from plowed A-horizon soil (0-20 cmbs) in the southwest quadrant of N171 W8. The second quartzite specimen is an unprovenienced surface find. The third quartzite specimen was recovered at Locus 2, from DSF matrix (103-115 cmbs) in the northwest quadrant of N130 E36. One chert specimen was recovered from the northwest quadrant of N102 E10, from unidentified feature soil (70-75 cmbs). There was no notable concentration of prehistoric artifacts in this vicinity, and the nature of the feature soil was never determined. The last point was recovered from the Locus I excavation block, where there appears to be a Middle Archaic site component. This chert specimen was recovered from DSF-1 matrix (70-80 cmbs) in the northeast quadrant of N49 W39. It was found approximately 4 meters from a charred wood fragment that yielded a Middle Holocene 14 C date of 7620 ± 60 BP (Beta-222825). Because this wood appeared to be in a 99 secondary/redeposited context (Fe.1.24), it cannot be definitively associated with the Middle Archaic component and may have been generated by a natural burn. In sum, occupation of Preston Plains by Middle Holocene foragers was relatively sparse if determined by the paucity of associated diagnostic points recovered. Three of the five recovered specimens were found within areas dominated by Late Archaic Period assemblages, and may represent early components of areas that were persistently used from the Middle to Late Holocene periods. The quartzite used to manufacture three of these was probably derived from Plainfield Formation outcrops in eastern Connecticut, reflecting exploitation of local lithic sources by bands whose territories had probably become more constricted than those associated with the preceding Piedmont Tradition. This finding is consistent with the increasingly localized mobility and resource exploitation patterns associated with the greater Atlantic Slope Macrotradition. 7.4. Late Holocene Occupation (post-5000 BP) As previously noted, an overwhelming majority of diagnostic points (89%) recovered from Preston Plains are types associated with the Late Archaic Period (ca. 5000-3000 BP). Vosburg and Brewerton points, the most frequently recovered Laurentian Tradition types, reflect a strong preference for local Plainfield Formation quartzite. Point types associated with the Narrow Stemmed Tradition, such as the Squibnocket Stemmed, Squibnocket Triangle, and Wading River points, are frequently made of local quartz, though quartzite was still used. The third most frequently used 100 material by Late Archaic knappers was argillite, probably derived from sources in the Narragansett Bay region. Though peripheral to the goals of this dissertation, one phenomenon is reported that may be of great interest to other New England archaeologists. Regarding the identification of triangular Late Archaic points from Preston Plains, the author found that using Brewerton and Squibnocket Triangle type classifications was problematic. Several specimens seemingly constitute intergrades between the two types, and specimens can be ordered in a way that illustrates a gradation between Brewerton and Squibnocket Triangle types (Figure 7-5) (Ives 2007d). So while Ritchie‟s seminal research in the Northeast (1965, 1969) suggested that these types were associated with two distinct archaeological traditions (ie. Laurentian versus Narrow Stemmed) the Preston Plains assemblage supports the proposal that the differences between these traditions in southern New England are “more subtle than were previously thought” (Hoffman 1990:115). Figure 7-5. Late Archaic triangular points from the Preston Plains Site illustrating a gradient between Brewerton (4 on the left) and Squibnocket (4 on the right) types. All are made of Plainfield Formation quartzite. 101 The greatest concentrations of Late Archaic Period cultural materials were identified at Locus 1, 2, 4, and 5. These appear to be areas that were used periodically throughout the Late Archaic Period, probably for short-term, seasonally-timed residence by small groups of foragers. Block excavations at Locus 2 revealed an overlapping series of DSFs, in addition to smaller, carbon-rich deposits that appear anthropogenic. Wood charcoal from two hearth-like deposits there dated to the Late Archaic Period. One, designated Fe.4, yielded a designated Fe.1, yielded a 14 C date of 4490 ± 70 BP (Beta-132294), while the other, 14 C date of 4650 ± 70 (Beta-132293). Block excavations at Locus 4 also revealed an overlapping series of DSFs. Wood charcoal from a hearth-like deposit there (Fe.2) yielded a 14 C date of 4310 ± 70 BP (Beta-132293). The block excavation at Locus 1 resulted in the identification of a series of Late Archaic Period 14C dated contexts, which will be described in detail in the following chapter. While no DSFs were detected at Locus 5, a hearth-like feature, designated Fe.6, was identified there. Wood charcoal from this feature yielded 14C date of 3990 ± 80 BP (Beta-136708), which fits with the Late Archaic Period artifact assemblage found there. Terminal Archaic Period occupation at Preston Plains is identified at Locus 5 by three diagnostic projectile points. These consist of a rhyolite Wayland Notched point (from N173 W9, SE quad) and two untyped chert Broadspears (from N167 W6, SW quad, and N175 W10) – all from topsoil contexts. The fact that these points were clustered at Locus 5 suggests that Terminal Archaic foragers continuted to use this location after their Late Archaic Period predecessors. An argillite Snook Kill point was also recovered, but this was an unprovenienced surface find. 102 Evidence of Early Woodland Period occupation is, for the most part, thinly distributed across the original Preston Plains Energy Center project area. One quartz Rossville point was recovered from plowed A-Horizon soil in Locus I (N52 W41, SE quad). Two other Rossville points, one of quartzite and another argillite, were recovered from topsoil contexts in Locus 5 (N170 W9, SE quad; N180 E0, SW quad). A fourth, made of quartz, was an unprovenienced surface find. A rhyolite Meadowood point was recovered from plowed A-Horizon soil in an area that is generally outside of any artifact concentration (N106 E10, SE quad) and may represent a lost projectile. However, a very robust hearth/roasting feature was discovered near Locus 2 during the 2006 machine trenching, and designated Fe.215 (Figure 7-6). This is the only Woodland Period feature known from the Preston Plains Site. Centered on excavation unit N131 E13 in the area labeled “Locus 2b” by Jones, it consisted of a circular pit filled with stone cobbles (manuports) that were visibly fire-cracked. Its matrix was charcoal-rich and contained many carbonized hickory nutshells. A sample of wood charcoal was sent for AMS dating and yielded a conventional 14C date of 2230 ± 40 BP (Beta-226147). Lithic debitage and a broken quartz biface were recovered from this feature. No diagnostic tool forms or radiocarbon data from Preston Plains can be confidently attributed to the Middle or Late Woodland Periods. While this place was likely used by local populations, or at least held meaning for them as part of a greater landscape, the effects of their activities are not archaeologically visible here. Archaeologically robust artifact deposits would not be generated here again here until the eighteenth-century. 103 Figure 7-6. Plan view of Fe.215, the only confirmed Woodland Period feature at the Preston Plains Site. Trowel points north. 7.5. DSFs at the Preston Plains Site Many DSFs were identified at Preston Plains from 1999-2003 under the oversight of Brian Jones, former archaeological field supervisor for the MPMRC. These were found to occur in clusters at Locus 1, 2, and 4, and considered to be anthropogenic features associated with Late Archaic Period occupations. Paula Coutre-Palmerino, a former University of Connecticut graduate student and teaching assistant for the University‟s Summer Archaeological Field School, was the first archaeologist involved with the Preston Plains Site to propose that these features represented the remains of semi-subterranean pit houses (personal communication, Coutre-Palmerino 2009). 104 To further investigate the distribution of DSFs across the site, as well as address a number of secondary issues, the author directed two episodes of machine trenching (2006 and 2008) within the revised (slightly smaller) Preston Plains Energy Center Project Area (see Figures 5-1) in addition to an area close to the edge of Avery Pond (see Figure 5-2). These were excavated by backhoe equipped with a 5-foot wide bucket, resulting in trenches that were approximately 1.5 m wide. The lengths of trenches ranged from 7 m to 45 m. Most trenches were excavated to depths of approximately 4 feet, while trenches exceeding this depth (MT-6, 9, 11) had one side graded back at an incline to reduce the risk of trench collapse. One end of every trench was similarly inclined to provide an egress. Though the distribution of machine trenches may appear somewhat random, each was located to address a particular question about the site‟s archaeology and/or geomorphology according to the evolving elements of the project. These work efforts identified 41 additional DSFs (Table 7-2). 105 Table 7-2. DSFs identified in machine trenches at the Preston Plains Site. Year Excavated 2006 2006 2006 2006 2006 2006 2006 2006 2006 2006 2006 2006 2008 2008 2008 2008 2008 DSFs present in trench? Yes No No No Yes Yes No No Yes Yes No No Yes Yes No No Yes Trench MT-1 MT-2 MT-3 MT-4 MT-5 MT-6 MT-7 MT-8 MT-9 MT-10 MT-11 MT-12 MT-13 MT-14 MT-15 MT-16 MT-17 Notes Aside from DSF-10 (first identified in Locus I excavation block), there were 3 DSFs identified in this trench. Aside from DSF-8 (first identified in Locus I excavation block), there were no DSFs identified in this trench. Aside from DSF-9 (identified in Locus I excavation block), there were no DSFs identified in this trench. There were 4 DSFs identified in this trench. There were overlapping DSFs (# unknown) in the east end of this trench, which appear to have been situated within the remnant glaciofluvial channel and previously sampled with a block excavation. 14 DSFs identified 4 DSFs identified If DSFs were present, they were obscured beyond recognition by a woodchuck burrow complex. 1 DSF identified 1 DSF identified 3 DSFs identified 10 DSFs identified Profile sections and descriptions of 23 specimens are included in this dissertation. Nine DSFs from the Locus I excavation block are reported in Chapter 8 while fourteen additional DSFs identified during machine trenching are reported in Chapter 9. These are sufficient to illustrate the morphological and stratigraphic diversity of DSFs at this site. Generalizing observations about the DSFs include: 1. While there is a range in the size of DSFs at Preston Plains, their morphology tends to be bimodal. They tend to occur in simple, basin-like forms and more complex “D-shaped pit” form, though intermediate forms occur as well, such as basins that are 106 slightly deeper at one end. Sizes vary, and they generally range from 2-5 meters across. The simple basin-like forms are usually shallower than the “D-shaped pit” form, which may extend up to 1.5 meters below ground surface. 2. The internal stratigraphy of DSFs at Preston Plains is variable, ranging from stratified to unstratified. Where stratification is discernable in “D-shaped pit” forms and forms that have a deeper end, strata are typically inclined from the shallow end towards the deeper end. 3. There is sometimes, but not always, a correlation between the locations of Late Archaic Period artifact concentrations and DSFs. In several cases, DSFs occur in close spatial association with concentrations of Late Archaic Period cultural materials, as was the case at Locus 1, 2, and 4. However, there is not always a correlation between the locations of prehistoric artifact concentrations and the locations of DSFs. For example, three DSFs were identified in MT-15 in an area that had very few artifacts. Additionally, a minor lithic concentration was detected in vicinity of MT-5 and a moderate concentration near MT-7, though no DSFs were identified in either of those trenches. Finally, a block excavation in the center of a multicomponent prehistoric site locus (Locus 5) detected no evidence of DSFs. 107 Chapter 8. Locus I: A Case Study of Late Archaic Period Occupation and a Cluster of DSFs This chapter presents data from a concentration of prehistoric archaeological deposits and 10 DSFs at Locus 1 (Figure 8-1). The majority of cultural deposits here are attributed to small groups of foragers associated with the Laurentian and NarrowStemmed Traditions who periodically established seasonally-scheduled, short-term residential sites here during the Late Archaic Period. Serial occupation here has resulted in a substantial aggregation of artifacts and cultural features. Twenty-four small, discrete features were identified, several of which were circular pit hearths penetrating the upper portions of DSF matrix while other hearths appeared to have been situated within, and were essentially part of, the shallow ends of DSFs. Hearth-related anthrosols were also recorded within DSF matrices. Small foraging groups, perhaps nuclear families, established short-term residential camps here in the summer and fall, exploiting local animal and plant resources and refurbishing their personal gear (sensu Binford 1982). While species-specific identifications are generally lacking, the majority of calcined bone recovered from heaths and hearth-related anthrosols is attributed to mammals. Wood identifications indicate the consistent use of broadleaf taxa for fuel. Two intact hearths dating to the latter portion of the 5th millennium BP (Fe.18 and 21) specifically indicate the use of American hornbeam, eastern hophornbeam, beech, birch, red oak, and white oak. The locally diverse forest canopy was exploited for its mast resources. The ubiquity of nutshells in 108 Figure 8-1. Plan of Locus 1 excavation block depicting all of the DSFs and most of the smaller, discrete features. hearth features indicates that that most “were formed in the fall and that collecting nuts was an important activity, perhaps carried out by women and children;” however, identified plant taxa also include species available in the summer (Largy and Demello 2010). Flintkapping activities at the site overwhelmingly reflect biface production, presumably to replace expired projectile points and knives that were discarded on site, while informal tools, including scrapers and flake tools, reflect organic material processing. The episodes of stone tool manufacture and/or maintenance composing the site aggregation appear to have been sporadic and short-term. The diversity and ubiquity of Late Archaic point types here, including Laurentian and subsequent Narrow-Stemmed 109 Tradition varieties, and the radiocarbon date series indicate repeated occupations that generally extend from the late 6th to early 4th millenniums. The nature and distribution of artifacts from across Locus 1 are discussed in the first section of this chapter (8.1.). The second section (8.2.) objectively reports the morphology and stratigraphy of DSFs while avoiding making interpretations about their genesis or function. The third section (8.3.) provides analysis and interpretation of discrete features, as far as one may go without an informed interpretation of DSFs here. These interpretations will be revisited and refined in Chapter 11 according to the model of DSF genesis and modification developed in therein. Several key questions emerge from review of the following data that cannot be effectively answered without an interpretation of DSF genesis: 1. Why are the oldest projectile points located within the matrix of DSFs 1-3? 2. Why are hearth-related anthrosols deeply buried within DSFs 1 and 2? 3. Why do hearths conform to the very edge of the shallow ends of DSFs 3, 4, and 7? 4. Why are cultural deposits and DSFs concentrated together? These questions are answered in the last portion of Chapter 11 (11.3) after the tree throw hypothesis is supported, and competing hypotheses discounted. The nature of interplay between DSFs and cultural deposits, as well as their cultural significance, is clearly apparent after DSFs are interpreted to have been initiated by naturally occurring tree throws. 110 8.1. Artifacts and their General Distribution For the convenience of analysis, Locus 1 is considered to be within the following grid area: N40-60 W30-50. Most of its total excavated area (108 sq m) is encompassed by the excavation block (105.5 sq m) (see Figure 8-1), which, according to systematic test pit data (see Figures 7-1 and 7-2), likely recovered most of the cultural material here. An additional 2.5 sq. m of soil was excavated in surrounding test pits at 5-m grid intervals (N40 W35, 40, 45; N45 W35, 40, 45, 50; N50 W45; N55 W35, 40) (see Figure 5-1). Lithic artifacts consist of debitage and a variety of tool forms (Table 8-1). Debitage is dominated by bifacial reduction flakes, though uniface reduction flakes are common. A total of 67 projectile points were recovered, nearly all of which exhibit use wear or damage (haft-wear, significantly reworked blades and/or broken elements). Most bifaces (n=85) consist of fragments or specimens that had not been reduced into formal tools. When taken together, the discard of expired projectile points, generation of biface reduction flakes, and presence of crude or broken (in-progress) bifaces suggests that formal, bifacial knives and/or projectile points were being refurbished and/or replaced as elements of personal gear. Animal or plant processing activities are also reflected by the recovery of scrapers (n=32), most of which are unifacial, and small uniface reduction/resharpeing flakes. Expedient flake tools were also recovered (n=34), which may have been used to process animal or plant resources. They may have used for the maintenance of the organic components of tools or weapons. The author suspects there are far more flake tools than have been identified in the assemblage thus far.as suggested 111 Table 8-1. Lithic tools and debitage, Locus 1. Class Variety Angular debris - small (1-3 cm) Angular debris - large (3-5 cm) Chunk (>5 cm) Core Flake (1-3 cm) Flake - large (>5 cm) Microdebitage (0-1 cm) Primary reduction debris Split cobble Tablet Untyped Anvil/pitted stone Biface Bifacial preform Flake knife Flake - retouched Flake - utilized Hammerstone Pendant Perforator Projectile point Scraper Count 346 48 14 5 4198 12 1359 13 3 17 528 1 85 6 9 12 13 5 1 1 67 32 22 Debitage Tools Untyped by the preponderance of projectile points dating to that period (Table 8-2, Figure 8-2). Most of the Late Archaic projectile points are attributed to the sequential, though partially overlapping, Laurentian and Narrow-Stemmed Traditions. Laurentian manifestations are generally thought to date to ca. 4800-4200 BP in Connecticut (McBride 1984), and represent the earliest expression of Late Archaic Period adaptation in southern New England. However, it should be noted that the earliest manifestation of the Laurentian tradition - the Vosburg Phase as expressed in the Laurentian Valley New York - dates to the mid-6th millennium BP (Ritchie 1965). The Narrow Stemmed Tradition appears to begin at ca. 4200 BP in southern New England (Ritchie 1969, McBride 1984, Pheiffer 1984), though its termination in Connecticut has been variably estimated at 3300 112 Table 8-2. Projectile points, Locus 1. Late Archaic types are italicized. Type Bare Island Beekman triangle Bifurcate (Kanawha-like) Brewerton corner notched Material quartzite quartzite quartz quartzite argillite quartzite argillite hornfels quartzite quartzite quartz argillite quartzite chert argillite quartz chert quartz quartzite quartz quartzite argillite chert quartz quartzite rhyolite quartzite quartz quartzite quartzite argillite quartzite unid. lithic quartz Count 1 1 1 1 1 5 1 1 3 1 1 2 1 1 1 1 1 1 5 1 1 2 2 4 8 1 1 1 3 1 1 9 1 3 Brewerton eared notched Brewerton side notched Lamoka Narrow-Stemmed Neville Otter Creek Rossville San Patrice Squibnocket triangle Squibnocket-stemmed Untyped Untyped lanceolate Untyped stemmed Untyped triangle Vosburg Wading River 113 Figure 8-2. Representative Late Archaic Period projectile points from Locus I, Preston Plains Site. Laurentian Tradition points are in the top row, while Narrow-Stemmed Tradition points are in the bottom row: a. Otter Creek (argillite); b-c. Vosburg (quartzite); d-e. Brewerton (quartzite); f. Bare Island (quartzite); g. Narrow-Stemmed (argillite); h. Narrow-Stemmed (quartzite); i. NarrowStemmed (quartz); j-k. Squibnocket Triangle (quartzite); Squibnocket Triangle (quartz). (Pheiffer 1984) and 2900 BP (McBride 1984). However, some evidence suggests that Narrow Stemmed points were used in southern New England as early as 4500 BP (Hoffman 1985; 1990). Thus, by liberal standards, this study assigns Laurentian points a date range of ca. 5500-4200 BP and Narrow Stemmed points a date range of ca. 45002900 BP. Ideally, the author would attempt to define individual activity areas across the Locus by “disaggregating” artifact scatters (e.g. Jones 2007, Sullivan III 1995). Unfortunately, two factors suggest that such attempts would not be productive. First, the 114 topsoil here has been graded to a depth of up to 25 centimeters along the north end of the Locus 1 excavation block, which is unfortunate as the majority of lithic artifacts were recovered from topsoil. The horizontal extent of this grading activity across the rest of Locus 1 is unknown, having not been generally recognized during most of the fieldwork, but is estimated to extend as far south as the N50 line. Thus, topsoil contexts here cannot be effectively separated into plowed, versus disturbed/graded contexts to facilitate such analysis. Second, the author assumes that DSF genesis shifts preexisting artifact concentrations, which further complicates such analysis, at least at a fine grained level. However, the author assumes that the majority of prehistoric materials recovered from topsoil contexts here are generally from the area, as the surface of the Locus 1 excavation block area still reflects glacial topography. Even without spatially disaggregating the lithic distribution, it is possible to loosely estimate the intensity of flintknapping activity represented. Borrowing from Jones (2007), it is estimated that a piece of debitage was produced at a rate of one every two seconds. According that that estimate, approximately 2.88 hours of flintknapping activity is represented at Locus 1 by a debitage total of 5184 (excludes microdebitage, some of which may have been misidentified). If we include microdebitage, for a total of 6543, the number of hours projected increased to 3.64 hours. Assuming that the debitage here is largely the product of serial occupation, the individual flintknapping sessions that contributed to it would have probably have only lasted for minutes each, as opposed to hours. 