Calupoh Estandar

S e asonalit y and
H um an M obilit y along
the G eorgia B ight
ELIZABETH J. REITZ,
IRVY R. QUITMYER, AND
DAVID HURST THOMAS
Editors and Contributors
American Museum of Natural History Anthropological Papers, Number 97

Scientific Publications of the American Museum of Natural History
American Museum Novitates
Bulletin of the American Museum of Natural History
Anthropological Papers of the American Museum of Natural History
Publications Committee
Robert S. Voss, Chair
Board of Editors
Jin Meng, Paleontology
Lorenzo Prendini, Invertebrate Zoology
Robert S. Voss, Vertebrate Zoology
Peter M. Whiteley, Anthropology
Managing Editor
Mary Knight
Submission procedures can be found at http://research.amnh.org/scipubs
All issues of Anthropological Papers of the American Museum of Natural History

are available on the web from
http://digitallibrary.amnh.org/dspace
Order printed copies from http://www.amnhshop.com or via standard mail from

American Museum of Natural History—Scientific Publications,
Central Park West at 79th St.,
New York, NY 10024
∞ This paper meets the requirements of ANSI/NISO Z39.48-1992 (permanence of paper).

2012 SEASONALITY AND HUMAN MOBILITY ALONG THE GEORGIA BIGHT 1
SEASONALITY AND HUMAN MOBILITY

ALONG THE GEORGIA BIGHT
ELIZABETH J. REITZ, IRVY R. QUITMYER,

AND
DAVID HURST THOMAS

EDITORS AND CONTRIBUTORS
WITH CONTRIBUTIONS BY
C. F red T. A ndrus , S arah G. B ergh, N icole R. Cannarozzi,
G wendolyn D. C arroll, Carol E. Colaninno, Brendan J.
C ulleton, Chester B. D eP ratter, K andace D. H ollenbach,
Douglas S. J ones, D eborah A nn K eene, D ouglas J. K ennett,
C harles W. L ambert, J.W. M oak , Bruce M. S aul, C. M argaret
S carry, G regory A. W aselkov, AND Elizabeth S. W ing
Proceedings of the Fifth Caldwell Conference

St. Catherines Island, Georgia
May 14–16, 2010
ANTHROPOLOGICAL PAPERS OF
THE AMERICAN MUSEUM OF NATURAL HISTORY
Number 97, 236 pages, 80 figures, 30 tables
Issued March 6, 2012
Copyright © American Museum of Natural History 2012

ISSN 0065-9452
2 ANTHROPOLOGICAL PAPERS AMERICAN MUSEUM OF NATURAL HISTORY NO. 97
CONTENTS
Abstract ................................................................................................................................................ 10
Participants in the Fifth Caldwell Conference ..................................................................................... 11
Preface. Elizabeth J. Reitz, Irvy R. Quitmyer, and David Hurst Thomas ......................................... 13
Acknowledgments ............................................................................................................................... 16
Chapter 1. Seasonality and mobility on the Georgia Bight: why should we care?
David Hurst Thomas ...................................................................................................................... 19
What’s the “Guale problem”? ......................................................................................................... 19
Conflicting views ........................................................................................................................ 19
Competing hypotheses and logical consequences ...................................................................... 20
Archaeology and the Guale problem .............................................................................................. 21
Developing multiscalar chronologies ............................................................................................. 23
The St. Catherines Island ceramic chronology: reading the hour hand....................................... 23
Radiocarbon dating: reading the minute hand ............................................................................ 23
Site seasonality: reading the second hand ................................................................................... 25
Seasonality and the Guale problem ............................................................................................ 27
But did the Jesuits get it wrong? .................................................................................................... 28
Some conclusions and implications ............................................................................................... 29
The St. Catherines Island research design .................................................................................. 31
Potential problems with proxy populations ................................................................................ 33
Chapter 2. A Bayesian chronological framework for determining site seasonality

and contemporaneity. Douglas J. Kennett and Brendan J. Culleton ......................................... 37
Bayesian essentials ......................................................................................................................... 37
Building site chronologies .............................................................................................................. 38
Trimming confidence intervals ....................................................................................................... 39
Toward building a Bayesian model for the McQueen Shell Ring .................................................. 41
A note on precision 14C and seasonality studies ............................................................................. 43
Contemporaneity and Archaic period shell rings ........................................................................... 44
Conclusions .................................................................................................................................... 49
Chapter 3. Interpreting seasonality from modern and archaeological fishes

on the Georgia Coast. Elizabeth J. Reitz, Bruce M. Saul, J.W. Moak,
Gwendolyn D. Carroll, and Charles W. Lambert ....................................................................... 51
Estuaries and fishes ........................................................................................................................ 52
Background premises ..................................................................................................................... 56
Methods .......................................................................................................................................... 56
Caveats ........................................................................................................................................... 61
Results ............................................................................................................................................ 64
Modern South Beach and Cumberland Sound data .................................................................... 64
Archaeological data .................................................................................................................... 71
Comparison ................................................................................................................................. 71
Discussion ...................................................................................................................................... 77
Conclusion ...................................................................................................................................... 81
Chapter 4. Evaluating δ18O profiles of hardhead catfish and Atlantic croaker otoliths
as a method of determining seasonal use of fishes. Carol E. Colaninno ...................................... 83
Methods .......................................................................................................................................... 84
δ18Ootolith fractionation factors and estuarine environments ............................................................ 86
3
δ18O of modern otoliths .................................................................................................................. 89

δ18O of archaeological otoliths ....................................................................................................... 91
The Cannon’s Point Ring ............................................................................................................ 92
The West Shell Ring ................................................................................................................... 92
The McQueen Shell Ring ........................................................................................................... 92
The St. Catherines Shell Ring ..................................................................................................... 92
Summary ..................................................................................................................................... 92
δ18Ootolith, estuarine environments, and archaeological implications .............................................. 92
Modern δ18Ootolith: problems and areas of future research ........................................................... 92
Archaeological implications ....................................................................................................... 97
Summary and conclusions .............................................................................................................. 98
Appendix 4.1. Modern otoliths sampled for isotopic analysis ....................................................... 99
Appendix 4.2. Archaeological otoliths sampled for isotopic analysis ......................................... 100
Chapter 5. Late prehistoric settlement patterns: zooarchaeological evidence from Back

Creek Village, St. Catherines Island. Sarah G. Bergh ................................................................ 103
Seasonality in faunal assemblages ............................................................................................... 103
The “Guale problem” ................................................................................................................... 104
Materials and methods .................................................................................................................. 105
Results .......................................................................................................................................... 108
Discussion ..................................................................................................................................... 117
Conclusion .................................................................................................................................... 121
Chapter 6. Molluscs as oxygen-isotope season-of-capture proxies in southeastern

United States archaeology. C. Fred T. Andrus ............................................................................. 123
Discussion .................................................................................................................................... 125
Oyster (Crassostrea virginica) ................................................................................................. 125
Quahog, or hard clam (Mercenaria spp.) ................................................................................... 127
Coquina clam (Donax variabilis) ............................................................................................. 128
Bay scallop (Argopecten irradians) .......................................................................................... 128
Ribbed mussel (Geukensia demissa) ........................................................................................ 129
Stout razor clam (Tagelus plebeius) .......................................................................................... 130
Marsh clam (Rangia cuneata and Polymesoda caroliniana) .................................................... 130
Freshwater mussels (Unionidae) ............................................................................................... 131
Gastropods ................................................................................................................................ 132
Conclusions .................................................................................................................................. 132
Chapter 7. Annual incremental shell growth patterns in hard clams (Mercenaria spp.)
from St. Catherines Island, Georgia: a record of seasonal and anthropogenic
impact on zooarchaeological resources. Irvy R. Quitmyer and Douglas S. Jones ..................... 135
Methods ........................................................................................................................................ 136
Modern hard clams ................................................................................................................... 136
Results .......................................................................................................................................... 139
Modern cycle of incremental shell growth ............................................................................... 139
Zooarchaeological hard clams—season of resource procurement ........................................... 141
Population dynamics—modern ontogenetic age and survivorship .......................................... 141
Population dynamics—zooarchaeological mean age and survivorship ................................... 142
Discussion .................................................................................................................................... 143
Conclusion .................................................................................................................................... 148
Chapter 8. Validation of annual shell increments and shifting population dynamics

in modern and zooarchaeological hard clams (Mercenaria mercenaria)
2012 SEASONALITY AND HUMAN CONTENTS
MOBILITY ALONG THE GEORGIA BIGHT 5
from the Litchfield Beach Region, South Carolina. Douglas S. Jones,

Irvy R. Quitmyer, and Chester B. DePratter ............................................................................ 149
Methods ........................................................................................................................................ 151
Archaeological sites .................................................................................................................. 151
Litchfield—modern proxy ........................................................................................................ 151
Isotopic validation of incremental shell growth patterns .......................................................... 154
Ontogenetic age ......................................................................................................................... 154
Results .......................................................................................................................................... 155
Observed incremental shell growth .......................................................................................... 155
Isotopic validation ..................................................................................................................... 155
Size vs age ................................................................................................................................. 157
Mean ontogenetic age and survivorship ................................................................................... 157
Discussion .................................................................................................................................... 157
Conclusions .................................................................................................................................. 164
Chapter 9. Reevaluating the use of impressed odostome (Boonea impressa) as a season

of capture indicator for oysters. Deborah Ann Keene .................................................................. 165
Location and environment ............................................................................................................ 165
Reproduction ................................................................................................................................ 166
Inconsistent growth patterns within populations .......................................................................... 167
Feeding habits ............................................................................................................................... 168
Current state of knowledge concerning Boonea impressa ........................................................... 169
Chapter 10. Estimating the season of harvest of the eastern oyster (Crassostrea
virginica) from the St. Catherines Shell Ring. Nicole R. Cannarozzi ........................................ 171
Oyster seasonality studies: a review ............................................................................................. 172
Visual methods .......................................................................................................................... 172
Geochemical methods ............................................................................................................... 173
Seasonality determination using impressed odostomes (Boonea impressa) ............................. 174
Oyster habitat on St. Catherines Island ........................................................................................ 174
Methods ........................................................................................................................................ 174
Stable isotope analysis .............................................................................................................. 174
Morphometric analysis of shell shape ...................................................................................... 176
Shell length measurements using impressed odostomes .......................................................... 177
Results .......................................................................................................................................... 177
Stable isotope geochemistry ..................................................................................................... 177
Morphometric analysis of oyster shell shape ............................................................................ 179
Shell length measurements of impressed odostomes ................................................................ 179
Discussion .................................................................................................................................... 181
Conclusions .................................................................................................................................. 185
Chapter 11. What can plants and plant data tell us about seasonality?
C. Margaret Scarry and Kandace D. Hollenbach .................................................................... 187
Using plants in analyses of seasonal resource use and mobility .................................................. 187
Optimal foraging models .......................................................................................................... 187
Operational chains .................................................................................................................... 189
Seasonal plant use in the Georgia Bight ....................................................................................... 193
Spring ........................................................................................................................................ 194
Summer ..................................................................................................................................... 197
Fall ............................................................................................................................................ 197
Winter ....................................................................................................................................... 197
Final thoughts ............................................................................................................................... 197
Chapter 12. Making a case for coastal subsistence seasonality. Gregory A. Waselkov .................. 199
Habitat analysis ............................................................................................................................ 203
Statistical analysis ........................................................................................................................ 205
Chapter 13. Discussion. Elizabeth S. Wing ...................................................................................... 207
References .......................................................................................................................................... 211
TABLES
1.1. Comparison of northern Georgia coast and St. Catherines Island chronologies ......................... 24
2.1. Radiocarbon dates from McQueen Shell Ring, N243E233 ......................................................... 40
3.1. Summary of archaeological sites and vertebrate collections ....................................................... 57
3.2. Seasonal total length ranges for high-ubiquity taxa from South Beach ...................................... 63
3.3. Total catch by year for North Beach and South Beach ................................................................ 64
3.4. Mean catch per unit effort by year for North Beach and South Beach ....................................... 65
3.5. Fishes with ubiquity ≥ 0.25, North Beach and South Beach data combined .............................. 65
3.6. Fishes collected from South Beach with ubiquity ≥ 0.25 ........................................................... 66
3.7. Seasonal ubiquity of fishes collected from South Beach ........................................................... 68
3.8. Comparison of archaeological and modern South Beach fish diversity ..................................... 68
3.9. Ubiquity of fishes identified in some Georgia Bight archaeological collections ........................ 72
4.1. Modern hardhead catfish otolith season of capture and estimated season of capture ................ 90
4.2. Modern Atlantic croaker otolith season of capture and estimated season of capture ................. 90
4.3. Archaeological otolith estimated season of capture ................................................................... 93
5.1. Fish size regression formulas ................................................................................................... 109
5.2. Back Creek Village, Midden A: species list ............................................................................... 110
5.3. Back Creek Village, Midden B: species list ............................................................................... 111
5.4. Back Creek Village, Midden C: species list ............................................................................... 112
5.5. Back Creek Village, Midden D: species list .............................................................................. 113
5.6. Back Creek Village, Midden F: species list ............................................................................... 114
5.7. Back Creek Village, Midden G: species list .............................................................................. 115
5.8. Back Creek Village, Midden H: species list .............................................................................. 116
5.9. Back Creek Village: summary .................................................................................................... 117
7.1. Modern and zooarchaeological survivorship data of Mercenaria spp. sampled
from St. Catherines Island, St. Simons Island, and Kings Bay, Georgia .................................... 146
8.1. Radiocarbon dates for hard clams (M. mercenaria) from Litchfield
Beach shell middens ..................................................................................................................... 153
8.2. Descriptive statistics of ontogenetic age of modern and zooarchaeological
hard clams from the Litchfield Beach region .............................................................................. 153
9.1. Size classes formulated by Russo (1991) ................................................................................... 167
10.1. Seasonal divisions and sample sizes for modern and archaeological impressed odostomes ....... 179
11.1. Decisions, activities, and artifacts associated with hickory nut use ........................................ 190
11.2. Plants identified from sites along the Georgia and Carolina coasts ........................................ 195
FIGURES
[Preface] Participants in the Fifth Caldwell Conference .................................................................... 12

1.1. The randomized transect research design employed in the island-wide
survey of St. Catherines Island ...................................................................................................... 22
1.2. Interpolated estimates of incremental growth stages for the modern control
sample of Mercenaria collected from St. Catherines Island ......................................................... 26
1.3. Position of growth surface within major increments at time of harvest: modern
control sample of Mercenaria collected between 1975 and 1984 on St. Catherines Island ......... 29
1.4. Map of probable seasonality on St. Catherines Island during the Irene period .......................... 30
2.1. Hypothetical seasonality and settlement data .............................................................................. 39
2.2. Calibrated results of a sequence of two 14C dates on hickory nuts from
McQueen Shell Ring using OxCal 3.01 ........................................................................................ 42
2.3. Calibrated results of a sequence of three 14C dates on Mercenaria shells
from McQueen Shell Ring using OxCal 3.01 ................................................................................ 42
2.4. Profile of the N243 trench north wall, McQueen Shell Ring, showing the
depositional sequence discussed in the text (courtesy of Matt Sanger) ........................................ 42
2.5. Schematic cross section of a hypothetical Mercenaria valve showing annual growth
increments and three 14C dates used in a defined sequence to wiggle match the terminal date ......... 45
2.6. Two sequences incorporating the most reliable 14C dates from the Archaic
in St. Catherines Shell Ring ........................................................................................................... 47
2.7. A phase of nested sequences incorporating the most reliable 14C dates
from the Archaic in McQueen Shell Ring ..................................................................................... 48
3.1. Location of sites mentioned in this study .................................................................................... 53
3.2. Monthly total catch, North and South beaches ............................................................................ 54
3.3. Map of St. Catherines Island, showing North and South beaches ............................................... 58
3.4. Aquatic biological sampling stations in Cumberland Sound (modified
from DEIS, 1978: D-187) .............................................................................................................. 59
3.5. Comparison of archaeological and South Beach fish diversity ................................................... 66
3.6. Abundance and biomass values for Cumberland Sound (DEIS, 1978: D-456) .......................... 69
3.7. Overall abundance for Cumberland Sound (DEIS, 1978: D-487) .............................................. 69
3.8. Seasonal percentages of fish biomass at Stations D and E, Cumberland
Sound (DEIS, 1978: D-458) .......................................................................................................... 69
3.9. Seasonal variations in fish diversity at Stations D and E, Cumberland
Sound (DEIS, 1978: D-489) .......................................................................................................... 70
3.10. Abundance of star drum (Stellifer lanceolatus) individuals per season .................................... 70
3.11. Comparison of South Beach and archaeological fish ubiquity .................................................. 76
3.12. Seasonal ubiquity of selected fish taxa in the South Beach collection ...................................... 76
3.13. Number of seasons in which fishes with archaeological ubiquity ≥ 0.68 are present
at Station E, Cumberland Sound (DEIS, 1978: D-438–D-445) .................................................... 77
3.14. Seasonal aspects of mean total length (in cm) of star drum (Stellifer lanceolatus)
in the South Beach collection ........................................................................................................ 78
4.1. Map of study area showing location of shell rings ...................................................................... 85
4.2. Cross section of hardhead catfish otolith, FLMNH-EA 11341 .................................................... 86
4.3. Cross section of Atlantic croaker otolith, FLMNH-EA 11319 .................................................... 87
4.4. Temperature and salinity measurements for Sapelo Island Marsh Landing ................................ 88
4.5. Annual cycle of temperature oscillation and salinity fluctuations for the McQueen
Inlet of St. Catherines Island, average and constant ...................................................................... 89
4.6. Annual cycle of temperature oscillation and salinity fluctuations for the McQueen
Inlet of St. Catherines Island, recorded ......................................................................................... 89
4.7. Modern hardhead catfish and Atlantic croaker δ18Ootolith ........................................................... 91
4.8. δ18O profile through ontogeny for selected archaeological otoliths ........................................... 95
4.9. Archaeological δ18Ootolith of an Atlantic croaker approximately 18.5 years of age ..................... 96
5.1. Location of Back Creek Village on St. Catherines Island, Georgia .......................................... 106
5.2. Topographic map of Back Creek Village showing the location of the
middens excavated in 2008 .......................................................................................................... 107
5.3. Richness, diversity, and equitability estimates for Back Creek Village, presented
for each midden and for the site as a whole ................................................................................. 118
5.4. Relative frequencies of final growth phases for hard clams (Mercenaria spp.)
from Back Creek Village, presented for each midden and for the site as a whole ...................... 119
5.5. Fish size estimates for selected fishes from Back Creek Village, presented for the
site as a whole and for each midden ............................................................................................. 120
6.1. Examples of taxa described in this chapter ................................................................................ 124

6.2. Idealized representation of seasonal δ18O oscillations in a mollusc through ontogeny ............ 125
6.3. Bisected oyster hinge in reflected light showing irregular alternating dense gray
and chalky white banding ............................................................................................................ 126
6.4. Bisected quahog valve in reflected light showing regular alternating light and dark bands ........ 128
6.5. Sequential δ18O data from R cuneata valve (0608.2) collected from
Chocolatta Bay where it enters Mobile Bay ................................................................................ 130
6.6. Archaeological Unionidae clam valve ....................................................................................... 131
7.1. Radial cross section of a hard clam (Mercenaria campechiensis) to expose
its growth increments and a back-lit thin section showing the alternating
opaque and translucent shell growth increments ......................................................................... 137
7.2. The location of the translucent and opaque shell growth structures showing a count
of the ontogenetic age of a specimen ........................................................................................... 138
7.3. Six-part division of translucent and opaque annual shell growth increments
in hard clams (Mercenaria spp.) shells ........................................................................................ 138
7.4. The relationship of sea surface temperature (°C) and monthly formation of translucent
and opaque growth in modern hard clams collected live from St. Catherines Island
and Kings Bay .............................................................................................................................. 140
7.5. Seasonal growth frequency patterns of modern hard clams (Mercenaria) collected
live from St. Catherines Island, 2007–2008 ................................................................................ 142
7.6. Growth frequency profiles of zooarchaeological hard clams collected
from St. Catherines Island and Cannon’s Point, representing the Archaic period ...................... 143
7.7. A comparison of shell length to ontogenetic age of modern and zooarchaeological
hard clams from St. Catherines Island ......................................................................................... 144
7.8. Survivorship curves characterizing modern and zooarchaeological hard clams
from St. Catherines Island, Cannon’s Point, and Kings Bay ....................................................... 144
7.9. Mean age and 95% confidence intervals plotted for modern and zooarchaeological
hard clams from St. Catherines Island, Cannon’s Point, and Kings Bay .................................... 145
8.1. Thin, radial cross section of hard clam (M. mercenaria) shell viewed under transmitted
and direct light ............................................................................................................................. 150
8.2. Location of the Litchfield Beach archaeological shell middens ................................................ 152
8.3. Monthly frequency of specimens forming translucent or opaque shell growth
increments from March 2005 to March 2007 in hard clams collected from the estuary
at Litchfield Beach ....................................................................................................................... 156
8.4. Percentage of translucent and opaque incremental shell growth identified in modern
and zooarchaeological hard clams from the Litchfield Beach region ......................................... 157
8.5. Variation in oxygen (18O/16O) and carbon (13C/12C) isotopic composition in the translucent
and opaque growth increments of modern hard clams from the Litchfield Beach region .......... 158
8.6. Oxygen (18O/16O) and carbon (13C/12C) isotopic composition of the translucent
and opaque growth increments from zooarchaeological hard clams (M. mercenaria)
from the Litchfield Beach region ................................................................................................. 159
8.7. Oxygen (18O/16O) and carbon (13C/12C) isotopic composition of the translucent and opaque
growth increments from zooarchaeological hard clams (M. mercenaria)
8.8. A comparison of age versus anterior-to-posterior (A/P) shell length in modern
hard clams (M. mercenaria) from the Litchfield Beach estuary ................................................ 161
8.9. The mean ontogenetic age and 95% confidence interval around the mean of modern
and zooarchaeological hard clams (M. mercenaria) from the Litchfield Beach region ............. 161
8.10. Survivorship curves of modern and zooarchaeological hard clams (M. mercenaria)
9.1. Drawing of Boonea impressa ..................................................................................................... 166
9.2. Length distribution data from Wells (1959) ............................................................................... 168
9.3. Season of capture for Grove’s Creek site shells using Russo method ...................................... 169
10.1. Map showing location of the St. Catherines Shell Ring and oyster collection site
(adapted from Thomas, 2008) ...................................................................................................... 175
10.2. Radial cross section of oyster showing sampling locations .................................................... 176
10.3. Left oyster valve height and length measurements for HLR calculation ................................ 177
10.4. Impressed odostome shell length measurement ....................................................................... 178
10.5. δ18O and δ13C profiles across oyster specimens from earliest to most recent growth ............ 180
10.6. Average monthly sea surface temperatures for the National Buoy Center Station,
41008 for 2006–2008 ................................................................................................................... 181
10.7. Salinity measurements for July 2006–July 2008 from Cemetery
Road Marsh collection site ........................................................................................................... 181
10.8. A plot of the mean and 95% confidence interval of the shell height-to-length ratio
(HLR) of oysters from modern, historic, and archaeological sites ............................................. 182
10.9. Frequency distribution of impressed odostomes length size classes from units
W83S2 and 789N801E plotted following Russo’s (1991) six-season division .......................... 182
10.10. Comparison of modern odostome collection data with archaeological measurements
on an annual scale, St. Catherines Island and Crescent Beach (Russo 1991) ......................... 183
11.1. Comparison of return rates for mulberries, hickory nuts, and deer ......................................... 188
11.2. Comparison of return rates for cutting and threshing versus hand-stripping
for chenopod and amaranth .......................................................................................................... 192
12.1. Results of a trial carbon and nitrogen isotope analysis of deer metapodial bone
collagen samples from two Late Woodland sites ........................................................................ 204
ABSTRACT
Some of the most enduring and fundamental questions in archaeology relate to site seasonality.
During which seasons did people occupy coastal archaeological sites? Why is “seasonality” important
to our understanding of human behavior? What does this knowledge tell us about life in dynamic es-
tuarine systems? What methods and technologies are available to address key issues of seasonality?
Archaeological seasonality is uniquely linked to settlement patterns, resource availability, environ-
mental relationships, anthropogenesis, landscapes, and social complexity.
Archaeologists working in coastal settings typically recover multiple biological proxies that are
well suited to explicating questions of human seasonal behavior. The Fifth Caldwell Conference was
convened to discuss and report on practiced methods for reading the seasonality record found in com-
mon biological proxies. These researchers spoke of how they are applying various methods grounded
in the natural sciences to estimate seasonality with particular reference to the archaeology of St. Cath-
erines Island and the Georgia Bight. These methods include stable isotope analysis, 14C dating, lon-
gitudinal studies of animals (molluscs and fishes), zooarchaeology, and archaeobotany. The research
shows that all plant and animal remains found in a midden contain a record of human behavior.
The authors of these 13 chapters agree that multiple indicators of site seasonality provide the
most robust picture of the annual settlement cycle. These papers were initially presented at the Fifth
Caldwell Conference, cosponsored by the American Museum of Natural History and the St. Cath-
erines Island Foundation, held on St. Catherines Island, Georgia, May 14–16, 2010.
10
PARTICIPANTS IN THE FIFTH CALDWELL CONFERENCE

ST. CATHERINES ISLAND, GEORGIA, MAY 14–16, 2010
C. Fred T. Andrus (Associate Professor, Department of Geological Sciences, University

of Alabama)
Sarah G. Bergh (Ph.D. student, Department of Anthropology, University of Georgia)
Elliot Blair (Ph.D. student, Department of Anthropology, University of California, Berkeley)
Rachel Cajigas (Material Culture Analyst, Nels Nelson North American Archaeology Laboratory,
American Museum of Natural History)
Nicole R. Cannarozzi (Ph.D. student, Department of Anthropology, University of Florida)
Carol E. Colaninno (Ph.D., Department of Anthropology, University of Georgia)
Christina Friberg (at the time, St. Catherines Island Assistant Archivist, American Museum
of Natural History)
Royce Hayes (Superintendent, St. Catherines Island)
Christa Hayes (St. Catherines Island)
Kandace D. Hollenbach (Ph.D., Archaeological Research Laboratory, Department of Anthropology,
University of Tennessee, Knoxville)
Douglas J. Kennett (Professor, Department of Anthropology, the Pennsylvania State University)
Matthew Napolitano (Assistant Laboratory Supervisor, Nels Nelson North American Archaeology
Laboratory, American Museum of Natural History)
Alexandra Parsons (Ph.D. student, Florida State University)
Lorann S.A. Pendleton (Director, Nels Nelson North American Archaeology Laboratory, American
Museum of Natural History)
Irvy R. Quitmyer (Environmental Archaeology Collection Manager, Florida Museum
of Natural History)
Elizabeth J. Reitz (Professor of Anthropology, Georgia Museum of Natural History,
University of Georgia)
David Hurst Thomas (Curator of Anthropology, American Museum of Natural History)
C. Margaret Scarry (Associate Professor and Director of Undergraduate Studies, Department
of Anthropology, University of North Carolina at Chapel Hill)
Gregory A. Waselkov (Professor, Department of Sociology, Anthropology, and Social Work,
University of South Alabama)
Elizabeth S. Wing (Curator Emeritus, Environmental Archaeology, Florida Museum
of Natural History)
11
Participants in the Fifth Caldwell Conference, standing in front of the Carriage House Archaeological
Laboratory on St. Catherines Island, May 14–16, 2010: (top row, left to right) Fred Andrus, Christina Friberg,
Sarah Bergh, Carol Colaninno-Meeks, Lori Pendleton, Margaret Scarry, Betsy Reitz, Matthew Napolitano,
Rachel Cajigas, and Elizabeth Wing; (bottom row, left to right) Kandace Hollenbach, Greg Waselkov, Alexandra
Parsons, Irvy Quitmyer, Royce Hayes, Christa Hayes, Doug Kennett, David Hurst Thomas, Nicole Cannarozzi,
and Elliot Blair.
12
PREFACE
E lizabeth J. R eitz , I rvy R. Q uitmyer , and D avid H urst T homas
Archaeological excavations at coastal sites vailing during that individual’s life or at the time
typically recover multiple biological proxies for of death. Judgments regarding season of avail-
seasonal behavior of plant and animal resources ability, procurement, and location (as defined
and of people. Questions of seasonality are em- by Monks, 1981) must rest on ecological anal-
bedded in much of archaeological research, with ogy, which typically requires monitoring mod-
answers linked to many aspects of cultures and ern patterns to correlate hard tissue growth and
environments, including why seasonality is im- other aspects of seasonality with annual cycles.
portant to the study of human behavior and what The fact that a particular clam died on St. Cathe-
seasonality can tell us about life in dynamic es- rines Island in November/December is, by itself,
tuarine systems. The Fifth Caldwell Conference of limited use in testing theories of seasonality
was organized to explore aspects of seasonality and mobility with regard to the use of resources
found in common biological remains and sub- and locations over time. To articulate season of
stantive applications with particular reference to death with broader inferences requires arguments
the archaeology of St. Catherines Island and the demonstrating that the death of an organism was
Georgia Bight (USA). contemporaneous with (and relevant to) specific
Issues of seasonality and mobility are fun- behaviors, cultural institutions, interactions with
damental to much archaeological research, even neighboring groups, site functions, and site his-
when not made explicit. These questions are linked tories. Without such linkage arguments, seasonal
to resource availability, subsistence strategies, estimates tell us very little about people.
economic institutions, and social complexity. But We believe it is important not to conflate pat-
the settlement pattern issue is too often reduced terns of seasonal site occupation (reflecting mo-
to a dichotomy between mobility on the one hand bility strategies), seasonal patterns of resource
and sedentism on the other, as though they were procurement (reflecting foraging and/or farming
mutually exclusive options. As Bar-Yosef and Ro- strategies), site function, and relationships with
cek (1998: 1, italics theirs) argue: “The fact is all people on neighboring islands and the mainland
societies have a mobility component; the issue is (e.g., Quitmyer et al., 1997: 826). These are
what the form of that mobility is, not whether it very different—if interrelated—facets of human
exists.” As with most human institutions, residen- behavior.
tial patterns as well as resource acquisition come The chapters in this volume address specific
in many variations for many different reasons. issues (such as inconsistencies in the ethnohis-
Archaeological excavations along the Geor- toric record, refinements in dating archaeologi-
gia Bight typically recover thousands of potential cal deposits, intersite and intrasite variability,
“seasonal indicators.” Aten (1981) identifies five and regional settlement patterns) associated with
characteristics commonly used to study seasonal- biological proxies used to assess aspects of sea-
ity in animals: presence or absence of skeletal ele- sonality and mobility along the Georgia Bight
ments (e.g., bones of migratory birds), demogra- generally, and on St. Catherines Island in par-
phy (e.g., changes in the sizes of estuarine fishes as ticular. They were originally presented at the
they mature through the annual cycle), morphol- Fifth Caldwell Conference, cosponsored by the
ogy (changes in shell contour through the annual American Museum of Natural History and the
cycle, e.g., marsh clams [Rangia cuneata]), struc- St. Catherines Island Foundation and held on St.
ture (changes in shell microstructure correlated Catherines Island, Georgia, May 14–16, 2010.
with the seasons of the year, e.g., growth phases The first chapter in the volume, by David
in hard clams [Mercenaria spp.]), and chemistry Hurst Thomas, discusses sampling and inferential
(changes in shell composition, e.g., shifting oxy- problems raised when linking seasonality studies
gen and carbon isotopes in Mercenaria). to our understanding of mobility patterns and
Regardless of approach, these techniques can paleoclimatic change. The author recaps how is-
reveal only when an organism died and some in- sues of seasonality and mobility have dominated
formation about environmental conditions pre- archaeological research and related discourse on
13
St. Catherines Island, as well as the rest of the J. Reitz, Bruce M. Saul, J.W. Moak, Gwendolyn
Georgia Bight. He specifically reviews efforts to D. Carroll, and Charles W. Lambert compare and
systematically collect archaeological evidence to contrast seasonal evidence obtained from research
resolve the conflicting ethnohistoric interpreta- on modern and archaeological fishes along the
tions of the Georgia coastline (the so-called Guale Georgia Coast. Estuarine fishes are a mixture of
problem). For the Late Archaic period, inquiry endemic, marine, and freshwater species attracted
has centered on whether the massive shell ring to estuaries as nursery grounds and feeding areas.
settlements represent episodic aggregations or The potential of estuaries to fill these roles var-
long-term, year-round occupations (or both). For ies seasonally, raising the possibility that fishes
the Late Prehistoric period, the “Guale problem” can serve as proxies for seasonal patterning in
has highlighted contrasting views of residential human fishing strategies. In this study, the spatial
and logistic mobility for these coastal people. and seasonal habits of modern and archaeologi-
Douglas J. Kennett and Brendan J. Culleton cal fishes are evaluated for evidence of seasonal
present a Bayesian chronological framework as fishing strategies. These authors find that fishing
a way to determine contemporaneity within and was highly selective and that some fish taxa in the
among sites, a procedure essential for assessing archaeological record are not markedly seasonal,
seasonal rounds, residential mobility, and deter- limiting the value of simple presence or absence
mining whether specific sites used during dif- to provide seasonal fishing information.
ferent or the same seasons are coeval. These au- In chapter 4, Carol E. Colaninno evaluates
thors note that seasonality studies depend upon data from oxygen isotopic profiles and season-
defining culturally meaningful analytical units ally sensitive vertebrate remains as evidence for
and analyzing and interpreting seasonality data year-round occupation at Late Archaic shell rings
within the broader context provided by these on the Georgia coast. Specifically, seasonal occu-
units. Chronology building is an essential part pation at five Late Archaic (4200–3100 b.p.) shell
of defining these analytical units. The authors rings is explored using oxygen isotopic evidence
observe that Bayesian analysis of radiocarbon from hardhead catfish (Ariopsis felis) and Atlan-
dates from archaeological sites is routine in tic croaker (Micropogonias undulatus) otoliths.
Britain, and programs like OxCal provide pre- Combined with knowledge of seasonal availabil-
packaged statistical tools with which to develop ity, Colaninno’s results demonstrate that these
robust archaeological site chronologies. Cus- fishes were captured during multiple seasons at
tom-built Bayesian statistical environments that each of the five Late Archaic shell rings; all four
analyze and model spatiotemporal patterns in the seasons are represented at several sites. This sug-
archaeological record are also being developed, gests that some portion of the human population
including tools potentially useful for integrating occupied these sites throughout the year during
the multiscalar chronological measures neces- this period.
sary for establishing site seasonality and infer- Sarah G. Bergh discusses intrasite variabil-
ring mobility patterns. The authors then provide ity in seasonal occupation at Back Creek Village
a new Bayesian chronological framework for (a.d. 1200–1600) on St. Catherines Island. These
interpreting seasonality data on St. Catherines archaeological shell middens likely accumulated
Island, combining stratigraphic information and outside structures occupied by sedentary house-
other prior knowledge with radiocarbon dates to holds. Bergh evaluates the degree to which in-
establish more precise and accurate chronologi- tra- and intermidden variation reflects specific
cal models. This paper concludes that changes modes of shell accumulation. This study uses an-
in material culture observed at the two Late Ar- nual cycles in fishes and hard clam (Mercenaria
chaic shell rings on St. Catherines Island could spp.) growth habits to explore seasonal patterns
have occurred within less than a decade. The of midden deposition, which can be linked to site
authors argue that more precise AMS 14C dates function and mobility. Bergh argues that villagers
could provide additional insights into the cul- relied heavily on vertebrate resources present in
tural processes in play. the estuary throughout the year, suggesting that
The next chapter addresses one of the major seasonal mobility to exploit animal resources
kinds of evidence available to archaeologists in- was not a factor in the settlement system.
terested in understanding seasonality along the In chapter 6, C. Fred T. Andrus presents an
Georgia coastline—fish. In chapter 3, Elizabeth integrated, programmatic approach to assessing
2012 SEASONALITY AND HUMAN MOBILITY
PREFACE ALONG THE GEORGIA BIGHT 15
seasonality and sedentism on the Georgia Bight. modern and zooarchaeological hard clams (Mer-
Andrus notes that most molluscan sclerochronol- cenaria spp.) from the Litchfield Beach Estuary,
ogy research in the southeastern United States South Carolina. They note that the periodicity of
focuses on Eastern oyster (Crassostrea virginica) incremental shell growth in Mercenaria spp. is
and quahog (or hard clam [Mercenaria spp.]), the a powerful tool for estimating the season of re-
two most common taxa in archaeological mid- source procurement and anthropogenic impact
dens from this area. But he argues that several on zooarchaeological resources. The seasonal
other molluscs likewise show varying degrees of pattern of shell formation in modern populations
promise as proxies for season of capture. While is well documented throughout their range, but
these other taxa may not be as abundant as oys- recent research suggests that changes in the ma-
ters and quahogs, their analysis may provide in- rine environment can quickly and significantly
sight into human behavior in a varied range of alter patterns of growth within a region. The re-
ecosystems and offer a more complete picture search analyzes the variability of oxygen isotope
of subsistence activities. This paper surveys rel- ratios (18O/16O) in shell carbonate to validate the
evant literature and presents new data concerning seasonal periodicity of incremental shell growth
the development of novel season of capture prox- in modern and zooarchaeological specimens re-
ies. Key areas of concern include taxon-specific covered from Litchfield Beach. The data verify
properties of biomineralization, shell structure, the hypothesis that the periodicity of incremental
seasonal growth, taphonomy, and geochemistry. shell growth has not changed over the past 2000
The goal is to foster expansion of the current years, allowing the population dynamics of mod-
suite of sclerochronological tools available to ar- ern and ancient populations in the region to be as-
chaeologists in this region and encourage modern sessed. The authors document significant changes
proof-of-concept research. in the composition of the ontogenetic age classes
In chapter 7, Irvy R. Quitmyer and Douglas of pre-Hispanic hard clam populations’ top down
S. Jones discuss annual incremental shell growth impact on hard clam beds.
patterns in hard clams (Mercenaria spp.) recov- In chapter 9, Deborah Ann Keene reevaluates
ered in archaeological contexts from St. Cath- the use of impressed odostomes (Boonea impres-
erines Island. Hard clams grow by mantle-me- sa) as proxies for season of capture for eastern
diated precipitation of calcium carbonate form- oysters (Crassostrea virginica) recovered from
ing a predictable alternating pattern of opaque middens along the Georgia Bight. The potential
(white) and translucent (gray) shell increments for zooarchaeological research was first explored
corresponding to seasonal changes in water tem- by Russo (1991) who hypothesized that deter-
perature. This research applies sclerochronologi- mining the season of death for Boonea impressa
cal techniques to document the annual periodic- would reveal the season of death for the oysters
ity of shell formation in modern populations of to which they were attached. By assuming that
hard clams from St. Catherines Island and applies most Boonea impressa are born at the same time
this model to the interpretation of Archaic-period of year, live for about one year, and grow steadily
archaeological assemblages. Their results dem- throughout that year, Russo argued that their sea-
onstrate that hard clams recovered from the St. son of death, as well as season of collection for
Catherines and the McQueen shell rings were oysters, could be estimated from odostome size.
intensively harvested during the late winter and To test his hypothesis, he collected and measured
spring seasons. These authors explicitly address modern samples of Boonea impressa from the
two companion issues associated with high lev- northeast coast of Florida for 14 months, then de-
els of sedentism: the extent to which natural and veloped a model of their yearly growth based on
anthropogenic processes impact these resources, shell length. Keene notes that whereas additional
and whether there is evidence of resource man- data have been collected on the growth, reproduc-
agement strategies. They clearly demonstrate that tion, and behavior of Boonea impressa in North
knowing something about cultural activity is to Carolina, we still do not understand how repro-
know something about the biology and ecology duction and growth are affected by environmen-
of the animals that were used. tal factors, meaning that studies using Boonea
Douglas S. Jones, Irvy R. Quitmyer, and Ches- impressa to determine season of capture require
ter B. DePratter present oxygen isotope validation deeper inquiry to establish a more reliable mod-
of annual macroscopic shell growth increments in el to assess the season of oyster harvest. Keene
proposes several potential ways to evaluate ac- earlier models suggested. But Waselkov warns
curacy, and concludes that the cost effectiveness that current archaeological methods employed
and ease of this method for determining season of to document and interpret subsistence seasonal-
death means that it should not yet be abandoned. ity (which is our principal portal to the multiple
Nicole R. Cannarozzi evaluates the potential facets of residential mobility and sedentism)
of the Eastern oyster (Crassostrea virginica) as have many shortcomings. Although even our best
a proxy for seasonality using zooarchaeological available methods have real limitations on sea-
collections from St. Catherines Island. She dis- sonal resolution, Waselkov suggests that a higher
cusses how determining the season of collection standard of analysis can provide a firmer basis for
of oysters in the Georgia Bight is complicated by models of coastal sedentism and mobility.
the dynamic nature of the estuaries these mol- In the concluding paper, Elizabeth S. Wing
luscs inhabit. Environmental changes may cause discusses the papers presented at the Fifth
unpredictable biological responses, including the Caldwell Conference, noting that many chal-
formation of multiple shell growth breaks. Cur- lenges exist in documenting the seasonal forag-
rently, stable isotope geochemistry is the most ing patterns of the past. She emphasizes the ex-
effective method for determining seasonality of pansion of archaeological techniques available
oyster harvest in the area. Morphological data for examining seasonal change over the past
show that different habitats were exploited and/ few years and encourages further integration of
or oyster habitats on St. Catherines have changed data from such diverse sources as charcoal, tree
considerably over time. She includes in her study rings, growth increments in mollusc shells and
measurements of impressed odostomes (Boonea fish otoliths, and fish faunal assemblages. Wing
impressa), which indicate year-round collection states that these techniques provide a more
with the greatest number of individuals collected complete understanding of human conditions
during cool months. in the past. She also underscores the difficulty
In chapter 11, C. Margaret Scarry and Kan- of integrating data from plants with those from
dace D. Hollenbach examine the question of animals due to the relatively poorer preservation
whether plant data derived from archaeological of plant remains, and suggests that the recovery
samples provide reliable evidence of seasonality of additional plant remains from waterlogged
in the past. In most environments, plants have conditions could help meld these two datasets.
predictable cycles of flowering and fruiting, for Wing notes that the bioarchaeological remains
example, and seem likely candidates for estimat- discussed in these papers derive from excavated
ing the season(s) of occupation. The potential to archaeological deposits and points to the im-
delay harvest and to store many plants, however, portance of documentary evidence in providing
complicates direct associations between harvest (sometimes conflicting) firsthand glances at the
seasons and seasonal residential patterns. The past foraging activities of people living along
authors, instead, consider seasonal subsistence the Georgia Bight.
rounds from the perspective of operational chains
involved in acquiring, processing, and consum- ACKNOWLEDGMENTS
ing key plant resources. To do this, they draw
upon species inventories from archaeological We express our thanks to the trustees of the
sites throughout the region as well as ethnograph- St. Catherines Island and Edward John Noble
ic and ethnohistoric information. They conclude foundations for their long-term support of the
that their hypothetical round of plant handling Caldwell Conference series, cosponsored by the
could be conducted from marshside settlements American Museum of Natural History and held on
on St. Catherines Island without requiring sea- St. Catherines Island, Georgia. We are especially
sonal settlement relocation. grateful to Ms. June Gibson, for her friendship,
Gregory A. Waselkov discusses more general guidance, and willingness to stick with us over
implications of seasonality and mobility in coast- the decades. We are also grateful to Mr. Jeremy
al research. He notes that coastal archaeologists Smith (President, St. Catherines Island Founda-
have refined their studies of subsistence remains tion), Mr. Noble Smith (Executive Director, St.
for evidence of resource intensification and sea- Catherines Island Foundation), Mr. David Smith
sons of resource use and residential patterns. Sea- (Director, St. Catherines Island Foundation), and
sonality, in particular, is far more complex than the late Mr. Bradford Smith (Co-Chair, St. Cath-
2012 SEASONALITY AND HUMAN MOBILITY
PREFACE ALONG THE GEORGIA BIGHT 17
erines Island Foundation); each has made unique Fifth Caldwell Conference and subsequent pub-
contributions to our archaeology program. lication operate so smoothly.
Special thanks go to Mr. Royce H. Hayes We thank Dr. Kenneth Ames, Dr. Geoff Bai-
(Superintendent of St. Catherines Island) who ley, Dr. Robert Cerrato, and Dr. Jon Erlandson for
has facilitated the Caldwell Conference series in their thoughtful and critical reviews of the entire
countless ways. We are likewise grateful to the draft manuscript.
members of the St. Catherines Island staff, past Mr. Irvy Quitmyer took the photo that is fea-
and present, who helped out in the hundreds of tured on the cover. We also thank Mr. Elliot Blair,
ways that made our archaeological investiga- Ms. Rachel Cajigas, Ms. Christina Friberg, Ms.
tions more productive (and more fun). We are Chelsea Graham, Ms. Ginessa Mahar, Mr. Mat-
also extremely grateful to the staff of the St. thew Napolitano, Ms. Lorann S.A. Pendleton,
Catherines Island Archaeological Project: Mr. Ms. Diana Rosenthal, Mr. Matthew Sanger, and
Elliot Blair, Ms. Rachel Cajigas, Ms. Christina Ms. Anna Semon for assistance with the manu-
Friberg, Ms. Chelsea Graham, Mr. Matthew Na- script production. We acknowledge the contribu-
politano, Ms. Lori Pendleton, Ms. Ginessa Ma- tions of Ms. Brenda Jones and Dr. Mary Knight
har, Ms. Diana Rosenthal, Mr. Matthew Sanger, for their editorial efforts and Ms. Jennifer Steffey
and Ms. Anna Semon; each made numerous for her skill and patience in designing the cover
behind-the-scenes contributions that made the and assisting with the final artwork.
2012 SEASONALITY AND MOBILITY ON THE GEORGIA BIGHT 19
CHAPTER 1
Seasonality and Mobility on the
Georgia Bight:
Why Should We Care?
D avid H urst T homas 1
Three years ago, our team published a three- jectives guiding the Fifth Caldwell Conference, it
volume, 1136-page monograph called Native seems worthwhile to explore the pivotal relation-
American Landscapes of St. Catherines Island ship between seasonality of site occupation and
(Thomas, 2008), presenting the data and synthe- human mobility strategies. To understand how
sizing all of our previous research on St. Cath- seasonality inferences were embedded within the
erines Island (except for the archaeology of Mis- St. Catherines Island archaeological program,
sion Santa Catalina de Guale). The American I’ll first summarize the previous findings, then
Museum of Natural History’s program in archae- deconstruct the Native American Landscapes
ology on St. Catherines Island has always been volumes specifically with respect to seasonality
about teamwork, and the Native American Land- (Thomas, 2008). The idea is to highlight some
scapes volumes clearly reflects this broad-based, shortcomings in our previous work and anticipate
interdisciplinary bias, with more than two dozen some ways to improve our understanding of site
contributors to the final publication. seasonality and its impact on the estuarine sys-
Four decades of archaeological research, en- tems of the Georgia Bight.
capsulated in Native American Landscapes, ad-
dressed aboriginal lifeways of indigenous St. WHAT’S THE ‘GUALE PROBLEM’?
Catherines Islanders in terms of four deceptively
simple questions (after Thomas, 2008: iii): The Guale Indians of the interior Georgia
Bight were among the first indigenous peoples
How and why did the human landscape (re- met by Europeans exploring north of Mexico.
flected in settlement patterns and land use) change After brief contact with the Spanish in 1526, this
through time? Muskhogean-speaking group later encountered the
To what extent were subsistence and settlement French in 1562–1563. Then, beginning in 1566,
patterns shaped by human population increase, in- the Guale experienced a sustained period of Span-
tensification, and competition for resources? ish colonization and missionization.2 By 1684, the
What factors can account for the emergence gradual southern withdrawal of the Spanish cou-
of social inequality in Georgia’s Sea Islands? pled with the southward expansion of the Carolina
Can systematically collected archaeological colony fostered relocation and reorganization of
evidence resolve the conflicting ethnohistoric in- the vastly reduced Guale population.3
terpretations of the aboriginal Georgia coastline
(the so-called Guale problem). Conflicting Views
Building on Swanton’s (1922) landmark
In this chapter, I focus on the fourth of these synthesis The Creek Indians and Their Neigh-
research objectives. bors, Lewis Larson (1969, 1980) contrasted the
A critical element in the scientific enterprise, Guale with apparently more sedentary peoples
I believe, is the right to be wrong. Given the ob- of Apalachee Florida. Concluding that coastal
Georgia environments seriously constrained and the empirical level I believe that [the conventional
hampered aboriginal lifeways, Larson argued wisdom] has led to an overstatement of the iso-
that only a highly dispersed, seasonally mobile lation of the Guale from the interior, the unpro-
population could have survived there. ductivity of Guale horticulture, and the scattered
Particularly critical to this interpretation were quality of Guale settlements” (Jones, 1978: 189).
ethnohistoric documents from the central Georgia Jones (1978: 179, 191, 194) believed that the Je-
Bight during the 1560s (Larson, 1969: 293–297), suit reports deliberately exaggerated the “misery”
especially the accounts by Fr. Jean Rogel at Orista of the land, and proposed instead that the Guale
(Sturtevant, 1964: 170), Fr. Antonio Sedeño (who lived in “dispersed towns … there is no doubt that
described coastal Georgia as “the most miserable there was a town center with a large round commu-
thing ever discovered” (Zubillaga, 1946: 424), nity building, a chunky field, and some residential
and René Laudonière (Bennett, 1975: 121), who structures. The regular presence of some form of
recorded similar impressions in 1564 at Outina (a mortuary structure is likely. The town center was
Timucuan settlement located just west of the St. surrounded by dispersed households practicing
Johns River in northeastern Florida (see Worth, shifting horticulture. Intergroup and intragroup
1998: 21 and Thomas, 2008: chap. 11). These economic exchanges and redistribution systems,
(and other) documents convinced Larson (1980: in a context of considerable resource variability
209, 218) that the infertile and patchy soils of contributed toward a dependable food supply that
coastal Georgia were “the primary reason for the probably required little regular seasonal residen-
scattered and small size of the agricultural pro- tial mobility” (Jones, 1978: 200).5
duction unit.” He concluded that along the Guale Jones (1978: 209) summarized his position
coast “permanent settlements were not the rule, this way: “Larson’s … reliance on the Jesuit view
for long seasonal junkets in pursuit of game miti- of the sixteenth century Guale … led him to con-
gated against a settled populace” (Larson, 1978: sider the Guale as an isolated, distinctive adapta-
122; see also 1969: 293–297, 1980: 206–209). tion. While it is significant that the Guale were
Despite some modifications in the mid-1980s adapted to special environmental conditions, I
(addressed in Thomas, 2008: chap. 11), Larson’s have argued that these conditions neither created
perspectives on Guale subsistence and settle- isolation nor impeded the development of a com-
ment were accepted and amplified by a number plex level of sociocultural integration.… Despite
of investigators (esp. Crook, 1984, 1986; see also their environmental and adaptive differences, the
Wallace, 1975: 265–271; Pearson, 1977: 62–63; Guale and the interior groups shared such ba-
Reitz and Scarry, 1985: 46; Reitz, 1988, 1991; sic features as chiefdoms, military federations,
see also Thomas, 2008: chaps. 11 and 35). matrilineality, and dual aspects of organization.
Likewise, arguing long-term, high residen- … [I]t is of considerable theoretical importance
tial mobility, Crook (1986: 17–20, fig. 2, 2004; that such an exploration eventually be attempt-
see also Crook, 1984: 260, 1986: 18–20; Larson, ed, for the Guale case seems to demonstrate that
1980) proposed a “purely aboriginal” fission- more than one set of adaptive conditions may
fusion settlement model for the precontact Guale well combine to create highly similar features.”
people. Crook defined the Guale wintertime These are vastly different interpretations of
settlement (mid-December through mid-March) the ethnohistoric record for the Georgia coast-
as “minimal settlements” consisting of a single line, and the so-called Guale problem became
matrilineage, dispersed “adjacent to tidal streams the central research question addressed in the
which permitted access to the estuarine system” Native American Landscapes volumes (Thom-
(1986: 22).4 Crook (1986: 53) summarized his as, 2008; see also 1987: 57–64, Worth, 1999;
“Annual Model” this way: “large villages primar- Saunders, 2000b; Ruhl, 2003: 188–189; Keene,
ily occupied during the summer, smaller settle- 2004: 672).
ments occupied either multi-seasonally or during
the fall–winter season, and small sites occupied Competing Hypotheses
for very short periods of time do exist” (see also and Logical Consequences
Thomas, 2008: table 11.6). The Guale problem turns on the twin issues
Ethnohistorian Grant Jones (1978, 1980) came of residential mobility and economic intensifica-
to a quite different conclusion after his own close tion, which I believe can be succinctly expressed
reading of the same documentary sources: “On in two competing hypotheses extracted from the
ethnohistoric record: most St. Catherines islanders maintained four-

The Jesuit hypothesis: Guale Indians were season residence in a single community. Hunting,
egalitarian, highly nomadic, mostly nonfarming fishing, and wood-collecting trips took individu-
foragers who lived in low-density, small, season- als away from the community on a temporary ba-
al aggregations; sis. Guale dispersed towns were located mostly
The Franciscan hypothesis: Guale Indians along the margins between maritime forest and
were a hierarchical, ranked society; they relied saltwater marshes. These forager farmers (or,
heavily on maize cultivation, rarely moved perhaps, farmer foragers), lived in large, relative-
their residences, and lived in high-density “dis- ly stable residential central places, out of which
persed towns.” logistical forays of small “task groups” brought
Native American Landscapes of St. Cathe- plants and animals back home. This strategy of
rines Island (Thomas, 2008) was all about testing minimal residential mobility should produce a
these two hypotheses. distinctive archaeological record, characterized
The truth, of course, is that scientists don’t re- by sustained, four-season occupations of marsh-
ally test their hypotheses—what we test are the side settlements, with only minimal evidence of
logical implications of our hypotheses. Essen- single- or biseasonal occupations elsewhere.
tially, hypotheses are abstract, inductive state-
ments that must be translated (through deductive ARCHAEOLOGY
reasoning) into logical, empirical, material con- AND THE GUALE PROBLEM
sequences. Such deductive arguments generally
take the form of “if ... then” statements: If the The American Museum of Natural History
hypothesis is true, then we will expect to observe harnessed a broad array of field and analytical
the following outcomes. Bridging the gap from if techniques to test the Jesuit and Franciscan hy-
to then is the tricky step. potheses (esp. Thomas, 2008: chaps. 7–12). We
With respect to the Guale problem, we framed situated this fieldwork within the general para-
the following logical consequences from the two digm of human behavioral ecology, grounded
competing hypotheses: in three basic models (Thomas, 2008: chaps.
The Jesuit hypothesis: If the Jesuit observa- 7–10). The diet-breadth (or prey choice) model
tions correctly reflect mid-17th century Guale addressed the issue of which foods an efficient
settlement patterns, then the resulting archaeo- forager should harvest from all those available on
logical record should reflect a palimpsest of some St. Catherines Island. Diet-breadth models predict
four-season settlements and several seasonally that foragers will optimize the time spent captur-
specific settlements. The most concentrated oc- ing prey, and employ the simplifying assump-
cupations should occur during the summer, with a tions that all resources are randomly distributed
dispersed series of archaeological accumulations (without patches) and that “capture/handling”
during the early fall, followed by moderate occu- and “search” times represent the sum total of all
pational intensity from the late fall through win- time spent foraging. We also applied the patch
ter. From the fall through springtime, the Guale choice model, which, combined with the cen-
were residentially mobile foragers who should tral limit theorem, predicts that foraging efforts
have produced an archaeological record of (1) will correlate directly with efficiency rank order,
dispersed oak forest settlements (fall and spring- meaning that foragers should spend more time
time occupations only) and (2) dispersed marsh- working the higher-ranked patches and less time
side settlements (late fall–winter and springtime in patches with lower energetic potential. Using
occupations only). the central place foraging model, we investigat-
The Franciscan hypothesis: If the Francis- ed the time/energy spent processing resources at
can observations correctly reflect mid-17th cen- temporary camps before transport to a residential
tury Guale settlement patterns, then the result- base. For several years, we have also conducted
ing archaeological record should reflect a basic a series of optimal foraging experiments, specifi-
strategy of low residential mobility centered on cally addressing procurement and return rates for
“dispersed towns,” generally positioned along key marine and terrestrial resources that would
the maritime forest/marsh margin. Whereas some have been available to aboriginal foragers on St.
Guale may have shifted their winter residence to Catherines Island.
be closer to hunting, fishing, or shellfish patches, We conducted a 20% probabilistic transect
survey of St. Catherines Island, walking and procedure generated a sample of 122 archaeo-
probing for buried sites across a series of 31 east– logical sites, which we tested with more than
west transects, each 100 m wide (fig. 1.1). This 400 1 m × 1 m units. Because the transect sam-
A6
A1
B6
B1
C6
C1
D6
D1
E6
E1
F6
F1
G6
G1
H6
H1
I6
I1
J6
J1
K6
K1
L6
L1
M6 N
M1
N6
N1 0 1 2
O6 KM
O1
P6 0 1 2
MI
Fig. 1.1. The randomized transect research design employed in the island-wide survey of St. Catherines Is-
land (after Thomas, 1987: fig. 22).
pling was heavily biased toward sites with ma- lifeways or addressing more processual matters.
rine shell, we also conducted a systematic shovel In our St. Catherines Island research, we have ap-
testing program and augmented these systematic proached chronology in the multiscalar fashion
surveys with a direct shoreline reconnaissance anticipated by Berthold Laufer a century ago:
(mostly following the late Holocene surfaces), The Hour Hand: The St. Catherines Island
recording an additional 84 shoreline sites. By ceramic sequence provided the “little hand” on
plotting the distribution of these known-age sites the clock, generally providing temporal controls
across the Holocene beach ridges, we developed at a submillennial level of resolution.
a detailed sequence documenting the prograda- The Minute Hand: The radiocarbon sequence
tion and erosion of beach ridge complexes adja- became the “big hand” on the clock, generating
cent to tidal estuaries and coastal shorelines on temporal controls at a century level of resolution.
St. Catherines Island. The Second Hand: Site seasonality studies
These data in hand, we turned next to the ana- offer microchronological controls at the resolu-
lytical tools necessary to make sense of this ar- tion of months or even weeks.
chaeological record, beginning with chronologi- Below, we discuss how this multiscalar ap-
cal controls. proach to chronology allowed us to evaluate the
Guale problem on St. Catherines Island, especial-
DEVELOPING MULTISCALAR ly as it relates to site seasonality studies.
CHRONOLOGIES
The St. Catherines Island Ceramic
Chronology is at the root of the matter, Chronology: Reading the Hour Hand
being the nerve electrifying the dead The regionalized random sampling of St.
body of history. Catherines Island generated more than 15,000
–Berthold Laufer (1913: 577) potsherds from 122 archaeological sites (Thom-
as, 2008: tables 20.3 and 20.5). Perceived as “in-
A century ago, addressing the annual meet- dex fossils,” these diverse ceramic assemblages
ing of the American Anthropological Associa- provided the initial (and most coarse-grained)
tion, ethnographer Berthold Laufer pretty much chronological controls.
got it right—chronology really does lie at the All of the aboriginal ceramics recovered from
heart of the matter. Archaeologist Alfred Tozzer St. Catherines Island were initially classified ac-
(1926: 283) subsequently expressed similar sen- cording to the northern Georgia coast ceramic
timents about archaeological data having “an chronology, as refined by Chester DePratter
inert quality, a certain spinelessness when unac- (1979: table 30; as updated in DePratter, 1991:
companied by a more or less definite chronolog- table 1; see also Thomas, 2008: chap. 14 and 15).
ical background” (see also O’Brien and Lyman, In this synthesis, DePratter relied on temper, sur-
1999: chap. 1). face decoration, rim form, and vessel form that
Laufer (1913: 576) went on to suggest that demonstrably vary “asynchronously” (DePratter,
whereas archaeology and ethnology are “insepa- 1979: 122). DePratter’s chronological sequence
rably one and the same—emanations of the same contained seven major cultural periods, subdi-
spirit, pursuing, as they do, the same ideal, and vided into nearly two dozen archaeological phas-
working to the same end,” they operate at differ- es (table 1.1.) This coarse-grained approach to
ent timeframes. “When archaeology and ethnol- chronology provided temporal estimates accurate
ogy have drawn up each its own chronology, then at the submillennial level.6
the two systems may be pieced together and col-
lated, and the result cannot fail to appear.” This Radiocarbon Dating:
observation is important because Laufer explicit- Reading the Minute Hand
ly recognized that chronology typically operates Monitoring radioactive emission from organic
at multiple levels—what today we might call a specimens by determining the current rate of 14C
“multiscalar” approach to chronology control. breakdown, the radiocarbon method estimates the
Today, we understand that archaeology’s ini- length of elapsed time since the death of a plant or
tial objective must always be chronological—re- animal. Within the last couple of decades, physi-
flecting the necessity of establishing a firm grasp cists have discovered that the atmospheric level
on time before attempting to reconstruct ancient of radiocarbon has changed somewhat over the
last several millennia, enabling archaeologists to tural) event of interest.

“correct” their radiocarbon dates using an absolute Throughout the St. Catherines Island research,
chronology based on radiocarbon dating of known- we have emphasized the importance of radiocar-
age bristlecone pine samples. New advances in bon dating in both archaeological and geological
accelerator-based radiocarbon methods permit perspectives. The St. Catherines Island archaeo-
archaeologists to use extremely small samples, logical research is currently grounded in a data-
vastly stretching the potential of the method.7 base of roughly 300 14C determinations, most of
By itself, the radiocarbon method tells us them processed on “cultural” samples recovered
nothing about human activities in the past. A 14C from documented archaeological contexts on St.
date merely estimates when a certain hickory nut Catherines Island,8 primarily burial mounds and
dropped to the ground or a specific hard clam shell middens. We have an additional suite of 14C
died. In each case, the event being dated must be dates from noncultural contexts, primarily organ-
demonstrated to be coeval with a behavioral (cul- ics and marine shell samples collected in conjunc-
TABLE 1.1
Comparison of Northern Georgia Coast and St. Catherines Island Chronologies
(Northern Georgia coast after DePratter, 1979: table 30, as modified by DePratter, 1991:
table 1. St. Catherines Island after Thomas, 2008: table 15.3.)
Northern Georgia Northern Georgia coast St. Catherines Island
Phases coast chronology age chronology age (calibrated) chronology age (calibrated)
(uncalibrated)
a.d. 17001 — a.d.17002
Altamaha
a .d . 1580 — a.d. 15802
Irene
a .d . 1325 a .d . 1310–1390 a.d. 1300
Savannah Savannah phase deleted
a .d . 1200 a.d. 1280 a.d. 1300
St. Catherines
a .d . 1000 a.d. 1050–1150 a .d . 800
Wilmington
a.d. 500 a .d . 630 a .d . 350
Deptford
400 b.c. 400 b.c. 350 b.c.
Refuge
1100 b.c. 1360 b.c. 1000 b.c.
St. Simons
2200 b.c. 2750–2860 b.c. 3000 b.c.
Beginning and ending age estimates for the Altamaha period in the northern Georgia coast chronology are
1
based on historical documentation, not 14C dating.

2
Uncalibrated.
tion with vibracore sampling and surface geolog- ia in 1975, a process that continued, somewhat
ical reconnaissance (see Bishop et al., 2011: ap- sporadically, over a nine-year interval; an inde-
pendix 1 for a compilation of these dates). Each pendent sample of modern Mercenaria was also
radiocarbon date was calibrated according to the collected between April 1994 and March 1995, in
conventions and protocols discussed in Thomas support of the oxygen isotope study (Andrus and
(2008: chap. 13). Crowe, 2008).
Radiocarbon dating provides temporal esti- Working with George R. Clark II (1979), we
mates generally accurate at the level of centuries, began a program of seasonal analysis by analyz-
sometimes even at the decades level. ing Mercenaria recovered from Johns Mound,
Marys Mound, and McLeod Mound (see also
Site Seasonality: Thomas and Larsen, 1979; Larsen and Thom-
Reading the Second Hand as, 1982: 338). Clark concluded that most hard
So far, we have emphasized the importance clams interred in these mortuary sites had been
of grounding St. Catherines Island archaeology harvested during the winter months, probably
within a solid macrochronological framework— December and January. But because these zoo-
an ordering of events within segments of millen- archaeological samples were recovered from a
nia, centuries, and even decades. But addressing secondary, nonmidden context, the complex for-
the Guale problem requires more fine-grained mation processes involved precluded actual sea-
temporal controls—certainly on the order of seasonal dating of the mortuary activities. Still, we
sons, perhaps even months or weeks. Literally were encouraged that seasonal patterns were in-
thousands of potential “seasonal indicators” were deed evident in the ancient Mercenaria samples
recovered during our excavations on St. Cath- recovered archaeologically, and we moved on to
erines Island and (following Aten, 1981) we ex- consider hard clam seasonality in the various oc-
plored several commonly employed methods for cupational sites of St. Catherines Island.
determining seasonality in such sites: presence Initial laboratory observations of growth in-
or absence of skeletal elements (such as bones crements along the ventral margins, compiled
from migratory species), demography (changing mostly between the late 1970s and mid-1980s,
size of estuarine fishes as they mature through were expressed in Clark’s descriptive terminolo-
the annual cycle), morphological changes in shell gy (e.g., “early gray,” “early-mid white,” “prob-
contour during the annual cycle, microstructural ably end of white,” and so forth). Since that
changes correlated with the seasons of the year, time, considerable progress has been made on
and chemical changes in shell composition (such the seasonal analysis of molluscs. O’Brien and
as shifting oxygen and carbon isotopes). Thomas (2008) made use of the then-standard-
The Incremental Growth Sequence in ized, six-part subdivision of annual shell growth
Mercenaria: As we developed the archaeological (Jones, 1980; Quitmyer et al., 1985, 1997: 830),
research design for St. Catherines Island, we rec- but converted to a four-stage scale, which was
ognized the potential for determining the season- then (fig. 1.2) correlated with approximate sea-
ality of harvest by analyzing growth increments son of harvest:
in the shell of hard clams (Mercenaria mercenar- O1–2 (initial to intermediate opaque incre-
ia), which occur in some abundance in the local ment): Winter (mid-December–mid-March)
shell middens (O’Brien and Thomas, 2008). Such O3 (terminal opaque increment): Early spring
studies were in their infancy in the mid-1970s, (mid-March–mid-April)
with only a limited literature then available (esp. T1 (initial translucent increment): Spring (mid-
Clark, 1968, 1974; Weide, 1969; Coutts, 1970, April–mid-June)
1975; Coutts and Higham, 1971; Ham and Irvine, T2–3 (intermediate to terminal translucent
1975; Kennish and Olsson, 1975; Koike, 1975). increment): Summer and fall (mid-June–mid-
Most of these pioneering studies emphasized December)
the importance of modern controls for under- Figure 1.2 summarizes the modern control
standing the variability introduced by changing sample, pooled from both St. Catherines Island
water temperatures and salinity, tides, preda- collection sites (after O’Brien and Thomas,
tion, spawning, and other environmental factors. 2008: fig. 17.4).
On St. Catherines Island, we began collecting a Mercenaria suitable for seasonal analysis
modern control sample of Mercenaria mercenar- were recovered from nearly 85% (110 of 130) of
INCREMENTAL STAGE
O1-2 O3 T1 T2-3
December/January 100%
N=3
January/February 100%
N=2
February/March 44% 56%

N = 16
March/April 10% 90%

N = 49
April/May 30% 80%

N = 10
May/June 4% 18% 78%

N = 28
June/July 100%
N=1
July/August 100%
N=7
August/September
33%
N=6
September/October
N=0
October/November
N = 11 100%
November/December 100%
N=8
Annual
20% 36% 22% 21%
N = 141
Fig. 1.2. Interpolated estimates of incremental growth stages for the modern control sample of Mercenaria
collected from St. Catherines Island.
the sites sampled in the island-wide survey, and scheme to narrow down the number of clams to
we saved every single undamaged clam ventral be analyzed, while still avoiding the introduction
margin for potential seasonal analysis. O’Brien of bias in the winnowing process. After the ce-
and Thomas (2008: 484–486) elaborate the vari- ramics from the survey sites were analyzed (us-
ous ways in which these valves were processed ing the St. Catherines Island ceramic chronology,
for analysis, including blind testing and addition- discussed above)—and we could classify most
al quality controls employed. sites according to archaeological period(s)—we
Because such analysis was time-consuming applied a series of sampling conventions to select
and labor intensive, we instituted a sampling appropriate Mercenaria for seasonal analysis:
Single-component sites: If fewer than 25 ary collection period (Russo, 1991: 219). Russo’s
readable clams were available, all clams were correlative study of seasonality in Mercenaria at
analyzed. If more than 25 suitable clams were Meeting House Field (Russo and Saunders, 2008)
recovered, the available valves (or fragments) employed the Kings Bay, Georgia control sample
were numbered sequentially, and a sample of 25 (Quitmyer, Hale, and Jones, 1985) to analyze
was selected using a table of random numbers. the zooarchaeological specimens from Meeting
Some “single component” sites also contained House Field. Overall, these data suggest a year-
evidence of secondary occupations during other round site occupation, though the evidence for
ceramic periods. When this happened, Mercenar- summer occupation is limited. Of particular in-
ia samples were taken whenever possible from terest was the difference in oyster collection and
“temporally discrete” test pits and/or excavation quahog collection in the fall, although both spe-
levels (from those units and levels containing cies were collected in varying amounts through-
only one ceramic complex) by randomly select- out the year.
ing from within these relatively homogeneous Assessing Season of Capture in Nonhu-
intrasite areas. man Vertebrate Remains: The island-wide
Double-component sites: Each component transect survey produced an extensive and di-
was sampled independently by targeting relative- verse set of vertebrate faunal remains collected
ly homogeneous test pits (and/or specific levels) systematically from archaeological sites tested
from each major temporal component. We then across the entire island. Elizabeth Reitz and her
selected up to 25 clams from each component colleagues analyzed this vertebrate faunal assem-
(randomly sampling in the case of n > 25). blage, which contains at least 586 individuals
Multiple-component sites: We wanted to represented by 14,970 vertebrate specimens (Re-
analyze 25 valves from each of the identified itz, 2008; Reitz and Dukes, 2008). Reitz (2008:
components, but in practice, we never recovered 623–624, 660–663, table 22.72) has discussed
sufficient Mercenaria to do this. As a result, we the presence of numerous seasonal indicators in
analyzed whatever clams were recovered and at- the vertebrate zooarchaeological samples recov-
tempted to determine the archaeological age of ered from archaeological sites on St. Catherines
each specimen by charting associated potsherds. Island—especially (1) unshed deer antlers, (2)
Undated components: Several sites con- juvenile deer dentition, (3) presence of shark re-
tained sufficient Mercenaria for seasonal analy- mains, and (4) presence of sea catfish remains.
sis, but too few potsherds to assign a probable pe-
riod of occupation. The seasonal estimates were Seasonality and the Guale Problem
included in the overall, island-wide total, but not With this multiscalar, three-tiered approach
in the period-by-period tallies. to chronology, we synthesized the evidence from
the island-wide archaeological survey, document-
Although this sampling procedure is, admit- ing residential mobility and human population
tedly, a bit mechanistic, we were seeking a way increase between 3000 cal b.c. and cal a.d. 1300
to generate a relatively uniform, representative (Thomas, 2008: chaps. 32–34). Overall, these
sample of hard-shelled clams selected for analy- data demonstrate (1) an exponential increase in
sis by reducing the overall number of analyzed human population through time, and (2) a low de-
specimens to approximately 2000 individual gree of residential mobility throughout the entire
Mercenaria valves (or fragments). Of these, 1771 aboriginal period. This pattern continued through
individual specimens (or fragments) provided us- the late prehistoric (Irene) period, characterized
able growth increment estimates, enabling us to by the largest and most frequent archaeological
address seasonal patterns during the 5000 years occupations recorded in the island-wide survey.
of human history (O’Brien and Thomas, 2008).9 Irene occupations accumulated at an extremely
Russo (1991) has also published estimates of rapid rate (34 occupations/century) and the num-
seasonality in oyster collection on St. Catherines ber of recorded archaeological components sky-
Island by employing measurements of impressed rocketed to 52 (for an average of 17.33 compo-
odostome (Boonea impressa) recovered from nents/century; Thomas, 2008: table 30.1). This is,
Meeting House Field. His analysis of Boonea by far, the densest concentration of archaeologi-
determined that oyster collection took place pri- cal remains recorded for any aboriginal period
marily during the fall, with summer as a second- on St. Catherines Island. The Irene period also
had the largest proportion of “large” sites and the newly available evidence, confirming and ampli-
smallest proportion of “small” sites (see defini- fying ethnohistorian Grant Jones’s (1978, 1980)
tions in Thomas, 2008: table 30.2) recorded dur- perception of the contact-period Guale people:
ing the probabilistic survey. These were largely sedentary foraging farmers
We generated seasonality estimates for 42 who lived in optimally positioned marshside
Irene-period components (representing 124 sea- dispersed towns, grew significant quantities of
sonally specific occupations; see fig. 1.2). More maize and other domesticated crops (at least late
than 40% of these Irene components had all four during the Irene period), and maintained a com-
seasons represented and another 36% were occu- plex chiefdom level of social organization with
pied in (at least) three seasons. Conversely, only centralized, inherited leadership and long-dis-
four Irene components represented a single sea- tance trade networks with the interior (Thomas,
son (each of these being a winter-only occupa- 2008: chap. 35).
tion). Throughout this discussion, we cautioned
against equating a four-season archaeological oc- BUT DID THE JESUITS GET IT WRONG?
cupation with “sedentism” (in the conventional
ethnographic and ethnohistoric usage).10 Given the compelling evidence for supporting
With respect to the implications derived from low residential mobility and significant maize
human behavioral ecology, the upshot was simple cultivation among the coastal Guale people, one
and conclusive: Optimal foraging considerations must ask: What happens to the French and Jesuit
strongly militate in favor of logistical, rather than accounts? These eyewitness accounts consistent-
residential, mobility. Even factoring in the shift- ly describe high residential mobility, seasonal
ing shape of St. Catherines Island over the past dispersal, infertile soils, and minimal horticultur-
five millennia, it is clear that all habitats on the al productivity during the 1560s along the central
island could have been systematically searched Georgia Bight.
and exploited by individual foragers who could Did the Jesuits and the French simply get it
easily return home daily. These biogeographic wrong? Probably not.
constraints suggest that St. Catherines Island Blanton and Thomas (2008) discussed the
foragers could have pursued a strategy of logis- relevance of recent paleoclimatic research on
tic procurement and low residential mobility (at bald cypress (Taxodium distichum) in the Amer-
least during times of relatively favorable climatic ican southeast (Stahle and Cleaveland, 1992;
conditions). Anderson, 1994: 277–289; Anderson, Stahle,
The vast preponderance of archaeological and Cleaveland, 1995; Stahle et al., 1998; Blan-
evidence supports a collector strategy of low ton, 2000, 2004). The period of extended dry-
residential mobility. For most of the time, Irene ness during the latter part of the 16th century is
populations apparently did live, year round in particularly relevant to the present discussion—
dispersed towns located along the forest-marsh a time when “megadrought” conditions plagued
margin (per Jones, 1978: 193–194). Expressed in much of North America (Stahle et al., 2000).
terms of expectations from central place foraging During the early European contact period,
theory, more than 80% of the archaeological com- Stahle et al. (1998: 545) document “a prolonged
ponents encountered on St. Catherines Island (for drought from 1562 through 1571 that was most
all temporal periods) fit the model of sustained severe from 1565 to 1569.” Whereas this in-
and multiseasonal marshside settlements. Only tensely warm and dry interval (between about
limited evidence exists for (1) single- or bisea- a.d. 1565 and 1569) has been little discussed
sonal occupations or (2) inland, special-purpose, in the recent literature, it signals an extraordi-
short-term settlements. narily difficult time for forager–farmers along
To conclude, the combined evidence from the the Georgia coastline—one of many challenges
island-wide archaeological survey, coupled with facing Europeans and Native Americans alike.
recent ethnohistoric interpretations and model- We can now see that the Jesuit missionaries of
ing from human behavioral ecology would seem Georgia and Carolina were facing a prolonged
to resolve the Guale problem, overwhelmingly drought, the driest interval of the entire 16th
rejecting the Jesuit hypothesis, at least on St. century (Worth, 1999; Saunders, 2000b).
Catherines Island. The Franciscan hypothesis, Still smarting from their public failures in
on the other hand, is entirely consistent with the Spanish Florida, the Jesuit friars may have ex-
aggerated their accounts regarding the poverty to understand how the logistic and residential
of the Georgia Bight (Jones, 1978, 1980; Worth, strategies might vary in different localized land-
1999). But when combined with the tree-ring evi- scapes across the Sea Islands and how they might
dence from this same area, these accounts gain respond to short-term climatic fluctuations—de-
considerable credibility because they document spite the fact that these coastal foragers pursued
how these coastal chiefdoms adapted their nor- identical hunt types across identical patch types
mal seasonal and annual routines to accommo- using identical technologies.11
date environmental challenges and social stress.
Judging from the combined tree-ring records Some Conclusions
and surviving ethnohistoric accounts, it seems and Implications
that the foraging farmers of Guale also adapted
their provisioning strategies, sometimes using This is where things stood with the publication
backup tactics to exploit relatively drought-re- of Native American Landscapes of St. Catherines
sistant prey taxa. Additional research is required Island (Thomas, 2008). We concluded that the
INCREMENTAL STAGE
O1-2 O3 T1 T2-3
Altamaha 50% 12% 8% 26%

N = 86
Irene 64% 16% 6% 14%

N = 864
St. Catherines/
Wilmington 81% 14% 7% 4%
N = 216
Wilmington 64% 19% 1% 16%

