Thesis Title

Home , Austrovenus stutchburyi, Conus araneosus, Conus lividus, Cypraecassis rufa, Drupa morum, Gyrineum pusillum

Archaeomalocological methods, forager decision-making, and intertidal ecosystems: Two millennia of mollusc exploitation on a remote Pacific atoll. Matthew Harris BA (Hons)

A thesis submitted for the degree of Doctor of Philosophy at The University of Queensland in 2017 School of Social Science

Abstract

Marine mollusc shells are excellent proxy records for human behaviour and environmental archives as they are ubiquitous in coastal archaeological deposits and preserve well compared with other marine fauna. Archaeomalacology, the study of molluscs from archaeological sites, has generated new data on the role of coastal environments in the human story, elucidating patterns of forager behaviour, human impacts to the environment, the role of marine foods in coastal palaeoeconomies, and responses to changes in climate and environments both within and outside the Pacific Islands. Molluscs are also critical to the functioning of coral reefs and intertidal ecosystems, and as such, can be useful in tracking long-term trajectories of change in marine environments. This thesis presents the first high-resolution study of the archaeomalacological record of Ebon Atoll, Republic of the Marshall Islands, eastern Micronesia, demonstrating that molluscs had been a stable component of the diet for two millennia.

Atolls, consisting primarily of unconsolidated biogenic sediments atop a narrow reef platform that surrounds a lagoon, have long been considered marginal environments for human habitation. A lack of standing fresh water, poor soils for agriculture, and exposure to storms and extreme weather due to low elevation present considerable challenges for inhabitants in the past as they do today. The relatively small land area is, however, bounded by an expansive reef platform which hosts a rich and diverse range of mollusc species, offering an easily accessible source of protein and other minerals not available in terrestrial foods.

Nevertheless, mollusc remains from Marshall Islands archaeological sites have been assessed only in broad terms as part of synthetic works on settlement and subsistence patterns in the archipelago. This thesis presents a detailed analysis of mollusc remains from a number of archaeological sites on Ebon Atoll, using a newly developed quantification protocol that incorporates a greater number of non-repetitive shell elements, and a new method for tracking forager decision-making in tropical intertidal settings. In addition, a review of the archaeological literature pertaining to human impacts to molluscs during the prehistoric period of the Pacific Islands facilitated investigation of these processes on Ebon Atoll. No discernible human impacts were noted, and mollusc assemblages from Ebon Atoll spanning two millennia of occupation were consistently rich, even, and diverse, incorporating a broad range of taxa from different habitats. Variation in assemblage composition is likely related to the configuration of intertidal habitats on windward and leeward exposed islets, rather than site function or alterations to marine environments. These results indicate that this generalised foraging strategy, low human populations and a productive marine environment i produced sustained yields of molluscs by spreading impact across trophic levels and functional groups. These data contest traditional perceptions of atolls, and are in line with current discourses that challenge traditional notions of small islands, and especially atolls as remote, isolated and marginal settings for human habitation.

Declaration by author

This thesis is composed of my original work, and contains no material previously published or written by another person except where due reference has been made in the text. I have clearly stated the contribution by others to jointly-authored works that I have included in my thesis.

I have clearly stated the contribution of others to my thesis as a whole, including statistical assistance, survey design, data analysis, significant technical procedures, professional editorial advice, and any other original research work used or reported in my thesis. The content of my thesis is the result of work I have carried out since the commencement of my research higher degree candidature and does not include a substantial part of work that has been submitted to qualify for the award of any other degree or diploma in any university or other tertiary institution. I have clearly stated which parts of my thesis, if any, have been submitted to qualify for another award.

I acknowledge that an electronic copy of my thesis must be lodged with the University Library and, subject to the policy and procedures of The University of Queensland, the thesis be made available for research and study in accordance with the Copyright Act 1968 unless a period of embargo has been approved by the Dean of the Graduate School.

I acknowledge that copyright of all material contained in my thesis resides with the copyright holder(s) of that material. Where appropriate I have obtained copyright permission from the copyright holder to reproduce material in this thesis.

iii

Publications during candidature

Peer-reviewed papers Harris, M. and M.I. Weisler. 2017. Two millennia of mollusc foraging on Ebon Atoll, Marshall Islands: sustained marine resource use on a Pacific atoll. Archaeology in Oceania (accepted Jun 2017). Harris, M., P. Faulkner and B. Asmussen. 2017. Macroscopic approaches to the identification of expedient bivalve tools: A case study investigating Polymesoda (=Geloina) coaxans (Bivalvia: Corbiculidae) shell valves from Princess Charlotte Bay, Queensland, Australia. Quaternary International 427, Part A:201-215. Harris, M. and M.I. Weisler. 2016. Prehistoric human impacts to marine molluscs and intertidal ecosystems in the Pacific Islands. Journal of Island and Coastal Archaeology (accepted Dec 2016). Harris, M. and M.I. Weisler. 2016. Intertidal foraging on atolls: prehistoric forager decision making at Ebon Atoll, Marshall Islands. Journal of Island and Coastal Archaeology 12(2):200-223. Harris, M., A.B.J. Lambrides and M.I. Weisler. 2016. Windward vs. leeward: Inter-site variation in marine resource exploitation on Ebon Atoll, Republic of the Marshall Islands. Journal of Archaeological Science: Reports 6:221-229. Harris, M., M.I. Weisler and P. Faulkner. 2015. A refined protocol for calculating MNI in archaeological molluscan shell assemblages: a Marshall Islands case study. Journal of Archaeological Science 57:168-179.

Publications included in this thesis

Several chapters within this thesis are composed of manuscripts accepted for publication in peer- reviewed journals. The research design, analysis of data, interpretation of results, writing, submission and revisions of each manuscript were performed primarily by myself.

Chapter 2: ‘Prehistoric human impacts to marine molluscs and intertidal ecosystems in the Pacific Islands’ has been peer-reviewed and accepted for publication in Journal of Island and Coastal Archaeology.

Contributor Statement of contribution M. Harris (Candidate) Collation and Review of literature (100%)

Wrote and edited the paper (70%) M.I. Weisler Wrote and edited the paper (30%)

Chapter 3: ‘A refined protocol for calculating MNI in archaeological molluscan shell assemblages: A Marshall Islands case study’ has been peer-reviewed and accepted for publication in Journal of Archaeological Science.

Contributor Statement of contribution M. Harris (Candidate) Research conception and design (60%) Analysis and interpretation of data (80%) Wrote and edited the paper (65%) M.I. Weisler Research conception and design (20%) Wrote and edited the paper (17.5%) P. Faulkner Research conception and design (20%) Analysis and interpretation of data (20%) Wrote and edited the paper (17.5%)

Chapter 4: ‘Intertidal foraging on atolls: Prehistoric forager decision making at Ebon Atoll, Marshall Islands.’ has been peer-reviewed and accepted for publication in Journal of Island and Coastal Archaeology.

Contributor Statement of contribution M. Harris (Candidate) Research conception and design (70%) Analysis and interpretation of data (100%) Wrote and edited the paper (80%) M.I. Weisler Research conception and design (30%) Wrote and edited the paper (20%)

Chapter 5: ‘Windward vs. leeward: Inter-site variation in marine resource exploitation on Ebon Atoll, Republic of the Marshall Islands.’ has been peer-reviewed and accepted for publication in Journal of Archaeological Science: Reports.

