RS 17 Human Adaptation in the Grand Marais - University of Arkansas

ARKANSAS ARCHEOLOGICAL SURVEY RESEARCH SERIES 17 

W. Fredrick Limp, Series Editor 

Fayetteville, Arkansas 

1982

Library of Congress #82-620028 

Printed in the U.S.A. at Fayetteville, Arkansas 

Production Editor: Mary Lynn Kennedy 

Graphic Artist: Jane Kellett 

Photographer: Pamela Ashford 

Typesetter: Louise Mullins 

2006 Digital Reprint: Lindi J. Holmes 

Report submitted to 

U.S. Army Corps of Engineers 

Vicksburg District

AbstrAct 

The Felsenthal Project is a multistage interdisciplinary research project carried out 

by the Arkansas Archeological Survey for the U.S. Army Corps of Engineers, Vicksburg 

District. The designated project area is the Felsenthal Navigation Pool, a portion of the 

Ouachita and lower Saline River floodplains below 65 feet elevation. A broad based research 

design was developed prior to fieldwork and emphasized four research domains: 

settlement characteristics, locational characteristics, culture history, and site transformations. 

Eight hypotheses relevant to regional prehistory and history are subsumed by 

these domains. 

The field portion of the project included overlapping intensive site survey and testing 

stages in late 1979; these stages were followed by intensive analysis at the Survey’s 

Contract Laboratory and by various specialists’ studies in 1980. 

The project area falls within the Grand Marais Lowland, which is closely related 

geohydrologically and ecologically to the Lower Mississippi Alluvial Valley. The floodplain 

here is an overflow bottomland consisting of an intricate mosaic of swamp and 

woodland forest, river channels, open lakes, bayous, and backswamp drainages. The 

Ouachita River is the lowest elevation in the state of Arkansas at the Louisiana line (6 

km south of the project area), and the prehistoric and historic sites located and studied 

in 1979 are the lowest known sites presently recorded in Arkansas. 

Our environmental and cultural data indicate that the floodplain contained abundant 

and diverse aquatic and riparian resources of fundamental importance to prehistoric 

populations aggregated on adjacent upland terraces. A distinctive characteristic of 

the floodplain ecosystem is the annual prolonged hydroperiod, or extensive deep submergence 

of lowlands during winter and spring. This hydroperiod “pulse” introduces 

and cycles nutrients to floodplain habitats, and imposes a strong seasonal dimension on 

biological (including human) activity. During the historic period, Euramericans never 

permanently settled or farmed the overflow bottomland, and it is today encompassed by 

the Felsenthal National Wildlife Refuge. 

Site survey activity, designed for wetland and forested conditions, consisted principally 

of floodplain transects and river or tributary bankline surveys. A total of 144 sites, 

including 132 prehistoric and 15 historic sites was recorded (these totals reflect the occurrence 

of both prehistoric and historic components on three sites). Prehistoric 

i

floodplain sites are predominantly small extractive camps with relatively sparse artifact 

and debris content. The nature of such extractive camps and their relationship to regional 

subsistence and settlement systems are a major focus of this report. About one-third 

of all floodplain sites provided depth data for occupational levels or artifacts in place, 

and a preliminary estimate of local rates of recent alluviation has been developed. About 

two-thirds of all floodplain sites furnished typological, technological, or stratigraphic 

data sufficient to identify cultural components. The cultural sequence extends from Late 

Archaic into the protohistoric and historic periods, or about 3000 years. Earlier sites are 

deeply buried or drowned by river level and high water table. The best available site 

sample comprises extractive camps of the Mississippi period (A.D. 1100-1700), including 

a dramatic concentration of such camps on the lower Saline River. One major conclusion 

of this report is the importance of small Mississippi period extractive sites, which are not 

farmsteads, as components of a regional subsistence-settlement system. Such sites have 

been little studied in eastern North America. 

Historic sites recorded during survey include steamboat landings, earth mounds 

built as raised platforms for buildings or lock construction facilities, an underwater 

steamboat wreck, and other sites related to riverine and extractive industry during the 

nineteenth and early twentieth centuries. 

The testing program involved eight prehistoric floodplain sites (and would have 

expanded to other prehistoric and historic sites, if not for the rise of river level above 

project contour in late November). Sites tested include a relatively deep stratified site 

on the Ouachita River with Poverty Point, Tchula period, and later occupation levels, 

two Mississippi period middens with evidence of bulk fish processing, two Mississippi 

period extractive sites with multiple activity loci, a small Baytown-Coles Creek period 

site, a protohistoric site with a Quapaw phase assemblage previously unknown in the 

region, and another possible Poverty Point period site. The last seven sites are all located 

on the lower Saline River. 

Tactics for site testing included controlled surface collection where riverbank sites 

were well exposed, subsurface testing of about 5% of known site area, and power auger 

or shovel testing to determine extent of buried occupation levels. Screening and 

wet screening recovery techniques were employed, and soil and organic samples were 

obtained for all sites tested. The results of the testing provided considerable data for the 

nature of floodplain extractive sites and variation among them, as well as an empirical 

basis for assessing significance. 

ii

Analyses of about 13,100 artifacts and items of debris from site survey and testing 

provided a variety of useful regional data, some preliminary in nature, but suggesting 

fruitful directions for future work. The more significant results of these analyses would 

include the following. 

1. partial temporal sequences for both arrow point and dart point styles 

2. functional studies of lithic reduction and lithic resources 

3. detailed analysis and comparison of Gran Marais phase (A.D. 1100-1400) ceramic 

assemblages from two sites, and 

4. ceramic seriation data for Mississippi period sites, based on temper differences. 

The limited program of soil and sediment analysis indicates that organic materials 

are deficient in most floodplain sites largely because of extreme acidity. High levels 

of organic carbon and phosphorus can occur in anthropogenically enriched soils or 

middens. Radiocarbon dating of floodplain sites has been markedly unsuccessful. Five 

samples from two sites gave ages from 500 to 3000 years earlier than expected. The dimensions 

of this special geochronological problem are discussed in detail. 

In the concluding chapter two topical areas of particular research interest, in the 

context of environment and culture in the Grand Marais Lowland, are expanded. The 

first area includes discussion of the archeological context and dimensions of seasonal or 

transient extractive camps and the material correlates of extractive site activity. These 

sites are distinguished by high research potential in many areas of current archeological 

and scientific interest. The second area discusses a seasonal model for extractive activity 

and resource use in an overflow bottomland environment. The variety of extractive 

activities, especially fishing, continues throughout the year, even during prolonged deep 

inundation of the floodplain. This model draws upon the best available project data 

from Mississippi period extractive camps on the Saline River and from ethnohistoric 

data for the Atchafalaya Basin in south-central Louisiana, an ecologically similar overflow 

bottomland environment. 

iii

tAble of contents 

Abstract ............................................................................................................................................i 

List of Tables .................................................................................................................................ix 

List of Figures ............................................................................................................................ xiii 

Management Summary .............................................................................................................xix 

Acknowledgments .................................................................................................................. xxiii 

Chapter 1. Introduction to the Felsenthal Project .................................................................. 1 

The Felsenthal Project and Its Research Potential .............................................. 1 

Elevation Constraint ......................................................................................... 2 

Definition of Study Area .................................................................................. 3 

Development of Research Design ......................................................................... 6 

Schedule of Investigations ............................................................................... 6 

Schedule of Hypotheses ................................................................................... 8 

Felsenthal Project Personnel ................................................................................ 10 

Chapter 2. Grand Marais Lowland Environment ................................................................ 13 

Southeastern Floodplain Ecosystems ................................................................. 15 

Riparian Ecosystems ....................................................................................... 15 

Aquatic Ecosystems ........................................................................................ 17 

Pleistocene Terrace Sequence .............................................................................. 19 

Recent Floodplain Geohydrology ....................................................................... 24 

Flood Characteristics ...................................................................................... 24 

Floodplain Features ........................................................................................ 25 

Topstratum/Substratum Deposits ................................................................ 28 

Floodplain Soils ............................................................................................... 29 

Channel Geometry and Channel Deposits .................................................. 31 

Riverbanks and Bank Erosion ....................................................................... 34 

Recent Floodplain Ecology .................................................................................. 36 

Hardwood and Swamp Forest ...................................................................... 36 

Vertebrates ....................................................................................................... 41 

Invertebrates .................................................................................................... 45 

Potential Paleoecological Studies ........................................................................ 46 

Potential Lowland Resources .............................................................................. 47 

Chapter 3. Prehistory of the Felsenthal Region, by John H. House .................................. 51 

Introduction ........................................................................................................... 51 

Previous Archeological Investigation ................................................................ 52 

Explorations ..................................................................................................... 52 

Recent Investigations ...................................................................................... 54 

v

The Sequence of Aboriginal Occupation ........................................................... 57 

Paleo-Indian Period, before 8000 B.C. .......................................................... 58 

Early Archaic Period, 8000-6000 B.C. ........................................................... 60 

Middle Archaic Period, 6000-4000 B.C. ........................................................ 61 

Late Archaic Period, 4000-1200 B.C. ............................................................. 62 

Poverty Point Period, 1200-500 B.C. ............................................................. 63 

Tchula Period, 500 B.C.-A.D. 1 ...................................................................... 64 

Marksville Period, A.D. 1-300 ....................................................................... 65 

Baytown Period, A.D. 300-700 ....................................................................... 66 

Coles Creek Period, A.D. 700-1100 ............................................................... 67 

Mississippi Period, A.D. 1100-1700 .............................................................. 68 

Discussion 

Chapter 4. History of the Felsenthal Region, by Beverly Watkins .................................... 79 

Introduction ........................................................................................................... 79 

European Exploration and Settlement ............................................................... 79 

American Exploration and Settlement ............................................................... 80 

Marie Saline Landing ..................................................................................... 83 

Disposition of Swamplands ................................................................................. 85 

River Commerce and Travel ................................................................................ 85 

Effects of the Civil War ......................................................................................... 87 

Improvements in Navigation .............................................................................. 87 

The Sternwheeler Lotawanna ......................................................................... 90 

Lumbering and Railroads .................................................................................... 94 

The Felsenthal Community ................................................................................. 96 

Conclusions ............................................................................................................ 97 

Chapter 5. Variability Among Floodplain Sites ................................................................... 99 

Survey Strategy and Tactics ............................................................................... 101 

Random and Judgmental Transects ........................................................... 105 

River Bankline Surveys ................................................................................ 112 

Tributary Bankline and Backswamp Surveys ........................................... 116 

Site Specific Surveys ..................................................................................... 117 

Site and Component Identification ............................................................ 117 

Survey Results ..................................................................................................... 121 

Unit 1: Ouachita River Above Mile 254 .................................................... 121 

Unit 2: Lower Saline River .......................................................................... 126 

Unit 3: Lower Eagle Creek .......................................................................... 130 

Unit 4: Ouachita River Below Mile 254 ..................................................... 130 

Unit 5: Lower Lapile Creek ........................................................................ 132 

Unit 6: Oxbow Lakes and Backswamp ..................................................... 133 

Unit 7: Pleistocene Terraces and Islands ................................................... 136 

Unit 8: Historic Sites .................................................................................... 138 

Evaluation of Survey Strategy and Tactics ...................................................... 146 

Site Locational and Historical Hypotheses ..................................................... 150 

vi

Chapter 6. Assessment of Floodplain Sites ......................................................................... 155 

Selection of Sites for Testing .............................................................................. 155 

Test Excavation Strategy and Tactics ................................................................ 157 

Marie Saline (3AS329) .................................................................................. 159 

False Indigo (3AS285) ................................................................................... 168 

Buttonbush (3BR58) ...................................................................................... 175 

One Cypress Point (3AS286) ....................................................................... 178 

Jug Point Cutoff (3BR76) .............................................................................. 185 

Jug Point 1 (3AS306) ..................................................................................... 187 

Jug Point 2 (3AS307) ..................................................................................... 191 

Mouth of Eagle Creek (3BR78) .................................................................... 194 

Evaluation of Test Excavation Strategy and Tactics ....................................... 194 

Summary Assessment of Floodplain Sites ....................................................... 197 

Site Functional Hypotheses ............................................................................... 197 

Variation Among Extractive Sites ............................................................... 201 

Chapter 7. Soils, Sediments, and Chronology .................................................................... 205 

Selected Properties of Floodplain Soils ............................................................ 205 

Sediment Textures and Floodplain Alluviation .............................................. 208 

Radiocarbon Dates for Floodplain Sites........................................................... 212 

Site Transformational Hypotheses .................................................................... 216 

Chapter 8. Lithic and Ceramic Analysis .............................................................................. 219 

Lithic Source Study ............................................................................................. 219 

Methods of Collection and Analysis .......................................................... 221 

Results of Analysis ........................................................................................ 222 

Lithic Analysis and Results ................................................................................ 224 

Arrow Points .................................................................................................. 224 

Dart Points ..................................................................................................... 230 

Bifacial Preforms and Other Bifaces ........................................................... 234 

Other Flaked Stone Tools ............................................................................. 237 

Cores and Debitage ....................................................................................... 240 

Cobble and Heavy Duty Tools .................................................................... 244 

Ground Stone Tools ...................................................................................... 245 

Mineral Pigments .......................................................................................... 250 

Inorganic Debris ............................................................................................ 250 

Ceramic Analysis and Results ........................................................................... 251 

Tchefuncte Complex ..................................................................................... 252 

Baytown Complex ........................................................................................ 254 

Gran Marais Complex .................................................................................. 255 

Shell-tempered Complex ............................................................................. 263 

Caddoan Imports .......................................................................................... 266 

Culture Historical Hypotheses .......................................................................... 266 

vii

Chapter 9. Human Adaptation in the Grand Marais Lowland ....................................... 269 

On the Nature of Extractive Sites ...................................................................... 269 

Site Size ........................................................................................................... 271 

Site Configuration ......................................................................................... 272 

Site Assemblage ............................................................................................. 272 

Site Residue .................................................................................................... 273 

Site Location .................................................................................................. 273 

A Model for Resource Use in an Overflow Bottomland ................................ 275 

References Cited ........................................................................................................................ 279 

Appendix A. Site Descriptive and Analytical Data .........................................................A-1 

Prehistoric Sites .............................................................................................A-1 

Historic Sites ..................................................................................................A-2 

Appendix B. Faunal and Floral Analysis .......................................................................... B-1 

Vertebrate Fauna ........................................................................................... B-1 

Molluscan Fauna ........................................................................................... B-4 

Carbonized Plant Remains .......................................................................... B-6 

Subfossil Wood .............................................................................................. B-6 

Appendix C. Recommended Program for Mitigation of Impacts on Archeological 

Resources in the Felsenthal Navigation Pool ............................................C-1 

Introduction ...................................................................................................C-1 

Significance of Floodplain Sites ..................................................................C-4 

National Register Evaluation ....................................................................C-10 

Assessment of Site Impacts .......................................................................C-14 

Recommended Program of Mitigation ....................................................C-17 

General Recommendations .......................................................................C-29 

viii

list of tAbles 

Table 1. Resource Potential of Floodplain Trees, Shrubs, and Vines ............................... 40 

Table 2. Floodplain Mammals in Order of Body Weight .................................................. 42 

Table 3. Selected Food Fishes and Habitat Preferences, Lower Ouachita River 

System in Arkansas .................................................................................................. 44 

Table 4. Comparison of Archeological Sequences in the Felsenthal, Bartholomew- 

Macon, Mid-Ouachita, and Tensas Regions ......................................................... 59 

Table 5. Power Equipment and Accessories Employed for Wetland Archeological 

Survey in the Felsenthal Project Area .................................................................. 102 

Table 6. Characteristics of Nine Judgmental Transects in the Grand Marais and 

Ouachita-Saline Channels Subareas .................................................................... 111 

Table 7. Channel Geometry and Prehistoric Site Distribution for Three Site Units 

on Ouachita and Saline Rivers ............................................................................. 123 

Table 8. Density of Prehistoric Riverbank Components in Three Site Units on 

Ouachita and Saline Rivers ................................................................................... 151 

Table 9. Summary of Test Excavation Tactics for Floodplain Sites ................................ 158 

Table 10. Total Recovery of Imperishable Artifacts and Debris by Occupation Area, 

Marie Saline Site (3AS329) .................................................................................... 167 


False Indigo (3AS285) ............................................................................................ 175 


One Cypress Point (3AS286) ................................................................................. 183 

ix


Jug Point 1 (3AS306) ............................................................................................... 191 


Jug Point 2 (3AS307) ............................................................................................... 193 

Table 15. Summary Assessment of Eight Floodplain Sites ............................................... 198 

Table 16. Proportions of Five Functional Categories of Imperishable Artifacts and 

Debris in Nine Prehistoric Components (Six Floodplain Sites) ....................... 203 

Table 17. Selected Properties of Soil Samples from Seven Floodplain Sites ................... 206 

Table 18. Particle Size Frequencies by Depth in Sediment Samples from Test Pits 1 

and 6, Area C, Marie Saline Site (3AS329) .......................................................... 209 

Table 19. Characteristics and Results of Age Determination for Radiocarbon Samples 

from Marie Saline (3AS329) and Jug Point 2 (3AS307) ..................................... 213 

Table 20. Proposed Sequence of Arrow Point Types in the Felsenthal Region .............. 226 

Table 21. Metric and Nonmetric Data for Dart Points and Proposed General Sequence 

of Dart Point Types in the Felsenthal Region ..................................................... 231 

Table 22. Comparison of Debitage by Reduction Category and Raw Material 

Frequency among Three Floodplain Site Assemblages .................................... 244 

Table 23. Descriptive Classification of Plainware from Two Sites of the Gran Marais 

Phase, Mississippi Period ...................................................................................... 260 

Table 24. Descriptive Classification of Decorated Sherds from Two Sites of the Gran 

Marais Phase, Mississippi Period ..................................................................261-262 

Table 25. Proportions of Shell and Clay Tempering in Ceramic Samples from Four 

Floodplain Sites of the Mississippi Period .......................................................... 266 

Table 26. Three Classifications of Settlement Type which Emphasize or Distinguish 

Small Extractive Sites ............................................................................................. 270 

x

Appendix A 

Table A-1. Prehistoric Sites (Units 1-7) ................................................................................A-4 

Table A-2. Historic Sites (Unit 8) ..........................................................................................A-9 

Table A-3. Occurrence of Temporally or Culturally Diagnostic Artifacts in 

Prehistoric Floodplain Sites .............................................................................A-10 

Appendix b 

Table B-1. Vertebrate Remains from Prehistoric Floodplain Sites .................................. B-2 

Table B-2. Molluscs Identified in Prehistoric and Historic Site Samples, and 

in a Modern Sample from Felsenthal National Wildlife Refuge .................. B-5 

Table B-3. Floral Remains from Prehistoric and Historic Floodplain Sites ................... B-7 

Table B-4. Wood and Subfossil Wood Specimens from Environmental Localities ...... B-8 

Appendix c 

Table C-1. Six Felsenthal Project Floodplain Sites Recommended for Extensive 

and Intensive Investigation................................................................................C-7 

Table C-2. Twenty-nine Felsenthal Project Floodplain Sites Recommended for 

Two-stage Mitigation ..........................................................................................C-8 

Table C-3. Summary of Site Characteristics Relevant to Research Potential and 

Scientific or Historic Significance for 117 Floodplain Sites .........................C-11 

Table C-4. Mitigation Budget Estimate (in field man days): 

PHASE 2: Mitigation of Six Prehistoric Floodplain Sites ...........................C-19 


PHASE 3: Two-Stage Mitigation, 23 Prehistoric Floodplain Sites..............C-20 


PHASE 3: Two-Stage Mitigation, Six Historic Sites .....................................C-21 

xi

xii

list of figures 

Figure 1. Location and features of Felsenthal National Wildlife Refuge in 

southern Arkansas .................................................................................................... 5 

Figure 2. Flow of information between research units from research design to proposed 

mitigation program for the Felsenthal project .......................................... 7 

Figure 3. Quaternary geology of the Ouachita River Valley in Arkansas ....................... 20 

Figure 4. Aerial view of the Saline River-Eagle Creek floodplain and Prairie Island ... 22 

Figure 5. Possible lowland prairie remnant on the Deweyville 2 Terrace near 

Redeye Lake ............................................................................................................. 23 

Figure 6. Hydrological features in the Felsenthal project area ......................................... 27 

Figure 7. Topstratum deposits exposed by the Ouachita River bankline ....................... 30 

Figure 8. Meandering channel characteristics and terminology ...................................... 32 

Figure 9. Typical riverbank profiles in the Felsenthal Project area .................................. 35 

Figure 10. Bottomland hardwood forest, predominantly mature red oaks, in the 

Ouachita River floodplain ..................................................................................... 37 

Figure 11. Hardwood and swamp forest in the Felsenthal Project area ........................... 38 

Figure 12. Cross-valley section in the Felsenthal Project area showing terraces, 

floodplain, forest types, and other features ........................................................ 48 

Figure 13. A portion of the Ouachita River Valley designated the Felsenthal 

region and locations of some important prehistoric sites ................................. 53 

xiii

Figure 14. The Mississippi Period mound and midden complex at Watts Field 

(3UN18) .................................................................................................................... 70 

Figure 15. Historic sites, settlements, and placenames in the Felsenthal Region ............ 82 

Figure 16. Hydrographic survey chart of Ouachita River in the Felsenthal Project 

area, showing locations of Marie Saline Landing (3AS299), Grand Mary 

(or Grand Marais) Landing (3UN122), and other features ............................... 84 

Figure 17. The sternwheeler Lotawanna fully loaded with cotton bales ............................ 89 

Figure 18. Portions of certificates of registry for the Lotawanna ......................................... 91 

Figure 19. Advertisement from unknown newspaper source (about 1868) 

announcing Lotawanna’s Camden-New Orleans trade ..................................... 93 

Figure 20. The Felsenthal Project area, subareas, and 13 floodplain transect locations .. 106 

Figure 21. Site survey tactics and equipment ...................................................................... 108 

Figure 22. Mounted air photo mosaics used for location finding and site plotting 

by boat survey and floodplain transect survey teams ..................................... 109 

Figure 23. Routes and extent of bankline survey in river channels, tributaries, 

open lakes, and backswamp areas accessible by boat ..................................... 113 

Figure 24. Site Unit 1, Ouachita River above mile 254, with prehistoric sites (dots) 

located in relation to channel geometry. Some prehistoric sites in Unit 6 

and historic sites in Unit 8 (rectangles) are also shown on this map. ........... 122 

Figure 25. Site 3UN162, consisting of fire-cracked rock exposed at 3 m depth in 

the Ouachita River bankline ................................................................................ 125 

xiv

Figure 26. Site Unit 2, Lower Saline River, and Site Unit 3, Lower Eagle Creek, 

with prehistoric sites (dots) located in relation to channel geometry. Two 

historic sites with prehistoric components (rectangles enclosing dots) 

in Unit 8 are also shown on this map. ............................................................... 127 

Figure 27. Mississippi period extractive sites on lower Saline River ............................... 129 

Figure 28. Site Unit 4, Ouachita River below mile 254, with prehistoric sites (dots) 

located in relation to channel geometry. Historic sites (rectangles) or 

historic sites with prehistoric components (rectangles enclosing dots) in 

Unit 8 are also shown on this map. .................................................................... 131 

Figure 29. Riverbank sites on Ouachita River ..................................................................... 133 

Figure 30. Site Unit 5, Lower Lapile Creek, with prehistoric sites (dots) located in 

relation to channel geometry. Two mound sites on the adjacent terrace, 

Locust Ridge (3UN8) and Shallow Lake (3UN9/52) in Unit 7 are also 

shown on this map. .............................................................................................. 135 

Figure 31. Site Unit 6, oxbow lakes and backswamp, with prehistoric sites located 

in relation to floodplain and terrace features ................................................... 137 

Figure 32. Historic artifacts from floodplain sites .............................................................. 140 

Figure 33. Hydrographic survey chart of Ouachita River in the Felsenthal Project 

area, showing location of “Wreck of Lotti Warner” (Lotawanna) ................... 147 

Figure 34. The spatial array of Mississippi period sites in or near the Felsenthal 

project area ............................................................................................................. 152 

Figure 35. Aerial view of the Ouachita River-Marais Saline Lake locale and the 

location of Marie Saline site ................................................................................ 160 

Figure 36. Marie Saline (3AS329), test excavations October 1979 .................................... 161 

xv

Figure 37. Marie Saline (3AS329), deep test pit profiles .................................................... 163 

Figure 38. Marie Saline (3AS329), profile records for Test Pits 1 and 6 (south walls) ... 164 

Figure 39. Cultural stratigraphy in Area C, Marie Saline (3AS329) ................................. 166 

Figure 40. False Indigo (3AS285), controlled surface collection and test excavations, 

November 1979 ..................................................................................................... 169 

Figure 41. False Indigo (3AS285), controlled surface collection in Area A with 

scatter map of sherds and other artifacts .......................................................... 170 

Figure 42. False Indigo (3AS285), profile with midden lens and other features in 

Test Pit 15 ............................................................................................................... 172 

Figure 43. False Indigo (3AS285), Area A showing 18 m 2 area of dense artifact 

concentration after surface collection and test excavation ............................. 173 

Figure 44. Buttonbush (3BR58), test excavations, November 1979 .................................. 176 

Figure 45. Buttonbush (3BR58), remains of two Baytown Plain vessels in Test Pit 2 .... 177 

Figure 46. Aerial view of the lower Saline River floodplain and the location of 

One Cypress Point (3AS286) ............................................................................... 179 

Figure 47. One Cypress Point (3AS286), controlled surface collections and test 

excavations, November 1979 ............................................................................... 180 

Figure 48. Flaked stone tools and cores from One Cypress Point (3AS286) ................... 181 

Figure 49. One Cypress Point (3AS286), graphic presentation of lithic assemblage ..... 184 

Figure 50. Jug Point Cutoff (3BR76), test excavations, November 1979 .......................... 186 

Figure 51. Jug Point 1 (3AS306), test excavations, November 1979 ................................. 188 

Figure 52. Jug Point 1 (3AS306), test pit profiles ................................................................. 189 

xvi

Figure 53. Jug Point 2 (3AS307), test excavations, November 1979 ................................. 192 

Figure 54. Mouth of Eagle Creek (3BR78), test excavations, November 1979 ................ 195 

Figure 55. Dendrogram generated by cluster analysis of artifacts and debris in 

nine prehistoric components (six floodplain sites) .......................................... 202 

Figure 56. Estimated rate of alluviation for the natural levees of Ouachita and 

lower Saline River in the Felsenthal Project area, based on selected 

riverbank sites ....................................................................................................... 210 

Figure 57. Histograms of size frequency in two samples of pebbles and small 

cobble from Ouachita Valley localities .............................................................. 223 

Figure 58. Arrow points from the Felsenthal Project area in proposed chronological 

sequence ........................................................................................................... 227 

Figure 59. Dart points from the Felsenthal Project area in proposed chronological 

sequence ................................................................................................................. 232 

Figure 60. Bifacial tools, preforms, and cores ..................................................................... 236 

Figure 61. Miscellaneous flaked stone tools ........................................................................ 239 

Figure 62. Schematic linear model for production of a bifacial tool and discard 

materials ................................................................................................................. 243 

Figure 63. Cobble tools ........................................................................................................... 246 

Figure 64. Groundstone tools and pigment ......................................................................... 247 

Figure 65. Tchefuncte complex ceramics ............................................................................. 253 

Figure 66. Baytown complex plainware ............................................................................... 256 

Figure 67. Miscellaneous sherds from floodplain sites ...................................................... 257 

Figure 68. Decorated rim sherds, Gran Marais ceramic complex, False Indigo 

(3AS285) ................................................................................................................. 264 

xvii

Figure 69. Decorated body sherds, Gran Marais ceramic complex, False Indigo 

(3AS285) ................................................................................................................. 265 

Figure 70. A seasonal model for subsistence activity and resource use in the 

Grand Marais Lowland ........................................................................................ 276 

xviii

MAnAgeMent suMMAry 

The Arkansas Archeological Survey conducted a multistage cultural resources 

survey and site testing project in the Felsenthal Navigation Pool on the Ouachita and 

lower Saline rivers for the U.S. Army Corps of Engineers, Vicksburg District, under 

terms of contract DACW 38-79-C-0002. Fieldwork was preceded by development of a 

broad-based, interdisciplinary research design and by detailed archeological, historical, 

ecological, and geohydrological background research. Existing or proposed National 

Register properties were not previously recorded in the Felsenthal Project area, which is 

an overflow bottomland never permanently settled or farmed in historic times and now 

encompassed by the Felsenthal National Wildlife Refuge. 

Multiple field phases began in mid-August 1979, shortly after winter-spring floodwaters 

receded below the upper project contour (65 feet MSL), and continued through 

late November 1979 when floodwaters again rose above this contour. During this threeand-one-half 

month field season, three floodplain survey teams operated continuously 

through a seven week field phase; test excavation then began with two testing crews 

operating continuously through much of the next seven week period. Some site survey 

activity continued on a reduced basis during the testing phase, so that site survey and 

testing phases actually overlapped. A field laboratory also operated continuously at the 

project headquarters throughout the field season. 

Site survey teams recorded 126 previously unknown archeological sites and revisited 

18 sites known from prior work in the area. These 144 sites are briefly described in 

Appendix A. Essentially all floodplain sites in the project area are buried, but many are 

exposed by riverbank erosion. Site survey tactics consisted of 13 random and judgmental 

transects traversing about 62 km, and 10 river or tributary bankline surveys totaling 

about 177 km. Various site specific surveys were also carried out, as well as collection 

of environmental data in conjunction with survey activity. Subsurface testing for buried 

sites included use of power auger, shovel testing, and close inspection of eroded banks 

or other exposed areas. All survey work was designed to contend with difficult wetland 

conditions and with extensive floodplain forest cover. 

xix

Site testing focused on eight prehistoric sites: a large, deep, and stratified site on 

the Ouachita River and seven shallow buried sites of relatively small size on the lower 

Saline River. Additional testing of selected prehistoric and historic sites was prohibited 

by rising floodwaters. Site testing tactics included random, judgmental, and systematic 

test pits which sampled about 5% of known site areas, and also numerous power auger 

or shovel holes to delineate site extent. Controlled surface collection was employed 

where riverbank erosion exposed discrete artifact scatters and arrays of debris. Each 

site was mapped by instrument and recorded by standard archeological methods. Site 

records filed at the Arkansas Archeological Survey run over 1000 pages. 

Extensive laboratory analysis, dealing with about 13,100 artifacts and items of 

debris and numerous other samples and residues, continued through 1980 at the Contract 

Laboratory, Arkansas Archeological Survey, or in various specialists’ laboratories. 

Results of these studies are reported in Chapters 7 and 8, and in Appendix B. 

All floodplain sites recorded by site survey or tested during 1979 were evaluated 

with regard to research potential in the context of regional or more broadly relevant 

problems in prehistory and history. This evaluation identified 37 floodplain sites as significant 

resources. A detailed assessment of navigation pool inundation and related impacts 

indicates that all significant floodplain sites will be adversely affected. One of the 

37 sites has been adequately mitigated by testing and another has recently been tested 

by Historic Preservation Associates. 

In compliance with the contract Scope of Services, the Arkansas Archeological 

Survey has prepared National Register nominations for 16 significant sites which have 

archeological deposits at or below the 65 foot contour interval. However, based on the 

evidence presented in this report, we have determined that a total of 35 floodplain sites 

appear eligible for nomination to the National Register of Historic Places, 19 of which 

are above 65 feet. These 35 floodplain sites, for effective management purposes, could be 

nominated as a Multiple Resource Area; a recommendation to this effect is made in Appendix 

C. 

Among mitigation alternatives considered, only mitigation by data recovery has 

been considered feasible. We have therefore recommended a multistage interdisciplinary 

program of mitigation. This program features extensive excavation and data recovery 

in six floodplain sites during a first field season, using concurrently operating field 

crews. A second full season of fieldwork would comprise two-stage mitigation of 29 

other significant floodplain sites, including both prehistoric and historic sites. These 29 

sites are divided into seven groups by cultural period; the flexible two-stage mitigation 

approach requires limited testing of all or most sites within each group, followed 

xx

y extensive excavation and data recovery at one or more selected sites within each 

group. Again, multiple field crews would operate concurrently in order to meet objectives 

of the mitigation program (Appendix C). It is recognized that the Corps maintains 

administrative responsibility only for these sites below 65 feet; nonetheless it is the 

Survey’s recommendation that the resources would be best served by a coordinated 

research approach to the entire Felsenthal project. 

xxi

xxii

AcknowledgMents 

Many individuals and organizations contributed significantly to the performance 

of the Felsenthal Archeological Project and to preparation of this report. It is a pleasure 

to acknowledge this assistance which includes administrative and scientific expertise, 

technical efforts or skill, and south Arkansas hospitality. 

First, we are most appreciative of the support and cooperation of the Vicksburg 

District, U.S. Army Corps of Engineers, under the direction of Col. Samuel P. Collins, Jr., 

District Engineer. Mr. St. Clair Thompson, Contracting Officer’s Representative and Ms. 

Shelia Lewis, Archeologist in the Environmental Branch, worked closely with us during 

the research design and field phases. Ms. Kate Yarbrough, Archeologist in the Environmental 

Branch, worked with us during later stages of the project. Among other contributions, 

the Vicksburg District provided us with excellent, recent air photo series and with 

hydrographic survey charts of the 1800s and early 1900s which proved to be invaluable. 

A number of colleagues at Arkansas Archeological Survey worked directly in various 

phases of the project and added immeasurably to its results. Dr. Charles R. McGimsey 

III, Director and Principal Investigator, assisted our work in numerous ways and at 

many times. Mr. Frank Rackerby, Contract Administrator, was responsible for planning, 

logistical, and personnel matters, and also worked closely with the principal author on 

recommendations for mitigation (Appendix C). Ms. Hester Davis, State Archeologist, 

visited field headquarters during site survey work and conferred with us and U.S. Army 

Engineers personnel on particular problems encountered and resources studied in the 

Felsenthal project area. Ms. Sylvia Medlin, Accountant, skillfully handled many crucial 

arrangements for housing and services. 

Dr. Sandra C. Parker, Systems Analyst, and Ms. Susan Houston, Research Assistant, 

performed a statistical analysis which appears in Chapter 6. Ms. Cathy Moore-Jansen, 

Registrar, and Ms. Carolyn Cox, Assistant Registrar, worked conscientiously with 

more than 1000 pages of site records. 

A special acknowledgement is due three colleagues who reviewed the draft report 

in detail and made expert recommendations for its improvement: Dr. Fred W. Limp, Assistant 

Director, Dr. Frank Schambach, station archeologist at Southern Arkansas University, 

and Dr. Marvin Jeter, station archeologist at University of Arkansas, Monticello. 

Mr. James E. Duncan skillfully supervised final processing and various artifact analyses 

at the Contract Laboratory in Fayetteville. The production staff is gratefully acknowledged 

for its skill and hard work during report preparation: Ms. Mary Lynn Kennedy, 

editor; Ms. Gwen Hamilton, typist; Ms. Louise Mullins, typist; Ms. Christina Spees, typist; 

Ms. Jane Kellett, illustrator; and Ms. Pam Ashford, photographer. 

xiii

The Felsenthal Project field crew endured very difficult wetland conditions and 

maintained high standards of work. Staff members include Jeyne Bennett, Katherine 

Dinnel, Andrew Hayes, and Clark Rogers; crew members include Paul Baumann, Diane 

Carpenter, John Paul Lane, Sheri Irvin Lane, Douglas Marsh, Jan Morrill, and William 

Radisch. Crew members employed for a portion of the project, or who rendered volunteer 

assistance during site testing, include Michael Hambacher, Douglas McKay, Henry 

McKelway, and Redus Montgomery. All are now veterans of southern floodplain forest 

and backswamp. 

A number of local residents were helpful to particular aspects of the Felsenthal 

Project or were gracious hosts, as indicated below: 

Mr. Chris Avery, Soil Scientist, Soil Conservation Service, El Dorado: project 

area soils information 

Mrs. Garland Brantley, Felsenthal: headquarters and laboratory space 

Mr. James Howe, Forester, Felsenthal National Wildlife Refuge: project area 

floral information 

Mr. Roland Johnson, Felsenthal: local historical information, including the 

tale of steamboat Lotawanna 

Mr. Bill Julian, Manager, Felsenthal National Wildlife Refuge: project area 

information 

Mr. Jack Lee, El Dorado: historical references and regional historical information 

Mr. Bud Poole, El Dorado: boat docking facilities and project area information 

Mr. John Robertson, Crossett: repair and servicing of all-terrain vehicles and 

outboard motors 

Mr. W. C. Saunders, Strong: crew housing 

Mr. Ken Short, Felsenthal: local archeological site information 

Mr. James Waterston, Felsenthal: crew housing 

Mr. and Mrs. Ed White, Hamburg: regional archeological site information 

Mr. Jack C. Womble, Assistant Manager, Felsenthal National Wildlife Refuge: 

project area information 

xxiv

In addition to these individuals, we are grateful to the group of professionals and 

scientists who provided analyses or materially assisted certain research efforts, including 

the following: 

Mr. Mark E. Gordon, Department of Zoology, University of Arkansas: mollusc 

identification 

Dr. Nancy G. McCartney, Curator of Botany, University of Arkansas Museum: 

wetland vegetation and ecology 

Mr. Charles Reeves, Federal Archives and Records Center: archival research 

Dr. E. Moye Rutledge, Department of Agronomy, University of Arkansas: 

sediment particle size analysis 

Mr. Allen R. Saltus Jr., Archeological Research and Survey: magnetometry 

data for steamboat Lotawanna 

Dr. Wayne E. Sabbe, Soil Testing and Diagnostic Laboratory, University of 

Arkansas: soil test data 

Dr. Frank Schambach, Arkansas Archeological Survey station archeologist at 

Southern Arkansas University: regional prehistory and ceramics 

Mr. David W. Stahle, Research Assistant, Arkansas Archeological Survey: 

wood and charcoal identification 

Dr. Jerry J. Stipp, Beta-Analytic, Inc.: radiocarbon assays 

Dr. Murray Tamers, Beta-Analytic, Inc.: radiocarbon assays 

Finally, the principal author expresses his thanks to the contributors to this volume, 

Mr. John H. House and Ms. Beverly Watkins, Research Assistants at Arkansas 

Archeological Survey, for their comprehensive reviews of prehistory and history in 

the Felsenthal region (Chapters 3 and 4). Their insights are a welcome addition to the 

emerging perspective of human adaptation in the Grand Marais Lowland environment, 

as presented in the following pages. 

xxv

chapter 1 

introduction to the felsenthAl project 

the felsenthAl contrAct And its reseArch potentiAl 

This report describes diverse results of a large scale interdisciplinary, archeological 

research and cultural resource management project in southern Arkansas. There is little 

question that we are telling the middle of a story, since our efforts were preceded by sporadic, 

but important, prior investigations (mostly in the last decade), and since the most fruitful 

archeological research and management efforts in this region lie ahead. One purpose of this 

introductory chapter is to alert the reader to important data and conclusions that will follow. 

Certainly the boldest conclusion is this—prehistoric and historic floodplain sites in the 

Felsenthal Navigation Pool lie within, and comprise a portion of, one of the richest and least 

disturbed archeological zones in North America. 

The contract for cultural resources survey and evaluation (DACW 38-79-C-0002) in the 

Felsenthal Navigation Pool was awarded by the U.S. Army Corps of Engineers, Vicksburg 

District, to the Arkansas Archeological Survey in October 1978. By authority of the River and 

Harbor Act of July 1960, the Corps is modifying the Ouachita River to provide a 9-foot (2.7 

m) deep navigation channel below Camden, Arkansas. Among four new 84 x 600 foot (25.6 

x 182.9 m) locks and dams to be constructed on the river, Felsenthal and Calion locks and 

dams in Arkansas are scheduled for completion in 1982 and 1983 respectively. (Our contract 

was concerned with both projects, but only Felsenthal Navigation Pool’s cultural resources 

and related studies are reported here; see Kelley 1981.) The Felsenthal Project Scope of 

Services (Appendix D) requires all of the following: (1) a research design, (2) background 

research, (3) a check for existing or proposed National Register properties, (4) intensive field 

survey, (5) testing of selected sites for extent and significance, (6) evaluation of National Register 

eligibility for all properties, (7) proposed mitigation program, (8) recommendations for 

interpretive displays and, of course, (9) a comprehensive report of these investigations. All 

of these requirements are addressed in this report in subsequent chapters and appendixes 

to the best of our ability. The table of contents is very detailed with respect to where these 

kinds of information are presented and discussed. 

Fieldwork was initiated by the Arkansas Archeological Survey in the Felsenthal Project 

area in mid-August and terminated in early December 1979. Our field season had to be 

delayed considerably until floodwaters subsided and was then curtailed by rising flood-

2 Hemmings 

waters in late November. Overlapping site survey and testing stages are described in Chapters 

5 and 6. One aspect of our work is repeatedly emphasized in this report; the project area 

lies within an overflow bottomland and, as indicated above, is subject to annual, prolonged 

inundation. This environmental dimension has affected the project at every stage—from 

research design through performance of fieldwork, analysis of site data, and proposal for 

mitigation measures. The 117 floodplain sites we investigated (among a total of 144 sites 

recorded) are the most lowlying archeological sites yet recorded in the state of Arkansas 

(see discussion of the Grand Marais Lowland in Chapter 2). Because of flood conditions, the 

floodplain here has never been developed for agriculture and no floodplain site has ever 

been plowed; also, to the best of our knowledge, these sites, because they are so unobtrusive, 

have never been collected or dug into by unauthorized individuals. Therefore our site 

survey and testing methods, and especially our logistical arrangements, were designed for 

unusual preservation and wetlands conditions. 

Moreover, the kinds of cultural resources we deal with most fully in this report are 

prehistoric, seasonal, specialized extractive camps that go under several meters of water 

every year (Chapter 9). If these floodplain sites are not unique, they are certainly members 

of a class of human settlements vastly underrepresented in current archeological studies. 

Based on site survey and testing data from the Felsenthal Project area, we believe these 

small extractive camps have high potential for regional research and broad anthropological 

problems (see also Research Design Outline in Appendix C). 

In Chapter 2 we emphasize ecological and geohydrological characteristics of this 

region and interrelationships with past human subsistence and settlement. The seasonal 

inundation (hydroperiod) in this lowland and its intricate mosaic of aquatic and riparian 

communities (which disperse or concentrate certain food resources) are keys to regional 

subsistence-settlement systems. However we identify a pressing need for conjunctive, 

environmental, archeological study in upland terrain of the region. Also the floodplain and 

riverine features enumerated in Chapter 2 are indicators of a dynamic geomorphic history 

which has barely been examined by qualified scientists. The point emphasized here is 

that great research potential exists for regional cultural-environmental study in the Grand 

Marais Lowland and the Felsenthal archeological region. 

elevation constraint 

The Felsenthal Project Scope of Services defined the navigation pool and the study 

area as beginning “at the Felsenthal Lock and Dam [and continuing] upstream to cover all 

areas below the 65-foot contour line” (Appendix D). The Arkansas Archeological Survey’s 

proposal (1978) interpreted this to mean that the Scope of Services “specifically excludes any 

land adjacent to or within the navigation pool that is at or above the 65-foot contour interval.” 

In point of fact, this constraint initially recognized by both parties to the contract, has 

caused numerous problems during all phases of the project. Although we have discussed 

these problems with the Corps of Engineers on several occasions and reached informal

Introduction 3 

agreement (especially during a field conference October 15, 1979), we would like to summarize 

them here as an introduction to many scientific and management considerations of 

floodplain elevation that appear throughout this report. 

1. Site survey transects crossed and recrossed terrain above 65 feet (19.8 m) 

because it is intricately interspersed with lower terrain and water courses; 

according to accepted archeological research standards, we recorded sites 

above 65 feet when encountered, as well as those below 65 feet (Chapter 5). 

2. Site survey emphasized riverbank exposures where frequently artifacts and 

debris were discovered from the upper portion of the bankline (above 65 

feet) to river level at 62 feet; according to accepted archeological research 

standards, we attempted to locate and record in situ cultural levels, whether 

or not above 65 feet (Chapter 5). 

3. Site survey and site testing activities delineated sites in which multiple 

occupation levels occurred above and below 65 feet elevation; in the most 

significant stratified site (3AS329) elevation necessarily began above and 

terminated below 65 feet; in other cases an occupation level was recorded 

above 65 feet, but a contemporary feature extended below 65 feet (Chapter 

6). 

4. Because of natural downvalley slope of the floodplain, a site at 70 feet elevation 

in the northern part of our project area could be ecologically and geohydrologically 

identical to a site at 65 foot in the southern part of our project 

area; however, research design and field methodology determined that the 

70-foot site would not likely be discovered unless well exposed in a riverbank 

(Chapter 2). 

5. The impacts of the Felsenthal Lock and Dam project and of the navigation 

and seasonally raised wildlife pools do not stop at the 65-foot contour, and 

we have attempted to assess a wide range of impacts in Appendix C. 

These detailed comments lead us to propose that the floodplain itself (terrain inundated 

annually) is the minimal, meaningful, ecological, geomorphic, and cultural unit of study 

in this area, and not some portion of it arbitrarily defined by a horizontal plane (the navigation 

pool). While we respect this engineering feature and consequent contract specification, 

we have also approached our floodplain work with the logic of science. 

Definition of Study Area 

Four geographical units of study are referred to frequently in this report, and their 

extent and relationship should be made clear at the outset.


1. Felsenthal Project area (or simply “project area”) includes all terrain below 

65 feet elevation in the Ouachita and Saline River floodplains above Felsenthal 

Lock and Dam. For practical purposes this terrain lies within the new 

navigation pool and surrounded by Felsenthal National Wildlife Refuge, 

and has been divided into two subareas by the U.S. Highway 82 crossing. 

Figure 20 in Chapter 5 best portrays the Felsenthal Project area. 

2. Ouachita and Saline River floodplain (or simply “floodplain”) includes all 

terrain and water courses within the project area defined above which are 

annually overflowed. This includes bottomland terrain at about 62-75 feet 

elevation; we occasionally refer to such terrain as “overflow bottom.” It consists 

primarily of Holocene floodplain as mapped by Fleetwood (1969), but 

includes also minor extensions of Deweyville 3 Terrace (Figure 3 in Chapter 

2). The floodplain so defined appears to be a useful cultural-environmental 

study unit for the past 3,000 years at least. Throughout this report we refer 

to floodplain environments and floodplain sites. 

3. Felsenthal National Wildlife Refuge (or “study area” as indicated in this report) 

is an irregular area of 26,305 ha (65,000 acres) established and operated 

by the U.S. Fish and Wildlife Service in southern Arkansas. Figure 1 indicates 

the boundaries and some of the features within the refuge, and also its 

proximity to the Louisiana border (see discussion of Grand Marais Lowland 

in Chapter 2). Because the refuge encompasses the range of environments 

and of upland and lowland cultural resources in the region, we have made 

intensive use of data pertaining to all refuge areas, although current field 

operations focused on the project area and floodplain. The refuge as a study 

area is a research concept used in this report, and is not the area of field 

study defined in the contract. 

4. Felsenthal archeological region (or simply the “region”) is defined by 

Schambach (1979; Rolingson and Schambach 1981:103-105) as follows: 

it lies within a zone of the Lower Mississippi Valley archeological area.... The 

northern [boundary] on the Ouachita River...I put at Camden, Arkansas... 

on good archeological and ecological evidence....The southern boundary... 

would be well down the Ouachita Valley in Louisiana, most probably between 

Ouachita City and Monroe....All major and minor tributaries of the 

Ouachita River between Ouachita City and Camden, except Bayou Bartholomew, 

were controlled all the way to their headwaters by Felsenthal region 

peoples. 

Figure 13 in Chapter 3 best portrays the extent of this region and the distribution 

of some important prehistoric sites. Schambach is of course follow-

Figure 1. Location and features of Felsenthal National Wildlife Refuge in southern Arkansas. 

5


ing the concept of an archeological region as a spatial division, where the 

“relationship between culture and a highly characteristic environment has 

been particularly close throughout the span of the archaeological period” 

(Willey and Phillips 1958:19). 

deVelopMent of reseArch design 

Since the research design is included in Appendix D, only its salient characteristics will 

be noted here. First, it should be acknowledged that our research design takes advantage of 

an earlier stimulating proposal for archeological research in the Felsenthal region by Raab 

(1976a), who in turn drew upon nearly a decade of varied research activity by the Arkansas 

Archeological Survey. However, where Raab assumed that a major segment of the Ouachita 

drainage basin (i.e., the Felsenthal region) would constitute the research area, we have necessarily 

adapted our research design to floodplain sites and environments (and to elevation 

constraints noted earlier). 

Second, the strategies for site survey and testing were conservatively devised for the 

formidable conditions of bottomland hardwood and swamp forest, interspersed with river 

channels, bayous, open lakes, and backswamps (Chapters 5, 6). Our research design also 

anticipated burial of sites in the floodplain topstratum and marked “invisibility” of sites 

buried deeply and those below water table or river stage. At several points in this report we 

point out that all prehistoric floodplain sites, and probably many historic sites, are buried 

through all or most of their extent in our project area. Under these conditions, we cannot 

claim to have located all cultural resources, although we elicit strong evidence for where 

buried sites occur frequently and where they do not (Chapter 5). 

Schedule of Investigations 

The Felsenthal Project was organized as a series of overlapping tasks with numerous 

pathways for information flow between research units (Figure 2). These tasks may be summarized 

as follows: 

Task 1—Background Research 

a. local and regional prehistory (Chapter 3) 

b. local and regional history (Chapter 4) 

c. regional ecology and geohydrology (Chapter 2) 

d. relevant theoretical and methodological perspectives (Appendix D) 

e. preparation of project research design (Appendix D) 

Task 2—Implementation of Site Survey 

a. transects (Chapter 5) 

b. bankline surveys (Chapter 5) 

c. initiation of field laboratory operations (Chapter 5) 

d. initial analysis of site locational data (Chapter 5) 

Task 3—Implementation of Site Testing 

a. ongoing site survey (Chapter 5)


Figure 2. Flow of information between research units from research design to proposed 

mitigation program for the Felsenthal Project (after Arkansas Archeological Survey 

1978:Figure 2).


b. initial assessment of potential significance and selection of sites for testing 

(Chapter 6) 

c. initiation of testing to establish site significance (Chapter 6) 

d. ongoing field laboratory operation (Chapter 5) 

e. initial analysis of site contextual data and research potential (Chapter 6) 

Task 4—Detailed Analysis of Project Research Data 

a. lithic and ceramic samples (Chapter 8) 

b. soil, sediment, and chronometric samples (Chapter 7) 

c. faunal and floral samples (Appendix B) 

d. other laboratory analyses (Chapter 8) 

e. integration of background and analytical data (Chapters 5-9) 

Task 5—Evaluation of Site Significance 

a. local and regional criteria (Appendix C) 

b. National Register criteria (Appendix C) 

c. preparation of recommended program of mitigation (Appendix C) 

Task 6—Preparation of Draft Report 

Schedule of Hypotheses 

The research design (Appendix D) provides a broad, flexible orientation by identifying 

four problem domains and presenting a series of eight hypotheses (and several subhypotheses) 

within these problem domains. These hypotheses are conservatively framed in 

regard to the kinds of data recovery expected and exigencies of site survey and testing in the 

overflow bottomland environment. In the research design we emphasized the near absence 

of current, systematic, environmental, or archeological studies in the project area and region. 

The problem domains and hypotheses are stated in full below, with a notation of where test 

data and discussion of hypotheses appear in the text of this report. (No data were obtained 

for Hypothesis 5 below.) 

Problem Domain 1: Settlement Characteristics (Chapter 6) 

H 1 If subsistence strategies employed wild plant and animal resources, impermanent 

specialized extractive sites would implement these strategies in true floodplain 

environments. 

Test Implications: Recognition of specialized extractive sites as settlement types, distinct 

from base settlements. 

Data Requirements: Floodplain site assemblages with high ratio of extractive/maintenance 

tool kits; faunal or floral indicators of seasonality and procurement strategy; 

absence of permanent structures for habitation, storage, or burial. 

Problem Domain 2: Locational Characteristics (Chapter 5) 

H 2 If specialized extractive sites were established for resource procurement, these 

sites would be located for most efficient acquisition and delivery to base settlements.


Subhypotheses: A. Such sites will correlate with specific resource zones. 

B. Such sites will correlate with access routes to base settlements. 

C. Such sites will correlate with both factors. 

D. None of the above, but other unknown factors. 

Test Implications: Significant correlations obtained for location factors; recognition of 

contemporary base settlements, access routes, and resource zones. 

Data Requirements: Map distributions; measures of spatial relationships; crossdating. 

Problem Domain 3: Culture History (Chapters 5, 8) 

H 3 If upland mound complexes represent a dominant regional culture, numerous 

sites of this culture should also be identified in the lowland setting. 

Test Implications: Discovery of culturally related sites in study area. 

Data Requirements: Common artifact styles, especially ceramics. 

H 4 If aboriginal Felsenthal populations occupied a frontier region, local styles of artifacts 

and settlement remains will occur mixed with extraregional styles. 

Test Implications: Recognition of local and alien styles. 

Data Requirements: Typological and comparative analyses disclosing introduced 

styles, especially in ceramics. 

H 5 If early historical sites are present in the floodplain, they were occupied briefly for 

hunting and trading activity (before 1780). 

Test Implications: Discovery of sites with simple technological assemblages. 

Data Requirements: Tools or discards primarily associated with hunting; trade items; 

datable objects; absence of permanent structures. 

H 6 If later historical sites are present in the floodplain, they were occupied in association 

with riverine activity and extractive industry (after 1780). 

Test Implications: Discovery of sites with complex technological assemblages. 

Data Requirements: Tools or components associated with engines.


Problem Domain 4: Site Transformations (Chapter 7) 

H 7 If geologically Recent topstratum deposits have aggraded 20 feet or more in the 

Ouachita Valley, methods of intensive survey in the floodplain will fail to sample 

adequately early occupation sites. 

Test Implications: Recognition of the age of sediments drowned and exposed; discovery 

of buried sites in datable context. 

Data Requirements: Significant disparity in observed and expected occurrence of 

early sites; geochronological controls. 

H 8 If differential preservation of organic matter occurs within or between sites, soil 

acidity has adversely affected the archeological context. 

Test Implications: Recognition of preservation gradients by sampling soil and refuse; 

demonstrable loss of organic matter. 

Data Requirements: Translation of preservation gradient to a numerical scale; significant 

positive correlation with pH (low soil pH—low degree of preservation). 

felsenthAl project personnel 

The following list describes key individuals involved in the Felsenthal Project from 

planning in 1978 to draft report preparation in 1981 (see also Acknowledgments for many 

other contributors). This list is presented here as an additional perspective on Felsenthal 

Project organization and expertise. 

Charles R. McGimsey III (Ph.D., Harvard, 1958), Director of Arkansas Archeological 

Survey, was Principal Investigator for the Felsenthal Project. He was similarly involved in all 

earlier Survey research activities in the Felsenthal region. He bears overall responsibility for 

all administrative and scientific aspects of the project. Dr. McGimsey is the author of several 

books and numerous articles in American archeology and cultural resource management. 

W. Fredrick Limp (Ph.D., Indiana, 1980), Assistant Director of Arkansas Archeological 

Survey, was involved with the project as research coordinator and principal reviewer of the 

draft report. In these capacities he contributed to many aspects of the research, especially 

to the proposed mitigation program in Appendix C. His research interests and publications 

include development of data recovery techniques such as flotation, research design, and 

quantitative analyses in archeology.


Frank Rackerby (ABD, Northwestern, 1968) is Contract Administrator at the Arkansas 

Archeological Survey, and served as the critical link between the Survey and the project 

sponsor. He was deeply involved in preparation of the proposal (Appendix D), complex 

scheduling and logistical arrangements for survey and testing, and personnel and fiscal 

matters. He has published various monographs and articles in Midwestern archeology and 

cultural resource management. 

E. Thomas Hemmings (Ph.D., Arizona, 1970), Associate Archeologist at the Arkansas 

Archeological Survey, was Felsenthal Project Archeologist, developed the research design 

(Appendix D), supervised all field and field laboratory operations, supervised or conducted 

analyses of project data in the Contract Laboratory, and prepared this draft report. He also 

provided the interdisciplinary dimension of the project at this stage by virtue of training 

and experience in both environmental archeology and geology/geochronology. Hemmings 

directly supervised testing of a deep stratified site on the Ouachita River, Marie Saline 

(3AS329), described in Chapter 6. Dr. Hemmings is the author of numerous monographs 

and articles in Southeastern and Ohio Valley archeology, and is actively involved in interdisciplinary 

research in Arkansas. 

Beverly Watkins (ABD, Auburn, 1980), Project Historian, performed historical background 

research which was essential to project planning and historic site survey work, and 

also prepared the regional historical summary in Chapter 4. Watkins was notably successful 

in obtaining archival data and oral history pertaining to the steamboat Lotawanna (3UN153), 

which burned and sank in 1874 on the Ouachita River within the project area. She is actively 

engaged in historical research on the nineteenth century period in Arkansas. 

John H. House (B.A., Arkansas, 1972), Research Assistant at Arkansas Archeological 

Survey, has extensive research experience in the Southeast United States including Arkansas 

and Louisiana. He recently prepared a syntheses of certain prehistoric periods in eastcentral 

and southeast Arkansas for the Arkansas State Plan (Davis 1981), and on this basis 

was invited to contribute a review of prehistory in the Felsenthal region, which appears as 

Chapter 3 of this report. 

Marvin Jeter (Ph.D., Arizona State, 1977), Arkansas Archeological Survey station 

archeologist at the University of Arkansas at Monticello, is an expert regional archeologist, 

and served as a project consultant. He visited ongoing field operations on several occasions 

and participated in testing of a significant Mississippi period extractive camp, False Indigo 

(3AS285), described in Chapter 6. Dr. Jeter has made many suggestions for improvement of 

this report. He has published articles and monographs in Southeastern and Southwestern 

archeology, and is currently active in prehistoric studies in southeast Arkansas. 

Clark Rogers (B.A., SUNY Geneseo, 1971) served as senior crew chief and was particularly 

concerned with all-terrain-vehicle and pedestrian survey transects, and with testing on 

sites 3AS285, 3AS306, and 3BR78 (Chapter 6).


Katherine Dinnel (M.A., Florida State, 1977) served as site survey crew chief and was 

particularly concerned with boat surveys on both the Ouachita and Saline rivers, and with 

testing on sites 3AS286 and 3AS307 (Chapter 6). Dinnel also held an appointed position as 

laboratory and research analyst following the fieldwork stage, and contributed significantly 

to this report in all these respects (Chapter 8). 

Andrew Hayes (B.A., Appalachian State, 1978) served as site survey crew chief and 

was particularly involved in all-terrain-vehicle and pedestrian survey transects, and with 

testing on sites 3BR58 and 3BR76 (Chapter 6). 

Jeyne Bennett (M.A., University of Arkansas, 1980) operated the field laboratory at 

Felsenthal and was concerned with processing and initial analysis of collections. She was 

additionally involved with personnel and logistical record keeping, and frequently served 

as communication link between field personnel and other individuals or agencies concerned 

with the Felsenthal Project.

chapter 2 

grAnd MArAis lowlAnd enVironMent 

In the West Gulf Coastal Plain of southern Arkansas, one of the most distinctive geographical 

features is the low trough of the Ouachita River Valley. At the Arkansas-Louisiana 

state line (about 33° N latitude) the floodplain is 40 feet (12.2 m) lower than the alluvial 

ridge of the Mississippi River at the same latitude. This floodplain corridor is the lowest 

terrain in the state of Arkansas and, not surprisingly, shares many environmental characteristics 

with the Lower Mississippi Alluvial Valley and with large tributary rivers and floodbasins 

in Louisiana. 

The southern character or affinity of the Felsenthal Project area is exemplified by a 

small brake of bald cypress draped with Spanish moss just northwest of the U.S. Highway 

82 bridge across the Ouachita River (locally called Mouton Brake). This occurrence is essentially 

the northern limit of Spanish moss in Arkansas (Tucker 1974). The journals and maps 

of Dunbar and Hunter who explored the Ouachita River in 1804-1805 make several references 

to the disappearance of “long moss/Tillandsia/above Latitude 33°” (Rowland 1930:309). 

Explorers, travelers, and especially geologists have commented on the low-lying floodplain 

and overflow conditions which prevail in the region of Felsenthal. We introduce some 

of those comments here as environmental and historical background to later sections of this 

report. 

The face of the Country begins to change [above present-day Alabama 

Landing, Louisiana]; the banks are low and steep, and the river generally deeper, 

and much contracted, being from 30 to 50 yards wide; this low Country is... 

liable to overflow 12 or 15 feet above the level of the land....It may be supposed 

that there existed a great Lake within the space now occupied by this alluvial 

tract....18 or 20 feet of soil perpendicular is yet wanting to render it a fit habitation 

for man; it appears never the less to be well peopled by the beasts of the forest....We 

now begin to see quantities of water fowl (Wm. Dunbar, November 14, 

1804, in Rowland 1930:240). 

A slight movement along the...fault line near Alabama Landing La., [caused] 

extreme swamping of the bottoms from that point to above the mouth of Bayou 

Moro in Arkansas....[Faulting] has amounted to 25 feet, and the movement is so


recent that, although occurring in the floodplain of a large river, it has not been 

perceptibly obliterated by sedimentation....The persistence of this break must be 

regarded as an extreme illustration of the small amount of sediment carried by 

Ouachita River (Veatch 1906:59, 65). 

Much of the territory bordering the Ouachita is low-lying and subject to inundation, 

and probably in aboriginal times, as at present, was not occupied for permanent 

habitation....From just above Ouachita City for about seventy-five miles up 

by water, is an almost uninhabited region (Moore 1909:16). 

Between the Saline River and Alabama Landing, the floodplain widens to about 

4 miles where it transects the Grand Marais lowland....The topographically 

depressed Grand Marais area is not structurally controlled, but rather is due to 

ponding of the lower portion of the river brought about during the formation of 

the relatively higher Arkansas River meander belt near Ouachita City (Fleetwood 

1969:Figure 1, facing page). 

Alluvial drowning...is manifested by the large, frequently inundated, swampy 

lowland or flood basin called Grand Marais...between Felsenthal and Alabama 

Landing, Louisiana. Within the flood basin the Ouachita River natural levees are 

poorly developed, the floodplain relief is minimal, and the river has abandoned 

meander belts more frequently than anywhere else....Faulting was postulated as 

the cause of the Grand Marais; however neither physiographic nor subsurface 

evidence has ever been discovered to substantiate this (Saucier and Fleetwood 

1970:878). 

Fleetwood and Saucier have thus introduced the concept of the Grand Marais Lowland as a 

geomorphic division of the Ouachita Basin and Lower Mississippi Alluvial Valley. The use 

of the historical placename “Grand Marais” is highly appropriate for this lowland, since in 

Mississippi Valley French a “marais” is more likely to refer to an open lake than a marsh or 

swamp (McDermott 1941:98). A profusion of oxbow lakes and abandoned river courses, reoccupied 

by bayous, is a salient characteristic of the floodplain in the Felsenthal Project area. 

Our concept of the Grand Marais is both geohydrological and ecological. The distinctive 

lowland flora and fauna have interacted with and profoundly influenced resident human 

populations over the past few thousand years. In advance of the summary of regional 

prehistory in Chapter 3 and site data presented later in this report, we note that a dense 

zone of prehistoric occupation sites coincides with the elevated outer edge (terrace) adjacent 

to the floodplain, and that an array of small impermanent camps coincides with channel 

margins within the floodplain. Many sites in both zones are firmly attributed to the Mississippi 

period (A.D. 1100-1700), but several millenia of earlier human occupation sites are also 

well represented. It is clear that the Grand Marais Lowland is the axis and core environment

Environment 15 

of the Felsenthal region, and that procurement of lowland resources was a major regional 

strategy of prehistoric human populations. 

southeAstern floodplAin ecosysteMs 

A limitation of our study is the lack of specific, detailed, ecological analyses in the 

Ouachita River Basin and in the Felsenthal Project area. Most available data are in the form 

of checklists and inventories (e.g., Arkansas Department of Planning 1974; U.S. Department 

of Agriculture 1978). Certain general statements can be made from the results of ecological 

studies in other regions of the southeastern United States. 

In broad terms the Felsenthal region can be subdivided as upland and lowland ecological 

zones; the latter zone, our focus of interest, is a distinctive juxtaposition of riparian and 

aquatic ecosystems. 

Riparian Ecosystems 

These wetlands are defined in part as follows: 

Riparian ecosystems usually occur as an ecotone between aquatic and 

upland ecosystems, but have distinct vegetation and soil characteristics....These 

ecosystems are most commonly recognized as bottomland hardwood and floodplain 

forests in the eastern and central U.S....Riparian ecosystems are uniquely 

characterized by the combination of high species diversity, high species densities, 

and high productivity. Continuous interactions occur betwen riparian, aquatic, 

and upland terrestrial ecosystems through exchanges of energy, nutrients, and 

species (Brown et al. 1979:18). 

Two salient characteristics that distinguish riparian ecosystems among other wetlands 

are the hydroperiod pulse (prolonged annually in the Grand Marais) and the high degree 

of connectedness with other ecosystems (Ewel 1979). The composition and structure of the 

bottomland hardwood forest is highly sensitive to the timing and duration of flooding; relatively 

fewer tree species, e.g., bald cypress, tolerate long hydroperiods and soil saturation 

extending into the growing season (Bedinger 1971; Huffman 1976). The standing biomass of 

bottomland hardwood and swamp forest is enormous, and contributes to high productivity 

and high carrying capacity for a variety of consumer organisms (Conner and Day 1976; Robertson 

and Weaver 1979). Among Southeastern ecosystems, floodplain forests rank among 

the highest in primary productivity with litterfall (leaf and twig debris) values exceeding 

those reported for upland forests (Wharton and Brinson 1979). This detritus is cycled laterally 

or downstream during flood stages and grazed by bacteria and invertebrates, which 

are in turn consumed by spawning or foraging fish and crayfish, seasonal inhabitants of the 

submerged bottomland forest.


The riparian ecosystem is connected or linked to its upland watershed, to wetlands 

upstream and downstream, and to distant ecosystems by a variety of processes. Important 

among these are long distance migrations of fish and eels, seasonally resident birds (swimming, 

wading, and perching birds of great diversity), and more restricted seasonal or periodic 

movements of terrestrial vertebrates (e.g., white-tailed deer, turkey). In the Grand 

Marais Lowland numerous terrestrial and semi-aquatic vertebrates inhabit or invade the 

bottomland forest and floodplain water courses on a seasonal schedule, withdrawing to upland 

margins during the winter-spring hydroperiod. Among these seasonal foragers is man 

himself. 

Smith (1978) skillfully relates the ecology of major river valleys to Mississippi period 

settlement and subsistence activity in terms broadly applicable to the eastern United States. 

In his view meander belt zones with closely spaced levees, oxbow lakes, and backswamps 

comprised the adaptive niche of Mississippian populations (Smith 1978:484). These populations 

subsisted by cultivating the aboriginal trinity of corn, beans, and squash (as well as 

secondary crops) on fertile, arable, but localized levee soils, and by exploiting five wild species 

groups, primarily (1) backwater fish species, (2) migratory waterfowl, (3) the “terrestrial 

trinity” of deer, raccoon, and turkey, (4) nuts, fruits, and berries; and (5) pioneer seed-bearing 

plants. 

Spring flooding provided an essential external energy subsidy to the meander belt 

zone, introducing and dispersing nutrients. Diverse meander belt features were bounded by 

maximum edge areas and sustained a high biomass of plants and animals. Regional “energy 

capture” variables of significance to Mississippian populations were the extent of arable 

land emergent for spring planting, and the extent of permanent or seasonally flooded lakes. 

Smith suggests that a dispersed pattern of small settlements was the optimum strategy for 

exploiting meander belt resources, but that aggregated patterns, adapted to maintenance 

and defense of the resource catchment was the most common Mississippian strategy. 

Lewis (1974) employed a similar ecological approach in a detailed study of Mississippian 

settlements in southeast Missouri. He emphasizes the constraint of flooding on permanent 

settlement in the Lower Mississippi Alluvial Valley, which may be resolved by one or 

more of the following strategies: 

1. locating settlements on meander belt sites least affected by inundation, 

2. incorporating “flood-proof” or elevated structures within flood-prone communities, 

and 

3. reducing flood risk by constructing levee and drainage systems. 

Lewis identifies the first strategy as that adopted by Mississippian groups in his study area.

Environment 17 

Although many of the ecological concepts and settlement data advanced by Smith and 

Lewis are relevant to our analysis of the Felsenthal Project area, we see here a significant regional 

variation of the Mississippian meander belt strategy. In our study area floodplain resources 

were heavily exploited by Mississippi period populations aggregated in settlements 

on elevated terraces adjacent to the floodplain (Chapters 5, 6). Floodplain farming was 

prohibited by infertile, poorly drained soils and by the prolonged annual period of overflow, 

but was presumably practiced on terrace lands adjacent to permanent settlements. These 

“flood-proof” settlements were “poised” at the outer edge of the floodplain to take advantage 

of slope, drainage, and access to lowland resources by means of creeks and bayous. We 

doubt whether this disposition of Mississippian settlements is unique to the Grand Marais 

and the Felsenthal region, but here the emerging pattern is distinctive (Chapter 5:Figure 34). 

With regard to settlement and flood risk, we may also make some observations for 

European and American inhabitants of the region during the historic period. Clearly, overflow 

conditions were a well known, practical matter to early hunters or traders and to later 

settlers, as indicated by the Dunbar-Hunter journals and other regional historical sources 

(e.g., U.S. Congress 1874). In the aboriginal fashion, permanent settlement and farming were 

relegated to upland terraces, and riverbank encampments for hunting, fishing, logging of 

hardwood, cutting staves, etc., were made during periods of normal flow. River landings 

and woodyards during the nineteenth century seem to have employed moored floating 

rafts, usable year-round, serving especially the high water steamboat traffic. In the twentieth 

century substantial fishing and hunting camps continue to employ floating rafts, pilesupported 

platforms, and, in a few cases, raised earth mounds. Low earth mounds were also 

constructed as refuges for pigs turned out to forage in the floodplain forest. In the Congaree 

River swamp of South Carolina, an environment markedly similar to the Grand Marais, 

Michie (1980:102-111) describes comparable, but larger, earth platforms called “cattle 

mounts.” In these lowland environments historic inhabitants have necessarily adapted and 

limited their activities in response to seasonal inundation. 

Aquatic Ecosystems 

Southeastern river and floodplain systems include a variety of aquatic habitats, differing 

for example in hydrodynamics, waterborne inorganic and organic matter, water hardness, 

acidity/alkalinity, penetration of sunlight, and, of course, abundance and diversity of 

organisms (Gosselink and Turner 1978; Wharton and Brinson 1979). These diverse habitats 

may be characterized by an energy gradient, with high energy (lentic) flowing water systems 

at one extreme and low energy (lotic) standing water at the other. Heterogeneity or 

niche differentiation and species richness increase with hydrologic energy. In the context 

of a meandering river system in an alluvial valley, heterogeneity is created by hydrological 

variation between reaches of a river (e.g., alternating deep pools and shallow riffles) and 

between the river, its tributary streams, oxbow lakes, and backswamps. These habitats may 

be interconnected as a broad pool during flood stage (an energy pulse) and isolated during


low flow. Among other aspects of these aquatic ecosystems, we are particularly concerned 

with river fisheries ecology (Welcomme 1979). How have human populations in the Grand 

Marais Lowland adapted to the spatial heterogeneity, marked seasonality, and diversity of 

fisheries resources? 

Both Smith (1978) and Limp and Reidhead (1979) emphasize the importance of fish as 

an aboriginal protein source, the seasonal concentration of fish biomass, and the efficiency 

of mass-capture techniques in the eastern United States. Limp and Reidhead experimentally 

“harvested” an oxbow lake remnant or slough during summer low water stage; gizzard 

shad, high in oil or fat and food energy value, was the predominant species (21.8 kg of the 

45.5 kg total catch in one experiment). They conclude that floodplain backwaters had high 

potential for low input/high output aboriginal fishing. In advance of our site survey and 

testing results, we note that significant evidence for intensive fishing was obtained from two 

Mississippi period sites in the Felsenthal Project area (see 3AS285 and 3BR76 in Chapter 6). 

However, these sites and a high proportion of all other floodplain sites are located on river 

channels rather than oxbow lakes or backswamp drainages. The local fisheries potential and 

relevant archeological data are discussed in more detail later in this chapter. 

A variety of potentially useful aquatic and semiaquatic animals, other than fish, inhabit 

the diverse niches of Southeastern aquatic ecosystems. Among these are furbearing mammals, 

alligators, turtles, frogs, crawfish, and molluscs. For most of these species we have 

little or no regional archeological or historical data, molluscs being the principal exception 

(see below and Appendix B). An analysis of commonly occurring freshwater mussels by 

Parmalee and Klippel (1974) indicates that these animals represent low food energy, supplementary 

food sources, even though they may be obtrusive constituents of an archeological 

deposit. None of the animals listed above are included in Smith’s (1978) five species groups 

exploited as primary food resources by Mississippian populations. But considering the 

ethno-historic use of some of these (e.g., crawfish and frogs) in the Atchafalaya Basin and 

other lowland regions of the southern Mississippi Alluvial Valley, one might predict a sixth 

group of important aquatic or semiaquatic species (Comeaux 1972). 

We noted earlier the limitation of specific ecological data for the Felsenthal Project 

area. Fortunately, it is not our task to delineate in all particulars an obviously complex 

Grand Marais Lowland ecosystem or to appraise and reconstruct all animal and plant populations 

in this region. Flannery (1968) and others note that human groups are not adapted to 

environments, but to particular animal and plant species in a systemic relationship. Subsistence 

strategies implemented by past human groups involved scheduling of seasonal procurement 

tasks and also technological choices for maximum efficiency. The following review 

of geohydrological and ecological data for the Felsenthal Project area leads to considerations 

of potential lowland resources and procurement strategies that may have been used by aboriginal 

and early historic populations.

pleistocene terrAce seQuence 


Above the lowlying floodplain, the valley walls and uplands consist of a steplike series 

of Pleistocene terraces (Figure 3). These terraces have been mapped and interpreted by Fleetwood 

(1969), Saucier and Fleetwood (1970), and Saucier (1974). Correlation with continental 

Pleistocene events and with Mississippi Alluvial Valley landforms and deposits is based on 

stratigraphic and other relative age data rather than chronometric data. (So far as we know, 

no radiocarbon dates for geological contexts have been obtained in the Ouachita Valley of 

Arkansas; archeological dates from this region are reported in Chapters 3 and 7.) Westward 

of the river about 5 km, Eocene Claiborne Group deposits are exposed at the surface, while 

eastward of the river these older sediments are almost completely buried by a huge expanse 

of the Prairie Terrace. Proceeding upstream on the Ouachita and Saline rivers, well beyond 

our project area, bluffs and hills of Claiborne sediments occasionally rise immediately above 

the modern channels (Haley 1976). 

The Prairie Terrace, a relict alluvial plain of the ancestral Arkansas and Ouachita rivers, 

has an average elevation of about 33.5 m (110 feet) above the natural low water stage 

of the Ouachita River (Saucier and Fleetwood 1970:873). Aggradation of the valley to this 

high level occurred during Sangamon Interglacial time or about 100,000-80,000 years B.P. 

(Saucier 1974:Figure 3). One of the characteristic features of this terrace surface is the occurrence 

of numerous low circular “prairie mounds” of presumed eolian/biologic origin several 

thousand years B.P. (Cain 1974; Saucier 1978). These mounds are well represented east and 

southeast of the Felsenthal Project area, not only on the Prairie Terrace, but also on younger 

surfaces. As the name implies, prairie grasslands interrupted the presettlement pine-oak forests 

of the Prairie Terrace uplands. These prairies were limited in extent, occurring south of 

the Arkansas River and north and east of the Felsenthal Project area. Remnants exist today 

in Ashley and Bradley counties. In Chapter 8 we describe a sample of chert gravel collected 

from Prairie Terrace alluvium near Sulphur Springs, Arkansas, for the purpose of studying 

local toolmaking raw materials. 

The Deweyville 1 Terrace, restricted in its extent, occurs east and north of the Felsenthal 

Project area at an average elevation of 16 m (52 feet) above natural low water stage 

(Saucier and Fleetwood 1970: 873). This terrace formed partly as an aggrading floodplain 

and partly as backbarrier flats of a large “finger lake” impounded by Arkansas River alluvium 

(Macon Ridge) at Monroe, Louisiana. Lake Monroe, mid-Wisconsin in age or about 

30,000-28,000 years B.P., covered an area of 500-700 square miles, including much of the 

Ouachita and lower Saline River valleys and tributary drainages within the Felsenthal Project 

area. The detailed evidence for this extinct lake is beyond the scope of this summary. The 

Deweyville 1 Terrace surface today supports poorly drained pine flatwoods in some areas. 

The Deweyville 2 Terrace flanks the Felsenthal Project area as an almost continuous, 

irregular band of terrain characterized by little surface relief. This terrace surface is inter-


Figure 3. Quaternary geology of the Ouachita River Valley in Arkansas (after Saucier and Fleetwood 1970:Figure 1).


preted as a lacustrine plain that developed during draining of Lake Monroe about 28,000- 

25,000 years B.P. (Saucier and Fleetwood 1970:883). The average elevation of the Deweyville 

2 Terrace surface above natural low water stage is about 12.2 m (40 feet). This terrace has 

special significance for understanding the record of past human use of the Grand Marais 

Lowland in several respects: 

1. On the west side of the Felsenthal Project area and to a lesser extent elsewhere 

in the valley, the inner edge of the Deweyville 2 Terrace is a prominent 

scarp rising 6-7 m above overflow bottomlands; the crest of this scarp 

(i.e., the inner edge of the terrace) was the preferred location for substantial 

permanent prehistoric settlements, and probably also was farmed in the 

vicinity of these settlements (see Rolingson 1972; Rolingson and Schambach 

1981; and Appendix A, Unit 7, for data pertaining to some major sites). 

2. Degradation of the terrace resulted in isolated remnants or islands above 

overflow bottomlands in some areas; examples in the Felsenthal Project area 

are Prairie Island (Figure 4) and Goulett Island on the Saline River. These 

outliers, commonly called “pine islands,” are also known to have been 

favored locations for prehistoric settlement and for later historic activity 

(Rolingson 1972; Appendix A, Unit 7). 

3. The terrace scarp or inner edge in the Felsenthal Project area marks the 

lower limit of loblolly pine forest and upper limit of bottomland hardwood 

forest; this edge reflects the control on seedling reproduction and other 

stress factors exerted by overflow conditions. Pine islands are bounded by 

the same edge between forest types. The elevation of this ecological and 

hydrological edge is about 75 feet MSL in our study area. 

4. The Deweyville 2 Terrace surface is ecologically diverse in other respects 

as well. As in the case of higher surfaces, presettlement pine-oak forest was 

interrupted by prairie openings of restricted extent. Prairie Island is an existing 

example. Another possible prairie remnant near Redeye Lake at 74 feet 

MSL is subject to frequent overflow (Figure 5). This oblong area was clear of 

trees by 1934 and remains unchanged to the present, based on topographic 

maps and air photos. Soil scientists mapped the clearing as Crevasse loamy 

fine sand, and referred to it as an “idle field,” but we believe it may be a 

natural prairie remnant (Gill et al. 1979:38). 

The Deweyville 3 Terrace is a relict floodplain formed at approximately 20,000 years 

B.P., and is distinguished by channel scars of an ancestral Ouachita River with much greater 

discharge than the present river (Saucier and Fleetwood 1970:876). Using relict channel dimensions 

and geometry, Saucier and Fleetwood calculated paleodischarge five or more

22 

Figure 4. Aerial view of the Saline River-Eagle Creek floodplain and Prairie Island (air photo December 1975, courtesy 

of U.S. Army Corps of Engineers) (AAS neg. 814148).


Figure 5. Possible lowland prairie remnant on the Deweyville 2 Terrace near Redeye 

Lake (AAS neg. 797152).


times greater than the present river, and attribute this flow to Late Wisconsin pluvial conditions. 

The average elevation of the terrace above low water is 10 m (33 feet), but throughout 

its extent the Deweyville 3 surface is barely elevated above the Recent floodplain or is 

buried by uppermost Recent sediments. For the purposes of our study the Deweyville 3 

Terrace is ecologically and hydrologically continuous with Recent floodplain and overflow 

bottomland, although Fleetwood (1969) has mapped within this unit some higher terrain, for 

example the vicinity of Eagle Lake Mounds (3BR4). Certainly, the terrace surface was diverse 

through its range in our study area, for lowlying prairies exist or formerly existed at several 

locations (e.g., Pine Prairie and Coffee Prairie). 

recent floodplAin geohydrology 

Post-Pleistocene sea level rise was accompanied by a complex sequence of events in 

the Lower Mississippi Alluvial Valley. A reduced sediment load, net aggradation of floodplain, 

and shifts of meander belt characterized the Mississippi River and its tributaries in 

various degrees. The sequence of events in lower courses of the Arkansas River had profound 

effects upstream on its major tributary, the Ouachita River. After about 10,000 years 

B.P., the Arkansas River formed and abandoned a series of meander belts now occupied 

by Bayou Macon, Boeuf River, and Bayou Bartholomew (Saucier and Fleetwood 1970:884; 

Saucier 1974:Figure 1). Above Monroe, Louisiana aggradation of the Bartholomew course 

from about 3000-1000 years B.P. brought about “extreme swamping” or alluvial drowning of 

the Grand Marais Lowland upstream. The final Arkansas River abandonment occurred here 

at about 1500 to 1000 years ago. These events provide a temporal and geological framework 

for assessing human use of the Recent floodplain of Ouachita River. Later in this report we 

introduce a record of buried archeological components and alluviation in the Felsenthal 

Project area which is reasonably complete for the past 3000 years. Earlier components in the 

Recent floodplain are deeply buried and “drowned” by the modern river (Chapter 7). 

Flood Characteristics 

Patterson (1971:183) provides a 56-year record of annual peak stages for the Ouachita 

River gauging station at Lock and Dam No. 6. These peak stages consistently occur from 

February to May, reflecting storms moving northeastward from the western Gulf Coast to 

the Ouachita Mountain region. For the period 1912-1968 bankful stage was exceeded in all 

but two years; in all other years peak floodwaters inundated the floodplain to at least 65 feet 

MSL locally, the upper project contour for our site survey work (see Chapters 1 and 5). Very 

high peak stages were recorded in 1927, 1932, 1945, and 1958, when floodwaters reached 

the 85-foot contour locally, barely overtopping the Deweyville 2 Terrace. These peak stages 

inundated the lower portions of major Mississippi period mound groups in the study area, 

and inundated a few mound groups throughout their extent (e.g., Big Mound Ridge, 3AS6). 

It is probable that these and other large permanent prehistoric settlements were subjected to 

comparable, but not greater, flood risk in the past. An assumption of similar flooding conditions 

in the recent past is based on Saucier and Fleetwood’s study and on our own observa-


tions with regard to floodplain topstratum deposits (Chapter 7). This assumption is made 

with some confidence for the past few centuries and may also be valid for the past 1500-3000 

years, but requires further work. 

Duration of flooding, not reported by Patterson, is of equal or greater importance for 

interpreting seasonal occupation or use of the floodplain. According to Bedinger (1971), 

the Ouachita River in our study area floods 35-40% of the year (during winter and spring) 

and greatly resembles the lower White River overflow bottomland in distribution of flood 

tolerant hardwoods. In 1979 we kept an intermittent record of river levels in order to plan 

and carry out our floodplain survey and testing work; we show bankful stage (about 65 feet 

MSL) exceeded for the first 5 1/2 months and last 1 1/2 months of 1979 (about 58% duration 

on a calendar year basis). Floodwaters peaked at about 83 feet MSL in early May. The 

Ouachita River is, however, regulated by dams in its mountainous headwaters which must 

affect peak stage and duration to some extent. Nevertheless, all available modern data lead 

one to predict that the floodplain in the Grand Marais Lowland will be overflowed in winter-spring 

and emergent in summer-fall portions of the year. 

We have not located many historical (pre-1912) records or specific accounts of flooding 

for the Ouachita or Saline rivers in southern Arkansas. Dunbar’s journal makes an oblique 

reference to unusual flooding here in 1799 (Rowland 1930:241). One authoritative report of 

Ouachita River conditions below Camden stated that “a large portion of the adjacent country 

is annually overflowed....Both banks are submerged at even moderate stages and vast 

tracts of bottomland, covered with forest and canebrake, are underwater five or six months 

of the year” (U.S. Congress 1874:12). 

Floodplain Features 

A variety of features associated with present or past regimes of the Ouachita River and 

its tributaries can be distinguished in the Grand Marais Lowland and in the Felsenthal Project 

area. Modern meander belt features occupy a zone 1-2 km wide defined by the sinuous 

channels of Ouachita and Saline rivers. The alluvial ridges or natural levees flanking these 

channels are more or less continuous downvalley, and subdued in relief, though in fact they 

are the highest terrain in the Recent floodplain. Typically, the greatest elevation occurs at the 

riverbank and the levee surface drops rapidly 1 or 2 m to lowlying backswamp. Just south of 

U.S. Highway 82 at an historic river landing location (Marie Saline Landing, 3AS299) the following 

elevations across the Ouachita River levee are noted (U.S. Army Corps of Engineers 

1979:Plate 1-1): 

Distance Eastward 

Elevation (MSL) from Left Bank 

70.7 feet (21.6 m) 10 feet (3.0 m) 

70.2 feet (21.4 m) 280 feet (85.4 m) 

67.8 feet (20.7 m) 730 feet (222.6 m) 

66.6 feet (20.3 m) 1170 feet (356.7 m)


This levee profile is typical of many site locations recorded during our survey work (Chapter 

5). Prehistoric sites frequently occur as small buried lenses in the levee, truncated and 

exposed at the riverbank. Historic sites often occur at or near the surface and adjacent to 

the bank. Both kinds of sites took advantage of the elevation and drainage of levee crests. 

In Chapter 7 we discuss the rate of alluviation in the Recent floodplain of the Ouachita and 

Saline rivers based on the occurrence of buried prehistoric sites in natural levees. 

On the inside of many concave bends a series of point bar ridges, separated by swales, 

has resulted from lateral accretion. These ridges may be best developed on the downstream 

arm or slipoff slope of a bend, but are uniformly low features of the modern meander belt 

on both rivers. In general our survey work does not extend to point bar ridges (which lie 

mostly above the 65-foot upper project contour), and we have recorded few sites in such a 

location. At least two prehistoric sites on the Saline River occupy a low ridge adjacent to the 

chute or high water passage across the neck of a bend (Jug Point Cutoff, 3BR76, and Persimmon 

3, 3AS315). 