N = 216
Refuge-
Deptford 86% 19% 2% 11%
N = 126
St. Simons 47% 50% 0% 3%

N = 32
TOTAL 64% 18% 5% 13%

N = 1642
Fig. 1.3. Position of growth surface within major increments at time of harvest: modern control sample of
Mercenaria collected between 1975 and 1984 on St. Catherines Island.
Probable Seasonality
December
September March
June
Meeting House
Field
Mound near South-

end Settlement (9Li3)
LEGEND
Archaeological site
0-9 sherds
10-99 sherds
100-999 sherds
> 1000 sherds
Burial mound
Irene Period
N
cal A.D. 1300–1580
1 .5 0 1
KM
Fig. 1.4. Map of probable seasonality on St. Catherines Island during the Irene period.
modeling from human behavior ecology, newer research designs require a thorough investiga-
ethnohistoric interpretations, and archaeologi- tion of site seasonality. This is why we believe it
cal evidence from the island-wide survey were necessary to deconstruct and to improve upon our
quite consistent with the Franciscan hypothesis: previous work on St. Catherines Island.
during the mid-16th century, Guale society was
hierarchical and ranked, relying heavily on maize The St. Catherines Island Research Design
cultivation and centered on relatively high den-
sity, low-mobility “dispersed towns.” (i.e., the Previous approaches to site seasonality on
Franciscan hypothesis). But we did suggest that St. Catherines Island can be improved in numer-
alternative, backup mobility strategies (consis- ous ways. For one thing, we used ¼ in. screens
tent with the Jesuit hypothesis) seemed possible throughout most of the island-wide archaeologi-
(even likely) during the dramatic drought epi- cal survey.12 As Reitz (2008: 656) and others have
sodes known to have taken place during the 16th noted, the use of such coarse-grained recovery
and 17th centuries. techniques discriminates against smaller taxa, es-
Since the publication of Native American pecially the smaller fishes (Reitz and Quitmyer,
Landscapes, we have continued intensive ar- 1988; Quitmyer and Reitz, 2006; Reitz et al.,
chaeological research on St. Catherines Island, 2010: 54). We should note that in all our excava-
pursuing two primary goals. One project flowed tions since 1982, we have only used 1⁄8 in. (or fin-
directly from our previous interest in the Guale er) screens, and in many cases, the deposits were
problem. Emphasizing remote sensing and large- water-screened as well. In our analysis of incre-
scale excavations, we are currently studying the mental growth in Mercenaria, we failed to distin-
settlement structure and household patterning at guish left from right valves, creating the possibil-
several late prehistoric (Irene) sites, especially ity that some of the randomly sampled specimens
Meeting House Field (9Li21). We think that these came from the same individuals (thereby weak-
large sites are likely the “dispersed Guale towns” ening the assumption of independent sampling).
discussed by Jones (1978). Although we have modified our procedures in
We have also launched a long-term program subsequent research on St. Catherines Island, we
investigating the McQueen Shell Ring (9Li1648) cannot currently assess the degree to which these
and St. Catherines Shell Ring (9Li231), two large shortcomings have biased the estimates of site
Late Archaic sites. The study of such Late Archa- seasonality in the Native American Landscapes
ic rings has provoked lively discussions within of St. Catherines Island (Thomas, 2008).
the archaeological community (e.g., Russo and We also created certain analytical ambigui-
Heide, 2001; Marquardt, 2010a, 2010b; Russo, ties when interpreting the specific incremental
2004, 2008; Sassaman, 2004, 2006; Thomas and patterns observed in the Mercenaria samples re-
Sanger, 2010). Ceramic production first occurs covered for seasonal analysis. When framing the
during the Late Archaic, spurring archaeologists temporal parameters for Native American Land-
to question the motives behind this technologi- scapes of St. Catherines Island, we generally fol-
cal advancement. Similarly, some of the earliest lowed Clark’s (1979) seasonal estimates, which
evidence for extended sedentism is found in such divided the annual growth cycle into phases ob-
Late Archaic sites, begging questions into the cul- servable through thin-section microscopy: win-
tural, sociological, and economic ramifications of ter (mid-December through mid-March), spring
such a shift in settlement pattern. The presumed (mid-March through mid-June), summer (mid-
planning and investment represented in Late Ar- June through mid-September), and fall (mid-
chaic shell rings also raises important questions September through mid-December). For phases
regarding power, control, and hierarchy. With of fast growth, our gradations of the “white” zone
these issues in mind, we have conducted detailed were translated to stages of “opaque” growth
mapping and extensive remote sensing opera- (scaled from O1–3); for episodes of slow growth,
tions at both sites, likewise excavating numer- our observations on the “gray” increments were
ous test units and several large block excavations expressed as increments of “translucent” zona-
(Sanger and Thomas, 2010; see also Thomas and tion. Even though we significantly expanded the
Sanger, 2010). modern Mercenaria control sample beyond what
Although the Late Archaic and late precontact was available to Clark and we switched over to the
periods are separated by three millennia, both six-stage growth incremental criteria of Quitmyer
et al. (1985), our regrouped and reconfigured sea- and zooarchaeological materials encountered,
sonal boundaries remain imperfect. Specifically collecting potential 14C samples in the process.
with reference to the St. Catherines Island con- We have already discussed the biases introduced
trol sample, we found that growth stages T2 and by our selection procedures in radiocarbon dating
T3 were almost entirely coterminous. This means (Thomas, 2008: 442–446). Specifically, when we
that T2 and T3 specimens significantly overlapped examined our motives in sampling the radiomet-
in samples collected between mid-August and ric record of St. Catherines Island, we found a
mid-December. Because of this overlap, we felt it significant bias created by our two rather basic
necessary to group these two incremental stages sampling strategies, attempting to pinpoint chro-
into a single analytical category (denoted as T2– nostratigraphic central tendencies and trying to
3
). Similarly, because we found almost complete define the temporal range of ceramic variability.
temporal overlap in growth stages O1 and O2, we We concluded that the very process of selecting
decided to group these readings into a single cat- appropriate radiocarbon samples introduced a
egory, denoted as O1–2. This is why we employed significant, nonrandom bias into the cumulative
a four-part subdivision of annual shell growth in 14
C histogram of all available radiocarbon dates,
Mercenaria from St. Catherines Island (Mayer likely influencing our attempt to define the tem-
and Thomas, 2008). This is an imperfect solution poral range of ceramic time-markers. This bias
and more recent studies (e.g., Quitmyer, Jones, could readily result in peaks that correspond to
and Andrus, 1997; Andrus and Crowe, 2008; the middle range of a ceramic type and a gap sep-
Thompson and Andrus, 2010) have employed arating temporally contiguous ceramic types.
more fine-grained and better-defined criteria, Accordingly, in 2006, we processed an ad-
standards that need to be applied throughout the ditional 49 radiocarbon determinations from
St. Catherines Island project. contexts in the island-wide survey, attempting to
Other issues of basic sampling also arise from “close the gaps” introduced by such biased sam-
the island-wide transect survey. By relying on ple selections (Thomas, 2008: 451–474). Among
probabilistic, randomized sampling strategies, other findings, we ran 15 additional 14C samples,
we were attempting to maximize the observed each apparently associated with fiber-tempered
variability while minimizing sampling bias ceramics. Surprisingly, only 40% of these de-
(Thomas, 2008: chaps. 11 and 19). Specifically, terminations fell into the expected age range,
we generated a large sample of archaeological with more than half of the marine shell samples
sites and loci, deliberately sampling across the producing significantly later ages than the St.
range of soil types, site sizes, and time periods. Simons period. We know this bias is unidirec-
We did this by using systematic transects, exca- tional because none of the additional 34 samples
vating hundreds of 1 × 1 m test pits, augment- associated with later ceramic types produced St.
ed by a shovel-testing program. Shovel-testing Simons age dates (Thomas, 2008: chap. 16). In
was a relatively unfamiliar technique during the other words, there is a clear-cut tendency of St.
late 1970s, when we conducted the island-wide Simons period ceramics to be commingled with
survey and, in retrospect, we should have sig- marine shell from later time periods. Many of the
nificantly expanded the shovel-testing aspect of marine shell samples apparently associated with
the survey (because, we believe, we overlooked St. Simons and early Refuge-Deptford period ce-
many nonshell sites, thus biasing the overall rep- ramics actually produce much later 14C age esti-
resentativeness of the sample). mates. This systematic error seems to reflect the
Reitz (2008: 617) has already commented on general lack of shell deposits dating to the time
the difficulties in computing Minimum Number span 1350 cal b.c.–200 cal b.c. (despite the pres-
of Individuals, for instance, in the “aggregated” ence of fiber-tempered and Refuge-Deptford pe-
samples resulting from the island-wide survey riod ceramics). The presence of Woodland period
sampling. Similarly, with respect to assessing site molluscs associated with Late Archaic ceramics
seasonality, our strategy of test-unit excavation creates obvious interpretive difficulties and bi-
also introduced an unanticipated degree of bias. ases when assessing matters of site seasonality.
When an archaeological “locus” was discovered, We now believe that similar biases could have
we excavated several (usually four to eight) 1 m2 been introduced in our analysis of growth incre-
test pits, working in 10 cm levels and proceeding ments in hard clams. Earlier in this chapter, we
down into sterile deposit. We saved all cultural reiterated the randomized procedures used to se-
lect specific Mercenaria for analysis. By deliber- atures, spawning, or storms. Because the control
ately attempting to assure “a relatively uniform sample was collected in conjunction with ongo-
distribution of hard-shell clams” (O’Brien and ing archaeological fieldwork, the collection was
Thomas, 2008: 490), we likely “averaged out” particularly strong for the late winter and spring
some of the seasonal variability we were seek- months; but our sampling strategy was sporad-
ing. That is, we probably introduced some bias, ic and somewhat seasonally biased, especially
as we did when selecting only “top and bottom” during the summer months, when our archaeo-
samples for 14C dating biased the macrochronological field crew was generally deployed else-
logical profiles resulting from radiometric dating. where for fieldwork, and also during the middle
We think the sample selection process for Mer- winter months, when we rarely excavated on
cenaria probably skewed some of the resulting St. Catherines Island. More recent studies (e.g.,
seasonal estimates for certain sites. Quitmyer, Jones, and Andrus, 1997; Andrus
In subsequent fieldwork, we have empha- and Crowe, 2008) have demonstrated the value
sized the importance of defining microstrati- of collecting larger, more systematic samples.
graphic units for both radiometric and season- Several chapters in this volume report on more
ality sampling, attempting to process multiple recent research that has considerably improved
samples throughout these small-scale units (en- the modern control sampling on St. Catherines
countered in both large-scale excavations and Island.
column samples). In this way, we hope to avoid Use of proxy population studies is also im-
the problem of “averaging out” the variability portant when employing nonhuman vertebrate
we are seeking. remains to reconstruct seasonality, on St. Cath-
Correlative problems arise with modern erines Island and elsewhere. Although noting
techniques of isotopic analysis. As the results that sharks and sea catfishes are expected to
in seasonality studies become increasingly so- be relatively abundant during warm weather
phisticated and fine grained, we run the risk of and are relatively rare in colder weather, Reitz
“learning more and more about less and less.” (2008: 660–663) cautioned that “Vertebrates are
As the precision increases, sample sizes become not the best seasonal indicators … the absence
progressively smaller, raising the risk, again, of a seasonal marker can in no way be interpret-
that we are seriously underestimating the sea- ed as evidence that the site was unoccupied at
sonal variability we are seeking. Ideally, the that time (Reitz and Wing, 1999: 259–261) …
fine-grained, small-sample methods of isotopic As with most research, negative evidence is not
analysis can be balanced against the larger-scale helpful.” There are also biases in preservation
(if less precise) methods of growth increment that militate against the recovery of juvenile
documentation. Thompson and Andrus (2010) specimens.
provide an excellent example of this procedure When estimating postencounter return rates
in practice, and we believe that their research for fish species common to the Georgia coast
provides a model for increasing precision while (Thomas, 2008: 113–135), I relied heavily on
simultaneously addressing the need for main- fisheries studies (esp. Dahlberg, 1972, 1975),
taining large sample sizes. significantly augmented by ichthyofaunal sur-
veys conducted for a dozen years (on a monthly
Potential Problems with basis) on St. Catherines Island under the direc-
Proxy Populations tion of Bruce Saul, of Augusta State University
Modern control samples are critical as proxy (Saul, 2002, 2003, 2004). In Native American
populations for calibrating estimates of seasonal Landscapes of St. Catherines Island, I discussed
harvest in the remote past. The control sample the seasonal availability of hardhead (Arius fe-
of Mercenaria collected in the waters surround- lis) and gafftopsail (Bagre marinus) catfish, not-
ing St. Catherines Island has some obvious ing Saul’s concern about the declining numbers
strengths and weaknesses (O’Brien and Thomas, between 2002 and 2004. Three years later, both
2008: 481). Because the Mercenaria were col- catfish had virtually disappeared from the wa-
lected over a nine-year interval, we hoped that ters surrounding St. Catherines Island (Lambert,
this “longitudinal aspect” would help buffer the 2007; Bruce Saul, personal commun.). This sur-
skewing effects of unique seasonal events, such prising turn of events prompted Lambert (2007)
as phases of exceptionally cold or warm temper- to ask “where have they gone and why?” Perhaps
this is a short-term natural cycle, or maybe a virus archaeological and paleoclimatic investigation
took out the population. Pollution is an obvious (Harding et al., 2010). By comparing bivalve
possibility, with some speculating that perhaps geochemistry (particularly oxygen isotope data)
intake of heavy metals might be responsible.13 between modern oysters to those discarded in
Whatever the cause, the mysterious disap- early 17th century wells, these investigators
pearance of catfish near St. Catherines Island have successfully quantified estuarine salin-
underscores the complex ecological dynamics ity, seasonality of oyster collection, and annual
involved in collecting control (proxy) samples shifts in drought conditions in the Chesapeake
for site seasonality studies. It is a mistake to Bay ecosystem.
assume that any single “modern collection” It is difficult for the modern observer to
is an accurate proxy reflection of species bio- appreciate the magnitude of the hydrological
geography or pattern of seasonal distribution. change over this past century. But by looking
There are also issues, I believe, that must be closely at the historical sources and tracing out
addressed with respect to all aspects of control/ the surviving geomorphological evidence, we
proxy collections. believe it is possible to reconstruct what St.
Consider, for example, the problems of Catherines Island looked like before the deep
changing hydrological regimes and their im- drilling so significantly changed the hydrology.
pact on levels of seawater salinity. Based on We think that shifting hydrologies over the
soil distributions and the dendritic pattern of past century have significantly impacted the
relic groundwater outflows that once drained St. American (or eastern) oyster, Crassostrea virgi-
Catherines Island, Hayes and Thomas (2008) nica, still found in abundance within intertidal
reconstructed a large freshwater lagoon they estuaries, along saltwater rivers and tidal creeks
believe once dominated the central reaches of that dissect the expansive Spartina marsh.
the island (see also Bishop et al., 2007: 40, fig. Within a given locality, oyster growth depends
6). During the aboriginal period, freshwater was on bottom conditions, degree of salinity, water
always abundant on St. Catherines, available temperature, and tidal movement. A century
in numerous places, except during periods of ago, coastal Georgia was among the world’s
extreme drought. The extensive central depres- leading oyster harvesters, rivaling the celebrat-
sion was powered by the Floridan aquifer, one ed, oyster-rich waters of the Chesapeake (e.g.,
of America’s most productive groundwater res- Irving, 1902; see also Thomas, 2008: chap. 7):
ervoirs, which extends from South Carolina to By the 1880s, overharvesting and pollution
Florida and reaches inland as far as Alabama. threatened Georgia’s shellfishery. The head-
The deep confined aquifer discharged artesian waters of these oyster-bearing rivers were also
water to the ground’s surface in many places, once fed by freshwater aquifers, which dried up
and elsewhere, a relatively shallow well could due to the lowering of water tables over the past
tap the surficial reservoir of nonartesian water. century. Walker and Cotton (2001) suggest that
This was the hydrological regime available this loss of freshwater headwater could account
to the first St. Catherines islanders. Discharge for gradual movement of oysters away from
from this system remained in approximate equi- places like the Duplin River (off St. Simons Is-
librium so long as the upper Floridan aquifer was land) over time.
recharged by rainfall in the interior of the coast- Cannarozzi and Quitmyer (2007) have dis-
al plain, where it lay near the ground surface. cussed the monthly collection of oysters from
These artesian conditions created natural seeps, St. Catherines Island, continuing since 2006. As
with water flowing to the surface in springs and these modern control samples are amassed for
seeping into rivers, ponds, wetlands, and other isotopic study, one must wonder about whether
surface-water bodies throughout most of coastal shifting hydrological and salinity patterns could
Georgia. But over the past century, groundwa- have impacted the biochemical processes in
ter pumping has significantly lowered the water modern oysters. There is also the possibility that
level in the upper Floridan aquifer throughout ongoing global climate change—especially with
the entire coastal area. respect to sea level and water temperature—
Recent research at colonial Jamestown (VA) could impact modern oyster growth patterning.
highlights the potential of studying estuarine One potential avenue of inquiry into mol-
drought conditions by coupling fine-grained luscan proxy populations might be available
through the study of century-old bivalves. In (I believe). But we must do much better … and
Native American Landscapes of St. Catherines that’s what the Fifth Caldwell Conference is all
Island (Thomas, 2008: chap. 13), we discussed about.
the issues surrounding reservoir correction cal-
culations, and the necessity of obtaining known- NOTES
age, prebomb mollusc samples for 14C dating.
We knew that a commercial oyster industry had 1. I gratefully acknowledge the assistance of Lorann
once flourished in the waters surrounding St. Pendleton Thomas, Chelsea Graham, and Diana Rosenthal in
Catherines Island (Thomas, 2008: chap. 6). In preparing this manuscript.
2. In 1562 and 1564, the French established two colonial
the late 19th century, Augustus Oemler erected forts (Charlesfort and Fort Caroline) at opposite ends of the
an oyster factory on the south end of St. Cath- central Georgia Bight. Each fort was occupied for about a
erines Island. Oysters, collected by hand from year, and the subsequent Spanish towns of St. Augustine
nearby creeks and marshes, were prepared in and especially Santa Elena—in roughly the same territory—
a large boiler connected to the southern end of continued an even more significant contact with local Indian
populations after 1565. Following a brief period of Jesuit
the island by a causeway. Two additional boilers mission activity (in 1569–1570), the Franciscans launched a
were added later. The apparently inexhaustible more concerted effort in 1574–1575. But Mocamo was not
supply of oysters disappeared during the 1920s, truly missionized until 1587, and the major Guale missions
forcing the once flourishing oyster factories of were established in 1595–1605 (Jones, 1978; Worth, 1998,
2004, 2007).
St. Catherines Island to close. Today, the rusting 3. The most important sources addressing the nature of
boilers and massive spoil heaps of oyster shell Guale coastal adaptations include Swanton (1922, 1946),
remain visible evidence of this industry. Since Sturtevant (1962), Larson (1969, 1978, 1980), Jones (1978),
virtually all of the shells within these factory and Worth (2004, 2007); see also Bushnell (1990), Crook
middens derived from Crassostrea individuals (1986), Jones (1873), Hann (1986a, 1986b), and Saunders
(2000a, 2000b).
harvested between about 1900 and 1920, we 4. By this interpretation, the Guale dispersed still further
collected numerous samples and used them in in the springtime into nuclear family settlements spread
the reservoir correction computations for St. throughout the oak forest, located near swidden plots of
Catherines Island. maize, beans, and squash. In Crook’s “Annual Model” the
Guale “resided in towns from the first part of July until the
We think that a similar approach might be middle of September when they dispersed to gather nuts”
useful in considering the effects of hydrologi- (Crook 1986: 20), after which they once again dispersed,
cal change and shifting salinity over the past probably into a “matrilineage segment with four or five
century. Not only could we collect oysters that nuclear families forming the social core,” several related
lived under the previous (artesian) hydrological families living together in dispersed settlements located
throughout the oak forest, subsisting on acorns, hickory nuts,
regimen, but we have recently discovered that and communal deer hunts, with occasional aggregation for
occasional Mercenaria were also (accidentally) feasting (Crook, 1986: 21). Fall settlements were basically
harvested at the oyster factories, providing an- chiefly compounds “defined by towns composed of tempo-
other potential source of samples to explore this rary and changing populations, as opposed to the seasonally
stable population of the summer towns.”
issue. For St. Catherines Island (and the rest of 5. Whereas Fr. Rogel reported that the Guale dispersed
the Georgia coastline), we have detailed salin- seasonally to gather acorns, Jones (1978: 193) questioned
ity measurements taken between October 1888 whether it was necessary to abandon the “dispersed town”
and February 1889 (Drake, 1891). We think a in order to harvest the mast. With respect to shellfish, Jones
comparison of modern and century-old salinity notes that whereas some important Guale towns lay beyond
the oyster beds, “there is no documentary evidence that
levels and mollusc proxy samples could help [the Guale] spent seasonal periods downstream or along
determine the nature of recent hydrological and the inland waterways to exploit the oysters.” Citing Robert
salinity shifts over the past century. It might be Sandford’s 1666 account for the North Edisto River (South
that the century-old molluscs provide more ac- Carolina), Jones suggests that concentrations of oysters were
sufficiently close to the maize fields “that the beds could be
curate “modern controls” than proxy specimens exploited without seasonal shifts in residence” (Jones, 1978:
collected in the past few years (see Thomas, n.d., 193).
for more recent data on this issue). 6. Specifically with regard to the island-wide survey
We pose these two case studies to suggest sites, we also compared the existing ceramic and 14C chro-
some potential new refinements in site seasonal- nologies for St. Catherines Island (Thomas, 2008: chap.
15). A total of 189 radiocarbon dates had been processed
ity studies. For decades, reconstructing site sea- on archaeological samples from St. Catherines Island, and
sonality has been the warp and the weft of St. 110 of these dates—from 31 distinct mortuary and midden
Catherines Island archaeology. So far, so good sites—were directly associated with datable ceramic assem-
blages from the aboriginal phase. This diverse sample of 14C 9. This study is reinforced by an oxygen isotope study of
dates, which spanned more than four millennia, provides a modern and ancient clams from St. Catherines Island (An-
workable set of radiometric controls on the ceramic chronol- drus and Crowe, 2008).
ogy currently available for St. Catherines Island. The results 10. “To repeat: The available archaeological evidence
fine-tune, yet overwhelmingly, confirm the previous research does not permit the conclusion that Irene populations were
on the ceramic chronology for Georgia’s north coast, par- sedentary (although we certainly believe that such was the
ticularly the work of Joseph Caldwell, Antonio Waring, and case). But sticking to the documented archaeological specif-
Chester DePratter. ics, the data regarding site seasonality are conclusive: (1)
7. Some years ago, we reported the results of our excava- single-season sites are extremely rare during Irene times (as
tions of several Refuge-Deptford-Wilmington burial mounds they are rare throughout the entire aboriginal period on St.
on St. Catherines Island (Thomas and Larsen, 1979). We Catherines Island) and (2) three-quarters of the known Irene
analyzed 29 radiocarbon dates, nearly one-quarter of them components on St. Catherines Island were occupied during
processed on marine shell. Although aware of potential prob- three or more seasons” (Thomas, 2008: 1098).
lems associated with radiocarbon dates on sea shells, we fol- 11. “Patch types” reflect the way in which food resources
lowed the lead of Joseph Caldwell, who had also worked on were clumped or aggregated across the landscape (Thomas,
St. Catherines Island. Caldwell concluded that “radiocarbon 2008: 63). Specific “hunt types” are variously associated with
determinations made from oyster shell do not appear to differ one or more particular prey species, particular microhabitats
significantly from determinations made from charred wood” (“patch type”), specialized methods of search or capture, spe-
(Caldwell, 1971: 1). Today, we understand that this assump- cialized transport or foraging technology, and certain seasons
tion was incorrect. A significant reservoir effect is operating or environmental conditions (Thomas, 2008: 71).
here because, relative to the atmosphere, ocean water is 12. On occasion, as at Little Camel New Ground Field
depleted in 14C, transmitting this deficiency to marine organ- 5 (9Li206), we did employ 1⁄16 in. screens to sample the
isms. This means that 14C determinations processed on marine extremely dense concentration of fish bones encountered
samples will routinely appear to be older (in many cases, (Thomas, 2008: 519). But ¼ in. hardware was primarily
several centuries older) than 14C dates run on contemporary used in the island-wide survey. We note, parenthetically, that
terrestrial samples. We found it necessary to derive a unique Rochelle Marrinan made extensive use of fine-mesh screens
reservoir correction for St. Catherines Island and the results in 1973–1975, while excavating two Late Archaic shell rings
of this fieldwork and analysis are reported in Thomas (2008: on St. Simons Island (Marrinan, 1975, 2010). In this respect,
chap. 13; see also Thomas, Sanger, and Hayes, n.d.). she was considerably ahead of her time.
8. This apparent discrepancy arises because, in order to 13. In chapter 3 in this volume, E.J. Reitz, B. Saul, J.W.
derive an accurate and reliable local reservoir correction, Moak, G.D. Carroll, and C.W. Lambert assess the value of
we processed a dozen radiocarbon dates on modern time the modern ichthyofaunal survey data as an archaeological
prebomb samples. proxy for St. Catherines Island.
2012 A Bayesian Chronological Framework 37
CHAPTER 2
A Bayesian Chronological Framework
for Determining Site Seasonality
and Contemporaneity
D ouglas J. K ennett and B rendan J. C ulleton
In the previous chapter, Thomas outlined a Bayesian statistical environments to analyze and
multiscalar approach for determining seasonal model spatiotemporal patterns in the archaeo-
resource exploitation strategies and, by exten- logical record are also being developed includ-
sion, for reconstructing mobility patterns in the ing tools potentially useful for integrating the
past. On St. Catherines Island this involves: (1) multiscalar chronological measures necessary for
ceramic typologies for establishing submillen- establishing site seasonality and inferring mobil-
nial temporal resolution (the hour hand); (2) ra- ity patterns (Buck, 2004; Steele, 2010; Winter-
diocarbon dating to control century-level resolu- halder et al., 2010). In this chapter we empha-
tion (the minute hand); and (3) site seasonality size the development of a Bayesian chronologi-
studies to provide seasonal or monthly resolution cal framework for interpreting seasonality data
(the second hand). This hierarchy of chronologi- on St. Catherines Island. We begin with a basic
cal measures is essential when inferring mobil- overview of the approach.
ity patterns from seasonality data from multiple
sites. The crux of the matter is whether or not Bayesian Essentials
sites used during different or the same seasons
are coeval. Chronology really does make a differ- Classical statistical analysis has dominated
ence when interpreting seasonality data and in- archaeological inquiry and is well suited to a
ferences regarding prehistoric mobility patterns. wide range of observations made by archaeolo-
If changes in mobility and resource extraction gists (Thomas, 1986; Shennan, 1997; Drennan,
are detected, then finer grained chronological in- 2010). However, there are certain contexts where
formation is also required to determine if these a Bayesian approach may be better suited and we
changes were influenced by paleoenvironmental argue that the types of data acquired in seasonal-
change and/or other mechanisms (e.g., popula- ity studies fall into this category. In contrast to
tion growth, resource intensification, or competi- classical statistics, Bayesian statistical analysis
tion for resources). derives posterior information (a posteriori) by
In this chapter we argue that a Bayesian statis- combining prior information (a priori), a like-
tical environment provides a coherent framework lihood function (a particular probability func-
for integrating these multiscalar chronological tion), and the available data (Buck and Millard,
measures. The Bayesian analysis of radiocarbon 2004b: VII). The best examples in archaeology
dates from archaeological sites is now becoming come from chronology building where a variety
routine in Britain (Bayliss and Bronk Ramsey, of nonquantitative contextual information (e.g.,
2004; Buck, 2004; Bayliss et al., 2007) and pro- stratigraphic position, diagnostic artifact assem-
grams like OxCal (Bronk Ramsey, 1995, 2001, blages) can be integrated with probability dis-
2005) provide a prepackaged set of Bayesian tributions from radiocarbon dates (Bayliss and
statistical tools to help develop more robust ar- Bronk Ramsey, 2004; see below).
chaeological site chronologies. Custom-built Emphasis is placed on chronological model
building based on prior knowledge and the avail- onstrates the importance of high-precision dating
able radiocarbon dates followed by iterative for interpreting seasonality data. The low-resolu-
data collection to make adjustments. The major tion site chronology displayed in the first panel
benefits of this approach are that: (1) a statisti- (A) shows six “contemporary” sites dating to be-
cal environment is created that incorporates a tween a.d. 800 and 1000. The one larger village
wider range of observations and knowledge (e.g., site shows evidence for hard clam (Mercenaria)
relative and absolute chronology), and (2) these exploitation and processing during all seasons
models can be used to direct research and make along with evidence for a variety of other activi-
sampling decisions. Using a priori information ties. The five other sites are more ephemeral shell
can make some researchers uneasy (see Steier middens showing hard clam harvesting and pro-
and Rom, 2000), but if care and transparency cessing during different seasons. One interpreta-
are used it provides a framework to formalize tion of this would be year round occupation of
assumptions and to build and test multiple mod- the larger site with logistical exploitation of hard
els with new data. Shell middens are particularly clams throughout the year by a certain segment
complex (Stein, 1992), but agreement indices (A) of the larger social group (B). A finer grained
provide a way of determining how each alterna- chronology is shown in panels C and D, revealing
tive model fits with the available data, and are an alternative model of higher residential mobil-
generated for the posterior distributions of each ity between a.d. 800 and 900 followed by greater
radiocarbon date in a model, as well as the overall sedentism at one locality between a.d. 900 and
model itself (Bronk Ramsey, 2000: 201). Agree- 1000. The combination of high-resolution chron-
ment indices falling below a critical value (A'c = ological and seasonality data has a major impact
60%) indicate a poor fit of data with the model, on the socioeconomic and evolutionary infer-
and can be used to identify potential outlier dates ences made from these data.
or problematic stratigraphic assumptions in the Establishing precise chronologies is no simple
model. It should be noted that, strictly speaking, matter and is done best within a Bayesian statis-
when A > A'c (i.e., there is agreement between tical framework. English Heritage has used the
model and dates) it does not mean that the model approach since the mid-1990s as a cost effective
assumptions and structure are correct. It simply way of building site chronologies (Bayliss and
tells us that we have no reason based on the data Bronk Ramsey, 2004). The computer program
at hand to reject the model as it stands. OxCal provides a preexisting Bayesian environ-
ment to build and refine chronologies. Model
Building Site Chronologies building and testing is completed in an interac-
tive fashion with new data informing and refin-
The first step in developing a coherent picture ing chronological models for individual sites.
of site seasonality is establishing a viable chro- Precise dating is dependent upon: (1) careful
nology and determining the contemporaneity of stratigraphic excavation and the exact recording
different sites. When combined with seasonality of 14C samples within the depositional sequence,
data this allows for the reconstruction of chang- (2) the selection of short-lived organisms for
ing patterns of resource extraction at one loca- AMS radiocarbon dating (e.g., carbonized seeds,
tion. Broader patterns of settlement and land use twigs, marine shells, or animal bones), (3) proper
are often inferred from these data when combined chemical protocols for processing samples, and
with other regional datasets. The accurate recon- (4) an understanding of taphonomic processes
struction of mobility patterns at the regional level affecting potential radiocarbon samples. Outdat-
is dependent upon whether sites were used or oc- ed radiocarbon dates with high error ranges from
cupied during the same interval. This highlights previous excavations may be used as an initial
one of the first-order decisions necessary when guide for model development along with strati-
building a Bayesian chronological model for sea- graphic information. They may contribute to the
sonality studies: finding a meaningful definition final chronological model, but they could also be
of contemporaneity (Bayliss and Bronk Ramsey, eliminated as outliers. This is largely dependent
2004). On St Catherines Island this could be de- upon the research question being asked. When
fined as the “Archaic period” or it could be taken high precision is required, as is the case with
to be the same century or even the same decade. seasonality studies, it is often necessary to start
The hypothetical example in figure 2.1 dem- over with a clear idea of site stratigraphy, sample
types and locations, and the low analytical er- Trimming Confidence Intervals
ror afforded by adaptations to the latest genera-
tion of Accelerator Mass Spectrometers (Beverly We start with a simple advantage afforded
et al., 2010). Even with the best technological by Bayesian analysis: the trimming of cali-
and procedural protocols in place, chronological brated 14C radiocarbon confidence intervals us-
control is ultimately constrained by the vagaries ing the stratigraphic position of samples. This
of the calibration curve, which afford different is based on the simple combination of relative
levels of precision throughout time depending and absolute chronological information. The
upon atmospheric radiocarbon production and variety of factors influencing measurement er-
ocean circulation (see e.g., Blackwell, Buck, and rors in radiocarbon dating and fluctuations in
Reimer, 2006). the calibration curve can result in substantial
A B
A. D. 800–1000 A. D. 800–1000
C D
A. D 800–900 A. D. 900–1000
Village (Year Round) SHELL MIDDENS

Winter
Spring
Summer
Fall
Fig. 2.1. Hypothetical seasonality and settlement data (A) along with differing interpretations of these data
based on low (B) and high (C and D) resolution chronologies for these sites.
overlap of calibrated distributions even if two the lowest of the two. The calibrated probabil-
events are known not to be contemporary based ity distributions of these dates also overlap in
on the stratigraphic record. Take, for example, figure 2.3. Applying the same stratigraphic as-
two radiocarbon dates on charred hickory nuts sumptions as above, the calibrated probability
from the McQueen Shell Ring on St. Catherines distributions are trimmed and the dates fall into
Island, table 2.1: one near the base of the shell an acceptable stratigraphic order between 2100
midden deposit (Beta-251767; 3680 ± 40 b.p., and 1900 cal b.c. (A = 83.8%). The smoother
N243E233; 40 cmbs) and the other closer to the probability distributions and larger calibrated
top (Beta-251761, 3710 ± 40 b.p., N243E233; ranges for marine shell, relative to associated
10 cmbs) (Sanger and Thomas, 2010). The cali- carbonized seeds, result from the mixing model
brated 2σ probability distributions for these two used to derive the Marine09 calibration curve
dates broadly overlap (outlined probability dis- from the IntCal09 atmospheric curve (which, for
tribution in fig. 2.2), but from the law of super- most of the Holocene, are identical to the 2004
position we can argue that the lowest part of the versions), and propagation of uncertainty in the
sequence accumulated first, followed by the up- ΔR estimate (following Stuiver, Pearson, and
per materials. This stratigraphic information can Braziunas, 1986).
be used to create a posteriori distributions for Considering the same three dates in two ad-
these same dates (filled probability distributions) ditional contexts can show the importance of
that assume that the hickory nut at 40 cmbs must reasoned stratigraphic interpretation for forming
be older than the one at 10 cmbs. The model fit the priors in Bayesian chronological modeling.
with the data satisfies the threshold for accep- First, let’s say that we had reason to infer that
tance (A = 99.9%, A′c = 60%). In other words the entire deposit was laid down in one event,
this model is viable and indicates that this part of or for practical purposes within a single year or
the deposit accumulated rapidly between 2100 a few years. In that case we would assume that
and 2000 cal b.c. all the shells should be the same age, and scat-
Probability distributions of marine shell dates ter in conventional 14C ages arises from simple
from the same deposit show a similar story (see measurement error. These dates could then be
table 2.1). The upper two dates are associated combined following the procedure of Ward and
with the radiocarbon dates on charred hickory Wilson (1978), with a resulting averaged age of
nuts and the third date is stratigraphically below 3860 ± 25 b.p. A chi-squared test indicates that
TABLE 2.1
Radiocarbon Dates from McQueen Shell Ring, N243E233
Conventional 2σ cal age as 2σ cal age as
Beta no. Provenience1 Material 2σ cal age
14
C age sequence phase
charred
251761 4.5–4.4 m 3710 ± 40 2280–1970 b.c. 2150–1960 b.c.
hickory nut
charred
251767 4.4–4.3 m 3680 ± 40 2200–1940 b.c. 2210–2030 b.c.
hickory nut
Mercenaria
251762 4.5–4.4 m 3820 ± 50 2180–1850 b.c. 2100–1820 b.c. 2150–1880 b.c.
shell
Mercenaria
251768 4.4–4.3 m 3910 ± 40 2280–1960 b.c. 2150–1940 b.c. 2220–1940 b.c.
shell
Mercenaria
251769 4.3–4.2 m 3830 ± 40 2160–1870 b.c. 2210–1980 b.c. 2150–1900 b.c.
shell
1
Note that elevations are above an arbitrary 0 m datum.
there are no outliers in this group of dates (χ2 = to avoid oversampling a particular interval at the
2.757; df = 2; T' = 5.991) and we can infer that expense of other units that may be more interest-
the shells were deposited sometime between ing or significant chronologically.
2180 and 1930 cal b.c. (at 2σ). Alternatively, if Figure 2.4 shows the north wall profile of a
the stratigraphic relationships among the three unit excavated on the eastern side of the Mc-
dates were unclear in the field, the dates could Queen Shell Ring (N243E233). The sequence
be modeled as a phase rather than a sequence, is between 40 and 50 cm deep in this area and
with boundaries placed as estimates of two un- dominated by a dense deposit of shell (~30 cm
dated events: the start and end of shell deposi- thick) overlying a medium brown sand contain-
tion. Modeling as a phase might be more appro- ing no shell and a thinner deposit of shell and
priate for dates within a stratum that appears to bone (labeled clam “floor”). The samples dis-
be distinct and in order within the larger strati- cussed above come from the thicker shell depos-
graphic sequence, but could have been mixed it and their stratigraphic position is shown along
before being buried. Another application would with our first attempt at building a chronological
be dates from multiple discrete shell lenses or model for this part of the site. In this model the
hearths deposited on a common living surface, dates are ordered in a sequence and the start and
where the depositional order among them can- end dates for the sequence are modeled from the
not be discerned from the stratigraphy. If the available data. Marine shell samples are avail-
three shell dates are modeled as an unordered able for dating from the shell deposit and the
phase, we find that the stratum was deposited clam “floor” and ongoing work will determine if
over a span potentially as long as 800 years, charred seeds or twigs will also be available for
from ca. 2600 to 1550 cal b.c. (A = 105.9%). study. In certain contexts 14C dating marine shell
All that differs between these three scenarios are is preferable to other materials and it is often the
the prior assumptions that are used to model the only material available from certain stratigraphic
three shell dates: the first assumes an ordered contexts (Kennett et al., 2002; Thomas, 2008). It
stratigraphic sequence; the second assumes a is also short-lived and material deposited during
single deposition event; the third assumes an a single year or season can be sampled (but see
unordered group of dates, or a phase. That the Culleton et al., 2006, for complicating factors in
resulting chronological data can differ so great- areas with high upwelling).
ly based on a simple assumption drives home This Bayesian chronological framework pro-
the fact that proper stratigraphic interpretation vides a guide for future work. One or two dates are
in the field is of primary importance. The agree- needed to determine the age of the clam “floor”
ment indices provide no guidance in these sce- and they will help constrain the age of the inter-
narios as to which assumptions are more reason- vening sterile medium brown sand lens (along
able (nor does Ward and Wilson’s χ2 test), so it with the shell date just above it [Beta-251769]).
rests upon the fundamental archaeological skill An additional date from the uppermost layers in
of interpreting stratigraphy and site-formation the shell deposit will help constrain the terminal
processes. age of the deposits in this part of the site. Dating
a marine shell may be preferable in this instance
Toward Building a Bayesian Model due to the contamination of surface deposits with
for the McQueen Shell Ring modern carbonized plant material (see discussion
of precision below). The presence of pitting in
Funding is often limited for archaeological other parts of the site suggests horizontal stra-
projects and precision AMS radiocarbon dating tigraphy that should also be explored with ad-
is expensive. How precise a chronology needs to ditional dates. The decision to add more data to
be is directly related to the research question at the modeled chronology is based on the ability of
hand and for site seasonality, high precision is new data to refine the model. The method calls
generally required if the ultimate goal is to re- for continued AMS radiocarbon dating of qual-
construct mobility patterns. Seasonality studies ity materials from known stratigraphic units until
themselves are also time consuming and expen- the model can no longer be improved. In reality,
sive. It makes sense to establish a precise site there is often a trade off between model quality
chronology prior to conducting seasonality work (precision and accuracy), available funds, and
in order to define the relevant sampling units and time to publication.
Sequence McQueen Shell Ring {A = 99.9% (A'c = 60.0%)}
N243 E233 4.5-4.4 m 99.4%
N243 E233 4.4-4.3 m 100.5%
3000 B.C. 2500 B.C. 2000 B.C. 1500 B.C.
CALIBRATED DATE
Fig. 2.2. Calibrated results of a sequence of two 14C dates on hickory nuts from McQueen Shell Ring using
OxCal 3.01. Standard calibration is shown in open outline (prior distribution); modeled calibration incorporat-
ing stratigraphic information (posterior distribution) is depicted as a solid fill. The agreement index (A = 99.9%)
above the critical value (A′c = 60.0%) indicates good agreement between the data and the model.
Sequence McQueen Shell Ring {A = 83.8%(A'c = 60.0%)}
N243 E233 4.5-4.4 m Shell 107.0%
N243 E233 4.4-4.3 m Shell 94.2%
N243 E233 4.3-4.2 m Shell 73.0%
3000 B.C. 2500 B.C. 2000 B.C. 1500 B.C.
CALIBRATED DATE
Fig. 2.3. Calibrated results of a sequence of three 14C dates on Mercenaria shells from McQueen Shell Ring
using OxCal 3.01. Standard calibration is shown in open outline (prior distribution) and modeled calibration in-
corporating stratigraphic information (posterior distribution) is depicted as a solid fill. The agreement index (A =
83.8%) higher than the critical value (A′c = 60.0%) indicates good agreement between the data and the model.
4.7 O
ELEVATION (m)
Ap Beta 251761 & 251762 SD 1

4.5
SD 2 Clam “Floor”
Beta 251768 & 251767
4.3 A2
4.1 B Beta 251769 MBS
E232 E233 E234 E235

Fig. 2.4. Profile of the N243 trench north wall, McQueen Shell Ring (AMNH 696, 9Li1648), showing the
depositional sequence discussed in the text (courtesy of Matt Sanger). O high organic, dark brown sand; Ap
plowed A, dark brown sand, shell flecking; A2 buried A, dark brown, no shell; B medium yellow brown, no shell;
MBS medium brown sand; SD shell deposit.
A Note on Precision 14C of hard clams has already provided an important

and Seasonality Studies source of seasonality data on St. Catherines Is-
land (O’Brien and Thomas, 2008). The benefits of
The larger age ranges evident in the marine working with hard clam shells include their ubiq-
shells from the McQueen Shell Ring point to the uity in island archaeological deposits of all ages,
importance of precision and accuracy in build- the resistance of their thick shells to decomposi-
ing site chronologies. Radiocarbon dates with tion, and a clear growth structure that responds
large analytical error reflect poor precision, but to water temperature and associated changes in
also undermine accuracy by increasing the range the marine system in predictable ways. These
of actual calendar ages that could produce the same characteristics provide an opportunity for
measured age. This compromises our ability to precision AMS 14C dating and therefore a link
determine if two sites with seasonality data are between the centennial resolution of radiocarbon
contemporary or not. One crucial way in which dating and the seasonal or monthly data provided
this occurs can be illustrated with the Ward and by seasonality studies.
Wilson test described earlier, which is essentially If greater precision and accuracy are required
a test of contemporaneity between conventional to establish a chronological model for an ar-
14
C ages. Taking two McQueen shell dates, Beta- chaeological site, then additional AMS 14C dates
251769, 3830 ± 40 b.p., and Beta-251768, 3910 within the same hard shell can be combined with
± 40 b.p., the χ2 test does not reject the sample the counting of annual growth increments to
as containing outliers, so they may be contem- obtain a more precise age. The known chrono-
porary (χ2 = 2.000; df = 1; T' = 3.841). However, logical separation between samples based on ring
if the same dates had better precision, say ± 20 counting is incorporated into OxCal and a defined
14
C yr, the same two dates are strongly rejected sequence is used to “wiggle match” the radio-
as unlikely to be contemporary (χ2 = 8.000; df = carbon ages to the calibration curve to improve
1; T' = 3.841). Put simply, low-precision dates precision and accuracy (Bronk Ramsey, van der
are more likely to appear contemporaneous, even Plicht, and Weninger, 2001). This approach has
with a generally robust statistical test, than high- been used to obtain a very precise and accurate
precision dates. Although Stuiver, Pearson, and age on a prehistorically constructed subcircular
Braziunas (1986) have argued that the relative ring of oak timbers discovered in the intertidal
smoothness of marine calibration curves (caused zone in Norfolk, England (Bayliss et al., 1999).
by propagating multiple layers of statistical un- In the Norfolk case, six contiguous samples of 20
certainty) reduces “wiggles” and the frequency years’ growth were precisely dated and a Bayes-
of multiple intercepts, it is at the cost of wider ian model incorporating these dates and the
calibrated ranges that make it more difficult to known gaps between them was used to establish
distinguish true contemporaneity. Analytical pre- the date of construction at 2050 cal b.c.
cision can be improved with modifications to the In the case of hard clams on St. Catherines Is-
latest generation of Accelerator Mass Spectrom- land, the time elapsed between radiocarbon sam-
eters (Beverly et al., 2010). Samples in the age ples in a single hard clam shell can be determined
range of the McQueen Shell Ring could be re- with annual growth increments. Modern specimens
duced to ± 15 and 20 14C years with this approach from protected areas along the southeastern coast
(Kennett et al., 2011). However, the precision are known to live up to 30 years (Ansell, 1968;
and accuracy of marine samples will always be Quitmyer and Jones, chap. 7, this volume), but the
lower than paired terrestrial samples because of maximum age of individuals from prehistoric con-
the mixing models and the ΔR estimates used in texts on St. Catherines is about 10 years with most
the calibration process. in the 3–4 year range (see fig. 7.9 in Quitmyer and
Bayesian statistical tools may provide one Jones, this volume). This increment just satisfies the
possible way of improving the precision and ac- minimum required to match the resolution of the
curacy of marine shell dates to make them more calibration curve (3–4 years; Bronk Ramsey, van
comparable to terrestrial samples. The approach der Plicht, and Weninger, 2001: 388). Hard clam
would rely on the visible and predictable incre- shells are composed of alternating white (opaque
mental changes in hard clam (M. mercenaria) under tungsten light) and dark (translucent under
shells and the “wiggles” in the marine calibration tungsten light) bands that represent sequential an-
curve. Visible incremental change in the growth nual cycles. Growth rates are faster in winter and
spring (white growth) and slower during the sum- same age and calibrated with the atmospheric
mer months (dark growth). Water temperature and calibration curve.
associated biotic shifts appear to influence growth Running multiple AMS 14C dates on the same
rates (Henry and Cerrato, 2007; Jones, Arthur, and clam shell is costly and this would only be war-
Allard, 1989; Jones and Quitmyer, 1996; Quitmy- ranted if associated carbonized twigs or seeds
er and Jones, chap. 7, this volume, Cannon, and were not available. However, marine shells are
Jones, 1985; Quitmyer, Jones, and Arnold, 1997) short lived, so dating the shells would be prefera-
and this has been confirmed with stable isotopic ble to 14C dating wood charcoal or bulk sediments
studies (Andrus and Crowe, 2008; Quitmyer and (see Sanger and Thomas, 2010, regarding prob-
Jones, chap. 7, this volume). The growth rings are lems with 14C dating bulk sediments). Depending
clearly visible and can be counted to determine a upon the age of the shell, and the character of the
clam’s age at the time of harvest. A couplet of one marine calibration curve at that time, a substan-
opaque and one translucent increment represents tial improvement may be obtained with just one
one year of life. additional measurement. In the case of wiggle
Figure 2.5 shows a cross section of a fictitious matching, more wiggly parts of the curve are ac-
hard clam shell displaying incremental growth. tually more of a boon than a bane because they
This clam was 10 years old when it was harvested constrain the match more tightly. To evaluate the
during the winter (O1, white [opaque], December potential for wiggle matching, terminal growth
through January). From the perspective of sea- margins should be 14C dated first and then a de-
sonality, we are interested in establishing a pre- fined sequence may be modeled with OxCal us-
cise age for this and associated hard clam shells. ing that date plus a number of simulated dates at
AMS 14C dating the carbonate near the terminal specified gaps to determine if (or how many) ad-
growth margin of the shell (sample A) will ac- ditional dates from the shell would be worthwhile
complish this, assuming that the marine reservoir (the above example was created in this way). The
does not fluctuate radically throughout the an- specific advantage of working with hard clams
nual cycle as it does in other parts of the world is that this provides a close linkage between the
(Kennett et al., 1997; Culleton et al., 2006). This subcentennial scale precision of AMS 14C dat-
assumption needs to be empirically tested, but ing and the seasonal data provided by hard clam
it seems a reasonable working hypothesis since growth increment studies.
marine upwelling (a source of old carbon) along
the coast of St. Catherines appears to be minimal Contemporaneity and Archaic
given the bathymetry, and the negative local res- Period Shell Rings
ervoir offset (ΔR = –134 ± 26 14C yr) indicates
better than average mixing of atmospheric car- Most applications of Bayesian models for
bon compared to the global marine model age. chronology building have focused on single lo-
We use a fictitious set of dates here based on the cations where the relationship between samples
marine shell dates from the McQueen Shell Ring is known and stratigraphic context is built into
discussed above. The calibrated age of the ter- the model. The ultimate goal of most seasonal-
minal growth margin ranges between 2090 and ity studies is to reconstruct prehistoric mobility
1960 cal b.c. (130 yrs at 1σ, shown as outline in across a landscape. This is dependent upon de-
fig. 2.5). Adding two additional dates (B and C) termining whether two archaeological sites are
5 and 10 years earlier in the growth of the clam contemporary or not. More sophisticated models
creates a sequence of dates that can be used to that add this spatial component are now being de-
wiggle match with the marine calibration curve veloped (Buck, 2004; Winterhalder et al., 2010),
to get a more precise age. In OxCal this is accom- but these types of studies are in their infancy and
plished by creating a defined sequence with an es- require custom-built Bayesian statistical environ-
tablished set of known age gaps based on counted ments not available in prepackaged programs
growth increments (Bronk Ramsey, 2001: 383). (e.g., OxCal).
The posterior or modeled age distribution for the As a starting point for work on St. Catherines,
terminal growth margin, based on wiggle match- we investigate the contemporaneity of the Mc-
ing, is reduced to 1980–1910 cal b.c. (70 yrs at Queen and St. Catherines shell rings within a
1σ, shown as a solid distribution in fig. 2.5). This Bayesian statistical framework. Shell rings are
is comparable in age to a carbonized seed of the relatively common in the American southeast and
are often the oldest sites found in coastal regions sites are contemporary and seasonality studies
(Russo, 2006). They generally date to the Late demonstrate year-round occupation, then this
Archaic period (4000 to 1000 cal b.c.). Two of observation has interesting social and political
these shell rings have been identified on St. Cath- implications.
erines Island: the St. Catherines and McQueen Summed probability distributions are useful
shell rings. The St. Catherines Shell Ring is po- heuristic devices and can be helpful in focusing
sitioned on the west coast of the island and has research questions, though their statistical mean-
been excavated extensively. A series of 14C dates ing actually remains poorly understood. We have
are available for study (Thomas, 2008: 370). The no metric to evaluate how much overlap in two
McQueen Shell Ring is located on the east coast summed distributions constitutes contemporane-
of the island and is currently being excavated, ity or temporal disjunction. In the present case, it
but preliminary results are now available for suggests the potential for simultaneous occupa-
comparison (Sanger and Thomas, 2010: 62–63). tion of the two shell rings over a broad period,
A comparison of 14C summed probability profiles which runs counter to the observation of differ-
from these two sites suggests that both were oc- ent material culture assemblages at each site that
cupied from ca. 2600 to 1800 cal b.c. However, could suggest two temporally distinct occupa-
the material culture differs substantially at these tions. With a Bayesian model, we can take the
sites, particularly in the types of ceramic deco- prior knowledge of the differing assemblages as
rations identified and the relative proportions of a basis for arguing that one shell ring precedes
groundstone and decorative items found at each the other, and there is no overlap in the occupa-
site (Sanger and Thomas, 2010: 67). If the two tions. At a first pass, we simply model the cul-
B: 5 yr before collection
3740 ± 20 B.P.
C: 10 yr before collection
3790 ± 20 B.P.
A: Terminal margin
3840 ± 20 B.P.
Prior Posterior
D_Sequence Mercenaria WM
(1σ cal) (1σ cal)
C: -10 yr, 3790 ± 20 B.P. 120.6% 2010-1900 B.C. 1990-1920 B.C.
Gap 5
B: -5 yr, 3740 ± 20 B.P. 76.5% 1960-1830 B.C. 1985-1915 B.C.
Gap 5
A: 0 yr, 3840 ± 20 B.P. 69.9% 2090-1960 B.C. 1980-1910 B.C.
2200 B.C. 2000 B.C. 1800 B.C. 1600 B.C.

CALIBRATED DATE
Fig. 2.5. Schematic cross section of a hypothetical Mercenaria valve showing annual growth increments and
three 14C dates used in a defined sequence to wiggle match the terminal date (A). Results of the wiggle match,
performed with OxCal 3.01, are depicted below. Standard calibrations are shown in open outline (prior distribu-
tion) and modeled calibrations incorporating known five-year gaps in the defined sequence (posterior distribu-
tion) are depicted as a solid fill.
turally relevant and reliable 14C dates from St. events, also provide interesting information. The
Catherines and McQueen as two phases (groups start of Pre-Ring phase cultural deposition is
of unordered dates) within a sequence that forces placed at 3100–2450 cal b.c., the beginning of
the St. Catherines phase to precede the McQueen Ring phase deposition at 2430–2200 cal b.c., and
phase. The advantage to this approach is that an the end of Ring phase deposition is at 2210–1780
agreement index (A) is calculated that gives a cal b.c. The Interior Plaza phase overlaps consid-
measure of model fit, and provides a statistical erably with both the Pre-Ring and Ring phases,
basis for accepting or rejecting the model and its being estimated to start between 2980 and 2500
assumptions. Perhaps unsurprisingly, this model cal b.c. and to end between 2400 and 1800 cal
thoroughly fails to accommodate the data (A = b.c. This result accords with descriptions of the
22.6%; A'c = 60.0%), which means we must re- Interior Plaza deposits as appearing to be more
ject the assumption of sequential and discrete oc- mixed, and potentially being a locus for deposi-
cupations of each shell ring. While this may seem tion through both phases of site use.
a minor insight in light of the overlap observed The depositional sequence of the McQueen
in the summed probabilities, it may also suggest Shell Ring is in some ways more complicated.
that the differing cultural assemblages in each Focusing on a series of dates from three exca-
shell ring may not be as temporally diagnostic as vation units (TPII, N243E233, and N272E200),
thought, or that their associations with the dated we can build a nested model of sequences and
materials could be reconsidered. phases for the shell ring. The overall period of
Further steps in refining this analysis de- deposition is considered a phase, which includes
pend on more detailed stratigraphic models for an unordered group of three sequences compris-
St. Catherines and McQueen shell rings. For ing the dates in each unit. We do this because
St. Catherines the 14C dates that are acceptable we are making no assumptions about the strati-
(based on the excavators’ knowledge of the de- graphic relationships between units, only within
posits) can be grouped into three categories: units. Within two of the unit sequences we have
Pre-Ring; Ring; and Interior Plaza. The first two pairs of shell and charcoal dates in a few levels;
groups have an intuitive temporal relationship these pairs are treated as phases within the unit
implied in their names, and the third represents sequences, again because they are unordered
a more temporally ambiguous group, where the with respect to each other, though they both
spatial association with the rest of the shell ring should postdate those below and predate those
doesn’t allow a firm stratigraphic argument to above. This is a somewhat involved model for
be made with respect to the other two groups of a relatively small number of dates, but its struc-
dates. Leaving the Interior Plaza aside for the ture actually reflects a fairly conservative set of
moment, we can treat the Pre-Ring and Ring stratigraphic assumptions, all of which are open
dates as phases in a sequence, similar to our to evaluation. Figure 2.7 shows the modeled re-
treatment of St. Catherines vs. McQueen shell sults, and we are immediately alerted by a poor
rings above. In this case, the assumed order is agreement index (A = 17.8%) for the overall
based on stratigraphic relationships rather than model. Phase models are more accommodating
diagnostic artifact assemblages, but the model than sequences, so we would expect the problems
structure is similar. Boundaries are modeled rep- to lie in one of the three unit sequences, and from
resenting the beginning of Pre-Ring deposition, the individual agreement indices it’s clear that
the transition from Pre-Ring to Ring deposition the upper and lower dates in TPII (A = 7.2% and
(i.e., the initial construction of the St. Catherines 10.9%, respectively) are problematic. The dates
Shell Ring), and the end of Ring deposition. The in this unit appear to be completely reversed,
Interior Plaza dates can be grouped into a sepa- and our assumed depositional sequence appears
rate phase independent of the other two, with to be incorrect. This information could direct the
boundaries for its beginning and end. The results excavators to reevaluate the context from which
of these models are depicted in figure 2.6 and the dated shells were taken, sample selection
from the agreement index of A = 84.6% we can criteria, or even to recheck the sample labels or
see that the assumption of all the Pre-Ring dates artifact inventories. For the time being, however,
preceding the Ring dates appears to be a reason- we can revise the model by removing these three
able one, based on the data at hand. The bound- dates and see if it behaves differently.
aries, which are estimates of additional undated Without the dates from TPII, the agreement
Sequence {A = 84.6%(A′c = 60.0%)}

Boundary End St. Caths Ring
Phase St. Caths Ring
215822 Crass 68.6%
21408 Merc 99.0%
215823 Crass 96.9%
229424 Shell 113.7%
229423 Shell 113.8%
238327 Hickory Nut 111.8%
238337 Burnt Wood 80.3%
Boundary Btwn Pre-Ring and Ring
Phase St. Caths PreRing
231334 Shell 96.5%
215824 Crass109.0%
215821 Crass107.1%
231335 Shell 104.2%
238336 Shell 62.6%
Boundary Start St. Caths
PreRing
Sequence {A = 94.6%(A′c = 60.0%)}

Boundary End St. Caths Interior Plaza
Phase Interior Plaza
233130 Shell 102.6%
238331 Burnt Wood 93.3%
238322 Hickory Nut 102.6%
238332 Burnt Wood 102.5%
239276 Charred material 100.5%
238328 Burnt Wood 86.2%
Boundary Start St. Caths
Interior Plaza
3500 B.C. 3000 B.C. 2500 B.C. 2000 B.C. 1500 B.C.
CALIBRATED DATE
Fig. 2.6. Two sequences incorporating the most reliable 14C dates from the Archaic in St. Catherines Shell
Ring. The first models two phases (unordered groups of dates) defined as “Pre-Ring” or “Ring” within a se-
quence that assumes no overlap between the two phases and are separated by a boundary. Standard calibration
is shown in open outline (prior distribution) and modeled calibration incorporating phase designations (posterior
distribution) is depicted as a solid fill. The agreement index (A = 84.6%) higher than the critical value (A′c =
60.0%) indicates good agreement between the data and the model. Below, a third group is placed in an indepen-
dent phase of Interior Plaza dates, with boundaries to estimate the beginning and end of that phase of deposition.
Substantial overlap with both Pre-Ring and Ring phases supports the view of continued deposition in the Interior
Plaza in both phases. See Sanger and Thomas, 2010, for raw data.
index for the revised McQueen model is A = Shell Ring deposition is estimated between 2520
85.4%, indicating a good fit between the observed and 2100 cal b.c., and the ending between 1970
data and the model. Also, the overall distribution and 1580 cal b.c., which overlaps substantially
of dates and boundaries is not greatly altered by with the Ring phase at St. Catherines (starting
removing the three problematic dates, other than at 2430–2200 cal b.c.) though McQueen likely
making the boundary estimates for the beginning persisted past the end of the Ring phase as cur-
and end of deposition a bit broader, as they are rently modeled. If the set of McQueen dates in-
constrained by fewer dates. The start of McQueen cluded in the model is representative of the site,
Sequence {A = 17.8%(A'c = 60.0%)}

Boundary Start McQueen
Phase McQueen Comprehensive
Sequence TPII
238325 TPII Bottom 10.9%
238326 TPII Middle 83.5%
238324 TPII Top 7.2%
Sequence N243E233
251769 4.3-4.2 m Shell 60.3%
Phase 4-4-4.3 m
251768 4.4-4.3 m Shell 103.8%
251767 4.4-4.3 m Hick 107.5%
Phase 4.5-4.4 m
251761 4.5-4.4 m Hick 90.0%
251762 4.5-4.4 m Shell 114.9%
Sequence N272E200
Phase 5.1-5.0 m
251765 5.1-5.0 m Shell 94.8%
251766 5.1-5.0 m Char 93.3%
251764 5.3-5.2 m Char 81.9%
Boundary End McQueen
2500 B.C. 2000 B.C. 1500 B.C.

CALIBRATED DATE
Fig. 2.7. A phase of nested sequences incorporating the most reliable 14C dates from the Archaic in McQueen
Shell Ring. Dates from within individual units can be placed in sequences based on their stratigraphic relation-
ships, but the three units cannot be cross-correlated, so they are collectively modeled as an unordered group of
sequences, i.e., a phase. Standard calibration is shown in open outline (prior distribution) and modeled calibra-
tion incorporating phase designations (posterior distribution) is depicted as a solid fill. The agreement index (A
= 17.8%) well below the critical value (A′c = 60.0%) indicates poor agreement between the data and the model.
Low individual agreement indices point to reversals in unit TPII, which should be removed from subsequent
analyses and the context reconsidered in the field. See Sanger and Thomas, 2010, for raw data.
it would appear that there is no occupation con- for determining if the various sites on the island
temporary with the St. Catherines Pre-Ring phase are contemporary or not, critical information
at McQueen. Further dates from McQueen and when comparing seasonality data and inferring
a refined stratigraphic picture may improve our past land use and mobility patterns.
interpretations, both by constraining the deposi- One of the benefits of the approach is that it
tional boundaries and adding more relationships provides a framework for future research and,
between dates in the various excavation units, most importantly, for establishing and testing
which should provide a more detailed sense of alternative chronological models. We have ex-
occupational changes through time. amined whether or not the two Late Archaic pe-
riod shell rings on St. Catherines are contempo-
Conclusions rary between 2430 and 2200 cal b.c. within this
framework. Based on the available data we argue
Seasonality studies depend upon defining cul- that they are most likely contemporary. This is
turally meaningful analytical units and analyz- a similar result to the summed probability ap-
ing and interpreting seasonality data within the proach used by Sanger and Thomas (2010), and
broader context provided by these units (Monks, adds strength to this hypothesis. However, the
1981: 223). Chronology building is an essential changes in material culture defined by Sanger
part of defining these analytical units. We argue and Thomas could have occurred in under a de-
that the Bayesian approach defined in this chap- cade during the Late Archaic and more precise
ter provides a viable statistical framework for AMS 14C dates could provide additional insights
combining stratigraphic information and other into the cultural processes in play. We argue
prior knowledge with radiocarbon dates to es- that these chronological improvements are best
tablish more precise and accurate chronological accomplished in an iterative fashion within the
models for archaeological sites on St. Catherines Bayesian statistical framework established in
Island. It also provides a set of tools and indices this chapter.
2012 INTERPRETING SEASONALITY FROM MODERN AND ARCHAEOLOGICAL FISHES 51
CHAPTER 3
INTERPRETING SEASONALITY
FROM MODERN AND ARCHAEOLOGICAL FISHES
ON THE GEORGIA COAST
E lizabeth J. R eitz , B ruce M. S aul , J.W. M oak ,
G wendolyn D. C arroll , and C harles W. L ambert 1
Estuarine fishes are a mixture of endemic, ma- location. He suggests that season of availability
rine, and freshwater species attracted to estuaries and season of activity are linked, but cautions
as nursery grounds and feeding areas. The poten- against associating these with season of location,
tial of estuaries to fill these roles varies season- specifically season of site occupation and settle-
ally, raising the possibility that fishes can serve as ment patterns.
indicators or ecological groups that can be used Despite Monks’s caution, it is likely that
to consider seasonal patterning in human fishing more seasonality studies have focused on season
strategies. In this chapter, modern spatial and sea- of location and seasonal components of politi-
sonal habits of fishes on the Georgia coast (USA) cal, economic, and other social institutions than
are used by analogy to assess seasonal fishing ac- on season of availability or activity. This leaves
tivities in the archaeological past. Despite prob- the impression that seasonal aspects of availabil-
lems common to all ecological analogies, we find ity and activity are resolved, whereas it is more
that fishing was highly selective, that the most likely that fishing strategies in the past specifi-
ubiquitous fish taxa in the archaeological record cally targeted fish communities and populations
of the Georgia Bight are not markedly seasonal, available throughout the year, albeit vulnerable
and that the overall record is marked by stabil- to a variety of techniques in different estuarine
ity rather than change. This limits the value of habitats. Questions of availability and activity are
fish presence and ubiquity to provide information subtle and unlikely to be answered by examining
about seasonality in human behavior and changes fish population characteristics such as seasonal
in that behavior over time. Although settlement variations in population structure, abundance,
patterns should not be inferred uncritically from biomass, diversity, dominance, distribution, or
season of availability, the evidence that the fish body size without collaborative evidence from
taxa most ubiquitous in the archaeological record other sources.
generally are available throughout the year with- The first of Monks’s (1981) three aspects of
in these estuaries raises doubts about the ability seasonality, the availability of fishes, is the focus
of unicausal economic, residential, and foraging of this chapter. Our goal is to associate the niches
models to capture the complexity of human life and habitat preferences of modern fishes with
in such environments. those of high-ubiquity fishes present in the ar-
In his seminal review of seasonality studies, chaeological record to derive season of availabil-
Gregory Monks (1981) critiques common meth- ity using modern data from St. Catherines Island
odologies used to study seasonality and offers a and Cumberland Sound, located behind Cumber-
framework for integrating such studies. He argues land Island (fig. 3.1). The modern data from these
that seasonality studies should begin by asking locations show a clear seasonal pattern in fish
the question: seasonality of what? Monks identi- availability measured by the number of fish in-
fies three major responses to this question: season dividuals (fig. 3.2). Thus, we approach the ques-
of availability, season of activity, and season of tion of seasonality of fishing on the premise that
fishing strategies were informed by seasonal mi- dicator taxon or indicator groups, but such com-
gratory patterns though these may be difficult to plex data are unavailable for the current study.
interpret when viewed through the lens of the ar- Interpretation of archaeological evidence re-
chaeological record. This leaves the season of ac- lies on ecological and ethnographic analogy; both
tivity and season of location unresolved, though of which should be used with caution. Ecological
we have some thoughts on those as well. analogy uses behaviors of organisms in response
Monks (1981) reviews much of the direct and to critical environmental parameters that can be
indirect evidence used to study seasonality. In his observed today to infer habitats and niches in the
discussion of indirect evidence, he critiques the past. Ethnographic analogies accomplish a simi-
many cultural variables that circumscribe season, lar goal using recent, observable human activi-
activity, and location (see also, Morales Muñiz, ties to interpret former behaviors. Changes in the
1998). His review of these variables is thorough, anthropogenic, biogeochemical, and hydrologic
and is not repeated here, other than to emphasize environments affect all aspects of the ecological
that the archaeological record is a cultural one. and archaeological record, and, consequently, all
Likewise, many of the biases associated with analogies must be viewed skeptically.
archaeofaunal data from sites associated with
aquatic ecosystems, especially coastal ones, are ESTUARIES AND FISHES
reviewed elsewhere (e.g., Erlandson, 2001; But-
ler and Campbell, 2004; Reitz and Wing, 2008; Our focus is on the section of the southern
Broughton, 2010; Campbell and Butler, 2010) U.S. coast of the Atlantic Ocean known as the
and these critiques are not repeated here. The Georgia Bight, a large embayment extending
complexity of seasonality in anthropogenic con- along the coast from Cape Fear, North Caroli-
texts extends far beyond basic, descriptive data. na, to Cape Canaveral, Florida (fig. 3.1; Hayes,
This study relies on Harry Kenward and Allan 1994). A series of barrier or sea islands forms a
Hall’s (1997: 665) definitions of indicator groups boundary between the coastal waters of the At-
and taxa. Indicator or ecological groups are com- lantic and lagoons (locally known as estuaries).
binations of organisms defined taxonomically or The distance between a barrier island and the
by some other common element (e.g., habitat, sea- mainland can be as much as 7 km. Barrier islands
sonal preference). Kenward and Hall (1997: 665) share similar Pleistocene and Holocene histories,
define an indicator taxon as “one which reliably as well as archaeological, biological, and phys-
carries the implication of the occurrence of some iographic characteristics (Hoyt, 1967; Hoyt and
event, activity, or ecological condition in the past” Hails, 1967; Johnson et al., 1974: 11; Thomas et
and an indicator group as “a natural grouping of al., 1978; Oertel, 1979; Howard and Frey, 1985;
organisms selected because it includes a range Frey and Howard, 1986; Hayes, 1994; Thomas,
of stenotopic species which together encompass 2008; Bishop, Rollins, and Thomas, 2011). Bar-
a wide spectrum of ecological conditions or hu- rier islands are between 5 and 17 km long and 1
man activities relevant to the aims of the study and 6 km wide (Hubbard, Oertel, and Nummedal,
being carried out.” Indicator taxa, single species 1979). Island elevations generally are less than 7
typical of, perhaps even restricted to, specific m, though individual dunes may be higher (John-
niches may not be as useful as indicator groups son et al., 1974: 11). St. Catherines and Cumber-
for studying environments or human behavior. land islands are two of the largest barrier islands
Estuarine organisms with strong preferences for in this area.
specific combinations of temperature, oxygen The central portion of the Georgia Bight, cor-
levels, water depths, or other characteristics are responding with the Georgia coast, is character-
more informative than are eurytopic organisms ized by large, complex, highly productive estuar-
with broad tolerances. ies. These semienclosed lagoons have access to
Monks advocates an integrated, multiproxy, the ocean through deep sounds that separate the
regional approach, termed by Kenward and Hall barrier islands from each other. Fresh waters from
(1997: 665) as indicator packages, which they large, perennial rivers draining the mainland mix
define as “a collection of recordable data of any with ocean water through tidal action, resulting in
kind which, when occurring together, can be ac- biogeochemical properties that are intermediate
cepted as evidence of some past state or activity.” between salt water and fresh water and change
Indicator packages are far more useful than an in- during each tidal cycle (Odum and Barrett, 2005:
Cape
Hatteras
NORTH CAROLINA
Cape Lookout
SOUTH CAROLINA Wilmington

Cape Fear
North Island
Sa
ine
va
L Cape Romain
nn
ll
ah
Fa
Ri
Charleston
ve
r
GEORGIA ATLANTIC
Savannah
Alta
OCEAN
mah Ossabaw Island
aR
iv St. Catherines Island
er Blackbeard Island
Sapelo Island
St. Simons Island
Cumberland Island
Kings Bay Locality
Ribault and Jacksonville Electric Company

S t . J o h n s River
FLORIDA Fountain of Youth

St. Augustine
Cape Canaveral
0 100
KM
Fig. 3.1. Location of sites mentioned in this study. Modern cities are indicated by blocks and archaeological
sites by circles.
421). This coastline is considered a low-energy erage of 20° C (Dahlberg, 1972; Dahlberg, 1975:
to moderate-energy, mesotidal system (Hubbard, 7–8). The turbid inshore waters of the Georgia
Oertel, and Nummedal, 1979). Estuaries in this Bight are generally no more than 10 m deep.
area have the lowest energy levels in the Georgia A distinction is made between estuaries and
Bight due to shoreline alignment, wind direction, coastal waters. Coastal waters are divided into
and the broad continental shelf (Johnson et al., inshore waters and offshore waters. Inshore wa-
1974: 89; Frey and Howard, 1986). Wave height ters extend eastward from the mouths of sounds
is less than 0.1 m, somewhat larger toward the and the sandy beaches of barrier islands (John-
northern and southern ends of the Bight (Hub- son et al., 1974: 86). Beaches bordering the sea-
bard, Oertel, and Nummedal, 1979). Inlets and ward sides of barrier islands extend east to ca.
marshes along the Georgia coast experience the 1–2 m depths. On St. Catherines Island, the beach
greatest tidal ranges in the Georgia Bight because zone extends seaward as much as 1000 m from
these back-barrier lagoons are larger and more North Beach and as little as 100 m from South
complex than those in South Carolina or Florida Beach (fig. 3.3). Temperatures range from 7° to
(Hubbard, Oertel, and Nummedal, 1979). The av- 29° C and salinities range from 25 to 31 ppt ex-
erage tidal range is 2 m with a range of 1 to 3 m cept when floodwaters reduce salinity (Dahlberg,
in spring (Schelske and Odum, 1961; Hubbard, 1972; Dahlberg, 1975: 113).
Oertel, and Nummedal, 1979; Howard and Frey, Estuaries are characterized by mosaics of
1985). A spring high tide may produce a 50% in- tidal creeks, sounds, salt marshes, and marsh is-
crease over mean tide level (Frey and Howard, lands. Estuaries are divided into major habitats
1986). A neap tide is absent in some areas, par- by salinity, water depth, dissolved oxygen, tem-
ticularly in high marshes adjacent to the main- perature, turbidity, and bottom topography, as
land. Surface water temperature extremes occur well as other biogeochemical and hydrological
in January and August. In the estuaries associated properties (Schelske and Odum, 1961; Dahlberg,
with Sapelo and St. Catherines Islands, water 1972; Dahlberg, 1975: 4–11; DEIS, 1978: D-411;
temperatures range from 8° to 32° C, with an av- Frey and Howard, 1986). One of these habitats
25,000
NUMBER OF FISH INDIVIDUALS
20,000
15,000
10,000
5000
0
JAN FEB MAR APR MAY JUN JUL AUG SEP OCT NOV DEC
Fig. 3.2. Monthly total catch, North and South beaches.
is the lower reach of each estuary, including the These habitats serve as nurseries and feed-
sounds that separate barrier islands from each ing areas for endemic, marine, and freshwater
other and provide access to the sea. These are species so that population structure, numbers of
the widest and deepest habitats. Waters are poly- species and individuals, and the average size of
haline; salinities range from 18 to 32 ppt and fishes vary seasonally within fish communities
temperatures from 8° to 32° C (Dahlberg, 1972; (Dahlberg and Odum, 1970; Conover and Ross,
Dahlberg, 1975: 113). The high marsh includes 1982; Rogers, Targett, and Vansant, 1984; Levin
the salt marshes and creeks of the upper littoral et al., 2001). Young fishes are more flexible and
zone. Salinity levels in high marsh range from 15 better able to tolerate the wide variations in es-
to 30 ppt (Dahlberg, 1975: 7). Middle and upper tuarine conditions than are adults, which may be
reaches of estuaries primarily are composed of able to use the estuary only for short periods of
narrow, shallow creeks and marshes with salinity time. Many species are not permanent residents
ranges between 0.3 and 29 ppt and temperature of estuaries; they spawn in the ocean or in fresh
ranges from 8° to 32° C (Dahlberg, 1972; DEIS, water and their young use estuaries as nursery
1978: D-412). Salinity increases in oligohaline grounds. Some species live offshore as adults
creeks with rising tides but is diluted by freshwa- and only use estuaries as feeding grounds. Very
ter runoff and rainfall. Salinity ranges from 0 to few spend their entire life cycle within an estuary.
14 ppt and temperatures range from 4° to 32° C The function of estuaries as nurseries for the an-
(Dahlberg, 1972; Dahlberg, 1975: 7, 114). Fresh- nual crop of juvenile fishes presents an important
water creeks and ponds lie beyond the tidal reach. feeding opportunity to predators, though proper-
These generally are on the mainland, though the ties such as salinity and temperature discourage
barrier islands also have freshwater sources. The some potential predators from taking advantage
final two habitats are low-salinity and high-salin- of this opportunity. Young fishes also are protect-
ity tidal pools. Low-salinity pools receive more ed from many diseases and parasites by the ex-
freshwater runoff and are flooded only during tremes presented within estuaries. The seasonal
high tide, whereas high-salinity pools are flooded dynamics of these roles reinforce the premise that
regularly by tides and rarely receive fresh water. fishes can serve as proxies for seasonal patterning
Salinity ranges in these pools are 0 to 34 ppt and in human fishing strategies.
temperatures range from 7° to 32° C (Dahlberg, The use of salt marshes and nearshore zones
1972). Some high-salinity pools are associated by fishes is often dependent on spawning cycles
with beaches, filling at high tide and perhaps iso- (Conover and Ross, 1982). Nearshore areas along
lated from the ocean at low tide, trapping some barrier island beaches are breeding grounds for
fish within them. small prey fishes, such as mummichogs (Fundu-
Characteristics that underlie seasonal patterns lus heteroclitus), which then attract larger preda-
in feeding, growth, and reproduction include tors (Martin et al., 2004). Food species, such as
water temperature, dissolved oxygen, turbidity, bay anchovies (Anchoa mitchilli) and Atlantic
salinity, predation, nutrients, and similar bio- menhadens (Brevoortia tyrannus), spawn either
geochemical and hydrological properties. Many in winter or early spring months, replenishing
are highly variable on daily, seasonal, and an- the food supply for larger predatory fishes that
nual cycles and organisms that live in estuaries move into estuaries during the warmer months
are adapted to significant and frequent changes (Reis and Dean, 1981; Conover and Ross, 1982;
in such properties. For example, when freshwa- Fives, Warlen, and Hoss, 1986). One of Georgia’s
ter discharge from mainland rivers is reduced important predatory species, both now and in the
by droughts, high-salinity waters and associated past, is the red drum (Sciaenops ocellatus), which
organisms may be found as far as 40 km inland exemplifies this behavior. Red drums spawn
(Frey and Howard, 1986) and the reverse is true offshore, inshore, and in estuaries between mid-
during freshwater floods. The lower reaches of August and October, typically in areas with sa-
estuaries, closer to the ocean, are more saline than linities above ca. 30 ppt (Lowerre-Barbieri et al.,
the upper reaches and less subject to the coldest 2008). Lowerre-Barbieri et al. (2008) report that
temperatures (Johnson et al., 1974: 86, 89; Dahl- many red drums are associated with deeper pools
berg, 1975: 7). Variations associated with each instead of shallower shoreline areas. Juvenile
tidal cycle approach seasonal and annual ranges red drums subsequently mature in estuarine and
(Dahlberg, 1972; DEIS, 1978). shallow nearshore waters while schools of adults
move between inshore and offshore areas. Rogers, Although we can define membership in this
Targett, and Vansant (1984) report that numerous indicator group, and do so in this chapter, we do
species use upper estuarine nursery areas near not know where the preferred fishing grounds
freshwater river discharges in the Ossabaw Sound were within each estuarine system when the ar-
estuary, at the northern end of Ossabaw Island. chaeological site under study was occupied be-
Other patterns may be related to river discharge, cause of the ever-changing nature of the estua-
temperature, food availability, or depth (Rogers, rine environment itself in response to changes in
Targett, and Vansant, 1984; Smith and Wenner, climate, sea level, sedimentation, spit formation,
1985; Jassby et al., 1995). St. Catherines is differ- dune building, freshwater hydrology, and other
ent from most barrier islands because it does not geomorphic processes that affect biological pro-
receive a significant influx of fresh water from a cesses (Blanton and Thomas, 2008; Bishop, Rol-
nearby large river, contrasting with such islands lins, and Thomas, 2011; chap. 1, this volume).
as Ossabaw, Sapelo, St. Simons, and Cumberland
that receive inputs of fresh water from rivers such METHODS
as the Ogeechee, Altamaha, and St. Marys (Oer-
tel, 1979; Howard and Frey, 1985). Modern fish samples from St. Catherines Is-
land are from collection sites on the North and
BACKGROUND PREMISES South beaches (fig. 3.3). These beaches lie on the
exposed seaward side of the island. The samples
Modern data from both St. Catherines and were taken between 1998 and 2009. Sites were
Cumberland islands show clear seasonal patterns selected to provide representative samples from
in the abundance of fish individuals, with different the entire maximum accessible area and to incor-
patterns for beaches compared to estuarine reach- porate as many differences in bottom topography
es that probably reflect the seasonal migrations of as possible. Initially four sites were selected on
juvenile and adult fishes. Thus, we approach the each beach. One site on North Beach was elimi-
question of seasonality and fishing based on the nated and another was occasionally inaccessible
background premise that fishing strategies were due to shifting mud, snags, and oyster beds, phe-
informed by migratory patterns that define sea- nomena that undoubtedly have always influenced
son of availability, though we cannot prove this fishing in this and other coastal areas. Each site
empirically. Archaeological indicator groups, or- was sampled on a monthly basis during the three-
ganisms with similar requirements, may provide hour period before low tide, allowing for collec-
evidence of temporal or spatial choices, or indi- tion in morning, midday, evening, night, and pre-
cate a specific collection strategy. dawn periods. Samples were collected under all
The other background premise is that a basic weather conditions, excluding lightning storms,
fishing strategy was followed within the Geor- extreme cold, and rough surf, which would en-
gia Bight and was a major aspect of economic danger the health and safety of the research team
life (see references in table 3.1). This strategy or make sampling impractical.
changed remarkably little over the millennia Species composition was assessed at each site
despite changes in Holocene climates and other by towing seines in a standard quarter-haul fash-
aspects of biogeochemical and hydrological pa- ion from the beach. Two hauls were made using a
rameters. The fishes that form the core of this 30.5 m monofilament seine, with a 2.5 cm mesh,
fishing strategy are present in most, if not all, ar- and two hauls were made with an 18.3 m cotton
chaeological collections. This indicator group of mesh bag seine of 0.3 cm mesh (Hillman, Davis,
fishes generally dominates archaeological collec- and Wennemer, 1977). Each tow was offset 9 to
tions in terms of minimum number of individu- 18 m from the previous one to reduce the effects
als (MNI) and biomass. Fishes in this indicator of population disbursement associated with the
group move within estuaries or between estuar- previous pull.
ies and offshore habitats in response to a suite of The 0.3 cm mesh net used in the bag seine was
seasonal variables associated with reproduction designed to collect all sizes of all possible game
and feeding. The presence of this indicator group and nongame species. Differences in nets in the
in an archaeological deposit could be evidence modern St. Catherines Island study likely do not
that specific environmental parameters prevailed pose a bias in this study. The 0.3 cm mesh used in
when, and where, these animals were captured. the modern study captured all but the very small-
TABLE 3.1
Summary of Archaeological Sites and Vertebrate Collections1
Sites Location Time period Screen NISP MNI No. No. fish
size (mm) taxa taxa
Back Creek Village, Mississippian (Irene) St. Catherines a.d.

9Li207 Island, GA 1300–1580 3.18 14,881 449 53 27
Bourbon Field, Mississippian (Savannah) Sapelo Island, a.d.
9Mc71 GA 1000–1350 6.35 15,331 563 35 19
Cathead Creek, Mississippian (Savannah) a.d.
9Mc360 Darien, GA 1200–1500 0.5 2,248 84 34 16
Cathead Creek, Woodland (Swift Creek) a.d.
9Mc360 Darien, GA 300–700 0.5 1,610 74 27 19
St. Augustine, 1450–500
Fountain of Youth Park, Archaic 8SJ31 FL b.c. 1.59 215 28 15 11
Fountain of Youth Park, Miss. (St. Johns St. Augustine, a.d.
IIc) 8SJ31 FL 1513–1565 1.59 14,891 218 36 27
Jacksonville Electric Authority, Miss. a.d.
(Savannah) 8Du634, 8Du669 Duval Co., FL 1000–1500 1.59 7,380 179 30 16
Kenan Field, Mississippian (Savannah) Sapelo Island, a.d.
9Mc67 GA 1000–1500 6.35 0 397 44 15
Kings Bay Locality, Mississippian Kings Bay a.d.

