Archaeological Investigations of the Sites of
Tatuba Cave (1-NMT-015), Muawai (1-NMT-014), and
Tawaleka (1-NMT-009)
District of Namataku, Navosa Province, Fiji Islands
By
Julie S. Field1
Report prepared for the Immigration Department,
Fiji Islands, the Fiji Museum,
and the Nadrogā / Navosa Provincial Office
1
Department of Anthropology,
University of Hawai‘i, Manoa
Honolulu, USA
December 1st, 2003
Acknowledgements
Funding for this research was provided by the National Science Foundation
(Dissertation Improvement Grant # BCS-0106221), the University of Hawai‘i Space
Grant College (Fellowship 1997-1998), the Honolulu Branch of the American
Association of University Women (Pacific Fellowship, 2001), and the Waikato
Radiocarbon Dating Laboratory (Waikato Archaeological Dating Fund, 2002). Special
thanks must go to the Roko Tui of the Nadrogā/Navosa Provincial Office, Lote
Naikasewa, for permitting archaeological research in Nadrogā and Navosa from August
2001 to March of 2002. Research permits and advising were provided by the Fiji
Museum, and my sincere thanks go to the Director, Sagale Buadromo, past director,
Tarisi Sorovi-Vunidilo, and also the Archaeology Department, in particular Jone
Naucabalavu and Sepeti Matararaba. Use of a 4WD vehicle was permitted by Australian
National University, and my gratitude goes out to Drs. Geoff Hope (ANU) and Paddy
Nunn (USP), and also Lyndall Fisher (FASANOC).
Research at the sites associated with Tatuba Cave would not have been possible
without the blessings of Sawene and Korovou villages. My special thanks go to the Tui
Namataku, Villieme Vatureba, and the turaga ni koro of Sawene and Korovou villages,
Epeli Kunavuli and Senivalati Dretoro. The results of the excavations at the site of
Tatuba Cave has provided important cultural and historical information concerning the
prehistory of Fiji, and I must express my immense gratitude for the opportunity to
investigate the site further. In addition, survey and excavation at the site of Tatuba Cave
was expertly conducted by Jope Kunavuli. My thanks goes to him for his diligent and
ii
careful work at the site, without which this project would have suffered greatly. Lastly,
housing was generously provided by Talatala Joeli Ratusala and Talatala Radini Sulueti
Ratusala (CMF).
iii
Abstract
This report describes a series of excavations and investigations performed upon
the site of Tatuba Cave, located in the district of Namataku, Navosa Province, Fiji. The
investigation of this site was part of a larger research project that explored the emergence
and consequences of competitive and cooperative strategies in Fijian prehistory. The
results of excavations conducted in 1972 (which also included surveys of the nearby sites
of Muawai and Tawaleka) are also included in this report. In brief, excavations indicate
that Tatuba was first occupied between BC 20 and AD 80, and was further occupied as a
fortification until the 19th century. This site is significant for all of Fijian history, as it
represents one of the earliest known habitations in the interior of Viti Levu. Ceramics,
ornaments, and other artifacts also suggest the emergence of a unique highland culture in
the region surrounding Tatuba Cave.
Environmental analyses are also compared to the archaeological record for other
sites in the Sigatoka Valley, and used to determine the presence of three modes of
habitation/subsistence: territorial strongholds, remote refuges, and agricultural production
sites. Archaeological testing of these classes in tandem with GIS-based environmental
research indicate that the Sigatoka Valley was initially occupied between Cal BC 20 –
Cal AD 80, in association with dense and predictable resources. Fortifications that
utilized natural topography, and also remote refuges, were established ca. AD 700, and
remained in use throughout the prehistoric period. Environmental refuges associated
with the effects of the LCO/LIA transition were established ca. AD 1300-1500.
Constructed fortifications that utilized an annular ditch, and which were located in the
valley bottom, appeared ca. AD 1700 – 1850
iv
Fijian Language Abstract
NA VEIKA E KUNE ENA VAKADIKEVI NI QARA VATU KO TATUBA
Oqo nai vakamacala ni veika e kunei ena vakadidike ni kelikeli ka a vakayacori ena loma
ni qara vatu ko Tatuba ka rawa ni kunei ena Tikina ko Namataku ,ena Yasana ko Navosa
ena noda vanua ko Viti.Na vakadidike oqo e tiki ga ni dua na vakadidike levu me dikevi
kina na veika e curu mai, ka mai vakavuna na kena dikevi me macala ka vakadeitaka nai
tukutuku makawa kei Viti.Na macala ni vakadidike ni kelikeli ka a vakayacori ena yabaki
1972(oqo e a sovea taki takega kina na yavu ni koro makawa ko Muawai kei Tawaleka )e
sa vakacurumi tale tikoga ena repote oqo.Me vakalekaleka taki ga na macala ni kelikeli
ka vakayacori mai Tatuba e vakaraitaka ni ko Tatuba e a sa tawani taumada ena yabaki
BC20 ki na AD80,ka a mai tawani me dua na koro ni valu se me tawani ena gauna ni
veivaluvaluti me yacova sara nai ka 19ni senitiuri.Oqo e dua nai vakaraitaki vinaka me
baleta nai tukutuku ni tawa vanua ena loma ni colo kei Viti Levu.Ni mai kune na tikitiki
ni kuro qele ,nai yaya ni sasauni,kei na veimataqali yaya vakamareqeti tale eso,e
vakaraitaka na kena curu mai nai tovo kei nai valavala ka kilai tauduaga kina na vanua ka
vakavolivolita na qara ko Tatuba.
Na kena vakadikevi na veikabula e vakavolivolita me vakatautauvatataki kei na veika e
kunei ena vakadidike ni kelikeli ka volai tukutukutaki tu me baleta na veikoro makawa
ena Bucabuca e Sigatoka ka vakadeitaki kina e tolu na ka,e dusia na tawani kei na ka era
bula kina ,na vanua era dau taqomaki se vakarurugi kina kei na veika era teivaka.E
vakadeitaka talega ni veika e kunei ena kelikeli ena kenai vakatagedegede ka muria ga na
kena vakadeitaki ena misini ka vakatokai na GIS,kei na veika e vakavolivoliti ira e
vakaraitaka ni a sa tawani taumada na Bucabuca e Sigatoka ena maliwa ni gauna mai na
20BC ki na AD80.
Na veikoro ni valu ka vakayagataki kina na veidelana kei na veibucabuca me bai ni valu
kai vakaruru ena gauna ni tiko yavavala era a tauyavutaki ni rauta na yabaki AD 700,ka
ra a tu mera vakayagataka ena loma ni gauna taucoko ni nodra bula ni se bera ni cabe na
lotu.Na kena vakayagataki na veika ga e tu vakavolivoliti ira e a tauyavutaki mai na
yabaki AD1300 ki na 1500.Na kena kelivaki nai keli me bai ni valu ena bucabuca se ena
vanua lolovira e a mai tauyavutaki ena yabaki AD1700 ki na 1850.
v
Table of Contents
Acknowledgments…………………………………………………………………… ii
Abstract……………………………………………………………………………… iv
List of Tables………………………………………………………………………… viii
List of Figures………………………………………………………………………… ix
Research in the Sigatoka Valley, Fiji Islands……………..…………………………… 1
The Upper Sigatoka Valley…………………………………………………. 5
Tatuba Cave (1-NMT-015)…………………………………………………………
6
The Tatuba Locale…………………………………………………………… 8
Cultural Features of Tatuba Cave……………………………………………… 11
The Cave System……………………………………………………………… 16
Historical Descriptions of the Cave System…………………………………… 19
The 1972 Investigation………………………………………………………………… 22
The Excavations of Tatuba Cave……………………………………………… 24
Excavations AC 1-2, ACc 1-2, and CD 1-2…………………………………… 26
Location B……………………………………………………………………… 40
Post-hole Features of Tatuba Cave, 1972 excavations………………………… 47
14
C dates from Tatuba Cave, 1972 excavations……………………………… 47
Artifacts recovered from Tatuba Cave, 1972 excavations…………………… 49
Indications for Site Function from 1972 Excavations………………………… 56
Tatuba Cave: 2002 Excavations……………………………………………………… 57
Test Unit 1: Surface to 10 cmbs……………………………………………… 59
Artifact Descriptions: 0-10 cmbs……………………………………………… 64
Test Unit 1: 10-20 cmbs……………………………………………………… 75
Artifact Descriptions: 10-20 cmbs…………………………………………… 77
Test Unit 1: 20-30 cmbs……………………………………………………… 80
Artifact Descriptions: 20-30 cmbs…………………………………………… 81
Test Unit 1: 30-40 cmbs……………………………………………………… 84
Artifact Descriptions: 30-40 cmbs…………………………………………… 85
Test Unit 1: Feature 1………………………………………………………
86
Radiometric Dating Results: Feature 1………………………………………… 88
Test Unit 1: 40-50 cmbs……………………………………………………… 90
Artifact Descriptions: 40-50 cmbs…………………………………………… 91
Test Unit 1: 50-60 cmbs……………………………………………………… 91
Artifact Descriptions: 50-60 cmbs…………………………………………… 92
AMS Dating Results from Test Unit 1: 50-60 cmbs…………………………… 92
Test Unit 1: 60 –70 cmbs……………………………………………………… 93
Artifact Descriptions: 60-70 cmbs…………………………………………… 94
Test Unit 1: 70-80 cmbs……………………………………………………… 94
Test Unit 1: 80-150 cmbs……………………………………………………… 97
AMS Dating Results: 80-150 cmbs………………………………………….. 99
Test Unit 1: 150-190 cmbs…………………………………………………… 100
The Chronology of Tatuba Cave…………………………………………………… 102
GIS-based Environmental Analyses in the Sigatoka Valley……………………….. 103
vi
Tatuba Cave: Site Function and Placement in the Sigatoka Valley
Chronology……………………………………………………………………………107
Thee Establishment of Territorial Fortifications and Refuges……………… 111
Relocation to Environmental Refuges During Periods of Disturbance……… 113
Constructed Fortifications and Undefended Production Sites………………. 115
Conclusion…………………………………………………………………………… 116
Appendix A: Fauna of Tatuba Cave (1-NMT-015)………………………………… 118
Appendix B - Ceramics of Tatuba Cave (1-NMT-015)…………………………… 119
Notes……………………………………………………………………………… 133
References…………………………………………………………………………… 134
vii
List of Tables
Table 1.1. Artifacts from the 1972 Excavations of Tatuba Cave, units
ACc1-2, AC1-2, and CD1-2…………………………………………………….……
28
Table 1.2. List of artifacts from the 1972 excavations of Location B………………
42
Table 2.1. List of artifacts from Test Unit 1…………………………………………
65
Table 2.2. List of faunal remains from Test Unit 1…………………………………
68
Table 2.3. AMS and radiometric dates from Tatuba Cave excavations
(including 1972)……………………………………………………………………
89
Table 2.4. List of human remains recovered from intrusive burial, Test Unit 1……
96
Table 4.1. List of decorative attributes recorded for the lips and rims of ceramic
vessels from Tatuba Cave……………………………………………………….
122
Table 4.2. Temper types and variations in density and grain-size recorded in the
classification of ceramic vessels for Tatuba Cave…………………………………
123
Table 4.3. Frequency of decorated ceramics by 10 cm level, Test Unit 1…………… 124
Table 4.4. List of morphological attributes for ceramic vessels recovered from
Test Unit 1, Tatuba Cave…………………………………………………………
viii
125
List of Figures
Figure 1.1. Archaeological sites of the Sigatoka Valley…………………………..
2
Figure 1.2. Agricultural features of the Sigatoka Valley…………………………… 3
Figure 1.3. Area map of the district of Navosa, with the location of
Tatuba Cave, with Muawai, and Tawaleka indicated………………………………
7
Figure 1.4. Aerial photo of Tatuba Cave and the associated sites of Muawai, Nalisa,
and Tawaleka………………………………………………………………………… 9
Figure 1.5. Maps of the settlements of Muawai and Nalisa, as mapped by
Palmer et. al., 1972…………………………………………………………………
10
Figure 1.6. Feature map of Tatuba Cave…………………………………………
12
Figure 1.7. Topographic map of Tatuba Cave. Elevation change is indicated
by shading……………………………………………………………………………. 13
Figure 1.8. Map of Tatuba Cave reproduced from Gilbert (1984)…………………… 17
Figure 1.9. Photo of the fortified entrance of Entrance 1, Tatuba Cave……………… 17
Figure 1.10. Map of fortifications at the mouth of Tatuba Cave, as mapped by
Palmer et al. in 1972…………………………………………………………………
23
Figure 1.11. Location of 1972 and 2002 excavation units within Entrance 1
of Tatuba Cave……………………………………………………………………
25
Figure 1.12. Diagram of 1972 excavations units ACc1-2, AC1-2,
and CD1-2, at 10 cmbs……………………………………………………………..
27
Figure 1.13. Diagram of 1972 excavations units ACc1-2, AC1-2,
and CD1-2, at 20 cmbs……………………………………………………………..
32
Figure 1.14. Diagram of 1972 excavations units ACc1-2, AC1-2,
and CD1-2, at 30 cmbs…………………………………………………………….
34
Figure 1.15. Diagram of 1972 excavations units ACc1-2, AC1-2,
and CD1-2, at 40cmbs……………………………………………………………..
37
Figure 1.16. Diagram of 1972 excavations units ACc1-2, AC1-2,
and CD1-2, at 50 cmbs……………………………………………………………..
39
ix
Figure 1.17. Diagram of 1972 excavations at Location B, 0-40 cmbs………….
41
Figure 1.18. Diagram of 1972 excavations at Location B, 50-63 cmbs…………
46
Figure 1.19. The distribution of post-holes throughout the 1972 excavation units…
48
Figure 1.20. Profiles of ceramic rims recovered from unit ACc1-2, 1972…………
51
Figure 1.21. Profiles of ceramic rims recovered from unit AC1-2, 1972…………
52
Figure 1.22. Profiles of ceramic rims recovered from unit CD1-2, 1972…………… 53
Figure 1.23. Profiles of ceramic rims recovered from unit Location B, 1972……… 54
Figure 1.24. Photo of a Nakoro style vessel from northern Serua (near the southern
border of Navosa)…………………………………………………………………… 55
Figure 2.1. Profile of south and west walls of Test Unit 1…………………………
60
Figure 2.2. Profile of north and east walls of Test Unit 1…………………………
61
Figure 2.3a. pH values and percentage of organic matter and CaCO3 in the stratigraphic
layers of Test Unit 1…………………………………………………………………… 62
Figure 2.3b. pH values and percentage of organic matter and CaCO3 in the stratigraphic
layers of Test Unit 1…………………………………………………………………… 62
Figure 2.3c. pH values and percentage of organic matter and CaCO3 in the stratigraphic
layers of Test Unit 1…………………………………………………………………… 63
Figure 2.4. Trochus spp. shell arm-band and dog-tooth pendant
from 0-10 cmbs, TU 1…………………………………………………………………. 67
Figure 2.5. Decorated ceramics from 0-10cmbs, Test Unit 1………………………… 70
Figure 2.6. Bowls from 0-10 cmbs, TU1, Tatuba Cave……………………………… 71
Figure 2.7. Jars from 0-10 cmbs, TU1, Tatuba Cave………………………………… 72
Figure 2.8. A sling-stone, recovered from 10 cmbs in TU1………………………… 74
Figure 2.9. Three fragmented adzes from 0-10 cmbs, TU1………………………… 74
Figure 2.10. Test Unit 1 at 20 cmbs………………………………………………… 76
x
Figure 2.11a. Jars from 10-20 cmbs, Test Unit 1………………………………………78
Figure 2.11b. Jars from 10-20 cmbs, Test Unit 1…………………………………… 79
Figure 2.12. Bowl from 20-30 cmbs, Test Unit 1…………………………………… 82
Figure 2.13. Unusual square-mouthed jar from 20-30 cmbs, Test Unit 1…………
82
Figure 2.14. Jars from 20-30 cmbs, TU1…………………………………………
83
Figure 2.15. Bowl from 30-40 cmbs, TU1………………………………………
86
Figure 2.16. Hearth (Feature 1) in Test Unit 1, at 40 cmbs……………………….
87
Figure 2.17. Profile of hearth (Feature 1)…………………………………………
87
Figure 2.18. Morphology of jar recovered from Feature 1, TU1…………………
89
Figure 2.19. Human remains from 60-70 cmbs…………………………………
95
Figure 2.20. Jar recovered from 70-80 cmbs, Test Unit 1………………………
96
Figure 2.21. Fragment of a wooden torch that emerged from the eastern
profile following the collapse of the unit wall at approximately 80 cmbs…………
98
Figure 2.22. Bowls from 80-150 cmbs, Test Unit 1………………………………
98
Figure 2.23. Jars from 80-150 cmbs, Test Unit 1…………………………………
100
Figure 2.24. Photos of the stratigraphic profiles from Test Unit 1………………
101
Figure 3.1. Environmental zones of the Sigatoka Valley…………………………… 105
Figure 3.2. Comparison of the environmental zones identified by soil/topographic
analyses………………………………………………………………………………. 108
Figure 3.3. Calibrated ages for the 2001-2002 Sigatoka Valley excavations……
110
Figure 2.4. The distribution of earliest occupations for each excavated site
according to temporal period……………………………………………………
112
Figure 4.1. Diagram of a ceramic jar, detailing the morphological
features (metric only) that were recorded for the ceramic assemblage of Tatuba
Cave………….…………….………………………………………………………
120
xi
Figure 4.2. Coded classification for the non-metric morphological attributes
recorded for the ceramic assemblage of Tatuba Cave……………………………
xii
121
Research in the Sigatoka Valley, Fiji Islands
Fiji is famous for its fortifications. Their ubiquitous presence throughout the
archipelago has led archaeologists and historians to privilege raiding and conflict as the
causal mechanism behind the development of Fijian society and history (e.g., Clunie
1977; Frost 1974; Green 1967). However, the causes of conflict have not been fully
investigated. Recent research in Fiji (e.g., Parry 1997; Nunn and Britton 2001, Nunn
2003) has suggested that environmental variations, including droughts related to the El
Niño Southern Oscillation (ENSO), and the transition between the Little Climatic
Optimum (LCO) and the Little Ice Age (LIA) ca. AD 1300 played a primary role in the
development of competitive settlement strategies in Fijian prehistory.
The research reported herein describes the archaeological investigation of
competition and conflict in Fijian prehistory. This research is focused upon the Sigatoka
Valley, located in the southwestern corner of the island of Viti Levu, Fiji.
Archaeological excavations at the mouth of the Sigatoka River in the 1960s by Birks
(1973) yielded evidence of the colonization of the region by 2640 ± 90 BP (GaK 946). In
addition, an excavation performed within the valley interior in 1972 has produced a
single radiocarbon date, that of 1000 ± 70 BP (GaK 4311) for a fortified occupation
(Palmer n.d., in Parry 1987:31). Geographical studies of fortified sites in the region were
also completed by Parry (1987), which provided the basis for subsequent work using
aerial photos, satellite imagery, and geographic information systems by the author (Field
1998, 2002, 2003). To date, this research has identified over 700 archaeological features
in the valley, ranging from fortified hill-top and lowland occupations, open villages, and
irrigated agricultural terraces (Figures 1.1. and 1.2). Given its large size, diverse
1
2
Figure 1.1. Archaeological sites of the Sigatoka Valley.
3
Figure 1.2. Agricultural features of the Sigatoka Valley.
environmental zones, and the quantity and richness of the archaeological record, the
Sigatoka Valley provides an excellent opportunity to investigate prehistoric settlement
patterns, social units, and the effects of environmental variability on the development and
persistence of warfare and conflict.
This research incorporates two bodies of data: GIS-based environmental
analyses, and archaeological excavations. Excavations were performed between August
2001 and March 2002 at 12 sites: Korokune (Conua District, Nadrogā), Bātiri (Sigatoka
District, Nadrogā), Nokonoko (Nokonoko District, Nadrogā), Korovatuma, Bukusia,
Korohewa (Mavua District, Nadrogā), Qoroqorovakatini (Qalimare District, Nadrogā),
Vitogo (Bemana District, Nadrogā), Tatuba Cave (Namataku District, Navosa), Malua
(Noikoro District, Navosa), Nadrogā (Noikoro District, Navosa), and Madraya (Noikoro
District, Navosa). The goal of excavation was to obtain samples for radiometric dating,
and thus determine the age of fortified and unfortified settlements in the Sigatoka Valley.
These results were then compared with environmental and climatic data in order to
determine the causes behind conflict in Fijian prehistory. The results from the excavation
of Tatuba Cave (1-NMT-015) are the focus of this report. Information pertaining to the
1972 surveys conducted at Muawai (1-NMT-014) and Tawaleka (1-NMT-009) are also
included. The place of these sites in the prehistory of the Sigatoka Valley, and the role
they may have played in the development of Fijian culture, are also discussed herein.
4
The Upper Sigatoka Valley
A few kilometers upriver from the settlements of Tonuve and Tuvu lies the
boundary between the tikina (districts) of Bemana and Namataku. This boundary also
marks the division between the southern and northern portions of the Province of
Nadrogā and Navosa, with the latter comprising an extensive region of mountains and
valleys that make up west-central Viti Levu. The Province of Nadrogā and Navosa is an
administrative unit that was formed during the colonial era, and which is subject to the
Commissioner for the Western Division of Fiji, headquartered in Lautoka. The
incorporation of both regions into a single province was the result of the activities leading
up to the Little War of 1875, which induced the colonial government to combine the two
provinces, and thus more easily administer governmental law and order amongst the
populace.
