Pluckhahn, Thomas J. 2011Households Making History: Household Change in the Late Woodland Period at Kolomoki (9ER1). Department of Anthropology, University of South Florida, Tampa.

Households Making History: Household Change in the Late Woodland Period at Kolomoki (9ER1)

edited by

Thomas J. Pluckhahn

Households Making History: Household Change in the Late Woodland Period at Kolomoki (9ER1)

Prepared for:

Georgia Department of Natural Resources34 Peachtree Street, NW

Suite 1600Atlanta, GA 30303

edited by

__________________Thomas J. Pluckhahn, Ph.D.

Principal Investigator

Department of AnthropologyThe University of South Florida4202 E. Fowler Ave, SOC 107

Tampa, FL 33620

August 29, 2011

Abstract

This report describes recent archaeological investigations of the Kolomoki site (9ER1) in EarlyCounty, Georgia. Kolomoki, one of the largest and most famous sites of the Woodland period, iscomprised of at least nine mounds. It is preserved as part of Kolomoki Mounds State Park and HistoricSite, managed by the Georgia Department of Natural Resources.

The project was sponsored by grants from the National Geographic Society and the HumanitiesInstitute at the University of South Florida (USF). Institutional support was provided by the USFDepartment of Anthropology and the Georgia Department of Natural Resources. This was not aSection 106 project, since no federal monies were expended on the research and no federal permits wererequired. The fieldwork for this project was conducted over the course of three field seasons from2006-2008 under the direction of Principal Investigator Dr. Thomas Pluckhahn. The field crewincluded volunteers and students from the University of South Florida, Florida Gulf Coast University,and Florida State University.

The goal of the study was to identify changes in households over the course of the LateWoodland period (ca. cal A.D. 400 to 1050). Previous excavations at the site had identified a householdfrom the early/middle Late Woodland (A.D. 400 to 600) period in area referred to as “Block A”(Pluckhahn 2003). The present research was directed toward the excavation of a household from thelate or terminal Late Woodland (ca. cal A.D. 600 to 1050), to complement the previous excavations. These excavations, in an area referred to as “Block D,” provided the data necessary to examine changesin household size, form, and economy over the course of the Late Woodland period.

Block D was located just south of the main park road, on a gentle, south- and southeast-facingslope above an active spring shovel tests and a test unit (Test Unit 18) excavated in this area in thecourse of previous research revealed higher concentrations of pottery, along with a buried A horizonand a high density of features (Pluckhahn 2003:110-125). Pottery recovered from this area includedrelatively high proportions of conspicuous Late Woodland types such as Indian Pass Incised, WeedenIsland Incised, and Napier Complicated Stamped. Including the previously excavated 2-x-2-m Test Unit18, we completed a total of 52 m2 of excavation in Block D. Of this total, 38 m2 were contiguous 1-x-1m units that together form a block about 8 m long (north-south) and 6 m wide (east-west).

Excavation of Block D (including Test Unit 18) resulted in the identification of 87 features fora density of approximately 1.7 features/m2. Smaller features representing post molds or small pits werethe most common feature type, with 48 examples identified. The 38 larger pit features included 18basin-shaped pits and 10 bell-shaped pits. These larger features had a combined volume of 5176 liters,demonstrating a considerable increase in storage associated with the late/terminal Late Woodlandperiod.

Feature patterning possibly indicative of a domestic structure was noted during the course ofexcavating Block D. The pattern is defined mainly by the distribution of post mold or small pit featuresin an oval pattern measuring about 7.3 m long and 5.2 m wide. The pattern is interpreted as the remainsof a house of single set post construction. Carbon dates from Block D, reported in greater detail inChapter 7, place the occupation of the house in the late or terminal Late Woodland period, primarilyfrom around cal A.D. 650 to 850 but perhaps continuing as late as 990. The presumed structures isconsistent with architectural patterns noted on other site sin the area dating to the same time period(Mickwee 2009; Milanich 1974). It is larger and of different construction than the structure identifiedin Block A at Kolomoki (Pluckhahn 2003).

i

Prehistoric ceramics are the most common artifact type represented in the assemblage fromBlock D. Including the previously excavated Test Unit 18, the pottery collection includes 21,637 sherds All of the pottery in the assemblage is associated with the Woodland period occupation of the BlockD area, with the exception of 17 sherds that appear to correspond more closely with Mississippian types.Excluding these and sherds too small to identify with confidence, the total identifiable Woodlandpottery assemblage from Block D consists of 9272 sherds. Plain wares—primarily sand temperedplain—predominate, making up a little more than three-quarters of the assemblage. Swift CreekComplicated Stamped is the most common named decorative type, but here makes up only about 10percent of the collection, compared with as much as 25 percent in earlier phases at Kolomoki. Lesscommon complicated stamped types include Napier and St. Andrews. Other pottery types of theWeeden Island series combine to form almost one-tenth of the Woodland pottery assemblage. Theseinclude, in descending order of frequency: Weeden Island Red, Carrabelle Punctate, Weeden IslandIncised, Mound Field Net Marked, Weeden Island Zoned Incised/Punctate, Keith Incised, Indian PassIncised, Carrabelle Incised, Wakulla Check Stamped, and Tucker Ridge Pinched.

Excavation of Block D resulted in the recovery of 6693 flaked stone artifacts. Debitage—broadly defined to include flakes and shatter/chunks—forms the majority of the flakedstone assemblage. The assemblage includes 84 bifaces or biface fragments, 38 of which retain enoughof the proximal hafting area element and display significant diagnostic attributes to be assigned toclassified to morphological clusters and named types. Perhaps most surprising was the recovery of fiveWoodland/Mississippian triangulars, a type not encountered in previous excavations at Kolomoki(Pluckhahn 2003).

Faunal and botanical remains from Block D were limited by poor preservation. Zooarchaeological analyses revealed an emphasis on terrestrial species, particularly deer. Macrobotanicalanalyses suggested a continued reliance on mast resources, with limited evidence for cultivation ofdomesticated crops. No maize macrobotanical remains were recovered from Block D, but we did findevidence for maize in the microbotanical remains. Microbotanical remains also revealed evidence forthe use of roots of an unidentified species of the arrowroot family.

Comparison of the archaeological households in Blocks A and D reveals changes in temporalchanges in households from the early/middle Late Woodland to the late/terminal Late Woodland. Most conspicuous are changes in the size and construction of houses, from the very small pit house inBlock A to the larger house of single-set post construction in Block D. Storage increased dramatically,as did the size of ceramic vessels. There appear to have been no major shifts in subsistence, but flakedstone technology changed with the adoption of the bow and arrow, from communal to moreindividualized hunting. In concert, the evidence suggests that late Late Woodland households exercisedgreater autonomy over domestic production and consumption.

ii

Acknowledgments

The successful completion of this project owes much to the help of a number of individuals andorganizations. First and foremost, I thank David Crass and Robert Entorf at the Georgia Departmentof Natural Resources for granting permission to work on Kolomoki Mounds State Historic Park; it hasbeen a privilege and a pleasure.

As always, our work at Kolomoki was greatly facilitated by the staff of Kolomoki State HistoricPark. I am particularly indebted to Park Manager Matt Bruner, as well as Billy Adams, Jason Harrison,Judy Moore, and Sid Sewell. The folks at Kolomoki have always made me and my students feelwelcome at Kolomoki, and have done much to facilitate this and previous research.

Grateful appreciation is also extended to the sponsors of the testing. This work was supported,in large part, by an Explorer Grant from the National Geographic Society. Additional support wasprovided by a grant from the Humanities Institute at the University of South Florida. Institutionalsupport was provided by the Department of Anthropology of the University of South Florida and theGeorgia Department of Natural Resources.

Finally, I am indebted to the students and volunteers who have participated in the field and labfor their hard work and for making our time at Kolomoki a fun and rewarding experience. Their namesare enumerated later in this report.

iii

iv

Table of Contents

Abstract . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . I

Acknowledgments . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . iii

Figures . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . vii

Tables . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . xi

Chapter 1: Introduction (Thomas J. Pluckhahn) . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 1

Research Design . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 2Environmental Setting . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 8

Chapter 2: Methods (Thomas J. Pluckhahn, Mary Theresa Bonhage-Freund, Matthew C. Compton, Linda Scott-Cummings, and Chad Yost) . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 11

Field Methods . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 11Laboratory Methods . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 13Curation . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 15

Chapter 3: Excavation Units (Thomas J. Pluckhahn) . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 23

Excavation Strategy . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 23Stratigraphy . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 25Artifacts . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 32

Chapter 4: Features (Thomas J. Pluckhahn) . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 51

Smaller Features: Postmolds and Small Pits . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 54Larger Pit Features . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 65Summary . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 88

Chapter 5: Artifacts (Thomas J. Pluckhahn) . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 91

Prehistoric Ceramics . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 91Prehistoric Flaked Stone . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 123Miscellaneous Prehistoric Artifacts . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 136Historic Artifacts . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 138Summary . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 139

Chapter 6: Faunal and Botanical Remains (Mary Theresa Bonhage-Freund, Matthew C. Compton,Linda Scott Cummings, Chad Yost, and Thomas J. Pluckhahn) . . . . . . . . . . . . . . . . . . . . . . . . . . 141

Faunal Remains . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 141Macrobotanical Remains . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 147Microbotanical Remains . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 170Summary . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 178

v

Table of Contents (Continued)

Chapter 7: Households Making History: Comparing Late Woodland Households in Blocks A andD(Thomas J. Pluckhahn) . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 179

Periods of Occupation . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 179Domestic Architecture . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 182Storage . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 188Subsistence and Environment . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 190Ceramics . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 197Flaked Stone . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 200Discussion . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 203Conclusion . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 208

References Cited . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 211

vi

Figures

Figure 1-1. Location of Kolomoki and other sites mentioned in the text . . . . . . . . . . . . . . . . . . 2Figure 1-2. Map of Kolomoki showing the locations of Blocks A and D . . . . . . . . . . . . . . . . . . 3Figure 1-3. Oblique aerial view of Kolomoki . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 9Figure 1-4. Topography of the Block D area . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 10Figure 2-1 The 2006 field crew at Kolomoki . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 11Figure 2-2 The 2007 field crew at Kolomoki . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 12Figure 2-3. The 2008 field crew at Kolomoki . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 12Figure 2-4. Bisection and profile mapping of a feature in Block D . . . . . . . . . . . . . . . . . . . . . . 13Figure 2-5. Flotation machine used for samples collected in the 2007

and 2008 field seasons . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 14Figure 2-6. Hafted biface measurement dimensions (adapted from

Andrefsky 1998:179) . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 15Figure 3-1. View to the southwest of the excavation of Block D . . . . . . . . . . . . . . . . . . . . . . . 23Figure 3-2. Map of Block D showing unit numbers and year of excavation . . . . . . . . . . . . . . . 24Figure 3-3. Sears’s (1956:Figure I) depiction of a buried midden deposit at Kolomoki . . . . . . 25Figure 3-4. Northern (top) and southern (bottom) profiles of the core group of contiguous

excavation units in Block D . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 26Figure 3-5. Plan view of Unit D9 at base of Stratum II, showing probable plow scars . . . . . . 27Figure 3-6. The West Profile of Units D3 and D4 . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 28Figure 3-7. The East Profile of Test Unit 18 . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 29Figure 3-8. The West Profile of Unit D1 . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 30Figure 3-9. The South Profile of Unit D5 . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 31Figure 3-10. The density of pottery in the core units of Block D . . . . . . . . . . . . . . . . . . . . . . . . 32Figure 3-11. The density of Swift Creek in the core units of Block D . . . . . . . . . . . . . . . . . . . . . 32Figure 3-12. The density of Weeden Island Red in the core units of Block D . . . . . . . . . . . . . . 39Figure 3-13. The density of Carrabelle Punctate in the core units of Block D . . . . . . . . . . . . . . 39Figure 3-14. The density of Lamar in the core units of Block D . . . . . . . . . . . . . . . . . . . . . . . . . 39Figure 3-15. The density of flaked stone in the core units of Block D . . . . . . . . . . . . . . . . . . . . 40Figure 3-16. The density of quartzite/sandstone in the core units of Block D . . . . . . . . . . . . . . 40Figure 3-17. The density of FCR in the core units of Block D . . . . . . . . . . . . . . . . . . . . . . . . . . 46Figure 3-18. The density of groundstone the core units of Block D . . . . . . . . . . . . . . . . . . . . . . 46Figure 3-19. The density of historic artifacts in the core units of Block D . . . . . . . . . . . . . . . . . 49Figure 4-1. Excavation of features in Block D, view to the northwest . . . . . . . . . . . . . . . . . . . 51Figure 4-2. Features in Block D . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 52Figure 4-3. Closer view of features in the core area of Block D . . . . . . . . . . . . . . . . . . . . . . . . 53Figure 4-4. Histogram of feature area . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 54Figure 4-5. Excavation of a postmold/small pit (Feature 151), view to the west . . . . . . . . . . . 54Figure 4-6. Profiles of postmolds and small pit features . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 56Figure 4-7. Outline of possible structure in Block D . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 59Figure 4-8. Depth of postmolds and small pit features in Block D . . . . . . . . . . . . . . . . . . . . . . 60Figure 4-9. The density of flaked stone in postmolds and small pit features in Block D . . . . . 64Figure 4-10. Larger bell- and basin-shaped pit features in Block D . . . . . . . . . . . . . . . . . . . . . . . 67Figure 4-11. The density of flaked stone in the larger pit features in the core area

of Block D . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 71Figure 4-12. The density of other modified stone (ground and pecked stone and FCR) in the larger

pit features in the core area of Block D . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 72Figure 4-13. The density of prehistoric ceramics in the larger pit features in the core

area of Block D . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 73Figure 4-14. The density of bone in the larger pit features in the core area of Block D . . . . . . . 74

vii

Figures (Continued)

Figure 4-15. The East Profile of Feature 192, a typical small basin-shaped pit . . . . . . . . . . . . . . 75Figure 4-16. Profiles of large, basin-shaped pit features . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 76Figure 4-17. Excavation of Feature 191A, view to the south . . . . . . . . . . . . . . . . . . . . . . . . . . . . 79Figure 4-18. Three views of Feature 171, a large basin-shaped pit . . . . . . . . . . . . . . . . . . . . . . . 80Figure 4-19. Three views of Feature 141A, B, and C . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 81Figure 4-20. Four views of Features 147A and 147B . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 82Figure 4-21. Feature 165, view to the north . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 83Figure 4-22. Profiles of large, bell-shaped pit features . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 84Figure 4-23. The North Profile of Feature 185 . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 86Figure 4-24. The North Profile of Feature 155 . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 87Figure 4-25. The East Profile of Feature 175 . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 88Figure 4-26. The South Profile of Feature 163 . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 88Figure 5-1. Profiles of vessels identified in MNV Analysis of Block D . . . . . . . . . . . . . . . . . . . 95Figure 5-2. Paddle Design D-1 (A) and sherds used in its reconstruction (B-E) . . . . . . . . . . 102Figure 5-3. Paddle Design D-6 (A) with sherds used in its reconstruction (B-D) and

potentially matching paddle-stamped sherds (E-G) . . . . . . . . . . . . . . . . . . . . . . . . 103Figure 5-4. Paddle Design D-3 (top) and sherd from Feature 171 used in its

reconstruction (bottom) . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 104Figure 5-5. Paddle Design D-4 (A) and sherds used in its reconstruction . . . . . . . . . . . . . . . . 105Figure 5-6. Paddle Design D-2 and sherds from Feature 171 used in its reconstruction . . . . 106Figure 5-7. Paddle Design D-5 and sherds from Feature 171 used in its reconstruction . . . . 107Figure 5-8. The distribution of Swift Creek pottery and paddle matches in pit features

in the core area of Block D . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 108Figure 5-9. The distribution of Napier pottery in pit features in the core area of Block D . . 110Figure 5-10. Selected Napier sherds from Block D . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 111Figure 5-11. Selected St. Andrews Complicated Stamped sherds from Block D . . . . . . . . . . . 112Figure 5-12. Selected Weeden Island Red and Zoned Red sherds from Block D . . . . . . . . . . . 113Figure 5-13. The distribution of Weeden Island Incised and Zoned Incised/Punctate

pottery in pit features in the core area of Block D . . . . . . . . . . . . . . . . . . . . . . . . . 115Figure 5-14. Selected Weeden Island Incised and Zoned Incised/Punctate sherds from

Block D . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 116Figure 5-15. Selected Carrabelle Punctate sherds from Block D . . . . . . . . . . . . . . . . . . . . . . . . 117Figure 5-16. Selected Carrabelle Incised sherds from Block D . . . . . . . . . . . . . . . . . . . . . . . . . 119Figure 5-17. Selected Keith Incised sherds from Block D . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 120Figure 5-18. Selected Indian Pass Incised sherds from Block D . . . . . . . . . . . . . . . . . . . . . . . . 121Figure 5-19. Selected other Weeden Island series sherds from Block D . . . . . . . . . . . . . . . . . . 122Figure 5-20. Ridge and Valley chert cobble/core (A) and Tallahatta Sandstone

cores (B-C) from Block D . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 124Figure 5-21. Selected chert cores from Block D . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 125Figure 5-22. Selected quartz cores and core-like artifacts from Block D . . . . . . . . . . . . . . . . . . 126Figure 5-23. The distribution of quartzite/sandstone in larger pit features in the core

area of Block D . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 127Figure 5-24. Comparison of the relative frequencies of debitage and tools by raw materials . . 128Figure 5-25. Selected flake tools from Block D . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 129Figure 5-26. Selected hafted bifaces from Block D . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 132Figure 5-27. Selected historic artifacts from Block D . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 139Figure 6-1. Number (count) of hickory, acorn (x 50) and seeds per liter of soil analyzed . . . 168Figure 6-2. Pollen diagram for microbotanical remains from Block D . . . . . . . . . . . . . . . . . . 170Figure 6-3. Phytolith diagram for microbotanical remains from Block D . . . . . . . . . . . . . . . . 173Figure 6-4. Selected phytoliths from Block D samples . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 174Figure 6-5. Comparison of archaeological (Block D) and reference Marantaceae phytoliths . 176

viii

Figures (Continued)

Figure 7-1. Plot of radiocarbon dates from Blocks A and D showing two sigmacalibrated ranges . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 180

Figure 7-2. Comparison of early/middle Late Woodland structure from Block A(left) and late/terminal Late Woodland structure from Block D (right) . . . . . . . . 183

Figure 7-3. Comparison of early/middle Late Woodland structures from the Catoma Creeksite (left) (after Shelby 2011:Figure 2) and Block A at Kolomoki (right) . . . . . . . 185

Figure 7-4. Comparison of late/terminal Late Woodland structures from the Sycamoresite (left) (after Milanich 1974:Figure 7) and Block D at Kolomoki (right) . . . . . 186

Figure 7-5. Comparison of floor area of structures in Blocks A and D . . . . . . . . . . . . . . . . . . 187Figure 7-6. Comparison of the density of pit features in Blocks A and D . . . . . . . . . . . . . . . 188Figure 7-7. Comparison of the volume of pit features in Blocks A and D . . . . . . . . . . . . . . . 188Figure 7-8. Comparison of the relative frequencies of raw materials in the flaked stone assemblages

from Blocks A and D . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 201Figure 7-9. Comparison of the relative frequencies of early and late stage debitage

and tools in the flaked stone assemblages from Blocks A and D . . . . . . . . . . . . . 201Figure 7-10. Comparison of the relative frequencies of morphological clusters

represented in the hafted biface assemblages from Blocks A and D . . . . . . . . . . 202

ix

x

Tables

Table 2.1. Vessel Form Categories Used in this MNV Analysis and Their Equivalentsin the Analyses of Hally (2009) and Wallis (2011) . . . . . . . . . . . . . . . . . . . . . . . . . . 16

Table 3-1. Prehistoric Ceramics from Unit Levels . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 33Table 3-2. Flaked Stone from Unit Levels . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 41Table 3-3. Miscellaneous Prehistoric Artifacts from Unit Levels . . . . . . . . . . . . . . . . . . . . . . . 47Table 3-4. Historic Artifacts from Unit Levels . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 49Table 4-1. Summary Data for Post Molds and Small Pit Features . . . . . . . . . . . . . . . . . . . . . . 55Table 4-2. Prehistoric Ceramics from Post Molds and Small Pit Features . . . . . . . . . . . . . . . . 61Table 4-3. Flaked Stone from Post Molds and Small Pit Features . . . . . . . . . . . . . . . . . . . . . . 62Table 4-4. Miscellaneous Artifacts from Post Molds and Small Pit Features . . . . . . . . . . . . . . 63Table 4-5. Summary Data for Larger Pit Features . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 66Table 4-6. Prehistoric Ceramics from Larger Pit Features . . . . . . . . . . . . . . . . . . . . . . . . . . . . 68Table 4-7. Flaked Stone from Larger Pit Features . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 69Table 4-8. Miscellaneous Artifacts from Larger Pit Features . . . . . . . . . . . . . . . . . . . . . . . . . . 70Table 5-1. Prehistoric Ceramics from Block D . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 92Table 5-2. Woodland Period Ceramics from Block D . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 93Table 5-3. Results of MNV Analysis of Ceramics from Block D . . . . . . . . . . . . . . . . . . . . . . . 94Table 5-4. Flaked Stone from Block D . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 123Table 5-5. Type and Metric Data for Hafted Bifaces from Block D . . . . . . . . . . . . . . . . . . . 131Table 5-6. Miscellaneous Prehistoric Artifacts from Block D . . . . . . . . . . . . . . . . . . . . . . . . 137Table 5-7. Historic Artifacts from Block D . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 138Table 6-1. Species List for Faunal Remains from Block D . . . . . . . . . . . . . . . . . . . . . . . . . . . 141Table 6-2. Proveniences in Block D with Faunal Remains and the Taxa Represented . . . . . 142Table 6-3. Summary of Taxa Identified in the Analysis of Faunal Remains from Block D . 144Table 6-4. Epiphyseal Fusion Noted on White-Tailed Deer Remains from Block D . . . . . . 146Table 6-5. Element Distribution for White-Tailed Deer in the Block D Faunal

Assemblage . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 146Table 6-6. Modifications of Bone Noted in the Block D Faunal Assemblage . . . . . . . . . . . . 147Table 6-7. Summary Data for Flotation Samples and Fractions from Block D . . . . . . . . . . . 148Table 6-8. Seeds (Counts) in Macrobotanical Remains from Block D . . . . . . . . . . . . . . . . . . 149Table 6-9. Nutshell/Nutmeat (Counts and Weights) in Macrobotanical Remains from

Block D . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 154Table 6-10. Wood Charcoal and Miscellaneous (Counts) in Macrobotanical Remains from

Block D . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 158Table 6-11. Wood Charcoal and Miscellaneous (g) in Macrobotanical Remains from

Block D . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 162Table 6-12. Habitats and Seasonality Associated with Macroplant Remains from Block D . . 169Table 6-13. Pollen and Starch Types Observed in Microbotanical Remains from Block D . . 171Table 6-14. Background Information for Plant Taxa Identified in Microbotanical

Remains from Block D . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 172Table 7-1. Radiocarbon Dates from Blocks A and D at Kolomoki . . . . . . . . . . . . . . . . . . . . 181Table 7-2. Comparison of the Faunal Assemblages from Blocks A and D . . . . . . . . . . . . . . 191Table 7-3. Comparison of Macrobotanical Assemblages from Blocks A and D . . . . . . . . . . 193Table 7-4. Comparison of Wood Charcoal Assemblages from Blocks A and D . . . . . . . . . . 196Table 7-5. Relative Frequencies of Surface Treatments in Identifiable Woodland Pottery

Assemblages from Blocks A and D . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 197Table 7-6. Relative Frequencies of Vessel Forms Identified in MNV Analysis of

Blocks A and D . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 198Table 7-7. Comparison of Mean Orifice Diameter for Vessel Forms in Blocks A and D . . 199Table 7-8. Comparison of the Relative Frequencies of Hafted Biface Types in Blocks

A and D . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 203

xi

xii

Chapter 1: Introduction

Thomas J. Pluckhahn

This report describes archaeological research aimed at understanding social change during theLate Woodland period (A.D. 400 to 1050), a critical juncture in the development of complex societiesin the southeastern United States. The focus of the study is the archaeological site of Kolomoki, in thelower Chattahoochee River valley of southwestern Georgia (Figure 1). Kolomoki, one of the largest andmost famous sites of the Woodland period, is comprised of at least nine mounds (Pluckhahn 2003). The site is preserved as part of Kolomoki Mounds State Park and Historic Site, managed by the GeorgiaDepartment of Natural Resources.

This study builds on previous research at Kolomoki. In 2000, the National Geographic Societyawarded funding to Stephen Kowalewski for his proposal Kolomoki: Learning About a Woodland CeremonialCenter. This research formed the core of a dissertation by Pluckhahn (2002), subsequently published bythe University of Alabama Press as Kolomoki: Settlement, Ceremony, and Status in the Deep South, A.D. 350to 750 (Pluckhahn 2003). Carbon dates were taken to develop a revised chronology for the site.Systematic sampling, augmented by test excavations and geophysical prospection, successfully definedchanges in the village plan over the course of the more than four hundred years of occupation. Briefly,during the Kolomoki I and II phases (from around cal A.D. 350 to 550) of the late Middle andearly/middle Late Woodland periods the occupation took the form of a large, formally-defined circularvillage centered on an immense plaza. In the Kolomoki III and IV phases (ca. cal A.D. 550 to 750) ofthe late Late Woodland period this formal village plan broke down, and households were widelydispersed. The shift in settlement plan was coincident with a breakdown in mound construction andceremony.

Although the previous study greatly improved our understanding of Kolomoki, a number ofimportant questions remain unanswered. Perhaps most intriguing is the way in which the broader socialchanges evident in settlement patterns played out at the household level. The final stages of theprevious study included the excavation of a small early/middle Late Woodland pit house in an area ofthe site designated “Block A” (see Figure 2). Limited test excavations in portions of the site that wereoccupied later in the Late Woodland suggest important changes in households took place in the courseof Kolomoki’s history, including a switch to larger, above-ground structures, and a privatization ofstorage facilities. However, due to time and budget constraints, excavations in later Late Woodlandoccupation areas were limited and the results are inconclusive. The present study remedied this throughan excavation of a household from the later phases of occupation at Kolomoki. This excavation, in anarea referred to as “Block D,” provided the data necessary to examine changes in household size,structure, and economy over the course of the Late Woodland period.

The project was sponsored by grants from the National Geographic Society and the HumanitiesInstitute at the University of South Florida (USF). Institutional support was provided by the USFDepartment of Anthropology and the Georgia Department of Natural Resources. This was not aSection 106 project, since no federal monies were expended on the research and no federal permits wererequired. The fieldwork for this project was conducted over the course of three field seasons between2006 and 2008. The field crew included volunteers and students from the University of South Florida,Florida Gulf Coast University, and Florida State University.

This chapter provides context for the project, including a review the research design andenvironmental setting. The methods that were employed on the project are described in Chapter 2. Theresults of the excavation are presented in Chapters 3 and 4. Chapter 5 presents an overview of theartifact assemblage, while Chapter 6 does the same for the faunal and botanical assemblages. Finally,Chapter 7 draws comparisons between the Late Woodland household excavated for this study and thepreviously excavated, earlier household.

1

Research Design

Lost Households of the Late Woodland

To what extent do households affect historical change? In her “subversive” account of one ofthe world’s great historical transformations—the origins of the modern world—the historian Mary S.Hartman (2004) has argued that a form of household relations particular to northwestern Europe priorto the modern era provided a “prior and distinctive development” that permitted the rise of capitalismand the origins of the nation-state after 1500. Hartman points to one characteristic of medievalnorthwestern European households in particular: the decision by many peasant women to postponemarriage in favor of a period of domestic servitude. This led to greater parity in household decision-making among men and women, permitted greater mobility, fostered more innovation, and allowed forgreater accumulation of capital.

Figure 1-1. Location of Kolomoki and other sites mentioned in the text.

2

Block A

Block D

!Kolomoki

0 500

Meterscontour interval = 1 m

(elevations from topographicsurvey and historical maps)

N

Figure 1-2. Map of Kolomoki showing the locations of Blocks A and D.

3

Hartman’s explanation stands in contrast to the “disembodied historical forces” historians havelong favored in their explanation of this and other major structural transformations (Hartman 2004:210,242), a tendency not limited to historians. Despite more than three decades of scholarly interest inarchaeological households, and more than two decades of archaeological studies invoking theimportance of agency in explaining the past, archaeologists have too seldom been willing or able to grantagency to households and their constituents in historical transformations.

The problem is exemplified by archaeological interpretations of one of the great junctures in theprehistory of southeastern North America: the Late Woodland period. On one side of this juncturestood the complex, but relatively egalitarian societies of the Middle Woodland period (300 B.C. to A.D.400) and on the other the hierarchically-organized societies of the Mississippian (A.D. 1050 to 1500). It is generally acknowledged that three major transformations in the Late Woodland contributed to thistransition. As described McElrath and colleagues (2000:23), these consist of:

1) a major population resettlement following the breakup of the Hopewell InteractionSphere and the general decline of the Middle Woodland lifestyle; 2) the widespreadadoption of the bow and arrow; and 3) the adoption of a maize-based agriculturaleconomy...

All of these changes clearly entail fundamental shifts in the organization of the domestic economy,but—with certain exceptions discussed in more detail below—they have rarely been attributed to theagency of households. The otherwise exhaustive survey of Late Woodland societies in the Midwest byEmerson and colleagues (2000), for example, contains only 15 references to the term “household” inits 26 separately authored chapters and more than 700 pages of text. The seminal edited volume on theLate Woodland Southeast by Nassaney and Cobb (1991), as well as the more recent overview of theWoodland Southeast by Anderson and Mainfort (2002), contain still fewer references to households.

There are practical reasons for the lack of attention granted to Late Woodland households. First, the Late Woodland period was omitted from many of the original cultural historical chronologiesfor the region, and earned only slight consideration in many of the second-generation syntheses. Griffin(1952:361-362), for example, made only passing reference to this as “...a period of rest and quiescence.” In some areas of the Southeast, archaeological understanding of the Late Woodland period hascontinued to suffer from the resulting lack of clear diagnostic markers and artifacts sequences (e.g.,Rudolph 1999).

More important, relatively few Late Woodland houses have been excavated (Steere 2011:79),perhaps because in many areas residential mobility was high and houses were lightly constructed, makingthe identification of archaeological households less secure. Cobb and Nassaney (2002:538-539), basedon a perceived lack of substantial houses or planned communities during the Middle and Late Woodlandperiods, argue that the “institutionalization” of domestic space—presumably including household-basedproduction and consumption—did not occur until the subsequent Mississippian period.

There are also methodological and conceptual issues that transcend the archaeology of the LateWoodland period. Households emerged as a topic of study among archaeologists within the context ofthe settlement pattern studies of the 1970s (e.g., Flannery 1976); early treatments fit squarely in theprocessual paradigm, with households viewed as basic building blocks of larger social formations, aspoints of articulation between societies and economic and ecological processes, and as windows onevolutionary change (Gerritsen 2004:142; Sabloff and Ashmore 2001:22; for reviews of householdarchaeology, see Nash 2009; Pluckhahn 2010b; Robin 2003). As Pauketat (2000; 2007:45-46) hasargued, this approach reduced households to “static and uniform organizational units.”

One might expect such conceptual problems to have been alleviated with the turn to agency inarchaeology within the last two decades, and there have certainly been keen advances in theunderstanding of households as divergent, internally-divided, and externally-connected social formations(e.g., Hendon 1996; Marcoux 2008; Rodning 2004; Rogers 1995; Souvatzi 2008; Wilson 2008). Yet the

4

problem persists because, as Johnson (2006:123) notes, many of archaeology’s most basic classificatoryconcepts—cultures, phases, and types—undermine the visibility of agency in the archaeological recordin their emphasis on similarity rather than variability. The problem is exacerbated by the conventionsof archaeological site reporting, which obscure agency through the use of generic categories of materialremains (Johnson 2006:124-125; see also Spector 1993).

Hartman (2006:31) suggests that the failure to grant greater agency to households may stemfrom a conceptual bind much broader than the terms of archaeological classification and discourse. Sheattributes the problem to the contemporary social milieu in which our research takes place:

Accustomed to seeing their own households as embattled and weakened, many woulddismiss the suggestion that there was ever a time when typical household settingsexercised immense influence, not only in structuring women’s and men’s daily lives butalso in generating lasting change beyond households and in setting crucial conditionson the nature of that change. In the contemporary world, after all, we are used to theidea that the arrow of change always moves from institutions back to households, thathouseholds are always reactors to outside developments and never places from whichfar-reaching transformation might emanate.

I would extend Hartman’s argument further: the denial of household agency arises not only from ourview of contemporary social institutions but also from the manner in which we conceive of thetemporality of history. Specifically, our tendency to emphasize long-term institutions and structures asthe determinants of historical change reduces the relevance of shorter term cycles and events (Sewell2005), including shorter-term social formations such as households. This tendency, most conspicuousin works that assume a universality of causality (be they marxist, evolutionary, or adaptionist), isunfortunate because as Gerritsen (2004:143) notes, the household potentially offers a “theoreticallyinformed counterweight” to the sort of large-scale systems and processes that are frequently invokedby archaeologists to explain social and cultural change—grand narratives that by definition refer totemporal and spatial scales largely meaningless to the people involved in those changes (see also Robin2003:308; Wilson 2008:8). Recognizing the agency of households may require adopting a more“eventful temporality” that recognizes the power of events in the process of historical change (Sewell2005), a perspective I review below.

Late Woodland Households and Mississippianization

Although archaeological study of the Late Woodland period remains relatively poorly developedfor much of the Southeast, several recent studies have granted greater attention to the role of LateWoodland households, specifically as they relate to the process of Mississippianization. Such researchis particularly active in the vicinity of Cahokia in the American Bottom region where the FAI-270excavations (Kelly 1990b; Kelly et al. 1987) were extensive enough to result in the identification ofnumerous houses and whole community patterns. Small semi-subterranean “keyhole” structuresbecame common in the American Bottom by the Patrick phase (A.D. 600 to 700) (Kelly 1990b). Thesekeyhole structures have been found arranged in clusters (Kelly 1990b; Kelly et al. 1987), which Peregrine(1992) interprets as evidence for the emergence of lineage compounds—several extended families fromthe same lineage functioning as a joint economic unit, and thus possibly functioning as a single, largehousehold. By the Early Mississippian period, larger rectangular houses were arranged linearly in villages(Kelly 1990a, 1990b; Kelly et al. 1989; see also Steere 2011), a pattern taken to represent the emergenceof individual nuclear or extended families as the basic economic unit (Pauketat 1998:135-136; Peregrine1992). Following Flannery (1972; see also Flannery 2002), Peregrine relates these changes to a generaltrend toward the attenuation of the household as the basic economic unit in the evolution of complexsocieties, arguing that this “promotes competition in production and the emergence of intensificationand social stratification” (Peregrine 1992:141-142).

5

Muller (1997:135-141) also considers households in his model of Mississippianization, focusingmore on their distribution than form. He sees the settlement dispersal in the Late Woodland as aconsequence of a natural tendency for households to seek greater autonomy, abetted by the adoptionof the more effective hunting technology of the bow and arrow. Increased dispersal brought greatercompetition for resources as populations expanded, eventually leading to the subordination of somehouseholds. Muller—somewhat like Peregrine, but without directly invoking social evolution—sees thisas a gradual process that played out with minor local variation across the Southeast.

Households also figure prominently in the historical processual rethinking ofMississippianization as a relatively rapid process emanating from a “Big Bang” at Cahokia around A.D.1050 (Blitz 2010; Pauketat 1994, 1997a, 2004a, 2004b, 2007). Pauketat (1997b; 2000:33-35) implicatesthe emergence of households as an economic unit disarticulated from larger kin groups (but nowattached to political patrons) as a transformative juncture. As part of this process, house constructionstyles changed abruptly; wall-trench houses, perhaps produced by work crews, replaced traditionalsingle-post structures (Pauketat 1994:130-140, 2004a:78-80). These and other elements of Mississippianculture are believed to have spread across the Southeast through the direct movement of people fromCahokia (Blitz 2010; Blitz and Lorenz 2006:124-125; Pauketat 2004a:119). Recent works have examinedthe ways in which households on the periphery of Cahokia resisted Mississippianization through theretention of traditional practices, including the continued use of semi-subterranean post structures(Emerson and Pauketat 2002:109; Pauketat and Alt 2003:166-167, 2005).

Beck and colleagues (2007) have discussed the Late Woodland-Mississippian transition in termsdrawn directly from Sewell’s development of eventful temporality, and thus more consistent withapproach taken here. In contrast with Pauketat, they argue that the reorganization of Cahokia aroundA.D. 1050—and by extension the forging of the cultural complex recognized archaeologically asMississippian—was not the product of a single “Big Bang” but instead a series of “creative solutionsto structural disjunctions” beginning during the Late Woodland period (Beck et al. 2007:844). Foremost among these disjunctions, according to Beck and colleagues, was a crisis for existing structuresof intergroup reciprocity as the interests of households began to precedence over those of thecommunity with increases in population and maize production.

These studies make significant advances in identifying changes in Late Woodland households. Yet they undervalue the agency of households in these changes, for reasons both particular and general. In general, because each of the models look backward from the Early Mississippian period at Cahokiaand the American Bottom, the transformative events of the Late Woodland period—and thus also theactions of Late Woodland households—are taken as an inevitability. This inevitably is redoubled in themodels put forth by Peregrine and Muller where the agency of households is reduced to either genericevolutionary processes or a mechanical responses to environmental pressures, respectively. Sewell(2005:84) has referred to these sorts of historical explanations as having a “teleologic temporality,”wherein the cause of a historical transformation is attributed “...neither to the actions and reactions thatconstitute that happening nor to concrete and specifiable conditions that shape or constrain the actionsand reactions but rather to abstract transhistorical processes leading to some future historical state.”

Pauketat, as well as Beck and colleagues, avoid this reductionism by recastingMississippianization as a unique historical process in which changes in households played a crucial role. But teleology is apparent in Pauketat’s work too in the invocation of some traditional practices asresistant to Mississippianization because they are “backward-looking” or “value some aspect of a givenperiod’s life and culture that the analyst, with her or his twenty-twenty hindsight, regards as doomed tothe dustbin of history” (Sewell 2005:84).

Beck et al. avoid the issue of resistance to Mississippianization and thus also this teleology. Yethere too the decisions made by Late Woodland households (e.g., to increase maize production) areevoked in a manner that makes them appear inevitable and unquestioned. This may stem in part fromtheir somewhat narrow conception of agency as “the potential to transform prevailing structures” (Becket al. 2007:845). As Gillespie (2007:847) points out, this would seem to remove agency “except in rare

6

moments of historical event.” Beck and colleagues (2007:856) answer this charge to the contrary, buttheir definition and exposition clearly mitigate the expression of agency by Late Woodland households,especially those further removed from the Mississippian event in time and space.

Finding Late Woodland Households

This report investigates changes in households during the Late Woodland period and theirimplications for the structural transformations marking the junctures with earlier (Middle Woodland)and later (Mississippian) periods. Kolomoki, the largest site of the Weeden Island complex of the GulfCoast and adjacent interior portions of Alabama, Florida, and Georgia, is ideally suited to such a study. Weeden Island has often been viewed as a bridge spanning the dark ages between the “climaxes” of theMiddle Woodland and Mississippian periods (e.g., Muller 1997:123; Nassaney and Cobb 1991:314;Willey 1966:289), a perspective stemming from the fact that many of the hallmarks of the formerperiod—mound construction and long-distance exchange—continued later here than elsewhere in theSoutheast. While the pattern was delayed, however, the same transformations noted by McElrath et al.(2000:23) for the Late Woodland period generally also took place at Kolomoki.

I utilize several strategies to emphasize households as agents of change in these structuraltransformations, rather than simply as reactors to external forces. First and most obviously, agency ishighlighted by defining the unit of analysis and reporting as the household (Johnson 2006:124-125). Ifocus specifically on the “archaeological household,” which Nash (2009:224) has described as a“...coresidential group that used the occupation surface, features, and the artifact assemblage of adwelling,” with dwelling defined to include one or more closely-related structures and both indoor andoutdoor spaces. Specifically, in this report, I summarize recent excavations of a domestic area atKolomoki referred to as Block D, containing the material remains and features of at least one householddating to the late/terminal Late Woodland.

Relatedly, greater appreciation of the agency of households in historical change is achieved by comparing the Block D household to another that is relatively closely related in time. I compare theLate Woodland household in Block D to an earlier Late Woodland household that was previouslyexcavated in Block A at Kolomoki (Pluckhahn 2003; Pluckhahn et al. 2006). I refer to the latter asdating to the early/middle Late Woodland and the Block D as dating to the late/terminal LateWoodland. Archaeological constructs such as phase and type cannot be entirely avoided, but theproblems with these are at least partially mitigated through comparison of sequentially occupiedhouseholds, each of which was occupied for a relatively brief interval and which together span a single,ca. 300 year-long period.

Finally, I conceive of historical change in terms of the “eventful temporality” described bySewell (2005) and introduced to archaeological contexts by Beck, Bolender, and colleagues (Beck et al.2007; contributors to Bolender 2010). Briefly, Sewell (2005:100) defines an eventful temporality as onethat “takes into account the transformation of structure by events.” Structure is conceived in termsrooted in the work of Giddens (1984) (i.e., “rules and resources”), but made more concrete throughseveral key conceptual advances. Perhaps most important, Sewell (2005:131) revises Giddens’s poorly-defined “rules” as schemas, defined as generalizable or transposable rules and procedures that areapplied in the social life. Sewell (2005:132-137) also redirects Giddens’s notion of resources away fromthe “virual” to include both actual (material) human and nonhuman resources.

As Sewell (2005:260-261) notes, events are complex and difficult to bound. They may overlapand interpenetrate. They are also fractal in character; what appears to be one event may well be said tobe comprised of several related events. The structural disjunctions recognizable archaeologically as the“collapse” of Middle Woodland practices and the “Big Bang” that ushered in a vastly different Mississippian social order were separated by three or four centuries, and in one sense may be conceivedas reflecting separate and discrete events. But events can also be conceptualized as “a sequence of

7

ruptures that dislocates and rearticulates structures” (Sewell 2005:261). In this sense, the two eventswere linked by ruptures stretching across the Late Woodland, and were thus interpenetrated.

I will argue that these ruptures included important changes in the organization of LateWoodland households. Briefly, households became smaller and more socially and economicallyautonomous. These changes that were rooted in the decisions made by household members. Theactions of households were thus integral to the structural transformations marked by the events at thebeginning and end of the Late Woodland.

Environmental Setting

Kolomoki is located in the lower Chattahoochee Valley of southwestern Georgia, immediatelysouth and west of Little Kolomoki Creek. Roughly 2 km to the northwest of the site, Little KolomokiCreek joins the North Prong of Kolomoki Creek to form Kolomoki Creek proper. Kolomoki Creekflows into the Chattahoochee River approximately 12 km northwest of the site.

Kolomoki lies within the Coastal Plain physiographic province. While the Coastal Plain is oftencharacterized as a monotonously low and level plain, in reality the situation is far more complex. Veatchand Stephenson (1911:28) divide the Coastal Plain province within Georgia into six major divisions. Kolomoki is situated near the border between the Fall Line Hills and the Dougherty Plain. However,the setting of the site amidst gently rolling hills is clearly more characteristic of the former area. AsVeatch and Stephenson (1911:29) note, the boundaries of the Fall Line Hills are not easily defined. Tothe north, the topography of the Fall Line Hills merges with that of the Piedmont Province, while tothe south there is a gradual transition to the more level Dougherty Plain and Altamaha Upland sectionsof the Coastal Plain Province. In general, the Fall Line Hills are highly dissected, with relief of 50 to 250ft (Hodler and Schretter 1986:74). There is little level land, save for marshy floodplains, narrow streamterraces, and occasional expanses of broad uplands (such as that at Kolomoki).

The topographic features of the Fall Line Hills are determined largely by erosion (Veatch andStephenson 1911:29). The upper portion of the region is characterized by the sand hills, which can bedescribed as flat ridges with a covering of gray or brown superficial sand. Soils in this area are largelyunproductive, and the tree growth consists primarily of stunted oak and long-leaf pine. The lowerportion of the Fall Line Hills, including the Kolomoki area, is principally composed of red hills (Veatchand Stephenson 1911:29). The surface soils here consist of bright red sand, residual material from theunderlying Eocene geologic formations. The sand layer of the red hills is easily eroded, and as a resultthe uplands are often cut by deep gullies. Wharton (1978:172-173) describes these gullies as theChattahoochee Ravines. He notes that the erosion through gorges on streams such as Kolomoki Creekhas exposed thinly laminated clays of the Tuscahoma formation, as well as chert nodules and fossils.

To the south of Kolomoki lies the Dougherty Plain, which consists of a flat to very gently rollingtopography (Hodler and Schretter 1986:171; Veatch and Stephenson 1911:30). This area hascomparatively few surface streams, and therefore lacks the erosion found in the red hills. TheDougherty Plain includes a number of lime-sinks, formed by the solution of the soft, underlyinglimestone layer.

The archaeological site of Kolomoki covers nearly a square kilometer. From the highestelevation at the southern end of the site, the landscape grades gradually downslope to the east and west. Moving north, there is a slight decline to a broad, level plain, where the plaza and most of the moundsare located (Figure 1-3). At the northern end of the site there is a increase in elevation to a northeast-southwest trending ridge. Block A, excavated in 2002 and used as a point of comparison in this report,is located on this ridge (see Figure 1-2).

8

Figure 1-3. Oblique aerial view of Kolomoki. View is to the east with Mound A in center backgroundand Mound D in center foreground.

9

The gently plain at the center of the site is broken on its edges by drainage heads—often referredto as “steepheads” (Northwest Florida Environmental Conservancy 2006). At the head of thesesteepheads are springs or seepages of water which constantly erode the gullies headward. The cool andmoist habitat at the bottom of these steepheads provides shelter for a number of rare plant species suchas: the Florida Torreya tree, the Florida Yew, Ashe Magnolia, Pyramid Magnolia, Mountain Laurel,various Rhododendron species, Florida Anise, Beech trees, White oak trees and a variety of othernorthern hardwood trees. The slopes of the steepheads provide habitat for plant species more oftenassociated with the Piedmont and Appalachian mountains, such as Trillium, Trout lilies, Wild Ginger,and a variety of orchids. Examples of steepheads at Kolomoki can be observed on the topographic map(see Figure 1-2) between Mounds D and E, between Mounds A and C, and east of Mound H. BlockD, which is the focus of this study, is located along the northern edge of the latter ravine-like area. Sears(1956:10) referred to this area as the “South Area” or “South Village.” I have elsewhere described it asthe “North Ravines,” in reference to the ravine or steephead just to the south (Pluckhahn 2003:73).

Block D is located just south of the mainpark road, on a gentle, south and southeast-facing slope above the steephead (Figure 1-4). A portion of this area is maintained as open andpark-like, but our work centered on the area justinside a tree line. Vegetation consists of a mixof hardwoods and pines with a moderatelydense understory.

Not surprisingly given the size ofKolomoki, soils within the site area are highlyvariable. The nearly level uplands at thesouthern margin of the site are characterized byGreeneville sandy loams. Pilkinton (1985:24)describes the typical profile for this soil type asconsisting of a 7 inch thick surface layer of darkreddish brown sandy loam over a dark red sandyclay. The gently sloping land at the center of thesite generally consists of Red Bay sandy loams. Here the surface layer is typically a dark reddishbrown sandy loam 8 inches thick (Pilkinton1985:37). Subsoil consists of a dark red sandyclay.

The lower elevations at Kolomoki,including the drainage area between Mounds Eand D and the gentle slope to the northwest ofMound A, are characterized by Americus loamysands. These soils are only moderately suited tofarming, largely because they consist ofsomewhat excessively drained sands (Pilkinton1985:17-18). The surface layer typically consists of a dark reddish brown loamy sand 8 inches thick. This is underlain by a few inches of dark reddish brown loamy sand, which grades into a dark red loamysand and, finally, a dark red sandy loam. Soils in Block D are described in Chapter 3.

Kolomoki witnesses long, hot summers as a result of the moist, tropical air of the Gulf ofMexico (Pilkinton 1985:1-2). The winters are cool and short. Precipitation is fairly heavy throughoutthe year, with more than half of the rainfall occurring in the spring and summer growing seasons fromApril through September.

Figure 1-4. Topography of the Block D area.

10

Chapter 2: Methods

Thomas J. Pluckhahn, Mary Theresa Bonhage-Freund, Matthew C. Compton, Linda Scott-Cummings, and Chad Yost

This chapter describes the methods that were employed on the project, including field,laboratory analysis, and curation methods. The descriptions of specialized laboratory methods aredrawn from the reports of Bonhage-Freund (2009) (macrobotanical analysis), Compton (2009) (faunalanalysis), and Yost and Scott-Cummings (microbotanical analysis), with minor editing by Pluckhahn.

Field Methods

The fieldwork for this project was conducted over the course of three field seasons in thesummers of 2006, 2007, and 2008. Work was completed under the direction of report author and P.I.Thomas Pluckhahn. The 2006 field crew (Figure 2-1) included volunteers and graduate students from USF and the University of Oklahoma. The 2007 and 2008 excavations were conducted primarily byfield school students from the USF (Figures 2-2 and 2-3, respectively), with additional assistance ofvolunteers. USF graduate students Cassandra (Rae) Harper and Elizabeth Heath served as fieldassistants in 2006 and 2007, respectively.

Figure 2-1. The 2006 field crew at Kolomoki. From left: Buck Brown, Bobby Butler, Elsbeth Fields,Rae Harper, and Sarah Carter. Not shown: Dick Brunelle.

11

Figure 2-2. The 2007 field crew at Kolomoki. From left: Todd Bogner, Liz Heath, DeannaGriffin, Elicia Kimble, Dianely Martin, Travis Couliette, Ashley Humphries, Catie Sheldon, JohnMeyer, Jennifer Vessels, Don Moysey. Not shown: Jen Boekenoogen, Dick Brunelle.

Figure 2-3. The 2008 field crew at Kolomoki. From left: Jenna Clevinger, Tria Marie Ellison,Erin Rosenthal, Kevin Hageman, and Stephanie Lonergan.

12

Proveniences in Block Dwere maintained using the site-specific grid system developed duringprevious work at Kolomoki(Pluckhahn 2003). Mapping wasaccomplished with a Leica totalstation. A slight error in thereconstruction of the grid systemfrom previous field seasons resultedin a minimal deviation from theorientation of Units D1-D4. Thiswas corrected so that all other unitswere oriented with the site grid,which is aligned with magnetic north.

Respecting the fact thatKolomoki is preserved as a statepark, the units in Block D—aselsewhere in recent work atKolomoki (Pluckhahn 1998, 2000,2003)—were excavated entirely byhand. Although we often laid outand numbered units as 1-x-2-m pits,each 1-x-1-m section was treated as aseparate unit. The units wereexcavated in 10 cm levels withinnatural soil layers when possible. Soil was screened through 1/4 inch(0.64 cm) mesh. Each unit wasdocumented thoroughly by writtennarrative and technical descriptionsof each stratigraphic level on project-specific test unit forms. Additionaldocumentation included mappingand photography of profiles and planviews.

Features were mapped on unit plan view maps and were usually photographed before and afterexcavation. We generally bisected features, excavated half, and drew a profile before excavating thesecond half (Figure 2-4). Soil samples were collected from the majority of the features determined tobe associated with the prehistoric occupation of the site. The rest of the fill was screened in the fieldthrough 1/4 inch (0.64 cm) mesh. Features were also documented with project-specific feature forms.

Laboratory Methods

Flotation

The majority of the soil samples from Block D features were floated, which was accomplishedwith SMAP-style machines most closely resembling the Siraf (Pearsall 1989:29-31, 52-58). Samples fromthe 2006 field season were processed using an older flotation machine at USF which separates into threefractions: A (1/4"), B (an 850 micron heavy fraction), and C (a 300 micron light fraction). Flotationsamples were processed bu USF graduate students Ryan Harke, Will Klinger, Rebecca O’Sullivan, andJennifer Poulssen.

Figure 2-4. Bisection and profile mapping of a feature in BlockD.

13

The samples from 2007and 2008 were processed using arecently constructed machinewhich uses 800 micron mesh forthe heavy fraction and 250 micronmesh for the light fraction (Figure2-5). In the case of these samples,the heavy fractions were screenedthrough 1/4 inch mesh. Thisflotation machine was constructedbecause the USF apparatusfrequently became clogged by thesite's sandy clay soils. The secondmachine features dual showerheads above a shelf lined with the800 micron mesh, providing moresurface area and agitation than theUSF device.

Flotation samples werecarefully processed by introducingonly about a liter of soil every fiveminutes into the respective system. Light and heavy fractions were thoroughly dried before packing inseparately labeled heavy-dutyzipper bags, and thorough recordsof provenience, feature types, andsample volume were maintained. Nearly all of the admittedly sparseseeds were captured in the lightfraction, with the heavy fractionconsisting mainly of relatively smallamounts of residual wood charcoaland resin, along with the bulk ofthe hickory shell.

Light fractions were roughsorted into general categories offaunal, floral, artifact, and other(mainly modern organic debris). Faunal and floral remains fromlight fractions were then analyzed

by Compton and Bonhage-Freund, respectively, as described in more detail in below. Bonhage-Freundalso examined a selection of heavy samples to ascertain the effectiveness of the flotation. Artifactsrecovered from the 1/4 inch were analyzed by Pluckhahn and graduate assistants at USF, also describedbelow.

As an additional check on the results of the flotation, some of the soil samples collected duringthe 2008 field season were submitted for pollen and phytolith analysis. The methods employed in thoseanalyses, which were conducted by Yost and Scott-Cummings, are presented below.

Figure 2-5. Flotation machine used for samples collected in the2007 and 2008 field seasons.

14

Analysis of Artifacts

All artifacts recovered during the survey were transported to the laboratory in the AnthropologyDepartment at USF for processing. The artifacts were analyzed by the P.I., with the assistance of USFgraduate students Jana Futch, Will Klinger, Shannon McVey, Rebecca O’Sullivan, and Jennifer Poulssen. Additional assistance was provided by several undergraduates at USF, particularly Todd Bogner, JohnMeyer, and Elizabeth Messer.

Flaked stone artifacts were sorted by raw material and stage of reduction or tool production bythe author. Technological analysis followed the lithic reductive model of Collins (1975) in which theprocesses of chipped stone tool manufacture and use are perceived as a series of five ordered stages: 1)acquisition of raw materials, 2) initial reduction, 3) primary flaking, 4) secondary flaking, and 5) useand/or recycling. Debitage, or detached pieces, were sorted into the following categories roughlyreflecting these stages: primary flakes (>90 percent cortex), secondary flakes (1-90 percent cortex),tertiary flakes (no cortex), shatter (angular pieces) and utilized flakes (exhibiting obvious use wear).

Objective pieces include cores and deliberately retouched tools. Tools were assigned to morespecific categories such as uniface and biface based on general morphology. Cultural and temporalaffiliations of diagnostic hafted biface were determined through comparison with published typedescriptions ( e.g., Bullen 1975; Cambron and Hulse 1978; Whatley 2002). Subsequently, Pluckhahn’s typeassignments were submitted to John Whatley and Lloyd Schroeder (personal communication 2008) fortheir opinions. The type assignments discussed herein reflect most, but not all of their suggestions fornomenclature and assignments of individual points. We thank Whatley and Schroeder for their opinionsbut we emphasize that they are not responsible for the conclusions presented here.

Our analysis of hafted bifaces emphasizes hafting areas to minimize effects of use wear andre-sharpening (Andrefsky 1998:178; Bacon 1977; Binford 1963). We also generally focus on ratio ofmeasurements (for example, the ratio of blade width to haft length), to accommodate the constraints onthe size of finished bifaces relative to raw material (although the are manufactured primarily from variouscherts of the Coastal Plain, there are also specimens made from cherts of the Piedmont and Ridge andValley, as well as Tallahatta Sandstones of southern Alabama). Hafted bifaces were measured using a dialcalipers to the nearest 0.1 mm and weighed to the nearest 0.1 g. Our measurements focused on 9 basicdimensions (Figure 2-6): Maximum Length (ML), Maximum Width (MW), Blade Length (BLL),

Figure 2-6. Hafted biface measurement dimensions (adapted from Andrefsky 1998:179).

15

Blade Width (BLW), Base Width (BW), Haft Length (HL), Neck Width (NW), Neck Height (NH), andMaximum Thickness (MT). These dimensions generally conform to those defined by relativelywell-defined hafting elements, as delineated by shoulders. Thus, triangular points by our definition lackhafting areas (although we recognize these were indeed hafted in most cases). Perhaps most important,we restrict the use of the term “neck” to an area of constriction below the shoulders of width roughlyequal to or less than that of the base. So defined, most proximally contracting stemmed and triangularpoints lack necks. We would note that our method also differs from Whatley (2002:10), who measureshaft width and length relative to an undefined point on the contracting stem.

The analysis of prehistoric ceramics was based on attributes of paste, tempering agents, andsurface modification or decoration. Pottery types were identified with reference to establishedchronologies for the region (Sears 1956; Wauchope 1966; Willey 1949). For each provenience, sherdswere counted and weighed (to the nearest 0.1 g) by type of general descriptive category. Small (less than2 cm), plain or eroded sherds were classified as “residual.”

A minimum number of vessel (MNV) analysis was conducted on larger rims sherds (>5 percentof orifice diameter) and diagnostic body sherds from the two excavation blocks to ascertain changes invessel form and size. Vessel forms were identified with respect to seven basic categories (Table 2-1)representing a compromise between the classifications recently devised by Hally (2009) and Wallis (2011)(the latter derived mainly from Willy [1949]). These forms include: neckless jar; collared jar; roundedbowl; unrestricted bowl/jar (combined here due to difficulty in differentiating on the basis of small rimsherds); cup; and dish/plate. Vessel size was measured with respect to orifice diameter, measured to thenearest centimeter using a template.

Although ceramics and flaked stone are the most common prehistoric artifacts, there are a numberof other miscellaneous prehistoric artifact categories. The ground and pecked stone tool categoriesinclude artifacts that display evidence of battering caused by repeated impacts, or which exhibit batteredor ground depressions on at least one surface. Tools of this type were weighed and counted. Heat-alteredrock was also weighed and counted. This class of artifact includes both fire cracked rock (FCR), and firereddened pebbles.

A limited quantity of historic artifacts were recovered on this project. Historic artifacts weresorted into functional categories or manufacturing type. Whenever possible, these artifacts were placedinto a chronological time frame.

Table 2.1. Vessel Form Categories Used in this MNV Analysis and Their Equivalents in the Analyses ofHally (2009) and Wallis (2011).

This Analysis Hally (2009) Wallis (2011)

neckless jar neckless jar flattened globular bowl

collared jar folded rim/tall folded rim jar collared jar

open bowl folded rim/rounded bowl open bowl

restricted bowl N/A restricted bowl

simple bowl/unrestricted jar unrestricted jar open/unrestricted pot

cup N/A small bowl/cup

dish/plate N/A shallow bowl

16

Analysis of Faunal Remains

Animal remains from the Block D excavations were identified using standard zooarchaeologicalmethods (Reitz and Wing 1999). All identifications were made by Matthew Compton (2009) using hispersonal comparative skeletal collection and the comparative skeletal collection of the University ofGeorgia Zooarchaeology Laboratory. Specimens are identified in terms of taxonomic classification,element represented, the portion recovered, and symmetry. Evidence for age at death, sex, seasonality,and modifications are noted where observed. Standard measurements are usually taken but none wereavailable in the Block D assemblage. The relative abundance of different taxa is presented in terms ofNumber of Identified Specimens (NISP), Minimum Number of Individuals (MNI), specimen weight, andbiomass. NISP is simply the count of specimens identified. Specimens that cross-mend are counted assingle specimens. Indeterminate vertebrate (Vertebrata) specimens are not counted due to theirfragmentary nature. MNI is a measure of how many individuals are required to account for the specimenspresent in a collection. MNI is estimated based on element portion, symmetry, size, and age. MNI isestimated at the lowest possible taxonomic level, usually the species.