(Savannah) 9Cam171a, 9Cam177 Locality, GA 1200–1500 1.59 36,667 903 57 26
Kings Bay Locality, Woodland (Swift Kings Bay a.d.
Creek) 9Cam171a, 9Cam177 Locality, GA 300–700 1.59 37,530 1,704 46 27
St. Catherines 2300–1950
McQueen Shell Ring, Archaic 9Li1648 Island, GA b.c. 3.18 43,240 1,153 58 36
Meeting House Field, Mississippian St. Catherines a.d.
(Irene) 9Li21 Island, GA 1300–1520 various 6,193 105 23 7
Mission Sta. Catalina de Guale (Historic) St. Catherines a.d.
17th
9Li13 Island, GA century various 43,206 204 54 16
Pueblo Sta. Catalina de Guale (Historic) St. Catherines a.d.
17th
9Li8 Island, GA century various 14,493 167 50 14
Ribault Clubhouse, Archaic 8Du76 Duval Co., FL 2000 b.c. 3.18 2,885 144 37 26
St. Catherines 2540–2030
St. Catherines Shell Ring, Archaic 9Li231 Island, GA b.c. 3.18 49,587 1,249 69 31
St. Simons Cannon’s Point (Marsh) Ring, St. Simons 2240–1815
Archaic 9Gn57 Island, GA b.c. 3.18 19,970 345 45 25
St. Simons 2320–1970
St. Simons West Ring, Archaic 9Gn76 Island, GA b.c. 3.18 9,453 251 31 16
Sapelo Shell Ring III, Unit 9, Archaic Sapelo Island, 2332–1740
9Mc23 GA b.c. 1.59 12,132 182 27 21
1
NISP refers to number of identified specimens (vertebrates only). MNI refers to minimum number of individuals (verte-
brates only); “No. taxa” refers to the total number of vertebrate taxa each collection; and “No. fish taxa” is the number of fish
taxa in each collection. Data are from the following sources: Bourbon Field (Crook, 1978, 1984; Reitz, 1982); Cathead Creek
site (Reitz and Quitmyer, 1988; Quitmyer and Reitz, 2006); Fountain of Youth site (Reitz, 1991); Jacksonville Electric Author-
ity site (Lee et al., 1984; Reitz, Quitmyer, and Marrinan, 2009); Kenan Field (Crook, 1978); Kings Bay Locality (Reitz and
Quitmyer, 1988; Quitmyer and Reitz, 2006); McQueen Shell Ring (Colaninno, 2010: 153–155; Sanger and Thomas, 2010);
Meeting House Field (Reitz and Dukes, 2008); Back Creek Village (Bergh, this volume); Mission (Eastern Plaza Complex)
Santa Catalina de Guale (Reitz et al., 2010); Pueblo Santa Catalina de Guale (Reitz et al., 2010); Ribault Clubhouse (Quitmyer
and LeFebvre, 2004; Reitz, Quitmyer, and Marrinan, 2009); St. Catherines Shell Ring (Colaninno, 2010: 116–119; Sanger and
Thomas, 2010); St. Simons Cannon’s Point (Marsh) and West Rings (Marrinan, 1975; Reitz, Quitmyer, and Marrinan, 2009;
Colaninno, 2010: 95–97, 100–101; Marrinan, 2010); and Sapelo Shell Ring (Thompson, 2006; Colaninno, 2010: 104–105).
Descriptions of each site, the methods used during excavation, and the original zooarchaeological studies of each collection are
found in the above sources. All of these materials were identified using the comparative skeletal collections at the Florida Mu-
seum of Natural History and the Georgia Museum of Natural History. Swift Creek is a Woodland period; Savannah, Irene, and
St. Johns IIc are Mississippian periods. Human MNI estimates are excluded.
est juvenile fishes. Zooarchaeological collections,

of course, only contain skeletal elements and not
live, whole fish; thus archaeological excavations
must use a smaller screen size to recover the re-
4 mains of the same small taxa that can be captured
live using a larger mesh. For example, the largest,
intact, skeletal element of a mummichog with a
total length of 86 mm is 12.5 mm. As a live fish,
this mummichog would be captured in the bag
seine used in the modern study.
The fishes in each haul were identified, enu-
merated, and measured. Larger specimens, par-
ticularly if part of a breeding cohort, were mea-
sured first and returned to the water to reduce
1 sampling mortality rates. Measurements were
2 taken using a Wildco fish measuring board with
1.0 cm divisions (Wildlife Supply Company, Yu-
lee, FL). Specimens that could not be identified
in the field were taken to the lab and preserved
in a 10% formalin solution. Fishes were identi-
fied to species using marine field guides (Breder,
1929; Dahlberg, 1975; Robins et al., 1986; Hoese
5
and Moore, 1998). A representative sample of
3
all sizes was measured when a large number of
specimens of a single species were collected (≥
6 25) and the remaining specimens were counted.
The following discussion generally focuses
on data from South Beach because, thus far, rela-
tively few archaeological sites have been found in
the North Beach area (Thomas, 2008: 546). Most
of the archaeological sites on the eastern side of
the island are closer to South Beach, though very
few sites are actually adjacent to either North or
7 South Beach.
Because South Beach data are from a sea-
ward beach, they are supplemented with data
from Cumberland Sound, a complex back-bar-
rier system landward of Cumberland Island, for
which similar quantified modern data are avail-
S.C.
able (figs. 3.1, 3.4). This enables us to assess
Georgia seasonal pulses in the shallow waters behind a
barrier island (DEIS, 1978). Fishes were collect-
magnetic
F.L. ed quarterly at four stations along each transect
using a 5 m semiballoon otter trawl and at shal-
north
1 MI 0 1 low water stations (A, C, X, D, E, F, G, I, J, K,

4000 FT 0 4000 L, M, N) using a 50 ft bag seine. Migratory fish
1 KM 0 1 collections were made in winter and spring quar-
Fig. 3.3. Map of St. Catherines Island, showing ters using three 100 ft gill nets with ¾ in., 1½
North and South beaches. Key to sites: 1, Meeting in., and 3 in. meshes set near transects D and E
House Field; 2, St. Catherines Shell Ring; 3, Santa for a complete tidal cycle (DEIS, 1978: D-192).
Catalina de Guale mission and pueblo; 4, North The methods used in the Cumberland Sound
Beach; 5, McQueen Shell Ring; 6, Back Creek Vil-
study are described in more detail in the Draft
lage; 7, South Beach.
Environmental Impact Statement for Cumber-
land Sound (DEIS, 1978: 186, 192–193). Occa- two locations. The modern collection strategies
sionally, specific reference is made to biological were not designed to replicate each other or hu-
sampling stations D and E, located in that part of man behavior in the past. The St. Catherines
the sound known as Kings Bay. Island modern data are from a collection local-
Differences in sampling location and sam- ity that may not have been frequently used by
pling gear for the modern St. Catherines Island islanders (seaward beaches), but the collection
and Cumberland Sound collections limit our abil- gear (manual seines) is probably similar to what
ity to directly compare modern data from these was used in the past. The Cumberland Sound data
N
A Crooked River B
00 KM
KM 55
D
X
LAN
D C
D IS
Ki
ng
E
sB
LAN
ay
ATLANTIC
BER
F
OCEAN
CUM
Mill G
Creek
R
J
I K
St. Marys River
H P Q
N O
L
AMELIA S
ISLAND
M
Fig. 3.4. Aquatic biological sampling stations in Cumberland Sound (modified from DEIS, 1978: D-187).
represent a collection locality similar to that used where, the animal was captured. The absence of
by people living in this region (back bay waters any resource, or the absence of evidence that a
and tidal creeks), but using a modern collection resource was used during a specific season, as
gear (mechanized trawls) unknown until recently. we all know, is more problematic (e.g., Balasse
From an archaeological perspective, it would be et al., 2003). To emphasize this point, we refer
highly desirable to collect modern samples from only to “presence” henceforth, using presence
archaeologically probable collection localities of taxa as reported in zooarchaeological reports
using probable collection gear. (table 3.1). We focus on those fishes that form
The archaeological collections are from 19 our hypothesized core group in the Georgia Bight
sites or temporal components of multicomponent fishing strategy (Quitmyer and Reitz, 2006; Re-
sites between St. Catherines Island and Anasta- itz, Quitmyer, and Marrinan, 2009; Reitz et al.,
sia Island, Florida (table 3.1; fig. 3.1). Five of 2010). Variables that affect presence will be re-
these sites (McQueen Shell Ring, Back Creek visited below.
Village, Bourbon Field, St. Simons Cannon’s An ideal study would combine presence with
Point [Marsh] Ring, and St. Simons West Ring) information about fish size and other indicators
are from the seaward side of barrier islands. The of age, such as increments. The tidal and sea-
Bourbon Field and the St. Simons Island sites are sonal movement of fishes within estuaries and
protected from the ocean by additional, smaller, between estuarine and coastal waters is closely
barrier islands (Blackbeard and Little St. Simons associated with age and body size. Age cohorts
islands) and two St. Catherines Island sites are are seasonal markers in fishes because habitats
adjacent to an inlet (McQueen) and a marsh and niches change as individuals grow larger
(Back Creek Village) on the seaward side of the and reach reproductive size. Members of an age
island. Other barrier island sites (St. Catherines cohort of a given species not only are roughly
Shell Ring, Meeting House Field, Mission and the same size at the same time, they also occupy
Pueblo Santa Catalina de Guale, Sapelo Ring III, similar habitats and niches at the same time, and
and Kenan Field) are located on the landward are susceptible to capture with similar fishing
sides of barrier islands. The remaining sites are gear. As their size changes with maturity, so too
mainland sites (Cathead Creek, Kings Bay, Rib- do the habitats and niches of the fishes, and the
ault, Jacksonville Electric Authority, and Foun- fishing strategy likely to efficiently capture them.
tain of Youth). Archaeological collections were These variables are likely to be somewhat differ-
deposited between ca. 2560 b.c. and a.d. 1680. In ent among those members of a taxon that occupy
the case of sites occupied prior to the onset of the ideal habitats and those that occupy the slightly
First Spanish period (a.d. 1565), it is not known less optimal edges of the species biogeographical
if any were contemporaneous or what relation- range. Sadly, too few measurements of archaeo-
ships among each site’s occupants might have oc- logical fish remains, or associations of size with
curred. Establishing contemporaneity is clearly age, are available for the Georgia Bight to use in
required to link seasonality in the resource base this study. Morphometric and incremental studies
with residential patterns (see Kennett and Cul- should become routine in all future archaeologi-
leton, chap. 2, this volume). cal projects.
Monks (1981) identifies presence-absence of Both the modern South Beach and the ar-
seasonally available species as the simplest, old- chaeological fishes are examined using ubiquity
est, and most common methodology and this is analysis and Shannon-Weaver diversity indices.
one of the methods used in the archaeological Ubiquity analysis of modern fishes involves di-
part of this study. Presence-absence relies upon viding the number of samples in which an indi-
ecological analogy, using the seasonal habits of vidual fish species is present by the total number
fishes today to infer seasonal behaviors in the of samples. Ubiquity analysis of archaeological
past. Fishes are expected to move within estuar- data involves dividing the number of archaeolog-
ies or between estuaries and either freshwater or ical collections in which a specific fish taxon is
offshore habitats in response to a suite of season- present by the total number of archaeological col-
ally sensitive variables associated with reproduc- lections in this application (N = 19). Only those
tion and feeding. The presence of a fish in an taxa for which MNI is estimated are used in the
archaeological deposit could be evidence that a archaeological ubiquity analysis. This approach
specific suite of parameters prevailed when, and controls for the diverse attribution levels in the
archaeological species lists and is similar to the cles. These have both short- and long-term out-
ubiquity analysis of modern collections, which is comes. Routine ecological processes (e.g., com-
based on individual fishes. petition, predation, community transformations)
The Shannon-Weaver diversity index is esti- alter the composition and physical location of
mated for each modern and archaeological site populations and communities. All of these pro-
using the following function: H' = –Σ Pi log Pi, cesses can occur without human intervention;
where Pi is the proportion of individuals in the but it is likely that the structure and productivity
i-th species (Dahlberg and Odum, 1970). Shan- of estuaries in the Georgia Bight have been im-
non-Weaver is also estimated for the combined pacted by human fishing strategies for millennia
archaeological assemblage by adding the indices (Reitz, 2004).
for all of the archaeological collections in table Thus, the suite of estuarine organisms found
3.1 together and dividing by the total number of at a specific time and place may change, either
collections (N = 18), excluding Kenan Field for for a long time or a short time, sometimes within
which MNI is unavailable. Shannon-Weaver al- a matter of hours. Organisms previously common
lows us to compare proportions and not numbers, in a specific fishing ground may now be rare and
accommodating characteristics in the data and taxa formerly absent from a location may now be
differences in techniques (see Reitz and Wing abundant there. In reference to the present study,
[2008: 110–113] for a review of this method). however, it is important to recognize that we
A t-test compares the overall archaeological di- may not know what the landscape at the time of
versity and modern diversity. Analysis of vari- occupation was or which specific habitats were
ance (ANOVA) compares seasonal differences of exploited from a specific site at a given point in
modern data and archaeological collections. time. This, of course, is the conundrum faced by
all diachronic and synchronic studies of seasonal
Caveats behaviors, ecosystems, and environments.
Biological surveys typically define seasons by
It bears repeating that comparisons of modern the lunar calendar: January–March, April–June,
and archaeological data rely on the premise that July–September, and October–December. Many
relationships exist among the organisms studied, characteristics are associated with seasons be-
their environmental preferences, environmental yond the lunar calendar and these vary within a
conditions at the site, and human behavior. This range from one year to the next. It is likely that
premise underlies the use of ecological analogies storms, climate cycles, and other dynamics in-
to infer relationships in the past and is basic to fluenced cultural and ecological systems at sev-
studies that associate season of availability with eral organizational, spatial, and temporal scales
other aspects of human behavior. In the Georgia during the millennia encompassed in this review
Bight, long-term processes operate within an (Redman, Grove, and Kuby, 2004; Reitz, Quit-
ecosystem in which populations and communi- myer, and Marrinan, 2009). Fishing strategies,
ties accommodate daily variations during each residential patterns, and behaviors of fishes them-
tidal cycle. Coastal landscapes are highly dynam- selves undoubtedly responded to these changes
ic, which means that environmental proxies and but might not occur precisely during the winter,
ecological analogies must be used cautiously. spring, summer, and fall months defined by the
Many analogies assume that organisms have Gregorian calendar.
not changed their habitats and niches during the This diachronic ambiguity is compounded by
intervening centuries, a premise that is unlikely synchronic Holocene changes in sea levels, sedi-
to be true (e.g., (Webb, Hedges, and Robinson, mentation, distance from fluvial sources, shore-
1998). In the case of St. Catherines Island, geo- line configurations, littoral drift, aeolian sand
logical and climatological processes have altered accumulations, migrations of tidal inlets, storm-
the structure and function of aquatic and terres- related topographic changes, biogeochemical
trial environments (Blanton and Thomas, 2008; properties, and sea surface temperatures, among
Bishop et al., 2011; Rich, Vega, and Vento, 2011). other changes (Dolan, Hayden, and Lins, 1980;
Many nonanthropogenic phenomena are associ- Liu, 2004; Reitz, Quitmyer, and Marrinan, 2009;
ated with environmental change, such as floods, Bishop, Rollins, and Thomas, 2011). The broad
storms, plant successions and other ecosystem general similarities among fish populations and
processes, coastal remodeling, and climatic cy- communities in both modern and archaeological
collections suggest that fish taxa represented in defined as having an adult total length of less than
today’s estuaries were able to endure synchronic ca. 250 mm today (Reitz, Quitmyer, and Marri-
changes just as they are able to endure diachronic nan, 2009; Reitz et al., 2010: 234–237). Large-
ones. Certainly, however, ecosystem structures bodied taxa are ones with an adult total length
and processes were impacted by synchronic over 250 mm. The mean total length (TL) of in-
changes, which are variables in this record of sea- land silversides (Menidia beryllina) is 79 mm,
sonal patterning; one of many reasons why it is bay anchovies 54 mm, striped killifishes (Fun-
difficult to associate seasonal aspects of resource dulus majalis) 62 mm, and striped anchovies (A.
availability with human behavior in the past us- hepsetus) 73 mm in the modern St. Catherines Is-
ing ecological analogies. land collection (table 3.2; Augusta State Univer-
The present study is only possible if one pre- sity field notes; see also, Jorgenson and Miller,
sumes that the overall estuarine characteristics 1968). It is unlikely that these four taxa would
required by aquatic animals were met someplace be represented in an archaeological collection if
within the environment, though the exact physi- a screen size of more than 3 mm was used during
cal and morphological aspects with regards to the archaeological fieldwork. Many of the archaeo-
archaeological sites investigated are unknown. logical collections were recovered using mesh
At the risk of a circular argument, while we can- sizes too large to catch the small bones of these
not define the temperature/salinity regime that small-bodied taxa. For this reason, and because
prevailed at a specific time and place, or specify of the susceptibility of small fish elements to site
where inlets, tidal creeks, sounds, salt marshes, formation processes, it is possible, if not probable,
marsh islands, and shallow waters were, the suite that very small taxa, as well as small individuals
of fishes recovered from St. Catherines Island of large-bodied taxa, are underrepresented in the
shows remarkable richness as well as continu- archaeological record. Without measurements
ity. The dominant fish taxa, those in the indicator and body size estimates for all fishes in all of the
group of fishes, are specifically those known for archaeological collections merged in this study, it
behavioral flexibility, a characteristic necessary is not possible to know which specific archaeo-
for all organic life in the Georgia Bight, includ- logical individuals were large or small, unless the
ing people. specimen is from a biologically small fish taxon.
Fish biologists identify most of their catch to Another limitation of all fishery studies is the
species and zooarchaeologists rarely do. It is spe- question about how people used small fishes, if
cies that respond to daily, seasonal, and annual they did. Today, we eat anchovies and other small
variations, not genera or families; but species- fishes, bones and all, and it is possible that peo-
level identifications often elude zooarchaeologi- ple did so in the past. Very small fishes, however,
cal studies. Thus, we may know we have a sea could have been used as bait instead of food. We
catfish (Ariidae) in an archaeological collec- question whether bait fish would be present at
tion, but not which of two possible sea catfish an archaeological site; it is more likely that bait
species (Ariopsis felis, Bagre marinus). Much was used shortly after capture, or tossed into the
more information is available about the habitat water at the end of the fishing expedition. Ad-
preferences of the hardhead catfish (A. felis), for ditionally, as this and other contributions in this
example, than for the family Ariidae. Likewise, volume demonstrate, small-bodied fishes can be
mullet (Mugil spp.) specimens in archaeological prominent components of archaeological collec-
collections almost always are attributed only to tions from the Georgia Bight, too prominent to
the genus Mugil because it is difficult, if not im- be dismissed as gut contents of predators, given
possible, to distinguish between striped mullets that many of these presumed predators are them-
(M. cephalus) and white mullets (M. curema). selves rare or absent in the archaeological record.
It is equally difficult to distinguish between the Regardless of whether or how small fishes were
two species of mullets in modern, live-collected used, the season of availability would be infor-
fishes, many of which are present in the modern mative of human behavior.
South Beach collection. Differences between modern seines and devic-
Recovery method is a well-known source of es used in the past are highly important sources of
bias in archaeological analyses. Four of the 10 bias. Such differences grow in significance when
fishes with ubiquity above 0.44 in the modern one recognizes that gear is tailored to prey, with
South Beach collection are small-bodied taxa, specific gear used for specific purposes. Gear dif-
TABLE 3.2
Seasonal Total Length (TL) Ranges (in mm) for high-ubiquity taxa from South Beach 2012
Winter Spring Summer Fall

Common name Scientific name Min Mean Max N Min Mean Max N Min Mean Max N Min Mean Max N
Inland silverside Menidia beryllina 6 82 122 2151 10 90 121 3327 6 71 121 1699 32 73 102 1246
Bay anchovy Anchoa mitchelli 40 58 85 1543 15 62 520 2384 24 48 480 2878 20 50 83 1860
Southern Menticirrhus 30 101 365 223 20 106 372 220 1 65 520 1776 19 68 370 544
kingfish americanus
Striped mullet Mugil cephalus 22 99 446 275 19 121 452 67 19 127 362 829 29 166 455 778
Striped killifish Fundulus majalis 39 57 81 56 44 70 135 188 20 59 113 870 34 61 95 100
Florida pompano Trachinotus 41 74 254 55 14 35 227 466 10 75 245 3954 15 63 419 1222
carolinus
Silver perch Bairdiella 116 143 232 113 45 148 270 203 129 160 199 18 112 143 195 84
chrysoura
Star drum Stellifer 21 43 97 49 19 47 160 42 19 78 440 420 26 105 167 22
lanceolatus
Striped anchovy Anchoa hepsetus 59 89 120 18 43 63 118 347 26 60 122 443 32 81 116 137
Hardhead catfish Ariopsis felis 292 292 292 1 226 331 421 69 42 130 550 345 280 333 384 11
Spot Leiostomus 82 188 271 9 60 100 182 16 70 154 231 49 115 208 262 32
xanthurus
Atlantic bumper Chloroscombrus 0 0 0 0 78 97 111 13 8 63 162 621 32 69 147 480
chrysurus
Red drum Sciaenops 370 509 723 33 405 507 640 14 20 367 535 15 350 508 680 15
ocellatus
Atlantic Brevoortia 114 179 235 24 125 164 443 11 61 134 212 150 142 177 283 40
menhaden tyrannus
Gulf kingfish Menticirrhus 64 254 467 40 29 199 370 21 62 152 262 18 60 255 411 13
littoralis
INTERPRETING SEASONALITY FROM MODERN AND ARCHAEOLOGICAL FISHES
Atlantic Strongylura 275 402 470 7 127 359 445 8 212 302 423 22 260 394 495 40
needlefish marina
Bluefish Pomatomus 29 73 160 3 18 61 345 95 35 150 280 20 34 172 257 3
saltatrix
63
ferences are particularly profound when modern by their nature, circular. This issue plagues all
trawl data, taken from sounds, are used to interpret ecological and ethnographic analogies.
archaeological fishing strategies. Anticipating the There are no obvious solutions to any of these
results of this study, there is no evidence that off- caveats; but it is in recognition of these issues
shore habitats or deep inshore sounds were used that statistical tests of significance are avoided;
in the past. Net size does influence catch, but nets they would confer on the results of this study an
used in the modern St. Catherines collection, at aura of legitimacy that is unwarranted. Clearly,
least, were selected purposefully to catch a large well-designed, regional, multiproxy studies de-
variety of species of all sizes. The merit of the signed to contribute to definitions of indicator
St. Catherines modern study is that seines were packages would make substantial contributions
used to sample two localities whose parameters to resolving some of the problems with ecologi-
are similar to some of those probably used in the cal analogies encountered in this study; but it
past: near-shore shallow waters into which stu- is likely that most cannot be resolved without a
dents could safely wade. The Cumberland Sound time machine. Having acknowledged this, how-
data are less than ideal in this regard. ever, it is clear from the data in this volume and
Ideally, modern fishes would be collected us- in other volumes in the American Museum of
ing gear more similar to that surmised for the past Natural History series on St. Catherines Island
in locations likely to have been routinely used in that we do know a great deal about life on this
the past. Early societies would have used multi- island and in the Georgia Bight.
ple methods of capture to maximize their efforts,
probably in shallow waters. Experiments that RESULTS
test the seasonal quantity and variety of fishes
that might be captured using fishing gear and Modern South Beach
locations hypothesized to have been used in the andCumberland Sound Data
past would greatly improve our understanding of A total of 134,895 individual fish representing
many aspects of this study. Arguments that infer 102 species were collected from North and South
archaeological fishing strategies in the archaeo- beaches between 1998 and 2007 (N = 208; table
logical past using modern fishing strategies are, 3.3). A total of 79 species were collected from
TABLE 3.3
Total Catch by Year for North Beach and South Beach (N = 208)
Year North (N = 107) South (N= 101) Total
1998 4,126 12,587 16,713
1999 9,088 9,423 18,511
2000 6,093 16,571 22,664
2001 5,392 10,783 16,175
2002 2,506 2,912 5,418
2003 2,484 1,865 4,349
2004 3,367 9,265 12,632
2005 5,954 9,696 15,650
2006 6,539 6,554 13,093
2007 6,609 3,081 9,690
Total 52,158 82,737 134,895
TABLE 3.4
Mean Catch per Unit Effort (CPUE) by Year for North Beach and South Beach
Year North (N = 1196) South (N = 1554) Average
1998 32.5 71.5 55.2
1999 64.5 54.5 59.0
2000 41.2 104.2 73.8
2001 37.4 56.5 48.3
2002 27.2 27.0 27.1
2003 32.3 16.1 22.5
2004 25.9 48.3 39.2
2005 46.9 55.1 51.7
2006 49.9 42.0 45.6
2007 83.7 28.8 52.1
Average 44.1 50.4 47.4
TABLE 3.5
Fishes with Ubiquity ≥ 0.25, North Beach and South Beach Data Combined
Family Common name Scientific name Ubiquity (N = 101)
Atherinopsidae Inland silverside Menidia beryllina 0.98
Engraulidae Bay anchovy Anchoa mitchelli 0.93
Sciaenidae Southern kingfish Menticirrhus americanus 0.89
Mugilidae Striped mullet Mugil cephalus 0.85
Fundulidae Striped killifish Fundulus majalis 0.71
Carangidae Florida pompano Trachinotus carolinus 0.67
Sciaenidae Silver perch Bairdiella chrysoura 0.51
Sciaenidae Star drum Stellifer lanceolatus 0.49
Engraulidae Striped anchovy Anchoa hepsetus 0.44
Ariidae Hardhead catfish Ariopsis felis 0.36
Sciaenidae Spot Leiostomus xanthurus 0.36
Carangidae Atlantic bumper Chloroscombrus chrysurus 0.32
Sciaenidae Red drum Sciaenops ocellatus 0.30
Clupeidae Atlantic menhaden Brevoortia tyrannus 0.29
Sciaenidae Gulf kingfish Menticirrhus littoralis 0.28
Belonidae Atlantic needlefish Strongylura marina 0.26
Pomatomidae Bluefish Pomatomus saltatrix 0.25
2.5
SHANNON-WIENER DIVERSITY (H')
2.0
1.5
1.0
0.5
0.0
Archaeological South Beach South Beach South Beach South Beach
Collections
Fig. 3.5. Comparison of archaeological and South Beach fish diversity.
TABLE 3.6
Fishes Collected from South Beach with Ubiquity ≥ 0.25
Family Common name Scientific name Ubiquity (N = 101)
Atherinopsidae Inland silverside Menidia beryllina 0.96
Engraulidae Bay anchovy Anchoa mitchelli 0.91
Sciaenidae Southern kingfish Menticirrhus americanus 0.89
Mugilidae Striped mullet Mugil cephalus 0.81
Fundulidae Striped killifish Fundulus majalis 0.72
Carangidae Florida pompano Trachinotus carolinus 0.67
Sciaenidae Red drum Scianops ocellatus 0.44
Engraulidae Striped anchovy Anchoa hepsetus 0.43
Sciaenidae Silver perch Bairdiella chrysoura 0.41
Carangidae Atlantic bumper Chloroscombrus chrysurus 0.33
Ariidae Hardhead catfish Ariopsis felis 0.31
Sciaenidae Spot Leiostomus xanthurus 0.30
Sciaenidae Star drum Stellifer lanceolatus 0.30
Sciaenidae Gulf kingfish Menticirrhus littoralis 0.29
Pomatomidae Bluefish Pomatomus saltatrix 0.25
North Beach and 66 species were collected from during summer and fall though they concentrate
South Beach. In terms of catch composition, 10 in shallow waters near shore in the summer and
species are present in over 95% of the total catch in midchannel locations in the fall. In winter,
and 75% of the total number of fishes caught fish are more abundant in the warmest, deepest
consists of four species: bay anchovies, inland parts of the estuary. Many of the species pres-
silversides, striped mullets, and Florida pom- ent in winter are year-round residents, but other
panos (Trachinotus carolinus). Mean catch per taxa leave the estuary altogether. In spring, fish
unit effort (CPUE), a measure of the number of abundance is greatest near freshwater creeks. The
fishes caught per seine haul, was slightly higher number of fishes present in the estuary, however,
for South Beach than for North Beach (table 3.4). is lowest during the colder months.
With respect to gear, 121,425 fish were caught This does not convey the complexity of fish
with the bag seine and 13,470 were caught with behavior because habitat preferences change as
the monofilament seine. The average total length individuals mature. Members of each age co-
(TL) of fish caught with the bag and monofila- hort have distinct habitat and niche preferences,
ment seines was 68 mm and 132 mm, respective- which change as animals mature during the an-
ly. No fish had a ubiquity of 1.0 (i.e., was present nual cycle. Most juveniles are more tolerant of
in every sample) in the combined North Beach estuarine conditions than are most adults, which
and South Beach data (table 3.5). Inland silver- prefer more stable offshore waters. Larger, more
sides and bay anchovies have the highest ubiq- mature fishes are present in fall collections, and
uity in the combined data, followed closely by dominate winter ones, because juveniles have left
southern kingfishes (Menticirrhus americanus) to mature elsewhere. Spring collections contain
and striped mullets. The number of fish individu- young of the year, in addition to larger individu-
als is lowest in the colder months and highest in als. The annual influx of juveniles reduces the
warm months (fig. 3.2). A large pulse is clearly average weight per fish in each sample. Juveniles
visible from April to September, followed by a grow larger as they mature, thereby increasing the
decline in October. The average size of fishes is average weight per individual. Thus, the average
another aspect of population structure related to weight per fish is highest in winter samples, de-
seasonality (table 3.2). clines in spring and summer samples, and is low-
Inland silversides and bay anchovies also est in fall samples (DEIS, 1978: D-480–481).
have the highest ubiquity in the South Beach col- The spawning cycle and use of Cumberland
lection, followed closely by southern kingfishes Sound by juvenile fishes are reflected in changes
and striped mullets (table 3.6). The ubiquity of in the average weight of fishes (biomass) dur-
fishes collected per quarter from South Beach ing the year at specific locations (fig. 3.8; DEIS,
has a distinctly seasonal pattern (table 3.7). Av- 1978: D-458, D-474–475, D-477–478). Changes
erage ubiquity is highest in summer, followed in biomass have a general association with the
by fall, spring, and winter. Species diversity on seasonal movement of adult and juvenile fishes
South Beach is highest in the spring collection in and out of the estuary. During summer, fish
and lowest in the fall collection (fig. 3.5). Sum- biomass is distributed throughout the estuary, but
mer and winter are the only seasons that show concentrated at the mouth of the estuary and near
significantly different diversity in the modern the confluence of two freshwater rivers. In fall the
data (table 3.8). greatest concentration of biomass is in and near
Cumberland Sound data represent fish season- Kings Bay, which is also where penaeid shrimp
ality in a back-barrier location. The stations near are abundant. The concentration of fish biomass
Kings Bay are the most productive areas in terms around Kings Bay also reflects the influx of juve-
of fish abundance and biomass (fig. 3.6; DEIS, nile fishes into this shallow part of the estuary.
1978: D-436, D-448–451, D-456). This may Community structure in the summer and fall
reflect the abundance of food and other nursery is essentially the same, but somewhat differ-
characteristics in these shallow waters. ent from that in spring and very different from
Cumberland data show seasonal pulses in the that in winter (DEIS, 1978: D-492; see also,
number of fish individuals as fishes in different Dahlberg and Odum, 1970). Larger fishes, more
parts of their life cycles migrate in and out of the mature ones, are present in fall (DEIS, 1978:
sound (fig. 3.7; DEIS, 1978: D-437, D-448–451). D-523). Larger fishes are much more common in
Large numbers of fish are uniformly distributed winter, when small juveniles have left the area
TABLE 3.7
Seasonal Ubiquity of Fishes Collected from South Beach
Common name Scientific name Winter Spring Summer Fall Average
Inland silverside Menidia beryllina 1.00 0.96 0.96 0.92 0.96
Bay anchovy Anchoa mitchelli 0.84 0.92 0.96 0.92 0.91
Southern kingfish Menticirrhus americanus 0.80 0.85 0.96 0.96 0.89
Striped mullet Mugil cephalus 0.88 0.58 0.88 0.92 0.81
Striped killifish Fundulus majalis 0.52 0.69 1.00 0.67 0.72
Florida pompano Trachinotus carolinus 0.20 0.50 1.00 1.00 0.68
Red drum Sciaenops ocellatus 0.52 0.42 0.46 0.33 0.43
Striped anchovy Anchoa hepsetus 0.16 0.38 0.65 0.50 0.42
Silver perch Bairdiella chrysoura 0.64 0.46 0.19 0.33 0.41
Atlantic bumper Chloroscombrus chrysurus 0.00 0.08 0.73 0.50 0.33
Hardhead catfish Ariopsis felis 0.04 0.31 0.62 0.25 0.30
Spot Leiostomus xanthurus 0.24 0.19 0.42 0.33 0.30
Star drum Stellifer lanceolatus 0.36 0.31 0.35 0.17 0.30
Gulf kingfish Menticirrhus littoralis 0.28 0.27 0.31 0.29 0.29
Bluefish Pomatomus saltatrix 0.08 0.42 0.35 0.13 0.24
Average ubiquity 0.44 0.49 0.66 0.55
TABLE 3.8
Comparison of Archaeological and Modern South Beach Fish Diversity
Comparison Diff. of means t P value Critical level Significant?
Archaeological vs. 1.030 4.54 <0.001 0.005 Yes
South Beach winter
Archaeological vs. 0.824 3.52 0.002 0.006 Yes
South Beach spring
South Beach summer 0.912 3.05 0.005 0.006 Yes
vs. South Beach winter
South Beach summer 0.697 2.33 0.028 0.007 No
vs. South Beach spring
Archaeological vs. 0.377 1.57 0.133 0.009 No
South Beach fall
South Beach fall vs. 0.593 1.98 0.059 0.01 No
South Beach winter
South Beach fall vs. 0.377 1.26 0.219 0.013 No
South Beach spring
South Beach summer 0.320 1.07 0.296 0.017 No
vs. South Beach fall
South Beach spring vs. 0.215 0.72 0.478 0.025 No
South Beach winter
Archaeological vs. 0.127 0.52 0.612 0.05 No
South Beach summer
Abundance Biomass, g
80,000
60,000
40,000
20,000
0
Upper reaches Kings Bay Lower reaches Ocean
Fig. 3.6. Abundance and biomass values for Cumberland Sound (DEIS, 1978: D-456). Upper reaches include
data from stations A, B, and C; Kings Bay includes data from stations X, D, E, and F; lower reaches include data
from stations G, K, N, and O; ocean includes data from stations P, Q, R, and S; shown on fig. 3.4. Data from
DEIS (1978: D-456).
35,000
NUMBER OF FISH INDIVIDUALS
30,000
25,000
20,000
15,000
10,000
5000
0
Fig. 3.7. Overall abundance for Cumberland Sound (DEIS, 1978: D-437).
100.0
90.0
PERCENT BIOMASS
80.0
70.0
60.0
50.0
40.0
30.0
20.0
10.0
0.0
Fig. 3.8. Seasonal percentages of fish biomass at Stations D and E, Cumberland Sound (DEIS, 1978: D-458).
to mature elsewhere. In spring, larger fishes are winter (10.9 g/fish), declines in spring (4.7 g/fish)
joined by small individuals, probably young of and summer (3.5 g/fish), and is lowest in fall (2.8
the year. The annual influx of juveniles reduces g/fish). Many of the species present in winter are
the average weight per fish; as these mature and year-round residents (DEIS, 1978: D-483).
grow larger the average weight increases. Thus, Diversity has a seasonal pattern as fishes re-
the average weight of individual fish is highest in spond to spawning cycle, salinity, water depth,
4.0 Station D Station E
3.0
DIVERSITY
2.0
1.0
0.0
Fig. 3.9. Seasonal variations in fish diversity at Stations D and E, Cumberland Sound (DEIS, 1978: D-489).
A
South Beach Star Drum (Stellifer lanceolatus)
NUMBER OF INDIVIDUALS
400
300
200
100
0
SEASON
B
Cumberland Sound Star Drum (Stellifer lanceolatus)
NUMBER OF INDIVIDUALS
16,000
12,000
8000
4000
0
SEASON
Fig. 3.10. Abundance of star drum (Stellifer lanceolatus) individuals per season: (A) South Beach and (B)
Cumberland Sound collections. Cumberland Sound data from DEIS (1978: D-439, D-441, D-443, D-445).
and water temperature (fig. 3.9; DEIS, 1978: months. Seasonal trends are most pronounced in
D-484–489). Although the nursery aspects of the the sounds and least in the upper marsh creeks
estuary are complex, there is a relationship be- (Dahlberg and Odum, 1970). Dahlberg and Odum
tween diversity and the spawning cycle. Young (1970) do not report significant differences in di-
fishes enter estuaries for the nursery aspects, versity of fishes in sounds and tidal creeks be-
adults enter to feed and spawn, and both adult hind St. Catherines Island, however. They did not
and newly mature fishes leave for inshore and specifically sample habitats for the smaller fish
offshore waters. During summer and fall, the species (e.g., silversides, mummichogs, and mul-
highest diversities are found in areas with the lets), or sample the seaward beach on the island.
lowest salinity values. This does not hold true Well-versed anglers understand these changes
for winter, when water temperatures decline and in fish behavior, as well as the accompanying
fishes move from shallower waters into deeper, variations in potential capture rates. Today, fish-
warmer areas. In spring, both low and high salin- ing strategies take advantage of local knowledge
ity areas have high species diversity, though es- of seasonal patterns, derived from decades of
tuarine waters may be too cool for some species, experience, oral traditions passed down through
which remain offshore in warmer waters. generations, and experts such as boat captains,
Modern data from South Beach and Cumber- guides, and recreational fishermen (Murray, Neis,
land Sound have different patterns of seasonal and Johnsen, 2006). These insights are used for
abundance (figs. 3.2, 3.7). Although these dif- subsistence, economic gain, and recreation (Mat-
ferences could be the result of different modern thews, 1928; McClanahan and Cinner, 2008).
catch technologies used in the South Beach and Such local knowledge undoubtedly informed
Cumberland Sound studies, they also can be earlier fishing strategies in the Georgia Bight and
explained by differences in habitat sampled. It can be used to develop current and historical con-
is likely that the more exposed, higher-salinity servation management plans (Murray, Neis, and
South Beach location represents a portion of life Johnsen, 2006; Silvano et al., 2006; McClanahan
cycles of high-ubiquity fishes that is less com- and Cinner, 2008).
mon in Cumberland Sound. This difference may
reflect the seasonal migration of fishes into and Archaeological Data
out of the back-barrier reaches lying between sea A total of 6912 fish individuals representing
islands and the mainland as part of their growth 72 taxa are present in the archaeological collec-
and reproduction cycles. Star drums (Stellifer lan- tions (table 3.9). The most ubiquitous fishes (≥
ceolatus) are most abundant in the South Beach 0.83) are, in order of ubiquity: seatrouts (Cyno-
catch in the summer (fig. 3.10A), but they are scion spp.), mullets, gars (Lepisosteus spp.), At-
most abundant in the Cumberland Sound catch lantic croakers (Micropogonias undulatus), hard-
in the fall (fig. 3.10B; DEIS, 1978: D-438–447, head catfishes (Ariopsis felis), gafftopsail cat-
D-459–473), for example. We use star drums fishes (Bagre marinus), flounders (Paralichthys
in this example because they constitute 27% of spp.), red drums, and star drums. The sea catfish
fishes individuals captured in Cumberland Sound family is as ubiquitous as seatrouts (1.0), but the
and 15% of the biomass in Cumberland Sound. archaeological reports of St. Simons Cannon’s
They are the most abundant species in Dahlberg Point and West rings and Bourbon Field record
and Odum’s (1970) study of Sapelo and St. Cath- these taxa at the family level, reducing the ubiq-
erines sounds and have an archaeological ubiquity of the two species.
uity of 0.84. This is always a small-bodied fish
(TL range 7.5–102.5 mm) despite indeterminate Comparison
growth (table 3.2; Dahlberg and Odum, 1970). Most high-ubiquity fishes in the modern South
Dahlberg and Odum (1970) elaborate upon Beach collection also are present in Georgia Bight
the seasonal aspect of fishes in the back-barrier archaeological collections (fig. 3.11). The taxa in
areas. They report that 29 of the 70 fish species figure 3.11 are arranged in order of their ubiquity
“were collected at estuarine stations in all seasons in the South Beach collection. The taxa labeled
or at least in both winter and summer” (Dahl- 1, 2, and 8 are small-bodied fishes (Menidia be-
berg and Odum, 1970: 384). Only eight of the ryllina, Anchoa mitchelli, Anchoa hepsetus) that
20 species were restricted to the colder months could be discriminated against by site formation
and seven species were collected only in warmer processes or during archaeological excavations.
72
TABLE 3.9
Ubiquity of Fishes Identified in Some Georgia Bight Archaeological Collectionsa
Sapelo St Cat Mc- Marsh West Ribault FOY A Cathead Kings Ke- Bour- Cat- Kings JEA FOY MHF Back EPC Pueb-
Ring Queen Ring Ring Ck Bay nan bon head Bay SJ Ck lo
III Ck
Ar- Ar- Ar- Ar- Ar- Ar- Ar- Sw Sw Sav Sav Sav Sav Sav SJ Irene Irene 17th 17th
chaic chaic chaic chaic chaic chaic chaic Creek Creek c c
Richness 71 21 31 36 25 16 26 11 19 27 15 19 16 26 16 27 7 26 16 14
Taxon Ubiq- Total
uity
Cynoscion 1.00 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 19
spp.
Mugil spp. 1.00 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 19
Lepisosteus 0.95 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 18
spp.
Micropo- 0.89 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 17
gonias
undulatus
Ariopsis felis 0.84 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 16
Bagre mari- 0.84 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 16
nus
Paralichthys 0.84 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 16
spp.
Sciaenops 0.84 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 16
ocellatus
Stellifer 0.84 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 16
lanceolatus
Archosargus 0.79 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 15
probato-
cephalus
Clupeidae 0.79 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 15
Pogonias 0.79 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 15
cromis
Bairdiella 0.74 1 1 1 1 1 1 1 1 1 1 1 1 1 1 14
chrysoura
ANTHROPOLOGICAL PAPERS AMERICAN MUSEUM OF NATURAL HISTORY
Leiostomus 0.68 1 1 1 1 1 1 1 1 1 1 1 1 1 13
xanthurus
Ariidae 0.47 1 1 1 1 1 1 1 1 1 9
Elops saurus 0.47 1 1 1 1 1 1 1 1 1 9
NO. 97
TABLE 3.9 — (Continued)
2012
III Ck
Richness 71 21 31 36 25 16 26 11 19 27 15 19 16 26 16 27 7 26 16 14
Taxon Ubiq- Total
uity
Menticirrhus 0.47 1 1 1 1 1 1 1 1 1 9
spp.
Pomatomus 0.42 1 1 1 1 1 1 1 1 8
saltatrix
Rajiformes 0.42 1 1 1 1 1 1 1 1 8
Carangidae 0.32 1 1 1 1 1 1 6
Carcharhini- 0.32 1 1 1 1 1 1 6
dae
Ictalurus spp. 0.32 1 1 1 1 1 1 6
Opsanus sp. 0.32 1 1 1 1 1 1 6
Amia calva 0.26 1 1 1 1 1 5
Anguilla 0.26 1 1 1 1 1 5
rostrata
Cyprinodon- 0.26 1 1 1 1 1 5
tidae
Dasyatidae 0.26 1 1 1 1 1 5
Fundulus sp. 0.26 1 1 1 1 1 5
Lagodon 0.26 1 1 1 1 1 5
rhomboides
Peprilus spp. 0.26 1 1 1 1 1 5
Ameiurus sp. 0.21 1 1 1 1 4
Dasyatus 0.21 1 1 1 1 4
spp.
Myliobatidae 0.21 1 1 1 1 4
Belonidae 0.16 1 1 1 3
Brevoortia 0.16 1 1 1 3
spp.
Chloroscom- 0.16 1 1 1 3
brus chrysurus
73
74
III Ck
Richness 71 21 31 36 25 16 26 11 19 27 15 19 16 26 16 27 7 26 16 14
Taxon Ubiq- Total
uity
Chondrich- 0.16 1 1 1 3
thyes
Citharichthys 0.16 1 1 1 3
spp.
Orthopristis 0.16 1 1 1 3
chrysoptera
Chaetodip- 0.11 1 1 2
terus faber
Diodontidae 0.11 1 1 2
Eleotridae 0.11 1 1 2
Galeocerdo 0.11 1 1 2
cuvier
Ictaluridae 0.11 1 1 2
Prionotus sp. 0.11 1 1 2
Acipenser 0.05 1 1
spp.
Carcharhi- 0.05 1 1
nus spp.
cf. Chlo- 0.05 1 1
roscombrus
chrysurus
cf. Lamnidae 0.05 1 1
Chaenobryt- 0.05 1 1
tus gulosus
Cyprinidae 0.05 1 1
Esox spp. 0.05 1 1

Larimus 0.05 1 1
fasciatus
NO. 97
2012
III Ck
Richness 71 21 31 36 25 16 26 11 19 27 15 19 16 26 16 27 7 26 16 14
Taxon Ubiq- Total
uity
Lepomis spp. 0.05 1 1
Lobotes suri- 0.05 1 1
namensis
Menticirrhus 0.05 1 1
americanus
Menticirrhus 0.05 1 1
saxatilis
Odontaspis 0.05 1 1
taurus
Ophidiidae 0.05 1 1
Peprilus 0.05 1 1
paru
Raja eglan- 0.05 1 1
teria
Rhinoptera 0.05 1 1
bonasus
Sardinella 0.05 1 1
sp.
Sciaenidae 0.05 1 1
Serranidae 0.05 1 1
Sparidae 0.05 1 1
Sphyrna sp. 0.05 1 1
Symphurus 0.05 1 1
spp.
Trichiurus 0.05 1 1
lepturus
Trinectes 0.05 1 1
spp.
Fish MNI 176 1121 1112 318 235 126 23 64 1672 231 398 61 853 157 204 32 334 75 54 7246
a
75
See table 3.1 for references.

However, taxon 5 is also a small-bodied fish (kil- and 8, which are absent or rare in archaeological
lifish, Fundulus majalis) and taxon 3 (southern collections. Thus, recovery technique is not the
kingfish) is generally a large fish when found in only explanation for this pattern.
Georgia Bight archaeological collections. Mul- Instead, it appears that many of the fishes
lets, taxon 4, are often small-bodied fishes in ar- most ubiquitous in the South Beach collection
chaeological collections (e.g., Colaninno 2010: are animals less commonly used in the past; sup-
51); though they grow to much larger sizes as porting our premise that fishing was highly se-
adults, the larger size is more typical of Hispanic lective; fishes with lower modern ubiquity were
collections from sites further south, such as those favored over fishes with higher modern ubiquity
from St. Augustine, than they are of pre-Hispanic by earlier fishing strategies. This conclusion is
collections. Mullets in many coastal archaeologi- underscored by taxon 23. Taxon 23 is gar, a fish
cal collections are about the size of taxa 1, 2, 5, found in 95% of the archaeological collections
South Beach Archaeological
1.0
0.8
UBIQUITY
0.6
0.4
0.2
0.0
1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23
TAXON
Fig. 3.11. Comparison of South Beach and archaeological fish ubiquity. 1, Menidia beryllina; 2, Anchoa
mitchelli; 3, Menticirrhus americanus; 4, Mugil cephalus; 5, Fundulus majalis; 6, Trachinotus carolinus; 7, Sci-
aenops ocellatus; 8, Anchoa hepsetus; 9, Bairdiella chrysoura; 10, Clupeidae; 11, Chloroscombrus chrysurus;
12, Ariopsis felis; 13, Leiostomus xanthurus; 14, Stellifer lanceolatus; 15, Menticirrhus littoralis; 16, Pomato-
mus saltatrix; 17, Cynoscion spp.; 18, Pogonias cromis; 19, Bagre marinus; 20, Micropogonias undulatus; 21,
Paralichthys spp.; 22, Archosargus probatocephalus; 23, Lepisosteus spp.
1.0
0.8
UBIQUITY
0.6
0.4
0.2
0.0
Striped Spot Star drum Red drum Hardhead Silver Seatrout Croaker
mullet catfish perch
Fig. 3.12. Seasonal ubiquity of selected fish taxa in the South Beach collection.
from every time period and every location within are not available for Cumberland Sound but most
the Georgia Bight. Gars are absent in the South taxa with ubiquity ≥ 0.68 in the archaeological
Beach catch and rare in the Cumberland catch. record are present during at least three, if not
Perhaps the conclusion to be drawn from figure four, of the seasons at Station E in the Kings Bay
3.11 is that people did not often fish on the sea- area (fig. 3.13; DEIS, 1978: D438–D445).
ward side of barrier islands, where gars are rare Species diversity of archaeological data (H' =
or absent, but they did fish often in back-barrier 2.06) is significantly higher than modern South
locations, where longnose gars (Lepisosteus os- Beach data (H' = 1.48). Diversity of the archaeo-
seus) are found throughout the year (Dahlberg, logical collections is significantly different than
1972). It also is possible that this difference is the diversity of modern winter and spring South
a reflection of the lower freshwater input at St. Beach collections, but not when compared with
Catherines, or may be an indication that more the fall and summer samples (table 3.8; fig. 3.5).
freshwater habitats once existed between St.
Catherines Island and the mainland and on the DISCUSSION
island itself, particularly if the island once ex-
tended further east than it does today. The purpose of this chapter is to consider evi-
Figure 3.12 shows the present-day seasonal dence for season of availability in modern fishes
ubiquity of eight fishes with high ubiquity in as a means of clarifying season of activity and
either the modern South Beach (≥ 0.30) or ar- location for people living in the Georgia Bight
chaeological (≥ 0.68) collections. The first obser- in the past. Although this study focuses on those
vation is that all but one of these fishes is pres- fishes that form the core of the Georgia Bight
ent throughout the year; the primary pattern is a fishery, many other organisms, including fungi,
seasonal shift in relative ubiquity. The standard plants, invertebrates, reptiles, mammals, and
seasonal markers that may be used elsewhere fail birds, supplemented these fishes. This flexible
in this context. The second observation is that subsistence strategy is reflected in a moderate
some fishes are more ubiquitous than are others vertebrate diversity (H' = 2.890, MNI) and high
during one or more season. Mullet, red drums, vertebrate equitability (V' = 0.841, MNI) before
and star drums are unlikely to be helpful as sea- the advent of the First Spanish period in a.d. 1565
sonal markers, whereas hardhead catfishes might (Reitz et al., 2010: 63, 72, 233–234).
be strong markers for warmer waters and silver In this study, modern data show clear season-
perches (Bairdiella chrysoura) moderately strong al patterns in the availability of fishes. Charac-
markers for cooler waters. Similar ubiquity data teristics such as abundance, biomass, diversity,
dominance, distribution, ubiquity, and body size,
however, are relative rather than absolute and
may depend on location and gear as much as on
season. Organisms characteristic of estuaries are
those best able to withstand the biogeochemi-
cal and hydrological conditions that define es-
ARCHAEOLOGICAL UBIQUITY
1.0 tuaries. Resource seasonality does occur in the

Georgia Bight, but many fishes identified in ar-
0.8
chaeological collections are present somewhere
0.6 in estuarine waters throughout the year, depend-
ing on whether they are young fishes using estu-
0.4 aries as nurseries or adults using them as feeding
grounds.
0.2 Fishing in the Georgia Bight was highly se-
lective, focused specifically on a core group
0.0 of indicator taxa that occupy estuarine waters
0 1 2 3 4 throughout the year. Many of these core fishes
NUMBER OF SEASONS either school or form large aggregations, mak-
Fig. 3.13. Number of seasons in which fishes with ing them susceptible to mass-capture techniques
archaeological ubiquity ≥ 0.68 are present at Station (Reitz, Quitmyer, and Marrinan, 2009; Reitz et
E, Cumberland Sound (DEIS, 1978: D-438–D-445). al., 2010: 234–237). These same high-ubiquity,
core fishes generally are not high-ubiquity taxa for season of availability.
in standard fishing samples taken under modern Modern studies suggest that shallow, warm,
conditions, however. Although this may be strong low-energy waters with salinity and nutrient lev-
evidence that major environmental changes in els typical of middle and upper reaches were pre-
estuarine ecosystem functions have occurred, it ferred fishing grounds. Beaches were used less
is also true that the ability of ecological analo- frequently than were back-barrier areas. Although
gies using fish presence and ubiquity to provide modern data clearly indicate when and where
answers to questions of season of human activ- fishing would be productive today, it is unlikely
ity and location is limited. The answer to season that these precise locations were equally produc-
of availability, however, is that taxa in the core tive in the past because the topography of each
fishing group were available somewhere in the estuary probably has changed. Identifying spe-
upper or lower reaches of the estuary throughout cific places where fishing occurred is unlikely.
the year. Where fishing took place and the gear used
Difficulties in correlating resource avail- undoubtedly changed during the annual cycle
ability with the synchronic biogeochemical and given the different locations the most ubiqui-
hydrological properties that define seasons are tous fishes occupy within estuaries during the
compounded by diachronic changes in the niche year. Fishing, likely using watercraft, probably
and habitat preferences of fish individuals as they followed the preferred, targeted fish taxa within
mature. Age cohorts are mobile and seek out pre- each estuary (e.g., Ames, 2002). The distances
ferred environmental properties throughout the are not great, probably no more than 10–12 km
estuary and beyond during an annual cycle. This depending on where the fishing party started and
produces patterns such as those observed in the where it was going. We suggest “party” intention-
mean total length of star drum on South Beach ally because many ubiquitous fishes are suscep-
(fig. 3.14). Mean total length of this small fish tible to technologies such as weirs and nets that
increases substantially in the fall, as mature indi- are most effective in shallow waters where tidal
viduals leave the estuary, even as their abundance ranges are large enough to overtop the enclosure
decreases (fig. 3.12). Young fishes are more tol- at high tide. Such facilities usually require several
erant of estuarine conditions than are most adult people to work and frequent repairs, but the labor
fishes. Combining presence data with measure- of people of all ages and abilities would be use-
ments of fishes may be more productive than ful. Although the distance from most places on
presence and ubiquity analysis in assessing sea- St. Catherines Island to daily or annually produc-
son of availability by considering the seasonal as- tive fishing grounds would not have been great,
pects of age cohorts, though one would still need fishing parties would have to be mindful of going
to know location of capture for larger and smaller against strong incoming or outgoing tides.
fishes and the influence of gear on this evidence Embedded in the archaeological literature is
the assumption that coastal life is difficult and
coastal residence must be forced upon people
by necessity (for reviews of these perceptions
see Erlandson, 2001 and Ames, 2002). Despite
Monks (1981) admonition that residential pat-
terns are not clearly related to season of availabil-
ity, it is typically presumed that people lived on
20
South Beach Star Drum (Stellifer lanceolatus) the coast only during the warmer months, resid-
MEAN TL (cm)
ing elsewhere during the remainder of the year.

15
The assumption that fishing communities must be
10 mobile fundamentally links season of availabil-
5 ity with season of location. Thus resource avail-
0 ability, subsistence schedules, residential habits,
Winter Spring Summer Fall political institutions, and social complexity are
SEASON merged in many models. Residential patterns
Fig. 3.14. Seasonal aspects of mean total length themselves are often reduced to a dichotomy of
(in cm) of star drum (Stellifer lanceolatus) in the mobility on the one hand and sedentism on the
South Beach collection. other, as though these were mutually exclusive
options. As Bar-Yosef and Rocek (1998: 1, ital- of a continuum with collecting at the other end.
ics theirs) argue: “The fact is all societies have a When used in this sense, foraging systems are
mobility component; the issue is what the form said to occur in areas with high productivity and
of that mobility is, not whether it exists.” As with resources that are seasonally and spatially homo-
most cultural institutions, the forms, causes, and geneous. Collector systems occur in areas with
consequences of resource acquisition and resi- low productivity and resources that are seasonally
dential patterns are diverse. or spatially uneven. Foragers acquire food each
Interpretations of coastal life labor under a day near their residential base and are residential-
particularly inflexible assumption that a farming ly mobile; all members move to another location
subsidy is required to support people through- when resources are judged to be inadequate. Such
out an annual cycle or to support complex social opportunistic foraging is only feasible if resourc-
institutions. Although it has been demonstrated es can be transported over a short distance. Col-
repeatedly that nonfarmers with complex social lectors, on the other hand, implement planned,
institutions did exist in coastal settings (e.g., logistic strategies that may exploit resources
Kennett and Kennett, 2000; Builth, 2006), the over a larger area, usually produce larger yields,
question remains unresolved in the minds of and are implemented by specialized task groups.
many. A quarter of all the fish taxa present in Such logistically mobile collectors bring resourc-
Georgia Bight archaeological collections is pres- es back to the residential base, where they may
ent in over 60% of the collections, regardless be stored. A foraging radius, therefore, is close
of whether the site is an Archaic shell ring or a to the base camp and a logistic/collecting radius
17th-century Spanish mission. This suggests that is further away. Collectors, too, may move their
the kinds of fish, though perhaps not the technol- residential bases seasonally, though storage may
ogy, have been little altered over the centuries, reduce the scope of this mobility. Coastal forag-
regardless of the cultural matrix of which fish- ers, therefore, should be characterized by low
ing was but one aspect (Reitz, 2004). This is not logistical mobility and high residential mobility
to suggest that there were no changes either in and coastal collectors by high logistical mobility
the resource base or in the cultural use of that and low residential mobility.
base; we know that there were. Throughout the Some research in coastal settings, however,
period, however, fishing was a major and consis- suggests that considerably more flexibility and
tent aspect of daily life; drawing upon the ability variety than a simple dichotomy between foragers
of both the fishes in the core group and people to and collectors characterize fishing in such areas
be flexible in an environment that changes with (Ames, 2002; Habu, 2002; Habu and Fitzhugh,
each tide. 2002; Butler and Campbell, 2004). Fishing in the
A suite of concepts draws upon human be- Georgia Bight appears to be another example not
havioral ecology, specifically optimal foraging easily accommodated in either of these models.
theory, to interpret fishing strategies (e.g., Habu, Generalizing broadly, the Georgia Bight could be
2002; Ugan, 2005; Winterhalder and Kennett, considered a high productive area with resources
2006; Culleton, Kennett, and Jones, 2009). In that are seasonally and spatially uneven.
archaeological applications, these concepts are The estuaries of the Georgia Bight are among
important because it is essential to distinguish the most productive and complex on the Atlantic
between anthropogenic and nonanthropogenic coast, with about four times as much marsh per
impacts on biological resources to assess causal- kilometer of coastline as other areas of the Atlan-
ity in environmental change and resilience in spe- tic seaboard (Schelske and Odum, 1961; Hayden
cific fisheries, important issues in conservation and Dolan, 1979: 1063; Frey and Howard, 1986).
biology and fisheries management (e.g., Butler Resources, however, are not seasonally or spa-
and Delacorte, 2004; Broughton, 2010). tially homogeneous; they are spatially and tem-
Among these concepts are those that distin- porally patchy in addition to being highly mobile
guish between subsistence-settlement systems in the case of vertebrates. Neither are they likely
of foragers and collectors (e.g., Binford, 1980; to be depleted within an annual cycle to such an
Bettinger, 1991: 68; Habu and Fitzhugh, 2002; extent that it would stimulate residential mobil-
Broughton, 2010). Although “forager” is often ity because the fishery is renewed with each tidal
used as a substitute for “hunter-gatherer,” in many cycle. It is entirely possible that the fishery could
optimal foraging models foraging forms one end be, and was, depressed by overfishing or other
drivers of long-term environmental change from have altered the fishing strategy in any measur-
time to time (e.g., Reitz, 2004; Quitmyer and Re- able way, though it may have impacted residen-
itz, 2006; Reitz, Quitmyer, and Marrinan, 2009), tial and fishing decisions as scheduling conflicts
but the broad features of the core fishing strategy were resolved (e.g., Flannery, 1968).
persisted, nonetheless. A similar conclusion is Sarah Campbell and Virginia Butler (2010)
reported for other fisheries (Butler and Camp- argue that beliefs and social institutions (such as
bell, 2004). ownership, regulations, rituals, and monitoring)
People likely did fish more or less each day might foster resilience in coastal ecosystems.
and return the food to a residential base but are They report a 7500-year-long record of stability
unlikely to have moved elsewhere if fishing dete- in a Pacific Northwest salmon fishery that may
riorated at a favored fishing ground; the distances be evidence for the regulation of salmon fish-
on any of the islands or within estuaries would be ing by cultural institutions in the region, thereby
well within the 30 km estimated for a coastal for- encouraging persistence in the fishery. A similar
aging radius (Ames, 2002). More likely the travel long-term fishery is seen in the Georgia Bight. It
distance for fishing was much less throughout the is probable that kin groups or other social units
year. It is entirely likely that fishing would be less owned the rights to specific fishing grounds and
productive at a specific location during the year the facilities constructed at each. We consider it
given the preferences of fish age cohorts, popu- unlikely that fishing networks would fail to safe-
lations, and communities for different locations guard their rights to such valuable assets, even
during the annual growth and reproductive cycles during brief absences for social or economic
of each species, but it is unlikely that all fishes in reasons (e.g., Nishimura, 1975; Builth, 2006).
the core strategy would move so far away that The possibility that members of these fishing
they could not be reached from any residential communities intentionally modified estuarine
base within most estuaries in the Georgia Bight. conditions to encourage a variety of estuarine
Likely only task groups were involved in each plants and animals, much as they might ma-
fishing foray. The potential exists for large quan- nipulate habitats preferred by terrestrial organ-
tities of fish to be taken, though storing fish may isms, is intriguing (e.g., Erickson, 2000; Builth,
not have been appealing in this subtropical, hu- 2006; Campbell and Butler, 2010). As Butler
mid region. and Campbell (2004: 373) report for Pacific
We suspect good places to build homes, se- coast salmon fishing, the core fishes in Georgia
cure water craft, shelter from wind and sand Bight archaeological collections are among the
fleas, and dry nets might be limited and highly most widespread and abundant fish prey. Their
prized. These would be abandoned with great role in the fishery persisted from ca. 2560 b.c.
reluctance in response to such larger factors as until the beginning of the First Spanish period,
a reconfiguration of the coastline after a storm, in fact, continuing to be prominent components
but not because of factors that are said to charac- of subsistence strategies in EuroAfrican colonial
terize either opportunistic foragers or logistical towns and plantations thereafter (Reitz, 1994;
collectors. Reitz et al., 2010). This fishery undoubtedly
Another key concept is the role of plant do- experienced cycles of crisis and recovery, but it
mestication. Plants and animals provide different was sufficiently flexible and resilient to endure
types of nutrients. It is unlikely that a sustainable despite cultural and environmental changes over
coastal economy was based exclusively on either the millennia.
plant or animal nutrients. Plants and animals also Other factors complicate studies of season of
provide different types of raw materials and play availability in the Georgia Bight. Many animals
different roles in social life. Thus, people living in used in the past are available throughout the year
the Georgia Bight, as broad-spectrum omnivores, at several different locations. Although they may
undoubtedly used both plants and animals, some move, reproduce, or grow seasonally, these de-
of which were terrestrial and some aquatic. When tails are difficult to determine from zooarchaeo-
or whether some plants were at least tended if not logical observations such as taxonomic identifi-
domesticated is poorly known for the Georgia cations, taxonomic frequencies, element distri-
Bight, but eventually some domestic plants did bution, measurements, age at sexual maturity,
enter the economy (see Scarry and Hollenbach, and season of death. Isotopic ratios and growth
chap. 11, this volume). This does not appear to increments may demonstrate that an organism
died during one part of the annual temperature bine observations for communities of organisms
or growth cycle; but not the contemporaneity of with information about the biogeochemical and
other sites that might be part of a seasonal round. hydrological properties preferred by ubiquitous
Technologies and other aspects of subsistence taxa. Indicator groups are useful for studies such
strategies can leave vertebrate evidence that di- as the one in this chapter, however, syntheses of
verges widely from background resource bases, biotic, abiotic, and cultural data to form indicator
as seen in this study. Storage, production, and packages are needed before we can address the
exchange systems could leave little unambigu- questions of where and how. Indicator packages
ous evidence at the site under study. Exchange are difficult to develop and seldom achieved. The
networks within an extended family or on a first step toward this goal is to ensure that a wide
broader scale have the purpose of rounding the range of evidence is included in the research
shortfalls associated with poor weather, poor project through thoughtful fieldwork and mul-
skills, and bad luck. tiproxy studies. The full suite of environmental
Estuaries are characterized by continuity and studies should be performed with equal rigor at
stasis rather than change in large part because of all sites that might be part of a seasonal round,
the inherent flexibility of organisms that use es- exchange network, kin group, or other social unit
tuaries as nursery grounds. Such taxa could with- to systematically explore how sites of different
stand significant exploitation, and alterations as- sizes and functions fit into the overall landscape,
sociated with top-down and bottom-up processes and how patterns of season, activity, and location
regardless of the cause. That many estuarine change through time and space.
organisms are no longer able to maintain former
population sizes and community structures, and
now are in decline, should be cause for alarm
NOTES
among coastal resource managers.
1. We gratefully acknowledge the long-term research
CONCLUSION support of the Edward John Noble Foundation and the
St. Catherines Island Foundation. We thank David Hurst
Estuarine fishes are attracted to back-barrier Thomas, American Museum of Natural History, for his sup-
port of zooarchaeology on St. Catherines Island, the oppor-
reaches as nursery grounds and feeding areas. tunity to study the St. Catherines Island animal remains, and
The potential of estuaries to fill these roles var- for facilitating the publication of this issue of the American
ies seasonally. Our analysis indicates that the fish Museum of Natural History Anthropological Papers. Fund-
communities and populations forming the core ing was provided in part by NSF Award No. BCS-1026167.
fishing strategy into the 17th century are avail- Over 100 Augusta State University undergraduate students
have helped with the St. Catherines Island sampling. We are
able in shallow waters throughout the year, albeit grateful to them all even if we cannot list them by name. We
in different locations and probably vulnerable to are also grateful to the diverse group of archaeologists who
different capture techniques. Seasonality in the provided the opportunities to study the vertebrate collections
Georgia Bight appears to be more a question of listed in table 3.1, particularly to Carol Colaninno-Meeks
and Sarah Bergh for permission to use unpublished disserta-
where and how rather than what or when. tion data. An earlier version of this paper was presented
In order to engender hope for determining at the 75th annual meeting of the Society for American
season of activity and location, studies must com- Archaeology, St. Louis.
2012 EVALUATING δ18O PROFILES OF HARDHEAD CATFISH 83
CHAPTER 4
EVALUATING δ O PROFILES OF HARDHEAD
18
CATFISH AND ATLANTIC CROAKER OTOLITHS

AS A METHOD OF DETERMINING
SEASONAL USE OF FISHES
C arol E. C olaninno 1
Documenting seasonal resource use among 2010: 54–57; Reitz et al., chap. 3, this volume).
human populations is a central aspect of archae- These resources include a diverse array of estua-
ological research. For coastal locations of the rine fishes along with numerous other aquatic and
southeastern United States, studies of seasonal terrestrial animals. Investigating seasonal use of
resource use have focused on molluscs, and for nonmolluscan taxa, particularly fishes, provides
good reason. Molluscs, including oysters (Cras- further information about past land and estuarine
sostrea virginica) and hard clams (Mercenaria use patterns, resource intensification, and social
spp.), generally are abundant at these locations; complexity. It is an area of research that largely is
there is a correlation between incremental for- unexplored in the southeastern United States.
mation and yearly, seasonal environmental os- This chapter examines seasonal use of two
cillations; and shell CaCO3 (calcium carbonate) fish species: hardhead catfish (Ariopsis felis)
precipitates in oxygen isotope equilibrium with and Atlantic croaker (Micropogonias undu-
the ambient water in which shell forms (Gross- latus). I examine δ18Ootolith profiles of modern
man and Ku, 1986; Quitmyer, Jones, and Ar- hardhead catfishes and Atlantic croakers to de-
nold, 1997; Andrus and Crowe, 2000; Andrus termine whether δ18Ootolith profiles track seasonal
and Crowe, 2008). This allows researchers to temperature oscillations. Seasonal temperature
examine patterns in incremental formations and oscillations observed in modern δ18Ootolith are
the ratio of stable oxygen isotopes (18O/16O, ex- applied to archaeological δ18Ootolith to determine
pressed as δ18O) at the terminal band and in prior past seasonal temperature oscillations and the
bands to determine the season in which people season in which people captured these fishes.
captured the organism (Quitmyer, Hale, and The archaeological otoliths are from four Late
Jones, 1985; Quitmyer, Jones, and Arnold, 1997; Archaic shell rings (2550–1500 cal b.c.) lo-
Andrus and Crowe, 2000, 2008; O’Brien and cated on the Georgia coast: the Cannon’s Point
Thomas, 2008; Thompson and Andrus, 2011). (9GN57) and West rings (9GN76) of St. Simons
These studies were instrumental in document- Island, and the St. Catherines (9Li231) and Mc-
ing year-round occupations, as well as locations Queen shell rings (9LIi648) of St. Catherines
used seasonally (Quitmyer, Jones, and Arnold, Island (fig. 4.1).
1997; Andrus and Crowe, 2008; Thompson and Shell rings are complex structures having in-
Andrus, 2011). tricate stratigraphic sequences representing mul-
Although oyster and hard clam studies are es- tiple cultural and natural processes (Marrinan,
sential for characterizing seasonal resource use, 2010: 80; Sanger and Thomas, 2010: 55; Kennett
they do so for only two organisms, oysters and and Culleton, chap. 2, this volume; Thompson
clams. Zooarchaeological research at coastal and Andrus, 2011). Radiocarbon dates indicate
southeastern sites shows that over a hundred dif- that otoliths from the Cannon’s Point Ring were
ferent animals were used in the past, in addition to deposited between 2760 ± 120 cal b.c. and 2200
oysters and clams (Colaninno, 2010; Reitz et al., ± 140 cal b.c. (Marrinan, 2010: 81). Dates from
the West Ring indicate deposition between 2320 were sampled. For the archaeological specimens
± 120 cal b.c. and 1970 ± 160 cal b.c. (Marri- this ensured that each otolith represents a dif-
nan, 2010: 81); however, dates were not obtained ferent individual. Unfortunately, these criteria
from locations where otoliths were sampled. De- excluded many otoliths, and more otoliths were
position from the location in which otoliths were sampled from some shell rings than from others
sampled from the St. Catherines Shell Ring dates (appendix 4.2).
between 2580 and 1770 cal b.c. (Sanger and The archaeological otoliths were immersed
Thomas, 2010: 62). Dates from the McQueen in distilled water for 24 hours and then brushed
Shell Ring suggest deposition between 2460 and with a nylon-polyester blend bristled toothbrush
1850 cal b.c. (Sanger and Thomas, 2010: 63); to remove adhering sediments from the exterior
however, materials were not dated from locations surface before they were cross-sectioned. I im-
where otoliths were sampled. mersed the modern otoliths in 30% H2O2 solu-
tion for 12 hours and rinsed them in distilled
METHODS water to remove organic contaminates. After all
otoliths were cleaned, I cross-sectioned each oto-
I chose hardhead catfishes and Atlantic croak- lith along the transverse plane through the cen-
ers because their otoliths are ubiquitous in these ter of the core with a Buehler Isomet low-speed
four shell ring collections (Colaninno, 2010). diamond-wafering saw model 11-1280-160 to
Additionally, these two species have relatively expose the otolith’s cross section. I then attached
large otoliths that are identifiable to species and each otolith section to a glass microscope slide
easily handled. with crystal bond, polished the otolith section
Seasonal use of archaeological hardhead to a flat surface, again rinsed the specimen with
catfishes and Atlantic croakers is estimated by distilled water, and allowed each specimen to air
analyzing the seasonal variation recorded in dry completely.
δ18Ootolith in relation to environmental oscilla- Isotopic analysis of fish otoliths was con-
tions. The original research design was to capture ducted at the University of Alabama stable iso-
modern hardhead catfishes and Atlantic croakers tope laboratory with a New Wave/Merchantek
each month for one year to ensure that δ18Ootolith micromill. Multiple drilling paths were interpo-
profiles track seasonal temperature fluctuations. lated across the cross-sectioned surface. Several
Unfortunately, monthly fishing trips did not al- drilling paths were sampled from some growth
ways result in the capture of a fish. To supple- bands, depending on the size of the band. In
ment the modern otolith study, I selected hard- some cases, however, growth bands were too
head catfish and Atlantic croaker otoliths for small to drill a single isolated band and two
isotopic analysis from individuals with known bands were included in one drilling path sample.
dates and locations of capture and biological Drilling paths were spaced across the cross-sec-
measurements curated in the zooarchaeologi- tioned surface of the otolith at regular intervals
cal comparative collection of the Georgia Mu- of approximately 0.1–0.2 mm. This method of
seum of Natural History and the environmental drilling sampled growth bands through reserved
archaeology program of the Florida Museum of ontogenic sequence; that is from the terminal
Natural History. I analyzed 24 modern otoliths, edge through the core (all δ18Ootolith are plotted
12 from each species (appendix 4.1). through ontogeny in the figures). For hardhead
I selected 42 archaeological otoliths for δ18O catfish otoliths, I sampled the entire cross-sec-
analysis using several criteria. Archaeologi- tioned surface (fig. 4.2). For Atlantic croaker
cal otoliths first were examined for evidence of otoliths, I sampled the medial surface (surface
burning or leaching. These taphonomic pro- with sulcus) to the center core (fig. 4.3). I did
cesses affect δ18O of preserved CaCO3 (Andrus not sample the lateral surface of Atlantic croak-
and Crowe, 2002). Additionally, only complete er otoliths because portions of the knobby tex-
otoliths with lengths greater than 7.0 mm were ture on the lateral surface often were not pres-
selected because it is difficult to section smaller ent on archaeological specimens. The complete
otoliths without destroying them. Otherwise, I length, width, and medial surfaces were present
sampled the full size range of otoliths in each on all studied otoliths. I noted the location of
archaeological collection. Only left otoliths of each drilling path within the otolith and within
both archaeological and modern individuals the growth band during isotopic sampling when
St. Catherines
Ring
Study Area
McQueen
Ring
Cannon’s Point
West Ring
0 5 10
KM
Fig. 4.1. Map of study area showing location of shell rings.
incremental bands clearly were definable and δ18Ootolith FRACTIONATION FACTORS

visible using the microscopic feature of the New AND ESTUARINE ENVIRONMENTS
Wave/Merchantek micromill.
Micromilling the sectioned surface of the Otoliths precipitate in oxygen isotopic equi-
otolith produced a powdered sample. I cleaned librium with ambient waters in which the otolith
the microdrill with compressed air after every forms (Devereux, 1967; Hoefs, 1997; Thorrold
sample to prevent cross-contamination. These et al., 1997). δ18Ootolith covaries with water tem-
samples were exposed to phosphoric acid to perature and δ18Owater. If δ18Owater can be held con-
generate CO2 gases, which were processed by a stant, δ18Ootolith can be used to estimate the water
Finnigan Mat Delta-E isotope ratio mass spec- temperature in which the fish lived (Thorrold et
trometer. Working standards also were processed al., 1997; Andrus et al., 2002). When δ18Ootolith is
after every seventh sample to ensure precision plotted through ontogeny, seasonal temperature
and accuracy of the mass spectrometer. oscillations should be observed, and the δ18Ootolith
Results of this analysis yielded raw oxygen samples prior to and at the terminal edge of
isotopic signatures that were processed and re- growth can determine the approximate tempera-
ported in per mil (‰) units with respect to VPDB ture the fish experienced when it was captured.
(Vienna Pee Dee Belemnite Standard). Precision For estuarine environments, such as those sur-
was estimated and data were calibrated to VPDB rounding Georgia Sea Islands, δ18Owater cannot be
through analysis of the NBS-19 standard. Preci- held constant and temperature calculations from
sion ranges from 0.08‰ and 0.37‰. Averaged δ18Ootolith will not be the actual temperature. This
precision of all samples is 0.17‰ (1σ). is because inputs of fresh water, with lower sa-
Temperatures (T) are calculated with δ18Ootolith linities compared to sea water, alter δ18Owater. Es-
using the Grossman and Ku (1986) aragonite tuaries are, by definition, areas of partly enclosed
precipitation equation where T is temperature in water with freshwater inputs. In some regions
degrees Celsius: of the world, freshwater inputs are seasonal and
δ18Ootolith (PDB) = [(T–20.6/–4.34)] + δ18Owater δ18Ootolith tracks these seasonal changes in salinity,
(SMOW)
. which can indicate season of capture (Kennett
I use several values for δ18Owater (SMOW), which and Voorhies, 1996). In other regions, freshwater
are stated in the various temperature estimates. influxes are erratic, complicating δ18Owater.
core drilling path

example
dark
band
light
band
drilling path example

Fig. 4.2. Cross section of hardhead catfish otolith, FLMNH-EA 11341. Black lines represent drilling path
examples for δ18O sampling. Multiple drilling paths were taken from each otolith. For example, 28 drilling paths
were sampled from FLMNH-EA 11341 across the entire surface of the otolith, within incremental bands.
drilling path
example
core
light band
dark band
Fig. 4.3. Cross section of Atlantic croaker otolith, FLMNH-EA 11319. Black lines represent drilling path exam-
ples for δ18O sampling. Multiple drilling paths were taken from each otolith. Nine drilling paths were sampled from
FLMNH-EA 11319 across the medial surface (surface with sulcus) to the center core, within incremental bands.
To understand oscillations in δ18Ootolith, mod- ple just prior to the terminal increment could es-
ern oscillations in temperature and salinity for tablish seasonal temperature fluctuations and the
Georgia estuaries must be considered. Water season in which people captured the fish. I dem-
temperatures of Georgia estuaries (fig. 4.4) gen- onstrate this using data from Andrus and Crowe
erally oscillate over a seasonal range of approxi- (2008) and the Grossman and Ku (1986) arago-
mately 20°C (http://cdmo.baruch.sc.edu/ data nite precipitation equation. With recorded tem-
from Marsh Landing, Sapelo Island water data perature, salinity, and a constant δ18Owater value as
year 2003; Andrus and Crowe, 2008). Warmest the average water value of –0.9‰ from monthly
temperatures occur from June to August, and collections, estimates of δ18Ootolith are calculated
coolest temperatures occur from December to (Andrus and Crowe, 2008).
February (http://cdmo.baruch.sc.edu/ data from As seen in figure 4.5, when δ18Owater is held
Marsh Landing, Sapelo Island water data year constant, δ18Ootolith accurately reflects seasonal
2003; Andrus and Crowe, 2008). Freshwater temperature oscillations. That is, the most en-
inputs and salinity also vary, both spatially and riched δ18Ootolith would precipitate during cool-
temporally. Monthly average salinity generally est months, winter, while the most depleted
ranges between approximately 16‰ and 31‰ δ18Ootolith would precipitate during warmest
(http://cdmo.baruch.sc.edu/ data from Marsh months, summer.
Landing, Sapelo Island, water data, years 1996 Figure 4.5 models the ideal circumstance for
and 2003), although lower salinity values are using δ18Ootolith to estimate season of capture, but
reported (Thompson and Andrus, 2011). Unlike this does not occur in estuarine waters of the
temperature oscillations, salinity does not vary Georgia coast. δ18Owater fluctuates due to both sea-
seasonally and no regular salinity cycle is ob- sonal temperature oscillations and erratic salinity
served, apart from those associated with daily fluctuations. Thus, the δ18Ootolith profiles are more
tidal cycles (fig. 4.4). complex than figure 4.5 suggests.
Unpredictable salinity fluctuations make ac- Using actual monthly measured δ18Owater
curate temperature estimations with δ18Ootolith (range from –2.2‰ to 0.4‰) from Andrus and
problematic. If δ18Owater values can be held con- Crowe (2008), a model (fig. 4.6) is generated
stant (e.g., seawater), δ18Ootolith could accurately based on the Grossman and Ku (1986) arago-
estimate temperatures. The δ18Ootolith sample of nite precipitation equation. From figure 4.6, a
the terminal increment combined with the sam- semisinusoidal pattern in estimated δ18Ootolith is
seen. Estimated δ18Ootolith values are depleted dur- δ18Ootolith, I examined modern otoliths for season-
ing warmest months, while enriched values form al δ18Ootolith oscillations, converting δ18Ootolith to
during coldest months. Salinity complicates esti- temperature in degrees Celsius, with a constant
mated δ18Ootolith values, however. For example, in δ18Owater of –0.9‰ (the yearly average δ18Owater
October, the temperature is 22.8°C, but δ18Owater from waters surrounding St. Catherines Island
is depleted due to low salinity levels from fresh [Andrus and Crowe, 2008]). These temperature
water, generating more depleted δ18Ootolith values. values are approximations of the actual tempera-
To examine effects of fluctuating δ18Owater on ture the fish experienced.
A 35
30
MEASURED TEMPERATURE º C
25
20
Max
15
Mean
10 Min
0
B 35
30
25
SALINITY ‰
20
Max
15
Mean
10 Min
0
Fig. 4.4. A. Mean temperature per month for 2003 with recorded temperature maximums and minimums at
Sapelo Island Marsh Landing. B. Mean salinity per month for 2003 with recorded salinity maximums and mini-
mums. See http://cdmo.baruch.sc.edu/ for Marsh Landing, Sapelo Island, water data.
Temperature º C Salinity ‰
Average δ Owater
18
Estimated δ18Ootolith
TEMPERATURE ºC AND SALINITY ‰
32 1.5
28 1.0
0.5
24
0.0
20 -0.5
δ18Ootolith‰
16 -1.0
12 -1.5
-2.0
8
-2.5
4 -3.0
0 -3.5
Jan Feb March April May June July Aug Sep Oct Nov Dec
Fig. 4.5. Annual cycle of temperature oscillation and salinity fluctuations for the McQueen Inlet of St.
Catherines Island, Georgia (Andrus and Crowe, 2008). Estimated δ18Ootolith is based on an average and constant
δ18Owater. Calculation based on Grossman and Ku (1986).
Temperature º C Salinity ‰
Measured δ Owater
18
TEMPERATURE º C AND SALINITY ‰
32 1.5
28 1.0
24 0.5
0.0
20
δ18Ootolith‰
-0.5
16
-1.0
12
-1.5
8
-2.0
4 -2.5
0 -3.0
Jan Feb March April May June July Aug Sep Oct Nov Dec
Fig. 4.6. Annual cycle of temperature oscillation and salinity fluctuations for the McQueen Inlet of St. Cath-
erines Island, Georgia (Andrus and Crowe, 2008). Estimated δ18Ootolith is based on a recorded δ18Owater (Andrus
and Crowe, 2008). Calculation based on Grossman and Ku (1986).
δ18O OF MODERN OTOLITHS fluctuations correlating with the expected sea-

son in which the hardhead catfish was captured
When the modern hardhead catfish δ18Ootolith (table 4.1).
are plotted through ontogeny, oscillations in Like the modern hardhead catfish δ18Ootolith,
δ18Ootolith are observed, reflecting seasonal tem- the modern Atlantic croaker δ18Ootolith show
perature oscillation (fig. 4.7A). In 75% of the semisinusoidal oscillations in δ18Ootolith when
examined otoliths, the terminal and penultimate plotted through ontogeny, recording seasonal
δ18Ootolith values record seasonal temperature temperature oscillations (fig. 4.7B). The modern
TABLE 4.1
Modern Hardhead Catfish Otolith Season of Capture and Estimated Season of Capture
Catalog no. Date captured Season captured Estimated season
captured δ18Ootolith
GMNH 2172 5/10/81 Spring Spring

GMNH 4457 6/27/09 Summer Spring
GMNH 1732 7/25/80 Summer Summer
FLMNH-EA 11162 8/9/79 Summer Summer
TABLE 4.2
Modern Atlantic Croaker Otolith Season of Capture and Estimated Season of Capture
Catalog no. Date captured Season captured Estimated season
captured δ18Ootolith
GMNH 1075 3/1/80 Winter Winter

GMNH 4471 9/18/09 Summer Indeterminate
A 1.0 35
0.5
30
ESTIMATED TEMPERATURE ºC
0.0
-0.5 25
δ18Ootolith‰(VPDB)
-1.0
20
-1.5
15
-2.0
-2.5 10
-3.0
5
-3.5
-4.0 0
δ18Ootolith Estimated Temperature
B 2.0 40
1.5
1.0 35
ESTIMATED TEMPERATURE ºC
0.5
30
0.0
-0.5
25
-1.0
-1.5 20
-2.0
-2.5 15
-3.0
10
-3.5
-4.0 5
-4.5
-5.0 0
Fig. 4.7. A. Modern hardhead catfish (FLMNH-EA 11271) δ Ootolith captured in August 1979, Walburg Creek,
18
St. Catherines Island. δ18Ootolith indicates that this individual was captured when water temperatures were warm.
B. Modern Atlantic croaker (GMNH-EA 4462) δ18Ootolith captured in June 2009, North Beach, St. Catherines Is-
land. δ18Ootolith indicates that this individual was captured when water temperatures were warm.
Atlantic croaker δ18Ootolith correctly predicted the modern δ18Ootolith accurately indicated the season
season of capture in 83% of the sample, which is in which the fish died. As such, δ18Ootolith pro-
a higher level of accuracy compared to hardhead files of archaeological specimens can indicate
catfish δ18Ootolith (table 4.2). the season in which people captured the fish.
The archaeological δ18Ootolith sample just before
δ18O OF ARCHAEOLOGICAL OTOLITHS terminal growth and at terminal growth, in addi-
tion to δ18Ootolith fluctuations experienced through
In 75% of modern hardhead catfish otoliths ontogeny, are used to characterize temperature
and 83% of modern Atlantic croaker otoliths, fluctuations occurring when the fish died. Little
to no temperature difference is noted between in winter (GMNH 02420227 and 02420261),

seasonal transitions, for example March 20 (the two were captured in spring (GMNH 02420057
last day of winter) and March 21 (the first day and 02420100), and one was captured in sum-
of spring) (http://cdmo.baruch.sc.edu/ data from mer (GMNH 02420187). These data suggest that
Marsh Landing, Sapelo Island water data years people fished during all four seasons.
1996 and 2003). The application of seasonal clas-
sification using archaeological δ18Ootoliths does not The St. Catherines Shell Ring
directly equate to the actual season, but rather, I sampled 16 hardhead catfish otoliths and
winter is indicated by cold temperatures, spring nine Atlantic croaker otoliths from the St. Cath-
is suggested by the transition of cold to warm erines Shell Ring for isotopic analysis (table
temperatures, summer is indicated by warm tem- 4.3). Sixteen of the archaeological δ18Ootolith
peratures, and fall is suggested by a warm to cold yielded profiles that provide evidence of sea-
temperature transition. If all four seasons are rep- son of capture. One δ18Ootolith from a hardhead
resented in the archaeological δ18Ootolith profiles catfish indicates that people captured this fish
from a site, fishing likely occurred during all four during winter (GMNH 02381373), four hard-
seasons. The Grossman and Ku (1986) aragonite head catfish (GMNH 02380115, 02381373, and
precipitation equation is used to convert δ18Ootolith two otoliths from 02380118) were captured in
to temperature in degrees Celsius, with the con- spring, and two were captured in fall (two oto-
stant δ18Owater of –0.9‰ for the following analysis liths from GMNH 02381373). δ18Ootolith from
(fig. 4.8A–D and 9). archaeological Atlantic croakers (fig. 4.8C) indi-
cate that people captured one individual in win-
The Cannon’s Point Ring ter (GMNH 02381348), two in spring (GMNH
Only one of the two otoliths examined from 02381477 and 02381554), four in summer
the Cannon’s Point Ring yielded conclusive re- (GMNH two from 02381444, and one each from
sults, FLMNH-EA 01550585 (table 4.3). This At- 02380999 and 02381886), and two were cap-
lantic croaker otolith indicates that the fish died tured in fall (GMNH 02381203 and 02381516).
while temperatures were cold, suggesting that These data suggest that people fished during all
people fished at least during winter (fig. 4.8A). four seasons.
The West Shell Ring Summary

