THE TO'AGA SITE
THREE MILLENNIA OF
/
POLYNESIAN OCCUPATION
2.
THE MANU'A ISLANDS,
I~~~~~~~~~N
AMERICAN SAMOA
_,I~~~~~~
- - -~ ~ ~ ~ ~ ~ ~ ~ ~ F
View of the completed 1987 main excavation at Totaga.
Cover photo: View of Ofu Island coastline fronting the To'aga site. The volcanic plug at Fa'ala'aga is
visible in the distance. (Photo by P.V. Kirch)
ISBN
1-882744-01-2
THE TO'AGA SITE
THREE MILLENNIA OF POLYNESIAN OCCUPATION
IN THE MANU'A ISLANDS,
AMERICAN SAMOA
R V. KIRCH AND T. L. HUNT
EDITORS
Number 51
Contributions of the University of Califomia
Archaeological Research Facility
Berkeley
Primary funding forthis research has been provided by the Historic Preservation Office of the
Department of Parks and Recreation, Government of Amenican Samoa, through grants from
the U. S. National Park Service, Archaeological Assistance Division, Department of the
Interior. However, the contents and opinions expressed do not necessarily reflect the views
or policies of the Department of the Interior.
Library of Congress Catalog Card Number 93-72699
ISBN 1-882744-01-2
© 1993 by the Regents of the University of California
Archaeological Research Facility
University of California at Berkeley
Printed in the United States of America
All rights reserved. No part of this publication may be reproduced or transmitted in any fonn
orby any means, electronic or mechanical, including photocopy, recording, or any information
storage or retrieval system, without permission in writing from the publisher.
Dedicated to the memory of
JOHN KNEUBUHL
(1920-1992)
Citizen of Two Cultures
Tagaloa e, tawnuli ai,
Tagaloafiamalolo;
E mapu i le lagi Tuli mai vasa;
Ta fili'a i peau a lao.
Fea le nu'u na lua'i tupu?
Manua-tele na mua'i tupu.
Tagaloa, who sits at the helm,
Tagaloa desires to rest;
Tulifrom the ocean must rest in the heavens;
These waves below affright my breast.
Where is the land whichfirst upsrang?
Great Manu'afirst uprose.
(Mead 1930:152)
Contents
Preface
ix
Chapter 1. INTRODUCHON AND RESEARCH DESIGN.
T. L. Hunt and P. V. Kirch
Background to Samoan Archaeology
Design of the Research
References Cited
1
2
5
6
Chapter 2. OFU ISLAND AND THE TO'AGA SITE:
DYNAMICS OF THE NATURAL AND CULTURAL
ENVIRONMENT. P. V. Kirch
9
Geology and Geomorphology of Ofu
Resources of the Island
The Cultural and Social Landscape
To'aga: Land Use and Vegetation Patterns
References Cited
11
12
17
18
21
Chapter 3. SURFACE ARCHAEOLOGICAL FEATURES OF THE
TO'AGA AREA. T. L. Hunt
References Cited
23
30
Chapter 4. THE TOAGA SITE: MODELLING THE MORPHODYNAMICS
OF THE LAND-SEA INTERFACE. P. V. Kirch
31
Geomorphology of the To'aga Coastal Terrace
Morphodynamic Processes: Sea Levels and Subsidence
Sediment Budgets: Sources and Modes of Deposition
Modelling the Morphodynamics of the To'aga Coastal Terrace
in the Mid- to Late-Holocene
To'aga: Expectations for Archaeological Site Formation Processes
References Cited
32
33
38
38
40
41
Chapter 5. EXCAVATIONS AT THE TOAGA SITE (AS-13-1).
P. V. Kirch and T. L. Hunt
43
Introduction
1986 Test Excavation
The 1987 Excavations
The 1989 Transect Excavations
Summary and Conclusions
References Cited
43
46
46
57
82
83
Chapter 6. RADIOCARBON CHRONOLOGY OF THE TO'AGA SITEa.
P. V. Kirch
Radiocarbon Dating Procedures: Corrections and Calibrations
The Radiocarbon Corpus from To'aga
v
85
86
87
Discussion
References Cited
89
91
Chapter 7. A GEOARCHAEOLOGICAL ANALYSIS OF SEDIMENT
SAMPLES FROM THE TO'AGA SITE EXCAVATIONS.
P. V. Kirch, E. Manning, and J. Tyler
93
Introduction
93
94
98
111
114
Methods
Results
Summary and Conclusions
References Cited
Chapter 8. NON-MARINE MOLLUSCS FROM THE TO'AGA
SITE SEDIMENTS AND THEIR IMPLICATIONS FOR
ENVIRONMENTAL CHANGE. P. V. Kirch
115
Introduction
115
116
116
Material and Methods
Systematic Review
Results
116
120
Conclusion
References Cited
120
Clupter 9. THE TO'AGA CERAMICS. T. L. Hunt and C. Eficelens
123
The Assemblage
Analytic Procedures
Results of Macroscopic Analysis
Ceramic Compositional Microanalysis
Conclusions
References Cited
123
125
128
137
145
148
Chapter 10. SAND TEMPER IN PREHISTORIC POTSHERDS FROM
THE TO'AGA SITE. W. R. Dickinson
To'aga Temper Variants
151
Temper Grain Types
Profuse Basaltic Temper
Sparse Basaltic Temper
Feldspathic Basaltic Temper
Mixed Temper Sand
151
152
152
153
154
154
Discussion: Temper Comparisons
References Cited
155
156
Chapter I 1. NON4ERallP
THE
T(YAGA
sum,
En.
157
j_157
-
~~~~~~~~~~~~~~~~~~~~~1
Ornaments
Miscellaneous Artifacts
Worked Shell
Unretouched Lithics
References Cited
162
164
164
165
166
Chapter 12. CHEMICAL CHARACTERIZATION AND PROVENANCE OF MANUA ADZ
MATERIAL USING A NON-DESTRUCTIVE X-RAY FLOURESCENCE
TECHNIQUE. Marshall Weisler
Intduction
167
167
168
170
175
185
185
185
The Collections
Methods
Results
Summary and Conclusions
Acknowledgements
References Cited
Chapter 13. FAUNAL ASSEMBLAGES FROM THE TOAGA SITE.
Lisa Nagaoka
189
189
190
197
208
208
214
214
214
Methods
Results
Analysis of Bulk Samples
Temporal Trends in the To'aga Assemblage
Regional Comparisons
Conclusions
Acknowledgements
References Cited
Chapter 14. BIRD BONES FROM THE TO'AGA SITE: PREHISTORIC
LOSS OF SEABIRDS AND NEGAPODES. David W. Steadman
Intduction
217
217
217
226
226
226
Species Accounts
Discussion
Acknowledgements
References Cited
Chapter 15. SYNTHESIS AND INTERPRETATIONS.
P. V. Kirch and T. L. Hunt
229
Cultural Sequence and Chronology
Geoarchaeology and Landscape Change
230
232
Ancestral Polynesian Culture
Subsistence Patterns
Inter-island Contacts and Exchange
The To'aga Site: Cultural Resource
Management Considerations
References Cited
236
240
243
243
246
vii
CONTRIBUTORS
William R. Dickinson, Department of Geosciences, University of Arizona, Tucson, AZ 85721
Conrad Erkelens, Department of Anthropology, University of Hawaii, Honolulu, HI 96822
Terry L. Hunt, Department of Anthropology, University of Hawaii, Honolulu, HI 96822
Patrick V. Kirch, Department of Anthropology, University of California, Berkeley, CA 94720
Elizabeth Manning, Department of Anthropology, University of California, Berkeley, CA 94720
Lisa Nagaoka, Department of Anthropology, University of Washington, Seattle, WA 98195
David W. Steadman, Biological Survey, New York State Museum, Albany, NY 12230
Jason Tyler, Department of Anthropology, University of Washington, Seattle, WA 98195
Marshall Weisler, Department of Anthropology, University of California, Berkeley, CA 94720
PREFACE
In 1985 I received a request from te Historic Preservation Officer of American Samoa, Mr. Stan Sorensen, to submit a proposal for archaeological survey work
in American Samoa. Realizing that the Manu'a Islands, lying at the extreme eastern
end of the Samoan archipelago, were the most neglected part of the Samoan group
from the viewpoint of prehistory, I suggested that we might undertake a reconnaissance of the three islands-Ta!u, Ofu, and Olosega-that make up Manu'a. A
previous survey in 1962, by Y. Sinoto and W. Kikuchi of the Bernice P. Bishop
Museum, had revealed the presence of various surface archaeological sites, but had
failed to establish any significant time depth for the prehistoric occupation of these
islands. The proposed reconnaissance was carried by Terry Hunt and myself from
June to August, 1986 and resulted in the discovery of the first pottery-bearing sites in
the Manu'a group, on Ta'u and Ofu, as well as a range of later prehistoric monuments
and artifacts. (Tbe results of the 1986 reconnaissance survey were subsequently
published by Hunt and myself in The Journal ofthe Polynesian Society, Vol. 97,
pages 153-83, 1988.)
From the results of our initial survey, it appeared that ceramic-bearing deposits
exposed by a Works Department landfill at To'aga on Ofu Island were particularly
promising for further archaeological investigations. Thus, in 1987 we proposed to the
Historic Preservation Office a second phase of work at To'aga. In this second phase,
we recommended a series of systematic subsurface test excavations in order to
determine more accurately the nature and extent of the To'aga site deposits. This
second field season in 1987 was again directed by Hunt and myself, with the assistance of Jason Tyler and Jean Gehman. (Preliminary results of this work were
published in Archaeology in Oceania, Volume 25, pages 1-15, 1990).
The 1987 excavations revealed the presence of a deep, well-stratified prehistoric record in the To'aga area, with subsurface deposits spanning most-if not allof Manu'an prehistory. These strata were so extensive, however, that we were not
able to determine their limits in the 1987 field season. Therefore, a third season was
proposed to define more precisely the full subsurface extent of the site. This work
was carried out in the summer of 1989, with fieldwork co-directed by Hunt and me,
and with the assistance of a team of students from the University of Hawaii and the
University of California at Berkeley.
This monograph constitutes the final report of our thre field seasons at the
To'aga site. While our core objectives were those of site survey and inventory (with
the aim of preparing a National Register of Historic Places nomination for To'aga),
we have endeavored to go well beyond these minimal goals in thle present work.
Thus, we have carried out extensive analyses of much of the excavated material,
including faunal remains, ceramics, and basalt artifacts. In addition, we have paid
considerable attention to the geomorphological and geoarchaeological problems of
site formation at To'aga, noting that thee also impinge on cultural resource manage1X
ment considerations. We trust that these efforts contribute significantly to our
knowledge of Western Polynesian prehistory as well as to site survey and inventory in
American Samoa.
During the course of our tree expeditions to Manu'a we have received the
support and assistance of a great many individuals and organizations. Our primary
source of funding has been the Historic Preservation Office of the Department of
Parks and Recreation, Government of American Samoa, through grants from the U. S.
National Park Service, Archaeological Assistance Division. Additional support has
been provided by the Burke Museum of the University of Washington (for the 1986
and 1987 seasons), and the Archaeological Research Facility of the University of
California at Berkeley (for the 1989 season). We are particularly grateful to Stan
Sorensen, the Historic Preservation Officer of American Samoa, for his interest and
support from the inception of this project. Anne Sauter of the Archaeological Research Facility at Berkeley assisted in various administrative matters. Tanya Smith
carefully edited and produced this final monograph.
Two respective district governors of Manu'a have lent their support to the
project: High Chief Aolaolagi Soli and High Chief Tufele Li'a. We are also pleased
to record the support of the people of Ofu Village, especially High Chief Misa'alefua,
High Talking Chief Faoa, Liulamaga Ta'ilele, Manu'a Peau, and Sina Peau. In 1987,
Tito and Margaret Malae assisted greatly with housing and other arrangments.
Likewise, in 1989 Manu'a Peau and her family provided housing, hospitality, and a
sense of being at home with family.
Our field crews of Ivala Live, Fuave'a Ta'ilele, Tillis Thompson, Pauesi
Malo, Eleloi Misa'alefua, Ele'ele Utuone, Paulo Su'e, and Opetaia Fa'amita were not
only dedicated workers, but interested co-investigators. We also thank Jason Tyler,
Jean Gehman, Elizabeth Manning, Conrad Erkelens, Melissa Kirkendal, Robert
Holsen, Lisa Nagaoka, and Ann Rowberg, who assisted in the 1987 and 1989 field-
work.
Prof. Roger C. Green of the University of Auckland kindly reviewed the
entire manuscript, and we are most grateful for his insightful comments.
Throughout all field seasons, John and Dorothy Kneubuhl ofTutuila
generously opened their home to us and to our assistants, and helped us in countless
ways. John's profound knowledge of Samoan culture and history was a constant
inspiron. His visit to our field site in 1989, arriving with only a suitcase crammed
full of two large legs of lamb, mint sauce, and a bottle of good scotch whiskey, was
an event not soon to be forgotten. It was with profound sadness that we learned of his
death in Tutuila in February of 1992. As a small token of ouresteem forJoln we
dedicate this volume to his memory.
Parick V. Kirch
Berkee,January 1991
x~~~~~~~~.
.^
LIST OF FIGURES
3
1.1. Map of the Samoan archipelago.
2.1. Map of the Manu'a Islands.
2.2. View of the exposed volcanic plugs at Fa'ala'aga from the beach at To'aga.
2.3. Sketch of basalt dikes exposed in the road cut across the Le'olo Ridge at Fa'ala'aga.
2.4. Map of Ofu Island.
2.5. Distribution of major floral dominants along Transect 7 at To'aga.
10
13
13
15
20
3.1. Map of the southern portion of Ofu Island showing site AS-13-1.
3.2. Plan of round-ended stone house foundation.
3.3. Large volcanic boulder at the base of the talus slope with twelve artificial grinding facets. 27
3.4. Plan and cross section of the Tui Ofu well at Muli'ulu.
29
3.5. Plan and cross section of the Tui Ofu Tia, stone mound/tomb at Muli'ulu.
25
25
4.1. A generalized transect across the coastal plain at To'aga.
35
4.2. A model for the sediment budget at To'aga, showing terrestrial and marine inputs.
4.3. A model of shoreline transgression and regression.
4.4. Map of the southwestern Pacific region showing locations with evidence for midto late-Holocene higher sea-level stands.
4.5. Time-elevation plot of sea levels in Polynesian archipelagoes
4.6. Time trends in four key variables affecting the morphodynamics of the To'aga coastal plain.
40
4.7. Retrodiction of the transgression-regression sequence along the To'aga coastline.
33
5.1. Example of the excavation recording form used in the 1987 and 1989 To'aga field seasons. 45
5.2. Elevation profile along the 1987 excavation transect.
48
5.3. Map of the 1987 excavation area, showing the locations of excavation units.
5.4. Excavation in progress in the 1987 main trench.
5.5. Stratigraphic section of the southwest face of the 1987 main excavation (Units 5-9). 50
5.6. Stratigraphic section of the south and west faces of Unit 3.
5.7. Stratigraphic section of the north and east faces of Unit 10.
5.8. Schematic map of the southern coastal flat of Ofu Island, from To'aga to Fa'ala'aga. 59
5.9. Elevation profile along Transect 3.
5.10. Stratigraphic profile of the east face of Unit 27.
5.11. Elevation profile along Transect 5.
5.12. Stratigraphic section of the west face of Unit 17.
5.13. Exploded stratigraphic section of all faces of Unit 17.
5.14. Stratigraphic section of the west face of Unit 16.
5.15. Stratigraphic section of the west face of Unit 15
5.16. Stratigraphic correlations between Units 15, 16, and 17 on Transect 5.
5.17. Elevation profile along Transect 7.
5.18. Stratigraphic section of the west face of Unit 19.
5.19. Stratigraphic section of the east and south faces of Unit 19.
5.20. Elevation profile along Transect 9.
28
35
36
37
39
47
49
52
53
60
61
62
63
64
65
66
68
69
70
71
72
5.21. Expansion of Unit 20 into Unit 23
5.22. Stratigraphic section of the west and north faces of Units 20/23.
5.23. View into Units 20/23.
5.24. Stratigraphic section of the south face of Unit 21.
5.25. Stratigraphic section of the north face of Unit 22.
5.26. Stratigraphic correlations between Units 20/23, 21, and 22.
5.27. Elevation profile along Transect 11.
5.28. Stratigraphic section of the west face of Unit 26.
5.29. Elevation profile along Transect 17.
5.30. Stratigraphic section of the west face of Unit 24.
5.31. Stratigraphic section of the north face of Unit 25.
6.1. Plot of radiocarbon age determinations from the To'aga site (cal B.P. at 1 sigma).
73
74
75
76
77
78
79
80
81
81
82
90
7.1. Photomicrograph (lOX) of the 1 phi size fraction of a modem control
sample of calcareous beach sand from Transect 7.
7.2. Terminology adopted for the To'aga site sediments.
7.3. Photomicrograph of a typical "salt-and-pepper" sand with mixed calcareous and
95
97
volcanic-lithic lithology.
99
7.4. Summary diagram for sediment samples from Unit 15, Transect 5.
7.5. Histogram plots of grain size distributions for sediment samples from Unit 15, Transect 5. 102
7.6. Photomicrographs (lOX) of the 0 phi size fractions of sediment samples
from Unit 15, Transect 5.
7.7. Summary diagram for sediment samples from Unit 19, Transect 7.
7.8. Histogram plots of grain size distributions for sediment samples from Unit 19, Transect 7. 106
7.9. Summary diagram for sediment samples from Unit 23, Transect 9.
7.10. Photomicrographs (lOX) of the 0 phi size fractions of sediment samples from
Unit 23, Transect 9.
7.11. Histogram plots of grain size distributions for sediment samples from Unit 23, Transect 9. 111
7.12. Summary diagrams of grain size and lithology for Units 21, 22, and 23, Transect 9. 112
7.13. Summary diagrams of grain size and lithology for Units 24 and 25, Transect 17.
113
7.14. Histogram plots of grain size distributions for sediment samples from Unit 24, Transect 17. 114
8.1. Photomicrographs of terrestrial gastropods recovered from the To'aga site sediments.
9.1. Plot of exterior sherd color (hue with value and chroma) for the late period.
9.2. Plot of exterior sherd color (hue with value and chroma) for the middle period.
9.3. Plot of exterior sherd color (hue with value and chroma) for the early period.
9.4. Thickness and temper size histogram for the To'aga ceramic assemblage.
9.5. Thickness histogram for late period ceramics.
9.6. Thickness histogram for middle period ceramics.
9.7. Thickness histogram for early period ceramics.
9.8. Cross sections of selected rim sherds from the To'aga assemblage.
9.9. Elemental spectrum of clay composition for Sherd 22 from To'aga.
9.10. Dendrogram of Ofu pottery and clay samples using the Average
Linkage (between groups) method.
xii
101
103
104
107
109
117
130
130
131
135
137
138
139
140
141
142
9.11. Dendrogram of Ofu pottery and clay samples using Ward's method.
9.12. First and second discriminant function scores for analysis based
on cluster (1-4) as grouping variable.
9.13. First and second discriminant function scores for analysis based on time period.
143
145
145
11.1. Basalt adzes from the To'aga site.
11.2. Turbo-shell fishhooks from the To'aga site.
11.3. Turbo-shell fishhooks and fishhook fragments from the 1987 excavation at To'aga.
11.4. Miscellaneous artifacts from the To'aga site.
11.5. Miscellaneous artifacts from the To'aga site.
12.1. Map of the Tatagamatau quarry complex, Tutuila Island, showing the location of
analyzed source rock samples.
12.2. A comparison of source rocks analyzed by non-destructive x-ray fluorescence.
12.3. A wide dispersion of data points results when ratios of rubidium, strontium, and
yttrium are used to assign specimens to possible sources.
12.4. Artifacts and source rocks plotted by ratios of zirconium, strontium, and niobium
resulting in many specimens plotting within the Tatagamatau source envelope.
159
160
161
163
165
171
172
175
182
197
13.1. Relative frequency of fish taxa from the To'aga site, including Diodontidae.
198
13.2. Relative frequency of fish taxa from the To'aga site, excluding Diodontidae.
13.3. Percentage composition of fish faunal assemblages from major Westem Polynesian and
Fijian sites.
205
14.1. Map of the Pacific Islands showing localities mentioned in the text.
14.2. Comparison of megapode bones.
219
223
15.1. Schematic map of the To'aga area showing inferred extent of ceramic-bearing deposits.
xiii
233
LIST OF TABLES
2.1. Environmental characteristics of the Manu'a Islands.
11
3.1. Surface archaeological features at To'aga.
26
7.1. Analytic data for To'aga control sediment samples (1989).
7.2. American Samoa, To'aga site 13-AS-I sediment analysis (1987).
7.3. Analytic data for Unit 15 sediment samples.
7.4. Analytic data for Unit 16 sediment samples.
7.5. Analytic data for Unit 17 sediment samples.
7.6. Analytic data for Unit 18 sediment samples.
7.7. Analytic data for Unit 19 sediment samples.
7.8. Analytic data for Unit 21 sediment samples.
7.9. Analytic data for Unit 22 sediment samples.
7.10. Analytic data for Unit 23 sediment samples.
7.11. Analytic data for Unit 24 sediment samples (micro-constituents).
7.12. Analytic data for Unit 24 sediment samples.
7.13. Analytic data for Unit 25 sediment samples.
95
96
100
100
101
104
105
105
107
108
108
110
110
8.1. Non-marine molluscs from 1987 Main Trench.
8.2. Non-marine molluscs from Unit 3.
119
119
9.1. Ceramics from the 1987 excavation units.
9.2. Ceramics from the 1989 excavation units.
9.3. Frequency of body and rim sherds.
9.4. Frequency of sherds by temper size modes.
9.5. Frequency of sherds by "Preferred" orientation of inclusions relative to vessel walls.
9.6. Frequency of sherds by oxidation-reduction pattern in cross section.
9.7. Frequency of sherds by exterior Mohs hardness.
9.8. Frequency of sherds by interior Mohs hardness.
9.9. Frequency of sherds by exterior surface treatment.
9.10. Frequency of sherds by interior surface treatment.
9.11. Frequency of sherds by interior anvil casts.
9.12. Frequency of sherds by exterior paddle marks.
9.13. Frequency of sherds by organic residue.
9.14. Sherd thickness statistics by analytic time periods.
9.15. Sherds and Ofu clays selected for SEM/EDS clay elemental and
sand temper petrographic analyses.
126
127
128
129
132
10.1. Frequency percentages of ferromagnesian mineral grains and
basaltic volcanic lithic fragments in sherds containing "profuse basaltic temper." 153
10.2. Frequency percentages of ferromagnesian mineral grains and basaltic volcanic
lithic fragments in sherds containing "sparse basaltic temper."
xiv
132
133
133
134
134
134
135
136
136
144
154
10.3. Frequency percentages of plagioclase feldspar and opaque iron oxide mineral grains and
basaltic volcanic lithic fragments in sherds containing "feldspathic basaltic temper." 155
10.4. Frequency percentages of calcareous grains and silicate-oxide mineral grains and lithic
fragments for sherds containing "mixed temper sand."
12.1. Artifacts analyzed by EDXRF.
12.2. Evaluation of analytical accuracy and precision for U.S.G.S. Standard RGM-1.
12.3. Variations of oxides and elements from the Mo'omomi adz quarry, Moloka'i, Hawaiian Islands.
12.4. Geochemistry of Tatagamatau source rock (fused disks).
12.5. Geochemistry of Mako ridge source rock (fused disks).
12.6. Geochemistry of Fa'ala'aga source rock (fused disks).
12.7. Geochemistry of Fa'ala'aga, Mako ridge, and Tatagamatau source rock (fused disks). 179
12.8. Geochemistry of Tatagamatau source rock (whole specimen).
12.9. Geochemistry of Mako ridge source rock (whole specimen).
12.10. Geochemistry of Fa'ala'aga source rock (whole specimen).
12.11. Geochemistry of Fa'ala'aga, Mako ridge, and Tatagamatau source rock (whole specimen).
12.12. Geochemistry of To'aga and Ofu Island artifacts (whole specimen)
13.1. Summary of To'aga faunal remains.
13.2. Fish fauna from the 1987 field season (NISP).
13.3. Fish fauna from the 1989 field season (NISP).
13.4. To'aga fish fauna: main trench, Units 1,4-9.
13.5. To'aga fish fauna: transect 5, Units 15/29/30 (NISP).
13.6. To'aga fish fauna: transect 9, Units 20/23 (NISP).
13.7. Modern Samoan fishing methods (after Hill 1986).
13.8. Non-fish vertebrate fauna for 1987 and 1989 excavations.
13.9. To'aga non-fish vertebrate fauna.
13.10. To'aga invertebrate data: transect 9, Units 20/23.
13.11. To'aga invertebrate data: 1987 main trench, Units 1,4-9.
13.12. To'aga invertebrate data: transect 5, Units 15/29/30.
13.13. Unit 20/23 bulk sample analysis.
13.14. Unit 30 bulk sample analysis.
13.15. Taxa represented in bulk sample from Units 20/23 (NISP).
13.16. Taxa represented in bulk sample from Layer II, Unit 30 (NISP).
13.17. Density of identified fish bone from Layer IIIA/B, Unit 20/23.
13.18. Density of identified fish bone from Layer II, Unit 30.
13.19. Summary of recovery and quantification techniques for Western Polynesian
faunal analyses.
13.20. Summary of Western Polynesian vertebrate faunal assemblages (NISP).
14.1. Birds from the To'aga site.
14.2. Measurements (in mm) of the femur and ulna of Megapodius.
14.3. Tarsal length (in mm) from skins of selected subspecies of Megapodiusfreycinet.
xv
155
169
173
174
176
177
178
180
181
182
183
184
190
191
192
193
194
195
196
199
200
201
203
205
207
207
208
209
210
210
212
213
218
224
225
10.3. Frequency percentages of plagioclase feldspar and opaque iron oxide mineral grains and
basaltic volcanic lithic fragments in sherds containing "feldspathic basaltic temper." 155
10.4. Frequency percentages of calcareous grains and silicate-oxide mineral grains and lithic
fragments for sherds containing "mixed temper sand."
12.1. Artifacts analyzed by EDXRF.
12.2. Evaluation of analytical accuracy and precision for U.S.G.S. Standard RGM-1.
12.3. Variations of oxides and elements from the Mo'omomi adz quarry, Moloka'i, Hawaiian Islands.
12.4. Geochemistry of Tatagamatau source rock (fused disks).
12.5. Geochemistry of Mako ridge source rock (fused disks).
12.6. Geochemistry of Fa'ala'aga source rock (fused disks).
12.7. Geochemistry of Fa'ala'aga, Mako ridge, and Tatagamatau source rock (fused disks). 179
12.8. Geochemistry of Tatagamatau source rock (whole specimen).
12.9. Geochemistry of Mako ridge source rock (whole specimen).
12.10. Geochemistry of Fa'ala'aga source rock (whole specimen).
12.11. Geochemistry of Fa'ala'aga, Mako ridge, and Tatagamatau source rock (whole specimen).
12.12. Geochemistry of To'aga and Ofu Island artifacts (whole specimen)
13.1. Summary of To'aga faunal remains.
13.2. Fish fauna from the 1987 field season (NISP).
13.3. Fish fauna from the 1989 field season (NISP).
13.4. To'aga fish fauna: main trench, Units 1,4-9.
13.5. To'aga fish fauna: transect 5, Units 15/29/30 (NISP).
13.6. To'aga fish fauna: transect 9, Units 20/23 (NISP).
13.7. Modern Samoan fishing methods (after Hill 1986).
13.8. Non-fish vertebrate fauna for 1987 and 1989 excavations.
13.9. To'aga non-fish vertebrate fauna.
13.10. To'aga invertebrate data: transect 9, Units 20/23.
13.11. To'aga invertebrate data: 1987 main trench, Units 1,4-9.
13.12. To'aga invertebrate data: transect 5, Units 15/29/30.
13.13. Unit 20/23 bulk sample analysis.
13.14. Unit 30 bulk sample analysis.
13.15. Taxa represented in bulk sample from Units 20/23 (NISP).
13.16. Taxa represented in bulk sample from Layer II, Unit 30 (NISP).
13.17. Density of identified fish bone from Layer IIIA/B, Unit 20/23.
13.18. Density of identified fish bone from Layer II, Unit 30.
13.19. Summary of recovery and quantification techniques for Western Polynesian
faunal analyses.
13.20. Summary of Western Polynesian vertebrate faunal assemblages (NISP).
14.1. Birds from the To'aga site.
14.2. Measurements (in mm) of the femur and ulna of Megapodius.
14.3. Tarsal length (in mm) from skins of selected subspecies of Megapodiusfreycinet.
xv
155
169
173
174
176
177
178
180
181
182
183
184
190
191
192
193
194
195
196
199
200
201
203
205
207
207
208
209
210
210
212
213
218
224
225
1
INTRODUCTION AND RESEARCH DESIGN
PATRICK V. KIRCH AND TERRY L. HUNT
1THE WESTERN POLYNESLAN ISLANDs-and
particularly the large Tongan and Samoan
archipelagoes-have long occupied a central focus
in scholarly endeavors to decode the history of
Polynesian origins and dispersals. Burrows (1939)
recognized and defined the distinctive cultural
pattems that set the Westem Polynesian societies
apart from those of Eastem Polynesia. With the
advent of modem stratigraphic excavation and of
radiocarbon dating in Polynesia during the 1950s, it
became apparent that Tonga and Samoa had been
settled considerably earlier than any other
Polynesian archipelagoes. Furthermore, pottery
assemblages from both Tonga and Samoa had
obvious relationships with early ceramic complexes in Melanesia to the west (Golson 1962).
Thus, when Suggs attempted his pioneering
synthesis of Polynesian prehistory, he could
rightly claim that "the islands of Westem
Polynesia were the earliest occupied of all the
Polynesian triangle and the source of all subsequent settlements in Polynesia" (1960:101).
As archaeological knowledge of the Westem
Polynesian archipelagoes continued to accumulate,
the relationship between the early Tongan and
Samoan materials occasioned debate and controversy (e.g. Green 1967, 1968; Groube 1971). The
presence of early (ca. 1200 B.C.) dentate-stamped,
Lapita-style pottery in Tonga and, in Samoa of only
later (ca. 200 B.C. tO A.D. 300) Polynesian Plain
Ware assemblages appeared to support the hypothesis of Tonga as the original Polynesia "homeland."
In 1973, the discovery of a submerged Lapita site at
Mulifanua off the 'Upolu coast in Westem
Samoa (Jennings 1974) brought the Samoan
sequenc back to a comparable antiquity to tat of
Tonga. Subsequently, ealy Lapita materials were
recovered from a number ofthe smaller Westem
Polynesian islands, including Futuna (Kirch 1981;
Frimigacci 1990), 'Uvea, and Niuatoputapu (Kirch
1988b). These discoveries indicate that the entire
Westem Polynesian region was rapidly explored and
colonized by Lapita people during the penultimate
centries of te second millennium B.C.
In our current conception, the adjacent
islands and archipelagoes of Western Polynesia
constituted a homeland region in which a
distinctive Ancestral Polynesian Culture
emerged out of its immediate Lapita ancestor
during the period from about 1000 to 500 B.C.
(Kirch 1984; Kirch and Green 1987). Presumably, the processes of cultural change and
differentiation during this time period were not
uniform across all islands. Regular inter-island
contact through exchange, however, kept each
local community from becoming isolated and
facilitated the spread of innovations. As Kirch
and Green (1987) have argued, the reconstruction
2
The To'aga Site
of Ancestral Polynesian Culture-not as a static
entity, but as a dynamic and changing configuration
over the course of 500 or more years-is a high
priority for Polynesian prehistory. Only by
understanding the technology, economy, settlement pattems, and socio-political organization of
this ancestral culture can we provide a secure
baseline for studying the subsequent development
and diversification of later Polynesian groups
throughout the vast Polynesian triangle. Reconstruction of Ancestral Polynesian Culture is
essential if we are to disentangle those traits and
institutions which are shared retentions, from
those which are later innovations in various island
societies.
The present monograph represents a modest
contribution toward the ultimate goal of tracing
the development of Ancestral Polynesian Culture
out of its Lapita roots. The To'aga site, situated on
the tiny island of Ofu in the Manu'a Group of
American Samoa, spans virtually the entire threemillennium-long sequence of Samoa, but is
especially rich in well-stratified materials dating to
the period from ca. 3200 to 1900 B.P. Our research at To'aga was initiated in order to meet
certain cultural resource management (CRM)
demands, as part of an archaeological survey and
inventory of American Samoan sites under the
auspices of the Office of Historic Preservation of
the Govenmment of American Samoa (Pago Pago).
Fortunately, the goals of this CRM project happily
meshed with the objectives of academically
oriented archaeological research, providing the
opportmity to add to our knowledge of the early
phases of Westem Polynesian prehistory while at
the same time addressing contemporary historic
preservation and land use management concems.
The To'aga site was discovered during our
1986 reconnaissance survey of the Manu'a Islands
(Hunt and Kirch 1987), one of two sites containing
pottery and thus dating to the earliest period of
Samoan prehistory. In our subsequent 1987 and
1989 field seasons we concentrated on the extensive To'aga site, seeking to define the spatial
extent of its deeply buried deposits, its chronology, stratigraphy, material culture, faunal assemblages, and other aspects of intra-site variation.
As our investigations progressed, it became
apparent that To'aga was of critical importance to
understanding the first millennium of Samoan-
and Westem Polynesian-prehistory. No other site
presently known in the Samoan archipelago
encapsulates such a continuous stratigraphic
record, nor one so rich in artifactual and faunal
materials. Although the scope and extent of our
excavations had to be limited by the CRM-funded
nature of our project, we have nonetheless made
every effort to push the analysis of our materials
farther than is usual in such reconnaissance surveys. For example, using a transect sampling
methodology, combined with detailed sediment
analysis and radiocarbon dating, we have been able
to test a model of shoreline progradation linked
with a mid-Holocene higher sea level. This model
has considerable geoarchaeological implications
for the formation and subsurface burial of early
Polynesian sites thrughout the tropical central
Pacific region. We have also pushed the analysis
of the archaeological record of Ancestral
Polynesian material culture at the To'aga site
through detailed studies of pottery (including
analysis of temper and of the chemical composition
of clays), adzes (including EDXRF analysis of
basalts), and other artifact classes. Likewise, the
samples of faunal materials have been subjected to
intensive studies, including extinct and extirpated
avifaunal remains, molluscan and fish faunal
assemblages, and terrestrial (synanthropic) land
snails. In the fourteen chapters that follow, we and
our collaborators present the detailed results of our
field and laboratory investigations of the To'aga
site.
BACKGROUND TO SAMOAN
ARCHAEOLOGY
Although Buck (1930) reported on stonework
and adzes, the first modem archaeological effort in
Samoa was that of Golson (n.d., 1962) in 1957,
resulting both in a general account of the range of
field monuments and in the discovery of prehistoric pottery dated to the first century A.D. at
Vailele, 'Upolu. The latter discovery was particularly significant in the then-emerging picture of
Polynesian origins as rooted in an earlier
Melanesian ceramic complex beginning to be
known by the term "Lapita" (Suggs 1961; Golson
1971; see also Kirch 1988a). Following Golson's
lead, a major archaeological program was organized for Westem Samoa under the direction of
Introduction and Research Design
Roger C. Green (University of Auckland), with
funding provided trugh the Bishop Museum's
Polynesian Culture History Program. Between
1962 and 1967, this project brought seventeen
archaeologists from nine institutions to Westem
Samoa for a coordinated series of investigations
including surface surveys and excavations of sites
spanning early settlement to the historic period.
Published in two large volumes (Green and
Davidson 1969, 1974), the results of this project
are a landmark in Polynesian archaeology.
Subsequent to the conclusion of the Westem
Samoa project in 1967, but in time to be incorporated in the second volume of results (Green and
Davidson 1974), an accidental discovery of
classically decorated Lapita pottery at Mulifanua,
'Upolu extended the Samoan sequence back to the
beginning of the first millennium B.C. (fig. 1.1).
Furthermore, the geomorphological context of this
find-a submerged site capped by nearly one meter
of reef rock-demonstrated that tectonically
induced changes in the Samoan landscape could
have significant implications for regional prehis-
tory.
In the 1970s, a project headed by Jesse
Jennings of the University of Utah completed two
seasons of archaeological research on 'Upolu
3
Island and on the adjacent small islet of Manono,
concentrating on settlement pattem surveys and
excavation of both plain ware and aceramic sites
(Jennings et al. 1976; Jennings and Holmer 1980).
Of particular note is their work on settlement
patterns in which they propose the concept of the
"household unit" as an analytical category
(Jennings et al. 1982).
As a result of these various projects, the
outline of a prehistoric cultural sequence for
Westem Samoa is reasonably well attested
(Davidson 1979). In conventional outline, this
sequence begins with the occupation of the archipelago between about 1200-1000 B.C. by makers of
classic, dentate-stamped Lapita pottery (represented by the presently submerged "Ferry-Berth
Site" at Mulifanua). The first millennium B.C.
witnessed changes in the composition of Samoan
ceramic assemblages, particularly the loss of
decoration and of more complex vessel shapes,
ending with Polynesian Plain Ware assemblages
around the time of Christ. Stone adzes and other
aspects of material culture also changed with the
pottery, and this sequence as a whole is viewed as
documenting the development of an Ancestral
Polynesian Culture out of an older Eastem Lapita
culture (Kirch and Green 1987).
Figure 1.1 The Samoan Archipelago.
4
The To'aga Site
Ceramnics ceased to be manufactued by about
A.D. 300, and e paucity of archaeological materials
during the next thousand years or so prompted
Davidson (1979) to tenn this the "Dark Ages" of
Westem Polynesian prehistory. In the final millennium of the Samoan sequence a number of new
developmenas are evidenced in the archaeological
record, especially the construction of several new
fonns of field monuments, such as the star mound
and ridge-top forifications. These large constructions are believed to reflect the rise of powerful
chiefdoms competing for land and resources.
The Samoan culurl sequence briefly oudined
above parallels-in many key aspects-the sequences developed by archaeologists for other
Westem Polynesian islands, such as Tongatapu
(Poulsen 1987), Niuatoputapu (Kirch 1988a), Futuna
(Kirch 1981; Frimigacci 1990), and 'Uvea
(Frimigacci, Siorat and Vienne 1984).
In contrast with Westem Samoa, the archaeology
and prehistory of American Samoa are less well
known. The first modem archaeological survey in
Amercan Samoa was caried out by W. Kikuchi
(1963, 1964) on Tutuila and Aunu'u Islands in 196162. Kikuchi's work provided an overview of the
main kinds of surface sites, but was neither intensive
nor systenatic in its coverage. In 1962, Kikuchi and
Y. Sinoto of the Bishop Museum extended the survey
to the Manu'a Group and conducted minor test
excavations on both Ta'u and Tutuila Islands.
Disappointed by their results, Sinoto decided to
abandon the Samnoan project, and shifted his locus of
field work to the Marquesas Islands of Eastem
Polynesia (Emory and Sinoto 1965). Prior to 1980,
the only other major field project conducted in
American Samoa was that of Janet Frost (1978), who
carred out limited test excavations at seven sites on
Tutuila. There have also been severl limited
cultual resource surveys, carried out under contract
to the National Park Service, the U.S. Anmy Corps of
Engineer, and the Department ofPublic Works
(Ladd and Moris 1970, Kikuchi, Palama, and Silva
1975; Silva and Palama 1975; McCoy 1977; Atxhns
1987; Hunt 1987). In 1980, J. Clark compiled a
summary of all recorded archaeological sites for the
American Samoan Historic Preservadon Commission, based on ree weeks in the field checking the
locations and status of many of these sites (Clark
1980). Clark summed up the status of archaeological
survey in Amercan Samoa as of 1980 in these
words: "the amount of land that has been intensively
and systematically covered is small indeed"
(1980:11).
Over the course of the past decade, archeological
knowledge of Amercan Samoa has increased
substantially due to the efforts of te Historc Preservadon Office ofthe Deparment of Pariks and Recreation, Government of American Samoa. With funds
made available by the U.S. National Park Service
(NPS), the Historic Preservation Officer has commissioned a variety of archaeological reconnaissance and
intensive survey projects in order to compile an
inventory of significant sites on Tutuila and in the
Manu'a Group. Given the rapid pace of economic
development in American Samoa, many of the sites
revealed through such survey projects have proved to
be threatened by current or projected land use
practices.
On Tutuila Island, a substantial multi-year
project was devoted to the recording and detailed
study of the Tataga-matau basalt adz quarry site,
situated on a complex of ridges inland of Leone
Village (Best, Leach, and Witter 1989; Leach and
Witter 1987; Leach and Witter 1990). This site is of
interest not only for its intrinsic importance to
understading the prehistory of Tutuila, but because
the adzes produced at this major quany have been
shown to have been distributed very widely throughout the Westem Pacific (Best, Sheppard, Green, and
Parker 1992). Also on Tutuila, Jeff Clark and his
associates have carried out a series of surveys and
test excavations which have greatly amplified our
understanding of the archaeological resources of
American Samoa (Clark 1989, in press; Clar and
Herdrich 1988, in press).
When the senior author was asked by te
American Samoan Historic Preservation Officer to
consider undertaking an archaeological survey in the
archipelago, we decided to focus on the thenneglected Manu'a Group. The selection of the
Manu'a Group as the focus of our project was
motivated by several factors. First, the larger island
ofTutuila was already receiving substantial attention
from several other archaeological field teams, leaving
Manu'a as a continuing lacuna. Second, because we
were particuladiy interested in seeking ceramicbearing sites dating to the first third of Samoan
prehistory, we preferred to focus on several smaller
islands where survey and subsurface testing could be
concentrated on likely areas of early occupation.
Introduction and Research Design
Third, as the most easterly and somewhat isolated
section of the Samoan archipelago, the prehistory and
archaeology of Manu'a could conceivably exhibit
significant differences from that of Tutuila, Upolu,
and Savai'i. While such differnces were predictable,
primary fieldwoik would be necessary for their
documentation.
DESIGN OF THE RESEARCH
Over the course of three field seasons, our
research focus in Manu'a evolved from extensive
reconnaissance survey of the surface archaeology to
intensive subsurface examnination of buried archaeological resources. Our field investigations began in
1986 with archaeological reconmaissance of he three
islands of the Manu'a Group: Ta'u, Olosega, and Ofu
(Hunt and Kirch 1987, 1988). In 1986 ceramicbearing sites were discovered at To'aga, Ofu Island
and at Ta'u Village, Ta!u Lsland. In the 1987 field
season, we focused on an intensive survey and
systematic excavation of the site at To'aga (Site AS13-1). Our msults from 1987 (Kirch et al. 1989,
1990) revealed a deeply stratified site containing a
long and continuous sequence of ceramics and dating
to more than 3,000 yearS B.P. (Kirch et al. 1990). A
site of this significance for either American or
Westem Samoa was previously unknown. A third
season at To'aga was therefore designed to determine
the naure and extent of bured archaeological
remains for the entire coastal flat of southem Ofu
(extending from To'aga to Fa'ala'aga). Thus, the
primary objectives of our 1987 and 1989 fieldwork
were to determine the nature, significance, and
spatial and stratigrphic extent of the deposits at site
AS-13-1. This fundamental step was accomplished
primarily for the purpose of cultural resource management, including preservation and public interpretation. The data collected will be used to nominate
site AS-13-1 to the National Register of Historic
Places and to assure its preservation in the face of
any future development plans.
While our Manu&a Project was designed first and
foremost to address the CRM concerns of the
Amercan Samnoa Historic Preservation Office, we
also regarded the project as an opportunity to tackle
several major research problems of Samoan archaeology and prehistory. Especially in the 1987 and 1989
field seasons, when we concentrated on the well-
5
stratified and extensive subsurface deposits at To'aga,
it was possible to design the field and laboratory
sttegies to address the following researh problems:
1. A major objective was the establishment of a
temporal framework and prehistoric sequence for the
Manu'a Islands. Prior to the commencement of our
project, it was uncertain whether the Manu'a Group
would prove to have been colonized at approximately
the same time as the other Samoan Islands, or
whether the main changes and trends in te Manu'a
sequence (such as te timing of ceramnic change and
eventual cessation of pottery manufacture) would
parallel ose on Tutuila, Upolu, and Savai'i.
2. A second research goal was to determine the
nature and magnitude of environmental change
during the period of prehistoric Polynesian occupation of the Manu'a Islands. Interdisciplinary research
on other central Pacific islands over the past two
decades had shown that human activities frequently
have resulted in major changes to the vegetation,
fauna, and landforms of island ecosystms (e.g.,
Kirch 1984:123-51; Kirch 1988b:247-50-, Kirch and
Yen 1982; Bayliss-Smith et al. 1988:1243). At te
same time, natural environmental changes such as
sea-level fluctuations, also were known to have
affected site distribution and the geomorphology of
coastal lowlands. At te To'aga site, we had te
opponunity to address such issues of human-induced
and natnal environmental changes over a 3,000+
year sequence. In particular, we were interested in
determining the geomorphological history of te
To'aga coastal plain, a narrow strip of intensively
used land (see chapter 2, Kirch). To this end, we
developed a morphodynamic model of landfonn
changes at To'aga (chapter 4, Kirch) which we were
able to test on te detailed data of site satigraphy
(chapter 5, Kirch and Hunt), radiocarbon chronology
(chapter 6, Kirch), and sedimentological analysis of
fte excavated deposits (chapter 7, Kirch, Manning,
and Tyler).
3. A third major issue concerned the reconstuction of certin aspects of Ancestral Polynesian
Culture, especially te nature of its settement
pattems and subsistence economy, which were rather
poorly evidenced on archaeological criteria (as
opposed to histonrcal linguistic reconstructions, see
Kirch 1984:53-67). Because the To'aga site spans the
entire firSt millennium B.C.-the period during which
Ancestrl Polynesian Culture developed out of its
6
The To'aga Site
Lapita ancestor-and because the site's calcareous
sandy deposits preserve a wide range of organic
materials, this was again an excellent situation in
which to tackle this researh issue. In particular, we
expended considerable effort in the analysis of the
extensive suite of vertebrate and invertebrate faunal
matetals rwcovered from the excavations (see
chapter 13, Nagaoka and chapter 14, Steadman).
4. A more specific research topic is the explanation of ceramic change in Westem Polynesia,
including aspects of technology, formal variation in
vessel shape and function, and te eventual disappearance of pottery. The To'aga site, with a particulariy long sequence of strtified deposits, yielded a
cermnic assemblage spaning virtually the entire
period of pottery manufacture in Samoa. We
fterefore decided to focus on a detailed examination
of these ceramics, from several analytical perspectives, in order to refine our understanding of ceramic
change in the Samoan archipelago (see chapter 9,
Hunt and Erkelens; and chapter 10, Dickinson).
5. A fifth research issue focussed on te role of
inter-island exchange of such matenal items as
ceranics and basalt adzes. Oceanic prehistorians
have increasingly come to realize that such exchange
was extremely important in maintaining contacts
between island societies (e.g. Kirch 1988a, b; Hunt
1989). In the case of materials frxm To'aga, we
resolved to investigate the role of inter-island exchange trugh an intensive study of the ceramic and
basalt adz artifact assemblages (see chapter 9, Hunt
and Erelens; and chapter 12, Weisler).
In sum, our field and laboratory studies of the
To'aga site was oriented by these five research issues,
combined with the prblems of site delineation and
significance-assessment dictated by the CRM nature
of our conta with the American Samoa Historic
Preservation Office. Because the funds for fieldwork
were limited, we knew that we would be unable to
carry out extensive subsurface excavations or
exposures of large areas. lTus, our field strategy had
to be designed to obtain the kinds of data relevant to
the research issues cited above in the most costeffective manner. In addition to intensive survey of
surface archaeological featumr located throughout
the Toaga area (Hunt, chapter 3), we concentaed on
a series of systematic trsct test excavations
following methods elaborated by Kirch in previous
fieldwork on Ticopia and Niuatoputapu (Kirch and
Yen 1982; Kirch 1988b). These unsect excavations
allowed us to define the spatial extent ofthe site,
especially the deeply buried, pottery-bearing deposits, to gain an overview of the statigraphic sequence
and to obtain sizable samples of ceramics, other
portable artifacts, and vertebrate and invertebrate
faunal remains. Furtfer details of our sampling and
excavation metiodology are provided in chapter 5.
As the various chapters to follow demonstrate,
the To'aga site is a unique and highly significant
archaeological resource, encapsulating dtree millennia of Samoan prehistory. As the pace of development and land use change quickens in American
Samoa, this site is likely to come under increased
dteat. It is our hope that in addition to contributing
to our understanding of Polynesian and Samoan
prehistory, this volume will serve to heighten awareness concerning this valuable archaeological resource. The To'aga site deserves to be carefully
protected and its non-replaceable resources managed
so that future generations will have the opportunity to
fuirter advance our knowledge and understanding of
the past
REFERENCES CITED
Athens, J. S. 1987. Archaeological survey of Leone
Bay, American Samoa. Unpublished manuscipt, Office of Historic Preservation, Pago
Pago.
Bayliss-Smith, T. P., R. Bedford, H. Brookfield, and
M. Latham 1988. Islands, Islanders, and the
World: The Colonial and Post-Colonial
Experience ofEastern Fiji. Cambridge: Cambridge University Press.
Best, S. B., H. Leach, and D. Witter 1989. Report on
the second phase of fieldwork at the TatagaMatau Site, American Samoa, July-August
1988. Unpublished manuscript, Office of
Historic Preservation, Pago Pago.
Best, S., P. Sheppard, R. Green, and R. Parier 1992.
Necromancing the stone: Archaeologists and
adzes in Samoa. Journal of the Polynesian
Society 101:45-85.
Buck, P. H. (Te Rangi Hiroa) 1930. Samoan Materill Culture. Honolulu: Bemice P. Bishop
Museum Bulletin 75.
Burrows, E. G. 1939. Westem Polynesia: A study in
cultural differentiation Ethnological Studies 7.
Gothenburg.
Introduction and Research Design
Clark, J. T. 1980. Historc preservation in Amencan
Samoa: Program evaluation and archaeological
site inventory. Unpublished manuscript,
Deparnment of Antrpology, Bishop Museum,
Honolulu.
. 1989. The Eastem Tutuila archaeological
project: 1988 final report. Unpublished manuscript, Office of Historic Preservation. Pago
Pago.
-. m press. Radiocarbon dates frm American
Samoa. Radiocarbon.
Clark, J. T., and D. J. Herdrich 1988. The Eastem
Tutuila archaeological project: 1986 final
report. Unpublished manuscript, Office of
Historic Preservation, Pago Pago.
in press. Prehistoric settlement system in Eastem
Tutila, American Samoa, Journal ofthe
Polynesian Society.
Davidson, J. 1979. Samoa and Tonga. IN J. Jennings,
ed., The Prehistory ofPolynesia, pp. 82-109.
Cambridge: Harvard University Press.
Emory, K. P., and Y. Sinoto 1965. Preliminary report
on the archaeological investigations in
Polynesia. Mimeographed report on file,
Department of Anthropology, Bishop Museum,
Honolulu.
Frimigacci, D. 1990. Aux Temps de la Terre Noire:
Ethno-ArchAologie des les Futuna etAlofi.
Paris: Peeters.
Frimigacci, D., J. P. Siorat, and B. Vienne 1984.
Inventaire et Fouille des Sites Archeologiques
etEthnohistoriques de l'lle d'Uvea. Documents
Provisoires, ORSTOM, Centre de Noumea,
New Caledonia.
Frost, J. 1978. Archaeological investigations on
Tutuila Island, American Samoa. Unpublished
Ph.D. diss., University of Oregon.
Golson, J. 1962. Report on New Zealand, Westem
Polynesia, New Caledonia, and Fiji. Asian
Perspectives 5:166-80.
-. 1971. Lapita ware and its transfonnations. IN R.
Green and M. Kelly, eds., Studies in Oceanic
Culture History, pp. 67-76. Pacific Andtopological Records 12. Honolulu: Bernice P.
Bishop Museum
nd. Report to TRIPP on archaeological fieldwork in Samoa and Tonga. Unpublished
manuscript, University of Auckland.
Green, R. C. 1967. The mmediate origins of the
7
Polynesians. IN G. A. Highland, et al., eds.,
Polynesian Culture History, pp. 21540.
Bemice P. Bishop Museum Special Publication
No. 56. Honolulu.
.1968. West Polynesian prehistory. IN I. Yawata
and Y. H. Sinoto, eds., Prehistoric Culture in
Oceania, pp. 99-110. Honolulu: B. P. Bishop
Museum.
Green, R. C., and J. M. Davidson, eds. 1969. Archaeology in Western Samoa, Vol. 1. Auckland
Institute and Museum Bulletin 6.
1974. Archaeology in Western Samoa, Vol. 2.
Auckland Institute and Museum Bulletin 7.
Groube, L. 1971. Tonga, Lapita pottery, and
Polynesian origins. Journal of the Polynesian
Society 80:278-316.
Hunt, T. L. 1987. Archaeological survey and assessment of the proposed Fiti'uta Airport site, Ta'u
Island, Manu'a Group, Amercan Samoa.
Unpublished manuscipt, Depailrent ofPublic
Woiks, Pago Pago.
1989. Lapita ceramic exchange in the Mussau
Islands, Papua New Guinea. Unpublished
Ph.D. diss., University of Washington, Seatde.
Hunt, T. L., and P. V. Kirch 1987. An archaeological
reconnaissance of the Manu'a Islands, Amercan
Samoa. Unpublished manuscipt, Office of
Historic Preservation, Pago Pago.
1988. An archaeological survey of the Manu'a
Islands, American Samoa. Journal of the
Polynesian Society 97:153-83.
Jennings, J. D. 1974. The Ferry-Berth Site,
Mulifanua District, Upolu. IN R.D. Green and J.
Davidson, eds., Archaeology in Western Samoa,
Vol. 2 pp. 176-78. Aukland Institute and
Museum Bulletin 7.
Jennings, J. D., R. N. Holmer, J. C. Janetski, and H.
L. Smith 1976. Excavations on Upolu, Western
Samoa. Pacific Anthrpological Records 25.
Honolulu: Bemice P. Bishop Museum
Jennings, J. D., and R. N. Holmer 1980. Archaeological Excavations in Western Samoa.
Pacific Anthropological Records 32. Honolulu:
Bernice P. Bishop Museum
Jennings, J. D., R. N. Holmer, and G. Jackmond
1982. Samoan village patems: Four examples.
Journal of the Polynesian Society 91:81-102.
Kikuchi, W. 1963. Archaeological surface ruins in
American Samoa. Unpublished M.A. thesis,
8
The To'aga Site
University of Hawaii, Honolulu.
-. 1964. Petroglyphs in Amencan Samoa. Journal
of the Polynesian Society 73:163-66.
Kikuchi, W., Palama, S., and T. Silva 1975. Archaeological reconnaissance survey, proposed
Ta'u Harbor at Fusi and quarry site between
Fusi and Fagamoto Ta'u Lsland, Manu'a Group,
Amencan Samoa. Unpublished manuscript,
Deparlment of Anthropolgy, University of
Hawaii, Honolulu.
Kirch, P. V. 1981. Lapitoid settlements of Futuna
and Alofi, Westem Polynesia. Archaeology in
Oceania 16:127-43.
1984. The Evoluion of the Polynesian
Chieldoms. Cambrdge: Cambridge University
Press.
1988a. A brief history of Lapita archaeology. IN
P. V. Kirch and T. L. Hunt, eds., Archaeology
of the Lapita Cultural Conmpex: A Critical
Review, pp. 1-8. Seatte: Tlomas Burke
Memorial Washington State Museum Research
Repoit No. 5.
-. 1988b. Niuatoputapu: The Prehistory of a
Polynesian Chiefdom. Thomas Burke Memorial Washington State Museum Monograph No.
5. Seatle.
Kirch, P. V., and R. C. Green 1987. History, phylogeny, and evolution in Polynesia. Current
Anthropology 28:431-56.
Kirch, P. V., T. L. Hunt, L. Nagaoka, and J. Tyler
1990. An ancestal Polynesian occupation site
at To'aga, Ofu Island, American Samoa.
Archaeology in Oceania 25:1-15.
Kirch, P. V., T. L. Hun, and J.Tyler 1989. A
radiocarbon sequence from te To'aga Site, Ofu
Isand, American Samoa. Radiocarbon 31:7-13.
Kirch, P. V., and D. E. Yen 1982. Tikopia: The
Prehistory andEcology of a Polynesian
Outlier. Honolulu: Bernice P. Bishop Museum
Bulletin 238.
Ladd, E. J., and D. K. Moris 1970. Archaeological
and ecological survey of 'Olovalu Crater, Island
of Tutuila, American Samoa. U.S. National
Pa& Service. Unpublished manuscript.
Leach, H. M., and D. C. Witter 1987. Tataga-matau
'rediscovered.' New Zealand Journal of
Archaeology 9:33-54.
. 1990. Further investigations at the Tataga-matau
Site, American Samoa. New Zealand Journal
ofArchaeology 12:51-83.
McCoy, P. C. 1977. Cultural reconnaissance survey,
Auasi Harbor Project, Auasi, Tutuila Island,
Amercan Samnoa. Unpublished manuscript,
Department of Anthrpology, University of
Hawaii, Honolulu.
Poulsen, J. 1987. Early Tongan Prehistory: The
Lapita Period on Tongatapu and its Relaionships. 2 vols. Terra Ausalis 12. Canberra:
Australian Natonal University.
Silva, T., and S. Palama 1975. Archaeological
reconnaissance survey, proposed shoreline and
highway improvements, Tutuila Island, and
Aunu'u Boat Harbor, Aunu'u Island, American
Samoa. Unpublished manuscrpt, Archaeological Research Center Hawaii.
Suggs, R C. 1960. The Island Civilizaions of
Polynesia. New York: Mentor Books.
. 1961. Archaeology of Nuku Hiva, Marquesas
Islands, French Polynesia. New York: Anthrpological Papers of the American Museum of
Natural History 49(1).
OFU ISLAND AND THE TO'AGA SITE:
DYNAMICS OF THE NATURAL AND
CULTURAL ENVIRONMENT
PATRICK V. KIRCH
Every Manuan thinksfirst ofhis own village, then ofhis island, then ofManua as a whole-a tiny archipelago
set offfrom all the rest of Samoa. There may be rivalry between villages; there have even been warsinternecine conflicts between Fitiuta and the three villages ofTau and between OfQu and Olosega-but against
the outside world Manua presents a solidfront.
Mead (1930:51)
THE MANU'A ISLANDS, ISOLATED from the
main Samoan islands by 100 kn of oftenturblent ocean, form a geographically and culturally
disinctive cluster at the eastem extreme of the
archipelago (fig. 2.1). Being closest to Tutuila,
Manu'ans had the greatest interaction with the
occupants of that island (a pattern that continues
today with Manu'a and Tutuila comprising the Terntory of American Samoa). Although sharing in most
respects the dassic characteristics of Samoan culture,
Manu'ans nonetheless regard themselves as different
and distinctive. Mead (1930:9) commented on this
distinctiveness, for example in the lack of emphasis
on "war, its paraphemalia, its ritual, and its gods."
She opined that "the chief historical value of Manua
lies in her easterly and isolated position, offering a
valuable check upon cultural traits which are intrusive in westem Samoa" (1930:9). Such isolation,
however, was relative and should not be overemphasized; as our archaeological investigations have
revealed, the prehistory of Ofu shares much in
common with the westerly islands of Samoa, and
prehistoric inter-island contacts can be documented
in the transport of basalt adzes (see Weisler, chapter
12), and perhaps of pottery as well (see Hunt and
Erkelens, chapter 9). Thus, despite considerable
isolation, Manu'a was never a "closed system."
The ihree islands of the Manu'a Group-Ta'u,
Olosega, and Ofu-form an intervisible cluster and
comprised a political unity in ancient times. Ta'u is
by far the largest island (table 2.1) and dominated
the group politically, being the seat of the Tui
Manu'a paramount line of chiefs. Olosega and Ofu
are practically adjacent to each other, separated only
by a narnow and shallow strait, now spanmed by a
concrete causeway. They are similar in size, Ofu
being slightly smaller (3.4 kM2), but rising to a
loftier summit (at 638 m) than Olosega.
Ofu and Olosega are visually exquisite, their
youthful volcanic profiles dthust abruptly out of dte
Pacific, rising to basalt pinnacles often shrouded in
misL Narrow ribbons of blindingly white coral sand
fonn a no-man's-land between the rank green of
tropical vegetation and the blue-green mosaic of
coral reefs. Sheer cliffs towering hundreds of meters
above To'aga, Sili, and other villages provide nesting
sites for the white-tailed tropic birds that ride the
trade winds battering the islands' mass.
10
The To'aga Site
Figure 2.1 Map of the Manu'a Isands.
Perhaps the most striking feature of Ofu's
environment is its precipitous topography, and the
scarcity of flat land. No less than ninety-one
percent of the total island area consists of slopes
greater than 30 degrees. Due to its geological
youth, the steep sides of the volcanic cone have not
yet been significantly incised by streams, the
principal weathering being major landslides
resulting from faulting, and creating sheer cliffs.
The only level terrain is found around the margins
of the island, at the critical land-sea interface,
where a narrow coastal terrace has been constructed
primarily from coral sand, with the addition of
some talus and colluvium from the slopes inland.
This coastal zone, ranging from 50 to 150 meters
wide, is unevenly distributed, being found along the
westem and southeastem coasts but not along the
north coast, except for a small strip at the northeastem point. The To'aga site occupies the longest
span of coastal terace land on the island, along the
southeastem shoreline. This coastal terrace was a
critical microenvironment for the Polynesian
occupants of Ofu, first as the setting for their
villages, and second as a major zone of garden
land. More importantly, the narrow coastal terrace
is a highly dynamic environment. Understanding
the morphodynanics of this land-sea interface is
essential not only to reconstructing the prehistoric
sequence of the To'aga site itself, but more broadly
for an understanding of Manu'an prehistory.
In this chapter, I outline the main characteristics and resources of the Ofu Island environment
that were significant to the prehistoric Polynesian
inhabitants with their horticultural and fishing
economy. I also pay considerable attention to those
aspects of the contemporary environment-both
physical and biotic-that inform us about the
dynamics of environmental change. Polynesian
archaeologists have now abandoned an earlier
perspective that viewed island ecosystems merely
as static backdrops to cultural developments.
Increasingly, archaeological investigations throughout Polynesia are yielding striking evidence of
ecological change: forest clearance and vegetational succession, extinction and extirpation of
birds and other biota, alluviation of valley floors,
sea level changes, and so forth (e.g., Kirch 1982,
1984; Steadman 1989; Flenley and King 1984;
Olson and James 1984; McGlone 1983). Some of
these changes resulted from natural processes;
others reflect the impact of human populations
themselves, colonizing isolated and biologicallyvulnerable islands for the first time (Fosberg 1963).
The prehistory encapsulated within the To'aga site
is as much a history of environmental dynamics as
of cultural changes in pottery Wpes or adz forms.
The Natural and Cultural Environment
11
Table 2.1
Environmental Characteristics of the Manu'a Islands
Characteristic
Ta'u
Olosega
Ofu
Area (kIn)
Highest Point (m)
Area<30% slope (%)
Coastline (kan)
Population (1980)
28.5
965
41
32.5
1,146
4.5
494
10
13.3
340
3.4
638
9
10.4
254
GEOLOGY AND GEOMORPHOLOGY
OF OFU ISLAND
Initial studies of Ofu geology by Friedlander
(1910), Daly (1924), and Stearns (1944) were
based on short reconmaissance surveys. Although
Steams spent only one or two days in Manu'a, he
was able to produce "a remarkably accurate
geologic sketch map" (Stice and McCoy
1968:429). The most intensive study to date was
carried out by Stice and McCoy (1968), whose
report and geologic map of Ofu we have drawn
upon for the following summary. Subsequent
work by Natland (1980) and McDougall (1985)
has corrected certain errors made by Stice and
McCoy in dating the island's geological formations.
Ofu and Olosega comprise an integral geological complex "of volcanic cones that have been
buried by lava flows from two coalescing shields"
(Stice and McCoy 1968:443). At least five cones
make up the Asaga Formation; these include a
"composite cone exposed in the cliff behind
To'aga," as well as "an explosion breccia cone
with an associated intrusive plug at Fatuaga Point"
(1968:443-44). The latter plug is exposed as the
visually impressive spire at the eastern point of
Ofu (fig. 2.2). Stice and McCoy believed that the
Asaga Formation was of lower Pliocene age,
although this has now been shown to be incorrect
(see discussion below). All of these Asaga
Fornation cones were subsequently buried by lava
of the Tuafanua Formation, from the A'ofa and Sili
coalescing shields. "After summit collapse of the
shields, volcanism decreased so that a sea cliff
about 300 feet high was cut around the islands"
(1968:456). In the To'aga area, the high cliff
seems to have resulted from a combination of
faulting and sea erosion. Finally, renewed volcanism resulted in the construction of the tuff cone of
Nu'utele Islet, as well as in several hawaiite and
olivine basalt flows that filled in deeply eroded
stream valleys on westem Ofu.
Whereas Stice and McCoy (1968) regarded
the geological history of Ofu to have extended
over a fairly long period, from the lower Pliocene
through to the late Pleistocene, recent work by
Natland (1980) and McDougall (1985) has confirmed that the island is very recent in age.
McDougall reports that "these volcanoes are quite
youthful, confirmed by unpublished K-Ar ages
from this laboratory [Austalian National University] averaging 0.3 Ma for Ofu/Olosega and less
than 0.1 Ma for Ta'u" (1985:318). As Natland
argues, "there can be little doubt that both Tau and
Ofu-Olosega are substantially younger than the
shield volcanoes of Tutuila, which were extinct
and extensively eroded before drowning of the
Pleistocene reefs" (1980:721). Indeed, the Samoan chain illustrates a typical instance of linear,
"hot-spot" progressive volcanism, with the islands
increasing in age from east to west. (Recent,
renewed volcanism on Savai'i is evidently due to
the proximity of the westem end of the archipelago to the Pacific Plate margin.) The youthful
age of Ofu and Olosega is of considerable importance to the geomorphology and geoarchaeology
of the To'aga site, as we shall argue in greater
detail in chapter 4. This is because the islands are
12
The To'aga Site
still tectonically unstable, due to point-loading on
the oceanic crust, and thus are undergoing a phase
of subsidence (Menard 1986).
The rocks of the island consist primarily of
olivine basalts, hawaiites, and ankaramites which
were extruded as both pahoehoe and aa lava flows,
interbedded with various pyroclastic tuffs and
breccias. Intrusive rocks consist primarily of
dikes, and of the plug at Fatuaga Point, a
"hypabyssal intrusion of ankaramite" (1968:449).
A major swarm of near vertical dikes rns through
the central spine of Ofu, and is clearly visible in
the cliffs behind the To'aga site. Stice and McCoy
report that "the razorback ridge of eastem Ofu is
the topographic expression of a dike complex
about 400 feet wide. The dikes are nearly vertical
... Most are dense basalt, although olivine basalt,
ankaramite, and feldspar-phyric basalt also are
present" (1968:449). We were particularly
interested in this dike complex as a possible source
of dense, fine-grained basalt that could have been
exploited by the prehistoric occupants of Ofu for
manufacture of adzes or other flake tools. The
recent road cutting at Fa'ala'aga provided an
opporunity to examine this dike cluster at close
hand, and to collect samples for XRF compositional analysis (see Weisler, chapter 12). As
shown in figure 2.3, the dikes are closely spaced,
cutting trugh older, weathered basalts of the
Asaga Formation. On the edges of several of these
dikes, where they came into contact with the older
basalts, we observed glassy "chills" of low-silica
volcanic glass ("obsidian"), up to 2 cm thick.
Although of poor quality, such volcanic glass
could have been exploited for the production of
small flakes. The appearance of local volcanic
glass flakes in archaeological contexts in the
Manu'a Group is thus a distinct possibility.
We did not observe other possible sources of
fine-grained basalts suitable for adz manufacture,
except for a rather weathered dike exposure on the
edge of Mako Ridge at about 350-400 m elevation
(also sampled for XRF analysis). It is entirely
possible, however, that prehistoric Manu'ans
exploited small exposures of suitable basalt or
hawaiite for adz manufacture. We are not aware
of any adz quaries in the Manu'a Group, but this
negative evidence certainly does not preclude local
adz manufacture.
There has been relatively little stream erosion
on Ofu, and the radial drainage pattem is poorly
developed. A number of shallow valleys drain to
the north and to the west from the slopes of
Tumutumu Mountain. All of these are intennit-
tent, flowing only after heavy rains. "'The stream
valleys are all youthful and nowhere exceed 50
feet in depth" (Stice and McCoy 1968:455).
The main erosional forces at work on Ofu
since the cessation of volcanism have been marine
wave attack and mass wasting, especially landslides. As noted above, an extensive sea cliff
more than 100 m high "was carved into the island
by the sea" (Stice and McCoy 1968:455) during
the late Pleistocene to early Holocene.. These high
cliffs, such as the one behind the To'aga site,
"originated by faulting and/or foundering," but
were certainly extended and modified by marine
erosion. Landslides continue to be active erosional forces, contributing to the talus screes that
border the inland edge of the coastal terrace.
"Individual blocks also work loose from the cliff
face and fall, fonning talus slopes that extend
almost continuously around the islands at the base
of these cliffs" (1968:455). One such massive
block fell between our 1987 and 1989 field
seasons: measuring at least 4 m in diameter, this
giant boulder crashed through the banana and
breadfruit orchards near the western edge of
To'aga to land by the roadside, a daily reminder of
the geological dynamism of Ofu. Similar large
blocks dot the surface of the To'aga site, and one
of these was modified in late prehistory for use as
an adz grinding stone (see Hunt, chapter 3). For
the prehistoric occupants of the To'aga site, such
falling boulders posed a continual hazard. As
Stice and McCoy relate, a local legend tells of a
young girl who "was killed by a large block that
rolled across the reef at Sili, where she was
fishing" (1968:456). We have no reason to doubt
the veracity of this story.
RESOURCES OF THE ISLAND
For
sive rootaso
arboricubw-.
b
taditionally
upon
inten-
bphoriculture, and on
s more than an abstract
The Natural and Cultural Enviromnent
Figure 2.2
13
View of the exposed volcanic plugs at Fa'ala'aga, from the beach at To'aga. The high mountain on the right
is the summit of Olosega Island.
constant, as rain is abundant and well distributed,
0
Figure 2.3
1
2m
Sketch of basalt dikes exposed in the road cut
across the Le'olo Ridge at Fa'ala'aga.
and as the islands are surrounded by the ocean, it
follows that the atmosphere is moist at all times
and seasons" (1930:17). Indeed, relative humidity
usually ranges between 80 and 86 percent, with the
temperature varying only between 25.7-26.20C.
Buxton also observed the effects of such humidity
on humans: "the perspiring entomologist soon
learns to recognize that his comfort depends
entirely on the wind, for the warm damp air does
not cool him unless it is in motion" (1930:18).
Green (1969:4) echoed these remarks from the
viewpoint of the field archaeologist, quoting Curry
(1962) to the effect that the Samoan climate is at
"the upper limit of thermal comfort for half-naked
men at rest."
concept. The Samoan archipelago lies within the
humid tropics. Buxton, who spent much time
studying the natural history of Samoa, summed up
the climate thusly: "As the temperature is nearly
There is, however, a distinct and significant
seasonality to the Samoan climate, expressed most
clearly in patterns of rainfall and of wind direction. We have no rain gauge records for Ofu itself,
14
The To'aga Site
but the records for Tutuila Airport (Nakamura
1984, table 1) are probably similar, indicating a
total annual average of 124 inches [3100 mm] (see
also Coulter 1941:10-11). (Buxton gives 2,738
mm [107.8 in] for Apia [1930:16].) There is a
distinctly drier season from about June through
September, in which the monthly rainfall averages
about 6 or 7 inches [175 mm], and a wet season
from October through May, with montly averages of from 11-14 inches [350 mm]. Even during
the "dry season," however, torrential downpours
may occur, as beset us all too frequently during the
1989 fieldwork at To'aga. As elsewhere in Western Polynesia, this seasonality played an important
role in scheduling the agricultural calendar,
particularly the clearing and planting of swidden
cultivations (Kirch 1978, forthcoming). In some
years, the dry season may be particularly acute or
lengthy, and the resulting drought can significantly
reduce agricultural yields. During the wet season,
the opposite may occur, with torrential rains and
serious flooding. "Some floods are associated
with hurricanes and tropical storms, but flooding
can occur at other times as well" (Nakamura
1984:3). The dry season also corresponds approximately to the period of prevailing trade winds
(from about April through September). During
October to March the winds are more variable, and
westerly reversals occur. This latter period was
important to early Polynesian voyagers, as it
allowed exploratory voyages from west to east
(Finney 1985; Irwin 1981, 1992).
In addition to the annual pattem of seasonality, there are stochastically recurring environmental hazards that seriously affect Samoan life. We
have already mentioned the problem of periodic
drought, which can inhibit agricultural production.
Even more severe are the hurricanes or tropical
cyclones that periodically lash the islands. Coulter
notes that "Samoa suffers hurricanes at irregular
intervals during the hotter season" (1941:12).
Visher (1925:27, table 6) indicates an average
annual frequency of two to three hurricanes in the
Samoan area, although not all of these are equally
intense, nor do their paths always cross the Manu'a
Group. When a hurricane bears directly down on
the islands, however, the effects are often devastating-to crops and orchards, to houses, and to
human life itself. The most recent hurricane to
lash Manu'a occurred early in 1987, between our
first and second field seasons. Ta'u Island was
particularly hard hit, with virtually every house
destroyed; the area was declared a Federal disaster
area. Flying in to Ta'u in June of 1987, one was
struck by the devestation of the forest cover of the
central volcanic cone which had not yet recovered.
Aside from their obvious significance to the
prehistoric Samoan population, hurrcanes have
considerable archaeological importance as agents
of landscape transformation, and of site formation
as well as destruction. The torrential rains unleashed during these events can cause severe
flooding and result in major landslides and in the
deposition of colluvium and alluvium. Storm
surges and high energy waves are capable of
moving large quantities of sand and larger clastics
(up to boulder size) in the coastal zone. During
our 1986 reconnaissance survey on Ta'u Island, we
observed a massive rampart of coral cobbles and
shingles at Saua which had been thrown up during
a previous storm surge. We shall give further
consideration to these high-energy processes
below, and elsewhere in this monograph (see
especially chapters 4 and 6), as they obviously
played a significant role in the geomorphological
history of the To'aga site.
Ofu soils are young and undeveloped, a
reflection of the island's geological youth. The
soils have been mapped and described by
Nakamura (1984). Most of the steep interior is
covered with "Ofu silty clay," a deep, well-drained
soil formed in volcanic materials. In the steeper
areas (slopes 40 to 70 percent) these silty clays are
covered in forest. Where slopes range between 15
and 40 percent, the Ofu silty clays provide the
main gardening soils. These gardening areas are
confined to two zones on the westem slopes, and
to one area on the northern slope of Tumutumu
Mountain (fig. 2.4). The very steep slopes and
talus regions lying inland of To'aga, and along the
northem coast, are described as "Fagasa familyLithic Hapludolls-Rock outcrop association, very
steep" (Nakamura 1984:1 1). At the inland edge of
the coastal terrace are found strips of "Aua very
stony silty clay loam" (1984:10), described as
"very deep, well drained soil ... on talus slopes . . .
formed in colluvium and alluvium derived dominantly from basic igneous rock." These areas
The Natural and Cultural Environment
15
5
;
0
'I
z
0
N-
a
0
-
I
*1
LI"
-
r44
R-
ct
:,31
N
0
I
4o
0D
ift
0
N.
13
I
E
CcmJl 0
0
0!
0
4c
CY
0
*4
C.
S
0
I
O-~
j4
z
16
The To'aga Site
are used for subsistence gardening, primarily of
tree crops such as breadfruit and bananas. Finally,
the coastal terrace itself, as at To'aga, consists
from the pedological viewpoint of "Ngedebus
mucky sand" (Nakamura 1984:15). This is a
"somewhat excessively drained soil ... derived
from coral and sea shells."
The natural vegetation of Manu'a has been
significantly modified in the lower elevations by
three millennia of human land use. The coastal
terrace, as well as the less precipitous mountain
slopes, comprise a mosaic of coconut stands,
breadfruit and banana orchards, and aroid gardens
interspersed with second growth (further discussion of coastal terrace vegetation below). On the
higher and steeper slopes, however, the original
ain forest vegetation persists. Yuncker, who
studied the flora of Manu'a, lists of total of 421
species for the three islands, including mosses and
pteridophytes as well as flowering plants (1945:4).
The Samoan root-tuber and tree crop complex, of
course, is wholly adventive to the archipelago,
having been introduced and established by the
early colonists.
The terrestrial fauna of the Manu'a Group is
very restricted in vertebrates, somewhat richer in
invertebrates (especially land molluscs and
insects). The only indigenous mammal is the fruit
bat, Pteropus samoensis. These are commonly
seen soaring over the forest canopy high above the
To'aga site, especially at dusk. The diminutive
Pacific rat, Ratus exulans, was introduced by
early Polynesian settlers, and its bones are common in the archaeological deposits at To'aga (see
Nagaoka, chapter 13). Also purposively introduced by the Polynesians were the domestic pig
(Sus scrofa) and dog (Canisfamiliaris).
The richest diversity of vertebrates is among
the birds, both native land birds and nesting
seabirds (Waading 1982). In the higher elevation
forests on Ofu are found the lupe or Pacific Pigeon
(Ducula pac4fica) and the Crimson-Crowned Fruit
Dove or manutagi (Ptilinopus porphyraceus); both
are occasionaUy taken for food. White-Collared
Kingfishers, ti'otala (Halcyon chloris manuae), of
which there is a distinct Manu'a subspecies, are
frequently perched high on poles or telephone wire
along the coast. Common in the coastal bush at
To'aga and elsewhere on the island is the ve'a or
Banded Rail (Gallirallus philippensis), often seen
making its characteristic headlong dash across the
dirt road from the security of one patch of undergrowth to another. The iao or Wattled Honeyeater
(Foulehaio carunculata) is abundant in the banana
groves at To'aga, where it feeds on banana flower
nectar. This bird is sometimes hunted by young
Samoan boys, who then pluck and roast the tiny
carcasses over a fire in the bush. (Having sampled
this delicacy myself, I can attest that while there is
little flesh, it is sweet and a delicious complement
to roasted bananas.) Also present on Ofu is the
Polynesian Starling, mitivao (Aplonis tabuensis
manuae).
A number of seabirds, and some migratory
species, also nest on Ofu. The White-Tailed
Tropic Bird, tava'e (Phaethon lepturus), nests in
the high cliffs towering over the To'aga site.
These elegant birds, soaring high ovetiead, were
always a wonderful visual diversion from the
perspective of a grimy test pit. A seasonal migrant, the Golden Plover or tuli (Pluvialis
dominicafulva) is also seen along the coast at
To'aga.
The contemporary avifauna of Manu'a is only
an impoverished remnant of the pre-human bird
life. Archaeological excavations throughout
Polynesia have revealed a significant pattem of
bird extinctions and extirpations due to humaninduced forest clearance and habitat destruction,
and to direct predation (Olson and James 1984;
Steadman 1989; Steadman, Pahlavan, and Kirch
1990; Steadman and Kirch 1990). As reported by
Steadman in chapter 14, the To'aga excavations
added further evidence of this widespread pattem
of avifaunal extinction and extirpation. These
faunal depletions are just one aspect of the dynamic nature of Polynesian ecosystems within the
span of human occupation.
The only other vertebrates indigenous to the
island are a number of lizard species in the families Geckonidae and Scincidae.
There is a rich endemic and indigenous insect
fauna, but this is of little archaeological relevance.
More important from the viewpoint of the prehistorian are the land snails, which include a number
of endemic and indigenous taxa, as well as several
species which have been introduced by humans,
prehistorically as well as after European contact.
The Natural and Cultural Environmen
Among the important families represented in
Samoa are Partulidae, Assimineidae, Tomatellinidae, Helicinidae, and Microcystinae. Land
snails frequently preserve well in archaeological
deposits, including those at To'aga, and are
excellent indicators of microenvironmental
change. Several synanthropic 'garden-snail'
species, including Lamellaxis gracillis, were
inadvertently transported around the Pacific basin
by prehistoric people, presumably with crop plants
and adhering soil. Their appearances in archaeological contexts are therefore important signals of
habitat modification, and indirect evidence for
horticulture. In chapter 8, I present an analysis of
land snails recovered from the To'aga excavations,
and their implications for environmental change at
the site.
Finally, we cannot ignore the marine environment, so crucial to the indigenous Samoan
economy. Ofu is surrounded by a fringing reef,
widest and most sheltered on the western side
(opposite Ofu Village). The reef is a complex
mosaic of micro-habitats and a source of shellfish
and fish for the human population. A diverse
array of molluscs occupy the reef flat and algal
crest, including various bivalves such as
Periglypta reticulata, Tridacna maxima, Hippopus
hippopus, and Asaphis violascens, and gastropods
such as Trochus maculatus, Turbo setosus, Nerita
spp., Cypraea spp., Drupa spp., Thais armigera,
and Conus spp. Many of these species were
heavily exploited by the occupants of the To'aga
site and occur in dense concentrations in the
midden deposits (see Nagaoka, chapter 13). Also
present on the reef are spiny lobsters (Panulirus
sp.), sea slugs (holothurians), sea urchins (echinoderms), octopus, and various edible seaweeds.
Approximately 800 species of inshore fishes occur
around Ofu (Jordan and Seale 1906) and are still
taken by the Samoans with a variety of traditional
fishing strategies, using spears, nets, hooks, and
other gear. Farrell's description of fishing activities in Westem Samoa is equally appropriate for
Ofu:
In the lagoon itself there is some activity at
almost any time, day or night. Hundreds of
yards from shore, near the edge of the reef,
women hunt for sea foods between the
breaking of the larger waves. Fishermen in
small canoes (paopao) equipped with goggles
17
and spear search the placid waters inshore,
while in shallower waters nets are tossed to
enshroud passing shoals of small fish.
Further out, midway between the shore and
the reef, fish traps are built of hard-riven,
close-spaced stakes and wire netting, or of
mounds of coral rock. Into these traps fish
are driven by 'beaters'; later the catch is
shared (1962:179).
Among the fish commonly taken are jacks
(Caranx spp.), parrot fish (Scarus spp.), wrasses
(Labridae), and acanthruids. Bones of these fishes
also occur in great frequency in the To'aga archaeological deposits. The open sea beyond the
reef is less heavily exploited but is the zone of the
prized tunas (Scombridae) and flying fish.
Marine turtles (Chelonia mydas and
Eretmochelys imbricata) are rarely sighted in the
waters off Ofu today but must have nested on the
island's sand beaches in substantial numbers prior
to early Polynesian settlement. The bones of these
turtles are one of the most commonly occurring
taxa in the To'aga faunal assemblages.
THE CULTURAL
AND SOCIAL LANDSCAPE
As in all Polynesian islands, the landscape of
Ofu is culturally and socially ordered. This
includes the system of land use, the pattem of land
tenure, and the village settlement pattem and its
intemal structure, all of which have evolved over
several millennia. One goal of our archaeological
investigations in Manu'a has been to contribute to
an understanding of how this distinctive cultural
and social landscape has developed over the
course of prehistory. A brief description of some
of the key aspects of the contemporary landscape
is therefore apposite, as an ethnographic reference
or 'endpoint.'
Farrell eloquently evokes the essence of the
Samoan settlement pattem of "villages ... strung
like beads unevenly along the thread of the
coastline" (1962:177). Mead elaborates: "The
cliff behind, the sea before it, defines the ground
plan of a Manuan village, which may spread out in
either direction as far as the land permits"
(1930:45). On Ofu today, there is only one such
village (with two named sectors, Ofu and Alaufau),
18
The To'aga Site
strung along the westem coastal terrace. TIhis village
lies conveniently between the widest and most
sheltered expanse of reef flat and lagoon (protected
by Nu'utele and Nuusilaelae Islets), and te largest
expanse of arable mountain slope inland. Previously, the zone of coastal villages extended around
most of the southem and southeastem parts of th
island, including the To'aga area (see Hunt, chapter
3). There is also limited evidence that some occupation may have extended inland into the intensive
gardening zone in late prehistorc times.
Although in the past decade or two most houses
on Ofu have been rebuilt using westem materialslargely concrete with corrugated roofs-the ground
plan of the village remains essentially traditional.
The central focus is the malae or "village green"
where important ceremonies and feasts may be held,
the church, and the guest house of the high chief
(which functions also as a council house for the
fono). The individual households, strips of land that
extend in principle from the beach to the mountain
slope, spread out on either side of fte sandy roadway. Traditionally, a household had thee main
structures: a guest house, the main dwelling, and fte
cookhouse (Handy and Handy 1924; Buck 1930:897). Mead aptly descrbed the Manu'an guest houses
"which stand by the sea, are round and high-ircles
of posts about four feet high topped by a twenty-foot
thathed circular cone. They stand upon a foundation of small stones which rises in slightly higher,
narrowing concentric circles, each terrace about five
or six inches high, and edged by larger stones"
(1930:46). The main dwelling or sleeping house was
traditionally the long house (fale o'o), with rounded
ends. The cookhouse or shed is thefale wnu,
literally house which shelters the wnu or earth oven
(Buck 1930:13). These are situated farthest from the
guest houses: "a small shack supported on four
pillars and roofed, not with sugar cane thatch but
with mats woven of palm leaves" (Mead 1930:48).
The layout of house types within the household
complex reflects a distinctive social and symbolic
structure to the village organization. The main axis
of orientation is perpendicular to the beach, the
interface of land and sea, and thus extends in two
directions: i tai, toward the sea, and i uta, inland.
The seaward direction is higher ranked, associated
with chiefs and persons of status; irnland is lower
ranked, associated with prduction and with the
economic basis of society. Mead describes the
seaward-irdand dichotomy thusly:
The term i tai (towards the sea) stands for the
optimum position; the village on the seashore,
the house on the sea side of the village, the place
of honor in the front of the house. And as the
trails lead back from the village over narrow
stils, through stony piaces, swamp pices, into
deep gulches, and up slippery inclines; so the
channel marks the way out to sea (1930:50).
This structural dichotomy between land and sea, and
the way in which it organizes the spatial stuctue of
the village, is a widespemad-and therefore probably
ancient-pattern within Westem Polynesia, including Fiji. Salins (1976:3745, fig. 6), for example,
has described this structure for Fiji, and in another
paper (1981) has outlined some of the cultural
associations between the sea and high-ranking
chiefs. The Polynesian Outlier of Tikopia (Firth
1936; Kirch and Yen 1982) likewise has a characteristic spatial organization very much like that described above for Samoa, with household units
differentiated along a seaward-landward axis, canoe
houses toward the beach, cookhouses inland, with
the main dwelling mediating between. (The
Tikopian dwelling itself is divided into seaward:male
and landward:female divisions; see Firth [1936].)
Similarly, on Niuatoputapu Lsland in northem
Tonga, the prehistoric coastal villages appear to
follow this pattem (Kirch 1988). And in Fuuna
Island, Burrows (1936) describes a village structure
very reminiscent of the Manu'an situation. In sum,
this kind of cultural and symbolic ordering of space
along a seaward-landward axis has a wide distribution in Westem Polynesia, and arguably has a deep
prehistory in the region as a "stucture of the long
run"'(Braudel 1980). Our tansect excavations at the
To'aga site provided some evidence of a coastal
village settement pattem ta may also have been
organized in this characteristic Westem Polynesian
model some 2,500 years ago.
TO'AGA: LAND USE
AND VEGETATION PATTERNS
While the fonmation of the coastal terrace at
To'aga was primarily a geomorphological phenomenon controlled by such extemal factors as sea level
change, ectonics, and their effect on sediment
budget (see Kirch, chapter 4), humans have also
The Natural and Cultal Environment
played an active role in the evolution of this landfonm. Inchatr5, for exampl, we s argue on
the basis of our igraphic data that the input of
teerial sediments onto the coastl terace increased after human colonization due to forest
clearance and agnicultral activities on the talus
slope and mounain. PeNaps the most obvious
effect of humans, however, results frm the pate s
of land use and vegetaion on the coastal terrace, for
the whole zone constitutes an an ogenic,
managed environment from the phytogeographic
percive. The coastal terrace of Ofu provides the
only flat land on the island, the setting for both
villages and for certain Inds of intsive horticultura and aiboncultual production. The development of this pattem of intensive land use is one
prbem that we have attempted to addrss in our
archaeological investigations at To'aga As backgound to this study, it is necessary to charactenize
te pattens of land use and vegetaton found on the
Toaga coastal terrace today.