Contributor Statement of contribution M. Harris (Candidate) Research conception and design (40%) Analysis and interpretation of data (50%) Wrote and edited the paper (50%) A.B.J. Lambrides Research conception and design (40%) Analysis and interpretation of data (50%) Wrote and edited the paper (40%) M.I. Weisler Research conception and design (20%) Wrote and edited the paper (10%)

Chapter 6: ‘Two millennia of mollusc foraging on Ebon Atoll, Marshall Islands: sustained marine resource use on a Pacific atoll.’ has been peer-reviewed and accepted for publication in Archaeology in Oceania.

Contributor Statement of contribution M. Harris (Candidate) Research conception and design (70%) Analysis and interpretation of data (90%) Wrote and edited the paper (70%) M.I. Weisler Research conception and design (30%) Analysis and interpretation of data (10%) Wrote and edited the paper (30%)

Contributions by others to the thesis

Professor Marshall Weisler was my principal supervisor and directs the Marshall Islands archaeological research project that this thesis forms a part of. This project was initiated in 1993, and Weisler formulated the initial research questions and project aims, and provided supervision, direction and feedback throughout the process of analysing the data and producing the peer- reviewed articles presented in this thesis.

Patrick Faulkner was my associate supervisor and provided feedback and guidance during publication of the peer-reviewed manuscripts presented in this thesis. Faulkner collaborated with myself and Weisler on a single manuscript (Chapter 3), and provided guidance on the identifcatoin protocols used in analysis of mollusc remains.

Ariana Lambrides was a member of the Marshall Islands archaeological research project initiated by Weisler. Lambrides collaborated with myself and Weisler on a single manuscript (Chapter 5) providing data from other faunal classes and contributing to the creation of the manuscript.

Statement of parts of the thesis submitted to qualify for the award of another degree

No part of this work was submitted to qualify for the award of another degree.

vii

Acknowledgements

I would like to thank first my academic supervisors, Professor Marshall Weisler and Dr. Patrick Faulkner for their assistance, guidance and mentorship throughout my candidature. This thesis forms part of a larger project in the Marshall Islands initiated by Professor Weisler, and I am grateful to have been able to contribute. I am exceptionally grateful to have been given the opportunity to work with Marshall, who has provided endless feedback and on manuscript drafts, opportunities for work, and helped me to develop my skills as a researcher and an illustrator.

I also thank Josepha Maddison of the Historic Preservation Office, Ministry of Internal Affairs, and Lajan Kabua, former Mayor of Ebon Atoll for their permission to conduct fieldwork on Ebon. I would also like to thank the people of Ebon for welcoming us, and especially to Lyn, Dom, Riem, Josen, Frankie, Nashir, Lister, Jolie, and Jin for their assistance in the field. This thesis was supported by The University of Queensland School of Social Science and an Australian Government Research Training Program Scholarship, the Donald Tugby Archaeological Research Award, and School of Social Science travel grants and awards. Ebon fieldwork was supported by a grant to Marshall Weisler from the Office of the Deputy Vice Chancellor (Research).

My fellow PhD candidates, especially Aleisha Buckler (and Judd, Brad, Sonya, Katie, Mark and Hannah), Dale Simpson, Emma James, Tam Smith, and Xavier Carah for their constant support, friendship, and occasional intellectual conversation. Without these friends around, this whole thing would have been a lot more difficult. Pat Faulkner has acted both as an academic supervisor and, but also as a great friend and mentor. Pat is a truly exceptional person and a world-class researcher, and has been a great support throughout my candidature. I must also thank Dr. Glenys McGowan for her support professionally by providing my opportunities to teach, but also personally for her constant support and encouragement throughout my candidature. Thanks also to Dr. Andrew Sneddon for providing me with opportunities for work and professional development throughout my candidature.

To my family - Lisa, Mike, Ross, Jen, Bonny, Laura, Doug, Chris, Pam, Jo-anne, Jack, Charlie, Billy and Ellen. Mum, Ross, Dad, and Jen, I can never thank you enough for everything you have done for me. To Mum and Dad for fostering my curiosity and for lugging a reef’s worth of shells and rocks collected from beaches around the country when we were travelling in ‘96. Nan, who helped me sneak so many of the shells and rocks back into the car or caravan after I was told that I already had too many, I can never thank you enough, for everything. We all get our good qualities from you. To Laura and Doug, thanks for being patient of all the rescheduled dinners and breakfasts when deadlines were

viii drawing near. I am constantly grateful and humbled to be the brother of someone so strong. To my newest family members, Chris, Pam, Jo-anne, Jack and Charlie. Thank you all for welcoming me into your family, and your home. I promise I’ll make Thursday dinner forever if you let me stick around. To Billy and Ellen, thanks for being incredible friends. Looking forward to lazy weekends in Lennox if you’ll have us. Finally, I would like to thank Ariana. I could never have managed it without you.

Keywords

Atoll archaeology, Pacific Islands, archaeomalacology, shell midden studies, island and coastal archaeology, marine subsistence, Marshall Islands, Micronesia, archaeological methods, zooarchaeology.

Australian and New Zealand Standard Research Classifications (ANZSRC)

ANZSRC code: 210106, Archaeology of New Guinea and Pacific Islands (excl. New Zealand), 50% ANZSRC code: 210102, Archaeological Science, 40% ANZSRC code: 210199, Archaeology not elsewhere classified, 10%

Fields of Research (FoR) Classification

FoR code: 2101, Archaeology, 100%

Table of Contents Abstract ...... i Declaration by author ...... iii Publications during candidature ...... iv Peer-reviewed papers ...... iv Publications included in this thesis ...... iv Contributions by others to the thesis ...... vii Statement of parts of the thesis submitted to qualify for the award of another degree ...... vii Acknowledgements ...... viii Keywords ...... x Australian and New Zealand Standard Research Classifications (ANZSRC) ...... x Fields of Research (FoR) Classification ...... x Chapter 1: Introduction ...... 1 Introduction ...... 1 Rationale ...... 3 Research Questions ...... 4 Thesis structure ...... 6 Chapter Summary ...... 6 References Cited ...... 7 Chapter 2: Prehistoric human impacts to marine molluscs and intertidal ecosystems in the Pacific Islands ...... 14 Abstract ...... 15 Introduction ...... 16 The Effects of Human Predation ...... 17 An Historical Ecology Approach ...... 19 Anthropogenic Extirpation of Mollusks in the Pacific Islands ...... 20 Changes in the Size/Age Structure of Mollusk Populations ...... 22 Trophic Alteration and Impacts to Species Richness, Abundance, and Diversity ...... 25 Non-Anthropogenic Alterations to Mollusk Assemblages ...... 30 Discussion ...... 30 Conclusion ...... 32 Acknowledgements ...... 33 References Cited ...... 34 Figures and tables ...... 48

