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