While MNI is a standard zooarchaeological quantification method, the measure has severalwell-known biases. For example, MNI emphasizes small species over larger ones. This can bedemonstrated in a hypothetical sample consisting of five eastern box turtles (Terrapene carolina) and onewhite-tailed deer (Odocoileus virginianus). Although five eastern box turtles might indicate emphasis oneastern box turtle, one white-tailed deer would supply more meat. Further, some elements are morereadily identifiable than others. The taxa represented by these diagnostic elements may achieve higherMNI estimates and therefore be incorrectly perceived as more significant to the diet than animals with lessdistinctive elements. Conversely, some taxa represented by large numbers of specimens may present fewpaired elements and hence the number of individuals for these species may be low. Basic to MNI is theassumption that the entire individual was utilized at the site. From ethnographic evidence, it is known thatthis is not always true (Binford 1978; Perkins and Daly 1968). This is particularly the case for largerindividuals, animals used for special purposes, and where food exchange was an important economicactivity (Thomas 1971; White 1953).

In addition to these primary biases, MNI is also subject to secondary bias introduced by the waysamples are aggregated during analysis. The aggregation of archaeological samples into analytical units(Grayson 1973) allows for a conservative estimate of MNI, while the "maximum distinction" method,applied when analysis discerns discrete sample units, results in a much larger MNI. For the KolomokiBlock D assemblage, a single analytical unit is used to estimate MNI values.

Biomass estimates are used to predict the dietary contribution of different taxa and are calculatedusing specimen weight and allometric formulae following Reitz et al. (1987) and Reitz and Wing (1999:72,224-231). Predictions of biomass are based on the allometric principle that the proportions of body mass,skeletal mass, and skeletal dimensions change with increasing body size (Reitz and Wing 1999:225-233).The relationship between body weight and skeletal weight is described by the allometric equation:

Y = aXb

(Simpson et al. 1960:397). In this equation, X is skeletal weight, Y is the biomass, b is the constant ofallometry (the slope of the line), and a is the Y-intercept for a log-log plot using the method of leastsquares regression and the best-fit line (Reitz et al. 1987; Reitz and Wing 1999:224-228). To solve forbiomass (Y), the archaeological specimen weight is substituted for skeletal weight (X). Many biologicalphenomena show allometry described by this formula (e.g., Gould 1966, 1971) so that a given quantityof skeletal material or a specific skeletal dimension represents a predictable amount of tissue or bodylength due to the effects of allometric growth. For more detail on the allometric formulae used in thisanalysis, consult Compton (2009:Appendix B). Biomass is not estimated for the indeterminate vertebrategroup.

17

In order to assess the relative abundance of different groups of animals identified, taxa aresummarized into faunal categories based on major vertebrate groups. These faunal categories arecompared in terms of MNI and biomass. The categories include Turtles, Birds, White-tailed deer, andOther mammals. The summary table includes biomass estimates only for those taxa for which MNI isestimated.

Relative ages of white-tailed deer are estimated based on observations of the degree of epiphysealfusion for diagnostic elements (Purdue 1983; Reitz and Wing 1999:71-76). During analysis, specimensare recorded as either fused or unfused and placed into one of three categories based on the age in whichfusion generally occurs. Elements in the Early-fusing category generally fuse before the animals firstbirthday (Purdue 1983). Unfused elements in this category are interpreted as evidence for juveniles. Elements of the Middle-fusing category typically fuse between one and two years in age (Purdue 1983). Unfused elements of the Middle-fusing category are interpreted as either subadults or juveniles whilefused elements are interpreted as subadults or adults. Late-fusing elements tend to fuse after the animalssecond birthday (Purdue 1983). Fused elements in this category are interpreted as representing adults;unfused elements are interpreted as subadults or juveniles. Because ages for epiphyseal fusion vary dueto sex, health status, climate, available forage, and individual variation among animals, assigning specificages from epiphyseal fusion data is difficult. However, general age ranges for the three age classes can beassigned in the following manner: juveniles <1 year, subadults 1-2.5 years, adults >2.5 years.

The presence or absence of elements in an archaeological assemblage provides data on animal usesuch as butchering practices, transport decisions, and food distribution (Reitz and Wing 1999:202-205). The white-tailed deer elements identified are summarized into categories by body parts. The Headcategory includes skull fragments, including mandibles, antlers, and teeth. The sternum, ribs, andvertebrae, excluding the sacrum, are placed into the Vertebra/Rib category. The Forequarter includes thescapula, humerus, radius, and ulna. The Hindquarter includes the innominate, sacrum, femur, tibia, andpatella. Carpal and metacarpal specimens are in the Forefoot category. The Hindfoot category includestarsal and metatarsal specimens. Metapodia and podia not identified to the Forefoot or Hindfootcategories, as well as sesamoids, phalanges, and bones of the minor digits are assigned to the Footcategory. The archaeological white-tailed deer element data are presented with a standard white-taileddeer skeleton for comparison. The standard white-tailed deer is based on the number of elements presentin a complete white-tailed deer skeleton. Elsewhere, Compton (2009:Appendix C) summarizes thenumber of elements present in an unmodified white-tailed deer skeleton. Here, elements that are fusedin adult white-tailed deer or that articulate in such a manner that they do not readily disarticulate in theabsence of soft tissue (e.g. elements of the skull) are counted as single elements.

The archaeological white-tailed deer element data are also compared to a standard white-taileddeer using a ratio diagram (Simpson 1941; Reitz and Wing 1999:212). Described by George Simpson(1941; Simpson et al. 1960:357-358), the formula is as follows:

d = logeX - logeY or d = loge(X/Y)

where d is the logged ratio, Y is percentage of each element category in the standard white-tailed deer, andX is the same percentage of this category in the archaeological collections. It does not matter to what basethe measurements are converted, though one should be consistent in order to remain comparable. AsSimpson (1941:23) describes this approach:

The basic purpose of the diagram is to represent each of a number of analogousobservations by a single entry and to plot them in such a way that the horizontal distancebetween any two of them will represent the ratio of either one of those two to the other.

In order to compare the archaeological data with the standard white-tailed deer, the percentagesof each element category for the standard white-tailed deer are converted into common logs, subtractedfrom the logged value of the same element category for the archaeological percentages, and plotted againstthe standard white-tailed deer represented by the horizontal line in the ratio diagram figures. Values on

18

the positive side of the standard’s horizontal line are over-represented and those on the negative side ofthe line are under-represented. A burial would present an essentially horizontal line compared to thestandard. Although the archaeological values are specimen counts and the values for the standardwhite-tailed deer are whole elements, the relationships in the ratio diagrams are similar to those found inunmodified histograms.

Modifications can indicate butchering methods as well as site formation processes. Modificationsare presented as counts and are classified here as burned and calcined (see Lyman 1994 and Reitz andWing 1999 for complete descriptions of modification categories). Burned specimens result from thecarbonization of bone collagen and are identified by their charred-black coloration (Lyman 1994:384-385). Burned specimens may result from exposure to fire when meat is roasted or when specimens areintentionally or unintentionally burned after discard. Heating bone at extreme temperatures ($600EC)can cause the specimen to become completely incinerated or calcined; calcined specimens are usuallyindicated by white or blue-gray discoloration (Lyman 1994:385-386).

Several sources of potential bias may have influenced the Block D faunal assemblage and shouldbe kept in mind in the following sections. First, the use of a 1/4-in (6.35 mm) screen to recover animalremains biases samples by reducing the number of smaller specimens collected. Numerous studies havedocumented the effects of the use of 1/4-in screen on the recovery of animal remains (Gordon 1993;Payne 1972; Shaffer and Sanchez 1994; Stahle 1996; Struever 1968). Most notable of these effects are theover-emphasis on the importance of large species and the reduced number of small species represented.

A second source of potential bias is differential preservation. Larger, denser bones tend topreserve more readily than smaller, less-dense bones (Lyman 1994:235-236, 397-398). This can causespecies with larger more robust bones to be more abundant than those with smaller more fragile bones. Further, the burning of bone and gnawing by rodents and carnivores may have caused the loss of anunknown amount of bone (Kent 1981; Lyman 1994:193-195, 205-216, 391). Differential preservationlikely played a role in biasing the Block D assemblage but to what degree cannot completely bedetermined.

A third source of potential bias is the context from which the materials were recovered. The BlockD faunal materials are assumed to represent domestic debris. However, the possibility exists that someof these materials were deposited due to a specialized activity not representative of the overall animal-usestrategy. Further, the excavated materials may have been deposited only during certain seasons of the yearand therefore are not representative of a full annual cycle. The materials might also be biased due to thefact that they may only sample the domestic refuse of a few households that are not representative of theentire population. Differences in the status and food preferences of individual households play a majorrole in food consumption habits and therefore may influence the interpretation of the faunal collections.

Finally, the small size of the faunal assemblage is a probable source of bias. The size of faunalsamples is known to influence measures of relative abundance, richness, and diversity indices (Cannon2001; Grayson 1981, 1984:116-117, 132, 158; Lyman and Ames 2004). It is likely that the Kolomoki BlockD assemblage is biased due to the small size of the sample.

Analysis of Macrobotanical Remains

Macrobotanical analysis was conducted on 51 light and 51 heavy fraction flotation samples, andone fine water-screened sample. The samples were taken from 40 unique features. The flotation samplingprogram was designed to test all feature types in as robust a manner as possible. The inventory includes16 large pits, 8 small pits, 15 postmolds, and one pottery cluster. The median pre-flotation sample volumewas 6.0 liters. Sample volume in the Block D excavations was largely a reflection of the feature sizes. Halfof the features consisted of pits and postmolds that were small in size. Multiple samples were taken fromlarge features resulting in smaller individual samples, but greater total floated material. Flotation samples

19

from large pits, exclusive of special thin highly organic layers, ranged from 4.4 to 12.0 liters, with themajority being at or above the median.

In the laboratory, each light fraction flotation sample was first weighed, and then passed throughnested geologic sieves (8.0 mm, 4.0 mm, 2.0 mm, 1.0 mm, 0.5 mm, and 0.125 mm). Each size-graded lightfraction was full sorted under low magnification (4-56x). All charred plant remains are considered to beof archaeological significance, but non-charred seeds are deemed modern. All charred seeds and othercharred plant remains collected in the 2.0 mm and larger sieves were removed and quantified them bymaterial type, count, and for nutshell, resin, and wood charcoal by count and weight. Fractions less than2.0 mm were visually scanned but only carbonized seeds were extracted. Uncharred plant remains andmychorrizae (fungal spores or fruiting bodies) were noted but these were not formally counted. Identifications of taxa were made by comparison to modern charred and uncharred reference collections,positively identified archaeological specimens, and standard reference volumes (Britton and Brown 1970;Delorit 1970; Martin and Barkeley 1973; USDA 1974; Montgomery 1977). Taxon determination focusedon morphological characteristics of the specimens, including size, texture, specialized structures (e.g.pores), and overall shape. In some cases important diagnostic traits were missing due to degradation, orbecause the specimen was incomplete.

These procedures were followed for the complete heavy fractions of flotation Samples 7, 9, 18,and 39 (see Table 6-6 for provenience information). It was quickly observed that the heavy fractions heldno seeds and very sparse to no charcoal, resin, or macroplant remains below the 2 mm level. Therefore,none of the remaining heavy fractions were examined below the 2 mm screen. The samples were furtherreduced prior to shipping to me by sifting through 0.25 inch screen. Thus, of these only the sizes > 2mmand < 0.25 inches were examined. With few exceptions the only plant remains in the heavy fractions werenut shell, resin, and wood charcoal. Therefore, to save time and expense without compromising results,20 of the already reduced heavy fractions were scanned with only the seeds and fragments of acorn shellbeing removed. These included flotation Samples 16, 17, 19, 20, 22, 23, 30, 31, 32, 33, 35, 36, 37, 38, 40,41, 42, 44, 49, 50, and two unnumbered samples from Features 147B and 181. The Feature 147B samplewas hand floated in a bucket and the Feature 181 sample was dry sieved and then water screened througha 125 micron sieve.

Wood charcoal is the dominant archaeological plant remain in every sample. Minnis and Ford(1977:82) propose that 20 pieces of charcoal from each provenience constitute a sample size that willadequately reflect the total frequencies of charcoal within a sample. However, the charcoal recovered byflotation from features and post molds in the southeastern United States is vastly different from samplesrecovered at Chimney Rock and other southwestern sites which were analyzed by Minnis and Ford (1977). Southeastern wood charcoal fragments are typically much smaller, most measuring between 2.0 and 4.0mm, or smaller, and being derived from a much richer and more diverse habitat than the mesa regions. In cases where the wood count was low, we attempted to identify all wood charcoal 2 mm or larger. Otherwise, wood charcoal fragments were evaluated in groups of 10. A minimum of two groups wasexamined from each feature or postmold having an adequate quality and quantity of wood charcoal. Extrapolating from the Minnis and Ford recommendation, this process continued until we encounteredone lot beyond the initial assemblage of 10 wood fragments, which contributed no new taxa to the sample. Examined specimens were selected using a method deemed most appropriate for the individual lot. Random pinch samples were used for samples estimated to contain 100 or more fragments of similar sizes. If a sample contained an estimated 50 or more fragments of varying samples an artist's brush was movedthrough the sample from top to bottom of the sample's center and again in an X pattern, from corner tocorner. Swept fragments were then randomly selected from the displaced specimens. Finally, in the caseof samples containing approximately 40 or fewer, but more than 20, wood charcoal fragments, every otherfragment was examined.

Wood specimens smaller than 2.0 mm in size were not evaluated. Any specimen smaller than 2.0mm is unlikely to possess all the structural features needed for a reasonable identification. In fact, mostfragments above 2.0 mm could only be identified to the genus, and some could only be classified as ring

20

porous, diffuse porous, hardwood, conifer, or monocot. The latter set provides limited, yet useful,information.

Wood taxa were identified by comparison with charred, natural, or published photographs oftransverse, tangential, and radial thin sections of modern wood, although the transverse view wasemphasized due to magnification limitations and size of the specimens (Panshin and de Zeeuw 1980).

Analysis of Microbotanical Remains

Soil samples from four pit features in Block D were submitted for pollen and phytolith analysis. The analysis was conducted by Yost and Scott-Cummings in the laboratory of PaleoResearch, Inc. (PRI).

Pollen. A chemical extraction technique based on flotation is the standard preparation techniqueused in this laboratory for the removal of the pollen from the large volume of sand, silt, and clay withwhich they are mixed. This particular process was developed for extraction of pollen from soils wherepreservation has been less than ideal and pollen density is lower than in peat. It is important to recognizethat it is not the repetition of specific and individual steps in the laboratory, but rather mastery of theconcepts of extraction and how the desired result is best achieved, given different sediment matrices, thatresults in successful recovery of pollen for analysis.

Hydrochloric acid (10%) is used to remove calcium carbonates present in the soil, after which thesamples are screened through 250 micron mesh. The samples are rinsed until neutral by adding water,letting the samples stand for 2 hours, then pouring off the supernatant. A small quantity of sodiumhexametaphosphate is added to each sample once it reaches neutrality, then the samples are allowed tosettle according to Stoke’s Law in settling columns. This process is repeated withethylenediaminetetraacetic acid (EDTA). These steps remove clay prior to heavy liquid separation. Thesamples are then freeze dried. Sodium polytungstate (SPT), with a density 1.8, is used for the flotationprocess. The samples are mixed with SPT and centrifuged at 1500 rpm for 10 minutes to separate organicfrom inorganic remains. The supernatant containing pollen and organic remains is decanted. Sodiumpolytungstate is again added to the inorganic fraction to repeat the separation process. The supernatantis decanted into the same tube as the supernatant from the first separation. This supernatant is thencentrifuged at 1500 rpm for 10 minutes to allow any silica remaining to be separated from the organics. Following this, the supernatant is decanted into a 50 ml conical tube and diluted with distilled water. These samples are centrifuged at 3000 rpm to concentrate the organic fraction in the bottom of the tube. After rinsing the pollen-rich organic fraction obtained by this separation, all samples receive a short (20-30minute) treatment in hot hydrofluoric acid to remove any remaining inorganic particles. The samples arethen acetolated for 3-5 minutes to remove any extraneous organic matter.

A light microscope is used to count pollen at a magnification of 500x. Pollen preservation in thesesamples varied from good to poor. Comparative reference material collected at the IntermountainHerbarium at Utah State University and the University of Colorado Herbarium was used to identify thepollen to the family, genus, and species level, where possible.

Pollen aggregates were recorded during identification of the pollen. Aggregates are clumps of asingle type of pollen and may be interpreted to represent pollen dispersal over short distances or theintroduction of portions of the plant represented into an archaeological setting. Aggregates were includedin the pollen counts as single grains, as is customary. The presence of aggregates is noted by an "A" nextto the pollen frequency on the pollen diagram. A plus (+) on the pollen diagram indicates that the pollentype was observed outside the regular count while scanning the remainder of the microscope slide. Pollendiagrams are produced using Tilia 2.0 and TGView 2.0.2. Total pollen concentrations are calculated inTilia using the quantity of sample processed in cubic centimeters (cc), the quantity of exotics (spores)added to the sample, the quantity of exotics counted, and the total pollen counted and expressed as pollenper cc of sediment. Indeterminate pollen includes pollen grains that are folded, mutilated, and otherwise

21

distorted beyond recognition. These grains are included in the total pollen count since they are part of thepollen record. The microscopic charcoal frequency registers the relationship between pollen and charcoal. The total number of microscopic charcoal fragments was divided by the pollen sum, resulting in a charcoalfrequency that reflects the quantity of microscopic charcoal fragments observed, normalized per 100pollen grains. Pollen analysis also includes examination for and identification of starch granules to generalcategories, if they are present. Starch granules are a plant's mechanism for storing carbohydrates. Starchesare found in numerous seeds, as well as in starchy roots and tubers. The primary categories of starchesinclude the following: with or without visible hila, hilumcentric or eccentric, hila patterns (dot, cracked,elongated), and shape of starch (angular, ellipse, circular, eccentric). Some of these starch categories aretypical of specific plants, while others are more common and tend to occur in many different types ofplants.

Phytoliths. Because of the highly oxidized nature of these sediments, 30% hydrochloric acid (Hcl)was first added to each 15 ml sample to remove excess iron (Fe) with can leach onto phytoliths and altertheir specific gravity, thus reducing recovery. After two days in HCl, the samples were rinsed to neutral. Sodium hypochlorite (bleach) was then used to destroy the organic fraction and to help breakdown someof the microscopic charcoal particles present in these samples. Once this reaction was complete, thesamples were rinsed to remove the bleach. A small quantity of sodium hexametaphosphate was addedto each sample once it reached neutrality, then the samples were allowed to settle according to Stoke’s Lawin 500 ml beakers. This process was repeated with EDTA. These steps remove clay particles and humatesprior to heavy liquid separation. Next, the samples were freeze dried under vacuum. The dried silts andsands were then mixed with sodium polytungstate (density 2.3) and centrifuged to separate the phytoliths,which will float, from the other silica, which will not. The samples were then rinsed with distilled water,then alcohol to remove the water. After several alcohol rinses, the samples were mounted in immersionoil for counting with a light microscope at a magnification of 500x. A phytolith diagram was producedusing Tilia and TGView 2.0.2.

Curation

Artifacts, field notes, maps, photographs, analysis forms, and other primary documents relatingto this project and analysis forms will be curated at the University of Georgia Laboratory of Archaeologyin Athens. The artifacts from each provenience were repackaged in acid free plastic bags with acid freepaper tags identifying provenience. Provenience information was also recorded on the exterior of thebags.

22

Chapter 3: Excavation Units

Thomas J. Pluckhahn

Excavation Strategy

As was noted previously, Block D is located just south of the main park road, on a gentle, south-and southeast-facing slope above an active spring (sometimes referred to as a “steephead”) (Figure 3-1).This area has long been noted as a productive portion of the site. A 1937 map of the site by Valliant(see Pluckhahn 2003:Figure 3.1) notes the presence of many sherds eroding from upturned trees. Searsdescribed buried midden layers in this area of the site, and reported the identification of several features(Pluckhahn 2003:73-78; Sears 1951:1-27).

Consistent with previous descriptions, shovel tests and a test unit (Test Unit 18) excavated inthis area in the course of previous research revealed higher concentrations of pottery, along with aburied A horizon and a high density of features (Pluckhahn 2003:110-125). Pottery recovered from thisarea included relatively high proportions of conspicuous Late Woodland types such as Indian PassIncised, Weeden Island Incised, and Napier Complicated Stamped.

With the goal of identifying and excavating a Late Woodland house, our strategy in theexcavation of Block D (as with previous block excavations at Kolomoki), was to begin by scattering testunits in the general area, before eventually concentrating contiguous units in an area where we judgedthe potential for household features to be highest, owing to the presence of a buried A horizon and ahigher density of features and artifacts. Figure 3-2 shows the locations of excavation units in Block D

Figure 3-1. View to the southwest of the excavation of Block D.

23

18

D18

D14

D8S

D1SD2SD3S

D9S

D5E

D4E

D7ED6E

D1ND2ND3N

D9ND8N

D6W

D5W

D4W

D7W

D15S D16S D10S

D24SD23SD21SD25S

D19S

D17SD11S

D12E

D22E

D20E

D13E

D25N D21N D23N D24N

D10ND16ND15ND19N

D17ND11N

D13W

D22W

D20W

D12W

Figure 3-2. Map of Block D showing unit numbers and year of excavation.

24

0 2 Meters¯

contour interval = 25 cmfield season:20002006

20072008

in greater. The map also shows the field seasons in which the test squares were excavated, thusillustrating how work on the block progressed. Including the previously excavated 2-x-2-m Test Unit18, we completed a total of 52 m2 of excavation in Block D. Of this total, 38 m2 were contiguous 1-x-1m units that together form a block about 8 m long (north-south) and 6 m wide (east-west). Our desireto expose a wide area of domestic features was hampered by the slow pace of hand excavation andscreening of the plow zone—necessary tasks, given that the site is preserved as a state park. The growthof large trees on this portion of the site furthered hindered the expansions of our excavations, and ledto the rather irregular shape of the contiguous block.

This chapter begins with an overview of the stratigraphy of Block D, then turns to a briefdescription of the density and distribution of artifacts in excavation units and the implications of thesefor formation processes. More detailed descriptions of the artifacts and features found in the excavationunits are reserved for the chapters that follow.

Stratigraphy

As noted in Chapter 1, soils at Kolomoki are variable, owing primarily to the large size of thesite and its position within the Fall Line Hills, a transitional zone between the Piedmont and CoastalPlain (Hodler and Schretter 1986:74; Veatch and Stephenson 1911:29). A long history of intensiveagriculture has further confounded the situation, with some portions of the site extensively deflatedfrom years of plowing and erosion. In general, soil profiles on the upland portions of the site areshallow, with a thin and eroded (10-to-20-cm thick) plow zone layer of dark reddish brown sandy loamcovering a subsoil consisting of a dark red sandy clay (Pilkinton 1985:37). However, there are areaswithin the upland portions of the site where the soil profile is deeper, thanks to less extensive plowing,the deposition of colluvium, or some combination of these factors. Sears (1956:8) noted the presenceof midden deposits preserved in some of the lower elevations at Kolomoki:

The undisturbed units of midden are, seemingly, those which occupied slight hollowsin the original surface or were protected by structures. From this point of view, theplow zone is not a midden, although it is at many points rich in artifacts. Figure I[reproduced as Figure 3-3] is an idealized cross section which would apply to almost anyof the midden units as we found them. For the sake of emphasis, however, I may repeatthat these basin shaped depressions are not structures in any sense of the word, but arethe fortuitous results of aggrading activities which have taken place since they werefilled.

Block D was intentionallypositioned to investigate one such buriedmidden deposit. Indeed, the block mayintersect the same, or a related buriedmidden deposit described by Sears as “Unit4.” This unit, somewhere on the samegeneral portion of the site (the unit was notmapped precisely), produced several largepit features and Weeden Island pottery(Sears 1956:9).

Figure 3-4 shows the profiles of the northern (top) and southern (bottom) walls of the coregroup of contiguous units. Here, the modern topsoil layer or A horizon (Stratum I) appeared as ayellowish red sand loam. This layer was thicker than is typical of upland areas at Kolomoki, extendingabout 30 cm below the ground surface in the northwestern corner of the block in Unit D25N (Figure3-3, top and left) and slightly less deep in units downslope to the south and east.

Figure 3-3. Sear’s (1956:Figure I) depiction of a buriedmidden horizon at Kolomoki.

25

Figure 3-4. Northern (top) and southern (bottom) profiles of the core group of contiguous excavation units in Block D.

26

N

D22W D22ED24ND23ND21ND25N

I

IIFeature 194

Feature 193Feature 205

Feature 195

D8SD6WD6ED7WD7ED9S

I

II

Feature 157 Feature 159 Feature 163


Feature 158

III (unexcavated)

III (unexcavated)

0 10 50

cm

0 10 50

cmN

I = 5YR4/6 yellowish sand loamII = 5YR3/4 dark yellowish brown sandy loam

III = 2.5YR5/6 dark red sandy clay

Below the topsoil layer, we encountered aslightly darker horizon (Stratum II), consisting ofa dark yellowish brown sandy loam. Although theboundaries between this and Stratum I werediffuse, we generally noted an increase in thequantity of both charcoal and bone flecking. Insome areas, Stratum II was also mottled with adark red dark red sandy clay—the same soil typeas the underlying subsoil (Stratum III). Stratum IIwas generally about 10 cm thick in the core blockof contiguous units.

Stratum II appears similar to the buriedmiddens described by Sears. This layer couldrepresent an unplowed remnant of the A horizonthat existed before Euro-American settlement inthe area in the 1700s. However, in the course ofexcavating Stratum II, we noted probable plowscars running northwest-southeast at theintersection with the underlying sand clay subsoil(Stratum III) (Figure 3-5). This is the directionone would expect plow scars to run if the fieldwas plowed parallel with the tree line. This,coupled with the subsoil mottling, suggests thatwhile Stratum II probably represents a remnant ofan earlier A horizon, it appears to have beenplowed.

The following hypothetical scenario mayaccount for the unusual soil profiles in Block D, and perhaps also for similar profiles encountered bySears. In the nineteenth century, virtually the entire site was cleared for agriculture. Pockets of forestcover remained only around the steepheads and drainages, as indicated by later aerial photographs andmaps (see Pluckhahn 2003:Figure 3.1, Figure 3.3). We may never know precisely how deep the topsoillayer may have been before farming began—as noted above, however, Sears (1956:8) guessed that it wasone to three feet thick. Within a few years of the onset of farming, soil likely began eroding down slopefrom the flat plain west of Mound A which, based on aerial photographs and soil profiles, appears tohave been most intensively farmed prior to the creation of the state park. This colluvium settled in lowspots (such as those described by Sears) and breaks in slope (such as the area of Block D), burying theoriginal Ap horizon (Stratum II) in these areas. Continued farming plowed this colluvium, as well asthe soils that built up naturally over time from the decomposition of organic matter. Soil formationwould have been greatest in areas such as Block D along the margins of the field where leaf litteraccumulated, resulting in deeper modern topsoil layers (Stratum I).

In Units D2, D3, and D4—located just to the northwest of the core block—the soil profileappeared very similar to those of the core block, with a darker buried A horizon (Stratum II)sandwiched between the modern A horizon (Stratum I) and the clay subsoils (Stratum III) (Figure 3-6). Test Unit 18, located about 4 m to the south of the main block, also exhibited these same basic soilhorizons (Figure 3-7). Soil profiles were slightly different with increasing distance from the core blockof contiguous units. In Unit D1, located about 9 m up slope to the west, the modern A horizon wasmuch thinner, extending only about 20 cm below ground surface (Figure 3-8). The older A horizon wasalso less substantial, and barely discernible from Stratum I in color and texture. This description is alsocharacteristic of Unit D5, positioned about 13 m to the southeast of the core block (Figure 3-9). Here,the modern topsoil layer measured only about 10-15 cm thick. In both Units D1 and D5, we continuedexcavation into the underlying red clay subsoil. Excavation was quickly terminated, as no additionalstrata or features were encountered and artifact counts dropped precipitously.

Figure 3-5. Plan view of Unit D9 at base ofStratum II, showing probable plow scars.

27

Figure 3-6. The West Profile of Units D3 and D4.

28

N0 10 50

cm

I = 5YR4/6 yellowish sand loamII = 5YR3/4 dark yellowish brown sandy loam


III

III (excavated) III (unexcavated)III (unexcavated)

D3ND3SD4W

Figure 3-7. The East Profile of Test Unit 18.

29

N

I = 2.5YR2.5/2 very dusky red sand loamII = 5YR2.5/2 dark reddish brown sandy loam

III = 10R3/4 dusky red sandy clay

III (excavated)III (unexcavated)

0 5 20

cm

II

I

III (excavated)

Features 38 and 39Disturbance

Figure 3-8. The West Profile of Unit D1.

30

N

I = 5YR4/6 yellowish red sand loamII = 2.5YR2.5/4 dark reddish brown sand loam


0 5 20

cm

I

II

III

Figure 3-9. The South Profile of Unit D5.

31

N

I = 2.5YR3/4 dark reddish brown sand loamII = 2.5YR4/4 dark reddish brown sand loam

III = 2.5YR4/6 red sandy clay

0 5 20

cm

III

III

Figure 3-10. The density of pottery in the coreunits of Block D.

Figure 3-11. The density of Swift Creek in thecore units of Block D.

Artifacts

A detailed description of the artifact assemblage from Block D is presented in Chapter 5. Here,I focus on patterns in the horizontal distribution of artifacts in excavation unit levels, omitting for themoment artifacts recovered from features.

Prehistoric Ceramics

The 52 m2 of excavation in Block D(including Test Unit 18) yielded 13,932 sherds(weighing 33,628.69 g) for a mean of around 268sherds per square meter. Table 3-1 summarizes thepottery from excavation unit levels. Not included inthe table, in addition to pottery from features, aresherds from mixed proveniences (such as grabcollections) and balk removal, as well as a few piece-plotted sherds. Focusing on the core units and thoselocated nearby, Figure 3-10 demonstrates that thedensity of pottery was greatest in Test Unit 18, wherea large number of residual sherds increased the totalconsiderably. Sherd density was also relatively highin units D2, D3, and D4, to the northwest. Both ofthese areas are located outside the presumedstructure described in the chapter that follows. Within the core group of contiguous units itself,pottery density was highest in three 1-x-1-m units(D15N, D16N, D19N) near the center of thepresumed structure, adjacent to a feature (Feature171) that may represent a hearth.

Most of the individual pottery types in theBlock D assemblage are represented by only a limitednumbers of sherds in our excavation levels (far fewerin most units), thus restricting the utility ofcomparing their spatial distributions. However, thereare a few exceptions. Three decorative types (SwiftCreek, Carrabelle Punctate, and Weeden Island Red)exhibit more total sherds (>50) and bear closerscrutiny. Swift Creek Complicated Stamped is themost common of the named pottery types identifiedin Block D. It is also widely distributed, withexamples present in the assemblages from everyexcavated square in the block. Swift Creek sherds aremost dense in Test Unit 18, just to the south of thecore group of units in Block D (Figure 3-11). Theyare also common in five of the six units to thenorthwest. Within the core group of contiguousunits, Swift Creek sherds are less dense in 1-x-1-msquares in the northeast section of the block.

32

Table 3-1. Prehistoric Ceramics from Unit Levels.U

nit

Lev

el

resi

dual

plai

n sa

nd

plai

n lim

esto

ne

St J

ohns

Pla

in

Swift

Cre

ek

Nap

ier

Car

rabe

lle I

ncis

ed

Car

rabe

lle P

unct

ate

Indi

an P

ass

Inci

sed

Mou

nd F

ield

Net

Mar

ked

Kei

th I

ncis

ed

Wes

t Flo

rida

Cor

d M

arke

d

Tuc

ker

Rid

ge P

inch

ed

WI

Inci

sed

WI

Red

Film

ed

Wak

ulla

Che

ck S

tam

ped

Wee

den

Isla

nd Z

oned

Inc

/Pun

ct

Oth

er W

oodl

and

UID

Inc

ised

UID

Sta

mpe

d

UID

Pun

ctat

e

Lam

ar P

lain

Lam

ar C

omp

Stam

ped

Tot

al

D1N

1 32 21 1 1 1 56

2 79 46 4 1 1 1 132

3 41 23 5 1 2 1 73

4 8 2 1 11

Total 160 92 9 4 1 3 0 1 1 1 272

D1S

1 65 41 8 1 1 116

2 47 31 7 1 1 87

3 60 21 7 1 1 90

4 11 7 5 23

Total 183 100 27 1 1 1 3 316

D2N

1 31 20 2 1 1 55

2 102 53 4 2 1 162

3 62 44 5 1 1 2 115

4 8 4 1 13

Total 203 121 11 2 1 4 3 345

D2S

1 35 22 1 1 1 60

2 81 58 3 1 5 1 2 151

3 85 54 1 5 2 4 1 152

4 29 26 1 1 1 3 1 62

Total 230 160 1 10 2 1 1 2 13 1 1 1 2 425

D3N

1 82 36 7 1 126

2 84 46 5 2 2 1 3 1 144

3 81 65 6 2 1 1 1 157

4 30 13 1 44

Total 277 160 19 4 1 1 2 3 3 1 471

D3S

1 67 23 2 1 2 95

2 119 63 5 3 2 5 197

3 111 48 1 1 2 163

4 76 26 1 3 2 1 19

Total 373 160 9 6 4 7 3 2 564

D4E

1 8 4 1 13

2 36 13 1 1 51

3 78 37 6 1 1 3 3 129

33


nit

Lev

el

resi

dual

plai

n sa

nd

plai

n lim

esto

ne

St J

ohns

Pla

in

Swift

Cre

ek

Nap

ier

Car

rabe

lle I

ncis

ed

Car

rabe

lle P

unct

ate

Indi

an P

ass

Inci

sed

Mou

nd F

ield

Net

Mar

ked

Kei

th I

ncis

ed

Wes

t Flo

rida

Cor

d M

arke

d

Tuc

ker

Rid

ge P

inch

ed

WI

Inci

sed

WI

Red

Film

ed

Wak

ulla

Che

ck S

tam

ped

Wee

den

Isla

nd Z

oned

Inc

/Pun

ct

Oth

er W

oodl

and

UID

Inc

ised

UID

Sta

mpe

d

UID

Pun

ctat

e

Lam

ar P

lain

Lam

ar C

omp

Stam

ped

Tot

al

D4E4 121 46 4 1 2 5 1 1 181

Total 243 100 12 3 2 1 8 1 4 374

D4W

1 17 4 1 1 1 24

2 55 19 3 3 8

3 103 30 1 1 2 1 138

4 31 8 1 40

Total 206 61 5 2 2 1 5 282

D5E

1 19 7 1 27

2 98 29 2 2 1 2 134

3 97 26 3 1 3 130

Total 214 62 6 2 1 3 3 291

D5W

1 45 38 2 1 1 1 1 89

2 110 53 1 1 1 4 1 171

Total 155 91 3 2 1 5 2 1 260

D6E

1 13 6 19

2 38 26 6 2 1 73

3 63 29 1 93

Total 114 61 6 2 2 185

D6W

1 18 17 3 38

2 77 11 3 1 1 1 94

Total 95 28 6 1 1 1 132

D7E

1 28 17 1 1 1 48

2 34 22 1 57

3 41 15 1 3 1 61

Total 103 54 1 1 1 4 1 1 166

D7W

1 27 7 5 1 1 41

2 38 21 5 1 1 66

3 51 17 2 2 1 2 75

Total 116 45 12 3 1 3 2 182

D8N

1 9 1 10

2 47 25 3 1 1 1 78

3 68 18 1 1 88

Total 124 43 4 1 2 1 1 176

34


nit

Lev

el

resi

dual

plai

n sa

nd

plai

n lim

esto

ne

St J

ohns

Pla

in

Swift

Cre

ek

Nap

ier

Car

rabe

lle I

ncis

ed

Car

rabe

lle P

unct

ate

Indi

an P

ass

Inci

sed

Mou

nd F

ield

Net

Mar

ked

Kei

th I

ncis

ed

Wes

t Flo

rida

Cor

d M

arke

d

Tuc

ker

Rid

ge P

inch

ed

WI

Inci

sed

WI

Red

Film

ed

Wak

ulla

Che

ck S

tam

ped

Wee

den

Isla

nd Z

oned

Inc

/Pun

ct

Oth

er W

oodl

and

UID

Inc

ised

UID

Sta

mpe

d

UID

Pun

ctat

e

Lam

ar P

lain

Lam

ar C

omp

Stam

ped

Tot

al

D8S

1 11 9 3 1 24

2 64 15 3 1 1 84

3 91 24 9 1 3 3 131

Total 166 48 15 1 4 4 1 239

D9N

1 23 11 1 35

2 85 29 1 1 2 118

3 59 22 3 1 1 1 2 89

Total 167 62 4 2 1 1 3 2 242

D9S

1 18 9 2 29

2 68 18 2 1 1 1 91

3 72 8 4 1 1 158

Total 158 107 8 2 1 1 1 278

D10N

1 79 35 2 3 2 4 2 127

2 62 45 4 2 113

Total 141 8 6 3 4 4 2 240

D10S

1 71 33 1 3 1 109

2 98 31 1 1 2 2 135

Total 169 64 1 1 1 5 3 244

D11N

1 6 2 5 2 87

2 66 26 1 1 3 1 1 108

3 58 33 4 1 3 99

Total 184 79 19 2 5 1 4 294

D11S

1 23 2 6 2 51

2 55 22 4 1 1 1 84

3 48 23 1 1 73

Total 126 65 11 1 4 1 208

D12E

1 4 1 5

2 32 19 8 1 60

3 59 23 5 2 1 5 95

Total 95 42 13 2 1 7 160

D12W

1 22 12 1 35

2 36 21 6 1 64

3 54 35 3 4 96

35


nit

Lev

el

resi

dual

plai

n sa

nd

plai

n lim

esto

ne

St J

ohns

Pla

in

Swift

Cre

ek

Nap

ier

Car

rabe

lle I

ncis

ed

Car

rabe

lle P

unct

ate

Indi

an P

ass

Inci

sed

Mou

nd F

ield

Net

Mar

ked

Kei

th I

ncis

ed

Wes

t Flo

rida

Cor

d M

arke

d

Tuc

ker

Rid

ge P

inch

ed

WI

Inci

sed

WI

Red

Film

ed

Wak

ulla

Che

ck S

tam

ped

Wee

den

Isla

nd Z

oned

Inc

/Pun

ct

Oth

er W

oodl

and

UID

Inc

ised

UID

Sta

mpe

d

UID

Pun

ctat

e

Lam

ar P

lain

Lam

ar C

omp

Stam

ped

Tot

al

D12W Total 112 68 9 1 4 1 195

D13E

1 7 32 1 1 1 1 3 118

2 48 32 2 1 1 2 2 88

Total 118 64 1 12 1 2 2 3 3 206

D13W

1 51 4 7 1 4 2 105

2 77 44 3 3 4 1 4 136

Total 128 84 1 1 3 8 1 6 241

D14

1 24 9 33

2 28 9 3 1 41

3 38 19 2 1 1 4 65

4 44 27 2 1 1 1 76

Total 134 64 7 2 2 1 1 4 215

D15N

1 193 65 5 1 3 1 268

2 46 24 2 2 1 75

3 8 6 1 15

Total 247 95 8 1 2 3 1 1 358

D15S

1 48 26 1 1 1 1 78

2 86 23 1 2 1 113

3 2 12 1 1 1 35

Total 154 61 2 2 3 1 1 1 1 226

D16N

1 113 5 4 2 1 1 4 175

2 82 32 3 2 5 2 126

Total 195 82 7 4 1 6 6 301

D16S

1 95 36 4 1 1 137

2 69 37 5 1 4 6 122

Total 164 73 9 2 5 6 259

D17N

1 3 31 6 2 1 70

2 49 28 1 1 79

3 2 14 1 4 39

Total 99 73 8 2 1 4 1 188

D17S

1 37 37 9 1 84

2 3 22 25

3 35 13 1 2 1 52

36


nit

Lev

el

resi

dual

plai

n sa

nd

plai

n lim

esto

ne

St J

ohns

Pla

in

Swift

Cre

ek

Nap

ier

Car

rabe

lle I

ncis

ed

Car

rabe

lle P

unct

ate

Indi

an P

ass

Inci

sed

Mou

nd F

ield

Net

Mar

ked

Kei

th I

ncis

ed

Wes

t Flo

rida

Cor

d M

arke

d

Tuc

ker

Rid

ge P

inch

ed

WI

Inci

sed

WI

Red

Film

ed

Wak

ulla

Che

ck S

tam

ped

Wee

den

Isla

nd Z

oned

Inc

/Pun

ct

Oth

er W

oodl

and

UID

Inc

ised

UID

Sta

mpe

d

UID

Pun

ctat

e

Lam

ar P

lain

Lam

ar C

omp

Stam

ped

Tot

al

D17S Total 75 72 1 3 1 161

D18

1 64 41 2 2 2 1 112

2 24 17 1 1 43

Total 88 58 2 1 2 1 2 1 155

D19N

1 107 55 5 1 2 1 171

2 91 64 3 4 3 1 166

Total 198 119 8 1 6 1 3 1 337

D19S

1 61 28 4 1 94

2 98 7 1 2 2 1 174

Total 159 98 5 2 3 1 268

D20E

1 4 22 2 1 2 67

2 91 48 1 1 5 3 1 1 151

3 9 12 2 1 24

Total 14 82 5 1 1 5 3 1 1 1 2 242

D20W

1 54 1 1 1 1 2 3 72

2 53 21 1 1 3 79

3 1 6 1 17

Total 117 37 2 1 2 1 2 6 168

D21N

1 26 15 1 42

2 51 34 2 1 4 92

3 11 24 3 38

Total 88 73 5 1 4 1 172

D21S

1 37 34 5 1 1 1 79

2 71 16 1 2 90

3 33 35 2 1 2 73

Total 141 85 5 3 1 4 3 242

D22E

1 103 34 1 2 1 2 1 144

2 86 19 3 2 2 112

189 53 4 2 3 2 3 256

D22W

1 87 38 4 1 1 3 1 135

2 62 24 1 4 1 3 95

Total 149 62 1 8 1 1 1 6 1 230

D23N 1 36 24 1 1 62

37


nit

Lev

el

resi

dual

plai

n sa

nd

plai

n lim

esto

ne

St J

ohns

Pla

in

Swift

Cre

ek

Nap

ier

Car

rabe

lle I

ncis

ed

Car

rabe

lle P

unct

ate

Indi

an P

ass

Inci

sed

Mou

nd F

ield

Net

Mar

ked

Kei

th I

ncis

ed

Wes

t Flo

rida

Cor

d M

arke

d

Tuc

ker

Rid

ge P

inch

ed

WI

Inci

sed

WI

Red

Film

ed

Wak

ulla

Che

ck S

tam

ped

Wee

den

Isla

nd Z

oned

Inc

/Pun

ct

Oth

er W

oodl

and

UID

Inc

ised

UID

Sta

mpe

d

UID

Pun

ctat

e

Lam

ar P

lain

Lam

ar C

omp

Stam

ped

Tot

al

D23N

2 106 44 3 2 2 157

3 9 7 1 1 2 20

Total 151 75 3 1 1 1 2 2 1 2 239

D23S

1 35 13 3 1 52

2 89 39 2 1 131

3 26 25 1 1 3 2 58

Total 15 77 3 1 1 6 3 241

D24N

1 47 13 3 1 1 65

2 105 31 1 4 2 2 145

3 63 21 1 3 2 1 91

Total 215 65 4 6 5 3 3 301

D24S

1 37 11 1 49

2 7 23 1 1 3 2 3 2 105

3 96 33 3 1 1 134

Total 203 67 5 1 1 4 2 3 2 288

D25N

1 5 62 5 4 2 3 126

2 55 33 4 1 93

3 24 13 1 1 39

Total 129 108 9 6 2 4 258

D25S

1 11 29 1 1 42

2 44 31 2 1 1 1 1 2 83

3 41 26 3 1 1 2 74

Total 96 86 6 1 1 1 2 3 1 2 199

18

1 343 183 28 1 1 1 3 2 2 5 1 570

2 33 185 2 1 2 1 1 6 1 1 548

3 267 134 11 1 3 1 417

4 39 37 9 1 86

Total 979 539 68 1 3 6 3 3 1 3 12 2 1 1621

38

Figure 3-12. The density of Weeden IslandRed in the core units in Block D.

Figure 3-13. The density of CarrabellePunctate in the core units in Block D.

Figure 3-14. The density of Lamar in thecore units in Block D.

Weeden Island Red displays some similaritiesto Swift Creek in its distribution (Figure 3-12). As withSwift Creek, concentrations are apparent in Test Unit18 and several of the 1-x-1-m squares to the northwestof the core block. However, in contrast with SwiftCreek, Weeden Island Red pottery is more evenlydistributed in squares in the central northeasternportion of the core block, within the presumedstructure described below. Several of the other typesof the Weeden Island series—particularly WeedenIsland Incised, Carrabelle Incised, and Tucker RidgePinched—also display slightly higher densities in unitsin this portion of the block. This pattern could havefunctional significance. As described in more detail inChapter 5, Weeden Island Red is more stronglyassociated with bowls, and serving vessels such asthese might be expected to concentrate within thestructure.

Carrabelle Punctate is the third most commonformal type in the ceramic assemblage from excavationunit levels. As indicated in Figure 3-13, sherds of thistype exhibit only marginal clustering in excavation unitsin the core block. Specifically, slight concentrations areapparent in four units scattered between the extremenorth, north-central, and extreme southwest. Outsidethe core block, there are slight concentrations ofCarrabelle sherds in two units to the northwest and inTest Unit 18 to the south. The wide, yet infrequent, distribution of Carrabelle Punctate sherds is consistentwith the hypothesis that this pottery type occurs insmall quantities throughout much of the occupation atKolomoki, as Pluckhahn has elsewhere suggested2003:Table 2.2, 95).

We recovered only 16 sherds of Lamar potteryfrom excavation unit levels, including 12 Lamar Plainand four Lamar Complicated Stamped. Lamar Potteryclusters primarily in 1-x-1-m units in the northernportion of the core block and in two of the six 1-x-1-m units to the northwest (Figure 3-14). Searsrecovered a relatively larger amount of Lamar potteryfrom his Unit 1, which appears to have been locateda short distance downslope to the east of Block D,closer to the creek (Pluckhahn 2003: 74-77; Sears1951). Thus, a light scatter of Lamar pottery in BlockD is not surprising. Overall, the limited quantity ofLamar pottery from Block D and elsewhere atKolomoki is consistent with Sears’s (1951:4)observation that this component represents “only asmall group, occupying a limited area for a shortperiod.”

39

Figure 3-15. The density of flaked stone inthe core units of Block D.

Figure 3-16. The density ofquartzite/sandstone in the core units ofBlock D.

Prehistoric Flaked Stone

We recovered 3892 pieces of flaked stoneweighing a collective 8584.3 g from the 52 m2 ofexcavation in Block D (including Test Unit 18). This yields a mean density of around 75 (or about165 g) flaked stone artifacts per square meter. Table3-2 summarizes the flaked stone from excavationunit levels. As with the pottery summary table(Table 3-1) this does not include artifacts fromfeatures, mixed proveniences (such as grabcollections), or balk removal. The density of flakedstone was remarkably even throughout the coreblock, with slightly higher densities in the isolatedunits to the northwest and south (Figure 3-15). Asnoted in the chapter that follows, we identified apossible structure in the core block; the relative lackof lithics in this area could indicate that toolmaintenance and manufacture were deliberatelyundertaken outside the structure, or that thestructure was periodically swept clean.

The flake stone assemblage from excavationunit levels is comprised principally of Coastal Plaincherts (N=2787, or 71.61 percent). Notsurprisingly, the distribution of Coastal Plain chertartifacts largely mirrors that of flaked stone ingeneral. Quartz is next in frequency (N=954, or24.51 percent), and also follows the same generalspatial pattern.

More interesting is the distribution ofquartzite/sandstone (N=148, or 3.80 percent)(Figure 3-16). In contrast with Coastal Plain chertand quartz, artifacts of this material are betterrepresented within the core block, perhaps indicativeof manufacture or maintenance of tools within thepresumed structure. Only a small quantity of Ridgeand Valley chert artifacts (N=5, or 0.13 percent)were recovered from excavation unit levels in BlockD; thus, a comparison of the spatial distribution ofthis raw material type would be of very limitedutility.

40

Table 3-2. Flaked Stone from Unit Levels.Unit Level quartz quartzite RV chert Coast Plain chert

prim

ary

seco

ndar

y

shat

ter/

chun

k

tert

iary

core

/cor

etoo

l

flake

tool

bifa

ce

prim

ary

seco

ndar

y

shat

ter/

chun

k

tert

iary

core

/cor

e to

ol

flake

tool

bifa

ce

tert

iary

bifa

ce

prim

ary

seco

ndar

y

shat

ter/

chun

k

tert

iary

core

/cor

e to

ol

flake

tool

bifa

ce

Tot

al

D1N

1 2 5 4 1 1 2 2 17

2 3 12 16 6 1 3 5 1 1 57

3 2 1 3 4 7 2 19

4 6 6

Total 3 14 23 11 1 7 1 25 5 99

D1S

1 3 2 6 3 2 2 5 23

2 5 3 3 1 4 7 1 24

3 4 6 6 2 2 13 33

4 1 1 4 6

Total 12 12 16 3 5 8 29 1 86

D2N

1 1 1 1 1 1 5 10

2 2 1 3 2 2 23 42

3 1 3 1 2 7 7 31 2 54

4 1 1

Total 4 14 5 1 2 9 1 6 2 107

D2S

1 2 3 1 1 6 13

2 2 4 1 1 1 1 6 4 13 1 34

3 1 5 2 8 1 1 1 3 2 5 25 54

4 2 1 5 3 12 23

Total 1 7 1 12 2 1 4 5 13 12 56 1 124

D3N

1 2 2 1 3 4 9 21

2 4 2 1 1 1 1 1 2 4 5 8 30

3 3 6 2 1 5 1 19 37

4 1 7 8

Total 5 7 9 1 1 3 1 4 12 1 43 96

D3S

1 5 3 1 2 1 2 12 26

2 6 8 1 3 5 37 1 61

3 2 1 1 2 2 1 49 67

4 1 2 1 1 2 24 31

Total 14 23 2 1 7 4 1 122 1 185

D4E

1 1 1 1 3

2 1 3 1 2 7

3 3 1 9 1 2 3 26 1 46

4 1 1 5 1 4 13 24 2 51

Total 1 3 3 17 1 1 7 17 53 1 3 107

D4W1 1 1 2

2 5 4 2 4 11 26

3 2 4 5 1 1 1 13 1 28

4 1 1 1 4 7 1 15

41


prim

ary

seco

ndar

y

shat

ter/

chun

k

tert

iary

core

/cor

etoo

l

flake

tool

bifa

ce

prim

ary

seco

ndar

y

shat

ter/

chun

k

tert

iary

core

/cor

e to

ol

flake

tool

bifa

ce

tert

iary

bifa

ce

prim

ary

seco

ndar

y

shat

ter/

chun

k

tert

iary

core

/cor

e to

ol

flake

tool

bifa

ce

Tot

al

D4W Total 2 1 11 1 3 2 9 31 2 71

D5E

1 1 2 1 4

2 1 1 2 1 14

3 1 4 1 1 1 2 7 17

Total 1 1 5 1 1 1 3 4 17 1 35

D5W

1 1 3 1 4 2 3 1 24

2 1 2 3 6 19 31

Total 1 3 1 1 6 5 9 29 55

D6E

1 1 1

2 1 2 5 1 1 1 4 7 22

3 1 2 1 2 1 2 5 13 27

Total 2 2 3 7 1 1 2 3 9 2 50

D6W

1 1 1 2 4

2 1 1 4 1 2 1 11 21

Total 2 2 4 1 4 1 11 25

D7E

1 2 3 2 1 1 1 4 4 1 19

2 2 5 1 2 4 2 9 25

3 1 1 5 14 1 22

Total 4 9 4 1 3 1 6 27 1 1 66

D7W

1 1 4 1 1 2 8 1 18

2 1 3 5 5 1 24

3 1 1 2 1 1 4 2 23 1 36

Total 1 1 4 8 1 1 1 9 41 2 78

D8N

1 1 1

2 2 3 3 1 2 3 1 24

3 1 2 2 3 11 19

Total 2 4 4 3 2 2 6 21 44

D8S

1 1 1 1 3

2 1 3 2 6 7 8 27

3 5 3 3 1 2 8 4 15 1 42

Total 6 6 5 2 1 2 14 11 24 1 72

D9N

1 1 1 2 5 1 10

2 2 5 2 4 2 1 34

3 2 4 4 4 14 28

Total 3 2 1 2 4 1 39 2 72

D9S

1 1 2 3 6

2 3 3 6 9 2 14 37

3 1 1 2 2 1 13 2 14 1 37

Total 1 4 5 8 1 1 24 4 31 1 80

42


prim

ary

seco

ndar

y

shat

ter/

chun

k

tert

iary

core

/cor

etoo

l

flake

tool

bifa

ce

prim

ary

seco

ndar

y

shat

ter/

chun

k

tert

iary

core

/cor

e to

ol

flake

tool

bifa

ce

tert

iary

bifa

ce

prim

ary

seco

ndar

y

shat

ter/

chun

k

tert

iary

core

/cor

e to

ol

flake

tool

bifa

ce

Tot

al

D10N1 1 2 2 1 2 1 2 5 6 11 33

2 1 3 3 1 2 1 7 2 12 32

Total 2 2 5 3 1 1 4 1 3 12 8 23 65

D10S1 1 1 1 4 9 3 6 1 26

2 2 3 2 1 3 6 26 1 44

Total 1 1 3 7 2 1 12 9 32 2 70

D11N

1 3 3 2 2 5 15

2 1 1 1 3 1 1 8 5 4 1 26

3 1 4 1 1 4 7 2 15 1 36

4 1 1 1 3

Total 2 2 4 11 1 1 1 5 17 9 25 2 80

D11S

1 3 2 1 2 8

2 1 6 5 1 6 3 7 1 30

3 1 5 3 2 2 2 8 4 18 45

Total 1 6 12 7 3 2 16 8 27 1 83

D12E

1 1 1

2 1 1 3 4 2 5 8 24

3 2 3 2 3 1 3 1 1 1 4 13 1 44

Total 3 4 5 7 1 3 1 1 12 9 22 1 69

D12W

1 1 1 2

2 1 2 2 1 3 1 11 21

3 1 2 3 1 1 1 1 7 5 16 38

Total 1 4 6 3 1 1 1 1 1 6 27 61

D13E1 1 4 3 3 1 3 4 1 11 31

2 1 2 5 1 1 3 2 12 27

Total 2 6 8 4 1 4 7 3 23 58

D13W1 1 3 3 4 1 2 8 6 11 1 40

2 3 4 2 1 1 2 3 2 25 3 46

Total 1 6 7 6 1 2 4 11 8 36 1 3 86

D14

1 1 2 1 1 5

2 2 1 1 1 4 9

3 1 5 7 1 1 6 3 11 35

4 2 2 1 1 4 2 4 16

Total 1 3 8 8 1 3 1 1 13 6 2 65

D15N

1 4 5 4 6 12 9 16 56

2 2 2 4 5 4 17

3 1 2 3 6

Total 4 7 7 6 18 14 23 79

D15S 1 1 3 2 8 14

43


prim

ary

seco

ndar

y

shat

ter/

chun

k

tert

iary

core

/cor

etoo

l

flake

tool

bifa

ce

prim

ary

seco

ndar

y

shat

ter/

chun

k

tert

iary

core

/cor

e to

ol

flake

tool

bifa

ce

tert

iary

bifa

ce

prim

ary

seco

ndar

y

shat

ter/

chun

k

tert

iary

core

/cor

e to

ol

flake

tool

bifa

ce

Tot

al

D15S2 1 2 2 1 5 16 1 28

3 1 1 2 1 1 5 6 4 21

Total 1 3 2 4 1 2 13 8 28 1 63

D16N1 3 6 3 7 18 1 38

2 3 5 1 2 2 8 12 33

Total 6 11 4 2 9 8 3 1 71

D16S1 1 6 3 2 1 5 1 9 1 29

2 1 3 5 4 9 7 17 46

Total 1 4 11 7 2 1 14 8 26 1 75

D17N

1 1 2 1 5 1 7 2 19 17

2 1 1 1 2 6 9 1 30

3 3 1 1 1 5 2 6 19

Total 1 3 4 6 1 2 1 3 18 13 35 1 88

D17S

1 4 6 1 2 8 1 22

2 1 1 4 3 5 14

3 1 2 3 3 5 1 1 25

Total 4 7 1 1 2 4 9 8 23 2 61

D18

1 1 4 1 1 3 3 2 5 1 21

2 1 2 1 3 7 14

Total 1 5 3 1 4 6 2 12 1 35

D19N1 1 5 5 3 6 5 13 1 39

2 4 1 1 1 2 1 9 8 33 1 61

Total 1 9 6 1 1 5 1 15 13 46 1 1 100

D19S1 1 1 3 1 4 2 4 16

2 3 4 1 2 1 2 1 8 2 51

Total 4 5 4 3 1 2 14 1 24 67

D20E

1 1 1 2 1 2 6 7 5 25

2 2 1 1 1 1 1 2 5 13 1 28

3 1 1 2 6 1 1 12

Total 3 2 4 1 1 1 4 1 18 19 1 1 65

D20W

1 1 4 2 3 4 11 1 26

2 2 1 2 2 2 2 14 1 26

3 1 1 2

Total 1 3 5 2 4 5 6 26 1 1 54

D21N

1 1 1 1 3 2 8

2 1 1 1 3 5 13 24

3 1 2 3 1 2 6 1 4 20

Total 2 3 4 2 1 3 9 9 19 52

D21S 1 1 2 1 1 1 6 1 3 16

44


prim

ary

seco

ndar

y

shat

ter/

chun

k

tert

iary

core

/cor

etoo

l

flake

tool

bifa

ce

prim

ary

seco

ndar

y

shat

ter/

chun

k

tert

iary

core

/cor

e to

ol

flake

tool

bifa

ce

tert

iary

bifa

ce

prim

ary

seco

ndar

y

shat

ter/

chun

k

tert

iary

core

/cor

e to

ol

flake

tool

bifa

ce

Tot

al

D21S2 2 1 2 1 1 1 2 1 20

3 3 1 7 1 1 4 6 14 1 47

Total 3 3 6 1 1 8 11 2 11 9 27 1 83

D22E1 1 2 3 1 1 2 1 1 1 1 6 6 16 2 44

2 1 2 1 2 5 9 16 36

Total 1 2 4 3 1 3 1 1 1 3 11 15 32 2 80

D22W1 3 5 4 1 5 1 28

2 3 1 1 7 5 16 33

Total 3 5 7 1 1 1 12 15 16 61

D23N

1 2 3 1 3 9 18

2 1 3 3 2 1 4 4 4 17 39

3 2 1 3 2 8

Total 1 5 6 4 3 4 7 7 28 65

D23S

1 1 2 3 1 3 2 12

2 2 4 2 1 1 4 5 6 1 26

3 1 3 1 1 2 1 9 18

Total 3 6 5 1 1 4 1 2 6 9 17 1 56

D24N

1 1 1 2 6 2 6 1 19

2 4 4 1 1 2 5 3 12 2 34

3 1 4 2 1 1 1 1 20

Total 1 9 5 1 1 6 12 6 28 1 3 73

D24S

1 1 1 1 1 2 6

2 3 1 1 2 5 5 1 1 19

3 6 2 5 1 1 6 4 19 44

Total 7 2 9 1 1 1 1 1 8 1 26 1 1 69

D25N

1 4 3 4 3 1 7 5 11 38

2 2 2 2 5 2 1 2 16 32

3 2 2 2 2 1 1 1 8 19

Total 6 7 8 1 2 1 4 9 7 35 89

D25S

1 3 2 4 2 1 1 7 20

2 4 5 1 1 3 2 2 4 22

3 1 1 4 5 9 7 1 28

Total 8 7 6 1 9 8 12 18 1 70

18

1 2 6 13 14 2 3 3 2 17 51 2 1 6 140

2 6 7 1 1 1 4 2 25 2 62 2 1 3 144

3 5 8 6 2 3 5 15 12 74 130

4 1 2 1 1 2 2 1 4 14

Total 2 17 29 32 3 1 3 11 12 62 5 191 4 2 9 428

45

Figure 3-17. The density of FCR in the coreunits of Block D.

Figure 3-18. The density of groundstone inthe core units of Block D.