Chapter 3: A refined protocol for calculating MNI in archaeological molluscan shell assemblages: a Marshall Islands case study ...... 49 Abstract ...... 50 Introduction ...... 51 Current methods of MNI calculation in mollusc shell assemblages ...... 52 Site Description ...... 53 Methods ...... 54 A refined protocol for calculating MNI ...... 54 Gastropod and bivalve shell features ...... 55 Additional NRE ...... 56 Calculating gastropod MNI ...... 56 Calculating bivalve MNI ...... 58 MNI Calculation ...... 60 Testing the influence of quantification protocol ...... 61 Results ...... 62 Total MNI ...... 62 Rank order abundance ...... 62 Species richness and evenness ...... 63 Element survivorship ...... 63 Discussion ...... 64 Conclusion ...... 66 Acknowledgements ...... 67 References Cited ...... 68 Figures and Tables ...... 74 Chapter 4: Intertidal foraging on atolls: prehistoric forager decision making at Ebon Atoll, Marshall Islands ...... 94 Abstract ...... 95 Introduction ...... 96 Previous archaeology in the Marshall Islands ...... 97 Mollusc assemblages from Marshall Islands archaeological sites ...... 98 Environmental Context ...... 99 Moniak islet ...... 100 Ebon islet ...... 101 Sites and Samples ...... 102

xii

Methods of mollusc identification and quantification ...... 102 Methodological framework ...... 103 Reconstructing foraging preferences ...... 103 Hierarchical classification scheme ...... 104 Data collection for molluscan zonation and ecology ...... 105 Molluscan ecology and habitat classification scheme usage ...... 105 Results ...... 106 MLEB-1 ...... 106 MLEB-31 ...... 107 Discussion ...... 108 Conclusion ...... 110 Acknowledgements ...... 110 References Cited ...... 112 Figures and Tables ...... 121 Chapter 5: Windward vs. leeward: Inter-site variation in marine resource exploitation on Ebon Atoll, Republic of the Marshall Islands ...... 132 Abstract ...... 133 Introduction ...... 134 Sites and Samples ...... 135 MLEb-1 ...... 137 MLEb-33 ...... 138 MLEb-31 ...... 138 Methods ...... 139 Identification and quantification protocols ...... 139 Statistical analyses ...... 139 Results ...... 141 Discussion ...... 143 Conclusion ...... 145 Acknowledgements ...... 146 References Cited ...... 147 Figures and Tables ...... 153 Chapter 6: Two millennia of mollusc foraging on Ebon Atoll, Marshall Islands: sustained marine resource use on a Pacific atoll...... 158 Abstract ...... 159

xiii

Introduction ...... 160 Traditional Marshall Islands Economy ...... 161 Sites and Samples ...... 162 Ebon Islet ...... 162 Enekoion Islet ...... 163 Moniak Islet ...... 163 Habitat Mapping Methods and Description of Marine Habitats ...... 164 Methods of Mollusc Analysis ...... 165 Results ...... 167 MLEb-5 ...... 168 MLEb-1 ...... 168 MLEb-33 ...... 170 MLEb-31 ...... 171 Tests for human impact ...... 171 Discussion ...... 172 Temporal and spatial trends in mollusc foraging ...... 172 Conclusion ...... 177 Acknowledgements ...... 178 References Cited ...... 179 Figures and Tables ...... 185 Chapter 7: Conclusion ...... 193 Introduction ...... 193 Overview of thesis results ...... 193 RQ1: Has human foraging for molluscs impacted mollusc populations and intertidal ecosystems in the Pacific Islands during the prehistoric period? ...... 193 RQ2: Does the inclusion of an increased number of non-repetitive shell elements (NREs) in quantification protocols influence measures of relative abundance and taxonomic heterogeneity? ...... 194 RQ3: What methods can be generated to explore and understand forager decision- making and habitat selection in atoll environments? ...... 196 RQ4: What is the influence of spatial (settlement patterns and local habitat) and temporal factors on the richness (number of species present), abundance (number of individuals of each species present), and diversity (richness and relative

xiv

distribution of individuals of each species within a population) of mollusc assemblages on Ebon Atoll? ...... 198 RQ5: Is there any indication that human foraging for molluscs directly impacted mollusc populations or had secondary, indirect impacts on intertidal ecosystems on Ebon Atoll over the two millennia of human occupation? ...... 199 Future research objectives ...... 201 Concluding Remarks ...... 201 References Cited ...... 203 Appendix A ...... 206 Appendix B ...... 208 Appendix C ...... 211 Appendix D ...... 218

Table of Figures Chapter 2: Prehistoric human impacts to marine molluscs and intertidal ecosystems in the Pacific Islands

Figure 1 Map of the Pacific Islands with sites mentioned in text ...... 48

Chapter 3: A refined protocol for calculating MNI in archaeological molluscan shell assemblages: a Marshall Islands case study

Figure 1 a. Stratigraphic section of MLEb-1, TP19 (left, 1 m wide) & TP18 (right, 1 m wide) south profile with a maximum depth of ~115cmbs. Scale is 1 m long.; b. North profile of TP7, with a maximum depth of 40cmbs at site MLEb-33 on Enekoion islet. The dense cultural deposit is ~35 cm thick. Scale is 1 m long. (Both photos, M. Weisler)...... 84

Figure 2 Gastropod terminology...... 85

Figure 3 Examples of gastropod shapes. a. globoid b. involute c. tubular d. trochoid e. turbinate f. patelliform g. disjunct h. turriform. Note the substantial variation in spire height between shell forms...... 86

Figure 4 Bivalve terminology...... 87

Figure 5 Examples of bivalve shapes a. orbicular b. alate c. auriculate d. subquadrate e. trigonal f.fan-shaped g.ensiform h. elongate-elliptical...... 88

Figure 6 Gastropod NRE (1 = spire; 2 = anterior canal; 3 = posterior canal; 4 = outer lip; 5 = aperture; 6 = operculum; 7 = umbilicus). Hatched areas represent areas of shell included in quantification of MNI. Note the presence of NRE on some shell forms, but not others...... 89

Figure 7 The tMNI method of gastropod MNI calculation. Hatched areas represent a fragment of shell. (Sp = spire; OL = outer lip; Ap = aperture; AC = anterior canal/notch; PC = posterior canal/notch; Um = umbilicus; Op = Operculum) ...... 90

Figure 8 Bivalve NRE (1 = umbo; 2 = posterior hinge; 3 = anterior hinge; 4 = posterior adductor muscle scar; 5 = anterior adductor muscle scar). Note the presence of only a single adductor muscle scar on the monomyarian shell valve of b...... 91

Figure 9 The tMNI method of bivalve MNI calculation. Hatched areas represent a shell fragment. (Um = umbo; AH = anterior hinge; PH = posterior hinge; AAMS = anterior adductor muscle scar; PAMS = posterior adductor muscle scar)...... 92

Figure 10 Additional NRE included in tMNI quantification. Hatched areas represent areas of shell included in quantification of MNI. a. view of Cypraeidae spp. aperture and base showing additional NRE; b. view of Neritidae spp. aperture and columellar deck showing location of additional NRE, I = anterior columellar deck/outer lip intersection II = posterior columellar deck/outer lip intersection...... 93

xvi

Chapter 4: Intertidal foraging on atolls: prehistoric forager decision making at Ebon Atoll, Marshall Islands

Figure 1 Map of the Marshall Islands and Ebon Atoll, showing archaeological sites on leeward (Ebon) and windward (Moniak) islets...... 125

Figure 2 Schematic cross section of a. leeward Ebon islet and b. windward Moniak islet highlighting patterns of intertidal zonation on atolls, the molluscan fauna characterisitc of each habitat, the relative exposure of each islet to winds and waves and traditional human settlement patterns (after Kendall et al. 2012; Merlin et al. 1994; Weisler 1999b; Wiens 1962). c. Ebon islet oceanside reef flat (Photo: M. Harris), d. Ebon islet lagoonside (Photo: M. Weisler), e. Moniak islet lagoonside (Photo: M. Weisler) f. Moniak Islet oceanside (Photo: M. Weisler)...... 126