Vegetion Transects
Yuncker's (1945) study of the Manu'an flora
enumerated the plant sources ofthe islands, but
neglected patterns ofplant distribution. In 1987 and
1989 we reoorded the horizontal distribution of
dominant species along three of our trasc across
the To'aga coastal terce that had been cleared for
archaeological subsurface sampling (see Kirch and
Hunt, chapter S for futer details of these trnsets).
One of xse, Trnsect 7, is graphically depicted in
figure 25. Although there are minor differences in
the distribution of species, the same overll pattem is
evident in all trsec . Whisler (1980) described
and ilustrated most of the species recorded in these
tran s. Three main vegetation zones can be
discerned, from seaward to ladward:
1. Strand Vegetation. Beginning at the highwater maik and extending to the seaward edge of the
coastal road is a zone dominated by halophytic,
liUttoral species. Ovehganing the beach are Scaevola
taccada and Messerschmidia argentea, with the
vines Canalia maritima and Jpomoea pes-caprae
trailing over the sands. Larger trees sunnounting the
beach ridge inlude Barringtonia asiatica (traditonally used for fish poisoning), Hernandia
nyphae4folia, Cocos nucifera, and Terminala
samoensis. Pandaiws tectorius and Hibiscus
19
tiliaceous shrubs dominate the iner edge of this
zone, lning the sandy radway.
2. Arborlulural-Horticultural Zone.
Commencing on the inland side of the road, and
extending acrmss the width of the coastal terrace to
the base ofthe talus slope is the main zone of
economic plants. Coconut palms dot the area,
especially toward the seaward halfofthe zone,
where the soil is sandier. The badfruit tree,
Artocarpus altilis, commences not far frm the
roadway and is the main upper story dominant
across the terace. Under and between the beadfruit
and coconuts are planted a number of fruit and root
crops, the most important being Ewnusa banas
and the large arid, Alocasia macrrorhiza. Some
taro, Colocasia esculent and the historically
inrduced American amid Xanthosoma
saggitolia, are also found, although in lower
frequency. The Alocasia aroids are often densely
planted in clearings which, after cropping, ae
secondarily planted in bananas. In sone areas, the
undertory beneath the coconut palms and badfruit
trees is a tangle of second growth shubs, dominated
by Hibiscus tiliaceous and Macaranga st4ulosa.
Othr useful trees occuring less frequently though
the zone include mosooi (Cananga odorata), the
flowers of which are used for scening coconut oil,
andfisoa (Colubrina asiatica), which has medicinal
value and can be used as a soap substitute (Whistler
1980:41).
This zone of tree, fruit, and root crps also
exhibits a high frequency of feral or natralized
species which are commonly cultivated in Oceanic
agricultural systems, and which unquestionably are
pesent in the Toaga area as suvivals frm an eailier
phase of (presumably) more intensive culdvation.
These include the t plant, Cordylinefruicosum, the
arrowrot, Tacca leontopetaloides, the bitter yam,
Dioscorea budbfera (which twines in gat abundance over fte trees and shrubs of this zone), and in
lesser quantities, the a'a or Pueraria lobata. All of
tese plants are mcognized by the Ofu people as
having edible subternean roots or tubers and are
regarded as potential famine resources. They are
part of the complex of plarts a Barru (1965) has
termed "witnss of the past," indicators of eatlier
cultivation praices in the Pacfic islands.
3. Talus Slope Vegetation. The main zone of
economic plants terminates abrupdy at the base of
e talus slope, strewn with large volcanic boulders.
The To'aga Site
20
4-4
Alocasia macrorrhiza
i
Colubrina asiatica
Morinda citrifolia
-4
Cordy/ine fruticosum
Artocarpus a/tilis
Macaranga sp.
i
4I
Tacca leontopetal/odes
4~ ~ ~ ~ 4-Dioscorea bulbifera
Musa hybrids
Hibiscus tiliaceous
Cocos nucifera
Pandanus tectorius
Barringtonia asiatica
@Hemandia nymphaeifo/ia
e
MassarsMhmida arnantna
Scaevola tacc4ada
8 *
0
-
4
--
v
-4
.
10 *\
#
i
D. ARBORICULTURAL ZONE
fi
Figure 2.5
Distribution of major floral dominants along Transect 7 at To'aga
A few breadfruit trees and the occasional banana
plant may be found extending a few meters up the
slope, but the main dominant here is Hibicus
tiliaceous, which fonns a dense tangle over the
boulders, and larger forest trees such as Erythrina
varigata and others.
Land Use and Site Formation
Although no longer a locus of pennanent
village habitation, the To'aga coastal terrace
remains an important zone of intensive horticulture and arboriculture for the Ofu Island population, as indicated by the analysis of vegetation
patterns. Indeed, on an island where flat land
comprises less than nine percent of the total area,
the economic importance of this coastal terrace to
the human population cannot be overemphasized.
It is certain that the pattem of intensive cultivation
of this area extended back at least into late prehistory, but it remains an archaeological problem to
deternine just when this pattem first developed.
The pattem of land use in prehistory must also
have had consequences for archaeological site
formation processes. For example, the clearance
of indigenous forest cover on the talus slope above
the flat would have exposed unstable soil and
rock, and thus accelerated erosion and deposition
of colluvial sediment onto the coastal terrace.
Cultivation on the flat itself would have resulted in
a continual reworking of the upper soil layer
(through the actions of digging sticks as well as
The Natural and Cultural Environmen
thrugh floral-turbation by plant roots and tubers).
The mixing of terruginous and calcareous sediments thugh cultivation would have created a
well-drained, highly fertile edaphic medium which
itself was probably more suitable for root crop
cultivation than either the heavy colluvial clays, or
the calcareous sands themselves (see Kirch and
Yen 1982).
In the chapters to follow, we pay particular
attention to several lines of evidence that point to
the gradual, historical development of the intensive land use pattem at To'aga. These lines
include evidence for changes in the rate of deposition of colluvial sediments, stratigraphic evidence
of buried soil surfaces, evidence of reworking of
soils, and the presence of several species of
synantrpic land snails that are markers of
Oceanic horticultural activities.
REFERENCES CITED
Barrau, J. 1965. Histoire et prdhistoire horticoles
de l'Ocdanie tropical. Journal de la Sociite
des Ocdanistes 21:55-78.
Braudel, F. 1980. On History. Chicago: University of Chicago Press.
Buck, P. 1930. Samoan Material Culture. Honolulu: Bernice P. Bishop Museum Bulletin 75.
Burrows, E. 1936. Ethnology of Futuna. Honolulu:
Bernice P. Bishop Museum Bulletin 138.
Buxton, P. A. 1930. Description of the environment. Insects of Samoa, Part IX, Fasc. 1.
London: The British Museum.
Coulter, J. W. 1941. Land Utilization in American
Samoa. Bemice P. Bishop Museum Bulletin
170. Honolulu.
Curry, L. 1962. Weather and climate. IN J. W. Fox
and K. B. Cumberland, eds., Western Samoa:
Land, Life, and Agriculture in Tropical
Polynesia, pp. 48-62. Christchurch: Whitcombe and Tombs.
Daly, R. A. 1924. The geology of American
Samoa. Carnegie Institution Publicaton
340:93-152.
Farrell, B. H. 1962. The village and its agrculture.
IN J. W. Fox and K. B. Cumberland, eds.,
Western Samoa: Land, Life, and Agriculture
in Tropical Polynesia, pp. 177-238.
Christchurch: Whitcombe and Tombs.
Finney, B. 1985. Anomalous westerlies, El Niflo,
21
and the colonization of Polynesia. American
Anthropologist 87:9-26.
Firth, R. 1936. We The Tikopia. London: George
Allen and Unwin.
Flenley, J., and M. King 1984. Late quatemary
pollen records from Easter Island. Nature
307:47-50.
Fosberg, R. 1963. Disturbance in island ecosystems. IN J. L. Gressitt, ed., Pacific Basin
Biogeography, pp. 557-61. Honolulu: Bishop
Museum Press.
Friedlander, I. 1910. Beitrage zur der Samoa
Inseln. K. Bayer. Akad. Wiss., Math-Phys.
El., Munich, Band 27:358-69.
Green, R. C. 1969. Archaeological investigation of
Westem Samoan prehistory. IN R. C. Green
and J. Davidson, eds., Archaeology in
Western Samoa, Vol. 1, pp. 3-11. Bulletin of
the Auckland Institute and Museum,
Auckland.
Handy, E. S. C., and W. C. Handy 1924. Samoan
House Building, Cooking, and Tattooing.
Bemice P. Bishop Museum Bulletin 15.
Honolulu.
Hiroa, Te Rangi (P. H. Buck) 1930. Samoan
Material Culture. Honolulu: Bemice P.
Bishop Museum Bulletin 75.
Irwin, G. 1981. How Lapita lost its pots. Journal
of the Polynesian Society 90:481-94.
1992. The Prehistoric Exploration and
colonisation of the Pacific. Cambridge:
Cambridge University Press.
Jordan, D. S., and A. Seale 1906. The Fishes of
Samoa. Bureau of Fisheres Bulletin 25, pp.
173455. Washington D.C.: Govemnment
Printing Office.
Kirch, P. V. 1978. Indigenous agriculture on 'Uvea
(Westem Polynesia). Economic Botany
32: 157-81.
1982. The impact of the prehistoric
Polynesians on the Hawaiian ecosystem.
Pacific Science 36:1-14.
1983. Man's role in modifying tropical and
subtropical Polynesian ecosystems. Archaeology in Oceania 18:26-3 1.
1984. The Evolution of the Polynesian
Chiefdoms. Cambridge: Cambridge University Press.
1988. Niuatoputapu: The Prehistory of a
22
The To'aga Site
Polynesian Chiefdom. Thomas Burke Memorial Washington State Museum Monograph
No. 5. Seattle: Burke Museum.
. forthcoming. The Wet and the Dry: Irrigation
and Agricultural Intensification in Polynesia.
Chicago: University of Chicago Press.
Kirch, P. V., and D. E. Yen 1982. Tikopia: The
Prehistory and Ecology of a Polynesian
Outlier. Honolulu: Bemice P. Bishop Museum Bulletin 238.
McDougall, I. 1985. Age and evolution of the
volcanoes of Tutuila, American Samoa.
Pacific Science 39:311-20.
McGlone, M. 1983. The Polynesian deforestation
of New Zealand: A preliminary synthesis.
Archaeology in Oceania 18:11-25.
Mead, M. 1930. Social Organization of Manua.
Honolulu: Bernice P. Bishop Museum
Bulletin 76.
Menard, H. W. 1986. Islands. New York: Scientific American Library.
Nakamura, S. 1984. Soil Survey ofAmerican
Samoa. Soil Conservation Service, U.S.
Deparunent of Agriculture. Washington
D.C.: Govermnent Printing Office.
Natland, J. H. 1980. The progression of volcanism
in the Samoan linear volcanic chain. American Journal of Science 280-A:709-35.
Olson, S., and H. James 1984. The role of
Polynesians in the extinction of the avifauna
of the Hawaiian Islands. IN P. S. Martin and
R. G. Kline, eds., Quaternary Extinctions, pp.
768-80. Tucson: University of Arizona Press.
Sahlins, M. D. 1976. Culture and Practical
Reason. Chicago: University of Chicago
Press.
. 1981. The stranger-king, or Dumezil among
the Fijians. Journal of Pacific History
16:107-32.
Steadman, D. 1989. Extinction of birds in Eastem
Polynesia: A review of the record, and
comparisons with other island groups.
Journal of Archaeological Science 16:177205.
Steadman, D. W., D. S. Pahlavan, and P. V. Kirch
1990. Extinction, biogeography, and human
exploitation of birds on Tikopia and Anuta,
Polynesian outliers in the Solomon Islands.
Bishop Museum Occasional Papers 30:11853.
Steadman, D. W., and P. V. Kirch 1990. Prehistoric extinction of birds on Mangaia, Cook
Islands, Polynesia. Proceedings of the
National Academy ofSciences U.S.A.
87:9605-9609.
Steams, H. T. 1944. Geology of the Samoan
Islands. Bulletin of the Geological Society of
America 55:1279-1332.
Stice, G. D., and F. W. McCoy, Jr. 1968. The
geology of the Manu'a Islands, Samoa.
Pacific Science 22:427-57.
Visher, S. S. 1925. Tropical Cyclones of the
Pacific. Bernice P. Bishop Museum Bulletin
20. Honolulu.
Wading, D. 1982. Birds of Fiji, Tonga, and
Samoa. Wellington: Millwood Press.
Whistler, W. A. 1980. Coastal Flowers of the
Tropical Pacific. Honolulu: Oriental Publishing Company.
Yuncker, T. G. 1945. Plants of the Manua Islands.
Honolulu: Bernice P. Bishop Museum
Bulletin 184.
3
SURFACE ARCHAEOLOGICAL
FEATURES OF TO'AGA
T. L. HuNT
boundaries rather an by the relative density of
~T'HE SAmoAN LANscAPE has been shaped by a
long history ofhuman setlement and resource
arifacts or architecturl remains. Thus, Site AS-13-1
exploitadon. Such evidence of settlement and land
comprises the area from the shoreline to the steep
use, both past and present, is distributed more-or-less slopes and cliffs, the coastal ter-ace described in
continuously across the island landscapes of Samoa.
detail in chapter 2.
Isolated arfifacts, feaus, sites, and site complexes
cover much of the Samoan Islands and vary signifiField Methods
cantly in relative density (e.g., Davidson 1974:242).
A systematic survey of the surface features at
The continuous nature of archaeological remains,
To'aga was accomplished using tanscts set frm the
often expressed in terms of "non-site archaeology"
beach to the steep slopes of the interior cliffs. In 1986
requires documenting the distributional pattems of
material culture in space (Thomas 1974; Dunnell and a baseline was set out along the coastal road with
Dancey 1983). This is in contrast to isolating "sites" lateral trans about 30 meters in width made at
as dense clusters and ignoring the lower density
every 100-meter interval. This resulted in 18 survey
distributions that infonn upon a varety of peistoric transects that covered approximately 4.4 hectares of
activities and their spadal distribution. The "non-site" the coastal terrace. Shrub vegetation, often dense,
was cleared from these transects. Without extensive
approach has proven useful elsewhere in Oceania
clearing of ground cover and other vegetation, some
where the vestige of entire settlement-subsistence
areas had very poor surface visibility. This means
patterns have been documented (e.g., Green 1980;
surface features were undoubtedly missed, even in
Kirch and Yen 1982; Weisler and Kirch 1985).
transects cleared of some vegetation.
Ideally, entire archaeological lamdscapes (spatial
AU surface feaues and artifacts discovered on
pattems) must be recorded in detail, although this
the
were recorded for provenience, deunsects
ultimate objective must usually be met though the
gradual accumulation of survey data over many years scrbed, and many were also mapped using a tape and
compass. Other features encountered during fieldof effort
work (in all three seasons) were wecorded in the same
This chapter describes the surface features
way.
Fmally, well-known features or monumental
recorded along systematic tansects on the coastal
constructions were located with local guides and
tenrace of To'aga (Site AS-13-1). As elsewhere in
recorded as well.
Manu'a, the "site" was defined by natural landscape
24
The To'aga Site
Trnsects oriented perpendicular to the
shoreline traverse the greatest range of geomorphic diversity that may correlate with variation in
settlement, land use, and age of surface features.
Surface data acquired along systematic transects
can be used to produce maps (with interpolaffons
as desired) that reveal pattem and structure in the
archaeological remains. In short, this kind of
systematic strategy provides a means to record, in
some detail, the spatial configuration and variety
of archaeological remains while producing an
informative, albeit partial, picture of the broader
archaeological landscape.
Results
Much of the coastal land of the south coast of
Ofu Island, from the areas known properly as
To'aga to Fa'ala'aga, appears to have a near
continuous distribution of archaeological remains
on the surface. In most areas where vegetation
pennitted visibility of the ground surface, evidence of habitation and land use was present.
Figure 3.1 shows the location of archaeological
surface features recorded in the survey transects.
These features, shown by consecutive numbers,
are listed and described in table 3.1.
The surface features recorded include basalt
and coral boulder alignments usually of curved
and oval form (house foundations) with coral and
basalt pebble paving ('il'ili, e.g., figure 3.2); small
oval or rectangular boulder alignments that appear
to mark burials or are simply the partial remnants
of house foundations; a massive, fine-grained talus
boulder used for multiple (twelve) grinding basins/
surfaces (figure 3.3); pits with associated boulder
slab lining in several places, probably the remains
from the production of fennented breadfruit (lua'i
mast); and long (inland-seaward oriented) single
course alignments which appear to have served
(and may continue to serve) as land boundary
markers.
Particularly noteworthy is a complex that
includes the locally well-known Tui Ofu Well
(Feature 23, figure 3.4) and Tui Ofu Tia (monumental tomb, Feature 24, figure 3.5). Situated
near the eastem end of the coastal flat of southem
Ofu (near Fa'ala'aga), these monumental constnrctions are traditionally associated with the highranking title Tui Ofu ("king of Ofu"). The Tui
Ofu Well is a relatively elaborate construction of
waterwom basalt boulders arranged in a rounded
form that includes a sloped concourse as well as a
small paved court around the excavated shaft. The
depth of the shaft from the sunrounding court is 2.1
m. The Tui Ofu Tia (tomb) comprises a crudely
terraced mound of basalt boulders (both rough and
waterwom) that is set in against the massive talus
and steep slope rising on the inland side of this
area. A small pit lies among the boulder rubble on
the uppermost terrace of the mound. No artifacts
or other cultural materials were observed at the
structure, except that omamental plants, such as
crotons, have been planted around the base of the
mound.
The survey along transects has yielded data on
only part of the surface record at To'aga. Other
features will be found with more intensive coverage of the area. A more complete survey of
surface remains at To'aga will require extensive
clearing of vegetation which restricts access to
parts of the coastal lands or obscures visibility.
The geomorphic evidence (described in
chapters 2, 4 and 5) suggests that the present
surface at To'aga is less than 1000 year B.P., and
very recent in some areas where colluvial and
eolian (i.e., primarily calcareous beach sediment)
deposition continues. Thus the surface evidence
represents late prehistoric and historic activities on
the coastal terrace. This observation is supported
by excavation and geomorphic data (Kirch and
Hunt, chapter 5). Many of the surface features are
probably pene-contemporaneous, at least in
archaeological time-frames.
The surface features at To'aga are similar in
their formal variability to those documented
elsewhere in Manu'a (Hunt and Kirch 1988) and
other islands of Samoa (e.g., Clark and Herdrch
1988; Davidson 1974; Green and Davidson 1969,
1974; Jennings et al. 1976, 1982; Jennings and
Holmer 1980). The architectural remains present
at To'aga do not reflect the full range known for
Swface Archaeological Features
Map of the southrn portion of Ofu Island, from To'aga to Fa'aia'aga, showing the location of
surface archaeological features (1-26) of site AS-13-1.
Figure 3.1
0 so
1
2m
AA
A
Figure 3.2
25
Plan of round-ended stone house foundation (Feature 19); the shaded area is paved with coral and
basalt gravel ('ili'ili).
The To'aga Site
26
53sq
*
19
)
WI)
6
(o
s
1-i
V-4
.1
x
Q
m
C-i
C>
In
S
r.
S
v)
o
O;
Q
0
)
_
_
esn
"
n
Q
o
'.4
10
s
C)
S 5
P-400
0. (:.mm
9
x
V-4
P-4
x
lqt
Cq
C. x x
63 C)04aC>
tn tnmW)0
J:
46(7A
Aen
V-4
v-4
0
W-4NW
S i S K iS
C = = 2 ai 7i °
c
V-4
s
Cf)
qqt
cc;
Q
xo
.o
t-
co
co
el
0~~~~~~~~~~~~~~~~~~~~~~~~~0
Eo
0~~~~~~~~~~~~~~~~0
co
>
] :t;;88r8r
6~~~~~~~~~~~~~~~c
en
ut
°E6Ew
wt
To
v
:s
I)
Q
*8
.
-
Po
P
0I
.
wosoo
_
_
= 3
0
ItC
:
m(mA
*_&
a
c
go
_
m_
cb
5
go
_
,
3
°
__t_
...
_
I
_
^__~~~~~~~.0.
aca£a.'O .*m
.*m.*m*m1S
...
a
m
ms m
m
m mm
m ON
__
__e
-
N
N
m
im Wm m
N
(4
N
N
Surface Archaeological Featwres
27
_~~~~~~~~~~
_~~~~~~~~~~~~~~~~~b
4)_
4)
0E
4)
0
4)
oU
_
j~ ~ ~ ~ ~ ~ ~ ~ ~ ~ ~ ~ ~ ~ ~ ~ ~ ~
4)
:U
4);
28
The To'aga Site
A
A
A'
r,.
Figure 3.4
Plan and cross section of the Tui Ofu well at Muli'ulu.
Surface Archaeological Features
A-
29
-A'
2r
A
n A'
Figure 3.5 Plan and cross
section of the the Tui Ofu Tia, stone mound/tomb at Muli'ulu.
30
The To'aga Site
Samoa but do include forns linked to domestic
activities. Only the Tui Ofu Well and Tui Ofu Tia
represent structures of specialized function, indeed
those associated with social rank.
The available evidence suggests a redundant
pattem of associated features which can be interpreted as the remains of domestic compounds.
The distribution of such habitation suggests
dispersed settlement organized on the basis of
descent groups, as represented in Samoan socioeconomic organization documented ethnographically (Mead 1930).
The distribution of structural features also
reveals late prehistoric and historic settlement
situated primarily on the stabilized dune ridge.
The surface and excavation evidence (Kirch and
Hunt, chapter 5) suggests that occupation has
centered on the high ground of stabilized sand
dune ridges. As progradation of the To'aga coastal
land occurred, settlements moved seaward. The
archaeological evidence on today's relatively
young surface reflects the last phase of habitation
at To'aga. Thus, the setdement pattem marks
continuity over To'aga's pre- and post-contact
history.
The primary focus of fieldwork at To'aga has
been the definition and sampling of the subsurface
deposits of the coastal terrace. Additional intensive surface survey is necessary to complete the
picture that has emerged so far.
REFERENCES CITED
Clark, J. T., and D. J. Herdrich 1988. The Eastern
Tutuila archaeological project: 1986 final
report Unpublished manuscript, Office of
Historic Preservation, Pago Pago.
Dunnell, R. C., and W. S. Dancey 1983. The
siteless survey: A regional scale data collec-
tion strategy. Advances in Archaeological
Method and Theory 6:267-87.
Davidson, J. M. 1974. Samoan structural remains
and settlement patterns. IN R. C. Green and J.
M. Davidson, eds., Archaeology in Western Samoa,
Vol. 1. Auckland Institute and Museum Bulletin 6.
Green, R. C., and J. M. Davidson, eds. 1969.
Archaeology in Western Samoa, Vol. 1.
Auckland Institute and Museum Bulletin 6.
Green, R. C., and J. M. Davidson, eds. 1974.
Archaeology in Western Samoa, Vol. 2.
Auckland Institute and Museum Bulletin 7.
Hunt, T. L., and P. V. Kirch 1988. An archaeological survey of the Manu'a Islands, American
Samoa. Journal of the Polynesian Society
97:153-83.
Jennings, J. D., R. N. Holmer, J. C. Janetski, and
H. L. Smith 1976. Excavations on Upolu,
Western Samoa. Pacific Anthropological
Records 25. Honolulu: Bernice P. Bishop
Museum.
Jennings, J. D., and R. N. Holmer, eds. 1980.
Archaeological Excavations in Western
Samoa. Pacific Anthropological Records 32.
Honolulu: Bernice P. Bishop Museum.
Jennings, J. D., R. N. Holmer, and G. Jackmond
1982. Samoan village pattems: Four examples.
Journal of the Polynesian Society 91:81-102.
Kirch, P. V., and D. E. Yen 1982. Tikopia: The
Prehistory and Ecology of a Polynesian
Outlier. Bernice P. Bishop Museum Bulletin
238. Honolulu.
Mead, M. 1930. Social Organization ofManua.
Bernice P. Bishop Museum Bulletin 76.
Honolulu.
Thomas, D. H. 1974. Predicting the Past. New
York: Holt, Reinhart, and Winston.
Weisler, M., and P. V. Kirch 1985. The structure
of settlement space in a Polynesian chiefdom:
Kawela, Moloka'i, Hawaiian Islands. New
Zealand Journal of Archaeology 7:129-58.
4
THE TO'AGA SITE:
MODELUNG THE MORPHODYNAMICS
OF THE LAND-SEA INTERFACE
PATRICK V. KIRCH
IN CHAP 2 wE described the economic and
cultual significance of the narrow coastal terrace
of Ofu which-a~the only flat land on the island-is
thus the main locus of settlement and a major zone of
subsistence prduction. This coastal tenrace, which
the To'aga site exemplifies well, must be undertood
as a dynamic geomorphological entity lying at the
cdtical land-sea interface. McLean has aptly described the imporance of such land-sea interfacial
zones on small tropical islands:
The atactiveness of the coastal lowlands for
setlement purposes is obvious.... On many
islands they comprise dte only flat country. It is
easy to move over and easy to build on. Soils,
though naurally of low fertlity and moise
statis, are easily worked. The sandy subsate,
beach sand and beachrock are readily available
for such uses as building matials, graves and
road 'metal.' That the coastal flats are multipurpose and multi-utilized resources goes
without saying. Moreover, their location at the
land-sea edge permits optimal access to a
complete range of terrestrial and imaine
environments and rsources: hilislope, valley
and forest; reef, lagoon and fish. They function
as convenient bases to exploit the local environs.
They also function as verandahs to the worlds
outside (1980:.129-30).
In their study of the prehistory and ecology of
Tikopia, Kirch and Yen (1982) similarly argued for
the importance of unraveling the morphodynamic
histories of tese interfacial enviromnents:
among these [environments] are the coastal/
marine interface, with implications for access to
rsources of reef and sea, and that of calcareous
plains and volcanic hills, with implications for
erosion and agricultural developmenL ... Both
of these interfaces have been active zones of
landscape change, with long-tern implications
for human adaptation (1982:17).
In particular, the Tikopia case revealed the agricultural significance of human-induced modifications to
the coastal zone: 'te plain and hill interface pmvides the opporunity for the mixing of volcanic and
calcareous soils to form edaphic media, more favorable for the cultivation of root crops than either
separate type" (Kirch and Yen 1982:17).
Prior experence in Tikopia (Kirch and Yen
1982), Niuatoputapu (Kirch 1988), and other small
tropical islands had sensitized us to the necessity of a
geomorphologically informed approach to the
archaeological study of the To'aga site. This was
critical both to an undemanding of depositional and
site-formation processes, as well as for contribudng to
our knowledge of the evolution of the island's larger
The To'aga Site
32
settlement paterns and production systems. This
chapter focuses on several aspects of the local
envinnt at To'aga which played key roles in te
morphodynamic evolution of the site. Chief anong
these are the ging configuration of sea level in
the mid-Holocene, and the tectonic instability of the
Manu'a Group. Equally sigficant, however, is the
role of humans tiemselves in shaping landscape
through physical manipulation of soil and rock,
forwst clearance, the inroduction of a "portnantea"
biota (cf. Crosby 1986), and the continual manipulation of the coastal enviro
to best suit human
needs. My aim here is to outline a model of these
morphodynamic processes that may then be tsted
agains the empirical evidence revealed thrugh our
field and laboratory investigations.
GEOMORPHOLOGY OF THE TO'AGA
COASTAL TERRACE
The key geomorphological features of the
coastal terrace at To'aga can best be descnbed along
a generalized transe ruing from the reef inland
across the flat up to the steep hillslope and cliff as in
figure 4.1. (Particular, surveyed transects are
illustraed and described in chapter 5, along with the
statigraphic evidence for temporal development) At
the seaward end is the reef flat, source of te calcarous sediment (composed of the detritus of coral
molluscs, and other reef-dwelling organisms) of
which the coastal terrace is primarily constructed.
This reef flat is a mosaic of microenvirornents,
including surge channels, the 'lithotamrnium ridge'
at the seaward edge, large coral heads, and sandy
patches.
A careful exanination of the beach slope reveals
several clues as to the dynamic processes presently
at wor along the land-sea interface. At the foot of
the beach slope in various places are exposed layers
of 'beach rock' (Wiens 1962:64-67) made up of sand
and coral rubble cemented with calcium carbonate
(CaC03). Such beach rock deposits can fonn fairly
quickly under tidal conditions of continual wetting
and drying witfin active beach ridges. Their
exposure along the To'aga beach front, however,
signals a cument phase of coastal erosion and net
sediment loss (or lateral transfer). This evidence is
reinforced by an examination of the vegetation line
at the top of the present beach ridge crest, where
large trees (such as Cocos, Cordia, and Hernandia)
have been undercut and eroded by wave action.
Moving inland, one mounts the crest of the
present beach ridge or benn, lying between the
beach slope and the sandy road. This nanrow zone is
covered with a thick tangle of vegetation, dominated
by Pandanus tectorius, Barringtonia asiatica,
Hernandia nymphaeifolia, Scaevola taccada,
Messerschmidia argentia, Cocos nucifera, and other
species that anchor the loose, unconsolidated
calcareous sands. Test pits dug into this beach ridge
revealed only a thin (ca. 5 cm) organic A horizon
directly overlying the young parent sands. The
absence of archaeological deposits suggests that the
present beach ridge is of no great antiquity.
Crossing the road and plunging thrugh a
narrow band of Pandanus and Hibiscus tliaceous
shrubs, one enters the main extent of the coastal
tenrace, a zone of intensive economic utilization
though arboriculture and root crop gardening (see
Kirch, chapter 2). The ground at first slopes down
slightly from the beach ridge, levels out, and then
begins to rise again gradually toward te talus slopes
inland. As one moves inland from the road through
this zone, the soil gradually becomes less sandy and
more clayey, reflecting the increased contribution of
volcanic colluvium which has been eroded frxm the
hillslopes and added to the calcareous sands that
form the main component of the substrate. The
mixing of tese calcareous and volcanic sedimentsthrogh the continual action of human cultivationhas made this zone of particular edaphic value for
Samoan horficulture. This is also the main zone of
archaeological feaures, both surface and subsurface, as described in chapters 3 and 5.
Moving farter irland toward the talus slope,
one begins to note volcanic cobbles and boulders
stewn over the land surface, signals of the dynamic
instability of the 500-m-high cliffs that tower over
the To'aga flaL Some of these boulders are several
meters in diameter and can fall with considerable
destructive force. [be ground surface closest to the
talus is entirely gravelley clay-loam with no calcareous component evident Rather abruptly, one arrives
at the base ofthe talus itself, an imposing and
unstable jumble ofboulders rising steeply toward the
cliffs, over which grows a tangle of Hibiscus
tiliaceous and scattered forest trees such as Erythrina
varigaa.
In sum, this geomorphological traverse across
Morphodynamics of the Land-Sea Inerface
STRAND VEGETATION
ARBORICULTUR4L ZONE
/
C_lluvium
33
Erosion Scarp
i
::
.:
: :
:: .::
Unconsolidated Calcareous Sands
>,Exposed Beach Rock
:.
Coral Reef Platform-
Figure 4.1
AgeneralizedtasectacrossthecoastalplainatTo'aga,showingthedistributionofcolluvialandcacareous
sediments, major vegetation associations, and geonnorphological featu .
To'aga coastal terrace reveals several key
feaues: (1) the terrace is constructed primarily of
marine biogenic sediments (calcareous sands and
larger clastics); (2) one finds a progressively greater
contribudon of terrigenous sediments as one approaches the volcanic mass; and (3) there is some
evidence for curnt coastal erosion and maine
transgression at the present time. These are features
that must be accounted for in any model of the
morphodynamics of the To'aga site over the past
several thousand years.
the
MORPHODYNAMIC PROCESSES:
SEA LEVELS AND SUBSIDENCE
As the archaeological deposits of the To'aga site
are integral sedimentary components of the coastal
terrace, any model of site formation processes must
first account for the geomorphological processes by
which the terrace was formed. Lying at the critical
and highly dynamic land-sea interface, any consideration of the morphodynamics of this zone must
begin with exanination of controlling processes
in shoreline formation. Among the most important
of these are: (1) glacio-eustatic sea-level changes;
and, (2) the local tectonic situation which-as we
an
fshall argue-is one of subsidence. The temporal
penod with which we are concemed is the mid- to
late-Holocene, from about 6 kyr B.P. to fte preet
This time span covers some 3 kyr piior to the human
colonization of Ofu as well as the subsequent
interval during which humans have added their input
to the development of the local lanscape.
Before examining the evidence for reladve sealevel change during this period, it is essential to
inrduce a few key concepts regarding coastal
change on small islands. In this I have drawn
primarily from Chappell (1982) and McLean
(1980b). The construction of a coastal tenrace such
as that at To'aga results from progradaion, the
"progressive formation of new land by sedimentation inespective of fte tendency of sea level movement" (Chappell 1982:71). Sea-level changes
themselves, whether due to glacio-eustacy or
tectonic movements, or both, are inportant as
controlling factors for the sediment budget, but alone
ftey do not provide a sufficient model of prgradation. As Chappell emphasizes, ". sea level
changes alone cannot be used to account for coastal
changes. In fact, for the last 6000 years, the sedimentary budget is the more important factor'
(1982:71). The sediment budget can be thought of
as te net sum of sediment input from both terrestrial
.
.
34
The To'aga Site
(talus, colluvium, etc.) and marine biogenic (calcareous sands and coral detritus) sources, minus the loss
of sediment from transport (fig. 4.2). In modelling
the Toaga coastal terrace formation, therefore, we
need to pay particular atention to anges that
would either increase or decrease the production of
sediment from te ial or marine sources.
Figure 4.3 graphicaly portrays the dynamic
effects of dcanges in relative sea level (due either to
glacio-eustatic or tectonic change), combined with
increases or decreases in sediment budget. In this
diagram, sediment budget is indicated along the x
axis, and relative sea-level change along the y axis.
The heavy diagonal line separates transgression or
coastal rera, frm regression or coastal advance.
Holocene Sea Levels in dhe South Pacific
A rapid rise in sea levels following the end of
the Pleistocene is a global phenomenon that has been
widely rognized (Fairbridge 1961; Shepard 1963).
More controversial-because they depend upon a
complexity of local conditons and processes-have
been the details of the eustatic sea-level curve in the
mid- to late-Holocene, especialy the matter of
whether there have been higher-than-present stands.
Bloom (1980, 1983) modelled some of the global
diversity in these Holocene curves and suggested
that a +1-2 m stand existed in the south Pacific
region during ts period. Substantial geomorphic
and radiometic evidence from a variety of islands
now supports this intepetation of a +1-2 m high sea
level during the period between about 4-2 kyr B.P.
In Fiji, for examnple, Nunn (1990:304) concluded that
the coasts "experienced a middle to late Holocene
sea-level maximum some 1-2 m above prset mean
sea level." Recent work by Miyata et al. (1990) also
supports this finding. Similar results are presented
by Ash (1987) for Viti Levu Island, and Yonkura et
al. (1988) report evidence for a +1.7 m stand between 3400-2900 B.P. on Mangaia Island in te
southem Cooks. In French Polynesia, Pirazzoli and
Montaggioni (1986,1988; Montaggioni and
Pirazzoli 1984) describe evidence from various
islands for a MSL between +0.8 and 1.0 m beginning about 6-5.5 kyr B.P. and lasting as late as 1.2
kyr B.P. In Westem Samoa, Rodda and his colleagues (1986; Sugimura et al. 1988) summarize
varous evidence for Holocene higher stands. Isla
(1989:361-63) discusses a comparable range of
evidence for several Pacific Islands. In figure 4.4,
the geographic distibudon ofmid- to late-Holocene
higher sea-level stands is plotted along with associated radiometric ages. Figure 4.5 shows a timeelevation plot of sea levels in various Polynesian
archipelagoes over the past 5 kyr B.P.
This widespread and consistent patem of
radiometrically dated shoreline features provides
strong evidence for a higher sea level ranging
between about +1-2 m over fte southwestem Pacific,
from at least 5 kyr B.P. and lasting until sometime
between 2-1 kyr B.P. After about 2 kyr B.P., sea level
fell (pedaps faidiy rapidly) to its present position.
TIis mid- to late-Holocene sea-level curve thus
provides one important dimension for a model of
coastal terrace formation at the To'aga site.
Subsidence in the Samoan Archipelago
Most ofthe volcanic archipelagoes of the centrl
Pacific region are aiayed lieiy along ageprogression sequences emanating from a "hot spot"
or magma plume on the floor of the Pacific Plate
(Menard 1986). As the Pacific Plate gradually
migmtes west or nortwest, active volcanic islands
move off te hot spot, and a new island is formed.
Thus, the islands in such an archipelago comprise a
"plume trace" of inceasig age from east to west
Classic instances of this pattem are the HawaiiEmperor chain and the Society Islands. The Samoan
archipelago presents a somewhat more complicated
geological picture, primarily due to reent volcanism
on Savaii Island which, at the westem end of the
chain, theoretically shodd be the oldest island
(Menard 1986:186-87). This unusual siution
confounded geologists for many years (Dana 1849;
Daly 1924; Steams 1944; Natland 1980). Recently,
however, K-Ar dating of rocks from various Samoan
islands has confirmed the basic pattem of an east to
west age pogression; the mean K-Ar ages are 0.1
myr for Ta'u, 0.3 myr for Ofu-Olosega, 1.26 myr for
Tutuila, and 2.2 myr for Upolu (McDougall
1985:318-19). Savai'i is presumably the oldest
island but has had renewed volcanism in recent
times. McDougall concluded ta "Upolu, Tuuila,
and the Manu'a Islands comprise a hot spot or plume
trace on te Pacific plate, similar to the Hawaiian
and Society Islands chains." Nevertheless, the
Morphodynamics of the Land-Sea Interface
MARINE
INPUT
TERRIGENOUS
INPUT
Landslides
Mass Wasting |
Surface Flow
SEDIMENT BUDGET
H
SrghEnergye
Biogenic
Volcanic Clastics
TRANSPORT LOSS
Colluvium
M(oral) Clastics
Calcareous Sand
Reef Detritus
Soil
Figure 4.2
A model for the sediment budget at To'aga, showing terrestial and marine inputs.
dominane of youthful volcanism on Savai'i remains
an egngma and may indeed reflect major rejuvenescence related to deformation of the Pacific Plate
adjacent to the Tonga Trench (1985:319; see also
Menard 1986:186-87).
This linear age-progression and the fact that
Sea Level
+
I
-
0
0
E
E
.0
co
.'
co
Sea Level
Figure 4.3
35
-
A model of shoreline transgression and regression as a function of relative changes in sea levels and in sediment bugt (after Chappell 1982).
Manu'a Islands are sll in a youfl stage of
geological evolution are essential factors in understading the local tectonic siion of Ofu Island.
The rapid constuction of a volcanic mass on the
oceanic crust of the Pacific Plate results in a phase of
subsidence, due to point loading on the thn and
flexible crst (Menard 1986:165-69). Such subsidence and associated crustal defornation is well
documented for the younger Hawaiian Islands. The
island of Hawai'i appears to be subsiding at an
average me of about 4.8 mm/yr.
For Ofu Lsland itself, we are not aware of
geological documentation of subsidence at the
present time, although we have strong indirect
reasons to believe that this is the case. The young
age of te island (only 0.3 myr) would itself suggest
that subsidence and crstal deformation (which lag
behind volcanism) have not yet weached equilibrium.
The evidence for active erosion of the island's
coastline, descrbed above, stmngly supports this
interprtation of active subsidence. Further evidence
is provided by the reefs along the To'aga area which
display active coral head growth and lack the
solution-pitted and eroded reef platforms typical of
islands that have been teconically stable during the
Holocene. These observations suggest that Ofu and
Olosega islands are still in a phase of subsidence due
to point loading on the oceanic crust resulting from
their initial volcanic construction.
Elsewhere in Samoa, dramatic evidence for
rapid subsidence was provided by the chance
36
The To'aga Site
o
*1
a)
0o
Is
I
0
I-
a)
0C)
.-
8t
04
0
c
.o
°
bo
Io
kI
,.I
qq
*s
Morphodynamics of the Land-Sea Interface
SOCIETY ISLANDS
o Scilly
" Mopelia
* Maupiti
v Tupai
v Bora Bora
& Raiatea
A Huahine
* Moorea
o Tahiti Nui
o Tahiti Iti
m
+2-
1.04
+1-
_~
.~~
_s
,_
p
L~~~~~o
so
m.s.I.- I
5
2
5~~~~~~I
1
0
-
km
t~~~~ t
' " '~P-M
~~~~~~
~
~
37
lwpgf_
1-164 1*
K)4PW4
-oo, .NM
r~~~~~~~~~~~~~~~~~66 09
I
.I
II
3I
3
2
4I
I
I
I
8
I
6
N.W.TUAMOTU
5
4
m
* Makatea
* Rangirioa
v Takapoto
* Mataiva
o Arutua
a
Kaukura
o Apataki
+1
s0.
-1.94
0 0
mm
P-6
--
:0:~
1*64
4
~0:
V
5.
GAMBIER ISLANDS
a Temoe
m
+1
COOK ISLANDS
o Mangaia
o Aitutaki
Rarotonga
E.TUAMOTU
w Vahitahi
20 Reao
-
104W
m.s.l9.
4
3
2
1
I
I
.+
1.04
WAO404
B..4
I
I
I
I
I
a
4
2
a
5
I
I
6
Age (Kyr)
Figure 4.5
Time-elevation plot of sea levels in Polynesian archipelagoes (source: Spencer 1989, fig. 13).
I.
38
The To'aga Site
discovery of the Mulifanua archaeological site
(Green and Davidson 1974; Green and Richards
and Gren 1989). Dating to 3251 ± 155
1975; L
B.P., Mulifanua is a Lapita-pottery bearing deposit
now situated -1.5 m below mean sea level, and
further capped by 0.75 m of reef rock (Leach and
Green 1989:324-26). The site was accidentally
discovered by dredging for a ferry berth Given
the site represents a village occupation on a former
beach or coast terrace, between 2.6-3 m of subsidence is indicated, or an approximate subsidence rae
of I mAlyr. (Davis [1928:249-53] also discusses
evidence for the submergence ofTutuila Island.)
In constructing a morphodynamic model for the
developnent ofthe To'aga cotal tenrace, we
assume that Ofu has been undergoing continual
subsidence trughout fte Holocene. Although no
precise rate can be empirically determined at this
time, we may use a woddng rate of 1 mAkyr based on
the Mulifanua siuation. (It is possible that the aul
rate of subsidence on Ofu could exceed this value.)
although this increases as one moves landward. The
primary source oftenestrial sediment is the steep
cliff which towers over the site, continually depositing talus and rockfall onto the coastal teface. Some
finer grained sediments wash over and down the cliff
during heavy rains, but there is no major alluvial
contribution of sediment Thus mass wasting
(principally landslides) and sheet erosion are the
major agents of terigenous sediment deposidon.
The sediment budget of the Toaga geomorphic
system would have fluctuated during the Holocene,
as the generation of biogenic sediments in particular
would vary with sea-level regimes. During periods
of rapid sea-level rise, when corals are actively
growing below mean sea level, the generaton of
sediment would be substantially reduced. When sea
level dropped or was stable for a perod of dme,
coral growth would have caught up with sea level
and would have been exposed to wave action,
resulting in erosion and te generation of calcareous
sediment
SEDIMENT BUDGETS: SOURCES AND
MODES OF DEPOSITION
MODELLING THE MORPHODYNAMICS
OF THE TO'AGA COASTAL TERRACE
IN THE MID- TO LATE-HOLOCENE
In addion
to the relative changes in sea level
resulting frm a combination of glacio-eustatic and
tectonic processes, we need to consider both the
sources of sediment which contributed to the
costruction of the To'aga coastal terrace and the
agents responsible for the deposition of these
sediments. The bulk of the Toaga coastal terrace is
constructed of calcareous sands and larger clastics
and chunks of reef conglomerate) of
(coral
marine biogenic origin. These biogenic sediments
are produced on the reef thogh a number of
mechanisms, including wave action and biologic
processes (such as the generation of sand by parrot
fish and other species that rasp and grind coral to
extract algae). These biogenic sediments are
transorted landward over te reef flat by wave
action. Episodes of high-energy storm surges,
associated with the cyclones
perodically lash
the island, are extremely important for the rapid
accumulaion of large quantities of sand and ofthe
larger size component of coral boulders and cobbles.
The contribudon of terigeos sediments to the
coastal terace appears to have been less significant,
e
The variables essential to a model of the formation of the coastal terace at To'aga in the mid- to
late-Holocene have been reviewed. In figure 4.6
these varables are diagrmmed along the same
temporal axis. Figure 4.6A shows the Holocene
glacio-eustic rise in sea level, reaching a +1-2 m
maxnimum between about 4-2 kyr B.P. Prior to about
5 kyr B.P., the rapid rise in sea level would have
continually drowned the shoreline, and e formation
of a stable coastal terrace would not have been
possible. Rather, the sea would have encrached
direcly against the island's volcanic mass, creating
the dramatic sea cliffbehind the To'aga site, as
described by Stice and McCoy (1968). Only after a
maximum sea level was achieved, between ca. 4-2
kyr B.P., could a coastal terrace have begun to have
formed along the base of the cliffs. At the sane
time, we presume ta Ofu has been subsiding at a
rte of about 1 m/kyr, as depicted in figure 4.6B.
Tberefore one takes into account these two controlling processes to determine the probable net change
in relative sea level as graphed in figure 4.6C. This
Morphodynamics of the Lad-Sea Interface
penod as the coral reef flroing the To'aga area was
exposed to wave action, especially stonn sures.
During the past I kyr B.P., as sea levels again
blized and local teconic subsidence coninued,
te sediment budget would again have decreased, so
that a phase of regression recommenced. This is
consistent with our field observations of active
shoreline erosion at the present time along the To'aga
site.
This model of the fonmation of the To'aga site
can also be diagrmmed as a temporal trajectory
along the two axes shown earlier in figure 4.3. In
figure 4.7, we have plotted a rtDdiction of the most
probable transgression-regression sequence for the
To'aga shoreline over the course of the past 6 kyr
B.P. In this diagram, Xt y-axis represents the net
sea-level change resulting from the combinaion of
both glacio-eustatic and local tectonic effects, while
fte x-axis represents fte changing sediment budget.
This model provides a wolking hypothesis for
te fomiation of e To'aga coastal terrace and its
archaeological deposits and may be tested against
fte strtgraphic and radionetric evidence denived
+2
A 0-2
EUSTATIC SEA LEVEL
-
B
SUBSIDENCE
-.
~~~NET SEA LEVEL
L
D+l
5
4
3
2
1
te
0
KYR B.P.
Figure 4.6 Time trends in four key variables affecting the
morphodynamics of the To'aga coastal plain.
te most likely period during
predicts
which the To'aga terrace could have prograded
through the rapid deposition of biogenic sediments
would have been between ca. 2-1 kyr B.P., when the
eustatcally contolled sea level dropped from its
mid-Holocene maximum down to modem levels.
Depending upon the extent of this drop (1-2 m), the
actual sea level on Ofu would have been either
stable, or there may have been a slight fall. This
would have resulted from te eustatic sea-level fall,
off-setting the continual effects of tectonic subsidence. As shown in figure 4.6D, the biogenic sediment budget would increase significantly during this
graph
of trasect excavations at various
points along site. Several predictions may be
generated based on the model: (1) The earliest
archaeological deposits should be located adjacent to
the cliffs and should date no earlier than about 4 kyr
B.P. (2) These early sediments are likely to have a
higher component of volcanic clastics, because talus
matenial would still have been readily available for
erosion by wave action. (3) There should be a fairly
rapid or even abrupt episode of coastal progradation
beginning sometime after about 2 kyr B.P. and
ending by about 1 kyr B.P. The sediments deposited
during this interval would consist almost wholly of
from our program
~~SEDIMENT BUDGET
I
6
39
marine biogenics.
This model was not developed prior to the actual
fieldwork, but evolved during the courae of our
investigations, as a dialectic between field observa-
tions and theoreical exercises. Nonetheless, it is
crucial to stress that the model in no way depends on
our archaeological data; it is wholly independent,
deriving from cumnt geological and geomorhological knowledge of coastal process in the
southwest Pacific. To briefly anticipate the results of
our field and laboratory studies presented in chapters
5, 6, and 7, we shall see dt the predictions of the
model developed here are substantially bome out
40
The To'aga Site
Relative
Sea Level +
I
.0-
._0
E
1
a)
E
0
0
CO
C)
Sea Level Figure 4.7 Retoiction of the tansgrssion-regression
sequence along the To'aga coastline fron 5 kyr
B.P. to the peent.
TO'AGA: SOME FURTHER
EXPECTATIONS FOR ARCHAEOLOGICAL
SITE FORMATION PROCESSES
Schiffer has avenred that "during human occupation of regions, natural processes, influenced by
cultual behavior, have created an ever-changing
landscape that the investigator perceives at just one
point in time. The contemporary region is a complex, the-dimensional mosaic consisting of natural
sediments, vegetation, modem artifacts and settlements, and archaeological remains. In order to find
sites and, especially, to understand how settlements
functioned in regional systems, one must endeavor
to infer or reconstruct changes in the landscape"
(1987:261). In this chapter, I have endeavored to
model some of the key processes which contributed
to the formation and modification of archaeological
deposits in the To'aga area. Lying at a fragile
interface between land and sea, the To'aga coastal
terrace is subject to envirornental influences of
several kinds, and can potentially undergo rapid
changes if any of these inputs vary. This interfacial
zone has also been subject to intensive human use
which has implications for both te formation and
modification of the archaeological record.
Among the most important implications of our
model of morphodynamics of the coastal terrace is
that the very existence of this interfacial microenvirorunent depends upon a sediment budget which has
fluctuated highly during the mid- to late-Holocene.
Since the archaeological deposits at To'aga are an
integral part of the geomorphological structure of
this zone, it is essential that we understand these
processes of coastal ten-ace construction and erosion.
To quote Schiffer again, "from the standpoint of
survey [and excavation] design, the archaeologist
needs to know where deposition and erosion are
occurring and where they have occunred during the
period of human occupation of the region"
(1987:257). Our model of relative sea level and
sedimentary budgets suggests that the To'aga coastal
ten-ace could not have begun to form or stabilize
until after the Holocene sea-level maximum of about
5-3 kyr B.P. Given that the Westem Polynesian
region was first colonized by the makers of Lapita
pottery at about 3.4-3.2 kyr B.P. (Kirch and Green
1987; Kirch and Hunt 1988), the area of coastal
terrace available for initial establishment of human
habitations would have been very restricted, probably consisting of little more than a beach ridge lying
directly under the steep cliff. As the coastal terrace
began to rapidly prograde after about 2 kyr B.P., the
area available for intensive land use would have
increased substantially. Such progradation, however, combined with deposition of colluvium,
probably would bury the earlier occupation deposits
deeply. Thus, our morphodynamic model also
implies that the earliest archaeological deposits will
be the most difficult to locate, being situated fartest
inland and possibly buried under substantial depths
of colluvium. The model predicts, therefore, that
archaeological deposits dating to the earliest phase of
Samoan prehistory-the phase marlced by te
presence of ceramic assemblages-are not likely to
be exposed on present ground surfaces or to be
encountered by surface survey alone, no matter how
intensive. Indeed, the initial archaeological reconnaissance of the island by Sinoto and Kikuchi
[Emory and Sinoto 1963] failed to locate a single site
Morphodynamics of the Lnd-Sea Interface
of this early period.
The model also has implications for the longertenn managemen of archaeological sites or 'cultual
resources' of the To'aga area. We explore these
implications in more detail in chapter 15, but briefly
note here that if, as our model predicts, the To'aga
area is now undergoing a phase of active erosion and
shoreline transgression, the archaeological sites of
this area will in time be eroded away. How quickly
these sites will be theatened by erosion depends on
several factors, especially the rate of subsidence and
of eustatic sea-level rise. The problem of 'global
waming' and associated sea-level ises (Geophysics
Study Committee 1990), however, could potentially
rsult in an acceleration of coastal erosion in the
To'aga area.
ACKNOWLEDGEMENTS
I am particularly grateful to Joanna Ellison of
the Depanment of Geography, University of California at Berkeley, for kindly supplying me with
vadous references on sea-level change in the
southwest Pacific, for the use of figure 4.4, and for
her comments on an earlier draft of this chapter.
REFERENCES CITED
Ash, J. 1987. Holocene sea levels in norerm Viti
Levu, Fiji. New Zealand Journal of Geology
and Geophysics 30:431-35.
Bloom, A. L. 1980. Late quatemary sea level change
on south Pacific coasts: A study in tectonic
diversity. IN N-A. Momer, ed., Earth Rheology, Isostasy and Eustasy, pp. 505-516. New
York: John Wiley and Sons.
1983. Sea level and coastal changes. IN H. E.
Wright, ed., Late Quaternary Enviroinents of
the United States, vol. 2, The Holocene, pp. 4253. Minneapolis: University of Minnesota
Press.
Chappell, J. 1982. Sea levels and sediments: Some
features ofthe context of coastal archaeological sites in the tropics. Archaeology in
Oceania 17:69-78.
Crosby, A. 1986. Ecological Imperialism. Cambridge: Cambridge University Press.
Daly, R. A. 1924. The geology of American
Samoa. Carnegie Institution Publication
41
340:93-152.
Dana, J. D. 1849. Geology. Vol. 10 of U. S.
Exploring Expedition Durng the Years 18381842 Under Cmmand of Charles Wilkes,
USN. Washington D.C.: Govenmnent Printing
Office.
Davis, W. M. 1928. The Coral ReefProblem.
Amercan Geophysical Society Special Publication No. 9. New Yoik.
Ellison, J. 1989. Pollen analysis of mangrove
sediments as a sea level indicator. Assessment
from Tongatapu, Tonga. Palaeogeography,
Palaeocibnatology, Pakaeoecology 74:327-41.
Fairtmidge, R. W. 1961. Eustatic changes in sea
level. Physics and Chemistry of the Earth 4:99185.
Farrell, B. H. 1962. The village and its agriculture.
IN J. W. Fox and K. B. Cumberland, eds.,
Western Samoa: Land, Life, and Agriculture in
Tropical Polynesia, pp. 177-238. Christchurch:
Whitcmbe and Tombs.
Friedlander, I. 1910. Beitrage zur der Samoa Inseln.
K. Bayer. Akad. Wiss., Math-Phys. El., Munich,
Band 27:358-69.
Geophysics Study Committee, 1990. Sea-Level
Change. Studies in Geophysics. Washington
D.C.: National Academy Press.
Green, R. C., and J. M. Davidson, eds. 1974.
Archaeology in Western Samoa, Vol. 11.
Bulletin of the Aucldand Institute and Museum.
Green, R. C., and H. G. Richards 1975. Lapita
pottery and a lower sea level in westem Samoa.
Pac#fc Science 29:309-315.
Isla, F. I. 1989. Holocene sea-level fluctuadon in the
southem hemisphere. Quaternary Science
Reviews 8:359-68.
Kirch, P. V. 1988. Niuatoputapu: The Prehistory of
a Polynesian Chiefdom. Thomas Burke
Memoral Washington State Museum Monograph No. 5. Seatte: Burke Museum.
Kirch, P. V., and R. C. Green 1987. History, phylogeny, and evoludon in Polynesia. Current
Anthropology 28:431-56.
Kirch, P. V., and T. L. Hunt, 1988. The spatial and
temporal boundaries of Lapita. IN P. V. Kirch
and T. L. Hunt, eds., Archaeology of the Lapita
Cultural Complex: A Critical Review, pp. 9-31.
Thomas Buike Memorial Washington State
Museum Research Report No. 5. Seattle: Burke
42
The To'aga Site
Musewn.
Kirdc, P. V., and D. E. Yen 1982. Tikopia: The
Pre/udtory and Ecology of a Polynesian
Outlier. Honolulu: Benice P. Bishop Museum
Buletin 238.
Leach, H., and R. C. Green 1989. New infonnadon
for the Ferry Berth Site, Mulifanua, Westem
Samoa. Journal of the Polynesian Society
98:319-29.
McDougall, I. 1985. Age and evoluton of te
volcanos of Tutuila, American Samoa. Pac#fc
Science 39:311-20.
McLean, R. F. 1980a. The land-sea interface of
small tropical islands: Morphodynamics and
man. IN H. Brookfield, ed., PopulationEnvironment Relaions in Tropical Islands: The
Case ofEastern Fiji. MAB Tehnical Note No.
13. Paris: UNESCO.
-. 1980b. Spatial and temporal vadability of
extenal physical controls on small island
ecosystems. IN H. Brokfield, ed., PopulaionEnvirovnent Rela s in Tropical Islands: The
Case ofEastern Fiji. MAB Technical Note No.
13. Paris: UNESCO.
Menard, H. W. 1986. Islands. New Yoik: Scientific
American Libray.
Miyata, T., Y. Maeda, E. Matsumoto, Y.
Matsushima, P. Rodda, A. Sugimum, and H.
Kayanne 1990. Evidence for a Holocene high
sea-level stand, Vanua Levu, Fiji. Quaternary
Research 33:352-59.
Montaggioni, L F., and P. A. Pirazzoli 1984. TMe
significance of exposed coral conglomerates
from Fenh Polynesia (Pacific Ocean) as
indicators of recnt rladve sea-level changes.
Coral Reefs 3:29-42.
Nakamura, S. 1984. Soil Survey ofAmerican Samoa.
Soil Conservation Service, U.S. Department of
Agriculue. Washington: Government Printing
Office.
Natand, J. H. 1980. The pgion of volcanism in
the Samoan linear volcanic chain. American
Journal of Science 280-A:709-735.
Nunn, P. D. 1990. Coastal processes and landfonns
of Fiji: Their bearing on Holocene sea-level
changes in the south and west Pacific. Journal
of Coastal Research 6:279-3 10.
Pirzzoli, P. A., and L. F. Montaggioni 1986. Late
Holoce sea-level cages in the northwest
Tuamotu Iands, French Polynesia. Quaternary Research 25:350-68.
1988. Holocene sea-level changes in French
Polynesia. Paleogeography, Paleoclimatology,
Paleoecology 68:153-75.
Rodda, P. 1988. Visit to Westem Samoa with
HIPAC team. IN N. Yonekura, ed., Sea Level
Change and Tectonics in the Middle Pac(i1c.
Report of the HIPAC Project in 1986 and 1987,
pp. 85-90. Deparunent of Geogrphy, University of Tokyo.
Shiffr, K B. 1987. Formation Processes of the
Archaeological Record. Albuquerque: University of New Mexico Press.
Schofield, J. C. 1969. Notes on late Quatemary sea
levels, Fiji and Ramtonga. New Zealad
Journal of Geology and Geophysics 13:199206.
Shepard, F. P. 1963. Thirty-five thousand years of
sea level. IN Essays in Marne Geology. Los
Angeles: University of Califonia Press.
Stearns, H. T. 1944. Geology of the Samoan Islands.
Bulletin of the Geological Society ofAmerica
55:1279-1332.
Stice, G. D., and F. W. McCoy, Jr. 1968. The
geology of the Manu'a Isands, Sanoa. Pacfic
Science 22:427-57.
Sugimum, A., Y. Maeha, Y. Matshima, and P.
Rodda 1988. Further report on sea-level
investigation in Westeem Samoa. IN N.
Yonra, ed., Sea Level Change and Tectonics in the Middle Pacific. Report of the HIPAC
Project in 1986 and 19879 pp. 79-84. Department of Geogray, University ofTokyo.
Visher, S. S. 1925. Tropical Cyclones of the Pacifc.
Bernice P. Bishop Museum Bulletin 20.
Honolulu.
Wiens, H. J. 1962. Atoll Environment and Ecology.
New Haven: Yale University Press.
Yonekur N., T. Ishii, Y. Saito, Y. Maeda, Y.
Matsushima, E. Matsunoto, and H. Kayanne
1988. Holocen fringing reefs and sea-level
change in Mangaia Island, Southern Cook
Islands. Palaeogeography, Palaeoclitology,
Palaeoecology 68:177-88.
5
EXCAVATIONS AT THE TO'AGA SITE (AS-13-1)
PATRICK V. KIRCH AND T. L. HUNT
INTRODUCTION
NTrr of fte Manu'a
of the archaeolknowledge
in
1986,
our
Project
ogy and prehistory of the Manu'a Islands was lImited
to a few late prehistoric architectal sites and to
surface finds of stone adzes and flakes (Kikuchi 1963;
Emory and Sinoto 1965; Clark 1980). No wellstratified occupation sites were known, and there was
no established sequence of occupation phases extending back to the Ancestral Polynesian period, as had
been developed for Westem Samoa (Green and
Davidson 1969, 1974). bdeed, previous archaeological reconmaissance in Manu'a had failed to locate any
sites that contained prehistonc pottery and thus might
date to the first millennium B.C., based on comparisons with the Samoan sequence as defined for Upolu
Island. A major goal ofour 1986 reconnaissance
survey of Ta'u, Olosega, and Cfiu islands was thus to
determine whether the appaent absence of ceramicbearing sites was simply an artifact of archaeological
sampling, or whether such early perod sites were
tnuly absent in the Manu'a Group.
The surface find of a Polynesian Plain Ware
sherd in Ta'u Village on the first day of our 1986
reconnaissance was followed up by test excavations,
revealing the presence of subsurface occupation
deposits dating to the Ancestral Polynesian period
(Hunt and Kirch 1988). This discovery was sufficient
to establish that the prehistory of the Manu'a Islands
PRIOR Tro Th coMmEN
would in general terms parallel that of the larger and
better known Westem Samoan group. Moving from
Ta'u to Ofu Island, we subsequently discovered a
pottery-bearing site at To'aga on the southem coast,
where the Public Wo*ks Department had bulldozed a
landfill dump and thereby exposed subsurface
deposits. This area was within the site AS-13-1
defined by aark (1980). We canied out a limited
test excavation at the To'aga landfill in 1986, eovering pottery and other artifacts in situ (see below).
Because of the significance of the To'aga materals
for elucidating Manu'a prehistory, we expanded our
excavations in 1987, revealing that the site was far
more extensive than originally thoughL A third
season of trnsect test excavations was carried out in
1989 to assess the full area of this extensive and
deeply staified site. In this chapter, we present the
details of these three excavation seasons, including
the specific excavadon pxcedures used, as well as
the srtigraphy and depositional sequences revealed.
Subsequent chapters provide detailed analyses of the
ceramics, other portable artifacts, and faunal materals, as well as discussions of radiocarbon chronology
and site geomorphology.
Field Methods: General Comments
Some aspects of field methodology which were
constant in all field seasons may be brefly summa-
rized. The prncipal sampling strategy used was that
44
The To'aga Site
of "systematic trasec " (Redman 1974; Kirch and
Yen 1982; Kirch 1988), with 1r-m excavation units
generally spaced at 10 m intervals. We oriented
these transets perpendicuarly to the coastine in
order to provide geomorphic proffles across the
coastal flat All trae pfeswee canied out to the
reefflat and coI to mean sea leveL Usng a
tlescopc level and stadia md, a srface elevato
prfile was obtained for each trmuect. Observan
were also made of soil vanation and vegetation
associations along the trasect, as thse data indicated cuntland use and provided im
clues as
to the geomorphic history of the site. This
strategy of acquiring data (surface and excavation) along systematic transects has the advantage
of producing results that can be mapped and
interpolated.
Whenever possible, excavation proceeded
according to natral stigraphy, although arbitray
subdivisions were made within thick strata Individual excavation blocks (refened to as "spits') were
designated within strata, but never cross-cut stratigraphic boundaries. Detailed rwcords were maintained on aarz recording forms (fig. 5.1), on
which horizontal feames were drawn to scale and all
localized finds were plotted according to x, y, and z
coo.rinates. Aftrexcavation, measuwd stratgrapic pofiles were dawn for each excavation
unit, and all sta were described in terms of
thickness, boundary, morphology, color, lithology,
cultual conte, and other caracteristics. Color
designations are from fte Munsell soil color charts
(Munsell 1988). Most proffile descriptions were
done by Kirch to assure consistency (in 1989, Hunt
described several units after Kirch left the field to
initiate a project in the CDook ls). Following the
drawing and description of each profile, a senres of
sediment samples were taken; sediment samples
were taken from within stata, never cross-cuttng
stratigraphic boundaiies. These samples were
subsequently analyzed in the University of California, Berkeley, geacaeology laboratory (see Kirch,
Manning, and Tyler, chapter 7).
All excavated sediment was sieved through 0.25
inch mesh screens. Although smaller mesh (particularly 0.125 inch mesh) would have enhanced the
recovery of minute faunal remains, we opted against
tfhis screening stategy in the inreests of covering a
greater area duing this testing phase. Given our
primary objective of detemuning t areal extent and
nature of the subsurface deposits at To'aga, sufficient
areal sanpling was judged to be a more important
consideation than complete faunal recovery.
During screening, however, we made perodic
checks of the small size fraction passing thrgh the
0.25-inch mesh screens and were generally satisfied
that the bulk of the faunal materals was being
retained. In addition, bulk sediment samples were
taken from strata in the areal excavation blocks, as
well as from some of the individual test pits (see
Nagaoka, chapter 13). These samples allowed us to
assess the frequency of minute faunal remains. All
vertebrate and invertebrate faunal materials caught in
the sieves were bagged and shipped back to the
laboratory for identification and analysis. Preliminary sorting of the vertebrate fauna, as well as
detailed identifications of the avifaunal remains,
were carried out by Dr. David Steadman of the New
York State Museum (see chapter 14). Furfter
analysis of the non-bird vertebrates and of the
invertebrate fauna was undertaken by Lisa Nagaoka
of the University of Washington (see chapter 13).
The numbering of stratigraphic units (distinguished by roman numerals) is based primarily on
sedimentological (rather than culturl) critera Each
sedimentological-lithological unit that was determined to have been deposited either as a single
event, or as several events all representing the samne
source and mode of deposition, was designated as a
"layer." Subunits withn these sedimentological
layers, including cultural occupadons, are designated
by letters. Hence, an occupation episode that is
wholly incorporated within a calcaeous sand beach
ridge deposit, Layer III for example, might be
designated as Layer HIB, with e lturally sterile,
but lithologically identical sands above and below
designated Layers hIA and IIIC. Thus, our layer
designations emphasize fte geomorphological site
formation processes rather than simply the pesence
or absence of cultural materials.
In describing the lithology of beach rdge
depositional units, we paid special attention to admixture of volcanic lithic grains with the dominant
calcareous grains. When volcanic grains are present,
these give the sediment a 'salt-and-pepper' appearance. The signficance of this 'salt-and-pepper'
lithology is that it reflects the availabiity of a
volcanic sediment source from which sand grains
Excavation
I
45
x
'' PM.
89
MNa 'a Project, 199
Sieve& DIy
cI4.et 0
&zt
Sit As-i-
A"11111m 'r-41s0o5
S'IddaW
Spit
F
;
1
Df~~
Depths: Surfae Ciiatum 0O
.L
..
..
.;
.t
End Ieve
-P
I
N
yI__
m
Mea En*cm|!r
|
Mean StartZ=
*_cm-
COMMENTS (Note sediment char~teristics, color; T
disturbances; samples taken; speclproblems):
0
(J STl &SALT' (tu44AaM foft(6
-tVt
tA~8ftC-4
.~~.)rL
-~~~~~
CliAkColF
-
u
S .b^t
i.*
OL~
jA
B,e~
$
-
*
,
6ete8)M}(~s
mVAIO
7
468(
74S,5
A ht5RP)
LAAYE tc.'oAn
1'
&OlCIWA I;"?
-SVAItrat
msScr''So6(
F'S
-T
AAz VAN,*w
-
_
Vw--To
Ftw&V
Rr'.AA-TtO%Z
d
Re
lb
o ~~~~~~~~~~CJ
#Ai
,r~u
lm
1-5.Ud'I¢°~W/
OLA. CbCgAC
Ceix
P
PC"'
Ct- 5LIRP1,^
e c: fS 5)g
fw'A Nr.v
T
->psli,o#
ShApSIOL. 7-*A7VST65CmA-
-
0
P
6C&AC-46
Recorder: C. mw,ae-ls
P
I
ALJW1
L
-m
Figure 5.1
Example of the excavation recording form used in the
is for level 5 of Unit 20 on Transect 9.
1987 and 1989 To'aga field
seasons.
This form
46
The To'aga Site
were produced by wave action along the fonner
shoreline. Today, fte sandy beaches fng To'aga
are almost pwely calcares. Volcanic grains are
presen only in a few spots, where volcanic headlands p
gh the coastal terrace and are
exposed to wave action. The frequent presence of
'salt-and-pepper' lithologies in deeply buried
deposits exposed by our excavation indicates a
landform stage whm the coastal ten-ace was much
narrower (prior to progradaton), and when substantially more volcanic material was available for
incorportion into the sediment budget Thus, along
most transects, there is a tran from deeper,
earlier 'salt-and-pepper' sand lithology to pudely
calcareous lithology, reflec coastalion
and removal of most volcan:ic source material from
the sediment system. In chapter 7 we use point
counting of sediment sampes to furtxr document
the precise qutities of volcanic lithic grains in
strata within various excavaton units.
containing coral and basalt pebbles and cobbles,
shell midden (mostly Turbo spp.), and ceramics
(four thickware sherds orly). Layer II rached a
maximum depth of 135 cm below surface. The
contact with Layer II is very diffuse. A single
adiocarbon date (Beta-19742) on Turbo shell
yielded a conventional radiocarbon age of 2350 ± 50
B.P. (cal 28 B.C.-A.D. 108 at one s ard deviation;
see chapter 6).
Layer m: No culturl material was pent in the
white (10 YR 8/2) calcareous sand of this layer.
Excavation of test Unit A reached a maximum depth
of 160 cm below surface.
Subsquent woik in 1987 revealed ta these
major stratigraphic zones fit well within the geomorphological sequence for the To'aga flat as a whole
(see below). Artifacts collected frxm the bulldozed
area of the To'aga landfill and from Unit A were
described and illustrated by Hunt and Kirch
(1988:169-76) and are included here in chapter 1 1.
1986 TEST EXCAVATION
THE 1987 EXCAVATIONS
TIe Toaga site (AS-13-1) was initially discovered duing te 1986 naissance survey, when
examination of a deep bulldozer cutting made by the
Public Woiks Department for a sanitary landfill
distbed a buned cultual deposit contning
Polynesian PaiOn Ware cermnics (Hunt and Kirch
1988:168). The bulldozercutting lies atthe edge of
a fan ofmassive talus bouldes. It appears that the
landfill site is at the inland-most location possible on
this paricular section of the coastal flat A single
test pit (1 m2) was placed dirty adjacent (3 m
west) to the bulldozed area of the To'aga landfill.
This locale had mained undistubed by the bulldozer activities.
Excavation of the test pit (designated Unit A)
revealed the following major stata:
Layer I: This dark (2.5 Y 2V0) mucky sand mixed
with coliuvial clay is eniched with organic maer
(contemporary A soil horizon). It contains coral
pebbles, shell midden (mosdy Turbo spp.), and
waterwom basalt pebbles and cobbles. Layer I
reached a maximum depth of 60-70 cm below
surface. The contactwith Layer 1is diffuse over a
1-3 cmzone.
Layer U: This layer had motted pale yellow to
grayish brown (2.5 Y 7i2 & 512) calcareous sand
The 1986 surface collections from the landfill
site, combined with the limited results frm the test
excavation, revealed the presence of early
Polynesian occupation deposits in the To'aga area.
Along with the discovery of pottery at Ta'u Village
on Tau Island (Hunt and Kirch 1988), this was the
first record of an early phase of occupation in the
Marn'a Islands. With e concurence ofthe
Amedcan Samoa Historic Preservation Officer, we
therefore determined ta a major objective for a
second season of fieldwork in Manu'a should be
more intensive investigation of the To'aga site, with
the aplicaton of subsurface systemadc transect
tting in order to determine whether undistubd
potteiy-bearing occupation deposits were pnt in
the vicinity of te landfill site. Tibs second phase of
work was caried out in 1987.
Excavation Procedures
Returning to the To'aga site in 1987, we
focused on the coastal flat immediately northeast
ofthe ladfill which had not been disurbed by
bulldozing (fig. 5.3). As can be seen in the transect
profle in figure 5.2, this flat is about 125 m wide
from the base of the steep colluvial-alus slope to the
Eavatio
47
oD
CL)
OC
E
00E
0
a.
*.
Beach ca.
:
Beach ca. 2000 B.P.
0
o
25m
350BP
VERTICAL SCALE X4
Figure 5.2
Elevation profile along the 1987 excavato transect, showing the positions of the main excavato
and 3, in relation to geomoaphic and pedologic feat.
shoreline. Using the systematic tranct
strtegy developed by Kirch for sub-surface sampling in similar coastal settings in Tikopia and
prese
Niuatoputapu islands (Kirch and Yen 1982; Kirch
1988),
we
laid out
a transect
baseline
across the Toaga coastal fiat,
extending
at a location 60 m
of the Ofu landfill. The first ree units
excavated along the transect (at 0, 15, and 45 m from
the base of the talus slope) revealed a complex and
deep cultual stratigraphy in Unit 1, but only shallow
cultual deposits overlying calcareous beach sands in
the more seaward Units 2 and 3. A shovel test at
105 m fardher seaward along this transect, near the
crst of the modem beach ridge, revealed a total
absence of culural deposits, with only calcareous
sand. These tests thus demonstated that the oldest
cultunal deposits were to be found close to the base
most of
of the steep talus and volcanic cliff, and
the coastal flat consisted of culturally sterile coral
sands and reef detritus, which had been deposited
during seaward progradation during the past 2-3 kyr
northe
B.P.
Following these iniitial tansect tests, Unit 1 was
expanded into a larger excavation in order to effectively sample the deep stratigraphic sequence,
including the in-situ deposits of Polynesian Plain
Ware ceramics. Units 4-9 were excavated, joining
with Unit 1 to fonn a T-shaped trench as shown in
figures 5.3 and 5.4. All units were dug trugh the
Layer II calcareous sand deposit containing pottery,
while Units and 6 were caried deeper into under-
and Units 2
lying Layers HI and IV (see Stratigraphy, below).
The third stage of our 1987 excavation strategy
was to detennine the lateral extent of the early
pottery-bearing deposits southwest and northast of
the main excavation, parallel to the base of the talus.
Unit 10 was thus laid out 45 m southwest of the
baseline transect, as close to the base of the talus as
feasible (acmally set in among several massive
rockfall boulders). This revealed a stigraphy
similar to that in the main trench, including deeply
buried cultural deposits with Polynesian Plain Ware
and one fine, thin-ware sherd. Unit 1 1 was then laid
out 45 m nonheast ofUnit 1. Unit 11 revealed
ceramic-bearing deposits, but these were tnncated
by a large, deep pit (probably a late prehistodc ua'i
masi or breadfruit-fermentation pit), and so an
adjacent square, Unit 14, was opened to clarify the
stratigraphy. In order to get as close as possible to
the base of the talus, another test was laid out 15 m
northwestof Unit 11, atthe footofthetalus and
designated Unit 12. In Unit 12, a massive deposit of
colluvium and large angular boulders had to be
penetated to a depth of 1.8-1.9 m before we were
able to reach a thin deposit of calcareous sand
containing four thin, fine-tempered potsherds (one
rim and four body sherds). These test units indicated
that early, pottery-bearing cultual deposits at To'aga
extend over a distance of at least 105 m northeast of
the Ofu landfill, in a narrow zone at the base of the
colluvial-talus slope.
A final t, Unit 13, was opened in the centerof
48
The To'aga Site
636
CM
.5
I-
0
I-
z
z
E
z
0
Uf)
G)
C:
0
F4
-0
0.
I-
z
l
a)
.8
0
-4
L..
0
Iz
A
CD
o
c
0
X
0
(a
As
'I
U
01 Lt
CM
I-
Ct)
I-
z
z
*s
g
Iza
d)
'3>
)&
Vol
0 n
C
4._
Co
bO
I)
0
ol:
0
O-
/.:N
z
-S
O..
:
6:2
....
1eI
wI
Ecavations
Figure 5.4
49
Excavation in progress in the 1987 main trench. Note the large talus boulder on the left. P. Kirch
is recording the stratigraphy of the south face of Unit 1.
a low pavement of waterwom pebbles (such pavements are called 'ili'ili in Samoan) apparently
marking a later prehistoiic house floor. Our objective in excavating Unit 13 was to obtain a sample of
this later prehistoric midden to contrast with the
older, ceramic-associated assemblage.
Stratigraphy of the 1987 Main Excavation
Although the stratigraphy of the 1987 excavations varied from unit to unit, the most complete
depositional sequence was revealed in the main
trench excavations (Units 1, 4-9), but was reflected
as well in Units 10, 11, 12, and 14. Seaward of the
main trnch, Units 2 and 3 displayed simpler
stratigraphic profiles, resulting from the later
progradation of calcareous beach sands.
The westem profile of the 1987 main trench is
fsown in figure 5.5, in which all of te principal
depositional units are represented. The stradgraphic
units follow:
Layer IA: The upper 15-20 cm portion of the upper
colluvium (10 YR 3/1.5) found here has been
heavily reworked by gardening. Various planting
pits or depressions are detectable in te secdon.
Layer IB: This is a massive deposit of reddish
brown colluvium (10 YR 312), very compact, with
no internal lensing or bedding evident. No charcoal
flecking was observed. The deposit incorporates
numerous angular to subangular weathered volcanic
lithic fragments in the gravel-to-pebble size range.
A small, lens-shaped pocket of slightly darker soil
containing charcoal flecking was noted within Layer
IB (designated Feature 1) and may represent a
garden bum feature. This would suggest that Layer
IB accumulated gradually and that the land surface
was internittendy gardened during its deposidon.
Layer IC: The basal 15-20 cm of the upper colluvium (10 YR 2/1) makes up this sm. It is
somewhat darker than e overlying Layer IB,
containing dispersed flecks and chunks (5-10 mm) of
charcoal. The deposit is associated with an earth
oven feature in Unit 9, and the charcoal (which is
concentrated in a zone around the oven) appears to
50
The To'aga Site
~ ~ ~ ~ ~ ~ ~ ~ ~ ~ ~ 0~ ~ ~ ~ ~ ~ ~ ~ ~ ~ ~ ~ ~.
fi
0 ,|
Il0_,
_
S *,b;''j].
I-
-
t (f _
W
1t_I- 5 -0\''
i
CD~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~C
t
-D
LO
W (w\ a 1ll @ _/z1 l,
W--<
tt X
|;
@ i
\
rib
-C
_'t'S
1- ox
t
§
:'-\
I
k l
0
°l|
< j~~~~~~~~~~~~~~~~~~~~~~-
Z~~~~~~~~~~~~~~~~~~
-
lo.
.!~~~~~~~~~~\1
.E0
U; '~~~~~~~~~~~~~olmnSmpeiu
o
±.1'm
Z ~~~~ Iii ~~~~~
°
0Csh.
I
0
IIgN
~~~~~~~~~~~~~
I-~
o
E-
z
ColumnSampleCut~~~~~~~~~~b
ci
c~~
Depth in Meters
c~
.
|a
Exavations
have derived from oven rake outs. Some thick,
coarse-tmpered pottery was also recovered from
this deposit. It does not appear to have been a
permanent occupation, however, and may represent
a short-term or intermittent occupation such as a
field shelter on a former land surface.
Layer H: This unit with its subdivisions represents
a period of active deposition of a calcareous beach
ridge.
Layer HA-1: A thin zone at the top of Layer HA,
slightly darker and organically enriched (10 YR 512),
this stratum contains anthrpophilic land snails and
represents a phase of stabilizaton and vegetation of
the Layer II sandy beach ridge. Some occupation in
the vicinity is suggested by the presence of a few
thick, coarse-tempered sherds and a thin scatter of
marine shell midden.
Layer HA: This deposit contains loose, calcareous
beach sand (10 YR 7/2), not compacted or cemented
and lacking culul materials. Although most of the
sediment consists of calcareous materials, there is a
subordinate quantity of basaltic lithic grains, giving
the sand a distinct 'salt-and-pepper' appearanc.
This indicates ta at the time of deposition there
were exposed volcanic headlands in the vicinity of
the beach, providing a source for the basaltic sand
grains.
Layer IIB: This principal pottery-bearing deposit
represents a perod during which the surface of the
actively accumulatng sandy beach ridge was
occupied. Lithologically, Layer IIB is similar to HA
and IIC, but with the addition of organic/cultural
materials due to human occupation making it both
daiker (7.5 YR 4/2) and more compacted. The
deposit contains shell and bone midden, large
quantities of small sea urchin spines and test fragments, ceramic sherds (prmarily of thick, coarsetempered ware), and other artifacts. It also contains
antrpophilic land snails. The deposit is nonconcentrated and probably accumulated over a fairly
brief span of time.
Layer [IC: This layer, the basal component of the
Layer II beach ridge consists, as with HA, of a 'saltand-pepper' lithology with dominant calcareous
grains and a subordinate quantity ofvolcanic lithic
fragments (10 YR 5/2). Toward the base of this
deposit are numerous large coral cobbles and some
angular volcanic cobbles, along with branch coral
fingers and coral rubble. This material indicates a
51
relatively high energy deposidonal environmaet,
such as storm activity along an exposed beach frnt.
Thirteen sherds ofthin, fine-tempered ware were
present, although te deposit showed no evidence of
being an in-situ occupation locale.
Layer II: Massive silty-clay colluvium (7.5 YR 3-
4/2) with some incorporated subanguar lithic
fragments makes up this layer. Occasional charcoal
flecks are present, pariucularly near te top of the
deposit. Nine thin, fine-tempered potsherds were
also incorporated in the deposit. It appears to
represent a single depositional event, resulting from
a combination of mass-wasting ad fluvial spot
of terigenous sediment from the colluvial slope
above te site. A lower zone, designated Layer IIIB,
incorporates some sand mixed frm Layer IV,
presumably at the time of deposition. The presence
of charcoal flecking indicates burning of this slope
prior to fte deposition of fte sediment, pediaps due
to gardening or oter human disturbance.
Layer IV: Here one encounters a mixed deposit of
fine-grained calcareous sand and reddish silt-clay
(7.5 YR 5/4), apparendy culturally sterile.
Layer V: This basal deposit of 'salt-and-pepper'
sand (10 YR 8/2) includes marine shell and reef
detritus; fairly compact, and showing the initial
stages of cementation probably due to frequent
ground-water wetting. The deposit represents an
active beach ridge depositional environment. It
yielded two sherds of thin, fine-tempered ware,
which may have derived from a primary occupation
locus behind the beach (which would now be bured
under at least 10 m of colluvium).
Excavation Units 2 and 3
The stratigraphies of Units 2 and 3, seaward of
the main excavation, were essentially identical; the
section for Unit 3 (fig. 5.6) is typical:
Layer I: This organically enriched, silty-clay loam
is presently under cultivation. Lithologically, the
deposit is a mixture of calcareous sand (77%), silt,
and clay-sized terigenous grains, the latter deriving
from sheet wash erosion from the colluvial slope.
Various planting depressions filled with loose,
rewoiked soil are visible in the section. ITh deposit
also contains fragments of marine shell midden from
occupation in the vicinity. An earth oven feature
was exposed in the west face of the unit
The To'aga Site
52
I
UNIT 3: SOUTH
I
UNIT 3: WEST
Adz Section
I
0.0
0
4'0
40
2
0.5 c
0.
0
1.0
Figure 5.6
Stratigraphic section of the south and west faces of Unit 3. "C" indicates coral, and "V"
indicates volcanic.
Layer II: Here one encounters loose, calcareous
beach sand which is culturally sterile. This deposit is
composed wholly of calcareous grains and lacks the
'salt-and-pepper' combination of calcareous and
basaltic grains found in the Layer II deposit in the
main excavation This absence indicates that by the
time of deposition of Layer 11 in Unit 3, the coastline
had prograded sufficiently to buq the volcanic
headlands.
Excavation Unit 10
As shown in figure 5.3, Unit 10 was located 45
m southwest of the 1987 baseline transect. The pit
was situated in a small clear space between several
very large boulders, at the base of the steep talus.
Our aim here was to get as close to the talus as
possible, in the hope of exposing deeply buried
cultural deposits. This unit revealed a staigraphic
sequence similar to that in the main excavation
trench. The north and east faces of Unit 10 are
shown in figure 5.7. The strata are descnrbed below:
Layer IA: This upper pordon of tfie colluvium (5
YR 3/2) is presently gardened and tius loose and
reworked. Its silty clay loam includes angular lithic
gravel. The contact with Layer IB is diffuse over a
2-3 cm zone.
Layer IB: This very compact deposit of clayey
colluvium (5 YR 4/4) contains subangular lithic
fragments. Occasional charcoal flecks are prsent
There is no evidence that this zone had been reworked by gardening, and it probably represents a
single depositional event. The contact with Layer IC
is diffuse over a 2-3 cm zone.
Layer IC: The basal porton of the colluvium (5 YR
312), this deposit is similar to IA, and may represent
an older land surface that was gardened. It is less
compact than Layer IB and contains some
anthropophilic land snails (e.g., Lamellaxis gracilis).
Also, some disturbances penetrate down into Layer
II. The contact with Layer nI is distinct.
Layer HA-1: This thin, discontinuous deposit of
calcareous sand is enriched with organic matter and
stained gray (color 5 YR 4/1). Land snails of several
genera are present, indicating the fonner presence of
a stable land surface with vegetative cover. This
deposit is a former paleosol fonned on the old beach
ridge represented by Layer H. It reflects a phase of
stability capping the earlier phase of beach ridge
accumulation. This paleosol must have formed after
pgration of the shoreline removed the sotve of
cal os sediment that had resulted in he deposiion
ofthe deeper Layer ldeposits. The cntact with Layer
IIA is diffuse and gradational over a 3-4 cm zone.
Excavations
UNIT 10: NORTH
53
UNIT 10: EAST
0.0
0.5
at
0
0
1.0 c.0.
0
a
0
-1.5
Stratigraphic section of the
north and east faces of Unit 10.
2.0
Layer IIA: A culturlly sterile deposit of 'salt-andpepper' sand (7.5 YR 7/2), this stratum is predominantly calcareous but incorporates some volcanic
grains indicating the exposure of volcanic headlands
near the site at the time of deposition. The deposit is
fairly loose and was probably developed rapidly
after the abandonment of the underlying Layer UB
occupation. The contact with Layer IIB is very
irregular but relatively sharp.
Layer HIB: This is a ceramic-bearing occupation
deposit (7.5 YR 5/2). The parent lithology is
essentially identical to Layer IHA. This layer represents a period of occupation on the beach ridge while
te active deposition of sandy sediment continued,
primarily through aeolian action. The deposit is
stained grey with finely dispersed organic material
but is very sparse in midden (not nearly so concentrated as Layer IIB in the main excavation trench).
The contact with Layer UC is diffuse.
Layer UC: The mixture of calcareous sand and
reddish clay (7.5 YR 5/4) found here prbably
results from fluvial/hydraulic tnsprt of the clay
from the talus upslope. It is culurally sterile. The
contact with Layer IIn is sharp but iregular.
Layer m: This compacted deposit of reddish
brown clay-silt (5 YR 3/4) includes no larger lithic
fragments, suggesting a fluvial rather than masswasdng mode of deposition. No charcoal flecks
were observed. This deposit appears to represent a
single depositonal event, perhaps associated with a
54
The To'aga Site
storm or cyclone. The contact with Layer IV is
somewhat diffuse and mixed.
Layer IV: The 'salt-and-pepper' beach sand (10 YR
812) of this layer incorporates larger coral cobbles
and waterwom branch coral fingers. The layer
represents a fairly exposed, high-energy beach nidge
depositional event.
Apart from shell and bone midden, Unit 10
yielded little artifacul material. A few small flakes
of obsidian were wecovered from Layer IIA-1, and a
single thinware body sherd (5.7 mm thick) with an
orange (5 YR 5/6) slipped exterior came from Layer
IIB.
Excavation Units 11114
As shown in figure 5.3, Units 11/14 were
located 45 m east of te baseline transect. This
location is among massive talus boulders where the
ground surface is relatively level. Excavation of
Unit 11 revealed a large pit feature over the entire
unit which extended from the lower colluvium (ca.
55 cm below surface), cutting down through calcareous sand layers to 2.25 m below surface. The fill of
this pit contained 181 thickware sherds and 4
thinware sherds (2 with red-slipped exterors). Field
observations indicated ta pottery came from the
lower deposits adjacent to the pit and not frm the
pit fill itself. As nonnal stratigraphy for this area
could not be discemed with the intnasion of the deep
pit, Unit l 1 was expanded into arectangulartrench
(1 by 2 m) with the excavation of Unit 14. Unit 14
exposed an undistrbed sraigraphic sequence, as
follows:
Layer IA: The upper 20-30 cm of colluvium found
here was dark brown (10 YR 312) and had been
reworked by gardening. This layer contains some
charcoal flecking (probably modem) and no pottery.