Between narrow natural levees and valley walls is a broad, lowlying floodbasin or 

backswamp zone, usually 3-5 km wide on both sides of the Ouachita River and somewhat 

less on the Saline River. This floodbasin zone is underlain partly by Recent sediments and 

partly by Pleistocene Deweyville 3 Terrace sediments, but the whole constitutes an expanse 

of terrain with imperceptible relief. A few irregular ridges (Gum Ridge, Pea Ridge) and pine 

islands rise above the backswamp, but the dominant features are hydrological in origin, 

either relict or modern. These hydrological features in the Felsenthal Project area are illustrated 

in Figure 6 and are listed below by name, with origin noted where ascertainable 

(Fleetwood 1969:Figure 3 and Felsenthal map): 

Open oxbow or cutoff lakes: 

Eagle Lake 

Pereogeethe Lake 

St. Marys Lake 

Open abandoned river courses: 

Jones Lake 

Fishtrap Lake 

Redeye Lake 

Horseshoe Slough 

Wildcat Lake 

Irregular open lakes and partly filled depressions: 

Marais Saline 

Boeuf Brake 

Mouton Brake 

Open Lake 

Grand Marais

Figure 6. Hydrological features in the Felsenthal Project area. 

Environment 27


Bayous partly reoccupying abandoned river courses: 

Eagle Creek 

Caney Bayou 

Lapoile Creek 

Lapile Bayou 

Sloughs, drains, crevasses, and backswamp distributaries: 

Carroll Slough 

Deep Slough 

Brushy Creek 

Many smaller hydrological features have only local names or lack names altogether. 

Those names which appear above commonly persist from eighteenth century French or 

have been anglicized; we find that some are correctly translated and others are consistently 

mistranslated by historical writers. Placenames are, of course, themselves a source of cultural 

and environmental information, as in the following few examples (see Branner 1899; 

McDermott 1941): 

Pereogeethe Lake—certainly from pirogue or periogue, “dugout canoe.” 

Boeuf Brake—shortened from boeuf sauvage, “buffalo.” 

Lapile Bayou—Hunter’s journal renders this as Bayu de la pelle, most correctly 

translated as “shovelnose sturgeon creek.” 

These names thus appear to recall an early type of watercraft and several animals no longer 

present in the region (see Buchanan 1974:78 on the shovelnose sturgeon). 

Our list of hydrological features emphasizes the prominence, diversity, complex distribution, 

and seasonality of aquatic habitats in the floodplain. All such habitats are interconnected 

during peak flood stage; many sluggish meandering sloughs or bayous increase their 

velocities, and may even reverse direction of flow, during rising and falling stages. Even at 

low water stage, circuitous drainages connect most backswamp areas with river channels 

and upland terrace edges. 

Topstratum/Substratum Deposits 

The Recent floodplain is underlain by a 35-m thick fill of clay, silt, sand, and gravel 

(Fleetwood 1969). As in the case of earlier terrace formations, Recent alluvium becomes 

finer-grained upward, and can be subdivided as substratum and topstratum deposits. The 

coarser substratum consists predominantly of channel or bedload materials while finer 

suspended load materials, deposited overbank, overlie the substratum (Allen 1965:127). The 

topstratum clays and silts in the Felsenthal Project area do not exceed about 8 m in thickness, 

and of these usually only 1 or 2 m, but not more than 5 m, are exposed above present


river level (lock stage at 62.2 feet MSL). Substratum deposits are exposed locally only in the 

deep excavation for the new Felsenthal Lock and Dam, which we have visited but not examined 

in detail. All of our site survey and testing work and all of the known geoarcheological 

record in the Felsenthal Project area pertain to the fine grained topstratum. 

Observations made during riverbank site survey and testing allow us to characterize 

the floodplain topstratum. First, these sediments are primarily clays and silts with minor 

occurrence of sands. Gravels, or dispersed gravel-sized particles, are exceedingly rare in 

the upper few meters of topstratum. At one location on the levee of Ouachita River we 

collected and analyzed 1.5 and 1.1 m column samples which consisted of clayey, silty, and 

sandy loams (Chapter 7). Second, the sediments ordinarily appear to be horizontally bedded 

without obtrusive breaks between units (Figure 7). Third, there is little variation in the drab 

grayish brown color of sediments. And last, organic components such as wood, bone, and 

shell are rarely preserved under the prevailing, strongly acid soil conditions (Chapter 7). 

(However, we know from published borehole data, from dredged spoil at a pipeline crossing, 

and from lock construction backdirt that subfossil wood, especially oak, is frequently 

encountered in water-logged levels of the topstratum; see specimens described in Appendix 

B.) We emphasize these characteristics of topstratum deposits because they are relevant to 

the Recent geological, archeological, and paleoecological record of the Felsenthal Project 

area, and suggest both limits and potential for future work. 

Floodplain Soils 

Published soil survey reports in the Felsenthal Project area are available for Ashley and 

Bradley counties (Larance 1961; Gill et al. 1979), but only a general soil map is available for 

Union County (U.S. Department of Agriculture 1968). The overflow bottomlands are uniformly 

mapped as Chastain series in Bradley and Union counties and as Guyton series in 

Ashley County. The latter series represents a revision of classification and terminology still 

incomplete for the region. We thus have the obvious inconsistency of soil series which coincide 

with political boundaries. For example, where a “panhandle” of Bradley County extends 

across the Saline River at Goulett Island one finds Bibb silt loam abutting Guyton soils 

along section lines. We are faced, however, with other sources of confusion. For example, a 

typical pedon of Guyton silt loam is described from a location mapped in the same report as 

the Pheba soil series (Gill et al. 1979:40, Sheet 41). This makes chemical and physical analyses 

reported for samples from this “typical pedon” doubtful. 

Nevertheless, we can briefly describe some properties of Guyton soils as an additional 

environmental characteristic of the Recent floodplain. These soils are strongly acid, siliceous, 

silt loams over finer textured subsoils, low in natural fertility and high in available water 

capacity. They are, of course, frequently flooded in winter and spring. A humic horizon or 

topsoil is weakly developed due to frequent prolonged flooding and repeated fluvial increments 

of silt and clay. Forest floor organisms which continuously degrade leaf litter and 

other debris in upland forests are limited by the flooding regimen. These lowlying floodplain 

soils, even on slightly elevated levees and ridges, very likely had no agricultural po-

30 

Figure 7. Topstratum deposits exposed by the Ouachita River bankline. a. partially 

obscured bankline with old cypress butts exposed; b. actively eroding bank 

section with bedding emphasized by a permeable sandy horizon (AAS neg. 

796513, 796544). 

A 

b


tential whatever for aboriginal human groups in the region. They have supported, however, 

a dense productive hardwood forest with many other useful wetland species of wild plants 

and animals. 

Channel Geometry and Channel Deposits 

Figure 8 clarifies the various terms which may be used to describe the channel and 

flow characteristics of a meandering river system. In the upper part of this figure are units 

of measurement which show, for example, how channel length (in river km or river miles) 

is related to downvalley distance. The sinuosity of a river is the ratio of channel length to 

downvalley distance (Leopold et al. 1964:281). In the lower part of this figure are lines of 

maximum flow and loci of maximum erosion and deposition with relation to the axis of 

a meander. The thalweg of a river is its low water channel; in a typical meandering channel 

the thalweg connects deep scour pools near the concave banks. In such a channel scour 

pools alternate with riffles or shoals at the crossover point between bends. In this brief 

discussion we shall employ such terms and concepts without delving more deeply into hydraulic 

processes; several terms explained by Figure 8 appear frequently in this report. 

The modern channels of the Ouachita and Saline rivers are sinuous patterns of meanders, 

usually asymmetrical in form, with rare straight reaches (Figure 6). An interesting 

characteristic of these channels is their relative stability. We were able to make a detailed 

comparison between General Land Office maps (1827) with section lines and modern 7.5 

minute topographic maps (1978), and this comparison shows virtually no measurable 

river channel changes in the Felsenthal Project area. (Sinuosity and stability are noted for 

particular reaches of these rivers in Chapter 5, Site Units 1, 2, and 4). While bank erosion 

has certainly proceeded, especially in the concave bank of meanders, this erosion has been 

markedly slow over the 151-year span of our maps. Lateral or down-valley migration of 

bends can hardly be detected, and no natural cutoffs can be identified for this period; bend 

for bend, the river channels of 1827 and 1978 are approximately congruent. Moreover, the 

various prominent oxbow lakes and abandoned courses in the project area were all abandoned 

before 1827. Certainly, this stability is attributable to a complex of factors, including 

magnitude of discharge, suspended load and bed load in transport, and cohesiveness of 

bed and bank materials (topstratum deposits; see Fisk 1947). It is clear that the major annual 

flood episodes on the Ouachita and lower Saline rivers are not also major erosional episodes; 

deep scouring and cutting of new channels is not indicated for the historic period. We 

are not so much concerned with these fluvial processes as with the fact of channel stability, 

for it furnishes two general approaches to evaluating riverbank archeological sites: 

1. site selection factors—we know, for example, that deep scour pools at the axis 

of bends and gravel shoals at the crossover point between bends have persisted 

for a century and a half; the spatial association of historic sites or late 

prehistoric sites with these channel environments may thus be related to site 

functions; Chapter 5 presents some site distributional data relevant to these 

locational questions.


Figure 8. Meandering channel characteristics and terminology (after Leopold 

et al. 1964:Figure 7-40). a. measurable channel features; b. 

flow characteristics and loci of erosion or deposition.


2. site preservation/destruction factors—we now have a scale for assessing 

lateral erosion (principally on concave banks) and deposition (principally on 

convex banks); in fact we perceive a surprisingly slow rate at which riverbank 

sites will become invisible or will be removed altogether by these processes; 

such an assessment is complicated by recent and prospective artificial controls 

on river regimen; bankline erosion is discussed in more detail later in this 

chapter. 

The bed load of a river consists of more or less coarse particles which move along or 

very near the bottom. These materials are transported principally during episodes of increased 

discharge and may otherwise be “stored” as gravel bars or shoals, in contrast to the 

more continuous transit of fine grained, suspended load particles. One of the principal river 

channel locations of storage has been noted, the crossover point between bends. The shallow 

riffles or shoals at such a location may be extensive, and may be broadly exposed at low water 

stage. The location and nature of shoals are of interest because they represent distinctive, 

shallow, and sometimes turbulent habitats preferred by certain fish and molluscs, among 

other species. The potential for human exploitation of these conditions is great, as for example 

by the use of fishweirs during seasonal spawning runs. We are also interested in gravel 

shoals as the probable source of stone for toolmaking and for various modes of cooking in 

an otherwise nearly stoneless alluvial plain. In Chapter 8 we describe a sample of bed load 

material from the Ouachita River collected for the purpose of testing these latter propositions. 

This sample was necessarily derived from dredged material since shoals are no longer 

exposed above river level (i.e., lock stage). 

The distribution of shoals and bars on the Ouachita River is well known from historical 

sources of the late nineteenth century when river improvements and navigational 

hazards were matters of great concern. Unfortunately, we lack the same data for the Saline 

River. From an authoritative report on Ouachita River conditions, we quote the following 

entries for the Felsenthal Project area (U.S. Congress 1874:“Table of Shoals from Camden, 

Ark. to the State Line”): 

Name Length Remarks 

Parrigeethe 2,500 feet (762 m) Gravel and sand... 

Eutaw Rapids 2,000 feet (609 m) Sand 

Caney Mary 8,000 feet (2439 m) Sand 

No name 16,500 feet (5030 m) Series of sand bars 

The U.S. Army Corps of Engineers hydrographic survey for 1873 locates these and other 

shoals by name, ordinarily also indicating the presence of gravel and the expected occurrence 

of shoals at the crossovers between bends. All of the shoals listed or mapped by these 

sources are found on the upper part of the Ouachita River (above the mouth of the Saline 

River), in contrast to the deeper channel characteristic of the Grand Marais Lowland


section downstream. One might expect, therefore, differences in archeological site density or 

distribution between these reaches of river (see Chapter 5, Site Units 1 and 4). 

Riverbanks and Bank Erosion 

During site survey work (described in Chapter 5), we examined riverbank exposures 

totaling over 100 km in length. One result of this work is an appraisal of ongoing riverbank 

erosion and variation in erosional processes related to channel pattern. Bankline erosion 

affects the initial exposure, dissection or truncation, and eventual destruction of numerous 

buried floodplain sites discussed in Chapters 5 and 6. Our observations are a preliminary 

attempt to understand present and future impacts on these cultural resources. 

Figure 9 represents two typical bank profiles which may be representative of many 

reaches on both Ouachita and Saline rivers and which may also represent erosional extremes. 

Numerous archeological sites were discovered on eroding banklines of each type, 

independent, we believe, of the intensity of erosion. Although we detected little map evidence 

of lateral erosion, there are many field indications that banks are retreating differentially. 

On the other hand, we noted few indications of lateral accretion except for obviously 

fresh sand bars at the axis of a bend or slightly downstream on the convex bank. Such 

prograding banklines are not subsumed by the profiles in Figure 9. The distinction between 

high intensity and low intensity profiles is borrowed and reinterpreted from a study of the 

Port Hudson Bluff, Louisiana (Brunsden and Kesel 1973). 

Our high intensity profile is steep, concave to near vertical, 1 to 5 m in height where 

observed. This profile is associated with the concave bank of meanders and is believed to 

be retreating at a relatively rapid rate for the Felsenthal Project area (but at a distinctly slow 

rate by comparison to other river channels like the Red River). The topstratum deposits are 

so uniformly fine grained and cohesive, that sapping, caving, slumping, and talus accumulation 

at the foot of the slope are rare (Fisk 1947:70). The shrink-swell tendency of these deposits 

causes blocky peds to dislodge periodically and fall from the upper slope, while a plume 

of muddy sediment washes away continuously at river level. Wetting and drying by rain or 

fluctuating water level (including wave wash) must accelerate this removal of particles. One 

of the most important aspects of these steep bank profiles is the potential exposure of deeply 

buried archeological materials at any elevation above river level, unobscured by talus (see 

3UN162 in Figure 25). A second important aspect is that sites so exposed are relatively more 

endangered by ongoing erosion. 

Our low intensity profile has low slope, slight convexity upward, and a height of 1 to 

3 m where observed. This profile is associated with convex to nearly straight reaches, and is 

believed to represent a slow rate of retreat, if not a steady state. The same uniform top-stratum 

deposits are exposed, and blocky peds are moving downslope to be removed at river 

level. However, sheet erosion is indicated by the frequent shallow exposure of tree roots 

(Figure 9b). Rills or gullies are also developed which must result from heavy rainfall and


Figure 9. Typical riverbank profiles in the Felsenthal Project area. a. high 

intensity, relatively rapidly eroding bank profile; b. low intensity, 

slowly degrading or nearly stable bank profile.


from the falling stages of flooding. Although little talus accumulates at the foot of these 

slopes, deeply buried archeological materials are not so well exposed. On the other hand 

we found that discovery and exposure of shallow buried sites was excellent near the upper 

slope and zone of exposed tree roots. In such sites the fine grained sediments may be winnowed 

away, leaving an array of artifacts and debris appropriate for controlled collection 

and precise plotting techniques (see False Indigo, 3AS285, and One Cypress Point, 3AS286, 

in Chapter 6). We believe these sites are threatened far less by bankline retreat than by sheet 

erosion and broad areal exposure of shallow remains. 

recent floodplAin ecology 

Our principal concern in this section is to identify and appraise potential human food 

resources of the overflow bottomland in the Felsenthal Project area and Grand Marais Lowland. 

Therefore, we exclude some animal and plant species found on terraces or floodplains 

where less rigorous hydrological conditions prevail. In doing so, we make an assumption 

that “extreme swamping” in this lowland, attested in nineteenth century and modern sources, 

also characterized the preceding several centuries when numerous prehistoric extractive 

camps were established in the floodplain (Chapters 5, 6). Regional archeological data, which 

do not now exist, are needed to synthesize and verify a model for late prehistoric human 

adaptation in the Grand Marais Lowland (Chapter 9). 

In order to present summaries of important animal and plant species we drew upon 

published sources of varying scope including: (1) broad areas of the southeast United States 

or Gulf Coastal Plain, (2) the Lower Mississippi Valley region, (3) the Ouachita Valley in 

southern Arkansas, and (4) localities near, but not in, the Felsenthal Project area. To remedy 

the lack of specific published sources, we also made limited field records of species occurrence 

in the overflow bottomland environment during site survey and testing work, and 

consulted with expert informants at Felsenthal National Wildlife Refuge and the University 

of Arkansas, Fayetteville. 

Hardwood and Swamp Forest 

Dense hardwood and swamp forest occupies more than 85% of the area of Recent 

floodplain, the remainder being open water. Species of the red oak group predominate in 

hardwood stands which are flooded approximately 40% of the year (Figure 10), but variation 

occurs among stands in response to hydrological and other environmental factors (Bedinger 

1971). Swamp forest refers to brakes and galleries of bald cypress and water tupelo, 

interspersed with hardwood forest, where inundation approaches 100% (Figure 11). The 

Ouachita and Saline River bottomlands in and near the Felsenthal Project area support some 

of the largest trees on record in the state, including the following (Moore 1960:22, 41, 55, 66, 

108):

Figure 10. Bottomland hardwood forest, predominantly mature red oaks, in the Ouachita River floodplain (U.S. 

Forest Service photo, July 1937, from Reynolds 1980:Figure 16) (AAS neg. 813844). 

37

38 

Figure 11. Hardwood and swamp forest in the Felsenthal Project area. a. bald cypress 

stand in an open lake; b. swamp privet, other shrubs, and saplings in the 

understory (AAS neg. 796506, 797155). 

A 

b


Bald cypress, Taxodium distichum, Union County, height 120 feet, dbh 9 feet 3 inches 

Water Tupelo, Nyssa aquatics, Ashley County, height 85 feet, dbh 5 feet 3.5 inches 

Pin Oak, Quercus palustris, Union County, height ?, dbh 4 feet 1.8 inches 

River Birch, Betula nigra, Ashley County, height 50 feet, dbh ? 

Water Elm, Planera aquatics, Union County, height 40 feet, dbh 15 inches 

During floodplain survey work, marked differences in composition of species were 

observed for various sites, e.g., river edges and bars, natural levees, oxbow lakes, sloughs, 

and backswamps. The following classification of forest types (American Society of Foresters 

1954) seems to encompass stands of similar composition in the Felsenthal Project area (only 

the dominant overstory species are listed): 

Hardwood forest types 

(92) Sweetgum-Nuttall Oak-Willow Oak 

(93) Sugarberry-American Elm-Green Ash 

(95) Black Willow 

(96) Overcup Oak-Water Hickory 

Swamp forest types 

(101-103) Bald Cypress-Water Tupelo 

In Table 1 we have compiled a list of overstory and understory species which occur 

in the project area; field identifications were made by reference to Maisenhelder (1958) and 

Moore (1960). Although this list is unlikely to be complete, we believe it is representative of 

the most common woody species which tolerate prolonged overflow in this floodplain area. 

More importantly, we annotated potential food resources, other technological resources, 

and certain wildlife uses by reference to Halls (1977) and other sources cited above. This 

summary suggests that bottomland forests of this type do not yield abundant, concentrated, 

wild foods for human consumption as do the loblolly pine-oak-hickory forests on elevated 

terraces. However, a detailed comparison of upland forest and bottomland forest resources 

has not been attempted for this region. A variety of technologically useful products are present 

(Table 1): wood, fuelwood, fiber, bark, and gum. Mast, fruits, and browse for wildlife 

consumption are evidently excellent during the summer-fall stage of low water. 

One understory species not listed in Table 1 is cane (Arundinaria gigantea). Although 

cane is an important element of floodplain vegetation in the Lower Mississippi Valley 

region, it does not tolerate prolonged overflow (Marsh 1977). Canebrakes or stands are not 

present today in the lowlying Felsenthal Project area, a fact also noted in 1804 by Dunbar 

and Hunter (Rowland 1930:245). The importance of canebrakes to wild turkeys and other 

floodplain animals in lower Arkansas River forests is noted by Meanley (1972:99).


Table 1. Resource Potential of Floodplain Trees, Shrubs, and Vines. 

Common Name Scientific Name Principal Uses and Seasonality 

OverstorySpecies: 

American Elm Ulmus americana bark can be peeled in summer; human use 

Baldcypress Taxodium distichum moisture- and decay-resistant wood; human use 

bitter pecan Carya aquatica nut crop in fall; minor wildlife use 

Black Willow Salix nigra fiber for cordage; human use 

carolina Ash Fraxinus caroliniana wood for implements; minor human use 

Cedar Elm Ulmus crassifolia 

Cherrybark Oak Quercus falcata var. pagodaefolia mast crop in fall; wildlife use 

Honey Locust Gleditsia triacanthos legume seeds and pulp in fall; human and wildlife use 

Nuttall Oak Quercus nuttalli excellent mast crop in fall; wildlife use 

Overcup Oak Quercus lyrata mast crop in fall; large sweet acorns; wildlife and human use? 

Pin Oak Quercus palustris mast crop in fall; wildlife use 

Persimmon Diospyros virginiana large fruits in fall; wildlife and human use 

Pumpkin Ash Fraxinus tomentosa wood used for tool handles; human use 

Red Maple Acer rubrum sugary sap in spring; human use 

silver Maple Acer saccharinum sugary sap in spring; human use 

Sugarberry Celtis laevigata small fruits in late summer; wildlife use 

Swamp Cottonwood Populus heterophylla 

Sweetgum Liquidambar styraciflua mast crop in late summer-fall; resinous gum used by humans 

Sycamore Platanus occidentalis 

Swamp Tupelo Nyssa sylvatica var. biflora small fruits in fall; wildlife use 

Water Oak Quercus nigra excellent mast crop in fall; wildlife use 

Water Locust Gleditsia aquatica legume seeds in fall; wildlife use? 

Water Tupelo Nyssa aquatica small fruits in fall; wildlife use 

Willow Oak Quercus phellos mast crop in fall; wildlife use 

Understory Species: 

American Hornbeam Carpinus caroliniana small nuts in late summer-early fall; wildlife use 

Boxelder Acer negundo sap is edible and available in spring; human use 

Buttonbush Cephalanthus occidentalis bark and twigs; used by beaver 

False Indigo Amorpha fruticosa legume seeds in fall; wildlife use 

Mayhaw Crataegus opaca small spring pomes in clusters; wildlife and human use 

Possumhaw Ilex decidua small fruits in early fall; wildlife use 

river birch Betula nigra wood used for implements, basketry; minor human use 

Swamp Privet Forestiera acuminata small fruits in summer; wildlife use? 

Water Elm Planera aquatica fuelwood; minor human use 

Vines: 

Wild Grape Vitis sp. clusters of fruit in late summer-fall; wildlife and human use 

Note: Many species of hardwoods listed above may be important as fuelwoods, e.g., oaks and maples

Vertebrates 


Lowlands of the Ouachita Valley support an abundant, diverse, vertebrate fauna 

including, for the most part, species common to forests and wetlands of the Lower Mississippi 

Valley and southeast United States (U.S. Department of Agriculture 1978:Exhibit 12). 

We will attempt to make a selective presentation of mammals, birds, reptiles, amphibians, 

and fishes which inhabit or regularly intrude riparian and aquatic habitats in the Felsenthal 

Project area, and which also appear to represent potentially important food resources for 

aboriginal human populations in the region. Many wildlife elements are mentioned in the 

Dunbar-Hunter journals of 1804-1805. These historical notes include some species rare or extinct 

in the region today, e.g., a large trumpeter (?) swan and alligators noted in the vicinity 

of Caney Marais (McDermott 1963:93) and buffalo on the upper Ouachita River (Rowland 

1930:264). Archeological records of vertebrates are few for the Felsenthal region as a whole, 

but a pertinent sample from the Shallow Lake site (3UN9/52) is described by Rolingson 

and Schambach (1981:Chapter 6). The extinct passenger pigeon is rare, but present, in this 

sample. 

Table 2 lists mammals of potential importance in order of body weight, omitting 

small rodents, shrews, and bats (all generally less than 50 g). Data included are taken from 

Sealander (1979) and other sources noted. This tabulation represents most of the large and 

medium sized mammals of the Ouachita Basin in Arkansas (55 species appear in a minimal 

modern checklist; U.S. Department of Agriculture 1978). The order of presentation by body 

weight is, of course, not necessarily the order of importance in an aboriginal subsistence 

system; many of these animals are solitary, nocturnal, wary, or relatively less palatable, and 

some are principally important as furbearers. Archeological data from other regions of the 

eastern United States suggest that white-tailed deer would far exceed all other mammals 

in importance as a food animal, with raccoons a distant second in importance (Smith 1975). 

From the Shallow Lake site (3UN9/52), immediately adjacent to the floodplain in the Felsenthal 

Project area, the following mammals are reported (Rolingson and Schambach 1981:Table 

4): black bear, white-tailed deer, bobcat, raccoon, gray fox, opossum, skunk, cottontail 

rabbit, fox squirrel, gray squirrel, mink, and small rodents. Some animals from both floodplain 

and upland habitats are certainly missing here because of the size and nature of the 

archeological sample. We conclude from modern and limited archeological data that deer 

and other mammals taken on the emergent floodplain during summer-fall, and occasionally 

also under flood conditions in winter-spring, would contribute a significant increment of 

protein and fat in the aboriginal diet. 

A modern checklist of birds for the Ouachita River Basin in Arkansas includes 62 

resident species, 59 summer and 56 winter migrants, and 97 transient species (U.S. Department 

of Agriculture 1978). A number of species formerly present in the region are omitted 

from this list, including passenger pigeon, Carolina parakeet, and ivory-billed woodpecker 

(James 1974). Numerous other species are endangered, including large wading birds, raptors, 

and some waterfowl characteristic of Grand Marais Lowland habitats. Again referring


Table 2. Floodplain Mammals in Order of Body Weight. 

Common Name Scientific Name Body Weight Potential Uses/Comments 

Buffalo or Bison Bison bison 1000-600 kg, flesh, hides; probably not present in numbers before 

rarely 1200 kg 1700s; last large herd killed in Saline River bottoms 

1808 (Sealander 1979:16); extinct in region 

Black Bear Ursus americanus 227-100 kg, flesh, fur, tallow or grease; prefers bottomland forrarely 

275 kg est and canebrakes; heavily exploited by French 

and Indian hunters in early historic period; probably 

taken commonly by aboriginal hunters in region 

White-Tailed Deer Odocoileus virginianus 136-40 kg flesh, hides, sinew; prefers bottomlands for mast, 

browse fleshy fruits (Stransky 1969); species of 

major importance to aboriginal hunters in region 

Mountain Lion or Felis concolor 103-36 kg flesh, fur; prefers heavily forested habitat with 

Cougar dense deer population; an endangered species still 

present in Ouachita and Saline River bottomlands 

Red Wolf Canis rufus 41-16 kg flesh, fur; wide-ranging species in Gulf Coastal 

Plain, including bottomland forest; extinct in region 

(Sealander and Gipson 1974:124) 

beaver Castor canadensis 35-11 kg flesh, fur; feeds on sweetgum, buttonbush, other 

trees and shrubs of bottomland forest 

Bobcat Lynx rufus 20-6 kg flesh, fur; probably not a heavily exploited species, 

but known to have been taken by aboriginal hunters 

in region 

river otter Lutra canadensis 14-45 kg flesh, fur; a possibly endangered species still present 

in Ouachita and Saline River bottomlands 

raccoon Procyon lotor 14-4 kg flesh, fur; prefers bottomland hardwood forest and 

stream or lake margins; species of major importance 

to aboriginal hunters in region 

Gray Fox Urocyon cinereoargenteus 6.4-2.5 kg flesh, fur; preferred habitat includes bottomland 

forest 

Opossum Didelphis virginiana 6-3 kg usually taken principally for flesh; prefers bottomland 

forest 

Swamp Rabbit Sylvilagus aquaticus 2.9-1.2 kg flesh, fur; prefers bottomland hardwood forest and 

canebrakes 

Muskrat Ondatra zibethicus 1.8-0.5 kg flesh, fur; prefers river and stream banks, as well as 

lakes and ponds 

Mink Mustela vison 1.6-0.8 kg flesh, fur; prefers river and stream banks 

Gray Squirrel Sciurus carolinensis 710-317 g flesh, fur; preferred habitat includes bottomland 

hardwood forest 

Long-tailed Weasel Mustela frenata 250-85 g flesh, fur; preferred habitat includes bottomland 

hardwood forest 

Southern Flying Glaucomys volans 98-40 g usually only taken for fur; preferred habitat includes 

Squirrel bottomland hardwood forest


to the archeological record from Shallow Lake (3UN9/52), we find a substantial representation 

of winter waterfowl such as Canada geese and ducks, turkey, and a single occurrence 

of passenger pigeon (Rolingson and Schambach 1981:Table 4). Other bird remains were 

unidentified, and sampling error here is likely to be present. Wild turkey, for example, is 

exceeded by waterfowl in this sample. In fact we would expect heavy use of turkey by the 

aboriginal residents of this region, as elsewhere in eastern United States (Smith 1975). 

The same modern checklist for the Ouachita River Basin includes 61 reptiles and 32 

amphibians, although many of these occur in upland habitats not under consideration here. 

Those animals occurring in riparian or aquatic habitats of the floodplain which can feasibly 

be exploited for human food include alligator, snapping turtles, sliders, softshells, map 

turtles, mud and musk turtles, water snakes, western cottonmouth, salamanders, siren, 

amphiuma, Red River waterdog, and the bullfrog. Alligators occasionally reached over 5 

m length prior to historic settlement and commercial exploitation in the region. Since that 

time the species has been eliminated, restocked, and remains endangered (Reagan 1974). 

Aboriginal uses for this animal included flesh, hides, oil, and ivorylike teeth. One very large 

alligator tooth has been recovered from the Gordon site (3AS152) near Saline River, just 

north of the Felsenthal Project area (Arkansas Archeological Survey site files). Many other 

species noted above have been recorded occasionally in archeological deposits. At Shallow 

Lake (3UN9/52) only snapping turtle, slider, and the upland box turtle have been identified 

(Rolingson and Schambach 1981: Table 4). We believe that the use of various aquatic turtles, 

alligator, and frogs could be more important than this limited record indicates. 

The final group of vertebrates considered here are fishes which may have constituted 

a highly important food source for the aboriginal inhabitants of the Grand Marais Lowland. 

In general, fishes here are abundant, diverse, high in food value, and susceptible to capture 

by simple techniques. Regional inventories and studies include those of Reynolds (1971), 

Buchanan (1974), and Robison (1975). Some records are from within the Felsenthal Project 

area. Table 3 summarizes the most important species or groups by body weight and habitat 

preference. Many of the smaller fishes such as darters and shiners are omitted. However, 

body weight is far less likely to determine aboriginal use than behavior of a species. Certain 

fishes are more easily exploited by mass capture techniques. Large differences in usable 

meat, palatability, and food value also exist among species. A historical account notes, for example, 

that buffalofish “floated” in large numbers during August and could be easily gigged 

(W. T. Martin in Etheridge 1959:xviii). Buffalofish also can be taken in quantity during spring 

flooding when spawning in shallow areas of the floodplain forest (Comeaux 1972). Mass 

capture of gizzard shad in a dwindling floodplain lake was described earlier in this chapter 

(Limp and Reidhead 1979). Shad and eel, among species listed in Table 3, are particularly 

high in fat content. Migratory and spawning species which ascend or descend the rivers


Table 3. Selected Food Fishes and Habitat Preferences, Lower Ouachita River System in 

Arkansas. 

Approximate Habitat Preferences 

Maximum River Sloughs/ Oxbows/ 

Common Name Scientific Name Weight (kg) Channels Bayous Open Lakes Comments 

Lake Sturgeon Acipenser fulvescens 135 x x Endangered species 

Paddlefish Polyodon spathula 68 x x x Endangered species 

Bowfin Amia calva - x x x 

Bigmouth Buffalo Ictiobus cyprinellus 30 x x 

Flathead Catfish Pylodictis olivaris 23 

Freshwater Drum Aplodinotus grunniens 20 x x 

Channel Catfish Ictalurus punctatus 9 x 

Catfish Ictalurus sp. - x 

Smallmouth Buffalo Ictiobus bubalus 9 x x 

gars Lepisosteus sp. - Alligator gar is depleted 

Walleye Stizostedion vitreum 5 x x 

Largemouth Bass Micropterus salmoides 4 x x 

American Eel Anguilla rostrata - x Catadromous species 

Shovelnose Sturgeon Scaphirynchus platorynchus 3 x Endangered species 

Bullheads Ictalurus sp. 1 x x 

crappie Pomoxis sp. 1 x x 

Pickerel Esox sp.


were susceptible to capture at shoals by gigging, spearing, or trapping in weirs. Species frequenting 

pools or ponds were sometimes poisoned or taken with nets, techniques that leave 

no direct evidence. 

Minimal data from the Felsenthal Project area are available for fish utilization during 

the prehistoric period. At Shallow Lake (3UN9/52) fishbone was abundant in refuse pits and 

midden; only catfish, gar, and drum are identified (Stacy 1976:61; Rolingson and Schambach 

1981:Table 4). In Chapter 6 we present test excavation data from two floodplain extractive 

sites of the early Mississippi period (A.D. 1100-1400) which have small, but dense, midden 

deposits with abundant fishbone (False Indigo, 3AS285, and Jug Point Cutoff, 3BR76). These 

small camps’ inhabitants were evidently intensively engaged in cooking or rendering fish 

and fish oil. 

Invertebrates 

Two groups of invertebrates are likely to have been exploited regularly by aboriginal 

inhabitants of the Grand Marais Lowland. The first group, freshwater mussels, was most 

accessible in river channels during summer-fall low water stage, and mussel shell remains 

have been preserved in numerous sites in the region. The second group, crawfish, were collected 

or trapped principally during winter-spring flooding on bottomland. 

In Appendix B we summarize a modern collection of molluscs from the Felsenthal 

Project area and limited archeological collections from both prehistoric and historic sites 

(5 species in each context). A slightly longer species list has been compiled for the Shallow 

Lake site (3UN9/52), but is equally unlikely to be a comprehensive sample; buckhorn, pigtoe, 

and three-ridge mussels predominated in this collection. Mussels now known from prehistoric 

sites represent a variety of habitats, including shoals, sloughs, and oxbow or open 

lakes. As noted earlier in this chapter, mussels would have been supplementary rather than 

staple food resources. A recently published checklist of unionid molluscs for the Ouachita 

River system in Arkansas indicates that the lower reaches remain poorly known (Gordon et 

al. 1979). 

Crawfish of several species occur frequently in riparian and aquatic habitats of the 

Grand Marais Lowland. The species of potential importance are Procambarus clarkii (swamp 

crawfish) and P. blandingi (river crawfish), both taken commercially in the Atchafalaya Basin 

and other regions of Louisiana (Comeaux 1972:Chapter 5). In this area young crawfish are 

baited with a stick or line or trapped during winter-spring flooding of bottomlands. The 

exoskeleton of this animal does not persist in archeological deposits, and its record is correspondingly 

rare or lacking. Almost certainly, crawfish were frequently used in the Felsenthal 

Project area during the prehistoric period. 

The last potential foodstuff considered here is that most desirable product of the honeybee. 

Hives and honey were (and still are) located in the hollows of many large trees above


the level of annual flooding. In historic times woodsmen were able to track honeybees to 

the bee tree and cut or burn it down, often taking a large quantity of honey and beeswax 

(Etheridge 1959:80). This low water, summer-fall activity became an important early historic 

industry in the Atchafalaya Basin and other areas of extensive bottomland forest (Comeaux 

1972:92). We would expect considerable use of this foodstuff by aboriginal inhabitants of the 

Grand Marais Lowland. Unfortunately, there is little possibility of an archeological record of 

its use. 

potentiAl pAleoecologicAl studies 

This chapter presented a summary of environmental data for the Felsenthal Project 

area and lower Ouachita Valley in Arkansas. In this section we acknowledge that paleoenvironmental 

data, which could support our view of regional prehistoric adaptations, are truly 

deficient. There are no palynological, paleontological, or geochronological studies within or 

near the project area. But the work of geologists cited earlier (chiefly mapping and map interpretation) 

indicated a dynamic geomorphic history for the last 100,000 years or so. Relict 

terraces, lacustrine features, abandoned river courses, cutoff lakes, and a sequence of alluvial 

deposits attest to this dynamic history. We are convinced that a wide range of paleoecological 

data are recoverable, and that interdisciplinary paleoenvironmental study is badly 

needed in the Felsenthal region, especially where modern land or river uses are modifying 

or effacing the paleoenvironmental record. 

One locality in the Felsenthal Project area, Goulett Island, may serve here as an example 

of potentially significant study sites. Goulett Island is a “pine island,” an erosional 

remnant of the Deweyville 2 Terrace, truncated by the modern channel of the Saline River. 

We refer to archeological and historical remains on Goulett Island at various points in this 

report (Chapter 5, Appendixes A, C). Here we are concerned with Goulett Island’s natural 

values and scientific potential. Goulett Island was in fact first studied by a geologist, David 

Dale Owen (1860), who described a riverbottom lignite ledge or outcrop mined for fuel by 

local inhabitants (see historic site 3BR8 in Chapter 5). This outcrop is no longer visible, even 

at low water stage. Goulett Island was also visited by the early archeological entrepreneur, 

C. B. Moore (1913), who made no useful observations beyond his perfunctory mound investigations 

(he does mention a “great flood of 1912” on the Saline River). We doubt whether 

geologists have studied the Goulett Island locality or published any observations pertaining 

to it since the time of Owen. 

During our own riverbank survey work, we examined the exposure of Deweyville 2 

Terrace sediments at Goulett Island. This exposure is a section of riverbank about 4 m high 

and 1000 m long. The topstratum deposits of the Deweyville 2 Terrace formation here consist 

of gray laminated silty clays and clays with charcoal and/or carbonized wood exposed at 

certain levels. No artifacts were observed in place, although there are abundant pre-


historic and historic artifacts eroding downslope or discarded overbank from the surface 

of Goulett Island. If artifacts were discovered in place they would presumably be mid-Wisconsin 

in age or about 30,000-25,000 years B.P., according to Saucier and Fleetwood’s (1970) 

interpretation of Ouachita Valley chronology. Moreover, the sediments exposed at Goulett 

Island may represent lake bottom or nearshore clays and silts deposited by Pleistocene Lake 

Monroe, at or near its upstream (northernmost) margin as proposed and mapped by Saucier 

and Fleetwood (1970:Figure 5). 

Goulett Island is probably the most significant outcrop of Lake Monroe sediments 

in the Felsenthal region, and could furnish samples to verify a lacustrine paleoenvironment 

and especially to corroborate (by radiocarbon assay) the age of Lake Monroe and the 

Deweyville 2 Terrace. Although we recommend a broad range of interdisciplinary study in 

conjunction with mitigation of archeological sites in the Felsenthal Project area (Appendix 

C), we specifically recommend detailed sedimentological and geochronological study of the 

Goulett Island locality. 

potentiAl lowlAnd resources 

In the preceding summary of biophysical characteristics in the Grand Marais Lowland 

and Felsenthal Project area we relied upon historical sources, published ecological and 

geohydrological data, regional archeological data, and our own field observations. Figure 

12 integrates some of these data in the form of a schematic cross section extending across 

the Felsenthal Project area and Ouachita Valley at about the latitude of Shallow Lake site 

(3UN9/52). Variation between upland pine-oak forest, locus of prehistoric permanent settlements, 

and lowland hardwood and swamp forest with numerous water courses, locus of 

impermanent extractive camps, is emphasized in this drawing. 

We have attempted to delineate potential floodplain food and other useful resources 

through limited modern and archeological information from the region (Tables 1-3). Many 

fundamental questions about local resource potential remain to be answered. Direct evidence, 

in the form of well preserved floral and faunal remains from floodplain sites in this 

region, is likely to be rare (and has in fact proved to be so; see Chapters 6 and 7). The following 

questions deal with the Felsenthal Project area specifically, but are broadly relevant to 

studying prehistoric subsistence systems in the region: 

1. What was the relative importance of floodplain floral and faunal resources? 

We are inclined to believe that fish was an exceedingly important protein source during the 

Mississippi period (A.D. 1100-1700), and probably also during earlier periods. Certain fish 

species were seasonally abundant or concentrated in floodplain aquatic habitats and were

48 

Figure 12. Cross-valley section in the Felsenthal Project area showing terraces, floodplain, forest types, and other features (geology 

adapted from Fleetwood 1969 and Saucier and Fleetwood 1970: Figure 8f; no horizontal or vertical scale).


quite likely harvested in bulk quantities. Floodplain mammals (especially white-tailed deer), 

birds, reptiles, amphibians, molluscs, and other invertebrates were certainly also supplementary 

protein sources, and many of these species were migratory and seasonally abundant, 

or were seasonally concentrated on pine islands by rising floodwaters. 

Among floodplain floral resources, nut crops were the major food supply available in 

both quantity and variety. The various red oaks, such as willow oak (Quercus phellos), occur 

in dense stands on the natural levees and produce substantial mast crops in fall, prior to inundation. 

Hilliard (1981) shows that red oak acorns were heavily used by prehistoric people 

of the Ozarks, could be collected, stored, and processed efficiently, and are higher than 

white oak acorns in caloric value. One very abundant nut tree of the floodplain forest, bitter 

pecan (Carya aquatica), does not appear to have any importance as a human food source, in 

spite of the productivity and usefulness of other hickories and pecans (Smith 1953). 

On the whole, we identify more “first line” animal foods than plant foods in the overflow 

bottomland environment, and we suggest that prehistoric subsistence strategies in the 

region focused on these lowland faunal resources. 

2. What was the relative importance of uplands (terrain above prolonged overflow) 

and lowlands (overflow bottomlands) as resource zones? 

This question is at least partly beyond the scope of our study, since we would have to control 

environmental and paleoenvironmental data from uplands in the region. In particular 

we would require knowledge of prehistoric agricultural systems in the region which is 

not now available. Schambach proposes some local agricultural adaptations, but there are 

presently no archeological data to support these propositions (Rolingson and Schambach 

1981:105): 

In late periods, if there was any horticulture, it must have been slash and burn horticulture 

in the uplands by people living in small hamlets or single farmsteads scattered 

along the tributary streams....Only extensive exploitation of the uplands...could have 

supported the populations [represented] in the Felsenthal region. 

Smith’s (1978) identification of major Mississippi period food resource groups, mentioned 

earlier in this chapter, might be more specifically adapted to the Felsenthal region as 

follows:


Major Resource Groups Primary Environments 

1 corn, beans, squash upland terraces 

2 channel, lake, and backwater fish species floodplain 

3 migratory waterfowl floodplain 

4 deer, raccoon, turkey terraces and floodplain 

5 nut, fruits, berries terraces and floodplain 

6 pioneer seed plants floodplain 

7 alligators, crawfish, frogs, turtles floodplain 

This simple tabulation suggests that overflow bottomlands were of enormous importance 

to the local Mississippi period population at any given period which had “two strings 

to its bow—wild and agricultural resource....time devoted to hunting and gathering [here 

primarily in floodplain environments]... was almost equal to that devoted to agriculture” [in 

uplands] (Peebles 1978:392). 

3. What are some “second line” food resources and nonfood materials specifically 

available or abundant in the floodplain? 

We cannot claim to have made a systematic appraisal of these resources, but the following 

items are suggested: 

Second Line Foods Other Raw Materials 

snakes, small turtles alluvial clays 

mussels, snails novaculite and other gravels 

beaver, otter, muskrat cypress and other implement woods 

honey, maple sap hardwood fuels 

persimmon, riverbank grapes, small fruits bark, fiber, gum 

seeds, shoots, greens furs, skins, feathers 

tallow, grease

chapter 3 

prehistory of the felsenthAl region 

by John H. House 

introduction 

The Felsenthal archeological region, recognized in the State Plan for the Conservation 

of Archeological Resources in Arkansas (Davis 1981) is a subdivision of the Lower Mississippi 

Valley archeological subarea of the Southeast. The Felsenthal region was given formal definition 

by Schambach (Rolingson and Schambach 1981:103-106) and assigned to the Lower 

Valley on the basis of physiographic and biogeographic continuities with the Lower Valley, 

as well as archeological evidence. Throughout much of the prehistoric and protohistoric 

aboriginal sequence, components in the Felsenthal region exhibit greater formal similarities 

to contemporary archeological culture units in the Lower Valley to the east than to culture 

units in the Caddo subarea to the west. 

The Felsenthal archeological region encompasses the Felsenthal Navigation Pool 

project area and the Felsenthal National Wildlife Refuge (see Chapter 1). The region centers 

on the Ouachita River between Camden, Arkansas, and Ouachita City, Louisiana, where 

the Ouachita River enters the Mississippi Alluvial Plain. It also extends up the Saline River 

to the vicinity of Rison, Arkansas and includes contiguous West Gulf Coastal Plain upland 

areas within the Ouachita River and Saline River watersheds. In Arkansas, adjoining archeological 

regions recognized in the State Plan (Davis 1981) are: the Great Bend of the Red River 

and the Little Missouri regions to the west, the Middle Ouachita (cf. Hodges and Hodges 

1945) and Middle Saline regions on the north, and the Bartholomew-Macon region on the 

east. 

The assignment of the Felsenthal region to the Lower Mississippi Valley subarea reverses 

the previous assignment of this portion of the Southeast to the Caddoan subarea (Davis 

1961; Phillips 1970:861; cf. Rolingson 1976a:100). Schambach (1979:21) has emphasized, 

however, that the Felsenthal region is on a frontier or border area between these two major 

archeological subareas of North America and that the archeological record in the region 

should reflect this frontier situation. 

The purpose of the present chapter is to review the history of archeological investigation 

in the Felsenthal region and briefly summarize our current knowledge of the cultural 

sequence and the nature of cultural development in the region during prehistoric and protohistoric 

times. The history of the region will be discussed by Watkins in Chapter 4 of this

52 House 

report. Locations of the principal archeological sites mentioned in this chapter are indicated 

on the map of the Felsenthal region, Figure 13. 

preVious ArcheologicAl inVestigAti0n 

exploration 

Although sporadic references to aboriginal mounds are made in the writings of early 

explorers in the region, notably Dunbar and Hunter (Rowland 1930; McDermott 1963; see 

also Chapter 4 of this report), the first scientific archeological investigation in or near the 

Felsenthal region appears to have been made by Cyrus Thomas (1894:250) at Pargoud Landing 

(16OU1) on the Ouachita River near Monroe, Louisiana. The first systematic reconnaissance 

in the region, however, was by Clarence B. Moore of the Academy of Natural Sciences 

of Philadelphia in the early decades of this century. In 1908-1909, Moore (1909) ascended 

the Ouachita River as far as Camden, Arkansas. On the southern margin of the region 

Moore briefly investigated Pargoud Landing and dug extensively at the Glendora site on 

the Ouachita and the nearby Keno site on the lower reaches of Bayou Bartholomew. In the 

Felsenthal region in Arkansas, Moore investigated the following sites in 1908-1909: 

Mound near Lock Number 6 (3AS6), Ashley County 

Mounds near Green Lake (3BR1), Bradley County 

Cemetery near Caryville Landing (3UN11), Union County 

Mounds near Pigeon Hill (3UN12), Union County 

Cemetery in Boytts Field (3UN13), Union County 

Mounds near Purdue Wood-Camp (no assigned state site number), Union County 

Cemetery at Bell Gin Landing (3UN99), Union County 

Mounds at Bell Landing (3UN51), Union County 

Mounds at Boone Place (3CA9), Calhoun County 

Mounds at Keller Place (3CA13), Calhoun County 

Mounds and sites near Pyles Landing (3CA26), Calhoun County 

Mound and cemetery at Kent (3OU6), Ouachita County 

Returning to the region in 1912-1913, Moore (1913) ascended the Saline River as far as 

the vicinity of Warren, Arkansas. Sites investigated by Moore on the Saline River in Arkansas 

are: 

Site on Goulett Island (3BR8), Bradley County 

Site at Hampton Landing (no assigned state site number), Bradley County 

Site near Gee Landing (3DR17), Drew County 

Site near Lowrie Landing (3BR5), Bradley County 

Mound near Wherry Landing (no assigned state site number), Bradley County

Figure 13. A portion of the Ouachita River Valley designated the Felsenthal region and 

locations of some important prehistoric sites. 

53

54 House 

Site near Wire Fence Landing (no assigned state site number), Bradley County 

Site near Sutton Ferry (3BR6), Bradley County 

Site near Brooks Landing (no assigned state site number), Bradley County 

The excavations in aboriginal cemeteries at Glendora (Moore 1909: 27-80) and Keno 

(Moore 1909:120-151) yielded polished and engraved ceramic bowls and bottles occurring 

with early European trade goods. Recovery of these vessels by Moore led to the subsequent 

long standing assignment of this portion of the Southeast to the Caddoan subarea (Davis 

1961). Observations on an extensive human skeletal series recovered by Moore at Boytts 

Field (3UN13) in Union County, Arkansas, were reported by Hrdlicka (1909). 

In the decades following Moore’s work, archeological investigations in the Ouachita 

River basin were largely confined to portions of the basin above the Felsenthal region 

(e.g., Harrington 1920; Hodges and Hodges 1945; Wood 1963; Schambach 1970) or below 

the Felsenthal region (e.g., Ford 1935, 1936; Walker 1936; Ford and Willey 1940). Price and 

Heartfield (1977:4) refer to unpublished investigations by James A. Ford at Pargoud Landing 

(16OU1) in 1935 resulting in recovery of a flexed burial. 

During the 1940s and 1950s, amateur archeologist Frank J. Soday and associates of 

El Dorado, Arkansas, excavated at a number of late prehistoric sites within the Felsenthal 

region primarily at Locust Ridge (3UN8), Watts Field (3UN18), and Jones Lake (3UN81). 

No report on these excavations has ever been made available but copies of site records were 

given by Soday to the Arkansas Archeological Survey. Frank Schambach and J. Cynthia 

Weber of the Arkansas Archeological Survey subsequently made photographic records of 

ceramic vessels with site and grave lot provenience in the Soday and Franz collections from 

the Felsenthal region. 

Recent Investigations 

Archeological reconnaissance and small scale excavations in the Felsenthal region by 

amateur members of the Arkansas Archeological Society began in the 1960s. After 1968 these 

efforts were augmented by reconnaissance and testing by Arkansas Archeological Survey 

station archeologists from Magnolia and Monticello. These activities have contributed to a 

growing regional data base in the Arkansas portion of the Felsenthal region. Excavations 

carried out at the Paw Paw site (3OU22) near Camden as a part of the Arkansas Archeological 

Society training session in 1971 yielded data on Fourche Maline and Archaic occupations 

at the northern margin of the Felsenthal region (Weber, excavation and analysis notes, 1973, 

Arkansas Archeological Survey).

Prehistory 55 

The history of recent archeological investigations along the Ouachita River in the Monroe, 

Louisiana, locality is recounted by Price and Heartfield (1977:6-7) and by Price (U.S. 

Army Corps of Engineers 1980). In the 1960s the Northeast Louisiana Archeological Society 

conducted excavations at Myatt’s Landing, one of the sites investigated by Moore, but no 

report on this work has been published. Excavations by the Northeast Louisiana Archeological 

Society and Northeastern Louisiana University at the suspected site of Colonial Spanish 

Fort Miro (16OU3) in downtown Monroe are reported by Greene et al. (1975). Follow-up excavations 

at this site were funded by the U.S. Army Corps of Engineers, Vicksburg District, 

supported by the city of Monroe and directed by archeologists from Northeast Louisiana 

University (Price et al. 1975). 

Subsequent investigations in the Monroe locality include small-scale cultural resource 

management surveys conducted by archeologists from Northeastern Louisiana University, 

as well as ongoing surveys and excavations sponsored by the U.S. Army Corps of Engineers, 

Vicksburg District in conjunction with navigation channel works on the Ouachita River (U.S. 

Army Corps of Engineers 1980). The latter efforts include salvage excavations at the T. E. 

Salsbury site (16OU15) where numerous Mississippi period burials accompanied by ceramic 

vessels, largely assignable to the Plaquemine Culture, were recovered (Price and Heartfield 

1977). 

Glen S. Green of Northeastern Louisiana University conducted further excavations at 

Pargoud Landing (16OU1) in 1975, 1976, 1977, and 1979. No report on these investigations 

is yet available but preliminary results are briefly summarized in a recent cultural resources 

survey report (U.S. Army Corps of Engineers 1980, Chapter 5). 

Investigations by the Arkansas Archeological Survey in the Felsenthal National Wildlife 

Refuge began with an initial survey or reconnaissance conducted by Rolingson and 

Schambach in 1971. This reconnaissance involved visits to previously recorded sites and 

recording of new sites (Rolingson 1972). The 65 sites on record at the conclusion of the 

project included a number of late prehistoric mound centers and habitation sites on terraces 

overlooking the floodplain and several small, low density prehistoric artifact scatters in the 

floodplain. Artifacts recovered suggested closer ties to the Coles Creek and Plaquemine 

cultures of the Lower Mississippi Valley than had been anticipated. Evidence of occupation 

dating to the Archaic, Poverty Point, Tchula, and Marksville periods was also recognized. 

The probability that many sites may be buried under recent alluvium in the floodplain was 

also noted (Rolingson 1972).

56 House 

The following year, test excavations were conducted at six sites in the Felsenthal 

National Wildlife Refuge by the Arkansas Archeological Survey under a cooperative agreement 

with the National Park Service. This project, directed by Lischka, included testing at 

the following sites: Shallow Lake (3UN9/52), Bent Tree (3UN75), Locust Ridge (3UN8), Watts 

Field (3UN18), Disturbed Cemetery (3UN84), and Oven (3UN53). The major components 

represented at these sites were attributed by Lischka (1973) to Coles Creek and Plaquemine 

culture occupation. Evidence of Marksville and Archaic period occupations was also encountered. 

The program of testing initiated by Lischka was continued by incoming Survey archeologist 

V. K. Pheriba Stacy in 1975. Her work, jointly sponsored by the Interagency Archeological 

Services, the U.S. Army Corps of Engineers, Vicksburg District, and the Arkansas 

Archeological Survey, consisted of excavation of two test pits at Pereogeethe Lake 1 (3BR52) 

and intensive excavation in Mound C at Shallow Lake (3UN9/52) which had been subject to 

vandalism. Stacy’s notes on these excavations also drew together data from Lischka’s work 

at Shallow Lake and other sites. A report on the Shallow Lake site was written by Martha 

Ann Rolingson and Frank F. Schambach (1981). Schambach had done considerable work in 

the Felsenthal region, and used the data as a vehicle for a regional summary and introduction 

of a new ceramic typology (Rolingson and Schambach 1981: 101-176). His ceramic classification 

is a response to difficulties previously encountered in describing and classifying 

Felsenthal region ceramics by either the current Lower Mississippi Valley (Phillips 1970) or 

Caddoan areas (Suhm and Jelks 1964; cf. Webb 1981) typologies. This analytical classification 

is discussed and applied to a large ceramic sample from a newly recorded floodplain site by 

Hemmings in Chapter 8 of this report. 

A cultural resources assessment of the projected Corps of Engineers Calion Lock and 

Dam project on the Ouachita River near Calion, Arkansas was carried out by David Kelley 

of the Arkansas Archeological Survey in 1978-1979. The fieldwork involved systematic subsurface 

testing in the entire 244.4 ha impact zone located in the Ouachita River floodplain. 

The study provided valuable new information on the Bell Gin Landing site (3UN99) which 

had been investigated by Moore in 1908-1909 and revealed one previously unrecorded site 

in the impact zone, an isolated find of a late Archaic point (Kelley 1979). 

A cultural resources reconnaissance of the Felsenthal National Wildlife Refuge was 

conducted by Heartfield, Price and Greene, Inc. of Monroe, Louisiana, in fall 1979 concurrent 

with present investigations. This study identified five additional archeological sites, 

four isolated finds, and nine standing structures in addition to the 63 archeological sites 

already on record within the refuge (Heartfield, Price and Greene, Inc. 1980).

Prehistory 57 

Recent and ongoing efforts to identify and evaluate archeological resources in proposed 

lignite strip mining areas in the West Gulf Coastal Plain uplands near Hampton in 

Calhoun County, Arkansas, have focused attention on the archeological record of prehistoric 

and historic occupation in the hinterlands of the Felsenthal region. An overview study 

conducted by Timothy C. Klinger (1979) involved reconnaissance and informant interview 

survey in the 16,194 ha Hampton Prospect. This resulted in the identification and recording 

of a number of prehistoric sites in spite of the difficult survey conditions encountered in the 

heavily wooded project area. 

Phase II assessment at Hampton (Lafferty et al. 1981) involved an intensive survey of 

10% of the redefined (and renamed) 14,175 ha Sparta Mine area. As a cumulative result of 

these efforts, 27 prehistoric sites have been identified in the project area. Although most of 

these sites were discovered by systematic shovel testing in wooded areas and yielded only 

very small samples of cultural material, they represent Archaic, Poverty Point, Tchula or 

Marksville, and Mississippi period occupations. Further analysis of the Sparta Mine data 

involved examination of the relationship between prehistoric site location and selected biophysical 

environmental variables in the project area and formulation of a predictive model 

of site location (to be tested in future work in the Sparta Mine area). 

Recent investigation in the Saline River watershed portion of the Felsenthal region 

has been limited to reconnaissance, testing, and small scale salvage excavation by members 

of the Arkansas Archeological Society and the Survey research station at Monticello and 

several small cultural resource management surveys. Published accounts include White’s 

(1970) report on salvage excavations at a very late prehistoric cemetery at Gee’s Landing 

(3DR17), one of the sites investigated by Moore in 1912-1913, and Reaves and McClurkan’s 

(1970) brief note on salvage work at the Dansbury Creek site (3CV24) near Rison, which also 

represents very late prehistoric or protohistoric occupation. 

the seQuence of AboriginAl occupAtion 

Knowledge of prehistoric and protohistoric cultural development in the Felsenthal archeological 

region is the result of slow accumulation of data from early explorations, reconnaissance, 

and small scale excavation by professional and amateur archeologists, and recent 

cultural resource management efforts. The following chronological survey of this topic relies 

heavily on a summary by Schambach in the Hampton Prospect (Sparta Mine) overview 

(Klinger 1979), and discussions by Schambach and Early in the Arkansas State Plan (Davis 

1981).

58 House 

Table 4 presents an outline of the Felsenthal archeological sequence, as it is currently 

known, and aligns it with the sequence of adjoining regions. The column on the left consists 

of archeological periods recognized in the Lower Mississippi Valley. For the latter part of 

the sequence, these periods follow Phillips (1970:Figure 2). The archeological culture units 

under the various regional headings, not in parentheses, are phases (Willey and Phillips 

1957:22). Entities in parentheses are artifact complexes, archeological cultures or, in the case 

of the Great Bend region, periods recognized in the Caddoan subarea. Intervals in regional 

sequences for which occupation is yet poorly defined archeologically are designated “Unnamed.” 

Corresponding intervals for which few or no published data exist are left blank. 

The Middle Ouachita region sequence follows Schambach (1970:Table 27). The Great Bend 

sequence follows Davis (1970) and subsequent discussions by Schambach in the Arkansas 

State Plan (Davis 1981). The Bartholomew sequence follows Rolingson (1973, 1976a). The 

Upper Tensas sequence follows Hally (1967) and Belmont (1979) and the discussion of phase 

distributions in Phillips (1970:861-976). 

Since archeological cultures and archeological periods in the Lower Mississippi Valley 

may have the same name, it will be important in the following discussion to rigorously 

distinguish between the two. Assignment of a given component to a certain archeological 

culture (e.g., Marksville culture) will imply a high degree of formal similarity on the total assemblage 

level to the named archeological culture. Assignment of a component to a certain 

period (e.g., Marksville period), on the other hand, will imply only that the component dates 

to the specified time interval; similarities on the total assemblage level may be slight. 

Paleo-Indian Period, before 8000 B.C. 

The archeological components of the Paleo-Indian period found throughout North 

America are regarded as representing hunter-gatherer populations living in terminal 

Pleistocene or transitional Pleistocene-Holocene environments. The fluted projectile point 

horizon represented by infrequent sites, continentwide, is dated to ca 9,000 B.C. while the 

Dalton culture, largely confined to the Southeast, appears to have flourished immediately after 

the Pleistocene-Holocene transition, ca 8500 B.C. (Goodyear et al. 1979:96-99). Dalton culture 

exhibits both continuities and discontinuities with earlier fluted point complexes; some 

investigators, emphasizing the latter, assign Dalton to the Early Archaic period (e.g., Goodyear 

et al. 1979:96). Scottsbluff and related projectile point forms may be roughly coeval with 

Dalton but have contrasting biogeographic distribution. The extent of Paleo-Indian period 

occupation in many regions is particularly difficult to assess because of alluvial burial of 

sites in riverine zones.

Prehistory 59 

Table 4. Comparison of Archeological Sequences in the Felsenthal and Adjacent Archeological 

Regions.

60 House 

Schambach (1979:24) notes that fluted points resembling the Clovis type have been 

recorded in small numbers from several counties in southeast Arkansas. The one specimen 

on record from the Felsenthal region is in an amateur’s collection from the Shallow Lake site 

(3UN9/52) on the Pleistocene Deweyville Terrace surface. 

Scottsbluff and Dalton points are comparatively frequent in southwest Arkansas. 

Schambach (1979:24) reports a Scottsbluff point from site 3OU148 on the Ouachita River 

across from the mouth of Champagnolle Creek. Within the Felsenthal National Wildlife Refuge 

area, a buried Dalton culture component appears present at Coon Island (3BR10) on the 

Ouachita River floodplain (Schambach 1979:26). A private collection from the Shallow Lake 

site includes six Dalton points (Rolingson and Schambach 1981:92). Rolingson and Schambach 

(1981:178) note that Dalton points in the Felsenthal region have been found as much as 

a meter deep in Pleistocene terrace deposits. 

The Snow Hill complex is the name given by Schambach (1979:24) to a reported—but 

unverified—discovery by a collector of apparent Paleo-Indian period artifacts at five sites 

near Smackover in Union County, Arkansas. The artifacts consist of some 50 Angostura-like 

points, termed “Snow Hill points” by Schambach, plus choppers, uniface scrapers, gravers, 

denticulates and polished magnetite “bola stones.” Although the specimens do appear to be 

authentic artifacts, they were excavated and collected in secrecy. No similar assemblage has 

ever been identified at any other site in southwest Arkansas. 

Early Archaic Period, 8000-6000 B.C. 

The Early Archaic period, together with the earlier Dalton culture horizon, is regarded 

as the period of initial adaptation to the varied postglacial environments of eastern North 

America. Dalton and Early Archaic sites are far more frequent than earlier fluted point sites 

and settlement pattern shifts are indicated; this period often represents the earliest occupation 

in many caves and alluvial sites with subsequent long prehistoric sequences. Technological 

discontinuities with the earlier fluted point complexes appear to become more 

pronounced through the time span of the Early Archaic period (Griffin 1967:178; Morse 

1973; Goodyear et al. 1979:97-99). 

The little known San Patrice culture (Webb et al. 1971) is localized in north Louisiana, 

south Arkansas, and east Texas. Its relationship to the Dalton culture is yet undetermined 

but Schambach (1979:26) suggests that it may locally represent the beginnings of “regional 

differentiation” of Southeastern Archaic cultures in post-Dalton times. San Patrice projectile 

points (Suhm and Jelks 1962:242), side notched points, and “Albany scrapers,” all attributes 

of the San Patrice culture, have been found in the Felsenthal National Wildlife Refuge at 

Watts Field (3UN18) and Shallow Lake (3UN9/52).

Middle Archaic Period, 6000-4000 B.C. 

Prehistory 61 

The Middle Archaic period in eastern North America is marked by some technological 

innovations such as stone atlatl weights and the earliest widespread use of ground and polished 

stone tools and by evidence of more intensive and prolonged occupation of individual 

sites with possible increased reliance on aquatic resources (Griffin 1967:178). This interval 

correlates generally with the Atlantic climatic episode (Bryson et al. 1970). Paleoenvironmental 

data suggest dramatic shifts toward drier and warmer climate in nearby portions of 

the East (Wood and McMillan 1976; King and Allen 1977). Response by human populations 

to this climatic change in some regions may have included dramatic demographic shifts 

(Morse 1975:191). 

Investigation in southwest Arkansas has resulted in recognition of Tom’s Brook culture. 

Diagnostic attributes of this culture, first defined as a phase by Schambach (1970:384- 

385) at the Cooper site in the Middle Ouachita region, include Johnson points (Scholtz 1967; 

Bartlett 1963), stemmed scrapers, and side-notched pebble “net sinkers.” Subsequent excavations 

at the Paw Paw site (3OU22) on the northern fringe of the Felsenthal region revealed a 

1 m thick, deeply buried Tom’s Brook culture midden and the earliest human burials identified 

to date in south Arkansas. These excavations also yielded a radiocarbon date of 4690 ± 

70 B.C. (TX-1550) for the Tom’s Brook culture (Weber excavation and analysis notes, 1973, 

Arkansas Archeological Survey). 

Tom’s Brook culture components in the Felsenthal region have been grouped into 

the Spoon Bend phase (Rolingson and Schambach 1981:175). A deeply buried Tom’s Brook 

culture component has been identified at the Short Brake site (3BR20) in the Ouachita River 

floodplain (Schambach 1979:26). 

The Tom’s Brook culture is considered by Schambach (1979:26-27) to be the earliest in 

this portion of the Southeast to show a strong riverine orientation. An upland component 

in the Felsenthal region has been identified at 3UN36. This upland site assemblage lacks net 

sinkers but includes grinding equipment and uniface scrapers. Components of the Tom’s 

Brook culture are widespread, occurring throughout the Ouachita River watershed in both 

the West Gulf Coastal Plain and Ouachita Mountains, the Arkansas Valley and the Ozarks 

(Bartlett 1963) and perhaps in the Arkansas River Lowland in the Lower Mississippi Alluvial 

Valley as well (House 1980). 

Side-notched Big Sandy projectile points (Lewis and Lewis 1961:34, 37), diagnostic of 

Schambach’s (1970:385-386) Crystal Mountain phase, have been identified in private collections 

from the Felsenthal region but no sites yielding these points in abundance or in good 

stratigraphic context have yet been discovered.

62 House 

In the Southeast there is considerable confusion about the correct temporal placement 

of the side-notched point horizon on horizons (cf. Fowler 1959:Figure 14; Dejarnette et al. 

1962; Bartlett 1963:Table 2; Schambach 1970:387; Goodyear et al. 1979:100-105) but a general 

Middle Archaic position seems likely. Schambach (1970:386) tentatively associates a bone atlatl 

hook and a full-grooved axe with the Crystal Mountain phase component at the Cooper 

site in the Middle Ouachita region. 

Late Archaic Period, 4000-1500 B.C. 

In much of the East it has proved useful to divide the 4000-500 B.C. interval, generally 

known as the Late Archaic, into an earlier and later segment. In the Lower Mississippi Valley 

the earlier segment, 4000-1200 B.C., is the Late Archaic proper while the latter segment 

is the Poverty Point period. The Late Archaic period throughout eastern North America is 

seen as a time of considerable human population growth, distinct regional adaptations, and 

interregional exchange of raw materials (Griffin 1967:178). Remains of both tropical and native 

North American cultigens have been recovered from some eastern North American sites 

(Chomko and Crawford 1978) as early as 2300 B.C. (Marquardt and Watson 1976). Recent 

data from Mississippi (Connaway et al. 1977) demonstrate that the beginnings of an elaborate 

lapidary industry in the Lower Valley dating to the 3000 B.C. time level or earlier. 

The White Oak phase in the Middle Ouachita region was defined by Schambach 

(1970:388) on the basis of the compact horizontal distribution of corner-notched Williams 

points (Suhm and Jelks 1962:259) at the Cooper site. Stratigraphic data support a ca 2000 

B.C. placement for this component (Schambach 1970:Table 27). This horizon is poorly defined 

but present throughout southwest Arkansas, including the Felsenthal region. Schambach 

(1970:388) suggests that Williams projectile points may be largely confined to upland 

areas in this part of the Southeast. Williams-like Big Creek points (Morse 1970), however, are 

markers of the Frierson phase in northeast Arkansas which can probably be assigned to the 

3000-1500 B.C. interval (Morse 1975:192). 

Schambach (in Davis 1981:57-59) defines the archeological Big Creek culture which 

subsumes the previously defined Dorcheat phase (Schambach 1970:389-390) of the Middle 

Ouachita region. Markers of the Big Creek culture are Bulverde (Suhm and Jelks 1962:404) 

and Evans (Ford and Webb 1956:64-65) projectile points. Schambach (1970:390) attributes a 

series of polished red slate items to the Dorcheat phase component at Cooper. The type site 

of the Big Creek culture is the Big Creek site near Rison, Arkansas on the northern fringe 

of the Felsenthal region. A single amateur collection from this site contains over 500 Evans 

points. Rolingson and Schambach (1981:179) note that the Big Creek culture is well represented 

in the Felsenthal region.

Poverty Point Period, 1200-500 B.C. 

Prehistory 63 

The interval 1200-500 B.C. brackets the major portion of the development of the 

Poverty Point culture (Webb 1977:Table 1). This archeological culture, recognized at sites 

throughout the Lower Mississippi Valley below the latitude of Memphis plus adjoining 

upland areas and Gulf Coast areas, is characterized by diagnostic forms of baked clay balls 

or “Poverty Point objects,” elaborate lapidary items, a microlith industry, far-flung interregional 

exchange in lithic raw materials, and—at the Poverty Point type site and some other 

sites—construction of mounds and earthworks (Webb 1968, 1977). This archeological culture 

is generally conceded to represent the most complex cultural development in North America 

on this time level. Its economic base is a topic of considerable interest and debate. The 

few data available at this time suggest that a continuation of the Archaic pattern of harvesting 

nondomesticated faunal and floral resources formed the basis of Poverty Point Culture 

subsistence (Webb 1977:3-4; Byrd and Neuman 1978). 

Several sites in and near the Felsenthal region have yielded such Poverty Point culture 

diagnostic artifacts as baked clay balls in several forms, hematite and magnetite plummets, 

steatite vessel fragments, microliths, and greenstone celts. These sites include Woodward 

Lake (3OU118), Calion (3UN51), Green Island (3BR1), Stringfellow (3CA15, Hoover Levee 

(3UN90), Bangs Slough (3CA3), and Coon Island (3BR10) in Arkansas and Monroe and 

Pargoud (16OU1) in Louisiana (Webb 1977:Figure 4; Schambach in Davis 1981). Schambach 

further notes that these diagnostic artifacts occur in sufficient number and variety that these 

components may readily be assigned to Poverty Point culture as well as to the Poverty Point 

period. This set of Felsenthal components has been assigned to the Calion phase by Rolingson 

and Schambach (1981:Table 19). 

Within the Felsenthal region Schambach (1979:27) notes that Poverty Point culture sites 

recorded to date tend to cluster in the Calion-Champagnolle-Moro Creek locality with only 

scattered occurrences of diagnostic artifacts in the Felsenthal National Wildlife Refuge area 

to the south; this record persists in spite of extensive survey work in the latter area (but see 

Chapters 5 and 6 of this report). Many of the sites in the Calion locality have yielded large 

quantities of unfinished magnetite plummets, worked magnetite debris and small unaltered 

pieces of magnetite. Schambach (1979:27-28) proposes that a yet-undiscovered natural occurrence 

of magnetite in gravel form in the uplands of the Champagnolle and Moro Creek 

watersheds may have been the resource that attracted Poverty Point culture people to the 

region. 

Large contracting stemmed points corresponding to Schambach’s (1970:188-191) Gary, 

variety Gary, may be the predominant point form associated with Poverty Point period components 

in the Felsenthal region. Webb (1977:37-38) notes that this form predominates at the

64 House 

Poverty Point site. In the Arkansas River Lowland to the east, use of Arkansas novaculite as 

a raw material for flaked stone tools appears to peak at this time (House 1980; cf. Giardino 

1979). 

During the course of the present investigation, testing at Marie Saline (3AS329) revealed 

an apparent Poverty Point culture component at a depth of about 1.1 m below surface. 

Recovered specimens included steatite sherds, muscovite schist, novaculite flakes, 

and fire-cracked rock (Chapter 6). The survey results also indicated a decided emphasis on 

novaculite versus chert on this time level in the project area (see Chapter 8). 

Tchula Period, 500 B.C.-A.D. 1 

Manufacture of ceramics became widespread in North America between 500 B.C. 

and A.D. 1. In the Lower Mississippi Valley this interval, known as the Tchula period, is 

characterized by the development of the Tchefuncte archeological culture in the south and 

the Lake Cormorant and possibly other cultures in the north (Phillips 1970:16). Data from 

throughout eastern North America establishes the early cultivation of certain tropical and 

native North American cultigens at this time. Of particular interest is the recovery of remains 

of squash and bottle gourd in Tchefuncte levels at Morton Shell Mound in Coastal 

Louisiana (Byrd and Neuman 1978:11). 

Attributes of the Tchefuncte Culture (Ford and Quimby 1945) include grog-tempered 

or “untempered” pottery (types Tchefuncte Plain, Tchefuncte Stamped, Tchefuncte Incised, 

Lake Borgne Incised), tetrapodal bases, and possibly construction of conical burial mounds. 

Sand-tempered Alexander series ceramics also occur in some Tchefuncte culture components 

(Gibson 1966; Phillips 1970:881). The Lake Cormorant culture in the northern Lower 

Mississippi Valley is equivalent to Phillips, Ford and Griffin’s (1951:432-433) Northern Tchula 

period complex which exhibits strong formal similarities to Early Woodland complexes 

in the Midwest. The nature of Tchula period occupation in the northern Lower Mississippi 

Valley west of the Mississippi is yet poorly understood. The few Tchula period components 

recognized in the Bartholomew-Macon region appear to exhibit closer similarities to the 

Tchefuncte culture to the south than to Lake Cormorant culture to the north (cf. Rolingson 

1981; Jeter 1981). 

One of the surprises of the recent survey and testing work in the Felsenthal National 

Wildlife Refuge has been the recognition of an apparent Tchefuncte culture component at 

Coon Island (3BR10). Artifacts include distinctive ceramic types Tchefuncte Incised, Lake 

Borgne Incised, Tchefuncte Stamped, and Tchefuncte Plain, and characteristic tetrapodal 

bases. Four other sites 3UN9/52, 3UN63, 3UN75, and 3UN86 have also yielded small numbers 

of Tchefuncte sherds (Rolingson 1972:3; Schambach 1979:29). These components have 

been grouped by Schambach (1979) into the Coon Island phase.

Prehistory 65 

No data pertaining to Tchula period occupation along the Ouachita River in the Monroe, 

Louisiana locality have been published. Moore (1909:Figure 4) illustrates a basal fragment 

of an apparent Tchefuncte vessel from Booth Landing on the lower reaches of the 

Ouachita in Catahoula Parish. 

During the course of the present survey, Tchula period components were identified at 

Marie Saline (3AS329) and False Indigo (3AS285) and minor components were identified at 

three additional sites. Ceramic variability among the components suggests that the Coon 

Island phase in the Felsenthal region may eventually be divided into earlier (500-250 B.C.) 

and later (250 B.C.-A.D. 1) subphases (see Chapter 8). 

Marksville Period A.D. 1-300 

The concept of Marksville culture and Marksville period have been important in 

Lower Mississippi Valley archeology since Setzler (1933, cited in Toth 1974:21) recognized 

the strong connection to Midwestern Hopewell in ceramic vessels recovered from the 

Marksville site in Avoyelles Parish, Louisiana. Marksville culture attributes include distinctive 

incised and stamped ceramics, mound and earthwork construction, and, in some cases, 

mortuary ceremonialism involving interment of large numbers of individuals on earthen 

platforms within mounds (Fowke 1928:420-421, cited in Toth 1974:18; Ford and Willey 1940). 

A chronological subdivision of the Marksville period into early or “classic” and late subperiods 

has been accomplished in some regions (Greengo 1964; Phillips 1970:10-11; Toth 

1974:101-131). Data from this interval in the Midwest indicate an economy based on intensive 

harvesting of nondomesticated resources, particularly fish, nuts, and starchy seeds of 

annual weeds, plus cultivation of native North American cultigens and limited cultivation of 

tropical cultigens (Asch et al. 1979). In the Lower Mississippi Valley, however, little progress 

has been achieved toward reconstruction of subsistence, settlement systems, or population 

biology during the Marksville period. 

In a 1939 article, Dickinson and Lemley recognized ceramics of the “Marksville complex” 

(cf. Ford 1936) at the Kirkham site (3CL29) in the Middle Ouachita region of Arkansas. 

Schambach (1970:396-399) notes the occurrence of Marksville types in the Lost Bayou Phase 

of this region. Hoffman (1970:151-153) reported additional data on Marksville period occupation 

in southwest Arkansas defining the Bellvue phase of Marksville culture. Schambach 

(in Davis 1981:88-89), subsequently has taken issue with Hoffman’s identification of 

Marksville culture, emphasizing the environmental contrasts between the Lower Mississippi 

Valley and Trans-Mississippi South and the fact that, in most regions in the latter area, 

Marksville attributes occur in only very low frequencies in assemblages that reflect Fourche 

Maline culture continuities. According to Schambach, this lack of comparability on the total 

assemblage level implies major systemic differences in culture and adaptation.

66 House 

The question of the presence of Marksville culture in the Felsenthal region, however, 

remains open. Marksville sherds have been found at sites throughout the Felsenthal region 

but no pure Marksville period components have been isolated (Schambach 1979:29). It is 

interesting to note that both ceramic and radiometric dating suggest placement for this 

occupation on a late Marksville period time level. Stratum 6 at Paw Paw (3OU22) yielded 

sherds tentatively classified as Marksville Incised, var. Yokena, and Charupa Punctated (Phillips 

1970), and uncorrected radiocarbon dates of A.D. 460 ± 40 (TX 1547) and A.D. 660 ± 60 

(TX-1552) (Weber Excavation and analysis notes, Arkansas Archeological Survey 1973). The 

Shallow Lake (3UN9/52) excavations likewise yielded sherds resembling Marksville Incised, 

var. Yokena and Goose Lake (Rolingson and Schambach 1981:181). 

Baytown Period, A.D. 300-700 

Phillips (1970:18) defines the Baytown culture period as the interval between the end 

of Hopewellian and emergence of Coles Creek culture. In the Lower Mississippi Valley, this 

interval is occupied by the archeological Baytown culture, a concept which—along with 

the Baytown period—is defined somewhat negatively and has proven difficult to characterize. 

Baytown period phases in portions of the Lower Mississippi Valley near the Felsenthal 

region (e.g., Dry Bayou [Rolingson 1973], Marsden [Phillips 1970:908] and Troyville [Phillips 

1970:908-910; cf. Walker 1936; Ford 1936, 1951]) are characterized by a predominance of Baytown 

Plain ceramics with consistent minorities of Mulberry Creek Cordmarked and painted, 

incised, and brushed ceramic types. Belmont and Williams (1981) suggest that horizon styles 

in grog-tempered painted ceramics may provide improved chronological controls within 

the Baytown period. Absolute dates for Baytown culture phases and comparable units in 

the Lower Valley cluster in the A.D. 600s (Phillips 1970:Table 18; Morse et al. 1980). The few 

available subsistence data for the Baytown period in the Lower Mississippi Valley indicate 

continued reliance on nondomesticated food resources with perhaps supplemental cultivation 

of native North American and tropical cultigens (Byrd and Neuman 1978). 

Contemporary components in the Middle Ouachita region are grouped into the Oak 

Grove phase of Fourche Maline culture (Schambach 1970:399-400). Like Baytown assemblages, 

Fourche Maline culture ceramic assemblages at this time level are dominated by grogtempered 

plainware (Williams Plain). 

The extent of Baytown period occupation in the Felsenthal region has proved difficult 

to determine—for the same reason that Baytown culture has proved a difficult concept to 

apply in the Lower Mississippi Valley. Schambach (in Davis 1981:91) notes that these components 

with plainware in the Felsenthal region can present a problem of distinguishing 

between Baytown and Fourche Maline cultures. He outlines two distinctions between the 

Baytown and Fourche Maline archeological cultures which should prove useful: (1) Fourche 

Maline ceramic assemblages are dominated by flat-based jar forms, whereas Baytown

Prehistory 67 

assemblages are characterized by high proportions of round-based bowls, and (2) Gary 

points (small varieties) are very abundant in Fourche Maline assemblages but infrequent in 

Baytown assemblages. In this report the concept of Baytown culture has proved most useful 

(Chapters 5, 6, 8). 

Coles Creek Period, A.D. 700-1100 

As Jeter (1981) observes, the Coles Creek period continues to be a “poorly defined concept 

applied to a number of poorly dated sites” (cf. Greengo 1964:14-15; Phillips 1970: 958; 

Schambach in Davis 1981:92). The Coles Creek period is characterized by Phillips (1970:20) 

as the interval of Coles Creek culture in the southern Lower Mississippi Valley. 

The Coles Creek archeological culture, first recognized as a ceramic complex by Ford 

(1935, 1936) has been widely used as a culture-level concept but is currently undergoing a 

somewhat painful process of redefinition (e.g., Rolingson 1979; Belmont 1979) and is only 

now beginning to be dated adequately (Phillips 1970:Table 13; Rolingson in Davis 1981:21; 

Schambach in Davis 1981:76; House and Jeter 1981). 

Evidence of Coles Creek period occupation is practically ubiquitous throughout the 

Lower Valley and adjacent south Arkansas, possibly indicative of increased populations 

over earlier periods. In the Lower Mississippi Valley, this interval is known as the Early 

Mississippi period (e.g., Morse 1980a) and is characterized not only by the appearance of 

shell-tempered ceramics but also by the earliest substantial occurrence of maize in archeological 

sites (Morse 1980b). In the southern Lower Mississippi Valley, ceramics continue 

to be overwhelmingly grog-tempered with characteristic types of incised and punctated 

decoration on rims. Corner-notched and stemmed arrow points are associated with Coles 

Creek period components in many regions (Rolingson 1979). Although maize agriculture is 

thought to have been important in the southern Lower Valley by this time (Haag 1971:26), 

no direct evidence has been reported. The few available data indicate continuing importance 

of nondomesticated plant foods (Byrd and Neuman 1978). Quite as significant, however, is 

evidence of construction of platform mounds at a few sites in each region (Belmont 1967; 

Rolingson 1979; House and Jeter 1981), implying the development of more hierarchically 

organized settlement systems than characterize the Baytown period. 

In southwest Arkansas, Coles Creek ceramics were first recognized by Dickinson 

and Lemley (1939) at the Kirkham Place (3CL29) in the Middle Ouachita region. Coles Creek 

ceramics were subsequently recognized by Wood (1963) at Crenshaw (3MI6), by Hoffman 

(1970b) in the Millwood Reservoir basin, and by Schambach (1970:401-406) in the Dutchman’s 

Garden phase in the Middle Ouachita region. Hoffman’s (1970a:153-155, 157-158) 

assignment of Coles Creek period components in the Little River region and at Crenshaw 

(3MI6) to Coles Creek culture has been questioned by Schambach (in Davis 1981:92-93) who 

cites major differences on the total assemblage level between Coles Creek period compo-

68 House 

nents in the southern Lower Mississippi Valley and those in the Trans-Mississippi South, 

again with the implication of major systemic differences in culture and adaptation. 

Ongoing investigation has revealed abundant evidence of Coles Creek period occupation 

in the Felsenthal Refuge area (Rolingson 1972; Lischka 1973) and elsewhere in the 

Felsenthal region (White 1970:14; Jeter 1981). Schambach (in Davis 1981:95) cites the near 

ubiquity of Coles Creek Incised ceramics in the region and suggests heavy occupation during 

the Coles Creek period. Schambach (1979:30) also points out that Coles Creek period 

middens are present at all of the major mound groups in the region and that sherd collections 

from mounds damaged by vandals and other destructive agents indicate that some 

mounds in each group were built or begun during the Coles Creek period. 

The taxonomic status of Coles Creek period components in the Felsenthal region remains 

to be addressed by isolation and description of total assemblages. Preliminary indications 

(Schambach in Davis 1981:92-94) are that the ceramics exhibit greater formal similarity 

to the Coles Creek culture in the southern Lower Mississippi Valley than to contemporary 

components of either the Plum Bayou culture of east and central Arkansas (Rolingson 1979) 

or to preCaddoan culture at Crenshaw and other sites in southwest Arkansas. The precise 

temporal placement of Coles Creek period components in the Felsenthal region is yet poorly 

known. Schambach (1979:30) states that the full temporal range of Coles Creek period occupation, 

from early to late, appears to be present. The Coles Creek period component at 

Shallow Lake (Rolingson and Schambach 1981:182-189) appears to include burials beneath 

Mound C. Ceramics from the overlying midden appear to represent Early Coles Creek period 

occupation. 

Mississippi Period, A.D.1100-1700 

The Mississippi period in the southern Lower Valley is generally divided into Early, 

ca A.D. 1100-1400, and Late, ca A.D. 1400-1700, subperiods. The two subperiods are characterized 

by the successive domination of the Plaquemine and Mississippian archeological 

cultures, respectively (Phillips 1970:20). Since the dividing line between the northern and 

southern portions of the Lower Valley is usually placed at about 33° north latitude (i.e., the 

Arkansas-Louisiana state line) the Felsenthal region during the Mississippi period may be 

seen to have been on the fluctuating frontiers of three archeological cultures, Mississippian, 

Plaquemine, and Caddoan (cf. Jeter 1981). 

Like the preceding Coles Creek culture, the Plaquemine culture, defined by Quimby 

(1951), persists in being a rather loose grouping of attributes. The culture is considered as 

“Mississippianized” Coles Creek by Brain (1978:345), who cites the development in some 

regions of impressive site plans characterized by multiple mounds dominated by a single 

mound as much as 20 m in height with distinct plaza areas (cf. Phillips 1970:560). In contrast 

to the ceramics of the northern Lower Valley during this interval, those of the southern 

Lower Valley Plaquemine culture persist in being predominantly grog-tempered (though

Prehistory 69 

perhaps with some technical improvements in paste over preceding Baytown-Coles Creek 

ceramics). Brushed, incised, and punctated decoration is frequently applied to body as well 

as rim portions of vessels. A new major vessel form, the bottle, appears as do additional vessel 

form modes in bowls and jars. Stemmed arrow points, called Ashley points by Rolingson 

(1971), occur in Early Mississippi period components in the Bartholomew-Macon region. 

Intensive agriculture involving both maize and beans as well as squash appears to 

have been widespread in eastern North America by this time (Griffin 1967:189) but no direct 

subsistence data from Plaquemine culture components are yet reported. Absolute dates for 

the Plaquemine culture are yet relatively few (Phillips 1970:Table 18; House and Jeter 1981) 

but progress has been made in some regions in chronologically subdividing the Plaquemine 

culture on the basis of artifactual attributes (Hally 1967; Phillips 1970:558, 560; cf. Jeter 1981). 

Early Mississippian period Plaquemine culture phases in the Lower Valley east of 

the Felsenthal region include the Routh phase of the Upper Tensas region (Hally 1967) and 

the Bartholomew phase of the Bartholomew-Macon region (Rolingson 1976a). To the west, 

the Caddoan culture Haley phase (Caddo II period) (Schambach in Davis 1980:107-108) in 

the Great Bend region is roughly coeval with the Plaquemine culture while in the Middle 

Ouachita region much of the Caddoan culture Mid-Ouachita phase (Hodges and Hodges 

1945) appears to date to this interval (Early in Davis 1981:109-110). 

The first recent investigations in the Felsenthal region (Rolingson 1972; Lischka 1973) 

identified Plaquemine culture occupation at a number of sites. Since the region had previously 

been considered part of the Caddoan archeological culture area, the low frequency of 

Caddoan attributes observed in the recovered ceramic assemblages was somewhat surprising. 