The higher altitude and harshness of the terrain have effectively isolated the
inhabitants of Navosa, few of whom have immediate access to the comforts of developed
countries, such as stable roads, electricity, and communication systems. However, it is
important to note the historic stature of Navosa, which traditionally was its own province
with distinctive linguistic dialects and customs, and its own chiefly lineages and
paramount chiefly title. Archaeological research in Navosa performed in the 1970s
(Palmer 1971, 1972) revealed unique house forms, ceremonial structures, and elements of
material culture, which suggested that Navosa had and long and diverse prehistoric
period, with varying rates of interaction with populations from the coast, interior, and the
high Nadrau plateau to the north. Perhaps most important, the excavations at Tatuba
Cave performed in 1972 revealed the presence of inhabitants as early as the 9th century
5
AD. This date clearly indicates that the interior was inhabited during the early-middle
periods of Fijian prehistory, and suggests that the use of caves as refuges or defensive
features may have had an early start in the interior as well. It also establishes that
populations resided in Navosa during the period, and likely founded the distinct culture
and traditions that are now typical of the region. Additional archaeological research in
Navosa in 2002 was focused on further investigating the early history of the region, and
also determining when the use of fortifications became prominent. As the site of Tatuba
had previously provided indications for early occupations, this site was selected for
additional investigation.
Tatuba Cave (1-NMT-015)
The site of Tatuba Cave is located within the district of Namataku, and directly
across the river from the modern village of Sawene (Figure 1.3). The number assigned to
this site (1-NMT-015) refers to its location in Navosa Province (1), Namataku District
(NMT), and the 15th site of a total of 46. This site is one of many in the area, however it
is the most famous due to its historical use as a distribution point for firearms. The site is
located amongst a substantial outcrop of the Naqalimare limestone formation that
encircles the eastern flanks of Talenalawe peak. In the early 1970s, researchers from the
Fiji Museum (Bruce Palmer, Fergus Clunie, and Kolinio Moce) set out to investigate the
historic and prehistoric aged sites of the Namataku region, and their work resulted in a
series of detailed but unpublished surveys and investigations of the outcrop that contains
Tatuba Cave. They determined via the collection of oral histories and ground-survey that
several portions of the outcrop were inhabited in prehistoric times, and these areas
6
Figure 1.3. Area map of the district of Navosa, with the location of Tatuba Cave,
Muawai, and Tawaleka indicated.
7
constituted distinct sites—some of which were contemporary with Tatuba Cave during
the mid-19th century. However, all the sites associated with the outcrop are within close
distance of one another, and may easily be treated as suburbs within a larger complex.
The Tatuba Locale
Figure 1.4 provides an aerial view of the topography of the Tatuba outcrop, and
also indicates the location of several associated sites. The most prominent feature is the
cliff-face, which extends northwest-southeast for approximately 1.5 km. The steeply
sloping ground that lies below the cliffs is covered with dense vegetation, and it is in this
area that most of the sites are located. The entrance to the main Tatuba Cave is located at
1899.376 Easting, and 3900.016 Northing. This places it directly beneath the highest
point in the cliff-line, and its exit appears 400 meters to the east within a forested sinkhole on the ridge above. Below the mouth of the cave the site of Tatuba is covered with a
multitude of walls and mounds, which will be described in more detail below.
Approximately 1 km to the north of Tatuba Cave is the site of Muawai (1-NMT-014),
which consists of a cluster of 20 stone-faced yavu and walls that lie at the top of a steep
slope, backed by the limestone cliff-face and defended by a large stone wall that flanks
the north end of the site. The features of this site were recorded by Moce and Clunie in
1971, however, no excavations were conducted. Oral histories from the Tui Namataku
suggest that the site was occupied for at least part of the time that Tatuba Cave was
8
Figure 1.4. Aerial photo of Tatuba Cave and the associated sites of Muawai, Nalisa, and
Tawaleka.
occupied, however its absence from Gordon’s notes indicates that the site may have been
abandoned prior to 1875.
Occupation sites atop the ridge above the cave entrance were also utilized during
the life of the Tatuba outcrop, and investigated by Clunie and Moce. Several mounds lie
along the crest of the ridge above the entrance to the cave, and to the east the small
suburb of Nalisa looks down over the sink-hole exit of Tatuba Cave (Figure 1.5). Thesite
9
Figure 1.5. Maps of the settlements of Muawai and Nalisa, as mapped by Palmer et. al.,
1972
10
consists of a few earthen mounds atop a grassy knoll, which is protected to the north by a
stone wall known as Korobulia. This wall transects the ridgeline that lays to the east of
the limestone-cliffs, and then extends down into the forested portion of the cliff face for a
distance of approximately 40 meters. Oral histories collected by Palmer et al. suggest
that the sites of Muawai and Nalisa were contemporary with Tatuba, the wall was
constructed during the time of the Little War (1875-1876) (Palmer et al. 1973).
Lastly, the remains of the historic site of Tawaleka (1-NMT-009) are located on
the flat river-terrace directly west of Tatuba Cave. This village was constructed in
September of 1876 following the end of the Little War, when the government banned the
use of fortifications and required all villages to be located along the valley bottom. The
populations of Tatuba Cave moved to Tawaleka, however it was only inhabited for a few
years due to its problems with flooding, and in the 1880’s they entire village moved and
established the present village of Sawene.
Cultural Features of Tatuba Cave
Figures 1.6 and 1.7 display the cultural features and topography associated with
Tatuba Cave. These maps were created by surveys conducted by Clunie and Moce in
1972, and was updated by the author in 2002. For analytical purposes, the site is
separated into three subsections that correspond roughly to distinctive topographical
features: Occupation Areas I, II, and III.
11
Figure 1.6. Feature map of Tatuba Cave.
12
Figure 1.7. Topographic map of Tatuba Cave. Elevation change is indicated by shading.
13
Occupation Areas I-III
Occupation Area I consists of the two caves within the limestone cliff-face, and
all of the features (yavu, burial mounds, terraces, and walls) located atop and inside of the
highest perimeter ridge that runs parallel the limestone cliff face. This ridge, which also
runs roughly northwest-southeast, forms a distinct enclosure to the west of the mouth of
Tatuba Cave. The yavu, mounds, and terraces included within Occupation Area I are
numbered from 1 to 38, and display an amazing range of forms and construction
methods—from round to square in shape, with and without stone-facing, and of heights
ranging from 30 cm to 2 m. A single 2 m high step-mound (#1) is also present, which
has a roughly square outline, and a lowered rectangular ramp extending off its northern
side. Perhaps even more impressive is the highest mound within Occupation Area I.
Located atop the perimeter ridge, this 1 m high oval mound is most likely the foundation
for what was a bure kalou (spirit house), which was later topped with a stone-faced burial
mound. From the ground level, this mound lies atop a ridge that rises to a height of
approximately 20 m and which is also flanked on both the north and south sides by a
complex series of stone-faced and earthen terraces (#s15-28).
Near the center of Occupation Area I is a large sink-hole, which was tested
archaeologically by Palmer and said to contain midden material to a depth of 1.2 meters
(Palmer et al. 1973). This sink hole is fed by an intermittent stream of water that pours
out of the lowest and most northerly cave entrance, and which forms a small gully below
the mouth of the higher fortified cave, which lies several meters above and atop a steep
slope. To the south of the gully and east of the sinkhole are a cluster of square and
rectangular mounds and terraces (#s 10-14) which abut the cliff face as it runs to the
14
south. Directly north of the sinkhole and the gully is another cluster of mounds and
terraces, which are of larger size and predominantly square/rectangular in shape (#s 3038). All of these mounds are stone-faced, and directly to the west they are flanked by a
heavy stone wall. This wall is approximately 1 m high, and serves as an extension of the
cliff face of the perimeter ridge. It runs northwards for approximately 40 m, and serves
as the northwest boundary for Occupation Area I. Immediately to the west of the wall,
the land drops away as a steep, boulder-strewn slope.
Occupation Area II consists of a cluster of features that is immediately below and
to the west of the perimeter ridge. At its most easterly end, two short stone walls serve to
block immediate access to the habitation area by cutting across the path that runs below
the perimeter ridge. To the north and west of these walls, a cluster of rectangular earthen
terraces extend laterally across the slope, forming a broad staircase of several long
platforms (#s 39-47). The eastern-most boundary of the largest terrace (# 46) is flanked
by a line of large limestone boulders, and the northernmost boundary of the next terrace
to the north is also bounded with a limestone outcrop, with was further extended with a
20 m stone wall. A single stone-faced oval mound also occurs amongst this cluster of
terraces (#43), and is located downslope and to the west.
Occupation Area III lies in the northwest corner of the site, and the two large
stone walls of the northern ends of Occupation Areas I and II form its southern and
eastern boundary. This area is quite steep and rocky, extending downslope and to the
east for approximately 150 meters. Three sets of earthen terraces are evident in
Occupation Area III. The highest is abutted by a bifurcated stone wall, which is clustered
with round and sub-rectangular mounds, all of approximately 1m height (#s 48-54). To
15
the north of this cluster stands a large stone faced terrace with a large boulder embedded
on its south side (# 55). Further north of this terrace, the slope drops slowly away and
down to a small creek.
The second terrace in Occupation Area III occurs approximately 20 m downslope,
and it consists of a long earthen embankment, that drops steeply through a rocky area.
Running roughly east-west, a stone wall bisects this terrace, and extends uphill through a
cluster of boulders and terminates at a square, stone-faced mound (#56). Downslope the
wall becomes discontinuous amidst a jumble of large boulders, and finally terminates at
the edge of the last and lowest earthen terrace. According to Palmer and Clunie’s notes
from 1972, a short tunnel extends from beneath this limestone boulder and runs towards
the creek to the north. This may have been constructed as a defensible route to the
stream, if water supplies at other portions of the site were to run out.
The Cave System
Thus far, subsurface archaeological investigation at Tatuba has focused almost
exclusively on the contents of the cave. As described for the sites of Bukusia,
Korovatuma, and Qoroqorovakatini, the caves of Tatuba are solution caves, being formed
by the constant flowing of water through an outcrop of the Naqalimare limestone. The
caves of Tatuba were surveyed by biologists and speleologists in the 1970s and 1980s
(Gilbert 1984; Watling and Pernetta 1977), which produced a rough sketch map of the
caverns and passages (Figure 1.8), as well as descriptions of its geological features. As
mentioned, the cave has two entrances that face to the west. The higher and more
southerly entrance (Entrance 1) was barricaded at some time in prehistory with a heavy
16
Figure 1.8. Map of Tatuba Cave reproduced from Gilbert (1984).
Figure 1.9. Photo of the fortified entrance of Entrance 1, Tatuba Cave.
17
stone wall, and the interior is relatively dry (Figure 1.9). This is the more ancient
entrance, as evinced by the thick deposits of guano—the source of which are the sheathtailed bat (Emballanura semicaudata) and the white-rumped swiftlet (Collocallia
spodiopygia) that use the cave as a roosting area (Appendix A). The constant deposition
of their excrement has created a flat floor within the cave as far as the light reaches. Past
the limit of daylight the passage continues on, transforming into a narrow crawlspace that
enters into the larger chambers beyond.
According to Gilbert (1984:114), the caverns of Tatuba extend for approximately
420 m, and in some instances the chamber ceilings reach a height of 25 m. At some point
in prehistory the flowing water that created Entrance 1 undercut its own bed, and flowed
northward to form the second entrance (Entrance 2) at a slightly lower elevation. This
opening lacks the guano deposits, and runs with water-flow throughout the year. Of note,
the bottom of this passage shows significant downcutting, as the water is gradually
cutting a new channel into the pre-existing cave floor. As shown in the diagram for
Tatuba Cave (Figure 1.8), the wet and dry entrance passages intersect at approximately
150 m into the interior of the limestone, forming a large vaulted chamber that contains
several stalactites and instances of walls encrusted with another form of precipitated
calcium carbonate, known as moonmilk. Two passages exit this chamber, the uppermost
one ascending a moonmilk-encrusted staircase, which then drops steeply down into
another high chamber. This chamber extends for approximately 50 meters, at which
point light can be seen from a number of cracks in the ceiling. At the end of the cave, the
exit appears as a jumble of rocks and boulders, which must be scaled in order to squeeze
out of the narrow exit. Throughout this portion of the cave, large amounts of ceramics
18
and stone-tool fragments are visible on the surface. It is likely these remains were
washed into the cave or were deposited here during the final occupation episode. The
exit is surrounded on all sides by the walls of the upper sink-hole, the surface of which
contains abundant amounts of ceramics, lithics, and animals bones.
Historical Descriptions of the Cave System
Histories that have been collected from the Namataku region indicate that Tatuba
was considered to be a very important defensive location in prehistoric times, and it had
always been fought over. Even the name “Tatuba” hints at a defensive history for the
site, as it is undoubtedly derived from the word “taubā”, which means “to anticipate the
enemy, having fenced in the town in readiness” (Capell 1991:222). According to oral
histories recounted by the Tui Namataku and his relatives, mountain tribes from as far as
the Wainimala River (approx. 50 km to the east) had sought control of the caves in the
past. Moreover, the Namataku themselves represent one of the later waves of migrants to
the region that had gained control of the region. Their own oral histories record a series
of migrations, wars, and group fissions that began near modern Nadi, and ended when
they advanced to the upper reaches of the Sigatoka River and were successful in gaining
control of Muawai and Tatuba. Genealogical information supplied by the Tui Namataku
and his relatives suggest that this occurred between AD 1780 and 1800.
Also of note, their connection to the tribes of Nadi allowed the Namataku to gain
access to firearms a few decades after their arrival in Navosa, and the advantages given
by these new weapons allegedly allowed the populations of Tatuba to further extend their
control of the region, and even execute raids of the Naqwāqwā, Navatusila, and
19
Nuyakoro tribes upriver (Clunie and Moce n.d., in Parry 1987:38-39). According to
Clunie, flintlock muskets began to appear in small numbers in the interior as early as
1840, and they increased in number and popularity between 1850 and 1875, replacing the
native bows and arrows. The firearm that was the most common during this period was a
muzzleloading, smoothbored flintlock musket, which was favored for its size and noisy
report by Fijian warriors. Even though repeating rifles and breech-loading rifles were
available after 1850 (and many of the native troops engaged in the Little War of 1875
were outfitted with these), they never replaced the flintlock varieties that were prevalent
in the highlands (Clunie 1977:82).
In the early decades of the 19th century the occupants of Tatuba maintained their
relationships with their relatives in Nadi, and this allowed for the influx of European
goods and ideas into the Upper Sigatoka valley. Missionaries also had an influence in
this area, and by the mid 1870s the Namataku of Tatuba had accepted Christianity, and
were supportive of new Colonial government. They were also loyal to the Colonial
garrison at Nasaucoko, and were one of the fort’s main food suppliers. This allegiance,
as well as their long history of domination in the region, encouraged retributive attacks
upon Tatuba and the Namataku people by the rebel tribes from Navatusila and
Naqwāqwā at the start of the 1876 rebellion. According to Gordon’s notes and the oral
histories of the Namataku, Tatuba was attacked and besieged in April of 1876 by musketbearing rebel forces, who stormed the ridge above the cave, near the location of the site
of Nalisa. Once they held this area they were able to burn the portion of the village
enclosed in Occupation Area 1 by tying burning brands to stones and then hurling these
down onto the village. As the village burned the Namataku inhabitants fled to the caves,
20
at which point the rebel forces were able to take up positions at both entrances and the
exit and commence firing. However, their siege of Tatuba Cave did not last long, as
reinforcements arrived from the fort at Nasaucoko the following day. The government
troops routed the rebels, who hastily fled back up the ridge towards Talenalawe peak.
The final battle for Tatuba occurred along the ridgeline north of Nalisa, where the rebels
had constructed a stone wall (Korobulia) to serve as protection in the case of retreat. The
government forces fired upon and then overran this wall, and the rebel forces fled into the
hills (Palmer et al. 1973).
One year later, John Archibald Boyd, an Australian adventurer and amateur
naturalist, journeyed from Nadi to the Upper Sigatoka region, and spent some time
exploring the recently abandoned caves. He described Entrance 1 as being “dry and
spacious, the mouth defended by two strong walls overlapping each [other] and with
loopholes for musketry” (Boyd 1877, in Clunie 1986:31). Boyd also created a campfire
and spent the night catching bats in the cave, and the next morning exhumed the remains
of one of the rebels from a grave located somewhere within Occupation Area 1. A few
years later other travelers entered the caves and noted the presence of grass bedding,
possibly left over from the final night of the cave’s occupation.
Full investigation of the cultural features of the caves did not occur until the
archaeological investigations of the 1970’s. The surveys and excavations of Palmer,
Clunie and Moce revealed the presence of several burials in the niches of the limestone
cliff to the north of Entrance 2. They also documented the location of previously
unknown petroglyphs that were carved onto the face of the south wall inside Entrance 2.
No photo exists for these petroglyphs, but according to Palmer et al. they were very
21
indistinct and consisted of abstract forms and a line of dots, as well as the figures of a
human and a canoe. Palmer et al. also described several features within Entrance 1 and
the main passage, including the presence of several fire hearths along the walls and also a
detailed map of the stone walls (Figure 1.10). Further up the passage they encountered
an area that was used for guano-mining by the inhabitants of Sawene, as well as a small
side chamber that was known to contain several Namataku burials. Throughout these
areas there were signs of occupation, such as fragments of bamboo and coconut, as well
as abundant pottery. They also noted the presence of large amounts of debris in the exit
of the cave, including oven stones, bamboo, ash, ceramics, glass fragments, and musket
balls. According to their Namataku informants, the bamboo fragments may have been
the remains of a palisade that was hastily constructed at the cave exit during the siege of
1876.
The 1972 Investigation
The abundance of cultural features in the cave encouraged Palmer to investigate
further, thus the first archaeological excavations were undertaken beginning in July of
1972. At the time of these excavations, the ceramic chronology for the Sigatoka Dunes
had only recently been established, but whether or not this sequence applied to the
interior of the islands was unknown. Moreover, the Lapita cultural phenomena was only
vaguely understood, and the hypothesis that Lapita peoples had settled the interior of the
main islands as well as the coasts had yet to be tested. This being the case, Palmer and
his assistants attempted to retrieve archaeological materials that were of Lapita age, as
well as draw comparisons between the Tatuba materials and the ceramic chronologies of
22
Figure 1.10. Map of fortifications at the mouth of Tatuba Cave, as mapped by Palmer et
al. in 1972. Much of the outer stone wall had collapsed and fallen downslope by 2002.
the coast. Excavations were conducted within the cave and the materials were partially
analyzed by Palmer et al. at the Fiji Museum. However, the final report concerning the
excavation of the cave and summaries of the findings never materialized due to the
untimely death of Bruce Palmer in 1973.
Despite this major setback, this volume will attempt to summarize the findings of
the 1972 excavations, and also incorporate the results with more recent investigations and
analyses. Unpublished documents including Palmer’s and Clunie’s notes and several
maps and diagrams housed in the archives of the Fiji Museum provide the means by
23
which this may be accomplished. These notes indicate that the excavations of 1972 were
extensive and detailed, but also problematic.
The Excavations of Tatuba Cave
The 1972 excavations of Tatuba Cave were primarily focused on the area
immediately inside the overhang of Entrance 1 (Figure 1.6). A single 2 x 2 m unit (AC 12) was established approximately 2 m east of the defensive stone wall that blocks the
entrance to the cave. Two additional units were extended to the immediate north and
south of AC 1-2, the adjoining walls maintained by 20 cm baulks. Unit ACc 1-2 lay to
the north, and was 1.8 x 2 m in size, while unit CD 1-2 abutted the southern wall of the
cave, and extended to approximately 60 cm at its maximum width. A fourth unit,
‘Location B’ was a 2 x 2 m unit that was located approximately 14 m inside the cave
passage, near the location of a scattered hearth feature (Figure 1.11).
Although the excavation method employed by Palmer et al. was focused on
identifying cultural deposits-- in particular occupational surfaces, the techniques that they
employed towards this goal are very problematic by modern standards. Excavation
proceeded in natural layers, however, there was little documentation of the vertical and
horizontal extent of individual deposits, and few diagrams exist indicating the precise
depths of sedimentary layers or features. Standard soil science descriptions and tests
were not employed, so most deposits have only rudimentary descriptions of color,
texture, and inclusions. Furthermore, artifacts were collected and placed into bags based
upon their provenience within different spits or layers, with little control for vertical
24
Figure 1.11. Location of 1972 and 2002 excavation units within Entrance 1 of Tatuba
Cave.
depth. Thus, it is difficult to compare the frequency of ceramics between different
deposits and excavation levels, as the extent of these features is unknown. Moreover, as
sieving screens were not employed, there is great potential for skewing of the frequency
data, as it is likely the smallest fragments were not fully recovered. Despite these
problems, Palmer et al’s excavations contain a great deal of information, much of which
25
would not be collected in an excavation that employed arbitrary 10 cm levels. The
following sections briefly describe Palmer et al.’s excavations, and also suggest a
rudimentary chronology for the activities that occurred in the cave. These findings are
compared with the results from the 2002 excavation described below.
Excavations AC 1-2, ACc 1-2, and CD 1-2
As mentioned, Palmer et al.’s excavations were extensive, covering
approximately 15 m2, of the cave floor. This mode of excavation allowed for the analysis
of deposits across a very large area, and identified several loci of activity. The primary
excavations at the mouth of the cave (AC1-2, ACc1-2, and CD1-2) can essentially be
treated as a single unit, as they were in close proximity and separated only by two narrow
baulks. According to Palmer’s notes, the surface of all of the units consisted of loose, dry
guano, which varied in depth from 3 to 5 cm. Below the loose surface layer they
encountered the compacted guano that constitutes the bulk of the fill in the cave. This
guano was reddish-brown in color, and was heavily flecked with charcoal fragments and
ash scatters. Limestone pebbles were also common, as were small lumps of chalk.