115 Lithic artifacts within DSF matrix and feature contexts escaped the reach of the plow (Figure 8-2). Lithic artifacts within DSF contexts tended to be most abundant in their upper layers, though some lithics were recovered from the deep ends of DSFs 1 and 2. These issues will be examined in detail in the following chapter. Figure 8-3. Lithic artifacts ≥1 cm recovered from beneath topsoil, Locus 1 excavation block (Surfer overlay by Mandy Ranslow). The counts depicted here may vary from those in profile drawings produced by the author in Chapter 9, which reflects minor differences in how Ranslow and the author sorted data. An interesting trend emerged regarding the distribution of diagnostic points from the Locus 1 excavation block. All diagnostic points recovered from topsoil contexts were 116 associated with the Late Archaic Period (Figure 8-4). While fewer diagnostic points were recovered from feature soil (DSFs and discrete features), they represent a much wider timeframe, one that extends back to the early Holocene (Figure 8-5). Specifically, DSFs 1, 2, and 3 contained points that are, on average, older than those found in topsoil contexts. Figure 8-4. Diagnostic points recovered from topsoil, Locus I excavation block (Surfer overlay by Mandy Ranslow). 117 Figure 8-5. Diagnostic points recovered from beneath topsoil, Locus I excavation block (Surfer overlay by Mandy Ranslow). The topsoil across the excavation block contained a relatively thin distribution of calcined bone. While most of this was probably deposited by prehistoric occupants, historic period additions cannot be ruled out. Historic period bone, much of which was burned but not calcined, was recovered from topsoil here and attributed to the 20th century component. Slightly higher amounts of calcined bone was recovered from feature soil contexts (DSF matrix and discrete features), which is probably be attributed to the fact that soil from these contexts were processed through flotation more often. Calcined bone from these contexts can be confidently attributed to prehistoric activity. 118 Bone identifications from smaller, discrete features are presented in section 8.3. of this chapter. Figure 8-6. Calcined bone recovered from beneath topsoil, by weight (grams), Locus 1 excavation block. 8.2. DSFs A series of 10 DSFs were identified within the Locus I excavation block. Machine trenches excavated to the east (MT-2), south (MT-3), and west (MT-4) of the excavation block (see Figure 5-1 and Table 7-2) showed that DSFs did not continue to occur in those directions, though DSFs were found to occur to the north of the excavation 119 block within MT-1. In sum, archaeological testing suggests that DSFs occur in the Locus 1 vicinity along a north-to-southeast running arc, much, if not most, of which is depicted in the Locus 1 excavation block plan. Some of these DSFs were simple basin-like forms, while others exhibited the classic “D-shaped pit” morphology so frequently described in the Middle Atlantic. This section describes the size, morphology, and stratigraphy of DSFs at Locus I. While these are interpreted to have been generated by tree throws in Chapter 11, they are described here objectively. DSF-1 exhibited “D-shaped pit” morphology (Figure 8-7). In plan, this DSF measured approximately 5 (E-W) by 4 (N-S) m. Its deep section (labeled “sub-pit” on the block plan) was oriented to the east and measured approximately 125 cmbs, while its shelf was oriented to the west and averaged approximately 60 cmbs. Its matrix appears to have been largely derived of glaciofluvial soils that naturally occur in this part of Preston Plains. DSF-1 is not robustly stratified, and most stratigraphic boundaries depicted in the soil profile represent very subtle changes in color or texture from one zone to the next. DSF-1 truncated DSF-2, and its upper portions were penetrated by smaller, discrete features. DSF-2 also exhibited “D-shaped pit” morphology (see Figure 8-7). In plan, this DSF measured approximately 4¼ (NE-SW) by 4 (NW-SE) m. Its deep section was oriented to the southwest and measured approximately 145 cmbs, while its shelf was oriented to the northeast and averaged approximately 70 cmbs. Its matrix appeared to have been largely derived of glaciofluvial soils that naturally occur in this part of Preston Plains. DSF-1 was not robustly stratified, though the strata that were discernable 120 Figure 8-7. Profile section drawing of DSF-1 and DSF-2. have been largely derived of glaciofluvial soils that naturally occur in this part of Preston Plains. DSF-1 was not robustly stratified, though the strata that were discernable (mostly from subtle patterns in gravel content) clearly plunged from the shallow end towards the deep end. DSF-2 was truncated by DSF-1. DSF-3 was a relatively shallow basin-shaped form that, in plan, measured approximately 5 (ENE-WSW) by 3 (NNW-SSE) m. While it was deepest (approx. 70 cmbs) in its east-northeastern end, it did not exhibit the robustly deepened section of a classic “D-shaped pit.” Its matrix was not robustly stratified; no profile drawings were made that depict a bisection of this DSF. DSF-3 appears to have truncated DSF-1, though the soil boundary used to make this determination was only vaguely recognizable. DSF-4 exhibited “D-shaped pit” morphology (Figure 8-8). It measured approximately 3 (NE-SW) by 2 (NW-SE) meters. Its deep section was oriented to the northeast and measured approximately 120 cmbs, while its shelf was oriented to the 121 southwest and averaged approximately 40 cmbs. DSF-4 matrix was largely homogenous, exhibiting no stratification whatsoever. Figure 8-8. Profile section drawing of DSF-4. DSF-5 was a simple, basin-shaped form with a maximum depth of 80 cmbs (Figure 8-9). In plan, it was an oval that is projected to have measured approximately 3½ (E-W) by 2½ (N-S) m. Its matrix appeared to have been largely derived of glaciofluvial soils that naturally occur in this part of Preston Plains. While the majority of its matrix was not stratified, its lower portions featured a distinct stratum of charred wood and heavily oxidized sand/gravel that closely match the contour of the feature‟s bottom (Fe.20). 122 Figure 8-9. Profile section drawing of DSF-5. DSF-6 was shallow basin-shaped form with a somewhat flattened bottom at 60 cmbs (Figure 8-10). Its horizontal dimensions have not been entirely determined, but it measured at least 4 m from east to west. Its matrix featured three distinct strata that plunged to the west. The uppermost of these strata contained what is suspected to be a remnant occupational surface that has been designated Fe.17. DSF-6 appears to have been truncated by DSF-10. 123 Figure 8-10. Profile section drawing of DSF-6. DSF-7 exhibited “D-shaped pit” morphology, though this assessment is based on only a partial excavation of this feature (Figures 8-11 and 8-12). The deepest recorded portion of it measured 70 cmbs, while the exposed shelf portion averaged 50 cmbs. Its matrix appeared to have been largely derived of glaciofluvial soils that naturally occur in this part of Preston Plains. DSF-7 truncated DSF-8 and DSF-9. 124 Figure 8-11. Profile section drawing of DSF-7 and DSF-8. Figure 8-12. Profile section of DSF-7 and DSF-9. 125 DSF-8 exhibited “D-shaped pit” morphology, though this assessment is based on only a partial excavation of this feature (see Figure 8-11). Its matrix appeared to have been largely derived of glaciofluvial soils that naturally occur in this part of Preston Plains. This DSF was truncated by DSF-4 and DSF-7. DSF-9 was a basin-shaped form (max depth = 75 cmbs), although this assessment is based on only a partial excavation of this feature (Figure 8-13). The maximum depth recorded for this feature was 75 cmbs, and its horizontal dimensions were not determined. Its matrix appeared to have been largely derived of glaciofluvial soils that naturally occur in this part of Preston Plains. This DSF was truncated by DSF-7. Only a small part of DSF-10 was recorded in the Locus I excavation block. Due to the small excavation sample size, its size, morphology, and stratigraphy were not assessed. This feature appeared to have truncated DSF-6. Figure 8-13. Profile section of DSF-9. 126 8.3. Discrete Features This section reports the provenience, physical character, artifact content, and radiocarbon data for each discrete feature. Four categories of artifacts are reported: charred botanicals, calcined bone, charred wood, and lithics. With the exception of a few shell fragments (noted in the “interpretation” section of each table), these categories account for all cultural material recovered from discrete features. Preliminary functional interpretations are provided when sufficient information is present. Bear in mind that there interpretations will be further informed by a fuller understanding of DSF formation and cultural significance presented in Chapter 11. Three main categories of features are present at Locus 1. The largest consists of feature soils contained within the matrix of DSFs 1, 2, and 3. These include indeterminate soil stains (Fe.2, 4, 5, 9, 11, 13, 15, and 16) and hearth-related anthrosols (Fe.10 and 12). While most of the indeterminate soil stains may also be hearth-related anthrosols, this determination cannot be confidently made according to the current analysis. However, two features (Fe.10 and 12) are confidently identified as hearthrelated anthrosols buried deep within DSF matrix. The second category consists of circular pit hearths that are clearly truncated by the plowed A-horizon (Fe.3, 6, 8, 14, 19, 21, and 25). Four of these penetrate DSF matrix (Fe.3, 6, 8, and 14) (Figure 8-14 [Fe.8]), while the other three (Fe.19, 21, and 25) penetrate C-horizon soil immediately beyond the edge of DSF-1 (see Figure 8-14 [Fe.21]). 127 The third category consists of hearths that conform to the shallow edge of DSFs (Fe.18, 22, 26). These all contain charred wood, thermally altered gravel, and conform to the very edge of their respective DSFs (see Figure 8-14 [Fe.18]). Other features were identified that do not fit these categories. Fe.17 is tentatively identified as a remnant occupational surface or anthropogenic feature excavated into the upper portion of DSF-6. Fe.24 appears to be a pit, possibly anthropogenic, that filled with natural and cultural deposits, some of which may be in secondary contexts. Fe.20 consists of a stratum of charred wood in the lower portion of DSF-5 that appears to represent a natural, in-situ burning episode. Data from hearths and hearth-related anthrosols indicate that terrestrial plant resources were consistently exploited, and in some cases, wetland plants were exploited as well. Meat consumption is frequently indicated by the recovery of calcined bone, and flintknapping activity is sometimes indicated by the presence of debitage. Wood fuel includes a variety of broadleaf trees, including beech, birch, oak, hickory, and hornbeam. Regarding seasonality, Largy and Demello (2010) state that “identified plant taxa indicate occupation beginning in early or mid-summer with the majority of taxa suggesting a fall occupation.” Hickory and hazelnut shells are the most ubiquitous fall season indicators at Locus 1. 128 Figure 8-14. Examples of primary hearth deposits at Locus 1, viewed in profile. Fe.8 (top left, view west) is a pit hearth penetrating DSF-1 matrix. Fe.21 (top right, view south) is a pit hearth penetrating C-Horizon soils immediately adjacent to DSF-1. Fe.18 (bottom, view south) is a hearth deposit within, and conforming to, the shallow end of DSF-4. Feature numbers 1 and 23 were not assigned. Some features are not depicted on the Locus 1 excavation block plan because the author was not sure of their precise 129 location and/or morphology while making the plan. Features 2-16 were excavated prior to the author‟s involvement with the Preston Plains Site, while Features 17-26 were excavated under his supervision. Artifact counts are provided for lithic, faunal and charred botanical items. Representative charred wood identifications are provided where possible. Summaries of each discrete feature are presented below. The nature of Fe.2 is indeterminate (Table 8-3). It was identified as a soil stain within the matrix of DSF-2, though it is unclear as to whether it is natural or anthropogenic. Table 8-3. Feature 2 summary. Unit(s) Occupied Horizontal Extent Vertical Morphology Stratigraphy Botanicals Calcined Bone Artifacts Charred Wood Lithics 14 C Data Interpretation N50 W37 ~15 cm diameter 40-76 cmbs “circular stain” internal stratigraphy unknown; situated within DSF-2 soil matrix --present none -Indeterminate Fe.3 was a pit hearth intruding DSF matrix, in the area where DSF-1 and 2 intersect (Table 8-4). This determination is based largely on this feature‟s distinct bowllike morphology and the presence of charred wood and thermally altered gravel lining its bottom. It contained a minimal amount of lithics, a small quantity of hickory nutshell, and white oak fuelwood. 130 Table 8-4. Feature 3 summary. Unit(s) Occupied Horizontal Extent Vertical Morphology Stratigraphy Botanicals Calcined Bone Charred Wood Lithics 14 C Data N49-50 W36-37 ~40 cm diameter 35-60 cmbs circular plan; bowl-shaped profile somewhat stratified; intrudes DSF soil matrix; truncated by Plowed AHorizon Nutshell: 4 Carya sp. (hickory) -includes white oak 1 core or steep-bitted scraper (quartz), 1 thermally altered gravel -Pit hearth intruding DSF-1/2 matrix. It has morphological integrity and its lower stratum contained a concentration of burnt/oxidized pebbles and charcoal. Artifacts Interpretation The nature of Fe.4 is indeterminate (Table 8-5). Its artifact content suggests that it may represent a zone of hearth-related anthrosols within the upper portions of DSF-1 martix; however, there is not sufficient information to confidently postulate this. Table 8-5. Feature 4 summary. Unit(s) Occupied Horizontal Extent Vertical Morphology Stratigraphy Botanicals Calcined Bone Artifacts Charred Wood Lithics 14 C Data Interpretation N49-50 W38 ~45 cm diameter 30-50 cmbs “circular stain” internal stratigraphy unknown; situated within DSF-2 soil matrix -1 lg mammal present 3 thermally altered stone, 5 flake -Indeterminate The nature of Fe.5 is indeterminate (Table 8-6). Its artifact content suggests that it may represent a zone of hearth-related anthrosols within the upper portions of DSF-2 martix; however, there is not sufficient information to confidently postulate this. 131 Table 8-6. Feature 5 summary. Unit(s) Occupied Horizontal Extent Vertical Morphology Stratigraphy Botanicals Calcined Bone Artifacts Charred Wood Lithics 14 C Data Interpretation N48 W40 40 (N/S) - x - 30 (E-W) cm 20-45 cmbs oval in plan internal stratigraphy unknown; situated within DSF-2 soil matrix -1 med-to-lg mammal, 1 lg mammal present 3 flake -Indeterminate Fe.6 was a pit hearth intruding DSF-1 matrix (Table 8-7). This determination was made according to its distinct bowl-like morphology and artifact contents. It contained a concentration of cobbles, suspected to be manuports, in addition to lithic debitage and charred wood and a hazelnut specimen. Table 8-7. Feature 6 summary. Unit(s) Occupied Horizontal Extent Vertical Morphology Stratigraphy Botanicals Calcined Bone Charred Wood Lithics 14 C Data N48-49 W40-41 60 (N/S) - x - 50 (E-W) cm 20-65 cmbs “circular stain” internal stratigraphy unknown; intrudes DSF-1 soil matrix; truncated by Plowed A-Horizon 1 Corylus sp. (hazelnut) meat -present 1 thermally altered stone, 13 flake, 1 small angular debris -Pit hearth intruding DSF-1 matrix. It has morphological integrity and contained a concentration of cobbles (suspected manuports). Artifacts Interpretation The nature of Fe.7 is indeterminate (Table 8-8). It was first speculated to be a post mold at the bottom of DSF-1, when all DSFs at Preston Plains were presumed to be pit house remnants. The author does not currently believe Fe.7 was a post mold and lacks sufficient data to develop alternate interpretations. 132 Table 8-8. Feature 7 summary. Unit(s) Occupied Horizontal Extent Vertical Morphology Stratigraphy Botanicals Calcined Bone Artifacts Charred Wood Lithics 14 C Data Interpretation N50 W39 25 cm diameter 80-87 cmbs “circular stain” internal stratigraphy unknown; situated within DSF-1 soil matrix -----Indeterminate. Speculated to be a post mold in the field but the author remains skeptical. Fe.8 was a pit hearth intruding DSF-1 matrix (Table 8-9) Review of site photos shows its distinct, bowl-shaped morphology, and its rich artifact content confirms its anthropogenic origin. Radiocarbon data places its genesis in the latter portion of the 5th millennium BP, which suggests a Laurentian or Narrow Stemmed Tradition association. Recovered nutshell fragments consist of hickory and unidentified specimens. Fe.8 is one of two features at Locus 1 that were analyzed for the presence of soft plant tissue by Dave Perry (2002). It was found to contain aquatic plant resources. The starchy roots of the Alisma and Typha may have been used as food, while the Scirpus may represent the remains of a mat (Perry and Jones 2002). Terrestrial plants are represented as well. The identification of ferns (Dryopteris and Peteridium) is particularly interesting. One ethnographic account describes the use of ferns to wrap tubers before placing them in earth-ovens, so perhaps they were used for food processing (Reeve and Forgacs 1999). The Medeola likely served as another starchy food source. 133 Table 8-9. Feature 8 summary. Unit(s) Occupied Horizontal Extent Vertical Morphology Stratigraphy N49-50 W41-42 60 (N/S) - x - 80 (E-W) cm 30-70 cmbs circular plan; bowl-shaped profile feature fill is stratified; intrudes DSF-1 soil matrix; truncated by Plowed A-Horizon Nutshell: 2 Carya sp. (hickory), 5 unid. Soft Tissue: 4 Alisma (water plantain), 18 Dryopteris (wood fern), 2 Medeola (Indian cucumber), 2 Poaecae-aquatic (wetland grass), 1 Peteridium (bracken fern), 1 Scirpus (bulrush), 1 Sparaganium (burreed), 42 Typha (cattail) 28 med-to-lg mammal, 1 vertebrate present 2 biface fragments (1 quartz, 1 chert), 10 flake, 2 large angular debris, 6 thermally altered stone wood charcoal yielded a 14C date of 4230±60 BP (Beta-130446) Pit hearth intruding DSF-1 matrix. Botanicals Artifacts Calcined Bone Charred Wood Lithics C Data Interpretation 14 The nature of Fe.9 is indeterminate (Table 8-10). While this deposit, located within DSF-1 matrix, contained lithic materials there is not sufficient information to determine whether or not it is anthropogenic. It may represent a zone of hearth-related anthrosols. Table 8-10. Feature 9 summary. Unit(s) Occupied Horizontal Extent Vertical Morphology Stratigraphy Botanicals Calcined Bone Artifacts Charred Wood Lithics 14 C Data Interpretation N50 W40 25 cm diameter 30-53 cmbs “circular stain” internal stratigraphy unknown; contained within DSF-1 soil matrix --present 1 biface fragment (quartzite), 10 flake -Indeterminate. Possibly hearth-related anthrosols within DSF-1 matrix. 134 Fe.10 consists of hearth-related anthrosols within DSF-2 matrix (Table 8-11). Review of site photos indicates that it contains a concentration of thermally altered stones that are probably manuports. It also contained charred wood, two lithic artifacts and hazelnut shells. It had a plunging orientation within DSF-2 matrix, and was inclined towards the deep end. Table 8-11. Feature 10 summary. Unit(s) Occupied Horizontal Extent Vertical Morphology Stratigraphy Botanicals Calcined Bone Charred Wood Lithics 14 C Data N50-51 W35-36 90 (N/S) - x - 120 (E/W) cm 60-75 cmbs circular plan; shallow, lens-shaped profile; plunged slightly to the southwest not internally stratified; contained within DSF-2 soil matrix and conforms to plunging slope of encompassing DSF-2 strata Nutshell: 5 Corylus sp. (hazelnut). -present 1 thermally altered stone, 1 flake, 1 untyped tool (quartz) -Hearth-related anthrosols within DSF-2 matrix. Contains a concentration of charcoal and thermally altered stones. It has a plunging orientation within DSF soil matrix, a slightly elongated plan, and lack of sharply defined horizontal boundaries. Artifacts Interpretation The nature of Fe.11 is indeterminate (Table 8-12). While it may represent a zone of hearth-related anthrosols within DSF-1 martix, there is not sufficient information to confidently postulate this. Fe.12 was a zone of hearth-related anthrosols in the deep end of DSF-1 (Table 813). While this feature was not morphologically distinct, it contained a significant quantity of charred wood, lithics, calcined bone, and botanicals. Radicarbon data places its genesis in the late 6th millennium BP, which may indicate a Laurentian Tradition association. 135 Table 8-12. Feature 11 summary. Unit(s) Occupied Horizontal Extent Vertical Morphology Stratigraphy Botanicals Calcined Bone Artifacts Charred Wood Lithics 14 C Data Interpretation N48 W41 50 cm diameter 45-65 cmbs circular stain internal stratigraphy unknown; contained within DSF-1 soil matrix -21 lg mammal, 1 vertebrate present 2 flake -Indeterminate Table 8-13. Feature 12 summary. Unit(s) Occupied Horizontal Extent Vertical Morphology Stratigraphy Botanicals Artifacts Calcined Bone Charred Wood Lithics 14 C Data Interpretation N48-49 W38 70 cm diameter 75-95 cmbs circular stain internal stratigraphy unknown; contained within DSF-1 soil matrix Nutshell: 8 Carya sp.(hickory), 87 Corylus sp. (hazelnut). Seed: 1 Cornus sp. (dogwood shrub), 8 Vitis sp. (grape) Soft Plant Tissue: 3 Dryopteris (wood fern), 1 Lycopodium (club moss), 2 Typha (cattail), 1 Zizania (wild rice). 8 vertebrate present 2 biface fragments (1 quartz, 1 quartzite), 10 thermally altered stone, 131 flake, 1 small angular debris wood charcoal yielded a 14C date of 5290±80 BP (Beta-131054) It was identified as a “possible disturbed hearth” during field investigations. It contained many nutshells, including hickory and hazelnut, and also contained 8 grape seeds. Fe.12 is one of two features at Locus 1 that were analyzed for the presence of soft plant tissue by Dave Perry (2002). His findings support the interpretation that the anthrosols of Fe.12 are hearth-related. It was found to contain two aquatic plant taxa. The starchy Typha roots and Zizania may have been used as food (Perry and Jones 2002). Two terrestrial plant taxa are represented as well. The fern may have been used to wrap roots tubers for preparation in an earth-oven. The function of the moss (Lycopodium) is 136 unknown, though it may have been used as tinder to start fires (Rossen and Dillehay 2001). The nature of Fe.13 is indeterminate (Table 8-14). This small stain may represent hearth-related anthrosols, but this is only speculation based on artifact content. Table 8-14. Feature 13 summary. Unit(s) Occupied Horizontal Extent Vertical Morphology Stratigraphy Botanicals Calcined Bone Artifacts Charred Wood Lithics 14 C Data Interpretation N49 W39 40 cm diameter 30-55 cmbs circular stain unknown Nutshell: 1 Corylus sp. (hazelnut). 