Figure 3 Mollusc species from TP17-20, MLEb-1, represented by 15 or more individuals...... 127

Figure 4 Habitats accounting for more than 20% of MNI MLEb-1, TP17-20; a. all taxa b. Conus spp., Monetaria moneta and Cypraeidae spp. removed. See Table 1 for classification scheme key...... 128

Figure 5 Mollusc species from TP17-20, MLEb-1 assigned to D/1/15 or D/1/16, represented by 15 or more individuals...... 129

Figure 6 Mollusc species from TP2-6, MLEb-31, represented by 15 or more individuals...... 130

Figure 7 Habitats accounting for more than 20% of MNI, TP2-6, MLEb-31 a. all taxa; b. Taxa assigned to D/1/18, represented by 15 or more individuals. See Table 1 for classification scheme key...... 131

Chapter 5: Windward vs. leeward: Inter-site variation in marine resource exploitation on Ebon Atoll, Republic of the Marshall Islands

Figure 1 Map of the Republic of the Marshall Islands, with Ebon Atoll and the location of sites MLEb-1, MLEb-31 and MLEb-33, and photos depicting intertidal marine habitats characteristic of each islet (a) Ebon Islet oceanside, view northwest showing expansive reef flat (Photo: A. Lambrides), (b) Enekoion Islet lagoonside, view northeast showing seagrass beds in the intertidal (Photo: M. Harris), (c) Moniak Islet oceanside, view east of coral cobble and boulder intertidal (Photo: M. Weisler)...... 155

Figure 2 The percent contribution to total MNI and NISP by taxon, site and screen for mollusc shell, 6.4 mm samples and fish bone 6.4 mm and 3.2 mm samples. Family level identifications, but note Selachii (modern sharks), which is a superorder/clade ...... 156

Figure 3 Correspondence analysis of taxonomic abundance. (a) 6.4 mm bivalve shell, (b) 6.4 mm gastropod shell and (c) 6.4 mm fish bone samples are displayed on separate plots for clarity, (d) 3.2 mm fish bone samples. Key taxa are annotated and distinct taxa are not displayed due to minimal contribution to total MNI at each site...... 157

xvii

Chapter 6: Two millennia of mollusc foraging on Ebon Atoll, Marshall Islands: sustained marine resource use on a Pacific atoll.

Figure 1 Map of the Republic of the Marshall Islands, with Ebon Atoll and the location of sites MLEb-1, MLEb-31 and MLEb-33 ...... 186

Figure 2 Representative mollusc taxa from Ebon Atoll archaeological deposits ...... 187

Figure 3 (a) Benthic Habitats mapped within a 2 km radius of MLEb-1 and MLEb-5 on Ebon Islet, MLEb-33 on Enekoion Islet, and MLEb-31 on Moniak Islet with photos depicting characteristic intertidal marine habitats (a) lagoonside, view north west showing seagrass beds north west of MLEb-1 (Photo: M. Weisler) (b) lagoonside, view north east of areas of coral growth adjacent to MLEb-5 (Photo: M. Harris) (c) oceanside, view northwest showing expansive reef flat (Photo: M. Harris) (d) lagoonside, view north east of seagrass beds (Photo: M. Harris) (e) oceanside, view north west showing rubble and boulder reef flat (Photo: M. Harris) (f) lagoonside, view south east showing coarse sands, Ebon Islet in background (Photo: M. Weisler) (g) oceanside, view east of areas of rubble and boulder dominated reef flat (Photo: M. Weisler) ...... 188

Figure 4 Summary of analysis for MLEb-1 TP6 showing a. evenness (E), dominance (1- D) and diversity (H’) b. relative abundance of top three ranked taxa overall c. predominant habitat categories, and d. feeding types...... 189

Figure 5 Summary of analysis for MLEb-1 TP 17-20 showing a. evenness (E), dominance (1-D) and diversity (H’) b. relative abundance of top three ranked taxa overall c. predominant habitat categories, and d. feeding types...... 190

Figure 6 Summary of analysis for all analysed test pits at MLEb-33 showing a. evenness (E), dominance (1-D) and diversity (H’) b. relative abundance of top three ranked taxa overall c. predominant habitat categories, and d. feeding types...... 191

Figure 7 Summary of analysis for MLEb-31 TP 2-6 showing a. evenness (E), dominance (1-D) and diversity (H’) b. relative abundance of top three ranked taxa overall c. predominant habitat categories, and d. feeding types...... 192

xviii

List of Tables Chapter 3: A refined protocol for calculating MNI in archaeological molluscan shell assemblages: a Marshall Islands case study

Table 1 Total MNI for all taxa, aggregated at test pit level ...... 74

Table 2 Total MNI for the top ten ranked taxa only, aggregated at test pit level ...... 75

Table 3 Unique family, genus and species counts for gastropods aggregated by test pit. For each test pit, the counts for all taxa and the top ten ranked taxa are presented...... 76

Table 4 Evenness and dominance measures. D = Simpson’s Dominance, 1-D = Simpson’s Evenness, H’ = Shannon’s index, H’/lnS = Shannon’s evenness ...... 77

Table 5 Eb-1 T18 & 19 NRE MNI and tMNI quantification results for gastropod taxa. * = change in rank order; ** = unique to rank order abundance for that method...... 78

Table 6 Eb-33 TP7 NRE MNI and tMNI quantification results for gastropod taxa. * = change in rank order; ** = unique to rank order abundance for that method...... 79

Table 7 Eb-33 TP7 NRE MNI and tMNI quantification results for bivalve taxa...... 80

Table 8 Eb-1 TP18 and 19 NRE MNI and tMNI quantification results for bivalve taxa...... 81

Table 9 Contribution of the pre-selected bivalve counting character for all taxa...... 82

Table 10 Contribution of the pre-selected gastropod counting character for all taxa...... 83

Chapter 4: Intertidal foraging on atolls: prehistoric forager decision making at Ebon Atoll, Marshall Islands

Table 1 Summary of previous analyses of molluscan assemblages from archaeological sites in the Marshall Islands from Majuro (Riley 1987), Arno (Dye 1987) and Utrōk (Weisler 2001) atolls. Habitat assignments from Baron (1992), Baron and Clavier (1992), Carpenter and Niem (1998), Demond (1957), Soemodihardjo and Matsukuma (1989), Thomas (2001), and Willan (1993)...... 121

Chapter 5: Windward vs. leeward: Inter-site variation in marine resource exploitation on Ebon Atoll, Republic of the Marshall Islands

xix

Table 1 Chord distance values for mollusc shell and fish bone assemblages retained in the 6.4 mm and 3.2 mm sieves for each site pair. 153

Chapter 6: Two millennia of mollusc foraging on Ebon Atoll, Marshall Islands: sustained marine resource use on a Pacific atoll

Table 1 Summary data for mollusc assemblages from MLEb-1, MLEb-5, MLEb-31, and MLEb-33; Wt. = weight ...... 185

Chapter 1: Introduction

Introduction Archaeomalacology, the study of mollusc remains from archaeological sites, has revealed that humans have been foraging for molluscs for at least 165,000 years (Jerardino and Marean 2010). Mollusc shells are excellent proxy records for human behaviour as they are ubiquitous in archaeological deposits and preserve well compared with other marine fauna. Molluscs are also critical components of tropical intertidal ecosystems and coral reefs and, as such, can be useful in tracking long-term trajectories of change in marine environments. Molluscs are effective regulators of algal cover (Hockey and Alison 1986), can filter water to improve quality (Hutchings et al. 2007; Klumpp and Lucas 1994), and can increase habitat complexity through ecosystem engineering or the accumulation of the shells of dead molluscs (Andréfouët et al. 2005; Gutiérrez et al. 2003).