Artifact density was also high, consisting of ceramic fragments, bones, riparian clam and
snail shells, cut wood fragments, glass fragments, adze flakes, and candlenut shells.
Within this initial compacted layer Palmer et al. encountered several hearth features, (H1H10) which are displayed in Figure 1.12. This diagram, which was reconstructed from
Palmer’s notes, also indicates the location of post-hole features and artifacts within the
deposits. Table 1.1 lists the artifacts that were recovered from each unit between 0 and
10 cmbs.
26
Figure 1.12. Diagram of 1972 excavations units ACc1-2, AC1-2, and CD1-2, at 10 cmbs.
27
Table 1.1. Artifacts from the 1972 Excavations of Tatuba Cave, units ACc1-2, AC1-2,
and CD1-2.
Unit
Depth Artifact Description
(cmbs)
ACc1-2 0-10
Beads
Adze Fragment
Wooden Spear Point
Pig Tusk Ornament
Wooden Spear Barb
AC1-2 0-10
Shell Bracelet
Beads
Iron Fragment
Musket Barrel
Botanical
Remains
Faunal
Remains
Number Weight
(g)
Aleurites
moluccana
CD1-2 0-10
Flake
Flint
Adze Fragment
Ceramic Rim Frags. (plain)
Ceramic Neck Frags. (plain)
Ceramic Body Frags. (plain)
Decorated Frags. (incised)
Decorated Frags. (applique)
Decorated Frags. (tool imp.)
Decorated Frags. (shell imp.)
146.29
Melanoides spp.
137.52
Batissa violacea
Unident. Bone
151.86
326.17
3.5
10.5
184
151.19
56
6971.3
95.68
31.84
24.5
7
3
1
3
Lapita-like Decorated Ceramic
Frags.
Glass
Worked Stone Disk
Bead
ACc1-2 10-20 Beads
Pearl Shell Fragments
Notched Shell Bracelet
Shell Bracelet
Ivory Bead
Worked Bone
2
AC1-2 10-20 Shell Bracelets
28
77.68
17.5
85.02
Unit
Depth Artifact Description
(cmbs)
Adzes
Grooved Stone
Botanical
Remains
Faunal
Remains
Number Weight
(g)
Aleurites
moluccana
Cibi nut
CD1-2 10-20
56.34
1.77
Melanoides spp.
Batissa violacea
44.11
120.33
Marine Shell
31.5
Unidentified Bone
21
119.68
120.02
116.18
3456
91
7
Flake
Ceramic Fragment (rim)
Ceramic Fragment (neck)
Ceramic Fragment (body)
Lapita-like Rim Sherd
Metal fragment
Decorated Ceramic Frags. (tool
imp.)
Decorated Ceramic Frag.
(applique)
Decorated Ceramic Frag. (incised)
Black Bead
ACc1-2 20-30 Worked Shell
Worked Bone
Shell Bracelets
Drilled Pearl Shell
Lapita-like Ceramic Rim
33
35
190
1
1
1
AC1-2 20-30 Shell Bracelet
Grinding Stone
Polished Bones
Ceramic Rim Frag. (plain)
Wooden post fragments
Aleurites
moluccana
CD1-2
Flakes
Melanoides spp.
42
8.77
Batissa violacea
56.34
77
29
Unit
Depth Artifact Description
(cmbs)
Botanical
Remains
Faunal
Remains
Adze
Ceramic Rim Frag. (plain)
Ceramic Neck Frag. (plain)
Ceramic Base Frag. (plain)
Ceramic Body Frag. (plain)
Decorated Ceramic Frag. (incised)
Decorated Ceramic Frag.
(applique)
Decorated Ceramic Frag. (mat
imp.)
Decorated Ceramic Frag. (shell
imp.)
Decorated Ceramic Frag. (crossed
paddle)
Unidentified Bone Frags.
Core
Worked Pearl Shell
ACc1-2 30-40 Necklace Fragment
Bracelet Fragment
Wooden Post Fragment
Perforated Breastplate Frag.
Number Weight
(g)
2
0
70.34
7
1338.54
1842.1
77.68
1
7
1
7
1
120.27
1
38.84
106.02
CD1-2 30-40 Flaked Core
Unidentified
seed
ACc1-2 40-50
Flint
Flakes
Ceramic Rim Frags. (plain)
Ceramic Neck Frags. (plain)
Ceramic Body Frags. (plain)
Ceramic Rim Frag. (incised)
Decorated Ceramic Frags. (incised)
Decorated Ceramic Frags.
(impressed)
Decorated Ceramic Frags.
(crossed-paddle)
Unidentified Bone
Wooden Post Frag.
Lapita-like Ceramic Rim Frags.
Bracelet
30
Melanoides spp.
Batissa violacea
7
24.77
35
Marine Shell
10.5
5
5
7
272.22
200.86
183.18
7565
14
17.5
99
8
7
1
Unit
Depth Artifact Description
(cmbs)
ACc1-2 50-60
Botanical
Remains
Faunal
Remains
Melanoides spp.
Batissa violacea
Number Weight
(g)
1.77
7
Bird Bone
7
Fish Bone
7
21
98
1125
Ceramic Rim Frags. (plain)
Ceramic Neck Frags. (plain)
Ceramic Body Frags. (plain)
Decorated Ceramic Frag. (crossed
paddle)
Lapita-like Ceramic Frag.
Human Bone
85.02
7
7
The ten distinct hearth features that were encountered varied in size and depth, but all
contained deposits of mixed ash and charcoal. Riparian clam shells (Batissa violacea)
were present in two of the more southerly hearth features (H9 and H6), and a scatter of
bones (unknown taxa) were recovered from the deposits near the hearth that abutted the
southern cave wall. The more northerly hearths (H1-H5) were associated with a variety
of cultural items, including a scattering of blue beads, portions of shell and boar’s tusk
jewelry, fragments from wooden spears, a fragmented adze, and pieces of iron and brass
that were probably part of a musket. Several post-holes of unknown function were
encountered along the western boundaries of units AC1-2 and CD 1-2 (P1, P4, P12, and
P11). There were also a great number of ceramics recovered during the excavation of
this surface level. For analytical purposes, the frequency of different vessel forms and
decorations from throughout the deposits excavated in 1972 will be discussed in
following sections.
31
Figure 1.13. Diagram of 1972 excavations units ACc1-2, AC1-2, and CD1-2, at 20 cmbs.
32
Between 10 and 20 cmbs, Palmer et al. continued to encounter compacted guano with
charcoal and limestone pebble inclusions (Figure 1.13). Small patches of orange-colored
sediment was also encountered at this level, which increased in frequency as the
excavation continued. Hearth features H1, H2, H5, H7, and H9 had bottomed out by 15
cmbs, however four features continued into this deeper level. By 15 cmbs, ACc1-2
revealed the extent of H4, which was a large and deep hearth deposit located along the
western boundary of the unit. Feature H3 also revealed itself to be a large ashy deposit
that contained fibrous fragments of bamboo. In unit AC1-2, the hearth features of H6
and H8 thinned out, although H6 contained two limestone cobbles that were likely used
as pot-stands. Feature H10 was the only hearth deposit located in unit CD1-2, however
this unit was predominated by deposits of loose brown fill and mixed ash. The loose
brown fill appeared to be a disturbed area, and contained many fragments of decorative
ceramics including tool-impressed, cross-paddled impressed, incised, and appliqué (Table
1.1).
The high artifact content of the deposits in the nearby units AC1-2 and ACc1-2
suggest the use of the cave at this level for occupation and ornament manufacturing. The
deposits between the hearth features in ACc1-2 contained fragments of shell bracelets,
worked bone and ivory beads, glass beads, and fragments ornamental pearl shell
(Pinctada margaritifera). Additional fragments of shell bracelets were recovered from
the mottled brown fill of AC1-2, which also contained several fragments of adzes and
stone tools. Several large-diameter post-holes also occur as new features in this level, the
largest occur in the eastern portion of AC1-2. The pattern of post-holes within the unit
suggest the location of either a fence or a light-weight structure, however the limited
33
Figure 1.14. Diagram of 1972 excavations units ACc1-2, AC1-2, and CD1-2, at 30 cmbs.
34
number of posts in this level make it difficult to determine a pattern. The distribution of
post-holes throughout the cave will be examined in greater detail below.
The next level (20-30 cmbs) was also dominated by the mottled brown guano fill,
and several deeper hearth features were encountered (Figure 1.14). A concentration of
ash and shell remains was noted in the northeast corner of ACc1-2 (H11), and also the
center of AC1-2 (H12). The artifacts recovered from the fill bounding these features
were predominantly worked pearl shell, shell bracelets, and bone ornaments (Table 1.1).
Of note was a scatter of bamboo fibers and fragments in unit AC1-2, and also the remains
of three in-situ post fragments. Hearth features H4 and H10 also continued to this depth,
consisting of thick deposits of ash and charcoal. H10 also contained a layer of waterworn
stones, which were probably used as pot stands.
The southern portions of units AC1-2 and CD1-2 were noteworthy as they
contained both sterile and disturbed deposits. At 30 cmbs, the loose brown fill located in
the center and southeast corner of CD1-2 revealed the remains of a human burial (B1).
The remains included the femora, tibiae, and smaller foot bones of an adult; the rest of
the skeleton lying buried in the sediment immediately east of CD1-2. The remains were
oriented along the axis of the cave and several stones placed across its length and
between the bones of the legs. Palmer et al. noted upon excavation that the burial pit
contained a very loose fill that dropped away to form a cavity, suggesting that the
individual had been wrapped within mats that had since disintegrated. The disturbed
nature of the sediments above this burial suggest that this entire feature is intrusive, and
likely originates from the later periods of the cave’s occupation. The bottom portion of
the burial was cut into a bright orange layer of sterile sediment, which contained neither
35
charcoal or limestone pebble inclusions, and was most likely an eroded limestone shelf.
Unlike the intrusive burial, hearth feature H10 cuts directly into this orange basal layer,
and was completely covered by later deposits. Despite its close proximity to burial B1,
the hearth feature likely predates it by several centuries.
Throughout the rest of the 20-30 cmbs deposits of CD1-2, the homogeneous
nature of the brown fill suggested that a great deal of disturbance occurred during the
burial process. The frequency of decorated ceramics from various time periods also
indicated mixing, as incised, appliqué, impressed, and cross-paddle impressed occurred
within the same deposit. Additional artifacts included worked pearl shell fragments and
chert flakes and cores. Of note, the yellow/orange basal layer also appears in the extreme
southern portion of AC1-2, directly beneath a large limestone boulder. The boulder
appears to be a fragment of roof-fall, which fell atop the ancient and eroding limestone
floor of the cave prior to human habitation.
The frequency of artifacts in the fill deposits began to drop as the excavation of
the three contiguous units reached depths between 30 and 40 cmbs. Hearth features and
lenses of ash and charcoal also became less common. The remnants of hearth features
H11 and H4 occurred in unit ACc1-2, with H11 consisting of an ashy deposit mixed with
riparian clam shell and fragments of bone artifacts (Figure 1.15). The deposits
surrounding this feature continued to consist of mottled brown fill and scatters of
ornamental artifacts, including fragments of shell bracelets and necklaces, and a section
of a perforated ivory breastplate (Table 1.1). To the south, unit AC1-2 contained the
sterile orange basal layer, with a large disturbed section in its center that was filled with a
loose brown deposit. Another hearth feature, H12, also occurred in this unit, and was cut
36
Figure 1.15. Diagram of 1972 excavations units ACc1-2, AC1-2, and CD1-2, at 40 cmbs.
37
into the orange layer to the immediate north of the large limestone boulder. No artifacts
were recovered in either the orange basal layer or the brown fill deposit located in the
center of the unit, however several large post-holes were noted in the orange fill. All of
these features had passed through upper layers prior to intruding into the lowest sterile
deposit.
This situation also occurred to the south, where a great number of post-holes were
noted in a layer of mixed ash and mottled yellow fill that lie directly above the sterile
basal layer. Excavation in this unit had continued to a maximum depth of 40 cmbs,
which revealed the extent of the orange basal layer along the cave wall, in addition to a
shallow channel of unknown purpose. Hearth feature H10 was also excavated to a
maximum depth of 40 cmbs, and contained large quantities of ash and waterworn stones,
which were capped with a layer of angular limestone cobbles.
Excavation continued to a maximum depth of 50 cmbs in units ACc1-2 and AC12 (Figure 1.16). Hearth H11 continued to approximately 54 cmbs, and contained several
waterworn stone pot stands and a thick layer of ash and charcoal. The base of this hearth
was carved into the top of the orange basal layer, which occurred at the deepest extent of
the excavation in the northeast corner of the unit. The rest of the unit continued to
contain the mottled brown fill deposit, which was pierced in several spots by post-holes
from previous layers. A scattering of waterworn stones also occurred near the center of
the unit, however, very little accompanying ash or charcoal was recovered. Therefore,
these stones may not be part of a hearth feature. Two intact wooden-posts and fragments
of shell ornaments were also encountered in this deposit, as were a high frequency of
38
Figure 1.16. Diagram of 1972 excavations units ACc1-2, AC1-2, and CD1-2, at 50 cmbs.
39
decorated ceramics (Table 1.1). Incised, impressed, and cross-paddle impressed wares
were all recovered, in addition to a large number of plain fragments.
To the south, unit AC1-2 was excavated to a maximum depth of 59 cmbs. The
deposit of disturbed brown fill in the center of the unit revealed itself to be a burial (B2),
which had been cut into the orange basal layer and then filled with a mixed deposit.
Several fragments of long bones, cranial fragments, and teeth, were recovered from the
burial, and two large post-holes of unknown function were located at the very base of the
burial. Hearth feature H12 also continued to a maximum depth of 50 cmbs, and
contained a layer of limestone cobbles at its base. The hearth deposit also contained large
amounts of ash and some fragments of bamboo, two of which were inserted into the
hearth in an upright formation. Few artifacts were recovered from this deposit.
Location B
The fourth unit, ‘Location B’ was located approximately 14 m east of the three
contiguous units, near the center of the cave passage and alongside a scattering of
waterworn stones that were interpreted to be the remains of a hearth feature (Figure 1.11).
The deposits of Location B were nearly identical to those encountered near the mouth of
the cave: the surface consisted of loose brown guano, which transitioned within 5 cm to
a compacted deposit with many lenses of ash and charcoal (Figure 1.17). Limestone and
chalk pebbles were also common throughout this deposit. Three large lenses of mixed
ash were encountered within the first 10 cmbs of Location B, however these layers were
only generally ashy, and lacked the large chunks of charcoal and heavy white ash that
was typical of hearth deposits from the previous excavations. The artifact frequency was
40
Figure 1.17. Diagram of 1972 excavations at Location B, 0-40 cmbs.
41
Table 1.2. List of artifacts from the 1972 excavations of Location B.
Depth Artifact Description
(cmbs)
Botanical
Remains
Faunal
Remains
Number Weight
(g)
0-10
Aleurites
moluccana
12.27
Melanoides spp.
29.77
Batissa violacea
Marine Shell
109.18
24.5
Pinctada spp.
Unident. Bone
Flake
Flint
Ceramic Rim Frags. (plain)
Ceramic Neck Frags. (plain)
Ceramic Body Frags. (plain)
Decorated Frags. (incised)
Decorated Frags. (tool imp.)
Decorated Frags. (shell imp.)
Lapita-like Ceramic
Glass
Iron
Metal
Shell Bracelets
Shell Ornament
Worked Shell
Gun Flint
Pearl Shell Fragment
5
2
2
Aleurites
moluccana
10-20
Flint
Flake
Ceramic Rim (plain)
Ceramic Neck (plain)
Ceramic Body (plain)
42
17.5
7.04
7
70.68
104.84
112.68
4350.02
14
28.34
7
14
21
7
14
15.77
Melanoides spp.
41.54
Batissa violacea
165.52
Pinctada spp.
Marine shell
Fish Bone
Unident. Bone
17.5
66.5
1.77
12.31
70.68
318.04
225.44
420.08
13964.53
Depth Artifact Description
(cmbs)
Decorated Ceramic Frag. (incised)
Decorated Ceramic Frag. (Tool Imp.)
Decorated Ceramic Frag. (Applique)
Decorated Ceramic Frag. (Shell Imp.)
Decorated Ceramic Frag. (crossed
paddle)
Lapita-like Rim Sherd
Glass
Iron Fragment
Metal Fragment
Iron Fragment
Adze
Worked Shell
Botanical
Remains
Faunal
Remains
8
4
Aleurites
moluccana
Flakes
Flint
Ceramic Rim Frags. (plain)
Ceramic Neck Frags. (plain)
Ceramic Body Frags. (plain)
Decorated Ceramic Frags. (incised)
Decorated Ceramic Frags. (applique)
Decorated Ceramic Frags. (tool imp.)
Decorated Ceramic Frags. (fiber imp.)
Decorated Ceramic Frags. (shell imp.)
Decorated Ceramic Frags. (crossedpaddle)
Lapita-like Ceramic Frag.
Imported European Ceramic Frag.
Bottle Glass
Metal Frag.
Charcoal
28.04
Melanoides spp.
12.02
Batissa violacea
129.84
Marine Shell
Fish Bone
31.5
1.77
477.88
42
158.77
322.81
11010.18
331.84
14
120.36
28.34
77.68
4
1
3
1
Aleurites
moluccana
30-40
Number Weight
(g)
34
593.59
4
2
14
14
56.68
31.5
1.77
3.5
24
17.5
5.27
Melanoides spp.
5.27
Batissa violacea 24.5
Marine Shell
17.5
Unidentified Bone 1.77
261.89
Flakes
43
49.59
198.21
24.5
63
103
56
Depth Artifact Description
Botanical
(cmbs)
Remains
Rock Crystal
Ceramic Rim Frag. (plain)
Ceramic Neck Frag. (plain)
Ceramic Body Frag. (plain)
Decorated Ceramic Frag. (incised)
Decorated Ceramic Frag. (wavy relief)
Decorated Ceramic Frag. (imp.)
Decorated Ceramic Frag. (crossedpaddle imp.
Faunal
Remains
5
1
2
Number Weight
(g)
63
49
1211.28
2180.5
38.84
7
14
9
289.04
Aleurites
moluccana
50-60
3.5
Melanoides spp. 7
Unidentified Bone 1.77
Worked Shell
49.34
108.8
836.03
1
42
1
7
Ceramic Rim Frag. (plain)
Ceramic Neck Frag. (plain)
Ceramic Body Frag. (plain)
Decorated Ceramic Frag. (incised)
Decorated Ceramic Frag. (imp.)
Decorated Ceramic Frag. (crossedpaddle)
Decorated Ceramic Frag. (mat imp.)
Lapita-like Ceramic Frag.
60-70 Decorated Ceramic Frag. (applique)
2
623.43
1
2
14
quite dense however, and consisted of fragments of shell bracelets, carved shell
ornaments, pearl shell, and a single gunflint (Table 1.2). A large number of ceramics
were also recovered, including decorated fragments with incised and impressed designs.
However, the most striking feature of Location B was the large number of post-holes,
which appeared to form a linear pattern along the south side of the unit. These features
also intruded into deeper layers, however the analyses of these post-hole distributions
will be fully discussed in the following section.
Similar deposits were encountered between 10 and 30 cmbs. Post-holes increased
in frequency, and scatters of worked shell and lithic fragments were recovered from near
the center of the unit (Figure 1.17). Ceramic frequencies also remained high, including
44
high numbers of incised pieces as well as a few fragments of appliqué, tool impressed,
and crossed-paddle impressed designs. After 35 cmbs however, ceramic and artifact
frequencies dropped off, and patches of the orange basal layer began to occur amongst
the mottled brown fill. At 40 cmbs, the orange basal layer was encountered in the
southeast corner of the unit, which extended to the north and west as the excavation
continued in depth (Figure 1.17). A semi-circle of limestone cobbles were encountered at
40 cmbs, which were embedded in a dark brown fill. This formation appeared to be a
rudimentary hearth, and the sediment surrounding it contained additional new post-holes
and also a small amount of shell ornament fragments, and decorated pottery.
At 40 cmbs, the orange basal layer extended across the southern 2/3rds of the unit,
and several post-holes were evident across its surface. Beneath the limestone cobbles at
the center of the unit lay another rudimentary hearth, this one with four waterworn
cobbles arranged in a semi-circle. This feature occurred at approximately 50 cmbs and
was located directly atop the orange basal layer, but it did not have significant amounts of
ash associated with the stone pot-stands. A fragment of ceramic with potentially an
impression of basketry upon it, as well as a fragment of worked riparian clamshell, were
recovered from between the stones (Table 1.2). Additionally, eight post-holes were
evident between the stones, and extended into the orange basal layer for several
centimeters. These post-holes may represent the use of wooden braces for the pot-stands,
or perhaps the construction of roasting or drying racks. Other scatters of post-holes and a
deposit of ash that contained several fragments of ceramics were also recovered from this
level. Excavation continued to a maximum depth of 70 mbs, at which point most of the
unit consisted of the sterile orange basal layer (Figure 1.18).
45
Figure 1.18. Diagram of 1972 excavations at Location B, 50-63 cmbs.