Miscellaneous Prehistoric Artifacts

Excavation of the 52 1-x-1-m units in BlockD produced a variety of other prehistoric artifacts,in addition to pottery and flaked stone (Table 3-3). I describe these in greater detail in Chapter 5, withthe focus here on distributional patterns in a few ofthe more prevalent artifact classes. Most prevalentis FCR, with a total of 142 fragments weighing5,398.6 g in the excavation unit levels in Block D. As indicated in Figure 3-17, several areas ofconcentration in the distribution of FCR areapparent within the core block. Two of thesecorrespond closely with the locations of featuresthat may have functioned for cooking, as describedin the chapter that follows. It should be noted,however, that FCR is difficult to differentiate fromnaturally-occurring rock at Kolomoki, and the datahere no doubt reflect some subjectivity in collection.

We recovered a modest quantity ofgroundstone from excavation unit levels (N=8,weighing 695.7 g), and all of these artifacts arefragmentary. Nevertheless, the distribution ofgroundstone in excavation unit levels (Figure 3-18)is intriguing relative to the possible structural patternand inferred feature functions described in thechapter that follows. Artifacts of this type wererestricted to units within the presumed structure. Inaddition, as with FCR, the occurrence ofgroundstone artifacts corresponds closely withseveral features that may be associated with cookingand food processing. Most of the groundstoneconsists of igneous rocks such as greenstone andgranite that are uncommon in the Coastal Plain,making the identification of this class of artifacts lesssubjective than for FCR. Nevertheless, some locallyoccurring sandstones are also present in thegroundstone category.

Another prominent category of miscellaneous prehistoric artifacts from our excavation unitlevels consists of quartz pebbles (N=28, 497.0 g). As described in more detail in Chapter 5, theseare small, water-worn rocks of clear or, less commonly, milky quartz of the same type as much ofthe quartz debitage. However, these pebbles are generally unworked, although a few are simplybroken. They may be raw material that was stored for later flaked stone tool production. Alternatively, or perhaps in addition, they may have been used for other purposes, although noneare extensively battered, cracked, or fire-reddened. We also recovered one quartz crystal (3.5 g).

A variety of other artifacts, presumably associated with the prehistoric occupation of theBlock D area, were recovered from excavation unit levels. These include: mica (6.8 g); pecked stone(N=4, 2804.9 g); fired clay (N=4, 47.7 g); possible pigment stone (N=2, 5.0 g); and possible workedsandstone (N=2, 2514 g).

46

Table 3-3. Miscellaneous Prehistoric Artifacts from Unit Levels.

Unit Level

mic

a

grou

ndst

one

peck

ed s

tone

limes

tone

FC

R

fired

cla

y

quar

tz p

ebbl

e

sand

ston

e ba

ll

quar

tz c

ryst

al

poss

ible

pig

men

t sto

ne

g N g N g N g N g N g N g N g N g N g

D1N 1

D1N 2 0.1 2 46.8

D1S 1 1 7.1

D1S 3 0.4

D2N 1 1 24.6

D2N 3 1 3.7

D2S 2 1 236

D3N 2 1 248.8

D3N 3 1 18.6

D3N 4 1 34

D3S 3 1 24.3

D3S 4 1 4.5 1 28

D4E 4 1 10.3 1 107.8 1 6.7

D4W 3 1 3.4

D4W 4 1 2

D5E 2 2 238 1 23.2

D5E 3 1 4

D5W 1 1 122.4 1 64.7

D5W 2 1 6.8

D6E 1 2 114.8

D6E 3 8 95.5

D7E 2 1 32.2

D7E 3 1 18.2

D7W 1 1 9.3

D7W 3 3 30.6 1 17.1

D8S 3

D9S 3 1 61.1

D10N 1 2 235

D10N 2 1 3.5

D11N 1

D11N 3 1 134.1

D11S 2 3 19.1

D12E 3 11 219.2

D12E 2 0.1

D12W 1 1 8.8

D12W 2 1 17.5

D12W 3 11 67.5

D13E 1 1 41.5 9 257.8

D13E 2 3 308 1 4.2

D13W 1 2 342.6 1 19

D13W 2 1 155.7

D14 2 1 115.8

D14 3 0.5 4 53.5 1 14.8

47

Table 3-3. Miscellaneous Prehistoric Artifacts from Unit Levels.

Unit Level

mic

a

grou

ndst

one

peck

ed s

tone

limes

tone

FC

R

fired

cla

y

quar

tz p

ebbl

e

sand

ston

e ba

ll

quar

tz c

ryst

al

poss

ible

pig

men

t sto

ne

g N g N g N g N g N g N g N g N g N g

D14 4 6 38.6 1 8.2

D15N 1 1 5 1 37.4

D15N 3 1 11.2

D15S 1 2 212

D15S 2 1 21

D15S 3 0.1 1 18.2

D16N 1 2.6 3 57

D16N 2 3 81.7

D16S 1 1 199.7

D16S 2 1 138.5

D17N 1 3 17.8

D17S 1 3 49.5

D17S 3 1 95.8 2 61.7

D18 2 6 49.4 1 2.1

D18 1 5 41.9

D19N 1 1 5.5 3 24.8

D19S 1 1 24.1

D20E 1

D20E 2 1 127.3

D20W 1 1 33.1 2 333.8

D21N 2 1 25.1

D21N 3 2 13.4 1 11.5

D22E 1 1 0.6 7 21.7

D22E 2 1 9.7

D22W 1

D23S 1 1 8

D23S 1 59.3

D24S 1 1 14.8

D24N 1 2 1.5

D25N 1 1 2495 2 112.4 1 96.5 1 2495

D25N 2 1 16.4

D25S 2 3 17

D25S 3 1 49.3

D25S 1 1 0.8

18 1 1 1 146 3 481 1 3 1

18 2 1 9 3 167

18 3 1 1 44 3 51

48

Figure 3-19. The density of historic artifactsin the core units in Block D.

Historic Artifacts

The Block D excavations produced anassemblage of historic artifacts that is surprisinglylarge and diverse for Kolomoki, albeit modest formost historic sites (Table 3-4). The majority ofthese artifacts were found in the first two excavationlevels, corresponding to the modern plowzone andcolluvium, as described above. There is no apparentclustering in the distribution of historic materials(Figure 3-19).

Bottle glass makes up the majority of thehistoric artifact assemblage, and most of this is oliveor dark green (N=6, 7.5 g). Only a few sherds ofamber (N=2, 1.5 g) and clear (N=1, 0.8 g) glass wererecovered.

Nails in the Block D assemblage includeexamples of both the cut (N=2, 4.0 g) and wire(N=1, 0.6 g) varieties. We also recovered a smallquantity of metal wire (N=2, 4.6 g) (perhaps from afence) and a small section of metal pipe (N=1, 47.7 g).

More interestingly, we recovered several artifacts relating to firearms. These include onepercussion cap (1.0 g) and two pieces (9.4 g) of lead waste or sprue. The latter was likely abyproduct of the manufacture of lead shot, produced either when excess lead was trimmed fromthe edge of a bullet mold or when lead dripped as it was poured into the mold (Blakney-Bailey2008:177). Lead waste is commonly found on historic sites (particularly historic-era Indian sites)from the eighteenth- and early nineteenth-century (Blakney-Bailey 2008:178; Fairbanks 1962:54). Its occurrence in Block D, coupled with the olive bottle glass, suggests the presence of an earlyhistoric settlement on this portion of the site, associated either with Creek Indians or Americans ofAfrican or European descent.

Table 3-4 Historic Artifacts from Unit Levels.

Unit Level

olive bottleglass

amber bottleglass

clear bottleglass

wire nail cut nail metal wire metal pipe percussioncap

lead sprue

N g N g N g N g N g N g N g N g N g

18 1 1 1

18 2 1 1

D1N 1 1 0.6

D8S 3 2 4

D11N 1 1 0.8

D11S 2 1 3.4

D12W 1 1 2.2 2 4.6

D12W 3 1 2.3

D16N 1 1 1.8

D18 2 1 0.5

D20E 1 1 47.7

D22W 1 1 2

D22E 2 1 6

D25N 2 2 0.8

D25S 2 1 0.6

49

50

Chapter 4: Features

Thomas J. Pluckhahn

Excavation of Block D (including Test Unit 18) resulted in the identification of 87 features(Figures 4-1, 4-2, and 4-3), for a density of approximately 1.7 features/m2. Put in other terms, the 87features had a combined area of 19.6 m2; thus, features comprised a considerable 37.7 percent of the52 m2 of excavated area in the block. This is a high density of features by almost any standard. Featuredensity was highest in the core block, the six 1-x-1-m units to the northwest, and Test Unit 18 (seeFigure 4-3). Outlying units in Block D produced only one feature (Feature 140, a post mold or smallpit in Unit D1N—see Figure 4-2).

Another measure of the prevalence of features is volume. The 87 features in Block D had acombined volume of roughly 5449.7 liters. This is equivalent to about 5.5 m3 or 194.2 ft3. As describedin more detail below and in Chapter 7, these figures demonstrate a considerable increase in storageassociated with the late/terminal Late Woodland Block D occupations vis-a-viz the earlier Woodlandcomponent represented in Block A.

This chapter provides a description of the features in Block D. As with the discussion ofexcavation units in the previous chapter, we defer a more detailed description of the artifact and ecofactassemblages from features to Chapters 5 and 6, respectively.

Figure 4-1. Excavation of features in Block D, view to the northwest.

51

0 1 2

meters

¯

post or small pit feature

larger pit feature

see Figure 4-3 for detail on the core block

140

Figure 4-2. Features in Block D.

52

0 1 2

meters

¯post or small pit feature

larger pit feature

Figure 4-3. Closer view of features in the core area of Block D.

142143

141C141A

144

141B

145

195

209

196 200

199

201

204

191B191A

198

197203

208

194

174178

179

206

193 205

192

207202

165147B

147B

171

185170B

170A

175

155

176

182

190

167168

160173

164172 180

177166184148

161

158

156

188

149

154

169

163B

189150

146

181

183153B

157B

157A

153A

159

186151

37

3534A

38

4133

3634B 40

39

53

163A

Our discussion is structured aroundtwo basic types of features, as defined on thebasis of surface area (Figure 4-4). Largerfeatures, probably representing large storagepits or cooking facilities, measure greater than0.21 m2 in area. Smaller features, probablyrepresenting post molds or smaller storagepits, measure less than 0.14 m2. Of course,this rule of thumb can not be applied directlyto features that extend beyond the limits ofour excavation, which were classifiedaccording to their presumed full size based onthe observed dimensions and shape in ourexcavation area.

Smaller Features: Post Molds and Small Pits

Smaller features representingpost molds or small pits (Figure 4-5)were the more common of the two basicfeature types in Block D, with 48examples identified. Table 4-1 presentssummary data for these features.

As noted above, this class offeatures was defined as having surfaceareas less than 0.14 m2. We notedsurface areas in the range of 0.01 to 0.14m2, with a mean of 0.05 m2. For atypical circular post mold, this meanequates to a feature with a diameter ofabout 22 cm.

Profile drawings of post andsmall pit features are provided in Figure4-6. As the profile drawings indicate,most of the smaller features arerelatively shallow, with walls that areeither straight-sided or basin-shaped. Depths ranged from 4 to 50 cm belowthe depth of plan view, with a mean of17.1 cm.

0.00 0.50 1.00 1.50 2.00

Area (square meters)

0

5

10

15

20

25

30

Figure 4-4. Histogram of feature area.

Figure 4-5. Excavation of a post mold/small pit (Feature151), view to the west.

54

Table 4-1. Summary Data for Post Molds and Small Pit Features.Feature Max Depth

(cm)Area(m2)

Volume(liters) Soil Description

34B 15 0.04 5.00 5YR3/2 dark reddish brown sand loam mottled with 2.5YR4/6 red sandy clay36 8 0.06 4.10 5YR3/2 dark reddish brown sand loam with charcoal39 32 0.07 10.30 2.5YR3/2 dusky red sand loam mottled with 2.5YR4/6 red sandy clay40 22 0.03 2.10 2.5YR2.5/2 very dark red sandy clay140 14 0.08 5.70 2.5YR2.5/3 dark reddish brown sand loam144 40 0.03 12.40 10YR4/6 dark yellowish brown sand loam145 30 0.06 13.90 2.5YR2.5/4 dark reddish brown sand loam148 50 0.06 18.90 5YR2.5/2 dark brown sandy clay149 10 0.06 3.50 10YR3/1 dark brown sandy clay150 12 0.07 5.80 2.5YR2.5/3 dark reddish brown sand loam151 16 0.06 6.70 5YR3/2 dark reddish brown sand loam

153B 20 0.03 3.60 5YR3/3 dark reddish brown sand loam160 13 0.04 3.50 2.5YR2.5/3 dark reddish brown sand loam161 12 0.03 1.80 5YR3/4 dark reddish brown sand loam

163B 20 0.03 3.40 5YR2.5/2 dark reddish brown sand loam164 25 0.09 14.40 2.5YR2.5/2 very dark red sand loam166 10 0.03 1.40 2.5YR3/3 dark reddish brown sand loam167 45 0.04 13.40 2.5YR2.5/3 dark reddish brown sand loam168 9 0.04 2.40 2.5YR3/3 dark reddish brown sand loam172 8 0.03 0.20 2.5YR3/3 dark reddish brown sand loam173 4 0.03 0.90 2.5YR3/3 dark reddish brown sand loam177 25 0.03 3.50 5YR3/1 very dark grey sand loam178 15 0.04 3.10 2.5YR3/4 dark reddish brown sand loam179 25 0.04 7.10 2.5YR2.5/4 dark reddish brown sand loam180 20 0.05 7.50 2.5YR3/3 dark reddish brown sand loam181 13 0.08 6.20 2.5YR2.5/1 reddish black sand loam182 7 0.04 2.40 2.5YR3/4 dark reddish brown sand loam183 25 0.05 6.40 2.5YR3/4 dark reddish brown sand loam184 7 0.04 2.70 2.5YR3/4 dark reddish brown sand loam186 35 0.03 6.70 2.5YR3/4 dark reddish brown sand loam187 15 0.03 2.90 2.5YR3/2 dark reddish brown sand loam188 25 0.04 6.00 2.5YR3/4 dark reddish brown sand loam189 8 0.03 1.40 2.5YR3/3 dark reddish brown sand loam190 15 0.05 4.40 2.5YR3/3 dark reddish brown silt loam

191B 25 0.10 14.40 7.5YR2.5/3 and 5YR3/4 dark reddish brown sand loam195 8 0.06 2.70 7.5YR3/4 dark brown sand loam196 5 0.06 2.10 2.5YR2.5/3 dark reddish brown sand loam197 5 0.05 1.30 5YR2.5/2 dark reddish brown sand loam198 15 0.05 4.60 2.5YR2.5/3 dark reddish brown sand loam199 12 0.02 1.40 2.5YR2.5/4 dark reddish brown sand loam200 9 0.08 6.50 2.5YR2.5/3 dark reddish brown sand loam201 8 0.02 1.10 2.5YR2.5/3 dark reddish brown sand loam202 17 0.14 21.20 2.5YR2.5/3 dark reddish brown sand loam204 15 0.07 5.20 2.5YR3/2 dark reddish brown sand loam206 11 0.04 2.90 2.5YR2.5/2 very dark red sand loam207 26 0.08 15.50 2.5YR3/3 dark reddish brown sand loam208 10 0.01 1.30 2.5YR3/4 dark reddish brown sand loam209 5 0.01 0.30 2.5YR2.5/3 dark reddish brown sand loam

55

Figure 4-6. Profiles of postmolds and small pit features.

56

0 5 20

cm

Feature 161

N

Feature 160

N

Feature 151

N

Feature 150

NN

Feature 149

N

Feature 148

N


N

Feature 140

NN

Feature 34B

N

Feature 36

N

Feature 40

N

Feature 39

F38

B (SEcornerXUD4)

Feature 164

N

N

Feature 166 Feature 172

N

Feature 167

N

Feature 168

N

Feature 173

N

Feature 177

N

Feature 178

N

Figure 4-6. Profiles of postmolds and small pit features (continued).

57

0 5 20

cm

Feature 179

N


N N N


N

Feature 184

N

Feature 187

N

Feature 188

N

Feature 189

N

N

Feature 183

Feature 190

N

Feature 191b

N

Feature 195

N N

Feature 196

N

Feature 197

N

Feature 198

N

Feature 199

N

Feature 200

N


N

Feature 204

N

Feature 206

N

Feature 207

N N

Feature 208

N

Feature 209

Possible patterning is apparent in the distribution of small features in the core area of Block D(Figure 4-7). In the initial stages of block excavation, we noted an arc of post features in excavationunits forming the southern half of the core block. The block was expanded north to follow this arcingpattern. Although the pattern is less definitive in the northern end of the core block, the overalldistribution of small features suggests the presence of an oval structure of single set posts measuringabout 7.3 m long and 5.2 m wide. This is consistent with the size of structures identified on severalLate Woodland sites elsewhere in the region (Mickwee 2009; Milanich 1974), as described in greaterdetail in Chapter 7.

Several other lines of evidence support the interpretation of a structure in Block D. First, thereis a regularity to the spacing of smaller features, particularly in the southern portion of the core block. Beginning with Feature 150 along the eastern edge of the block, probable post features are distributedevery 1.0 to 1.7 m in an arc extending first southwest to Features 153B, 151, and 149, then northwestto Features 148/184 (a possible double post) and 160, and then northeast to Features 168 and 182. Asnoted above, the pattern is less apparent in the northern half of the block, but even here an equivalent1 to 1.7 m interval spacing is evident on the northwestern edge of the presumed structure betweenFeature 198 and Features 196 and 200. Additional smaller features between these could representintermediate support posts in the exterior wall. Alternatively, or perhaps in addition, there is somesuggestion of rebuilding in the southern part of the core block, as evidenced by the presence of severalpost holes in the immediate interior of the structural pattern noted above.

In addition to the relatively even horizontal spacing of post features, there is also some regularityto the distribution of posts relative to their depth. We noted above the range of depths in smallfeatures; using the Natural Breaks algorithm in ArcGIS, the smaller features in Block D can be dividedinto two classes based on their depth below plan view: one class with depths between 4 and 20 cm andthe second with depths between 22 and 50 cm. As indicted in Figure 4-8, these two classes do notappear to be randomly distributed across the core block; deeper posts are more common in the interiorof the structure, while all of the posts defining the exterior wall of the presumed structure are shallower. In addition, four of the eight posts in the deeper class (Features 191B, 207, 183, and 164) are distributedat strategic points inside of the presumed structure, where interior support posts might be expected. Deeper holes might be expected for interior support posts, which would have borne more weight ofthe roof.

Tables 4-2, 4-3, and 4-4 document the prehistoric ceramics, flaked stone, and miscellaneousartifacts (respectively) from post and small pit features in Block D. Artifact distribution in smallerfeatures also offers some support for the hypothesized structural pattern, in the form of a discrepancyrelative to the position of features with respect to the structure. The density of artifacts in smallerfeatures was generally light: sherd density ranged from 0 to 8.8 g of pottery/liter with a mean of 1.3,while the density of flaked stone ranged from 0 to 2.9 flaked stone artifacts (n)/liter with a mean of 0.3. In general, however, smaller features within the possible structure exhibited high artifact densities thanthose outside the structure. As an example, Figure 4-9 documents the density of flaked stone in postand small pit features. All of the smaller features with higher artifacts density are located within or onthe edge of the hypothesized structure. The pattern is almost identical for the density of sherds in smallfeatures. The reasons for this pattern are not clear, but it seems reasonable to suggest that it could relateto the accumulation of debris on interior floors, especially around interior posts and along the edges ofthe structure.

58

0 1 2

meters

¯ post or small pit feature

larger pit feature

Figure 4-7. Outline of possible structure in Block D. Postmolds and small pit features mentionedin the text are numbered.

196200

198

206

182

168

160

164

184148 149

150

153B151

59

outline of possible structure

0 1 2

meters

¯

Figure 4-8. Depth of postmolds and small pit features in Block D. Features mentioned in the textare numbered.

191B207

164183

60


depth of smaller features (cm below plan view):

4.0 - 20.0

22.0 - 50.0

larger pit feature

Table 4-2. Prehistoric Ceramics from Post Molds and Small Pit Features.Fe

atur

e

resi

dual

plai

n sa

nd

Swift

Cre

ek

Nap

ier

Car

rabe

lle I

ncis

ed

Car

rabe

lle P

unct

ate

Indi

an P

ass

Inci

sed

Mou

nd F

ield

Net

Mar

ked

Kei

th I

ncis

ed

Wes

t Flo

rida

Cor

d M

arke

d

Tuc

ker

Rid

ge P

inch

ed

WI

Inci

sed

WI

Red

Film

ed

St A

ndre

ws

Wak

ulla

Che

ck S

tam

ped

Wee

den

Isla

nd Z

oned

Inc

/Pun

ct

Wee

den

Isla

nd Z

oned

Red

UID

Inc

ised

UID

Sta

mpe

d

UID

Pun

ctat

e

Wee

den

Isla

nd Z

oned

Red

Oth

er

Tot

al

34B 036 3 1 1 539 2 7 1 1040 2 2 4

140 0144 1 1145 2 2 1 1 6148 0149 2 2150 4 7 1 1 13151 4 2 6

153B 0160 2 2161 0

163B 0164 3 3 1 7166 5 5167 7 1 1 1 10168 0172 0173 0177 0178 0179 11 11180 8 6 14181 1 1182 4 4183 3 1 1 5184 1 1186 1 2 1 4187 1 1188 1 3 4189 0190 2 1 3

191B 1 1 1 3195 0196 0197 0198 0199 0200 1 1201 0202 3 1 4204 1 1 1 3205 1 4 1 6206 0207 9 4 1 1 15208 0209 0

61

Table 4-3. Flaked Stone from Post Molds and Small Pit Features.Feature quartz quartzite Coast Plain chert Total

prim

ary

seco

ndar

y

shat

ter/

chun

k

tert

iary

core

/cor

etoo

l

flake

tool

bifa

ce

prim

ary

seco

ndar

y

shat

ter/

chun

k

tert

iary

core

/cor

e to

ol

flake

tool

bifa

ce

prim

ary

seco

ndar

y

shat

ter/

chun

k

tert

iary

core

/cor

e to

ol

flake

tool

bifa

ce

34B 0

36 1 1

39 1 1

40 1 1 2

140 0

144 1 1

145 1 1

148 0

149 1 1

150 1 1 3 1 11 17

151 1 1 3 5

153B 0

160 1 1 2

161 0

163B 0

164 3 0

166 1 1

167 1 1

168 1 1

172 0

173 0

177 1 1 2

178 0

179 0

180 1 1 1 3 6

181 0

182 0

183 1 1 1 3 6

184 1 1 2

186 3 3

187 0

188 0

189 0

190 1 1

191B 1 1

195 0

196 0

197 0

198 0

199 0

200 1 1

201 1 1 2

202 5 2 8 15

204 1 1

206 0

207 1 1

208 0

209 0

62

Table 4-4. Miscellaneous Artifacts from Post Molds and Small Pit Features.

Feature

mic

a

grou

ndst

one

peck

ed s

tone

limes

tone

FC

R

fired

cla

y

quar

tz p

ebbl

e

sand

ston

e ba

ll

quar

tz c

ryst

al

poss

ible

pig

men

t sto

ne

cera

mic

pip

e bo

wl

cher

t cob

ble

g N g N g N g N g N g N g N g N g N g N g N g

34B 13639 140140 2 321144145 0.1148149150151160161

163B164166167 1 86168172173177178179180181182183184186187188189190

191B192 3 236193194 1 23 1 5195196197198199200 1 16201202 2 17204206207 1 17 1 188208209

63

0 1 2

meters

¯

Figure 4-9. The density of flaked stone in postmolds and small pit features in Block D.

64


larger pit feature

lithic density in small features (n/liter):

0 - 0.6

0.7 - 2.9

Larger Pit Features

Summary data for larger pit features are presented in Table 4-5. Although less numerous thanpost molds at 0.73/m2, larger pits were nevertheless quite common in Block D. More important, thesepits account for a much larger share of the total feature area and volume. The 38 larger pit features hada combined surface area of 17.25 m2, constituting about 33 percent of the total excavated area. Thesefeatures had a combined volume of 5176 liters.

Larger pit features can be divided into two categories based on shape: basin-shaped pits and bell-shaped pits. The latter are defined by profiles wherein at least one side breaks to the exterior (orundercuts), in contrast with basin-shaped features where the feature walls slope more or lesscontinuously to the bottom center of the feature. Figure 4-10 shows the distribution of these two typesof features in Block D. Before considering these two types of features in more detail, it is convenientto discuss artifact distribution patterning in larger features more generally, particularly in reference tothe possible structure described above. Table 4-6 summaries prehistoric ceramics from these features,while Table 4-7 documents flaked stone artifacts. Table 4-8 presents totals for miscellaneous artifactsfrom larger pit features.

The density of flaked stone in larger features ranged from 0 to 4.7 artifacts (n)/liter, with amean of 0.6. Classification into high, medium, and low densities reveals no unambiguous patterningwith respect to the structure, although several features within or immediately adjacent to the structureexhibit higher lithic density (Figure 4-11). The highest lithic density was observed in Feature 170A, abasin shaped pit immediately west of the possible house.

Figure 4-12 displays the density of other types of modified stone—including ground and peckedstone and fire cracked rock. Notably, the highest density is found in Feature 171, a basin-shaped pitnear the center of the presumed structure. This suggests that Feature 171 may have functioned as ahearth or oven, although the higher density here owes much to a single large rock with pitting possiblyindicative of service as an anvil or nutting stone. Moderate densities of these other types of modifiedstone were found in two deep, bell-shaped pits (Features 147B and 175) within the structure, as well asseveral basin-shaped pits (Features 170A, 170B, 37 and 41) elsewhere in the block.

Sherd density in larger pit features in Block D ranged from 0 to 18.8 g/liter, with a mean of 6.1g/liter. While sherd density was high in several of the larger pit features in the core area of the block,six of the 11 large pits with high sherd density are located within the structure (Figure 4-13). Moreover,the four pits nearest the center of the structure (Feature 155, 171, 175, and 176) all exhibit high sherddensity. Three of these are bell-shaped pits, while one (Feature 171) is basin shaped. The higher densityof sherds in these pits could reflect the use and breakage of pots relating to cooking and storage of foodnear the center of the structure, although deliberate disposal of broken pottery in pits formerly used forother purposes (such as storage) is also possible.

The preservation of bone was generally poor in Block D, as elsewhere at Kolomoki. Nevertheless, some features produced more bone than others, suggesting the possibility that theyfunctioned for different purposes. In general way, the density of bone in features mirrors that of sherds,with concentration in features near the center of the presumed structure (Figure 4-14). Specifically, thethree bell-shaped pits designated as Features 155, 175, and 176 exhibited high densities of bone. Feature171, the basin shaped pit that it was suggested above to have possibly functioned as a hearth, exhibitedmoderate density of bone. Most of the other pit features with moderate or high density of bone arelocated outside the presumed structure. This suggests that the processing and cooking of animal foodstook place both within and outside the house.

The remainder of ths section is devoted to a discussion of the two basic types of larger pitfeatures in turn. We begin with an overview of the basin-shaped pit features.

65

Table 4-5. Summary Data for Larger Pit Features.Feature Max Depth

(cm)Area(m2)

Volume(liters) Soil Description

Large, Basin-Shaped Pit Features33 15 0.07 2.60 2.5YR3/3 dark reddish brown sand loam

34A 40 0.40 96.90 5YR3/2 dark reddish brown sand loam mottled with 2.5YR4/6 red sandy clay35 37 0.25 45.30 5YR3/3 dark reddish brown sand loam37 37 0.24 37.50 5YR3/2 dark reddish brown sand loam38 37 0.24 65.20 2.5YR3/2 dusky red sand loam mottled with 2.5YR4/6 red sandy clay41 38 0.56 109.31 5YR3/3 dark reddish brown sand loam

141A 40 1.78 424.90 5YR3/2 dark reddish brown sand loam141C 60 0.64 207.60 5YR3/4 dark reddish brown sand loam142 8 0.06 31.00 5YR3/3 dark reddish brown sand loam143 10 0.03 26.90 5YR3/3 dark reddish brown sand loam146 40 0.30 81.40 2.5YR2.5/3 dark reddish brown sand loam

147A 100 0.56 270.50 5YR3/2 dark reddish brown sand loam153A 30 0.38 81.70 5YR3/3 dark reddish brown sand loam154 40 0.38 85.90 5YR3/2 dark reddish brown sand loam156 100 0.52 113.50 5YR3/2 dark reddish brown sand loam158 30 0.23 41.80 5YR3/2 dark reddish brown sand loam159 50 0.44 111.30 5YR2.5/2 dark brown sand loam169 30 0.40 62.40 5YR2.5/2 dark brown sand loam

170A 20 0.36 46.50 2.5YR2.5/3 dark reddish brown sand loam170B 20 0.29 46.50 2.5YR2.5/3 dark reddish brown sand loam171 50 0.91 277.60 5YR3/4 dark reddish brown sand loam

191A 50 0.84 237.60 7.5YR2.5/3 and 5YR3/4 dark reddish brown sand loam192 35 0.72 148.10 2.5YR3/3 dark reddish brown sand loam193 35 0.08 16.60 2.5YR3/6 dark red clay loam194 32 0.21 48.10 2.5YR3/3 dark reddish brown sand loam203 50 0.41 116.80 5YR2.5/2 dark reddish brown sand loam205 100 0.11 14.20 2.5YR3/3 dark reddish brown sand loam

Large, Bell-Shaped Pit Features

141B 100 0.78 385.80 Zone A: 10YR2/2 very dark brown sand loam with charcoalZone B: 10YR3/2 very dark grey brown with 10YR3/6 dark red sand loam

147B 90 0.74 395.10 5YR3/2 dark reddish brown sand loam155 100 0.79 244.90 5YR3/2 dark reddish brown sand loam

157A 100 0.33 34.40 2.5YR2.5/2 very dark red sand loam157B 80 0.06 126.60 2.5YR2.5/2 very dark red sand loam163A 100 0.22 73.60 5YR2.5/2 dark brown sand loam165 40 0.26 83.20 5YR3/3 dark reddish brown sand loam174 50 0.38 93.90 2.5YR3/4 dark reddish brown sand loam175 100 0.69 229.90 2.5YR2.5/3 dark reddish brown sand loam176 40 0.23 52.00 2.5YR2.5/3 dark reddish brown sand loam185 75 1.36 608.40 Zone A: 2.5YR2.5/3 dark reddish brown sand loam

Zone B: 2.5YR2.5/2 very dusky red sand loamZone C: 2.5YR3/4 dark reddish brown sand loamZone D: 2.5YR3/3 dark reddish brown sand loam

66

0 1 2

meters

¯

Figure 4-10. Larger bell- and basin-shaped pit features in Block D.

142143

141C141A

141B

191A203

194

174

193 205

192

165147B

147B

171

185170B

170A

175

155

176

182

158

156

154

169

146

157B

157A

153A

159

37

3534A

38

4133

67

basin-shaped pit

bell-shaped pit

Table 4-6. Prehistoric Ceramics from Larger Pit Features.Fe

atur

e

resi

dual

plai

n sa

nd

Swift

Cre

ek

Nap

ier

Car

rabe

lle I

ncis

ed

Car

rabe

lle P

unct

ate

Indi

an P

ass

Inci

sed

Mou

nd F

ield

Net

Mar

ked

Kei

th I

ncis

ed

Wes

t Flo

rida

Cor

d M

arke

d

Tuc

ker

Rid

ge P

inch

ed

WI

Inci

sed

WI

Red

Film

ed

St A

ndre

ws

Wak

ulla

Che

ck S

tam

ped

Wee

den

Isla

nd Z

oned

Inc

/Pun

ct

Wee

den

Isla

nd Z

oned

Red

UID

Inc

ised

UID

Sta

mpe

d

UID

Pun

ctat

e

Wee

den

Isla

nd Z

oned

Red

Oth

er

Tot

al

Large, Basin-Shaped Pit Features

33 2 5 7

34A 45 59 39 1 1 1 2 2 150

35 27 18 12 2 59

37 50 52 6 1 3 3 1 116

38 42 31 4 1 4 82

41 40 45 20 105

141A 345 304 38 2 20 4 2 5 14 1 1 13 2 1 752

141C 126 69 12 3 1 1 6 3 221

142 0

143 5 5

146 25 38 1 64

147A 63 58 4 2 2 4 1 1 135

153A 36 37 6 2 3 3 2 89

154 91 67 8 1 4 1 6 3 2 183

156 42 53 2 2 1 5 2 1 108

158 6 2 8

159 34 49 14 1 1 4 1 1 1 106

169 46 59 1 5 10 1 5 2 1 130

170A 38 47 3 5 3 1 97

170B 30 20 2 2 2 1 57

171 261 206 62 1 3 27 1 3 1 565

191A 75 69 6 3 3 3 8 1 1 1 2 172

192 48 40 1 3 1 1 2 1 97

193 1 2 3

194 29 27 1 1 3 61

203 63 36 6 4 2 1 2 1 5 3 123

205 1 4 1 6


141B 437 183 28 1 2 13 7 13 23 2 9 2 720

147B 342 355 54 2 2 33 1 23 4 4 14 1 1 836

155 208 216 15 3 2 6 7 3 1 5 36 1 1 3 1 508

157A 17 14 4 35

157B 36 13 3 1 53

163A 41 36 2 1 4 1 85

165 62 57 2 1 1 2 1 2 7 135

174 18 13 1 1 1 1 35

175 273 169 41 16 1 44 544

176 71 52 4 1 3 2 1 1 1 1 1 137

185 455 218 36 2 13 3 10 1 2 10 750

68

Table 4-7. Flaked Stone from Larger Pit Features.Feature quartz quartzite Ridge/

Valleychert

Coast Plain chert Total

prim

ary

seco

ndar

y

shat

ter/

chun

k

tert

iary

core

/cor

etoo

l

flake

tool

bifa

ce

prim

ary

seco

ndar

y

shat

ter/

chun

k

tert

iary

core

/cor

e to

ol

flake

tool

bifa

ce

bifa

ce

prim

ary

seco

ndar

y

shat

ter/

chun

k

tert

iary

core

/cor

e to

ol

flake

tool

bifa

ce


33 4 1 5

34A 1 2 1 6 1 14 1 1 27

35 1 2 1 2 1 12 1 20

37 3 2 2 1 2 2 3 23 38

38 1 2 6 14 2 25

41 1 1 2 9 1 14

141A 2 7 10 8 1 5 1 8 31 22 167 2 4 268

141C 3 4 3 3 14 11 104 1 143

142 0

143 1 1

146 1 1 2 1 1 4 6 8 14 1 39

147A 1 3 3 3 1 4 11 13 47 1 87

153A 1 1 1 4 3 18 1 29

154 4 2 4 6 8 53 1 1 79

156 2 1 8 5 27 1 44

158 1 1 2 4

159 1 2 1 4 3 3 7 21

169 2 2 1 17 2 63 87

170A 1 1 1 16 48 22 129 218

170B 1 2 10 5 21 1 40

171 1 4 5 3 1 1 1 2 5 22 20 66 2 133

186 3 3

187 0

188 0

189 0

190 1 1

191A 1 5 4 3 5 7 11 12 29 12 105 2 196

191B 1 1

192 2 2 1 1 3 3 4 15 31

193 1 1 2

194 1 5 1 5 1 2 14

203 1 1 2 1 2 6 2 19 1 35

205 0


141B 1 1 5 2 2 1 1 5 23 28 85 2 1 1 157

147B 2 6 7 2 1 1 3 13 25 25 122 4 211

155 3 5 6 1 1 8 18 17 99 2 160

157A 1 6 7

157B 1 1 3 1 4 10

163A 1 1 6 11 9 33 1 1 63

165 1 1 1 2 1 1 1 4 21 12 74 1 120

174 1 1 2 6 8 18

175 1 1 1 4 1 1 6 15 8 34 2 74

176 1 1 1 1 2 9 8 31 54

185 1 3 14 6 1 2 2 2 6 23 15 83 1 5 164

69

Table 4-8. Miscellaneous Artifacts from Larger Pit Features.

Feature

mic

a

grou

ndst

one

peck

ed s

tone

limes

tone

FC

R

fired

cla

y

quar

tz p

ebbl

e

sand

ston

e ba

ll

quar

tz c

ryst

al

poss

ible

pig

men

t sto

ne

cera

mic

pip

e bo

wl

cher

t cob

ble

g N g N g N g N g N g N g N g N g N g N g N g


33

34B 1

35 1 1 2

37 3 148

38 1 11 1 17

41 1 1 240 1 359 4 394 5 18

141A 2 46 15 447 354 2 124

141C 0.1 2 12 1 37

142 4 497

143

146 7 102

147A 6 513

153A 2 37

154 0.1 2 56 1 23

156 3 18

158 0.1

159 1.4 4 65

169 4 16

170A 0.1 5 338 1 8

170B 3 202 1 5

171 0.2 3 4750 1 2 4 65 1 14

186

187

188

190

191A 2 360 4 192

191B

192 3 236

193

194 1 23 1 5

203 1 36 1 4

205


141B 10 143 1 1 1 14

147B 0.5 1 20 2 13 8 2167 6 100

155 1 57 1 5 1 66

157A 0.1

157B 0.2

163A 2 11 1 32

165 3 46

174

175 0.1 1 212 1 113 10 744 2 16

176 0.1

185 0.1 2 511 7 1254 1 2 3 64

70

0 1 2

meters

¯

Figure 4-11. The density of flaked stone in the larger pit features in the core area of Block D.

142143

141C141A

141B

191A203

194

174

193 205

192

165 147B147B

171

185170B

170A

175

155

176

158

156

154

169

146

157B

157A

153A

159

37

3534A

38

4133

71

density of flaked stone inlarger pit features (n/liter):


163A

0 - 0.7

0.8 - 1.9

4.7

0 1 2

meters

¯

Figure 4-12. The density of other modified stone (ground and pecked stone and FCR) in the largerpit features in the core area of Block D.

142143

141C141A

141B

191A203

194

174

193 205

192

165 147B147B

171

185170B

170A

175

155

176

158

156

154

169

146

157B

157A

153A

159

37

3534A

38

4133

72

density of other modified stonein larger pit features (g/liter):


163A

0 - 2.9

4.0 - 9.1

17.1

0 1 2

meters

¯

Figure 4-13. The density of prehistoric ceramics in the larger pit features in the core area of Block D.

142143

141C141A

141B

191A203

194

174

193 205

192

165147B

147B

171

185170B

170A

175

155

176

158

156

154

169

146

157B

157A

153A

159

37

3534A

38

4133

73

3535

3535

sherd density in largerpit features (g/liter):


0 - 2.33.8 - 8.49.2 - 18.8

0 1 2

meters

¯

Figure 4-14. The density of bone in the larger pit features in the core area of Block D.

142143

141C141A

141B

191A203

194

174

193 205

192

165147B

147B

171

185170B

170A

175

155

176

158

156

154

169

146

157B

157A

153A

159

37

3534A

38

4133

74

3535

3535

density of bone in largerpit features (g/liter):


0 - 0.20.3 - 0.71.0 - 1.8

Large, Basin-Shaped Pits

Most of the larger pit features are basin shaped in cross section; we identified 28 of these inBlock D. As indicated in Figure 4-10, basin-shaped pits are widely distributed across Block D, but alsoappear to be non-randomly positioned with respect to the presumed structure: only two of the 28 basin-shaped pits lie wholly within the structure. Portions of another seven lie partially within the structure,but three of these would have been positioned outside the structure as well prior to the hypothesizedexpansion of the house to the south (as described above).

Feature 192, which was locatedjust to the northeast of the structure, istypical of the small and medium sizedfeatures that make up the majority ofthe basin-shaped pits in Block D. Asindicated in the profile photographreproduced as Figure 4-15, the darkbrown fill of the feature was easilydistinguished from the red sandy claysubsoil. The pit had a clearly defined,circular shape in plan, and measuredslightly less than 1 m in diameter. Excavation revealed that the walls of thepit sloped gently in to the base of the pitat about 35 cm below plan view depth. Although it is possible that a portion ofthe pit was truncated by the plow zone,the feature must have been relativelyshallow. Artifacts were relativelyuncommon in the fill of the pit. However, microbotanical analysis of asoil sample from the feature producedidentified starch grains diagnostic ofmaize, as described in Chapter 6. Theabsence of any maize pollen orphytoliths suggests that prepared foodscontaining maize may have been storedor discarded in the pit.

Profiles of basin-shaped pits are provided in Figure 4-16. Small and medium basin-shaped pitssuch as Feature 192 likely served mainly as processing facilities—that is, as basins for cooking, plantprocessing, and similar activities. Although dense concentrations of charcoal were generally lacking,scattered flecks and larger chunks of burned wood were common in the fill of these basin-shaped pits,consistent with their use as cooking facilities. Some of the smaller basin-shaped pits may have alsofunctioned for storage, although their capacity would have been limited for storage of bulkier items, andtheir wide opening would have made them more susceptible to flooding, inadvertent damage fromtrampling, and scavenging by both people and other animals.

Table 4-5 presents summary data for the basin-shaped pits that were identified in Block D. Profiles of basin-shaped pits are provided in Figure 4-16. In terms of surface area at plan view, thesepits ranged from 0.03 to 1.8 m2, with a mean of 0.4 m2. Depths below plan view ranged from just 8 cmto 100 cm or more. Volume varied considerably as a function of both surface area and depth, with arange of 2.6 to 424.9 liters. The mean volume was 106.2 liters.

Figure 4-15. The East Profile of Feature 192, a typical smallbasin shaped pit.

75

0 5 20

cm

Feature 33

N

2.5YR3/3 dark reddishbrown sand loam

Feature 35

N

5YR3/3 dark reddishbrown sand loam

Feature 34A

N

5YR3/2 dark reddish brownsand loam mottled with2.5YR4/6 red sandy clay

Feature 37

N

5YR3/2 dark reddish brown sand loam

Feature 38

N

2.5YR3/2 dusky red sand loammottled with 2.5YR4/6 red sandy clay

F39

Feature 41

N


Feature 141A

N

5YR3/2 dark reddishbrown clay loam


2.5YR3/4 dark reddishbrown clay loam

Feature 141C

NW

Feature 142

NW corner of XUD3

N

S wallTU18

Figure 4-16. Profiles of large, basin-shaped pit features.

76


Feature 154

Figure 4-16. Profiles of large, basin shaped pit features (continued).

77

Feature 143

N

Feature 146

N

5YR3/2 dark reddish brown sand loam

N

Feature 147A

0 5 20

cm

Feature 153a

N

S wall of XUD6

N

Feature 156

N

W wallof XUD8

Feature 158

N


2.5YR3/3 dark reddish brown sand loam



Feature 159

N

Feature 169

N5YR2.5/2 dark reddish brownsand loam with charcoal

2.5YR4/6 red clay loam


2.5YR2.5/3 dark reddishbrown sand loam




Figure 4-16. Profiles of large, basin-shaped pit features (continued).

78


F170A

F170B

2.5YR3/6 dark redsand loam

Features 170A and 170B

N


largerock

Feature 171

NE

7.5YR2.5/3 very dark brownclay loam with charcoal


Feature 191A

0 5 20

cm

Nedge ofblock


Feature 192

N

Feature 193

N

2.5YR3/6 dark redsandy clay 2.5YR3/3 dark reddish

brown sand loam

Feature 194

N

Feature 205

N


5YR2.5/2 dark reddishbrown sand loam

Feature 203

N

Several of the larger basin-shaped pitsmerit more detailed description. Feature191A, located just inside the inferredstructure, appeared in plan view as a circularstain about one meter in diameter (Figure 4-17). Excavation revealed that the walls of thepit sloped relatively steeply to the base of thefeature at a depth of around 50 cm. Arelatively large post feature (Feature 191B)adjoined the pit on its northeastern edge.

The fill of Feature 191A wasrelatively rich in small fragments of charcoal. A soil sample from the feature was submittedfor pollen and phytolith analysis. Asdiscussed in greater detail in Chapter 6, ascan of the sample revealed the presence ofmaize pollen, as well as phytoliths tentativelyidentified as a variety of arrowroot. Feature191A also produced a diverse assortment ofLate Woodland pottery, including examplesof the Napier, Carrabelle Incised, CarrabellePunctate, Mound Field Net Marked, KeithIncised, Tucker Ridge Pinched, WeedenIsland Incised, and Weeden Island RedFilmed types, among others. A sample ofhickory nutshell from Feature 191Aproduced a two sigma calibrated date of A.D.890 to 1020, as described in more detail inChapter 7.

As noted above, Feature 171 is intriguing for several reasons. First, it is located near the centerof the presumed house. Next, it contained a high density of modified stone other than flaked stone; thatis, fire cracked rock, ground stone, and pecked stone. Near the base of the feature, observable in theprofile in Figure 4-18a, was a large rock. Pitting on the surface of the rock suggests it may have servedas an anvil or nutting stone. Alternatively, assuming the feature functioned as a fire pit, the rock mayhave functioned as a pot support or andiron.

We observed a cluster of pottery in the northeastern half of Feature 171 (Figure 4-18b). Muchof the pottery was from a very friable, thin-walled, Swift Creek Complicated Stamped vessel. The potwas fired poorly—probably not far from Block D—and discarded in the pit feature. Fragments of atleast two other Swift Creek vessels were also present in the feature (see discussion of the MNV analysisin Chapter 5). We also noted fragments of various other pottery types of the Weeden Island series (seeTable 4-6). In the northeastern half of the pit—opposite the pottery cluster—was a small cluster ofdeer bone (Figure 4-18c). Botanical analysis of a flotation sample from the feature revealed seeds ofbean and goosefoot (see Chapter 6). A sample of hickory nutshell from Feature 171 produced a two-sigma calibrated date of A.D. 780 to 990, as described in more detail in Chapter 7.

Several of the largest basin-shaped pits in Block D were partially intruded by other pits. Features141A and 141C were overlapping basin-shaped pits. The former appears to have been intruded byFeature 141B, a bell-shaped pit. These three features together formed a complex series of stains thatwere difficult to separate in plan view (Figure 4-19a), but which became better defined as the featurefill was removed, exposing the red sandy clay subsoil (Figure 4-19b and c). All three features were richin artifacts, suggesting they became useful places to dispose of trash after the pits outlived

Figure 4-17. Excavation of Feature 191A, view to thesouth.

79

Figure 4-18. Three views of Feature 171, a large basin-shaped pit. A) Southeast Profile; B) clusterof pottery; C) cluster of deer bone.

80

A

B C

Figure 4-19. Three views of Feature 141A, B, and C. A) plan view to south before excavation; B) view to west after excavation (Feature 141C,a basin-shaped pit, is visible in profile); C) view to east after excavation (Feature 141B, a bell-shaped pit, is visible in profile).

81

A

B

C

Figure 4-20. Four views of Feature 147A and 147B. A) plan view to west after excavation; B) west profile of Feature 147B before block wasexpanded west to expose remainder of feature; C) excavation of Feature 147B (the excavator is standing in the bell shaped half); d) east profileof Feature 147A.

82

AC

D

B

their usefulness for storage or, perhaps less likely, as fire or roasting pits. Feature 141A, in particular,produced a number of large ceramic vessel fragments, including examples identified as Weeden IslandRed and Zoned Red (see Chapter 5). Features 141A and 141B also produced several diagnostic haftedbifaces (also documented in greater detail in Chapter 5).

Feature 147B, on the eastern margin of the possible structure, includes both a large, basin-shaped pit and—below this—the lower half of a bell-shaped pit (both pits were subsumed under thesingle feature number) (Figure 4-20a). In this case, it is apparent from the profile (Figure 4-20b) (takenbefore the unit was expended) that whichever pit intruded the other, the basin-shaped pit filled last, asevidenced by the dark lenses of fill lying on top of the bell-shaped pit below. Feature 147B producedan impressive artifact collection. The pottery assemblage includes over 800 sherds, with a variety of LateWoodland diagnostic types represented. We also recovered three diagnostic hafted bifaces, includingtwo Late Woodland/Mississippian triangulars and a Tampa point. A sample of hickory nutshell fromthe feature produced a two sigma calibrated date of A.D. 780 to 980, as described in more detail inChapter 7.

The bell-shaped portion of Feature 147B, which extended about a meter below plan view(Figure 4-20c), also adjoined Feature 147A, another basin-shaped pit to the east. This was a relativelyshallow basin, (Figure 4-20d), suggesting uses other than storage. Compared with Feature 147B, thedensity of artifacts was lighter in this basin-shaped pit. Nevertheless, Feature 147A produced severaltemporally diagnostic artifacts, including a Swan Lake biface and Weeden Island Incised and WakullaCheck Stamped pottery.

Bell-Shaped Pits

Bell shaped pit features, summarized inTable 4-3 and profiled in Figure 4-22, were lesscommon than basin shaped pits in Block D. What these pits lack in number, however, theymake up for in size. The 11 bell shaped pits hada combined volume of 2201.2 liters. This worksout to an average of 200.1 liters, nearly double themean volume of basin shaped pits. Bell shapedpits were also larger in surface area on average,with a mean of 0.53 m2. This compares with amean of 0.4 m2 for basin shaped pits.

As discussed above, bell shaped pitsappear to be non-randomly distributed in BlockD. Seven of the 11 bell shaped pits are locatedwithin the hypothesized structure. Moreover, thebell shaped pits within the structure form a ringroughly midway between the inferred hearth(Feature 171) near the center of the structure andthe exterior walls. The remaining bell shaped pitscluster immediately to the southeast of thestructure.

Mention has already been made ofFeature 141B and 147B, two of the large, bell shaped pit features in Block D. Features 165 and 174,both located in the northern interior of the structure, represent the smaller end of the range of sizes ofbell shaped pits. The former feature was particularly well defined (Figure 4-21). At plan view, itappeared as a nearly perfect, 60-cm-diameter, circle of darker brown soil. The walls of the pit enlargedwith depth before contracting to meet at the bottom of the pit around 40 cm below plan view. The

Figure 4-21. Feature 165, view to the north.

83

Figure 4-22. Profiles of large, bell-shaped pit features.

10YR3/2 very dark greyish brownmottled with 10YR3/6 dark red loam

Feature 141b

N

Feature 147B

N


5YR3/2 dark reddish brown sand loammottled with 2.5YR4/8 dark red sandy clay

5YR3/2 dark reddish brown sand loam with charcoal flecking

edge ofunit

Feature 155

N


2.5YR2.5/4 darkreddish brown

clay loam

Feature 157

N

2.5YR2.5/2 very dark redsand loam

F157B

F157A

edge ofunit

Feature 163A

N

5YR2.5/2 dark brownsand loam

5YR3/3 dark reddish brownsand loam

Feature 165

0 5 20

cm

N

Feature 174

N

2.5YR3/4 darkreddish brown

sand loam

84

5YR3/3 dark reddish brown mottledwith 2.5YR3/6 dark red loam

Figure 4-22. Profiles of large, bell-shaped pit features (continued).

85

2.5YR2.5/3 clay loam

2.5YR2.5/4 dark reddishbrown clay loam

Feature 175

N

Feature 176

N

2.5YR2.5/3 dark reddishbrown sand loam

2.5YR2.5/3 dark reddish brown clay loam

2.5YR2.5/2 very dusky red clay loam

2.5YR3/3 dark reddish brown clay loam

2.5YR3/4 clay loam

2.5YR3/6 dark red clay loam


Feature 185

0 5 20

cm

N

small size and clean definition of this pit, as well as Feature 174, suggest that it was filled in the earlystages of its potential use life. In some cases, smaller bell-shaped pits like Features 165 and 174 mayhave been gradually expanded---either deliberately or unintentionally through continuous removal ofcontents or cleaning—to form larger bell-shaped pits. There is some evidence for this in the moreirregular shape of the larger bell-shaped pits in Block D.

Although artifact density was generally lower in the smaller bell-shaped pits, several of thesefeatures produced interesting artifact assemblages. Feature 165, for example, yielded several fragmentsof Weeden Island Zoned Incised/Punctate pottery and a spike PP/K. Feature 174 contained examplesof Indian Pass Incised and Napier Complicated Stamped pottery.

Feature 185 is the largest feature—bell-shaped or otherwise—in Block D, with an excavatedvolume of over 600 liters. This represents only about two-thirds of the actual volume of the pit,however, since an estimated one-third of the feature extends beyond the limits of the block and was notexcavated. The irregular shape of the feature both in plan and profile suggests that it may have beenexpanded several times. In contrast with many of the other features in Block D, Feature 185 exhibiteda stratified fill that included several distinct layers, some with relatively high densities of charcoal. Thepresence of this layered fill suggests the feature filled gradually, either during the course of occupationof the house in Block D or another, later structure somewhere else in the vicinity of Block D.

The excavated fill from Feature 185 produced 750 sherds. The pottery collection includes avariety of Late Woodland types, such as Napier Complicated Stamped, Wakulla Check Stamped,Weeden Island Incised, and Weeden Island Zoned Incised/Punctate. We also recovered five chertbifaces or biface fragments from the pit, although unfortunately none of these is temporally diagnostic.

Figure 4-23. The North Profile of Feature 185.

86

Feature 155 was another sizable bell-shaped pit in Block D. Located near the center of thestructure, it extended beyond the limits of our excavation. The portion of Feature 155 that weexcavated had a volume of approximately 245 liters. A smaller bell shaped pit (Feature 176) adjoinedFeature 155 on the east side.

As with the previously described Feature 185, the profile of Feature 155 also displays a numberof distinct zones (Figure 4-24), indicative of a gradual accumulation of refuse after the feature hadoutlived its uselife, probably as a storage facility. This refuse included a relatively high density of faunalremains and ceramics, as noted above and as visible in profile (see Figure 4-24). The faunal assemblagefrom this pit is notable for the presence of bear.

Over 500 sherds were recovered from the excavated portion of Feature 155. The assemblageis diverse, including at least of example of virtually every Late Woodland diagnostic type representedin the Block D collection as a whole. Included in the assemblage are several large vessel fragments,including examples identified as Keith Incised and Napier Complicated Stamped, as described in moredetail in the discussion of the MNV analysis in the chapter that follows. Feature 176, the adjoining bell-shaped pit, produced a fragment of a Weeden Island Zoned Red plate, a ceramic type which is relativelyrare at Kolomoki and which would presumably been used for serving foods in rituals or other specialoccasions.

Figure 4-24. The North Profile of Feature 155.

87

Also meriting more detailed description isFeature 175 (Figure 4-25), another bell-shaped pit withinthe presumed structure. This pit was somewhat irregularin both plan and profile, suggesting it was enlarged orexpanded. The western portion of the pit was relativelyshallow and basin shaped. In the eastern portion, the pitwas deep and bell shaped. Less dark (although stilldistinct) and more homogenous fill in this pit suggeststhat it filled relatively rapidly.

Artifacts were plentiful in Feature 175, with highdensities of FCR, bone, and ceramics. The potterycollection includes more than 500 sherds, includingexamples of the Carrabelle Punctate and Mound FieldNet Marked types. Notably absent from the assemblagefrom this feature are any of the incised Weeden Islandtypes or Napier Complicated Stamped. This suggeststhe possibility that Feature 175 dates a little earlier thanseveral of the other large bell-shaped pits in Block D,and perhaps earlier than the structure.

I would also note Feature 163A (Figure 4-26), abell-shaped pit that extended into the southern wall ofthe core block and was thus located just outside thepresumed structure. As described in Chapter 6, aflotation sample from this pit produced macrobotanical evidence for Ilex. One variety of this genus,youpon holly (Ilex vomitoria), was widely used to make black drink in the later historic era Southeast.

Summary

Excavation of Block D (including Test Unit 18)resulted in the identification of 87 features: 48 postmolds or small pits and 39 larger pits. Feature patterningpossibly indicative of a domestic structure was notedduring the course of excavating Block D. The pattern isdefined mainly by the distribution of post mold or smallpit features in an oval pattern measuring about 7.3 mlong and 5.2 m wide. As described in more detail inChapter 7, this is almost precisely the same length asLate Woodland structures identified at the Sycamore(Milanich 1974) and Woodland Terrace (Mickwee 2009)sites to the south of Kolomoki. The pattern isinterpreted as the remains of a house of single set postconstruction.

Several lines of evidence support thisinterpretation. First, there is a regularity to the spacingof exterior posts, especially along the southern half ofthe structure. There is also regularity to the spacing ofinterior support posts. Moreover these interior posts aredeeper than those on the exterior, as might be expectedif they supported the weight of the roof.

Figure 4-25. The East Profile of Feature 175.

Figure 4-26. The South Profile of Feature163.

88

Pit features appear to be non-randomly positioned with respect to the structure. Specifically,larger pit features—especially large, bell shaped pits—are more common in the interior of the structure. Further, pit features in the interior generally display higher densities of artifacts than those outside thestructure. At the center of the oval pattern and presumed house, Feature 171 consists of a basin-shapedpit with relatively high density of FCR, ground, and pecked stone. It may have served as a fire pit. Thisfeature is ringed by several large bell shaped pits that likely functioned for storage.

Carbon dates from Block D, reported in greater detail in Chapter, place the occupation of thehouse in the late Late Woodland period, from around cal A.D. 750 to 850. The number and variety offeatures in Block D stands in marked contrast to the remains of the earlier Late Woodland householdpreviously excavated in Block A. I highlight this contrast in more detail in Chapter 7. First, however,we present an overview of the artifact and ecofact assemblages from Block D in Chapters 5 and 6,respectively.

89

90

Chapter 5: Artifacts

Thomas J. Pluckhahn

Artifact density was high in the excavation units levels and features in Block D. This chapterpresents a summary of the artifact assemblage, with particular attention to temporally diagnostic lithicsand ceramics. Faunal and botanical remains, which are less extensive, are described in the chapter thatfollows.

Prehistoric Ceramics

Prehistoric ceramics are the most common artifact type represented in the assemblage fromBlock D. Including the previously excavated Test Unit 18, the pottery collection includes 21,637 sherdsweighing a collective 69,458.6 g (Table 5-1). Around 64 percent (by count) of the sherds were recoveredfrom excavation unit levels, while about 35 percent were found in feature fill. The remaining 1 percentwere retrieved from miscellaneous proveniences such as grab collections and profile cleaning.

All of the pottery in the assemblage is associated with the Woodland period occupation of theBlock D area, with the exception of 17 sherds (representing less than 0.1 percent of the total collectionby count) that appear to correspond more closely with Mississippian types. These 17 sherds include 12Lamar Plain and 4 Lamar Complicated Stamped, distinguished primarily on the basis of abundant andrelatively large quartz grit temper. Sherds of the latter type were additionally distinguished bycomplicated stamping that differs from the more predominant Swift Creek stamping in execution andmotif. All of the Lamar sherds were recovered from excavation unit levels rather than features, thussupporting the interpretation that they were deposited after the more intensive occupation of the areain the Late Woodland.

The final possible Missississippian sherd consists of a relatively small sherd from Feature 147Bthat we have tentatively identified as an example of Fort Walton Incised. This was distinguished fromWoodland incised types by the quality and spacing of the incising. Fort Walton developed from lateWeeden Island precedents around A.D. 900-1000 and persisted until the 1500s (Milanich 1994:358,2002). The Fort Walton Incised sherd could thus be associated with either the Late Woodland WeedenIsland occupation of Block D or the much later Lamar period (Mississippian) settlement centeredfurther downslope near Little Kolomoki Creek, as documented by Sears (Pluckhahn 2003: 74-77; Sears1951a).

Leaving aside these Mississippian sherds, the Block D Woodland pottery assemblage includes21,616 sherds. However, the residual category—consisting of sherds that are too small or eroded (orboth) to securely classify—makes up more than half this total. Leaving aside these residual sherds aswell, the total identifiable Woodland pottery assemblage from Block D consists of 9272 sherds. Table5-2 documents the breakdown of Woodland pottery categories and types in the Block D assemblage. The relative frequencies are generally in keeping with Pluckhahn’s (2003:Table 2.2) definition of theKolomoki IV phase. Plain wares—primarily sand tempered plain—predominate, making up a littlemore than three-quarters of the assemblage. Swift Creek Complicated Stamped is still the mostcommon named decorative type, but here makes up only about 10 percent of the collection, comparedwith as much as 25 percent in earlier phases at Kolomoki. Less common complicated stamped typesinclude Napier and St. Andrews. Both of these are good indicators of the Late Woodland period, asdescribed in more detail below.