Analysis by Rolingson and Schambach (1981) of material from Stacy’s 1975 excavation 

of a circular structure beneath Mound C at Shallow Lake (3UN9/52) have resulted in definition 

of the Gran Marais phase of the Early Mississippi period, culture unknown. Ceramics 

include Baytown Plain, var. Shallow Lake, Harrison Bayou Incised, Kiam Incised, var. Redeye 

Lake and Pargoud Incised, var. Monroe. Both a radiocarbon date and an archeomagnetic 

date place this occupation in the A.D. 1250-1300 range. This date is supported by an archeomagnetic 

date of A.D. 1350 for Watts Field (3UN18) Mound D. Analysis of faunal and 

floral specimens from the Coles Creek period and Gran Marais phase midden at Shallow 

Lake Mound C indicated exploitation of both terrestrial and aquatic habitats in the locality 

and probable multiseasonal occupation. Surprisingly, in light of the date of the context, no 

remains of cultigens were recovered (Rolingson 1981). 

Gran Marais phase components appear to be present at every one of the large mound 

groups in the Felsenthal region. Figure 14 presents the plan of one of these mound groups,

70 House 

Figure 14. The Mississippi period mound and midden complex at Watts Field 

(3UN18) (after Lischka 1973 and Arkansas Archeological Survey notes on 

file).

Prehistory 71 

Watts Field (3UN18). These components, however, appear primarily to represent ceremonial 

and mortuary behavior rather than intensive habitation, leading Schambach (1979:30) 

to propose that the Early Mississippi period settlement pattern in the Felsenthal region is 

characterized by sites of highly dispersed hamlets and farmsteads. 

During the course of the present investigation, Gran Marais phase components were 

identified at False Indigo (3AS285) and Jug Point Cutoff (3BR76). Application of Schambach’s 

analytical classification system (Rolingson and Schambach 1981:101-176) to the large 

ceramic sample from False Indigo Area A revealed the sample to be strikingly similar to the 

Gran Marais phase assemblage from Shallow Lake (3UN9/52) Mound C. The contrasting 

higher percentage of shell tempering in the ceramic sample from Jug Point Cutoff suggests 

placement of this component later in the span of the Gran Marais phase (Chapter 8). In 

Chapter 9 of this report, Hemmings presents a model of human adaptation to the Ouachita 

River floodplain by Gran Marais phase people, incorporating information on the structure 

of the environment and data from the present and previous investigations in the Felsenthal 

region. 

On the southern fringes of the Felsenthal region, recent excavations in Area I of the 

badly-damaged T. E. Salsbury site (16OU15) in Ouachita Parish, Louisiana (Price and Heartfield 

1977) yielded ceramics resembling those of the (subsequently-defined) Gran Marais 

phase in association with aboriginal burials and a village midden. A small amount of Caddoan 

ceramics were also recovered in the cemetery and midden area. Subsequent radiocarbon 

dates (TX-2839 through TX-2846) indicate this occupation to date between A.D. 1200 

and 1500 (U.S. Army Corps of Engineers 1980:5:7). An important Plaquemine time-level 

component also appears to be present at the Pargoud Landing site (16OU1) near Monroe, 

though no report on recent work at that site is yet available (U.S. Army Corps of Engineers 

1980:5:6). 

The Late Mississippi subperiod (A.D. 1400-1700) is characterized by the domination of 

Mississippian culture throughout much of the southern Lower Valley as well as in the north. 

In this sense, Mississippian culture may (with some simplification) be equated with shelltempered 

pottery (cf. Holmes 1903). Alternative definitions of “Mississippian” as a type of 

adaptation (Griffin 1967:189; Smith 1978:486), however, have also been proposed. 

Frequencies of shell-tempered pottery appear to increase gradually through time in 

some regions (Hally 1967) and more abruptly in others (Phillips 1970:501-567). In much of 

the Lower Valley east of the Felsenthal region, the late shell-tempered complexes equate 

with the “Tunican” complex identified by Ford (1936:98-106) in the Lower Yazoo basin (cf. 

Lemley and Dickinson 1936). Important decorated types on this horizon include Winterville 

Incised, Leland Incised, and Parkin Punctated. In many Lower Valley sites on this time level, 

Caddoan ceramics—or at least polished and engraved types whose distribution extends into 

the Caddoan area—are part of predominantly shell-tempered “Mississippian” assemblages.

72 House 

Willow Leaf or Nodena and triangular arrowpoints tend to replace stemmed forms in many 

regions by this time (Phillips et al. 1951:450). 

The Late Mississippi period lifeway in the Lower Valley appears to have focused on 

intensive agriculture involving maize, beans, and squash but considerable utilization of nondomesticated 

floral and faunal resources persisted (Byrd and Neuman 1978). Dramatic settlement 

pattern and demographic shifts are perceived in some regions after A.D. 1400 (Brain 

1978; Morse 1981; House 1981) but their nature is not yet understood. The Late Mississippi 

period in the Lower Valley extends through the 1541-1543 De Soto expedition dateline to the 

time of Early French contact, ca A.D. 1700. This contact had drastic and wide-ranging effects 

on aboriginal populations and cultural development in this portion of the Southeast but 

documentation derived from these episodes (Swanton 1911, 1942, 1946) provides ethnohistoric 

data pertinent to many phenomena observed archeologically in prehistory. There are 

also many sites of aboriginal occupation, particularly in the southern Lower Valley, at which 

contact is attested to by the occurrence of objects of European manufacture in metal and 

glass. 

Prehistoric Late Mississippi period phases in the Lower Mississippi Valley east of the 

Felsenthal region include the Wilmot (Rolingson 1973) and Bellaire (Phillips 1970:944; Jeter 

1981) phases in the Bartholomew-Macon region of Arkansas and the subsequent Fitzhugh 

and Transylvania phases of the Upper Tensas Basin in Louisiana (Rally 1967; cf. Phillips 

1970:945). In the latter region, Hally emphasizes the essential continuity of Plaquemine culture 

despite the increasing frequency of shell-tempered pottery. To the west, the Texarkana 

(Krieger 1946) and Belcher (Webb 1959) phases of the Caddo III period on the Red River are 

roughly contemporary units. 

Defined phases which include European trade goods and clearly date to the Contact 

Historic horizon include the Quapaw (Phillips 1970:443-444; Hoffman 1977; House 

and McKelway 1980) and Tensas (Phillips 1970: 945) phases. On the southern fringe of the 

Felsenthal region, in Ouachita Parish, Louisiana, Moore (1909:27-80, 120-151) recovered 

abundant European trade goods including glass beads and discs and cones of sheet brass 

from aboriginal burials in association with elaborate painted, polished, engraved, and incised 

ceramics at Glendora and Keno. Only a single Caddoan site in the Great Bend region, 

Rosenbrough Lake in Bowie County, Texas, has produced early European trade goods but 

contemporary occupations are seen at a number of other sites in the region (Schambach in 

Davis 1981:124). 

On the basis of the findings of two amateurs’ 1965 salvage excavation of an aboriginal 

cemetery at Shallow Lake (3UN9/52), Schambach and Rolingson (1981:193-198) have defined 

the Caney Bayou phase of the Late Mississippi period, culture unknown. The cemetery 

yielded incomplete bundle burials accompanied by both shell-tempered and grog-tempered 

vessels in the types Baytown Plain, var. Shallow Lake, Coleman Incised, Cowhide Stamped,

Prehistory 73 

Glassell Engraved, Parkin Punctated, and Winterville Incised. Additional grave goods 

include discoidals and a cannel coal pendant. Three ceramic elbow pipes were found in the 

cemetery area but not in direct grave association. No European trade goods were recovered. 

This occupation is placed in the A.D. 1500-1700 range. 

Schambach (1979:30) notes that other late prehistoric pottery vessels have reportedly 

been dug by collectors from shallow intrusive graves in the tops of mounds in the Felsenthal 

region. Schambach and Rolingson (1981:201) propose that the apparent absence of large sites 

dating to this time level is due to a post-De Soto depopulation or to a highly dispersed settlement 

pattern. They suggest that the Caney Bayou phase population may have visited the 

preexisting mound centers not to build or maintain mounds but to bury their dead in small 

cemeteries, some of which were located on top of mounds. The present survey has identified 

probable Caney Bayou phase components, possibly representing riverine extraction camps, 

at the Jug Point 2 (3AS307) site (see below, Chapter 8). 

Ceramics which can be roughly dated to the Late Mississippi period were recovered 

by Clarence B. Moore at a number of sites in the Felsenthal region. These include Boytt’s 

Field (3UN13), Keller (3CA13), and Kent (3OU6) (Moore 1909:82-89, 91-96, 96-100) and Wire 

Fence Landing on the Saline (Moore 1913:91-92). 

Ceramic vessels recovered from an aboriginal cemetery at Gee’s Landing (3DR17) on 

the Saline River in 1967-68 include shell-tempered and grog-tempered vessels with incised, 

engraved, and ridge-pinched decorations. Collectively, this series of vessels resembles 

those recovered by Moore at Glendora and Keno and series from contact horizon sites in 

the Lower Mississippi Valley. No European trade goods, however, were recovered at Gee’s 

Landing (3DR17) (White 1970). Another component on the northern fringes of the Felsenthal 

region in the Saline River watershed which dates to this time level was recognized at Dansbury 

Creek (3CV24) in Cleveland County, Arkansas (Reaves and McClurkan 1970). 

In the Felsenthal Navigation Pool area investigated by the present survey, an additional 

very Late Mississippi period complex, called “terminal Mississippi” in this report, has 

been identified at One Cypress Point (3AS286) on the Saline River. The lithic assemblage recovered 

at this site (see Chapter 6, below) includes Nodena or Willow Leaf points and small 

hafted end scrapers, both attributes of the Quapaw phase on the Lower Arkansas (Ford 

1961:156-158; House and McKelway 1980; cf. Phillips 1970:945; Williams 1980). 

No ethnohistoric sources from the sixteenth and seventeenth centuries appear to refer 

directly to the Felsenthal region. Swanton (1946:55) suggests that the aboriginal town of 

Utiangue, where the De Soto expedition wintered in 1541-42, was on the Ouachita River 

near Camden or Calion but it has generally proved impossible to positively identify any 

regions west of the Mississippi mentioned in the De Soto expedition chronicles. In 1700, the

74 House 

Washita, identified by Swanton (1946:204) as a Caddoan group, were encountered on the 

lower Ouachita near present-day Columbia, Louisiana by French explorer Bienville. Bienville 

reported the Washita town at this location to consist of five cabins with about 70 

men. Somewhat confusing late sixteenth century documentary data reviewed by Swanton 

(1911:329-330) place part of the Tunican-speaking Koroa on the west side of the Mississippi 

River, perhaps near lower Bayou Bartholomew or the Ouachita River (Swanton 1911:Plate 

I). The possible presence of Tunican-speakers near the Felsenthal region in the sixteenth and 

seventeenth centuries (cf. Rolingson and Schambach 1981:106) is especially interesting in 

light of the marked similarities between Felsenthal region ceramics on this time level and 

contemporary Late Mississippi period ceramics in the Yazoo Basin east of the Mississippi 

River. 

discussion 

The Felsenthal archeological region is the same territory which Schambach in 1971 

described in these terms: 

Few areas of Arkansas are as wild, as rich archeologically and as little studied 

as this portion of the Ouachita Valley. When Clarence B. Moore (1909) traveled 

up the river to the vicinity of Camden in 1908, he found a few ‘plantations’ 

and some wood camps for the steamboat traffic carved out of the dense forest 

along the banks. Since then most of these clearings have reverted to timber and 

for the most part the Ouachita bottoms south of Camden are wilder than they 

were in Moore’s time. (Schambach 1971) 

Since 1971 the Felsenthal region has undergone considerable archeological investigation, 

much of it in conjunction with the development of the Felsenthal National Wildlife 

Refuge and U.S. Army Corps of Engineers navigation channel improvements on the Ouachita 

River. This investigation has revealed a long prehistoric and contact historic aboriginal 

sequence extending from Paleo-Indian through the era of earliest French exploration. One 

of the intriguing results of this investigation are indications that, throughout much of this 

aboriginal sequence, components in the Felsenthal region exhibit much greater similarity to 

cultures in the Lower Mississippi Valley to the east than to cultures in the Caddoan area to 

the west. Indeed, these Felsenthal region occupations appear to exhibit greater similarities 

to cultures of the southern Lower Mississippi Valley (e.g., Tchefuncte, “classic” Coles Creek, 

Plaquemine) than to geographically close cultures of the northern Lower Mississippi Valley. 

In spite of the impressive advances of the last decade, however, the archeological record 

of the Felsenthal region is only beginning to be explored on the regional level. Because 

of present-day land use patterns, the archeological resource base in the region appears to be 

especially well-preserved compared with the resource base in the Lower Mississippi Valley.

Prehistory 75 

Investigation in the Ouachita River floodplain particularly, reveals the presence of alluvially 

buried and stratified sites. Such contexts are of limited occurrence in the Lower Mississippi 

Valley proper. The present survey of the Felsenthal Navigation Pool project area has been 

especially productive of information on buried and stratified floodplain sites and associated 

geomorphological processes (Chapter 7). 

Ongoing prehistoric archeology in the Felsenthal region has been heavily involved 

with taxonomy, on both the artifact and culture unit levels. By explicitly addressing measurement 

of prehistoric social interaction and delineation of prehistoric sociocultural 

boundaries (e.g., Schambach in Davis 1981:88-98), this taxonomic emphasis may be seen 

to transcend traditional Southeastern preoccupations with chronology and culture history. 

Particularly, Schambach’s design-mode based classification of aboriginal ceramics from the 

region (Rolingson and Schambach 1981: 106-176) holds the promise of allowing quantification 

of formal similarity/dissimilarity between components, at least on the intraregional 

level. 

The development of independent chronological controls in the region has only begun 

in recent years but dated components in the region range from Middle Archaic through 

Mississippi in time and from the Camden locality to the Monroe locality in space. Contexts 

suitable for archeomagnetic dating are widespread in the region. Archeomagnetic dating 

of samples from Shallow Lake (3UN9/52) and Watts Field (3UN18) have already provided 

valuable corroboration of radiocarbon dates for the Gran Marais phase of the Early Mississippi 

period (Rolingson and Schambach 1981:52-53). During the present investigation, attempts 

to radiocarbon-date samples recovered from low-lying floodplain sites encountered 

unanticipated technical problems and yielded disappointing results (see Chapter 7, below). 

Biocultural anthropology in the Felsenthal region can be said to have begun with 

Hrdlicka’s (1909) analysis of human skeletal series from Boytt’s Field (3UN13) in Union 

County, Arkansas. Since this time, aboriginal human remains have been unearthed at several 

sites in the region but only the probable Coles Creek period series from beneath Shallow 

Lake Mound C has been analyzed and reported (Rolingson and Schambach 1981: 48, 51-52, 

89). Though a biological picture of past aboriginal populations in the Felsenthal region is yet 

to emerge, these studies demonstrate that a great potential exists. 

The present-day environment in the Felsenthal archeological region, in both the uplands 

and lowlands, presents formidable challenges to intensive and systematic site survey. 

The Sparta Mine (Hampton) project in the West Gulf Coastal Plain uplands (Lafferty et al. 

1981) and the present study in the Grand Marais Lowland (Chapter 5) both represent innovative 

attempts to meet this challenge and both have yielded encouraging results. 

Our current site sample in the Felsenthal region, a cumulative result of both intensive 

and systematic site survey in recent years and traditional reconnaissance and collector leads

76 House 

spanning decades, encompasses a wide range of habitat types and significant functional 

variability. The number and magnitude of prehistoric mound centers in this hitherto obscure 

region of the Southeast is particularly impressive. These mound centers appear in every 

case to have complex histories and perhaps changing functions through time. Though data 

on specific community patterns are yet very limited, there appears to be a shift from more 

nucleated settlements in the Coles Creek period to more dispersed settlement in subsequent 

times. The proposed seasonal occupation and extractive function of late prehistoric sites on 

the floodplain (Rolingson 1972) receives considerable support from the results of the present 

investigation (see Chapter 6 below). 

Recovery and analysis of large assemblages of ecofactual remains is thus far confined 

to Shallow Lake Mound C (Rolingson and Schambach 1981: 55-68) though analysis of a 

small sample of faunal remains from badly-disturbed contexts at the T. E. Salsbury site 

(16OU15) is presented by Price and Heartfield (1977:92-94). Small samples of floral and 

faunal remains recovered by testing at floodplain sites during the present investigation are 

described in Appendix B of this report. Until extensive series of faunal and floral samples 

from sites in a variety of habitats have been analyzed and reported, the most basic questions 

about paleoenvironments and aboriginal economies will remain unanswered. The floral and 

faunal specimens recovered by the present survey are particularly valuable in demonstrating 

that even floodplain extractive sites may have a potential to yield direct subsistence data. 

On the macroregional level, Schambach (Rolingson and Schambach 1931:103-106) has 

addressed the Lower Mississippi Valley affinities of aboriginal components in the Felsenthal 

region in terms of hydrological, topographic, and biogeographic similarities between 

the Felsenthal lowland environment and that in the Lower Mississippi Valley. To bring this 

problem into sharper focus, it is necessary to explicitly acknowledge that the Lower Mississippi 

Valley is not a homogeneous set of regions and to begin comparing the Felsenthal 

region to specific Lower Mississippi Valley regions in terms of resources and their seasonality 

and abundance. In Chapter 2 of this report, Hemmings provides a detailed discussion of 

the Gran Marais Lowland environment in these terms. 

It may ultimately prove especially important that the aboriginal components in the 

Felsenthal region exhibit strong similarities to units in the southern rather than the northern 

Lower Mississippi Valley. The northern/southern dichotomy in the Mississippi Alluvial Valley 

has long been recognized. In a recent review of Lower Mississippi Valley faunal studies, 

Springer (1980) noted that in the southern Lower Mississippi Valley fish tend to exhibit a 

markedly greater importance relative to terrestrial species in comparison to the northern 

Lower Mississippi Valley. Furthermore, this geographic dichotomy in archeologically-recovered 

faunal assemblages appears to override localized habitat variability. Thus, suggests 

Springer, the observed north-south cultural dichotomy in the Lower Mississippi Valley may

Prehistory 77 

have important underlying ecological dimensions. It will be interesting to see what faunal 

resource exploitation patterns ultimately emerge from investigation of sites in the Felsenthal 

region. This question is obviously of direct relevance to socioculture, anthropology’s current 

concern with the ecological basis of ethnicity, and the adaptive significance of cultural variability 

(cf. Barth 1969; Smith 1978; Hardesty 1980:162-163).

Chapter 4 

history of the felsenthAl region 

by Beverly Watkins 


This chapter examines the history of south-central Arkansas along the Ouachita River. 

Focusing on the importance of the river itself to the history of the region, this study follows 

the changing economy of the area from hunting, to cotton, to lumbering, and traces fluctuations 

in river commerce as a result of these changes. 

Two archeological sites that are closely tied to these themes are discussed in some detail. 

Marie Saline Landing was established as a cotton shipping point and died when lumber 

replaced cotton as the area’s most important product. The Lotawanna, which burned and 

sank in the Felsenthal Project area, was the largest sternwheel steamboat on the Ouachita 

River in the 1870s, and carried both passengers and freight for the Blanks Steamboat Line. 

These two sites are representative of the area in the second half of the nineteenth century at 

the peak of its cotton economy. 

europeAn explorAtion And settleMent 

The Ouachita River Valley was visited by Europeans on their earliest explorations west 

of the Mississippi River. Traversed by the De Soto expedition in 1541-1542, the next recorded 

visits were those of Henri de Tonti in 1689 and the Sieur De Bienville in 1700. The first European 

settlements on the Ouachita River were started about 1718 near Monroe, Louisiana, but 

the river valley in Arkansas was known only to hunters and trappers (Mitchell and Calhoun 

1937:291-293). 

The European population grew, and a census was ordered soon after Spain took possession 

of Louisiana under the Treaty of 1762, which ended the French and Indian War. The 

count was made in 1769. At that time there were 110 white persons living in the Washita 

district of south Arkansas and northeast Louisiana (Mitchell and Calhoun 1937: 294). This 

figure may have included the families that founded Longview on the Saline River at about 

this same date (Goodspeed 1890:876). The area came under the supervision of the Commandant 

of the Natchitoches District, Athanase de Mezieres, who was ordered to clear the 

Ouachita River of the vagabonds (wanderers) living there. He accomplished this in 1774 

with the help of the Caddo Indians, and reported the removal of at least 14 people, including 

Pedro Champagnolle and the Le Boeuf family (Bolton 1914:114).

80 Watkins 

Throughout the 1770s Spain was concerned by the growth of American settlements 

along the Ohio River, and in Kentucky and Tennessee, in defiance of the Proclamation of 

1763 restricting English settlements to the lands east of the Appalachian Mountains. When 

the United States defeated England in the War of the American Revolution and was given 

title to the lands east of the Mississippi River, Esteban Miro, the new governor of Louisiana, 

decided that the time had come to establish a post and settlements along the Ouachita River 

as a foil to any further expansion the Americans might have in mind. He commissioned Jean 

Baptiste Filhoil as the first commandant of the Ouachita District, and sent him up the river 

to gather the European population into settlements. Filhoil established his first post at Ecore 

a Fabre (Camden, Arkansas), but moved two years later to Prairie des Canots (now Monroe, 

Louisiana) where he was granted a large tract of land (Mitchell and Calhoun 1937:295; 

Sprague 1974:194-213). The population of the area grew so that by 1790 there were 242 persons 

(including 19 slaves) in the Ouachita District, more than double the 1770 count (Greene 

et al. 1975:7-13). 

Louisiana was returned to France in 1800, but through an amicable arrangement the 

Spanish officials continued to administer the affairs of the province. By early 1803 the French 

colonial administration in New Orleans had become interested in the Ouachita District. 

Charles F. Adrien le Paulinier, the Chevalier D’Anemours, former Consul General of France 

to the United States, who had settled at Ouachita Post in 1796, submitted a lengthy report on 

the area. He described the rivers, land, trees, and minerals of the District. In listing agricultural 

products he stated that the residents grew corn, sweet potatoes, Irish potatoes, peas, 

and other vegetables for their own use, and tobacco for trade. Hunters brought in the skins 

of deer, bear, beaver, and otter, and bear grease and tallow. D’Anemours did not say how 

many farmers there were, but that there were about 60 white hunters and 600 to 800 Indian 

hunters living in the Ouachita District (D’Anemours 1803:26, 34). 

AMericAn explorAtion And settleMent 

After the United States bought Louisiana from France in 1803, a number of expeditions 

were organized to explore the new territory. One of these, led by William Dunbar and 

George Hunter, investigated the Ouachita River as far north as Hot Springs to learn something 

of the interior of the country between the Mississippi River and the Red River. Dunbar, 

a well known naturalist from Natchez, and Hunter, a chemist and astronomer, were well 

trained to make observations on the plants, animals, and geology of the area as well as to analyze 

springs and salt licks and to record locations so that a map of the new territory could 

be drawn. 

The expedition left Fort Miro (Monroe) late on the afternoon of November 11, 1804, 

and entered the Felsenthal Project area four days later. Their guide was a man who had 

lived in the area for over ten years and who had traveled to the hot springs several times. On

History 81 

November 15 they passed the entrance to Grand Marais and camped just below Lapile 

Creek. The next day they passed Marais Saline and the Saline River and camped near St. 

Mary’s Lake (Figure 15). During the day they passed several hunters’ cabins, but all were 

empty. On November 17 they passed Caney Marais and made their camp outside the project 

area. The following day, November 18, just below Moro Bay, Dunbar and Hunter met two 

French hunters, part of a larger party using dogs to hunt bear. On the nineteenth Hunter 

noted now low the surrounding country was as they passed through the area of the new 

Calion lock and dam; after passing the hills of Champagnolle, they made camp just above 

Champagnolle Creek. 

As the expedition returned through the area on its way downriver, Hunter recorded 

on January 12, 1805, that they passed Champagnolle Creek, then John Skinner’s camp, before 

camping two leagues below the Champagnolle hills. (A league equals about 3 marine 

miles or less than 5 km.) The next day after they passed Moro Bay and the place used by 

Tulipe for storing his hides, they came to a clearing called the Eagle (L’Aigle) and to a place 

known as la pirague d’Auguste (now corrupted to Perrogeethee and other forms). They 

camped 1.5 leagues farther down the river. On the fourteenth Hunter noted both Lapoile 

and Lapile Creeks before they left the project area and returned to Louisiana (McDermott 

1963:91-94, 111-112; Rowland 1930:242-246, 310-312). 

The camps Dunbar and Hunter saw, and the names of the places they visited emphasize 

the importance of hunters to the area. Hunter described a typical camp that they passed 

in Louisiana as having a bark cabin one-story high, about 15 feet square with no windows 

and a grass and mud chimney. The furnishings for the family of six were a bed, three 

blocks of wood for stools, a trough for pounding corn, a rifle, and a shot pouch (McDermott 

1963:90). 

Settlement of the lands near the project area proceeded slowly. The basic pattern of 

growth can be seen in the creation of new counties. Union County, which covered all of 

south-central Arkansas, was formed in 1829. Bradley County was created from parts of 

Union and Chicot counties in 1840. Ouachita County was formed from the northern part of 

Union County in 1842. Ashley County was added in 1848, and Calhoun County was created 

in 1852. 

As the population grew, landings on the Ouachita River became trade centers for the 

interior. Champagnolle, which was established before 1830 by Lawrence, John, and Silas 

Scarborough, became a prosperous town and the most important shipping point on the river 

between Camden and the Louisiana line. The county seat was at Champagnolle from 1838 to 

1844, and a government land office was opened there in 1845 shortly after the surveys were 

finished. The town had several stores, a tavern, and a jail, and several of the houses had 

plastered walls and formal gardens (Cordell 1951:37-42).


Figure 15. Historic sites, settlements, and place-names in the Felsenthal region.

History 83 

The landings at Pigeon Hill, Wilmington, and Caryville (also called Burk’s Landing) 

gained importance in the late 1840s. Even though they were closer than Champagnolle, they 

still were not conve nient to the people who had settled in southeast Union County in what 

was called the “Dark Corner” until Lapile Township was created in 1846. William Rowland, 

who had settled in 1843 near present-day Huttig met the need for a landing by building a 

floating dock at the nearest point on the river. Known as Rowland’s Raft, this landing helped 

the Lapile community grow. Wesley Rowland and his brother William Lorenzo started the 

first store there about 1850, and John V. McHenry built the first gin about the same time. The 

community had Methodist and Primitive Baptist churches, and a school named the Lapile 

Academy (Cordell 1947:275-281). 

Other improvements were also needed as the population grew away from the river. 

The Black family who settled southeast of Burk’s Landing in the 1830s had to cut a road 

from Farmerville, Louisiana to their new land. This road was later extended to the river at 

Burk’s Landing and was eventually made a part of a road from Camden to Monroe (Spencer 

1953:228, 233-234). Other roads ran from Pigeon Hill to Black’s, from Wallace’s (east of New 

London) to Lapile Creek, and from Wallace’s to Watt’s Landing on Lapoile Creek (Union 

County Court Record B:98). Similar improvements were being made on the east side of the 

river (General Land Office 1845). 

Ferries were established to make river crossings easier. The first license on this part 

of the river was issued to William Burk in 1834 for his ferry at the mouth of Moro Creek 

(Union County Court Record A:27). He later established the landing and ferry that became 

Caryville. Other ferries were established on the Saline River (Bradley County Court Record 

A:25, 56-58; Union County Court Record A:55). 

Marie Saline Landing 

The community with the most direct impact on the project area was Marie Saline 

Landing (see Site 3AS299 in Chapter 5) on the Ouachita River (Figure 16). The landing 

and ferry at Marie Saline were established in 1852 by Nathanial M. Mulholland and Joseph 

D. Christian on the Ouachita River at the crossing for the Hamburg-El Dorado Road 

(State Land Office Tract Book:18S 10W). Mulholland soon sold his interest in the landing 

to James A. Mason, and Mason and Christian received a license to run a ferry at the landing 

(Etheridge 1959:31, 136). The landing served a community that stretched along the road 

from Hamburg to the river, and along the road from Fountain Hill in the northern part of 

Ashley County to Extra Landing on the river below Lock 6. Warehouses were built to store 

cotton and merchandise, and a canal was dug so that boats could reach the warehouses 

during low water. During high water, flatboats were used to transport goods from the 

warehouses to wagons waiting on higher ground (Etheridge 1959:136). By 1856 the landing 

served enough people that a post office was established with Joseph Christian as postmaster 

(U.S. Post Office n.d.). By 1860 the extended community included 56 families, including doctors, 

surveyors, clerks, carpenters, mechanics, and a teacher (Etheridge 1959:135-136).


Figure 16. Hydrographic survey chart of Ouachita 

River in the Felsenthal Project area, showing 

locations of Marie Saline Landing (3AS299), 

Grand Mary (or Grand Marais) Landing 

(3UN122), and other features (U.S. Army 

Corps of Engineers 1871:Sheet 6).

History 85 

Trade on the river almost ceased during the Civil War, but by 1870 Marie Saline Landing 

was again a major shipping point for cotton. In 1871, 5,000 bales were shipped from 

there, making it the third largest shipper on the Ouachita River in Arkansas, after Camden 

and Caryville (U.S. Congress 1872:5). A map made the same year (Figure 16) shows three 

buildings just north of the canal (Raynolds 1871:Sheet 6). The landing changed owners several 

times in the late 1800s, going first to Fred Erickson and then to Dock Bird, who also ran 

Extra Landing. Marie Saline Landing went out of business soon after the railroad reached 

the river, and the economy shifted from cotton to timber (Etheridge 1959:137). 

disposition of swAMplAnds 

In contrast to the growth of the surrounding areas, few people settled in lowlying 

lands of the Felsenthal Project area, primarily because there was still plenty of good land 

available that was not subject to flooding. In 1850 Congress gave all unclaimed land that was 

swamp or subject to overflow to the states. Called swamplands, the proceeds from the sale 

of these lands was to be used to reclaim more land by building levees and drains. Arkansas 

eventually got 8,600,000 acres of swampland including most of the land in the project 

area (Harrison and Kollmorgen 1947:369-371). Arkansas found it very difficult to sell its 

swamplands, so in the late 1850s discussion began on donating swamplands to a number of 

railroads in exchange for stock, the justification being that railroad beds and trestles acted 

as levees and drains. The interest from the stock was to be used to reclaim lands as called 

for in the original act. The lands in the project area, along with all of the swamplands in the 

Champagnolle Land District, were given to the Mississippi, Ouachita, and Red River Railroad. 

This clouded title to the lands and prevented private ownership of these lands until 

problems with this grant were resolved in 1877 (Watkins 1981:343-44). Consequently, by 

1860 only 1200 acres within the area now designated Felsenthal National Wildlife Refuge 

had been claimed, mostly the higher ground at the edges of the area (State Land Office Tract 

Books). 

riVer coMMerce And trAVel 

Central to the study of the history of the Felsenthal Project area is the Ouachita River 

itself. The river was used for commerce as soon as there were settlers to serve, first by flatboats 

and keelboats, then in the 1820s by steamboats. The first steamboats on the Ouachita 

River in Arkansas were the Dime owned by Jacob Barkman, and the Natchitoches and Enterprise 

owned by John Nunn (WPA Author n.d.). The first to navigate the lower Saline River 

was the Howard piloted by Captain Bob Withers (Martin 1936). Steamboating on these rivers 

was seasonal, being dependent on high water, and the boats were usually shallow draft 

sidewheelers, which could be easily maneuvered through narrow channels and around tight 

bends. In spite of the difficulties, as the population grew and the production of cotton in-


creased, river commerce prospered. By 1855 there were nine boats making regular trips to 

Camden (Camden in 1855:10-11). 

The trade centered around cotton, corn, and merchandise. The higher lands along the 

Ouachita River were prime agricultural lands, well suited to a plantation economy of slaves 

and cotton. As early as 1840, Union County had one of the ten largest slave populations in 

the state. By 1850 Union County had the largest slave population in the state (4,767), with 

Ouachita County in fourth place (3,304). By 1860, Union County had dropped to fourth 

place (6,331), Ouachita County had dropped to seventh (4,478), and Ashley County had 

moved into tenth place (3,761) (Taylor 1958:51-52). Obviously most of these slaves did not 

live and work immediately along the river, but the products of their labor were shipped to 

New Orleans by boat. 

At this same time, Camden became an important commercial center, second only to 

Little Rock in population. By 1860, the town had 56 merchants, nine of them doing business 

in excess of $30,000, as well as three newspapers, three hotels, a woolen mill, a brick factory, 

a foundry, a sash factory, and a tannery. In addition the town boasted 21 lawyers, 11 doctors, 

six churches, and three schools (Kellam 1970:17-18). In 1861 roughly 20,000 bales of cotton 

(at 400 pounds per bale) were shipped down the river from Camden (Kellam 1971:18); the 

return freights were merchandise and corn. The other communities along the river contributed 

to the trade in proportion to their size, and all of this trade passed through the Felsenthal 

Project area. 

By the 1840s most boats were also carrying passengers, and a few had begun to specialize 

in this business. The accommodations varied. In 1854 the Louisa, owned by Frank 

Keeling, was advertised as “elegantly fitted with modern improvements...having new furniture 

and carpets” (Griffin 1932:65). The J. F. Pargoud was described in 1860, shortly after it 

was built, as having rosewood furniture, lace curtains, chandeliers, silverware, bathrooms, 

and “all other conveniences” (Thatcher 1970:6). It became a standard way of life to travel by 

boat for both business and pleasure (Williamson and Williamson 1939:208-212). 

River travel was not without its hazards, however. Boats hit snags, floating debris, or 

sandbars, and sank. Boilers exploded, boats burned, or were trapped by falling river levels. 

The first loss on the Ouachita River in Arkansas was the Fair Star which sank at Pereogeethee 

Shoals in 1827. She was followed by the Chieftain which hit a snag at Walnut Hills 20 

miles (32 km) below Camden in 1835, the Hadie sank at Ragged Island (now Haidee Shoals) 

in 1850, the Romeo which went down above Horsehead Shoals in 1851, and the John Ray 

which burned at Pine Prairie in 1859 (Williamson and Williamson 1939:218; Arkansas Gazette 

April 28, 1835).

effects of the ciVil wAr 

History 87 

The coming of the Civil War brought trade on the Ouachita River almost to a standstill. 

The blockade of New Orleans and its capture by the Union Navy in 1862 left no market 

for the sale of cotton, nor any place to buy other merchandise for trading. Although some 

planters continued to grow cotton, more attention was given to corn, wheat, and vegetables. 

Some trade was maintained by the smaller boats; the larger boats served the Confederacy 

as transports and cotton-clad gunships. Some were sunk near Vicksburg as obstacles to the 

Union gunboats (Owens 1979:63-103). On the Ouachita River, a few boats were beached or 

abandoned to keep them out of Union hands. One of these, the Beauregard, escaped from 

New Orleans with a load of sugar and molasses, and was left at Pereogeethee Shoals. After 

she was recovered, she served the Confederacy by hauling soldiers and corn on Bayou Macon 

(McMahan [1913]). 

There were no battles or skirmishes in the Felsenthal Project area. The Union Army 

came no further south than Camden, although they did capture the steamer Homer (Green 

1864). The Union Navy came only as far north as Ouachita City, stopping at Monroe to burn 

the courthouse and the railroad bridge and depot (Foster 1864). 

iMproVeMents in nAVigAtion 

Following the Civil War, the Ouachita River was once again the focus of the commercial 

life of south-central Arkansas. Planters who had stored cotton during the war were 

anxious to ship it to New Orleans quickly to get the best price. In 1865 there were 45,000 

bales of cotton at Camden, and as many as 11 steamboats at the wharf at one time (McMahan 

[1913]). Cotton prices fell because of the surplus available, but even so it remained the 

principal agricultural product as tenant farming and sharecropping agreements forced farmers 

to concentrate on a cash crop in order to pay their debts. 

Although the railroad reached Arkadelphia in 1873, for farmers farther south it was 

easier to ship their cotton on the river. Approximately 68,000 bales were shipped each year 

from Camden and other landings on the Ouachita River in Arkansas, including 5,000 bales 

from Marie Saline (U.S. Congress 1872:5). During the 1872 season 19 steamboats made 67 

trips up the river to Camden (U.S. Congress 1874:7). In 1876, eight steamboats made 43 trips 

to Camden (U.S. Congress 1877:479). And in the 1899 season, just before the railroads were 

extended into the extreme southern part of the state, eight boats made 24 trips to Camden, 

plus 74 trips to other landings in Arkansas (U.S. Congress 1900:19-20). Altogether in the 

30 years following the Civil War, 105 steamboats operated on the Ouachita River (Sifford 

n.d.:Vol. 4).


As river trade prospered, sternwheel boats replaced sidewheel boats because they 

had a larger cargo area and more power. Two brothers from Monroe, Fred and Jack Blanks, 

established a steamboat line to provide regular service from New Orleans to Monroe and as 

far north as Arkadelphia with connections with smaller boats up various tributaries of the 

Ouachita River (Ouachita Telegraph May 1, 1874). Their largest boat was the Ouachita Belle 

which called at landings in the project area (see Site 3UN155 in Chapter 5) during the 1870s. 

In the 1880s the Blanks Line was joined by Captain L. V. Cooley, whose boats had dominated 

the Tensas River trade, and the new company was named the Ouachita Consolidated Company 

(Williamson and Williamson 1939:221-223). 

Steamboating on the Ouachita River continued to be a hazardous occupation. Among 

the boats lost were the Willie sunk above Camden in 1868; the Native sunk at Camden in 

1872; the McCullough burned at Rowland’s Raft in 1873; and the Lotawanna (Figure 17) 

burned below Grand Marais Landing in 1874 (see below and Chapter 5) (Williamson and 

Williamson 1939:218-219). The greatest hazards were still snags and shoals. 

As early as 1840 citizens of Clark County had asked Congress for funds to improve 

navigation on the Ouachita River (Arkansas Gazette April 15, 1840) but nothing came of the 

request. Following the Civil War, local businessmen felt that it would be possible to build 

a series of locks and dams that would make the river navigable all year, thereby increasing 

trade and bringing a return of prosperity to the area. The first appropriation to study this 

idea was made in 1870, for a survey from Arkadelphia to the Louisiana line. The work was 

done by Justin Strazser of the Corps of Engineers, and resulted in the recommendation that 

$56,000 be spent to remove snags and leaning trees (Scott 1940:7). In 1871 a civil engineer, 

Clement Smith, was hired to survey the river from Trinity, Louisiana, to Camden. He recommended 

a series of five locks and dams to provide a 4-foot deep channel, and offered cost estimates 

from $1.1 to $2.1 million depending on whether the locks were of timber or masonry 

(U.S. Congress 1872:6-7). The project was approved by Congress and a contract was signed, 

but doubts soon arose about the accuracy of the survey and the wisdom of the recommendations. 

The Chief of Engineers then authorized a new survey of the river. This work was done 

in 1872 by C. W. Durham under the direction of W. H. H. Benyaurd, who agreed that a system 

of locks and dams was the only way to make the river navigable year-round. He found, 

however, that the locks recommended by Smith were poorly placed and were too small for 

the largest boats. Benyaurd outlined several alternative plans using from seven to ten locks 

and dams, with the cost ranging from $3 to $6 million, depending on the option chosen and 

the type of construction used. However, he felt that the amount of trade did not justify this 

much expense and recommended a special snagboat be built instead (U.S. Congress 1874:3- 

9; Mills 1978:43-44). 

Snagging operations had begun in 1871 on the Ouachita River using two flatboats fitted 

with cranes, before Strazser submitted his report. These were hampered by a lack of

Figure 17. The sternwheeler Lotawanna fully loaded with cotton bales (photo 

courtesy of Harry P. Fischer Collection, Marietta College Library) 

(AAS neg. 803785). 

History 89


power and by the depth of water required. In 1874, using funds from the cancelled lock and 

dam project, Benyaurd had an iron-hulled, light draft snagboat, the O. G. Wagner, built and 

snagging began in earnest in 1875 (Mills 1978:44-49). Benyaurd also tried to improve navigation 

by the construction of wing dams at the worst shoals. These concentrated the flow of 

water into one channel, which would then be deeper at low water levels. Two of these dams 

were built in Arkansas at Spoon Camp and at Buffalo Flats (Mills 1978:49; U.S. Congress 

1877:478). 

During the 1880s work on the river was limited to snagging operations to keep the 

natural channels open. New surveys were made in 1882 and 1887, but each one concluded 

that the needed system of locks and dams would cost too much. In 1889 a new snagboat, the 

Hooker, came to south Arkansas, but her wooden hull only lasted a few years (Scott 1940:10). 

At this same time, trade on the river was increasing and businessmen in Camden and 

Arkadelphia were anxious to provide for year-round navigation. In May 1893 the Ouachita 

Valley Improvement Association was formed at Camden. Led by Colonel Asa S. Morgan 

and Adolph Felsenthal, the group pressured Congress to authorize a slack water navigation 

system as far up the Ouachita River as possible. Congress responded with funds for a new, 

intensive, and scientific survey (Thatcher 1970:18; Sifford n.d.; Scott 1940:10). 

The new survey was done in 1895 and 1896 by M. H. Marshall whose report, issued in 

1902, recommended a series of nine locks and dams, including Lock 6 at Rowland’s Raft and 

Lock 7 at Pereogeethee Shoals, to provide a 6.5-foot slack water channel. These plans were 

enthusiastically supported by the Corps of Engineers, and Congress authorized the project 

in June 1902. Work then began on Lock 6 in 1904 and on Lock 7 in 1910. However, modifications 

to Lock 6 eliminated the need for Lock 7, and work on that installation ended in 1912 

(see Site 3BR32 in Chapter 5). Although Lock 6 was finished in 1915, the entire project was 

not completed until 1925 (Thatcher 1970:13-14; Mills 1978:92). 

The Sternwheeler Lotawanna 

The Lotawanna was built in 1867 at Marietta, Ohio for Henry J. Brinker of New Orleans. 

A temporary license, called an enrollment certificate, issued on October 29, 1867 at Wheeling, 

West Virginia, described her as a steamboat with three decks, no mast, a wheel stern, 

and a cabin on deck (Figure 18). The official measurements given were 155 feet long, 35 

feet wide, and 4 feet deep (drawing four feet of water). She was rated as being able to carry 

479.22 tons of cargo (U.S. Bureau of Marine Inspection and Navigation). Although tonnage 

ratings were never very accurate, this would give her a capacity of over 2,000 bales of cotton 

at 400 pounds per bale.

A b 

History 91 

Figure 18. Portions of certificates of registry for the Lotawanna. a. temporary 

certificate issued at Wheeling, West Virginia, October 27, 1867; b. 

permanent certificate issued at New Orleans, Louisiana, November 

27, 1867 (U.S. Bureau of Marine Inspection and Navigation) (AAS 

neg. 803787, 803789).


Brinker then took her to New Orleans where her Permanent Enrollment Certificate 

was issued on November 27, 1867, and she began trading on the Ouachita River (U.S. Bureau 

of Marine Inspection and Navigation). In an advertisement the following year Brinker 

described the Lotawanna as a “splendid new and light draught Packet, built expressly for the 

(Camden-New Orleans) trade,” and promised punctuality and comfortable accommodations 

(Sifford n.d.:Vol. 1:50; Figure 19). 

The Lotawanna had several owners. In 1869 Brinker sold her to Phillip A. Work, his 

clerk, and Joseph Morgan. They ran her for just over a year before selling her to H. G. McComas 

in January 1871. In May 1871 McComas mortgaged the boat for $13,613.53, which he apparently 

was not able to pay, because in August she became the property of William O’Hern, 

Michael Shelly, and Arthur Poincy (with E. R. Hart joining the partnership in September). In 

June 1872 Poincy, O’Hern, and Shelly bought out Hart’s interest before selling the Lotawanna 

to F. A. Blanks, of the Blanks Steamboat Line, in December (U.S. Bureau of Marine Inspection 

and Navigation). 

She was the largest sternwheeler on the river in the early 1870s. During the 1872-1873 

season she made seven trips to Camden from New Orleans, exceeded only by the Mayflower 

and Sabine (eight trips each) and the Bertha Bruner (nine trips) (U.S. Congress 1874:7). She 

also did some trading on the Yazoo River (Owens 1979:169). In 1874, as one of the stars of 

the Blanks Line, the Ouachita Belle being the other, the Lotawanna was advertised as a passenger 

packet leaving New Orleans once a week and connecting with the tributary steamers 

Bertha Bruner to Farmerville, Ora on Bayou Bartholomew, Stella Block on the Saline River, and 

the Willie from Monroe to Camden in low water (Ouachita Telegraph May 1, 1874). 

In May 1874 the Lotawanna was headed up to Camden with 25-30 passengers. Because 

the river was too hazardous to navigate after dark, she tied up for the night just below 

Grand Marais Landing. During the night the boat caught fire, and was destroyed, stranding 

the people on a small bar where they had tied up. There was a young man in the crew 

who was learning to be a pilot, and the next morning he volunteered to swim ashore and get 

help. He did so, finding wagon tracks and following them to the home of Arnold Nash (near 

present-day Dollar Junction). Nash and a neighbor, Jim Dollar, rode through the neighborhood 

gathering bedding and other camping equipment which they took out to the people in 

a flatboat. The passengers were rescued when a steamboat came down from Camden to pick 

them up (Johnson 1979). 

The wreck of the Lotawanna was a hazard to navigation, especially during low water, 

so Captain Justin Straszer of the U.S. Snagboat O. G. Wagner was ordered to remove it. He 

visited the wreck the week of June 14, 1875 but found the water too high to make work practical. 

The following week he had better luck and was able to salvage the following articles:

Figure 19. Advertisement from unknown newspaper source 

(about 1868) announcing Lotawanna’s Camden- 

New Orleans trade (after Sifford n.d.; Vol. 1, p. 50) 

(AAS neg. 803780). 

History 93


1. one shaft and cranks—the axle which supported the sternwheel and the 

mechanisms at each end for attaching the shaft to the pitmans. The shaft was 

wrought iron and had recently been replaced at the Leeds foundry in New 

Orleans for $1600. 

2. two pitmans—the long rods connecting the engine to the sternwheel and, with 

the cranks, changing the linear motion of the cylinder into circular motion; 

usually made of wood with metal ends. 

3. one engine—the cylinder which turned steam pressure into linear motion, a 

sternwheeler would have two engines, one connected to each end of the wheel 

by a pitman. 

4. doctor—a small auxiliary engine run by steam for operating winches and 

hoists. 

5. capstan gearing—a winch. 

6. gaspipes—pipes used to run steam from the boilers to the engines. 

7. hog chains—truss rods with turnbuckles running from bow to stern up over 

the top of the boat to keep the hull from breaking from the weight of the boilers 

in the bow and the engines and wheel at the stern. 

8. smaller articles 

Straszer took the salvaged materials to Monroe to be sold at auction. He estimated the sale 

would bring $600 to $700 if all articles sold for scrap (at $.02/pound for wrought iron and 

$.01/pound for cast iron) (Straszer to Benyaurd, June 20, June 25, August 4; U.S. Army Corps 

of Engineers 1875). 

This salvage operation did not clear the Lotawanna from the channel because Straszer 

was only able to work on the parts of the wreck that were out of the water or just below 

the surface. He returned to the wreck later in the year and blew up the forward part of the 

hull with gunpowder (U.S. Congress 1876:592). However, the wreck remained enough of 

an obstruction that it was marked on river maps as late as 1896, and the decking of the hull 

could still be seen during low water as late as about 1904 (U.S. Army Corps of Engineers 

1896:Sheet 14; Johnson 1979). 

luMbering And rAilroAds 

In the 55 years that passed between the first survey and the completion of the lock and 

dam system, great changes had taken place in the economy of the Ouachita River Valley in

History 95 

south Arkansas in general, and in the project area in particular. Even before the Civil War, 

attempts had been made to develop the area’s timber and mineral resources. Timber was 

rafted down the river as early as 1826; lignite was mined from the bed of the Saline River 

in the 1850s; and axle grease was made from an oil seep at Wilmington (Etheridge 1959:25; 

Owen 1860: 139; Spencer 1957). However, the lack of reliable transportation, the distance to 

markets, and the scarcity of labor caused these enterprises to fail. 

In the 1870s the economy was dependent on cotton, while timbering and sawmilling 

were local activities by farmers who owned boilers for gins or grist mills and wanted 

to make a little extra money by keeping their mills running all year long (Curry 1960:115). 

The first group to ship finished wood products from the Felsenthal area were the “Slavonians.” 

Originally from the mountains of Yugoslavia, these men came to the United States in 

the early 1880s to work in the steel mills. Every winter a group of 8 to 25 men would come 

to south Arkansas to cut pipe staves and claret staves. The men lived in shelters of red oak 

(which was free) and cut the staves from white oak (which they bought). The rough staves 

were cut with a broadaxe and trimmed with a draw knife. During one season the men 

could cut 90,000 staves, which they floated to New Orleans on barges in the spring (Curry 

1960:117-118). Their venture was successful enough that others followed their example. In 

1894 the H. M. Townsend left Camden with 50,000 staves and expected to pick up a barge 

of staves at the Saline River (Sifford n.d. Vol. 1). Stave cutting continued into the 1920s, at 

which time the standard price was $.03 per stave (Johnson 1979). 

Just as the new Ouachita River navigation project was getting underway, the economic 

focus of the area shifted away from the river. The St. Louis, Iron Mountain, and Southern 

Railroad (now the Missouri and Pacific) reached El Dorado in 1891, and extended its tracks 

to Bastrop, Louisiana, in 1904, crossing the Ouachita River near the new town of Felsenthal. 

Its coming made practical the building of logging railroads which opened the entire area to 

extensive timbering. Logging began at the railroad and advanced to new areas by building 

tramways (Curry 1960:118; see also Watkins 1981:54-56). 

Between 1900 and 1904, 17 sawmills and lumber companies were chartered in south 

Arkansas (Crockett 1904:359-399). The Crossett Lumber Company was the largest on the 

east side of the river. It established Crossett as its company town in 1902 and shipped out 

its products over the Rock Island Railroad (Buckner 1979). On the west side of the river the 

largest company was the Union Sawmill which established Huttig in 1902. This company 

rafted logs down the river in winter (7 million logs in 1904), but relied on the railroad for 

bringing the logs into the mill and for shipping out finished products (Felsenthal Press March 

4, 1905). 

During this early twentieth century period timbering in the floodplain concentrated on 

bald cypress, oaks, and other hardwoods, while both hardwoods and pine were cut in up-


land areas. Extensive upland clearing and pine planting began in the 1920s and 1930s (Buckner 

1979; Reynolds 1980). Trees were cut by two-man teams with large bucksaws and skidded 

by mules or oxen teams to tramways or to the riverbank for transportation to mills at 

Crossett and Huttig. 

The shift in the economy from cotton to timber and from river to railroad was reflected 

in the changing status of the towns. Communities like Marie Saline Landing and Champagnolle 

which were dependent on river trade died from lack of business when the railroad 

passed them by (Etheridge 1959:137; Cordell 1951:43-44). New communities (Bolding, 

Strong, Dollar Junction) grew up along the railroad to replace them (Cordell 1947:284). 

the felsenthAl coMMunity 

The town of Felsenthal is the best example of changes that took place in and around 

the project area. There had been a community associated with Lake Landing, where Grand 

Marais Lake entered the Ouachita River, at least as early as 1899 (U.S. Congress 1900:19). 

Adolph Felsenthal and his brother Ike bought property and laid out the town of Felsenthal 

as a manufacturing and commercial center for nearby Huttig, a company town. The first 

public sale of lots was held on April 27, 1904. At that time the town already had the Lake 

Lumber Company sawmill, a stave factory, and a brick factory. The day of the sale was a big 

celebration with excursion trains coming from Gurdon, Camden, and Monroe. Two brass 

bands and a speech by Governor Jeff Davis, as well as races, contests, and “other unique 

amusements” were provided for entertainment; and an auctioneer was brought in from 

Atlanta, Georgia, to make sure the sale was properly conducted. The big selling points for 

the new “Magic City of South Arkansas” were that it was at the junction of four railroads 

(the Little Rock and Monroe Railroad and three branches of the St. Louis, Iron Mountain, 

and Southern) and that the government had appropriated funds to build a lock on the river 

nearby (Arkansas Gazette April 24, 1904). 

The sale apparently went well, because Felsenthal prospered. In its prime it had a 

bank, a newspaper, several hotels, grocery stores, and drug stores. The brick factory was 

joined by a bottling plant and a gar factory that ground up rough fish for dog food (Johnson 

1979). The town reached its height in the 1920s while the timbering industry was at its peak 

and when the completion of the navigation project brought increased river trade. 

In the meantime events just west of El Dorado signaled another shift in economic emphasis. 

On January 10, 1921 an oil well owned by Dr. Samuel T. Busey, came in as a gusher 

and started the south Arkansas oil boom (Green 1954:51-53). As the hopeful and the curious 

flocked to El Dorado (which grew from 4,000 to 32,000), areas at the eastern end of the 

county, where most of the timber had already been logged out, lost population.

History 97 

Trade on the Ouachita River continued to grow. By the 1930s sternwheel steamboats 

had been replaced by gasoline or diesel powered towboats capable of pushing several 

large barges. With the start of World War II, however, the river trade came to a temporary 

halt—partly because the barges were needed elsewhere, partly because the channel was not 

maintained (Thatcher 1970:16). After the war businessmen all along the river were interested 

in reestablishing this trade, and this interest, through the work of the Ouachita River Valley 

Association, has brought about the current 9-Foot Navigation Project (Mills 1978:165-169). 

conclusions 

The history of the Felsenthal Project area is closely tied to the history of trade and 

travel on the Ouachita River. Beginning as a highway for hunters, trappers, and traders, 

the river was the only practical means of shipping agricultural products and merchandise 

throughout the nineteenth century. The importance of this trade was reflected in the number 

of communities and landings on the river which served as shipping points for the inland 

farms and plantations. 

The building of railroads into the area early in the twentieth century brought a change 

of economic focus. Shipping and commerce shifted inland to the railroad, and new towns 

took the place of landings which went out of business as the volume of trade on the river 

declined. At the same time, the railroads opened the area to lumbering, especially in bottomlands 

which had never been cleared because they were not suitable for farming. 

Permanent improvements in navigation kept the river open for commerce, but the end 

of intensive lumbering and the oil boom near El Dorado caused the area to lose population. 

Today towboats pass through the project area on their way to Camden, but do not stop, and 

the bottomlands are once again the preserve of hunters and fishermen.

Chapter 5 

VAriAbility AMong floodplAin sites 

River levels in the Felsenthal Project area remained above the upper project contour 

(65 feet elevation) for the first half of 1979, dropping to lock stage (62.2 feet) finally in midsummer. 

In August we assembled our field personnel and a large quantity of equipment and 

supplies at the project headquarters in Felsenthal, conducted a three-day training session, 

and sent out survey teams into the field on August 20. The intensive survey stage of the 

Felsenthal Project, which deployed three survey teams simultaneously, then proceeded essentially 

without interruption for seven weeks until the testing stage was initiated on October 

6. After this date, a single small survey team continued to operate on an occasional basis, 

concurrent with testing of selected sites (Chapter 6), until rising river levels terminated all 

field activity in the floodplain on November 24. A few site-specific surveys on higher terrain 

were conducted in the first week of December, but all aspects of fieldwork had ended 

by December 7. The termination of work in the floodplain in late November was about three 

weeks earlier than initially projected, affecting the objectives of survey only in minor respects, 

but curtailing testing more markedly. 

In this chapter we present both the methods and results of intensive survey, focusing 

on variability in the physical context (and to a lesser extent cultural content) of floodplain 

sites deduced from site survey data and subsequent analyses of these data. During the 

1979 Felsenthal Project, 126 previously unknown archeological sites were recorded and 18 

sites known from prior work were revisited. The Arkansas Archeological Survey site forms 

generated by this survey activity are voluminous, averaging seven pages for each of the 144 

sites, or more than 1,000 pages total. A limited selection of these site descriptive and analytical 

data are presented as two tables (Prehistoric and Historic Sites) in Appendix A. For 

the purposes of data presentation, comparison, and generalization we have organized the 

survey results in this chapter and in Appendix A as a series of eight site units as follows: 

Unit 1: Ouachita River above Mile 254 (18 sites) 

Unit 2: Lower Saline River (40 sites) 

Unit 3: Lower Eagle Creek (8 sites) 

Unit 4: Ouachita River below Mile 254 (20 sites) 

Unit 5: Lower Lapile Creek (8 sites) 

Unit 6: Oxbow Lakes and Backswamp (11 sites) 

Unit 7: Pleistocene Terraces and Islands (27 sites)


Unit 8: Historic Sites (15 sites, 3 of which have prehistoric components 

listed in other site units above) 

These units are not survey subareas, although they reflect the differences among various 

methods employed (see Survey Strategy and Tactics below), but were devised after all 

analyses of site data were completed. They do represent geomorphic units which vary by 

drainage area within the Recent floodplain (Units 1-6) or occur on older landforms above 

the floodplain (Unit 7). The division of the Ouachita River at mile 254 is not an arbitrary one, 

but represents the approximate present location of the mouth of Saline River. Historic sites 

have been arbitrarily grouped as Unit 8, but all occur on or below the Recent floodplain. As 

an additional organizing device, sites and site data are presented in an order from upstream 

to downstream location within each of the eight site units segregated for discussion. 

The environmental and administrative constraints on performance of archeological 

survey in the Felsenthal Project area were discussed in detail in Chapter 1. In this chapter 

we define two major subareas divided by the Ouachita River bridge at U.S. Highway 82: 

the Ouachita-Saline Channels subarea on the north and the Grand Marais Lowland subarea 

on the south (Figure 20). In the northernmost subarea the upper project contour (65 feet) is 

for the most part restricted to the banklines of the rivers and their tributaries, and even the 

lowest floodplain surfaces (in backswamp) lie above this elevation. Our survey efforts necessarily 

employed boat transportation in this subarea. The southern-most subarea, by contrast, 

includes 9330 acres (3776 ha) of floodplain terrain below 65 feet and above lock stage at 62.2 

feet, most of this terrain representing backswamp. In this Grand Marais Lowland subarea 

we have employed boat, vehicle, and pedestrian tactics in a variety of ways. For both subareas 

the environmental constraints on mobility include the following: 

1. There are no permanent or improved roads or boat landings in the floodplain 

except at the U.S. Highway 82 bridge and at the new lock construction site. 

2. The floodplain surface is hardwood and swamp forest complexly interspersed 

with river channels, creeks, sloughs, and open lakes. 

3. There are no agricultural clearings, fencelines, or artificial levees in the floodplain. 

4. The modern cultural features that do occupy the floodplain are widely scattered 

and ordinarily occur on terrain above 65 feet elevation (4-wheel drive 

tracks, fishcamps and deercamps, tramways, a pipeline right-of-way, an abandoned 

railroad crossing, and the Highway 82 crossing).

Variability Among Floodplain Sites 101 

The Felsenthal Project area as a whole, therefore, is an extensive, forested overflow bottom 

where access to floodplain terrain is prohibited, or at least sharply restricted, by high water 

and wet weather, and where the logistical basis for archeological survey must be specially 

adapted to wetland conditions. 

Table 5 summarizes the principal items of power equipment and accessories employed 

by our survey teams in various modes of survey described below. During the four months of 

fieldwork some items of power equipment were operated more or less continuously under 

adverse conditions. For the purpose of routine maintenance and repair we had technical 

manuals and an assortment of tools and spare parts. In the event of breakdowns we were 

usually able to substitute a backup item and keep the survey teams in the field. Fortunately, 

also, arrangements were made for almost overnight mechanical repair service for trucks, 

all-terrain-vehicles (ATVs), and outboard motors, and these arrangements were called upon 

frequently. Almost all of the power equipment and some of the accessory items listed in 

Table 5 had failures of some sort, but could be put back into service in a day or two. 

The diverse kinds of hand operated testing equipment, recording equipment, miscellaneous 

items and supplies used by the Felsenthal Project are largely those that would be 

employed also in large scale upland survey and need not be tabulated here. We should, 

however, emphasize the necessity for special protective or safety equipment and for waterproof 

clothing, footwear, and containers or coverings of various kinds for the purpose 

of survey in a remote wetland. Our survey teams, for example, were provided with snake 

bite and first aid kits, snake leggings, hard hats, blaze orange and flotation vests, machetes, 

compasses, canteens, daypacks, and voyager bags, and these items were used extensively 

throughout the floodplain and river channel survey effort. 

The three-day orientation and training session at the outset of the survey stage was 

an important mechanism for introducing project information and tactics, outlining safety 

responsibilities and measures, and demonstrating correct use of power equipment. Approximately 

one half of the field crew had never before entered a Southern floodplain forest. Survey 

teams had numerous close encounters with poisonous snakes and a few minor accidents 

which did not lead to significant difficulties or loss of time. 

surVey strAtegy And tActics 

The strategy for cultural resources survey in the Felsenthal Navigation Pool was 

outlined in Arkansas Archeological Survey’s proposal (1978) and a subsequent research 

design (Chapter 1, Appendix D). Our general objective was to perform an “intensive, on-theground 

survey and testing of [the] area sufficient to determine the number and extent of the 

resources present, their cultural and scientific importance, and to estimate the time and cost


Table 5. Power Equipment and Accessories Employed for Wetland Archeological Survey 

in the Felsenthal Project Area. 

Power Equipment No. of Items Accessory Equipment Principal Applications 

4WD Dodge trucks 2 winches, camper shells, moving personnel and 

tow cables, flatbed equipment to floodplain 

trailer, tool kits, first- staging areas 

aid kits 

All-Terrain-Vehicles 2 come-alongs, jerry cans, hauling personnel and 

(Hustler and Coot) tool kits, first-aid kits equipment on floodplain 

transects 

Jon boats 2 flotation and safety gear, hauling personnel and 

(16-foot and 12-foot) boat trailer, anchors, equipment on river channel 

mooring lines, and and tributary surveys 

paddles 

Outboard motors 4 auxiliary tanks, spare as above 

(25 HP and 7.5 HP) parts, tool kits 

Power augers 3 bits, extensions, jerry subsurface testing on floodcans, 

ear protectors plain transects 

Chain saw 1 gas can, safety goggles, brush clearing as needed 

ear protectors


for preserving, recovering, or otherwise mitigating adverse effects on them” (33CFR Part 

305). The strategy adopted for this intensive, on-the-ground survey is a conservative one, 

based strongly on the results of nearly a decade of difficult fieldwork and critical evaluation 

by Arkansas Archeological Survey in this portion of the Felsenthal region (Rolingson 

1972; Lishka 1973; Raab 1976b; Stacy 1976; Schambach 1976, 1979; Rolingson and Schambach 

1981). While much of this earlier work was restricted to elevated Pleistocene terraces, a brief 

floodplain reconnaissance in 1976 and the sum total of work in the area indicated that archeological 

survey in the overflow bottoms would contend with formidable environmental 

conditions. These conditions were expected to reduce both the mobility of survey teams and 

success of site discovery in the floodplain, which is largely backswamp, but not on navigable 

river channels and tributaries. Our strategy adjusted to this expectation by employing 

successive stages of survey, which integrated newly obtained results, and by employing 

varying methods and intensities of survey in each of the two project subareas. 

The Grand Marais Lowland subarea is an almost enclosed basin, 5.8 x 6.4 km in maximum 

dimensions (Figure 20). Using the best available maps (USGS Felsenthal NW, NE, 

SW and SE 7.5 minute advance workprints 1978) and an acreage grid overlay, the following 

proportions of terrain and water were calculated: 

Area of all water surface at 62.2 feet MSL 976.9 ha (2414 acres or 15.1%) 

(lock stage) 

Area of terrain at or below 65 feet MSL 3775.8 ha (9330 acres or 58.5%) 

(survey area) 

Area of terrain above 65 feet MSL 1705.0 ha (4213 acres or 26.4%) 

(above survey area) 

Total Area of Grand Marais Lowland subarea 6457.7 ha (15,957 acres or 100.0%) 

The potential survey area of 3775.8 ha is, however, widely and intricately interspersed with 

bayous, lakes, and sloughs. Except where interrupted by open water, this floodplain was 

expected to be (and indeed was) 100% hardwood or swamp forest with essentially 100% 

ground cover of leaf litter. Our maps showed a maximum relief of about 8 feet (62-70 feet 

MSL) in this expanse of floodplain, with few clearly definable elevated features excepting 

an irregular, narrow, low natural levee zone adjoining the modern river channel. We were 

concerned that habitable rises, ridges, or hummocks of little extent, not indicated on maps 

with a 5-foot contour interval, could be present, but in fact there appear to be no such small 

scale features in this lowland subarea. 

Previously recorded archeological sites in this subarea included three small prehistoric 

artifact scatters (3AS201 which was not relocated, 3UN123, and 3UN121) and one historic


site (3UN122), all on the Ouachita River bankline (see Units 4 and 8 below). Very little was 

known about the size, depth, content, or age of these sites. The survey tactics employed by 

us in the Grand Marais Lowland subarea thus assumed all of the following conditions: 

1. no floodplain zone or feature can be ruled out as a potential site location, 

2. most sites must be discovered by subsurface testing, except on exposed 

banklines, and 

3. mobility and location finding by survey teams will be very difficult. 

The last condition, in particular, required the choice of transect, as opposed to quadrat, 

methods in the floodplain, and these transects were combined with extensive bankline surveys. 

The Ouachita-Saline Channels subarea contrasts with the previously described subarea 

in many respects, most importantly in the disposition of potential survey area. In this subarea, 

extending upstream from the U.S. Highway 82 bridge, terrain below the 65-foot contour 

is restricted to the banklines of the rivers and their tributaries, and to the margins of lakes, 

swamps, and sloughs. The extent of potential survey area is more realistically represented 

by lineal rather than areal calculations (see River Bankline Surveys, and Tributary Bankline 

and Backswamp Surveys below). For this subarea we had prior knowledge of the relative 

stability of the river channels since the early 1800s (Chapter 2), and a record of three small 

prehistoric artifact scatters (3BR33, 3UN87, 3UN93, the last two not relocated) and three 

historic sites (3BR32, 3UN88, 3UN92), again all located on the Ouachita River bankline (see 

Units 1, 4, and 8 below). For the prehistoric sites little was known about size, depth, content, 

or age. The survey tactics devised for the Ouachita-Saline Channels subarea thus assumed 

all of the following conditions: 

1. no reach of a river or tributary can be ruled out as a potential site location, 

2. most sites can be discovered by careful inspection of exposed banklines, and 

3. mobility and location finding by boat survey teams will present no extraordinary 

problems in the accessible waterways. 

Extensive bankline surveys, combined with two judgmental transects in backswamp areas, 

were the methods chosen for use in this subarea. Bankline surveys are themselves a variation 

of the transect method where the intensity of survey depends principally on the degree 

and pattern of subsurface exposure. 

One of the fundamental findings of the 1979 Felsenthal Project was derived from the 

nature and relationships of buried archeological sites. Our research design anticipated the


aggrading regime of the floodplain and the need for subsurface testing methods and evaluation 

of these methods. Based on all available site data from the project area and analyses of 

these data, we can say with some confidence that all floodplain archeological sites greater 

than 50 years in age are buried. The only exception, and it is a minor one, is the occurrence 

of substantial above-ground structures in historic sites (e.g., structural earth mounds at 

3BR32 and 3UN92). The alluvial stratigraphic context of floodplain sites is examined further 

in Chapters 6 and 7. In this chapter we evaluate the use of power auger and shovel test 

methods, and more importantly the method of bankline survey. Regrettably, we have not 

had the opportunity here to evaluate the use of backhoe trenching, core drilling, grading, 

plowing, or other mechanical techniques of site location and examination. These techniques 

are presently regarded as too logistically complex, too costly, or too destructive for employment 

in an overflow bottom environment (which is also a wildlife refuge). Moreover, in 

the context of the Felsenthal Project area, only a thin slice of terrain is accessible to testing, 

below 65 feet and above 62.2 feet MSL. Excavations of any significant depth would require 

pumping if they are to be entered and examined in detail. We have therefore implemented 

relatively commonplace subsurface testing methods, augering, shovel testing, and bank 

inspection, which are distinguished chiefly by mobility. We have no great or detailed knowledge 

of deep floodplain and underwater sites, while there is virtual certainty that these are 

present (see Units 1 and 8 below). 

Random and Judgmental Transects 

In the Grand Marais Lowland subarea the initial survey stage consisted of four, randomly 

placed, east-west transects (Figure 20). This subarea was stratified on the basis of 

preliminary map observations which indicated that the floodplain varies more or less 

continuously in terrain and drainage features along the downvalley direction (N-S) and 

also perpendicular to this direction (E-W). Essentially all of the potential survey area is 

backswamp environment within the Recent floodplain, characterized by a single soil series 

and by bottomland hardwood forest, undifferentiated on available maps (U.S. Department 

of Agriculture 1968; Fleetwood 1969; Gill et al. 1979). Open lakes and sloughs comprise a 

greater proportion of the floodplain area on the south (the Grand Marais itself and Wildcat 

Lake). This hydrographic variation probably existed before impoundment by Felsenthal 

Lock and Dam in 1915, as indicated by historical sources. 

For the purpose of stratification, the subarea was divided east-to-west by the Ouachita 

River and north-to-south by an imaginary line across the valley at its midpoint. Each of the 

four strata defined in this manner was subdivided into 250 potential east-west transects of 

20-m width and variable length, and one transect was selected from each by use of a table of 

random numbers. Transects 1 to 4 (Figure 20) are 4.9, 4.5, 2.8, and 3.0 km in length respectively, 

and appear to traverse all known variation in floodplain terrain.


Figure 20. The Felsenthal Project area, subareas, and 13 floodplain transect locations.


The designated tactic for exploration of these random transects was to send the survey 

teams from a point on higher terrain into the floodplain on a due-east or due-west compass 

course, flagging and testing at 50 m intervals by power auger (10 cm diameter by 1 m depth) 

below the 65-foot contour, except where prohibited by water. Two random transect teams 

were deployed simultaneously, one proceeding on an odd numbered transect and one on 

an even numbered transect (i.e., on opposite sides of the valley). Each team (three to four 

persons) was responsible for inspecting all tree-throw disturbances, banklines, game trails, 

or other rare exposures of soil throughout the 20 m wide transect swath, and could test or 

inspect beyond these limits at the discretion of the crew chief. An all-terrain vehicle (ATV) 

was completely essential for ferrying personnel and equipment across sloughs, tributary 

channels, and even portions of lakes intersected by the transect route. We planned to extend 

and complete each of the four initial transects over a few days’ time, moving into and withdrawing 

from the floodplain by the same route each day, until the Ouachita River termination 

was reached. A 1976 aerial photo mosaic at 1:10,000 scale was prepared and marked 

appropriately, mounted on masonite, covered by plastic film, and carried into the field as the 

principal locating and plotting device. Figures 21 and 22 illustrate these items of equipment 

as employed for random transect surveys. 

The results of these random transects were largely negative; no sites were recorded in 

the backswamp portion of the floodplain and the method of proceeding by ATV was necessarily 

modified. Two sites were located on or near the riverbank and several others on higher 

terrain near the origin of transects, all by inspection rather than augering. In the case of 

all four transects we discovered that uninterrupted, end-to-end, ATV travel was impossible 

because of localized dense shrub growth and cypress knees, and that crossings of expanses 

of open water by ATV were too slow and cumbersome. Clearing of woody growth was not 

considered feasible and attempts to make short circumventions of these obstacles usually 

failed. We therefore attempted to complete the random transects in three segments each: an 

outer floodplain segment reached by ATV from the elevated terrace, an inner floodplain segment 

reached by boat and pedestrian travel from the riverbank, and a midtransect segment 

reached by ATV on the best available, circuitous, floodplain route. These attempts were 

successful to the extent that 50% to 80% of each transect route was completed, the remainder 

representing water surface or saturated terrain areas. 

Random transect work exceeded the time tentatively scheduled for this initial stage, 

and riverbank survey was briefly delayed by the competing need for boat transportation. 

The two teams emplaced a total of 390 auger holes to 1 m depth, proportionately more on 

the longer and drier transects (Tl and T2). These auger holes each required only a few minutes 

to complete, but none recovered an artifact or trace of organic remains which could be 

attributed to cultural origin (e.g., shell, bone, charcoal). One of the detrimental qualities of 

auger testing was the tendency to bring up a cohesive mass or fillet of wet clay in water-


Figure 21. Site survey tactics and equipment. a. off-loading the Hustler all terrain 

vehicle (ATV) from a flatbed trailer at a point of access to the floodplain; 

b. subsurface testing by means of power augering in bottomland hardwood 

forest (AAS neg. 796415, 796537). 

A 

b


Figure 22. Mounted air photo mosaics used for location finding and site plotting 

by boat survey and floodplain transect survey teams (AAS neg. 814156, 

814153, 813887, 813888, 814154).


logged areas, in which small sparse artifacts or organic remains would be obscured. The 

negative results of these four random transects may indicate that no sites are present in 

backswamp, but more likely that sites are too few, too small, scattered, or deeply buried to 

be located by this method. Great difficulty in completing these transects convinced us to 

ensure greater mobility in the succeeding survey stage. 

The series of six judgmental transects next employed in the Grand Marais Lowland 

subarea was considerably different in concept and execution. Three other judgmental 

transects (T8, T9, and T10) in the Ouachita-Saline Channels subarea are similar in execution. 

These nine transects (Figure 20), comprising the second stage of floodplain survey, were 

equally based on the need for greater mobility and the opportunity to explore a variety of 

localities and geohydrological settings within the floodplain. To ensure mobility we planned 

transect routes to originate on higher terrain and to follow closely 4-wheel drive tracks 

shown on air photos, the toe of elevated terraces, and existing borrow pits, all routes traversing 

floodplain but varying considerably in extent of terrain below the 65-foot contour. These 

routes minimized crossings of open water, and were surveyed by pedestrian teams using the 

ATV capability only when necessary. Characteristics of these judgmental transects, including 

the diverse floodplain areas and geohydrological settings targeted for survey, are summarized 

in Table 6. 

The principal tactic for these judgmental transects was to send the survey teams from 

a point on higher terrain into the floodplain on designated pedestrian routes, shovel testing 

at 50 m intervals (25 cm diameter by 50 cm depth) below the 65 foot contour, except where 

tree-throw disturbance or deeply rutted vehicle tracks provided comparable exposure or 

where prohibited by water. These shovel tests were closely inspected, but not screened. 

Each team (three to four persons) was again responsible for close inspection of a 20 m wide 

transect swath, and carried aerial photo mosaics at 1:10,000 scale with the route marked 

(Figure 22). Two survey teams were deployed simultaneously, in most cases operating on 

opposite sides of the floodplain. Judgmental transects were designed to proceed rapidly, 

and were in fact able to do so. Depending on length of transect, difficulty in water crossing, 

and degree of subsurface exposure, these transects required one-half to four days to complete. 

This relatively rapid method essentially sacrificed the capability of discovering sites at 

greater than 50 cm depth, but increased the extent of survey markedly for this second stage 

and presumably the ability to discover shallow buried sites. 

Transects 9 and 10 have special importance for achieving and evaluating results during 

this second survey stage. In carrying out these transects we took advantage of existing, 

linear, water-filled, borrow pits which parallel U.S. Highway 82 and extend nearly across the 

Recent floodplain. These borrow pits have steep eroded walls which extend from the water 

surface at 62.2 feet to the modern floodplain surface at 65-70-foot elevations. In effect there is 

an excellent, almost continuous exposure of floodplain sediments about 1 m high and 8 km


Table 6. Characteristics of Nine Judgmental Transects in the Grand Marais and Ouachita- 

Saline Channels Subareas. 

Transect Length 

Designation (Kilometers) Survey Target Area* Tactics Employed 

T5 3.3 Ouachita River levee and backslope to shovel test and inspect 

Open Lake 

T6 7.7 backswamp at Redeye Lake and Horseshoe shovel test and inspect 

Lake; Ouachita River levee at Henderson 

Bend 

T7 2.8 Gum Ridge in the Grand Marais backswamp shovel test and inspect 

T8 5.5 Brushy Creek-Marais Saline backswamp; shovel test and inspect 

Saline River levee at mile 1.0 

T9 6.0 Ouachita River levee at mile 252.1; Pea inspect subsurface exposure 

Ridge-Lapoile Creek backswamp in borrow pit 

T10 2.0 Ouachita River levee at mile 252.1; terrace- inspect subsurface exposure 

floodplain transition in borrow pit 

T11 8.4 terrace-floodplain transition; Grand Marais- shovel test and inspect 

Lapile Creek backswamp 

T12 7.3 terrace-floodplain transition; backswamp at shovel test and inspect 

Wildcat Lake 

T13 3.7 Ouachita River levee and backslope to the shovel test and inspect 

Grand Marais backswamp 

* Place-names and river mileages from USGS Felsenthal, Ark.-La. 15 minute (1939); floodplain geomorphic divisions 

from Fleetwood (1969).


in length. Three small sites were recorded here (3UN170, 3AS295, 3AS294; see Unit 6 in 

Appendix A) which are, of course, shallow buried components disturbed by borrow pit 

excavation. It was of great interest and importance to know at this stage that so few buried 

sites are present in an extensive transect of backswamp and in the upper meter of sediments 

exposed. 

The results of nine judgmental transects in both subareas are on the whole more positive 

and more useful than those of the preceding random four. All but one of the second 

stage transects recorded one or more sites; these sites were most commonly located on 

terrace-floodplain transition areas, but a few shallow buried sites in backswamp environments 

were also located. While we deplore the lack of an efficient and reliable method of 

subsurface survey in the backswamp, we are more confident that the number and density of 

shallow buried sites here are low. The backswamp surveys carried out by boat, to be described 

later, tend to corroborate this conclusion. 

River Bankline Surveys 

Background research and some limited prior site data indicated that riverbank sites 

could be expected to occur in both subareas. Accordingly, we devoted a major effort to 

bankline survey, employing a boat survey team (three to four persons) continuously for 

seven weeks and intermittently thereafter on both river courses. A secondary function of this 

boat survey team was to acquire floodplain environmental data of several kinds (Chapters 

2, 7). River bankline surveys, for example, provided a major source of information about the 

alluvial stratigraphic context of buried sites in the project area. The assumption made about 

potential site location in any reach of river channel was noted earlier. This assumption and 

others regarding the relationship of sites and site discovery to bank forms and channel geometry 

could be thoroughly tested against survey results (see Units 1, 2, and 4). The extent 

of river bankline survey completed is summarized below for both subareas (see Figure 23): 

Ouachita-Saline Channels subarea: 

Ouachita River above U.S. Highway 82 bridge = 37.8 river miles or 60.8 km 

(total for left and right banks) 

Lower Saline River = 22.0 river miles or 35.4 km (total for left and right banks) 

Grand Marais Lowland subarea: 

Ouachita River below U.S. Highway 82 bridge = 19.6 river miles or 31.6 km 

(total for left and right banks) 

These figures were calculated by doubling navigational distances indicated on the USGS 

Felsenthal, Ark.-La. 15 minute quadrangle (1939). Our upstream and downstream limits of 

bankline survey were miles 271.0 and 242.3 on the Ouachita River, and miles 11.0 and 0.0


Figure 23. Routes and extent of bankline survey in river channels, tributaries, 

open lakes, and backswamp areas accessible by boat.


on the lower Saline River, for a total navigational distance of 39.7 river miles (63.9 km). The 

limits here were partly determined by the practical daily range of boat travel and survey 

activity. We found it essential to operate boats from the improved launching area on the 

right bank just above U.S. Highway 82. This arrangement provided us with a secure docking 

facility, centrally located with respect to the entire Felsenthal Project area. 

River bankline survey tactics were specifically devised to meet field conditions at the 

outset of this work, and were consistently practiced throughout the seven weeks of continuous 

survey and thereafter. The major requirement was continuous close scrutiny of all accessible 

bankline by observers on foot, usually working as a pair or team. The only inaccessible 

areas were low sloping banklines with impenetrable undergrowth at the water’s edge, 

usually found on the slipoff slope of prominent bends and at the entrance to sloughs, and, 

more rarely, vertical cutbanks with no footing. We estimate that only 10 to 20% of Ouachita 

or Saline River banklines was so thoroughly inaccessible. In the remaining 80 to 90% survey 

teams inspected the beach or water’s edge, the face or slope of the bank, and the summit of 

the bank, comprising at least a 10 m wide swath extending back from the shoreline. At the 

discretion of the crew chief, boat survey personnel could, and frequently did, extend the 

survey into coves, sloughs, crevasses, or clearings adjacent to the channel (bankline survey 

in tributary drainages is described below). This degree of scrutiny is necessary because of 

the nature of floodplain sites in this region—under the best conditions of riverbank exposure 

they have extremely low visibility. 

It should be emphasized here also that bankline survey tactics encompassed the 

entire riverbank exposure and not just a portion below the 65-foot contour. Archeological 

remains occurred in situ, or eroded and transported, at any elevation or on any portion of 

an exposed bankline, and no case can be made for ignorance of the sources or stratigraphic 

relationships of these remains. This introduces a second major requirement of bankline 

survey—exposed artifacts and debris which remained in situ were so identified, and depth 

below modern surface was recorded by actual taped measurement as an indicator of relative 

age and site significance (see Chapter 7 and Appendix C). Site elevations were recorded by 

survey teams referring to river stage (read on a river guage daily), and were checked against 

U.S. Geological Survey advance work prints (1978) for the project area. 

For both the Ouachita and Saline River bankline surveys, aerial photo mosaics identical 

to those employed in floodplain transect work were the principal locating or site plotting 

device (Figure 22). Site locations and extents were marked with grease pencil on the clear 

plastic overlay of these air photos, and later transferred to project base maps. The upstreamdownstream 

length of artifact scatters was easily measured by taping or pacing along the 

riverbank (Appendix A, Table A-1, column 6). The perpendicular dimension or site width, 

however, proved to be elusive beyond the eroded bankline, for, as noted earlier, all floodplain 

sites are buried. Considerable experimentation with power augering and shovel testing 

at 5 m and 2.5 m intervals did not achieve the desired result where occupation levels


were sparse, or deep, or both of these (and indeed most sites were markedly sparse). Therefore, 

in order to make timely and efficient progress, we abandoned the attempt to trace 

site width below the surface (thus also precise buried site area), and the linear configuration 

recorded for most floodplain sites (as width x length in column 6, Table A-1) is partly 

a measure of exposure and exploration technique. However, we know from various sites 

with substantial exposure, and from others tested extensively for site limits (Chapter 6), that 

floodplain sites, regardless of overall size, consistently do exhibit a pronounced linear configuration, 

where the length along the shoreline far exceeds the width. Testing also demonstrated 

that linear configuration in these sites was a function of cultural activity rather than 

scouring and downstream transport of artifacts or debris by floodwaters (Chapter 6). 

Since riverbank exposures are a kind of sinuous transect, and one that can be effectively 

surveyed (that is, sites intersected by the river have a high probability of discovery), the 

results of our bankline site survey provide some controlled distributional data. The existing 

navigation pool enters this problem by varying the amount of riverbank exposure accessible 

to survey. As the floodplain slopes downstream, the navigation pool’s horizontal surface 

increasingly “drowns” the riverbank exposures. In our project area Ouachita River bankline 

heights decrease more or less continuously downstream from about 5 m to 1 m above lock 

stage, and Saline River bankline heights decrease from over 2 m to over 1 m. Thus, if archeological 

sites were randomly distributed in the irregular prism of floodplain sediments, the 

number of such sites intersected (and recorded) would steadily decrease downstream. Shifting 

of channel reaches and destruction of sites in these reaches would not affect the overall 

steady decrease. In advance of the results presented later in this chapter (Units 1, 2, and 4), 

we can say that random distribution is all but excluded, and that other distributional factors 

are indicated. The same drowning of riverbank exposure downstream should also cause 

older, more deeply buried sites to decrease in that direction, and this relationship is corroborated 

by our data (compare Site Units 1 and 4 later in this chapter). 

Bankline survey on both the Ouachita and Saline rivers produced significant results in 

terms of newly discovered sites, our understanding of their physical context, and our application 

and evaluation of survey methods. Records of 68 buried prehistoric sites on the 

narrow natural levees of these rivers, 64 previously unknown, are essential to both scientific 

and resource management goals of the Felsenthal Project. All 40 floodplain sites recorded 

on the lower Saline River, a dramatic and important concentration, were previously unknown. 

Most historic sites (Unit 8) also occurred on or near the riverbanks, and were newly 

recorded or reinvestigated by boat survey teams, sometimes using also archival maps or 

other historical information. One fundamental contribution of riverbank survey is an emerging 

body of data pertaining to floodplain prehistoric extractive sites (Chapter 6) and to the 

systematic relationship of site depth, age, and alluviation in the natural levee zone of this 

lowland region (Chapter 7).


Tributary Bankline and Backswamp Surveys 

Boat or pedestrian survey of banklines in tributary waterways was conducted using 

the principles and most of the tactics described above. All of this work was scheduled late in 

the seven-week period of intensive survey, when floodplain transect crews could be reassigned, 

and when substantial results from riverbank survey were available. These surveys 

thus represent a third stage of exploratory work, although they continued to emphasize 

bankline inspection within the backswamp portion of the floodplain. A major determinant of 

the extent of survey was again accessibility and the practical range of boat travel. For much 

of the work considered here a smaller (12-foot) boat with a two-person survey team was employed, 

although a larger pedestrian crew was used in some areas accessible overland. The 

extent of accessible tributary channels and backswamp waterways surveyed is characterized 

below by subarea (Figures 6, 23): 

Ouachita Saline Channels subarea: 

Eagle Creek (total for left and right banks) 5.6 km 

Jones Lake (south bank) 1.0 km 

St. Mary’s Lake (west bank) 2.4 km 

Grand Marais Lowland subarea: 

Marais Saline Lake (perimeter) 2.2 km 

Lapile Creek (total for left and right banks) 5.8 km 

Lapile Bayou, Fishtrap Lake, and backswamp 

distributaries (total for left and right banks) 21.6 km 

Wildcat Lake and distributaries (perimeter 

and irregular route) 4.5 km 

Grand Marais and distributaries (perimeter 

and irregular route) 6.5 km 

Bankline exposure in these smaller channels and backswamp areas was generally poor 

with dense undergrowth often obscuring low shorelines. We estimate that less than 10% of 

the routes summarized above were well enough exposed for effective bankline survey with 

one major exception. This higher degree of exposure was well developed on Lapile Bayou in 

backswamp and greatly exceeded on the Lapile Creek segment which extends into the Pleistocene 

terrace (Figure 23). Heavy vehicle and foot traffic here by fishermen and consequent 

bankline erosion have thoroughly exposed the margins of this tributary channel (see Unit 5 

below). 

Our survey efforts in tributary waterways proceeded fairly rapidly, chiefly because the 

most accessible routes were selected and because good bankline exposure was rare. Relatively 

few sites were recorded in an actual backswamp environment, and these were usually 

exceedingly small and sparse (Unit 6), paralleling results from floodplain transect work. We 

can state with greater assurance that shallow buried sites are few and scattered in the back-

Variability Among Floodplain Sites 117 

swamp zone of the floodplain. If additional time were available for investigation of less accessible 

areas, we would have extended the survey to the Caney Bayou-Lapoile Creek channels 

and to large oxbow lakes on the north. At Pereogeethe Lake one previously recorded 

prehistoric site (3BR19) was tested by Lishka (1973) with modest, but interesting, results 

from our floodplain perspective. 