Layer IB: This massive lower colluvium is very
poorly sorted and includes angular volcanic rock
fragments with no evident bedding or lenses. No
charcoal or other cultual materials were observed in
this dark brown (10 YR 312) layer.
Layer IC: This sm contained a dark grayishbrown (10 YR 3/1) mixture of clayey colluvium and
calcareous sand with abundant midden, including
shell, bone, pottery (309 sherds, with 304 [98%]
thick-coarse, and 5 [2%] tiinware), and charcoal.
This deposit was the source of the fill of the large pit
described for Unit 1 1, but in Unit 14, it made up a
discrete area between 54-75 cm below surface. This
layer appears to represent a depositional event
contemporaneous with Layer IC in the main trench
(see above).
Layer U: This calcareous sand deposit is enriched
with organic matter and some eluviation of clay
particles from the clay-rich Layer IC above. Layer II
is brown (10 YR 5/3) and contains pottery (15
thickware sherds) and a relatively sparse amount of
shell midden. This stratum is comparable to Layer II
in the main area excavation.
Layer Im: Calcareous sand (10 YR 7-812) with
waterwom coral cobbles and unwom branch coral
makes up this layer. It contains little cultural
material and appears to represent a high energy
depositional environment. Layer III compares in
lithology and content to Layer IIC in the main
trench.
Layer IV: This brown (7.5 YR 4/2) mixture of
silty-clay colluvium and calcareous sand has some
organic matter. Few larger lithic cobbles or boulders
are present, suggesting a sediment that probably
represents sorting with surficial flow of the terrigenous source, and wind (saltational) ransport of the
calcareous (beach ridge) source. A small hearth
comprised of waterwom coral boulders surrounding
a shallow ash concentration was excavated at the
surface of this layer in Unit 14. Five thinware sherds
were ecovered from this stratigraphic boundary.
Excavation was discontinued at the surface of Layer
IV due to time constraints at the end of dte 1987
season.
Excavation Unit 12
Unit 12 was located 16 m north and 45 m east of
Unit I inthe maintrench (see figure 5.3). he unit is
situated on the colluvial slope amidst huge talus
boulders. Excavation of this unit (chosen to penetrate the subsurface deposits as close to the talus
slope as possible) proved to be hazardous given te
large boulders protruding from the unstable
sidewalls of the pit, especially as the clayey sediment
dried and became friable. The stratigraphy of Unit
12 was as follows:
Layer IA: The upper colluvium (0-20/30 cm below
surface) consists of an organically enriched, A
horizon soil. The layer is dark brown (10 YR 3/1)
Exavtions
and has been rewoiked by recent gardening. No
cultural material was observed.
Layer IB: The massive lower colluvium, dark
brown (10 YR 3/2) in color, consists of very poorly
sorted, large, angular rock fragments, with cobbleand boulder-sized clastics, in a clay-silt matrix. No
bedding or lenses of finer-grained sediment were
observed, nor was any cultural material noted.
Layer IC: This is the base of the colluvial deposit,
distinguished by an incmase in fine-grained sediment The boundary with Layer IB is gradual,
reflecfing a continuous icreae ofclay and silt conte
toward fte base of Layer I. This day-loam aains
some calcamous sand (thoroughly intermixed) and
but these are reangular volcanic rck frag
stricted to cobble-sized. Charcoal flecks and
waterwom coral gravel are present in small amounts.
Small amounts of bone and shell were recovered
from this layer, but pottery was absent. Layer I
extends from the surface to 2 m below surface.
Layer IC-1: This is a brown (10 YR 3/3) clay-loam
with calcareous sand thoroughly intermixed. Layer
IC- I can be distinguished by both its greater calcareous sand content and the presence of pottery. The
pottery in this layer is predominantly thinware (25
sherds, with 2 thick-coarse and 23 thinware) and
includes duee direct rims, one of them with red slip
and an impressed lip. Shell and charcoal were
recovered in small quantities. The deepest (2.6 m)
portions of this layer had no cultural materials.
The stratigraphy of Unit 12 was almost entirely
colluvial in origin. The upper pordons, free of
cultural materials, represent one or more rapid
depositional events, i.e., mass wasting of material
from the nearby steep volcanic cliffs. The clay-loam
of the lowest portions of the sequence represents
gradual accumulation of colluvial sheet wash during
a perod of occupation. Calcareous sand entered the
deposit through wind trasport (saltation) from the
active beach ridge, which was then much more
closely adjacent to this location.
,
Excavation Unit 13
Located 45 m east and 37 m south from Unit 1
in the main trench, Unit 13 was near the center of a
stone-faced, waterwom gravel-filled ('ili'ili) mound
typical of Samoan house foundation construction.
The gravel mound measures 15 by 16 m across its
55
center and is nearly round in plan view. The mound
is elevated approximately 25-30 cm above the
surrounding organically enriched and stable sandy
substrate. Excavation of Unit 13 revealed the
following stratigraphy:
Layer I: This deposit of waterwom coral gravel
('ili'ili) is mixed with a small prporton of
waterwom volcanic gravel and with a finer, daick
organically enriched, sandy loam matrix (7.5 YR 2/
0). A bottle glass sherd was excavated from this
layer, possibly indicating occupation of the house
mound into the Historic period.
Layer IB: This is a lighter colored (10 YR 5t2)
sediment of the same texture as Layer I. The color
difference is due to a lower percentage of organic
matter. Ash, shell, and numerous sea urchin spines
were mixed into the gravel paving stones. Excavation of the lower part (3048 cm below surface) of
this stratum exposed a small pordon of a human
burial cut from the floor surface (Layer I) of the
house mound and extending into Layer II below.
Only bones from the feet were recovered, the
remainder of the burial lying outside of the excavation uit.
Layer H: This stratum contains culrally sterile,
white (10 YR 8/2). calcareous sand. Excavation was
completed with the testing of this layer to 68 cm
below surface.
Radiocarbon Chronology of the 1987 Excavations
Seven samples of charcoal and shell from the
1987 excavation units at To'aga provide the basis for
a radiocarbon chronology of the deposits in the
vicinity of the Ofu landfill. The 1987 dates were
reported in full by Kirch, Hunt, and Tyler (1989) and
are discussed in detail in chapter 6.
The oldest dates are those for samples Beta25035 and -25673, both frm Layer V of the main
trench, the basal calcareous beach sand which
yielded two thin, fine-tempered sherds. Both
samples consisted of unweathered marine shell
which were probably deposited at approximately the
same time as the sherds (details of these samples are
provided in Kirch, Hunt, and Tyler 1989:11-12; see
also Kirch, chapter 6). The samples yielded ovedapping ages at two standard deviations which indicates
a calibrated time range of between 3700-3300 cal
B.P. for the deposition of dtis beach sand which
56
The To'aga Site
contained the thin, fine-tempered ware pottery
sherds. While this age range is certainly early for
human occupation in Westem Polynesia, it is not
out-of-line with the earliest known dates for Lapita
sites in the region (Kirch and Hunt 1988; Kirch
1988, table 48).
Three samples are in direct association with the
Polynesian Plain Ware assemblage (Beta-25034,
-26464, and -25033), denrving from th Layer II
occupation in the main trench and from its corelated
deposit in Unit 10. All three samples agree well in
age and indicate a calibraed time range of between
2500-1900 cal B.P. for Xt main Polynian Plain
Ware occupation.
A terminus post quem for the use of ceramics at
the To'aga site is provided by sample Beta-26463,
which derives from the base of an aceramic occupation deposit which statigraphically postdates Layer
n in the main trench. This sample yielded a calibrated age of 1389-1287 cal B.P.
The final sample (Beta-26465) was obtained
from the base of the aceramnic house mound tested by
Unit 13 and yielded a calibrated age of 1122-950 cal
B.P. Since the platform is consructed on the present
land surface, it is evident that the sequence of coastal
progradation and of colluvial deposition in the
vicinity of th main excavation had stabilized by the
beginning of the second millenniUM A.D.
The Depositional Sequence in the 1987 Main Trench
The depositional sequence in the 1987 main
trench excavation at To'aga can be summadzed as
te following series of stages:
Stage 1 (3700-3300 Cal B.P.): The calcareous sand
beach represented by Layer V of the main trench
was formed at a time when the island's shoreline
was close to the base of the talus slope. Thus,
volcanic headlands were exposed to active wave
erosion, yielding mixed calcareous-basaltic lithology
sands. Human occupation in the vicinity of the
beach (presumably now buried under a considerable
depth of colluvium and talus) is suggested by the
presence of tin fine-tempered ware sherds.
Stage 2 (> 2500 cal B.P.): The older beach ridge was
buried at this time by terigenous silt-clay from
upslope, with erosion initiated in part by human
clearance of the vegetation with fire, indicated by
charcoal flecking in the erosional deposit.
Stage 3a (ca. 2500 cal B.P.): Accumulation of the
calcareous beach ridge was renewed during this
stage, while the shoreline was still close to the
talus slope (revealed again by mixed calcareousbasaltic grain suites). Exposure to the active
shoreline is indicated by the large coral cobbles
and reef detritus, representing one or more storm
events.
Stage 3b (2500-1900 cal B.P.): Humans inhabited
the active beach ridge surface (Layer IIB of the main
trench) during this time, resulting in the deposition
of ceramics and midden.
Stage 4 (2500-1900 cal B.P.): In this stage, the
beach ridge was abandoned as an occupation locus,
and the accumulation of the beach ridge (Layer HA)
continued.
Stage 5 (ca. 1900 cal B.P.): The old beach ridge
surface stabilized during this stage. This is indicated
by the formation of an A soil horizon and the
deposition of anhopophilic garden snails (Layer
IIA-I of the main trench). The stabilization presumably resulted from seaward progradation of the
shoreline, thus removing the immediate source of
calcareous sediment. At roughly this time, the basal
sands represented by Layer II in Units 2 and 3 were
deposited.
Stage 6 (< 1900 cal B.P.): During Stage 6 there was
additional erosion and deposition of temgenous
sediments from upslope in the area of the main
trench. The presence of ceramics and an earth oven
indicate a brief occupation event.
Stage 7 (< 1900-1000 Cal B.P.): The progradation of
the shoreline continued to its present position about
125 m seaward of the main excavation locus.
Gradual deposition of fine-grained terigenous
sediments over the surface of the newly formed
coastal flat and the rewoiing of these deposits
through gardening activities also occunred during
this stage.
Stage 8 (ca. 1000- 100 cal B.P.): Human habitation
dispersed across the expanded coastal flat during
Stage 8, represented by the 'ili'ili pebble pavement
tested by Unit 13 and by other surface cultural
features.
in sum, the depositional sequence revealed in
the 1987 excavations begins late in the second
millennium B.C. with a coastal teriace only a few
tens of meters wide, and with a reef platform
substantially wider than at present. This geomorphic
Excavations
situation fits well with evidence for a +1-2 m higher
stand of the sea in the southwest Pacific at ca. 3 kyr
B.P., as argued in chapter4. By about 1900 B.P., the
shoreline had begun to rapidly prograde, associated
with a drop in sea level to its present stand and with
exposure of the reef crest leading to a higher rate of
calcareous sediment production. The development
of a coastal terrace more than 100 m wide was
accompanied by the addition of fine terigenous
sediments due to erosion and sheet-wash of the
higher colluvial slopes, forming a highly prducfive
zone for intensive cultivation and habitation. lherefore, the morphodynamic model of coastal terace
fonnation outlined above in chapter 4 is closely
supported by the stratigraphic and radiometric data
from the 1987 excavations
Swnmary f the 1987 Excavaions
The program of excavations carried out in 1987
was successful in establishing the presence of
undisturbed, well-stratified prehistonic occupation
deposits in the To'aga area. The basal occupation
layers contained Polynesian Plain Ware ceramics,
including a fine, thinware pottery dating to the midfirst millennium B.C. The 1987 excavations also
established a sequence of geomorphological change
in the To'aga area, indicating substantial coastal
progradation and consequently deep burial of the
pottery-bearing deposits under recent calcareous
dune ridge and colluvial slope-wash deposits.
By the close of the 1987 season, it had become
clear that the subsurface archaeological resources of
the To'aga area-especially te deeply buried,
pottery-bearing deposits-were far more extensive
than originally thought. The 1987 excavation units
had revealed the presence of culural deposits over
an area of at least 4,000 i2, but without reaching the
horizontal boundaries of these layers. Because of the
great significance of the To'aga site for Samoan
prehistory, it was clearly essential to establish the
full areal extent of the subsurface archaeological
features at this extensive and well-staified site. A
third phase of excavations was therefore required,
applying the systematic tansect strategy of subsurface testing over the whole extent of the southeem
coastal plain on Ofu Island, extending from te Ofu
landfill site and adjacent 1987 excavations eastward
towards Fa'ala'aga. Potentially, this entire coastal
57
strip neady 2 km long and varying between 50 to
100 m wide, might contain subsurface archaeological deposits; this could only be determined drugh
subsurface testing. Such a testing prgrmn was
proposed to the Historic Preservation Office of
American Sanoa in 1988 and implemented by a
third field phase of the Manu'a Project in 1989.
THE 1989 TRANSECT EXCAVATIONS
1989 Excavation Procedures
The prmary objectives of the 1989 field season
at To'aga included-mong other things-the areal
definition of the fuill extent of subsurface archaeological deposits witin the coastal terrace and the
furter elaboration of a morphodynamic model of
site fonnation processes. These goals dictated the
continued use of a systemadc tract sanpling
strategy, with excavation transects spaced along the
full extent of fte coastal terrace (as far to the northeast as Fa'ala'aga). In order to maintain accurate
spatial control on the test units to be excavated along
these ransects, our first task was to establish a
baseline. This was staked out at 100 m intervals,
following the course of the dirt road that paallels the
shoreline trughout te To'aga area. A zero or
origin point was established where this baseline
bisected tX 1987 trmnsect (which was designated
Transect 1); from this 0-point, the baseline extended
300 m to the southwest, and 1500 m to the norheast.
Each 100-m interval along the baseline was identified as a potential transect-intersection point for
subsurface testing, and tansects were labeled fron 1
to 19. Transects 2-3 extended southwest from the
origin point (Transect 1), while tansects 4-19
extended to the norheasL
Choice of specific trans for test excavation
was dictated by several factors, including permission
from local landowners, locations of gardens (we did
not wish to disturb active Alocasia garden sites),
presence of 'sacred' sites (in particular, the Tui Ofu
tomb complex), as well as te usual constraints of
time and budget. Over the course of the field season,
we were able to excavate test units along six of the
potential sampling tansects: Transects 3 (200 m
W), 5 (100 m E), 7 (300 m E), 9 (500 m E), 11 (700
m E), and 17 (1300 m E). The locations of these
transects, and their relationship to te baseline, am
58
The To'aga Site
diagrammed in figure 5.8. (Note that in this figure
the scales for the baseline and te trinsects are
different.) These trants thus cover a total area
extending 1.5 kn from southwest to nonheast, and
in every case across te lateral extent of the coastal
terrace.
Prior to the commencement of excavation, each
selected traect was cleared of obscuring vegetation, and a transe baseline staked out at 10-m
intervals. A continuous elevation profile along the
Utnsect was then surveyed using a telescopic level
and stadia rod. These surveyed proffles provided
valuable infonnation on the geomorphic structure of
the coastal terrace. In order to ascenain the relationships between surface and subsurface geomorphic
features and modem sea level, each tansect profile
was also surveyed down to the shoreline, and out
onto the reef fiat.
A total of 16 -iM2 units were excavated along
the six tranwcs listed above. Virtually all units
extended to at least 1.5 m below surface, and some
units reached depths as great as 2.9 m before culturally sterile sediments were exposed. Thus fte total
excavated volume was approximately 32 m3.
Excavation units were numbered sequentially in the
order that they were iniidated, beginning with Unit 15
and ending with Unit 30.
In the following pages, the 1989 excavations are
described by trasect, beginning with Transect 3 at
the southwestem end of the coastal terrace and
progressing towards the norheast to Transect 17.
Transect 3
Transect 3 is situated 200 m to the west of the
1987 excavation area (Transect 1); between Transect
3 and Transect 1 lies the bulldozed landfill site
where pottery was first discovered during the 1986
reconnaissance survey. The surveyed elevation
profile of Transect 3 is shown in figure 5.9, with the
location of the single st, Unit 27, excavated along
this transect. The ground surface along Transect 3 is
composed of clayey colluvium with numerous
volcanic cobbles and boulders scattered over te
surface. The vegetation consists predominantly of
cultivated bananas.
Excavation Unit 27
This test excavation is situated 42 m inland of
the road. The profile of the east face is shown in
figure 5.10. The following strata were identified:
Layer I: This dari brown colluvial clay contains
angular and subangular basalt cobbles. Charcoal
flecks were present thughout in low frequency.
Layer I: An orange-brown-stained calcareous sand
makes up this layer. The lower part ofthis stm
contains many coral and basalt pebbles and cobbles.
The distribution of ese larger clastics suggests a
storm or other high-energy depositional event,
followed by a more gradual accumulation of the
finer-grained sands in the upper portion of the
deposit.
Layer MIIA: The brown-to-tan calcareous sand in
this statum exhibits some "peppenrng" of volcanic
grains. The straum also contains many angular and
subangular basalt cobbles and boulders, shell, and
coral rubble.
Layer hIB: This layer is composed of brown sandy
clay with angular to subangular basalt cobbles,
rounded coral cobbles, and shell. The clay component of this sand is higher than in Layer lIA,
although the boundary separating the two strata is
diffuse and unclear. Layer IIIB produced the bulk of
the cultural material in Unit 27, between about 175214 cm below surface. This cultual materal
included more than 6 kg of shell midden, 120 fish
bones, and some turtle (see Nagaoka, chapter 13).
The lower part of Layer EB was sterile, and excavation was discontinued at 225 cm below surface.
Unit 27 yielded a variety of prehistorc arifacts.
iMese include pottery, pieces ofobsidian (probably
natural), basalt (from Layers H and IIA), and an adz
butt fragment from Layer IRA. Layer I did not
contain any artifacts. A total of 95 sherds was
recovered, with the following strafigraphic distribution: Layer 11, 5 thickware sherds; Layer IIA, 80
sherds (5 [23%] thinware, 57 [71%] thickware);
Layer HIB, 10 sherds (3 [309] thinware, 7 [70%]
thickware).
No radiocarbon dates were processed from Unit
27. On the basis of the ceramics recovered, however, the occupation in Layer HsB could date to
avationh
59
q~~~~~~~~~~~~~~~Y
LL~~~~~~~~~~~~~~~~~~~~~~~~~
0 ~~~~~0-
LL~~~~~~~~i
-l)
I
*U, S ~~~~
gioesuleji
4
\%*%%N%~~~j<(I) 1;ISUSJ1L
V~~~~~~~~~~~~~~~~~~
>
co
)
0)
LL
S
~
~
~
~
~
~
~
E0
0~~
EM0
0~~~~~~i
The To'aga Site
60
w
0.
0
4
0
W
cnD
-l 0) CD~~C) C
2co
CM
C
A
-92
C~~~~~
0
0
CD)0
0
w
Q.
x
wCO
12
LI.
LI.
-3
-4
-5
Figure 5.9
t'
w
0
10
20 m
TRAN[SECT 3
Elevation profile along Transect 3, showing the location of Unit 27 in relation to geomorphic features.
approximately 1000 B.C. Only a single occupation
phase is indicated (in Layers IIIA and IIIB), followed by fte accumulation of calcareous sands,
mixed with increasing quantities of termgenous, claysilt sediment in the higher parts of the stratigraphic
column. The concentration oflarge clastics in the
lower part of Layer II is probably the result of a
high-energy storm event. In the most recent depositional phase, the locality has been capped by 60-70
cm of colluvium.
Transect 5
Transect 5 lies 100 m east of the 1987 baseline
which incorporates the main excavation trench
(Units 1, 4-9) and Units 2 and 3. In 1989, Transect 5
was largely in second growth vegetation, with some
Alocasia and bananas near the talus slope. The
elevation profile along Traect 5 is shown in figure
5.11.
Excavations along Transect 5 commenced with
three units, designated Units 15, 16, and 17, spaced
10 m apart After these units had been completed
and their statigraphic proffles correlated, it became
apparent that the most inland test, Unit 15, had
exposed a deeply buried, thin cultnal deposit
containing a few sherds of thin, orange-slipped
pottery. Seeking to sample more of this early
deposit, another excavation (Unit 28) was opened 10
m farther inland, where the talus slope began to rise
steeply. However, in this pit the massive upper
colluvial deposit extended filly 2.4 m deep, making
exposure of the underlying calcareous sandy strata
that contained cultual materials exceedingly
difficult. (Nonetheless, we did manage to carry the
excavation down to 3.6 m, a risky matter in an unreinforced lxI m pit!) We therefore retmned to the
Unit 15 excavation, and expanded it with 1I-m
extensions north (Unit 29) and west (Unit 30). The
excavation of these extensions was complicated by
the partial collapse of the expanded pit walls at one
point. Notwithstanding this setback, we carried te
excavation of these extensions down to 2.5 m,
thereby exposing a larger sample of the early culural
deposits. Our persistence in seeking an enlarged
sample of this deeply buried material was justified
subsequently by the msults of 14C dating of a
charcoal sample from Layer nI of Unit 28. The result
of 3257-2879 cal B.P. is the oldest date for unquestionably in-situ cultural materal obtained from the
To'aga site. Indeed, this sanple is penecontemporaneous with the submerged Lapita site of Mulifanua
on Upolu Island.
Excavation Unit 17
Unit 17 was the most seaward excavation along
TrnsectS, situated 56 m inland from the baseline
Excavations
here in an ashy, brown (10 YR 5/3), sandy matix.
This appears to be a tick, artificialy laid house
floor deposit and contains well-preserved shell
midden. Four post molds visible in te exploded
section (fig. 5.13) are filled with Layer II material
and were cut thrugh the undedying Layer III
pavements. In addition, a hearth visible in the east
and south profiles lies at the contact of Layers II and
UNIT 27: EAST FACE
0.0
0.5
Ill.
~~II
' "
1.0 0
.'
0
0
a
0.
1.5
'): ok )
.,,O
Figure 5.10
~~~Sherds
:?
D
61
2.0
Stratigraphic profile of the east face of
Unit 27.
along the road. A detailed profile of te west face is
shown in figure 5.12, and an 'exploded' profile of
the entire unit in figure 5.13. This unit was of
particular note because of the exposure of a sequence
of gravel house floor pavements ('iliill) in Layer III,
associated with several post molds and a hearth. The
stratigraphy as wecorded on the west face was as
follows:
Layer I: This 'greasy' midden deposit of sandy clay
loam contains much dispersed coral 'iliili gravel and
shll middenm It has a very dark gray color (10 YR 3/1).
Ihecontact with Layer is gradational over 2-3 an.
Layer fl: A compact deposit of 'ili'ili gravel occurs
Layer m: This thick deposit is made up of dutee
successive ili'ili pavements, laid directly over each
other. The paving matenals range from small
waterwom coral pebbles (ca 0.5-1 cm diameter)
down to very coarse coral sand (1-2 mm diamneter).
There are discontinuous bands of black ash and fine
charcoal trughout the deposit as well, probably
deriving frm hearth rake-out events. The lowest
pavement (Layer HIC) lies dircly on Layer IVA,
with a very sharp, abrupt, flat contac All of the
pavements were pale brown (10 YR 7/3). The
individual sub-components were: IIA, faidty clean
coral gravel; HIB, coral gravel mixed with sand and
ashy beds; and HIC, coarse coral sand and gravel in
fine beds. These ree gravel deposits presumably
represent successive repavements of a house floor.
Buck (1930:68) describes the traditional Samoan
practice of "levelling off the upper surface of the
platform within the house and covering it with small
stones ('iitiI). ....he larger coral gravel was picked
out on the beach and carried up in baskets. It often
took some time to get a sufficient quantity. A
Samoan woman pickdng over the coral gravel might
remind one of a woman of a higher culture selecting
a carpet."
Layer IVA: This deposit consists of strong brown
(7.5 YR 5/6) calcareous sand with a mixture of fine
volcanic silt-clay. The admixture of silt-clay
suggests surficial alluvial deposition of fine-grained
sediment emanating from one of the colluvial fans
inland, perhaps after a torrential rain. Some charcoal
flecking is also present
Layer IVB: The grayish brown (10 YR 5/2)
calcareous sand of this deposit has small dispersed
charcoal flecks. The deposit had been distubed by
the deposition of Layer IVA, and is therefore
discontinuous. It represents an old paleosol horizon
on top of the beach ridge following stabilization.
The contact with Layer V is iregular and diffuse.
Layer V: This is a massive, structureless deposit of
The To'aga Site
62
w
E
6
4
C
0
-W
co
2
0ccn
-J0
(0)
0)
CO) cD
.)
-j
Cu
co
(0)
C*
L(
aI')
D
_:
c
Lll
co)I
I
qm
0)
D
LL
w
w
0-
(0--4
O
I
2 -t
20
0
&
Figure 5.11
a
.
40m
a.
TRANSECT 5
Elevation profile along Transect 5, showing the locations of Units 28, 15, 16, and 17 in relation to
geomophic features.
fine to medium-graired calcareous sand (white, 10
YR 812) with a minor component of volcanic lithic
fragments. It is cultally sterile. The contact with
Layer VI is gradational.
Layer VI: This layer contains coarse-grained
calcareous sand (white, 10 YR 812) with a greater
frequency of volcanic lithic grins. The deposit also
incorporates two distinct bands of fist-sized coral
cobbles.
Layer VII: Composed of fine-grained sand with a
distinctly 'salt-and-pepper' mixture of calcareous
and volcanic grains, this layer maiked te base of the
excavation which was reached at 225 cm below
surface.
Unit 17 did not produce a large quantity of
cultural materials. One bone of Sus scrofa from
Layer I was the only identifiable vertebrate faunal
specimen. Shell midden was present in Layers I-Ill,
with the highest concentration in Layer III (see
chapter 13 for details).
Unit 17 yielded a single fishhook fragment (the
point of a jabbing hook) from Layer II. No pottery
was recovered from this unit, suggesting deposition
of dtese strata after the abandornent of pottery
production in Manu'a.
Excavation Unit 16
Unit 16 was positioned 10 m farther inland
along Transect 5 from Unit 17. The straigraphic
profile of the west face, shown in figure 5.14,
follows:
Layer IA: This is the A horizon-garden soil. On the
surface one finds much leaf litter with numerous
anthropophilic snails in the genera Subulina,
Lamellaxis, Succinea, and Pleuropoma.
Layer IB: The very dark gray (10 YR 3/1) sandy
loam here is a cultural midden deposit. Parent
materal consists primarily of calcareous sand, with a
minor component of fine clay-silt. The dar1k gray
color results from the inclusion of much finely
dispersed charcoal and other organic material. The
layer contains much shell midden (9.4 kg) and some
dispersed 'ili'ili gravel. The contact with Layer IC is
gradational.
Layer IC: The very daik grayish brown (10 YR 3/
2) calcareous sand in this stratum has some fine
charoal flecking. The layer appears to represent an
Exavataons
UNIT 17: WEST FACE
0.0
* :. I X .
- ;
;
f...','0,.- 0~~~~~~~~~~~
0.5
WBWA
0
1.0 c
._
0
a
Branch Coral
*
VI.
1.5
Coral CoWes
.
' '
*
Figure 5.12
*. .
VII..~~~~*..
2.0
Stratigraphic section of the west face of
Unit 17.
old stable surface (paleosol A horizon) of the former
beach ridge. The surface is somewhat disturbed and
rewoiked by the overlying Layer IB occupation.
The contact with Layer II is irnegular, and root casts
penetat from Layer IC into Layer II, indicative of
vegetation on the former stable surface.
Layer II: This white (10 YR 8/2), massive, stucureless deposit of very fine-grained calcareous sand
incorporates some fine volcanic lithic fragments. At
the base of Layer II is a zone of fist- to head-sized,
waterwom coral cobbles indicative of a high-energy
stonn event. The contact with Layer Ill is very
unclear and gradational.
Layer m: This is a very pale brown (10 YR 7/3),
63
extremely non-concentrated culturl deposit. It
yielded a few sherds during excavation, and one
sherd was visible in situ in the west wail during
profile recording (see figure 5.14). Several volcanic
fire-altered oven stones were also noted in this layer.
Except for its slightly darker color, this deposit is
lithologically hardly distinguishable from Layer II.
It probably represents the old beach slope fronting
either the Layer IIB or Layer UD occupadons in Unit
15. The lower contact with Layer IV is gradational
and unclear.
Layer IV: The white (10 YR 8/2) fine grained
calcareous sand in this deposit contains volcanic
lithic grains ('salt-and-pepper' sand).
Unit 16 produced considerable culual material,
including 741 fish bones, as well as rat, turtle, and
bird bones. Shell midden was most heavily concentrated in Layer I (9.4 kg), with lesser quantities in the
deeper layers. Full details of the faunal analysis are
presented in chapter 13.
Unit 16 yielded a worked Tridacna shell
(possibly an adz preform) and a shell scraper from
Layer IB. Artifacts from Layer III include a shell
bracelet fragment, a Turbo shell fishhook fragrnent,
obsidian (including a red and black banded specimen), and basalt flakes. In all, forty-four sherds was
recovered from Unit 16. These include twelve
thickware sherds in Layers I and H, one of which (in
Layer I) has a parallel-rib, paddle-impressed exterior
surface. Layer III yielded twelve (37%) thinware
sherds and twenty (63%) thickware sherds.
Excavation Units 15129130
As indicated above, Unit 15 was excavated
along with Units 16 and 17 on Trnsect 5. When a
deeply buried, thin cultural deposit containing thin,
red-slipped pottery was encountered, we decided to
expand Unit 15 by excavating two adjacent squares,
designated Units 29 and 30. The straigraphy
described below, and shown in.figure 5.15, is that of
the west face of Unit 15.
Layer IA: The upper 10 cm of Layer I is composed
of dark reddish brown (2.5 YR 3/4) colluvium. This
upper zone is slightly humic, with an A soil horizon
at the surface. The deposit is a poorly sorted, very
rubbly mixture of volcanic sand and clay, full of
angular to subangular gravel and small cobbles.
This deposit is the tongue of a colluvial fan emanating
64
The To'aga Site
UNIT 17
I
w
N
E
S
I
0.0
0.5
co
0
0
0
1.0 c
~
~
~
~
Po tm ld
c
1.5
VI
2.0
V.
Figure 5.13
Exploded stratigraphic section of all faces of Unit 17, showing postmolds and hearth associated with
Layer II.
from a small hanging valley above the Le'olo cliff.
Layer IA grades into IB.
Layer IB: Although almost identical to Layer IA,
this layer lacks the organic, humic component. The
lower 5-10 cm of IB is noticeably more clayey, with
a lower frequency of larger clastics.
Layer II: This 'greasy' midden deposit is a very
dark gray (5 YR 3/1). The parent material is primarily calcareous sand with some silt-clay admixture.
The dark color and greasy textu
result from the
incorporation of much finely dispersed ash, charcoal,
and other organic material. The deposit contains
abundant fist-sized, volcanic, fire-altered oven
stones. Shell midden is also present (3 kg) but is
somewhat chalky and chemically degraded. The
contact with Layer IB is straight and abrupt; that
with Layer lIIA-I is irregular, varying from abrupt to
gradational. The high carbon content and the
presence of oven stones suggest that this layer
represents a cookhouse activity area.
Layer IIA-1: The reddish-yellow (7.5 YR 6/6)
calcareous sand of this deposit has pockets of
pinkish-gray sand (7.5 YR 6/2). In addition to
considerable mottling, various irregulanties ad
disturbaces appantly resulted from the Layer
occupation on an old, stabilized beach ridge surface
represented by this layer. Root casts extend from
IIIA-l into iIIA. The contact with Layer IIIA is
shaip but highly irregular. This layer is interpreted
as a formerly vegetated, stabilized dune surface
(paleosol).
Layer IIA: This sterile deposit of white (10 YR 8/
2) calcareous sand is well-sorted with a mediumsized (2 phi) mode and has a minor admixture of
volcanic lithic grains ('salt-and-pepper'). The
contact with Layer IUB is very sharp but highly
irregular ('wiggly').
Layer IUB: This pale brown (10 YR 6/3), very
non-concentrated occupation deposit is indicated
primarily by its slightly darker color and the
Excavations
UNIT 16: WEST FACE
...S
,.IB:
0.0
r
Voicanlci Boule
0.5
IC
.
I
*
*
.
.U
Root Mold
II1.0
to
0
._
a
0
Coral Cobbles
a
1.5
(9
Vocanic Oven Stones
*
IV*
.
2.0
6
Iv.
Water Table
2.5
Figure 5.14
Stratigraphic section of the west face
of Unit 16. Black squares indicate
sediment sampling locations.
Thinware
Layer II
Layer IIIA
Layer IIIB
Layer IIID
3
8
7
19
5%
24%
13%
38%
65
presence of some marine shell midden. The
deposit is medium- to coarse-grained, and less
well-sorted than Layer IIIA. The lower contact
with Layer IIIC is irregular and slightly more
gradational than the top contact. Lithologically,
this is the same as Layer IIIA.
Layer mc: The white (10 YR 8(2). calcareous
sand of this stratum is essentially identical to
Layer IIIA but contains a higher frequency of
coarse grains (I phi).
Layer IUD: In Unit 15, this deposit appeared
only discontinuously, as pockets in the west face
of the section. These were pale brown (10 YR 6/
3), very similar to the Layer IIIB deposit. In the
east face of Unit 15, the zone was continuous.
This is the deposit that yielded thin, fine-tempered, red-slipped pottery, together with
thickware, during excavation.
Layer IV: The coarse- to very coarse-grained,
white (10 YR 8(2) calcareous sand that is found
here is poorly sorted, having a considerable
mixture of volcanic lithic grains ('salt-andpepper' appearance) and some small volcanic
pebbles. The deposit also contains considerable quantities of branch coral fingers (water
rolled) and coral pebbles. The presence of these
larger grained sediments and the lack of sorting
suggest that this was a relatively high energy, exposed beach depositional environment.
Excavation of Units 15(29/30 yielded an
of
array artifacts that include two shell fishhook
tabs (Layers II and IIID), cut pearlshell (Layer
II), a shell bracelet fragment (Layer IIIB), an
unfinished fishhook (Layer IIIB), and numerous
flakes of basalt and obsidian (Layers II, IIIB, and
IIID).
A total of 200 sherds was recovered from
these units. The distribution of these sherds by
layer was as follows:
Thickware
58 95%
26 76%
48 87%
31 62%
Total
61
34
55
50
The To'aga Site
66
UNIT 15: WEST FACE
.Trkdacnai
,'Oven
.
0
,
Stones
.I' A.-
U/
_L~fB>
!
mc.
0
.IV
v
-k
Figure 5.15
v
Q
M.
0
,
Stratigraphic section of the west face
of Unit 15. Black squares indicate
sediment sampling locations.
Two thickware direct rims from
Layer HIA have
impressed lips, which appear to have been made
with a carved paddle. Red slip is resticted to
thinware (2 sherds only) from Layer II). As the
frequencies above indicate, there is a generl trend
toward increasing relative frequency of thickware in
proporton to thinware over time in this unit.
Excavation Unit 28
Unit 28 was excavated 10 m farther inland from
Unit 15 in an effort to trace the deep Layer IIID
occupation in Unit 15, inland under te steeply nsing
colluvial fan. Clearly, the depth of colluvial
overburden which would have to be removed was
substantial, making this the most inland position in
which excavation could be attempted witout the use
of heavy machinery. The stratigraphy described
below is that of the norh face.
Layer I: This massive deposit of dark reddishbrown (25 YR 312) clayey colluvium is between
2.2-2.4 m thick and contains numerous, large,
angular volcanic cobbles and boulders. No bedding
or lenses were present, although clay content
increases toward the base of the stratum. Free of
cultural material, this colluvial deposit represents
one or more events of mass wasting (landslides)
from the adjacent volcanic cliffs.
Layer HA: This discontinuous zone (pockets) of
calcareous sand has been discolored (7.5 YR 6/6) by
eluviation of clay from the colluvium above (Layer
I). A small amount of cultural material was recovered, including pottery, shell and bone midden, and
charcoal.
Layer TIB: The white (10 YR 812) calcareous sand
in this stratum has waterwom coral cobbles and a
few coral boulders. This layer contains a minor
admixture of volcanic sand grains ('salt-and-pepper'
lithology) as well as cultural material.
Layer IC: The pale brown (10 YR 6/3) calcareous
sand of Layer HC is similar to IIB in origin, but is
distinguished by a darker color explained by te
addition of organic matter and some clay, resulting
from human occupation. Its relatively abundant
cultural material includes midden ad pottery (see
below). Layer IIC of Unit 28 is comparable to Layer
IlD of Units 15129/30.
Layer mI: This layer is a white (10 YR 812). coarsegrained calcareous sand with abundant waterwom
coral cobbles and fresh (unwom) brnch coral
fingers. Layer III yielded less cultural material than
IIC and clearly is culurally sterile at its lower depths
(3.3-3.4 m below surface).
Artifacts from Unit 28 came primarily from
Layer IIC. These include worked shell, basalt flakes,
obsidian flakes, an echinoid-spine abrader, and duee
Cypraea shell dorsa found artificially nested together, which may be parts of an octopus lure
apparatus (see Kirch, chapter 1 1).
A total of 127 sherds was recovered, with the
following stratigraphic distribution:
Excavatio
Thinware
Layer IIA
Layer IIB
Layer IIC
7
51
30%
49%
decline in thinware in proportion to thickware
the overll decrease in
density of ceramics. Included in the above sherd
counts are two thinware rims from Layer IIB, and
nine thinware and five thickware rms from UC.
One of the thickware rims has an impressed lip.
Eight thinware sherds, all from Layer IIC, carry a red
or orange slip. Since the ceramic assemblage from
Layer HB dates to 1308-930 cal B.C. (Beta-35601),
Layer UC must be of equal or possibly greater
antiquity. This suggests that in the Manu'a Islands,
te ceramic trnsition from Early Eastem Lapita
(calTying dentate-stamped decoration) to thin,
plainware must have occurred relatively rapidly at
the end of the second millennium B.C.
The
over time is notable, as is
Radiocarbon Dates from Transect S
Three 14C age deterninations were obtained
from Transect 5 samples. Ihe oldest date (Beta35601) comes from charcoal at the Layer IVIll
interface in Unit 28, associated with the early, thin,
red-slipped pottery. This sanple has an age of 32572879 cal B.P. at one standard deviation. A sample of
marine shell midden (Turbo setosus) from a
cookhouse activity zone, Layer II in Unit 15, is
associated with thick, coarse-tempered pottery. This
yielded an age of 1631-1477 cal B.P. at one standard
deviation (Beta-35924). The third sample (Beta35600) was obtained from the Layer Ill 'ili'ili
pavement in Unit 17. This sample, which had no
ceramics associated, yielded an age of 1256-1007 cal
B.P. at one standard deviation. Further details of
these radiocarbon dates and their implications are
provided in chapter 6.
Transect5: Sunmary of the Depositional Sequence
The correlation diagram of the Transect 5
excavation units (fig. 5.16) also shows the relationship of these units to contemporary shoreline and sea
level features. The general depositional sequence
along Transect 5 can be reconstructed as follows:
Stage 1 (> 3200-2800 cal B.P.): A narrow coastal
Thickware
100%
1
16
70%
52
51%
67
Total
1
23
103
bench at the base of the steep talus was fomied. The
active shoreline at this time would have been in the
vicinity of Units 16-17, considerably inland of the
modem shoreline. The 'salt-and-pepper' lithology
of the beach ridge sediments indicates exposure of
volcanic headlands along the coasline, providing a
source of volcanic lithic grains. In addition, the
presence of larger clastics (coral cobbles, branch
coral fingers) indicates a fairly high energy shoreline.
Stage 2 (ca. 3200-2800 cal B.P.): Humans began to
1,
resulting in non-concentrated midden deposits that
contain thin, fine-tempered, orange- or red-slipped
pottery in Units 28 and 15129/30. The main area of
occupation was probably farther inland from Unit
28, and thus is now deeply bured under talus
rockfall and colluvium. The deposits exposed in
Units 28 and 15/29/30 appear to represent the
seaward periphery of such an occupation, down the
slope of the former beach ridge toward the old
shoreline. Archaeological exposure of the putative
main occupation zone would require the use of
heavy machinery, since as much as 5-15 m
overburden of boulder talus and colluvium would
probably have to be removed.
Stage 3 (ca. 2800-2000 cal B.P.?): Deposition of
calcareous sands onto the beach ridge continued,
with significant seaward progradation of the shoreline occuning late in this phase which resulted in the
deposition of the basal deposits in Units 16 and 17.
During this stage, a second occupation phase
resulted in the midden deposit of Layer IIIB in Units
occupy the narrw bench formed during Stage
15129/30.
Stage 4 (ca. 2000-1600 cal B.P. ?): A stabilized
land surface formed during these four centuries
over the now wider and prograded coastal terrace,
marked by the paleosol horizon (Layer IIIA-I in
Units 15/29/30; Layer IC in Unit 16; Layer IVB
in Unit 17).
Stage SA (ca. 1600-1400 cal B.P.): The stabilized
coastal terrace in the vicinity of Units 16 and 15/29/
30 was occupied during this terminal phase of
ceramic manufacture and use on Ofu Island. The
68
The To'aga Site
UNIT 15
UNIT 16
UNIT 17
0.0
-
0.5
1.0
1.5
e
0
0
2
2.0 *r
2.5
3.0
3.5
Figure 5.16
Stratigraphic correlations between Units 15, 16, and 17 on Transect 5.
Layer II deposit in Units 15129/30 is interpreted as a
cookhouse activity area, while the contemporaneous
Layer lB in Unit 16 is a concentrated midden.
Stage SB (ca. 1300-1000 cal B.P.): Acermic
occupation on the coastal terrace in the vicinity of
Unit 17 resulted in the construction of a low house
mound fonned by several successive gravel ('ili'ili)
pavements.
Stage 6 (< 1000 cal B.P.): A tongue of clay-silt
colluvium was deposited out onto the coastal ten-ace,
probably due to increased up-slope forest clearance,
agriculural acdvity, and subsequent erosion. At this
time the coastal terrace was used for tree-cropping
and shiffing cultivadon, continuing into te present
era.
stratigraphic sequence along Trnsect 5 and
its corelation to modem sea level, as well as the key
topographic features of the contemporary coastline
(reef flat, active beach ridge) also provide key
evidence for testing the morphodynamic model
developed in chapter 4. As indicated in figure 5.16,
the oldest cultural deposit at TransectS is the Layer
IIID occupation in Unit 15, yielding thin, finetempered, orange-slipped pottery, between 180-200
cm below surface. In Unit 28, this occupation
appeared at the base of Layer II, between 290-300
cm below surface, and was 14C dated to 3257-2879
The
cal B.P. When corelated to modem sea level, this
occupation zone lies at virtually the same elevation
as mean high tide and about 1.8 m above the reef
flat. These elevation relationships provide incontrovertible evidence for tectonic subsidence, as the in
situ cultural matenials in Units 15 and 28 were
cleatly deposited on a narrow terrace or beach ridge
that must have been at least 1 m, and more likely 2
m, above te sea level at 3 kyr B.P. Given a +1-2 m
high sea level stand at 3 kyr B.P., this means ftat fte
To'aga site has undergone between 2-3 m of tectonic
subsidence over fte past three thousand years, as
suggested by the morphodynamic model developed
in chapter 4. In other words, if we retrodict the
elevation of te early beach-ridge occupation at 3
kyr B.P. to +3 m (based on our hypothesized subsidence rate of 1 m/kyr), this would put the occupation
surface between 1-2 m above the 3 kyr B.P. sea level
(which itself was +1-2 m above the present level).
The altemative hypothesis-that there was no
tectonic subsidenc-would require the deposition
of the occupation deposit under water, a physical
impossibility given the sedimentological evidence.
In short, the Transect 5 stratigraphic profile strongly
confilrs several key elements of the morphodynamic model of coastal terrace fonration developed in chapter 4.
Excavation
Transect 7
Transect 7 is situated 300 m east of the 1987
excavation area. The coastal terrace here is fairly
narrow, only about 65 m from the crest of the active
beach ridge to the base of the steep talus slope. The
elevation profile along Transect 7 is shown in figure
5.17 (a vegetation transect is given in chapter 2,
fig. 2.5). The area was primarily in banana
gardens under coconut and breadfruit trees in
1989.
Two tests were excavated along this transect, 10
m aparL Unit 18 was located at the base ofthe talus
slope, while Unit 19 was situated on the flat depression in the center of the arboricultural zone. Unit 18
was excavated to 1.4 m below surface, at which
depth the presence of large basalt boulders forced us
to tenninate the excavation. Unit 19, on the other
hand, was taken to a depth of 2.2 m, and a thin, nonconcentrated cultural deposit was exposed near the
base.
Excavadon Unit 18
The following strata were discemible in Unit 18:
69
Layer IA: This massive deposit of colluvial clay
has numerous inclusions of subangular to angular
gravel and pebbles. It is weak red to dusky red
(2.5 YR 3-412) and very compact and sticky.
Layer IB: This layer contains the same material as
Layer IA, but with an admixture of snall quantities
of calcareous sand grains. Some charcoal flecking
(arge chunks) was observed in a 10 cm thick zone
between Layers IA and lB. The boundary between
IA and IB is gradational over this zone of charcoal
flecking. The contact with underlying Layer II is
diffuse and inregular, over a 2-3 cm zone.
Layer II: Fie-grained calcareous sand with some
clay admixture due to eluviation from Layer IB
makes up Layer l. This pale brown (10 YR 6/3) layer
incorpomtes massive volcanic cobbles and boudders
which forced a temiination ofthe excavation.
Excavation Unit 19
Unit 19 was positioned 45 m inland of the
coastal road, in a relatively level area, sunrounded by
a few large talus boulders. The stratigraphy of the
west face is described below and is illustrated in
figure 5.18:
6
w
0.
4
c w
co
CIO
a
Co CC
W
D
E
Cu
2
O2
<
co
IV,,
0)
9-
:L-l
a) I-1
4-0
c
c
:3
111
0
c
0 -2
-m
-4
-6J
w
w
cr.
10
0
20m
TRANSECT 7
I
Figure 5.17
Elevation profile along Transect 7, showing the locations of Units 18 and 19 in relation to geomorphic
features.
The To'aga Site
70
UNIT 19: WEST FACE
.
.~~IB
InIY
Turtio
~
VI
Figure 5.18 Stratigraphic section of the west face of
Unit 19. Black squares indicate sediment
sampling locations.
Layer IA: This dark reddish brown (5 YR 3/2),
poorly sorted colluvial deposit has some subangular
volcanic pebbles, but the general absence of gravel
suggests deposidon of the sand to clay-sized particles by sheet wash at the margin of a colluvial fan.
Layer IA grades into IB over a 5 cm zone.
Layer IB: This strum contains reddish-brown (5
YR 5/3), fine clay mixed with calcareous sand. The
calcareous sand component increases with depth.
The contact with Layer II is sharp and only slightly
irregular.
Layer U: The very pale brown (10 YR 7/3), very
fine-grained calcareous sand of this layer is cultur-
ally stenle. This deposit appears to represent a faidy
long period of beach ridge stability, probably under
vegetation. (The very pale brown color is virtually
identical to that of beach ridge deposits under
Pannus and Hibiscus just inland of the modem
beach.) The contact with Layer III is diffuse and
irregular.
Layer IH: This stratulm is composed of pale brown
(10 YR 6/3), medium- to fine-grained calcareous
sand which is well-sorted with scattered coral and
volcanic pebbles and cobbles. Layer III is barely
distinguishable from Layers n and IVA. The contact
with Layer IVA is very irregular. Layer IIn appears
to represent another phase of beach ridge stability.
Layer IVA: The white (10 YR 8/2), medium- to
fine-grained calcareous sand of this layer is wellsorted, lacks larger inclusions, and is cultually
sterile.
Layer IVB: This layer with its very pale brown (10
YR 7/3), medium- to fine-grained calcareous sand,
probably represents a period of beach ridge stability.
Layer V: The color of this cultural deposit ranges
from dark grayish-brown (10 YR 4/2) to grayishbrown (10 YR 512). This non-concentrated midden
consists of medium- to fine-grained calcareous sand
stained with charcoal and ash. There are considerable quantities of dispersed charcoal flecks and
charred wood fragments, some quite large (5-10 cm
diameter). A Turbo shell fishhook was excavated
from Layer V at 164 cm below surface. No pottery
was present The contacts with Layer IVB and with
VI are both quite irregular but fairly distict.
Layer VI: This layer contains white (10 YR 8/2),
coarse- to fine-grained calcareous sand and is
cultally sterile. Excavation was terminated at 220
cm below surface. An auger was then used to core
to a depth of 385 cm below surface, at which point
the water table was encountered. No cultual
materials were observed in the augured sediments.
Unit 19 was notable also for the presence of a
large pit, sectioned in the east and south faces of the
square, as shown in figure 5.19. The pit, which was
cut from Layer I down trugh Layers II and HI into
Layer IV, has a mixed fill incorporating Layer IB
sediment. The pit is straight-sided, with a flat base.
There is no evidence of burning or use as an oven
pit, nor was the feature filled with midden or trash.
In shape and size, the pit is consistent with Iua'i
masi, subterranea silos for the fernentation and
Excavation
71
UNIT 19
I
E
I
S
I
0.0
0
L.
0
Figure 5.19
Stratigraphic section of
the east and south faces
of Unit 19, showing the
large pit cut from Layer I
into Layers II-IV.
0
-1.0 *E
0
-2.0
storage of breadfruit paste or masi (Buck 1930:13233; Kirch 1984:132-35; Cox 1980). This function is
also suggested by the presence of large volcanic
cobbles in the pit fill; such cobbles are used to cap
lua'i masi pits after filling with breadfruit. The
To'aga area is dotted with shallow circular depressions, ranging from 1-2 m in diameter, which our
infonmants described as fonner masi pits.
Unit 19 produced a modest faunal sample (see
Nagaoka, chapter 13), and a Turbo shell fishhook
from Layer V. No other arfifacts or pottery were
present. Although charcoal and Turbo shell samples
were collected from Layer V for radiocarbon dating,
these have not been processed due to budgetary
limitations. We are thus uncertain wheter te
absence of pottery in Layer V reflects sampling
enror, or whether this deposit post-dates the cessation
of pottery use on Ofu (i.e., after about 1500 cal B.P.).
Transect 9
Transect 9 was positioned 500 m east of the
baseline origin point. Inland of the road, the coastal
tenrace here was in dense banana plantations under
breadfruit and coconut. No surface archaeological
features were evident, other thwa a filled-in lua'i
masi pit (1.5 m diameter, 30 cm deep depression)
between Units 21 and 22. The elevation profile
along Trnsect 9 is shown in figure 5.20.
We initially laid out three units for subsurface
sampling along Transect 9: Units 20, 21, and 22,
situated at 10 m intervals. Unit 20 was posidoned at
the base ofthe talus slope, with Units 21 and 22 on
the flat between the talus and the road. Unit 20
penentated a thick series of cultual deposits between
about 1-2.6 m below surface, yielding an aray of
thin, fine-tempered ceramics (some with notched
iims) and thickware, as well as fistooks, onaman,
and odter antifacts. We therefore decided t expand this
test in order to enlarge our sample of the comparatively rare cultual materals recovered from these
deposits. Unit 23 was dterefore opened adjacent to
Unit 20 on the north (inland) side (fig. 5.21). Thus a
total of 4 m2 was excavated along Transect 9.
The To'aga Site
72
6
C
0
1C
-
cO
2
2
0c
'IW
I-0
C
-
2
co
-L
L
F-J
w
w
Cc
-4
-6
Figure 5.20
\lc
u
1U
20m
rI,T A XTOVrP A
:ia
11AtIobrikl
Elevation profile along Transect 9, showing the locations of Units 20/23, 21, and 22 in relation to
geomorphic features.
Excavation Units 20/23
The stratigraphy of this 1x2 m excavation was
recorded along the west and north faces, as depicted
in figure 5.22, and described below:
Layer IA: This stratum consists of dusky red (2.5
YR 312) colluvium; the upper 10 cm is humic garden
soil. The very poorly sorted, compact, sticky, very
plastic sand-to-clay material has abundant
subangular volcanic gravel- to cobble-sized inclusions. The contact with Layer IB is gradational.
Layer IB: This dark reddish-brown (5 YR 3/2)
massive colluvial deposit is poorly sorted and has
abundant larger volcanic clastics. Calcareous sand
grains and occasional charcoal flecks are dispersed
th}oughout. The deposit also contains a low frequency of dispersed, chalky shell midden (due to
chemical decomposition in the humic soil matrix)
and occasional waterwom MiMi'ili coral pebbles.
These cultural materials suggest some low intensity
use of the area during the perod of colluvial deposition. This deposition was presumably gradual and
incremental, and not a single event. The contact
with Layer IC is gradational over a 2 cm zone.
Layer IC: The dark reddish-brown (5 YR 3/3),
compact, sticky silty-clay of this stratum lacks larger
pebble-sized inclusions. The clay appears to represent a single depositional event, presumably result-
ing from sheet wash (fluvial mode of deposition)
following heavy rains. The contact with Layer IIA is
abrupt and slightly irregular.
Layer IHA: This stratum is composed of dark
reddish-brown to reddish-brown (5 YR 34/4),
coarse- to fine-grained, poody sorted, calcareous
sand mixed with fine clay-silt and is marked by
abundant charcoal flecking. This appears to represent a paleosol or fonmer stable surface of the beach
ridge. The contact with Layer IIB is gradational.
Layer UB: This layer contains yellowish-red (5 YR
5/6) to reddish-yellow (7.5 YR 7/6) well-sorted,
medium- to coarse-grained, structureless calcareous
sand mixed with reddish clay and is culturally sterile.
The clay content decreases with depth, with the base
of the layer consisting of nearly pure, mediumgrained sand. The contact with Layer IIIA is abrupt
and irregular.
Layer MA: This cultural midden deposit is very
daik gray to very dark grayish-brown (10 YR 3/1-2)
and has a sandy loam texture. The dark coloration of
the calcareous sand matrix results from a high
carbon and organic content. This layer was associated with a large earth oven feature, from which a
14C sample was obtained (see below). The oven and
adjacent sediment contained heavy concentrations
(more than 2 kg) of the spines of the large slatepencil sea urchin (Heterocentrotus mamnillatus) and
Excavations
Figure 5.21
73
Expansion of Unit 20 into Unit 23; removal
of the thick colluvial deposit in progress.
of fire-altered volcanic oven stones. This stratum
clearly represents a cookhouse or food preparation
activity area The contact with Layer IIIB is gradational.
Layer IUB: In this layer, a dark brown to brown
(10 YR 4/3), thick, cultural midden deposit of
calcareous sandy loam incorporates numerous, large,
subangular, volcanic cobbles. Near the top of the
layer is a distinct lens of compact grayish-white ash,
about 1 cm thick. This ash lens was discontinuous
over the lx 2m excavation unit. The ash lens is
capped by 1-2 cm of clean white beach sand,
probably an artificially deposited house floor. Layer
ilB contned a heavy concentration of shell and bone
midden, ceramics, and other portable artifacts. The
contact with Layer IIIC is gradational over 5-10 cm.
Layer mC: Subangular volcanic pebbles and
cobbles are contained in the light yellowish-brown
(10 YR 6/4) calcareous sand midden of this stratum.
The matrix consists of coarse- to medium-grained
calcareous sand. This layer is essentially a lower
facies of Layer IIIB, distinguished primarily by a
lower concentration of midden and organic staining.
The contact with Layer IV is gradational.
Layer IV: The white (10 YR 8/2), coarse- to
medium-grained, 'salt-and-pepper' sand (calcareous
grains with an admixture of volcanic liftic grains) of
Layer IV contains abundant large clastics consisting
of branch coral fingers, and waterwom volcanic
pebbles and gravel. It is culturally sterile.
An earth oven was exposed in Layer IIIA of
Unit 23, as noted above, and is illustrated in figure
74
The To'aga Site
UNIT 20
UNIT 23
w
UNIT 23
N
w
IA
I
X
I
1
1
0.0
1v
1
IN
0
0
0
LBE
I
K'
K'~~~~~~~Ah
K
1.0 c
en
2.0
v~~~~~~
.2.0
C3~~~~
Figure 5.22
Stratigraphic section of the west and north faces of Units 20f23. Black squares indicate sediment
sampling locations. "C" indicates coral; "V" indicates volcanic.
5.23. The oven consisted of a circular pit, with a
diameter in excess of 80 cm, and a depth of 38 cm
from the rim to the base. The
oven
pit
fill consisted
of fire-cracked volcanic oven stones (fist-sized) with
a few larger stones at the base, charcoal, and ash.
Both the oven fill and the sunTounding midden
contained large quantities of echinodenm spines and
Layer IIB
Layer IIIA
Layer IIIB
Layer IIIC
Thinware
5
10%
8%
8
50 21%
32 28%
test fragments, particularly of the slate-pencil sea
urchin (Heterocentrotus mammllats), and several
smaller echinoid species.
Units 20/23 yielded a large sample of pottery
and a number of noteworthy portable artifacts. A
total of 499 sherds was recovered, with the following
stratigraphic distribution:
Thickware
45 90%
87 92%
191 79%
81 72%
Total
50
95
241
113
Excavations
Figure 5.23
75
View into Units 20/23, showing circular pit or
postmold features filled with Layer IIB sediment, exposed at the contact with Layer IIIC.
These frequencies confinn a decrease in thinware
relative to thickware also noted from other tmnsects
(see above). The sherd assemblage from Units 20/23
includes one Wtckware rim from Layer IIB, six
thickware rims and one thinware rim from Layer
hIA, twelve thickware and six thinware rims from
Layer HIB, and five thinware rims from Layer HIC.
None of the thickware rims have impressed lips, and
all were recovered from Layer IIIB. Of 499 sherds,
15 are red-slipped (on thinware only) and were
recovered from Layers IIIA, IIIB, and IIIC. One
small thinware body sherd from Layer IIIC has
incised lines on the exterior, but no pattem is
discemible.
Several Turbo-shell fishhooks were recovered
from Layer III, including one with a distinctly bent
shank similar to early Eastem Polynesian fonns (see
Kirch, chapter 1 1 for further discussion). The butt of
a plano-convex section adz (Samoan Type V) was
excavated from Layer IIIB. Layer IIIC produced an
unfinished ring of Tridacna shell which evidently
broke in half during the process of manufacture. At
the contact of Layers IIIB and IIIC, we also recovered an abrading stone designed for the manufacture
of small Conus-shell beads. All of these finds are
further described and illustrated in chapter 11.
Excavation Unit 21
Unit 21 was positioned 10 m seaward of Units
20/23, at essentially the saime elevation above sea
level. The profile of the south face is illustrated in
76
The To'aga Site
figure 5.24, and described below:
Layer IA: This dusky red to weak red (2.5 YR 34/
2) massive colluvial deposit is pnimanly clay with
some large volcanic cobbles and talus boulders and
some smaller subangular volcanic inclusions. It is
very sticky and plastic. Small quantities of chalky
shell midden are dispersed thrughout the deposit.
The contact with Layer IB is very gradational over
10-15 cm.
Layer IB: In this mixture of fine clay and calcareous sand, the sand content increases with depth.
Some shell midden was noted near the top of the
deposit, but the layer becomes culturally sterle near
the bottom. There are many discretely dispersed
flecks and chunks of wood charcoal, probably
deposited with the clay, and suggestive of forest
clearance/burning up slope. The contact with Layer
IIA is quite irregular and somewhat diffuse.
Layer HA: This stratum is composed of very pale
brown (10 YR 7/3) calcareous sand. There is a
minor lithological component of volcanic lithic
grains.
Layer UB: The sandy midden deposit of Layer IIB
is grayish-brown to dark grayish-brown (10 YR 4-5/
2). Contacts with both Layers IIA and IIC are
irregular (wavy) but fairly distinct. This deposit
contains large quantities of Heterocentrotus
mamillas spines, some fire-altered volcanic
stones, shell midden, bone, and ceramic sherds. This
layer appears to correlate with the thick Layer III
midden deposit in Units 20/23, being a seaward
'pinching out' of this occupation zone in the direction of the former beach slope.
Layer IIC: The white (10 YR 8/2), fine-grained,
calcareous sand of this stratum has a minor component of volcanic lithic grains. The layer is culturally
sterile and represents a beach slope depositional
environment.
The Layer IIB midden yielded a large faunal
assemblage, including 13.1 kg of marine shell, 1,803
fish bones (including a large number of spines of the
puffer fish Diodon hystrix), rat, bird, and turtle (see
Nagaoka, chapter 13 for further details).
Unit 21 yielded several artifacts of worked shell
and a shell omnament (see Kirch, chapter 11), all
from Layer IIB. Pottery from IIB includes three
(15%) thinware and 17 (85%) thickware sherds. No
rims, decorated sherds, or red-slipped sherds were
present.
UNIT 21: S FACE
0.0
I7IhIA
nic Slab
Turb
l
IA
/+I -ItE
2 i/+
-0.5
1.0
to
Volcanic
f~~~/,
:7|
I._
0
0.
0
-1.5
ImA
HA .
.ER.
-2.0
Oven Stones
0
* A2
uc
2.5
Coral
,:~
.~~~~
K-
Figure 5.24
Stratigraphic section of the south face
of Unit 21. Black squares indicate
sediment sampling locations.
Excavation Unit 22
Unit 22 was positioned 10 m further seaward
from Unit 21, and at a slightly higher ground
Excavation
elevation (on the inland slope of the present beach
ridge). The noith face of Unit 22 is illustated in
figure 5.25 and described below:
Layer I: The very daik grayish-brown (10 YR 3/2),
sandy loam of this sm has a 'greasy' texture.
The deposit consists primarily of calcareous sand
with a minor component of fine silt-clay. Dispersed
trughout is chalky shell midden and some 'ili'ili
gravel. The deposit becomes daiker toward the base,
with increasing charcoal and ash content. The
contact with Layer II is diffuse and irregular.
Layer U: Considerable coral and basalt waterwom
'ili'ili gravel is dispersed throughout this brown (7.5
YR 5/4) sandy midden. There is some charcoal
flecking, especially in the bottom of a shallow pit
feature at the base of the deposit. The contact with
Layer III is sharp but irregular.
Layer m: The white (10 YR 8/2), medium-grained,
calcareous beach sand of this stratum with its minor
component of volcanic lithic grains is culturally
sterile.
77
No ceramics were present in this unit.
Radiocarbon Dates from Transect 9
Three 14C age detenminations were obtained
from Transect 9, all on samples excavated from Unit
23. Charcoal from the large earth oven in Layer IIA
(Beta-35602) yielded an age range of 2845-2612 cal
B.P. at one standard deviation. This sample was in
direct association with a ceramic assemblage
composed of 8% thinware and 92% tickware. A
second sample (Beta-35603) of dispersed charcoal
flecks from d sratigraphically older Layer IIB
cultual deposit yielded an age range of 2917-2382
Cal B.P. at one stadard deviation. A large single
valve of Tridacna maxima shell (Beta-35604) from
Layer IIlB was also dated, yielding an age range of
2444-2289 cal B.P. at one standard deviation. All
dte samples from Unit 23 ovetlap at one standard
deviation, suggesting that the deposition of Layers
IIIB and IIIA occurred fairly rapidly, possibly over a
span of only one or two hundred years. Furter
details on all samples are provided in chapter 6.
UNIT 22: N FACE
rt.r~F-.--
. 1..r.
nic
Swnmary of the Transect 9 Depositional Sequence
I 0.0
ni
V
16
c
Tur60
olcani V~~~~
0.5
.SCharcoal
*
Concentration
O)
*
. ~
.
>.1..
.
~
%
.
.=
.
1.0
. . b
Charcoal
mI
Figure 5.25
1.5
Stratigraphic section of the no rth face
of Unit 22.
Figure 5.26 is a correlation diagram of the strata
exposed in the Transect 9 excavation units. This
diagram also shows the elevational relationships
between these strata and the modem sea level and
reef-shoreline features. Evidence from the
Transect 9 excavation data suggests the following
depositional sequence:
Stage 1 (> 2800 Cal B.P.): A narrow coastal terrace
or bench with a mixed calcareous-volcanic lithic
sand lithology was fonned. At this time the active
shoreline was probably in the vicinity of Units 2122.
Stage 2 (ca 2800-2300 cal B.P.): The occupaion of
the nanow coastal terrace formed in Stage 1 resulted
in the accumulation of the thick pottery-bearing
Layer III strata in Units 20/23. The Layer IIB
deposit in Unit 21 represents a seaward diminution
of this occupation down the former beach slope.
The shoreline itself was presumably no more thm
10-20 m funher seaward of Unit 21 at this stage.
Stage 3 (ca 2300-1900 cal B.P. ?): The rapid
progradation of the coastal terrace was initiated by
an increase in the biogenic/calcareous sediment
The To'aga Site
78
0
&fi
v-
CNJ
CN4
ci
,Higher Berm
IVegetation
Line
2
o2
-1
-2
Zone
Ceramic Zone
-
Post- 1900 BP Progradation
e
tv
¢
Approx. Shoreline
at 3000 BP
-3
Figure 5.26
Sigraphic correlations between Units 20/23, 21, and 22 shown in relation to a schematic elevation
Transect 9. The horizontal dimension along this transect is schematic and arbitrary, while
the vertical dimension is expressed in meters above and below mean high water.
profile of
budget. Ihis resulted in the sterile calcareous Layer
UB sand deposit in Units 20q23 (conrelated with
Layer HA in Unit 21 and Layer HI in Unit 22),
capping the earlier pottery-bearing occupation.
During this stage, the shoreline prograded substantially, and active accumulation of calcareous sediments in the vicinity of Units 20-21-22 subsequently
ceased.
Stage 4 (ca 1900 cal B.P. ?): Following progradation
of the shoreline, the coastal terrace in the vicinity of
the excavation units was stabilized, with the surface
covered by vegetation. This is indicated by the
stable paleosol surface represented by Layer HA in
Units 20123.
Stage 5 (< 1900 cal B.P.): At this time colluvial
clays and larger volcanic clasfics were deposited
onto the coastal tenrace. The presence of charcoal in
these clays suggests forest cleane and buming on
the talus slope inland of the site. These human
activities thus initiated increased rates of erosion.
Stage 6 (< 1900 cal B.P.): The mixed colluvialcalcareous soil of the coastal terrace was used for
subsistence gardening and for dispersed habitation
during Stage 6.
In sum, the depositional sequence of Transect 9
closely replicates that of the 1987 excavation
locality, and that of Transect 5, as described
above.
The elevational data from Transect 9 also fully
support the morphodynamic model outlined in
chapter 4, and argued above for Transect 5. As can
be seen in the correlation diagram (fig. 5.26), the
Stage 2 occupation is no more than I m above the
modem high water mark, and some 3.5 m below the
modem beach ridge. Given a + 1-2 m higher sea
level stand at 2-3 kyr B.P., the Stage 2 habitations
would have been awash if they were not at a higher
elevation than at present. Retrodicting a subsidence
rate of about 1 m/kyr puts these habitation surfaces
at 1-2 m above the then sea level, which also fits the
lithological-sedimentological data (see Kirch,
Manning, and Tyler, chapter 7). Thus, the incorporation into our morphodynamic model of a tectonically induced subsidence rate for Ofu Island of ca. I
m/kyr is supported by the Tansect 9 stratigraphic
data.
Transect 11
Transect 1 1 lies 700 m east of the baseline
origin-point (the 1987 excavation locality). Here the
coastal tenrace is quite broad and flat, neatly 100 m
wide from te present shoreline to the base of the
talus slope. The elevation profile is depicted in
figure 5.27. Only a single excavation, Unit 26, was
located here, some 60 m inland from the baseline.
The soil along Transect I I is more sandy than at
Transects 9 or 5, lacking the colluvial clay component except very close to the talus slope. The
vegetation consists of tree crops (coconut and
Excavation
79
w
-a
3
'C
-
)
cc
co
0
m
m
O-1-
0
4
I
0
-2
0
I
CO -3
U.
w
-4
0
10
20m
TRANSECT 11
Figure 5.27
Elevation profile along Transect 11, showing the location of Unit 26 in relation to geomorphic
features.
breadfruit) with bananas, but aroids are absent,
reflecting tfie sandy, edaphic conditions.
The stratigraphy of the west face of Unit 26,
shown in figure 5.28, follows:
Layer IA: The dark grayish-brown (10 YR 3/2)
sandy loam in this layer is culturally sterile. A thick
root mass was present in the upper 5 cm. This zone
contains abundant shells of several terestrial
molluscs, including Subulina, Lamellaxis, and
Pleuropoma. The contact with Layer IB is gradational over 3-4 cm.
Layer IB: This very dark gray (10 YR 3/1), nonconcentrated midden deposit is lithologically
identical to Layer IA. There is some charcoal
flecking, and the texture is slightly 'greasier' than
Layer IA due to carbon and organic material. Some
fist-sized, volcanic fire-altered stones were noted.
The contact with Layer IHA is iregular but distinct
Layer HA: This stratum contains a white (10 YR 8/
2), massive, stnuctureless deposit of cultually sterile
calcareous, medium-grained sand. There is an
irregular, diffuse contact with Layer IIB.
Layer JIB: The pale brown (10 YR 6/3), faintly
darker calcareous sand in this zone is apparently an
old, stabilized A horizon paleosol within the beach
ridge depositional sequence represented by Layer II.
Layer UC: The white (10 YR 8/2), coarse-grained,
culturally sterile sand ofthis layer has a 'salt-andpepper' lithology (mixture of calcareous and volcanic grains).
Little cultural material was recovered durng the
excavation of Unit 26. Layer lB, the major source,
yielded slightly more than 2 kg of shell midden. No
pottery was present in any of the layers. One broken
Turbo shell fishhook was recovered at the Layer IB/
IIA contact. No radiocarbon dates were obtained
from this unit. The Layer IIB deposit may possibly
correlate with the phase of progradation and subsequent coastal terrace stability at ca. 1900 cal B.P.
identified fiom other trnsects. This could not be
verfied in the absence of radiometric dates, however. If a ceramic period occupation were ever
present in the vicinity of Transect 1 1, it would have
had to have been inland of Unit 26, and therefore
now deeply buried beneath the steep talus rockfall.
Transect 17
This was the most easterly of our 1989 excavation transects, situated 1300 m east of the baseline
origin-point. Transect 17 mns just slightly west of
the low gap over the island's central ridge at
Fa'ala'aga. At this easterly end of the To'aga coastal
terrace, the reef flat is quite narrow; thus the coastline is exposed to greater wave energy, particularly
during storms and cyclones. The beach fronting
Transect 17 consists of coarse sand and gravel, rather
tha
the finer-grined sands typical of locations
farther to the west Also,.the modem beach sediments at Fa'ala'aga contain a high percentage of
80
.Xe: .:~ (9.:
The To'aga Site
Excavation Unit 24
UNIT 26: W FACE
......1 . .1.. . 1 1*
~
*
~
~
.
.
.
.
0
. ...
_
..110.-.
Figure 5.28
Straigraphic section of the west fact
of Unit 26. Black squares indicate
sediment sampling locations.
volcanic clastics, derived from the exposed volcanic
headland at the eastem tip of Ofu (a control sample
of this modem beach sediment is described in
chapter 7). These particular geomorphological
features are likewise reflected in te litology of te
two units, 24 and 25, excavated along Trnsect 17.
The elevation profile along the trnset is shown in
figure 5.29, indicating the positions of the two
excavated pits just seaward of the talus slope.
Unit 24 was situated 1 14 m inland of the
baseline (along the coastal road) and about 10 m
from te base of tX talus slope. The talus here
consists entirely of large subangular boulders,
without a colluvial clay component. The ground
surface at Unit 24 was sandy loam with a cover of
coconut and breadfruit. The stratigraphy of the north
face of Unit 24, shown in figure 5.30, follows:
Layer IA: This stratum contains dark brown (10
YR 3/3) sandy loam (A honzon) developed under
Hibiscus tiliaceus, banana, breadfiuit, and other
vegetation. The parent material consists wholly of
coarse- to medium-grained calcareous sand. The
contact with Layer IB is irregular and slightly
diffuse.
Layer IB: Light yellowish-brown (10 YR 6/4),
coarse- to medium-grained, calcareous sand with a
very minor volcanic component makes up this
stratum. Also, there is slight humic staining. The
contact with Layer IC is gradational.
Layer IC: This white (10 YR 812). very coarse- to
coare-grained sand contains lenses of waterwom
coral, volcanic gravel, and waterwom branch coral
fingers. The sand is dominantly calcareous but with
a strong secondary mode of volcanic lithic grains.
TMe contact with Layer ID is fairly sharp and
regular.
Layer ID: Coral, gravel, and rubble with waterwom
branch coral fragments, in a matrix of coarse-grained
sand make up this stratum which appears to represent a single, high-energy depositional event. The
contact with Layer IE is sharp and distinct.
Layer IE: The coarse- to medium-grained 'salt-andpepper' sand of this layer has scattered waterwom
coral and volcanic gravel- to cobble-sized clastics.
Layer IF: This stratum contains volcanic, lithic,
coarse-grained sand with a slightly dominant
calcareous mode. The deposit incorporates large
cobbles of waterwom basalt and cobble-sized coral
head shingles. It represents a very high-energy
beach, with constant input of volcanic source
materials to the sediment budget.
The depositional sequence in Unit 24 reveals a
gradual transition from a very high-energy, exposed
beach directly at the base of the talus slope, then to a
lower-energy beach resulting from progradation,
and finally to the formation of a stable coastal tenrace
Excavatio,
81
W-0
W
0
0
0-
cor
:30
6
U-
e 4
I-
6.2
-2
-Jw
U-.
W
a:
o -4
20
0
40m
Elevation profile along Transect 17, showing the locations of Units 24 and 25 in relation to geomorphic features.
Figure 5.29
UNIT 24: W FACE
with soil development under vegetation. Three
ceranic body sherds (1 thickware, 2 thinware) were
recovered from Layer LB of Unit 24. One of these is
particularly unusual in that it is thick, coametempered pottery yet with a very clear red-slipped
exterior surface. While thickware is obviously
contemporaneous with early thinware assemblages,
ilufing those witi red slipping, this is the only
thickware sherd from the To'aga site with a red slip.
0.0
*IA
Root Mold
LB
0.5
'.
TRANSECT 17
0
O
2
IC
o
a
Excavation Unit 25
4._
9
~E
Unit 25 was positioned 10 m seaward of Unit
24. The stratigraphy of the north face, shown in
figure 5.31, is:
Layer IA: The very dark grayish-brown (10 YR 3/
2) sandy loam of this stratum has some subangular,
fist-sized, volcanic stone inclusions (not fire altered).
The contact with Layer LB is gradational over 2-4
cm.
Figure 5.30
St
secto
of
dte west fae of
Unit 24. Black squares indicate sediment
sampling
"V" i
lbcatio. "4C" indicae coal;
vocanic.
Layer IB: The black (10 YR 2/1) sandy loam of this
layer is a very 'greasy' textued cultural deposit of
calcareous sand with a heavy carbon and organic
materal component Shell midden and fire-altered
volcanic stones are present. The contact with Layer
IHA is fairly disdnct and slightly inregular.
Layer IC: The daik grayish-brown (10 YR 4/2) to
very dark grayish-brown (10 YR 312) sandy loam of
this deposit appears to be identical to Layer IA. It
probably represents a fairly long period of coastal
82
The To'aga Site
SUMMARY AND CONCLUSIONS
UNIT 25: N FACE
0.0
IA
4*
TB
:o.:
1.5
Base:of Excavation at2.1 ma.
U 25Graveli
Bl;/ squares
*
.
of Coral
and Branch Coral
O
*
o °
/
-
°
*
%'0''
,
~~~~1.5
***Base of Excavation at 2. 10 m
Figure 5.31
Sltlga_
secd of th ot fac
of Unit 25. Black squaresidct
sediment sampling loations.
terrace stability under vegetation, with gradual
accumulation. The contact with Layer IIA is distinct
and slightily irregular.
Layer HlA: This layer contains pale brown (10 YR
6/3), coarse-graied sand with a marked 'salt-andpepper' appearance due to a heavy mixture of
volcanic grains. In the northeast comer of the unit
there is a facies of sand matrix incorporafing waterrolled coral and volcanic gravelpebbles. The
contact with Layer IEB is sharp but slightly irregular.
Layer JIB: The white (10 YR 812), coarse-grained,
'salt-and-pepper' sand ofthis stratum has a lens of
water-rolled coral rubble and gravel (incorporating
branch coral fingers) at about 145 cm below surface.
The heavy carbon content ofLayer IB is suggestive of cookhouse activity. This interpretation is also
supported by the presence of fire-altenred volcanic
oven stones. No pottery was present in Unit 25.
A detailed discussion of the results of our duee
seasons of excavations at the To'aga site is best
defened to te concluding chapter of this volume
(see Kirch and Hunt, chapter 15). Here we confine
ourselves to a brief summary ofseveral key points:
1. Areal extent qfthe To'aga site. One of our
pnmary objectives during the 1989 excavation
season was to determine the full horizontal extent of
the subsurface cultural deposits at To'aga. Through
the use of systematic tnsect sampling, we have
been able to ascertain that buiied cultural deposits
dating as early as ca. 3000 cal B.P. extend contnuously from Transect 3 to Transect 9, a distance of
some 700 m. Between Transects 9 and Ilthe
density of cultual materials declines significantly,
but some culturl materals were present in Unit 26
along Trauect 1 1. Transt 17 also yielded a very
low density of culual materials. We were restricted
by the local landowners from testing the 600-m gap
between Transects 11 and 17 because of the presence of the sacred-and rather feard-Tui Ofu
tomb and other surface cultural remains in this area.
Based on our transect results, we can now state ta
subsurface archaeological deposits are present
thrughout the entire To'aga coastal flat as far as
Fa'ala'aga, but that the main pottety-bearing, concentrated midden deposits are restricted to the area
between Transects 3 and 9. The bured deposits are
concentrated in a narrow band, extending no more
than about 30 m seaward from the present base of
the talus slope. Shovel tests in vanous localities in
the present beach ridge (on both sides of the road)
revealed no archaeological materials. We are also
certain that cultual deposits extend inland under the
steep talus for an unknown extent. The great
thickness of e talus and colluvium, however,
prevents testing of these deposits without the aid of
heavy machinery which was not available to us. It
would probably require fte removal of as much as
15 m of massive talus-olluvial overburden to
expose some of these buried occupation deposits.
2. Testing o.f the morphodynnic model. The
morphodynamic model of coastal terrace fonnation
outlined in chapter 4 has been substantively supported by our stratigraphic and radiometric data from
the various tansect excavations. In particular, we
have been able to confirm that Ofu Island has been
Excavations
undergoing subsidence at a rate of 1-2 m/kyr during
the late Holocene. Rapid progradation and fonnation of the coastal terrace occurred during the period
1900-1000 cal B.P., largely due to an increase in the
marine-biogenic sediment budget. Thus, occupation
deposits dating earlier than 2000 cal B.P. are confined to a very narrow, fonner bench or beach ridge
situated at the base of the talus slope.
3. The O.fu cultural sequence. The cultural
materials recovered from our varous transect
excavations also provide the basis for outlining a
cultural sequence for Ofu Island. This sequence
commences at about 3000 cal B.P. with colonization
of the island by makers of a ceramic complex
including both thin and thickware vessels. While no
classic dentate-stamped Lapita pottery has been
recovered at To'aga yet, the early ceramic assemblage from AS-13-1 fits well within the Lapitoid
ceramic series (see Kirch 1988). Other portable
artifacts, such as Type V basalt adzes and shell
annbands, are also similar to Early Eastem Lapita
assemblages from Tonga and Fiji.
Over the period from 3000-1900 cal B.P., the
To'aga site was continuously occupied, and the
ceramic assemblages reveal a gradual shift in
dominance from thinware to thickware. At around
1900 cal B.P., a rapid progradation of the shoreline
commenced, resulting in the construction of the
present coastal terrace. Also, the deposition of
terrigenous sediments onto the coastal terrace
increased, creating a prime zone for horticultural and
arboricultural activities. During the past two
millennia, the To'aga coastal terrace has been heavily
utilized for such subsistence pursuits, although
habitations dispersed over the terrace continued to be
occupied, evidenced by such features as the 'ili'ili
house mound tested by Unit 13. In chapter 15 we
discuss this sequence in greater detail, and compare
it with the sequence defined for Westem Samoa.
REFERENCES CITED
Buck, P. H. [Te Rangi Hiroa] 1930. Samoan Material Culture. Honolulu: Bemice P. Bishop
Museum Bulletin 75.
Clark, J. T. 1980. Historic preservation in American
Samoa: Program evaluation and archaeological
83
site inventory. Report prepared for the Amencan Samoa Government. Anthrpology
Deparment, Bernice P. Bishop Museum,
Honolulu.
Cox, P. A. 1980. Masi and tanu 'eli: Ancient
Polynesian tehnologies for the prservation
and concealment of food. Pacfic Tropical
Botanical Garden Bulletin 10:81-93.
Emory, K. P., and Y. H. Sinoto 1965. Preliminary
report on the archaeological investigations in
Polynesia. Report prepared for the National
Science Foundation. Anthropology Department, Bemice P. Bishop Museum, Honolulu.
Green, R. C., and J. M. Davidson, eds. 1969.
Archaeology in Western Samoa, Vol. 1. Bulletin
of the Auckdand Insfitute and Museum.
. 1974. Archaeology in Western Samoa, Vol. 11.
Bulletin of the Auckland Institute and Museum.
Hunt, T. L., and P. V. Kirch 1988. An archaeological
survey of the Manu'a Islands, American
Samoa. Journal of the Polynesian Society
97:153-83.
Kikuchi, W. K. 1963. Archaeological surface ruins
in American Samoa. Unpublished M.A. thesis,
University of Hawaii, Honolulu.
Kirch, P. V. 1984. The Evolution of the Polynesian
Chiefdoms. Cambridge: Cambridge University
Press.
1988. Niuatoputapu: The Prehistory of a
Polynesian Chiefdom. Thomas Burke Memoral Washington State Museum Monograph No.
5. Seattle.
Kirch, P. V., T. L. Hunt, and J. Tyler 1989. A
radiocarbon sequence from the To'aga site, Ofu
Island, American Samnoa. Radiocarbon 31:713.
Kirch, P. V., and D. E. Yen 1982. Tikopia: The
Prehistory and Ecology of a Polynesian
Outlier. Honolulu.: Bemice P. Bishop Museum Bulletin 238.
Munsell Soil Color Charts 1988. Munsell Color,
MacBeth Division of Kollmorgen Instruments
Corporation. Baltimore.
Redman, C. L. 1974. Archaeological Sampling
Strategies. Reading, Mass.: Addison-Wesley
Module in Antwpology, No. 55.
6
RADIOCARBON CHRONOLOGY
OF THE TO'AGA SITE
PATRICK V. KIRCH
T HE AGE OF ARCHAEOLOGIcAL deposits at To'aga
may be roughly estimated by typological
comparison of te excavated cermnics and basalt
adzes with dated assemblages of similar artifacts
from Westem Samoa (Green and Davidson 1974),
and from elsewhere in Westem Polynesia (e.g.,
Kirch 1988; Poulsen 1987). Such comparisons
suggest that prehistoric occupation at To'aga spanned
much of the first millennium B.C. and on into te
Christian era Typological cross-dating is always
suspect, however, as it depends upon the assumption
that stylistically similar artifacts frm different
islands were used contemporaneously, and thus
ignores the possibilities of cultral lag or of independent rates ofchange. For example, there is no
particular reason to assume, a prioin, that the change
from Early Eastem Lapita pottery (characterized by
dentate stamnped designs) to Late Eastem Lapita
should have occurred at the same time trughout
the Samoan archipelago. hIneed, the specific aspects
of ceramic change may themselves have varied
spatially over the Samoan Islands. Likewise, the
cessation of ceramic manufacture and use was
probably not a contemporaneous event trughout
Westem Polynesia, as indicated by relatively late
dates for pottery from Niuatoputapu (Kirch
1988:13942). In order to obtain a more precise
chronology for the To'aga site, and an independent
assessment ofthe age of the varous arfifact assemblages sampled from the To'aga deposits, fourteen
samples of charcoal and shell were radiocaibn
dated. This chapter psnts provenience and
sample details for each of ftese dates, and discusses
some of the broader implications of the To'aga site
chmnology for Samoan prehistory.
An initial assessment of the age of potterybearing deposits at To'aga was obained by dating a
sample of Turbo shell midden obtained from the
1986 reconnaissance test pit adjacent to the landfill
dump. This sample (Beta-19742) and another
pottery-associated sample from Ta'u Island (Beta1974 1) were first reported by Hunt and Kirch
(1987). The sanples were close in age, with a
weighted average of 2605 ± 56.6 B.P., caibraed to
360-50 cal B.C. aftr corrcion for marine rervoir
effects. Hunt and Kirch noted that these dates from
Manu'a were "in close contemporaneity with
corrected radiocarbn assays of wood charcoal
associated with stylistically and technologically
similar ceramic assemblages from tUpolu Island in
Westem Sanoa" (1987:417). At the conclusion of
the 1987 field season, an additional seven samples
(six shell, one charcoal) were submitted to Beta
Analytic for 14C assay. These dates were reported
by Kirch, Hunt, and Tyler (1989), and provided "a
statigphically consistent chrnologic sequence for
86
The To'aga Site
human occupation of Ofu Island, spaning the
period from ca. 3700-3300 B.P. up to the modem
era" (1989:10). At the end of the 1989 excavation
season, six more samples (two shell, four charcoal)
were submitted to Beta Analytic for 14C age assessment These 14C ages ar reported here fortfhfirst
time, along with their contexual details. The 1989
samples all proved to be irnally consistent, both
saigraphically and in comparison with the 1987
14C series.
In swn, a suite of fourteen radiocarbon dates are
now avalable for the To'aga site, ranging in age
from the second millanium B.C. up to fte end of fte
first millennium A.D. All dates are stratigraphically
consistent, and in aggregate provide a firm basis for
establishing the chrnlogy of human occupation
and of geomorphological change at To'aga. Because
this series spans viually the entire "ceramnic period"
of Samoan prehistory, it is also of considerable
significance for the cultural history of the archipelago as a whole.
RADIOCARBON DATING PROCEDURES:
CORRECTIONS AND CAUBRATIONS
hi selecting samples for radiocarbon dating,
emphasis was placed on material which was in firm
stratigraphic context (with no apparent distubance
or inversion), and in obvious association with
artifacts or featur. When possible, charcoal
samples were obtained from hearth or oven features,
although in four cases finely dispersed charcoal
flecks had to be dated in order to ascertain the age of
features of particular interest. The charcoal samples
consisted of wood charcoal (species not determined)
mixed in some cases with carbonized fragments of
coconut endocarp (Cocos nucifera). With the
exception of two shell sanples from the
stratigrahically oldest beach deposits, the shell
selected for 14C dating was culurlly-modified
midden material about which there was no uncertainty as to culturl associations. All dated shell is of
marine origin, consisting of shallow-water, reefdwelling gastrpods and bivalves. Whenever
possible, we seleced the species Turbo setosus for
dating, as a further control on consistency of results.
This reef gastpod is one of the dominant components of the shell midden assemblages from the
To'aga site (see Nagaoka, chapter 13).
All 14C measurements on samples from Toaga
were performed by Beta Analytic, Ic. Shell
samples were pretrated by etching away the outer
layers with dilute acid, to remove any adhering
contaminants. The smples were then atacked with
further acid to produce carbon dioxide, which was
used as Xt carbon source for dating. The charcoal
samples were first picked manually for rootlets, and
then given a series of acid, alkali, and acid ings
to remove carbonates and humic acids; they were
then rinsed repeatedly to neutality. One sample
(Beta-35600), consisting of very finely dispersed
organics (chamoal and ofter organic material), was
picked for rootlets and then dispersed in hot acid to
eliminate carbonates. All benzene syntheses and
countings proceeded nonnally (M. Tamers, pers.
comms., 21 July 1988; 6, 13 March 1990). Four
samples (three of charcoal and the organic sample)
had only small quantities of carbon remaining after
ptreatment, and thus were accorded extended
counting times (four times the normal duration) in
order to reduce statistical eror as much as practicable. The 13C/14C ratios were measured for all
samples to establish 13C adjusted, "conventional 14C
ages" following the wecommendations of Stuiver and
Polach (1977). The nonadjusted "'Libby ages,"
conventional 13C adjusted ages, and a13C values for
all samples are presented below.