91

Table 5-1. Prehistoric Ceramics from Block D.Type Unit Levels Miscellaneous Features Total

N g N g N g N g % by count

residual 8,620 9,135.5 119 163.6 3,609 3,849.7 12,348 13,148.8 57.07

plain sand tempered 4,305 20,096.2 76 604.1 2,773 21,611.4 7,154 42,311.7 33.06

plain limestone tempered 1 12.4 1 12.4 <0.01

St. Johns Plain 1 2.3 1 2.3 <0.01

Swift Creek Comp Stamped 434 1,802.5 10 50.9 448 3,886.4 892 5,739.8 4.12

Napier Comp Stamped 6 30.0 28 629.5 34 659.5 0.16

Carrabelle Incised 7 32.7 13 201.6 20 234.3 0.09

Carrabelle Punctate 112 559.2 2 7.5 152 1,109.8 266 1,676.5 1.23

Indian Pass Incised 7 34.8 16 155.8 23 190.6 0.11

Mound Field Net Marked 11 63.1 1 3.5 24 326.2 36 392.8 0.17

Keith Incised 15 61.8 10 97.7 25 159.5 0.12

West Florida Cord Marked 3 19.1 1 3.3 4 22.4 0.02

Tucker Ridge Pinched 4 27.6 4 16.5 8 44.1 0.04

Weeden Island Incised 19 115.4 33 290.1 52 405.5 0.24

Weeden Island Red Filmed 200 972.09 1 0.7 254 1,861.9 455 2,834.7 2.10

St. Andrews Comp Stamped 3 88.0 3 88.0 0.01

Wakulla Check Stamped 4 32.4 9 49.8 13 82.2 0.06

Weeden Island ZonedIncised/Punctate

10 70.8 1 67.5 16 241.2 27 379.5 0.12

Weeden Island Zoned Red 1 30.1 1 30.1 <0.01

unidentified incised 112 305.4 5 21.8 77 342.8 194 670 0.90

unidentified stamped 10 39.8 1 1.6 5 71.4 16 112.8 0.07

unidentified punctate 35 119.7 12 26.4 47 146.1 0.22

Lamar Plain 12 78.2 12 78.2 0.06

Lamar Comp Stamped 4 17.7 4 17.7 0.02

Fort Walton Incised 1 19.1 1 19.1 <0.01

TOTAL 13, 932 33,628.7 216 921.2 7,489 34,908.7 21,637 69,458.6 100.00

The main difference between the Block D collection and the Kolomoki IV phase as describedby Pluckhahn (2003:Table 2-2) is the diversity and high relative frequencies of other pottery types of theWeeden Island series in the Block D assemblage. Combined, these form almost one-tenth of theWoodland pottery assemblage. The most common Weeden Island types include, in descending orderof frequency: Weeden Island Red, Carrabelle Punctate, Weeden Island Incised, Mound Field NetMarked, Weeden Island Zoned Incised/Punctate, Keith Incised, Indian Pass Incised, Carrabelle Incised,Wakulla Check Stamped, and Tucker Ridge Pinched.

92

Table 5-2. Woodland Period Ceramics from Block D.Type Unit Levels Miscellaneous Features Total

N g N g N g N g % by count

plain sand tempered 4,305 20,096.2 76 604.1 2,773 21,611.4 7,154 42,311.7 77.16

plain limestone tempered 1 12.4 1 12.4 0.01

St. Johns Plain 1 2.3 1 2.3 0.01

Swift Creek Comp Stamped 434 1,802.5 10 50.9 448 3,886.4 892 5,739.8 9.62

Napier Comp Stamped 6 30.0 28 629.5 34 659.5 0.37

Carrabelle Incised 7 32.7 13 201.6 20 234.3 0.22

Carrabelle Punctate 112 559.2 2 7.5 152 1,109.8 266 1,676.5 2.87

Indian Pass Incised 7 34.8 16 155.8 23 190.6 0.25

Mound Field Net Marked 11 63.1 1 3.5 24 326.2 36 392.8 0.39

Keith Incised 15 61.8 10 97.7 25 159.5 0.27

West Florida Cord Marked 3 19.1 1 3.3 4 22.4 0.04

Tucker Ridge Pinched 4 27.6 4 16.5 8 44.1 0.09

Weeden Island Incised 19 115.4 33 290.1 52 405.5 0.56

Weeden Island Red Filmed 200 972.09 1 0.7 254 1,861.9 455 2,834.7 4.91

St. Andrews Comp Stamped 3 88.0 3 88.0 0.03

Wakulla Check Stamped 4 32.4 9 49.8 13 82.2 0.14

Weeden Island ZonedIncised/Punctate

10 70.8 1 67.5 16 241.2 27 379.5 0.29

Weeden Island Zoned Red 1 30.1 1 30.1 0.01

unidentified incised 112 305.4 5 21.8 77 342.8 194 670 2.09

unidentified stamped 10 39.8 1 1.6 5 71.4 16 112.8 0.17

unidentified punctate 35 119.7 12 26.4 47 146.1 0.51

TOTAL 5,296 24,397.3 97 757.6 3,879 31,039.9 9,272 56,194.8 100.00

Table 5-3 summarizes the MNV analysis of pottery from Block D. We identified 57 vesselsfrom rim sherds. As noted in Chapter 2 and discussed in more detail below, seven basic vessel formsare evident in the Block D assemblage. Figure 5-1 documents vessel profiles.

As with sherds more generally, plain sand tempered vessels are the most common categoryrepresented in the MNV analysis, with 24 vessels forming 42.11 percent of the MNV assemblage. Alsoin keeping with the larger sherd assemblage, Swift Creek vessels are next in frequency (N=7, or 12.28percent). Various other types of the Weeden Island series make up the majority of the remainingvessels. These include, in descending order of frequency: Weeden Island Red (N=8, or 14.06 percent);Carrabelle Punctate (N=4, or 7.02 percent); Carrabelle Incised, Keith Incised, and Weeden Island ZonedIncised/Punctate (each N=3, or 5.26 percent); and Weeden Island Incised, Weeden Island Zoned Red, and Indian Pass Incised (each N=1, or 1.82 percent). Two Napier Complicated Stamped vessels roundout the remaining 3.51 percent of the vessel assemblage.

93

Table 5-3. Results of MNV Analysis of Ceramics from Block D.

Vessel Provenience Type formOrifice

Diameter(cm)

Rim Type Rim Fold(mm) Lip

1214 XU D2S, L3 Carrabelle Incised open bowl 18 folded 7.5 rounded1203 F147B Carrabelle Incised restricted bowl 14 folded 6.3 rounded1237 F147B Carrabelle Incised simple bowl/unrestricted jar 26 direct rounded1210 XU D2S, L4 Carrabelle Punctate neckless jar 22 direct rounded1211 F141B Carrabelle Punctate neckless jar 12 folded 12.1 rounded1225 F203 Carrabelle Punctate neckless jar 22 folded 10.7 rounded1205 F153 Carrabelle Punctate simple bowl/unrestricted jar 24 folded 13.2 rounded1124 F37 Indian Pass Incised restricted bowl 16 double folded 6.5, 12.2 rounded1117 F34 Keith Incised folded rim jar 14 folded 12.6 rounded1223 F155 Keith Incised folded rim jar 16 folded 7.1 rounded1231 F155 Keith Incised neckless jar 28 double folded 12.7, 27.5 rounded1204 F155 Napier restricted bowl 16 direct rounded1208 F203 Napier restricted bowl 20 folded 9.7 thinned interior1119 F34 plain sand tempered cup 12 wedge flat1215 F141A plain sand tempered dish/plate 12 folded 13 rounded1123 F37 plain sand tempered folded rim jar 30 direct rounded1235 F169 plain sand tempered folded rim jar 14 folded 16.2 rounded1115 TU18, L2 plain sand tempered open bowl 20 folded 27.1 flat1202 F175 plain sand tempered open bowl 26 wedge flat1212 F141A plain sand tempered open bowl 20 direct flat1218 F141A plain sand tempered open bowl 30 direct flat1220 F141A plain sand tempered open bowl 22 direct flat1221 XU D3S, L2 plain sand tempered open bowl 24 false fold 13 rounded1222 F141A plain sand tempered open bowl 28 direct pointed1114 TU18, L1 plain sand tempered restricted bowl 24 folded 18 flat1201 F170A plain sand tempered restricted bowl 14 folded 7.1 rounded1209 F141B plain sand tempered restricted bowl 20 wedge flat1226 F191A plain sand tempered restricted bowl 18 direct rounded1126 TU18, L4 plain sand tempered simple bowl/unrestricted jar 32 folded 8.4 rounded1113 TU18, L1 plain sand tempered neckless jar 34 folded 24.3 flat1116 TU18, L3 plain sand tempered simple bowl/unrestricted jar 26 direct rounded1125 F38 plain sand tempered simple bowl/unrestricted jar 30 folded 23.9 flat1128 F41 plain sand tempered simple bowl/unrestricted jar 28 wedge flat1213 F141A plain sand tempered simple bowl/unrestricted jar 24 direct flat1224 F185 plain sand tempered simple bowl/unrestricted jar 22 folded 23.7 flat1227 F141A plain sand tempered simple bowl/unrestricted jar 30 folded 5.4 flat1234 F141A plain sand tempered simple bowl/unrestricted jar 22 folded 6.7 pointed1120 F34 Swift Creek folded rim jar 24 folded 18.2 flat1200 F171 Swift Creek folded rim jar 26 folded 19.3 rounded1207 F171 Swift Creek folded rim jar 30 folded 19 rounded1206 F171 Swift Creek folded rim jar 22 folded 20.1 rounded1229 F147B Swift Creek folded rim jar 40 folded 19.6 pointed1127 TU18, L4 Swift Creek restricted bowl 24 folded 15.6 rounded1230 F141B Swift Creek simple bowl/unrestricted jar 30 direct flat1196 F165 Weeden Island Incised dish/plate 24 rounded1232 F155 Weeden Island Red dish/plate folded 16.4 rounded1216 F141A Weeden Island Red open bowl 22 folded 13.5 rounded1228 F141B Weeden Island Red open bowl 26 folded 12.4 rounded1118 F34 Weeden Island Red restricted bowl 30 folded 23.3 flat1233 F175 Weeden Island Red restricted bowl 28 folded 21 flat1197 F175 Weeden Island Red simple bowl/unrestricted jar 26 folded 19.2 flat1198 F174 Weeden Island Red simple bowl/unrestricted jar 26 folded 20.3 flat1217 F141A Weeden Island Red simple bowl/unrestricted jar 38 folded 22.1 flat1238 F176 Weeden Island Zoned Red dish/plate1239 F141A Weeden Island Zoned Incised/Punctate dish/plate1219 XUD4 Weeden Island Zoned Incised/Punctate neckless jar 18 double folded 5.7, 20.6 rounded1236 F147B Weeden Island Zoned Incised/Punctate neckless jar 16 direct rounded

94

95

Figure 5-1. Profiles of vessels identified in MNV analysis of Block D. Top and bottom rows: simple bowls/unrestricted jars. Shown approximately actual size, with interiors to right.

1237

1234

1205

1198

1197

12131217

1224

1227

1230

11131116

1125

1128

1236

1210

1211

1225 1219

1231

1223

1207

1200

1235

1117

1120

1123

1206

1229

1126

97

Figure 5-1. Profiles of vessels identified in MNV analysis of Block D (continued). Top row: neckless jars.Middle and bottom rows: folded rim jars. Shown approximately actual size, with interiors to right.

1233

1203

1226

1209

1208

1204

1201

1114

11181124

1127

98

Figure 5-1. Profiles of vessels identified in MNV analysis of Block D (continued). Top and bottom rows:restricted bowls. Shown approximately actual size, with interiors to right.

96

Figure 5-1. Profiles of vessels identified in MNV analysis of Block D (continued). Top and middle rows:open bowls. Bottom row: plates/dishes and cup. Shown approximately actual size, with interiors to right.

1214

1212

1202

1216

1220

1218

12211222

12281115

1196 1215 1232 1119

The remainder of this section consists of a detailed analysis of the pottery assemblage, beginningwith a discussion of vessel forms. This is followed by a focus on particular pottery types and categories.

Vessel Forms

Simple bowls and unrestricted jars. These two forms—combined in this analysis because of theuncertainty in differentiating them on the basis of smaller rim sherds—constitute the most commoncategory of vessel forms (N=14). Orifice diameters are relatively large, ranging from 22 to 38 cm witha mean of 27.6 cm. This is comparable to the range noted by Hally (2009) for unrestricted jars, and—aswith his sample—may include several size classes. The rims on these vessels are primarily folded (N=9),but also in some cases direct (N=4) or wedged (N=1). Folded rims average 17.6 mm, an exaggeratedwidth that is typical of the later Woodland period at Kolomoki (Pluckhahn 2003:Table 2-2). Lips areprimarily flat (N=10) but occasionally rounded (N=3) or pointed (N=1). In keeping with presumedutilitarian function of these vessels for cooking, serving, and short-term storage (Hally 2009), the simplebowls/unrestricted jars in the Block D assemblage are primarily plain (N=8). However, there are alsoexamples of Weeden Island Red (N=3), Swift Creek Complicated Stamped (N=1), Carrabelle Incised(N=1) and Carrabelle Punctate (N=1).

Restricted bowls. This form, presumably associated with serving and short term storage, is nextin frequency (N=11). Plain (N=4) is again the most common surface treatment, but is less commonin a relative sense given the presence of two Napier Complicated Stamped, two Weeden Island RedFilmed, and single examples of Swift Creek Complicated Stamped, Carrabelle Incised, and Indian PassIncised. Orifice diameters range from 14 to 30 cm with a mean of 20.4 cm. The rims on these vesselsare mostly folded (N=7) but there are rare examples of direct (N=2), wedged (N=1), or double folded(N=1). Folded rims on these vessels were generally narrower than those on the simplebowls/unrestricted jars, with a mean of 14.2 mm. Lips are rounded (N=6), flat (N=4), and thinned onthe interior (N=1).

Open bowls. This bowl form is slightly less common (N=10) than restricted bowls. These aremostly plain (N=7), but there are also two Weeden Island Red and one Carrabelle Incised. Orificediameters range from 18 to 30 cm with a mean of 23.6 cm, thus showing less variation in size thenHally’s rounded bowls. Rims are equally divided between folded (N=4) and direct (N=4) with isolatedexamples of false fold and wedged. Rim folds here, as on restricted bowls, are more narrow than thoseon simple bowls/unrestricted jars, with a mean of 14.7 mm. Lips are mainly flat (N=5) or rounded(N=4), or (less commonly) pointed (N=1).

Folded rim jars. The folded rim jar form is next in frequency (N=9). In keeping with the name,these are mainly folded, but one example of this form has a direct rim. Rim folds are generally wide,with a mean of 16.5 mm. Lips are almost exclusively rounded (N=7), but occasionally flat (N=1) orpointed (N=1). Orifice diameters are narrow relative to simple bowls/unrestricted jars, varying from14 to 40 cm (two size classes may be represented) with a mean of 24 cm. Swift Creek ComplicatedStamping is the most common surface treatment on these jars (N=5), followed by plain (N=2) andKeith Incised (N=2).

Neckless jars. These jars are only slightly less common (N=7) than the folded rim variety. Theyoverlap in sizes, with orifice diameter ranging from 12 to 32 cm and averaging 21.4 mm. Also like thefolded rim variety of jar, the rims are mainly folded (N=3). However, double folded (N=2), and direct(N=2) rims are not uncommon. Lips are exclusively rounded. Rim folds are variable: narrow on singlefolds, but wide on double folds. The mean for this class is 15.9 mm, intermediate between the simplebowls/unrestricted jars and the other bowl forms. Neckless jars, which appear to be used mainly forserving and short term storage, are almost invariably decorated, sometimes elaborately so. We notedtwo examples of Weeden Island Zoned Incised/Punctate, two of Carrabelle Punctate, and one of KeithIncised. Only one neckless jar is plain.

99

Cup. Only one cup was identified in the MNV analysis. It is from a plain vessel with a wedgedrim and an orifice 12 cm in diameter.

Dishes/plates. These forms, probably associated exclusively with serving, round out theassemblage, with five examples identified. One of the plates is a plain vessel with a folded, rounded lip. It has a diameter of 12 cm. Another of these vessels was identified from a Weeden Island Incised rimsuggestive of a plate approximately 24 cm in diameter. The third was identified from the rim of ashallow Weeden Island Red dish of a four or more sided squared form of indeterminate diameter. Thisvessel had a rim fold of 16.4 mm.

In addition to these three dish/plate forms identified from rims, we have counted here twoplates identified from diagnostic body sherds; each has incising on the interior of the plate, a form ofdecoration rare enough that these can confidently be identified as distinct vessels. One is Weeden IslandZoned Red. The other is Weeden Island Zoned Incised/Punctate.

Ceramic Types/Categories

Plain. As noted above, plain pottery makes up the majority of Woodland ceramic assemblage. The vast majority of the plain pottery is tempered with sand—generally abundant, fine sand. Mica isalso common in the paste of plain and other pottery types, but was presumably not added as temper(Pluckhahn and Cordell 2010). We recovered only one sherd each of limestone tempered and “temperless” (St. Johns-like) pottery. The paucity of these other temper types in the Block Dassemblage is consistent with the assumption that they are more common earlier in the occupationalhistory at Kolomoki (Pluckhahn 2003:Table 2-2). Much of the plain pottery in the Block D assemblageis smoothed or burnished, although this was not differentiated in our analysis.

Plain pottery was found in 37 of the 38 (97.4 percent) pit features in Block D. Consistent withits overall ubiquity, plain pottery is produced in virtually the entire range of vessel forms apparent in theBlock D assemblage, including the cup (N=1), dish/plate (N=1), neckless jar (N=1), folded rim jar(N=2), open bowl (N=7), restricted bowl (N=4), and simple bowl/unrestricted jar (N=8) categories. These include vessels suited to serving, storage, and cooking.

Plain pottery also runs the full gamut of rim and lip treatments. With regard to rims, we notedfolded (N=11), direct (N=8), wedged (N=4), and false folded (N=1). Lips include flat (N=14), rounded(N=8), and pointed (N=2).

Swift Creek Complicated Stamped. Swift Creek pottery is the most common decorated potterytype in terms of overall numbers and weight. It also exhibits a high ubiquity in features; we found SwiftCreek pottery in 32 (84.2 percent) of pit features.

We identified seven Swift Creek vessels in the Block D assemblage. The folded rim jar is thepreferred vessel form for this type (N=5), but there are also single examples of the simplebowls/unrestricted jar and the restricted bowl. These forms are associated primarily with cooking andstorage. Rims are almost invariably folded (N=6); the single exception is one direct rim. Lips arerounded (N=4), flat (N=2), and pointed (N=1).

The Swift Creek pottery from Block D bears impressions generally typical of the Late Woodlandperiod, with relatively narrow lands and grooves. I would conservatively estimate that around 20 distinctSwift Creek paddle designs are represented in the assemblage from Block D. Unfortunately, most ofthe Swift Creek sherds are much too small for the reconstruction of paddle designs, or even to matchwith paddle designs that have been previously reconstructed from Kolomoki and other sites (Snow2007). However, there are several intriguing exceptions that warrant more extended comment.

100

Figure 5-2 illustrates a partial paddle design (herein designated Design D-1) reconstructed fromseveral larger sherds from Feature 171. This is Vessel 1206 in the MNV analysis. Sherds bearing thispaddle design were most numerous in Feature 171, but possible matches were identified with sherdsfrom Features 153, 157B, 175, 183, and 185, and a level in XUD2. Some of these could be from thesame vessel, but it is also possible that some of these represent fragments of other vessels bearing thesame paddle stamp. The design has a central circular element set within a four-pointed star. A usefuldiagnostic feature is provided by a discrepancy in the number of lands and grooves in two opposingpoints of the star; one point has two lands (one quite uneven) and two grooves, while the other pointhas three lands and three grooves.

A similar, yet distinct, paddle design (Design D-6) was partially reconstructed from several largersherds from Feature 185 (Figure 5-3). These consisted only of body sherds, so no MNV number wasassigned. Possible matches were identified with other body sherds from Feature 171 and Level 3 in UnitD2N. Although the general motif represented in Design D-6 is similar to that in Design D-1, the twopaddle designs differ in the size and shape of the points that surround the central circular element. differs from the previously described design in

The partial paddle design (Design D-3) in Figure 5-4 was reconstructed from the sherdillustrated in the same figure. This is Vessel 1200, recovered from Feature 171. The paddle design isclearly impressed several times on this single sherd. Unfortunately, the two distinct circular elementson the sherd do not match well, even when the orientation is changed. It is possible that the paddleitself had two circular elements, but it is difficult to discern how these were connected.

The paddle design (Design D-4) in Figure 5-5 was reconstructed from fragments of a singleSwift Creek vessel (Vessel 1229) from Feature 147B. This paddle matches one identified from sherdsfound on the Vaughn and Mill Creek sites in the Big Bend region of the Ocmulgee River of southernGeorgia (Frankie Snow, personal communication 2007) (see Design #81 in Snow 2007). This is theonly obvious external contact represented in the assemblage from Block D, although a more thoroughcomparison would undoubtedly reveal other examples.

Figure 5-6 documents a large vessel fragment (Vessel 1207) from Feature 171 that bears anotherunique paddle design (Design D-2)—perhaps a variety of figure eight, but only a portion of the paddledesign is recognizable on the sherds we recovered due to the relatively heavy over-stamping. However,the distinctive nature of central element of the teardrop motif, possibly representing a flaw in the paddle,should facilitate efforts to match these sherds. Not represented in the design reconstruction is anotherpossible flaw in the paddle that connects two of the concentric lines.

Large fragments of another vessel from Feature 171 (Figure 5-7) are stamped with anotherdistinct paddle, herein designated as Design D-5. There is one recognizable element—asnowshoe—that is distinct from the previous paddle designs. The stamped design here is obscured bythe feint and overstamped paddle impressions. This pot is represented only by body sherds, so novessel number was assigned in the MNV analysis.

Figure 5-8 illustrates the potential connections between features based on paddle matches. Previous research suggests that Swift Creek paddles may have sometimes been used for many decadesor even several centuries (Snow and Stephenson 1992). Nevertheless, paddle matches demonstrate therough contemporaneity of features. Although it is difficult to account for sampling error in an analysisof paddle designs, the connections between features near the center of the presumed structure is strikingand provides another line of evidence in support of this interpretation.

101

Figure 5-2. Paddle Design D-1 (A) and sherds used in its reconstruction (B-E). B-D: Feature 171; E:Feature 185. Shown approximately actual size.

102

Figure 5-3. Paddle Design D-6 (A), with sherds used in the reconstruction (B-D), and sherds withpotentially matching paddle stamps (E-G). A-D, F: Feature 185; E: XUD2N, L3; G: Feature 171.Shown approximately actual size.

103

Figure 5-4. Paddle Design D-3 (top) and sherd from Feature 171 used in its reconstruction (bottom). Shown approximately actual size.

104

Figure 5-5. Paddle Design D-4 (A) and sherds used in its reconstruction (B-C). B-C: Feature 147B. Shown approximately actual size.

105

Figure 5-6. Paddle Design D-2 and sherds from Feature 171 used in its reconstruction. Shownapproximately actual size.

106

Figure 5-7. Paddle Design D-5 and sherds from Feature 171 used in its reconstruction. Shownapproximately actual size.

107

0 1 2

meters

¯

Figure 5-8. The distribution of Swift Creek pottery and paddle matches in pit features in the corearea of Block D.

171

185

155

183

157B153A

108

distribution of Swift Creekpottery in pit features:


no sherds1 or more sherds

175

paddle matches betweenSwift Creek sherds in features

Napier Complicated Stamped. Apart from Swift Creek, Napier is the only other complicatedstamped pottery type represented in appreciable quantities. We recovered 34 sherds of Napier pottery. While this quantity is not large, Napier sherds were nevertheless relatively common in pit features; 14(36.84 percent) of the 38 pit features produced at least one Napier sherd. Pit features with Napierpottery are strongly associated with the presumed structure in Block D, as indicated in Figure 5-9. Allof the pit features yielding Napier pottery are within or in immediate proximity to the structure, and onlyone of the pits completely within the house failed to produce any Napier sherds. This pattern pointsto the rough contemporaneity of these pits, and provides an additional line of support for theinterpretation of the structural pattern. Napier pottery dates to the interval from around A.D. 800 to1000 (Cobb and Garrow 1996; Markin 2010), consistent with the presumed occupation of Block Dbetween ca. 750 and 900 A.D.

The Napier pottery from Block D is differentiated from Swift Creek by narrower lands andgrooves and rectilinear design motifs (Figure 5-10). Lands generally range from around 1.0 to 1.5 mm,and grooves from approximately 1.5 to 2.0 mm. This former is similar to the range of 1-2 mm forlands on Napier at McKeithen (Milanich et al. 1997:68). However, Milanich and colleagues report awider range (3-4 mm) for grooves.

At least two distinct Napier design motifs, representing at least two different paddles, are presentin the assemblage from Block D. The most numerous and most widely distributed motif consists ofinterlocking line-filled triangles with borders defined by 5 lands and 4 grooves (see Figure 5-10D-F). Vessel 1208, identified from a sherd in Feature 203 (see Figure 5-10D), bears this motif. Otherexamples were noted on sherds from Features 141B, 154, 155, 165, 169, 171, and 185. All of thesefeatures are within or near the house. The small size of many of the sherds and the use of over-stamping and smoothing make it difficult to say if the sherds bearing this design motif were all stampedwith the same paddle.

The other variety of Napier stamping includes a design field set off by parallel, straight linesdefined by 3 lands and 2 grooves (see Figure 5-10A-C). Within these straight lines are triangular- orchevron-shaped elements. This motif appears similar to that found on Napier sherds at the McKeithensite, as described by Milanich and colleagues (1997:68). It is not precisely the same as motifs describedfor Napier pottery in northern Georgia (Wauchope 1966:Figure 15), but the size of the lines andgrooves is consistent. Examples of this motif were noted on sherds found in Features 155, 159, and191A. Again, given the size of the sherds and the nature of the stamping, it is difficult to say if morethan one paddle is represented.

Two Napier vessels were identified in the Block D assemblage. One of these was identifiedfrom sherds in Feature 203 and the other from sherds in Feature 155. Both vessels are restricted bowlsthat probably functioned primarily for serving or short term storage. One has an orifice diameter of 20cm, a folded rim 9.7 mm wide, and a pointed lip. The other is only 16 cm in diameter at the orifice, hasa direct rim with a lip that is rounded. These orifice diameters are comparable to the mean of 19.3 cmfor Napier vessels at McKeithen (Milanich et al. 1997:68).

St. Andrews Complicated Stamped. Three sherds excavated from Test Unit 18 in 2002 wereidentified as possible examples St. Andrews Complicated Stamped. One sherd was recovered fromFeature 37 and two were found in the fill of Feature 34A.

The two larger sherds from Feature 34A are illustrated in Figure 5-11. These sherds appear tphave been stamped with the same paddle. The motif is a line block comprised of 5 lands and 4 grooves. The lands are generally 1.0-1.5 mm, comparable to those on the Napier pottery from Block D. However, the spacing between the lands is wider; grooves on the two sherds from Feature 34A measurebetween 2 and 2.5 mm. The motif is generally consistent with Willey’s (1949:385) description of theearly variety of St. Andrews. All of the St. Andrews sherds from Block D are body fragments, thus no vessels were identified.

109

0 1 2

meters

¯

Figure 5-9. The distribution of Napier pottery in pit features in the core area of Block D.

191A203

174 165147B

171

185

175

155

176

154

169 153A

159

110

distribution of Napierpottery in pit features:


141B


Figure 5-10. Selected Napier sherds from Block D. A: Feature 155 (Vessel 1204); B: Feature 159; C:Feature 191A; D: Feature 203 (Vessel 1208); E: Feature 169; F: Feature 185. Shown approximatelyactual size.

111

Weeden Island Red and Zoned Red. Weeden Island Red pottery was relatively common inBlock D, forming about 5 percent of all the Woodland pottery. It was also fairly ubiquitous, with oneor more sherds found in 29 (76.32 percent) of the 38 pit features. In contrast, only one sherd of theZoned Red type was identified. This was recovered from Feature 176, a small bell-shaped pit feature(see description in the preceding chapter).

Figure 5-12 illustrates selected examples of the Weeden Island Red and Zoned Red pottery fromBlock D. Red filming is generally apparent on both the interior and exterior surfaces of most of thesherds of this type. On the single Zoned Red sherd, the filming is set off by incised and punctate lines(Figure 5-12A).

Eight Weeden Island Red vessels were identified in the MNV analysis. Servingvessels—particularly bowl forms—predominate. We identified three simple bowls/unrestricted jars,two restricted bowls, two open bowls, and one dish or plate. As described above, the latter is a four ormore sided rectangular dish. The single Zoned Red vessel is identified from a body sherd and consistsof a plate with incising on the interior surface.

Orifice diameters on the on the Weeden Island Red vessels are generally large, with an averageof 28 cm. This is consistent with their use as serving vessels. However, the Block D vessels averagesignificantly larger than the mean of 20 cm noted for Weeden Island Red vessels at McKeithen(Milanich et al. 1997:63), possibly representing an increase in vessel size from Middle to Late Woodland(as described in more detail in Chapter 7). Rims on Weeden Island Red vessels are invariably folded andthe folds are generally large, with a mean of 18.5 mm.

Figure 5-11. Selected St. Andrews Complicated Stamped sherds from Block D. Both sherds fromFeature 34A in TU18. Shown approximately actual size.

112

Figure 5-12. Selected Weeden Island Red (B-E) and Zoned Red (A) sherds from Block D. A: Feature176 (Vessel 1238); B: Feature 141B (Vessel 1228); C: Feature 175 (Vessel 1233); D: Feature 155 (Vessel1232); E: Feature 175 (Vessel 1197). Shown approximately actual size.

113

Weeden Island Incised and Zoned Incised/Punctate. Weeden Island Incised (N=52) and ZonedIncised/Punctate (N=27) are relatively uncommon in the Block D assemblage, combining to form onlyabout one percent of the Woodland pottery collection. Nevertheless, pottery of this type was relativelyubiquitous, with one or more sherds identified in 13 (34.21 percent) of the pit features in Block D(Figure 5-13). As with Napier, the presence of Weeden Island Incised appears to be correlated with thehypothesized house. Most of the pit features within the structure produced at least one sherd ofWeeden Island Incised or Zoned Incised/Punctate, and most of the features producing sherds of thesetypes are located within or immediately adjacent to the structure.

Selected Weeden Island Incised sherds are illustrated in Figure 5-14. Incised lines on thesesherds are typically around 1.0 mm. The separation between incised lines is variable, but typically doesnot exceed 3.0 mm. The quality of the execution is very high, albeit not always quite as careful as withIndian Pass Incised.

The MNV analysis resulted in the identification of one Weeden Island Incised vessel and threeWeeden Island Zoned Incised/Punctate vessels. The former is represented by a fragmentary rim sherdfrom a plate or dish. The rim itself is incised; little of the body is represented, so this too may beincised and punctate.

One of the Weeden Island Zoned Incised/Punctate vessels is also a plate or dish. This vesselis represented only by two body sherds, but the shape of these is diagnostic of the form. In this casethe incising and punctation are on the interior, suggesting this was a serving dish associated withhousehold ritual. The other two of the Zoned Incised/Punctate vessels consist of neckless jars withsmall orifice diameters (16 and 18 cm). These measurements are consistent with Weeden Island Incisedand Punctate vessels at McKeithen, which exhibited means of 13.7 cm and 16.4 cm, respectively(Milanich et al. 1997:64-65). One of the Zoned Incised/Punctate vessels in the Block D assemblage hasa double folded rim, while the other is direct. The lips on both vessels are rounded.

Carrabelle Punctate. Carrabelle Punctate is relatively common in the Block D assemblage, witha total of 266 sherds forming roughly 3 percent of the Woodland pottery assemblage. In addition, thereare large quantities of unidentified punctate pottery in the assemblage, and much of this is probably ofthe Carrabelle type. As might be expected given its overall high frequency of occurrence, CarrabellePunctate was ubiquitous in its distribution in Block D, with at least one sherd found in 23 (60.53percent) of the 38 pit features. In contrast with Napier and Weeden Island Incised, the distribution ofCarrabelle types (both punctate and incised) does not favor pits in or near the presumed structure.

The Carrabelle Punctate sherds from Block D display an amazing diversity in terms of the shape,size, and spacing of punctations (Figure 5-15). Some clearly appear to have been made with small sticksof hollow cane, others with wedge shaped instruments, and others possibly with fingernails. Particularlyintriguing is the specimen in 5-15I, which was decorated with a distinctive circular tool with severalirregular “bumps” in the center.

We identified four Carrabelle Punctate vessels in the MNV analysis. Three of these are necklessjars, while the fourth is a simple bowl/unrestricted jar. Mean orifice diameter on these vessels is 20 cm,comparable to the mean of 16.4 cm for pottery of this type at McKeithen (Milanich et al. 1997:65-66). Three of the four vessels have folded rims with relatively narrow rim folds (average of 12.0 mm). Thefourth has a direct rim. Lips are exclusively rounded.

114

0 1 2

meters

¯

Figure 5-13. The distribution of Weeden Island Incised and Zoned Incised/Punctate pottery in pitfeatures in the core area of Block D.

191A

141A

192

165147B

171

185

147A

155

154

169

159

115

distribution of Weeden IslandIncised and Zoned Incised/Punctate pottery in pit features:


141B


Figure 5-14. Selected Weeden Island Incised and Zoned Incised/Punctate sherds from Block D. A:XU D4, E Profile (Vessel 1219); B: Feature 155; C: Feature 141B; D: Feature 192; E: XU D2S, L4; F:Feature 147B (Vessel 1236); G: Feature 141A; H: XU D4, L4 (G and H are Vessel 1239). Shownapproximately actual size.

116

Figure 5-15. Selected Carrabelle Punctate sherds from Block D. A: Feature 191A; B: Feature 147B; C:Feature 175; D: Feature 203 (Vessel 1225); E: Feature 141B (Vessel 1211); F: XU D2S, L4 (Vessel1210); G: Feature 153 (Vessel 1205); H: Feature 155; I: Feature 165; J: Feature 170B. Shownapproximately actual size.

117

Carrabelle Incised. The Carrabelle Incised type, in contrast with the more plentiful punctatevariety, is represented by only 20 sherds constituting just 0.22 percent of all the Woodland pottery fromBlock D. Somewhat unexpectedly, Carrabelle Incised was nevertheless relatively ubiquitous, with atleast one sherd in 7 (18.42 percent) of all pit features. As noted above in reference to CarrabellePunctate type, there does not appear to be a correlation of the Carrabelle types with the presumedstructure.

By definition, the incised lines on Carrabelle Incised pottery are less carefully inscribed than onsherds of the Weeden Island and Indian Pass Incised types. The Carrabelle sherds from Block D aretypical in this regard (Figure 5-16). The incised lines are generally on the order of 1.0 mm, and thus notdissimilar from the other two types. However, the line spacing is generally greater (on the order of 4-5mm) and the lines are less perfectly parallel. There are exceptions, however, such as the sherd illustratedin Figure 5-16B, which could instead be classified as Indian Pass Incised (although the incised patternis typical of the Carrabelle type, which is why it was so classified).

Three Carrabelle Incised vessels in the Block D assemblage are comprised of one open bowl,one restricted bowl, and one simple bowl/unrestricted jar. Orifice diameters average 19.3 cm,comparable to the mean of 17.4 cm at McKeithen (Milanich et al. 1997:66). However, the three vesselsfrom Block D vary widely, from 14 cm for the restricted bowl to 26 cm for the simplebowl/unrestricted jar. Two of the vessels have folded rims, while the rim on the third is direct. Thefolded rims are very narrow (7.5 and 6.3 mm).

Keith Incised. Twenty-five Keith Incised sherds, account for just 0.27 percent of the Woodlandpottery from Block D. Despite the low quantity of pottery of this type, however, it is surprisinglyubiquitous; Keith Incised sherds were recovered from 6 (15.79 percent) of 38 pit features. Thedistribution shows no preference for features in or near the structure.

Figure 5-17 documents some of the variety in a selected sample of Keith Incised sherds fromBlock D. Three types of incising are apparent. One variety is tightly cross-hatched, with one directionof incising predominating (see Figure 5-17A). Another variety, nicely executed, has much broader cross-hatching with small punctations between the lines (see Figure 5-17B-C). Finally, there are sherds withbroader cross-hatched incised lines but lacking the punctations and generally less finely executed.

The MNV analysis revealed evidence for three Keith Incised vessels: two folded rim jars andone neckless jar. These likely functioned primarily for storage and, in the case of folded rim jars,perhaps also for cooking. The two folded rim jars have narrow orifice diameters (14 and 16 cm) andnarrow folds (12.6 and 7.1 mm). The neckless jar has a wide orifice (28 cm) and a double folded rim. The lips on all three vessels are rounded. One of the folded rim jars and the single neckless jar wererecovered from the same large bell-shaped pit (Feature 155) within the structure.

Indian Pass Incised. This was another minority type of the Weeden Island series, representedby 23 sherds that form 0.25 percent of the Woodland pottery from Block D. Indian Pass Incised sherdswere found in five (13.16 percent) of the 38 pits in Block D. These pits are relatively widely scatteredacross the block. Only one is located within the structure.

MNV analysis identified one Indian Pass Incised vessel, consisting of a restricted bowl withnarrow orifice diamter (16 cm) and double folded rim. A large fragment of this vessel was recoveredfrom Feature 37 in Test Unit 18, excavated in 2001.

As is typical for pottery of this type (Willey 1949:425-427), the incised lines are carefully drawn. Specifically, the lines are thin (less than 1.0 mm), parallel, and consistently spaced. The restricted bowlform is also typical.

118

Figure 5-16. Selected Carrabelle Incised sherds from Block D. A: Feature 170A; B: XU D2S, L3(Vessel 1214); C: Feature 147B (Vessel 1237); D: Feature 141B; E: Feature 147B (Vessel 1203); F: XUD15N, L1; G: XU D2N, L1; H: Feature 165. Shown approximately actual size.

119

Figure 5-17. Selected Keith Incised sherds from Block D. A: Feature 155 (Vessel 1223); B: Feature 154;C: Feature 169; D: Feature 157B; E: Feature 141C; F: XU D1S, L1; G: Feature 155 (Vessel 1231). Shown approximately actual size.

120

Other Pottery of the Weeden Island Series. Several other types of the Weeden Island series arepresent in trace quantities in the Block D assemblage. These include, in descending order of frequency,Mound Field Net Marked (N=36), Wakulla Check Stamped (N=13), Tucker Ridge Pinched (N=8), andWest Florida Cord Marked (N=4). Each of these represents less than 0.5 percent of all Woodlandpottery, and combined they form less than 1 percent of the Woodland pottery. None of these types isrepresented in the MNV analysis; although several large sherds of these types were recovered, noneconsisted or rim sherds.

Figure 5-19 illustrates selected examples of these types. The limited number of sherds of thesetypes is not too surprising. However, as noted above, previous work at Kolomoki had suggested thatseveral of these—particularly Tucker Ridge Pinched—might be at least slightly more common inKolomoki IV phase assemblages (Pluckhahn 2003:Table 2.2). The paucity of these types suggest thatthey are less common in this phase than was previously assumed. It also suggests that ceramic typeschanged relatively rapidly during the Late Woodland period, as discussed in more detail in Chapter 7and elsewhere by Pluckhahn (2010).

Figure 5-18. Selected Indian Pass Incised sherds from Block D. A-B: Feature 37 (Vessel 1124); C:Feature 174. Shown approximately actual size.

121

Figure 5-19. Selected other Weeden Island series sherds from Block D. A-C: Mound Field Net Marked;D-F: Wakulla Check Stamped; G-I: West Florida Cord Marked; J-L: Tucker Ridge Pinched. A: Feature141C; B: Feature 141; C, E: Feature 155; D: Feature 147B; F: XU D13W, L2; G: Feature 163; H: XUD25S, L2; I: XUD7E, L2; J: Feature 156; K: XU D5S, L2; L: XU D23N, L3. Shown approximatelyactual size.

122

Prehistoric Flaked Stone

Excavation of Block D resulted in the recovery of 6693 flaked stone artifacts weighing acombined 11,388 g (Table 5-4). More than half of this total (58.14 percent) was recovered from unitlevels. Features produced about 40.63 percent of the total. The remaining 1.23 percent was recoveredfrom miscellaneous other proveniences, such as profile cleaning and grab collections.

Raw Materials

The collection can be divided into four general categories of raw materials. Finally, a very smallfraction of the flaked stone from Block D bears the distinctive dark color and lustrous texture of chertsof the Ridge and Valley Province of northwestern Georgia, northeastern Alabama, and southeasternTennessee. The five artifacts of this material have a combined weight of 15 g and make up only 0.01percent of the flaked stone by count. However, we also recovered one unworked cobble of a dark greychert that could have been obtained from a Ridge and Valley source (Figure 5-20A).

Table 5-4. Flaked Stone Totals for Block D.Raw Material Type Unit Levels Miscellaneous Features Total

N N N N %

quartz primary flake 48 1 11 60 5.03

secondary flake 208 5 48 261 21.90

shatter/chunk 333 4 83 420 35.24

tertiary flake 350 3 75 428 35.91

core/core tool 13 1 7 21 1.76

flake tool 2 2 0.17

Total Quartz 954 14 224 1192 17.80

quartzite/sandstone

primary flake 12 1 9 22 9.40


shatter/chunk 36 1 17 54 23.08

tertiary flake 51 4 32 87 37.18

core/core tool 4 1 5 2.14

biface 1 1 2 0.86

Total Quartzite 148 10 76 234 3.50

Ridge/Valleychert

tertiary flake 1 1 20.00

biface 2 2 4 80.00

Total Ridge/Valley cherts 3 2 5 0.08

Coastal Plaincherts

primary flake 142 5 127 274 5.20


shatter/chunk 450 2 282 719 13.66

tertiary flake 1592 39 1560 3191 60.61

core/core tool 24 3 15 42 0.80

flake tool 1 1 3 5 0.10

biface 45 1 33 79 1.50

Total Coastal Plain cherts 2787 59 2419 5265 78.63

TOTAL FLAKED STONE 3891 83 2719 6696 100.00

123

Coastal Plain cherts are most common (N=5,265, weighing 11,388 g), accounting for more thanthree-quarters of the assemblage by both count (78.63 percent) and weight (78.08 percent). Most of thischert is white, yellow, brown, red, or pink in color. Much of this material closely resembles what hasbeen described as Ocala or Tallahatta Cherts (University of South Alabama [USA] 2004). However,cherts subsumed within this general category no doubt come from a variety of sources, including someon- or near-site and others more distant. Figure 5-21 illustrates selected chert cores from Block D. These are typical; with the exception of a few late stage preforms, the chert cores are generallyamorphous and blocky. They were likely used in the production of flakes for flake tools and bifaces.

Quartz (N=1192, weighing 1863 g) makes up 17.81 percent of the flaked stone by count. Therelative frequency is less by weight (12.77 percent), however, due to the small size of the quartz coresand debitage (as described below). Most of the quartz from Block D is of a colorless, translucent varietythat resembles crystal or glass. Indeed, we recovered a few quartz crystals that may have been used ascores. However, some of the quartz is opaque or semi-translucent and “milky” in color. We recovereda number of small pebbles of the same type, some worked and others not. These were probablyprocured from elsewhere, given that quartz is uncommon in the Coastal Plain. It seems possible thatquartz pebbles could be found in gravel bars on the Chattahoochee or Flint Rivers, which originate inthe quartz-rich Piedmont. Figure 5-22 illustrates selected quartz cores and other “core-like” artifactsfrom Block D. All of the cores are small and amorphous. They would be well suited to the productionof small flake tools, but would have been poorly suited for the production of larger tools.

Finally, a small but significant share (3.50 percent) of the flaked stone from Block D matcheswhat has generally been termed “Tallahatta Quartzite,” but which is more correctly described as asandstone (USA 2004) (N=234, weighing 1319 g). This material is coarser grained than the quartz inBlock D. It often has a grey or brownish tint and a “snowflake pattern.” Tallahatta Sandstone quarries are commonly found across southwestern Alabama, but may extend east to the Chattahoochee River. Figure 5-20C-D illustrates two cores of this material that were recovered from Block D.

As noted in the previous chapter, the density of flaked stone in larger features ranged from 0to 4.7 artifacts (n)/liter, with a mean of 0.6. As might be expected, Coastal Plain chert and quartz wereubiquitous in larger pit features (34 and 29 of 38 total, respectively). Because of their overall ubiquity,these types reveal no unambiguous patterning with respect to the presumed structure. However,quartzite/sandstone—with a surprisingly high ubiquity of 23 of 38 large pit features—does demonstratea preference for features in or near the structure (Figure 5-23).

Figure 5-20. Ridge and Valley Chert Cobble/Core (A)and Tallahatta Sandstone Cores (B-C) from BlockD. A: Feature 141C; B: Feature 171; C: XU D14, L4. Shown approximately actual size.

124

Figure 5-21. Selected chert cores from Block D. A: XU D19N, L2; B: XU D22E, L1; C: Feature 175;D: Feature 193. Shown approximately actual size.

125

Figure 5-22. Selected quartz cores and “core-like” artifacts from Block D. A-B: quartz crystals possiblyused as cores; C-E: cores; F-N: unworked or minimally flaked quartz pebbles. A: XU D8N, L3; B: XUD10N, L2; C-D, I-K: Feature 147B; E: Block D Clean-up; F: Feature 154; G: XU D13W, L1; H: Feature185; L: Feature 163; M: XU D3, L3; N: XUD17, L3. Shown approximately actual size.

126

0 1 2

meters

¯

Figure 5-23. The distribution of quartzite/sandstone in larger pit features in the core area of BlockD.

142143

141C141A

141B

191A203

194

174

193 205

192

165 147B147B

171

185170B

170A

175

155

176

158

156

154

169

146

157B

157A

153A

159

37

3534A

38

4133

127

number of quartzite/sandstonein larger pit features:


163A

0

1 or more

Debitage

Debitage—broadly defined to include flakes and shatter/chunks—forms the majority of theflaked stone assemblage (97.34 percent). This figure is probably significantly inflated, however, giventhat no attempt was made to account for flakes that may have been utilized as tools unless deliberateretouch was noted (as described in Chapter 2). No doubt much of the debitage from Block D was usedas expedient tools.

Figure 5-24 contrasts the percentages of early stage debris (primary and secondary flakes andshatter/chunks), late stage debitage (tertiary flakes), and tools (cores/core/tools, flake tools, and bifaces)for each variety of raw material. Quartz and quartzite/sandstone display remarkably similar patterns,with roughly 60 percent early stage debitage, 36-37 percent late stage debitage, and 2-3 percent tools. The high frequency of early stage debitage is consistent with reduction of cores on site. It would alsogenerally be considered evidence of local procurement of raw materials, but—in the case of quartz andperhaps also quartzite/sandstone—the cores and unworked manuports that we recovered (see Figures5-20 and 5-22) are small enough to have little cost of transportation, and thus could have been gatheredfrom more distant sources. Notably, although the percentages of early stage debitage are comparablefor quartz and quartzite/sandstone, they differ in how this debitage is comprised. Specifically, a muchlarger percentage of the quartz debitage consists of shatter/chunks. This reflects the fact the fact thatthe small quartz cores must have been worked with a bipolar reduction technique, resulting in greaterspalling and shatter. Quartz is also more difficult to predictably flake. Presumably, much of the quartzdebitage was utilized as flake tools, but this must await more detailed study with a microscope.

In contrast with quartz and quartzite/sandstone, the Coastal Plain chert from Block D consistsprimarily of late stage debitage (60.61 percent). This is somewhat surprising, given that such chert isavailable locally and would undoubtedly have been reduced on site with some regularity. However, asignificant fraction of the Coastal Plain chert was probably procured from more distant quarries, whereit was reduced to cores that could be more easily transported. In addition, chert was the preferredmaterial for bifaces, which require more late stage finishing and, eventually in many cases, re-sharpening.

Ridge and Valley chert is heavily biased towards tools over debitage, and late stage debris overthat from the early stages of core reduction. This is consistent with the fact that this chert was importedfrom some distance, probably in the form of finished tools.

0

10

20

30

40

50

60

70

80

90

quartz quartzite/sandstone Ridge/Valley chert Coastal Plain cherts

rela

tive

freq

uenc

y

early debitage late debitage tools

Figure 5-24. Comparison of the relative frequencies of types of debitage and tools by raw material.

128

There is little patterning to the distribution of early and late stage debitage in larger pit features. However, there is a slight tendency for both types of debris to show higher densities in pit featuresoutside the structure than within. This would be consistent with more intensive flintknapping in thearea outside the structure, as might be expected. Feature 170A, immediately west of the structure,exhibits the highest density of both early and late stage debitage.

Cores, Core Tools, and Flake Tools

The flake stone assemblage from Block D includes a number of tools, including both flake toolsand bifaces. Also included in this discussion are cores and core tools, in that these are objective, ratherthan detached pieces (i.e., debitage).

We recovered 68 cores or core fragments from Block D. Not surprisingly, most of these areCoastal Plain chert (N=42), but the relative frequency of cores for this raw material type is lower thanfor quartz (N=21) and quartzite (N=5). This is consistent with previously mentioned hypothesis thatmuch of the chert was initially reduced elsewhere. A small number of cores (N=4) show obviousflaking or wear indicative of use as tools. However, a microscopic analysis would undoubtedly revealthat more of the cores in the collection were so used.

The same is true of flakes, as noted several times previously. The number of flake tools (N=10)is clearly a very conservative estimate, reflecting only with clear and presumed deliberate retouch (ratherthan use-wear) on one or both sides. Thus, blade-like flakes (of which there are a very limited numberin the assemblage) are also excluded. The paucity of formal flake tools and blades suggests that muchof the emphasis was on more expedient flake tools that could be easily produced from amorphous cores. Figure 5-25 illustrates selected examples of flake tools from Block D.

Figure 5-25. Selected flake tools from Block D. A-E: unifacial flake tools; F: bifacial flake tool. A: XUD12W, L3; B: XU D10S, L2; C: XU D3N, L2; D: Feature 141B; E: Feature 147B; F: XU D20W, L2. Shown approximately actual size.

129

Formal Bifacial Tools

The Block D excavations produced 84 bifaces or biface fragments. The majority of these aremanufactured from Coastal Plain cherts (N=79). We recovered no quartz bifaces, but found one madefrom quartzite/sandstone and four from Ridge/Valley chert.

Thirty-eight of the bifaces from Block D retain enough of the proximal hafting area element anddisplay significant diagnostic attributes to be assigned to classified to morphological clusters and namedtypes. Table 5-5 presents summary data for these 38 hafted bifaces, which are illustrated in Figure 5-26.

Pluckhahn and Norman (2011), following the example of Farr’s (2006) analysis of Archaic pointsin Florida, have suggested that it may be useful to focus more on general morphological categories or“clusters” than the multitude of types that have been proposed to describe Woodland points in Georgia,Alabama, and Florida (Baker 1995; Bullen 1975; Cambron and Hulse 1975; Schroder 2006; Whatley2002). Specifically, they suggest the hafted bifaces from Kolomoki fall in three general morphologicalclusters: proximally contracting, proximally straight/expanding, and triangular. These divisions are usedto organize the discussion that follows, although I also discuss particular types that are represented.

Proximally Contracting Cluster. This cluster is composed of bifaces with clearly discerniblehafting areas but which lack true necks in the sense of constrictions in the haft area that are less thanor equal to the width of the base. Thus, haft length is greater than zero, while neck height and neckwidth are not recorded. Included here, in addition to points typically identified as contracting stemmed,are lanceolate, ovate, and spike forms. Slightly less than one-half (N=18, or 47.37 percent) of the 38points in the Block D assemblage are assigned to this cluster. This contrasts with the overall assemblageof hafted bifaces from Kolomoki, of which the contracting stemmed cluster comprises only about one-third (Pluckhahn and Norman 2011). The higher relative frequency in Block D suggests that thisgeneral form was more commonly utilized in the later stages of the occupation of Kolomoki.

Five named types can be identified in the proximally contracting cluster. Most common is theNew Market type represented by six specimens, all manufactured from Coastal Plain chert. The six NewMarket points exhibit a mean ML of 31.1 mm and mean MW of 15.8 mm, somewhat smaller than thevalues cited by (Cambron and Hulse 1975:96).

Five Woodland Spikes are present in the Block D assemblage. Following Pluckhahn andNorman (2011), the spikes were sorted in to two varieties. Two of the spikes are defined as Variety 1,with long haft areas but relative wide bases. I have opted not to assign these to one of the manyvarieties of named spikes and lanceolates, but they would probably correspond best with lanceolate typessuch as Benjamin (Cambron and Hulse 1975:11) and Flint River Spike (Cambron and Hulse 1975:53;DeJarnette et al. 1962). Generally, this variety of spikes is more common at Kolomoki (Pluckhahn andNorman 2011); their low relatively frequency in Block D suggests their use may have declined by thelate or terminal Late Woodland. However, we recovered one example from Feature 155.

Three of the spikes more closely match the second variety of spikes, with short hafting areas andwide bases. These three points account for one-third of total hafted bifaces of this type at Kolomoki;the high relative frequency here thus suggests that these were more common in the late or terminal LateWoodland. Two of the three points were found in feature contexts (Features 165, 194).

Four hafted bifacess in the proximally contracting cluster approximate the Ebenezer typedescribed by Cambron and Hulse (1975:42). These account for all but one of the bifaces of this typein the total Kolomoki assemblage, suggesting a strong association with the later Late Woodland period. Three of the Ebenezer points were found in feature contexts, including two from Feature 141A and onefrom Feature 141C. All of the Ebenezer points are manufactured from Coastal Plain cherts. Theyaverage 44.8 mm in ML and 18.4 mm in MW, generally in keeping with the type (Cambron and Hulse1975:42).

130

Table 5-5. Type and Metric Data for Hafted Bifaces from Block D.PPK TypeName TULevel ML

(mm)MW(mm)

BLL(mm)

BLW(mm)

BW(mm)

HL(mm)

NW(mm)

NH(mm)

MT(mm)

Weight(g) Form Material Color Diapheneity Comments

proximally contracting cluster255 Ebenezer XUD10N, L2 42.3 20.2 34.0 20.2 5.3 8.3 8.8 6.6 plano-convex Coastal Plain chert yellow/brown opaque slight break at tip256 Ebenezer F141A 47.7 17.0 40.5 17.0 7.0 7.2 8.3 5.5 biconvex Coastal Plain chert yellow/brown opaque257 Ebenezer F141C 40.6 19.8 31.6 19.8 7.3 9.0 8.7 4.5 biconvex Coastal Plain chert grey opaque274 Ebenezer F141A 48.5 16.7 42.4 16.7 6.2 6.7 6.6 4.5 biconvex Coastal Plain chert yellow/brown opaque130 New Market TU18, L1 33.9 16.0 25.2 16.0 11.3 11.4 6.3 2.0 biconvex Coastal Plain chert yellow/brown opaque131 New Market TU18, L2 33.2 14.6 23.3 14.6 9.3 10.0 5.0 2.0 plano-convex Coastal Plain chert yellow semi translucent248 New Market XUD13W, L2 23.6 14.0 15.5 14.0 7.9 7.6 6.0 1.6 biconvex Coastal Plain chert yellow/red opaque slight break at tip250 New Market F185 37.0 17.7 29.6 17.7 7.5 9.2 7.0 3.1 biconvex Coastal Plain chert yellow/brown semi-translucent252 New Market XUD10S, L2 34.1 17.8 25.0 17.8 7.4 9.1 6.9 3.5 biconvex Coastal Plain chert pink opaque253 New Market F141A 25.0 14.6 16.3 14.3 5.5 8.7 7.0 2.0 biconvex Coastal Plain chert yellow opaque249 Spike, Variety 1 XUD3S, L2 42.0 15.5 28.7 14.9 5.3 13.3 9.2 5.5 biconvex Ridge/Valley chert black/grey opaque259 Spike, Variety 1 F155 39.1 13.6 33.4 13.6 5.5 9.1 6.2 3.3 biconvex Coastal Plain chert yellow semi translucent260 Spike, Variety 2 XUD7W, L3 48.6 16.3 42.4 16.3 7.4 10.4 9.4 5.0 plano-convex Coastal Plain chert white semi translucent slight break at stem266 Spike, Variety 2 F165 42.1 15.7 36.0 15.7 7.9 9.8 8.5 5.0 plano-convex Coastal Plain chert white opaque slight break at stem280 Spike, Variety 2 F194 48.0 15.7 39.7 15.7 8.7 8.3 8.2 5.0 biconvex Coastal Plain chert yellow opaque142 Swannanoa TU18, L1 34.0 20.3 27.2 20.3 9.4 9.4 5.0 3.0 plano-convex Coastal Plain chert white opaque258 Swannanoa XUD6W, L3 27.7 13.5 19.8 13.3 6.9 9.3 6.6 2.0 biconvex Coastal Plain chert yellow/brown opaque slight break at base281 Tampa F147B 30.8 14.6 25.9 14.6 5.4 4.6 6.5 2.2 Plano-convex Coastal Plain chert yellow/brown opaque

triangular cluster272 WMT F147B 26.1 16.6 26.1 16.6 16.6 4.0 1.3 biconvex Coastal Plain chert yellow/brown semi translucent

273 WMT XUD12E, L3 17.0 13.3 17.0 13.3 13.3 3.0 0.5 biconvex Coastal Plain chert pink opaque slight break at tip, major break at base

275 WMT XUD20E, L1 24.7 16.7 24.7 16.7 16.1 5.5 2.3 biconvex Ridge/Valley chert black/grey opaque277 WMT F147B 22.6 14.6 22.6 14.6 14.6 4.5 1.1 biconvex Coastal Plain chert yellow/brown opaque slight break at tip and base279 WMT F194 27.1 17.0 27.1 17.0 14.5 5.0 1.8 biconvex Ridge/Valley chert black/grey opaque platform on side near base

proximally straight and expanding cluster4 Bakers Creek F33 39.7 17.3 28.6 17.3 15.2 12.8 14.5 7.2 7.9 5.0 biconvex Coastal Plain chert yellow/brown semi translucent slight break at tip

105 Bakers Creek TU18, L1 41.1 17.6 25.4 17.6 14.4 15.4 13.3 7.4 7.2 5.0 biconvex Coastal Plain chert yellow/brown opaque243 Bakers Creek XUD1N, L1 33.6 19.1 25.6 19.1 12.6 8.8 12.0 8.8 6.6 3.5 biconvex Coastal Plain chert yellow/brown semi translucent267 Bakers Creek XUD16S, L1 41.1 21.0 35.1 21.0 18.4 9.8 16.3 7.2 6.0 4.7 plano-convex Coastal Plain chert yellow/red semi translucent269 Bakers Creek F147B 45.1 21.0 36.8 21.0 14.5 10.1 12.7 7.4 6.7 4.8 biconvex Coastal Plain chert yellow/brown opaque246 Broward XUD2N, L3 49.7 22.2 38.7 22.2 9.7 13.1 11.5 7.0 10.7 7.5 biconvex Coastal Plain chert grey opaque slight break at stem54 Duval Type 2 F37 40.7 17.3 34.4 17.3 10.0 8.1 9.5 2.8 7.6 5.0 biconvex Coastal Plain chert yellow/brown semi translucent slight break at tip

251 Duval Type 2 F141B 42.7 16.6 32.6 16.6 6.1 10.1 16.6 10.1 8.9 5.5 biconvex Ridge/Valley chert black/grey semi translucent slight break at tip263 Duval Type 2 F163 48.3 20.2 38.9 20.1 10.2 10.2 9.9 6.0 8.6 6.8 plano-convex Coastal Plain chert yellow/brown semi translucent148 Duval Type 3 TU18, L1 52.4 18.0 41.0 18.0 11.4 10.7 9.9 7.3 9.7 7.0 biconvex Coastal Plain chert red semi translucent276 Jacks Reef C XUD4E, L1 37.6 16.7 31.5 16.0 13.3 6.2 10.6 6.1 3.9 2.8 plano-convex unidentified chert brown semi translucent247 Swan Lake F147A 40.3 16.3 30.3 16.3 12.8 9.0 10.6 6.2 6.4 3.9 biconvex Coastal Plain chert yellow/red semi translucent261 Swan Lake XUD9N, L1 40.6 15.6 30.8 15.6 12.4 12.5 10.4 5.7 6.2 3.0 biconvex Coastal Plain chert yellow/brown opaque slight break at tip265 Swan Lake F153 36.4 14.8 27.9 14.6 14.8 10.5 13.1 5.9 5.7 3.0 biconvex Coastal Plain chert yellow/brown semi translucent very slight break at tip254 WISS F141B 32.5 15.8 23.7 15.8 10.4 8.8 15.8 8.8 5.7 2.0 biconvex Coastal Plain chert white semi translucent

Figure 5-26. Hafted bifaces from Block D. Numbers correspond to Table 5-5. Shown approximatelyactual size. Top row: Ebenezer. Second row: New Market. Third row: Spikes. Fourth row: Swannanoa(left two) and Tampa (right).