I analyzed two hardhead catfish and six Atlan- δ18Ootolith profiles of archaeological specimens
tic croaker otoliths for δ18Ootolith profiles from the indicate that people fished during all four seasons
West Ring (table 4.3). People who formed this at the West, McQueen, and St. Catherines shell
ring captured one hardhead catfish during sum- rings. The limited sample from the Cannon’s
mer (FLMNH-EA 01610161) and one during fall Point Ring prevents a comprehensive understand-
(FLMNH-EA 01610027). One Atlantic croaker ing of seasonal fishing behaviors at this site.
otolith yielded inconclusive results (FLMNH-
EA 01610231). Three Atlantic croakers were δ18Ootolith, ESTUARINE ENVIRONMENTS,
captured during winter (FLMNH-EA 01610169, AND ARCHAEOLOGICAL IMPLICATIONS
01610194, and 01610214), one was captured in
the spring (FLMNH-EA 01610214), and one was Modern δ18Ootolith
captured during fall (FLMNH-EA 01610169). Problems and Areas of Future Research
These data suggest that fishing occurred during In most cases, modern δ18Ootolith accurately de-
all four seasons (fig. 4.8B, D). termined the season in which the fish died, how-
ever, there are some problems.
The McQueen Shell Ring Shackleton (1973) outlines six requirements
Two hardhead catfishes and five Atlantic for a successful archaeological application of
croakers were analyzed for δ18Ootolith profiles. All δ18Ocarbonate to determine season of capture. To
otoliths from the McQueen Shell Ring yielded paraphrase Shackleton’s requirements, he notes
seasonal results (table 4.3). One hardhead catfish that: (a) the species must form their carbonate
was captured during spring (GMNH 02420216), structure in isotopic equilibrium with the ambient
while the other was captured during fall (GMNH water; (b) the isotopic composition of the water in
02420047). Two Atlantic croakers were captured which the carbonate structure forms must remain
TABLE 4.3
Archaeological Otolith Estimated Season of Capture
δ18O ‰(VPDB)
Catalog no. Ring Species Otolith no. One before Terminal Estimated season
terminal captured
FLMNH-EA Cannon’s Point A. felis Cannon’s-1 −0.93 — Indeterminate
01551061
FLMNH-EA Cannon’s Point M. undulatus Cannon’s-3 0.47 1.28 Cold
01550585
FLMNH-EA West A. felis West-1 −0.70 −0.70 Warm/cold
01610027 transition
FLMNH-EA West A. felis West-2 −0.44 −0.96 Warm
01610161
FLMNH-EA West M. undulatus West-3 −1.90 −1.28 Warm/cold
01610169 transition
FLMNH-EA West M. undulatus West-4 −0.30 1.51 Cold
01610169
01610194
FLMNH-EA West M. undulatus West-8 0.21 −0.71 Cold/warm
01610214 transition
01610214
FLMNH-EA West M. undulatus West-10 0.84 0.62 Indeterminate
01610231
GMNH 02420216 McQueen A. felis McQueen-1 −0.31 −0.43 Cold/warm
transition
GMNH 02420047 McQueen A. felis McQueen-3 −0.86 −0.68 Warm/cold
transition
GMNH 02420227 McQueen M. undulatus McQueen-2 1.05 0.31 Cold
GMNH 02420187 McQueen M. undulatus McQueen-4 −1.52 −1.62 Warm
GMNH 02420261 McQueen M. undulatus McQueen-5 −0.75 0.69 Cold
GMNH 02420057 McQueen M. undulatus McQueen-6 0.28 −0.81 Cold/warm
transition
GMNH 02420100 McQueen M. undulatus McQueen-7 0.51 −0.45 Cold/warm
transition
GMNH 02380115 St. Catherines A. felis 0115-O1 −0.73 −0.54 Indeterminate
GMNH 02380115 St. Catherines A. felis 0115-O2 −0.39 −0.95 Cold/warm
transition
transition
GMNH 02380118 St. Catherines A. felis 0118-O4 −1.39 — Indeterminate
GMNH 02380118 St. Catherines A. felis 0118-O5 −0.59 −0.46 Indeterminate
GMNH 02380118 St. Catherines A. felis 0118-O6 0.02 −1.01 Cold/warm
transition
GMNH 02380120 St. Catherines A. felis 0120-O8 0.06 — Indeterminate
GMNH 02381373 St. Catherines A. felis 1373-O4 −0.14 −0.18 Warm/cold
transition

δ18O ‰(VPDB)
Catalog no. Ring Species Otolith no. One before Terminal Estimated season
terminal captured
transition
GMNH 02381373 St. Catherines A. felis 1373-O6 −0.30 −0.28 Warm/cold
transition
GMNH 02381373 St. Catherines A. felis 1373-O7 0.12 0.40 Cold
GMNH 02381437 St. Catherines A. felis 1437-O8 — −0.08 Indeterminate
GMNH 02381037 St. Catherines A. felis 1037-O19 — −1.11 Indeterminate
GMNH 02381348 St. Catherines M. undulatus 1348-O3 −0.17 0.32 Cold
GMNH 02381444 St. Catherines M. undulatus 1444-O9 −1.97 −1.90 Warm
GMNH 02381477 St. Catherines M. undulatus 1477-O11 −0.81 −0.74 Cold/warm
transition
GMNH 02381516 St. Catherines M. undulatus 1516-O12 −1.34 −0.75 Warm/cold
transition
GMNH 02381554 St. Catherines M. undulatus 1554-O13 0.63 −0.74 Cold/warm
transition
GMNH 02381203 St. Catherines M. undulatus 1203-O25 −1.41 −0.08 Warm/cold
transition
constant through a season; (c) the water tempera- leton, 1973: requirement b). It is clear from the
ture in which the carbonate structure forms must zooarchaeological record of the Georgia coast
be sea temperature; (d) the organism must depos- that estuarine fishes were an important subsis-
it the carbonate structure throughout the year; (e) tence resource (Reitz, Quitmyer, and Marrinan,
growth in the carbonate structure must be great 2009; Colaninno, 2010; Reitz et al., 2010: 54–
enough to allow for discrete sampling of incre- 57). People fished estuarine fishes heavily, but
ments covering no more than a few weeks; and they did not fish in offshore areas where δ18Owater
(f) the habitat in which the organism lives must values are more stable. To characterize seasonal
undergo large and regular seasonal temperature fishing efforts in inshore settings, this require-
oscillations. Although hardhead catfish and At- ment must be met. It is possible that a more
lantic croaker otoliths meet requirements a, c, comprehensive environmental understanding of
and f, otoliths from these two species do not meet Georgia estuaries can generate accurate δ18Ootolith
requirements b, d, and e. This may explain why models that account for temperature, salinity,
some of the modern δ18Ootolith did not indicate the and δ18Owater (see Surge and Walker, 2005; An-
correct season of capture. drus and Crowe, 2008; Culleton, Kennett, and
Firstly, δ18Owater in which the fish lives does Jones, 2009). Monthly analysis of δ18Owater in
not remain constant through the season (Shack- conjunction with δ18Ootolith from several locations
surrounding St. Catherines Island may resolve other fish species (Jones and Campana, 2009)
this issue. and could be practiced by hardhead catfishes
Fishes are mobile animals and movement of and Atlantic croakers. Developing better mod-
fishes confounds fluctuations in δ18Owater, again els to predict δ18Ootolith through further study of
violating Shackleton’s (1973) requirement b. δ18Owater, temperature, and salinity may resolve
Both hardhead catfishes and Atlantic croak- some of the confounding factors of temperature
ers move within estuaries and outside estuaries and salinity fluctuations on δ18Owater, but seasonal
(Dahlberg, 1972; Muncy and Wingo, 1983; Hare temperature oscillations will be difficult to detect
and Able, 2007). Fish movement complicates in fishes with thermoregulatory movement.
documenting environmental oscillations within Shackleton (1973: requirement d) also notes
Georgia estuaries. Each individual fish may ex- that formation of the carbonate structure used to
perience different environmental parameters with determine seasonality must form throughout the
movement. Alternatively, these species may have year. Growth in fish otoliths may violate this re-
a specific set of preferred environmental condi- quirement. Researchers note stoppage of otolith
tions, including temperature and salinity, within growth in other fish species (Fowler, 1995) and
which the fish swims as these parameters move this could occur in hardhead catfishes and Atlantic
with tides and seasons. This behavior, known as croakers. This is particularly likely for hardhead
thermoregulatory movement, is documented in catfishes. Males of this species gestate their eggs
A B
Estimated Temperature º C
1.0
30 1
30
0.0 0
-1.0 20 -1 20
-2.0 -2
10 10
-3.0 -3
-4.0 0 -4 0
C D
40 1.0 40
1.0
0.0
0.0 30 30
-1.0 -1.0
20 -2.0 20
-2.0
-3.0 -3.0 10
10
-4.0 -4.0
-5.0 0 0

Fig. 4.8. δ Ootolith through ontogeny. A. Atlantic croaker (FLMNH-EA 01550585) δ18Ootolith profile indicating
18
cold temperatures, or winter season of death from the Cannon’s Point Ring. B. Atlantic croaker (FLMNH-EA
01610214) δ18Ootolith profile indicating a cold to warm temperature transition, or spring season of death from the
West Ring. C. Atlantic croaker (GMNH 02381886) δ18Ootolith profile indicating a warm, or summer season of
death from the St. Catherines Shell Ring. D. Atlantic croaker (FLMNH-EA 01610169) δ18Ootolith profile indicating
a warm to cold temperature transition, or fall season of death from the West Ring.
and young in their mouth for approximately 60 through ontogeny in response to environmental
to 80 days after spawning (Dahlberg, 1972). Oral and biological conditions such as temperature,
gestation occurs during the late spring through food availability, and reproductive efforts (Pan-
early fall (Gunter, 1947; Ward, 1957). During this nella, 1971, 1980; Wootton, 1998:111–116; Van
period, males do not eat and somatic growth may Neer, Löugas, and Rijnsdorp, 2004). Major fluc-
not occur. The fact that only one archaeological tuations in environmental and biological condi-
hardhead catfish otolith had a terminal increment tions that the fish experiences generally occur
that formed in summer further suggests that oto- annually, causing fish to grow quickly and then
lith growth is not continuous throughout the year. slowly within an annual cycle. This process pro-
Hardhead catfish are most abundant in Georgia duces two visually distinct bands per year (dark
estuaries during summer (Dahlberg, 1972), mak- and light growth bands). As such, the number of
ing the limited evidence for summer-captured bands in an otolith should correlate to the number
hardhead catfish surprising. Nonetheless, mod- of isotopic peaks in δ18Ootolith profile when plotted
ern hardhead catfish δ18Ootolith generally do form through ontogeny (figs. 4.7, 8A–D, and 9). This
otolith increments in warmer months. Additional is not always the case, however. In some otoliths,
modern sampling of both males and females, in the number of increments is equal to the number
known states of reproduction, is required to un- of seasonal peaks in the δ18Ootolith profile. In other
derstand consequences of reproduction on so- otoliths, there are more increments than peaks.
matic growth and otolith formation. This suggests that either too little growth oc-
Shackleton (1973: requirement e) also ob- curred in these otoliths to discretely sample each
serves that growth rates of the carbonate struc- season using the sampling method I employed,
ture must be large enough to sample isolated or increment formation does not correlate to sea-
increments covering no more than a few weeks sonal temperature oscillations in every case.
of growth. Some archaeological otoliths do not The effect of growth increments that cover
meet this requirement, particularly otoliths from no more than a few weeks also is seen in FLM-
older fishes. This problem is noted when peaks NH-EA 01610231 (fig. 4.9). This individual was
in δ18Ootolith profiles are examined in conjunction 18.5 ontogenic years at death and formed 1.07
with the number of incremental bands in each mm of noncore otolith (from the terminal edge
otolith. Theoretically, fish form growth bands of the otolith to the outer edge of the core). As-
1
30
Estimated Temperature ºC
0
-1 20
-2
10
-3
-4 0
Fig. 4.9. Archaeological δ18Ootolith of an Atlantic croaker (FLMNH-EA 01610231) approximately 18.5 years of
age. δ18Ootolith during 17.5 years of noncore growth ranges between –0.25‰ and 1.05‰ or a 5.7°C difference in
experienced temperature.
suming an average growth rate over the 17.5 tions fishes experienced just prior to and at death
years of noncore growth, only 0.06 mm of otolith and thus, season of capture. Further research is
formation occurred per year or 0.01 mm every needed to understand the full seasonal range of
three months (i.e., per season). I spaced isotopic environmental oscillations, particularly correla-
drilling paths for both modern and archaeologi- tions among temperature, salinity, and δ18Owater
cal otoliths between 0.1 and 0.2 mm apart. It is in Georgia estuaries, and biological and environ-
impossible to isolate season of capture for older mental factors that affect otolith growth and in-
individuals with such little growth each season. cremental banding in these two species.
Because fish generally grow most rapidly during
their first few years of life (Wootton, 1998: 116), Archaeological Implications
otoliths from younger individuals should provide Researchers have conducted analyses of clam
higher-resolution trends in temperature oscilla- and oyster incremental formation and δ18O pro-
tions experienced by the fish. files at several Late Archaic shell rings, includ-
Researchers also use visual characterization ing the shell rings in this otolith study (Quitmyer,
of the terminal band type of otoliths to determine Hale, and Jones, 1985; O’Brien and Thomas,
seasonality. If the type of banding, that is dark or 2008; Thompson and Andrus, 2011). These mol-
light bands, is known to occur during specific sea- lusc studies concluded that people harvested
sons for a species, the terminal band can be used shellfish during multiple seasons (Marrinan,
to infer the season in which the fish was captured 1975: 99; Thompson and Andrus, 2011), and in
(Smith, 1983; Hales and Reitz, 1992; Erlandson, some instances, throughout the year (Quitmyer,
1994: 103; Higham and Horn, 2000). Currently, Hale, and Jones, 1985; Thompson, 2006; Thomas,
this method yielded inconclusive results for both 2008: 927; Thompson and Andrus, 2011). These
hardhead catfishes and Atlantic croakers. For data suggest a model of year-round shellfish har-
modern hardhead catfishes, individuals were cap- vesting at most, although not necessarily at all,
tured only during five months of the year (May, shell rings (Thompson and Andrus, 2011).
June, July, August, and September), representing This archaeological δ18Ootolith study also sug-
two seasons: spring and summer. All hardhead gests a year-round model for fishing. At the
catfishes in the sample had dark or fast, terminal West, McQueen, and St. Catherines shell rings,
growth bands when they were captured. Although δ18Ootolith profiles just prior to and at death indicate
this indicates that hardhead catfish likely form dark that people captured these two fish species during
growth bands in the spring and summer, these data all four seasons in conjunction with multiseason
do not indicate whether dark growth occurs con- harvesting of shellfishes (table 4.3). Researchers
tinuously throughout the spring and summer and in other coastal areas document a similar pattern
when light growth bands form. Season of death for of year-round resource use by coastal peoples
Atlantic croakers in the modern sample includes in the past (Yesner, 1980; Russo, 1998; Keene,
croakers that died in spring and summer. Only one 2004; Orr, 2007).
croaker captured in spring was in noncore growth Although δ18Ootolith profiles define a pattern
and this individual was forming a dark growth of year-round fishing at the West, McQueen, and
band. Otoliths from Atlantic croakers captured in St. Catherines shell rings, this does not mean the
summer had both dark and light growth terminal entire population associated with each shell ring
bands. One otolith was in light growth formation lived there continuously. Segments of the popu-
and four were in dark growth formation. These lation could be absent occasionally on collecting
data indicate that dark bands generally form dur- and hunting trips, social visits, or ritual activities
ing spring and summer; however, without direct for longer or shorter lengths of time. Such sea-
evidence that light bands occur in winter and fall, sonal patterns of movement by only segments
it is impossible to draw this conclusion. of a population are documented among modern
Characterizing the terminal band in otoliths hunter-gatherer populations (Meehan, 1982: 31–
may provide an additional method to determine 41; Kelly, Poyer, and Tucker, 2005) and likely
season of capture, but more modern sampling is characterize the behavior of archaeological popu-
needed to understand seasonal timing of band lations.
formation in these two species. Overall results Admittedly, evidence that people fished Geor-
from the modern δ18Ootolith indicate that oscilla- gia estuaries and deposited the fish remain at
tions in δ18Ootolith can estimate temperature oscilla- these shell rings during all four seasons does not
directly equate to a year-round, sedentary popula- head catfishes and Atlantic croakers during mul-
tion occupying shell rings. Temperature regimes tiple seasons. At the West, McQueen, and St.
at time of capture represent a discrete time inter- Catherines shell rings, people fished during all
val during a period of three to four months. Ar- four seasons. With these data, future archaeologi-
chaeological δ18Ootolith profiles are not sufficiently cal research can focus on patterns in the horizon-
precise to identify patterned, but brief, absences tal and vertical distributions of seasonal deposits
or presences of fishing activities. It is possible within each shell ring and across the region, as
that people only fished and deposited fish remains well as other spatial, temporal, and behavioral
at these shell rings for a day, week, or month of scales. Additionally, seasonal data from molluscs
the three to four month period defined using ar- can be examined in conjunction with data from
chaeological δ18Ootolith. Given the large quantity fish otoliths to view seasonal resource use involv-
of faunal remains (Colaninno, 2010), depth of ing several taxa (Walker and Surge, 2006). Multi-
deposits at these sites, and the multiseason har- proxy studies may provide a better prospective on
vesting practices documented in mollusc samples Late Archaic seasonal resource use patterns and
(Quitmyer, Hale, and Jones, 1985; O’Brien and settlement patterns than single-proxy studies.
Thomas, 2008; Thompson and Andrus, 2011), an
extended, year-round occupation is probable. NOTES
SUMMARY AND CONCLUSIONS 1. Isotopic analysis of modern, Cannon’s Point, West,

and McQueen otoliths is funded by the National Science
Foundation Doctoral Dissertation Improvement Grant
This study is one of the first to use archaeo- BCS-0912176. Isotopic analysis for the St. Catherines Shell
logical fish otoliths to determine the seasonality Ring is funded by the American Museum of Natural History
of fishing in the southeastern United States (for and Edward J. Noble Foundation. I would like to thank all
seasonal studies that used otoliths in other re- those who have helped me with this research: David Hurst
Thomas, Matthew C. Sanger, Ginessa Mahar, and the crew
gions see, Erlandson, 1994: 103; Van Neer, Löu- at the American Museum of Natural History for allowing me
gas, and Rijnsdorp, 1999; Van Neer et al., 2004; to work with the St. Catherines Island collections, Rochelle
Hufthammer et al., 2010). Although initial results Marrinan for allowing me to work with the St. Simons
of this analysis are promising, further research is material, Bruce Saul, Charles Lambert, and their students at
needed. Factors that control δ18Ootolith for Georgia Augusta State University for collecting fishes, Irvy Quitmyer
for helping me collect fishes and aiding me at the Florida
estuaries, as well as biological and environmen- Museum of Natural History, Fred Andrus for allowing me to
tal factors that control otolith formation must be work at the University of Alabama, Joe Lambert for process-
better understood. Despite these shortcomings, ing my isotopic samples, John A. Turck for the map, Julia
modern δ18Ootolith profiles indicate that δ18Ootolith K. Orr for processing the modern fishes, and Betsy Reitz
for the insightful comments. This research could not have
determines the season in which people captured taken place without your support, so thank you. The author
the fish in most cases. is responsible for all contents, including the errors and any
These data indicate that people fished hard- misinterpretation of the data herein.
APPENDIX 4.1
Modern Otoliths Sampled for Isotopic Analysis
Catalog # Location Captured Collector Length, Breadth, Width, Thickness, Weight,
mm mm mm mm g
GMNH 2172 Edisto Island, SC 5/10/81 Lisa Osteen 11.53 10.23 9.62 4.97 0.57
GMNH 4457 St. Catherines Island, 6/27/09 Bruce Saul, 8.61 7.68 6.72 3.59 0.22
Ga Carol Colaninno
GMNH 1732 Sapelo Island, 7/25/80 Besty Reitz 11.28 10.12 9.72 4.65 0.54
Caretta Beach, Ga
GMNH 4269 Camachee Island, St. 7/31/03 Kelly Orr 10.81 9.79 8.83 4.15 0.40
Johns County, Fl
FLMNH 11162 St. Catherines Sound, 8/9/79 Richard Casteel 9.55 8.21 8.01 3.72 0.29
SCI2, Ga
FLMNH 11271 St. Catherines 8/9/79 Richard Casteel 10.73 9.58 9.13 3.82 0.39
Walburg Creek, Ga
FLMNH 11341 St. Catherines Island 8/9/79 Richard Casteel 10.20 8.85 8.33 3.85 0.33
SCI1, Ga
Ga Carol Colaninno
Ga Carol Colaninno
GMNH 4295 Vilano Bridge, St. 9/16/03 Kelly Orr 12.10 10.62 9.98 5.70 0.75
Johns County, Fl
North Beach Seine, Ga Carol Colaninno
North Beach Seine, Ga Carol Colaninno
GMNH 1075 Skidaway Island, Ga 3/1/80 D. Miller 8.91 6.63 3.60 0.17
GMNH 4475 Ossabaw Estuary 4/25/09 Carol Colaninno 7.72 5.65 3.40 0.11
Trawl, Ga
GMNH 4458 St. Catherines Island 5/23/09 Carol Colaninno 11.39 8.56 4.84 0.39
Trawl, Ga
GMNH 3762 Gloucester, 5/29/87 J. Gayner 11.04 8.14 5.19 0.39
Middlesex Co., Va
GMNH 4462 St. Catherines Island, 6/27/09 Carol Colaninno 8.98 6.64 3.63 0.19
North Beach Seine, Ga
FLMNH 11318 St. Catherines 8/8/79 Richard Casteel 7.36 5.39 3.42 0.10
McQueen Inlet, Ga
FLMNH 11319 St. Catherines 8/8/79 Richard Casteel 7.28 5.17 2.80 0.08
McQueen Inlet, Ga
FLMNH 11313 St. Catherines Island 8/9/79 Richard Casteel 7.46 5.55 3.29 0.10
SCI2, Ga
GMNH 4473 St. Catherines Island, 8/27/09 Bruce Saul, 8.57 6.37 3.61 0.17
Ga Carol Colaninno
GMNH 4474 St. Catherines Island, 8/27/09 Bruce Saul, 9.99 7.23 4.32 0.25
Ga Carol Colaninno
GMNH 4471 St. Simons Estuary 9/18/09 Bruce Saul, 7.98 5.60 3.42 0.13
trawl, Ga Carol Colaninno
GMNH 4476 St. Simons Estuary 9/18/09 Bruce Saul, 7.91 5.56 3.22 0.12
trawl, Ga Carol Colaninno
APPENDIX 4.2
Archaeological Otoliths Sampled for Isotopic Analysis 100
Catalog no. Ring Species Otolith no. Unit Level Length, mm Breadth, mm Width, mm Thickness, mm Weight, g
FLMNH Cannon’s Point A. felis Cannon’s 1 18N, 0E 25 9.85 8.80 8.19 3.61 0.294
01551061
FLMNH Cannon’s Point M. undulatus Cannon’s 3 18N, 0E 12 9.39 7.30 4.48 0.253
01550585
FLMNH West A. felis West 1 5S, 30E ZII L1 10.52 9.34 8.80 4.52 0.442
01610027
FLMNH West A. felis West 2 5S, 30E 6 10.71 9.49 8.88 4.05 0.407
01610161
FLMNH West M. undulatus West 3 5S, 30E 6 8.60 6.57 4.06 0.176
01610169
01610169
01610194
01610214
01610214
01610231
GMNH 02420216 McQueen A. felis McQueen 1 N263 E179 5.6–5.5 8.85 7.93 7.5 3.52 0.247
GMNH 02420227 McQueen M. undulatus McQueen 2 N263 E179 5.6–5.5 8.14 5.98 3.33 0.142
GMNH 02420047 McQueen A. felis McQueen 3 N213 E179 4.6–4.5 8.08 7.25 6.78 3.72 0.215
GMNH 02380115 St. Catherines A. felis 0115-O1 N789 E801 10–20 cm 9.36 8.29 7.82 3.56 0.270
NO. 97
APPENDIX 4.2 — (Continued)
2012
Catalog no. Ring Species Otolith no. Unit Level Length, mm Breadth, mm Width, mm Thickness, mm Weight, g
GMNH 02381348 St. Catherines M. undulatus 1348-O3 N784 E797 10–20 cm 8.70 6.82 3.69 0.178
GMNH 02381554 St. Catherines M. undulatus 1554-O13 N784 E797 70–75.5 cm 10.32 7.62 4.35 0.284
EVALUATING δ18O PROFILES OF HARDHEAD CATFISH
GMNH 02381886 St. Catherines M. undulatus 1886-O26 N771 E819 2.5–2.4 9.24 7.06 4.24 0.207
101
2012 LATE PREHISTORIC SETTLEMENT PATTERNS: ZOOARCHAEOLOGICAL EVIDENCE 103
CHAPTER 5
LATE PREHISTORIC SETTLEMENT PATTERNS:
ZOOARCHAEOLOGICAL EVIDENCE FROM BACK
CREEK VILLAGE, ST. CATHERINES ISLAND
S arah G. B ergh 1
Despite more than three decades of archaeo- waters only during certain seasons. Sea turtles,
logical research on St. Catherines Island, there for example, spend most of the year offshore, but
are still questions about the residential mobility females lay their eggs on the beaches during the
of late prehistoric populations. Ethnohistoric ac- summer months. Sharks, too, tend to stay offshore
counts of late 16th-century coastal Georgia soci- except for the warmer months when they enter
eties are open to interpretation; some researchers estuaries and nearshore waters. Some migratory
posit that the Guale people moved residences on birds stop along the Georgia coast in the spring
a seasonal basis following available resources, and fall to rest and refuel. Reproductive cycles
while others suggest Guale lived in dispersed but of many species are driven by seasonal events.
permanent settlements. This debate also applies The white-tailed deer (Odocoileus virginianus)
to the late prehistoric period (a.d. 1300–1580), breeding season is generally October–January
when it is assumed Guale or their direct ances- and young are born in May and June. Fishes have
tors occupied the coast. This chapter presents the specific spawning seasons, which vary by spe-
results of zooarchaeological analysis of recently cies. Fish of the same species born in the same
excavated materials from Back Creek Village, a year form a cohort, and each age cohort may have
large, late prehistoric site on St. Catherines Is- different food and habitat preferences. The loca-
land. These data are applied to the “Guale Prob- tion of preferred habitats and food may change
lem” (Thomas, chap. 1, this volume), and provide on a seasonal basis, causing cohorts to move to
new insight, as the materials were fine-screened, different parts of the estuary or out of the estu-
recovered from multiple middens at a single site, ary. Mollusc growth is mostly dependent on en-
and yield a large sample of vertebrate remains vironmental conditions, such as temperature and
and multiple seasonal indicators that speak to salinity, and these often change on a seasonal
seasonality of site occupation. basis. All these phenological events can be used
to interpret the season in which an animal in the
SEASONALITY archaeological record died. It is important to re-
IN FAUNAL ASSEMBLAGES member, however, that there will always be indi-
vidual variation and that the environmental con-
Zooarchaeological remains are often used ditions that trigger phenological events may not
to address seasonal occupation of sites (Monks, always occur on a regular, seasonal schedule—
1981). Life histories of many animal species in- storms and unusually warm, cold, wet, or dry sea-
volve behavioral and physiological events that sons can alter animal behavior and physiology.
occur in response to seasonal changes in temper- The assumptions that follow the interpretation
atures, vegetation, sunlight, and rainfall, among of an archaeological animal’s season of death—
other factors. These phenological events are often that humans killed the animal, the animal’s re-
visible in the archaeological record. Some spe- mains were deposited at the site about the time
cies follow seasonal rounds, occupying estuarine the animal was killed, and the site was therefore
occupied by humans about the time the animal out the rest of the year, households dispersed into
was killed—can be disputed in some cases. On small sets of households (i.e., matrilineages),
St. Catherines Island, however, some evidence moving residences seasonally among habitats
supports these assumptions. Abundant cultural with the most productive resources—oak forests
material throughout the middens reported here in the fall to collect mast and hunt deer, marsh-
supports the assumption that humans killed the side in the winter to collect molluscs, and near
animals and deposited the remains. Although a swidden plots in the spring for planting maize
more tenuous assumption, the bones and shells (Crook, 1986: 18). The model proposed by Jones
deposited at most sites were likely collected and (1978: 190–194), based on his interpretation of
deposited while people occupied the site. Al- Jesuit accounts as exaggerated and other descrip-
though storage of meat is certainly possible, it tions of southeastern Indians, suggests that Guale
is less likely that meat was stored on the bone lived in dispersed, permanent settlements, each of
or shell and transported among sites on a regular which had access to mast and shellfish resources
basis. The bones and shells are heavy and bulky without the need for seasonal, residential mobil-
and the same animals are available in most other ity. In addition, although Jones also understood
locations along the coast. It is important to re- Guale to use a swidden system of maize cultiva-
member, however, that the presence of animals tion, he proposed that maize plots were scattered
killed in all seasons does not necessarily mean around the settlement, new fields being cleared
that the site was occupied permanently through- in the vicinity when necessary, without moving
out the year; it is possible that the site was reoc- to a new site.
cupied intermittently over the course of a single This debate about the Guale settlement-sub-
or multiple years. The absence of animals killed sistence system also applies to the late prehistoric
in a season also is not necessarily evidence that period, as it is assumed that Guale or their direct
the site was abandoned during that season, only ancestors inhabited the coast at this time. A better
that the archaeological excavations failed to understanding of the late prehistoric settlement-
unearth any animals that definitely died in that subsistence system may, therefore, shed light on
season. It is only possible to estimate season of the “Guale Problem.” The 1970s transect survey
death for a small portion of the individuals re- of the island addressed settlement-subsistence
covered at a site. systems on an island-wide scale (Thomas, 2008).
Sites, in general, and large sites in particular, are
THE ‘GUALE PROBLEM’ more numerous during the late prehistoric period
than earlier (Thomas, 2008: 877), suggesting
The “Guale Problem” essentially sets up two larger communities, and more of them. The sites
different models, defined by economic potential tend to be located at the margins of marsh and
and residential mobility (Thomas, 2008: 239). forest habitats (Thomas 2008: 929–933). This
This chapter does not address economic poten- would seem to support the Jones model. Faunal
tial, in terms of maize production, as there is samples, however, were small, not fine-screened,
no quantified study of botanicals from the late and the samples of seasonal indicators from each
prehistoric period on St. Catherines. One maize site were also small (Reitz, 2008: 617, 656–663).
kernel was found at Back Creek Village, the site The data generated by the survey were appropri-
discussed in this chapter, so the plant was utilized ate for addressing macroscale questions, but not
to some extent during the late prehistoric period for site-level questions about intrasite variability
by the people who lived at the site (Ruhl, 2008). in subsistence and seasonality. The recent exca-
The faunal evidence, instead, speaks to patterns vations at Back Creek Village were designed to
of residential mobility. The model proposed by address these site-level questions. Fine-screened
Larson (1980: 206–209, 221–228) and Crook faunal collections and the resulting large samples
(1986: 11–28), based on their interpretation of Je- of vertebrates and clams from multiple middens
suit missionary accounts for the region, suggests from Back Creek are used to address the question
a residentially mobile society that was unable to of whether late prehistoric sites were occupied on
support permanent settlements with maize grown a seasonal basis for the purpose of exploiting sea-
on small patches of poor agricultural soils. The sonally available resources, or if they were oc-
maize harvest in late summer allowed the popula- cupied on a permanent basis. The latter scenario
tion to aggregate for a time in towns, but through- allows for the movement of groups of people for
subsistence, political, or social activities, but not NISP (number of identified specimens), MNI
residential movement of entire communities on a (minimum number of individuals), and biomass
seasonal basis. were estimated for each midden. MNI, based on
symmetry and age, was estimated for the lowest
MATERIALS AND METHODS taxonomic level possible, usually at the species
or genus level, except in cases where more in-
Back Creek Village (9Li207) is located on dividuals were present at the family level. NISP
the seaward side of the island, just inland from assumes that each specimen is from a different
the marsh formed by McQueen Inlet (fig. 5.1). animal, which overestimates the contribution
Thomas (2008: 520, 534) classified the site as of taxa that have large numbers of elements or
large, defined as having an inferred subsurface particularly identifiable elements. MNI assumes
area greater than 500 m². Numerous discrete that the entire animal was present at the site and
shell middens are visible, and others have been tends to overemphasize the relative importance
located through a probe survey. The middens of small species since it gives equal weight to
surround a large depression that was possibly all species, though small species have less meat
man-made and may once have held water. Mid- to contribute to the diet (Reitz and Wing, 2008:
dens vary in size—most are several meters in di- 212–213). Biomass was estimated from speci-
ameter, though others are more amorphous and men weight using allometric formulas published
cover a larger area. The site was identified dur- in Reitz and Wing (2008: 68). This measure is
ing the 1970s transect survey of the island and based on the allometric principle that skeletal
five middens were tested (Thomas, 2008: 584). weight scales to body size as an animal gets
In the 1990s, another midden was tested. The larger, and may be a more accurate reflection
materials recovered in those excavations indi- of dietary contribution because it only assumes
cate that the site was occupied during the Irene the presence at the site of the bones found, and
phase (a.d. 1300–1580). Excavations in 2008 re- estimates meat contribution from those elements
tested five of the previously excavated middens (Reitz and Wing, 2008: 239). It must be noted,
as well as two additional middens, one of which however, that skeletal weight is subject to di-
was a large, amorphous mound (fig. 5.2). Two agenic processes, so specimen weight does not
one-by-one meter units were excavated into each necessarily reflect the weight of the specimen
of the seven middens. The tested middens were when the animal was alive.
given the letter designations A–H, though E was Data were further summarized in a number of
not excavated in 2008. Deposits were water ways. They were assigned to higher taxonomic
screened through nested screens, 1⁄8 inch mesh categories based largely on class and order; this
being the smallest. Vertebrate remains were summary was done at the site level, merging the
sorted in the field and sent to the Zooarchaeol- data for each midden. Richness, diversity, and
ogy Laboratory, Georgia Museum of Natural equitability were estimated for each midden and
History, University of Georgia, for analysis. All for the site as a whole. Richness is the number of
whole hard clam valves (Mercenaria spp.) also species present in the species list, not including
were retained for analysis. commensals (Anura, Caudata, Lacertilia, Sorici-
Identification of the vertebrate remains and dae, Talpidae, Sigmodontinae). Diversity, using
clam valves used the comparative collection of the Shannon-Weaver index, is an index of the
the Georgia Museum of Natural History, fol- relative contribution of each species to the over-
lowing standards established in Reitz and Wing all diet and equitability measures the degree of
(2008: 151–181). Specimens were counted and evenness with which the species were used (Reitz
weighed to aid in establishing relative abundanc- and Wing, 2008: 110–113). The diversity index
es of taxa. Fish otoliths and atlases were mea- ranges from 0 to 5 and is calculated using the for-
sured using lab standards; namely, greatest width mula: H' = –Σ (pi) (logepi), where pi is the relative
for atlases, and greatest length, greatest width, abundance of each taxon, in this case MNI and
and thickness for otoliths. Age estimates for biomass estimated from specimen weight. The
deer were based on epiphyseal fusion and tooth equitability index ranges from 0 to 1 and is calcu-
eruption. Sex was not evaluated, as none of the lated using the formula: V' = H'/logeS where S is
elements appropriate for such designation were the number of taxa for which MNI and biomass
present in the sample. were estimated.
483000 m.e.
3506000 m.n.
Back Creek Village (9Li207)
3495000 m.n.
487000 m.e.
S.C.
Georgia
magnetic
north F.L.
1 MI 0 1
4000 FT 0 4000
1 KM 0 1
Fig. 5.1. Location of Back Creek Village on St. Catherines Island, Georgia.
Hard clams grow incrementally throughout (Quitmyer, Jones, and Arnold, 1997; Andrus
the year by adding carbonate to the margins of and Crowe, 2008; O’Brien and Thomas, 2008;
each valve. In southeastern populations, an an- Quitmyer and Jones, chap. 7), it is possible to
nual cycle is represented by a translucent, slow estimate season of death for the clams recov-
growth increment laid down in the summer and ered at Back Creek Village. Andrus and Crowe
an opaque, fast growth increment laid down in (2008: 507–517) found that opaque growth
the late winter–spring (Quitmyer, Jones, and began around November or December in St.
Arnold, 1997). Using an established six-part Catherines clams and translucent growth began
sequence of growth (Jones, 1980; Quitmyer, around March. Growth is most influenced by
Jones, and Arnold, 1985; Quitmyer, Jones, and temperature, though other environmental vari-
Arnold, 1997; Quitmyer and Jones, chap. 7, this ables, such as salinity, dissolved oxygen, and
volume) and modern seasonal growth patterns food availability, can cause variations in growth
A D
B E
C
F
0 20
m
Fig. 5.2. Topographic map of Back Creek Village showing the location of the middens excavated in 2008.
patterns (Andrus and Crowe, 2008: 499–502). es (Ariidae), mullets (Mugil spp.), silver perch
In addition, there are physiological conditions (Bairdiella chrysoura), seatrout (Cynoscion spp.),
that cause variations in this pattern—most spe- spot (Leiostomus xanthurus), kingfish (Menticir-
cifically, older individuals lay down much thin- rhus spp.), Atlantic croaker (Micropogonias un-
ner increments and may be more sensitive to en- dulatus), black drum (Pogonias cromis), and red
vironmental stresses (Andrus and Crowe, 2008: drum (Sciaenops ocellatus). Standard length was
499). Looking at multiple clams from a single estimated for the most common element of each
context can help account for individual varia- taxon present in each unit.
tion within an archaeological population.
Seventy-five percent of the left valves from RESULTS
each unit were selected for analysis using a ran-
dom number table. In units with 10 or fewer left A total of 14,881 vertebrate specimens were
valves, all left valves were selected. Each valve identified, for a total weight of 2049.296 g, and a
was sectioned radially from the umbo to the total of 449 individuals. The results are present-
edge, along the margin of greatest growth, with ed by midden in tables 5.2 through 5.8. Hard-
a Buehler Isomet low-speed saw. Thick sections head catfish (Ariopsis felis), mullets, killifishes
were visually inspected to determine the final (Cyprinodontidae), snakes (Serpentes), rabbits
growth phase. Histograms comparing the per- (Sylvilagus spp.), and deer are 100% ubiquitous.
centage of specimens assigned to each growth Herrings and shad (Clupeidae), silver perch, sea-
phase were constructed for each unit and for the trout, spots, red drums, flounders, mud turtles
site as a whole, to illustrate the range and distri- (Kinosternidae), diamondback terrapin (Mala-
bution of growth phases present in each context clemys terrapin), birds (Aves), and raccoon (Pro-
Fish size estimation can be useful for deter- cyon lotor) are present in over half the middens.
mining season of death in two ways. Modern trawl Overall, fishes dominate the MNI (84%) and
data record the frequency of different size classes contribute a large portion of the biomass (37%)
of fishes caught on a monthly basis. The size of (table 5.9). Deer contribute the most biomass
archaeological fishes can then be compared to (43%), but only a small percentage of the MNI.
the modern data to estimate the time of year the Turtles and other wild mammals contribute more
fishes were captured. Trawl data are available for to the MNI than deer, though pond turtles are the
a number of localities along the Georgia Coast only taxon besides fishes and deer to contribute
(DEIS, 1978; Nelson et al., 1991) including the a significant portion of the biomass. Birds and
landward side of St. Catherines Island (Dahlberg, snakes contribute a small percentage of the MNI
1972). Back Creek Village, however, is located and biomass. Killifishes contribute the largest
on the seaward side and the modern data may not number of individuals to the fish MNI, though
be directly comparable. Fish size is generally re- sea catfishes contribute the largest portion of the
lated to fish age. If growth rates are known for fish biomass. Drums make up the next largest
the species in a region, then individuals can be portion of fish biomass, but sea catfishes, drums,
classified as juveniles, subadults, or adults. This and mullets contribute roughly similar numbers
information can then be used to estimate season- of individuals. Other fishes are not an insignifi-
ality for species in which different age cohorts cant portion of the MNI, but contribute little to
have distinct seasonal movements within or out the total biomass.
of the estuary. Richness, diversity, and equitability results are
Standard length for Back Creek Village fishes presented in figure 5.3. Richness for the whole
was reconstructed from allometric formulas (ta- site was 36, but for the middens it ranged from 14
ble 5.1) using measurements of otoliths and at- in midden G to 26 in midden D, with a mean of
lases. Growth of these elements scales allometri- 19. Diversity, estimated from MNI, ranged from
cally with fish size. Modern comparative collec- 1.89 in midden F to 2.46 for midden D, and for
tions were used to generate regression formulas the site as a whole, was 2.55. Diversity, estimated
relating otolith length or width and atlas width from biomass, ranged from 1.49 in midden F to
to standard length that were applied to the ar- 2.18 in midden C, and for the site as a whole, was
chaeological measurements (Colaninno, personal 1.81. Equitability, estimated from MNI, was 0.71
commun.; Reitz and Wing, 2008: 68). Standard for the site, and ranged from 0.68 in midden F
length was estimated for nine taxa: sea catfish- to 0.85 in midden B. Equitability, estimated from
TABLE 5.1
Fish Size Regression Formulas
Otolith Width (mm) to Standard Length (mm)
Taxon N Slope (b) Y-intercept (log a) r2
Ariidae 337 1.13 1.31 0.7
Bairdiella chrysoura 68 1.09 1.32 0.92
Otolith Length (mm) to Standard Length (mm)

Cynoscion spp. 94 1.27 0.94 0.95
Menticirrhus spp. 45 1.27 1.13 0.97
Micropogonias undulatus 65 1.19 1.06 0.95
Pogonias cromis, 75 1.238 1.05 0.91
Sciaenops ocellatus
Leiostomus xanthurus 86 1.39 1.06 0.86
Atlas Width (mm) to Standard Length (mm)

Mugil spp. 55 0.852 1.803 0.96
Sciaenidae 152 0.61 1.93 0.65
biomass, was 0.5 for the site, and ranged from fishes (fig. 5.5). Age was estimated using bio-
0.48 in midden H to 0.81 in midden G. logical studies of growth rates and size and age
A total of 331 clam valves were sectioned for at maturity for each taxon. Size was estimated for
this study, but the growth phase of the final incre- a total of 41 sea catfishes from seven middens.
ment could only be determined for 308 valves. The size ranges from 137 mm to 403 mm, with
All growth phases were represented (T1–T3 and an average of 260 mm. Based on modern data for
O1–O3) in the observed final increments (fig. sea catfishes, these are all at least in their second
5.4). For the site as a whole there is a normal dis- year of life, and most are probably mature adults
tribution around O1 growth, and the percentage (Muncy and Wingo, 1983). Size was estimated
in each phase ranges from 10% to 23%, a rela- for a total of 62 mullets from six middens. The
tively equal representation of each phase. A simi- size ranges from 88 mm to 201 mm, with an aver-
lar distribution is seen in middens C (N = 26) and age of 135 mm. These individuals are likely un-
G (N = 48). All growth phases are also present in der three years of age, and not mature (Collins,
middens D (N = 53) and A (N = 66), though these 1985). Size was estimated for a total of 17 silver
middens have a small spike in T2 growth and an- perch from five middens. The size ranges from
other in O1 or O2 growth. Midden F (N = 27) is 95 mm to 180 mm, with an average of 140 mm.
dominated by O1 growth and T2 is absent, mid- No data are available to address age. Size was
den B (N = 43) is dominated by O2 growth and estimated for a total of 20 seatrout from five mid-
T2 is absent, and midden H (N = 39) is dominated dens. The size ranges from 149 mm to 521 mm,
by O2 though T1 growth and T2 is absent. with an average of 352 mm. Seatrout can reach
Fish size was estimated for a total of 155 150 mm by the end of the first year, and even 200
TABLE 5.2
Back Creek Village, Midden A: Species List
Taxa NISP MNI % MNI Weight (g) Biomass (kg)
Actinopterygii Indeterminate bony fishes 1185 — — 18.951 0.356
Elops saurus Ladyfish 10 1 1.3 0.082 0.004
Anguilla rostrata American eel 2 1 1.3 0.050 0.003
Siluriformes Catfishes 44 — — 3.551 0.069
Ameiurus natalis Yellow bullhead 1 1 1.3 0.039 0.001
Ariidae Sea catfishes 17 — — 7.012 0.131
Ariopsis felis Hardhead catfish 27 1 1.3 4.051 0.046
Bagre marinus Gafftopsail catfish 23 2 2.6 2.788 0.055
Mugil spp. Mullets 480 10 12.8 7.427 0.171
Cyprinodontidae Killifishes 152 16 20.5 1.207 0.041
Sciaenidae Drums 83 — — 3.645 0.104
Bairdiella chrysoura Silver perch 1 1 1.3 0.006 0.001
Cynoscion spp. Seatrout 241 15 19.2 25.374 0.559
Leiostomus xanthurus Spot 1 1 1.3 0.050 0.004
Menticirrhus spp. Kingfish 3 1 1.3 0.243 0.014
Micropogonias undulatus Atlantic croaker 1 1 1.3 0.270 0.015
Sciaenops ocellatus Red drum 2 1 1.3 0.376 0.019
Paralichthyidae Flounders 8 1 1.3 0.485 0.015
Anura Frogs and toads 37 — — 0.656 —
Anaxyrus spp. North American toads 10 2 2.6 0.301 —
Scaphiopus holbrookii Eastern spadefoot toad 22 3 3.8 0.648 —
Rana sp. Bullfrog 1 1 1.3 0.010 —
Caudata Newts and salamanders 6 1 1.3 0.128 —
Testudines Indeterminate turtles 150 — — 27.224 0.362
Kinosternidae Mud turtles 10 2 2.6 2.340 0.070
Emydidae Pond turtles 15 — — 13.957 0.227
Malaclemys terrapin Diamondback terrapin 31 3 3.8 19.740 0.290
Lacertilia Indeterminate lizards 14 4 5.1 0.136 —
Anolis carolinensis Green anole 1 (2) — 0.019 —
Serpentes Indeterminate snakes 43 — — 2.443 0.034
Nerodia spp. Water snakes 14 1 1.3 2.262 0.031
Zenaida macroura Mourning dove 1 1 1.3 0.052 0.001
Passeriformes Perching birds 6 1 1.3 0.116 0.003
Mammalia Indeterminate mammals 193 — — 73.483 1.346
Scalopus aquaticus Eastern mole 1 1 1.3 0.097 0.003
Sylvilagus spp. Cottontail rabbit 4 1 1.3 0.703 0.019
Sigmodontinae New World mice and rats 11 — — 0.173 0.006
Sigmodon hispidus Hispid cotton rat 1 1 1.3 0.040 0.001
Procyon lotor Raccoon 2 1 1.3 4.153 0.096
Odocoileus virginianus White-tailed deer 29 2 2.6 95.908 1.708
Vertebrata Indeterminate vertebrates — — — 52.786 —
Total 2883 78 100 372.982 5.805
TABLE 5.3
Back Creek Village, Midden B: Species List
Clupeidae Herrings and shads 4 1 2.2 0.015 0.002
Ariidae Sea catfishes 51 7 15.6 47.294 0.789
Ariopsis felis Hardhead catfish 7 (1) — 0.903 0.018
Bagre marinus Gafftopsail catfish 142 (4) — 14.558 0.257
Orthopristis chrysoptera Pigfish 1 1 2.2 0.021 0.001
Pomatomus saltatrix Bluefish 1 1 2.2 0.021 0.001
Bairdiella/Stellifer Small drums 11 — — 0.086 0.007
Pogonias cromis Black drum 2 1 2.2 0.010 0.001
Serpentes Indeterminate snakes 59 1 2.2 0.645 0.009
Soricidae Shrews 2 1 2.2 0.014 0.001
Sylvilagus sp. Cottontail rabbit 1 1 2.2 0.158 0.005
Total 1112 45 100 130.187 2.172
TABLE 5.4
Back Creek Village, Midden C: Species List
Lepisosteus spp. Gars 15 1 1.5 0.549 0.020
Cyprinidae Carp and minnows 1 1 1.5 0.014 0.001
Lagodon rhomboides Pinfish 1 1 1.5 0.005 0.000
Bairdiella/Stellifer Small drums 12 2 3.1 0.159 0.012
Bairdiella chrysoura Silver perch 8 (1) — 0.366 0.020
Paralichthyidae Flounders 1 1 — 0.011 0.000
Nerodia spp. Water snakes 2 1 1.5 0.027 0.000
Aves Indeterminate birds 30 1 1.5 2.067 0.040
Didelphis virginiana Opossum 1 1 1.5 0.229 0.007
Sylvilagus spp. Cottontail rabbits 2 1 1.5 0.562 0.016
Sciurus sp. Squirrels 1 1 1.5 0.126 0.004
Sigmodontinae New World mice and rats 3 1 1.5 0.023 0.001
Total 1542 65 100 124.777 2.138
TABLE 5.5
Back Creek Village, Midden D: Species List
Lepisosteus spp. Gars 3 1 1.1 0.141 0.006
Anguilla rostrata American eel 4 1 1.1 0.051 0.003
Opsanus sp. Toadfish 1 1 1.1 0.023 0.001
Belonidae Needlefish 1 1 1.1 0.013 0.001
Carangidae Jacks 1 1 1.1 0.033 0.002
Bairdiella/Stellifer Small drums 2 — — 0.020 0.002
Rana sp. Bullfrog 1 1 1.1 0.033 —
Kinosternidae Mud and musk turtles 1 1 1.1 0.061 0.005
Colubridae Nonvenomous snakes 2 1 1.1 0.008 0.000
Viperidae Venomous snakes 1 1 1.1 0.012 0.000
Aves Indeterminate birds 3 1 1.1 0.666 0.015
Didelphis virginiana Opossum 1 1 1.1 0.173 0.005
Sigmodon hispidus Hispid cotton rat 1 1 1.1 0.015 0.001
Total 3170 95 100 537.243 6.203
TABLE 5.6
Back Creek Village, Midden F: Species List
Lepisosteus sp. Gars 1 1 1.6 0.126 0.005
Kinosternon spp. Mud turtles 3 1 1.6 0.276 0.013
Aves Indeterminate birds 1 — — 0.165 0.004
Rallidae Coots and rails 4 2 3.3 0.492 0.011
Sigmodontinae New World mice and rats 1 1 1.6 0.006 0.000
Total 1431 61 100 438.91 5.454
TABLE 5.7
Back Creek Village, Midden G: Species List
Ariidae Sea catfishes 36 8 20.5 7.435 0.138
Ariopsis felis Hardhead catfish 57 (4) — 4.438 0.085
Scaphiopus holbrookii Eastern spadefoot tad 5 1 2.6 0.184 —
Kinosternon spp. Mud turtles 4 1 — 1.646 0.044
Serpentes Indeterminate snakes 32 1 2.6 0.743 0.010
Total 1782 39 100 88.704 1.479

TABLE 5.8
Back Creek Village, Midden H: Species List
Archosaurgus probatocephalus Sheepshead 4 1 1.5 0.618 0.010
Lagodon rhomboides Pinfish 1 1 1.5 0.006 0.000
Cynoscion spp. Seatrouts 23 2 3.0 2.074 0.087
Menticirrhus spp. Kingfish 40 3 4.5 1.849 0.073
Kinosternidae Mud and musk turtles 11 1 1.5 0.522 0.029
Soricidae Shrews 9 1 1.5 0.247 0.007
Sciurus spp. Squirrels 3 1 1.5 1.084 0.028
Peromyscus spp. Deer mice 20 2 3.0 0.377 0.011
Total 2961 66 100 356.493 5.457
TABLE 5.9
Back Creek Village: Summary
NISP MNI % MNI Biomass (kg) % Biomass
Deer 102 9 2.0 6.09 43.2
Other wild mammals 28 15 3.3 0.32 2.3
Birds 47 8 1.8 0.07 0.5
Snakes 503 8 1.8 0.32 2.3
Pond turtles 358 15 3.3 1.90 13.5
Mud turtles 29 6 1.3 0.09 0.6
Sea catfishes 1370 55 12.2 3.44 24.4
Mullets 2043 73 16.3 0.46 3.3
Killifishes 883 115 25.6 0.13 0.9
Drums 593 61 13.6 1.11 7.8
Other fishes 137 30 6.7 0.11 0.8
Commensal taxa 1
467 54 12.0 0.07 0.5
Total 6560 449 100 14.11 100

1
Biomass not calculated for frogs, toads, lizards, and salamanders.
mm, depending on the environment. These indi- a juvenile, probably about half a year old (Sutter,
viduals probably range in age from one year to Waller, and McIlwain, 1986). Size was estimated
six years (Johnson and Seaman, 1986). Size was for two red drums from two middens. The stan-
estimated for a total of five spots from two mid- dard lengths for these two individuals were 378
dens. The size ranges from 137 mm to 187 mm, mm and 400 mm. These individuals are likely
with an average of 157 mm. Based on modern mature adults (Reagan, 1985).
data, these individuals are likely in their second
year, and most are probably not mature (Hales DISCUSSION
and Van Den Avyle, 1989). Size was estimated
for a total of two kingfish from two middens. The The vertebrate taxa recovered from Back
standard lengths for these two individuals were Creek Village are available year round in the es-
263 mm and 326 mm. There are no data avail- tuary. Seasonally available resources, such as sea
able to address age, though based on growth rates turtles, sharks, and migratory birds are complete-
for other drums, these are probably adults. Size ly absent. Terrestrial mammals, such as deer, rac-
was estimated for a total of five Atlantic croakers coon, squirrels, opossums, and rabbits, live on the
from three middens. The size range was 145 mm island, and although the mainland is close enough
to 203 mm, with an average of 180 mm. These for some exchange between populations, there is
individuals are probably in their second year, and no population-scale migration of these taxa off
likely not yet mature (Lassuy, 1983). Size was es- the island at any time of year. Modern studies of
timated for one black drum. The standard length southeastern estuarine fishes demonstrate that
for this individual was 87 mm. This individual is the most common fish taxa from Back Creek—
3.0 40
35
2.5
EQUITABILITY AND DIVERSITY
30
2.0
25
RICHNESS
1.5 20
15
1.0
10
0.5
5
0.0 0
A B C D F G H Site
MIDDEN
Richness MNI Diversity MNI Equitability Biomass Diversity Biomass Equitability
Fig. 5.3. Richness, diversity, and equitability estimates for Back Creek Village, presented for each midden
and for the site as a whole.
sea catfishes, mullets, killifishes, drums, and tributed most of the meat to the diet. Equitability
flounders—are present in the estuary throughout for MNI is quite high in all middens (over 0.68),
the year. The age cohorts and locations of each indicating that the hunting strategy targeted a
taxon may change, but the taxon itself is pres- range of taxa on a relatively equal basis. Equi-
ent all year some place in the estuary (Dahlberg, tability estimated from biomass is lower in most
1972; DEIS, 1978; Nelson et al., 1991). The less cases because deer, the largest animal identified
abundant fishes in the Back Creek collection are in the collection, dominates the biomass. In mid-
also available for at least most of the year. This dens C and G, equitability is the same for MNI
suggests a subsistence strategy that targeted con- and biomass. In midden C, this is because both
sistently available animal resources. MNI and biomass are dominated by a single tax-
The subsistence strategy also appears to have on (killifishes in the case of MNI and deer in the
been generalized, not focused on single taxa or case of biomass), and in midden G, this appears
habitats. Mammals, fishes, shellfishes, estuarine to be because no taxa dominate the MNI or bio-
and mud turtles, and terrestrial and water birds mass. The variability among middens can prob-
are all found in the same middens. Richness var- ably be accounted for by the different sample
ies among middens, but it roughly correlates with sizes recovered and by the fact that the specific
NISP for the midden, and the same common taxa species caught on any given fishing or hunting
are present in all or most middens. Diversity es- trip will vary.
timated from MNI is consistently higher than di- The fish size and age data suggest a focus on
versity estimated from biomass, suggesting that fishing during the warmer months, though size
although a range of taxa were exploited in any could only be estimated for a handful of the fish
given midden, a small number of these taxa con- individuals present at the site and the modern
0.6 0.6
Entire Site Midden A
RELATIVE FREQUENCY
RELATIVE FREQUENCY
0.5 0.5
0.4 0.4
0.3 0.3
0.2 0.2
0.1 0.1
0 0
T1 T2 T3 O1 O2 O3 T1 T2 T3 O1 O2 O3
GROWTH PHASE GROWTH PHASE
Midden B Midden C
RELATIVE FREQUENCY
RELATIVE FREQUENCY
0.6 0.6
0.5 0.5
0.4 0.4
0.3 0.3
0.2 0.2
0.1 0.1
0 0
T1 T2 T3 O1 O2 O3 T1 T2 T3 O1 O2 O3
Midden D Midden F
RELATIVE FREQUENCY
RELATIVE FREQUENCY
0.6 0.6
0.5 0.5
0.4 0.4
0.3 0.3
0.2 0.2
0.1 0.1
0 0
T1 T2 T3 O1 O2 O3 T1 T2 T3 O1 O2 O3
0.6 0.6
RELATIVE FREQUENCY
RELATIVE FREQUENCY
0.5 Midden G 0.5

Midden H
0.4 0.4
0.3 0.3
0.2 0.2
0.1 0.1
0 0
T1 T2 T3 O1 O2 O3 T1 T2 T3 O1 O2 O3

Fig. 5.4. Relative frequencies of final growth phases for hard clams (Mercenaria spp.) from Back Creek Vil-
lage, presented for the site as a whole (top left graph) and for each midden.
40
30
COUNT
20
10
0
80− 120− 160− 200− 240− 280− 320− 360− 400− 440− 480− 520−
120 160 200 240 280 320 360 400 440 480 520 560
STANDARD LENGTH (mm)

Sea catfishes Mullet Silver perch Seatrout Spot Croaker
Fig. 5.5. Fish size estimates for selected fishes from Back Creek Village, presented for the site as a whole.
data used to estimate age and seasonality are not als in middens D and H are in the range caught all
precise. The size of the black drum individual year (Dahlberg, 1972). The spots in midden G are
from midden D suggests an age of about half a in the range of those caught in the summer and
year. The black drum spawning season is in the fall in modern trawls, while those in midden G
late winter/early spring (Sutter, Waller, and McIl- are into the range of those caught all year (Dahl-
wain, 1986), which could suggest a late-summer berg, 1972). The size of the kingfish at Back
or fall occupation for midden D, though juvenile Creek is larger than those caught in the modern
black drums are present in the estuary year-round trawls, but is closest to those caught in the fall
(Nelson et al., 1991). Size was estimated for sea and early winter (Dahlberg, 1972). The Atlantic
catfishes at the family level, since otoliths of the croaker individuals from middens A, B, and D
two species can be very similar. The individuals are in or above the size range captured during the
at Back Creek are all probably over a year old. summer and early fall (Dahlberg, 1972). Adult
They are all within the size ranges of hardhead red drums, croakers, kingfish, and spots are pres-
catfish captured during the summer in modern ent in the estuary at least from April to December
trawls (Dahlberg, 1972). Size was estimated for (Nelson et al., 1991). Size was estimated for mul-
seatrout at the genus level, as otoliths and atlases lets at the genus level because the skeletal ele-
can look very similar. However, most of the indi- ments of the species found on the Georgia coast
viduals are probably spotted seatrout (Cynoscion are very similar. The mullets ranged from under
nebulosus), and all but one are probably over two a year to two years. As striped mullet juveniles
years of age. Juveniles and adults of this species and adults are present in the estuary year-round
are present in the estuary year-round (Nelson et (Dahlberg, 1972; Nelson et al., 1991), this does
al., 1991). Adults of the other species, weakfish not provide any seasonal evidence. Back Creek
(Cynoscion regalis), are present from April to is located on the seaward marsh, and the modern
December, and juveniles are present all year. Al- trawl data are from the landward marsh, so the
though there are no age estimates for the silver comparability of the two data sets is uncertain.
perch, the size of individuals from middens B, C, In addition, trawling as a method of capture may
and G are in the range, or larger, than those caught not replicate the methods or locations used by
only in summer in modern trawls. The individu- past human populations. The consistency among
the middens and among the fish taxa, however, middens were utilized during multiple seasons.
does suggest fishing was especially important Although this is not proof that people occupied
during the warmer months. the site on a permanent basis, there is no evidence
The results from clam incremental analysis they occupied it on a strictly seasonal basis. The
indicate that clams were collected during all sea- middens themselves may represent contemporary
sons at Back Creek Village. The distribution of households or different occupations by one, or a
final increments for the site as a whole is relative- few, households over time, but they do not ap-
ly even, though weighted a little toward the cold pear to be a series of seasonal occupations. Ani-
season, with a peak at O1 growth. Middens C mal taxa in a temperate estuarine environment
and G have a similar distribution. Middens A and vary in their abundance and location on a vari-
D are equally weighted toward cold and warm ety of temporal scales. Daily tides, fluctuations
season growth, with small peaks at T2 and O1 in timing and strength of tides during the year,
or O2 growth, also suggesting that clams were rainy or dry periods in the interior or on the coast,
collected in all seasons. Midden F is dominated stormy times of year, and temperature changes all
by O1 growth and may suggest more of a fall– influence resource abundance. Different taxa are
winter collection. Midden B is dominated by O2 affected by and respond to these environmental
and O3 growth, and might represent more of a changes differently. The Georgia coast, then, is
winter–spring collection. Midden H is missing not a place where seasonality of resource avail-
only T2 growth and may be a fall–spring collec- ability is simply a matter of fall, winter, spring,
tion. In all cases, except for midden B, both cold and summer. Subsistence-settlement systems,
and warm season growth are represented. This instead, exploited the constant changes in avail-
may suggest a focus on molluscs, such as clams, ability of each resource—focusing on taxa that
during the colder seasons when many other ani- are reliably present in the estuary and emphasiz-
mals are lean or less abundant and plant materi- ing those that are most abundant during a time of
als are less abundant. year, such as fishes in the summer and fall and
invertebrates in the winter and spring.
CONCLUSION
The zooarchaeological evidence is more con- NOTES

sistent with the Jones model of settlement than
with the Larson and Crook model. The late pre- 1. I gratefully acknowledge the funding of archaeologi-
historic subsistence strategy represented at Back cal research on St. Catherines Island by the Edward John
Noble Foundation and the St. Catherines Island Founda-
Creek Village relied heavily on vertebrate re- tion. I also want to acknowledge Dr. Elizabeth J. Reitz, Dr.
sources that are present in the estuary throughout David Hurst Thomas, and Irvy R. Quitmyer for their sup-
the year, suggesting that seasonal movement to port in this research, Carol Colaninno-Meeks and Siavash
exploit animal resources was not a structuring Samei for their help in the lab, Maran Little for her enor-
mous help sectioning clams, and Lori Pendleton Thomas,
principle of the settlement system. The same Elliot Blair, Anna Semon, and the rest of the 2008 St. Cath-
suite of vertebrate resources is present in all mid- erines Island archaeology staff and crew for excavating and
dens and the seasonal indicators suggest that all sorting these materials.
2012 MollusCs as Oxygen Isotope Season of Capture Proxies 123
CHAPTER 6
MollusCs as Oxygen-Isotope Season-
of-Capture Proxies in Southeastern
United States Archaeology
C. F red T. A ndrus
Oxygen isotope measurements in mollusc assessing SOC (e.g., Andrus and Crowe, 2008;
shells (fig. 6.1) are increasingly used to determine Thompson and Andrus, 2011). In these cases, a
season of capture (SOC) in southeastern U.S. ar- seasonal trend in δ18Oshell must be reconstructed
chaeological sites (e.g., Keene, 2004; Quitmyer, through ontogeny and the most recent δ18O value
Jones, and Andrus, 2005; Andrus and Crowe, (at the growing shell edge) can be assessed rela-
2008; Thompson and Andrus, 2011). The basic tive to the particular seasonal cycle experienced
rationale was initially outlined by Shackleton by the shell as it grew (for a more complete ex-
(1969; 1973, drawing from Emiliani et al., 1964). planation of one method using this approach,
The method relies on the systematic variation of see Andrus and Crowe, 2008, and Thompson
oxygen isotopes (expressed as δ18O values in parts and Andrus, 2011. This approach is depicted in
per mil: ‰) in shells during ontogeny as a func- fig. 6.2). In most habitats along the southeast-
tion of water temperature and local water δ18O. ern coast of North America (Atlantic and Gulf
In turn, local water δ18O values are a function of of Mexico), local δ18Owater is highly variable and
fresh and salt water mixing and/or evaporation. is not reliably seasonal (see data at http://cdmo.
Therefore, seasonal oscillations in shell δ18O can baruch.sc.edu/ for examples from the Atlantic
be detected in areas with consistent variation in coast, and http://www.mymobilebay.com/ for
seasonal water temperature and/or precipitation examples from the Gulf coast). Therefore, the
(assuming that these two variables do not interact more expensive and time-consuming sequential
to obscure a seasonal signal in a shell). sampling approach is required.
If only one of those two variables controls Early in the use of isotope sclerochronology,
shell δ18O in a particular habitat, then SOC it was noted that careful study of modern living
analysis is fairly simple (i.e., along desert coasts examples of each species (or close analogs) is
with almost no variation in δ18Owater or tropical necessary to ensure correct interpretation of ar-
areas with little temperature variation but a pro- chaeological SOC data (e.g., Bailey, Deith, and
nounced rainy season leading to seasonal changes Shackleton, 1983). Different mollusc species
in δ18Owater). In those comparatively rare coastal have characteristics that may render their isotope
environments, a small number of δ18O samples profiles more or less useful as seasonal proxies
near the edge of the shell can be analyzed for (i.e., if the shape of the δ18O curve becomes non-
both absolute value and trend, and then a SOC sinusoidal, seasons cannot be determined). δ18O
estimate can be made (e.g., Jones et al., 2008). values in most biogenic carbonates are influenced
Most other coastal mollusc habitats are subject by multiple variables in addition to temperature
to regular variation in temperature and irregular and ambient δ18Owater. These factors may include
variation in δ18Owater, thus complicating interpre- metabolic/kinetic effects (e.g., Owen, Kennedy,
tation of shell δ18O data. In many regions tidal and Richardson, 2002), shell mineralogy (e.g.,
variation between geographically adjacent zones Cusack et al., 2008), shell structure (e.g., Cusack
makes absolute δ18O values nearly useless for et al., 2008; Jones, 2010), variable growth rate,
Fig. 6.1. Examples of taxa described in this chapter. A, American oyster (Crassostrea virginica); B, quahog
(Mercenaria mercenaria); C, coquina (Donax variabilis); D, bay scallop (Argopecten irradians); E, ribbed mus-
sel (Geukensia demissa); F, stout razor clam (Tagelus plebeius); G, marsh clam (Rangia cuneata); H, Carolina
marsh clam (Polymesoda caroliniana); I, periwinkle (Littorina irrorata); J, knobbed whelk (Busycon carica); K,
freshwater mussel (Unionidae).
and growth cessations (e.g., Goodwin, Schöne, in the interpretation of δ18O profiles. Others may
and Dettman, 2003) to name a few. be arrhythmic (e.g., Cobb, Andrus, and Etayo-
To illustrate, consider that molluscs com- Cadvid, 2009) and may cause significant ambi-
monly cease growing for portions of each year guity in δ18O interpretations.
(e.g., Goodwin, Schöne, and Dettman, 2003; These and other complications need to be as-
Schöne, 2008; Jones et al., 2010a). If molluscs sessed for each taxon prior to making SOC es-
in a midden were collected during a prolonged timates from archaeological contexts. This is
period in which there was no shell growth, then usually accomplished by detailed growth and
the SOC estimate would be inaccurate. If the geochemical analyses of modern controlled col-
growth cessation periods are brief and do not lections of specimens. Such baseline data are
interrupt the general seasonal trends in δ18O, time-consuming and expensive to collect, thus the
then the shell isotope profile can still be a useful accumulation of SOC proxy methods for species
SOC proxy. This may, however, limit the preci- in a region may be slow. A number of useful SOC
sion of the estimates. Large numbers of growth taxa are needed to assess a wide range of human
breaks may preclude effective use of the shell subsistence practices and to reconstruct possible
as an SOC indicator by distorting the seasonal seasonal-round movements across habitats. For
δ18O oscillation to the point where it is not rec- example, it must be remembered that an SOC es-
ognizably sinusoidal. Some growth cessations timate from a particular taxon does not necessari-
are regular and linked to a particular time of the ly equate to the total season(s) of site occupation.
year (e.g., Jones et al., 2010a), and thus may aid It is plausible that some resources would only be
exploited seasonally in an otherwise permanent coastal zone. Each relevant taxon (fig. 6.1) will
occupation, perhaps due to dietary prohibitions, be discussed individually. In most cases these
comparative abundance of other food sources, taxa have not been used to measure SOC yet,
seasonal changes in taste and condition of mol- so their potential utility will be evaluated based
luscs, or a number of other factors. In such cases, on published literature and new data pertinent to
a single-taxon SOC study would yield errone- key variables such as biogeography, taphonomy,
ous conclusions. Furthermore, multispecies SOC shell structure and mineralogy, seasonal growth,
data would also be useful in assessing regional and geochemistry. This is not meant to be a com-
subsistence strategies. For example, prehistoric plete list of all molluscs found in southeastern
people in the coastal southeastern United States middens, but rather focuses on taxa that are com-
exploited a wide range of habitats (e.g., Thomas, paratively common in archaeological sites, dis-
2008). Few mollusc species are present in all of play some potential for use as SOC indicators,
these habitats. To gain a more complete picture of and cover a wide range of ecosystems relevant
regional site occupation patterns, data would be to coastal subsistence. Not all of these species
required from sites from all habitat zones. are present in St. Catherines Island middens, but
To accomplish this broader approach in the all are found within nearby coastal ecosystems.
southeastern United States, not only do we need For the purposes of this chapter, the “southeast-
an understanding of the abundant coastal marine/ ern U.S. coast” includes all tidally affected water
lower estuarine species like oyster and quahog, from the northernmost barrier islands of the Car-
but also brackish and freshwater taxa. If reliable olinas south to the semitropical zones of southern
methods can be established for multiple taxa, then Florida, and in the Gulf of Mexico coast from
widespread and more cost-effective application semitropical Florida west to Mexico.
of SOC estimates can be made, thus building a
regional understanding of subsistence practices. Oyster (Crassostrea virginica)
The oyster (see A in fig. 6.1) is the most com-
Discussion mon mollusc found in southeastern U.S. coastal
middens (for examples see Pearson, 1977; Crook,
The goal of this chapter is to review pub- 1992; Russo and Heide, 2001; Keene 2002; Quit-
lished literature pertinent to oxygen isotope SOC myer and Reitz, 2006; Thomas, 2008; Thomp-
assessment methods in the southeastern U.S. son and Andrus, 2011 ). Five studies have been
Winter
δ18O ‰
Transitional seasons
Summer
SAMPLES FOLLOWING ONTOGENY

Fig. 6.2. Idealized representation of seasonal δ18O oscillations in a mollusc through ontogeny.
published on oxygen isotope analyses in modern sonal pattern to increments in an estuary on the
populations, Kirby, Soniat, and Spero (1998) An- Gulf coast of Florida. Based on a sample of oyster
drus and Crowe (2000) and Surge, Lohmann, and shells from Georgia, Andrus and Crowe (2000)
Dettman (2001), Harding et al. (2010) and Cann- argued that while most increments seen in cross
arozzi (chap. 10, this volume), as well as three section were related to seasonal extremes, bands
isotopic SOC analyses of archaeological sites, were also associated with nonseasonal events.
Keene (2004), Harding et al. (2010), and Thomp- Kirby, Soniat, and Spero (1998) found that liga-
son and Andrus (2011). mental increments were seasonal in Mississippi
Oysters present several significant challenges and Chesapeake Bay samples. Other studies in
unique among common southeastern U.S. mid- the Chesapeake Bay area have also shown regular
den molluscs. Most obvious is their highly vari- banding (e.g., Kent, 1992; Herbert and Steponai-
able morphology. Whereas most molluscs grow tis, 1998). It may be that there are regional differ-
in generally predictable rates and shapes like ences in growth patterns. Further studies across a
Mercenaria spp. (see Quitmyer, Jones, and Ar- broad geographic range are needed to clarify this
nold, 1997; compare A and B in fig. 6.1), oysters discrepancy. As such, it is difficult to confidently
do not have a specific growth geometry (see Kent, estimate age and growth rate prior to sequential
1982). As such, oysters are not typically sampled isotope sampling, thus limiting estimates of time-
for oxygen isotopes across the valve length as is averaging and sampling resolution.
the case in most taxa, but rather are bisected along This sampling difficulty in turn may con-
the chondrophore (fig. 6.3) and then sequentially tribute in part to nonsinusoidal δ18O profiles,
sampled along the hinge area (e.g., Surge, Lohm- sometimes described as “saw-tooth” in pattern.
ann, and Dettman, 2001; Keene, 2004; Thomp- In Andrus and Crowe (2000) these patterns were
son and Andrus, 2011). likely exaggerated by the laser sampling method
Another complication is their lack of reliably whereby a single pit was ablated into each incre-
rhythmic growth increments. In cross section ment, leaving significant spaces of unanalyzed
along the chondrophore, dark and dense calcite shell in between adjacent pits, thus increasing
alternates with light and chalky calcite in what the likelihood of signal aliasing and the appear-
appear to be growth increments. Surge, Lohm- ance of sudden variation in δ18O. Likewise, the
ann, and Dettman (2001) found no regular sea- laser method and the microenvironments at that
Fig. 6.3. Bisected oyster hinge in reflected light showing irregular alternating dense gray and chalky
white banding.
collection site likely contributed to wider ampli- in the southeastern United States with respect to
tudes of seasonal oscillations and more variable oxygen isotopes and growth increments. Modern
absolute values due to comparatively poor ana- baseline incremental oxygen isotope analyses in
lytical precision and standardization. This meth- hard clams were pioneered by Jones, Arthur, and
od has since been abandoned, yet these saw-tooth Allard (1989) and Jones and Quitmyer (1996).
patterns and variability between profiles are still Other modern baseline studies followed, includ-
seen even in finely micromilled samples ana- ing Elliot et al. (2003), Walker and Surge (2006),
lyzed with high precision (see Surge, Lohmann, Surge and Walker (2006), and Andrus and Crowe
and Dettman, 2001, Keene 2002; Harding et al. (2008). Comparisons of isotope data to visual
2010; Thompson and Andrus, 2011). There may growth band analysis and modern control col-
be several related explanations for this. lections showed that, unlike some other taxa in
Periods of growth cessation are noted in oys- the region, quahog growth increments form with
ters (e.g.,Kirby, Soniat, and Spero, 1998; Surge, good seasonal regularity (fig. 6.4) and can be used
Lohmann, and Dettman, 2001), and even short to assess SOC (Clark, 1979; Quitmyer, Hale, and
gaps may lead to abrupt changes in δ18O values Jones, 1985, 1997; Arnold et al., 1991; Jones
if temperature and δ18Owater change rapidly. Also, and Quitmyer, 1996; Andrus and Crowe, 2008;
unlike most other taxa discussed here, oysters are O’Brien and Thomas, 2008; Parsons, 2008; Rus-
completely sessile, epibenthic, and frequently so and Saunders, 2008; Thompson and Andrus,
found in intertidal zones. This means that oys- 2011; Quitmyer and Jones, chap. 7; and Jones,
ters are not insulated by any sediment, have no Quitmyer, and DePratter, chap. 8 in this volume).
means of avoiding temperature and salinity ex- The two species and their hybrids are difficult to
tremes other than closure, and may be exposed to differentiate using only shell morphology; how-
diurnal and tidal temperature extremes that create ever, they can be safely treated as identical in
abrupt changes in δ18O and frequent brief growth terms of their δ18O records according to Surge
cessations. Furthermore, oysters cluster and cre- et al. (2008), who made detailed comparisons of
ate growing beds in ways that may promote the each and found no differences that would compli-
formation of changeable microenvironments. cate their use as SOC proxies.
The net result of these issues in terms of SOC These clams are arguably the best and most
estimates is that nonsinusoidal δ18O profiles may reliable sources of sclerochronological SOC data
be commonly seen and be difficult or impossible in the region. This is due to their well-studied,
to interpret. As a result, a comparatively high per- regular growth patterns, temperature-dependent
centage of analyzed shells may not yield trust- oxygen isotope fractionation, large size that per-
worthy data, thus adding uncertainty, cost, and mits easy sampling, good preservation, abun-
time to the overall site analysis (e.g., Thompson dance in many middens in the region, and wide
and Andrus, 2011). It may also lead to decreased environmental tolerance. The principal limita-
overall accuracy in SOC estimates. tions of these clams are that they are less abun-
In summary, even though several modern dant than oysters in archaeological sites in the
studies exist and midden oysters are already region, and that they do not range into brackish
in use as SOC proxies, more baseline data are and freshwater environments. Geriatric individu-
needed to better understand many key aspects of als may also present challenges in that banding at
oyster growth and geochemistry. Such additional the shell margin is very thin and may not permit
work is warranted because oysters are the domi- adequate sampling density to resolve seasonal
nant taxon in middens in the region and cover a shifts in δ18O (e.g., Andrus and Crowe, 2008;
wide range of habitats. The research presented Parsons, 2008).
in this volume by Cannarozzi (chap. 10) repre- δ18O data can be used to augment incremental
sents an important contribution to this progress (structural) SOC methods (Quitmyer, Jones, and
and adds detail and new data to the description Arnold, 1997; Andrus and Crowe, 2008; Jones,
given here. Quitmyer, and DePratter, chap. 8, this volume) in
that latitudinal (Jones and Quitmyer, 1996) and
Quahog, or Hard Clam (Mercenaria spp.) temporal (Henry and Cerratto, 2007) variation
Quahogs, or hard clams (Mercenaria merce- in increment precipitation has been noted. δ18O
naria, M. campechiensis, and hybrids [B in fig. profiles compared to increments in midden clams
6.1]), are the most studied of all midden molluscs can assess the stability of growth patterns through
Fig. 6.4. Bisected quahog valve in reflected light showing regular alternating light and dark bands.
time and space. Thus, large numbers of shells can also been made using these clams (Quitmyer,
be visually analyzed cheaply and quickly with a Jones, and Andrus, 2005; Price, 2008).
few complementary isotope profiles to build con- The small size of these clams (often 1–2 cm
fidence in the method. Application of this general along the longest axis) presents some challenges
approach is growing (e.g., Andrus and Crowe, in sampling. Micromilling an intact valve from
2008; O’Brien and Thomas, 2008; Thompson and the outer surface is the method most often used
Andrus, 2011) and will hopefully expand over (Jones, Quitmyer, and Andrus, 2004, 2005; Quit-
time. Quitmyer and Jones (chap. 7) and Jones, myer, Jones, and Andrus, 2005; Etayo-Cadavid,
Quitmyer, and DePratter (chap. 8, this volume) 2010) and this fine-scale sampling can result in
expand on these baseline data and provide more submonthly resolution. The resulting δ18O pro-
detail on the analysis of clam shells. files generally produce smooth sinusoidal curves
amenable to seasonal interpretation. In many
Coquina Clam (Donax variabilis) respects δ18O profiles in these clams represent
Coquinas are not as common in most mid- ideal SOC proxies and useful coastal marine pa-
dens as oysters and quahogs but are very abun- leoclimate proxies. The principal limitation ap-
dant in some archaeological sites (e.g., Quitmyer, pears to be their comparative rarity in sites in
Jones, and Andrus, 2005), typically in sites near the region and their restricted habitat. Additional
the clam’s littoral habitat. These are small clams research into the periodicity of increments, shell
(C in fig. 6.1) that inhabit the surf zones of open structure, and environmental tolerances may im-
beaches, living by burrowing in the sand and rid- prove this proxy.
ing waves in and out with the tides. This habitat
drew the attention of researchers because δ18Owater Bay Scallop (Argopecten irradians)
in beach environments is typically far more stable Scallops (D in fig. 6.1) are not common in all
than the habitats of other midden molluscs (Jones, ecosystems across this region; thus their appear-
Quitmyer, and Andrus, 2004). Three papers have ance in middens is somewhat localized (Russo
been published on modern and archaeological co- and Quitmyer, 1996, 2008; Withers and Hubner,
quinas (Jones, Quitmyer, and Andrus, 2004, 2005; 2009). Argopecten irradians concentricus is the
Quitmyer, Jones, and Andrus, 2005). These papers dominant subspecies found in the southeastern
took advantage of the stable δ18Owater to reconstruct U.S. Atlantic and Gulf coasts (MacKenzie, 2008).
past sea surface temperatures (Jones, Quitmyer, This scallop is extensively well studied with re-
and Andrus, 2004, 2005). SOC estimates have spect to basic biology (see MacKenzie, 2008, for
a recent review), but no published analysis of oxygen isotope distributions. Collectively, these
oxygen isotope shell data exists. Archaeologi- studies demonstrate that making isotopic esti-
cal analyses of these scallops have focused on mates of SOC is likely possible, but as with the
SOC estimates using shell length measurements bay scallops, there are several potential problems
(Russo and Quitmyer, 1996, 2008) and changes that need to be addressed before these mussels
in shell morphology (Marelli and Arnold, 2001). can be analyzed with confidence.
Examination of the shell structure and min- First among these concerns is a need to have
eralogy suggests that there may be several chal- a detailed understanding of the shell microstruc-
lenges to overcome when measuring δ18O pro- ture and mineralogy. Lécuyer, Reynard, and Mar-
files in these scallops. For example, the shell is tineau (2004) performed Raman analyses on one
composed of two calcite layers separated by an shell and detected only aragonite; however, some
aragonite layer. Such a configuration makes mi- mussels like Mytilus edulis (Vander Putten et al.,
cromilling single-mineral samples difficult. Pri- 2000) and Choromytilus chorus (Jones, 2010)
or knowledge of sample mineralogy is necessary also contain calcite. More detailed mineralogical
to generate and interpret δ18O data from bimin- and microstructural analysis is required because
eralic shells because aragonite and calcite have δ18O data from bimineralic mussels presents
differing phosphoric acid fraction factors (Kim, a similar situation as described above for bay
Mucci, and Taylor, 2007) and potentially differ- scallops. Furthermore, in some other mussels,
ent original isotopic enrichments (Cusack et al., different textures and crystal habits of the same
2008). Undetected mixing of mineral samples mineral are present across shell layers and con-
could alter the seasonal δ18O profile and compli- tain different isotopic values (Cusack et al., 2008;
cate assignment of SOC. Jones, 2010).
Additional alteration of scallop δ18O profiles Shell growth rate in this species is not uniform
can be due to seasonally episodic growth. Signifi- throughout the year in the southeastern United
cant winter mortality occurs in this species in at States (Borrero and Hilbish, 1988). Growth ces-
least part of its range after the first year of growth sations and significant changes in growth rate
(Russo and Quitmyer, 1996, 2008). This suggests over ontogeny have been seen to impact δ18O pro-
that sublethal exposure to seasonal temperature files in related taxa (e.g., Jones, 2010). Studies of
extremes may lead to diminished or paused shell modern populations of ribbed mussels are needed
growth and result in truncation of seasonal δ18O to fully assess whether growth rate impacts their
oscillations. Although these growth and structur- utility as SOC indicators. Seasonal δ18O profiles
al issues need to be addressed, sequential isotopic may also be impacted by damage to the fragile
analysis in morphologically similar scallops has valves during life (Hillard and Walters, 2009).
been successful in measuring seasonal environ- If absolute temperature reconstructions are de-
mental variation (e.g., Jones et al., 2007). sired, it may also be necessary to develop a spe-
cies-specific isotope temperature equation. Other
Ribbed Mussel (Geukensia demissa) mussels such as Mytilus edulis (Wanamaker et
Although ribbed mussels (E in fig. 6.1) are al., 2006, 2007) and Mytilus californiensis (Ford
found in southeastern U.S. shell middens (e.g., et al., 2010) have slightly different slopes and
Keene, 2002; Quitmyer and Reitz, 2006; Parsons, significantly different y-intercepts to their respec-
2008) and are abundant in regional salt marshes, tive isotope/temperature relationships. The only
no sequential oxygen isotope data have been pub- published δ18O data from this species (Lécuyer,
lished for this species. However, fairly extensive Reynard, and Martineau, 2004) lack adequate en-
geochemical, mineralogical, and microstructural vironmental monitoring to assess which, if any,
analysis has been performed on closely related published equation is suitable.
species from similar coastal environments (Jones The biggest obstacle to the use of ribbed
and Kennett, 1999; Vander Putten, 2000; Gillikin mussels for SOC analysis may be taphonomy.
et al., 2006; Wanamaker et al., 2006, 2007; Cu- Compared to oysters and quahogs, their shells
sack et al., 2008; Jones et al., 2008; Ford et al., are quite thin and fragile. They are typically ex-
2010; Jones, 2010). One study (Lécuyer, Rey- cavated in fragments that may be too small to
nard, and Martineau, 2004) analyzed bulk (no generate a seasonal sinusoidal curve with which
time series data) δ18O samples from Caribbean to compare the edge values. Because they live in
ribbed mussels and noted temperature-dependent similar habitats to the oyster and thus experience
highly variable δ18Owater, such complete sinusoids Lomovaskya, Gutiérrez, and Iribarnea (2005)
are necessary to measure SOC. note that specimens of T. plebeius commonly dis-
play evidence of shell damage and repair around
Stout Razor Clam (Tagelus plebeius) the valve edge (>70% in their samples). Similar
Razor clams (F in fig. 6.1) are often present to periods of growth cessation or diminishment,
in middens in the Southeast (e.g., Pearson, 1977; this may have significant impact on the shape of
Keene, 2002; Quitmyer and Reitz, 2006; Parsons, the seasonal δ18O profiles and should be assessed
2008), but little SOC research has been conduct- with controlled modern studies.
ed on them. No published studies exist regarding
seasonal growth in their shells or oxygen isotope Marsh Clam (Rangia cuneata
distributions. These large, robust clam valves pre- and Polymesoda caroliniana)
serve comparatively well in middens and would Unlike the other taxa thus far discussed, marsh
be technically straightforward to sample sequen- clams live in brackish to nearly fresh water envi-
tially. Morphologically similar razor clams have ronments and thus may provide insight into sub-
been analyzed for sequential isotope data in Peru- sistence activities and seasonal timing outside the
vian and Chilean middens (e.g., Carré et al., 2005, immediate coast. Rangia cuneata (G in fig. 6.1) is
2009; Jones et al., 2010b). These South American the dominant marsh clam in habitats and middens
shells produced regular, seasonal sinusoidal pat- along the Gulf of Mexico coast, and Polymesoda
terns, with abrupt summer shifts in shell δ18O caroliniana (H in fig. 6.1) is more abundant on
values indicating subseasonal warm temperature the Atlantic coast. While R. cuneata is frequently
growth cessations (Carré et al., 2005; Jones et al., the most abundant species in Gulf middens (e.g.,
2010b). The Pacific coast of South America has a Aten, 1981; Carlson, 1987), P. caroliniana is less
far narrower range in seasonal water temperatures often found in either coast.
and very little variation in δ18Owater as compared There are three published δ18O studies of
to the southeastern United States; therefore, di- R. cuneata (Lécuyer, Reynard, and Martineau,
rect analysis of modern T. plebeius is required to 2004; Andrus and Rich, 2008; Cobb, Andrus,
assess the seasonality of shell growth. and Etayo-Cadavid, 2009); however, they do
Research on modern populations in South not focus directly on measuring SOC; rather,
Carolina by Holland and Dean (1977) suggest they examine basic sclerochronological con-
that these razor clams contain annual growth cerns such as isotope fractionation and growth
bands and grow most rapidly in summer and fall rates and patterns. R. cuneata precipitates ara-
with variation between habitats and tidal zones. gonite shells in or near oxygen isotope equilib-
-2
-3
-4
δ18Ο‰
-5
-6
-7
-8
EQUIDISTANT SAMPLES FOLLOWING ONTOGENY
Fig. 6.5. Sequential δ18O data from R cuneata valve (0608.2) collected from Chocolatta Bay where it enters
Mobile Bay, AL (30°40′36″S and 87°58′21″W) collected on June 8, 2006. Ontogeny runs left to right, with
sample from valve edge on right of plot. Dashed lines divide the seasonal range in δ18O into equal thirds. δ18O
analytical precision is finer than ± 0.1‰ (based on analysis of NBS-19 average across all runs). Following the
methods outlined in figure 6.2, the SOC estimate is spring.
rium, which supports its use for SOC determina- elemental geochemistry of shell fragments in
tions (Lécuyer, Reynard, and Martineau, 2004; ceramics as a means to examine paleoenviron-
Andrus and Rich, 2008). However, it appears mental conditions, but no work focused on SOC
that these shells are prone to frequent growth measurements has been done. δ18O analysis has
cessations that are neither rhythmic nor con- been conducted in some western U.S. and Euro-
sistent within a population (Cobb, Andrus, and pean unionids (e.g., Dettman, Reische, and Lo-
Etayo-Cadavid, 2009). These growth cessations hmann, 1999; Goewert et al., 2007; Versteegh et
are often visible in the shell structure and are al., 2009, 2010). These studies suggest that while
described as partly seasonal in some locations unionid δ18O values reflect seasonal environmen-
(Fairbanks, 1963; Carlson, 1987). This irregular tal changes, there are sometimes significant peri-
growth may be a result of the highly variable ods in which shell growth stops. Long cessations
environments of upper estuaries where tempera- may limit these mussels’ utility as SOC proxies,
ture, salinity, turbidity, and other factors rapidly but further baseline studies are needed before this
and unpredictably change. These growth cessa- can be concluded.
tions may result in nonsinusoidal δ18O profiles Other challenges may impede the use of these
that complicate SOC assessments. However, species as well. These shells are very fragile and
δ18O profiles thus far analyzed (fig. 6.5) contain are often deposited in acidic soils that contribute
predictable sinusoidal oscillations. The irregular to recrystallization of their aragonite. Although
growth breaks likely contribute to subseasonal recrystallization is relatively easy to detect, oth-
abrupt changes in δ18O but may not obscure the er taphonomic concerns create additional prob-
overall seasonal trends. lems. In particular, these shells tend to fragment
In contrast, no published δ18O studied of P. and spall along growth increments, and inner
caroliniana exist, but similar growth habit and layers of the shells foliate cleanly, leaving what
environments suggest that it may share general appears to be a pristine shell that in fact is miss-
sclerochronological properties with R. cuneata ing its innermost layers. This property is com-
and other closely related clams. For example, mon enough that it has been exploited to aid in
Polymesoda radiata was measured for shell δ18O ontogenetic sampling of mussel shells (Nelson,
to assess seasonality of rainfall and SOC in mid- 1964; Sterrett and Sayville, 1974). Such break-
den samples from Pacific Mexico (Kennett and
Voorhies, 1995, 1996).
While more research may be required before
confident SOC estimates can be made from up-
per estuarine species such as these, the results
may warrant the effort. Currently, sclerochrono-
logical analysis in the southeastern United States
focuses on marine and lower estuarine species,
thus human subsistence activities in fresh and
brackish water habitats are less well understood.
This means archaeologists may be blind to a sig-
nificant portion of a seasonal subsistence round
or resource area.
Freshwater Mussels (Unionidae)