Calibration of conventional 14C ages for secular
effect, and in the case of marine shell samples for
reservoir effect as well, were made following the
rece work of Stuiver, Pearson, and Braziunas
(1986) for shell samples, and of Stuiver and Becker
(1986) for fte tenestrial charcoal and organic
samples. All calibrations and probability estimates
were obtained with the use of the CALIB microcomputer program (Rev. 2.0) developed by Stuiver and
Reimer (1986).
As discussed by Kirch and Hunt (1988b:23), the
dating of marine shell samples from southwestem
Pacific archaeological sites has posed some problems of calibration due to the "reservoir effect" of
older carbon present in the wodrd's oceans. Archaeologists have recognized that marine shell samples
frequently give ages somewhat older than those of
charcoal from identical archaeological contexts, and
several attempts have been made to develop correction factors for Pacific island shell dates (Gillespie
and Polach 1979; Gillespie and Swadling 1979;
Jansen 1984; Atxhns 1986; Kirch 1989:139). These
Radiocarbon Chronology
87
attempts suggested that southwest Pacific samples of
shell from shallow reef envirornents ranged frm
about 400-550 years older than the actual age.
Recently, Stuiver, Pearson, and Braziunas (1986)
produced a first, approximate, global model of txse
reservoir effects for both an "'upper mixed layer" of
the ocean (relevant to the To'aga samples) and for
the deep ocean. Their model takes into account
time-dependent changes in 14C activity, and allows
for local geographic variations in reservoir effect
(e.g., te effect ofupwelling from the deep ocean
along continental shelves). The local geographic
variations are accounted for by a A-R correction
factor. Ideally, such a A-R factor should be independently determined for Samoa by 14C assay of
modem, pre-bomb marine shell. As yet, however,
this has not been possible. We therefore have used
as a A-R correction factor a weighted average of
empirically determined A-R values from several
mid-Pacific islands, specifically Eniwetok, Hawaii,
and Tahiti and Mo'orea in the Society group
(Stuiver, Pearson, and Braziunas 1986, table 1).
This pooled A-R value of 100 ± 24 represents a
reasonable working value for tropical, central Pacific
island samples of shallow-water marine organisms.
We have also used this A-R value of 100 ± 24 in
calibrating radiocarbon dates from a wide range of
Lapita sites (Kirch and Hunt 1988a, b). It is significant that following calibration of the marine shell
samples according to the above procedure, these are
entirely consistent with the calibrated charcoal dates.
This is shown for example by samples Beta-35603
and -35604, both from the same excavation level of
Unit 28 in Transect 9, which overlap at one standard
deviation. (hat the charcoal sample appears to be
slightly older may reflect the use of wood from a tree
that had been growing for a century or more prior to
cutting and buming.)
face type, along with the D13C value in paits per
thousand. Calibrations for each date are given at one
standard deviation; values within parenteses
indicate intercepts on the calibration curve.
THE RADIOCARBON CORPUS
FROM TO'AGA
Beta-25035. 3370 ± 70 B.P.; 3820 ± 70 B.P.; a13C =
+ 2.4 %/o.
The fourteen 14C age determinations from
To'aga are listed below, by laboratory number, with
all relevant information on provenience, associations, samnple details, and calibrations. The
uncorrected "Libby age" is given first, followed by
the a13C correCted "conventional 14C age" in bold-
Marine shell (2 specimens: Asaphis violascens
and Lunella cincerea, total weight 48 g) from 1987
main excavation trench, Unit 6, Layer V, 314 cm
below surface. Both specimens retained their
surface coloration and were not water-rolled, thus
indicating deposition in Layer V soon after death.
These shells consist of naturally deposited marine
Beta-19742. 1890 ± 50 B.P.; 2350 ± 50 B.P.; a3C=
+ 2.9 0/o0.
Marine shell (Turbo setosus) from the initial test
pit excavation adjacent to the Ofu landfill dump
(Layer D, Level 10). The Turbo shell consisted of
midden refuse in an organicaly-enriched midden
deposit, with calcareous sand matrix. The sample
was direcly associated with small quantities of
thick, coarse-tempered pottery. Cal 28 B.C. (A.D. 45)
A.D. 108 at 1 a; cal B.P. 1977 (1905) 1842 at 1 a.
Beta-25033. 2190 ± 80 B.P.; 2640 ± 80 B.P.; al3C =
+ 2.3 %/o.
Marine shell (Turbo setosus, 71 g) from 1987
main excavation trench, Unit 6, Layer IIA-1. The
sample consisted of cultrally deposited shell
midden in direct association with an earth oven
feature, and with small quantities of Polynesian
Plainware ceramics. Cal B.C. 362 (244) 145 at 1 a;
cal B.P. 2311 (2193) 2094 at 1 a.
Beta-25034. 2120 ± 80 B.P.; 2570 ± 80 B.P.; al3C =
+ 2.5 %oo.
Marine shell (Turbo setosus, 70 g) from 1987
main excavation trench, Unit 6, Layer IIB. The
sample consisted of culurally-deposited shell
midden from the pnincipal occupation deposit in
Layer II, in association with Polynesian Plainware
ceramics and with Turbo-shell fishhooks. Cal B.C.
295 (161) 58 at 1 a; cal B.P. 2244 (2110)2007 at 1
a.
88
The To'aga Site
cultural food refuse) in a calcareous
shell (rather
beach deposit also containing isolated tn-fine ware
ceramic sherds. Cal B.C. 1765 (1682) 1600 at 1 a;
cal B.P. 3714 (3631) 3549 at 1 a.
Beta-25673. 3170 ± 80 B.P.; 3620 80 B.P.; al3C =
0/oo.
+ 2.2
Marine shell (Phalium sp.,45 g) from 1987
main excavation trench, Unit 1, Layer V, 290 cm
below surface. The dated specimen was a single
whole gastpod, not waterwom and retaining
original surface coloration, thus indicating rapid
deposition in the Layer V beach deposit soon after
death. Layer V contained isolated thin, fine-tempered ceramic sherds. Cal B.C. 1526 (1441) 1377 at
1 ai; cal B.P. 3475 (3390) 3326 at 1 a.
Beta-26463. 1460 ± 50 B.P.; 1910
+ 2.5
50
B.P.; al3C =
%1oo.
Marine shell (Turbo setosus, 72 g) from Unit 3
of the 1987 excavation tranet; in Layer 11,40-70
cm below surface. The dated specimen was a single
large Turbo shell, with the apenural margin displaying chipping due to cultual removal of the operculum (presumably to extract the edible soft pails).
The specimen was stradgraphically situated at the
basal contact of the cultural, aceramic midden
deposit and the underlying sterile calcareous beach
sand. The specimen's context thus indicates a time
period after the commencement of rapid
progradation of the To'aga coastal terrace. Cal A.D.
561 (620) 663 at 1 a; cal B.P. 1389 (1330) 1287 at 1
Beta-26464. 2660 ± 140 B.P.; 2620 140 B.P.; a13C
--27.8 %1oo.
Charcoal flecks from 1987 excavation Unit 10,
Layer HB, at 70-80 cm below surface. The charcoal
was in association with Polynesian Plainware in the
Layer IIB occupation deposit. The field sample
weighed ca. 1 g, yielding 0.2 g of carbon after
pretreatment in the laboratory; the sample was
accorded extended counting time. Cal B.C. 967
(801) 454 at 1 a; cal B.P. 2916 (2750) 2403 at 1
a.
Beta-2646S. 1160 ± 70 B.P.; 1600 70 B.P.; a13C =
+ 2.0 %1oo.
Marine shell (Turbo setosus, 66.4 g) from 1987
excavation Unit 13, Layer III, 35-45 cm below
surface. The dated sample consisted of 1 nearly
complete shell and 2 small fragmentary shells, all
displaying cultually-induced fractures and chipping
(possibly from ardfact manufactre). The Turbo
shells were in direct association with an aceramic
cultural midden near the base of a pebble-paved
house platfoim, in which Unit 13 was excavated.
Cal A.D. 828 (914) 1000 at 1 a; cal B.P. 1122 (1036)
950 at 1 a.
Beta-35600. 1200 ± 70 B.P.; 1190 ± 70 B.P.; a13C =
- 26.1 %/o.
Finely dispersed charcoal and ash ("organics')
from Trnsect 5, Unit 17,53 cm below surface in the
southeast comer of the unit. The field sample
weighed 510 g. The charcoal and ash were dispersed in the interstices of a gravel ('ili'iIi) pavement
layer (see chapter 5), presumably a dwelling house
floor. Cal A.D. 694 (781, 789, 805, 821, 829, 839,
862) 943 at 1 a; cal B.P. 1256 (1169, 1161,1145,
1129, 1121, 1111, 1088) 1007 at 1 a;.
Beta-35601. 2950 ± 110 B.P.; 2900 ± 110 B.P.; al3C
= - 27.8 %1o.
Charcoal flecks from Transect 5, Unit 28, base
of Layer 11,290-300 cm below surface. The field
sample weighed 12 g, and produced 0.6 g of carbon
after laboratory pretreatment; it was accorded
extended counting time. The charcoal was from a
non-concentrated midden deposit containing several
sherds of a red-slipped, thin, fine-tempered ceramic
ware. Cal B.C. 1308 (1188, 1184, 1127, 1126, 1107,
1105, 1083, 1059, 1054) 930 at 1 a; cal B.P. 3257
(3137, 3133,3076, 3075, 3056,3054, 3032, 3008,
3003) 2879 at 1 a.
Beta-35602. 2660 ±100 B.P.; 2630 ± 100 B.P.; a13C
= - 26.9 0/oo.
Charcoal in an ash and sediment matrix (470 g)
fm Transect 9, Unit 23. The charcoal was contained within an earth oven feaure cut from the
upper pan of Layer HIA into Layer IlIB, which
contained predominantly thickware pottery, with a
small quantity of thinware. Cal B.C. 896 (803) 663
at 1 a; cal B.P. 2845 (2752) 2612 at 1 a.
Radiocarbon Chronology
89
Beta-35603. 2660 ± 170 B.P.; 2600 ± 170 B.P.; a13C
stratigraphically consistent series of dates s
= - 28.4 0/oo.
the period from ca. 3700 to 1000 cal B.P. In figure
6.1 these dates are plotted as calibraed age ranges at
one standard deviation, in chronological order. The
cultual associations of the samples are indicated on
the right-hand side of the diagram. The eau¶iest two
dates (Beta-25035 and -25673) are frm the Layer V
beach deposit of te 1987 main trench excavation
and are associated with thin, fine-tempered ceramics.
Sanple Beta-35601, also associated with this early
type of pottery, is from an in situ cultrl deposit. A
suite of four dates (Beta-35602, -26464, -35603, and
-35604) is associated with assemblages having
significant frequencies of thin, fine-tempered
pottery, fishhooks, and other artifacts as well as
coarse-tempered thickware. Another suite of four
dates (Beta-25033, -25034, -19742, and -35924) is in
association with predominantly thick, coarsetempered pottery, of the sort known in Westem
Samoa as "Polynesian Plainware" (Green 1974).
Some thinware, however, continues to be represented in these later deposits.
These ages indicate a time span of ca. 24001500 cal B.P. for the phase of dominant use of thick
pottery. The final three dates (Beta-26463, -35600,
and -26465) are from aceramic cultural contexts, and
all are younger than 1500 cal B.P. Although we have
no 14C dates from To'aga younger than about 1000
Cal B.P., it is lkely that the site continued to be
occupied at least until European contact, based on
historic artifacts and varous surface archaeological
indications (house mounds, grindstones, masi pits,
etc.).
DISCUSSION
Charcoal (70g) from Tnsect9, Unit 23, from
the base of Layer IUB, 190-206 cm below surface.
The field sample of 70 g yielded 0.15 g of carbon
after labortory patment, and was accorded
extended counting time. The charcoal flecks were
collected from a culturl occupation deposit containing thin, fine-tempered pottery, including some rnms
with notched decortion. This sample was from the
same depositional context as shell sample Beta35604. Cal B.C. 968 (797) 433 at 1 a; cal B.P. 2917
(2746) 2382 at 1 a.
Beta-35604. 2330 ± 80 B.P.; 2770 ± 80 B.P.; D13C =
+ 1.7 0/0o.
Marine shell (Tridacna maxima, 455 g) from
Transect 9, Unit 23, Layer IIB, 198 cm below
surface. The dated specimen consisted of a single
valve of T. maxina, showing no signs of water
abrasion or rounding; the shell was presumably
collected live for food and deposited directly in the
cultual matrix. It was in direct association with
ceramics and other portable artifacts, in the sane
stratigraphic position as charcoal sample Beta35603. Cal B.C. 495 (385) 340 at 1 a;; cal B.P. 2444
(2334) 2289 at 1 a.
Beta-35924. 1640 ± 70 B.P.; 2100 ± 70 B.P.; a13C =
+ 2.7 /co.
Marine shell (Turbo setosus, 400 g) from
Transect 5, Unit 15, Layer II, 60-70 cm below
surface. The dated sample consisted of numerous
cultually modified (chipped and broken) pieces of
T. setosus, representing food refuse and probably
also detritus from the manufacture of Turbo shell
fishhooks. The shell was dispersed in a dark grey,
'greasy' (carbon-nich) deposit with much fire-cracked
y a cookhouse culural deposit. The
rock,
deposit also contained thick, coarse-tempered
ceranics. This is the youngest 14C age from To'aga
in direct association with pottery. Cal A.D. 319 (410)
473 at 1 a; cal B.P. 1631 (1540) 1477 at 1 a.
The To'aga Sequence
The fourteen samples described above provide a
The radiocarbon sequence from the To'aga site
is the largest suite of dates from any single ardhaeological site in the Samoan archipelago, and provides
an imporant baseline not only for establishing the
cultural chronology of the Manu'a Islands, but for
comparisons with sequences from Tutuila and
Westem Samoa. An initial radiocarbon and stratigraphic sequence for the Samoan archipelago was
established by Green and Davidson (1974) on the
basis of forty-five dates from a number of sites on
'Upolu and from two sites on Savai'i. The University
of Utah Samnoan Archaeological Program (Jennings
and Holmer 1980) obtained another thirty-four 14C
dates from dteir excavations of sites on 'Upolu and
90
The To'aga Site
B-26465
-4
*
B-35600
4
w
o-. B-26483
-4 B-35924
__)
'-4 B-19742
i
4 B-25034
i- 4 B-25033
i
;---
i B-35603
i
I B-26464
i-
I--4
'-
B-35604
w
B-35602
B-35*01
B-25673
'4|B-25035
I
I
2
3
4
I
1
KYR CAL B.P.
Figwre 6.1
Plot of radiocarbon age determinations from the To'aga site (cal B.P. at 1 sigma).
Manono Islet. These seventy-nine dates from
Westem Samoa provide a firm basis for an 'absolute' chronology.
In American Samoa, the situation is less well
developed. Frost (1978) obtained five radiocarbon
dates from tee sites (AS-21-1, AS-21-2, AS-25-1)
excavated by her. One sample, from Tulauta Village
(AS-21-1), yielded a very early date of 810 ± 140
B.C. along with a significandy later date of A.D. 1320
± 70. The other sites all dated to the last thousand
years. Best et al. (1989) reported seven radiocarbon
dates from the Tataga-matau adz quarry on Tutuila,
indicating a sequence of use spaning the past 1000
years. Clark (in press) reviews an additional fifteen
radiocarbon dates, largely from Aoa and Leone
valleys on Tutuila Island. These include a date of
3389-2749 cal B.P. for initial occupation of the Aoa
Valley, as well as very late dates for the use of
ceramnics. The late dates associated with ceramics
have caused Clark (in press) to argue that "pottery
was not abandoned uniformly and wholesale
trughout the archipelago."
Dating Initial Hwnan Colonization
On archaeological (ceramic) critera, t oldest
known site in te Samoan archipelago is the submerged Early Eastem Lapita occupation at
Mulifanu'a, Upolu (Green and Davidson 1974;
Leach and Green 1989). A sample of marne shell
Radiocabon Chronology
(NZ-1958) from this site was originally reportd as
having an age of 2980 ± 80 B.P. (Green and Richards
1975:317). More recently, Leach and Green have
indicated that the correct convendonal 14C age for
this sample should be 3251 ± 155 B.P. (1989:319).
They suggest that the "most probable age" for this
sample is 3116 cal B.P. (1 166 cal B.C.), after taking
into accou te marine resevoir effect (Using the
100 ± 24 A-R value applied to the To'aga samples,
the Mulifanua date would be calibrated to 1280-800
B.C.) This is consistent with other dates on Early
Eastem Lapita assemblages from Fiji, Tongatapu,
and Niuatoputapu (Kirch 1988, table 48), suggesting
initial Lapita colonization of the Fiji-Westem
Polynesian region by about cal 1200 B.C. (see also
Kirch and Hunt 1988a, b).
The earliest two 14C detemninations from the
To'aga site, from the Layer V beach deposit in the
1987 main trench excavation, range between about
3700-3300 B.P. Although small numbers of thin,
fine-ware sherds are present in this beach deposit, we
must be cautious in interpreting these dates, because
the deposidonal context is not specifically cultural.
As argued in chapter 5, it is likely that the sherds
derive from an in situ occupation situated on a beach
ridge somewhat inland of the Layer V beach, and
now buried under several meters of talus and
colluvium. That the shell samples used for dating
showed no signs of water-rolling or prolonged
weathering is noteworthy, indicating that they were
deposited in Layer V soon after death. It is conceivable, however, that the sherds became incorporated
into the beach deposit somewhat later in time.
Although these Layer V dates are not inconsistent
with the earliest radiocarbon ages for Lapita colonization in the region, they do fall at the early end.
Sample Beta-35601, from a deep test unit
directly against the talus slope, is unquestionably in
primary, cultural context and is associated with thin,
fine-tempered ware. It dates to 3257-2870 cal B.P.
(1308-930 cal B.C.). Thus, we can be certain that
Ofu Island, and the To'aga site, were settled by the
end of the second millennium B.C. as part of te
proess of discovery and colonization of the FijiWestem Polynesian region by Lapita populations
(Kirch and Green 1987).
Calibrating the Morphodynamic Model
A second objective of some importance in our
91
work has been to calibrate th sequence of shoreline
change and progradation esponsible for the creation
of the To'aga coastal ten-ace. Based on the model
developed in chapter 4 (Kirch), it was suggested that
the coastal tenace would not have begun to prograde
significantly until sometime after about 2000 cal
B.P., when te Holocene sea-level maximum began
to drop to its modem level. This hypothesis is
substantially confinned by the series of 14C dates
from To'aga which reveals at the coastal terrace
was very narrow and confined to a zone adjacent to
the high cliffs until sometime early in the second
millennium A.D. Sample Beta-26463, at 1389-1287
cal B.P., comes from a straigraphic context ta postdates fte onset of significant prograation of fte
coastal terrace.
The Samoan Ceramic Sequence
Fmally, the To'aga radiocarbon suite provides a
finn chronology for the changes in the ceramic
sequence revealed by our excavations and analyses
(see chapter 5). Although classic dentate-stamped
Early Eastem Lapita pottery was not recovered from
To'aga, the red- and orange-slipped thin, finetempered ware recovered from the deepest units
would appear to be contemporary with the
Mulifanu'a Lapita assemblage from rUpolu. The
manufacture and use of thin, fine-tempered ware,
some decorated with rim notching, spwanned the
period from about 2800-2400 cal B.P. This was
followed by a phase in which the manufacture and
use of thicker, coarse-tenpered pottery became
dominant between about 2400-1500 cal B.P. The use
of pottery appears to have been discontinued entirely
by about 1500 cal B.P., thus matching closely the
chronology for 'Upolu derived by Green and
Davidson (1974). Furtr details of this ceramic
chronology are presented in chapter 9 by Hunt and
Erkelens.
REFERENCES CITED
Athens, S. 1986. Archaeological investigations at the
Tarague Beach site, Guam. Report prepared for
the Department of the Air Force. San Francisco, Califomia.
Best, S., H. Leach, and D. Witter 1989. Report on
the second phase of fieldwork at the Tatagamatau site, American Samoa, July-August
92
The To'aga Site
1988. Department of Anthropology, University
of Otago, New Zealand.
Clarlc, J. T. 1989. The Eastem Tutila archaeological
project: 1988 final report Report prepared for
the Govenmnent of American Samoa.
in press. Radiocarbon dates from American
Samoa. Radiocarbon.
Frst, J. 1978. Archaeological investigations on
Tutuila Island, American Samoa. Unpublished
Ph.D. dissetation, University of Oregon,
Eugene.
Gillespie, R., and H. A. Polach 1979. The suitability
of marine shells for radiocarbon dating of
Australian prehistory. IN R. Berger and H.
Suess, eds., Proceedings of the Ninth International Coniference on Radiocarbon Dating, pp.
404-421. Berkeley: University of Califomia
Press.
Gillespie, R., and P. Swadling 1979. Marine shells
give reliable ages for middens. Search 10:9293.
Green, R. C., and J. M. Davidson 1974. A radiocarbon and stratigraphic sequence for Samoa. IN
R. C. Green and J. M. Davidson, eds., Archaeology in Western Samoa, Vol. 11, pp. 212-24.
Auckland stitute and Museum Bulletin 7.
Hunt, T. L., and P. V. Kirch 1987. Radiocarbon
dates from two coastal sites in the Manu'a
Group, American Samoa. Radiocarbon 29:41719.
Jansen, H. S. 1984. Inshtie of Nuclear Sciences
INS-R-328: Radiocarbon Datngfor Contributors. Lower Hut, New Zealand: New Zealand
Institute of Nuclear Sciences.
Jennings, J. D., and R. N. Holmer 1980. Archaeological Excavations in Western Samoa. Pacific
Antrpological Records 32. Honolulu:
Bemice P. Bishop Museum.
Kirch, P. V. 1988. Niuatoputapu: The Prehistory of
a Polynesian Chiefdom. Thomas Burke
Memorial Washington State Museum Monograph No. 5. Seattle.
Kirch, P. V., and R. C. Green 1987. History, phylogeny, and evolution in Polynesia. Current
Anthropology 28:431-56.
Kirch, P. V., and T. L. Hunt 1988a. Radiocarbon
dates from the Mussau Islands and the Lapita
colonization of the southwestem Pacific.
Radiocarbon 30:161-69.
1988b. The spatial and temporal boundaries of
Lapita. IN P. V. Kirch and T. L. Hun, eds.,
Archaeology of the Lapita Cultural Complex: A
Critical Review, pp. 9-32. Burke Memorial
Museum Research Report No. 5. Seattle.
Kirch, P. V., T. L. Hunt, and J. Tyler 1989. A
radiocarbon sequence from the To'aga site, Ofu
Island, American Samoa. Radiocarbon 31:713.
Leach, H., and R. C. Green 1989. New information
for the Ferry Berth Site, Mulifanua, Westem
Samoa. Journal of the Polynesian Society
98:319-29.
Poulsen, J. 1987. Early Tongan Prehistory: The
Lapita Period on Tongatapu and its Relonships. 2 vols. Terra Australis 12. Deparunent of
Prehistory, Australian National University.
Canberra.
Stuiver, M., and B. Becker 1986. High-precision
decadal calibraion of the radiocarbon time
scale, A.D. 1950-2500 B.C., Radiocarbon
28(2B):863-910.
Stuiver, M., G. W. Pearson, and T. Braziunas 1986.
Radiocarbon age calibration of marine samples
back to 9000 cal B.P. Radiocarbon 28(2B):9801021.
Stuiver, M., and H. A. Polach 1977. Discussion:
Reporting of 14C data. Radiocarbon 19:355-33.
Stuiver, M., and P. J. Reimer 1986. A computer
program for radiocarbon age calibration.
Radiocarbon 28(2B): 1022-1030.
7
A GEOARCHAEOLOGICAL ANALYSIS OF
SEDIMENT SAMPLES FROM
THE TO'AGA SITE EXCAVATIONS
PATRIcK V. KIRCH, ELIZABETH MANNING, AND JASON TYLER
INTRODUCTION
A
MAJOR THUSr OF OUR research at the To'aga
site has been the investigation of the geomorphological and sedimentary context of the site's
extensive buried archaeological deposits. Such an
approach is essential both for the underading of
the "formation processes" by which the site was
created (Schiffer 1987), and for the broader issues of
landscape change in relation to human ecology on
the Manu'a Islands. An understanding of the site's
geomorphology also relates directly to cultural
resource management concems for, as we shall
argue, there are reasons to believe that te To'aga
coastal terrace is now in a stage of active coastal
erosion. While the deeply bunred archaeological
deposits at To'aga are not immediately theatened by
shoreline regression, they could become endangered
should this process intensify, particularly through the
effects of global warming and sea-level rise (Geophysics Study Committee 1990).
In chapter 4, a model of the morphodynamics of
the To'aga coastal terrace was deduced from the
general pattem of Holocene sea-level regimes in the
southwestem Pacific, and from certain features of
Samoan geological history, especially evidence of
tectonic subsidence. The stratigraphic results of our
transect excavations, presented in detail in chapter 5,
substantially confinned this model, while the
radiometric chronology presented in chapter 6
provided a temporal calibration of the stages of
coastal terrace formation. In this chapter, we focus
on another line of evidence for testing the
morphodynamic model: that of geoarchaeological
analysis of sediment samples obtained from the
stratigraphic profiles of the varous transect excavation units.
Our approach here follows long-established
principles of sedimentology (Blatt et al. 1972; Folk
1974; Krumbein and Sloss 1963; Twenhofel 1950,
Reineck and Singh 1980) in the context of a
geoarchaeological perspective (Stein and Fan-and
1985; Shackley 1975). In particular, our aim is to
augment our field stratigraphic descriptions and
interpretations through laboratory analysis of
sediment samples, using a "life history" approach to
sediment interpretation. Any sediment, for example
a sand layer in one of our transect units, can be
understood in tenns of four "life history" stages.
Each sediment has: (1) a source; (2) a tansport
history; (3) an environment of deposition; and, (4)
post-depositional alterations (Stein 1985:6-7). These
stages may be infenred through the application of
various analytical techniques including particle-size
analysis; grain lithology and mineralogy; pointcounting of grain composition; chemical and
physical detennination of carbonates, organic matter,
and pH; colorimetric analysis; and oher techniques.
94
The To'aga Site
Our geoarchaeological investigation of sediment
samples from To'aga was conducted in two phases.
In 1987, a series of archaeological and control
samples was obtained to check aspects of our
statigraphic interpretation. ibese samples were
analyzed by Tyler at the University of Washington
geoarchaeology laboratory, following procedures
outlined by Stein (1985). In 1989, a much more
extensive set of samples (both archaeological and
contrls) was obtained from our trLansect excavations. These 1989 samples were analyzed at the
University of Califonia, Berkeley, Archaeological
Research Facility by Mmaning, under the direction of
Kirch. The analytical methods applied to the 1989
samples differed slightly from those used in 1987,
largely because in the latter case we were able to
define more precisely the specific research questions
to be addiessed by laboratory analysis. For examnple,
in 1989 we recognized that time-onsuming pipette
determination of the fine-fcion (< 4 phi) was
unnecessary to determine the environment of
deposition for these predominately coarse-grained,
calcareous sediments. lherefore, we concentrated
our efforts on mechanical sieving and textual
analysis of the pebble- to fine-sand-sized components, with additional determination of grain lithology (basalt versus carbonates) in the -1 to 1 phi (¢)
size ranges. In other aspects, however, such as pH,
organic matter, and carbonate determination, our
1987 and 1989 laboratory tecniques were identical.
As indicated above, Tyler and Manming were
responsible for the laboratory work upon which this
chapter is based, while Kirch authored the text and is
responsible for the final interpretations and data
presentation.
directly into heavy plastic bags and sealed for
shipment back to the laboratory. As discussed by
Stein (1985:7-9, fig. 1), sampling strategies can vary
depending upon the questions to be asked in laboratory analysis. Since we were interested pimarily in
depositional events, our samples were taken from
within individual strata, avoiding boundares or
contacts between layers. Where large clasdcs were
present (e.g., > -2 0) which could not be adequately
sampled, these were noted in stratigraphic descriptions. Subsampling in the laboratory, for various
analytic tests, was caried out with the use of a Jones
sample splitter.
A total of 32 sediment samples (including 3
controls) was obtained and analyzed in the 1987
season. Another 70 samples (including 5 controls)
were taken and analyzed from the 1989 excavations.
Of the 1989 samples, those from Units 15, 19, and
23 were accorded a full analysis, while samples from
other units were analyzed only for color, pH, parficle
size, and lithologic composition.
Controls
Several control samples were obtained in the
To'aga area from modem depositional envimrnments,
in order to assist in the interpretation of the archaeological sediments. These included samples of active
colluvial material from the slope inland of the 1987
excavations, sand from contemporary beach ridges,
sand from te beach in front of Transect 7 (fig. 7.1),
and a high-energy, sand-and-gravel beach at
Fa'ala'aga. Details of tese control samrles are
provided in tables 7.1 and 7.2.
Particle-Size Analysis
METHODS
A primary aim of laboratory analysis was to
Field Sampling
interpret the environment of deposition of sediments.
Sediment samples were systematically taken
from all units following th completion of excavation, in conjunction with the drawing and description
of stratigraphic profiles. In both 1987 and 1989, the
exact positions of sampling blocks were recorded on
the satigraphic profiles, and these have been
indicated on the various section drawings reproduced in chapter 5. Sample blocks, which usually
measured 10 x 15 x 10 cm, were cut into the cleaned
section walls with a trwel. Samples were placed
This requires a knowledge of the particle-size
distribution, since the size ranges and degree of
sorting of a sediment reflect the energy levels in that
envirnment. For example, low-energy beaches will
be characterized by well-sorted (stmngly unimodal)
sands in the medium- to very fine-grained size range.
Higher energy beaches, on the other hand, display
less sorting and a higher frequency of coarse-grained
to granule-pebble-sized clastics. Similatly, colluvial
sediments deposited by mass wasting can be expected to be very poorly sorted, exhibiting a ful size
Geoarchaeological Analysis of Sediment Samples
Figure 7.1
Photomicrograph (lOX) of the 1 phi size fraction of a modem control sample of calcareous beach sand
from Transect 7.
grains together (Kunze 1965). Organic materials
were removed by pretreatment with hydrogen
peroxide, H202 (Jackson 1969). A 1:1 solution of
sediment and distilled water was prepared, to which
H202 was gradually added until frothing and
foaming subsided. The beaker was then placed in a
water bath at 80 OC, with additional H202 added
range from silts and clays up through very large
clastics (cobbles to boulders).
Unconsolidated sands generally required no
pretreatment prior to parficle size analysis. The
colluvial samples, however, required pretreatment
for the removal of organics, and in some cases, for
iron oxides which would otherwise bind individual
Table 7.1
Analytic Data for To'aga Control Sediment Samples (1989)
Sample
No.
89-131
89-132
89-133
89-134
89-135
95
%
Gravel
Location
Fa'ala'aga Beach
Fa'ala'aga Beach
Fa'ala'aga Beach
T-7, A horizon,
beach ridge
T-7, beach sand
%
Sand
% Textural
Class
Silt
pH
%
Calc
%
Basalt
2.3
43.0
97.4
1.7
97.7
56.9
2.5
97.0
0
<0.1
<0.1
1.3
S
sG
G
S
9.01
8.68
8.95
8.28
97
90
89
100
3
10
0.2
99.8
0
S
8.64
100
0
11
0
96
The To'aga Site
N _-
0
SW
a
en Q 'N
C!
N., Q en
1-- 1-..
m m C4 WI) t- W) en
C4 w 04 04 04 N 04
>, 0>,
- . >4
tn 4n
00
t t
V-4 7-4
"a
U
>4
lw i
:t t:
en
C14
'r_oo
4)0
en CtO
>4
t t_
W
04 -O: 04
-o
r4 No t-o
lq
vx
IS
CN 00
O
g
t', .
v
:sco
E-4
t
(N
en
0
E W)
C4
-4
er
*4 >.0
I. 1:
s'
cI(NI r) r)
-
0> 004
>
r-4 -
_
>0
"t >rX O;
oo
1.
<
0 0
0x ~c o o\
O ~o t_
0%O-t
W)
C1
~o C14"I "-4 - N-4
c t 00 t
"t 'tt enON
t-
00
x- t- qt cr
0
as
Cu
:t -.
O WO
44
) O>4 W w W
tn ---
odo40' e
0
-
1-
W:
Rr)tfL0 C4
^k
;soo-seno
'-qo*o4
0
_;~W C sfu00x N 00
s*;o-en*r in
r*6
^ t
_-
\0
>
()
_
4
6
t
-
- 0-0
In a a
1014
>,
Q
00
!qQ
en
V-
m
4
N \0
o
T--4
m
r- ON t ONi ON
m r xur ON N
t)
~o
CS _o-x ON oN
W) xh on xn
C C1
C -4
47
00 n C1 ot W) en r
vx
cr
00
r-
u N N
a
m
00 (Vs e
1
-
~o e
W-
~u
:3f
00
i
06~ 60 06 6t:0
_
tt oo
(NI
C,'
Cu
V-
o
c;O
en
V
W
o,- Q
xy4
6~ 6O0
60
e
tn en
_ot
e
>o _"
o
Q w00
V-4
*O
9
0 0 0
t
0*-->_ N-oa_t
R
W
o
m o 8
o
_
_
W~~~~~~~~1 0 o&
0_
00N 0oen
~o
Wo ON
_
r:6
0 0
0
t~- 0 0r 00 00 00
ro_
Os00_V- om
tn
C)
-
~o_
O _
N
0
QN
eon
r
ro t
_ tn
CZ ON
0_t
ON
r-4
o0o0
t
\O00'iO~~~~~~~~l
00e
O\O\NI 0--
00
.o
N
8
vo
Wo e %v
_oo
-
o7Nr
e
t\eOo%
O
N
N
Co
O t O
oO
t oO N
-_
_
X
O
O
ON
0
cr o
e
xO\O x0t cr
\ cO Ft
ON
_ ox
x o'r
r0
f-400en
oo
ON
t- t
r- O 00 (In en t N
06
li t--:
r- W) 00
t-
O
oO
O-
0
st0
0
c.)
g0
.o
U
C.)
u
co
D
g
ci
c)
i
i
M
U
0-4 011,
.
.o oN
4-0
m u --
c > ._
CS
'-4 -4-
-
C
4..#
in
.:) ::)
4-0
.9.4
.W
=
.9-= , .4
_
_=
_
V-4 -
m
--
-
>
-
-,
.1.4
.""
t fi
4
0.
_
S:S:4
.4
=
o
o o_
3>3V4 &
-
(A
1-
t
v4
*s:.
_
*.= 4"4
*-
*
_
-P
.*
*-
.4
z
§4
.
rAZI
0
Geoarchaelogical Analysis of Sedime SWples
Pretreated samples were mechanically shaken
through a nested set of geological reens ofmesh
sies -2 to 4 # (plus pan) for fifteen minues.
Weights of each subsample were deteimined with a
high-precision digital electic balnce. The 1987
samples were futher analyzed by the pipette metod
to determine the particle size distribution of the fine
frtion; this was not done for the 1989 samples.
Following mechanical sieving and weighing, the
percentages and cumulative percentages of all*
classes were calculated, as was the sand:mud atio
(mud is defined as all material > 4 *, i.e., silt plus
clay). Tese stastics permit the textual classification ofthe sediments according to the system outlined by Folk (1974:27-30, table 1). This system
utlizes a triangular diagram with apices of gravel,
sand, and mud, reproduced here as figure 72. The
dominant mode is indicated by a capital letter,
supplemented by lower case letters for secondary
modes. For exanple, a sandy gravel is designated
sG, while a slightly gravelly muddy sand would be
untl all reaction had ceased. The sanple was then
cooled, rinsed witfi distilled water, and centnfuged at
1700 rpm for fifteen minutes. Superatant liquid
was thn decaned, and the ctifuging procedure
repeated until the soludon was dlear.
The metd for removal of inn oxides was as
follows: sediment which had lready been pretrat for removal of organics was placed in a 250
ml centrifuge tube to which about 200 ml of citrte
buffer soludon (sodium citrate dihydrate; sodium
bicarbonate NaHCO3; sodium chloride NaCI) was
added. Samples were then warmed to 80 OC in a
water bath within the fume hood. Four grams of
sodium dithionite (Na2S204) were slowly added
while stiring coantdy for one minute. Following
a fifteen-minute digestion period, 10 ml of sated
sodium chlonde (NaCI) was added to flocculate te
sample, which was then centrifuged at 1700 rpm for
fifteen minutes; the supematant liquid was decanted
and discarded. The procedure was repeated until
samples appeared light gey or white in color, indicating the mnoval of mostor all offe free imn oxides.
Figure 7.2
(g)mS.
Gravel
(>2n11im)
Terminology adopted for the
To'aga site sediments, in terms
of the relative contribution of
gravel, sand, and mud size fractions, expresd asa trary plot
8(1
(Souce: G-dinerandDackombe
1983:108).
Z\
MUDl)Y
a
a
tGRAVEL | a
(mG)
a<
0
4.
...i
A
GRAVELLY \7
MUDDY \
GRAVELLY MUD
sIGhTLY
-
v
SuIIY
G;RAVELLY
SANDY
-
tf
(ginS)
Mtl)
((Wg)M)h Hf
--Iv
Irl-v
xx
SAND
(gM)
GRAVLLY
MUD
-
SLIGIITL Y
GRAVELLY
MUI)DY SANI)
l(igmS)
IC
I
SLIGHTLY
GRAVELLY
SAND ((g) S)
A
IL
/ UMUI(M)/ SANDY MUD (sM) MUDDY SAND (mS)\ SAND (S\
0
Mlud
(<0.06 mm)
I :9
97
a
I
1: 1
9:1
Sand: mud ratio
Sand
(0.06-2mm)
98
The To'aga Site
pH
Point-Counting (Lithology and Micro-artifacts)
The pH (acidity, neutrality, alkalinity) of
samples was detennined witfi a Mettler automatic
pH meter. A 20 g sample was p din a 1:1
solution with distfilled water, and standards set to pH
7 and 10. Tlree trials were made for each sample,
with the repoited value being the average of these
readings
The pH values from the To'aga sediments are of
intres prmarily for the assessmentof preservadon
of organic culural materials, such as bone and shell
faunal remains. pH values for the upper colluvial
deposits, or of contemporary A soil horizons with
humic materials, tended to range between 6.4-8.
Deeper sediments, generally calcareous, had more
alkalin pH values ranging frxm 8 to 9.5. This
alkalinity strngly favors the preservation of bone
and shell, reflected in the generally excellent condition of tX faunal assemblage. In contast, the
relative neutrality of the shallower deposits probably
reflects th presence of humic acids. This is evident
in te "chalky" nate of shell midden in these upper
During the field description of sedimentary
units, we observed that the more deeply buiied and
older staa tended to have sands with a mixed
calcareous-volcanic lithology, designated 'salt-andpepper' sands (fig. 7.3), whereas higher and more
strata.
Organic Matter and Carbonates
Ihe p=sence of organic matter and carbonates
in sanples was determined by the 'loss-on-ignition"
method (Dean 1974; Stein 1984). This is based on
the principle that organic materials will begin to
ignite at 200 OC and will bum completely when the
temperue raches 550 OC. Calcium carbonate
(CaCO3) evolves to carbon dioxide gas when heated
to 800 OC and is eliminated at 850 OC. Thus te
loss-on-ignition procedure involves the controlled
burning of samples with precision weighing before
ignitions after 500 °C, and after 1000 OC bums. All
samples were processed in a Thermolyne muffle
fumace.
Color
Sediment colors were recorded in situ during the
description of profiles, using the Munsell Soil Color
charts, and have been reported in chapter 5. In
addition, the colors of laboratory sediment samples
wecorded both dry and moist, also using the
Munsell system.
were
rcent sands were
largely
or
wholly calcareous.
These differences were believed to reflect the
process ofcoastal progradation and termce formation, which gradually removed sources of volcanic
sediment from the To'aga coastal sediment budget.
These sources originally would have been volcanic
headlands and talus boulden which when exposed to
high-energy wave action would generate volcaniclithic sand grains. In order to quantify more precisely these field observations, we undertook
controlled point-counting of samnples in the laboratory, following procedures first outined by
Galehouse (1971). The -1, 0, and 1 + size classes
were selected for point-counting following mechanical sieving. Subsamples of these + classes were
mounted on petrographic glass slides, which could
then be examined under a Nikon stereozoom
microscope. An average of 100 grains was counted
for each sample, with grains being classified as
calcareous (consisting of coral sand, foraminifera,
marine shell fragments, etc.) or volcanic lithic (basalt
or similar materials). These counts were then
calculated as percentages.
For Units 15, 19, and 23, we also used the pointcounting method to determine the frequencies of the
following classes of micro-artifactual or faunal
materials: charcoal, marine shell, bone, sea urchin,
and terrestial gastrpods. (In order for marine shell
to be counted as a micro-artifact constituent, it had to
have sharp or fractured edges, as opposed to waterworn or rolled edges, the latter indicating a natural
constituent of the sand matrix.) The terrestrial gastropods are of particular interest as they consist primarily of synanthrpic species; these are the subject of
a separate and detailed analysis in chapter 8.
RESULTS
The 1987 Excavations
Thirty-two sediment samples were obtained
from the 1987 excavations (including fhree control
Geoarchaeological Analysis of Sediment Samples
Figure 7.3
99
Photomicrograph of a typical "salt-and-pepper" sand with mixed calcareous and volcanic-lithic
lithology; from Layer IF of Unit 24 (1 phi size fraction; lOX).
samples). Samples were taken from Units 2 and 3
along the initial transect, from Units 5, 6, and 9 in
the main trench, and from Unit 10. Analytical data
from all of these samples are reported in table 7.2.
The sedimentary sequence in the main excavation trench is summanzed by samples 13-23 listed in
table 7.2. The colluvial units (Layers I and III) are
similar in their higher frequencies of silt and clay,
contrasting markedly with Layers II and IV which
are dominated by sand-sized particles. Similar
contrasts are expressed in the percent of calcium
carbonates in these strata.
The 1989 Transects
Transect S
Units 15, 16, and 17 were selected for sediment
analysis along Transect 5, with a complete set of
analyses for Unit 15, and particle-size and
lithological analyses only for Units 16 and 17.
Complete analyfical data for these units are presented in tables 7.3 to 7.5.
The sedimentary sequence in Unit 15 is graphically depicted in figure 7.4, and a more detailed
graphic presentation of the grain-size distribution
results is provided in figure 7.5. In figure 7.4 several
overall trends are apparent. The lower sediments
(Layers Ill to IV) are dominated by sands, with
significant gravel components confined to the upper
deposits (Layers I and II). This change reflects the
shift from a littoral depositional environment to a
mass wasting (colluvial) depositional environment.
The pH shifts from neutral to somewhat alkaline
between Layers IA and IIIA and remains alkaline to
the base of the section. Organic matter is greatest in
Layers IA and II. The shift from littoral to ternigenous mass wasting depositional environments is
also clearly indicated in the percent of CaCO3 and in
the lithological composition of basalt and calcareous
grains determined by point-counting of the 1 0 size
100
The To'aga Site
00 NO00
0
S
0
Q zt C4 >- >4 >-
_- t._. _.
"-4
RQI
tv
0'r)0
Qo
OO0
sri
o.
-
00 N- 00
0; 00I
NN0C
I'
qA
to
z
>0
e
00
cir
ad
r- (1 -
e
0
w0 od)00
04 rC
0
A
~ -C.I Q
"4H
2
-1 00
-H
e "a
o 0oo
-
0 R o 0t r
w ON C C O C bON
--4-b
i
cl
m
Cu
2
0
Y r-'%o
ON * Nt t
O-_N -4
CR
-O R
ON ON 0000 ON
t
u
V"
014
0'
IGIN
00 00
O CIA ON
C
u
*CI0
4-
co
E0tc;
Cu
*s:
Cn Cu
c,
w) -
N
w) r- c7
o
00000000
XN
X0
i.
Iw
;a
-4
Cu
e4
V-
tn
V--
0
ONO
0*
PC
U
qtC
)
,t"TC C) r-
NO 00
ON 4 N
\0 £ o CN GII r-G
ooCK
a en
Ce
0
NN
0
VN
cI
O
"-4
d
~0.t-ONt--00
(V
Wm cf \0
0 r-c- <t),
ONt00
_. C4)
OO
c\-C\ C\b
'w
Ev
c o
m0--c _e 0<-4
_>
-.
CX
w
riA
N-00OGN
C
m
- m
0000 00
&..
-
-
--
-
0.4
00
ON0
0
,it
)RI~R
ON
00 0000
000000
00 0000
0
_-4
00
00
00
C0oo
Geoarchaeological Analysis of Sdet Se
101
s
Table 7.5
Analytic Data for Unit 17 Sediment Samples
Munsefl Color
1+
Sample
9
No.
Layer Gravel
89-66
89-67
89-68
89-69
89-70
89-71
89-72
89-73
89-74
89-75
I
II
IIIA
IIIB
IIIC
IVA
IVB
V
VI
VII
GRAIN SIZE
49.1
80.7
83.5
83.4
63.0
1.9
1.4
2.3
11.5
0.4
%
Sand
% Textural
Silt Class
47.8
17.4
14.6
15.1
35.4
95.9
94.9
97.5
88.4
99.5
%
Basalt
Dry
7.84
8.42
8.51
8.47
8.46
8.48
8.57
8.54
8.90
9.50
99
100
100
99
100
99
76
100
100
93
1
0
0
1
0
1
24
0
0
7
7.5YR 3/2
7.5YR 5/2
10YR 5/3
7.5YR 5/2
10YR 5/3
7.5YR 5/8
7.5YR 6/2
10YR 8/3
7.5YR 8/4
7.5YR 8/2
ORGANIC MATTER
pH
8
sG
G
G
G
sG
S
S
S
gS
S
3.1
1.8
1.9
1.5
1.6
2.1
3.6
0.2
0.04
0.05
pH
%
Calc
9
o
i
5
10%
CARBONATES
0
I
I
Moist
5YR 2.5/2
7.5YR 5/2
7.5YR 5/2
7.5YR 4/2
10YR 5/3
7.5YR 5/6
7.5YR 5/2
10YR 8/3
7.5YR 8/4
7.5YR 8/2
LITHOLOGY (1+)
100%
_-
MA-1
TRANSECT 5, UNIT 15
Figure 7.4
Summary diag of grain size, pH, organic matte, carbonates, apd lithology for sdiment
samples from Unit 15, Transect 5.
The To'aga Site
102
2
IA
30
50
50
0
0
-~~~~~MIII
0
80
70
70
0~~~~~m
50
~~~mc
L
X
t
50
0
-2
-1
0
I
2
GRAN SIZE
Figure 7.5
3
44 'P'
t
Hstogram plots of grain size distribu
dons for sedinent samples from Unit
15, Transect 5.
class. PhotomicrograPhs of the 0 + size frcions for
sediment samples from Unit 15 are siown in figure
7.6; tese illusae the vadations in lithology
between the various cultl and natrl deposits.
The detailed grain-size distnbutions shown in
figue 7.5 likewise document changes in the environment of deposition. The basal sand, Layer IV,
consists of rather poorly sorted sands (dominant
mode = coarse sand) with a small granule-sized
component, suggestive of a relatively high-enrgy
beach environment In Layen HID to lIIA-1 there is
a steady shift to well sorted sands dominated by
particles in the medium sand range (2 0) with
granule-sized particles generally lacking. This shift
is presumably correlated with the progradation of the
active beach seaward of the Unit 15 locus. The
Layer II midden deposit displays a very poorly
sorted sediment, reflecting the incorporation of
cultual materials (including oven stones and other
large clastics) into a medium-sized sand matrix.
Layers lB and IA are also poorly sorted, with
dominant modes in he pebble-to-cobble size ranges
(-2 to -I O). These upper sediments are typical of
young colluvium denived from mass wasting of the
cliffs and volcanic slopes immediately inland of te
site.
In sum, the Unit 15 sedimentary sequence
reveals the following geomorphological evolution.
The intmdal enviromnent of deposidon (at ca. 3,000
Cal B.P., prior to human colonizadon) was a highenergy beach, indicating an active shoreline very
close to Unit 15. The source of sediment was
primarily the calcareous reef flat, with te minor
addition of volcanic lithic grains derived from
exposed headlands and/or talus rockfall exposed to
wave actdon. During the period of human occupation represented by Layers HID to IIIA-l (ca. 30001600 cal B.P.), the Unit 15 locality shifted to a beach
ridge depositional enviromnent, with trasport of
calcareous sediment primarily by aeolian processes
(e.g., saltation of grains inland). Following the
Layer H midden occupation on top of a vegetated
beach ridge, the Unit 15 locality began to be covered
by terrigenous, poorly sorted, volcanic sediments
deriving from the steep talus and colluvial slopes
inland. The encraclhnent of xse volcanic sediments probably reflects increasing slope instability
due to forest clearance and agricultual modifica-
tions.
Geoarchaeological Analysis of Sediment Samples
Figure 7.6
103
Photomicrographs (lOX) of the 0 phi size fractions of sediment samples trom Unit 15, Transect 5: A,
Layer IB; B, Layer II; C, Layer IIIA; D, Layer IIA-1; E, Layer IIIB; F, Layer IIIB; G, Layer IIID; and
H, Layer IV. (e.s. = echinoid spine; Lb. = fishbone; s.g. = synanthropic gastropod).
104
The To'aga Site
Transect 7
Sediment analyses were perfonned on samples
from Units 18 and 19 along Transect 7. Tables 7.6
and 7.7 summanze the detailed analytical results.
A graphical summaiy of the Unit 19 sedimentary sequence is shown in figure 7.7, and the detailed
grain-size distnbutions are depicted in figure 7.8.
The overall sequence minors ta described above
for Unit 15, with such features as the shift from
neutral to alkaline pH, the rapid decease in organic
matter, and the relative contributions of basalt and
Transect 9
Units 21, 22, and 23 were selected for sediment
analysis along Transect 9. Tables 7.8 to 7.10 present
Table 7.6
Analytic Data for Unit 18 Sediment Samples
1+
Sample
%
No.
Layer Gravel
89-86
89-87
89-88
43.8
27.3
1.1
IA
IB
II
GRAIN SIZE
0
50
IB
50.1
64.9
95.6
7
9
7.8
3.3
msG
gmS
(m)S
7.42
8.46
8.63
95
99
Munseil Color
%
Basalt
Dry
63
5
1
5YR 3/3
IOYR 3/4
IOYR 5/4
37
CARBONATES
0
50
I
100%
No Data
44 0
IA
40
Sand
Moist
5YR 3/3
10YR 3/4
5YR 4/6
LITHOLOGY (1+)
o
so
40
I
4p
TRANSECT 7, UNIT 19
Figure 7.7
100%
a
Calcareous
IVB
VI
6.1
pH
%
Calc
I
m
V
% Textural
Silt Class
ORGANIC MATTER
pH
100%
IATG
%
Sand
i1~
carbonate sediment grains. Layers VI to II are all
reladvely well-sorted sands dominated by a mediumsand (2 0) mode, with the exception of Layer V, in
which some larger (-1 and -2 0) clastics represent the
incorporation of cultural materials into the sand
matrix. The Layer IA colluvium is typically poorly
sorted.
Summary diagram of grain size, pH, organic matter, carbonates, and lithology for sediment samples
from Unit 19, Transect 7.
4
Geoarchaeological Analysis of Sediment Swpks
00 t00 00
lx IY od o
t.o
0
Qr: 00
oo
00tt00
-- >00I 0
40
1- V-
" - -4
"au
Qrtn[o
a
u
x00 000
i
(
eir-:
2m
gO
', 00
a
Ck (2 -N
6 tr; t
105
A
a
) 00
Ia
-4 -4 --_
000
r- Os
Q4> I Q
en
d 4-
0
2
N
It C4
I'- - t- -
-
CU
ce
~u
cr4
0
-
C
I
0
0
r-00 N% o t)
m 00 0000
0X
0% 0000
ICu
cr,
Eu
r-:00
vo m m tl
00
00
1
00
00
'a
00
00
co "a
CO) C
"A
CT C
F*
to trxocraxo
aN
*a:
*so
co
%,wI
c'
c
6
enCq4-4
-
-
"
0|
>e
as
ON as
q
|t
t
00
en in
c
r-4 -4
_s 0oe
1-m
-m
-t
N
P0%
0
-
--~
0
13
m
S
.
> > >
&o
04
_0 04
a
0 6-
S<
>
S1
WI
ON o _- N en
m o
x ON ON ON ON O ON
oEoao b oo o o
Oa
o00 0
00
00 00
o-~ VI
The To'aga Site
106
the detailed analytical results.
A graphic swumary of the Unit 23 deposidonal
sequence is shown in figure 7.9, and the detailed
grain-size distributions are illuste in figure 7.1 1.
otomicrographs of the 00 size fraction for
sediment samples from Unit 23 are shown in figure
7.10. As seen in figure 7.9, the overall pattem of
E2
.2
U .~
30
so
0
60
I
70
IVA
0
60
l
.
0-
0
~~~~~IVB
soV
-2
'
-1
1
' 2 ' 3 ' 4 ' PN'
GRAIN SIZE
Figure 7.8
(#)
Hisgram plots of grain size distributions for sediment samples from Unit
19, Transect 7.
change is much like that described above for
Transects 5 and 7. Of particular note is the prsence
of basalt grains in the basal sands (Layers IV and
IIIC), reflecting the 'salt-and-pepper' lithology
observed in the field. The detailed grn-sze
histograms (figure 7.1 1) indicate that Layer IV is a
coarse-to-medium-grained sand with a 'tail' extending into the larger clastic size ranges, suggestive of a
higher-energy beach depositional environment In
Layers MIB and IIA, the larger dastics are cultual
materials (such as oven stones, marine shells, and
large sea urhin spines) which were added to a
coarse-to-medium-grained sand matrix. Layer HA,
intepreted in the field as a paleosol horizon, is less
well sorted than the underlying sands and begins to
incorporate fine-grained terigenous sediments. The
Layers IC to IA colluvial deposits are all very poody
sorted, with ineasing quantities of terrgenous
(basalt) materals.
Figue 7.12 presents a graphical summary of
grain-size distribution and lithology for Units 21, 22,
and 23. These results clearly reflect the temporal
shift from (1) high-energy beach depositional
environments in the basal levels of Units 23 and 21
to (2) beach ridge depositional environments with
human occupations, to (3) the encroachment of
terigenous colluvial deposits in the vicinity of Unit
23.
A micro-artifact constituent analysis of sediment
samples from Unit 23 is presented in table 7.1 1, with
frequencies for five categores of culural material
indicated by grain size (-1 to 10 size classes).
Charcoal was represented only in the upper colluvial
deposit Layer IB, presumably reflecting humaninduced buming up-slope, associated with agricultral activities. Other cultural materials, mainly sea
urchin and bone, were confined to Layers IIA to
IUB.
Transect 17
Sediment analyses were performed for samples
from Units 24 and 25 along Transect 17. Analytical
results are summarzed in tables 7.12 and 7.13.
In figure 7.13, the results of grain-size distribution and lithology are presented for both Units 24
and 25. A shift from high to lower energy depositional environments is indicated by the grain-size
distribution, while a decreasing contribution of
volcanic sand grains is reflected in the 10 lithology.
Geoarchaeological Analysis of Sediment Sampks
GRAIN SIZE
pH
ORGANIC MATTER
5
10
CARBONATES
107
LITHOLOGY (10)
15
TRANSECT 9, UNIT 23
Figure 7.9
diapm of grain size, pH, organic matter, carbonates, and lithology for sediment sanples
Unit 23, Transect 9.
Summary
from
Table 7.9
Analytic Data for Unit 22 Sediment Samples
Sample
No.
Layer
%
Gravel
%
Sand
89-111
89-112
89-113
I
II
III
15.5
31.5
0.1
79.8
67.0
99.6
This indices the effects of coastal progradation,
and the gradual eliminadon of volcanic boulders and
headlands as a source of sand sediment.
detailed grain-size distribution for Unit 24
clearly reflects the changing depositional environments at Transect 17 (figure 7.14). Layers IF
hough IC are indicative of very high-energy
beaches with substantial components in the granulepebble-cobble size modes (O to -2 O). Layer ID, in
The
% Textural
Silt Class
4.7
1.5
0.2
gS
sG
S
pH
8.36
8.64
8.84
1¢
%
%
Basalt
Calc
100
100
100
0
0
0
particular, probably represents a major stonn event,
possibly associated with a tropical cyclone. Layer
IB reflects a rapid shift to a lower energy beach,
dominated by coarse-to-medium-grained sands,
while the upper Layer IA reflects the development of
a stable beach rdge depositional environment
There was no significant encroachment of colluvium
over Transect 17.
It is clear that for most of its late Holocene
The To'aga Site
108
Iid
0
0
*9
r5o
94
.-
a
I
a
-
I
a
I
1
I
I
C'4
I
I
I
I
I
I
a
I
I
I
I
I
I
I
I
0
I
I
I
U)
t_ -
> t .-
Cfi
I
ln
tn W
t
a
I
Ia
I
a
-
I
WI
WI WI
hia ;,
I
~oc
*
He |
*.
a
I
a
0
a
I
sn"
0
m
0%sr4n
te
'8It
0
C14 %o
0^
0o00
"4
C J
%O0
2
;~aco
la
q
i
N~
E- 00 00 00
V-
riC5,,
6
"-4
| cl
s.
I
a
.
"4^
e4
-IL -
N
I
I
I
I
I
I
6
"4
*a
as
c;
U)
*sA
m
en
00
00 00
00
0
a)
*=
3U
0
I
I
I
I
a
I
I
I
I
I
I
I
a
I
I
I
I
I
I
I
a (NI
I
I
I
I
a
I
I
I
I
I
I
en
I
I
I
.
I'
I
000000Oa
V) o)n _4V
4C
o
I
0+
*;.
Q
0a"
x x o
*0>N
-D""0
o
I
8
i
-ItV
-
*Eoo
'I
6< ma4
DI
as
;5 f<<
c;
a- 4
0-
rt0m
o
cl
4
14 6
4fiw
v-
%t
V-4
en
6< m C; 6 m <
C
t
V-
%4
oa
"04
>,
8
0-~-~
e
t-
m u >
-N
0x
00
00
lid
m u
"-4
aao
000000000
00
00
8
x0 oo
an
"-4
0
"-4
00
0
o4
00
0
0
x4
"-4
00
a
00
*
Geoarchaeological Analysis of Sediment Samples
109
I
I.
I
I.
I
Figure 7.10
Photomicrographs (lOX) of the 0 phi size fractions of sediment samples from Unit 23, Transect 9: A,
Layer IB; B, Layer IC; C, Layer IIA; D, Layer IIB; E, Layer IIIA; F, Layer IIIB; G, Layer IIIC; and
H, Layer IV.
The To'aga Site
110
I"
It
0
Li
Ig
&W
0
0
-
_B
--
~u
0
I-
2
cn
0
v
Q
en
CZ
r-0000 00
C4
0
9114
>~~~~~~~~~~~~~~~~~ o~~ ~~~~~~~~~~~~~~~~~=
'-4
V--4 -4 -4
*N
t
-
Cu
en
Q '.4
1-m W)
Cd C'i '0e
c
o 0
en o _t
>-, Cd
0
W) >4
t- W)
00
it
en en
*
,
00
I-
I----
C4r:400
od
Q C) b-C> so
C1
U)
v
;._
4)
112
~aN
- -
.Ude
u
exi
r
CF oo CD ON Ch
0000000en00000
Ono
\0 0-00
C00
C
N IT m £ r- 0
0x 0x 0x 0o 0o
od
eu
CX
V1)
Li N)
*s:
00
Ul
(jA
t-0 *&
C
00 00
i2
N
co
r
cnCA
1-4
co
Cu
s N x00N N
0
0CC
b0t
co
*a
'0N
U)
Q
lqt'-(14 IOCN00
QON \
-C1
co
PC
00 00
rj,
C1400 \OC1 000C
00
o00 CI40WIr)
bi : ^
0
'U
'Ut
4)
NW
04)
Nx
'r) \or'
eNcNw
0000 00
00 CO
o
.< m U
O.4-40-
;~ a e
-4e
N V--N
c
oo~
00 00 00
-
X0 00 00 X0 00
C0
00
'_~
40
40
30
_
0
~L
Geoarchaeological Analysis of Sedcm Sanples
I
history, Transect 17 has been an area of high-energy
beach deposition, which conelates with the lack of
significant cal deposits. The stable coastal
ten-ace probably did not extend into this Fa'alapaga
area unil the last 1-1.5 kyr B.P.
I~~I
SUMMARY AND CONCLUSIONS
IA
I0
I
o-
0
40
0
IC
HIA
.
lA
m
_r_
50
0
50
50
o~~~~
rm
0
-2 '-I 'PAOIs
2
~~~IV
3'
GRAIN SIZE (4)
Figure 7.11
Hisg plots of grain size distributions for sediment samples from Unit
23, Transect 19.
111
Th results of detailed sdiment analyses frm
the To'aga excavations provide a set ofindependen
data with which to t th morphodynamic model of
coasal change developed in chapter 4. These rsts
ar highly consistent between tasects and confirm
that dtere has been a significant shift in source,
mode, and environment of deposition at To'aga
which conmlates with shoreine prg ion and
formation of the coasal terrace. Pidor to, and at the
time of initial human colonization of Ofu, the
shoreline along the southem pail of the island was
much closer to the volcanic cliffs, in the vicinity of
the present surficial contact between the talus slope
and coastal terrace. These basal deposits are consistently reflective of high-eergy beacrs, and the
admixture of basalt with calcareous grains indicates
ta volcanic headlands along with the coral reef
provided sources of sediment A shift to low-energy
beach ridge depositional environments by 1900 cal
B.P. reflects coastal progradation, as predicted by our
morphodynamic model. Subsequently, inrased
deposition of young, pootly sorted, temgenous
colluvial sedimets onto the calcareous coastal
terrace suggests inreased human disuibance
(burning and agricultural activity) on the interor
volcanic slopes after about 1900 cal B.P. Further
implications of ftese results are explored in grater
detail in chapter 15.
112
The To'aga Site
UNIT 21
UNIT 23
100
UNIT 22
100
0
LA
IA
IB
IB
IC
HA
I1A
fIB
LB
RC
I
MLI
mA
GRAIN SIZE (%)
mc
wV
INLAND
46
0I
IA
IA
IB
IB
IC
IIA
IIA
fB
fiB
'IC
50
100
SEAWARD
50
0
100
I
Calcareous
II
Calcareous
m
"Salt and Pepper" Sands
MfA
MB
mc
IV
Figwe 7.12
"Salt and Pepper" Sands
GRAIN LITHOLOGY
I
(%)
Summary diagrams of grain size and lithology for Units 21, 22, and 23, Transect 9.
GeoarchaeologicalAnalysis of Sediment Swmp1es
UNIT 25
UNIT 24
GRAIN SIZE
GAIN SIZE
LllTHOLOGY (1)
0
50
100%
0
IA
IA
IB
IB
IC
IC
ID
HIA
is
LIf)
EF
IIB
INLAND
Figure 7.13
-'
*
50
LITHOLOGY (14)
100% 0
50
100%
SEAWARD
Summary diagrams of grain size and lithology for Units 24 and 25, Transect 17.
113
The To'aga Site
114
Q '5|
t~~
E3 A
siH
Jill~'
I
40
r
~~
O~
70 -
40
IC
so
40
50
L:
IE
0
so
phological Field Manual. London: Allen and
Unwin.
Dean, W. E., Jr. 1974. Detennination of carbonate
and organic matter in calcareous sedimentary
rocks by loss-on-ignition: Comparison with
other methods. Journal ofSedimentary Petrology 44:242-48.
Folk, R. L. 1974. Petrology ofSedimentary Rocks.
Austin: Hemnphll Publishing Co.
Galehouse, J. S. 1971. Point counting. IN R. Carver,
ed., Procedures in Sedimentary Petrology, pp.
385-407. New York: Wiley and Sons.
Geophysics Study Committee 1990. Sea Level
Change. Studies in Geophysics. Washington
D.C.: National Academy Press.
Jadcson, M. L. 1969. Soil Chemical Analysis: An
Advanced Course. 2nd ed., Department of Soil
Science, University of Wisconsin, Madison.
Krumbein, W. C., and L. L. Sloss 1963. Stratgraphy
and Sedirnentation. San Frandsco: Freeman.
Kunze, G. W. 1965. Pretreatment for mineralogical
analysis. IN C. A. Black, et al., eds., Methods of
SoilAnalysis, pp. 569-77. Madison: American
Society of Agronomists.
Reineck, H. E., and I. B. Singh 1980. Depositional
Sedimentary Environments. New Yodk:
Springer-Verlag.
IF
0
-2 -1-I
F
0
1
GRAIN SIZE
Figure 7.14
3
2
4N
(#)
Hisgraam plots of grain size distributions for sediment samples from Unit
24, Transect 17.
REFERENCES CITED
Blatt, H., G. Middleton, and R Murray 1972. Origin
of Sedimentary Rocks. New Jersey: Prentice
Hall.
Dackombe, R V., and V. Garliner,1983. Geomor-
Schiffer, M. B. 1987. Formation Processes of the
Archaeological Record. Albuquerque: University of New Mexico Press.
Shackley, M. L. 1975. ArchaeologicalSediments.
New York: Wlley and Sons.
Stein, J. K. 1984. Organic matter and carbonates in
archaeological sites. Journal of Field Archaeology 11:239-46.
1985. Interpreting sediments in cultural settings.
IN J. K. Stein and W. R. Farrand, eds., Archaeological Sediments in Context, pp. 5-20. Orono:
Center for the Study of Early Man, University of
Maine.
Stein, J. K., and W. R. Farrand 1985. Archaeological
Sediments in Context. Orono: Center for the
Study of Eai¶y Man, University of Maine.
Twenhofel, W. H. 1950. Principles of Sedimentation. New York: McGraw-Hil.
8
NON-MARINE MOLLUSCS
FROM THE TO'AGA SITE SEDIMENTS AND
THEIR IMPLICATIONS FOR ENVIRONMENTAL CHANGE
PATRIcK V. KIRCH
INTRODUCTION
NON-MARINE MOLLUScs, or "land snails," can
occurin high frequencies at archaeological
sites where deposidonal environments are favorable
to their preservation (e.g., calcareous sands such as
at the To'aga site). In general, these snails are
diminutve (often in fte 1-3 mm size range), and
their presence in archaeological deposits results from
natual ratherthan cultual deposition. Since land
snails are highly sensidve to microenvironmental
conditions such as vegetation cover, changing
fruencies of snail shells in sratdfied archaeological
deposits can provide a proxy measure ofenvironmental change. Snail analysis for the purposes of
envionmental reconstruction has been carided out in
English and European archaeology for many years
(Evans 1972), but has been relatively neglected in
the Pacific. Beginng in the 1970s, Kirch and
Christensen began to apply snail analysis in Oceanic
arhaeology with fruitful results (Kirch 1975, 1989;
Christensen and Kirch 1981, 1986; Christensen
1983). bse studies have demonstrated that nonmarine molluscs in Pacific archaeological sites can
provide significant infoimation with regard to (1)
local environmental change (especially in vegetaton), and (2) the intrducdon of exotic biota by
preistoric humans. Local enviromnental changes
are demostrated trugh changing assemblages of
endemic and indigenous snail species which are
sensitive to microenvironmental conditions. The
introduction of exodc biota, on the other hand, is
signalled by the presence of one or more species of
synanthropic (or anthropophilic) snails that are
closely associated with human habitations, gardens,
and disuibed environments.
Pacific malacologists recognized some time ago
that a number of widely disseminated snail species in
te Pacific had almost cenainly been dispersed
trugh prehistoric human agency. C. Montague
Cooke, who spent a lifetime reearching Pacific snail
faunas, wrote that '"ere is no doubt that about a
dozen species were carned by Polynesians in their
migrations. At least four species were carried by the
latter to the Hawaiian Islands. These for the most
part are minute species of snails that are always
found in situations just above high-water mark and
are fairly uniformly distributed wherever
Polynesians live" (Cooke 1926:2279). Among these
synanthropic snail species are Gastrocopta
pediculus, La,nellidea pusilla, Liardetia samoensas,
Assiminea niida, and Lmnellaxis gracilis. Because
tese widely dispersed species are highly characteristic of atolls, they have sometimes been described as
an "atoll fauna" (Reigle 1964; Harry 1966), although
they are just as common in lowland, antrpogenic
habitats on high islands.
During fte 1987 excavations at To'aga, I
observed te presence of non-marine mollusc shells
in a number of sntaigraphic contexts, and therefore
undertook systematic sampling of fte 1987 main
trench and several oter test units for snail shells.
116
The To'aga Site
This chapter presents the results of analysis of these
samples and discusses the implications of our results
for paleoenvironerntal change and for the inrduction of exotic biota at To'aga.
MATERIAL AND METHODS
Sediment samnples for non-marine mollusc
analysis were bained from two columns at site AS13-1, one from the main trench of the 1987 excavation and the other from Unit 3. The main trch
column was taken from the cleaned west face of Unit
9, and fte nine column samples spanned the stratigraphic sequence from Layer IIC thrugh Layer IIA1. No land snail shells were observed below Layer
IIC or in the acidic colluvial sediments of Layer I.
The position of the land snail column in the main
tench is depicted on the main stratigraphic section
shown in figure 5.5.
For the most panl, the methods used in the
analysis of non-manne molluscs at To'aga follow
those developed and reported by Christensen and
Kirch (1986:55-56) for use in Hawaiian archaeological sites. Sediment samples for snail extraction were
taken as continuous columns. Prior to sieving, the
sediment samples were air-dried, te weighed and
measured volumetrically in order to assure quandtative comparability between samples. In the laboratory, the sediment samples were wet-sieved through
4, 2, and 1-mm mesh nested geological screens, and
the washed residues air-dried. Snail shells were then
extacted by hand-sorting under a lOX binocular
microscope. Snail shells were identified by referrng
to published taxonomic monographs and to reference
material in the collections of the Benice P. Bishop
Museum (Honolulu).
In the next section the various species of nonmarine molluscs present in the To'aga samples are
listed in taxonomic order, with remarks on their
distribution and ecology. This is followed by the
presentation of quantitative data on snail species
frequency from the two stratigraphic columns.
Family Assimineidae
Assuimnea cf. nitida (Pease)
The genus Assuimnea includes a number of
estuarine and standline-dwelling species (Abbou
1958). The most abundant land-snail species in the
To'aga site sediments is an Assuimnea (fig. 8.1)
which very closely resembles A. nida (Pease). This
species is known to be widely distibuted from
Soufteast Asia eastwards into Oceania.
Family Achatinellidae
Lamnellidea pusilla (Gould)
Lamellidea pusilla is (fig. 8.1) one of several
species in the Achadnellidae that have extensive
distributions throughout the inner Pacific. Cooke
and Kondo remarked that "'there is little doubt that
the wide distribution of L. pusilla is due to human
agency. It was probably tnsported from island to
island during the Polynesian migrations" (1960:188).
The species is distributed from the Maranas and
Palau groups in the westem Pacific and as far
eastward as Mangareva (Cooke and Kondo 1960:
fig. 81).
Family Pupillidae
Gastrocopta pediculus (Shuttleworh)
This small pupilHid (fig. 8. 1) is also abu tin
the To'aga samples. Pilsbry made the following
remarks on the distribution and geographic origin of
Gastrocopta pediculus:
Adaptability to life around habitations has, no
doubt, led to the vast Polynesian distribution of G.
pedculus. I infer that it has been carried from
island to island, sticking to native impedimenta,
cocoanuts, or other food mateials, in the thousand
years or more of inter-island canoe voyages of the
Polynesians. Its original habitat may have been
somewhere between the Philippines and New
Caledonia (1916-18:140-41).
Generally distributed in coastal environments, G.
pediculus shows a strong liking for coconut groves,
often being found in large numbers "under leaves
and sprouting cocoanuts" (Pilsbry 1916-18:148).
SYSTEMATIC REVIEW
Family Succineidae
Family Helicinidae
Pkuropoma sp.
A few specimens of Pleuropoma are present in
e To'aga samples. It has not been possible to
identify these to species level.
Succinea sp.
Only a single example of this species was
encountered in the To'aga site sediments, in Unit 3,
even though living Succinea sp. were observed in
leaf liUtter samples in the arboricultural zone.
Non-Marine Molluscs
:
:9!Ek
v
500rt7% °
igure 8.1
?
3 mm
Photomicrographs of terrestrial gastropods recovered from the To'aga
site sediments: a, Sinployea cf. allecta; b, Gastrocopta pediculus, c,
Lamellidea pusilla; d, Assirrinea cf. nitida; e, Lamellaxis gracilis.
117
118
The To'aga Site
Family Caropidae
Sinployea cf. allecta (Cox)
A few specimens of a caropid snail of the genus
Sinployea are present in several samples (fig. 8.1).
This is most likely S. allecta, although the closely
related species S. clausa (Solem) is also present in
the Manu'a Group. Solem (1982:127) has described
a subspecies, Sinployea allecta tauensis, from Ta'u
Island which closely matches the remains found at
the To'aga site. These species are apparently all
lowland dwellers, primarily in leaf litter.
Family Helicarionidae
Liardetia samoensis (Mousson)
This species occurs rarely in the To'aga samples.
Liardetia samoensis is the most widely distributed
member of the genus, ranging from te Bismarcks
eastward to the Marquesas (Solem 1959:96). Baker,
who monographed the zonitid snails of the Pacific,
observed that the genus Liardetia "contains species
that have been widely disseminated, probably
trugh the agency of man" (1938:12), among them
Liardetia samoensis. Solem (1959) regards the
probable center of distribution of Liardetia as
Indonesia.
Family Subulinidae
Lamellaxis gracilis (Hutton)
This distinctive species (fig. 8.1) is reasonably
well represented in te To'aga samples from both the
1987 main trench and from Unit 3. Pilsbry (1906)
remaked
L. gracilis is "probably the most
widely distributed land snail in the world," and
suggested that its dispersal throughout the inner
Pacific may have been due to the Polynesians. Other
malacologists (Cooke 1934; Solem 1978) regarded
its wide distribution as a result of modem commerce.
Christensen and Kirch (1981) demonstrated, however, that Lamellaxis gracilis was present on the
Polynesian outlier of Tikopia by 900 B.C., where
Lapita colonists presumably introduced it inadvertently. Subsequently, specimens of L. gracilis have
been identified from prehistorc contexts in Tonga
(Niuatoputapu, in association with Lapita sites,
Kirch 1988:233-34), the Marquesas (Rolett 1989,
table 5.14), and Hawai'i (Christensen and Kirch
1986:60). The dissemination of this synanthrmpic
species probably began with the Lapita expansion
and continued with the Polynesian dispersal into the
far eastem Pacific.
RESULTS
The frequencies of non-marine mollusc species
in the various sample units from the 1987 main
excavation, and from Unit 3, are presented in tables
8.1 and 8.2. In addition to the raw frequency counts
per sample, the tables provide infornation on sample
weight and volume, and on te density of snail shells
per cubic liter of sediment.
The nine sediment samples from the main trench
(Unit 9) spanned te Layer II depositional sequence
as follows: Samples 1-3, Layer IIC; Samples 4-6,
Layer UB, the main occupation horizon; Samples 78, Layer IIA; and Sample 9, Layer IIA-l. Several
significant pattems are evident from te data presented in table 8.1. First, the assemblage is dominated by a suite of synanthropic species, particularly
Assiminea cf. nitida, Gastrocopta pediculus,
Lamellidea pusilla, and Lamellaxis gracilis, but with
Liardetia samoensis also represented. Indeed, the
dominance of synanthropic species in xse samples
is remarkable, greater than any oter archaeological
example that I am aware of from the Pacific. The
two indigenous/endemic species present,
Pleuropoma sp. and Sinployea sp., are represented
only in very limited numbers.
The temporal distribution of species is also
striking. The oldest sample contains only one
species, Assiminea cf. nitida. This species is joined
by Lamellidea pusilla and Gastrocopta pediculus in
Sample 2, and by Lamellaxis gracilis and Liardetia
samoensis in Sample 3. Thus, by the time of
occupation represented in Layer IIB, ca. 2500 cal
B.P., five species of synanthropic, human-transported
land snails had been introduced to, and become
established on, the coastal terrace of Ofu Island.
Because Lamellads gracilis is particularly associated with human gardening sites, its appearance by
2500 cal B.P. is important in tenns of the economic
prehistory of the To'aga site.
As can be seen from table 8. 1, the density of
non-marine molluscs is extremely low (8 snaiLs/I3) in
the lower part of Layer IIC, rising rapidly to a peak
density of 400 snails/I3 in the middle of Layer IIB,
the occupation horizon. Shell density then drops
Non-Mari'ne Molluscs
119
Table 8.1
Non-marine Molluscs from the 1987
Main Trench, To'aga Site
Sample No.
Taxon
1
2
3
4
5
6
7
8
9
Pleuropoma sp.
Assiminea sp.
Lamnellidea pusilla
3
5
4
9
1
17
26
26
1
19
12
37
51
10
40
1
59
5
17
1
46
4
18
2
47
1
8
3
84
14
19
6
75
269
9
110
400
8
90
321
4
18
65
1
5
76
276
73
203
3
64
Gastrocopta pediculus
Sinployea sp.
Liardetia samoensis
Lamellaxis gracilis
Total snails counted
3
8
Snails/13*
*
197
11
128
366
standardized density, adjusted for sample volumes.
somewhat in Layer IHA, but rises again to 366
mails/13 in the Layer IIA-I deposit. This distribution
is sigficant, since it correlates well with the sedimentological history of the column and is precisely
the sort of overall density distribution that would be
expected with the establishment of human habitations in Layer IIB and of a stable vegetated soil
surface represented by Layer IIA-1.
The land snail data for Unit 3 are presented in
table 8.2. The overall pattem resembles that in the
Table 8.2
Non-marine Molluscs from Unit 3,
To'aga Site
Sample No.
Taxon
Pleuropoma sp.
Assiminea sp.
Lamellidea pusilla
2
10
2
Gastrocopta pediculus 6
Succinea sp.
Sinployea sp.
Liardetia samoensis
3
Lamellaxis gracilis
Total snails counted 21
71
Snails/13*
*
3
4
1
27
6
9
44
3
7
1
2
8
52
184
10
66
280
5
1
110
18
27
1
4
1
13
175
735
standardized density, adjusted for suample volumes.
6
15
3
18
68
120
The To'aga Site
main trench, except that the suite of synanthropic
snails is already established at the base of the column
(Sample 2). The increasing density of land-snail
shells and the dominance of synanthropic species
suggest the establishment of a stable, anthropogenic
vegetation over the To'aga coastal terrace following
human occupation of the island.
CONCLUSION
The analysis of non-marine molluscs from the
To'aga site provides important conroborative evidence for several interpretations made elsewhere in
this volume. That the snail assemblages at To'aga
are dominated by a suite of synanthropic species (the
so-called "atoll fauna') known to have been widely
disseminated by human agency highlights our
interpretation of the To'aga coastal terrace as a
strongly anthropogenic habitat. The absence of
endemicfindigenous species below Layer IIC in the
main trench supports our interpretation of the coastal
terrace as a narrow, exposed, calcareous depositional
environment prior to human colonization. After
human settlement and the establishment of crop
plants and other introduced vegetation at the beginning of the first millennium B.C., the coastal terrace
began to sustain an adventive snail fauna of
synanthrpic species. The high concentration of
these snails in Layer HA- I is also consistent with our
interpretation of this deposit as a fonner stable,
vegetated A1 horizon paleosol.
The snail assemblages from To'aga further
document of the early spread of synanthropic species
into central Polynesia. Such species as Lamellaxis
gracilis, associated with Polynesian gardening
environments, were already known to have been
present in Lapita contexts (Christensen and Kirch
1986; Kirch 1988). The presence of this and four
other species at To'aga reflects the propensity of
early Pacific colonizers to create "transported
landscapes."
REFERENCES CITED
Abbott, R. T. 1958. The gastropod genus Assiminea
in the Philippines. Proceedings ofthe Academy
of Natural Sciences ofPhiladelphia 110:21378.
Baker, H. B. 1938. Zonitid Snailsfrom Pacific
Islands, Part 1. Bemice P. Bishop Museum
Bulletin 158. Honolulu.
Christensen, C. C. 1983. Analysis of land snails. IN
Archaeological Investigations of the WaimeaKawaihae-Mudlane Road Corridor, Island of
Hawaii: An Interdisciplinary Study of an
Environmental Transect, eds. J. T. ClaIk and P.
V. Kirch, pp. 449-71. Departnental Report 831, Anthropology Departnent, Honolulu:
Bernice P. Bishop Museum.
Christensen, C. C., and P. V. Kirch 1981. Nonmarine
mollusks from archaeological sites on Tikopia,
Southeastem Solomon Islands. Pacific Science
35:75-88.
. 1986. Nonmarine mollusks and ecological
change at Barbers Point, O'ahu, Hawai'i.
Bishop Museum Occasional Papers 26:52-80.
Cooke, C. M., Jr. 1926. Notes on Pacific land snails.
Proceedings ofthe Third Pan-Pacific Science
Congress, pp. 2276-84.
. 1934 Land shells of Makatea. Occasional
Papers of the Bernice P. Bishop Museum 10:111.
Cooke, C. Montague, Jr., and Y. Kondo 1960.
Revision of Tornatellinidae and Achatinellidae
(Gastropoda, Pulmonata). Honolulu: Bemice
P. Bishop Museum Bulletin 221.
Evans, J. G. 1972. Land Snails in Archaeology.
London: Seminar Press.
Harry, H. A. 1966. Land snails of Ulithi Atoll,
Caroline Islands: A study of snails accidentally
distributed by man. Pacific Science 20:212-23.
Kirch, P. V. 1988. Niuatoputapu: The Prehistory of
a Polynesian Chiefdom. Thomas Burke
Memorial Washington State Museum Monograph No. 5. Seattle: Burke Museum.
-. 1989. Non-marine molluscs from the
rockshelter sediments. IN Prehistoric Hawaiian Occupation in the Anahulu Valley, O'ahu
Island: Excavations in Three Inland
Rockshelters, ed. P. V. Kirch, pp. 73-82.
Contributions of the University of Califomia
Archaeological Research Facility No. 47.
Berkeley.
Pilsbry, H. A. 1906. Manual of Conchology, Second
Series: Pulmonata, Vol. XVIII, Achatinidae:
Stenogyrinae and Coeliaxinae. Philadelphia:
Academy of Natural Sciences.
. 1916-18. Manual of Conchology, Second Series:
Non-Marine MoUuscs
Pubdonata, Vol. XXIV, Pupillidae
(Gastrocopdnae). Philadelphia: Academy of
Natua Sciences.
Rdgle N.J. 1964. Nonmaine mollusks of Rogeap
AUlMaral slanids Pac#c Science 18:126-29.
Rolct, B. V. 1989. Hanai: Changing subsistence
and ecology in the prehistory of Tahuata
(Marquesas Islands, French Polynesia). Un-
121
published PtLD. dissertation, Yale University.
Solem, A. 1959. Systemadcs and Zoogeography of
the Land and Fresh-water Mollusca of dte New
Hebrides. Fieldiana: Zoology 43. Chicago.
. 1982. Endodontid Land Snailsfrom Pacfic
Islands, Part l, Familes Punctidae and
Charopidae, Zoogeography. Qiicago: Field
Museum of Natuml History.
9
THE TO'AGA CERAMICS
T. L. HUNT AND C. L. ERKELENS
CsAmi HAvE PLAYED a ciitical role in understdig prehistory in Samoa and West
Polynesia. They are usually well peserved,
archaeologically visible, and carry a large amount of
infoimation on variation in style (temporal and
spatial), technology, function, and raw material.
While ceramics have proven useful for making
culture-iistorical inferences about Samoa and the
region, ftey also present several interesting problems
to be resolved in their own righlt First, what is the
nature of ceramic variability (temporal and spatial)
in Samoa? Second, what kinds of change (stylistic,
techoogicalimaterial, and functional) occuned over
the duation of ceramic prodution in Samoa? And
third, why did ceramics disapear in Samoa after an
appoximate thousand-year sequence of production?
In this chapter we describe the ceramic assemblage
from the To'aga excavations and begin to address
hese questions through analyses of ceramic clay
composition, technology, function, and style.
THE ASSEMBLAGE
The 1986 excavations on Ofu and neighboring
Ta'u produced a total of 147 sherds (32 sherds from
To'aga), and were analyzed and reported in Hunt and
Kirch (1988:169-71). [be 1987 field season added
1464 sherds to the To'aga assemblage. These sherds
were described by Kirch et al. (1990:7-8) and are
further analyzed here. Excavations in 1989 provided
an additional 938 sherds to the previous total. In
sum, hre seasons of fieldwoik atTo'aga (AS-13-1)
have yielded a total ceramic assemblage of 2434
sherds.
The To'aga ceramic assemblage is significant in
several respects. The sherds come from a deep,
well-stratified site dated with several radiocarbon
determinatons (see chapters 5 and 6). The assemblage spans the full duration known for ceramic
production in Samoa, and consists of quantities of
the pottery distinguished as "Ihickware" and
"'thinware" (see Green 1974; Holmer 1980, Clark
and Herdrich 1988). Pottery from To'aga, Samoa's
eastemnmost ceramic-bearing site, dates to the firt
millennium B.C. Finally, the To'aga assemblage is
among the largest excavated in Samoa-only the
SU-SA-3 (Green 1974) assemblage from Vpolu is
larger-and thus provides an adequate sample to
assess several dimensions of variability.
The entire ceramic assemblage has been cataloged, with individual sherds enumerated. In the
most general terms, the assemblage can be divided
into broad classes of thickware and thinware sherds.
This division of wares is based on sherd thickness,
temper size, and paste texture. Such distinctions are
qualitative (and somewhat impresionistic), however, and do not accurately reflect the range of
variability present in the assemblage,-hence the
need for detailed analysis. Vessel parts include only
direct rims (no necks) and body serds, indicative of
a single class ofvessel forms comprising open,
round-based bowls. While occurring in only minor
124
The Tolaga Site
frupencies, classes ofdecoration include red-
slipped, carved paddle-imprsed, incised, and rm
sherds with notched, impesd, or crenelated lips.
Frequencies of sherds by these broad classes in
stratigraphc context are sunmarized in tables 9.1-2.
Detailed analysis of ceramics is extremely labor
intesive. Observations and measurements on
individual sherds can require as much as five to ten
minutes eacL This problem requires selecting a
representative sample from the larger assemblage.
In the case of To'aga, we chose excavation units with
largersamples. Also, pottery was selected from
excavation units to span the full temporal sequence
represented at To'aga. This stagy enabled us to
assess change across the fu sequence of pottery
manufacture at Toaga.
Sherd samples were drawn from the main areal
excavation of 1987 (units 1 and 4-9, see Kirch et al.
1990 and chapters) and from units 27 and 28
excavated in 1990. The total sample selected for
intensive analysis is 737 sherds.
The main areal excavation of 1987 (units l and
4-9) contained a large number of sherds (n=527)
within a well-dated, statified context The primary
ceramic-bearing occupation layers are dated (Beta25033, 25034) to an averaged coneted range of
306-138 B.C. at one stard deviation (see chapter 6
and Kirch et al. 1990). Units 27 and 28, excavated
in 1989, were selected because they provided large
samples from early contexts (see chapters 5 and 6).
Based on radiocarbon dates and statigraphic
correlations, the sheris for intensive analysis can be
divided into early, middle, and late periods for
comparative prposes. The early ceramics, from
layer III in units 1, 5-7, 9, and 28, range from
approximately 1250-500 B.C. Middle period ceramics, from layer I (B and C) in units 1, 4-7, 9, 27, and
28, range from about 500 B.C. to the beginning of the
Chuistian era. The late period sample, from layer II
(A) in units 4-9 and 27, dates from the time of Christ
and may span the first 200-300 years A.D. These are
not "ceramic periods" or "hases," but simply
represen a three-pan division devised to analyze
change in the sample.