3 vertebrate includes American hornbeam and birch 2 thermally altered stone, 5 flake -Indeterminate Fe.14 was a pit hearth intruding DSF-3 matrix (Table 8-15). It contained a concentration of charred wood, thermally altered stone, and lithics. The pitted stone was probably used for nut-processing activities, as Fe.14 contained thousands of hickory nutshell specimens in addition to lesser quantities of hazelnut and oak. Fruit seeds were also recovered, consisting of huckleberry and grape. Calcined bone specimens were also recovered. The nature of Fe.15 is indeterminate (Table 8-16). Review of site photos suggests that it is hearth-related because it contained what appears to be a concentration of thermally altered stones. Perhaps it was a pit hearth intruding DSF-3 matrix or hearthrelated anthrosols therein. Its morphology was ambiguous and it contained few artifacts. 137 Table 8-15. Feature 14 summary. Unit(s) Occupied Horizontal Extent Vertical Morphology Stratigraphy N51-52 W42 40 cm diameter 30-60 cmbs circular stain unknown 1 possible nutmeat Nutshell: 3278 Carya sp. (hickory), 45 Corylus sp. (hazelnut), 4 Quercus sp. (oak), 38 unid. Seed: 2 Artemisia sp. (wormwood), 1 Cornus sp. (dogwood shrub), 1 Gaulussacia sp. (Huckleberry), 2 Rhus radicans sp. (poison ivy), 1 Vitius sp. (grape), 1 unid. 1 vertebrate, 2 possible turtle includes white oak and beech 1 anvil stone (pitted cobble), 7 thermally altered stone, 16 flake, 1 small angular debris -Indeterminate. 3 unid. shell fragments recovered as well. Botanicals Artifacts Calcined Bone Charred Wood Lithics C Data Interpretation 14 Table 8-16. Feature 15 summary. Unit(s) Occupied Horizontal Extent Vertical Morphology Stratigraphy Botanicals Calcined Bone Artifacts Charred Wood Lithics 14 C Data Interpretation N52 W41 15 (N/S) - x - 30 (E/W) cm 34-60 cmbs ovoid stain unknown Nutshell: 22 Corylus sp. (hazelnut) 1 unid. mammal present 1 small angular debris -Indeterminate. 1 unid. shell fragment recovered as well. The nature of Fe.16 is indeterminate, though it appears to have contained heathrelated anthrosols (Table 8-17). It may be an individual hearth or perhaps an extension of Fe.15. 138 Table 8-17. Feature 16 summary. Unit(s) Occupied Horizontal Extent Vertical Morphology Stratigraphy Botanicals Artifacts Calcined Bone Charred Wood Lithics 14 C Data Interpretation N52 W40-41 30 cm diameter 50-60 cmbs circular stain unknown Nutshell: 155 Corylus sp. (hazelnut), 4 Carya sp. (hickory) Seed: 1 cotyledon, 1 Gaylussacia sp. (Huckleberry), 4 monocot, 5 unid. 2 vertebrate present 5 flake -Indeterminate Fe.17 is tentatively identified as an occupational surface within a depression in DSF-6 (Table 8-18). It contained a high quantity of lithic materials and a large manuport. Faunal remains were recovered as well, in addition to hazelnut shell fragments and other, unidentified specimens. Table 8-18. Feature 17 summary. Unit(s) Occupied Horizontal Extent Vertical Morphology Stratigraphy Botanicals Artifacts Calcined Bone Charred Wood Lithics 14 C Data N54-55 W39 50 cm diameter 25-60 cmbs circular stain, bowl-shaped not robustly stratified; contained within DSF-6 matrix Nutshell: 5 Corylus sp. (hazelnut), 7 unid. Seed: 1 possible seed 6 unid. mammal, 2 med-to-lg mammal present 2 thermally altered stone, 54 flake, 24 small angular debris -This feature represents a concentration of botanical, faunal, and lithic material occurring within DSF-6 matrix. In the field, this was interpreted to be a separate feature that was dug into DSF-6, or a layer of DSF 6 that was occupied as a living surface. Interpretation 139 Fe.18 was a hearth conforming to the shallow end of DSF-4 (Table 8-19). Its matrix contained a high proportion of charred wood while its bottom was lined by thermally altered gravel. The charred wood is interpreted to be fuelwood, and includes American hornbeam, birch, white oak and red oak. concentration of calcined mammal bone. Its matrix contained a small The southwestern edge of this feature conformed to the edge of the DSF that cradled it, so their formations appear to be closely related. Radiocarbon data indicates that this feature was deposited near the end of the 5th millennium BP, which probably indicates a Narrow Stemmed Tradition association. This feature is similar in form and context to Fe.22 and 26. Table 8-19. Feature 18 summary. Unit(s) Occupied Horizontal Extent Vertical Morphology Stratigraphy Botanicals Artifacts Calcined Bone Charred Wood Lithics 14 C Data N46-47 W35-36 100 (NW/SE) - x - 75 (NE/SW) cm 25-54 cmbs basin shape; irregular in plan not robustly stratified Nutshell: 1 unid. Seed: 1 Potentilla sp. (resembles cinquefoil), 1 Vitis sp. (grape), 1 possible seed 6 lg mammal includes American hornbeam, white oak, red oak, and birch 30 thermally altered stone wood charcoal yielded a 14C date of 4120±60 BP (Beta-219137) Hearth in shallow end of DSF-4. Its morphology suggests it is a part of DSF-4 and that their depositions were interrelated. Interpretation Fe.19 was a pit hearth located immediately west of DSF-1 (Table 8-20). This determination was made according to its distinct bowl-like morphology, artifact contents, and the fact that it contained thermally altered gravel. It contained a few nutshell fragments, including oak and hazelnut. It contained a variety of fuelwoods, including white oak, red oak, birch, and an unidentified conifer. Radiocarbon dating of a wood 140 charcoal fragment places its genesis in the latter portion of the 5th millennium BP, which probably indicates a Laurentian or Narrow Stemmed Tradition association. Table 8-20. Feature 19 summary. Unit(s) Occupied Horizontal Extent Vertical Morphology Stratigraphy Botanicals Artifacts Calcined Bone Charred Wood Lithics 14 C Data N49 W42-43 100 cm diameter 20-35 cmbs shallow, round basin not stratified Nutshell: 1 Quercus sp. (oak), 1 Corylus sp. (hazelnut), 3 unid. Seed: 1 Gaylossacia sp. (huckleberry) 2 vertebrate includes white oak, red oak, birch, and unidentified conifer 203 thermally altered stone wood charcoal yielded a 14C date of 4270±50 BP (Beta-219138) This feature was identified in the field as a pit hearth due to its distinct morphology and charcoal content. Interpretation The nature of Fe.20 is indeterminate, though it appears to represent an open surface where burning occurred (Table 8-21). It was recorded in the field as a thin, charcoal-rich stratum near the bottom of DSF-5. It conforms to the bowl-shaped morphology of this feature and the consistent distribution of thermally altered gravel that lines its bottom suggests in-situ modification. Most of the larger charred wood chunks appeared to be of the same species, identified through laboratory analysis as oak. Nearly all of the cultural material from DSF-5 was recovered from its uppermost portions, well above the Fe.20 burn layer. 141 Table 8-21. Feature 20 summary. Unit(s) Occupied Horizontal Extent Vertical Morphology Stratigraphy Botanicals Artifacts Calcined Bone Charred Wood Lithics 14 C Data N54-55 W42-43 80 (NW/SE) - x - 150 (NE/SW) cm 60-84 cmbs ovate, bowl-shaped this feature is a distinct stratum of DSF-5 Nutshell: 3 Corylus sp. (hazelnut), 2 unid. Seed: 2 Rubus sp. (raspberry/blackberry/dewberry), 1 Papaveraceae (poppy family), 2 unid. -includes oak 1827 thermally altered stone, 4 small angular debris wood charcoal yielded a 14C date of 4710±50 BP (Beta-219139) This feature is a thin, charcoal-rich stratum within DSF-5 matrix. It consists of a charcoal layer near the bottom of this DSF that conforms to surrounding stratigraphy; therefore, Fe.20 and DSF-5 are thought to have formed coincidentally. Interpretation Fe.21 was a pit hearth located adjacent to DSF-1 (Table 8-22). This determination was made in the field according to its distinct morphology, charred wood, and thermally altered gravel content. It contained nutshell fragments, including hazelnut and hickory, in addition to fruit seeds consisting of grape and Rubus sp. Fuelwood included American hornbeam, eastern hophornbeam, oak, and birch. Radiocarbon data from this feature places its genesis at the end of the 5th millennium BP, which suggests a Narrow Stemmed Tradition association. Fe.22 was a hearth conforming to the shallow end of DSF-3 (Table 8-23). Only a portion of this feature was recognized during excavations in 2006, so its total morphology is unknown. Like Fe.18 and 26, it contained a high proportion of charred wood, its bottom was lined with thermally altered gravel, and portions of its edge conformed to the shallow end of a DSF. Because and edge of this feature conformed to the edge of the 142 DSF that cradled it, their formations are assumed to be closely related. Radiocarbon data indicates Table 8-22. Feature 21 summary. Unit(s) Occupied Horizontal Extent Vertical Morphology Stratigraphy Botanicals Artifacts Calcined Bone Charred Wood Lithics 14 C Data N46-47 W41-42 85 cm diameter 30-73 cmbs deep, circular pit stratified, with charred wood concentrated on bottom Nutshell: 16 Corylus sp. (hazelnut), 2 Carya sp. (hickory), 10 unid. Seed: 1 Artemisia (wormwood [may not be charred]), 7 Vitis sp. (grape), 1 Rubus sp. (raspberry/blackberry/dewberry), 3 unid. 1 vertebrate includes American hornbeam, eastern hophornbeam, oak, and birch 934 thermally altered stone, 10 flake wood charcoal yielded a 14C date of 4070±50 BP (Beta-219140) This feature was identified as a pit hearth according to its distinct morphology and charred wood content. Interpretation Table 8-23. Feature 22 summary. Unit(s) Occupied Horizontal Extent Vertical Morphology Stratigraphy Botanicals Artifacts Calcined Bone Charred Wood Lithics 14 C Data Interpretation N53 W41-42 unknown 30-50 cmbs (in the archaeologically recognized portion) unknown unknown; edge of DSF-3 Nutshell: 1 Corylus sp. (hazelnut), 599 Carya (hickory), 9 unid. Seed: 1 unid. 2 unid. includes white oak and red oak 2 thermally altered stone, 2 flake 1 carbonized hickory (Carya sp.) nut shell frag. yielded a 14C date of 5030±50 BP (Beta-222824) This appears to be a hearth within the perimeter of DSF-3. Oxidized (possibly heat-altered) pebbles are present, in addition to charred organics. Fe.22 and DSF-3 are thought to have formed coincidentally. that this feature was deposited near the beginning of the 5th millennium BP, which suggests a Laurentian Tradition association. It contained a large quantity of hickory nutshells and very few lithics and calcined bone. 143 The nature of Fe.24 is indeterminate, though it is suspected to have been a small pit (possibly anthropogenic) in DSF-3 matrix that filled naturally (Table 8-24). The matrix of this circular pit was very fine-grained (sands and silt) and appeared to be largely washed-in. It contained a small quantity of charred wood but significant quantities of lithic debitage, calcined bone, and botanical materials. The one radiocarbon date from this feature is attributed to the Middle Archaic Period, which pre-dates the suspected Late Archaic Period deposition of the DSF that it penetrates. Thus, this date is not thought to correlate with the genesis of Fe.24, and corroborates the speculation that at least some of its artifact content consists of secondary depositions from previous components. While the nutshells and fruit seeds may be associated with human consumption, most of the taxa identified “do not seem to be among the commonly known economic plants” and have been suggested to “have been preserved...perhaps as the result of a [natural] burn.” Fe.25 was tentatively identified as a pit hearth located immediately adjacent to DSF-1 (Table 8-25). It exhibited shallow, bowl-shaped morphology, contained a large proportion of charred wood, and was bordered by thermally altered sand and gravel. It contained few artifacts. 144 Table 8-24. Feature 24 summary. Unit(s) Occupied Horizontal Extent Vertical Morphology Stratigraphy N53 W40-41 60 cm diameter 31-77 cmbs deep, circular pit stratified; intrudes into DSF-3 Nutshell: 3 Carya sp. (hickory), 20 Corylus sp. (hazelnut), 3 Quercus sp. (oak), 8 unid. Seed: 1 Gaylussacia sp. (huckleberry), 1 Artemisia sp. (wormwood), 9 Rubus sp. (blackberry/raspberry/dewberry), 1 Ceanthus [americanus] (Jersey Tea), 2 Lysimachia sp. (yellow [whorled] loosestrife), 1 Leguminoseae (Legume family), 1 Solanaceae (nightshade family), 3 Vitis sp., 23 unid./tentatively identified. 6 vertebrate, 3 lg mammal, 5 med-to-lg mammal includes shagbark hickory, oak, and unid. conifer 27 flake, 16 small angular debris, 6 untyped debitage charred wood frag. yielded a 14C date of 7620±60 BP (Beta-222825) This feature was a pit intruding DSF-3 and contained stratified silt and sand thought to have washed-in. Charred botanical objects recovered from its matrix likely secondary depositions, a hypothesis supported by the fact that it yielded a charred wood fragment that pre-dates DSF-3 (dated via Fe.22). While this feature may have been created by an animal, its regular form suggests it is anthropogenic, though its contents are secondary depositions. Botanicals Artifacts Calcined Bone Charred Wood Lithics 14 C Data Interpretation Table 8-25. Feature 25 summary. Unit(s) Occupied Horizontal Extent Vertical Morphology Stratigraphy Botanicals Calcined Bone Artifacts Charred Wood Lithics 14 C Data Interpretation N47-48 W43 80 (N/S) - x - 65 (E/W) cm 20-35 cmbs ovate; bowl-shaped not stratified Nutshell: 2 Corylus sp. (hazelnut) -present 1 flake -This feature was identified as a pit hearth according to its distinct morphology and charred wood content. Fe.26 was a hearth conforming to the shallow end of DSF-7 (Table 8-26). Its matrix contained a high proportion of charred wood and it also contained thermally 145 altered gravel. The southern edge of this feature conformed to the edge of the DSF that cradled it, so their formations appear to be closely related. Radiocarbon data indicates that this feature was deposited near the end of the 5th millennium BP, which suggests a Narrow Stemmed Tradition association. This feature is similar in form and context to Fe.18 and 22. Table 8-26. Feature 26 summary. Unit(s) Occupied Horizontal Extent Vertical Morphology Stratigraphy Botanicals Calcined Bone Artifacts Charred Wood Lithics 14 C Data Interpretation N45 W33-34 40 (N/S) - x - 100 (E/W) cm 25-40 cmbs irregular not stratified Nutshell: 1 hazelnut -present 8 thermally altered stone wood charcoal yielded a 14C date of 4090±40 BP (Beta-222826) This feature consists of a very charcoal-rich soil package occurring at the shallow end of DSF-7. Its morphology suggests it is a part of DSF7 and that their deposition coincided or was, at the least, closely interrelated. 146 Chapter 9. Testing a Hypothesis: Artifact Concentrations are Not Present on the Bottoms of DSFs at Preston Plains In this chapter, the hypothesis that there are no artifact concentrations on the bottoms of most DSFs at Preston Plains is tested, the purpose of which is to discount the possibility that these features were created by Native Americans for use as pit houses. It is assumed that if humans used the bottoms of DSFs as habitation surfaces this behavior would be reflected by remnant artifact concentrations. Quantitatively demonstrating the absence of such concentrations allows us to more confidently support the natural tree throw hypothesis for DSF genesis featured in Chapter 11. Furthermore, the data reviewed in this chapter offers additional insights regarding the distribution of cultural material at Preston Plains that contribute to the analysis in Chapter 11. Two different data sets are used to test the hypothesis, the first of which first is drawn from the TUs (column samples) excavated through various soil contexts across the greater site area. Eleven of the 25 TUs were used to sample DSFs, and are directly relevant to the hypothesis test. Data from the remaining TUs (except for TU-8, which is omitted due to its irrelevancy) is also reviewed because it reveals additional information regarding the relationship between cultural materials and soil contexts across the site. The second data set consists of lithic artifact distributions from 7 DSFs in the Locus I excavation block. Together, these two data sets test include 18 DSFs. As will be shown, no artifact concentrations were found on the bottoms of the 18 DSFs. It was expected that artifacts most frequently occur in their upper or middle-to147 upper portions, and that their quantities diminish as depth increases. This expectation was generally confirmed, which supports the interpretation that artifacts tend to be deposited on the ground surface and transported downward into DSF matrix through natural mechanisms. This analysis also recognized the possibility of identifying distinct artifact concentrations/stratums in upper or middle portions of DSFs. This pattern appears to be evident for one DSF (DSF-6, at Locus 1), which has what appears to be a remnant occupational surface. This prompts us to seriously consider the question of whether or not humans used or modified upper portions of DSFs in the final interpretation. Other patterns revealed in this analysis include the tendency for artifacts to become deeply buried in various types of feature contexts (DSFs, the remnant glaciofluvial channel, and an extensive soil complex in MT-17) but remain otherwise confined to the shallow topsoil layer across the greater Preston Plains landform. Furthermore, some DSFs are virtually devoid of artifacts and located are outside of artifact concentrations, which demonstrate that the presence of DSFs does not always correspond with the locations of prehistoric Native American artifact concentrations at Preston Plains. 9.1. Testing Results for DSFs Identified in Machine Trenches in 2008 The first data set consists of artifact distributions in TUs excavated in 2008 during machine trenching. A variety of features were identified during this work effort, summarized in Table 9-1, and most were sampled using TUs according to the 148 methodology described in Chapter 5. The results of these samples are summarized in this sub-chapter. A total of 11 DSFs were sampled, and the results of the hypothesis test are printed in italics. Some of the TUs penetrated the remnant glaciofluvial channel (n=5), while others (n=4) were control samples penetrating non-feature contexts. Table 9-1. Summary of features identified in machine trenches at Preston Plains in 2008 (Class abbreviations: DSF=Deep Soil Feature; RC=Remnant Channel; U=unidentified). Exposed Feature Dimensions Width Depth (cm) (cmbs) 150 230 250 900 >600 300 160 110 250 300 670 >300 80 150 300 300 80 200 250 >400 >700 250 500 150 130 100 50 75 105 >120 >120 80 40 80 70 100 >120 >130 45 70 70 70 70 50 100 110 100 80 100 90 80 90 General Feature Descriptions Class Morphology Basin Basin Basin Channel-like Basin Basin Ambiguous Basin Basin Channel-like Pit-like Flat bottom basin Flat bottom basin Flat bottom basin Ambiguous Flat bottom Basin, with poss. shelf Deep & broad Basin Complex basin Concave bottom Concave bottom Concave bottom Soil Matrix Non-stratified; fine-grained Non-stratified; fine-grained Coarse-grained lower fill; fine-grained upper fill Stratified, varying textures Coarse-grained lower fill; fine-grained upper fill Non-Stratified; B-horizonlike Coarse-grained Coarse-grained Coarse-grained lower fill, with fine-grained upper fill Stratified, varying textures Non-stratified; fine-grained Various jumbled textures Various jumbled textures Various jumbled textures Charcoal-rich; various textures Non-stratified; fine-grained Non-stratified; fine-grained Vaguely stratified; generally fine-grained Varying textures Varying textures Non-stratified; fine-grained Non-stratified; fine-grained; charcoal-rich; oxidized bottom Non-stratified; fine-grained Notes Further exposed in a MT lateral Further exposed in a MT lateral Further exposed in a MT lateral Buried A1 and Bhorizons Further exposed in a MT lateral Feature remnant truncated by Fe.304? Further exposed in a MT lateral Buried A1 and Bhorizons Small soil anomaly Part of series Part of series Part of series Possible root burn or turbated pit hearth remnant Possibly part of Fe 315 Further exposed in a MT lateral Possible channel and/or DSF amalgam Part of series Part of series Further exposed in a MT lateral Further exposed in a MT lateral Possibly part of series DSF DSF DSF RC MT Fe.# Trench Wall E E W S N S 300 15 301 302 303 14 304 N 305 306 S N S S S N SE SE SE SE SE SE SE NW NW NW NW NW NW DSF U DSF RC U DSF DSF DSF U DSF DSF U DSF DSF DSF DSF DSF 13 307 308 309 310 311 312 313 314 17 315 316 323 317 318 319 320 321 149 Vertical distributions of three widely recovered index materials are addressed: charred wood, calcined bone, and lithic debitage. Finding these in mutually associated stratigraphic concentrations is assumed to closely reflect anthropogenic discard patterns. The first is charred wood. This material was selected because occupational layers at ancient human habitation sites are often marked by higher concentrations of charred wood, usually the byproduct of fuelwood burning. Due to time constraints, other charred botanicals from TU contexts, such as nutshells and seeds, were not submitted for identification and quantification by an expert paleobotanist, thus they are omitted from this analysis. The second index material is calcined bone, all of which was identified by faunal analyst Randy Nokes. Though his analysis includes more detailed information, these tables only list the count, type, and weight of identified calcined bone lots for brevity. It is assumed that all calcined bone is anthropogenic, as natural surface fires rarely generate the conditions necessary to reduce bone to its mineral constituents. The third index material consists of lithics. The author identified all lithics from TU flotation samples, including microdebitage. All lithics are accounted for in the forthcoming tables with the exception of quartz microdebitage (pieces under 1-cm long) because this is frequently difficult to distinguish from naturally occurring angular quartz and plow-generated shatter. Data consists primarily of counts and material types; weights are not included. All lithic materials are debitage unless otherwise noted in the table. 150 This analysis begins with TUs excavated in MT-13. MT-13 was excavated near the southern edge of the artifact concentration referred to as Locus 2 by Brian Jones. This trench exposed a cross-section of the remnant glaciofluvial channel, which appears to be bifurcated in this locality (Figure 9-1). The western sub-channel was designated Fe.307, while the eastern sub-channel was designated Fe.308. designated Fe.306, was identified. One robust DSF, TUs were excavated in all of these features to Quantities of chipping debris investigate vertical distributions of artifacts therein. recovered during test pit excavation were often substantial at Locus 2, exceeding 40 pieces per square meter in some test pits. Figure 9-1. Profile of MT-13 showing location of TUs. TU-9 was excavated to sample a central portion of the DSF designated Fe.306 (see Figure 9-1). This feature contained trace quantities of charred wood from its middleto-upper portions and one piece of calcined bone (Table 9-2). Four pieces of lithic debitage were recovered from the overlying plowed A-horizon and its interface with feature soil. TU data indicates no concentration of cultural material at the bottom of Fe.306 (a DSF). 