46
Post-hole Features of Tatuba Cave, 1972 excavations
Two hundred and nine posthole features of varying size and depth were
encountered during the 1972 excavations, and over half were located within the deposits
of Location B. The preliminary analyses of the post-holes performed by Palmer et al.
suggested that the configuration of post-holes throughout the levels may have been the
result of the construction of drying racks or small enclosures. More recent analyses of
the post-holes by the author indicate that the postholes with the largest diameters and
deepest extents may in fact be the remains of a wooden or bamboo palisade. Figure 1.19
displays the distribution of postholes that were approximately 10 cm or greater in
diameter, and which also extended to 35 and 50+cm in depth. These large, deep posts
indicate the desire to create robust and durable structures, and their distribution in linear
patterns that lie perpendicular to the axis of the chamber are suggestive of a fence
structure. All of these postholes originated in layers between 10 and 20cmbs, thus
indicating the construction of the palisade during the later periods of the cave’s
prehistory, perhaps prior to the construction of the stone wall system. The plethora of
narrow and shallow post-holes from various depths also demonstrate some linear
formations, and these may be due to a variety of activities, including racks for storage or
cooking, or even pig enclosures.
14
C dates from Tatuba Cave, 1972 excavations
A single fragment of wood was submitted for radiocarbon dating from the 1972
Tatuba Cave excavations. Palmer’s notes indicate that the fragment came from a wooden
stake (either P127 or P128) that was embedded in the hearth that lay at the bottom of
47
Figure 1.19. The distribution of post-holes throughout the 1972 excavation units. Dots
indicate the location of post-holes that were both wide in diameter, and which extended
deep (30-50 cm) into the substrata.
48
Location B, at a depth of 54 cmbs. This hearth lay directly above the sterile orange basal
layer, and the stake extended into this layer to a depth of 7 cm. This was the deepest
hearth deposit encountered during the 1972 excavations, and presumably its presence
atop the sterile layer was indicative of the earliest human presence in the cave. The wood
fragment (GaK 4311) was submitted to Gakushuin laboratory in Japan, and the resulting
date of 1000 ± 70 years BP suggested that the occupations of Tatuba Cave that lay atop
the sterile orange layer were considerably younger than the Lapita aged settlements of
Sigatoka Dunes (ca. 2900 BP). With modern calibration, the wood sample dates the
hearth deposits at 54 cmbs to between Cal AD 970 and 1160 (at 1σ). A more
conservative calibration (2σ) suggests that there is a 95% probability that the wood
sample dates to a cutting date between Cal AD 890 and 1220.
Artifacts recovered from Tatuba Cave, 1972 excavations
The date from the bottom of unit Location B suggests that all of the deposits
encountered in the 1972 excavations are generally post-AD 1000 in age. Despite the fact
that the artifacts recovered from the 1972 excavations of Tatuba Cave were not fully
analyzed by Bruce Palmer or Fiji Museum staff, an examination of Palmer’s notes
suggests that the remains are typical of post-AD 1000 material culture. Although it is
difficult to provide any assessment of the cultural and temporal diversity represented by
the assemblage without a full physical examination, it should be noted that the
assemblage of artifacts listed by Palmer appears to be exceptionally rich. Future scholars
will undoubtedly find the 1972 Tatuba collection to be a valuable resource for the study
of highland art and decoration, including shell jewelry, bone ornaments, and even
49
decorated ceramics. In addition, the great number of lithic fragments, especially adzes,
may also provide future scholars with detailed information concerning adze material
exchange between Fiji and Tonga. The exchange of items within Fiji between highland
and coastal populations may also be revealed in the sourcing of ceramics and shell items
from the site. Also of note is Palmer’s investigation of ceramic rim forms from the
Tatuba Cave deposits. Prior to the dating of the wooden stake from Location B, Palmer
hypothesized that Lapita-aged ceramics occurred in the deeper deposits of the cave.
Although the AD 1000 date rejected this hypothesis, Palmer’s analysis of ceramic forms
does suggest that there was a great degree of diversity in the ceramic traditions of the
region, some of which undoubtedly descended from early (ca. AD 0-500 ) occupations in
Fiji. The presence of occupations of this age at Tatuba was later confirmed with dates
from the 2002 excavation.
Figures 1.20, 1.21, 1.22, and 1.23, have been adapted from Palmer’s notes and
diagrams, and detail the morphological variability between ceramic rim forms from units
AC1-2, ACc1-2, CD1-2, and Location B. Based upon Palmer’s drawings, the ceramics
of the cave appear to be very diverse in lip shape, rim profile, and rim course, and this
diversity appears to occur throughout the deposits of the cave, and in all locations
(Appendix A)1 Perhaps the most common rim form is the ‘Nakoro’ style, which refers to
manufactured in the village of Nakoro, and consist of ovoid-bodied pots with tall rims
which curve outwards slightly to form rounded or slightly pointed lips (Figure 1.24).
Occasionally these vessels are decorated around their shoulders with impressions created
by a grass stem (Palmer and Shaw 1968). The great number of these rims throughout the
deposits of the 1972 excavations suggests that the style has a long history in the region,
50
Figure 1.20. Profiles of ceramic rims recovered from unit ACc1-2, 1972. Adapted from
original drawings by Palmer et al. n.d. A: fiber impressed; B: finger-gouged; C: ridges;
D: finger-gouged; E: tool impressed; F: perforated; G: perforated; H: tool impressed; I:
impressed lip; J: incised lip; K: impressed lip; L: finger impressed.
51
Figure 1.21. Profiles of ceramic rims recovered from unit AC1-2, 1972. Adapted from
original drawings by Palmer et al. n.d. A: finger impressed; B: shell impressed; C: tool
impressed; D: appliqué band.
52
Figure 1.22. Profiles of ceramic rims recovered from unit CD1-2, 1972. Adapted from
original drawings by Palmer et al. n.d. A: finger-gouged; B: tool impressed lip.
53
Figure 1.23. Profiles of ceramic rims recovered from unit Location B, 1972. Adapted
from original drawings by Palmer et al. n.d. A: tool impressed; B: incised lip; C: tool
impressed lip; D: perforation; E: appliqué; F: interior lip impressed; G: appliqué nubbins;
H: appliqué nubbins.
54
Figure 1.24. Photo of a Nakoro style vessel from northern Serua (near the southern
border of Navosa). Vessel is missing its pointed rim, which would have extended sharply
outwards. Photo credit: Palmer 1971.
and was perhaps the predominant tradition in the area. The other styles (i.e., straight rim
the ceramic tradition of eastern Navosa province. These vessels continue to be courses
and profiles, square lips, and expanding/contracting rims) occur throughout the
laterphases of Fijian prehistory, and are also found throughout the Sigatoka Valley.
The darker-shaded rims in Figures 1.20, 1.21, 1.22 and 1.23 were identified by
Palmer as “Lapita-like”. These forms are distinctly different from the other rims in the
assemblage in that they have tightly curved rim profiles and square lips, and are also
quite thick. Cursory examination of Palmer’s diagrams of these specimens by Dr. David
Burley (Simon Frasier University) led to the suggestion that these forms may be related
to the “Fijian Plainware” tradition, which may have developed in Fiji between AD 0 and
300. However, the 1000 ± 70 date for the deepest cultural deposits encountered in the
1972 excavation strongly suggest that these forms post-date the Plainware period.
55
Therefore, it seems more likely that these square-lipped forms represent a potting style
that was imported into the upper Sigatoka Valley at some point between AD 1000 and
1500.
Moreover, the presence of decorated ceramics throughout the deposits suggests
that the deposits dated to between AD 1000 and the contact period. Although the exact
number of decorated ceramics was not recorded in Palmer’s notes, is appears that small
amounts of incised, tool impressed, shell impressed, crossed-paddle impressed, and
appliqué designs occurred on ceramic fragments recovered from all of the excavation
units.
Indications for Site Function from 1972 Excavations
Based upon the presence of imported shell, complete and incomplete shell
ornaments, and food remains at the cave, Palmer hypothesized that Tatuba was both an
occupation site and a center for shell-ornament manufacture. Whether or not the site was
a ‘center’ is difficult to ascertain at this time, although the lack of pearl shell at other sites
in the Sigatoka Valley does suggest that the resource was restricted in some way. The
analysis of post-holes suggests that a wooden palisade once blocked access to the mouth
of the cave, and the stone wall was probably built shortly after. Tatuba Cave may have
also been used for housing pigs, as suggested by the alignment of smaller post-holes
along the cave walls. In addition, the site was used for burials, as evinced by the
presence of human remains in units AC1-2 and CD1-2.
56
Tatuba Cave: 2002 Excavations
The excavation performed in 2002 was focused primarily on establishing a more
secure chronologic sequence for the occupation of the cave and the Tatuba locale in
general. The excavation was also intended to sample the deepest deposits of the cave,
and determine the depositional history of both the natural and cultural deposits. Prior to
establishing this excavation unit, the formation of Entrance 2 was examined. As this
entrance is an active solution cave, it lacks the deep deposits of guano that hide the
underlying limestone formations. The immediate entrance area of this cave revealed the
presence of a 1m deep gutter, which had been carved into the floor of the ancient
limestone chamber as the water exited the cave. Also of note, the gutter formation in
Entrance 2 created a sharply cut limestone bench, which extended into the cave for some
distance. Upon examination, Entrance 2 appeared to be a “keyhole passage”, which is a
particular limestone cave formation that occurs when water is forced between two
bedding planes, creating what is called a called a “phreatic tube”. Later on the water flow
erodes a new channel in the bottom of the passage, thus creating the “keyhole” shape of a
keyhole passage. The floor of the cave also contained evidence of roof collapse, as
numerous boulders and rubble piles littered both the gutter and the limestone bench.
Using Entrance 2 as a model, it was assumed that Entrance 1 would contain a
similar “keyhole passage” formation, and thus the deepest deposits would lie near the
center of the cave chamber, perhaps amidst rubble from the collapsing roof. Thus, Unit 1
was positioned in between the 1972 excavations of Location B and ACc1-2 – CD1-2, and
also along the centerline of the chamber, in hopes of reaching the very deepest cultural
and natural deposits (Figure 1.11). Unlike the 1972 excavations, the 2002 excavation
57
employed arbitrary 10 cm levels, and all of the deposits were sieved using a 1/8th inch
screen. Several fragments of human remains from a burial were also encountered during
the excavation, and these fragments were reburied at the site. Additional fragments of
human bone were recovered from the deeper levels of the excavation, but were not
identified as such until analysis in the laboratory. These remains, and all other artifacts,
were collected from the excavation.
The ceramic fragments encountered in the 2002 excavations were analyzed and
compared to existing classifications for early, middle, and late vessel forms and designs.
Other artifacts such as shell, bone, and plant remains were classified according to taxa.
The potential formation processes at the site were also investigated: standard soil science
terminology was used to describe the layers that were encountered, and samples were
taken for later geoarchaeological analyses, including the measurement of pH levels,
CaCO3 content, organic matter content, and grain size distributions. In addition,
drawings and photos of the layers and cultural features visible at each 10 cm layer were
generated, as were stratigraphic profiles for all the walls of the unit. Perhaps most
important, five organic samples, including a fragment of human bone, were submitted for
AMS and standard radiometric dating. When combined, these data indicate the processes
that were involved in the natural formation of the cave and its deposits, and also indicate
the chronology of human occupation at the site. These data provide new insights into the
depositional history of the cave, and also serve to enrich the results from the 1972
excavations, generating a greater understanding of the cave and its history.
58
Test Unit 1: Surface to 10 cmbs
Two deposits were encountered during the excavation of the first 10cm of
sediment that lay below the surface. As with Palmer et al.’s excavations, the surface
layer (Layer I) consisted of dry, loose, guano-laden deposit, with the consistency of cakeflour. This deposit was approximately 5-8 cm thick, and extended as a horizon across the
entire test unit (Figures 2.1 and 2.2). The analysis of this deposit in both the field and
laboratory suggested that material was a sandy loam, 24% of which was calcium
carbonate. An additional 20% of the sample was organic matter, most likely the
excretions of the bats and birds that occupy this cave. This being the case, the deposit
was also acidic, with a pH level of 7.6. The chemical and physical profiles of this deposit
(Layer I) are summarized in Figure 2.3a, b, and c (charts of pH, organic matter, and
CaCO3). Within 5-8 cm Layer II was encountered, which also extended across the entire
surface of the unit. This deposit was more compacted, and it also contained a high
proportion of charcoal fragments and many thin lenses of ash. The sheer number and
thinness of these lenses made it impossible to excavate them as features, however they
were highly visible in the profiles of the unit. Layer II was also a sandy loam, and
appeared in character to be very similar to Layer I. However, the lower boundary of this
layer was extremely irregular, as it contained several postholes (Figure 2.1).
Four waterworn stones were encountered in the southwestern portion of the unit,
at the boundary between Layers I and II. These stones may have been employed as potstands during the later periods of the cave’s occupation. A smaller stone also from this
level was identified as a possible sling-stone, and was collected as an artifact.
Throughout the first 10 cm of the excavation, the artifact content was very high, and
59
Figure 2.1. Profile of south and west walls of Test Unit 1. White square: location of sediment samples; red triangles: AMS samples.
60
Figure 2.2. Profile of north and east walls of Test Unit 1. Red triangles indicate the
location of dated samples.
61
Figure 2.3a. pH values for Test Unit 1, Tatuba Cave.
Figure 2.3b. Percentage of organic matter content for Test Unit 1, Tatuba Cave.
62
Figure 2.3c. Percentage of CaCO3 in the stratigraphic layers of Test Unit 1.
63
ceramics were particularly common, constituting approximately 30% of the fill (Table
2.1). Twenty-five grams of charcoal were also collected from throughout the charcoal
lens deposits.
Artifact Descriptions: 0-10 cmbs
Faunal Remains
As in the 1972 excavations, ornaments and ornament fragments were recovered
from the uppermost levels of the excavation. These included a fragment of a shell armband (possibly Trochus niloticus) and a pendant made from a dog canine (Figure 2.4).
Other faunal remains more typical of prehistoric Fijian sites were collected as well (Table
2.2). Seventy-two grams of riparian clam and snail shells (Batissa violacea, Physastra
nasuta, and Melanoides aspirans) occurred in this level (Appendix A). The presence of
the riparian clam Batissa violacea, which is native to the Lower Sigatoka Valley,
indicates that the inhabitants of Tatuba had access to distant resources either through
exchange or travel. Ninety-nine grams of toto nutshell (Aleurites moluccana) was also
recovered, and these remains most likely reflect the use of the oily nuts for lamps. An
assortment of rat (Rattus spp.), sheath-tailed bat (Emballonura semicaudata), and pig
(Sus scrofa) bones were also recovered in the fill. These fragments (especially the pig)
may represent food remains, but may also be naturally occurring fauna.
Ceramics
The ceramic fragments recovered from Test Unit 1 were analyzed with a
classification system that incorporates the morphological variation of vessel form and
64
Table 2.1. List of artifacts from Test Unit 1.
Layer
Level Depth Artifact
(cmbs) Description
Weight
(g)
I
1
ceramic
charcoal
misc. shell
adze fragments
lithic fragments
toto nut shell
glass
rifle cartridge
dog tooth pendant
shell (Trochus niloticus) arm band
3300
24.9
72
226.2
10
99.9
6.5
11.1
0.9
5.9
rat (Rattus spp.) bone
bat (Emballonura semicaudata) bone
human bone
pig (Sus scrofa) bone
sling stone
charcoal
unident. bird bone
unident. fish bone
rat (Rattus spp.) bone
lithic fragments
adze fragments
charcoal
misc. shell
toto nut shell
ceramic
ceramic
lithic core
lithic flagments
groundstone fragment
misc. shell
charcoal
ceramic
misc. shell
toto nut shell
adze fragments
0.8
0.2
7.3
4.3
211.5
14.3
0.2
0.2
0.2
118
ceramic
charcoal
unident. fish bone
misc. shell
toto nut shell
825
29.9
0.1
3.6
6.8
II/III/IV/IX
III/IV/V/IX
2
2
3
II/IV/V/VI/IX 4
Feature 1
0-10
15
10-20
20-30
30-40
none 45-55
III/IV/V/VI/IX 5
40-50
65
28.5
27.6
12.6
2660
1150
301.3
7.9
41.2
1.8
1.6
22.2
Layer
Weight
(g)
7.8
1
3.2
207.2
1.2
16.7
0.1
135.5
52
0.4
0.3
III/IV?
Level Depth Artifact
(cmbs) Description
charcoal
lithic
misc. shell
ceramic
6
50-60 misc. shell
charcoal
toto nut shell
ceramic
7
60-70 ceramic
toto nut shell
human bone
matting
charcoal
8
70-80 unidentified bone
human hair
lithic
matting
misc. shell
ceramic
charcoal
toto nut shell
9
80-90 human hair
human bone
misc. shell
toto nut shell
ceramic
charcoal
10 90-100 toto nut shell
ceramic
misc. shell
human bone
human hair
charcoal
11 100-110 charcoal
ceramic
human bone
toto nut shell
none 30-100 wooden torch frag.
IX
12
IX
13
99.3
15.7
9
0.8
86.5
3.9
III/IV/IX
III/IV/IX
IX
IX
IX
IX
110-120 ceramic
charcoal
toto nut shell
misc. shell
120-130 ceramic
toto nut shell
66
12.7
0.3
0.5
2
1.1
0.4
53.4
6
1
0.6
3.3
0
1.1
27
6.6
1.9
31.4
0.7
3.7
0.2
3
6.5
46.5
7.7
3.8
2.3
Layer
Level Depth Artifact
(cmbs) Description
adze fragment
lithic fragment
charcoal
14 130-140 charcoal
ceramic
toto nut shell
human bone
human hair
15 140-150 toto nut shell
misc. shell
charcoal
ceramic
Weight
(g)
7.1
1
2.1
3.4
26.9
2
6
0
1.6
0.3
5.8
32
IX
16
IX
17
IX
18
IX
IX
19
20
0.1
3.7
1.3
22
1.2
0.1
0.1
2.4
1.5
17.2
1
0.2
0.7
0.2
0.3
IX
IX
150-160 misc. shell
charcoal
toto nut shell
ceramic
wood fragment
160-170 misc. shell
unidentified nut shell
ceramic
charcoal
wood fragment
170-180 charcoal
toto nut shell
wood fragment
180-190 charcoal
190-200 charcoal
Figure 2.4. Trochus spp. shell arm-band and dog-tooth pendant from 0-10 cmbs, TU 1.
67
Table 2.2. List of faunal remains from Test Unit 1
Layer
Molluscan
Level Depth Species
I
1
0-10
Mammallian
Species
Avian
Species
Batissa
violacea
Physastra
nasuta
Melanoides spp.
unidentified
shell
Rat (Rattus spp.)
mandible
Bat (Emballonura
semicaudata) os
coxae
crania frag.
(unidentified)
Pig (Sus scrofa)
rib frag.
II/III/IV/IX 2
10-20
Batissa
violacea
Melanoides spp.
Rat (Rattus spp.)
mandible
Unident. Bird
bone
III/IV/V/IX 3
20-30
III/IV
V/VI/IX
30-40
4
Feature 1
45-55
Batissa
violacea
Melanoides spp.
Melanoides spp.
Batissa
violacea
Melanoides spp.
Modiola tulipa
(kuku)
Piscean
Species
Number Weight
(g)
30
50
1
45
1
20
1
1
1
0.2
1
0.2
1
2
12
33
20
7.6
1
0.2
1
0.5
Unident. Fish
bone
1
0.2
3
4
5
2.9
2
1.8
4
3
1.6
1
7
1
Unident. Fish
1
bone
III/IV
V/VI/IX
5
40-50
III/IV/IX
IX
6
8
50-60
70-80
Melanoides spp.
Batissa
violacea
Melanoides spp.
Melanoides spp.
68
0.2
4
1
2
2
1
2.2
1.2
0.4
Layer
Molluscan
Level Depth Species
Mammallian
Species
Unident. Mammal
bones
IX
9
IX
10
IX
12
IX
13
IX
15
IX
16
IX
17
80-90
Melanoides spp.
Batissa
90-100 violacea
110Melanoides spp.
120
Batissa
violacea
120Melanoides spp.
130
140150
Melanoides spp.
150Batissa
160
violacea
160Melanoides spp.
170
Avian
Species
Piscean
Species
Number Weight
2
1
2
1
1
0.7
2
0.2
2
0.6
1
1
1
0.3
1
0.1
1
0.1
also the presence or absence of decoration (Appendix B). The result of analysis indicates
that of the 416 ceramic fragments recovered from Level 1, 9.8% were decorated with
crossed paddle impressed, parallel paddle impressed, incised, end-tool, and wipedneckdesigns (Figure 2.5). The designs are very diverse, but typical of late-prehistoric
period decorations.
Thirty-two of the fragments were from the rim and necks of vessels, and of these,
10 were complete enough to allow for the morphology of the parent vessel to be
reconstructed and analyzed. Figures 2.6 and 2.7 reveal the rim course, rim profile, and
lip shape of the bowls and jars recovered from Layer 1. The bowls are predominantly
square-lipped, although there is a great deal of variation in the profile of the lip shape,
which include parallel and contracting forms. The bowls are also quite variable in the
depths, with both shallow (#s 272, 281) and deep forms (#s 275, 277). The temper used
in the fabric of the bowls is alluvial sand, and most particles are 1-2 mm in size and with
69
Figure 2.5. Decorated ceramics from 0-10cmbs, Test Unit 1. A: tool impressed; B:
finger-gouge; C: parallel paddle impressed; D: crossed paddle impressed; E: incised.