Site Specific Surveys 

Site specific survey was employed as a means of acquiring regional comparative data, 

both archeological and environmental, rather than as a technique for locating previously 

unknown sites. This work was conducted by the project archeologist, usually assisted by a 

survey team, but in some cases working with consulting personnel or informants. The following 

site specific work contributed diverse kinds of data and results: 

1. Observed new lock construction site, lock chamber excavation, and spoil areas. 

2. Collected samples for lithic source study at Recent floodplain and Pleistocene 

terrace locations (Chapter 8). 

3. Visited and recorded historic sawmill site with informant (3BR55). 

4. Visited and recorded newly disturbed prehistoric site on Pleistocene terrace 

(3AS328). 

5. Visited and recorded natural prairie area and associated archeological sites on 

Pleistocene terrace (3AS160, 3AS293). 

6. Revisited and recorded major mound groups on Pleistocene terraces and islands 

with Arkansas Archeological Survey station personnel (3BR4, 3BR8, 3UN18, 

3UN8, 3UN9/52, 3AS1, 3AS6). 

7. Located and recorded underwater steamboat wreck with marine survey archeologist 

(3UN153). 

Site and Component Identification 

Our record of floodplain sites is a product not only of specific survey tactics, but also 

of operational definitions and levels of analysis. We shall briefly introduce certain of those 

definitions and levels here, and note that differences may exist with other large scale cultural 

resource surveys (e.g., House and Schiffer 1975). 

The principal unit of observation for Felsenthal Project survey teams was an artifact 

or artifact scatter. Debris of presumed cultural origin or of uncertain origin was similarly 

recorded. In the stoneless floodplain topstratum, any medium to large pebble or coarse rock 

fragment was regarded as a potential “manuport,” and recorded and collected, but was 

subsequently judged as to modification and context. Survey teams made 100% surface collections 

on prehistoric artifact scatters, and either 100% or select representative collections 

on historic scatters. A series of consecutive project numbers was assigned for use by each 

survey team, and ultimately about 200 numbers were used, although some refer to environ-


mental localities or samples. At this level of operation (survey team recording) no decision 

had been made about “sites.” 

In recording historic artifacts, scatters, or structures, we specified that “modern” 

remains be rejected, and those remains 50 years of age or greater be recorded and collected. 

Doubtful examples were recorded, collected, and judged subsequently. In practice this eliminated 

all the “ongoing” remains of fishcamps, hunters’ camps, vehicle tracks, and timber 

harvesting activities. While historical remains less than 50 years of age can have regional 

and even scientific importance, recording these clearly becomes an endless task for specialists. 

A contemporary survey in the Felsenthal Project area differs on this point (Heartfield et 

al. 1980). 

Our second level of operation consisted of processing, cataloging, and limited study of 

collections at the Felsenthal Project field laboratory. Individual objects, collections, or localities 

were occasionally eliminated when no archeological or environmental importance could 

be established. All archeological collections were now entered in the permanent cataloging 

system of Arkansas Archeological Survey, while unassociated environmental localities and 

samples retained only project numbers. Again no decision about “sites” had been made, 

except by eliminating certain localities or collections as indicated. 

The third level of operation consisted of detailed analysis of both field data and collections, 

conducted upon our return to permanent laboratory and research facilities. At this 

stage the numbered floodplain loci of artifacts or other cultural remains were designated as 

sites, or combined as sites, based on all survey and analytical results. We now had to distinguish 

single finds (an apparently isolated artifact or item of debris firmly attributed to 

human activity) from sparse scatters of such items, and these from more measurable, dense, 

discrete clusters, usually having diameter of a few tens of meters. A “50 m rule” was developed 

at this point—scattered finds or discrete scatters less than 50 m apart could be combined 

as sites when no other data mitigated against this assignment. A single fire-cracked 

rock fragment could be designated a site (3UN162 is just such a site) when careful inspection 

produced no other artifacts within 50 m, and so also could two flint flakes 49 m apart. Given 

the widely scattered nature of our finds, this rule was rarely needed, but Unit 5 (discussed 

below) is a most notable exception. In Chapter 6 relatively dense concentrations of artifacts 

(Figure 41) and closely spaced activity loci (Figures 47, 51, 53) are illustrated. We do not recommend 

use of a 50 m rule in any other regional setting without careful consideration. 

For each of the 126 previously unknown sites distinguished at this level, standard Arkansas 

Archeological Survey site forms were completed, checked, and distributed to station 

archeologists for comments and assignment of trinomial site designations. Site names were 

also assigned, except in the case of single finds, by the project archeologist. These site names


were drawn from local place-names where available, and from prominent landscape features 

or common names of local woody species (Appendix A, Table A-1, column 10). For 

each of the 18 sites known from prior work, updated site forms were completed, and similarly 

checked and distributed to station archeologists. 

We now define a site as “any spatially discrete locus of remains of past human activity.” 

These remains may or may not be preserved in place on the locus of use and abandonment, 

and they may or may not result from a single episode of past human activity. These 

questions and others must be answered by detailed study of association or context for any 

designated site, and by application of general principles of site formation and transformation 

process (Schiffer 1976; Binford 1980). In the case of Felsenthal floodplain sites, the loci of 

remains are comparatively very small in size and the site assemblages are small in quantity 

and low in diversity. As a class, such small sites have frequently been attributed to brief, 

one-time use by a task unit or small social group (Talmage and Chesler 1977). These sites 

have high potential for behavioral analysis in general, and for settlement and subsistence 

studies in the Felsenthal region (Chapter 6). 

Essentially the last operation performed with site data was assignment of prehistoric 

components, where possible. A “component” is a contemporaneous assemblage of remains, 

sometimes a level, within a specific site (Willey and Phillips 1958:21). Typological and 

stratigraphic correlation permits assignment of related components to “phases” in a locality 

or region. Two difficulties beset our assignments: (1) the Felsenthal regional sequence 

is still in an early stage of formulation, based principally on large upland sites (Chapter 3), 

and (2) floodplain site artifact samples are usually small and restricted in variety (Appendix 

A). However, our floodplain site data include numerous discrete clusters of artifacts and 

some buried “floors” where co-occurrence of artifacts and raw materials can be observed 

and verified again and again. Most artifact finds were made on river banks with more or less 

slope and which expose generally flat-lying sedimentary units. Artifacts or debris still firmly 

embedded in sediment, essentially flat-lying and arranged along a bedding plane or thin 

sedimentary unit, were assumed to be in place. Artifacts reposing at angles on the surface 

of the bank, scattered in location or elevation, were assumed to be transported (usually a 

short vertical distance downslope). The following attributes of components in floodplain 

sites have been determined, or at least are strongly suggested, by our site data (see site and 

artifact data in Chapters 3, 6, 7, 8, and Appendix A): 

Mississippi Period, Caney Bayou Phase 

1. buried levels at less than 0.2 m depth 

2. shell tempering in 75-100% of ceramics 

3. Eagle Creek, Ashley, Bassett and other arrow point styles 

4. high proportion of brown chert among debitage


Mississippi Period, Gran Marais Phase 

1. buried levels at about 0.2 m depth 

2. Gran Marais ceramic complex featuring clay-tempered vessels with diverse 

rim and body decoration (incising, punctating, and other); shell tempering in 

only 5-15% of ceramics 

3. Ashley arrow point style (a variable type cluster) 

4. high proportion of brown chert among debitage 

Baytown-Coles Creek Period, phases undifferentiated 

1. buried levels at about 0.2-0.6 m depth 

2. high proportion of clay-tempered plain vessels with disc bases; occurrence of 

incised rims in Coles Creek style; no shell-tempered ceramics 

3. small variety of Gary dart point, arrow points (?), and other dart point styles 

4. both novaculite and chert well represented in debitage 

Marksville Period (no floodplain site data) 

Marksville sites may be few, but present, deeply buried, and sparse, as are Tchula sites. 

Such sites would probably lack distinctive decorated sherds (such as Marksville Stamped) 

and have only plainware, in contrast to Tchula sites which are ceramically distinctive. Thus, 

Marksville sites may have been missed due to sampling error or are misclassified as deeply 

buried Baytown sites. 

Tchula Period, Coon Island Phase 


2. incised, stamped, or drag-and-jab decoration on clay-tempered or temperless(?) 

vessels 

3. Gary dart points and perhaps other dart point styles 

4. high proportion of novaculite among debitage 

Poverty Point Period, Calion Phase 

1. buried levels at 1.0 m or greater depth 

2. no ceramics 

3. large Gary dart points and other dart point styles 

4. high proportion of light colored novaculite among debitage 

5. exotic materials including steatite, schist 

6. high frequency of fire-cracked rock compared to ceramic sites 

Late Archaic Period, phases undifferentiated 


2. no ceramics 

3. large stemmed dart points 

Middle and Early Archaic Periods (no floodplain site data)

Paleo-Indian Period (no floodplain site data) 


In Appendix A, Table A-1, components represented in prehistoric sites are identified 

where sufficient comparative data are available. All such identifications represent inferences 

where confidence decreases with fewer typological, technological, or stratigraphic observations. 

Nevertheless, one or a few diagnostic artifacts were assumed to be good criteria of 

identification in the context of small floodplain sites, just as they may be in small upland 

sites (Hemmings 1975). We make an attempt to identify cultural period, where phase cannot 

be distinguished; or to indicate an undifferentiated span (“Archaic”) in some cases. About 

one third of all prehistoric floodplain components (Units 1-7) remain in an “unknown” category. 

Large mound complexes on Pleistocene terraces (Unit 8) are known from existing site 

records to have a long sequence of site occupation in most cases, and sequences of cultural 

periods are indicated in Table A-1 (“Archaic-Mississippi”). 

In Appendix A, Table A-2, historic components are assigned to decade spans within 

the nineteenth and twentieth century, based on regional historical data (Chapter 4), or more 

rarely, diagnostic historic artifacts. Archival data for more precise dates is available in a few 

cases (Unit 8). 

surVey results 

Unit 1: Ouachita River above Mile 254 

Unit 1 includes 18 prehistoric sites on a sinuous 17 mile (27.4 km) stretch of Ouachita 

River (Figure 24). In this stretch the river proceeds only 8 miles (12.9 km) in the down-valley 

direction, and is marked by prominent bends and oxbow lakes, the latter cut off prior to the 

Dunbar-Hunter expedition of 1804-1805 (Rowland 1930; McDermott 1963). Dunbar’s map, as 

well as the General Land Office maps of 1827 and later, show that the river has maintained 

its course, bend for bend, during the last 175 years. The U.S. Army Corps of Engineers’ hydrographic 

surveys of 1873 and later also indicate that low water shoals of gravel and sand 

were extensive in this stretch of river, in contrast to the stretch immediately downstream 

(Unit 4). The floodplain here also exhibits greater relief, with distinctive point bar ridges 

inside bends and inside oxbow lakes. Dunbar and Hunter, traveling upstream, noted the 

appearance of small canes and a few pines, on or near the riverbank, attributing this growth 

to reduced extent and duration of flooding. Thus, in certain respects the river channel and 

its margin in Unit 1 represent less extreme overflow conditions than are found in the Grand 

Marais Lowland subarea downstream. 

Beginning with this group of 18 prehistoric sites, we will examine the relationship of 

site distribution to river channel geometry. The record of riverbank sites may be a product of 

all the following factors:


Figure 24. Site Unit 1, Ouachita River above mile 254, with prehistoric sites (dots) 

located in relation to channel geometry. Some prehistoric sites in Unit 6 

and historic sites in Unit 8 (rectangles) are also shown on this map.


1. Preferential site location, e.g., adjacent to shoals in the crossover point between 

bends or on higher levee crests adjacent to the concave bank of major bends. 

2. Differential site destruction, e.g., on concave cutbanks of migrating bends. 

3. Differential site visibility, e.g., on eroding cutbanks versus prograding slipoff 

slopes obscured by vegetation. 

4. Artificially imposed bias, e.g., when survey teams give unequal consideration 

to left or right banks or to various reaches. 

Using meander terminology introduced in Chapter 2 (Figure 8), we can classify all riverbank 

site locations in eight possible categories, rate these categories by expected site density (from 

the first three criteria above), and compare the observed site distributions. Table 7 makes 

this comparison for Unit 1 and two other site units on the Ouachita and Saline rivers. In Unit 

1, riverbank sites are most frequently represented on the downstream arm (below the axis) 

Table 7. Channel Geometry and Prehistoric Site Distribution for Three Site Units on 

Ouachita and Saline Rivers. 

Expected Observed Site Density 

River Channel Location Site Density* Unit 1 Unit 2 Unit 4 Total 

Concave bank above bend axis high 1 16 7 24 

Concave bank at axis moderate 3 3 1 7 

Concave bank below axis high 9 6 5 20 

Crossover point between bends moderate 0 7 4 11 

Convex bank above bend axis moderate 5 2 2 9 

Convex bank at axis low 0 3 1 4 

Convex bank below axis low 0 1 0 1 

Straight reach low 0 2 0 2 

Site totals 18 40 20 78 

* Based on processes observed or inferred for the relatively stable courses of these rivers and not 

necessarily applicable to other meandering river regimens.


of the concave or convex bank. In certain respects the Unit 1 site distribution does not agree 

with “expected site density” or with distributions observed in Units 2 and 4. One possible 

explanation of this discordance is the high frequency of Archaic sites in Unit 1, as discussed 

further below. These older sites may well have been established on or near river courses 

which vary markedly from the “modern” channel serving as our geometric reference. 

It is of interest to note that the 18 prehistoric sites recorded in Unit 1 are equally distributed 

on the left and right banks (Figure 24). We suspect that extensive shifts of channel 

during the past several centuries or imbalance in survey intensity would result in an asymmetrical 

distribution. 

The following components are represented in Unit 1: Archaic 8, Baytown-Coles Creek 

3, Tchula/Coon Island 1, and unknown 6 (Appendix A, Table A-1). Three of 18 sites recorded 

are single finds. The large proportion of Archaic sites is very likely a function of high banks 

and exposure of deep alluvium in Unit 1. Site 3UN162, consisting of a single fire-cracked 

cobble in place at 2.95 m below the levee surface, is the deepest artifact occurrence recorded 

by us in the Felsenthal Project area (Figure 25). Shovel scraping along the alluvial horizon 

indicated by this artifact did not produce other cultural material. This site exemplifies the 

problem of locating and assessing significance of sparse buried sites in an alluvial floodplain. 

We have no basis for deciding whether it represents the vestige of an eroded occupation 

level, the initial indicator of an undisturbed level, or an isolated object transported from 

some other location. 

Prehistoric sites in Unit 1 are small, sparse, and relatively deep compared with other 

site units (Appendix A, Table A-1). Few survey collections exceed a dozen objects, and inorganic 

debris (usually fire-cracked rock) is the major constituent of these collections. Diagnostic 

ceramics and stone tools are characteristically rare. A large novaculite Macon dart point 

from Small Cane 2 (3BR73) is shown in Figure 59, and Tchefuncte Stamped sherds from 

Hunter’s Swan (3BR70) are shown in Figure 65f. Although Unit 1 sites are sparse, buried 

materials are known to occur in place in some of these and may be expected to occur in others. 

The probability of Middle Archaic occupations (before about 4000 B.C.) in deeply buried 

zones is good, and even older occupations could be present (above river stage) as well. Identifiable 

Late Archaic and early ceramic occupations are known, while late prehistoric sites 

(after A.D. 1100) are curiously absent, or at least presently unknown. 

Monitoring of the reach of Ouachita River presented by Unit l is desirable in the 

future. Three of the presently recorded sites have been recommended for mitigation (sites 

3BR72, 3BR73, 3UN166 in Appendix C). Ongoing riverbank erosion could reveal sites to be 

more substantial than they now appear, disclose unrecorded sites, and clearly could remove 

the ephemeral traces of some sites now known. The functional characteristics of these sites


Figure 25. Site 3UN162, consisting of fire-cracked rock exposed at 3 m depth in the 

Ouachita River bankline (AAS neg. 796519).


remain to be established through testing and analysis of site data, as we have attempted to 

do in Units 2 and 4 (Chapter 6). One of the interesting possibilities is that prehistoric fishing 

stations here were established at riverbank locations adjacent to extensive shoals. 

Unit 2: Lower Saline River 

Unit 2 comprises the lower 11 miles (17.7 km) of the Saline River where 40 riverbank 

sites, all previously unknown, were recorded (Figure 26). The Saline River’s course is 

slightly less sinuous than the Ouachita River in Unit 1, making about 6.5 miles (20.1 km) in 

the down-valley direction through this reach. Meander length, amplitude, and radius, and 

channel width, are all smaller than observed for the Ouachita River, reflecting the Saline 

River’s lower discharge. Relatively many spits, lined by cypress trees and buttonbush, and 

elongate coves or bays give this lower Saline River course a drowned appearance (Figure 

46). Bank height decreases from greater than 2.0 m at river mile 11.0 to greater than 1.0 m at 

the mouth. 

The General Land Office maps of 1827 and later years indicate that the lower Saline 

River channel has also remained stable for at least a century and a half. Historic and modern 

maps of the channel can be correlated bend for bend, with a single exception. This exception 

is the cutoff of a small bend at the mouth of the Saline River about 1895, probably artificially 

assisted, as shown on a U.S. Army Corps of Engineers hydrographic survey (Sheet 11) 

for that year. The riverbank setting and perhaps the integrity of two sites (3BR77 in Unit 1, 

3AS284 in Unit 2) are affected by this cutoff (Figures 24, 26). 

Table 8 shows considerable agreement between expected and observed site densities 

among the eight possible channel locations previously discussed. Sites were most frequently 

recorded along concave banklines above the axis of a bend. Relatively many sites were also 

recorded at crossover points between bends. One or more sites occur in all eight categories. 

The distribution of sites by left and right bank is slightly asymmetrical (23/17), which could 

reflect survey team bias. The boat survey team ordinarily proceeded upstream, working 

along the left bank, in the morning, and downstream along the right bank in the afternoon. 

The most interesting aspect of site distribution in this Saline River unit is the increasing site 

density downstream, contrary to the probability of site discovery as bank height and exposure 

decrease (Figure 26). Twenty-eight of 40 sites recorded lie along the 5 mile (8 km) reach 

above the mouth. This is a dramatic concentration of floodplain sites unequalled elsewhere 

in the project area. The downstream increase in site density is believed to be significant and 

real, not an artifact of survey intensity (see Site Locational Hypotheses, this chapter). 

The following 41 prehistoric components are represented in Unit 2 (Appendix A, Table 

A-1):


Figure 26. Site Unit 2, lower Saline River, and Site Unit 3, lower Eagle Creek, with 

prehistoric sites (dots) located in relation to channel geometry. Two 

historic sites with prehistoric components (rectangles enclosing dots) in 

Unit 8 are also shown on this map.


Mississippi/Terminal 2 

Mississippi/Caney Bayou 4 

Mississippi/Gran Marais 6 

Mississippi 4 

Baytown-Coles Creek 4 

Tchula/Coon Island 1 

Poverty Point/Calion 2 

Archaic 4 

Unknown 14 

Only one multicomponent site has definitely been identified; False Indigo (3AS285) has 

a Gran Marais phase component overlying a Coon Island phase component, hence the total 

of 41 components reported above for 40 sites. This site and six others in Unit 2 were systematically 

tested, and are reported in detail in Chapter 6. It is important to note that 40% of 

components identified here are Mississippi period, while no component of this period was 

identified in Unit 1. The lower reach of the Saline River thus appears to have been selectively 

utilized by Mississippi period groups, a point we return to in discussing other site units 

below. 

While the riverbank sites in Unit 2 are generally of small size, one-fourth of these 

exceed 50 m in length. These latter sites frequently include multiple discrete artifact scatters. 

Jug Point 1 (3AS306) and Jug Point 2 (3AS307), for example, circumscribe five such scatters 

within a 225 m stretch of Saline River bankline (Figures 51, 53). In some cases these scatters 

appear to be coeval activity loci and in others merely successive occupations of closely 

spaced loci (Chapter 6). The typical lower Saline River site is evidently a miniscule 10-20 m 

in length, but relatively densely marked by artifacts and debris where shallow buried occupation 

levels are eroding (Figure 27). 

Collections of artifacts and debris obtained in these Saline River sites range from a 

single item (five sites) to nearly 5,000 items (One Cypress Point, 3AS286). Mississippi period 

sites furnished the greatest proportion of this material, and our analyses of ceramics, arrow 

points, other stone tools, and both organic and inorganic debris have been the fundamental 

basis for understanding Mississippian exploitative activities in the floodplain (Chapters 6, 8, 

Appendix B). 

Unit 2 is a unique and significant group of 40 prehistoric sites, all buried and none 

plowed or disturbed to any extent except by processes of bankline erosion. No other site 

unit in the Felsenthal Project area has produced the quantity and quality of data represented 

by this group. Extensive further investigation and measures of mitigation have been 

proposed for eight sites in Unit 2 (Appendix C). In the next chapter we use test excavation 

data from Saline River sites to investigate functional characteristics of floodplain extractive 

camps and also to examine variation among extractive sites of the Mississippi period.


Figure 27. Mississippi period extractive sites on lower Saline River. a. Jug Point 1 

(3AS306); b. One Cypress Point (3AS286) (AAS neg. 797088, 797133). 

b 

A


Unit 3: Lower Eagle Creek 

All eight sites recorded in this unit are clustered near the mouth of Eagle Creek, although 

boat survey was extended throughout the accessible 1.7 mile (2.8 km) lower reach 

(Figure 26). While Fleetwood (1969) shows Eagle Creek as an abandoned Ouachita River 

course, the lower channel is so atypically straight from NNW to SSE that some structural 

control may be suspected. No Quaternary faulting is known to have occurred in this region 

(Saucier and Fleetwood 1970:870). 

Lower Eagle Creek sites are essentially within the slightly elevated natural levee zone 

of the Saline River, and these sites might easily be considered ecologically equivalent in location 

to Unit 2. The eight recorded sites occurred equally on concave and convex bank-lines, 

but are asymmetrically distributed with regard to left and right banks (2/6); two sites on the 

right bank are single finds, however (Appendix A, Table A-1). 

The components identified in Unit 3 appear to extend the dense cluster of Mississippi 

period components on the adjacent reach of Saline River. Unit 3 components include the 

following (Appendix A, Table A-1): Mississippi/Caney Bayou-one, Mississippi/Gran Maraisone, 

Mississippi-four, and unknown-two. These components are generally shallow, small 

in size, with moderate amounts of artifacts and debris eroding out, including occasional 

organic remains. Three of these small but substantial sites, 3BR65, 3BR66, and 3BR79, have 

been recommended for mitigation (Appendix C). 

As in the case of Unit 2, the lower Eagle Creek components are largely undisturbed 

and may have high potential for investigating the nature of floodplain extractive sites. 

Unit 4: Ouachita River below Mile 254 

The 20 prehistoric sites in Unit 4 occur along the lower 11.7 mile (18.8 km) stretch of 

Ouachita River below the mouth of Saline River (Figure 28). The river here is just slightly 

less sinuous than the reach immediately upstream (Unit 1), proceeding 6.0 miles (9.7 km) in 

the downvalley direction. The prominent bends of the present channel are found also on the 

Government Land Office maps of 1827 and later years, and on Dunbar’s 1804-1805 map with 

the exception of one bend (at mile 251.0) on the latter. The omission of this unnamed bend 

on the Dunbar map appears merely to be a discrepancy rather than rapid development of a 

meander between 1805 and 1827. The floodplain in Unit 4 is characterized by low, narrow 

natural levees adjoining the stable channel, and extensive backswamp areas with bayous 

and open lakes. 

Table 8 shows a distribution of bankline sites in Unit 4 much like that of the lower Saline 

River, but less dense. Sites were recorded most frequently on concave banklines and on 

crossover points between bends, as expected. There is rather marked asymmetry in the


Figure 28. Site Unit 4, Ouachita River below mile 254, with prehistoric sites 

(dots) located in relation to channel geometry. Historic sites (rectangles) 

or historic sites with prehistoric components (rectangles 

enclosing dots) in Unit 8 are also shown on this map.


occurrence of sites on left and right banks (14/6). We cannot presently attribute this asymmetry 

either to survey intensity or to a shift in river channel. The dearth of sites on the right 

bank above Lapile Creek was noted while boat survey proceeded, and was checked by a second 

boat survey team without positive results. Given the extreme low visibility of riverbank 

sites in this area, we would expect some sites to be recorded here by future efforts under 

ideal conditions of bankline survey. 

Riverbank sites in Unit 4 compare in small size and sparse content to our sample from 

Unit 1 particularly. Only one site, Marie Saline (3AS329), exceeds 50 m in length. This multicomponent 

stratified site is reported in detail in Chapter 6, based on text excavation results. 

Ceramics are relatively well represented in collections from Unit 4 sites, while all other categories 

of tools and debris are rare (Appendix A, Table A-1). Five of the 20 sites recorded are 

single finds. In Appendix C five ceramic sites with relatively substantial artifact assemblages 

(3AS287, 3AS321, 3AS322, 3AS327, 3UN168) have been recommended for mitigation, in addition 

to extensive further work at Marie Saline (3AS329). 

In contrast to results of survey in Unit 1, few Archaic and many later ceramic sites 

were recorded in Unit 4. This is certainly attributable to low bank height and limited exposure 

of deeper alluvium (Figure 29). Mississippi period components are also much less 

frequent than observed for Unit 2. The following 23 components from 20 sites in Unit 4 are 

identified: 

Mississippi/Caney Bayou 1 

Mississippi/Gran Marais 2 

Mississippi/Caddo II 1 

Mississippi 1 




Late Archaic 1 

Unknown 5 

The small buried riverbank sites of Unit 4 are a significant sample of floodplain sites 

in the region. The depths of relatively many components are known (0.1 to 2.0 m below 

the modern levee surface), and correlate well with typological ages of these components. 

In Chapter 6 we argue that stratified site data from Marie Saline (3AS329) are fundamental 

to understanding prehistory in the Felsenthal region. Unit 4 sites are essentially the lowest 

known prehistoric sites in the state of Arkansas, by reason of their occurrence at the downstream 

end of the project area and near the Ouachita River crossing of the state line. 

Unit 5: Lower Lapile Creek 

The eight prehistoric sites in this unit lie along the margins of Lapile Creek just upstream 

from the Recent floodplain of Ouachita River (Figure 30). This location represents a


Figure 29. Riverbank sites on Ouachita River. a. Marie Saline (3AS329); b. mouth 

of Lapile Creek (3UN123) (AAS neg. 796403, 797124). 

A 

b


1 km wide tongue of low floodplain, bordered north and south by higher Pleistocene terraces 

(Fleetwood 1369). Although we regard this unit as an extension of the Grand Marais 

Lowland subarea, it varies somewhat in environmental character being hemmed in by 

higher terraces and in site distribution from all other areas surveyed. 

Lapile Creek sites were resolved (by the 50 m rule) from a great many scattered finds, 

amounting in some cases to almost continuous distribution of artifacts and debris along 

both banklines. The many project numbers, large site dimensions, and high totals for artifacts 

and debris listed in Appendix A (Table A-1, Unit 5) testify to this complex situation. At 

least six of these sites include multiple activity loci which impinge on or overlap each other. 

More than one component is identified in two sites, based on survey collections available, 

but others may remain unidentified. The Lapile Creek bankline is today heavily used by 

fishermen, extensively denuded of ground cover, and may be subject to casual artifact collecting 

activity. In addition to the complex problems of density and distribution of artifact 

scatters along lower Lapile Creek, we are not satisfied with the field records furnished by 

our survey team in two respects: precision of linear dimensions within and between scatters, 

and absence of depth observations for buried material. Therefore, the record of prehistoric 

sites in Unit 5 is a preliminary one, not completely understood, and in need of additional 

carefully designed survey work. In particular, this survey work could be extended upstream 

from the known site concentration. 

The following 10 components were identified from eight sites recorded in Unit 5: 

Mississippi 1 


Archaic 2 

Unknown 2 

Two Unit 5 sites of relatively large size and substantial content have been recommended 

for mitigation (3UN151 and 3UN157, in Appendix C). Ceramics, stone tools, and inorganic 

debris of several kinds, including daub, are well represented in the collections from these 

sites. 

Lower Lapile Creek sites recorded by us are located in close proximity to many substantial 

habitation sites on the elevated Pleistocene terraces (cf. Rolingson 1972). One interesting 

aspect of these floodplain sites is their functional relationship to upland settlements. 

Intensive use of the Lapile Creek channel margin, even though subject to frequent prolonged 

overflow, is indicated by existing site data. 

Unit 6: Oxbow Lakes and Backswamp 

Unit 6 comprises 11 prehistoric sites in both subareas which occur in low backswamp 

areas, more or less remote from the modern channel and levee of Ouachita River (Figure 31).


Figure 30. Site Unit 5, lower Lapile Creek, with prehistoric sites (dots) located in 

relation to channel geometry. Two mound sites on the adjacent terrace, 

Locust Ridge (3UN8) and Shallow Lake (3UN9/52) in Unit 7 are also 

shown on this map.


The extent and difficulty of backswamp survey were noted earlier in this chapter. Except to 

note that exceedingly small, sparse, artifact scatters were recorded, it is not possible to generalize 

from these backswamp sites. One Mississippi period and two Baytown-Coles Creek 

period components were identified; eight other prehistoric components remain unidentified 

(Appendix A, Table A-1). 

Although backswamp sites are few, difficult to locate, and usually nondescript in content, 

we do not assume that all such sites are insignificant. Small fishing stations, in particular, 

could be represented at favorable locations on oxbow lakes and bayous. Sites recorded 

on St. Mary’s Lake and Fishtrap Lake (the lower bayou of Lapile Creek) may represent such 

stations. Two sites on Fishtrap Lake (3UN179 and 3UN180) have been recommended for 

mitigation (Appendix C). 

Unit 7: Pleistocene Terraces and Islands 

The 27 sites included in this unit occur on elevated Pleistocene terraces or terrace 

remnants (“islands”) above the Recent floodplain, and their discovery or reinvestigation was 

secondary to the survey objectives discussed earlier in this chapter (see Elevation Constraint 

in Chapter 1). Project records for these sites are summarized in Appendix A, Table A-1, and 

are discussed briefly here. Detailed comparison among these sites or with floodplain site 

data is beyond the scope of this report. 

Major sites on elevated terrace surfaces are rarely subject to overflow, and are not alluviated 

as are all floodplain sites. Depth of material in these sites is not well known, but early 

components may underlie later ones as the result of local colluvial and cultural deposition. 

Mound and midden complexes at seven terrace or island locations were noted in Chapter 

3 (3BR4, 3BR8, 3UN18, 3UN8, 3UN9/52, 3AS1, and 3AS6). Well drained silt or sandy loams 

characterize the terrace edges, in contrast to poorly drained clay loams of the Recent floodplain. 

Soil conditions and overflow characteristics ensure the growth of oak-pine forest on 

terraces and islands, as opposed to the bottomland hardwood forest (Chapter 2). 

From the report of Rolingson (1972) and to a lesser extent from our own observations 

(Appendix A, Table A-1), it is clear that a wide variety of prehistoric sites occupy terrace 

locations in the Felsenthal region. Large multicomponent habitation sites are conspicuous, 

but smaller components, varying in size and intensity of occupation, are much more numerous 

and widespread in distribution. Systematic survey or testing in these small upland sites 

is regrettably lacking, and no good basis for assessing significance has been established as 

it has for certain large sites (Rolingson and Schambach 1981). The impact of surface disturbance 

of various kinds poses a much greater threat to small terrace surface sites than to 

buried floodplain sites, so that additional intensive survey of upland areas is badly needed 

in this region (Appendix C:General Recommendations).


Figure 31. Site Unit 6, oxbow lakes and backswamp, with prehistoric sites located 

in relation to floodplain and terrace features.


unit 8: historic sites 

A total of 15 historic sites was recorded for the entire project area by combined use of 

survey techniques, archival research, informants, and prior site records (Chapter 4). This 

very low number for so large an area reflects the limited use made of overflow bottomlands 

during the historic period and also the difficulty of locating small impermanent floodplain 

camps. In fact we were unable to record any early historic sites (before about 1780) and can 

bring no data to bear on Hypothesis 5 regarding such hunting or trading camps (Chapter 1). 

The sample of later historic sites (after 1780) associated with riverine activity and extractive 

industry is discussed in relation to Hypothesis 6 later in this chapter, although we have very 

little pertinent artifact data. 

Each of the 15 historic sites is summarized in Appendix A (Table A-2), and is described 

briefly below in terms of observed content, relevant archival or documentary sources, and 

research potential. 

Ganer Mound (3BR32). This square, flat-topped earth mound (ca. 50 x 50 x 3.5 m) was 

built at Stormhole Bend (Pereogeethe Shoals) about 1910-1912 as a staging area for construction 

of Lock No. 7, and presumably to raise the lockkeeper’s residence above winter-spring 

flood stage (Figure 24). Modification of Lock No. 6 at Felsenthal in 1912 eliminated the need 

for Lock No. 7, and no further improvements were made (Thatcher 1952:19). The government 

tract at Stormhole Bend is shown on the 1895 Ouachita River hydrographic survey 

(Sheet 10). The earth mound itself is shown only on the most recent U.S. Geological Survey 

topographic map at 1:24,000 (USGS Felsenthal NW, 1978 advance workprint). 

Ganer Mound was first recorded by Arkansas Archeological Survey in 1971, and was 

revisited by our boat survey team. A few oyster shells were collected on the mound flanks, 

and a massive cast iron machine component was noted. A modern corrugated metal cabin 

and associated debris occupy the mound summit. Other artifacts of the 1910-1912 construction 

period may be present at the water’s edge or adjacent to the mound and buried by silt 

and leaf litter. Ganer Mound pertains to river improvement and engineering activity early 

in this century, and thus has some interest for regional history of the Ouachita Valley. No 

mitigation recommendation has been made for this site, but destruction should be avoided. 

Caney Marie Landing (3BR56). This river landing of the late 1800s and early 1900s 

was located on a concave bank at the axis of a prominent bend (Figure 24). The place-name, 

spelled variously (“Marais de Cannes” was recorded by Hunter in 1804), once applied to a 

cane marsh, the riverbend, the landing itself, and shoals below the landing. We note here the 

location of four other riverboat landings at the axis of bends (3BR81, 3UN155, 3UN88, and 

3UN122). Apparently this location conferred an advantage by virtue of the deep scour pool 

near the concave bank.


Caney Marie Landing is shown on U.S. Army Corps of Engineers river survey sheets 

and index maps for 1873, 1895, and 1897, usually as a single small onshore rectangle (structure?). 

Caney Mary Shoal downstream was described as 8000 feet in length, 0.6 feet deep at 

low water (U.S. Congress 1874). This document also noted dangerous snags near the landing. 

We have no map or documentary references for continued use of the landing early in 

the twentieth century. However, our boat survey team recovered a 1905 Liberty Head dime 

and stoneware bottle sherd on the eroding bank, precisely at the nineteenth century map 

location (Figure 32a, b). A sandstone block eroding out of the bank was also noted, but no 

other traces of the structure(s) which may have existed here. Mitigation by data recovery, 

including additional archival research, is needed for this site which retains some evidence of 

early Ouachita River commerce (Appendix C). 

Goulett Island (3BR8). Goulett Island on the Saline River (Figure 26) has significant 

prehistoric remains known since the time of C. B. Moore (1913), but also a variety of historic 

structures and remains which are essentially unstudied. Arkansas Archeological Survey revisited 

and recorded Moore’s site in 1961 and 1971, and our survey team made another brief 

visit and supplementary site record in 1979. However, the bulk of Goulett Island lies above 

our project contour and we were not able to justify a comprehensive search or testing of this 

important locality. 

Goulett Island is the first high terrain, immediately adjoining the channel, which one 

encounters proceeding upstream on Saline River. One suspects that historic settlement began 

early, or may even have been continuous, after the prehistoric occupation of this “pine 

island.” However, Government Land Office maps of 1827 and 1830 show no settlement or 

agricultural fields until Godfrey Landing is reached 3 miles (4.8 km) upstream. David Dale 

Owen, the geologist, made note in 1860 of a lignite outcrop in the bed of Saline River at 

Goulett Island which was quarried for fuel by local inhabitants at low water. Land adjoining 

the river, evidently including Goulett Island, was owned by Arkansas Governor Elias N. 

Conway (served two terms 1852-1860) at this time (Ferguson and Atkinson 1966). Sections 

23 and 24 in T17S, R10W, which encompass Goulett Island and adjacent floodplain terrain, 

remained an isolated “panhandle” of Bradley County on the left bank of Saline River when 

Ashley County was formed in 1846. Conway’s influence may have affected this boundary. 

Archival research for this early to middle nineteenth century settlement and use of Goulett 

Island is badly needed. 

Stave cutting (white oak) by “Slavonians” was an important winter activity in the 

lower Saline River area from about 1890 to 1920 (Chapter 4). Logging of cypress first, and 

then pine and hardwoods, began in the early 1800s; logging continues today in various Saline 

River floodplain and terrace areas (Etheridge 1959; Reynolds 1980). This activity is likely 

to have involved Goulett Island as a logging camp and river landing.


Figure 32. Historic artifacts from floodplain sites. a. 1905 Liberty Head dime, 

Caney Marie Landing (3BR56); b. stoneware bottle sherd, Caney 

Marie Landing (3BR56); c. cast iron chain drive link, Keelboat Brake 

Sawmill (3BR55); d. whiteware plate fragment, Marie Saline Ferry 

(3AS331); e. stoneware jug handle, South of Highway 82 Bridge 

(3AS300) (AAS neg. 813882).


Our survey team recorded abundant, diverse historic refuse along the eroding riverbank 

at Goulett Island, including domestic utensils and debris greater than 50 years in age. 

Only a few selected items were collected (Appendix A, Table A-2). We also observed and 

photographed a small pine log barn with hand split boards on gable ends and shake roof, 

situated on a low aboriginal (?) earth mound near the riverbank. U.S. Geological Survey 

maps of the 1930s indicate one centrally located structure and agricultural (?) clearing of 

the island. We recommend that Goulett Island be intensively surveyed and tested in order 

to assess the significance of the complex of prehistoric and historic remains that occur here 

(Appendix C). The historic refuse on the riverbank is considered significant and is recommended 

for mitigation (Appendix C). This locality is evidently slated for development of 

boat access facilities by the Felsenthal National Wildlife Refuge (U.S. Fish and Wildlife Service 

1979). In Chapter 2 (Potential Paleoecological Studies) we noted also the importance of 

the geological exposure at Goulett Island. 

Prairie Island Landing (3BR81). This landing downstream from Goulett Island on the 

Saline River also contains both prehistoric and historic components (Figure 26). The historic 

component comprises a point at the axis of a bend where a track from Prairie Island 

intersects the riverbank (Deeter et al. 1925), and where at least one structure is located on 

topographic maps of the 1930s. Since the riverbank lies above our project contour, no intensive 

search was made for early historic remains and only modern fishermen’s debris was 

observed. 

Prairie Island Landing does not appear on Government Land Office maps of the early 

1800s, and we have no other nineteenth century Saline River maps available. Nor does it appear 

on C. B. Moore’s map of Saline River, which is, however, deficient of most detail (Moore 

1913:80). Prairie Island itself, a “pine island” elevated well above the floodplain, may be 

significant archeologically, historically, and even ecologically, and should be intensively surveyed 

(Chapter 2). At the present time we have no knowledge of significant historic remains 

at the riverbank landing, only indications from early twentieth century maps. 

Keelboat Brake Sawmill (3BR55). The remains of cypress logging, a sawmill, and a 

tramway leading to the sawmill in a backswamp area were pointed out to us by a local 

trapper, David Marshall (Figure 24). Cypress was being rafted out of this area, the “forks of 

Saline and Ouachita,” in 1855 and probably a decade or two earlier (Etheridge 1959:25). Old 

cypress butts and tops are lying along the margin of Keelboat Brake. This brake shows on 

Government Land Office maps of 1827 and later years. The sawmill remains, a burned slab 

pile 10 m in diameter and about 0.5 m high, were shovel tested and produced only charcoal. 

A cast iron, drive chain link was collected on the surface nearby (Figure 32c), and other artifacts 

and debris may be silted over in the vicinity. The low tramway road extending many 

miles north and northeast to the Chicago, Rock Island, and Pacific rail line shows up well on 

1951 air photos (Larance 1961:Sheet 39). This sawmill operation was evidently very small in


scale and was used decades later than the period of cypress rafting, perhaps in the early 

1900s. Most hardwood logs could have been skidded or hauled out of the floodplain via the 

tramway. The sawmill and associated remains at Keelboat Brake are potentially significant 

to the history of the development of forestry in this region (Reynolds 1980). We have recommended 

mitigation of the site, since it pertains to regional research questions concerning 

extractive industry (Appendix C). 

Ouachita Belle Landing (3UN155). This riverboat landing is known presently from 

map data only. The U.S. Army Corps of Engineers’ hydrographic survey for 1873 (Sheet 11) 

identifies the landing on the concave bank of Five Mile Bend, indicating possibly an offshore 

landing stage or raft. Choice of this landing location probably reflects the occurrence 

of a deep scour pool along the concave bank and the presence of high terrain less than a 

mile southwestward (Figure 24). Our survey team found no surface indication of the 1870s 

steamboat landing. 

The Ouachita Belle of the Blanks Line, largest steamboat on the river in the 1870s, was 

a sidewheeler 250 feet long, 67.5 feet at the beam, drawing 8.5 feet of water (Thatcher 1952). 

She made the New Orleans-Camden run only during high water, and was considered the 

“pride of the Ouachita.” The landing site should be monitored and protected from surface 

disturbance, since undiscovered remains may be present or may be revealed by ongoing 

erosion. 

Fletchers Landing and Woodyard (3UN88). This mussel shell concentration on the 

riverbank was recorded by Arkansas Archeological Survey in 1971 and revisited by our 

survey team. U.S. Army Corps of Engineers’ hydrographic surveys for 1873 and 1895 indicate 

the presence of a landing and woodyard here with one offshore landing stage or raft (?). 

The location on the concave bank at the axis of the bend, and in proximity to higher terrain 

southwest (Figure 24), is similar to that noted above for Ouachita Belle Landing. 

The only remains known at 3UN88 are eroding shallow lenses of mussel shell with 

rare pebbles and rocks, none attributable to prehistoric occupation. Four or more species of 

mussel have been identified (Appendix B). Modern fishcamp debris and logging roads and 

clearings are also present. The shell lenses are discontinuous over an area of about 40 x 60 

m adjoining the riverbank and a small tributary slough or crevasse. These shell patches are 

clearly not modern, but are not certainly associated with the late nineteenth century river 

landing either. This site principally occupies the levee crest above our project contour, and 

was not extensively searched or tested. We recommend continued monitoring and protection 

from surface disturbance. 

Brown Camp (3UN92). The historic site known as Brown Camp or Hogan Property 

was recorded by Arkansas Archeological Survey in 1971 and revisited by our survey team


which also recorded prehistoric remains (Figure 28). Brown Camp consists of a large earth 

mound surmounted by a log house near the riverbank, a borrow pit, and a second smaller 

earth mound about 150 m southward. The owner of this property, including adjacent boat 

launching facilities and riverbank lots, is Mrs. Louise M. Hogan of Little Rock, Arkansas. According 

to our informant, Mr. Bud Poole, who operates the boat landing, these mounds were 

built about 1930 when the approaches to the Highway 82 bridge were under construction, 

and the log house was built soon thereafter. Brown Camp lies above our project contour, and 

was not examined closely by the survey team. We did note prehistoric artifacts on the slopes 

of the larger mound which are redeposited with fill from the borrow pit nearby. At the present 

time neither the historic structures nor the prehistoric artifacts are considered to have 

research potential, although one or more buried prehistoric components, partly disturbed, 

may be present. Additional construction on this site should be preceded by subsurface testing. 

Marie Saline Landing (3AS299) and associated historic sites. The most important historic 

site in the Felsenthal Project area in terms of nineteenth century settlement and commerce 

was Marie Saline Landing on the left bank of the Ouachita River where U.S. Highway 

82 now crosses (Figure 28). This landing is listed in an early steamboat directory as having 

a population of 300 (Lloyd 1856). Habitation and farming, however, would have been on 

higher terrain nearby. The history of Marie Saline Landing was recounted briefly in Chapter 

4. Here we describe the limited historic remains which were recorded by our survey team at 

or near the landing location shown on historic maps, and note the significance of the locale. 

Site 3AS299 is directly identified with the landing, but represents only a small part of 

the area once utilized on and near the riverbank. It consists of a canal, 4.5 m wide, extending 

200 m due east from the river, partly obscured by a mud bar at the mouth. Waterlogged 

tree trunks, limbs, and debris are submerged in its shallow depression. This canal seems 

to be indicated on the U.S. Army Corps of Engineers hydrographic survey (1871, Sheet 6), 

reproduced as our Figure 16; one large rectangular structure and two smaller structures 

(?) were evidently situated near the canal’s north bank. Etheridge (1959:136) briefly noted 

this canal “leading from the river to each [warehouse] so that boats might reach [them] for 

loading and unloading.” The exact date and details of canal construction are not presently 

known. There is a good possibility that submerged artifacts, including even wood or other 

organic materials, are preserved in the mud fill of the canal depression. The adjacent terrain 

lies above our project contour and was not investigated extensively or tested. No standing 

remains of warehouses or other contemporary structures or surface refuse are known for 

the canal’s vicinity, although buried remains could be present. Much of the surroundings are 

forested, but dirt roads, clearings, and modern fishcamps are scattered through this area. As 

indicated in Appendix A, Table A-2, we regard the canal site as a significant vestige of Marie 

Saline Landing with excellent potential for regional archeological and historical study. The 

canal itself warrants further investigation and mitigation (Appendix C).


First Slough (3AS330) north of the Highway 82 Bridge is a small shoreline mussel shell 

lens with rusted iron, ceramics, and glass among the shell. Three mussel species have been 

identified (Appendix B). The historic artifacts collected are few and fragmentary, but are 

likely to have been contemporary with the nineteenth century occupation of Marie Saline 

Landing. Several heavy iron spikes and large rusted objects suggest that the remains at First 

Slough are not merely domestic refuse, but pertain to riverboat or landing construction of 

some sort. This mussel shell lens with artifacts is potentially associated with Marie Saline 

Landing, and should be protected from disturbance. 

Marie Saline Ferry (3AS331), not far downstream from First Slough on the same 

bankline (Figure 28), is the approximate location of a ferry crossing shown on 1873 and 1895 

Ouachita River hydrographic survey maps. Etheridge (1959:31) indicates that the owners of 

Marie Saline Landing were licensed to operate a ferry here in 1852. This ferry connected the 

Hamburg and El Dorado roads, and must have been one of the busiest crossings in the region 

(Chapter 4). Our survey team observed about 10 sandstone blocks at the water’s edge, 

and collected one large whiteware sherd (Figure 32d) of probable early twentieth century 

manufacture on the eroding bank. Much modern fishcamp debris is strewn along the shoreline 

at this location, and surface disturbance is more extensive than elsewhere in the vicinity. 

We cannot be certain that the remains noted pertain directly to the operation of the ferry, or 

that other remains of ferry and landing activity are not present. As with other areas of activity 

associated with Marie Saline Landing, this site should be protected and may also warrant 

more intensive survey. 

Site 3AS300, between the U.S. Highway 82 Bridge and the Marie Saline canal (3AS299), 

consists of a small mussel shell lens on the riverbank with an early twentieth century soft 

drink bottle buried among the shell and also one stoneware jug handle found on the eroding 

bank nearby (Figure 32e). There is no basis for attributing these remains directly to activity 

at Marie Saline Landing except that it falls within the landing area generally delimited by 

nineteenth century hydrographic survey maps. Modern fishcamps and debris are extensive 

along this shoreline. This small site is potentially significant if it can be linked with the history 

of the landing, and on that basis warrants protection and additional study. 

Our study of Marie Saline Landing rests on results of riverbank survey and preliminary 

search of historical sources. Except for the canal depression, all of the natural levee of 

Ouachita River, where nineteenth century landing structures and roads were situated, lies 

several feet above our project contour. We were thus unable to perform intensive survey or 

testing on this levee. The same may be said for elevated Pleistocene terrace surface, knownlocally 

as Parker Ridge, just eastward of the river, where portions of the Marie Saline settle-


ment must have stood. Much of the locale of Marie Saline Landing is today traversed by 

a highway, borrow pits, and dirt roads, but much more remains forested. The canal itself 

(3AS299) appears to be partly filled, but otherwise undisturbed. Submerged remains may be 

expected within its depression. We recommend that a program of intensive survey and testing, 

encompassing the extent of sites and structures associated with Marie Saline Landing, 

be implemented in the near future. Additional construction on the Highway 82 bridge and 

eastern approach, but especially a proposed “Crossett Harbor” project, pose serious threats 

to the remains of Marie Saline Landing (U.S. Army Corps of Engineers 1979). In addition to 

the historic remains, we have recorded two small prehistoric sites (3AS301, 3AS319) on the 

riverbank within this area (Figure 20). The four historic and two prehistoric riverbank sites 

in this reach of Ouachita River were all previously unknown. 

Artesian Well (3UN174). An artesian well in the Ouachita River floodplain near Open 

Lake (Figure 28) is indicated on U.S. Geological survey maps of 1934 and later years. Our 

survey team visited the location in the course of completing Transect 1. The well site consists 

of a low, cylindrical, brick and concrete tank (some inscribed bricks were recorded; see site 

form for 3UN174). A metal shed and modern debris are also present near this brick structure. 

The date of development and intended use of this well are unknown, although local 

drillers’ records might contain such information. This well does not appear in a compilation 

of wells and springs by Veatch (1906:Chapter 5). Insofar as it is now known, this well site 

does not appear to have historical significance. 

Grand Marais Landing (3UN122). This late nineteenth century riverboat landing, at or 

just above the axis of Coon Glory Bend on the right concave bank (Figures 16, 28) was first 

recorded by Arkansas Archeological Survey in 1976, but its nature and significance were 

only established by our work with various historical sources. U.S. Army Corps of Engineers’ 

hydrographic survey maps for 1871, 1873, 1895, and 1897 show “Grand Mary Landing,” 

usually with one or two riverbank structures just above the axis of the bend. The deep scour 

pool would have provided sufficient water for steamboats to tie up along the concave bank. 

Higher, permanently settled terrain (present-day Felsenthal), was less than 2 km southwestward, 

across the Grand Marais backswamp. 

An artificial depression, bricks, and mussel shell were reported during the initial site 

visit, but not relocated by our survey team. However, we did examine, record, and shovel 

test a thin shallow buried organic lens, 25 x 40 m, partially exposed in the cutbank. This 

lens at about 64.0 feet MSL lies immediately adjacent to the upper approach channel to the 

new Felsenthal Lock and Dam, and may be affected by construction, as may other unknown 

buried remains. One split pine board or stave fragment (Appendix B) was recovered from 

a shovel test, but no other artifacts were recovered here. No surface indications of the landing’s 

structures were noted in the vicinity of the bend.


Systematic subsurface testing of this area was scheduled by us for fall 1979, but was 

prohibited by rising river level (Chapter 6). We believe that other remains of Grand Marais 

Landing may be present below the levee surface, and that historical significance of the site is 

established by our testing and archival research (Appendix C). A specific recommendation 

to this effect has been submitted to the U.S. Army Corps of Engineers (C. R. McGimsey letter 

to Shelia Lewis 29 August 1980). 

Wreck of the Lotawanna (3UN153). The 1874 loss of this steamboat in the Felsenthal 

Project area was discovered from historical sources prior to fieldwork, and has subsequently 

been confirmed and investigated by various methods. Only after much other field and archival 

work was undertaken did we acquire a hydrographic survey map which pinpoints the 

underwater wreck about 200 m below the axis of Coon Glory Bend (Figure 33). In September 

1979 we interviewed an elderly Felsenthal man, Roland Johnson, who had considerable 

knowledge of local river landings and who had seen the submerged wreckage of Lotawanna 

as a child (transcript on file at Arkansas Archeological Survey; see also Chapter 4). In May 

1980 one of us (Hemmings) then accompanied a marine survey archeologist who obtained 

magnetometer data for the wreck site (a marked anomaly matching the map data, presumed 

to be the Lotawanna in about 24 feet of water; information from Allen R. Saltus, Jr., Archeological 

Research and Survey, Prairieville, Louisiana). The wreck has not yet been directly 

observed. 

The underwater Lotawanna wreck site is located nearly within or just below the upstream 

approach to the new Felsenthal Lock and Dam, so far as we can tell now. Precise 

mapping of this wreck in relation to river channel and construction right-of-way has been 

undertaken by Saltus, using both magnetometer and bathymetric data. We also have examined 

a 1976 construction plan with bathymetric contours which may indicate the wreckage. 

It is expected that lock construction and deep inundation will have an adverse impact on the 

remains of Lotawanna (Appendix C). On this basis we have recommended to the U.S. Army 

Corps of Engineers that a qualified diver-archeologist inspect the wreck site, establish its extent 

and integrity, evaluate impact, and recommend mitigation procedures (C. R. McGimsey 

letter to Shelia Lewis August 29, 1980). At the present time the Lotawanna is the only steamboat 

wreck, pinpointed and archeologically recorded, on the Ouachita River in Arkansas. 

eVAluAtion of surVey strAtegy And tActics 

Intensive archeological survey was undertaken in the Felsenthal Project area in order 

to locate cultural resources present, assess their cultural and scientific importance, and 

identify means for mitigating adverse effects of a 65-foot navigation pool and other related 

impacts. A seven-week survey stage was first completed, involving a successive series of 

survey procedures and target areas in the floodplain; survey activity then continued on a 

reduced basis during a seven-week test excavation stage reported in Chapter 6. A systematic


Figure 33. Hydrographic survey chart of Ouachita River in the Felsenthal Project area, showing location of “Wreck of 

Lottie Warner” (Lotawanna) (U.S. Army Corps of Engineers 1896:Sheet 14).


body of survey data for 144 prehistoric and historic sites has been summarized in this chapter 

and in Appendix A. Much new substantive data and a variety of new interpretations 

have resulted from this work. These results pertain to both scientific and cultural resource 

management objectives of the Felsenthal Project, and to many aspects of prehistory and 

history in the Felsenthal region. In this section we briefly evaluate our strategy and tactics, 

especially in regard to survey capability in a remote wetland environment where resources 

lie buried beneath the modern surface. 

1. The project area was stratified into two subareas, and one of these subdivided 

as four sampling units for the purpose of random transect survey. This stratification 

was based principally on the constraint of project contour (terrain below 

65 feet MSL) and secondarily on geohydrological variation between floodplain 

areas. Ideally, a continuous block or segment of floodplain terrain would serve 

as the project area, rather than the mosaic of patches or ribbons of terrain with 

which we have had to deal (see Chapter 1). 

2. Environmental variation within the floodplain was poorly known at the outset 

of this study and was integrated into survey strategy with difficulty. Detailed 

studies or maps of biotic and abiotic variables in the Felsenthal floodplain do 

not exist. Guyton silty clay loam, bottomland hardwood or swamp forest, and 

minimal relief characterize the entire project area. However, a variety of drainage 

features of geologically Recent origin are present, and are differentiated 

generally by Fleetwood (1969). We now perceive significant microenvironmental 

variation between the inner valley natural levee zone, poorly developed as 

it is, and the outer valley backswamp zone, and also between the upvalley and 

downvalley sections of the floodplain. Site distributions are believed to reflect 

this variation (see Site Locational Hypotheses below). 

3. Floodplain transects involved point sampling (augering or shovel testing) to 

0.5 or 1.0 m depth and at 50 m intervals. Four transects were randomly located 

and nine others judgmentally located. These transects largely intersected backswamp 

portions of the floodplain and proceeded with difficulty. From results 

of all floodplain survey work we now know that eight of 10 sites are likely to 

extend less than 50 m in greatest dimension, and also that some sites are buried 

at depths greater than 0.5 m. The deficiency in survey intensity is obvious. However, 

increased intensity and therefore cost of survey would not be justified on 

the basis of our present understanding of backswamp sites and site distribution. 

Transects 9 and 10 which provided continuous exposure of sediments to at least 

1.0 m depth, 8 km across the floodplain, indicate that backswamp sites are few, 

small, and sparse in content.


4. Adequate procedure for rapid, extensive, subsurface testing in this floodplain 

context has yet to be developed. We are convinced that power augers are inefficient 

devices for locating sparse remains to 1.0 m depth, although they penetrated 

tough clayey sediments with ease. Shovel testing proved effective to 0.5 

or 0.6 m depth, but decisions to employ hundreds of time-consuming shovel 

tests should be carefully considered. Manually operated posthole diggers, 

said to be effective to 1.5 m or more (Wood 1976), are by no means well suited 

to tough clays. Lack of quick, effective, and nondestructive testing procedure 

prevented us from obtaining precise areal extent of buried sites in most cases. 

Portability is a requirement for any testing device used extensively in these 

overflow bottoms. 

5. Riverbank survey tactics were carefully conceived and consistently carried 

out with good results. An understanding of meander activity, prograding, 

degrading, or stable banklines, alluviation, and other local floodplain or river 

channel processes is essential to designing and evaluating riverbank surveys. 

The procedures we employed in the Felsenthal Project area would not be well 

suited to the dynamic floodplain regime in the Great Bend region of the Red 

River (Hemmings 1980). We were not able to devise a precise workable method 

of evaluating and recording degree of bankline exposure within various channel 

reaches. The record of small ephemeral riverbank sites of low visibility is 

presumably strongly biased by exposure. 

6. Among other site survey data, emphasis was placed on recording the depth of 

buried artifacts or debris in place on prehistoric riverbank sites. About onethird 

of all floodplain sites furnished depth observations, and about two-thirds 

of all floodplain sites furnished typological, technological, or stratigraphic data 

sufficient to identify cultural components. We are able to advance a sequence of 

floodplain sites and components and general rates of floodplain alluviation for 

the past 3000 years (Chapter 7). We believe this is a significant contribution to 

regional prehistory. 

7. We do not calculate a sample fraction from the extent of transect and river-bank 

survey and the total extent of project area. Given the mosaiclike distribution 

of terrain below 65 feet elevation, the uniform occurrence of buried sites, and 

the tactics of survey such a calculation would be intricate and spurious. Nevertheless, 

the empirical data for distribution of archeological sites in the upper 

1.0 meters of floodplain sediments, especially in the stable natural levee zone 

of Ouachita and Saline rivers, provides a systematic basis for evaluating these 

resources.


site locAtionAl And historicAl hypotheses 

Site survey data, especially those data obtained by riverbank survey, are relevant to 

most of the hypotheses proposed in the research design (Chapter 1). We shall here examine 

two hypotheses regarding the location of prehistoric sites and the function of historic sites in 

the floodplain. The problem domains and hypotheses were stated as follows: 

Problem Domain 2: Locational Characteristics 

H 2 

If specialized extractive sites were established for resource procurement, these sites 

would be located for most efficient acquisition and delivery to base settlements. 

Subhypotheses: A. Such sites will correlate with specific resource zones. 

B. Such sites will correlate with access routes to base settlements. 

C. Such sites will correlate with both factors. 

D. None of the above, but other unknown factors. 

Test Implications: Significant correlations obtained for location factors; recognition of 

contemporary base settlements, access routes, and resource zones. 

Data Requirements: Map distributions; measures of spatial relationships; cross-dating. 

Problem Domain 3: Culture History 

H 6 

If later historical sites are present in the floodplain, they were occupied in association 

with riverine activity and extractive industry (after 1780). 

Test Implications: Discovery of sites with complex technological assemblages. 

Data Requirements: Tools or components associated with engines. 

Hypothesis 2 seeks to explain the distribution of floodplain sites in terms of how a 

resident local prehistoric population might best acquire riparian and aquatic resources. Multiple 

variables are involved in the several subhypotheses and also in test data we can bring 

to bear, so that elegant proof of this hypothesis is not presently possible. We believe that 

nonrandom distribution of riverbank sites can be established and that locational factors can 

at least be systematically examined. The most marked departures from random distribution 

are relatively high frequencies of Archaic components in Unit 1 (at least partly due to deep 

exposure), of Tchula and Baytown-Coles Creek components in Unit 4, and of Mississippi 

period components in Unit 2. By far the best available site contextual data pertain to the 

Mississippi period, where:


1. visibility of shallow buried components is good, 

2. more floodplain components can be assigned to this period or its three phases, 

3. the nature of extractive sites has been investigated (Chapter 6), 

4. the presumption of stable channels and stable levee environment over the past 

several centuries is relatively strong, and 

5. some understanding of base settlements on terraces and islands can be advanced. 

In order to emphasize the striking concentration of Mississippi period components on 

the lower Saline River we present some simple density calculations in Table 8. On the average, 

nine such components occur here within a 10 km reach of the river. We know also from 

other survey and test excavation data (Chapter 6) that multiple activity loci characterize 

some of these Mississippi period components on the lower Saline River. The densities for 

Ouachita River components attributed to this period, however, are markedly lower where 

such components have been discovered at all. Our locational hypothesis and any other that 

might be formulated must contend with this dense cluster within a complex regional system 

of Mississippi period settlements. 

Table 8. Density of Prehistoric Riverbank Components in Three Site Units on Ouachita and 

Saline Rivers. 

All Prehistoric Linear Density Mississippi Period Linear Density 

Site Unit and Length of Unit Components (Components/ Components (Components/ 

Location (river km) Recorded 10 river km) Identified river km) 

unit 1 

(Ouachita River 

above Mile 254) 27.4 18 6.6 0 0.0 

Unit 4 

(Ouachita River 

below Mile 254) 18.8 23 12.2 5 2.7 

unit 2 

(Lower Saline River) 17.7 41 23.2 16 9.0 

Figure 34 illustrates the spatial array of Mississippi period components in or near the 

Felsenthal Project area. In Chapter 3 we proposed an hierarchy of Mississippi period settlements 

where major and minor mound centers on terraces and islands are reasonably well 

known, but where habitation sites (villages, hamlets, and farmsteads) on terraces are very 

poorly known. We therefore have incomplete knowledge of the Mississippian settlement 

hierarchy and of the nature and distribution of base settlements, and this bias is reflected in 

Figure 34. The seven mound centers indicated around the periphery of the floodplain are


Figure 34. The spatial array of Mississippi period sites in or near the Felsenthal 

Project area.


not merely ceremonial loci for a dispersed population, but evidently were themselves occupied 

by a sizable population segment as indicated by extensive dense refuse and, in some 

cases, multiple cemeteries (Rolingson and Schambach 1981; Arkansas Archeological Survey 

site records on file). The very large Mississippi mound and midden complex at Watts Field 

(3UN18) was illustrated in Figure 14. 

Inspection of the spatial array in Figure 34 reveals several very interesting characteristics. 

First, there is a pattern of base settlements arranged around a cluster of extractive 

camps (our Unit 2), rather than extractive camps arranged about base settlements. This “centripetal” 

pattern is clear with respect to floodplain site data, but lacks data for upland extractive 

sites. Second, there may be regularity in spacing among Mississippi period settlements; 

the seven mound centers average about 7 km distance between each pair, and the shortest 

distance between any mound center and the lower Saline River cluster of extractive sites 

also averages about 7 km. Variations from these average distances are not especially great, 

and are apparent in Figure 34. We suggest that cultural factors primarily, not environmental 

differences, have operated to produce this spatial array. 

Some, and conceivably all, mound centers in the project area were occupied contemporaneously 

during the Gran Marais phase (A.D. 1100-1400), and interacted within a regional 

sociopolitical and economic network. At least some of the larger mound centers seem to 

have been occupied also during the Caney Bayou phase (A.D. 1400-1600). When these base 

settlements (or others along the terrace edges) undertook to exploit floodplain resources 

within a radius of a few kilometers, no temporary camp would be required; when the radius 

was extended to approximately 7 km (or more) a work party encamped in the floodplain 

to maximize its extractive and processing activity, and minimize unproductive daily travel. 

That is to say, a principle of least effort for maximum return is reflected in the spacing of 

floodplain extractive sites and base settlements on adjacent terrace edges. Cooperative procurement 

activities among base settlements might also have influenced this spacing. 

Returning now to Hypothesis 2, we would argue that test implications are generally 

supported by the Mississippi period settlement data. Extractive sites are highly correlated 

with a resource zone, the river channel and its levee margins, where important faunal and 

floral resources are concentrated (Chapter 2). River channels and their tributaries provided 

easiest access, across intervening backswamp, to base settlements on elevated terraces. 

An “unknown factor” is also implied by the concentration of extractive sites on the lower 

Saline River, a relatively remote interior floodplain zone with respect to the peripheral ring 

of base settlements. We tentatively attribute this “centripetal pattern” to resource procurement 

efficiency. Recognition of this pattern is one of the most important research results of 

our floodplain survey work. Biases and deficiencies in regional Mississippian settlement 

data are great, but a productive direction for site survey work and spatial analysis is clearly 

indicated.


In Hypothesis 6 we proposed that later historical sites in the floodplain would reflect 

riverine and extractive activity. The data requirement, tools or components associated with 

engines, is minimally met by the metallic remains collected or observed at Ganer Mound 

(3BR32), Keelboat Brake Sawmill (3BR55), and First Slough (3AS330). However, the location 

and function of most other historic sites is documented by maps and other historical sources; 

steamboat travel, logging, or other riverine commerce is clearly associated with nine 

of 15 historic sites recorded and not associated with two sites (Brown Camp [3UN92] and 

Artesian Well [3UN174]). The hypothesis is well supported by all available data, but not by 

the few tools or engine components we were able to locate and recover in floodplain historic 

sites. 

One other result of historic site survey was the correlation of river landings with 

meander geometry. These landings were consistently located on the concave bank at the 

axis of a bend, in order to take advantage of the deep scour pool adjoining this bank. Bends 

were selected for their proximity to upland terrain, which minimized the difficult crossing 

of backswamp. From nineteenth century hydrographic surveys and other maps we believe 

these preferred landing locations are repeated upstream on the Ouachita and Saline rivers.

Chapter 6 

AssessMent of floodplAin sites 

On October 6, 1979 the testing stage of the Felsenthal Project was initiated with a daylong 

training session for all personnel on test excavation objectives, procedures, and equipment. 

Survey activity was to continue on a reduced basis on the lower Saline River and the 

Ouachita River below mile 254, consistent with the needs of site testing. The testing stage 

was projected to end in mid-December, but on November 24 the rivers rose above 65 feet 

and never again dropped below 64 feet during 1979. This curtailed the testing stage, prevented 

satisfactory completion of testing then underway on one site, and, of course, reduced 

the number of sites tested and data available for floodplain site assessment. 

Testing proceeded without interruption for seven weeks, with one to three test excavation 

crews and one boat survey crew deployed, often simultaneously. Nearly all floodplain 

sites were reached most efficiently by boat, and are essentially otherwise inaccessible in wet 

weather. We therefore employed two boats on a busy schedule to deliver crews and equipment 

and to extend bankline surveys. We also necessarily scheduled simultaneous testing 

operations on sites close to one another, although no site was selected for testing on this 

basis. Considerations of safety and logistics are involved in moving the testing operation 

between site clusters, that is, in centralizing effort at intervals within the testing stage. 

In the following section we are at pains to describe the process of selection of sites for 

testing and limitations of this process. Clearly the results of testing are critically important to 

assessment and management of floodplain sites, including the direction that future archeological 

work may take in this project area and region. 

selection of sites for testing 

Decisions about selection of sites for testing were made when substantial survey results 

were available, including both field observations and field laboratory processing of collections. 

In accordance with contract requirements and our project research design, testing 

was conducted to assess the significance of selected archeological resources in the context of 

certain research questions and hypotheses. Our criteria for selection were based on site survey 

data recovered and also on artificial constraints of location and accessibility. Thus any


site selected for testing has produced or is expected to produce at least three of the following: 

1. cultural remains below the 65-foot contour, whether or not this material is 

known to be in place 

2. one or more substantial samples of artifacts and debris 

3. temporally or functionally diagnostic artifacts 

4. superposed assemblages of artifacts 

5. one or more samples of organic remains 

6. subsurface or above-ground structural remains or features 

7. data representative of a larger functional, temporal, or areal site group 

8. data evidently unique to the locality or region 

9. data bearing on specific research questions 

10. also any site selected must be reasonably and safely accessible to personnel 

and equipment. 

Consultation with the principal investigator, station archeologists, and other personnel 

closely involved with the project contributed to the decision process. Although criteria 1 and 

10 were considered necessary, no other weight or order of importance was ascribed, so that 

judgment of the project archeologist and staff was ultimately exercised. 

At the outset of the testing stage 16 floodplain sites were selected and information 

pertaining to these sites (location, elevation, size, observed content, proposed work) was 

submitted to the U.S. Army Corps of Engineers. The total number of sites indicated for testing 

was based on our initial proposal, our experience with floodplain mobility and logistics, 

and on the length of time scheduled for testing (curtailed as noted above). It should be 

emphasized here that many other floodplain sites meet some of the criteria, and a few sites 

have emerged as potentially significant resources during the last survey efforts (e.g., Marie 

Saline Landing, 3AS299) or as analysis proceeded. One such site, the Wreck of the Lotawanna 

(3UN153), described previously (Chapters 4, 5) has now also been recommended to the U.S. 

Army Corps of Engineers for assessment by special techniques. The standing list of 17 floodplain 

sites selected for testing, including those eight sites (indicated by asterisks) described 

in the remainder of this chapter, is as follows: 


*Mouth of Eagle Creek (3BR78) 

*Jug Point 1 (3AS307) 

*Jug Point 1 (3AS306) 

*Jug Point Cutoff (3BR76) 

*One Cypress Point (3AS286) 

*Buttonbush (3BR58) 

*False Indigo (3AS285) 


Eagle Creek 4 (3BR66)

Unit 4: Ouachita River below Mile 254 

*Marie Saline (3AS329) 

Water Elm 3 (3AS287) 

Coon Glory Bend (3UN168) 


Lapile Creek 5 (3UN157) 

Lapile Creek 7 (3UN151) 

Unit 6: Oxbow Lake and Backswamp 

St. Mary’s Lake North (3UN187) 

Unit 8: Historic Sites 

Keelboat Brake Sawmill (3BR55) 

Grand Marais Landing (3UN122) 

Wreck of the Lotawanna (3UN153) 

Assessment of Floodplain Sites 157 

Under ideal field conditions, we would have expected to test minimally six to eight 

more sites, not including the Lotawanna, since most of the remaining ones (not indicated by 

asterisks) are small in extent and relatively accessible. Also, it would have been desirable to 

accomplish testing on the historic sawmill and landing sites. 

test excAVAtion strAtegy And tActics 

A consistent set of test excavation procedures was briefly outlined in the research 

design, introduced during a field training session, and implemented on the first site tested 

(3AS329) under the direction of the project archeologist. All succeeding test excavations employed 

these procedures under the direction of crew chiefs with minor adjustments for field 

conditions. Testing strategy required not only judicious selection of sites, but also efficient 

allocation of testing effort within sites and management of excavator’s bias. These requirements 

are relevant to all testing work, but assume added importance where sites are buried 

through most of their extent, as are in fact all prehistoric floodplain sites recorded by us 

(Chapter 5). Table 9 summarizes test excavation procedures as proposed in outline and then 

implemented or adjusted in the field. (This summary does not address the contingencies of 

test excavation in historic sites with structural or otherwise complex remains.) 

We turn now to summaries of results of test excavation for eight floodplain sites in 

the order in which testing proceeded. All of these sites are newly discovered and recorded. 

These summaries emphasize data pertaining to the physical context, lithic and ceramic 

analyses, and other specialized analyses, as well as inferences about site functions and age 

of components represented. Additional site data may be found in Appendixes A and B and 

in Chapters 7 and 8. Location maps for these eight sites include Figures 26 and 28.


Table 9. Summary of Test Excavation Tactics for Floodplain Sites. 

Tasks Proposed Tactics Implemented 

1. Area tested to approximate 5% of site Area tested approximated 5% of maximum site extent 

extent known from survey 

2. Site boundaries to be established by Site boundaries and depths established approximately by 

augering shovel testing (60 cm) at close 

regular intervals; contents of all holes screened 

3. 1 m 2 test pits located randomly within 1 m 2 test pits located systematically along a baseline, 

site boundaries randomly within site boundaries, and judgmentally where 

subsurface remains already exposed (on small sites only 

judgmental or systematic test pits employed) 

4. Test pits to be dug in 10 cm levels, Test pits dug to sterile soil or water table as proposed; 

screening all fill through ¼ inch (6.4 mm) because of clay content, all fill screened thoroughly and 

hardware cloth coarse residue searched and dumped 

5. Collection of radiocarbon and other Opportunistic collection of charcoal for radiocarbon and 

datable samples wood identification; no archeomagnetic samples available 

from any site 

6. Collection of soil and constant volume Limited program of soil, sediment, and wet screen or sieve 

flotation samples samples from all sites 

7. Field laboratory processing Field laboratory washing, sorting, and cataloging of all 

collections as received 

8. Instrument mapping Surface features, artifact concentrations, baseline, and test 

excavation units mapped for each site with light portable 

transit and tape; elevations determined approximately by 

reference to river level and gauge reading; semipermanent 

benchmarks established 

9. Standard recording Site form, field notes, level forms, profiles, site maps, photo 

record, lab record compiled for all sites

Marie Saline (3AS329) 


The Marie Saline site (3AS329), located on the low natural levee of the Ouachita River 

(Figure 35) below the entrance to the Marais Saline Lake (named by the French before 1804) 

and its intricate system of sloughs, is a relatively large, deep, multicomponent, and stratified 

site. It is in fact the largest and most complex floodplain site now known in the project area. 

We suggest the possibility that Marais Saline Lake represents the former mouth of the Saline 

River, and that early occupation here between about 1200 B.C. and A.D. 1 was contemporary 

with this confluence of rivers. Although Fleetwood (1969) does not make this distinction, he 

does present geomorphic and subsurface data for former courses of the Saline River at or 

near Marais Saline. Future work in this site should address this geohydrological problem 

(Appendix C). 

During initial survey and later testing, the Marie Saline site was recorded and collected 

as three adjacent and distinctive occupation areas (Figure 36). These artifact scatters extend 

over 200 m along the riverbank and constitute a very narrow elongate configuration insofar 

as now known. The bankline has a maximum height of 1.6 m (62.2-67.5 feet MSL). Artifacts 

or debris have been recovered in place throughout this vertical span, albeit sparsely, and 

from the modern levee surface, especially in Area C upstream. The following archeological 

periods and phases appear to be represented by one or more components at the Marie Saline 

site (other components may be identified in future work): 

Mississippi period, Caney Bayou phase (A.D. 1400-1600) 

Baytown-Coles Creek period, phases unnamed (A.D. 300-1100) 

Tchula period, Coon Island phase (500 B.C.-A.D. 1) 

Poverty Point period, Calion phase (1200-500 B.C.) 

Area A. The presumed extent of this area was sampled by means of systematic (10, 

11) random (14, 15) and judgmental test pits (12) to sterile soil at about 0.6 m and six auger 

holes to 1.0 m depth (Figure 36). All of the test pits produced sparse artifacts and debris 

from the upper half meter of levee deposit which are best attributed to Baytown-Coles Creek 

occupations of low intensity. A novaculite dart point stem, Baytown Plain sherds, and chert, 

novaculite, and clear crystal quartz debitage were the principal materials recovered. No 

organic remains were noted or recovered except in Test Pit 14 which intersected a modern 

disturbance. All auger holes were sterile, presumably because this technique is ineffective 

under conditions of low artifact density. Based on test excavation and good riverbank exposure 

from levee surface to river stage, no cultural materials are known to be present below 

0.6 m depth. 

Area B. This small area produced two notable concentrations of chert debitage, one 

broadly scattered by erosion and disturbance, but another dense cluster remaining largely 

in place at about 0.5 to 0.6 m depth. A systematic test pit (9), a judgmental test pit (13) in the 

remaining cluster of chert debitage, and four auger holes were emplaced in Area B (Figure 

36). Surface collection from the eroded scatter (shown to the north of Test Pit 9) recovered a


Figure 35. Aerial view of the Ouachita River-Marais Saline Lake locale and the 

location of Marie Saline site (3AS329) (air photo December 1195, 

courtesy of U.S. Army Corps of Engineers) (AAS neg. 814150).


Figure 36. Marie Saline (3AS329), test excavations, October 1979.


Late Archaic dart point and two Mississippi period arrow points, chert debitage, and Baytown 

Plain sherds (Figures 58e, 59q in Chapter 8). Test Pit 9 produced only a few novaculite 

flakes, while all auger holes were again sterile to 1.0 m depth. This heterogeneous assemblage 

suggests to us localized occupation and associated flintknapping during Baytown- 

Coles Creek time, and subsequent low intensity use late in the Mississippi period. No 

cultural remains were observed below 0.7 m depth, and the Archaic dart point is quite likely 

out of context. Area B is partly disturbed by vehicle and boat launching tracks as indicated 

in Figure 36. 

Area C. Area C furnished key stratigraphic data for this site and for the project area 

as a whole. During initial survey, fire-cracked rock and debitage were observed in the riverbank 

at or near river level, extending intermittently the length of Area C (130 m). This material 

was subsequently mapped and elevations obtained (Figure 36). Similar debris and two 

dart point fragments were found eroded out and concentrated at the water’s edge. Testing 

by means of six systematic test pits (1-6), two random test pits (7, 8), and 15 auger holes was 

designed to intersect a presumed Archaic horizon at 1.0 m or greater depth below the levee 

surface. Younger and stratigraphically higher ceramic components were expected on the 

basis of a few Baytown Plain sherds from the riverbank and two shell-tempered plain sherds 

eroding out at less than 0.2 m depth. 

All Area C test pits were excavated to depths of 1.2 to 1.5 m where ground water was 

encountered, and all produced artifacts or manuports in the deepest levels. The nature of 

topstratum deposits and stratigraphic detail were similar among test pits and will be summarized 

here by reference to Test Pits 1 and 6, for which soil and sediment analyses are 

available (Chapter 7). These test pits are separated by 100 m along the axis of the levee 

surface, which lies about 0.4 m lower at Test Pit 6 (Figures 37, 38). The topstratum deposits 

intersected are fine grained loams and clays, predominantly gray in color, and massive 

to weakly laminated in appearance. Erosional or nondepositional breaks, paleosoils, and 

obtrusive cultural zones or “floors” are missing in these test pit profiles (and others) with 

one major exception. This exception is a thin, dark gray, organically stained zone, associated 

with relatively dense, white novaculite debitage, at about 0.9 to 1.0 m in Test Pit 6 (Figure 

38). (Novaculite is defined in Chapter 8.) Small quantities of fire-cracked rock, fired clay, and 

wood charcoal, but no sherds, have been recovered from this zone. Comparison of the depth 

and artifact content with other test pits indicate that the age of this zone is either late Poverty 

Point or early Tchula period. 

Published soil survey data contribute some additional insights to our observations 

on topstratum deposits. Marie Saline occupies Guyton soils, which are extensive, poorly 

drained, slowly permeable, siliceous silt loams and silty clay loams derived from forested, 

Coastal Plain uplands (Gill et al. 1979:Sheet 40). They are strongly or very strongly acid, 

moderate to low in organic content, and low in natural fertility. Gleying, or reduction and 

transfer of iron, and subsoil mottling resulting from intermittent waterlogging are charac-

A b 


Figure 37. Marie Saline (3AS329), deep test pits profiles. a. Test Pit 1; b. Test Pit (AAS neg. 

797140, 797146).


Figure 38. Marie Saline (3AS329), profile records for Test Pits 1 and 6 (south walls).


teristic, and are evident in our profiles. Soil horizons are faint and weakly developed because 

sediments are of recent origin and are deposited at frequent intervals (Larance 1961: 

60). 

The cultural stratigraphy obtained by excavation of Area C test pits is excellent, but 

not complete in all details. Figure 39 schematically presents this stratigraphy by means of 

artifact and debris densities in 10 cm levels. Five categories of imperishable artifacts and 

debris, which occur frequently in this and other floodplain sites, are indicated: sherds, stone 

tools, debitage, fire-cracked rock, and particles of fired clay. (We will utilize these categories 

here and elsewhere in this report for comparison within and between sites. Frequencies of 

material in these categories and identification of temporally or functionally distinctive artifact 

types are employed with other conjunctive data in these comparisons.) An additional 

category of material, manuports, is omitted here (and in later comparisons) because human 

introduction of these objects, usually unmodified pebbles, cannot be established with certainty 

(see also Chapter 8). We can say that occurrence of 273 pebble “manuports” recovered 

in Test Pits 1-8 is strongly, but not perfectly, correlated with other artifacts and debris. 

The interpretation of cultural stratigraphy in Area C is therefore summarized as follows: 

1. A preceramic zone is probably present in all test pits below about 1.1 or 1.2 

m and lies partly below ground water or river stage; the principal artifacts 

recovered are two steatite sherds in Test Pit 8 and a muscovite schist slab in 

Test Pit 3, these materials being of presumed Appalachian origin (Figure 64a, 

b); novaculite flakes and fire-cracked rock are again associated; the dates for 

preceramic occupation here should lie between 1200 and 500 B.C. 

2. A definite early ceramic zone appears at about 0.8 to 1.2 m in Test Pits 1 and 

7 (shaded levels in Figure 39); this zone is marked by eroded, soft paste, 

Tchefuncte sherds, probably including Tchefuncte Incised and other types 

(Figure 65e); the low frequency and eroded, friable condition of sherds suggest 

that this zone is present, but not distinguished by us, in other test pits; a 

few novaculite flakes and considerable fire-cracked rock are associated; the 

dates for early ceramic occupation here should lie between 500 B.C. and A.D. 

1. 

3. A late ceramic zone occurs as a diffuse scatter of Baytown Plain sherds and a 

few chert and novaculite flakes from about 0.2 to 0.7 m in depth (dashed line 

in various test pits, Figure 39); a contracting stemmed dart point (Figure 59b) 

is evidently associated with a single Baytown Plain sherd in Test Pit 5 at 75 

cm; the dates for sporadic occupation here during the Baytown-Coles Creek 

period should range from about A.D. 300-1100.


Figure 39. Cultural stratigraphy in Area C, Marie Saline (3AS329). West walls of Test Pits 1-8 indicated with artifacts 

and debris counts by 10 cm levels. Within each test pit profile the five columns of figures represent sherds, 

stone tools, debitage, fire-cracked rock, and fired clay particles respectively. Superscripts on artifact counts as 

follows: P=dart point; M=muscovite or talc schist slab; S=steatite sherds (these items appear in Figures 59b, c; 

64a); dashed horizontal lines equal the position of lowermost Baytown Plain sherds. Hachured blocks in Test 

Pits 1 and 7 represent probable earliest ceramic occupation levels with Tchefuncte sherds (one of these sherds 

appears in Figure 65e). NP=new navigation pool elevation (meters).


4. The sparse Mississippi period occupation known to be present at about 0.2 

m (from surface finds) is not distinguishable in these test pits: dates for this 

occupation are believed to be late (Caney Bayou phase) or about A.D. 1400- 

1600. 

In summarizing Area C cultural stratigraphy we have not specifically identified the 

component represented obtrusively by organic matter and novaculite debitage in Test Pit 6 

at 0.9-1.0 m. On the basis of correlation among test pits, this zone may either be preceramic 

or early ceramic in age. It has not presently produced Tchefuncte sherds. Four radiocarbon 

samples were selected from among the 30 small charcoal samples obtained in Area C. Results 

of age determination were for the most part unacceptably early; this problem is discussed 

fully in Chapter 7. 

Two trial wet screen (1/8 inch or 3.1 mm) samples of about 2 liters volume each were 

processed in an attempt to evaluate the potential for recovery of organic remains. These 

samples were collected from Test Pit 1, 1.1-1.2 m (presumed Calion phase occupation) and 

from Test Pit 6, 0.9-1.0 m (organic zone with debitage noted above). The results were equivocal, 

but not promising. The coherent silty clay matrix must be dispersed or deflocculated in 

this procedure. Both samples yielded rare wood charcoal particles and organic (?) matter not 

positively identified, but no carbonized seeds or nuts and no animal bone. The general experience 

of test pit excavation is that organic remains are exceedingly sparse, where present, in 

all components represented at Marie Saline. 

In Area C, 15 auger holes were emplaced to 1.0 m depth in order to delimit area of 

occupation. One hole produced two white novaculite flakes from about 0.8-1.0 m, while all 

others were apparently sterile (Figure 36). Since some auger holes between test pits would 

certainly intersect cultural “horizons,” we have reason to believe that 4-inch (10 cm) auger 

bits are ineffective in locating diffusely distributed artifacts and debris. 

Overview. Marie Saline (3AS329) contains a record of intermittent occupation and 

reoccupation covering a 3000-year span. Artifacts and debris of all kinds are scarce by the 

standards of upland habitation sites in this region. Table 10 below summarizes the collections 

of tools and imperishable materials recovered from three occupation areas by survey 

Table 10. Total Recovery of Imperishable Artifacts and Debris by Occupation Area, Marie 

Saline (3AS329). 

Occupation Area Sherds Stone Tools Debitage Fire-cracked Rock Fired Clay 

Area A 10 3 26 1 0 

Area B 19 8 292 2 0 

Area c 68 21 159 213 134 

Site Totals 97 32 477 216 134


and test excavation techniques. The deep occupation zones now identified as Poverty Point 

and Tchula periods in Area C appear to represent more intensive and extensive site use than 

do overlying zones. We have introduced the possibility that Marie Saline occupied a natural 

levee location at the mouth of the Saline River during early occupation. In fact, the initial use 

of this site lies below river stage, and has not yet been investigated (see Appendix C). 

False Indigo (3AS285) 

The False Indigo site (3AS285) is located on the low natural levee of the Saline River 

0.3 river miles (less than one-half kilometer) above the present mouth. The site name is 

taken from a leguminous shrub, Amorpha fruticosa, here growing profusely on the exposed 

bankline, but absent from the floodplain forest understory. This bankline reaches a maximum 

height of 2.1 m above river level (62.2 to 69.9 feet MSL). The topstratum deposits 

exposed are similar in appearance to the Marie Saline cutbank and test pit profiles. 

When discovered by our survey team, False Indigo produced a minor scatter of 

sherds and lithic debris along the steep bank and water’s edge for 40 m (Figure 40), but also 

produced an impressive and dense sherd cluster on the levee surface where degraded by 

sheet erosion. Much of this ceramic array could be attributed to the Gran Marais ceramic 

complex and Gran Marais phase (A.D. 1100-1400). A journal entry of that initial discovery 

states that “the remaining sherd concentration on the brow of the bank literally looks like 

it was dumped yesterday rather than five or more centuries ago.” These observations led 

us to plan and implement a controlled surface collection at False Indigo, focusing on an 18 

m 2 area of dense artifact occurrence (Figures 40, 41). We shall return to the objectives and 

results of this effort below. 

Upstream from the dense sherd cluster, evidently separated by a few meters of sterile 

bankline, a second small sherd concentration was located in the bank at 0.6-0.8 m depth. 