Moving further up the estuaries of the region,
marine and brackish species give way to freshwa-
ter mussels (K in fig. 6.1). These clams thus be-
come more common in archaeological sites deep- Fig. 6.6. Archaeological Unionidae clam valve.
The preservation of this specimen is typical of sites
er into the interior coastal plain. However, there
in the region. Note the spalling occurring on the
has been little research into δ18O sclerochronolo- right. Entire shell layers have been noted to foliate
gy in these taxa. Carroll, Romanek, and Paddock along growth increments, potentially leaving what
(2006) measured oxygen and hydrogen isotope appears to be an intact valve. Such taphonomic
variation in Savannah River watershed mussels issues could lead to erroneous season-of-capture
and Peacock and Seltzer (2008) examined the estimates.
age may occur during site formation or excava- sess provenance (Claassen and Sigmann, 1993).
tion, and if undetected may lead to erroneous While these other uses may complicate their in-
SOC estimates (fig. 6.6). terpretation as subsistence remains in middens,
Species-specific issues may exist as well. it magnifies the importance of their potential as
Some unionids are quite long-lived and grow seasonality indicators in that insight into trade
slowly late in life with very thin increments that and tool procurement practices may be possible.
make sequential sampling at seasonal resolution Unfortunately, little baseline data exist concern-
difficult (e.g., Schöne et al., 2004; Versteegh, ing the specifics of whelk isotope geochemistry
2009). However, sampling resolution adequate or sclerochronology.
to assess SOC has been achieved in some fresh- The work that has been done underlines the
water mussels (Goewert et al., 2007). Even if importance of baseline studies. For example,
species-specific baseline studies are conducted mark and recapture data suggest that shell growth
to address concerns of differential growth rate rates are highly variable, that some shells may not
and cessation, species identification is sometimes grow for extended periods, and that shell abra-
difficult in archaeological mussel shells. This is sion is common along the aperture edge (e.g.,
especially true of fragmentary specimens, which Eversole, Anderson, and Isely, 2008). Growth
might be expected in middens or those that have rates may also vary between sexes (Power, Sell-
eroded or broken hinge areas. ers, and Walker, 2009), and habitat (Walker et
Habitat-specific concerns exist as well. al., 2008). Unlike most bivalves, whelks in this
δ18Owater values can vary according to a number study area migrate, often with different pat-
of factors that may complicate interpretation of terns between species (e.g., Walker et al., 2008).
mussel δ18O profiles. For example, small, shad- These factors can complicate the interpretation
ed, spring-fed streams may have little differenc- of seasonal oscillations in δ18O by altering both
es in δ18Owater from season to season, but slower, the seawater temperature and the δ18Owater ex-
shallow, and exposed bodies of water subject to perienced by the snails. Furthermore, other gas-
seasonal evaporation may have wide variation tropods, notably the queen conch (Strombus gi-
in δ18Owater. gas) are offset from oxygen isotope equilibrium
In summary, freshwater mussels represent (Wefer and Killingley, 1980). Such offsets do not
one of the most complex challenges in terms of preclude SOC estimates, but suggest the need for
developing new taxa for isotopic SOC studies. species-specific baseline analyses.
However, the lack of alternative taxa in interior Periwinkles (Littorina irrorata) are common
sites may warrant the effort and provide insight at some sites (e.g., Quitmyer and Reitz, 2006), but
into links between coastal and inland subsis- no isotope sclerochronology has been performed
tence strategies. on this species. However, the calcite shells of the
closely related Littorina littorea have been stud-
Gastropods ied in both modern and archaeological contexts,
There are no published oxygen isotope studies and the resultant δ18O records show promise
of archaeological southeastern U.S. gastropods. because the periwinkle oxygen isotope fraction-
Some relatively common midden gastropods in- ation is temperature dependent and at least some
clude the whelks Busycon carica (J in fig. 6.1), populations record the full seasonal range of tem-
Busycon sinistrum, Busycon spiratus, Busycoty- peratures (Andreasson, Schmitz, and Jönsson,
pus canaliclatum, and the periwinkle, Littorina 1999; Burman and Schmitz, 2005; Burman and
irrorata (I in fig. 6.1). Gastropods have been ana- Påsse, 2008). Key questions remain concerning
lyzed for oxygen isotopes from other archaeo- the continuity of growth, shell margin damage,
logical contexts in other regions (e.g., Mannino, and other issues that need to be addressed with
Spiro, and Thomas, 2003; Mannino et al., 2007; local modern baseline data.
Colonese et al., 2009), and presumably the south-
eastern U.S. species would be similar in at least CONCLUSIONS
some respects.
Busycon whelks are particularly important in This review illustrates some of the reasons
southeastern archaeology as they were modified why the expansion of isotope sclerochronology
and traded as utilitarian objects and gorgets, and in midden analysis is slow and expensive. Each
as such their geochemistry has been used to as- potentially useful species requires fundamental
baseline studies before archaeological applica- my mind, the most important knowledge gap to
tion can be considered reliable. Even some of the address is in freshwater habitats. Unfortunately,
most extensively studied molluscs such as qua- molluscs in these environments may also present
hogs and oysters still demand further research. some of the greatest analytical challenges.
Progress toward expanding the range of SOC in- Finally, I hope this review highlights the need
dicator species will likely occur piecemeal, with to conduct comprehensive analyses of human di-
different taxa being studied according to relative etary remains. While isotope sclerochronology
abundance in key regions and their utility in an- may be too expensive and specialized for many
swering archaeological questions. archaeological purposes, simple and low-cost ap-
Currently, oysters, quahogs, and coquina proaches may yield similar conclusions. Mollusc
clams have been studied adequately enough to shells need not be relegated to the back-dirt pile
allow at least some application of SOC analyses but should be retained, identified, and at least
in the southeastern United States. Rangia clams roughly quantified. These data can be greatly
will likely soon join this list because new studies strengthened when integrated into vertebrate and
are underway. An effective means to expand the botanical analyses. By having a large repertoire
number of SOC indicator species applicable to of seasonality indicators, with different biases
the southeastern United States would be to target and sources of error, we may more confidently
those organisms having close relatives that have reconstruct not only season of exploitation, but
been used in other regions. For example, scal- also more complex questions of occupation pat-
lops, razor clams, and mussels from other parts terns, site formation processes, and overall sub-
of the world have been successfully utilized in sistence strategies. This integrated approach has
archaeological research. Building on these pre- yielded broadly significant results on St. Cathe-
existing studies may partially limit the extent rines Island over the past decades and may serve
of needed baseline research on local species. To as a template for others.
2012 Annual Incremental Shell Growth Patterns in Hard Clams 135
chapter 7
Annual Incremental Shell Growth Patterns
in Hard Clams (Mercenaria spp.) from
St. Catherines Island, Georgia: A Record
of Seasonal and Anthropogenic Impact
on Zooarchaeological Resources
I rvy R. Q uitmyer and D ouglas S. J ones
Sclerochronology is the study of physical seized on sclerochronology as a tool to address

and chemical variations in the accretion- one of the fundamental questions posed by ar-
ary hard tissues of organisms, and the chaeologists—during which seasons did people
temporal context in which they formed. occupy a given archaeological site (Weide, 1969;
Sclerochronology focuses primarily Coutts, 1970, 1975; Coutts and Higham, 1971;
upon growth patterns reflecting annual, Ham and Irvine, 1975; Koike, 1975)? The answer
monthly, fortnightly, tidal, daily, and sub- to such a question helps to define settlement pat-
daily increments of time entrained by a terns, and subsistence strategies. Such definitions
host of environmental and astronomical expand our understanding of cultural complexity
pacemakers. Familiar examples include found in the archaeological record.
daily banding in reef coral skeletons or There is an emerging body of zooarchaeo-
annual growth rings in mollusc shells. logical inquiry that is examining the intensity of
Sclerochronology is analogous to den- resource use, resource management strategies,
drochronology, the study of annual rings and natural and anthropogenic changes in the en-
in trees, and equally seeks to deduce or- vironment. Methods grounded in sclerochronol-
ganismal life history traits as well as to ogy are helping to explore whether prehistoric
reconstruct records of environmental and peoples were good ecologists, conservationists,
climatic change through space and time. both, or neither (Quitmyer, Hale, and Jones,
—1st International Sclerochronology 1985; Thomas, 1987; Quitmyer and Jones, 2000;
Conference, 2007 Marcello and Thomas, 2002; Quitmyer, 2003;
Peacock, Haag, and Warren, 2004; Hames 2007;
Historically it has been recognized that a Rick and Erlandson, 2008; Erlandson et al., 2008;
diverse number of plant and animal taxa form Szabó and Quitmyer, 2008)? The literature also
repeating growth structures in their skeletons identifies instances where there are changes in
(Rhoads and Pannella, 1970; Rhoads and Lutz, zooarchaeological assemblages that are not an-
1980; Jones and Gould, 1999). Many times these thropogenic in origin, but related to a changing
structures represent physiological responses to environment (e.g., Bailey and Craighead, 2003).
environmental stimuli that can be regarded as In southeastern North America, sclerochronol-
biorecordings of the life and times of an organ- ogy of hard clams (Mercenaria spp.) as a zooar-
ism (Jones and Quitmyer, 1996). It was not until chaeological indicator of season of resource har-
the last half of the 20th century that scientists re- vest first emerged from archaeological research
alized the potential of asking and answering natu- conducted on St. Catherines Island, Georgia (Clark,
ral history questions through the study of these 1976a, 1976b; O’Brien and Thomas, 2008). These
structures (Wells, 1963; Buddemeier, Maragos, early researchers showed that hard clams are suit-
and Knutson, 1974; Hudson et al., 1976). In con- able taxa for such work and are ubiquitous in the
cert with natural historians, zooarchaeologists shell middens. Sclerochronology of hard clams has
become an integral part of St. Catherines Island seen in cross sections of their shells using direct
archaeological research and has influenced zoo- (fig. 7.1A) or transmitted light (fig. 7.1B). Back-
archaeological research throughout southeastern lit radial thin sections of the shells show that the
North America (Claassen, 1982; Quitmyer, Hale, white increment is opaque while the gray (or
and Jones, 1985; Quitmyer, Jones, and Arnold, purple) increment is translucent, allowing the
1997; Andrus and Crowe, 2008). transmission of light (fig. 7.1B). The two incre-
This research represents the third year of ments represent distinct differences in the shell
a five-year study of modern incremental shell microstructures that are known to be periodic and
growth in St. Catherines Island hard clams (Mer- associated with physiological responses to envi-
cenaria mercenaria) that was undertaken by ronmental changes in the annual cycle (Clark,
the authors in collaboration with David Hurst 1976a; Quitmyer, Jones, and Arnold, 1997; Fritz,
Thomas, American Museum of Natural History. 2001: 60–61). We use the terms translucent and
Our first goal is to establish longitudinal growth opaque to characterize the shell growth incre-
frequency profiles of incremental shell growth ments in this chapter (figs. 7.1B and 7.2).
among living populations of St. Catherine Island We follow the methods of Quitmyer, Jones,
hard clams. We seek to assemble a robust sam- and Arnold (1997) to establish a chronology of
ple that is systematically collected to document annual incremental shell formation in living St.
the annual periodicity of incremental shell for- Catherines Island hard clams. Approximately
mation. These data augment earlier research by 40 living hard clams were collected monthly for
O’Brien and Thomas (2008) that started during one year (2007–2008) from a shallow tidal creek
the mid to late 1970s and more recent research in King New Ground that flows into McQueen
conducted by Andrus and Crowe (2008). We ap- Creek. The specimens were collected by hand
ply the modern model of incremental shell growth from sandy mud and shell substrate around the
to zooarchaeological hard clams excavated from base of oyster bars. This represents typical hard
the Late Archaic at St. Catherines Shell Ring (ca. clam habitat as reported in the literature (Walker
4500 b.p.), and McQueen Shell Ring (4450 b.p.) and Tenore, 1984; Fegley, 2001). The specimens
(general levels and “floor” feature). We note that were transported to the St. Catherines Island
the McQueen “floor” archaeological feature is compound where they were quickly frozen in a
constructed of a sand base covered with a modest commercial grade freezer. Following transport
number of hard clam shells (Thomas, personal back to the Florida Museum of Natural History,
commun.). Relative to the general levels of the the specimens were thawed and eviscerated. The
two St. Catherines Island Archaic period sites, valves were washed, dried, numbered, and stored
the “floor” feature represents a distinct period for later analysis.
in time at McQueen. We also include previously Historical sea surface temperature (SST) and
published data from nearby Cannon’s Point Shell salinity (PSU) data were obtained from the Na-
Ring (ca. 4500 b.p.), St. Simons Island, GA (Mar- tional Estuarine Research Reserve System moni-
rinan, 1975; Quitmyer, Hale, and Jones, 1985). toring station (SAQG1) at nearby Sapelo Island,
Our second research objective is to examine Georgia. These data represent a close approxima-
the population dynamics of the three Archaic tion of estuarine SST and salinity for the region
period assemblages. We apply commonly used during 2008.
fisheries statistics that help to evaluate and man- The size and age data provide significant in-
age living hard clam populations by quantifying formation about the natural history of hard clams
their size classes, survivorship, and ontogenetic and the interpretation of their use by Native
age (Fegley, 2001). In so doing we are able to Americans (see Jones, Quitmyer, and DePratter,
assess relative harvesting pressure during the chap. 8, this volume). The anterior to posterior
Archaic period on St. Catherines Island and at length of the right valve of each specimen was
Cannon’s Point. measured using digital calipers attached to a per-
sonal computer (Jones, Quitmyer, and DePratter,
METHODS chap. 8).
A radial cross section of the right valve, from
Modern Hard Clams the umbo to the ventral margin, was facilitated
Hard clams form alternating white and gray with a water-cooled lapidary saw equipped with
(or purple) shell growth increments that may be a Mark V alumina oxide blade (Quitmyer, Jones,
Fig. 7.1. Radial cross section of a hard clam (Mercenaria campechiensis) to expose its growth increments
(A) and a backlit thin section (B) showing the alternating opaque and translucent shell growth increments.
Translucent increment
Annual growth cycle
Opaque increment
Outer shell layer
Middle shell layer
Inner shell layer

Umbo
0 5
Ventral
margin CM
Fig. 7.2. The location of the translucent and opaque shell growth structures showing a count of the ontoge-
netic age (years) of a specimen.
Radial CS,
hinge plate Radial CS, Description of
Phase margin ventral margin growth phases
T1 T just forming on margin

Translucent
Opaque
T2 T ½ complete
T3 T complete
O1 O just forming on margin
O2 O ½ complete
O3 O complete
Abbreviations: CS = cross section; T = translucent shell

increment; O = opaque shell increment.
Fig. 7.3. Six-part division of translucent and opaque annual shell growth increments in hard clam (Merce-
naria spp.) shells.
and Arnold, 1997). The blade produced a polished histogram for each of the four seasons. This sim-
surface where the incremental growth structures ple, straightforward technique accurately charac-
could be evaluated with the unaided eye and/or terizes seasonal incremental shell growth (Claas-
microscopically (10–20×) with transmitted light sen, 1990; Quitmyer, Jones, and Arnold, 1997).
(Rhoads and Lutz, 1980; Quitmyer, Hale, and The zooarchaeological samples were prepared
Jones, 1985; Ropes, 1987; Jones et al., 1990: and analyzed in accordance with the methods
217; Quitmyer and Jones, 1992). outlined for the modern comparative collection
It is generally agreed that hard clams living in of hard clams. The St. Catherines Island modern
the southeastern part of their range form a cou- analog of incremental shell growth was used to
plet of opaque and translucent shell growth incre- evaluate the zooarchaeological assemblages.
ments each year (fig. 7.2; Jones and Quitmyer, We note that zooarchaeological season of har-
1996; Quitmyer, Jones, and Arnold, 1997; Fritz, vest does not rest on the shell growth of a single
2001; Andrus and Crowe, 2008). A count of the hard clam, but on a population approach where
paired increments provides a direct method of as- all individuals contribute to a seasonal pattern of
sessing the longevity of hard clam assemblages incremental shell growth. The growth frequency
(Jones and Quitmyer, 1996; Fegley, 2001; Jones, profiles constructed from our 2007–2008 longi-
Quitmyer, and DePratter, chap. 8, this volume). tudinal study are used to evaluate zooarchaeo-
At a finer seasonal scale, the size of opaque or logical hard clams excavated from St. Catherines
translucent shell increments at the growing mar- Shell Ring, McQueen Shell Ring (general levels
gin of the shell may be used to estimate the sea- and “floor” feature), and Cannon’s Point Shell
son in which a hard clam was harvested (Jones Ring, St. Simons Island, Georgia.
and Quitmyer, 1996; Quitmyer, Jones, and Ar-
nold, 1997). A six-part subdivision of the annual RESULTS
shell growth cycle was used to establish a season-
al growth model in order to resolve the season of Modern Cycle
harvest of the zooarchaeological hard clams (fig. of Incremental Shell Growth
7.3). The formation of the translucent growth in- Our results are based on a numerically robust
crement was divided into three stages or phases: sample of a population of modern St. Catherines
(a) translucent 1 (T1)—translucent increment just Island hard clams. Because the various phases of
starting to form; (b) translucent 2 (T2)—translu- individual incremental shell formation may be
cent increment approximately one-half the size observed in most months of the year, an individ-
of the previous year’s translucent increment; (c) ual’s shell growth does not describe the overall
translucent 3 (T3)—translucent increment equal pattern of growth in a modern or zooarchaeologi-
to or greater than the previous translucent incre- cal population. It is tempting to use these proxy
ment. The formation of the opaque growth in- data to interpret the zooarchaeological shells at
crement was similarly divided into three growth monthly or subseasonal levels (e.g., late spring);
phases: (a) opaque 1 (O1)—opaque increment however, hard clam populations are responsive
just starting to form; (b) opaque 2 (O2)—opaque to various environmental stimuli that are variable
increment approximately one-half the size of the across space and time. Such is the nature of most
previous year’s opaque increment; (c) opaque 3 biological populations.
(O3)—increment nearly completed, almost equal Figure 7.4 presents a sclerochronologi-
in size to the previous year’s opaque increment. It cal comparison of opaque and translucent shell
should be noted that early in ontogeny the T3 and growth through an annual cycle. This is simply
O3 growth phases are relatively more common. meant to show the monthly progression of the
With increasing age it becomes increasingly less two shell growth increments in the population.
likely that T3 and O3 phases would exceed those Translucent shell growth is present in at least
of the previous year. Nonetheless, this provides some of the individuals during every month; it
a temporal profile of incremental shell growth in becomes the dominant increment during the
the hard clam population. summer and fall seasons. Opaque shell growth
The frequency of individuals in each of the is most prevalent during the winter and extends
various phases of incremental shell growth was into the spring. This pattern correlates with previ-
calculated for each month and each season, to ous studies in southeastern North America, such
document an annual pattern that is presented as a as Kings Bay, Georgia (fig. 7.4; Quitmyer, Jones,
100 = translucent
N = 194
= opaque
50
PERCENT (%)
50
SCI Modern Mercenaria
2007−2008
100
D J F M A M J J A S O N
MONTH
100
N = 65
50
PERCENT (%)
50
Kings Bay Mercenaria
1983−1984
100
MONTH
35
30
TEMPERATURE (°C)
25
20
15
10 SST Mean
PSU Mean NOAA SAQG1−
5
Georgia
0
MONTH
Fig. 7.4. The relationship of sea surface temperature (°C) and monthly formation of translucent and opaque
growth in modern hard clams collected live from St. Catherines Island and Kings Bay.
and Arnold, 1997; Fritz, 2001; Henry and Cer- data from the four sites show that the majority of
rato, 2007) and Litchfield Beach, South Carolina the zooarchaeological specimens were forming
(Jones, Quitmyer, and DePratter, chap. 8, this opaque shell growth (figs. 7.5 and 7.6). The best
volume). Previous researchers have reported that fit with the modern growth frequency profiles
the sequential formation of opaque and translu- suggests the zooarchaeological hard clams were
cent shell is a physiological response to seasonal harvested in late winter and spring (fig. 7.5). Rel-
changes in temperature or factors related to tem- ative to St. Catherines and McQueen, the greater
perature (Ansell, 1968). Temperatures ranging frequency of the T3 shell growth phase in the
between 15° C and 20° C are regarded as optimal: Cannon’s Point assemblage suggests a harvest
the range in which the animal attains maximum period that was more active during the winter,
growth rates (Fegley, 2001). Near the thermal albeit still in the winter/spring seasons. This is
maximum and minimum, 31° C and 7° C, respec- the time when the modern specimens have tran-
tively, growth abates, then stops. A similar rela- sitioned out of T3 shell growth phase and un-
tionship between temperature and shell growth dergo rapid opaque shell formation (fig. 7.4). It
exists in the St. Catherines population (fig. 7.4). is also the time when the population begins their
Opaque shell growth is most prevalent when SST major spawning event.
is at the animal’s thermal optimum during the
late winter and spring (fig. 7.4). As SST reaches Population Dynamics—Modern
the thermal maximum (31° C) for the organism, Ontogenetic Age and Survivorship
translucent shell growth dominates the seasonal As we note above, a combination of one
profile. In fact, Quitmyer, Jones, and Arnold opaque and one translucent or opaque incre-
(1997) observed this same shell growth profile in ment represents one year of life (fig. 7.2). Thus,
five modern populations occurring along the At- a count of the translucent growth increments
lantic and Gulf coasts between Litchfield Beach, yields an accurate account of the ontogenetic age
South Carolina, and Cedar Key, Florida (also see (years) of each individual. The question arises,
Andrus and Crowe, 2008). why go to the trouble of sectioning and counting
At the seasonal scale, 34.8% of the St. Cath- the growth increments to assess hard clam popu-
erines Island population is forming T3 shell lation dynamics? Why not use size as a proxy for
growth in the winter, while around 65% is in the age and simply measure each specimen? Figure
O1 (17.4%), O2 (21.7%), and O3 (26.1%) shell 7.7 presents a plot of shell length, anterior to pos-
growth phases (fig. 7.5). In the spring season T3 terior (mm), against ontogenetic age (years) for
shell growth is observed in 30% of the speci- the modern St. Catherines and zooarchaeological
mens, while 70% of the individuals are forming McQueen hard clams. As we have observed else-
the opaque growth increments: O1, 8.3%; O2, where (Jones, Quitmyer, and DePratter, chap. 8,
31.7%; and O3, 30%. By the summer season 87% this volume) there is considerable overlap in size
of the population has reached T3 shell growth versus age relationships in hard clams (see also
and a similar pattern extends into the fall season. Walker and Tenore, 1984; Fegley, 2001). For ex-
Contrary to previous studies where the six-part ample, the shell lengths of a six-year-old modern
subdivision of seasonal growth has been used to cohort overlap those of 12-year-old clams (fig.
distinguish all four seasons, only two halves of 7.7). Further, a 30-year-old clam can be the same
the year can be confidently identified in the St. size as a three-year-old clam. Researchers have
Catherines population: winter/spring and sum- found that the growth of same-aged individuals
mer/fall (fig. 7.5). may be affected by genetic and environmental
variables (Fegley, 2001). Environmental vari-
Zooarchaeological Hard Clams— ables may include water quality, substrate type,
Season of Resource Procurement and predator-prey relationships. At best, size is a
The modern SCI six-part subdivision of in- coarse proxy for age and direct age-determined
cremental shell growth (fig. 7.5) is used as a models provide greater confidence in assessing
model to evaluate the zooarchaeological season hard clam population dynamics (fig 7.7; Fegley,
of hard clam collection from St. Catherines Shell 2001). Because of the overlap in size and age ob-
Ring (N = 117), McQueen Shell Ring (general served in the samples, the use of population age
levels N = 68; “floor” feature, N = 54), and Can- dynamics dispenses with the possibility of mod-
non’s Point Shell Ring (N = 35) (fig. 7.6). The ern or zooarchaeological size-specific collection
inefficiency that may occur during the course of populations can expect to live at least 10 years.
harvesting (Haddon, 2001). There is also great similarity in the mean ages of
Survivorship curves are commonly used in the St. Catherines Island (9.4 years) and Kings
fisheries biology to examine the health of animal Bay (10.4 years) populations (fig. 7.9). In fact, a
populations (Fegley, 2001). The percentage of plot of the 95% confidence intervals around the
hard clams surviving each year for living popu- means overlap and indicate that the two sample
lations collected from Kings Bay, and St. Cath- assemblages could have been derived from the
erines Island, Georgia, serve as an example (fig. same population (fig. 7.9). These data also con-
7.8; table 7.1). Like many other hard clam popu- firm previous research showing that hard clam
lations living along the Georgia coast, the Kings beds are typically dominated by older, larger in-
Bay assemblage may be considered unimpacted dividuals (Walker and Tenore, 1984).
or in equilibrium (Walker, 1989). The Kings Bay
specimens were sampled (1983–1984) from a Population Dynamics—Zooarchaeological
military reserve where harvesting was prohibited Mean Age and Survivorship
for over 40 years. The Kings Bay and St. Cath- The curves for the zooarchaeological speci-
erines Island hard clam survivorship curves show mens from the St. Catherines Shell Ring and Mc-
great similarity with a gradual loss of individu- Queen Shell Ring show a dramatic loss of indi-
als with time. Between 31% and 40% of the two viduals during the first four years of life (fig. 7.8).
100 100
N = 23 N = 60
Winter Spring
80 80
PERCENT (%)
PERCENT (%)
60 60
40 34.8 40 31.7
26.1 30.0 30.0
21.7
17.4
20 20
8.3
0.0 0.0 0.0 0.0
0 0
T1 T2 T3 O1 O2 O3 T1 T2 T3 O1 O2 O3
GROWTH FREQUENCY PHASES GROWTH FREQUENCY PHASES
100 100
87.3 N = 55 87.5 N = 56
Summer Fall
80 80
PERCENT (%)
PERCENT (%)
60 60
40 40
20 20 8.9
5.5 7.3
0.0 0.0 0.0 0.0 1.8 0.0 1.8
0 0
T1 T2 T3 O1 O2 O3 T1 T2 T3 O1 O2 O3
GROWTH FREQUENCY PHASES GROWTH FREQUENCY PHASES
Fig. 7.5. Seasonal growth frequency patterns of modern hard clams (Mercenaria) collected live from St.
Catherines Island, 2007–2008.
100 100
McQueen Shell Ring N = 69 St. Catherines N = 116
80 Floor Deposit 80 Shell Ring
PERCENT (%)
PERCENT (%)
60 60 57.8
40 38.9 40
24.1
20.4 17.2
20 20 12.1
9.3 5.6 9.5
1.9 1.7 1.7
0 0
T1 T2 T3 O1 O2 O3 T1 T2 T3 O1 O2 O3
SHELL GROWTH PHASES SHELL GROWTH PHASES
100 100
McQueen Shell Ring N = 54 N = 35
Cannon’s Point
80 80 Shell Ring
PERCENT (%)
PERCENT (%)
60 60
49.3
40 40 36.7
30.4
26.7 26.7
20 11.6 20
1.4 4.3 2.9 3.3 3.3 3.3
0 0
T1 T2 T3 O1 O2 O3 T1 T2 T3 O1 O2 O3
SHELL GROWTH PHASES SHELL GROWTH PHASES
Fig. 7.6. Growth frequency profiles of zooarchaeological hard clams collected from St. Catherines Island and
Cannon’s Point, representing the Archaic period.
In contrast to the two modern populations, fewer zooarchaeological record can be interpreted.
than 11% of the zooarchaeological hard clams at- Zooarchaeologists sometimes erroneously refer
tained 10 years of age (fig. 7.8). The mean on- to these as site seasonality studies; but, seasonal
togenetic ages of all three St. Catherines Island site occupation and seasonal patterns of resource
zooarchaeological samples is around three years procurement (e.g., shellfish collection) represent
(fig. 7.9). The 95% confidence intervals around two different kinds of human behavior (Deith,
the means all overlap, indicating that they could 1983: 423). Unless a single taxon represents all
have been collected from the same living popula- seasons of resource procurement, site seasonal-
tion (fig. 7.9). The mean age of the St. Simons ity cannot be confidently determined. Repeated
shells is six years and the 95% confidence inter- sampling and analysis of multiple indicators of
val does not overlap with the modern or St. Cath- resource procurement may provide a more con-
erines Island assemblages (fig. 7.9). fident indication of seasonal or year-round site
occupation (Quitmyer, Jones, and Arnold, 1997).
DISCUSSION This research is intended to extend our knowl-
edge of the sclerochronology of the hard clam
In the archaeological literature the term sea- as one tool that may be used to interpret human
sonality refers to questions of variability in the interaction with the environment.
seasonal round of settlement behavior. Such The application of the six-part subdivision
studies are in their infancy (Russo and Quit- of incremental shell growth in monthly collec-
myer, 1996) and they are typically biologically tions of St. Catherines Island hard clams does
based. As such, they require long-term field and not provide unique profiles of shell growth that
laboratory analysis of modern taxa before the can be used to identify all of the four seasons of
120
100
SHELL LENGTH (A/P) (MM)
80
60
40
20 Modern St. Catherines Island, N = 217
Archaic McQueen Shell Ring, N = 186

0
0 2 4 6 8 10 12 14 16 18 20 22 24 26 28 30
ONTOGENETIC AGE (YEARS)

Fig. 7.7. A comparison of shell length (mm) to ontogenetic age (years) of modern and zooarchaeological hard
clams from St. Catherines Island.
100
St. Catherines Island, N = 221
80 Kings Bay, N = 65
Cannon’s Point Shell Ring, N = 30
St. Catherines Shell Ring, N =117

60
PERCENT (%)
McQueen Shell Ring Floor, N = 54
McQueen Shell Ring, N = 69

40
20
0
1 3 5 7 9 11 13 15 17 19 21 23 25 27 29
Fig. 7.8. Survivorship curves characterizing modern and zooarchaeological hard clams from St. Catherines
Island, Cannon’s Point, and Kings Bay.
the year (fig. 7.5). However, the modern analog These represent a diverse number of archaeologi-
provides a basis to distinguish winter/spring from cal cultures that date between 550 and 4500 b.p.
summer/fall collection of zooarchaeological hard We do not attempt to explain this long period of
clams. When the St. Catherines Island modern preferentially harvesting hard clams during the
profile of seasonal shell growth is applied to the winter and spring seasons in this chapter. This is
zooarchaeological assemblages from St. Cath- a subject for future inquiry. However, it is clear
erines Shell Ring, McQueen Shell Ring general that during the Archaic period on St. Catherines
levels, McQueen Shell Ring floor deposit, and Island and at Cannon’s Point, hard clams were in-
from Cannon’s Point Shell Ring, a clear winter/ tensively harvested during the winter and spring
spring pattern of hard clam harvest is evident. seasons.
In fact, all four of the zooarchaeological growth Here we question how the population dynam-
frequency profiles have elevated levels of ics of the Archaic period hard clams might have
Opaque 1 growth phase (fig. 7.6) relative to the been affected by intensive harvesting pressure
other growth phases. This represents the onset of as seen through their biology and ecology. First,
rapid shell growth during the winter and spring we note that the modern Kings Bay population
seasons where water temperature approaches is at or very near equilibrium because that lo-
optimal (15° C–20° C) (Ansell, 1968). Such a cality has been closed to harvesting for several
pattern suggests an intensive period of harvest as decades. Although technically the St. Catherines
opaque shell growth forms. population is not closed to harvest, it experienc-
This winter/spring pattern of hard clam pro- es negligible recreational clamming as a private
curement is prevalent across space and time for island and as a consequence has a similar age
zooarchaeological assemblages along the south- class frequency distribution when compared to
eastern Atlantic coast. Nearly half of the 26 sites the Kings Bay assemblage. In fact, many beds
(48%) reported by Quitmyer, Jones, and Arnold throughout coastal Georgia have experienced
(1997) contained hard clams that were specifi- very little harvesting and have some common-
cally harvested during the winter/spring seasons. ality with our two modern populations (Walker,
13.5
11.5
9.5
YEARS
7.5
Modern Zooarchaeological
5.5
3.5
N = 221 N = 65 N = 35 N = 117 N = 54 N = 69
1.5
St. Kings St. St. McQueen McQueen
Catherines Bay Simons Catherines Floor General
Fig. 7.9. Mean age (years) and 95% confidence intervals plotted for modern and zooarchaeological hard
clams from St. Catherines Island, Cannon’s Point, and Kings Bay.
TABLE 7.1
Modern and Zooarchaeological Survivorship Data of Mercenaria spp. Sampled 146
from St. Catherines Island, St. Simons Island, and Kings Bay, Georgia
Ontogenetic age (years) 1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 24 25 26 27 28 29 30
Modern St. Catherines Island

(N = 221)
Total number of deaths per
year 0 2 1 15 19 26 21 27 23 18 13 14 7 6 9 4 5 3 2 5 0 0 0 0 0 0 0 0 0 1
Total number remaining in
each age group 221 219 218 203 184 158 137 110 87 69 56 42 35 29 20 16 11 8 6 1 1 1 1 1 1 1 1 1 1 0
Percent of individuals
remaining in each age group 100 99.1 98.6 91.9 83.3 71.5 62.0 49.8 39.4 31.2 25.3 19.0 15.8 13.1 9.1 7.2 5.0 3.6 2.7 0.5 0.5 0.5 0.5 0.5 0.5 0.5 0.5 0.5 0.5 0.0
Modern Kings Bay (N = 65)

year 1 2 2 7 7 4 4 4 3 5 4 0 1 2 3 5 1 1 1 3 1 1 0 1 1 0 1
each age group 64 62 60 53 46 42 38 34 31 26 22 22 21 19 16 11 10 9 8 5 4 3 3 2 1 1 0 0 0 0
remaining in each age group 98.5 95.4 92.3 81.5 70.8 64.6 58.5 52.3 47.7 40.0 33.8 33.8 32.3 29.2 24.6 16.9 15.4 13.8 12.3 7.7 6.2 4.6 4.6 3.1 1.5 1.5
Archaic Period, St. Simons

(N = 35)
year 0 2 6 5 8 2 1 4 0 3 0 1 0 1 0 0 0 0 0 1 0 0 0 0 0 0 0 0 0 1
each age group 35 33 27 22 14 12 11 7 7 4 4 3 3 2 2 2 2 2 2 1 1 1 1 1 1 1 1 1 1 0
remaining in each age group 100 94.3 77.1 62.9 40.0 34.3 31.4 20.0 20.0 11.4 11.4 8.6 8.6 5.7 5.7 5.7 5.7 5.7 5.7 2.9 2.9 2.9 2.9 2.9 2.9 2.9 2.9 2.9 2.9
Archaic, St. Catherines Shell

Ring (N = 117)
year 6 36 36 15 13 5 1 1 1 0 2 0 0 0 0 0 0 0 1 0 0 0 0 0 0 0 0 0 0 0
each age group 111 75 39 24 11 6 5 4 3 3 1 1 1 1 1 1 1 1 0 0 0 0 0 0 0 0 0 0 0 0

remaining in each age group 94.9 64.1 33.3 20.5 9.4 5.1 4.3 3.4 2.6 2.6 0.9 0.9 0.9 0.9 0.9 0.9 0.9 0.9
NO. 97
2012

Ontogenetic age (years) 1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 24 25 26 27 28 29 30
Archaic, McQueen Shell

Ring floor deposit (N = 54)
year 5 22 9 8 5 1 2 1 1 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0
each age group 49 27 18 10 5 4 2 1 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0
remaining in each age group 90.7 50.0 33.3 18.5 9.3 7.4 3.7 1.9
Archaic, McQueen Shell

Ring general levels (N = 69)
year 4 34 18 8 3 0 1 0 1 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0
each age group 65 31 13 5 2 2 1 1 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0
remaining in each age group 94.2 44.9 18.8 7.2 2.9 2.9 1.4 1.4 0.0 0.0
Annual Incremental Shell Growth Patterns in Hard Clams
147
1989; Fegley, 2001). For these reasons our two individuals to replenish the population. Size-
modern populations serve as baselines when we class selection does not appear to be the strategy
assess the age frequency of the four zooarchaeo- that was used in the harvest of hard clams at St.
logical assemblages. Catherines Island. For example, the McQueen
Our age-based analysis of the three St. Cath- hard clams are dominated by individuals that
erines Island zooarchaeological populations are are between two and six years old (figs. 7.7 and
all dominated by individuals that are between 7.8), but their sizes exceed the maximum and
two and six years old (fig. 7.7). In contrast to our minimum of all age classes, up to 30 years old
two modern baselines and previously published in the modern St. Catherines assemblage (fig.
biological data for the Georgia coast (Fegley, 7.7). Again, there is a great deal of overlap in the
2001), the zooarchaeological hard clams are physical size of individuals within a population
nearly devoid of specimens older than six years of multiple ages. Protracted and intensive har-
(fig. 7.7). In fact, specimens from the two modern vest would have eventually removed the older
baseline populations average around 10 years old individuals in the population.
while the St. Catherines Island zooarchaeological
specimens are approximately three years old (fig. CONCLUSION
7.9). The zooarchaeological survivorship curves
are representative of a population where there is On St. Catherines Island, the timing of mod-
great mortality in early in life with very few indi- ern hard clam incremental shell growth correlates
viduals reaching full longevity (fig. 7.8; Deevey, with the winter/spring (opaque growth) and sum-
1947). Such a pattern is not typical of hard clam mer/fall (translucent growth) seasons and may
populations in southeastern North America that be used to effectively assess the season of hard
may be considered in equilibrium. These data clam harvest during the Archaic period. Zooar-
suggest a population where intensive harvesting chaeological hard clams from Cannon’s Point,
has occurred. St. Catherines Shell Ring, and McQueen Shell
As we have outlined in another chapter in Ring were collected during the winter/spring sea-
this volume (Jones, Quitmyer, and DePratter, sons when opaque growth was at its highest fre-
chap. 8), the biology and ecology of hard clams quency. This is a pervasive pattern that has been
make them susceptible to measurable harvesting documented among many cultural groups living
pressure. In southeastern North America, most along the southeastern coast of North America.
populations occur in densely circumscribed beds At this level of research, we cannot account for
that are easily collected from their preferred the choice to collect or harvest during the winter/
habitats in shallow tidal creeks. These beds are spring seasons.
usually dominated by older, larger specimens The life curves and ontogenetic age data pro-
that are easily located by hand collection. Once vide evidence for intensive exploitation of hard
the larger specimens are removed, the smaller clams from McQueen and Cannon’s Point shell
individuals become more susceptible to natural rings. All of the zooarchaeological assemblages
predation. Intensive harvest of hard clams during represent populations where large, older speci-
their peak spawning period (spring) would have mens have been removed from the populations
further deleterious impact on the beds. and there is high mortality early in life. This pat-
In many cases, humans manage their subsis- tern is not characteristic of most southeastern
tence resources by limiting the sizes of animals hard clam populations that are in equilibrium.
that they hunt and gather (Kraeuter and Cast- These data are indicative of intensive harvest of
agna, 2001), thus leaving mature and productive the resource.
2012 Validation of Annual Shell Increments and Shifting Population 149
CHAPTER 8
Validation of Annual Shell Increments and
Shifting Population Dynamics in Modern and Zoo-
archaeological Hard Clams (Mercenaria mercenaria)
from the Litchfield Beach Region, South Carolina
D ouglas S. J ones , I rvy R. Q uitmyer ,

and C hester B. D e P ratter
Hard clams (Mercenaria spp.) form an alter- sociated with ambient water temperature. Ansell
nating pattern of opaque (light) and translucent (1968) demonstrated that optimal shell growth
(dark) growth increments that can be seen in ra- occurs at ~20° C and growth slows as water tem-
dial cross sections of their shells (fig. 8.1). These peratures approach 9° C or 32° C. The seasonal
increments are thought to represent an annual timing of translucent and opaque incremental
cycle of growth and reflect changes in shell mi- shell formation is also known to change across
crostructure and chemistry associated with physi- the geographic range of the animal from the Gulf
ological responses to variations in water temper- of St. Lawrence to the Bay of Campeche (Fritz,
ature or factors related to temperature (Ansell, 2001). In the southern part of its range (e.g.,
1968; Jones and Quitmyer, 1996). South Carolina to Florida) hard clams form the
Biologists, paleobiologists, and zooarchaeol- translucent increment during the summer and fall,
ogists alike have documented the annual period- while opaque shell growth occurs during the late
icity of incremental shell growth in Mercenaria winter and early spring. In the northern latitudes
mercenaria populations from the Atlantic coast of (e.g., mid-Atlantic to New England) this pattern
North America (Jones, Arthur, and Allard, 1989; is reversed—translucent increments form in the
Quitmyer, Jones, and Arnold, 1997; Fritz, 2001). winter and opaque in the summer. Although the
These seasonally formed shell increments have pattern of shell increment differs between north-
been used to estimate the season of hard clam ern and southern populations, the preponderance
procurement in archaeological sites—zooar- of baseline research demonstrates that a couplet
chaeological seasonality (Clark, 1979; Quitmyer, of one opaque plus one translucent increment
Jones, and Arnold, 1997; O’Brien and Thomas, comprises one year of life.
2008). Less commonly attempted in zooarchaeol- Within the last decade, some exceptions to
ogy are analyses of ancient hard clam population these observations have been reported (Fritz,
dynamics based on annual shell increments. This 2001). In a zone between New Jersey and Virgin-
approach provides an empirical way to assess ia researchers discovered hard clam populations
potential anthropogenic effects of hard clam har- that deposit four increments during an annual pe-
vest on natural populations (Quitmyer, Hale, and riod during certain years—one opaque increment
Jones, 1985; Quitmyer and Jones, 2000). Such in the spring and another in the fall, plus one
techniques are widely used to monitor the effects translucent increment in the summer and another
of harvesting on modern shellfish beds (Deevey, in the winter (Fritz, 2001). The conventional wis-
1947; Hallam, 1972; Fegley, 2001). In this inves- dom of the shell growth literature would seem to
tigation we explore whether the natural distribu- indicate that annual increment formation in hard
tion of ontogenetic age classes was disrupted by clams could be used to assess the population dy-
pre-Hispanic harvesting pressure on hard clam namics and season of harvest of zooarchaeologi-
beds in the Litchfield Beach region. cal hard clams throughout their latitudinal range.
As noted above, shell growth is strongly as- However, reports of variability in the periodicity
of shell formation across space and time indicate (1989) from Narragansett Bay, Rhode Island,
a need for caution (Fritz, 2001). reporting translucent shell increment formation
In a more recent study Henry and Cerrato during the winter and opaque increment forma-
(2007) reported that the timing of formation of tion during the summer. These patterns were con-
shell increments can quickly change within a firmed by the analysis of oxygen isotope ratios
single locality. Their work reviewed hard clam (18O/16O) found in the translucent and opaque
research published by Jones, Arthur, and Allard. increments (Jones, Arthur, and Allard, 1989). In
Transmitted Light
Translucent Increment Opaque Increment

(Summer/Autumn) (Winter /Spring)
Direct Light
Fig. 8.1. Thin, radial cross section of hard clam (M. mercenaria) shell viewed under transmitted and direct light.
this case, a count of the translucent increments assemblages sampled from the marshes of the
yields the ontogenetic age (in years) of each ani- Litchfield Beach, South Carolina (Georgetown
mal. A size analysis of the age-corrected annual County). A sixth sample excavated from nearby
shell increments documented the growth of Nar- Sewee, South Carolina (Charlestown County) is
ragansett Bay specimens living between 1958 included. These shell middens date between 1690
and 1983 (Jones, Arthur, and Allard, 1989). b.p. and 250 b.p. (table 8.1).
Henry and Cerrato (2007) found a quite dif- In light of the findings by Henry and Cerrato
ferent annual periodic pattern in incremental (2007), we validate the periodicity of the opaque
shell formation in their analysis of shell oxy- and translucent shell increments in the modern
gen isotopes collected from Narragansett Bay in and zooarchaeological specimens by analyzing
2005–2006: a translucent increment formed dur- the ratio of oxygen isotopes (18O/16O) measured
ing the winter, an opaque in the spring, a second in the cross-sectioned shells. We use the annual
translucent in the summer, and a second opaque incremental pattern established from this study to
in the fall. They hypothesize that the changes estimate the composition of the ontogenetic age
in periodic shell formation resulted from an in- classes of the hard clams contained in the modern
crease in water temperature in Narraganset Bay and six zooarchaeological assemblages to chart
(Henry and Cerrato, 2007). If they had relied on temporal changes in their populations.
the Jones, Arthur, and Allard (1989) baseline in-
terpretation, a count of the translucent increments METHODS
would have resulted in an overestimate of the
mean age of the population. They conclude: Archaeological Sites
Chester DePratter and Jim Legg (South Caro-
Future studies, even if they are conduct- lina Institute of Archaeology and Anthropology)
ed in geographic regions where the an- surveyed 18 shell middens located in the tidal
nual growth pattern of M. mercenaria has marshes of the Litchfield Beach estuary (fig. 8.2)
been previously defined, should confirm in 2005. With the exception of one eastern oyster
that the annual pattern has not changed shell (Crassostrea virginica) midden (8GE572),
over time due to local environmental all of the sites were composed of hard clam shells
fluctuations (Henry and Cerrato, 2007). with the isolated remains of Atlantic ribbed mus-
sel (Geukensia demissa) and stout tagelus clam
Their conclusion is particularly relevant to the (Tagelus plebeius). Pottery sherds were rarely
study of zooarchaeological assemblages where encountered, while carbon lenses were observed
baseline patterns of periodic shell growth have during the excavations. The middens are located
the potential of changing in response to local and in the salt marsh and some extend below the pres-
worldwide climatic shifts such as those experi- ent high tide line, indicating that sea level was at
enced during the Holocene by Native Americans a lower stand than it is today.
who populated the coastal environments of the We report on only those sites where hard clam
Southeast (Fagan, 2000, 2005). Given the right sample size is greater than 30 interpretable speci-
environmental conditions, the periodicity of mens and where the deposits were radiocarbon
incremental shell growth could have changed dated (tables 8.1 and 8.2). This represents a total
many times. Recent reports suggest that estab- of six archaeological sites in the Litchfield Beach
lished baseline patterns of increment formation locality (fig. 8.2; table 8.2). Twelve sites remain
can even change within a single researcher’s to be dated and analyzed (fig. 8.2).
lifetime (Jones, Arthur, and Allard, 1989; Fritz,
2001; Henry and Cerrato, 2007). Litchfield—Modern Proxy
In this study we establish a seasonal profile A two-year reference collection of modern
of incremental shell growth using cross sections hard clams (N = 1055) was assembled by Chester
of Litchfield Beach hard clam shells collected and Kalla DePratter who collected approximately
monthly in 2005–2007. These data help to frame 45 live specimens each month over the two-year
the population dynamics of modern hard clams interval, March 2005–March 2007. The purpose
living in the Litchfield Beach estuary (fig. 8.2). of the study was to associate the timing and peri-
We also examine hard clam shell growth in spec- odicity of incremental shell growth with the sea-
imens from five zooarchaeological hard clam sons of the year and to document the population
dynamics (ontogenetic age distribution) of the are then used to provide a temporal assessment
hard clams living in the Litchfield Beach estu- series of zooarchaeological hard clams from the
ary. The annual patterns observed in these data five Litchfield Beach sites.
Virginia Lit
ch
Dri field
ve
North Carolina
38GE581
South Carolina Myrtle Beach 38GE580
Litchfield Beach 38GE572
38GE579
Sewee
Charleston 38GE574
Georgia
38GE578
ATL ANTIC N
38GE585
OCEAN 38GE573
Florida 0 50 100
7
38GE588
y1
MILES
wa
38GE589
igh
.H
U.S
h
eac
ld B
h fi e
38GE587
Litc
38GE586
38GE571
ek
38GE584
Cre
AT L A N T I C
se
OCEAN
ou
bh
Clu
38GE583
38GE582
38GE570
N
Sewee 38CH44
Midway
Inlet
Pawleys 0 500
Island
METERS
Fig. 8.2. Location of the Litchfield Beach archaeological shell middens.

TABLE 8.1
Radiocarbon Dates for Hard Clams (M. mercenaria)
from Litchfield Beach Shell Middens
Site no. Beta no. 14
C dates 1 σ cal. 14
C dates (intercepts) C sample depths
14
38GE570 220164 a .d . 850–1010 (1100–940 b.p.) a .d . 920 (1030 b.p.) 30–50 cm bs

38GE572 209876 a .d . 180–360 (1170–1590 b.p.) a .d . 260 (1690 b.p.) 1.4 m bs
38GE579 223880 a .d . 760–910 (1180–1040 b.p.) a .d . 830 (1120 b.p.) 45–48 cm bs
38GE586 220165 220166 a .d . 1420–1490 (530–460 b.p.) a .d . 1450 (500 b.p.) 20–40 cm bs
a .d . 1310–1420 (640–530 b.p.) a .d . 1390 (560 b.p.) 40–60 cm bs
38GE588 223881 a .d . 1670–1810 (280–140 b.p.) a .d . 1700 (250 b.p.) ST, 35 cm bs
38CH44 229579 a .d . 1420–1480 (530–470 b.p.) a .d . 1450 (500 b.p.) 45–50 cm bs
Sewee clam
TABLE 8.2
Descriptive Statistics of Ontogenetic Age (years) of Modern
and Zooarchaeological Hard Clams from the Litchfield Beach Region
Litchfield 38GE588 38CH44 38GE586 38GE570 38GE579 38GE 572
Modern 250 b.p. 500 b.p. 560 b.p. 1030 b.p. 1120 b.p. 1690 b.p.
Mean 7.88 7.48 6.83 3.17 4.95 5.00 7.75
Standard error 0.15 0.39 0.44 0.20 0.25 0.32 0.50
Median 8 7 6 3 5 5 8
Mode 7 7 5 2 5 4 8
Standard deviation 2.63 2.63 3.05 1.18 1.68 1.97 3.00
Sample variance 6.90 6.93 9.29 1.40 2.84 3.89 8.99
Range 13 11 14 4 6 9 12
Minimum 3 3 2 2 2 3 4
Maximum 16 14 16 6 8 12 16
Count 305 46 48 36 44 38 36
Confidence interval (95.0%) 0.30 0.78 0.89 0.40 0.51 0.65 1.01
Modern hard clams were hand collected live the modern Litchfield Beach assemblage may
in the shallows of Clubhouse Creek. Scientific closely resemble a natural population because
Permit #0984 was issued by South Carolina De- there had been no harvesting pressure over an
partment of Natural Resources because over the extended period.
past 20 years the water quality of the estuary The specimens were frozen, allowed to thaw,
has not allowed for healthy commercial or rec- cleaned of their soft tissue, and the shells were
reational harvest of hard clams. We suggest that numbered and stored. Eventually the shells were
transported to the Florida Museum of Natural Six zooarchaeological and two modern shells
History, Gainesville, for measurement and analy- were randomly selected for isotopic analysis.
sis. The anterior-to-posterior length of the right The shells were radially cross-sectioned from
valve of each specimen was measured using digi- the umbo to the ventral margin and glued to a
tal calipers attached to a personal computer. This glass petrographic microscope slide using two-
allowed for the direct input of shell measurement part epoxy. Each specimen was then mounted
data into Microsoft Excel spreadsheets where all on a Buehler Isomet saw equipped with a low
descriptive statistics were calculated. concentration diamond wafering blade. A sec-
We followed the methods of Quitmyer, Jones, ond radial cross section was made which pro-
and Arnold (1997) to expose and characterize duced a thick section for sampling. Between
the incremental shell growth patterns. The right one and three samples of shell carbonate were
valve of each specimen was radially cross-sec- milled from the opaque and translucent incre-
tioned along the greatest growth axis, from the ments using a Merchantek Micromill equipped
umbo to the ventral margin, using a water-cooled with a tungsten carbide bit. The samples were
lapidary saw fitted with a Mark V alumina oxide taken in ontogenetic sequence from the young-
blade (Quitmyer, Jones, and Arnold, 1997). The est to oldest part of the shell (i.e., umbo to ven-
alumina oxide blade produced a polished cross- tral margin).
sectional surface that was examined with the un- The isotopic analyses were conducted at the
aided eye and/or microscopically (10×, 20×) with light stable isotope mass spectrometry labora-
transmitted light. tory, Department of Geological Sciences, Uni-
For the purposes of our study, the frequency versity of Florida. The powdered shell samples
of individuals in either the translucent or opaque were analyzed according to standard techniques,
phase of shell growth was calculated for each which involved an initial reaction in vacuo with
month to document a pattern, or annual profile 100% orthophosphoric acid at 70° C for 10 min-
that is presented as a histogram for the modern utes. An online, automated, carbonate-prepara-
hard clam assemblage. All of the zooarchaeo- tion system (Kiel III) facilitated the production
logical samples were prepared and analyzed in and purification of the evolved CO2 gas. The iso-
accordance with the methods outlined for the topic differences between the derived CO2 gas
modern comparative collection of hard clams and the VPDB standard were determined with a
(Quitmyer, Jones, and Arnold, 1997; Quitmyer Finnigan-MAT 252 isotope ratio mass spectrom-
and Jones, 2000). eter equipped with a Kiel III carbonate prepara-
tion system. All values are reported in standard
Isotopic Validation of Incremental δ notation where:
Shell Growth Patterns δ18O = [(18O/16O)sample/(18O/16O)standard-1] × 103
The seasonal variability in the oxygen isotope per mil
record contained in modern and zooarchaeologi- The weight of the individual microsamples
cal hard clams provides an independent method was so small that replicates of unknowns could
of validating the annual periodicity of incremen- not be run. However, variation among standards
tal shell growth. Bivalves generally produce shell run before and after sample strings was less than
(CaCO3) at or very near oxygen isotopic (18O/16O) ± 0.03 per mil (‰).
equilibrium with the ambient seawater (Wefer
and Berger, 1991). The exchange reaction is Ontogenetic Age
temperature dependent; thus, where the isotopic Based on the annual cycle of incremental
composition of ambient seawater remains rela- shell growth observed in the shell cross section
tively uniform, warm seawater temperatures are and validated from the isotopic study, the onto-
indicated where the oxygen isotopic composition genetic age (in years) of each modern and zooar-
(δ18O) is depleted, while cooler temperatures are chaeological hard clam was determined by count-
indicated as the oxygen isotopic composition being the number of translucent shell increments
comes enriched (Epstein and Lowenstam, 1953; seen in the radial cross section of the shells. The
Grossman and Ku, 1986). It is for this reason that mean age and 95% confidence interval around
we plot the isotopic values in an inverted fashion the mean was calculated for each archaeological
with negative (warm, up) to positive (cool, down) site. It was then possible to ascertain which of
(Epstein et al., 1953). the sample sites were statistically different (p <
0.05) by noting whether or not their confidence tween –1.75 and 1.48 in the 2005 specimen and
intervals overlapped. This technique is straight- between –2.40 and 1.63 in the 2007 shell. Figure
forward, easily interpretable, and conservative. 8.5 shows a pattern of growth where each trans-
Once the ontogenetic age was determined for lucent increment is characterized by depleted or
every specimen, the rate of survivorship was cal- light δ18O values (warm SST) while the opaque
culated as the percentage of the population sur- increments are relatively enriched (cooler SST).
viving during each year of life. These dynamic These data confirm the monthly observations of
data were plotted as survivorship curves that fa- marginal shell growth, which indicate that trans-
cilitate the comparison of the modern and zoo- lucent increments are seasonally formed during
archaeological samples (Deevey, 1947; Hallam, the summer and autumn (figs. 8.1 and 8.3). Be-
1972; Cerrato, 1980). cause only one translucent increment forms per
year, the total number of translucent increments
RESULTS counted in a shell cross section yields the ontoge-
netic age (in years) of each specimen.
Observed Incremental Shell Growth Carbon isotope (δ13C) profiles from the two
Modern Hard Clams: Figure 8.3 presents the modern shells range between –4.06 and –0.36.
monthly frequency of individual clams (N = 507) The δ13C profiles weakly track the δ18O values
forming the translucent and opaque increments in (fig. 8.5).
relation to annual monthly sea surface tempera- The results of the isotopic analyses from four
ture (SST) and precipitation. The data show that zooarchaeological hard clams, excavated from
during each month some proportion of the hard three sites dating between 250 and 560 b.p., are
clam population is forming the translucent incre- presented in figure 8.6. The δ18O values range
ment in their shells. Opaque shell growth is most between –2.20 and 1.51. In all cases, the translu-
prevalent between February and June. Thereafter, cent increment represents warmer SST (depleted
there is a distinct decline in opaque shell growth δ18O), while the opaque increments formed in
as SST increases toward summer maximum cooler conditions (enriched δ18O). δ13C variation
values. We note that precipitation in the region shows a positive correlation with the seasonal
tracks SST and also reaches its maximum during changes in the δ18O composition (fig. 8.6). The
the warm season. No opaque shell growth was correlation is much stronger than observed in the
observed in August and September when SST data from the modern specimens.
reached the annual maximum (August SST = 28° The three zooarchaeological hard clams from
C). When all of the monthly data are pooled (fig. the sites that date between 1040 b.p. and 1690 b.p.
8.3), annual translucent shell growth, relative to share a similar oxygen isotopic pattern (fig. 8.7)
opaque, dominates the modern Litchfield Beach with the modern shells and the other zooarchaeo-
hard clam population. logical shells. Among the three hard clams (fig.
Zooarchaeological Hard Clams: In 8.7) the δ18O values range between –2.04 and
contrast to the modern hard clams, the opaque 1.70. The oxygen isotope data indicate that all
shell increment is most frequently identified in translucent increments formed during the warm
all of the zooarchaeological assemblages (fig. season and the opaque increments during cooler
8.4). Because the δ18O composition of the mod- SST conditions. The δ13C profiles also show a
ern and zooarchaeological shells (figs. 8.5–8.7) positive correlation with δ18O as seen in the pre-
indicate comparable temperature regimes (iso- vious zooarchaeological specimens.
topic values ranging between –2.00 and +1.00 The isotopic data suggest there is a difference
‰), such a pattern most likely resulted from in the relationship between δ13C and δ18O in the
the preferential, seasonal harvest of hard clams modern versus the zooarchaeological shells. In
during the winter and spring when opaque shell all but one of the zooarchaeological shells (figs.
growth is most prevalent. 8.6 and 8.7), the correlation between δ13C and
δ18O is stronger than in the modern specimens
Isotopic Validation (fig. 8.5). This reflects the natural seasonal input
The translucent and opaque shell growth in- of freshwater into the estuary, primarily during
crements were microsampled from two modern the wet season. In contrast, the modern estuary
hard clams collected live in March 2005 and receives abundant freshwater year-round from
January 2007 (fig. 8.5). δ18O values ranged be- septic tanks, irrigation, and storm water runoff.
100
60
PERCENT (%)
20
20
N = 507
Translucent
60
Opaque
100
MONTH
30 18
25
SEA SURFACE TEMPERATURE (°C)
14
PRECIPITATION (CM)
20
10
15
10 6
Temperature
5 Precipitation
2
0 0
MONTH
Fig. 8.3. Monthly frequency of specimens forming translucent or opaque shell growth increments from March
2005 to March 2007 in hard clams collected from the estuary at Litchfield Beach, SC. Sea surface temperature
(°C) and precipitation (cm) from nearby North Myrtle Beach, SC.
Size vs. Age age and 95% C.I. of the seven samples, in tempo-
Hard clam populations living in the Southeast ral order, shows a significant decline in the mean
display a great deal of variation in age among age between 1690 b.p. and 560 b.p. and a subse-
similar-sized individuals (Walker and Tenore, quent rebound to the present (table 8.2; fig. 8.9).
1984; Quitmyer and Jones, 2000; Fegley, 2001). The 95% C.I. of the three youngest samples and
This pattern also exists in the modern Litchfield 38GE52 (1690 b.p.) overlap and are statistically
Beach hard clam population that was collected in indistinguishable; however, the samples from
2005–2007. For example, the anterior to poste- 1120 b.p. to 560 b.p. are statistically different.
rior range in size of the three-year-old clams en- A plot of the survivorship curves also shows
compasses a sizable proportion of all other age temporal changes in the survivorship of the seven
classes in the sample (see box in fig. 8.8). The hard clam assemblages (fig. 8.10). The survivor-
shells range in age between three and 16 years ship curves for the three samples dating between
(N = 305, mean = 7.9, S.D. ± 0.15), while the 1120 b.p. and 560 b.p. (solid symbols in fig. 8.10)
anterior-to-posterior length ranges between 64.3 show a substantial reduction in the mean survi-
and 84.5 mm (mean = 64.3 mm, S.D. ± 0.41). vorship and the right tail of the curve is truncated
relative to the younger samples (500 b.p. to the
Mean Ontogenetic Age and Survivorship present) or the earliest sample (1690 b.p.).
As noted above there is considerable overlap
in the size range of consecutive age classes in DISCUSSION
each hard clam population. It is for this reason
that we rely on ontogenetic age as a measure of Documentation of the yearly pattern of
the population dynamics of the modern and zoo- growth increment formation in the shells of mod-
archaeological assemblages. A plot of the mean ern hard clams establishes a baseline proxy of an-
100
60
T
PERCENT (%)
20
20
N = 50
O
60
N = 40 N = 44
N = 38
N = 38 N = 38
N = 47
100
Modern 38GE588 38CH44 38GE586 38GE570 38GE579 38GE572
LITCHFIELD BEACH LOCALITY

Fig. 8.4. Percentage of translucent and opaque incremental shell growth identified in modern and zooar-
chaeological hard clams from the Litchfield Beach region. Abbreviations: T = translucent shell growth; O =
opaque shell growth.
nual incremental shell growth that can be used to Beach region. The shell microstructural data in-
evaluate the population dynamics of zooarchaeo- dicate that in the modern population the opaque
logical hard clams excavated from the Litchfield increments form during a short seasonal interval
5
3-21-2005 r = 0.24
4
δ18O and δ13C (VPDB)
2
1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20
SAMPLE NUMBER
5
1-21-2007 r = 0.37
4
δ18O and δ13C (VPDB)
2
1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20
SAMPLE NUMBER
Carbon Translucent
Oxygen Opaque
Fig. 8.5. Variation in oxygen ( O/ O) and carbon ( C/ C) isotopic composition in the translucent and
18 16 13 12
opaque growth increments of modern hard clams from the Litchfield Beach region.
between late winter and spring. These alternate (chemical and microstructural) is consistent with
with translucent shell growth increments that are other modern seasonal and zooarchaeological
most pervasive during the summer and autumn as studies of hard clams reported from the Southeast
seawater temperature warms. (Clark, 1979; Quitmyer, Jones, and Arnold, 1997;
Analyses of oxygen isotopes (δ18O) in the Fritz, 2001; Andrus and Crowe, 2008; O’Brien
modern and zooarchaeological shells indepen- and Thomas, 2008). These findings verify that a
dently verify the seasonal periodicity of the two couplet of one opaque and one translucent shell
alternating shell increments. In all cases, enriched growth increment marks one year of life and
values of δ18O indicative of cool water conditions justifies its use in characterizing the population
characterize the opaque increment and depleted dynamics of modern and zooarchaeological hard
δ18O values (warm conditions) characterize the clam assemblages. Further, the data indicate that
translucent increment. This periodic pattern this pattern has remained constant in the Litchfield
5 5
Litchfield Beach r = 0.51 Seewee r = 0.75
4 4 (38CH44) − 500 B . P .
(38GE588) − 250 B . P .
δ 18 O & δ C (VPDB)
3 3
2 2
13
13
1 1
0 0
1 1
2 2
1 3 5 7 9 11 13 15 17 19 1 3 5 7 9 11 13 15 17 19
SAMPLE NUMBER SAMPLE NUMBER
5 5
Litchfield Beach r = 0.66 Litchfield Beach r = 0.27
4 4
(38GE586) − 560 B . P . (38GE586) − 560 B . P .
3 3
2 2
13
13
1 1
0 0
1 1
2 2
1 3 5 7 9 11 13 15 17 19 1 3 5 7 9 11 13 15 17 19
SAMPLE NUMBER SAMPLE NUMBER
Carbon Translucent
Oxygen Opaque
Fig. 8.6. Oxygen (18O/16O) and carbon (13C/12C) isotopic composition of the translucent and opaque growth
increments from zooarchaeological hard clams (M. mercenaria) from the Litchfield Beach region.
5
Litchfield Beach r = 0.51
4 (38GE570) − 1040 B.P.
δ18 O & δ13 C (VPDB)
3
2
1 3 5 7 9 11 13 15 17 19
5
Litchfield Beach r = 0.66
4 (38GE579) − 1120 B.P.
2
1 3 5 7 9 11 13 15 17 19
5
Murrels Inlet r = 0.44
4 (38GE572) − 1690 B.P.
2
1 3 5 7 9 11 13 15 17 19
Carbon Translucent
Oxygen Opaque
Fig. 8.7. Oxygen (18O/16O) and carbon (13C/12C) isotopic composition of the translucent and opaque growth
increments from zooarchaeological hard clams (M. mercenaria) from the Litchfield Beach region.
90
80
70
SHELL LENGTH (A/P) (mm)
60
50
40
30
20
10
N = 305
0
0 2 4 6 8 10 12 14 16 18
AGE (YEARS)
Fig. 8.8. A comparison of age (years) versus anterior-to-posterior (A/P) shell length in modern hard clams (M.
mercenaria) from the Litchfield Beach estuary.
9.5
N = 36
8.5 N = 305 N = 46
N = 48
7.5
6.5
N = 38
N = 44
5.5
4.5
N = 36
3.5
2.5
Modern 250 B.P. 500 B.P. 560 B.P. 1030 B.P. 1120 B.P. 1690 B.P.
Fig. 8.9. The mean ontogenetic age and 95% confidence interval around the mean of modern and zooarchaeo-
logical hard clams (M. mercenaria) from the Litchfield Beach region.
Beach region for at least the last 1700 years. in soft substrates (Fegley, 2001). In the Southeast
The population dynamics of an organism like they generally occupy the intertidal zone of es-
the hard clam are influenced by its ontogenetic tuarine creeks where they tend to be restricted to
age composition, growth rate, mortality (survi- heterogeneous substrates (Walker, 1989). They
vorship), and recruitment. Human harvesting are typically found in densely packed, circum-
pressure has been shown to alter the population scribed beds among oysters, shell deposits, and to
dynamics of hard clam beds, in particular the a lesser extent sand, mud, and sandy mud (Walker
mean ontogenetic age and survivorship (Walker, and Tenore, 1984; Fegley, 2001). Such a configu-
1989). For example, Walker (1989) has shown ration facilitates rapid, intensive harvesting.
that small, densely populated hard clam beds are Recruitment is a gradual process that is in-
easily overfished. In 1981 a small bed (90 m2) con- consistent from year to year where very few
taining 49 clams per square meter was illegally individuals enter the adult stage (Malinowski,
harvested (hand collecting) in the Wassaw Island 1985). Because of sparse, sporadic settling of
National Wildlife Refuge of coastal Georgia. In spat, major sets are rare and hard clam popula-
a one-week period, the population declined to 22 tions are dominated by larger individuals (Walk-
individuals per square meter. In fact, hard clam er, 1989). Commercial harvesting along the
populations of the southeastern United States are Georgia and South Carolina coasts represents
vulnerable to harvesting pressure because of their a minimal activity. Because of light harvesting
habitat preferences, recruitment, and distribution pressure, it has been suggested that many of the
within those habitats. beds resemble populations that are in equilib-
Hard clams are one of the most abundant, rium (Fegley, 2001). This may extend to other
large-bodied, infaunal suspension feeders living areas of the Southeast where harvesting is lim-
100
Modern, N = 305
38GE588−250 B.P., N = 46
38CH44−500 B.P., N = 48
80
PERCENT OF SURVIVORS (%)
38GE586−560 B.P., N = 36
38GE570−1030 B.P., N = 44
38GE579−1120 B.P., N = 38
60
38GE572−1690 B.P., N = 36
40
20
0
1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17
AGE (YEARS)
Fig. 8.10. Survivorship curves of modern and zooarchaeological hard clams (M. mercenaria) from the Litch-
field Beach region.
ited. The Litchfield Beach hard clam population the population (Walker, 1989). Intensification of
is one such example where harvesting has been natural predation pressure would therefore have
limited or nonexistent for decades due to poor the net opposite effect of what is observed in the
water quality. zooarchaeological assemblages by increasing the
Where hard clam beds are fished, larger in- mean ontogenetic age (size) of the population. In
dividuals are at greater risk than small to medi- contrast, human exploitation that blindly extracts
um-sized individuals (Fegley, 2001). It has been clams from the substrate puts the largest (gener-
suggested that the larger hard clam shells shelter ally oldest) individuals at greater risk of being
the smaller individuals from predators (Walker, collected. In the zooarchaeological assemblages
1989). As larger individuals are removed from that appear to have been impacted by intensive
their beds there may be an increase in predation harvesting, the older members of the population
on smaller specimens (Walker, 1989), which fur- are missing from the zooarchaeological assem-
ther reduces the number of individuals that might blages (fig. 8.10).
reach maximum age. We cannot fully reject unseen environmental
Although hard clam beds tend to be dominat- conditions that could have affected the dynamics
ed by larger individuals, size classes may vary of the various Litchfield Beach hard clam popula-
with habitat type, predator diversity, and density tions. However, previous research has shown that
(Walker and Tenore, 1984). As we have shown the pattern of intensive hard clam exploitation is
above, there is considerable scatter in the size- nonuniform over space and time in southeastern
versus-age relationship (Fegley, 2001). Because North America (Quitmyer and Jones, 2000). At
of this, variations in mean shell size are not a reli- some localities, the age class composition is di-
able indicator of hard clam population dynamics minished, while at other sites dating to the same
and their relationship to harvesting pressure. time period, there are samples that seem to have
In light of our understanding of hard clam been collected from unharvested beds. Quitmyer
biology and ecology, we can consider the onto- and Jones (2000) have also documented the reduc-
genetic age composition and survivorship of the tion in the zooarchaeological age classes of hard
hard clam assemblages that span nearly 1700 clams in some locations in archaeological sites,
years at Litchfield Beach. The mean ontogenetic while in other areas of the same sites there is no
age of the earliest zooarchaeological assemblage evidence for concentrated harvesting. Exposure
(1690 b.p.) resembles that of the modern, unhar- time to humans and intensive harvesting are the
vested beds. Between 1120 b.p. and 560 b.p. the two common variables that exist when heavy ex-
data suggest the zooarchaeological hard clams ploitation of hard clams is identified. Similar ob-
were intensively harvested and the mean age of servations are well documented in the zooarchae-
the population declined significantly. The survi- ological record (Quitmyer, Hale, and Jones, 1985;
vorship curves also show a loss of younger in- Szabó and Quitmyer, 2008; Quitmyer and Jones
dividuals ranging between one and four years of 2000; Marcello and Thomas, 2002; Quitmyer,
life and fewer individuals reaching maximum 2003; Peacock, Haag, and Warren, 2004; Erland-
age. After 560 b.p. there is a rebound in the on- son et al., 2008; Rick and Erlandson, 2008).
togenetic age composition and survivorship that This is the same pattern observed in the Li-
resembles the assemblages dating between 500 tchfield Beach region. With exposure time to
b.p. and the present. These are also similar to the humans and evidence for intensive collecting,
earliest zooarchaeological sample (1690 b.p.). there is a clear and precipitous decline in the
The question might naturally be posed as to mean ontogenetic age (in years) between 1690
whether the diminished mean age composition b.p. and 560 b.p. We cannot account for the sub-
of the ancient hard clam populations observed in sequent increase in the mean age of the popula-
this study could be the result of natural causes tions after 560 b.p., but it is clear that harvesting
as opposed to human harvesting pressure (i.e., pressure was removed and the age class compo-
anthropogenic impact on the beds). Like many sition returned to what may be considered in-
infaunal bivalve molluscs, hard clams refuge dicative of naturally occurring populations. This
themselves in size from nonhuman predation pattern may have been the result of intensive
pressures (e.g., fish, crabs, sea stars, boring gas- harvest in the region and a subsequent aban-
tropods, whelks), which are typically concen- donment of the beds, thus allowing their age
trated upon the smaller, juvenile components of class compositions to equilibrate. Alternatively,
could harvesting pressure have been abated by With these data in hand we can evaluate the
Hispanic incursion into the region? Shortly after season of hard clam harvest in the middens, onto-
560 b.p., Hispanic explorers entered the South- genetic age, and survivorship of the modern and
east, resulting in a profound disruption of the zooarchaeological populations in the Litchfield
Native American population. Beach region. The data show that all of the zooar-
chaeological harvesting occurred during the winter
CONCLUSIONS and spring seasons. There was intensive harvest-
ing pressure between 1690 b.p. and 560 b.p. where
In this research we combine microstructural hard clams were consumed at an intensive rate.
and oxygen isotopic approaches to document and The mean ontogenetic age class composition de-
validate the fundamental pattern of periodic, in- clined from 7.8 years to 3.2 years. After 560 b.p.,
cremental shell growth for hard clams living in harvesting pressure was removed and the survi-
the Litchfield Beach region. The two-year longi- vorship of the hard clam assemblages rebounded
tudinal study of incremental shell growth (obser- with the mean ontogenetic age rising to 7.9 years.
vational data) demonstrates that the opaque (light) This pattern of intense exploitation appears to
growth increment forms in the late winter and have occurred more than one time and in other
spring when sea surface temperatures approach geographic locations in the Southeast. Like many
the optimum for hard clam growth (~20° C) and molluscan taxa, hard clams are sensitive indica-
the translucent (dark) increment forms as the sea tors of harvesting pressure. Where there is evi-
surface temperature increases toward the summer dence for dense human populations, evidence for
maximum (~32° C). The oxygen isotopic compo- sedentism, or in places where hard clams are a
sition of the opaque and translucent shell incre- major focus of subsistence behavior, changes in
ments validate the observational data. A couplet the ontogenetic age classes and survivorship can
of one opaque and one translucent shell growth be identified.
increment represents one year of life. These data Pre-Hispanic harvest of hard clams extended
also confirm a similar pattern in the zooarchaeo- across a wide temporal and geographic range.
logical hard clam assemblages for the past 1700 The decisions that resulted in the changes to
years in the region. This periodic pattern of incre- the zooarchaeological hard clam population dy-
mental shell growth is consistent with previous namics appear to be more intrinsic to the human
research from the southeastern Atlantic coast of species rather than associated with the level of
the United States. cultural complexity.
2012 Reevaluating the use of impressed odostome Boonea impressa 165
CHAPTER 9
Reevaluating the use of impressed
odostome (Boonea impressa) as a Season-
of-Capture Indicator for Oysters
D eborah A nn K eene
Boonea impressa are small (less than 1 cm) pressa are born at one time of the year, a smaller
parasitic gastropods that feed on oysters and percentage is born throughout the year. Using a
are found along the Atlantic and Gulf coasts of modal range theoretically isolates the larger birth
North America (fig. 9.1). Although B. impressa cohort and negates any effects of the other birth
were probably not deliberately gathered by hu- cohorts in determining the season of death. He
mans as a food source, the oysters to which they did not test his model on known samples, but did
were attached were gathered in large amounts. B. apply them to archaeological assemblages. Rus-
impressa are frequently found in archaeological so (1991) has carefully considered many of the
middens, although a small screen size (0.7 mm) drawbacks of assessing seasonality (i.e., mean
must be used to recover them all. length vs. mode size categories); however, there
The use of Boonea impressa as a season-of- are several additional problems that must be ad-
capture indicator for archaeological oysters was dressed. These include variation between species
first devised by Russo (1991) and his ideas have in different environments, constant reproduction
laid the groundwork for others using this species throughout the year, inconsistent growth patterns,
(see Cannarozzi, chap. 10, this volume). Russo feeding habits of juveniles, and lack of specific
(1991) hypothesized that determining the season knowledge of many aspects of growth, reproduc-
of death for B. impressa would reveal the season tion, and behavior. Russo’s application and these
of death for the oysters to which they were at- issues are the focus of this chapter.
tached. His hypothesis rests on the assumptions
that the majority of B. impressa are born at the Location and Environment
same time of year, live only for about one year,
and increase in size throughout the year. There- Morphological characteristics of Boonea
fore, one could determine the season in which differ among geographic locations and environ-
they died by their size. To test his hypothesis, he ments. White, Kitting, and Powell (1985) col-
collected and measured modern samples of B. lected samples of Boonea impressa in October,
impressa from the northeastern coast of Florida December, March, May, and July (1981–1982)
for 14 months. He used these data to develop a on Mud Island, Texas. These samples were com-
model of yearly growth based on shell length. pared to shells collected at Virginia Creek and
The growth curve is divided into shell length size Williston Creek, North Carolina. Multiple mea-
classes representing spring, summer, autumn, late surements of these collections indicated that the
autumn, winter, and late winter (table 9.1). The North Carolina specimens were larger overall
shell length size classes were determined from (White, Kitting, and Powell, 1985: 42). However,
the modal range, or the most common lengths no cause is suggested for this size difference. It
collected during a particular season, rather than may be due to temperature difference between
the average length of all the shells collected. North Carolina and Texas, or it may be that one
Russo did this because, although most B. im- collection was made in 1981–1982 while the oth-
er collections were made in the 1970s. in the same geographic location but in different
White, Kitting, and Powell (1985) counted environments within that location. Porter, How-
sperm and oocytes in a laboratory sample of ie, and Deriso (1979) compared B. impressa and
Boonea impressa collected from Big Slough, Boonea seminuda specimens from North Caro-
Texas, and found that sperm was present in all lina estuaries with B. seminuda specimens from
months but December. The authors suggest that offshore environments. The estuarine samples of
the lack of sperm in December may be due to cold B. impressa and B. seminuda shared more select-
temperatures. If the cold temperatures in Texas ed characteristics than the B. seminuda from dif-
can affect sperm production, it is likely that B. ferent environments did with each other (Porter,
impressa in more northerly climates would not Howie, and Deriso, 1979: 44).
have the same birth seasons as populations in Both the White, Kitting, and Powell (1985)
Texas and Florida. This means that B. impressa and Porter, Howie, and Deriso (1979) studies in-
size classes would not be universal in determin- dicate that morphological differences in Boonea
ing season of capture. This is illustrated by Cann- can be caused by environment and geography.
arozzi’s (chap. 10) work from St. Catherines Is- However, the extent of these differences is not
land. Her size classes are markedly different from yet known, and more work needs to be done, es-
Russo’s (1991) Florida size classes. pecially with the effects of temperature on repro-
Morphological differences have also been ob- duction and growth. The available data indicate
served in Boonea impressa populations that live that size ranges to determine season of capture
may only be useful in a small geographical area,
highlighting the importance of developing eco-
logical analogies using data collected as near
as possible to the sites being studied (see Cann-
arozzi, chap. 10).
Reproduction
160
6 The Boonea impressa method relies on the as-
sumption that the majority of individuals are born
in late spring/early summer and live for approxi-
mately one year (Russo, 1991: 209). However,
5 this assumption may not be supported by avail-
120 able data. Russo (1991: 208) cites two studies of
B. impressa in addition to his own. The first is
Wells (1959), and the second is White, Kitting,
4
and Powell (1985).
The Wells (1959: 142) article states that “the
largest collection of young were made in June and
80
3
July, when they formed by far the most numerous
class in the population.” However, Wells (1959)
does not report specific percentages. White, Kit-
ting, and Powell (1985: 39) report percentages of
2 young (0.5–1.0 mm wide) collected each month.
July contained the largest number of young at
40 55%, however, December contained 22%, Oc-
tober contained 10%, and March 5%. The July
1 sample may have been affected by climatic con-
ditions and may be lower than normal, but this
simply illustrates the point that the Boonea im-
pressa demography is dynamic from one year to
0 0 the next. Regardless, White, Kitting, and Pow-
mm in ell (1985) show that a third of the population is
Fig. 9.1. Drawing of Boonea impressa. born in seasons other than summer, and there is a
fairly large birth rate increase in December, albeit fer from one another, they will not be included in
smaller than the June/July increase. this chapter. The mean length in May 1955 is ap-
This second reproductive peak in the annual proximately 2.5 mm, and the mean in May 1956 is
growth curve (December) would cause many approximately 4.25 mm. The mean length in June
problems in interpreting archaeological assem- 1955 is approximately 1.5 mm and the mean in
blages by using Boonea impressa to determine June 1956 is approximately 2.5 mm. The mean for
season of capture. An archaeological assemblage September is about the same for both years. In Oc-
consisting of a December population would con- tober 1955, however, the mean is slightly over 3.0
sist of about 25% young B. impressa, about 55% mm, while in October 1956, the mean is nearly 4.0
middle-aged individuals, and about 20% of the mm. Clearly, the size means vary between years, in
individuals would reflect a variety of other age this case up to 1.75 mm, and this is precisely why
groups. This distribution would result in a bi- Russo (1991) uses modal size ranges rather than
modal growth curve suggesting spring and win- mean. Modal sizes would do a good job of isolating
ter collection rather than just December. As these a specific cohort, but if that cohort measured as lit-
populations grow, the bimodal curve will con- tle as 0.2–0.4 mm larger than it did in the previous
tinue to shift, producing a bimodal distribution year, Russo’s (1991) particular modal size ranges
that does not reflect the actual collection times would not accurately predict the season of capture.
of the B. impressa. The White, Kitting, and Powell (1985: 38)
studies of Boonea impressa from Mud Island,
Inconsistent Growth Patterns Texas, contain data indicating that the growth
Within Populations patterns of their population of B. impressa do not
match with the growth patterns seen in Russo’s
Two of the key assumptions needed when us- (1991) population. Although the width of the
ing Boonea impressa as a season-of-capture indica- sample populations collected in October, Decem-
tor is that “the average size … increases through- ber, March, May, and July increase throughout
out the year” and that the lengths are comparable the year, a chi-square test indicated that there
across time (Russo, 1991: 209). In Wells (1959), was not a significant difference in size between
the collection of B. impressa takes place over a pe- the March and May populations (p < 0.05). So
riod of 18 months and 16 collections are made (fig. although Russo’s (1991) assumption that B. im-
9.2). Collections were taken during two successive pressa generally grow larger throughout the year
years for the months of May, June, July, August, is supported, the month in which the older popu-
September, and October. One collection was taken lation dies and the new one takes over is variable.
for July/August in 1956; separate July and Au- This conclusion is supported by the differences
gust collections were taken in 1955. As the length between Russo’s (1991) and Cannarozzi’s (chap.
means from these collections would certainly dif- 10) size classes.
TABLE 9.1
Size Classes Formulated by Russo (1991)
Season designation Calendar months included Shell length (mm)
Spring May, June 1.1–2.5
Summer July, August 2.6–3.5
Autumn September, October 3.6–4.0
Late autumn October, November 4.1–4.5
Winter November, December, January 4.6–5.0
Late winter February, March, April 5.1–5.5
It is clear from these data that a population instance indicating that juveniles and adults have
of Boonea impressa from the same area will different feeding habits, it demonstrates a general
have variable growth characteristics from year behavior that could affect distribution of B. im-
to year. If the growth of B. impressa is related to pressa in archaeological assemblages. Feeding
temperature, as suggested earlier, this would ex- on any species other than oyster could skew the
plain the changes in size from year to year and number of juveniles in a random sample, because
would seriously impact the use of B. impressa the smallest B. impressa might not be brought to
size to determine season of capture. the archaeological site unless that species was
present. Thus, the smallest size class would be
Feeding Habits underrepresented and larger individuals, those
preferring oysters, would be numerically more
Another variable is the feeding habits of ju- common. This would result in consistently find-
venile Boonea impressa. Powell et al. (1987) ob- ing a larger number of mature individuals.
served juvenile B. impressa frequently feeding on In fact, most applications that use Boonea
Crepidula plana and other Crepidula species in impressa do find a larger number of mature indi-
Texas. During a controlled experiment in which viduals. When Russo (1991) applied the method
juvenile B. impressa were placed in a tank with to archaeological samples, all 13 of the features
oyster spat and C. plana for five days, all juve- and middens contained B. impressa size classes
niles attached to the C. plana by the fifth day and representing late autumn or winter components
none were found on the oyster spat. Adults always (his larger size classes). Two features had small-
preferred the oyster. Although this is a particular er spring components, one had a smaller summer
4
LENGTH (mm)
0
M J J A S O N D J F M A M J J A S O
1955 1956
Fig. 9.2. Length distribution data from Wells (1959). Reprinted with permission. Colored highlights added
by the author.
component, and one had a larger summer com- capture (Keene, 2004). However, this discrep-
ponent. Fradkin (2008) measured over 6000 B. ancy may be the result of small sample size (N
impressa from the Greenfield site, Florida, to de- = 8 oyster shells).
termine season of capture. Again, all the compo- All four applications of the method described
nents included autumn and late autumn captures above find larger percentages of Boonea impres-
(using Russo’s [1991] size classes) with one hav- sa in the 4–6 mm size class. This supports the
ing an additional summer capture. Cannarozzi hypothesis that primarily larger B. impressa are
(chap. 10) created her own size classes using lo- being recovered at greater rate than smaller B.
cal B. impressa near St. Catherines Island. The impressa. Recovery bias may be due to feeding
most numerous size class for both her archaeo- habits, but it may also be due to increased break-
logical samples was spring, which corresponds age of younger B. impressa, or recovery meth-
roughly in size to Russo’s (1991) autumn, late ods. With only four studies, these conclusions
autumn, and winter size classes (approximately are preliminary. The Russo (1991) and Fradkin
4–6 mm in length). (2008) assemblages do contain significant num-
Russo’s (1991) size classes were used by the bers of spring and summer size classes. However,
author on samples of Boonea impressa taken this could be due to reasons other than a spring
from Grove’s Creek Site, Skidaway Island, Geor- or summer collection of oysters. First, multiple
gia. The B. impressa shells were sorted from flo- years might be represented in the archaeologi-
tation samples taken from a 1 × 1 m midden unit cal sample, with one year producing smaller or
and passed through 0.50 mm mesh. All shells larger adult B. impressa than another year. Sec-
were examined under a low-power microscope ond, a bimodal distribution of larger and smaller
to confirm that they were unbroken. The length individuals might be caused by a population col-
of each shell was measured with a pair of digital lected during the season when there is an influx
calipers. The measurements were taken from the of young B. impressa into an existing population
apex to the abapical end and were divided into of mature B. impressa.
the modal length size classes outlined in Russo
(1991). The data (N = 79) indicated primarily Current State of Knowledge
autumn season of capture (fig. 9.3). It should be Concerning Boonea impressa
noted that stable isotope analysis was performed
on oysters recovered from the same excavation The last issue has been mentioned through-
unit and none indicated an autumn season of out, and that is general lack of information about
30
25
PERCENTAGE
20
15
10
0
Spring Summer Autumn Winter
Fig. 9.3. Season of capture for Grove’s Creek site Boonea impressa shells using Russo method with a 0.50
mm screen.
Boonea impressa growth, reproduction, and would be to recreate an archaeological midden