The pol for intensive analysis was designed
to assess variability in raw materials, technology,
style, and fuction. In addition to recoding provenience (by unit, stratum, and level) and catalog
number for each sherd, the following analytic
protocol was used and observations or measurements
coded for analysis (using SPSS- PC+; Nonrsis
1986):
1) Exterior surface Etment:
0. eroded ("missing data")
1. plain (untextred)
2. wiped (striations present)
3. puddled
4. slipped
5. carved paddle impressed
6. residue obscuring ("missing data")
2) terior surface Ltatment:
(same criteria as above)
3) Orientation ofinclusions ("preference') relative to
vessel walls:
1. indeterminate (i.e. no definitive long axis to
gais)
2. random orientation
3. paalel
4. perpendicular
4) Interior anvil casts:
0. indeterminate ("missing data")
1. present
2. absent
5) Exterior paddle marks casts:
0. indeterminate ("missing data")
1. present
2. absent
6) Exterior hardness (Mohs scale)
7) Interior hardness (Mohs scale)
8) Exterior surface color (Munsell):
0. erded ("missing data")
1. 75R (& value & chroma for all modes)
6. 7.5YR
2. SR
7. 10YR
3. IOR
4.2.5YR 8.2.5Y
9.5Y
5.5YR
9) Interior surface color (Munsell):
(recorded same as above)
10) Oxidation/reduction pattern (using pattern
template)
11) Organic residue:
0. absent
1. interior
2. exterior
3. both surfaces
12) Weight of sherd (grams)
Ceramics
13) Mean sherd thickness (mm, 3 measuements/3)
14) Variance in sherd thickness (mm, maximum
minus minimum value)
15) Size modality of temper (Wentworth scale using
grain size template, sherd viewed under lOX
magnification):
0. canrKot determine temper size modality ("miss-
ing data")
1. granule (-1 phi; >2.0 mm)
2. bimodal granule, very coase sand
3. very coae sand (0 phi; >1.0 mm)
4. bimodal very coarse, medium sand
5. medium sand (2 phi; >0.25 mm)
6. bimodal medium, fine sand
7. fine sand (3 phi; >0.125 mm)
8. bimodal fine, very fine sand
9. very fine sand (4 phi; >0.0625 mm)
16) Rank order of temper by material in hand
specimen only (fist temper by rank, second,
tIrd); materials code:
0. indeterminate ("missing data")
1. black tachyte
2. green olivine
3. gray basalt
4. clear tanslucent crystals (quartz)
5. calcaous sand
6. ferous peds
7. opaque feldspathic crystals
17) Decoration technique (note: carved paddleimpressed and red slip are included under the
dimension of surface treatment):
0. undecorated (or eroded, "'missing data")
1. tool impressed
2. incised
Rims Only
18) Rim course (angle to central vertical axis):
1.direc
2. inverted
3. evented
19) Cross-section of lip:
1.onded
2. square
3. poned
>2) Rim profile (degree of thickening):
1. noe, parallel walls
2. tined
3. thickened, exterior
4. thiickted,inteior
125
21) Mean thickness of rim at lip (mm, 3 measures/
3)
22) Variance in thickness of rim at lip (mm, maximum minus minimum value)
23) Estimated orifice diameter (10 cm intervals)
ANALYTIC PROCEDURES
Observation of exterior and interior surfaces
allows one to determine the final finishing metxhds
used on vessels prior to firing. Where sherds have
an eroded exterior or interior surface, such observations could not be made and were coded as "missing
data." Plain surfaces were simply smooth, with no
other finishing techniques evident. Wiped surfaces
were recognized by fine, parallel striations that were
made in the clay pror to firing. Puddled surfaces,
sometimes called "self-slipped" (Rye 1981:57), are
those formed by wetting the surface, thus bringing
the finest clay particles to the surface of the paste.
This technique is recognized by textural differences
on the sherd surface and in cross-section. Slip, a
thin surface coating created from a fluid suspension
of clay in water, commonly has a different color than
the body (Rye 1981:41). Also, slip diffesfrom the
sherd body in texture. Carved, paddle-impressed
surface treatment is usually produced with secondary
forming using a paddle and anvil. The textured
(patterned) paddle leaves a raised design on the
vessel surface.
The "preferred" orientation of sand temper
inclusions relates to techniques of primary forming
(Rye 1981). Slab building and coiling can be
detected, in part, by differences in orientation. A
definitive long axis for the particles is necessary to
observe orientation.
Interior anvil casts and exterior paddle (carved
and plain) maiks were noted as present/absent
These traits reveal the use of the paddle and anvil
technique in secondary forming in the ceramic
production sequence.
Hardness was measured on the interior and
exterior surfaces of sherds by a scratch test using the
Mohs scale. Hardness, measured on this ordinal
scale, may relate to ceramic strength, raw materials,
firing, and post-depositional diagenesis.
Sherd color was measured on the exterior and
interior surfaces using the Munsell Soil Color Chart
(1988). Ceramic color, while complex, reflects
126
The Tolaga Site
Table 9.1
Ceramics from the 1987 Excavation Units
Unit Layer
Thin
n (%)
Thick
n (%)
Rims
Other
Sherds
Red
Slip
Temporal
Analytic
Period
Total
1
IIB
IIC
III
2 (3)
0 (0)
2 (67)
66 (97)
0 (0)
1 (33)
10
1
0
Middle
Middle
Early
68
1
3
4
IIA
IIB
0 (0)
4 (44)
7 (100)
5 (56)
0
0
Late
Middle
7
9
5
IIA
IIB
III
4 (14)
12 (27)
8 (100)
24 (86)
32 (73)
0 (0)
0
4
3
Late
Middle
Early
28
44
8
6
IIA
IIB
IIC
III
2 (8)
2 (13)
17 (74)
4 (80)
22 (92)
13 (87)
6 (26)
1 (20)
3
0
0
2
Late
Middle
Middle
Early
24
15
23
5
7
IIA
IIB
IIC
III
1
5
5
4
(13)
(71)
(83)
(80)
7 (87)
2 (29)
1 (17)
1 (20)
1
0
0
.1
Late
Middle
Middle
Early
8
7
6
5
Late
27
Late
Middle
Early
69
22
3
Late
Middle
4
1
Late
Middle
Middle
8
98
2
Early
20
Late
Middle
182
95
8
IIA
0 (0)
27 (100)
0
9
IIA
IIB
III
2 (3)
2 (9)
1 (33)
67 (97)
20 (91)
2 (67)
11
IIA
IIB
1 (25)
1 (50)
3 (75)
0 (0)
0
0
11
IIA
IIB
IIC
3 (38)
8 (8)
1 (50)
5 (62)
90 (92)
1 (50)
2
11
0
12
III
20 (100)
0
3
14
IIA
IIB
7 (4)
10 (11)
175 (96)
85 (89)
10
0
128 (15)
663 (85)
(62)
10
Totals*
*
Rims, slipped, and decorated
0
Incised (1)
0
1
Notched rim (1)
791
sherds included with thick/thin sherd counts; eroded sherds not included in total.
Ceramics
127
Table 9.2
Ceramics from the 1989 Excavation Units
Temporal
Thin
Unit Layer
15
23
24
9 (90)
22 (73)
13 (87)
2 (50)
(0)
(0)
12 (40)
9 (100)
3 (100)
18 (60)
IIIC
4 (7)
12 (9)
10 (22)
IIA
IIB
IIIA
IIIB
I
0
0
Rims
Slip
1
0
2
0
Incised (1)
0
53 (93)
120 (91)
36 (78)
4
10
10
1 (25)
2 (12)
3 (75)
14 (88)
0
IIIC
8 (13)
32 (35)
17 (28)
54 (87)
60 (65)
44 (72)
10
3
IB
1 (33)
2 (67)
0
II
(0)
(8)
3 (30)
IIIA
IIIB
IIA
IIB
IIC
29
II
IIIA
5
(100)
0
5
57 (92)
7 (70)
3
1
(100)
16 (70)
52 (50)
(0)
12 (100)
2 (22)
0
7 (78)
1
Impressed (6)
0
4
2
1
Analytic
Period
Total
Middle
Early
Early
Early
10
30
Late
Middle
9
3
30
Early
3
0
0 (0)
7 (30)
51 (50)
Other
Sherds
3
2
1
IIA
IIIB
4
57
132
Middle
Middle
4
16
Early
Early
46
1
Early
62
92
61
1
Middle
3
Late
Middle
Early
62
Decorated (4)
5
10
0
Late
1
2
13
Middle
Middle
23
103
1
Middle
Early
12
9
34
39
0
9
Impressed (1)
II
2 (6)
32 (94)
28 (72)
1
Middle
11 (28)
6
Early
198
674
(68)
Totals*
15
Early
Early
Early
IIIA
30
*
1
Red
2
27
28
(%)
(10)
II
III
21
n
8 (27)
2 (13)
2 (50)
II
IIIA
IIIB
IIID
16
20
n (%)
Thick
872
Rims, slipped, and decorated sherds included within thick/thin sherd counts; eroded sherds not included
in total.
128
The To'aga Site
vaniability in raw material, pyrotechnology, and use.
The oxidation/reduction pattern was recorded
for sherd cross-sections using an inductively derived
pattern template (e.g., Rye 1981:116). These
patterns are indicators of the atmosphere and
temperature of firing (Rye 1981:115-18).
The presence or absence of organic (carbonaceous) residue was recorded. Residues supply
important clues of ceramic function.
Variation in sherd thickness has proven significant in Samoan pottery studies. Sherd thickness is
estimated as a mean value from three measurements
on different parts of the sherd. This mean value is a
more reliable measure of sherd thickness than a
single measurement (Barry 1978). Three thickness
measures also provide a range expressing the
variance in sherd thickness. Variance measures the
uniformity or evenness of sherd thickness.
Temper was examined in terms of its rank by raw
material as estimated from hand specimens only.
Also, the size modality of the sand temper grains
was recorded using a template produced with sand
samples of varying phi sizes on the Wentworth scale.
Sherds were viewed under lOX magnification.
Dickinson's study of sand temper petrography offers
a much more detailed and reliable (some materials
are difficult to determine in hand specimen alone)
analysis of a selection of sherds from the To'aga
assemblage (see chapter 10).
For sherds with one or more intact surfaces,
decoration technique was recorded as tool-impressed, incised, or undecorated (plain). Carved
paddle-impressed and red slip, while decorative
techniques, are included under the dimension of
surface treatment.
Rim sherds were identified in terms of their
course relative to the central vertical axis of the pot
(e.g., direct, everted). The cross-section of the lip
(form) and the profile of the rim, or the degree of
thickening, were also classified. Finally, the thickness of each rim was measured, and orifice diameters were estimated.
Of the 737 sherds sampled for intensive analysis, 583 sherds retained both interior and exterior
surfaces enabling measurement or observation of
the characteristics listed above. The results of
intensive macroscopic analysis are summarized and
discussed below. These results, together with
compositional results (below and chapter 10),
provide the basis for reconstructing aspects of
ceramic raw material use, production technology,
style, and function.
RESULTS OF MACROSCOPIC ANALYSIS
Many of the results used for intensive analysis
in tables 9.3-13. Specific characteristics that relate to material use, forming techniques,
pyrotechnology, style, and function are discussed
below.
The analysis of sand temper composition and
grain size modality (see table 9.4) revealed the
following: temper ranges in size from grains
measuring approximately 2.5 4 mm (i.e., granules
are summarized
-
Table 9.3
Frequency of Body and Rim sherds
Frequency
Percent
Cum.
Percent
Body
Rim
683
54
92.7
7.3
92.7
100.0
Total
737
100.0
Sherd
(Valid cases
=
737; Missing cases = 0)
Ceramics
129
Table 9.4
Frequency of Sherds by Temper Size Mode(s)
Temper Size Modes
Frequency
Percent
Cum.
Percent
granule
granule, very coarse
very coarse
very coarse, medium
medium
medium, fine
fine
30
189
140
114
81
91
92
4.1
25.6
19.0
15.5
11.0
12.3
12.5
4.1
29.7
48.7
64.2
75.2
87.5
100.0
Total
737
100.0
(Valid cases = 737; Missing cases = 0)
< -1 phi) in many sherds, to particles visible only
under magnification (i.e., fine to very fine sand, > 3
phi). No attempt was made to observe modes of
finer (e.g., silt) particles.
Most of the sherds in the sample assemblage
(86%) had a mixture of sands and contained more
than one temper compositional class. Thicker sherds
tended to have coarser temper and thinner sherds
tended to have finer temper. Temper size correlates
to temper composition (x2 = 237.21, df= 9, p >
.0001) in that well-rounded, calcareous sand grains
ame smaller than the very coarse (often angular) sand
class (O phi). Grain size and temper composition are
also associated with sherd thickness. Calcareous
sad temper occurs most often in thinware (73% of
this temper occurs in thinware). Yet, thinware
cannot be described as predominantly calcareous
sand tempered because only 26% of the thinware has
a ped
anceof calcareous sand temper. On the
oter hand the largest tempers, grains of glassy black
tm~hyte, dull grey basaltic, rounded olivine, pale
p feldspars, clear crystalline, and red ferrous,
rounded grains were resented in the smallest size
classes as well.
Color variability was assessed by plotting values
in ascateplot to check for trends or grouping
tendenies. Bivariate plots for a three dimensional
cassification were achieved by plotting Munsell hue
against value-chroma (figures 9.1-3; e.g., see Bishop
et aL 1988). These plots reveal that color variability
in the To'aga ceramics is comparable for all hree
time periods and for both thick- and thinwares.
Sherds range from red (lOR) to yellow-red (lOYR)
and cover an array ofvalue and chroma The
majority of the assemblage is reddish brown or gray
in color. Early ceramics include red-slipped sherds
(10 R 5/6 and 2.5 YR 5/6) distinctive in coloration
from the west of the assemblage.
Although analysis showed that interior and
exterior surface colors have similar distribution, the
conrelation on individual sherds was poor (Pearson's
r = 0.143, p > 0.001). Oxidation/reduction patterns
(table 95) play some part in this correlation, since
24% of the sherds have interior surfaces that were
not fully oxidized as compared with only 3% for the
exterior surfaces. As a result, more sherds have
interior surfaces that are darker than their exterior
surfaces. Most of the sherds (68%), however, were
fully oxidized during firing.
Following analysis, the entire assemblage was
inspected to ceck for further variation in color (or
other unrecorded differences in temper, etc.). Five
sherds, from units 15129/30 were anomalous in color.
These sherds (5 YR 6/6 and 7.5 YR 6/3) are described respectively as reddish yellow and light
brown in the Munsell system. The former group of
thickware sherds (5 YR 6/6) is similar in color to the
predominate color of the sberds from Upolu Island
designated "Falemoa Tan" by Holmer (1980:114).
Sherd hardness, measured on the Mohs scale,
130
The Tolaga Site
I
I
I
I
I
C
I
I
I
I
I
I
1
6-
1
3
1
E
7B
C
H
U
K
1 1
4.5*815
327I
I
I
1
I
I
27
I
36
I
45
I
1
I
I
I
I
63
54
I
I
VALUE AND CHROMA
Figure 9.1
Plot of exterior sherd color (hue with value and chroma) for the late period (note to figures 9.1-3: number
of cases in each position is shown in sequential order by 1-9; A-; 10-35 cases; and *, more than 36 cases).
I
I
I
5
I
I
I
I
1
* 22
I
I
I
1
2
6-
I
K 1
B
H
0
U
3 1
4J
Q*
4.5-
*1 4
1
1
3I
I
2
I
27
9
I
3
36
I
4
I
45
5
54
9
I
6
I
63
VALUE AND CHROMA
Figure 9.2
Plot of exterior sherd color (hue with value and chroma) for the middle period (see
note to fig. 9.1).
I
I
Ceramics
I
I
1
I
I
I
I
I
I
I
I
131
I
1
X 11
6-
1
1
IR
1D
H
2
U
1 1
B
4.5R7 2
3I
*
2
*
27
3I
I
*
*
*
I
36
4
*
*
I
5
*
54
45
*
I
6
*
-
lF
63
VALUE AND CHROMA
Figure 9.3
Plot of exterior sherd color (hue with value and chroma) for the early period (see note to fig. 9.1).
ranged from 2.0 to 8.0 (tables 9.6-7). This is a
remaikable range ofhardness, but reflects the
difficulty of measuring paste hardness with a scratch
test in sherds with abunant temper (which is
consistently hard). Ceramic tiles manufactured from
colluvial clay samples and fired in an open fire (for
appoximately fifteen minutes), or in the muffle
furnace (5000C for fifteen minutes) were uniformly
3 in hardness. There is a strong correlation
(r = 0.8467, p > .0001) between the interior and
exterior hardness of individual sherds.
The sample analyzed (n = 737) had 583 sherds
which retained both surfaces (i.e., non-eroded),
allowing the measurement of mean thickness for
each sherd. Results show that sherds range from
4.20 mm to 16.97 mm in thickness. The total sample
mean sherd thickness is 9.42 mm (a = 2.84), however the distribution is not normal, but has hree
modes (figure 9.4). The late portion of the sample
(n = 196) with 140 measurable sherds, ranges from
5.60 mm to 14.55
mm
in thickness.
mean sherd thickness is
late
sample
10.14 mm (a = 1.77) (figure
The
middle period sample (n = 41 1) with 323
9.5).
measurable sherds, ranges from 4.20mm to 16.97
mm in mean sherd thickness. The sample mean
thickness is 9.97 mm (a = 3.10) (figure 9.6). The
early ceramics (n = 130) had 120 measurable sherds
with a mean thickness of 7.09 mm (a = 1.68). These
sherds range in thickness from 4.30 mm to 14.52
mm in a somewhat normally distributed range of
measurements (kurtosis 4.357, skewness 1.783), but
skewed toward the timner sherds (figure 9.7).
Differences in sherd thickness are summanzed by
temporal-analytic periods in table 9.14.
Rims for the entire To'aga assemblage were
analyzed and identified to class (see protocol).
These forms are illustrated in figure 9.8. One large
reconstructed sherd provided a measurable portion
of a rim
(9% of the estimated total)
so
that the
diameter is reconstructed as 48 cm. About 89% of
the rims are oriented 900 to a central vertical axis of
the vessel. Other rims have angles approximating
800. All rim courses are direct, with the majority
(80%) having a squared lip cross section. The
remaining rims (20%) are rounded in cross section.
The only decorated rims are from early contexts,
comprising approximately 7% of the total. The lips
of these decorated rims are impressed with narrow
tools forming U-shaped notches, repeating parallel
lines perpendicular to the rim, or in one case, oblique
to the rim. A rim with a crenelated lip was also
recovered.
Other decorated sherds are small in number.
Only 30 slipped sherds and 1 1 other decorated
132
The To'aga Site
Table 9.5
Frequency of Sherds by
"Preferred" Orientation of Inclusions
Relative to Vessel Walls
Inclusion
Orientation
Frequency
Percent
Cum.
Percent
Indeterminate
Random
Parallel
199
458
80
27.0
62.1
10.9
27.0
89.1
100.0
Total
737
100.0
(Valid cases = 737; Missing cases = 0)
Table 9.6
Frequency of Sherds by Oxidation-Reduction
Pattern in Cross-Section
Pattern
Fully oxidized
Core oxidized
Ext. oxidized
it. surf. reduced
Fully reducd
Ext. surf. reduced
Surfs. rediced
Ext. reduced
Ext. surf.reduced
int. reduced
Total
Percent
Valid
Percent
Cum.
Percent
380
51.6
67.9
67.9
10
68
1.4
9.2
7.5
1.8
12.1
9.8
69.6
81.8
91.6
Frequency
55
9
14
2
7
1.2
1.6
93.2
1.9
.3
.9
2.5
.4
1.3
95.7
96.1
97.3
100.0
15
2.0
2.7
177
24.0
Missing
737
100.0
100.0
(Valid cases = 560; Missing cases = 177)
Ceramics
Table 9.7
Frequency of Sherds by
Exterior Hardness (Mohs Scale)
Exterior
Hardness
Frequency
Percent
Valid
Percent
Cum.
Percent
2
3
4
5
6
7
8
Eroded
31
113
109
154
145
53
18
114
4.2
15.3
14.8
20.9
19.7
7.2
2.4
15.5
5.0
18.1
17.5
24.7
23.3
8.5
2.9
Missing
5.0
23.1
40.6
65.3
88.6
97.1
100.0
Total
737
100.0
100.0
(Valid cases = 623; Missing cases = 114)
Table 9.8
Frequency of Sherds by
Interior Hardness (Mohs Scale)
Frequency
Percent
Valid
Percent
Cum.
Percent
2
3
4
5
6
7
8
Eroded
24
110
122
141
158
52
31
99
3.3
14.9
16.6
19.1
21.4
7.1
4.2
13.4
3.8
17.2
19.1
22.1
24.8
8.2
4.8
Missing
3.8
21.0
40.1
62.2
87.0
95.1
100.0
Total
737
100.0
100.0
Interior
Hardness
(Valid cases = 638; Missing cases = 99)
133
134
The To'aga Site
Table 9.9
Frequency of Sherds by Exterior Surface Treatment
Surface
Treatment
Frequency
Percent
Valid
Percent
Cum.
Percent
Plain
Wiped
Puddled
Slipped
Eroded
128
13
481
4
111
17.4
1.8
65.3
.5
15.0
20.4
2.1
76.8
.6
Missing
20.4
22.5
99.4
100.0
Total
737
100.0
100.0
(Valid cases = 626; Missing cases = 111)
Table 9.10
Frequency of Sherds by Interior Surface Treatment
Surface
Treatment
Valid
Percent
Cum.
Percent
18.9
2.1
77.3
1.4
0.3
18.9
21.0
98.3
99.7
100.0
Frequency
Percent
Residue
Eroded
124
14
508
9
2
80
16.8
1.9
68.9
1.2
0.3
10.9
Missing
Total
737
100.0
100.0
Plain
Wiped
Puddled
Slipped
(Valid cases = 657; Missing cases = 80)
Table 9.11
Frequency of Sherds by Interior Anvil Casts
Anvil
Casts
Frequency
Percent
Valid
Percent
Cum.
Percent
Absent
Present
Indeterminate
85
297
355
11.5
40.3
48.2
22.3
77.7
Missing
22.3
100.0
Total
737
100.0
100.0
(Valid cases = 382; Missing cases = 355)
Cerenics
Table 9.12
Frequency of Sherds by Exterior Paddle Marks
Frequency
Percent
Valid
Percent
Cum.
Percent
Absent
Present
Indeterminate
54
282
401
7.3
38.3
54.4
16.1
83.9
Missing
16.1
100.0
Total
737
100.0
100.0
Paddle
Marks
(Valid cases = 336; Missing cases = 401)
courht
25
_~ granule,
_~
20
-
-
10
--
5
-
09A
v. coarse,
medium
medium, fine
a=
15
v. coarse
4
5
6
7
8
9
10
11
12
fine, v. fine
13
14
15
16
thickness (mm)
Figur 9.4
Thickiess and temper size histogram for the Toaga ceramic assemblage (n = 538). Temper
classes are indited by tew variable shading of the histam bars.
135
The To'aga Site
136
N
d
E
*
U,
*IC2
0
oo: o! oqfio
v
00 -401
N 00"()o
*aC
Ch _
at
o
;
00
e
6
6
0
00
(%.-
x
0
0
8._o
m
EI"
0
N
I
(A
00
C'
;Ob
0*
W
~u
"
0%.
0
si
U
_m
N
N
.2~~~~~~C
e
6
9
6
00
I%
0
3
11
U
00
O r
~~~~F
co
*
oa
a,
.0
to
*r
W
0
U~~~~~~)
0
W
ml
1
0%
§t
,
w
0%
0
0~
as
N
co
es
0
ON
8o
0%.
o
xo
O
_-
t.-
x~
0
a,
.0
T-
s
Ceramics
137
count
25
20
15
10
5
0
4
5
6
Figure 9.5
7
8
10
11
12
9
thickness (mm)
13
14
15
16
Thickness histogram for late period ceramics (n = 196).
sherds were recovered. The color of the slip is red
(10 R 5/6 and 2.5 YR 4/6). This slip occurred on
both the interior and exterior of the sherds and was
found on both rim and body sherds. Three body
herds from the assemblage are decorated. Two
sherds have incised lines, although both are small
making the overall patterns indistinct. The third is a
body sherd with parallel-ribbed, carved paddle
essions on the exterior surface of the vessel.
This sherd was from an early context (unit 20, layer
IB), and compares to other paddle-impressed
sherds known from Westem Samoa (Green and
Davidson 1969:pl. 17).
All sherds recovered during excavations were
examined for residue in the field, (i.e., prior to any
cleaning). Ten sherds with substantial quantities of
carbonized residue were discovered; seven with
residue on the interior and exterior, and three on the
interior alone. This residue has yet to be identified
but is probably the result of cooking starchy foods
(see Hill et al. 1985).
CERAMIC COMPOSITIONAL
MICROANALYSIS
Analysis of macroscopic ceramic traits supports
many research objectives, especially those examined
here. Documenting raw material vanability requires
additional work, especially with respect to temper
and clay of the ceramic fabric. The temper component has been analyzed and discussed by Dicdinson
(chapter 10). Here we present the compositional
microanalysis of thee clay samples and the clay
portion of the ceramics from To'aga. These results
allow us to address questions of clay variability (as a
part of technology) and the potential for ceramic
exchange in Samoa.
A sample of twenty-nine sherds was chosen for
their visual differences in thickness, temper, and
paste in hand section. Also, this variety of sherds
came from excavation contexts that could be inferred
to be of different ages (table 9.1). Age differences
138
The To'aga Site
count
30
25
20
15
10
5
0
4
5
6
Figure 9.6
7
8
9
10
11
12
thickness (inm)
Thickness
histogram
correspond to the analytic divisions made for the
larger sample (i.e., early, middle, and late). The
sherd samples were also analyzed by Dickinson
(chapter 10) for their sand temper petrography.
Examination of the To'aga ceramics by both sand
temper petrography and elemental analysis takes
advantage of the strengths of each approach (Hunt
1988).
In addition to the shends, three clay samples
were collected from colluvium near the base of the
cliff at To'aga on transects 1, 5, and 9. These clays
were fired in a furnace (at 500(C) for fifteen minutes
to produce ceramic tiles resembling sherds. One of
these clay samples (from Transect 9) was prepared
in the laboratory as a fired ceramic tile (sherd). It
was also analyzed by Dickinson (chapter 10) to
compare "self-tempered" sherd petrography.
The elemental microanalysis was accomplished
using an energy-dispersive spectrometer (EDS)
integrated with a scanning electron microscope
for middle
13
period ceramics (n
14
=
15
16
411).
(SEM). SEM/EDS microanalysis is described in this
chapter, and these results are integrated with those
from the petrographic analysis. The distinct advantage of the SEM/EDS is in the selectivity afforded
by the microscope component of the instrunent.
Using the SEM in conjunction with an x-ray analyzer, it is possible to characterize the clay matrix
alone, or individual inclusions, slips, and residues
(e.g., Hunt 1989). Analyses described here were
conducted by one of us (TLH) on a JEOL model
JSM-840A SEM fitted with a Tracor Northem
energy-dispersive x-ray detector housed at the
University of Washington.
Selective elemental microanalysis of pottery is
possible by the coupling of an x-ray analyzer with
the SEM. In the simplest terms, dte SEM provides a
source of electrons of appropriate energy that
impinge on a sample and cause the emission of xrays. The x-rays emitted have energies and relative
abundances that reflect the elemental composition of
139
Cereaics
count
20
15
10
5
0
4
5
6
Figure 9.7
7
8
9
10
11
12
thickness (inm)
13
14
15
16
Thickness histogram for early period ceramics (n = 130).
the sample. The characteristic x-rays are deted by
a lithium-drifted silicon Si(Li) crystal that-together
with electic amplifiers and signal processors-
collects and electrically sorts all the energies from
the x-rays emitted. Under nonnal operating conditions, elements with atomic numbers above 10 (Na =
11), and below 100 (Es = 99) are deteable.
The conversion of x-ray emissions into a
compositional spectrum (figure 9.9) and potential
quantitative data is achieved through a series of
electic components described in some detail by
Postek et al. (1980) and Goldstein et al. (1981:22224). Qualitative and quantitative analysis of the xray spectrum for the composition of a particular
sample is complex, yet well understood (Goldstein et
al. 1981:275-392). As in other recent studies, (e.g.,
Dunnell and Hunt 1990, Graves et al. 1990, Hunt
1989), quantitative analysis of To'aga ceramic clays
used the ZAF conection method (see Goldstein et al.
1981:308). The final values calculated are quantities
of elements (by weight and atomic percents) present
on the cross-section surface of the sample at the
point/area impinged by the electr beam. A
goodness of fit between those elements quantified
and stand intensities is evaluated by a chi-square
test This test provides an objective criterion to
evaluate the goodness of fit for the peak-fitting
algorithms used in each particular application
(Goldsteinet al. 1981:411-12). The analystcan
judge the success or failure of x-ray collection for
goodness of fit for a particular specimen on statistical criteria for each spectrum and quantitative
analysis generated. All these features are available
through the Tracor Northem software used.
Minimum elemental detection limits for energydispersive microanalysis are below 0.1% under ideal
settings, and typically less than 19%, with a relative
precision of 1-5% throughout the elemental range
detected (Hunt 1989). Rice (1987:375) notes the
general concentration range for x-ray analyses as
140
The Tolaga Site
L J~ ~ ~ o
--w.
-
I.If
i
-f-
A~~~~~~1-'O
E
~~~~~~t
_-
o '
F;,
K~~~~~~~~~~~
_
8
I
X~~~~~~~~~~~~~~~I
Ceramics
Figure 9.9 Elemental spectrum of clay composition for
sherd 22 firn To'aga (Univ. of Washington
SEM, 27 Dec. 1989, 1OKV, l200X).
detection of major, minor, and trace elements (>100
ppm)
Specimens were viewed on the CRT of the SEM
while undergoing x-ray analysis. An area that
appeared to be only day was selected at low magnification (30- 1OX) for analysis. This area was
isolated by a step-wise increase in the magnification,
allowing careful inspection of the region for inclusions or other anomalies. X-rays were collected
from areas that appeared at higher magnification to
be clay matrix only. All other specific sample
ation and analytic procedures follow hose
described in detail by Hunt (1989:155-59).
Twelve major and minor elements, Na, Mg, Al,
Si, P, Cl, K, Ca, Ti, Cr, Mn, and Fe, were selected
for analysis. These multivariate data for thirty-two
specimens (sherds and To'aga clays) form the data
matrix used to search for compositional structure in
the Toaga ceramics.
Quantitative Analysis of Clay Elemental Data
The goal of analysis of elemental data is orgaor reducing compositional variability into
archaeologically meaningful groups (Armold et al.
1991; Harbottle 1976:42). Ideally, such groups
represent discrete clay sources, and thus, the minimum number of production locales represented in
the ceramic assemblage. While individual clay
sources are not always distinguishable, the method
has the potential to sort clays from different regions.
nizing
141
Islands offer an especially good setting for sorting
ceramic provenance because they vary in age and
geologic orgin. In addition, clays come from small
drainages and do not mix as in continental deposits.
Sorting the clay elemental data matrix into
meaningful groups generally requires deductive tests
using multivanate statistics (e.g., Bishop and Neff
1989; Davis 1986). Such analyses are usually
directed at infernng a probable number of distinguishable clays represented in an assemblage.
Distinguishable clays might be used to infer production locales for prehistoric pottery. Some studies are
successful at linking prehistoric pottery to specific
(known) clay sources, either on quantitative criteria
that range from ordinal tests to multivariate ones
(e.g., Neff et al. 1988; Topping and MacKenzie
1988). Yet, as Arnold et al. (1991:85) have pointed
out, individual clay deposits may not be easily
distinguished, and in Oceania an island-or even a
group of islands-as a unit of geographic space may
form a single "source" in compositional space.
The To'aga compositional data matrix was
analyzed for its grouping tendency with two different clustering algorithms: average linkage between
groups and Ward's method (Norusis 1986). These
algorithms are agglomeradve and build groups or
clusters from the individual specimens (sherds and
clays).
Given the problems of evaluating cluster
dendrograms, a rigorous solution is in the use of
different algorithms. Then, only those clusters
which arise independently in different analyses are
considered valid, or accurately descriptive
(Alenderfer and Blashfield 1984:65; Dunnell
1983:146; Sokal and Sneath 1963:166). This
strategy will work in data sets in which clay groups
(i.e., not necessarily individual "sources", see Arnold
et al. 1991) are chemically distinctive and can be
detected by statistical measures.
Discriminant function analysis was also used to
examine compositional stnrcture in the To'aga data.
Discriminant function analysis offers a deductive
tool for testing structure in a multivanate data set
(Bishop and Neff 1989; Davis 1986; Hunt 1989). As
Bishop and Neff (1989) imply, ceramic compositional data sets are complexly multivariate and
require going beyond an inductive search for
structure (see also Harbottle 1976).
142
The To'aga Site
Compositional Results
Results of these clustering procedures were
plotted as dendrograms (figures 9.10-1 1). From
these dendrograms, comparable clusters of sherds
(and clay sample tiles) occurring in both solutions
can be deduced. Comparison of the Ward and
average linkage between group cluster dendrograms
reveals four identical clusters. Table 9.15 provides
summary data and results for the sample.
Cluster assignments from the two solutions (1-4,
and two cases unassigned) were analyzed and plotted
against first and second discriminant functions. The
plot (figure 9.12) of discriminant function scores
illustrates the distribution of the groups in multidimensional space.
Cluster 4 includes seven sherds of both thickand thinware as well as the three clay samples
collected from To'aga colluvium. This match
suggests that local colluvial clay from To'aga was
used in some pottery manufacture. The remaining
three (1-3) clusters represent clay compositional
groups as yet unmatched to samples from Ofu, or
elsewhere. These unmatched clays are similar in
composition to those of the local colluvial clay
sample, and may come from other unknown sources/
source areas on Ofu, elsewhere in Manu'a, or
beyond. A determination of local versus exotic
provenance for the unmatched sherds would be
premature; additional sampling of clays is necessary.
Sherds from other islands in Samoa should also be
tested for their compositional similarity to those of
Rescaled Distance Cluster Combine
C A S E
Seq
Label
20
22
05
25
09
21
23
06
15
26
27
08
24
07
14
17
16
13
19
29
18
11
12
30
03
04
02
01
10
32
31
Figure 9.10
20
22
5
25
9
21
23
6
15
26
27
8
24
7
14
17
16
13
19
28
18
11
12
29
3
4
2
1
10
31
30
0
5
10
I
i
I
I
i
15
20
i
Il
I
i
25
I
I
I
I
I
Dendrogram of Ofu pottery and clay samples using the Average Linkage (between groups) method.
Ceramics
143
Rescaled Distance Cluster Combine
C A SE
Label Seq
20
22
05
25
09
24
29
18
13
19
14
17
16
21
23
06
15
26
27
08
07
10
32
31
03
04
02
11
12
30
01
0
5
10
15
i
I
i
iI
I
20
25
I
20
22
5
25
9
24
28
18
13
19
14
17
16
21
23
6
15
26
27
8
7
10
31
30
3
4
2
11
12
29
1
Figure 9.11
Denrgm of Ofu pottery and clay samples using Ward's method.
the To'aga assemblage.
The association of clay compositional groups
with thickware, thinware, red-slipped ware, the
paddle-impressed sherd, and the three colluvial clay
samples from To'aga (table 9.15) shows that thickand thinware cannot be separated compositionally.
All compositional groups are represented in thickand thinware. Discrhminant function analysis, using
ware as the grouping variable, confirmed this
observation. Scatterplots revealed little separation
along the first and second discriminant functions
(figure 9.13). The red-slipped pottery (n = 2) falls
into groups 2 and 3, although a larger sample is
needed to assess compositional variation in this
class. The To'aga colluvial clays (in group 4) match
sherds of thickware, thinware, and the carved
paddle-impressed sherd.
Compansons of ware with temper groups (table
9.15) identified by Dickinson (chapter 10) reveal that
all four temper groups are represented in thickware.
The profuse basaltic temper is found only in
thickware, for this sample. Thinware contains sparse
basaltic (temper group 2), feldspathic (temper group
3), and mixed (temper group 4) tempers (see chapter
10). The two red-slipped sherds in the sample have
sparse basaltic temper. The paddle-impressed sherd
and the analyzed "self-tempered" clay sample (from
Transect 9 colluvium) have the mixed (temper group
4) temper, including calcareous sand. Calcareous
sand in the colluvium suggests saltational transport
of grains from the coast over the previously shorter
distance to the colluvial deposits where mixing could
144
The To'aga Site
Table 9.15
Sherds and Ofu Clays Selected for
SEM/EDS Clay Elemental and Sand
Temper Petrographic Analyses
Specimen
No.
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
31
32
Provenience
Class
Unit6 IIa
Unit 6 IIa
Unit 6 Ila
Unit6 IIa
Unit 6 Ilb
Unit 6 Ilb
Unit 6 lIb
Unit 6 HIb
Unit 6 IIb
Unit 6 Ilc
Unit 6 IIc
Unit 6 IIc
Unit 20 Ilb
Unit 20 HIb
Unit 20 Ilb
Unit 20 IIIa
Unit 20 I11a
Unit 20 I11a
Unit 20 1I1b
Unit 20 IlIb
Unit 20 T11b
Unit 20 IIIa
Unit 20 IIIa
Unit 20 IHIb
Unit 20 Tub
Unit 20 IlIc
Unit 20 IHIc
Unit 20 HlIc
Unit 24 II
Transect 9
Transect 5
Transect 1
Thick
Thick
Thick
Thick
Thick
Thick
Thick
Thick
Thick
Thick
Thick
Thick
Thick
Thick
Thick
Thin
Thick
Thick
Temper
Group
Clay
Cluster
1
1
1
1
1
2
2
4
3?
2
4
4
2
1
3
2
3
3
2
2
2
2
2
4
4
4
2
2
2
4
4
4
4
1
3
4?
4
1
4
4
4
2
2
3
2
2
2
2
1
3
1
3
4?
1
3
3
Thin'
Thin
Thin2
Thick
Thick
Thin3
Thin
Thin
Thick
Thin
Thick4
Clay
Clay
Clay
1 Square rim with an impressed lip
2 Red-slipped
3 Carved paddle-impressed (parallel rib motif)
4 "Thickware" with red-slipped exterior
-
4
-
-
2
4
4
4
Period
Late
Late
Late
Late
Middle
Middle
Middle
Middle
Middle
Early
Early
Early
Middle
Middle
Middle
Early
Early
Early
Early
Early
Early
Early
Early
Early
Early
Early
Early
Early
Early?
---------
Ceramics
x
x
x
4.0
4
1
*
1
13
.0
2
11
3 3
3*2
2
2
44
4
4
*4 4
*
2
2
4
-4.0 +
x
X
x
-6.0 -4.0 -2.0
x
4.0
6.0
First and second discriminant function
scores for analysis based on cluster (1-4)
as grouping variable; * indicates group
center, # indicates unclassified as to
cluster.
Figure 9.12
x
2.0
.0
I
x
I
x
4.0
3
3
3323 3 31
332*3 1 3
*
2 * 333
3
2
.0
1
-4.0
x
Y.
A-
I
-6. 0 -4.0 -2.0
Figure 9.13
X
I
I
.0
2.0
I
4.0
..AV
Profuse basaltic temper (Dickinson's group 1) is
associated with cluster 4 (that includes the three
"self-tempered" colluvial clays) more than expected
by chance alone (expected = 1.7, observed = 4).
Furthermore, temper groups 1, 2, and 4 occur with
clay compositionally indistinct from the To'aga
colluvium samples. All temper groups are associated with clay group 1. It is also noteworthy ta the
overall association is otherwise somewhat well
dispersed. This observation suggests that in many of
the specimens, temper and clay composition vary
independently.
Compositional groups tabulated by their analytic
time penods show a pattern of decline. Compositional variation reflected by temper groups reveals a
similar pattern of decline. The early time period
(1250-500 B.C.) ceramics fall into all fourclay
compositional groups, reflecting the greatest variety
of clay (source) use. One of the clay sources in use
during the early period is the local colluvial clay
from To'aga, and it is in both thick- and thinware.
The middle period (500 B.C.- A.D. 0) may show a
decline to time clay groups, and does not include the
colluvial clay from To'aga. The late (A.D. 0-300?)
ceramics are only represented by the To'aga colluvial
clay in thickware samples. A decline in compositional variability reflects the general simplification
and homogenization of the total To'aga assemblage
with time. This potential trend, however, could
simply result from the smaller samples in the middle
and late time periods. Additional samples must be
analyzed to test a hypothesis of change in the
compositional variability in the To'aga sequence.
CONCLUSIONS
6.0
First and second discriminant function
scores for analysis based on time period
(early [1], middle [2], and late [3]) as
grouping variable; * indicates group
center.
occur natualy. Temper with calcareous sand could
also represent purposefil mixing on the part of
ancient potters.
The association of clay compositional groups
temper groups (table 9.15) defined by
Dickinson reveals some maiked correspondence.
with
145
Ceramics provide a critical source of information for building chronologies and inferring cultural
relatedness because they vary in style. Recent
advances in physico-chemical and archaeological
analyses (Rice 1987) open the door to many new
questions in ceramic studies. In this study, we
focused on ceramics in terms of material composition and provenance, production technology, style,
and function. The To'aga ceramic assemblage is
particularly valuable for this kind of detailed study.
The assemblage is large and comes from a stratified
site where a long chronology of pottery production
can be delineated. In the discussion that follows, we
146
The Tolaga Site
offer some partial answers to the questions posed at
the beginning of this chapter.
Ceramic Provenance and Production Technology
The day microanalytic and sand temper petrographic results provide a basis for several conclusions and new hypotheses conceming ceramic
provenance and production technology. Thickware,
thinware, and a carved paddle-impressed sherd from
To'aga can be inferred to be of local production,
using colluvial "self-tempered" clay source(s) from
Leolo Ridge on Ofu. Such clay(s) could be gathered
near the base of the cliff at To'aga, immediately
adjacent to the prehistoric occupation. Processing of
such clay appears to have been minimal. The
colluvial source(s) accounts for the greatest amount
of pottery in the EDS/petrographic sample.
One class of pottery, the red-slipped (thinware
and one thick, red-slipped sherd), does not match the
local clays as presently known. The red-slipped
pottery in the EDS/petro-graphic sample also
contains only sparse basaltic temper (group 2).
These distinctions may suggest an exotic provenance
for red-slipped ware that arrived on Of through
inter-island exchange. This hypothesis requires a
larger sample to test further.
In sum, the available compositional evidence
suggests at least four hypotheses that may be
confirmed or falsified with additional research:
1) The decline in compositional groups
(both temper and clay groups) merely reflects
sample size differences for the time periods (cf.
Grayson 1984, 1989; Kintigh 1989). This is, in
a sense, the null hypothesis suggesting that
with larger samples, the association of time
period and clay or temper group will become
more even (random in the statistical sense).
2) Local colluvial clay(s) provided a source
for most pottery production at To'aga. Such
clay(s) underwent little, if any, processing by
potters. In most cases the colluvial clays could
be described as "self-tempered."
3) Red-slipped pottery does not conform to
the clay compositional data known for To'aga
(clays and pottery). This ware, and others of
similar clay-temper composition may be exotic
to Ofu, and represent inter-island exchange.
4) The To'aga ceramic sequence is marked
by a decline in the diversity of clays used in
production (which in this case is not a product
of sample size effects). This decline reflects
change in the use or availability of the clay
sources. Such a trend might also denote a
decline in exchange, including that from other
islands (see Hunt 1989; Kirch 1988, 1990).
Reconstruction of production technology is
supported by the macroscopic ceramic analysis
outlined above. Vessel form (bowls) and the
observation of some laminar fracturing in the sherds
point to slab-building as the primary forming
technique. The analysis of orientation angle was
designed to provide evidence for primary forming
(table 9.5). However, due to the absence of grains
with a definitive long-axis, and perhaps the difficulty
of determining "random," "indeterminate," or
"preferred orientation," the vestige of possible slabbuilding is not reflected in attempts to analyze
particle orientation.
Secondary forming is indicated by paddle
impressions (visible on 17% of the sherds) and anvil
marks (present on 23% of the sheeds). Two sherds
show the unmistakable impression of a finger used
for the same purpose.
The majority of the surface treatment is puddling
(77%), a finishing technique using water and wiping
to bring the finest clay particles to the surface of the
paste. Wiping was also evident, occurring commonly on the rim sherds (tables 9.9-10).
Approximately 25% of the sherds in the sample
display a pattern of incomplete oxidation adjacent to
the interior surface in contrast to 4% with incomplete
oxidation at the exterior surface. Only 2% of the
sherds show little or no oxidation present About
67% of the sherds are completely oxidized. Based
on this, and other evidence described (hardness and
comparison of experimentally fired-clay tiles), we
suggest that pottery was fired in open conditions of
temperatures reaching approximately 500-600( C.
The fact that interior surfaces were darker (less
oxidized), suggests that bowls were placed up-sidedown for firing. This technique is similar to some
documented edtnographically in Melanesia, where
pottery is still made in many locations (e.g., Irwin
1985; May and Tuckson 1982).
Explaining the abandonment of pottery production in Samoa remains unresolved. Our To'aga
analyses show that diversity of material use may
Ceramics
have declined over the period of ceramic production.
It could be hypothesized that changes in raw materials, for example the use of "self-tempered" colluvial
clays, resulted in a ceramic product of marginal
quality. This hypothesis requires additional study
(e.g. see Feathers 1990).
Styk
The To'aga assemblage is simple in form and
carries very little decoration. Vessel parts present
(direct rims and body sherds only) indicate tat only
forms of unr cted orifice (bowls) were produced.
There is no evidence in the To'aga assemblage of
globular pots, jars, plates, or other complex vessel
forms. Decorative attributes are restricted to impressing and notching on the lip, red-slip, carved
paddle impression, and incision. Such a short roster
departs dramatically from assemblages of comparable age from Mulifanua, 'Upolu, and from assemblages in Tonga and Fiji.
Style can be defined for analytic purposes in
archaeology as traits ta are free to vary independent of function (Dunnell 1978). This definition
emphasizes style as governed by stochastic processes, and distributional frequencies that behave
accordingly. In this perspective, thickness might be
treated as a "stylistic" tait. Sherd thickness has
received much attention in previous attempts to
understand diachronic change in Samoan ceramics
(Claik and Herdrich 1988; Green 1974; Hunt and
Kirch 1988; Jennings and Holmer 1980, Kirch et al.
1990). The changing (declining) frequencies of
thinware, in particular, might be a reflection of
homology ("style"), and its independence from
functional constraints. The frequency distributions
(see tables 9.1-2) of thinware (defined as <7.5 mm)
and thickware (>75 mm) from To'aga allow the
following conclusions:
1) Thickware is present in the earliest deposits, and
its abundance over time is relatively stable.
2) Thinware is never dominant in the assemblage
but occurs in roughly equal percentages to
thickware in the earliest deposits.
3) The presence of thinware declines in real and
relative values over time but persists perhaps as
long as pottery production itself.
4) Pottery declines in abundance early in the
ChIstian era and then its production disappears
147
entirely.
The evidence firm Western Samoa is similar in
many, but not all, rspects. In spite of early dates for
To'aga (i.e., contemporaneous with the Mulifanua
Lapita site), no dentate-stamped Lapita pottery is
known for this site, or elsewhere in Manu'a. This
absence may be paralleled in the cases from Tikopia
(Kirch and Yen 1982) and Anuta (Kirch and
Rosendahl 1973) where assemblages of pottery date
to early times, yet do not share the degree of decoration known elsewhere in the southwestern Pacific.
Perhaps this reflects isolation from a larger
intensland network that shared ideas of designs, or
pots themselves. Manu'a may have simply been far
enough away to incur such isolation from other
islands of Samoa, Tonga, and Fiji.
Green (1974) proposed a sequence of ceramic
change for Samoa. His chronological analysis from
Upolu was based on a short occupation sequence,
with radiocarbon dates ranging from 1840 ± 100 B.P.
(GaK-1441) in the lowest cultural layer (V) to 1800
± 80 B.P. (GaK-1341) in the layer (IV) above (Green
1974:115). These dates overlap at one standard
deviation. When calibrated and averaged together,
these two dates yield a calibrated age range at one
standard deviation of A.D. 1 17-254 (Stuiver and
Reimer 1986). Based on his analysis of over 7400
sherds, Green (1974:130) concluded that, "thin and
thick ware sherds occur in association in both layers"
and that over time the end is for thickware to
predominate but not totally replace thinware
(1974:248). Inspection of Holmer's (1980:116)
data, and his comparison to other Western Samoan
assemblages, reveals a similar trend for thin- and
thickware frequencies.
Function
Vessel function in the To'aga assemblage is
suggested by fonn and the presence of residues. rhe
single vessel form (bowls) might have served
functions of storage, cooking, and serving dishes.
Carbonaceous residues suggest cooking, at least in a
small number of the vessels. Microanalysis for
concentrations of phosphorus (P) was performed on
three sherds in an experimental effort (Dunnell and
Hunt 1990). These and other test case results were
varied, and revealed that functional inferences based
on P concentrations in pottery are unreliable
148
The Tolaga Site
(Dunnell and Hunt 1990).
The ceramics from To'aga are among the best
studied in the Samoan Ilands. Addressing difficult
issues beyond questions of chronology and cultural
affinities demand detailed studies as we have
attempted here. With regard to the questions posed
at the start of this chapter, our study contributes to
answers that will necessarily come from several
studies ofcomparable detail and scope conducted
with assemblages from throughout Samoa and the
larger region.
REFERENCES CITED
Aldenderfer, M. S., and R. K. Blashfield 1984.
Cluster Analysis. Beverly Hills: Sage Publications.
Arnold, D. E., H. Neff, and R. L. Bishop 1991.
Compositional analysis and "sources" of
pottery: An ehnoarchaeological approach.
American Anthropologist 93:70-90.
Barry, B. A. 1978. Errors in Practical Measurement
in Science, Engineering, and Technology. New
York: John Wiley & Sons.
Bishop, R. L., and H. Neff 1989. Compositional data
analysis in archaeology. Archaeological
Chemistry IV:57-86.
Bishop, R L., V. Canouts, S. P. DeAtley, A.
Qoyawayma, and C. W. Aikins 1988. The
formation of ceramic analytic groups: Hopi
pottery production and exchange, A.C. 13001600. Journal ofField Archaeology 15:317-37.
Clark, J., and D. Herdrich 1988. The Eastern Tutuila
archaeological project 1986, final report
Prepared for the government of American
Samoa, Office of Historic Preservation, Dept
of Parks and Recreation. Pago Pago, American
Samoa.
Davis, J. C. 1986. Statistics and Data Analysis in
Geology. Second Edition, New York: John
Wiley & Sons.
Dunnell, R C. 1978. Style and function: A fundamental dichotomy. American Antiquity 43:192-
202.
1983. Aspects of the spatial structure of the
Mayo Site (15-JO- 14), Johnson County,
Kentucky. IN R. C. Dunnell and D. K.
Grayson, eds., Lulu Linear Punctated: Essays
in Honor of George Irving Quimby, pp. 109-
165. Anthropological Papers No. 72, Museum
of Anthropology, University of Michigan, Ann
Arbor.
Dunnell, R C., and T. L. Hunt 1990. Elemental
composition and inference of ceramic vessel
function. CurrentAnthropology 31(3):330-36.
Feathers, J. K 1990. Explaining the evolution of
prehistoric ceramics in southeastern Missouri.
Ph.D. dissertation, University of Washington,
Anm Arbor: University Microfilms.
Goldstein, J. I., D. E. Newbury, P. Echlin, D. C. Joy,
C. Fiori, and E. Lifshin 1981. Scanning Electron Microscopy and X-ray Microanalysis.
New York: Plenum Press.
Graves, M., T. L. Hunt, and D. Moore 1990. Ceramic production in Mariana Islands: Explaining change and diversity in prehistoric interaction and exchange. Asian Perspectives 29:21133.
Grayson, D. K 1984. Quantitative Zooarchaeology:
Topics in the Analysis of Archaeological
Faunas. New York: Academic Press.
1989. Sample size and relative abundance in
archaeological analysis: Illustrations from spiral
fractures and seriation. IN R. D. Leonard and
G. T. Jones, eds., pp.79-84,Quandfying Diversity in Archaeology. Cambridge: Cambridge
University Press.
Green, R. C. 1974. Excavation of the prehistoric
occupation of SU-SA-3. IN R. C. Green and J.
Davidson, eds., Archaeology in Western
Samoa, Vol. II, pp. 108-154. Bulletin of the
Auckland Institute and Museum No. 7.
Gren, R. C., and J. Davidson, eds., 1969 Archaeology in Western Samoa, Vol. I. Bulleting of the
Auckland Institute and Museum No. 6.
Harbottle, G. 1976. Activation analysis in archaeology. Radiochemistry 3:33-72.
Hill, H. E., J. Evans, and M. Card 1985. Organic
residues on 3000-year-old potsherds from
Natunuku, Fiji. New Zealand Journal of
Archaeology 7:125-28.
Holmer, R 1980. Samoan ceramic analysis. IN J.
Jennings and R. Holmer, eds., Archaeological
Excavations in Western Samoa, pp. 104-116.
Pacific Anthrpological Records 32. Bernice P.
Bishop Museum, Honolulu.
Hunt, T. L. 1988. Lapita ceramic technological and
compositional studies: A critical review. IN P.
Ceramics
V. Kirch and T. L. Hunt, eds., Archaeology of
the Lapita Cultural Complex: A Critical
Review. pp. 48-59. Thomas Burke Memorial
Washington State Museum Research Report
No. 5. Seattle.
. 1989. Lapita Ceramic Exchange in the Mussau
Islands, Papua New Guinea. Ph.D. dissertation,
University of Washington, Ann Arbor: University Microfilms.
Hunt, T. L., and P. V. Kirch 1988. An archaeological
survey of the Manu'a Islands, American Samoa.
Journal of the Polynesian Society 97(2):153-83.
Irwin, G. J. 1985. The Emergence ofMailu. Terra
Austalis, No. 10, Department of Prehistory,
Research School of Pacific Studies. Canberra:
Australian National University.
Jennings, J., and R. Holmer 1980. Archaeological
Excavations in Western Samoa. Pacific Andtpological Records 32. Bernice P. Bishop
Museum, Honolulu.
Kintigh, K. W. 1989. Sample size, significance, and
measures of diversity. IN R D. Leonard and G.
T. Jones, eds., Quantifying Diversity in Archaeology, pp. 25-36. Cambridge: Cambridge
University Press.
Kirch, P. V. 1988. Long-distance exchange and
island colonization. Norwegian Archaeological
Review 21:103-117.
. 1990. Specialization and exchange in the Lapita
Complex of Oceania (1600-500 B.C.). Asian
Perspectives 29:117-33.
Kirch, P. V., and P. Rosendahl 1973. Archaeological
investigation of Anuta. IN D. E. Yen and J.
Gordon, eds., Anuta: A Polynesian Outlier in
the Solonon Islands, pp. 25-108. Pacific
Anthropological Records No. 21 Honolulu:
Bernice P. Bishop Museum
149
Kirch, P. V., and D. E. Yen 1982. Tikopia: The
Prehistory and Ecology of a Polynesian
Outlier. Bishop Museum Bulletin 238, Honolulu.
Kirch, P. V., T. L. Hunt, L. Nagaoka, and J. Tyler
1990. An ancestral Polynesian occupation site
at To'aga, OfN Island, Ameuican Samoa.
Archaeology in Oceania 25(1):1-15.
May, P., and M. Tuckson 1982. The Traditional
Pottery ofPapua New Guinea. Sydney: Bay
Books.
Neff, H., Bishop, R. L., and Arnold, D. E. 1988.
Reconstructing ceramic production from
ceramic compositional data: A Guatemalan
example. Journal of Field Archaeology 15:33948.
Norusis, M. J. 1986. SPSS/PC+for the IBM PC/XT/
AT. Chicago: SPSS Inc.
Postek, M. T., K. S. Howard, A. H. Johnson, and K
L. McMichael 1980. Scanning Electron
Microscopy: A Student's Handbook. Ladd
Research Industries, Inc.
Rice, P. M. 1987. Pottery Analysis: A Sourcebook.
Chicago: University of Chicago Press.
Rye, 0. S. 1981. Pottery Technology: Principles and
Reconstruction. Washington, D.C.: Taraxacum.
Sokal, R. R., and H. A. Sneath 1963. Principles of
Numerical Taxonomy. San Francisco: W. H.
Freeman.
Stuiver, M., and P. Reimer 1986. A computer
program for radiocarbon age calibration.
Radiocarbon 28:1022-1030.
Topping, P. G., and A. B. MacKenzie 1988. A test of
the use of neutron activation analysis for clay
source characterization. Archaeometry 30:92101.
10
SAND TEMPER IN PREHISTORIC POTSHERDS
FROM THE TO'AGA SITE
W. R. D cKINSON
TWENTY-NI
RETATIVE sherds selected by
T. L. Hunt from the collection of artifacts
excavated at the To'aga site near the south coast of
Ofu Island in the Manu'a Group of American Samoa
were examined pergraically in thin section. As
there is no current rason to suspect that any of the
To'aga ceramic ware was made elsewhere, the
purpose of the study was to provide baseline information about Manu'a temper sands. All the sherds
examined contain volcanic sand as temper, aithough
this basaltic detritus is mixed with calcareous grains
derived from reef sources in some of the sherds. As
would be predicted for Manu'a and other parts of
Samoa, the volcanic sand is typical of the oceanic
basalt tempers common to intra-oceanic Pacific
archipelagoes (Dickinson and Shutler 1968, 1971,
1979). Several variants of temper sand are present in
different sets of sherds, and available information is
inadequate to pinpoint their
ive sources. All
could p ly have been collected on Ofu or nearby
Olosega Island, but derivation of some from Ta'u in
Manu'a or even elsewhere in Samoa is not precluded
by the petrgric data. Their petrologic compatibility with Samoan lavas, however, and their overall
resemblance to the spectrum of basaltic tempera
studied to date from Samoa, makes importation from
outside Samoa quite unlikely. As none of the temper
variants are identical to tempers known from Tutuila
or Upolu, all are regarded provisionally as indigenous To'aga temper, with the proviso that petrographic evidence alone cannot indicate how far
afield ancient potters may have gone in their seach
for suitable clay and temper within Manu'a.
TO'AGA TEMPER VARIANTS
The following variants ofbasaltic temper sand
are all pr in varying numbers of To'aga sherds,
and each is described in detail in subsequent passages:
(a) Profise Basaltic Temper: Seven sherds (1-5,
7, 14) contain ferromagnesian basaltic sand so
abundant that it forms 50-60 percent ofthe sherd
bodies. The proportions of grain ypes in six (1-5, 7)
of the sherds are statistically indiuis e, but
the seventh (14) contains a related volcanic sand of
slightly different composition.
(b) Sparse Basaltic Temper: Twelve sherds (6,
10, 13, 16, 19-23, 27-29) contain sparse feremnagnesian basaltic sand of somewhat different composition
and texture. The mineral and lithic grains form only
5-15 percent of the sherd bodies, but about a third
(25-40 percent) of the temper grains used may have
been fragments ofbroken pottery.
(c) Feldpathic Basaltic Temper: Three sherds
(15, 17, 18) contain feldspathic basaltic sand, which
152
The To'aga Site
fonns a normal prport (30-40 percen) of the
sherd bodies and is both mineralogically and texturt the two ferroma
ally disinc from
an variants of
Toaga tper. Another sherd (9) contins a similar
but sparser temper sa (A15
i percent of body) that
appears to be a hybrid sand with a significant
e of detritus from the kinds of bedrock
sources that yielded the feromagesia basaltic
sands.
(d) Mixed Temper Sand: Seven sherds (8, 11,
12, 24-26, 30) contain mixed teper sands composed of both basaltic and reefdetitus in varying
poportions. Fcrwmagnesian mineral grai and
basaltic litc fragmens do the volcanic sand
components of the mixed tempers, but their proportions are highly variable and plagioclase feldspar
grains are al pesent in some sherds.
temmed "slabwor') display blocky to prismatic
subhedral pyroxene crystals iergrown with aggregated and multiply twined plagioclase crystals of
slabmike aspect. Lithic fragments of intermediate
grain size (here termed "latiwotk) display disorented mosaics oftwinned plagioclase laths with
equant and largely anhedral pyroxene grains within
their interstices. As all gradations are seemingly
present between "slabworkW and 'lathwodc" and
'tachylitic" grains, the distinction made among hem
is useful in a qualitatve sense only. Groundmass
iron oxides in lithic fragmes range from equant or
granularto skelet or elongate in form without
apparent regard to other aspes of internal texture.
TEMPER GRAIN TYPES
The fenomagnesian volcanic sands in sherds
with profuse basaltic temper are moderately sorted
assemblages of subangular to subounded grains
with a texture suggestive of alluvial orgin. Unmistable rounding of the edges of many gains
indicates n rally occuming sand, rather than
artificially cnmshed aggregate, and local ravine
streams may have provided the sources of the
temper. As would be expected for such a setting,
lithic fragmens are generally but not uniformly
larger than mineral grains. Abundance of
subangular silty basalt detritus within the clayey
paste in which the temper sand is imbedded suggests
that potters collected naturally tempered sandy clay.
This circumstance may account for the superabundance of temper sand in proportions higher than
typically encountered in Pacific Island sherds.
Proportions of grain types in most sherds containing
this alluvial temper are quite consistent (table I0.1):
half pyroxene, a quarter lithic fragments, a fifth
olivine, and a trace of opaque iron oxides. All
micrhenocrysts in lithic fragments are pyroxene
and olivine. As the inteal texture of lithic fragments is somewhat variable, being ree-quarters
'lathwok" in two sherds (1, 7) and two-thirds
"slabwok" in four others (2, 5), collecting sites were
evidently closely related but not identical. Nevertheless, the average temper composition (table 10.1) for
the six sherds in which proportions of grain types are
essentially the same is taken here to be the best
estimate, petrlogically speaking, of proportions of
consituents for characteristic To'aga temper.
The non-calcareous components of all the
To'aga temper types are without exception composed
of mineral grains and lithic fragments derived
entirely from basaltic bedrock sources, eiter lavas
or pyroclastic deposits, together mi some cases with
fragments of broken pottery. The mineral grains,
originally phenocrysts or micrphenocrysts in basalt,
include clinopyroxene, olivine, opaque iron oxides
(magnetite anid/or ilmenite), and plagioclase feldspar.
The lithic fr s, represing aphanitic groundmass of basaltic lava or tephra, display a spectrum of
internal textures reflecting an inherent range of
constituent grain sizes. Microphanocrysts in lithic
fragments include all the mineral species that were
also p t in the temper sands as separate mineral
gams
Routine distinction between pyroxene and
olivine in thin section was based upon key diagnostic
features visible for each grin, and their identifications were checked by observations ofoptic axial
angle and birefringence on suitably oriented grains.
Pyroxene grains generally display faint green tints,
and many show either cleavage or prismatic shapes.
Untinted olivine grains are brighter in plane light,
and many are altered along edges and fIracues to
bright reddish iddingsite. Most basalt lithic fragments have an intergranuar internal texture, although the finest gained (here tenmed "tachylitic")
ae intersertal with plagioclase microlites set in black
basaltic glass (tachylite). The coarsest grained (here
PROFUSE BASALTIC TEMPER
Sand Temper in Prehistoric Potsherds
153
Table 10.1
Frequency Percentages of Ferromagnesian Mineral
Grains' and Basaltic Volcanic Lithic Fragments (VRF)
in Sherds Containing "Profuse Basaltic Temper"
Sherd
1
2
3
4
5
7
Ave
14
n
Py
01
Fe
VRF
Py(Py+01)
130
160
210
105
185
260
51
50
53
51
50
54
52
36
21
23
20
17
24
23
21
14
2
2
1
4
4
1
2
3
26
25
26
28
22
22
25
47
0.71
0.68
0.73
0.75
0.78
0.71
0.71
0.72
'Py, clinopyroxene; 01, olivine; Fe, opaque iron oxides
Note: n=number of grains counted in each sherd and average (Ave) composition is calculated for sherds
1-5 plus 7 but not 14.
A seventh sherd (14) contains distinctly more
lithic volcanic sand (table 10.1), although its pyroxene/olivine ratio is very close similar to that of the
other sherds. Lithic fragments, mostly "lathwork,"
are also more irregular in shape and some are
microvesicular. Curved re-entrants on some lithic
fragments and the presence of a few grains of
microvesicular brownish basaltic glass suggest a
pyroclastic component lacking in the more characteristic six sherds whose tempers were probably
derived entirely from bedrock lava sources.
SPARSE BASALTIC TEMPER
The ferromagnesian volcanic sands in sherds
with sparse basaltic temper are well sorted aggregates of subrounded grains with a texture suggestive
of beach origin. The lack of finer grained grit within
the clayey paste suggests that potters added artificial
temper to clay bodies. Dark angular blotches within
the clayey paste are probably ghosts of broken
pottery fragments also added as pain of the tempering
process. Although recognition of this grog constituent is equivocal in some sherds owing to indistinct
outlines of the pottery fragments, its presence may
account for the low overall proportion of volcanic
sand, which amounts alone to much sparser temper
than typically encountered in Pacific Island sherds.
Frequency percentages of grain types are highly
variable for different sherds (table 10.2), but so few
grains are present in each sherd (average of only 20
per sherd in 10 of the sherds) that the statistical
significance of individual counts is questionable.
Consequently, all grains (250 total) were summed
from all sherds counted to yield net frequency
percentages, but net and average temper compositions are almost identical (table 10.2). The fact that
both measures of bulk composition are similar to
values for the single sherd (22) containing the most
grains (60) gives confidence that either measure is a
valid estimate of the overall temper composition.
The sparse basaltic temper, probably beach sand, is
less pyroxenic and more lithic than the average
composition of the dominant alluvial variant of the
profuse basaltic temper, but grain proportions closely
resemble those in the more lithic variant of alluvial
sand. In general, differences are not great enough to
suggest wholly different provenance except for the
contrast between stream and beach collecting sites.
Proportons of lithic grain types are quite variable
from sherd to sherd, but all three types are present in
subequal amounts within the suite of twelve sherds
as a whole. The net pyroxene/olivine ratio (0.63) is
only slightly lower than in the alluvial sands (0.72),
and may have been reduced marginally by preferential cleaving of pyroxene grains and winnowing of
resulting cleavage fragments during prolonged
reworking in a beach environment.
154
The To'aga Site
Table 10.2
Frequency Percentages of Ferromagnesian Mineral Grains
and Basaltic Volcanic Lithic Fragments (VRF) in
Sherds Containing "Sparse Basaltic Temper"
Sherd
6
10
16
19
20
21
22
23
27
28
29
Net
Ave
n
Py
01
Fe
VRF
20
10
15
30
10
10
60
25
30
20
20
250'
5
40
47
33
30
40
42
32
40
25
25
34
33
10
20
33
13
20
20
20
20
40
5
15
20
20
5
10
7
4
10
10
2
8
3
5
5
4
6
80
20
13
50
40
30
36
40
17
55
55
42
41
'Summation of n for 11 sherds listed.
Note: n=number of grains counted in each sherd (note that net and average compositions
are essentially the same). Sherd 13 too weatherd to allow accurate count.
FELDSPATHIC BASALTIC TEMPER
The volcanic sand in sherds (15, 17, 18) containing feldspathic basaltic temper essentially lacks
fenrmagnesian mineral grains (table 10.3), and
nearly all microphenocrysts in lithic fragments are
plagioclase rather than pyroxene or olivine. One
olivine grain is present, however, in one sherd (18),
and one olivine microphenocryst is present in
another (15). Lithic fragments are consistently
larger than separate plagioclase mineral grains,
although the two are present in about the same
frequency (table 10.3), and the sand overall is only
moderately sorted. Most lithic fragments are
"tachylitic," many have smoothly curved re-entrants
typical of tephra clasts, and some are microvesicular.
The textural features of the sand jointly suggest that
scoiaceous basaltic as, possibly reworked locally,
was added as artificial temper to the clay body by
potters having some selective aim in using such a
tempering material The feldspathic basaltic temper
shows no compositional overlap with the fenomagnesian basaltic tempers, but the geograhic separa-
tion of their respective sources need not have been
great. As if to underscore that point, one sherd (9)
contains well-sorted and subrounded temper,
probably a beach sand, that apparently represents a
mixture of feldspathic and ferromagnesian volcanic
sands. This anomalous sherd has an apparently
sparse temper (- .15% of body) but also includes a
few fragments ofbroken pottery as part of its overall
temper component.
MIXED TEMPER SAND
The volcanic sands in sherds containing an
admixture of reef-derived calcareous grains (15-75
percent) are highly variable in mineralogical composition (table 10.4). Although all are dominantly
ferromagnesian volcanic sands, nearly half contain
feldspathic components as well. Coupled with the
presence of the calcareous grains, the good sorting
and rounding of the sands is diagnostic of coastal
origin on beaches where mixing of detritus from
multiple sources is to be expected. Proportions of
temper sand vary from 10-25 percent (8, 11, 12, 24)
155
Sand Temper in Prehistoric Potsherds
Table 10.3
Frequency Percentages of Plagioclase Feldspar (P1) and Opaque Iron Oxide
(Fe) Mineral Grains and Basaltic Volcanic Lithics Fragments (VRF) in
Sherds Containing "Feldspathic Basaltic Temper"
Sherd
15
17
18
Ave
9
n
P1
Fe
VRF
60
35
125
43
50
56
50
18
8
6
4
6
7
49
44
40
44
59
60
Py
01
-
-
-
-
-
-
-
-
13
3
Note: n=number of grains counted in each sherd and average (Ave) composition is calculated for sherds 15
plus 17-18 but not 9 (Py and 01 are clinopyroxene and olivine mineral grains in sherd 9).
Table 10.4
Frequency Percentages of Calcareous Grains (calc), Silicate-Oxide
Mineral Grains, and Lithic Fragments for Sherds Containing "Mixed
Temper Sand"
Sherd
8
11
12
24
25
26
30
Calc
n
Py
01
Fe
P1
VRF
13
22
20
15
44
18
77
65
35
12
30
70
160
60
32
14
33
27
21
22
39
32
7
17
23
17
18
14
3
7
25
-
-
1
1
2
1
5
20
33
72
25
50
60
54
25
Note: n=number of non-calcareous grains counted in each sherd (percentages reported sum to 100
exclusive of calcareous grains)
to 40-60 percent (25, 26, 30), the "tachylitic" variety
of lithic fragments form about half the lithic popula-
DISCUSSION: TEMPER COMPARISONS
mon, and the overall pyroxene/olivine ratio ('0.60) is
similar to that in the other sherds thought to contain
beach sand temper. Computation of an average or
net composition for the sherds containing mixed
sand temper would be meaningless, given their
inherent compositional variability, but their volcanic
sands fit broadly within the spectrum of temper types
present in other sherds. Sherds containing only
sparse mixed temper sand also contain fragments of
broken pottery in uncertain amounts.
Although each of the To'aga temper types has
clear distinguishing characteristics, compositional
links argue that they form a related temper suite that
is presumably indigenous to Manu'a. The two
ferromagnesian basaltic tempers have contrasting
textures tat reflect different sedimentological
origins such as stream and beach sands, but the same
grain types are present in both in slightly different
proportons. The admixture of similar fenomagnesian constituents in one ofthe sherds containing
156
The To'aga Site
feldspathic basaltic temper suggests hat bedrock
sources for the fenomagnesian and feldspathic
volcanic sands exist not far apart This inference is
strengtned by the observation that mixed beach
sands containing calcareous grains contain a varied
spectrum of fenomagnesian and feldspathic constituents.
Prehistoric sherds examined previously from
Upolu (Dickinson 1969,1974,1976) contain
generally similar basaltic temper sands composed of
the same basic grain types, but none of the 'Upolu
tempers is identical in texture or composition to the
To'aga tempers. Fenromagnesian basaltic tempers
from 'Upolu commonly contain a higher proportion
oflithic fragments, typically have a higher ratio of
olivine to pyroxene, and generally contain a subordinate proportion of brownish basaltic glass particles
not present in To'aga tempers. Moreover, Upolu
ferromagnesian tempers
sherds with a
contain well-sorted coastal sands texturally unlike
the stream sands evident in well tempered To'aga
sherds. Feldspathic basaltic tempers from 'Upolu are
broadly similar to their To'aga analogs, but internal
textures of lithic fragments differ in being coarser
gained in the lJpolu sherds studied to date. Feldspathic trachytic tempers present in all available
Upolu collections apparently have no counterparts at
To'aga.
The generic resemblance of all Samoan temper
types examined to date pennits the strong inference
that the tempers in To'aga sherds are indigenous to
Samoa. On balance, there is no reason to suppose
on petrographic evidence that any of the To'aga
temper types were derived from sites elsewhere in
Samoa. The fragments ofbroken pottery present in
about half the To'aga sherds are not common constituents ofPacific Island wares, but do occur in sherds
from the Ryukyu Islands, Palau, and the Nan Madol
site on Ponape (Dickinson and Shutler 1979; Dickinson et al. 1990). Their presence at To'aga presumably reflects a common paucity of suitable local
temper sand, rather than any close cultural relationship between the Caroline Islands region and Samoa.
REFERENCES CITED
Dickinson, W. R 1969. Temper sands in prehistoric
potsherds from Vailele and Falefa. IN R. C.
Green and J. M. Davidson, eds., Archaeology
in Western Samo, Vol. I: pp. 271-73,
Auckland Institute Museum Bulletin 6.
. 1974. Temper sands in shends from Mulifanua
and comparison with similar tempers at Vailele
and Sasoa'a (Falefa). IN R. C. Green and J. M.
Davidson, eds., Archaeology in Western
Samoa, Vol. II: pp. 179-80, Auckland Institute
Museum Bulletin 7.
1976. Mineralogy and petrology of sand
tempers in sherds from the Feriy Beth site,
Paradise site, and Jane's Camp. IN J. D.
Jennings, R N. Holmer, J. C. Janetski, and H.
L. Smith, eds., Excavatios on 'Upolu, Western
Samo, pp. 99-103. Pacific Anthrpological
Records 25. Honolulu: Bernice P. Bishop
Museum.
Dickinson, W. R, and R. Shutler, Jr. 1968. Insular
sand tempers of prehistoric pottery from the
southwest Pacific. IN I. Yawata and Y. H.
Sinoto, eds., Prehistoric Culture in Oceania,
pp. 29-37. Honolulu: Bishop Museum Press.
1971. Temper sands in prehistoric pottery of the
Pacific islands. Archaeology and Physical
Anthropology in Oceania 6:191-203.
-.1979. Petrography of sand tempers in Pacific
islands potsherds. Geological Society of
America Bulletin 90 (Part I, summary):993-95;
and Part II (microfiche), No. 1 1, Card 1, pp.
1644-1701.
Dickinson, W. R. J. Takayamna, E. A. Snow, and R.
Shuder, Jr. 1990. Sand temper of probable
Fijian origin in prehistoric potsherds from
Tuvalu. Antiquity 64:307-312.
11
NON-CERAMIC PORTABLE
ARTIFACTS FROM TO'AGA
PATRiCK V. KIRCH
A soOM ThE SBsrANTiALquatities of pottery
above in chapter 9, the To'aga site
excavations yielded a small but typologically diverse
assemblage of non-ceramic portable artifacts.
Because the site's alkaine, calcareous depositional
environent (particularly in the lowerlevels) favors
the preservation of bone, shell, and sea urchin, a
variety of artifacts made from these organic materials was recovered, in addition to objects of basalt
and coral. This contasts with most early Samoan
archaeological sites, such as Vailele or Sasoa'a on
Upolu (Green and Davidson 1969, 1974), in which
the acidic soils did not preserve a wide range of
materials. Prior to our work at To'aga, only the
Potsa and Falemoa sites on Manono Islet (Janetski
1980) had yielded a significant array of artifacts of
shell, bone, and sea-urchin spine in association with
Samoan ceramics. Thus, our knowledge of early
Samoan material culture was largely restricted to
basalt adzes, non-retouched lithics, and ceramics
(Green 1974). This was in contrast to the situation
with sites of comparable age in Tonga, where
excavations on Tongatapu (Poulsen 1987),
Niuatoputapu (Kirch 1988), and Ha'apai (Dye 1987)
had produced a diverse array of material culture
dating to the Ancestral Polynesian period. Hence,
the To'aga artifact assemblage, described in full
below, significantly expands our knowledge of the
Samoan variants of Ancestral Polynesian material
culture in the first millennium B.C.
The non-ceramic artifacts from To'aga are
described below according to broad functional
classes in general use by Polynesian archaeologists.
Comparisons are also made between the To'aga
assemblage and other assemblages from Ancestral
Polynesian period sites in Samoa and elsewhere in
Westem Polynesia.
STONE ADZES
Six adzes which were either whole or sufficiently intact to be classified were excavated,
primarily from pottery-bearing contexts. These
adzes are classified according to the system devised
by Green and Davidson (1969) for adzes from
Westem Samoa. In addition, we recovered five
small flakes with ground or polished surfaces, which
appear to have been derived from adzes during use
or bevel resharpening. Most of these diagnostic
specimens were petro-chemically analyzed by M.
Weisler in order to determine the range in quarry
sources utilized. Weisler used the nondestructive
XRF technique and presents the results of his study
in chapter 12.
From Layer HA-I in Unit 9 we excavated a
finely ground and polished, complete adz of Samoan
Type V (Green and Davidson 1969:24-26). The adz
is of a very fine-grained, light grey basalt or
andesite, and most of the original flaking scars have
been removed by extensive polishing (fig. 11.1, c).
158
The Tolaga Site
The bevel is curved, and the poll shows distinct
battering, indicating use as a hammer while hafted.
ile adz is 136.8 mm long, 54 mm wide, and 35.2
mm thick at the midpoint It weighs 422 g. Type V
adzes are commonly associated with plainware
ceramics both in Samoa (Green 1974) and in other
early Western Polynesian contexts (Kirch 1988:192,
203).
An incomplete section of another Type V adz,
consisting of the bevel to the midsection, was found
in Layer IIIB of Unit 20 (fig. 11.1, a). The adz is of
a light-grey, fine-grained basalt The bevel is curved
and very highly polished, while other paris exhibit
remnant flake scars. The plano-convex section is
rather high. The incomplete length of the adz is 78.4
mm; the width, 38.2 mm; and the thickness at
midsection. 29.1 mm.
Another partial adz ofType V, consisting of the
butt to midsection, was found in Layer IIIB of Unit
23 (fig. I 1.1, b). Made of greyish basalt, the adz has
a low (flattened) plano-convex cross section. In plan
view, it also distinctly narrows toward the butt. The
front and sides are partially ground and polished, but
some flake scars remain. The midsection break
displays considerable battering, indicating that the
specimen was used as a hammerstone after breaking.
The incomplete length is 68.1 mm; its width at the
butt, 29.0 mm; the width at midsection, 53.3 mm;
and the thickness at midsection, 27.3 mm.
A small adz of fine-grained basalt was recovered
from the disturbed landfill site at To'aga during the
1986 reconnaissance. The adz has a sub-triangular
cross section, and thus would be classified as Type
VI in the Green and Davidson (1969) system.
However, it has been well ground on the front,
removing the original flaked ridge (and thus rounding off the apex of the triangle). Hence, in some
respects, the adz resembles a Type V.
A rather battered remnant section of an adz,
possibly of Type V or another type with a subquadrangular section, was excavated from Layer
iIIA of Unit 27. This specimen is of dark grey basalt
and has polished front and back surfaces. The butt is
largely intact, but the artifact has been heavily
battered from use as a hammerstone. The thickness
is 25.4 mm, and the incomplete length, 75.4 mm.
From Unit 3, in an aceramic depositional
context we recovered the midsection of a partially
ground, fine-grained basalt or andesite adz with
trapezoidal cross section, probably of Samoan type
IV (Green and Davidson 1969:24). The midsection
is 24.2 mm thick, with the width ranging from 30.6
to 50.5 mm. Petrohemical analysis by non-destructive XRF, described further in chapter 12, suggests
that this adz was manufactured at the large
Tatagamatau quarry site on Tutuila Island (Best et aL
1989, 1992; Leach and Witter 1987, 1990). This is
noteworthy, since most of the Manu'a adzes assignable to the Tatagamatau quarry were surface finds,
also of trpezoidal sectioned types typical oflater
Samoan prehistory. This adz from Unit 3 is associated with a 14C date of 1389-1287 cal B.P., Which
indicates that adzes from the Tatagamatau quarry
were being distributed as far as the Manu'a Group by
at least the mid-first millennium A.D.
In Layer HIlA of Unit 27 we excavated a flaked,
tabular piece of daik gray basalt, extensively flaked,
but retaining some cortex on one surface. The flake
which measures 63.9 by 58.6 mm, and is 16.4 mm
thick, may be a large decortication or trimming flake
from adz manufacture.
In addition to the large diagnostic specimens
described above, we excavated five small flakes,
each with one or more ground and polished facets.
These are from Units 16, 17, 20, and 22 and all
derive from adzes, either from use or resharpening.
Four flakes were analyzed by XRF (see Weisler,
chapter 12). Two of these can be ascribed to te
Tatagamatau quarry site on Tutuila Island.
SHELL ADZ
A small adz of heavy shell, possibly Cassis sp.,
was found in association with plainware pottery at
the landfill site during the 1986 reconnaissance. The
adz is rectangular in shape with a slightly rounded
bevel. Shell adzes are very me in Samoa and may
have been restricted to the earlier ceramic period.
Buck (1930:353-54) records only two shell adzes in
the Bishop Museum collection from Samoa.
HAMMERSTONES
Two hammerstones, both from Layer IIA-1 of
Unit 9 in the main trench, were excavated. One is an
ovoid cobble of porphyritic igneous stone (with
abundant feldspars). It is 30 mm thick, has a
diameter of 93-105 mm, flat sides, and distinct
Non-Ceramic Portable Artifacts
0
5 cm
C
Figure 11.1
Basalt adzes from the ToWaga site: a, bevel section from Unit 20, Layer IIIB; b, butt section from Unit
23, Layer hUB; c, complete Type V adz from Unit 9, Layer HA-I (drawings by J. Ogden).
159
160
The Tolaga Site
pecking or damage along the margins. One face
appears to be ground smooth, peitaps during use as
an abrading or polishing stone. The second specimen is an elongate basalt cobble, beach-worn, with
pecking damage on the broader end. The cobble
measures 147 mm long (max. width 70 mm), and the
damaged surface has an area of 17.1 by 26.9 mm.
shell fragments. Thus, the total fishhook assemblage
from To'aga now stands at twenty-eight whole or
incomplete specimens, not including tabs and
unfinished fragments. This is by far the largest
assemblage of prehistoric fishing gear recovered
from Samoa and is a major addition to our knowledge of early Polynesian fishing.
The To'aga fishhook assemblage is remarkably
uniform in size and morphology, with only minor
variations. The hooks were all manufactured from
the body whorls of Turbo setosus, a gastropod
common on the reef edge of Ofu Island. The various
midden deposits contained large quantities of T.
setosus shell, some of which was probably manufacture debris (see Nagaoka, chapter 13). One worked
fragment from Layer IIB of the 1987 trench, probably an unfinished hook tab, was of the larger
and less commonly occurring species Turbo
narmoratus.
Examples of the hooks are illustrated in figures
11.2 and 11.3. They are small, andrather delicate,
and were probably used to take smaller reef fish.
The complete hooks have shank heights ranging
from 13.1 to 30.4 mm. Hook widths range from
10.1 to >20.8 mm. Most hooks appear to have been
FISHING GEAR
Turbo-Shell Fishhooks
Samoan archaeological sites have been notoriously poor in the preservation of bone or shell
artifacts, and only a few specimens of fishing gear
have ever been excavated (Green and Davidson
1969, p1. 23; Janetski 1980). The same has been true
of other Western Polynesian sites in Tonga and
Futuna (Kirch and Dye 1979). In our 1986 test
excavation at To'aga, two fragments of small Turboshell one-piece fishhooks were recovered (Hunt and
Kirch 1988:175, fig. 8, b-c). In 1987 the expanded
excavations yielded four neatly complete hooks and
fourteen hook fragments. In 1989, we recovered an
additional eight hooks or hook fragments, and a
large number of prepared tabs and unfinished Turbo
1.
f--l
-
b
a
c
A-
d
f
e
9
0
L
Figure 11.2
a
5cm
.-
a
.
Turbo-shell fishhooks from the To'aga site: a, roughed-out fishhook tab from Unit 21, Layer IIB; b,
well-ground fishhook tab from Unit 23, Layer IIIB; c, ground and perforated tab from Unit 30, Layer
II; d, unfinished fishhook from Unit 15, Layer U; e, head and shank fixxn Unit 27, Layer ILA; f,
complete hook from Unit 23, Layer IIIB; and g, hook with sharply inturned shank and head from Unit
20, Layer IIIB (drawings by J. Ogden).
Non-Ceramic Portable Artifacts
2: 3
4
161
5cm
U]
11 -1
Figure 11.3
Turbo-shell fishhooks and fishhook fragments from the 1987 excavation at ToWaga.
rotating in form, although one hook is technically of
the jabbing variety. The bends have an 'O' or 'U'
shape. Three hooks have a distinctive in-curved or
"bent" shank, strongly reminiscent of some early
Marquesan hooks (Suggs 1961:8 1, fig. 26). Two
specimens have an inner shank knob, presumably to
assist in line attachment. Several other shanks have
small notches or grooves on the outer shank face,
also for line attachment.
Turbo-Shell Fishhook Tabs
The manufacture of fishhooks from Turbo
setosus shell is well attested in the To'aga site by the
presence of numerous preforms or tabs roughed out
of the body whorls of this gastropod as well as by
worked shell fragments and many kinds of abrading
tools (see below). Several examples of fishhook tabs
are illustrated in figure 11.2. Of particular note is a
specimen from Unit 23, Layer IIIB, which has been
carefully shaped and fully ground on the exterior
surface and around the margins (fig. 11.2, b). This
tab measures 21.7 by 16.4 mm. Another specimen
(fig. '1.2, c), from Layer II of Unit 30, represents yet
a further stage in manufacture, with the entire center
of the tab removed by drilling and filing. This
specimen measures 19.8 by 13.4 mm. These tabs
indicate that the reduction procedure for the manufacture of Turbo-shell hooks at To'aga was as
follows: (1) a tab was first roughed out of the body
whorl of Turbo; (2) this roughout was then ground
flat on the exterior surface and carefully shaped by
grinding around the margins; (3) the interior was
then removed by drilling and filing; and (4) the gap
between the shank and point was opened last by
cutting and filing. Sinoto (1967:353, table 3)
remarks that "simple drilling" and "chipping and
filing" were the methods used by early Marquesans
in hook manufacture. Thus, not only the forns of
the To'aga hooks, but thle specific manufacture
methods, are consistent with the Marquesan hooks
for which the To'aga specimens may have been
162
The To'aga Site
prototypes.
During the 1987 excavations we did not make a
special effort to distinguish Turbo-shell tabs from
Turbo-shell midden or worked debris, and no exact
counts are therefore available. For the 1989 materials, however, shaped tabs were carefully separated
from the shell midden during the laboratory study of
faunal materials by L. Nagaoka. The following
counts by unit indicate the frequency of such
prepared tabs:
2 tabs
Unit 15:
1 tab
Unit 16:
7
tabs
Unit 20:
11 tabs
Unit21:
Unit 23:
9 tabs
Unit 30:
1 tab
Cypraea-Shell Caps
The caps or dorsa of large Cypraea shells
(especially C. tigris) comprised one component of
the Samoan octopus lure. Buck (1930:434-38, fig.
257, pl. XLI, B) describes and illustrates this apparatus. Three such dorsa were recovered together in
Layer III of Unit 28 and were possibly part of such
an octopus lure rig.
ABRADING TOOLS
Coral Abrader
From Layer III of Unit 11 we recovered a tabular shaped abrader of Porites sp. coral. The abraded
facet has a surface area measuring 50 x 60 mm.
Echinoid-Spine Abraders
The long spines of the slate-pencil sea-urchin
(Heterocentrotus mammillatus) have a natural
abrasiveness and thus were used throughout most of
Polynesia to manufacture fishhooks and other
objects of shell and bone. Two such abraders were
excavated from Layer IIA-1 in the main trench.
Both have distally abraded facets at an angle to the
longitudinal axis, as do the abraders reported by
Janetski (1980, fig. 43, g-i) from the early Falemoa
site in Western Samoa. A complete spine which has
been slightly faceted at the distal tip was excavated
in Layer II of Unit 15. From Layer IIIC in Unit 23
we recovered a sea urchin spine abrader which had
been distally abraded to a point (circular section),
presumably from use as a drill in the manufacture of
Turbo-shell fishhooks (fig. 11.4, a). The tip only of
a circularly abraded echinoid spine was also found in
Layer IIB of Unit 20. A particularly interesting
echinoid abrader was found in Layer IIB of Unit 28,
and is illustrated in figure 1 1.4, j. This spine, 73.2
mm long, has been equally reduced on two sides
from the distal end to form a thin, saw-like blade. It
would appear that this blade edge was purposefully
produced in order to cut shell or bone objects. Also
from Unit 28 (Layer III) was a small fragment of sea
urchin spine which was abraded laterally to form a
flat surface. All of these abraders were likely used to
manufacture the Turbo hooks and other artifacts of
shell.