132

Figure 5-26. Hafted bifaces from Block D (continued). Numbers correspond to Table 5-5. Shownapproximately actual size. Top row: Woodland/Mississippian triangular (WMT). Second row: BakersCreek. Third row: Broward. Bottom row: Duval Type 2.

133

Figure 5-26. Hafted bifaces from Block D (continued). Numbers correspond to Table 5-5. Shownapproximately actual size. Top row: Duval Type 3. Second row: Jacks Reef Corner Notched. Thirdrow: Swan Lake. Bottom row: Weeden Island Straight Stemmed (WISS).

134

Two points in this cluster are identified as examples of the Swannanoa type described by Keel(1976:196). Consistent with the type description, these have weak shoulders and excurvate blades. Bothof the of this type from Block D are manufactured from cherts of the Coastal Plain. However, in theoverall Kolomoki assemblage there are several examples made from cherts of the Ridge and Valleyprovince of Tennessee and northern Georgia and Alabama, where the type name is more frequentlyemployed. Swannanoa is generally assumed to be an Early Woodland type (Whatley 2002:114). Whilenone of the Kolomoki specimens are from securely dated contexts, their presence at the site, andparticularly in Block D, suggests that there temporal range extends to the Middle and Late Woodlandas well.

The Tampa type is the final variety represented in the proximally contracting cluster. Oneexample of this type was recovered from Feature 147B, dated to cal A.D. 780 to 980. This type inuncommon at Kolomoki; the Block D specimen represents one of only six in the total assemblage ofmore than 200 points (Pluckhahn and Norman 2011). The Tampa point from Block D is consistentwith the description and range of measurements cited by Bullen (1975:10), except that it is thick and hasstraight blade edges. It is possible that this point represents an unfinished Woodland/Mississippiantriangular, two examples of which were also recovered from this feature.

Proximally Straight and Expanding Cluster. This cluster is defined by bifaces that evidenceconspicuous hafting areas set off by shoulders, as well as some semblance of a true neck in the senseof a constriction narrower than or roughly equal to the base. Thus HL, NH, and NW are all greaterthan zero. This cluster includes points typically described as straight-stemmed, expanding-stemmed,and notched. The 15 points in this cluster make up around 40 percent of the Block D assemblage. Inthe total hafted biface assemblage from Kolomoki, this cluster is better represented (forming around60 percent) (Pluckhahn and Norman 2011), suggesting that, in general, the use of points of this formalso declined by the late or terminal Late Woodland. However, some varieties within this cluster mayhave been used more in the later Late Woodland, as described in more detail below.

The proximally straight and expanding cluster includes seven named types. Most numerous isthe Bakers Creek type (DeJarnette et al. 1962) (N=5). This point type is generally more common atKolomoki, however. All of the Bakers Creek points from Block D are made from Coastal Plain cherts. The points average 40.1 mm in ML and 19.2 mm in MW. These means are lower than the range ofmeasurements cited by Cambron and Hulse (1975:8). However, the Kolomoki sample appearsconsistent with those from other Woodland sites in Georgia (Whatley 2002:18-19). In fact, the rangesand means conform very closely with those for a sample of Bakers Creek points from the Leake site innorthwestern Georgia (Pluckhahn 1998:156-157). Bakers Creek points were found in feature contextsin two cases (Features 33 and 147B). The latter feature was dated to cal A.D. 780 to 980.

Three examples of the side notched Swan Lake type (Cambron and Hulse 1975:120) wereretrieved from Bock D. Two of these were found in pit features (Features 147a and 153). This typeis not common at Kolomoki as a whole; in fact, the three specimens from Block D constitute 60 percentof all of the points of this type in the total assemblage (Pluckhahn and Norman 2011). This suggeststhis type was most prevalent in the late or terminal Late Woodland period.

Three points from Block D match the description Duval Subtype 2 (Bullen 1975:13), with anearly straight and narrow base. Notably, all three were found in features, suggesting a secure associationwith the late or terminal Late Woodland. All three points were made from chert; two of Coastal Plainvarieties and one of Ridge and Valley types.

The four remaining point types in this cluster are represented by only one specimen each. Oneof these is an example of third variety of the Duval type, with a base that expands slightly but remainsnarrower than the blade (Bullen 1975:13). Another matches Bullen’s (1975:15) Broward type, withstraight to excurvate blade edges and fairly sharp blade corners. One may be an example of the JacksReef Corner Notched type (Ritchie 1961). Finally, one point is tentatively assigned to Winston Baker’s(1995:428) Weeden Island Straight Stemmed type. All four of these points are manufactured from

135

cherts of the Coastal Plain, although the color and texture of the Jacks Reef point suggests that the chertmay be exotic. The Weeden Island Straight Stemmed point was the only one example recovered fromfeature contexts (Feature 141B).

Triangular Cluster. Triangular bifaces, defined as points with triangular forms and lacking necksand clearly defined hafting areas, account for only about 3 percent of all the hafted bifaces fromKolomoki (Pluckhahn and Norman 2011). They are far more common in Block D, with five examplesforming roughly 13 percent of the assemblage.

The five triangulars from Block D are all relatively small in terms of ML and MW and thus canbe classified under the generic type Woodland/Mississippian triangular (WMT). All five of these pointshave widths less than 18 mm, and thus would be classified as Mississippian triangulars under the ruleof thumb devised by Sassaman and colleagues (1990:165) for the Savannah River Site. Whatley(2002:64), however, puts the threshold between the Late Woodland and Mississippian varieties at 17-20mm, and all four of these points fall within this range (one is smaller, but falls in this range whenbreakage is accounted for).

The Block D specimens represent the only small triangulars in a total assemblage of more than200 hafted bifaces from Kolomoki (Pluckhahn and Norman 2011). Although it is possible that otherpoint types in the assemblage were used as arrow points, small triangular forms like these are widelyconsidered the first type that was specifically manufactured for use with an arrow (Blitz 1988; Nassaneyand Pyle 1999). Metric analysis of the Kolomoki points lends support to this assumption; the smalltriangular is the only type at Kolomoki that consistently meets criteria for arrows (e.g., Nassaney andPyle 1999, Shott 1997; see Pluckhahn and Norman 2011). Given that Block D is the only context atKolomoki from which small triangular points have been recovered, it would appear that bow and arrowtechnology was adopted at the time of Block D was occupied, in the late Late Woodland. In furthersupport of this interpretation, three of the points were found in secure contexts in pits: two in Feature147B and one in Feature 194. As noted above, the former feature was dated to cal A.D. 780 to 980. The implications of the discovery of triangular points in Block D are discussed in more detail in Chapter7.

Miscellaneous Prehistoric Artifacts

The Block D excavations produced a number of miscellaneous prehistoric artifacts, as describedabove with regard to excavation units (Chapter 3) and features (Chapter 4). The most common classof miscellaneous artifacts by weight is FCR. We recovered a total of 13,135.5 g of FCR from Block D. As discussed in Chapter 3, we noted several areas of concentration in the distribution of FCR acrossexcavation units. It later became apparent that these were near the center of the presumed structure andin the general area of features that also exhibited high densities of FCR. As mentioned above, however,FCR is difficult to differentiate from naturally-occurring rock at Kolomoki, and the data here no doubtreflect some subjectivity in collection. Most of the FCR consists of sandstone, although some is quartzand quartzite.

Pecked stone (N=15, weighing 9105.8 g) stone is considerably less common than FCR. Asnoted in Chapter 4, one of the pecked stones—a large anvil or nutting stone from Feature171—accounts for much of the total weight for this class of artifacts. Most of the other pecked stoneappears to have been used as hammerstones. Like FCR, the pecked stone includes sandstone, quartz,and quartzites.

Groundstone (N=11, weighing 1164.4 g) is also relatively uncommon in the Block Dassemblage. This is not surprising given that these are comprised primarily of igneous rocks such asgranite that are mostly exotic to the Coastal Plain. Unfortunately, all of the groundstone is fragmentary,but several are large enough to suggest they were parts of axes or adzes.

136

Table 5-6. Summary Data for Miscellaneous Prehistoric Artifacts from Block D. Artifact Type Measure Unit Levels Features Total

mica g 6.8 7.2 14.0

groundstone N 8 3 11

g 695.7 468.7 1164.4

pecked stone N 4 11 15

g 2804.9 6300.9 9105.8

limestone N 4 4 8

g 47.7 58.5 106.2

FCR N 143 116 259

g 5437.3 7698.2 13135.5

fired clay N 1 14 15

g 3.0 880.0 883.0

quartz pebble N 27 26 53

g 505.9 562.9 1068.8

sandstone ball N 1 - 1

g 18.6 - 18.6

quartz crystal N 1 4 5

g 3.5 74.9 78.4

possible pigment stone N 2 1 3

g 5.0 17.0 22.0

Prehistoric ceramic pipe bowl N - 1 1

g - 17.0 17.0

Chert cobble (possiblyRidge/Valley), unworked

N - 1 1

g - 14.0 14.0

The relatively small quantity of fired clay (N=15, weighing 83 g) in Block D suggests that thehypothesized structure was not daubed. Instead, the fired clay was probably a product of cooking, asindicated by the fact that most of this came from features.

Quartz pebbles (N=53, weighing 1068.8 g) are the most common class of the miscellaneousartifacts by count. Although the examples included here were not flaked, these appear identical to thequartz cores that were recovered (as describe din more detail above). Thus, these might have beenintended for use as cores. Some are reddened, however, suggesting they might have been used forcooking.

Also as noted above, we recovered a small number of quartz crystals (N=5, weighing 78.4 g). These too might have been flaked or used as cutting tools. However, it is also possible that these wereused as ornaments or ritual objects. A small sandstone “ball” (19 g) could have used for similarpurposes. Small quantities of limestone (N=8, weighing 106 g) and possible pigment stone (N=3,weighing 22 g) might also have been used for ornamentation. The latter consist of small nodules ofwhat appear to be hematite and limonite.

137

The recovery of 14 g of mica from Block D suggests that this material was still occasionallyworked into ornaments, a number of which were recovered by Sears (1956) from burial mound contextsat Kolomoki. However, the quantity of mica from Block D is quite small in comparison to that fromBlock A and other, earlier contexts at Kolomoki (Pluckhahn 1998, 2003), suggesting that this industrywas in decline by the late Late Woodland.

Finally, we retrieved one small fragment of a prehistoric ceramic pipe bowl. This was retrievedfrom Feature 38 in Test Unit 18. Ceramic pipes appear to have been uncommon at Kolomoki ingeneral; relatively few were recovered during earlier work at the site either (Pluckhahn 2003; Sears 1956).

Historic Artifacts

The Block D excavations produced an assemblage of historic artifacts that would be consideredvery modest for most historic sites, but which is surprisingly large and diverse for Kolomoki (Table 5-7). All of the historic material was recovered from excavation unit levels, rather than features. The majorityof these artifacts were found in the first two excavation levels, corresponding to the modern plowzoneand colluvium, as described in Chapter 3. There is no apparent clustering in the distribution of historicmaterials. Selected historic artifacts from Block D are illustrated in Figure 5-27.

Bottle glass makes up the majority of the historic artifact assemblage, and most of this is oliveor dark green (N=6, 7.5 g). We recovered a few sherds of amber (N=2, 1.5 g) and clear (N=1, 0.8 g)glass as well.

Nails in the Block D assemblage include examples of both the cut (N=2, 4.0 g) and wire (N=1,0.6 g) varieties. We also recovered a small quantity of metal wire (N=2, 4.6 g) (perhaps from a fence)and a small section of metal pipe (N=1, 47.7 g).

We recovered several artifacts relating to guns. These include one percussion cap (1.0 g) andtwo pieces (9.4 g) of lead waste or sprue. The latter was likely a byproduct of the manufacture of leadshot, produced either when excess lead was trimmed from the edge of a bullet mold or when leaddripped as it was poured into the mold (Blakney-Bailey 2008:177). Lead waste is commonly found onhistoric sites (particularly historic-era Indian sites) from the eighteenth- and early nineteenth-century(Blakney-Bailey 2008:178; Fairbanks 1962:54). Its occurrence in Block D, coupled with the olive bottleglass, suggests the presence of an early historic settlement on this portion of the site, associated eitherwith Creek Indians or Americans of African or European descent.

Table 5-7. Summary Data for Historic Artifacts from Block D.olive bottle

glassamberbottleglass

clear bottleglass

wire nails cut nails metal wire metal pipe percussioncap

leadsprue

N g N g N g N g N g N g N g N g N g

Unit Levels(Total)

7 9.7 2 1.5 1 0.8 1 0.6 2 4.0 2 4.6 1 47.7 1 1.0 2 9.4

138

Summary

The excavation of Block D resulted in the recovery of over 20,000 pieces of pottery and morethan 10,000 fragments of flaked stone. We also retrieved lesser quantities of miscellaneous prehistoricand historic artifacts.

The pottery assemblage includes a veritable smorgasbord of Late Woodland types, includingmost of the Weeden Island series types more common to the south along the Gulf Coast, as well astypes such as Swift Creek and Napier Complicated Stamped more common to the north and east. Wealso recovered small quantities of Mississippian period Lamar pottery.

An MNV analysis of the Woodland pottery revealed the presence of 57 vessels taking a varietyof forms and sizes. Perhaps most impressive is the identification of Weeden Island Zoned Red Filmedand Weeden Island Zoned Incised/Punctate plates or dishes. These have been recovered elsewhere inoff-mound areas at Kolomoki (Pluckhahn 2003), but only in small quantities. At the McKeithen site,pottery of this type was primarily associated with a high status burial on top of one of the mounds(Milanich et al. 1997). The possible implications of the increase (vis-a-viz Block A) in the frequency ofzoned incised, zoned red filmed pottery, and other varieties of Weeden Island ceramics more generallyare discussed in Chapter 7.

The flaked stone assemblage is comprised mainly of debitage. Most of this is Coastal Plainchert, but quartz and quartzite/sandstone are also well represented and there are trace amounts of chertsfrom the Ridge and Valley province. Tools are less common than debitage, but include 38 nearly wholeprojectile points. These were sorted into three general morphological categories and 12 different types. Probably most significant is the discovery of five small triangular points, a type not previouslyencountered at Kolomoki. The recovery of these points signals the adoption of new—or at leastimproved—bow and arrow technology at the time of the Block D occupation, as discussed in moredetail in Chapter 7.

The distributional patterns of several classes of artifacts lend additional support to theinterpretation of a structure in Block D. Swift Creek paddle matches connect several features near thecenter of the presumed structure. All of the pit features yielding Napier pottery are within or inimmediate proximity to the structure, and only one of the pits completely within the house failed toproduce any Napier sherds. The presence of Weeden Island Incised also appears to be correlated withthe hypothesized house; most of the pit features within the structure produced at least one sherd of

Figure 5-27. Selected historic artifacts from Block D. A: olive bottle glass; B-C: lead sprue. A: XUD12W, L3; B: XU D11S, L2; C: XU D22E, L2. Shown approximately actual size.

139

Weeden Island Incised or Zoned Incised/Punctate, and most of the features producing sherds of thesetypes are located within or immediately adjacent to the structure. With regard to flaked stone,quartzite/sandstone also demonstrates a preference for features in or near the structure. On the otherhand, there is a slight tendency for both early and late stage debitage to show higher densities in pitfeatures outside the structure than within.

140

Chapter 6: Faunal and Botanical Remains

Mary Theresa Bonhage-Freund, Matthew C. Compton, Linda Scott Cummings, Chad Yost, and Thomas J.Pluckhahn

As noted in Chapter 2, where the methods of analysis are described in detail, faunal andbotanical remains were analyzed by specialists, whose reports are summarized here largely verbatim. The section on faunal remains is drawn from Compton’s (2009) report, with minor additions byPluckhahn. The discussion of macobotanical remains is drawn from a report by Bonhage-Freund(2009), again with minor additions by Pluckhahn. Finally, the microbotanical section is based largelyon a report by Yost and Cummings (2010), as edited for reproduction here by Pluckhahn.

Faunal Remains

The Block D faunal assemblage is relatively small, including 1,595 specimens weighing 1,501.1g (Table 6-1). Four taxa including at least 13 individuals are represented. All of the species indentifiedare terrestrial taxa common to the Coastal Plain environments surrounding Kolomoki.

Table 6-2 summarizes the faunal remains by major provenience (some related proveniences,such as halves of features, have been combined in the interest of space). Forty-one of the featuresproduced measurable quantities of bone. We also recovered faunal remains from several excavationunits.

Turtles are represented by only five specimens. Of these, one specimen is identified as easternbox turtle. The eastern box turtle is most frequently found in open woodlands but also occurs in moistmeadows and floodplains (Behler and King 1997:468; Ernst and Barbour 1972:88). The carapace andplastron of box turtles are made up of individual bones that fuse together as the turtle ages. Thecarapace specimen present in the Kolomoki assemblage is fused, indicating that it was from an olderindividual. Turtles represent only a small proportion of the total MNI and less than one percent of thetotal biomass represented in the assemblage (Table 6-3).

Table 6-1. Species List for Faunal Remains from Block D.Taxon Description NISP MNI Weight Biomass

N % g kg

Testudines Indeterminate turtle 4 0.6 0.022

Terrapene carolina Eastern box turtle 1 1 9.1 0.9 0.029

Aves Indeterminate bird 9 0.8 0.017

Meleagris gallopavo Wild turkey 3 1 9.1 0.6 0.013

Mammalia Indeterminate mammal 1435 749.7 10.172

Ursus americanus Black bear 1 1 9.1 0.1 0.003

Odocoileus virginianus White-tailed deer 142 10 90.9 580.8 8.084

Vertebrata Indeterminate vertebrate 167.6

TOTAL 1595 13 1501.1

141

Table 6-2. Proveniences in Block D with Faunal Remains and the Taxa Represented. Provenience Recovery Strategy Taxon NISP Weight (g)

XUD2S, L1 >1/4 inch Mammalia 0.28

XUD2N, L3 >1/4 inch Mammalia 2 0.22

Vertebrata 5 0.04

XUD2S, L3 >1/4 inch Mammalia 4 1.37

XUD2S, L4 >1/4 inch

Mammalia 1 0.18

Mammalia 5.74

Vertebrata 1.14

XUD2 Disturbed Area >1/4 inchMammalia 0.96

Vertebrata 0.25

XUD4E, L4 >1/4 inchMammalia 0.16

Vertebrata 0.23

XUD9S, L3 >1/4 inchMammalia 17 3.83

Vertebrata 31 0.34


XUD12E, L2 >1/4 inch Mammalia 0.46

XUD12E, L3 >1/4 inchVertebrata 0.01

Mammalia 0.45

XUD12W, L3 >1/4 inch Mammalia 1.21

XUD13E, L1 >1/4 inch Mammalia 0.14

XUD13E, L2 >1/4 inchMammalia 0.22

Vertebrata 0.02

XUD13W, L2 >1/4 inch Mammalia 0.52

XUD15S, L1 >1/4 inch Vertebrata 0.07


XUD19N, L2 >1/4 inch Mammalia 0.52

XUD20W, L3 >1/4 inch Vertebrata 1 0.21

XUD21S, L3 >1/4 inch Mammalia 9 4.25

Vertebrata 15 0.3

XUD21N, L3 >1/4 inch Odocoileus virginianus 1 0.27

XUD24N, L3 >1/4 inch Mammalia 1 0.56

XUD21, Balk removal >1/4 inchMammalia 28 40.47

Vertebrata 36 0.64

F141A >1/4 inch

Aves 0.2

Vertebrata 4 13.94

Mammalia 45.71

Odocoileus virginianus 9.58

Meleagris gallopavo 0.36

F141B

>1/4 inch

Mammalia 15 58.18

Vertebrata 5 9.45



Heavy fraction, >1/4 inchMammalia 8 6.04

Vertebrata 22 0.84

F141C>1/4 inch

Mammalia 5 31.76


Vertebrata 3 6.11

Heavy fraction, >1/4 inch Vertebrata 8 0.18

142


F146 >1/4 inch


Vertebrata 1.41

Mammalia 7.03

F147A >1/4 inch


Mammalia 0.74

Vertebrata 0.42

F147B>1/4 inch

Terrapene carolina 0.85

Mammalia 19.65

Vertebrata 5.96



F149 >1/4 inch Mammalia 0.08

F150

>1/4 inch Vertebrata 4 0.2


Vertebrata 4 0.09

F153 >1/4 inchVertebrata 6 0.98

Mammalia 4 1.49

F154

>1/4 inch


Vertebrata 1.20

Mammalia 9.15


Testudines 1 0.09

Vertebrata 2 0.07

F155>1/4 inch

Testudines 0.51

Vertebrata 32.08

Mammalia 159.78


Aves 0.1

Ursus americanus 0.13


Vertebrata 2 0.02

F156>1/4 inch

Mammalia 1.3

Vertebrata 0.22

Heavy fraction, >1/4 inch Mammalia 4 0.89

F157A >1/4 inch


Vertebrata 1.39

Mammalia 4.54

F157B>1/4 inch

Aves 0.18


Vertebrata 1 3.89

Mammalia 5.63


F158 >1/4 inchMammalia 5.82

Vertebrata 0.55

F159 >1/4 inch


Vertebrata 5.88

Mammalia 10.86

F160 >1/4 inch Vertebrata 0.06

F163>1/4 inch

Vertebrata 1.32

Mammalia 11.07


143


F165>1/4 inch

Vertebrata 1.31

Mammalia 0.49




Vertebrata 0.22


F170A >1/4 inchMammalia 1.47

Vertebrata 0.38

F170B >1/4 inchVertebrata 0.64

Mammalia 0.2

F171 >1/4 inch

Aves 0.16


Mammalia 44.39

Vertebrata 15.32


F175

>1/4 inch


Vertebrata 23.89

Mammalia 101.83


Vertebrata 8 0.48

F176

>1/4 inch

Vertebrata 9.2

Mammalia 37.55



Vertebrata 3 0.1

F180 >1/4 inch Mammalia 1.92

F185 >1/4 inch

Vertebrata 7 4.92

Mammalia 11.73



Vertebrata 0.2

F191A >1/4 inchOdocoileus virginianus 2 15.42

Mammalia 21 7.15

Vertebrata 141 5.58

F191B >1/4 inchMammalia 3 0.19

Vertebrata 11 0.13

F192 >1/4 inch

Odocoileus virginianus 3 38.56

Mammalia 45 25.48

Vertebrata 153 2.75

F193 >1/4 inch Aves 1 0.16

F194 >1/4 inch


Mammalia 33 21.59

Vertebrata 134 4.37

F198 >1/4 inch Mammalia 1 0.7

F199 >1/4 inch Vertebrata 7 0.71


F203 >1/4 inch


Mammalia 37 16.6

Vertebrata 104 2.7


F207 >1/4 inchMammalia 2 0.62

Vertebrata 13 1.04

144

Table 6-3. Summary of Taxa Identified in the Analysis of Faunal Remains from Block D.MNI Biomass

N % kg %

Turtles 1 7.7 0.029 0.4

Birds 1 7.7 0.013 0.2

White-tailed deer 10 76.9 8.084 99.4

Other mammals 1 7.7 0.003 <0.1

TOTAL 13 8.129

Of the 12 bird specimens identified, three are identified as wild turkey (Meleagris gallopavo). Thewild turkey is common to the pine-oak forests of the Coastal Plain and would have been readily availableto the inhabitants of Kolomoki (Bull and Farrand 1996:451). Birds however, do not contributesignificant proportions of the total MNI or biomass to this assemblage.

Mammalian taxa identified include the black bear (Ursus americanus) and the white-tailed deer. Black bear is represented by a single upper fourth premolar. The lack of wear on this tooth indicatesthe bear was relatively young at the time of death. In the eastern United States the black bear is mostfrequently found in forested areas and swamps, particularly areas with dense cover (Whitaker 1997:704). Black bear is the only mammal included in the 13 Other Mammals category which contributes oneindividual to the total MNI and less than one percent of the total biomass.

White-tailed deer is by far the most abundant species present in the assemblage, accounting forover 99 percent of the total estimated biomass and 77 percent of the individuals represented. An MNIof 10 is estimated for white-tailed deer based on the presence of 10 distal right humeri. All 10 of thesehumeri are fused indicating all of the individuals were at least subadults or adults at the time of death(Table 6-4). Two of these individuals can be classified as adults based on the presence of several LateFusing elements with fused epiphyses. The presence of an unfused distal tibia signifies that at least oneindividual was a subadult or juvenile at the time of death. With only two unfused white-tailed deerbones recorded, the Block D assemblage seems to point to a relatively old deer population.

Elements represented for white-tailed deer are presented in Table 6-5. When compared to thestandard white-tailed deer skeleton (see Compton 2009:Appendix C), specimens of the Forequarter andHindquarter are over-represented, while elements of the Head and Hindfoot are only slightlyover-represented. Specimens of the Vertebra/Rib, Forefoot, and Foot categories are under-represented.

A total of 886 modifications are present in the Block D assemblage (Table 6-6). Only two typesof modifications are recorded: burning and calcination. The majority of burned and calcined specimensare identified as indeterminate mammal (Mammalia) or indeterminate vertebrate (Vertebrata). Twenty-six percent of indeterminate mammal specimens exhibit burning or calcination. This highpercentage suggests burning of refuse took place in the vicinity of Block D

As discussed previously a number of factors may bias the Block D faunal assemblage andconsequently the interpretation of Late Woodland animal use practices at Kolomoki. Of particularconcern is the role differential preservation played in altering the original skeletal assemblage. Block Dspecimens are generally in a poor state of preservation and are highly fragmented. The fragmentarynature of the assemblage is illustrated by the high proportion of specimens that could only be identifiedas indeterminate mammal or indeterminate vertebrate. The poor condition of the specimens also likelyaccounts for the lack of modifications identified other than burning and calcination. Modifications suchas butchery marks and animal gnawing occur on the outer surface of bone and can be difficult to

145

Table 6-4. Epiphyseal Fusion Noted on White-tailed Deer Remains from Block D.Unfused Fused

Early Fusing Metapodia, proximal (prenatal)

Cervical, Thoracic, Lumbar vertebra, centrum body

Radius, proximal

Innominate, acetabulum 3

2nd phalanx, proximal

Humerus, distal 13

Atlas, dorsal

1st phalanx, proximal 2

Scapula, distal

Middle Fusing Tibia, distal 1 7

Calcaneus

Metapodia, distal 1

Late Fusing Sacrum

Ulna, proximal

Femur, proximal 1

Radius, distal 2

Femur, distal 4

Tibia, proximal 1 2

Lumbar veterbra, centrum epiphysis

Ulna, distal

Cervical vertebra, centrum epiphysis

Pubic symphysis

Thoracic vertebra, centrum epiphysis

Humerus, proximal 2

TOTAL 2 37

Note: Elements are presented in general order of age at fusion following Purdue (1983). Elements have overlapping ranges of age at fusion and the orderof elements should not be considered the exact sequence of fusion.

Table 6-5. Element Distribution for White-tailed Deer in the Block D Faunal Assemblage.Block D Standard Deer

NISP % NISP %

Head 50 35.2 45 18.6

Vertebra/Rib 2 1.4 74 30.6

Forequarter 37 26.1 8 3.3

Hindquarter 37 26.1 9 3.7

Forefoot 1 0.7 16 6.6

Hindfoot 10 7.0 14 5.8

Foot 5 3.5 76 31.4

TOTAL 142 242

146

Table 6-6. Modifications of Bone Noted in the Block D Faunal Assemblage.Burned Calcined

Indeterminate turtle 1

Indeterminate bird 1

Wild turkey 2

Indeterminate mammal 70 302

White-tailed deer 3

Indeterminate vertebrate 132 375

TOTAL 205 681

identify if the bone surface is in poor condition. In addition, differential preservation may account forthe dominance of white-tailed deer in the assemblage.

The small size of the Block D assemblage is also a source of potential bias. Although the sampleconsists of 1,595 specimens, only 147 of these specimens are identified beyond taxonomic class and/ororder. The small size of the assemblage and the effects of differential preservation make the Block Dassemblage an unlikely candidate for accurate interpretations of Late Woodland animal use practices atKolomoki. However, when these data are compared to the previously studied early/middle Late Woodland Block A assemblage (Pluckhahn et al. 2006) some similarities are observed, as described inthe chapter that follows.

Macrobotanical Remains

Analysis of macrobotanical remains focused on 51 light and 51 heavy fraction flotation samples,as well as one fine water-screened sample (Table 6-7). These samples come from 35 distinct featuresin Block D.

Seed preservation is poor in these samples. This is the usual pattern for Coastal Plain sites inGeorgia, and Branch-Raymer and Bonhage-Freund (2011) have recently demonstrated that this is at leastin part due to soil acidity. Nutshell—especially hickory—and wood charcoal and resin tend to berelatively abundant in these samples, probably due to their more robust constitution.

Due to the low density of macroplant remains, in this analysis ubiquity is emphasized overcounts and weight for nutshell, and mere presence/absence is noted for seeds. Ubiquity measures theproportion of unique features that contain macroplant remains of any given taxon. The assumption isthat higher the ubiquity the greater the initial abundance of the taxon because it is present in morelocations.

Results

Seeds. For convenience, seeds, fruits, and seed coats are all subsumed under the title of "seeds"in this analysis. Only 43 seeds and seed fragments, or 0.146 seeds per liter of floated soil, wererecovered (Table 6-8). The 43 seeds in this archaeobotanical assemblage are scattered across the featuretypes. Without regard to taxa, seed ubiquity is 50 percent for large pits, 42.5 percent for small pits, and28.6 percent for post molds. A single large pit (Feature 146) yielded 32.6 percent of the total seedassemblage including 9 grass seeds, 1 grape family fruit, and 4 unidentifiable seed fragments.

147

Table 6-7. Summary Data for Flotation Samples and Fractions from Block D.Sample Feature Description Sample

Volume (l)Weight of Light

Fraction (g)Weight of Heavy

Fraction (kg) Comments

1 141B South 1/2 12.00 34.61 18.52 141B Near base (100 cmbd) 10.00 57.15 24.093 144 North 1/2 1.25 4.32 5.054 141 Southwest 1/2 10 34.33 17.09 Sample taken before features divided5 143 3.5 7.89 13.466 140 1.75 25.55 6.177 141A 9.50 72.28 2.3428 185 Near base, South 1/2 9.50 41.7 33.529 146 North 1/2 7.50 35.27 2.73510 147B West 1/2 1.25 24.76 50.0611 147B West 1/2 6.30 46.33 1.92312 147B Near base 5.70 19.48 42.2813 150 North 1/2 5.30 9.78 39.114 151 North 1/2 5.00 28.77 63.9415 153 North 1/2 8.80 22.75 46.716 154 Near top, East 1/2 21.45 50.02 volume not recorded17 154 Near base, East 1/2 6.90 18.43 42.8618 155 6.50 45.83 2.34319 155 Northeast 1/2 6.00 11.76 36.9520 155 South 1/2 6.00 13.18 41.7821 156 North 1/2 8.30 17.43 53.6522 157 8.20 14.57 42.34 Sample taken before features divided23 157B 6.30 15.93 32.7524 158 West 1/2 7.80 11.85 27.5925 159 4.90 10.77 24.5726 160 North 1/2 3.20 1.67 11.8727 162 North 1/2 7.20 10.4 34.5128 163 North 1/2 6.00 20.95 29.5929 164 3.70 11.94 18.2430 165 Near top, South 1/2 6.40 16.38 35.7431 165 Near base, North 1/2 7.60 23.11 77.2832 166 North 1/2 2.30 4.1 10.9133 167 East 1/2 7.00 8.97 33.1634 168 East 1/2 1.40 0.88 4.735 169 Near top, South 1/2 5.50 5.15 4336 169 Near base, South 1/2 8.00 6.1 48.8437 170A West 1/2 5.70 11.05 44.6938 170B West 1/2 5.92 6.92 32.3939 171 Northwest 1/2 6.90 26.04 2146.1940 171 Southeast 1/2 6.00 9.31 46.8141 175 Near top, South 1/2 6.70 13.06 30.6742 175 South 1/2 4.80 34.36 39.943 176 Near top, South 1/2 5.00 6.23 25.6344 176 Near base 4.40 36.2 34.9645 177 East 1/2 4.80 28.02 24.7746 179 4.20 2.54 14.9147 180 South 1/2 4.20 4.42 2748 180 North 1/2 3.60 4.74 18.2249 185 West 1/2 5.40 8.28 35.8850 185 South 1/2 2.90 4.14 26.4

none 147B 6.50 253.47 107.08 hand floated

none 181 1.50 848.26 dry sieved, water screened through125 micron sieve

148

Table 6-8. Seeds (Counts) in Macrobotanical Remains from Block D.Sa

mpl

e

Fea

ture

Fra

ctio

n

Scan

ned

only

domesticates starchy/herbaceous wild fleshy fruits medicinal unknown or unidentifiable

Che

nopo

dium

sp.

see

d co

at(t

runc

ated

) (go

osef

oot)

Phas

eolu

s vu

lgaris

or

polys

tayc

hus

(par

tial c

otyl

edon

) (c

omm

onbe

an o

r "w

ild b

ean"

)

c.f.

Bar

bare

a sp

. (ye

llow

rock

et) (

ct0

Che

nopo

dium

sp.

(wild

type

)

Poae

ae c

t

Am

pelo

psis

sp. (

pepp

ervi

ne)

Ilex

sp. (

holly

)

Nys

sa sy

lvat

ica a

nd c

.f. N

yssa

sylv

atica

seed

(bla

ckgu

m)

c,f,

Vita

ceae

fru

it (g

rape

fam

ily)

unid

entif

ied

seed

, c.f.

Bra

ssica

sp.

Aca

lypha

c.f.

virg

inia

na (4

-see

ded

mer

cury)

c.f.

Cam

panu

la s

p. (b

ellfl

ower

)

Leg

umin

osae

Leg

umin

osae

, rou

nd c

otyl

ydon

unid

entif

ied

seed

(no

dist

ingu

ishi

ng c

hara

cter

istic

s)

unid

entif

ied

seed

fra

g

test

a fr

agm

ents

Unk

now

n s

eed

(hol

low

)

1 141Blight

heavy

2 141Blight

heavy

3 144light

heavy

4 141Clight

heavy

5 143light

heavy

6 140light

heavy

7 141Alight 1

heavy

8 185Wlight 1

heavy

9 146light 9 3

heavy 1 1

10 147Blight

heavy

11 147Blight 4

heavy

12 147Blight

heavy

13 150light

heavy 1 2

14 151light

heavy

15 153light

heavy

16 154light

heavy x

17 154light

heavy x

149


mpl

e

Fea

ture

Fra

ctio

n

Scan

ned

only


Che

nopo

dium

sp.

see

d co

at(t

runc

ated

) (go

osef

oot)

Phas

eolu

s vu

lgaris

or

polys

tayc

hus

(par

tial c

otyl

edon

) (c

omm

onbe

an o

r "w

ild b

ean"

)

c.f.

Bar

bare

a sp

. (ye

llow

rock

et) (

ct0

Che

nopo

dium

sp.

(wild

type

)

Poae

ae c

t

Am

pelo

psis

sp. (

pepp

ervi

ne)

Ilex

sp. (

holly

)

Nys

sa sy

lvat

ica a

nd c

.f. N

yssa

sylv

atica

seed

(bla

ckgu

m)

c,f,

Vita

ceae

fru

it (g

rape

fam

ily)

unid

entif

ied

seed

, c.f.

Bra

ssica

sp.

Aca

lypha

c.f.

virg

inia

na (4

-see

ded

mer

cury)

c.f.

Cam

panu

la s

p. (b

ellfl

ower

)

Leg

umin

osae

Leg

umin

osae

, rou

nd c

otyl

ydon

unid

entif

ied

seed

(no

dist

ingu

ishi

ng c

hara

cter

istic

s)

unid

entif

ied

seed

fra

g

test

a fr

agm

ents

Unk

now

n s

eed

(hol

low

)

18 155light

heavy 1

19 155light

heavy x

20 155light

heavy x

21 156light

heavy

22 157light

heavy X

23 157Blight 1

heavy x

24 158light

heavy

25 159light

heavy

26 160light

heavy

27 162light

heavy

28 163light 1

heavy 1 1

29 164light

heavy

30 165light 1

heavy x

31 165light

heavy x

32 166light

heavy

33 167light

heavy x

34 168light

heavy

150


mpl

e

Fea

ture

Fra

ctio

n

Scan

ned

only


Che

nopo

dium

sp.

see

d co

at(t

runc

ated

) (go

osef

oot)

Phas

eolu

s vu

lgaris

or

polys

tayc

hus

(par

tial c

otyl

edon

) (c

omm

onbe

an o

r "w

ild b

ean"

)

c.f.

Bar

bare

a sp

. (ye

llow

rock

et) (

ct0

Che

nopo

dium

sp.

(wild

type

)

Poae

ae c

t

Am

pelo

psis

sp. (

pepp

ervi

ne)

Ilex

sp. (

holly

)

Nys

sa sy

lvat

ica a

nd c

.f. N

yssa

sylv

atica

seed

(bla

ckgu

m)

c,f,

Vita

ceae

fru

it (g

rape

fam

ily)

unid

entif

ied

seed

, c.f.

Bra

ssica

sp.

Aca

lypha

c.f.

virg

inia

na (4

-see

ded

mer

cury)

c.f.

Cam

panu

la s

p. (b

ellfl

ower

)

Leg

umin

osae

Leg

umin

osae

, rou

nd c

otyl

ydon

unid

entif

ied

seed

(no

dist

ingu

ishi

ng c

hara

cter

istic

s)

unid

entif

ied

seed

fra

g

test

a fr

agm

ents

Unk

now

n s

eed

(hol

low

)

35 169light

heavy x

36 169light

heavy x

37 170Alight

heavy x

38 170Blight

heavy x

39 171light 1 1

heavy 1 1

40 171light 1

heavy x

41 175light 1

heavy x

42 175light 1

heavy x

43 176light

heavy

44 176light

heavy x

45 177light

heavy

46 179light

heavy

47 180light

heavy

48 183light

heavy

49 185light

heavy x

50 185light

heavy x

147Blight 1 4 2

heavy x

181lightand

heavyTOTAL COUNT 1 1 1 1 9 1 1 3 1 1 2 1 4 2 5 2 6 1

151

Domesticates are conspicuously rare in the Block D macrobotanical assemblage. One each ofa truncated goosefoot seed (Chenopodium sp.) and a partial bean cotyledon (Phaseolus c.f. vulgaris) are theonly potential garden species recovered. The latter was recovered from Feature 171, the possible hearthlocated near the center of the structure. Common bean is not known in this geographic area until atleast the late prehistoric period. Since this cotyledon is incomplete and small in size, there is a possibilitythat it is actually Phaseolus polystachyus, a wild bean, possibly either deliberately planted or "encouraged"in some areas of Georgia (Bonhage-Freund 1997). An alternative explanation for the presence of thisspecimen is that it is an intrusion from a Lamar component located primarily downslope from BlockD (Pluckhahn 2003:74-77; Sears 1951), although this seems unlikely given the complete absence ofLamar pottery from features in Block D in general, and Feature 171 in particular.

The goosefoot seed was recovered from Feature 147B, one of the large pits northwest of thestructure. Goosefoot was a common Woodland period domesticate throughout the Midwest andSoutheast. While the specimen recovered from Block D features the truncated margin typical of adomesticated variety, it consists only of half of a seedcoat (testa). The only way to be certain that agoosefoot seed is domesticated is to measure the thickness of the seedcoat, using a scanning electronmicroscope.

Naturally occurring seeds can be divided into four categories: starchy and herbaceous, fleshyfruit, medicinal, and not identifiable. One goosefoot (Chenopodium sp.) and nine unidentified grass seeds(Poaceae) comprise the naturally occurring starchy seeds. Non-domesticated goosefoot seeds aregenerally smaller than the domesticated varieties, with a low food to seed coat ratio. However, naturallyoccurring goosefoot is also used as a "potherb" or "wild spinach." As a "green" it would have requiredvery little effort to harvest and prepare, and while low in calories, would provide iron and otherimportant nutrients. This is a pioneering species usually found near human habitation sites, even today,and particularly attracted to nitrogen-rich garbage deposits. In addition, traditional farmers often uprootthese plants, tossing them to the sides of garden plots where they flourish and serve as a no-maintenancecrop (Bonhage-Freund 1997). While the grass seeds have not yet been identified, all grasses are edibleand it is possible that these were harvested for food. There is an even more remote possibility that thesegrasses were cultivated or "encouraged" as a crop. Yellowrocket (Barbarea sp.) was used as a potherbby historic Cherokee and also raw to purify the blood (Moerman 1998: 121). Possible Brassica sp. alsoserved as pot herb and medicine (Moerman:128). These food and medicine traditions, and others tobe described below, may have been in place during the Late Woodland Period.

Evidence of fleshy fruits is scant, but all identified genera were exploited for food, medicine, orceremonial purposes in the historic Southeast. One peppervine (Ampelopsis sp.) and a single possiblefruit represent the grape family (Vitaceae, 5 percent ubiquity combined). Peppervine is a favorite foodof birds, which in turn distribute its seeds. While this single specimen has been identified only to thegenus, the bark of Ampelopsis cordata (heartleaf peppervine) was used by historic Cherokee to treat urinarytroubles (Moerman 1998:70). Across North America, various species of grape (Vitis spp.) wereharvested by Indians from the wild and some species were cultivated for food and medicine duringhistoric times (Moerman 1998: 98-600). Three blackgum or probable blackgum (c.f. Nyssa sylvatica, 5percent ubiquity) seeds were recovered. While these, too, are dispersed by birds, they are rich innutrients and could have been incorporated into the diet. In historic times, Cherokee and Creek usedthe bark or roots in decoctions and teas to treat a variety of digestive tract ailments, for tuberculosis,and as eye drops (Moerman 1998:360). One holly seed (Ilex sp., 2.5 percent ubiquity) was recoveredfrom Feature 163. Throughout the Southeast the bark of various hollies was used as an emetic and mostfamously as a major ingredient of the ritual purgative "black drink." Holly bark and root decoctionsserved less well-known medicinal functions such as the treating of hay fever, eye problems, sore musclesand nightmares (Moerman 1998: 273). All of these taxa were most likely uncultivated plants, yet humandisturbance of the landscape initiated the succession cycles that allowed for their growth.

152

In addition to medicinal taxa already cited, two three-seeded mercury seeds (Acalypha c.f. virginiana) and one bellflower seed (Campanula sp.) represent naturally occurring species of somemedicinal value. Historic Cherokee used three-seeded mercury root to treat kidney problems, while aninfusion of bellflower was taken for diarrhea (Moerman 1998: 37, 135).

A single small round legume is probably a non-native clover, and an incidental inclusion, whilethe remaining Leguminosae were not able to be identified. A total of 13 seeds lacked necessarydistinguishing characteristics or were too fragmented or incomplete to identify. One partial hollow seed also could not be identified, despite extensive research.

Nutshell. Hickory (Carya spp.) nutshell is present in 100 percent of features (Table 6-9). Acorn(Quercus spp.) shell is found in 20 features (50 percent ubiquity). Differential preservation most likelyaccounts for much of his discrepancy and acorn is generally grossly under-represented in thearchaeological record (Lopinot 1983). While not every heavy fraction was fully sorted at or above the2.00 mm fraction, it was clear that the bulk of hickory shell appeared in the heavy fractions, althoughthe light fractions yielded most of the wood charcoal.

In order to reduce distorting effects of both original bulk of hickory versus acorn nutshellrelative to the nutmeat itself, and the differential preservation, it is best to assess the importance of nutsby count, rather than weight. This does not eliminate the distortion but does mitigate it to some extent.These data should be further refined by multiplying the recovered acorn shell by a factor of 50 toachieve a more accurate acorn shell: hickory shell ratio (Yarnell and Black 1983). A ratio of acorn shellcount (x 50) to hickory shell count based on all features for which both light and heavy fractions weresorted and not merely scanned, is 2.93. This indicates that while hickory shell is more abundant in themacrobotanical assemblage, acorn was probably the more important resource. Furthermore, Lopinot(1983) suggests that, depending on the species under consideration, acorns yield between 5 and 200times more "nutmeat" than hickory, relative to the amount of nutshell.

Calculation of the ratio of nutshell to remnant fuel wood is used to compare the relativeabundance of nutshell in different contexts across a site where it is fairly certain of the origin of woodcharcoal (Miller 1988). Applying this tool to those cases in which both the heavy and light fractionsof all given feature samples were fully sorted to the 2.00 mm level it is observed that in nine large pits(Features 141A, 141B, 141C, 146, 156, 158, 159, 171, and 185) the ratio of the number of hickoryfragments to wood charcoal weight (hickory [N]: wood weight [g]) is 10.75, while the ratio of acorn shellfragments to wood charcoal weight (acorn [N x 50]: wood weight [g]) is 38.39. Three fully sortedsamples from small pits (Features 140, 153, and 163) yielded a ratio of 3.2 hickory count to woodcharcoal weight (g) and 2.24 ratio of acorn shell count to wood charcoal weight (g). Since the wood ofpost molds is presumed to be composed in part of construction wood, this ratio cannot be used in thatcontext.

As a point of comparison, if these same ratios are calculated based on all fully sorted samples,without considering if both heavy and light fractions were fully sorted, there are somewhat differentresults. In this case, the ratios of hickory to wood charcoal is 5.90 for large pits and 2.48 for small pits. The acorn to wood charcoal ratio is 28.6 for large pits and 12.22 for small pits. The difference betweenthese ratios is clearly related to the fact that hickory is more often recovered in the heavy fraction, whileacorn shell is more common in the light fractions. When the heavy fractions are not all fully analyzed,much of the hickory is precluded from the ratio, and in some cases acorn increases in prominence.

153

Table 6-9. Nutshell/Nutmeat (Counts and Weights) in Macrobotanical Remains from Block D.Sa

mpl

e

Fea

ture

Fra

ctio

n

Scan

ned

Onl

y

Car

ya s

p. (N

) - h

icko

ry

Car

ya s

p (g

)

c.f.

Jugla

ns n

igra

(N) -

bla

ck w

alnu

t

c.f.

Jugla

ns n

igra

(g)

Jugl

anda

ceae

(N) w

alnu

t fam

ily

Jugl

anda

ceae

(g)

c.f.

Que

rcus

sp

(oak

) aco

rn n

ut m

eat(

N)

c.f.

Que

rcus

sp

(oak

) aco

rn n

ut m

eat(

g)

Que

rcus

sp

( oak

) aco

rn s

hell

(N)

Que

rcus

sp

( oak

) aco

rn s

hell

(g)

c.f.

Que

rcus

sp

( oak

) aco

rn s

hell

(N)

c.f.

Que

rcus

sp

( oak

) aco

rn s

hell

(g)

Uni

dent

ified

nut

mea

t (N

)

Uni

dent

ified

nut

mea

t (g)

1 141Blight

heavy 13 0.31

2 141Blight 2 0.02 1 0.01

heavy 30 0.51

3 144light 6 0.08

heavy 7 0.06

4 141Clight 1 0.01

heavy 8 0.16

5 143light 1 0.01

heavy 2 0.05

6 140light

heavy 1 0.01

7 141Alight 3 0.05

heavy 202 4.91 4 0.05

8 185light 1 0.03

heavy 24 0.18

9 146light 2 0.05

heavy 97 1.77 8 0.06

10 147Blight 6 0.12 2 0.01

heavy 135 2.02 1 0.01

11 147Blight 2 0.07

heavy 240 7.38 8 0.06

12 147Blight 1 0.02

heavy 163 5.02 8 0.04

13 150light

heavy 74 1.48 1 0.00

14 151light

heavy 32 0.50

15 153light

heavy 35 0.81 1 0.10

16 154light

heavy x 15 0.08

17 154light 3 0.11 2 0.02

heavy x

18 155light 8 0.21 11 0.05

heavy 137 1.67 1 0.00

154


mpl

e

Fea

ture

Fra

ctio

n

Scan

ned

Onl

y

Car

ya s

p. (N

) - h

icko

ry

Car

ya s

p (g

)

c.f.

Jugla

ns n

igra

(N) -

bla

ck w

alnu

t

c.f.

Jugla

ns n

igra

(g)

Jugl

anda

ceae

(N) w

alnu

t fam

ily

Jugl

anda

ceae

(g)

c.f.

Que

rcus

sp

(oak

) aco

rn n

ut m

eat(

N)

c.f.

Que

rcus

sp

(oak

) aco

rn n

ut m

eat(

g)

Que

rcus

sp

( oak

) aco

rn s

hell

(N)

Que

rcus

sp

( oak

) aco

rn s

hell

(g)

c.f.

Que

rcus

sp

( oak

) aco

rn s

hell

(N)

c.f.

Que

rcus

sp

( oak

) aco

rn s

hell

(g)

Uni

dent

ified

nut

mea

t (N

)

Uni

dent

ified

nut

mea

t (g)

19 155light 2 0.02

heavy x

20 155light

heavy x 4 0.02

21 156light 1 0.06

heavy 22 0.25 2 0.06

22 157light

heavy X

23 157Blight 2 0.08

heavy x

24 158light 1 0.01

heavy 23 0.54

25 159light

heavy 27 0.41

26 160light

heavy 11 0.12

27 162light 2 0.01

heavy 44 0.65

28 163light

heavy 36 0.48

29 164light 2 0.01 1 0.00

heavy 11 0.15

30 165light 1 0.01

heavy x 1 0.01

31 165light

heavy x 7 0.05

32 166light

heavy 6 0.12 2 0.12

33 167light 1 0.01

heavy x

34 168light

heavy 19 0.37

35 169light

heavy x 1 0.01

36 169light

heavy x

155


mpl

e

Fea

ture

Fra

ctio

n

Scan

ned

Onl

y

Car

ya s

p. (N

) - h

icko

ry

Car

ya s

p (g

)

c.f.

Jugla

ns n

igra

(N) -

bla

ck w

alnu

t

c.f.

Jugla

ns n

igra

(g)

Jugl

anda

ceae

(N) w

alnu

t fam

ily

Jugl

anda

ceae

(g)

c.f.

Que

rcus

sp

(oak

) aco

rn n

ut m

eat(

N)

c.f.

Que

rcus

sp

(oak

) aco

rn n

ut m

eat(

g)

Que

rcus

sp

( oak

) aco

rn s

hell

(N)

Que

rcus

sp

( oak

) aco

rn s

hell

(g)

c.f.

Que

rcus

sp

( oak

) aco

rn s

hell

(N)

c.f.

Que

rcus

sp

( oak

) aco

rn s

hell

(g)

Uni

dent

ified

nut

mea

t (N

)

Uni

dent

ified

nut

mea

t (g)

37 170Alight

heavy x

38 170Blight 2 0.02

heavy x 1 0.00

39 171light 16 0.16

heavy 548 27.42 35 0.27

40 171light 5 0.15 1 0.01

heavy x 4 0.04

41 175light 3 0.05 1 0.10 1 0.00

heavy x 2 0.01

42 175light 2 0.03

heavy x 3 0.02

43 176light

heavy

44 176light 3 0.21

heavy x 4 0.02

45 177light

heavy 9 0.12

46 179light

heavy 4 0.08

47 180light 1

heavy 41 0.48

48 183light

heavy 35 0.78

49 185light 1 0.03

heavy x 5 0.07

50 185light

heavy x 31 0.14

147Blight 66 1.30 2 0.01

heavy x

181lightand

heavyTOTAL COUNT 2150 2 6 4 162 29 1

TOTAL WEIGHT 61.47 0.01 0.08 0.16 1.25 0.21 0.01

156

Wood Charcoal. Wood charcoal generally comprises the bulk of any light fraction flotationsample from open sites, and this was true in all of the Kolomoki samples. Every feature containedwood charcoal as well as the exploded resins generally associated with burned pine.

An attempt was made to identify 1904 wood charcoal fragments, including two fragments ofbark (one pine, one indistinguishable), two of unidentifiable root and one completely degraded andunidentifiable fragment. The remaining 1900 fragments were at identified to some level and arecomprised of 19 percent hardwood (N=357), 75 percent softwood (N=1432), and 6 percent monocot(N=111). These are very similar to the proportions observed in the earlier Late Woodland habitationarea in Block A (19 percent hardwood, 67 percent softwood, and 9 percent monocot) (Bonhage-Freund2003).

As in most Southeastern sites, pine (Pinus spp.) dominates the wood charcoal assemblage. Anaverage of 99 percent of all softwood was pine or pine bark with the remainder consisting of easternredcedar (Juniperus virginiana, N=12) and 2 fragments identifiable only as conifer. Similarly, in the BlockA assemblage, pine comprised 98 percent of the softwood (Bonhage-Freund 2003).

Several combined oak species comprise 58 percent by count of the 227 hardwood fragments thatwere identifiable to at least the level of genus. This is in contrast to earlier Block A assemblage, whereno individual hardwood count was significantly higher than any other (Bonhage-Freund 2003). Pluckhahn et al. (2006) suggest that some of the diversity of charcoal in Block A could be attributableto the unusual nature of the deposit as refuse associated with feasting. This difference could also beaccounted for by natural maturing of the local forest where oak, hickory, and pine are expected todominate (Wharton 1978).

In the Block D assemblage, the proportions of hardwood, softwood, and monocots (cane andother grasses) did not vary much between large and small pit features, and in every feature classsoftwoods dominate hardwoods. This pattern is not unusual for sites in the region. Among the reasonsfor a consistently lower proportion of hardwood to softwood in southern assemblages is the fact thathardwood is more likely to burn to ash with little residual, compared to pine (Bonhage-Freund 2005). In the case of post molds in Block D, there is a somewhat greater proportion of hardwood tosoftwood. At least some post molds were likely associated with structures and thus higher proportionof hardwood which might well represent floor or plaza sweepings containing remnants of thoroughlyburned fireplace residue, further damaged by abrasion. In other situations the "post molds" mightactually be small fire-associated features. However, none of the post molds or small pits displayed thecharacteristics of a smudge pit (Binford 1967; Bonhage-Freund 2005).

Nineteen distinct hardwood genera were identified in these samples including maple, Americanchestnut, hickory, flowering dogwood, honey locust, butternut, black walnut, probable blackgum,ironwood, sycamore, cottonwood, wild cherry, white oak, red oak, post oak, black locust, willow,basswood, and elm. In addition, some specimens were assigned to the generic walnut, oak, and elm/hackberry classifications. All types of features demonstrated comparable species breadth. Large pits,small pits, and post molds yielded 14, 15, and 12 different hardwood species, respectively. Cane andgrasses (monocots) are widely distributed across Block D, but in small amounts. These monocots wouldbe locally abundant and served many purposes such as matting, cordage, roof and wall construction,deliberate creation of smoke, among others.

157

Table 6-10. Wood Charcoal and Miscellaneous (Counts) in Macrobotanical Remains from Block D.

Sam

ple

Fea

ture

Fra

ctio

n

Scan

ned

Onl

y

Ace

r sp.

(map

le)

Cas

tane

a de

ntat

a (A

mer

ican

che

stnu

t)

c.f.

Car

ya s

p. (h

icko

ry)

Car

ya s

p. (h

icko

ry)

c.f.

Cor

nus f

lorid

a (f

low

erin

g do

gwoo

d)

Gled

etsia

tria

ncan

thos

(hon

ey lo

cust

)

Jugla

ns ci

nere

a (b

utte

rnut

)

Jugla

ns n

igra

(bla

ck w

alnu

t)

c.f.

Jugla

ns s

p. (w

alnu

t fam

ily)

c.f.

Nys

sa sy

lvat

ica (b

lack

gum

)

c.f.

Nys

sa s

p. (b

lack

gum

)

c.f.

Ostr

ya v

irgin

iana

(iro

nwoo

d)

Plat

anus

occi

dent

alis

(syc

amor

e)

c.f.

Plat

anus

occi

dent

alis

(syc

amor

e)

c.f.

Popu

lus d

eltoi

des (

cott

onw

ood)

c.f.

Prun

us s

p. (

wild

che

rry)

Que

rcus

alb

a. (w

hite

oak

)

c.f.

Que

rcus

alb

a. (w

hite

oak

)

Que

rcus

rubr

a (r

ed o

ak)

Que

rcus

c.f.

rubr

a (r

ed o

ak)

Que

rcus

stell

ata

(pos

t oak

)

c.f.

Que

rcus

sp.

(oak

)

cf. R

obin

ia p

seud

oaca

cia (b

lack

locu

st)

c.f.

Salix

sp.

(will

ow)

c.f.

Tili

a am

erica

na (b

assw

ood)

Ulm

us ru

bra

(elm

)

Ulm

us s

p./

Celt

is sp

. (el

m/h

ackb

erry

)

ring

poro

us

diff

use

poro

us

Har

dwoo

d (n

ot id

entif

iabl

e)

Har

dwoo

d ro

ot

Juni

peru

s virg

inia

na (e

aste

rn r

edce

dar)

Pinu

s sp.

(pin

e)

Pinu

s sp.

bar

k (p

ine)

Con

ifer

Aru

ndin

aria

sp.

(ca

ne)

mon

ocot

c.f.

Aru

ndin

aria

sp.

(ca

ne)

mon

ocot

c.f.

mon

ocot

bark

c.f.

root

Not

iden

tifia

ble

resi

n

sam

ple

type

1 141B light 9 2 9 1 6 pinch

heavy 9 12

2 141B light 2 1 2 4 43 4 pinch

heavy 1 19 12

3 144 light 3 1

heavy 2

4 141C light 1 3 15 1 2

heavy 1 4 2

5 143 light 9 8 2

heavy 8 2 2

6

140 light 1 10 1 17 odd #s

heavy 13 7 6

7 141A light 4 4 1 1 28 5

heavy 2 1 41 2 31

8 185W light 1 1 15 3 1 sweep

heavy 1 17 2 8 pinch

9 146 light 1 2 2 32 3 6

heavy 2 1 19 171

10 147B light 1 4 13 15

heavy 2 18 83 pinch

11 147B light 1 7 1 1 39 30 pinch


12 147B light 2 18 3

heavy 20 84 pinch

13 150 light 1 1 6 2 5 14 1 1 3



Sam

ple

Fea

ture

Fra

ctio

n

Scan

ned

Onl

y

Ace

r sp.

(map

le)

Cas

tane

a de

ntat

a (A

mer

ican

che

stnu

t)

c.f.

Car

ya s

p. (h

icko

ry)

Car

ya s

p. (h

icko

ry)

c.f.

Cor

nus f

lorid

a (f

low

erin

g do

gwoo

d)

Gled

etsia

tria

ncan

thos

(hon

ey lo

cust

)

Jugla

ns ci

nere

a (b

utte

rnut

)

Jugla

ns n

igra

(bla

ck w

alnu

t)

c.f.

Jugla

ns s

p. (w

alnu

t fam

ily)

c.f.

Nys

sa sy

lvat

ica (b

lack

gum

)

c.f.

Nys

sa s

p. (b

lack

gum

)

c.f.

Ostr

ya v

irgin

iana

(iro

nwoo

d)

Plat

anus

occi

dent

alis

(syc

amor

e)

c.f.

Plat

anus

occi

dent

alis

(syc

amor

e)

c.f.

Popu

lus d

eltoi

des (

cott

onw

ood)

c.f.

Prun

us s

p. (

wild

che

rry)

Que

rcus

alb

a. (w

hite

oak

)

c.f.

Que

rcus

alb

a. (w

hite

oak

)

Que

rcus

rubr

a (r

ed o

ak)

Que

rcus

c.f.

rubr

a (r

ed o

ak)

Que

rcus

stell

ata

(pos

t oak

)

c.f.

Que

rcus

sp.

(oak

)

cf. R

obin

ia p

seud

oaca

cia (b

lack

locu

st)

c.f.

Salix

sp.

(will

ow)

c.f.

Tili

a am

erica

na (b

assw

ood)

Ulm

us ru

bra

(elm

)

Ulm

us s

p./

Celt

is sp

. (el

m/h

ackb

erry

)

ring

poro

us

diff

use

poro

us

Har

dwoo

d (n

ot id

entif

iabl

e)

Har

dwoo

d ro

ot

Juni

peru

s virg

inia

na (e

aste

rn r

edce

dar)

Pinu

s sp.

(pin

e)

Pinu

s sp.

bar

k (p

ine)

Con

ifer

Aru

ndin

aria

sp.