behavior. The major studies were conducted in sample with Boonea impressa from known col-
North Carolina (Wells, 1959; Porter, Howie, and lection dates. Shells from different years and
Deriso, 1979) and Texas (White, Kitting, and months could be mixed and then measured by
Powell, 1985; Powell et al., 1987). Some data a person who does not know which seasons are
indicate that these populations are morphologi- included. Shells from the same month and dif-
cally distinct, but are insufficient to explain the ferent years should be tested similarly. Increased
degree of difference or why there are differenc- overall sample size as compared to previous stud-
es. Reproductive behavior may be affected by ies is key. Modern collections need to be taken in
temperature (White, Kitting, and Powell, 1985), several environmental zones over several years
but this is inconclusive. Until we know how re- to record clinal variations in growth. Cannarozzi
production and growth are affected by environ- (chap. 10) has already done a multiyear collection
mental factors, using B. impressa as a method to at St. Catherines Island. If molluscan researchers
determine season of capture for oysters may not and archaeologists from coastal universities col-
be reliable, and it is clear that size classes are laborate, this project would not be as daunting as
not universal. it may seem. Although there are certainly issues
There are several avenues that can be taken to that need to be addressed, the cost effectiveness
evaluate the merits of this method for indirectly and ease of this method for determining season of
assessing season of capture for oysters. The first capture indicate that it should not be abandoned.
2012 Estimating the Season of Harvest of the Eastern Oyster 171
chapter 10
Estimating the Season of Harvest of the
Eastern Oyster (Crassostrea virginica)
from St. Catherines Shell Ring
N icole R. C annarozzi
Determining seasonal of use of the animals on where studies have shown that these oysters de-
St. Catherines Island is an important step toward posit identifiable winter growth breaks (Custer and
understanding settlement and subsistence strat- Doms, 1990; Kent, 1992; Herbert and Steponaitis,
egies through time. White-tailed deer, sharks, 1998; Kirby, Soniat, and Spero, 1998). Techniques
catfishes, hard clams (Mercenaria spp.), and that measure growth structures are unsuitable for
impressed odostomes (Boonea impressa) have use on intertidal oyster populations in the South-
been used as proxies for seasonal procurement east because of the sensitivity of this species to ex-
of resources on St. Catherines Island (Thomas, treme changes in the environment (Russo, 1991).
2008: 877). Evidence from one or more of these Changes in water temperature and salinity, storm
animals for year-round use of resources, if not events, and even spawning may cause deposition
residence, is present for all cultural periods on of multiple growth breaks throughout the year
St. Catherines Island. However, evidence for sea- (Kent, 1992; Shumway, 1996; Andrus and Crowe,
sonal resource procurement when the St. Cath- 2000; McLusky and Elliott, 2004). Previous stud-
erines Shell Ring was occupied is very rare for ies of oyster seasonality in the Southeast used shell
the St. Simons cultural period (3000 b.c.–1000 length measurements of the impressed odostome,
b.c.) (Thomas, 2008: 1001). a predatory gastropod found among oyster reefs of
One of the primary ways in which the question the Gulf and Atlantic coasts (Russo, 1991; Frad-
of seasonal settlement of coastal sites has been kin, 2008). This demographic approach (size at
addressed is through the measurement of annual season) measures shell lengths as estimates of
growth increments of bivalve species such as the oyster season of death. Recent studies on oysters
hard clam and, less frequently, the eastern oyster in the Southeast have yielded positive results using
(Crassostrea virginica) (Claassen, 1986; Quit- stable isotope geochemistry (Kirby, 1998; Andrus
myer, Hale, and Jones, 1985; Quitmyer, Jones, and Crowe, 2000; Surge, Lohmann and Dettman,
and Arnold, 1997; Quitmyer, Jones, and Andrus, 2001). However, morphological seasonal indica-
2005). Oysters are often the most numerous com- tors have not been consistently correlated with iso-
ponent in southeastern coastal shell middens, but topic profiles in oysters (Andrus and Crowe, 2000;
have not been considered reliable proxies for sea- Surge, Lohmann and Dettman, 2001).
sonality because morphological features have not Environmental changes are recorded as
been found to correlate with seasons. Additionally, changes in shell chemistry, shell microstructure,
difficulties are encountered in interpreting the iso- and physical morphology (Rhoads and Lutz,
topic record of estuarine organisms due to mixing 1980). Certain criteria are required for both visual
of freshwater and ocean water with varying sa- (structural) and isotopic analyses of modern and
linities and oxygen isotopic signatures. Season of archaeological shells. The primary concerns are
death studies using visual analysis of shell growth clear, visual delineation of growth increments,
structures have been successful on subtidal and ar- age of the animal, precipitation of CaCO3 in iso-
chaeological oysters in the Chesapeake Bay region topic equilibrium with water, minimal diagenetic
effects, and growth throughout the annual tem- described as light and dark bands, respectively
perature range (Shackleton, 1973; Killingley and (see fig. 10.2) (Andrus and Crowe, 2000, Carriker,
Berger, 1979; Claassen, 1993). 1996). However, these bands have not been corre-
The primary goal of this research is to deter- lated consistently to environmental causes such as
mine whether intertidal oysters off the Georgia temperature. Stable isotope geochemistry is cur-
coast can be used as reliable proxies for estimat- rently the best method to understand and correlate
ing season of death in archaeological specimens growth band formation to environmental factors.
on St. Catherines Island. Methods used include This section reviews various methods applied in
direct and indirect season of death estimations. oyster seasonality studies.
Geochemical analysis of carbon and oxygen iso-
topes from modern and archaeological samples Visual Methods
is used as a direct measure of oyster season of In his seminal work on subtidal oysters from the
death. Shell length measurements of impressed Chesapeake Bay in Maryland, Kent (1992) uses a
odostomes from modern and archaeological con- method of analysis in which acetate peels magnify
texts are indirect and are compared to geochemi- growth structures so that winter growth breaks can
cal data. Results on season of capture estimates be distinguished from other breaks on the hinge area
using odostome shell length are also compared to (Kennish, 1980). This produces an index that rep-
results obtained by Russo (1991) for modern and resents a measure of growth since the last growth
archaeological sites. Morphometric analysis of break or winter season and can be used to retrod-
oyster shell shape was applied to modern, histor- ict season of death. Custer and Doms (1990) used
ic, and archaeological shells to determine types this method on a modern control population in the
of habitats from which Archaic oysters were col- Chesapeake Bay area and archaeological specimens
lected (Kent, 1992). in Delmarva Spring. Measurement of the most re-
cent growth is divided by an average of growth
Oyster Seasonality over three or more previous years. Herbert and Ste-
Studies: A Review ponaitis (1998) also used the acetate peel method
for modern collections in Maryland and Early and
Oysters deposit carbonate in concentric rings Late Woodland archaeological assemblages with
along their shell margin. The oxygen isotopic success. Most oyster seasonality studies on mod-
composition of skeletal carbonate is primarily in- ern and archaeological samples in the United States
fluenced by the isotopic composition of the wa- have followed this method. The growth index is
ter which is constrained by freshwater input and the most widely used method but has not been ap-
temperature. Shell carbon is influenced by dis- plied to oysters in the southeastern United States. It
solved inorganic carbon in the source water and has been suggested that the growth index method
metabolic processes (McConaughey and Gillikin, may not be applicable to oysters in the Southeast
2008). Therefore, carbonate samples taken from because the fast growth observed in warmer waters
the dorsal to ventral margins of the umbo provide a allows oysters to reach edible size at younger ages
record of environmental conditions and growth of (Shumway, 1996). Use of the growth index method
that organism. Isotope sclerochronology has been requires three to five years of growth and oyster
extensively used to interpret past environmental populations in the Southeast may reach edible size
conditions for many molluscs (Jones and Quit- within two years. Southeastern oysters may even
myer, 1986; Quitmyer and Jones 2005, see also deposit breaks multiple times within a year during
Andrus, chap. 6, this volume). Determination of all seasons associated with extremes of heat and
season of death for oysters, however, has proved cold while exposed at low tide (Russo, 1991). The
to be a more formidable endeavor. While clams timing of increment formation varies by species and
exhibit a clear record of growth in the form of dis- latitude. Kent (1992) notes that the interpretation of
tinct light and dark bands, oysters do not. Oysters growth increments for southern oysters would be
are composed primarily of calcite with layers of reversed—that distinctive breaks would be associ-
aragonite occurring on the hinge in the area of lig- ated with heat stress. This phenomenon has been
ament attachment and also in the adductor muscle reported for hard clams and oysters in the South-
scar (Stenzel, 1694). In cross section, calcitic foli- east (Jones and Quitmyer, 1996; Surge, Lohmann,
ated shell microstructure is interrupted by irregu- and Goodfriend, 2003). Modern and archaeological
larly sized, chalky shell islands, which have been oysters from Nueces Bay, on the central Texas coast,
do not exhibit this reverse patterning, however. In peake, Virginia, the authors found that oysters
Nueces Bay, oysters tend to form distinctive winter exhibit a seasonal record of water temperature.
growth breaks due to water temperatures that aver- Additionally, they could establish that external
age 8° C lower than the Florida Gulf coast (Cox, growth structures on the hinge correlated with
1994). Cox (1994) posits that cooler water tempera- seasonal temperature changes. This work estab-
tures produce distinctive winter growth breaks and lished some important baselines:
that this difference in average winter water temper- (1) Oysters precipitate shell in isotopic equi-
atures may explain the lack of winter growth breaks librium with respect to the ambient environment.
in southern oyster populations. (2) Growth increments are formed annually and
Intertidal organisms are adapted to the condi- reflect changes in seasonal water temperature.
tions characteristic of their environments. Oysters (3) For this population, fast growth occurs dur-
are poikilothermic, euryhaline, and ecomorphic ing the spring/summer months, whereas growth is
and thus able to tolerate variable extremes of tem- slowed during winter months.
perature and salinity. Their variable morphology Andrus and Crowe (2000) established similar
and wide geographical range make them highly baselines for intertidal oysters on Little Egg Is-
adaptable to various environments (Van Sickle et land, Georgia. Sampling methods differed from
al., 1976). It has been suggested that the rate of those of Kirby and colleagues because southeast-
temperature change is more important for meta- ern intertidal oysters do not exhibit similar mor-
bolic activities than is the level of temperature phology in the hinge plate as those from the Mis-
(Shumway, 1996). There are few studies on the sissippi Delta. Using laser ablation technology,
effects of temperature and salinity on respiration Andrus and Crowe analyzed carbonate samples
but it has been reported that oyster tissues respond from alternating light and dark bands visible in
differently to stressors and environmental condi- cross section. They found that light bands were
tions (McConnaughey and Gillikin, 2008). It is re- formed during warm months and dark bands dur-
ported that respiration for all tissues increases with ing cool months. They noted that the primary fac-
warmer temperatures, and decreases with cooler tor affecting the oxygen isotope composition of
temperatures (Shumway, 1996). However, gill oyster shell is temperature.
function, the primary organ of respiration, is opti- Surge, Lohmann, and Dettman (2001) studied
mum at 25°–26° C (77–79° F) and stops complete- intertidal oysters from the Blackwater River near
ly at 5°–7° C (41°–45° F) (Galstoff, 1964; Eble Naples, Florida, to establish chemical controls on
and Scro, 1996). Intertidal oysters in the Southeast oyster shell chemistry. They established that all ar-
may record heat stress while exposed during low eas of the hinge are suitable for isotopic sampling,
tides (Russo, 1991). Oysters are facultative anaer- including chalky layers, as there was no statistical
obes capable of closing their valves and reducing difference in the carbon or oxygen values. They
O2 consumption to zero if necessary (Hammen, reported that in these oysters, the chalky layers
1969). During the period of time that oysters are provide a larger sampling surface. Like Kirby, they
exposed at low tide, they may not respire but con- found that shells precipitated in isotopic equilib-
serve O2 until immersed again. If these events are rium with water; however, morphological features
recorded in oyster valves, it would be reflected in did not correlate with season. Also, in Naples sam-
daily or subdaily tidal growth structures. ples, oysters exhibit fast growth in the winter, and
slow growth during late summer and fall, though
Geochemical Methods Surge, Lohmann, and Dettman (2001) did not con-
Because of the problems associated with vi- fidently attribute the cause of growth cessation to
sual analysis, methods of determining season of temperature, as this is also the time that spawning
death in southeastern oysters have relied on mea- occurs. They also found that temperature is more
surements of organisms associated with oyster important than salinity in determining the oxygen
populations, like impressed odostomes or stable isotope composition of oyster shell. These differ-
isotope analysis of oyster shell carbonate. Kirby, ences in methods and results outlined above rein-
Soniat, and Spero (1998) applied stable isotope force the need to acquire historical baseline data of
geochemistry to modern and Pleistocene-aged the environment. It is most desirable that historical
oysters to reconstruct past estuarine conditions. data are collected as close as possible to the area of
Using modern oysters from the Mississippi Del- study due to regional variability in shell formation
ta as a proxy for Pleistocene oysters from Chesa- in these animals.
Seasonality Determination the island are classified as high marsh (flooded

Using Impressed Odostomes twice monthly by spring tides) and low marsh
The impressed odostome is a small predatory (flooded twice daily by high tides). In addi-
gastropod common in oyster beds in the Atlantic tion, there are interior marshes that experience
and Gulf coasts of the United States (White and high tides twice daily but lack freshwater in-
Wilson, 1996: 571). These animals attach to the put other than local precipitation (Andrus and
soft tissues of adult and juvenile oysters and feed Crowe, 2008: 503). Oysters were collected for
on body fluids. While these animals may parasitize this study from an interior marsh. The status of
a number of species, oysters are the primary host St. Catherines Island as a conservation and re-
species (Powell et al., 1987). Impressed odostomes search island means that no appreciable human
are reported to have a life span of approximately pressure has been placed on these animals. As a
one year and research conducted over a wide geo- result, they are generally longer lived than those
graphical range agrees that spawning occurs con- in regularly harvested shellfisheries. This is an
tinuously with peak reproduction and recruitment advantage when seeking seasonal patterns over
in late spring and early summer (Wells, 1959; multiple years of growth.
White, Powell, and Kitting 1984; White, Kitting,
and Powell, 1985; Russo, 1991). This initial co- Methods
hort makes up the majority of the odostome popu-
lation and members of the cohort increase in size Stable Isotope Analysis
throughout the year until death in spring of the fol- Approximately 30 modern oysters were
lowing year. This expected growth pattern will be hand collected at low tide from Cemetery Road
reflected in measurements of shell length through Marsh during the middle of each month from
the year. The short life span, predictable growth July 2006 to July 2008 to establish a modern
rate, and close association with oyster populations analog for comparison. Cemetery Road Marsh
make impressed odostomes a suitable proxy for is an inland marsh located on the Atlantic side
the seasonality of oyster collection. Methods fol- of St. Catherines (fig. 10.1). Specimens were
lowing Russo’s (1991) seminal work on this spe- immediately frozen and subsequently trans-
cies are applied to odostomes on St. Catherines ported to the Florida Museum of Natural His-
Island. Keene (chap. 9, this volume) highlights tory for processing and storage. They were hand
important points to consider when using this tech- cleaned with a brush and water and separated
nique. These are primarily related to establishing into singles from clusters when necessary. Soft
a sound understanding of growth, reproduction, tissue was removed manually and shells were
and behavior of odostomes throughout their geo- dried in a desiccator. Measurements of water
graphical range. In addition, the applicability of temperature and salinity were recorded and
this method outside of northern Florida is called water samples taken at the collection site at the
into question due to problems that arise when Rus- time of oyster collection. Salinity levels were
so’s data are applied to populations outside of this measured using a refractometer and temperature
region (Keene, chap. 9, this volume). This study was recorded using a digital thermometer. Two
attempts to resolve some of these uncertainties by well-preserved archaeological specimens were
applying modern proxy data collected on St. Cath- selected from column samples excavated from
erines to archaeological odostome shells from St. St. Catherines Shell Ring. Specimens chosen
Catherines Shell Ring. were complete and free of evidence of fouling
organisms and predation.
Oyster habitat Prior to sampling for isotopic analysis, the
on St. Catherines Island left valve of each specimen was radially cross
sectioned and mounted on glass slides with JB
Oysters are located in inland and estuarine KWIK Weld and fixed to the sample stage of a
salt marshes on both the mainland and coastal Merchantek EO Micromill at the Florida Muse-
sides of St. Catherines Island. These environ- um of Natural History stable isotope laboratory.
ments differ in that the tidal fluctuations in Carbonate samples were drilled in ontogenetic
the estuarine salt marsh average greater than 2 (oldest to youngest) sequence from the calcitic,
m and occur two times a day (Thomas, 2008: foliated layers of the bisected surface of the umbo
255–256). The marshes on the mainland side of only. Chalky layers were avoided because of the
St. Catherines Sound
St. Catherines
Shell Ring
Cemetery Road
Marsh
1 .6 0 1
KM
Sapelo Sound
Fig. 10.1. Map showing location of the St. Catherines Shell Ring and oyster collection site (adapted from
Thomas 2008).
irregularity in size and difficulty of microsam- Historical temperature data for the two-year
pling these areas. Each drill hole measured ap- collecting period (July 2006–July 2008) were
proximately 50 µm in depth and the holes were 1 obtained from the National Data Buoy Center
mm apart (fig. 10.2). The isotopic analyses were station, 41008, located off the coast of Savan-
conducted in the light stable isotope mass spec- nah (http://www.ndbc.noaa.gov/station_page.
trometry laboratory, Department of Geological php?station=41008). Temperature measurements
Sciences, University of Florida. All samples were taken at the time of collection are not used be-
analyzed according to standard techniques (Jones cause of consistent thermometer malfunction
and Quitmyer, 1996). All values are reported in which did not allow for accurate temperature
standard notation where: readings for some months.
δ18O = [(18O/16O)sample/(18O/16O)standard-1] × 103
per mil (‰). Morphometric Analysis of Shell Shape
The temperature of the water in which the Oyster habitat can give insight into interpre-
shell carbonate formed was calculated using the tations of the carbon isotopic signatures. Oyster
paleotemperature equation of Craig (1965) for shell shape is heavily influenced by the habitat
the temperature-dependent fractionation of cal- in which it grows. The substrate, density of the
cite in molluscs relative to seawater: oyster bed, the type and number of epibionts,
T(°C) = 16.9 – 4.2 (shell δ18Ocalcite – δwSMOW) + as well as turbidity, salinity, and water depth all
0.13*(δ18Ocalcite – δwater)2. influence shell shape (Kent, 1992). The mor-
Sample Locations
Dorsal
5 mm
Ventral
Fig. 10.2. Radial cross section of oyster showing sampling locations.

phometric analysis of oyster shells outlined by formation. Low oxygen isotope values indicate
Kent (1992) provides a method with which to warmer temperatures while high values indicate
correlate oyster shell shape to habitat. Calcula- cooler temperatures. All samples were adequate
tion of the height-length ratio (HLR) of the left for analysis with the exception of sample 11 in
valve is the simplest way to interpret habitat. specimen arch-1, which was too small. The δ18O
The HLR is the height of the valve divided by profile from archaeological specimen arch-1 (N =
the length (fig. 10.3). Mean HLR values of a 27) shows a semisinusoidal pattern of five com-
sample population provide an accurate estimate plete cycles with values ranging from –1.35 to
of the habitat in which oysters grew. Measure- 1‰. Similarly, the δ18O profile from archaeologi-
ments of oysters from modern, historical, and cal specimen arch-2 (N = 22) shows five complete
archaeological sites from the island were mea- cycles with values between –1.87 and 1.30‰.
sured to determine habitat differences through Modern oyster CRM-July (N = 26) shows four
time. Modern oysters measured were collected complete cycles with δ18O values between –1.97
from Cemetery Road Marsh (N = 30) and Kings and 1.24‰ and CRM-February (N = 16) shows
New Ground Marsh (N = 30). Oysters from two complete cycles with δ18O values between
Hokes Dock (N = 28) are remnants of an early –1.18 and 1.38‰.
20th-century oyster boiler. Oysters (N = 481) The calculated temperature for modern oys-
from St. Catherines Shell Ring were obtained ters correlates well with the water temperature
from column samples excavated by the author, recorded by the buoy at the time of collection
during the 2007 field season, from excavation
units W82S3 and 789N801E. Meeting House
Field (9Li21) is an archaeological site that dates
to the Irene cultural period (N = 71).
Shell Length Measurements Length

Using Boonea impressa
Oyster clumps collected monthly from July
2006 to July 2008 (25 months) from the Cem-
etery Road Marsh site for stable isotope analy-
sis were cleaned in buckets of water and sieved
through 1.168 mm geological screens and dried.
Archaeological specimens were obtained from
the 1⁄16 in. portion of column samples excavated
from two units (W83S2 and 789N801E) from
St. Catherines Shell Ring. The shell length is
described as the measurement from the tip of
Height
the apex to the abapical end (fig. 10.4). Shell
length measurements were obtained using elec-
tronic calipers attached to a desktop computer
(see Jones, Quitmyer, and DePratter, chap. 8,
this volume). Following Russo (1991), data are
presented as a seasonal frequency distribution
of length size classes. For purposes of compari-
son with Russo’s data, his six-season division
has been applied to modern data from St. Cath-
erines (table 10.1).
Results
Stable Isotope Geochemistry

Figure 10.5 (A–D) shows the δ18O and δ13C
values of the two modern and two archaeologi- Fig. 10.3. Left oyster valve height and length
cal shells. Each exhibits a cyclical pattern of shell measurements for HLR calculation.
with a difference of 4.85° C. Based on a com- Arch-1 was collected during the spring, while
parison of modern sea surface temperatures with arch-2 was collected during the late summer/
paleotemperatures, both archaeological samples early autumn (fig. 10.6). All oysters show a pat-
were collected during warm months of the year. tern of fast growth during the warm months and
Apex
1 mm
Abapical end
Fig. 10.4. Impressed odostome shell length measurement.
TABLE 10.1
Seasonal Divisions and Sample Sizes for Modern
and Archaeological Impressed Odostomes
Season Months Sample Size
Spring May, June 73
Summer July, August 20
Fall September, October 1147
Late Fall October, November 456
Winter November, December, January 726
Late Winter February, March, April 309
W83S2 196
Excavation Unit
789N801E 132
slowed growth during cool months. deep channels and are frequently intertidal. This
The carbon profiles of the modern and archae- is confirmed by the habitat from which mod-
ological specimens are markedly different. The ern oysters were collected. The Cemetery Road
carbon profiles in the archaeological specimens Marsh collecting site is intertidal, with oysters
track closely with oxygen. Values range from densely clustered in soft mud. There was little
–1.44 to 0.51‰ in arch-1 and –2.03 to 0.60‰ in difference (HLR = 0.01) between the HLR of the
arch-2, an approximate variation of 2‰. No such oysters in the units 789N801E and W83S2 from
pattern exists in the modern specimens where St. Catherines Shell Ring, thus the HLR was av-
values range from –0.95 to 0.23‰ in CRM 2006 eraged. Based on the mean HLR, archaeological
and –1.89 to 2.51‰ in CRM 2007, an approxi- oyster samples are classified as bed oysters. Bed
mate variation of 1‰. oysters have an HLR between 1.3 and 2.0 and
Salinity measurements are consistent with a occur in muddy sand in loose clusters or singly
highly saline environment with the exception of (Kent, 1992).
two points in September 2007 and January 2008
that are lower than expected (fig. 10.7). All speci- Shell Length Measurements
mens appear to accurately record ambient tem- of Impressed Odostomes
peratures throughout the growth period. Measurements of modern impressed odos-
tomes from St. Catherines Island follow a growth
Morphometric Analysis pattern of increasing shell length from the summer
of Oyster Shell Shape to spring months (fig. 10.9). Peaks in frequency
The mean and 95% confidence interval of the of each size class represent the growth of the ini-
HLR of oysters from the modern and archaeo- tial cohort through time. All size classes are rep-
logical sites were calculated for comparison (fig. resented in the archaeological samples, therefore
10.8). The 95% confidence interval on the mean all seasons are represented. However, approxi-
HLR shows that the difference between modern, mately 30% of individuals from unit W83S2 and
historic, and archaeological oysters is statisti- 27% of individuals from 789N801E correspond
cally significant. Based on HLR measurements, to winter/late winter/spring collections. When
modern collected oysters are classified as chan- compared to Russo’s data from Florida, on an an-
nel/reef oysters. As summarized by Kent (1992), nual scale of warm and cool months, both data
these oysters are densely clustered and elongat- sets show archaeological oysters were collected
ed with an HLR greater than 2.0. They occur in during cool months (fig. 10.10).
A -2.50
Warm
-3.00
-2.00 -2.50
-1.50 -2.00
-1.00
δ13C‰
δ18O‰
-1.50
-0.50
-1.00
0.00
0.50 -0.50
1.00 0.00
Cool
1.50 0.50
1 3 5 7 9 11 13 15 17 19 21 23 25
SAMPLE NUMBER
B -2.50 -3.00
-2.00
-2.50
-1.50
-1.00 -2.00
δ18O‰
δ13C‰
-0.50
-1.50
0.00
0.50 -1.00
1.00
-0.50
1.50
2.00 0.00
1 3 5 7 9 11 13 15
SAMPLE NUMBER
C -2.50 -3.00
-2.00 -2.50
-1.50
-1.00 -2.00
δ18O‰
δ13C‰
Fig. 10.5. δ18O and δ13C -0.50 -1.50
profiles across oyster speci- 0.00
mens from earliest to most re- -1.00
0.50
cent growth (dorsal to ventral -0.50
1.00
edge of the umbo). Solid lines
represent δ18O values, dashed 1.50 0.00
1 3 5 7 9 11 13 15 17 19 21 23 25 27
lines represent δ13C values.
(A) Modern oyster collected SAMPLE NUMBER
July 2006 from the Cemetery
Road Marsh collection site. D -2.50 -2.50
-2.00
(B) Modern oyster collected -2.00
-1.50
February 2007 from the Cem-
-1.00
etery Road Marsh collection
δ13C‰
δ18O‰
-1.50
site. (C) Archaeological oys- -0.50
-1.00
ter from unit WS83S2 of St. 0.00
Catherines Shell Ring (arch- 0.50
-0.50
1). (D) Archaeological oyster 1.00
from the St. Catherines Shell 1.50 0.00
Ring unit number 789N801E 1 3 5 7 9 11 13 15 17 19 21
(arch-2). SAMPLE NUMBER
Discussion result is similar to that of Surge, Lohmann, and

Dettman (2001) in their study of oysters from
A record of ambient temperature can be de- southwest Florida with the exception that seasons
tected chemically in the shells of modern and of fast and slow growth are reversed. Water tem-
archaeological oysters from St. Catherines. This perature is the primary variable controlling iso-
arch 1= 22.87 °C arch 2 = 28.69 °C

40
WINTER SPRING SUMMER FALL WINTER
28.50 28.50
30 24.41 26.79 28.11
24.86
TEMPERATURE °C
19.34 19.24
20
15.93 16.49
14.80
13.66
10
0
J F M A M J J A S O N D
MONTH
Fig. 10.6. Average monthly sea surface temperatures for the National Data Buoy Center Station, 41008 for
2006–2008.
40
35
30
SALINITY (‰)
25
20
15
10
0
Jul Sep Nov Jan M ar M ay Jul Sep Nov Jan M ar M ay Jul
2006 2007 2008
MONTH-YEAR
Fig. 10.7. Salinity measurements for July 2006–July 2008 from Cemetery Road Marsh collection site.
3.5
N = 30
Channel/Reef
3.0
N = 30 N = 28
HEIGHT/LENGTH RATIO
2.5
N = 71
2.0
Bed N = 481
1.5
1.0
Sand
0.5
0.0
Cemetery Road King New Hokes Dock St. Catherines Meeting House
Marsh Ground Marsh Shell Ring Field
S ITE
Fig. 10.8. A plot of the mean and 95% confidence interval of the shell height-to-length ratio (HLR) of oysters
from modern, historic, and archaeological sites (after Kent, 1992).
70
Fall
60
50
Spring
PERCENTAGE (%)
Summer
40
Winter
Late Winter
30 Late Fall 789N801E
W83S2
20
10
0
1 2 3 4 5 6 7
LENGTH (mm)
Fig. 10.9. Frequency distribution of impressed odostomes length size classes from units W83S2 and
789N801E plotted following Russo’s (1991) six-season division.
topic variation in the oyster shell. The observed ranges are similar to those during the Archaic oc-
temperature at the time of collection is consistent cupation of St. Catherines Shell Ring.
with calculated temperatures and oxygen iso- Salinity is an important factor affecting oyster
tope variation in modern shells. The temperature growth, particularly in estuarine environments
range of all specimens does not exceed the range where salinity can fluctuate rapidly and severely.
of modern monthly water temperature averages, Of particular concern in estuarine environments
which suggests that modern seasonal temperature is lowered salinity which results in a decrease in
A 70
Warm
60
Cool
PERCENTAGE (% )
50
W83S2
40
789N801E
30
20
10
0
1 2 3 4 5 6 7
LENGTH (mm)
B 70
Warm
60
Cool
50
PERCENTAGE (% )
W83S2
40
789N801E
30
20
10
0
1 2 3 4 5 6 7
LENGTH (mm)
Fig. 10.10. Comparison of modern odostome collection data with archaeological measurements on an annual
scale. A. St. Catherines Island. B. Crescent Beach (Russo, 1991).
the carbon and oxygen isotope ratios, producing Comparing the carbon isotope profiles reveals
a false temperature reading (Mook and Vogel, possible differences in water chemistry between
1968). Decreased salinity occurs with the intro- the modern collection site and the archaeological
duction of freshwater by rivers, streams, and pre- setting. Modern signatures correlate to a stable
cipitation in the estuary. Oysters are adapted to a habitat while the archaeological specimens may
wide range of salinities. Optimal salinity ranges be from habitats that experience greater seasonal
exist though these vary geographically (Van freshwater mixing. The incorporation of carbon-
Sickle et al., 1976; McLusky and Elliott, 2004). ate carbon in the mollusc shell is complicated by
Lowered salinity in the estuary can invoke a dual metabolic and environmental effects and, as
number of physical and physiological responses such, has not been considered as reliable as oxy-
in different classes of organisms, but for bivalves, gen as an environmental proxy (Grossman and
particularly sessile ones like oysters, the response Ku, 1986). Recent studies have confirmed that
is typically to seal themselves within their shell. aquatic molluscs generally build shell carbonate
The ability of oysters to tightly close their shells from ambient dissolved inorganic carbon and,
allows the organism to protect itself from ad- when carefully interpreted, shell carbonate can
verse environmental conditions, provided these be an effective proxy for environmental condi-
are temporary (Galstoff, 1964). Upon closing tions like salinity and magnitude of freshwater
the valves, an oyster can reduce its oxygen con- mixing (Fry, 2002; McConnaughey and Gillikin,
sumption until oxygen becomes available again 2008). It is possible that the carbon signatures in
(Hammen, 1969). It is possible that during times oysters from the shell ring reflect environments
of stress such as salinity or temperature extremes, that experience greater freshwater input through
or when exposed to air, oysters respire anaerobi- streams or rivers. Modern oysters were collected
cally. These events are not likely to be recorded from the Atlantic side of the island, which re-
in the microstructure of the shell and therefore ceives little freshwater input except in the form
would have no bearing on isotopic composition of precipitation (Andrus and Crowe, 2008). This
(Schöne, 2008; Lutz and Rhoads, 1977). Andrus consistently saline environment could account
and Crowe (2000) suggest that variation in salin- for the weak correlation between carbon and oxy-
ity could produce oxygen isotopic variation simi- gen profiles of modern oysters. Oysters from the
lar to what is produced by temperature but salin- St. Catherines Shell Ring, located on the western
ity would have to be seasonally variable. Salinity side of the island, could have come from habi-
measurements from the Cemetery Road Marsh tats closer to the ring that experience seasonal
collecting site did not show seasonal variation. differences in freshwater input. The differences
While oysters may record tidal cycles in their in carbon signatures between the modern and ar-
shells, the samples presented here show varia- chaeological oysters cannot be interpreted with-
tion on a larger scale and fluctuations in salinity out comparative water data from the western side
over tidal cycles are much more fine grained than of the island.
what is represented in the oxygen isotope profiles Morphometric analysis of shell shape indi-
for these specimens. The modern samples taken cates that modern and archaeological oysters
from the oyster bed on St. Catherines indicate lived in different habitats. The difference in shell
a constant, highly saline environment with the morphology between the modern and archaeo-
exception of measurements made in September logical oysters is supported by the carbon data
2007 and January 2008. Precipitation records that may also indicate different habitats. Dif-
from Sapelo Island, located just south of St. ferences in shell shape may reflect changes in
Catherines, indicate that the island received 6.9 island habitats due to anthropogenic effects. It
cm of rain over the two days prior to the Sep- may not be that Archaic peoples collected oys-
tember collection date and 3.4 cm of rain a day ters from different habitats on the island, but that
before the January collection date (climatological habitats on St. Catherines have changed through
data for Georgia, National Climatic Data Center, time. Rollins and Thomas (2011: 324–325) sug-
2009). The changes in salinity that occur during gest that environmental degradation and disease
the tidal cycles in the estuary are not significant have likely affected the oyster populations on and
enough in duration or severity to account for the around St. Catherines but the authors note prob-
oxygen isotopic variation seen in the modern lems with interpretation of shell size changes
oyster shells. in archaeological contexts. Future studies will
couple morphometric analysis with additional volume). Shell length measurements of impressed
habitat indicators such as type and frequency of odostomes from St. Catherines are suitable for de-
oyster shell predation markers by boring sponges termining archaeological season of oyster collec-
and other organisms. tion, provided appropriate proxy data are carefully
Shell length measurements of the impressed applied.
odostomes indicate cool season collections for
archaeological oysters. The combination of geo- ConclusionS
chemical data and odostome measurements in-
dicate year-round oyster collection for the St. Modern and archaeological oysters were
Catherines Shell Ring. In comparison to Russo’s evaluated as proxies for environmental condi-
data from Florida, the seasonal growth patterns of tions and season of oyster harvest on the Geor-
odostomes from St. Catherines are different, but gia coast using direct and indirect methods of
both populations exhibit a trend of increasing size season of death estimation. Geochemical re-
throughout the year. Additionally, modern winter cords indicate that temperature is the primary
collections in both data sets are represented by variable influencing oxygen isotopic variation
short, broad distributions. Russo explains that this in the oyster shell. Although variations in δ18O
is due to the secondary cohort that widens the size are consistent with variations in water tempera-
range of the most numerous size classes as the ini- ture, more localized historical water chemistry
tial cohort dies off. This accounts for the absence data are needed to estimate seasonal tempera-
of a single peak in the cool season sample. The ture ranges and the range of variation that may
increase in shell length from summer to spring occur within individual oysters and oyster beds.
for the 25-month period supports the hypothesis Future studies will be based on larger sample
that a predictable pattern of annual growth ex- sizes and localized historical water tempera-
ists for odostomes on St. Catherines Island. This tures for specimens currently being collected
hypothesis is challenged by Keene (chap. 9, this from St. Catherines Island. Analysis of mod-
volume), who cites the need to better constrain key ern oyster samples from the western side of the
life history parameters such as variation in shell island, closer to the location of St. Catherines
morphology, growth rate, reproduction, and be- Shell Ring, are needed and may contribute to a
havior. While this is certainly true, collecting mod- better understanding of the carbon signatures.
ern proxy data from environments closest to the Morphological data show that different habitats
archaeological site is essential for gathering such were exploited and/or that oyster habitats on
data. Variation in shell morphology, growth rate, St. Catherines have changed considerably over
and reproductive behavior are expected across time. Measurements of impressed odostomes
geographic ranges and through time. Intraspecific indicate year-round collection, with the greatest
shell shape plasticity associated with environ- number of individuals collected in spring. Mea-
mentally mediated growth patterns is common surements of this species remain the simplest
and documented for many molluscs (Kemp and and most cost-effective method for determining
Bertness, 1983, Martin-Mora et al., 1995; Zieritz, season of oyster collection.
2010; Marquez and van der Molen, 2011). Fur- It is possible to estimate the season of death
thermore, growth patterns in species occupying of oysters from archaeological sites with appro-
the same locality can change over short periods of priate proxy data. Currently, general estimates of
time (Henry and Cerrato, 2007). Proxy data col- seasonal use of oysters at the St. Catherines Shell
lected over at least a full year, during all months of Ring cannot be made due to the small sample
the year, are needed to assess growth patterns. Two size of oysters studied, however, future analysis
or more years of collection will capture variation of modern and archaeological specimens from
that occurs within a population over time. Once a the island will contribute to a better understand-
local pattern of growth has been established, suit- ing of the role of oysters in the Archaic subsis-
ability of the proxy can be determined. However, tence economy. Although geochemical methods
frequent validation of known life history param- remain the most effective for determining season
eters over time and space is necessary if we are of death of oysters, future work should focus on
to confidently apply proxy species data to past understanding the timing of visible growth struc-
animal populations (Jones, Quitmyer, and Andrus, tures on the interior umbo and correlating growth
2004, Jones, Quitmyer, and DePratter, chap. 8, this structures to geochemical signatures.
2012 What Can Plants And Plant Data Tell Us About Seasonality? 187
CHAPTER 11
What Can Plants And Plant Data
Tell Us About Seasonality?
C. M argaret S carry and K andace D. H ollenbach
Contradictory descriptions by 16th- and USING PLANTS IN ANALYSES

17th-century Jesuit and Franciscan mission- OF SEASONAL RESOURCE USE
aries, combined with comparatively sparse AND MOBILITY
archaeological investigations, have led to on-
going debates among historians and archae- Plants have predictable cycles of flowering
ologists about the degree to which the Guale and fruiting and generally grow in relatively
and their predecessors were seasonally mobile distinct habitats. Thus, plants seem obvious
(Thomas, 2008). Were the Guale and other Na- sources of evidence about seasonality and mo-
tive American communities living along the bility. We can certainly construct charts show-
Georgia Bight mobile foragers, who visited the ing when plants identified in archaeological
coast as part of their seasonal round? Or, were assemblages were ready for harvest and where
they sedentary central-base foragers—and in gatherers could find those plants. However, in-
later times forager-farmers, who spent much, if ferring seasonal use or mobility patterns from
not all, of the year on the coast or barrier is- such charts is far from straightforward. Some
lands? The American Museum of Natural His- plant resources must be harvested as soon as
tory’s program of systematic excavations and they ripen, while others persist into later seasons
ecological studies, directed by David Hurst and may be gathered when time allows. More
Thomas, on St. Catherines Island, has produced problematic, most plant foods can be stored for
a corpus of data that we can bring to bear on later use. Some remains found in archaeologi-
questions relating to the Guale and their ances- cal sites probably derive from plants deposited
tors (Thomas, 2008; Reitz et al., 2010). But to shortly after they ripened, but other remains
understand fully the seasonal activities and mo- may derive from plants processed or consumed
bility patterns of the island’s residents, we need long after they were harvested. Thus, ripening
fine-grained data about the plants they collected dates and habitat preferences by themselves are
or cultivated and consumed. Unfortunately, we insufficient to explain people’s seasonal uses of
lack systematic analyses of archaeobotanical plants or how these relate to settlement and mo-
data from St. Catherines Island (Thomas, 2008: bility patterns.
978) and there are relatively few detailed analy-
ses from elsewhere along the coast to help fill in Optimal Foraging Models
the picture. Nonetheless, it is possible to draw A more promising approach is to consider
on patterns of plant use from the lower South- the full range of decisions and activities people
east and the broader Georgia Bight to offer undertake as they obtain, process, and consume
some thoughts about how plant data can con- plants and animals. Gatherers do not simply em-
tribute to understanding seasonal resource use bark on forays to collect whatever resources they
and mobility patterns on St. Catherines Island happen upon during their day’s walk. Instead
and elsewhere along the coast. they consider which plant foods are ready for
harvest or which animal resources are prime for however, require detailed information about nu-
collecting, which patches would be best to visit trient composition that may not be available for
first, which individuals should join a particular many wild foods.
foraging party, and whose efforts might best be Plant foods fare better in optimal foraging
spent at camp processing plant and animal foods models that separate activities of hunters, pre-
gathered previously. If we take into account the sumably men, from those of gatherers, presum-
seasonal and spatial restrictions of plant resourc- ably women and children (see Hawkes [1996]
es as well as the practices and decisions involved and Kelly [1995], for further discussions of the
in monitoring, cultivating, harvesting, collecting, division of labor). These models incorporate as-
processing, cooking, and storing plants—and an- sumptions that gatherers target foods that may
imals—then we can model the seasonal activities deliver lower return rates but are predictable.
and movements of gatherers. Elston and Zeanah (2002) use this approach to
Optimal foraging models are one avenue for construct diet breadth models for pre-Archaic
addressing decisions and activities pertaining hunter-gatherers in Railroad Valley, Nevada.
to resource choices. Models of optimal forag- Their results suggest that men’s hunting oppor-
ing differ in assumptions, variables, currencies, tunities determined residential mobility, while
and computational formulas (Bettinger, 1991; women’s foraging opportunities determined
Kaplan and Hill, 1992; Kelly, 1995). With the site locations (Elston and Zeanah, 2002: 115).
exception of nuts, plants tend to rank low in diet Closer to our area of interest, Hollenbach (2009:
breadth models that use only calories or simpli- 69–97) applies a central-base foraging model to
fied return rates for predicting which resources Late Paleoindian and Early Archaic data from
should be targeted. More detailed linear pro- Alabama. Among other things, she demonstrates
gramming models often include micronutrients that plants requiring minimal processing, such as
in their consideration of resource value (Betting- early spring greens and summer fruits, can have
er, 1991: 116–118; Kelly, 1995: 74–78). Mod- high return rates when people collect them near
els that incorporate micronutrients highlight the camp but return rates decay rapidly with distance
importance of fruits and greens because of the (fig. 11.1). As return rates for large game do not
vitamins and minerals they contain, and depict drop precipitously with distance, she argues that
relatively diverse diets, which appear to be pre- site locations and mobility patterns were orga-
ferred by most primates (Milton, 1993; Addessi nized around the seasonal and spatial availability
et al., 2010).1 Such linear programming models, of gathered resources.
8000
RETURN RATE (KCAL/HR)
7000
6000
5000 Mulberry
4000 Deer, fall
3000 Hickory nuts
2000
1000
0
0 15 30 45 60 75
DISTANCE (KM)
Fig. 11.1. Comparison of return rates for mulberries, hickory nuts, and deer (see Hollenbach, 2009, for full
discussion).
Operational Chains dressed to construct operational chains for plants.

Some of the difficulty in applying optimal for- We first need a list of potential foods and in-
aging models, or more broadly, human behavioral formation about their distribution, abundance,
ecology models, stems from the problem that rela- and predictability. We can assemble our list by
tively little information about the decisions and ac- combining species inventories from archaeolog-
tivities associated with the use of plant resources ical sites with ethnographic or ethnohistoric in-
is available or employed. One way we can rem- formation about plant use, and knowledge about
edy this is by constructing operational chains for edible plants available in local fields and forests.
potential foods. The study of operational chains People’s use of diverse sets of seemingly low-
(chaînes opératoires) was pioneered by Leroi- ranking plants becomes more comprehensible
Gourhan (1964), who argued that operational se- when we take seasonal cycles into account and
quences were deeply embedded in human behav- recognize that choices are based on what food
ior. Material culture, the organization of space, and resources are currently available and what can
subsistence strategies are the products of multiple be stored for later use. Other factors, such as a
technical choices (Leroi-Gourhan, 1993; Stark, predilection for a varied diet or human social
1998). Operational chains have been widely used considerations, also certainly affect people’s
to study the manufacture and use of stone tools food selections.
and ceramics (e.g., Gosselain, 1998; Lemonier, Monitoring/Cultivating: An initial step in
1986). A similar approach—though not always building operational chains is to consider the lev-
referred to as operational chains—is commonly el of human involvement in the growth of each
used by archaeobotanists to study crop process- plant. At the simplest level, in the course of their
ing activities in Europe and Asia (Hillman, 1984; daily activities, gatherers monitor plants from
Jones, 1984; Fuller, Korisettar, and Venkatasubba- flowering to fruiting. They make their decisions
iah, 2001). While operational chains have not been based on knowledge of what is ripe, where alter-
used much to examine the food-related activities natives are located, and what will remain avail-
of Native Americans, we think they would yield able in the future. Gatherers might also tend plants
important insights. in various ways, including clearing underbrush,
In constructing operational chains for subsis- pruning, girdling or removing competing trees or
tence activities, the goal is to identify the choices, plants, coppicing, or even transplanting (Shipek,
activities, resources, tools, and refuse associated 1989; Yen, 1989; Fowler, 2008). Acorn (Quercus
with each step in acquiring, processing, and con- sp.) and hickory mast (Gardner, 1997; Scarry,
suming a particular food. Such an exercise can 2003) were important resources for southeast-
identify pertinent activities and decisions as well ern foragers and remained important after crops
as the material signatures they would generate. were added to their repertoire. People gathering
It can also identify potential scheduling conflicts firewood, hunting, or simply passing through oak
when decisions must be made to gather or pursue and hickory groves likely kept an eye on the im-
one plant or animal and forego another. Construct- pact of spring frosts, summer droughts, etc. on
ing operational chains for all of the food resources the flowers and developing nuts. They would
available to the inhabitants of St. Catherines Island know well before fall whether it was a good mast
through time is beyond the scope of this chapter. year and which trees and groves were most pro-
Here, we discuss some of the most relevant deci- ductive. Several researchers have suggested that
sions and activities associated with key plant re- people not only monitored nut groves but also
sources used by prehistoric peoples living in the managed them to enhance productivity (Munson,
Southeast. We then draw on what archaeobotani- 1986; Gardner, 1997; Scarry, 2003).
cal data are available from the Georgia Bight to Intensive agriculture stands at the other end
illustrate some of the seasonal activities and de- of the spectrum, but cultivation of plants can be
cisions associated with the use of these various described as a continuum (Ingold, 1996; Smith,
plant resources. Table 11.1 sketches out activities, 2001). People’s investment in cultivated plants
resources, and tools for gathering and processing may range from simply broadcasting seeds and
hickory (Carya sp.) nuts, which were prominent harvesting those plants that survive through the
foods for Native Americans in the Southeast. In growing season, to the much more intensive
the following paragraphs, we give a more general work of preparing beds, sowing seeds individual-
description of the considerations that must be ad- ly, weeding, and warding off predators. There are
ethnohistoric accounts of southeastern farmers plant resources will be removed and which, if
engaging in the full range of cultivation activi- any, will be allowed to remain. Descriptions of
ties: from casually sowing chenopod (Chenopo- native fields (Bartram, 1928: 57) suggest that
dium berlandieri) on exposed mudflats (Gilmore, people left fruit trees such as plum (Prunus sp.)
1931; Smith, 1992) to intensively preparing fields and persimmon (Diospyros virginiana) standing
for maize (Zea mays) and other crops (Scarry, when they cleared the land. We speculate that
2008; Swanton, 1946: 268, 274, 289, 292). The they also avoided placing gardens in favored nut
decision to sow on newly exposed ground has groves or at least left prolific trees to continue
little impact on other plant resources. Preparation producing mast. Weeding was also likely selec-
of more formal plots, however, involves choices tive; people probably tolerated and even tended
about how to allocate time and labor as well as plants that produced greens, medicines, or fruit
choices about where to locate the gardens, what (Scarry and Yarnell, 2006).
TABLE 11.1
Decisions, Activities, and Artifacts Associated with Hickory Nut Use1
Associated Potential opportunity
Season Task Application to St. Catherines
tools/features cost
Likely embedded in spring
Monitor spring Negligible if embedded
Spring forays to gather/collect/trap/fish
flowering in other activities other resources
Likely embedded in summer
Late Monitor nut de- Negligible if embedded forays to gather/collect/trap/fish
summer velopment in other activities other resources
Likely embedded in early fall
Early Monitor timing Negligible if embedded forays to gather/collect/trap/fish
fall of harvest in other activities other resources
Hickories are interspersed
Baskets, poles for
Gather hickory among oaks on the island both
knocking nuts from Gather acorns
nuts nuts could be gathered simul-
branches taneously
Decide size of Can shellfish collection be de-
Collect shellfish
work group layed, or performed by others?
Return to logis-
tical camp or Hunt/trap animals that Can hunting/trapping be de-
home base each are putting on winter fat layed, or performed by others?
night?
Store hickory Large storage pits, Other food procurement Can storage tasks be performed
nuts baskets, bags tasks by a relatively small group?
Fall
Store above or Other maintenance tasks
below ground?
Minimal if hearth-dried,
Decide to parch Baskets, pots, other food procurement
first? griddles tasks if parched
Decide near lo- Base camps likely within easy
gistical or base travel distance to tree groves
camp?
Decide who
controls stores— Storage pits may be located in
groups or fami- public or private locales
lies?

Associated Potential opportunity
Season Task Application to St. Catherines
tools/features cost
Stores are likely kept near base
Process hickory Other food procurement camp due to small travel dis-
nuts tasks tances on the island
Decide size/ Can processing tasks be per-
composition of Other maintenance tasks formed by a relatively small
work group? group?
Decide size of Other foods to eat
batch?
Groundstone tools are likely
Decide tools Anvil, hammerstone, to be recovered and may have
used for initial basket sieve, large residues amenable to lipid
cracking nutshell fragments analysis
Groundstone tools are likely
Decide tools Mortar, pestle, small- to be recovered and may have
used for addi- er nutshell fragments residues amenable to lipid
tional smashing analysis
Carbonized nutshell is likely
Decide discard to be preserved and recovered
of nutshell, pri- Carbonized nutshell Other fuels to burn through flotation; size of frag-
marily through fragments
Fall/ ments may reflect processing
burning
winter/ stage
spring What other foodstuffs (fresh or
Cook hickory Other foods to eat stored) are available? What is
nuts their nutritive/flavor content?
Can cooking tasks be per-
Decide size of Other maintenance tasks formed by a relatively small
batch group?
Separate nut-
meats, or melt
into oil
Subsurface features are likely
Decide method to be recognized during exca-
of heating: hot Fire-cracked rock, vation but may have been used
rock boiling or hearths/cooking pits for various cooking/heating
direct heat tasks
Skin-lined pit, ce- Ceramic vessels are likely to be
Decide tools
used for cooking to skimvessel,
ramic scoop recovered and may have resi-
nutmeats dues amenable to lipid analysis
Discard nut- Carbonized plant remains are
Carbonized nutshell
shell, primarily Other fuels to burn likely to be preserved and re-
through burning fragments covered through floatation
1
Sources: Swanton, 1946; Talalay, Keller, and Munson, 1984; Gardner, 1997; Fritz, Drywater Whitekiller, and
McIntosh, 2001; Thomas, 2008; Roger Cain, personal commun. 2010.
The investment that gatherers make in tend- able crops every two to five years (Schopmeyer,
ing or cultivating plant foods may vary from year 1974), they may devote more effort to the prepa-
to year, depending on the availability of other ration and upkeep of garden plots or maize fields.
foodstuffs. If gatherers expect a poor yield from Alternatively, if people anticipate poor crops be-
oak and hickory trees, which only produce size- cause of drought, flood, frost or other inclement
events, they may monitor and tend wild resources movements and locations of game are dependent
more carefully or plan foraging ventures that take on the plants they eat (Petruso and Wickens,
them farther afield than usual. 1984; Talalay, Keller, and Munson, 1984; Scarry,
Harvesting/Collecting: Another con- 2003; Hollenbach, 2009).
sideration is when and for how long the edible We also need to consider how the various
portions of plants are available. When we model seeds, nuts, fruits, and roots were collected and
gatherers’ and farmers’ activities, to the extent transported (Bettinger, Malhi, and McCarthy,
possible, our estimates of when foods can be 1997). Do the desired parts remain on trees,
harvested should be based on local data. Besides shrubs, and stalks or do they drop to the ground?
when they can be collected, it is key to know how How readily can they be seen and gathered if they
long foods stay edible and whether they are at- are on the ground? Are tools needed for harvest-
tractive to other animals. Knowledge about du- ing or can the plant foods be picked, stripped, or
rability and competition guides gatherers’ deci- uprooted by hand? What kinds of containers are
sions about what they must collect immediately needed for collection and transport? Often there
and what plants they can delay collecting. Fleshy are alternatives to be weighed. For example,
fruits such as mulberries (Morus rubra) will spoil Hollenbach (2009: 87) shows that distance af-
or be eaten by other animals, whereas dry fruits fects whether it makes more sense for gatherers
such as cabbage palm (Sabal palmetto) berries to cut chenopod plants, tie them in bundles, and
may linger on plants for weeks or months. Sweet thresh them on return to camp or hand strip the
acorns are prone to mold and insect infestations seeds into baskets or bags and not transport the
and they are favored browse for deer, turkeys, inedible stalks (fig. 11.2). For acorns and hickory
and other wildlife. Bitter acorns and thick-shelled nuts, gatherers must consider not only the cost
hickories are more rot resistant and will be by- of transporting the nuts in the shell but also loss
passed by wildlife when sweet acorns are avail- of “shelf life” if they choose to shell nuts at the
able. Gatherers can defer collection of bitter and grove to reduce transport weight (see below).
thicker-shelled nuts while they harvest those that Processing/Storage: It is also important to
need immediate attention. Knowledge of ripen- know what processing is required for a given
ing periods of plants that are favored by wildlife foodstuff. This question is inextricably connect-
also is important for hunters because the seasonal ed to whether the food will be eaten immediately
1400
1200
RETURN RATE (KCAL/HR)
1000
800 Amaranth, cut

Amaranth, strip
Chenopod, cut
600
Chenopod, strip
400
200
0
0 15 30 45 60 75
DISTANCE (KM)
Fig. 11.2. Comparison of return rates for cutting and threshing versus hand-stripping for chenopod and ama-
ranth (see Hollenbach, 2009, for full discussion).
or stored. If gatherers plan to use the food in the is taken out of storage and cooked. Gremillion
near future, their processing considerations re- (2002) notes that starchy and oily seeds such as
volve around what needs to be done to make the chenopod, maygrass (Phalaris caroliniana), sun-
food edible or palatable and where to perform flower (Helianthus annuus), and sumpweed (Iva
these activities. Do they eat the food on loca- annua) have relatively low return rates; however,
tion or return to camp to cook it? If they decide they can be stored with chaff or hulls attached
to return to camp, do gatherers remove inedible and processed in the winter when there are few
portions at the collection site to reduce trans- foods to gather. She argues that this led not only
port costs or maximize collection time and save to the regular incorporation of these foods into
“cleaning” for camp? southeastern people’s diets but also to the sub-
If gatherers intend to store their harvest, then sequent domestication of these starchy and oily
their decisions about how, when, and where to seeded plants. The ability to store plants and de-
process the food may be different. Decisions fer some labor was an advantage for the people
about how to handle foods destined for storage who collected, prepared, and ate those foods.
balance considerations about what, if anything, Processing/Cooking: Finally, we need to
is needed to prevent spoilage, whether the shelf consider how foods are cooked. What sorts of
life is longer for a minimally handled or derived vessels or other tools are required? Are there
product, and whether deferring processing evens multiple ways a food might be prepared and
workloads or solves scheduling conflicts. Re- consumed? Depending on the desired dish, pro-
moving nutshells or grain chaff at the collection cessing activities, cooking equipment, and labor
site may save transport costs but generally reduc- demands may vary. Cracking hickory nuts to ex-
es the time a food can be kept without spoiling tract the meat is a labor-intensive task, which re-
and eliminates the option of deferring labor. For quires a nutting stone, a pick, and patience. Hick-
example, to limit mold and insect damage, acorns ory nuts can be processed much more efficiently
must be dried or parched before they can be by pounding, dumping the mix of shell and meat
stored, but additional processing to make acorn into a vessel of water, boiling, and skimming to
flour, porridge, or soup can be delayed (Scarry, extract the oil. Or they can be pounded, sieved
2003). Thick-shelled hickory nuts are more im- to remove the larger pieces of shell, pulverized,
pervious to infestations; they can be gathered and then formed into kenuchee balls, which can
and stored with little or no processing. Indeed, be added to soups and stews. While the raw food
while the nuts may be placed near a fire to dry, is the same in all “recipes,” the equipment, labor
it is better to defer other processing because the demands—and quite likely archaeological signa-
shelf life of hickory nut products is significantly tures—vary (Swanton, 1946; Talalay, Keller, and
shorter than that of unshelled nuts (Scarry, 2003). Munson, 1984; Gardner, 1997; Fritz, Drywater
Once foodstuffs are in storage, there are further Whitekiller, and McIntosh, 2001; Thomas, 2008;
decisions to make. Are small quantities of foods Roger Cain, personal commun., 2010).
removed and processed as they are needed for
meals? Or are some episodically processed in SEASONAL PLANT USE
bulk generating several weeks’ supply? House- IN THE GEORGIA BIGHT
holds make such decisions based on the labor
they have available and on what other activities Constructing operational chains for multiple
require attention (Fuller, Korisettar, and Venkata- food items is a substantial challenge. If we are
subbaiah, 2001). going to incorporate plant foods more fully into
Because they can be stored for extended pe- existing models of seasonal subsistence rounds
riods, many plant foods have value beyond their for St. Catherines Island or the broader Georgia
immediate caloric returns. Storage can even out Bight, then there is a lot of work to be done. First
food availability and provide for anticipated lean and foremost, we need robust data about the plants
seasons. As noted above, storage can also even that were collected, cultivated, and consumed.
out labor demands and help solve scheduling Unfortunately, while there are detailed analyses
conflicts. Harvested nuts, grains, and fruits may of animal remains from multiple sites and tem-
need to be dried or parched to prevent spoilage, poral contexts from St. Catherines Island, the
but further processing can often be deferred un- only plant data come from a preliminary report
til there is a lull in other activities or the food on the mission assemblage (Thomas, 2008: 978).
Moreover, there have been far fewer systematic are high (Scarry, 2003). Despite the fact that pres-
analyses of plant assemblages from elsewhere on ervation of roots and tubers is generally uncom-
the Atlantic coast than from the interior South- mon, we have reports of greenbrier (Smilax sp.)
east. Thus, we lack the archaeological basis for (Thomas, 2008), Indian turnip (Arisaema triphyl-
assessing the relative importance of various plant lum) (Jones, Espenshade, and Kennedy, 1997;
foods or for building detailed models of plant use Scarry and Scarry, 1997), and an unidentifiable
on St. Catherines Island. We can, however, be- tuber (Detwiler and Keene, 2003) from sites in our
gin by compiling an inventory of plants that have inventory. This suggests that roots were important
been recovered from sites along the Carolina and resources for people living along the Carolina and
Georgia coasts. This gives us a list of plants that Georgia coasts. Digging roots in the spring requires
might have been used on St. Catherines Island knowledge about where they can be found even
and provides a starting point for thinking about when aboveground (or above water) leaves and
what foods were available to gather or take out of vines are absent. It is also labor intensive and po-
storage over the course of the year and what deci- tentially uncomfortable when reaching the plants
sions and tasks were associated with their use. requires wading in cold mud or water. Gatherers’
For this chapter, we have culled presence data activities and conversations while digging aquatic
from Ruhl’s plant list for Mission-period contexts or terrestrial roots likely disturbed game, fish, and
on St. Catherines Island, with the omission of Eu- fowl, making active hunting by the party unpro-
ropean-introduced species (Thomas, 2008: 978); ductive. Trips to the marsh side could, however,
from Hollenbach’s analysis of late prehistoric be combined with setting or checking traps and
plant remains from Keene’s (2004) excavations trotlines for fish, turtles, and crabs as well as with
of the Grove’s Creek site (9CH71) on Skidaway digging clams or other molluscs. Once gathered,
Island, Georgia (Detwiler and Keene, 2003); and some tubers (e.g., duck potato) can be cooked with
from 19 prehistoric or contact-era sites on the little processing, while others (e.g., arum) require
North Carolina coast that are reported in the con- pounding or grating before they are cooked (Mess-
tract literature (Scarry and Scarry, 1997) and in ner, 2011: 20–26). Regardless of the labor, roots
Kimberly Schaefer’s (2011) dissertation. Most of would provide a welcome source of carbohydrates
the sites date after people began to grow maize on at a time when there were few alternatives. In ear-
the coast, although six of the North Carolina sites ly spring, gatherers also could pick tender greens
also have earlier components. We do not present from plants such as chenopod, purslane (Portulaca
quantitative analyses or even calculate ubiquity, oleracea), and poke (Phytolacca americana) that
because only Grove’s Creek and seven of the favor open ground and sprout in dormant garden
North Carolina sites have plant assemblages that plots and other disturbed soils. Greens eaten raw
were systematically collected (using flotation) or stewed provide few calories but are important
and fully analyzed. Table 11.2 lists plants identi- sources of vitamins and minerals and undoubtedly
fied from the 21 sites. added welcome variety to meals.
Even a cursory inspection of table 11.2 is As spring progressed, people would monitor
informative. We use the tabulations of species the growth of grasses and the flowering of fruit
present and the probable collection dates for the and nut trees, anticipating and estimating the po-
various plant foods to sketch out a hypothetical tentials for future gathering excursions. We have
seasonal round of plant-related activities and de- scant evidence for spring-ripening grains such
cisions for forager-farmers living on the lower as maygrass, little barley (Hordeum pusillum),
Atlantic coast or one of the islands. or other grasses and sedges. Whether these were
much used at coastal sites remains to be seen. If
Spring spring grasses were cultivated or gathered, then
Few plant foods are available for collection in people must have made time for threshing and
the early spring. At this time gatherers likely con- parching the grains. Once people began growing
centrated on plants with underground storage or- maize, beans (Phaseolus vulgaris), and squash
gans such as arum (Peltandra virginica), duck po- (Cucurbita pepo), fields would need to be pre-
tato (Sagittaria sp.), and cattail (Typha sp.). Roots pared so that planting could begin by mid-April.
and tubers can be collected at virtually anytime, As they prepared for sowing and tended their ear-
but are best harvested in winter or early spring ly crops, women could continue to gather greens
when they are most tender and starch reserves from volunteer plants.
TABLE 11.2
Plants Identified from Sites along the Georgia and Carolina Coasts
Common name Taxonomic name Season available St. Cath- North
Grove’s
erines Carolina
Creek
Island coast1
Crops
Maize Zea mays summer x x x
Bean Phaseolus vulgaris summer/fall x x x
Squash Cucurbita pepo summer/fall x x x
Bottle gourd Lagenaria siceraria summer/fall x
Starchy Seeds
Little barley Hordeum pusillum spring/early summer x
Maygrass Phalaris caroliniana spring/early summer x x
Vetch Vicia sp. summer x
mid summer/late
Wild rice Zizania aquatica x
summer
Amaranth Amaranthus sp. late summer/fall2 x x x
Chenopod Chenopodium sp. late summer/fall2 x x x
Knotweed Polygonum sp. late summer/fall2 x
Wild bean Strophostyles sp. late summer/fall x x
Sedge family Cyperaceae mid summer/fall x x
Grass family Poaceae x x x
Oily Seeds
Bearsfoot Smallanthus uvedalius late summer/fall x x
Ragweed Ambrosia sp. late summer/fall x
Sumpweed Iva annua late summer/fall x
Sunflower Helianthus annuus late summer/fall x x
Nuts
Acorn Quercus sp. fall x x x
Hickory Carya sp. fall x x x
Walnut Juglans nigra fall x
Fruits
Blackberry Rubus sp. summer x x
Blueberry Vaccinium sp. summer x
Creeping cucumber Melothria pendula summer x
Grape Vitis sp. summer x x x
Maypops Passiflora incarnata summer x x x
Mulberry Morus sp. summer x
Plum/cherry Prunus sp. summer x
Elderberry Sambucus sp. late summer/fall x
Prickly pear Opuntia sp. late summer/fall x
Wax myrtle Morella caroliniensis late summer/fall x x

Common name Taxonomic name Season available St. Cath- North
Grove’s
erines Carolina
Creek
Island coast1
Black gum Nyssa sylvatica fall x x
Cabbage palm Sabal palmetto fall x
Saw palmetto Serenoa repens fall x
Palm family Arecaceae fall x
Persimmon Diospyros virginiana fall x
Dogwood Cornus sp. fall x
Sumac Rhus sp. fall x x
Yaupon holly Ilex vomitoria fall x
Hackberry Celtis sp. fall/winter x x
Roots and Tubers
Indian turnip Arisaema triphyllum all year3 x
Greenbrier Smilax sp. all year3 x
Tuber or root all year3 x
Greens and Miscellaneous
Bedstraw Galium sp. x x
Carpetweed Molluga sp. x
Copperleaf Acalypha sp. x x
Gromwell Lithospermum sp. x
Morning glory Ipomoea/Convolvulus sp. x x
Mustard Brassica sp. x x
Pokewee Phytolacca americana x x x
Purslane Portulaca sp. x x
Spurge Euphorbia sp. x x
Stargrass Hypoxis sp. x
1
Presence data derived from 19 sites along the North Carolina coast.
2
Season indicated is for ripe seeds. Greens would be available in spring/summer.
3
Available all year but highest starch content is in winter/early spring.
4
Sources: Abbott et al., 1999; Crites, 1999; Detwiler and Scarry, 1999; Gardner, 1984, 1990; Glazier, 1987;
Green, 1984, 1986; Jones, Espenshade, and Kennedy, 1997; Loftfield, 1979; Marshall, 1986; Payne and Dahlin,
1987; Scarry and Scarry, 1997; Schaefer, 2011.
Summer ties may have differed depending on the size of