Archaeomalacological analyses have elucidated patterns of forager behaviour (Thomas 1999, 2009), exploitation and management of subsistence resources (Gutiérrez-Zugasti 2011a; Langejans et al. 2012; Milner et al. 2007; Whitaker 2008), human impacts, and the difficulties of disentangling these processes from non-anthropogenic causes (Braje et al. 2007; Erlandson et al. 2008; Faulkner 2013; Giovas 2016; Giovas et al. 2013; Harris and Weisler in press-a and references therein; Jones 2007; Morrison and Hunt 2007; Spennemann 1989; Thakar 2011; Thakar et al. in press). Other researchers have focussed on human responses to environmental or climatic changes (Amesbury 1999, 2007; Parkington et al. 2013) taphonomy (Faulkner 2010, 2011; Gutiérrez-Zugasti 2011b; Sommerville-Ryan 1998), mollusc shells as raw material (Harris et al. 2017; Perrette 2011; Spennemann 1993; Szabó and Koppel 2015) and tracking patterns in long-term human interaction with the marine environment (Fitzpatrick et al. 2011; Szabó and Amesbury 2011). In addition, given that molluscs are important for archaeologists and neo-ecologists alike, studies that combine archaeological data with historical and neo-ecological data have proven utility for understanding how coral reefs have changed over time, and might change in the future (Fitzpatrick and Intoh 2009; O'Dea et al. 2014; Thomas 2001, 2009, 2014; Wiley et al. 2013).

Atolls, consisting primarily of unconsolidated biogenic sediments atop a narrow reef platform surrounding a lagoon, have long been considered marginal environments vulnerable to human impacts and challenging landscapes for sustained habitation. This is due to a lack of standing fresh water, nutrient poor soils for agriculture, and exposure to storms and extreme weather due to low elevation.

The long-term survival of human populations in the atoll archipelago of the Marshall Islands was underpinned by a system of pit-cultivation techniques for giant swamp taro (iraij, Cyrtosperma chamissonis), rain-fed production of arrowroot (makmōk, Tacca leontopetaloides), and arboriculture including pandanus (bōb, Pandanus tectorius), coconut (ni, Cocos nucifera), and breadfruit (mā, Artocarpus altilis) (Weisler 1999a, b) and marine subsistence (Weisler 1999b, 2001b). Much of the reef platform is intertidal, and can be easily accessed for the exploitation of finfish, molluscs and other marine resources that provided protein and other vitamins and minerals not available or limited in terrestrial foods (Erlandson 1988, 2001; Weisler 2001b). The Marshallese language has at least 64 words describing different fishing techniques, and about 45 words for different kinds of mollusc, specific words that divide the intertidal into habitats and zones, and sophisticated seafaring technology and navigation techniques attesting to a deep an intimate knowledge of the ocean (Abo et al. 1976; Merlin et al. 1994). A single language spoken across the entire archipelago, an area of nearly 2,000,000 km2, suggests continuous inter-atoll contact throughout the prehistoric period (Sudo 1984).

For most of the 20th century Pacific atolls were relatively neglected compared to other island types as it was thought that archaeological deposits would not survive in situ (Davidson 1968, 1971). Archaeological survey and limited excavation of atolls in the Marshall Islands beginning in the 1970’s (Dye 1987a) revealed at least two millennia of human occupation. Subsequent work in the archipelago has increased archaeological understanding of the lifeways of people living on atolls during the precontract period (Christensen and Weisler 2013; Harris et al. 2016; Harris and Weisler 2016; Harris and Weisler in press-a, b; Harris et al. 2015; Horrocks and Weisler 2006; Lambrides and Weisler in press; Pregill and Weisler 2007; Sommerville-Ryan 1998; Weisler 1999a, b, 2000, 2001a, b, 2002; Weisler et al. 2000; Weisler and Swindler 2002; Weisler et al. 2012; Weisler et al. in prep.; Yamaguchi et al. 2009). Archaeological analyses have demonstrated that molluscs were a consistent component of subsistence systems in the Marshall Islands from the time of initial colonisation soon after atoll emergence (Kayanne et al. 2011; Weisler 1999b, 2001b). However, these faunae have received little in-depth analysis in the archaeological discourses of the archipelago. Broad syntheses of the place of molluscs in prehistoric period Marshallese economies has been presented by Weisler (2001b), but there have been few dedicated studies of archaeological molluscs from the archipelago prior to the research presented here. This thesis explores quantification methods (Harris et al. 2015), analytical methods (Harris and Weisler 2016), and conduct a spatio-temporal analysis of mollusc foraging during the prehistoric period on Ebon Atoll, Republic of the Marshall Islands (RMI) (Harris et al. 2016; Harris and Weisler in press-b)

Rationale Long-term patterns of mollusc foraging on atolls and the potential for human impacts to coral reefs in the region are not well understood. Recently, archaeologists and neo-ecologists have demonstrated the critical contribution of historical data for comprehensively understanding the state of ecosystems due to the time-lag between disturbance and ecosystem change (Alleway and Connell 2015; Braje et al. 2005; Briggs et al. 2006; Hayashida 2005:45; Rick and Lockwood 2013). Recent studies of finfish remains from sites in the RMI have demonstrated that Marshall Islanders have exploited these resources for at least 2000 years (Weisler 1999a, 2000), potentially with little negative impact (Lambrides and Weisler in press; Weisler 2001b). These studies have also demonstrated some variability in patterns of mollusc exploitation across the windward-leeward exposure gradient on atolls, and variation in foraging patterns between site type (i.e. village v. ephemeral campsites) (Weisler 2001b).

On atolls, the configuration of intertidal environments and related faunal and floral communities are strongly structured by the degree of exposure to wave action (Drumm 2005), and archaeological studies in the Marshall Islands have demonstrated that larger, leeward islets that are sheltered from the wind and waves are more likely to be permanently inhabited than smaller, windward islets (Dye 1987a; Weisler 1999b, 2001b, 2002). Small, often ephemeral windward islets are generally less suitable for permanent settlement due to a smaller, often saline subterranean Ghyben-Herzberg fresh water lens, making them generally unsuitable for agriculture. Variation in site use and the taxonomic composition of mollusc assemblages relating to windward or leeward exposure has been documented for other Pacific Islands (Bayman and Dye 2013; Bedford 2007; Kirch and Dye 1979; Morrison and Hunt 2007; Szabó and Anderson 2012), but has not been systematically investigated for the southern Marshall Islands (but see Weisler 2001b for a discussion of inter-islet variation on Utrō k Atoll, northern RMI). Broad syntheses of mollusc data from Utrō k, Arno, Majuro and Ebon Atolls (Dye 1987b; Riley 1987; Weisler 2001b), has revealed that most assemblages are rich (consisting of many taxa), even (all taxa represented in relatively equal proportions) and diverse (many taxa with many individuals from each taxon). Weisler (2001b:116) documented no indications of human impacts to molluscs on Utrō k, inferred to be the result of low human populations and expansive reef flats. The mollusc assemblages recovered from Ebon Atoll provide an opportunity to perform, for the first time in the RMI, a dedicated, fine-grained analysis of human foraging for molluscs and human-environment interactions using high-resolution archaeomalacological techniques.