70
Bowl # 264, TU1, 0-10 cmbs
Temper Class: 1
Bowl # 266, TU1, 0-10 cmbs
Temper Class: 3
Bowl # 272, TU1, 0-10 cmbs
Temper Class: 1
Bowl # 275, TU1, 0-10 cmbs
Temper Class: 1
Bowl # 277, TU1, 0-10 cmbs
Temper Class: 1
Bowl # 281, TU1, 0-10 cmbs
Temper Class: 1
Figure 2.6. Bowls from 0-10 cmbs, TU1, Tatuba Cave.
71
Jar # 263, TU1, 0-10 cmbs
Temper Class: 3
Jar # 270, TU1, 0-10 cmbs
Temper Class: 1
Jar # 278, TU1, 0-10 cmbs
Temper Class: 1
Jar # 283, TU1, 0-10 cmbs
Temper Class: 1
Figure 2.7. Jars from 0-10 cmbs, TU1, Tatuba Cave.
a density of 10-20%. In general the bowls resemble modern dari in size and form
(averaging 30 cm in diameter), and may have been used for serving food or beverages.
The jars are equally diverse in form. Two vessels (#s 270, 278) are similar to the
‘Nakoro’ style pot, which was identified by Palmer as having an ovoid body, elongated
straight neck, and a sharply curved and pointed rim (Figure 1.24). The remaining two
72
vessels have square lip forms, but vary between straight and curved in their rim course.
All the jars are approximately the same size, with orifices averaging 25 cm, which would
have allowed easy access during cooking. The vessels are also manufactured with
alluvial sand of a small (1-2 mm) grain size.
Lithics
Several fragments of stone tools and also lithic debitage were recovered from the
fill of Level 1. As mentioned, a single sling-stone was recovered. This artifact consists
of a lightly pecked river cobble, which is approximately 6 cm in diameter (Figure 2.8).
Seven adze fragments were also recovered, all of which were composed of dense
aphanitic basalt. Adzes appear to have been both quadrangular and lenticular in crosssection, and have polished surfaces. A single adze also has been girdled with a series of
grooves that allowed the implement to be more easily hafted to a handle (Figure 2.9).
Seven lithic fragments were also recovered from the fill of Level 1. A single
fragment was composed of chert, and the rest consisted of primary flakes of aphanitic
basalt. These lithics did not show evidence of retouching (sharpening or shaping).
Historic Artifacts
Two fragments of green bottle glass and a cartridge from a repeating rifle were
also recovered from Level 1. The glass is green in color and appears to be from a modern
(20th century) beer bottle. The cartridge (also called a “shot shell”) is very corroded, but
is likely 20th century in age.
73
Figure 2.8. A sling-stone, recovered from 10 cmbs in TU1.
Figure 2.9. Three fragmented adzes from 0-10 cmbs, TU1. A: quadrangular adze butt;
B: lenticular adze; C: quadrangular adze with incised hafting scars.
74
Test Unit 1: 10-20 cmbs
Between 10 and 20 cmbs, the excavation encountered three additional Layers—
III, IV, and IX, which occurred within different portions of the unit. As indicated in
Figure 2.10, the deposits at 20 cmbs were extremely mottled, with many scatters of
charcoal and ash, as well as medium-sized pebbles of limestone and chalk. This was due
to the mixing of deposits that occurred during the various occupation periods of the cave.
Layer III was encountered in the southern portion of the unit, and consisted of a layer of
dark grey sandy loam that was virtually indistinguishable from layers I and II. However,
this deposit was slightly darker in color, and had higher levels of calcium carbonate
(29%) and slightly less organics (14%) (Figure 2.3). As the excavation continued in the
southern portion of the unit, a single feature began to appear at the bottom of Layer III,
consisting of a deposit of thick ash and charcoal (Feature 1, Layer VI).
Throughout the central portion of the unit at depths between 10 and 20 cmbs, the
excavation encountered Layer IV. This deposit was also a sandy loam, but had a reddish
brown color and fewer instances of charcoal inclusions. Chemical analyses indicate that
this deposit had a much lower percentage of calcium carbonate (3%), despite the fact that
the organic level remained high, at 17%. Layer IV was approximately 10cm thick in the
southern portion of the unit, but increased in thickness in the center and northern
portions, as presumably the floor of the cave dipped down in this area, allowing for a
deeper deposit. As indicated in the profiles from the west and east walls (Figures 2.1 and
2.2), Layer IV underwent a great deal of mixing and disturbance in this portion of the
cave, and the intrusion of several deposits from above and below are visible. However,
75
Figure 2.10. Test Unit 1 at 20 cmbs.
as excavation continued in the northern portion of the unit, intact deposits from Layer IV
were encountered from depths of 70 cmbs in the western portion of the unit, and at
120cmbs in the eastern portion.
Layer IX, which appeared to be the remains of an intrusive burial, was also
encountered between 10 and 20 cmbs, in the extreme northwest corner of the unit. This
deposit is entirely mixed and disturbed, and was not further tested with either chemical or
physical analyses. It continued to extend in depth throughout the excavation, and reached
a maximum depth of 140 cmbs.
76
Artifact Descriptions: 10-20cmbs
Faunal Remains
The artifacts recovered from between 10 and 20 cmbs were retrieved from Layers
I, II, III, IV, and IX. The frequencies were slightly less than those of the preceding level,
but are largely similar. Twenty-seven grams of riparian clam and snail shell (Batissa
violacea and Melanoides spp.) was recovered. In addition, the mandible of a rat (Rattus
spp.) and unidentified fragments of bird and fish bone occurred in the fill. Twelve grams
of toto nutshell were also recovered.
Ceramics
Two hundred and seventy five ceramic fragments were recovered, 37 of which
were decorated with parallel and crossed paddle impressed designs, incising, neck
wiping, end-tool and shell impressing, and grass impressing. A few of the incised
fragments appeared to be from a single tool impressed vessel, portions of which had been
recovered in Level 1 (Figure 2.5, A). All of the decorations are typical of the late
prehistoric period.
Ten of the rim fragments were complete enough to allow the reconstruction of the
morphology of the parent vessels (Figure 2.11a and 2.11b). All of these fragments were
from jars, and the variability in the forms is similar to those recovered in Level 1. One
fragment is similar to the ‘Nakoro’ style identified by Palmer (# 313). The remaining
nine vessels have both straight and curving rim courses, parallel and expanding profiles,
and vary in lip form from square to rounded. These vessels also have large orifices
(averaging 25-30 cm in diameter), and are tempered with alluvial sand.
77
Jar # 312, TU1, 10-20 cmbs
Temper Class: 1
Jar # 313, TU1, 10-20 cmbs
Temper Class: 2
Jar # 315, TU1, 10-20 cmbs
Temper Class: 2
Jar # 317, TU1, 10-20 cmbs
Temper Class: 1
Jar # 318, TU1, 10-20 cmbs
Temper Class: 1
Figure 2.11a. Jars from 10-20 cmbs, Test Unit 1.
78
Jar # 326, TU1, 10-20 cmbs
Temper Class: 2
Jar # 331, TU1, 10-20 cmbs
Temper Class: 2
Jar # 342, TU1, 10-20 cmbs
Temper Class: 1
Jar # 343, 10-20 cmbs
Temper Class: 1
Jar # 346, 10-20 cmbs
Temper Class: 4
Figure 2.11b. Jars from 10-20 cmbs, Test Unit 1.
79
Lithics
Seven basalt fragments were recovered from between 10 and 20 cmbs, and these
fragments had the abraded and polished surfaces that are typical of stone adzes.
However, the original shape of the artifacts could not be ascertained from the fragments.
In addition, two fragments from stone pounders, and 10 fragments of basal and chert
debitage were recovered. None of the debitage fragments showed signs of sharpening or
shaping.light, charcoal fleck ash. By 30 cmbs, this hearth feature (Feature 1) extended
northwards.
Test Unit 1: 20-30 cmbs
Excavation into the deposits between 20 and 30 cmbs cut into and through the
bottoms of layers II, III, IX, and also encountered a new layer, Layer V. The sediment of
each of these layers remained mottled, and charcoal and chalk fragments were frequent
throughout. Patches of orange sediment increase in frequency, and also emerged as a
lens along the southern wall of the unit, at a depth of 25 cmbs (Layer VII). Cutting into
this orange deposit was a well defined hearth feature, the top of which consisted of very
into the unit by 20 cm, creating a semi-circular shape. No other features were
encountered in this level, and artifact frequencies began to drop but still remained
substantial.
80
Artifact Descriptions: 20-30 cmbs
Faunal Remains
Faunal remains were very sparse between 20 and 30 cmbs. Only 8 g of riparian
clam and snail shell (Batissa violacea and Melanoides aspirans) were recovered.
Ceramics
Over 1 kilogram of ceramics were recovered from this level, consisting of 136
fragments. Thirty-seven (27.2%)of these fragments were decorated with wiped necks,
crossed-paddle impressed, parallel-paddle impressed, and incised designs. A single
fragment of a bowl rim and six jar rims were also recovered and morphologically
reconstructed. The single bowl (# 417) was of average size, and had a square lip and
straight rim course (Figure 2.12). The jars were extremely variable. Jar # 379 was very
unique, in that it was of small size, with a rectangular-shaped mouth and tall straight rim,
and a squat body that was impressed with crossed-paddle designs (Figure 2.13). This
vessel may have been used for pouring liquid, or storing some other kind of liquid or oil.
The other jars were typical of cooking vessels, but varied in their rim and lip
morphology. Two vessels (#s 400, 401) had inward curving rims (Figure 2.14). All were
tempered with alluvial sand.
Lithics
A single fragment of ground basalt and a chert core was recovered from Level 3.
In addition, two flakes, one each of basalt and chert, were encountered in the fill deposits.
These fragments appear to be debitage from stone tool manufacture.
81
Bowl # 417, 20-30cmbs
Temper Class: 2
Figure 2.12. Bowl from 20-30 cmbs, Test Unit 1.
Figure 2.13. Unusual square-mouthed jar from 20-30 cmbs, Test Unit 1.
82
Jar # 400, TU 1, 20-30 cmbs
Temper Class: 2
Jar # 401, TU1, 20-30 cmbs
Temper Class: 1
Jar # 404, TU1, 20-30 cmbs
Temper Class: 2
Jar # 405, TU1, 20-30 cmbs
Temper Class: 1
Jar # 411, TU1, 20-30 cmbs
Temper Class: 1
Figure 2.14. Jars from 20-30 cmbs, TU1.
83
Test Unit 1: 30-40 cmbs
As the excavation continued into the deposits that lay between 30 and 40 cmbs,
the orange lens (Layer VII) from the southern portion of the unit was observed to extend
to the north for approximately 25 cm, at which point it dipped sharply down and
disappeared under the overlying layer IV that constituted the rest of the unit. Also of
note, another layer of unknown origin was identified as lying directly atop the orange
horizon, and it consisted of a thin layer (approx. 2 cm) of reddish-brown sediment. This
deposit (Layer V) was very soft and loose, and could be brushed away by hand. As the
excavation of the orange deposits surrounding the hearth feature continued in the
southern portion of the unit, this thin reddish-brown layer could be observed lining the
carved basin that served as the foundation for the hearth. Laboratory analyses of Layer V
indicated that this deposit was a loamy sand, with a calcium carbonate level of 7.8%, an
organic content of 13%, and a pH level of 7. These results suggest that this thin lens
represents an occupational layer that was deposited immediately prior to the construction
of the hearth feature.
Also of importance is the character of the basal layer, Layer VII. No artifacts
were encountered in the orange deposit, and its inclusion-free matrix further suggested
that this sediment may actually be eroded limestone, as Palmer had suggested when they
encountered this basal layer. Chemical analyses of this deposit indicate that this is likely
the case, as only 2.5% of the calcium carbonate remained in this material. In addition,
the pH level for this deposit was very acidic (4), which further suggests that the orange
basal layer was the original limestone floor of the cave, which has since eroded away into
a sandy conglomerate of clay minerals. A small amount of riparian shell, ceramics, and
84
lithics were recovered from between 30-40 cmbs. These artifacts were encountered
within the fill immediately above the eroded limestone, and most likely are derived from
layers IV and V in the central portion of the unit, and also the large disturbed deposit
(Layer IX), located in the extreme northwest corner (Figure 2.1).
Artifact Descriptions: 30-40 cmbs
Faunal Remains
A trace amount (2 g) of riparian snail shell (Melanoides spp.) and toto nutshell
(Aleurites moluccana) (2 g) was recovered from the general deposits of Level 4.
Ceramics
Fifty-two ceramic fragments were recovered from Level 4. Only 11 (21.1%) of
these fragments contained decorations, including finger-gouging, crossed-paddle
impressing, parallel paddle impressing, incising, and wiping. Like previous levels, these
designs are typical of late prehistoric Fijian ceramics.
Only a single bowl rim was recovered that was complete enough to allow for the
analysis of the morphology of the parent vessel. Illustrated in Figure 2.15, this bowl was
shallow in form, with a square lip and diameter of approximately 25 cmbs.
Lithics
Two adze fragments were recovered from the deposits that lay between 30 –40
cmbs. Both adzes were composed of aphanitic basalt, but were too fragmented to allow
the determination of their original cross-section morphologies. The largest adze fragment
had been retouched to create a sharp serrated edge, and was probably used as a cutting
tool.
85
Bowl # 433, TU1, 30-40 cmbs
Temper Class: 1
Figure 2.15. Bowl from 30-40 cmbs, TU1.
Test Unit 1: Feature 1
As described, Feature 1 consisted of an ash-filled hearth that was approximately
40 cm wide, and extended out from the south wall of the unit to reveal the most northerly
20 cm of the feature (Figure 2.16). In profile, the top of the ash in the hearth deposit
occurred at approximately 20 cmbs, although much of this material was removed with the
excavation of previous levels. During the occupation of the cave, the base of the hearth
was carved into the sterile orange basal layer, reaching a maximum depth of 55 cmbs.
However, this feature is intrusive into Layers IV, V, and VII, as is clearly seen in the
profile for the south wall of the unit (Figure 2.17). This being the case, the deposits
within the hearth feature correspond to occupations that are associated with Layer II,
which extends as a horizon across the surface of the cave at a depth of approximately 15
cmbs.
Feature 1 was excavated separately, and all the materials within the hearth were
removed prior to continuing the excavation of deeper levels. The bulk of the fill within
the hearth was ash, 41% was calcium carbonate that was probably deposited in the hearth
following the burning of organic materials over limestone cooking stones. Organic
content was also high, at 14% (Figure 2.3). A portion of a globular vessel (# 335) (Figure
86
Figure 2.16. Hearth (Feature 1) in Test Unit 1, at 40cmbs. Hearth was created by the
excavation of a small pit in the sterile eroded limestone layer. Hearth was filled with ash
and charcoal, and also the remains of a single cooking vessel (Jar # 335).
Figure 2.17. Profile of hearth (Feature 1). Image shows from surface to approximately
65 cmbs.
87
2.18) was visible in the ash at 44 cmbs. Nineteen plain body fragments, likely from the
same vessel, were also recovered from the deposit. This vessel was undoubtedly used for
cooking, and its body had a diameter of approximately 36 cm. The neck of the vessel
was curved slightly inward, and its rim was contracted in profile, and slightly pointed at
the lip.
The hearth feature also contained 4 g of riparian clam and snail shell (Batissa
violacea and Melanoides spp.) amongst the ash deposits, as well as the shell of an
unidentified marine mussel, possibly kuku (Modiola tulipa) (Appendix B). A single large
fish bone from an unidentified species was also recovered from the bottom of the hearth.
Radiometric Dating Results: Feature 1
The largest charcoal fragment recovered from Feature 1 was submitted for 14C
dating as sample Wk-11137 (Table 2.3). The resulting date suggests that the wood
fragments deposited in the hearth were cut at 968 ± 52 years BP. After calibration, this
result indicates that the hearth has a 68.2% probability (1σ) of dating to between Cal AD
1020 and 1160, and a 95% probability (2σ) of dating to between Cal AD 1000 and 1220.
This date corroborates the single date for the deepest cultural deposits of the 1972
excavation, which suggested that Tatuba Cave was occupied as early as AD 1000. This
date also indicates that habitations and hearths along the wall of the cave were placed
atop the eroded limestone ‘bench’, which by AD 1000 was soft enough to be excavated
and used for hearth foundations.
88
Jar # 334, TU 1, Feature 1 (35 cmbs)
Temper Class: 1
Figure 2.18. Morphology of jar recovered from Feature 1, TU1.
Table 2.3. AMS and radiometric dates from Tatuba Cave excavations (including 1972).
Lab No.
Provenience
Material
Calibrated Age
Calibrated Age
Measured 13C/12C Range Probability Range Probability
14
C Age BP Ratio
Distributions (1σ)* Distributions (2σ)*
GaK4311
Loc. B, 54
cmbs
Wood
1000 ± 70 ?
Cal AD 970 - 1160 Cal AD 890 - 1220
968 ± 52
Cal AD 1020 - 1160 Cal AD 1000 - 1220
Wood
Wk-11137 TU1, Feature 1 Charcoal
(45-55 cmbs)
-28
AA50298
TU1, 5060cmbs
Wood
Charcoal
1294 ± 40 -25.6
Cal AD 685 - 780
AA50294
TU1, 80-90
Wood
Charcoal
1993 ± 35 -26
Cal BC 20 - Cal AD Cal BC 50 – Cal AD
80
130
AA50293
TU 1, 130-140 Human Bone 339 ± 40
(intrusive
cmbs
burial)
AA50292
Wood
TU1, 140-150 Charcoal
cmbs
Wood
**AA50291 TU1, 190-200 Charcoal
cmbs
-20.6
Cal AD 660 - 880
Cal AD 1510 - 1650 Cal AD 1480 - 1660
1802 ± 34 -25.5
Cal AD 210 – 340
207 ± 31
Cal AD 1650 – 1950 Cal AD 1640 – 1950
-28.1
Cal AD 130 - 380
* All determinations were calibrated using the terrestrial calibration curve of Stuiver et al.
(1998), with 27 ± 5 years subtracted from the CRA to account for the southern
hemisphere offset in 14C (McCormac et al. 1998).
** This date was has been rejected from the chronology due to the fact it is a contaminant
from a later strata.
89
Test Unit 1: 40-50 cmbs
Past 40 cmbs, excavation continued into the sterile Layer VII, which extended
across the southern portion of the unit, and also into Layer IV, which was present in the
central and northeastern portions of the unit (Figures 2.1, 2.2). Throughout the unit the
mottled character and consistency of Layer IV remained the same, with similar amounts
of charcoal and chalk. However, rocks and cobble sized pebbles increased to 20% of the
fill, and they were especially common in the northern portion of the unit. As indicated by
Figure 2.2, these cobbles were probably associated with the intrusive burial deposit
(Layer IX) that extended downwards through the north and west profiles. However, the
bulk of the deposits from Layer IV located in the center and eastern portion of the unit
appeared to be intact. Unfortunately the boundary of the intrusive burial was difficult to
detect during the excavation, thus the artifacts that were contained within its mixed
deposits were also collected during the excavation of each 10cm level.
Between 40 and 50 cmbs, Layer IV appeared to be extremely mottled, as lenses of
the orange basal layer appeared to be mixed with an intrusive post-hole, which is visible
on the west profile. Chalk and charcoal continued to be common, as were small
limestone pebbles in the mixed fill. However, artifact frequencies collected from this
level continued to suggest a steady decline, although a variety of material types were
encountered.
90
Artifact Descriptions: 40-50 cmbs
Faunal Remains
As in previous levels, 7 g of riparian clam and snail shell (Batissa violacea and
Melanoides spp.), and 7 g of toto nutshell (Aleurites moluccana) were recovered from the
deposits between 40-50 cmbs (Table 2.2).
Ceramics
Ceramic frequencies are also similar to previous levels. Twenty-five ceramic
fragments were recovered, 5 of which were decorated with incising, wiping, and crossedpaddle impressed designs. However, none were complete enough to allow for the
analysis of vessel morphology.
Lithics
Stone tool fragments were also increasingly scarce between 40-50 cmbs. Only a
single flake of grey chert was recovered from these deposits (Table 2.1).
Test Unit 1: 50-60 cmbs
Between 50 and 60 cmbs, Layer IV retained its mottled character, and also
demonstrated an increase in the number of limestone pebbles and cobbles. The sediment
was also increasingly loose, and excavation could be easily accomplished with a brush.
Of note, the eastern portion of the unit contained a higher proportion of charcoal-stained
sediment, which appeared to be due to the increased crumbling of a hearth deposit that
lay to the east of the unit. The profile of this hearth deposit in the eastern wall of the unit
became more visible as the excavation progressed to 60 cmbs, and it was at this depth
that this profile began to erode into the unit.
91
Artifact Descriptions: 50-60 cmbs
Faunal Remains
Artifact frequencies collected from between 50 and 60 cmbs were similar to the
previous level. Two fragments of riparian snailshell (Melanoides spp.), as well as a few
fragments of toto nutshell (Aleurites moluccana) were also retrieved from the deposit
(Table 2.2).
Ceramics
Twenty-four ceramic fragments recovered from Level 5. Of these, two were
decorated with crossed-paddle impressed designs, and one had a wiped surface.
However, none were complete enough to allow for the reconstruction of the morphology
of the parent vessel.