These sherds were linear punctated or “jabbed incised” on Tchefuncte paste, therefore a variety 

of Lake Borgne Incised (Figure 65a-c). We now distinguish two occupation areas, Area 

A downstream and Area B upstream (Figure 40), which contain younger and older components 

and are separated vertically by about 0.4 m of topstratum deposition; these components 

represent two periods of the site use as follows: 

Mississippi period, Gran Marais phase (A.D. 1100-1400) 

Tchula period, Coon Island phase (500 B.C.-A.D. 1) 

Area A. The Gran Marais phase occupation in Area A was investigated by systematic 

(100%) surface collection, one judgmental test pit (1), one random test pit (15), and 17 shovel 

tests to about 0.4 m depth. This work delimited a compact, 12 x 15 m area of artifacts, midden, 

and debris at about 0.1 to 0.3 m depth below the levee surface, partly exposed by sheet 

erosion near the bankline. Figure 41 is compiled from two episodes of systematic surface


Figure 40. False Indigo (3AS285), controlled surface collection and test excavations, November 1979.

170 Hemmings 

Figure 41. False Indigo (3AS285) controlled surface collection in Area A with scatter map of sherds and other artifacts.

Assessment of Floodplain Sites 171 

collection and “piece plotting” which attempted to distinguish differential distribution of 

artifacts and debris. Such patterns could be related to onsite behavior, for example, men’s or 

women’s activities in or near dwellings or merely refuse disposal. Our collection and plotting 

activities were separated by two month’s time between site discovery and initiation 

of testing; ongoing erosion uncovered about 100 additional items during this interval. The 

second surface collection effort also involved superficial troweling of leaf litter and 2-3 cm 

of silt from the previous spring flood on the 18 m 2 area of densest artifact occurrence. We 

now think that sheet erosion of rather homogeneous midden is reflected by this distribution 

(Figure 41), but in fact additional spatial analyses of this scatter are possible. 

Test Pit 1 intersected an area lowered by erosion, but was nevertheless rich in artifacts 

to 0.2 m depth. An Ashley arrow point (Figure 58 l), Gran Marais sherds, chert, and clear 

crystal quartz debitage, and about 2900 fired clay particles were recovered here. Fired clay 

on the surface had in fact suggested the existence of a hearth, but none was distinguishable. 

An irregular shallow pit, Feature 1, extending into the subsoil, may merely be the remnant 

basal level of midden. A 2-liter wet screen sample yielded many small wood charcoal fragments, 

modest amounts of calcined bone, and much imperishable residue, chiefly fired clay 

and debitage. The numerous fired clay particles from this test pit are variable in size up to 

40 mm, usually iregular in form, but include a few examples with smoothed or prepared 

surfaces (not including the wattle impressions of true daub). 

Test Pit 15, selected at random from our 18 m 2 area of debris, had little surface occurrence 

of artifacts, but intersected well preserved midden and features from less than 0.1 to 

0.5 m depth (Figures 42, 43). Feature 2 is an irregular midden lens with a conspicuous downward 

protrusion into subsoil, either an odd postmold or tree root disturbance. A 2-liter wet 

screen sample from this lens was relatively more productive of organic remains than that 

noted above for Feature 1. This sample contained abundant tiny wood charcoal fragments, 

burned and calcined bone, and imperishable residue. Fish vertebrae and spines, drum teeth, 

and gar scales are present; in fact nearly all bone fragments may be those of fish. Charred 

nutshell or seed fragments may be present in minute amounts. Wood ash is believed to be 

a constitutent, based on midden appearance and soil pH (5.9) contrasting with nonmidden 

samples (Chapter 7, Table 8). 

Feature 3 in Test Pit 15 is the submidden surface, an irregular occupation floor marked 

by sizable sherds and buried by varying thicknesses of midden. From various data we 

believe this is indeed a structure floor or outdoor work area, not merely a locus of refuse 

disposal. The principal artifacts recovered in Test Pit 15 are an Ashley arrow point (Figure 

58l), chert, novaculite, and clear crystal quartz debitage, and about 600 fired clay particles of 

small size. Traces of glossy, black, cinderlike residue from both features may pertain to the 

burning process which occurred here (in addition to wood charcoal, ash, and fired clay).


Figure 42. False Indigo (3AS285), profile with midden lens and other features 

in Test Pit 15 (AAS neg. 797097).


Figure 43. False Indigo (3AS285), Area A, showing 18 m 2 area of dense artifact 

concentration after surface collection and test excavation; Saline 

River bankline is in the foreground (AAS neg. 797100).


One additional 2-liter wet screen sample from subsoil below Feature 3 was processed for 

organic remains and produced traces of calcined bone and wood charcoal as tiny fragments. 

Nine of 17 shovel tests in Area A produced artifacts or debris between about 0.1 and 

0.4 m depth; at least one of these (Figure 40) intersected dark colored midden at about 0.1- 

0.2 m. One other shovel test recovered a sandstone cobble anvil, the only heavy duty tool 

(Figure 63c). These tests together indicate an irregularly oval occupation area, unlike the 

linear configuration which characterizes most riverbank sites recorded. 

Area A contained a ceramic sample of about 1600 sherds which are entirely attributable 

to the Gran Marais ceramic complex. The following frequencies are noted (see also 

Chapter 8): clay-tempered plain, 58.4%; clay-tempered decorated, 9.3%; shell-tempered 

plain, 2.5%; shell-tempered decorated, 0.3%; bone-tempered plain, 0.1%; indeterminate 

(small weathered sherds), 29.4%. 

Area B. Figure 40 indicates the position of Test Pit 19 and three shovel tests in Area B. 

These tests establish the presence of a Tchula period component with Lake Borgne Incised 

and other Tchefuncte sherds at about 0.6 to 0.8 m below the modern levee surface. A white 

novaculite dart point fragment and six white novaculite flakes were associated with this 

level in Test Pit 19. Minor amounts of Baytown Plain sherds and chert or novaculite debitage 

occurred in upper levels of Test Pit 19 and two shovel tests. From these test excavations and 

riverbank exposure, the Area B early ceramic component is evidently localized and sparse, 

lacking the fire-cracked rock and organic remains seen at Marie Saline. Depth of this material 

and predominance of a late Tchula ceramic type may indicate occupation late in the Coon 

Island phase, between about 250 B.C. and A.D. 1. 

Overview. False Indigo (3AS285) has early and late ceramic components separated by 

0.4 m of silty clay and more than a thousand years of time. The setting and nature of Coon 

Island phase occupation cannot be interpreted from test excavation data. In contrast, the 

Gran Marais phase occupation was clearly oriented to a Saline River course much like that 

of the present; the contents of a shallow refuse deposit indicate fishing activity and wet 

cookery. Well preserved organic midden deposits were not known to exist in floodplain sites 

of our study area prior to test excavation at False Indigo. Relatively high densities of artifacts 

and debris including organic material, and indications of structural remains (daublike 

material, postmold?) occur here in a spatially restricted area. An upland site of these characteristics 

would be identified as a single house site of some permanence. The paucity of lithic 

tools and debitage as compared to ceramic remains (Table 11), lowland setting, and midden 

content (fishbone and residue from burning) imply more specialized site use.


Table 11. Total Recovery of Imperishable Artifacts and Debris by Occupation Area, False 

Indigo (3AS285). 


Entire Site Initial Collection 90 0 4 2 11 

Area A 1490 7 83 7 4068 

Area b 83 2 10 1 0 

Site Totals 1663 9 97 10 4079 

Buttonbush (3BR58) 

This small ceramic site is located 0.8 river miles (1.3 km) upstream from the present 

mouth of the Saline River. The bankline is degraded by sheet erosion and slopes to the 

water’s edge. Our survey team recorded Buttonbush as a very sparse lithic artifact scatter 

extending 98 m along the shoreline and a small dense cluster of Baytown Plain sherds eroding 

out 1.5 m above river level near the upstream end of the site (Figure 44). 

Test excavation consisted of two judgmental test pits and 19 shovel tests. Test Pit 1 recovered 

fired clay particles to 0.1 m depth, but was otherwise sterile. Test Pit 2 investigated 

the dense sherd cluster and was expanded to 1.5 x 1.5 m in order to encompass all additional 

sherds encountered. Analysis indicates that the entire group of 454 sherds, tightly massed 

and nested at or just below the modern surface, is the remains of two Baytown Plain vessels 

(Figure 45). These vessels were relatively large bowls with straight flaring walls and flat disc 

bases (Figure 66h, i). Most of the rim sections are missing or have been weathered beyond 

recognition. Test Pit 2 encountered no other artifacts or debris and was sterile below 0.1 m 

depth. A soil sample collected beneath the sherd cluster is reported in Chapter 7 (Table 17). 

Shovel tests at upstream and downstream ends of the site were sterile except in one instance. 

This test about 4 m west of the sherd cluster (Figure 44) produced charcoal at slightly 

over 0.2 m depth, which may better reflect the extent of alluviation following site use than 

does the exposed sherd cluster. 

The principal lithic artifacts collected from the shoreline are a small chert Gary point 

(Figure 59a), two chert and novaculite preforms, and minor amounts of chert and novaculite 

debitage. Most of this material was scattered 30 to 50 m downstream from the sherd cluster, 

and may or may not be contemporary. Since the Baytown Plain vessels can most likely be 

ascribed to the long span of the Baytown-Coles Creek period (A.D. 400-1100), the lithic artifacts 

may in fact be associated. More precise dating is desirable, but beyond the capability of 

existing site data.


Figure 44. Buttonbush (3BR58), test excavations, November 1979 (four shovel tests are outside map area).


Figure 45. Buttonbush (3BR58), remains of two Baytown Plain vessels in Test 

Pit 2 (AAS neg. 797105).


Buttonbush (3BR58) appears to represent a small collecting (?) station on the Saline 

River bank whose occupants made use of large open-mouthed ceramic vessels. For comparative 

purposes, the total sample of imperishable artifacts and debris is recorded here: 454 

sherds, 3 stone tools, 8 items of debitage, 5 fire-cracked rock fragments, and 9 particles of 

fired clay. A similar site at similar elevation, Water Elm 3 (3AS287), was recorded during the 

Ouachita River bankline survey (recommended for further work in Appendix C). These sites 

now seem to consist chiefly of the remains of one or two cooking or collecting vessels broken, 

discarded, alluviated, and reexposed by sheet erosion near a relatively stable bankline. 

One Cypress Point (3AS286) 

One Cypress Point (3AS286) occupies the low natural levee of the Saline River at the 

entrance to a small lake, the scar of an earlier river channel (Figure 46). Many older channel 

scars and sloughs weave through the adjacent backswamps. In the site description and 

analysis that follows we will attribute site use to a terminal phase of the Mississippi period 

(ca A.D. 1600-1700). Government Land Office maps of 1827 and 1844 indicate that the river 

channel has not shifted appreciably at this location for a century and a half. This stability 

suggests that the river and lake existed also at the time of occupation, or about three centuries 

ago, rather than the possibility that a cutoff lake formed during this interval. 

Discovery and recording of One Cypress Point quickly established the dimensions and 

content of this important site. An elongate 5 m wide band of chert tools, debitage, and an occasional 

shell-tempered sherd were exposed on the surface by sheet erosion from the Saline 

River bank 60 m northward along the lakeshore (Figure 47). This band of debris extended 

from the levee surface nearly to the water’s edge, a vertical distance of 1.7 m (62.2-68.4 feet 

MSL), on a gently sloping bankline. An apparent sterile zone intervened near the midpoint 

of this elongate surface scatter, and has provided the basis for distinguishing two occupation 

areas, Area A adjoining the Saline River bank and Area B along the lakeshore. These occupation 

areas have been investigated by systematic surface collection and mapping on two occasions 

(site discovery and initiation of testing) and by means of seven test pits and 28 shovel 

tests. The typological and stratigraphic data now available do not justify any temporal or 

cultural distinction between Areas A and B. They are residential or functional loci of a single 

late Mississippi or protohistoric period component. 

One Cypress Point’s stone tool assemblage is distinctive, previously unknown in the 

Felsenthal region, and likely to be intrusive. The hallmarks of this assemblage are 15 Nodena 

arrow points and fragments of two tiny endscrapers (Figure 48). These tools are entirely 

like their counterparts of the Nodena phase (A.D. 1400-1700) in northeast Arkansas and the 

Quapaw phase (ca A.D. 1600-1824) in the lower Arkansas River drainage (Ford 1961:Plate 

29; McGimsey 1964, 1969; Morse 1973:Figure 9; Hoffman 1977:31; see also Williams 1980 on 

the protohistoric Armorel phase). Other flaked stone tools are generally consistent with this


Figure 46. Aerial view of the lower Saline River floodplain and the location 

of One Cypress Point (3AS286) (air photo December 1975, courtesy 

of U.S. Army Corps of Engineers) (AAS neg. 814152).


Figure 47. One Cypress Point (AS286), controlled surface collection and test excavations, November 1979.

g 

A 

h 


Figure 48. Flaked stone tools and cores from One Cypress Point (3AS286). a. arrow points; 

b. arrow point fragments; c. end scrapers; d. cylindrical drill; e, f. bifacial preforms 

and fragments; g. bifacial cutting tool; h. perforators; i. retouched and 

utilized flakes and bladelike flakes; j. pebble cores (AAS neg. 813883). 

e 

f 

j 

b 

i 

c d


Quapaw attribution, as are the shell-tempered ceramics (which are too fragmentary to 

classify on a specific level). No European trade items have been recovered in our work at 

One Cypress Point, but limited testing does not rule out their possible existence here. We 

are aware of the complex problem of ascribing Quapaw phase archeological remains to the 

historic Quapaw people (House and McKelway 1980). However, we are confident that One 

Cypress Point contains the remains of an occupation by Quapaw phase people, regardless of 

their linguistic identity and point of origin. It thus becomes of great interest to examine the 

temporal and functional nature of this small settlement or camp. 

Area A. The larger of two activity or residence loci was examined by two systematic 

(2, 3) and two judgmental (4, 5) test pits and by 13 shovel tests, all excavated to sterile soil at 

about 0.3 m depth. Test pits near the bankline produced chert flakes and other material in 

the first level primarily, while Test Pit 2 on slightly higher terrain was most productive in the 

second 10 cm level excavated. Nodena arrow point fragments, shell-tempered sherds, chert 

debitage, and other debris characterize this shallow buried zone at about 0.1-0.2 m depth. 

No organic remains and no indications of features were uncovered, although irregular fired 

clay particles and sparse fire-cracked rock fragments were present in these test pits. Chapter 

7, Table 18, contains soil test data for three samples from Test Pit 2, where soil conditions are 

extremely acid. Shovel tests extend the dimensions of Area A and suggest that much more 

of the shallow buried horizon remains here, unaffected by sheet or bankline erosion (Figure 

47). A significant collection of stone tools, debitage, sherds, and other imperishable debris 

is the principal result of the testing work. This collection and that from Area B have been 

analyzed in considerable detail, and the results will be summarized below and in Chapter 8. 

Area B. Test excavation in this smaller locus consisted of two systematic test pits (1, 6), 

one random test pit (7), and 15 shovel tests to sterile soil at 0.3 m depth. The results obtained 

are entirely similar to those in Area B. The shallow buried horizon at 0.1-0.2 m depth is best 

represented in Test Pit 7 where all of the following materials were recovered: chert arrow 

point tip, chert microcore, chert utilized and retouched flakes, chert and novaculite debitage, 

fire-cracked sandstone, traces of burned bone, and one charred persimmon seed. Again, 

these remains suggest that an occupation zone at shallow depth is partly unaffected by erosion 

and extends well beyond the bankline artifact scatter. 

Overview. One Cypress Point (3AS286) is a shallow buried site believed to have been 

occupied by terminal Mississippi period or Quapaw phase hunters between about A.D. 1600 

and 1700. The settlement or camp extended over an area of about 700 m 2 , so far as we can 

tell, and included two loci of activity on the Saline River bank and adjacent lakeshore. The 

remains of this camp are a rich assemblage of tools and imperishable debris, but evidently 

contain few traces of structures or features and only rare organic debris. Table 12 below 

summarizes this site assemblage.

Assessment of Floodplain Site 183 

Table 12. Total Recovery of Imperishable Artifacts and Debris by Occupation Area, One 

Cypress Point (3AS286). 


Area A 79 133 847 5 124 

Area b 5 45 237 2 9 

Site Totals 84 178 1084 7 133 

Chert and novaculite tools and toolmaking debris from One Cypress Point appear 

to constitute the best available sample from a floodplain site in our project area. The lithic 

assemblage has been analyzed and tabulated by functional categories for tools (Figure 49a) 

and by reduction categories for debitage (Figure 49b). Expediently utilized flakes, usually of 

small size, with one or more cutting edges sheared or nibbled by use, comprise 64% of this 

assemblage. Bifaces of several presumed functional types, including arrow points and arrow 

point preforms, comprise an additional 31%. Endscrapers (Figure 48c) which must have 

been employed in hidework, and other maintenance tools are sparsely represented. The sole 

occurrence of a ground stone tool is a hematite plummet from Area A (Figure 64c), which is 

quite likely out of context among the Quapaw phase materials. 

Although Nodena arrow points predominate, single examples of other styles are present 

(Figure 58a-c). These include triangular, concave-based willow leaf, and stemmed arrow 

point with barbs. The barbed arrow point (fragment not illustrated) is not “typical” of Nodena 

phase and Quapaw phase assemblages, and may be a local style or trace of earlier Mississippi 

period site use. Nodena points are not known in the Shallow Lake site assemblage 

(Rolingson and Schambach 1981), and were previously unknown in the Felsenthal region. In 

addition to One Cypress Point, we have recorded a single Nodena point find at 3AS310, 1.9 

river miles (3 km) upstream on the Saline River. 

Figure 49b compares total samples of debitage from Areas A and B by means of cumulative 

frequencies for lithic reduction categories. This comparison indicates that reduction of 

pebble chert and novaculite in these areas was highly similar with regard to the debris produced 

and discarded. Production of bifaces was an important task in both areas. The choice 

of raw material was also similar—chert of upland terrace origin comprises 89.5 and 92.0% of 

Area A and B samples of debitage (Chapter 8). Finally, approximately 10-25% of cores and 

debitage from all subsamples in both areas exhibit obtrusive heat alteration and imply heat 

treatment of chert and novaculite raw material to improve its knapping quality. 

One Cypress Point furnishes an excellent sample of flaked stone tools and debris which 

may pertain to a single brief seventeenth century occupation. Moreover, these remains are


Figure 49. One Cypress Point (3AS286), graphic presentation of lithic assemblage. a. 

pie diagram showing relative proportions of flaked stone tools; b. cumulative 

frequency curves for cores and categories of debitage in Areas A 

and B.


likely to derive from the all male activity of a seasonal hunting camp in the Saline River 

floodplain. This statement, is, of course, subject to future field and laboratory findings at 

3AS286 (see Appendix C). 

Jug Point Cutoff (3BR76) 

This small midden site is located on the “gooseneck” of Jug Point Bend at the entrance 

to a chute or slough. During rising and falling flood stages, the chute conveys a strong current 

across the gooseneck. Eventually neck cutoff here will produce a sizable oxbow lake, 

but Government Land Office maps of 1827 and 1844 again indicate stability of the present 

Saline River channel during the past century and a half. 

Jug Point Cutoff was discovered as a small dense sherd scatter on the eroded bankline 

of the slough from the water’s edge almost to the modern levee surface, a 2.2 m vertical 

distance (62.2-69.1 feet MSL). The single component identified from surface collection and 

testing is evidently much like that present in Area A of the False Indigo site (3AS285), that is 

a shallow buried midden of the Mississippi period, Gran Marais phase (ca A.D. 1100-1400). 

Test excavation at Jug Point Cutoff consisted of one judgmental test pit to 0.4 m depth 

and 21 shovel tests ranging from 0.4 to 0.6 m depth (Figure 50). Test Pit 1 intersected a dense 

organic midden lens from about 0.1 to 0.4 m in depth, sloping eastward slightly toward the 

slough. Abundant mussel shell (four or more species) and alkaline midden soil are among 

the chief constituents of this refuse deposit (Appendix B, Chapter 7). Other contents recovered 

in Test Pit 1 are a chert Ashley arrow point (Figure 58i), a single novaculite flake, abundant 

sherds, and fired clay particles. 

Three wet screen samples of about 2 liters volume each were taken from dense midden 

in Test Pit 1 and shovel tests nearby at 0.1 to 0.2 m depth. No quantitative results were 

obtained, but the qualitative results are consistent among all samples. Fishbone occurs in 

profusion as small burned and calcined fragments, including vertebrae, spines, otoliths, and 

other elements. Gar scales and drum teeth are numerous. Considerable variation in size of 

elements (and species ?) may indicate that nets or traps were employed. Other organic midden 

contents, generally of less frequency, are small animal bone, crawfish exoskeleton (?), 

wood charcoal, carbonized seed fragment (?), siliceous sinter (?), herbaceous plant stem (?), 

and several animal coprolites with bone or fiber content. Under the binocular microscope, 

both plant and animal matter are coated with a cement or precipitate, and some bone fragments 

are rounded or altered by soil acidity or partial digestion. On the whole, however, 

this deposit has high quality preservation of fishbone and other organic debris. 

Shovel tests extend the dimensions of surface artifact scatter and subsurface midden 

as indicated in Figure 50. Only three of 21 such tests were sterile; all others produced sherds 

and other debris sometimes in appreciable quantity.


Figure 50. Jug Point Cutoff (3BR76), test excavations November 1979. Hachured area is presumed extent 

of subsurface midden lens.


Jug Point Cutoff (3BR76) now appears as a shallow buried occupation area, 11 x 17 

m and irregular in plan, enclosing a smaller circular midden deposit. The following totals 

for imperishable artifact and debris are recorded: 250 sherds, 2 stone tools, 6 items of debitage, 

2 fire-cracked rock fragments, and 599 particles of fired clay. Many of the observations 

noted for Gran Marais occupation at False Indigo (3AS285) apply equally well to Jug Point 

Cutoff—here a shallow organic midden with high density of artifacts and diverse debris 

indicates fishing activity and wet cookery in a spatially restricted area. 

The Gran Marais ceramic complex is exclusively represented and the following frequencies 

are noted: clay-tempered plain, 56.8%; clay-tempered decorated, 3.6%; bone-tempered 

plain, 1.2%; indeterminate sherd fragments, 16.8%. Relatively more shell-tempered 

ceramics and other differences suggest that Jug Point Cutoff was occupied slightly later than 

the Gran Marais phase component at False Indigo (Chapter 8, Table 26). Independent dating 

evidence is needed to corroborate this sequence. 

Jug Point 1 (3AS306) 

This site is located on the low natural levee of the Saline River at Jug Point Bend. It 

appears to be well exposed by cutbank and sheet erosion 75 m along the shoreline, and was 

recorded as two discrete artifact scatters (Figure 51). Area A downstream is evidently exclusively 

a lithic scatter, while Area B upstream contains both shell-tempered sherds and lithic 

debris. Both artifact scatters included occasional debris at the water’s edge or on the cutbank, 

and concentrations exposed by erosion near the levee surface. The bankline here has 

a vertical span of 1.8 m (62.2-68.1 feet MSL). An incised shell-tempered sherd from Area B is 

shown in Figure 67b. The concentration in Area A also has produced the most unusual artifact 

now known in our project area, a perforated disc of fossil amber (?), discussed further in 

Chapter 8. We will present some data from surface collection and test excavation indicating 

that Areas A and B are contemporaneous activity loci, and represent brief occupation during 

the Late Mississippi period, Caney Bayou phase (A.D. 1400-1600). 

Area A. Test excavations in Area A consisted of one judgmental test pit and 18 shovel 

tests to about 0.4 m depth. Test Pit 1 was emplaced at the locus of productive shovel tests 

and also of an arrow point surface find. This chert arrow point has a distinctive “gar scale” 

silhouette, and has been provisionally described as an Eagle Creek point (with two other 

examples recovered at Jug Point 1; see Chapter 8 and Figure 58f). Chert and novaculite 

debitage was concentrated in Test Pit 1 at 0.1 to 0.2 m depth. Flat-lying flakes may mark an 

otherwise imperceptible “floor” at 19 cm depth (Figure 2). No organic remains were noted 

in this test pit. Chert tools from the productive level include a bifacial cutting tool and two


Figure 51. Jug Point 1 (3AS306), test excavations, November 1979.


Figure 52. Jug Point 1 (3AS206), test pit profiles.


utilized flakes. Soil samples collected at and near the presumed floor are characterized in 

Chapter 7, Table 18. 

Ten of the 18 shovel tests in Area A were sterile; all others produced from one to seven 

chert or novaculite flakes at less than 0.3 m in depth. These shovel tests assisted us in defining 

an elongate oval occupation zone, marked principally by debitage. 

Area B. A single judgmental test pit was emplaced near productive shovel tests which 

yielded shell-tempered sherds and chert and novaculite debitage at less than 0.3 m depth. 

Test Pit 2 intersected an occupation level or “floor” at 17-19 cm below the levee surface and 

a small refuse filled pit (Feature 1) extending into subsoil from this level ( Figure 52). The 

contents of Feature 1 included the stem portion of an Eagle Creek point, three chert and 

novaculite biface fragments, five shell-tempered sherds, abundant chert and novaculite 

debitage, calcined mammal bone fragments, wood charcoal, carbonized seeds and nutshell 

(?), fired clay particles, and other residue. A 2-liter wet screen sample produced similar 

materials, and especially a small quantity of carbonize seeds. No fishbone has been noted in 

Feature 1 refuse. The relatively large seeds and seed fragments are coated by a sediment film 

or precipitate and obscured, but are distinctive in size and shape and are tentatively identified 

as honey locust and persimmon. This suggests that Jug Point 1 occupation and disposal 

of food materials in Feature 1 occurred in fall. 

Only six of 14 shovel tests in Area B were sterile. Others yielded chert and novaculite 

debitage, shell-tempered sherds, and other debris at less than 0.3 m depth. Area B has a 

surface scatter of historic artifacts and debris which can be ascribed to a 1930-1940 hunter’s 

or fisherman’s camp (Figure 51). 

Overview. Two small discrete activity loci of the Caney Bayou phase (A.D. 1400-1600) 

are separated by about 25 m on the Saline River bank. Eagle Creek arrow points are sparsely 

represented in both loci, while shell-tempered sherds, mostly plainware, are restricted to 

Area B. Shallow occupation floors were intersected by test pits at about the same depth (17- 

19 cm) in each locus. Arrow point style and depth of alluviation suggest that these areas are 

coeval. 

Feature 1 in Area B furnished ordinary food refuse which is important to the interpretation 

of seasonal use at Jug Point 1. Calcined mammal bone and charred seeds of honey 

locus and persimmon are intermixed with other debris in a small pit. Nutshell may also be 

present in small quantities. The seed pods, nuts, and fruits of these trees were undoubtedly 

available in the natural levee forest during fall. 

The content of Areas A and B, known from surface collection and limited testing, presents 

an interesting contrast. Table 13 summarizes the occurrence of imperishable artifacts 

and debris in these areas. It may be suggested that flaked stone tools and debitage in Area 

A represent male items and male activity, and that ceramics and food preparation refuse in 

Area B represent female items and debris (although stone tools and debitage are present


here too). A family group or task unit of this kind would reflect quite different strategy for 

exploiting floodplain resources than that proposed earlier for One Cypress Point (3AS286). 

Table 13. Total Recovery of Imperishable Artifacts and Debris by Occupation Area, Jug 

Point 1 (AS306) 


Entire Site Surface Collection 2 4 74 1 6 

Area A 0 12 164 0 0 

Area b 60 7 439 7 107 

Site Totals 62 23 677 8 113 

Finally, there is the highly unusual find at Jug Point 1, Area A, of an amber (?) disc or 

bead among otherwise ordinary debitage. Detailed discussion of this specimen is presented 

in Chapter 8. 


Jug Point 2 (3AS307) occupies the same bank of the Saline River about 70 m upstream 

from Jug Point 1. The bankline, about 1.9 in height (62.2-68.3 feet MSL), slopes gently upward 

from the water’s edge. This site was first recorded as a sparse artifact scatter extending 

80 m along the shoreline, but was subsequently collected, tested, and, mapped as three 

discrete and closely spaced activity or residential loci (Figure 53). Area A downstream is a 

sparse lithic scatter, while Areas B and C upstream contain both shell-tempered and claytempered 

ceramics and lithic debris. This site is therefore extremely similar in layout and 

content to Jug Point 1, and one might ask if these two sites are a single Mississippi period 

component consisting of five successive bankline activity loci. Ceramic and other artifact 

data, however, suggest to us that Jug Point 2 was occupied early in the Caney Bayou phase, 

between about A.D. 140 and 1500, and prior to occupation at Jug Point 1. A single radiocarbon 

sample taken from Area C did not provide the chronological control needed to answer 

this question (Chapter 7). 

Area A. This area is known only from surface collection. The principal artifacts recovered 

here are a novaculite Ashley point (Figure 58h), and chert and silicified wood debitage. 

Thirteen shovel tests to about 0.3 m depth were sterile. 

Area B. A single judgmental test pit and 10 shovel tests to 0.3 m depth were emplaced 

in Area B. Test Pit 2 produced a few Baytown Plain sherds and manuports at 0.2 to 0.3 m 

depth, and was otherwise sterile. Surface finds from this area include a novaculite Ashley 

point (Figure 58h), novaculite drill (Figure 61f), chert debitage, shell-tempered sherds, and 

fired clay particles. One shovel test recovered a plain shell-tempered sherd, while nine others 

were sterile.


Figure 53. Jug Point 2 (3AS307), test excavations, November 1979.


Area C. Area C is a larger and more diverse artifact scatter, and has been investigated 

by means of one judgmental test pit and 1 shovel tests to about 0.3 m depth. Test Pit 1 

disclosed a subtle, but visible, occupation level at 14 cm below the levee surface, extending 

4 to 8 cm into subsoil. This thin zone is organically stained and contains artifacts, calcined 

mammal (?) bone fragments, sparse wood charcoal, and a few carbonized seeds, probably 

all honey locust. A collection of charcoal fragments from the entire test square (14-22 cm 

organic zone) was combined as one radiocarbon sample (see Chapter 8). The principal artifacts 

from this level are a chert Bassett arrow point (Figure 58g), a chert spokeshave, chert 

and novaculite debitage, and a small group of thin plain body sherds (about equally divided 

between clay and shell temper). Characteristics of a soil sample from the occupation level 

are tabulated in Chapter 7. 

Six of 15 shovel tests in Area C produced artifacts or debris above 0.2 m depth, and 

these tests extend the known dimensions of this shallow occupation level to about 8 x 15 m. 

Overview. Jug Point 2 (3AS307) consists of three small activity or residential loci within 

an 80 m stretch of the Saline River bankline at Jug Point Bend. These loci differ in size and 

content, and may or may not be strictly contemporary. All occur at very shallow depth, less 

than 0.2 m, beneath the modern levee surface. The co-occurrence of Ashley arrow points in 

Areas A and B, and of shell-tempered sherds with a thin late variety of Baytown (?) plainware 

in Areas B and C leads us to place the occupation of Jug Point 2 early in the Caney 

Bayou phase, or about A.D. 1400 (see also Chapter 8:Table 26). 

The presumed functional contrast among activity loci noted at Jug Point 1 is evidently 

repeated at Jug Point 2 60 m upstream. Area A contains male-produced lithic items, while 

Areas C contains a mixed assemblage of ceramic and lithic materials with an increment of 

organic refuse. Food remains include small quantities of calcined mammal bone and carbonized 

honey locust seeds, the latter indicating site use during fall. Table 14 summarizes the 

contrast between activity loci in terms of imperishable materials. 

Table 14. Total Recovery of Imperishable Artifacts and Debris by Occupation Area, Jug 

Point 2 (3AS307). 


Entire Site Surface Collection 67 2 44 1 2 

Area A 0 1 2 1 0 

Area b 10 2 0 0 2 

Area c 21 3 20 2 6 

Site Totals 98 8 66 4 10


Mouth of Eagle Creek (3BR78) 

This small site lies on the low natural levee of the Saline River at the mouth of Eagle 

Creek. During initial survey, a large novaculite Gary point (Figure 59e), small quantities of 

novaculite and chert debitage, and fire-cracked rock were found on the eroded bankline. 

Some recent historic debris, but no prehistoric artifacts, were found in a clearing on the levee 

surface which reaches a height of 1.8 m (62.2-68.0 feet MSL) above river stage. The kinds of 

artifacts and debris recovered led us to believe that a relatively deep preceramic component 

is represented at Mouth of Eagle Creek. 

Testing consisted of troweling and examination of the bank profile, emplacement of 

two judgmental test pits (which were not excavated below 0.5 m due to rise of river level), 

and 15 shovel tests to 0.6 m depth. The general results of this work are indications that a 

thin, charcoal-stained lens or “floor” with sparse fire-cracked rock and flake is exposed at 

about 0.6-0.7 m depth, running at least 9 m along the bank; other finds of fire-cracked rock 

and flakes were made from about 0.1 to 0.7 m depth at various locations within a 50 m 

stretch of riverbank or in adjacent test units (Figure 54). Two of 15 shovel tests produced one 

flake each above 0.6 m; all others were sterile. Neither Test Pits 1 or 2 reached the level of the 

charcoal-stained lens appearing in the riverbank, and therefore we have insufficient information 

about the age, content, and degree of preservation of this relatively deep occupation 

level. 

The principal occupation at Mouth of Eagle Creek (3BR78) was probably Poverty Point 

period, Calion phase (1200-500 B.C.) on slim evidence and on the basis of comparison to 

this early component at Marie Saline (3AS329). Gary dart points, white novaculite debitage, 

and fire-cracked rock are to be expected in such a component. For comparative purposes the 

following total sample of imperishable artifacts and debris is noted: 1 stone tool, 32 items of 

debitage, and 17 fire-cracked rock fragments. Very sparse remains at shallow depth indicate 

occasional low intensity site use at Mouth of Eagle Creek during subsequent ceramic 

periods, although we have not presently recovered sherds or other diagnostic late artifacts. 

Completion of testing, and perhaps extensive mitigation, are needed to investigate the 

nature of the early component, obtain samples of organic material for dating and identification, 

and verify the presence of younger superposed components (see Appendix C). 

eVAluAtion of test excAVAtion strAtegy And tActics 

Test excavation was undertaken at selected sites in order to assess their significance as 

cultural resources and to acquire objective data as a basis for recommendations about management 

of these resources. A large, systematic body of data has in fact resulted from this 

work. These data and interpretations placed on them are, we believe, fundamental contributions 

to regional archeological research and to management needs in the Felsenthal Project 

area, although they are preliminary in nature. The following brief evaluation of our work


Figure 54. Mouth of Eagle Creek (3BR78) test excavations, November 1979.


notes only the particular successes and difficulties of the testing stage, perceived in retrospect. 

1. Testing effort concentrated on one major Ouachita River prehistoric site and 

a cluster of lower Saline River sites; these are shown to be regionally significant, 

but testing of additional prehistoric and historic sites throughout the 

project area would greatly reinforce the data base and validity of the results. 

2. Vertical control on individual sites was maintained scrupulously so that we 

could differentiate superposed assemblages and relate buried components 

to the proposed navigation pool; however, existing lock stage and river level 

were employed as a horizontal datum (slackwater pool) when some slope 

of the water surface must be present, and errors in tenths of feet are likely to 

affect our site records; average slopes of reaches on the Ouachita and Saline 

rivers could have been determined by instrument and introduced as correction 

factors (Marie Saline site elevations refer directly to readings on a river 

gauge and thus are considered accurate). 

3. The baseline method of horizontal control, used with systematic, judgmental, 

and random test pits, was quick and efficient on our floodplain sites characterized 

by small size, elongate configuration, and shallow buried components; 

systematic surface collection at several levels of precision was also 

employed to discriminate activity loci and site boundaries in shallow components; 

since none of these sites is disturbed by plowing (or by other modern 

processes to significant extent), we think a presumption of spatial patterning 

is necessary, and even more rigorous controlled collection techniques could 

justifiably be tried and evaluated. 

4. Excavation, screening, and wet screening techniques were uniformly applied, 

but difficult, because of relatively high clay content in floodplain deposits 

(Chapter 7, Tables 8 and 19); experimentation is needed here to determine 

the most efficient and reliable techniques of recovery of materials, including 

delicate faunal and floral remains; flotation experimentation is needed so that 

preservation can be rigorously assessed and organic remains identified and 

quantified on a statistical basis; dispersion or deflocculation of matrix will be 

required when flotation sampling is implemented. 

5. The experience of both survey and test excavation is that power augering, 

while very quick, is ineffective in locating buried components with low density 

of inorganic and organic debris, as characterizes most floodplain sites in 

the project area; the slower method of shovel testing does prove effective to 

about 0.5 m depth, presumably because it greatly increases the volume of


soil extracted by each test; although we have not employed deep testing by 

means of core drilling or backhoe trenching in the project area, we believe 

such equipment could be effective in carefully designed modes of use. 

suMMAry AssessMent of floodplAin sites 

Results of test excavation in eight floodplain sites (described in this chapter) and 

subsequent analyses of materials recovered (Chapters 7, 8, Appendixes A, B) are the basis 

for evaluating these sites for potential eligibility and for nomination to the National Register 

of Historic Places. Our findings and recommendations are summarized in Table 16. More 

detailed discussion of significance, adverse effects, and proposed mitigation is contained in 

Appendix C. All sites selected for testing are buried components with relatively high integrity, 

having a degree of alteration or loss by riverbank erosion, but no significant historical 

disturbance. No site tested has been truncated or superficially disturbed by plowing. 

Elevations cited in Table 16 are those obtained by methods described earlier in this 

chapter for prehistoric cultural remains of diverse kinds in primary context (persisting essentially 

at the original locus of use). In every case artifacts and debris were also recovered 

during initial survey at the water’s edge and on eroded banklines at various elevations 

above river level. Testing has shown lowlying cultural remains to be eroded and transported 

(in secondary context) from contour in the case of six sites. We believe that all sites tested 

will be adversely affected by the 65-foot navigation pool, when cultural remains are eroded 

and removed at the shoreline, regardless of any case where primary context occurs above 65 

feet (see also Appendix C). 

As Table 15 indicates, all eight floodplain sites tests are scientifically important and 

“have yielded or may be likely to yield information important in prehistory” (36 CFR 

1202.6). We therefore recommend seven sites be nominated to the National Register of Historic 

Places; the remaining site (Buttonbush 3BR58) is not eligible by virtue of the fact that all 

known or expected cultural remains and related site data were recovered during testing. 

site functionAl hypotheses 

Test excavation in floodplain sites furnished data relevant to six of eight hypotheses 

stated in the research design (Chapter 1), but we are here concerned with one key hypothesis 

regarding the functional nature of such sites. The quality of data also permits us to go 

beyond this initial hypothesis and refine or present new site functional hypotheses pertaining 

to the Mississippi and earlier periods. Our initial hypothesis was stated as follows:


Table 15. Summary Assessment of Eight Floodplain Sites. 

Significance National Register General 

Site Name and Elevation* of Eligibility Mitigation 

Designation Components) Site Specific Data Research Potential Recommendation Recommendation** 

Marie Saline

Problem Domain 1: Settlement Characteristics 


H 1 If subsistence strategies employed wild plant and animal resources, impermanent 

specialized extractive sites would implement these strategies in true floodplain 

environments. 

Test Implications: Recognition of specialized extractive sites as settlement types, 

distinct from base settlements. 

Data Requirements: Floodplain site assemblages with high ratio of extractive/ 

maintenance tool kits; faunal or floral indicators of seasonality and procurement 

strategy; absence of permanent structures for habitation, storage, or burial. 

In the preceding summaries of test excavation results, 12 components were isolated 

and identified in eight floodplain sites; half of these components represent Mississippi 

period occupation and the other half are attributed equally to Baytown-Coles Creek, Tchula, 

and Poverty Point period occupations. Data obtained in Mississippi period components is 

more diverse and more complete in several respects. It seems then that we can best examine 

the functional nature of floodplain sites established between about A.D. 100 and 1700, drawing 

on pre-Mississippi period components for comparison and contrast. 

Our results are potentially of great importance in the context of recent ecologically 

oriented studies of Mississippi period settlement and subsistence (Lewis 1974; Smith 1975; 

contributors in Smith 1978). Many of these excellent studies employ site locational and size 

data with internal characteristics to derive settlement hierarchies in which farmsteads are 

the minimal economic and residential units. Where very small extractive camps are identified 

in regional Mississippi period settlement hierarchies, the analysts seem to doubt their 

functional typologies. Green and Munson (1978:319), for example, can distinguish relatively 

many Angel phase “hunting and/or gathering camps” in the Ohio River floodplain, terrace, 

and upland locations, but believe that “additional research will reveal many of the sites 

located in the river valley [as] farmsteads.” And Muller (1978:283) concludes for numerous 

small sites in the Black Bottom surrounding Kincaid that “there is really no good evidence 

to support an idea of widespread specialized extractive sites.” These views may be partly inconsistent 

with the strategy of Mississippian subsistence in floodplain environments, which 

had “two strings to its bow—wild and agricultural resources” (Peebles 1978:392). 

Before presenting data bearing on Hypothesis 1, we should amplify some of the terms 

which have been incorporated. 

1. A subsistence strategy is a schedule of decisions about, and priorities for, obtaining 

seasonally and spatially restricted food resources. 

2. Impermanent sites are loci of brief occupation or transient activity on a scale 

of days or weeks.


3. Specialized extractive sites are loci of one or few procurement activity sets 

(not including horticultural activities) which are operated by few personnel or 

small task units. 

4. True floodplain environments here refer to annually overflowed bottomlands 

which impose seasonal limits on site occupancy, as well as on seasonal availability 

of resources. 

The first data requirement is a high ratio of extractive to maintenance tools in site assemblages. 

Two principal sampling biases will affect our comparison: nonpreservation of 

perishable equipment and curate behavior (conserving and removing useful tools). With regard 

to stone tools recovered in test excavations, we note first that they are scarce (N= 256), 

ranging from two at 3BR58 to 178 at 3AS286, and that they are nearly all projectile points, 

bifacial preforms, bifacial cutting tools, and utilized flakes with cutting edges. We note also 

that one or more drills, scrapers, perforators, beads, plummets, pitted cobbles, and steatite 

sherds are present, but numerically insignificant in any site tested and in the group as a 

whole. There are no woodworking tools, no agricultural tools, and no grinding implements. 

With regard to ceramic vessels, we note again their scarcity (N=2708 sherds, representing 

far fewer vessels) in ceramic components, ranging from 62 at 3AS306 to 1663 at 3AS285, 

and that two Gran Marais phase components with fishbone midden (3AS285 and 3BR76) 

furnished 70% of the sherd sample. In these components particularly, the systemic context 

of ceramic vessels may have been processing of fish meat and oil for later distribution and 

consumption. Therefore we believe that extractive tool kits are dominant features of floodplain 

site assemblages, although variability exists among these tool kits in the proportions 

of weapons, items associated with refurbishment of weapons, and utensils or containers 

directly employed to butcher, render, boil, or otherwise process foodstuffs. 

The second data requirement is the discovery of faunal and floral indicators of seasonality 

and procurement strategy. The limited evidence can be quickly summarized as follows 

(see Potential Lowland Resources in Chapter 2). 

3AS285 remains of gar, drum, and other fish (late spring, summer) 

3AS286 animal remains, persimmon (fall) 

3BR76 remains of gar, drum, other fish, crawfish, and mussels (late spring, summer) 

3AS306 mammal remains, honey locust, persimmon (fall) 

3AS307 mammal remains, honey locust (fall) 

Suitable indicators thus appear to be present, while quality of data is presently inadequate 

and cannot strongly support the hypothesis. 

The last data requirement is an absence of permanent structures for habitation, storage, 

and burial. On the basis of bankline exposure and 5% subsurface investigation these 

structural remains are indeed absent. We doubt whether any well preserved Mississippi 

period farmstead could be similarly tested without producing substantial indications of


these kinds of facilities. Such farmsteads are distinguished in our view as relatively permanent 

base settlements, regardless of small size (Green and Munson [1978:310] rate Mississippi 

period farmsteads and camp similarly in respect to small size, “less than .25 ha”). 

An overall conclusion, qualified at various points in the preceding discussion, is that 

data acquired in test excavation of eight floodplain sites tend to support the hypothesis: 

impermanent specialized sites were established in the overflow bottoms of the Ouachita and 

Saline rivers to extract wild plant and animal resources. 

Variation Among Extractive Sites 

A final objective of our functional analysis is to consider (in brief and preliminary fashion) 

the implications of variation in extractive sites as a class. We shall look at the proportions 

of tools in assemblages, as Winters (1969) and others have done, but will also introduce 

three categories of debris. These debris categories are presumed to reflect activities as do 

tool kits and are much less likely to be “curated.” Table 17 presents the proportional data 

for imperishable artifacts and debris in nine components (six floodplain sites) where total 

samples exceed about 100 items. These components are easily assigned to periods and phases, 

except in the case of 3AS329, Area C, which incorporates both Calion and Coon Island 

phase assemblages (see Figure 39). Two of our Mississippi period components incorporate 

discrete activity loci (3AS306 and 3AS307), while the two relatively prolific loci in 3AS286 

are kept separate in order to examine their relationship. The next step in examining these 

proportional data was performed by cluster analysis using the Numerical Taxonomy System 

(NTSYS) with unweighted arithmetic averaging an coefficients of Euclidean distance (Rohlf 

et al. 1972). 

In the resulting dendrogram, we distinguish two clusters of components which are 

joined at levels of distance less than 0.2 (hachured areas in Figure 55). These clusters are of 

particular interest in the light of total available archeological evidence from our floodplain 

sites. Our inferences about these results are as follows: 

1. Cluster A components were heavily involved in hunting; debitage and stone 

tools predominate in these assemblages; mammal bone and seeds of fall fruits 

are present. 

2. Cluster B components were specially, perhaps exclusively, involved in fishing 

and fish processing; sherds and fired clay predominate in these Gran Marais 

phase assemblages which are all but identical in proportions (Table 16). 

3. Three components are dissimilar and intermediate to these clusters and dissimilar 

among themselves; we cannot say that they are simply hunting/fishing 

components; the heterogeneity in proportions of artifacts and debris may 

be taken as a tendency toward characteristics of base settlements; the earliest 

component represented (3AS329, Area C) especially should be tested more 

extensively as a possible base settlement (Appendix C).


Figure 55. Dendrogram generated by cluster analysis of artifacts and debris in nine prehistoric components 

(six floodplain sites). Clusters A and B (distinguished by hachuring) represent extractive sites 

associated with hunting and fishing respectively (see text discussion).


Table 16. Proportions of Five Functional Categories of Imperishable Artifacts and Debris in 

Nine Prehistoric Components (six floodplain sites). 

Percents of Assemblage Total* 

Fire- 

Provenience by Cultural Stone Cracked Fired Assemblage 

Site and Locus Period/Phase Sherds Tools Debitage Rock Clay Total 

3AS329-B Baytown-Coles 5.9 2.5 91.0 0.6 0.0 321 

Marie Saline Creek/unnamed 

3AS286-A Mississippi/ 6.6 11.2 71.3 4.2 10.4 1188 

One Cypress Point terminal 

3AS306 Mississippi/ 7.0 2.6 76.7 0.9 12.8 883 

Jug Point 1 Caney Bayou 

3AS286-B Mississippi/ 16.8 15.1 79.5 0.7 3.0 298 

One Cypress Point terminal 

3AS329-C Poverty Point- 11.4 3.5 26.7 35.8 22.5 595 

Marie Saline Tchula/Calion- 

Coon Island 

3AS285-B Tchula/Coon 86.5 2.1 10.4 1.0 0.0 96 

False Indigo Island 

3AS307 Mississippi/ 52.7 4.3 35.5 2.2 5.4 186 

Jug Point 2 Caney Bayou 

3AS285-A Mississippi/ 26.3 0.1 1.5 0.1 71.9 5655 

False Indigo Gran Marais 

3BR76 Mississippi 29.1 0.2 0.7 0.2 69.7 859 

Jug Point Cutoff Gran Marais 

* Percentages rounded off from five places; order of components in table based on results of cluster analysis 

(Figure 55).


Data and results of analyses presented in this chapter can lead, we think, to new approaches 

and new conclusions regarding early human adaptation in the Felsenthal region. 

Much additional, carefully designed research is needed. One necessary step will be to 

reformulate and test rigorously hypotheses about floodplain extractive site functions within 

particular cultural systems. Without attempting to characterize these hypotheses in every 

detail, we suggest the direction they can usefully take: 

H 1 Transient fall camp in the overflow bottom used by male hunting party to 

procure deer and other game (terminal phase of the Mississippi period) 

Data Base: One Cypress Point (3AS286) 

H 2 Transient late spring and summer camp in the overflow bottom used by specialist 

male and female task unit to harvest and process fish on a “bulk” basis 

(Gran Marais phase of the Mississippi period) 

Data Base: False Indigo (3AS285); Jug Point Cutoff (3 R76) 

H 3 Semisedentary late spring-fall base settlement in the overflow bottom used 

by multiple family unit to exploit riparian resources (Calion phase of Poverty 

Point period, Coon Island phase of Tchula period) 

Data Base: Marie Saline (3AS329) 

Finally, we believe that results of testing reported in this chapter affirm the research 

potential and quality of significance that may be present in small limited activity sites (Talmage 

and Chester 1977; Ward 1978). This perspective is absolutely necessary for managing 

resources in this Arkansas region, as it is elsewhere. Small sites investigated by us were not 

located in marginal upland settings and, for the most part, cannot be attributed to huntergatherer 

settlement systems. In fact, our most dramatic results pertain to Mississippi period 

societies who were part-time agriculturalists.

Chapter 7 

SOILS, SEDIMENTS, AND CHRONOLOGy 

A limited program of soil testing and radiocarbon age determination was carried out 

in conjunction with testing of floodplain sites. Results are reported below and are applied to 

site transformational hypotheses at the end of this chapter. Radiocarbon dates have proved 

to be an unusual problem, since our results are markedly earlier than expected; we discuss 

this problem fully near the end of this chapter. As the chapter title implies we deal with 

properties of soils and sediments, so it is well to distinguish between these terms. According 

to Haynes (1978:8)1 a soil is 

a modification of earth material by physical, chemical, and biological agents...[it 

is not] regolith, which as all unconsolidated material over bedrock can support a 

soil [and it is not] sediment but an alteration thereof and other rock types as well. 

We pointed out in Chapter 2 that floodplain topstratum deposits in the Felsenthal 

Project area are mapped as Guyton soils which are strongly acid, siliceous, silt loams over 

finer textured subsoils. Floodplain topsoil (the A horizon or humus) is weakly developed 

because flooding removes leaf litter and retards degradation by forest floor organisms, and 

because annual increments of silt and clay are deposited (overbank primarily). It is generally 

true that buried soils or paleosols are not present, or at least are not visible as organically 

enriched horizons, in riverbank exposures of the Felsenthal floodplain topstratum. 

An exception to the latter statement occurs where buried archeological levels are enriched 

(anthropogenic soils, including middens). Rich middens are recorded in only two floodplain 

sites, False Indigo (3AS285) and Jug Point Cutoff (3BR76); a thin anthropogenic soil zone 

was noted at Jug Point 2 (3AS307) in Test Pit 1 (Chapter 6). (Test pit profiles presented in 

Chapter 6 include Figures 37, 38, 42, and 52.) We report soil test results from these and other 

floodplain sites below. 

selected properties of floodplAin soil sAMples 

Table 17 presents the soil test data for seven floodplain sites which are described in 

Chapter 6. Analyses were performed by the Soil Testing and Research Laboratory, Agronomy 

Department, University of Arkansas. Soil samples of about 480 cc volume were collected 

in standard containers furnished by the Agronomy Department.


Table 17. Selected Properties of Soil Samples from Seven Floodplain Sites. 

Soil Observed Organic 

Site Name and Horizontal Depth Below Sample Texture Munsell Soil Re- Organic Phosphorus 

Designation Provenience Surface (cm) Number Remarks Color 2 action (pH )3 Carbon (%) mg/g 

Marie Saline Test Pit 1 10 1415-1 vfsl 10yR6/2 5.0 0.55 70 

(3AS329) “ 40 1415-2 sic 10yR5/1 4.1 0.45 90 

“ 70 1415-3 sicl 10yR6/1 4.9 0.39 60 

“ 100 1415-4 c 10yR5/1 4.9 0.27 50 

“ 130 1415-5 fsl 10yR5/1 5.7 0.09 25 

“ 150 1415-6 cl 10yR5/1 5.8 0.08 10 

Test Pit 6 5 1415-7 l 10yR5/2 4.9 1.69 50 

“ 8 1415-8 l (humus) 10yR3/1 5.1 5.86 65 

“ 25 1415-9 sicl 10yR6/2 4.4 0.59 60 

“ 55 1415-10 sil 10yR6/1 4.1 0.41 15 

“ 88-94 1415-11 cl (charcoal) 10yR5/1 3.9 0.20 25 

“ 110 1415-12 cl 10yR5/1 4.0 0.20 10 

False Indigo Test Pit 15 3-6 1433-1 si 10yR5/2 4.7 0.98 1250 

(3AS285) “ 11-15 1433-2 midden 10yR2/1 5.2 1.68 100 

“ 25-32 1433-3 sic 10yR6/1 5.8 0.64 75 

“ 38-42 1433-4 c 10yR6/1 5.9 0.19 30 

“ 14-17 1433-5 midden 10yR2/1 5.9 1.98 750 

“ 29-34 1433-6 midden 10yR2/1 5.9 3.78 875 

Shovel Test 1 0-30 1433-7 sic 10yR7/1 5.3 2.38 350 

Buttonbush Test Pit 2 0-10 1435-1 sic (under sherd 10yR6/1 4.5 1.44 250 

(3BR58) concentration) 

One Cypress Test Pit 2 8 1434-1 cl (humus) 10yR5/2 4.5 2.64 165 

Point (3AS286) “ 20 1434-2 c 10yR6/2 4.4 1.10 580 

“ 36 1434-3 c 10yR6/2 4.4 0.84 60 

Jug Point Cut- Test Pit 1 10-20 1439-1 midden — 7.0 2.55 1500 

off (3BR76) Shovel Test 16 10-20 1439-2 midden — 7.0 2.47 250 

Jug Point 1 Test Pit 1 10 1436-1 cl 10yR4/2 4.3 1.35 160 

(3AS306) “ 20 1436-2 sic (occupation 10yR4/2 4.2 0.83 195 

level) 

“ 30 1436-3 sic 10yR4/2 4.2 0.69 135 

Test Pit 2 5 1436-4 sic (humus) 10yR5/1 4.4 1.70 185 

“ 12 1436-5 c 10yR5/8 4.4 1.06 215 

“ 22 1436-6 sic (feature) 10yR4/2 4.5 0.70 75 

“ 30 1436-7 c 10yR5/2 4.1 0.45 90 

Jug Point 2 Test Pit 1 10-20 1437-1 sic (occupation 10yR4/1 4.4 0.69 40 

(3AS307) level) 

Notes: (1) Texture abbreviations as follows: v=very, f=fine, c=clay, l=loam, si=silt, s=sand; see Table 19 for dimensions of particle size 

(texture) classes; texture estimated in the field except for all 12 samples from Marie Saline (3AS329) which were determined in 

the laboratory; the association of soil samples with visible natural or cultural features or zones noted in this column; association 

with artifacts is not indicated, but only the following samples represent “sterile” subsoil: 1433-4, 1434-3, 1436-3, 1436-7. 

(2) Munsell Soil Color determined in the field from moist profiles. 

(3) Soil reaction scale as follows:

Soils, Sediments, and Chronology 207 

We chose to measure (or in some cases estimate) soil properties which can indicate 

cultural activity when site stratigraphy and soil variability are reasonably well understood 

(color, texture, pH, organic carbon, organic phosphorus, and other characteristics). In this 

case test pit profiles (the source of most soil samples) have been carefully studied, but testing 

was limited on all sites, and data for only 33 samples from seven sites are reported in 

Table 17. 

Our general conclusions from available soil test data are as follows: 

1. Samples from Ouachita River and Saline River floodplain sites, which are not 

modified by cultural activity, are probably homogeneous and indistinguishable 

Guyton soils; however we have not done any offsite soil sampling to support 

this observation. 

2. Soil reaction (pH) values are generally in the extremely acid to strongly acid 

range, except where organically enriched midden deposits range from medium 

acid to neutral values. 

3. High percentages (1-4%) of organic carbon among these samples generally 

distinguish midden deposits and humic soil zones; wood charcoal is probably 

the principal constituent in midden samples measured by this test (ignition 

loss); not all visibly charcoal stained features and occupation levels produced 

high percentages. 

4. High concentrations of organic phosphorus (>75 µg/g) are clearly associated 

with midden deposits and cultural features in Saline River sites; this concentration 

presumably reflects deposition and decay of human wastes and/or 

animal residues; very high concentrations of organic phosphorus (750-1500 

µg/g) characterize midden samples from False Indigo (3AS285) and Jug Point 

Cutoff (3BR76), which we have interpreted as fish processing camps; this is an 

interesting result which requires further tests and comparisons. 

5. It is disappointing to note that Marie Saline (3AS329 , a stratified multicomponent 

site on the Ouachita River, does not furnish good indications of early 

cultural levels or features based on soil samples and soil properties tested 

(Table 18); in fact this site is notably deficient in organic matter (although we 

have observed and collected charcoal concentrations); there are two medium 

acid pH values at 130 and 150 cm in Test Pit 1 which could reflect known early 

occupation levels, but no high or even moderate carbon or phosphorus values 

are recorded for these same samples as would be expected. Marie Saline 

site soils, generally strongly acid, may be deficient in organic matter because 

cultural occupations were of low intensity and principal occupation zones are 

greater than 2000 years of age.


sediMent textures And floodplAin AlluViAtion 

Field observations and soil tests suggest to us that local and regional processes of 

floodplain development can be investigate and elucidated in conjunction with archeological 

site survey, testing, and excavation. Based on information presented in Chapter 2 and elsewhere, 

we propose that channel patterns in the Felsenthal Project area have remained stable 

for centuries, that the Ouachita River has carried a relatively small load of fine suspended 

materials, and that overbank sedimentation, rather than lateral accretion, is the dominant 

process of floodplain development. These are key concepts for understanding why regional 

floodplain sites are buried, why they are not destroyed by lateral erosion, and why “floors,” 

or activity loci with arrays of artifacts and debris reflecting human behavior many centuries 

ago, may persist in relatively undisturbed condition. 

In the case of Marie Saline (3AS329) we had limited opportunity (Test Pits 1-8) to look 

at sediments to 1.5 m depth below the modern surface of the Ouachita River levee. We know 

this site was occupied or reoccupied sporadically for about 3000 years, possibly more, and 

we have briefly suggested that Marais Saline Lake nearby (Figure 3 ) was the mouth of the 

Saline River early in this interval (Chapter 6). Therefore, do the sediments enclosing cultural 

remains at Marie Saline record processes of floodplain development or changes in these 

processes (and in site environment) through time? 

The limited attempt we made to investigate these questions included collection of 

sediment samples from carefully studied test pit profiles and particle size analysis of these 

samples (among other tests reported earlier in this chapter). These analyses were run at the 

Soil Characterization Laboratory, Department of Agronomy, University of Arkansas, and 

data are presented in Table 18 and Figure 56. The provenience of the 12 samples is Test Pits 1 

and 6, which are 100 m apart near the modern bankline, and vary 40 cm in surface elevation 

(Figures 36, 38). 

Our general conclusions from data in Table 19 are as follows: 

1. Samples are generally fine grained loams, or even finer silty clay or clay; the 

two coarsest samples are sandy loams, one surficial and modern (1415-1), the 

other at 130 cm depth (1415-5); all other samples contain less than 52% sand, 

usually fine sand or very fine sand; on the whole these sediments and their 

fabrics seen in profile walls are attributed to overbank accretion; point bar 

deposits (variable in particle size range, sorting and fabric) do not appear in 

these test pit profiles, so far as we can tell. 

2. Test Pit 6 samples (all loams) are somewhat more homogeneous in texture 

than Test Pit 1 samples; also no distinctive sediment lens or break in the sequence 

of sediments can be correlated between test pits (although we correlate 

cultural stratigraphy in Figure 39).

Table 18. Particle Size Frequencies by Depth in Sediment Samples from Test Pits 1 and 6, Area C, Marie Saline (3AS329). 

Sand (mm) Silt (um) Clay 

Sample VCS CS MS FS VFS TS CSI MSI FSI TSI TC TEX 

Depth (cm) Number 2-1 1-.5 .5-.25 .25-.1 .1-.05 2-.05 50-20 20-5 5-2 50-2


Figure 56. Estimated rate of alluviation for the natural levees of the Ouachita and lower Saline 

rivers in the Felsenthal Project area, based on riverbank sites. The hatchured 

zone includes the best available, but slightly scattered, data points, which generally 

describe a curve of decreasing slope. Local conditions of alluviation may 

cause some scattering of data, but the trend is apparent.


3. There may be a slight trend for coarser grain size upward in these sample columns, 

but it is not clear or consistent; this trend, if real, could be due to changing 

fluvial processes, but it also could be due to degradation of older, more 

deeply buried particles; surficial samples (


3. Most importantly, in the floodplain and natural levee environment of the 

Felsenthal Project area, depths are consistent with attributed ages of cultural 

assemblages, and depth observed or measured should predict the age of an 

newly discovered site or site level in this environmental context. 

We reach similar conclusions with regard to channel stability and floodplain development 

in the Felsenthal Project area (especially for the Marie Saline site) as those stated by 

Ritter et al. (1973:375) for the Faucett site on the Delaware River. 

The Faucett site sequence is devoid of point-bar sediment...Evidence of abundant 

scour and fill, normal in point-bar deposits, has not been observed....It becomes 

clear that the Delaware River has maintained its present valley-bottom position 

for an extremely long time [otherwise] older sediments would have been disturbed 

and the earlier records of human occupation rendered unintelligible....The 

Faucett sequence demonstrates that a natural channel in unconsolidated material 

can be fixed for very long periods of time and that under those conditions overbank 

accretion becomes the dominant process involved in the development of 

the floodplain. 

rAdiocArbon dAtes for floodplAin sites 

As indicated earlier in this chapter, the radiocarbon date obtained from floodplain sites 

gave ages much greater than expected, and the results are in fact unacceptable to us in light 

of present knowledge of regional prehistory. We suspect that contamination of samples is 

involved, but the ordinary outcome of contamination is to produce younger than expected 

ages and erratic or inconsistent results where samples were obtained in a stratified context. 

The dates reported here are both too old and inconsistent with stratigraphy. We have attempted 

to examine the various possibilities for contamination with older carbon, and discussed 

these possibilities with the laboratory involved; our queries and suggestions to the 

laboratory scientists and their replies are reproduced later in this section. 

Table 19 characterizes the five samples assayed and presents both the laboratory determinations 

and our estimates of accurate sample age. (The radiocarbon laboratory which 

furnished the results was Beta Analytic, Inc., Coral Gables, Florida.) Since charcoal and other 

organic residues are exceedingly sparse or rare in occurrence throughout the project area 

and in nearly all floodplain sites examined or tested, we submitted a small series of samples, 

each sample also of relatively small size. All samples were examined by us under a microscope, 

concentrated or “picked” for extraneous matter in preliminary fashion, and identified 

as wood charcoal of one or another taxon or group (see Appendix B. Those samples (1415

Table 19. Characteristics and Results of Age Determination for Radiocarbon Samples from Marie Saline (3AS329) and Jug Point 2 

(3AS307). 

Age Determination 

Site Name Project Sample Sample Laboratory (C-14 years 

and Designation Number Provenience Description Number B.P. uncorrected) Expected Age* 

Marie Saline 1415-5-4 Collected from riverbank adjacent Wood charcoal in 8 x 25 cm lens; Beta-1880 2945 ± 100 Early Tchula period, 

(3AS329) to Test Pit 4, Area C, at 1.0 m depth no direct association of artifacts (or 995 B.C.) ca. 500 B.C.± 

(19.4 m elevation) 

Marie Saline 1415-11-7 Test Pit 6, Area C, at 0.8 m depth Wood charcoal fragment or chunk Beta-1882 5390 ± 170 Early Tchula period, 

(3AS329) (19.4 m elevation) associated with novaculite flakes (or 3440 B.C.) ca. 500 B.C.± 

Marie Saline 1415-3-1 Test Pit 3, Area C, at 0.7 m depth Wood charcoal in small concentra- Beta-1879 2650 ± 215 Early Baytown-Coles 

(3AS329) (19.5 m elevation) tion; no association of artifacts (or 700 B.C.) Creek period, ca. 

A.D. 300± 


Marie Saline 1415-7-6 Test pit 7, Area C, 0.4-0.8 m depth Scattered wood charcoal associ- Beta-1881 3410 ± 200 Baytown-Coles Creek 

(3AS329) (19.4-19.8 m elevation) ated with novaculite flakes (or 1460 B.C.) period, ca. A.D. 500± 

Jug Point 2 1437-7-2 Test Pit 1, Area C, 0.2 m depth Wood charcoal scattered in thin Beta-1883 1540 ± 85 Mississippi period, 

3AS307) (20.6 m elevation) occupation level; associated with (or A.D. 410) early Caney Bayou 

Bassett point and both clay- and phase, ca. A.D. 1400± 

shell-tempered sherds 

* Expected age based on natural and cultural stratigraphy within sites and on external correlation (Chapter 3)


series) selected from stratified components at Marie Saline (3AS329) were essentially the 

only adequate ones available here in terms of sample size and association; this site is sparse 

in artifact content and features and deficient in organic matter, based on results of testing, 

but is potentially one of the most important multicomponent sites in the region (Chapter 6, 

Appendix C). We intended the four samples from Marie Saline to verify stratigraphy observed 

and to date the major ceramic components represented here. These objectives failed, 

as the age determinations appear to run from 500 to 3,000 years too old. 

The sample selected from Jug Point 2 (3AS307) was well associated with a presumed 

early Caney Bayou phase occupation level and artifact assemblage. We examined and concentrated 

this sample in the manner indicated above, and it appeared to be quite adequate 

in size. Sample 1437-7-2 was intended to provide the first date for this late Mississippi 

period phase in the region. In this case also, the age determination appears to fall 1,000 years 

too early (Table 20). 

In the following section we quote correspondence (E. T. Hemmings letter to Murray 

Tamers, December 3, 1980) and paraphrase a telephone conversation (Jerry Stipp to Hemmings, 

December 15, 1980) regarding the problematical radiocarbon results from these two 

floodplain sites. 

1. All five samples are from buried levels in an overflow bottomland where 

deep, prolonged inundation has occurred annually for a long time; the soils 

are siliceous silty clay loams, gleyed or reduced, extremely acid (pH—4.0-4.5); 

is some unusual soil chemical process occurring that could introduce old carbon? 

Laboratory reply: We have no specific knowledge of such processes; in one 

case two water-logged stumps at the same level in Mississippi River alluvium 

produced inconsistent dates, and these dates also could not be replicated. 

2. Oil and gas fields are mapped all around this bottomland area and upstream; 

does methane gas (or any other dead carbon compound) move vertically or 

laterally in recent alluvium, river water, or ground water? No liquid or solid 

petroleum residues have been noted in our work in floodplain sediments. 

Laboratory reply: Again, we cannot specifically evaluate such processes; we 

do know that methane or marsh gas can be produced in wetlands and swamps 

by decomposition of vegetal matter. We also know of one site on the Gulf of 

Mexico shoreline where pitch residues produced inconsistent results; in this 

case special extraction of these residues was possible. 

3. Lignite outcrops in the riverbed some miles upstream, but we believe that 

all carbon materials examined in our excavations (and under the microscope 

later) are just wood charcoal of cultural origin. Redeposited lignite has not


been observed in our work in this floodplain, but I suppose finely divided 

material could be present? 

Laboratory reply: Lignite particles could occur in alluvium and could resemble 

more recent wood charcoal strongly; also, hypothetically, there could 

be soluble carbonaceous extracts of lignite as a contaminant in archeological 

samples. 

4. All five samples were collected and stored in amber colored plastic vials for 

a year. These vials have not been used for C-14 samples by me previously. 

Could the vials be degassing some antique carbon? 

Laboratory reply: The few molecules of carbon released in such a process 

would be an insignificant source of contamination (aluminum foil, used widely 

to collect and store samples has a thin oily film which is evidently insignificant 

as a contaminant). 

5. Redeposition of older charcoal or reuse of older wool is not likely to be a factor, 

I think. The local floodplain sediments are devoid of wood and carbonized 

wood, except possibly below the permanent water table and river stage. 

Laboratory reply: Nevertheless, transportation and redeposition of older 

charcoal or wood can occur in floodplains; because of this possibility scattered 

particles or fragments are poor sampling material compared with obvious 

concentrations or features in place. 

6. Laboratory error resulting in older dates? I have no idea what to suspect here. 

Can you discuss this possibility in relation to these samples specifically? The 

actual counting times for each sample should be included in my report. 

Laboratory reply: The samples were given our usual acid-alkali-acid pretreatment. 

They were all small. We obtained less than one gram benzene for four 

samples, while Beta-1883 was borderline (0.9676 g benzene). Counting times 

were extended (2500-2900 minutes for the smallest samples) and 2400 minutes 

for Beta-1883. We have no reason to suspect laboratory error produced older 

dates. 

7. The ages reported are accurate and the archeological data are incorrectly interpreted? 

As I indicated...this is not likely to be the case. We have abundant correlative 

age data from the Lower Mississippi Valley area and also some from 

this region that simply make these dates too early. 

Laboratory reply: You, of course, are in the position to make this judgment;


the laboratory always respects the information obtained from field context and 

regional study by the archeologist or geologist. 

We have introduced these detailed comments to show that special problems of contamination 

may exist in the Felsenthal Project area. A well designed problem oriented, chronometric 

study should be integrated with future work in floodplain sites (Appendix C). 

The failure of our series of dates means that we lack fine scaled chronometric data for 

some of the most important project research results, including the following: 

1. chronometry of multiple components at Marie Saline (3AS329) (Chapter 6) 

2. chronometrically established rates of alluviation (Figure 56) 

3. chronometric data for recent floodplain and riverine features (Chapter 2) 

4. chronometrically established arrow point, dart point, and ceramic sequences 

(Chapter 8) 

5. independent age data for seriation of ceramic temper types and differential 

use of chert/novaculite raw materials (Chapter 8). 

However, we have attempted to make maximum use of stratigraphy, seriation, and 

cross-dating (regional and extraregional correlation) to provide temporal ordering for sites, 

artifacts, and assemblages in the Felsenthal Project area, as indicated at various points in this 

report. This ordering will undoubtedly be tested and refined by future work. 

site trAnsforMAtionAl hypotheses 

Data presented in this chapter and earlier in this report (Chapters 5, 6) are relevant to 

the last problem domain stated in our research design, site transformations. The transformations 

of concern here are certain “postdepositional changes in site and artifact morphology 

caused by noncultural processes” (Schiffer 1976:15). In the context of floodplain sites in our 

project area, we predicted the outcome of alluviation on site survey results (Hypothesis 7) 

and of soil chemical processes on site content (Hypothesis 8). These hypotheses were stated 

in full as follows: 

Problem Domain 4: Site transformations 

H 7 If geologically Recent topstratum deposits have aggraded 20 feet or more in 

the Ouachita Valley, methods of intensive survey in the floodplain will fail to 

sample adequately early occupation sites. 

Test Implications: Recognition of age of sediments drowned and exposed; discovery 

of buried sites in datable context. 

Data Requirements: Significant disparity in observed and expected occurrence of 

early sites; geochronological controls.

Soils, Sediments, and Chronology 217 

H 8 If differential preservation of organic matter occurs within or between sites, 

soil acidity has adversely affected the archeological context. 

Test Implications: Recognition of preservation gradients by sampling soil and refuse; 

demonstrable loss of organic matter. 

Date Requirements: Translation of preservation gradient to a numerical scale; significant 

positive correlation with pH (low soil pH—low degree of preservation). 

With regard to Hypothesis 7, we have compiled considerable data on the occurrence of 

buried prehistoric sites and identification of components represented (Chapter 5, Appendix 

A). In this chapter we discussed the consistent relationship between depth of components 

in natural levee deposits and ages assigned by virtue of diagnostic artifacts or assemblages 

(i.e., concordance between stratigraphy and cross-dating). This concordance is limited by 

available site data to the upper 1.5-2.0 m of topstratum deposits and the last 2000-3000 years. 

Site Units 1 and 4 (described in Chapter 5) comprise 27.4 and 18.8 km reaches of the 

Ouachita River adjoining at river mile 254. Site Unit 1 was longer, had substantially higher 

banks, and had deeper exposure of topstratum deposits (because of the existing navigation 

pool) than Site Unit 4 downstream. The following components were recorded in each (Chapter 

5): 

Site Unit 1 

Mississippi 0 




Archaic 8 

Unknown 6 

18 

Site Unit 4 

Mississippi 5 




Late Archaic 1 

Unknown 5 

23 

This compilation suggests that preceramic components (before Tchula period or about 500 

B.C.) are relatively more abundant in Site Unit 1 even though total components are fewer. 

However, unidentified components in both site units obscure this comparison. One of the 

Archaic components in Site Unit 1 is 3UN162 (Figure 25), the deepest artifact occurrence 

recorded (fire-cracked rock at 2.95 m depth). This site points up a weakness in our attempt


to test Hypothesis 7, because we identified such deeply buried components without ceramics 

as Archaic, even where diagnostic dart points were absent. On the whole, however, we 

are confident that older, deeper sites are exposed in Site Unit 1, and that their counterparts 

may be present in Site 4, largely below river level and therefore unrecorded. 

It is also evident from site survey data, that no Paleo-Indian and Early or Middle Archaic 

components (i.e., sites dating before 4000 B.C.) were recorded in the Felsenthal floodplain, 

although such components are known in adjacent upland sites. We can suggest four 

possible reasons for this hiatus: 

1. Such sites are so few they were missed by site survey teams. 

2. No human occupation or use of the floodplain occurred before 4000 B.C. 

3. Fluvial processes have removed such sites, or 

4. Sites occupied before 4000 B.C. are deeply buried and drowned by existing 

river levels. 

We are inclined to reject the second statement and admit that all others may be causative 

factors. The last two statements describe natural transformations of the archeological record, 

including the agent of ongoing alluviation and site burial proposed in our hypothesis. Although 

many of the data cited above support this hypothesis, complex problems of interpretation 

remain to be solved (and early floodplain sites remain to be discovered). Among these 

problems are the 1ack of reliable chronometric data, although we believe that geochronological 

interpretation of floodplain development and site burial processes has been advanced 

by our work. 

Hypothesis 8 predicts that loss of organic materials (faunal and floral remains) in 

floodplain sites will increase with soil acidity. Our limited soil test data indicate that extreme 

acidity recurs in soil samples from floodplain sites and that organic residues are absent or 

rare in such sites except where very local anthropogenic enrichment (middens, pit fills, occupation 

“floors”) has occurred (Table 18). Ideally, this hypothesis should be tested by means 

of intensive soil and faunal/floral sampling within a single component, rather than by few 

samples from many sites. Our observations from local soil test data might be summarized as 

follows: 

pH 3.9-4.3 generally no visible organic remains 

pH 4.4-5.2 traces of charred faunal or floral matter may be present 

pH 5.3-7.0 visibly enriched by organic material, usually charred faunal and 

floral matter, including “black dirt midden” deposits 

We cannot deal more adequately with Hypothesis 8, but suggest that carefully designed soil 

testing during mitigation work (Appendix C) can elicit a better understanding of floodplain 

site transformation, including anthropogenic soil processes.

chapter 8 

lithic And cerAMic AnAlyses 

In this chapter we summarize and report laboratory studies of about 9,400 lithic and 

3,700 ceramic specimens recovered during site survey and test excavation work. Analyses of 

these materials were performed by the author, project staff members, and various personnel 

of the Contract Laboratory, Arkansas Archeological Survey, during 1980 (see Acknowledgments). 

Clearly, we have not been able to carry these analyses as far as we would have liked 

because of time constraints. For example, the binocular microscope was our basic tool of 

lithic and ceramic study, but we have not attempted high magnification use-wear analysis 

of flaked stone tools nor petrographic study of sherd composition and fabric. Many useful 

directions for future work are suggested by our initial studies of floodplain site collections. 

Information presented in this chapter should lead to a better understanding of the 

kinds of tools and debris (preserved and recoverable) which occur in floodplain extractive 

sites, their implications for extractive site behavior, sources of raw material and artifact 

forms indigenous or alien to the region, and other functional or technological considerations. 

Functional analysis was the primary emphasis of laboratory study, and results obtained 

were integrated with site data and analyses presented elsewhere in this report, when 

possible, to explore the archeological dimensions of extractive sites. 

We have also attempted typological and related analyses of projectile points and 

decorated ceramics as a basis for understanding intersite and interregional interactions and 

changes in these interactions through time. In fact, some fundamental data pertaining to 

technological and stylistic change in floodplain site assemblages have emerged, and must 

now be tested in larger upland sites of the region. We know, for example, the proportions of 

chert and novaculite debitage and of clay-tempered and shell-tempered ceramics to expect 

in floodplain sites of various periods. At the end of this chapter two hypotheses pertaining 

to regional culture history are briefly examined and tested through artifact data. For the 

most part, such hypothesis testing will require much more systematic data from sites in both 

upland and lowland settings. 

lithic source study 

From the outset of site survey work and through later stages of test excavation and 

field laboratory operation, we could see that flint-knapping activity on prehistoric flood-


plain sites featured reduction of rounded pebbles and cobbles. Among the larger collections 

of debitage from Mississippi period sites, discussed later in this chapter, brown pebble 

chert is especially prevalent in the form of cortical flakes and occasionally also pebble cores. 

Among collections of debitage from Archaic and early ceramic sites, gray or white novaculite 

flakes, less obviously attributable to pebble origin, are prevalent. As these observations 

accumulated, we designed and carried out a small-scale study of potential local resources 

for stone toolmaking. This study involved the investigation of graveliferous deposits at two 

localities and comparison of results with archeological site data. 

In Chapter 2 we emphasized the fine grained texture of topstratum deposits exposed 

in the floodplains of Ouachita and Saline rivers. These floodplains are virtually “stoneless,” 

without exposed gravel lenses and with small scattered pebbles exceedingly rare. However, 

we know from nineteenth century hydrographic surveys of Ouachita River that gravel 

bars were exposed at low water in predictable locations, especially at the crossover point 

between bends. These riffles, often colorfully named, were among the serious obstructions 

and hazards to low water navigation, but were eliminated in this century by lock and dam 

construction and by dredging. Such channel bars represent the coarse bedload of the river, 

material “stored” for long periods or transported downvalley at a slow rate. Channel bars 

are proposed here as one important source of lithic raw materials for aboriginal inhabitants 

of the floodplain in this region. Also, there may well be a close association between locations 

of certain kinds of prehistoric sites and past locations of gravel bars. Not only toolmaking 

materials, but also rocks for stone boiling and earth oven cookery could be easily obtained 

on gravel bars. 

A second potential source of pebbles or cobbles comprises the series of Pleistocene 

terraces adjoining the project area (Chapter 2), especially where substratum materials are 

exposed or concentrated by erosion. In general, older, more weathered graveliferous deposits 

are exposed at increasing distances and elevations from the modern river channel. One 

may suggest that terrace pebbles or cobbles collected and introduced into the floodplain for 

aboriginal use would be obtained from a proximal, accessible source. However, the intricate 

series of floodplain drainages makes many terrace edges and many upland areas accessible 

by water travel, and it is likely that numerous expedient sources were exploited in this manner. 

Our Ouachita River Valley localities were selected as study and sampling stations 

principally by virtue of accessibility to the Felsenthal Project area. We were unsuccessful in 

locating a gravel deposit in or near the lower Saline River floodplain, where many prehistoric 

sites have been recorded, and where some sites have been tested and analyzed in detail 

(Chapters 5, 6). The place-names at Gravel Ridge, which we have not visited, and Goulett 

Island (galet = pebble, gravel, or gravel bank in Mississippi Valley French) where we have

Lithic and Ceramic Analyses 221 

done minimal survey work, suggest possible lithic sampling stations for future work in the 

lower Saline River area. 

Station 1, Ouachita River Bedload Material. This locality consists of a pipeline crossing 

at river mile 262.6 where dredged bedload material has been deposited on the left bank 

(SW 1/4, Sec. 30, T17S, R11W). The hydrographic survey of 1873 (Ouachita River Survey, 

Sheet No. 11) shows “Cany Mary Shoals” upstream and unnamed shoals, gravel shoals, and 

rapids extending from this location downstream to Five Mile Bend. The dredged spoil examined 

by us included chiefly novaculite pebbles or cobbles and a few sandstone boulders; 

quartz, quartzite, shale, and silicified wood materials were also present, but little or no chert. 

Station 2, Pleistocene Terrace Alluvium. This locality consists of a gravel pit at the inner 

edge of the Prairie Terrace (Fleetwood 1969) about 10 km east of the present Ouachita 

River channel (NW 1/4, Sec. 34, T18S, R9W). The pit walls expose a deep red, sandy clay 

subsoil with gravel lenses increasing in extent with depth. These lenses contained primarily 

chert, sandstone, and quartzite pebbles or cobbles; novaculite, oolitic chert, silicified wood, 

and banded agate were present in lesser quantities or as traces. Chert and other pebbles 

tended to be far more weathered than those from Station 1, having thick cortical zones of 

alteration. 

Methods of Collection and Analysis 

At each station a 1 x 1 m square was placed at any surface area where coarse material 

was visibly concentrated or well represented, and all surface particles of 40 mm or greater 

diameter were collected without excavating. A second square was similarly placed, but at a 

minimum distance of 10 m from the first. Then a grab sample of about 30 pebbles of any size 

was collected in the vicinity, choosing apparently high quality lithic materials or unusual 

materials. This collecting procedure was intended to reproduce potential aboriginal use of 

surficial deposits in two respects. First, gravel bars, or lag gravel concentrated on upland 

surfaces may have fine particles winnowed away, leaving a pavement of coarse particles 

highly accessible. Second, the minimum functional size of pebbles for production of most 

stone tools and useful flakes is about 40 mm, so far as we can tell from actual pebble cores 

and tools in site collections. (The factor of pebble or cobble shape was not considered in our 

procedure.) 

In the laboratory each pebble was measured on the long axis, broken with a rock hammer; 

identified as to silica mineral (Frondel 1962) or rock type (Travis 1955), and subjectively 

evaluated as to knapping quality. The size frequency data and composition for conchoidally 

fracturing materials are distinguished for each sampling station in Figure 57. The 1 x 1 m 

subsamples collected at each station were found to be essentially identical in size and com-


position, and have been presented as a single sample (N=75 in each); grab samples (N=30 in 

each) are not included in Figure 57). 

The following operating definitions were employed in the sorting and identification 

procedure: 

Chert is a silica mineral commonly originating as nodules or beds in marine 

limestone or other sedimentary rocks; generally it is light colored with matte 

to lustrous surfaces; minute fossil remains are common constituents; cortex on 

weathered chert pebbles is variable in color and texture (Frondel 1962:221). 

Novaculite is a fine-grained silica rock derived from chert by thermal metamorphism; 

milky white to gray colors and matte grainy surfaces predominate; fossil 

remains are generally not preserved; “polygonal triple-point texture” is characteristic 

when novaculite is examined by scanning electron microscope (Frondel 

1962:222; Keller et al. 1977). 

Pebbles and cobbles are 2-64 mm and 64-256 mm in maximum diameter respectively 

in terms of the Wentworth size scale (Compton 1962: 213). 

Results of Analysis 

The Ouachita River bedload and Pleistocene terrace samples are similar with regard 

to prevalence of pebbles and small cobbles of conchoidally fracturing siliceous materials in 

the 40-85 mm size range, but the latter sample has larger modal sizes and poorer sorting. 

The composition of these samples is drastically different in that the first sample is entirely 

novaculite and the second largely chert (when sorted according to definitions and attributes 

given above). Many of the freshly broken pebbles exhibit color, texture, and other raw material 

characteristics which can be matched (on a gross visual comparative basis) with flaked 

stone tools or debitage in collections from prehistoric floodplain sites in the project area. 

On the other hand, distinctive novaculite, chert, silicified wood, and crystal quartz varieties 

which recur in several archeological collections are not represented in the samples collected 

and described above. On a subjective basis each sample of 75 pebbles and small cobbles included 

only one or two relatively unweathered, homogeneous specimens which a flintknapper 

might select for reduction. The novaculite tended to be laced with veins and fractures, 

while the chert was deeply weathered with heterogeneous inclusions. We conclude from our 

analysis of both samples that similar local gravel deposits may well have been exploited for 

toolmaking raw materials, but that selectivity, by means of trial flaking tests, would have 

been employed by the aboriginal flintknapper. We make the presumption that our dredged 

and gravel pit samples are representative of past channel bar and Pleistocene terrace gravels 

widely distributed and exposed in the project area.


Figure 57. Histograms of size frequency in two samples of pebbles and small 

cobbles from Ouachita Valley localities.


If this is so, numerous late prehistoric floodplain camps contain pebble chert brought 

from upland source areas. In contrast, Archaic and early ceramic floodplain camps could obtain 

novaculite toolmaking raw material and granular rocks for cookery from channel bars, 

and might be expected to favor riverbank locations adjacent to these bars (Chapter 6). 

lithic AnAlysis And results 

Analyses of stone tools and a variety of inorganic materials proceeded in conjunction 

with other studies of Felsenthal site collections. Only about 400 stone tools or fragments 

were collected, but many of those were recovered in single component sites and in stratigraphic 

context. We have studied these tools intensively, employing functional concepts and 

a classification similar to House’s (1975) Cache Project lithic analysis. Approximately 3,000 

flakes and 6,000 other items of inorganic debris have also been studied from this functional 

point of view. A series of artifact data sheets were devised to meet the particular needs and 

goals of analysis, including the following: 

Lithic Analysis (Flaked Stone Tools) 

Lithic Analysis (Ground Stone Tools) 

Lithic Analysis (Debitage) 

Inorganic Debris Analysis 

These records are retained at the Arkansas Archeological Survey, and results are summarized 

and reported in this section. 

In addition to functional analyses of artifacts and site collections, projectile points were 

examined typologically, and projectile point data were recorded and analyzed. We propose 

sequences or partial sequences of arrow point and dart point types for the region in this 

chapter, based on about 70 specimens from floodplain sites. These sequences are sketchy 

and incomplete, and clearly require further work, but we hope they will be fundamentally 

useful for organizing regional archeological data. 

Arrow Points 

Arrow points are bilaterally symmetrical bifaces with a pointed tip at one end and a 

haft modification (except in some triangular and ovate forms) at the opposite end (House 

1975:60). In large samples of American Indian arrow points, both ethnographic and archeological 

in origin, the range of weight is from about 0.3 to 4.0 g, although heavier examples 

occur (Fenenga 1953; Thomas 1978). The majority of archeological specimens of appropriate 

morphology and size/weight can be presumed to have been hafted and propelled by bow, 

and were therefore associated with hunting, bow fishing, or warfare. Sociotechnic variants


of arrow point styles, distinct in morphology, size, raw material, and craftmanship, are also 

known in much of North America, usually from cache or burial contexts (e.g., in the Great 

Bend of Red River and Arkansas Valley Caddoan regions). 