Shell-Bead Abrader
An abrader of Porites coral, specifically adapted
for grinding small Conus-shell beads, was recovered
from Layer IIIB of Unit 23 (fig. 11.4, h). The
abrader consists of a naturally waterworn coral
pebble (68.6 by 46.4 mm and 19 mm thick) which
has been flattened on one face by grinding. In the
center of this face is a single depression or "cupule"
with a diameter of 9.1 mm, about 1-2 mm deep.
This depression has a central "nipple" which results
from positioning a Conus-shell spire in the depression, and then using the abrader to grind the shell
against a larger grindstone. Such specialized Conusshell bead abraders had been reported from Vanuatu
(Garanger 1972) and from Vanikoro in the Santa
Cruz Islands of eastern Melanesia (Kirch 1983:102104, fig. 16), but were previously unknown from
Western Polynesia. Recently, however, Sand (pers.
comm.) excavated such an abrader from the Asipani
Lapita site on Futuna Island.
ORNAMENTS
Conus-Shell Beads
From Layer IIB in the main trench are two
delicate beads of Conus sp., very well ground, with
diameters of 5.6 and 5.9 mm, and thicknesses of 1.9
and 2.1 mm (fig. 11.5). A slightly larger bead or
ring of Conus, complete and very well ground (fig.
11.4, g), was found in Layer IIIC of Unit 20. This has
an extemal diameter of 15.8 mm and is 2.4mm thick.
Conus-Shell Rings
Layer IIB in the 1987 main trench produced two
fragments of larger Conus sp. rings, very well
ground, with original diameters of about 50 mm (fig.
Non-Ceramic Portabk Artifacts
Ia
163
f
e
cK
(d
:a
-M
9
I
b
-10-11
I11
I
I
1
i
0
Figure 11.4
5cm
Miscellaneus artifacts from the To'aga site: a, echninoid spine abrader from Unit 23, Layer HIC; b,
bone point from Unit 27, Layer IIIB; c, Conus-shell ring fragment from Unit 29, Layer ITIB; d, Conusshell ring fragment from Unit 29, layer IUB; e, drilled shark's tooth from Unit 21, Layer HI; f, Conusshell bead from Unit 30, Layer II; g, Conus-shell bead from Unit 20, Layer IIIC; h, coral abrading
stone for grinding shell beads, from Unit 23, Layer IB; i, unfinishedTridacna-shell ring from Unit 23,
Layer IIIC; and j, echinoid-spine abrader from Unit 28, Layer UB (drawings by J. Ogden).
11.5). One fragment had been sharpened to a point
after breaking. Layer II in Unit I I produced a
fragment of a large shell ring or anmband, made
either of a large species of Conms, or possibly of
Tridacna. The ring fragment is 7.3 by 11.6 mm in
thicness, and has a reconstructed diameter of about
70 mm (fig. 11.5). A similar annband fragment
from the Falemoa site is illustrated by Janetski
(1980: fig. 45, b). From Layer Im of Unit 16 we
recovered a fragment of a Conus-shell ring with a
cross section measuring 3.5 by 4.8 mm, and a
reconstructed diameter of about 35 mm. Layer rIB
of Unit 29 produced another Conus-shell ring
fragment (fig. 1 1.4, d) with a roughly rectangular
164
The To'aga Site
cross section (5.3 by 4.8 mm), which would have
had an original diameter of about 30 mm. Based on
Buck's extensive compilation of Samoan material
culture (1930), Conus-shell rings were not a part of
the Samoan ornamental repertoire in historic times.
Indeed, they were probably associated only with the
early ceramic period.
Bone Point
From Layer IIIB of Unit 27 was recovered a facet bone "point" of unknown function (fig. 11.4, b).
The bone, of either dog or pig, has been carefully
faceted to a chisel-like tip, across which were
abraded a series of fine grooves.
WORKED SHELL
Ufinished Tridacna-Shell Ring
Approximately one-half of a Tridacna-shell ring
which broke during the process of manufacture was
found in Layer IIIC of Unit 23 (fig. 11.4, i). The
Tridacna valve incorporates part of the hinge. It was
worked by chipping and pecking to create a central
perforation. Presumably the artifact broke during
this chipping process, prior to the initiation of
grinding. The specimen has an outer diameter of 73
mm, and the central perforation is 15 mm in diameter.
Nerita-Shell Beads
From Layer IIB in the 1987 main trench were
two Nerita sp. shells with artificial perforations in
the basal whorl, perhaps for stringing as beads. A
third specimen was found in Unit 21. Buck
(1930:638) mentions the use of sea shells as beads
but does not illustrate examples or provide further
details.
Gastropod Bead
A small gastropod (species unknown) from
Layer II of Unit 30 has had both the spire and basal
whorl removed by grinding (fig. 11.4, f), leaving
only the midsection of the shell as a bead. It has a
diameter of 14.0 mm.
Echinoid-Spine Bead
In Layer IIIC of Unit 20 we found a unique bead
made from a section ofHeterocentrotus mammilat
spine which was double-drilled to form a central
perforation. The bead measures 13.5 mm in diameter and is 10.5 mm thick.
MISCELLANEOUS ARTIFACTS
Drilled Shark's Tooth
A small shark's tooth (14.7 mm high) was found
in Layer IIB of Unit 21 (fig. 11.4, e). This has been
drilled (hole diameter 2.6 mm), presumably in order
to lash the tooth to a handle.
A large piece of Tridacna shell (possibly from T.
gigas) which has been chipped around the edges to a
roughly rectangular shape (measuring 170 by 135
mm) was found in Layer IIIB of Unit 29. This may
have been intended as a Tridacna adz preform or
may have been for the manufacture of some other
object, such as a shell ring.
From Layer III of Unit 28 we recovered two
matching pieces of worked Conus shell. These
consist of part of the main body whorl, with a cut
and beveled edge near the spire. These are presumably rejected material resulting from the removal of
a large Conus spire, as a part of the manufacture
process for Conus-shell rings.
The chipped basal whorl section of a species of
Trochus or Textus shell was found in Unit 21. This
may have been intended to be a ring or armband.
Various small pieces of worked shell were
recovered throughout the excavations. Most of these
are of Turbo spp. and relate to fishhook manufacture.
In the 1987 excavated material these were not
distinguished from the Turbo-shell midden. In 1989,
however, all worked Turbo shell was carefully
segregated during faunal analysis by L. Nagaoka,
yielding the following frequencies by excavation
unit:
2 specimen(s)
Unit 15
5
Unit 16
1
Unit 17
14
Unit 20
3
Unit 21
1
Unit 22
9
Unit 23
2
Unit 25
1
Unit 26
5
Unit 28
1
Unit 30
Five specimens of worked pearl shell (Pinctada
sp.), were also recovered from Units I 1 22, 23, 29,
and 30. A triangular-shaped specimen from Layer
Non-Ceramic Portable Artifacts
Figure 11.5
165
Miscellaneous artifacts from the To'aga site: left to right, Conus-shell ring fragments, Conus-shell
beads, and echinoid-spine abraders.
111B of Unit 23 is of special interest because it shows
distinct filing or cutting marks on all three margins.
This piece, measuring 24.9 by 16.9 mm, is probably
detritus from the manufacture of some other object,
rather than a preform.
UNRETOUCHED LITHICS
Basalt Flakes
Flakes of basalt were surprisingly uncommon in
the Totaga excavations. During the 1989 excavations, when particular attention was paid toward the
recovery of such lithics during screening, only
eighteen flakes were noted. Eleven of these are from
Unit 23 [Layer III], suggesting that this may have
been a locus of basalt flaking activity. The other
flakes are from Unit 28 (three flakes) and Unit 29
(four flakes). Most of these are rather small and
could derive from adz use, although they do not
show polished surfaces.
Obsidian Flakes
A number of very small flakes of an opaque,
black, low-silica volcanic glass or obsidian were
found from various excavation contexts. Most of
these are less than 5 mm in size. As the dike complex of Leolo Ridge overlooking the To'aga site has
many glassy chills along the dike margins (see chapter 2), it is most probable that these "flakes" are natural
and simply derive from the talus rockfall above the site.
One small core from Layer IIIC of Unit 23 is
completely different, however, from the other
obsidian specimens. This is of a reddish-brown
color, with black spots and banding. The "core"
measures 12.2 by 13.3 mm. Its geological provenance is unknown, but it is very likely an import to
Ofu.
REFERENCES CITED
Best, S., H. Leach, and D. Witter 1989. Report on
166
The To'aga Site
the second phase of fieldwork at the Tatagamatau site, American Samoa, July-August
1988. Department of Antrpology, University of Otago, Dunedin.
Bests S., P. Sheppard, R. Green, and R. Parker 1992.
Necromancing the stone: Archaeologists
and adzes in Samoa. Journal of the
Polynesian Society 101:45-85.
Buck, P. H. [Te Rangi Hiroa] 1930. Samoan Material Culture. Bernice P. Bishop Museum
Bulletin 75. Honolulu.
Dye, T. S. 1987. Social and cultural change in the
prehistory of the ancestral Polynesian
homeland. Unpublished Ph.D. dissertation,
Yale University, New Haven.
Emory, K. P., W. J. Bonk, and Y. H. Sinoto 1959.
Hawaiian Archaeology: Fishhooks. Bernice
P. Bishop Museum Special Publication No.
47. Honolulu: Bishop Museum Press.
Garanger, J. 1972. Arch!ologie des NouvellesHebrides. Publications de la Sociat des
Ocdanistes No. 30. Paris.
Green, R. C. 1974. A review of portable artifacts
from Westem Samoa. IN R. C. Green and J.
Davidson, eds., Archaeology in Western
Samoa, Vol. II, pp. 108-154. Auckland
Institute and Museum Bulletin 7.
Green, R. C., and J. Davidson 1969. Descrpton and
classification of Samoan adzes. IN R. C.
Green and J. Davidson, eds., Archaeology in
Western Samoa, Vol. I, pp. 21-32. Auckland
Institute and Museum Bulletin 6.
Green, R. C., and J. Davidson, eds. 1969. Archaeology in Western Samoa, Vol. 1. Bulletin of
the Auckland Institute and Museum.
. 1974. Archaeology in Western Samoa, Vol. II.
Bulletin of the Auckland Institute and
Museum.
Janetski, J. 1980. Shell, bone, coral, and urchin spine
artifacts. IN J. D. Jennings and R. N.
Holmer, eds., Archaeological Excavations
in Western Samoa, pp. 123-31. Pacific
Andhpological Records 32. Honolulu:
Bishop Museum.
Kirch, P. V. 1983. An archaeological exploration of
Vanikoro, Santa Cruz Islands, Eastern
Melanesia. New Zealand Journal ofArchaeology 5:69-113.
-. 1988. Niuatoputapu: The Prehistory of a
Polynesian Chiefdom. Thomas Burke
Memonal Washington State Museum
Monograph No. 5. Seattle.
Kirch, P. V., and T. Dye 1979. Ethnoarchaeology
and the development of Polynesian fishing
strategies. Journal of the Polynesian Society
88:53-76.
Leach, H., and D. C. Witter 1987. Tataga Matau
rediscovered. New Zealand Journal of
Archaeology 9:33-54.
. 1990. Further investigations at the Tatagamatau
site, American Samoa. New Zeaad
Journal ofArchaeology 12:51-83.
Poulsen, J. 1987. Early Tongan Prehistory: The
Lapita Period on Tongatapu and Its Relationships. 2 vols. Terra Australis 12.
Canberra: Department of Prehistory,
Australian National University.
Sinoto, Y. H. 1967. Artifacts from excavated sites in
the Hawaiian, Marquesas, and Society
Islands: A comparative study. IN G. A.
Highland et al., eds., Polynesian Culture
History: Essays in Honor ofKenneth P.
Emory, pp. 341-62. Bernice P. Bishop
Museum Special Publication No. 56.
Honolulu: Bishop Museum Press.
Suggs, R C. 1961. The Archaeology ofNuku Hiva,
Marquesas Islands, French Polynesia.
Anthropological Paprs of the American
Museum of Natural Hisory 49(1). New York
12
CHEMICAL CHARACTERIZATION AND
PROVENANCE OF MANU'A ADZ MATERIAL USING A
NON-DESTRUCTIVE X-RAY FLUORESCENCE TECHNIQUE
MARSHALL WEISLER
INTRODUCTION
THE COLONIZATION SiRATEGIEs employed during
the settlement of Polynesia and the subsequent
diversification of island societies are key issues in
Oceanic prehistory. Yet it is only recently that
lsearchers have begun to amass empirical evidence
of inter-island communication-throughout island
sequences -that undoubtedly influenced the direcdon and tempo of post-settlement island histories. I
refer here to recent finds of Tongan pottery in the
Cook Islands (Walter and Dickinson 1989:465), of
Samoan adz material in Fiji (Best 1984; Best et al.
1992) and the Cook Islands (Weisler 1993), and of
inter-island transport of volcanic glass and finegrained basalt in Hawaii during late prehistory
(Weisler 1990; et al. In prep. [For a review of
Polynesian basalt adz provenance studies, see
Weisler In press a.]). Isolation has been invoked as
'"the most fundamental" mechanism of divergence in
Polynesia (Kirch and Green 1987:440), but "how . . .
different extremes of isolation have influenced the
evolution of human diversity from island to islandand perhaps even on the same island" (Terrell
1986:122-23) remains a problem that must be
addressed for each island sequence. Dissimilarities
may also originate because of continuing contact
(see Terrell 1986:147). Isolation may have contrib-
uted to regional variations in portable artifacts,
architecture, and language in New Zealand (Prickett
1982), Society Islands (Emory 1933), and Hawaii
(Kirch 1985, 1990). In contrast, inter-island communication is inferred by parallel histories in subsistence and technological transformations in the
northemrn and southemrn Marquesas (Rolett 1990:363)
and similar changes in ceramic styles and adz
evolution in the Samoa-Tonga-Fiji area at similar
times imply continuous inter-island contact
(Davidson 1977, 1978, 1979).
Tracking the spread and subsequent communication of prehistoric Polynesian societies has been
difficult without the widespread occurrence of
pottery and obsidian, artifact classes that have
proven especially useful for demonstrating interaction in the southwest Pacific. Without inferring
contact from similar styles of fishhooks, adzes, and
architecture, Polynesian archaeologists are left with
few items to track empirically the intra- and interisland movement of things. Connecting stylistic
nodes, however, may only reflect convergent
evolution (Kirch and Green 1987). I may risk being
accused of taking a hard-line empiricist view of
culture contact and change, but I think it crucial to
confront the problem withseparating stylistic from
functional dimensions (Dunnell 1971:26-30) in
relation to documenting the movement of exotic raw
168
The To'aga Site
materials and finished objects (e.g., peairlshell and
industrial stone) between and within island groups.
The spatial and temporal dimensions of contact
spheres (Irwin 1990:92; Walter 1990; Weisler In
press b), can be delimited more accurately using
exotic materials than by connecting artifact styles.
Consequently, the distribution of exotic materials
provides a better framework for evaluating the role
of isolation and communication in shaping historical
developments of island societies.
This preliminary effort to document inter-island
communication during the prehistory of the Manu'a
Islands should take into account the region's
geography. These islands are situated 100 km east
of the most important source of adz material in
central Polynesia, the Tatagamatau adz quarry,
Tutuila (Leach and Witter 1987, 1990). The distance and often turbulent ocean conditions between
Tutuila and the Manu'a Group may have reduced
the frequency of contacts (Hunt and Kirch
1988:155).
THE COLLECTIONS
Two questions are addressed by XRF analysis of
the Ta'u and Ofu island assemblages and source
material: (1) Did any artifacts originate at the
Tatagamatau adz quarry complex on Tutuila Island
located 100 km to the west?; and (2) What was the
geochemical variation of adz material and
unmodified basalt flakes from Manu'a sites? Although it seemed unlikely that I would be able to
assign a provenance to all artifacts, I could determine
if specific geochemical groups were correlated with
certain tool classes. Interaction spheres could be
delineated by identifying similar rock at different
sites and, perhaps, different temporal periods. An
additional goal of the XRF analysis was to generate
geochemical data for the important Tatagamatau
quarry complex on Tutuila and two local To'aga
sources that may have been used.
The Archaeological Sample
The archaeological contexts of the Ta'u and Ofu
island assemblages are described elsewhere (Hunt
and Kirch 1988; Kirch et al. 1990; chapters 5 and 11,
this volume) and are not reiterated here. The
artifacts selected for compositional analysis consisted of all whole or fragmentary adzes, all flakes
exhibiting polished surfaces and assumed to be adz
fragments, and a sample of unmodified basalt flakes
which represented the macroscopic vanriability
present. Table 12.1 presents characteristics of the
analyzed artifacts. Specimens ranged from 1.2 to
421.9 gmins and 22.08 to 136.30 mm in length. The
thinnest sample was 3.53 mm.
Source Material
During their 1987 field season, Kirch and Hunt
collected nine samples of source rock from the
Tatagamatau quarry complex on Tutuila (Best et al.
1989; Leach and Witter 1987). On Ofu Island, they
collected four samples from dike swarms at Mako
Ridge and three samples from Fa'ala'aga. The
Tatagamatau quarry complex consists of numerous
cut-and-fill terraces, stone-working areas, pits, and
possible fortifications along several ridges and scree
slopes which together may cover more than 110
acres (Best et al. 1989; Leach and Witter 1987,
1990). The raw material derives from a dike complex at the foot of a waterfall at the head of a small
gulch (P. Kirch, personal communication 1990,
Leach and Witter 1987:39). Fine-grained dike rock
may have been removed from dikes but most likely
was prised from the soil and rock scree of slopes and
from streambeds. Although I have not visited
Tatagamatau, the geological setting appears to be
similar to the Waiahole quarry on windward O'ahu
(Dye et al. 1985), albeit on a much smaller scale.
Two collections of source rock were made by Kirch
and Hunt during a visit on July 3, 1989. Eight
samples, consisting of large primary or secondary
flakes with obvious bulbs of percussion, were
collected at three locations along the main spur of
the complex which trends NE-SW, and an additional
flake was retrieved from Area 3 (figure 12.1). The
exterior color of the source rocks varied from dark
brown (1OYR 3/3), dark gray (10YR 4/1), to dark
grayish-brown (2.5Y 4/2) due to weathering and
contact with lateritic soil. Fresh saw-cut surfaces are
mostly dark bluish-gray (Gley card, SB 4/1) grading
to dark gray (7.5R 4/0 and 2.5Y 4/0). Even prior to
geochemical analysis, it was not likely that nine
samples would document the total variability of
quarry rock from so large a quarry, but this
collection did contribute significantly towards
developing a strategy for collecting additional
material.
Chemical Characterization of Manu'a Adz Material
Table 12.1
Basalt Artifacts Analyzed by EDXRF
Lab
Number
89-22*
89-25
89-26*
89-31
89-30*
89-28
90-30*
90-33
89-18
89-34*
89-33*
89-32
89-19
89-17*
90-22
90-23
90-42
90-46
9045
90-38
90-27*
90-37
90-41
90-34
90-48
90-29
90-36
9040
9043
9049
9044
90-25
90-24*
90-26
90-31
90-39
90-47
90-35
*
=
Artifact
Number
Weight (g)
Fiti'uta 1
11-1-S-4
11-2-S- 1
11-2-S-2
11-2-S-3
11-2-S-8
11-2-9
11-52-S-9
13-S-3
13-S-5
13-S-6
13-S-9
13-1-S-36
13-2-3-1-1
13-1-9-2-4
13-1-13-87
13-1-16-1
13-1-16-11
13-1-16-11-81
13-1-16-12-83
13-1-20-4-118
13-1-20-5-121
13-1-20-5-122
13-1-20-5-123
13-1-20-6-139
13-1-20-6-140
13-1-20-6-141
13-1-20-7-143
13-1-20-9-147
13-1-21-7-151
13-1-22-2-171
13-1-22-2-173
13-1-22-2-174
13-1-23-7-194
13-1-27-5-39
13-1-27-5-40
13-1-27-5-41
13-1-27-6-42
Probable source is Tatagamatau.
24.8
70.9
41.4
35.3
21.4
33.1
60.9
133.1
48.4
44.7
33.8
44.4
63.9
63.4
421.9
0.7
5.5
1.5
127.9
20.5
3.4
1.2
4.4
11.3
11.4
128.4
56.7
3.2
20.6
59.3
3.7
2.1
0.4
211.4
89.4
128.1
2.1
9.5
Description
Quadrangular, reworked adz
Quadrangular adz
Quadrangular adz, front section
Quadrangular adz, butt section
Quadrangular adz, front section
Quadrangular adz, front section
Quadrangular adz, butt section
Quadrangular adz, front section
Quadrangular adz, front section
Quadranglulkar (?) adz, front (?) section
Quadrangular adz, front section
Quadrangular adz
Plano-convex adz
Quadrangular adz, midsection
Plano-convex adz
Adz flake
Adz fragment
Unmodified flake
Retouched flake
Retouched flake
Adz flake
Unmodified flake
Unmodified flake
JUnmnodified flake
Unmodified flake
Plano-convex adz, front section
Retouched flake
Unmodified flake
Unmodified flake
Used flake
Unmodified flake
Adz flake
Adz flake
Plano-convex (?) adz
Quadrangular (?) adz
Quadrangular (?) adz, butt (?) section
Unmodified flake
Unmodified flake
169
170
The To'aga Site
Situated at about 350 m elevation and exposed
by a recent road-cut, the Mako Ridge dike swarm on
Ofu consists of relatively dense basalt. Weathering
has smoothed the normally angular, loose dike rock
into sub-rounded cobbles which lie buried within a
thick soil matrix. It is doubtful that this material was
used for making stone tools, but four samples were
collected for chemical analysis. The weathered
exterior surface was gray in color (7.5YR 5/0) and
fresh breaks, dark gray (2.5Y 4/0).
A large concentration of dikes is found along a
steep ridge towards the east end of Ofu. The
Fa'ala'aga dike swarm (Stice and McCoy 1968) may
have been a more important source of local, medium-grained basalt during prehistory. Angular rock
from the eroding dikes form a large scree slope
which descends to the To'aga coastal zone where
stone is readily available (see Kirch, chapter 2).
Three fresh samples were removed from dikes
exposed by a modem road-cut (see fig. 12.2) to
provide the general chemical composition of the dike
swarm although many dozens of dikes may exist
each with varying chemical compositions. Surface
color of these rock samples is very dark gray (2.5Y
3/0) with fresh breaks being gray (2.5Y 5/0).
METHODS
Because this specific technique of non-destructive, energy-dispersive XRF had not been used
previously with adz material from Oceania, each
individual source sample was divided and analyzed
by two techniques: destructive XRF using pelletized
samples and non-destructive XRF using whole
specimens. Destructive XRF analyzes a wider range
of elements, and results are fully quantitative for
most elements and comparable to other data sets
now being developed for adz source rock in
Polynesia The non-destructive technique focused
on those elements that are more reliably detected
with this procedure, and results presented are
considered here as semi-quantitative.
Destructive XRF
Source samples from Tatagamatau, Mako
Ridge, and Fa'ala'aga were analyzed by Dr. Peter R.
Hooper at the Department of Geology, Washington
State University, on 16 and 17 October 1989. The
equipment and operating procedure are paraphrased
from Hooper and Johnson (1987; this manuscript is
available from Dr. Hooper or myself). Twentyseven elements were analyzed on an automatic
Rigaku 3370 spectrometer. For the oxides, SiO2,
A1203, TiO2, FeO*, MnO, CaO, MgO, K20, Na20,
and P205, the weight percents are presented. For the
other seventeen elements, the weight percent or ppm
(parts-per-million) of the element are provided. All
elements are analyzed on a single 2:1 lithium
tetraborate:rock powder fused disk. Each element
analysis is fully corrected for line interference and
matrix effects of all other analyzed elements. The
results are normalized and printed out with total iron
expressed as FeO*.
Whole rock samples were reduced with a
tungsten carbide jaw crusher to small chips generally
about 1 cm in average size. Chips were hand picked
and reduced further in a Tema swingmill shatterbox
with tungsten carbide surfaces where they were
milled for two minutes. Rock powder weighing 3.5
grams was mixed with pure lithium tetraborate and
emptied into a 34.9 mm-intemal-diameter, graphite
crucible. The samples were fused in a muffle
furnace at 1000'C. After grinding, the samples were
fused a second time. The lower, flat surfaces of the
fused disks were ground on coarse (240) grit and
fine (600) grit. Then, they were washed in an
ultrasonic cleaner, dried, and loaded into the XRF
spectrometer.
The x-ray intensity of each element from
unknown samples was measured and compared to
values of two beads from each of eight international
rock standards (U.S. Geological Survey [U.S.G.S.]
standards PCC- 1, BCR-1, BIR-1, DNC- 1, W-2,
AGV-1, GSP-1, and G-2). Standards were run and
recalibrated after processing between one and two
hundred unknown samples. U.S.G.S. BCRP-84 and
GSP-1 were used as internal standards and routinely
run after every twenty-eight samples to check
instrument performance. Drift between standard
calibrations was due almost entirely to iron which
produced slightly higher values for other elements
presented as oxides, weight %. When thirty analyses
of the U.S.G.S. standard BCRP were compared to
accepted values, the precision or standard deviation
of oxide values to known measures is 0.22% ppm.
Values reported for thirteen trace elements in ppm
were below 6% ppm at one standard deviation,
except barium (16.2%) and vanadium (10.0%o). MThe
highest precision was achieved with Rb, Sr, Zr, Y,
Chemical Characterization of Manu'a Adz Material
Figure 12.1
171
Map of Tatagamatau quarry showing locations of source samples (after Best et al. 1989, fig. 21).
Oa,Cu, and Zn. Ni, Cr, Sc, V, and Ba should be
xgarded as semi-quantitative below the 30 ppm
kvel. Rb, Sr, Zr, Nb, and Y have satisfactory
Non-destructive Energy Dispersive XRF
precisions and accuracies down to one to three ppm,
while Y and Nb could be measured to 0.1 ppm.
Evaluation of accuracy suggested that variation
between different samples of standard powder or
onhomogeneity resulting from sample preparation
is greater than inaccuracies caused by inadequate
Portions or splits of all source samples from
Tatagamatau, Mako Ridge, and Fa'ala'aga were
analyzed by Weisler along with thirty-eight artifacts
during four runs in late 1989 and early 1990. Sample preparation consisted of submerging specimens
matrix and interference correction.
172
The To'aga Site
70T
Mako (x)
0
Tatagamata[
0
1"
Co
X
A
Tatagamatau
50
subsource
L.
Mako (x)
A
60
s
s
A
Fa'ala'aga (x)
A
Fa'ala'aga (X)
=ubsource
40
N
40
40
50
50
60
60
70
70
80
80
9
0
90
100
10
110
110
120
120
130
Nb/Sr X 1000
Figure 122
A compaison of source rocks analyzed by non-destuctive x-ray flourescence using whole specimans
and XRF with fused disks (indicated by italic type).
in a sonic bath of distilled water for up to one hour
then air drying. Some artifacts, deeply stained with
lateritic soil, were scrubbed with brushes as well.
Carbonate encrustations on a few artifacts were
removed with a 10% solution of HCI and then rinsed
in distilled water. The EDXRF instrument is limited
to the maximum weight and size of specimen that
can be analyzed. Although the opening of the
sample holder where the x-ray beam is directed
through to the specimen is 32 mm in diameter,
careful placement of the artifact or rock in the
EDXRF instrument may accommodate samples (as
in this analysis) up to 421 gmns and 136 mm long.
Molding clay can be used to secure specimens on the
tray which can hold up to twenty samples. Each
artifact was carefully examined to locate the flattest
surface for analysis that could be accommodated
within the space parameters of the sample chamber
(see illustration in Bouey 1991:fig. 5). Source
samples were cut to size and only fresh saw-cut
surfaces were analyzed.
Laboratory facilities and equipment were
provided by the Department of Geology and Geophysics, University of Califomrnia, Berkeley. The
same equipment and nearly identical operating
conditions were used as reported by Hughes
(1986:esp. 25-30) in his comprehensive study of
California and Oregon artifactual and source obsidian. The XRF spectrometer consisted of a Spectrace
440 energy dispersive machine and 572 power
supply (50kV, lmA), 534-1 pulsed tube control, 588
bias/protection module, 514 pulse processor or
amplifier, Tracor Northemrn 1221 100 mHz ADC
converter, and a Tracor Northern 2000 computer
based analyzer with an LSI-11 microcomputer
(Hughes 1986:25). The Si(Li) solid state detector
with 144 eV resolution (FWHM) at 5.9 keV in a 30
mm squared area was used for detecting all x-ray
intensities. For analysis of trace elements in the midz energy range, a rhodium (Rh) x-ray tube was used
for primary x-ray excitation at 30.0 keV, .20 mA
pulsed, with a .05 mm Rh primary beam filter in an
air path at 300 seconds livetime. Analytical lines
used for analysis were Ni (Ka), Cu (Ka), Zn (Ka),
Ga (Ka), Pb (Lb), Th (La), Rb (Ka), Sr (Ka), Y (Ka),
Zr (Ka), and Nb (Ka). For rare earth elements, an
Am241 100 mCi radioscope source was used in the
20-60 keV range at 300 seconds livetime (Hughes
1986:26), and analytical values were derived from
the Ka lines of Cs, Ba, La, Ce, Pr, Nd, and Sm. The
x-ray tube was operated at 15.0 kV, .40 mA pulsed,
with an Al primary beam filter in a vacuum path at
200 seconds livetime for Ka lines of Fe, Mn, and
The elemental data as reported represent one
analysis per specimen. While Bouey (1991) has
demonstrated some variability with multiple analyses of the same obsidian specimens when values are
presented in strictly quantitative (ppm) data, these
Ti.
Chemical Characterization of Manu'a Adz Material
effects are minimized when presenting data as ratios
(see, for example, Jack and Carmichael 1969).
The
XRF
technique has its
greatest accuracy in
detecting elements in the mid-z range (e.g.,
bidium, strontium, yttrium, zirconium, and niobium)
which was confirmnned by this study for both destructive and non-destructive sample preparations
analyzed on their respective instruments. It is
fortunamte that these elements are of particular interest
to igneous petrologists (Cox et al. 1979:332). For
pressed pellets or fused disks these particular mid-z
elements have satisfactory precisions and accuracies
down to 1 3 ppm, while niobium and yttrium could
probably be measured to 0.1 ppm (Hooper and
Johnson 1987). In basalt, mid-z elements are also
present in sufficient concentrations to be easily
measured, and detection in whole, unaltered specimens ranges from 10 ppm to 100% with an accuracy
of ± 2 5% under favorable conditions (Parkes
1986:153). Lighter elements are not only harder to
detect and measure by XRF, but they are usually not
present in great abundance. Iron, titanium, and
magnesium values, reported here, were detected
under vacuum. Lead (Pb), although detected during
the mid-z analysis, was not used to discriminate
sources or characterize artifacts due to its presence in
low abundance and its susceptibility to atmospheric
contamination (Flanagan 1969:82).
Rock standards have been used in routine XRF
ru-
-
-
analyses for about twenty-five years (Flanagan
1969), and currently at least 272 geostandards are in
use worldwide (Govindaraju 1989). Standards are
important for calibrating the XRF instrumentation
for matrix effect corrections and for monitoring the
precision and accuracy during analysis (Germanique
and Briand 1985). Using standards thought to be
close in composition to the unknowns limits the
efficacy of matrix correction programs to a restricted
range of elements. Conversely, many different
standards provide a greater range of elements and
values for calibrating the program and evaluating
unknowns. Therefore, more than ten internal
standards were used to calibrate the Spectrace 440
machine used in this study. U.S.G.S. standard
RGM-1, a rhyolite from Glass Mountain, California,
was used to monitor precision and accuracy during
analysis and the results are presented in table 12.2.
Ppm values reported by Govindaraju (1984, 1989)
ar preferred "working values" which are the
average of at least forty results from more than four
techniques of analysis (1989:7). The accuracy and
precision values reported for this study in ppm and
selected ratios are reasonably close to accepted
standard "working values."
Due to its cryptocrystalline texture and homogeneous distribution of elemental abundances, it is not
surprising that obsidian has garnered most analytical
attention in XRF studies. Large grain sizes can
Table 12.2
Evaluation of Analytical Accuracy and Precision
for U.S.G.S. Standard RGM-1
Rb
Sr
Zr
Y
Nb
Pb
Th
Zr/Sr
Nb/Sr
Govindaraju
Govindaraju
1984
1989
155
100
200
25
9.4
21
15
2
0.094
149
108
219
25
8.9
24
15.1
2.03
0.082
173
This study (n=4)
Precision
Accuracy
159.9
107.3
227.2
27.7
10.3
28.6
20.4
2.12
0.096
156.3-163
106-108
225.3-229.9
25.9-29.2
9-11.7
27.1-31.1
18.9-22.1
Govindaraju (1984, 1989) pressed powder samples; this study, whole specimens analyzed.
174
The Tolaga Site
distort XRF analysis of whole samples. Pressed
pellet samples are prepared by reducing whole rock
to a fine powder estimated to be 90% less than 400
mesh (Bice 1980:19) or ca. 40 microns. However,
very fine-grained basalt may have more than 130
grains
per mm
(ca.
160 microns
per grain).
Table 12.3
Variation of Oxides and
Elements from the Mo'omomi
Basalt Quarry, Hawaiian Islands
While
this grain-size is early four times larger than pressed
pellet samples, it has not been demonstrated that
grains of this size adversely effect XRF analysis.
This subject should be investigated further. The
distribution of elemental abundances within a single
basalt specimen may not be as homogeneous as
obsidian, but a comparison of the date in table 12.3
is quite instructive. Here, the distribution of selected
elements within a quarry (1.8 hectares in size) are
indeed quite regular, and perhaps this is more so for
individual rocks or artifacts.
Another factor which can affect XRF analysis is
an uneven specimen surface. Analyzing adz material has a distinct advantage, however, especially
over bifacially flaked obsidian artifacts, because,
almost by definition, most adz surfaces are extremely flat and, in many cases, are ground to a near
mirror-like finish. Recalling that fused disks are
finished by 240 (coarse) and 600 (fine) grit prior to
XRF analysis (Hooper and Johnson 1987), examined under 10-40X magnification source rock from
eight west Moloka'i Island basalt quarries whose
material exhibited a range of textures and phenocryst
sizes and densities. These samples were polished
with 600 grit mesh and compared to the adz material
in the present study. The prepared specimens had
uniform, smooth surfaces and occasional striations
formed by disintegrating phenocrysts that had been
trapped between the rock and grinding plate. Although the artifacts were not as uniform in contour
and had many more striations, portions of the
artifacts were as smooth-if not more finely polished-4han the prepared specimens. Therefore,
careful selection of artifact surfaces for EDXRF
analysis may limit the amount of analytical distortion
caused by uneven sample surfaces.
To reduce or eliminate the effects of uneven
sample surface, many researchers have advocated
presenting elemental abundance values as ratios.
"In spite of variations in effective sample surface
of randomly broken pieces or loosely packed grains,
relative intensities may be very precisely determined" (Jack and Carmichael 1969:30; see also
Range
Mean
n=10
Oxide (weight %)
46.16±+0.12
SIO2
15.72 0.04
AL203
4.17 0.01
TIO2
0.20
14.21
FEO*
0.19 0.01
MNO
8.53 0.03
CAO
6.32 0.08
MGO
0.92 0.02
K20
3.52 0.07
NA20
0.62 0.01
P205
±
±
±
±
±
±
±
±
Element (ppm)
Ni
68.30 1.90
5.40 1.69
Cr
18.50 2.58
Sc
302.20 7.70
V
215.50 20.69
Ba
14.40 0.92
Rb
786.00 6.62
Sr
Zr
272.50 2.91
Y
37.80 0.98
27.09 1.19
Nb
23.30 1.79
Ga
11.80 5.62
Cu
Zn
133.30 1.90
4.30 1.35
Pb
18.80 9.05
La
62.00 9.89
Ce
0.80 0.98
Th
45.9646.36
15.66-15.81
4.157-4.183
13.91-14.04
0.178-0.202
8.50-8.58
6.18-6.42
0.90-0.95
3.40-3.67
0.613-0.634
±
64-70
3-8
15-23
283-310
187-258
12-15
775-798
268-278
37-40
25.0-28.9
21-26
4-21
±
131-137
±
±
±
±
±
±
±
±
±
±
±
±
±
±
2-6
2-31
49-80
0-3
*total iron
Analyst: Dr. Peter R. Hooper, Dept. of
Geology, Washington State University
Stross et al. 1968:82). Sheets et al. (1990:149-50)
concur that errors introduced by sample size and
shape are largely cancelled by "the use of abundance
ratios of elements having nearly the same energy
(e.g., Rb, Sr, and Zr)." These observations were
175
Chemical Characterization of Manula Adz Material
taken into consideration for the present study.
Ever since the pioneering work of Parks and
Tieh (1966), selection of elements for obsidian
analysis has been fairly standard. Regarding data
reduction, however, two "camps" have emerged
somewhat recently. For reasons cited above,
elemental abundances are usually presented in semiquantitative ratio form including temrnary diagrams
(e.g., Best 1984; Jack and Carmichael 1969;
Shackley 1988; Stross et al. 1968) or scatter plots
(see fig. 12.3, this study; Jack and Carmichael 1969).
Hughes (1986, 1988a, b) suggests that simple
bivariate plots of zirconium and rubidium in ppm are
sufficient for distinguishing raw material and
determining artifact provenance in some regions.
However, Bouey (1991) demonstrated that multiple
analyses of the same specimens produced "widely
divergent determinations in ppm concentrations,"
while ratio level presentation reduces this error
significantly (cf. Jack and Carmichael 1969; Sheets
et al. 1990). Hughes (1986) and Walter (1990) have
applied statistical clustering programs to define
geochemical groups. Multivariate statistics are quite
appropriate for some problems, however their use
eliminates any consideration of the elemental
abundances (either in ppm or at a nominal level) as
having any geological value. For example, on west
Moloka'i, Hawaiian Islands, relatively high values of
Y signal one particular quarry (David Clague,
personal communication, 1989) and, simple bivariate
plots of silica and total alkalis can be very informative for determining sources and characterizing
geochemical groups. Statistical clustering techniques should only be used after the data have been
examined for geological information that can inform
on the "provenance environment," that is, the
geology of the suspected interaction sphere. Conversely, groups created by statistical manipulation
are an end in themselves.
RESULTS
The results of the fused disk XRF analysis of
sixteen source rocks from Tatagamatau, Mako
Ridge, and Fa'ala'aga are presented in tables 12.4-6
and summarized in table 12.7. According to a recent
comprehensive and international evaluation of the
systematics of igneous rocks (Le Maitre 1989), the
Tatagamatau and Mako Ridge rocks are classified
.125T
0
.115
.105
.095
.085
CO,
Tatagamatau Fa'ala'aga (x)
subsource
.075
A/
0
..0
D
f~-
r
* A
.065
0
0
.055
A
0
*
Mako (W)
.045
I
.
-I
.045
.055
.065
.075
.085
Y/Sr
Figure 12.3
A wide dispersion of data points results
when ratios of rubidium, strontium,
and yttrium are used to assign specimens to possible sources.
chemically as hawaiites and the Fa'ala'aga rock as
basalt. Rock names were assigned by plotting silica
values between total alkalis. Between sources,
marked differences are found with the oxides,
A1203, FeO, CaO, and P205, and most trace elements. The samples from the Tatagamatau source
reveal important intra-source variability between
Areas 1 and 3. Oxide values for AI203, MnO,
Na20, and P205 demonstrate marked differences as
well as the trace elements Ni, Cr, and Cu. Until
additional samples collected from several areas of
the 110-acre Tatagamatau quaTrry complex are
analyzed and the geochemical variability of source
rock is understood in greater detail, Ni and Cr may
well be significant trace elements for demonstrating
intra-source variability; that is, at least between
Areas I and 3.
176
The To'aga Site
U
_
t- Ct. .l
NO tn
e
qt
qC
C-4-
Wo C4 Co o
r
"4
Rt
4 en
i
t-
-i
cn
ooo
%enao%0
%A
o4
r-
t-
t-
m
X
e-O
m W)0i
A
^
ml C14
1",
° ° ""8 tx
T
,r
§ A"'EDn
. 0\ w C2
q0
-
O-" t
__)
9 00
o
(C:
cne
A
co
X Xo? -4
o,
t-Oo
1¢
V-
00
Ir)
V"4
%o
t
o
8
7-
"4
A
(n
"4
"M4
0
"I0"-4.rr
0a
(
t_4
00
ON
Q£ N en %n
qt
-
."-4
en
fo
-
oi
00C)C
in C1I4 t O 00
r
q en
_en 4 i*b
CD
-
A
" 4n
-4 teIc- _)
m nx
w
x)
_- M x
"o
V0%
PC
w
_4 - ,
a
4E.
t- N 'v
O-
NA
xob
ICoooo
%At}
xNb£l_ or_
WI--,,
en
00
N -O Oso 0
CoS
co
en)"
"400
0 ~ie~
en
8v4
C
00%o
00
000%
o
"4
w
f~'% n-t
0 X 4)
%0
p.1.
as
N
e4
* -
qt *)
X
0%
00
8
N
wi
a-
W)
X
00 00
_4
wr
t-
t..:
w
V
4
t-)
qt
e VC'4
z
0%
A
O*
"
e4
0
O-> Bo
"I
00
Eco0
"-
8
W) t--
r
00 O
C
-g
S
m t-
mW
F mO
W)
-n -OO-~o
"-4
"-4
U)
00
l
m
CNqt
00
t
en
00
N en
NW^
o4
x ) o
o m Ch
"
en
t
_
o
0t
o 0 Rt
o
N 4 o \0 C4 t oW
0
t-X
t4 Wo tn ON O
"400C~~0~%00
.
_rT ^
.
_
C
o0
-j
_
on
o
en tl-W
J;
e
en N
N
O
El>
en
n
qt
0
cx
oor
;x 14o7A 0x
(sOut
;^xe
"
en
W-O
I
:a
r-
0% 0%
_4*j
00
%0t
%n
c4
c4
4
4
0D
0
_V- o
0%
0%N.
00I
"40
0
o
O-
to
Xou
>. _ t
-O
en
LL
A
t o
iu i izso
eq
ONI
w
o
t"9
a
ne
w
TO
4
*i;t
* s;
Chemical Characterization of Manu'a Adz Material
177
Table 12.5
Geochemistry of Mako Ridge Source Rock (Fused Disks)
Sample No.
89-1
89-2
89-3
89-4
Oxide (Weight %)
SiO2
46.48
A1203
18.22
TiO2
3.552
FeO*
12.10
MnO
0.21
CaO
6.65
4.44
MgO
K20
1.89
Na20
3.45
P205
0.988
Total
97.980
49.84
17.22
3.238
10.68
0.193
8.42
4.41
1.79
4.18
0.897
100.868
43.44
19.72
3.886
13.20
0.231
5.50
4.42
1.98
2.55
1.070
95.997
48.12
18.33
3.473
0.208
7.29
4.24
1.37
3.84
0.960
99.311
0
0
0
0
8
19
246
471
52
481
469
61
104.8
39
0
188
7
62
144
8
1
12
204
450
14
856
429
58
89.3
35
Element (PPM)
Ni
0
Cr
3
Sc
23
V
Ba
Rb
Sr
Zr
Y
Nb
Ga
Cu
Zn
Pb
La
Ce
Th
228
433
46
660
433
13
206
421
39
883
397
97
96.3
36
0
163
7
86
141
7
82.7
30
2
154
3
68
121
7
46
11.48
0
167
6
87
161
6
Range
Mean
43.44-49.84
17.22-19.72
3.238-3.886
10.68-13.20
0.193-0.210
5.50-7.29
4.24-4.44
1.37-1.98
2.55-4.18
0.897-1.070
46.97
18.35
3.537
11.87
0.211
6.966
4.38
1.76
3.51
0.979
98.533
0
0-8
12-23
204-246
421-471
14-52
481-883
397-469
46-97
82.7-104.8
0
30-39
0-2
154-188
3-7
62-87
121-161
6-8
3
16.75
221
443.75
37.75
720
432
65.5
93.275
35
0.5
168
5.75
75.75
141.75
7
*
total iron
Analyst: Dr. Peter R. Hooper, Dept. of Geology, Washington State University, 16-17 October 1989.
EDXRF analysis of Tatagamatau, Mako Ridge,
and Fa'ala'aga source rock are presented in tables
12.8-10; mean and range for sources are summarized
in table 12.11. Elemental values are presented with
one standard deviation which represents the uncertainty of counting statistics at 300 seconds livetime.
A comparison of EDXRF and fused disk samples for
source means and the Tatagamatau source envelopes
is illustrated in figure 12.2. Zr/Sr values are very
comparable, whereas Nb/Sr ratios are higher for
fused disk values. This probably corresponds to the
greater detection efficiency (machine sensitivity) for
niobium by the Rigaku spectrometer (Hooper and
Johnson 1989); or it could be a sample thickness
problem since niobium x-rays are excited as deep as
5-7 mm into the specimen (David Clague, written
178
The To'aga Site
Table 12.6
Geochemistry of Fa'ala'aga Source Rock (Fused Disks)
Sample No.
89-14
Range
Mean
89-15
89-16
0.674
99.808
46.69
13.79
5.426
12.93
0.178
10.56
4.8
1.73
3.13
0.684
99.918
46.79
13.80
5.335
13.06
0.174
10.75
4.76
1.60
2.97
0.627
99.866
46.69-46.89
13.68-13.80
5.335-5.426
12.93-13.06
0.172-0.178
10.54-10.75
4.72-4.80
1.60-1.73
2.97-3.13
0.627-0.684
46.79
13.76
5.391
13.01
0.175
10.62
3.16
1.67
3.03
0.662
98.268
Element (PPM)
Ni
57
24
Cr
21
Sc
V
344
Ba
300
Rb
45
644
Sr
Zr
323
34
Y
Nb
68
28
Ga
104
Cu
Zn
128
5
Pb
La
40
Ce
115
4
ITh
88
25
25
357
305
46
642
330
36
71.1
30
117
138
7
45
108
6
72
38
19
359
307
42
651
313
35
66.6
29
101
129
6
31
104
3
57-88
24-38
19-25
344-359
300-307
42-46
642-651
313-330
34-36
66.6-71.1
28-30
101-117
128-138
5-7
31-45
104-115
3-6
72.3
29
21.7
353.3
304
44.3
645.7
322
35
68.57
29
107.3
131.7
6
38.7
109
4.3
Oxide (Weight %)
46.89
SiO2
A1203
TiO2
FeO*
MnO
CaO
MgO
K20
Na20
P205
Total
*
13.68
5.412
13.04
0.172
10.54
4.72
1.68
3.00
total iron
Analyst: Dr. Peter R. Hooper, Dept. of Geology, Washington State University, 16-17
October 1989.
communication, 1990). The larger source envelope
for EDXRF may relate to the greater variability in
specimen surface.
After selecting elements for analysis (based on
analytical precision, accuracy, and sufficient elemental concentrations), assigning artifacts to source was
facilitated initially by trial and error. Figure 12.3
illustrates a wide dispersion of data points with most
artifacts plotting outside the source envelope when
Rb/Sr and Y/Sr ratios are used. Figure 12.4, however, using ratios of Zr/Sr and Nb/Sr, is much mome
useful for assigning artifacts to the Tatagamatau
source. The source envelope is delimited by taking
into account the variability in analytical precision.
Chemical Characterization of Manu'a Adz Material
179
Table 12.7
Geochemistry of Fa'ala'aga, Mako Ridge,
and Tatagamatau Source Rock (Fused Disks)
Fa'ala'aga (n=3)
Range
Mean
Mako Ridge (n=4)
Range
Mean
Tatagamatau (n=8)
Tatagamatau
sub-source
Range
Mean
(n=1)
50.02
15.45
3.413
12.68
0.177
7.63
4.69
1.58
4.02
0.800
49.32
13.86
3.649
12.40
0.167
7.80
7.06
1.79
3.45
0.710
Oxide (Weight %)
SiO2
A1203
46.69-46.89
13.68-13.80
TiO2
5.335-5.426
FeO*
MnO
CaO
12.93-13.06
0.172-0.178
10.54-10.75
4.72-4.80
1.60-1.73
2.97-3.13
0.627-0.684
MgO
K20
Na20
P205
Element (PPM)
Ni
57-88
Cr
24-38
Sc
19-25
V
344-359
Ba
300-307
Rb
42-46
Sr
642-651
Zr
313-330
Y
34-36
Nb
66.6-71.1
Ga
28-30
Cu
101-117
Zn
128-138
Pb
5-7
La
31-45
Ce
104-115
Th
3-6
46.79
13.76
5.391
13.01
0.175
10.62
3.16
1.67
3.03
0.662
43.44-49.84 46.97
17.22-19.72
3.238-3.886
10.68-13.20
0.193-0.210
5.50-7.29
4.244.44
1.37-1.98
2.554.18
0.897-1.070
18.35
3.537
11.87
0.211
6.966
4.38
1.76
3.51
0.979
49.56-50.56
15.35-15.62
3.385-3.467
12.34-12.90
0.173-0.179
7.56-7.70
4.55-4.77
1.54-1.62
3.95-4.09
0.783-0.814
0
0
0
0
-,,.--J
694-708
348-362
45-49
52.4-56.0
27-31
0-9
165-178
4-8
13-50
88-110
3-7
1.1
18.8
207.8
290.5
42.9
702.9
355.6
47.6
54.5
29.4
1.5
171.3
6.1
35.3
100.3
5.25
72.3
645.7
322
35
68.57
29
107.3
131.7
6
38.7
109
4.3
1t-'.JL
1 .0O
481-883
720
432
65.5
93.3
35
0.5
168
5.8
75.8
141.8
7
397469
46-97
82.7-104.8
30-39
0-2
154-188
3-7
62-87
121-161
6-8
173
181
27
233
331
50
683
325
37
47.9
26
31
153
5
43
99
2
* =
total iron
Analyst: Dr. Peter R. Hooper, Dept. of Geology, Washington State University, 16-17 October 1989.
Nine artifacts fall within the envelope and four
others are very close. Taken together, fifty percent
of the adzes and other artifacts with one or more
polished surfaces (adz flakes) can be assigned to the
Tatagamatau quarry on Tutuila. Weathered surfaces
of these specimens were gray (2.5Y 5/0; 2.5YR 5/0;
7.5YR 5/0), dark gray (2.5Y 4/0), to very dark gray
(7.5R 3/0; 2.5Y 3/i, 7.5Y 3/0). Freshly broken
surfaces on two specimens were very dark gray
(2.5Y 3/0; 7.5Y 3/0).
Additional samples from this quarry will
undoubtedly define a much larger source envelope
since the Tatagamatau sub-source is well outside the
limits of the eight samples used to define the geochemical dimensions of the quarry. This underscores the need to collect sufficient samples to define
the geochemical variability of adz quarry sources.
The diagonal line in figure 12.4 separates
+
ur
*i
6
+
+1
The To'aga Site
180
0
I
I
eq
M en.ON en.en
M It
tn.C.
M
-4
M
In C4
r
go
mg
12,
WI
V-4
en
67,e
C4
.s
V-48
.
_- en x
~o
,
C14
m
V-
-4 qt
n en CD t -4 Q
4
%
ent
en ,
t- en
C1
V-4
66
en en %n 0
e
etiM
t-
0
9Cs +1M
t.o
1-
te
+1
~o
0 in
00 ^o 4 e
en t^
.-
n7
-4xv
V
oN
ON x -
4en
ot4
o, C4
+1
(:
m
o"
g}
+l
,t°
+l
ON tn
eq
+l. +l
+
%n W r-
+
+1
WI)x
6 od
eneW)
in in
~oC'
+l '
q
CDC
(
q
0 t- m
(D t-
>o t*4 F_
C. +D
w
C. :;V-
IC
.6
,4
-H
+1
m
'H
+l t+l 00
C%
m
o
V-4
C114
V
O
00
6
C-4
Go
Cf oo
+1
s0 0
c>WqN t t-
o
Is
v-
+1
+H
w'
*o
e4
TNO
t- V4 en oo1
ql -V
0
:aco [..
F-4
eq.
WI,
C1
.2
00 ,H +l +l
r eq
oq.
'n:.H+~
t Mt 0e
F* a
+l
en +1
06
C>r-
em
ok
+l +l
't00 ,N t..
m ° C4 +
oo:~ cr
I00
no4
en
,--.
*
0
o,,oen
-
cr
W-
W)
6
+1
t-: o6
T-4
_
,t
>
+l
0
0
O-4
o ot
C>W
-4. cr) -t o) o o Q
M
+
*
+l;+It+l
I1
0 r- 1*
er
r.8°H
en cli
*
-H
+.l ,
+l
qq
c5s
+
cl
os t
*1
W)
V)
in WIt-
Q M
o m o m;
.1 +
+1oo1
r-
C> -
..
WI
O'
0
n ON
--
t_ _
i
*
0 It 0 0 14t
+1
6
+1
00
W)
oP-4
,N.I
0o/
+I
tn
Af
·0
en
o
It=,,~
ko 00 C1
ot
A
m
+lv"
.H
j"
X-
C It
C114
0
p
-2IL,20=
Vt
.H+I
._
C. q1t
C;
X=t,
qqtC4 C1
c+1 kn
o V)
t
14
*
(I
+1
cri
00
~=,=~
+1
10
0
in tn
I
Chemical Characterization of Manu'a Adz Material
O XD tn CN 00 W N 00 00 e
0 ren rren C
.4
It
m
CN
b E
V-
en ,q
N
t-
O~N
D
N-v00
"-
tmO
m
0\NXN0
M
--
N
6 o x0
~o
00
V0 OQ-
181
*
ofi
a uN
'r(..if0
tno
ur 0%
F-~t
,it re
x- N
"-
-
e. Ce i
0_
'R
r-W
N
I'0
-
06t
%
er-
(
Oe
--
_4
-O
6
,^
0
0 r-
le
U,
co
._
41
'
+00
U
414444cq
N- 44414
0
0; -o
N t
n
Cf) CN
m
41
i
ON
W-
V
Ch
o
r-
eni
Wo<
i
X 't W 'o00
c
_ +I - 't
+1 + vo +1 +1
_- t_ O
C1 _~ t-
N
V-4
en..4
C_ te
m )N
00
Q
+
'H
_.
c'4
V-
.H .Hooi
oof.
~o r,. . '0
I-
c~
00
0% c~
Is
am'
O9
4~
0
c~
o .,t cn
en
0
t.4
-v"
et)o
-4
. 1414m
- X00 o 00 O
t l("
t"'4v c-4
4
14141
414n
.H.H
0H
4
41
-
/00
*
+4 H00
Sf
Co-
0
m
z
Oa
ONOO
ON
d+
W)
01 41~
Eas
,; fi
·
q
*
rt
en
-cr
4-141 tn
W~
.H
V-
>
8'Hc,,
ON aN
0
)
*
0
H .H
en m -'0
414
H 'H
00'
Vt-
t- -- q ~t
r 4 004s
M
It
IR
0C0000
o.0
+141
41
Cx00
i
en
c
m
so
t- 00
"_
-
-
N
ON et
qq
11-
oN
00
cn W)
0%
-_
ON
*
00
(Ni
h
04
41
(Ni4VNI
H .H
C'4 v'~
x-
*
N e
"i4
cV'l cN
I.H
*
(14"t
r
*
cf CN
_ ) It M 0
N
414+1
r~
4141414141414
'0 o v * - wi
-r
x
V
At
o
0
41
cs
0
(4
v
r-
It
/3
o.
.7..
P
v. 2
w z m 1:
(n
t¢ >.
z
0u
9
u
u
;zQtaC
<
182
The To'aga Site
70.
Tatagamatau source rock from adz material deriving
.
00.
8|
-
()
/ ~~~A
.
·0 *c*
from the dike sources of Fa'ala'aga and Mako Ridge
on Ofu and from all unmodified flakes. The source
for the flakes was probably the scree slope below the
Mako M
Tatagamatau
X_ ss
subsource.D
":
A
N
**A Fa'ala'aga R)
.
*|~~~~~~~~~ subsor:e * *
40
Fa'ala'aga dike swarm. There are indeed several
flakes that plot close to this source mean. The dike
*source rock and flakes recovered from the site are all
medium- to coarse-grained, and eight specimens
(61.5% of the total flakes) have one or more natu-
dike rock
*.0~~~~~~~ ~rally flat surfaces characteristic of dike rocks.
50 60 ~70 80
~~Nb/Sr
1~
margin is smoothed from use. These medium~distal
to coarse-grained rocks were probably not conducive
X
Figure 12A.4
Several flakes, however, exhibit retouching and one
Anifas and source rocks plotted by ratios of
zirconiu, strontium, and niobium resulting
in many specimns ploing widin the
Tatagamatau source envelope.
to adz manufacture, and it is perhaps noteworthy that
all fine-grained rocks were manufactured into, or are
Table 12.10
Geochemistry of Fa'ala'aga Source Rock (Whole Specimen)
Sample No.
(PPM)
Oxide
TiO2
FeO
MnO
Element
Ni
Ba
Rb
Sr
Zr
Y
Nb
Ga
Cu
Zn
Pb
La
Ce
Nd
Cs
Fr
Th
Co
As
89-14
Range
Mean
89-15
89-16
50476.8 ± 358
50757.1 ± 366.8
168791.1 ± 1153.7 171686.3 ± 1191.4
2391.4 ± 44.6
2396 ± 45.5
47928.1 ± 334
156480.4 ± 1047.1
1978.4 ± 40
47928.1-50757.1
156480.4-171686.3
1978.4-2396
49720.7
165652.6
2255.3
89.7 ± 7.1
233.3 ± 6.1
42.9 ± 2.4
704.8 ± 6.3
341.7 ± 5.2
39.1 ± 2.7
52.3 ± 3.6
33.9 ± 2.5
134.4 ± 5.7
115.9 ± 4.6
7.6 ± 1.8
26.8 ± 3.2
91.2 ± 3.9
34.4 ± 3.9
0
9.2 ± 3.5
0
0
1.6 + 0.4
84.6-103.2
230.3-234.4
92.5
232.67
45.57
690.57
354.43
38.4
54.7
29.9
135
132.2
7.33
35.53
80.9
39.63
0
8.47
0
0
1.7
84.6 ± 7.9
234.4 ± 6.1
45.4 ± 2.4
691.6 ± 6.2
362.4 ± 5.3
40 ± 2.8
51.8 ± 3.7
25.7 ± 2.7
135.9 ± 6.5
143.1 ± 5.5
7 ± 1.8
40.1 ± 3.3
70.2 ± 3.9
43.7 ± 3.9
0
8.3 ± 3.5
0
0
1.7 + 0.5
103.2 ± 8.1
230.3 ± 7.1
48.4 ± 2.6
675.3 ± 6.6
359.2 ± 5.7
36.1 ± 3
60 ± 4
30.1 ± 2.6
134.7 ± 6.4
137.6 ± 5.4
7.4 ± 2
39t.7 ± 3.9
81.3 ± 4.5
40.8 ± 4.4
0
7.9 ± 4
0
0
1.8 ± 0.4
Analyst: Marshall Weisler, September 1989.
42.9-48A.4
675.3-704.8
341.7-362.4
36.1-40
51.8-60
25.7-33.9
134.4-135.9
115.9-143.1
7-7.6
26.8-40.1
70.2-91.2
34.4-43.7
0
7.9-9.2
0
0
1.6-1.8
Chemical Characterization of Manus'a Adz Material
183
Table 12.11
Geochemistry of Fa'ala'aga, Mako Ridge,
and Tatagamatau Source Rock (Whole Specimen)
Mako Ridge (n=4)
Fa'ala'aga (n=3)
(PPM)
Oxide
i02*
FeO*
MnO*
Element
Ni
Ba
Rb
Sr
Zr
Y
Nb
Ga
Cu
Zn
Pb
La
Ce
Nd
Cs
Fr
Th
Co
As
*x
Mean
Range
Mean
Range
Tatagamatau (n=8)
Range
Mean
(n=l)
3A.414
3.033-3.393
13.65-14.78
0.238-0.285
3.147
14.33
0.258
3.160-3.490
14.60-16.85
0.206-0.243
3.320
16.00
0.230
nd
nd
340.0
40.4
791.9
477.7
66.1
71.2
33.0
16.2
168.7
7.8
57.3
121.8
68.8
0.6
11.6
3.5
nd
1.6
nd
209.2-262.4
36.4-46.0
718.1-769.7
376.4-409.6
47.1-53.1
41.0-47.3
27.3-36.8
19.1-95.3
159.0-191.1
0.0-13.1
28.3-41.4
70.5-84.5
34.1-52.3
nd
0.0-19.4
0.0-13.1
nd
0.8-2.1
nd
226.7
40.7
750.3
395.8
49.7
44.4
30.8
42.0
177.7
8.0
33.9
75.7
45.1
nd
8.9
1.6
nd
1.3
4.793-5.076
15.65-17.17
0.198-0.240
4.972
16.57
0.226
84.6-103.3
230.3-234.4
42.9-48.4
675.3-704.8
341.7-362.4
36.1-40.0
51.8-60.0
25.7-33.9
134.4-135.9
115.9-143.1
7.0-7.6
26.8-47.6
70.2-81.3
34.443.7
92.5
232.6
45.6
690.6
354.4
38.4
54.7
29.9
135.0
132.2
7.3
38.0
77.6
39.6
nd
nd
7.9-9.2
nd
nd
1.6-1.8
8.5
nd
327.1-354.2
14.6-52.2
551.6-959.7
466.1-497.1
48.2-94.8
66.4-75.3
32.1-33.9
8.3-26.4
161.6-178.3
5.9-9.8
47.2-66.5
115.2-131.3
60.2-75.3
0.0-2.4
10.1-15.1
0.0-13.8
nd
1.7
1.2-2.2
nd
Tatagamatau
sub-source
15.55
0.220
188.0
259.2
56.5
738.6
357.0
38.8
32.7
28.3
61.2
149.2
nd
33.1
75.9
38.3
nd
nd
nd
nd
1.5
1i4
Analyst: Marshall Weisler, September 1989.
the by-products of, adz production. Geochemical
data for all artifacts are presented in table 12.12.
N'me artifacts assigned to the Tatagamatau adz
quarry complex on Tutuila are from three archaeological sites on Tau Island and two localities within
the large To'aga coastal habitation area (see table
12.1). Unformmnately, only three artifacts are from
excavated contexts, and the remaining are surface
finds. All stylistically diagnostic specimens are
quadrangular-sectioned adzes or fragments dating to
the late prehistoric penriod (Green 1974; Green and
Davidson 1969). By connecting the six sites with
adz material that originated from the Tatagamatau
quarry, we can document several nodes of an
interaction sphere between the islands of Tutuila,
Ta'u, and Ofu. These.results seem quite promising
for additional provenance studies of the kind employed here.
184
The To'aga Site
Table 12.12
Geochemistry of To'aga and Ofu Island Basalt Artifacts (Whole Specimen)
Trace Element Concentrations (PPM)
Artifact
Number
Fiti'uta 1
11-1-S-4
11-2-S-1
11-2-S-2
11-2-S-3
11-2-S-8
11-2-9
11-52-S-9
13-S-3
13-S-5
13-S-6
13-S-9
13-1-S-36
13-1-3-1-1
13-1-9-2-4
13-1-13-87
13-1-16-1-84
13-1-16-11
13-1-16-11-81
13-1-16-12-83
13-1-20-4-118
13-1-20-5-121
13-1-20-5-122
13-1-20-5-123
13-1-20-6-139
13-1-20-6-140
13-1-20-6-141
13-1-20-7-143
13-1-20-9-147
13-1-21-7-151
13-1-22-2-171
13-1-22-2-173
13-1-22-2-174
13-1-23-7-194
13-1-27-5-39
13-1-27-5-40
13-1-27-5-41
13-1-27-6-42
Lab
Number
89-22
89-25
89-26
89-31
89-30
89-28
90-30
90-33
89-18
89-34
89-33
89-32
89-19
89-17
90-22
90-23
90-42
90-46
9045
90-38
90-27
90-37
9041
90-34
9048
90-29
90-36
9040
9043
9049
90-44
90-25
90-24
90-26
90-31
90-39
9047
90-35
Pb
11.3 :e 1.3
11.1 ± 2
11.0 ± 1.6
7.7 + 1.7
11.1 ± 1.7
0.0
10.4 ± 2.1
8.9 ± 1.8
6.5 ± 1.6
11.7 ± 1.8
6.6 ± 2
7.6 ± 2
5.3 ± 1.9
9.2 ± 1.6
0.0
12.0 ± 2.1
8.3 + 1.8
9.6 ± 2
9.6 + 2.2
9.7 ± 1.8
6.9 ± 1.7
10.9 ± 2.1
0.0
9.1 ± 1.7
8.0 ± 1.8
7.4 ± 1.6
11.8 ± 2
9.1 ± 1.7
11.4 ± 1.9
8.7 ±t 2
8.8 ± 1.7
8.7 ± 1.9
12.8 ± 2.4
7.6 ± 1.9
8.6 ± 1.9
7.6 ± 1.6
7.1 ± 1.9
9.1 ± 1.8
Th
0.0
0.0
0.0
0.0
0.0
0.0
0.0
0.0
0.0
0.0
0.0
0.0
0.0
0.0
0.0
0.0
0.0
0.0
0.0
0.0
0.0
0.0
0.0
0.0
0.0
0.0
0.0
0.0
0.0
0.0
0.0
0.0
0.0
0.0
0.0
0.0
0.0
0.0
Rb
Sr
Y
Zr
Nb
65.4 ± 2.4
68.1 ± 6
52.4 ± 2.2
49.0 ± 2.2
50.5 ± 2.3
30.2 ± 2.1
57.8 ± 2.4
59.3 ± 2.5
48.7 ± 2.3
49.0 ±t 2.2
43.0 ± 2.3
46.3 ± 2.4
38.9 ± 2.4
58.9 ± 2.4
48.9 ± 2.5
26.4 ± 2.2
31.5 ± 2.3
32.5 ± 2.2
40.8 ± 2.6
37.3 ± 2.3
48.5 + 2.3
33.5 ±t 2.3
38.7 ± 2.4
37.2 ± 2.2
40.1 ± 2.4
46.4 ± 2.1
31.1 ±t 2.3
35.6 + 2.2
832.5 ± 6.3
724.0 ± 6
812.8 ± 6.1
782.3 ± 6
816.8 ± 6.2
578.0 ± 5.3
815.7 ± 6.4
750.7 ± 6.2
823.3 ± 6.5
824.4 ± 6.1
772.1 ± 6.2
691.2 ± 6
635.8 ± 6.2
854.1 ± 6.5
894.5 ± 7.2
660.0 ± 5.8
545.5 ± 5.6
638.5 ± 5.8
856.3 ± 7.5
967.8 ± 7.3
772.8 + 6.2
691.7 ± 6
733.0 ± 6.7
700.9 + 5.9
739.8 ± 6.5
724.4 ± 5.5
758.6 ± 6.6
694.4 ± 6.2
969.8 ± 7.9
524.1 ± 5.5
786.8 ± 6.3
722.4 ± 7
712.3 ± 6.8
689.7 ± 5.6
793.6 ± 6.2
729.7 ± 5.7
676.3 + 5.7
703.6 ± 6.4
52.0 ±: 2.7
51.9 ± 2.7
49.6 ± 2.6
41.1 ± 2.6
40.3 ± 2.6
39.2 ± 2.6
51.1 ± 2.8
48.0 ± 2.8
43.7 ± 2.7
50.4 ± 2.6
44.0 ± 2.7
34.7 ± 2.7
39.1 ± 2.9
50.9 ± 2.7
49.0 ± 3
31.6 ± 2.6
33.6 ± 2.8
32.4 ± 2.6
35.3 + 3.1
34.7 + 2.6
40.8 ± 2.7
36.8 ± 2.6
29.3 ± 2.8
33.9 ± 2.6
33.2 ± 2.7
35.7 ± 2.4
37.0 + 2.8
27.9 ± 2.7
44.7 ± 3.1
24.4 ± 2.8
37.5 ± 2.7
45.2 ± 3.1
49.6 ± 3.1
47.1 ± 2.6
50.2 ± 2.7
50.0 ± 2.6
30.8 ± 2.5
40.7 ± 2.8
438.8 ± 5.2
449.1 ±t 5.4
397.6 ± 5
369.4 ± 4.9
422.8 ± 5.2
339.1 ± 4.8
427.4 ± 5.4
457.0 ± 5.5
406.9 ± 5.3
438.0 ± 5.1
413.1 ± 5.3
322.5 ± 5
365.1 ± 5.6
464.6 ± 5.5
527.1 ± 8.2
292.1 ± 4.8
329.7 ± 5.1
293.2 ± 4.8
333.8 ± 5.7
331.0 ± 5.2
386.0 ± 5.1
317.7 ± 4.9
321.3 ± 5.4
324.5 ± 4.9
335.7 ± 5.2
368.9 ± 4.6
342.1 ±t 5.3
314.2 + 5.1
401.7 ± 6
264.6 ± 4.9
385.8 ± 5.2
335.4 ± 5.7
361.9 ± 5.8
415.3 ± 5
455.4 ± 5.4
440.0 ± 5.1
298.1 ± 4.6
392.7 ± 5.6
50.5 ± 3.3
51.1 ± 3.4
49.4 ± 3.2
43.2 + 3.2
46.9 ± 3.3
36.7 ± 3.3
49.9 ± 3.5
53.1 + 3.5
44.3 ± 3.4
50.9 ± 3.3
43.9 ± 3.4
37.4 ± 3.4
53.0 ± 3.8
54.3 ± 3.4
64.5 ± 3.8
27.8 ± 3.3
34.1 ± 3.6
49.7 ± 3.5
53.0 ± 4
51.1 ± 3.5
44.4 + 3.4
48.3 ± 3.5
58.5 ± 3.8
56.6 ± 3.5
52.1 ± 3.7
36.0 ± 3
61.7 ± 3.8
52.0 ± 3.6
63.4 ±4
41.4 ± 3.7
49.6 ± 3.5
40.8 ± 3.9
43.6 ± 3.9
45.4 ± 3.3
53.7 ± 3.4
47.3 ± 3.2
48.9 ± 3.3
43.0 ± 3.6
69.0 ±
49.8 +
52.3 ±
46.1 ±
41.5 ±
48.2 ±
49.6 ±
50.5 ±
38.1 ±
43.4 ±
2.8
2.5
2.4
2.7
2.7
2.2
2.3
2.2
2.2
2.4
Analyst: Marshall Weisler, September 1989 to Febr,-uy 1990.
Chemical Characterization of Manula Adz Material
SUMMARY AND CONCLUSIONS
Defining the spatial and temporal dimensions of
intra- and inter-island communication is an important precursor for evaluating historical developments
on insular landscapes. Because Polynesia generally
lacks the "footprints of pottery" throughout sequences of most island groups, and large quantities
of obsidian are absent in this geologic province, finegrained basalt manufactured into adzes and widely
distributed must of necessity formnn an important basis
for tracking regional interaction. Knowledge of the
"provenance environment" or geologic features and
their chemical signatures are essential aspects for
identifying and predicting the occurrence of finegrained, stone-tool-quality basalt. Two material
characterization and provenance techniques have
been described and evaluated, and geochemical
analysis is argued to be the most profitable for longterm and regional-scale distributional studies as: (1)
the results are reproducible; (2) instrument operating
conditions can be reported in full facilitating comparison of regional databases; (3) identification of
elements is not subject to human error as with thinsection descriptions; (4) elemental abundances can
be reported with precision and accuracy values for
specimens and standards; and (5) geochemical
sampling locales on specimens more closely represent the population rather than petrographic thinsections which are limited by two-dimensional
surfaces. The efficacy of non-destructive x-ray
fluorescence spectroscopy has been demonstrated
and its continued use for distributional studies of
Polynesian adz material is warranted. Although the
technique, as applied here, is limited to elements that
can be analyzed with high precision and accuracy
(those in the mid-z range), future applications should
seek to expand these limits. A major benefit of
EDXRF is the wide range of specimen sizes that can
be analyzed without destruction. In this study,
artifacts ranged from as small as 2.1 grams up to
421.9 grams with lengths of 22 to 136 mm. Accommodating a wide range of specimen sizes can be
advantageous when analyzing assemblages with a
high proportion of small adz flakes, a common
situation with most collections of adz material. In
fact, with nearly 1000 m2 of archaeological excavations on the island of Moloka'i, small adz flakes
outnumber whole or stylistically diagnostic adzes by
more than 10 to 1. Moreover, most adzes are surface
185
finds whereas adz flakes are more often from
excavated contexts.
Application of the non-destructive, energydispersive technique to source matenrial and artifact
assemblages from several sites on Ta'u and Ofu
islands has permitted delineation of an interaction
sphere during late prehistory that links habitation
complexes on these islands to the large adz quarry
complex at Tatagamatau, Tutuila Island, some 100
kmn to the west. Local dike rock from Ofu Island is
chemically similar to most of the non-polished and
unground flakes. Geochemical analysis has shown
that these medium- to coarse-grained rocks were not
fashioned into adzes, but this material was restricted
to a few retouched, "awl-like" tools and one flake
with use-wear along a distal margin. The vast
majority, however, were unmodified.
EDXRF is not the answer to all provenance
studies, but the results reported here suggest that the
technique merits further use and refinement with adz
material from Polynesia and possibly throughout
Oceania (e.g., Weisler 1993; Weisler et al. In prep.).
As the ownership and preservation of archaeological
sites and museum collections become increasingly
the focus of controversy between the scientific and
native communities, discovering more about the past
without destroying the evidence itself, may help to
bridge our common goals.
ACKNOWLEDGMENTS
I thank David Clague (Director, Hawaii Volcanoes Observatory), Joachim Hampel (Department of
Geology, University of Califomrnia, Berkeley), M. S.
Shackley (Lowie Museum, University of Califomrnia,
Berkeley), Pat Kirch, and Roger Green for comments on an earlier version of this paper. I am
grateful to Joachim Hampel who was especially
helpful during my XRF analyses, and I thank Pat
Kirch and Terry Hunt for inviting me to participate
in the Manu'a archaeological project.
REFERENCES CITED
Best, S. 1984. Lakeba: The prehistory of a Fijian
island. Unpublished Ph.D. dissertation, Department of Anthropology, University of Auckland.
Best, S., H. Leach, and D. Witter. 1989. Report on
the second phase of fieldwork at the Tatagamatau site, American Samoa, July-August
186
The To'aga Site
1988. Report on file, Historic Preservation
Office, American Samoa.
Best, S., P. Sheppard, R. Green, and R. Parker 1992.
Necromancing the stone: Archaeologists and
adzes in Samoa. Journal of the Polynesian
Society 101:45-85.
Bice, D. 1980. XRF data storage/matrix correction
programs. Ms. on file, Department of Geology
and Geophysics, University of California,
Berkeley.
Bouey, P. 1991. Recognizing the limits of archaeological applications of non-destructive energydispersive x-ray fluorescence analysis of
obsidians. Materials Research Society Symposiwn Proceedings 185, Materials Issues in Art
and Archaeology ll. pp. 309-320. Pittsburgh.
Cox, K., J. Bell, and R. Pankhurst 1979. The Interpretation ofIgneous Rocks. London: George
Allen and Unwin.
Davidson, J. 1977. Westem Polynesia and Fiji:
Prehistoric contact, diffusion and differentiation
in adjacent archipelagos. World Archaeology
9:82-94.
.1978. Western Polynesia and Fiji: The archaeological evidence. Manlakind 11:383-90.
. 1979. Samoa and Tonga. IN J. Jennings, ed.,
The Prehistory of Polynesia, pp. 82-109.
Cambridge: Harvard University Press.
Dunnell, R 1971. Systematics in Prehistory. The
Free Press, New York.
Dye, T., M. Weisler, and M. Riford 1985. Adz
quarries on Moloka'i and Oahu, Hawaiian
Islands. Report on file, Historic Sites Section,
Department of Land and Natural Resources,
Honolulu, Hawaii.
Emory, K. 1933. Stone Remains in the Society
Islands. Bernice P. Bishop Museum Bulletin
116. Honolulu.
Flanagan, F. 1969. U. S. Geological Survey standards-11. First compilation of data for the new
U.S.G.S. rocks. Geochimica et Cosmochimnica
Acta 33:81-120.
Gemianique, C., and B. Briand 1985. XRF determination of Zr, Nb, Y, Sr, Rb, Zn, Pb, in fifteen
international geochemical reference samples.
Geostandards Newsletter 9:31-34.
Govindaraju, K. 1984. 1984 compilation of working
values and sample description for 170 international reference samples of mainly silicate
rocks and minerals. Geostandards Newsletter
Special Issue 8.
. 1989. 1989 compilation of working values and
sample description for 272 geostandards.
Geostandards Newsletter Special Issue 13:1113.
Green, R. 1974. Review of portable artifacts from
Westemrn Samoa. IN R Green and J. Davidson,
eds., Archaeology in Western Samoa Vol. II,
pp. 245-75. Bulletin of the Auckland Institute
and Museum 7.
Green, R., and J. Davidson, eds. 1969. Archaeology
in Western Samoa. Bulletin of the Auckland
Institute and Museum 6. Auckland, New
Zealand.
Hooper, P., and D. Johnson 1987. Major and trace
element analyses of rocks and minerals by
automatic x-ray spectrometry. Ms. on file,
Department of Geology, Washington State
University, Pullman.
Hughes, R. 1986. Diachronic variability in obsidian
procurement patterns in northeastern California
and southcentral Oregon. University ofCalifornia Publications in Anthropology 17. Berkeley:
University of California Press.
. 1988a. Archaeological significance of geochemical contrasts among southwestern New
Mexico obsidians. The Texas Journal of
Science 40:297-307.
-. 1988b. The Coso volcanic field reexamined:
Implications for obsidian sourcing and hydration dating research. Geoarchaeology 3:253-65.
Hunt, T., and P. Kirch 1988. An archaeological
survey of the Manu'a Islands, American
Samoa. Journal of the Polynesian Society
97:153-83.
Irwin, G. 1990. Human colonisation and change in
the remote Pacific. Current Anthropology
31:90-94.
Jack, R., and I. Carmichael 1969. The chemical
'fingerprinting' of acid volcanic rocks. Special
Report 100:17-32. California Division of Mines
and Geology, Sacramento.
Kirch, P. 1985. Feathered Gods and Fishhooks: An
Introduction to Hawaiian Archaeology and
Prehistory. Honolulu: University of Hawaii
Press.
. 1990. Regional variation and local style: A
neglected dimension in Hawaiian prehistory.
Chemical Characterization of Manu'a Adz Material
Pacific Studies 13:41-54.
Kirch, P. V., and R. Green 1987. History, phylogeny, and evolution in Polynesia. Current
Anthropology 28:431-56.
Kirch, P. V., T. Hunt, L. Nagaoka, and J. Tyler.
1990. An ancestral Polynesian occupation site
at To'aga, Ofu Island, American Samoa.
Archaeology in Oceania 25:1-15.
Leach, H., and D. Witter 1987. Tataga-matau
"rediscovered." New Zealand Journal of
Archaeology 9:33-54.
. 1990. Further investigations at the Tataga-matau
site, American Samoa. New Zealand Journal of
Archaeology 12:51-83.
Le Maitre, R. ed. 1989. A Classification ofIgneous
Rocks and Glossary of Termns, Recommendations of the International Union of Geological
Sciences Subcommission on the Systematic of
Igneous Rocks. Oxford: Blackwell Scientific
Publications.
Parkes, P. 1986. Current Scientific Techniques in
Archaeology. London: Croom Helm.
Parks, G., and T. Tieh 1966. Identifying the geographical source of artifact obsidian. Nature
211:289-90.
Prickett, N., ed. 1982. The First Thousand Years:
Regional Perspectives in New Zealand Archaeology. New Zealand Archaeological Association, Palmerston North, New Zealand.
Rolett, B. 1989. Hanamiai changing subsistence and
ecology in the prehistory ofTahuata
(Marquesas Islands, French Polynesia).
Unpublished Ph.D. dissertation, Yale University, New Haven.
Shackley, M. 1988. Sources of archaeological
obsidian in the Southwest: An archaeological,
petrological, and geochemical study. American
Antiquity 53:752-72.
Sheets, P., K. Hirth, F. Lange, F. Stross, F. Asaro,
and H. Michel 1990. Obsidian sources and
elemental analyses of artifacts in southemrn
Mesoamerica and the Northem Intermediate
Area. American Antiquity 55:144-58.
187
Slice, G., and F. McCoy 1968. The geology of the
Manu'a Islands, Samoa. Pacific Science
XXII:427-57.
Stross, F., J. Weaver, G. Wyld, R. Heizer, and J.
Graham 1968. Analysis of American obsidians
by x-ray fluorescence and neutron activation
analysis. Contributions of the University of
California Archaeological Research Facility
5:59-79. Berkeley.
Terrell, J. 1986. Prehistory in the Pacific Islands.
Cambridge: Cambridge University Press.
Walter, R. 1990. The southern Cook Islands in
eastern Polynesian prehistory. Unpublished
Ph.D. dissertation, University of Auckland,
New Zealand.
Walter, R., and W. Dickinson 1989. A ceramic sherd
from Ma'uke in the southern Cook Islands.
Journal of the Polynesian Society 98:465-70.
Weaver, J., and F. Stross 1965. Analysis by x-ray
fluorescence of some American obsidians.
Contributions of the University ofCalifornia
Archaeological Research Facility 1:89-93.
Berkeley.
Weisler, M. 1990. Sources and sourcing of volcanic
glass in Hawai'i: Implications for exchange
studies. Archaeology in Oceania 25:16-23.
. 1993. Long-distance interaction in prehistoric
Polynesia: Three case studies. Unpublished
Ph.D. dissertation,University of California,
Berkeley.
In press a. Provenance studies of Polynesian
adze material: A review and suggestions for
improving regional data bases. Asian Perspectives 32(1).
In press b. The settlement of marginal
Polynesia: New evidence from Henderson
Island. Journal of Field Archaeology 24(1).
Weisler, M., P. Kirch, and J. Endicott In press. The
Mata'are basalt source: Implications for
prehistoric interaction studies in the Cook
Islands. Journal of the Polynesian Society.
Weisler, M., J. Sinton, and Y. Sinoto In prep. Mfh
Polynesian basalt artifact geochemical database.
13
FAUNAL ASSEMBLAGES FROM
THE TO'AGA SITE
LISA NAGAOKA
THE CALCAREOUS SAND depositional environment at To'aga favored the preservation of
faunal matenial (table 13.1). Over 165 kilograms of
invertebrate remains were recovered, represening
more than forty families. The To'aga fish-bone
sample-the largest in Western Polynesia contains
2,196 identified bones across twenty-two taxa. Pig,
chicken, rat, marine mammal, urle, and bird
comprse the 687 bones of the non-fish vertebrate
sample. Each component of the faunal assemblage
is described in detail below.
The problem of recovery bias is addressed here
through the analysis of bulk sediment samples from
the To'aga excavation. Generally, the use of smallersized screens increases the size of the faunal sample
and the number of taxa recovered. The To'aga bulk
samples were sieved through different screen sizes to
determine the effects of screen size on the composition of the faunal assemblage.
Cuwent knowledge of Western Polynesian
subsistence practices is limited since few
zooarcheological studies have been conducted in the
region. In this context, the To'aga faunal assemblage
is important for adding new information to our
understanding of regional subsistence trends. The
long temporal sequence at To'aga allows for an
assessment of changing subsistence patterns.
Despite the small sample of Western Polynesian
sites, comparisons of the To'aga assemblage with
other regional faunal assemblages may yield information about subsistence patterns.
METHODS
The faunal remains were recovered by dryscreening all excavated earth (except the clayey
colluvial sediment) through 1/4" mesh. To determine the feasibility of screening the colluvium,
Layer I of Unit 20 was screened rough 1/4"
screens. Only one poody preserved Turbo shell was
recovered from the 0.8 m3 sieved. A decision was
made not to screen the colluvial layer in other units
because of the difficulty in dry-screening the matrix,
and because of the low density and poor preservation
of faunal and other material in this clayey deposit
(See chapter 7 for further discussion of the pH and
other aspects of the To'aga site sediments.) During
the 1987 field season, most of the shell recovered
was identified, weighed, and discarded in the field.
Voucher samples and the remaining unidentified
shell and bone were shipped back to the laboratory.
All faunal materials recovered from the 1989 field
season were washed in the field and returned to the
laboratory for identification and analysis.
Fish remains were identified to the family level
using reference collections from the Bishop Museum, and the personal collections of Patrick Kirch
(U.C. Beikeley), Melinda Allen (University of
190
The To'aga Site
Table 13.1
Summary of To'aga Site Faunal Remains
Faunal Class
Tota Shel (kg)
Identified Shell (kg)
Tota Non-fish (NISP)
Tota Fish (NISP)
Identified Fish (NISP)
Washington) and the author. Reference collections
from the Bishop Museum were used to identify rat,
dog, pig, manine mammal, and marine turtle. The
bird component was identified by David Steadman
of the New York State Museum (see chapter 14).
Although many Pacific faunal analysts use MNI
(minimum number of individuals) to quantify
vertebrate remains (e.g., Leach 1986; Anderson
1986; Green 1986), we use NISP (number of
identified specimens) for the To'aga vertebrate
assemblage. The problems with both measures have
been discussed extensively elsewhere (see Chaplin
1971; Grayson 1979, 1984; Payne 1972). NISP was
chosen here because the effects of aggregation make
MNI an inconsistent measure. Although the problem of interdependence affects NISP, the measure is
constant across aggregation units.
The invertebrate faunal component was identified using standard shell identification guides (Abbot
and Dance 1986; Hinton 1972; Eisenberg 1981).
Tentative identifications were confirmed using
reference collections at the Burike Museum in Seattle
and at the Bishop Museum in Honolulu. Invertebrate remains were quantified by weight. As with
MNI and NISP, use of weight has its drawbacks
because of variations in size and density of different
shell taxa. For example, Tridacna maxima, the giant
clam, has a very dense shell, and one large individual
may weigh more than 1 kg. On the other hand,
shells such as limpets (Patellidae) are very light, so
that many individuals may account for a small
Excavation Units
1-14
15-30
Total
50.291
50.291
118.367
115.669
168.658
165.960
322
365
687
3462
6062
9524
723
1473
2196
amount of weight. As a result, heavy shells may be
overrepresented and light shells undernepresented in
any sample. Post-depositional alterations such as
leaching and fossilization also may distort shell
weight
RESULTS
The To'aga faunal data are presented in three
categories: the vertebrate component, which is
subdivided into fish and nonfish, and the invertebrate
component Of the thirty excavation units, complete
data by stratigraphic layer from thre areal excavations (1987 Main Trench, Units 20/23, and Units 15/
29/30) are presented in the text for comparison.
These units were chosen to represent the site because
they comprise a larger sample than the individual
units. Data for the remaining excavation units are
presented only in the summary tables for the separate
faunal categories. The complete faunal data by
stratigraphic layers from all excavation units are
available from the author on request.
Fish Remains
The To'aga excavations yielded 2,196 identified
fish bones representing twenty taxa (tables 13.2-6).
Although this is the largest archaeological fish bone
assemblage from Western Polynesia, an average of
only 73 bones were identified for each excavation
unit. Acanthuridae (surgeonfish), Diodontidae
FawuaIAsbages
191
Table 13.2
Fish Fauna Recovered from the 1987 Excavations
(NISP)
Excavations Units
Taxa
1, 4-9
3
2
10
12
11
13
14
Total
-
4
9
328
80
--
1
13
67
2
--
2
15
67
1
6
--
1
5
--
1
3
--
--
6
--
--
--
2
--
--
2
38
32
28
--
--
--
--
--
1
17
1
--
--
--
--
2
-
--
--
--
7
1
--
--
1
Diodontidae
Holocentridae
289
64
1
Acanthuridae
48
47
24
22
24
15
9
5
9
8
5
3
2
2
1
1
--
--
4
--
--
1
1
--
--
Seffanidae
Scaridae
Carangidae
Balistidae
Muraenidae
Labridae
Ostaciidae
Lutjanidae
Aulostomidae
Congridae
Elasmobranchii
Lethrinidae
Belonidae
Kyphosidae
Sphyraenidae
Scombridae
Bothidae
TOTAL IDENTIFIED
UNIDENTIFIED
TOTAL
4
--
15
--
3
1
--
--
-
--
~
--
--
-
--
--
--
--
4
577
2003
16
2580
20
(spiny puffers), Holocentridae (squinelfish),
Serranidae (groupersfcods) and Scaridae (paotfish)
comprise appoximately 78% of the identified fish
remains (fig. 13.1). These taxa are usually the most
almiant across time and space at the To'aga site.