AMS Dating Results from Test Unit 1: 50-60 cmbs
Despite the low artifact frequencies after 50 cmbs, charcoal remained abundant
throughout all layers. Sixteen grams of charcoal were recovered from Layer IV at this
level, and the largest fragment was submitted for AMS dating as sample AA50298. The
resulting date indicates that this charcoal is 1294 ± 40 years BP in age (Table 2.3). At 1σ
calibration, this sample indicates an age range of Cal AD 685 to 780 for the deposits that
occurred between 50 and 60 cmbs. The analysis of the sediments from Layer IV at this
depth also supports a cultural origin for this deposit, as 20% of it was organic in content
(Figure 2.3). Thus, this date indicates that occupations in Tatuba Cave pre-dated the AD
1000 date that was established by Palmer et al. for the occupations atop the eroded
limestone shelf. Moreover, the location of this date in Layer IV suggests that the central
92
and deepest portion of the cave contains cultural deposits that are considerably older than
the occupations along the cave walls, and atop the ancient eroded limestone shelf.
Test Unit 1: 60 –70 cmbs
As excavation continued past 60 cmbs, an additional deposit of sterile, eroded
limestone was encountered in the southern portion of the unit. This layer (Layer VIII)
had a very strong grade, and medium/coarse sized peds that were round and bubble-like
in form. It was also very lightly pigmented, with a Munsell color rating of 10YR 7/4
(very pale brown). Analysis in the laboratory determined that this deposit had an
extremely small amount of calcium carbonate (1.8%), and a very acidic pH level of 3.5.
These results suggest that this deposit was also once limestone, and over the ages the
precipitation of acid through the profile has eroded away the CaCO3 that formerly
constituted the bulk of the limestone. The top of this deposit lies directly below the base
of the hearth feature, and also the orange Layer VII in the southern portion of the unit.
However, as with Layer VII this deposit did not extend further into the northern portion
of the unit, and plunged steeply downwards towards the bottom of the unit.
Excavation continued however in the center and northern portion of the unit, and
the fill of Layers IV and IX at 60 to 70 cmbs continued to be mottled and laden with
pebbles and cobbles. The fill at this level of the excavation was very loose and dry, and
continued to have a high organic content of 20%. In addition, calcium carbonate
continued within the deposits despite the high pH level of 2.9. Human remains from the
intrusive burial in the northern portion of the unit were also encountered in this unit, and
are described below.
93
Artifact Descriptions: 60-70 cmbs
Human Remains
Fragments of human bone and fiber mat were recovered from Layer IX at a depth
of 60-70 cmbs (Figure 2.19). These remains (excluding the hair and mat fragments) were
photographed and then reburied at the site. The remains included fragments of the
cranium and femur, and preliminary analyses indicated that the remains consisted of a
single adult of undetermined sex. Fragments of a fiber mat adhered to the shaft of the
femur, which suggests that the individual had been wrapped within a map prior to burial.
Faunal Remains and Ceramics
As with previous levels, the fill of Layers IV and IX at 60-70 cmbs had a small
amount of cultural material. This included 12 g of charcoal, 1g of toto nutshell
(Aleurites moluccana), and 17 ceramic fragments (Tables 2.1, 2.2). Six of the fragments
were decorated with crossed-paddle impressed and incised designs, however none were
complete enough to allow for the analysis of the parent vessel’s morphology.
Test Unit 1: 70-80 cmbs
Between 70 and 80 cmbs, the deposits of Test Unit 1 remained similar to those of
the previous two levels, however a large amount of materials associated with the intrusive
burial were recovered. At this point the excavation was almost exclusively confined to
Layer IX, which continued downwards in the northern portion of the unit. The southern
portion of the unit was almost entirely composed of rock rubble and boulders,
presumably from roof-collapse. As before, small amounts of toto nutshell, charcoal, and
riparian snailshell were recovered. Two fragments of unidentified bone (possibly
94
Figure 2.19. Human remains from 60-70 cmbs. Long bone fragment is encrusted with a
fiber mat.
associated with the burial) were also recovered. A single fragment of chert was also
present in the deposits, and 14 ceramic fragments, 4 of which were decorated with paddle
impressed and incised designs (Table 2.1). A single rim fragment was complete enough
to allow for morphological analysis. As indicated in Figure 2.20, the jar had a thin wall,
and curving rim profile that ended in a square lip.
The human remains recovered from Level 8 were considerably more numerous
than the previous level (Table 2.4). Although not thoroughly studied prior to reburial, the
fragments appeared to include portions of a tibia, the cranium, a clavicle, rib, and
zygomatic arch. In addition to these items were several fragments of mat fiber, and also
human hair. The small amounts of hair were black in color and slightly curled, and
approximately 1.5 cm in length.
95
Jar # 456, TU1, 70-80 cmbs
Temper Class: 1
Figure 2.20. Jar recovered from 70-80 cmbs, Test Unit 1.
Table 2.4. List of human remains recovered from intrusive burial, Test Unit 1.
Layer
Level Depth Element
(cmbs)
Number Weight
(g)
III/IV/IX
7
1
1
1
18
1
1
1
1.2
IX
9
2
3
IX
10
*crania fragments
*femur fragment
70-80 *tibia fragment
*crania fragments
*clavicle fragment
*rib fragment
*zygomatic arch fragment
hair
80-90 *crania fragments
hair
90-100 *crania fragments
hair
*mandible fragments
*2 molars
*vertebrae
*os coxae fragments
1
1
IX
IX
11
14
100-110 *crania fragment
130-140 hair
*crania fragments
*vertebrae
4
5
60-70
* remains reburied at site
96
Test Unit 1: 80-150 cmbs
As mentioned, by 70 cmbs large boulders and rubble began to appear in the
southern portion of Test Unit 1, and between 80 and 150 cmbs these rocks completely
filled the bottom of the unit, and extended downwards in a staircase fashion to the
northern portion of the unit. At 120 cmbs several large boulders were removed from the
fill, and it was noted that the eroded limestone of Layer VIII lie directly atop this layer of
boulders and cobbles. By this point in the excavation, the walls of the unit became very
unstable and all except the southern wall collapsed to some extent. The materials that fell
into the unit were discarded; however, of note was a large (30 cm) wooden torch or stake,
which fell out from beneath the hearth feature in the eastern wall (Figure 2.21). This
artifact quickly disintegrated, but the burned end and cut surface of the torch was clearly
visible.
In the northern and central portion of the unit the deposits continued to consist of
mottled brown fill from the lowest portion of Layer IV, and also the loose cobble-laden
fill of the intrusive burial (Layer IX) in the northwest corner of the unit. The looseness of
the deposits, and also the presence of large amounts of limestone pebbles and cobbles
indicated that the northern portion of the unit was disturbed throughout. Bone fragments
(later identified in the laboratory as human remains), charcoal, toto nutshell, a polished
adze fragment, ceramics, lithics, and riparian clam and snail shell continued to occur in
small amounts throughout the deposit, and the frequencies of these artifacts are indicated
in Tables 2.1, 2.2, and 2.4.
Figure 2.22 illustrates the variety of bowls that occurred within the disturbed
deposits of Layer IX. Bowl #s 465, 475, and 491 all have square lips, although #491 has
97
Figure 2.21. Fragment of a wooden torch that emerged from the eastern profile following
the collapse of the unit wall at approximately 80cmbs. The torch may have been located
at a depth of 30-50cmbs.
Bowl # 465, TU1, 90-100 cmbs
Temper Class:1
Bowl # 475, TU1, 110-120 cmbs
Temper Class: 2
Bowl # 491, TU1, 130-140 cmbs
Temper Class: 2
Figure 2.22. Bowls from 80-150 cmbs, Test Unit 1.
98
a slightly expanding rim profile. These vessels are all alluvial sand tempered, although
with larger (2-4 mm) sized grains. Jar #s 473 and 489 are quite different, with square and
rounded lips and a great degree of difference between the inclination angle of the neck
(Figure 2.23). Jar #489 has a much sharper neck, and is also tempered with fragments of
limestone, as opposed to alluvial sand.
AMS Dating Results: 80-150 cmbs
Due to the disturbed nature of these deposits, several fragments of charcoal and
human bone from the intrusive burial were submitted for AMS dating. The results
indicate that the charcoal fragments from between 80 – 90 cmbs, and 130-140 cmbs were
stratigraphically inverted: sample AA50294 from 90 cmbs resulted in a date of 1993 ±
35, which calibrates at 2σ to between Cal BC 50 and 130 Cal AD (Table 2.3). Of similar
age, sample AA50293 from 150 cmbs resulted in a date of 1802 ± 34, which calibrates at
2σ to between Cal AD 130 and 380. The fact that these samples are inverted and also
separated by 60 cm suggests that they are the remnants of what was likely the earliest
occupation layer in the cave, which was later disturbed and redeposited by the intrusive
burial in the northern portion of the unit. Also of note, a single fragment of human
cranium from the burial was recovered from 140 cmbs. This fragment (sample
AA50293) was submitted for AMS dating, and the results indicate that the death of this
individual occurred at 339 ± 40 BP. After calibration, this date indicates that the burial
and subsequent disturbance of the deposits occurred between Cal AD 1480 and 1660
(2σ).
99
Jar # 473, TU1, 110-120 cmbs
Temper Class: 1
Jar# 489, TU1, 120-130 cmbs
Temper Class: 7
Figure 2.23. Jars from 80-150 cmbs, Test Unit 1.
Test Unit 1: 150-190 cmbs
Figure 2.24 indicates the visible stratigraphy at 150 cmbs. Excavation continued
with great difficulty past this depth. The sediment was extremely loose and dry, and
voids between large cobbles and boulders were very common. The percentage of rocks
and boulders on the floor of the unit increased as the excavation continued in depth, and
the area of excavation steadily decreased to only a small portion in the northern part of
the unit. Ceramics, toto nutshell, charcoal, wood fragments, and riparian clam and
snailshell continued throughout the deposits until a maximum depth of 170 cmbs (Table
2.1). Portions of the walls continued to collapse, and in several instances the deposits
from the profiles collapsed into the bottom of the unit and had to be removed. By
100
Figure 2.24. Photos of the stratigraphic profiles from Test Unit 1. Collapse of the
stratigraphy was most problematic in the northern and eastern walls, as shown above.
190 cmbs, the excavation was restricted to the extreme northern portion of the unit, and
only fragments of charcoal and toto nutshell were recovered. However, it seems likely
that these fragments were contaminants from collapsing upper layers (i.e., Layers II and
III), as a single fragment of charcoal collected from the bottom of the unit provided an
AMS date of 207 ± 31 BP (Cal AD 1640-1950, sample AA50291). This date for Tatuba
Cave has been discarded as it most likely originated in much later deposits.
101
The Chronology of Tatuba Cave
The results of the 2002 excavation reveal that Tatuba Cave has been occupied for
the past two millenia. Evidence for the earliest occupations in contained in Layer IV. If
conditions within Entrance 2 serve as a guide, the limestone base of Entrance 1 is also
‘keyhole’ shaped, and thus Layer IV lies within the base ‘gutter’ of this formation.
Occupations first occurred upon this deposit ca. Cal BC 20. Later centuries saw the slow
accumulation of guano in the cave, which was acidic enough to decompose the limestone
floor, and thus produce the eroded benches that parallel the wall of the cave (Layers VII
and VIII). Although the disturbance caused by the burial (Layer IX) makes it difficult to
determine if occupation was continuous through time from Cal BC 20 onwards, the
cultural deposits and dates from Layer IV suggest occupations were present during the 7th
century AD. In addition, the cultural deposits that top Layer VII ( i.e., the hearth feature,
and Layers I, II, and III), can be securely dated to ca. AD 1000. In addition, the presence
of ceramics, lithics, faunal remains, and the burial during the 16th century AD also
suggest that occupation at the cave was constant. Future excavations focused on the
features outside the cave will undoubtedly show evidence of additional occupations
during these time periods.
These data, in tandem with the 1972 excavations, indicate that Tatuba Cave was
perhaps the most ancient and important population center in the region. Cultural
materials suggest that the inhabitants had access to diverse food items, such as riparian
clams, which must have been imported from the Sigatoka or Bā river delta, and marine
mussels and pearl shell. The inhabitants also manufactured or exchanged for ornaments,
such as marine-shell armbands, pearl shell plates, and perforated dog-teeth. The
102
extensive size of Tatuba and its associated sites, and also the number and variety of yavu
and architecture, indicate that Tatuba supported a sizeable population. Without doubt,
the sheltering capabilities of the cave attracted early inhabitants, however, the locale is
also blessed with a variety of environmental resources that would have been desirable to
newcomers, and also encouraged continued occupations.
GIS-based Environmental Analyses in the Sigatoka Valley
Land classifications performed by the Fijian Dept. of Agriculture indicate that
most of the soils of the valley exist on steep and mountainous slopes, although the upper
reaches of the drainage are considerably more rugged than the Sigatoka delta. The river
has carved a variety of geological formations through the valley’s interior, including
raised planation surfaces in the upper valley, and expanses of alluvial terraces in the
lower valley. Much of the vegetation of the region consists of grasses and shrubs,
although pockets of forest exist in the drainages, and atop limestone peaks.
In prehistoric times, the inhabitants of the Sigatoka Valley subsisted on
hydrophilic crops (plants that grow in partially submerged conditions) such as dalo
(Colocasia esculenta) and via kana ‘giant swamp taro’ (Cyrtosperma chamissonis), and
dryland crops such as uvi (Dioscorea spp.), and dryland taro (Barrau 1961; Kirch 1994;
Spriggs 1982). These crops require particular conditions for high yield cultivation. With
adequate amounts of moisture and silty soils, dalo can yield approximately 20+
tons/hectare/year, and can be planted year round, and harvested every 8-10 months
(Brookfield 1979, in Spriggs 1984:129). Earthworks, such as ponded-fields (vuci) and
terraces with irrigation (tabawai) provided by nearby streams, were often installed to take
103
advantage of naturally productive environments and encourage vibrant growth. In
contrast, uvi and other dryland crops were cultivated in swiddens (slash and burn
gardens) that were well-drained and occasionally terraced to retard erosion. With
adequate amounts of rainfall, uvi and dryland taro yield between 5-15 tons/hectare/year
(Kirch 1994:8; Tindall 1983:203). However, unlike dalo, uvi are only planted at the
beginning of the wet season, and can be harvested 7-9 months later.
Sedimentological analyses of the Sigatoka Valley indicate that most soils would
have provided enough nutrients to allow for cultivation of both uvi and dalo. However,
the severity of slope, in particular the higher elevations and portions of the upper valley,
would have significantly reduced the amount of land available for cultivation. In
addition, the effects of the dry season would have made the cultivation of dalo impossible
in all but the most well-watered portions of the valley. Figure 3.1 provides a summary
image of the environmental zones of the Sigatoka Valley, based upon soil fertility and
dry season intensity. These data also suggest the existence of three zones of productivity
and agricultural risk: Zone 1) lands suited to dryland cultivation, and which also
experienced shortfalls due to severe dry seasons; Zone 2) lands suited to mixed wetland
and dryland cultivation, and which also experienced mild shortfalls due to moderate dry
seasons; and Zone 3) lands suited to wetland cultivation, and which also experienced
minimal shortfalls due to weak dry seasons.
In the short view, the results of these analyses simply suggest that the inhabitants
of the valley must have privileged one mode of cultivation over the other. However, it
also suggests that portions of the Sigatoka Valley were more restricted in terms of
cultivation strategies, and more at risk from severe, long-term droughts. These results
104
105
Figure 3.1. Environmental zones of the Sigatoka Valley, as identified by soil fertility and severity of dry
season.
suggest that Zones 2 and 3 would have allowed for more temporally stable forms of
agricultural production, and Zone 1 would have been more susceptible to episodic
environment-based shortfalls. This variation is particularly important in light of recent
observations of global climatic phenomena, in particular the El Niño Southern Oscillation
(ENSO). Archaeological and historical investigations from other parts of the world (e.g.,
South America, the Middle East) have traced ENSO-related events to as early as 6000
BP, and suggest that frequencies assumed essentially modern ranges ca. 3000 BP.
(Caviedes 2001; Sandweiss 2002). According to climatologists (e.g., Nichols and
Wong1990; Salinger et al. 1995) Fiji is strongly affected by ENSO-related disturbances,
and modern droughts and cyclones (such as the 1982/83 and 1997/98 seasons) provide
models for the severity of ENSO cycles in prehistory. During these periods, rainfall was
22-42% of normal, and national production loss for sugarcane was 50%. Economic
analyses also suggest that the average income for subsistence farmers dropped from
F$3500 to F%1500 per annum (Kaloumaira 2000). If modern experienced serves as any
guide, the Sigatoka Valley (in particular Zone 1) was subject to devastation during a
severe ENSO episode. The antiquity of ENSO in the Pacific further suggests that this
cycle of devastation repeated itself throughout Fijian prehistory, providing the impetus
for human conflict over resources.
Paloeclimatic data from New Zealand caves as well as shell conglomerates in
Fijian shorelines indicate that dramatic environmental perturbations occurred in Fiji
approximately 700 years ago. These events mark the transition period between the
warmer Little Climatic Optimum (1250-700 BP) and cooler Little Ice Age (700-200 BP).
According to recent analyses in Fiji and Niue by Nunn and others (Nunn 2000a, 2000b,
106
2003; Nunn et al. 2003; Nunn and Britton 2001), the disturbances of the transition were
quite dramatic, and were marked by cooler weather and sea temperatures, and a rapid
lowering of sea level by over one meter between 730-525 BP. Nunn and Britton
proposes that these changes were devastating for both coastal and interior populations:
sea-level fall would have converted embayments to brackish wetlands, and also caused
the submersion of groundwater for the interiors of the Fiji Islands. Extensive incising of
the riverbeds and streambeds, as well as decreased soil moisture levels in many of the
alluvial terraces that supported uvi cultivation, would have resulted in dramatic reduction
of uvi yields in the lowlands. Torrential rains and floods may have also occurred in the
higher elevations.
Figure 3.2 combines the results of the analyses described above and allow for the
formulation of the following conclusions: 1) Zone 1 would have experienced the most
severe effects of ENSO-related droughts and floods, and also suffered from sea-level fall
during the LCO/LIA transition; 2) Zone 2 would have experienced moderate ENSOrelated droughts, and been affected to a limited degree by the landslides and erosion
associated with the LCO/LIA transition; and 3) Zone 3 would have only been weakly
affected by ENSO-related droughts or floods, and experienced a high/moderate amount
of erosion during the LCO/LIA transition.
Tatuba Cave: Site Functions and Placement within the Sigatoka Valley Chronology
Tatuba Cave had immediate access to small tracts of land suited to dryland
cultivation. However, Tatuba is also within a three kilometers of much larger expanses
of alluvial terraces in the vicinity of Sawene and the confluence of the Namada river. In
107
Figure 3.2 Comparison of the environmental zones identified by soil/topographic
analyses with the extent of disturbance related to ENSO and the LCO/LIA transition.
would have experienced severe ENSO-related droughts.
108
this case, it may be more accurate to suggest that Tatuba Cave was associated with
distant (e.g., 2-3 km) production areas. These areas are divided by soil quality and
rainfall patterns in Zones 1 and 2, which suggests that production would have been
susceptible to decreased yields during severe ENSO floods, but largely buffered against
drought. The presence of extensive archaeological remains of vuci (ponded-fields) in this
region also indicates that the topography and hydrology was amenable to high-yield
production of dalo. Additional areas suited to cultivation lie to the immediate north and
east of the Tatuba, atop elevated planation surfaces.
The extent and quality of resources near Tatuba, and also the natural topographic
defenses of the location, suggest that the site served as an ideal base for occupation and
defense. Hence, Tatuba fits the expectations of a territory as defined by the evolutionary
ecology model, and served as a stronghold in association with dense and predictable
resources. In addition, the conditions of the LCO/LIA transition may not have been as
severe in this region, as Davies’ research suggests that groundwater is much closer to the
surface in this portion of the valley (Davies 1992). Thus, agricultural production may
have remained relative stable during this period. However, the effects of the transition on
the entire Sigatoka Valley are recorded at Tatuba, as it was during this time that the cave
was fortified with a wooden palisade.
Figure 3.3 displays the range of dates associated with the 12 other sites excavated
in the Sigatoka Valley. Four temporal periods are identified: T1 (2000-1500 BP), T2
(1500-1000 BP), T3 (1000-500 BP), and T4 (500-150 BP). The period proposed for the
transition between the Little Climatic Optimum (LCO) and the Little Ice Age (LIA) is
also indicated in the diagram. The location of these sites in the Sigatoka Valley, and their
109
Figure 3.3. Calibrated ages for the 2001-2002 Sigatoka Valley excavations. Temporal
periods T1-T4 and the position of the LCO/LIA transition are indicated. All
determinations were calibrated using the terrestrial calibration curve of Stuiver et al.
(1998), with 27 ± 5 years subtracted from the CRA to account for the southern
hemisphere offset in 14C (McCormac et al. 1998).
110
placement into one of four temporal categories, is indicated in Figure 3.4 In brief, the
results of archaeological investigations and environmental analyses suggest the following
chronology for the Sigatoka Valley:
The Establishment of Territorial Fortifications and Refuges
The earliest occupation in the Sigatoka Valley occurred at Tatuba Cave between
Cal BC 20 and Cal AD 80. GIS-based environmental analyses demonstrate that Tatuba
Cave is located within a region that had moderately high soil fertility, and which also
experienced a moderate dry season. In addition, Tatuba Cave had immediate access to
small tracts of land suited to dryland cultivation, and was also within a few kilometers of
much larger expanses of alluvial terraces suitable for vuci construction and dalo
cultivation. Of note, the site is located within Zone 2, but a substantial portion of its
resource base is located in Zone 1. Although Zone 1 would have been susceptible to
flooding and drought during severe ENSO cycles, the surrounding region was generally
buffered against extreme climatic phenomena. The size of the entire Tatuba Cave locale
(area, and also number of yavu and architectural features) also suggests that the area
remained desirable as an occupation site during the following two millennia (T1-T4), and
the strategy of territorialism persisted in the region.