Introduction of the bow and arrow in various parts of eastern North America, including 

Arkansas, occurred within a few centuries of A.D. 500. In the Felsenthal Project area, this 

introduction must have occurred during the Baytown-Coles Creek period, but no single or 

multiple component floodplain site yet investigated has produced evidence which could 

identify or date this event. Most arrow point styles which recur here in floodplain sites are 

good markers of Mississippi period extractive camps. 

A brief study of arrow point manufacture, based on Shallow Lake site (3UN9/52) lithic 

materials, is reported by Flenniken (in Stacy 1976:Appendix A). Many aspects of this study 

require updating. However, Flenniken experimentally replicated six arrow points of generalized 

Alba and Ashley type, beginning with 42-70 mm long chert and novaculite pebbles 

collected on the site, producing flakes and flake preforms by percussion and pressure flaking, 

and finally producing completed arrow points by pressure flaking. Each arrow point required 

about six minutes to complete from a detached flake (one failure occurred in the final 

stage of pressure flaking). We suspect that some or even many arrow points recovered in the 

Felsenthal region were produced by bifacial reduction from pebble cores, while some arrow 

points were certainly produced on flake blanks in the manner of Flenniken’s experiment. 

More experimentation of this kind is clearly needed to understand regional lithic technologies. 

Our arrow point sample (N = 44) includes 28 complete or nearly complete specimens 

as well as basal and tip fragments. Three-quarters of the identifiable specimens are from 

floodplain sites which have been tested, so that assignment to a component is generally possible 

(Chapter 6). We are able to propose an arrow point sequence which extends throughout 

the Mississippi period (A.D. 1100-1700); only a few arrow points can be attributed to the 

preceding Baytown-Coles Creek period and these few present certain problems of classification 

and context. Table 20 is the first attempt to organize a sequence of arrow point styles for 

the Felsenthal region, and this sequence may well be supplemented and modified by future 

work. Figure 58 portrays this sequence by means of actual specimens recovered in site 

survey or testing work. The recent report on Shallow Lake site (3UN9/52) by Rolingson and 

Schambach (1981) indicates that Catahoula and Rockwall arrow points may also occur in the 

regional sequence, but no examples were obtained in our work. 

The following paragraphs provide brief summaries of descriptive data, typological 

discussion, selected references, and other comments for each of the arrow point types enumerated 

in Table 21. From among more extensive records of our projectile point study, only 

selected attributes are recorded here: N = sample size; L = maximum length or range ob-


Table 20. Proposed Sequence of Arrow Point Types in the Felsenthal Region. 

Regional Type Number of 

(Extraregional Type) Cultural Period/Phase Floodplain Sites Terrace/Island Sites Specimens 

Nodena Terminal Mississippi One Cypress Point (3AS286) 6 

Terminal Mississippi No name (3AS310) 1 

Nodena, Banks var. (Shetley) Terminal Mississippi One Cypress Point (3AS286) 1 

Madison (Fresno) Terminal Mississippi One Cypress Point (3AS286) 1 

Terminal Mississippi Marie Saline (3AS329) 1 

Terminal Mississippi Lapile Creek 4 (3UN158) 1 

Washita (Cahokia) Mississippi/Caney Bayou Marie Saline (3AS329) 1 

Eagle Creek Mississippi/Caney Bayou Jug Point 1 (3AS306) 3 

Bassett Mississippi/Caney Bayou Jug Point 2 (3AS307) 1 

Ashley (Colbert) Mississippi/Caney Bayou Jug Point 2 (3AS307) 2 

Ashley (Colbert) Mississippi/Gran Marais Jug Point Cutoff (3BR76) 1 

Mississippi/Gran Marais False Indigo (3AS285) 2 

Mississippi/Gran Marais Jones Lake North (3UN81) 1 

Mississippi/Gran Marais No name (3AS326) 1 

Alba Mississippi/Gran Marais False Indigo (3AS285) 1 

Homan Baytown/Coles Creek* Goulett Island (3BR8) 1 

Baytown/Coles Creek* Pine Ridge 1 (3UN177) 1 

Honey Creek Baytown/Coles Creek* Small Cane 1 (3BR72) 1 

Baytown/Coles Creek* One Cypress Point (3AS286) 1 

* Homan and Honey Creek arrowpoints may pertain to unnamed phase(s) late in this period; four sites listed here presently lack adequate 

evidence for Coles Creek components.

e 


A b c d 

f 

g h 

i j k l M 

n o p Q 

Figure 58. Arrow points from the Felsenthal Project area in proposed chronological 

sequence. a. Nodena (3AS286); b. Nodena, Banks variant (3AS286); 

c. Madison (3AS286); d. Madison (3UN158); e. Washita (3AS329); f. 

Eagle Creek (3AS306); g. Bassett (3AS307); h. Ashley (3AS307); i. Ashley 

(3BR76); j. Ashley (3UN81); k. Ashley (3AS326); l. Ashley (3AS285); m. 

Alba (3AS285); n. Homan (3BR8); o. Homan (3UN177); p. Honey Creek 

(3BR72); q. Honey Creek (3AS286) (AAS neg. 813868).


served; HL = haft length or range; W = maximum width or range; WT = weight or range; RM 

= raw material count; TA = thermal alteration observed; ( ) = measurement of slightly incomplete 

specimen, not an estimate if specimen were complete. Dimensions are in millimeters 

and weights in grams and tenths. All arrow points recorded measured between 3 and 7 mm 

in maximum thickness. 

Nodena (Bell 1958:64; Morse 1973:Figure 9; Hoffman 1977:Figure 8). N = 6, L = 

(17)-(32), W = 12-14, WT = (0.7)-(2.2), RM = 6 chert, TA = 0. 

Slender, willow leaf shaped points without haft modification, associated particularly 

with late prehistoric/protohistoric cultures in eastern Arkansas, including the historic Quapaw. 

The principal occurrence in our project area is at One Cypress Point (3AS286) on the 

Saline River, evidently a Quapaw (?) site-unit intrusion, ca A.D. 1600-1700. No Nodena 

points have been reported previously from the Felsenthal region. 

Nodena, Banks variant (Morse 1973:Figure 9c). N = 1, L = 22, W = 12, WT = 1.0, 

RM = chert, TA = 0. 

This specimen is a variant of Nodena with concave basal edge, associated with the One 

Cypress Point (3AS286) assemblage. See also Perino (1971:92) on the Caddoan Shetley point. 

Madison (Perino 1968:52). N = 3, L = 17-24, W = 12-17, WT = 0.5-2.5, RM = 2 chert, 

1 novaculite, TA = 2. 

Widely distributed triangular arrow point of the Mississippi period. See also Bell 

(1960:44) on the Fresno point. One of our chert specimens is associated with the Nodena 

point assemblage at One Cypress Point (3AS286) and is comparable in raw material, symmetry, 

and flaking technique. 

Washita (Bell 1958:98; Brown 1976:105). N = 1, W = 19, RM = chert, TA = 1. 

Widely distributed, side-notched, triangular point of the Mississippi period. See also 

Perino (1968:12) on the Cahokia point and variants. 

Eagle Creek (provisional type described here). N = 3, L = 21-15, HL = 8-9, W = 10- 

13, WT = (0.5)-1.3, RM = 3 chert, TA = 0. 

A small slender point with contracting stem, convex blade edges, and slight shoulders. 

Our three specimens are from a single site, Jug Point 4 (3AS306) on the Saline River, 

assigned to the Caney Bayou phase (Chapter 6). At least one contracting stemmed arrow 

point recovered at Shallow Lake (3UN9/52) may also belong in this provisional type (Rolingson 

and Schambach 1981:Figure 15n). A small series of contracting stemmed arrow points, 

unidentified to type, was recovered at the Salsbury site (16OU15) on Ouachita River below 

Monroe, Louisiana (Price and Heartfield 1977:78). No contracting stem arrow point type 

presently defined subsumes the attributes noted above, although the silhouette resembles 

modified gar scale arrow points shown by Suhm and Krieger (1954: Plate 134E). The specimens 

from Jug Point 1 were evidently made on bifacial preforms rather than flakes, are


relatively thick even on haft elements, and could have been inserted in a socket rather than 

a split or notched shaft. Haft element length (HL) approximates one third of the total length 

(L) in these points. 

Bassett (Bell 1958:10; Webb 1980b:3). N = 1, L = 41, HL = 5, W = 22, WT = 3.0, RM = 

chert, TA = 1. 

Basally modified, triangular arrow point with small pointed stem isolated by notching, 

distributed principally in the Caddoan Red River region. Our rather large specimen from 

Jug Point 2 (3AS307) may date early in the Caney Bayou phase or about A.D. 1400-1500, contemporary 

with latest arrow points of this type in the Caddoan area. 

Ashley (Rolingson 1971; Brown 1976:79). N = 7, L = (16)-31, HL = 4-6, W = 13-15, 

WR = (0.6)-1.3, RM = 4 chert, 3 novaculite, TA = 2. 

Triangular bladed arrow point, commonly with recurved blade edges and tiny graverlike 

tip; prominent, expanding, “bulbous” stem or haft element is one fourth total length 

(less in long slender forms). This arrow point type is a marker for the Bartholomew phase, 

A.D. 1100-1400, in southeast Arkansas and the contemporary Gran Marais phase in the 

Felsenthal region, but persists into the early Caney Bayou phase in the project area so far 

as we can tell. Brown (1976:79) extends the type name into the Arkansas Valley Caddoan 

region (Harlan phase) and notes the confusion with Perino’s (1968:34) description of earlier 

Homan points. Ashley points in Arkansas are evidently contemporary with Colbert points 

in northwest Louisiana and overlap morphologically with, or are derived from, earlier Caddoan 

arrow point styles, including Homan (see Wood 1963; Webb 1980a, 1980b). 

Alba (Bell 1958:8; Brown 1976:61; Webb 1980b:2). N = 1, L;= 23, HL = 6, W = 14, 

WT = 0.8, RM = chert, TA = 0. 

Small arrow point, commonly having the blade form described above (Ashley) but 

with essentially rectangular stem; widely distributed in the Caddoan area eastward to the 

Lower Mississippi Valley. Our specimen from False Indigo (3AS285) may be later than the 

A.D. 750-1100 time range commonly ascribed this type, since it is apparently associated with 

Ashley points in a Gran Marais phase assemblage. 

Homan (Wood 1963; Perino 1968:34; Brown 1976:92; Webb 1980b:4). N = 2, L = 14- 

21, HL = 6-7, W = 13-14, WT = 0.3=0.9, RM = 1 chert, 1 novaculite, TA = 1. 

Deeply and relatively broadly corner-notched arrow points having convex basal edge 

and relatively broad triangular blade; blade edges straight, slightly convex, or slightly recurved. 

This arrow point style is quite early, ca A.D. 800-900, in the Ouachita Mountain and 

Red River Caddoan regions, but Brown indicates a later date in the Arkansas Valley. The 

typological confusion discussed by Brown was noted above. Our two specimens were recovered 

on upland sites without known stratigraphic context, and also may not be representative 

of late Coles Creek or Fourche Maline arrow points elsewhere attributed to this type.


Honey Creek (Derley 1979). N = 2, L = (24)-27, W = 16-20, WT = (1.2)-(1.9), RM = 2 

chert, TA = 0. 

This is a provisional type, minimally described from the Honey Creek site (3PR28) 

near the lower White River. Numerous specimens which may be assigned to this type have 

also been recovered at Toltec Mounds (3LN42), Dawson Mound (3CE68), and the Alexander 

site (3CN117) in the central Arkansas Valley, and at Roland Mound (3AR30) on the lower 

White River. Honey Creek points are thin, “lozenge-shaped” bifaces with straight or slightly 

concave blade edges and a convex base without hafting modification. In sites enumerated 

above they are most likely associated with a Coles Creek occupation, between about A.D. 

700 and 1100, and some or all forms may be advanced preforms for arrow points of Scallorn- 

Rockwall style. In the Felsenthal region thin well-made bifaces of Honey Creek form can 

either be completed arrow points or advanced preforms for arrow points, lacking only the 

notches. Only additional work on larger better controlled collections will resolve the functional 

and classificatory problems posed by these bifaces. 

dart points 

Dart points have the symmetry and haft modification (except in relatively few lanceolate 

or ovate forms) noted above for arrow points, but almost invariably exceed 4.0 g in 

weight. Archeological specimens generally date before A.D. 500 in eastern United States, 

and can be presumed to have been fixed to a dart foreshaft or simple, light spear shaft which 

was cast by means of an atlatl. Some special forms may have been cast javelin-style or thrust 

as lances. Dart points were thus components of equipment associated principally with 

hunting and occasionally, perhaps, with warfare. As in the case of arrow points, sociotechnic 

variants may also occur (e.g., Ohio Valley Adena stemmed points in mortuary context). 

Moreover, dart points of many styles were quite often refurbished, or modified and recycled, 

largely by virtue of the large mass of raw material still available in damaged points. 

Our sample of classifiable dart points is limited (N=26), chiefly because of the relatively 

few preceramic and early ceramic components recorded or investigated in the project 

area (Chapter 5). In Table 22 we summarize data for 20 complete or nearly complete dart 

points in typological categories and in general temporal groupings. Figure 59 illustrates 

this sample by means of specimens recovered in site survey and testing work. Most of our 

specimens are readily classified in types well known from the Lower Mississippi Valley 

(Ford and Webb 1956) or from the West Gulf Coastal Plain and Southern Plains in Louisiana 

and Texas (Suhm and Krieger 1954; Webb 1980b; see also Bell 1958, 1960; Perino 1968, 1971). 

In general, precise dates for these types have not been obtained and some types (e.g., Gary) 

persisted for centuries or even several millenia. Our broad temporal groupings in Table 21 

are thus based on extraregional data of little precision, but also, in a few cases noted below, 

on stratigraphic or site assemblage data from our own floodplain sites:

Table 21. Metric and Nonmetric Data for Dart Points and Proposed General Sequence of Dart Point Types in the Felsenthal Region. 

Descriptive Data* Provenience 

Regional Types L HL W T WT RM TA G Floodplain Sites Terrace/Island Sites 

Ceramic (ca. 500 B.C. - A.D. 500) 

Gary (small) 37 13 17 11 5.0 Ch Buttonbush (3BR58) 

Langtry (38) 13 28 9 (8.9) Ch X Marie Saline (3AS329) 

Steuben 48 15 24 10 9.1 Ch X Marie Saline (3AS329) 

Preceramic/Ceramic (ca. 2000 B.C. - A.D. 1) 

Gary - 19 (27) 9 (6.4) No X X Marie Saline (3AS329) 

Gary (52) 16 33 7 (10.7) No X Mouth of Eagle Creek (3BR78) 

Gary (39) - 24 9 (7.7) Ch X Lapile Creek 4 (3UN158) 

Gary (35) 14 21 7 (5.1) Ch Lapile Creek 5 (3UN157) 

Gary (37) - 20 a (5.6) Ch X Lapile Creek 5 (3UN157) 

Gary 47 20 25 9 8.5 Ch Lapile Creek 7 (3UN151) 

Gary (43) 17 32 9 (14.5) No X X Bolding Road (3UN175) 

Ellis 60 9 33 10 16.8 No X No name (3AS324) 

Ellis 37 11 26 8 6.2 Ch Burnie Creek (3UN132) 

Ellis 36 9 27 8 6.0 Ch Cross Road near Burnt Bridge (3UN72) 

Marcos (32) 7 30 10 (5.8) Ch Kelly Spring Creek (3UN173) 

Delhi 52 7 29 10 10.9 No X X Potlatch Field (3AS160) 

Macon 77 18 28 12 21.0 No S Small Cane 2 (3BR73) 


Preceramic (ca. 3000 -1000 B.C.) 

Big Creek (Williams) 48 15 37 10 14.0 Ch CoulettIsland (3BR8) 

yarbrough 50 20 27 11 14.2 No X Marie Saline (3AS329) 

Morhiss - 17 28 11 (15.5) No X MarieSaline (3AS329) 

Carrollton 34 11 22 8 5.0 Ch X Lapile Creek 7 (3UN151) 

* L = maximum length in mm; HL = haft length in mm; W = maximum width in mm; T = maximum thickness in mm; WT = weight in grams; RM = raw material, either chert (Ch) or 

novaculite (No); TA = thermal alteration observed; G = grinding on basal edge, stem edges, or notches


A b c d e 

f 

g 

j k l M 

o 

Figure 59. Dart points from the Felsenthal Project area in proposed chronological 

sequence. Specimens are shown in the same order as Table 22. 

(AAA neg. 813869) 

h 

p Q r s 

i 

n


1. A small Gary point surface find (Figure 59a) from Buttonbush (3BR58) may be 

associated with Baytown Plain ceramics and may represent occupation early 

in the Baytown-Coles Creek period (A.D. 300-1100); this specimen compares 

closely with Schambach’s (1970:200) Gary var. Camden assigned to the Oak 

Grove phase (300 B.C.-A.D. 300) in the Mid-Ouachita region (see also Rolingson 

and Schambach 1981:99). 

2. A small dart point, essentially the Langtry type but with markedly slender 

stem (Figure 59b), was recovered at Marie Saline (3AS329) in a deep undisturbed 

level (70-80 cm) with one Baytown Plain sherd, and may also represent 

occupation early in the Baytown/Coles Creek period; Langtry points are related 

to the variable cluster of Gary point forms which persist into the first millenium 

A.D. in Oklahoma (Baerreis et al. 1958) and elsewhere. 

3. An expanding stemmed point (Figure 59c), highly similar to the Steuben type, 

was recovered in yet a deeper level (90-100 cm) at Marie Saline (3AS329) without 

other diagnostic artifacts directly associated, and could pertain to an otherwise 

“invisible” Marksville period (ca A.D. 1-300) occupation (on the basis of 

depth alone we would assign this point to the preceding Tchula period); Steuben 

points were described by Morse (1963) in late Middle Woodland context 

(ca A.D. 500) in Illinois, but none have been identified previously in Arkansas. 

4. A large, well-made Gary point (Figure 59e) surface find at Mouth of Eagle 

Creek (3BR78) may pertain to preceramic Poverty Point period (1200-500 B.C.) 

occupation. 

5. A crude corner-notched Ellis point (Figure 59j) was recovered in place at 2.0 m 

depth without associated artifacts at riverbank site 3AS324, and may pertain to 

the Late Archaic period (ca 4000-1200 B.C.). 

6. A long, well-made Macon point (Figure 59o) recovered on the surface with 

fire-cracked rock at Small Cane 2 (3BR73) may pertain to Poverty Point or Late 

Archaic period occupation. 

7. A Big Creek (or Williams) corner-notched dart point (Figure 59p) from the 

surface at Goulett Island (3BR8) may pertain to Poverty Point or Late Archaic 

period occupation; Big Creek points are estimated to date at about 750 B.C. and 

Williams points at about 2500 B.C. (Morse 1970; Schambach 1970:152; Perino 

1971:10; Rolingson and Schambach 1981:99). There seems to be little basis for 

distinguishing these as regionally distinct types of great age difference. 

Rolingson and Schambach (1981:Table 15) describe a series of dart points from the 

Shallow Lake site (3UN9/52) in a private collection and also provide a tentative sequence of


dart point types for the Felsenthal region. There is relatively little overlap in types identified 

and ordered between their work and ours, partly because of sampling error and partly 

because we have necessarily dealt with the last 3500 years of such a sequence. However, it is 

clear that a detailed regional sequence is emerging. 

In Figure 60 we show flaked stone tools with the haft modification of dart points. Two 

specimens represent asymmetrical blades adapted to cutting tasks, one has a scraping edge 

prepared on a large thick blade, and two specimens have blades markedly reduced by resharpening. 

Bifacial Preforms and Other Bifaces 

Bifacial artifacts other than projectile points (N = 88) were frequently recovered in site 

survey and testing work. One site collection has an excellent series of these items (3AS286), 

but most were scattered finds in small sites of all time periods (or of unknown age). A particular 

problem of biface analysis is the high frequency of fragments for which only limited 

observations can be made. We can distinguish preforms as an important functional category, 

but other classes of bifacial tools and also multipurpose tools are present in floodplain sites. 

Preforms are symmetrical, generally ovate or trianguloid bifaces which lack the fine 

marginal/facial retouch and haft modification of projectile points, and also lack prepared 

and/or utilized edges of bifacial cutting tools. Cortex and thick cross section may distinguish 

early from advanced stages of preform reduction. Aborted preforms commonly exhibit raw 

material irregularities and breakage (House 1975:67). The presence of preforms in archeological 

assemblages indicates manufacture, or at least initial stages of manufacture, of bifacial 

tools such as projectile points. 

Bifacial cutting tools may or may not exhibit bilateral symmetry and haft modification, 

but will commonly have sinuous to regular lateral edges with edge angles less than 35° and 

evidence of use-wear on these edges (Wilmsen 1968:156). Depending on results of use-wear 

analysis, assemblages with such cutting tools may reflect butchery activity. 

Bifacial choppers are relatively heavy bifaces or cobbles with one or more cutting 

edges, having edge angles in the 45-65° range and evidence of pronounced edge wear or impact 

damage (House 1975:62). Such tools are likely to have been produced and employed in 

expedient fashion for multiple tasks, including food processing and working of tough plant 

or animal materials. 

Figure 60 illustrates bifacial preforms and other bifacial tools from various sites in the 

project area. In Chapter 6 the sample of flaked stone tools from One Cypress Point (3AS286)


Figure 60. Bifacial tools, preforms, and cores. a. hafted knife, Shallow Lake West 

(3UN154); b. hafted knife, Lapile Creek 8 (3UN152); c. hafted scraper, 

Small Cane 1 (3BR72); d. resharpened dart point, Lapile Creek 2 (3UN160); 

e. resharpened dart point, Goulett Island (3BR8); f. arrow point preform, 

Goulett Island (3BR8); g. arrow point preform, Persimmon 3 (3AS315); h. 

dart point preform, Lapile Creek 4 (3UN158); i. dart point preform, Lapile 

Creek 8 (3UN152); j. preform, Willow Oak (3AS312); k. preform, Persimmon 

2 (3AS314); l. preform/chopper, unnamed site (3BR88); m. chert pebble core, 

Goulett Island (3BR8); n. novaculite pebble core, Marie Saline (3AS329); o. 

silicified wood tabular core, Goulett Island (3BR8) (AAS neg. 813870).


A 

h 

l 

b 

i 

c d e 

n 

M 

j 

f g 

o 

k


includes arrow point preforms and other bifaces pertaining to a single terminal Mississippi 

period component (Figure 48). 

Other Flaked Stone Tools 

A variety of other modified flaked stone tools, both bifacial and unifacial, was recovered 

in site survey and testing work; the number from any single site is small and the total 

number considered here is also relatively small (N = 41). The collection from One Cypress 

Point (3AS286) again contains a greater variety of specialized flaked stone tools, examples of 

which are shown in Figure 48. A generalization we can draw from all floodplain site collections 

is that projectile points, point fragments, and preforms which reflect use, manufacture, 

or refurbishing of weapons are relatively more numerous than maintenance types of tools 

(Binford and Binford 1966). Four functional types of flaked stone tools are described below. 

Scrapers are edged tools which may or may not have haft modification, but include 

a moderately steep working edge, generally in the 45-55° edge angle range (Wilmsen 

1968:156). Placement and shape of working edge define certain common scraper 

types: end scraper, side scraper, spokeshave, plane, and other (see House 1975: 62). 

Also, both bifacial and unifacial scraper forms occur, although plano-convex cross 

sections and unifacially prepared working edges predominate. Scrapers are associated 

with a variety of processing and maintenance activities, including hide working, 

retting of fibrous vegetal material, and shaving of wood, bone, and antler. The 

few identified scrapers or scraperlike tools are listed below: 

1. Two small chert endscrapers associated with the terminal Mississippi period 

component at One Cypress Point (3AS286), and considered to be diagnostic of 

Nodena or Quapaw phases in other areas (Chapter 6, Figure 48c). 

2. Side scraper on a chert flake, Marie Saline (3AS329), probably Poverty Point or 

Tchula period (Figure 61a). 

3. Discoidal scraper on a chert flake, Eagle Creek 6 (3BR79) probably Mississippi 

period (Figure 61b). 

4. Spokeshave on a novaculite flake, Pine Island Road (3AS290), probably Archaic 

(Figure 61c). 

5. Plane or heavy scraping tool on chert cobble, Mud Lake Bend 2 (3UN166), 

probably Archaic. 

Drills and perforators are bifacial or sometimes minimally retouched flake tools 

with an elongate bit, usually bearing traces off rotary wear. Bit wear is a function 

of drilling implement hardness relative to raw material hardness and also duration 

of implement use; transverse bit breakage is frequently observed. Drills


commonly have haft modification and a bit approaching cylindrical form; perforators 

generally lack haft modification and have a short, contracting bit with 

alternately beveled edges. This class of tools can be associated with fabricating 

tasks, including piercing or reaming wood, bone, shell, and other moderately 

hard materials. Only three drills and five perforators were recovered in site 

survey and testing work and most of these are from One Cypress Point (3AS286). 

The following specimens can be assigned to particular components: 

1. One cylindrical chert drill and a cylindrical chert drill tip associated with the 

terminal Mississippi period component at One Cypress Point (3AS286); the 

cylindrical drill form without haft modification appears to be diagnostic of 

Nodena or Quapaw phases in other areas (Chapter 6, Figure 48d). 

2. Three perforators on chert flakes associated with the same component at 

3AS286 (Figure 48h). 

3. Perforator on a chert flake; Caney Bayou phase component at Jug Point Cutoff 

(3BR76) (Figure 61d). 

4. An expanded base drill on a novaculite flake associated with the Caney Bayou 

phase component (ca A.D. 1400-1500) at Jug Point 2 (3AS307) (Figure 61f). 

Gravers are minimally retouched flakes with one or more small sharp points or 

spurs isolated and formed by unifacial retouch. Hafting modification is absent. 

Use-wear, when distinguishable, consists of back-and-forth striae or rotary wear 

or both (cf. House 1975:64). Gravers are therefore associated with fabricating 

tasks, including engraving, incising, or punctating moderately hard raw materials 

such as wood, bone, and shell. Only one graving tool, a small chert flake with 

two delicate spurs (Figure 61g) was recovered in our work. This specimen is 

tentatively associated with a Caney Bayou phase component at Overcup Oak 3 

(3AS305). 

Retouched flakes have one or more margins trimmed or regularly flaked to 

produce a working edge or to refurbish a previously dulled edge. Such edges 

may have been employed in a variety of cutting/scraping activities, may show 

heavy to undetectable use-wear, and may be especially difficult to distinguish 

from unintentional edge-wear which takes a regular form (shearing). Moderately 

steep working edges intergrade with those of tools identified as scrapers. Retouched 

flakes are most reliably associated with processing or fabricating tasks 

when a large series is available for study within a single site assemblage. Our 

sample (N = 22) does not meet this requirement in that specimens were recovered 

infrequently in widely scattered sites. Examples are illustrated in Figure 48i and 

Figure 61g-i.

A b c d e 

f 

g 

h 


Figure 61. Miscellaneous flaked stone tools. a. side scraper, Marie Saline (3AS329); b. 

discoidal scraper, Eagle Creek 6 (3BR79); c. spokeshave, Pine Island Road 

(3AS290); d. perforator, Jug Point Cutoff (3BR76); e. perforator, Redeye Lake 

Prairie (3AS293); f. expanded base drill, Jug Point 2 (3AS307); g. graver, 

Overcup Oak 3 (3AS305); h. retouched flake, Pine Island Road (3AS290); i. 

retouched blade or bladelike flake, Marie Saline (3AS329); j. utilized flakes, 

Lapile Creek 7 (3UN151). (AAS neg. 813871) 

i 

j


Utilized flakes are characterized by irregular edge damage and/or use-wear 

which cannot be attributed to purposeful modification of tool form. Utilization 

can be so brief and traces of this utilization so minimal that presumed expedient 

tools intergrade with waste flakes modified by cultural or natural processes not 

directly related to tool use, e.g., postdepositional dumping and trampling (Flenniken 

and Haggarty 1979). Our sample of utilized flakes (N = 188) is large, variable, 

and difficult to associate with particular tasks. Utilized flakes constituted 

64% of the flaked stone tool assemblage recovered at One Cypress Point (3AS286) 

where small thin chert flakes or bladelike flakes with nibbled edges are abundant 

(Chapter 6, Figure 48i). We can characterize these tools from the terminal Mississippi 

period component by type of flake selected as follows (see also Cores and 

Debitage below): 

primary flakes 19 

secondary flakes 47 

interior flakes 24 

flakes of bifacial retouch 24 

Total 114 

These utilized flakes are generally of markedly small size (maximum dimension 

= 11-30 mm, mean maximum dimension = 18.9 mm). These and other data for the 

One Cypress Point lithic assemblage suggest that flakes were randomly selected 

from the population of waste flakes, and expediently employed as delicate, disposable, 

cutting tools. 

Utilized flakes were isolated by inspection during sorting and were examined under a 

binocular microscope. Detailed study at high magnification of use-wear on tool edges was 

beyond the scope of this study. Other examples of utilized flakes or bladelike flakes from 

floodplain sites are shown in Figure 61j. The distinction between utilization and postdepositional 

damage among these specimens requires additional microscopic examination and 

analysis. 

Cores and Debitage 

Lithic waste resulting from flaked stone tool production or refurbishing is termed debitage, 

and any substantial assemblage of this material may include three subcategories: 

Cores or masses of conchoidally fractured raw material which have two or more 

flake removals, recur in a variety of distinctive forms, and exhibit core preparation 

but not use-wear; such cores occur in archeological assemblages as the result 

of discard (rejected and exhausted cores) or merely through loss (House 1975:65).


Waste flakes of conchoidally fractured raw material which have one or more attributes 

associated with controlled flake removal (platforms, dorsal and ventral 

surfaces, etc.); such flakes may occur complete or as fragments, but lack evidence 

of use-wear; an assemblage of waste flakes attests to flintknapping activity in 

general, and variability within this assemblage may adduce the range of raw 

material types in use, forms undergoing reduction, and sequence of knapping 

techniques employed (House 1975:67). 

Shatter, or blocky fragments of conchoidally fracturing raw material which lack 

all flake characteristics; such fragments are ordinarily present as a minor fraction 

of any substantial assemblage of debitage and are presumed to originate during 

pretreatment or reduction of raw material. 

Debitage (N = 3256) was nearly ubiquitous on floodplain sites and other prehistoric 

sites examined during survey and testing work (Appendix A, Table A-1). We will attempt to 

characterize this material generally, noting that more than half of the sample was recovered 

in two late Mississippi period sites (3AS286 and 3AS306), and that relatively little material 

attributable to Archaic components was collected. 

Cores are chiefly pebbles or small cobbles reduced bifacially or in irregular fashion, 

but some tabular specimens also occur (Figure 60m-o). In the best available series, 12 pebble 

cores from One Cypress Point (3AS286), maximum dimensions ranged between 28 and 66 

mm, the smaller specimens seemingly split and rejected (Figure 48j). At least half of these 

pebble cores were thermally altered prior to the flake removal. Eleven cores are brown chert 

and one is novaculite. Direct freehand percussion and, perhaps more rarely, bipolar percussion 

were techniques used to reduce these cores. Bifacially flaked pebble cores and early 

bifacial preforms intergrade in this assemblage (Figure 48). Among the many other pebble 

cores from scattered sites, we could find no distinctive differences from this group, although 

examples from Archaic or early ceramic sites were generally lacking. 

Debitage and cores were classified as heat altered when some combination of the following 

attributes occurred: (1) discoloration and differential discoloration, generally toward 

pink or gray, (2) waxy or vitreous luster where flakes are detached after heating, (3) crazing, 

crenated fracture, or unopened fractures, and (4) pot lids or spalls. Heat alteration may 

imply heat treatment to improve knapping quality; however, any collection may contain 

inadvertently heat altered specimens or heat treated specimens not distinguished by the 

obtrusive attributes cited above.


Waste flakes from all sites were sorted by raw material (see chert and novaculite 

above) and by reduction categories, defined as follows: 

Primary flakes have 50% or more of outer (dorsal) surface covered by cortex, and 

generally have hard hammer flaking characteristics (prominent bulb, ripples, etc.) 

on inner (ventral) surface; 

Secondary flakes have less than 50% cortex on dorsal surface; 

Interior flakes lack cortex; and 

Flakes of bifacial retouch are thin expanding flakes without cortex, generally having 

soft hammer characteristics (diffuse bulb, lipped lenticular platform, etc.) on 

the ventral surface. 

These categories reflect in a general way the reduction sequence from pebble cores 

through several stages of bifacial preforms to completed bifacial tools, represented schematically 

in Figure 62 (see also Shafer 1973:43; House 1975:Figure 1). In Chapter 6 (Figure 49b) 

we compared horizontally separated subsamples from One Cypress Point (3AS286) in terms 

of these general reduction categories for cores, waste flakes, and shatter. 

Almost all site survey collections contain less than 50 items of debitage, and most 

contain only a few waste flakes (Appendix A). Qualitative observations from these small collections 

suggested that brown pebble chert reduction was highly associated with Mississippi 

period sites and gray or white novaculite reduction with Tchula period and all preceramic 

sites. Table 22 examines these observations for three site collections of adequate size (from 

combined surface collections and test excavation) and of reasonably well known age; see 

Chapter 6). For these samples there are in fact striking differences in chert/novaculite ratios 

and also in proportions of flakes in various reduction categories. The earlier sample in this 

comparison, Marie Saline, Area C (3AS329), stands out by virtue of the high frequency of 

novaculite and low frequency of primary and secondary flakes (which bear cortex). Evidently 

novaculite bifaces or bifacial preforms were being reduced here, rather than pebble cores, 

during the early occupation at Marie Saline. We suggest that these raw material preferences 

and reduction strategies have regional significance, and should be tested more thoroughly 

in future work, with larger site assemblages.


Figure 62. Schematic linear model for production of a bifacial tool and discard materials


Table 22. Comparison of Debitage by Reduction Category and Raw Material Frequency 

among Three Floodplain Site Assemblages. 

Reduction Categories 

Bifacial Chert/ 

Sample Primary Secondary Interior Retouch Novaculite 

Identification of Component Size Cores Flakes Flakes Flakes Flakes Shatter Ratio 

One Cypress Point (3AS286) 1084 12 222 343 173 318 16 9:1 

(ca. A.D. 1600-1700) 

Jug Point 1 (3AS306) 677 4 100 185 193 180 15 3:1 

(ca. A.D. 1400-1600) 

Marie Saline, Area C (3AS329) 159 5 2 5 24 111 12 1:10 

(ca. 1200 B.C.-A.D. 1) 

The following list of conchoidally fracturing raw materials describes generally all 

types which occur among debitage and have been distinguished in Felsenthal Project site 

collections. 

Major or abundant types: chert (pale brown, brown, yellowish brown, reddened); 

novaculite (white, gray, yellowish brown, pink, reddened). 

Minor or poorly represented types: chert (oolitic); novaculite (translucent and 

dark veined); clear crystal quartz; quartz (milky, rose); quartzite (off-white, gray); 

silicified wood (varicolored and grained). 

Some rarely occurring, but distinctive, raw materials are associated with particular site assemblages, 

and may eventually prove to have regional cultural or temporal significance. 

Clear crystal quartz, for example, may have been acquired from the central Arkansas- 

Ouachita Mountain area through trade or long range procurement. 

Cobble and Heavy Duty Tools 

This class of tools includes rounded to tabular cobbles modified by specific modes of 

use, resulting in (1) battered or pitted edges and surfaces produced by percussion, (2) abraded, 

faceted, or grooved edges and surfaces produced by grinding, and (3) combinations of 

these. The use-wear patterns are distinctive for certain tool forms, such as hammerstones 

and pitted cobbles or anvils, which are easily separated from naturally modified cobbles in 

most cases. However, where use-wear is dubious, we have assigned the specimen to a


manuport category (see Inorganic Debris below). Cobble tools were evidently employed in a 

variety of “tough” extractive and maintenance tasks. Some pitted cobbles, for example, may 

be associated with nut processing and others with bipolar core reduction (House 1975:71). 

In spite of the accessibility of cobbles and the apparent utility of cobbles tools, we 

found such tools to be exceedingly rare in the project area. In fact, there are only 18 examples 

of hammerstones and other modified cobbles in all site survey and test excavation 

collections. For the floodplain sites in our sample, mostly late prehistoric extractive camps, 

expedient cobbles tools were either unimportant, were conserved for further use at a base 

settlement, or were not sufficiently modified by use to permit differentiation from manuports. 

We have the impression that relatively more cobble and heavy duty tools occur in Archaic 

sites, but no presently available data can be relied upon to substantiate this observation. The 

following examples illustrate the range of cobble tools: 

1. battered novaculite hammerstone with abraded facets on opposing ends, Mud 

Lake Bend 2 (3UN166), probably Archaic (Figure 63a). 

2. edge-abraded and facially abraded sandstone cobble mano(?), unnamed site 

3UN169 (Figure 63b). 

3. facially abraded and pitted sandstone cobble anvil, False Indigo (3AS285), associated 

with Gran Marais component (Figure 63c). 

4. two edge-battered and facially pitted sandstone cobble anvils from a buried 

feature at Swamp Privet (3AS327), probably Baytown-Coles Creek period (Figure 

63d). 

Ground Stone Tools 

This group of seven artifacts is very small and diverse. Some warrant little more than 

brief mention, but several are highly informative in a particular site context or are extremely 

unusual finds in the region. 

1. Two steatite vessel sherds of small size (Figure 64a) were recovered at Marie 

Saline (3AS329), Test Pit 8, 120 cm below the surface. These sherds are 17 

mm thick, rough in appearance with tool marks plainly visible, on soft, gray, 

coarsely schistose steatite. Although no other diagnostic artifacts were directly 

associated, stratigraphic context indicates that this deep level pertains to a 

Poverty Point period occupation, ca 1200-500 B.C. Steatite vessels are well represented 

at the Poverty Point site in Louisiana, and the source of steatite has


A 

Figure 63. Cobble tools. a. hammerstone, Mud Lake Bend 2 (3UN166); b. mano?, 

unnamed site (3UN169); c. pitted cobble or anvil, False Indigo 

(3AS285); d. pitted cobble anvils, Swamp Privet (3AS327). (AAS neg. 

813872) 

b 

d e 

c

A 

c 

b 


f 

e 

d 

Figure 64. Ground stone tools and pigment. a. steatite sherds, Marie Saline (3AS329); b. schist slab, Marie Saline 

(3AS329); c. hematite plummet, One Cypress Point (3AS286); d. shale gorget fragment, Persimmon 1 

(3AS313); e. amber(?) disk fragment, Jug Point 1 (3AS306); f. red ochre or earthy hematite, Marie Saline 

(3AS329). Length of c is 62 mm (note millimeter scale for item e) (AAS neg. 813873, 806895)


been attributed to the Southern Appalachian region (Ford and Webb 1956). 

The infrequent occurrence of steatite vessels and fragments in southern Arkansas 

has been summarized by Schambach (1974). It is clear that these vessels 

did move long distances through a Poverty Point exchange network. 

2. A tabular section of muscovite schist or talc schist (Figure 64b) was also recovered 

at Marie Saline (3AS329), Test Pit 3, 110-120 cm below the surface. 

This specimen is about 18 mm thick and is probably modified in shape, but 

lacks distinguishing tool marks. It is conceivably a vessel fragment. Again, we 

believe that Poverty Point occupants introduced this object, the raw material 

probably being originally derived from the Southern Appalachian region. 

3. A plummet of massive hematite (Figure 64c) was a surface find at One Cypress 

Point (3AS286). Although associated with terminal Mississippi period occupation 

debris, it is stylistically and technologically a much older artifact, strongly 

resembling magnetite and hematite plummets from Poverty Point (Ford and 

Webb 1956:Figure 33). The form is an elongated ovoid, broken through the 

perforation at one end and subsequently smoothed over. The surface is weathered 

and etched. An important source area for this mineral is at Magnet Cove 

on the headwaters of the Ouachita River in central Arkansas. Ford and Webb 

attribute plummets to use of the bolas in hunting fowl and small game. The 

late Mississippi period occupants of One Cypress Point need not have reused 

the plummet in this manner, but perhaps were merely “curating” the mass of 

hematite for other eventual use. 

4. A small fragment of shale gorget (Figure 64d) was retrieved from the surface 

at Persimmon 1 (3AS313), but may derive from a buried Poverty Point period 

occupation. The fragment retains a rounded edge, ground and polished face, 

and two remnants of drill holes. Raw material is a soft, fissile, gray shale of 

unknown origin. 

5. A tiny chip or flake from an unidentified ground stone tool was recovered on 

the surface at Mud Lake Bend 1 (3UN167), a probable Archaic component. The 

material is light colored siltstone. 

6. A perforated disc fragment, possibly a bead of fossil amber (Figure 64e), was 

found on the surface at Jug Point 1, Area A (3AS306), and is attributed to 

Caney Bayou phase occupation (ca A.D. 1400-1600) at that site. This artifact is 

clearly a unique find and warrants detailed treatment, but only some preliminary 

observations can be presented here in regard to outward appearance and


indications of raw material and source. Final results of a complex analysis of 

the amber (?) are not yet available, but some preliminary findings and methods 

used will be noted. 

7. The remaining fragment represents about one third of a disc 24 mm in diameter 

and 4.5 mm in maximum thickness, centrally perforated as shown in 

Figure 64e. The disc was not flat, but gently concave; that is, the obverse face 

shown is concave and the reverse face is parallel and convex. The obverse face 

bears the only other modification evident, and this is a broad shallow groove 

(0.9 mm deep) closely encircling the central perforation. The effect of this 

groove is to leave a raised rim about the interior and exterior margins. The 

form is thus reminiscent of European style buttons, but matches none that we 

have seen in all details. Nor does it reproduce in detail the form of any shell 

disc bead we have seen. 

The raw material is soft, markedly lightweight, amber colored, and translucent, 

although a weathered crust makes it appear opaque in reflected light. 

No manufacturing marks are visible, perhaps because of the crust. The broken 

edges do not reveal much of the nature of fracture in fresher material. We easily 

melted a corner with a quick pass of a match’s flame. Most of these qualitative 

observations led us to make the comparison with amber or fossil resin 

which can be characterized as follows (Langenheim 1964, 1969): 

(a) Honey colored to molasses colored commonly, but nearly colorless, ruby 

red, and even “blue” varieties are well known. 

(b) Specific gravity commonly 1.05-1.08 (not quite floating, but easily transported 

by water). 

(c) Hardness 1 to 3 Mohs scale (scratched by fingernail to calcite). 

(d) Melting point 150° to 420° C (a variable quality like color). 

(e) Consisting of resin acids, resin alcohols, and resene (insoluble in water 

and decay-resistant in alluvial environments, although progressive oxidation 

and polymerization occur). 

Langenheim and Beck (1968) have published a catalogue of infrared spectra of fossil 

ambers, which chemically “fingerprints” fossil and modern resins and assists in identifying 

botanical sources. An important occurrence of amber in Eocene Gulf Coastal Plain lignitic 

sediments at Malvern, Arkansas, has been described by Saunders et al. (1974). 

Our approach to analyzing the disc fragment has involved infrared spectrophotometric 

study (at the Department of Chemistry, Vassar College, by Prof. Curt W. Beck, and also at 

the University of Arkansas). At the present time we have identified the material as a resin, 

but with an admixture of silica not characteristic of known amber sources. Certainly, our


specimen does not match the published Malvern, Arkansas, amber spectra. If we are able 

to confirm the artifact from Jug Point 1 as fossil amber, it will represent the first record of 

aboriginal use of this material in southeastern United States, and one of relatively few such 

records for the continent. 

Mineral Pigments 

Scattered finds of earthy hematite and limonite in the project area indicate that at 

least these two potential sources of mineral pigment were locally available. Yellow-brown 

limonite, in particular, occurs as concretions, coatings, and minor lenses in floodplain sediments. 

Dark red ocherous masses of hematite were found more rarely, and may have been 

introduced by aboriginal occupants of floodplain sites, although none was distinguished by 

facets or striations from pigment preparation. The Magnet Cove area was cited above as a 

potential source of hematite. Two red ocher masses, illustrated in Figure 64f, were excavated 

in a deep level at Marie Saline (3AS329) and were presumably associated with the Poverty 

Point period occupation at that site. 

Inorganic Debris 

Inorganic waste materials produced and discarded on an aboriginal site may furnish 

important functional data since they (1) are resistant to decay, (2) were rarely removed or recycled 

compared to many classes of tools, and (3) can sometimes be associated with specific 

kinds of activity. In analyses of site assemblages presented in Chapter 6 we included considerations 

of debris, as well as ceramics and stone tools. Three categories of debris which recur 

repeatedly in floodplain sites, but in significantly varying frequency, are discussed here. 

Fire-cracked rock includes granular and sometimes conchoidally fracturing 

cobbles and cobble fragments in an archeological assemblage which are characterized 

by (1) irregular or created fracture surfaces and crazing or incomplete 

fractures, (2) pink, red, or gray discoloration and differential discoloration within 

a cobble, and (3) absence of breakage features which can be attributed to percussion. 

House and Smith (1975) report on the experimental replication of firecracked 

rock, and problem of distinguishing it from other debris. Fire-cracked 

rock from floodplain sites in the Felsenthal Project area includes sandstone, novaculite, 

chert, quartzite, and quartz. We noted a high frequency of fire-cracked 

rock in preceramic sites and low frequency or absence in ceramic sites. This 

observation is emphasized in the analysis and comparison of excavated Poverty 

Point/Tchula period levels at Marie Saline (3AS329, Area C). Abundant firecracked 

rock in preceramic floodplain sites is believed to reflect the employment 

of stone boiling and earth oven cookery as principal food preparation techniques.


Fired clay particles are hardened clay masses of irregular form and variable, but 

generally small size, which are light colored (oxidized), dark colored (reduced), 

or clouded due to contact with wood-fueled fires. Specimens from floodplain 

sites are few and scattered in occurrence, and very few have the distinctive 

vegetal impressions of daub or other prepared surfaces. No modeled “Poverty 

Point objects” are known from our sites, but good examples are reported from 

other sites in the region (Rolingson and Schambach 1981:179). Most clay particles 

examined by us can probably be attributed to inadvertent firing of floodplain 

silty clays in and around hearths, rather than to burned structures or prepared 

clay features. Only close association with artifacts and other cultural debris separates 

the majority of these items from naturally occurring fired clay fragments. 

We noted an unusually high frequency of fired clay particles in two Mississippi 

period extractive camps (3AS285 and 3BR76), which are believed to be bulk fish 

processing stations (Chapter 6, Table 17). Among the mostly irregular masses or 

lumps from these sites are a series of small ovoid pellets and several daublike 

specimens which warrant further study. Substantial amounts of daub or daublike 

material were recovered on the surface at Lapile Creek 4 (3UN158), a multicomponent 

ceramic site. 

Manuports are pebbles, cobbles, blocks, or slabs of any rock material which are 

evidently unmodified or exhibit simple breakage of unknown origin, but which 

occur in a context which implies human introduction and manipulation. In 

our site survey work many such objects were recorded and collected from fine 

grained floodplain deposits where other artifacts or suspected artifact were present. 

At Marie Saline (3AS329), 273 pebble manuports were positively correlated 

with other artifact classes in excavated levels within Area C (Chapter 6). Pebble 

raw materials are those known to be present in the bedload gravels of the Ouachita 

River, although some chert pebbles may have been introduced from Pleistocene 

Terrace gravels. Manuports presently appear to be the least useful class of 

debris for functional analysis, although they can contribute to site identification 

(Chapter 5). 

cerAMic AnAlysis And results 

Felsenthal Project ceramic collections include about 3,700 sherds from 60 sites. A single 

floodplain site, False Indigo (3AS285) produced 45% of the total collection, and another, 

Buttonbush (3BR58), produced 454 sherds attributed to two Baytown Plain vessels. Most site 

collections are markedly small, containing less than 15 sherds, and many sherds are highly 

weathered, even when recovered from buried context by test excavation. Our analysis took 

the form of a detailed inventory and comparison in the case of small sherd samples, but in 

the case of False Indigo we have attempted a more complex stylistic analysis (see Gran


Marais complex below). Ceramic data were recorded on a standard form entitled “Ceramic 

Analysis (Sherds),” except that additional records were generated for the stylistic analysis 

noted above. These records are retained at the Arkansas Archeological Survey. 

We were fortunate to have Schambach’s unpublished ceramic analysis and detailed 

notes for the Shallow Lake site (3UN9/52) as a current source for regional ceramic data and 

classification (report now in print; see Rolingson and Schambach 1981). Dr. Schambach 

served as a ceramic consultant and examined samples of sherds as the project proceeded. 

Regional site reports, published before 1979 or unpublished, are few and contain ceramic 

descriptions and data which are inconsistent and must be used with caution. Schambach’s 

study provides a new set of methods and concepts, and also indicates that systematic ceramic 

analysis in the Felsenthal region has much potential for future work. 

In the following sections we briefly characterize, compare, and illustrate four ceramic 

complexes best known from floodplain site collections. “Complexes” here recognized may 

correspond with phases, periods, or even greater temporal units (cf. Phillips 1970:30). There 

are no floodplain site data for diagnostic Marksville ceramic types, little data for definitive 

Coles Creek ceramics and few examples of obviously introduced Caddoan ceramic types. 

Presently, it seems clear that floodplain extractive sites lack the ceramic abundance and diversity 

of upland habitation sites in the region. However, many of our small sherd samples 

are useful “time capsules” and indicators of site activity, as well as sources of technological 

and stylistic ceramic data. 

Tchefuncte Complex 

The earliest known ceramics in the Felsenthal region are low fired, untempered or 

clay-tempered wares with distinctive “contorted laminated texture” (Phillips 1970:163), 

which may be indistinguishable from Lower Mississippi Valley Tchefuncte ceramics. Schambach 

(1981:181) notes the presence of Tchefuncte Plain, Tchefuncte Incised, Tchefuncte 

Stamped, and Lake Borgne Incised types at Shallow Lake (3UN9/52) and several other sites 

in the Felsenthal region. Our small sample of Tchefuncte ceramics is taken from five newly 

recorded floodplain sites, as follows (see also Figure 65): 


Lake Borgne Incised 20 sherds 

Tchefuncte Incised 18 sherds 

Tchefuncte Plain ? 9 sherds 

47 sherds 

River Birch 2 (3AS321) 

Tchefuncte Stamped 1 sherd 

Tchefuncte Incised ? 1 sherd 

2 sherds

A 

A 

b 

c 


f 

d e 

Figure 65. Tchefuncte complex ceramics. a. Lake Borgne Incised, False Indigo (3AS285); b. Tchefuncte Incised, False Indigo 

(3AS285); c. Tchefuncte Plain?, False Indigo (3AS285); d. Tchefuncte Stamped, River Birch 2 (3AS321); e. 

Tchefuncte Incised?, Marie Saline (3AS329); f. Tchefuncte Stamped, Hunter’s Swan (3BR70). (AAS neg. 813875)



Tchefuncte Incised 14 sherds 

Eagle Creek (3BR65) 

Tchefuncte Stamped 1 sherd 

Hunter’s Swan (3BR70) 

Tchefuncte Stamped 2 sherds 

Local varieties of these Tchefuncte types have not been named, and we cannot make 

such distinctions based on the few uneroded sherds available. We can suggest that an early 

Coon Island phase component (ca 500-250 B.C.) is relatively well represented at Marie 

Saline (3AS329) from 80-120 cm depth, and that a later Coon Island phase component (ca 

250 B.C.-A.D. 1) is represented at False Indigo (3AS285) from 60-0 cm depth. Tchefuncte 

Incised may predominate in the earlier component, and Lake Borgne Incised in the later 

one. There are no depth data for Tchefuncte Stamped sherds in our floodplain site sample. 

Phillips (1970:165) and other workers in Louisiana (Weinstein and Rivet 1978:50) recognize 

the temporal priority of rocker stamped over incised or drag-and-jab Tchefuncte decorative 

treatments. Two or more Tchula period phases with distinctive ceramic assemblages may 

eventually be identified in the Arkansas portion of the Felsenthal region. 

Baytown Complex 

A long-lived clay-tempered plainware tradition in the Felsenthal region has been identified 

most closely with the Baytown culture of the Lower Mississippi Alluvial Valley rather 

than with pre-Caddo or Fourche Maline cultures to the west. According to Schambach (in 

Rolingson and Schambach 1981:149), the Fourche Maline plainware type, Williams Plain, 

has a low ratio of rim sherds to basal sherds in certain site samples, while Baytown Plain 

has a higher rim/base ratio at Shallow Lake and other sites in the Felsenthal region. Presumably, 

this ratio reflects the prevalence of narrow-mouthed beaker forms in Williams Plain 

and open-mouthed jars and bowls in Baytown Plain. Both plainwares have square and disc 

bases, although the frequency of these may also vary among sites and regions. Disc bases 

are predominant in our collections. 

Baytown Plain utility vessels were obviously important items of equipment in floodplain 

extractive sites. Numerous rim sherds and body sherds are recorded for sites of several 

time periods. One site alluded to earlier, Buttonbush (3BR58) on the Saline River, consisted 

principally of the fragments of two wide-mouthed, Baytown Plain vessels with disc bases, 

broken and concentrated in a small sherd cluster (Figure 45). A second similar site on the 

Ouachita River, Water Elm 3 (3AS287), contained a cluster of 298 plain and four incised 

sherds, representing one or two Baytown vessels with disc bases. Marie Saline (3AS329), 

Mo-Pac (3UN121) and other riverbank sites have produced Baytown Plain sherds in place at


80-90 cm and shallower depths. A selection of Baytown Plain sherds from these and other 

floodplain sites is illustrated in Figure 66. 

Baytown plainware appears to dominate ceramic assemblages from about A.D. 1 to 

1400, to diminish in proportion to decorated wares after this date, and to recur in trace 

amounts with shell-tempered ceramics after A.D. 1500. We have been unable to identify 

Marksville period sites because no diagnostic decorated Marksville ceramic types have been 

recovered (Schambach reports a site collection with 1,532 plain and 18 decorated sherds 

upstream from the project area; cf. Rolingson and Schambach 1981:182). We also have been 

generally unable to differentiate ceramic collections from Baytown (A.D. 300-700) and Coles 

Creek periods (A.D. 700-1100) because of the uniformity of plainwares and rarity of decorated 

wares. It is probable that Baytown plainwares varied in temper admixtures, surface 

finishes, and vessel forms and sizes through the 1500-year span of this ceramic tradition, 

but no thoroughgoing analysis of plainwares has been attempted for the Felsenthal region. 

Depth data for floodplain sites with Baytown plainware (Appendix A, Table A-1) may eventually 

provide the key to distinguishing early and late phases within a Baytown-Coles Creek 

continuum. 

Floodplain sites recorded by us at about 30-50 cm depth may possibly represent Coles 

Creek occupations, ca A.D. 700-1100. Such sites which contained ceramics included only 

plainware and weathered, indeterminate sherds. We are therefore unable to describe a Coles 

Creek ceramic complex, the obvious link with earlier Baytown plainware and with later 

ceramic complexes having abundant, diverse and decorative treatment. The reader may 

find an introductory statement on the regional Coles Creek ceramic complex in Rolingson 

and Schambach (1981:182-189). Figure 67a illustrates a Coles Creek Incised rim sherd from 

Eutaw Rapids (3BR75) which was recovered on the Ouachita River shoreline (a buried shell 

lens nearly 25 cm in depth may have been the place of origin). This large bowl sherd has two 

lip lines and a haphazard line at the base of the rim fold, resembling Phillips’s (1970:72) Yazoo 

Basin, var. Ely. Much of the Coles Creek decorative treatment is strongly continued and 

elaborated in the Gran Marais ceramic complex, discussed below. 

Gran Marais Complex 

Late in the Coles Creek period or at the outset of the Gran Marais phase, Mississippi 

period (A.D. 1100-1400), marked changes occurred within the indigenous Baytown-Coles 

Creek ceramic tradition. Schambach characterizes these changes as follows (Rolingson and 

Schambach 1981: 107): 

they were foreign to the Lower Mississippi Valley and they resulted in a ceramic 

tradition that defies classification according to the Phillips system. The main innovations 

were to decorate vessel bodies as well as the rims or shoulders


A 

f 

c 

h i 

Figure 66. Baytown complex plainware. a. rim, Mo-Pac (3UN121); b. bowl rims, False 

Indigo (3AS285); c. carinated bowl rims, False Indigo (3AS285); d. rim castellation, 

False Indigo (3AS285); e. body sherd with shoulder, False Indigo 

(3AS285); f. disk base, False Indigo (3AS285); g. base and wall, Water Elm 

3 (3AS287); h. body sherd, Buttonbush (3BR58); i. disk base, Buttonbush 

(3BR58). (AAS neg. 813876) 

d 

b 

g 

e

c 

A b 


Figure 67. Miscellaneous sherds from floodplain sites. a. Coles Creek 

Incised rim, Eutaw Rapids (3BR75); b. shell-tempered incised 

body, Jug Point 1 (3AS306); c, d. Sinner Linear Punctated rim 

and neck, Water Elm 1 (3AS322). (AAS neg. 813877) 

d


(probably this was a Caddoan idea), and to treat rims and bodies as different 

decorative fields. . .this is more than an occasional occurrence; it is a general characteristic 

of the Felsenthal region pottery of the Mississippi period. 

Schambach convincingly shows why these Mississippi period ceramics, when they occur as 

sherds or as whole vessels, often cannot correctly be classified in the Lower Valley or Caddoan 

typologies (Suhm and Krieger 1954; Phillips 1970). He then sets out in detail the basis 

for a new “descriptive classification system” for rim-and-body decorated ware, especially 

relevant to clay-tempered ceramics of the Gran Marais phase. 

At this point we must refer the close reader to Schambach’s ceramic presentation in 

the Shallow Lake site report (Rolingson and Schambach 1981). However, three hierarchical 

concepts are necessary to classify sherds or vessels in the descriptive system, and they are 

briefly summarized here as a basis for discussion of the Gran Marais complex below: 

Class is a general category based on a major motif and decorative technique (e.g., 

Class B includes rectilinear or curvilinear, incised line designs placed horizontally, 

or composed of horizontal lines). Five classes (lettered A through E) are 

defined by Schambach presently. 

Patterns are groups of similar designs within classes, and are based on surface 

finish, line quality, and contrasts in decorative technique. Numerous patterns 

are distinguished by Schambach within Classes A to C, and are named “Afton, 

Anthony, Butler, Concord, Cornell,” etc. 

Designs are variations within patterns, usually minor variations or modes, such 

as lip notching, lip lines, etc., and are numbered after the pattern name, e. g., 

“Afton 1, Afton 2.” 

Also a notation system for sherds and vessels consists of rim pattern or design notation 

separated from body pattern or design notation by the symbol “::” (again, the reader should 

consult Rolingson and Schambach 1981:106-131). Depending on completeness and preservation, 

all sherds can be descriptively classified at some level of specificity, for example, “Afton 

:: plain, Butler 1 :: Coker 2, Butler 15 :: missing, missing :: Alma 3,” and so on. 

As the result of site survey in the Felsenthal Project area (Chapter 5) we recorded many 

small Gran Marais phase components, and have classified many small sherd samples from 

these sites. One very important and relatively large collection was obtained from False Indigo 

(3AS285) on the Saline River. This site contains a 15 x 18 m artifact scatter (Area A) and 

small midden deposited during a brief interval within the Gran Marais phase. The ceramic 

sample (1,130 identifiable sherds) from False Indigo, Area A, thus represents a “pure” as-


semblage which can be used to characterize, analyze, and compare Gran Marais complex 

pottery used in a floodplain extractive site. 

Schambach’s analysis of Shallow Lake ceramics includes a sample from a structure 

(Unit X) underlying Mound C, which has furnished an important basis for describing Gran 

Marais complex pottery used in a ceremonial site (Rolingson and Schambach 1981:Table 20). 

This structure is well dated at about A.D. 1250-1300. In the following analysis we implement 

Schambach’s descriptive classification system and compare Shallow Lake and False Indigo ceramics. 

We know that variability among the two samples could reflect sociological and functional 

differences between sites, and could also reflect temporal change in the Gran Marais 

complex. The two samples are sufficiently large and “unmixed,” so far as we can determine, 

to furnish a valid comparative basis. They may be generally characterized as follows: 

Plain Decorated 

Shallow Lake, Mound C, 1,148 (81.6%) 258 (18%) 

Structure Unit X 

(N=1,406) 

False Indigo, Area A 976 (86.4%) 154 (13%) 

(N=1,130) 

Tables 23 and 24 present the raw sherd counts in terms of the descriptive classification 

system. We note that there is general similarity in the proportion of plainware and, by inference, 

in vessel form among the two samples, except that polished plainware is not present 

at False Indigo. Among decorated wares there is also general similarity, except that proportions 

of class vary as follows: 

Shallow Lake, Mound C, False Indigo, 

Structure Unit X Area A 

Class A 26.8% 24.2% 

Class A or B 29.4% 21.4% 

Class B 11.6% 29.2% 

Class C 31.8% 21.4% 

All other 0.4% 3.8% 

100.0% 100.0% 

The contrast is greatest for classes B and C (Class B includes horizontally placed, incised 

elements and Class C includes punctated, pinched, and notched elements). One interpretation 

of this difference is that False Indigo is an earlier assemblage, retaining more of the 

Coles Creek decorative style, and that Shallow Lake, Mound C, is a later assemblage with 

increasing use of rim and body punctation. This interpretation assumes that functional


Table 23. Descriptive Classification of Plainware from Two Sites of the Gran 

Marais Phase, Mississippi Period. 

Shallow Lake (3UN9/52) False Indigo (AS285) 


Clay Tempered 

Plain Rims 82 76 

Plain Body Sherds 960 843 

Polished Plain Body Sherds 85 0 

Lugs and Suspension Holes 2 1 

Carinated Bases 4 0 

Convex Bases 1 9 

Flat Circular Bases 2 12 

Indeterminate Bases 9 0 

Shell Tempered 



Bone Tempered 



TOTALS 1,148 982


Table 24. Descriptive Classification of Decorated Sherds from Two Sites of the 

Gran Marais Phase, Mississippi Period. 

Shallow Lake (3UN9/52), False Indigo (3AS285), 


Afton 1 :: missing 1 missing :: Adelphi 5 

Afton 2 :: missing 1 Afton :: missing 4 

Afton 10 :: missing 1 missing :: Alma 4 1 

missing :: Alma 3 1 missing :: Alma 3 

Anthony 1 :: plain 2 Anthony 1 :: missing 2 

Anthony 1 :: missing 2 Anthony :: class A 1 

Anthony 3 :: missing 2 Anthony :: missing 1 

Anthony 4 :: missing 1 missing :: Ashland 1 

Anthony :: missing 11 

Class A rims/bodies 47 Class A rims/bodies 19 

Class A or B rims/bodies 75 Class A or B bodies 33 

Boise 10 :: missing 1 Barnard 3 :: missing 1 

Boise :: missing 2 Barnard 4 :: missing 2 

Butler 1 or 2 :: plain 1 Barnard :: plain 3 

Butler 1 or 2 :: Coker 1 Barnard :: missing 3 

Butler 2 :: missing 1 Barrington :: missing 2 

Butler 3 or 5 :: plain 1 Bates :: missing 1 

Butler 3 or 5 :: Coker 1 missing :: Buffalo 1 

Butler 3 or 5 :: missing 3 Butler 1 :: Coker 2 2 

Butler 3 or 17 :: missing 1 Butler 1 :: missing 7 

Butler 15 :: missing 1 Butler 4 :: missing 4 

Butler 1, 8, or 9 :: missing 10 Butler 17 :: missing 2 

Butler :: missing 6 Butler :: plain 1 

Butler :: missing 2 

Class B rims/bodies 1 Class B rims/bodies 14 

Class B or C rims/bodies 2 

Canisius :: or :: Creston 1 Cameron 2 :: missing 2 

missing :: Casper 2 missing :: Caney 10 

plain :: Castleton 2 1 missing :: Carver 1 

:: Catawba or Caldwell :: 3 Cedar 2 :: missing 1 

Cisco 5 :: missing 1 Cedar :: missing 1 

Cisco :: or :: Claflin 1 missing :: Coker 1 

(continued)


Table 24. (concluded) Descriptive Classification of Decorated Sherds from Two 

Sites of the Gran Marais Phase, Mississippi Period. 

Shallow Lake (3UN9/52), False Indigo (3AS285), 


missing :: Claflin 2 Concord 6 :: missing 1 

missing :: Coker 1 3 Cornell 2 :: plain 1 

missing :: Coker 4 1 Cornell :: plain 1 

missing :: Coker 5 1 

missing :: Coker 6 

Concord :: or :: Coker 30 

Concord :: missing 1 

cornell 1 :: plain 8 

Cornell 2 :: plain 9 

cornell 3 :: plain 1 

Culver :: missing 1 

missing :: Curry 2 1 

missing :: Curry 3 

Curtis 1 :: missing 1 



Class C rims/bodies 1 Class C rims/bodies 14 

Class D rims/bodies 1 Class D rims/bodies 2 

Class E rims/bodies 1 Class E rims/bodies 2 

TOTALS 258 154


differences among contemporary assemblages is not the significant factor, but of course we 

do not know the truth of this statement. A reliable set of dates for False Indigo occupation 

could test this interpretation of ceramic differences. At this point we regard the variation as 

likely to reflect both functional and temporal differences between ceramic samples. 

Since some rim-and-body classifications designated by Schambach can be translated 

into types and varieties, we will also attempt to compare assemblages on this basis. The following 

list of occurrence is minimal, and more complete sherds or vessels would undoubtedly 

add a few other types and varieties (SL=present or probably present in Shallow Lake, 

Mound C, Structure Unit X, sample; FI=present or probably present in False Indigo, Area A, 

sample): 

Avoyelles Punctated, var. Wolf Slough SL 

Baytown Plain, var. unspecified SL, FI 

Baytown Plain, var. Shallow Lake SL, FI 

Coles Creek Incised, var. Kemp SL 

Evansville Punctated, var. Pine Island SL, FI 

Harrison Bayou Incised, var. Strong SL 

Kiam Incised, var. Redeye Lake SL, FI 

Kiam Incised, var. Sowell SL, FI 

Lapile Incised, var. unspecified SL 

Pargoud Incised, var. Monroe FI 

This list comprises a large part of the indigenous pottery types now known to characterize 

the Gran Marais complex in the Felsenthal region. Figures 68 and 69 illustrate decorated 

sherds of the Gran Marais complex from False Indigo, Area A. Baytown Plain sherds 

from this site are illustrated in Figure 66. 

Shell-tempered Complex 

Shell-tempered ceramics were undoubtedly present throughout the Mississippi period 

in the Felsenthal region, but the use of crushed shell prevailed, and eventually nearly 

excluded clay temper, during the Caney Bayou phase (A.D. 1400-1600). Table 25 records this 

transition from the standpoint of ceramic samples obtained in small floodplain extractive 

sites, which appear in presumed chronological order from early to late Mississippi period 

(see also Chapter 6). 

Although we have many sherd samples pertaining to the Caney Bayou phase, and 

one pertaining to terminal Mississippi period occupation at One Cypress Point (3AS286), 

these samples contain only small, weathered, and often leached and exfoliated, sherds. We 

therefore cannot characterize shell-tempered ceramics except to say that plain and incised 

vessels were present in floodplain extractive sites, especially those occupied after about A.D. 

1400. Schambach defines and illustrates the Caney Bayou ceramic complex at Shallow Lake 

(Rolingson and Schambach 1981:193-195). Indigenous plain and decorated ceramics, as well


Figure 68. Decorated rim sherds, Gran Marais ceramic complex, False Indigo (3AS285) (AAS neg. 813879).


Figure 69. Decorated body sherds, Gran Marais ceramic complex, False Indigo (3AS285) (AAS neg. 813881).


as various Caddoan types and Winterville Incised, are well represented in this complex. Figure 

67b illustrates one relatively well preserved shell-tempered body sherd from Jug Point 1 

(3AS306). This sherd appears to represent a vessel with diagonally opposed, incised lines on 

the body just below the rim panel. 

Table 25. Proportions of Shell and Clay Tempering in Ceramic Samples from Four 

Floodplain Sites of the Mississippi Period. 

Archeological Sample Shell Clay 

Site Name and Designation Phase Size Temper Temper 

Jug Point 1 (3AS306) Caney Bayou (late) 62 98.4% 1.6% 

Jug Point 2 (3AS307) Caney Bayou (early) 98 74.5% 25.5% 

Jug Point Cutoff (3BR76) Gran Marais (late) 250 22.8% 76.0% 

False Indigo, Area A (3AS285) Gran Marais (early) 1130 4.0% 95.9% 

Caddoan Imports 

Caddoan vessels (or sherds) which can most likely be attributed to manufacture 

outside the Felsenthal region or to manufacture by ethnic Caddoan potters are evidently 

present at Shallow Lake and other upland habitation sites. These ceramics include various 

polished, engraved, pseudolinear punctated, stamped, and brushed types. Sherds from such 

vessels are exceedingly rare in our floodplain site samples, and isolated small sherds are 

difficult to identify or distinguish from local wares. Figure 67c, d illustrates rim and body 

sherds from a single vessel of Sinner Linear Punctated, var. Sinner, recovered at Water Elm 

1 (3AS322) on the Ouachita River. The sherds were eroding out of the riverbank at about 40 

cm depth and were not associated with other artifacts or debris. It is possible that Water Elm 

1 represents a Caddo Indian transient camp or bivouac on the riverbank, but Schambach’s 

comment on the frequency of var. Sinner indicates that it may also be locally manufactured 

(Rolingson and Schambach 1981:170). 

culture historicAl hypotheses 

Analysis of artifacts from site survey and test excavation has provided data relevant to 

two hypotheses stated in the research design. In general we have not obtained the quantity 

or quality of data which could produce a rigorous test of these hypotheses, and we shall 

advance only a preliminary statement and discussion of possibilities. The culture historical 

hypotheses were framed as follows:

Problem Domain 3: Culture History 


H 3 If upland mound complexes represent a dominant regional culture, numerous 

sites of this culture should also be identified in the lowland setting. 

Test Implications: Discovery of culturally related sites in study area. 

Data Requirements: Common artifact styles, especially ceramics. 

H 4 If aboriginal Felsenthal populations occupied a frontier region, local styles of 

artifacts and settlement remains will occur mixed with extraregional styles. 

Test Implications: Recognition of local and alien styles. 

Data Requirements: Typological and comparative analyses disclosing introduced 

styles, especially in ceramics. 

Hypothesis 3 seeks to provide the physical link between upland population centers 

and transient floodplain camps through shared artifact styles. We have in fact identified 

numerous “culturally related sites” on the basis of ceramics primarily, but also on the basis 

of projectile points and lithic raw materials. Our comparisons are almost always constrained 

by the small, and sometimes weathered, artifact samples furnished by floodplain sites and 

by the dearth of controlled and detailed artifact data for upland sites in the region. In the 

preceding section of this chapter a detailed ceramic comparison was made for Gran Marais 

phase assemblages at Shallow Lake (3UN9/52) and False Indigo (3AS285). These assemblages 

are markedly similar and must derive from a single decorative style, and probably 

also from a single technological tradition. Stylistic similarities are present among simple elements 

or techniques and also among complex designs and patterns. Minor variation is likely 

to reflect functional or temporal differences among these dissimilar site units, and behavioral 

differences among social units. 

Although we cannot prove Hypothesis 3 in any formal sense, it would seem to be well 

supported by available data. We could easily envision the work party, which deposited its 

ceramic waste in the False Indigo midden, actually being deployed from the Shallow Lake 

center or its satellite communities. Ceramic analysis has not achieved the level of analytical 

sophistication which could justify such a statement, but perhaps that level is attainable 

through stylistic and technological approaches now being developed.


Hypothesis 4 also relies on stylistic analysis of ceramics and other remains to determine 

whether settlements were characteristic of a frontier region. Floodplain sites studied 

by us do not appear to furnish significant and relevant test data. If the broken linear punctated 

vessel from Water Elm 1 (3AS322) is in fact an “alien” Caddo artifact, no indigenous 

materials were associated or “mixed” with it. Other examples of imports are too few and 

too uncertain. We believe that upland sites of several types may furnish evidence of frontier 

character when more detailed analytical work is accomplished in this region.

Chapter 9 

huMAn AdAptAtion in the grAnd MArAis lowlAnd 

This concluding chapter does not attempt to summarize all important archeological 

and historical results of the Felsenthal Project; it does expand and discuss two topics of continuing 

research interest. From development of research design through all successive stages 

of work, we emphasized the potential and distribution of floodplain resource and the physical 

and spatial record of prehistoric human activity in this overflow bottomland (extractive 

sites in their environmental context). 

Our concluding comments deal first with the archeological and systemic qualities of 

extractive sites and then with a hypothetical seasonal model for resource use in an overflow 

bottomland environment. These observations stem from regional environmental data (the 

Grand Marais Lowland and its floodplain) and from our most complete site data (specialized 

extractive sites of the Mississippi period, A.D. 1100-1700), but there are broad implications 

for subsistence-settlement studies in all forested river valleys of the eastern United 

States and in all periods. 

on the nAture of extrActiVe sites 

While a concern with large multiple activity or complex sites has prevailed in American 

archeology, many students of settlement systems distinguish a class, occasionally a functionally 

differentiated series, of small limited activity sites. Small site studies frequently, but 

not always, pertain to hunter-gatherer cultural systems (Hayden 1965; Winters 1969; Moseley 

and Mackey 1972; Dillehay 1973; Talmage and Chester 1977; Ward 1978; Binford 1980). 

Table 26 compares the terminology and site characteristics developed in three relevant studies; 

hierarchical classifications of Mississippian settlement types were omitted here because 

they generally emphasize complex centers and large habitation sites, grouping all small site 

units as “farmsteads” or “camps.” We noted examples of this emphasis on small Mississippi 

period sites in Chapter 6 (e.g., Muller 1978:283). Klinger (1975:51) at least recognizes the potential 

of such sites for the Parkin phase (ca A.D. 1400-1600) in northeast Arkansas: “the fifth 

order of settlement is composed of small hamlets and farmsteads.... Other activity loci such 

as quarry sites, butchering stations or overnight campsites would be included [but none] 

have been recognized in the St. Francis Basin.” It is of interest to note that a recent intensive 

survey of a 1 km wide area around the Parkin site on the St. Francis River did not confirm 

the existence of fifth order sites as proposed (see Morse 1981:41).


Table 26. Three Classifications of Settlement Type which Emphasize or Distinguish Small 

Extractive Sites. 

Winters (1969): Wabash Valley Fitzhugh (1972): Coastal Labrador Binford (1980): Collectors/Foragers 

1. Settlement–2-3 acres, per- 1. Gathering Site–large size; sea- 1. Residential Base–one of multiple camps 

manent houses, storage pits, sonally reoccupied by band. occupied annually by mobile foraging or 

burials, many fabricating, collecting group, “archeological visibility” 

processing, and domestic tools 2. Base Camp–smaller size; occupied varies with group size, duration of stay 

by several families for extended and reoccupation of site; processing, 

2. Transient Camp–2-3 acres; period in a resource area. manufacturing, and maintenance activifew 

or no houses, storage pits ties; daily foraging originates here. 

or burials; many weapons and 3. Exploitation Camp, intensive–a 

processing tools. single dwelling occupied by family 2. location–a place where extractive tasks 

or extended family during several are exclusively carried out by foragers or 

3. Base Camp–2 acres; no weeks of hunting or fishing. collectors; “archeological visibility” is low 

houses, few storage pits because of low bulk procurement, brief 

or burials; many weapons 4. Exploitation Camp, light–single site use, infrequent reoccupation; expediand 

fabricating tools. dwelling or hearth attributed to ent extractive tools obtained and abana 

family briefly engaged in food doned. 

4. Hunting Camp–small area; gathering or hunting. 

only weapons and processing 3. Field Camp–temporary camp occupied 

tools. 5. Bivouac–transient camp with mini- by collectors’ task group to procure and 

mal activities; few structures or process particular resources; “archeo- 

5. Gathering Camp–small area; tools. logical visibility” may be high because of 

processing tools. bulk procurement and reoccupation of 

6. Specialized Camp, internal– preferred location. 

6. Bivouac–about .01 acres or limited activity site within band 

less; debitage territory. 4. station–information gathering points, 

such as hunting stand, used by logistically 

7. Specialized Camp, external– organized collectors’ task group. 

trading or procurement sites 

outside band territory. 5. cache–temporary field storage facility 

used by collectors’ task group.

Human Adaptation 271 

We think Binford’s (1980:10) discussion of “logistically organized” resource procurement 

strategies has relevance to Mississippi period subsistence and settlement in our study 

area (even though he generalizes from the Nuniamut Eskimo). 

Logistical strategies...are accomodations to the situation where consumers are 

near to one critical resource but far from another equally critical resource.... Special 

task groups may leave a residential location and establish a field camp or station 

from which food procurement operations may be planned and executed.... 

The obtained food may be field processed to facilitate transport and then moved 

to the consumers....Logistically organized task groups are generally small and 

composed of skilled and knowledgeable individuals. 

This is a fair description, we think, of the Mississippi period subsistence-settlement system 

described from site survey data in Chapter 5. In our system a peripheral ring of mound-village 

complexes adjoins one critical resource—farmland on upland terraces—and is distant 

from another—fish and other aquatic/riparian species in the interior of the floodplain. We 

called this a “centripetal pattern,” and we think “logistically organized” task groups operated 

the Mississippi period extractive sites (“field camps” and “locations” in Table 28) concentrated 

on the lower Saline River. 

If we distill from our experience with floodplain sites and from various studies cited 

earlier, we can suggest a series of qualities which should distinguish extractive sites from 

other settlement types, and further, should distinguish variations among extractive sites 

which may occur in the archeological record of a region. Our presentation will provide 

information about the archeological context, i.e.,, “artifacts, featues, and residues...no longer 

participating in a behavioral system,” and the systemic context, the past inferred behavioral 

system which contributed to this observed array of materials (Schiffer 1976:28). In other 

words, we will attempt to establish some correlates of human behavior on extractive sites 

(but see Schiffer [1976] and Wood and Johnson [1978] on site disturbance processes). 

1. Site size. Although we frequently refer to “small” extractive sites, one might expect 

considerable variation in this quality, depending on task group size, nature of extractive, 

processing, and maintenance activities, nature of discard behavior, duration of occupation, 

“redundancy” of site use, and other factors. In our sample of floodplain extractive sites, 

which admittedly are not well known in regard to extent because of site burial, area of artifact 

and debris scatter ranged from 1-900 m 2 and maximum linear dimensions rarely exceeded 

60 m. (About half of our floodplain sites had maximum linear dimensions less than 20 

m.) These sites are a fraction of a hectare and should be distinguishable from “hamlets” and 

“farmsteads” on size alone. 

Without a doubt much more extensive extractive sites occur in the archeological record, 

e.g., Ouachita Mountain novaculite quarries studied by Baker (1974, 1975) and a mass 

bison kill studied by Wheat (1972). The first case presumably results chiefly from redundant


site use, but the latter represents a single event where extractive activity featured high bulk 

procurement. 

Correlates: Extractive site size relates to multiple variables; in the case of discrete activity 

loci, size, number of personnel, and intensity of task-related behavior should be positively 

correlated. 

2. Site Configuration. In very general terms, an extractive site is observed archeologically 

as a continuous distribution of artifacts and remains having some outer limit which 

defines its shape, or as an apparent cluster of artifact scatters (loci) arranged in some configuration. 

Loci themselves may vary in shape. Again, considerable variation in configuration 

is to be expected, depending on task group size and composition, nature of target 

resources, differentiation or repetition of tasks, nature of discard behavior, redundancy of 

site use, terrain characteristics, and other factors. In our floodplain site sample, many extractive 

sites have pronounced linear configuration, whether observed as continuous scatters or 

as successive discrete loci. These configurations are congruent with narrow levees and river 

banklines, and imply that performance of tasks and discard of wastes were strongly determined 

by terrain characteristics. Discrete linear scatters observed by us may also frequently 

result from single brief episodes of occupation, since artifact and debris density was light. In 

contrast, we noted compact oval or more irregular riverbank artifact scatters with centrally 

located midden lenses at 3AS285 and 3BR76, where bulk fish processing tasks and more 

prolonged site occupation are inferred. At this point, we might suggest that defensibility 

could also affect extractive site configuration (if we are correct about sites 3AS285 and 3BR76 

as bulk facilities, they might require protection). 

Other variations in extractive site configuration are likely to be many (e.g., fishweirs, 

quarries, salt-making stations, and butchering sites) and we will not attempt to expand our 

discussion of this quality. 

Correlates: Extractive site configuration relates to many variables; specific configurations 

of discrete loci should correlate with specific tasks, e.g., fan-shaped and other arrays of 

debris left by a flintworker (Callahan 1980; Newcomer and Sieveking 1980). 

3. Site Assemblage. Assemblage here refers to tools and debris from toolmaking, 

refurbishing, or recycling, and also facilities for extractive or processing tasks, which persist 

in archeological context. Extractive site assemblages can vary greatly in quantity of such 

materials and in proportions of various functional categories of material. This variation is 

affected by size and composition of the task group, nature of target resources, intensity of 

extractive, processing, or maintenance activity, duration of occupation, redundancy of site 

use, discard and abandonment behavior, seasonality, and other factors. In our floodplain site 

sample, substantial permanent structures for habitation, storage, and burial appeared to be 

absent, but other task-related or maintenance facilities were occasionally identified (hearths,


pits, floors, post structures?). These characteristics may be associated with a “fine grained 

assemblage,” or the remains of specific events that “accumulated over a short period of time, 

for example, a two-day camp” (Binford 1980:17). Also, the numbers of tools and task-related 

items in our sample ranged from one to about 5,000 items. In Chapter 6 (Site Functional Hypotheses) 

we noted the small size of assemblages recovered from sites tested, and also lack 

of diversity in these assemblages (high ratio of extractive/maintenance tool kits). The floodplain 

sites we studied may be highly biased toward small task groups, specificity of target 

resources, and transient site use during seasonal low river levels. These sites have unusually 

low “archeological visibility.” 

Very much larger, more intensively used, and reoccupied extractive sites are well 

known in the archeological record of many regions, e.g., flint quarries and salt-making stations. 

Binford’s (1980:17) “coarse grained” distinction applies to these assemblages in the 

sense that “the resolution between archeological remains and specific events is poor.” 

Correlates: Multiple variables determine the size and content of an extractive site 

assemblage; in fine grained assemblages differentiation in task group personnel and tasks 

performed should be directly related to assemblage variability (e.g., male vs female tasks, 

hunting vs fishing activity, and so on). 

4. Site Residue. Under favorable conditions of preservation, the remains of extractive 

sites include organic or inorganic residues which are the byproducts of extractive or processing 

activity. These residues may be expected to vary with seasonality, target resources, 

processing techniques, and discard behavior, among other factors. Target resources may 

be concentrated or diffuse in occurrence. Also, quantities of residue vary with the nature of 

processing tasks, e.g., where high bulk procurement strategies are employed. 

In our floodplain site sample, residues were usually associated with burning. Many 

extractive, processing, or maintenance tasks required heat or flame, and a wide range of 

residues resulted (wood charcoal, ash, cinder, charred faunal and floral matter, heat-altered 

chert and novaculite, fire-cracked rock, and fired clay). Analysis of these residues may furnish 

critical data for interpreting site function, since they are much less likely to be “curated” 

than tools or tool fragments. However, where target resources were foodstuffs, analysis 

of faunal and floral residues provides empirical evidence for site function. 

Correlates: Extractive site residues may vary in identity and quantity, but an adequate 

site sample should relate directly to target resources processed on that site and to procurement 

in low bulk or high bulk amounts. 

5. Site Location. The final and often most difficult quality of extractive sites to interpret 

is location. Location includes the spatial relationship of an extractive site with its contemporary 

residential base and with its contemporary target resources or zone of resources.


Since an extractive site may be a point of assembly and processing of bulk resources for 

delivery to base settlements (Hypothesis 2 in Chapter 5), access routes can be an important 

locational factor. Other factors include the spacing of contemporary base settlements 

and their satellite communities and extractive camps within the region. In our study of 

floodplain sites, we had insufficient data to establish contemporaneity among upland base 

settlements, lowland extractive site , and access routes (presumably the latter refers to river 

channels, bayous, and backswamp tributaries which are not static features of the landscape). 

As indicated at the end of Chapter 5 (Figure 34) and earlier in this chapter, we have made a 

substantial advance in delineating a “centripetal pattern” of Mississippi period settlement in 

the Felsenthal Project area, but much work remains to be done here. 

One aspect of extractive site location has received little attention. Not only base settlements, 

but small seasonal camps themselves, have catchment areas (Vita-Finzi and Higgs 

1970). These areas encompass resources which can be efficiently exploited from a central 

site locus; archeological models of site catchments are usually presented as circles of a few 

kilometers radius or as concentric rings. Flannery (1976:116) notes with regard to the early 

Mesoamerican village and its subsistence system that “one must consider not only its own 

immediate catchment area, but those of its associated seasonal camps.” We suggest that extractive 

site catchment analysis is a fruitful area for future work in our study area. However, 

where linearity of floodplain and meander belt features prevails and where canoe travel 

may be the most efficient means of access to resources, circular catchment areas may be 

totally inappropriate for extractive site analyses. 

Correlates: Extractive site location is related to multiple environmental and cultural 

variables; among other factors, location is highly related to the distribution of target resources. 

Once again we would affirm the research potential and quality of significance that may 

distinguish small extractive sites. Our brief review of their attributes and potential as archeological 

resources leads to the following conclusions. 

1. These sites may represent brief time intervals or events and can contribute to studies 

emphasizing temporal ordering (seriation, cross-dating) 

2. Site assemblages may be culturally “pure” or unadulterated by redundant site use, 

and contribute to taxonomic studies (typology, stylistic analysis). 

3. Artifacts and other remains may pertain directly and exclusively to exploitative 

activity (environmental archeology, subsistence and settlement systems).


4. Task group personnel may have been culturally or ethnically homogeneous (social 

organization, ethnohistory). 

5. The array of tools, facilities, and debris may pertain to specific activities and to 

discard and abandonment processes (behavioral archeology, functional analysis). 

6. Small extractive sites are underrepresented in current archeological studies (research 

design, sampling strategies). 

7. Small sites are manageable units in terms of cost for comprehensive data recovery 

or for alternative mitigation measures (cultural resource management studies). 

A Model for resource use in An oVerflow bottoM 

In order to understand the abundance and nature of archeological resources in the 

Felsenthal Project area we have necessarily examined the salient environmental characteristics 

of an overflow bottomland (Chapter 2). In this floodplain environment, potential human 

food resources are predominantly aquatic and riparian species, complexly distributed or 

concentrated in space and responsive to seasonal changes, especially the annual hydroperiod. 

Prehistoric Native Americans of this region developed and implemented subsistence 

strategies which were at least partly determined by upland-lowland environmental contrast 

and by annual prolonged flooding of the lowlands. Intensive archeological research in the 

Felsenthal region has not yet adduced the subsistence-settlement system for any prehistoric 

culture or period (Chapter 3). Certainly regional substantive data for Mississippi period 

populations are approaching the level needed to attain this goal (Rolingson and Schambach 

1981, and this report). 