(ca

ne)

mon

ocot

c.f.

Aru

ndin

aria

sp.

(ca

ne)

mon

ocot

c.f.

mon

ocot

bark

c.f.

root

Not

iden

tifia

ble

resi

n

sam

ple

type

14 151light 22 2 1

heavy 3 3 11 2 17

15 153light 3 1 10 23 4

heavy 20 73 pinch

16 154light 2 1 27 1 sweep

heavy x

17 154light 3 23 4 3

heavy x

18 155light 1 6 1 31 1 20 pinch

heavy 19 1 100 every 10th

19 155light 1 2 2 1 14 2 sweep

heavy x

20 155light 1 3 29 2 1 pinch

heavy x

21 156light 1 19 1

heavy 20 44

22 157light 3 15 3 sweep

heavy x

23 157Blight 5 15 pinch

heavy x

24 158light 2 16 2 5

heavy 1 19 15 sweep

25 159light 1 3 23 2 1 2 sweep

heavy 17 72

26 160light 1 1 2 2

heavy 1 20 3


Sam

ple

Fea

ture

Fra

ctio

n

Scan

ned

Onl

y

Ace

r sp.

(map

le)

Cas

tane

a de

ntat

a (A

mer

ican

che

stnu

t)

c.f.

Car

ya s

p. (h

icko

ry)

Car

ya s

p. (h

icko

ry)

c.f.

Cor

nus f

lorid

a (f

low

erin

g do

gwoo

d)

Gled

etsia

tria

ncan

thos

(hon

ey lo

cust

)

Jugla

ns ci

nere

a (b

utte

rnut

)

Jugla

ns n

igra

(bla

ck w

alnu

t)

c.f.

Jugla

ns s

p. (w

alnu

t fam

ily)

c.f.

Nys

sa sy

lvat

ica (b

lack

gum

)

c.f.

Nys

sa s

p. (b

lack

gum

)

c.f.

Ostr

ya v

irgin

iana

(iro

nwoo

d)

Plat

anus

occi

dent

alis

(syc

amor

e)

c.f.

Plat

anus

occi

dent

alis

(syc

amor

e)

c.f.

Popu

lus d

eltoi

des (

cott

onw

ood)

c.f.

Prun

us s

p. (

wild

che

rry)

Que

rcus

alb

a. (w

hite

oak

)

c.f.

Que

rcus

alb

a. (w

hite

oak

)

Que

rcus

rubr

a (r

ed o

ak)

Que

rcus

c.f.

rubr

a (r

ed o

ak)

Que

rcus

stell

ata

(pos

t oak

)

c.f.

Que

rcus

sp.

(oak

)

cf. R

obin

ia p

seud

oaca

cia (b

lack

locu

st)

c.f.

Salix

sp.

(will

ow)

c.f.

Tili

a am

erica

na (b

assw

ood)

Ulm

us ru

bra

(elm

)

Ulm

us s

p./

Celt

is sp

. (el

m/h

ackb

erry

)

ring

poro

us

diff

use

poro

us

Har

dwoo

d (n

ot id

entif

iabl

e)

Har

dwoo

d ro

ot

Juni

peru

s virg

inia

na (e

aste

rn r

edce

dar)

Pinu

s sp.

(pin

e)

Pinu

s sp.

bar

k (p

ine)

Con

ifer

Aru

ndin

aria

sp.

(ca

ne)

mon

ocot

c.f.

Aru

ndin

aria

sp.

(ca

ne)

mon

ocot

c.f.

mon

ocot

bark

c.f.

root

Not

iden

tifia

ble

resi

n

sam

ple

type

27 162light 2 2 9 3 15 0

heavy 4 1 1 23 20

28 163light 3 6 10 1 9 pinch

heavy 20 33 pinch

29 164light 2 1 5 9 13 2


30 165light 3 1 4 21

heavy x

31 165light 1 1 2 16 21

heavy x

32 166light 7 1 2 8 1 1 pinch

heavy x 20 12 every other

33 167light 18 1 6

heavy x

34 168light 2 1 2

heavy 1 14 8

35 169light 2 1 3 13 3

heavy x

36 169light 2 1 30 3

heavy x

37 170Alight 2 1 1 2 24 1

heavy x

38 170Blight 1 2 1 4 2 3 28 6 1 1 4 every other

heavy x

39 171light 1 39 1

heavy 30 176


Sam

ple

Fea

ture

Fra

ctio

n

Scan

ned

Onl

y

Ace

r sp.

(map

le)

Cas

tane

a de

ntat

a (A

mer

ican

che

stnu

t)

c.f.

Car

ya s

p. (h

icko

ry)

Car

ya s

p. (h

icko

ry)

c.f.

Cor

nus f

lorid

a (f

low

erin

g do

gwoo

d)

Gled

etsia

tria

ncan

thos

(hon

ey lo

cust

)

Jugla

ns ci

nere

a (b

utte

rnut

)

Jugla

ns n

igra

(bla

ck w

alnu

t)

c.f.

Jugla

ns s

p. (w

alnu

t fam

ily)

c.f.

Nys

sa sy

lvat

ica (b

lack

gum

)

c.f.

Nys

sa s

p. (b

lack

gum

)

c.f.

Ostr

ya v

irgin

iana

(iro

nwoo

d)

Plat

anus

occi

dent

alis

(syc

amor

e)

c.f.

Plat

anus

occi

dent

alis

(syc

amor

e)

c.f.

Popu

lus d

eltoi

des (

cott

onw

ood)

c.f.

Prun

us s

p. (

wild

che

rry)

Que

rcus

alb

a. (w

hite

oak

)

c.f.

Que

rcus

alb

a. (w

hite

oak

)

Que

rcus

rubr

a (r

ed o

ak)

Que

rcus

c.f.

rubr

a (r

ed o

ak)

Que

rcus

stell

ata

(pos

t oak

)

c.f.

Que

rcus

sp.

(oak

)

cf. R

obin

ia p

seud

oaca

cia (b

lack

locu

st)

c.f.

Salix

sp.

(will

ow)

c.f.

Tili

a am

erica

na (b

assw

ood)

Ulm

us ru

bra

(elm

)

Ulm

us s

p./

Celt

is sp

. (el

m/h

ackb

erry

)

ring

poro

us

diff

use

poro

us

Har

dwoo

d (n

ot id

entif

iabl

e)

Har

dwoo

d ro

ot

Juni

peru

s virg

inia

na (e

aste

rn r

edce

dar)

Pinu

s sp.

(pin

e)

Pinu

s sp.

bar

k (p

ine)

Con

ifer

Aru

ndin

aria

sp.

(ca

ne)

mon

ocot

c.f.

Aru

ndin

aria

sp.

(ca

ne)

mon

ocot

c.f.

mon

ocot

bark

c.f.

root

Not

iden

tifia

ble

resi

n

sam

ple

type

40 171light 2 1 1 1 13 1 1 pinch

heavy x

41 175light 1 1 2 23 3 11

heavy x

42 175light 1 1 18 6

heavy x

43 176light 2 1 2 4 20 1 1

heavy 1 19 12 pinch

44 176light 5 15 3 pinch

heavy x

45 177light 1 12 1 6 pinch

heavy 2 13 3

46 179light 1 1 4 7 9 1

heavy 1 1 16 1 3 pinch

47 180light 1 2 3 15 4

heavy 20 22 pinch

48 183light 1 1 4 1 4 8

heavy 3 17 11

49 185light 2 15 3 3

heavy x

50 185light 3 16 5

heavy x pinch

147Blight 1 18 1 160

heavy x

181 light x

Total Count 20 1 3 10 3 2 1 5 5 1 4 4 1 4 10 12 21 6 38 41 3 22 2 2 2 3 1 8 22 98 3 12 1437 1 2 6 56 18 31 1 2 1 1762

Table 6-11. Wood Charcoal and Miscellaneous (g) in Macrobotanical Remains from Block D. See Table 6-10 for sample and scan information.

Sam

ple

Fea

ture

Fra

ctio

n

Ace

r sp.

(map

le)

Cas

tane

a de

ntat

a (A

mer

ican

che

stnu

t)

c.f.

Car

ya s

p. (h

icko

ry)

Car

ya s

p. (h

icko

ry)

c.f.

Cor

nus f

lorid

a (f

low

erin

g do

gwoo

d)

Gled

etsia

tria

ncan

thos

(hon

ey lo

cust

)

Jugla

ns ci

nere

a (b

utte

rnut

)

Jugla

ns n

igra

(bla

ck w

alnu

t)

c.f.

Jugla

ns s

p. (w

alnu

t fam

ily)

c.f.

Nys

sa sy

lvat

ica. (

blac

kgum

)

c.f.

Nys

sa s

p. (b

lack

gum

)

c.f.

Ostr

ya v

irgin

iana

(iro

nwoo

d)

Plat

anus

occi

dent

alis

(syc

amor

e)

c.f.

Plat

anus

occi

dent

alis

(syc

amor

e)

c.f.

Popu

lus d

eltoi

des (

cott

onw

ood)

c.f.

Prun

us s

p. (

wild

che

rry)

Que

rcus

alb

a. (w

hite

oak

)

c.f.

Que

rcus

alb

a. (w

hite

oak

)

Que

rcus

rubr

a (r

ed o

ak)

Que

rcus

c.f.

rubr

a (r

ed o

ak)

Que

rcus

stell

ata

(pos

t oak

)

c.f.

Que

rcus

sp.

(oak

)

cf. R

obin

ia p

seud

oaca

cia (b

lack

locu

st)

c.f.

Salix

sp.

(will

ow)

c.f.

Tili

a am

erica

na (b

assw

ood)

Ulm

us ru

bra

(elm

)

Ulm

us s

p./C

eltis

sp. (

elm

/hac

kber

ry)

ring

poro

us

diff

use

poro

us

Har

dwoo

d (n

ot id

entif

iabl

e)

Har

dwoo

d ro

ot

Juni

peru

s virg

inia

na (e

aste

rn r

edce

dar)

Pinu

s sp.

(pin

e)

Pinu

s sp.

bar

k (p

ine)

Con

ifer

Aru

ndin

aria

sp.

(ca

ne)

mon

ocot

c.f.

Aru

ndin

aria

sp.

(ca

ne)

mon

ocot

c.f.

mon

ocot

bark

c.f.

root

Not

iden

tifia

ble

resi

n

1 141Blight 0.7 0.04 0.7 0.01

heavy 0.15

2 141Blight 0.02 0.01 0.02 0.03 1.1

heavy 0.02 0.32

3 144light 0.01 0.01

heavy 0.01

4 141Clight 0.01 0.04 0.39 0.01

heavy 0.01 0.01

5 143light 0.07 0.13 0.01

heavy 0.16 0.06

6 140light 0.01 0.01 0.01 0.2

heavy 0.22 0.02

7 141Alight 0.07 0.02 0.01 0.00 0.39

heavy 0.02 0 0.44 0.01

8 185Wlight 0 0 0.15 0.03

heavy 0.02 0.21 0.02

9 146light 0.01 0.01 0.01 0.47 0

heavy 0.03 0.01 0.2

10 147Blight 0.06 0.06 0.17

heavy 0.02 0.22

11 147Blight 0.01 0.04 0.01 0.03 5.33

heavy 0.02 1.32 0.12

12 147Blight 0.01 0.3

heavy 0.18

13 150light 0.01 0.01 0.05 0.01 0.04 0.41 0.01 0.01

heavy 0.01 0.26 0.01


Sam

ple

Fea

ture

Fra

ctio

n

Ace

r sp.

(map

le)

Cas

tane

a de

ntat

a (A

mer

ican

che

stnu

t)

c.f.

Car

ya s

p. (h

icko

ry)

Car

ya s

p. (h

icko

ry)

c.f.

Cor

nus f

lorid

a (f

low

erin

g do

gwoo

d)

Gled

etsia

tria

ncan

thos

(hon

ey lo

cust

)

Jugla

ns ci

nere

a (b

utte

rnut

)

Jugla

ns n

igra

(bla

ck w

alnu

t)

c.f.

Jugla

ns s

p. (w

alnu

t fam

ily)

c.f.

Nys

sa sy

lvat

ica. (

blac

kgum

)

c.f.

Nys

sa s

p. (b

lack

gum

)

c.f.

Ostr

ya v

irgin

iana

(iro

nwoo

d)

Plat

anus

occi

dent

alis

(syc

amor

e)

c.f.

Plat

anus

occi

dent

alis

(syc

amor

e)

c.f.

Popu

lus d

eltoi

des (

cott

onw

ood)

c.f.

Prun

us s

p. (

wild

che

rry)

Que

rcus

alb

a. (w

hite

oak

)

c.f.

Que

rcus

alb

a. (w

hite

oak

)

Que

rcus

rubr

a (r

ed o

ak)

Que

rcus

c.f.

rubr

a (r

ed o

ak)

Que

rcus

stell

ata

(pos

t oak

)

c.f.

Que

rcus

sp.

(oak

)

cf. R

obin

ia p

seud

oaca

cia (b

lack

locu

st)

c.f.

Salix

sp.

(will

ow)

c.f.

Tili

a am

erica

na (b

assw

ood)

Ulm

us ru

bra

(elm

)

Ulm

us s

p./C

eltis

sp. (

elm

/hac

kber

ry)

ring

poro

us

diff

use

poro

us

Har

dwoo

d (n

ot id

entif

iabl

e)

Har

dwoo

d ro

ot

Juni

peru

s virg

inia

na (e

aste

rn r

edce

dar)

Pinu

s sp.

(pin

e)

Pinu

s sp.

bar

k (p

ine)

Con

ifer

Aru

ndin

aria

sp.

(ca

ne)

mon

ocot

c.f.

Aru

ndin

aria

sp.

(ca

ne)

mon

ocot

c.f.

mon

ocot

bark

c.f.

root

Not

iden

tifia

ble

resi

n

14 151light 0.68 0.02 0

heavy 0.02 0.02 0.13 0.02

15 153light 0.03 0 0.02 0.3

heavy 0.42

16 154light 0.03 0.01 0.35

heavy

17 154light 0.01 0.22 0.04

heavy

18 155light 0.07 0.06 0.01 0.06 0.03

heavy 0.17 0.01

19 155light 0.01 0.01 0.01 0.00 0.19

heavy

20 155light 0.05 0.01 0.3 0.01

heavy

21 156light 0.01 0.61

heavy 0.25

22 157light 0.03 0.16

heavy

23 157Blight 0.05 0.31

heavy

24 158light 0.01 0.15 0.01

heavy 0.11 0.21

25 159light 0.01 0.06 0.5 0.02 0.02

heavy 0.2

26 160light 0.03 0.01 0.01 0.01

heavy 0.00 0.3


Sam

ple

Fea

ture

Fra

ctio

n

Ace

r sp.

(map

le)

Cas

tane

a de

ntat

a (A

mer

ican

che

stnu

t)

c.f.

Car

ya s

p. (h

icko

ry)

Car

ya s

p. (h

icko

ry)

c.f.

Cor

nus f

lorid

a (f

low

erin

g do

gwoo

d)

Gled

etsia

tria

ncan

thos

(hon

ey lo

cust

)

Jugla

ns ci

nere

a (b

utte

rnut

)

Jugla

ns n

igra

(bla

ck w

alnu

t)

c.f.

Jugla

ns s

p. (w

alnu

t fam

ily)

c.f.

Nys

sa sy

lvat

ica. (

blac

kgum

)

c.f.

Nys

sa s

p. (b

lack

gum

)

c.f.

Ostr

ya v

irgin

iana

(iro

nwoo

d)

Plat

anus

occi

dent

alis

(syc

amor

e)

c.f.

Plat

anus

occi

dent

alis

(syc

amor

e)

c.f.

Popu

lus d

eltoi

des (

cott

onw

ood)

c.f.

Prun

us s

p. (

wild

che

rry)

Que

rcus

alb

a. (w

hite

oak

)

c.f.

Que

rcus

alb

a. (w

hite

oak

)

Que

rcus

rubr

a (r

ed o

ak)

Que

rcus

c.f.

rubr

a (r

ed o

ak)

Que

rcus

stell

ata

(pos

t oak

)

c.f.

Que

rcus

sp.

(oak

)

cf. R

obin

ia p

seud

oaca

cia (b

lack

locu

st)

c.f.

Salix

sp.

(will

ow)

c.f.

Tili

a am

erica

na (b

assw

ood)

Ulm

us ru

bra

(elm

)

Ulm

us s

p./C

eltis

sp. (

elm

/hac

kber

ry)

ring

poro

us

diff

use

poro

us

Har

dwoo

d (n

ot id

entif

iabl

e)

Har

dwoo

d ro

ot

Juni

peru

s virg

inia

na (e

aste

rn r

edce

dar)

Pinu

s sp.

(pin

e)

Pinu

s sp.

bar

k (p

ine)

Con

ifer

Aru

ndin

aria

sp.

(ca

ne)

mon

ocot

c.f.

Aru

ndin

aria

sp.

(ca

ne)

mon

ocot

c.f.

mon

ocot

bark

c.f.

root

Not

iden

tifia

ble

resi

n

27 162light 0.01 0 0.05 0 0.09 0.09

heavy 0.02 0.01 0.01 0.36

28 163light 0.19 0.03 0.27 0.01

heavy 0.35

29 164light 0 0 0.03 0.08 0.12

heavy 0.01 0.1 0.04

30 165light 0.11 0 0.02 0.32

heavy

31 165light 0.01 0.01 0.03 0.56

heavy

32 166light 0.04 0.00 0.02 0.11 0

heavy 0.28

33 167light 0.16 0.01

heavy

34 168light 0.02 0.04 0.02

heavy 0.00 0.15

35 169light 0.07 0 0.03 0.19

heavy

36 169light 0.01 0 0.68 0.03

heavy

37 170Alight 0.04 0.07 0.01 0.01 0.43

heavy

38 170Blight 0 0.01 0.01 0.06 0.05 0.02 0.47 0.04 0.01 0.02

heavy

39 171light 0.01 0.93

heavy 1.27


Sam

ple

Fea

ture

Fra

ctio

n

Ace

r sp.

(map

le)

Cas

tane

a de

ntat

a (A

mer

ican

che

stnu

t)

c.f.

Car

ya s

p. (h

icko

ry)

Car

ya s

p. (h

icko

ry)

c.f.

Cor

nus f

lorid

a (f

low

erin

g do

gwoo

d)

Gled

etsia

tria

ncan

thos

(hon

ey lo

cust

)

Jugla

ns ci

nere

a (b

utte

rnut

)

Jugla

ns n

igra

(bla

ck w

alnu

t)

c.f.

Jugla

ns s

p. (w

alnu

t fam

ily)

c.f.

Nys

sa sy

lvat

ica. (

blac

kgum

)

c.f.

Nys

sa s

p. (b

lack

gum

)

c.f.

Ostr

ya v

irgin

iana

(iro

nwoo

d)

Plat

anus

occi

dent

alis

(syc

amor

e)

c.f.

Plat

anus

occi

dent

alis

(syc

amor

e)

c.f.

Popu

lus d

eltoi

des (

cott

onw

ood)

c.f.

Prun

us s

p. (

wild

che

rry)

Que

rcus

alb

a. (w

hite

oak

)

c.f.

Que

rcus

alb

a. (w

hite

oak

)

Que

rcus

rubr

a (r

ed o

ak)

Que

rcus

c.f.

rubr

a (r

ed o

ak)

Que

rcus

stell

ata

(pos

t oak

)

c.f.

Que

rcus

sp.

(oak

)

cf. R

obin

ia p

seud

oaca

cia (b

lack

locu

st)

c.f.

Salix

sp.

(will

ow)

c.f.

Tili

a am

erica

na (b

assw

ood)

Ulm

us ru

bra

(elm

)

Ulm

us s

p./C

eltis

sp. (

elm

/hac

kber

ry)

ring

poro

us

diff

use

poro

us

Har

dwoo

d (n

ot id

entif

iabl

e)

Har

dwoo

d ro

ot

Juni

peru

s virg

inia

na (e

aste

rn r

edce

dar)

Pinu

s sp.

(pin

e)

Pinu

s sp.

bar

k (p

ine)

Con

ifer

Aru

ndin

aria

sp.

(ca

ne)

mon

ocot

c.f.

Aru

ndin

aria

sp.

(ca

ne)

mon

ocot

c.f.

mon

ocot

bark

c.f.

root

Not

iden

tifia

ble

resi

n

40 171light 0.01 0 0.01 0.01 0.15 0.02

heavy

41 175light 0 0 0.02 0.57 0.03

heavy

42 175light 0.01 0.01 0.38

heavy

43 176light 0.02 0.01 0.01 0.02 0.21 0.02 0.01

heavy 0.01 0.2

44 176light 0.03 0.34

heavy

45 177light 0.01 0.15 0 0.03

heavy 0.01 0.06

46 179light 0.01 0.04 0.03 0.15 0.17 0.01

heavy 0.01 0.01 0.23 0.02

47 180light 0 0.02 0.01 0.02

heavy 0.25

48 183light 0 0 0.04 0.01 0.04 0.17

heavy 0.06 0.14

49 185light 0.01 0.21 0.05

heavy

50 185light 0.06 0.2 0.02

heavy

147Blight 0 0.14 0.14

heavy

181lightand

heavy

Total (g) 0.29 0.01 0.02 0.13 0.19 0.07 0.01 0.07 0.06 0.01 0.03 0.05 0 0.09 0.11 0.73 0.09 0.1 0.3 0.46 0.07 0.68 0.02 0.04 0.05 0.01 0.01 0.06 0.21 0.78 0.03 0.06 28.5 0.12 0.03 0.06 0.45 0.28 0.21 0 0.01 0.03 0.11

Interpretation

Under the best of conditions macroplant analysis does not recover all taxa used at a site, noreven necessarily proportions that reflect the initial abundance of any taxon. Rather, flotation analysisis a starting point. There is no guarantee that all recovered macroplant remains were actually used atthe site, and undoubtedly many more taxa were exploited than find their way into the archeologicalrecord. Nevertheless, with careful analysis, much can often be learned about the past relations betweenplants and people. Owing to the limitations of the assemblage from Block D, however, only the tentative interpretations can be offered.

Subsistence. In considering the Kolomoki subsistence system it is important to recognize the

limitations of the data. Overall, macrobotanical recovery has been so sparse that it is not possible todevelop a subsistence model for the Late Woodland period. Furthermore, even under the bestpreservation conditions, some plant remains are more likely to be preserved than others, due todifferential discard patterns, durability, and opportunity for charring. Charring is the means by whichall of the Kolomoki macroplant remains were preserved. Plant foods that are cooked beforeconsumption have an increased chance of preservation due to accidental charring, as in cookingaccidents. Similarly, those parts of the plant that were considered to be waste have a much higherchance of being preserved than those that were not. Finally, hickory shell was sometimes used as fueland as such would have a higher incidence of preservation. In trying to tease out a picture ofsubsistence ubiquity and even a mere presence are often nearly as important as count and weight of anygiven taxon.

With 99 percent of the samples containing nutshell, it is likely that locally abundant nut mastplayed an important role in Late Woodland Period subsistence. While hickory was the most abundantof all macroplant remains by both count and weight, it has been explained above that oak acorn isalmost certainly under-represented in the assemblage and was probably more important than hickory. Archaeological evidence indicates that hickory mast was a dietary staple in the Eastern Woodlands asearly as 8300 B.C., but acorn was probably the most important plant food in the Southeast up to andthroughout the Late Woodland Period (Gremillion 1998; Yarnell and Black 1983). Both hickory andoak are reasonably abundant locally in the Coastal Plain (Wharton 1978) and these are the only two nuttaxa that appear in any abundance in the assemblage. Based on both wood charcoal remains, and theexpected proportion of species in the local forest, acorn was probably more abundant in the area ofKolomoki during the Late Woodland Period. However acorn mast is produced in two or three yearcycles, depending on the species and is also sensitive to drought and other factors. Thus hickory waslikely a valued resource, for which there is much evidence.

Use of hickory as a staple makes sense in terms of energetic return. Based on native harvestingand processing methods, hickory is estimated to return 2565 kcal per hour, compared to 1107 kcal perhour for acorn (Gremillion 1998). Hickory was important in both the Middle and Late Woodlandoccupations at Kolomoki, and throughout the Southeast continued to increase in importance over timethrough the Mississippian period, even when maize was available (Bonhage-Freund 1999, 2003;Gremillion 1998; Yarnell and Black 1983; ).

The number and variety of starchy seeds, fleshy fruits, and possible "greens" or "pot herbs" inthe Block D macrobotanical assemblage is very small. Nevertheless, it is likely that seasonal wild plantfoods such as these were integral to the diet, adding variety and nutrients. A small amount of seedsrelated to plants of medicinal use suggests an intimate knowledge of and relationship to, the local plantcommunities.

A probable domesticated chenopod and possible common bean, together with the often"encouraged" wild type chenopodium suggest that gardening might have been practiced, but the numberof seeds are so small that these could just as easily be intrusions, and in any case they are statisticallyinsignificant. The presence of unidentified grass seeds also raise the possibility of grain being cultivatedor at least encouraged.

166

Although small amounts of maize were identified in the early/middle Late Woodland occupationof Block A at Kolomoki (Bonhage-Freund 2003), no maize was evident in the later Late Woodlandassemblage from Block D. As discussed below, the analysis of microplant remains suggests that maizecultivation was indeed associated with the occupation of Block D. The absence of maize macroplantremains may in part be attributed to the generally poor floral preservation in Block D. Additionally,there may have been unrecognized problems in the flotation process. This latter explanation seemsunlikely, however, as two different systems were employed, and another two samples were completelyhand-processed in small batches, and yet few or no seeds were yielded in those samples. In addition, thecontrol sample yielded a sizable number and variety of seeds and other small plant remains, using thesame equipment that was used to process the archaeological flotation samples.

Although gardening cannot be confirmed by the macrobotanical remains recovered from BlockD, it seems likely—given the results of the microbotanical evidence described below—that thesubsistence regimen at Kolomoki remained relatively unchanged during the Late Woodland Period, andcontinued to emphasis limited cultivation of domesticates as a supplement to a diet focused primarilyon wild resources. Since the 1980s, research suggests that Woodland groups of the Lower MississippiValley were not dependent on intensified agricultural production, and that cultigens comprised only aminor component of the diet, with nut mast serving as the plant-based staple (Caldwell 1958; Fritz andKidder 1993; Gremillion 2002; Springer 1980; Webb 1981). Gremillion (2002) argues that cultivationplayed a minor role in Woodland subsistence because it could not compete successfully with ahunter-gatherer economy based on abundant, seasonally predictable forest resources.

Although the Block D macrobotanical assemblage is limited, it is consistent with thischaracterization, and at odds with previous assumptions that Weeden Island societies became moreheavily dependent on agriculture during the Late Woodland period (Kohler 1991:105; Milanich et al.1997:76). The macrobotanical remains suggest that hickory and acorn mast were important staples. More generally, it would appear that wild resources provided the basis of the Late Woodland periodsubsistence.

Wood Use. Pine is the most abundant genus in the forests of the southeastern Coastal Plain,so its ubiquity is anticipated. It was commonly used as a construction material throughout the Southeasteven into early historic times (Bonhage-Freund 1997). Pine produces a hot, smoky, resin-filled fire, idealfor smudging or deterring insects, but by contemporary Western standards, not for cooking. It wasprobably used to start fires, and even as a fuel when absolutely necessary. Some types of pine are lessmessy than others and could be used as fuel if needed. Hardwoods of all kinds, but particularly generalike oak and maple provide long-burning coals with even heat. These hardwoods frequently burncompletely to ash, in contrast to pine that often leaves some residual charcoal (Bonhage-Freund 1994). This also contributes to the abundance of pine charcoal relative to hardwood charcoal. It is presumedthat the hardwood charcoal at Kolomoki was primarily fuel wood.

Distribution. Understanding the relative abundance and distribution of taxa can aid in theinterpretation of activity areas. In this assessment liters of floated soil are used as the leveling device,because fuel wood cannot be separated from structural wood in the case of post molds, and it isdesirable to include all feature types in this study. Every analyzed sample is included in assessing thisseed assemblage because every light and heavy fraction was scanned for seeds. In evaluating nutshell,only those features where both light and heavy fractions were fully sorted are included for reasonsdiscussed above.

When acorn shell is equalized with hickory by multiplying by a factor of 50, it dominates thelarge pit features (Figure 6-1). Since these are most likely to be trash features, this supports those studiesthat indicate the dominance of acorn in Woodland period diets. Hickory and acorn are fairly similar innumber in small pits, which may have been processing facilities, and also fairly close in number in postmolds. Again, some post molds may actually be small processing facilities and thus should show asimilar pattern to small pit features. Those post molds associated with a house structure, particularlythose near the door, would be expected to collect debris from in and around the house. These are areas

167

where food preparation would be anticipated. The greatest representation of seeds is found in postmolds.

In historic times acorn meat was removed from the shell in designated areas, usually on preparedoutcroppings of rock (Parker 1991). Therefore acorn shell would more likely occur in trash pits nearthe nutting stone areas, than close to houses, and this pattern seems to be reflected in the distributionof nutshell across the features.

With the bulk of the seeds being recovered in the large pit features, it is likely that these weretrash facilities. In particular, Feature 146, containing all 9 grass seeds, and a total of 33% of the seedassemblage is an ideal candidate for the storage pit classification.

One interesting observation is the fact that two of the three blackgum seeds and the single hollyseed came from the same small pit (Feature 163). This feature may have been used to prepare medicinefor body or spirit as both have healing properties, and holly is as an emetic. Holly leaves featureprominently in the "Black Drink" of Southeastern ritual (Hudson 1976; Swanton 1969.)

Environment and Site Seasonality. Many of the plants represented in this assemblage favor orrequire specific environments (Table 6-12). Due to this fact, some of the local habitats around theKolomoki site can be identified. The wood identifications suggest proximity to both wetlands (e.g.,sycamore) and drier better-drained soils (e.g., pine). Cane favors damp or wet soils. Monocots may beeither grass or cane and usually require open sunny areas. Many grasses can survive droughts by goinginto a kind of hibernation until water becomes sufficient to stimulate growth, but cane prefers mesic towet soils. Various grasses can also withstand fire, some even being "fire adapted", and Native peopleof the Southeast often burned fields to maintain grasslands for grazing game (Hudson 1976; Swanton1969 [1946]). Blackgum is likewise fire adapted, relying on fire to reduce competition and releasenutrients into the soil, and both seeds and wood of this taxon were identified. An abundance of pine,a pioneering genus, is another sign of constant renewal of the local forest by fire. Some types of pinerequire fire for seed release.

0

5

10

15

20

25

large pits small pits postmolds

hickory acorn (x 50) seeds

Figure 6-1. Number (count) of hickory, acorn (x 50), and seeds per liter of soil analyzed.

168

Table 6-12. Habitats and Seasonality Associated with Macroplant Remains from Block D. Based onBritton and Brown (1970), Fernald (1950), and Martin and Barkley (2000).

Taxon Scientific Nomenclature

habitat

ripens

agric

ultu

ral f

ield

/gar

den

dist

urbe

d ar

eas

dry

soils

open

fie

lds

old

field

s

edge

zon

es

open

woo

ds

rich

woo

ds

bott

omla

nds

wet

land

s

river

bank

s

peppervine Ampelopsis sp. x x x June-July

Holly Ilex sp. x x x May-JulyBlackgum Nyssa sylvatica x x Sept-Oct

Grape Vitis sp. (notdomesticate V. vinifera) x x Aug.-Sept.

grass family Gramineae x x x Summer-fall

3-seeded mercury Acalypha sp. x x x x x Sept.-Oct.

bellflower Campanula sp. x x Aug.-Sept.

goosefoot Chenopodium sp. x x Sept.

yellow rocket Barbarea sp. x June

mustard Brassica sp. x x July-Oct.

pulse family Leguminosae x x x x x x x summer-fall

Note: Habitat and seasonality depend on species.

The macroplant remains identified in this assemblage indicate a habitat containing some of thesame taxa found in the modern control sample or identified in the wood charcoal. Blackgum, pine, andgrasses all suggest a fire-adapted or dependent forest and field association, dominated by pioneeringspecies. Plants of the grape family, and 3-seeded mercury are common to both edge zone and old, orabandoned fields in early stages of succession, and black gum can be found here as well. All of therecovered species can become established in more than one environment, but the intersecting habitatsindicate an oak-hickory-pine forest with sub-canopy or understory species like blackgum or holly. Grape family would live on trees at the forest edge and spread through the edge zone into open fields. A variety of herbaceous plants, some only tentatively identified, would inhabit the zones from forestedge through newer disturbed and older regenerating fields. The juxtaposition of a variety of habitats,particularly forest and edge-zone, typically provides a rich mixture of food and medicinal plants. Whilethere is clear evidence that forest mast was harvested in the Late Woodland period, it can only behypothesized that additional species were also exploited.

The sparse seed assemblage does provide another window to the past. Because only one taxonwas identified to the species level, the seasons of seed or fruit ripening can only be expressed by a rangewhich encompasses the ripening schedule of all possible species of a given genus. Even so, the majorityof taxa ripen in late summer through early fall. This also coincides with the nut mast harvest andprocessing. Prior to storage all nuts would be smoked or otherwise processed to kill larvae and insects,and while hickory could be eaten immediately, acorns would be stored for at least a year after smoking(Ortiz and Parker 1991). If Block D was a nut processing site used for the autumn harvest, it wouldhelp to explain the near total absence of other plant foods. Other foods and their debris would bedisposed of closer to their processing point. Even if the entire archaeological seed assemblage wouldbe considered as "seed rain," at least 75 percent of the taxa produce fruits and seeds from Augustthrough October. This would indicate that this is the season when the features were being filled, andthere is minimal evidence to indicate their use in other seasons.

169

Microbotanical Remains

As a follow up to the negative recovery of maize in the macrobotanical remains from flotationsamples, soil samples from four features were submitted for pollen, starch, and phytolith analysis. Thesamples were retrieved from Features 191A, 192, 194, and 203. These are all large, basin-shaped pitsin the northern portion of the core block, within or adjacent to the structure.

Results

Pollen. Pollen analysis of these samples yielded differing records for the four pits represented. Quercus pollen was noted in all four pits and was particularly abundant in the sample from Feature 194(Table 6-13, Figure 6-2). This is consistent with the presence of an oak forest in the vicinity of BlockD. In addition, moderate quantities of Low-spine Asteraceae and Poaceae and small quantities of Carya,Pinus, Anacardiaceae, and Cheno-am pollen indicate that weedy plants of the sunflower family thatinclude marshelder, grasses, hickory trees, pine, a member of the sumac family (possibly poison ivy), andcheno-ams were part of the local vegetation (see Table 6-14 for background information on these andother taxa in the microbotanical assemblage). This sample contained a small quantity of Apiaceae starch,suggesting that roots/tubers of plants in the umbel family were stored in this feature or possibly thatthese roots/tubers were discarded as part of the trash that filled the pit. No evidence of Zea mays wasobserved in this sample.

The sample from Feature 203 also was dominated by Quercus pollen. Moderate quantities ofLow-spine Asteraceae and Poaceae pollen reflect local weedy members of the sunflower family andgrasses. Recovery of small quantities of Pinus, Tilia, High-spine Asteraceae, Brassicaceae, Cephalanthus,Cheno-am, Cyperaceae, and a slightly elevated frequency of Typha angustifolia-type pollen indicate localpresence of pine, basswood, other members of the sunflower family, members of the mustard family,buttonbush, cheno-ams, sedges, and cattails. It is possible that the presence of Highspine Asteraceae,Brassicaceae, Cheno-am, and Cyperaceae pollen represents weedy plants growing in the vicinity of thefeature or perhaps in the trash fill of the pit. Typha pollen might be present through wind transport ofpollen from cattails growing nearby in a wetland or drainage. This sample contained a small starch witheccentric hilum, representing a root or tuber that has not been identified. In addition, two starches withlinear hila were noted. These represent grass seeds from grasses such as little barley grass or wild rye.

Figure 6-2. Pollen diagram for microbotanical remains from Block D.

170

Table 6-13. Pollen and Starch Types Observed in Microbotanical Remains from Block D.Scientific Name Common Name

Arboreal Pollen Carya Hickory, Pecan

Magnolia Magnolia

Pinus Pine

Quercus Oak

Tilia Linden, Basswood

Non-Arboreal Pollen Anacardiaceae Sumac family

Asteraceae Sunflower family

Low-spine Includes ragweed, cocklebur, sumpweed

High-spine

Brassicaceae Mustard or cabbage family

Cephalanthus Buttonbush

Cheno-am Includes the goosefoot family and amaranth

Cyperaceae Sedge family

Eriogonum Wild buckwheat

Opuntia Prickly pear cactus

Poaceae Grass family

Smilax Catbrier, Greenbrier, Prickly-ivy, Sarsaparilla,Zarzaparrilla, Sarsparilla and Smilax

Typha angustifolia-type Cattail

Cultigens Zea mays Maize, Corn

Indeterminate Too badly deteriorated to identify

Starches Apiaceae

Small Eccentric hilum Root starch

Linear hilum Legume family or little barley grass, wild rye, andwheat grass seeds

Zea-type Starch Typical of starches produced by maize

Algae Botryococcus Algal body

Fungal Spores Concentricyste Fungal spore, indicator of wet, oxidized conditions

Recovery of these starches suggest that wild foods were either stored in this pit or discarded as part ofthe trash that filled the pit. No evidence of Zea mays was observed.

The Feature 192 sample exhibits only a moderate quantity of Quercus pollen. This sample wasdominated by Low-spine Asteraceae pollen, probably representing weedy plants from the sunflowerfamily, such as marshelder, growing in the trash fill of the pit. Moderate to small quantities of Pinus,Anacardiaceae, High-spine Asteraceae, Brassicaceae, and Poaceae pollen indicate the local presence of

171

Table 6-14. Background Information for Plant Taxa Identified in Microbotanical Remains fromBlock D.

Native Plants

Apiaceae (ParsleyFamily)

Several members of the parsley family were used for food, medicines, and charms by eastern tribes. Many plants were utilized for their greens, cookedas potherbs, and roots, which were most often boiled. These include Heracleum (cow parsnip), Angelica (angelica), Pastinaca (wild parsnip), Cryptotaenia(honewort), Carum (caraway), Daucus (wild carrot), Sium (water-parsnip), Osmorhiza (sweet cicely), and Erigenia (harbinger-of-spring). The seeds ofseveral of these plants also may be used as a seasoning (Peterson 1977:38-42).Seeds of these and other plants such as Taenidia (yellow pimpernel) alsowere smoked as hunting and fishing charms. Cicuta (water-hemlock) is noted to be poisonous, but medicinal and contraceptive uses are reported.Angelica, Thaspium (meadow parsnip), Sanicula (black snakeroot), Erigenia, Pimpinella (anise), Heracleum, and Sium also provided medicinal resources(Hamel and Chiltoskey 1975:23, 27, 31, 48, 55; Yarnell 1964:164, 171, 178, 180).

Arecaceae (PalmFamily)

Palms may be shrubs, vines, or trees. They are an important part of the economy in tropical and subtropical areas. The most well known palms includedate and coconut palms. Palms may be used for their fiber, for building and thatching, making baskets and mats, and for their edible fruits and oilcontent (Hickey and King 1981). The two palms important in the economy of the southeastern United States are discussed below.

Sabal minor(DwarfPalmetto)

Dwarf palmetto is a perennial shrub or tree of the palm family that was a staple food to native groups in the southeastern U.S.The fresh root slices were baked and eaten as bread, and the terminal bud or “heart” of this palm was specially valued, having adelicate asparagus-like flavor. A decoction of the dried root was taken by southern coastal native groups as a treatment for highblood pressure, kidney trouble, and “swimming in head”. The small roots, when fresh, were crushed and the juice rubbed intosore eyes as a counterirritant. Sabal minor grows in the southeastern coastal U.S. from Texas east to the Carolinas, then north toOklahoma and Arkansas, were it grows in floodplains, levees, riverbanks, and swamps. It has also been observed growing in muchdrier conditions inland in Texas and Mexico (Kiple and Ornelas 2000:1784; Moerman 1998:499).

Serenoa repens(Saw Palmetto)

Serenoa repens is the most abundant native palm in the U.S. Saw palmetto grows as a shrub (2-6 feet) or a small tree (20-25 feet)with stems that run parallel to the ground. Leaves are fan-like, with sword-shaped leaf blades radiating out from a central stalk.Two forms are recognized, the more common yellow-green color, and the blue-green or “silver” form, which is found in acontinuous belt on a strip of Florida’s coast and occasionally inland. To the pre-columbian Glades Cultures, saw palmetto appearsto have been one of the most important wild food sources, as well as to the later Seminole and Miccosukee. Leaf stems were usedto make medicine baskets. The plant was also used in the making of brushes, rope, dance fans, rattles, fish drags, fire fans, anddolls. Saw palmetto is found growing in low pine woods, thickets, savannas and coastal plains in the southeastern U.S. from SouthCarolina to southeastern Louisiana, including the Florida peninsula. It is a hardy, slow-growing, and fire resistant plant that canlive, especially in Florida, for hundreds of years (Foster and Duke 1990:228; Milanich 1998:42,124; Moerman 1998:527-528;Peterson 1977:170).

Arundinaria gigantea (River cane) River cane or giant cane is a bamboo-like, woody, hollow-stemmed grass native to the eastern U.S., found from Delaware toFlorida, Texas to Illinois. The Seminole used the stalks of this plant to make blowguns, knives, arrows, bows, spears, flutes,and blowing tubes. River cane was also used as a building material, in basket-making, and as fuel. The young shoots can beprepared and eaten like bamboo shoots. The large seeds can be used as a cereal or ground into flour. A decoction of the rootwas used to stimulate the kidneys, renew strength, ease breast pain, and as a cathartic. Arundinaria gigantea is found growing indense stands on low ground along river and streambanks and in swamps (Foster and Duke 1990:312; Moerman 1998:104;Peterson 1977:228).

Marantaceae(arrowroot family)

This is a group of perennial herbs with starchy tuberous rhizomes. The Marantaceae produce distinctive and diagnostic phytoliths and starch grains.Important economic plants within the family cultivated for food and ornamental use include Maranta, Calathea and Thalia. It should be noted that theMarantaceae are closely allied with the canna family (Cannaceae) and are recognized as a sister taxa. The Marantaceae are pantropical, and most diversein the American tropics. Some genera extend into subtropical and warm-temperate regions, and can be found naturally in moist forest habitats andswamps. While Maranta and Calathea occur in Mexico, only two species of Thalia are native to North America north of Mexico. Thalia geniculata is extanttoday in Louisiana and Florida, flowers June through December, fruits August through January, and can be found in swamps, marshes, cypress sloughs,along streams and lakes, often in regions with a pronounced dry season. Thalia dealbata is extant today in portions of Texas, Oklahoma, Missouri,Illinois, Arkansas, Louisiana, Georgia, and the Carolinas. Thalia dealbata flowers May through September and fruits June through October. Thaliadealbata is the only Marantaceae endemic to North America and the only one not found in the tropics. It is hypothesized that Thalia dealbata is the resultof an early, chance long-distance dispersal event from a South American population (Kennedy 2000; Zomlefer 1994). Marantaceae subsistenceutilization is not well understood. Moerman (1998) lists one account of Seminole use of Thalia geniculata as a food. Most of the Marantaceae subsistencerecord is derived from use of arrowroot (Maranta arundinacea) and Guinea arrowroot, also called sweet-corm root and leren (Calathea allouia) in southernMexico, Central and South America. Maranta is cultivated for its starchy roots which yield an edible starch that makes both flour and a very smoothjelly or paste. The flour is almost all starch, easily digested, and often employed as a thickener. Because Native Americans believed that the rootabsorbed poison from arrow wounds, the plant has long been thought to have medicinal properties (Atran, et al. 2003; Breedlove and Laughlin2000:224; Kiple and Ornelas 2000:1721). Calathea allouia has been cultivated by indigenous farmers in its tropical American habitat for more than athousand years. The roots are a rich source of many essential amino acids and are valued for their texture, which remains crisp even after cooking,and their flavor, which is said to resemble green maize (Kiple and Ornelas 2000:1782). Young spikes are also reported as being edible (Atran etal.2003:134), and leaves are used to wrap various foods, including meat (Breedlove and Laughlin 2000:224).

Poaceae (GrassFamily)

Members of the Poaceae (grass) family, such as Elymus (rye grass), Oryzopsis (ricegrass), Panicum (panic-grass) and Phalaris caroliniana (maygrass), wereused as a food resource by native groups. The seeds often were parched and ground into meal to make various mushes and cakes. The young shootsand leaves might have been cooked as greens. Grass stems also are reported to have been used for baskets, mats, etc. Grass seeds ripen during thesummer and fall (Fernald 1950; Medsger 1966:128-129; Reidhead 1981:238).

Cultivated Plants

Zea mays (Maize, Corn) Maize is a New World cultigen that evolved in the southern highlands of Mexico from annual teosinte. From this area, maize isbelieved to have reached the southwestern United States around 3,000 years ago and the eastern United States about A.D. 200.Once maize was introduced to eastern North America, the Great Plains acted as a barrier to gene flow; therefore, maize in easternNorth America is enzymatically different than maize of the Southwest or Mexico. After maize was introduced to the East, it isbelieved to only have been a minor crop and grown in small amounts for hundreds of years. By the Mississippian period, however,maize had become a principal food crop. Seeds were planted in hills. After the maize had sprouted, beans were often planted inthe corn hills. Watchtowers and platforms were erected so that women and children could guard the crops against birds, especiallycrows. Snares and deadfalls were used to trap raccoons, woodchucks, and deer. Two crops of maize were often planted. Theearly-ripening corn was picked while still green and roasted on the cob, and the late-ripening corn was dried and stored for futureuse. After harvesting, ears of corn were braided and hung inside the house to dry. Shelled maize was stored in bark containers,large baskets, and/or in underground storage pits. The husks and stalks were often saved and used in a variety of ways. Driedmaize was boiled, often with meat and dried pumpkin or squash (succotash), or ground into a meal that was used to make bread,mush, or dumplings (Hurt 1987; King 1987; MacNeish 1992). There are five different types of maize determined by theendosperm composition. Pop and flint corn have a high protein content and a hard starch. Dent corn has a deposit of soft, waxystarch at the crown of the kernel. Flour corn contains little protein and mostly waxy starch, while sweet corn stores more sugarthan starch (McGee 1984:241).

172

pine, a member of the sumac family (possibly poison ivy), other members of the sunflower family,members of the mustard family, and grasses. This is the only sample that contained angular starches thatare considered to be diagnostic of maize. Recovery of this starch without confirming evidence of pollenor phytoliths suggests that prepared food that contained maize was discarded here. Only one of theseangular starches was observed in the portion of the sample examined, providing relatively weak evidencefor the presence of maize. This sample also yielded evidence for the presence of water in the recoveryof a Botryococcus fragment, representing an algae, and a few concentricyste forms, which represent analgae that thrives in moist, oxidative conditions.

Finally, the sample from Feature 191A yielded the most diverse pollen record. There is no single,overwhelming pollen type, however, Poaceae pollen frequency dominates this record. Moderatequantities of Low-spine Asteraceae and High-spine Asteraceae pollen were noted, probably representingweedy plants growing in the vicinity of the feature. Recovery of small quantities of Magnolia, Pinus,Quercus, Anacardiaceae, Brassicaceae, Cheno-am, Cyperaceae, Erigonum, Opuntia, and Smilax pollenindicates local growth of magnolia, pine, oak, a member of the sumac family (possibly poison ivy), a member of the mustard family, cheno-ams, sedges, wild buckwheat, prickly pear cactus, and greenbriar(respectively). This sample yielded two large Poaceae pollen. One measured approximately 35 microns,which is the correct size for Agropyron (wheat grass) or Arundinaria (cane). The other pollen measured90 microns, and clearly represents Zea mays. This pollen was observed while scanning the microscopeslide in search of Zea mays, not within the pollen count. Recovery of Zea mays pollen in this sample isthe best evidence for the processing of maize in association with the Block D occupation.

Phytoliths. Phytolith analysis of these samples yielded a highly concentrated assemblage ofphytolith morphotypes (Figure 6-3 ). Overall, phytolith preservation was good; however, evidence ofdissolution (the dissolving of silica due to a soil pH level of above neutral) was observed on some ofthe phytoliths. All of the samples were overwhelmingly dominated by globular echinate phytolithsdiagnostic of the palm family (Arecaceae; see Figure 6-4A). For southwest Georgia, three species ofpalm are noted as occurring: needle palm (Rhapidophyllum hystrix), dwarf palmetto (Sabal minor), and sawpalmetto (Serenoa repens). Needle palms and dwarf palmetto are found in swampy woods, bottomlands,and along streams. Saw palmetto is an infrequent plant found along sandy banks of streams, sandy oakbarrens, and moist, sandy pinelands.

Since palms are not a dominant vegetation type in southwestern Georgia, the overwhelmingdominance of palms in the phytolith record (40% to 70% relative abundance) indicates that material

Figure 6-3. Phyolith diagram for microbotanical remains from Block D.

173

from members of the palm family were likely utilized intensely and then discarded at this site. Inaddition, the absence of Aracaceae (palm family) pollen from this site supports this interpretation.Aracaceae pollen is expected when palms are common in the local vegetation community. Moerman(1998) lists a wide variety of uses for palms in the southeast, which include medicinal, subsistence, and technological applications. Since palm globular echinate phytoliths are produced in all parts of palmplants (leaves, stems, bark, roots, etc.) the palm phytoliths observed in these samples may be derivedfrom almost any plant part.

Phytoliths derived from a wide variety of grasses were also very abundant in these samples. Forthe grass phytoliths, long saddles, diagnostic of the grass subfamily Bambusoideae, were the mostabundant morphotypes for the samples from Features 191A and 192, but were also present in thesamples from Features 194 and 203. Since cane grass (Arundinaria) is the only Bambusoideae native tothe United States, the long saddles observed in these samples can be ascribed to Arundinaria. There aretwo species of Arundinaria found in the Southeast, river cane (Arundinaria gigantea) being the species mostlikely to have been growing in this area during the time of occupation. Arundinaria gigantea formsextensive colonies in low woods, moist ground, and along river banks. It was once widespread in thesoutheastern United States, but cultivation, burning, and overgrazing have destroyed many stands.Moerman (1998) lists many examples of medicinal and technological utilization of Arundinaria gigantea.

Figure 6-4. Selected phytoliths from Block D samples (micrographs at approximately 400xmagnification). A) Several globular echinate phytoliths (1) diagnostic of the palm family (Arecaceae),and given the location of this site, are most likely derived from Serenoa repens and/or Sabal minor. Also in micrograph A is a trapeziform sinuate (2) diagnostic of the cool-season grass subfamilyPooideae. Many pooid grasses are also shade tolerant and mesophytic. B) Calathea-type phytolithdiagnostic of the arrowroot family (Marantaceae) inside view and C) in top view from sample fromFeature 203. D) Opaque perforated plate derived from inflorescence material from a member of thesunflower family (Asteraceae). E) Silicified cone cell from the achene of a member of the sedge family(Cyperaceae) (this morphotype is very similar to those produced by the genus Cyperus).

174

Phytoliths diagnostic of cool-season/shade tolerant grasses were numerous in these samples as well, andmost likely derived from the grass subfamily Pooideae. In the Southeast, pooid grasses (brome,bluegrass, etc.) can be found growing in shady wooded areas and in open moist areas, especially alongstreams. Phytoliths diagnostic of warm season Panicoideae grasses were present, but relatively rare inthese samples. Panicoid grasses typically require warm, humid and full-sun conditions. Panicoidphytoliths were most abundant in the samples from Features 191A and 192, and rare in the samplesfrom Features 194 and 203. Since maize is a panicoid, extra attention was given to observed panicoidphytoliths; however, none of these were indicative of maize.

Two other notable phytolith morphotypes observed were opaque perforated plates (Figure 6-4D) diagnostic of inflorescence material from a member of the sunflower family (Asteraceae), andsilicified cone cells (Figure 6-4E) from the achenes of a member of the sedge family (Cyperaceae),possibly from the genus Cyperus. These two phytolith types were only observed in the sample fromFeature 192, and indicate the possible utilization of sunflower and sedge seeds for subsistence. Thissample was dominated by Low-spine Asteraceae pollen, which includes marshelder. It is possible thatthe phytoliths from Asteraceae inflorescence represent the same plants that contributed the Low-spineAsteraceae pollen.

The most interesting phytolith observation was the recovery of phytoliths derived from thearrowroot family (Marantaceae) in the samples from Features 191A, 192, and 203 (Figure 6-4B and C).The Marantaceae phytoliths observed in these samples exhibit a morphology that is consistent withthose derived from fruits and roots with a rind (pericarp) or hard outer layer and a soft tissue inner area.The part of the phytolith extending out from the rind of a fruit, or outer, scaley area of a root orrhizome, is typically smooth and more dense with silica (Figure 6-5B, 1). The part of the phytolithextending into the soft interior of a fruit or rhizome is often ciliate or beaded in appearance (Figure 6-5 B, 2). The length of the ciliate portion extending into the soft tissue can often vary depending on thesize of the fruit or rhizome. This ciliate portion is lightly silicified and susceptible to dissolution andbreakage from the more heavily silicified “cap”.

Previous studies by PRI have identified these same phytolith morphotypes in archaeologicalsamples from New Mexico, Texas, Illinois and Indiana. Consultation with Dr. Dolores Piperno, fromthe Smithsonian Tropical Research Institute, and Dr. Jose Iriarte, from the University of Exeter, hasconfirmed the origin of these phytoliths within the Marantaceae family, most likely within the Calatheagroup, as the phytoliths observed here are unlike those produced by Maranta, the type genus of thefamily. Thalia dealbata and Thalia geniculata are the only two species of Marantaceae native to NorthAmerica north of Mexico. Thalia is an aquatic emergent plant of the southeastern United States foundin ponds, swamps, marshes, and along the margins of streams and lakes, often in areas with pronounceddry seasons. Thalia dealbata is found as far west as eastern Texas and Oklahoma. Phytoliths have beenexamined in leaves from both species of Thalia; however, no suitable matches for the Marantaceaephytoliths found in these samples have been observed. Phytoliths from the fruit and seed of Thalia havenot yet been examined, as these are still under cultivation at PRI; however, published images ofphytoliths from Thalia seeds do not bear any resemblance with the phytoliths observed here. Thus, themost likely source for these Marantaceae phytoliths would be the starchy rhizome of a species ofCalathea, to which these phytolith bear a resemblance (see Figure 6-5A and B for a comparison to areference Calathea allouia phytolith).

However, it should be cautioned that phytolith production within the Marantaceae (especiallyfor Thalia) is an ongoing study by PRI, and thus, we cannot yet ascribe the Marantaceae phytolithsobserved from these feature samples unequivocally to the genus Calathea. See Table 6-14 for moreinformation on the Marantaceae.

175

Interpretation

Overall, samples 191A, 192, and 203 exhibited a phytolith assemblage that was very similar.Sample 194 was a bit anomalous in that it had a very high percentage of amorphous and blockyphytoliths derived from woody material, no Marantaceae phytoliths, and the highest amount ofmicroscopic charcoal particles. Samples 191A and 192 were similar in the fact that they both had arelatively high abundance of Bambusoideae (cf. Arundinaria gigantea) saddle phytoliths. And finally, thesample from Feature 203 was unique in that it had the highest relative abundance of palm (Arecaceae)phytoliths, most likely derived from dwarf palmetto and/or saw palmetto, and the highest number ofMarantaceae phytoliths. No phytoliths diagnostic or even suggestive of maize or squash were observedin these pit feature samples.

Because of preservation differences in micro- and macrobotanical remains, analysis of thesesamples highlights the need to employ multiple lines of evidence (pollen, phytolith, macrofloral, etc.)to obtain the best possible interpretation of plant resource and subsistence utilization fromarchaeological features. Pollen and phytolith analysis yielded very different pictures of the localvegetation and possible exploitation of plants at this site. The pollen record provides evidence of a localoak forest that might have changed in density over time of these features represent more than one timeperiod. Alternatively, the samples might reflect seasonal pollen accumulation. Recovery of a largequantity of Low-spine Asteraceae pollen corresponds with recovery of Asteraceae charred perforated

Figure 6-5. Comparison of archaeological (Block D) and reference Marantaceae phytoliths. A)Marantaceae phytolith recovered in the sample from Feature 203. B) Reference phytolith from the edibleroots of Calathea allouia (scale bar equals 10 ìm). The Marantaceae phytoliths observed in these samplesexhibit a morphology that is similar with those derived from the roots of Calathea allouia. The part ofthe phytolith extending out from the outer, scaley area of the rhizome is smooth and more dense withsilica (B, 1). The part of the phytolith extending into the soft interior of the rhizome is often ciliate orbeaded in appearance (B, 2). This ciliate portion is lightly silicified and susceptible to dissolution andbreakage from the more heavily silicified “cap” portion, especially when disarticulated from the plantmatter and deposited in soil. When found in archaeological samples, the ciliate portion is often missing,or severely dissolved. The archaeological Marantaceae phytolith from the Feature 203 sample (A) is nota perfect match with Calathea allouia; however, different growing conditions, especially environmentalstress, can affect phytolith morphology. The Marantaceae phytoliths observed here are likely derivedfrom the Calathea clad of the family; however, more reference work needs to be completed.

176

plates. Pollen evidence for maize is limited to Feature 191A, which contained Zea mays pollen andFeature 192, which yielded an angular starch typical of those produced by Zea maize. Recovery of thispollen and starch indicate that even during this later occupation, people living at this site wereprocessing maize. No evidence for Zea mays was recovered from Features 194 and 203, which indicatesthat each feature must be evaluated separately.

Failure to recover pollen from palms suggests that they did not grow in abundance at the siteand that the palms represented in the phytolith record were harvested and brought to the site. Palmpollen is readily transported and is expected to be obvious in a pollen record from an area supportinga significant number of palms in the local vegetation communities. None of the starches recovered fromthis sample are typical of those produced by Marantaceae roots.

Although phytolith evidence for maize and squash was not observed in these samples, otherinteresting results were obtained. The phytolith assemblages recovered from these pit features suggestthat plant material from palm plants was intensively used for medicinal, subsistence, and/ortechnological applications. Phytolith analysis also indicates that cane grass (Arundinaria) may have alsobeen utilized for medicinal (roots) or technological applications. The recovery of Marantaceaephytoliths from three of the feature samples suggests that plant material, most likely the roots from amember of the arrowroot family, may have been used for medicinal or subsistence purposes. Theseroots are rich in starches and are easily processed to recover starch, making them a valuable subsistenceresource. It is possible that a member of the Marantaceae (Calathea, Maranta, etc.) may have beencultivated, as some microhabitats in the southeastern United States are similar to those in southernMexico and Central America, where members of the Marantaceae family thrive. Marantaceae trade (seed,dried root, processed starch) with people living in the Carribean and Yucatan via Florida is also apossibility. The only Marantaceae native to North America north of Mexico is represented by the genusThalia, which does not produce a suitable phytolith match with the Marantaceae phytoliths observed inthis study and from other locations within the United States. Modern phytolith reference work withinthe Marantaceae is ongoing at PRI, as the recovery of these phytoliths from several archaeological sitesin North America has raised some intriguing questions concerning trade, trade routes, and use of exoticfoods and now the possibility that this part of Florida might be part of a trade network or a corridoralong which trade goods were moved.

Macrofloral remains from this site indicate consumption of mast (nuts), a small quantity ofmaize, domesticated sunflower seeds, wild seeds such as goosefoot and amaranth, fruits such ashackberry, huckleberry, and possibly roots/tubers. Seeds from medicinal plants such as bedstraw, St.Johnswort, checkermallow, pokeweed, violet, and rush also are reported (Pluckhahn, et al. 2006). Pollenand starch analysis expands this list of plants to include the availability of cattail, as well as exploitationof roots/tubers from a member of the umbel family and grass seeds of a particular group of grasses(little barley grass, wild rye, wheat grasses). Maize is also represented. The phytolith record indicates thatoccupants of this site were exploiting palms, which are not evident as part of the local vegetationcommunity in the pollen record, meaning that they were not abundant locally or were grown to beharvested while they were young, perhaps the immature shoots known as “hearts of palm”, and werenot yet pollinating. Most significantly, a member of the Marantaceae was either cultivated at this site ortraded into the area and prepared at this site, since phytoliths specific to roots of a plant in this familywere recovered from three of the four pits examined. This is the first report of this family at any site insouthwest Georgia that reflects use of this resource, perhaps as part of a complex of foods associatedwith feasting, as discussed by (Pluckhahn et al. 2006).