151 Table 9-2. TU-9 soil flotation summary. Depth (cmbd) 12-20 20-30 30-40 40-45 45-50 50-60 60-70 70-80 80-90 90-100 100-110 110-120 Soil Type Ap Ap Ap Ap Ap,Fe.306 Fe.306 Fe.306 Fe.306 Fe.306 Fe.306 Fe.306 Fe.306,C Sample Vol. (l) 11 13 15 9 9 18 16 14 18 9 10 8 Charred Wood (g) -0.12 0.10 -0.11 0.39 0.41 0.20 0.51 0.12 --Calcined Bone ------1 vertebrate (0.005g) -----Lithic Artifacts --1 argillite, 2 quartzite -1 quartzite -------- TU-10 was excavated to sample the edge of a western sub-channel (Fe.307) of the remnant glaciofluvial channel (see Figure 9-1). Trace quantities of charred wood were recovered from the plowed A-horizon and the upper half of Fe.307 matrix (Table 9-3). A few pieces of lithic debitage were recovered from the plowed A-horizon and Fe.307 matrix, but not in any significant concentrations. No artifacts were recovered from the bottom of Fe.307 in this locality. Table 9-3. TU-10 soil flotation summary. Depth (cmbd) 40-50 50-60 60-70 65-70 70-80 80-90 90-100 100-110 110-120 120-130 130-140 140-150 Soil Type Ap Ap Ap Fe.307 Fe.307 Fe.307 Fe.307 Fe.307 Fe.307 Fe.307,C Fe.307,C C Sample Vol. (l) 18 18 21 4 20 16 16 12 17 8 15 16 Charred Wood (g) -0.15 0.25 0.28 1.06 0.89 1.80 -----Calcined Bone -1 unid. (0.129g) 1 vertebrate (0.012g) ---------Lithic Artifacts 1 argillite, 1 quartzite, 1 jasper -1 chert --1 argillite, 1 quartz --1 quartzite ---- TU-11 was excavated to sample a central portion of a western sub-channel (Fe.307) of the remnant glaciofluvial channel (see Figure 9-1). The bottom of the channel could not be sampled here because it was deeper than the bottom depth of the 152 trench. Trace quantities of charred wood were recovered from the plowed A-horizon and most of the Fe.307 matrix, with a significant increase in abundance in the lowest level that was sampled (Table 9-4). Two pieces of lithic debitage were recovered from the plowed A-horizon and one additional piece was recovered from Fe.307 matrix, but these did not constitute significant concentrations. recovered as well, all from the plowed A-horizon. Three pieces of calcined bone were Table 9-4. TU-11 soil flotation summary. Depth (cmbd) 40-50 50-60 60-70 70-75 75-85 85-90 90-100 100-110 110-120 120-130 130-140 140-150 150-160 Soil Type Ap Ap Ap Ap Buried A Fe.307 Fe.307 Fe.307 Fe.307 Fe.307 Fe.307 Fe.307 Fe.307 Sample Vol. (l) 14 22 26 ? 22 1 21 19 15 13 10 9 16 Charred Wood (g) 0.10 0.16 0.32 0.16 0.26 -0.99 0.48 0.25 0.27 0.25 0.61 2.71 Calcined Bone --1 med-to-sm mammal (0.01g) 1 vertebrate (0.081g) 1 med-to-lg mammal (0.119g) ----------1 chert 1 chert ---1 argillite ------Lithic Artifacts TU-12 was excavated to sample a central portion of an eastern sub-channel (Fe.308) of the remnant glaciofluvial channel (see Figure 9-1). The bottom of the channel could not be sampled here because it was deeper than the bottom of the trench. Trace quantities of charred wood were recovered from the plowed A-horizon and most of the Fe.308 matrix (Table 9-5). Two pieces of lithic debitage were recovered from the plowed A-horizon and one additional piece was recovered from Fe.307 matrix, but these did not constitute significant concentrations. Limited quantities of lithic debitage and 153 claimed bone fragments were intermittently detected in the soil column, but no particularly high concentrations were evident. Table 9-5. TU-12 soil flotation summary. Depth (cmbd) 43-50 50-60 60-70 70-74 74-80 80-90 90-100 100-110 110-120 120-128 128-140 140-150 150-160 160-170 Soil Type Ap Ap Ap Ap Buried A Buried A Fe.308 Fe.308 Fe.308 Fe.308 Fe.308 Fe.308 Fe.308 Fe.308 Sample Vol. (l) 16 29 26 ? 13 27 18 23 11 14 21 16 8 17 Charred Wood (g) 0.12 0.45 0.14 -0.37 0.91 0.29 1.12 0.80 1.40 1.18 0.73 0.22 0.29 Calcined Bone ----1 vertebrate (0.015g) ---2 mammal (0.037g) -1 mammal (0.008g) -1 mammal (0.012g) -Lithic Artifacts 1 argillite, 1 quartzite 2 quartz, 1 quartzite 3 quartzite, 1 unid. lithic --1 chert, 1 quartzite 1 quartzite No data 1 chert -3 quartz, 5 quartzite, 1 unid. lithic 1 quartz --- MT-14 exposed a cross-section of the remnant glaciofluvial channel, which was designated Fe.303 in this locality (Figure 9-2). A robustly visible DSF, designated Fe.304, was also exposed. TUs were excavated to investigate artifact distributions in both of these features. An additional TU was excavated in a non-feature context to provide a control sample. MT-14 was excavated through an area of Preston Plains that exhibited a very minor concentration of lithic debitage during test pit excavation, with maximal concentrations not exceeding 20 pieces per square meter. Figure 9-2. Profile of MT-14 showing location of TUs. 154 TU-4 was excavated in a non-feature context to provide a control sample (see Figure 9-2). Here, the plowed topsoil transitioned directly to C-horizon soils. A small quantity (n=4) of artifacts were recovered from the plowed topsoil (Table 9-6). Table 9-6. TU-4 soil flotation summary. Depth (cmbd) 60-70 70-78 78-90 90-100 100-110 110-120 120-130 Soil Type Ap Ap C C C C C Sample Vol. (l) 28 23 24 14 20 15 20 Charred Wood 0.37 -0.55 ----Calcined Bone -1 mammal (0.034g) ------3 quartzite -----Lithic Artifacts TU-5 was excavated to sample the central portion of the DSF designated Fe.304 (see Figure 9-2). No artifact concentration was found on the bottom of Fe.304 (a DSF). A small quantity of materials, consisting of chipping debris, calcined bone, and charred wood, were recovered from the plowed A-horizon and the uppermost portion of Fe.304 (Table 9-7). Table 9-7. TU-5 soil flotation summary. Depth (cmbd) 60-70 70-80 80-90 90-100 100-110 110-120 110-120 120-130 130-140 140-150 Soil Type Ap Ap,Fe.304 Fe.304 Fe.304 Fe.304 Fe.304 Fe.304 Fe.304,C Fe.304,C C Sample Vol. (l) 17 18 18 17 15 4 13 6 9 8 Charred Wood (g) 0.16 0.10 1.40 0.60 ----No data -Calcined Bone -2 lg mammal (0.109g) 1 vertebrate (0.136g) 2 mammal (0.016g) 1 med-to-lg mammal (0.151g) -------2 quartz 1 quartz --1 quartzite ----Lithic Artifacts 155 TU-6 was excavated to sample an eastern margin of the section of remnant glaciofluvial channel designated Fe.303 (see Figure 9-2). It is not known for certain if the bottom of the channel was exposed in this area because the stratigraphy here was somewhat ambiguous. Trace amounts of charred wood were recovered, generally from A-horizon contexts (Table 9-8). Four fragments of calcined bone were recovered from the upper portion of the plowed A-horizon. A small quantity of lithics were recovered (n=8) that were confined to the Buried A and B-horizons. Table 9-8. TU-6: soil flotation summary. Depth (cmbd) 42-50 50-60 60-70 70-74 74-80 80-90 90-93 93-100 100-110 110-120 120-130 130-140 140-147 147-150 150-160 160-170 Soil Type Ap Ap Ap Ap Buried A Buried A Buried A Fe.303/B Fe.303/B Fe.303 Fe.303 Fe.303 Fe.303 Fe.303/C? Fe.303/C? Fe.303/C? Sample Vol. (l) 17 ? 19 ? 11 30 6 14 21 20 22 18 20 ? 23 16 Charred Wood (g) -0.20 -0.15 -0.78 0.69 0.36 No data -------Calcined Bone 1 vertabrate (0.062g) 3 mammal (0.186g) -------------------1 quartz 1 quartzite 1 quartz 5 argillite -------Lithic Artifacts TU-7 was excavated to sample a western margin of the remnant glaciofluvial channel exposure designated Fe.303 (see Figure 9-2). There was a trace amount of charred wood recovered throughout the upper and medial portions of the soil column, with the most recovered in the buried A-horizon (Table 9-9). Only one artifact was recovered, a calcined bone fragment from a medial portion of the channel fill. 156 Table 9-9. TU-7 soil flotation summary. Depth (cmbd) 40-50 50-60 60-70 70-75 75-80 80-90 90-100 100-110 110-120 120-130 130-140 140-150 150-160 Soil Type Ap Ap Buried A Buried A Fe.303 Fe.303 Fe.303 Fe.303 Fe.303 Fe.303 Fe.303 Fe.303 Fe.303,C Sample Vol. (l) 17 18 20 3 4 14 13 10 11 10 12 9.5 10 Charred Wood (g) -0.11 2.18 0.46 0.91 0.65 0.41 0.74 0.42 0.90 ---Calcined Bone ------1 med-to-lg mammal (0.116g) ------No data ------------Lithic Artifacts MT-15 was excavated through an area of the site that exhibited an extremely thin scattering of lithic debitage, with abundances never exceeding 8 pieces per square meter. This trench exposed a cross-section of three DSFs, one of which (Fe.302) was sampled by excavating a TU (Figure 9-3). TUs were excavated to collect control samples through an ice wedge cast (labeled “C-horizon soil structure” on profile) and a profile that contained no soil anomalies. Figure 9-3. Profile of MT-15 showing location of TUs. TU-1 was excavated through an ice wedge cast that was not assigned a feature number (see Figure 9-3). Trace amounts of lithic debitage and calcined bone were recovered from overlying topsoil, and calcined bone fragments were recovered from the 157 thinly developed overlying B-horizon (Table 9-10). The matrix filling the ice wedge cast, suspected to be yedoma, feature felt dense and homogenous during excavation and is assumed to be a pristine Late Pleistocene deposit. It contained no artifacts. Table 9-10. TU-1: soil flotation summary. Depth (cmbd) 20-30 30-40 40-48 48-50 50-60 60-70 70-80 80-90 90-100 100-110 110-120 120-130 Soil Type Ap Ap Ap B? B?,C B?,C C C C C C C Sample Vol. (l) 16 ? 14 8 19 15 21 15 No data 17 9 25 Charred Wood (g) 0.13 -0.30 0.14 0.27 0.24 0.12 ----Calcined Bone 1 vertebrate (0.067g) 1 mammal (0.044g) -3 mammal (0.038g) -------Lithic Artifacts 1 argillite, 2 chert ------------ TU-2 was excavated at the southern edge of Fe.302 (a DSF) in a location that generally exhibits non-feature stratigraphy (see Figure 9-3). A single piece of calcined bone was recovered from plowed topsoil, in addition to a single piece of lithic debitage (Table 9-11). A second piece of lithic debitage was recovered from B-horizon-like soil. Table 9-11. TU-2 soil flotation summary. Depth (cmbd) 20-30 30-40 40-46 46-50 50-60 60-70 70-80 80-90 Soil Type Ap Ap Ap B? B? B? C C Sample Vol. (l) 16 24 19 7 21 18 12 15 Charred Wood (g) 0.10 -0.12 -0.22 ---Calcined Bone 1 med-to-lg mammal (0.015g) -------1 quartz ---1 quartz ---Lithic Artifacts 158 TU-3 was excavated through the DSF designated Fe.302. (see Figure 9-3). Two artifacts were recovered from the plowed A-horizon and its interface with Fe.302 matrix (Table 9-12). No artifact concentration was found on the bottom of Fe.302 (a DSF). Table 9-12. TU-3 soil flotation summary. Depth (cmbd) 20-30 30-40 40-47 47-50 50-60 60-70 70-80 80-90 90-100 100-110 110-120 Soil Type Ap Ap Ap Ap,B,Fe.302 B,Fe.302 B,Fe.302 B,Fe.302 Fe.302 Fe.302 Fe.302,C C Sample Vol. (l) 23 32 21 13 28 23 ? 16 15 9 11 Charred Wood (g) 0.19 -0.37 1.10 2.80 3.40 0.71 0.78 ---Calcined Bone -------------1 quartzite 1 quartzite -------Lithic Artifacts MT-17 was excavated outside of the revised Preston Plains Energy Center Project area, near the edge of Avery Pond (Figure 9-4). During the 2007 field school, Phase I test pits excavated at 10-m intervals in this vicinity resulted in the recovery of Late Archaic period lithic tools and debitage. Feature soils were also identified in three of these test pits. In 2008, MT-17 was excavated in alignment with these three test pits to expose feature soils for further examination and sampling. The exposed profiles revealed a variety of overlapping soil structures that were designated as features. Some of these exhibited robust, basin-like morphology and were classified as DSFs (Fe.319, 320, 315). Features in the northeastern end of the trench were tentatively classified as DSFs, though they extensively overlapped and constitute a greater complex (Fe.310, 311, 312, 317, 318). A disturbed pit hearth or root burn was also revealed (Fe.313), in addition to a massive area of feature soil on the southwestern end of the trench that appeared largely non-stratified (Fe.316, 323). TUs excavated 159 through several of these contexts, in addition to a control (non-feature) context, are hereby reviewed. Figure 9-4. Profile of MT-17 showing location of TUs. TU-13 was excavated through the DSF designated Fe.320 (see Figure 8-4). Its basin morphology was confirmed with the excavation of a machine trench lateral. Three pieces of lithic debitage and a steep-bitted scraper were recovered from the plowed Ahorizon, in addition to several pieces of calcined bone and traces of charred wood (Table 160 9-13). Similar proportions of calcined bone and charred wood were recovered from the central portion of Fe.320. However, no artifact concentration was found on the bottom of Fe.320 (a DSF). The light fraction of the float sample from the lowest level was lost in the laboratory; therefore this statement is only qualified by the absence of lithic materials in the heavy fraction. . Table 9-13. TU-13 soil flotation summary. Depth (cmbd) 90-100 100-110 Soil Type Ap Ap Sample Vol. (l) ? 22 Charred Wood (g) 0.05 0.42 Calcined Bone -1 med-to-lg mammal (0.152g) 2 vertebrate (0.061g) 2 mammal (0.024g) 1 lg mammal (0.046g) --1 mammal (0.077g) 1 lg mammal (0.319g) 1 vertebrate (0.037g) 1 med-to-lg mammal (0.06g) 1 sm mammal (0.031g) -----1 argillite, 1 quartzite 2 quartz (includes 1 steep-bitted scraper) --------Lithic Artifacts 110-123 123-130 130-140 141-152 152-160 156-167 162-170 170-180 180-190 Ap,Fe.320 Fe.320 Fe.320 Fe.320 Fe.320 Fe.320 C C C 17 4 17 10 12 No data 8 14 12 0.29 0.21 4.19 2.30 1.64 ---- TU-14 was excavated through the DSF designated Fe.319 (see Figure 9-4). Its basin morphology was confirmed through the excavation of a machine trench lateral. The plowed topsoil yielded trace amounts of lithic debitage, calcined bone, and charred wood (Table 9-14). Fe.319 matrix yielded moderate quantities of charred wood, and four additional artifacts from central portions. However, no artifact concentration was found on the bottom of Fe.319 (a DSF). 161 Table 9-14. TU-14 soil flotation summary. Depth (cmbd) 90-100 100-110 115-120 120-130 130-140 140-150 150-160 160-170 170-180 180-185 185-200 Soil Type Ap Ap Fe.319 Fe.319 Fe.319 Fe.319 Fe.319 Fe.319 Fe.319 Fe.319 C Sample Vol. (l) 23 17 3 15 12 20 12 13 9 4 13 Charred Wood (g) 0.14 0.15 0.41 4.06 6.54 -3.54 3.85 3.47 0.45 0.07 Calcined Bone 4 med-to-lg mammal (0.081g) 1 mammal (0.017g) --2 mammal (0.031g) -------2 quartz --1 quartzite -1 quartz -----Lithic Artifacts TU-15 was excavated in a non-feature context to provide a control sample (see Figure 9-4). Here, the plowed topsoil transitioned directly to C-horizon soils (Table 915). A small quantity (n=4) of artifacts were recovered from the plowed topsoil, consisting of nine pieces of lithic debitage, two fragments of calcined bone, and traces of charred wood. Table 9-15. TU-15 soil flotation summary. Depth (cmbd) 83-90 90-100 100-110 110-120 120-129 129-140 Soil Type Ap Ap Ap C C C Sample Vol. (l) 12 ? 16 10 10 19 Charred Wood (g) 0.13 -0.26 0.15 ----1 vertebrate (0.076g) 1 turtle (0.069g) ---Calcined Bone Lithic Artifacts 2 chert, 1 quartz, 2 quartzite, 1 unid. lithic No data 1 argillite, 1 chert, 1 quartz ---- TU-16 was excavated through the DSF designated Fe.315 (see Figure 9-4). This feature was identified as a basin, which appears to have a shallower “shelf” extending to the southwest. This morphology was confirmed through the excavation of a machine trench lateral. The plowed topsoil and upper-to-middle portions of Fe.315 matrix yielded relatively consistent amounts of chipping debris, calcined bone, and charred wood (the 162 seemingly larger quantities recovered from 140-170 cmbs reflect larger sample volumes.) The only material recovered from the bottom of Fe.319 matrix was diminishing quantities of wood charcoal (Table 9-16). No artifact concentration was found on the bottom of Fe.315 (a DSF). Table 9-16. TU-16 soil flotation summary. Depth (cmbd) 104-110 110-120 120-130 130-136 136-140 140-150 150-160 160-170 170-177 177-180 180-190 190-198 198-210 Soil Type Ap Ap Ap Ap Fe.315 Fe.315 Fe.315 Fe.315 Fe.315 Fe.315 Fe.315 Fe.315,C C Sample Vol. (l) 18 30 16 20 7 38 18 35 10 14 19 15 22 Charred Wood (g) 0.12 0.34 0.24 5.44 2.00 7.91 7.61 8.29 3.48 5.50 3.53 2.70 0.22 Calcined Bone 1 mammal (0.022g) 1 vertebrate (0.074g) -1 mammal (0.01g) 2 vertebrate (0.014g) -2 mammal (0.201g) 2 med-to-lg mammal (0.08g) 1 sm-to-med mammal (0.019g) ------Lithic Artifacts 1 unid. lithic 1 quartzite, 3 rhyolite 1 felsite -1 quartz 2 quartz, 1 quartzite, 10 rhyolite 1 quartz, 2 quartzite 2 quartzite, 2 rhyolite (includes 1 biface fragment) 1 quartz unifacial core-scraper ----- TU-17 was excavated to sample part of the massive area of feature soil on the southwestern end of the trench (Figure 9-4). This feature was designated Fe.316 on the southern trench wall, and its bottom could not be sampled here because it was deeper than the trench. Chipping debris was recovered from plowed A-horizon soil and upper portions of Fe.316 matrix, and diminished in abundance as depth increased (Table 9-17). Calcined bone was recovered from plowed A-horizon soil and the uppermost portion of Fe.316. Traces of charcoal were recovered from all levels, with the highest amounts coming from central portions of the feature. 163 Table 9-17. TU-17 soil flotation summary. Depth (cmbd) 120-130 130-140 140-152 152-160 160-170 170-180 180-190 190-200 200-210 210-220 Soil Type Ap Ap Ap Fe.316 Fe.316 Fe.316 Fe.316 Fe.316 Fe.316 Fe.316 Sample Vol. (l) 21 23 20 18 ? 22 ? 23 17 9 Charred Wood (g) 0.23 0.40 0.28 0.69 2.09 2.79 1.84 6.16 4.59 1.86 Calcined Bone 3 mammal (0.025g) 1 vertebrate (0.118g) 3 med-to-lg mammal (0.297g) 1 mammal (0.016g) 1 vertebrate (0.006g) ------Lithic Artifacts 2 quartz, 1 rhyolite 2 argillite, 2 quartz, 1 unid. lithic 1 argillite, 3 quartzite, 2 rhyolite 1 argillite 1 quartz, 2 quartzite 1 quartz, 3 quartzite 1 quartzite 1 quartzite -1 quartzite TU-18 was excavated through a small feature designated Fe.313, which may constitute a turbated pit hearth or root burn (see Figure 9-4). Lithic debitage and calcined bone were recovered from the plowed A-horizon above Fe.313 (Table 9-18). While the upper level of Fe. 313 contained two pieces of lithic debitage, its matrix was distinguished by its large quantity of charred wood. Table 9-18. TU-18 soil flotation summary. Depth (cmbd) 80-90 90-100 100-110 110-120 120-130 130-140 140-150 150-160 Soil Type Ap Ap Ap Ap Fe.313 Fe.313 Fe.313 C Sample Vol. (l) 16 ? 15 11 24 17 18 16 Charred Wood (g) 0.39 0.43 0.35 1.20 85.20 42.60 0.24 -Calcined Bone -2 mammal (0.125g) 3 vertebrate (0.103g) 3 lg mammal (0.107g) 1 med-to-lg mammal (0.031g) 1 vertebrate (0.104g) ------2 chert 1 argillite, 1 quartz 1 quartzite 1 argillite, 1 chert ---Lithic Artifacts TU-19 was excavated through Fe.310, which was tentatively identified as a DSF belonging to a greater complex of feature soil (Figure 9-4). While artifacts were recovered from throughout the plowed A-horizon and Fe.310 matrix, the highest quantity of wood, calcined bone, and lithic debitage all came from an upper layer (40-50 cmbd) of 164 Fe.310 matrix (Table 9-19). No artifact concentration was found on the bottom of Fe.310, a DSF. Table 9-19. TU-19 soil flotation summary. Depth (cmbd) 20-30 30-35 35-40 Soil Type Ap Ap Fe.310 Sample Vol. (l) 20 17 2 Charred Wood (g) -0.40 0.13 Calcined Bone 7 vertebrate (0.007g) 7 mammal (0.123g) 3 large mammal (0.143g) 3 med-to-lg mammal (0.288g) 6 mammal (0.19g) 1 med-to-lg mammal (0.029g) 100 mammal (1.2g) 19 med-to-lg mammal (1.175g) 1 lg mammal (0.225g) 4 turtle (0.09g) 20 vertebrate (0.384g) 4 med-to-lg mammal (0.083g) -1 mammal (0.048g) 3 med-to-lg mammal (0.072g) 5 vertebrate (0.106g) 2 med-to-lg mammal (0.037g) --1 argillite 1 argillite, 1 chert 3 quartzite Lithic Artifacts 40-50 Fe.310 14 1.75 1 chert, 2 quartz, 7 quartzite 50-60 60-70 70-80 80-90 90-100 100-110 Fe.310 Fe.310 Fe.310 Fe.310,C C C 20 18 18 10 10 5 0.34 0.46 1.20 ---- -2 quartz, 1 quartzite 1 chert, 2 quartz, 2 quartzite 1 quartz --- TU-20 was excavated through Fe.323, which was tentatively identified as a DSF belonging to a greater complex of feature soil (see Figure 9-4). The highest quantities of calcined bone and lithic debitage were recovered from the plowed A-horizon, while charred wood was most abundant in the upper half of Fe.312 matrix (Table 9-20). Several pieces of lithic debitage and one calcined bone fragment were recovered from the central portion of Fe.310 matrix. However, no artifact concentration was found on the bottom of Fe.323, a DSF. TU-21 was excavated through Fe.311, which was tentatively identified as a DSF belonging to a greater complex of feature soil (see Figure 9-4). Charred wood, calcined bone, and lithic debitage were present in virtually every level of TU-21 (Table 9-21). The highest concentrations of these materials was generally in the middle/central levels 165 of Fe.311 matrix, with quantities diminishing towards the bottom of the feature. Therefore, no artifact concentration was found on the bottom of Fe.311, a DSF. Table 9-20. TU-20 soil flotation summary. Depth (cmbd) 46-50 50-60 60-69 69-80 80-90 90-100 100-111 111-120 Soil Type Ap Ap Ap Fe.312 Fe.312 Fe.312 Fe.312 C Sample Vol. (l) 10 16 16 16 17 14 16 14 Charred Wood (g) --0.40 2.50 2.20 1.40 -0.17 Calcined Bone -4 med-to-lg mammal (0.107g) 1 turtle (0.025g) 1 vertebrate (0.092g) 2 med-to-lg mammal (0.243g) 1 mammal (0.086g) -1 mammal (0.028g) ----7 chert, 3 quartzite 2 chert, 3 quartz, 3 quartzite -2 argillite, 1 quartz, 4 quartzite 1 quartz, 2 quartzite --Lithic Artifacts Table 9-21. TU-21 soil flotation summary. Depth (cmbd) 23-30 30-40 Soil Type Ap Ap Sample Vol. (l) 12 20 Charred Wood (g) -1.20 Calcined Bone (and Shell*) 1 vertebrate (0.001g) 3 mammal (0.026g) 15 mammal (0.196g) 1 lg mammal (0.124g) 7 med-to-lg mammal (0.261g) 3 sm-to-med mammal (0.099g) 34 mammal (0.413g) 7 med-to-lg mammal (0.249g) 2 lg mammal (0.163g) 3 med-to-lg mammal (0.183g) 2 vertebrate (0.092g) 24 mammal (0.333g) 5 lg mammal (0.745g) 9 med-to-lg mammal (0.279g) 13 vertebrate (0.69g) 17 mammal (0.439g) 1 lg mammal (0.125g) 5 med-to-lg mammal (0.151g) 1 med-to-lg mammal (0.072g) 2 lg mammal (0.149g) 3 mammal (0.117g) 5 lg mammal (0.263g) 1 shell* (0.001g) -Lithic Artifacts 5 quartz 1 chert, 1 quartz, 3 quartzite (includes 1 Narrow-Stemmed point) 2 quartz, 8 quartzite (includes 1 Brewerton/Vosburg point corner fragment) 40-50 Fe.311 16 0.60 50-60 Fe.311 17 1.60 1 chert, 2 quartzite 60-70 Fe.311 20 3.60 1 chert 70-80 80-90 90-100 100-110 Fe.311 Fe.311 Fe.311 C 22 16 13 2 4.80 0.31 0.70 -- 2 chert, 5 quartz (includes 1 biface), 7 quartzite 2 quartzite --- TU-22 was excavated through Fe.312, which was tentatively identified as a DSF belonging to a greater complex of feature soil (see Figure 9-4). Charred wood, calcined bone, and lithic debitage was most heavily concentrated in the mid-to-upper portions of 166 Fe.312 matrix (Table 9-22). Quantities of artifacts diminished towards the bottom of the feature. Therefore, no artifact concentration was found on the bottom of Fe.312, a DSF. Table 9-22. TU-22 soil flotation summary. Depth (cmbd) 37-40 40-50 50-56 56-60 Soil Type Ap Ap Ap Fe.312 Sample Vol. (l) ? 