In Figure 70 we present a general seasonal model for extractive activity in the Grand 

Marais Lowland. This model pertains most closely to the Mississippi period (A.D. 1100- 

1700), but may be applicable on a general level to earlier cultural periods when lowland 

environments were similar. Not only Felsenthal Project data, but also Comeaux’s (1972: 

Chapter 10) discussion of Indian and Euramerican “swamp life” in the Atchafalaya Basin, 

south-central Louisiana, are incorporated in Figure 70. The Atchafalaya River overflows 

its banks in spring, inundating the entire basin, and here fishing has always dominated 

“swamp life.” 

Our model emphasizes the dual nature of floodplain activity. During the prolonged 

winter-spring hydroperiod, fishing occurs not only in deep channels or lakes, but “in the 

woods” where concentrated foraging and spawning species are available. Some trapping 

and collecting activity focuses on the margins of the flooded basin. When the floodplain 

emerges in summer-fall, fishing continues in permanent channels and lakes, but is supplemented 

markedly by hunting and collecting activity. We noted earlier the potential for


Figure 70. A seasonal model for subsistence activity and resource use in the 

Grand Marais Lowland. The bold outer ring represents the annual 

hydroperiod. The more labor intensive and productive extractive 

activities are shown at increasing distance from the center.


mass summer harvesting of fish in dwindling lakes and pools as floodwaters receded (Chapter 

2). The counterpart condition, rising floodwaters in late fall temporarily concentrated 

terrestrial animals on ridges and islands, where they also were susceptible to mass capture 

techniques. It therefore seems likely that extractive activity intensified during these episodes 

of rising and falling floodwaters. 

Other times of intensive activity would be determined by seasonal migration, movement, 

rutting, spawning, feeding, foraging, and other behavior of important food animals 

and by ripening of floodplain nut or seed crops. Buffalo fishing in August may serve here as 

an example; an early historical account states rather colorfully, “I have seen the Saline river 

black with Buffalo fish in August, their floating time, which [Jack Fogle, a Frenchman born 

in the area in 1818] would kill with a gig to feed his hogs on, and give us all we wanted...” 

(W. T. Martin in Etheridge 1959:xviii). 

Although the floodplain in the Grand Marais Lowland had distinctive annual emergent 

(summer-fall) and inundated (winter-spring) phases, which could be nearly of equal 

length, it is clear from our model that extractive activity could be continuous throughout the 

year. During winter-spring all such activity was implemented by dugout canoe or by foot 

along the accessible margins of the basin. 

Moreover, extractive sites in the floodplain, concentrated along the low narrow levees 

of the rivers and their tributaries, can only represent summer-fall camps (although dry 

years with reduced hydroperiod could extend these occupations). The prehistoric extractive 

camps in the Felsenthal Project area, at about 20 m above sea level, were just as distinctively 

seasonal as their functional counterparts in a high mountain range.

References Cited 

Allen, J. R. L. 

1965 A Review of the Origin and Characteristics of Recent Alluvial Sediments. Sedimentology 

5:89-191. 

Arkansas Archeological Survey 

1978 Proposal for a Cultural Resources Survey and Evaluation: A) Calion Lock and 

Dam Site; B) Felsenthal Navigation Pool, South-central Arkansas. Ms. on file, Arkansas 

Archeological Survey, Fayetteville. 

Arkansas Department of Planning 

1974 Arkansas Natural Area Plan. Arkansas Department of Planning, Little Rock. 

Arkansas Gazette 

Newspaper published at Little Rock, Arkansas. Available on microfilm at the Arkansas 

History Commission, Little Rock. 

Asch, David L., Kenneth B. Farnsworth, and Nancy B. Asch 

1979 Woodland Subsistence and Settlement, West Central Illinois. Ms. on file, Foundation 

for Illinois Archeology. 

Baker, Charles M. 

1974 A Study of Aboriginal Novaculite Exploitation in the Ouachita Mountains of Southcentral 

Arkansas. Unpublished M. A. thesis, University of Arkansas. 

1975 Site Abandonment and the Archeological Record: An Empirical Case for Anticipated 

Return. Arkansas Academy of Science Proceedings 29:10-11. 

Baerris, David A., Joan E. Freeman, and James V. Wright 

1958 The Contracting Stem Projectile Point in Eastern Oklahoma. Bulletin of the Oklahoma 

Anthropological Society 6:61-82. 

Barth, Fredrick (editor) 

1969 Ethnic Groups and Boundaries. Little-Brown, Boston. 

Bartlett, Charles S. Jr. 

1963 The Tom’s Brook Site–3JO1: A Preliminary Report. In Arkansas Archeology, 1962, 

edited by Charles R. McGimsey III, pp. 15-67. Arkansas Archeological Society, Fayetteville.

280 

Bell, Robert E. 

1958 Guide to the Identification of Certain American Indian Projectile Points. Oklahoma 

Anthropological Society, Special Bulletin 1. 



Belmont, John S., and Stephen Williams 

1981 Painted Pottery Horizons in the Southern Mississippi Valley. Geoscience and Man 

22:19-42. 

Binford, Lewis R. 

1980 Willow Smoke and Dogs’ Tails: Hunter-gatherer Settlement-Systems and Archeological 

Site Formation. American Antiquity 45:4-20. 

Binford, Lewis R., and Sally R. Binford 

1966 A Preliminary Analysis of Functional Variability in the Mousterian of Levallois 

Facies. In Recent Studies in Paleoanthropology, edited by J. D. Clarkland and F. C. Howell, 

pp. 238-295. American Anthropologist 68(2). 

Bolton, Herbert Eugene 

1914 Athanase de Mezieres and the Louisiana Texas Frontier, 1768-1780 (Vol. 2). Arthur J. 

Clarke, Cleveland. 

Bradley County 

County Court Records. Bradley County Courthouse, Warren. 

Brain, Jeffry P. 

1978 Late Prehistoric Settlement Patterning in the Yahoo Basin and Natchez Bluffs 

Regions of the Lower Mississippi Valley. In Mississippian Settlement Patterns, edited by 

Bruce D. Smith, pp. 331-368. Academic Press, New York. 

Branner, John C. 

1899 Some Old French Place Names in Arkansas. Modern Language Notes 14:33-40. 

Brown, James A. 

1976 Spiro Studies: The Artifacts (Vol. 4). Stovall Museum of Science and History and 

University of Oklahoma Resarch Institute, Norman.

Brown, Sandra, Mark M. Brinson, and Ariel E. Lugo 

1979 Structure and Function of Riparian Wetlands. In Strategies for Protection and Management 

of Floodplain Wetlands and Other Riparian Ecosystems, coordinated by R. Roy 

Johnson and J. Frank McCormick, pp. 17-31. U.S. Department of Agriculture, General 

Technical Report WO-12. 

Brunsden, D., and R. Kesel 

1974 Slope Development on a Mississippi River Bluff in Historic Time. Journal of Geology 

81:576-598. 

Bryson, Reid A., David A. Baerreis, and Wayne 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 Wakefield Dort, Jr. and J. 

Knox Jones, Jr. Department of Geology, University of Kansas Special Publication 3:53- 

74. 

Buchanan, Thomas M. 

1973 Key to the Fishes of Arkansas. Arkansas Game and Fish Commission, Little Rock. 

1974 Threatened Native Fishes of Arkansas. In Arkansas Natural Area Plan, pp. 67-92. 

Arkansas Department of Planning, Little Rock. 

Buckner, John W. 

1979 Wilderness Lady: A History of Crossett, Arkansas. Rose Publishing, Little Rock. 

Byrd, Kathleen M., and Robert H. Neuman 

1978 Archaeological Data Relative to Prehistoric Subsistence in the Lower Mississippi 

Alluvial Valley. Geoscience and Man 19:9-21. 

Cain, Robert H. 

1974 Pimple Mounds: A New Viewpoint. Ecology 55:178-182 

Callahan, Errett 

1980 Spatial Organization of the Work Areas of Three Contemporary Flintknappers. 

Quarterly Bulletin of the Archeological Society of Virginia 35:101-108. 

Camden in 1855 

1969 Unidentified newspaper clipping. Reprinted in the Ouachita County Historical 

Quarterly 1(4):10-12. 

281

282 

Chomko, Stephen A., and Gary W. Crawford 

1978 Plant Husbandry in Prehistoric Eastern North America: New Evidence for its 

Development. American Antiquity 44:405-408. 

Comeaux, Malcolm L. 

1972 Atchafalaya Swamp Life, Settlement and Folk Occupations. Geoscience and Man 2. 

Compton, Robert R. 

1962 Manual of Field Geology. John Wiley and Sons, New York. 

Conner, William H., and John W. Day 

1976 Productivity and Composition of a Bald Cypress-Water Tupelo Site and a Bottomland 

Hardwood Site in a Louisiana Swamp. American Journal of Botany 63(10):1354- 

1364. 

Connaway, John M., Samuel O. Brooks, and Samuel O. McGahey 

1977 The Denton Site: A Middle Archaic Occupation in the Northern Yazoo Basin, Mississippi. 

Mississippi Department Of Archives and History, Archaeological Report 4. 

Cordell, Anna H. 

1951 Champagnolle: A Pioneer River Town. Arkansas Historical Quarterly 10:37-47. 

Cordell, Mrs. Robert V. M. 

1947 Lapile: A Pioneer Community. Arkansas Historical Quarterly 6:273-285. 

Crockett, John W. 

1904 Biennial Report of the Secretary of State. Little Rock, Arkansas. 

Curry, Corliss C. 

1960 Early Timber Operations in Southeast Arkansas. Arkansas Historical Quarterly 

19:111-118. 

D’Anemours, Chevalier Paulinier 

1803 Memoire sur le District de Ouachita dans la Provinc de la Louisianne. Ms. on file, 

American Philosophical Society Library, Philadelphia. 

Davis, E. Mott, editor 

1961 Proceedings of the Fifth Conference on Caddoan Archeology. Texas Archeological 

Society Bulletin 31:77-143. 

1970 Texas. In Archeological and Historical Resources of the Red Red River Basin, edited by 

Hester A. Davis. Arkansas Archeological Survey Research Series 1:27-68.

Davis, Hester A. 

1981 A State Plan for the Conservation of Archeological Resources in Arkansas. Volume 

II: Research Study Units. Ms. on file, Arkansas Archeological Survey, Fayetteville. 

Deeter, E. B., W. J. Latimer, and C. E. Born 

1925 Soil Survey of Bradley County, Arkansas. U.S. Department of Agriculture, Soil Survey 

Report 7. 

DeJarnette, David L., Edward B. Kurjack, and James W. Cambron 

1962 Stanfield-Worley Bluff Shelter Excavations. Journal of Alabama Archaeology 8:1-111. 

Derley, Jim 

1979 The Honey Creek Point: A Prairie County Original? Field Notes, Newsletter of the 

Arkansas Archeological Society 170:5-6. 

Dickinson, S. D., and Harry J. Lemley 

1939 Evidences of the Marksville and Coles Creek Complexes at the Kirkham Site, 

Clark County, Arkansas. Texas Archeleogical and Paleontological Society Bulletin 11:139- 

189. Reprinted 1963 in The Arkansas Archeologist 8(4):57-76. 

Dillehay, Tom D. 

1973 Small Archeological Site Investigations for Interpretation of Site Activities. Texas 

Archeological Society Bulletin 44:169-177. 

Etheridge, Y. W. 

1959 History of Ashley County, Arkansas. Arkansas Historical Series 8. Press-Argus, Van 

Buren. 

Ewel, Katherine Carter 

1979 Riparian Ecosystems: Conservation of Their Unique Characteristics. In Strategies 

for Protection and Management of Floodplain Wetlands and Other Riparian Ecosystems, 

coordinated by R. Roy Johnson and J. Frank McCormick, pp. 56-62. U.S. Department of 

Agriculture, General Technical Report WO-12. 

Felsenthal Press 

Newspaper published at Felsenthal, Arkansas. Available on microfilm at the Arkansas 

History Commission, Little Rock. 

Fenenga, Franklin 

1953 The Weights of Chipped Stone Points: A Clue to Their Functions. Southwestern 

Journal of Anthropology 9:309-323. 

283

284 

Fisk, Harold N. 

1944 Geological Investigation of the Alluvial Valley of the Lower Mississippi River. 

War Department, U.S. Army Corps of Engineers, Publication 52. 

1947 Fine-grained Alluvial Deposits and Their Effects on Mississippi River Activity. 

War Department, Corps of Engineers, Waterways Experiment Station, Vol. 1. 

Fitzhugh, William W. 

1972 Environmental Archeology and Cultural Systems in Hamilton Inlet, Labrador: A 

Survey of the Central Labrador Coast from 3000 B.C. to the Present. Smithsonian Contributions 

to Anthropology 16. 

Flannery, Kent V. 

1968 Archaeological Systems Theory and Early-Mesoamerica. In Anthropological Archeology 

in the Americas, edited by Betty J. Meggers, pp. 67-87. Anthropological Society of 

Washington, Washington D. C. 

1976 The Early Mesoamerican Village. Academic Press, New York. (editor) 

Fleetwood, A. R. 

1969 Geological Investigation of the Ouachita River Area, Lower Mississippi Valley. 

U.S. Army Corps of Engineers, Technical Report 81-69-2. Vicksburg. 

Flenniken, J. Jeffrey, and James C. Haggarty 

1979 Trampling as an Agency in the Formation of Edge Damage: An Experiment in 

Lithic Technology. Northwest Anthropological Research Notes 13(2):208-214. 

Ford, James A. 

1935 Ceramic Decoration Sequence at an Old Indian Village Site near Sicily Island, 

Louisiana. Louisiana Department of Conservation, Louisiana Geological Survey Anthropological 

Study 1. 

1936 Analysis of Indian Village Site Collections from Louisiana and Mississippi. Department 

of Conservation, Louisiana Geological Survey, Anthropological Study 2. 

1951 Greenhouse: A Troyville-Coles Creek Period Site in Avoyelles Parish, Louisiana. 

American Museum of Natural History, Anthropological Papers 44(Part 1):1-132.

1961 Menard Site: The Quapaw Village of Osotouy on the Arkansas River. Anthropological 

Papers of the American Museum of Natural History 48:Part 2. 

Ford, James A., and George I. Quimby, Jr. 

1945 The Tchefuncte Culture, an Early Occupation of the Lower Mississippi Valley. 

Society for American Archaeology Memoirs 2. (Kraus Reprint 1974). 

Ford, James A., and Clarence H. Webb 

1956 Poverty Point, a Late Archaic Site in Louisiana. Anthropological Papers of the American 

Museum of Natural History 46:Part 1. 

Ford, James A., and Gordon R. Willey 

1940 Crooks Site, a Marksville Period Burial Mound in La Salle Parish, Louisiana. Department 

of Conservation, Louisiana Geological Survey, Anthropological Study 3. 

Foster, James P. 

1864 Report to Adm. David D. Porter, April 15, 1864. In Official Records of the Union and 

Confederate Navies in the War of the Rebellion, Series 1, 26(Part 1):236-237. 

Fowke, Gerard 

1928 Archeological Investigations II. Bureau of American Ethnology, Annual Report 

44:399-540. 

Fowler, Melvin L. 

1959 Summary Report of Modoc Rock Shelter. Illinois State Museum Report of Investigations 

8. 

Frondel, Clifford 

1962 The System of Minerology of James Dwight Dana and Edward Salisbury Dana, Yale University 

1837-1892, Vol. III, Silica Minerals. John Wiley and Sons, New York. 

General Land Office 

1827 Plat maps: T17S R11W; T17S R10W; T18S R10W; T18S R9W. General Land Office, 

Little Rock. 

1842 Plat maps: T19S R10W; T19S R9W. General Land Office, Little Rock. 

1843 Plat maps: T17S R10W; T18S R9W. General Land Office, Little Rock. 

1844 Plat maps: T17S R11W; T18S R11W; T18S R10W; T19S R11W; T19S R9W. General 

Land Office, Little Rock. 

285

286 

1845 Plat maps: T17S R11W; T17S R10W; T18S R11W; T19S R11W; T19S R10W; T19S 

R9W. General Land Office, Little Rock. 

Giardino, Marco J. 

1979 Cultural Resources Survey and Evaluation of the Cane Creek R. C. & D. Measure, 

Lincoln County, Arkansas. New World Research Report of Investigations 16. 

Gill, Hiram V., Don C. Avery, Fred C. Larance, and Charles L. Fultz 

1979 Soil Survey of Ashley County, Arkansas. U.S. Department of Agriculture, Soil Conservation 

Service. 

Goodspeed 

1890 Biographical and Historical Memoirs of Southern Arkansas. Goodspeed Publishing, 

Chicago. 

Goodyear, Albert C., John H. House, and Neal W. Ackerly 

1979 Laurens-Anderson: An Archeological Study of the Inter-riverine Piedmont. Institute 

of Archeology and Anthropology, University of South Carolina Anthropological Studies 

4. 

Gordon, M. E., L. Russert Kraemer, and A. V. Brown 

1979 Unionaceae of Arkansas: Historical Review, Checklist, and Observations on Distributional 

Patterns. Bulletin of the American Malecological Union 1979:31-37. 

Gosselink, J. G., and R. E. Turner 

1978 The Role of Hydrology in Freshwater Wetland Ecosystems. In Freshwater Wetlands: 

Ecological Processes and Management Potential, edited by Ralph E. Good, Dennis 

E. Whigham, and Robert L. Simpson, pp. 63-77. Academic Press, New York. 

Green, Juanita Whitaker 

1954 The History of Union County. (Mimeograph copies) Jack Lee, Eldorado, Arkansas. 

Green, Thomas J., and Cheryl A. Munson 

1978 Mississippian Settlement Patterns in Southwestern Indiana. In Mississippian Settlement 

Patterns, edited by Bruce D. Smith, pp. 293-330. Academic Press, New York. 

Green, W. D. 

1864 Letter to Col. Powell Clayton, April 20, 1864. In War of the Rebellion, the Official 

Records of the Union and Confederate Armies, Series 1, 34(Part 3):236-237. 

Greene, Glen S., Ernest R. Williams, and G. R. Dennis Price 

1975 The Search for Fort Miro: A Colonial Spanish Fort in the Ouachita Valley. Northeast 

Louisiana Archeological Society Popular Series 1.

Greengo, Robert E. 

1964 Issaquena: An Archeological Phase in the Yazoo Basin of the Lower Mississippi 

Valley. Memoirs of the Society for American Archaeology 18. 

Halls, Lowell K. (editor) 

1977 Southern Fruit-producing Woody Plants Used by Wildlife. U.S. Department of 

Agriculture, Southern Forest Experiment Station, General Technical Report SO-16. 

Hally, David J. 

1967 Post-Coles Creek Cultural Development in the Upper Tensas Basin of Louisiana. 

Proceedings of the Twenty-third Southeastern Archaeological Conference, SEAC Bulletin 

6:36-39. 

Hardesty, Donald L. 

1980 The Use of General Ecological Principles in Archaeology. In Advances in Archaeological 

Method and Theory (Volume 3), edited by Michael B. Schiffer, pp. 158-188. Academic 

Press, New York. 

Harrington, Mark R. 

1920 Certain Caddo Sites in Arkansas. Museum of the American Indian Heye Foundation, 

Indian Notes and Monographs 10. 

Hayden, Julian D. 

1965 Fragile Pattern Areas. American Antiquity 31:272-276. 

Haynes, C. Vance 

1973 Soil Redefined. Geotimes 18:8. 

Heartfield, Lorraine, and G. R. Dennis Price 

1976 An Archeological Survey of Portions of the Monroe to Sandy Bayou Levee, 

Ouachita River Levees, Louisiana. Ms. on file, U.S. Army Corps of Engineers, Vicksburg. 

Heartfield, Price, and Greene, Inc. 

1980 A Cultural Resources Reconnaissance of the Felsenthal National Wildlife Refuge, 

Arkansas. Ms. on file, Interagency Archeological Services, Heritage Conservation and 

Recreation Service, Atlanta. 

Hemmings, E. Thomas 

1969 Excavation and Analysis of Living Floors in Alluvial Sites. Southeastern Archaeological 

Conference Bulletin 11:41-45. 

287

288 

1975 An Archaeological Survey of the South Prong of the Alafia River, Florida. Florida 

Anthropologist 28:41-52. 

1979 The Felsenthal Project: Archeological Survey and Testing in the Ouachita Valley, 

South Arkansas. Paper presented at the 36th Annual Meeting of the Southeastern 

Archaeological Conference, Atlanta. 

1980 Spirit Lake (3LA83): Test Excavations in a Late Caddo Site on the Red River, 

Southwest Arkansas. Ms. on file, Arkansas Archeological Survey, Fayetteville. 

Hemmings, E. Thomas, and Katherine Dinnel 

1980 Analysis of a Quapaw Hunting Camp on the Saline River, Southeast Arkansas. 

Paper presented at the 37th Annual Meeting of the Southeastern Archaeological Conference, 

New Orleans. 

Hilliard, Jerry E. 

1981 Selection and Use of Acorn Species by Late Prehistoric Ozark Inhabitants. Paper 

presented at the 46th Annual Meeting of the Society for American Archaeology, San 

Diego. 

Hodges, T. L., and Mrs. T. L. Hodges 

1945 Suggestions for the Identification of Certain Mid-Ouachita Pottery as Cahinnio 

Caddo. Texas Archeological and Paleontological Society Bulletin 16:98-116. Reprinted 1963 

in The Arkansas Archeologist 4(8):1-12. 

Hoffman, Michael P. 

1970a Arkansas. In Archeological and Historical Resources of the Red River Basin, edited 

by Hester A. Davis. Arkansas Archeological Survey Research Series 1:137-187. 

1970b Three Sites in Millwood Reservoir: The Hutt Site, the Stark Site, and the Beard 

Lake Site. Arkansas Archeological Survey Research Series 2. 

1977 The Kinkead-Mainard Site, 3PU2: A Late Prehistoric Quapaw Phase Site near 

Little Rock, Arkansas. The Arkansas Archeologist 16-18:1-41. 

Holmes, W. H. 

1903 Aboriginal Pottery of the Eastern United States. Bureau of American Ethnology, Annual 

Report 20:1-237. 

House, John H. 

1975 A Functional Typology for Cache Project Surface Collections. In the Cache River 

Archeological Project: An Experiment in Contract Archeology, edited by Michael B. 

Schiffer and John H. House. Arkansas Archeological Survey Research Series 8:55-73.

1980 Archaic Occupation in the Arkansas River Lowland. Arkansas Archeological Society, 

Field Notes 171:5-10. 

1981 Evolution of Complex Society in East-central Arkansas: An Overview of Environments 

and Regional Databases. In Arkansas Archeology in Review: A Symposium, edited 

by Marvin D. Jeter and Neal L. Trubowitz. Arkansas Archeological Survey Research 

Series (in press). 

House, John H., and Marvin D. Jeter (editors) 

1981 Archeological Investigations in the Boydell Locality, Southeast Arkansas. Ms. on 

file, Arkansas Archeological Survey, Monticello. 

House, John H., and Henry McKelway 

1980 Mississippian and Quapaw on the Lower Arkansas. Ms. on file, Arkansas Archeological 

Survey, Fayetteville. 

House, John H., and Michael B. Schiffer 

1975 Archeological Survey in the Cache River Basin. In The Cache River Archeological 

Project: An Experiment in Contract Archeology, edited by Michael B. Schiffer and 

John H. House. Arkansas Archeological Survey Research Series 8:37-53. 

House, John H., and James W. Smith 

1975 Experiments in Replication of Fire-cracked Rock. In The Cache River Archeological 

Project: An Experiment in Contract Archeology, edited by Michael B. Schiffer and 

John H. House. Arkansas Archeological Survey Research Series 8:75-80. 

Hrdlicka, Ales 

1909 Report on an Additional Collection of Skeletal Remains from Arkansas and Louisiana. 

Journal of the Academy of Natural Sciences of Philadelphia (second series) 14:171- 

249. 

Huffman, R. T. 

1976 The Relation of Flood Duration Patterns to Dominant Forest Species Associations Occurring 

on Selected First Bottom Sites of the Ouachita River Drainage Basin in Southern Arkansas. 

Unpublished Ph.D. dissertation, University of Arkansas. 

Huxtable, J., M. J. Aitken, and J. C. Weber 

1972 Thermoluminescent Dating of Baked Clay Balls of the Poverty Point Culture. 

Archaeometry 14(2):269-275. 

James, Frances C. 

1974 Threatened Native Birds of Arkansas. In Arkansas Natural Area Plan, pp. 107-122. 

Arkansas Department of Planning, Little Rock. 

289

290 

Jeter, Marvin D. 

1980 The Archeology of Southeast Arkansas: An Overview for the 1980s. Paper presented 

at the 45th Annual Meeting of the Society for American Archaeology, Philadelphia. 

Jeter, Marvin D., David B. Kelley, and George P. Kelley 

1979 The Kelley-Grimes Site: A Mississippi Period Burial Mound, Southeast Arkansas, 

Excavated in 1936. Arkansas Archeologist 20:1-51. 

Johnson, Roland 

1979 Interview at Felsenthal, Arkansas, September 21, by Beverly Watkins and E. 

Thomas Hemmings. Transcribed tape on file, Arkansas Archeological Survey, Fayetteville. 

Kellam, Robert F. 

1970 Diary, September 1-December 31, 1860. Ouachita County Historical Quarterly 2(2):6- 

18. 

1971 Diary, January 1-March 31, 1861. Ouachita County Historical Quarterly 2(3):11-20. 

Keller, W. D. George W. Viele, and Clayton H. Johnson 

1977 Texture of Arkansas Novaculite Indicates Thermally Induced Metamorphism. 

Journal of Sedimentary Petrology 47:834-843. 

Kelley, David B. 

1979 Cultural Resources Assessment of the Calion Lock and Dam, Union County, Arkansas. 

Ms. on file. Arkansas Archeological Survey, Fayetteville. 

Kelley, Lucretia S. 

1979 Animal Resource Exploitation by Early Cahokia Populations on the Merrill Tract. 

Illinois Archaeological Survey, Circular 4. 

King, James E., and W. H. Allen, Jr. 

1977 A Holocene Vegetation Record from the Mississippi River Valley, Southeastern 

Missouri. Quarternary Research 8:307-323. 

Klinger, Timothy C. (assembler) 

1975 Mississippian Communities in the St. Francis Basin: A Central Place Model. Arkansas 

Academy of Science Proceedings 29:50-51. 

1979 Hampton: An Archeological and Historical Overview of a Proposed Strip Mine 

Tract in South Central Arkansas. Arkansas Archeological Survey Research Reports 19.

Krieger, Alex D. 

1946 Culture Complexes and Chronology in Northeastern Texas. University of Texas 

Publication 4640. 

Lafferty, Robert H. III, Jeffrey L. Otinger, Sandra Clements S. Scholtz, W. Fredrick Limp, Beverly 

Watkins, and Robert D. Jones 

1981 Settlement Prediction in Sparta. Arkansas Archeological Survey Research Series 14 (in 

press). 

Langenheim, Jean H. 

1964 Present Status of Botanical Studies of Ambers. Harvard University Botanical Museum 

Leaflets 20(8):225-28. 

1969 Amber: A Botanical Inquiry. Science 163:1157-1169. 

Langenheim, Jean H., and Curt W. Beck 

1968 Catalogue of Infrared Spectra of Fossil Resins (Ambers), I. North and South 

America. Harvard University Botanical Museum Leaflets 22(3):65-120. 

Larance, Red C. 

1961 Soil Survey of Bradley County, Arkansas. U.S. Department of Agriculture, Soil Conservation 

Services Series 1958(1). 

Lemley, Harry J., and S. D. Dickinson 

1937 Archeological Investigations on Bayou Macon in Arkansas. Texas Archeological and 

Paleontological Society Bulletin 9:11-47. Reprinted 1964 in Arkansas Archeologist 5(2):21-39. 

Leopold, Luna B., M. Gordon Wolman, and John P. Miller 

1964 Fluvial Processes in Geomorphology. W. H. Freeman, San Francisco. 

Lewis, R. Barry 

1974 Mississippian Exploitative Strategies: A Southeast Missouri Example. Missouri 

Archeological Society Research Series 11. 

Lewis, Thomas M. N., and Madeline Knebeg Lewis 

1961 Eva: An Archaic Site. University of Tennessee Press, Knoxville. 

Limp, W. Fredrick, and Van A. Reidhead 

1979 An Economic Evaluation of the Potential of Fish Utilization in Riverine Environments. 

American Antiquity 44:70-78. 

Lischka, Joseph J. 

1973 Preliminary Report on Test Excavations of Prehistoric Sites in the Felsenthal National 

Wildlife Refuge, 1972. Ms. on file, Arkansas Archeological Survey, Fayetteville. 

291

292 

Lloyd, James T. 

1856 Lloyd’s Steamboat Directory and Disasters on the Western Rivers. Jas. T. Lloyd, 

Cincinnati. 

Maisenhelder, Louis C. 

1958 Understory Plants of Bottomland Forests. U.S. Department of Agriculture, Southern 

Forest Experiment Station, Occasional Paper 165. 

Marquardt, William H., and Patty Jo Watson 

1976 Excavation and Recovery of Biological Remains from Two Archaic Shell Middens 

in Western Kentucky. Paper presented at the Southeastern Archaeological Conference, 

Tuscaloosa, Alabama. 

Marsh, Daniel L. 

1977 The Distribution of Cane, Arundinaria gigantia (Poaceae: Bambusoideae). Arkansas 

Academy of Science Proceedings 31:72-74. 

Martin, W. T. 

1936 Early History. Works Projects Administration, Federal Writers Project. Ms. on file, 

Ashley County Place File, Arkansas History Commission, Little Rock. 

McDermott, John F. 

1941 A Glossary of Mississippi Valley French, 1673-1850. Washington University Studies 

Language and Literature 12. 

1963 The Western Journals of Dr. George Hunter, 1796-1805. Transactions of the American 

Philosophical Society 53(Part 4). 

McGimsey, Charles R. III 

1964 Quapaw Sites on the Lower Arkansas River. Ms. on file, Arkansas Archeological 


1969 Indians of Arkansas. Arkansas Archeological Survey Popular Series 1. 

McMahan, Captain Jesse T. 

[1913] Memories of Steamboat Life on the Ouachita. Clipping in Grinstead Chapter History, 

United Daughters of the Confederacy, compiled by Mr. J. T. Sifford, pp. 85-85. 

Public Library of Camden and Ouachita County, Camden, Arkansas. 

McManus, J. W. 

1963 The McElroy Site: A Coles Creek-Caddoan Transitional Site in Union County, 

Arkansas. Arkansas Archeologist 4(3):1-5.

Meanley, Brooke 

1972 Swamps, River Bottoms, and Canebrakes. Barre Publications, Barre, Massachusetts. 

Michie, James L. 

1980 An Archeological Survey of Congaree Swamp: Cultural Resources Survey and 

Assessment of a Bottomland Environment in Central South Carolina. Institute of Anthropology 

and Archeology, University of South Carolina, Research Manuscript Series 163. 

Mills, Gary B. 

1978 Of Men and Rivers: The Story of the Vicksburg District. U.S. Army Engineer District, 

Vicksburg. 

Mitchell, Jennie O’Kelly, and Robert Dabney Calhoun 

1937 The Marquis de Maison Rouge, the Baron de Bastrop, and Colonel Abraham Morhouse 

— Three Ouachita Valley Soldiers of Fortune. The Maison Rouge and Bastrop 

Spanish Land “Grants.” Louisiana Historical Quarterly 20:289-462. 

Moore, Clarence B. 

1909 Antiquities of the Ouachita Valley. Journal of the Academy of Natural Sciences of 

Philadelphia 14. 

1913 Some Aboriginal Sites in Louisiana and in Arkansas. Journal of the Academy of 

Natural Sciences of Philadelphia 16(Part 1). 

Moore, Dwight M. 

1960 Trees of Arkansas. Arkansas Forestry Commission, Little Rock. 

Morse, Dan F. 

1963 The Steuben Village and Mounds: A Multicomponent Late Hopewell Site in Illinois. 

Museum of Anthropology, University of Michigan, Anthropological Papers 21. 

1970 The Big Creek Point. Central States Archaeological Journal 17:20-23. 

1973 Nodena: An Account of 75 Years of Archeological Investigation in Southeast Mississippi 

County, Arkansas. Arkansas Archeological Survey Research Series 4. (editor) 

1975 Research Potential in Terms of Questions of Regional Prehistory. In The Cache 

River Archeological Project: An Experiment in Contract Archeology, assembled by 

Michael B. Schiffer and John H. House. Arkansas Archeological Survey Research Series 

8:187-198. 

293

294 

1980a Cultural Problems. In Excavation, Data Interpretation and Report on the Zebree 

Homestead Site, Mississippi County, Arkansas, edited by Dan F. Morse and Phyllis A. 

Morse (Chapter 4). Ms. on file. Memphis District, U.S. Army Corps of Engineers. 

1980b The Big Lake Household and Community. In Excavation, Data Interpretation and 

Report on the Zebree Homestead Site, Mississippi County, Arkansas, edited by Dan 

F. Morse and Phyllis A. Morse (Chapter 21). Ms. on file. Memphis District, U.S. Army 

Corps of Engineers. 

1981 Regional Overview of Northeast Arkansas. In Arkansas Archeology in Review. 

Arkansas Archeological Survey Research Series (in press). 

Morse, Dan F., Daniel Wolfman, and Herbert Haas 

1980 Radiocarbon Dating. In Excavation, Data Interpretation and Report on the Zebree 

Homestead Site, Mississippi County, Arkansas, edited by Dan F. Morse and Phyllis A. 

Morse (Chapter 11). Ms. on file. Memphis District, U.S. Army Corps of Engineers. 

Moseley, M. E., and C. J. Mackey 

1972 Peruvian Settlement Pattern Studies and Small Site Methodology. American Antiquity 

37:67-81. 

Muller, Jon 

1978 The Kinkead System: Mississippian Settlement in the Environs of a Large Site. 

In Mississippian Settlement Patterns, edited by Bruce D. Smith, pp. 269-292. Academic 


Newcomer, M. H., and G. de G. Sieveking 

1980 Experimental Flake-scatter Patterns: A New Interpretive Technique. Journal of Field 

Archeology 7:345-352. 

Ouachita Telegraph 

Newspaper published at Monroe, Louisiana. Available on microfilm, Northeast Louisiana 

University, Monroe. 

Owen, David Dale 

1860 Second Report of a Geological Reconnaissance of the Middle and Southern Counties of 

Arkansas. C. Sherman and Son, Philadelphia. 

Owens, Harry P. 

1979 An Assessment of Historic Period Cultural Resources along the Yazoo River 

Between River Miles 75.6 and 273.0, Including Tchula Lake and Honey Island, Mississippi. 

Ms. on file, Center for Archaeological Resources, University of Mississippi.

Parmalee, Paul W., and Walter E. Klippel 

1974 Freshwater Mussels as a Prehistoric Food Resource. American Antiquity 39:421- 

434. 

Patterson, James L. 

1971 Floods in Arkansas, Magnitude and Frequency Characteristics through 1968. Arkansas 

Geological Commission, Water Resources Circular 11. 

Peebles, Christopher S. 

1978 Determinants of Settlement Size and Location in the Moundville Phase. In Mississippian 

Settlement Patterns, edited by Bruce D. Smith, pp. 369-416. Academic Press, 

New York. 

Perino, Gregory 





Phillips, Philip 

1970 Archeological Survey in the Lower Yazoo Basin, Mississippi, 1949-1955. Papers of 

the Peabody Museum of Archaeology and Ethnology 60. Harvard University. 

Phillips, Philip, James A. Ford, and James B. Griffin 

1951 Archaeological Survey in the Lower Mississippi Alluvial Valley, 1940-1947. Papers 

of the Peabody Museum of American Archaeology and Ethnology 25. Harvard University. 

Price, G. R. Dennis 

1979 Cultural Resources Survey Report: Appendix to Design Memorandum No. 58, 

Nine-foot Navigation Channel, Jonesville and Columbia Pools. Ms. on file, U.S. Army 

Corp of Engineers, Vicksburg. 

Price, G. R. Dennis, and Lorraine Heartfield 

1977 Archeological Test Excavation of Site 16OU15 Located on the Monroe to Sandy 

Bayou Levee, Ouachita River Levees, Ouachita Parish, Louisiana. Ms. on file, U.S. 

Army Corps of Engineers, Vicksburg. 

Price, G. R. Dennis et al. 

1975 Excavation at the Site of Historic Fort Miro (16OU3) in Monroe, Louisiana. Ms. on 

file, U.S. Army Corps of Engineers, Vicksburg District. 

295

296 

Quimby, George I,. 

1951 The Fedora Site, West Baton Rouge Parish, Louisiana. Field Museum of Natural History 

Anthropological Series 24(2):81-135. 

Raab, L. Mark 

1976a A Proposal for Archeological Investigations in the Felsenthal Project Area, Arkansas. 

Ms. on file, Arkansas Archeological Survey, Fayetteville. 

1976b A Report of Archeological Reconnaissance in the Felsenthal Project Area, Arkansas. 

Ms. on file, Arkansas Archeological Survey, Fayetteville. 

Raynolds, Lt. Col. W. F. 

1871 Topographical Sketch of Ouachita River in Arkansas from Arkadelphia to the Southern 

State Line. U.S. Army Corps of Engineers, Vicksburg. 

Reagan, Douglas P. 

1974 Threatened Native Reptiles of Arkansas. In Arkansas Natural Area Plan, pp. 101- 

105. Arkansas Department of Planning, Little Rock. 

Reaves, Robert G., and Burney McClurkan 

1970 The Dansbury Creek Site (3CV24). Arkansas Archeological Society Field Notes 70:3-4. 

Reynolds, R. 

1980 The Crossett Story: The Beginning of Forestry in Southern Arkansas and Northern 

Louisiana. U.S. Department of Agriculture, Southern Forest Experiment Station, General 

Technical Report SO-32. 

Ritter, Dale F., W. Fred Kinsey, and Marvin E. Kaufman 

1973 Overbank Sedimentation in the Delaware River Valley During the Last 6000 

Years. Science 179:374-375. 

Robertson, Philip A., and George T. Weaver 

1979 Vegetation and Tree Species Patterns near the Northern Terminus of the Southern 

Floodplain Forest. Ecological Monographs 48:249-267. 

Robinson, Henry W. 

1975 New Distributional Records of Fishes from the Lower Ouachita River System in 

Arkansas. Arkansas Academy of Science Proceedings 29:54-56. 

Rohlf, F. James, John Kishpaugh, and David Kirk 

1972 Numerical Taxonomy System of Multivariate Statistical Programs. State University of 

New York at Stony Brook.

Rolingson, Martha A. 

1970 Archeological Investigation of Bayou Bartholomew, 1969. Arkansas Academy of Science 

Proceedings 24:13-16. 

1971 The Ashley Point. Arkansas Archeologist 12(3):50-52. 

1972 Report on the Preliminary Site Survey of the Felsenthal National Wildlife Refuge, 

1971. Ms. on file, Arkansas Archeological Survey, Fayetteville. 

1973 Initial Draft of the Bartholomew Sequence. Ms. on file, Arkansas Archeological 


1976a The Bartholomew Phase: A Plaquemine Adaptation in the Mississippi Valley. 

In Cultural Continuity and Change: Essays in Honor of James Bennett Griffin, edited by 

Charles C. Cleland, pp. 99-119. Academic Press, New York. 

1976b Forests and Floods: Clues to Mississippi Valley Floodplain Environment in Southeastern 

Arkansas and Significance for Prehistoric Settlement. Paper presented at the 

33rd Southeastern Archaeological Conference, Tuscaloosa, Alabama. 

1979 Emerging Cultural Patterns at the Toltec Site. Paper presented at the 36th Southeastern 

Archaeological Conference, Atlanta, Georgia, November 1979. 

1981 SE Study Units 3, 4, 5 and 6: Tchula, Marksville, Baytown and Northern Coles 

Creek. In A State Plan for the Conservation of Archeological Resources in Arkansas, 

Volume II: Research Study Units. Ms. on file, Arkansas Archeological Survey. 

Rolingson, Martha A., and Frank F. Schambach 

1981 The Shallow Lake Site (3UN9/52) and Its Place in Regional Prehistory. Arkansas 

Archeological Survey Research Series 12. 

Rowland, Eron (editor) 

1930 Life, Letters and Papers of William Dunbar, 1749-1810. National Society of Colonial 

Dames, Jackson, Mississippi. 

Saucier, Roger T. 

1974 Quaternary Geology of the Lower Mississippi Valley. Arkansas Archeological Survey 

Research Series 6. 

Saucier, Roger T., and A. R. Fleetwood 

1970 Origin and Chronologic Significance of Late Quaternary Terraces, Ouachita River, 

Arkansas and Louisiana. Bulletin of the Geological Society of America 81:869-890. 

297

298 

Saunders, W. B., R. H. Mapes, F. M. Carpenter, and W. C. Elsik 

1974 Fossiliferous Amber from the Eocene (Claiborne) of the Gulf Coastal Plain. Geological 

Society of America Bulletin 85:979-984. 

Schambach, Frank F. 

1970 Pre-Caddoan Cultures in the Trans-Mississippi South: A Beginning Sequence. Unpublished 

Ph.D. dissertation, Harvard University. 

1974 A Unique Engraved Steatite Bowl from Southwest Arkansas. Arkansas Archeologist 

15:40-44. 

1976 Some Suggestions for 1976-77 Operations on the Calion-Felsenthal Project. Ms. on 

file, Arkansas Archeological Survey, Fayetteville. 

1979 Summary of Local and Regional Archeology. In Hampton: An Archeological and 

Historical Overview of a Proposed Strip Mine Tract in South Central Arkansas, assembled 

by Timothy C. Klinger. Arkansas Archeological Survey Research Report 19:21-30. 

Schiffer, Michael B. 

1976 Behavioral Archeology. Academic Press, New York. 

Scholtz, James A. 

1967 An Archeological Survey of the Beaver Reservoir in Northwest Arkansas. M. A. thesis, 

University of Arkansas, Fayetteville. 

Scott, Walter 

1940 Surveying the Ouachita. Arkansas Gazette Sunday Magazine, July 14, Little Rock. 

Sealander, John A. 

1979 A Guide to Arkansas Mammals. River Road Press, Conway, Arkansas. 

Setzler, Frank 

1933 Pottery of the Hopewell Type from Louisiana. U.S. National Museum, Proceedings 

28(art. 22). 

Shafer, H. G. 

1973 Lithic Technology at the George C. Davis Site, Cherokee County, Texas. Unpublished 

Ph.D., dissertation, University of Texas. University Microfilms, Ann Arbor. 

Sifford, Mrs. J. T. (compiler) 

n.d. History of Camden scrapbooks. Public Library of Camden and Ouachita County, 

Camden, Arkansas.

Smith, Bruce D. 

1975 Middle Mississippi Exploitation of Animal Populations. Museum of Anthropology, 

University of Michigan, Anthrpological Papers 57. 

1978 Variation in Mississippian Settlement Patterns. In Mississippian Settlement Patterns, 

edited by Bruce D. Smith, pp. 479-503. Academic Press, New York. 

Smith, J. Russell 

1953 Tree Crops: A Permanent Agriculture. Devin-Adair, New York. 

Society of American Foresters 

1954 Forest Cover Types of North America (exclusive of Mexico). Society of American Foresters, 

Washington, D. C. 

Spencer, Annie L. 

1953 The Blacks of Union County. Arkansas Historical Quarterly 12:225-252. 

Sprague, Marshall 

1974 So Vast So Beautiful a Land: Louisiana and the Purchase. Little, Brown, Boston. 

Springer, James W. 

1980 An Analysis of Prehistoric Food Remains from the Bruly St. Martin Site, Louisiana, 

with a Comparative Discussion of Mississippi Valley Faunal Studies. Midcontinenta1 

Journal of Archaeology 5:193-224. 

Stacy, V. K. Pheriba 

1976 Summary Report on Archeological Survey and Excavations in the Felsenthal National 

Wildlife Refuge. Ms. on file, Arkansas Archeological Survey, Fayetteville. 

State Land Office 

Tract Books. State Land Office, Little Rock. 

Suhm, Dee Ann, and Edward B. Jelks (editors) 

1962 Handbook of Texas Archeology: Type Descriptions. Texas Archeological Society, 

Special Publication 1 and Texas Memorial Museum, Bulletin 4. 

Suhm, Dee Ann, and Alex D. Krieger 

1954 An Introductory Handbook of Texas Archeology. Bulletin of the Texas Archeological 

Society 25. 

Swanton, John R. 

1911 Indian Tribes of the Lower Mississippi Valley and Ajacent Coast of the Gulf of 

Mexico. Bureau of American Ethnology Bulletin 43. 

299

300 

Travis, Russell B. 

1955 Classification of Rocks. Quarterly of the Colorado School of Mines 50(1). 

Tucker, Gary E. 

1974 Threatened Native Plants of Arkansas. In Arkansas Natural Area Plan, pp. 39-65. 

Arkansas Department of Planing, Little Rock. 

Union County 

County Court Records. Union County Courthouse, El Dorado, Arkansas. 

U.S. Army Corps of Engineers 

1871 Topographic Sketch of Ouachita River in Arkansas from Arkadelphia to the Southern 

State Line; Made under the Orders of Lieut. Colonel W. F. Raynolds by Justin Straszer, C. E. 

(6 sheets; scale printed on maps). Vicksburg. 

1873 Ouachita River Survey (1:10,000; 34 sheets). Vicksburg. 

1875 Letters received by officer in charge of Ouachita River improvements stationed at 

Monroe, Louisiana, Vicksburg, Mississippi, and Memphis, Tennessee. PI entry 1719, 

RG77, Federal Archives and Records Center, East Point, Georgia. 

1890 Map of Ouachita River Surveyed under the Direction of Gen. W. F. Raynolds, U.S. A., by 

Clement Smith, C. E. (1:30,000; 7 sheets). Vicksburg. 

1894 Basin of Ouachita River and Tributaries from Arkadelphia to Red River (1:200,000; 3 

sheets). Vicksburg. 

1895- Ouachita River Survey (1:10,000; 48 sheets). Vicksburg. 

1897 

1896 Ouachita River Hydrographic Survey Maps. Vicksburg. 

1897 Index Map of Ouachita River Survey, Louisiana and Arkansas, Camden to Mouth in Five 

Sheets (scale printed on maps). Vicksburg. 

1979 Feasibility Report, Ouachita-Black Rivers Nine-foot Navigation Project, Crossett Harbor, 

Ashley County, Arkansas. Vicksburg. 

1980 Cultural Resources Survey Report: Appendix to Design Memorandum No. 58. 

Nine foot Navigation Channel, Jonesville and Columbia Pools. Vicksburg District. 

n.d. Preliminary Plan for Slackwater Navigation, Ouachita River and Tributaries (1:300,000). 

Vicksburg.

U.S. Bureau of Marine Inspection and Navigation 

Enrollment Certificates, RG41. National Archives, Washington D.C. 

U.S. Congress 

1872 Survey of Certain Rivers in the Southern States. 2nd Congress, 2nd Session, House 

Executive Document 25 . 

1874 Ouachita River. 43rd Congress, 1st Session, House Executive Document 259. 

1876 Report of the Chief of Engineers. 44th Congress, 2nd Session, House Executive 

Document 1(1):Part 2. 

1877 Report of the Chief of Engineers. 45th Congress, 2nd Session, House Executive 

Document 1(2):Part 2. 

1900 Ouachita and Black Rivers, Arkansas and Louisiana. 56th Congress, 1st Session, 

House Document 480. 

U.S. Department of Agriculture 

1951 Soil Survey Manual, Handbook No. 18. U.S. Government Printing Office, Washington 

D. C. 

1968 General Soil Map, Union County, Arkansas (1:126, 20). U.S. Department of Agriculture, 

Soil Conservation Service. 

1978 Special Report, Fish and Wildlife Ouachita River Basin Study, Arkansas and Louisiana. 

U.S. Department of Agriculture, Alexandria, Louisiana. 

U.S. Geological Survey 

1939 Felsenthal, Arkansas-Louisiana (1:62,500). U.S. Geological Survey, Washington, D. C. 

1978 Felsenthal NE, NW, SE, and SW, Advance Workprints (1:24,000). U.S. Geological Survey, 

Reston. 

U.S. Post Office 

n.d. Appointments of Postmasters, Arkansas. Ms. on microfilm, Arkansas History 

Commission, Little Rock. 

Veatch, A. C. 

1906 Geology and Underground Water Resources of Northern Louisiana and Southern 

Arkansas. U.S. Geological Survey Professional Paper 46. 

Vita-Finzi, C., and E. S. Higgs 

1970 Prehistoric Economy in the Mount Carmel Area of Palestine: Site Catchment 

Analysis. Proceedings of the Prehistoric Society 36:1-37. 

301

302 

Walker, Winslow 

1936 The Troyville Mounds, Catahoula Parish, Louisiana. Bureau of American Ethnology, 

Bulletin 133. 

Watkins, Beverly 

1981 History and Historical Documentation. In Settlement Prediction in Sparta: A Location 

Analysis and Cultural Resource Assessment in the Uplands of Calhoun County, 

Arkansas, by Robert H. Lafferty III et al., Arkansas Archeological Survey Research 

Series 14. 

Weber, J. Cynthia 

1973 The Paw-Paw Site. Ms. on file, Arkansas Archeological Survey, Fayetteville. 

Webb, Clarence H. 

1959 The Belcher Mound: A Stratified Caddoan Site in Caddo Parish, Louisiana. Society 

for American Archaeology, Memoirs 11. 

1968 The Extent and Content of Poverty Point Culture. American Antiquity 33:297-321. 

1977 The Poverty Point Culture. Geoscience and Man 12. 

1980 Stone Tools and Points of Northwestern Louisiana. Ms. on file with author, 

Shreveport. 

1981 Review of the Shallow Lake Site (3UN9/52) and Its Place in Regional Prehistory 

by Martha A. Rolingson and Frank F. Schambach, Appendix C. Arkansas Archeological 

Survey Research Series 12. 

Webb, Clarence H., Joel Shiner, and E. Wayne Roberts 

1971 The John Pearce Site (16CD56): A San Patrice Site in Caddo Parish, Louisiana. Bulletin 

of the Texas Archeological Society 42:1-47. 

Weinstein, Richard A., and Philip G. Rivet 

1978 Beau Mire: A Late Tchula Period Site of the Tchefuncte Culture, Ascension Parish, 

Louisiana. Louisiana Archeological Survey and Antiquities Commission, Anthropological 

Report 1. 

Welcomme, Robin L. 

1979 Fisheries Ecology of Floodplain Rivers. Longmans, New York. 

Wharton, Charles H., and Mark M. Brinson 

1979 Characteristics of Southeastern River Systems. Strategies for Protection and 

Management of Floodplain Wetlands and Other Riparian Ecosystems, coordinate by 

R. Roy Johnson and J. Frank McCormick, pp. 32-40. U.S. Department of Agriculture, 

General Technical Report WO-12.

Wheat, Joe Ben 

1972 The Olsen-Chubbuck Site: A Paleo-Indian Bison Kill. Society for American Archaeology 

Memoir 26. 

White, Patsy 

1970 Investigation of the Cemetery at the Gee’s Landing Site, 3DR17. Arkansas Archeologist 

11(1):1-20. 

Willey, Gordon R., and Philip Phillips 

1958 Method and Theory in American Archaeology. University of Chicago Press. 

Williams, Stephen 

1980 Armorel: A Very Late Phase in the Lower Mississippi Valley. Southeastern Archaeological 


Williamson, Frederick W., and Lillian Herron Williamson 

1939 Northeast Louisiana, a Narrative History of the Ouachita River Valley and the Concordia 

Country. Historical Record Association, Monroe, Louisiana. 

Wilmsen, Edwin H. 

1968 Lithic Analysis in Paleoanthropology. Science 161:982-987. 

Winters, Howard Dalton 

1969 The Riverton Culture: A Second Millenium Occupation in the Central Wabash 

Valley. Illinois Archaeological Survey, Monograph 1, and Illinois State Museum, Reports of 

Investigations 13. 

Wood, W. Dean 

1976 A Sampling Scheme for Subsurface Archeological Survey. Southeastern Archaeological 


Wood, W. Raymond 

1963 Two New Projectile Points: Homan and Agee Points. Arkansas Archeologist 4(2):1- 

6. 

Wood, W. Raymond, and Donald Lee Johnson 

1978 A Survey of Disturbance Processes in Archeological Site Formation. In Advances in 

Archaeological Method and Theory Volume 1, edited by Michael B. Schiffer, pp. 315-381. 

Academic Press, New York. 

Wood, W. Raymond, and R. Bruce McMillan 

1976 Prehistoric Man and His Environments: A Case Study in the Ozark Highlands. Academic 


303

304 

WPA Author 

n.d. Ouachita County History by Works Project Administration Federal Writers 

Project. Ms. on file, Ouachita County Place File, Arkansas History Commission, Little 

Rock.

Appendix A 

site descriptiVe And AnAlyticAl dAtA 

This appendix presents a summary tabulation of field data and results of analyses 

for 144 archeological sites recorded or revisited by the 1979 Felsenthal Project. These 

data are drawn from Arkansas Archeological Site Survey Forms which run to seven or 

more page for each of the 144 sites, and entries therefore represent the briefest characterization 

and assessment of each site. 

Table A-1 summarizes 132 prehistoric sites in the order introduced in Chapter 5; 

that is, by seven drainage and geomorphic units, and from upstream to downstream 

locations within each unit. Table A-2 summarizes 15 historic sites in Unit 8, also in order 

from upstream to downstream location. Three sites having both prehistoric and historic 

components appear in both tables (3BR8, 3BR81, 3UN92). The following explanatory 

notes may assist the use of these tables (also see Chapter 5). 

TABLE A-1, PREHISTORIC SITES 

Column (1): Trinomial site designations assigned by Arkansas Archeological 

Survey; sites marked by asterisk were previously recorded and then revisited 

by the 1979 Felsenthal Project. 

Column (2): Site names from previous records or assigned by project except 

to single finds; some apparent clusters of sites indicated by a name and consecutive 

numbers, e.g., Jug Point 1, Jug Point 2. 

Column (3): For all buried sites, field observations and 1978 topographic 

maps were combined to provide the best estimate of elevation for upper 

and lower boundaries of cultural levels; elevations followed by the letter “I” 

indicate that cultural remains are known only to derive from some unknown 

level or levels within this interval; for all sites in Unit 7 map elevations are 

given for the surface extent, regardless of buried remains that may be present. 

Column (4): River channel mileages above mouths of Black or Saline rivers to 

nearest tenth as interpolated on USGS Felsenthal 15’ Ark-La (1939). 

Column (5): See discussions of landforms and channel geometry in Chapters 

2 and 5.

A-2 

Column (6): Dimensions of artifact scatters recorded in the field, rounded to 

nearest meter; for riverbank sites a width perpendicular to bankline is reported 

first, followed by a length parallel to bankline; no reliable dimensions 

were available for large complex sites previously recorded (Unit 7). 

Column (7): Depths of artifacts or cultural levels, measured in place from 

modern surface, rounded to nearest tenth of a meter. 

Column (8): Project numbers assigned in the field to artifacts scatters or single 

finds, later combined as sites in some cases. 

Columns (9)-(13): Totals for prehistoric artifact and debris categories, discussed 

in Chapters 6 and 8, and site totals. 

Column (14): Organic materials observed in newly recorded sites or indicated 

in existing records for previously recorded sites. 

Column (15): Best estimate of component or components represented (see 

Chapter 5). 

Column (16): Recommended status for nomination to the National Register of 

Historic Places, based on 1979 Felsenthal Project survey and testing; one site, 

3UN121 in Site Unit 4, is listed as “indeterminate” because it has been tested 

for significance subsequent to our work and results are not yet available (see 

Appendix C). 

TABLE A-2, HISTORIC SITES 

Column (1): Trinominal site designations assigned by Arkansas Archeological 

Survey; sites marked by asterisk were previously recorded and revisited by 

1979 Felsenthal Project. 

Column (2): Site names from previous records or assigned by project; historic 

placenames used where firm identifications made; e.g., Marie Saline Landing. 

Column (3): Elevations to nearest foot for surface extent of artifacts scatters 

and structures obtained from 1978 topographic maps; underwater remains 

indicated by elevations below 62.0 feet. 

Column (4): River channel mileage as in Table A-1. 

Column (5): Topographic locations as in Table A-1. 

Column (6): Dimensions of artifact scatters or, in some cases, standing structures 

recorded in the field, rounded to nearest meter.

Column (7): Project numbers assigned in the field to artifact scatters and 

structures, later combined as sites in some cases. 

Columns (8)-(12): Totals for historic artifact and debris categories and site 

totals. 

Column (13): Organic materials observed in field. 

Column (14): Historic components with best estimate of occupation or use 

on a decade-by-decade basis; more precise dates are available in a few cases 

(Chapter 5). 

Column (15): Recommended status for nomination to the National Register of 

Historic Places based on 1979 survey results ( (see Appendix C). 

A-3

A-4 

Table A-1. Prehistoric Sites. 

Unit 1: Ouachita River Above Mile 254 

(1) (2) (3) (4) (5) (6) (7) (8) (9) (10) (11) (12) (13) (14) (15) (16) 

State Topographic Linear Prehistoric Component(s) National 

Site Elevation Location or River Dimensions Depth Project Stone Inorganic Organic Identified (cultural Register 

Number Site Name (feet MSL) (river mileage) Channel Location (meters) (meters) Number(s) Ceramics Tools Debitage Debris Total Remains period/phase) Eligibility 

3UN167 Mud Lake Bend 1 68.0-76.01 270.1 right concave bank 1 x 5 093 0 1 4 3 8 Archaic Not eligible 

3UN166 Mud Lake Bend 2 71.1-72.4 269.9 right concave bank 1 x 50 1.1-1.5 094, 095 0 3 6 52 61 Archaic Eligible 

3UN165 Mud Lake Bend 3 63.0-76.01 269.7 right concave bank 1 x 100 096 0 0 0 7 7 Archaic Not eligible 

3UN164 Stormhole Bend 62.0-65.01 268.9 right concave bank 1 x 1 104 0 0 0 2 2 unknown Not eligible 

3BR71 Parrigeethe Shoals 1 63.0-70.01 268.2 left concave bank 1 x 2 105 0 0 0 21 21 Archaic Not eligible 

3BR33* Parrigeethe Shoals 2 63.0-75.01 268.1 left concave bank 1 x 25 106 0 1 0 6 7 unknown Not eligible 

3UN163 – 71.5 267.7 right convex bank 1 x 1 1.0 121 0 0 1 0 1 Archaic Not eligible 

3BR75 Eutaw Rapids 74.0 267.5 left concave bank 1 x 13 0.3 111 1 0 0 0 1 shell lens Baytown-Coles Creek Not eligible 

3UN162 – 65.3 267.3 right concave bank 1 x 1 3.0 123 0 0 0 1 1 Archaic Not eligible 

3BR73 Small Cane 2 71.0 266.2 left concave bank 2 x 23 1.2 109 0 1 1 10 12 late Archaic Eligible 

3BR74 – 72.0-75.01 266.1 left concave book 1 x 1 110 0 0 1 0 1 unknown Not eligible 

3BR72 Small Cane 1 70.0-73.51 266.0 left concave bank 2 x 35 107, 108 0 3 1 11 15 charcoal Archaic Eligible 

3UN161 Fletchers Landing 74.0 265.4 right concave bank 1 x 12 0.3 124 1 0 1 1 3 charcoal Baytown-Coles Creek Not eligible 

3BR70 Hunter’s Swan 70.0-73.01 265.0 left convex bank 1 x 1 091 8 0 0 0 8 Tchula/Coon Island Not eligible 

3BR57 Little Bay Rapids 63.0-70.01 261.8 left concave bank 1 x 6 012 0 0 1 1 2 unknown Not eligible 

3UN188 Five Mile Bend 62.0-70.01 257.2 right concave bank 1 x 5 007 5 0 0 1 6 Baytown-Coles Creek Not eligible 

3UN185 – 63.0-71.01 255.0 right concave bank 1 x 1 023 0 1 0 0 1 unknown Not eligible 

3BR77 Lower Saline Cutoff 67.0-70.07 254.3 left concave bank 2 x 42 128 0 0 1 4 5 unknown Not eligible

Table A-1 (continued). Prehistoric Sites. 


(1) (2) (3) (4) (5) (6) (7) (8) (9) (10) (11) (12) (13) (14) (15) (16) 




3BR92 Water Locust 65.0-67.0 9.9 left concave bank 2 x 105 0.8-1.1 233 0 0 4 37 41 Archaic Eligible 

3BR91 – 63.0-70.01 9.1 left concave bank 1 x 1 216 0 0 0 1 1 unknown Not eligible 

3BR90 Round Turn 63.0-66.01 8.3 right concave bank 1 x 40 212, 213 0 0 1 1 2 Archaic Not eligible 

3BR89 Mouth of Miller Creek 65.0-70.01 8.0 right concave bank 1 x 20 211 1 0 0 1 2 Baytown-Coles Creek Not eligible 

3BR88 – 67.0-70.01 7.8 left concave bank 1 x 1 150 0 1 0 0 1 unknown Not eligible 

3BR87 Horton Island Landing 68.0-70.01 7.3 right concave bank 2 x 19 149 8 0 2 5 15 Miss./Gran Marais Not eligible 

3BR86 Sweetgum 2 67.5 7.0 right concave bank 10 x 200 0.2-0.3 146, 147 15 2 3 32 52 Mississippi Eligible 

3BR85 Sweetgum 1 63.0-69.01 6.6 right convex bank 1 x 1 145 0 1 1 1 3 Archaic Not eligible 

3AS318 Hawthorn 3 63.0-70.01 6.3 left concave bank 2 x 24 144 0 0 2 2 4 unknown Not eligible 

3AS317 Hawthorn 2 63.5-64.5 6.0 left concave bank 2 x 10 1.7-1.9 142 0 1 0 7 7 Archaic Not eligible 

3AS316 Hawthorn 1 64.0-67.1 5.7 left concave bank 2 x 27 0.4-1.3 141 9 0 0 1 10 charcoal Baytown-Coles Creek Eligible 

3BR81 Prairie Island Landing 66.0-68.01 5.1 right concave bank 1 x 1 137 0 0 1 2 3 unknown Not eligible 

3BR80 Prairie Island Bend 67.0-68.0 5.0 right concave bank 2 x 40 0.2 136 1 0 2 13 16 Mississippi Not eligible 

3AS315 Persimmon 3 68.0-70.01 4.9 levee near left bank 3 x 90 135 0 2 24 29 55 Mississippi Eligible 

3BR82 Bitter Pecan 1 68.0-69.0 4.4 right convex bank 3 x 45 0.3 138 10 0 4 12 26 charcoal/bone Baytown-Coles Creek Eligible 

3AS314 Persimmon 2 63.0-70.01 4.4 left concave bank 17 x 20 134 0 2 2 20 24 Mississippi Not eligible 

3BR84 Bitter Pecan 3 63.0-69.01 4.4 levee near right bank 1 x 5 140 1 1 0 1 3 unknown Not eligible 

3BR83 Bitter Pecan 2 63.0-69.01 4.2 right straight bank 2 x 50 139 0 0 0 3 3 unknown Not eligible 

3AS313 Persimmon 1 63.0-66.01 4.1 left concave bank 3 x 75 133 0 2 7 4 13 Poverty Point/Calion Not eligible 

3BR78 Mouth of Eagle Creek 63.2-68.0 3.9 right concave bank 5 x 49 0.1-0.7 131 0 1 32 17* 50 charcoal Poverty Point/Calion Eligible (tested) 

3AS312 Willow Oak 66.0-70.01 3.7 left concave bank 1 x 2 130 0 1 0 1 2 unknown Not eligible 

3AS311 Tupelo Gum 64.0-68.01 2.9 levee near left bank 2 x 20 129 0 0 0 5 5 unknown Not eligible 

3AS309 Jug Point 4 63.0-68.01 2.7 left concave bank 3 x 20 117 0 0 4 0 4 unknown Not eligible 

3AS310 – 67.5 2.7 left concave bank 1 x 1 0.1 118 0 1 0 0 1 Miss./Terminal Not eligible 

3AS308 Jug Point 3 66.0-67.01 2.6 left convex bank 3 x 18 116 4 0 16 1 21 Miss./Caney Bayou Not eligible 

3AS307 Jug Point 2 67.2-68.3 2.6 left concave bank 8 x 80 0.1-0.3 115 98 8 66 14* 186 charcoal/bone Miss./Caney Bayou Eligible (tested) 

3AS306 Jug Point 1 67.2-68.1 2.5 left concave bank 9 x 75 0.1-0.3 114 62 23 677 121* 883 charcoal/bone Miss./Caney Bayou Eligible (tested) 

3BR76 Jug Point Cutoff 67.6-68.4 2.1 right concave bank 11 x 17 0.1-0.4 119 250 2 6 601* 859 midden Miss./Gran Marais Eligible (tested) 

3AS305 Overcup Oak 3 66.0-67.07 1.6 left concave bank 13 x 51 103 35 5 84 15 139 Miss./Caney Bayou Eligible 

3AS304 Overcup Oak 2 66.0-67.01 1.3 left convex bank 3 x 15 102 12 0 7 0 10 Miss./Gran Marais Not eligible 

3AS303 Overcup Oak 1 66.0-67.01 1.1 left convex bank 5 x 60 101 14 5 27 10 56 Miss./Gran Marais Not eligible 

3AS302 Sycamore 66.0-67.01 0.9 levee near left bank 1 x 2 100 0 0 1 1 2 unknown Not eligible 

3AS286 One Cypress Point 66.6-68.4 0.8 levee near left bank 16 x 58 0.1-0.2 024 84 178 1084 140* 1486 floral/bone Miss./Terminal Eligible (tested) 

3BR59 Possumhaw 64.0-66.01 0.8 levee near right bank 2 x 3 026 0 0 3 0 3 unknown Not eligible 

3BR58 Buttonbush 67.2-67.3 0.8 right convex bank 4 x 98 0.2 025 454 3 8 14* 479 Baytown-Coles Creek Not eligible 

(tested) 

3BR60 – 61.0-64.01 0.6 right straight bank 1 x 1 027 0 0 0 1 1 unknown Not eligible 

3BR61 – 63.0-64.01 0.4 levee near right bank 1 x 1 028 0 0 1 0 1 unknown Not eligible 

3BR62 Mouth of Long Slough 65.0-66.0 0.4 right convex bank 7 x 10 0.1-0.3 029 13 0 1 2 16 charcoal/bone Miss./Gran Marais Eligible 

3AS285 False Indigo 66.9-69.9 0.3 left straight bank 12 x 40 0.1-0.5 021 1663 9 97 4089* 4848 midden Tchula/Coon Island Eligible (tested) 

Miss./Gran Marais 

3AS284 Mouth of Saline River 63.0-67.01 0.0 left convex bank 2 x 10 019 0 0 1 4 5 unknown Not eligible 

A-5

A-6 



(1) (2) (3) (4) (5) (6) (7) (8) (9) (10) (11) (12) (13) (14) (15) (16) 




3BR68 – 66.0-68.01 – right concave bank 1 x 1 065 0 1 0 0 1 unknown Not eligible 

3BR67 – 66.0-68.01 – right concave hank 1 x 1 084 1 0 0 0 1 Mississippi Not eligible 

38866 Eagle Creek 4 66.0-68.01 – right convex bank 15 x 15 083 56 0 0 41 97 charcoal/bone Miss./Gran Marais Eligible 

3BR69 Eagle Creek 5 66.0-68.01 – left concave bank 5 x 50 086 4 1 1 0 6 Mississippi Not eligible 

3BR65 Eagle Creek 3 67.5 – right convex bank 15 x 80 0.1 058 16 0 15 29 60 charcoal Miss./Caney Bayou Eligible 

38864 Eagle Creek 2 63.0-69.01 – right concave bank 1 x 40 056, 057 0 0 2 0 2 unknown Not eligible 

38863 Eagle Creek 1 63.0-69.01 – right convex bank 3 x 125 055 5 1 9 7 21 Mississippi Not eligible 

3BR79 Eagle Creek 6 65.0 – left concave bank 3 x 20 0.1-0.3 132 0 3 26 12 41 Mississippi Eligible 

Unit 4: Ouachita River Below Mile 254 

3UN184 – 68.0-70.01 254.1 right concave bank 1 x 1 030 1 0 0 0 1 Baytown-Coles Creek Not eligible 

3UN92* Brown Camp below 71.01 252.9 levee near right bank 022 4 1 4 0 9 Mississippi Not eligible 

3AS329 Marie Saline 62.0-67.5 252.5 left concave bank 14 x 200 0.0-1.5 001 97 32 477 350• 956 charcoal Poverty Point/Calion Eligible 

Tchula/Coon Island (tested) 

Baytown-Coles Creek 

Miss./Caney Bayou 

3AS326 – 68.0-69.01 252.4 left concave bank 1 x 1 231 0 1 0 0 1 Miss./Gran Marais Not eligible 

3A5301 Honey Locust 66.0-69.01 252.0 left concave bank 2 x 20 090, 219 4 0 0 2 6 Baytown-Coles Creek Not eligible 

3AS319 – 66.0 251.9 left concave bank 1 x 1 0.3 220 0 0 0 1 1 unknown Not eligible 

3AS320 River Birch 1 63.0-67.01 251.5 left concave bank 1 x 3 222 0 0 0 4 4 unknown Not eligible 

3AS321 River Birch 2 63.0-67.01 251.4 left straight bank 2 x 40 223 8 1 1 1 11 Tchula/Coon Island Eligible 

3AS288 Red Oak 67.0-68.01 250.5 levee near left bank 2 x 4 032 0 0 0 2 2 unknown Not eligible 

3AS322 Water Elm 1 67.5 250.4 left concave bank 1 x 1 0.4 225 8 0 0 0 8 Mississippi/Caddo 2 Eligible 

3AS332 Water Elm 2 63.0-70.01 250.4 left concave bank 2 x 50 018 0 0 1 3 4 unknown Not eligible 

3A5287 Water Elm 3 67.0 250.2 left concave bank 3 x 5 0.1 031 277 0 1 4 302 Baytown-Coles Creek Eligible 

3AS323 Water Elm 4 68.0 250.0 left convex bank 1 x 1 0.5-0.6 226 52 0 0 0 52 Baytown-Coles Creek Not eligible 

3A5324 – 63.0 247.7 left convex bank 1 x 1 2.0 227 0 1 0 0 1 late Archaic Not eligible 

3UN123* Mouth of Lapile Creek 62.0-65.01 245.4 right concave bank 2 x 30 077 27 0 1 0 28 Baytown-Coles Creek Not eligible 

3AS327 Swamp Privet 66.0 245.0 levee near left bank 2 x 17 0.5 232 1 3 16 1 21 Baytown-Coles Creek Eligible 

3A5325 – 65.5 244.8 left concave bank 1 x 1 0.6 229 1 0 0 0 1 Baytown-Coles Creek Not eligible 

3UN169 – 64.0-65.01 244.4 levee near right bank 1 x 1 078 0 1 0 0 1 unknown Not eligible 

3UN168 Coon Glory Bend 63.0-65.01 244.2 levee near right bank 20 x 25 080 14 1 67 1 83 Mississippi/Gran Eligible 

3UN121* Mo-Pac 67.5 242.3 right concave bank 1 x 10 0.7-0.9 230 5 0 0 2 7 charcoal Baytown-Coles Creek Indeterminate



(1) (2) (3) (4) (5) (6) (7) (8) (9) (10) (11) (12) (13) (14) (15) (16) 




3UN178 Lapile Creek 1 68.0-72.01 – left concave bank 30 x 60 059 1 2 11 0 14 unknown Not eligible 

3UN160 Lapile Creek 2 68.0-72.01 – left concave bank 10 x 320 151-154, 175 1 3 16 2 22 Baytown-Coles Creek Not eligible 

3UN159 Lapile Creek 3 68.0-72.01 – opposite concave/ 10 x 160 155, 168-169 25 0 1 0 26 Baytown-Coles Creek Not eligible 

convex banks 

3UN158 Lapile Creek 4 68.0-71.01 – opposite concave/ 10 x 300 156-159, 170, 5 6 29 8 48 Baytown-Coles Creek, Not eligible 

convex banks 174 Mississippi 

3UN157 Lapile Creek 5 65.0-69.01 – opposite concave/ 10 x 165 160-161, 171 26 4 10 5 45 Baytown-Coles Creek Eligible 

– convex banks 

3UN156 Lapile Creek 6 65.0-68.01 – opposite concave/ 165 x 190 162-164, 172 0 1 24 15 40 unknown Not eligible 

convex banks 

3UN152 Lapile Creek 8 64.0-67.01 – right concave bank 1 x 3 173 0 2 1 0 3 Archaic Not eligible 

3UN151 Lapile Creek 7 64.0-67.01 – left concave bank 5 x 200 165-167 105 5 33 2 145 Archaic, Baytown- Eligible 

Coles Creek 


3UN187 St. Mary’s Lake North 69.5 – oxbow lake bank 5 x 40 0.4 013 0 0 4 5 9 unknown Not eligble 

3UN186 – 64.0-71.01 – oxbow lake bank 1 x 1 014 0 0 1 0 1 unknown Not eligible 

3UN170 Borrow Pit near 63.0-65.01 – back swamp 1 x 10 076 0 0 3 2 5 unknown Not eligible 

Lapoile Creek 

3UN183 Lapoile Creek 62.0-64.01 – opposite concave/ 1 x 2 046, 047 0 1 1 1 3 unknown Not eligible 

– convex banks 

3UNI82 Fishtrap Lake 1 63.0-65.01 – right concave bank 5 x 5 048 0 1 4 1 6 unknown Not eligible 

3UN180 Fishtrap Lake 2 63.0-66.01 – left concave bank 2 x 3 049 4 0 1 1 6 Baytown-Coles Creek Eligible 

3UN181 Fishtrap Lake 3 63.0-65.01 – right convex bank 1 x 100 050 0 0 0 3 3 unknown Not eligible 

3UN179 Fishtrap Lake 4 63.0-68.01 – left concave bank 2 x 2 051 7 0 0 1 8 Mississippi Eligible 

3AS295 Borrow Pit near 82 68.0-71.01 – backswamp/terrace 3 x 30 043, 044 4 2 4 2 12 Baytown-Coles Creek Not eligible 

Bridge 1 edge 

3AS294 Borrow Pit near 82 62.0-70.01 – backswamp/terrace 2 x 3 045 0 0 5 3 8 unknown Not eligible 

Bridge 2 edge 

3AS292 Wildcat Lake 62.0-63.01 – oxbow lake bank 5 x 15 019 0 0 4 10 14 unknown Not eligible 

A-7

A-8 

Table A-1 (concluded). Prehistoric Sites. 

Unit 7: Pleistocene Terraces and Islands 

(1) (2) (3) (4) (5) (6) (7) (8) (9) (10) (11) (12) (13) (14) (15) (16) 




3BR4* Eagle Lake Mounds 85.0-95.0 – Deweyville 3 Terrace 237 midden Mississippi (see Appen- 

3BR8* Goulett Island 63.0-86.0 – Deweyville 2 Terrace 217, 218 midden Archaic-Mississippi dix C). 

3UN81* Jones Lake North 66.0-90.0 – Deweyville 2/3 Terrace 081, 082 midden Archaic-Mississippi 

3AS296 Brushy Creek 1 73.0-74.0 – Deweyville 2 Terrace 3 x 10 072 0 0 5 1 6 unknown 

3AS297 Brushy Creek 2 67.0-70.0 – Deweyville 3 Terrace 3 x 200 073 1 0 12 1 13 Baytown-Coles Creek 

3AS298 Brushy Creek 3 68.0 – Deweyville 2/3 Terrace 3 x 12 074 0 0 3 4 7 unknown 

3UN18* WattsField 80.0-90.0 – Deweyville 2 Terrace 180 midden Archaic-Mississippi 

3UN8* Locust Ridge 90.0 – Deweyville 2 Terrace 179 midden Archaic-Mississippi 

3AS328 Greenbrier 115.0 – Deweyville 1 Terrace 60 x120 234 45 3 7 5 60 midden Miss./Gran Marais 

3AS289 Parker Ridge 70.0 – Deweyville 3 Terrace 15 x 150 034 9 3 20 16 48 Baytown-Coles Creek 

3AS160* Potlatch Field 69.0-74.0 – Deweyville 3 Terrace 037, 038 Archaic-Mississippi 

3AS293 Redeye Lake Prairie 73.0-80.0 – Deweyville 2/3 Terrace 125 x 200 041, 042 91 6 108 42 247 Miss./Gran Marais 

3UN72* Crossroad near Burnt 75.0-83.0 – Deweyville 2 Terrace 3 x 150 065 2 2 6 2 12 Archaic, Baytown- 

Bridge Coles Creek 

3UN177 Pine Ridge 1 85.0-86.0 – Deweyville 2 Terrace 3 x 150 061 1 3 7 1 12 Miss./Gran Marais 

3UN176 Pine Ridge 2 85.0 – Deweyville 2 Terrace 3 x 50 062 1 2 12 0 15 unknown 

3UN175 Bolding Road 80.0 – Deweyville 2 Terrace 3 x 10 064 0 1 0 1 2 late Archaic 

3UN9/52* Shallow Lake 85.0-87.0 – Deweyville 2 Terrace 054, 063, 067 midden Archaic-Mississippi 

3UN154 Shallow Lake West 85.0 – Deweyville 2 Terrace 5 x 20 177 0 1 2 0 3 Archaic 

3UN34* Pendleton Camp 76.0-82.0 – Deweyville 2 Terrace 052, 053 midden Archaic-Mississippi 

3AS291 Black Slough 67.0-68.0 – Deweyville 3 Terrace 3 x 5 036 0 1 0 2 3 Archaic 

3AS290 Pine Island Road 73.0-74.0 – Deweyville 3 Terrace 20 x 110 035 0 8 47 41 96 Archaic 

3UN173 Kelly Spring Creek 1 65.0-66.0 – Deweyville 3 Terrace 3 x 90 068, 075 2 2 41 2 47 Baytown-Coles Creek, 

3UN172 Kelly Spring Creek 2 73.0 – Deweyville 2 Terrace 3 x 10 069 1 0 3 1 5 Mississippi 

y unknown-Coles Creek 

3UN171 Gum Ridge Road 76.0 – Deweyville 2 Terrace 3 x 25 070 0 0 5 1 6 unknown 

3UN132* Burnie Creek 73.0-78.0 – Deweyville 2 Terrace 071 Archaic-Mississippi 

3AS1* Sulphur Springs Mound 96.0 – Prairie Terrace 236 Baytown-Coles Creek, 

Mississippi 

3A56* Big Mound Ridge 73.0-75.0 – Deweyville 3 Terrace 238 midden Archaic-Mississippi

Table A-2. Historic Sites. 

Unit 8: Historic Sites 

(1) (2) (3) (4) (5) (6) (7) (8) (9) (10) (11) (12) (13) (14) (15) 

State Linear Historic National 

Site Elevation Location River Dimensions Project Organic Component(s) Register 

Number Site Name (feet MSL) (river mileage) Channel Location (meters) Number(s) Ceramics Glass Metal Lithic Total Remains Identified Eligibility 

3BR32* Ganer Mound 75.0-87.0 267.9 levee near left bank 50 x 50 097 0 0 1 0 1 oyster shell American 1910s Not eligible 

3BR56 Caney Marie Landing 74.0-76.0 263.5 left concave bank 1 x 1 008 1 0 1 1 3 American late 1800s-early 1900s Eligible 

3BR8* Goulett Island 63.0-86.0 9.4 Deweyville 2 Terrace 217, 218 1 1 1 1 4 structures, American 1850s-present Eligible 

refuse 

3BR81 Prairie Island Landing 70.0 5.1 right concave bank 137 0 0 0 0 0 American 1920s-1930s Not eligible 

3BR55 Keelboat Brake Sawmill 66.0-68.0 – backswamp 004 0 0 1 0 1 charcoal American late 1800s-early 1900s Eligible 

3UN155 Ouachita Belle Landing 70.0 257.8 right concave bank 178 0 0 0 0 0 American 1870s Not eligible 

3UN88* Fletchers Landing and 63.0-75.0 265.7 right concave bank 40 x 60 125, 126 0 0 0 6 6 shell American 1870s-1890s Not eligible 

Woodyard 

3UN92* Brown Camp 71.0-90.0 252.9 levee near right bank 022 0 0 0 0 0 American 1920s-present Not eligible 

3AS330 First Slough 64.0-65.0 252.3 left concave bank 4 x 5 002 1 1 14 5 21 shell American late 1800s-early 1900s Not eligible 

3AS331 Marie Saline Ferry 62.0-68.0 252.2 left concave bank 2 x 4 003 1 0 0 0 1 American late 1800s-early 1900s Not eligible 

3AS300 South of Highway 82 Bridge 69.0-70.0 252.0 left concave bank 1 x 40 088, 089 1 1 0 0 2 shell American late 1800s-early 1900s Not eligible 

3AS299 Marie Saline Landing 61.0-69.0 251.9 left concave bank 5 x 200 087 0 0 0 0 0 American 1850s-early 1900s Eligible 

3UN174 Artesian Well 65.0 – backswamp 2 x 2 066 0 0 0 0 0 American early 1900s Not eligible 

3UN122* Grand Marais Landing 64.0-65.0 244.0 right concave bank 25 x 40 079 0 0 0 0 0 wood American late 1800s Eligible 

3UN153 Wreck of the Lotawanna 34.0-38.0 243.9 river bottom 176 0 0 0 0 0 American 1860s-1870s Eligible 

A-9

A-10 

Table A-3. Occurrence of Temporally or Culturally Diagnostic Artifacts in Prehistoric Floodplain Sites. 

Ceramic Complexes Projectile Points 

Other Distinctive 

Number Site Name Tchefuncte Baytown Gran Marais Shell tempered Other Ceramics Arrowpoints Dart Points Artifacts or Debris 

Unit l: Ouachita River Above Mile 254 

3UN167 Mud Lake Bend 1 1 ground stone tool fragment 

3UN166 Mud Lake Bend 2 1 heavy scraper or plane, 

36 fire-cracked rock 

3UN165 Mud Lake Bend 5 fire-cracked rock 

3UN164 Stormhole Bend 1 fire-cracked rock 

3BR71 Parrigeethe Shoals 1 12 fire-cracked rock 

3BR33 Parrigeethe Shoals 2 3 fire-cracked rock 

3UN163 – 

3BR75 Eutaw Rapids 1 Coles Creek 

Incised rim 

3UN162 – 1 fire-cracked rock 

3BR73 Small Cane 2 1 Macon 10 fire-cracked rock 

3BR74 – 

3BR72 Small Cane 1 1 Honey Creek 1 hafted scraper, 

1 hammerstone, 


3UN161 Fletchers Landing 1 Baytown Plain 1 fire-cracked rock 

body 

3BR70 Hunter’s Swan 2 Tchefuncte 

Stamped body 

3BR57 Little Bay Rapids 

3UN188 Five Mile Bend 3 Baytown Plain 1 untyped plain 1 untyped 1 fire-cracked rock 

body body incised rim 

3UN185 – 1 abrader 

3BR77 Lower Saline Cutoff 4 fire-cracked rock 


3BR92 Water Locust 35 fire-cracked rock 

3BR91 – 1 fire-cracked rock 

3BR90 Round Turn 1 fire-cracked rock 

3BR89 Mouth of Miller Creek 1 Baytown Plain 1 fire-cracked rock 

body 

3BR88 – 1 bifacial preform 

or chopper

Table A-3 (continued). Occurrence of Artifacts. 