1
7
45
1
52
13
12
11
8
7
5
3
2
1
1
1
1
2
---
--
1
19
2
--
--
--
--
--
38
190
228
--
1
--
1
1
--
--
---
--
--
---
3
2
5
1
10
19
29
82
464
546
722
2739
3461
The stucture and composition of the Toaga
fish-bone data pwbably reflect a conbination of
methodological, environmental, and cultuml factors.
Methodological factors include recovely bias and
problems in identification and quantification. Bias
192
The To'aga Site
3ma
4
>! 8
oo
)
(71,
w
m xn
N
n
x
Ch
m
m
_N
O
_
N
_
o
_
o
_
o
N
-
:
C-
',
N
ct 4 U)
O
F
m
_
_xo x
%D on o
vo
It
N
m
2
m
~o o
vo
t
_4
_
I
mQ
7
_;
tn
Q2 1. t S:
ON
m
9
X0
m
e4
C4
e
N
V-4
V4
-_
_
0
_-
a _
1.
!4
-
oo
A
r)o 00
9-
V-
:
:
:
-
N
:
-
00
V-
00
W
0
e4
la A0
II
"t tn
6
e
_-
I
I
I
I
-
-
I
I
I
I
t
,m
A
"-4
II
C
CU
_-4
kn
Cie;IaC
en
;O.
-
I
I
tI
I
II
"I
:
II
:
I
I
:
I
I
I
I
:
I
:
I
I
-
a
-
:
!s)
Cu
0x
-
"-4
en
lqt
I
I
C4
ON
I.t
I
I
-4
(I
"14
00
0 )
00
.o
o
Wo 0
T-4
1
e>
-
C4
V-4
00
ar
-
-
V-4
a
"4
a
*;0
-
v
a
a
a
:
:
~
a
C
V-4
en
Cf)
dl~t
-
en
en
-
-
|
U 0~
--
4]
3V3
8
-4
-
0
;
"-4
tn
r
Faunal Assemblages
193
Table 13.4
Fish Fauna from the 1987 Main Trench (Units 1, 4-9)
Taxa
Diodontidae
Holocentridae
Acanthuridae
Serranidae
Scaridae
Balistidae
Carangidae
Muraenidae
Labridae
Lutjanidae
Aulostomidae
Ostraciidae
Congridae
Elasmobranchii
Lethrinidae
Belonidae
Kyphosidae
TOTAL IDENTIFIED
UNIDENTIFIED
TOTAL
HA-I
4
3
3
5
2
4
2
---
2
2
---
---
Layers
HA
HB
25
17
13
11
1
8
4
4
3
1
1
1
154
25
20
17
21
10
10
4
2
5
4
3
1
HC
Total
106
19
12
14
1
2
6
7
2
1
3
1
3
1
289
64
48
47
25
24
22
15
9
9
8
5
5
3
2
2
1
578
2003
2581
-----
1
2
---
---
2
---
1
1
---
---
---
-----
28
191
219
93
535
628
---
in the recovery process is shown to affect the sample
size and the taxa represented in the assemblage (see
"Bulk Samples" section). Problems in the identificaion process include the quality of the reference
collection, which can limit the accuracy of the
identifications and the number of taxa represented.
For the To'aga assemblage, several distinctive mouth
parts could not be identified using the reference
collection at hand. With a better reference collecton, subfamily identifications may also be possible.
Another methodological problem is the inclusion of "special bones" in the NISP count. A few
taxa are identified mainly by special bones that can
number up to 300 per individual, thus greatly
inflating he NISP count. This is especially true for
1
279
827
1106
178
450
628
Diodontidae, which can have more than 250 spines
per individual, and to a lesser extent for Ostraciidae,
Elasmobanchii, and Balistidae. Of the 923
Diodontidae bones identified at To'aga, 901 were
spines and only 22 were mouth parts, most being
concentrated in Unit 21 and in the 1987 main
excavation trench. If the Diodontidae spines are
removed from the NISP count, the ranking of
diodonts drops from one to thirteen, and the shape of
the graph changes (fig. 13.2). Although the presence
of this poisonous fish may seem odd, its remains are
common in middens across the Pacific (e.g., Allen
1990; Butler 1987; Masse 1989). Moreover, the fish
is still eaten by some modem Pacific populations
(Bagnis 1972, Masse 1986).
194
The To'aga Site
Table 13.5
Fish Fauna from Transect 5, Units 15/29/30
(NISP)
Layers
Taxa
H HA-1
Diodontidae
Lutjanidae
Labridae
9
9
11
3
3
3
6
Ostraciidae
---
Lethrinidae
Muraenidae
IIIB
9
8
4
HID Total
---
10
3
2
2
1
16
15
7
5
8
5
1
3
2
---
---
---
---
---
---
Carangidae
---
---
1
Mullidae
Balistidae
---
---
3
1
2
2
---
---
---
AcanthuridM
Serranidae
Holocentridae
Scaridae
Kyphosidae
Bothidae
Congridae
TOTAL IDENTIFIED
UNIDENTIFIED
TOTAL
---
1
---
1
---
1
-----
---
1
1
---
---
---
---
---
---
48
201
249
3
16
19
42
250
292
1
67
297
364
The structure of the To'aga fish-bone data also
may reflect natural distributions and abundances of
fish taxa. Most of the To'aga assemblage can be
classified as inshore fishes, although a few families
such as Serranidae, Lutjanidae, and Carangidae
cover a wide range of habitats. Reef ecosystems are
generally more diverse and have a higher pwductivity rate than open ocean environments; therefore, the
abundance of inshore versus pelagic fish may reflect
the natural diversity of the different environments.
Fishing strategies may also be reflected in the
35
32
23
18
15
10
9
4
4
3
2
2
2
1
1
1
160
764
924
fish-bone data. The fishhooks recovered from the
site (see Kirch, chapter 1 1) may have been used to
catch serranids, holocentrids, and lutjanids, but
probably not scarids or acanthunds which are more
likely to be caught by netting or spearing. A comparison of modem Samoan reef exploitation (Hill
1986) and the To'aga fish data shows that the most
abundant taxa in the archaeological assemblage can
be caught by several fishing techniques (table 13.7).
These taxa may have had more opportunity to be
caught than taxa for which only one stsegy was used.
FawalAssemblages
195
Table 13.6
Fish Fauna from Transect 9 (Units 20/23)
(NISP)
Layers
Taxa
Acanthuridae
Serranidae
Diodontidae
Muraenidae
Holocentridae
Scaridae
Labridae
Ostraciidae
Carangidae
Lutjanidae
Congridae
Aulostomidae
Lethrinidae
Balistidae
Scombridae
Elasmobranchii
Bothidae
Kyphosidae
Mullidae
TOTAL IDENTIFIED
UNIDENTIFIED
TOTAL
IIB
HIA
HIIB
uic
IV
Total
3
2
3
3
1
2
30
18
---
53
5
14
37
13
9
---
53
3
23
3
14
---
1
14
---
1
15
3
9
9
5
---
1
1
2
2
2
1
12
9
2
44
---
38
8
9
1
1
27
25
6
---
6
13
2
---
---
---
22
22
14
12
2
---
7
5
1
2
3
---
---
---
1
4
2
-
16
63
79
Non-Fish Vertebrate Remtains
The non-fish vertebrate sample of 687 bones is
small, averaging only 23 bones per excavation unit
(tables 13.8, 13.9). About half the sample consists of
Rattus exulans, the Pacific Rat, with nearly half the
rat bones coming from Layer II of the 1987 Main
-----
5
4
3
1
---
1
1
1
34
191
90
161
610
801
195
6
1
1
10
7
414
17
46
348
1294
504
63
1642
Trench. Bird bones were the second most abundant,
with 139 bones. Steadman
of
analysis
the 72 identified bird
bones presents
in chapteran14.
Fifty-six
marine turtle-bone fragments were scattered
throughout the excavations with one-third of the
sample concentrated in Layer IIIB of Unit 20, dating
to about 2900-2400 B.P. From this same time period,
196
The To'aga Site
Table 13.7
Modern Samoan Fishing Methods (after Hill 1986)
Taxa
Gleaning
Linefishing
Diving/
Spearing
Gillnetting
Throw-
netting
Holocentridae
Serranidae
Acanthuridae
Mullidae
Lutjanidae
Muraenidae
Letbrinidae
Carangidae
Scaridae
Mugilidae
Labridae
three-fourths of the marine mammal bones were
recovered from Layer HIC of Unit 15. Another
concentration of marine mammal bones was associated with the 'ili'ili paving found in Layers I and II of
Unit 22. No fruit bat bones (Pteropus sp.) were
identified from the site, although the fruit bat is
present on Ofu Island today.
Of the domesticated animals, only 1 pig (Sus
scrofa) tooth and 15 chicken (Gallus gallus) bones
were identified. The chicken bones are concentrated
in Layer IIIB of Unit 20123. Generally, the non-fish
vertebrate bones were very fragmented and difficult
to identify to species or even class. Thus, 44 bones
were placed in the "general vertebrate" category and
40 in the "general mammal" category. Many of the
bones placed in the mammal category may be either
pig or dog, but a distinction between the two could
not be made.
Invertebrate Remains
As is true for most Pacific island faunal assemblages, the invertebrate component dominated the
To'aga faunal assemblage. The densest concentra-
tions of shell midden were recovered from layers
that dated to two periods of time and contained eitlr
'ili'ili paving or features inteipreted as food preparation areas. For the period of 2500-1900 cal B.P., the
instances of concentrated midden are dispersed
across the site. Unit 28, Layer RC of Transect 5 had
a shell density of over 7.8 kg/n3. Along Transect 9,
about 27 kg of shell midden with a density of 11.9
kg/M3 were recovered from Layer III of Unit 20123
(table 13.10), along with one-third of the marine
turtle remains for the site and half the chicken bones.
The upper portion of this layer contained a large
earth oven. The extension of Layer IIn into Unit 21
contained over 13 kg of shell midden (8.2 kg/M3).
The densest midden in the 1987 Main Trench, in
Layers IIB and UC, also dated to this time period
(table 13.11).
In Transect 5, Layer II of Unit 15129/30, interpreted as a cookhouse activity area dating to the
period 1641-1477 cal B.P., contained nearly 12
kilograms of midden, a density of 7.4 kg/M3 (table
13.12). This layer is contemporaneous with Layer I
of Unit 16 of the same transect which was associated with a dispersed distribution of 'ili'iMi gravel and
Fauna! Assemblages
197
45-
40-
30-f .--
t
25- ------- f-
E
- ------ t
D~ 20f-
15-t-------tio-It-. .
I
t
5-
,
-----------------------------------------------------------------------------------
rl r.l ,mm
I I I II I
.
I
I
.
1
I
--
I
I
1
nv h. .
-
.
2
3
I
4
5
6
7
8
rl ,§ tl
owu Oman
.
9 10 11 12 13 14 15 16 17 is 19 20 21
22
I
Tuaa
Figure 13.1
Relative frequency of fish taxa from the To'aga site, including Diodontidae.
contained the densest concentration of midden in the
site (13.4 kg/m3). A midden with 7.9 kg/n3 density
was also present in Layer IIC of Unit 28, Transect 5.
Only a few families make up the majority of the
invertebrate assemblage. Over 76% of the 165
kilograms of identified shell consisted of three
families, Turbinidae, Trochidae, and Tridacnidae,
with Turbo setosus by far the most abundant species.
Besides the shell, more than 14 kilograms of slatepecil sea urchin (Heterocentronus mamilaus),
comprising over 8.5% of the invertebrates, were
recovered from the site. Most of the sea urchins
were concentrated in Units 20-24 along Transect 9;
about half were associated with the earth oven in
Layer III, Units 20/23.
The rank order of the invertebrate taxa varies
little across time and space. Turbinidae is by far the
major taxon in the assemblage with Echinoidea,
Trochidae, Tridacnidae, Conidae, Cypraeidae,
Muricidae, and Neritidae as secondary taxa. The
remaining thirty-seven taxa are minor components,
contributing less than 1% each to the assemblage.
This high diversity may reflect both cultural and
environmental factors. Food choice in foraging
often reflects the natural abundance and distribution
of resources. However, some of the most abundant
taxa in the assemblage, such as Turbinidae,
Tridacnidae, Echinoidea, and Conidae, were also
used as raw material in the manufacture of artifacts.
The abundance of these taxa therefore may reflect
these dual uses and species may have been selected
disproportionately to their natural distributions. A
comparison of natural and archaeological invertebrate distributions through modem marine survey
infonmation would be useful in sorting out the
influence of environment versus cultural effects on
the invertebrate taxa represented archaeologically at
To'aga.
ANALYSIS OF BULK SAMPLES
Archaeologists screen sediments in order to
increase comparability within and between sites by
systematically sampling the archaeological record.
The To'aga Site
198
OWN-01
zu-
f-
18-4
16-4 -----1--
144- --------i
ci) 12-It
10-fr
1-..
-
01) 8-
4-0 I........
."
.
..
.
.
..
6X::.......... : 1 : I : 1 1
1
"
-
I
-
I
.
.................................................
"'
-
..
.
-
-
I
-
-
I
-
-
I
-
I
-
I .-=.-.
C
2
3
4
5
6
7
8 9
10
11
12
13
14
15
16
17- 18- 19- 20 21
-
22
Taxa
Figure 132
Relative frequency of fish taxa from the Todaga site, excluding Diodontidae.
Recovery methods can greatly influence the kinds
and the amount of matenal retrieved from excavation (Grayson 1984). As with other sampling
techniques, the size of the screen used is determined
by the research problem. Pacific island archaeology
has been oriented toward the recovery of artifacts,
such as pottery, that can be readily recovered by 1/4"
screens. Unfortunately, the consistent use of 1/4"
screens does not always sufficiently sample other
classes of archaeological material such as smaller
faunal remains.
Experiments on the differential recovery of
faunal material show that screen size affects the
sample size and the number and kind of taxa represented (Thomas 1969; Casteel 1972; Butler 1987).
Larger screen sizes bias the sample toward taxa with
larger body sizes. The use of smaller screens
increases the sample size and retrieves smaller taxa
that would otherwise be lost though the larger
screens.
To determine how our recovery methods
influenced the composition of the To'aga faunal
assemblage, bulk samples of ca. 5 kilograms (ca. 500
cm3) was taken from Layer IIIA/B of Units 20123
and from Layer II of Unit 30. These layers were
chosen because they contained dense midden
concentrations. Thus the two samples may not
represent the site in general since the recovery rate
may be less for areas with a lower midden density.
The bulk samples were wet-sieved thrugh
window screen in the field to reduce the bulk of the
sediment for shipping. They were dry-screened in
the laboratory though nested -3 phi (8mm), -2 phi
(4mm), -I phi (2mm), and 0 phi (Imm) geological
sieves. The contents from each phi size were
separated into gross categories (rocks, coral, shell,
sea urchin, crab, and bone) and weighed (tables
13.13 and 13.14). The bone was then identified and
quantified using NISP.
The recovery of bone was affected by screen
size more than shell. Almost all the bone was
recovered in phi sizes -2 and smaller. The shell
recovered by screen sizes less than -3 phi was
difficult to identify. Furthennore, the weight of the
Fawual Assemblages
I
0
F-
0%
.o
fC
S
N1
Cl
6
c1
I
N
i tt
Stn
t-
%n
C
N
t
t
"<
I
I
It
I
:
>
-4
t
t
N-
I
I
I
I
00
'
C%
II
I
z
II
a
0o
I
-
13
iI
cd
_l
"14
-
Is
00
"
II
I
0
0)
Nw
t-
I
'
N
-4
e4
"4
'Cuo
A
oo
.00
C-
0
-
as
X
II
I
F*
0-
E
"4
00
I
"-4
I
II
I
en
"4
II
I
II
I0i
Al
C
Clq
I
II
I
I
II
I
I
I
II
I
II
0
II
I
I
II
II
I
9-4
V-4
%O
C4
t-
"4
II
I
4)
"4
C!
en
0ON
I
II
I
II
I
eq
II
I
ClCl
Cl
Cl4
(4
e4
cO,
Cu
Cl
go
en
eq
II
I
II
I
C4
II
I
II
I
"-4
iI
I
I
II
I
II
I
II
I
II
0
00
II
I
"-4
Cl1
"-4
ON
V-4
II
a
SrI
I
I
4)
N
40
*
"4
II
"-4
2
5-a,
I
cuoo
Cl4
Cl
1^
- a-
N8
1
i
Cu/
Il
I
m
_
Cl-
it
I
I
1
"-4
"
4.
'0
-4
Il
I
T-4
I~"
v
N
"-4
e
a
en
.
tN
F
_
N
00
C4
"4
0
A
-
:
00
Cl
V-4
-
WI
Cl
N
V
asd
N
so
"4
Co
0
co
cn
I
q
m
II~l
0
"_4
_-
Ft
Cl11
0
V)
"
-4
a
ON
"-4
N
_4
-4
F0
199
i
I
t1i,
I
i
Cl4
11
1
V-4
W
"
V"
4
01
1
V-
9-4
kn
C
"4
l41
I
I
I
en
.o
V4
"-4
"4
No
0%
Cl4
II
1
1 7
II
I
-4
V-4
L
00
t"
M
!F-o
m
A4 ! (.- !!
"I
s
CZ>
0,
c
%)
%
Q
-t
14
to
x
Cs*
_F-
bC
-
2 2
_
-s
m
2 0
,It
200
The Todaga Site
Table 13.9
Non-Fish Vertebrate Fauna from the To'aga Site
A. 1987 Main Trench, Units 1, 4-9
Layers
Taxa
HA-1
hA
HIB
HIC
III
Total
9
63
--
5
--
13
1
1
78
28
1
14
--
--
88
2
21
2
--
2
45
--
1
188
3
54
3
3
251
RaNus exulans
Mamnmal
Bird
Marine turtle
Vertebrate
TOTAL
--
14
113
--
--
1
--
B. Transect 5, Units 15/29/30
Layers
HIA-1
IIIO
HIC
HID
Total
2
22
12
1
--
--
--
--
-----2
--
18
2
2
1
--25
--
--
--
1
11
42
6
4
47
8
20
2
11
8
11
107
IIB
IIIA
Layers
HIB
HIC
IV
Total
5
2
4
5
20
2
--
--
--
7
4
18
--
--
2
3
1
--
--
--
--
34
25
3
34
4
8
15
22
2
85
II
Taxa
Rattus exulans
Mammal
Marine mammal
Gallus gallus
Bird
Marine turtle
Vertebrate
TOTAL
10
8
---
4
3
--
25
--
13
C. Transect 9, Units 20/23
Taxa
Rattus exulans
Mammal
Gallus gallus
Bird
Marine turtle
Vertebrate
TOTAL
--
1
6
---
12
--
2
1
2
11
Faunal Assemblages
Table 13.10
Invertebrate Fauna from Transect 9 (Units 20/23)
(weight in grams)
Layers
Taxa
GASTROPODA
Patellidae
Trochus maculatus
Trochus niloticus
Trochus spp.
Tectus pyramis
Turbo crassus
Turbo setosus
Turbo spp.
Turbo operculae
Astrea stellare
Lunella cinereus
Nerita albicilla
Nerita picea
Nerita plicata
Nerita polita
Nerita spp.
Neritina spp.
Tectarius grandinatus
Cerithium nodulosum
Cerithium spp.
Clypeomorus spp.
Strombus mutablis
Strombus spp.
Hipponix spp.
Cypraea annulus
Cypraea arabica
Cypraea caputserpentis
Cypraea mappa
Cypraea moneta
Cypraea tigris
Cypraea spp.
Natica spp.
Tonna spp.
Cassidae
Cymatium nicobarium
Cymnatium spp.
Bursa granularis
Bursa spp.
Drupa ricina
Drupa morum
Drupa spp.
Morula uva
Nassa spp.
Thais armigera
Thais tuberosa
(continued next page)
I
IUB
32.5
0.3
118.4
---
---
48.2
263.9
8.3
89.9
276.4
792.8
114.6
138.1
2.0
0.4
HLiA
---
235.8
1.2
3.9
21.1
932.9
1561.0
140.8
228.7
7.5
1.0
HIB
HIC
0.2
642.6
0.5
23.8
46.5
1734.6
3980.6
135.1
1632.0
34.1
9.1
0.4
394.6
IV
Total
0.9
1.2
---
4.7
---
---
500.2
2275.0
125.3
894.4
8.9
11.8
58.4
100.9
0.7
10.6
---
1.1
---
1.0
5.2
1.0
0.6
3.3
2.2
0.9
1.2
1.5
4.3
2.8
23.2
47.4
38.0
2.7
29.3
44.9
11.4
--
1.2
8.0
5.6
0.1
1.6
7.5
54.8
19.0
1.7
21.9
15.1
4.7
26.4
---
---
---
7.7
9.2
16.0
13.7
0.6
2.4
1.3
12.7
6.0
14.4
19.2
68.7
9.0
3.1
6.5
29.9
89.9
62.0
66.3
80.4
23.8
57.6
206.6
---
---
---
51.5
---
---
9.8
9.4
---
---
2.2
37.1
11.3
9.0
---
41.3
---
---
---
-------
10.9
141.9
10.9
2.0
473.2
0.5
0.5
6.6
30.2
7.5
---
50.6
---
---
11.0
---
---
11.0
1.0
---
---
3.3
---
2.5
4.2
70.1
17.1
0.6
---
2.4
---
15.9
---
---
13.1
---
---
1.1
7.6
63.3
107.9
56.9
0.1
1.6
68.5
78.6
4.1
33.0
74.3
22.0
52.0
19.2
101.6
171.7
175.4
6.3
3.6
5.9
---
1425.1
1.7
32.4
67.6
3550.7
8974.2
524.8
2993.7
52.5
22.3
18.0
8.1
67.8
---
4.0
19.6
10.2
12.9
6.9
25.4
---
16.9
91.6
29.6
---
40.3
5.9
1.4
1.2
0.8
48.0
-
---
---
2.5
12.8
89.7
70.9
19.4
6.9
51.8
1.2
16.9
238.5
37.7
201
202
The To'aga Site
Table 13.10 (continued)
Layers
Taxa
Th'a-is
spp.
Cantharus undosa
Nassarius spp.
Vasum ceramicum
Conus argus
Conus chaldeus
Conus eburneus
Conus cf. maculifera
Conus spp.
Bulla spp.
Dolabella spp.
Melampus spp.
Pythia spp.
PELEYCYPODA
Mytilidae
Isognomon spp.
Chama spp.
Chlamys spp.
Cardiidae
Periglypta reticulata
Tridacna maxima
Quidnipagus palatam
Scutarcopagia scobinata
Trapezium oblongum
Asaphis violaceus
Pinna spp.
I
IIB
LiA
fIB
HIC
---
6.3
27.0
0.3
12.6
46.9
5.1
2.9
22.4
43.4
10.9
1.6
40.9
---
---
---
---
---
-- - -1.2
2.3
20.3
17.1
16.7
6.7
14.5
38.2
13.9
19.8
1.3
---
---
0.5
---
27.4
0.2
---
---
0.4
1.6
110.1
1.4
0.8
11.9
---
---
---
---
20.0
---
6.7
0.7
23.9
---
4.8
---
---
---
2.2
15.1
14.9
4.7
17.9
10.6
9.0
---
3.4
81.1
---
55.2
3.2
5.7
28.0
6.8
1.9
1.6
7.2
1.8
---
1.1
---
---
---
---
283.5
18.1
5.1
13.7
12.5
234.4
70.3
75.4
---35.4
0.3
2.2
-------
---
----
147.1
18.3
5.5
21.3
3.3
---
IV
3A
-----
3.6
11.1
Total
113.7
36.2
11.5
117.4
38.2
16.2
41.3
0.5
348.0
23.1
12.0
66.8
21.2
52.5
2.3
8.3
16.9
14.9
0.8
205.4
31.8
29.8
1.8
2.6
0.2
7.6
34.1
1.8
5.5
-----
822.3
132.6
121.5
13.7
61.5
0.5
---
919.8
6.8
3052.9
30.7
3742.2
39.3
395.5
54.4
Unidentified
14.2
48.8
116.4
38.4
67.9
285.7
TOTAL (g)
672.5
2676.5
7281.7 13703.0
5875.7
256.7 30466.1
1.90
0.35
0.85
3.15
ECHINOIDEA
CRUSTACEA
VOLUME (m3)
DENSrrY (kg/m3)
25.0
0.40
18.20
0.95
14.24
0.90
6.53
0.30
0.86
8169.5
133.0
5.30
5.75
Faunal Assemblages
203
Table 13.11
Invertebrate Fauna from the 1987 Main Trench, Units 1, 4-9
(weight in grams)
Layers
Taxa
GASTROPODA
Haliotis ovina
Patellidae
Trochus maculatus
Trochus spp.
Turbo crassus
Turbo setosus
Turbo spp.
Turbo operculae
Astrea stellare
Astrea rhodostoma
Lunella cinereus
Nerita plicata
Nerita polita
Nerita spp.
Cerithium nodulosum
Cerithiidae
Strombus cf. maculatus
Strombus cf. mutablis
Strombus spp.
Hipponix conicus
Cypraea arabica
Cypraea annulus
Cypraea caputserpentis
Cypraea mappa
Cypraea noneta
Cypraea cf. tigris
Cypraea spp.
Policines spp.
Cymatium nicobarium
Cymatfidae
Tonnidae
Bursidae
Drupa grossolaria
Drupa ricina
Drupa spp.
Nassa spp.
Thais armigera
Thais spp.
Muricidae
Cantharus undosus
Nassarius spp.
Vasum ceramicum
Conidae
Bulla sp.
Melampus fasciatus
Melampidae
Pythia scarabeus
(continued next page)
IC
HA-I
HA
UB
UC
13.9
2.0
66.8 145.4 136.9
20.0
65.0
71.3
383.1 1274.5 667.7
973.1 4646.3 2850.6
235.0 313.1
189.5
872.3 1116.1 2045.2
--
---
---
4.3
7.6
---
---
---
3.2
0.3
6.7
20.0
---
---
2.0
15.0
1.2
9.2
9.0
---
---
11.0
---
3.0
---
---
---
---
1.0
11.6
2.6
0.9
7.9
65.4
20.0
34.9
60.0
2.2
15.3
25.6
---
1.0
0.4
53.4
5.9
25.2
69.8
1.1
---
1.5
15.2
---
276.0
3.7
24.1
4.2
19.5
---
---
2.3
1.6
13.7
6.5
5.9
1.4
48.8
70.0
183.0
2.7
---
---
---
---
---
0.9
79.2
1.6
0.5
---
---
---
---
---
---
8.1
4.5
8.7
16.8
43.4
166.5
2.7
65.0
1.4
1.0
1.3
14.9
2.4
15.9
31.6
24.3
9.7
255.4
0.7
6.8
2.0
17.2
16.1
1.7
13.2
3.9
5.0
---
---
185.8
32.0
18.1
8.4
3.3
0.5
4.1
---
---
4.6
13.9
3.1
406.3
162.3
2892.8
12.9
---
---
1.1
20.5
1.0
44.9
1.1
0.7
4.3
9.0
---
---
---
Total
5.0
---
---
IV
---
----
36.7
5.0
20.0 498.0
60.0 2072.3
--215.2
347.3
HI
11.5
1.1
0.5
15.0
9.3
14.5
1.4
---
11.0
64.4
48.7
9.5
0.2
30.0
2.2
1.8
285.0
5.5
6.4
68.1
82.5
1.2
1.2
37.0
4.8
0.9
1.5
10788.1
992.9
4403.5
24.8
5.0
13.0
1.9
51.6
139.1
186.5
48.8
149.0
5.0
2.2
45.2
25.6
20.5
10.0
24.3
42.3
0.4
660.2
5.0
19.5
49.3
22.9
64.0
1.7
8.8
10.2
1.4
64.7
2.7
617.2
7.6
6.4
141.8
162.0
13.2
2.8
93.0
1.5
The To'aga Site
204
Table 13.11 (continued)
Layers
Taxa
PELECYPODA
Anadara sp.
Arca spp.
Mytilidae
Isognomon spp.
Chama spp.
Codakia divergens
Gafrarium spp.
Lucinidae
Periglypta reticulata
Tridacna maxims
Hippopus hippopus
Quidnipagus palatam
Scutarcopagia scobinata
Tellinidae
Asaphis violaceus
IC
HA-1
HA
IHB
HC
HI
IV
Total
1.0
6.0
226.6
0.7
31.1
0.9
65.0
10.0
59.8
---
---
1.0
---
---
---
---
---
---
---
-----
2.0
75.4
69.5
---
---
---
38.5
--30.0
6.0
40.2
0.7
1.1
0.9
1.0
---
---
---
---
---
---
---
---
---
---
---
---
---
---
-------
-----
3.0
303.5
22.3
---
---
---
---------
0.3
3.3
8.0
3.1
1.9
10.0
ECHINOIDEA
---
CRUSTACEA
---
21.6
4.8
TOTAL
88.0
3862.7
VOLUME (i3)
DENSITY (kg/m3)
0.10
0.88
0.65
5.94
---
65.0
7.0
---
---
---
---
---
---
0.6
1698.8
215.0
13.2
10.9
75.0
---
59.2
1320.4
121.4
10.3
7.5
57.0
0.9
---
1.4
-----------------
32.8
5.1
27.5
28.4
52.2
1.9
17.9
0.3
5.3
152.0
45.8
2949.7 10305.2
9255.7
429.8
51.5
26942.6
1.85
5.57
1.85
5.00
1.75
0.25
1.05
0.05
8.20
3.29
---
---
0.95
3.10
shell recovered by the smaller screens added rela-
tively little to the shell recovered from the larger
screens.
The bulk samples show that the size of the
vertebrate sample greatly incmases as the screen size
decreases. Only one unidentifiable bone was
recovered from the -3 phi screen. The -2 phi screen
recovered only a fraction of the material recovered
from the -I and 0 phi screens (tables 13.15 and
13.16). Comparisons of the density of identifiable
fish bone obtained from the bulk samples to that
from the excavation unit illustrate the amount of
material being lost trugh the 1/4" screens (tables
13.17 and 13.18). While the standardization of the
volume to a cubic meter exaggerates the recovery
rate for the bulk samples, it shows that a significant
amount of bone may be lost through 1/4" screens.
The smaller screen sizes also increase the
sample's richness through the addition of new taxa.
---
40.7
-----
4.8
---
---
3385.7
336.4
25.7
26.5
150.0
5.4
Balistidae and lizard were not recovered in either
excavation unit. Along with Balistidae, four other
fish families (Ostraciidae, Muraenidae, Carangidae,
and Apogonidae) were added to the Unit 30 data
through fine screening. Most of these are smallbodied taxa with small diagnostic skeletal elements
that are less likely to be recovered by 1/4" screens.
Although the To'aga fish sample is the largest in
Westem Polynesia, the analysis of the bulk samples
shows that sample size, taxonomic richness, and thus
sample representativeness can greatly increase
through the consistent use of smaller screens and
bulk samples. This point is especially relevant for
Pacific island archaeology where the vertebrate
samples from most sites have been small. Because
the representativeness of the 1/4" sample is suspect,
the validity of interpretations based on measures of
diversity, such as richness (the number of taxa
present) and evenness (the distribution of abundance
Faunal Assemblages
Table 13.12
Invertebrate Fauna from Transect 5, Units 15/29/30
(weight in grams)
Layers
Taxa
GASTROPODA
Haliotis spp.
Patellidae
Trochus maculatus
Trochus niloticus
Tectus pyramis
Turbo crassus
Turbo setosus
Turbo spp.
Turbo operculae
Astrea stellare
Lunella cinereus
Nerita albicilla
Nerita picea
Nerita plicata
Nerita polita
Nerita spp.
Cerithium nodulosum
Cerithium columna
Cerithium spp.
Clypeomorus spp.
Strombus mutablis
Strombus spp.
Hipponix conicus
Hipponix sp.
Cypraea annulus
Cypraea caputserpentis
Cypraea eburneus
Cypraea mappa
Cypraea mauritania
Cypraea moneta
Cypraea tigris
Cypraea spp.
Policines spp.
Naticidae
Tonna spp.
Cymatium nicobarium
Cymatium spp.
Bursa granularis
Bursa spp.
Drupa ricina
Drupa morum
Drupa rubusidaceus
Drupa spp.
Morula sp.
(continued next page)
H
HIA-1
HIUB
HID
Total
2.7
0.4
95.4
1.3
1.9
151.6
872.8
138.9
396.5
57.4
7.8
2.7
3.4
684.2
43.8
14.3
1330.2
5665.1
372.7
3434.5
83.1
11.1
4.0
10.7
45.4
123.6
91.5
100.2
2.6
14.4
35.5
10.4
52.1
8.3
10.6
2.5
54.1
9.8
92.9
0.8
27.6
0.9
383.7
3.1
6.7
---
---
---
0.8
504.5
42.5
12.4
988.0
4213.8
172.8
2721.2
21.3
2.9
4.0
1.9
29.1
36.0
23.7
56.8
--43.4
2.2
40.9
---
-----
---
45.0
170.0
13.6
153.6
3.2
145.6
408.5
47.4
163.2
1.2
0.4
-----
---
---
---
8.5
13.8
82.6
51.7
17.4
2.6
9.2
8.6
5.2
11.2
2.9
1.8
0.3
8.6
---
0.3
2.5
3.9
16.1
26.0
---
---
---
5.0
19.9
--29.8
5.4
7.8
1.7
37.5
9.8
24.8
0.8
10.1
---
0.2
6.5
5.2
8.4
--1.1
---
0.5
---
2.7
---
---
----1.3
--0.8
---
1.0
0.5
6.7
---
---
3.8
---
63.5
---
---
17.5
---
---
241.6
9.8
---
0.9
22.6
2.3
1.8
0.1
7.3
---
0.3
11.6
7.2
6.2
42.5
5.8
34.9
29.0
13.6
8.3
2.8
-------
-----
-----
109.7
0.8
4.6
5.4
3.5
4.5
-----
---
---
9.7
---
9.7
6.8
---
---
---
---
---
---
1.1
---
---
4.5
1.2
17.1
18.0
10.7
42.5
5.8
54.3
35.8
13.6
13.9
4.0
205
206
The To'aga Site
Table 13.12 (continued)
Layers
Taxa
Nassa sp.
Thais armigera
Thais tuberosa
Thais spp.
Cantharus undosa
Nassarius gaudiosis
Latirus filamentosa
Vasum ceramicum
Conus chaldeus
Conus eburneus
Conus spp.
Terebra sp.
Bulla spp.
Dolabella spp.
Pythia spp.
Melampus spp.
II
HIA-1
IB
IIID
Total
2.3
88.2
111.7
124.5
--7.6
1.0
---
12.9
25.4
1.9
---
5.5
156.0
142.0
193.5
34.3
0.6
66.8
53.4
0.5
12.0
401.5
5.7
29.0
1.4
51.4
14.9
11.0
---
---
---
66.8
42.2
---
---
---
----0.4
3.6
----15.0
2.2
60.2
17.4
33.6
21.4
0.6
--7.6
0.5
0.8
36.6
---
---
---
0.8
7.2
5.3
---
---
---
0.4
0.2
3.0
3.5
48.0
10.2
0.7
1.2
---
9.7
1.6
14.8
1.4
26.4
11.2
349.5
5.7
15.7
1.4
---
1.0
10.0
---
PELECYPODA
Arca spp.
Mytilidae
Isognomon spp.
Chama spp.
Codakia spp.
Periglypta reticulata
Tridacna maxima
Hippopus hippopus
Quidnipagus palatam
Scutarcopagia scobinata
Trapezium oblongum
Asaphis violaceus
---
---
26.9
1170.1
---
7.7
1.0
0.3
11.0
215.2
---
---
---
28.9
51.2
4.2
50.7
1.6
2.4
4.2
13.1
13.6
3.6
10.3
1.1
7.1
3.6
34.6
1.7
30.1
0.3
41.5
1703.8
13.6
38.3
77.0
4.2
60.4
POLYPLACOPHORA
ECHINOIDEA
CRUSTACEA
---
---
---
72.6
17.7
5.0
1.0
28.7
11.1
5.3
215.6
8.8
5.3
321.9
38.6
Unidentified
198.4
19.4
30.5
162.4
410.7
TOTAL (g)
11853.2
514.6
1326.0
3212.5
16906.3
1.60
7.41
0.10
5.15
1.50
0.88
1.85
1.74
5.05
3.35
VOLUME (in3)
DENSITY (kg/m3)
1.3
8.9
0.3
2.7
-----
---
---
3.6
310.8
---
1.5
Fauna! Assemblages
Table 13.13
Units 20/23 Bulk Sample Analysis
(weight in grams)
Screen Size
Class
>-3
-2
Rock
Coral
522.5
234.4
Shell
56.6
30.7
34.1
19.6
Crab
Bone
TOTAL
8.6
0.4
950.2
-10
1
133.6
103.3
2.9
136.5
2.5
105.8
---
1.0
104.5
Table 13.14
Unit 30 Bulk Sample Analysis
(weight in grams)
Screen Size
Class
>-3
Rock
Shell
138.4
102.2
57.8
Bone
---
Coral
TOTAL
299.4
-2
49.7
68.5
24.5
0.8
143.5
-10i
10.9
30.4
30.4
3.9
14.8
2.5
32.9
values) is also questionable (Gordon 1991). Ideally,
a faunal assemblage should reflect the larger target
population of the archaeological record, not simply
te archaeological recovery techniques used.
TEMPORAL TRENDS IN THE
207
change which has been described for some Pacific
island sites is a quantitative shift from the exploitation of wild or naturally occuning resources to a
dominance on horticultural production (e.g., the
Tikopia case documented by Kirch and Yen [1982]
or the Mangaia case described by Steadman and
Kirch [1990]). Temporal increases in the frequency
of pig, dog, and chicken and decreases in wild
vertebrate taxa such as birds, turtle, and marine
mammal are taken to indicate this trend. In contrast,
the character of the To'aga assemblage changes little
over time and does not strongly reflect this kind of
shift. Wild taxa are found thmughout the site, and
the sample of domesticated animals is too small to
draw any firm conclusions. Although much of the
wild taxa (especially the birds) are represented in
early contexts, most of the chicken bone is also
found in those early layers. Thus, there is no clear
cut shift from one type of resource use to the other.
A corollary of the wild to domesticated fauna
hypothesis is the reduction of marine resources with
increasing reliance on horticulture (e.g., Janetski
1976, 1980; Kirch 1982, 1988). Resource exploitation and environmental degradation by humans are
also suggested to contribute to the decline in marine
resources, with decreases in the density of shellfish
used to support this hypothesis. Invertebrate density
varies at To'aga, increasing then decreasing over
time (tables 13.10-13.12). However, the use of
density measures may be misleading since changes
in density may result from other factor, such as
changing rates of sedimentation or shifts in settlement pattern.
In sum, the composition of the To'aga assemblage changes little over time. The invertebrate
assemblage best illustrates this with a few taxa
dominating the assemblage across time and space. A
similar trend appears to be evident for the fish
assemblage as well. At present, the cause of this
pattern is not evident. Some possible causes include
the exploitation of naturally abundant taxa from a
temporally stable environment, a lack of change in
subsistence practices, or a combination of both
factors.
TO'AGA ASSEMBLAGE
REGIONAL COMPARISONS
One goal of faunal analysis is the description
and interpretation of temporal change in prehistoric
subsistence patterns. A pattern of subsistence
Comparisons of faunal assemblages from
different areas or islands allow for the assessment of
208
The To'aga Site
Table 13.15
Vertebrate Taxa Represented in the
Bulk Sample From Units 20/23 (NISP)
Screen Size
Taxa
-24
-10i
Balistidae
7
5
Ostraciidae
Serranidae
3
3
Labridae
1
Holocentridae
1
4
2
2
2
Diodontidae
1
Muraenidhae
1
1
Acanthuridae
Scaridae
1
2
Rattus sp.
Bird
1
Lizards
TOTAL IDENTIFIED
1
18
UNIDENTIFIED
10
196
TOTAL
11
214
*
456
475
Not found in regular 0.25 inch screened samples from this
excavation unit.
regional trends. To'aga may be compared with other
assemblages from well-documented sites in
Western Samoa (Green and Davidson 1969, 1974;
Janetski 1976, 1980; Lohse 1980; Smith 1976),
Tonga (Kirch 1988; Poulsen 1987), and Fiji (Best
1981, 1984; Hunt 1980; Kay 1984). First, the issue
of data comparability is addressed to determine the
quality of the regional data base. Differences in
recovery, identification, and quantification techniques can seriously affect the comparability of data
across assemblages (Butler 1988; Nagaoka 1988). If
data sets are not comparable, differences between
them may reflect methodological rather than regional differences. Once these issues have been
addressed, the faunal data are then examined for the
faunal
1
19
invertebrate, fish, and non-fish vertebrate categories.
For Western Polynesian faunal assemblages,
recovery and quantification techniques vary considerably across sites (table 13.19). As was shown in
the analysis of the bulk samples from To'aga, screen
size influences the kind and the size of the faunal
sample. Screen size differences can even change
data at a nominal level since smaller screen sizes add
taxa. Quarter-inch screens have been the most
commonly used although for some sites screen size
was not reported. In other cases, several screen sizes
were used, but when and where the different sizes
were used was not reported. This lack of information makes it difficult to evaluate the comparability
of the data.
Faunal Assemblages
209
Table 13.16
Vertebrate Taxa Represented in the Bulk Sample
from Layer II, Unit 30 (NISP)
Screen Size
Taxa
-20
-10
0
-
Balistidae
6
Diodontidae
8
Labridae
2
7
4
Ostraciidae
5
Serranidae
Muraenidae
3
2
Scaridae
1
Lutjanidae
1
Carangidae
1
Apogonidhae
1
Rattus sp.
2
3
Lizard
3
1
TOTAL IDENTIFIED
4
34
13
UNIDENTIFIED
18
246
532
TOTAL
22
280
546
*
Not found in 0.25 inch screened samples from this excavation
unit.
NISP was the common technique for quantificadon of the vertebrate component, except for the Fiji
sites where MNI was used. For the invertebrates,
weight was used except for Lakeba and Naigani.
These differences in quantification may not be as
severe as screen size differences since, in many
cases, there is little difference between quantification
techniques at an ordinal level (Grayson 1984;
Jaretski 1980). If the data are considered in terms of
the rankings of taxa, it may still be possible to make
valid comparisons.
In the identification process, differences in the
reference collections and the diagnostic elements
used can also affect the data. Kirch (1988) and Best
(1984) noted that their fish reference collections
inadequate, limiting the number of possible
taxa that could be identified. This problem also
exists for the To'aga assemblage. Publication of
reference collections and the elements used would
help evaluate how these factors have biased the data.
Some differences among the faunal samples
may be due to the range of diagnostic elements used
to identify the assemblages, especially for the fish
assemblages. Kirch, Poulsen, and Best used mainly
the premaxilla, dentary, and special bones for their
fish identifications. For the To'aga assemblage three
additional elements, the articular, maxilla, and
quadrate, were used. This increased the size of the
To'aga sample about fifteen percent and added two
were
taxa.
210
The To'aga Site
Table 13.17
Density of Identified Fish Bone
from Layer III A/B, Units 20/23
No. of
Identified Fish
Excavation Unit
Bulk Samunple
Sample
Volume (m3)
Density
(NISP/m3)
221
1.3
170
34
0.0005
68,000
Given the problems in data comparability of the
Western Polynesian faunal assemblages, only
general comparisons between the data sets can be
made. The issue of comparability is important for
future faunal work in the area. Ideally, a faunal data
base would be created in which data from different
sites could be easily assimilated into one body of
knowledge with new data continually adding to our
knowledge of subsistence patterns.
Fish
For many of the Western Polynesian sites either
little faunal material was recovered or the data are
poorly reported. The data on fish bones from
Western Samoa consists of brief notes on their
presence in the sites. Janetski (1976, 1980) mentions
10 fish bones identified from Potusa and an unknown quantity from Falemoa and Jane's Camp.
Over 174 grams of fish bone were recovered from
Lotofaga (Davidson 1969), but no other data are
presented.
The best reported and largest samples of fish
come from Tongatapu, Niuatoputapu, Lakeba, and
To'aga. Kirch (1988) recovered a sample of 231
NISP across 11 taxa from Niuatoputapu. From
Tongatapu, Poulsen (1987) identified 15 fish taxa
containing 179 NISP. Lakeba produced 323 MI or
1782 NISP, and 21 taxa from the four sites for which
the fish component was analyzed (Best 1984). The
To'aga assemblage is comparable in size to Lakeba
with 2196 NISP and 22 taxa represented.
The most abundant fish taxa are inshore/reef
fishes, and a few taxa make up the majority of the
assemblage (fig. 13.3). The cause of this distribution
of fish taxa may be cultural (fishing strategies, food
preferences) or environmental (natural abundances
and distributions). Unfortunately, the biases created
by the recovery techniques, differential preservation,
and identifiability may have influenced these
distributions.
The most common fish families across sites are
Scaridae, Lethrinidae, Serranidae, Acanthuridae, and
Diodontidae. The dominance of taxa, such as
Scaridae, Lethrinidae, and Diodontidae, may be due
to preservation and identification bias as much as
subsistence patterns. The diagnostic elements of
these taxa are very robust and easily identified. It is
Table 13.18
Density of Identified Fish Bone
from Layer II, Unit 30
No. of
Identified Fish
Excavation Unit
Bulk
Samunple
29
52
Sample
Volume
(m3)
0.8
0.0005
Density
(NISP/M3)
36
104,000
Faunal Assemblages
211
To'aga
To'aga
-a
Other (22.0%)-
riranidae (17.7%)
Other (28.0%).
(42.1%)
Scaridae (6.6%)-
-Acanthuridae (18.2%)
Muraenidae
Hobcentridae (8.1 %)-
(6.0%)I
Carangidae (5.2%)Acanthuridae (10.7%)-
Fblocentridae (13.7%)
Scaridae (112%)
'-Serranidae (10.4%)
Lakeba
Lakeba
101/7/196
101/7/197
Other (6.6%)
Balistidae (5.3%)
Other (16.7/o)-
Scaridae (25.0%)
Serranidae (6.6%%)
(39.5%)
Diodontidae (6.6%
Acanthuridae (6.7%)Labridae
(10.0%)
Lethrinidae (15.0%)
Diodontidae (1 1.7%)-
Lethrinidae (35.5%)
LBalistidae (15.0%)
Niuatoputapu
Tongatapu
Other (22.3°h)-
Scaridae (24.6%)
Other (1 7.7h)-
Acanthuridae (7.8°h)-
Labridae (7.8h)-
-Lethrinidae (17.3%)
Serranidae (8.4%)-
Diodontidae (9.5%)-
Figure 13.3
'-Gireilidae (10.1%)
Elasmobranchii (10.0°h)-
Diodontidae (22.5%)-'
Percentage composition of fish faunal assemblages from major Western
Polynesian and Fijian sites; the two diagrams for the To'aga site
are
with and without Diondontidae.
(42.0%)
The To'aga Site
212
Table 13.19
Summary of Recovery and Quantification Techniques
for Western Polynesian Faunal Analyses
Vertebrate
Invertebrate
Screen Size Quantification Quantification
Site
SAMOA
Potusa
Falemoa
Jane's Camp
Lotofaga
TONGA
Niuatoputapu
Tongatapu
FUI
Lakeba
Naigani
Yanuca
Reference
1/4"
1/2", 3/16"
1/4"
P/A, NISP
P/A, NISP
P/A, NISP
P/A, Weight
Weight
Weight
Weight
Weight
Jennings et al. 1980
Jennings et al. 1980
Jennings et al. 1976
Green & Davidson 1969, 1974
1/4"
1/4"
NISP
NISP
Weight
Weight
Kirch 1989
Poulsen 1987
2.5, 5, 9mm
2.5, 3.5, 7.1mm
N/A
MNI
MNI
MNI
MNI
NISP
N/A
N/A, information not available
P/A, presence/absence
therefore more likely that these taxa will be preserved and identified than taxa with less robust and
distinctive skeletal elements.
The quality of the fish reference collections has
also influenced the presence or absence of taxa in the
assemblages. Based on edthoarchaeological data,
Mullidae and Pomacentridae are among the most
abundant fish caught on Niuatoputapu, but none are
recorded archaeologically (Kirch and Dye 1979).
Kirch (1988:225) suggests that the differences
between the modem and archaeological assemblages
may be due to the poor quality of the reference
collection. The lack of an adequate reference
collection is also a factor in the composition of the
Lakeba (Best 1981:497) and To'aga data. Finally the
use of 1/4" screens may have resulted in the absence
of Pomacentridae from te Niuatoputapu archaeological assemblage since these fish have small
diagnostic elements.
---
Best 1984
Best 1981; Kay 1984
Hunt 1980
NISP, number of identified specimens
MNI, minimum number of individuals
Tongatapu and To'aga. Fruit bat was recovered from
the Fijian sites, from Falemoa, and from
Niuatoputapu. The 'wild' vertebrate fauna tend to
be from earlier instead of later sites. Some of the
largest amounts of turtle and bird were from the
Lapita sites, TO-2, NT-90, 101M7/196 and 1017/197.
Chicken is the most common of the three
domesticated animals. Dog and pig are less abundant, possibly due to the problem of distinguishing
between tex two species when the bone is fragmented. The evidence for the presence of the pig,
dog, and chicken from initial colonization is scant.
Pig is present throughout the Lotofaga sequence, but
the basal date of the site is about A.D. 1000. At
Tongatapu, only chicken is found in the early site,
TO-2. Dog and chicken, as well as a bone tentatively identified as pig, were recovered from the
Lapita layers of Yanuca. The best evidence for dte
early presence of all three domesticated animals
comes from NT-90, on Niuatoputapu.
Non-Fish Vertebrates
Invertebrates
Compared to the fish, the non-fish vertebrate
sample is smaller, but better reported (table 13.20).
Rat, bird, and marine turtle are found at most sites.
Marine mammal was identified at only two sites,
The invertebrate component comprises a large
proportion of Western Polynesian faunal assemblages. The most abundant taxa vary across sites,
Fawsal Assemblages
213
Table 13.20
Summary of Western Polynesian Vertebrate
Faunal Assemblages (NISP)
Site
To'aga
Jane's Camp
Falemoa
Potusa
Lotofaga
Locus A
Locus B
Locus C
Tongatapu
TO-I
TO-2
TO-3
TO-4
TO-5
TO-6
Total
Niuatoputapu
NT-90
NT-91
NT-93
NT-100
Pig
1
2
37
p
p
p
NT-163
Total
Lakeba
1o0n/196
1o0n1/97
ioin/47
olin/i32
Chicken
15
p
---
380
Marine
Turtle
Bird
56
25
139
p
p
p
p
p
p
p
Fish
27
p
p
3
1
74
2
198
505
12
1
3
16
189
202
4
294
9
2
1
p
10
>39g
>74g
>61g
17
404
19
1
18
15
474
125
109
42
2
73
167
504
2
72
43
1
1
1
4
27
36
179
71
31
10
1
---
3
7
3
6
7
1
37
16
---
6
---
10
---
---
2
1
3
1
7
15
46
---
3
14
3
26
92
19
2
1
3
---
---
17
20
---
11
2
1
11
4
1
---
19
33
1
2196
87
p
p
p
3
1
93
28
8
22
3
ioin/135
4
5
52
2
69
1
---
12
34
231
76
60
180
7
1
iOin/2b
1o0n//66
Total
Naigani
Yanuca
Marine
Mammal
p
47
7
2
Fruit
Bat
Rat
p
10
NT-l10
NT-112
NT-113
NT-135
Dog
4
4
1
1
5
1
2
P = present in unknown quantities
21
2
20
3
4
11
5
2
61
12
9
72.
4
11
323
18
23
214
The To'aga Site
but the dominant taxa tend to reflect the marine
environment near the site. For example, the exploitation of a sheltered lagoon is reflected in the
abundance of bivalves in the Tongatapu assemblage.
Other sites contain mainly Turbo or other taxa
reflecting the exploitation of a fringing reef environment. Changes in the dominant taxa at
Niuatoputatpu, Tongatapu, and Lakeba are also used
to indicate changing marine environments.
While it appears that the most abundant taxa are
good indicators of environment, the influence of
environmental versus cultural factors still needs to be
determined. As at To'aga, the bulk of the invertebrate assemblages is concentrated in a few taxa.
Whether this uneven distribution reflects the exploitation ofnaturally abundant taxa or cultural preferences will need to addressed in future studies.
CONCLUSIONS
The analysis of the To'aga faunal assemblage
has generated a number ofnew questions. Despite
the time depth represented at the To'aga site, there is
a striking lack of change in the resources exploited.
A few taxa comprise a large percentage of the fish
and invertebrate components of the assemblage. The
overall pattem of high diversity may reflect the
exploitation of naturally abundant taxa or culturally
preferred taxa. Population studies of marine environments off To'aga would be useful for creating a
baseline of natural distributions which could then be
compared to the archaeological data.
Addressing these and other faunal questions
requires data robust enough to compile into a
cumulative data base and to use at a level higher than
nominal. The To'aga analysis has shown that
methodological biases introduced during excavation
and analysis can severely affect the data, reducing its
robustness. Thus, interpretations must be made
cautiously. The problems created by these biases are
compounded when data are compiled from different
sites into a regional data base. Variability in and
among data sets may be attributed to differences in
recovery or analytical techniques rather than prehistoric cultural patterning. Faunal analysis can be a
useful tool for understanding subsistence practices,
an important aspect of prehistoric culture. Its utility
in future studies, however, depends upon the commitment of faunal analysts and archaeologists to
create quality faunal data.
ACKNOWLEDGEMENTS
M. S. Allen, P. V. Kirch, and the Bishop
Museum kindly allowed the use of their reference
collections. V. L. Butler assisted in the fish identifications. E. Marshall at the Burke Museum and R
Kawamoto at the Bishop Museum assisted in
confirming shell identifications. E. Gordon, K.
Stark, and P. Kirch read the numerous versions of
this chapter. Funding was provided by Sigma Xi
Grants-in-Aid of Research for the development of
the author's fish reference collection.
REFERENCES CITED
Abbot, R. T., and S. P. Dance 1986. Compendium of
Seashells. Amen can Malacologists,
Melboume, Florida
Allen, M. S. 1990. Excavations at the Ureia site,
Aitutaki, Cook Islands: Preliminary results.
Archaeology in Oceania 25:24-37.
Anderson, A. 1986. Mahinga ika o te moana:
Selection in the pre-European fish catch of
southeem New Zealand. In A. Anderson, ed.,
Traditional Fishing in the Pacific: Ethnographical and Archaeological Papersfrom the
15th Pacific Science Congress, pp. 151-66.
Pacific Anthropological Records 37. Honolulu:
Bernice P. Bishop Museum.
Best, S. 1981. Excavations at Site VL21/5, Naigani
Island, Fiji, a preliminary report. Department of
Anthropology, University of Auckland.
-. 1984. Lakeba: The prehistory of a Fijian Island.
Ph.D. dissertation. University of Auckland.
Butler, V. 1987. Fish remains. IN D. E. Lewarch,
ed., The Duwamish No. I Site: 1986 Data
Recovery, pp. 1-37. URS Corporation and
BOAS Inc., Seattle.
1988. Lapita fishing strategies: The faunal
analysis. IN P. V. Kirch and T. L. Hunt, eds.,
Archaeology of the Lapita Cultural Complex: A
Critical Review, pp. 99-116. Thomas Burke
Memorial Washington State Museum Research
Report No. 5, Burke Museum, Seattle.
Casteel, R. 1972. Some of the biases in recovery of
archeological faunal remains. Proceedings of
the Prehistoric Society 36:382-88.
Faunal Assemblages
Chaplin, R. E. 1971. The Study ofAnimal Bones
from Archaeological Sites. New York: Seminar
Press.
Eisenberg, J. M. 1981. A Collector's Guide to
Seashells ofthe World. New York: McGrawHill.
Gordon, E. 1991. Differential faunal recovery and
models of subsistence change: A Hawaiian
example. Paper presented at the 1991 Annual
Meeting, Society for American Archaeology,
New Orleans, Louisiana
Grayson, D. K. 1979. On the quantification of
vertebrate archaeofaunas. IN M. B. Schiffer,
ed., Advances in Archaeological Method and
Theory, vol 2, pp. 199-237. New York: Academic Press.
Grayson, D. K. 1984. Quantitative Zooarchaeology.
New York: Academic Press.
Green, R. C. 1986. Lapita fishing: The evidence
from sites in the Reef/Santa Cruz group,
southeast Solomons. IN A. Anderson, ed.,
Traditional Fishing in the Pacific: Ethnographical and Archaeological Papersfrom the
15th Pacific Science Congress, pp. 1 19-36.
Pacific Anthropological Records 37. Bishop
Museum, Honolulu.
Green, R. C., and J. Davidson 1969. Archaeology in
Western Samoa, Vol. 1. Auckland Institute and
Museum Bulletin 6.
1974. Archaeology in Western Samoa, Vol. 2.
Auckland Institute and Museum Bulletin 7.
H.
Hill, B. 1978. The use of nearshore marine life as
afood resource by American Samoans. Unpublished MA. thesis, University of Hawaii,
Honolulu.
Hinton, A. G. 1972. Shells ofNew Guinea and the
Central Indo-Pacific. Rutland (Vermont):
Tuttle.
Hunt, T. L. 1980. Toward Fiji's past: Archaeological research on southwestern VidtLevu.
Unpublished M.A. thesis. University of
Auckland.
Janetski, J. C. 1976. Dietary remains from Jane's
Camp-a midden site. IN Jesse Jennings, ed.,
Excavations on Upolu Western Samoa, pp. 7582. Pacific Anthropological Records 25,
Department of Anthropology, Bishop Museum,
Honolulu.
1980. Analysis of dietary remains from Potusa
215
and Falemoa. IN J. Jennings and R. Holmer,
eds., Archaeological Excavations in Western
Samoa, pp. 117-22. Pacific Anthropological
Records 32, Department of Anthropology,
Bishop Museum, Honolulu.
Kay, R. M. A. 1984. Analysis ofarchaeological
materialfroom Naigani. Unpublished M.A.
thesis. University of Auckland.
Kirch, P. V. 1976. Ethno-archaeological investigations in Futuna and 'Uvea (Western Polynesia):
A preliminary report. Journal of the Polynesian
Society 85:27-69.
1988. Niuatoputapu: The Prehistory of a
Polynesian Chiefdom. Thomas Burke Memorial Washington State Museum Monograph No.
5. Burke Museum, Seattle.
Kirch, P. V., and T. S. Dye 1979. Edmoarchaeology
and the development of Polynesian fishing
strategies. Journal of the Polynesian Society
88:53-76.
Kirch, P. V., and D. E. Yen 1982. Tikopia: The
Prehistory and Ecology of a Polynesian
Outlier. Bishop Museum Bulletin 238. Bishop
Museum, Honolulu.
Leach, F. 1986. A method for the analysis of Pacific
Island fishbone assemblage and an associated
database management system. Journal of
Archaeological Science 13:147-59.
Lohse, E. 1980. Excavations on Manono Islet. IN J.
Jennings and R. N. Holmer, eds., Archaeological Excavations in Western Samoa, pp. 21-32.
Pacific Anthropological Records 32, Department of Anthropology, Bishop Museum,
Honolulu.
Masse, B. 1989. The archaeology and ecology of
fishing in the Belau Islands, Micronesia.
Unpublished Ph.D. dissertation, Department of
Anthropology, Southern Illinois University,
Carbondale.
Nagaoka, L. 1988. Lapita subsistence: The evidence
of non-fish archaeofaunal remains. IN P. V.
Kirch and T. L. Hunt, eds., Archaeology of the
Lapita Cultural Complex: A Critical Review,
pp. 117-134. Thomas Burke Washington State
Memorial Museum Research Report No. 5.
Seattle.
Payne, S. 1972. On the interpretation of bone
samples from archaeological sites. IN E. S.
Higgs, ed., Papers in Economic Prehistory, pp.
216
The To'aga Site
65-81. Cambridge: Cambridge University
Press.
Poulsen, J. 1987. Early Tongan Prehistory: The
Lapita Period on Tongatapu and its Relaionships, Vols. 1 and 2. Terra Ausais 12.
Department of Prehistory, Research School of
Pacific Studies, Australia National University,
Canberra.
Smith H. L. 1976. Jane's Camp (SUFl-I). IN J.
Jennings, ed., Excavations on Upolu, Western
Samoa, pp. 61-74. Pacific Anthropological
Records 25. Department of Antrpology,
Bishop Museum, Honolulu.
Thomas, D. H. 1969. Great Basin hunting pattems:
A quantitative method for treating faunal
remains. American Antiquity 34:393401.
14
BIRD BONES FROM THE TO'AGA SITE:
PREHISTORIC LOSS OF SEABIRDS AND MEGAPODES
DAVID W. STEADMAN
INTRODUCTION
A S PART OF A LONG-TERM program to reconstrUct
the natural distribution and diversity of birdlife in the South Pacific (figure 14.1), I have sought
bird bones from archaeological sites in many
different archipelagos. Until recently, the islands of
Samoa have not been represented in this data base.
In 1987, T. L. Hunt and P. V. Kirch conducted
excavations at the To'aga site (AS-13-1), Ofu Island,
American Samoa yielding a small sample of bird
bones representng at least six taxa (Steadman 1990).
This first glimpse ofthe prehistoric avifauna of
Samoa showed that at least two species (a shearwater and petrel) had been lost on Ofu since the first
anrival of humans more than 3000 years ago. Hunt
and Kirch expanded the excavations at To'aga in
1989. This paper reports the entire bird bone
assemblage from both the 1987 and 1989 field
seasons. The sample of bird bones from To'aga now
consists of at least fifteen taxa (table 14.1) and
provides a more thorough, although still far from
complete, portrayal of the birdlife of ancient Samoa.
The bird bones from To'aga have been catalogued in the University of Washington Burke
Museum (UWBM) Fossil Bird Collection. Comparative skeletal or oological specimens are from
UWBM, the British Museum (Natural History)
(BMNH), the New York State Museum (NYSM),
and the United States National Museum of Natural
History, Smithsonian Institution (USNM). hi the
species accounts that follow, "Unit" refers to meter
square excavations (designated with Arabic numerals). Roman numerals refer to stratigraphic layers.
For details of the stratigraphy, chronology, material
culture, and non-bird faunal assemblages ofthe
To'aga site, see Hunt and Kirch (1988), Kiivh et al.
(1989, 1990), and various chapters in this volume.
Unless stated otherwise, the modem distribudons of species within American Samoa are taken
from the excellent surveys conducted in 1975-1976
by Amerson et al. (1982ab) and in 1986 by
Engbring and Ramsey (1989). Modem distributions
for elsewhere in Polynesia are taken from Pratt et al.
(1987). The prehistoric records are from Steadman
(1989a) and from unpublished data from my recent
research; islands preceded by an asterisk (*) represent extirpated populations.
SPECIES ACCOUNTS
Order Procellariiformes
Family Procellanidae
Pwffinuspacificus (Wedge-tailed Shearwater)
MATERIAL. Humerus (UWBM 1680), T9/
500E, Unit 21, fiB. Five ulnae (UWBM 1244, 1251,
218
The To'aga Site
Table 14.1
Bird Bones from the To'aga Site
Taxa
Number of Identified Bones
Seabirds
*Puffinus pacificus
*Puffnus iherminieri
*Puffinus griseus
*Pterodroma rostrata
*Pterodroma sp., size of P. heraldica
(*)Procellariidae sp.
*Sula sula
Fregata sp.
Anous stolidus
Gygis alba
Steminae sp.
Landbirds
Egretta sacra
Numenius tahitiensis (M)
Gallus gallus (I)
*Megapodius sp.
Gallicolumba stairii
TOTAL
Total Seabirds (species/bones)
Total Landbirds (species/bones)
Total Native Landbirds, without IM (species/bones)
% of Bones from Extirpated Species (without IM)
11
2
15
6
2
9
1
2
1
1
1
1
1
16
2
3
15n4
10/51
5/23
4/7
85%
I=Introduced species
M=Migrant Species
* Extirpated on Ofu
(*) represents extirpated species, but not necessarily different from those already listed.
1256, 1678, 1679), Unit 9, IIB; Unit 5, IIC; Unit 14,
llIa,4; 19/500E, Unit 21, IIB (2 bones). Radius
(UWBM 1664), T9/500E, Unit 20, IIIB.
Carpometacarpus (UWBM 1630), T3I200W. Unit
27, fI1A. Tibiotarsus (UWBM 1642), T5/100E, Unit
28, HIB. Two pedal phalanges (UWBM 1246, 1248),
Unit 4, IIB; Unit 1, IIB.
REMARKS. This tropical shearwater is rarely
noted at sea today in American Samoa It may nest
on Pola Islet (off Tutuila) and on Ta'u, but this has
not been confirmed. There are no previous rcords
from Ofu. It is widespread in Polynesia today,
although nesting islands are few. Other Polynesian
archaeological records of PwJnuspac~lcus are
from *Ua Huka and *Tahuata (Marquesas),
*HuAhine (Society Islands), *Lifuka and * Eua
(Tonga), and Tikopia and Anuta (Solomon Islands).
Puojinus lherminieri (Audubon's Shearwater)
MATERIAL. 2 ulnae (UWBM 1651, 1671),
T9/500E, Unit 23, IIB; T9/500E, Unit 21, IIB.
REMARKS. Within American Samoa, Puffinus
iherminueri nests only on Tau, where at least 200
birds breed in the cloud forest, and on Tutila
(colony size unknown). It is uncommon at sea and
has not been recorded previously on Ofu. Like P.
pacificus, P. lherminieri nests today on relatively
few islands scattered across Polynesia. Other
Polynesian archaeological records of P. Iherminieri
are from *Nuku Hiva, Ua Pou, *Ua Huka, *Hiva Ok,
and *Tahuata (Marquesas), *Huahine (Society
Prehistoric Loss ofSeabirds and Megapodes
I
I
0C
C
C,5
N
0
o
4.
*0
J
.
0~~~
z~~~~~~~
o
w -J ~~~~~~~~~~-J
C)(
00J
*. (4
t
zes*
C
X
*
4
z~~~~~
0
-0
Z
.00
I-.a~~~~~~~~
a
.
:
Y
0
J
-
u0
Z
'0
w
0
Z
;'=
-
4
0I
| *oa
a-J
4
uot~~~~~~~~~~; J
z~~~~~
0.
-
-
i
z
z
w
Z
9~~~~~~~~~c
*r.
219
220
The Tolaga Site
Islands), *Mangaia (Cook Islands), Eua (Tonga),
and Tikopia and *Anuta (Solomon Islands).
Puffinus griseus (Sooty Shearwater)
MATERIAL. Sternum (UWBM 1681), T9/
500E, Unit 21, IIB. Coracoid (UWBM 1645), T5/
100E, Unit 29, RIB. Scapula (UWBM 1245), Unit
9, IIB. Three humeri (UWBM 1641, 1643, 1659),
T5/100E, Unit 29,1; T5/100E, Unit 28, IIB; Unit
20, RB. Three ulnae (UWBM 1240, 1241, 1647),
Unit 1 1, I-8 (2 bones); T5/100E, Unit 30,1H.
Carpometacarpus (UWBM 1259), Unit 7, hA.
Manus digit (UWBM 1260), Unit 7, HA. Pelvis
(UWBM 1674), T9/500E, Unit 21, IIB. Three
tibiotarsi (UWBM 1252, 1253, 1652), Unit 6, hA (2
bones); T9/500E, Unit 23, IRA.
REMARKS. As reported by Steadman (1990),
these bones are larger than those of any species of
shearwater that resides today in tropical Polynesia.
They agree in all osteological details with bones of
Puffinus griseus, a species that probably migrates
through the Samoan region today (Harrison
1983:260, 420, Pratt et al. 1987:55), although there
are no records from American Samoa. The Sooty
Shearwater nests today only on islands off New
Zealand, southern Australia, and extreme southern
South America (Harrison 1983:260, 420). Most of
the nesting localities of P. griseus are temperate or
subantarctic, although in Australia nesting occurs at
least as far north as 320 40' S on the subtropical
Broughton Islands.
Three possible explanations for the unexpected
presence of Paufinus griseus on Ofu were proposed
by Steadman (1990): (1) the bones represent
migrant birds taken at sea by fishermen; (2) P.
griseus once resided on Ofu and, like other
procellarids, was extirpated through predation in
their nesting burrows by humans and rats; (3) the
bones represent an extinct, resident shearwater that
differs specifically or subspecifically from modem
P. griseiss but is osteologically very similar. Regarding the first explanation, I am aware of no ethnographic accounts of Polynesians capturing seabirds
while fishing. The data now available, particularly
that fifteen bones (nine different skeletal elements)
are osteologically indistinguishable from modem
skeletons of P. griseus, suggest that the second
explanation may be corect, even though the oceanographic conditions near Ofu are wanmer than those
of the wannest localities where P. griseus nests
today. That this large shearwater (or an osteologically indistinguishable subspecies [explanation 3 is
not necessarily incompatible with explanation 2])
was once resident rather than migrant on Ofu is
indicated by the presence at To'aga of two bones of
P. griseus (UWBM 1645,1653) that, based upon
porosity of the external surface, are from volant
juveniles unlikely to have dispersed far from their
place of birth. The fonner residency of P. griseus on
Ofu is supported furtxr by the abun e of its
bones at To'aga (15 of 74, or 20% of identifiable
bones). Sooty Shearwaters are represented as well
among the few bird bones from a Lapita site on
*Niuatoputapu (Tonga), southwest of American
Samoa (Steadman 1990).
Pterodroma rostrata (Tahiti Petrel)
MATERIAL. Two mandibles (UWBM 1250,
1682), Unit 4, RB; T9/500E, Unit 21, IIB. Scapula
(UWBM 1684), 19/500E, Unit 21, IIB. Humenrs
(UWBM 1657), T9/500E, Unit 20, IIB. Ulna
(UWBM 1242), Unit 5, HA. Tarsometatarsus
(UWBM 1247), Unit 4, IIB.
REMARKS. This large petrel nests today at a
few widely scattered localities in te Marquesas
Islands, Tahiti, New Caledonia, and American
Samoa. Its presence on the latter is based upon a
colony of about 500 individuals discovered in 1976
in the montarn cloud forest of Ta'u, a single nesting
bird discovered on Tutuila in 1986, and a few birds
heard on Olosega in 1986. An onshore record in
1972 from Taveuni (Clunie et al. 1978) suggests th
Pterodroma rostrata may nest in Fiji as well.
Elsewhere in Polynesia, P. rostrata has been idendfied from archaeological sites on *Ua Huka and
Tahuata (Marquesas), *Huahine (Society Islands),
*Aitutaki (Cook Islands), *Lifuka and *Eua
(Tonga), Tikopia (Solomon Islands), and New
Caledonia.
Pterodroma sp. (unknown petrel,
size of P. heraldica)
MATERIAL Radius (UWBM 1670), T9/500E,
Unit 21, RB. Femur (UWBM 1676), Unit 21, RIB.
REMARKS. These specimens, although not
adequate for species-level identification, represent a
species ofPterodroma in the approximate size rae
Prehistoric Loss of Seabirds and Megapodes
of P. heraldica (Herald Petrel, which often is
considered conspecific with P. arminjonana) or
perhaps the slightly larger P. exlerna (Juan
Fernandez Petrel), both of which are significantly
smaller than P. rostrata. In American Samoa, P.
eterna is known ordy from uncommon sightings at
sea, while P. heraldica nests on TaVu. Regardless of
species-level identification, UWBM 1670 and 1676
indicate that a species of petrel other than P. rostrata
probably once nested on Ofu.
Procelladiidae sp. (unknown petrel/shearwater)
REMARKS. These specimens cannot be
distinguished from those of Fregata minor or F.
ariel, both of which visit but do not nest on Ofu
today (total of thirteen birds counted in 1975-76; still
recorded as a visitor in 1986). The bones ofF.
minor males and F. ariel females are similar in size
and difficult to distinguish from each other
(Steadman et al. 1990). Both species of Fregata
occur fairly commonly in Polynesian archaeological
sites.
Order Charadriifonues
Family Laridae
MATERIAL. Coracoid (UWBM 1633), T5/
100E, Unit 16,11. Five ulnae (UWBM 1243,1258,
1634, 1662, 1669), Unit 9, IIB; Unit 5, IIB; T5/
100E, Unit 16, I; T9/500E, Unit 20, IIIB; T9/500E,
Unit 21, fIB. Radius (UWBM 1249), Unit 4, IIB.
Two carpometacarpi (UWBM 1239, 1677), Unit 9,
UB; T9/500E, Unit 21, IIB.
REMARKS. These fragmentary bones cannot
be identified beyond the family level. They do not
represent a taxon separate from those already
identified.
Order Pelecaniformes
Family Sulidae
Sula sula (Red-footed Booby)
221
Anous stolidus (Brown Noddy)
MATERIAL. Humerus (UWBM 1668),T9/
500E, TP 21, IIB.
REMARKS. This tern is common and widespread today in American Samoa as well as most of
Polynesia. The Ofu population was ca. 500 in 197576 and was not accurately estimated in 1986. Other
Polynesian archaeological records of Anous stolidus
are from Henderson (Pitcairn Group); Ua Pou, Ua
Huka, and Tahuata (Marquesas); Huahine (Society
Islands); Mangaia and Aitutaki (Cook Islands);
Niuatoputapu, Lifuka, and 'Eua (Tonga); 'Upolu
(Western Samoa); Tikopia and Anuta (Solomon
Islands); and Mussau (Papua New Guinea).
MATERIAL. Radius (UWBM 1656), T9/500E,
Unit 20, IIIC.
REMARKS. Sula sula nested on Ofu as
recently as 1975-76 (only twenty-five birds), but the
small colony no longer existed in 1986. This booby
still nested in 1986 on Tutuila and Rose islands. It is
widespread in tropical oceans. Other Polynesian
archaeological records of S. sula are from
*Henderson (Pitcaim Group); *Ua Huka, *Hiva Oa,
and *Tahuata (Marquesas); *Huahine (Society
Islands); *Aitutaki (Cook Islands); Niuatoputapu
(Tonga); and *Tikopia and *Anuta (Solomon
MATERIAL. Humerus (UWBM 1631), T5/
100E, Unit 16,I.
REMARKS. This distinctive tern is common
and widespread today in most of Polynesia including
American Samoa. The Ofu population was ca. 100
in 1975-76, and at least 500 in 1986. Other
Polynesian archaeological records of Gygis alba are
from Henderson Island (Pitcairn Group), Ua Huka
and Tahuata (Marquesas), Huahine (Society Islands),
Mangaia (Cook Islands), and Niuatoputapu and 'Eua
Islands).
(Tonga).
Gygis alba (Common Fairy-Tern)
Sterninae sp. (unknown tern)
Family Fregatidae
Fregata sp. (unknown frigatebird)
MATERIAL. Ulna (UWBM 1257), Unit 14,
llb-5.
MATERIAL. Humerus (UWBM 1254), Unit
14, lIa-4. Ulna (UWBM 1675), T9/500E, Unit 21,
BB.
REMARKS. This eroded, fragmentary specimen represents a tern smaller than either of the
above species. UWBM 1257 is approximately te
222
The To'aga Site
size of the ulna in Sterna swnatrana orAnous
minus, both of which occur today in American
Samoa, although only the latter nests on Ofu (just ten
birds in 1975-76, and perhaps about the same in
1986).
Order Clconiifonmes
Family Ardeidae
Egretta sacra (Pacific Reef-Heron)
MATERIAL. Coracoid (UWBM 1655), T9/
500E, Unit 22, 1.
REMARKS. Egreta sacra resides throughout
American Samoa and most of Polynesia today. It is
uncommon on Ofu, where only dark-phase birds
have been recorded. Other archaeological records
are from Nuku Hiva and Ua Huka (Marquesas),
Huahine (Society Islands), Mangaia and Aitutaki
(Cook Islands), and 'Eua (Tonga).
Order Charadriiformes
Family Scolopacidae
Nwnenius tahidensis (Bristle-thighed Curlew)
MATERIAL. Corcoid (UWBM 1635, 1636;
originally believed to be two separate bones), T5/
100E, Unit 15, IIID.
REMARKS. This widespread but rather
uncommon migrant shorebird has not been recorded
previously from Ofu, although undoubtedly it still
occurs there occasionally. American Samoan
records ofNwnenius tahlitensis are confined to
Tutuila, Ta'u, Swains, Rose, and Olosega. Other
Polynesian archaeological records for N. tahidensis
are from Henderson Island (Pitcaim Group), Ua
Huka (Marquesas), Huahine (Society Islands),
Mangaia and Aitutaki (Cook Islands), and Tikopia
(Solomon Islands).
During the autumn wing molt, thirteen of
twenty-nine individuals ofN. tahtidensis captured on
Laysan (Leeward Hawaiian Islands) were flightless
(Marks et al. 1990). This adaptation would seem to
be viable only in a predator-free setting. It may have
led to reductions in the distribution and abundance of
N. ahitiensis following the human colonization of
Polynesia.
Order Galliformes
Family Megapodiidae
Megapodius sp. (Unknown Megapode)
MATERIAL. Ulna (UWBM 1637), T5/100E,
Unit 15, HID. Femur (UWBM 1649), Unit 30, HID.
REMARKS. These two fragmentary specimens
(fig. 14.2) are most similar quantitatively and
qualitatively to modem specimens of Megapodius
freycinet, a widespread species that now reaches its
eastern limit in Vanuatu. UWBM 1649 is indistinguishable qualitatively from the femur in modem M.
freycinet, while UWBM 1637 differs from the ulna
of M. freycinet in having a slightly deeper sulcus
radialis and in lacking a diagonal trough on the
cranial surface of tuberculum carpale.
Like the archaeological specimens of
Megapodiusfreycinet from *Tikopia (Solomon
Islands; Steadman et al. 1990), the specimens of
Megapodius from To'aga are at the extreme small
end of the size range ofM.freycinet (table 14.2).
Among living subspecies of M.freycinet, there is no
indication that individual body size decreases
eastward in Oceania. For example, the tarsus (in
skins) of the Vanuatu population is not smaller han
that from Australia (table 14.3). The bones from
To'aga are larger than in M. pritchardi (confined to
Niuafo'ou, Tonga) and M. laperouse (found orny in
Palau and Mariana Islands), but much smaller than
in two extinct species recently described from late
Holocene archaeological and paleontological sitesM. molistructor of New Caledonia and Lifuka and
M. alimentum of Lifuka and 'Eua (Balouet and Olson
1989; Steadman 1989b, pers. obs.). Megapodius
pritchardi is the only species of megapode that
survives east of Vanuatu.
Both To'aga specimens of Megapodius are from
one of the site's deepest and oldest strata, Layer IIID
in units 15/29/30. This suggests that megapodes
may have been lost from Ofu not long after prehistoric colonization of the island. These bones provide
the first unequivocal, well-documented record of a
megapode from Samoa. There is, however, complicated historical evidence that a megapode, described
as M. stain by Gray (1861), may have existed in tbe
mid-i 800s on Upolu or Savai'i (Western Samoa).
Gray (1861) also described M. burnabyi from
Prehistoric Loss of Seabirds and Megapodes
223
C%
A
Figure 14.2
B
Comparison of megapode bones: ulna in ventral aspect
(A, B) and femur in lateral aspect (C, D). A, C.
Megapodius cf.freycinet, archaeological specimens
(UWBM 1637, 1649), To'aga site, Ofu, American Samoa.
B, D. M. freycinet, modem specimen, USNM 557015,
male, Halmahera, Northern Moluccas. Scale = 20 mm.
Ha'apai (Tonga). The true identities of M. stairi and
Mf. burnabyi are uncertain because only a single egg
-ver was collected of each, and these two eggs
cannot be assigned unequivocally to M. pritchardi,
M.freycinet, or any other species of megapode
(Steadman 1991). One or both of these eggs may
represent the same species as the bones from Ofu,
which also are at the very lower limit of the size
range of M. freycinet and larger than, or at the
uppermost size limit of, M. pritchardi.
The original, handwritten data slip with the
holotypical egg of Megapodius burnabyi in the
British Museum notes that this egg was "called the
chief s egg' as they are only allowed to eat them."
Such a chiefly tabu might suggest rarity of the bird
and a knowledge that overexploitation, which
probably was the cause of the rarity, would eventu-
ally lead to extinction. Alternatively, being called
the "chief's egg" could suggest that megapode eggs
were prestigious trade items brought to Ha'apai from
another island. An extensive exchange network
operated among Fiji, Tonga, and Samoa in late
prehistoric and early historic times (Kirch 1984:23842; 1988:257-60). The Samoan voyages recorded by
Stair (1895) included Fiji and Tonga as well as much
of East Polynesia. Bennett (1862:247) noted that the
nesting grounds of M. pritchardi on Niuafo'ou were
"under the protection of the king or chief, and by his
permission only can the birds or eggs be procured."
Even if megapodes of the Fiji/Tonga/Samoa region
were confined by the nineteenth century to
Niuafo'ou, they would have been known to Tongans
in Ha'apai, as well as to Samoans.
Unless additional evidence comes forth, neither
224
The To'aga Site
Table 14.2
Measurements (in mm) of the Femur and Ulna of Megapodius, with Mean,
Range, and Sample Size
Specimen
Locality
Megapodious sp.
UWBM 1637 (U)
Samoa: Ofu: To'aga Site
Megapodius sp.
UWBM 1649 (U)
Samoa: Ofu: To'aga Site
M. f cf. eremita
Lab #: 66, 68 (U)
New Ireland
M. f freycinet
USNM 556995-557002
557016, 557017 (M)
N. Moluccas: Halmahera
M. f freycinet
USNM 557006-557008
6.5
1
7.7
1
C
D
E
4.4
1
3.3
1
6.8
1
4.9
1
8.4
1
7.3
1
8.2
7.6-9.1
10
3.8
5.0
4.7-5.3 3.3-4.1
10
10
7.4
7.0-7.9
N. Moluccas: Halmahera
8.2
7.6-8.8
10
8.3
7.7-8.8
10
3.7
4.9
4.6-5.4 3.5-4.1
10
10
7.3
7.0-7.8
10
Sulawesi
7.2
6.9-7.5
2
7.4
7.3-7.6
2
3.2
4.3
4.3 3.2-3.3
2
2
6.5
6.5
2
8.1
1
8.0
1
5.2
1
4.1
1
7.4
1
8.3
8.0-8.6
2
8.3
8.2-8.4
2
5.0
3.8
4.9-5.0 3.6-4.0
2
2
7.8
7.6-7.9
2
5.6
5.4-5.7
2
6.0
5.9-6.1
2
2.8
3.4
3.3-3.5 2.8-2.9
2
2
5.2
5.0-5.3
2
USNM 226175, 226176 (F, M)
M. f cumingi
BM(NHl) 1862.2.10.2 (U)
Philippines: island
unknown
M. f nicobariensis
USNM 19686, 19700 (M)
Nicobar Islands
M. pritchardi
tJSNM 319633, 319634 (IJ)
Tonga: Niuafo'ou
M. laperouse
USNM unnumbered (U)
N. Mariana Islands:
Rota: Payapai Cave
M. alimenturm
UWBM 2100 (IJ)
Tonga: 'Eua:
'Anatu (Ground-Dove Cave)
F = Female, M = Male, U = Sex unknown
Column Headings:
A. Femur: width at deepest proximo-lateral muscle scar
B. Femur: depth at deepest proximo-lateral muscle scar
C. Ulna: minimum width of shaft
D. Ulna: minimum depth of shaft
E. Ulna: width of distal end
B
8.2
7.5-8.7
10
557010-557013, 557019-557022 (F)
M. f gilberti
A
3.4
-
1
2.5
1
6.0
1
4.4
1
10
5.4
1
Prehistoric Loss of Seabirds and Megapodes
225
Table 14.3
Tarsal Length (in mm) from Skins of Selected
Subspecies of Megapodius freycinet, with Mean,
Range, and Sample Size. Based upon Specimens
from BM(NH).
Subspecies
Tarsal Length
M. f layardi
Vanuatu: Santo, Vate
(F, 5M)
73.8 (71-76), 6
M. f eremita
Solomon Is.: San Cristobal,
Guadacanal, Bouganville
72.2 (67-77), 6
M. f eremita
Papua New Guinea: New Britain
(3F, 2M, 3U)
67.1 (62-71), 8
(3F, 2M, tI
M. f yorki
Australia: Queensland
(2F, 4M, 4tJ)
71.6 (70-74), 10
F = female, M = male, U = sex unknown
Megapodius stairi orM. burnabyi should be regarded as certain records of indigenous populations
ofmegapodes in nineteenth century Samoa or
Ha'apai, although this cannot be ruled out. The
survival of M. pritchardi on Niuafo'ou has been due
to chiefly control of exploiting eggs and birds at the
conspicuous nest mounds, as first described by
Bennett (1862). The people of Niuafo'ou must have
realized that conserving megapodes, which probably
occurred nowhere else in the region, would help to
maintain their share of commerce in the SamoaTonga-Fiji trade network.
Family Phasianidae
Gallus gallus (Chicken)
MATERIAL. Sternum (UWBM 1261), Unit 11,
1. Coracoid (UWBM 1654), T9/500E, Unit 23, IIIB.
Scapula (UWBM 1663), T9/500E, Unit 20, HIB.
Ulna (UWBM 1255), Unit 14, Illa-4. Radius
(UJWBM 1660), T9/500E, Unit 20, IhB. Two pelves
(UWBM 1665, 1686, 1688; the last two originally
believed to be separate bones), T9/500E, Unit 20,
RIB; T9/500E, Unit 21, IIB. Two femora (UWBM
1666, 1667), T9/500E, Unit 20, hIB. Four tibiotarsi
(IJWBM 1632, 1644, 1658, 1661), T5/100E, Unit
16, 1; T5/1OOE, Unit 29, IIIB; T9/500E, Unit 20, IIB;
T9/500E, Unit 20, IhB. Two tarsometatarsi
(UWBM 1646, 1683), Unit 29, rIB; Unit 21, II.
REMARKS. Feral and/or domestic populations
of Gallus gallus occur neatly throughout Polynesia,
including all inhabited Samoan islands. All chickens
recorded on Ofu in 1986 were near human habitation
and not from deep within forests. This non-native
species has been found thugh virtually all of
Polynesia in archaeological sites of any age, except
that it is absent from all sites on Henderson Island
(Schubel and Steadman 1989). Chicken bones occur
throughout the To'aga sequence, but are especially
well represented in the Layer IIIB occupation in
Units 20/23 along Transect 9: This indicates that G.
gallus was a commonly eaten bird during the
Ancestral Polynesian phase, ca. 2500 yr B.P.
226
The To'aga Site
Order Columbiformes
Family Columbidae
Gallicolwnba stairii (Shy Ground-Dove)
MATERIAL. Humerus (UWBM 1638), T5/
100E, Unit 15,111D. Ulna (UWBM 1658), Unit 30,
IIID. Tarsometatarsus (UWBM 1690), T5/100E,
Unit 29, IIID.
REMARKS. This species occurs only in very
old deposits at To'aga, primarily Layer IIID in Units
29/30. Gallicolwnba stairii is extremely rare on Ofu
today, with a very roughly estimated 100 birds
surviving in 1975-76. Only two or three grounddoves were seen on Ofu during the 1986 surveys; no
population estimate was made. Within American
Samoa, only perhaps on Olosega does another small
population of G. stairii survive. Similar declines or
losses of populations of G. stairii have occurred in
Tonga, where the only other archaeological record of
G. stairii is from *'Eua.
DISCUSSION
Virtually all of the bird bones from To'aga are
broken, often with both articular ends missing. Most
of these breaks are not fresh, although often it is
difficult to distinguish whether human or sedimentary processes have caused the breakage. A few
bones are rounded, suggesting some post-mortem
sedimentary transport. The breakage and rounding
might indicate that the calcareous sands at To'aga
represent a somewhat higher energy deposit than the
calcareous sands at certain other Polynesian archaeological sites, such as Hane (Ua Huka, Marquesas) or
Tongoleleka (Lifuka, Tonga). Among the shearwaters and petrels, however, a systematic butchering
technique is suggested by the fairly consistent
pattern of both ends of the humerus, ulna, and
tibiotarsus being broken off. Two of the chicken
bones had been chewed by rats. None of the bird
bones seems to have been modified into recognizable artifacts.
Among indigenous, resident species recorded
from the To'aga site, five of ten seabirds and one of
three landbirds are extirpated on Ofu (table 14.1). At
least two of the surviving species (Sterninae sp.,
Gallicolwnba stairit) exist today on Ofu only in very
small, threatened populations. Should these species
be lost from Ofu, the proportion of bones of extir-
pated species at To'aga would increase from 85%
(table 14.1) to 93%.