15 12 5 Charred Wood (g) --0.50 0.34 Calcined Bone (and Shell*) ----3 turtle (0.034g) 14 mammal (0.299g) 5 sm-to-med mammal (0.128g) 6 med-to-lg mammal (0.223g) 1 lg mammal (0.131g) 5 vertebrate (0.097g) 3 turtle (0.699g) 10 mammal (0.154g) 1 vertebrate (0.121g) 1 mammal (0.005g) 1 shell* (0.09) -Lithic Artifacts -1 argillite, 2 quartzite 1 chert, 3 quartz, 1 quartzite, 1 unid. lithic -- 60-70 Fe.312 16 7.60 1 chert, 4 quartz, 10 quartzite 70-80 80-90 90-107 107-120 Fe.312 Fe.312 Fe.312 C 16 16 18 13 5.78 2.40 0.84 0.15 1 argillite, 1 rhyolite 1 chert, 3 quartz, 6 quartzite, 1 rhyolite 1 chert -- TU-23 was excavated through Fe.314, which may be part of a DSF (see Figure 94). Trace quantities of charred wood, calcined bone, and lithic debitage were recovered from plowed A-horizon soil and Fe.314 matrix (Table 9-23). Table 9-23. TU-23 soil flotation summary. Depth (cmbd) 95-100 100-110 110-120 120-128 128-140 140-150 150-160 160-170 Soil Type Ap Ap Ap Ap Fe.314 Fe.314 C C Sample Vol. (l) 14 29 29 17 26 23 24 ? Charred Wood (g) 0.24 0.33 1.08 0.39 3.10 1.30 0.04 0.38 Calcined Bone -1 med-to-lg mammal (0.025g) 3 vertebrate (0.024g) 1 vertebrate (0.004g) 1 mammal (0.24g) ---Lithic Artifacts -2 quartzite 2 chert 1 argillite, 1 quartz -2 quartzite --- TU-24 was excavated through what it thought to be the shallower “shelf” of Fe.315, a DSF (see Figure 9-4). Trace quantities of charred wood, calcined bone, and 167 lithic debitage were recovered from this TU (Table 9-24). No artifact concentration was found on the bottom of Fe.315, a DSF. Table 9-24. TU-24 soil flotation summary. Depth (cmbd) 102-110 110-120 120-130 130-140 140-150 150-160 160-170 160-170 170-180 Soil Type Ap Ap Ap Ap Fe.315 Fe.315 Fe.315 C C Sample Vol. (l) ? 21 22 28 25 23 10 11 18 Charred Wood (g) -0.33 -0.79 3.42 1.40 0.12 --Calcined Bone ---1 mammal (0.016g) -----Lithic Artifacts -2 chert, 1 quartz -1 quartz 2 quartz 1 quartz ---- TU-25 was excavated to sample a second portion of the massive area of feature soil on the southwestern end of the trench (see Figure 9-4). The feature bottom was reached here, at about 214 cmbd. While the plowed A-horizon was virtually devoid of charred wood here, significant quantities were recovered from middle-to-upper portions of Fe.316 matrix (Table 9-25). Small quantity of chipping debris and a trace quantity of calcined bone was recovered from the middle-to-upper portions of Fe.316 matrix. Artifact quantities diminished in the deepest layers of feature matrix. Table 9-25. TU-25 soil flotation summary. Depth (cmbd) 103-110 110-120 120-130 130-142 142-150 150-160 160-170 170-180 180-190 190-200 200-210 210-214 Soil Type Ap Ap Ap Ap Fe.316 Fe.316 Fe.316 Fe.316 Fe.316 Fe.316 Fe.316,C Fe.316,C Sample Vol. (l) 8 20 22 4 20 22 30 17 24 24 20 21 Charred Wood (g) ----6.36 7.50 5.63 Not weighed 4.58 2.04 1.50 -Calcined Bone ---1 lg mammal (0.118g) -1 vertebrate (0.005g) 2 vertebrate (0.008g) -1 vertebrate (0.004g) ---Lithic Artifacts -2 rhyolite -1 quartz 2 quartz 1 quartz, 1 quartzite, 1 rhyolite 2 quartzite 1 quartz, 1 quartzite 2 quartz, 1 quartzite -1 quartz -- 168 9.2. Testing Results for DSFs Identified in Locus I Excavation Block Seven DSF from Locus I are hereby analyzed to further test the hypothesis that artifact concentrations are not present on the bottoms of DSFs at Preston Plains. The author has selected only one class of artifacts to use for this analysis because it was the most consistently recovered from all contexts – lithics measuring at least 1-cm (ie. 1-cm long in at least one dimension). The data used to test this hypothesis is presented as a series of analytical graphics. These show lithic artifact counts from seven 50-cm wide trench sections superimposed onto adjacent stratigraphic profiles of DSFs. Each number on a profile represents the total number of lithic artifacts recovered from a 10-cm range of a given EU quadrant, and is superimposed on the approximate center of that quadrant. By reviewing these lithic abundances relative to depth and strata, one may easily confirm or deny the presence of an artifact concentration on the bottom of a DSF, in addition to recognizing potential concentrations in medial or upper strata. It is reasonable to assume that lithics of this size were consistently recovered because they were easily visible to archaeological technicians using a ¼ inch screen mesh, and they were easily recognized in flotation samples. They closely reflect cultural activity in a way that is quantifiably meaningful due to their high rate of preservation and resistance to fragmentation. Microdebitage (shatter, microflakes, and flake fragments) is discounted in this analysis because it is assumed to have been recovered less consistently. This reveals the bias that concentrations of microdebitage, which could form through cultural agency or natural size-sorting processes, would not be evident in this analysis. 169 Heat-altered organic artifact distributions (ie. bone, seeds/nuts) are discounted here. The majority of these materials are fragmented to less than 1-cm, which are only effectively recovered from flotation samples. Screening efforts in the field likely missed the majority of such materials. And while soil from discrete features was floted, most of the soil from DSF matrices was only screened. In sum, heat-altered organic artifacts were not collected in a way that lends itself to rigorous spatial analysis across the locus. DSF-1: The bottommost strata of this DSF were largely devoid of lithics (Figure 9-5). Moderate amounts of lithics occurred in its upper portions, with the highest concentrations near or at the topsoil interface. A relatively high concentration of lithics was recovered from within, and immediately surrounding, Feature 8, which are likely associated. Figure 9-5. Lithic distribution in a cross-section of DSFs 1 and 2. DSF-2: This DSF contained a moderate concentration of lithics that generally followed a steeply inclined stratigraphic boundary running from the shallow end to the deep end (see Figure 9-5). The bottommost portions of this DSF yielded lithics in places 170 where this boundary “bottomed out” against C-horizon soil, but did not consistently exhibit distinct artifact concentrations across the C-horizon soil boundary. A very distinct vertical concentration of artifacts is evident immediately inside the deep end of this DSF (between W36.5-37). DSF-4: The matrix of this DSF was largely devoid of lithics, and no concentrations were evident on its bottom (Figure 9-6). A few lithics were recovered from the uppermost level of its matrix, and from immediately overlying topsoil. Figure 9-6. Lithic distribution in a cross-section of DSF 4. DSF-5: This DSF was also largely devoid of lithics, and contained no concentrations on its bottom (Figure 9-7). A few lithics were recovered from the uppermost level of its matrix, and a moderate quantity was recovered from immediately 171 overlying topsoil. There was a very concentrated stratum of charred wood near the bottom of this feature, but no artifact concentration. Figure 9-7. Lithic distribution in a cross-section of DSF-5. DSF-6: This DSF did not contain a consistent artifact layer along its bottommost contour (Figure 9-8). However, it contained a high concentration of lithics within a particular section - along the bottom of a stratigraphic contour called Feature 17. This contour was subtly evidenced by a soil color change (which is only approximated by a dotted line in Figure 9-8), and more robustly evidenced by a marked increase in lithic abundance and the presence of a large angular stone (visible in profile). Such stones do not naturally occur in Preston Plains soils, and this specimen is presumed to be a manuport. It was laying in an orientation consistent with the overall Feature 17 contour 172 and may be a primary deposition. During excavation, this was tentatively interpreted to be a remnant occupational surface. Figure 9-8. Lithic distribution in a cross-section of DSF-6. DSF-7: This DSF was virtually devoid of lithics (Figure 9-9). Only a few lithics were recovered from immediately overlying topsoil. DSF-9: This DSF was also virtually devoid of lithics (Figure 9-10). Only a few lithics were recovered from immediately overlying topsoil. DSF-3 and DSF-8 are not specifically reviewed here because complete stratigraphic profiles were never recorded. However, analytical graphics generated by Mandy Ranslow indicate that lithics in DSF-3 are contained mostly in its upper portions, and that DSF-8 is virtually devoid of lithics. Artifact distributions within DSF-10 cannot be addressed because only a small portion of this feature was excavated. 173 Figure 9-9. Lithic distribution in a cross-section of DSF-7. Figure 9-10. Lithic distribution in a cross-section of DSF-9. 9.3. Summary of Observations In conclusion, this chapter‟s data clearly supports the hypothesis that artifact concentrations are not present on the bottoms of DSFs at Preston Plains. Of the 18 DSFs sampled, none exhibited an artifact concentration on the bottom. Additional trends have 174 also been revealed that illuminate the spatial relationship of cultural material and soil types at Preston Plains. 1. Cultural material occurs most frequently in topsoil, and is only deeply buried in feature soil contexts or soil anomalies. The excavation of TU control samples confirms that artifacts are largely confined to the plowed A-horizon when there is no feature soil underneath. Cultural materials are found to extend to greater depths than surrounding topsoil when buried in certain DSF contexts, the remnant glaciofluvial channel, and in the massive, non-stratified feature soil area (Fe.316, 323) in the southwestern portion of MT17. These results solidly confirm expectations based on the geomorphological assessment of the site (presented in Chapter 6). 2. Some DSFs (Fe.302, 304, 306) are virtually devoid of artifacts, and are not situated within artifact-rich areas. Therefore, the locations of DSFs do not always correspond with the locations of prehistoric Native American activity areas as identified by artifact concentrations. 3. DSFs that contain notable amounts of cultural materials tend to have these distributed in their middle-to-upper levels, with few to no artifacts in their bottom levels. TU data reflects this pattern in relatively distinct DSFs (Fe.315, 319, 320) as well as DSFs that appear to overlap, forming a larger feature soil complex (Fe.310, 311, 312). Furthermore, lithic distributions from Locus I indicate that there were no lithic concentrations on the bottoms of the seven DSFs examined (DSF-1, 2, 4, 5, 6, 7, and 9). Some were nearly devoid of lithics (DSF-4, 5, 7, 9), and three of these had very few 175 lithics in overlying A-horizon soils (DSF-4, 7, 9). One had artifacts concentrated in its mid-to upper portions (DSF-1), while another had a stratigraphically plunging distribution of lithics in its interior (DSF-2). Another (DSF-6) had a stratigraphic concentration of lithics along a bowl-shaped contour that is thought to be anthropogenic. 176 Chapter 10. Local Tree Throw Examination Background research suggests that tree throws would have been significant formative agents at the Preston Plains Site when it was forested, prior to historic period clearing and farming. To further understand the dynamics of this agency, contemporary tree throws were excavated at nearby archaeological sites where similar soil conditions existed. The author identified these according to their diagnostic pit-and-mound topography, and though they are located at archaeological sites they are not considered to be archaeological features because they are not associated with artifact concentrations. Data from these analogous environmental contexts informs the model of DSF genesis at Preston Plains presented in the next chapter. The author selected three tree throws for study; one from the Avery Pond Site (CT 114-106) and two from the Sandy Hill Site (CT 72-97). He recognized them while supervising archaeological excavations at these sites, and investigated them by collecting descriptive, photographic, and metric information. Each tree throw was assigned a number (ie. Tree Throw 1, 2, 3). The pit-and-mound topography of each tree throw was digitally photographed and a scaled plan map was drawn that depicts topography and the orientation of trunk remnants, if still extant. A bisection line was established along the approximate longitudinal centerline of the tree throw. Along this line, a narrow trench, measuring approx 50-cm wide, was excavated that was of sufficient length and depth to expose each tree throw‟s subterranean elements. Profile exposures were digitally 177 photographed and drawn according to scale. Data collected from these investigations is reported and contextualized below. Three key statements can be made are made in view of this data. First, this data supports the assumption that trees of the same species thrown under similar conditions may leave consistent subsurface signatures. Second, subterranean elements of tree throws occurring in glacial outwash may resemble the D-shaped variety of DSFs. Third, tree throws in glacial outwash are likely to rotate soil masses, leaving inclined-to-inverted stratigraphy. 10.1. Tree Throw 1: Avery Pond Site The Avery Pond Site is located immediately north of the Preston Plains Site, and has been the subject of Phase I and Phase II archaeological investigations by the MPMRC that have revealed significant Late Archaic Period deposits in close spatial association with two DSFs (Ives 2007a, 2007b). The Avery Pond Site is situated on a geological feature known as a kame – an irregularly shaped mound of stratified sand and gravel deposited during glacial retreat. This kame was likely deposited in contact with an ice mass that formed the kettle hole of Avery Pond. The isolation of this kame in the swampland and its steeply sloping sides appear to have discouraged historic tilling/plowing, as no plowed A-horizons were encountered in test pits here. However, a mid-twentieth century cabin was built on the northeastern end of this landform to accommodate vacationers who likely enjoyed a view of Avery Pond from a relatively 178 private setting. Mixed conifer and broadleaf forest covers this landform, with the largest trees being white pines (Pinus strobus). The USDA classifies soil on this kame as Hinckley gravely sandy loam 15 to 35 percent slopes (Crouch 1983). “Typically, this Hinckley soil has a dark brown, gravelly sandy loam surface layer 2 inches thick. The subsoil is yellowish brown gravelly loamy sand 20 inches thick. The substratum is brownish yellow very gravelly coarse sand to a depth of 60 inches or more” (ibid: 20). This deep, excessively drained soil type occurs on stream terraces, outwash plains, kames, and eskers, and mapped areas are typically irregular in shape. Available water capacity is low and runoff is rapid. Steepness of slopes is a limiting factor for community development and cultivated crops. Tree Throw 1 was identified at the southwestern end of the site according to the presence of a pit and mound, and investigated in 2008. The relative age of this tree throw is unknown, and the trunk had completely decayed. The tree fall occurred in a southerly direction away from the edge of Avery Pond, perhaps in response to fall/winter gales generated by artic air masses. 179 Figure 10-1. Profile section drawing of Tree Throw 1, Avery Pond Site. Figure 10-2. Tree Throw 1, Avery Pond Site (10-cm increment scale rod). 180 In plan, the pit-and-mound pair are relatively D-shaped (Figure 10-1), which is a common characteristic of tree throws. The soil profile reveals a simple, bowl-shaped soil rotation with slightly inverted (ie. rotated past 90 degrees) stratigraphy that is consistent with trees that fall down slope (see Figure 10-1, Figure 10-2). Topsoil was transported downward and concentrated at the leeward end, while subsoil was lifted and concentrated at the windward end. A small cobble appears to have been displaced upward at the windward end of this feature, probably having been dislodged from the upper B2horizon. There is very little pedogenesis on the mound, probably due to constant erosion. The pit exhibits a thickened A-1 horizon that has probably been contributed to by windblown and washed-in silt and organics. 10.2. Tree Throws 2 and 3: Sandy Hill Site The Sandy Hill Site (CT 72-97), located at the southern end of Mashantucket‟s Great Cedar Swamp, has been the subject of data recovery investigations by the MPMRC that have vitally contributed to New England‟s early Holocene archaeology (Forrest 1999; Jones and Forrest 2003). The site is situated on a complex of glaciodeltaic sands that gently prograde to the south before becoming submerged beneath the swamp (Thorson and Webb 1991). These sands were deposited in contact with a remnant glacial ice mass, slumping after it melted. The USDA classified soils here as a Hinckley gravelly sandy loam with a 15 to 45% slope. The two tree throws examined here (Tree Throws 2 and 3) are on a section of white pine-dominated (Pinus strobus) forest. Topography here is interrupted at fairly 181 regular intervals by robust tree throw remnants. Degraded trunks are extant on many of these, and the fact that nearly all tend to point to the northwest is consistent with a single, hurricane-induced blowdown. Review of Connecticut Ariel Photographs indicates that this particular blowdown was caused by The New England Hurricane of 1938. The fact that many of the white pine trunks have not entirely decayed, even after 70 years, can be attributed to the preservative effects of pine sap (Hays 1910:124). In plan, the pit-and-mound pair of Tree Throw 2 (investigated in 2008) constitutes overlapping ovals (Figure 10-3). In profile, the outline morphology of the tree throw shows a shelf-like section in the windward end and a deepened section in the leeward end (Figure 10-4). Stratigraphy has been rotated so that the interface between the plowed Ahorizon and B1 horizon is nearly vertical. There is very little pedogenesis on the mound, probably due to constant erosion. The pit exhibits developing A-horizons that have probably been partially formed by windblown and washed-in silt and organics. There is a small uplift in the C-horizon immediately beneath the tree throw (upward “bulge” of Stratum 1) that may have been caused by the uprooting of a taproot. The mound of Tree Throw 3 (investigated in 2009) exhibits ovate outline morphology while the pit is somewhat irregular (Figure 10-5). In profile, this tree throw is nearly identical to Tree Throw 2 (Figure 10-6) in terms of stratigraphy and morphology. It is, however, slightly larger in magnitude, consistent with its larger trunk and root mass remnants. 182 Figure 10-3. Profile section drawing of Tree Throw 2, Sandy Hill Site. Figure 10-4. Tree Throw 2, Sandy Hill Site (10-cm increment scale rod). 183 Figure 10-5. Profile section drawing of Tree Throw 3, Sandy Hill Site. Figure 10-6. Tree Throw 3, Sandy Hill Site (10-cm increment scale rod). 184 10.3. Summary of Observations These investigations resulted in three key observations regarding the subterranean elements of tree throws in local glacial outwash settings. These contribute to the model of DSF formation presented in the next chapter. 1. Trees of the same species thrown under similar conditions may leave consistent subsurface signatures, as suggested by the similar morphologies and stratigraphies in Tree Throws 1 and 2 (both white pine). As gleaned from the background research, root systems of different trees vary from one species to the next but remain consistent within a species; thus, it seems logical to assume that the subterranean remnants of tree throws may exhibit species-specific character when they occur under similar conditions. This minor investigation lends weight to this principle, and may help explain modal forms of DSFs at Preston Plains. 2. Subterranean elements of tree throws occurring in glacial outwash may resemble the D-shaped variety of DSFs. Tree Throws 2 and 3 both had “shelves” on their windward sides and “pits” on their leeward sides, though the differences in depth between the two were not as pronounced as in many Preston Plains DSFs. It is reasonable to conclude that nature can create such seemingly complex shapes. While published profiles of natural (known) tree throws capture a wide range of morphologies, D-shaped morphology has not been clearly documented or described as a natural phenomenon. 185 3. Tree throws in glacial outwash are likely to rotate soil masses, leaving inclined-to-inverted stratigraphy. This process may result in localized concentrations of topsoil in one end of the tree throw, and higher concentrations of subsoil in the other. Background research revealed this tendency, but excavating these examples dramatically illustrates it. This pattern should be discernable long after treethrown topography has become obscured, and should be interpreted as a diagnostic feature of ancient tree throws even in the absence of topographic anomalies. 186 Chapter 11. Genesis and Cultural Significance of DSFs at Preston Plains This chapter articulates key information from throughout the previous chapters to achieve three central goals. The first is to decisively support, above competing explanations, the hypothesis that DSFs at Preston Plains began as tree throws. The second is to present a simple and flexible model that explains how the subterranean remnants of tree throws become modified over time into the variety of expressions identified as DSFs. The third is to determine the cultural significance of DSFs at Preston Plains by reviewing archaeological patterns at Locus 1 according to the hypothesis that DSFs were initiated by tree throws. 