177

Summary

The faunal and botanical collections from Block D are limited. Thus, our interpretations ofthese data must be qualified. Nevertheless, the faunal and botanical remains provide some useful pointsof comparison with the earlier Late Woodland assemblages from Block A, as alluded to here anddescribed more fully in Chapter 7.

The faunal assemblage from Block D is dominated by white-tailed deer. Other taxa includingeastern box turtle, wild turkey, and black bear are also present in the assemblage though in much lowerfrequencies. The small size of the Block D assemblage and its poor state of preservation make itunsuitable for estimating dietary contribution of different taxa; but it does indicate the Late Woodlandinhabitants of Kolomoki were relying on white-tailed deer and other terrestrial taxa for their animalresources. The low taxonomic richness, dominance of white-tailed deer, and abundance of Forequarterand Hindquarter elements observed in the Block D assemblage are strikingly similar to characteristicsof the previously studied Block A assemblage (Pluckhahn et al. 2006). The similarities between the twomay be the result of several contributing factors including the small sample size and poor preservationconditions of Block D, similar taphonomic conditions experienced by both, or similar animal usestrategies practiced by the early and late Late Woodland inhabitants of Kolomoki.

The study of botanical remains in Block D included both macrobotanical analysis of flotationsamples and microbotanical analysis of a smaller number of soil samples. In many ways, themacrobotanical results compare closely to earlier assays of flotation samples from Block A. As in BlockA, hickory nutshell dominated acorn in the samples from Block D, but acorn was present in smallamounts and was likely of equal prevalence when differential preservation is taken into account(Bonhage-Freund 2003). Thus it is likely that acorn and hickory were both staples throughout the LateWoodland period at Kolomoki.

Maize was found in the early/middle Late Woodland macrobotanical assemblage from BlockA but was not noted in the assemblage from Block D. Other cultigens are also rare or absent in theBlock D macrobotanical assemblage. Maize pollen was noted in the microbotanical analysis, however,suggesting that cultivation of this domesticate continued through the later Late Woodland. Still, maizepollen was not common in the microbotanical samples. Moreover, phytolith evidence for maize wasnot observed in these samples

The microbotanical remains suggest that plant material from palm plants was intensively usedby the residents of Block D for medical, subsistence, or technological applications, or some combinationof these. Phytolith analysis also points to the use of cane grass. Perhaps most interesting, the recoveryof Marantaceae phytoliths from three of the four microbotanical samples suggests that plant material,most likely the roots from a member of the arrowroot family, may have been used for medicinal orsubsistence purposes. These roots are rich in starches and are easily processed to recover starch, makingthem a valuable subsistence resource.

178

Chapter 7: Households Making History: Comparing Late Woodland Households in Blocks A and D

Thomas J. Pluckhahn

As Johnson (2006:124-125) notes, citing the example of North American historic archaeologists,one means of highlighting the agency of households is through comparison of assemblages from houselots. Previous chapters have summarized the features and artifacts found in Block D at Kolomoki,presenting evidence that these represent the remains of an archaeological household. In this chapter,the remains from Block D are compared to those from another archaeological household in Block Aat Kolomoki, excavated in 2001.

As described in greater detail elsewhere (Pluckhahn 2003; Pluckhahn et al. 2006), the Block Adesignation was applied to a cluster of 29 non-contiguous 1 × 1 m excavation units in the northeasternportion of the site, about 300 m north of Block D (see Figure 1-2). I compare the two excavationblocks and their respective archaeological households across five main variables: periods of occupation;domestic architecture; storage; ceramics; subsistence and environment; and flaked stone.

Unfortunately, as discussed in Chapter 1 (see also Steere 2011:79), the household archaeologyof the Late Woodland period societies of the Southeast is poorly developed. Obviously, in that theanalysis is based on only one excavated archaeological household from each period, the results of ourcomparison cannot generalized uncritically to the community at Kolomoki as a whole, let alone to thelower Chattahoochee or the wider region. Nevertheless, I present evidence that the changes identifiedin the comparison of Blocks A and D may hold for other households at Kolomoki, as well as othercommunities in the region.

Periods of Occupation

Five radiocarbon assays have been taken on materials from Block A (Table 7-1 and Figure 7-1). Most of these dates have been reported elsewhere, but one is reported for the first time here. These fivedates have a wide temporal spread. However, the youngest of the five dates—which was taken on bonefrom the hearth (Feature 171)—was obviously contaminated by small root hairs that could not beremoved during pretreatment. This date, as well as one taken on wood charcoal from the same featurethat exhibits a relatively wide margin of error, probably do not accurately reflect the period ofoccupation. The three remaining dates from Block A were taken on a maize kernel and Carya nutshell,materials which should have shorter use-lives and thus may better represent the occupational historyof the block. In keeping with this observation, these three dates are more precise and cluster closelytogether in time, with two sigma calibrated ranges extending from A.D. 420 to 660 and overlappingbetween A.D. 570 and 610. In the discussion to follow, I adopt cal A.D. 550 to 650 as a slightly moreconservative estimate for the occupation of the Block A archaeological household.

Four radiocarbon dates have been retrieved from Block D, including one previously reporteddate from Feature 34 in Test Unit 18 (Pluckhahn 2003:Table 2-3). As with those from Block A, the twosigma calibrated ranges on the dates from Block D exhibit a wide temporal span, from A.D. 650 to1020. Reconciling these ranges is somewhat difficult. On one hand, the two sigma ranges for the three

179

youngest dates overlap between A.D. 890 and 980. On the other hand, the ranges for the three oldestdates overlap in the interval from A.D. 780 to 880 and two of the dates—from separate features in ornear the presumed house—have nearly identical two sigma ranges that overlap between A.D. 780 and980. These older ranges are more consistent with the ceramic assemblage from Block D, which wouldseem to place the Block D occupation before around cal A.D. 750 to 800—when check stamped potterybegins to dominate assemblages in the area (Mickwee 2009; Milanich 1974). I adopt an estimate of calA.D. 750 to 850 for the occupation of the archaeological household in Block D, although it mayreasonable to assume that this area may have been occupied intermittently to as late as cal A.D. 950.

Thus, the archaeological households in Blocks A and D would appear to mark occupations thatwere sequential, or at least closely-related temporally (i.e., within a century or two of one another). Theformer dates to the early/middle Late Woodland, probably between cal A.D. 550 to 650. Thearchaeological household in Block D dates principally between cal A.D. 750 to 850; I refer to this as thelate Late Woodland, despite the cumbersome repetition. The occupation here may have continued intothe terminal Late Woodland, or what has sometimes been referred to as the “Emergent Mississippian”period (Kelly 1980; but for critiques of this term, see Cobb and Garrow 1996; Fortier and McElrath2002).

Although the two archaeological households are not far removed in time or space, they appearto straddle an important social and historical divide. At the time Block A was inhabited, the settlementplan at Kolomoki took the form of a large, formally-defined circular village centered on an immense

Figure 7-1. Plot of radiocarbon dates from Blocks A and D showing two sigma calibrated ranges. Generated using OxCal 4.0. Calibrations based on Reimer et al. (2004) and Stuiver and Reimer (1993).

180

Table 7-1. Radiocarbon Dates from Blocks A and D at Kolomoki.

Number Context Service Material 13C/12CRatio

ConventionalRadiocarbon

Age

2 Sigma CalibratedResults

Reference

Block A

Beta-165118 Feature 131, Zone A AMS bone -20.7 o/oo 1160±40 BP A.D. 780 to 980 Pluckhahn 2003:Table 2-3

Beta-161791 Feature 131, Zone B radiometric wood charcoal -25.0 o/oo 1280±70 BP A.D. 640 to 900 Pluckhahn 2003:Table 2-3

Beta-234443 Feature 57, Zone B AMS maize kernel -27.4 o/oo 1420±40 BP A.D. 570 to 660 this report

Beta-206785 Feature 57, Zone A radiometric Carya nutshell -26.1 o/oo 1480±40 BP A.D. 540 to 660 Pluckhahn et al. 2006: Table 1

Beta-206786 Feature 57, Zone B radiometric Carya nutshell -25.3 o/oo 1550±40 BP A.D. 420 to 610 Pluckhahn et al. 2006: Table 1

Block D

Beta-284228 Feature 191A AMS Carya nutshell -23.8 o/oo 1060±40 BP A.D. 890 to 1020 this report

Beta-242563 Feature 171 radiometric Carya nutshell -23.3 o/oo 1140±40 BP A.D. 780 to 990 this report

Beta-284227 Feature 147B, Zone B AMS Carya nutshell -23.4 o/oo 1150±40 BP A.D. 780 to 980 this report

Beta-161790 TU18, Feature 34 radiometric wood charcoal -25.0 o/oo 1290±60 BP A.D. 650 to 880 Pluckhahn 2003:Table 2-3

plaza. By the time Block D was occupied in the late Late Woodland period, this formal village planbroke down, and households were dispersed into less regular clusters around drainages.

The shift in settlement plan was coincident with a breakdown in mound construction andceremony. Much of the mound construction at Kolomoki—including the two major burial mounds(Mounds D and E)—took place during the Middle Woodland, predating either of the two households(Pluckhahn 2003:185-201). However, some degree of mound construction continued into the early andmiddle Late Woodland, concurrent with the Block A household. Specifically, radiocarbon datingsuggests that several of the smaller mounds at the site, including, two small dome-shaped mounds ofuncertain purpose (Mounds B and C) and two low platform mounds (Mounds F and H), date to thisinterval (Pluckhahn 2003:207-215). Thus, the Block A household dates to a time when public ceremonywas still active at Kolomoki, if on the wane. By the time the Block D household was occupied in thelate Late Woodland, however, public ceremony seems to have ceased.

Domestic Architecture

Architecture provides the most obvious difference between the archaeological households inBlocks A and D. In the case of the former, there is unequivocal evidence for a very small pit house. Evidence of domestic architecture in Block D is less certain, but points to the presence of a larger, butless substantial, structure of single set post construction.

The pit house in Block A is defined by the pit (Feature 57), which measured approximately 2.5to 3 m square at plan view (Figure 7-2). It had steeply sloping sides and a flat bottom that extended 30to 50 cm below the base of the plowzone (approximately 30 cm deep in this area). An entrance rampprojected about 2 m to the northeast from the center of the east side of the structure. The floor of thestructure was flat and virtually devoid of features, with two exceptions. First, there was a shallowrectangular depression in the interior of the structure at the base of the entrance ramp. Next, there wasa fire pit (Feature 131) near the center of the structure. This fire pit, which measured about 75 cm indiameter, extended about 40 cm below the floor of the pit house. The base of the fire pit was lined witha yellowish brown clay unlike the red sandy clay subsoil typical of the area immediate to the house. Given its small size, semi-subterranean construction, and the presence of an interior fire pit, thestructure clearly served as a cold weather dwelling. A concentration of post features to the south of thestructure could indicate the presence of a more lightly constructed, warm weather dwelling of single setpost construction, but no definitive post patterns are apparent.

The Block A structure is nearly identical to "keyhole" structures identified on several LateWoodland sites in the Midwest (Binford et al. 1970; Kelly et al. 1987; Kelly et al. 1990). These havebeen interpreted as small bent pole or mat-covered arbor structures (Binford et al. 1970:23). The wallposts were presumably located along the exterior margin of the pit and, with the exception of a few ofthe deeper posts, have been plowed away. Although keyhole structures such as that in Block A are bestknown from Late Woodland sites in the Midwest, semi-subterranean structures of roughly equivalentsize, shape, and construction have been identified elsewhere in Georgia and Alabama (e.g., Espenshadeet al. 1998; Jenkins and Ensor 1981:139; Price 1999; Shelby 2011), and at least one other possibleexample has been identified at Kolomoki (Pluckhahn 2003:156, 171-172). These appear to dateprimarily or exclusively to the Middle Woodland or early Late Woodland periods, roughly coeval withBlock A.

182

0 1 2

meters

¯

Figure 7-2. Comparison of early Late Woodland structure from Block A (left) and late Late Woodland structure from Block D (right).

183


house pit

hearth or fire pit


larger pit feature

Perhaps most relevant to the discussion of the Block A household is the pit house identified atthe Catoma Creek site in eastern Alabama, only about 100 km removed from Kolomoki (Figure 7-3). Like the example in Block A, this structure consisted of a small (ca. 2.7 m diameter), sub-rectangular,shallow pit. Like the house in Block A, it had a central hearth; however, it lacked an entrance ramp. A single carbon date with a calibrated range of A.D. 410 to 640 places the pit house at Catoma Creekcoeval with the Block A household in the early or middle Late Woodland (Price 1999; Shelby 2011).

The evidence suggests that a very different form of domestic architecture became prevalent inthe late and terminal Late Woodland, as evidence by Block D at Kolomoki and other sites in the region. As described in Chapter 4, an oval pattern is apparent in the distribution of post features in the core areaof Block D (see Figure 4-7). Along the southern edge of the oval pattern, post features are distributedat relatively regular intervals of 1.0 to 1.7 m. The pattern is less definitive in the northern end of thecore block, but even here an equivalent interval is evident between post features on the northwesternedge of the presumed structure. The overall distribution of post features suggests the presence of anoval structure of single set posts measuring about 7.3 m long and 5.2 m wide. Additional posts alongthe southern margins of the pattern suggest at least one possible episode of repair or rebuilding.

Several lines of evidence support the interpretation of a structure in Block D. First, as notedin Chapter 4, there is a regularity to the spacing of both exterior posts and interior support posts. Moreover, interior posts are deeper than those on the exterior, as might be expected if they supportedthe weight of the roof. As also discussed in Chapter 4, pit features appear to be non-randomlypositioned with respect to the structure. Specifically, larger pit features—especially large, bell-shapedpits—are more common in the interior of the structure. Further, pit features in the interior generallydisplay higher densities of artifacts than those outside the structure.

Finally, as described in Chapter 5, the distributional patterns of several classes of artifacts showdiscrepancies between areas inside and outside the structure. For example, all of the pit features yieldingNapier pottery are within or in immediate proximity to the structure, and only one of the pits completelywithin the house failed to produce any Napier sherds. The presence of Weeden Island Incised alsoappears to be correlated with the hypothesized house; most of the pit features within the structureproduced at least one sherd of Weeden Island Incised or Zoned Incised/Punctate, and most of thefeatures producing sherds of these types are located within or immediately adjacent to the structure. With regard to flaked stone, quartzite/sandstone also demonstrates a preference for features in or nearthe structure. On the other hand, there is a slight tendency for both early and late stage debitage toshow higher densities in pit features outside the structure than within.

Like that in Block A, the structure in Block D may have served, at least in part, as a cold weatherdwelling. At the center of the oval post pattern and presumed house was Feature 171, a basin-shapedpit with relatively high densities of FCR, ground, and pecked stone, that may have served as a fire pit. Unlike the structure in Block A, however, there is no evidence that this house was set in a pit. It alsodoes not appear to have been daubed.

The structure in Block D corresponds closely in size, shape, and method of construction withlate/terminal Late Woodland structures excavated at the Sycamore (Milanich 1974) and WoodlandTerrace (Mickwee 2009) sites, about 100 km to the south and southwest of Kolomoki, respectively. Thestructure at Sycamore was perhaps the more definitive of these two. It was oval and measured about6 x 9.5 m, judging from post features and a shallow depression marking the presumed living floor(Figure 7-4). Milanich (personal communication, 2010) now believes this may have been a cold-weather

184

Figure 7-3. Comparison of early/middle Late Woodland structures from the Catoma Creek site (left) (after Shelby 2011:Figure 2) andBlock A at Kolomoki (right).

185

house pit

hearth or fire pit


larger pit feature

house pit

0 1 2

meters

¯

Figure 7-4. Comparison of late Late Woodland structures from Sycamore site (left) (after Milanich 1974) andBlock D at Kolomoki (right).

186


hearth or fire pit


larger pit feature


slight depression(possible living surface)

structure with a possible “summer house” represented in an incomplete post hole pattern in a separateexcavation to the northeast. Based on seven radiocarbon dates, Milanich (1974:Figure 10) places theoccupation of the Sycamore house in the range from around A.D. 800 to 900. My own re-calibrationof these dates using OxCal 4.0 and the calibration curves of Reimer et al. (2004) and Stuiver and Reimer(1993) yields two sigma calibrated ranges running from A.D. 600 to 1260. The seven dates overlap inthe interval from cal A.D. 900 to 950.

The differences in domestic architecture between Blocks A and D point to possible changes insettlement. Gilman’s (1987) cross-cultural review demonstrates that households that occupy pit housesare sedentary at least during the season of structure use. This is perhaps not surprising given that pithouse construction is relatively labor intensive, and thus would likely not be undertaken if theoccupation was intended to be short-lived. The formality of the Block A pit house, with its projectingentrance passage and prepared hearth, would also seem consistent with a relatively permanentoccupation. By comparison, the less formal and less labor intensive single set post construction in BlockD would seem to suggest less permanent settlement—perhaps shifting seasonally or in longer-termcycles.

Differences in the floor areas of the twostructures (Figure 7-5) suggest possible temporalchanges in the size of the co-resident groups. Thepithouse in Block A, with a floor area of 8.2 m2

(not including entrance ramp), is about one-quarter the size of the structure in Block D (34.2m2). The floor area of a house does notnecessarily correlate directly with the number ofoccupants (see discussions in Goody 1958; Narrol1962; Wiessner 1974). However, it may bereasonable to hypothesize an increase in the sizeof the co-resident group in the transition from

early to late Late Woodland. I present additional evidence in support of this hypothesis below, in regardto discussions of ceramics and storage.

Of course, the co-resident group does not necessarily correspond with the household, however,since all the members of a household need not live under the same roof (Wilk 1983). In the early LateWoodland Patrick phase (A.D. 600 to 700) of the American Bottom, keyhole structures such as thatfrom Block A have been found arranged in clusters (Kelly 1990b), which Peregrine (1992) interprets aslineage compounds—several extended families from the same lineage functioning as a joint economicunit, and thus forming a single, large household (for a general ethnographic analogy along these lines,see Wilk 1988:).

There is currently no structural evidence to support the existence of such house clusters in BlockA or elsewhere at Kolomoki. Nevertheless, it seems likely—based on the small size of the early LateWoodland house—that some domestic activities were organized above the level of this very small co-resident group. Pluckhahn and colleagues (2006) have suggested that the faunal and botanical remainsrepresented in the fill of the pit house represents production and consumption by more people than thesmall pithouse could have accommodated. Perhaps, as Peregrine suggests for the American Bottom,the residents of two or more of these small houses—perhaps with each housing separate families of thesame lineage—functioned together as a single household.

0

10

20

30

40

Block A Block D

squa

re m

eter

s

Figure 7-5. Comparison of the floor area ofstructures in Blocks A and D.

187

It also seems possible that production and consumption were coordinated at a larger level thana lineage divided among two or more small houses, perhaps even by the village as a whole. As notedabove, the village during the early Late Woodland consisted of a seemingly very formal U-shapedarrangement of houses around a central plaza (Pluckhahn 2003), a pattern common in the WeedenIsland area (Russo et al. 2006). Peregrine (1992) and Flannery (1972, 2002) have interpreted this sortof spatial configuration to indicate economic cooperation at the village level. The paucity of storagefeatures in Block A and other Middle and early Late Woodland domestic contexts, as described in moredetail below, would be consistent with this interpretation. While houses were obviously present atKolomoki in this interval, perhaps as Cobb and Nassaney (2002:538-539) suggest, domestic space wasnot yet institutionalized.

Large structures such as the late/terminal Late Woodland house in Block D are often interpretedas the residences of individual, extended families (Flannery 1972, 2002; Peregrine 1992). As notedabove, by this time, the formal village arrangement at Kolomoki appears to have given to a seeminglymore haphazard scatter of houses (Pluckhahn 2003). This pattern may indicate less supra-householdorganization of production and consumption (Peregrine 1992), an interpretation also consistent withthe increase in domestic storage (as described below). Storage

The archaeological households in Blocks A andD obviously differ in periods of occupation andarchitecture. Equally obvious is the difference instorage capacity. We identified only seven pits(including storage, processing, and fire pits) in the 31m2 in Block A, for a density of 0.23 pits/m2. In BlockD, on the other hand, we identified 38 pits in 52 m2, fora density of 0.73 pits/m2. These figures reflect a morethan three-fold increase in the number and density ofstorage pits (Figure 7-6).

Pits in Block D were not only more numerous,but also much larger both on average and in total. The seven pits in Block A had a combined volume of205.2 liters. With the exception of the fire pit at thecenter of the house (Feature 171)—which accountsfor more than half of the total volume (132.5 liters)for the block—the pits here were small, with a meanvolume of just 29.3 liters. In contrast, the 38 pits inBlock D had a combined volume of 5175.5 liters anda mean volume of 136.2 liters. Equalizing for area ofexcavation (Figure 7-7), Block A had 6.6 litersstorage/m2. The storage capacity in Block D wasmore than ten times this, at 99.5 liters/m2. The

context of storage features is also notable. There are no storage pits in the interior of the early Late

0

0.1

0.2

0.3

0.4

0.5

0.6

0.7

0.8

Block A Block D

pit

feat

ures

/sq

m

Figure 7-6. Comparison of the density of pitfeatures in Blocks A and D.

0

20

40

60

80

100

120

Block A Block D

pit

volu

me

(lit

ers)

/sq

m

Figure 7-7. Comparison of the volume of pitfeatures in Blocks A and D.

188

Woodland structure in Block A. In contrast, there are several large bell- and basin-shaped pits in theinterior of the late Late Woodland house in Block D.

There is evidence to suggest that temporal changes in the capacity and context of storage atKolomoki were not limited to the two archaeological households under consideration here. Testexcavations elsewhere in the village have consistently demonstrated a paucity of pit features in Middleand early Late Woodland habitation areas (Pluckhahn 2003:126-179). The few pit features that havebeen identified in such contexts have been small and located outside any presumed structures. Comparative data from elsewhere in the region are hard to come by and variable. However, the fewMiddle and early Late Woodland pit houses that have been identified at other sites are similarly deficientin domestic storage, and generally lack any internal storage facilities (Shelby 2011). In contrast, at thelate Late Woodland Sycamore site, storage pits were common and present both within and outside thestructure (Milanich 1974).

Why the dramatic increase in storage across the Late Woodland period at Kolomoki and othersettlements in the region? Obviously, an increase in storage could expected with larger domestic groups,and this may partly explain the pattern. However, the increase in storage capacity from early to late LateWoodland as evidenced from Blocks A and D far outpaces the increases in floor area. Indeed, it seemspossible that houses became larger, at least in part, to accommodate greater interior storage.

As discussed below, the differences do not appear to be related a shift in diet; the subsistenceregimen throughout the Late Woodland period centered on wild plant and animal resources,supplemented by the growing of some cultigens. There does not appear to have been an increase inmaize consumption, as is the case in some parts of the Southeast and Midwest during the later LateWoodland period (McElrath et al. 2000:18-20). No phytoliths of Zea mays were identified from the pitswe sampled for microbotanical remains, suggesting these features not used for maize storage. Giventhe prevalence of nutshell and oak pollen in Block D features, the storage of mast resources seems amore likely interpretation. The recovery of phyotliths from an unidentified species of the arrowrootfamily in several samples provides an intriguing hint that some of the pits in Block D were also used forstoring tubers.

The increase in storage from early/middle to late/terminal Late Woodland could also be relatedto less residential stability, a trend hypothesized above in regard to architecture. As DeBoer (1988)notes, storage pits may be used at less permanent settlements to cache food and other goods. Thelimited macrobotanical evidence (described below) suggests that habitation of Block D continued acrossthe major seasons. However, it is still possible that the Block D house was unoccupied or at leastminimally occupied at certain times of the year when task groups were dispatched further afield in searchof game and plant foods.

Finally, it seems likely that the increase in storage from early/middle to late Late Woodland wasrelated to changes in the organization of production and consumption. Specifically, the limited storagecapacity and lack of private storage within the early Late Woodland house in Block A is consistent withthe supposition raised above that these activities were organized, at least in part, at a supra-householdlevel. Conversely, the presence of many pits in Block D, some internal to the structure, accords wellwith the idea that households such as this one were exercising greater autonomy over production andconsumption (DeBoer 1988; Hendon 2000; Wesson 1999, 2008).

189

Subsistence and Environment

While there are differences in the faunal and botanical assemblages from Blocks A and D, theseassemblages are fundamentally similar. This points to no major changes in subsistence or environmentover the course of the Late Woodland period at Kolomoki, unlike some other areas of the Southeastand Midwest where the cultivation of maize and other domesticates increased markedly after aroundcal A.D. 800 (McElrath et al. 2000:18-20). The assemblages from Blocks A and D are limited byrelatively poor preservation and recovery, however, so these conclusions must be considered tentative.

Faunal Remains

The faunal assemblages from Blocks A and D differ greatly in size. Block A includes over 6,000identified specimens, nearly five times the number identified for the Block D assemblage. In terms ofspecimen weight, Block A is over three times the size of the Block D assemblage. There is also adifference in the condition of the two assemblages. During analysis, Compton noted that the Block Dassemblage seemed to be in a poorer state of preservation than Block A. This subjective observationis supported by a comparison of the average size of white-tailed deer specimens for the twoassemblages. Here, specimen weight is used as a proxy for size. For the Block A assemblage theaverage weight of white-tailed deer specimens is 6.32 g. White-tailed deer specimens in the Block Dassemblage are considerably smaller with an average weight of 4.09 g. This indicates greaterfragmentation of the Block D assemblage.

In spite of the differences in size and preservation of the two assemblages, they are generallysimilar in taxonomic richness. Like Block A, the Block D assemblage is dominated by white-tailed deerwith lesser amounts of primarily terrestrial taxa represented (Table 7-2). Eastern box turtle and wildturkey are present in both assemblages indicating these two species were important to both early andlate Late Woodland inhabitants of Kolomoki. Black bear is the only species present in Block D that wasnot identified in the Block A assemblage.

In addition to the similarities in taxa represented, the two assemblages also have similarwhite-tailed deer element distribution profiles (see Figure 6-1). In both assemblages the mostover-represented categories are the meatier Forequarter and Hindquarter categories. The 16Vertebra/Rib and Foot categories are under represented in both assemblages. The Head, Forefoot, andHindfoot categories are less similar between the two assemblages but all are less represented than theForequarter and Hindquarter categories.

The prevalence of terrestrial taxa in both of the assemblages is not surprising considering thatKolomoki is not located near any major aquatic habitats. By extension it is logical that the large-bodiedterrestrial white-tailed deer would be the dominant source of meat for both the early and late LateWoodland inhabitants of Kolomoki. This fact is illustrated by the dominance of white-tailed deer tothe total estimated biomasses of Blocks A (45.8 percent) and D (99 percent). However, the exclusivedominance of white-tailed deer to these two assemblages is not considered to be representative of theoverall animal use strategy at Kolomoki. Although venison is commonly cited as the single mostimportant source of meat to prehistoric Native Americans living in the Southeast, it was but oneimportant resource on a long list of animal species utilized.

In the case of Block A, Pluckhahn et al. (2006:267-268) interpret the low richness of theassemblage and the dominance of white-tailed deer as the result of a specific activity or event rather than

190

Table 7-2. Comparison of the Faunal Assemblages from Blocks A and D.Taxon Description NISP MNI Weight (g) Biomass (kg)

Block A Block D Block A Block D Block A Block D Block A Block D

N % N % kg % kg %

Osteichthyes Indeterminate bony fish 1 1 6.3 0.01 <0.01 0

Testudines Indeterminate turtle 8 4 1.08 0.6 0.03 0.1 0.02 0.1

Terrapene carolina Eastern box turtle 1 1 1 6.3 1 9.1 0.45 0.9 0.02 0 0.03 0.2

Aves Indeterminate bird 85 9 9.5 0.8 0.16 0.3 0.02 0.1

Meleagris gallopavo Wild turkey 6 3 1 6.3 1 9.1 2.33 0.6 0.04 0.1 0.01 0.1

Mammalia Indeterminate mammal 5,943 1435 2,298.26 749.7 27.88 53.6 10.17 55.5

Sciurus niger Fox Squirrel 1 1 6.3 0.36 0.01 0

Urocyon cinereoargenteus Gray Fox 1 1 6.3 0.60 0.02 0

Ursus americanus Black bear 1 1 9.1 0.1 <0.01 0

Procyon lotor Raccoon 1 1 6.3 1.22 0.03 0.1

Odocoileus virginianus White-tailed deer 308 142 10 62.5 10 90.9 1,926.21 580.8 23.78 45.8 8.08 44.1

Vertebrata Indeterminate vertebrate 79.60 167.6

TOTAL 6,355 1,595 16 13 4,319.62 1501.1 51.97 18.33

the overall animal use strategy of the inhabitants of Kolomoki. They suggest one or two small-scalefeasts facilitated by successful communal deer hunts could have produced the abundance of white-taileddeer in the assemblage. They further support this hypothesis with white-tailed deer element distributiondata that indicate the meatier portions of the carcass were transported to the community more oftenthan the less meaty portions. This pattern might be expected if numerous deer were killed some distancefrom the community and only the more valuable portions of the carcasses were carried back.

While the Block D assemblage is similar to that of Block A in terms of its low taxonomicrichness, the dominance of white-tailed deer, and an over representation of the meatier Forequarter andHindquarter portions of the white-tailed deer carcass, there are differences in the context of theassemblage that argue against an interpretation of feasting in this case. The Block A faunal remainscame almost exclusively from the fill of the house pit and associated unit levels (Pluckhahn et al. 2006). The house pit fill appeared to have been deposited relatively rapidly. In contrast, the faunal remainsfrom Block D were spread across a number of features that were likely filled over the course of manyyears. Thus, the assemblage is more likely to represent general trends in faunal procurement rather thanone or two unusual episodes. Those general trends appear to consist of an emphasis on white-taileddeer gathered some distance from the village. This would be consistent with the interpretation ofgreater seasonality of occupation during the late Late Woodland, as noted above in regard to architectureand storage.

Floral Remains

The identification of plant remains from Blocks A and D was primarily limited to the analysisof macrobotanical remains from flotation samples. The sampling and analysis regimes for the twoexcavations were similar. For Block A, we processed 52 flotation samples with a combined volume ofapproximately 211 liters from 33 features (Pluckhahn 2003; Pluckhahn et al. 2006). For Block D, weprocessed 52 samples with a combined volume of approximately 295 liters from 35 distinct features, asdiscussed in Chapter 6.

The quantity of macrobotanical remains that was recovered from both excavation blocks is verylimited. Such poor recovery is typical of sites in the sandy Coastal Plain soils of the Southeast (Branch-Raymer and Bonhage-Freund 2011). Nevertheless, the botanical collection from Block A includes 15identifiable taxa (excluding wood charcoal). The richness of the collection from Block D is morelimited; 10 identifiable taxa (excluding wood charcoal) are represented. Table 7-3 compares themacrobotanical assemblages from Blocks A and D. To facilitate comparison, I focus on the remainsfound in flotation sample light fractions.

The two assemblages are generally comparable. For both excavation blocks, mastresources—particularly hickory—predominate in terms of both counts and ubiquity. Acorn is thesecond most common taxon represented in both count and ubiquity for both assemblages. Bothhickory and oak are reasonably abundant in the Coastal Plain of the Southeast (Black et al. 2002;Delcourt and Delcourt 1977; Foster et al. 2004; Godfrey 1988; Scarry 2003: 89; Wharton 1978) Thepredominance of hickory among other mast resources is common for assemblages from the lowerSoutheast (Scarry 2003:89).

Both assemblages include a limited number and variety of seeds from edible starchy/herbaceousplants, wild fruits, and potential medicinal plants. The assemblage from Block A stands out for the

192

Table 7-3. Comparison of Macrobotanical Assemblages from Blocks A and D. Block A Block D

Count Ubiquity Count Ubiquity

samples features samples features

mast acorn (Quercus sp.) nutshell, cap 9 9.62 15.15 22 13.46 17.14

hickory (Carya sp.) nutshell 361 78.85 69.70 140 46.15 54.29

walnut (Juglandaceae sp) nutshell 8 7.69 11.43

hazelnut (Corylus sp.) nutshell 2 1.92 3.03

domesticates sunflower (Helianthus annuus) 1 1.92 3.03

maize (Zea mays) cob, kernel 10 9.62 6.06

starchy/herbaceous pigweed (Amaranthus sp.) seed 2 3.85 3.03

yellowrocket (Barbarea sp.) seed 1 1.92 2.86

unidentified grass (Poaeae) seed 9 1.92 2.86

goosefoot (Chenopodium sp.) seed 1 1.92 3.03 1 1.92 2.86

wild fleshy fruits hackberry (Celtis occidentalis) seed 1 1.92 3.03

peppervine (Ampelopsis sp.) seed 1 1.92 2.86

blackgum (Nyssa sylvatica ) seed 1 1.92 2.86

huckleberry (Gaylussacia sp.) seed 3 3.85 3.03

medicinal plants bedstraw (Galium sp.) seed 1 1.92 3.03

Saint Johnswort (Hypericaceae) seed 43 21.15 15.15

checkermallow (Sidalcea neomexicana) seed 2 3.85 3.03

pokeweed (Phytolacca americana) seed 1 1.92 3.03

violet (Viola sp.) seed 1 1.92 3.03

unidentified seed, c.f. Brassica sp. 1 1.92 2.86

4-seeded mercury (Acalypha virginiana) seed 2 3.85 5.71

rush (Juncus sp.) seed 2 1.92 3.03

variety of taxa identified as potential medicinal plants. As previously noted, that this may be the resultof the unusual formation of the fill of the pit house in Block A, perhaps in association with one or twoepisodes of small-scale feasting (Pluckhahn et al. 2006)

The most obvious difference between the two assemblages is in the representation ofdomesticated plant taxa. In Block A, maize (Zea mays) was present in five of the 52 samples, for anubiquity of 9.62 among samples. Most of the samples that produced maize, however, were from thehouse pit, however, so the ubiquity of maize in the features is slightly lower (6.06). These maize remainsfrom Block A include one glume, two cob fragments, and seven kernel fragments. As indicated in Table7-1, one of the kernel fragments was dated and returned a two sigma calibrated date of A.D. 570 to 660.

In addition to maize, the Block A assemblage includes evidence for another domesticate,common sunflower (Helianthus annuus). The simple presence of sunflower constitutes insufficientevidence of domestication (Asch and Asch 1985:165). However, the size of the Kolomoki specimen,

193

when adjusted for breakage and shrinkage due to carbonization, is consistent with domesticated varietiesas described by Heiser (1985: 60) and Yarnell (1978: 296).

Various authors have suggested that cultigens contributed little or nothing to the diet of theWoodland societies of the Deep South (Fritz and Kidder 1993; Gremillion 2002, 2003; Scarry 2003;Springer 1980; Webb 1981). The limited quantity of maize and other cultigens in the Block Aassemblage and from other excavations at the site (see Bonhage-Freund 1998, 2001; Pluckhahn 2003)does not contradict this argument, but it does suggest that cultivation may have formed a larger part ofthe economy than many have allowed.

Given the recovery of at least limited quantities of cultigens from Block A, the complete—orat least virtually complete—absence of domesticated plants from the light fractions from Block D issurprising. We recovered no macrobotanical remains identifiable as maize or sunflower. One goosefoot(Chenopodium sp.) seed was recovered. Unfortunately, however, while the goosefoot specimen recoveredfrom Block D appears to represent a domesticated variety given its truncated margin, it consists onlyof half of a seedcoat (testa) and thus the thickness of the seedcoat cannot be measured to confirm this.

The absence of domesticates from Block D could be due to differential preservation or recovery. However, while the counts of seeds and other macroplant remains from Block D are lower than forBlock A, the differences are not dramatic. Still, to test the possibility that the absence of domesticatesfrom Block D is due to sample or recovery biases, we submitted four soil samples from an equal numberof features for pollen, starch, and phytolith analysis, as described in Chapter 6. Oak pollen wasubiquitous in the four samples; also prevalent were phytoliths of an unidentified species of thearrowroot family, suggesting storage of tubers. Two of the samples produced no evidence of maize. One of the samples produced possible evidence for maize in the form angular starch typical of thoseproduced by Zea mays. The fourth sample contained pollen definitively identified as that of Zea mays.

Recovery of this pollen and starch indicate that the production of maize continued into the lateLate Woodland period at Kolomoki. However, it seems unlikely that maize production increased duringthe Late Woodland period at Kolomoki, as some have suggested for Weeden Island societies moregenerally (Kohler 1991:105; Milanich et al. 1997:76). Indeed, given the absence of maize from the BlockD macrobotanical assemblage, coupled with the dearth of maize phytoliths and the limited evidence formaize pollen and starches, it seems possible that maize became less, rather than more important, in thelater Late Woodland diet.

Although a reduction in maize consumption seems counterintuitive, there are several possibleexplanations. The consumption of maize may have been associated primarily or exclusively withcommunity rituals which, judging from the drastic reduction in mound construction (Pluckhahn 2003),appear to have declined in scale and frequency by the time of the Block D occupation. However, therecovery of maize from the early Late Woodland household in Block A indicates that maize was notassociated exclusively with large-scale public ceremony. Moreover, there is evidence that household-based food rituals continued into the late Late Woodland, as evidenced by the recovery from Block Dof several elaborate serving vessels (described in more detail below).

Increased seasonality of occupation is another possible explanation. As noted in Chapter 6, themacrobotanical remains from Block D appear to have come primarily or exclusively from taxa whose

194

edible fruits, seeds, and nuts ripen in the late summer or fall. If maize production took place elsewhereduring the late Late Woodland, perhaps at settlements in the surrounding area, it would help explain itsapparent reduced presence in Block D. However, the presence of maize pollen indicates some degreeof both summer occupation and maize production at Kolomoki. This, in turn, would suggest year-round occupation, or at least no major seasonal abandonments.

It is possible that the climate came more poorly suited to the cultivation of maize during the lateLate Woodland period. Smith (2009:176) has implicated drought as a factor in the changes that tookplace in the lower Chattahoochee Valley during the Late Woodland. However, the differences in themacrobotanical assemblages are not as pronounced as one might expect if drought was a serious factor. The taxa represented in wood charcoal from the two blocks also undermine the suggestion ofpronounced drought as a causative factor (Table 7-4). Pine dominates both charcoal assemblages in

Table 7-4. Comparison of Wood Charcoal Assemblages from Blocks A and D.Block A Block D

Count Ubiquity Count Ubiquitysamples features samples features

oak (Quercus sp.) 6 9.62 15.15 22 1.92 2.86

red oak (Quercus rubra) 9 7.69 6.06 63 42.31 54.29

white oak (Quercus alba) 16 7.69 12.12 27 17.31 25.71

post oak (Quercus stellata) 3 5.77 8.57

Hickory (Carya sp.) 1 1.92 3.03 9 13.46 20.00

butternut/walnut (Juglans sp.) 8 9.62 9.09 9 3.85 5.71

maple (Acer sp.) 2 3.85 6.06 20 15.39 22.86

American chestnut (Castanea dentata) 1 1.92 3.03 1 1.92 2.86

common persimmon (Diospyros) virginiana) 1 1.92 3.03black gum(Nyssa sp.) 5 5.77 8.57

ash (Fraxinus sp.) 3 1.92 3.03

eastern hophornbeam (Ostrya virginiana) 1 1.92 3.03flowering dogwood (Cornus florida) 1 1.92 2.86

sycamore (Platanus occidentalis) 6 5.77 9.09 3 1.92 2.86

cherry (Prunus sp.) 5 1.92 3.03 12 5.77 5.71

elm/hackberry (Ulmaceae) 7 7.69 9.09 3 3.85 5.71

basswood (Tilia americana) 2 1.92 2.86

willow (Salix sp.) 2 1.92 2.86

cottonwood (Populus deltoides) 10 9.62 11.43

ironwood (Ostrya virginiana) 4 5.77 8.57

honey locust (Gledetsia triancanthos) 2 1.92 2.86

black locust (Robinia pseudoacacia) 1 1.92 3.03 2 3.85 5.71

unidentifiable hardwood 157 76.92 72.73 87 40.39 51.43

eastern redcedar (Juniperus virginiana) 16 13.46 12.12 12 7.69 11.43

pine (Pinus sp.) 791 98.08 96.97 873 94.23 100.00

cane (Arundinaria sp.) 57 25.00 30.30 51 34.62 48.57

195

terms of overall counts and ubiquity. While there are differences between the two assemblages,particularly in the abundance and ubiquity of various species of oaks, these are relatively minor. Ifanything, the assemblage from Block D includes a richer array and higher ubiquities of taxa adapted tomoist conditions, including sycamore, cherry, hackberry, basswood, willow, cottonwood, maple, andwillow.

The final alternative is that Late Woodland households at Kolomoki became more, rather thanless, dependent on a wide spectrum of plant and animal resources due to the introduction of newtechnologies that increased their economic productivity and autonomy, as Muller (1997:129) hassuggested for the Midwest. He credits the change in subsistence largely to the adoption of the bow andarrow which, in combination with improvements in cooking technology, “...might have given LateWoodland households the ability to break free of social and economic entanglements, andinterdependence that bound them to floodplains” (Muller 1997:129-130).

Judging from the site’s upland location, households at Kolomoki were never bound tofloodplains or floodplain-grown crops to the extent suggested by Muller. Nevertheless, they may havebeen socially and economically interdependent, as argued above for the early Late Woodland from thehousehold remains in Block A. The decision to reduce maize production and consumption may havebeen tied to a weakening of supra-household institutions—especially as they encouraged sharing—andan increase in household autonomy, as Muller suggests for the Midwest. Whether new ceramic andflaked stone technologies were a causal factor in the process, however, or if instead households weresimply choosing to assert their autonomy over production and consumption and added thesetechnologies secondarily, is a topic taken up in the sections that follow.

Ceramics

Blocks A and D produced ceramic assemblages of roughly comparable size (Table 7-5). BlockA yielded over 18,000 sherds, of which around 8,515 could be confidently identified to type or moregeneral categories of surface decoration (Pluckhahn 2003:148-165). All of these appear to be associatedwith the Woodland period occupation of Kolomoki. Excavations in Block D produced 21,637 sherds. Discounting a small number of Mississippian sherds and a larger quantity of pottery too small or erodedto confidently identify leaves a sample of 9,272 identifiable Woodland sherds.

With the exception of plain wares of various tempers (overwhelmingly sand), the assemblagefrom Block A is dominated at roughly 40 percent by complicated stamped pottery. This category iscomprised principally of Swift Creek Complicated Stamped, with far lesser amounts of BlakelyComplicated Stamped (an earlier variety). Decorative attributes characteristic of the GulfTradition—including those associated with both Santa Rosa Swift Creek and Weeden Islandwares—each contribute less than 1 percent of the total assemblage of identifiable pottery.

In contrast, complicated stamped pottery makes up only about 10 percent of the assemblagefrom Block D. As with Block A, most of this Swift Creek Complicated Stamped. Blakely ComplicatedStamped is not present, but Napier Complicated Stamped pottery—a hallmark of the Late Woodlandperiod in northern Georgia (Wauchope 1966)—now appears in small, but not insignificant, numbers.Net marked and ridge pinched types of the Gulf Tradition appear for the first time, albeit in lower

196

Table 7-5. Relative Frequencies of Surface Treatments in Identifiable Woodland Pottery Assemblagesfrom Blocks A and D.

Surface Treatment Block A (n=8,515)percent

Block D (n=9,272)percent

plain 59.92 77.18

complicated stamped 39.32 10.19

dentate stamped 0.03 0

incised 0.12 3.68

punctate 0.07 3.38

net marked 0 0.39

ridge pinched 0 0.08

red filmed 0.42 4.92

check stamped 0 0.14

cord marked 0 0.04

other 0.12 0

frequencies. More relevant for the discussion here are the increases in incised, punctate, and red filmed wares, from well less than 1 percent each in Block A to over 3 percent each in Block D. A chi-squaretest indicates very significant differences between the two assemblages with respect to decorativeattributes (χ2=1282.123, p < 0.001) (the ridge pinched, cord marked, check stamped, and dentatecategories were collapsed into “other,” since the low expected values (<5) violate the assumptions ofthis test).

A minimum number of vessel (MNV) analysis was conducted on larger rims sherds (>5 percentof orifice diameter) and diagnostic body sherds from the two excavation blocks to ascertain changes invessel form and size. As with sherd counts, the MNV samples from the two contexts are comparablein size, with a total MNV sample of 55 for Block A and 57 for Block D.

The MNV analysis reveals temporal changes in the frequencies of vessel forms (Table 7-6). These changes are difficult to evaluate in statistical terms due to the low (<5) expected values for severalcategories. Nevertheless, it is difficult to see the differences between several categories arising fromchance, particularly given the continuity in other forms. For example, the relative frequencies of openbowls and simple bowls/unrestricted jars are nearly identical. On the other hand, while collared jarspredominate in the Block A assemblage, forming about 30 percent of all vessels, this form accounts foronly about 16 percent of the vessels in Block D. The decline in collared jars is accompanied bypronounced increases in the relative frequencies of several other vessel forms, particularly restrictedbowls (from 12.73 percent in Block A to 19.3 percent in Block D) and dishes/plates (from 1.82 percentin Block A to 8.77 percent in Block D).

197

Table 7-6. Relative Frequencies of Vessel Forms Identified in MNV Analysis of Blocks A and D.

Vessel Form Block A (n=55)percent

Block D (n=57)percent

simple bowl/unrestricted jar 23.64 24.56

restricted bowl 12.73 19.30

open bowl 18.18 17.54

collared jar 30.91 15.79

neckless jar 10.91 12.28

cup 1.82 1.75

dish/plate 1.82 8.77

Comparison of the principal vessel forms from Blocks A and D reveals that the mean orificediameter for each vessel form increased through time (Table 7-7). There are statistically significantincreases in the mean orifice diameters of simple bowls/unrestricted jars (t=3.0606, .01 > p > .005),collared jars (t=2.4473, .05 > p > .02), and open bowls (t=2.3479, .05 > p > .02).

Thus, the ceramic assemblages from the archaeological households in Blocks A and D point tosignificant changes in ceramic decoration and vessel form and size in the transition between the earlyand late Late Woodland periods. Changes in decoration are the most dramatic and must be consideredin the context of broader trends in material culture across the region, particularly the waning of the SwiftCreek Complicated Stamped type. Wallis (2011) has recently argued that on the Atlantic Coast, theexchange of Swift Creek vessels was inextricably linked with mortuary ceremonies at mound centers,and more specifically with gift-giving associated with marriage alliances. This remains to bedemonstrated for southwestern Georgia, but it is probably not an unreasonable extrapolation, given theseemingly disproportionate representation of Swift Creek pottery at Kolomoki during the Middle andearly Late Woodland periods (Pluckhahn 2003:Table 2.2). As noted above, Pluckhahn (2003:207-210)has suggested that community ceremony—as evidenced mainly by mound construction—declined atKolomoki after around A.D. 550 and may have ceased entirely after approximately A.D. 650. This is

Table 7-7. Comparison of Mean Orifice Diameter for Vessel Forms in Blocks A and D.

Form mean orifice diameter (cm)

Block A Block D

collared jar 18.0 24.0

simple bowl/unrestricted jar 21.4 27.3

open bowl 20.2 23.6

restricted bowl 19.4 20.4

neckless jar 16.3 21.4

198

roughly coincident with the precipitous decline in Swift Creek evident in the transition between theoccupations in Blocks A and D. If true, the decline in mortuary ceremonialism and the reducedfrequency of Swift Creek may be inextricably linked.

Greater mobility during the late Late Woodland may have contributed to the changes indomestic ceramic assemblages by bringing more regular and intensified interaction among householdsof different ceramic traditions. Potters from late Late Woodland households at Kolomoki may haveemulated the more elaborate net marked and incised decorations they observed on vessels being madeand used by households they encountered on more distant forays. Increased interactions would havealso brought more opportunities to trade for vessels of other ceramic traditions. In contrast with earlierLate Woodland and Middle Woodland times, there do not appear to have been any supra-householdinstitutions discouraging the manufacture, acquisition, or use of foreign or otherwise elaboratelydecorated vessels in domestic contexts.

The heightened interaction among households might have included participation in domestic,food-related rituals, a shift from the more public feasting commonly inferred for the earlier LateWoodland and Middle Woodland periods (Knight 1990, 2001). It is noteworthy in this regard that twoof the vessel forms that become more common in domestic assemblages in the late Late Woodland atKolomoki—restricted bowls and plates/dishes—were presumably associated primarily or exclusivelywith serving. Relative to other vessel forms, these more commonly bear elaborate and labor intensivedecoration such as zoned punctation and zoned red filming, making them well-suited to such ritualoccasions. In regard to the possibility of an elaboration of household ritual, it is worth noting that oneholly seed (Ilex sp., 2.5 percent ubiquity) was recovered from a feature in Block D; the leaves of variousholly species were used as an emetic and most famously as a major ingredient of the ritual purgative"black drink."

Changes in vessel form and size may be linked to the waning in popularity of complicatedstamping as a decorative attribute. Swift Creek Complicated Stamped is strongly associated with thecollared (or folded rim) jar (Hally 2009; Pluckhahn 2010), so it is hardly surprising that the decline inthis vessel form parallels the decrease in complicated stamping. However, it seems likely that otherfactors were at work as well, given that changes in vessel form and increases in vessel size are notrestricted to those bearing complicated stamping. Perhaps most obvious, larger vessels could beassociated with the increase in the size of the domestic group, as evidenced by the dramatic increase infloor space cited above between the pit structure in Block A and the single set post structure in BlockD. The production of larger vessels may also relate to the inferred shift to greater household autonomyover production and consumption, since households would have been responsible for more of theirown food preparation, storage, and serving.

Flaked Stone

Although the differences are not as dramatic as with ceramics, there are important differencesin the flaked stone assemblages from Blocks A and D. The assemblage from Block A includes 35,054flaked stone artifacts; the assemblage from Block D is notably smaller (N=6696), suggesting that agreater share of lithic production and maintenance was conducted elsewhere. This is consistent withhigher residential mobility during the late/terminal Late Woodland.

199

Figure 7-8. Comparison of the relative frequencies of raw material types in the flaked stone assemblagesfrom Blocks A and D.

Figure 7-9. Comparison of the relative frequencies of early andlate stage debitage and tools in the flaked stone assemblagesfrom Blocks A and D.

The two assemblages do not differ greatly with respect to the relative frequencies of rawmaterials (Figure 7-8). Coast Plain cherts account for the bulk of both assemblages, although quartz isalso common. Quartzites/sandstones and Ridge and Valley cherts are slightly better represented in theassemblage from Block D. The differences are minor but consistent with the greater diversity notedin ceramic types. As suggested above for ceramics, the trend may be indicative of greater mobility andheightened interaction among late Late Woodland households. Raw material variability is generallyassumed to be greater on sites of shorter or less permanent occupation (Andrefsky 1998:233-234)

The two assemblages also donot differ greatly with respect to therepresentation of general types ofdebitage and tools. Figure 7-9compares the two blocks in regard tothe relative frequencies of early-stagedebris (cortical flakes and angularshatter), late-stage debris (non-cortical flakes), and cores/tools. This sort of aggregate analysis isobviously problematic (Andrefsky1998:140). It is also important tobear in mind that our analysisprobably greatly understates thenumber of expedient tools, as discussed in Chapters 2 and 5. Still, some potentially meaningful generaltrends are worth noting. In particular, late-stage debris associated with the final stages of toolmanufacture and tool maintenance is proportionately better represented in Block A than Block D. These sorts of activities are probably more common on sites associated with greater residential stability(Parry and Kelly 1987; Sievert and Wise 2001).

200

Figure 7-10. Comparison of the relativefrequencies of morphological clusters representedin the hafted biface assemblages from Blocks Aand D

Temporal differences in flaked stonetechnology are most apparent with regard tohafted bifaces. As described in Chapter 5, theflaked stone assemblage from Block D includes38 hafted bifaces that were sufficiently whole anddistinctive to classify to type. The assemblagefrom Block A consists of 31 bifaces. To facilitatecomparison, the bifaces were sorted into generalmorphological clusters based primarily ondiagnostic attributes of the hafting area(Pluckhahn and Norman 2011). As indicated inFigure 7-10 there are pronounced differences inthe relative frequencies of general morphologicalclusters. Specifically, in Block A the cluster ofbifaces with hafting areas that are straight orexpanding cluster makes up roughly three-quarters of the assemblage. Proximally-

contracting bifaces make up the remaining one-quarter of the collection. In Block D, on the other hand,proximally contracting points comprise slightly less than one-half the assemblage, and proximallystraight/expanding points decrease in relative frequency to around 40 percent. Most notable, however,is the increase in the relative frequency of triangular points, from 0 percent in Block A to around 13percent in the late Late Woodland collection from Block D.

Table 7-8 compares the relative frequencies of the individual point types represented in theassemblages from Blocks A and D. Owing to the diversity of the assemblages and the low relativefrequencies of most types, the differences are not pronounced in most cases. However, there areexceptions. Consistent with the trends identified for the morphological clusters described above, theterminal Late Woodland occupation of Block D witnessed significant declines in the manufacture ofseveral straight/expanding types such as Bakers Creek and Broward. Conversely, the manufacture anduse of contracting types such as New Market and Swannanoa appear to have increased. Most striking,however, is the large increase in the number and relative frequency of Woodland and MississippianTriangulars.

Pluckhahn and Norman (2011) conducted a functional analysis of hafted bifaces from Kolomoki(including those from Blocks A and D), applying the criteria suggested by Thomas (1978), Schott (1997),and Nassaney and Pyle (1999) for discriminating dart and arrow points. Their analysis suggested thatwhile a number of individual points in the collection could have functioned as arrows,Woodland/Mississippian triangulars were the only type consisted classified as such. This accords wellwith the widely held notion that the small triangular form is the earliest arrow point in the Southeast(Blitz 1988; Nassaney and Pyle 1999). The five examples of the Woodland\Mississippian triangular type from Block D are the only identified from Kolomoki to date (from an assemblage of more than200 hafted bifaces). This strongly suggests that the Block D occupation was coincident with theintroduction of new, or at least much improved, arrow technology at Kolomoki. Thus, arrow pointswere not adopted until around A.D. 750 at Kolomoki and—at least judging from their relativefrequency—for the next 50-100 years remained a relatively minor addition to the long-establishedtradition of spear-thrown darts.

201

Table 7-8. Comparison of the Relative Frequencies of Hafted Biface Types in Blocks A and D.

Type Block A Block D

Bakers Creek 25.81 13.16

Broward 12.90 2.63

Duval Type 1 3.23 0

Duval Type 2 6.45 7.90

Duval Type 3 3.23 2.63

Ebenezer 3.23 10.53

Florida Copena (lanceolate/triangular variety) 6.45 0

Florida Copena (notched variety) 9.68 0

Jacks Reef 0 2.63

New Market 3.23 15.80

Swan Lake 3.23 7.90

Swannanoa 3.23 5.26

Tampa 0 2.63

Weeden Island Straight Stemmed 6.45 2.63

Woodland/Mississippian Triangular 0 13.16

provisional type 3.23 0

TOTAL 100.00 100.00

Judging from work elsewhere in the region, households at Kolomoki appear to have beenrelatively slow in adopting improved bow and arrow technology. In the Midwest, arrow points becomecommon in the archaeological record after cal A.D. 600 (McElrath et al. 2000:18). Milanich andcolleagues (1997:188) argue that arrow points were present at McKeithen by A.D. 500—severalcenturies earlier than the Block D occupation. The identification of a triangular arrow point in the boneof a woman buried on top of one of the mounds at McKeithen provides evidence that these arrowpoints were not used solely for hunting game.

Discussion

The Late Woodland period is bounded by two great structural transformations: on one end the“collapse” of the Middle Woodland lifestyle on the other the development of a new Mississippian socialorder and the subsequent Mississippianization of the greater Southeast. Beck et al. (2007), building onthe work of Pauketat (Pauketat 1994, 1997a, 1997b, 2004a, 2004b, 2007), have argued that the lattertransformation was precipitated by an event: the A.D. 1050 ‘Big Bang” at Cahokia. McElrath andcolleagues (2000:14-16; see also: Carr and Case 2008:28; Yerkes 1988:1) speak of the transition between

202

the Middle and Late Woodland between A.D. 300/400 and 500 in terms also befitting an event: that is,as a sudden and dramatic “collapse” (but see Dunnell and Greenlee 1999). However, as Sewell hassuggested, and as Beck and colleagues argue for the A.D. 1050 “Big Bang,” events may be conceivedof as sequences of ruptures that reorganize structures (Sewell 2005:261). The argument advanced hereis that these “bookend” events were interpenetrated by a series of ruptures across the Late Woodland,that these ruptures included major changes in domestic organization, and that these changes were largelythe product of the agency of households themselves.

In the Midwest, the first such rupture consisted of a shift in settlement from floodplain touplands, accompanied by a marked decline in the construction of burial mounds (McElrath et al.2000:14-16). Various explanations have been offered for this Middle Woodland “collapse.” Dunnelland Greenlee (1999) have reviewed the social and political factors invoked by archaeologists to explainthe transformation of Middle Woodland societies and have found all of them deficient, either becausethey have few testable implications (cultural fatigue, outside influence, internal sociopolitical strife) orbecause their expectations do not match the archaeological record (warfare, agriculture, bow and arrow). Citing the ubiquity and relatively simultaneity of settlement shifts across the Midwest, McElrath et al.(2000:15) posit an environmental change—perhaps a shift in flooding regimes—as a possible trigger. Carr and Case (2008:28), however, reject environmental explanations for the collapse of Ohio Hopewell;they suggest instead that a “perceived spiritual difficulty” precipitated a division in a key ceremonialalliance.

As noted above, these changes are also true of Kolomoki and portions of the Weeden Islandarea, although the timing was delayed. However, burial mound construction declined here too byaround A.D. 500 and a dispersed settlement system was in place throughout much of the area by aroundA.D. 800 (Pluckhahn 2003:185-201; White 1981:60). Given its upland location, Kolomoki would havebeen little affected by the flooding implicated by McElrath and colleagues. Smith (2009:176) suggeststhat periodic drought may have been the principal causal factor behind the eventual settlement shiftsin this region; there were also changes in sea level that may have affected Weeden Island groups closerto the coast (Marquardt 2010). However, as noted above, there are no good indications of climatechange in the botanical assemblages from Blocks A and D. This, coupled with the persistence ofmound building for several centuries after the Midwestern Middle Woodland collapse, suggest that localchanges were not simply a response to broad-scale environmental crises or social upheavals.

Whatever the ultimate causes, the decline in burial mound construction surely reflects afundamental rift in the schemas that held sway through the Middle Woodland period. It is nowcommonly agreed that Middle Woodland mortuary ceremonies intertwined themes relating to mourning,world renewal, and rites of passage (Carr 2006b:475-476, 2008; Carr and Case 2008; Hall 1979:259-261;Romain 2000:167-197) and were organized around both extended kin groups such as clans and sodalitiesthat cross-cut kinship and residences (Carr 2006a, 2008; Carr and Case 2008). These sodalities, as wellas the more distant social ties they often manifested, probably declined in importance as the ceremoniesthey sponsored and supervised became smaller, less extravagant, and less frequent. Along with thesesocial relationships, there must have been fundamental changes in the schemas that dictated theconstitution of kin and community.

The themes that underwrote Middle Woodland mortuary ceremony were not likely cast asideor forgotten, however, even if they were less often expressed in public ceremony. As noted above, smallplatform mounds were constructed at Kolomoki in the early/middle Late Woodland. Knight (1990,

203

2001) has argued that activities that were undertaken on the summits of these and similar mounds inthe region—including the repetitive placement of large posts and the displaying of meat onscaffolding—were directed to communal rites of world renewal.

Given the persistence of these practices, it is perhaps not surprising that early and middle LateWoodland households at Kolomoki remained economically interdependent. The paucity of storagefacilities in Block A, combined with the evidence for communal hunting and feasting, indicate that muchof production and consumption was organized at organizational level above the co-resident group,perhaps by lineages or the community as a whole (Pluckhahn et al. 2006). These households thus likelycontinued to maintain schemas that encouraged pooling of effort and resources. The value placed onshared labor may have been strong enough to discourage—at least for time being—the adoption of thebow and arrow, even as it was coming into more frequent use in surrounding area.

Domestic production and consumption appear to have been radically transformed in the intervalbetween the occupations in Blocks A and D, probably between around cal A.D. 650-750, as evidencedfirst and most dramatically by the increase in the number and size of storage pits. McElrath andcolleagues (2000:18) note that the increase in storage is a widespread pattern in the Midwest during theLate Woodland. The change is sometimes attributed to intensified maize production (Steere 2011:197). However, maize was clearly not the impetus for increased storage at Kolomoki, where pits appear tohave been employed mainly for storing and processing mast and roots. Even in the Midwest, wheremaize would become more important in the Late Woodland, the emphasis on domestic storage appearsto predate intensive maize cultivation (McElrath et al. 2000:18).