It is critical that patterns of human behaviour and an elucidation of foraging practices must be based on reliable, transparent identification and quantification protocols. Comprehensive and clearly reported quantification protocols are critical to data quality, comparability and replicability. The advantages and disadvantages of utilising Number of Individual Specimens (NISP) counts, shell weight, or calculating Minimum Numbers of Individuals (MNI) has been widely debated in the archaeomalacological and zooarchaeological literature (Driver 2011; Gutiérrez-Zugasti 2011b; Jerardino et al. 2016; Popejoy et al. 2016). MNI values are commonly considered the most accurate representation of taxonomic abundance in archaeological contexts, but previous quantification protocols have relied on a very limited range of non-repetitive elements (NRE) to derive MNI values, such as the apex or spire of gastropods and the umbo of bivalves (Claassen 1998, 2000; Glassow 2000; Mason et al. 1998). Similarly, methods for investigating human foraging behaviour have commonly utilised deterministic methods for reconstructing habitat selection based on the life history of mollusc taxa (Allen 1992; Morrison and Addison 2008; Morrison and Hunt 2007; Spennemann 1987; Szabó 2009; Thomas 2002; Weisler et al. 2010), where a single taxon is assigned to a single habitat. This thesis explores an alternate probabilistic method for reconstructing human foraging practices by assigning each taxon to multiple combinations of intertidal zone (reef flat, shoreline, etc.), geomorphological structure (hard v. soft bottom) and benthic cover (sand, rubble, coral, etc.). This method allows exploration of the most likely areas where molluscs were collected, while acknowledging that particular taxa may be collected from multiple areas of the reef. Research questions pertaining to these identified gaps in literature and observations of areas where methods could be improved or altered are presented below.

Research Questions The major goal of this research is to examine temporal and spatial variation in mollusc foraging on Ebon Atoll, and whether prehistoric patterns of resource depression are visible in the archaeological record. Five major questions will be addressed, broadly separated into two themes. First, archaeomalacological methods for quantifying mollusc remains from archaeological sites will be examined, a new method for investigating forager decision-making will be presented, and the current state of knowledge of human impacts to molluscs in the prehistory of the Pacific Islands will be explored. Second, these methods will be applied to archaeological datasets from Ebon Atoll to investigate long-term human interaction with the marine environment, foraging behaviours, and potential human impacts to coral reefs.

Research Question (RQ1) reviews literature pertaining to human impacts to molluscs during the prehistoric period of the Pacific Islands. RQ1 also provides a broad, regional context for the investigation of human impacts to molluscs and intertidal ecosystems on Ebon Atoll. RQ2 examines current methods for quantifying mollusc remains from archaeological sites, and how the protocols influence derived measures of relative abundance and taxonomic composition and heterogeneity. Clear and transparent quantification protocols that do not inherently bias data towards one faunal class over another are critical for developing high-resolution, replicable and reliable understandings of past human behaviour (Driver 2011). RQ3 explores current methods for inferring human foraging behaviour from mollusc remains, and proposes a new method that links archaeomalacological data to a hierarchical classification scheme for benthic habitats in tropical intertidal environments. RQ1, RQ2, and RQ3 are as follows:

RQ1: Has human foraging for molluscs impacted mollusc populations and intertidal ecosystems in the Pacific Islands during the prehistoric period?

RQ2: Does the inclusion of an increased number of non-repetitive shell elements (NREs) in quantification protocols influence measures of relative abundance and taxonomic heterogeneity?

RQ3: What methods can be generated to explore and understand forager decision-making and habitat selection in atoll environments?

Archaeomalacological data has proven utility for enhancing understanding of past human subsistence behaviours, and human-environment interactions in coastal and island archaeology. Using the information gathered to address RQ1, and methods developed in addressing RQ2 and RQ3, RQ4 and RQ5 examine human foraging patterns and potential human impacts to molluscs on Ebon Atoll. RQ4 and RQ5 are as follows:

RQ4: What is the influence of spatial (settlement patterns and local marine habitats) and temporal factors on the richness (number of species present), abundance (number of individuals of each species present), and diversity (richness and relative distribution of individuals of each species within a population) of mollusc assemblages on Ebon Atoll?

RQ5: Is there any indication that human foraging for molluscs directly impacted mollusc populations or had secondary, indirect impacts on intertidal ecosystems on Ebon Atoll over the two millennia of human occupation?

Thesis structure This thesis consists of seven chapters: an Introduction, five published papers (Chapter 2 to Chapter 6), and a Conclusion. Chapter Two reviews the current state of knowledge regarding prehistoric human impacts to molluscs, and potential secondary impacts to intertidal ecosystems in the Pacific Islands resulting from human foraging for molluscs. Chapter Three reviews current methods for calculating MNI for molluscs from archaeological sites, presenting a new method that uses a range of non-repetitive shell elements to calculate MNI, rather than the few traditionally used. Chapter Four presents an analysis of foraging behaviour by tracking the range of habitats likely accessed by prehistoric foragers on Ebon Atoll. Chapter Five compares site-level data for mollusc and fishbone assemblages from a windward and leeward islet on Ebon Atoll to understand the influence of site location and site function on overall Ebon Atoll marine subsistence patterns. Chapter Six presents the results of a comprehensive temporal and spatial analysis of mollusc remains from Ebon Atoll, and investigates the potential for human impacts to molluscs and intertidal ecosystems in the prehistoric period. Chapter Seven summarises the results of each published paper in relation to the research aims and research questions, discusses future research objectives, and provides some concluding remarks.

Chapter Summary This chapter presented the primary aims of this thesis, which are to develop comprehensive methods for quantifying archaeological mollusc shells from Ebon Atoll, Marshall Islands, which are transparent, replicable and reliable. This thesis will also investigate long term patterns of human interaction with molluscs in intertidal environments in the Pacific Islands, and focus intensively on human foraging practices for molluscs on Ebon Atoll, RMI to understand the variation in foraging across the windward-leeward atoll gradient, the influence of site function, and investigate potential human impacts. Mollusc foraging on atolls is a relatively unexplored area of archaeological research, and has the potential to contribute useful data to archaeological discourses on human interaction with the ocean, human colonisation and long-term survival on small islands, and current discourses on the global threats to coral reefs and marine ecosystems.