11.1. DSFs at Preston Plains Began as Tree Throws This study‟s evidence strongly supports the interpretation that DSFs are tree throw-induced features. As previously established (Chapter 9), artifact concentrations were not present on the bottoms of 18 DSFs investigated at Preston Plains, effectively ruling out the possibility that their bottoms served as the floors of residential structures as hypothesized by archaeologists investigating analogous features (Barnes 1972: 81-2; Custer 1994). Multiple lines of evidence indicate that DSFs were never open pits, but were initiated by tree throws as hypothesized by others (Mueller and Cavallo 1995; Thomas and Payne 1981, Thomas 1982). One common aspect of DSFs at Preston Plains indicates that their basin-like morphologies never constituted open pits. It is unlikely that the steep and sharply187 defined edges prevalent in the vast majority of these features would have remained intact for long if they were ever open surfaces. For example, the edges of DSF-4 at Locus 1 exhibit strikingly steep and well-defined edges that sharply truncate bedded glaciofluvial deposits (Figure 11-1). When units were excavated through adjacent glaciofluvial deposits, their walls rapidly eroded, causing disorganized sand and gravel to accumulate at their bases. No such colluvium was recorded at the bottoms of any DSFs at Preston Plains, which strongly suggests that they were never open pits. Figure 11-1. Section of DSF-4, Locus 1, Preston Plains Site. Note the very steep and sharply defined boundary between feature matrix and bedded glaciofluvial deposits. DSF genesis is best explained as a process of in-situ soil rearrangement, rather than one of excavation and infilling. In accordance with Occam‟s Razor, tree throws are identified as the prime formative agent responsible for creating these basin-shaped 188 features. Literature review (Chapter 4) has established that tree throws may cause deep soil disturbances in unconsolidated soils. Furthermore, deep tree throws tend to cause rotational disturbances that re-arrange soil materials, often leaving inclined stratigraphic patterns beneath the ground surface. Contemporary tree throw excavations (Chapter 10) confirm that tree throws occurring in soils similar to those at Preston Plains generate substantial subsurface disturbances characterized by soil rotation within basin-like morphologies. Some DSFs at Preston Plains are morphologically and stratigraphically consistent with naturally occurring tree throws. The best examples are relatively isolated DSFs such as Features 304, 300, and 302. Longitudinal sections of these DSFs reveal a pattern that can be attributed to the downward thrusting of topsoil during throwing (Figure 11-2). Upturned subsoils are evident in their midsections, and their shallow ends are loamy, perhaps from initial spilling of topsoil into their tree throw pits followed by gradual colluviation. Comparison of these DSFs with a relatively modern tree throw excavated at the nearby Hidden Creek Site (72-163) in Mashantucket further illustrates an undeniable similarity between certain DSFs and modern tree throws (Figure 11-3). 189 Figure 11-2. Longitudinal sections of Fe.304 (top left), Fe.300 (top right), and Fe.302 (bottom), Preston Plains Site. These relatively isolated DSFs retain patterns of inclined stratigraphy consistent with naturally occurring tree throws (10-cm increment scale rod). Figure 11-3. Longitudinal section of modern tree throw, Hidden Creek Site. Note its striking similarity to Features 304 and 302 (DSFs) at the Preston Plains Site, as depicted in Figure 11-2. 190 There is no evidence to suggest that the DSF-inducing tree throws of Preston Plains were anything other than natural occurrences. Egghart‟s anthropogenic tree throw model (2005) is not supported by the Preston Plains data. His model predicts the presence of charred wood concentrations on the bottoms of the deep ends of D-shaped pit features, which were hypothesized to be “grubbing trenches” where tree roots were exposed and burned. Particularly high concentrations of charred wood were not present in these locations within Preston Plains DSFs. In fact, the most carbon-rich sections of DSF matrix were discovered in the shallow ends of DSFs 3, 4, and 7 at Locus I. Several other observations at Preston Plains support the interpretation that DSFs are the result of naturally occurring tree throws. For instance, a by-product of catastrophic soil rotation was preserved along a margin of Fe.306, where this DSF‟s steep edge is lined with an upthrust of fine, sandy substrate (Figure 11-4). Also, several DSFs (310, 311, 312, 317, 318) in machine Trench 17 penetrated soft, sandy outwash and “bottomed out” against coarser underlying gravel (Figure 11-5). This indicates that the maximum depth of some DSFs is determined by the differing potentials of glaciofluvial layers to resist mechanical stress exerted by a rotating root mass. Furthermore, DSFs at Preston Plains were not confined to areas of cultural deposits, a trend that has been repeatedly recorded at sites containing similar features in the Middle Atlantic Region (Egghart 2005). 191 Figure 11-4. View west of a section of Fe.306 in MT-13, Preston Plains Site. Note the upthrust of fine sandy substrate that reflects a process of catastrophic soil rotation (10-cm increment scale rod). Figure 11-5. Section of Fe.318 in MT-17, Preston Plains Site. Note how this DSF penetrates finegrained sands and bottoms out against a gravel substrate. The morphological variability of DSFs at Preston Plains is partially explained by the fact that different tree species can produce differently shaped soil disturbances when thrown. This was indicated by background research (Chapter 4), and primary data (Chapter 10). Bisection of a tree throw of an unidentified species at the Avery Pond Site 192 revealed a simple, bowl-shaped morphology, similar to the one recorded at the Hidden Creek Site. In contrast, bisections of two white pine tree throws at the Sandy Hill Site revealed identical “D-shaped pit” morphologies, though the deep ends were not as dramatically deep as those recorded in DSFs 1 and 2 at Locus 1 of Preston Plains. While tree throw morphologies tend to be described by scholars as either rotational (deep) or hinge (shallow/plate-like), the tree throws at Sandy Hill could be characterized as both at once. Thus, it seems quite plausible that nature can produce such forms as the “D-shaped pit” (Figure 11-6), though it is not known which tree taxa are implicated. But there are many other factors outlined in Chapter 4 that could contribute to variability in tree throw signatures, including a tree‟s age, size, and health, in addition to external factors including soil, topographic, and meteorological conditions. In other words, while modes in form, stratigraphy, and magnitude should be expected within a group of tree throws, variability should also be expected. However, variability in DSFs is not simply a function of the conditions that cause tree throws to occur. According to the author‟s vision, tree throw remnants are modified through time to become the phenomena that he identifies as DSFs. 193 Figure 11-6: Hypothesized formation of “D-Shaped pit” morphology by a tree throw that involves deep soil rotation and shallow plate removal. 194 11.2. DSF Variability Reflects Individual Formation Histories It would be a mistake to view the DSFs of Preston Plains as mere echoes of catastrophic soil displacement. While each was initiated in a momentary episode, each subsequently experienced a unique formative history, having been modified through time according to multiple natural factors. To acknowledge these complex histories, the author thinks it is fitting and appropriate to continue calling their manifestations DSFs, rather than tree throws. Though the author believes it impossible to identify all factors influencing the modification of a DSF, much less accurately gauge their magnitudes of effect, some key factors are hereby discussed. Soil water capacity is a key factor in the modification of tree throws remnants. Test data presented in Chapter 6 (see Table 6-1) indicate that glaciofluvial sand and gravel at Preston Plains retains very little moisture, which explains why it was consistently free of root-related turbation across the site. However, samples of DSF fill generally retained more water, which explains the fact that they were selectively penetrated by root systems when trees or shrubs were in close proximity (such as at Locus I and MT-17). On the excessively drained Preston Plains landform, tree throws create dynamic subterranean microenvironments that are highly attractive to roots. First, they loosen soil, making it easier for roots to intrude and expand. Second, they transport fine-grained (aeolian) fractions and organics from the ground surface downward, which increases water retention and prolongs the mobility of ions within DSF matrix. The degree to which root systems penetrate and modify DSF matrix must partially reflect the individual 195 life histories of plants that grow in proximity, not to mention the myriad effects of faunal activity that they attract. So while the absence of trees across most of Preston Plains has facilitated the archaeological investigation of DSFs, it has also prevented us from witnessing a fuller snapshot of their dynamism. Clusters of DSFs in some areas of Preston Plains probably reflect complex, synergistic formative trends. The author believes these trends are driven by a principle: the degree to which a DSF is modified over time is positively influenced by its proximity to subsequently occurring tree throws or to other water-retaining soil features. As noted above, relatively isolated DSFs exhibit stratigraphy that resembles that of recently occurring tree throws. DSFs that occur in clusters are more likely to lack clear stratification and/or form overlapping-to-amorphous composite features. One factor contributing to this synergy may be a tendency for serial tree throwing to occur. For example, the topography across much of the Sandy Hill Site is dominated by the pit-and-mound topography of a single blowdown attributed to The New England Hurricane of 1938. There is a strong tendency for new trees to grow on the loam-rich mounds here (Figure 11-6). If any of these secondary-growth trees were thrown, they would turbate strata within developing DSFs and possibly contribute to the expansion of their forms. This tendency may have contributed to the pattern of feature soils in MT-17, where a series of DSFs were identified in addition to an extensive soil mass that may have been formed by overlapping DSFs (Figure 11-7). 196 Figure 11-7. Tree throw at Sandy Hill, with a new tree growing on the loam-rich mound (10-cm increment scale rod). Figure 11-8. North wall of MT-17, Preston Plains Site. The clustered occurrence of DSFs (such as Fe.319, 320, 321) may eventually result in the development of extensive, unstratified soil masses (such as Fe.323). 197 Also, as noted in Chapter 6, DSFs were heavily concentrated along the remnant glaciofluvial channel. This channel represents a massive water-retaining feature that doubtlessly attracted the growth of large trees earlier and more frequently than surrounding locales. It should come as no surprise that deep tree throws occurred here during prehistory, and that they are manifested by numerous, and often overlapping, DSFs. A similar synergistic model, referred to as the model of self-reinforcing pedologic influences of trees (SRPIT), was developed through the study of the pedologic influence of trees in the Ouachita Mountains, Arkansas (Phillips and Marion 2003). It explains the heterogeny of local forest soils as the product non-random impacts of individual trees on the landscape. The preferential and persistent attraction of trees to certain microsites results in those sites becoming nutrient rich through a variety of biomechanical processes, including tree throws. The author envisions a comparable process governing pedogenesis across greater Preston Plains. An effect of such synergies at Preston Plains is an increased rate of destratification within DSF forms that are clustered (Figures 11-8 and 11-9). Hypothetically speaking, tight clusters of DSFs should support a greater vegetal biomass than surrounding areas. This, in turn, means a greater volume of root systems are available to turbate the matrices of clustered DSFs through processes such as root growth and decay, tree sway, and subsequent tree throws. Over time, these processes cause inclined stratigraphic signatures (originally consisting of rotated and/or slumped topsoil and subsoil horizons) within DSF forms to become indistinct as their matrix 198 homogenizes. Destratification certainly occurs in isolated DSFs as well, but is expected to proceed at slower rates according to a lack of synergy. Figure 11-9. Detail of Fe.319, MT-17, Preston Plains Site. Note the virtual absence of stratification. Figure 11-10. West view of a section of DSF-2, Locus 1, Preston Plains Site. Note the lack of robust stratification within DSF matrix. 199 In sum, soil dynamics that are sensitive to water capacity and proximity-driven synergistic tendencies are thought to influence the rate of DSF modification. These dynamics are assumed to be reflected in innumerable processes of bioturbation and weathering, the individual effects of which may be impossible to isolate or measure on a feature-by-feature basis. This discussion has doubtlessly oversimplified the factors contributing to DSF formation and modification at Preston Plains, and each of these features should be recognized as having an individual formation history. Nonetheless, this model provides a flexible, plausible, and well-supported vision of how the underground remnants of ancient tree throws can transform into the variety of expressions recognized as DSFs. 11.3. Cultural Significance of DSFs at Locus I: Hummocked Topography as an Element of Site Selection During the Late Archaic Period By applying the above interpretations to archaeological data from Locus I, this study strongly supports Mueller and Cavallo‟s hypothesis that DSFs were generated by tree throws that, in some instances, were utilized or modified by prehistoric populations (1995). During the Late Archaic Period, tree throw-induced pit-and-mound features were highly visible, enduring elements of the Locus 1 site setting, a setting that small groups of forgers repeatedly selected for the establishment of short term camps in the summer and fall. Site occupants discarded artifacts directly onto these features and modified their surfaces by excavating formal pit hearths in their soft matrix. It also appears that in three 200 instances (DSF-3, 4, and 7) the tree throw pits themselves (spatially corresponding to the shallow ends of DSFs) were used, and/or slightly modified for use, as fireplaces. In light of data gathered from the excavation of modern tree throws (Chapter 10), we may infer that each DSF at Locus I marks the previous location of a topographically robust pit-and-mound feature. Despite erosional and depositional processes, each probably endured as a highly visible element of the Locus 1 area for quite some time. To illustrate, Tree Throw #3 (discussed in Chapter 10) at the Sandy Hill Site was induced by The New England Hurricane of 1938, and the vertical range of its pit-and-mound topography exceeded 1.5 meters when it was recorded in 2009, 71 years later! This should come as little surprise, as this study‟s background research established that visually recognizable mounds attributed to tree throws in temperate forest settings may be centuries old. Evidence indicates that these tree throws occurred in series for over a millennium, and were not the result of a single blowdown. The tendency for some DSFs to truncate others reflects this serial formation. Furthermore, the radiocarbon date series spans much of the Late Archaic Period. Therefore, it may be reasonably postulated that hummocked topography was a persistent aspect of the Locus 1 locale throughout the 5th millennium, though it was doubtlessly in a constant state of change. It can be confidently inferred that Late Archaic foraging populations who periodically visited this location recognized the pit-and-mound topography that was concentrated here. Table 11-1 estimates when treethrown topography existed based on stratigraphy and radiocarbon data. A conservative 201 estimate of 200 years is provided for the longevity of a robust pit-and-mound, though their eroding morphologies were likely discernable for much longer. Table 11-1: Estimated tree throw dates and longevities of pit-and-mound pairs at Locus 1. DSF 1 2 3 4 5 6 7 8 9 10 Date Associated Tree Throw Occurred After ca. 5290 BP (based on Fe.12 date), Before ca. 5030 BP (based on truncation by DSF-3). By averaging these dates, genesis is estimated at ca. 5160 Before DSF-1 At ca. 5030 BP, when its pit was used as a fireplace (Fe.22) At ca. 4120 BP, when its pit was used as a fireplace (Fe.18) At ca. 4710, when temporarily open pit experienced burning No date, but it is capped by a Narrow-Stemmed Tradition component. At ca. 4090 BP, when its pit was used as a fireplace (Fe.26) Unknown, but pre-4120 BP as it is truncated by DSF-4 Unknown, but pre-4090 BP as it is truncated by DSF-7 Unknown Duration of Pit-and-Mound ca, 5160-4960 ca. late 6th millennium? ca. 5030-4830 BP ca. 4120-3920 BP ca. 4710-4510 BP ca. 5th or 4th millennium ca. 4090-3890 BP ca. 5th millennium? ca. 5th millennium? unknown Artifact distributions at Locus 1 are particularly informative when interpreted according to Bubel‟s (2003) hypothetical scenarios for the disposition of archaeological materials in tree thrown contexts (summarized in Chapter 4). One notable trend was the tendency to recover older projectile points from the matrix of DSFs 1-3, while younger points were recovered mostly from overlying topsoil contexts. DSFs 1-3 contained the only Early-to-Middle Holocene point types recovered from this locus. They also contained Late Archaic Period points, but only the earlier Laurentian varieties. All diagnostic point types recovered from topsoil above these DSFs are attributed to the Late Archaic Period, including types attributed to the Laurentian and subsequent Narrow Stemmed Tradition. Overall, this trend indicates that DSFs 1-3 were induced by tree throws that occurred during the late 6th and 5th millenniums, burying points associated with preexisting components. The vertical distribution of artifacts in the deep end of 202 DSF-2 probably resulted from tree throw-induced soil rotation that is consistent with Bubel‟s Case 2 scenario. The tree throws that initiated DSFs 1-3 also buried pre-existing hearth deposits. Fe.10 and 12 are hearth-related anthrosols that appear to have been plunged into the deep ends of DSFs 2 and 1, respectively. In all probability, these were once hearths associated with early Laurentian Tradition occupations. Some of the suspected hearth-related soil stains within DSFs 1 and 2 may represent similarly displaced deposits as well. This suggests that foragers were already attracted to the Locus 1 site setting by the opening of the Late Archaic, probably because of its slightly elevated topography, though tree throws may have enhanced its appeal. After the tree throws associated with DSFs 1-3 occurred, Late Archaic foragers associated with both the Laurentian and Narrow Stemmed Traditions continued to occupy the site, depositing additional materials on hummocked surface contexts and creating additional hearth features. This interpretation is bolstered by the general distribution of all lithic artifacts, including debitage, in DSFs 1 and 2 (see Figure 9-5). Most lithics are located in the uppermost portions of DSF matrix, which probably reflects discard patterns on the ground surface. This is consistent with Bubel‟s Case 3 scenario. It is critical to note here that DSF matrix at Locus 1 was much softer than intact glaciofluvial sand and gravel deposits and far easier to excavate. It should come as little surprise that loosened, tree-thrown soil provided an attractive medium for the excavation of formal pit heaths such as Fe.3, 6, 8, and 14. 203 So, the evidence thus far strongly supports the interpretation that foragers periodically inhabited the evolving, hummocked topography that existed at Locus 1 during the Late Archaic Period. But were they using the pit-and-mound morphologies themselves for particular purposes? According to background research in Chapter 4, archaeologists have already suggested that prehistoric foragers expediently used tree throws as site furniture. The strongest evidence of this behavior at Locus 1 is reflected in Fe.18, 22, and 26. These hearths are all situated along the very edge of the shallow ends of their respective DSFs. If we accept that these DSFs began as tree throws, we should recognize their pits as offering excellent natural fireplaces that require little or no modification by site occupants. All three of these hearth features conformed to the edge morphologies of their respective DSFs. The burning that occurred within them appears to be in-situ burning, indicated by the presence of heat-altered gravel; therefore, it seems unlikely that their cultural contents are merely the result of windblown or washed-in fractions from other portions of the site. People appear to have used the shallow ends of DSFs 3, 4, and 7 as fireplaces, and the author suspects that this occurred within a century or two of the tree throws that initiated them. No post molds were identified during this study, so the presence of living structures here cannot be confirmed. There are two interesting phenomena at Locus 1 that are still not fully understood. The apparent burning of a surface (Fe. 20) in the lower portions of DSF-5 suggests that its associated tree throw‟s uplift created a substantially deep basin feature for a short period of time. If so, this is an atypical expression of tree throwing at Preston Plains. 204 Though cultural influences cannot be ruled out, the lack of an associated artifact concentration suggests this was a natural burn that was subsequently capped by the collapse or rapid erosion of uplift matrix. The artifact concentration identified as Fe.17 is also not fully understood. While it is thought to represent a remnant occupational surface or feature remnant intruding DSF-6 matrix, this interpretation is inconclusive. 