At Kolomoki, the increase in household storage is coincident with the collapse of the traditionof mound building. I argue that the two trends are related and reflect another major rupture in thestructure: the social rules and mechanisms for sharing were fundamentally transformed as communalceremony withered and households exercised greater control of their own food production andconsumption. This is not to say that households were completely independent economically. Instead,I suggest that the reciprocal social obligations that were exercised and reinforced in public ceremonygave way to the development of less formal relationships between individual households. Theserelationships may have been materialized in household-based food rituals, as indicated by the increasedfrequency of specialized serving vessels in the ceramic assemblage from Block D.

Also coincident with these changes is the first appearance of small triangular points, markingthe adoption of new—or at least improved—bow and arrow hunting technology (Blitz 1988; Nassaneyand Pyle 1999). These points are relatively uncommon in Block D, suggesting the transition to thistechnology was both delayed and gradual at Kolomoki relative to other areas of the Southeast andMidwest.

Various authors have discussed the advantages of the bow and arrow for hunting. Relative tothe spearthrower, the bow and arrow is generally credited with improved hunting efficiency owing toits greater range, velocity, and accuracy (Blitz 1993; Muller 1997:129; Seeman 1992:42; but see Shott1993). Bettinger (1999) has argued that the greater accuracy of the bow and the ability it conferred tohunters to stay more still during release facilitated individual hunting and negated the advantages ofhunting in groups.

204

Given the apparent superiority of the bow and arrow, why were households at Kolomoki slowto adopt the new technology? Seeman (1992:42) notes that there are production costs associated withbow and arrows relative to spearthrowers; they have more component parts, require a wider range ofmaterials to manufacture, require more skill to produce, and also require more maintenance costs (dueprimarily to the higher rate of arrow loss). None of these production costs precluded the rapid adoptionof the bow and arrow in most areas of eastern North America by around cal A.D. 700, however (Blitz1988; McElrath et al. 2000:5; Nassaney and Pyle 1999; Shott 1993).

Rarely considered in previous discussions of the adoption of the bow and arrow are the potentialsocial costs of this new technology for communally-organized societies. By permitting increasedefficiency in individual hunting, the bow and arrow reduced the need for the sort of coordinated drives that were presumably more common when hunting with darts (Bettinger 1999; Hall 1980). Suchcoordinated productive activities, as well as the shared consumption of the spoils, must have constitutedan important social tie among households. For the Great Basin, Bettinger (1999) has argued that theintroduction of the bow and arrow around 1500 BP had dramatic and far-reaching effects on theorganization of production. Specifically, he proposes that with the switch to individual hunting and itsaccompanying higher returns, the meat of larger game may still have been shared, but there would havebeen a reduction in the social pressures to share less valued resources, including plants. Hence, “thesocial relations of production were transformed from a system in which all resources were treated aspublic goods, to one in which some resources, notably plant resources...were regarded as privateproperty” (Bettinger 1999:73).

As noted above (see also Pluckhahn et al. 2006), there is evidence for such communal huntingin the faunal remains associated with the early Late Woodland archaeological household in Block A, inthe form of fill layers with high MNI for white-tailed deer and a disproportionate representation of themeatier cuts of deer. The pattern is similarly for Block D, but here the faunal remains are dispersedacross a number of features and a longer time interval, consistent with repeated individual hunting, withmany of these forays extending some distance from the community and thus necessitating return withonly select cuts.

I suggest that the social costs associated with the adoption of the bow and arrow may havediscouraged the adoption of this technology at Kolomoki while supra-household institutions were stillstrong, in the Middle and early Late Woodland periods. However, as households began to assert greaterautonomy over production and consumption during the late Late Woodland, new technologiessupporting more individualized hunting would have represented a more attractive option forhouseholds. It is worth emphasizing here that I am reversing the order of causality in the relationshipbetween the bow and arrow and household autonomy as discussed by previous authors authors (e.g.,Bettinger 1999; Muller 1997:127). Given that arrows form a decided minority of the points in BlockD, even while there is evidence for increased household autonomy in other aspects of material culturediscussed above (from storage to ceramics), it would appear that households chose to adopt the newtechnology only after they had achieved greater independence from the supra-household institutionsthat bound them together in the Middle and early Late Woodland periods.

Thus, by around cal A.D. 800-850, households at Kolomoki appear to have made many of thesame decisions as Late Woodland households elsewhere in the Midwest and Southeast: they removedthemselves from formalized collective ritual, developed greater storage capacity, and adopted newtechnologies that permitted greater efficiency in hunting (McElrath et al. 2000). Like households

205

everywhere they were still embedded in larger social networks, but compared with households of theMiddle Woodland period they exercised greater autonomy over their own production and consumption.

However, late and terminal Late Woodland households at Kolomoki differed from those in theAmerican Bottom region in several key respects, and these differences help to explain the divergenttrajectories societies in these two regions would take over the next two centuries. First, Late Woodlandhouseholds at Kolomoki chose not to intensify the effort devoted to horticulture. Maize pollen andstarch grains from the Block D household attest to continued maize cultivation at Kolomoki, but onlyin modest quantities judging from the absence of macobotanical remains. Indeed, maize cultivation mayhave even declined at Kolomoki during the late and terminal Late Woodland. The pattern is notunusual for the Late Woodland period in the Weeden Island area; maize was also uncommon atSycamore (Milanich 1974) and not represented at all at Woodland Terrace (Mickwee 2009). A reviewof the archaeological literature by Ashley and Rolland (2009) suggests that maize was not present inpeninsular Florida until Middle Mississippian times. In contrast, in the American Bottom region, theubiquity of maize increased to around 40 percent by around cal A.D. 800/900 (Fortier and McElrath2002; McElrath et al. 2000:18).

The choice not to step up the effort devoted to farming probably played a crucial role in twoother points of divergence in the trajectories taken by late/terminal Late Woodland households atKolomoki relative to those in the American Bottom. Perhaps most immediate, it meant that householdsat Kolomoki remained less residentially stable; they probably regularly dispatched task groups and mayhave shifted residence seasonally or in longer term cycles. Koldehoff and Galloy (2005) suggest thatPatrick phase (A.D. 650 to 900) households in the American Bottom were also less sedentary than manyarchaeologists have assumed. Here, however, sedentism probably increased with the intensification ofmaize production after cal A.D. 900 (McElrath et al. 2000:18).

The residential instability of Late Woodland households at Kolomoki and elsewhere in theregion was probably an impediment to population increase. At the same time, it may have provided anincentive to keep extended kin together in a single household, to accommodate the demands ofscheduling extended foraging trips. At Kolomoki and other sites in the region, co-resident groupsappear to have become larger, as indicated by the greater size of houses, as well as by an increase in theaverage size of ceramic vessels. They may have begun more commonly incorporating extended familiesunder a single roof (Peregrine 1992; Steere 2011:72-73). Larger, extended family households are favoredunder circumstances where the demands of scheduling are great, because if parents are away there areother adults in residence who can assist with childcare and other household duties (Pasternak andEmber 1976; see also Wilk and Rathje 1982; Williams and Williams 1965). Constraints on organizationmay also favor such large households where community-level institutions are weak (Johnson 1982).

The larger households at Kolomoki and sites in its immediate region do not appear to have beenunusual. Steere (2010) has recently completed a comprehensive review of domestic architecture fromarchaeological sites across the Southeast. For the Late Woodland period, he notes that sites in theeastern portion of the Southeast are relatively large and widely spaced (Steere 2010:72-73). As arguedhere for Kolomoki, he credits this pattern to the prevalence of extended family households.

As households at Kolomoki and throughout much of the region became larger in the late andterminal Late Woodland, those in the American Bottom and a few other areas may have become moreattenuated (Steere 2010:72-73). Fortier and McElrath (2002) have reviewed evidence for domestic

206

architecture and community organization in the American Bottom region during the terminal LateWoodland period; they document far greater diversity than is accounted for in simple models. Nevertheless, it is clear that the predominant pattern during the terminal Late Woodland consisted ofvery small houses, sometimes occurring in relative isolation and other times clustered around smallcourtyards (Fortier and McElrath 2002; Kelly et al. 1990; Steere 2010). Although the arrangement ofsome of these small houses in clusters suggests some degree of sharing in household activities, thepresence of internal hearths and storage pits in many of the houses (Kelly et al. 1990) suggests they alsofunctioned independently in some household tasks. This is the type of pattern one would expect witha mixture of independent, nuclear family households and multiple-family household groups, in whichthe conjugal couple maintained some degree of economic autonomy and thus perhaps could beconsidered separate households (Wilk 1988:139).

Of course, the diversity of houses in the terminal Late Woodland American Bottom suggeststhat households took a variety of forms, a pattern which ethnographic accounts suggest is not unusualfor societies more generally (Wilk 1988). Still, I argue that the nuclear family emerged as what Wilk(1988:137) has referred to as “normative household;” that is, a form held as an ideal, if frequentlysubject to creative manipulation and variation based on economics and the domestic developmentalcycle. It is also true that households do not necessarily act with a single voice or in concert (Blanton1995; Hendon 1996; Wilk 1989; Yanagisako 1979). At Kolomoki, for example, if we assume thatceramics and food storage were primarily the domain of women, the rapid change in these facets ofmaterial culture relative to hafted bifaces might suggest that the female members of households weremaking many of the key decisions discussed above. However, we need not deny internal differencesin interest or authority to show that households can employ effective strategies when faced with certainopportunities (Hartman 2004:95).

Various authors have implicated the emergence of nuclear family households as a keycomponent of the Mississippian transformation in the American Bottom. However, these authors havetended to view this household form as a development that followed the initial push towards aMississippian lifestyle, whether that push is conceptualized in terms that are evolutionary and materialist(Peregrine 1992) or historic and political (Pauketat 1998:135-136). In contrast, I suggest that thedevelopment of smaller, more residentially stable, and more autonomous nuclear family householdsconstituted a “prior and distinctive development” (Hartman 2004) that permitted the rapidentrenchment of a new Mississippian social and political order in the American Bottom after the A.D.1050 event. Conversely, the absence or relative uncommonness of this household form in theKolomoki area helps explain why the a Mississippian lifestyle did not develop in this region until aroundA.D. 1100-1200, when it was introduced by migrants from elsewhere in the Mississippianized Southeast(Blitz and Lorenz 2006:136-137).

Why might the development of this form of household been key to the Mississippiantransformation in the American Bottom? Peregrine (1992) argues that the sort of smaller householdsthat were typical of the area promoted competition, thus fostering intensification of maize productionand ultimately greater social stratification. This may be true even if the logic is not clearly articulatedby Peregrine; Wilk and Rathje (1982:623-624) observe that intensified agricultural production favorssmaller households, since more can be produced per worker from a given unit of land (as long as thereis no need for larger land modifications such as irrigation systems, which would require larger taskgroups).

207

Nevertheless, I suggest that the development of this household form had other, moresubtle—yet potentially equally important—ramifications for the process of Mississippianization. First,because they are smaller, nuclear households are more physically mobile (Hartman 2004:105-106; Wilkand Rathje 1982:632). In the case of the development of the modern nation state, the nuclearhouseholds of western Europe were better able to position themselves with respect to emergingemployment opportunities (Hartman 2004). In the case of the Mississippian event, the householdswould have been able more easily relocate as the geo-political landscape shifted, first with thereorganization of space in the “Big Bang” at Cahokia, then with the founding of new communities onCahokia’s periphery, and ultimately with the migration of Mississippian households away from theAmerican Bottom.

Next, the structure of smaller, nuclear family households may have made them more amenableto new and novel forms of socio-political organization. Hartman (2004:105-106) makes this case inregard to the pattern of nuclear households that took hold in northern and western Europe in the yearspreceding the development of modern nation states. She notes that this system:

...enables and encourages greater flexibility of response to new situations, inviting acertain creativity. As the sole married adults in residence, couples...do not live daily withthe accumulated weight of generations of practice passed down within the husband’sfamily. Neither spouse is obliged to engage in what may become decades of deferenceto the wishes of the resident elder generation.

In contrast, the extended household pattern more common in eastern and southern Europe provideda strategic advantage for scheduling household tasks, but emphasized tradition over innovation, sincemarried couples face long apprenticeships in household management under members of an oldergeneration (Hartman 2004:105). But as Hartman further notes, the nuclear form of household alsocomes with a substantial price, in that with only one married adult couple in residence, there is nocushion of support in regard to the scheduling of household chores or the provisioning of thehousehold in the event of death or disability of one of the conjugal pair (see also Laslett 1988). Insocieties where nuclear families predominate, this relative insecurity in the face of weakened familial tiesis often mitigated by increased emphasis on collective social and political institutions. Extrapolatingthese observations to the Mississippian example, the nuclear family households of the American Bottommay have been more amenable to life under new forms of political authority and within larger-scalesocial formations, even if by doing so they would eventually---and largely unwittingly—forego some ofthe autonomy they had gained over the course of the Late Woodland (Pauketat 2000).

Conclusion

McElrath and colleagues (2000:23) have noted that “by viewing the Late Woodland simply asa logical evolutionary step between Middle Woodland and Mississippian...evolutionists have forced thisperiod into a stereotypic tribal transition to chiefship and...have muddied the water for interpreting allthree periods.” Their call for a more historical view has been answered with recent scholarship positingthat the structural transformations at the transitions between the Middle and early Late Woodland (Carrand Case 2008:28; McElrath et al. 2000:14-16; Yerkes 1988:1) and again between the terminal LateWoodland and Early Mississippian (Beck et al. 2007; Pauketat 1994, 1997a, 1997b, 2004a, 2004b, 2007)

208

are best viewed as historical events. I agree with these approaches, but argue that viewing these asdiscrete historical events reinforces the notion of the intervening Late Woodland period as a relativelystatic or “uneventful” interval (e.g., Griffin 1952:361-362).

I suggest instead that, following Nassaney and Cobb, “...the undoing of Middle Woodland andall it entailed is simultaneously the emergence of Mississippian...” (Nassaney and Cobb 1991:314, emphasisin original). Specifically, the events that bounded the Late Woodland period were interpenetrated bya series of ruptures that reorganized structures associated with the organization of households and thedomestic economy. First, the collapse of public ceremony and dispersal of settlement near the end ofthe Middle Woodland period reflect fundamental changes to the resources and schemas by whichcommunities and extended kin networks were conceived and constituted. The subsequentreorganization of domestic economy—as indicated most dramatically by increases in storage andproliferation of new biface forms—manifest shifts in resources and schemas regarding the sharing oflabor and resources. While still embedded in wider social and political relationships, late Late Woodlandhouseholds were less beholden to community and extended kin, and exercised greater autonomy of theirown production and consumption.

These changes were fundamental to the “undoing” of the Middle Woodland. But the emergenceof Mississippian would depend on an additional rupture in the schemas and resources associated withdomestic economy, including the very manner in which family and household were defined. Thischange was less pervasive geographically, however. Late and terminal Late Woodland households atKolomoki and across much of the Southeast chose not to invest heavily in the growing of maize. Relatedly, they mostly retained an extended household structure that was well-suited to the schedulingdemands of a relatively sedentary, primarily hunting-gathering lifestyle. This was a household structurethat favored the continuation of existing cultural practices and traditions by keeping generationstogether.

In contrast, by around cal A.D. 850, households in the American Bottom began to intensify thecultivation of maize. In the process, they appear to have more frequently adopted a smaller form ofhousehold that was more efficient for horticultural production. These nuclear households werepotentially more mobile. Moreover, lacking both the support and constraints afforded by having anolder generation in residence, they may have been more favorably inclined to the new and novel formsof collective social and political organization that arose in the American Bottom around A.D. 1050.

I do not argue that these changes in households directly caused either the collapse of MiddleWoodland ceremonialism or the A.D. 1050 “Big Bang.” However, transformations in domesticeconomy were clearly crucial to both events. Variation in the timing and intensity of changes atKolomoki and across the region suggest that these transformations were not simply the result of eitherevolutionary process or disembodied historical forces, but were instead filtered through the agency ofhouseholds. In this sense, the households of the Late Woodland period were truly “making history.”

209

210

References Cited

Anderson and Mainfort (editors)2002 The Woodland Southeast. University of Alabama Press, Tuscaloosa.

Andrefsky, William, Jr.1998 Lithics: Macroscopic Approaches to Analysis. Cambridge Manuals in Archaeology. Cambridge

University Press, Cambridge.

Asch, Nancy B. and David L. Asch1985 Prehistoric Plant Cultivation in West-Central Illinois. In Prehistoric Food Production in North

America, Anthropological Papers No. 75, edited by Richard I. Ford, pp. 149-204. Museum ofAnthropology, University of Michigan, Ann Arbor.

Ashley, Keith, and Vicki Rolland2009 Where is the Corn in Precolumbian Peninsular Florida? Paper presented at the 74th annual

meeting of the Society for American Archaeology, Atlanta.

Atran, Scott, Ximena Lois and Edilberto Ucan Ek'2003 Plants of the Peten Itza' Maya. Memoirs, Number 38. University of Michigan, Museum of

Anthropology, Ann Arbor, Michigan.

Bacon, W.S.1977 Projectile Point Typology: The Basic Base. Archaeology of Eastern North America 5:107-121.

Baker, Winston H.1995 A Hypothetical Classification of Some of the Flaked Stone Projectiles, Tools and Ceremonials from the

Southeastern United States. Williams Printing, Quincy, Massachusetts.

Beck, Robin A., Jr., Douglas J. Bolender, James A. Brown, and Timothy K. Earle2007 Eventful Archaeology: The Place of Space in Structural Transformation. Current Anthropology

48(6)833-860.

Behler, John L. and F. Wayne King1997 National Audubon Society Field Guide to North American Reptiles and Amphibians. Alfred A. Knopf,

New York.

Bettinger, Robert L.1999 What Happened in the Medithermal. In Models for the Millennium, edited by C. Beck, pp. 62-74.

University of Utah Press, Salt Lake City.

Binford, Lewis R.1963 A Proposed Attribute List for the Description and Classification of Projectile Points. In

Miscellaneous Studies in Typology and Classification, edited by Anta M. White, Lewis R. Binford, and Mark L. Papworth, pp. 193-221. Anthropological Papers, Museum of Anthropology, University of Michigan, No. 19., Ann Arbor.

211

1967 Smudge Pits and Hide Smoking: The Use of Analogy in Archaeological Reasoning. AmericanAntiquity 32(1):1-12.

1978 Nunamiut Ethnoarchaeology. Academic Press, New York.

Black, Bryan, H. Thomas Foster, II, and Marc Abrams2002 Combining Environmentally Dependent and Environmentally Independent Analysis of Witness

Tree Data in East-Central Alabama. Canadian Journal of Forestry Research 32: 2060-2075.

Blakney-Bailey, Jane Anne2008 An Analysis of Seminole Artifacts from the Paynes Town Site (8AL366), Alachua County,

Florida. The Florida Anthropologist 61(3-4): 167-188.

Blanton, Richard E.1995 The Cultural Foundations of Inequality in Households. In Foundations of Social Inequality, edited

by T. Douglas Price and Gary M. Feinman, pp. 105-127. Plenum Press, New York.

Blitz, John H.1988 Adoption of the Bow in Prehistoric North America. North American Archaeologist 9:123-145.

2010 New Perspectives in Mississippian Archaeology. Journal of Archaeological Research 18(1):1-39.

Blitz, John H., and Karl G. Lorenz2006 The Chattahoochee Chiefdoms. University of Alabama Press, Tuscaloosa.

Bolender, Douglas J. (editor)2010 Eventful Archaeologies: New Approaches to Social Transformation in the Archaeological Record. State

University of New York Press, Albany.

Bonhage-Freund, Mary Theresa1997 Paleoethnobotany of the Georgia Piedmont: Four Lamar Period Farmsteads in the Middle

Oconee Uplands. Unpublished Ph. D. dissertation, Department of Anthropology, ThePennsylvania State University, University Park.

1998 Paleoethnobotanical Analysis of the Kolomoki Site (9ER1). LEAF Archaeological Research Services,Farmington Hills, Michigan. Report submitted to the LAMAR Institute, Watkinsville, GA.

2000 Kolomoki 2000. Intermet Associates, Alma, Michigan. Report submitted to the Department ofAnthropology, University of Georgia, Athens.

2003 Final Paleoethnobotanical Analysis of Flotation Samples from the Kolomoki 2001 Project (9ER1). IntermetAssociates, Alma, Michigan. Report submitted to the Department of Anthropology, Universityof Georgia, Athens.

2005 An Experimental Approach to the Analysis of Two Maize Cob-Filled Features from Etowah(9BR1). Early Georgia 33(2):131-141.

212

2010 Paleoethnobotanical Analysis of Flotation Samples, NGS II Kolomoki Project (9ER1). IntermetAssociates, Alma, Michigan. Submitted to the Department of Anthropology, University of SouthFlorida, Tampa.

Branch-Raymer, Leslie, and Mary Theresa Bonhage-Freund2011 Middle Woodland Gardening in the Etowah River Valley, Northwest Georgia: The Hardin

Bridge Site (9BR34). Early Georgia, in preparation.

Breedlove, Dennis E., and Robert M. Laughlin2000 The Flowering of Man: A Tzotzil Botany of Zinacantan. Abridged edition. Smithsonian Institution

Press, Washington, DC.

Britton, Nathaniel Lord, and Hon. Addison Brown1970 [1913] Illustrated Flora of the Northern United States and Canada. Second edition. Dover

Publications, Inc., New York.

Bull, John, and John Farrand, Jr.1996 National Audubon Society Field Guide to North American Birds. Alfred A. Knopf, New York.

Bullen, Ripley P.1975 A Guide to the Identification of Florida Projectile Points. Revised edition. Kendall Books, Gainesville,

Florida.

Caldwell, Joseph1958 Trend and Tradition in the Prehistory of the Eastern United States. American Anthropological

Association Memoir 88, Salt Lake City.

Cambron, J.W., and D.C. Hulse1978 Handbook of Alabama Archaeology: Part I Point Types. Alabama Archaeological Society, Huntsville.

Cannon, Michael D.2001 Archaeofaunal Relative Abundance, Sample Size, and Statistical Methods. Journal of Archaeological

Science 28:185-195.

Carr, Christopher2006a Salient Issues in the Social and Political Organizations of Northern Hopewellian Peoples:

Contextualizing, Personalizing, and Generating Hopewell. In Gathering Hopewell: Society, Ritual,and Ritual Interaction, edited by Christopher Carr and D. Troy Case, pp. 73-118. Springer, NewYork.

2006b Scioto Hopewell Ritual Gatherings: A Review and Discussion of Previous Interpretations andData. In Gathering Hopewell: Society, Ritual, and Ritual Interaction, edited by Christopher Carr andD. Troy Case, pp. 463-479. Springer Science+Business Media, Inc., New York.

2008 Social and Ritual Organization. In The Scioto Hopewell and Their Neighbors: BioarchaeologicalDocumentation and Cultural Understanding, edited by D. Troy Case and Christopher Carr, pp. 151-288. Springer Science+Business Media, Inc., New York.

213

Carr, Christopher, and D. Troy Case2008 Documenting the Lives of Ohio Hopewell People: Philosophical and Empirical Foundations.

In The Scioto Hopewell and Their Neighbors: Bioarchaeological Documentation and Cultural Understanding,edited by D. Troy Case and Christopher Carr, pp. 3-34. Springer Science+Business Media, Inc.,New York.

Cobb, Charles R., and Patrick H. Garrow1996 Woodstock Culture and the Question of Mississippian Emergence. American Antiquity 61(1):21-

37.

Cobb, Charles R., and Michael S. Nassaney2002 Domesticating Self and Society in the Woodland Southeast. In The Woodland Southeast, edited

by David G. Anderson and Robert C. Mainfort, Jr., pp. 525-539. University of Alabama Press,Tuscaloosa.

Collins, Michael B. 1975 Lithic Technology as a Means of Processual Inference. In Lithic Technology: Making and Using

Stone Tools, edited by Earl Swanson, pp. 15–34. Mouton Publishers, The Hague.

Compton, J. Matthew2009 Vertebrate Fauna from the Kolomoki (9ER1) Block D Excavations, Early County, Georgia. Report

submitted to the Department of Anthropology, University of South Florida, Tampa.

DeBoer, Warren1988 Subterranean Storage and the Organization of Surplus: the View from Eastern North America.

Southeastern Archaeology 7(1):1-20.

DeJarnette, David L., Edward B. Kurjack, and James W. Cambron1962 Excavations at the Stanfield-Worley Bluff. Journal of Alabama Archaeology 8(1-2):1-124.

Delcourt, H.R., and P.A. Delcourt1977 Presettlement Magnolia-Beech Climax of the Gulf Coastal Plain: Quantitative Evidence from

the Apalachicola River Bluffs, North Central Florida. Ecology 58: 1085-1093.

Delorit, R.J.1970 Illustrated Taxonomy of Weed Seeds. Agronomy Publications, River Falls, Wisconsin.

Dunnell, Robert C., and Diana M. Greenlee1999 Late Woodland Period “Waste” Reduction in the Ohio River Valley. Journal of Anthropological

Archaeology 18:376–395.

Emerson, Thomas E., Dale L. McElrath, and Andrew C. Fortier (editors)2000 Late Woodland Societies: Tradition and Transformation Across the Midcontinent. University of Nebraska

Press, Lincoln.

214

Emerson, Thomas E., and Timothy R. Pauketat2002 Embodying Power and Resistance at Cahokia. In The Dynamics of Power, edited by M.

O’Donovan, pp. 105-125. Occasional Paper No. 30, Center for Archaeological Investigations,Southern Illinois University, Carbondale.

Ernst, C. H. and R. W. Barbour1972 Turtles of the United States. University Press of Kentucky, Lexington.

Espenshade, Christopher T., Linda Kennedy, William F. Stanyard, and David S. Leigh1998 The Prehistoric Occupation of the Shoal Creek Reservoir Basin: Data Recovery Investigations at 9HY95,

9HY98, and 9HY104 in Henry County, Georgia. TRC Cultural Resource Group, Atlanta, Georgia.Prepared for Clayton County Water Authority, Morrow, Georgia.

Fairbanks, Charles H.1962 Excavations at Horseshoe Bend, Alabama. The Florida Anthropologist 15(2):41-56.

Fernald, M.L.1950 Gray's Manual of Botany. 8 ed. American Book Company, New York, New York.

Flannery, Kent V.1972 The Origins of the Village as a Settlement Type in Mesoamerica and the Near East: A

Comparative Study. In Man, Settlement, and Urbanism, edited by P. Ucko, R. Tringham, and G.Dimbleby, G., pp. 23-53. Duckworth, London.

2002 The Origins of the Village Revisited: From Nuclear to Extended Households. AmericanAntiquity 67: 417-433.

Flannery, Kent V. (editor)1976 The Early Mesoamerican Village. Academic Press, Orlando.

Fortier, Andrew C., and Dale L. McElrath2002 Deconstructing the Emergent Mississippian Concept: The Case for the Terminal Late

Woodland in the American Bottom. Midcontinental Journal of Archaeology 27(2):172-215.

Foster, H. Thomas, II, Bryan Black, and Marc D. Abrams2004 A Witness Tree Analysis of the Effects of Native American Indians on the Pre-European

Settlement Forests in East-Central Alabama. Human Ecology 32: 27-47.

Foster, Steven, and James A. Duke1990 A Field Guide to Medicinal Plants: Eastern and Central North America. Houghton Mifflin Company,

Boston.

Fritz, Gayle J., and Tristram R. Kidder1993 Recent Investigations into Prehistoric Agriculture in the Lower Mississippi Valley. Southeastern

Archaeology 12(1): 1-14.

Fritz, Gayle J., Virginia Drywater Whitekiller, and James W. McIntosh2001 Ethnobotany of Ku-Nu-Che: Cherokee Hickory Nut Soup. Journal of Ethnobiology 21(2): 1-27.

215

Gerritsen, F.2004 Archaeological Perspectives on Local Communities. In A Companion to Archaeology, edited by J.

Bintliff, pp. 141-154. Blackwell Publishing, Malden, Massachusetts.

Giddens, Anthony1984 The Constitution of Society: Outline of a Theory of Structuration. University of California Press,

Berkeley.

Gillespie, Susan D.2007 Comments on “Eventful Archaeology: The Place of Space in Structural Transformation” by

Robin A. Beck, Jr., Douglas J. Bolender, James A. Brown, and Timothy K. Earle. CurrentAnthropology 48(6)846-847.

Gilman, Patricia1987 Architecture as Artifact: Pit Structures and Pueblos in the American Southwest. American

Antiquity 52: 538-564.

Godfrey, R.K.1988 Trees, Shrubs, and Woody Vines of Northern Florida and Adjacent Georgia and Alabama. University of

Georgia Press, Athens.

Goody, Jack1958 The Fission of Domestic Groups among the Lodagaba. In The Development Cycle of Domestic

Groups, edited by J. Goody, pp. 53-91. Cambridge University Press, Cambridge.

Gordon, Elizabeth A.1993 Screen Size and Differential Faunal Recovery: A Hawaiian Example. Journal of Field Archaeology

20(4):453-460.

Gould, S. J.1966 Allometry and Size in Ontogeny and Phylogeny. Biological Review of the Cambridge Philosophical

Society 41:587-640.

1971 Geometric Similarity in Allometric Growth: A Contribution to the Problem of Scaling in theEvolution of Size. The American Naturalist 105(942):113-137.

Grayson, Donald K.1973 On the Methodology of Faunal Analysis. American Antiquity 38(4):432-439.

1981 The Effects of Sample Size on Some Derived Measures in Vertebrate Faunal Analysis. Journalof Archaeological Science 8(1):77-88.

1984 Quantitative Zooarchaeology. Academic Press, New York.

Gremillion, Kristen1985 The Development and Dispersal of Agricultural Systems in the Woodland Period Southeast.

In The Woodland Southeast, edited by David G. Anderson and Robert C. Mainfort, Jr., pp 483-501. The University of Alabama Press, Tuscaloosa.

216

1998 Changing Roles of Wild and Cultivated Plant Resources Among Early Farmers of EasternKentucky. Southeastern Archaeology 17(2):140-157.

2002 The Development and Dispersal of Agricultural Systems in the Woodland Period Southeast. In The Woodland Southeast, edited by D. G. Anderson and R. C. Mainfort, Jr. pp. 483-501. University of Alabama Press, Tuscaloosa.

2003 Eastern Woodlands Overview. In People and Plants in Eastern North America, edited by Paul E.Minnis, pp. 17-49. Smithsonian Institution Press, Washington, D.C.

Griffin, James B.1952 Culture Periods in Eastern United States Archaeology. In Archaeology of the Eastern United States,

edited by J.B. Griffin, pp352-364. University of Chicago Press, Chicago.

Hall, Robert1979 In Search of the Ideology of Adena-Hopewell Climax. In Hopewell Archaeology: The Chillicothe

Conference, edited by D.S. Brose and N. Greber, pp. 258-265. Kent State University Press, Kent,Ohio.

1980 An Interpretation of the Two-Climax Model of Illinois Prehistory. In Early Native Americans:Prehistoric Demography, Economy, and Technology, edited by David Broman, pp. 401-462. Mouton,The Hague.

1997 Archaeology of the Soul: North American Indian Belief and Ritual. University of Illinois Press, Urbana.

Hally, David J.2009 Late Woodland to Mississippian Vessel Assemblage Change as a Response to Changing

Foodways. Paper presented at the Annual Meeting of the Society for American Archaeology,Atlanta.

Hamel, Paul B. and Mary U. Chiltoskey1975 Cherokee Plants and Their Uses--A 400 Year History. Herald Publishing Co., Sylva, North Carolina. Hartman, Mary S.2004 The Household and the Making of History: A Subversive View of the Western Past. Cambridge University

Press, New York.

Heiser, Charles B., Jr.1985 Some Botanical Considerations of the Early Domesticated Plants North of Mexico. In Prehistoric

Food Production in North America, Anthropological Papers No. 75, edited by Richard I. Ford, pp.19-56. Museum of Anthropology, University of Michigan, Ann Arbor.

Hendon, Julia A. 1996 Archaeological Approaches to the Organization of Domestic Labor: Household Practice and

Domestic Relations. Annual Review of Anthropology 25: 45-61.

2000 Having and Holding: Storage, Memory, Knowledge, and Social Relations. American Anthropologist102(1): 42-53.

217

Hickey, Michael and Clive J. King1981 100 Families of Flowering Plants. Cambridge University Press, Cambridge.

Hodler, Thomas W., and Harold A. Schretter1986 The Atlas of Georgia. The Institute of Community and Area Development, University of Georgia,

Athens.

Hudson, Charles1976 The Southeastern Indians. University of Tennessee Press, Knoxville.

Hurt, R. Douglas1987 Indian Agriculture in America, Prehistory to the Present. University Press of Kansas, Lawrence, Kansas.

Jenkins, Ned J., and Blaine Ensor1981 The Gainesville Lake Area Excavations. Archaeological Investigations of the Gainesville Lake Area, Volume

1. Report of Investigations 11. Office of Archaeological Research, University of Alabama,Moundville.

Johnson, Gregory A.1982 Organizational Structure and Scalar Stress. In Theory and Explanation in Archaeology, edited by C.

Renfrew, M. Rowlands and B. Segraves, pp. 389-421. Academic Press, New York.

Johnson, Matthew H.2006 On the Nature of Theoretical Archaeology and Archaeological Theory. Archaeological Dialogues

13(2):117-132.

Kay, Daisy E.1987 Root Crops. Second edition. Tropical Development and Research Institute, London.

Keel, Bennie C.1987 Cherokee Archaeology: A Study of the Appalachian Summit. University of Tennessee Press, Knoxville.

Kelly, John E.1990a The Emergence of Mississippian Culture in the American Bottom Region. In The Mississippian

Emergence, edited by B.D. Smith, pp. 113-152. Smithsonian Institution Press, Washington, D.C.

1990b Range Site Community Patterns and the Mississippian Emergence. In The MississippianEmergence, edited by B.D. Smith, pp. 67-112. Smithsonian Institution Press, Washington, D.C.

Kelly, John E., Andrew C. Fortier, Stephen J. Ozuk, and Joyce A. Williams1987 The Range Site: Archaic Through Late Woodland Occupations. American Bottom Archaeology FAI-

270 Site Reports, Volume 16. University of Illinois Press, Urbana.

Kennedy, Helen2000 Marantaceae. In Flora of North America North of Mexico, Volume 22, edited by Flora of North

America Editorial Committees, pp. 315-319. Oxford University Press, New York.

218

Kent, Susan1981 The Dog: An Archaeologist's Best Friend or Worst Enemy- the Spatial Distribution of Faunal

Remains. Journal of Field Archaeology 8(3):367-372.

King, Chester1987 Ethnohistoric Reconstruction of Subsistence-Settlement Systems in the Vicinity of Burton Mesa.

Manuscript on file with the author.

Kiple, Kenneth F. and Kriemhild Conee Ornelas (editors)2000 The Cambridge World History of Food. Volume 2. Cambridge University Press, Cambridge.

Knight, Vernon J., Jr.1990 Excavation of the Truncated Mound at the Walling Site: Middle Woodland Culture and Copena in the

Tennessee Valley. Report of Investigations 56. Division of Archaeology, Alabama State Museumof Natural History, University of Alabama, Tuscaloosa.

2001 Feasting and the Emergence of Platform Mound Ceremonialism in Eastern North America. In Feasts: Archaeological and Ethnographic Perspectives on Food, Politics, and Power, edited by MichaelDietler and Brian Hayden, pp. 311-333. Smithsonian Institution Press, Washington, DC.

Kohler, Timothy A.1991 The Demise of Weeden Island, and Post-Weeden Island Cultural Stability, in

Non-Mississippianized Northern Florida. In Stability, Transformation, and Variation: The LateWoodland Southeast, edited by M.S. Nassaney and C.R. Cobb, pp. 91-110. Plenum Press, NewYork.

Koldehoff, Brad, and Joseph M. Galloy2005 Late Woodland Frontiers in the American Bottom Region. Southeastern Archaeology 25:275-300

Lopinot, Neal H.1983 Analysis of Flotation Sample Materials from the Late Archaic Horizon. In The 1982 Excavations

at the Cahokia Interpretive Center Tract, St. Clair County, Illinois, by Richard W. Jefferies, pp. 77-108. Research Paper No. 37, Center for Archaeological Investigations, Southern Illinois University,Carbondale.

Lyman, R. Lee1994 Vertebrate Taphonomy. Cambridge University Press, Cambridge.

Lyman, R. Lee, and Kenneth M. Ames2004 Sampling to Redundancy in Zooarchaeology: Lessons from the Portland Basin, Northwestern

Oregon and Southwestern Washington. Journal of Ethnobiology 24(2):329-346.

MacNeish, Richard S.1992 The Origins of Agriculture and Settled Life. University of Oklahoma Press, Norman, Oklahoma.

219

McElrath, Dale L., Thomas E. Emerson, and Andrew C. Fortier2000 Social Evolution or Social Response? A Fresh Look at the "Good Gray Cultures" After Four

Decades of Midwest Research. In Late Woodland Societies: Tradition and Transformation across theMidcontinent, edited by T.E. Emerson, D.L. McElrath, and A.C. Fortier, pp. 3-36. University ofNebraska Press, Lincoln.

McGee, Harold1984 On Food and Cooking. Charles Scribner's Sons, New York, New York.

Marcoux, J.B.2008 Cherokee Households and Communities in the English Contact Period, A.D. 1670-1740. Unpublished

Ph.D. dissertation, Department of Anthropology, University of North Carolina, Chapel Hill.

Markin, Julie G.2010 The Importance of Woodstock Complicated Stamped Fifty Years After Caldwell. Early Georgia

38(2):169-186.

Martin, Alexander C., and William D. Barkley1973 Seed Identification Manual. University of California Press, Berkeley.

Marquardt, William H. 2010 Mounds, Middens, and Rapid Climate Change During the Archaic-woodland Transition in the

Southeastern United States. In Trend, Tradition, and Turmoil: What Happened to the SoutheasternArchaic?, edited by D.H. Thomas and M.C. Sanger, pp. pp. 253-271. Proceedings of the ThirdCaldwell Conference, St. Catherines Island, Georgia, May 9-11, 2008. Anthropological Papersof the American Museum of Natural History, No. 93, New York.

Medsger, Oliver Perry1966 Edible Wild Plants. MacMillan Publishing Company, Inc., Racine, Wisconsin.

Mickwee, Christopher L.2009 Wakulla in the Sandhills: Analysis of a Late Weeden Island Occupation in the Northwest Florida

Interior Uplands. Paper presented at the Annual Meeting of the Florida AnthropologicalSociety, Pensacola.

Milanich, Jerald T.1974 Life in a 9th Century Indian Household, a Weeden Island Fall-Winter Site on the Upper Apalachicola River,

Florida. Florida Bureau of Historic Sites and Properties Bulletin 4:1-44.

1994 Archaeology of Precolumbian Florida. University Press of Florida, Gainesville.

1998 Florida's Indians from Ancient Times to the Present. University Press of Florida.

2002 Weeden Island Cultures. In The Woodland Southeast, edited by D.G. Anderson and R.C. Mainfort,Jr., pp. 352-372. The University of Alabama Press, Tuscaloosa.

220

Milanich, Jerald T., Ann S. Cordell, Vernon J. Knight, Jr., Timothy A. Kohler, and Brenda J.Sigler-Lavelle1997 Archaeology of Northern Florida, A.D. 200-900. Academic Press, New York. Originally published

1984.

Miller, Naomi1988 Ratios in Paleoethnobotanical Analysis. In Current Paleoethnobotany, edited by C. Hastorf and V.

Popper, pp. 72-85. University of Chicago Press, Chicago.

Minnis, Paul E. and Richard I. Ford1977 Analysis of Plant Remains from Chimney Rock Mesa. In Archaeological Investigations at Chimney

Rock Mesa: 1970-1972, edited by F.W. Eddy, pp. 81-91. Memoir 1. Colorado ArchaeologicalSociety, Denver.

Moerman, Daniel E.1998 Native American Ethnobotany. Timber Press, Portland, Oregon.

Montgomery, F.H.1974 Seeds and Fruits of Plants of Eastern Canada and the Northeastern United States. University of Toronto

Press.

Muller, Jon1997 Mississippian Political Economy. Plenum Press, New York.

Narrol, R.1962 Floor Area and Settlement Population. American Antiquity 27:587-89.

Nash, D. J.2009 Household Archaeology in the Andes. Journal of Archaeological Research 17: 205-261.

Nassaney, Michael S.1992 Communal Societies and the Emergence of Elites in the Prehistoric American Southeast. In

Lords of the Southeast: Social Inequality and the Native Elites of Southeastern North America, edited byA. Barker and T. Pauketat, pp. 111-143. Archaeological Papers of the AmericanAnthropological Association Number 3, Washington, D.C.

2000 The Late Woodland Southeast. In Late Woodland Societies: Tradition and Transformation across theMidcontinent, edited by T.E. Emerson, D.L. McElrath, and A.C. Fortier, pp. 713-30. Universityof Nebraska Press, Lincoln.

2001 A Historical-Processual Development of Late Woodland Societies. In The Archaeology ofTraditions: Agency and History Before and After Columbus, edited by T.R. Pauketat, pp. 157-173. University Press of Florida, Gainesville.

Nassaney, Michael S., and Charles R. Cobb (editors)1991 Stability, Transformation, and Variation: The Late Woodland Southeast. Plenum Press, New York.

221

Nassaney, Michael S., and Charles R. Cobb1991 Patterns and Processes of Late Woodland Development in the Greater Southeastern United

States. In Stability, Transformation, and Variation: The Late Woodland Southeast, edited by M.S.Nassaney and C.R. Cobb, pp. 285-322. Plenum Press, New York.

Nassaney, Michael S., and Kendra Pyle1999 The Adoption of the Bow and Arrow in Eastern North America: A View from Central

Arkansas. American Antiquity 64(2):243-263.

Northwest Florida Environmental Conservancy2006 Steepheads. Electronic document, http://www.nwflec.com/northwestfloridaenvironmental

conservancypart2/id12.html, accessed 2010.

Ortiz, Beverly R. and Julia F. Parker1991 It Will Live Forever: Traditional Yosemite Indian Acorn Preparation. Heyday Books, Berkeley.

Panshin, A.J. and Carl de Zeeuw1980 Textbook of Wood Technology. Fourth edition. McGraw-Hill, New York.

Parry, William J., and Robert L. Kelly1987 Expedient Core Technology and Sedentism. In The Organization of Core Technology, ed. J. K.

Johnson and C. A. Morrow, pp. 285–304. Westview Press, Boulder, Colorado.

Pasternak, Burton, Carol R. Ember, Melvin Ember1976 On the Conditions Favoring Extended Family Households. Journal of Anthropological Research

32(2):109-123.

Pauketat, Timothy R. 1994 The Ascent of Chiefs: Cahokia and Mississippian Politics in Native North America. University of

Alabama Press, Tuscaloosa.

1997a Cahokian Political Economy. In Cahokia: Domination and Ideology in The Mississipian World, editedby T.R. Pauketat and T.E. Emerson, pp. 30-51. University of Nebraska Press, Lincoln.

1997b Mississippian Ups and Downs. Book review essay of Political Structure and Change in the PrehistoricSoutheastern United States (edited by J. F. Scarry) and Mississippian Communities and Households (editedby J. D. Rogers and B. D. Smith). American Anthropologist 99: 634-636.

2000 Politicization and Community in the Precolumbian Mississippi Valley. In The Archaeology ofCommunities: A New World Perspective, edited by M.A. Canuto J. Yaeger, pp. 16-43. Routledge,London.

2004a Ancient Cahokia and the Mississippians. Cambridge University Press, Cambridge.

2004b the Economy of the Moment: Cultural Practices and Mississippian Chiefdoms. In ArchaeologicalPerspectives on Political Economies, edited by G.M. Feinman and L.M. Nicholas, pp. 25-39. University of Utah Press, Salt Lake City.

222

2007 Chiefdoms and Other Archaeological Delusions. AltaMira Press, Lanham, Maryland.

Pauketat, Timothy R., and Susan M. Alt2003 Mounds, memory, and contested Mississippian history. In Archaeologies of Memory, edited by R.M.

Van Dyke and S.E. Alcock, pp. 151-179. Blackwell Publishing, Malden, Massachusetts.

2005 Agency in a Postmold? Physicality and the Archaeology of Culture Making. Journal ofArchaeological Method and Theory 12: 213-236.

Payne, S.1972 Partial Recovery and Sample Bias: The Results of Some Sieving Experiments. In Papers in

Economic Prehistory, edited by E. S. Higgs, pp. 49-63. Cambridge University Press, Cambridge.

Pearsall, Deborah M.1989 Paleoethnobotany: A Handbook of Procedures. Academic Press, San Diego.

Peregrine, P. N.1992 Social Change in the Woodland-mississippian Transition: A Case Study of Household and

Community Patterns in the American Bottom. North American Archaeologist 13: 131-147.

Perkins, D., Jr., and P. Daly1968 A Hunter's Village in Neolithic Turkey. Scientific American 219(5):96-106.

Peterson, Lee A.1977 Edible Wild Plants. Collier Books, New York, New York.

Pilkinton, Jerry A.1985 Soil Survey of Calhoun and Early Counties, Georgia. United States Department of Agriculture, Soil

Conservation Service, Washington, DC.

Pluckhahn, Thomas J.1998 Kolomoki Revisited: The 1998 Field Season. Publication 46. LAMAR Institute, Watkinsville,

Georgia.

2000 Fifty Years Since Sears: Deconstructing the Domestic Sphere at Kolomoki. SoutheasternArchaeology 19(2):145-155.

2003 Kolomoki: Settlement, Ceremony, and Status in the Deep South, A.D. 350-750. The University ofAlabama Press, Tuscaloosa.

2010a "Gulfization" Revisitied: Household Change in the Late Woodland Period at Kolomoki (9ER1). Early Georgia 38(2):207-220.

2010b Household Archaeology in the Southeastern United States: History, Trends, and Challenges. Journal of Archaeological Research 18:331-385.

223

Pluckhahn, Thomas J., J. Matthew Compton and Mary Theresa Bonhage-Freund2006 Evidence of Small-Scale Feasting from the Woodland Period Site of Kolomoki, Georgia. Journal

of Field Archaeology 31(3):263-284.

Pluckhahn, Thomas J., and Sean P. Norman2011 Typological, Functional, and Comparative Contextual Analyses of Woodland Hafted Bifaces

from Kolomoki (9ER1). The Florida Anthropologist, in preparation.

Price, George D.1999 A Case for Increasing Sedentism in the Middle Woodland Prehistory of East Central Alabama.

Unpublished M.A. Thesis, Department of Anthropology, University of Alabama-Birmingham,Birmingham.

Purdue, James R.1983 Epiphyseal Closure in White-Tailed Deer. Journal of Wildlife Management 47(4):1207-1213.

Reidhead, Van A.1981 A Linear Programming Model of Prehistoric Subsistence Optimization: A Southeastern Indiana

Example. Prehistory Research Series VI (1). Indiana Historical Society, Indianapolis, Indiana.

Reimer, P. J., Baillie, M. G. L., Bard, E., Bayliss, A., Beck, J. W., Bertrand, C. J. H., Blackwell, P. G.,Buck, C. E., Burr, G. S., Cutler, K. B., Damon, P. E., Edwards, R. L., Fairbanks, R. G., Friedrich, M.,Guilderson, T. P., Hogg, A. G., Hughen, K. A., Kromer, B., McCormac, F. G., Manning, S. W., Ramsey,C. B., Reimer, R. W., Remmele, S., Southon, J. R., Stuiver, M., Talamo, S., Taylor, F. W., van der Plicht,J., and Weyhenmeyer, C. E.2004 IntCal04 Terrestrial Radiocarbon Age Calibration, 26 - 0 ka BP. Radiocarbon 46:1029-1058.

Reitz, Elizabeth J., I. R. Quitmyer, H. S. Hale, S. J. Scudder, and E. S. Wing1987 Application of Allometry to Zooarchaeology. American Antiquity 52(2):304-317.

Reitz, Elizabeth J. and Elizabeth S. Wing1999 Zooarchaeology. Cambridge University Press, Cambridge.

Richie, William A.1961 A Typology and Nomenclature for New York Projectile Points. Bulletin No. 384, New York State

Museum and Science Service, Albany.

Robin, Cynthia2003 New Directions in Classic Maya Household Archaeology. Journal of Archaeological Research 11:

307-356.

Rodning, Christopher B.2004 The Cherokee Town at Coweeta Creek. Unpublished Ph.D. dissertation, Department of

Anthropology, University of North Carolina, Chapel Hill.

224

Rogers, J. D.1995 Dispersed Communities and Integrated Households: A Perspective from Spiro and the Arkansas

Basin. In Mississippian Communities and Households, edited by J.D. Rogers B.D. Smith, pp. 81-98.University of Alabama Press, Tuscaloosa,

Romain, William F.2000 Mysteries Of The Hopewell. University Of Akron Press, Akron

2009 Shamans of the Lost World: A Cognitive Approach to the Prehistoric Religion of the Ohio Hopewell. Altamira Press, Lanham, Maryland.

Rudolph, Teresa P.1991 The Late Woodland "Problem" in North Georgia. In Stability, Transformation and Variation: The

Late Woodland Southeast, edited by M. S. Nasaaney and Charles R. Cobb, pp. 259-283. PlenumPress, New York.

Russo, Michael, Margo Schwadron, and Emily M. Yates2006 Archeological Investigation of the Bayview Site (8BY137), A Weeden Island Ring Midden, Tyndall Air Force

Base, Panama City, Florida. National Park Service, Southeast Archeological Center, Tallahassee. Submitted to Tyndall Air Force Base, Panama City, Florida.

Sabloff, Jeremy A., and Wendy Ashmore2001 An Aspect of Archaeology’s Recent past and its Relevance in the New Millenium. In Archaeology

at the Millennium: A Sourcebook, edited by G.M. Feinman and T.D. Price, pp. 11-32. KluwerAcademic/Plenum Publishers, New York.

Sassaman, Kenneth E., Mark J. Brooks, Glen T. Hanson, and David G. Anderson1990 Native American Prehistory of the Middle Savannah River Valley: A Synthesis of Archaeological Investigations

on the Savannah River Site, Aiken and Bamwell Counties, South Carolina. Savannah RiverArchaeological Research Papers 1. Occasional Papers of the Savannah River ArchaeologicalResearch Program, South Carolina Institute of Archaeology and Anthropology, University ofSouth Carolina, Columbia.

Scarry, Margaret2003 Patterns of Wild Plant Utilization in the Prehistoric Eastern Woodlands. In People and Plants in

Eastern North America, edited by Paul E. Minnis, pp. 50-104. Smithsonian Institution Press,Washington, D.C.

Schroder, Lloyd E.2006 The Anthropology of Florida Points and Blades (Revised). Cafe Press, San Mateo, California

Sears, William H.1951 Excavations at Kolomoki: Season I – 1948. University of Georgia Press, Athens.

1956 Excavations at Kolomoki: Final Report. University of Georgia Press, Athens.

225

Seeman, Mark F.1992 The Bow and Arrow, The Intrusive Mound Complex, and a Late Woodland Jack’s Reef Horizon

in the Mid-Ohio Valley. In Cultural Variability in Context: Woodland Settlements of the Mid-OhioValley, edited by Mark F. Seeman, pp. 41-51. Midcontinental Journal of Archaeology SpecialPublication No. 7, Kent State University Press, Kent, Ohio.

Sewell, William H., Jr.2005 Logics of History: Social Theory and Social Transformation. The University of Chicago Press, Chicago.

Shaffer, Brian S., and Julia L. J. Sanchez1994 Comparison of 1/8"- and 1/4"-inch Mesh Recovery of Controlled Samples of

Small-to-Medium-Sized Mammals. American Antiquity 59(3):525-530.

Shelby, Thomas M. 2011 The Cobbs Swamp Phase and the Middle Woodland Period of East-Central Alabama: The View

from Catoma Creek. Early Georgia, in preparation.

Shott, Michael J.1997 Stones and Shaft Redux: The Metric Discrimination of Chipped-Stone Dart and Arrow Points.

American Antiquity 62(1):86-101.

Sievert, A.K., and K. Wise2001 A Generalized Technology for a Specialized Economy: Archaic Period Chipped Stone at

Kilometer 4 Perú. In Lithic Debitage: Context, Form, Meaning, edited by W. Andrefsky, Jr., pp. 80-105. University of Utah Press, Salt Lake City.

Simpson, George Gaylord1941 Large Pleistocene Felines of North America. American Museum Novitates 1136:1-27.

Simpson, George G., A. Roe, and R. C. Lewontin1960 Quantitative Zoology. Harcourt, Brace, and Co., New York.

Smith, Karen Y.2009 Middle and Late Woodland Period Cultural Transmission, Residential Mobility, and Aggregation

in the Deep South. Unpublished Ph.D. dissertation, Department of Anthropology, Universityof Missouri, Columbia.

Snow, Frankie, and Keith Stephenson1992 Salvage Excavations at Hartford (9PU1): A 4th Century Swift Creek Site on the Ocmulgee River

in Pulaski County, Georgia. Draft manuscript in possession of the authors.

Souvatzi, Stella G.2008 A Social Archaeology of Households in Neolithic Greece: An Anthropological Approach. Cambridge

University Press, New York.

Spector, Janet1993 What This Awl Means: Feminist Archaeology at a Wahpeton Dakota Village. Minnesota Historical

Society Press, St. Paul.

226

Springer, James W.1980 Analysis of Prehistoric Food Remains from the Bruly St. Martin Site, Louisiana, with a

Comparative Discussion of Mississippi Valley Faunal Studies. Midcontinental Journal of Archaeology5: 193-224.

Stahle, Peter W.1996 The Recovery and Interpretation of Microvertebrate Bone Assemblages from Archaeological

Contexts. Journal of Archaeological Method and Theory 3(1):31-75.

Steere, Benjamin A.2011 The Archaeology of Houses and Households in the Native Southeast. Unpublished Ph.D.

dissertation, Department of Anthropology, the University of Georgia, Athens.

Struever, Stuart1968 Flotation Techniques for the Recovery of Small-scale Archaeological Remains. American

Antiquity 33(3):353-362.

Stuiver, M., and Reimer, P. J.1993 Extended 14C Database and Revised CALIB Radiocarbon Calibration Program. Radiocarbon

35:215-230.

Swanton, John R.1969 [1946] Indians of the Southeastern United States. United States Government Printing Office,

Washington, D.C.

Thomas, David Hurst1971 On Distinguishing Natural from Cultural Bone in Archaeological Sites. American Antiquity

36:366-371.

1978 Arrowheads and Atlatl Darts: How the Stones Got the Shaft. American Antiquity 43(3):461-472.

United States Department of Agriculture1974 Seeds of Woody Plants in the United States. USDA Forest Service, Agriculture Handbook 45. United

States Department of Agriculture, Washington, D.C.

University of South Alabama2004 Geoarchaeology at South Alabama. Electronic document, http://www.geoarchaeology.south

alabama.edu/index.html, accessed July, 2011.

Veatch, O. and L.W. Stephenson 1911 Geology of the Coastal Plain. Bulletin 26, Geological Survey of Georgia, Atlanta.

Wagner, Gail E.1981 Testing Flotation Recovery Rates. American Antiquity 47: 127-132.

Wallis, Neill J.2011 The Swift Creek Gift: Vessel Exchange on the Atlantic Coast. University of Alabama Press,

Tuscaloosa.

227

Wauchope, Robert1966 Archaeological Survey of Northern Georgia. Society for American Archaeology Memoir 21, Salt lake

City.

Webb, Malcum C.1982 The Neutral Calorie? On the Maintenance of Ranked Societies in the "Agriculturally Deficient

Environment of Gulf Coastal Louisiana. Louisiana Archaeology 8:1-19.

Wesson, Cameron B.1999 Chiefly Power and Food Storage in Southeastern North America. World Archaeology 31(1):145-

164.

2008 Households and Hegemony: Early Creek Prestige Goods, Symbolic Capital, and Social Power. Universityof Nebraska Press, Lincoln.

Wharton, Charles H.1978 The Natural Environments of Georgia. Geological and Water Resources Division, and Resource

Planning Section, Office of Planning and Research, Georgia Department of Natural Resources,Atlanta.

Whatley, John S.2002 An Overview of Georgia Projectile Points and Selected Cutting Tools. Early Georgia 30(1).

Whitaker, John O., Jr.1997 National Audubon Society Field Guide to North American Mammals. Alfred A. Knopf, New York.

White, Nancy Marie1981 Archaeological Survey at Lake Seminole, Jackson and Gadsden Counties, Florida, Seminole and

Decatur Counties, Georgia. Cleveland Museum of Natural History Archaeological ResearchReport No. 29.

White, T. E.1953 Method of Calculating the Dietary Percentages of Various Food Animals Utilized by Aboriginal

Peoples. American Antiquity 19(2):396-398.

Wiessner, P.1974 A Functional Estimator of Population from Floor Area. American Antiquity 39:343-50.

Wilk, Richard W. 1983 Little House in the Jungle: The Causes of Variation in House Size among Modern Kekchi Maya.

Journal of Anthropological Archaeology 2: 99-116.

1988 Maya Household Organization: Evidence and Analogies. In Household and Community in theMesoamerican Past, edited by R. Wilk and W. Ashmore, pp. 135-152. University of New MexicoPress, Albuquerque.

228

1989 Decision Making and Resource Flows Within the Household: Beyond the Black Box. In TheHousehold Economy: Reconsidering the Domestic Mode of Production, edited by R.W. Wilke, pp. 23-52. Westview Press, Boulder.

Wilk, Richard R., and William L. Rathje1982 Household Archaeology. American Behavioral Scientist 25(6):617-639.

Willey, Gordon R.1949 Archeology of the Florida Gulf Coast. Smithsonian Miscellaneous Collections 113, Smithsonian

Institution, Washington, DC.

1966 An Introduction to American Archaeology, Volume One: North and Middle America. Prentice-Hall, Inc.,Englewod Cliffs, New Jersey.

Willoughby, Lynn2007 Flowing Through Time: The Chattahoochee Throughout History. Sherpa Guides, Lenz,

Georgia. http://www.sherpaguides.com/georgia/chattahoochee/. Accessed 27 May 2009.

Wilson, Gregory D.2008 The Archaeology of Everyday Life at Early Moundville, University of Alabama Press, Tuscaloosa.

Yanagisako, Sylvia1979 Family and Household: The Analysis of Domestic Groups. Annual Review of Anthropology 8: 161-

205.

Yarnell, Richard A.1964 Aboriginal Relationships Between Culture and Plant Life in the Upper Great Lakes Region.

Anthropological Papers, Museum of Anthropology, University of Michigan, No. 23. TheUniversity of Michigan, Ann Arbor, Michigan.

1978 Domestication of Sunflower and Sumpweed in Eastern North America. In The Nature and Statusof Ethnobotany, Anthropological Papers No. 67, edited by Richard I. Ford, pp. 298-300. Museumof Anthropology, University of Michigan, Ann Arbor.

Yarnell, Richard A., and M. Jean Black1983 Temporal Trends Indicated by a Survey of Prehistoric Plant Food Remains from Southeastern

North America. Paper presented at the 40th Annual Meeting of the SoutheasternArchaeological Conference, Columbia, South Carolina.

Yerkes, R. W.1988 Introduction. In Interpretations of Culture Change in the Eastern Woodlands During the Late Woodland

Period, edited by R. W. Yerkes, pp. 1–6. The Ohio State University, Columbus.

Yost, Chad, and Linda Scott-Cummings2010 Phytolith and Pollen Analysis of Pit Features from the Kolomoki Mounds Site (9er1), Southwestern Georgia.

PaleoResearch Institute, Golden, Colorado. Prepared for Department of Anthropology,University of Southern Florida, Tampa, Florida

229

Zomlefer, Wendy B.1994 Guide to Flowering Plants Families. The University of North Carolina Press, Chapel Hill and

London.

230

Pluckhahn, Thomas J. 2011Households Making History: Household Change in the Late Woodland Period at Kolomoki (9ER1). Department of Anthropology, University of South Florida, Tampa.

Documents

earlymiddle late woodland

woodland period occupation

block d area

excavation of block

kolomoki pluckhahn

course of excavating

test unit test unit

kolomoki 9er1