Similarly, the sites of Nokonoko and Qoroqorovakatini suggest an association
between the founding of territorial strongholds and refuges as a response to dense and
predictable resources and environmental shortfalls. Both of these sites consist of remote
and naturally defended peaks that are over 240 m above the valley bottom, and contain
complex archaeological features and deposits of substantial antiquity. Nokonoko
111
112
Figure 3.4. The distribution of earliest occupations for each excavated site according to temporal period.
was established as early as the 6th century AD (1492 ± 43 years BP, or T1) and
Qoroqorovakatini was constructed several centuries later (974 ± 43 years BP, or T2). In
the case of Nokonoko, the range of dates, the high frequency of yavu, and also the deep
middens of clamshell and ceramics suggest that the site was occupied throughout the
prehistoric period. Thus, this site most closely represents a territorial stronghold that was
occupied year-round. In contrast, Qoroqorovakatini contains deposits that indicate
episodic occupations, hence the site may represent a combination of remote refuge and
territorial stronghold. However, the long chronologies of Nokonoko and
Qoroqorovakatini reflect the persistence of a territorial and refuge strategy through all
temporal periods (T2-T4) relating to environmental limitations and climatic perturbations
in Fiji. Of note, the site of Korokune reflects a similar adherence to the strategy of
territorial stronghold, although this site was not established until the third temporal
period.
Relocation to Environmental Refuges During Periods of Disturbance
Four of the 12 excavated sites were established during the third temporal period
(1000-500 BP), and also within the period proposed for the transition between the
LittleClimatic Optimum and Little Ice Age (1250 – 700 BP). The increase in site
frequency during this period suggests significant population growth between periods T1
and T3. More importantly, the location of three newly founded habitations in Zones 2
and 3 suggests the development of a new strategy of habitation and subsistence that is
directly related to environmental disturbance. In particular, the environmental conditions
surrounding these sites, their topographic positions in remote drainages, and also the
113
presence of faunal material indicating a diversified resource base, are suggestive of a
distinctive and new settlement pattern focused on environmental buffering.
Malua, and Korovatuma (and perhaps also Bukusia) are emblematic of this new
trend. Both were established between AD 1300-1630, and are located atop dramatic
outcrops that provided natural defense. The surrounding lands were of moderate quality
but extraordinarily rugged. Cultivation of dalo was certainly possible, but the arable land
was restricted to ‘pocket gardens,’ (i.e., small isolated patches along minor creeks).
Neither site had access to the large tracts of alluvial deposits that were available in earlier
periods (e.g., Nokonoko, Tatuba Cave, and Qoroqorovakatini). However, environmental
analyses place these sites within Zone 2, which indicates that the resource base
surrounding these sites would not have been affected by severe drought or a reduction of
soil moisture due to sea-level fall predicted for the LCO/LIA transition. Unlike
significant portions of the valley bottom (Zone 1), cultivation of dalo and uvi in Zones 2
and 3 would have continued as normal. The timing of the founding of the sites, as well as
quality of the surrounding environment, strongly suggest that these new habitations
represent a strategy of environmental refuge relating to the effects of the LCO/LIA
transition. Cumulatively, these data suggest that the LCO/LIA transition encouraged
populations to optimize in a new way: 1) less investment in large but risk-laden
agricultural surplus, and increased focus on temporally stable resources; 2) variable
patterns of population aggregation and dispersal, perhaps coupled with mobility; 3)
diversification of the resource base; and 4) exchange or travel for non-local food items.
114
Constructed Fortifications and Undefended Production Sites
The final phase of Sigatoka Valley prehistory (period T4) witnessed the
persistence of previously established strategies (territorial strongholds, remote refuges,
and environmental refuges), and also the emergence of a new form of fortification on the
valley bottom: villages surrounded by a constructed moat and ditch. The sudden
appearance of this kind of fortification suggests the diffusion of new ideas concerning
defensive construction, and also an increasingly large population. The sites of Vitogo
and Bātiri are a case in point. Both of these sites were constructed between 260-150 BP
and are of the ‘ring-ditch’ type; (i.e., protected with an annular ditch, bank, and palisade).
Unlike previous centuries, these villages did not make use of any topographical features
for defense, but were located within swampy, riparian environments that would have
been prime locations for growing dalo. The size of annular ditch and bank construction
for these sites suggests a significant investment and labor pool: in the case of Bātiri, the
ditch was 10 m wide with a circumference of 390 m. This is a marked difference from
the fortifications of previous centuries, which employed naturally fortified positions that
required minimal initial investment, and which also did not require a sizeable population
for defense or construction tasks. Therefore, the restriction of manufactured fortifications
in the latest period is almost certainly a factor of population size, and perhaps also related
to the transmission of this innovative design throughout Fiji in the late prehistoric period.
The emergence of undefended habitation sites in the valley bottom may also be
related to population growth, and perhaps the integration of small communities with
larger territorial strongholds. The sites of Nadrogā and Korohewa are representative of
this strategy. These habitations were of small size, late age, and were located on the
115
valley bottom. No visible defenses were encountered at Korohewa, and only a modest
wall unrelated to defense was present at Nadrogā. The lack of defenses at these sites
breaks nearly two millennia of tradition for the Sigatoka Valley; for the first time,
villages did not employ topography as the first line of defense against invaders. This
omission strongly suggests that these populations were integrated with regional centers,
or were protected by another form of defense-- armed warriors. Moreover, the presence
of undefended habitation sites may be related to cooperative strategies that were
established during the establishment LCO/LIA transition. In this case, undefended
settlements may have been linked to fortifications maintained by close relatives.
Conclusion
In conclusion, the results of analyses in the Sigatoka Valley suggest that: 1)
fortifications are the result of competitive strategies that originated early in Fijian
prehistory (ca. AD 700) as a response to spatio-temporal environmental variability,
specifically the effects of the El Niño Southern Oscillation; 2) fortifications utilized
natural topography from the earliest period onwards, and this pattern persisted throughout
all temporal periods, with annular ditch style fortifications only occurring in the latest
temporal period (ca. AD 1700); 3) changes in settlement patterns ca. AD 1300 can be
linked to the LCO/LIA transition, as opposed to the influx of migrants (cf. Frost 1974).
The analytical capabilities of a GIS, in particular the analysis of landscapes via slope,
aspect, and viewshed, are shown in this research to have particular value to spatial
analysis, and can be modified to incorporate a temporal aspect. When coupled with
archaeological investigations, this approach allows for extensive analyses of the variables
116
that impinge upon subsistence and habitation. Increasing resolution of the data involved
in these kinds of analyses will undoubtedly multiply the analytical potential for
archaeological studies, and contribute more to studies of Fijian prehistory.
117
Appendix A: Fauna of Tatuba Cave (1-NMT-015)
The following is a list of species and identifying photos for the floral and faunal
remains recovered from the excavations at Tatuba Cave. All photos were taken by the
author unless otherwise specified. Descriptions of molluscan characteristics were
obtained from Parkinson (1982) and Haynes (2001).
Mollusca
Common Name: Tave or Kai waidranu
(Fijian)
Scientific Name: Batissa violacea
Family: Unionidae
Adult Size: 30-90mm
Depth: brackish and freshwater
Location: Common in the lower
Sigatoka, Rewa, Ba, Nadi, Navua,
Korovou, Dreketi, Labasa, and Waikoro
rivers.
Common Name: Sici moto (Fijian)
Scientific Name: Melanoides spp.
Family: Thiaridae
Adult Size: 50mm
Depth: 2-4 feet
Location: Freshwater streams, sandy to
rocky
118
Aves
Chicken (Gallus gallus) or Fijian heron (Egretta
sacra)
119
Appendix B - Ceramics of Tatuba Cave (1-NMT-015)
Ceramic Classification
The ceramics recovered from the excavations at Tatuba Cave were analyzed with
a system that has been developed by previous researchers in Fiji (e.g., Best 1984, Clark
1999, Crosby 1988, Hunt 1980). This classification relies upon the identification of
decorative and morphological attributes of ceramic vessels. It also allows for the
separation of vessels into at least four categories: jars, bowls, trays, and vessels of
unknown form. As outlined in Figures 4.1 and 4.2, the classification records the presence
and morphology of diagnostic features, including rim orientation and rim-body contour
(which indicate the overall form of the vessel as either inverted or everted bowls and
jars), rim course, rim profile, and lip shape (which indicate the morphology of the rim),
rim orientation angle, neck inclination angle, rim height (which describe the morphology
of jar forms), and rim thickness, orifice diameter, and temper type (which indicate the
overall size, fabric, and coarseness of the vessel). Although the latter classes are metric
measurements, numeric codes representing their variability were substituted in the
classification. In addition, the presence or absence of a variety of surface decorations, as
well as their positions on either the lip, rim, or body of the vessel, were included in the
classification (Table 4.1).
When possible, the orientation of the vessel was determined by placing the
fragment on its rim atop a white piece of paper, and then using a bright light to aid the
120
Figure 4.1. Diagram of a ceramic jar, detailing the morphological features (metric only)
that were recorded for the ceramic assemblage of Tatuba Cave.
121
Figure 4.2. Coded classification for the non-metric morphological attributes recorded for
the ceramic assemblage of Tatuba Cave.
122
Table 4.1. List of decorative attributes recorded for the lips and rims of ceramic vessels
from Tatuba Cave.
Decoration - Lip
Plain
End-tool, in rows
Appliqué
Incised design
Decoration – Rim
Plain
Carved paddle, diamond
Carved paddle, square
Carved paddle, parallel
End-tool, in rows
Finger-nail impression
Side-tool notch
Incised design, lines
Appliqué
Finger-gouge/pinch
Grass-impressed
Wiped
Shell-end impressed
Code
0
73
100
101
Code
0
64
65
66
73
77
86
90
100
106
107
108
109
angling of the sherd (indicated by the presence or loss of light under the rim) until the
correct orientation of the rim could be determined (e.g., Joukowsky 1980:423). All of the
diagnostic sherds were drawn in profile, and the measurement of rim orientation and neck
angles was performed upon the drawings with a compass. Sherds that were incomplete
were excluded from this analysis. Sherd thicknesses and rim heights were measured with
vernier calipers, and vessel diameters were determined by comparison of rim curvatures
to a rim diameter chart. Sherd cross-sections were also examined with a 10x hand-lens,
and the size, type, and density of temper grains recorded. Descriptions of the recognized
tempers and included in Table 4.2. The frequency of decorated ceramics by 10 cm level
is indicated in Table 4.3. A list of the morphological and decorative attributes for all the
vessels recovered from Tatuba Cave is included in Table 4.4.
123
Table 4.2. Temper types and variations in density and grain-size recorded in the
classification of ceramic vessels for Tatuba Cave.
Temper type, Density, and Grain-size
Alluvial, 10-20%, 1-2mm
Alluvial, 10-20%, 2-4mm
Alluvial, 20+%, 1-2mm
Alluvial, 20+%, 2-4mm
Pyroxene
Alluvial mixed with beach sand
Alluvial mixed with limestone
Code
1
2
3
4
5
6
7
Table 4.3 Frequency of decorated ceramics from Test Unit 1, Tatuba Cave.
Paddle
Depth Imp.
(cmbs) Parallel
Paddle
Imp.
Crossed
EndIncised tool
Imp.
Shell-end Wiped Finger Grass Total Total Percent
Imp.
Gouge Pressed Dec. Sherds Dec.
0-10
10-20
20-30
30-40
45-55
40-50
50-60
60-70
70-80
80-90
90-100
100110
110120
120130
130140
140150
150160
160170
3
2
1
1
0
0
0
0
0
0
0
5
11
26
6
0
2
2
6
3
4
4
11
6
3
1
0
1
0
1
1
0
0
1
1
0
0
0
0
0
0
0
0
0
0
1
0
0
0
0
0
0
0
0
0
20
15
7
2
0
2
1
0
0
0
0
1
0
0
1
0
0
0
0
0
0
0
0
1
0
0
0
0
0
0
0
0
0
41
37
37
11
0
5
3
7
4
4
4
416
275
136
52
19
25
24
17
14
4
10
9.8
13.4
27.2
21.1
0
20
12.5
41.1
28.5
100
40
0
3
0
0
0
0
0
0
3
10
30
0
10
0
0
0
0
0
0
10
32
31.2
0
0
0
0
0
0
0
0
0
5
0
0
2
0
0
0
0
0
0
2
12
16.6
0
2
0
0
0
0
0
0
2
11
18.1
0
1
0
0
0
0
0
0
1
9
11.1
0
0
0
0
0
0
0
0
0
1
0
124
Table 4.4. Morphological attributes for ceramic vessel rim/neck fragments recovered from Test Unit 1, Tatuba Cave.
Vessel Vessel
#
Form
Rim
Depth Portion Rim Rim-body Rim Rim Lip Lip Rim Body Orient.
(cmbs)
Orient. Cont.
Course Prof. Shape Dec. Dec. Dec. Angle
Rim Rim Neck Orifice Temp.
Thick. Height Incl. Diam. Class
259
260
261
262
263
264
265
266
267
268
269
270
271
272
273
274
275
276
277
278
279
280
281
282
283
0-10
0-10
0-10
0-10
0-10
0-10
0-10
0-10
0-10
0-10
0-10
0-10
0-10
0-10
0-10
0-10
0-10
0-10
0-10
0-10
0-10
0-10
0-10
0-10
0-10
0
0
0
0
0.72
0.51
0
0.61
0
0
0
0.72
0
0.78
0
0
0.51
0.52
0.95
0.8
0
0
0.35
0
0.43
jar
jar
jar
jar
jar
bowl
jar
bowl
jar
jar
jar
jar
jar
bowl
jar
jar
bowl
jar
bowl
jar
jar
jar
bowl
jar
jar
neck
neck
neck
neck
rim
rim
neck
rim
neck
neck
neck
rim
neck
rim
neck
neck
rim
rim
rim
rim
neck
neck
rim
neck
rim
17
17
17
0
17
17
17
17
17
17
17
17
17
17
17
17
17
17
17
17
17
17
17
17
17
24
24
24
0
24
23
24
23
24
24
24
24
24
23
24
24
23
24
23
24
24
24
23
24
24
0
0
0
0
27
26
0
26
0
0
0
28
0
26
0
0
26
0
26
28
0
0
26
0
26
0
0
0
0
35
35
0
35
0
0
0
30
0
30
0
0
31
0
30
30
0
0
35
0
30
0
0
0
0
40
39
0
50
0
0
0
47
0
40
0
0
39
0
39
47
0
0
47
0
40
125
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
0
0
0
33
0
0
0
0
0
0
0
0
0
0
0
0
0
0
60
0
0
0
0
0
0
0
0
0
2.12
0
0
0
0
0
0
0
0
0
0
0
0
0
0
2.6
0
0
0
0
0
0
0
0
0
115
0
0
0
0
0
0
0
0
0
0
0
0
0
0
120
0
0
0
0
0
0
0
0
0
20
32
0
0
0
0
0
0
0
0
0
0
0
0
32
14
0
0
0
0
0
1
1
1
1
3
1
1
3
1
1
1
1
1
1
1
1
1
1
1
1
1
1
1
1
1
Vessel
#
284
285
286
287
288
289
290
291
292
293
294
295
296
297
298
299
300
301
302
303
304
305
306
307
308
309
310
311
Vessel
Form
jar
jar
jar
jar
jar
jar
unknown
unknown
unknown
unknown
unknown
unknown
unknown
unknown
unknown
unknown
unknown
unknown
unknown
unknown
unknown
unknown
unknown
unknown
unknown
unknown
unknown
unknown
Depth
(cmbs)
0-10
0-10
0-10
0-10
0-10
0-10
0-10
0-10
0-10
0-10
0-10
0-10
0-10
0-10
0-10
0-10
0-10
0-10
0-10
0-10
0-10
0-10
0-10
0-10
0-10
0-10
0-10
0-10
Portion Rim Rim-body Rim Rim
Orient. Cont.
Course Prof.
neck
17
24
0
0
neck
17
24
0
0
neck
17
24
0
0
neck
17
24
0
0
neck
17
24
0
0
neck
17
24
0
0
rim
0
0
0
0
neck
0
0
0
0
body
0
0
0
0
body
0
0
0
0
body
0
0
0
0
body
0
0
0
0
body
0
0
0
0
body
0
0
0
0
body
0
0
0
0
body
0
0
0
0
body
0
0
0
0
body
0
0
0
0
body
0
0
0
0
body
0
0
0
0
body
0
0
0
0
body
0
0
0
0
body
0
0
0
0
body
0
0
0
0
body
0
0
0
0
body
0
0
0
0
body
0
0
0
0
body
0
0
0
0
Lip
Shape
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
126
Lip
Dec.
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
Rim
Dec.
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
Body
Dec.
0
0
0
0
0
0
0
66
90
90
90
64
66
106
65
90
90
90
90
66
65
90
65
90
90
90
64
90
Rim
Orient.
Angle
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
Rim
Thick.
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
Rim
Height
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
Neck
Incl.
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
Orifice
Diam.
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
Temp.
Class
2
1
1
1
1
1
1
1
1
1
1
1
1
1
1
1
1
1
1
1
1
1
1
1
1
1
1
1
Vessel
#
312
313
314
315
316
317
318
319
320
322
323
324
325
326
327
328
329
330
331
332
333
334
335
336
337
338
339
340
Vessel
Form
jar
jar
jar
jar
jar
jar
jar
jar
jar
jar
jar
jar
jar
jar
jar
jar
jar
jar
jar
jar
jar
jar
jar
jar
jar
jar
jar
jar
Depth
(cmbs)
10-20
10-20
10-20
10-20
10-20
10-20
10-20
10-20
10-20
10-20
10-20
10-20
10-20
10-20
10-20
10-20
10-20
10-20
10-20
10-20
10-20
10-20
40-50
10-20
10-20
10-20
10-20
10-20
Portion Rim Rim-body Rim Rim
Orient. Cont.
Course Prof.
rim
17
24
26
30
rim/neck 17
24
28
35
rim/neck 17
24
0
0
rim/neck 17
24
27
32
neck
17
24
0
0
rim
17
24
0
30
rim
17
24
27
31
neck
17
24
0
0
neck
17
24
0
0
neck
17
24
0
0
rim/neck 17
24
0
0
rim/neck 17
24
0
0
neck
17
24
0
0
rim
17
24
27
31
rim
17
24
0
0
neck
17
24
0
0
neck
17
24
0
0
neck
17
24
0
0
rim/neck 17
24
27
30
neck
17
24
0
0
neck
17
24
0
0
neck
17
24
0
0
body
17
24
26
35
neck
17
24
0
0
neck
17
24
0
0
neck
17
24
0
0
neck
17
24
0
0
neck
17
24
0
0
Lip
Shape
40
47
0
39
0
40
40
0
0
0
0
0
0
47
0
0
0
0
40
0
0
0
39
0
0
0
0
0
127
Lip
Dec.
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
Rim
Dec.
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
Body
Dec.
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
Rim
Orient.
Angle
0
65
0
29
0
70
28
0
0
0
0
0
0
40
0
0
0
0
41
0
0
0
30
0
0
0
0
0
Rim
Thick.
0.53
0.8
0
1.19
0
0.6
0.72
0
0
0
0
0
0
0.75
0
0
0
0
0.67
0
0
0
0.51
0
0
0
0
0
Rim
Height
0
3
0
4.7
0
0
2.1
0
0
0
0
0
0
3.5
0
0
0
0
3.37
0
0
0
3.31
0
0
0
0
0
Neck
Incl.
0
110
0
120
0
0
127
0
0
0
0
0
0
104
0
0
0
0
105
0
0
0
100
0
0
0
0
0
Orifice
Diam.
0
28
0
20
0
40
30
0
0
0
0
0
0
20
0
0
0
0
20
0
0
0
36
0
0
0
0
0
Temp.
Class
1
2
1
2
1
1
1
2
4
1
1
1
1
2
1
1
2
1
2
1
2
2
1
1
1
2
4
1
Vessel
#
341
342
343
344
345
346
347
348
349
350
351
352
353
354
355
356
357
358
359
360
361
362
363
364
365
366
367
368
Vessel
Form
jar
jar
jar
jar
jar
jar
jar
jar
jar
jar
jar
jar
unknown
unknown
unknown
jar
unknown
unknown
unknown
unknown
unknown
unknown
unknown
unknown
unknown
unknown
unknown
unknown
Depth
(cmbs)
10-20
10-20
10-20
10-20
10-20
10-20
10-20
10-20
10-20
10-20
10-20
10-20
10-20
10-20
10-20
10-20
10-20
10-20
10-20
10-20
10-20
10-20
10-20
10-20
10-20
10-20
10-20
10-20
Portion Rim Rim-body Rim Rim
Orient. Cont.
Course Prof.
neck
17
24
0
0
rim
17
24
27
30
rim
17
24
26
30
neck
17
24
0
0
neck
17
24
0
0
rim
17
24
27
30
neck
17
24
0
0
neck
17
24
0
0
neck
17
24
0
0
neck
17
24
0
0
neck
17
24
0
0
neck
17
24
0
0
body
0
0
0
0
body
0
0
0
0
body
0
0
0
0
neck
17
24
0
0
body
0
0
0
0
body
0
0
0
0
body
0
0
0
0
body
0
0
0
0
body
0
0
0
0
body
0
0
0
0
body
0
0
0
0
body
0
0
0
0
body
0
0
0
0
body
0
0
0
0
body
0
0
0
0
body
0
0
0
0
Lip
Shape
0
39
39
0
0
39
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
128
Lip
Dec.