The majority of bird bones from To'aga (46 of
74, or 62%) are of at least five species ofpetrels or
shearwaters (table 14.1), none of which nests on OfN
today (Amerson et al. 1982a:90). Only two of hse
species (Audubon's Shearwater and Tahiti Petrel)
are known certainly to nest today anywhere in
American Samoa. As seems to be case throughout
Polynesia (Steadman 1989a, Dye and Steadman
1990), the island of Ofu had a diverse and probably
abundant seabird fauna when humans first arrived.
In the case of Ofu, not a single species of procellariid
has survived the three millennia of human occupation.
Compared to avian assemblages from other
Polynesian archaeological sites, the dominance of
procellariids at To'aga would characterize a fairly
early site, i.e., one that dates to within the first
thousand years of human occupation (Dye and
Steadman 1990). When compared to sites that seem
to represent the initial human occupation of an
island, however, such as the Hane site (Ua Huka,
Marquesas; Steadman 1989a) orTongoleleka site
(Lifuka, Tonga; Steadman 1989b), the To'aga site's
lower percentage and diversity of bones from native
landbirds and higher percentage of chicken bones
would suggest that this site may not represent the
first 500 years of human occupation of Ofu. This
suggestion is compatible with the radiocarbon
chronology at To'aga which indicates occupation of
the site from about 2800 to 1900 yr B.P. (see Kirch,
chapter 6).
Although the bird bones from To'aga provide
much new data on the prehistoric distribution of
Samoan birds, a bone sample about an order of
magnitude larger would be necessary to provide a
fairly complete picture of the past birdlife of Ofu.
That six of the fifteen taxa of birds from To'aga are
represented by only a single bone indicates that more
species await discovery if a larger bone sample were
available. The 1987 sample of twenty-three bones
yielded six taxa, which increased to fifteen taxa with
the addition of fifty-one bones from the 1989
excavations. The point of diminishing returns is
difficult to determine, however, from comparison
with assemblages from other sites. For example,
thirty-five species of birds are represented in a
sample of ca. 350 bones from the Fa'ahia site
(Huahine, Society Islands; Steadman and Pahlavan
Prehistoric Loss of Seabirds and Megapodes
1992), while thirty-nine species of birds ae represented in a sample of ca. 12,000 bones from the
Hane site (Ua Huka, Marquesas) (Steadman 1989a,
pers. obs.). Twelve of the species from Hane are
known from either one or two bones. Another gauge
of the incompleteness of the To'aga avifauna is the
low percentage of species in the modem avifauna of
Ofu tha are represented at the site. In this case, only
two of sixteen possible species of resident landbirds
(12.5%) ar represented.
The To'aga site has given us an intriguing
introduction to Samoa's prehistoric birdlife. Our
understanding of the relationship between native
birds and the first human inhabitants of Samoa
undoubtedly will improve as more early archaeological sites are discovered and carefully excavated.
ACKNOWLEDGEMENTS
This research was supported in part by National
Science Foundation grant BSR-8607535. I than T.
L. Hunt and P. V. Kirch for inviting me to study the
bird bones from To'aga. L. Nagaoka and D. S.
Pahlavan assisted in curation. S. L. Olson and M. P.
Walters kindly shared information on megapodes.
For comments on the manuscript, I thank T. L. Hunt,
P. V. Kirch, N. G. Miller, S. L. Olson, and M. P.
Walters. This is contribution number 670 of the
New York State Museum and Science Service.
REFERENCES CITED
Amerson, A. B., Jr., W. A. Whistler, and T. D.
Schwaner 1982a. Wildlife and Wildlife Habitat
ofAmerican Samoa. 1. Environment and
Ecology. U.S. Dept. of Interior, Fish and
Wildlife Service, Washington D.C.
. 1982b. Wild!ife and Wild0lfe Habitat ofAmerican Samoa. H. Accounts ofFlora and Fauna.
U. S. Dept. of Interior, Fish and Wildlife
Service, Washington D.C.
Balouet, J. C., and S. L. Olson 1989. Fossil birds
from late Quatemary deposits in New
Caledonia. Smithsonian Contributions to
Zoology 469.
Bennett, G. 1862. [Letter from Dr. G. Bennett on
Didunculus, Tallegalla, and Megapodius.]
Proceedings of the Zoological Society of
London 1862:24648.
Clunie, F., F. C. Kinsky, and J. A. F. Jenkins 1978.
227
New bird records from the Fiji Archipelago.
Notornis 25:118-27.
Dye, T., and D. W. Steadman 1990. Polynesian
ancestors and their animal world. American
Scientist 78:207-215.
Engbring, J., and F. L. Ramsey 1989. A 1986 Survey
of the Forest Birds of American Samoa. U.S.
Dept. of Interior, Fish and Wildlife Service,
Washington D.C.
Gray, G. R 1861. [pub. April 1862] List of species
composing the family Megapodiidae, with
descriptions of new species, and some account
of the habits ofthe species. Proceedings of the
Zoological Society ofLondon 1861:288-96.
Hanison, P. 1983. Seabirds: An Identification Guide.
Houghton Mifflin Co., Boston.
Hunt, T. L., and P. V. Kirch 1988. An archaeological
survey of the Manu'a Islands, American
Samoa. Journal of the Polynesian Society
97:153-83.
Kirch, P. V. 1984. The Evolution of the Polynesian
Chiefdoms. Cambridge: Cambridge University
Press.
1988. Niuatoputapu: The Prehistory of a
Polynesian Chiefdom. Thomas Burke Memorial Washington State Museum Monograph No.
5.
Kirch, P. V., T. L. Hunt, and J. Tyler 1989. A
radiocarbon sequence from the To'aga site, Ofu
Island, American Samoa. Radiocarbon 31:713.
Kirch, P. V., T. L. Hunt, L. Nagaoka, and J. Tyler
1990. An ancestral Polynesian occupation site
at To'aga, Ofu Island, American Samoa.
Archaeology in Oceania 25:1-15.
Marks, J. S., R. L. Redmond, P. Hendricks, R. B.
Clapp, and R. E. Gill, Jr. 1990. Notes on
longevity and Sightlessness in Bristle-thighed
Curlews. Auk 107:779-81.
Pratt, H. D., P. L. Bruner, and D. G. Berrett 1987. A
Field Guide to the Birds ofHawaii and the
Tropical Pacific. Princeton: Princeton University Press.
Schubel, S. E., and D. W. Steadman 1989. More bird
bones from Polynesian archaeological sites on
Henderson Island, Pitcaim Group, South
Pacific. Atoll Research Bulletin 325:1-18.
Stair, J. B. 1895. Flotsam and jetsam from the great
Ocean; or, summary of early Samoan voyages
and settlements. Journal of the Polynesian
228
The Tolaga Site
Society 4:99-131.
Steadman, D. W. 1989a. Extinction of birds in
Eastern Polynesia: A review of the record, and
comparisons with other Pacific island groups.
Journal ofArchaeological Science 16:177-205.
1989b. New species and records ofbirds (Aves:
Megapodiidae, Columbidae) from an archeological site on Lifuka, Tonga. Proceedings of
the Biological Society of Washington 102:53752.
. 1990. Archaeological bird bones from Ofu
[Manu'a, American Samoa]: Extirpation of
shearwaters and petrels. Archaeology in
Oceania 25:14-15.
1991. The identity and taxonomic status of
Megapodius stairi and M. burnabyi (Aves:
Megapodiidae). Proceedings of the Biological
Society of Washington 104:870-77.
Steadman, D. W., and D. S. Pahlavan 1992. Prehistoric exploitation and extinction of birds on
Huahine, Society Islands, French Polynesia.
Geoarchaeology 7:449-83.
Steadman, D. W., D. S. Pahlavan, and P. V. Kirch
1990. Extinction, biogeography, and human
exploitation of birds on Tikopia and Anuta,
Polynesian outliers in the Solomon Islands.
Occasional Papers of the Bishop Museum
30:118-53. Honolulu: Bemice P. Bishop
Museum.
15w
SYNTHESIS AND
INTERPRETATIONS
P. V. KIRCH AND T. L. HUNT
PRIOR TO THE MANU'A PROJECIr in 1986, our
knowledge of the archaeology and prehistory
ofthe most easterly Samoan islands-Ta'u, Olosega,
and Ofu-was rudimentary indeed. Limited reconnaissance survey had revealed the presence of stoneand-earth monuments, especially house platfonns,
typical of Samoan settlement pattems on the larger
islands (Kikuchi 1963). Attempts at excavation on
Ta'u Island by a Bishop Museum expedition in 1962
(in one cave site and two open "cooking-house
sites') were disappointing (Emory and Sinoto
1965:40-48), and no stratified materals were
discovered that might be used to outline a cultual
sequence. Although the Bishop Museum team
collected 201 basalt adzes and coconut graters from
Manu'a, these were all from surface contexts (Emory
and Sinoto 1965, table 2). In short, at the commencement of our project, knowledge of Manu'an
archaeology was limited to bdef descriptions of
urface sites and stone tools, with no time depth.
Against this background, our 1986 reconnaissance work thrughout the Manu'a Group yielded a
number of significant advances (Hunt and Kirch
1988). The number and range of surface monuments
were extended considerably. Peitaps most importantly, several stratified sites were discovered, two of
which on test excavation yielded ceramic assemblages radiocarbon dated to between 1950-1850 cal
B.P. (Hunt and Kirch 1987). These results confirned
our initial expectations that the Manu'a Group would
prove to have a deep prehistoric sequence extending
back to the Ancestral Polynesian phase, quite
probably in parallel with the sequence defined for
Westem Samoa (Green and Davidson 1969, 1974).
The preceding fourteen chapters have presented
the detailed results of two subsequent seasons of
concentated field and laboratory research at the
most promising of the ceranic-bearing sites discovered during the 1986 reconnaissance: the To'aga site
on Ofu Island. In 1987 and again in 1989 we carried
out a progrwn of systematic subsurface sampling of
te To'aga area, producing an areal and stratigraphic
definition of what is currently the largest and most
deeply stratified early site known for the entire
Samnoan archipelago. Although our research design
was of necessity orented first and forenost to
considerations of culrl resource management (i.e.,
the spatial definifion of the site and assessment of its
significance), we have also been able to use this
research opportunity to expand on certain aspects of
our knowledge of Manu'an-and indeed general
Samoan-prehistory. For example, we have
fonnulated and tested a model of morphodynamic
laape change witfi implications for other coastal
sites dtoughout the archipelago. Similarly, our
various analyses of artifactual and faunal materials
230
The To'aga Site
from the
systematic
tra
excavations have
enhanced the weconstncdons of early Samoan
material cultu, subience economy, and interisland exchange. In this concluding chapter, we
attempt to integrate these significant new results
from the To'aga site with the existing reconstrucdons
and interetations of Samoan prehistory, deriving
primarily from work in Westem Samoa (Green and
Davidson 1969, 1974; Jennings et al. 1976; Jennings
and Holmer 1980) and to a limited extent from
rcent work
Tutila Island (e.g., Claik and
Herdrich 1988; Cark 1989; Leach and Witter 1987;
Best et al. 1989).
In presenting this synthesis, we are acutely
aware that our fieldwork at To'aga has barly begun
to tap te immense archaeological potential of this
large and complex site. Our excavated sample of 31
m2, while sufficient to give some idea of the extent
and range of subsurface deposits and assemblages,
represents a very smaU porton of the estimated
21,000 m2 of buried, ceramic-bearing deposits
present in the To'aga area. It will take a much larger
effort, including the application of time-consuming
and costdy horizontal excavation methods, to begin
to exploit fully the potential ofthis site to reveal
unknown aspects of the Samoan past.
on
CHRONOLOGY AND
CULTURAL SEQUENCE
The suite of fourteen radiocarbon age determinations from the To'aga site (see Kirch, chapter 6)
constitutes the largest set of dates from a single
excavation locality anywhere in the Sanoan archipelago. These 14C dates define a two-millennium
long sequence of coastal tenace formation and
occupation beginning ca. 3600 cal B.P. and continuing witout pause up to ca. 1000 cal B.P. Although
no radiocarbon dates younger than about 1000 cal
B.P. were obtained durng our fieldwork, this does
not necessarily imply site abandonment dunng the
last millenium. Rather, the distribution of radiocarbon dates from To'aga reflects our emphasis on
testing and dating the earlier, deeply buried occupation deposits, as opposed to later surface features.
Given the presence of vanous aceramic ocwpation
pavements and mounds ('ili'ili pavements), grinding
sues, bua'i masi breadfruit storage pits, other
surface featues dtoughout the To'aga area, and the
presence of historic artifacts (see Hunt, chapter 3), it
seems likely that To'aga was in fact continuously
occupied by prehistoric Polynesians for a full hee
millennia ndeed, this is the only site ecorded to
date in the archipelago that appears to encapsulate
dre entire prehistoric record of the islands, from
initial settlement to historc contact
The iming of initial human settlement in
Westem Polynesia has been a matter of some
contention over the years and is direcdy relevant to
cunent debates over the rate of dispersal and colonization of the southwestem Pacific by the makers of
Lapita pottery (Kirch and Hunt 1988a,b; Spriggs
1990). Kirch (1988:244, table 48) summarized the
radiocarbon evidence for initial settlement of Fiji,
Tonga, and Samoa, based on such Lapita sites as
Natunuku, Yanuca, Lakeba (Site 197), Naigani, To.
2, and Mulifanua. "Almost all of these [14C] ages
cluster between about the twelfth and ninth centuies
B.C." (1988:244). The ordy site in Samoa which has
yielded dentate-stanped Early Eastem Lapita pottery
is Mulifanua, which has a single radiocarbon age of
1280-800 Cal B.C. (based on a A-R value of 100 ± 24,
and correcion for the oceanic reservoir effect; see
Leach and Green 1989).
The To'aga site provides additional new evidence for the timing of early, if not iuntial, settlement. (This qualification is necessary, because we
cannot be certain whether our systematic trascts
actually exposed the earliest occupation deposits in
the To'aga area. It is a distinct possibility that earlier
strata are preserved under the deep colluvium and
talus at te inland edge of the site, which could not
be penetrated without the use of heavy mechanical
equipment.) Our oldest 14C sample in direct, unquestionable association with ceramics is Beta-35601
from Unit 28, at 1308-930 cal B.C. This age is
essentially identical with that from the Mulifanua
site and entirely consistent with the suite of dates
from other early Westem Polynesian localities. Ihe
radiocarbon sample was associated with thinware
and thickware pottery, although no dentate-stanped
sherds were present. The absence of dentatestamping in this early context may simply be a
function of sampling error (only 8 percent of te
Mulifanua cermnics were decorated). Alternatively,
it is possible that a stratum bearing dentate-stamped
pottery lies furter inland, under the impenetrable
Synthesis and Interpretations
talus. A third possibility is that e decoraton of
pottery with dentate-stamping had ceased prior to the
colonization of Ofu. Only furter woik at the To'aga
site will be able to discriminate among these altemative hypotheses.
Despite the absence of "clasic" dentatestamped pottery, it is clear that Ofu Island and the
Manu'a Group were colonized within the same
general time penod, between ca. 1200-900 cal B.C.,
as were Eariy Eastem Lapita sites in Westem
Samoa, Niuatoputapu, Tongatapu, Futuna, and Fiji.
Given that Manu'a lies at the eastem extreme of the
Westem Polynesia region, this finding is quite
significant. It implies that the Lapita colonization of
the entire Fiji-Westem Polynesian region was
accomplished rapidly, with no appreciable lag
between sites at the westem and eastem boundaries.
The settlement of Manu'a by the close of te second
millennium B.C. also implies that Lapita populations
were poised at the threshold of the vast eastem
Pacific, thus raising again te question of wheher
ftere was tuly a "long pause" between fte setdement of Westem and Eastem Polynesia (Irwin 198 1;
Kirch 1986; Terrell 1986). If highly successful and
rapidly advancing island colonizers had moved as far
to the east as Manu'a by 1000 B.C., it seems strange
that they did not continue eastward into central
Polynesia (dte Cook, Society, and Austral archipelagoes) within the next two or thee hundred years.
Yet, at present we have no confinned human
habitaion sites in central Polynesia dated to the first
millennium B.C. (The earliest sites are stfill those in
te Marquesas, probably dating to as early as 400200 B.C., although this remains controversial, see
Kirch 1986.) In sum, while our radiocarbon dates
from To'aga confirm the pattem of rapid colonization of Westem Polynesia by the end of the second
millennium B.C., they also heighten the controversy
surrounding te subsequent phase of human settlement of the eastem Pacific.
Another chronological issue of some concem
within Westem Polynesia has been the timing of the
cessation of pottery manufacture and use. Poulsen's
initial claims for a long sequence of pottery manufacte in Tongatapu have now been revised, and
ceranic use appears to have ceased somewhere
between about 400 cal B.C. and cal A.D. 50 (Poulsen
1987:83; see also Groube 1971). Based on a large
number of 14C dates from several pottery-bearing
231
sites on 'Upolu, Green and Davidson (1974) put the
date of pottery disappearance in Samoa after A.D.
300, somewhat later than the Tongatapu sequence.
On Niuatoputapu Island, situated between Tonga
and Samoa, radiocarbon dates frm Sites NT-93 and
-100, containing plainwares, may indicate the
persistence of ceramic manufacture and use as late as
A.D. 800-900 (Kirch 1988:142, 246). Thus, while
tere is a consistent pattem of pottery decline and
eventual loss throughout Westem Polynesia, the
timing of this process may not be contemporaneous
in all locations.
At To'aga, the youngest radiocarbon date in
direct association with pottery (predominately
coarse, thickware) is Beta-35924, at cal A.D. 319473. Sanple Beta-26463, which comes from the
base of an acermnic midden deposit (in Unit 3), dates
tO Cal A.D. 561-663, while Beta-35600 from an
aceramic 'ili'ili pavement, dates to cal A.D. 694-943.
Thus, the radiometric evidence from To'aga indicates
that the cessation of pottery manufacture and use in
the Manu'a Group occurred during te fifth-sixth
centuries A.D. If Green and Davidson's dating of
pottery cessation in 'Upolu to approximately A.D.
300 iscorrect, there was a lag of 100-200 years in
the Manu'a Group. Such chronological differences
are certainly not surprising, especially in light of the
known ethnographic differences between Manu'a
and the rest of Samoa in historic times (Mead 1930,
see Kirch, chapter 2).
The radiometric chronology developed for the
To'aga site now allows the Manu'a Group to be
incorporated into a cultual sequence for the Samoan
archipelago as a whole. The initial settlement of
Manu'a was pene-contemporneous with that at
Mulifanua, suggesting that all islands of the archipelago were settled faidy rapidly at the close of the
second milleMium B.C. Parallel changes in material
culture during the following 1500 years are also
evidenced (see furiher discussion below), with the
cessation of ceramic manufacture in Manu'a possibly
lagging behind 'Upolu by one or two centures. One
chronological issue not addressed at To'aga is the
development of large monumental architecture (such
as star mounds and tombs). (We were not pennitted
by the landowners to excavate in or near the Tui Ofu
monumental complex situated within the To'aga
area.) We suspect, however, that as in Westem
Sanoa, these features will prove to date to the last
232
The To'aga Site
millenniun of Manu'a prehistory.
GEOARCHAEOLOGY AND
LANDSCAPE CHANGE
A major thstof ourrsearhprogam at To'aga
has been the application of a geoarchaeological and
geomorphological approach, in order to address
questions of site fomnation prcesses. Given that
most of fte archaeological deposits and features at
To'aga are deeply buried, a geomorphological
approach was essential even for such basic objectives as location and subsurface mapping of archaeological deposits. Our aim, however, has been to go
furthr and to explore the evidence for both natural
and culatal processes of landscape change at To'aga
over the thee millennia represented in the prhistonc sequence. To this end, a morphodynamic
model for the formation of the To'aga coastal terrace
was developed (see Kirch, chapter 4) and has been
explicitly tested on (1) stratigraphic relationships
(see Kirch and Hunt, chapter 5), (2) radiocarbon
dates (see Kirch, chapter 6), and (3) sedimentological analyses (see Kirch, Marning, and Tyler, chapter
7). Here we briefly summarize te major results of
this effort, with particular attention to the broader
implications for Westem Polynesian prehistory.
One specific result of our strategy of systematic
trasect excavations at To'aga was the areal definition of te extent of subsurface archaeological
deposits dating to the ceramic period. Although our
excavation sample is admitedly small, the highly
consistent distribution of pottery-bearing deposits in
a narrow zone at the base of the talus slope allows us
to predict with considerable accuracy the probable
extent of these subsurface deposits. These early
occupation layers are all associated with former
beach-ridge environments, at a time when the
shoreline was much closer to the cliff and talus, and
when sea level was apparently at a +1-2 m stand.
The infenred distribution of pottery-bearing occupation deposits (dating to between ca. 3000-1900 cal
B.P.) is plotted in figure 15.1, which also depicts the
probable location of the Ofu shoreline during the
first milleniuM B.C. It is important to stress,
however, that we were unable to determine precisely
the inland boundary of these pottery-bearing deposits, as this lies somewhere under the modem talus
slope. Ling heavy mechanical equipmnen, it was
impossible for us to penerate fte compact mass of
colluvium and talus boulders (estimated to be as
thick as 10-15 m in places) which caps the inland
portons of these early layers. Probably these
deposits do not extend any farther inland than 2040
m from our inland-most test units.
Based on the distribution depicted in figure 15.1,
te total area of buried subsurface archaeological
deposits containing ceramics covers a minimum of
21,000 M2, and a maximum of 35,000 M2. By any
Polynesian standards, this is a large site with considerable potential for future horizontal excavations.
As noted earlier in this volume (see Kirch,
chapter 2), the coastal terace nnning along the
southem side of Ofu from To'aga to Fa'ala'aga
comprises an important land use and resource zone
for the island's Polynesian occupants. Given the
island's very steep and rugged topography, this
stretch of flat land is one of the few areas suitable for
intensive arborcultur and for habitation. Yet, our
geoarchaeological studies at To'aga clearly demonstrate that this coastal tenrace is a highly dynamic
accumulation form which developed into its present
configuration only within the past 2,000 years-well
after initial human colonization of the island. At the
time of initial settlement, the coastal tenrace was very
nanrow, consisting of little more than a beach ridge
situated directly beneath fte steep oveihanging cliffs
and talus. A phase of active progradation of this
shoreline did not commence until about 1.9 kyr B.P.
This prgradation appean to have been initiated by a
rapid fall in sea level (presumably a eustatic fall
which is evidenced over much of fte southwestem
Pacific at this time, see chapter 4) equalling or
slightly exceeding the rate of local subsidence, and
thus exposing fte reef flat to erosion and storm
surges. The littoral, calcareous contribution to the
sediment budget was thereby increased, producing a
sufficient volume of sediment to prograde the
shoreline between 40-100 meters from its location
prior to 1.9 kyr B.P. Following th classification of
"accumulation forms" proposed by Zenkovich
(1971:95-97, fig. 4.1), the To'aga tenrace is an
attached form, specifically type c, a "'tefface formed
by the infiling of a concavity (supplied laterally)."
In this case, the concavity was formed by the marine
cliffinland ofthe site, which was deepest at the
southwestem end of the coastal stip. The infilling
Synthesis and Interpretations
233
C)
\
*I
a
4*
.0
*bO
8
4o
H0
*r
0
C
9o
w
EC0
_w
[I
0
Ct
0
E
0
0
.O
234
The To'aga Site
and progradation of the coastal terrace proceeded
from southwest to norheast This sequence of
infifling is confirmed by our tansect excavations and
conrelates withi the early pottery-bearing deposits
(daing to 3-1.9 ky B.P.) which are confined to th
southwesem part of the To'aga coastal tenre. The
norteastem end of the terrace, in the vicinity of
Fa'ala'aga Cransect 17), remained a high-energy
beach until relatively recently.
The morphodynamic model developed and
tested for the To'aga site has wider implications for
coastal archaeology elsewhere in Samoa and, indeed,
on other volcanic oceanic islands (i.e., ose situated
on te Pacific Plate). In Manu'a, we would predict
that coastl terraces on Olosega and Ta'u islands (for
example, the Faga and Saua aeas on Ta'u Island)
will prove to have similar dynamic geomorphological histories to that evidenced at To'aga. lTus, any
effort to discover occupation sites dating to the
period between ca. 3-2 kyr B.P. will rquire the use
of subsurface testing to locate the early beach ridge
environments that are prdictably buried under later
prograded sediments. We believe that it is likely that
such bured sites do exist on Olosega and Ta'u.
Indeed, a ceramic-bearing site in precisely this kind
of bured beach ridge envirnment was identified by
Hunt and Kirch at the base of the marine cliff behind
Ta'u Village during our 1986 rconnaissance survey
(1988:166-67).
The morphodynamic model of the To'aga site is
also likely to apply, pediaps with minor modifications, to Tutuila Island. (One modification to the
model which may be required is the rate of subsidence. Tutuila is somewhat older than Manu'a, and
thus may already have passed trugh its phase of
rapid subsidence due to point-loading of the thin
oceanic crust.) Claik and Herdrich (1988) demonstrat that the 'Aoa Valey has been substantially infilled since initial human occupation, indicated by a
pottery-bearing site situated along the interior edge
of the modem valley floor and by other geomorphological signs of a former shoreline now well inland
of the present coast. They listed several hypotheses
to account for this sequence of infilling (1988:17375), including "a lowering of sea level" of about 1-2
m. We would suggest that this hypothesis is the
most likely of the alternaives they list, although the
contribution of terrigenous sediment to the 'Aoa
sediment budget (resulting from human-induced
forest clearance and erosion in the interior valley
slopes) was doubdess greater ta at To'aga. In any
event, the failure of early archaeological efforts to
locate ceramic-bearing sites on Tubuila (e.g., Kikuchi
1963; Emory and Sinoto 1965; Frost 1978) was due
to the lack of a geomorphologically informed
approach to site discovery. It is probable that most if
not all of the valley floors as well as the coastal
tenrc ofTutuila Lsland have undergone significant
infifling and progradation during the past two
millennia Thus, early coastal sites are unlikely to be
exposed on the surface. It will be essential to woik
out local geomorphological sequences of infilling
and progradation as an integral part of archaeological
survey in these cases.
We also predict that similar sequences of
progradation and burial of early archaeological sites
will be found trughout many of the volcanic high
islands of centr Polynesia, particularly in the
southem Cooks, Society Islands, and probably the
Austals. These are all "hot spot" linear volcanic
chains situated on te Pacific Plate, which typically
undergo point-loading induced subsidence (Menard
1986:95-99). Steams (1978:286), for example,
points to geological and geomorphological evidence
for rapid subsidence of the Society Islands during the
late Pleistocene (and probably continuing into the
Holocene). Furtermore, there is now considerable
radiometrically dated evidence for a +1-2 m higher
sea level at ca. 4-2 kyr B.P. throughout French
Polynesia with a rapid fall to modem level after
about 2 kyr B.P. (see Kirch, chapter 4). In short, both
ofthe major controing processes that resulted in
coastal terrace construcdon on Ofu (subsidence and
sea level change) were probably also operating in
central Polynesia. This suggests that early occupation sites will not be easy to discover using conventional archaeological surface survey methods.
Indeed, the presence of buried, and even partially
submerged, archaeological deposits is known for
Huahine (Sinoto 1979) and for Mo'orea (Green et al.
1967) in the Society Islands. On Mo'omea, Lepofsky
recently discovered anaerobically prerved coconuts buried under 2 m of recent alluvium in the
interor of the Opunohu Valley (Lepofsky, Haries,
and Kellun 1992).
These problems of site survey and discovery
raise serious issues for Eastem Polynesian prehistory. Specifically, given the probability of deep
Synthesis and Interpretations
burial of early coastal sites, te archaeological record
for central Polynesia as curnently defined is likely to
be highly biased toward later prehistoric sites. It is
entirely possible that the early phases of occupation
in the Societies and Cooks have yet to be discovered.
Just how far back in time we may be able to extend
these chronologies after a program of geomorphologically informed, systematic subsurface sampling
is launched is impossible to say. It seems conceivable, however, that the "long pause" currently
identified between the settlement of Westem
Polynesia and the movement of people into Eastern
Polynesia (Irwin 1981; Kirch 1986; Terrell 1986)
may prove to be an artifact of archaeological sampling bias. Only a concerted effort to work out the
late Holocene dynamics of coastal landforms in
central Eastem Polynesia, and to search for early
sites in these depositional contexts, will provide a
definitive answer to this problem.
Although the sequence of coastal tenrace
accumulation at To'aga was largely controlled
rough the interaction of sea level change and
subsidence, it would be a mistake to attribute the
entire pattem of landscape change to natural processes. At To'aga, as throughout much of te
tropical Pacific, humans have played a major role in
modifying and shaping island environments (see
Kirch 1982, 1983, 1984:123-51). One process
which is probably due largely, if not wholly, to
human interference in the Ofu ecosystem is the
erosion and deposition of substantial volumes of
colluvium. Our wansect excavations revealed a
consistent pattem over most of the To'aga area
(except at Transect 17 which has an almost exclusively boulder talus underlying the narrow nidge) of
increasing rates of colluvial deposition onto the
coastal terrace after about 2 kyr B.P. This deposition
occurred as a series of small, overlapping colluvial
fans emanating out of intermittent watercourses or
small ravines inland of the site. These fans cap the
early calcareous beach-ridge deposits containing
pottery-bearing occupations, and in our inland-most
test units frequently exceeded 1 m in depth. The
fans rapidly pinch out as they extend onto the coastal
tenrace, with the larger angular clastics decreasing in
frequency, and thin "tongues" of fine-grained silt
and clay extending out onto the coastal flat. That
humans played a key role in initiating this incrased
rate of erosion and deposition of colluvium after
235
about 2 kyr B.P. is suggested by the presence of
charcoal in most of these colluvial sediments.
Natural forest fires are an extremely rare occurrnce
in te humid tropics, and thus the presence of
charcoal flecking in these sediments is almost
certainly a signal of human buming (see Kirch and
Yen 1982:154, 351-52 for a discussion of this
phenomenon on Tikopia Island). This buming most
likely was associated with forest clearance for
agriculture, specifically shifting cultivation of root
and tuber crops on the steep hillsides inland of the
site. Once the native forest was disturbed and
opened up, erosion of the youthfil volcanic soils on
the steep slopes would have increased drnadcally.
The effects of this erosion and deposition of
colluvium on the newly prograded coastal terrace
were by no means negative from the human viewpoint of potential land use. Rather, the addition of
highly fertile, young volcanic sediments to the welldrained calcareous terrace created a mixed edaphic
environment that was more suited to the cultivation
of tuber and tree crops (such as Dioscorea yams,
Alocasia aroids, breadfruit, and coconut), than either
the volcanic or calcareous sediments by themselves.
Hence, as the coastal terrace itself prograded and
expanded in area after about 2 kyr B.P., its potential
as a zone of intensive agricultural production was
significantly enhanced by the human-induced
deposition of volcanic sediments. We can only
speculate that the creation of this coastal zone,
edaphically well-suited to intensive agriculture as a
result of human actions, may have occurred at a time
when the island's population was likely to have been
increasing after a millennium or so of settlement.
Here too, it may not be at all coincidental that the
traditional site of chiefly power on Ofu Island-the
Tui Ofu monument complex-is situated in the
approximate center of this rich resource zone (see
Hunt, chapter 3). The emerging chiefly polity of the
island could be expected to have exercised its hegemony by seizing control of this newly created and
highly productive resource zone.
Other indications of human impacts on te Ofu
Island ecosystem in the archaeological record at
To'aga include te extinction or extirpation of bird
populations (see Steadman, chapter 14) and the
intduction of adventive species of terrestrial
molluscs associated with Polynesian horficulture
(see Kirch, chapter 8). Consideration of this evi-
236
The To'aga Site
dence, however, will be deferred to the discussion of
te prehistoric subsistence systen in a later section
of this chapter.
Befoe closing this discussion of the sequence of
landscape change at the To'aga site some brief
comparisons with other documented sequences frm
tropical Polynesia are wamnted. The pattems of
change that dramatically wansformed the lowland
environment of To'aga between 3 kyr B.P. and e
present are not unique to Ofu Island. Similar
sequences have been atested for other islands in the
southwestem Pacific, including Tikopia (Kirch and
Yen 1982), Aneityum (Spriggs 1986), Lakeba
(Hughes et al. 1979; Bayliss-Smith et al. 1988),
Futuna (Kirch 1975, forthcoming), Mangaia (Kirch
et al. 1992) and Niuatoputapu (Kirch 1988). The
specific similarties include: (1) the progradation of
cosal lowlands between ca 3-1.5 kyr B.P., presumably due to sea-level changes in the area; (2) humaninduced forest clearance and erosion resulting from
shifting cultivation; (3) deposition of alluvial and
colluvial sediments in lowland landforms; and (4)
edaphic enhancement of lowland environments as a
consequence of (3). Also attested in these sequences
are human impacts on the endemic and indigenous
biota. In short, the To'aga site adds yet another case
to the growing catalog of significant human modifications to the island ecosystems of the central
Pacific.
ANCESTRAL POLYNESIAN CULTURE
The derivation of Polynesian cultures from a
Lapita ancestor is now well attested in the archaeological sequences of Westem Polynesia. Kirch
(1984) and Kirch and Green (1987) have used the
tenn "Ancestral Polynesian Culture" (or "Society,"
depending upon the frame of reference) for the
culture which emerged in the archipelagoes of
Westem Polynesia during the middle of the first
millennium B.C. The reconstruction of Ancestral
Polynesian Culture (which doubtess was not
unifom troughout the region) is arguably a key
task for Polynesian prehistorians, because it provides
the baseline against which subsequent cultural
divergence, evolution, or transformation can be
measured. Such reconstruction may be attempted
from at least three different, but complementary,
sources of evidence: comparative ehnography,
historical linguistics, and archaeology. Green (1986)
has oudined this "tiangulation" approach in gater
detail using the example of Ancestal Polynesian
settlement systans.
The To'aga site is potenially a major source of
archaeological evidence for the reconstruction of
Ancestral Polynesian Culture, because the site's
stratigraphic sequence spans the whole of the first
millennium B.C., the period duMg which tis
Ancestral culue emerged out of its Lapita predecessor. Because our objectives and temporal-fiscal
limitations necessitated a testing stategy of excavation, our present evidence from To'aga is primarily
in the realms of material culure and of faunal
material bearing on dte subsistence economy. In the
fuure, however, expanded horizontal excavations at
To'aga may reveal much new evidence regarding the
structure and spatial organization of Ancestl
Polynesian settlements.
Ceramic studies have been an important part of
understanding prehistory and culure change in
Samoa and West Polynesia. The assemblage from
the To'aga excavations is especially significant as it
is large, excavated from a well-stratified, well-dat
deposit represents the full duration of pottery
manufacture in Samoa, and has been studied in detail
(chapter 15).
Green (1974) perfonned the first detailed
analysis of Samoan ceramics from the SU-Sa-3 site
at Sasoa'a, Upolu. He defined one "type" with two
varieties of plainware: a coarse-tempered thickware,
and a fine-tempered thinware. Green (1974) followed a methodology for classification developed in
the culture-historical paradigm of American archaeology (i.e., especially in the work of J. Ford, A.
Kreiger, and I. Rouse; see Dunnell 1986a, 1986b).
Contrary to crticisms from proponents of the New
Archaeology (1950s-60s), classification fonnulated
by culture historians is deductive, problem-oriented,
and based largely on a paradigmatic structure
(Dunnell 1971). Culture historians understood the
explanatory meaning of variability-their methodology was founded in a materialist ontology (Dunnell
1986b). The debate between Ford and Spaulding in
the mid-1950s is illustrative of the contrasts between
culture history and the aspirations of the New
Archaeology (see Dunnell 1986b).
Following the culture historians' lead, Geen
(1974) defined two classes of pottery by the
Synthesis and Interpretations
combination of dimensions in a paradigmatic
structre (see Dunnell 1971) for thickness and
temper size. Thick- and thinware, as vanreties of one
type (in the type-variety system; see Dunnell
1986b: 174), are ideational units or classes. Ideational units "are tools of our construction the
purpose of which is to allow us to recognize and
describe these things about which empirical claims
are made" (Dunnell 1986b: 151). Classes are based
on intentional definitions, where a specific set of
features fonns the necessary and sufficient conditions formembership in a unit (Dunnell 1971:16).
Once Geen (1974) defined classes for analytic
purposes, he described the empirical variability
within them. His descriptions include observations
such as color, paste texture, temper composition, and
surface treatments-attributes that were not part of
the defining criteria. Dunnell (1971) has distinguished definitions from descrptions as the contrast
between class and group. Classes comprise lists of
criteria; groups are sets of things (Dunnell
1986b:181). Thus, defining classes and describing
groups (empirical entities), which he called "categories," characterizes Green's analytic approach.
In his analysis, Green (1974) attempted to use
the vessel, rather an sherds, as the basic counting
unit comprising the archaeological assemblage. This
innovative approach to establish a minimum number
of vessels based on similarity in sherd color, temper,
and other attributes has been attempted by others
(e.g., Brose 1970, Sullivan 1983). However, as
Feathers (1990:139) explains, attempts to transfonn
sherds into vessels are met with some problems.
And, such attempts are not always necessary, since
variability between assemblages will be reflected in
attributes of sherds as well as vessels. As Feathers
(1990:13940) puts it, "sherd information is not
inferior to vessel infornation. It is just different
And because of the problematic representation of
vessels archaeologically, sherd data is [sic.] te best
source of assemblage infonnation." Green's (1974)
minimum number of vessels approach does not
lessen the usefulness of his analysis. While later
statistical analyses (e.g., Clark and Herdrich 1988;
Hunt and Eikelens, chapter 9) revealed problems in
the intuitive definition of thick- and thinware,
Green's rich descriptions provide data suitable for
comparative analysis.
In a subsequent study, Smith (1976) took
237
assemblages from the Mulifanua Lapita site, Jane's
Canp, and the Paradise site for ceramic analysis.
Smith used principal components analysis, a multivariate grouping procedure, to (1) evaluate Green's
conclusions, (2) compare Lapita with later Samoan
cermnics, and (3) in his words, "attempt a meaningful and useful classification of the pesent Samoan
ceramic material using a wide range of both stylistic
and techological variables" (1976:83-84). Focusing on attributes that appeared to vary temporally,
Smith (1976:86) initially soiled sherds into three
kinds based on thickness and paste texture: a thick
coarse-textured (tempered?) ware; a thinner, finertextured ware; and an extremely fine textured ware.
The distinctions of thickness and paste texture
(temper size?) appear identical to those made by
Green (1974), with the addition of a finer ware (no
doubt due to earlier ceramics in te sample).
Smith's (1976) principal component analysis,
like Green's pottery descriptions, is a means to
delineate the empirical varability of groups.
Smith's principal components analysis is not a
classificatory tool, however, but a method that
analyzes the grouping tendency of the sherds
studied. His results show that pottery from these
assemblages is relatively homogeneous with respect
to the attributes of paste and color. The variables of
paste and color are represented in components I and
n of his analysis. A third component includes
thickness, exterior and interior evenness (variance in
thickness?), "filler type" (temper composition?), and
"filler size" (temper size?) (Smith 1976:90). While
Smith (1976:90) states that "it does not appear
possible to interpret this combination of variables in
any meaningful fashion," these variables clearly
relate to the criteria used to define thick- and
thinwares (e.g., Green 1974; Hunt and Erkelens,
chapter 9).
In analytic terms, Smith (1976) conflates the
distinction of group and clas (Dumell 1971). He
has not created a classification, but has only shown
the grouping tendency of sherds from three assemblages. Urdike paradigmatic classes, groups (as
statistical summaries) change with the addition of
every new case (sherd or assemblage). Smith
(1976:92) proposes that his groups forn the basis for
a new "typology" of Samoan ceramics. He points
out, however, that his "types" would serve as
descriptive categories only as they do not correspond
238
The To'aga Site
to either chronology or spatial distribution. Smith
(1976: 93) implicitly recognizes the problem of
confladng class and group as he notes that the
discreteness of "'types" will disappear with additonal
analyses of Samnoan ceramics (see Dunnell 1971,
especially pp. 87-110, figure 8). Clearly, analysis of
groups does not provide the basis for a useful
classification. Istead statisdcal grouping techniques
allow one to examine variability among defined
classes.
Holmer (1980) re-analyzed sherds from
Mulifanua, Jane's Camp, and te Paradise site with
sherds from two new sites, Potusa and Falemoa, both
on Manono Lsland. He attempted factor analysis on
unspecified variables-apparently similar or identical to those used by Smith (1976). When factor
analysis failed to produce groups (clusters), a new
strategy was attempted. Holmer subjectively
(implicitly) classified sherds into seven "types."
Varables were then selected for observation/
measurement, and data from the analysis of sherds
was used in discriminant funcdon analysis. In this
analytic procedure, Holmer (1980) simply confirms
(statistically) his "subjective" sorting criteria (implicit a priori class definitions). As with Smith's
(1976) analysis, the ideational (classes) and empirical entities (groups) are confused. Holmer (1980) is
thus left with 'lypes" that will constantly change
with every new case analyzed.
These problems with confusing group and class,
evident in Smith (1976) and Holmer's (1980) work,
are not merely an academic issue. The conflation of
these observational and analydc steps is the reason
why Smith and Holmer did not succeed in producing
classificatory systems, contrary to their stated
objectives. This is because: (1) Analysis of objects
is necessarily based on an a priori (most often
implicit) classificatory system (i.e., to observe "x" is
to observe "x" as a case of something). Grouping
procedures are data manipulations performed on
unanalyzed and implicit classifications; they are
inductive and often formulated as "problem-free."
(2) Groups are based on empirical sets (e.g., statistical summaries) that change with every case added,
thus an object cannot be assigned to a pre-existing
group without altering the "definition" of the unit
(3) Grouping provides descriptions of phenomena,
but not definitions stipulating necessary and sufficient conditions for membership (althoughpost hoc
definitions might be extracted). Group membership
is based on similarity which vanes in degree. This
means that individual objects in a group may share
many, some, or in extreme cases, no traits in common (see Dunnell 1971:fig. 8).
Clark and Herdrich (1988) recently called
attention to the problem of Green's (1974) original
formulation of coarse-tempered thickware and finetempered thinware. They point to the lack of explicit
crteria for what is thick and thin, fine or coarse.
Clark and Herdrich (1988) illustrate the variability of
these dimensions in an assemblage from 'Aoa,
eastem Tutuila. While preliminary in nature, their
examination of ceramics points to the importance of
classification as a means to document change in te
Sanoan sequence.
The To'aga ceramic study reported in this
volume is the most intensive analysis of assemblage
from Samoa to date. The analytic protocol was
designed to support the definition of numerous
classes (i.e., from two or more of the dimensions)
deduced to address particular research questions of
the assemblage. While the To'aga assemblage is
especially well-studied, much wor remains to
answer the larger questions posed by Hunt and
Erkelens (chapter 9) for ceramic evolution. Results
from To'aga show tha thickware is present in the
earliest deposits, and its abundance (in actual
numbers) over time is relatively stable. Thinware is
never clearly dominant at To'aga. Thinware declines
in real and relative abundance over time but persists
perhaps as long as pottery production itself. At
To'aga, as elsewhere in Samoa, pottery declines in
abundance early in the Christian era and then
disappears entirely.
Results from ceramic compositional analyses
show that the bulk of pottery, including both thickand thinwares, and some carved paddle-impressed
ceramics, was produced locally with colluvial "selftempered" clay source(s) from Leolo Ridge on Ofu.
Red-slipped pottery from To'aga does not match the
local colluvial clays presently known. Based on this,
the red-slipped ware may have an exotic provenance,
amving on Ofu through inter-island exchange.
Finally, results also suggest that the diversity of raw
materials declined over time. Such a pattem if
substantiated with furher work, suggests changes in
availability and/or procurement of raw material. A
decline and eventual end to inter-island exchange of
Synthesis and Interpretations
ceramics might also be hypothesized.
The To'aga assemblage, like most others from
Sanoa, is simple in foni and includes very little
decoration. Only bowls are represented. Decoration
is restricted to impressing and notching on te lip,
and on body sherds, red-slip, carved paddle-impression, and incision. This simplicity stands in marked
contrast to Lapita assemblages of comparable age
from Mulifanua, 'Upolu, and assemblages known
from Fuuna, Tonga, and Fiji. Isolation of Manu'a
from communities beyond Samoa might account for
this stylistic divergence.
The To'aga excavations also yielded a small but
important set of stone adzes in association with the
ceramics just discussed. Significantly, these are all
variants of the plano-convex sectioned Type V adz
described by Green and Davidson (1969). A
dominance of Type V adzes in ceramic-bearing
contexts in Westem Samoa was noted by Green
(1974:257-58). The Manu'a results confirm this
pattem for the eastem part of the Samoan archipelago. Indeed, Type V appears to have been a
widespread and common form throughout the
Ancestral Polynesian region, given the presence of
this form in sites in Futuna (Kirch 1981),
Niutoputapu (Kirch 1988:192), and Tongatapu
(Poulsen 1987:170). Type V was dropped frm the
Samoan adz inventory early in the first millennium
A.D., and our surface collections from Manu'a are
dominated by adzes with quadrangular or trapezoidal
cross sections. One of these trapezoidal forms was
recovered from a late, aceramic depositional context
in Unit 3; this particular specimen appears to have
been manufactured at the Tatagamatau quarry on
Tutila Island (see Weisler, chapter 12).
Pottery and stone adzes are the best documented
classes of Ancestral Polynesian portable artifacts.
For both of these artifacts, the development of
uniquely Ancestral Polynesian forms out of Lapita
prototypes has been archaeologically demonstrated
(e.g., Green 1971, 1974). This is not the case,
however, with another important Polynesian artifact
class: the one-piece fishhook. When archaeologists
first commenced stratigraphic excavations in the
Eastern Polynesian archipelagoes of Hawaii, the
Marquesas, Societies, and New Zealand in the 1950s
and 60s, fishhooks proved to be among tX most
ubiquitous arfifacts. Indeed, in the absence of
pottery, Polynesian archaeologists applied their skills
239
at classification and seriation to fishing gear, in an
effort to establish chronological sequences (e.g.,
Emory, Bonk, and Sinoto 1959; Suggs 1961). Thlus,
when efforts were directed at the Westem
Polynesian islands of Tonga and Samoa, the initial
expectation was that similarly rich assemblages of
fishing gear would be recovered. These expecations
were quickly thwarted by an almost complete
absence of fishhooks in Westem Polynesian sites.
Poulsen recovered only one "certain specimen" of
one-piece hook in his excavations of six sites on
Tongatapu (1987:186), while the major Samoan
archaeological programme of Green and Davidson
(1969, 1974) recovered but a single fragment of a
Turbo-shell hook from the Lotofaga midden (Green
and Davidson 1969, pl. 23). Kirch's excavations on
Niutaputapu fared only slightly better, with four onepiece hooks out of thirteen sites sampled (1988:204,
fig. 124).
The extreme paucity of fishing gear in Westem
Polynesian sites-in contrast with the typically high
density of fishhooks in Eastem Polynesian sitesraised a number of questions. Given the extensive
scope of Westem Polynesian excavations, sampling
enfor alone can be ruled out. Rather, it appears that
te rarity of hooks in early Westem Polynesian sites
(i.e., those dating to the Ancestral Polynesian period)
is an accurate reflection of the relative unimportance
of angling gear (Green 1986:131). The faunal
assemblages from these sites, however, clearly
indicate that inshore fishing was a major component
of the subsistence economy. Were hooks being
made of pershable materials (such as wood) and
hence not preserved in the archaeological record?
Were other fishing strategies, such as netting,
spearing, or poisoning, prefened over angling by
Ancestal Polynesian fishermen?
To'aga is the first Ancestral Polynesian site to
produce a large assemblage of one-piece fishing
gear, and thus demonstrates anoter kind of variability in eady Polynesian culture. As described in
chapter 11, a total of twenty-eight whole or partial
Turbo-shell hooks were recovered from our excavations, along with another thirty-one prefonns or tabs.
This is a fishhook density level much more in
keeping with Eastem Polynesian sites. Why should
To'aga produce such an assemblage of one-piece
hooks when other Samoan and Westem Polynesian
sites are devoid of these aftifacts? We suggest that
240
the
The To'aga Site
answer lies in the differenidal marine environ-
ments of the vanous islands. The Westem Samoan
islands of 'Upolu and Savai'i, as well as
Niuatoputapu and Tongatapu, are all characterized
by extensive barrier reef and lagoon ecosystens. In
these kinds of coastal environments, the most
effective fishing stategies are usually those involving nets (seines, dip nets, nets used with weirs, and
other techniques). This was well documented, for
example, in ethoarchaeological studies of contenporary fishing on Niuatoputapu Island (Kirch and
Dye 1979; Dye 1980). In contastL the marine
environment of Ofu is that of a relatively nanow
fringing reef, lacking a broad protected lagoon. In
such fringing reef environments, angling becomes a
far more significant fishing stategy, to exploit the
dominant fish populations of the reef crest and outer
slope.
The hypothesis that a greater emphasis on
angling gear is conrelated with frnging mef (as
opposed to barrier reef-lagoon) environments
receives some support from fte general Oceanic
picre of archaeological fishing gear distribution.
For exanple, both Tikopia (Kirch and Yen 1982)
and Anuta (Kirch and Rosendahl 1973)-small high
islands with narrow fringing reefs-yielded high
frquencies of one-piece fishhooks in teir archaeological sites. The same is true of Hawai'i and the
Marquesas, where reefs are finging or even lacking
altogether. On the other hand, the Society Islands
which have extensive lagoons have produced
relatively low densities of fishing gear in comparison
with other Eastem Polynesian sites. Consequently,
we would argue that the Manu'a Group is one area
within Westem Polynesia where the marne ecological conditions favored the use of angling gear.
The Turbo-shell fishhook assemblage from
To'aga is of some interest from a morphologicalstylistic perspective, in addition to its ecologicalfuncdonal implications. As noted in chapter 11,
several of the hooks exhibit morphological features
similar to those in eatly Eastem Polynesian fishhooks. These include the strongly incurved or
"bent" shank and the single-notched, line-lashing
devices. Thus, the To'aga hooks can madily be
identified as a "prototype" stage from which the
greater diversity of Eastem Polynesian forms was
subsequently developed.
THE SUBSISTENCE
ECONOMY OF EARLY SAMOA
The To'aga excavations produced one of te
largest and best preserved faunal assemblages ever
recovered from a Westem Polynesian site: 10,209
vertebrate bones and approximately 166.5 kg of
invertebrate materials. Largely due to poor prsrvation, most previously excavated Samoan sites
yielded very poor faunal collections. In Westem
Samnoa, only the late prehistoric Lotofaga midden
reported by Davidson (1969) and the three coastal
sites analyzed by Janetski (1976, 1980)-ceramicbearing Potusa, Falemoa, and Jane's Camp-have
well-preserved vertebrate and invertebrate faunal
materials. lTus, the To'aga materials provide the
first extensive sample of fauna frm well-stratified
and dated contexts spanning the first half of the
Samoan sequence. The faunal data have been
presented and analyzed by Nagaoka and Steadman
in chapters 13 and 14, respectively. Here we expand
on their analyses with several general observations
and with comparisons to other Samoan and Westem
Polynesian sites.
One problem that has concerned archaeologists
in Westem Polynesia is whether the Polynesian triad
of domestic animals-pig, dog, and chicken-was
introduced at the time of initial settlement and
colonization (Groube 1971; Hunt 1981; Best 1984;
Kirch 1979, 1988). At To'aga, only the chicken
(Gallus galus) is well represented in our faunal
suites. Chicken is actually the most frequent bird
species represented in the avifaunal material, with 16
NISP (see Steadman, chapter 14). Chicken bones
were especially well represented in the Layer III
deposits in Units 20123, dating to ca. 2800-2300 cal
B.P. Pig, however, is unambiguously represented
only in later contexts (in Layer I of Unit 17). Some
of the unidentifiable mammal bone from eadier
strata may indeed be of pig or dog-or both-so that
the absence of pig and dog in early contexts is not
certain. Nonetheless, given the large vertebrate
faunal sample and excellent preservation, it is cerain
that neither of these domestic animals was ever
present in large numbers at the To'aga site.
Another adventive species introduced (presumably as an inadvertent "stowaway" on voyaging
canoes) at the time of initial colonization is the
Synthesis and Interpretations
Polynesian or Pacific rat, Rattus exulans. This
species is ubiquitous in the To'aga strata and occurs
in the earliest dated depositional context (Layer HID
in Units l5129/30). Tate (1951) discusses the very
widespread dispersal of this synanthropic species.
Another group of human-intrduced organisms
appearing in the early To'aga deposits is the set of
five synantrpic terrestrial molluscs discussed in
detail in chapter 8: Assiminea cf. nitda, Lanellidea
pusilla, Gastrocopta pediculus, Liardetia samoensis,
and Lauellaxis gracilis. These species are closely
commensal with humans, their preferred habitats
being gardens and disurbed environments adjacent
to habitation sites. The species are all minutevisible to the human eye only on close inspectionand can only have been tansported inadvertently.
Many of these species have also been identified from
early, Lapita-associated archaeological contexts on
Niuatoputapu Island (Kirch 1988:233-35) and on
Tikopia (Kirch and Yen 1982:308-309). As discussed in chapter 8, the most likely mechanism for
their inter-island transfer, and intrduction to the
Manu'a Islands, was with economic plants and
adhering soil media. In this regard, these snails
provide indirect evidence for early plant introductions to the island. Indeed, in the absence of direct
ethnobotanical evidence for cultigens, the suite of
synanthDpic snails is the best clue that the early
Polynesian colonists intrduced a complex of
economic plants to the island, along with the domestic chicken (and possibly also pigs or dogs). Future
excavations at To'aga should test this hypothesis
through identification of charcoal, carbonized
parenchyma, and other carbonized plant materials
from earth ovens, hearths, and other stratigraphic
contexts. Recent developments in the identification
of such carbonized materals by J. Hather (Institute
of Archaeology, London; pers. comm., 1991) and
others, not available to us at the time the To'aga
excavations were undertaken, now make the possibilities for such paleoethnobotanical studies possible.
Two kinds of larger marine animals are represented in the vertebrate faunal collections: sea
urtes and unidentified marine mammal. The sea
trtles probably consist mostly (if not exclusively) of
the Green Sea Turtle, Chelonia mydas, but definite
identifications on the post-cranial skeleton are
virtually impossible. Turtle bones were fairly
common and were especially frequent in Layer IIIB
241
of Units 20123 and Layer IIB of Unit 19. The
marine mammal bone is most likely from one or
more species of porpoise. A relatively high frequency (18 NISP) of marine mammal bone was
found in the early Layer IIIC deposit in Units 15/29/
30.
The bird bones are of particular interest, for they
reveal exinctions and extirpations consistent with a
pattem of avifaunal change from early sites thoughout Polynesia (Steadman 1989; Steadman, Pahlavan,
and Kirch 1990, Steadman and Kirch 1990). An
unexpected discovery was the presence of two
Megapodius sp. bones from Layer IIID of Units 15/
29/30, the oldest dated layer at the site. Megapodes
were not fonnerly known to have been present in
Samoa, and this find thus represents an eastern
extension of the prehistoric range of this taxon.
Given the restriction of this taxon to the earliest
stratum, it is likely that the species was rapidly
overexploited-to the point of extinction-by the
early colonizers of Ofu. Also striking is the presence
of bones of six species of seabirds which no longer
occur on Ofu Island, including Puwinus pacificus,
Puinus lherninieri, Puffinus griseus, Pterodroma
rostrata, Pterodroma sp., and Sula sula. The loss of
these species from the island within the span of
human occupation most likely reflects both direct
predation by humans and habitat disturbance.
Ninety-four percent of the vertebrate fauna from
To'aga consists of fishbone, of which 2,229 NISP
were identifiable to family-level taxa (see Nagaoka,
chapter 13). Although there are frequency differences in taxa witiin different excavation units, there
is remarkable consistency overall in the rank-order
dominance of particular fishes. Four families
dominate the faunal assemblages: Diodontidae,
Serranidae, Acanthuridae, and Holocentridae. These
families include numerous species, most of which
occur on the reef flat or immediately off the reef
edge. They may be taken with a variety of fishing
strategies including netting, spearing, poisoning, and
angling. The acanthurids, holocentrids, and serranids especially, can be taken with hook-and-line,
and it is very likely that the small one-piece Turboshell fishhooks recovered from the To'aga site were
used to capture these taxa. Of considerable interest
is the high frequency of Diodontidae (primarly
Diodon hystrix), the porcupinefish. These fishes are
known to carry tetradontoxin which can cause severe
242
The To'aga Site
illness or even death when ingested by humans.
That such a dangerous fish should dominate the
To'aga faunal assemblages is cunous, although not
inconsistent with pattems in other early Pacific sites
(see Green 1986:132). InTikopia, Diodon hystrix
was extremely plentiful in the early middens (Kirch
and Yen 1982: 292, table 42), as it was also in the
Early Eastem Lapita site of NT-90 on Niuatoputapu
Island (Kirch 1988:223, table 29).
A second tier of fish taxa, in terms of fteir rankorder abundances, comprses the following families:
Scaiidae (parrotfish), Carangidae (jacks), Labridae
(wrsses), Lutjanidae (snappers), Muraenidae
(moray eels), Balistidae (triggerfish), and Osaciidae
(boxfish). Again, these are all inshore, reef or reef
edge fishes, represented by a large number of
species. A varety of fishing strategies were doubtless employed to take these fishes. A number of
other taxa are less commonly represented among te
fish fauna assemblages from various excavation
units. These are again prmarily inshore fishes, but
several exanples from the family Scombridae (nas
and mackerels) are present. This is significant, for it
does indicate the prctice of pelagic fishing, probably with pearl-shell trolling lures. Trolling for tuna,
however, was clearly a minor fishing strategy in
terms of its contribution to the total fish catch.
In terms of sheer bulk, invertebrates (and
especially molluscs) comprise the majority of the
faual materials from the To'aga site. (Nonetheless,
their contribution of meat to the prehistoric Samoan
diet was probably less than that of fish.) In tenns of
rank-order abundances based on weighLt a few taxa
dominate the assemblages. Consistently the most
abundant species is the reef gastropod Turbo setosus;
the closely related species T. crassus is also quite
common. Only slightly less common is the bivalve
Tridacna maxima, which occupies the reef platform.
Turbo spp. gastropods comprise on average about 62
percent of the invertebrate faunal suite from the
To'aga midden deposits, while Tridacna bivalves
constitute anodter 7 percent Other commonly
represented taxa include: Trochus maculus, Tectus
pyramis, Cypraea spp., Conus spp., Vasum
ceramicwn, Cerithium noduloswn, Strombus
maculatu, Thais armigera, Asaphis violaseus, and
Nerita spp. All of these molluscs occur on the reef
platform and reef crest fronting the To'aga site.
Sea urchins of several species are also repre-
sented in the To'aga middens. The smaller-spined
taxa are doubdess underrepresented in our samples,
because the spines usually are not retained in the
0.25-inch mesh sieves that we employed. Some
indication of their presence was provided by the bulk
samples and micro-aftifact analyses of selected
sediments (see Kirch, Manning, and Tyler, chapter
7; Nagaoka, chapter 13). The Layer IIIA/IhB
occupation in Units 20/23 was noteworthy for an
unusually dense concentration of the large slatepencil sea urchin (Heterocentrotus mamllatus).
More an 6 kg of these spines and test fragments
were overed from these strata, pardy in associadon with an earth oven feature.
In sum, the To'aga excavations have provided
significant new information on which to base
reconstnuctions of Ancestral Polynesian subsistence
economy, and of the impacts of these early island
colonists to the biota of remote Pacific islands. An
economic strategy integrating brad-spectrum
exploitation of natural faunal resources (marine and
tenrestrial) with agricultural producdon is indicated
by the To'aga evidence, reinforcing reconstmctions
based on other early Fijian and Westem Polynesian
sites (Kirch 1984; Kirch and Green 1987). The
presence of an extinct or extirpated species of
megapode, of six species of extirpated seabirds, and
of marine twutle, all in the earliest deposits at To'aga
(especially in Layer IhB of Units 20/23), suggest
that iniidal exploitation of the island's larger faunal
resources may have exceeded te capacity of these
naural populations to survive or reproduce under the
pressures ofintensive human predation. In addition
to these early impacts on fte naltura biota, however,
te Polynesians radically altered the To'aga area,
transforming te coastal environment in particular
into a ftoroughly anthropogenic landscape. The
purposive inroduction of domestic animals and
economic plants, and dte inadvertent introduction of
rats, terrestrial snails, and other organisms were the
first stages in the conversion of the Ofu ecosystem
into a cultumal landscape capable of supporting a
dense human population. Following progradation of
fte coastal plain after about 1900 B.P., te To'aga
area was developed into a highly intensive
arboricultural production zone, dominated by
economic plants of Polynesian inroduction (coconut, breadfruit, aroids, yams, arrowroot, and othes).
Even the steep volcanic slopes inland were cleared
Synthesis and Interprdatios
243
of native forest and converted to zones of shifting
cultivation. Such interior slope modification resulted
as well in incrased rates of soil erosion and deposition onto the coastal flats, enhancing the edaphic
condition of e latter zone for cmp production. In
this regard, the To'aga data add another instance in
the rapidly accumulating repertoire of archaeological
evidence for prehistoric human transfonnation of
Pacific island environments (Bayliss-Smith et al.
1988; Kirch 1983; Steadman 1989).
Hunt and Eikelens (chapter 15) conclude that
red-slipped ware, and perhaps other pottery made of
clay distinctive from the locally Ikown colluvial
sources on Ofu, may reflect imports to the island.
The decline in compositional diversity hypothesized
for the ceramic assemblage might also indicate that
exchange diminished in importance over time.
Additional research on sherds from Manu'a and
elsewhere in the region is necessary, however, to
fully test hypotheses for inter-island ceramic exchange.
INTER-ISLAND CONTACTS
THE TO'AGA SITE:
AND EXCHANGE
CULTURAL RESOURCE
MANAGEMENT CONSIDERATIONS
Archaeological research in fte Fiji-Westem
Polynesia region has produced some evidence for
inter-island contacts and exchange (e.g., Best 1984;
Davidson 1977; Kirch 1988). While inter-archipelago contacts and exchange are known ethohistorically, extensive exchange of ceramics and
other materials appears to have occurred in the
earliest perod of the region's history. Best (1984),
for example, shows that a substantial propordon of
early ceramics on Lakeba was imported to the island.
While te Tatagamatau basalt quany site on Tutuila
(Leach and Witter 1987; Best et al. 1989) is a likely
center for adz export, little is known about potential
pattems of exchange in Sanoa and the quany's
place in a regional system.
Excavations at To'aga produced a range of
materials for which provenance can be deduced or
inferred. Compositional analysis ofvolcanic rock
lithics and adzes, temper and clay of ceramics, and
raw materials from potential sources offer a means to
explore questions of prehistoric exchange.
Weisler (chapter 12) used non-destmctive XRF
analysis on ardfacts and potential source rocks from
Manu'a and Tatagamatau. His results show that the
composition of rock in finished adzes (and in flakes
from adzes showing polish) of fine-grained basalt
cluster with those from Tatagamatau on Tutuila. In
contrast, a relatively coarse-grained dike stone from
Fa'ala'aga on Ofu Island is represented only in
debitage and simple flake tools. These results
suggest that perhaps much, or all local stone was not
suitable for adz producdon. Exchange, at least with
Tutuila some 100 kn to the west, brought adzes and
probably other materials to Manu'a.
As he initial objective of the Manu'a Project
was the identification of archaeological sites for
purposes of cultural resource management-under
contract to the Historc Preservation Office of the
Government of American Samoa-it is appropriate
to conclude this monograph with a discussion of the
significance of the To'aga site, the current status of
the site, and the potential impacts which may
hreaten its integrity in future years.
Site AS-13-1 is unquestionably one of the most
significant archaeological sites yet discovered in
American Samoa, and indeed, in the Samoan
archipelago as a whole. Within American Samoa, it
certainly ranks with the extensive Tatagamatau adz
quanry complex on Tutuila in tenns of its potential to
yield infonnation on the prehistory of the archipelago. Some specific aspects of the To'aga site that
collectively contribute to its archaeological significance are enumerated below:
1. Site AS-13-1 incorporates the largest continuous area of subsurface archaeological deposits dating
to the ceramic phase of Samoan prehistory of any
site yet discovered in the archipelago. These
deposits are estimated to cover between 21,000 and
35,000 m2 and appear to represent a seres of
domestic household units.
2. Site AS-13-1 is well statified, and thus has
the potential to yield a finely-detailed chronological
sequence of cultral change for the Manu'a Islands.
The remarkably deep stratification in parts of the
To'aga site contains three or more meters of culural
deposits. Because of this stratigraphic record, the
244
The Toaga Site
To'aga site presens excellent opporunities for
recovering a detailed sequence of artifacal, faunal,
and settlement information.
3. Site AS-13-1 spans virully the entire
prehistoric sequence of the Samoan archipelago.
Initial occupation of the site began around the close
of the second millennium B.C., contemporaneous
with Xt Mulifanua Lapita site on 'Upolu Island. The
stigraphic wecord from ca. 3000 B.C. tO A.D. 800
has been well documerned by our systematic transect
excavations, detailed in this monograph. Other
archaeological feaus dating to the last one thousand years are also present in the area, altough they
have not yet been intensively investigated or radiocartbon dated. No other single site locality in
American Samoa has yet produced such a continuous occupation sequence spanning the whole of
regional prehistory.
4. The preservation of both ardfacts and faunal
materals in Site AS-13-1 is excellent, especially in
the deeper stratigraphic units, where calcareous
(alkaline) sedimentary conditions prevail. The
majodty of Sanoan archaeological sites are characterzed by acidic soil conditions which do not favor
the preservation of such organic materals as bone,
shell, or sea urchin spines. In such acidic contexts,
cultal materials are usually limited to pottery and
stone artifacts. At To'aga, the excellent preservation
conditions yield not only ceramic and stone artifacts,
but extensive assemblages of bone and shell faunal
materials, as well as arfifacts of shell, bone, and sea
urchin spine. As a result, our knowledge of early
Samoan materal culture and economy has been
significantly expanded by te materials from Site
AS-13-1. A particular example is the complex of
Turbo-shell fishing gear, which for te first time has
given us some in-depth information on Samoan
angling strategies in the first millennium B.C.
Similarly, the faunal assemblages from the To'aga
site are the largest-in terms of both numbers and
taxonomic richness-from any site yet excavated in
the archipelago.
5. Site AS-13-1 also incorporates a number of
features of considerable cultural significance to te
people of Ofu Island. In particular, To'aga is the
taditional seat of the Tui Ofu chiefship, represented
by the Tui Ofu well and burial mound (see Hunt,
chapter 3). These monuments are held in considerable awe by the people of Ofu Island and are directly
conected to a body of oral tradions (see Mead
1930).
Togefter, all of te aspects of Site AS-13-1
enumerated above combine to make this archaeological complex one of the most significant cultural
resources in American Samoa, and indeed, in the
Samoan Lslands as a whole. The site has already
yielded much imporant new infonnation on te
prehistory of the Manu'a Group and the Samoan
archipelago, and its potential has hardly been tapped.
Because of its outsanding significance, we have
nominated Site AS-13-1 to the National Register of
Historic Places, in conjunction with the Historic
Preservation Office of the Govenunent of American
Samoa.
From te culurl resource management viewpoint, it is important to assess any potential hreats to
te To'aga site. As described in chapter 2, the
present mode of land use over most of the site is
subsistence gardening in a more-or-less traditional
manner (arboriculture and Alocasia aroid
swiddening). This relatively low intensity land use
does not seriously heaten the integrity of the site,
other than for minor impacts to surface archaeological features such as 'ili'il pavements or lua'i masi
pits.
There has already been some significant damage
to the site, however, trugh the construction of the
Public Woiks Department landfill at the southwestem edge of the site. (Ironically, it was this landfill
that led to the original discovery of surface potterybearing deposits during the 1986 reconnaissance
survey.) It appears that most of the bulldozed landfill
pit lay outside of the area of deeply stratified archaeological deposits, but these were disturbed along
the inner edge of te bulldozer cut. While the 1986
bulldozing probably did not greatly impact the total
site area, it is extremely important that no furxer
expansion of this landfill operation occur without
prior consultation with the Historic Preservation
Office. It should be possible to plan for future
landfill needs by situating such landfill pits in the
seaward portions of the To'aga coastal fiat tiat do
not contain subsurface archaeological deposits. We
strongly recommend ta prior to any future landfill
operations, the Public Woiks Deparument consult
with the Historc Preservation Officer and arrange
for limited test excavations to assure that landfill
bulldozing take place outside of the zone of buried
Synthesis and Interpretations
archaeological features.
We are unaware, at present, of any other planned
developments or constuction in the To'aga area, but
if such projects arise, ftey could also have te
potential to threaten the integrity of the site. For this
reason, it is important that AS-13-1 be placed on the
National Register of Historic Places, and that the
American Sanoa Historic Preservation Officer
attempt to monitor any proposed land use actions in
the vicinity of the site.
Over the longer tern, te entire archaeological
complex at To'aga could be serously thratened by
natural coastal erosion. According to the
morphodynamic model developed in chapter 4, and
tested through various field and laboratory observations, it would appear tha the southem Ofu coastline
may have entered a phase of sea level rnsgression.
This appears to be reflected along the modem
coasdine by active erosion of the beach ridge and
associated vegetation line, and by exposure of beach
rock. Given that the subsurface archaeological
deposits are situated well inland of the present
shoreline (between about 40 to 60 m, depending
upon the particular locality), such erosion is not at
present a serious concem. However, should this
transgression phase continue or intensify, for
example as a result of global wanming and consequent sea level rise (Geophysics Study Committee
1990), the integrity of Site AS-13-1 would ultimately
be affected. Continued tectonic subsidence of Ofu
Island may itself eventually result in the natural
erosion and destruction of the site, but this pocess
could be seriously intensified and quickened by
rapid sea level rise. Obviously, these are not problems requiring immediate attention, but they should
not be wholly ignored either.
Fnally, we wish to conclude by briefly drawing
attention to some further research possibilities at the
To'aga site. While we have been able to use the
opporunity of subsurface testing and site survey at
AS-13-1 to addess a number of research problems
in Samoan archaeology, our investigation of the
To'aga site in 1987 and 1989 must be regarded as no
more than a pioneering phase. This site has enormous potential to add to our knowledge and undersanding of the prehistoiy not only of Samoa, but of
the Westem Polynesian region as a whole. The site
is also of sufficient size, with an estimated 21,00035,000 m2 of stratified deposits dating to the ceramic
245
phase of Samoan prehistory, that even a large scale
excavation program would not remove more than a
small percentage of the total area, leaving the
majority of the site as an "archaeological bank" for
future research.
We suggest that the next logical phase of
archaeological research at To'aga might be to
employ a horizontal excavation strategy to expose
one or more larger areas (on the order of 100- 150
m2) within the zone of subsurface deposits dating to
the ceramic phase of Samoan prehistory. Our
systematic transect sampling suggests that this zone
is made up of clusters of domesfic or household
residential units, probably consisting of series of
dwelling, cookhouse, and possibly other special
purpose activity areas. Horizontal exposure of one
or more of xse residential units could provide the
first clear picture of the setflement layout and spatial
anangement of an early, Ancestral Polynesian
community. Such a project would be of considerable interest not only for Samoan prehistory, but for
expanding our knowledge of Ancestral Polynesian
culture, which is a critical baseline for the development of later variants of Polynesian culture oughout the whole Polynesian triangle (Kirch and Green
1987).
Obviously, such an expanded excavation
program would also need to be combined with oher
kinds of rsearch objectives and methods. For
example, Weisler's trial application of the nondestructive XRF method of chaacterizing and
sourcing basalt artifacts (see chapter 12) could be
followed up with a more intensive study, with much
potential to reveal patterns of long-distance exchange at all periods of Manu'an prehistory. Similarly, the varation in ceramnics noted at To'aga, such
as the distinction between fine thinware and coarse
thickware, could be explored along avenues other
than just chronological changes in frequency. It may
be that these ceramic wares reflect functional, or
social, pattems in early Samoan society. These can
only be explored through the use ofhorizontal
excavation strategies in which the distribution of
ceramics can be closely mapped in comparison to
the spatial layout of households.
These suggestions are not meant to be an
exhaustive catalog of research problems that might
be addressed at the To'aga site but simply some
possible research directions that we feel would
246
The To'aga Site
contnbute significantly to cunrent issues in Polynesian archaeology and prehistory. We are pleased
that our pioneering phase of research at AS-13-1 has
already been able to yield inporant new insights on
Sanoan prehistory and look forward to the contributions dt future work in the deeply stratified sands
of To'aga will doubtless bring.
REFERENCES CITED
Bayliss-Smith, T., R. Bedford, H. Bookfield, and M.
Latham 1988. Islands, Isainders and the World.
Cambridge: Cambridge University Press.
Best, S. B. 1984 Lakeba: The prehistory of a Fijian
island. Unpublished PhD. dissertation,
University of Auckland.
Best, S., H. Leach, and D. Witter 1989. Report on
the second phase of fieldwork at the Tatagamatau site, American Samoa, July-August
1988. Department of Anthropology, University
of Otago.
Brose, D. S. 1970. The Archaeology of Summer
Island: Changing Settlement Pattern Systems in
Northern Lake Michigan. Anthropological
Papers No. 41, University of Michigan Museum of Anthropology, Ann Arbor.
Clark, J. T. 1989. The Eastem Tutuila archaeological
project: 1988 final report Report prepared for
the Government of American Samoa. Department of Sociology-Anthropology, North
Dakota State University, Fargo.
Clark, J. T., and D. J. Herdrich 1988. The Eastem
Tutuila archaeological project, 1986 final
report Report prepared for the Government of
Amencan Samoa. Department of SociologyAnthrpology, North Dakota State University,
Fargo.
Davidson, J. M. 1977. Westem Polynesia and Fiji:
Prehistoric contact, diffusion and differentiation
in adjacent archipelagos. WorldArchaeology
9:82-94.
Dunnell, R C. 1971. Systatics in Prehistory.
New York: The Free Press.
1986a. Five decades of American archaeology.
IN D. J. Meltzer, D. D. Fowler, and J. A.
Sabloff, eds., American Archaeology Past and
Future. Washington D.C.: Smithsonian
istitution Press.
1986b. Medtodological issues in Americanist
artifact classification. Advances in Archaeological Method and Theory 9:149-297.
Dye, T. S. 1983. Fish and fishing on Niuatoputapu,
Tonga. Oceania 53:242-71.
Emory, K. P., W. Bonk Jr., and Y. H. Sinoto 1959.
Hawaiian Archaeology: Fishhooks. Bemice P.
Bishop Museum Special Publication 47.
Honolulu.
Emory, K. P., and Y. H. Sinoto 1965. Prliminary
report on the archaeological investigations in
Polynesia. Report prepared for the National
Science Foundation. Department of Anthrpology, Bemice P. Bishop Museum, Honolulu.
Feathers, J. K. 1990. Explaining the evolution of
prehistoric ceramics in southeastern Missouri.
Ph.D. dissertation, University of Washington,
University Microfilms, Ann Arbor.
Frost, J. 1978. Archaeological investigations on
Tutuila Island, American Samoa. Unpublished
PhD. dissertation, University of Oregon.
Geophysics Study Committee, 1990. Sea-Level
Change. Studies in Geophysics. Washington
D.C.: National Academy Press.
Green, R. C. 1971. Evidence for the development of
te early Polynesian adz kit. New Zealnd
Archaeological Association Newsletter 14:1244.
1974. Excavation of the prehistoric occupations
of Su-Sa-3. IN R. C. Green and J. Davidson,
eds., Archaeology in Western Samoa, Volume
II. Bulletin of the Auckland Institute and
Museum.
1986a. Lapita fishing: The evidence of site SERF-2 from the Main Reef Islands, Santa Cruz
Group, Solomons. IN A. Anderson, ed.,
Traditional Fishing in the Pacific, pp. 119-35.
Pacific Anthropological Records 37. Honolulu:
Benice P. Bishop Museum.
1986b. Some basic components of the ancestral
Polynesian settlement system: Building blocks
for more complex Polynesian societies. IN P.
V. Kirch, ed., Island Sociedes: Archaeological
Approaches to Evolution and Transformation,
pp. 50-54. Cambridge: Cambridge University
Press.
Green, R. C., and J. M. Davidson, eds., 1969.
Archaeology in Western Samoa, Vol. 1. Bulletin
of the Auckland Institute and Museum.
1974. Archaeology in Western Samoa, Vol. 11.
Synthesis and Interpretations
Bulletin of the Auckland Institute and Museum.
Gren, R. C., K. Gren, K Rappapoit, A. Rappaport,
and J. Davidson 1967. Archaeology on the
Island ofMo'orea, French Polynesia. Antropological Papers of the Amedcan Museum of
Natual History 51(2). New Yor.
Groube, L. 1971. Tonga, Lapita potery, and Polynesian oiigins. Journal of the Polynesian Society
80:278-316.
Holmer, R 1980. Samoan ceramic analysis. IN J.
Jennings and R. Holmer, eds., Archaeological
Excavations in Western Samoa, pp. 104-116.
Pacific Anthropological Records 32. Honolulu:
Bemice P. Bishop Museum.
Hughes, P., G. Hope, M. Latham, and M. Brookfield
1979. Prehistoric man-induced degradation of
te Lakeba landscape: Evidence from two
inland swanps. IN H. Brookfield, ed., Lakeba:
Environmental Change, Population Dynamics,
and Resource Use, pp. 93-110. Paris:
UNESCO.
Hunt, T. L. 1981. New evidence for early horticulhur in Fiji. Journal of the Polynesian Society
90:259-69.
Hunt, T. L., and P. V. Kirch 1987. Radiocaftn
dates from two coastal sites in the Manu'a
Group, American Samoa. Radiocarbon 29:41719.
-. 1988. An archaeological survey of the Manu'a
Islands, American Samoa. Journal ofthe
Polynesian Society 97:153-83.
Irwin G. 1981. How Lapita lost its pots: The
question of continuity in the colonisation of
Oceania. Journal ofthe Polynesian Society
90:481-94.
Janetski, J. C. 1976. Dietary remains from Jane's
Camp-a midden site. IN Jesse Jennings, ed.,
Excavations on 'Upolu Western Samoa, pp. 7582. Pacific Antrpological Records 25,
Department of Antrpology, Honolulu:
Bernice P. Bishop Museum.
1980. Analysis of dietary remains from Potusa
and Falemoa. IN J. Jennings and R Holmer,
eds., Archaeological Excavations in Western
Samoa, pp. 117-22. Pacific Anthrpological
Records 32, Deparument of Antropology,
Honolulu: Bemice P. Bishop Museum.
Jennings, J. D., and R. N. Holmer, eds., 1980.
Archaeological Excavations in Western Samoa.
247
Pacific Anthropological Records 32. B. P.
Bishop Museum, Honolulu.
Jennings, J. D., R. N. Holmer, J. C. Janetski, and H.
L. Smith 1976. Excavations on 'Upolu, Western
Samo. Pacific Anthropological Records 25.
Dept of Anthropology, Honolulu: Benice P.
Bishop Museum.
Kikuchi, W. K. 1963. Archaeological surface ruins
in American Samo. Unpublished M.A. thesis,
University of Hawaii, Honolulu.
Kirch, P. V. 1975. Cultural adation and ecology
in Western Polynesia. Unpublished Ph.D.
disseration, Yale University, New Haven.
1979. Ecology and subsistence. IN J. Jennings,
ed., The Prehistory ofPolynesia, pp. 286-307.
Cambridge: Harvard University Press.
1981. Lapitoid settlements of Futuna and Alofi,
Westem Polynesia. Archaeology in Oceania
16:127-43.
1982. The impact of the prehistoric Polynesians
on the Hawaiian ecosystem. Pacific Science
36:1-14.
1983. Man's role in modifying tropical and
subtropical Polynesian ecosystems. Archaeology in Oceania 18:26-31.
1984. The Evolution of the Polynesian
Chiefdoms. Cambridge: Cambridge University
Press.
. 1986. Retirning East Polynesian prehistory.
Journal of the Polynesian Society 95:940.
1988. Niuatoputapu: The Prehistory of a
Polynesian Chiefdom. Burke Museum Monograph No. 5, Seattle.
. forthcoming, The Wet and the Dry: Irrigation
and Agricultural Intensilcation in Polynesia.
Chicago: University of Chicago Press.
Kirch, P. V., and T. S. Dye 1979. Etnarchaeology
and the development of Polynesian fishing
strategies. Journal of the Polynesian Society
88:53-76.
Kirch, P. V., J. Flenley, D. W. Steadman, F. Lainont,
and S. Dawson 1992. Prehistoric human
impacts on an island ecosystem: Mangaia,
Central Polynesia. National Geographic
Research and Exploration 8:166-79.
Kirch, P. V., and R. C. Green 1987. History, phylogeny, and evolution in Polynesia. Current
Anthropology 28:431-56.
Kirch, P. V., and T. L. Hunt 1988a. Radiocarbon
248
The Toaga Site
dates from the Mussau Islands and the Lapita
colonization of the southwestem Pacific.
Radiocarbon 30:161-69.
1988b. The spadal and temporal boundaries of
Lapita IN P. V. Kirch and T. L. Hunt eds.,
Archaeology ofthe Lapita Cultural Complex: A
Critical Review, pp. 9-32. Burke Memorial
Museum Research Report No. 5. Seattle.
Kirch, P. V., and P. Rosendahl 1973. Archaeological
investigation of Anuta. IN D. E. Yen and J.
Gordon, eds., Anuta: A Polynesian Outlier in
the SolomnIslands, pp. 25-108. Pacific
Antropological Records 21. Honolulu:
Benice P. Bishop Museum.
Kirch, P. V., and D. E. Yen 1982. Tikopia: The
Prehistory and Ecology of a Polynesian
Outlier. Benice P. Bishop Museum Bulletin
238. Honolulu.
Leach, H., and R. C. Green 1989. New information
for the Ferry Berth Site, Mulifanua, Westem
Samoa. Journal ofthe Polynesian Society
98:319-29.
Leach, H. M., and D. Witter 1987. Tataga-matau
"'ediscovered."' NewZealandJournalof
Archaeology 9:33-54.
Lepofsky, D., H. C. Harries, and M. Kellum 1992.
Early coconuts on Mo'orea Island, French
Polynesia Journal of the Polynesian Society
101:299-308.
Mead, M. 1930. Social Organization ofManua.
Bernice P. Bishop Museum Bulletin 76.
Honolulu.
Menard, H. W. 1986. Islands. New Yori: Scientific
American Library.
Poulsen, J. 1987. Early Tongan Prehistory: The
Lapita Period on Tongatapu and its Relaionships, Vols. 1 and 2. Terra Austalis 12.
Department of Prehistory, Research School of
Pacific Studies, Australia National University,
Canberra.
Sinoto, Y. H. 1979. Excavations on Huahine, French
Polynesia Pacific Studies 3:1-40.
Smith, H. L. 1976. Jane's Camp (SUF1-1). IN J.
Jennings, ed., Excavations on 'Upolu, Western
Samoa, pp. 61-74. Pacific Anthropological
Records 25. Department of Anthropology,
Honolulu: Bernice P. Bishop Museum.
Spriggs, M. 1986. Landscape, land use, and political
transformation in southeem Melanesia. IN P. V.
Kirch, ed., Island Societies: Archaeological
Approaches to Evolution and Transformaon,
pp. 6-19. Camnbridge: Cambridge University
Press.
-. 1990. Dating Lapita: Another view. IN M.
Spniggs, ed., Lapita Design, Form and Composition, pp. 6-27. Occasional Papers in Prehistory No. 19. Departnent of Prehistory, Australian National University, Canberm
Steadman, D. W. 1989. Exinction of birds in eastem
Polynesia: A review of the record, and comparsons with other Pacific island groups.
Journal ofArchaeological Science 16:177-205.
Steadman, D. W., and P. V. Kirch 1990. Prehistoric
extinction of birds on Mangaia, Cook Islands,
Polynesia Proceedings of the U. S. National
Academy of Sciences 87:9605-9609.
Steadman, D. W., D. S. Pahlavan, and P. V. Kirch
1990. Exdtion, biogeography, and human
exploitation of birds on Tikopia and Anuta,
Polynesian outliers in the Solomon Islands.
Bishop Museum Occasional Papers 30:118-53.
Honolulu.
Steams, H. T. 1978. Comparison of the geology of
the Society and the Hawaiian islands. Bernice
P. Bishop Museum Occasional Papers
24(14):275-90.
Suggs, R C. 1961. Archaeology of Nuku Hiva,
Marquesas Islanus, French Polynesia. Antrpological Papers of te American Museum of
Natal History 49(1). New Yoik.
Sullivan, A P., III 1893. Storage, non-edible resource
processing, and the interpretaton of sherd and
lithic scatters in Sonoran Desert lowlands.
Journal of Field Archaeology 21:101-114.
Tate, G. H. H. 1951. The Rodents ofAustralia and
New Guinea. American Museum of Natural
History Bulletin 97:187-430.
Terrell, J. 1986. Prehistory in the Pacific Islands.
Camnbridge: Cambridge University Press.
Zenkovich V. P. 1971. A theory of the development
of accumulation fonms in the coastal zone. IN J.
A. Steers, ed., Introduction to Coastline
Development, pp. 94-116. Cambridge: M. I. T.
Press.
.z
Ing
0
am
M.w
Exploradons on the Makran Coast, Pakistan: A Search for Paradise George F. Dales and Carl P. Lipo. Drawing
upon the diary of his 1960 expedition to the remote Makran coast, Dales recounts the search for evidence of
commercial and cultural contacts between the ancient Indus civilization (2500 to 1900 B.C.) and the Near
East. 1992. paper, 288 pp., 96 plates, 68 figures, 11 tables, and 1 foldout map, No. 50
24.00
The Archaeology and Ethnohistory of Fort Ross, Cal#fornia Vol. 1. K. G. Lightfoot, T. A. Wake, and A. M.
Schiff. This new series on the archaeology and ethnohistory of Fort Ross examines the responses of
Native Americans to Russian mercantile adctivities in northern California. This volume focuses on
the interactions of Native Californians with the Russians and the Native Alaskans.
1991. paper, 250 pp., 33 maps, 12 appendices, 6 pp. of illustrations, No. 49
$18.00
Current Directons in Calfornia Obsidian Studies. Editor Richard E. Hughes, with contributions by Jonathon
E. Ericson, Christopher M. Stevenson and Barry E. Scheetz, M. C. Hall and R. J. Jackson,
Robert L. Bettinger, Thomas M. Origer, David A. Fredrickson and Mark E. Basgall. State-of-theart research on sourcing and hydration rate studies in California. 1990. 126 pp., No. 48
$14.00
Prehistoric Hawaiuan Occupation in the Anahulu Valky, O'Ahu Island: Excavations in Three Inland
Rockshelters. Editor Patrick Kirch, with contributions by Terry L. Hunt, Sara Collins, Melinda S.
Allen and Gail M. Murakami. Impact of Hawaiian occupation, circa A.D. 1300, on the local
environment of the Anahulu Valley. Detailed studies of rockshelter sediments, archaeobotanical
remains, charcoal, and landsnails. 1989. 130 pp., 30 plates, No.47
$12.00
Prehistoric Hunter-Gatherers of Shelter Island, New York: An Archaeological Study of the Mashomack
Preserve. Authors Kent G. Lightfoot, Robert Kalin and James Moore. Case study of prehistoric
subsistence and settlement patterns of Shelter Island, New York. The authors evaluate whether or
not horticulture and sedentary lifeways were adopted widely by coastal hunter-gatherers.
1987. 224 pp., No.46
$9.00
Methods in Artfact Analysis: A Study of Upper Paleolithic Burins.
Author Richard N. Dreiman. 1979. 79 pp., No.42
$5.00
Studks in Ancknt Mesoamerica, IV. Editor John A. Graham, with contributions by M. Johnson, E.M.
Shook, M.P. Hatch, J.K. Donaldson, P. Mathews, D.M. Pendergast, D.C. Pring, D.S. Rice and
P.M. Rice. Collection of papers on the archaeology, architecture and epigraphy of the
Maya and the Olmec. 1979. 277 pp., 14 plates, No. 41
$8.00
An Archaeological Assay on Dry Creek, Sonoma County, California.
Authors M.A. Baumhoff and Robert I. Orlins. 1979. 244 pp., No. 40
Il
$8.00
For a complete list of titles and ordering information, please write to: Administrator, Archaeological Research
Facility, Anthropology Department, University of Califomia, Berkeley, CA 94720.
A