205 Chapter 12. Significance of this Study to the Cultural Resource Management Industry and the Academy This final chapter outlines the significance of this study to researchers in the cultural resource management (CRM) industry as well as the academy. Practical advice is offered to those investigating DSF complexes in New England and the Middle Atlantic under the umbrella of public archaeology. Namely, flexible testing strategies may be necessary to detect the presence such features, and key attributes should be called upon to positively identify them as DSFs. It is also appropriate to distinguish and bypass culturally insignificant DSFs during expensive archaeological data recovery excavations. On a scale of regional relevance, this study indicates that archaeologically relevant DSFs hold the potential to contribute valuable information regarding the reconstruction of prehistoric settlement systems, and/or may harbor valuable archaeological data sets that are not preserved in plowed contexts. On a scale of global relevance, scholarship geared toward understanding prehistoric human behavioral ecology in forested environments should be vigilant regarding the identification and evaluation of ancient tree-thrown topographies as potential site elements. Results of this research are particularly relevant to CRM professionals. As development projects proceed in eastern North America additional archaeological complexes containing DSFs will likely be encountered, and CRM professionals cannot provide convincing management recommendations to address them until effective avenues of inquiry have been developed to assess their cultural significance. Over the 206 past two decades, the Delaware Department of Transportation has substantially invested in the recovery of archaeological data from several sites containing DSF complexes as planning elements of development projects. Doucette and Flynn (2008) recently completed data recovery investigations at the J.T. Berry Site in North Reading, Massachusetts, which involved the excavation of several DSFs. This was carried out as a planning element in the development of a residential and professional complex. Excavations at Preston Plains were carried out primarily to provide the MPTN Planning Department with management recommendations regarding the Preston Plains Energy Center Project and any future development projects in the vicinity, though they were also tailored to satisfy the academic goals of this dissertation. DSF complexes on culturally significant (ie. eligible for listing in the National Register of Historic Places) archaeological sites that require mitigation pose a bipartite problem. First, complete excavation of DSFs may not be possible according to the financial and temporal constraints of most development projects. Also, while data recovery efforts guided by sampling strategies are often appropriate to implement regarding cultural resources that are well understood, devising an informed sampling strategy to excavate DSFs is problematic because archaeologists do not agree as to what these features represent or how culturally significant they are. In regard to interpreting the genesis and cultural significance of DSFs at Preston Plains, this study does not offer the CRM industry a simple or definitive set of conclusions that could (or should) be universally applied to the interpretation of similar features from all sites in the Northeast or Middle Atlantic. That would conflict with the 207 author‟s perspective of Preston Plains as a unique and historically particular landscape, fail to acknowledge that equifinial processes may generate similar features, and discount the creativity of other researchers who may devise better ways of addressing these issues. Instead, by drawing on empirically and scientifically based evidence, this study lends weight and clarity to existing modes of interpretation in the scholarly discourse on DSFs that recognize that naturally occurring tree throws are more likely to generate these features than any other known mechanism. In doing so, it is hoped to bring the archaeological community that much closer to agreeing upon the factors that contribute to the formation and modification of DSFs, and closer to understanding the ways in which humans may have used them in the prehistoric past. Several practical insights based on the author‟s experience at Preston Plains are hereby shared, some of which may be instrumental to other researchers investigating DSF complexes. First, it would be appropriate to incorporate flexible excavation strategies into the research design of data recovery investigations. DSFs were often very difficult to recognize at Preston Plains, especially when they contained few or no artifacts. DSF matrix, when penetrated by test units as large as 1-m square, often appeared to constitute little more than an abnormally deep B-horizon soil. Furthermore, DSF morphology was often more difficult to assess in plan than in profile. Determining the best way to reveal the morphology and stratigraphy of these features may be an organic process that unfolds differently depending on what site, or what part of a site, one is excavating. Though DSFs vary in character, two traits are useful to distinguish them from anthropogenic pit features, such as pit houses, roasting pits, or storage pits. First, though 208 some DSFs appear non-stratified, many exhibit patterns of inclined or vertical stratigraphy that is characteristic of tree throws. These patterns may be robust, or they may be reflected only by subtle changes in gravel content between zones of matrix. Either way, such patterns should be recognized as potentially critical aspects of feature identification. Furthermore, if such patterns are found among a few DSFs in a given complex, this finding may be used to support the interpretation that other DSFs (even if they are entirely non-stratified) likely exhibited these patterns earlier in their formation histories. Second, artifact concentrations should not be present on the very bottoms of DSFs. The tree throws that initiate DSFs appear to cause deep soil rotations that provide no opportunity for an artifact concentration to accumulate on their bottoms. While artifacts may occur throughout the matrix of a DSF, the presence of a distinct artifact layer on the bottom of a basin-shaped feature should alert the investigator to the possibility that it may not be a DSF, but an actual basin that has filled. Data recovery research designs for DSF complexes should distinguish between potentially archaeologically significant DSFs and non-significant DSFs. DSF complexes may extend beyond the boundaries of an archaeological site, which supports the logical assumption that not all tree thrown topography was targeted for use by prehistoric peoples. Therefore, it would be inappropriate for any archaeological mitigation to completely excavate every DSF in a given complex if they did not all occur in close spatial association with archaeological materials. All DSFs that occur in close spatial association with archaeological materials should be considered potentially culturally significant pending the results of a data recovery investigation. That investigation should 209 sufficiently assess the disposition of artifacts and anthrosols within and surrounding DSFs to determine their primary cultural significance, which is hereby proposed to be centered on the use pit-and-mound topography within prehistoric forested landscapes. Furthermore, there is a secondary context in which DSFs are culturally significant. Tree throws may transport pre-existing archaeological deposits to depths that are beyond the destructive reach of the plow, thus serving as natural preservation mechanisms. As mentioned in Chapter 4, the author realized this while excavating the Stones Throw Site in New Hampshire (Ives 2006b). This site is located on a sandy outwash terrace where most lithic artifacts were recovered from plowed topsoil. However, at the epicenter of a small Late Paleoindian lithic scatter a probable hearth remnant was discovered beneath the plowzone in the matrix of an ancient tree throw (which the author would now classify as a DSF). This feature, and its valuable radiocarbon data, would probably not have been evident if a tree throw had not fortuitously transported feature matrix to greater depth. In areas that have been deeply plowed, DSFs may offer the only soil contexts where prehistoric feature remnants would be encountered. This study‟s findings make a modest contribution to the study of southern New England‟s Late Archaic Period settlement systems, revealing the repeated use of a forested, inland outwash plain by small, mobile groups in the summer-to-fall seasons. While Late Archaic sites not difficult to find, figuring out how to articulate them into broader, and potentially flexible, patterns of population movements and reorganizations remains challenging. The high density of sites and almost exclusive reliance on locally 210 available lithic materials suggests higher populations during the Late Archaic than in preceding or subsequent periods (Dincauze 1975). These populations clearly made more frequent use of a wider variety of environmental zones (Hoffman 1985; Sgarlata 2009), which probably reflects a regional packing effect (sensu Binford 1983). The Narrow Stemmed Tradition, the most prominent cultural manifestation of the Late Archaic in southern New England, is thought to reflect a central-based wandering pattern (sensu Beardsley et al. 1956) (Snow 1980: 230) where numerous, but relatively small, communities exploited a wide variety of settings according to a diffuse adaptation and complex pattern of seasonal movements (Dincauze 1975). There is little agreement as to when or where seasonal base camps are typically found, which suggests regional variability during this time (McBride 1984). Focal points in the settlement system, where base camps are more likely to occur, may include lakeside winter (Dincauze 1974), riverine summer/fall (McBride 1978), and coastal settings (Cozzone and Hartenberger 2009; Snow 1980). Smaller, temporary camps are anticipated in a variety of microenvironments away from these base camps (McBride 1984). Preston Plains clearly represents such an environment – specifically, a forested, inland outwash plain in southeastern Connecticut. Its palimpsest of short-term residential occupations reflects the exploitation of plant and animal resources by small groups of foragers, perhaps nuclear families, during the summer-to-fall seasons. Interestingly, this occupational pattern was already in place by the time Narrow Stemmed occupants arrived at Preston Plains. Foragers associated with the preceding, but partially coeval, Laurentian Tradition, also used a central-based wandering system 211 (Snow 1980) and their material attributes clearly ranged beyond their cultural heartland into southern New England. Laurentian Tradition manifestations are poorly understood in Connecticut compared to those of the Narrow Stemmed Tradition (Pfeiffer 1984), but recognition of Laurentian components on Connecticut sites has increased substantially over the past two decades. The Preston Plains projectile point assemblage and radiocarbon date series, especially at Locus 1, substantiate the presence of Laurentian occupants in eastern Connecticut perhaps as early as the late 6th millenium. The continuity of the Late Archaic foraging signature at Locus 1 suggests that Laurentian and Narrow Stemmed occupants used this site in approximately the same fashion. So while Connecticut could be seen as part of a “tension belt” (sensu Snow 1980) between coeval traditions or cultural adaptations to specific environmental zones, Preston Plains suggests a local continuity between the two traditions. As noted in chapter 7, this continuity is also suggested by the morphological gradation between Brewerton (Laurentian Tradition) and Squibnocket (Narrow Stemmed Tradition) points from Preston Plains. On a scale of global relevance, this study alerts scholars of prehistoric human behavioral ecology to some of the more subtle dynamics of life in forested landscapes that are archaeologically visible. Specifically, archaeological studies of ancient forested landscapes should be vigilant regarding the identification and evaluation of tree-thrown topographies as potential site elements. It has long been speculated that tree throws create topographic features that were opportunistically exploited by prehistoric people, but we are now beyond speculation. Evidence strongly indicates that final Paleolithic-toMesolithic foragers of northern Belgium made use of tree throw pits according to the 212 deposition of flints therein (Crombé 1993). Additionally, the shifting cultivators of early Neolithic England are found to have generated domestic middens within tree throw pits (Evans et al. 1999). This study has shown that Late Archaic Period foragers of southeastern Connecticut establishing short-term residential camps used tree throw pits as fireplaces in some instances, while in other instances they modified the tree throw‟s soft matrix to create formal pit hearths. Confirming the prehistoric exploitation of tree thrown topography on both sides of the Atlantic is an exciting development and the author hopes that it captures the imaginations of scholars who will, in turn, promote related research inquiries. Perhaps the most critical insight revealed by this research is the simple recognition of DSFs as marking previous locations of robust pit-and-mound topographies within ancient cultural landscapes. As framed by Evans, Pollard and Knight (1999), fallen trees created “substantial landscape features” that “should not be underestimated.” In addition to being highly visible to prehistoric peoples inhabiting forests, they were likely perceived as meaningful on both practical and symbolic levels. If any archaeologist is fortunate enough to discover a context where cultural deposits and DSFs are interwoven, their relationship should be dissected to determine whether or not their concurrence reflects the exploitation of hummocked topography by prehistoric peoples. Bubel (2003) has outlined a clear framework for dissecting this interrelationship, and this dissertation has demonstrated its successful application as a key element of a holistic interpretive approach. 213 Sources Cited Alley, R.B., P.A. Mayewski, T. Sowers, M. Stuiver, K.C. Taylor, and P.U. Clark 1997 Holocene Climactic Instability: A Prominent, Widespread Event 8200 Yr Ago. Geology 25(6):483-486. Araujo, A.G., and J.C. Marcelino 2003 The Role of Armadillos in the Movement of Archaeologial Materials: An Experimental Approach. Geoarchaeology 18(4):433-460. Barnes, R.C. 1972 Bear Swamp II: A Preliminary Report. Unpublished Ph.D. Dissertation, Anthropology Department, University of Pennsylvania. Beardsley, R.K, P. Holder, A.D. Krieger, B.J. Meggers, J.B. Rinaldo, and P. Hutsche 1956 Functional and Evolutionary Implications of Community Patterning. Seminars in Archaeology: 1955. Society for American Archaeology, Memoir 11:129-155. Beatty, S.W., and E.L. Stone 1986 The Variety of Microsites Created by Tree Falls. Canadian Journal of Forestry Research 16:539-548. Bennie, A.T.P. 1996 Growth and Mechanical Impedence. In Plant Roots: The Hidden Half, Second Edition, edited by Y. Waisel, A. Eshel and U. Kafkafi, pp. 453-471. Marcel Decker Inc., New York. Bernstein, D.J. 1990 Prehistoric Seasonality Studies in Coastal Southern New England. American Anthropologist 92(1):96-115. Binford, L. R. 1979 Organization and Formation Processes: Looking at Curated Technologies. Journal of Anthropological Research 35: 255-273. 1983 In Pursuit of the Past: Decoding the Archaeological Record. Thames and Hudson, New York. Boeck, B. 1986 Rodent Ecology and Burrowing Behavior: Predicted Effects on Archaeological Site Formation. American Antiquity 51:589-603. Boisvert, R.A. 1999 PaleoIndian Occupation of the White Mountains, New Hampshire. Geographie physique et Quaternaire 53(1):159-174. Borns, H.W., Jr., L.A. Doner, C.C. Dorion, G.L. Jacobson Jr., M.R. Kaplan, K.J. Kreutz, T.V. Lowell, W.B. Thompson, and T.K. Weddle 2004 The Deglaciation of Maine, U.S.A. In Quaternary Glaciations – Extent and Chronology, Part II: North America, edited by J. Ehlers and P.L. Gibbard, pp. 89-109. Elsevier, Amsterdam. Bourque, B.J. 2001 Twelve Thousand Years: American Indians in Maine. University of Nebraska Press: Lincoln. Braun, D.P. 214 Explanatory Models for the Evolution of Coastal Adaptation in Prehistoric New England. American Antiquity 39:582-596. Brennan, L.A. 1974 The Lower Hudson: A Decade of Shell Middens, Archaeology of Eastern North America 2(1):81-93. 1977 The Lower Hudson: The Archaic. In Amerinds and Their Paleoenvironments in Northeastern North America, edited by W.S. Newmn and B. Salwen. Annals of the New York Academy of Sciences, No. 288. Brewer, R., and P.G. Merritt 1978 Wind Throw and Tree Replacement in a Climax Beech-Maple Forest. Oikos 30:149-145. Bryson, R.A., D.A. Baerreis, and W.M. Wendland 1970 The Character of Late-Glacial and Post-Glacial Climatic Changes. In Pleistocene and Recent Environments of the Central Great Plains edited by W. Dort, Jr. and J. K. Knox, Jr. pp. 53-74. University of Kansas Press. Bubel, S. 2002 Bioturbation and its Effects on the Archaeological Record. Unpublished dissertation, Departement Archaologie, Kunstwetenschappen en Musicologie Adeling Archeologie. Katholieke Universiteit Leuven. Bunker, V. 1992 Stratified Components of the Gulf of Maine Archaic Tradition at the Eddy Site, Amoskeag Falls. In Early Holocene Occupation in Northern New England, edited by B. Robinson, J. Petersen and A. Robinson, pp. 135-148. Occasional Papers in Maine Archaeology 9. Maine Historic Preservation Commission, Augusta. Butzer, K.W. 1982 Archaeology as Human Ecology: Method and Theory for a Contextual Approach. Cambridge University Press, Cambridge. Ceci, L. 1980 Maize Cultivation in Coastal New York: The Archaeological, Agronomical, and Documentary Evidence. North American Archaeologist 1(1):45-74. Connecticut Development Commission 1963 The Physical Geography of Connecticut. Connecticut Interregional Planning Program, Technical Report 111. Coutre-Palmerino, P. 1999 Personal Communication. Crombé, P. 1993 Tree-Fall Features on Final Palaeolithic and Mesolithic Sites Situated on Sandy Soils: How to Deal With It. Helinium 33:50-66. Crouch, M.H. 1983 Soil Survey of New London County, Connecticut. United States Department of Agriculture, Soil Conservation Service. Custer, J.F. 215 1974 Prehistoric Cultures of the Delmarva Peninsula: An Archaeological Study. University of Delaware Press, Newark. 1994 Stability, Storage and Culture Change in Prehistoric Delaware: The Woodland I Period (300 B.C. – A.D. 1000). On file at the Delaware State Historic Preservation Office, Division of Historical and Cultural Affairs, Dover. Custer, J.F., L. Riley, and G. Mellin 1994 Final Archaeological Investigations at the Leipsic Site (7K-C-194A) State Route 1 Corridor, Kent County, Delaware. University of Delware Department of Anthropology Center for Archaeological Research, Newark. Custer, J.F., and B.H. Silber 1994 Final Archaeological Investigations at the Pollack Site (7K-C-194A) State Route 1 Corridor, Kent County, Delaware. University of Delaware Department of Anthropology Center for Archaeological Research, Newark. Custer, J.F., S.C. Watson, and B.H. Silber 1995 Final Archaeological Investigations at the Snapp Site (7NC-G-101), State Route 1 Corridor, Chesapeake and Delaware Canal Section, New Castle County, Delaware. Delaware Department of Transportation Archaeology Series No. 122. Dover. Cuzzone, H., and B. Hartenberger 2009 Late Archaic and Late Woodland occupations at Cove River, West Haven: New Data on Subsistence and the Narrow Point Tradition. Bulletin of the Archaeological Society of Connecticut 71:17-39. Davis, M.B. 1969 Climactic Changes in Southern Connecticut Recorded by Pollen Deposition at Roger Lake. Ecology 50(3):409-422. 1983 Holocene Vegetation History of the Eastern United States. In Late Quaternary Environments of the United States: The Holocene, edited by H.E. Wright, Jr., pp. 166-181. University of Minnesota, Minneapolis. Davis, M.B., R.W. Spear, and L.C.K. Shane 1980 Holocene Climate of New England. Quaternary Research 14:240-250. Davis, R.B., T.E. Bradstreet, R. Studkenrath, and H.W. Borns, Jr. 1975 Vegetation and Associated Environments during the Past 14000 Years at Moulton Pond. Maine Quaternary Research 5:436-465. Davis, R.B., and G.L. Jacobson, Jr. 1985 Late Glacial and Early Holocene Landscapes in Northern New England and Adjacent Areas of Canada. Quaternary Research 23:341-368. Day, W.R. 1950 The Soil Conditions Which Determine Windthrow in Forests. Forestry 23:9095. Decker, Z. 2006 Summary Report of Field Investigations at Site 114-93, Fall 2006. On file at the Mashantucket Pequot Museum and Research Center. Deevey, Jr., E.S., and R.F. Flint 1957 Postglacial Hypsithermal Interval. Science 125:182-184. 216 1989 Delcourt, P.A., and H.R. Delcourt 1987 Late Quaternary Dynamics of Temperate Forests: Applications of Paleoecology to issues of Global Environmental Change. Quaternary Science Reviews 6(2):129-146. Denny, C.S., and J.C. Goodlett 1956 Microreleif Resulting from Fallen Trees. In Surficial Geology and Geomorphology of Potter County, Pennsylvania, edited by C.S. Denny, US Geological Survey Professional Paper 288, pp. 59-66. Dincauze, D.F. 1968 Cremation Cemeteries in Eastern Massachusetts. Papers of the Peabody Museum of Archaeology and Ethnology 59(1). Cambridge. 1971 An Archaic Sequence for Southern New England. American Antiquity 36(2):194-198. 1974 An Introduction to the Archaeology of the Greater Boston Area. Archaeology of Eastern North America 2(1):39-67. 1975 The Late Archaic Period in Southern New England. Artic Anthropology 12(2):23-24. 1976 The Neville Site: 8000 Years at Amoskeag Falls. Peabody Museum Monographs 4. Harvard University, Cambridge, MA. Dincauze, D., and M.T. Mulholland 1977 Early and Middle Archaic Site Distributions and Habitats in Southern New England. In Amerinds and their Paleoenvironments, edited by Walter S. Newman and Bert Salwen, pp. 439-456. Annals of the New York Academy of Sciences Vol. 288. Doyle, R.A., N.D. Hamilton, J.B. Petersen, and D. Sanger 1985 Paleo-Indian Remains from Maine and Their Correlations in Northeastern Prehistory. Archaeology of Eastern North America 28:81-112. Doucette, D.L., and J.R. Cross 1997 Annasnappet Pond Archaeological District, North Carver, Massachusetts. An Archaeological Data Recovery Program. The Public Archaeology Laboratory, Inc. Report No. 580. Prepared for US Department of Transportation, Federal Highway Administration and Massachusetts Highway Department. Doucette, D.L., and E. Flynn 2008 Technical Report: Archaeological Data Recovery, J.T. Berry Site, Locus 1, Locus 2, Locus 4, Locus 6, and Locus 9, J.T.Berry Rehabilitation Center, North Reading and Wilmington, Massachusetts. PAL Report No. 1851.02. Submitted to Symmes Maini & McKee Associates, Cambridge, MA. On file at the Massachusetts Historical Commission, Dorchester, MA. Dragoo, D.W. 1976 Some Aspects of Eastern North American Prehistory: A Review 1975. American Antiquity 41:3-27. Dumais, P. 2000 The La Martre and Mitis Late Paleoindian Sites: A Reflection on the Peopling of Southeastern Quebec. Archaeology of Eastern North America 28:81-112. 217 Egghart, C. 2005 Prehistoric Modification of the Forest Cover in the Delaware Coastal Plain: An Interpretive Model for Large D-Shaped Pit Formation. Journal of Middle Atlantic Archaeology 21:83-92. Evans, C., J. Pollard, and M. Knight 1999 Life in Woods: Tree-Throws, „Settlement‟ and Forest Cognition. Oxford Journal of Archaeology 18(3). Farb, P. 1961 The Forest. Time Incorporated: New York. Faison, E. K., D. R. Foster, W.W. Oswald, E.D. Doughty, and B.C.S. Hansen 2006 Early-Holocene Openlands in Southern New England. Ecology 87:2537-2547. Fiedel, S.J. 2001 What Happened in the Early Woodland? Archaeology of Eastern North America 29:101-142. Fitting, J. 1968 Environmental Potential and the Postglacial Readaptation in Eastern North America. American Antiquity 33(4):441-445. 1978 Regional Cultural Development, 300 B.C. to A.D. 1000. In Handbook of North American Indians, Vol.15:44-57. Forrest, D.T. 1999 Beyond Presence and Absence: Establishing Diversity in Connecticut‟s Early Holocene Archaeological Record. Bulletin of the Archaeological Society of Connecticut 62:79-99. Foster, D.R., and J.D. Aber 2004 Forests in Time: The Environmental Consequences of 1,000 Years of Change in New England. Yale University Press: New Haven. Foster, D.R., W.W. Oswald, E.K. Faison, E.D. Doughty, and B.C.S. Hansen 2006 A Climactic Driver for Abrupt Mid-Holocene Vegetation Dynamics and the Hemlock Decline in New England. Ecology 87:2959-2966. Fraser, A.I. 1962 The Roots and Soil as Factors in Tree Stablility. Forestry 35:117-27. Fritzsche, K. 1933 Sturmgefahr und Anpassung. Tharanther Forstliches Jahrbuch 84:1–94. Fuller, J.L. 1998 Ecologial Impact of the Mid-Holocene Hemlock Decline in Southern Ontario, Canada. Ecology 79(7):2337-2351. Funk, R.E. 1988 The Laurentian Concept: A Review. Archaeology of Eastern North America 16:1-42. 