References Cited Abo, T., B.W. Bender, A. Capelle and T. DeBrum. 1976. Marshallese-English Dictionary. Honolulu: University of Hawai'i Press. Allen, M.S. 1992. Dynamic Landscapes and Human Subsistence: Archaeological Investigations on Aitutaki Islands, Southern Cook Islands. Ph.D. thesis. Washington: University of Washington. Alleway, H.K. and S.D. Connell. 2015. Loss of an ecological baseline through the eradication of oyster reefs from coastal ecosystems and human memory. Conservation Biology 29(3):795- 804. Amesbury, J.R. 1999. Changes in species composition of archaeological marine shell assemblages in Guam. Micronesica 31(2):347-366. ---. 2007. Mollusk collecting and environmental change during the Prehistoric Period in the Mariana Islands. Coral Reefs 26(4):947-958. Andréfouët, S., A. Gilbert, L. Yan, G. Remoissenet, C. Payri and Y. Chancerelle. 2005. The remarkable population size of the endangered clam Tridacna maxima assessed in Fangatau Atoll (Eastern Tuamotu, French Polynesia) using in situ and remote sensing data. ICES Journal of Marine Science: Journal du Conseil 62(6):1037-1048. Bayman, J.M. and T.S. Dye. 2013. Hawaii's Past in a World of Pacific Islands. Washington D.C.: The SAA Press. Bedford, S. 2007. Terra Australis 23: Pieces of the Vanuatu Puzzle: Archaeology of the North, South and Centre. Canberra: ANU E Press. Braje, T.J., J.M. Erlandson, D.J. Kennett, T.C. Rick and J.E. Peterson. 2005. Deep history: Using archaeology and historical ecology to promote marine conservation. Alternate Routes 21:5- 17. Braje, T.J., D.J. Kennett, J.M. Erlandson and B.J. Culleton. 2007. Human impacts on nearshore shellfish taxa: A 7,000 year record from Santa Rosa Island, California. American Antiquity 72(4):735-756. Briggs, J.M., K.A. Spielmann, H. Schaafsma, K.W. Kintigh, M. Kruse, K. Morehouse and K. Schollmeyer. 2006. Why ecology needs archaeologists and archaeology needs ecologists. Frontiers in Ecology and the Environment 4(4):180-188. Christensen, C.C. and M.I. Weisler. 2013. Land snails from archaeological sites in the Marshall Islands, with remarks on prehistoric translocations in tropical Oceania. Pacific Science 67(1):81-104. Claassen, C. 1998. Shells. Cambridge: Cambridge University Press.

---. 2000. Quantifying shell: Comments on Mason, Peterson, and Tiffany. American Antiquity 65(2):415-418. Davidson, J. 1968. Nukuoro: Archaeology on a Polynesian outlier in Micronesia. In Prehistoric Culture in Oceania (I. Yawata and Y.H. Sinoto, eds):51-66. Honolulu: Bishop Museum Press. ---. 1971. Archaeology on Nukuoro Atoll. Bulletin of the Auckland Institute and Museum 9. Driver, J.C. 2011. Identification, classification and zooarchaeology. Ethnobiology Letters 2:19-39. Drumm, D.J. 2005. Habitats and Macroinvertebrate Fauna of the Reef-top of Rarotonga, Cook Islands: Implications for Fisheries and Conservation Management. Ph.D. thesis. Dunedin: University of Otago. Dye, T. 1987a. Marshall Islands Archaeology. Honolulu: Bernice Pauahi Bishop Museum. ---. 1987b. Report 3: Archaeological survey and test excavations on Arno Atoll, Marshall Islands. In Marshall Islands Archaeology (T. Dye, ed.):271-394. Honolulu: Pacific Anthroplogical Records: Bernice Pauahi Bishop Museum. Erlandson, J.M. 1988. The role of shellfish in prehistoric economies: A protein perspective. American Antiquity 53(1):102-109. ---. 2001. The archaeology of aquatic adaptations: Paradigms for a new millennium. Journal of Archaeological Research 9(4):287-350. Erlandson, J.M., T.C. Rick, T.J. Braje, A. Steinberg and R.L. Vellanoweth. 2008. Human impacts on ancient shellfish: a 10,000 year record from San Miguel Island, California. Journal of Archaeological Science 35(8):2144-2152. Faulkner, P. 2010. Morphometric and taphonomic analysis of granular ark (Anadara granosa) dominated shell deposits of Blue Mud Bay, northern Australia. Journal of Archaeological Science 37(8):1942-1952. ---. 2011. Quantifying shell weight loss in archaeological deposits. Archaeology in Oceania 46(3):118-129. ---. 2013. Terra Australis 38: Life on the Margins: An Archaeological Investigation of Late Holocene Economic Variability, Blue Mud Bay, Northern Australia. Canberra ANU E Press Fitzpatrick, S.M., C.M. Giovas and O. Kataoka. 2011. Temporal trends in prehistoric fishing in Palau, Micronesia over the last 1500 years. Archaeology in Oceania 46(1):6-16. Fitzpatrick, S.M. and M. Intoh. 2009. Introduction: Archaeology and historical ecology in the Pacific basin. Pacific Science 63(4):463-464.

Giovas, C.M. 2016. Though she be but little: Resource resilience, Amerindian foraging, and long- term adaptive strategies in the Grenadines, West Indies. The Journal of Island and Coastal Archaeology 11(2):238-263. Giovas, C.M., M. Clark, S.M. Fitzpatrick and J. Stone. 2013. Intensifying collection and size increase of the tessellated nerite snail (Nerita tessellata) at the Coconut Walk site, Nevis, northern Lesser Antilles, AD 890–1440. Journal of Archaeological Science 40(11):4024- 4038. Glassow, M.A. 2000. Weighing vs. counting shellfish remains: A comment on Mason, Peterson, and Tiffany. American Antiquity 65(2):407-414. Gutiérrez-Zugasti, I. 2011a. Coastal resource intensification across the Pleistocene–Holocene transition in Northern Spain: Evidence from shell size and age distributions of marine gastropods. Quaternary International 244(1):54-66. ---. 2011b. Shell fragmentation as a tool for quantification and identification of taphonomic processes in archaeomalacological analysis: The case of the Cantabrian region (Northern Spain). Archaeometry 53(3):614-630. Gutiérrez, J.L., C.G. Jones, D.L. Strayer and O.O. Iribarne. 2003. Mollusks as ecosystem engineers: the role of shell production in aquatic habitats. Oikos 101(1):79-90. Harris, M., P. Faulkner and B. Asmussen. 2017. Macroscopic approaches to the identification of expedient bivalve tools: A case study investigating Polymesoda (=Geloina) coaxans (Bivalvia: Corbiculidae) shell valves from Princess Charlotte Bay, Queensland, Australia. Quaternary International 427, Part A:201-215. Harris, M., A.B.J. Lambrides and M.I. Weisler. 2016. Windward vs. leeward: Inter-site variation in marine resource exploitation on Ebon Atoll, Republic of the Marshall Islands. Journal of Archaeological Science: Reports 6:221-229. Harris, M. and M.I. Weisler. 2016. Intertidal foraging on atolls: Prehistoric forager decision making at Ebon Atoll, Marshall Islands. The Journal of Island and Coastal Archaeology 12(2):200- 223. Harris, M. and M.I. Weisler. in press-a. Prehistoric human impacts to marine molluscs and intertidal ecosystems in the Pacific Islands. The Journal of Island and Coastal Archaeology. ---. in press-b. Two millennia of mollusc foraging on Ebon Atoll, Marshall Islands: Sustained marine resource use on a Pacific atoll. Archaeology in Oceania. Harris, M., M.I. Weisler and P. Faulkner. 2015. A refined protocol for calculating MNI in archaeological molluscan shell assemblages: A Marshall Islands case study. Journal of Archaeological Science 57:168-179.