During the summer, people would continue the nut crop that year.
to cultivate their crops. Depending on planting To prevent spoilage, acorns would need to
dates, maize, beans, and squash ripen from late be parched for storage but shelling, leaching,
June into August. Women would pick and cook and pounding into flour could be saved for lat-
some of the crops as soon as they were ripe. A er. Hickory nuts could be set in baskets near the
significant portion of the crop, however, was hearth to dry but other processing (see table 11.1)
likely left until fully mature. Maize would be al- could be deferred. Late fall fruits would include
lowed to dry on the stalk then picked; seed corn cabbage palm, saw palmetto, and persimmon.
would be set aside, and the rest prepared for stor-
age. Husking and shelling the maize would re- Winter
duce storage bulk but it would increase immedi- Winter would bring few new plant foods,
ate labor demands and decrease shelf life. Beans though roots, tubers, and persistent seeds and
would be picked and shelled or the vines might be fruits could be gathered. Most attention at this
pulled and threshed to release the seeds. Fleshy time of year would turn to husking, shelling,
squash would need to be cut and dried and the cracking, and otherwise processing and cooking
seeds parched or toasted. While waiting for their stored plant foods.
crops to ripen and after harvesting and preparing
them for storage, people could pick and eat or FINAL THOUGHTS
dry a variety of summer-ripening fruit including
blackberries (Rubus sp.), blueberries (Vaccinium Much of the seasonal round we describe
sp.), mulberries, grapes (Vitis sp.), plums, and could fit almost any location in the Southeast.
maypops (Passiflora incarnata). These have to For the most part, we lack qualitative and quan-
be gathered as soon as they ripen or they rot or titative data necessary to refine this picture for
are eaten by animal competitors. St. Catherines Island and distinguish the coast
from the interior. We can offer several thoughts,
Fall however. First, while there is considerable de-
Late summer and early fall would bring an bate about how heavily late prehistoric coastal
abundance of plant foods and work. Harvesting people relied on crops, maize is consistently
and preparing crops would continue into early present and sometimes abundant at coastal sites.
fall. Chenopod, wild rice (Zizania aquatica), Second, there are several plant foods, notably
sunflower, and bearsfoot (Smallanthus uvedalius) bearsfoot nutlets and the fruits of black gum,
would offer starchy and oily seeds that could be cabbage palm, saw palmetto, and wax myrtle,
harvested and stored to be processed and eaten which are seldom reported for interior sites but
later. Fruits such as elderberry (Sambucus sp.), are common in coastal assemblages. The dietary
saw palmetto (Serenoa repens), cabbage palm, role and processing requirements of these plants
wax myrtle (Morella caroliniensis), and black need attention. Third, our seasonal sketch fo-
gum (Nyssa sylvatica) would also ripen. Wild cuses on plant foods, but gatherers often collect
rice would need to be gathered before its seeds small game, shallow-water fish, and shellfish.
dropped and were dispersed by water. Elderber- During the summer and fall, people would have
ries would soon rot or disappear. Chenopod, the to weigh the merits of collecting fruit, nuts,
palms, wax myrtle, and gum have fruits that are turtles, shellfish, etc. When the mulberries or
persistent and less attractive to wildlife. Collec- blackberries were ripe did people forgo a trip to
tion of these could probably be deferred. the shore or berry patch? Or did they split their
In mid- to late fall, gatherers’ attention would work parties and gather both fruit and clams?
shift to nut mast with priority in collection and In this vein, we are intrigued by the data on
drying given to sweet acorns over thick-shelled shellfish collection (see chaps. 7, 8, and 10) that
hickories. Because of the dietary importance of seem to indicate collection of clams and oysters
these nuts, as well as their relatively short period in late fall and early spring when there were few
of availability, nut-collecting parties may have plants to harvest and gatherers may have been
been relatively large and included women, chil- collecting tubers from the marsh edges. Last
dren, and men in order to maximize the group’s but not least, St. Catherines Island, and quite
gains. The size and composition of these task par- possibly other island and coastal settings along
the Georgia Bight, presents an interesting case to juxtapose the plants and animals targeted by
for considering central-base foraging patterns. gatherers, gardeners, and hunters. To engage in
Thomas (2008) has argued that the size of St. such an exercise, however, we need additional,
Catherines Island and its geological and ecolog- detailed, quantified analyses of plant remains,
ical configuration allowed camps to be located such as Hollenbach’s (Detwiler and Keene,
where they would provide access to key terres- 2003) study of the Grove’s Creek assemblage
trial and aquatic hunting and gathering grounds and Schaefer’s (2011) analyses of coastal North
as well as agricultural soils without the need for Carolina assemblages.
seasonal relocation. People could have followed
our hypothetical round of plant handling from
marshside settlements without making more NOTES
than brief forays away from camp.
The seasonal round we have sketched is a
1. Recent experiments indicate that capuchins have a
far cry from what might be possible if we prac- preference for dietary diversity. When offered choices, they
ticed what we preached and examined opera- select a variety of items over a greater quantity of their fa-
tional chains and used optimal foraging models vorite food (Addessi et al., 2010).
2012 Making a Case for Coastal Subsistence Seasonality 199
CHAPTER 12
Making a Case for Coastal
Subsistence Seasonality
G regory A. W aselkov 1
This chapter represents for me something of sial) “Man the Hunter” symposium held in 1966
a rebirth as a coastal archaeologist. After disser- at the University of Chicago, Lee and DeVore
tation research on a Potomac River oyster mid- famously offered “two basic assumptions about
den and publication of a worldwide review of hunters and gatherers: (1) they live in small
shellfish gathering and shell midden archaeology groups and (2) they move around a lot” (Lee and
(Waselkov, 1987), my career took a decades-long DeVore, 1968: 11). Both generalizations pro-
detour into the ethnohistory of southeastern North voked criticism and an extensive literature now
America and the archaeology of French colonists documents great variability in group size and mo-
and historic Creek Indians. When an opportunity bility observed ethnographically among hunter-
arose a few years ago to do some developer- gatherers (Lee and Daly, 1999). Archaeologists
funded investigations of Woodland shell mid- participated in the ensuing debate in two ways,
dens on the Alabama coast, I felt akin to Rip van by developing the theoretical means to distin-
Winkle reawakening to a world changed greatly guish various kinds of residential mobility in the
in the interim. While there are surely drawbacks archaeological record and by recognizing that the
to such a circuitous career path, my return after archaeological record contains evidence for ap-
a long preoccupation elsewhere in the discipline proaches to mobility unobserved in the modern
has allowed me to see coastal archaeology with world (Yesner, 1980).
fresh eyes and envision some potentially fruitful Binford (1980, 1990) contributed substan-
paths forward. The invitation to join the Caldwell tially by contrasting the residential mobility of
V conference on seasonality, where participants foragers who move as a group to their resources
were immersed for a few days amid one of the with the logistical mobility of collectors who ex-
most productive and innovative coastal archaeo- ploit resources as individuals or specially orga-
logical projects underway anywhere in the world, nized task groups. He conceived of logistical and
offered the perfect venue for introspection. residential variability not “as opposing principles
Understanding seasonality seems, more than … but as organizational alternatives which may
ever, to be essential to unraveling the rhythms be employed in varying mixes in different set-
and rationales of lives lived in coastal landscapes. tings” (Binford, 1980: 19). Nevertheless, many
More than an end in itself, though, seasonality is archaeologists have employed the collector/for-
a portal through which we can grapple with ques- ager concepts as typological opposites or two ex-
tions of residential mobility, exploitative strate- tremes on a continuum without recognizing the
gies, and many other facets of human existence multidimensional nature of the organizational
along the world’s estuaries. A good place to be- possibilities they imply. Similarly Kelly (1992:
gin, therefore, is with a glance back at the shifting 50, 60) has criticized archaeologists for the ten-
significance placed on seasonality by archaeolo- dency to think in terms of a single scale of mo-
gists concerned with coastal settings. bility leading toward sedentism that conflates the
In the wake of the influential (and controver- independent variables of individual and group
mobilities (each with its own factors of age, gen- relatively sedentary village sites of hunter-gath-
der, kin relations, and abilities), territorial shifts, erers who occupied areas with abundant resourc-
and migrations. Among the many dimensions of es. In most parts of the world, it is now apparent,
mobility, he enumerates “seasonal movement of sedentism and cultural complexity preceded plant
the residential base camp, movement of individ- and animal domestication and in fact may have
uals around and between residences, movement been preconditions for serious reliance upon do-
of a group’s yearly range or aggregation site ..., mesticates. Not too many years ago, archaeolo-
and permanence of facilities such as houses and gists considered coastal resources less reliable
fish weirs” (Kelly, 1995: 148–149; cf. Builth, and less abundant than terrestrial and riverine re-
2006: 14–15). sources (Cohen, 1977; Osborn, 1977), but that er-
Ames has explored the transformative role roneous attitude was based on misunderstandings
of watercraft in aquatic resource exploitation, of estuarine environments, which are some of the
particularly the possibilities for food production most productive on earth (Rowley-Conwy, 1983,
by easing acquisition and transport, as in open- 1998; Brown, 1985; Marquardt, 1996; Costanza
ing new offshore habitats to exploitation and et al., 1997; Erlandson, 2001). Studies of cultural
expediting collection of large amounts of small- complexity in coastal locations are hampered to
sized taxa, which might not be a cost-effective some extent by Holocene marine transgressions
strategy without boats even over short distances that inundated early coastlines. But worldwide
(Winterhalder, 2001; Ames, 2002: 34, 47). Selec- evidence is revealing how substantial populations
tive resource exploitation assisted by watercraft exploited estuarine resources, in some cases cre-
could, of course, have seasonal mobility implica- ating large sites interpreted as permanent, year-
tions by enabling access to widespread habitats round villages of coastal hunter-gatherers orga-
through broad dispersal of task groups or village nized in complex societies (Russo and Quitmyer,
movement. In some watery environments, canoes 1996; Russo, 1998; Keene, 2004; Thompson and
may at times have so transformed mobility that Andrus, in press). Archaeologists need to under-
they became the residence, the ultimate in mobile stand how coastal strategies differed from inte-
central-place foraging. rior strategies, which in so many instances led to
A number of researchers have noted that reliance on agriculture. Brian Hayden concisely
gender and age distinctions may be reflected in states the dilemma:
archaeological food remains at coastal sites. In
modern societies where shellfishing is conducted One of the major conundrums of the do-
at a domestic nonindustrial scale, women and mestication problem … is why domestica-
children are the most active shellfish gatherers tion did not occur in certain neighboring
(Meehan, 1982; Waselkov, 1987: 96–99; Claas- regions or areas that seem to have been
sen, 1998: 175–182; Klein, 1999; Bird and Bliege just as well endowed and as rich and envi-
Bird, 2000). Ethnographic studies have repeated- ronmentally variable as the initial hearths
ly documented such age and gender disparities in of domestication. In the eastern United
foraging and collecting emphasis, with important States, for example, while people in the
dietary implications (particularly in protein and Ohio and neighboring valleys actively
fat contributions to diet) for segments of society domesticated a number of species, noth-
with limited access to hunted sources of meat (Bi- ing appears to have happened around the
galke, 1973; Meehan, 1982; de Boer, Pereira, and Louisiana floodplain or the Gulf Coast
Guissamule, 2000; Thomas, 2002; Bird, Bliege region until the introduction of efficient
Bird, and Richardson, 2004; cf. Voorhies, 2004: maize cultivation thousands of years later.
129–141, for a nondomestic example). Precisely California provides another example of a
how this generalization translates to specific sea- problem area with many diverse habitats
sonal mobility patterns remains an open question and great resource variability. Australia
for archaeologists, but one that should be consid- is a problem area of even greater magni-
ered in the development of seasonality models. tude…. (Hayden, 1995: 297)
Coastal seasonality is playing an important
role in reevaluations of the origins of cultural Archaeologists working along coastlines are
complexity. In recent decades, archaeologists rethinking this issue. In California, for instance,
have found early evidence of complexity in the Kroeber’s long-accepted view, that the natu-
ral complementary bounties of oak groves and sion of Mississippian maize farmers around a.d.
coast made domesticates unnecessary, is being 1250 (Brown, 2003). These thousands of years
replaced by a more realistic grasp of local envi- of nonagricultural existence were not insignifi-
ronmental uncertainties and a better appreciation cant footnotes to the rise of agriculture. Rather,
of settlement and subsistence strategies that re- they represent remarkably effective alternatives
lied on diversification rather than specialization to the reliance on domesticates practiced by their
(Kroeber, 1925: 919–926; Raab, 1996; Glas- neighbors and constitute an important but ne-
sow, 1997; Luby, Drescher, and Lightfoot, 2006; glected part of the human past that we need to
Jones et al., 2008; Lightfoot and Parrish, 2009). understand on its own terms.
Builth (2006) argues for the development of so- But my task here is to focus on the problems
cial complexity along the temperate south coast facing archaeologists who wish to understand
of Australia based on evidence of landscape man- coastal seasonality. Seasonality of resource use
agement and the development of a storage econ- and residential patterns is fundamental to many
omy and trade in wild foods. Cases like these significant human behaviors. Managing or ac-
begin to dismantle long-standing assumptions commodating seasonal aspects of plant and ani-
about the rise of social complexity and challenge mal physiology and habits underlies such diverse
the presumed exceptionalism of California and behaviors as rites of passage, calendrical rituals,
Australia as ethnographic oddities, places where political and economic organizations, harvest
people inexplicably rejected agriculture and the cycles, architectural designs, settlement patterns,
monumentality that often followed the rise of so- and the density and size of human populations.
cial complexity elsewhere in the world. Seasonality is fundamental to these phenomena
I suggest that coastal regions routinely sup- because most of the organisms important for hu-
ported nonagricultural forms of resource use and man life respond to cycles of temperature, mois-
permitted different trajectories of social complex- ture, and other variables that define seasons.
ity than are documented in noncoastal regions Unraveling the complexities of seasonality
(also see Hayden, 1990; Jochim, 2006). One of using archaeological data has proven extremely
Binford’s principal conclusions in Constructing difficult. It is not uncommon for modern studies
Frames of Reference, his massive analysis of to extrapolate overall seasonality of resource use
ethnographically documented hunter-gatherer for entire sites from evidence provided by one
behavior, is his observation that subsistence in- or two species (each typically represented by a
tensification by hunter-gatherers eventually leads handful of specimens). In simplest form, season-
to one of two outcomes: dependence on aquatic al preferences of species present in a sample are
resources or dependence on plant domesticates cited to establish which seasons are represented
(Binford, 2001: 201, 357, 368, 443–444). In the in the assemblage and which are not, thereby
absence of evidence for agriculture, archaeolo- linking seasonal behavior of plants and animals
gists must refocus their energies on interpreting with human residential patterns. This approach
those plant and animal remains that reflect not to deriving seasonality and reconstructing en-
only the changing targets of use, but the chang- vironments relies on analogy to modern spe-
ing seasonalities of exploitation and settlement. cies physiology and behavior, and on a further
Hayden’s comment above, that “nothing appears (doubtful) assumption, that climates, ecosys-
to have happened around the Louisiana flood- tems, and organisms have remained unchanged
plain or the Gulf Coast region until the introduc- throughout the Holocene. This approach fails to
tion of efficient maize cultivation thousands of capture the complexity, diversity, and inherent
years later,” of course reflects our collective ig- flexibility of organisms, including people, on
norance of what did happen in those areas during a daily, seasonal, annual, and long-term basis.
several thousand years of increasing population Nonetheless, this remains the dominant archaeo-
density and social complexity. We now know the logical method to assess season of site occupa-
earliest mound ceremonialism in North America tion in many parts of the world.
appeared in the Mississippi floodplain around In addition to problems inherent in using
5300 b.p. (Saunders et al., 1997), and the north- modern ecological analogies, other problems
ern Gulf supported a series of vibrant nonagri- exist in assessing seasonality at coastal archaeo-
cultural societies culminating in Weeden Island, logical sites using species presence/absence,
which disappeared coincident with the expan- quantified or not. In those cases where more
than one site is studied, neither site locations nor round, exchange network, kin group, or some
artifact types provide the precision needed to other economic or social unit in order to assess
verify or refute the presence of an asynchronous variability in coastal settlement patterns. Ideally
residential pattern due to seasonal availability these methods ought to be performed in tandem
of resources on a regional scale (e.g., Thomas, at coastal and inland sites to explore entire settle-
2008). Radiocarbon dating has not yet provided ment patterns and understand how sites of dif-
fine enough temporal control to establish two ferent sizes and functions in different locations
or more sites as part of a seasonal round, or fit into the overall landscape. In practical terms,
the contemporaneity of complementary special however, as expressed by Bar-Yosef and Rocek
function sites, or the sequential reuse of a site (1995: 2), “the likelihood of recovering all of the
on a year-to-year basis. (However, see Kennett sites in a settlement round is vanishingly small
and Culleton, chap. 2, this volume, for a pos- ... we must give up the idea of reconstructing the
sible resolution of the problem.) entire settlement pattern” of any ancient society. I
Faunal evidence is equally challenging. Many share their skepticism on that score, given current
animals are (and were) available throughout the limits of analytical methods, disparities in fau-
year at different places along subtropical and nal preservation often observed between coastal
tropical coasts. Although they may move, repro- and inland sites, and losses to modern develop-
duce, and grow seasonally, the broad behavioral ment that disproportionately impact the world’s
and physiological patterns visible in modern coastal sites. Yet I maintain that we should apply
species are often unclear from archaeological our best analytical procedures more rigorously
taxonomic identifications, element distributions, and more broadly to learn, as best we can, how
measurements, ages at sexual maturity, and sea- patterns of subsistence and residential mobility
sons of death, particularly when the ancient prey changed over time.
lived in different environmental conditions or You may take issue with my assessment of the
different climatic regimes than exist currently. progress archaeologists have made in determin-
Ethnobotanists have similar problems. It does not ing seasonality of site occupancy, especially in
help that fruits and grains have “seasons” when light of recent widespread advances in the analy-
they are ripe, because in tropical and subtropi- sis of growth structures in shellfishes, fishes, and
cal environments these seasons are fluid and can mammals. In fact, I am heartened that archaeolo-
only be broadly defined. Both zooarchaeologists gists are developing ever more powerful tools to
and ethnobotanists argue, too, that identified taxa establish seasonality of capture, and many of the
could have been caught or collected in one season Caldwell V conference participants have con-
and stored for later use, or processed elsewhere in tributed to this effort (e.g., Quitmyer, Jones, and
such a way that no evidence of their use reached Arnold, 1997; Waselkov et al., 1998; Andrus and
the site being studied. Crowe, 2000, 2008; Quitmyer, Jones, and An-
In sum, after more than a century of archaeo- drus, 2005; Reitz, Andrus, and Sandweiss, 2008;
logical effort, we cannot yet prove that people Thomas, 2008; Culleton, Kennett, and Jones,
lived at two different sites within the same year, 2009). But most of this progress has been spe-
let alone parse out the precise season in which cies specific. The biology of each target species is
each settlement was occupied, for what purpose, complex enough that the various techniques (e.g.,
and by how many people. However, our inabil- those based on periodic changes in ontogenetic
ity to replicate the kind of observations made by growth structures, oxygen isotopic profiles, pop-
ethnologists does not mean that archaeologists ulation availability due to migrations) have had
cannot address the seasonal nature of foraging to be carefully crafted and tested in multiple loca-
behavior in the past. We do need to face up to tions to demonstrate their efficacy. Consequently,
current limitations of our analytical procedures, few archaeologists have applied more than one
reassess the possibilities of our data, and consider or two of these new analytical techniques to an
where we can make better progress. assemblage, and they typically examine very few
It is increasingly obvious that multiple lines specimens per site. We, as a profession, have
of evidence (e.g., from radiocarbon dating, isoto- so far missed the opportunity to gather a broad
pic profiles, zooarchaeology, paleoethnobotany) range of evidence for complex seasonal activities
ought to be pursued with equal rigor on a variety undertaken by people who could plan and carry
of sites that might have been part of a seasonal out more than one task at a time, different tasks
in rapid succession, and tasks that changed in re- Analysis of stable isotope ratios of carbon and
sponse not only to seasonal variation but to envi- nitrogen in organic remains is based on their roles
ronmental fluctuations at longer time scales. in two major biogeochemical cycles, the carbon
I will mention just two areas of inquiry I think cycle and the nitrogen cycle. A great many bio-
are amenable to immediate improvement: (1) logical and chemical processes lead to variation
seasonality modeling that considers evidence of in the relative proportions of these isotopes (for
ontogenetic movement (that is, during the life- the basic science, see Hedges, Stevens, and Rich-
time of an individual animal) between habitats; ards, 2004, 2005; Dawson and Siegwolf, 2007;
and (2) increasing resolution and confidence in Ferrio et al., 2007; Pollard et al., 2007: 176–189;
seasonality estimates by statistical analysis of Crawford, McDonald, and Bearhop, 2008). In
oxygen isotope and other temperature proxies. particular, differences in human and animal bone
collagen ratios of 12C to 13C (δ13C) and 14N to 15N
Habitat Analysis (δ15N) have been interpreted as dietary indicators
(DeNiro, 1985; Bocherens et al., 1999; Evershed
While oxygen isotope analysis has gained et al., 2007; Goffer, 2007: 307–309; Koch, 2007).
wide acceptance as a reliable means of determin- Both are closely tied to trophic level, so, for in-
ing season of capture for animals with calcium stance, carnivores are enriched in 15N compared
carbonate structures, the potential for stable iso- to their herbivore prey. Terrestrial herbivores
tope and elemental analyses to provide informa- from different habitats are frequently distinguish-
tion on location of capture is barely explored, able on the basis of δ13C and δ15N values, which
although this is obviously an aspect of season of fluctuate in the plants upon which they feed. Ar-
exploitation (see Ishimaru et al., 2006; Lynch, chaeologists have mostly used this approach to
Hamilton, and Hedges, 2008). Biologists are ac- document the importance of domesticates, par-
tively pursuing this line of inquiry for their own ticularly maize, in the diets of humans (or, indi-
purposes (such as tracking migrations or trac- rectly, in their dogs; White et al., 2001), but there
ing habitat environmental changes; see Hobson, are other potential applications.
1999), and some of their findings and procedures Five deer bone samples—three from the
have relevance and applicability to archaeology. coastal Bayou St. John site and two from the Late
Though not necessarily aimed squarely at sea- Woodland Corps site situated inland in the Mo-
sonal questions, determining habitat of acquisi- bile-Tensaw delta—were submitted to Beta Ana-
tion can inform seasonal interpretations in many lytic for collagen extraction and δ13C and δ15N
contexts. Archaeological application is still ex- analysis using procedures (Brown et al., 1988)
perimental, but I think entirely feasible, as I’ve standard for AMS radiocarbon dating of bone.
tried to demonstrate with a small pilot study fo- The results (fig. 12.1) indicate 13C depletion and
cusing on white-tailed deer that examines carbon 15
N enrichment in bones from the inland site com-
and nitrogen isotope analysis of bone collagen. pared to two bones from the coastal site. A third
My original intent was to seek evidence of bone from the coastal site isotopically resembles
transport of deer elements between inland and the inland bones, suggesting transport of some
coastal locations. Suspicion that such transport deer parts to the coast, as anticipated. With addi-
occurred was raised by the presence at the Late tional testing, this approach may prove a valuable
Woodland Bayou St. John site on the Alabama analytical tool for investigating the transport or
coast of hundreds of tools, mostly awls and spat- exchange of subsistence materials between set-
ulas, made almost exclusively from deer metapo- tlements, data that might otherwise be interpreted
dials (Price and Waselkov, 2009). Analysis of all as evidence for seasonal task group activity.
deer bones, worked and unworked, from Bayou A second approach to habitat analysis, ap-
St. John revealed disproportionate numbers of plicable to shellfish valves and fish otoliths, in-
metapodials compared to other skeletal parts. volves elemental and isotopic analysis by laser
Since movement of subsistence remains has ablation inductively coupled plasma mass spec-
important implications for the interpretation of trometry (LA-ICPMS) to obtain information on
site subsistence, seasonality of occupation, and ontogenetic habitat changes of individual mobile
regional economies, I devised an experiment to fish (with their own seasonality implications) and
distinguish bones of coastal deer from bones of to identify exploited habitats of mobile and ses-
deer hunted at inland locations. sile shellfish. Elements of interest include Ca,
Mn, Sr, Ba, Pb, the lanthanide rare earths, Th, and Gulf of Mexico, Livingston (1975, 1982, 1985)
U. Oxygen and carbon isotopes have also been pioneered studies of estuaries, which he found to
used for habitat studies. be not only geologically unstable, but highly dy-
Multiple elemental and isotopic analyses of namic in terms of seasonal and annual variations
shellfish valves primarily reveal data on habitat in temperature, salinity, nutrient flow, and grass-
of capture. Archaeologists previously have had bed compositions. Most fish become increasingly
little access to this kind of information. For in- specialized in their food habits as they develop
stance, surface features and valve shape (“eco- in estuarine nurseries, migrate to the open gulf,
morphism”) of oysters have traditionally been and return to the estuaries as adults. Livingston
regarded as indicative of habitat (i.e., bed, sand showed a progression of species entering and
bottom, channel, and reef) (Galtsoff, 1964: 21– leaving estuaries through the year, and others
32; Kent, 1988: 28–38), but these can be ambigu- have noted seasonal habitat preferences of differ-
ous, and most species do not vary by habitat in ent species (Rozas and Zimmerman, 2000; Rozas
simple morphological ways. Some biological and Minello, 2001).
research suggests that Ba/Ca profiles may reveal In recent years biologists have explored stable
differences in environment of origin, and δ18O isotope relationships to these variables and have
analyzed for salinity may prove enlightening be- found, for instance, that δ15N and δ87Sr are en-
cause salinity in estuaries should correlate with riched in fish moving from estuarine to deep wa-
distance from passes to open water (Surge, Lo- ter habitats, and δ13C is enriched in seagrass habi-
hmann, and Dettman, 2001; Barats et al., 2009; tats (Hobson, 1999: 320–321; Nagelkerken and
Hobson et al., 2009). van der Velde, 2004). Thus a chemical signature
Analysis of fish otoliths is likely to lead in un- record of an animal’s habitat changes is recorded
expected directions. Within the last half century, in otoliths. A combination of light-stable isotope
biologists have built ever more complex fish life analysis (δ13C, δ15N, δ18O) and trace element anal-
histories documenting migrations between shal- ysis is likely to reveal information about habitats
low coastal estuarine habitats and open waters. In preferred for exploitation. Identifying collection
my research universe on the northern coast of the habitats, and not ontogenetic movements, is typi-
-22.60
1BA21
-22.40
1CK56
-22.20
-22.00
δ13C (0⁄00)
-21.80
-21.60
-21.40
-21.20
-21.00
-20.80
+0.00 +1.00 +2.00 +3.00 +4.00 +5.00 +6.00 +7.00
δ15N (0⁄00)
Fig. 12.1. Results of a trial carbon and nitrogen isotope analysis of deer metapodial bone collagen samples
from two Late Woodland sites, the coastal Bayou St. John site (1BA21) and the inland Corps site (1CK56).
cally the focus of archaeologists, since they re- of harvest so conservatively because the natu-
late most directly to human residential mobility. ral variation was not well understood. Appar-
But archaeological data on changes in fish spe- ent differences in paleotemperatures at harvest
cies habitat preferences may correlate with data were interpreted as evidence of harvest across
being gathered by marine biologists (see France, seasons. But modern control studies of shellfish
1995; Hanson, Koenig, and Zdanowicz, 2004; harvested monthly at many locations now thor-
Dorval et al., 2004; Kennedy et al., 2005; Martin oughly document how a single day’s harvest can
and Thorrold, 2005; Surge and Walker, 2005; Co- be represented by a substantial range in δ18O
myns et al., 2008; Fodrie and Herzka, 2008; Fod- values. Clearly we need to refine our seasonal-
rie et al., 2010; cf. Chittaro et al., 2005, 2006). In ity interpretations in light of larger-scale mod-
collaboration with Betsy Reitz and Fred Andrus, ern control and archaeological studies, to narrow
I have proposed a trial study of hardhead cat- our seasonality estimates for specific collection
fish and spotted seatrout otoliths. Catfish otolith events rather than continue to follow an overly
chemistry is expected to reflect a limited habitat conservative approach that yields three-month or
range of high-salinity mud/sand flats and chan- six-month interpretations of archaeological de-
nels, while seatrout could have been collected posits that, on the basis of other evidence, must
from a wider range of seagrass, oyster reef, chan- have accumulated much more rapidly.
nel, and salt-marsh habitats. In this sort of study, It is worth recalling that the preferred archae-
oxygen isotope analysis will enable us to cali- ological contexts for seasonality sampling were
brate the seasonal timing of habitat exploitation identified decades ago (Koike, 1979; Deith, 1986:
(reflected in other chemical signatures). The two 69; Waselkov, 1987: 142–144; Stein, 1992: 77;
analytical methods work in concert, which makes Stein, Deo, and Phillips, 2003; Claassen, 1998:
this a very powerful interpretive approach. As an 152). Gorski (2005) defined the smallest unit of
ancillary benefit, gaining a means of analyzing analysis as the microstratum, the smallest visible
shellfish and fish remains for habitat of capture “natural” stratum (a pocket or lens of midden).
should finally enable archaeologists to document In deposits consisting largely of mollusc valves
and quantify dependency on canoe transport. these are often interpreted as single episodes of
cultural deposition—the contents of a fish boil or
Statistical Analysis debris from a single shellfish roast or some other
processing event, discarded as primary refuse and
Climatologists (e.g., Grimes et al., 2003) forming a discrete feature representing a moment
have suggested a need to apply statistical pro- in time. These micro contexts rightly turn our at-
cedures to seasonality data, something both bi- tention from a site as a generalized whole to the
ologists and archaeologists have been reluctant individual activities that, multiplied a thousand
to do. δ18O data profiles of individual mollusc fold, created a site’s deposits. By focusing analy-
valves and otoliths are generally interpreted in- sis on micro contexts, they can give us meaning-
dividually and very conservatively by attribution ful insights on real behavior and help us move
to broad seasonal divisions: spring, summer, fall, away from the deceptively simple (and simply
winter. I suspect this hesitancy to employ statis- deceptive) generalized “behavior” we think we
tics is a consequence of the very small sample see when our unit of seasonality analysis is an
sizes available until recently in most seasonality accreted midden or an entire site.
studies. In the past it has not been uncommon for By comparison to modern control studies,
seasonality interpretations to be based on tem- we should be able (simplifying greatly) to calcu-
perature proxy curves obtained from samples of late probability values that correlate increments
one, two, or three shells per stratum or even per of the δ18O proxy curve to water temperature
site. As our analytical methods mature and we values. When we compare probability curves
acquire suites of data from larger archaeological between species, the problem becomes more
samples, we are now beginning to see that tem- complex. However, we have an archaeological
perature data profiles for individuals of the same precedent for such an analysis in the method
species differ in ways that at least in part reflect used to calibrate radiocarbon years to calendar
natural variation around a mean with a calcula- years, which is an application of Bayesian sta-
ble standard deviation. Earlier studies that relied tistical analysis (see Buck, Litton, and Smith,
on a few shells tended to interpret seasonality 1992; Buck, Cavanagh, and Litton, 1996; Buck
and Millard, 2004a; Whittle and Bayliss, 2007). I Specifically, we need to apply traditional zooar-
have only begun to develop this idea, but Bayes- chaeological methods, extensive oxygen isotope
ian statistics offer a way to combine evidences analyses (correlated by statistical analysis), in-
of seasonality from our multiproxy studies. The cremental analysis, morphometrics, and related
protocol of monthly interval sampling followed approaches to well-dated assemblages. Our goal
in the most exacting oxygen isotope analyses al- should be to apply and refine the best techniques
lows conversion to rank-order data suitable for currently available to adequately sample and pre-
modeling (Heuzé and Braga, 2008). A very early cisely define and model with confidence the sea-
application of Bayesian analysis to season of sonal component of coastal settlements, provid-
capture questions directed at fish otoliths (Eng- ing a platform upon which subsequent studies of
lish and Freeman, 1981) demonstrates the appro- other cultural variables and other environmental
priateness of the approach, but modern computer processes can be built.
modeling of micromilled samples promises much
finer seasonal resolution with an estimate of as-
sociated uncertainties (cf. Parnell et al., 2008). NOTES
In sum, we need thorough investigations of
1. Portions of this chapter are based upon work sup-
subsistence seasonality in specific locales oc- ported by the National Science Foundation under collab-
cupied for limited periods of time to test the orative research grants BCS-1026166, BCS-1026167, and
limits of our current range of analytical skills. BCS-1026168.
2012 Discussion 207
CHAPTER 13
Discussion
E lizabeth S. W ing
The Fifth Caldwell Conference was held in a growth and reproductive cycles of game ani-
gorgeous setting free from commercial distrac- mals, and fruiting of plants. These present-day
tions, making it a pleasant and intellectually stim- observations will, of course, show fluctuations
ulating experience. The intent of the conference from year to year in response to environmental
was to provide a forum for discussion of differ- changes such as temperature and rainfall varia-
ent aspects of the archaeological research being tion, much as they did in the past. Consequently,
conducted on the island and directed by David it may not be possible to estimate seasons of
Hurst Thomas over many years. The research has harvest with great precision. However, the more
involved a large number of students and archae- we know about variability in life histories of the
ologists committed to detailed longitudinal stud- plants and animals used by people, the better our
ies of the full array of materials excavated from understanding of human adaptations to the dy-
archaeological sites on the island. St. Catherines namic coastal setting.
Island was occupied for 5000 years, from the two People are generally flexible in adapting to
Late Archaic period shell rings right up to the changes in resource availability (Reitz, Quitmy-
17th-century Mission Santa Catalina de Guale er, and Marrinan, 2009). Despite environmental
and then into the plantation era. Archaeological variations, “data demonstrate the antiquity, flex-
deposits on this coastal island provide evidence ibility, and richness of a well-established dy-
for changing environmental conditions and suc- namic coastal fishing and hunting tradition in the
cessive cultures of people who inhabited it. southern Georgia Bight that existed for millennia
The focus of this year’s conference was sea- before the 17th century” (Reitz et al., 2010: 76).
sonality and mobility along the Georgia Bight. The persistence of the Guale tradition is demon-
To be most effective in approaching these issues strated by faunal assemblages deposited during
the research requires a multidisciplinary ap- the Mission period indicating that the “Spaniards
proach marshalling the specialized disciplines of altered their diet toward local Guale subsistence
biologists, zooarchaeologists, archaeobotanists, patterns far more than Guale members of the
geochemists, sclerochronologists, and, of course, community altered theirs to conform with Span-
archaeologists. Such research requires detailed ish practices.” (Reitz et al., 2010: 131).
study of the local conditions of the island and its Timing of events such as those rituals dic-
surrounding waters as well as year-long studies tated by ecclesiastical calendars requires preci-
of seasonal changes in plant and animal growth sion to the month and day of an archaeological
to provide baseline data for interpreting evidence deposit and may not be possible based on season-
of seasonal changes seen in the archaeological ality studies alone. Evidence of a feast in an ar-
remains of animals. The baseline studies would chaeological deposit might correspond to a ritual
include, but not be limited to, shellfish incremen- event celebrating Christian commemoration of
tal growth patterns and sizes, fish species com- the holy days of Christmas or Easter. The sig-
positional changes throughout the year, annual nature in a deposit for a feast has to be carefully
evaluated though it might be expected to include availability of a whole array of resources. Know-
a large source of meat such as pig (Sus scrofa) or ing when in a normal year buck deer shed their
white-tailed deer (Odocoileus virginianus) and antlers, or fawns are born, or persimmons (Dio-
accompanied by sweet fruits and nuts. spyros virginiana) ripen, provides clues about
Many challenges exist in documenting the the season of capture or collection when the
seasonal foraging patterns of the past and, there- remains are recovered from archaeological de-
fore, it is particularly important to have the posits. Sarah Bergh also presented information
opportunity provided by this conference to ex- on multiple seasonal indicators including data
change ideas and describe new and innovative from hard clam sclerochronology, deer fusion
approaches to the complex issues of seasonal- sequences, and fish age and size classes, which
ity. Eleven papers, all involved with efforts to the author suggests indicates year-round forag-
understand seasons of capture or collection of ing of these animals. In addition to advocating
different resources based on material excavated use of multiple lines of evidence, Fred Andrus
from archaeological deposits, were presented at and Gregory Waselkov have legitimate concerns
the 2010 conference. Most importantly, the past about adequacy of the sample sizes and the cost
conferences have resulted in four publications in of some analyses.
the American Museum of Natural History An- The conference clearly achieved its purposes
thropological Papers series. The papers present- in stimulating lively debate and discussion about
ed at this conference are now published in this the accuracy of the methods used to estimate
fifth volume of the series. These are landmarks seasonal acquisition and uses of resources. It is,
in the progress of research on the archaeology of therefore, necessary to take the next step, as many
St. Catherines Island. of the authors have, and apply these techniques to
Five of the papers are concerned with scle- gain a better understanding of the seasonal round
rochronology of the growth increments in the of the people who lived on the island. The occu-
shells of bivalves or otoliths of fishes. Irvy Quit- pation at some sites was apparently of short dura-
myer and Douglas Jones applied these studies tion. However, animal remains from these sites
to the hard clam (Mercenaria mercenaria) and show use of resources throughout the year, in-
validated the season of the increments by oxy- dicating long-term occupations during each cul-
gen isotope analysis. Nicole Cannarozzi studied tural period (Thomas, 2008: 878). Though clam
growth increments with stable oxygen isotope sclerochronology data indicate that clams were
analysis along the hinges of oysters (Crassostrea gathered predominantly during the winter (mid-
virginica) and supporting evidence from the December to mid-March), a few clams were also
growth measurements of the impressed odos- gathered during other parts of the year (Thomas,
tome (Boonea impressa). Carol Colaninno ap- chap. 1, this volume: fig. 1.3). When sources of
plied the techniques of sclerochronology to seasonal information such as age of deer and the
otoliths of hardhead catfish (Ariopsis felis) and composition of the fish fauna in addition to the
Atlantic croaker (Micropogonias undulatus). clam data are examined and combined for all cul-
Douglas Kennett and Brendan Culleton applied tural time periods, these catches appear to gradu-
a Bayesian statistical framework for determin- ally decline through the year (seasonal indica-
ing site seasonality and contemporaneity. They tors of winter catch are 32%, spring 28%, sum-
advocate combining stratigraphic information mer 24%, and fall 17% [Thomas, 2008: 878]).
with multiple AMS 14C dates. They apply this These samples are small, especially for the St.
approach to the examination of the occupations Simons period, and need further substantiation.
of the two Archaic shell rings, St. Catherines and However, if this is an accurate trend, one might
McQueen, and their contemporaneity. expect a relatively greater dependence on plant
Once clear criteria for seasonal changes are resources collected during the summer and fall.
established, most participants advocated the use Edible plants reported from the Mission era de-
of multiple lines of evidence to better understand posits are: chenopod (Chenopodium sp.), grape
seasonal changes in foraging resources during (Vitis sp.), blackberry (Rubus sp.), and elder-
pre-Hispanic and Mission periods. Multiple lines berry (Sambucus sp.), available in the summer,
of evidence can strengthen conclusions or show and maize (Zea mays), acorns (Quercus sp.), and
variation in seasonal indicators. Elizabeth Re- hickory (Carya sp.) nuts available in late sum-
itz and Margaret Scarry examined the seasonal mer and fall (Ruhl, 1993: table 15-11).
2012 Discussion 209
The difficulty in integrating data from plants garden hunting which might have made hunting
with those from vertebrates and invertebrates more successful. The maize crop would also be
lies in the relatively poorer preservation of plant subject to plunder from many animals and guard-
remains. The only hopes for recovery of plant ing the produce would be garden hunting in the
remains are when they were deposited in water- original sense of the term (Reitz et al., 2010).
logged conditions of wet sites such as a well During the Spanish period, sea catfishes (Arii-
or when they have been burned, which renders dae) were very abundant in the faunal samples.
them into bits of fragile charcoal. The anatomi- As scavengers, they were probably attracted to
cal features of the plant may still be preserved the Spanish mission period disposal of trash in
in the charcoal. Integrating these data with those the estuary, another form of garden hunting (Re-
from the generally robust remains of shell and itz et al., 2010: 162).
bone is difficult. Nevertheless, some assessment Access to fishery resources greatly expands
of the relative contribution of all remains to the the carrying capacity of the dry land because
past diet provides a more complete view of con- tides and currents bring nutrients to fishing
ditions in the past. grounds from coastal waters (Odum, 1971).
Other plants that must have been used but Likewise, domestic animals such as pigs con-
whose remains are invisible are fiber plants es- centrate nutrients making them available to
sential for making nets, baskets, and weirs. These people who consume the meat. Few remains
must have been used to catch the small fishes and of domestic mammals were recovered from the
those species that are reluctant to bite a hook such Mission period deposits, however (Reitz et al.,
as mullet (Mugil sp.) found in the faunal remains. 2010). Most surprising is the scarcity of dog
Some pottery has net impressions on it, further (Canis familiaris) remains in the deposits. Dogs
substantiating the presence of netting in the past would be expected in both pre-Hispanic and
(Royce Hayes, personal commun.). There is a Hispanic deposits and might have been useful
long tradition by the Gullah people living on the hunting companions and assistants.
Georgia Sea Islands for making baskets out of All of the remains discussed so far are based
sweet grass (Muhlenbergia filipes) stitched to- on excavated archaeological deposits. Every ar-
gether with palm leaves (Sabal palmetto). These chaeologist wishes at some point to have a first-
plants are available, as is bear grass (Nolina sp. hand look at the activities of people who lived in
or Yucca sp.), which may have been used for net- the past. Documents written by Jesuits during the
ting (Royce and Christa Hayes, personal com- Mission period on the island described the situa-
mun.). Weirs and baskets might also have been tion as the most miserable thing ever discovered.
made using cane (Arundanaria gigantea), which The impression was that the people wandered
apparently became quite abundant in response to about and that the soil was too poor to support a
clearing during the Mission period (Donna Ruhl, crop. The Jesuit mission attempt coincided with
personal commun.). Using fiber plants to make an extreme and extended drought, which resulted
baskets and nets would have been essential to in a deteriorating resource base. The disruption
catch small fishes, and to carry and store plant of the normal subsistence strategies and the de-
and animal products. mands of the mission resulted in deteriorating
In addressing issues of mobility, many im- health, disease, social and physiological stress,
portant resources known to have been used and demographic collapse (Reitz et al., 2010:
throughout the year would have been within easy 133). The short-lived Jesuit mission was fol-
reach from the forest and shore margin. Or they lowed by a more sustained Franciscan mission.
would have been within the economic foraging The Franciscan view of the Guale people was
distance of 10 km (Thomas, 2008: 245). Further- quite different. They were portrayed as living in
more, transport would have been much easier ranked society in sedentary towns and cultivat-
with nets and baskets. Along the marine edge, ing maize (Thomas, chap. 1, this volume). These
access to fishes, shellfish, diamondback terrapin two very different eyewitness accounts may both
(Malaclemys terrapin), raccoon (Procyon lotor), be accurate and reflect the conditions the writers
and white-tailed deer is documented (Reitz et saw during the very difficult climatic disruption
al., 2010). During the fall, the mast crop would while the Jesuits were on the island and the re-
have attracted both people and mammals such as turn to more normal, less threatening time during
deer and raccoon, creating an opportunity akin to the Franciscan mission (Thomas, 2008).
The expansion of archaeological techniques of fish faunal assemblages are providing ever
such as those examining seasonal change in ani- more complete understanding of human con-
mals discussed at this conference and the further ditions in the past. Eyewitness accounts are,
integration of data from such diverse sources as of course, most valuable to have but must be
plant charcoal, tree rings, growth increments “ground truthed” by archaeology for biases and
of mollusc shells and fish otoliths, and remains partial understanding.
2012 REFERENCES 211
REFERENCES
Abbott, L.E., Jr., E.E. Sanborn, L.E. Raymer, L.D. Island, Georgia. II. The data. Anthropological
O’Steen, W. Cleary, and G.C. Turner. 1999. Papers of the American Museum of Natural
Data recovery at 31CB114, Columbus Coun- History 88: 498–518.
ty, North Carolina: prehistoric settlement and Andrus, C.F.T., and K.W. Rich. 2008. A pre-
subsistence practices within the Lower Cape liminary assessment of oxygen isotope frac-
Fear River valley. Manuscript on file at the tionation and growth increment periodicity
Office of State Archaeology, Raleigh, NC. in the estuarine clam Rangia cuneata. Geo-
Addessi, E., A. Mancini, L. Crescimbene, D. Ari- Marine Letters 28: 301–308. http://dx.doi.
ely, and E. Visalberghi. 2010. How to spend org/10.1007/s00367-008-0109-3
a token? trade-offs between food variety Andrus, C.F.T., D.E. Crowe, D.H. Sand-
and food preference in tufted capuchin mon- weiss, E.J. Reitz, and C.S. Romanek. 2002.
keys (Cebus apella). Behavioural Process- Otolith δ18O record of Mid-Holocene sea
es 83 (3): 267. http://dx.doi.org/10.1016/j. surface temperatures in Peru. Science
beproc.2009.12.012 295: 1508–1511. http://dx.doi.org/10.1126/
Ames, K.M. 2002. Going by boat: the forager- science.1062004
collector continuum at sea. In B. Fitzhugh and Ansell, A.D. 1968. The rate of growth of the hard
J. Habu (editors), Beyond foraging and col- clam Mercenaria mercenaria (L.) throughout
lecting: evolutionary change in hunter gath- the geographical range. Journal du Conseil –
erer settlement systems: 19–52. New York: Conseil International pour l’Exploration de la
Kluwer Academic. Mer 31: 364–409.
Anderson, D.G. 1994. The Savannah River Arnold, W.S., D.C. Marelli, T.M. Bert, D.S.
chiefdoms: political change in the late pre- Jones, and I.R. Quitmyer. 1991. Habitat
historic Southeast. Tuscaloosa: University of specific growth of hard clams Mercenaria
Alabama Press. mercenaria (L.) from the Indian River, Flor-
Anderson, D.G., D.W. Stahle, and M.K. Cleave- ida. Journal of Experimental Marine Biol-
land. 1995. Paleoclimate and the potential ogy and Ecology 147: 245–265. http://dx.doi.
food reserves of Mississippian societies: a org/10.1016/0022-0981(91)90185-Y
case study from the Savannah River Valley. Aten, L.E. 1981. Determining seasonality of
American Antiquity 60 (2): 258–286. http:// Rangia cuneata from Gulf Coast shell mid-
dx.doi.org/10.2307/282140 dens. Bulletin of the Texas Archaeological
Andreasson, F.P., B. Schmitz, and E. Jönsson. Society 52: 179–200.
1999. Surface-water seasonality from stable Bailey, G.N., and A.S. Craighead. 2003. Late
isotope profiles of Littorina littorea shells: im- Pleistocene and Holocene coastal palaeo-
plications for paleoenvironmental reconstruc- economies: a reconsideration of the mol-
tions of coastal areas. Palaios 14: 273–281. luscan evidence from northern Spain. Geo-
http://dx.doi.org/10.2307/3515439 archaeology 18 (2): 175–204. http://dx.doi.
Andrus, C.F.T., and D.E. Crowe. 2000. Geo- org/10.1002/gea.10057
chemical analysis of Crassostrea virginica Bailey, G.N., M.R. Deith, and N.J. Shackleton.
as a method to determine season of capture. 1983. Oxygen isotope analysis and seasonality
Journal of Archaeological Science 27: 33–42. determinations: limits and potential of a new
http://dx.doi.org/10.1006/jasc.1999.0417 technique. American Antiquity 48: 390–398.
Andrus, C.F.T., and D.E. Crowe. 2002. Altera- http://dx.doi.org/10.2307/280460
tion of otolith aragonite: effects of prehistoric Balasse, M., A.B. Smith, S.H. Ambrose, and
cooking methods on otolith chemistry. Jour- S.R. Leigh. 2003. Determining sheep birth
nal of Archaeological Science 29: 291–299. seasonality by analysis of tooth enamel oxy-
http://dx.doi.org/10.1006/jasc.2001.0694 gen isotope ratios: the Late Stone Age site
Andrus, C.F.T., and D.E. Crowe. 2008. Isotope of Kasteelberg (South Africa). Journal of
analysis as a means for determining season Archaeological Science 30: 205–215. http://
of capture for Mercenaria. In D.H. Thomas, dx.doi.org/10.1006/jasc.2002.0833
Native American landscapes of St. Catherines Barats, A., D. Amouroux, L. Chauvaud, C.
Pécheyran, A. Lorrain, J. Thébault, T.M. Binford, L.R. 1980. Willow smoke and dogs’
Church, and O.F.X. Donard. 2009. High fre- tails: hunter-gatherer settlement systems and
quency barium profiles in shells of the great archaeological site formation. American An-
scallop Pecten maximus: a methodological tiquity 45: 4–20. http://dx.doi.org/10.2307/
long-term and multi-site survey in Western 279653
Europe. Biogeosciences 6: 157–170. http:// Binford, L.R. 1990. Mobility, housing, and envi-
dx.doi.org/10.5194/bg-6-157-2009 ronment: a comparative study. Journal of An-
Bartram, W. 1928. Travels of William Bartram. thropological Research 46 (2): 119–152.
New York: Macy-Macius. [Reprinted in 1955 Bird, D.W., and R. Bliege Bird. 2000. The ethno-
by Dover Publications, New York.] archaeology of juvenile foragers: shellfishing
Bar-Yosef, O., and. T.R. Rocek. 1998. Introduc- strategies among Meriam children. Journal of
tion. In T.R. Rocek and O. Bar-Yosef (editors), Anthropological Archaeology 19: 461–476.
Seasonality and sedentism: archaeological http://dx.doi.org/10.1006/jaar.2000.0367
perspectives from Old and New World sites: Bird, D.W., R. Bliege Bird, and J.L. Richardson.
1–8. Peabody Museum of Archaeology and 2004. Meriam ethnoarchaeology: shellfishing
Ethnology Bulletin No. 6, Cambridge, MA: and shellmiddens. Memoirs of the Queen-
Harvard University. sland Museum, Cultural Heritage Series 3
Bayliss, A., and C. Bronk Ramsey. 2004. Prag- (1): 183–197.
matic Bayesians: a decade of integrating ra- Bishop, G.A., R.H. Hayes, B.K. Meyer, H.B.
diocarbon dates into chronological models. In Rollins, F.J. Rich, D.H. Thomas, and R.K.
C.E. Buck and A.R. Millard (editors), Tools Vance. 2007. Transgressive barrier island fea-
for constructing chronologies: crossing disci- tures of St. Catherines Island, Georgia. In F.J.
plinary boundaries: 25–41. Lecture notes in Rich (editor), Fieldtrips of the 2007 annual
statistics 177. New York: Springer. meeting, Southeastern Section, Geological
Bayliss, A., C. Groves, G. McCormac, M. Baillie, Society of America: 39–86. Statesboro, Geor-
D. Brown, and M. Brennand. 1999. Precise gia Southern University.
dating of the Norfolk timber circle. Nature Bishop, G.A., B.K. Meyer, R.K. Vance, and F.J.
402: 479. http://dx.doi.org/10.1038/44993 Rich. 2011. Geoarchaeological research at
Bayliss, A., C. Bronk Ramsey, J. van der Plicht, St. Catherines Island: defining the geological
and A. Whittle. 2007. Bradshaw and Bayes: foundation. In G.A. Bishop, H.B. Rollins, and
toward a timetable for the Neolithic. Cam- D.H. Thomas (editors), Geoarchaeology of
bridge Archaeological Journal 17: 1–28. http:// St. Catherines Island, Georgia. Anthropologi-
dx.doi.org/10.1017/S0959774307000145 cal Papers of the American Museum of Natu-
Bennett, C.E. (translator). 1975. Three voyages— ral History 94: 79–98.
René Laudonière. Gainesville: University of Bishop, G.A., H.B. Rollins, and D.H. Thomas
Florida Press. (editors). 2011. Geoarchaeology of St. Cathe-
Bettinger, R.L. 1991. Hunter-gatherers: archaeo- rines Island, Georgia. Anthropological Papers
logical and evolutionary theory. New York: of the American Museum of Natural History
Plenum Press. 94.
Bettinger, R.L., R. Malhi, and H. McCarthy. Blackwell, P.G., C.E. Buck, and P.J. Reimer.
1997. Central place models of acorn and mus- 2006. Important features of the new radio-
sel processing. Journal of Archaeological Sci- carbon calibration curves. Quaternary Sci-
ence 24: 887–899. http://dx.doi.org/10.1006/ ence Reviews 25: 408–413. http://dx.doi.
jasc.1996.0168 org/10.1016/j.quascirev.2005.12.001
Beverly, R.K., W. Beaumont, D. Tauz, K.M. Blanton, D.B. 2000. Drought as a factor in the
Ormsby, K.F. von Reden, G.M. Santos, and Jamestown colony. Historical Archaeology 34
J.R. Southon. 2010. The Keck Carbon Cycle. (4): 1606–1612.
AMS Laboratory, University of California, Ir- Blanton, D.B. 2004. The climatic factor in late
vine: Status Report. Radiocarbon 52. prehistoric and post-contact human affairs.
Bigalke, E.H. 1973. The exploitation of shellfish In D.B. Blanton and J.A. King (editors), In-
by coastal tribesmen of the Transkei. Annual dian and European contact: the mid-Atlantic
of the Cape Province Museum (Natural His- region: 6–21. Gainesville: University Press of
tory) 9: 159–175. Florida.
2012 REFERENCES 213
Blanton, D.B., and D.H. Thomas. 2008. Paleo- versity of Alabama Press.
climates and human responses along the cen- Brown, J.E. 1985. Long-term trends to sedentism
tral Georgia coast: a tree-ring perspective. In and the emergence of complexity in the Amer-
D.H. Thomas, Native American landscapes of ican Midwest. In T.D. Price and J.A. Brown
St. Catherines Island, Georgia. II. The data. (editors), Prehistoric hunter-gatherers: the
Anthropological Papers of the American Mu- emergence of cultural complexity: 201–231.
seum of Natural History 88: 799–806. San Diego, CA: Academic Press.
Bocherens, H., D. Billiou, A. Mariotti, M. Patou- Brown, T.A., D.E. Nelson, J.S. Vogel, and J.R.
Mathis, M. Otte, D. Bonjean, and M. Toussaint. Southon. 1988. Improved collagen extraction
1999. Palaeoenvironmental and palaeodietary by modified Longin method. Radiocarbon 30
implications of isotopic biogeochemistry of (2): 171–177.
last interglacial Neanderthal and mammal Buck, C.E. 2004. Bayesian chronological data
bones in Scladina Cave (Belgium). Journal interpretation: where now? In C.E. Buck and
of Archaeological Science 26 (6): 599–607. A.R. Millard (editors), Tools for constructing
http://dx.doi.org/10.1006/jasc.1998.0377 chronologies: crossing disciplinary boundar-
Borrero, F.J., and T.J. Hilbish. 1988. Temporal ies: 1–24. Lecture notes in statistics 177. Lon-
variation in shell and soft tissue growth of don: Springer-Verlag.
the mussel Geukensia demissa. Marine Ecol- Buck, C.E., and A.R. Millard (editors). 2004a.
ogy Progress Series 42: 9–15. http://dx.doi. Tools for constructing chronologies: crossing
org/10.3354/meps042009 disciplinary boundaries. London: Springer-
Breder, C.M. 1929. Field book of marine fishes Verlag.
of the Atlantic coast from Labrador to Texas; Buck, C.E., and A.R. Millard. 2004b. Introduc-
being a short description of their characteris- tion. In C.E. Buck and A.R. Millard (editors),
tics and habits with keys for their identifica- Tools for constructing chronologies: crossing
tion. New York: G.P. Putnam & Sons. disciplinary boundaries: I–VI. Lecture notes
Bronk Ramsey, C. 1995. Radiocarbon calibration in statistics 177. London: Springer-Verlag.
and analysis of stratigraphy: the OxCal pro- Buck, C.E., C.D. Litton, and A.F.M. Smith. 1992.
gram. Radiocarbon 37 (2): 461–474. Calibration of radiocarbon results pertaining
Bronk Ramsey, C. 2000. Comment on ‘The use to related archaeological events. Journal of Ar-
of Bayesian statistics for 14C dates of chrono- chaeological Science 19 (5): 497–512. http://
logically ordered samples: a critical analysis.’ dx.doi.org/10.1016/0305-4403(92)90025-X
Radiocarbon 42: 199–202. Buck, C.E., W.G. Cavanagh, and C.D. Litton.
Bronk Ramsey, C. 2001. Development of the ra- 1996. Bayesian approach to interpreting ar-
diocarbon program OxCal. Radiocarbon 43 chaeological data. Chichester, UK: Wiley.
(2A): 355–363. Buddemeier, R.W., J.E. Maragos, and D.W.
Bronk Ramsey, C. 2005. OxCal version 3.10. Inter- Knutson. 1974. Radiographic studies of reef
net resource (http://c14.arch.ox.ac.uk/embed. coral exoskeletons: rates and patterns of coral
php?File=oxcal.html), accessed on December growth. Journal of Experimental Marine Biol-
8, 2011. ogy and Ecology 14: 179–199. http://dx.doi.
Bronk Ramsey, C., J. van der Plicht, and B. org/10.1016/0022-0981(74)90024-0
Weninger. 2001. ‘Wiggle Matching’ radiocar- Builth, H. 2006. Gunditjmara environmental
bon dates. Radiocarbon 43: 381–389. management: the development of a fisher-
Broughton, J.M. 2010. Prey spatial structure gatherer-hunter society in temperate Austra-
and behavior affect archaeological tests of lia. In C. Grier, J. Kim, and J. Uchiyama (edi-
optimal foraging models: examples from the tors), Beyond affluent foragers: rethinking
Emeryville Shellmound vertebrate fauna. In hunter-gatherer complexity: 4–23. Oxford:
J.M. Broughton and M.D. Cannon (editors), Oxbow Books.
Evolutionary ecology and archaeology: appli- Burman, J., and T. Påsse. 2008. Oceanography in
cations to problems of human evolution and northwestern Europe during the last interglacial
prehistory: 192–207. Salt Lake City: Univer- from intrashell δ18O ranges in Littorina littorea
sity of Utah Press. gastropods. Quaternary Research 70: 121–
Brown, I.W. 2003. Bottle Creek: a Pensacola cul- 128. http://dx.doi.org/10.1016/j.yqres.2008.
ture site in south Alabama. Tuscaloosa: Uni- 03.007
Burman, J., and B. Schmitz. 2005. Periwinkle Insights into early Holocene hunter-gatherer
(Littorina littorea) intrashell δ18O and δ13C mobility on the Peruvian southern coast from
records from the mid-Holocene Limfjord mollusk gathering seasonality. Journal of Ar-
region, Denmark: a new high-resolution chaeological Science 36: 1173–1178. http://
palaeoenvironmental proxy approach. Ho- dx.doi.org/10.1016/j.jas.2009.01.005
locene 15: 567–575. http://dx.doi.org/10. Carriker, M.R. 1996. The shell and ligament. In
1191/0959683605hl831rp V.S. Kennedy, R.I.E. Newell, and A.R. Eble
Bushnell, A.T. 1990. The sacramental impera- (editors), The eastern oyster Crassostrea vir-
tive: Catholic ritual and Indian sedentism in ginica: 75–168. College Park: Maryland Sea
the provinces of Florida. In D.H. Thomas, Grant College.
Columbian consequences, archaeological and Carroll, M., C.S. Romanek, and L. Paddock. 2006.
historical perspectives on the Spanish border- The relationship between the hydrogen and
lands east: 475–490. Washington, DC: Smith- oxygen isotopes of freshwater bivalve shells
sonian Institution Press. and their home streams. Chemical Geology
Butler, V.L., and S.K. Campbell. 2004. Resource 234: 211–222. http://dx.doi.org/10.1016/j.
intensification and resource depression in the chemgeo.2006.04.012
Pacific Northwest of North America: a zooar- Cerrato, R.M. 1980. Demographic analysis of bi-
chaeological review. Journal of World Prehis- valve populations. In D.C. Rhodes and R.A.
tory 18: 327–405. http://dx.doi.org/10.1007/ Lutz (editors), Skeletal growth of aquatic or-
s10963-004-5622-3 ganisms: biological records of environmental
Butler, V.L., and M.G. Delacorte. 2004. Doing change: 417–463. New York: Plenum Press.
zooarchaeology as if it mattered: use of faunal Chittaro, P.M., P. Usseglio, B.J. Fryer, and P.F.
data to address current issues in fish conserva- Sale. 2005. Using otolith microchemistry
tion biology in Owens Valley, California. In of Haemulon flavolineatum (French Grunt)
R.L. Lyman and K.P. Cannon (editors), Zoo- to characterize mangroves and coral reefs
archaeology and conservation biology: 25–44. throughout Turneffe Atoll, Belize: difficul-
Salt Lake City: University of Utah Press. ties at small spatial scales. Estuaries 28
Caldwell, J.R. 1971. Chronology of the Georgia (3): 373–381. http://dx.doi.org/10.1007/
coast. Southeastern Archaeological Confer- BF02693920
ence Bulletin 13: 89–91. Chittaro, P.M., P. Usseglio, B.J. Fryer, and P.F.
Campbell, S.K., and V.L. Butler. 2010. Archae- Sale. 2006. Spatial variation in otolith chem-
ological evidence for resilience of Pacific istry of Lutjanus apodus at Turneffe Atoll,
Northwest salmon populations and the socio- Belize. Estuarine, Coastal and Shelf Science
ecological system over the last ~7,500 years. 67: 673–680. http://dx.doi.org/10.1016/j.ecss.
Ecology and Society 15 (1): 17. 2005.12.014
Cannarozzi, N., and I. Quitmyer. 2007. Oyster Claassen, C.P. 1982. Shellfishing patterns: an an-
schlerochronology helps archaeologists date alytical study of prehistoric shell from North
sites. No Wake Zone 2 (1): 1. Carolina coastal middens. Ph.D. dissertation,
Carlson D.L. 1987. Rangia cuneata as a seasonal Department of Anthropology, Harvard Uni-
indicator for coastal archaeological sites in versity, Cambridge.
Texas. Bulletin of the Texas Archaeological Claassen, C. 1986. Shellfishing seasons in the
Society 58: 201–214. prehistoric southeastern United States. Amer-
Carré, M., I. Bentaleb, D. Blamart, N. Ogle, F. ican Antiquity 51 (1): 21–37. http://dx.doi.
Cardenas, S. Zevallos, R.M. Kalin, L. Ortlieb, org/10.2307/280391
and M. Fontugne. 2005. Stable isotopes and Claassen, C.P. 1990. Investigation of monthly
sclerochronology of the bivalve Mesodesma growth in shellfish for application to archae-
donacium: potential application to Peruvian ology. Final Report, National Science Foun-
paleoceanographic reconstructions. Palaeo- dation Grant # BNS-8507714.
geography, Palaeoclimatology, Palaeoecol- Claassen, C. 1993. Problems and choices in shell
ogy 228: 4–25. http://dx.doi.org/10.1016/j. seasonality studies and their impact on results.
palaeo.2005.03.045 Archaeozoologica 5(2): 55–75.
Carré, M., L. Klaric, D. Lavallée, M. Julien, I. Claassen, C. 1998. Shells. Cambridge, UK: Cam-
Bentaleb, M. Fontugne, and O. Kawka. 2009. bridge University Press.
2012 REFERENCES 215
Claassen, C., and S. Sigmann. 1993. Sourcing http://dx.doi.org/10.1016/j.jas.2009.04.021

Busycon artifacts of the eastern United States. Comyns, B.H., C.F. Rakocinski, M.S. Peterson,
American Antiquity 58: 333–347. http://dx. and A.M. Shiller. 2008. Otolith chemistry of
doi.org/10.2307/281974 juvenile spotted seatrout Cynoscion nebulo-
Clark, G.R., II. 1968. Mollusk shell: daily growth sus reflects local natal regions of coastal Mis-
lines. Science 161: 800–802. http://dx.doi. sissippi, USA. Marine Ecology Progress Se-
org/10.1126/science.161.3843.800 ries 371: 243–252. http://dx.doi.org/10.3354/
Clark, G.R., II. 1974. Growth lines in inverte- meps07604
brate skeletons. Annual Review of Earth and Conover, D.O., and M.R. Ross. 1982. Patterns
Planetary Sciences 2: 77–99. http://dx.doi. in seasonal abundance, growth and biomass
org/10.1146/annurev.ea.02.050174.000453 of the Atlantic silverside, Menidia menidia,
Clark, G.R., II. 1976a. Seasonal growth varia- in a New England estuary. Estuaries 5 (4):
tions in bivalve shells as indicators of human 275–286. http://dx.doi.org/10.2307/1351750
occupation patterns on St. Catherines Island, Costanza, R., R. d’Arge, R. de Groot, S. Far-
Georgia. Interim report New York, American ber, M. Grasso, B. Hannon, K. Limburg, S.
Museum of Natural History. Naeem, R.V. O’Neill, J. Paruelo, R.G. Raskin,
Clark, G.R., II. 1976b. Interpretation of season P. Sutton, and M. van den Belt. 1997. The
of death of mollusk shells associated with value of the world’s ecosystem services and
burial mounds on St. Catherines Island, natural capital. Nature 387: 253–260. http://
Georgia. Proposal. American Museum of dx.doi.org/10.1038/387253a0
Natural History. Coutts, P.J.F. 1970. Bivalve-growth pattern-
Clark G.R., II. 1979. Seasonal growth variations ing as a method for seasonal dating in ar-
in the shells of recent and prehistoric speci- chaeology. Nature 226: 874. http://dx.doi.
mens of Mercenaria mercenaria from St. org/10.1038/226874a0
Catherines Island, Georgia. In D.H. Thomas Coutts, P.J.F. 1975. The seasonal perspective of
and C.S. Larsen (editors), The anthropol- marine-oriented prehistoric hunter–gatherers.
ogy of St. Catherines Island: 2. The Refuge- In G.D. Rosenberg and S.K. Runcorn (edi-
Deptford mortuary complex. Anthropological tors), Growth rhythms and the history of the
Papers of the American Museum of Natural earth’s rotation: 243–252. London: Wiley.
History 56 (1): 161–179. Coutts, P.J.F., and T. Higham. 1971. The seasonal
Cobb, R.M., C.F.T. Andrus, and M.F. Etayo- factor in prehistoric New Zealand. World Ar-
Cadavid. 2009. Rangia cuneata shells as an chaeology 2 (3): 266–277. http://dx.doi.org/1
environmental proxy: variations in elemental 0.1080/00438243.1971.9979480
concentrations within populations. University Cox, K.A. 1994. Oysters as ecofacts. Bulle-
of Alabama McNair Journal 9: 45–58. tin of the Texas Archaeological Society 62:
Cohen, M.N. 1977. The food crisis in prehistory: 219–247.
overpopulation and the origins of agriculture. Craig, H. 1965. Measurement of oxygen isotope
New Haven, CT: Yale University Press. paleotemperatures. In E. Tongiori (editor),
Colaninno, C.E. 2010. Zooarchaeological analy- Stable isotopes in oceanographic studies:
sis of vertebrate remains from five late Archaic 161–182. Consiglio Nazionale delle Ricerche,
shell rings of the Georgia coast, USA. Ph.D. Spoleto, Italy.
dissertation, University of Georgia, Athens. Crawford, K., R.A. McDonald, and S. Bearhop.
Collins, M.R. 1985. Species profiles: life histo- 2008. Applications of stable isotope tech-
ries and environmental requirements of coast- niques to the ecology of mammals. Mam-
al fishes and invertebrates (South Florida): mal Review 38 (1): 87–107. http://dx.doi.
striped mullet. U.S. Fish and Wildlife Service, org/10.1111/j.1365-2907.2008.00120.x
Biological Report 82 (11.34). Crites, G.D. 1999. Long Point (31JN2) plant
Colonese, A.C., S. Troelstra, P. Ziveri, F. Martini, remains. In M.S. Shumate and P. Evans-
D. Lo Vetro and S. Tommasini. 2009. Meso- Shumate, Archaeological investigations at
lithic shellfish exploitation in SW Italy: sea- Long Point (31JN2) and Haywood Landing
sonal evidence from the oxygen isotopic com- (31JN3) on the Croatan National Forest, Jones
position of Osilinus turbinatus shells. Journal County, North Carolina. Ms. on file at the Of-
of Archaeological Science 36: 1935–1944. fice of State Archaeology, Raleigh, NC.
Crook, M.R., Jr. 1978. Mississippian period com- American Midland Naturalist 83: 382–392.
munity organizations on the Georgia coast. http://dx.doi.org/10.2307/2423951
Ph.D. dissertation, University of Florida, Dawson, T.E., and R.T.W. Siegwolf. 2007. Using
Gainesville. stable isotopes as indicators, tracers, and re-
Crook, M.R., Jr. 1984. Evolving community or- corders of ecological change. In T.E. Dawson
ganization on the Georgia coast. Journal of and R.T.W. Siegwolf (editors), Stable isotopes
Field Archaeology 11 (3): 247–263. http:// as indicators of ecological change: 3–18. Am-
dx.doi.org/10.1179/009346984791535485 sterdam: Elsevier.
Crook, M.R., Jr. 1986. Mississippi period archae- de Boer, W.F., T. Pereira, and A. Guissamulo.
ology of the Georgia coastal zone. University 2000. Comparing recent and abandoned shell
of Georgia Laboratory of Archaeology, series middens to detect the impact of human ex-
report 23. ploitation on the intertidal ecosystem. Aquat-
Crook, M.R., Jr. 1992. Oyster sources and their ic Ecology 34 (3): 287–297. http://dx.doi.
prehistoric use on the Georgia coast. Journal org/10.1023/A:1009957409421
of Archaeological Science 19: 483–496. http:// Deevey, E.S., Jr. 1947. Life tables of natu-
dx.doi.org/10.1016/0305-4403(92)90024-W ral populations of animals. Quarterly Re-
Crook, M.R., Jr. 2004. Late prehistoric settle- view of Biology 22: 283–314. http://dx.doi.
ment mobility on the Georgia coast. Unpub- org/10.1086/395888
lished ms. DEIS. 1978. Draft environmental impact state-
Culleton, B.J., D.J. Kennett, B.L. Ingram, J. Er- ment for preferred alternative location for a
landson, and J. Southon. 2006. Intra-shell fleet ballistic missile (FBM) submarine sup-
radiocarbon variability in marine mollusks. port base. Appendix 2. Washington, DC: De-
Radiocarbon 43 (3): 387–400. partment of the Navy.
Culleton, B.J., D.J. Kennett, and T.L. Jones. Deith, M.R. 1983. Molluscan calendars: the use
2009. Oxygen isotope seasonality in a temper- of growth-line analysis to establish season-
ate estuarine shell midden: a case study from ality of shellfish collection at the Mesolithic
CA-ALA-17 on the San Francisco Bay, Cali- site of Morton, Fife. Journal of Archaeo-
fornia. Journal of Archaeological Science 36 logical Science 10: 423–440. http://dx.doi.
(7): 1354–1363. http://dx.doi.org/10.1016/j. org/10.1016/0305-4403(83)90058-4
jas.2009.01.021 Deith, M.R. 1986. Subsistence strategies at a Me-
Cusack, M., D. Parkinson, A. Freer, A. Pérez- solithic camp site: evidence from stable iso-
Huerta, A.E. Fallick, and G.B. Curry. 2008. tope analyses of shells. Journal of Archaeo-
Oxygen isotope composition in Modiolus logical Science 13 (1): 61–78. http://dx.doi.
modiolus aragonite in the context of biologi- org/10.1016/0305-4403(86)90027-0
cal and crystallographic control. Mineral- DeNiro, M.J. 1985. Postmortem preservation
ogical Magazine 72: 569–577. http://dx.doi. and alteration of in vivo bone collagen iso-
org/10.1180/minmag.2008.072.2.569 tope ratios in relation to paleodietary recon-
Custer, J.F., and K.R. Doms. 1990. Analysis of struction. Nature 317: 806–809. http://dx.doi.
microgrowth patterns of American oyster org/10.1038/317806a0
(Crassostrea virginica) in the Middle Atlantic DePratter, C.B. 1979. Ceramics. In D.H. Thomas
Region of Eastern North America: archaeo- and C.S. Larsen (editors), The anthropology of
logical applications. Journal of Archaeo- St. Catherines Island: 2. The Refuge-Deptford
logical Science 17: 151–160. http://dx.doi. mortuary complex. Anthropological Papers of
org/10.1016/0305-4403(90)90056-B the American Museum of Natural History 56
Dahlberg, M.D. 1972. An ecological study of (1): 109–132.
Georgia coastal fishes. Fishery Bulletin 70 DePratter, C.B. 1991. W.P.A. archaeological ex-
(2): 323–353. cavations in Chatham County, Georgia:
Dahlberg, M.D. 1975. Guide to coastal fishes of 1937–1942. University of Georgia, Labora-
Georgia and nearby states. Athens: University tory of Archaeology Series, Report 29.
of Georgia Press. Dettman, D.L., A.K. Reische, and K.C. Lohmann.
Dahlberg, M.D., and E.P. Odum. 1970. Annual 1999. Controls on the stable isotope composi-
cycles of species occurrence, abundance, and tion of seasonal growth bands in aragonitic
diversity in Georgia estuarine fish populations. fresh-water bivalves (Unionidae). Geochimica
2012 REFERENCES 217
et Cosmochimica Acta 63: 1049–1057. http:// 00438240220134287

dx.doi.org/10.1016/S0016-7037(99)00020-4 Emiliani, C., L. Cardini, T. Mayeda, C.B.M. Mc-
Detwiler, K., and D. Keene. 2003. Rethinking Burney, and E. Tongiorgi. 1964. Palaeotem-
late prehistoric subsistence on the Georgia perature analysis of marine molluscs (food
coast: evidence for agriculture at the Grove’s refuse) from the site of Arene Candide Cave,
Creek Site (09CH71). Paper presented at the Italy and the Haua Fteah Cave, Cyrenaica. In
26th Annual Conference of the Society of H. Craig, S.L. Miller, and G. J. Wasserburg,
Ethnobiology, Seattle. (editors), Isotopic and cosmic chemistry: 133–
Detwiler, K. and C.M. Scarry. 1999. Plant re- 156. Amsterdam: North Holland Pub. Co.
mains. In R. Daniel, Archaeological excava- Epstein, S., and H.A. Lowenstam. 1953. Tempera-
tions at Hammocks Beach West (31ON665): ture–shell-growth relations of recent and inter-
a Woodland shell midden on the North Caro- glacial Pleistocene shoal-water biota from Ber-
lina coast: 149–156. Occasional Papers of the muda. Journal of Geology 61: 424–130. http://
Phelps Archaeology Laboratory No. 1. East dx.doi.org/10.1086/626110
Carolina University. Epstein, S., R. Buchsbaum, H.A. Lowenstam,
Devereux, I. 1967. Temperature measurements and H.C. Urey. 1953. Revised carbonate-
from oxygen isotope ratios of fish otoliths. water isotopic temperature scale. Bulletin
Science 155: 1684–1685. http://dx.doi. of the Geological Society of America 64:
org/10.1126/science.155.3770.1684 1315–1326. http://dx.doi.org/10.1130/0016-
Dolan, R., B. Hayden, and H. Lins. 1980. Barrier 7606(1953)64[1315:RCITS]2.0.CO;2
islands. American Scientist 68: 16–25. Erickson, C.L. 2000. An artificial landscape-scale
Dorval, E., C.M. Jones, R. Hannigan, and J. van fishery in the Bolivian Amazon. Nature 408:
Montfrans. 2004. Can otolith chemistry be 190–193. http://dx.doi.org/10.1038/35041555
used for identifying essential seagrass habi- Erlandson, J.M. 1994. Early hunter-gatherers of
tats for juvenile spotted seatrout, Cynoscion the California coast. New York: Plenum Press.
nebulosus, in Chesapeake Bay? Marine and Erlandson, J.M. 2001. The archaeology of
Freshwater Research 56 (5): 645–653. http:// aquatic adaptations: paradigms for a new
dx.doi.org/10.1071/MF04179 millennium. Journal of Archaeological Re-
Drake, J.C. 1891. On the sounds and estuaries search 9 (4): 287–350. http://dx.doi.org/10.
of Georgia with reference to oyster culture. 1023/A:1013062712695
U.S. Coast and Geodetic Survey Bulletin 19: Erlandson, J.M., T.C. Rick, T.J. Braje, A. Stein-
179–209. berg, and R.L. Vellanoweth. 2008. Human im-
Drennan, R.D. 2010. Statistics for archaeologists, pacts on ancient shellfish: a 10,000 year record
2nd edition: a common sense approach (inter- from San Miguel Island, California. Journal
disciplinary contributions to archaeology). of Archaeological Science 35 (8): 2144–2152.
New York: Plenum Press. http://dx.doi.org/10.1016/j.jas.2008.01.014
Eble, A.F., and R. Scro. 1996. General anatomy. Etayo-Cadavid, M.F. 2010. Peruvian mollusk
In V.S. Kennedy, R.I.E. Newell, and A.F. Eble shells as multi-proxy archives: Late Holocene
(editors), The eastern oyster Crassostrea vir- upwelling variation and El Niño–induced
ginica: 19–73. Maryland Sea Grant College, biomineralization effects on trace elements.
College Park. Ph.D. dissertation, University of Alabama,
Elliot, M., P.B. deMenocal, B.K. Linsley, and Tuscaloosa.
S.S. Howe. 2003. Environmental controls on Evershed, R.P., I.D. Bull, L.T. Corr, Z.M. Cross-
the stable isotopic composition of Mercenaria man, B.E. Van Dongen, C.J. Evans, S. Jim,
mercenaria: potential application to paleoen- H.R. Mottram, A.J. Mukherjee, and R.D. Pan-
vironmental studies. Geochemistry Geo- cost. 2007. Compound-specific stable isotope
physics Geosystems 4: 1–16. http://dx.doi. analysis in ecology and paleoecology. In R. Mi-
org/10.1029/2002GC000425 chener and K. Lajtha (editors), Stable isotopes
Elston, R.G., and D. Zeanah. 2002. Thinking in ecology and environmental science, 2nd
outside the box: a new perspective on diet ed.: 480–540. Malden, MA: Blackwell. http://
breadth and sexual division of labor in the dx.doi.org/10.1002/9780470691854.ch14
prearchaic Great Basin. World Archaeol- Eversole, A.G., W.D. Anderson, and J.J. Isely.
ogy 34: 103–130. http://dx.doi.org/10.1080/ 2008. Age and growth of the knobbed
whelk Busycon carica (Gmelin 1791) org/10.1029/2008PA001677

in South Carolina subtidal waters. Jour- Fowler, A.J. 1995. Annulus formation in otoliths
nal of Shellfish Research 27: 423–426. of coral reef fish—a review. In D.H. Secor,
http://dx.doi.org/10.2983/0730-8000(2008) J.M. Dean, and S.E. Campana (editors), Re-
27[423:AAGOTK]2.0.CO;2 cent developments in fish otolith research:
Fagan, B.M. 2000. The Little Ice Age: how cli- 45–63. University of South Carolina Press,
mate made history 1300–1850. New York: Columbia.
Basic Books. Fowler, C.S. 2008. Historical perspectives on
Fagan, B.M. 2005. The long summer: how climate Timbisha Shoshone land management prac-
changed civilization. New York: Basic Books. tices, Death Valley, California. In E.J. Reitz,
Fairbanks, L.D. 1963. Biodemographic studies of C.M. Scarry, S. Scudder (editors), Case stud-
the clam Rangia cuneata Gray. Tulane Stud- ies in environmental archaeology, 2nd ed.:
ies in Zoology 10: 3–47. 43–57. New York: Springer. http://dx.doi.
Fegley, S.R. 2001. Demography and dynamics org/10.1007/978-0-387-71303-8_3
of hard clam populations. In J.N. Kraeuter Fradkin, A. 2008. Applying a measure for deter-
and M. Castagna, Biology of the hard clam: mining season of oyster harvesting: a case
381–422. Amsterdam: Elsevier. study of a pre-Columbian Florida site, USA.
Ferrio, J.P., J. Voltas, N. Alonso, and J.L. Araus. Archaeofauna 17: 147–156.
2007. Reconstruction of climate and crop con- France, R. 1995. Stable nitrogen isotopes in fish:
ditions in the past based on the carbon isotope literature synthesis on the influence of ecoton-
signature of archaeobotanical remains. In T.E. al coupling. Estuarine, Coastal and Shelf Sci-
Dawson and R.T.W. Siegwolf (editors), Stable ence 41: 737–742. http://dx.doi.org/10.1006/
isotopes as indicators of ecological change: ecss.1995.0087
319–332. Amsterdam: Elsevier. Frey, R.W., and J.D. Howard. 1986. Mesotidal
Fives, J.M., S.M. Warlen, and D.E. Hoss. 1986. estuarine sequences: a perspective from the
Aging and growth of larval bay anchovy, An- Georgia Bight. Journal of Sedimentary Pe-
choa mitchilli, from the Newport River estu- trology 56 (6): 911–924.
ary, North Carolina. Estuaries 9 (4): 362–367. Fritz, G.J., V. Drywater Whitekiller, and J.W.
http://dx.doi.org/10.2307/1351416 McIntosh. 2001. Ethnobotany of Ku-nu-che:
Flannery, K.V. 1968. Archaeological systems Cherokee hickory nut soup. Journal of Ethno-
theory and early Mesoamerica. In B.J. Meg- biology 21 (2):1–27.
gers (editor), Anthropological archeology in Fritz, L.W. 2001. Shell structure and age deter-
the Americas: 67–87. Washington DC: An- mination. In J.N. Kraeuter and M. Castagna
thropological Society of Washington. (editors), Biology of the hard clam: 53–76.
Fodrie, F.J., and S.Z. Herzka. 2008. Tracking London, New York: Elsevier.
juvenile fish movement and nursery contri- Fry, B. 2002. Conservative mixing of stable
bution within arid coastal embayments via isotopes across estuarine salinity gradients:
otolith microchemistry. Marine Ecology a conceptual framework for monitoring wa-
Progress Series 361: 253–265. http://dx.doi. tershed influences on downstream fisheries
org/10.3354/meps07390 production. Estuaries 25 (2): 264–271. http://
Fodrie, F.J., K.L. Heck, J.P. Powers, W.M. Gra- dx.doi.org/10.1007/BF02691313
ham, and K.L. Robinson. 2010. Climate- Fuller, D.Q., R. Korisettar, and P.C. Venkatasub-
related, decadal-scale assemblage changes baiah. 2001. Southern Neolithic cultivation
of seagrass-associated fishes in the northern systems: reconstruction based on archaeo-
Gulf of Mexico. Global Change Biology 16 botanical evidence. South Asian Studies 17:
(1): 48–59. http://dx.doi.org/10.1111/j.1365- 171–187.
2486.2009.01889.x Galtsoff, P.S. 1964. The American oyster Cras-
Ford, H. L., S.A. Schellenberg, B.J. Becker, D.L. sostrea virginica Gmelin. U.S. Department
Deutschman, K.A. Dyck, and P.L. Koch. of the Interior, Fish and Wildlife Service,
2010. Evaluating the skeletal chemistry of Bureau of Commercial Fisheries, Fishery
Mytilus californianus as a temperature proxy: Bulletin 64.
effects of microenvironment and ontogeny. Gardner, P.S. 1984. Floral analysis. Appendix E.
Paleoceanography 25: PA1203. http://dx.doi. In M. Wilde-Ramsing, Archaeological survey
2012 REFERENCES 219
and testing on prehistoric shell midden sites gen isotopes as paleotemperature proxies in
in New Hanover County, North Carolina: bivalve mollusk shells: models and obser-
156–158. Unpublished M.A. thesis, Depart- vations. Palaios 18: 110–125. http://dx.doi.
ment of Anthropology, Catholic University of org/10.1669/0883-1351(2003)18<110:RAF
America. OOI>2.0.CO;2
Gardner, P.S. 1990. Excavations at the Amity Gorski, L. 2005. Microstratigraphic analysis at
Site: final report of the Pomeiooc Project: the Carlston Annis site, Butler County, Ken-
1984–1989. Archaeological Research Report tucky. In W.H. Marquardt and P.J. Watson
7 (OSA # 2771). East Carolina University, (editors), Archaeology of the Middle Green
Greenville. River Region, Kentucky: 153–166. Gaines-
Gardner, P.S. 1997. The ecological structure and ville: Institute of Archaeology and Paleoenvi-
behavioral implications of mast exploitation ronmental Studies Monograph 5, University
strategies. In K. Gremillion (editor), People, of Florida.
plants, and landscapes: 161–178. Tuscaloosa: Gosselain, O.P. 1998. Social and technical iden-
University of Alabama Press. tity in a clay crystal ball. In M.T. Stark (edi-
Gillikin, D.P., F. Dehairs, A. Lorrain, D. Steen- tor), The archaeology of social boundaries:
mans, W. Baeyens, and L. André. 2006. Bari- 78–106. Washington, DC: Smithsonian Insti-
um uptake into the shells of the common mus- tution Press.
sel (Mytilus edulis) and the potential for es- Green, P.R. 1984. Appendix C. Ethnobotanical
tuarine paleo-chemistry reconstruction. Geo- analysis of samples from the Wanchese Site,
chimica et Cosmochimica Acta 70: 395–407. 31Dr35. In D.S. Phelps, The archaeology of
http://dx.doi.org/10.1016/j.gca.2005.09.015 the Tillett Site: the first fishing community at
Gilmore, M.R. 1931. Dispersal by Indians, a fac- Wanchese, Roanoke Island: 103–107 (OSA
tor in the extension of discontinuous distribu- # 1627). Archaeological Research Report 6.
tion of certain species of native plants. Papers East Carolina University, Greenville, NC.
of the Michigan Academy of Science, Arts Green, P.R. 1986. The archaeology of “Chowa-
and Letters 13: 89–94 noke”: results of the 1983–1984 investiga-
Glassow, M.A. 1997. Research issues of impor- tions at Mount Pleasant and Liberty Hill,
tance to coastal California archaeology of Hertford County, North Carolina (OSA #
the middle Holocene. In J.M. Erlandson and 3279). Ms. on file, Office of State Archaeol-
M.A. Glassow (editors), Archaeology of the ogy, Raleigh, NC.
California coast during the middle Holocene: Gremillion, K.J. 2002. Foraging theory and hy-
151–161. Los Angeles: Institute of Archaeol- pothesis testing in archaeology: an explora-
ogy, University of California. tion of methodological problems and solu-
Glazier, E. 1987. Ethnobotanical analysis: Late tions. Journal of Anthropological Archaeol-
Woodland plant food utilization along the ogy 21: 142–164. http://dx.doi.org/10.1006/
Central North Carolina Coast. Appendix 1. In jaar.2001.0391
T.C. Loftfield, Excavations at 31On305, the Grimes, S.T., D.P. Mattey, J.J. Hooker, and M.E.
Flynt Site at Sneads Ferry, North Carolina Collinson. 2003. Paleogene paleoclimate re-
(OSA # 2190). Manuscript on file, Office of construction using oxygen isotopes from land
State Archaeology, Raleigh, NC. and freshwater organisms: the use of multiple
Goewert, A., D. Surge, S.J. Carpenter, and J. paleoproxies. Geochimica et Cosmochimi-
Downing. 2007. Oxygen and carbon isotope ca Acta 67 (21): 4033–4047. http://dx.doi.
ratios of Lampsilis cardium (Unionidae) from org/10.1016/S0016-7037(03)00173-X
two streams in agricultural watersheds of Grossman, E.L. and T.-L. Ku. 1986. Oxygen and
Iowa, USA. Palaeogeography, Palaeoclima- carbon isotopic fractionation in biogenic ara-
tology, Palaeoecology 252: 637–648. http:// gonite: temperature effects. Chemical Geolo-
dx.doi.org/10.1016/j.palaeo.2007.06.002 gy 59: 59–74. http://dx.doi.org/10.1016/0168-
Goffer, Z. 2007. Archaeological chemistry, 2nd 9622(86)90057-6
ed. Hoboken, NJ: Wiley-Interscience. http:// Gunter, G. 1947. Observations on breeding of
dx.doi.org/10.1002/0471915254 the marine catfish, Galeichthys felis (Lin-
Goodwin, D. H., B.R. Schöne, and D.L. Dett- naeus). Copeia 4: 217–223. http://dx.doi.
man. 2003. Resolution and fidelity of oxy- org/10.2307/1438916
Habu, J. 2002. Jomon collectors and foragers. meps267253