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
Rim
Dec.
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
Body
Dec.
0
0
0
0
0
0
0
0
0
0
0
0
66
66
65
0
90
107
90
90
64
66
65
109
90
64
66
65
Rim
Orient.
Angle
0
51
0
0
0
30
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
Rim
Thick.
0
0.77
0.47
0
0
0.79
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
Rim
Height
0
1.77
0
0
0
2.51
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
Neck
Incl.
0
82
0
0
0
128
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
Orifice
Diam.
0
18
0
0
0
20
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
Temp.
Class
1
1
1
1
2
4
1
1
1
1
1
1
1
1
1
1
1
1
1
1
1
1
1
1
1
2
1
1
Vessel
#
369
370
371
372
373
374
375
376
377
378
379
380
381
382
383
384
385
386
387
388
389
390
391
392
393
394
395
396
Vessel
Form
unknown
unknown
unknown
unknown
unknown
unknown
unknown
unknown
unknown
unknown
jar
unknown
unknown
unknown
unknown
unknown
unknown
unknown
unknown
unknown
unknown
unknown
unknown
unknown
unknown
unknown
unknown
unknown
Depth
(cmbs)
10-20
10-20
10-20
10-20
10-20
10-20
10-20
20-30
20-30
20-30
20-30
20-30
20-30
20-30
20-30
20-30
20-30
20-30
20-30
20-30
20-30
20-30
20-30
20-30
20-30
20-30
20-30
20-30
Portion Rim Rim-body Rim Rim
Orient. Cont.
Course Prof.
body
0
0
0
0
body
0
0
0
0
body
0
0
0
0
body
0
0
0
0
body
0
0
0
0
body
0
0
0
0
body
0
0
0
0
body
0
0
0
0
body
0
0
0
0
body
0
0
0
0
rim/neck 17
24
26
30
body
0
0
0
0
body
0
0
0
0
body
0
0
0
0
body
0
0
0
0
body
0
0
0
0
body
0
0
0
0
body
0
0
0
0
body
0
0
0
0
body
0
0
0
0
body
0
0
0
0
body
0
0
0
0
body
0
0
0
0
body
0
0
0
0
body
0
0
0
0
body
0
0
0
0
body
0
0
0
0
body
0
0
0
0
Lip
Shape
0
0
0
0
0
0
0
0
0
0
39
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
129
Lip
Dec.
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
Rim
Dec.
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
Body
Dec.
64
90
66
65
66
64
64
65
64
65
64
90
64
64
64
64
64
65
65
90
64
64
65
65
64
65
100
65
Rim
Orient.
Angle
0
0
0
0
0
0
0
0
0
0
2
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
Rim
Thick.
0
0
0
0
0
0
0
0
0
0
0.4
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
Rim
Height
0
0
0
0
0
0
0
0
0
0
1.81
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
Neck
Incl.
0
0
0
0
0
0
0
0
0
0
112
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
Orifice
Diam.
0
0
0
0
0
0
0
0
0
0
10
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
Temp.
Class
1
1
2
1
2
1
1
3
1
1
1
1
3
1
1
1
1
1
4
3
1
2
2
1
1
2
2
1
Vessel
#
397
398
399
400
401
402
403
404
405
406
407
408
409
410
411
412
413
414
415
416
417
418
419
420
421
422
423
424
Vessel
Form
unknown
unknown
unknown
jar
jar
jar
jar
jar
jar
jar
jar
jar
jar
jar
jar
jar
jar
jar
jar
bowl
unknown
unknown
unknown
unknown
unknown
unknown
unknown
unknown
Depth
(cmbs)
20-30
20-30
20-30
20-30
20-30
20-30
20-30
20-30
20-30
20-30
20-30
20-30
20-30
20-30
20-30
20-30
20-30
20-30
20-30
020-30
20-30
30-40
30-40
30-40
30-40
30-40
30-40
30-40
Portion Rim Rim-body Rim Rim
Orient. Cont.
Course Prof.
body
0
0
0
0
body
0
0
0
0
body
0
0
0
0
rim
17
24
26
30
rim/neck 17
24
29
35
neck
17
24
0
0
neck
17
24
0
0
rim
17
24
27
30
rim/neck 17
24
27
30
neck
17
24
0
0
neck
17
24
0
0
neck
17
24
0
0
neck
17
24
0
0
neck
17
24
0
0
rim
17
24
27
30
neck
17
24
0
0
neck
17
24
0
0
neck
17
24
0
0
neck
17
24
0
0
rim
17
24
26
30
body
0
0
0
0
body
0
0
0
0
body
0
0
0
0
body
0
0
0
0
body
0
0
0
0
body
0
0
0
0
body
0
0
0
0
body
0
0
0
0
Lip
Shape
0
0
0
40
47
0
0
50
39
0
0
0
0
0
50
0
0
0
0
39
0
0
0
0
0
0
0
0
130
Lip
Dec.
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
Rim
Dec.
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
Body
Dec.
65
90
0
0
0
0
0
0
0
0
0
0
0
0
0
0
90
0
0
0
0
66
106
66
65
90
65
108
Rim
Orient.
Angle
0
0
0
27
31
0
0
70
30
0
0
0
0
0
37
0
0
0
0
0
0
0
0
0
0
0
0
0
Rim
Thick.
0
0
0.35
0.68
0.63
0
0
0.67
0.49
0
0
0
0
0
0.7
0
0
0
0
0.5
0
0
0
0
0
0
0
0
Rim
Height
0
0
0
2.37
6.01
0
0
1.69
3.1
0
0
0
0
0
1.8
0
0
0
0
0
0
0
0
0
0
0
0
0
Neck
Incl.
0
0
0
100
107
0
0
109
112
0
0
0
0
0
108
0
0
0
0
0
0
0
0
0
0
0
0
0
Orifice
Diam.
0
0
0
24
24
0
0
18
18
0
0
0
0
0
18
0
0
0
0
0
0
0
0
0
0
0
0
0
Temp.
Class
1
3
7
2
1
1
1
2
1
1
1
1
1
1
1
1
1
1
2
1
2
3
3
1
1
1
1
1
Vessel
#
425
426
427
428
429
430
431
432
433
434
435
436
437
438
439
440
441
442
443
444
445
446
447
448
449
450
451
452
Vessel
Form
unknown
unknown
unknown
unknown
jar
jar
jar
unknown
bowl
jar
jar
unknown
jar
jar
jar
jar
unknown
unknown
unknown
unknown
unknown
unknown
unknown
jar
unknown
unknown
unknown
unknown
Depth
(cmbs)
30-40
30-40
30-40
30-40
30-40
30-40
30-40
30-40
030-40
30-40
30-40
30-40
40-50
40-50
40-50
40-50
40-50
40-50
40-50
50-60
50-60
50-60
50-60
60-70
60-70
60-70
60-70
60-70
Portion Rim Rim-body Rim Rim
Orient. Cont.
Course Prof.
body
0
0
0
0
body
0
0
0
0
body
0
0
0
0
body
0
0
0
0
neck
17
24
0
0
neck
17
24
0
0
neck
17
24
0
0
body
0
0
0
0
rim
17
23
26
30
neck
17
24
0
0
neck
17
24
0
0
body
0
0
0
0
neck
17
24
27
0
neck
17
24
0
0
neck
17
24
0
0
neck
17
24
0
0
body
0
0
0
0
body
0
0
0
0
body
0
0
0
0
body
0
0
0
0
body
0
0
0
0
body
0
0
0
0
body
0
0
0
0
neck
17
24
0
0
body
0
0
0
0
body
0
0
0
0
body
0
0
0
0
body
0
0
0
0
Lip
Shape
0
0
0
0
0
0
0
0
39
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
131
Lip
Dec.
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
Rim
Dec.
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
Body
Dec.
65
0
64
65
0
0
0
0
0
0
0
107
0
108
0
0
90
65
65
64
64
107
0
0
64
64
64
64
Rim
Orient.
Angle
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
Rim
Thick.
0
0
0
0
0
0
0
0
0.58
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
Rim
Height
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
Neck
Incl.
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
Orifice
Diam.
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
Temp.
Class
1
1
2
1
3
2
1
1
1
1
1
1
1
1
1
1
2
2
2
1
1
1
1
1
1
1
1
1
Vessel
#
453
454
455
456
457
458
459
460
461
462
463
464
465
466
467
468
469
470
471
472
473
474
475
476
477
478
479
480
Vessel
Form
unknown
unknown
unknown
jar
unknown
unknown
unknown
unknown
unknown
unknown
unknown
unknown
bowl
unknown
unknown
unknown
unknown
unknown
unknown
jar
jar
unknown
bowl
unknown
unknown
unknown
unknown
unknown
Depth
(cmbs)
60-70
60-70
60-70
70-80
70-80
70-80
70-80
70-80
80-90
80-90
80-90
80-90
090-100
90-100
90-100
90-100
90-100
100-110
100-110
100-110
110-120
110-120
110-120
110-120
110-120
110-120
110-120
110-120
Portion Rim Rim-body Rim Rim
Orient. Cont.
Course Prof.
body
0
0
0
0
body
0
0
0
0
body
0
0
0
0
rim
17
24
27
30
body
0
0
0
0
body
0
0
0
0
body
0
0
0
0
body
0
0
0
0
body
0
0
0
0
body
0
0
0
0
body
0
0
0
0
body
0
0
0
0
rim
17
23
26
31
body
0
0
0
0
body
0
0
0
0
body
0
0
0
0
body
0
0
0
0
body
0
0
0
0
body
0
0
0
0
neck
17
24
0
0
rim
17
24
26
30
rim
17
0
26
26
rim
17
23
26
30
rim
17
24
0
0
body
0
0
0
0
body
0
0
0
0
body
0
0
0
0
neck
0
0
0
0
Lip
Shape
0
0
0
39
0
0
0
0
0
0
0
0
39
0
0
0
0
0
0
0
39
39
39
50
0
0
0
0
132
Lip
Dec.
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
Rim
Dec.
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
Body
Dec.
90
64
64
0
90
65
65
64
65
64
64
65
0
64
64
0
64
65
107
0
0
0
0
0
65
65
65
0
Rim
Orient.
Angle
0
0
0
27
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
24
0
0
0
0
0
0
0
Rim
Thick.
0
0
0
0.39
0
0
0
0
0
0
0
0
0.61
0
0
0
0
0
0
0
0.55
0.5
0.85
0.62
0
0
0
0
Rim
Height
0
0
0
3.15
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
3.1
0
0
0
0
0
0
0
Neck
Incl.
0
0
0
124
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
118
0
0
0
0
0
0
0
Orifice
Diam.
0
0
0
24
0
0
0
0
0
0
0
0
20
0
0
0
0
0
0
0
22
0
20
0
0
0
0
0
Temp.
Class
1
1
1
1
1
7
7
1
1
1
1
2
1
1
1
1
1
2
1
1
1
1
2
1
1
1
2
1
Vessel
#
481
482
483
484
485
486
487
488
489
490
491
492
493
494
495
496
497
498
Vessel
Form
unknown
unknown
unknown
unknown
jar
unknown
unknown
unknown
jar
jar
bowl
unknown
unknown
jar
jar
unknown
unknown
unknown
Depth
(cmbs)
110-120
110-120
110-120
110-120
110-120
110-120
110-120
110-120
120-130
130-140
130-140
130-140
130-140
140-150
140-150
140-150
140-150
150-160
Portion Rim Rim-body Rim Rim
Orient. Cont.
Course Prof.
neck
0
0
0
0
body
0
0
0
0
body
0
0
0
0
body
0
0
0
0
neck
17
24
0
0
body
0
0
0
0
body
0
0
0
0
body
0
0
0
0
rim
17
24
27
30
neck
17
24
0
0
rim
17
23
26
30
body
0
0
0
0
body
0
0
0
0
neck
17
24
0
0
rim
17
24
0
0
body
0
0
0
0
body
0
0
0
0
body
0
0
0
0
Lip
Shape
0
0
0
0
0
0
0
0
40
0
39
0
0
0
50
0
0
0
133
Lip
Dec.
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
Rim
Dec.
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
Body
Dec.
0
64
65
65
0
66
64
65
0
0
0
65
65
0
0
64
64
64
Rim
Orient.
Angle
0
0
0
0
0
0
0
0
42
0
0
0
0
0
0
0
0
0
Rim
Thick.
0
0
0
0
0
0
0
0
0.42
0
0.82
0
0
0
0
0
0
0
Rim
Height
0
0
0
0
0
0
0
0
3.15
0
0
0
0
0
0
0
0
0
Neck
Incl.
0
0
0
0
0
0
0
0
97
0
0
0
0
0
0
0
0
0
Orifice
Diam.
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
Temp.
Class
2
1
1
1
1
1
1
2
7
1
2
1
1
1
1
1
1
1
Notes
1
The ceramic rims illustrated by Palmer were not fully analyzed with a morphological
classification by this author due to the fact that the actual artifacts were unavailable for
examination in 2002.
135
References
Barrau, J.
1961 Subsistence Agriculture in Polynesia and Micronesia. Bernice P. Bishop
Museum Publication 219. Bishop Museum Press, Honolulu.
Best, S.
1984 Lakeba: The Prehistory of a Fijian Island. Unpublished PhD thesis,
University of Auckland.
Birks, L.
1973 Archaeological Excavations at Sigatoka Dune Site, Fiji. Bulletin of the Fiji
Museum No. 1. The Fiji Museum, Suva.
Brewster, A. B.
1922 The Hill Tribes of Fiji. Seeley Service and Co., London.
Brookfield, H. C.
1979 Lakeba: Environmental Change, Population Dynamics and Resource Use.
The UNESCO/UNFPA. Population and Environment Project in the Eastern
Islands of Fiji, Island Reports No. 5. Australian National University Press,
Canberra.
Capell, A.
1991
The Fijian Dictionary. Government Printer, Suva, Fiji.
Caviedes, C. N.
2001 El Niño in History: Storming Through the Ages. University Press of Florida,
Gainsville.
Clark, G. R.
1999b Post-Lapita Fiji: Cultural Transformation in the Mid-Sequence.
Unpublished PhD thesis, Australian National University, Canberra.
Clunie, F.
1977 Fijian Weapons and Warfare. Bulletin of the Fiji Museum, No. 2. The Fiji
Museum, Suva.
1986 Letters from the Highlands of Vitilevu, 1877 (John Archibald Boyd).
Domodomo 1986 IV:20-46.
Crosby, A.
1988 Beqa: Archaeology, Structure, and History of Fiji.
thesis,University of Auckland, Auckland.
136
Unpublished MA
Field, J. S.
1998 Natural and Constructed Defenses in Fijian Fortifications.
Asian
Perspectives 37: 32-58.
2002 GIS-based Analyses of Agricultural Production and Habitation in the
Sigatoka Valley, Fiji. In Pacific Landscapes: Archaeological Approaches, edited
by T. Ladefoged and M.W. Graves, pp. 97-124. Easter Island Foundation Press,
Los Osos.
2003
The Evolution of Competition and Cooperation in Fijian Prehistory:
Archaeological Research in the Sigatoka Valley, Fiji. Unpublished PhD
dissertation. University of Hawai‘i, Manoa.
Fijian Dept. of Forestry
1996 A Guide to Some Indigenous Fijian Trees. Department of Forestry,
Suva, Fiji.
Fowler, H. W.
1955 Archaeological Fishbones Collected by E. W. Gifford in Fiji. Bulletin
of the Bernice P. Bishop Museum, No. 214. BPBM, Honolulu.
Frost, E. L.
1974 Archaeological Excavations of Fortified Sites on Taveuni, Fiji. Asian and
Pacific Archaeology Series No. 6, University of Hawai‘i Press, Honolulu.
Gifford, E. W.
1952 Tribes of Viti Levu and their Origin Places. Anthropological Records
13(5):337-341. University of California Press, Berkeley and Lost Angeles.
Gilbert, T.
1984 Limestone and Volcanic Caves of the Fiji Islands. Transactions of the
British Cave Research Association 11(2):105-118.
Gordon, A.
1879 Letters and Notes Written During the Disturbances in the Highlands
(known as the ‘devil country’) of Viti Levu, Fiji, 1876. R. and R. Clark,
Edinburgh.
Green, R.
1967 Fortifications in Other Parts of Tropical Polynesia. New Zealand
Archaeological Newsletter, 10:96-113.
Haynes, A.
2001 Freshwater Snails of the Tropical Pacific Islands. Institute of Applied
Sciences, University of the South Pacific, Suva.
Hunt, T. L.
137
1980 Toward Fiji's Past: Archaeological Research on Southwestern Viti Levu.
Unpublished MA thesis, University of Auckland, New Zealand.
Joukowsky, M.
1980 A Complete Manual of Field Archaeology. Tools and Techniques of Field
Work for Archaeologists. Prentice Hall, New Jersey.
Kaloumaira, A.
2000 Reducing the Impact of Environmental Emergencies Through Early Warning
and Preparedness- The case of El Niño-Southern Oscillation (ENSO). South
Pacific Applied Geosciences Commission < http://www.esig.ucar.edu/un/fiji.html
(June 16th, 2003).
Kirch, P. V.
1994 The Wet and the Dry: Irrigation and Agricultural Intensification in
Polynesia. The University of Chicago Press, Chicago.
McCormac, F. G. A. G. Hogg, T. F. G. Higham, J. Lynch-Steiglitz, W. S. Broeker, M.
G. L. Baille, J. G. Palmer, L. Xiong, J. R. Pilcher, D. Brown, and S. T. Hooper.
1998 Temporal Variation in the Interhemisphere 14C Offset. Geophysical
Research Letters 25(9):1321-4.
Nichols, N. and K. K. Wong
1990 Dependence of Rainfall Variability on Mean Rainfall, Latitude, and the
Southern Oscillation. Journal of Climate 3:163-170.
Nunn, P. D.
2000a Environmental Catastrophe in the Pacific Islands about AD 1300.
Geoarchaeology 15: 715-740.
2000b Illuminating Sea-Level Fall around AD 1220-1510 (730-440 cal yr BP) in
the Pacific Islands: Implications for Environmental Change and Cultural
Transformation. New Zealand Geographer 56:46-54.
2003 Revising Ideas About Environmental Determinism: Human-Environment
Relations in the Pacific Islands. Asia Pacific Viewpoint 44, 63-72.
Nunn, P., J. Britton
2001 Human-Environmental Relationships in the Pacific Islands around A.D.
1300. Environment and History 7, 3-22.
Nunn, P., J. Stone, and R. Kumar
2003 Corroborating the ‘AD-1300 sea-level fall’ in the South Pacific: New Dates
from Niue and Fiji. South Pacific Journal of Natural Science 20: in press.
Palmer, B.
1966 Site Survey- Sigatoka Valley, 1965-66. Preliminary Report I, Sigatoka
138
Research Project. Unpublished ms., Fiji Museum Archives.
1969a Ring-ditch Fortifications on Windward Viti Levu, Fiji. Archaeology and
Physical Anthropology in Oceania 4, 182-197.
1969b Fortified Sites on Ridge Junctions, Fiji. New Zealand Archaeological
Association Newsletter 12, 15-19.
1971 Naga Ceremonial Sites in Navosa Upper Sigatoka Valley. Final Report
No. 1. Records of the Fiji Museum 1(5):92-106.
1972 Tatuba Cave Excavations, 1972. Unpublished field notes, Fiji Museum
Archives.
Palmer, B., F. Clunie, K. Moce
1973 Excavations at Tatuba Cave. Unpublished ms., Fiji Museum Archives
Palmer, B., and E. Shaw
1968 Pottery-Making at Nakoro, Fiji. Records of the Fiji Museum I(4):80-90.
Parkinson, B. J.
1982 The Specimen Shell Resources of Fiji. South Pacific Commission, Noumea.
Parry, J.
1987 The Sigatoka Valley- Pathways into Prehistory. Fiji Museum Bulletin No.
9.The Fiji Museum, Suva.
1997 The North Coast of Viti Levu Bā to Rā: Air Photo Archaeology and
Ethnohistory. Bulletin of the Fiji Museum, No. 10. The Fiji Museum, Suva.
Salinger, M. J., R. Basher, B. Fitzharris, J. Hay, P. Jones, J. McVeigh, and
Schmidely-Leleu.
1995 Climate Trends in the South-West Pacific. International Journal of
Climatology 15: 285-302.
I.
Sandweiss, D. H.
2002 The Prehistory of El Nino: An Archaeological Perspective. Presented at the
Smithsonian Tropical Research Institute Annual meeting for the Association for
Tropical Biology, Panama City, Panama.
Severns, M.
2000 Hawaiian Seashells. Island Heritage Publishing, Aiea.
Spriggs, M.
1982 Taro Cropping Systems in the Southeast Asian-Pacific Region.
Archaeology in Oceania 17: 7-15.
1984 Taro Irrigation Techniques in the Pacific. In Edible Aroids, edited by S.
Chandra, pp. 123-55. Clarendon Press, Oxford.
139
Stuiver M., P. Reimer, E. Bard, J. Beck, G. Burr, K. Hughen, B. Kromer, G.
McCormac, J. van der Plicht, and M. Spurk
1998 INTCAL98 Radiocarbon Age Calibration, 24000-0 Cal BP. Radiocarbon
40: 1041-1083.
Tindall, H. D.
1983 Vegetables in the Tropics. The Macmillan Press, Ltd., New Hampshire.
Watling, R., J. C. Pernetta
1977 Limestone Caves in the Sigatoka Valley, Viti Levu, Fiji. Studies in
Speleology 3(2):78-86.
140