3BR87 Horton Island Landing 2 Baytown Plain body 1 fire-cracked rock 

3BR86 Sweetgum 2 8 Baytown Plain body 1 Madison 10 fire-cracked rock 

3BR85 Sweetgum 1 1 hammerstone 

3AS318 Hawthorn 3 1 fire-cracked rock 

3AS317 Hawthorn 2 7 fire-cracked rock 

3AS316 Hawthorn l 8 Baytown Plain body 

3BR81 Prairie Island Landing 2 fire-cracked rock 

3BR80 Prairie Island Bend 1 untyped 3 fire-cracked rock 

incised body 

3AS315 Persimmon 3 1 untyped arrow- 1 arrowpoint preform, 

point fragment 26 fire-cracked rock 

3BR82 Bitter Pecan 1 5 Baytown Plain body, 9 fire-cracked rock 

4 Baytown Plain rim 

3AS314 Persimmon 2 1 untyped arrow- 17 fire-cracked rock 

point fragment 

3BR84 Bitter Pecan 3 1 hammerstone 

3BR83 Bitter Pecan 2 3 fire-cracked rock 

3AS313 Persimmon 1 1 gorget fragment, 

1 hammerstone, 


3BR78 Mouth of Eagle Creek 1 Gary 3 fire-cracked rock 

(see also Chapter 6) 

3AS312 Willow Oak 1 bifacial preform 

3AS311 Tupelo Gum 2 fire-cracked rock 

3AS309 Jug Point 4 

3AS310 – 1 Nodena 

3AS308 Jug Point 3 4 untyped plain 

body 

3AS307 Jug Point 2 25 Baytown Plain body 73 untyped plain 2 Ashley, 1 expanded base drill 

body 1 Bassett 4 fire-cracked rock 

(see also Chapter 6) 

3AS306 Jug Point 1 1 Baytown Plain body 58 untyped plain 2 Eagle Creek, 1 amber (?) bead, 

body 1 untyped 8 fire-cracked rock 

arrowpoint fragment (see also Chapter 6) 

3BR76 Jug Point Cutoff 171 Baytown Plain body, 13 untyped in- 45 untyped plain 2 untyped bone 1 Ashley 1 perforator, 

3 Baytown Plain rim cised or punctated body, 2 untyped tempered plain 2 fire-cracked rock 

body, 3 untyped plain rim, 10 un- body (see also Chapter 6) 

incised rim typed incised or 

punctated body 

A-11

A-12 





3AS305 Overcup Oak 3 1 Baytown Plain 4 untyped plain 1 untyped arrow- 1 graver 

body body point fragment 

3AS304 Overcup Oak 2 7 Baytown Plain 

body 

3AS303 Overcup Oak 1 5 Baytown Plain 1 untyped in- 1 hammerstone 

body cised body 6 fire-cracked rock 

3AS302 Sycamore 

3AS286 One Cypress Point 5 Baytown Plain 78 untyped Plain 14 Nodena, 1 Nodena 2 end scrapers 

body body, 5 untyped Banks var., 1 Madison, 1 cylindrical drill, 

incised body 1 Honey Creek, 1 un- 1 plummet 

typed arrowpoint frag. (see also Chapter 6) 

3BR59 Possumhaw 

3BR58 Buttonbush 454 Baytown Plain 1 Gary 5 fire-cracked rock 

body Small 

3BR60 – 

3BR61 – 

3BR62 Mouth of Long Slough 13 Baytown Plain 

body 

3AS285 False Indigo 20 Lake Borgne 965 Baytown Plain 106 untyped in- 38 untyped plain 1 untyped bone 2 Ashley, 1 pitted cobble, 

Incised body, body, 76 Baytown cised or punctated body, 2 untyped tempered plain 1 Alba 10 fire-cracked rock 

18 Tchefuncte Plain rim body, 48 untyped plain rim body (see also Chapter 6) 

Incised body, incised or punc- 

9 Tchefuncte tated rim (see 

Plain body also Chapter 8) 

3AS284 Mouth of Saline River 3 fire-cracked rock 


3BR68 – 

3BR67 – 1 Baytown Plain 

body 

3BR66 Eagle Creek 4 54 Baytown Plain 1 untyped punc- 1 fire-cracked rock 

body, 3 Baytown tated body, 1 

Plain rim untyped pinched 

body





3BR69 Eagle Creek 5 1 Baytown Plain body 1 untyped plain 1 hammerstone 

body 

3BR65 Eagle Creek 3 1 Tchefuncte 11 untyped plain 3 fire-cracked rock 

Stamped body body, 2 untyped 

plain rim 

3BR64 Eagle Creek 2 

3BR63 Eagle C:reek 1 1 Baytown Plain body 1 untyped plain 4 fire-cracked rock 

body 

3BR79 Eagle Creek 6 1 discoidal scraper, 


Unit 4: Ouachita River Below Mile 254 

A-13 

3UN184 – 1 Baytown Plain body 

3UN184 Brown Camp 2 Baytown Plain body 1 untyped in- 1 untyped plain 1 Washita, 1 Morhiss, 2 steatite sherds, 

cised body body 1 Madison 1 yarbrough, 1 schist slab, 216 

1 Gary, 1 Steuben, fire-cracked rock 

1 Langtry (see (see also Chapter 6) 

Chapter 8) 

3AS329 Marie Saline 14 Tchefuncte 66 Baytown Plain body 2 untyped plain 

Incised body, body 

15 untyped body 

3AS126 – 1 Ashley 

3AS301 Honey Locust 4 Baytown Plain body 

3AS319 – 1 fire-cracked rock 

3AS320 River Birch 1 3 fire-cracked rock 

3AS321 River Birch 2 1 Tchefuncte 1 hammerstone 

Stamped body, 

1 Tchefuncte 

Incised body, 6 

untyped body 

3AS288 Red Oak 2 fire-cracked rock 

3AS322 Water Elm 1 7 Sinner Linear 

Punctated body, 

1 sinner linear 

Punctated rim 

3AS332 Water Elm 2 2 fire-cracked rock 

3AS287 Water Elm 3 239 Baytown Plain 2 untyped inbody, 

34 Baytown cised body, 2 un- 

Plain rim typed incised rim 

3AS323 Water Elm 4 52 Baytown Plain body

A-14 





3AS324 – 1 Ellis 

3UN123 Mouth of Lapile 27 Baytown Plain 

Creek body 

3AS327 Swamp Privet 1 Baytown Plain 2 cobble anvils 

body (see Chapter 8) 

3AS325 – 1 Baytown Plain 

body 

3UN169 – 1 cobble mano 

3UN168 Coon Glory Bend 14 Baytown Plain 

body 

3UN121 Mo-Pac 4 Baytown Plain 

body, 1 Baytown 

Plain rim 


3UN178 Lapile Creek 1 1 Baytown Plain 1 bifacial preform 

body 

3UN160 Lapile Creek 2 1 Baytown Plain 1 untyped 

body stemmed point 

3UN159 Lapile Creek 3 25 Baytown Plain 

body 

3UN158 Lapile Creek 4 5 Baytown Plain 1 Madison 1 Gary 1 bifacial preform 

body 

3UN157 Lapile Creek 5 26 Baytown Plain 2 Gary 1 bifacial preform, 

body 1 hammerstone, 


3UN156 Lapile Creek 6 5 fire-cracked rock 

3UN152 Lapile Creek 8 1 hafted knife, 

1 bifacial preform 

3UN151 Lapile Creek 7 104 Baytown Plain 1 Carrollton, 1 fire-cracked rock 

body, 1 Baytown 1 Gary 

Plain rim 


3UN187 St. Mary’s Lake 2 fire-cracked rock 

north 

3UN186 – 

3UN170 Borrow Pit near 2 fire-cracked rock 

Lapoile Creek





3UN183 Lapoile Creek 

3UN182 Fishtrap Lake 1 

3UN180 Fishtrap Lake 2 4 Baytown Plain 

body 

3UN181 Fishtrap Lake 3 

3UN179 Fishtrap Lake 4 7 untyped plain 

body 

3AS295 Borrow Pit near 4 Baytown Plain 3 fire-cracked rock 

82 Bridge 1 body 

3AS294 Borrow Pit near 2 fire-cracked rock 

82 Bridge 2 

3AS292 Wildcat Lake 6 fire-cracked rock 

A-15

Appendix B 

fAunAl And florAl AnAlysis 

This appendix describes, primarily by means of tables, faunal and floral samples 

from floodplain sites investigated during site survey and testing work (Chapters 5, 6). 

Also, the occurrence of wood and subfossil wood from floodplain deposits or localities, 

not representing archeological contexts but relevant to paleoenvironmental study (Chapter 

2), is discussed briefly below. 

VertebrAte fAunA 

Table B-1 characterizes the small sample of vertebrate remains from eight prehistoric 

floodplain sites. Six of these sites were tested and appear in the order they are 

reported in Chapter 6. The other two samples represent the rare recovery of animal remains 

from site survey work done. None of the historic sites recorded in the project area 

produced vertebrate remains, although none has been tested presently. 

As a general observation, the extremely acid soils and annual flooding regime are 

not favorable to preservation of animal remains in these floodplain sites. The archeological 

specimens in Table B-1 are almost all calcined or burned to some extent; probably 

inorganic bone constituents are persisting in these specimens, with some loss or replacement 

of organic content. However, none of the bones or fragments examined actually 

appear mineralized. Two sites of particular interest, False Indigo (3AS285) and Jug Point 

Cutoff (3BR76), have midden deposits or anthropogenic soils with medium acid to neutral 

pH values (Chapter 7, Table 19), and some unburned bone may be preserved in these 

deposits (especially 3BR76). Unfortunately, the stratified multiple component site, Marie 

Saline (3AS329), tested extensively to 1.5 m depth, is essentially devoid of animal bone 

so far as we can tell. 

Faunal samples were examined under a binocular microscope and identifications 

made by the project archeologist. Typically, the specimens recovered are minute, fragmentary, 

and calcined, so that specific identifications are impossible, and even broad 

classification (mammal, bird, turtle, fish, and so on) was often difficult. Among a total 

of 10 2-liter soil samples from occupation levels or features washed through graded fine 

screens (1/16-inch or .16 mm mesh was the finest used), only samples from midden soil 

at 3AS285 and 3BR76 were truly productive (Table B-1). In these Mississippi period sites 

on the lower Saline River (Chapter 6), fishing activity and fish processing are indicated 

by an abundance of spines, scales, teeth, otoliths, and other elements. Only drumfish and 

gar have been identified, but a dozen otoliths from 3BR76, potentially identifiable, are 

likely to represent additional species in the 5-30 kg size range (Chapter 2, Table 3).

B-2 

Table B-1. Vertebrate Remains from Prehistoric Floodplain Sites. 

Site Number Site Name Sample Number Provenience Faunal Content Taxa Identified Recovery Technique 

3AS285 False indigo 1433-1-2-6 Test Pit 1, 0-10 cm 4 small calcined bone fragments dry screened 

1433-1-5-7 Test Pit 1, 10-20 cm 4 small calcined bone fragments dry screened 

1433-14-2 Test Pit 15, 0-10 cm many small burned or calcined bone Aplodinotus wet screened 

fragments, including fish spines; 1 fish grunniens 

vertebra; 1 drumfish tooth 

1433-14-7 Test Pit 15, Feature 2 several small burned or calcined bone wet screened 

fragments; 1 fish spine; 1 mammal? 

bone fragment 

1433-15-4 Shovel Test 1, 0-30 cm trace of calcined bone dry screened 

718-4-3 Grid Square 2, surface 3 small calcined bone fragments controlled surface collection 

1433-13-1-2 Grid Square 14, surface 1 calcined mammal? bone fragment controlled surface collection 

3BR58 Buttonbush 1435-1-1-2 Test Pit 1, 0-10 cm trace of calcined bone dry screened 

3A5286 One Cypress Point 1434-18-3-6 Test Pit 7, 10-20 cm 2 small calcined bone fragments dry screened 

3BR76 Jug Point Cutoff 1439-16-5-6 Test Pit 1, 10-20 cm 1 fish otolith dry screened 

1439-16-2-5 Test Pit 1, 10-20 cm many small burned or calcined bone Aplodinotus wet screened 

fragments, including fish spines and grunniens, 

vertebrae, drum teeth, gar scales Lepisosteus sp. 

1439-16-2-5 Test Pit 1, 15-20 cm many small burned or calcined bone wet screened 

fragments, including fish spines and 

vertebrae; 3 fish otoliths 

1439-16-2-6 Test Pit 1, 20-30 cm 1 fish otolith; 1 fish vertebra dry screened 

1439-16-4 Test Pit 1, 20-30 cm 3 fish otoliths; 3 fish vertebrae; 4 dry screened 

large bird? limb bone fragments; 2 

mammal? bone fragments; ca 8 sections 

of animal coprolites? 

1439-16-3 Test Pit 1, depth unknown 1 fragmentary large animal incisor hand picked 

tooth

Table B-1 (concluded). Vertebrate Remains. 

Site Number Site Name Sample Number Provenience Faunal Content Taxa Identified Recovery Technique 

1439-5-4 Shovel Test 5, 13-30 cm 1 fish otolith, many small calcined dry screened 

bone fragments 

1439-5-4-3-2 Shovel Test 5, 13-30 cm many small burned or calcined bone Lepisosteus sp. wet screened 

fragments including fish spines and 

vertebrae, gar scales; 2 fish otoliths; 

1 animal coprolite or dropping? 

1439-13-5 Shovel Test 16, 13-30 cm 2 fish otoliths; 2 fish spines; 1 small dry screened 

animal long bone fragment 

1439-17-4 Surface 1 charred, fragmentary, small animal surface collection 

bone 

3AS306 Jug Point 1 1436-20-1-7 Test Pit 2, 10-20 cm trace of calcined bone dry screened 

1436-20-3-5 Test Pit 2, Feature 1 ca. 200 small burned or calcined bone dry screened 

fragments, including large or medium 

sized mammal 

1436-20-5-5 Test Pit 2, Feature 1 2 small calcined bone fragments wet screened 

1436-1-1-6 Surface 1 calcined turtle shell fragment surface collection 

3AS307 Jug Point 2 1437-7-2-6 Test Pit 1, 10-20 cm ca 12 small calcined bone fragments dry screened 

3BR62 Mouth of Long Slough 1270-4 Surface 4 calcined turtle shell fragments surface collection 

3BR66 Eagle Creek 4 1379-4 Surface 2 calcined mammal? bone fragment surface collection 

B-3

B-4 

Among all prehistoric floodplain sites, these two Mississippi period components have 

the greatest potential for intensive faunal analysis (Appendix C). 

By way of contrast, the faunal remains from Jug Point 1 (3AS306) include mammal 

bone in fragmentary, calcined condition; no fish elements have been identified in dry or 

wet screen samples (Table B-1). On present evidence, this site must be associated with 

hunting activity and fall occupation (Chapter 6). 

MolluscAn fAunA 

Table B-2 presents rather limited archeological data for mollusc remains in both 

historic and prehistoric floodplain sites. The modern sample consists of about 110 mussels, 

usually pairs of valves, collected as dead shells along the shorelines of the Ouachita 

and Saline rivers and in backswamp areas during 1979 site survey work. These most 

likely represent animals stranded by lowered river levels earlier in that year, or recent 

death through some other cause. The dredged sample in Table B-2 represents another six 

mussels collected from dredged riverbottom sediment at a pipeline crossing (mile 262.5) 

on the Ouachita River. These animals may have died as the result of dredging here in 

about 1974-1975. Both modern comparative samples are deposited in the University of 

Arkansas Museum collection (accession numbers 80-99 and 80-100). 

Identification of modern and archeological molluscs in Table B-2 was made by 

Mark E. Gordon, Department of Zoology, University of Arkansas, and the classification 

follows a recent checklist of Unionacea for this and other regions in Arkansas (Gordon et 

al. 1979). 

As in the case of animal bone, mollusc shell is not likely to be preserved in floodplain 

sites. Specimens collected here are usually fragmentary or eroded, and may be 

calcined. Archeological samples in Table B-2 consist of 4-30 valves in the case of historic 

sites 3UN88 and 3AS330, and 1-2 valves plus fragments in the case of prehistoric sites 

3BR75 and 3BR76. Unidentifiable traces of mussel and snail shell were also noted in deep 

excavated levels at Marie Saline (3AS329), including Test Pit 5, 100-110 cm, and Test Pit 

7, 80-90 cm. 

Four species of mussels have been identified in prehistoric site samples (Table B-2) 

and presumably represent supplementary food resources (Parmalee and Klippel 1974). 

Five species of mussels have been identified in historic sites samples and may have been 

eaten or used as a bait source. No mussel shell or “pearl” button industry is indicated by 

historic floodplain sites in this area. All of the archeological species are known or expected 

in the region today, although the modern molluscan fauna of the lower Ouachita 

River and the Saline River in Arkansas remains poorly known (Gordon et al. 1979).

Table B-2. Molluscs Identified in Prehistoric and Historic Site Samples and a Modern 

Sample from Felsenthal National Wildlife Refuge. 

Modern Dredged 

Sample Sample Historic Sites Prehistoric Sites 

Molluscan Species (1979) (1974-1975) 3UN88 3AS330 3BR75 3BR76 

Unionidae 

Fusconaia flava x x 

Fusconaia ebena x x x x x 

Amblema plicata x x x x x 

Quadrula pustulosa x x x? 

Quadrula quadrula x 

Quadrula nodulata 

Quadrula metanerva x 

Quadrula cylindrica x 

Tritogonia verrucosa x 

Plectomerus dombeyanus x 

Pleurobema clava x x? 

Anodonta grandis x 

Anodonta imbecilis x 

Ptychobranchus occidentalis x 

Cyprogenia aberti 

Obovaria jacksoniana x 

Truncilla truncata 

Leptodea fragilis x 

Proptera purpurata x 

Carunculina texasensis x 

Ligumia subrostrata x 

Lampsilis anodontoides x 

Lampsilis hydiana x x? 

Lampsilis ovata 

Corbicula cf. fluminea x 

Viviparidae 

Viviparus intertextus x 

Campeloma subsolidum x 

Planorbidae 

Helisoma trivolvis x 

B-5

B-6 

cArboniZed plAnt reMAins 

Identifications and other data for charcoal and carbonized seeds from prehistoric 

floodplain sites are presented in Table B-3. Charcoal specimens were prepared and 

examined under a binocular microscope by David W. Stahle, dendrochronologist and 

Research Assistant at Arkansas Archeological Survey. One historic wood specimen from 

Grand Marais Landing (3UN122) was similarly examined and is reported in Table B-3. 

Charcoal or wood taxa were identifed by reference to a wood characteristics manual 

(Panshin and de Zeeuw 1970). About two dozen charcoal samples from floodplain sites, 

not tabulated, were too small or too poorly preserved for identification of taxa. The few 

carbonized seeds recovered were identified by the project archeologist using a seed manual 

(U.S. Department of Agriculture 1948) and modern specimens collected in the project 

area. Surprisingly, charred nutshell occurred only in trace amounts in these samples, and 

prehistoric use of floodplain nut crops remains an open question. 

The prehistoric charcoal specimens, presumably representing fuelwood, could, 

for the most part, be derived from bottomland hardwood and swamp forest species 

found in the project area today (Chapter 2, Table 1). Red oak species include cherrybark, 

Nuttall, water, and willow oak; the only white oak of this overflow bottomland noted 

during site survey was overcup. Bald cypress is present in channels, sloughs, and lakes 

throughout the area. Pine charcoal and pitch or tar, tentatively identified in a midden 

sample from False Indigo (3AS285), may represent wood introduced from an upland or 

terrace location, since no pines occur in the bottomland forest as the result of overflow 

conditions. This introduction could be intentional or could merely represent use of pine 

occurring as driftwood. At Gran Marais Landing (3UN122), a split, worked section of 

pine most likely represents nineteenth century activity at this riverboat landing. 

The few carbonized seeds of honey locust and persimmon recovered at One Cypress 

Point (3AS286), Jug Point 1 (3AS306), and Jug Point 2 (3AS307) most likely represent 

foodstuffs in use on these sites of the Late Mississippi period (Chapter 6). Both trees 

grow today on the same natural levees of the Saline River where these sites are located. 

Their seedpods and fruits ripen in fall, and the carbonized remains should be an indicator 

of fall occupation, especially where overflow may extend through winter and spring 

(Chapter 9, Figure 70). 

subfossil wood 

Table B-4 summarizes characteristics of several wood samples from floodplain 

deposits or nonarcheological localities of environmental interest. These specimens were 

prepared, examined, and identified by David W. Stahle in conjunction with study of 

archeological charcoal specimens. They are included here to show that organic material 

can persist in floodplain sediments under continuously saturated conditions. Such specimens 

are of paleoenvironmental importance if they can be identified and dated by

Table B-3. Floral Remains from Prehistoric and Historic Floodplain Sites. 

Site Other Carbonized 

Number Site Name Sample Number Provenience Sample Description Charcoal or Wood Taxa Floral Material Remarks 

3AS329 Marie Saline 1415-2-9-2 Test Pit 2, 114 cm several small woody fragments red oak group — ring porous species 

1415-10-8-5 Test Pit 5, 100-110 cm many small woody fragments cf. baldcypress — diffuse porous species 

1415-11-8-5 Test Pit 6, 90-100 cm many small woody fragments, red oak group — Feature1 (see Chapter 

(wet screened) trace of nutshell? 6) 

3AS285 False Indigo 1433-4-3 Test Pit 15, 10-18 cm one small woody fragment cf. pine pitch or tar? Feature 1 (see Chapter 

adhering 6) 

1433-15-3 Shovel Test 1, 0-30 cm several small woody fragments cf. bald cypress or — diffuse and ring porous 

pine and red oak group species 

1433-21-5 Shovel Test 8, 0-20 cm several small woody fragments red oak group — ring porous species 

3AS286 One Cypress Point 1434-18-3-5 Test Pit 7, 10-20 cm one seed ca. 2 x 6 x 14 mm — honey locust (see Chapter 6) 

3BR76 Jug Point Cutoff 1439-8-3 Shovel Test 8, 0-50 cm many small woody fragments red oak group — 

1439-16-2-5 Test Pit 1, 10-20 cm several small woody fragments cf. baldcypress — associated with midden 

(wet screened) (see Chapter 6) 

3AS306 Jug Point 1 1436-20-1-6 Test Pit 2, 10-20 cm 2 small woody fragments, 1 seed, red oak group honeylocust 

other seed fragments 

1436-20-3-4 Test Pit 2, 19-30 cm several small woody fragments, white oak group honey locust, Feature 1 

ca. 3 seeds, trace of nutshell? persimmon (see Chapter 6) 

1436-20-5-5 Test Pit 2, Feature 1 fill 2 seeds ca. 2 x 8 x 12mm — honey locust (see Chapter 6) 

(wet screened) 

3AS307 Jug Point 2 1437-7-2-5 Test Pit 1, 10-20 cm 5 seeds ca. 2 x 8 x 12mm — honey locust (see Chapter 6) 

3UN122 Grand Marais 1372-3 unnumbered split section of wood pine — (not charred) 

Landing shovel test, 0-40 cm ca. 3 x 3 x 37 cm 

Note: All specimens listed above are carbonized except sample 1372-3. Samples recovered by dry screening except as indicated. 

B-7

B-8 

Table B-4. Wood and Subfossil Wood Specimens from Environmental Localities. 

Project 

Description of Locality Sample Number Description of Specimens Wood Taxa Identified 

Eroded cypress butts, 295-120 2 sections of light, dry wood; bald cypress 

Ouachita River bankline 3 x 6 x 26 cm and 3 x 7 x 53 cm; 

at mile 266.6 sensitive annual ring series may be 

(see Figure 7a) present; diffuse porous 

Dredged spoil from — 1 dark-colored, wood section; white oak group 

pipeline crossing, peeling in thin sheets along rays; 

Ouachita River bottom 4 x 6 x 54 cm; ring porous 

at mile 262.5 

— 1 dark-colored wood section, white oak group 

2 x 9 x 22 cm; slightly distorted 

ring porous 

— 1 dark-colored, hard, wood section; white oak group 

6 x 7 x 11 cm; highly distorted; 

ring porous 

Excavated spoil from new — 6 x 9 x 10 cm; undistorted; about 

Ouachita River floodplain 60 rings present; ring porous 

at miles 242-243 

— 1 dark-colored, wood section, cf. white oak group 

3 x 4 x 6 cm, distorted, hard, and 

brittle; ring porous

dendrochronological or radiocarbon methods. In this group of samples white oak predominates, 

in contrast to results of charcoal identification (Table B-3). The cause of this 

prevalence is not known, and too few specimens have been collected presently. All of the 

white oak specimens are discolored, distorted, or otherwise altered by burial and subsequent 

exposure. 

references cited 

Gordon, M. E., L. Russert Kraemer, and A. V. Brown 

1979 Unionaceae of Arkansas: Historical Review, Checklist, and Observations on 

Distributional Patterns. Bulletin of the American Malecological Union 1979:31-37. 

Panshin, A. J., and Carl de Zeeuw 

1970 Textbook of Wood Technology. Volume 1 (third ed.). McGraw-Hill, New York. 

U.S. Department of Agriculture 

1948 Woody-plant Seed Manual. Miscellaneous Publication 654. Washington, D.C. 

B-9

Appendix C 

recoMMended progrAM for MitigAtion of 

iMpActs on ArcheologicAl resources in the 

felsenthAl nAVigAtion pool 


The Felsenthal Navigation Project survey, carried out by the Arkansas Archeological 

Survey during the last half of 1979, recorded 126 previously unknown archeological 

sites and revisited 18 others known from prior work. These 144 sites are summarized in 

Chapter 5 and Appendix A, both of which are organized as eight “site units.” Seven of 

these site units include prehistoric and historic sites within the floodplain of the Ouachita 

and Saline rivers; the other site unit (Pleistocene Terraces and Islands) includes 27 

sites above the floodplain. 

In the mitigation proposal presented here we address all known floodplain sites 

(recorded or revisited), noting their relationship to the 65-foot contour (or prospective 

navigation pool elevation). The reader should refer back to the discussion of the 65-foot 

contour (as a constraint on survey area) in Chapter 1 and to the narrative of methodology 

presented in Chapter 5. The point we emphasize here is that a number of floodplain 

sites evaluated and recommended for mitigation in this appendix lie partly or wholly 

above the 65-foot contour. Also our appraisal of project impacts is not limited strictly to 

the new navigation pool, but includes floodplain areas above 65 feet elevation. In dealing 

with all known floodplain sites and floodplain site impacts, we have departed from 

the project Scope of Services in certain respects which have been the subject of previous 

discussion with the U.S. Army Corps of Engineers, Vicksburg District. 

Previous site data and maps from the Felsenthal Project were forwarded to the 

Vicksburg District in 1980 as an aid to their planning process (E. T. Hemmings, letter to 

Shelia Lewis, 5 August 1980). We were cautious about the use of site data in advance of 

completed analyses and full discussion in a final report, stating some of the limitations 

as follows: 

Most, but not all, of the 740 sites [now 144 sites] plotted on this map 

are within the floodplain project area. Other sites at higher elevations were 

recorded or revisited when survey crews were entering and exiting the floodplain. 

This map represents the 1979 work of our project and does not begin to 

include all known sites at higher elevations under Fish and Wildlife jurisdiction. 

It is complete through 1979 for the floodplain, given the limitations of

C-2 

survey in that environment to be discussed and evaluated in our report to 

you. It cannot, of course, be used as an archeological site map of the entire 

Felsenthal National Wildlife Refuge. Lastly, avoidance of buried archeological 

sites in the floodplain cannot be assured by use of this map, as for example in 

constructing bendways and boat channels. 

Having completed all analyses of site data and site significance, we think it is advisable 

to state these limitations of project results once again in succinct form, as an aid to management 

decisions: 

1. All prehistoric sites, and probably all or parts of some historic sites, are buried 

in the Ouachita and Saline River floodplain study area. 

2. No agricultural development has taken place in this floodplain area, which 

is an overflow bottomland, and no sites have been, or can be, discovered as 

the result of such activity (Chapter 5). 

3. Subsurface survey methods in the bottomland hardwood and swamp forest 

environments, constituting about 85% of the floodplain area, are deficient, 

and buried cultural resources are likely to remain undiscovered (Chapter 5). 

4. Underwater survey methods were not employed in the river channel, open 

lake, and bayou environments which constitute about 15% of the floodplain 

area; underwater cultural resources are likely to remain undiscovered (the 

steamboat Lotawanna is a notable exception—see Chapter 4, 5). 

5. No large scale, systematic, cultural resource survey has been carried out on 

floodplain terrain over 65 feet elevation; this report does identify resources 

above 65 feet in certain transect and levee locations, and more importantly, 

concludes that a markedly low density of sites can be expected to occur in 

the extensive backswamp zones of Ouachita and Saline rivers. 

6. Site specific surveys of certain floodplain areas (new Felsenthal Lock and 

Dam construction area and new Crossett Harbor Facility area, among others) 

are, we believe, now scheduled; the results of these surveys are not known 

to us and are not considered in the report. 

7. The results of the 1979 Felsenthal Project deal briefly with sites above 

floodplain level, but no systematic survey of such terrain was required or 

performed by us; recommendations of a general nature for management of 

nonfloodplain resources in the Felsenthal National Wildlife Refuge are presented 

at the end of this appendix. 

In addition to site data and maps forwarded to the Corps of Engineers in 1980, 

some specific recommendations with regard to sites endangered or possibly endangered

y construction were also furnished by Arkansas Archeological Survey. These recommendations 

were based on preliminary assessment of site significance and project 

impacts, and they do not differ from final assessments presented in this appendix except 

in the matter of distinction we make between terrain above and below 65 feet elevation. 

The essential portions of these recommendations are reproduced here, as they continue 

to be relevant to the mitigation proposal presented below. 

the following information and recommendations should be forwarded 

to you prior to completion of our contract. One previously recorded site, 

Grand Marais Landing (3UN122) and one newly recorded site, Wreck of the 

Lotawanna (3UN153), are located within or near the upstream end of Felsenthal 

Lock and Dam channel construction....As far as we can determine now, 

both sites will be adversely affected. Also, based on present knowledge both 

sites have regional historical significance and are potentially eligible for 

nomination to the National Register of Historic Places. However, the determination 

of eligibility here should include systematic information on the extent 

and integrity of historic remains, in one case below ground and in the other 

underwater. Clearly, this information should be obtained before construction 

proceeds into the area of the landing or steamboat wreck. 

We recommend that the Corps arrange to have a qualified diver-archeologist 

inspect the underwater Lotawanna wreck site, pinpoint the location 

and extent, and assess the integrity of remains. 

We also recommend limited subsurface testing of the riverbank area 

recorded as Grand Marais Landing (3UN122), including the problematical 

physical features examined and recorded by us in 1976 and 1979, as well as 

sites of structures nearby indicated only by nineteenth century river maps (C. 

R. McGimsey III, letter to Shelia Lewis, 29 August 1980). 

...actual construction rights-of-way and impacts are not known to us in detail. 

I have in hand the Crossett Harbor Feasibility Report (September 1979) and 

a Lock and Dam construction map (April 1976). On the basis of these, six 

sites on record are probably subject to direct or indirect impacts of construction. 

Those sites are 3AS159, 3AS299, 3AS300, 3AS301, and 3AS319 at or 

near Crossett Harbor and 3UN121 at the closure dam....Both Crossett Harbor 

and the closure dam will largely impact terrain above 65 feet which has not 

been intensively surveyed by us or anyone else. With regard to the Crossett 

Harbor area,...Hester Davis, Arkansas State Archeologist,...noted that buried 

sites would be expected to fall in this area and that a more thorough survey 

with shovel testing should be carried out. We continue to hold this view, and 

moreover we call attention now to the probability that significant historic 

remains of Marie Saline Landing (3AS299) are likely to be present and undiscovered 

in terrain areas above 65 feet. The 200-meter long canal at Marie 

Saline Landing was recorded in detail by our survey team because it lies at 

and below 65 feet elevation. We know that obtrusive historic remains are not 

present on the surface near the canal, but buried remains may be present. 

C-3

C-4 

Also the canal itself may have a submerged bottom deposit with artifacts 

pertaining to the 19th century construction and use. More work is certainly 

needed in this vicinity. See also the site records for four other sites that may 

be affected. 

With regard to the closure dam area, the one site on record here, 

3UN121, is a buried prehistoric site of apparent small size, within the rightof-way 

according to my maps. This site lies above the 65 foot contour and was 

not tested by our survey team. Subsurface testing should be carried out here 

prior to construction (E. T. Hemmings, letter to Joyce Williams, 10 April 1981). 

All of the floodplain sites mentioned in these excerpts have now been evaluated 

on the basis of all available project information. Recommendations for certain of these 

which appear below simply build upon these earlier appraisals. Also, we would note 

briefly mitigative actions by the Corps which have been taken recently with regard to 

two of the sites of concern: 

1. Wreck of the Lotawanna (3UN153) on the Ouachita River bottom has 

probably been pinpointed by magnetometry and bathymetry, as noted in 

Chapter 5; additional recommendations appear below. 

2. Mo-Pac site (3UN121) in the closure dam right-of-way has now been 

tested by Historic Preservation Associates of Fayetteville; no additional 

recommendations appear below. 

significAnce of floodplAin sites 

The concept of significance in cultural resource management has been scrutinized 

and debated for over a decade. Several dimensions of significance have been proposed 

and discussed, among which scientific significance, or the quality of research potential, 

is most relevant to our evaluation of floodplain sites (Schiffer and Gumerman 1977:241; 

Advisory Council on Historic Preservation 1980:6). Although this concept is very useful, 

it must be adapted to the particular cultural-environmental context of the Felsenthal 

Project area. In contrast to many other archeological localities and regions in North 

America, our floodplain sites are: 

1. markedly small in size or extent 

2. often markedly sparse in content of artifacts and debris 

3. buried by overbank silty clays 

4. never plowed, and probably never dug or collected, prior to our work in 

1979, and 

5. set in an overflow bottomland and annually overflowed by several meters 

of water. 

We have dealt with the research potential of small sites in Chapter 6, and again in 

Chapter 9. As Schiffer and Gumerman (1977:242) note:

there is no a priori reason to think that [information potential] would correlate 

directly with site size....With the advent of settlement system analysis 

and other approaches in modern archeology the smallest site in a region may 

be as important or more so than the largest.... 

The summary assessment of eight prehistoric floodplain sites, tested extensively by 

us, indicates that specific relevant site data were adduced in all the following domains 

(Chapter 6, Table 16): 

1. cultural and technological change 

2. interregional contact 

3. paleoecology 

4. procurement technology 

5. specialized procurement technology (focus on one or a few target resources) 

6. settlement/subsistence systems 

7. trade 

Scientific Significance and Research Potential 

Using this list as a starting point, and considering now all known floodplain sites 

as a data base, we can suggest also a series of more specific research problems of current 

methodological and theoretical interest, but oriented to representative samples of specialized 

extractive sites and transient camps: 

1. energetics of an overflow bottomland environment in relation to human 

adaptations (prehistoric and historic) 

2. formation and transformation processes of alluvial sites in an overflow 

bottomland setting 

3. recent geohydrological history of the Grand Marais Lowland 

4. geochronological problems and chronometric potential of alluvial sites in 

an overflow bottomland setting 

5. archeological context and systemic context in specialized extractive sites 

(prehistoric) 

6. material correlates of intrasite extractive tasks and maintenance tasks (prehistoric) 

7. social or ethnic units which operated extractive sites (and variability 

among these units) 

8. variability in discard and curate behavior among prehistoric floodplain 

site assemblages 

9. locational analysis of riverbank/levee extractive sites or transient camps, 

and of historic landings and stations 

10. catchment analysis of specialized extractive sites (prehistoric) 

11. fisheries ecology, seasonality, fish procurement, and fish processing in the 

Grand Marais Lowland (aboriginal and Euramerican) 

12. structure and function of a nineteenth century riverboat landing (especially 

Marie Saline Landing) 

C-5

C-6 

13. site discovery strategies and investigation methodologies in an overflow 

bottomland environment (including remote sensing) 

14. small site methodologies 

15. lithic behavioral systems in specialized extractive sites (prehistoric) 

16. ceramic behavioral systems in specialized extractive sites (especially in 

regard to fishing and fish processing) 

17. Baytown ceramic technology and variability 

18. Grand Marais ceramic styles as sociobehavioral correlates 

19. submerged deposits and preservation of saturated materials (prehistoric 

and historic) 

20. impacts of artificially modified and controlled inundation on cultural resources 

of the Grand Marais Lowland 

This latter list emphasizes, and even dramatizes, the broad range of relevant data 

and current problems which we associate with the range of floodplain sites identified, 

and analyzed in Chapters 5 and 6 (see also Appendix A). Floodplain sites, in our view, 

can and do have qualities of anthropological and historical significance which supplement 

or exceed regional research problems; we will also refer to public significance later 

in this chapter. Many archeologists and cultural resource managers will know, as we 

do, that this list is incomplete. Also, it is important to note that the list of specific research 

problems emphasizes what we perceive as the “particular cultural-environmental 

context” of the Felsenthal Project area. Finally, the quality of research potential among 

known floodplain sites ranges from very little to very great. 

We nevertheless have attempted rigorous, if difficult, identification of significant 

floodplain sites which warrant mitigation recommendations of one kind or another. As 

Schiffer and House (1977:252) noted in the case of Cache Basin cultural resources “we 

cannot now, or perhaps ever, assess the substantive significance [of these resources] mechanically.” 

Our process of assessment involved review of 117 floodplain sites (plus one 

upland site, Goulett Island, 3BR8, which has extensive refuse downslope to floodplain 

level). This process featured: 

1. evaluation of all site survey, test excavation, and analytical data contained in 

relevant records and in this report 

2. review of the 1979 project research design (Appendix D), the research design 

outline for proposed mitigation presented below, the Arkansas State Plan, Volume 

II, and other regional or general sources, and 

3. consultation among the Project Archeologist, Principal Investigator, Project 

Historian, Station Archeologists, and other staff members at the Arkansas Archeological 

Survey. 

We now identify 37 floodplain sites as scientifically significant, and among these 

some are also anthropologically or historically significant. One site, Buttonbush (3BR58),

has been tested extensively, and because of small size and sparse content, is not believed 

to contain additional cultural data (Chapter 6). We therefore state in Table 16 for Buttonbush: 

“no additional work [is needed]; mitigated by testing.” (It follows that this site 

is not eligible for the National Register by reason of its removal.) A second site, Mo-Pac 

(3UN121), was recommended for testing, and has in fact now been tested for significance 

by Historic Preservation Associates (see comments at beginning of this appendix). In this 

case we have no additional data, and cannot properly comment further on significance, 

eligibility, or requirements for mitigation. We are therefore recommending 35 (of 37 significant 

floodplain sites identified) as eligible for nomination to the National Register of 

Historic Places, as discussed further below. These 35 sites are enumerated in Appendix A 

and in Tables C-1 and C-2 below. 

Table C-1. Six Felsenthal Project Floodplain Sites Recommended for Extensive and 

Intensive Investigation. 

Site Elevation Location by 

Number Site Name (feet MSL) Drainage Topographic Location Impacts 

3AS329 Marie Saline 62.0-67.5 Ouachita River riverbank/levee near open lake 1, 2, 3, 4 

3AS285 False Indigo 66.9-69.9 Saline River riverbank/levee 2, 3, 4 

3AS286 One Cypress Point 66.6-68.4 Saline River riverbank/levee and old 2, 3, 4 

meander scar 

3BR76 Jug Point Cutoff 67.6-68.4 Saline River riverbank/levee near chute 2, 3, 4 

3AS306 Jug Point 1 67.2-68.1 Saline River riverbank/levee 2, 3, 4 

3AS307 Jug Point 2 67.2-68.0 Saline River riverbank/levee 2, 3, 4 

Note: Impacts are described on pages C-15 to C-17. 

Through the same process of site data review, evaluation, and consultation, the 

remaining 81 floodplain sites were found to be not scientifically, anthropologically, or 

historically significant, and not eligible for nomination to the National Register. Prehistoric 

floodplain sites in this group generally lacked these kinds of research potential: 

C-7 

1. useful samples of artifacts, debris, or organic remains, 

2. in situ features, structures, occupation levels, or multiple occupation levels, and 

3. items or data unique to the locality or region. 

This statement is based on site survey methods (Chapter 5) and data recovered (Appendix 

A).

C-8 

Table C-2. Twenty-nine Felsenthal Project Floodplain Sites Recommended for Twostage 

Mitigation. 

Site Elevation Location by Topographic Probable 

Number Site Name (feet MSL) Drainage Location Impacts 

Group 1 - Prehistoric Floodplain Sites, Gran Marais and Caney Bayou Phases, Mississippi Period 

3AS305 Overcup Oak 3 66.0-67.01 Saline River riverbank/levee 2, 3, 4 

3BR62 Mouth of Long Slough 65.0-66.0 Saline River riverbank/levee 2, 3, 4 

38865 Eagle Creek 3 67.5 Eagle Creek creekbank/levee 2, 3, 4 

3BR66 Eagle Creek 4 66.0-68.01 Eagle Creek creekbank/levee 2, 3, 4 

3UN168 Coon Glory Bend 63.0-65.01 Ouachita River levee near riverbank 1, 2 

Group 2 - Prehistoric Floodplain Sites, Unknown Phases, Mississippi Period 

3AS315 Persimmon 3 68.0-70.01 Saline River levee near riverbank 3, 4 

3AS322 Water Elm 1 67.5 Ouachita River riverbank/levee 2, 3, 4 

3BR79 Eagle Creek 6 65.0 Eagle Creek creekbank/levee 2, 3, 4 

3BR86 Sweetgum 2 67.5 Saline River riverbank/levee 2, 3, 4 

3UN179 Fishtrap Lake 4 63.0-68.01 Lapile Bayou bayou bank 2, 3, 4 

Group 3 - Prehistoric Floodplain Sites, Unknown Phases, Baytown-Coles Creek Period 

3AS287 Water Elm 3 67.0 Ouachita River riverbank/levee 2, 3, 4 

3AS316 Hawthorn 1 64.0-67.0 Saline River riverbank/levee 1, 2, 3, 4 

3AS327 Swamp Privet 66.0 Ouachita River levee near riverbank 2, 3, 4 

3BR82 Bitter Pecan 1 68.0-69.0 Saline River riverbank/levee 3, 4 

3UN151 Lapile Creek 7 64.0-67.01 Lapile Creek creekbank 2, 3, 4 

3UN157 Lapile Creek 5 65.0-69.01 Lapile Creek creekbank 2, 3, 4 

3UN180 Fishtrap Lake 2 63.0-66.01 Lapile Bayou bayou bank 2, 3, 4 

Group 4 - Prehistoric Floodplain Sites, Coon Island Phase, Tchula Period 

3AS321 River Birch 2 63.0-67.01 Ouachita River riverbank/levee 2, 3, 4 

Group 5 - Prehistoric Floodplain Sites, Calion Phase, Poverty Point Period 

3BR78 Mouth of Eagle Creek 63.2-68.0 Saline River riverbank/levee 1, 2, 3, 4 

Group 6 - Prehistoric Floodplain Sites, Unknown Phases, Archaic Periods 

3BR72 Small Cane 1 70.0-73.51 Ouachita River riverbank/levee 3, 4 

3BR73 Small Cane 2 71.0 Ouachita River riverbank/levee 3, 4 

3BR92 Water Locust 65.0-67.0 Saline River riverbank/levee 2, 3, 4 

3UN166 Mud Lake Bend 2 71.1-72.4 Ouachita River riverbank/levee 3, 4 

Group 7 - Historic Floodplain and River Channel Sites, Nineteenth and Early Twentieth Century American 

3AS299 Marie Saline Landing 61.0-69.0 Ouachita River riverbank/levee 1, 2, 4, 5 

3BR8 Goulett Island 63.0-86.0 Saline River Deweyville 2 Terrace 1, 2, 3, 4 

3BR55 Keelboat Brake Sawmill 66.0-68.0 Ouachita River backswamp 2, 3, 4 

3UN122 Grand Marais Landing 64.0-65.0 Ouachita River riverbank/levee 1, 2, 5 

3UN153 Wreck of Lotawanna 34.0-38.0 Ouachita River river bottom 1, 5 

3BR56 Caney Marie Landing 74.0-76.0 Ouachita River riverbank/levee 3, 4 

Note: Impacts are described on pages C-15 to C-17.

Historic floodplain sites among the 81 sites not found to be significant generally lacked 

the following kinds of research potential: 

1. useful samples of artifacts, debris, or organic remains 

2. standing structures or subsurface features 

3. data bearing on historic research problems in any of our research design 

sources, and 

4. association with an important historic event or person. 

This statement is based on site survey methods (Chapter 5), data recovered (Appendix 

A), and historical or archival research pertaining to specific sites, localities, and the 

Felsenthal Region (Chapter 4). 

The determination of “signficant/not significant” has been applied here in a polythetic 

analytical framework (cf. Clarke 1968:35ff). In short, no single criterion was both 

sufficient and necessary for a site to be classed as significant. Rather specific polythetic 

sets of the 26 criteria, as presented in Table C-3, determined whether or not the site is 

assessed as significant. Further, it is recognized that there are patterned regularities in 

site types and properties. Given these regularities, it is possible to recognize only part of 

a set of similar sites as representative of the entire set. Additional study is not needed in 

members of a given grouping—instead examination of representative sites will gather 

sufficient information to mitigate the impacts at all the sites. In a sense while we can 

state that all members of the group are “significant,” for the purposes of the management 

of those particular resources, selected members of the group can be and have been 

selected as representative. It is for these that nomination documentation has been prepared. 

The remainder of the group are then seen, from a resource management perspective, 

as redundant sites and thereby deemed not eligible for the National Register of 

Historic Places. 

Of the various factors considered and enumerated in Table C-3, a relatively large 

number of artifacts (diagnostic ceramics present) factors #1 and #2 proved to be particularly 

critical to determination of a prehistoric site significance. Only three sites 

determined to be significant did not have #1, all of which had #2 plus other important 

factors; only eight sites considered significant had #1 but did not have #2 and all but one 

of these had other important factors. That is, all sites considered significant had #1 or #2 

(two-thirds had both) and all had other factors to add to the decision making equation. 

Of these other factors, five proved most critical: diagnostic lithics, carbonized materials, 

in situ deposits, features present, and distinctive environmental association. All prehistoric 

sites which did have the above characteristics, but were not considered significant, 

were excluded because they were believed to be redundant to others which had similar 

characteristics and which displayed them more obtrusively or in greater quantity, or had 

other combinations of important factors present. 

The historic sites developed no patterns with respect to the importance of particular 

criteria or combinations of criteria in determining significance. Each site was more or 

less unique, perhaps because of the relatively small number of sites. 

C-9

C-10 

Table C-3 summarizes the qualities of significance found to be present or lacking 

in each of 117 floodplain sites surveyed or tested by the Felsenthal Project in 1979. These 

sites are listed by site unit in the order they appear in Chapter 5 and Appendix A, Table 

A-1. Although this table gives the appearance of a mechanical assessment of significance, 

the preceding discussion has explained the more complex process by which significance, 

or lack of significance, was identified. 

nAtionAl register eVAluAtion 

The 35 sites found to be scientifically significant are recommended for nomination 

to the National Register of Historic Places (36 CFR 1202 and 1204). Each can be said to 

“have yielded or [is] likely to yield information important in prehistory or history.” This 

National Register criterion is frequently applied to archeological properties, but warrants 

more detailed treatment, as in our discussion of significance. 

In compliance with the contract Scope of Services, individual National Register 

nominations have been prepared for 16 significant sites that have archeological deposits 

at or below the 65 foot elevation. 

Recommendation for Multiple Resource Area Nomination 

We suggest that a Multiple Resource Area, called “Cultural Resources of the Grand 

Marais Floodplain Area, Arkansas,” be nominated to the National Register (Interim 

Guidelines: How to Complete National Register Multiple Resource Nomination Forms, 

1977). This MRA shall include the 35 known eligible properties, including 29 prehistoric 

and six historic floodplain sites. (Goulett Island, 3BR8, has significant prehistoric and 

historic components above and below floodplain level, but we are not attempting here 

to nominate the entire elevated “island” itself because of survey and study limitations; 

our survey data established that prehistoric and historic artifacts and refuse extended 

downslope to floodplain and river level, and these deposits are identified as significant). 

The geographical extent of the proposed MRA is all of the Recent floodplain of 

Ouachita and Saline rivers and their tributaries within the Felsenthal National Wildlife 

Refuge. (This floodplain and some adjacent lowlying Pleistocene terrace surface are an 

environmental unit we have informally called “overflow bottomland.”) We can suggest 

and provide map series which will spatially define the proposed MRA. Based on our 

intensive cultural resources survey and testing in floodplain terrain (but not comprehensive 

survey undertaken to identify all resources), the MRA includes 35 sites deemed 

eligible, 82 sites presently deemed not eligible, and one site (3UN121) awaiting eligibility 

recommendation. It is probable that more sites of significance will be discovered and 

added at a future time. It may therefore be appropriate to append the notation “partial 

inventory” to our MRA.

Table C-3. Summary of Site Characteristics Relevant to Research Potential and 

Scientific or Historic Significance for 117 Floodplain Sites. 

Content Context Significance Determination 

Site Number Site Name (see explanatory note at end of Table) 

UNIT 1: Ouachita River above Mile 254 

3UN167 Mud Lake Bend 1 B 

3UN166 Mud Lake Bend 2 1 7 A 

3UN165 Mud Lake Bend 3 B 

3UN164 Stormhole Bend B 

3BR71 Parrigeethe Shoals 1 1 F 

3BR33 Parrigeethe Shoals 2 B 

3UN163 — B 

3BR75 Eutaw Rapids 2, 5 7 F 

3UN162 — B 

3BR73 Small Cane 2 1, 3 7 A 

3BR74 — B 

3BR72 Small Cane 1 1, 3, 4 A 

3UN161 Fletchers Landing 4 7 C 

3BR70 Hunter’s Swan 2 C 

3BR57 Little Bay Rapids B 

3UN188 Five Mile Bend 2 C 

3UN185 — B 

3BR77 Lower Saline Cutoff E 

UNIT 2: Lower Saline River 

3BR92 Water Locust 1 7 A 

3BR91 — B 

3BR90 Round Turn B 

3BR89 Mouth of Miller Creek 2 C 

3BR88 — B 

3BR87 Horton Island Landing 2 C 

3BR86 Sweetgum 2 1, 2 7 A 

3BR85 Sweetgum 1 B 

3AS318 Hawthorn 3 B 

3AS317 Hawthorn 2 B 

3AS316 Hawthorn 1 2, 4 7, 10 A 

3BR81 Prairie Island Landing B 

3BR80 Prairie Island Bend 2 7 F 

3AS315 Persimmon 3 1 A 

3br82 Bitter Pecan 1 1, 2, 4, 5 7, 10 A 

3AS314 Persimmon 2 1 F 

3BR84 Bitter Pecan 3 B 

3br83 bitter pecan 2 b 

3AS313 Persimmon 1 B 

3BR78 Mouth of Eagle Creek 1, 3, 4 7, 8, 10 A 

3AS312 Willow Oak B 

3AS311 Tupelo Gum B 

3AS309 Jug Point 4 B 

3AS310 — B 

3AS308 Jug Point 3 1, 2 F 

3AS307 Jug Point 2 1, 2, 3, 4, 5 7, 8 A 

3AS306 Jug Point 1 1, 2, 3, 4, 5, 6 7, 8, 10 A 

3BR76 Jug Point Cutoff 1, 2, 3, 4, 5 7, 8, 11, 12 A 

3AS305 Overcup Oak 3 1, 2 A 

3AS304 Overcup Oak 2 2 C 

3AS303 Overcup Oak 1 1, 2 F 

3AS302 Sycamore B 

3AS286 One Cypress Point 1, 2, 3, 4, 5 7, 8, 12 A 

3BR59 Possumhaw B 

3BR58 Buttonbush 1, 2, 3 7, 10 A 

3BR60 — B 

3BR61 — B 

3BR62 Mouth of Long Slough 2, 4, 5 7 A 

3AS285 False Indigo 1, 2, 3, 4, 5 7, 8, 10, 11 A 

3AS284 Mouth of Saline River B 

C-11

C-12 

Table C-3 (continued). Summary of Site Characteristics Relevant to Research Potential 

and Scientific or Historic Significance for 117 Floodplain Sites. 



UNIT 3: Lower Eagle Creek 

3BR68 — B 

3BR67 — 2 C 

3BR66 Eagle Creek 4 1, 2, 4, 5 10 A 

3BR69 Eagle Creek 5 2 C 

3BR65 Eagle Creek 3 1, 2, 4 7 A 

3BR64 Eagle Creek 2 B 

3BR63 Eagle Creek 1 1, 2 F 

3BR79 Eagle Creek 6 1 7 A 

UNIT 4: Ouachita River below Mile 254 

3UN184 — 2 B 

3UN92 Brown Camp 2 C 

3AS329 Marie Saline 1, 2, 3, 4, 6 7, 8, 9, 10, 12 A 

3AS326 — 3 C 

3AS301 Honey Locust 2 C 

3AS319 — 7 C 

3As320 river birch 1 b 

3As321 River Birch 2 1, 2 A 

3AS288 Red Oak B 

3As322 Water Elm 1 2 7 A 

3AS332 Water Elm 2 B 

3AS287 Water Elm 3 1, 2 7, 10 A 

3AS323 Water Elm 4 1, 2 7 F 

3AS324 — 3 7 C 

3UN123 Mouth of Lapile Creek 1, 2 F 

3AS327 Swamp Privet 1, 2 7, 10 A 

3AS325 — 2 7 F 

3UN169 — B 

3UN168 Coon Glory Bend 1, 2 A 

3UN121 Mo-Pac 2, 4 7 F 

UNIT 5: Lower Lapile Creek 

3UN178 Lapile Creek 1 B 

3UN160 Lapile Creek 2 2 C 

3UN159 Lapile Creek 3 2 C 

3UN158 Lapile Creek 4 1, 2 F 

3UN157 Lapile Creek 5 1, 2, 3 A 

3UN156 Lapile Creek 6 1 F 

3UN152 Lapile Creek 8 B 

3UN151 Lapile Creek 7 1, 2, 3 A 

UNIT 6: Oxbow Lakes and Backswamp 

3UN187 St. Mary’s Lake North 1 7, 12 C 

3UN186 — B 

3UN170 Borrow Pit near Lapoile Creek D 

3UN183 Lapoile Creek B 

3UN182 Fishtrap Lake 1 1 12 F 

3un180 Fishtrap Lake 2 1 12 A 

3UN181 Fishtrap Lake 3 12 C 

3UN179 Fishtrap Lake 4 1, 2 12 A 

3AS295 Borrow Pit near 82 Bridge 1 1, 2 D 

3AS294 Borrow Pit near 82 Bridge 2 1 D 

3AS292 Wildcat Lake 1 12 F

Table C-3 (continued). Summary of Site Characteristics Relevant to Research Potential 




UNIT 8: Historic Sites 

3BR32 Ganer Construction Mound 15, 17 20, 23 G 

3BR56 Caney Marie Landing 13, 14 19, 20 A 

3br8 Goulett Island 13, 14, 16 20, 22, 23 A 

3BR81 Prairie Island Landing B 

3BR55 Keelboat Brake Sawmill 15, 16 20, 22 A 

3UN155 Ouachita Belle Landing B 

3UN88 Fletchers Landing and Woodyard 13, 17 C 

3UN92 Brown Camp 20 C 

3AS330 First Slough 13, 17 22 G 

3AS331 Marie Saline Ferry 13, 14 23 G 

3AS300 South of Highway 82 Bridge 14, 17 22 G 

3AS299 Marie Saline Landing 20, 23 A 

3UN174 Artesian Well B 

3un122 Grand Marais Landing 16 19, 21, 23 A 

3UN153 Wreck of the Lotawanna 23, 24 A 

Notes: Historic sites 3UN81 and 3UN92 have prehistoric components and appear twice in this table. The following explanatory 

notes pertain to columns 3-5 in this table: 

prehistoric content 

(1) Sample of artifacts and/or debris recovered was substantial relative to site size, exposure, extent of testing, or other factors. 

(2) One or more temporally or culturally diagnostic ceramic sherds were recovered. 

(3) One or more temporally or culturally disgnostic projectile points were recovered. 

(4) Carbonized seeds or charcoal were recovered. 

(5) Animal bone or mussel shell were recovered. 

(6) Artifacts evidently unique to the locality or region were recovered. 

prehistoric context 

(7) One or more buried artifacts or items of debris were recorded in place. 

(8) A visible buried occupation level marked by artifacts and/or organic stain was observed. 

(9) Superposed buried occupation levels were observed. 

(10) One or more subsurface features were recorded. 

(11) A midden or refuse deposit was recorded. 

(12) Association of site with a relict floodplain feature or other distinctive environmental feature was observed. 

historic content 

(13) Sample of artifacts and/or debris recovered was substantial relative to site size, exposure, or other factors. 

(14) One or more temporally diagnostic artifacts were recovered. 

(15) One or more technologically diagnostic artifacts were recovered. 

(16) Wood or charcoal was recovered. 

(17) Animal bone, mussel shell, or oyster shell was recovered. 

(18) Artifacts evidently unique to the locality or region were recovered. 

historic context 

(19) One or more buried artifacts or items of debris were recorded in place. 

(20) Above-ground structural remains were observed. 

(21) One or more subsurface features were recorded. 

(22) A refuse deposit was recorded. 

(23) Site is associated with archival documents or secondary sources. 

(24) Site is associated with an important event or person. 

C-13

C-14 

Table C-3 (concluded). Summary of Site Characteristics Relevant to Research Potential 


significAnce deterMinAtion 

(A) Significant (See discussion on page C- 4). 

(B) Not significant because no diagnostic criteria present. 

(C) Not significant because insufficient criteria present. 

(D) Not significant because riverbank erosion or some other natural disturbance is believed to have destroyed a large part of the 

site. 

(E) Not significant because historic or modern cultural disturbance is believed to have destroyed a large part of the site. 

(F) Not significant because redundant—other similar sites determined eligible have greater potential for producing significant 

data. 

(G) Not significant because does not conform to National Register guidelines. 

With regard to the proposed MRA, all 35 eligible properties are believed to be 

under federal ownership presently (either the U.S. Army Corps of Engineers or the U.S. 

Fish and Wildlife Service), but some properties deemed not eligible and some intrusive 

properties not recorded as archeological or historical sites are privately owned presently 

(e.g., modern fishing camps and boat landings). 

Finally, we believe that choice of the Multiple Resource Area concept will lead to 

practical management of the nomination process, now and in the future, and that National 

Register status of the proposed “Grand Marais Floodplain Area, Arkansas” will 

truly enhance recognition and wise management of significant cultural resources in this 

area and region. 

AssessMent of site iMpActs 

Felsenthal Project contract requirements include discussion of probable impact on 

cultural resources through proposed work of the U.S. Army Corps of Engineers (Scope 

of Services, item 7b). Schiffer and Gumerman (1977:291), among others, point out that 

responsible and informed forecast of impacts is of extreme importance to sponsors as 

resource managers. Through review of various construction maps provided to us, informal 

discussions with the Contracting Officer’s representatives, and review of current 

archeological and management documents pertaining to reservoir inundation, we have 

prepared a statement of probable impacts which can be adjudged on a site-by-site basis. 

In fact, in Tables C-1 and C-2 below we do predict one or more impacts in the case of 

each site deemed eligible for the National Register of Historic Places.

C-15 

Our statement of probable impacts recognizes that inundation studies for the purpose 

of conserving or mitigating effects on cultural resources require expert assistance 

from engineers, geologists, hydrologists, and other scientists. We would therefore welcome 

review and assistance in this area. For present purposes we have relied on certain 

inundation studies, many of which are general models or preliminary reports of work in 

particular reservoirs on few sites (Garrison 1975, 1977; Carell et al. 1976; Schnell and Tyler 

1976; Garrison et al. 1977; Lenihan et al. 1977). As in the case of assessing significance, 

we believe there is a particular cultural-environmental context in the Felsenthal Project 

area, against which impacts should be assessed. Some preliminary conclusions include 

the following: 

1. All floodplain sites now known undergo prolonged submergence each 

year as the result of previous artificially raised river level (Lock and Dam 

No. 6), as well as the natural annual hydroperiod (Chapter 2). 

2. Many known floodplain sites lie on riverbanks and levees along meandering 

channels which are stable, or have a markedly low rate of lateral erosion 

compared to other southeast U.S. rivers (Chapter 2). 

3. The net effect of increased or permanent inundation caused by new navigation 

pool and wildlife pool levels will be adverse on floodplain site 

content and context, not merely preservation under a water column (see 

inundation studies cited above). 

4. A variety of chemical, mechanical, and biological effects of inundation on 

archeological resources have been studied or methods of study are being 

developed (e.g., Lenihan et al. 1977). 

5. We are considering also noninundation impacts, such as construction in 

floodplain areas and recreational effects, but our basis for assessment of 

these impacts is insufficient. 

On the basis of these comments and considerations, the following five categories of 

probable impacts are presented (in special reference to variation among floodplain areas 

and sites by elevation). 

1. Permanent Inundation Impacts 

All sites or portions of sites lying at or below 65 feet elevation will be permanently 

inundated. Direct mechanical, chemical, or biological impacts may be expected to affect 

all such sites in varying degrees. On a site-by-site basis impacts may vary in the following 

dimensions:

C-16 

Site location relative to section of new navigation pool (upstream, downstream, 

midpool) 

Site location relative to existing channels, bends, and floodplain features 

Site physical conditions, including sediment and vegetation cover, and existing 

exposure 

Site content, including soil matrix, features, artifacts, organic and inorganic debris 

Depth of submergence by new navigation pool 

2. Navigation Pool Shoreline Impacts 

All sites or portion of sites lying at or near 65 feet elevation will be subject to shoreline 

disturbance processes, including any or all of the following: 

Wave current erosion produced by large commercial watercraft or smaller pleasure 

craft, and also by wind 

Scouring by downstream flow and by seasonal or occasional changes in flow 

Biological activity, including the addition or removal of plant and animal populations 

which inhabit or visit the shoreline 

These impacts will vary among sites located immediately upon a river or bayou 

bankline and sites located away from such banklines. 

3. Wildlife Pool Shoreline Impacts 

All sites or portions of sites from 65 feet to 70 feet elevation and for some distance 

above 70 feet will be subject to the effects of a temporary or seasonal pool level and 

seasonally reduced pool level. Chemical, mechanical, and biological impacts associated 

with inundation and shoreline disturbance will vary among sites, depending on location, 

elevation, context, and other factors. 

4. Recreational Activity Impacts 

Increased or shifting recreational use of shorelines, levees adjoining channels, 

backswamp areas, and elevated “islands” will affect floodplain terrain at and near the 

new pool levels and shorelines. Based on existing uses, the mechanical impacts on sites 

in these areas could include boat landing and launching, four-wheel drive vehicle operation, 

bankline fishing, hunting, and camping. Collecting of artifacts exposed by erosion 

or digging in floodplain sites may be stimulated by raised and lowered pool levels and 

their mechanical effects.

5. Direct Construction Impacts 

C-17 

All sites within or, in some cases near, construction rights-of-way will be impacted 

by excavation, spoil deposit, heavy equipment operation, and other similar factors. 

These impacts are anticipated primarily for the new Felsenthal Lock and Dam structures 

and for the Crossett Harbor facility. Future construction of bendways, boat launching 

facilities, access roads, boat channels, or other structures may affect floodplain sites. 

recoMMended progrAM of MitigAtion 

Our recommended program of mitigation is based on several general considerations. 

First, that we have assembled and interpreted an adequate and sound body 

of data regarding cultural resources in the Felsenthal Project area (despite the many 

limitations discussed throughout this report). Second, that we have made an adequate 

appraisal of site significance, based on current regional knowledge and broad based 

research goals. Third, that 35 floodplain sites recommended for nomination to the National 

Register of Historic Place (through the mechanism of a Multiple Resource Area) 

will all be adversely affected to some degree by Felsenthal Navigation Pool structures 

or modifications of floodplain terrain. And fourth, that mitigative actions proposed here 

will need to proceed during an advanced stage of project implementation (construction 

activity) rather than early in the planning process. Obviously alternatives for mitigation 

of impact must be considered in the light of all these sources of information and practical 

considerations. 

Mitigation by Data Recovery and Alternatives 

Lipe (1974), Schiffer and Gumerman (1977), and others generally rank mitigative 

measures in terms of the goals of modern conservation archeology (see also Advisory 

Council on Historic Preservation 1980). We concur with this ranking, and shall briefly 

here examine the feasibility of alternatives. 

Avoidance. The shorelines at 65 and 70 feet are taken as invariable, and all 35 sites 

deemed eligible are either below or very near these elevations. Avoidance is here considered 

infeasible (but see General Recommendations below). 

Preservation-in-Place. Many techniques for site protection with pavement, plastic, 

ripraps, cofferdams, etc., have been suggested (e.g., Garrison 1975:290). Examples of

C-18 

active preservation are increasing in all parts of the country, although long term benefits 

are unknown. In the case of inundated sites, mechanical impacts can be reduced and 

perhaps nearly eliminated, but chemical, analytical, and contextual impacts of submersion 

are not well understood and cannot be controlled or eliminated presently (Lenihan 

and others 1977). Any decision to preserve sites in place must contend with a high cost 

factor. We are not prepared to recommend this alternative for even the smallest site of 

significance, although we approve of experimentation on active preservation in other 

cases. Also, we would value discussion with engineers and stabilization experts on the 

feasibility of preserving floodplain sites in place in the Felsenthal Project area. 

Mitigation by Data Recovery. Scientific excavation involving specific, appropriate, 

and rigorous data recovery techniques is the last alternative, and the one we can recommend 

in this case without reservation. We turn now to a summary of the proposed 

mitigation program which is designed to maximize data recovery in the “particular cultural-environmental 

context” of the Felsenthal Project area. Mitigation costs, presented 

as recommended man-day expenditures for field activity, are presented on Tables C-3, 

C-4, and C-5. It is recommended that these activities be conducted over two field seasons 

due to limits of optimal time for fieldwork (July-November). 

I. Research Design Outline (Phase 1) 

The research design shall be interdisciplinary in scope and shall incorporate a flexible, 

multistage program of data recovery. Broad-based research domains shall include, 

but not necessarily be limited to, significant, relevant problems identified in the 1979- 

1980 Felsenthal Project: 

Domain 1—Floodplain geoarcheology 

a. fluvial geomorphology in the Grand Marais Lowland (processes and change) 

b. geochronometry in alluvial sites and localities (problems and potential) 

Data Base: Marais Saline Lake, Marie Saline site (3AS329), Goulett Island, and 

other terrace and floodplain localities. 

Domain 2—Floodplain environmental archeology 

a. function and variability in Mississippi period extractive sites (including locational 

analysis and comparative ethnohistorical research) 

b. function and variability in Woodland period extractive sites 

c. function and variability in extractive/habitation sites of the Archaic period 

d. overflow bottomland focal resources 

e. locational and catchment analysis of extractive sites 

Data Base: False Indigo (3AS285), Jug Point Cutoff (3BR76), and other sites listed 

in Tables C-1 and C-2.

Table C-4. Mitigation Budget Estimate (in field man days): PHASE 2: Mitigation of Six Prehistoric Floodplain Sites. 

Number Number Total Volume of Average Sump Water Other 

Site Crew of of Manpower Site to Excavated Backhoe Pump Screening Required 

Number Size* Supervisors Weeks Effort** be Excavated Per Day Time Needed Flotation Expertise 

(in days) (cu meters) (cu meters) (hours) 

3AS329 18 2 16 1600 806 .504 40 + + geomorphology, 

chronometrics 

3AS285 7 1 3 120 54 .450 + zooarcheology 

3AS286 7 1 14 560 278 .496 + ethnohistory 

3BR76 7 1 4 160 75 .469 + — 

3AS306 7 1 6 240 129 .538 + — 

3AS307 7 1 2 80 37 .463 + — 

— 1379 .487 

2760 

man days 

* a total of three crews (one of 18 members and two of seven members) could conduct this work in a single four month field period. This model assumes 

an average excavation effort of ca .5 sq m per person per day 

** does not include Project Archeologist 

C-19

C-20 

Table C-5. Mitigation Budget Estimate (in field man days): PHASE 3: Two-Stage Mitigation, 23 Prehistoric Floodplain 

Sites. 

∑ 

Crew Number of Number of Number of Number of Surface Area 5% Average Total Volume Average 

Size Supervisors Crews Weeks Man Days 23 Sites Sample Depth to be Excavated cu m/man day 

Stage One 3 1 5 6 600 8,693 m2 435 .5 217 m3 .36* 

Excavate 5% 

sample of 23 

sites 

Stage Two 7 1 6 240 Total manpower effort 130 m3 .54 

Excavate ex- per site per site** 

tensively at 1440 

least six sites 

* greater per day manpower cost due to boat travel and logistics associated with the dispersed nature of the sites 

** Site 3AS306 is considered representative of site size and mitigation effort required

Table C-6. Mitigation Budget Estimate (in field man days): PHASE 3: Two Stage Mitigation—Six Historic Sites. 

Surface Metal Controlled Subsurface 

Site Historic Crew Number of Manpower Effort Survey, Detector Surface Testing Backhoe Underwater 

Number Archeologist Size Weeks in Man Days Site Type Mapping Survey Collection (hand) Testing Study 

Stage One 

3AS299 1 3 2 40 canal/landing x x x x x 

3BR8 1 3 1 20 dump x x x 

3BR55 1 3 1 20 sawmill x x x 

3UN122 1 3 2 40 landing x x x 

3BR56 1 3 1 10 landing x x x 

3UN153 1 3 1 30 wreck x 

1 3 7 140 

Stage Two: additional archival work and further field investigations may be required on Stage Two mitigation; these work requirements cannot be 

estimated until after Stage One results are evaluated. 

C-21

C-22 

Domain 3—Floodplain behavioral archeology 

a. interassemblage variability among floodplain sites 

b. material correlates of low bulk and high bulk procurement sites 

c. material correlates of transient site use, reuse, and anticipated return (discard and 

curate behavior) 

d. activity sets and activity areas 

Data Base: Especially those sites listed in Table C-1. 

Domain 4—Ethnohistorical analysis 

a. composition and identity of social group occupying an extractive site (tested 

against a Quapaw model) 

b. seasonality and function of an extractive site (tested against a Quapaw model) 

Data Base: One Cypress Point (3AS286). 

Domain 5—Historic site analysis 

a. structure and function of a nineteenth century riverboat landing 

b. structure and function of lowland sawmills, logging facilities, and logging techniques 

(prior to 1930) 

c. depositional and postdepositional history of a sunken steamboat 

d. functional and socioeconomic analysis of riverbank, riverbottom, and canal bottom 

refuse 

Data Base: Marie Saline Landing (3AS329), Keelboat Brake Sawmill (3BR55), Goulett 

Island (3BR8), Lotawanna (3UN153) and other historic sites in Table 

C-2. 

Domain 6—Floodplain site transformations 

a. mechanical, chemical, and biological effects of inundation 

b. small site visibility and methodology 

Data Base: All sites in Tables C-1 and C-2. 

II. Extensive and Intensive Investigation of Six Floodplain Sites (Phase 2) 

Six of eight sites tested extensively in 1979 and reported in Chapter 6 are known 

to have relatively great research potential. Data relevant to a variety of research problems 

cited in this research design outline are adduced for these sites, as well as a variety 

of adverse impacts (Table C-1). A brief statement of proposed research scope and field 

methodology is presented below for each site, but the reader must refer to Chapter 6 for 

rationale. 


1. Stratigraphic excavation of two deep blocks, each about 15 x 15 m in size, at upstream 

and downstream ends of Area C, emphasizing cultural levels from about 

0.5 m depth to below river stage or sterile soil; extensive use of water screening in 

all levels and pumping of deep levels (below about 1.5 m).

2. Precisely controlled excavation of deep occupation floors, if identified 

3. Specially designed chronometric sampling and dating of deep levels 

4. Backhoe trenching primarily for geological stratigraphy and soil sediment sampling 

keyed to profiles 

5. Minimal offsite testing or testing in Areas A and B 

6. Geomorphological study of site environs, including Marais Saline Lake 


1. Total excavation of a 12 x 15 m midden in Area A to about 0.3 m depth; water 

screening and/or flotation of all midden levels 

2. Precisely controlled excavation of features or traces of structures, if identified 

3. Chronometric sampling and dating of midden 

4. Minimal offsite testing or testing in Area B 

One Cypress Point (3AS286) 

1. Precisely controlled horizontal excavation of a 16 x 58 m area (which may circumscribe 

two activity or residential loci) to sterile soil at 0.3 m depth 

2. Precisely controlled excavation of features, if identified 

3. Water screening and/or flotation on a sampling basis 

4. Dating, if suitable material recovered 

5. Comparative ethnohistorical research 

Jug Point Cutoff (3BR76) 

1. Total excavation of an 11 x 17 m occupation area and small midden to about 0.4 m 

depth; water screening and/or flotation of all midden levels 

2. Precisely controlled excavation of features or traces of structures, if identified 

3. Chronometric sampling and dating of midden 

4. Minimal offsite testing 


1. Precisely controlled horizontal excavation of two discrete activity or residential 

loci, Area A (8 x 20 m) and Area B (9 x 30 m), to 0.3 m depth 




5. Dating, if suitable material recovered 

C-23 


1. Precisely controlled horizontal excavation of three discrete activity loci, Area A (3 x 

8 m), Area B (2 x 5 m), and Area C (6 x 15 m), to about 0.3 m depth 




5. Dating, if suitable material recovered

C-24 

III. Two-stage mitigation of 29 floodplain sites (Phase 3) 

This part of our proposal represents a much more flexible approach to mitigating 

impacts on certain floodplain sites, and also represents we believe, an efficient means of 

attaining stated research goals. Schiffer and Gumerman (1977:325), among others, suggest 

that the two-stage mitigation strategy is well suited to programs of the scope we are 

proposing. In this strategy exploratory testing, and additional survey if needed, precede 

selection of sites for intensive excavation and data recovery. Two-stage mitigation 

assumes that among a large group of significant sites, testing will identify those with 

greatest research potential and probably also those with maximum predicted impact. 

Sites can be stratified on these (or other) bases and one or more optimal sites within each 

stratum intensively mitigated. As in any multistage field program, the best choices are 

based on maximal, current site data and assessment. 

In the case of Felsenthal Project floodplain sites, we have identified 29 that are well 

investigated on the basis of site survey methodology (Chapter 5), and that are recommended 

for nomination to the National Register (Table C-2). We know, for example, 

much about the location, elevation, extent, depth, content, integrity, and often the cultural 

period or phase for each (Appendix A). However, we also expect that research 

potential varies among these 29 sites, both prehistoric and historic, and that redundancy 

in research potential may be encompassed by this group. Since all prehistoric sites, and 

all or parts of historic sites are buried, and since subsurface exploration is limited to 

shovel or auger holes and examination of riverbank exposures, we cannot know the full 

research potential in all cases. 

We therefore propose to stratify or group these sites by cultural period (Table C-2), 

and to conduct limited testing within each group on the most flexible and efficient basis 

to identify those sites with greatest research potential. Intensive excavation or other appropriate 

mitigation measures for selected sites can then proceed as rapidly as possible. 

While recommending flexibility and responsibility on the part of the project archeologist 

and his field staff, we also propose the following guidelines for planning and cost estimation 

purposes: 

1. Within each designated group of prehistoric floodplain sites, one or more sites 

(and perhaps all) shall be tested by appropriate subsurface techniques, not to 

exceed 5% of mapped site area (we believe limited testing of this extent was 

demonstrably effective during our 1979 site testing stage). 

2. Within each designated group of prehistoric floodplain sites at least one site 

(and ordinarily no more) shall be extensively mitigated by appropriate techniques, 

including stratigraphic or horizontal excavation strategies, and water 

screening and/or flotation (as in Phase 2 above). 

3. Appropriate testing of historic riverboat landings shall include in the case of 

Marie Saline Landing (3AS299), the following:

C-25 

a. shovel testing and metal detection in the vicinity of the 200 m long canal 

b. limited backhoe or dredge tests of the submerged canal-bottom deposit 

c. backhoe trenches in certain accessible areas in the vicinity of the canal, if 

indicated by other subsurface or documentary evidence. 

4. Appropriate testing at Goulett Island (3BR8) should include 

a. sampling the riverbank refuse deposits (>50 years) 

b. geomorphological study of the Goulett Island riverbank exposure (see Research 

Design outline) 

c. see General Recommendations below for island resources above floodplain 

level 

5. Wreck of the Lotawanna (3UN153) shall be located, inspected, and studied by a 

qualified diver-archeologist 

6. Extensive and intensive mitigation of historic sites including additional archival 

research, shall be designed and implemented on the basis of Phase 3 testing 

results—if necessary. 

7. The project archeologist shall regularly and formally consult with Contracting 

Officer’s representative and Principal Investigator as two-stage mitigation 

proceeds 

IV. Analysis (Phase 4) 

A field laboratory will operate concurrently with Phases 2 and 3, accomplishing all 

processing of collections and samples and initiating analysis. Both routine and specialized 

analyses following the field phases shall employ experts in the following fields: 

prehistoric archeology, geomorphology, ecology, historical archeology, ethnohistory, 

statistics, and computer science. 

V. Report of Results (Phase 5) 

A comprehensive technical report of mitigation program results, including all 

interdisciplinary efforts shall be prepared and submitted to the sponsor in a timely manner. 

After a review process, these results shall be revised or supplemented and published 

for the use of the scientific and resource management communites. Interpretation and 

dissemination of results for the lay community is discussed below (Recommendations 

for Interpretive Displays).

C-26 

Logistics and Scheduling 

Field phases proposed above must proceed during low water and relatively dry 

weather, normally late summer and fall, especially when Ouachita and Saline River levels 

remain at present lock stage (62.2 feet elevation). 

Multiple field crews must operate concurrently in order to accomplish data recovery 

phases in two field seasons. We recommend that Phase 2 (sites of demonstrated 

research potential) be accomplished during the first field season, and Phase 3 (two-stage 

mitigation) be accomplished during a second field season. However, establishment of 

the navigation pool at 65 feet elevation will prevent access to some sites, and the construction 

schedule must be considered here. 

Extensive boat travel is required to reach many, or even most, floodplain sites here 

designated for fieldwork, and several types or sizes of boats will be required for personnel 

and equipment transport, as well as for special needs of collaborating scientists. 

Underwater exploration and study of Lotawanna (3UN153) clearly requires specialized 

diving and recording equipment. Water pumps, water screens, and flotation devices are 

required for each of the intensive mitigation tasks. 

The use of backhoe trenching on a limited basis in particular site contexts (Marie 

Saline, 3AS329, and Marie Saline Landing, 3AS299) is anticipated. 

Recommendations for Interpretive Displays 

The Felsenthal Project Scope of Services requires that we furnish recommendations 

for interpretive displays, and such recommendations are appropriate to the philosophy 

of mitigation shared by conservation archeologists. Schiffer and Gumerman (1977:326), 

for example, state that “an excavation program [can lead] not only to scientific results, 

but also to enhancement of public educational and recreational values.” We believe that 

systematic scientific information derived from our survey and testing project, earlier 

work in or near the project area, and our proposed archeological mitigation program can 

lead to a beneficial public interpretation program, equal in quality to any existing in the 

state of Arkansas. In this section we propose two kinds of information: themes which are 

appropriate to the cultural-environmental context of our study area, and forms or interpretive 

media through which these themes can be transmitted. 

However, we should acknowledge first that the floodplain area and resources 

studied by us are circumscribed by the Felsenthal National Wildlife Refuge and that a 

master plan for wildlife and public use of this area has been designed by the U.S. Fish 

and Wildlife Service (1979). It seems clear that recommendations we make here should 

be compatible with objectives of the Wildlife Refuge, and also that interpretive efforts

(as well as resource management efforts) should be closely coordinated by the U.S. 

Army Corps of Engineers and the U.S. Fish and Wildlife Service. 

C-27 

The master plan referred to above (the version available to us is a general refuge 

map with limited, but useful, planning information) estimates that annual visitors to 

the Wildlife Refuge will increase from the 1979 level of 50,000, summarizes the wildlife 

management program which includes increased levels of hunting and fishing, and lists 

various planned facilities for public use, including a visitor station, boat access sites, 

roads, foot trails and canoe trails, and support structures. The plan states briefly that 

“the refuge also offers protection to some of the most significant archeological resources 

located in the state of Arkansas.” All professional archeologists knowledgeable about the 

Felsenthal region, with whom we have conferred, agree that numerous sites of great scientific 

and public significance exist within the boundaries of Felsenthal National Wildlife 

Refuge. It has been our task to locate and evaluate floodplain sites, but some of our 

information clearly pertains to sites and environmental features throughout the refuge 

area. 

Themes. The following themes of various scope are suggested as particularly relevant 

to the study area and suitable for public educational and recreational presentation. 

Some themes refer to Indian prehistory or ethnohistory, some to European or American 

presence, and some to unique ecological or geohydrological features (an unusual culturalenvironmental 

context has been emphasized in this report): 

1. Native Americans of the Ouachita Valley: Ways of Life in the Grand Marais 

Lowland 

2. Hunters and Fishermen in the Ouachita River Valley before You 

3. Indian Earthworks, Mounds, and Mound Centers 

4. Indian Villages, Farmsteads, and Camps 

5. Prehistoric Trade Networks 

6. Caddo, Koroa, and Quapaw Indians in the Ouachita Valley 

7. DeSoto’s Expedition: First Europeans to Enter the Ouachita Valley 

8. French Hunters and Trappers in the 1700s 

9. Dunbar-Hunter Exploration, 1804-1805 

10. Early American Settlements on Ouachita and Saline Rivers 

11. Improving the Ouachita River for Navigation 

12. Steamboating on the Ouachita and Saline Rivers 

13. Riverboat Landings and Commerce in the 1800s 

14. Growth of the Lumbering Industry and Company Towns 

15. What is a Bottomland Hardwood Forest? Hydric Tree and Shrub Associations 

16. Energy Flow and Nutrient Cycling in the Ouachita River Floodplain 

17. Floodplain Forest Biomass: It’s Enormous 

18. Wildlife in Bayous, Brakes, Backswamps, and River Channels 

19. Wildlife in Oak-Pine Forests and Lowland Prairies 

20. Preserving Endangered Species: Alligator, Eagle, and Red-cockaded Woodpecker

C-28 

21. Spanish Moss in Arkansas 

22. Native Fishes and Molluscs of the Ouachita River System 

23. Poisonous Snakes of the Ouachita Valley 

24. What is an Oxbow Lake? Meandering River Processes 

25. Wetlands in Arkansas: The Ecological Imperative 

Forms. The means of interpretation best suited, in our opinion, to presenting local 

cultural and environmental themes are noted below, with examples of specific resources 

or optimal locations for interpretive displays: 

1. Visitor Center. Relatively detailed themes and related or sequential themes can 

best be presented in the specially designed exhibit area of a visitor center. Many excellent 

visitor centers in this country were built near, and designed to harmonize with, 

major Indian mound sites (for example at Toltec Mounds State Park, Arkansas). It is essential 

that intensive archeological survey be a part of site planning so that visitor center 

buildings, roads, parking lots, and trails do not destroy portions of the cultural resource 

they are interpreting and protecting. The most accessible, and also the most thoroughly 

studied, major mound site in the refuge is Shallow Lake (3UN9/52), but two less accessible 

mound groups are far more dramatic in extent and locale (Watts Field, 3UN18, and 

Eagle Lake Mounds, 3BR4). State archeologists are deeply concerned with protecting the 

integrity of all three sites mentioned above. 

There are no monumentally constructed or dramatic prehistoric sites in the floodplain 

which could serve as the focus of a visitor center, and we doubt the feasibility of 

developing a lowlying site because of annual inundation. (Big Mound Ridge, 3AS6, is 

outside the refuge, but is an important lowlying mound site.) 

2. Standing Historic Structures. The few historic structures known to us include 

twentieth century mounds at 3BR32 and 3UN92. The former is related to early lock and 

dam construction (Theme 11 above), but the latter was not found to be historically or 

regionally sigificant (Chapter 5). These sites do not appear to merit development and 

interpretation. 

Goulett Island (3BR8) includes one or more standing structures related to nineteenth 

and early twenteeth century farming and logging activity (we briefly describe a 

primitive log barn in Chapter 5). Unfortunately, neither historic or prehistoric resources 

of Goulett Island have been systematically surveyed or recorded in modern times—we 

have emphasized the potential significance of Goulett Island at several points in the 

report (see Chapters 2, 5, and General Recommendations below). This locality may have 

both cultural and natural values which should be protected and interpreted, but it is in a 

remote area of the Refuge. 

3. In-Place Site Exhibits. Such exhibits normally include prehistoric or historic 

archeological remains, exposed in the ground, stabilized, and interpreted by display

panels or markers. This type of exhibit could be developed at major mound sites and at 

numerous other sites in the refuge (e.g., Jones Lake, 3UN81, and Locust Ridge, 3UN8) 

which are located near boat access facilities or trails. Such development requires a carefully 

designed program of excavation, and also an informed decision to excavate rather 

than preserve the site intact (i.e., public significance is thought to be great). Site exhibits 

also present major security problems, unless continuously guarded. This form of interpretation 

does not seem to be feasible for floodplain sites that will be overflowed annually. 

C-29 

4. Outdoor Exhibit Panels and Kiosks. Single or multiple display panels at important 

visitor locations or at special points of interest are currently a major form of 

interpretation in outdoor recreation areas. Many of the themes suggested above are 

adapted to panel display. Such displays must be vandal-proof, and must not include real 

prehistoric or historic artifacts, since they will not be attended. Although we believe that 

such outdoor exhibits are well adapted to the future interpretive program in the refuge, 

we strongly recommend that they do not provide specific archeological site location 

information. We presume that certain environmental features require similar protection. 

Again, this form of interpretation is not feasible for overflow bottomland locations. 

5. Canoe and Foot Trails. Evidently some marked trails of these kinds are planned 

for the refuge, and will intersect points of interest in upland and floodplain areas. Some 

of the themes enumerated above could be indicated, but not explicated, by markers, e.g., 

“The explorers Wm. Dunbar and George Hunter camped near here on November 16, 

1804.” The general presence and nature of prehistoric floodplain sites (but not specific 

locations), which we have investigated in this report, could be indicated by markers 

along riverbanks or bayous, e.g., “Numerous prehistoric camps of hunters and fishermen 

have been located by archeologists along the lower Saline River.” Such markers 

need to be vandal-proof and resistant to the effects of annual flooding. 

6. Popular Publications, Brochures, or Leaflets. All of the themes suggested by us 

can be made available to the public in this form, but without the impact of exhibits discussed 

above. We recommend that the array of important scientific information from our 

project and other work be made availabe in popular published form, covering unique 

cultural and ecological characteristics of the lower Ouachita Valley in Arkansas. 

generAl recoMMendAtions 

The following recommendations are not specific to the Felsenthal Project area, and 

are not properly a part of our program of mitigation as required by contract with the 

U.S. Army Corps of Engineers.

C-30 

1. We recommend that a comprehensive management plan for cultural resources 

in the Felsenthal National Wildlife Refuge be prepared by the U.S. Fish and Wildlife 

Service with the assistance and coordination of the U.S. Army Corps of Engineers, Vicksburg 

District. Such a management plan would require more extensive site survey and 

site assessment in certain areas than has now been accomplished. These areas should 

include Goulett Island, Prairie Island, Horton Island, Pine Island, Cedar Ridge, and adjacent 

floodplain areas at about 70 feet elevation in the northern portion of the refuge. The 

management plan should also summarize and appraise the results of the several large 

scale and small scale survey and testing projects which have been performed within refuge 

boundaries to date. We suggest that the Multiple Resource Area concept, which was 

proposed for significant floodplain sites in this report, can be expanded to other terrain 

or all terrain within the refuge. Comprehensive inventory and management of cultural 

resources in the Felsenthal National Wildlife Refuge would protect one of the richest and 

least disturbed archeological zones in the United States. 

2. We recommend that federal agencies responsible for terrain and waterways 

in the lower Ouachita River Valley in Arkansas consider the acquisition of significant 

cultural properties just outside refuge boundaries, and privately owned. Among these 

are Big Mound Ridge (3AS6) and Sulphur Springs Mound (3AS1) which are about 3.5 

and 8 km respectively east of the new lock and dam (Appendix A). These major mound 

sites are within timber lands and may be threatened by total destruction in the next few 

years.

eferences cited 

C-31 

Advisory Council on Historic Preservation 

1980 Treatment of Archeological Properties: A Handbook. Advisory Council on Historic 

Preservation, Washington D.C. 

Carrell, Toni, Sandra Rayl, and Daniel Lenihan 

1976 The Effects of Freshwater Inundation of Archeological Sites Through Reservoir Construction: 

A Literature Search. National Park Service, U.S. Department of the Interior. 

Clarke, David L. 

1968 Analytical Archaeology. Methuen, London. 

Garrison, Ervan G. 

1975 A Qualitative Model for Inundation Studies in Archeological Research and Resource 

Conservation: An Example from Arkansas. Plains Anthropologist 20:279-296. 

1977 Modeling Inundation Effects for Planning and Prediction. In Conservation 

Archaeology: A Guide for Cultural Resource Management Studies, edited by Michael B. 

Schiffer and George J. Gumerman, pp. 151-156. Academic Press, New York. 

Garrison, Ervan G., Alan May, Jeffrey Newsom, and Alf Sjoberg 

1977 Progress Report on the Effects of Inundation on Cultural Resources: Table 

Rock Reservoir, Missouri. Ms. on file, Department of Anthropology, University of 

Missouri, Columbia. 

Lenihan, Daniel J., Toni L. Carrell, Thomas S. Hopkins, A. Wayne Prokopetz, Sandra L. 

Rayl, and Cathryn S. Tarasovic 

1977 The Preliminary Report of the National Reservoir Inundation Study. National Park 

Service, U.S. Department of the Interior. 

Lipe, W. D. 

1974 A Conservation Model for American Archaeology. Kiva 39(4):213-245. 

National Park Service 

1977 How to Complete National Register Multiple Resource Nomination Forms: 

Interim Guidelines. National Park Service, U.S. Department of the Interior. 

Schiffer, Michael B., and George J. Gumerman (editors) 

1977 Conservation Archaeology: A Guide for Cultural Resource Management Studies. 

Academic Press, New York. 

Schiffer, Michael B., and John H. House 

1977 An Approach to Assessing Scientific Significance. In Conservation Archaeology: 

A Guide for Cultural Resource Management Studies, edited by Michael B. Schiffer and 

George J. Gumerman, pp. 249-257. Academic Press, New York.

C-32 

Schnell, Frank, and Jack Tyler 

1976 Hydrology and Archaeological Site Conservation. Southeast Archaeological Conference 

Bulletin 19:37-38.

RS 17 Human Adaptation in the Grand Marais - University of Arkansas

Create successful ePaper yourself

Delete template?

Save as template?