In B. Fitzhugh and J. Habu (editors), Beyond Harding, J.M., H.J. Spero, R. Mann, G.S. Her-
foraging and collecting: evolutionary change bert, and J.L. Sliko. 2010. Reconstructing
in hunter-gatherer settlement systems: 53–72. early 17th century estuarine drought condi-
New York: Kluwer Academic. tions from Jamestown oysters. Proceedings
Habu, J., and B. Fitzhugh. 2002. Introduction: of the National Academy of Sciences 107:
beyond foraging and collecting: evolutionary 10549–10554. http://dx.doi.org/10.1073/
change in hunter-gatherer settlement systems. pnas.1001052107
In B. Fitzhugh and J. Habu (editors), Beyond Hare, J.A., and K.W. Able. 2007. Mechanistic
foraging and collecting: evolutionary change links between climate and fisheries along
in hunter-gatherer settlement systems: 1–14. the east coast of the United States: explain-
New York: Kluwer Academic. ing population outbursts of Atlantic croak-
Haddon, M. 2001. Modeling and quantitative er (Micropogonias undulatus). Fisheries
methods in fisheries. Boca Raton, FL: Chap- Oceanography 16 (1): 31–45. http://dx.doi.
man and Hall/CRC. 406 p. org/10.1111/j.1365-2419.2006.00407.x
Hallam, A. 1972. Models involving population Hawkes, K. 1996. Foraging differences between
dynamics. In T.J.M. Schoph (editor), Models men and women: behavioural ecology of the
in paleobiology: 62–81. San Francisco: Free- sexual division of labour. In J. Steele and S.
man, Cooper. Shennan (editors), The archaeology of human
Hales, L.S., Jr., and E.J. Reitz. 1992. Historical ancestry: 283–305. New York: Routledge.
changes in age and growth of Atlantic croaker, Hayden, B. 1990. Nimrods, piscators, pluckers,
Micropogonias undulatus (Perciformes: Sci- and planters: the emergence of food produc-
aenidae). Journal of Archaeological Science tion. Journal of Anthropological Archaeology
19: 73–99. http://dx.doi.org/10.1016/0305- 9: 31–69. http://dx.doi.org/10.1016/0278-
4403(92)90008-Q 4165(90)90005-X
Hales, L.S., and M.J. Van Den Avyle. 1989. Spe- Hayden, B. 1995. A new overview of domestica-
cies profiles: life histories and environmental tion. In T.D. Price and A.B. Gebauer (editors),
requirements of coastal fishes and invertebrates Last hunters—first farmers: new perspectives
(South Atlantic): spot. U.S. Fish and Wildlife on the prehistoric transition to agriculture:
Service, Biological Report 82 (11.91). 273–299. Santa Fe: School of American Re-
Ham, L.C., and M. Irvine. 1975. Techniques for search Press.
determining seasonality of shell middens from Hayden, B.P., and R. Dolan. 1979. Barrier islands,
marine mollusc remains. Syesis 8: 363–373. lagoons, and marshes. Journal of Sedimentary
Hames, R. 2007. The ecologically noble savage Petrology 49: 1061–1072.
debate. Annual Review of Anthropology 36: Hayes, M.O. 1994. The Georgia Bight barrier
177–190. http://dx.doi.org/10.1146/annurev. system. In R.A. Davis, Jr. (editor), Geology
anthro.35.081705.123321 of Holocene barrier island systems: 233–304.
Hammen, C.S. 1969. Metabolism of the oyster, Berlin: Springer-Verlag.
Crassostrea virginica. American Zoologist 9 Hayes, R.H., and D.H. Thomas. 2008. Hydrol-
(2): 309–318. ogy of St. Catherines Island. In D.H. Thom-
Hann, J.H. 1986a. Translation of Alonso de Letu- as, Native American landscapes of St. Cath-
riondo’s memorial to the king of Spain; AGI, erines Island, Georgia. I. The theoretical
SD 853, SC. Florida Archaeology 2: 165–225. framework. Anthropological Papers of the
Hann, J.H. 1986b. Demographic patterns and American Museum of Natural History 88:
changes in mid-seventeenth century Timucua 56–58.
and Apalachee. Florida Historical Quarterly Hedges, R.E.M., R.E. Stevens, and M.P. Rich-
64 (4): 371–392. ards. 2004. Bone as a stable isotope archive
Hanson, P.J., C.C. Koenig, and V.S. Zdanowicz. for local climatic information. Quaternary
2004. Elemental composition of otoliths used Science Reviews 23: 959–965. http://dx.doi.
to trace estuarine habitats of juvenile gag My- org/10.1016/j.quascirev.2003.06.022
cteroperca microlepsis along the west coast Hedges, R.E.M., R.E. Stevens, and P.L. Koch.
of Florida. Marine Ecology Progress Series 2005. Isotopes in bones and teeth. In M.J.
267: 253–265. http://dx.doi.org/10.3354/ Leng (editor), Isotopes in palaeoenvironmen-
2012 REFERENCES 221
tal research: 117–145. Dordrecht: Springer. Applied Ecology 46: 572–581. http://dx.doi.
Henry, K.M., and R.M. Cerrato. 2007. The org/10.1111/j.1365-2664.2009.01651.x
annual macroscopic growth pattern of Hoefs, J. 1997. Stable isotope geochemistry. New
the northern quahog (hard clam, Mer- York: Springer.
cenaria mercenaria [L.]), in Narragan- Hoese, H.D., and R.H. Moore. 1998. Fishes of
sett Bay, Rhode Island. Journal of Shell- the Gulf of Mexico: Texas, Louisiana, and ad-
fish Research 26 (4): 985–993. http://dx. jacent waters. 2nd ed. College Station: Texas
doi.org/10.2983/0730-8000(2007)26[985: A&M Press.
TAMGPO]2.0.CO;2 Holland, A.F., and J.M. Dean. 1977. The biol-
Herbert, J.M., and L.C. Steponaitis. 1998. Esti- ogy of the stout razor clam Tagelus plebeius:
mating the season of harvest of eastern oys- II. Some aspects of the population dynam-
ters (Crassostrea virginica) with shells from ics. Chesapeake Science 18: 188–196. http://
the Chesapeake Bay. Southeastern Archaeol- dx.doi.org/10.2307/1350860
ogy 17 (1): 53–71. Hollenbach, K.D. 2009. Foraging in the Tennes-
Heuzé, Y., and J. Braga. 2008. Application of see River Valley: 12,500 to 8,000 years ago.
non-adult Bayesian dental age assessment Tuscaloosa: University of Alabama Press.
methods to skeletal remains: the Spitalfields Howard, J.D., and R.W. Frey. 1985. Physical
collection. Journal of Archaeological Science and biogenic aspects of backbarrier sedi-
35 (2): 368–375. http://dx.doi.org/10.1016/j. mentary sequences, Georgia coast, U.S.A.
jas.2007.04.003 Marine Geology 63: 77–127. http://dx.doi.
Higham, T.F.G., and P.L. Horn. 2000. Seasonal org/10.1016/0025-3227(85)90080-5
dating using fish otoliths: results from the Hoyt, J.H. 1967. Barrier island formation. Geo-
Shag River Mouth Site, New Zealand. Jour- logical Society of America Bulletin 78 (9):
nal of Archaeological Science 27: 439–375. 1125–1136. http://dx.doi.org/10.1130/0016-
http://dx.doi.org/10.1006/jasc.1999.0473 7606(1967)78[1125:BIF]2.0.CO;2
Hillard, R., and K. Walters. 2009. Prevalence, Hoyt, J.H., and J.R. Hails. 1967. Pleistocene
patterns, and effects of shell damage on Geu- shoreline sediments in coastal Georgia: de-
kensia demissa in South Carolina estuarine position and modification. Science 155:
habitats. Marine Biology 156: 2149–375. 1541–1543. http://dx.doi.org/10.1126/sci-
http://dx.doi.org/10.1007/s00227-009-1245-7 ence.155.3769.1541
Hillman, G. 1984. Interpretation of archaeologi- Hubbard, D.K., G. Oertel, and D. Nummedal.
cal plant remains: the application of ethno- 1979. The role of waves and tidal currents in
graphic models from Turkey. In W. Van Zeist the development of tidal-inlet sedimentary
and W.A. Casparie (editors), Plants and an- structures and sand body geometry: examples
cient man: studies in paleoethnobotany: 1–42. from North Carolina, South Carolina, and
Rotterdam: A.A. Balkans. Georgia. Journal of Sedimentary Petrology
Hillman, R.E., N.W. Davis, and J. Wennemer. 49 (4): 1073–1092.
1977. Abundance, diversity, and stability in Hudson, J.H., E.A. Shinn, R.B. Halley, and B.
shore-zone fish communities in an area of Lidz. 1976. Sclerochronology: a tool for in-
Long Island Sound affected by the thermal terpreting past environments. Geology 4 (6):
discharge of a nuclear power station. Es- 361–364. http://dx.doi.org/10.1130/0091-
tuarine and Coastal Marine Science 5 (3): 7613(1976)4<361:SATFIP>2.0.CO;2
355–381. http://dx.doi.org/10.1016/0302- Hufthammer, A.K., H. Høie, A. Folkvord, A.J.
3524(77)90062-7 Geffen, C. Andersson, and U.S. Ninnemann.
Hobson, K.A. 1999. Tracing origins and migra- 2010. Seasonality of human site occupa-
tion of wildlife using stable isotopes: a re- tion based on stable oxygen isotope ratios of
view. Oecologia 120: 314–326. http://dx.doi. cod otoliths. Journal of Archaeological Sci-
org/10.1007/s004420050865 ence 37: 78–83. http://dx.doi.org/10.1016/j.
Hobson, K.A., H. Lormée, S.L. Van Wilgenburg, jas.2009.09.001
L.I. Wassenaar, and J.M. Boutin. 2009. Stable Ingold, T. 1996. Growing plants and raising ani-
isotopes (δD) delineate the origins and migra- mals: an anthropological perspective on do-
tory connectivity of harvested animals: the mestication. In D. Harris (editor), The origins
case of European woodpigeons. Journal of and spread of agriculture and pastoralism in
Eurasia: 12–24. London: UCL Press. topes and season of annual increment for-
Irving, W.S. 1902. Brunswick and Glynn Coun- mation. Palaios 11: 340–346. http://dx.doi.
ty, Georgia. Issued by the Board of Trade, org/10.2307/3515244
Brunswick, GA. Jones, D.S., and S.J. Gould. 1999. Direct mea-
Ishimaru, E., T. Umino, M. Yoneda, and I. Taya- surement of age in fossil Gryphaea: the so-
su. 2006. Identification of origin of faunal re- lution to a classic problem in heterochrony.
mains: carbon and nitrogen isotope analysis Paleobiology 25 (2): 158–187.
of marine fishes. Mexico City, Mexico: Inter- Jones, D.S., M.A. Arthur, and D.J. Allard. 1989.
national Council for Archaeozoology. Internet Sclerochronological records of temperature
resource (http://www.alexandriaarchive.org/ and growth from shells of Mercenaria merce-
icaz/pdf/poster_ishimaru_etal.pdf), accessed naria from Narragansett Bay, Rhode Island.
on December 14, 2011. Marine Biology 102: 225–234. http://dx.doi.
Jassby, A.D.,W.J. Kimmerer, S.G. Monismith, org/10.1007/BF00428284
C. Armor, J.E. Cloern, T.M. Powell, J.R. Jones, D.S., I.R. Quitmyer, W.S. Arnold, and D.C.
Schubel, and T.J. Vendlinski. 1995. Isohaline Marelli. 1990. Annual shell banding, age, and
position as a habitat indicator for estuarine growth rate of the hard shell clams (Merce-
populations. Ecological Applications 5 (1): naria spp.) from Florida. Journal of Shellfish
272–289. http://dx.doi.org/10.2307/1942069 Research 9 (1): 215–225.
Jochim, M. 2006. The implications of inter-group Jones, D.S., I.R. Quitmyer, and C.F.T. Andrus.
food exchange for hunter-gatherer affluence 2004. Seasonal shell growth and longevity in
and complexity. In C. Grier, J. Kim, and J. Donax variabilis from northeastern Florida:
Uchiyama (editors), Beyond affluent forag- evidence from oxygen isotopes. Journal of
ers: rethinking hunter-gatherer complexity: Shellfish Research 23: 707–714.
80–89. Oxford: Oxbow Books. Jones, D.S., I.R. Quitmyer, and C.F.T. An-
Johnson, A.S., H.O. Hillestad, S.F. Shanholtzer, drus. 2005. Oxygen isotopic evidence for
and G.F. Shanholtzer. 1974. An ecological greater seasonality in Holocene shells of
survey of the coastal region of Georgia. Na- Donax variabilis from Florida. Palaeogeog-
tional Park Service Scientific Monograph raphy, Palaeoclimatology, Palaeoecology
Series 3. 228: 96–108. http://dx.doi.org/10.1016/j.pa-
Johnson, D.R., and W. Seaman, Jr. 1986. Species laeo.2005.03.046
profiles: life histories and environmental Jones, G.D. 1978. The ethnohistory of the Guale
requirements of coastal fishes and inverte- coast through 1684. In D. H. Thomas, G.D.
brates (South Florida): spotted seatrout. U.S. Jones, R.S. Durham, and C.S. Larsen, The an-
Fish and Wildlife Service Biological Report thropology of St. Catherines Island. 1. Natural
82 (11.43). and cultural history. Anthropological Papers
Jones, C.A. 2010. Mineralogy and seasonal growth of the American Museum of Natural History
of South Pacific mussel valves. Master’s the- 55 (2): 178–210.
sis, University of Alabama, Tuscaloosa. Jones, G.D. 1980. Guale Indians of the southeast-
Jones, C.C., Jr. 1873. Antiquities of the southern ern United States coast. In J.D. Howard, C.B.
Indians, particularly of the Georgia tribes. DePratter, and R.W. Frey (editors), Georgia
New York: D. Appleton. Geological Survey Guidebook: 215–224.
Jones, D.C., C.T. Espenshade, and L. Ken- Jones, G.E.M. 1984. Interpretation of archaeo-
nedy. 1997. Archaeological investigations logical plant remains: ethnographic models
at 31ON190, Cape Island, Onslow County, from Greece. In W. Van Zeist and W.A. Cas-
North Carolina. Ms. on file (OSA # 4069), parie (editors), Plants and ancient man: stud-
North Carolina Office of State Archaeology, ies in paleoethnobotany: 43–62. Rotterdam:
Raleigh, NC. A.A. Balkans.
Jones, D.S. 1980. Annual cycle of shell growth Jones, J.B., and S.E. Campana. 2009. Stable oxy-
increment formation in two continental shelf gen isotope reconstruction of ambient tem-
bivalves and its paleoecologic significance. perature during the collapse of a cod (Gadus
Paleobiology 6 (3): 331–340. morhua) fishery. Ecological Applications 19
Jones, D.S., and I.R. Quitmyer. 1996. Mark- (6): 1500–1514. http://dx.doi.org/10.1890/07-
ing time with bivalve shells: oxygen iso- 2002.1
2012 REFERENCES 223
Jones, K.B., G.W.L. Hodgins, D. Dettman, C.F. among Madagascar’s Mikae. American An-
T. Andrus, A. Nelson, A., and M.F. Etayo- thropologist 107 (3): 403–416. http://dx.doi.
Cadavid. 2007. Seasonal variations in Peru- org/10.1525/aa.2005.107.3.403
vian marine reservoir age from pre-bomb Ar- Kemp, P. and M.B. Bertness. 1984. Snail shape
gopecten purpuratus shell carbonate. Radio- and growth rates: evidence for plastic shell
carbon 49: 877–888. allometry in Littorina littorea. Proceedings
Jones, K.B., G.W.L. Hodgins, C.F.T. Andrus, and of the National Academy of Sciences of the
M.F. Etayo-Cadavid. 2010a. Modeling mol- United States of America 81: 811–813. http://
luscan marine reservoir ages in a variable- dx.doi.org/10.1073/pnas.81.3.811
upwelling environment. Palaios 25: 126–131. Kennedy, B.P., C.P. Chamberlain, J.D. Blum,
http://dx.doi.org/10.2110/palo.2009.p09-050r K.H. Nislow, and C.L. Folt. 2005. Compar-
Jones, K.B., G.W.L. Hodgins, M.F. Etayo-Ca- ing naturally occurring stable isotopes of
david, C.F.T. Andrus, and D.H. Sandweiss. nitrogen, carbon, and strontium as markers
2010b. Centuries of marine radiocarbon res- for the rearing locations of Atlantic salmon
ervoir age variation within archaeological (Salmo salar). Canadian Journal of Fisheries
Mesodesma donacium shells from southern and Aquatic Science 62: 48–57. http://dx.doi.
Peru. Radiocarbon 52 (3): 1207–1214. org/10.1139/f04-184
Jones, T.L. and D.J. Kennett. 1999. Late Holo- Kennett, D.J., and J.P. Kennett. 2000. Competi-
cene sea temperatures along the central Cali- tive and cooperative responses to climatic in-
fornia coast. Quaternary Research 51: 74–82. stability in coastal southern California. Amer-
http://dx.doi.org/10.1006/qres.1998.2000 ican Antiquity 65: 379–395. http://dx.doi.
Jones, T.L., D.J. Kennett, J.P. Kennett, and B.F. org/10.2307/2694065
Codding. 2008. Seasonal stability in Late Kennett, D. J., and B. Voorhies. 1995. Middle Ho-
Holocene shellfish harvesting on the central locene periodicities in rainfall inferred from
California coast. Journal of Archaeological oxygen and carbon isotopic fluctuations in
Science 35 (8): 2286–2294. http://dx.doi. prehistoric tropical estuarine mollusc shells.
org/10.1016/j.jas.2008.03.002 Archaeometry 37: 157–170. http://dx.doi.org/
Jorgenson, S.C., and G.L. Miller. 1968. Length 10.1111/j.1475-4754.1995.tb00734.x
relations of some marine fishes from coastal Kennett, D. J., and B. Voorhies. 1996. Oxygen
Georgia. Brunswick, GA: Bureau of Com- isotopic analysis of archaeological shells
mercial Fishes Biological Laboratory Con- to detect seasonal use of wetlands on the
tribution 93. southern Pacific coast of Mexico. Journal of
Kaplan, H., and K. Hill. 1992. The evolutionary Archaeological Science 23: 689–704. http://
ecology of food acquisition. In E.A. Smith dx.doi.org/10.1006/jasc.1996.0065
and B. Winterhalder (editors), Evolutionary Kennett, D.J., B.L. Ingram, J. Erlandson, and P.
ecology and human behavior: 167–202. New Walker. 1997. Evidence for temporal fluc-
York: Aldine de Gruyter. tuations of marine radiocarbon reservoir
Keene, D.A. 2004. Reevaluating Late Prehistoric ages in the Santa Barbara Channel region,
coastal subsistence and settlement strategies: California. Journal of Archaeological Science
new data from Grove’s Creek Site, Skidaway 24: 1051–1059. http://dx.doi.org/10.1006/
Island, Georgia. American Antiquity 69 (4): jasc.1996.0184
671–688. http://dx.doi.org/10.2307/4128443 Kennett, D.J., B.L Ingram, J.R. Southon, and K.
Kelly, R.L. 1992. Mobility/sedentism: con- Wise. 2002. Differences in 14C age between
cepts, archaeological measures, and ef- stratigraphically associated charcoal and
fects. Annual Review of Anthropology 21: marine shell from the Archaic period site of
43–66. http://dx.doi.org/10.1146/annurev.an. Kilometer 4, southern Peru: old wood or old
21.100192.000355 water? Radiocarbon 44 (1): 53–58.
Kelly, R.L. 1995. The foraging spectrum: diver- Kennett, D.J., B.J. Culleton, B. Voorhies, and
sity in hunter-gatherer lifeways. Washington, J. Southon. 2011. Bayesian analysis of high
DC: Smithsonian Institution Press. precision AMS 14C dates from a prehistoric
Kelly, R.L., L. Poyer, and B. Tucker. 2005. An Mexican shellmound. Radiocarbon.
ethnoarchaeological study of mobility, ar- Kennish, M.J. 1980. Shell microgrowth analy-
chitectural investment, and food sharing sis: Mercenaria mercenaria as type example
for research in population dynamics. In D.D. Koike, H. 1979. Seasonal dating and the valve-
Rhoads and R.A. Lutz (editors), Skeletal pairing technique in shell-midden analysis.
growth of aquatic organisms: 255–294. New Journal of Archaeological Science 6 (1):
York: Plenum Press. 63–74. http://dx.doi.org/10.1016/0305-4403
Kennish, M.J., and R.K. Olsson. 1975. Effects (79)90033-5
of thermal discharges on the microstructural Kraeuter, J.N., and M. Castagna. 2001. Biology
growth of Mercenaria mercenaria. Envi- of the hard clam. New York: Elsevier.
ronmental Geology 1: 41–64. http://dx.doi. Kroeber, A.L. 1925. Handbook of the Indians of
org/10.1007/BF02426940 California. Bureau of American Ethnology
Kent, B. 1988. Making dead oysters talk: tech- Bulletin 78.
niques for analyzing oysters from archaeo- Lambert, C. 2007. Comings and goings on the
logical sites. St. Mary’s City: Maryland His- seashore. No Wake Zone 2 (12): 2.
torical Trust. Larsen, C.S., and D.H. Thomas. 1982. The an-
Kent, B.W. 1992. Making dead oysters talk: thropology of St. Catherines Island: 4. The St.
techniques for analyzing oysters from 595 Catherines period mortuary complex. Anthro-
archaeological sites. Crownsville: Maryland pological Papers of the American Museum of
Historical and Cultural Publications. Natural History 57 (4): 271–342.
Kenward, H., and A. Hall. 1997. Enhancing bio- Larson, L.H., Jr. 1969. Aboriginal subsistence
archaeological interpretation using indicator technology on the Southeastern Coastal
groups: stable manure as a paradigm. Jour- Plain during the late prehistoric period.
nal of Archaeological Science 24: 663–673. Ms., University of Michigan, Department
http://dx.doi.org/10.1006/jasc.1996.0149 of Anthropology.
Killingley, J.S., and W.H. Berger. 1979. Stable Larson, L.H., Jr. 1978. Historic Guale Indians of
isotopes in a mollusk shell: detection of up- the Georgia coast and the impact of the Span-
welling events. Science 205: 186–188. http:// ish mission effort. In J. Milanich and S. Proc-
dx.doi.org/10.1126/science.205.4402.186 tor (editors), Tacachale: essays of the Indians
Kim, S.T., A. Mucci, and B. Taylor. 2007. Phos- of Florida and southeastern Georgia during
phoric acid fractionation factors for calcite the historic period: 120–140. Gainesville:
and aragonite between 25 and 75 °C: revis- University of Florida Press.
ited. Chemical Geology 246: 135–146. http:// Larson, L.H., Jr. 1980. Aboriginal subsistence
dx.doi.org/10.1016/j.chemgeo.2007.08.005 technology on the Southeastern Coastal Plain
Kirby, M.X., T.M. Soniat, and H.J. Spero. 1998. during the late prehistoric period. Gainesville:
Stable isotope sclerochronology of Pleisto- University of Florida Press.
cene and recent oyster shells (Crassostrea Lassuy, D.R. 1983. Species profiles: life histories
virginica). Palaois 13: 560–569. http://dx.doi. and environmental requirements of coastal
org/10.2307/3515347 fishes and invertebrates (Gulf of Mexico): At-
Klein, M.J. 1999. Shell midden archaeology: lantic croaker. U.S. Fish and Wildlife Service,
gender, labor, and stone artifacts. In N.L. Biological Report 82 (11.3).
Wicker and B. Arnold (editors), From the Laufer, B. 1913. The relationship of archeology
ground up: beyond gender theory in archaeol- to ethnology: remarks. American Anthropolo-
ogy: 143–154. BAR International Series 812. gist 15: 573–577.
Koch, P.L. 2007. Isotopic study of the biology of Lécuyer, C., B. Reynard, and F. Martineau. 2004.
modern and fossil vertebrates. In R. Michen- Stable isotope fractionation between mollusc
er and K. Lajtha (editors), Stable isotopes shells and marine waters from Martinique Is-
in ecology and environmental science, 2nd land. Chemical Geology 213: 293–305. http://
ed.: 99–154. Malden, MA: Blackwell. http:// dx.doi.org/10.1016/j.chemgeo.2004.02.001
dx.doi.org/10.1002/9780470691854.ch5 Lee, C.H., I.R. Quitmyer, C.T. Espenshade, and
Koike, H. 1975. The use of daily and annual R.E. Johnson. 1984. Estuarine adaptations
growth lines of the clam Meretrix lusoria during the Late Prehistoric period: archaeol-
in estimating seasons of Jomon period shell ogy of two shell midden sites on the St. Johns
gathering. In R.P. Suggate and M.M. Cress- River. Pensacola, FL: University of West
well (editors), Quaternary studies: 189–193. Florida, Office of Cultural and Archaeologi-
Wellington: Royal Society of New Zealand. cal Research, Report of Investigations 5.
2012 REFERENCES 225
Lee, R.B., and R. Daly (editors). 1999. The Cam- ogy, Raleigh, NC.
bridge encyclopedia of hunters and gatherers. Lomovaskya, B.L., J.L. Gutiérrez, and O.O. Irib-
Cambridge: Cambridge University Press. arnea. 2005. Identifying repaired shell dam-
Lee, R.B., and I. DeVore. 1968. Problems in the age and abnormal calcification in the stout
study of hunters and gatherers. In R.B. Lee razor clam Tagelus plebeius as a tool to in-
and I. DeVore (editors), Man the hunter: 3–12. vestigate its ecological interactions. Journal
Chicago: Aldine-Atherton. of Sea Research 54: 163–175. http://dx.doi.
Lemonier, P. 1986. The study of material cul- org/10.1016/j.seares.2005.01.004
ture today: toward an anthropology of tech- Lowerre-Barbieri, S.K., L.R. Barbieri, J.R. Flan-
nical systems. Journal of Anthropological ders, A.G. Woodward, C.F. Cotton, and K.
Archaeology 5 (2): 147–186. http://dx.doi. Knowlton. 2008. Use of passive acoustics
org/10.1016/0278-4165(86)90012-7 to determine red drum spawning in Georgia
Leroi-Gourhan, A. 1964. Le geste et la parole, waters. Transactions of the American Fisher-
tome 1: technique et language. Paris: Albin ies Society 137 (2): 562–575. http://dx.doi.
Michel. org/10.1577/T04-226.1
Leroi-Gourhan, A. 1993. Gesture and speech (Le Luby, E.M., C.D. Drescher, and K.G. Lightfoot.
geste et la parole). Translated from French by 2006. Shell mounds and mounded landscapes
Anna Bostock Berger. Cambridge: MIT Press. in the San Francisco Bay area: an integrat-
Levin, L.A., D.F. Boesch, A. Covich, C. Dahm, ed approach. Journal of Island and Coastal
C. Erseus, K.C. Ewel, R.T. Kneib, A. Mold- Archaeology 1: 191–214. http://dx.doi.
enke, M.A. Palmer, P. Snelgrove, D. Strayer, org/10.1080/15564890600935365
and J.M. Weslawski. 2001. The function of Lutz, R.A. and D.C. Rhoads.1977. Anaerobiosis
marine critical transition zones and the impor- and a theory of growth line formation. Sci-
tance of sediment biodiversity. Ecosystems ence 198 (4323): 1222–1227. http://dx.doi.
4 (5): 430–451. http://dx.doi.org/10.1007/ org/10.1126/science.198.4323.1222
s10021-001-0021-4 Lynch, A.H., J. Hamilton, and R.E.M. Hedges.
Lightfoot, K.G., and O. Parrish. 2009. California 2008. Where the wild things are: aurochs
Indians and their environment: an introduc- and cattle in England. Antiquity 82 (318):
tion. Berkeley: University of California Press. 1025–1039.
Liu, K.-B. 2004. Paleotempestology: principles, MacKenzie, C.L., Jr. 2008. History of the bay
methods, and examples from Gulf coast lake scallop, Argopecten irradians, fisheries and
sediments. In R.J. Murnane and K.-B. Liu habitats in eastern North America, Massachu-
(editors), Hurricanes and typhoons: past, pres- setts through northeastern Mexico. Marine
ent, and future: 13–95. New York: Columbia Fisheries Review 70 (3–4): 1–5.
University Press. Malinowski, S.R. 1985. The population ecology
Livingston, R.J. 1975. Impact of Kraft pulp-mill of the hard clam, Mercenaria mercenaria, in
effluents on estuarine and coastal fishes in eastern Long Island Sound. Ph.D. disserta-
Apalachee Bay, Florida, USA. Marine Biol- tion. University of Connecticut, Storrs.
ogy 32 (1): 19–48. http://dx.doi.org/10.1007/ Mannino, M.A., and K.D. Thomas. 2002. De-
BF00395158 pletion of a resource? the impact of prehis-
Livingston, R.J. 1982. Trophic organization toric human foraging on intertidal mollusc
of fishes in a coastal seagrass system. Ma- communities and its significance for human
rine Ecology Progress Series 7: 1–12. http:// settlement, mobility and dispersal. World
dx.doi.org/10.3354/meps007001 Archaeology 33 (3): 452–474. http://dx.doi.
Livingston, R.J. 1985. Organization of fishes in org/10.1080/00438240120107477
coastal seagrass system: the response to stress. Mannino M.A., B.F. Spiro, and K.D. Thomas.
In A. Yáñez-Arancibia (editor), Fish commu- 2003. Sampling shells for seasonality: oxy-
nity ecology in estuaries and coastal lagoons: gen isotope analysis on shell carbonates of
towards an ecosystem integration: 367–382. the inter-tidal gastropod Monodonta lineata
Mexico City, Mexico: UNAM Press. (da Costa) from populations across its mod-
Loftfield, T.C. 1979. Excavations at 31 Onv 33: a ern range and from a Mesolithic site in south-
Late Woodland seasonal village (OSA # 915). ern Britain. Journal of Archaeological Sci-
Manuscript on file, Office of State Archaeol- ence 30: 667–679. http://dx.doi.org/10.1016/
S0305-4403(02)00238-8 ganese, and barium incorporation in otoliths

Mannino M.A., K.D. Thomas, M.J. Leng, M. of larval and early juvenile spot Leiostomus
Piperno, S. Tusa, and A. Tagliacozzo. 2007. xanthurus. Marine Ecology Progress Series
Marine resources in the Mesolithic and Neo- 293: 223–232. http://dx.doi.org/10.3354/
lithic at the Grotta dell’Uzzo (Sicily): evi- meps293223
dence from isotope analyses of marine shells. Martin, K.L.M., R. C.V. Winkle, J.E. Drais, and
Archaeometry 49: 117–133. http://dx.doi. H. Lakisic. 2004. Beach-spawning fishes, ter-
org/10.1111/j.1475-4754.2007.00291.x restrial eggs, and air breathing. Physiological
Marelli, D.C., and W.S. Arnold. 2001. Shell mor- and Biochemical Zoology 77 (5): 750–759.
phologies of bay scallops, Argopecten irradi- http://dx.doi.org/10.1086/421755
ans, from extant and prehistoric populations Martin-Mora, E., E.C. James, and A.W. Stoner.
from the Florida Gulf Coast: implications 1995. Developmental plasticity in the shell of
for the biology of past and present metapo- the queen conch Strombus gigas. Ecology 76:
pulations. Journal of Archaeological Sci- 981–994.
ence 28: 577–586. http://dx.doi.org/10.1006/ Matthews, J.H. 1928. Fisheries of the south Atlan-
jasc.2000.0565 tic and Gulf states. Economic Geography 4 (4):
Marquardt, W.H. 1996. Four discoveries: envi- 323–348. http://dx.doi.org/10.2307/140392
ronmental archaeology in southwest Florida. McClanahan, T.R., and J.E. Cinner. 2008. A
In E.J. Reitz, L.A. Newsom, and S.J. Scudder framework for adaptive gear and ecosystem-
(editors), Case studies in environmental ar- based management in the artisanal coral reef
chaeology: 17–34. New York: Plenum Press. fishery of Papua New Guinea. Aquatic Con-
Marquardt, W.H. 2010a. Mounds, middens, and servation: Marine and Freshwater Ecosys-
rapid climate change during the Archaic- tems 18: 493–507. http://dx.doi.org/10.1002/
Woodland transition in the southeastern Unit- aqc.874
ed States. In D.H. Thomas and M.C. Sanger McConnaughey, T.A., and D.P. Gillikin.2008.
(editors and contributors), Trend, tradition, Carbon isotopes in mollusk shell carbon-
and turmoil: what happened to the southeast- ates. Geo-Marine Letters 28: 287–299. http://
ern Archaic? Anthropological Papers of the dx.doi.org/10.1007/s00367-008-0116-4
American Museum of Natural History 93: McLusky, D.S., and M. Elliott. 2004. The es-
253–271. tuarine ecosystem: ecology, threats, and
Marquardt, W.H. 2010b. Shell mounds in the management. Oxford University Press,
Southeast: middens, monuments, temple Oxford. http://dx.doi.org/10.1093/acprof:o
mounds, rings, or works? American Antiquity so/9780198525080.001.0001
75 (3): 551–570. Meehan, B. 1982. Shell bed to shell midden.
Marrinan, R.A. 1975. Molluscs, ceramics and Canberra: Australian Institute of Aboriginal
sedentism: the Late Archaic period on the Studies.
Georgia Coast. Ph.D. dissertation, Depart- Messner, T.C. 2011. Acorns and bitter roots:
ment of Anthropology, University of Florida, starch grain research in the prehistoric East-
Gainesville. ern Woodlands. Tuscaloosa: University of
Marrinan, R.A. 2010. Two Late Archaic period Alabama Press.
shell rings, St. Simon’s Island, Georgia. In Milton, K. 1993. Diet and primate evolution. Sci-
D.H. Thomas and M.C. Sanger (editors and entific American 269 (2): 86–93. http://dx.doi.
contributors), Trend, tradition, and turmoil: org/10.1038/scientificamerican0893-86
what happened to the southeastern Archaic? Monks, G.G. 1981. Seasonality studies. In M.B.
Anthropological Papers of the American Mu- Schiffer (editor), Advances in archaeological
seum of Natural History 93: 71–102. method and theory 4: 177–240. New York:
Marshall, R.R. 1986. A compilation of materials Academic Press.
relating to the proposed Stump Sound archae- Mook, W.G., and J.C. Vogel. 1968. Isotopic equi-
ological district, Onslow County, North Caro- librium between shells and their environment.
lina (OSA #2193). Manuscript on file, Office Science 159 (3817): 874–875. http://dx.doi.
of State Archaeology, Raleigh, NC. org/10.1126/science.159.3817.874
Martin, G.B., and S.R. Thorrold. 2005. Tempera- Morales Muñiz, A. 1998. The mobile faunas:
ture and salinity effects on magnesium, man- reliable seasonal indicators for archaeozoolo-
2012 REFERENCES 227
gists? In T.R. Rocek and O. Bar-Yosef (edi- Academic.

tors), Seasonality and sedentism: archaeologi- Odum, E.P., and G.W. Barrett. 2005. Fundamen-
cal perspectives from Old and New World tals of ecology. 5th ed. Belmont, California:
sites: 25–39. Peabody Museum of Archaeolo- Thomson Brooks/Cole.
gy and Ethnology Bulletin No. 6, Cambridge, Odum, H.T. 1971. Economy, power, and society.
MA: Harvard University. New York: Wiley-Interscience.
Muncy, R.J., and W.M. Wingo. 1983. Species Oertel, G.F. 1979. Barrier island development
profiles: life histories and environmental re- during the Holocene recession, southeastern
quirements of coastal fishes and invertebrates United States. In S.P. Leatherman (editor),
(Gulf of Mexico): sea catfish and gafftopsail Barrier islands: from the Gulf of St. Lawrence
catfish. U.S. Fish and Wildlife Service, Bio- to the Gulf of Mexico: 273–290. New York:
logical Report 82 (11.5). Academic Press.
Munson, P.J. 1986. Hickory silvaculture: a sub- Orr, K.L. 2007. Coastal Weeden Island subsis-
sistence revolution in the prehistory of eastern tence adaptations: zooarchaeological evidence
North America. Paper presented at the Con- from Bayou St. John (1BA21), Alabama.
ference on Emergent Horticultural Economies Ph.D. dissertation, Department of Anthropol-
of the Eastern Woodlands. Center for Archae- ogy, University of Georgia, Athens.
ological Investigations. Urbana: Southern Il- Osborn, A.J. 1977. Strandloopers, mermaids,
linois University. and other fairy tales: ecological determinants
Murray, G., B. Neis, and J.P. Johnsen. 2006. of marine resource utilization—the Peruvian
Lessons learned from reconstructing interac- case. In L.R. Binford (editor), For theory
tions between local ecological knowledge, building in archaeology: 157–250. New York:
fisheries science, and fisheries management Academic Press.
in the commercial fisheries of Newfound- Owen, R., H. Kennedy, and C. Richardson.
land and Labrador, Canada. Human Ecology 2002. Isotopic partitioning between scal-
34 (4): 549–571. http://dx.doi.org/10.1007/ lop shell calcite and seawater: effect of
s10745-006-9010-8 shell growth rate. Geochimica et Cosmo-
Nagelkerken, I., and G. van der Velde. 2004. chimica Acta 66: 1727–1737. http://dx.doi.
Are Caribbean mangroves important feeding org/10.1016/S0016-7037(01)00882-1
grounds for juvenile reef fish from adjacent Pannella, G. 1971. Fish otoliths: daily growth
seagrass beds? Marine Ecology Progress Se- layers and periodical patterns. Science 173:
ries 274: 143–151. http://dx.doi.org/10.3354/ 1124–1126. http://dx.doi.org/10.1126/sci-
meps274143 ence.173.4002.1124
Nelson, D.M., E.A. Irlandi, L.R. Settle, M.E. Pannella, G. 1980. Growth patterns in fish sag-
Monaco, and L. Coston-Clements. 1991. Dis- ittae. In D.C. Rhoads and R.A. Lutz (edi-
tribution and abundance of fishes and inverte- tors), Skeletal growth of aquatic organisms:
brates in Southeast estuaries. ELMR Report biological records of environmental change:
9. Silver Spring, MD: NOAA/NOS Strategic 519–560. New York: Plenum Press.
Environmental Assessments Division. Parnell, A.C., J. Haslett, J.R.M. Allen, C.E. Buck,
Nishimura, A. 1975. Cultural and social change in and B. Huntley. 2008. A flexible approach to
the modes of ownership of stone tidal weirs. assessing synchroneity of past events using
In R. W. Casteel and G. I. Quimby (editors), Bayesian reconstructions of sedimentation
Maritime adaptations of the Pacific: 77–88. history. Quaternary Science Reviews 27:
The Hague: Mouton Publishers. 1872–1885. http://dx.doi.org/10.1016/j.quas-
O’Brien, D.M., and D.H. Thomas. 2008. The cirev.2008.07.009
molluscan incremental sequence. In D.H. Parsons, A. 2008. The Grand Shell Ring: a study
Thomas, Native American landscapes of St. of site seasonality, hard clam exploitation, and
Catherines Island, Georgia. II. The data. An- resource scheduling. Master’s thesis. Florida
thropological Papers of the American Muse- State University Tallahassee.
um of Natural History 88: 475–497. Payne, T.M., and B.H. Dahlin. 1987. A phase I
O’Brien, M.J., and R.L. Lyman. 1999. Seriation, archaeological survey 31BF115 and 31BF117
stratigraphy, and index fossils: the backbone and a phase II archaeological survey 31BF115
of archaeological dating. New York: Kluwer Texasgulf, Bath Creek, North Carolina (OSA
# 2208). Ms. on file, Office of State Archaeol- Price, S.E., and G.A. Waselkov. 2009. Bone and
ogy, Raleigh, NC. shell tools. In S.E. Price (editor), Archaeol-
Peacock, E., and J.L. Seltzer. 2008. A comparison ogy at Orange Beach: phase III data recovery
of multiple proxy data sets for paleoenviron- at 1BA21, the Bayou St. John site, Orange
mental conditions as derived from freshwater Beach, Baldwin County, Alabama: 144–169.
bivalve (Unionid) shell. Journal of Archaeo- Mobile, AL: Center for Archaeological Stud-
logical Science 35: 2557–2565. http://dx.doi. ies, University of South Alabama. Internet
org/10.1016/j.jas.2008.04.006 resource (http://www.usouthal.edu/archaeol-
Peacock, E., W.R. Haag, and M.L. Warren, Jr. ogy/temp-file-access/001.pdf), accessed on
2004. Prehistoric decline in freshwater mus- December 14, 2011.
sels coincident with the advent of maize Quitmyer, I.R. 2003. Zooarchaeological remains
agriculture. Conservation Biology 19 (2): from Bottle Creek. In I.W. Brown (editor),
547–551. http://dx.doi.org/10.1111/j.1523- Bottle Creek: a Pensacola culture site in south
1739.2005.00036.x Alabama: 130–155. Tuscaloosa: University of
Pearson, C.E. 1977. Analysis of Late Prehistoric Alabama Press.
settlement on Ossabaw Island, Georgia. Uni- Quitmyer I.R., and D.S. Jones. 1992. Calendars
versity of Georgia, Laboratory of Archaeol- of the coast: seasonal growth increment pat-
ogy, Series Report 12. terns in shells of modern and archaeological
Petruso, K.M., and J.M. Wickens. 1984. The southern quahogs, Mercenaria campechien-
acorn in aboriginal subsistence in eastern sis, from Charlotte Harbor, Florida. In W.H.
North America: a report on miscellaneous Marquardt, Culture and the environment in
experiments. In P.J. Munson (editor), Experi- the domain of the Calusa: 247–262. Gaines-
ments and observations on aboriginal wild ville, Institute of Archaeology and Paleoenvi-
plant food utilization in eastern North Ameri- ronmental Studies.
can: 360–378. Prehistoric Research Series VI Quitmyer, I.R., and D.S. Jones. 2000. The over
(2). Indianapolis: Indiana Historical Society. exploitation of hard clams (Mercenaria spp.)
Pollard, M., C. Batt, B. Stern, and S.M.M. from five archaeological sites in the south-
Young. 2007. Analytical chemistry in ar- eastern United States. Florida Anthropologist
chaeology. Cambridge: Cambridge Uni- 53 (2–3): 160–167.
versity Press. http://dx.doi.org/10.1017/ Quitmyer, I.R., and M.J. LeFebvre. 2004. The
CBO9780511607431 zooarchaeology of Ribault Clubhouse Mid-
Porter, H.J., L.A. Howie, and R.B. Deriso. 1979. den (8DU76), Duval County, Florida. Ms. on
Morphometric character variation in Boonea file, Environmental Archaeology Laboratory,
impressa (Say) and B. seminuda (C.B. Ad- Florida Museum of Natural History, Univer-
ams)—family Pyramidellae. Bulletin of the sity of Florida, Gainesville.
American Malacological Union for 1979: Quitmyer, I.R., and E.J. Reitz. 2006. Marine
43–48. trophic levels targeted between a.d. 300 and
Powell, E.N., M.E. White, E.A. Wilson, and S.M. 1500 on the Georgia coast, USA. Journal of
Ray. 1987. Change in host preference with age Archaeological Science 33: 806–822.
in the ectoparasitic pyramidellid snail Boonea Quitmyer, I.R., H.S. Hale, and D.S. Jones. 1985.
impressa (Say). Journal of Molluscan Studies Paleoseasonality determination based on
53: 285–286. http://dx.doi.org/10.1093/mol- incremental shell growth in the hard clam,
lus/53.3.285 Mercenaria mercenaria, and its implications
Power, A. J., C.J. Sellers, and R.L. Walker. 2009. for future analysis of three Georgia coastal
Growth and sexual maturity of the knobbed shell middens. Southeastern Archaeology 4
whelk, Busycon carica (Gmelin, 1791), from (1): 27–40.
a commercially harvested population in coast- Quitmyer, I.R., D.S. Jones, and W.S. Arnold.
al Georgia. Occasional Papers of the Univer- 1997. The sclerochronology of hard clams,
sity of Georgia Marine Extension Service 4. Mercenaria spp., from the south-eastern
Price, S.E. 2008. Phase III archaeology at Plash U.S.A.: a method of elucidating the zooar-
Island archaeological site 1BA134 in Bald- chaeological records of seasonal resource pro-
win County, Alabama. Mobile: University of curement and seasonality in prehistoric shell
South Alabama. middens. Journal of Archaeological Science
2012 REFERENCES 229
24 (9): 825–840. http://dx.doi.org/10.1006/ dx.doi.org/10.2307/4128348

jasc.1996.0163 Reitz, E.J. 2008. Nonhuman vertebrate remains.
Quitmyer, I.R., D.S. Jones, and C.F.T. Andrus. In D.H. Thomas, Native American land-
2005. Seasonal collection of coquina clams scapes of St. Catherines Island, Georgia.
(Donax variabilis Say, 1822) during the Late II. The data. Anthropological Papers of the
Archaic and St. Johns periods in coastal north- American Museum of Natural History 88:
east Florida. In D.E. Bar-Yosef Mayer (edi- 615–665.
tor), Archaeomalacology: molluscs in former Reitz, E.J., and J. Dukes. 2008. Change and
environments of human behavior. Proceed- stability in vertebrate use between the Irene
ings of the Archaeomalacology Symposium Period and the Mission period: nonhuman
of the 9th International Council for Zooar- vertebrate remains from Meeting House Field
chaeology: 18–28. Oxford: Oxbow Press. and Fallen Tree. In D.H. Thomas, Native
Raab, L.M. 1996. Debating prehistory in coastal American landscapes of St. Catherines Island,
southern California: resource intensification Georgia. II. The data. Anthropological Papers
versus political economy. Journal of Cali- of the American Museum of Natural History
fornia and Great Basin Anthropology 18 88: 778–798.
(1): 64–80. Reitz, E.J., and I.R. Quitmyer. 1988. Faunal
Reagan, R.E. 1985. Species profiles: life histories remains from two coastal Georgia Swift
and environmental requirements of coastal Creek sites. Southeastern Archaeology 7
fishes and invertebrates (Gulf of Mexico): red (2): 95–108.
drum. U.S. Fish and Wildlife Service. Biolog- Reitz, E.J., and C.M. Scarry. 1985. Reconstruct-
ical Report 82 (11.36). ing historic subsistence with an example
Redman, C.L., J.M. Grove, and L.H. Kuby. 2004. from sixteenth-century Spanish Florida. So-
Integrating social science into the Long-Term ciety for Historical Archaeology 3 (special
Ecological Research (LTER) network: social publications).
dimensions of ecological change and eco- Reitz, E.J., and E.S. Wing. 1999. Zooarchaeolo-
logical dimensions of social change. Ecosys- gy. Cambridge: Cambridge University Press.
tems 7: 161–171. http://dx.doi.org/10.1007/ Reitz, E.J., and E.S. Wing. 2008. Zooarchaeol-
s10021-003-0215-z ogy. 2nd ed. Cambridge: Cambridge Univer-
Reis, R.R., and J.M. Dean. 1981. Temporal varia- sity Press.
tion in the utilization of an intertidal creek by Reitz, E.J., C.F.T. Andrus, and D.H. Sandweiss.
the bay anchovy (Anchoa mitchilli). Estuaries 2008. Ancient fisheries and marine ecology of
4: 16–23. http://dx.doi.org/10.2307/1351539 coastal Peru. In T.C. Rick and J.M. Erlandson
Reitz, E.J. 1982. Vertebrate fauna from four (editors), Human impacts on ancient marine
coastal Mississippian sites. Journal of Ethno- ecosystems: a global perspective: 125–145.
biology 2: 39–61. Berkeley: University of California Press.
Reitz, E.J. 1988. Evidence for coastal adaptations Reitz. E.J., I.R. Quitmyer, and R.A. Marrinan.
in Georgia and South Carolina. Archaeology 2009. What are we measuring in the zooarchae-
of Eastern North America 16: 137–158. ological record of prehispanic fishing strate-
Reitz, E.J. 1991. Animal use and culture change in gies in the Georgia Bight, USA? Journal of Is-
Spanish Florida. In P.J. Crabtree and K. Ryan land and Coastal Archaeology 4: 2–36. http://
(editors), Animal use and culture change: 62– dx.doi.org/10.1080/15564890802349894
77. Philadelphia: University of Pennsylvania, Reitz, E.J., B. Pavao-Zuckerman, D.C. Weinand,
Museum of Archaeology and Anthropology and G.A. Duncan. 2010. Mission and Pueblo
Research Papers in Science and Archaeology, of Santa Catalina de Guale, St. Catherines
MASCA 8. Island, Georgia: a comparative zooarchaeo-
Reitz, E.J. 1994. Zooarchaeological analysis of a logical analysis. Anthropological Papers of
free African community: Gracia Real de San- the American Museum of Natural History 91:
ta Teresa de Mose. Historical Archaeology 28 1–275.
(1): 23–40. Rhoads, D.C., and R.A. Lutz (editors). 1980.
Reitz, E.J. 2004. “Fishing down the food web”: a Skeletal growth of aquatic organisms. New
case study from St. Augustine, Florida, USA. York: Plenum Press.
American Antiquity 69 (1): 63–83. http:// Rhoads, D.C., and G. Pannella. 1970. The use
of molluscan shell growth patterns in ecol- Sheffield Academic Press.

ogy and paleoecology. Lethaia 3: 143–161. Rozas, L.P., and T.J. Minello. 2001. Marsh
http://dx.doi.org/10.1111/j.1502-3931.1970. terracing as a wetland restoration tool for
tb01854.x creating fishery habitat. Wetlands 21 (3):
Rice, M.A., and J.A. Pechenik. 1992. A review 327–341. http://dx.doi.org/10.1672/0277-
of the factors influencing the growth of the 5212(2001)021[0327:MTAAWR]2.0.CO;2
northern quahog, Mercenaria mercenaria Rozas, L.P., and R.J. Zimmerman. 2000. Small-
(Linnaeus, 1758). Journal of Shellfish Re- scale patterns of nekton use among marsh and
search 11: 279–287. adjacent shallow nonvegetated areas of the
Rice, M.A., C. Hickox, and I. Zehra. 1989. Ef- Galveston Bay estuary, Texas (USA). Marine
fects of intensive fishing effort on the popula- Ecology Progress Series 193: 217–239. http://
tion structure of quahogs, Mercenaria merce- dx.doi.org/10.3354/meps193217
naria (Linnaeus 1758), in Narragansett Bay. Ruhl, D.L. 1993. Customs and traditions in new
Journal of Shellfish Research 8: 345–354. terrain: sixteenth and seventeenth century ar-
Rich, F.J., A. Vega, and F.J. Vento. 2011. Evo- chaeobotanical data from La Florida. In C.
lution of Late Pleistocene-Holocene climates M. Scarry (editor), Foraging and farming in
and environments of St. Catherines Island and the eastern woodlands: 255–283.
the Georgia Bight. In G.A. Bishop, H.B. Roll- Ruhl, D.L. 2003. Imagining sixteenth and seven-
ins, and D.H. Thomas (editors), Geoarchaeol- teenth century Native American and Hispanic
ogy of St. Catherines Island, Georgia. Anthro- transformations of the Georgia Bight land-
pological Papers of the American Museum of scapes. Bulletin of Florida Museum of Natu-
Natural History 94: 67–77. ral History 44 (1): 183–198.
Rick, T.C., and J.M. Erlandson (editors). 2008. Ruhl, D.L. 2008. Bridging the 20th and 21st cen-
Human impacts on ancient marine ecosys- turies: St. Catherines Island archaeobotany
tems: a global perspective. Berkeley: Univer- from missions to shell rings. Paper presented
sity of California Press. at 65th Annual Meeting of the Southeastern
Robins, C.R., G.C. Ray, J. Douglass, R. Freund, Archaeological Conference, Charlotte, NC.
National Audubon Society, and National Russo, M. 1991. A method for the measurement
Wildlife Federation. 1986. A field guide to At- of season and duration of oyster collection:
lantic coast fishes of North America. Boston: two case studies from the prehistoric south-
Houghton Mifflin. east U.S. Coast. Journal of Archaeologi-
Rogers, S.G., T.E. Targett, and S.B. Vansant. cal Science 18 (2): 205–221. http://dx.doi.
1984. Fish-nursery use in Georgia salt- org/10.1016/0305-4403(91)90049-U
marsh estuaries: the influence of spring- Russo, M. 1998. Measuring sedentism with
time freshwater conditions. Transactions fauna: Archaic cultures along the southwest
of the American Fisheries Society 113: Florida coast. In T.R. Rocek and O. Bar-Yosef
595–606. http://dx.doi.org/10.1577/1548- (editors), Seasonality and sedentism: archaeo-
8659(1984)113<595:FUIGSE>2.0.CO;2 logical perspectives from Old and New World
Ropes, J.W. 1987. Preparation of acetate peels sites. Peabody Museum of Archaeology and
of valves from the ocean quahog, Artica is- Ethnology, Harvard University, Peabody Mu-
landica, for age determinations. Washington, seum Bulletin 6: 143–164.
DC: U.S. Department of Commerce, NOAA Russo, M. 2004. Measuring shell rings for social
Technical Report. inequality. In J.L. Gibson and P.J. Carr (edi-
Rowley-Conwy, P. 1983. Sedentary hunters: tors), Signs of power: the rise of cultural com-
the Ertebølle example. In G. Bailey (editor) plexity in the Southeast: 26–70. Tuscaloosa:
Hunter-gatherer economy in prehistory: a University of Alabama Press.
European perspective: 111–126. Cambridge: Russo, M. 2006. Archaic shell rings of the south-
Cambridge University Press. east U.S. theme study. Tallahassee, FL: Na-
Rowley-Conwy, P. 1998. Cemeteries, seasonality tional Park Service, National Register of
and complexity in the Ertebølle of southern Historic Places, Southeastern Archaeologi-
Scandinavia. In M. Zvelebil, R. Dennell, and cal Center.
L. Domanska (editors), Harvesting the sea, Russo, M. 2008. Late Archaic shell rings and so-
farming the forest: 193–202. Sheffield, UK: ciety in the southeast U.S. SAA Archaeologi-
2012 REFERENCES 231
cal Record 8 (5):18–22. ence 277 (5333): 1796–1799. http://dx.doi.

Russo, M., and G. Heide. 2001. Shell rings of the org/10.1126/science.277.5333.1796
southeast U.S. Antiquity 75: 491–492. Saunders, R. 2000a. Stability and change in
Russo, M. and I.R. Quitmyer. 1996. Sedentism in Guale Indian pottery a.d. 1300–1702. Tusca-
coastal populations of south Florida. In E.J. loosa: University of Alabama Press.
Reitz, L.A. Newsom, and S.J. Scudder (edi- Saunders, R. 2000b. The Guale Indians of the
tors), Case studies in environmental archaeol- Lower Atlantic Coast: change and continu-
ogy: 215–231. New York: Plenum Press. ity. In B.G. McEwan (editor), Indians of the
Russo, M. and I.R. Quitmyer. 2008. Develop- greater Southeast: historical archaeology and
ing models of settlement for the Florida Gulf ethnohistory: 26–56. Gainesville: University
coast. In E.J. Reitz, C.M. Scarry, and S.J. Scud- of Florida Press.
der (editors), Case studies in environmental Scarry, C.M. 2003. Patterns of wild plant utiliza-
archaeology, 2nd ed.: 235–254. New York, tion in the prehistoric Eastern Woodlands. In
Plenum Press. http://dx.doi.org/10.1007/978- P. Minnis (editor), People and plants in ancient
0-387-71303-8_12 eastern North America: 50–104. Washington,
Russo, M.A., and R. Saunders. 2008. Quahog DC: Smithsonian Institution Press.
(Mercenaria mercenaria) seasonality at Scarry, C.M. 2008. Crop husbandry practices in
Meeting House Field. In D.H. Thomas, Na- North America’s Eastern Woodlands. In E.
tive American landscapes of St. Catherines Reitz, C.M. Scarry, and S. Scudder (editors),
Island, Georgia. II. The data. Anthropologi- Case studies in environmental archaeology,
cal Papers of the American Museum of Natu- 2nd ed.: 391–404. New York: Springer. http://
ral History 88: 715–726. dx.doi.org/10.1007/978-0-387-71303-8_20
Sanger, M.C., and D.H. Thomas. 2010. The two Scarry, C.M., and R.A. Yarnell. 2006. Domes-
rings of St. Catherines Island: some prelimi- tication of plants in the east. In D. Ubelaker
nary results from the St. Catherines and Mc- (editor), Environment, origins, and popula-
Queen shell rings. In D.H. Thomas and M.C. tion, volume 3 of the Handbook of North
Sanger (editors and contributors), Trend, American Indians: 428–436. Washington DC:
tradition, and turmoil: what happened to the Smithsonian Institution Press.
southeastern Archaic? Anthropological Pa- Scarry, J.F., and C.M. Scarry. 1997. Subsistence
pers of the American Museum of Natural His- remains from prehistoric North Carolina
tory 93: 44–69. sites. Raleigh: North Carolina Office of State
Sassaman, K.E. 2004. Complex hunter-gatherers Archaeology. Internet resource (http://www.
in evolution and history: a North American archaeology.ncdcr.gov/subsist/subsis.htm),
perspective. Journal of Anthropological Re- accessed on December 14, 2011.
search 12: 227–280. Schaefer, K. 2011. Prehistoric subsistence on the
Sassaman, K.E. 2006. People of the shoals: Stall- coast of North Carolina: an archaeobotanical
ings culture of the Savannah River Valley. study. Ph.D. dissertation, Department of An-
Gainesville: University Press of Florida. thropology, University of North Carolina at
Saul, B.M. 2002. Ichthyofaunal survey of St. Chapel Hill.
Catherines Island, Report 2002. American Schelske, C.L., and E.P. Odum. 1961. Mecha-
Museum of Natural History, Division of An- nisms maintaining high productivity in Geor-
thropology, ms. on file. gia estuaries. Proceedings of the Gulf and Ca-
Saul, B.M. 2003. Ichthyofaunal survey of St. ribbean Fisheries Institute 1961: 75–80.
Catherines Island, Report 2003. American Schöne, B.R. 2008. The curse of physiology—
Museum of Natural History, Division of An- challenges and opportunities in the interpreta-
thropology, ms. on file. tion of geochemical data from mollusk shells.
Saul, B.M. 2004. Ichthyofaunal survey of St. Geo-Marine Letters 28 (5–6): 269–285. http://
Catherines Island, Report 2004. American dx.doi.org/10.1007/s00367-008-0114-6
Museum of Natural History, Division of An- Schöne, B.R., E. Dunca, H. Mutvei, and U.
thropology, ms. on file. Norlund. 2004. A 217-year record of sum-
Saunders, J.W., R.D. Mandel, R.T. Saucier, et mer air temperature reconstructed from
al. 1997. A mound complex in Louisiana freshwater pearl mussels (M. margarifitera,
5400–5000 years before the present. Sci- Sweden). Quaternary Science Reviews 23:
1803–1816. http://dx.doi.org/10.1016/j.quas- Stahle, D.W., and M.K. Cleaveland. 1992. Re-

cirev.2004.02.017 construction and analysis of spring rainfall
Schopmeyer, C.S. 1974. Seeds of woody plants. over the southeastern U.S. for the past 1000
Agricultural Handbook 450. Washington, DC: years. Bulletin of the American Meteoro-
United States Department of Agriculture. logical Society 73: 1947–1961. http://dx.doi.
Semon, A., G. Mahar, M.C. Sanger, M. Napoli- org/10.1175/1520-0477(1992)073<1947:RA
tano, S. Bergh, and E. Blair. 2008. Trip Report AOSR>2.0.CO;2
Spring 2008. Ms. on file, Nels Nelson North Stahle, D.W., M.K. Cleaveland, D.B. Blanton,
American Archaeology Lab, American Mu- M.D. Therrell, and D.A. Gay. 1998. The lost
seum of Natural History, New York. colony and Jamestown droughts. Science
Shackleton, N.J. 1973. Oxygen isotope analy- 280: 564–567. http://dx.doi.org/10.1126/sci-
sis as a means of determining season of oc- ence.280.5363.564
cupation of prehistoric midden sites. Ar- Stahle, D.W., E.R. Cook, M.D. Therrell, D.M.
chaeometry 15 (1): 133–141. http://dx.doi. Meko, H.D. Grissino-Mayer, E. Watson, and
org/10.1111/j.1475-4754.1973.tb00082.x B.H. Luckman. 2000. Tree-ring data docu-
Shennan, S. 1997. Quantifying archaeology. Ed- ment 16th century megadrought over North
inburgh: University Press. America. EOS, Transactions of the American
Shipek, F.C. 1989. An example of intensive plant Geophysical Union 81 (12): 212, 125.
husbandry: the Kumeyaay of southern Cali- Stark, M.T. 1998. Technical choices and social
fornia. In D. Harris and G. Hillman (editors), boundaries in material culture patterning: an
Foraging and farming: the evolution of plant introduction. In M.T. Stark (editor), The ar-
exploitation: 159–170. Boston: Unwin Hy- chaeology of social boundaries: 1–11. Wash-
man. ington DC: Smithsonian Institution Press.
Shumway, S.E. 1996. Natural environmental fac- Steele, J. 2010. Radiocarbon dates as data: quan-
tors. In V.S. Kennedy, R. Newell, I.E. and A.F. titative strategies for estimating coloniza-
Eble (editors), The eastern oyster: Crassostrea tion front speeds and event densities. Journal
virginica: 467–503. College Park: Maryland of Archaeological Science 37: 2017–2030.
Sea Grant College. http://dx.doi.org/10.1016/j.jas.2010.03.007
Silvano, R.A.M., P.F.L. MacCord, R.V. Lima, Steier, P., and W. Rom. 2000. The use of Bayes-
and A. Begossi. 2006. When does this fish ian statistics for 14C dates of chronologically
spawn? Fishermen’s local knowledge of mi- ordered samples: a critical analysis. Radiocar-
gration and reproduction of Brazilian coastal bon 42: 183–198.
fishes. Environmental Biology of Fishes 76 Stein, J.K. 1992. Interpreting stratification of a
(2–4): 371–386. http://dx.doi.org/10.1007/ shell midden. In J.K. Stein (editor), Decipher-
s10641-006-9043-2 ing a shell midden: 71–93. San Diego: Aca-
Smith, B.D. 1992. The role of Chenopodium as demic Press.
a domesticate in premaize garden systems of Stein, J.K., J.N. Deo, and L.S. Phillips. 2003.
the eastern United States. In B.D. Smith (edi- Big sites—short times: accumulation rates in
tor), Rivers of change: 103–131. Washington, archaeological sites. Journal of Archaeologi-
DC: Smithsonian Institution Press. cal Science 30 (3): 297–316. http://dx.doi.
Smith, B.D. 2001. Low-level food production. org/10.1006/jasc.2002.0838
Journal of Archaeological Research 9:1–43. Stenzel, H.B. 1964. Oysters: composition
http://dx.doi.org/10.1023/A:1009436110049 of the larval shell. Science 145 (3628):
Smith, H.A. 1983. Determination of seasonality 155–156. http://dx.doi.org/10.1126/sci-
in archaeological sites through examination ence.145.3628.155
of fish otoliths: a case study. Journal of Field Stuiver, M., G.W. Pearson, and T.F. Braziunas.
Archaeology 10: 498–500. 1986. Radiocarbon calibration of marine
Smith, J.W., and C.A. Wenner. 1985. Biology of samples back to 9000 cal yr bp. Radiocarbon
the southern kingfish in the South Atlantic 28 (2B): 980–1021.
Bight. Transactions of the American Fish- Sturtevant, W.C. 1962. Spanish-Indian relations
eries Society 114: 356–366. http://dx.doi. in southeastern North America. Ethnohistory 9
org/10.1577/1548-8659(1985)114<356:BOT (1): 41–94. http://dx.doi.org/10.2307/480785
SKI>2.0.CO;2 Sturtevant, W.C. 1964. Document 42: A Jesuit
2012 REFERENCES 233
missionary in South Carolina, 1569–70. In cology working group, International Council

W.E. Washburn (editor), The Indian and the for Archaeozoology: Papers of the First In-
white man: 167–175. New York: Anchor augural Meeting, Gainesville, Florida, 2005.
Books. International Journal of Archaeofauna 17.
Surge, D., and K.J. Walker. 2005. Oxygen iso- Talalay, L.E., D.R. Keller, and P.J. Munson. 1984.
tope composition of modern and archaeo- Hickory nuts, walnuts, butternuts, and ha-
logical otoliths from the estuarine hardhead zelnuts: observations and experiments rele-
catfish (Ariopsis felis) and their potential to vant to their aboriginal exploitation in east-
record low-latitude climate change. Palaeo- ern North America. In P.J. Munson (editor),
geography, Palaeoclimatology, Palaeoecolo- Experiments and observations on aboriginal
gy, 228: 179–191. http://dx.doi.org/10.1016/j. wild plant food utilization in eastern North
palaeo.2005.03.051 America: 338–359. Prehistoric Research
Surge, D., and K.J. Walker. 2006. Geochemical Series VI(2). Indianapolis: Indiana Histori-
variation in microstructural shell layers of the cal Society.
southern quahog (Mercenaria campechien- Thomas, D.H. 1986. Refiguring anthropology: first
sis): implications for reconstructing season- principles of probability and statistics. Prospect
ality. Palaeogeography, Palaeoclimatology, Heights, IL: Waveland Press, Inc. 532 p.
Palaeoecology 237: 182–190. http://dx.doi. Thomas, D.H. 1987. The archaeology of mis-
org/10.1016/j.palaeo.2005.11.016 sion Santa Catalina de Guale: 1. Search and
Surge, D., K.C. Lohmann, and D.L. Dettman. discovery. Anthropological Papers of the
2001. Controls on isotopic chemistry of the American Museum of Natural History 63
American oyster, Crassostrea virginica: im- (2): 47–161.
plications for growth patterns. Palaeogeogra- Thomas, D.H. 2008. Native American landscapes
phy, Palaeoclimatology, Palaeoecology 172 of St. Catherines Island, Georgia. Anthropo-
(3–4): 283–296. http://dx.doi.org/10.1016/ logical Papers of the American Museum of
S0031-0182(01)00303-0 Natural History 88: 1–1136.
Surge, D., K.C. Lohmann, and G.A. Goodfriend. Thomas, D.H., and C.S. Larsen. 1979. The an-
2003. Reconstructing estuarine conditions: thropology of St. Catherines Island: 2. The
oyster shells as recorders of environmental Refuge-Deptford mortuary complex. Anthro-
change, southwest Florida. Estuarine, Coastal pological Papers of the American Museum of
and Shelf Science 57: 737–756. http://dx.doi. Natural History 56 (1): 1–179.
org/10.1016/S0272-7714(02)00370-0 Thomas, D.H., and M.C. Sanger (editors and con-
Surge, D., G. Kelly, W.S. Arnold, K.J. Walker, and tributors). 2010. Trend, tradition, and turmoil:
A. Goewert. 2008. Isotope sclerochronology what happened to the southeastern Archaic?
of Mercenaria mercenaria, M. campechien- Anthropological Papers of the American Mu-
sis, and their natural hybrid form (Bivalvia): seum of Natural History 93: 1–341.
does genotype matter? Palaios 23: 559–565. Thomas, D.H., G.D. Jones, R.S. Durham, and
http://dx.doi.org/10.2110/palo.2007.p07-056r C.S. Larsen. 1978. The anthropology of St.
Sutter, F.C., R.S. Waller, and T.D. McIlwain. Catherines Island: 1. Natural and cultural his-
1986. Species profiles: life histories and envi- tory. Anthropological Papers of the American
ronmental requirements of coastal fishes and Museum of Natural History 55 (2): 155–248.
invertebrates (Gulf of Mexico): black drum. Thomas, D.H., M.C. Sanger, and R.H. Hayes.
U.S. Fish and Wildlife Service. Biological n.d. Revising the 14C reservoir correction for St.
Report 82 (11.51). Catherines Island, Georgia. Unpublished ms.
Swanton, J.R. 1922. Early history of the Creek Thomas, F.R. 2002. An evaluation of central-
Indians and their neighbors. Bureau of Ameri- place foraging among mollusk gatherers in
can Ethnology Bulletin 73. Western Kiribati, Micronesia: linking behav-
Swanton, J.R. 1946. The Indians of the southeast- ioral ecology with ethnoarchaeology. World
ern United States. Bureau of American Eth- Archaeology 34: 182–208. http://dx.doi.
nology Bulletin 137. Washington, DC: Gov- org/10.1080/00438240220134313
ernment Printing Office. Thomas, K.D. 1987. Prehistoric coastal ecolo-
Szabó, K., and I.R. Quitmyer (editors). 2008. gies: a view from outside Franchthi Cave,
World archaeomalacology: the archaeomala- Greece. Geoarchaeology 2 (3): 231–240.
http://dx.doi.org/10.1002/gea.3340020305 Gulick. 1976. Barataria Basin: salinity chang-

Thompson, V.D. 2006. Questioning complexity: es and oyster distribution. Baton Rouge: Lou-
the prehistoric hunter-gatherers of Sapelo Is- isiana State University.
land, Georgia. Ph.D. dissertation, Department Versteegh, E.A.A. 2009. Silent witnesses: fresh-
of Anthropology, University of Kentucky, water bivalves as archives of environmental
Lexington. variability in the Rhine-Meuse delta. Ph.D.
Thompson, V.D., and C.F.T. Andrus. 2011. Eval- dissertation. University of Amsterdam, Neth-
uating mobility, monumentality, and feasting erlands.
at the Sapelo Island Shell Ring Complex. Versteegh, E.A.A., S.R. Troelstra, H.B. Vonhof,
American Antiquity 76 (2): 315–344. and D. Kroon. 2009. Oxygen isotope compo-
Thorrold, S.R., S.E. Campana, C.M. Jones, and sition of bivalve seasonal growth increments
P.K. Swart. 1997. Factors determining δ13C and ambient water in the rivers Rhine and
and δ18O fractionation in aragonitic otoliths Meuse. Palaios 24: 497–504. http://dx.doi.
of marine fish. Geochimica et Cosmochim- org/10.2110/palo.2008.p08-071r
ica Acta 61: 2909–2919. http://dx.doi. Versteegh, E.A.A., H.B. Vonhof, S.R. Troelstra,
org/10.1016/S0016-7037(97)00141-5 and D. Kroon. 2010. A molluscan perspective
Tozzer, A. 1926. Chronological aspects of Amer- on hydrological cycle dynamics in northwest-
ican archaeology. Proceedings of the Massa- ern Europe. Netherlands Journal of Geosci-
chusetts Historical Society 59: 283–292. ences 89: 51–60.
Ugan, A. 2005. Does size matter? Body size, Voorhies, B. 2004. Coastal collectors in the Holo-
mass collecting, and their implications for cene: the Chantuto people of southwest Mexi-
understanding prehistoric foraging behavior. co. Gainesville: University Press of Florida.
American Antiquity 70: 75–89. http://dx.doi. Walker, K.J., and D. Surge. 2006. Developing
org/10.2307/40035269 oxygen isotope proxies from archaeological
University of Florida Institute of Food and Ag- sources for the study of Late Holocene human–
ricultural Services. 2006. Proceedings from climate interactions in coastal southwest Flori-
the 1st International Sclerochronology Con- da. Quaternary International 150: 3–11. http://
ference, July 17–21, 2007, St. Petersburg, dx.doi.org/10.1016/j.quaint.2006.01.022
Florida. Internet resource (http://conference. Walker, R.L. 1989. Exploited and unexploited
ifas.ufl.edu/sclerochronology/#overview), ac- hard clam, Mercenaria mercenaria (L.), pop-
cessed on November 17, 2010. ulations in coastal Georgia. Contributions in
Vander Putten, E., F. Dehairs, E. Keppens, and Marine Science 31: 61–75.
W. Baeyens. 2000. High resolution distribu- Walker, R.L., and C. Cotton. 2001. Oyster bed
tion of trace elements in the calcite shell layer distribution as a long term environmental in-
of modern Mytilus edulis: environmental dicator for the Duplin River, Sapelo Island,
and biological controls. Geochimica et Cos- National Estuarine Research Reserve. Univer-
mochimica Acta 64: 997–1011. http://dx.doi. sity of Georgia, School of Marine Programs,
org/10.1016/S0016-7037(99)00380-4 Marine Technical Report 01–1.
Van Neer, W., L. Löugas, and A.D. Rijnsdorp. Walker, R.L., and K.R. Tenore. 1984. The dis-
1999. Reconstructing age distribution, season tribution and production of the hard clam,
of capture, and growth rate of fish from ar- Mercenaria mercenaria, in Wassaw Sound,
chaeological sites based on otoliths and ver- Georgia. Estuaries 7(1): 19–27. http://dx.doi.
tebrae. International Journal of Osteoarchae- org/10.2307/1351953
ology 9: 116–130. http://dx.doi.org/10.1002/ Walker R.L., A.J. Power, M. Sweeney-Reeves, E.
(SICI)1099-1212(199903/04)9:2<116::AID- Covington, M. Mitchell, and T. Recicar. 2008.
OA465>3.0.CO;2-H Growth, migration, population structure, and
Van Neer, W., A. Ervynck, L.J. Bolle, and R.S. sex ratio of four species of whelks (family
Millner. 2004. Seasonality only works in cer- Melongenidae) within Wassaw Sound, Geor-
tain parts of the year: the reconstruction of gia. Occasional Papers of the University of
fishing seasons through otolith analysis. In- Georgia Marine Extension Service 1.
ternational Journal of Osteoarchaeology 14: Wallace, R.L. 1975. An archaeological, ethno-
457–474. http://dx.doi.org/10.1002/oa.727 historic, and biochemical investigation of the
Van Sickle, V.R., B.B. Barrett, T.B. Ford, and L.J. Guale aborigines of the Georgia coastal strand.
2012 REFERENCES 235
Ph.D. dissertation, Department of Anthropol- org/10.1007/BF00389156

ogy, University of Florida, Gainesville. Weide, M.L. 1969. Seasonality of Pismo clam
Wanamaker, A.D., Jr., H.W. Borns, Jr., D.S. In- collecting at Ora-82. Archaeology Survey An-
trone, S. Feindel, and B.J. Barber. 2006. An nual Report 11: 127–141. Los Angeles: Uni-
aquaculture-based method for calibrated bi- versity of California.
valve isotope paleothermometry. Geochemis- Wells, H.W. 1959. Notes on Odostomia impressa
try Geophysics Geosystems 7: Q09011. http:// (Say). Nautilus 72: 140–144.
dx.doi.org/10.1029/2005GC001189 Wells, J.W. 1963. Coral growth and geochro-
Wanamaker, A.D., Jr., K.J. Kreutz, H.W. Borns, nometry. Nature 197: 948–950. http://dx.doi.
Jr., D.S. Introne, S. Feindel, S. Funder, P.D. org/10.1038/197948a0
Rawson, and B.J. Barber. 2007. Experimental White, C.D., M.E.D. Pohl, H.P. Schwarcz, and
determination of salinity, temperature, growth, F.J. Longstaffe. 2001. Isotopic evidence for
and metabolic effects on shell isotope chemis- Maya patterns of deer and dog use at Preclas-
try of Mytilus edulis collected from Maine and sic Colha. Journal of Archaeological Science
Greenland. Paleoceanography 22: PA2217. 28 (1): 89–107. http://dx.doi.org/10.1006/
http://dx.doi.org/10.1029/2006PA001352 jasc.1999.0560
Ward, G.K., and S.R. Wilson. 1978. Procedures White, M.E., and E.A. Wilson. 1996. Predators,
for comparing and combining radiocar- pests, and competitors. In V.S. Kennedy, R.I.E.
bon age determination: a critique. Archae- Newell, and A.F. Eble (editors), The eastern
ometry 20: 19–31. http://dx.doi.org/10.1111/ oyster Crassostrea virginica: 559–579. Col-
j.1475-4754.1978.tb00208.x lege Park: Maryland Sea Grant College.
Ward, J.W. 1957. The reproduction and early de- White, M.E, E.N. Powell, and C.L. Kitting. 1984.
velopment of the sea catfish, Galeichthys fe- The ectoparasitic gastropod Boonea (= Odos-
lis, in the Biloxi (Mississippi) Bay. Copeia 4: tomia) impressa: population ecology and the
295–298. http://dx.doi.org/10.2307/1439158 influence of parasitism on oyster growth rates.
Waselkov, G.A. 1987. Shellfish gathering and Marine Ecology 5 (3): 283–299. http://dx.doi.
shell midden archaeology. In M.G. Schiffer org/10.1111/j.1439-0485.1984.tb00126.x
(editor), Advances in archaeological meth- White, M.E., C.L. Kitting, and E.N. Powell.
od and theory 4: 93–210. New York: Aca- 1985. Aspects of reproduction, larval de-
demic Press. velopment, and morphometrics in the pyra-
Waselkov, G.A., L.C. Steponaitis, E. Mueller, midellid, Boonea impressa (= Odostomia
J.M. Herbert, and G.R. Abbe. 1998. Seasonal- impressa), (Gastropoda: Opisthobranchia).
ity determinations from archaeological oyster Veliger 28 (1): 37–51.
shells: an experimental evaluation of incre- Whittle, A., and A. Bayliss. 2007. The times of
mental shell growth. Report to the National their lives: from chronological precision to
Science Foundation, award #9421140. Mo- kinds of history and change. Cambridge Ar-
bile: University of South Alabama. chaeological Journal 17 (1): 21–28. http://
Webb, S.C., R.E.M. Hedges, and R. Robinson. dx.doi.org/10.1017/S0959774307000030
1998. The seaweed fly Thoracochaeta zos- Winterhalder, B. 2001. The behavioural ecol-
terae (Hal.) (Diptera: Sphaerocidae) in inland ogy of hunter-gatherers. In C. Panter-Brick,
archaeological contexts: δ13C and δ15N solves R.H. Layton, and P. Rowley-Conwy (edi-
the puzzle. Journal of Archaeological Science tors), Hunter-gatherers: an interdisciplinary
25: 1253–1257. http://dx.doi.org/10.1006/ perspective: 12–38. Cambridge: Cambridge
jasc.1998.0298 University Press.
Wefer, G., and W.H. Berger. 1991. Isotope pale- Winterhalder, B., and D.J. Kennett. 2006. Behav-
ontology: growth and composition of extant ioral ecology and the transition from hunting
calcareous species. Marine Geology 100 (1– and gathering to agriculture. In D.J. Kennett
4): 207–248. http://dx.doi.org/10.1016/0025- and B. Winterhalder (editors), Behavioral
3227(91)90234-U ecology and the transition to agriculture, 1–21.
Wefer, G., and J.S. Killingley. 1980. Growth Berkeley: University of California Press.
histories of strombid snails from Bermuda Winterhalder, B., D.J. Kennett, M. Grote, and
recorded in their O-18 and C-13 profiles. J. Bartruff. 2010. Ideal free settlement of
Marine Biology 60: 129–135. http://dx.doi. California’s Northern Channel Islands. Jour-
nal of Anthropological Archaeology 29 tor), Handbook of North American Indians

(4): 469–490. http://dx.doi.org/10.1016/j. [vol. 14]. Southeast: 238–244. Washington,
jaa.2010.07.001 DC: Smithsonian Institution Press.
Withers, K., and M. Hubner. 2009. Bay scallops, Worth, J.E. 2007. The struggle for the Georgia
Argopecten irradians, in the northwestern coast. Tuscaloosa: University of Alabama
Gulf of Mexico (Alabama, Mississippi, Loui- Press.
siana, and Texas). Marine Fisheries Review Yen, D.E. 1989. The domestication of environ-
71 (3): 1–5. ment. In D. Harris and G. Hillman (editors),
Wootton, R.J. 1998. Ecology of teleost fishes. Foraging and farming: the evolution of plant
Dordrecht: Kluwer Academic. exploitation: 55–75. Boston: Unwin Hyman.
Worth, J.E. 1998. Timucuan chiefdoms of Span- Yesner, D.R. 1980. Maritime hunter-gather-
ish Florida [vol. 1]. Assimilation. Gainesville: ers: ecology and prehistory. Current An-
University of Florida Press. thropology 21 (6): 727–750. http://dx.doi.
Worth, J.E. 1999. Coastal chiefdoms and the org/10.1086/202568
question of agriculture: an ethnohistorical Zubillaga, F. 1946. Monumenta antiquae Floridae
overview. Fifty-sixth Annual Southeastern (1566–1572). Monumenta historica societatis
Archaeological Conference, Pensacola, FL. Iesu 69. Monumenta Missionum Societatis
Worth, J.E. 2004. Guale. In R.D. Fogelson (edi- Iesu 3. Roma.
Cover Design: Jennifer Steffey
Book Design and Layout: Kevin B. Havener
On the cover: Backlit thin sect ion of a quahog (Mercenaria campechiensis) show-
ing its alternat ing opaque and translucent shell grow th increments. Quahogs, or
hard clams (M. mercenaria, M. campechiensis, and hybrids), are the most st udied
of all midden molluscs in the southeaster n United States w ith respect to ox ygen
isotopes and shell grow th increments. T hese increments may be seen in cross sec-
t ions of the shells; in the backlit radial thin sect ion of the shell on the cover, trans-
mit ted light shows opaque and translucent increments that ref lect the periodic
dif ferences in the shell microstr uct ures associated w ith physiological responses to
env ironmental changes in the annual c ycle. Quahog grow th increments form regu -
larly according to the season and can be used to assess the t ime during which the
bivalves were capt ured at southeaster n U. S. archaeological sites, and sur veys of
quahog season - of- capt ure data and their localit ies can elucidate the movement of
human populat ions throughout an annual c ycle. Photograph by Ir v y R . Quit myer.
The St. Catherines Island research ... is a stellar example of what long-term, detailed, and interdisciplinary
work can contribute to our understanding of the history and human ecology of coastal regions around
the world. The Caldwell V conference proceedings extend earlier findings ... and the result is an impres-
sive, informative, and well-focused volume about the archaeology of St. Catherines Island and the Georgia
Bight—a book that will be of interest to a broad audience of coastal archaeologists interested in recent
advances in seasonality and related issues in coastal settings.
Jon M. Erlandson, Professor of Anthropology, University of Oregon
The volume’s strengths arise from its focus on a single set of problems, competing hypotheses about mo-
bility patterns on St. Catherines Island, especially issues of seasonality and chronology. The volume will be
very useful to any archaeologist grappling with mobility and seasonality since it presents a multistranded
approach to data and the creation of evidence.
Kenneth M. Ames, Professor of Anthropology, Portland State University
A rchaeological excavations at coastal sites

typically recover multiple biological proxies
for seasonal behavior of resources and of people.
originally presented at the Fifth Caldwell Con-
ference, cosponsored by the American Museum
of Natural History and the St. Catherines Island
Questions of seasonality are embedded in much Foundation and held on St. Catherines Island
of archaeological research, with answers linked to (Georgia), May 14–16, 2010.
many aspects of cultures and environments, such
as: why is seasonality important to the study of ELIZABETH J. REITZ is a research associate of the
human behavior? What does this knowledge tell American Museum of Natural History and Profes-
us about life in dynamic estuarine systems? sor of Anthropology at the Georgia Museum of
Contributors to Seasonality and Human Mobil- Natural History, University of Georgia.
ity along the Georgia Bight believe that one must
distinguish between patterns of seasonal site oc- IRVY R. QUITMYER is a specialist in the zooarchaeol-
cupation (reflecting mobility strategies), seasonal ogy of marine people and serves as Environmental
patterns of resource procurement (reflecting Archaeology Collection Manager at the Florida
foraging and/or farming strategies), site function, Museum of Natural History, Gainesville, Florida.
and relationships with people on neighboring
islands and the mainland. These are very differ- DAVID HURST THOMAS is a curator in the Division of
ent—if closely interrelated—components of hu- Anthropology at the American Museum of Natural
man behavior. History. He has conducted archaeological research
The chapters in the volume presented here on St. Catherines Island since 1974.
address specific methodological issues (such as
sample size, intersite and intrasite variability, dat-
ing, and defining regional settlement patterns)
that remain to be resolved as investigators pro-
ceed with substantive applications to the archae-
ological record of the Georgia Bight. They were

Calupoh Estandar

Caricato da

Informazioni sul documento

Titolo originale

Copyright

Formati disponibili

Condividi questo documento

Condividi o incorpora il documento

Opzioni di condivisione

Hai trovato utile questo documento?

Questo contenuto è inappropriato?

Copyright:

Formati disponibili

Calupoh Estandar

Caricato da

Copyright:

Formati disponibili

S e asonalit y and

Editors and Contributors

American Museum of Natural History Anthropological Papers, Number 97

Submission procedures can be found at http://research.amnh.org/scipubs

All issues of Anthropological Papers of the American Museum of Natural History

Order printed copies from http://www.amnhshop.com or via standard mail from

∞ This paper meets the requirements of ANSI/NISO Z39.48-1992 (permanence of paper).

SEASONALITY AND HUMAN MOBILITY

ELIZABETH J. REITZ, IRVY R. QUITMYER,

DAVID HURST THOMAS

Proceedings of the Fifth Caldwell Conference

Copyright © American Museum of Natural History 2012

Chapter 2. A Bayesian chronological framework for determining site seasonality

Chapter 3. Interpreting seasonality from modern and archaeological fishes

δ18O of modern otoliths .................................................................................................................. 89

Chapter 5. Late prehistoric settlement patterns: zooarchaeological evidence from Back

Chapter 6. Molluscs as oxygen-isotope season-of-capture proxies in southeastern

Chapter 8. Validation of annual shell increments and shifting population dynamics

from the Litchfield Beach Region, South Carolina. Douglas S. Jones,

Chapter 9. Reevaluating the use of impressed odostome (Boonea impressa) as a season

Chapter 13. Discussion. Elizabeth S. Wing ...................................................................................... 207

References .......................................................................................................................................... 211

[Preface] Participants in the Fifth Caldwell Conference .................................................................... 12

6.1. Examples of taxa described in this chapter ................................................................................ 124

PARTICIPANTS IN THE FIFTH CALDWELL CONFERENCE

C. Fred T. Andrus (Associate Professor, Department of Geological Sciences, University

ethnohistoric record: most St. Catherines islanders maintained four-

last several millennia, enabling archaeologists to tural) event of interest.

based on historical documentation, not 14C dating.

February/March 44% 56%

March/April 10% 90%

April/May 30% 80%

May/June 4% 18% 78%

Altamaha 50% 12% 8% 26%

Irene 64% 16% 6% 14%

Wilmington 64% 19% 1% 16%

St. Simons 47% 50% 0% 3%

TOTAL 64% 18% 5% 13%

Mound near South-

Village (Year Round) SHELL MIDDENS

Sequence McQueen Shell Ring {A = 99.9% (A'c = 60.0%)}

N243 E233 4.5-4.4 m 99.4%

N243 E233 4.4-4.3 m 100.5%

3000 B.C. 2500 B.C. 2000 B.C. 1500 B.C.

Sequence McQueen Shell Ring {A = 83.8%(A'c = 60.0%)}

N243 E233 4.5-4.4 m Shell 107.0%

N243 E233 4.4-4.3 m Shell 94.2%

N243 E233 4.3-4.2 m Shell 73.0%

3000 B.C. 2500 B.C. 2000 B.C. 1500 B.C.

Ap Beta 251761 & 251762 SD 1

E232 E233 E234 E235

A Note on Precision 14C of hard clams has already provided an important

2200 B.C. 2000 B.C. 1800 B.C. 1600 B.C.

Sequence {A = 84.6%(A′c = 60.0%)}

Sequence {A = 94.6%(A′c = 60.0%)}

Sequence {A = 17.8%(A'c = 60.0%)}

2500 B.C. 2000 B.C. 1500 B.C.

SOUTH CAROLINA Wilmington

Ribault and Jacksonville Electric Company

FLORIDA Fountain of Youth

Back Creek Village, Mississippian (Irene) St. Catherines a.d.

Kings Bay Locality, Mississippian Kings Bay a.d.

est juvenile fishes. Zooarchaeological collections,