1996 Holocene or Hollow Scene? The Search for the Earliest Archaic Cultures in New York State. The Review of Archaeology 17:11-24. Gaudreau, D.C. 1988 The Distribution of Late Quaternary Forest Regions in the Northeast: Pollen Data, Physiography, and the Prehistoric Period. In Holocene Human Ecology in 218 Northerastern North America, edited by G. P. Nicholas, pp. 215-256. Plenum: New York. Goodby, R. 1988 The Seabrook Phase and Post-Hopewellian Interaction in the Northeast. Unpublished Master‟s thesis, Department of Anthropology, Brown University, Providence, RI. Gramley, R.M. 1988 The Adkins Site: A Palaeo-Indian Habitation and Associated Stone Structure. Persimmon Press: Buffalo. Gribbin J. 1990 Hothouse Earth: The Greenhouse Effect and Gaia. Grove Weidenfield: New York. Griffin, J. B. 1964 The Northeast Woodlands Area. In Prehistoric Man in the New World, edited by J.D. Jennings and E. Norbeck, pp. 223-258. University of Chicago Press, Chicago. 1983 The Midlands. In Ancient North Americans, edited by J.D. Jennings, pp. 243301. W.H. Freeman and Co., New York. Hall, J.D. 1988 Treethrow as a Mass Wasting and Soil Forming Process in the Boston Mountains of Northern Franklin County, Arkansas. M.A. Thesis, University of Arkansas- Fayetteville. Hays, W.M. 1910 Farm Development. New York: Orange Judd Company. Herbster, H. 2004 Technical Report, Archaeological Data Recovery Program, East Pasture Site (19-DK-238), Aquinnah, Massachusetts. PAL Report No. 1530.01. Submitted to Lang Gerhard, Greenbrae, CA. Herbster, H., and S.G. Cherau 1999 Fieldwork Completion Report, Archaeological Data Recovery Program within the Herring Creek Archaeological District, Wampanoag Aquinnah Shellfish Hatchery (WASH) Project, Aquinnah, Massachusetts. Submitted to the Wampanoag Tribe of Gay Head (Aquinnah), Aquinnah, MA. 2001 Technical Report, Data Recovery Program, Lot 5/6 Site (19-DK-138), Lot 6 Building Envelope, Squibnocket Ridge, Chilmark, Massachusetts. PAL Report No. 1176. Submitted to Rosalie Catlin, Cambridge, MA. 2003 Technical Report, Intensive (Locational) Archaeological Survey and Data Recovery Program, Gerhard Site, Aquinnah Massachusetts. PAL Report No. 1100. Submitted to Lang Gerhard, Chilmark, MA. Hoffman, C.R. 1983 A Dated Feature Complex from Charlestown Meadows and its Implications for Regional Prehistory. Bulletin of the Massachusetts Archaeological Society 44(2):55-61. 219 Revising the Late Archaic Period in Southern New England. Archaeology of Eastern North America 13:58-78. 1990 People of the Fresh Water Lake. Peter Lang Publishing, Inc.: New York. Holmes, C.E., B.A. Potter, J.D. Reuther, O.K. Mason, R.M. Thorson, and P.M. Bowers 2008 Geological and Cultural Context of the Nogahabara I Site. American Antiquity 73(4)781-790. Ives, T.H. 2005 Busy in the Shadow of the Ossipee Mountains: Archaic Hornfels Workshops and a Paleoindian Site in Tamworth, NH. The New Hampshire Archaeologist 45(1):1-29. 2006a Summary Report of Field Investigations Conducted at Site 114-93, Locus I, Summer 2006. On file at the Mashantucket Pequot Museum and Research Center. 2006b The Stones Throw Site: A Late Paleoindian Site in East-Central New Hampshire. Northeast Anthropology 71:65-85. 2007a Summary Report: Phase I Archaeological Survey of 455 and 495 Norwich Westerly Road (Route 2), Preston, CT, Summer 2007. On file at the Mashantucket Pequot Museum and Research Center. 2007b Summary Report: Phase II Archaeological Investigation of Site 114-106 in Preston, CT, Summer 2007. Draft on file at the Mashantucket Pequot Museum and Research Center. 2007c Excavations at the Preston Plains Archaeological Site. Invited lecture presented at the March Meeting of the Massachusetts Archaeological Society, Central Massachusetts Chapter, Worchester, MA. 2007d Deep Holes and Deeper Questions: Ongoing Investigations of a Late Archaic Pit Feature Complex at Southeastern Connecticut‟s Preston Plains Site. Paper presented at the 72nd annual meeting of the Society for American Archaeology, Austin. 2009 Large Pit Feature Formation and Use at the Late Archaic Preston Plains Site in Southeastern Connecticut. Paper presented at the 74th annual meeting of the Society for American Archaeology, Atlanta. Johnson, D.L. 1989 Subsurface Stone Lines, Stone Zones, Artifact-Manuport Layers, and Biomantles produced by Bioturbation via Pocket Gophers. (Thomomys bottae). American Antiquity 54:370-389. Johnson, D.L., and D. Watson-Stegner 1990 The Soil-Evolution Model as a Framework for Evaluating Pedoturbation in Archaeological Site Formation. In Archaeological Geology of North America, Centennial Special Volume 4, edited by N.P. Lasca and J. Donahue, pp. 541560. Geological Society of America: Boulder. Johnson, E.S. 1993 Bifurcate Base Projectile Points in Eastern and Central Massachusetts: Distribution and Raw material. Bulletin of the Massachusetts Archaeological Society 54(2):46-55. 220 1985 Jones, B.D. 1999 Preston Plain Energy Center, Interim Report of Preliminary Archaeological Testing. Memorandum on file at the Mashantucket Pequot Museum and Research Center. 2000 Continuity Versus Change during the Last Three Millennia at Mashantucket. Northeast Anthropology 64: 17-29. 2004 Paleoindian Population Dynamics in New England: Possible Typological Consequences. In Hunters and Gatherers in Theory and Archaeology, edited by G.M. Crothers, pp. 48-67. Center for Archaeological Investigations, Occassional Paper No. 31, Southern Illinois University. 2006 A 3-Culture System Model for Early Holocene New England. Paper presented at the 46th annual meeting of the Northeastern Anthropological Association. 2007 The Simple and the Complex: Two Middle Archaic Small Upland Lithic Sites in North Stonington, Connecticut. In Current Approaches to the Analysis and Interpretation of Small Lithic Sites in the Northeast, edited by C. B. Reith, pp 77-88. New York State Museum: Albany. Jones, B.D., and D.T. Forrest 2003 Life in a Postglacial Landscape: Settlement-Subsistence Change During the Pleistocene-Holocene Transition in Southern New England. In Geoarchaeology of Landscapes in the Glaciated Northeast, pp. 75-89. New York State Museum Bulletin 497, edited by D.L. Cremeens and J.P. Hart, University of the State of New York, The State Education Department, Albany. Juli, H.D., and K.A. McBride 1984 The Early and Middle Woodland Periods of Connecticut Prehistory: Focus on the Lower Connecticut River Valley. Bulletin of the Archaeological Society of Connecticut 47:89-98. Kooi, P.B. 1974 De Orkann Van 15 November 1972 en Het Ontstaan Van “Hoefijzervormige” Grondsporen. Helinium 14:57-65. Krigström, A. 1962 Geomorphological Studies of Sandur Plains and Their Braided Rivers in Iceland. Geogra ska Annaler 44:328–346. Langhor, R. 1993 Types of Tree Windthrow, Their Impact on the Environment and Their Importance for Understanding Archaeological Excavation Data. Helinium 33:36-49. Largy, T., and D. DeMello 2010 Plant Remains from the Preston Plains Site, Locus 1, New London County, Preston, Connecticut. Report on file at the Mashantucket Pequot Museum and Research Center. Lavin, L. 1988 Coastal Adaptations in Southern New England and Southern New York. Archaeology of Eastern North America 16:101-120. Lavin, L., and L.W. Russell 221 Excavation of the Burwell-Karako Site: New data on Cultural Sequences and Artifact Typologies in Southern New England. Bulletin of the Archaeological Society of Connecticut 48:45-87. Leveillee, A. 2002 An Old Place, Safe and Quiet: A Blackstone River Valley Cremation Burial Site. Bergin & Garvey, Westport, CT. Luedtke, B.E. 1987 The Pennsylvania Connection: Jasper in Massachusetts Archaeological Sites. Bulletin of the Massachusetts Archaeological Society 48(1):37-47. Lutz, H.J. 1940 Disturbance of Forest Soil Resulting from the Uprooting of Trees. Yale School of Forestry Bulletin 45. 1960 Movement of Rocks by Uprooting of Forest Trees: American Journal of Science 258:752-756. McBride, K.A. 1978 Archaic Subsistence in the Lower Connecticut River Valley: Evidence from Woodchuck Knoll. Man in the Northeast 5:124-131. 1984 Middle and Late Archaic Periods in the Connecticut River Valley: A Reexamination. Bulletin of the Archaeological Society of Connecticut 47: 55-72. MacDonald, G.F. 1968 The Debert Paleoindian Site in Central Nova Scotia. Anthropology Paper No. 16. National Museums of Canada, Ottawa. Malde, H.E. 1964 The Ecological Significance of Some Unfamiliar Geologic Processes. In The Reconstruction of Past Environments, assembled by J.J. Hester and J. Schoenwetter, pp. 7-13. Publication of the Fort Burgwin Research Center, No. 3. Taos: New Mexico. McWeeney, L. 1999 A Review of Late Pleistocene and Holocene Climate Changes in Southern New England. Bulletin of the Archaeological Society of Connecticut 62:3-18. Mergen, F. 1954 Mechanical Aspects of Wind-Breakage and Wind-Firmness. Journal of Forestry 52:119-125. Mueller, R.G., and J.C. Cavallo 1995 Multiple Working Hypotheses: An Analysis, Pit Houses or Tree Throws. Paper presented at the 1995 ESAF Conference, Wilmington, Delaware. Mueller, R.G., K. Steck, and J. Godin 1997 Bioturbation by Tree-throw: An Alternative Explanation for Pit House Depressions in the Mid-Atlantic Region. Presented at the annual meeting of the Society for American Archaeology, Nashville, Tennessee. Mueller, O.P., and M.G. Cline 1959 Effects of Mechanical Soil Barriers and Soil Wetness on Rooting of Trees and Soil-Mixing by Blow-down in Central New York. Soil Science 88:107-111. Newell, R. 222 1985 Mesolithic Dwelling Structures: Fact or Fantasy. Veröffentlichungen des Museums für Ur- und Frühgeschichte Potsdam 14/15:235-284. Nicholas, G.P. 1987 Rethinking the Early Archaic. Archaeology of Eastern North America 15:99123. 1988 Ecological Leveling: The Archaeologial and Environmental Dynamics of Early Postglacial Land Use. In Holocene Human Ecology in Northeastern North America, edited by George P. Nicholas, pp.257-288. Plenum Press, NY. Nixon, S. 2004 Marine Resources and the Human Carrying Capacity of Coastal Ecosystems in Southern New England before European Contact. Northeast Anthropology 68: 1-23. Ouimet, W.B., T. Sibley, and J.R. Swanson, Jr. 2010 Tree Throw Erosion in Western Massachusetts Associated with the December 2008 Ice Storm. Paper No.12-1. Presented at the Northeastern Section (45th Annual) and Southeastern Section (59th Annual) Joint Meeting (13-16 March 2010) of the Geological Society of America. Oswald, W.W., E.K. Faison, D.R. Foster, E.D. Doughty, B. Hall, and B.C.S. Hansen 2007 Post-Glacial Changes in Spatial Patterns of Vegetation Across Southern New England. Journal of Biogeography 34:900-913. Patton, P.C., and J.M. Kent 1992 A Moveable Shore. Duke University Press: Durham. Pelletier, B.G., and B. Robinson 2005 Tundra, Ice and A Pleistocene Cape on the Gulf of Maine: A Case of Paleoindian Transhumance. Archaeology of Eastern North America 33:163176. Perry, D. 2002 A Synopsis of Recent Identifications of Soft Plant Tissues from Prehistoric Contexts at Mashantucket. Report on file at the Mashantucket Pequot Museum and Research Center. Perry, D., and B.D. Jones 2002 A Synopsis of Recent Identifications of Soft Plant Tissues from Prehistoric Contexts at Mashantucket. Paper Presented at the 25th Annual Ethnobiology Conference, Storrs, CT. Petersen, J.B. 1991 Archaeological Testing at the Sharrow Site: A Deeply Stratified Early to Late Holocene Cultural Sequence in Central Maine. Occasional Publications in Maine Archaeology 8. Maine Historic Preservation Commission and Maine Archaeological Society. Augusta, ME. Petersen, J.B., and D.E. Putnam 1992 Early Holocene Occupation in the Central Gulf of Maine Region. In Early Holocene Occupation in Northern New England, edited by B. Robinson, J. Petersen and A. Robinson, pp. 13-62. Occasional Papers in Maine Archaeology 9. Maine Historic Preservation Commission, Augusta. 223 1981 Petersen, J.B., N.D. Hamilton, R.A. Doyle, and D. Sanger 1986 Late Paleoindian Remains from Maine. Current Research in the Pleistocene 3:19-21. Petersen, J.B., R.N. Bartone, and B.J. Cox 2000 The Varney Farm Site and the Late Paleoindian Period in Northeastern North America. Archaeology of Eastern North America 28:113-140. Petraglia, M.D., and K. Cunningham 2006 Native American Collaboration in the Delmarva: New Meanings and an Expanded Approach to Delaware Archaeology. In Cross Cultural Collaboration: Native Peoples and Archaeology in the Northeastern United States, edited by J.E. Kerber, pp 213-232. University of Nebraska Press: Lincoln. Petraglia, M.D., S.L. Bupp, S.P. Fitzell and K.W. Cunningham. 2005 Hickory Bluff: Changing Perceptions of Delmarva Archaeology. Delaware Department of Transportation Archaeology Series No. 175. Pfeiffer, J.E. 1984 The Late and Terminal Archaic Periods of Connecticut Prehistory. Bulletin of the Archaeological Society of Connecticut 47:73-88. 1986 Dill Farm Locus 1: Early and Middle Archaic Components in Southern New England. Bulletin of the Archaeological Society of Connecticut 49:19-35. Phillips, J.D., and D.A. Marion 2003 Pedological Memory in Forest Development. Forest Ecology and Management 188(1-3):363-380. Putz, F.E. 1983 Treefall Pits and Mounds, Buried Seeds, and the Importance of Soil Disturbance to Pioneer Trees on Barro Colorado Island, Panama. Ecology 64:1069-1074. Pyddoke, E. 1961 Stratification for the Archaeologist. Phoenix House, London. Reeve, S.A., and K. Forgacs 1999 Connecticut Radiocarbon Dates: A Study of Prehistoric Cultural Chronologies and Population Trends. Bulletin of the Archaeological Society of Connecticut 62: 19-66. Renfrew, C. 1976 Archaeology and the Earth Sciences. In Geoarchaeology: Earth Science and the Past, edited by Rover G. Walker, pp 119-140. Geoscience Canada, Reprint Series 1. Geological Association of Canada, Toronto. Richard, P.J.H., and S. Occhietti 2005 14C Chronology for Ice Retreat and Inception of Champlaign Sea in the St. Lawrence Lowlands, Canada. Quaternary Research 63:353-358. Ritchie, W.A. 1961 A Typology and Nomenclature for New York Projectile Points. New York State Museum and Science Service Bulletin, 384. 1965 The Archaeology of New York State. Natural History Press: Garden City, NY. 224 The Archaeology of Martha’s Vineyard. Natural History Press: Garden City, NY. Rittenour, M.T n.d. Glacial Lake Hitchcock. http://www.bio.umass.edu/biology/conn.river/hitchcock.html Downloaded from the World Wide Web on April 15, 2007. Robinson, B.S. 1992 Early and Middle Archaic Period Occupation in the Gulf of Maine Region: Mortuary and Technological Patterning. In Early Holocene Occupation in Northern New England, edited by B. Robinson, J. Petersen and A. Robinson, pp 63-116. Occasional Papers in Maine Archaeology 9. Augusta. 1996 A Regional Analysis of the Morehead Burial Tradition: 8500-3700 B.P. Archaeology of Eastern North America 24:95-148. 2003 Multiple Boundaries of the Moorehead Burial Tradition. Northeast Anthropology 66:15-27. Robinson, B.S., and J.B. Petersen 1993 Perceptions of Marginality: The Case of the Early Holocene in Northern New England. Northeast Anthropology 46:61-75. Robinson, B.S., J.C. Ort, W.A. Eldridge, A.L. Burke, and B.G. Pelletier 2009 Paleoindian Aggregation and Social Organization at Bull Brook. American Antiquity 74(4):423-447. Rossen, J., and T. Dillehay 2001 Seasonality and Scheduling of Sedentary Hunter-Gatherers: The Case of Monte Verde, Chile. Paper Presented at the 18th Annual Visiting Scholar's Conference "Hunters and Gatherers in Theory and Archaeology." Center for Archaeological Investigations, Southern Illinois University Carbondale, March 23-24. Rowlands, W. 1939 A Study of Damage Caused by the 1938 Hurricane in the Vicinity of Petersham, Massachusetts. Harvard Forest Archives: .HF 1939-05 Sanger, D. 1973 Who Were the Red Paints? Maine Archaeological Society Bulletin 13(1):3-12. Sanger, D., W.R Belcher, and D.C. Kellogg 1992 Early Holocene Occupation at the Blackman Stream Site, Central Maine. In Early Holocene Occupation in Northern New England, edited by B. Robinson, J. Petersen and A. Robinson, pp 149-161. Occasional Papers in Maine Archaeology 9. Augusta. Schaetzl, R.J. 1990 Effects of Treethrow Microtopography on the Characteristics and Genesis of Spodosols, Michigan, USA. Catena 17:111-126. Schaetzl, R.J., and L.R. Follmer 1990 Longevity of Treethrow Microtopography: Implications for Mass Wasting. Geomorphology 3:113-123. Schaetzl, R.J., S.F. Burns, T.W. Small, and D.L. Johnson 1969 225 Tree Uprooting: Review of Types and Patterns of Soil Disturbance. Physical Geography 11:277-291. Schiffer, M.B. 1987 Formation Processes of the Archaeological Record. University of New Mexico Press: Albuquerque. Schoenwetter, J. 1981 Prologue to a Contextual Archaeology. Journal of Archaeological Science 8:367-369. Sgarlata, C. 2009 The Spatial Analysis of Mobile Hunter-Gatherers Campsites in Southern New England and the Potential Impact of Open Source Resources. Bulletin of the Archaeological Society of Connecticut 71:41-60. Snow, D.R. 1972 Rising Sea Level and Prehistoric Cultural Ecology in New England. American Antiquity 37(2):211-221. 1980 The Archaeology of New England. Academic Press: New York. Speiss, A.E., and R.A. Lewis 2001 The Turner Farm Fauna: 500 Years of Hunting and Fishing in Penobscot Bay, Maine. Occasional Publications in Maine Archaeology, No. 11. Augusta, Maine. Speiss, A., D. Wilson, and J. Bradley 1998 PaleoIndian Occupation in the New England-Maritimes Region: Beyond Cultural Ecology. Archaeology of Eastern North America 26:201-264. Stokes, A., A.H. Fitter and M.P. Cloutts 1995 Responses of Young Trees to Wind and Shading Effects on Root Architecture. Journal of Experimental Botany 45(290):1139-1146. Stone et al. 1998 Glacial Lake Hitchcock. Quarternary Geologic Map of Connecticut and Long Island Sound. U.S. Geological Survey Open File Report OF 98-371. Strauss, A.E. 1978 Nature‟s Transformations and Other Pitfalls: Toward a Better Understanding of Post-Occupational Changes in Archaeological Site morphology in the Northeast. Part I: Vegetation. Bulletin of the Massachusetts Archaeological Society 42(1):2-11. 1992 Argillite Blade Cache Found Near Sewall‟s Falls (NH 31-30). The New Hampshire Archaeologist 26(1):81-86. Streuver, S. and K. Vickery 1973 The Beginnings of Cultivation in the Midwest-Riverine Area of the United States. American Anthropologist 75:1197-1219. Sullivan III, A.P. 1995 Artifact Scatters and Subsistence Organization. Journal of Field Archaeology 22(1):49-64. Swanson, F.J., R.L. Fredrickson, and F.M. McCorison 226 1990 Material Transfer in Western Oregon Forested Watershed. In Analysis of Coniferous Forest Ecosystems in the Western United States, edited by R.L. Edmonds. Hutchinson Ross Publication Company. Taylor, W.B. 1976 A Bifurcated-Point Concentration. Bulletin of the Massachusetts Archaeological Society 37(3):36-41. Thomas, R.A. 1981 Excavations at the Delaware Park Site (7NC-E-41). Delaware Department of Transportation Archaeological Series No. 11, Dover. Thomas, R.A., and T.M. Payne 1981 Archaeological Data Recovery at the Hollingsworth Farm Site (18CE29), Cecil County, Maryland. On file, Maryland State Historic Preservation Office. Thorson, R.M. 1990 Geoarchaeological Contexts of Archaeological Sites in Beringia. In Archaeological Geology of North America, edited by Norman P. Lasca and Jack Donahue, pp.399-420, Centennial Special Volume 4, Geological Society of America, Boulder. Thorson, R.M., and K. McBride 1988 The Bolton Spring Site, Connecticut: Early Holocene Human Occupation and Environmental Change in Southern New England. Geoarchaeology 3:221-234. Thorson, R.M., and R.S. Webb 1991 Postglacial History of a Cedar Swamp in Southeastern Connecticut. Journal of Paleolimnology 6:17-35. Tomirdiaro, S.V. 1978 Cryogenous-Eolian Genesis of Yedoma Deposits. Pp. 817-18 in Permafrost: Second International Conference, edited by Frederick J. Sanger and Peter J. Hyde. National Academy of Sciences: Washington, D.C. Troedsson, T., and W.H. Lyford 1973 Biological Disturbance and Small-Scale Spatial Variation in a Forest Soil Near Garpenberg, Sweden. Studia Forestalia Suecica 109:1-23. Turnbaugh, W. 1975 Toward and Explanation of the Broadpoint Dispersal in Eastern North America. Journal of Anthropological Research 31:51-68. Tuck, J.A. 1978 Regional Cultural Development, 3000 to 300 B.C. In Handbook of North American Indians, vol. 15, edited by Bruce G. Trigger, pp 28-43. Smithsonian Institution, Washington, D.C. Uchupi, E., N. Driscoll, R.D. Ballard, and S.T. Bolmer 2001 Drainage of Late Wisconsin Glacial Lakes and the Morphology and Late Quaternary Stratigraphy of the New Jersey - Southern New England Continental Shelf and Slope: Marine Geology 172:117-145. Veneman, P.L.M., P.V. Jacke, and S.M. Bodine 1984 Soil Formation as Affected by Pit and Mound Microrelief in Massachusetts, USA. Geoderma 33:89-99. 227 1982 Walter, K.M., M.E. Edwards, G. Grosse, S.A. Zimov, and F.S. Chapin III 2007 Thermokarst Lakes as a Source of Atmospheric C14 During the Last Deglaciation.. Science 318:633-636. Wandsnider, L. 1987 Natural Formation Process Experimentation and Archaeological Analysis. In Natural Formation Processes and the Archaeological Record, edited by D.T. Nash and M.D. Petraglia, BAR International Series 352: 150-185. Waters, M.R. 1992 Principles of Geoarchaeology: A North American Perspective. University of Arizona Press: Tucson. Wendorf, M. 1982 The Fire Areas of Santa Rosa Island: An Interpretation. North American Archaeologist 3:173-180. Wessels, T. 1997 Reading the Forested Landscape: A Natural History of New England. Countryman Press: Woodstock, VT. White, J.W.C., and L.K. Barlow 1993 The “Flickering Switch” of Late Pleistocene Climate Change. Nature 361:432436. Whittaker, W.E., M.T. Dunne, J.A. Artz, S.E. Horgen, and M.L. Anderson 2007 Edgewater Park: A Late Archaic Campsite along the Iowa River. Midcontinental Journal of Archaeology 31(1):5-46. Winkler, M.G. 1985 A 12,000 Year History of Vegetation and Climate Change for Cape Cod, Massachusetts. Quaternary Research 23:301-312. Wood, W.R., and D.L. Johnson 1978 A Survey of Disturbance Procecees in Archaeological Site Formation. In Advances in Archaeological Method and Theory, vol. 1, edited by M.B. Schiffer, pp.315-381. Academic Press, New York. Yu, Z., and U. Eicher 1998 Abrupt Climate Oscillations during the Last Deglaciation in Central North America. Science 282:2235-2238. Yu, Z., J. H. McAndrews, and U. Eicher 1997 Middle Holocene Dry Climate Caused by Change in Atmospheric Circulation Patterns: Evidence from Lake Levels and Stable Isotopes. Geology 25:251-254. 228