Hayashida, F.M. 2005. Archaeology, ecological history, and conservation. Annual Review of Anthropology 34:43-65. Hockey, P.A.R. and L.B. Alison. 1986. Man as an intertidal predator in Transkei: Disturbance, community convergence and management of a natural food resource. Oikos 46(1):3-14. Horrocks, M. and M.I. Weisler. 2006. Analysis of plant microfossils in archaeological deposits from two remote archipelagos: The Marshall Islands, eastern Micronesia, and the Pitcairn Group, southeast Polynesia. Pacific Science 60(2):261-280. Hutchings, P., S. Ahyong, M. Byrne, R. Przeslawski, G. Wörheide and Great Barrier Reef Marine Park Authority. 2007. Chapter 11: Vulnerability of benthic invertebrates of the Great Barrier Reef to climate change. In Climate Change and the Great Barrier Reef: A Vulnerability Assessment (J.E. Johnson and P.A. Marshall, eds):309-356. Townsville: The Great Barrier Reef Marine Park Authorty. Jerardino, A., P. Faulkner and C. Flores. 2016. Current methodological issues in archaeomalacological studies. Volume 427(Part A):1-4. Jerardino, A. and C.W. Marean. 2010. Shellfish gathering, marine paleoecology and modern human behavior: Perspectives from cave PP13B, Pinnacle Point, South Africa. Journal of Human Evolution 59(3–4):412-424. Jones, S. 2007. Human impacts on ancient Marine environments of Fiji's Lau Group: Current ethnoarchaeological and archaeological research. The Journal of Island and Coastal Archaeology 2(2):239-244. Kayanne, H., T. Yasukochi, T. Yamaguchi, H. Yamano and M. Yoneda. 2011. Rapid settlement of Majuro Atoll, central Pacific, following its emergence at 2000 years CalBP. Geophysical Research Letters 38(20):L20405. Kirch, P.V. and T.S. Dye. 1979. Ethno-archaeology and the development of Polynesian fishing strategies. The Journal of the Polynesian Society 88(1):53-76. Klumpp, D.W. and J.J. Lucas. 1994. Nutritional ecology of the giant clams, Tridacna tevoroa and T. derasa from Tonga: Influence of light on filter-feeding and photosynthesis. Marine Ecology Progress Series 107:147-156. Lambrides, A.B.J. and M.I. Weisler. in press. Alterations in the late Holocene Marshall Islands archaeological tuna fishery provide proxy evidence for western and central Pacific Ocean ENSO variability. The Journal of Island and Coastal Archaeology. Langejans, G.H.J., K.L. van Niekerk, G.L. Dusseldorp and J.F. Thackeray. 2012. Middle Stone Age shellfish exploitation: Potential indications for mass collecting and resource intensification at Blombos Cave and Klasies River, South Africa. Quaternary International 270:80-94.

Mason, R.D., M.L. Peterson and J.A. Tiffany. 1998. Weighing vs. counting: Measurement reliability and the California school of midden analysis. American Antiquity 63(2):303-324. Merlin, M., A. Capelle, T. Keene, J. Juvik and J. Maragos. 1994. Plants and Environments of the Marshall Islands. Honolulu: East West Center. Milner, N., J. Barrett and J. Welsh. 2007. Marine resource intensification in Viking Age Europe: the molluscan evidence from Quoygrew, Orkney. Journal of Archaeological Science 34(9):1461-1472. Morrison, A.E. and D.J. Addison. 2008. Assessing the role of climate change and human predation on marine resources at the Fatu-ma-Futi site, Tutuila Island, American Samoa: An agent based model. Archaeology in Oceania 42(4):22-34. Morrison, A.E. and T.L. Hunt. 2007. Human Impacts on the nearshore environment: An archaeological case study from Kaua'i, Hawaiian Islands. Pacific Science 61(3):325- 328,331-345. O'Dea, A., M.L. Shaffer, D.R. Doughty, T.A. Wake and F.A. Rodriguez. 2014. Evidence of size- selective evolution in the fighting conch from prehistoric subsistence harvesting. Proceedings of the Royal Society of London B: Biological Sciences 281(1782):1-9. Parkington, J., J.W. Fisher and K. Kyriacou. 2013. Limpet gathering strategies in the later Stone Age along the Cape West Coast, South Africa. The Journal of Island and Coastal Archaeology 8(1):91-107. Perrette, C. 2011. Value and Shell Artefacts in Melanesia: Analysis of the Assemblage of Bourewa (Viti Levu, Fiji), Volume I. Woolongong: Université de Bourgogne (Dijon) and University of Wollongong (Australia) Popejoy, T., S. Wolverton, L. Nagaoka and C.R. Randklev. 2016. An interpretive framework for assessing freshwater mussel taxonomic abundances in zooarchaeological faunas. Quaternary International Volume 427(Part A):36-46. Pregill, G.K. and M.I. Weisler. 2007. Lizards from prehistoric sites on Ebon Atoll, Marshall Islands. Micronesica 39(2):107-115. Rick, T.C. and R. Lockwood. 2013. Integrating paleobiology, archeology, and history to inform biological conservation. Conservation Biology 27(1):45-54. Riley, T.J. 1987. Report 2: Archaeological survey and testing, Majuro Atoll, Marshall Islands. In Marshall Islands Archaeology (T. Dye, ed.):169-270. Honolulu: Pacific Anthroplogical Records: Bernice Pauahi Bishop Museum. Sommerville-Ryan, G. 1998. The Taphonomy of a Marshall Islands' Shell Midden. M.A. thesis. Dunedin: University of Otago.

Spennemann, D.H.R. 1987. Availability of shellfish resources on prehistoric Tongatapu, Tonga: Effects of human predation and changing environment. Archaeology in Oceania 22(3):81- 96. ---. 1989. Effects of human predation and changing environment on some mollusc species on Tongatapu, Tonga. In The Walking Larder: Patterns of Domestication, Pastoralism, and Predation (J. Clutton-Brock, ed.):326-335. Sydney: Unwin Hyman. ---. 1993. Cowrie shell tools: Fact or fiction? Archaeology in Oceania 28(1):40-49. Sudo, K. 1984. Social organization and types of sea tenure in Micronesia. In Maritime Institutions in the Western Pacific (K. Ruddle and T. Akimichi, eds):203-230. Szabó, K. 2009. Molluscan remains from Fiji. In Terra Australis 31: The Early Prehistory of Fiji (G. Clark and A.J. Anderson, eds):183-212. Canberra: ANU E Press. Szabó, K. and J.R. Amesbury. 2011. Molluscs in a world of islands: The use of shellfish as a food resource in the tropical island Asia-Pacific region. Quaternary International 239(1–2):8-18. Szabó, K. and A. Anderson. 2012. The Tangarutu invertebrate fauna. In Terra Australis 37: Taking the High Ground: The Archaeology of Rapa, a Fortified Island in Remote East Polynesia (A. Anderson and D.J. Kennett, eds):135-144. Canberra: ANU E Press. Szabó, K. and B. Koppel. 2015. Limpet shells as unmodified tools in Pleistocene southeast Asia: an experimental approach to assessing fracture and modification. Journal of Archaeological Science 54:64-76. Thakar, H.B. 2011. Intensification of shellfish exploitation: Evidence of species-specific deviation from traditional expectations. Journal of Archaeological Science 38(10):2596-2605. Thakar, H.B., M.A. Glassow and C. Blanchette. in press. Reconsidering evidence of human impacts: Implications of within-site variation of growth rates in Mytilus californianus along tidal gradients. Quaternary International 427(Part A):151-159. Thomas, F.R. 1999. Optimal Foraging and Conservation: The Anthropology of Mollusk Gathering Strategies in the Gilbert Islands Group, Kiribati. Ph.D. thesis. Honolulu: University of Hawaii at Manoa. ---. 2001. Mollusk habitats and fisheries in Kiribati: An assessment from the Gilbert Islands. Pacific Science 55(1):77-97. ---. 2002. An evaluation of central-place foraging among mollusk gatherers in western Kiribati, Micronesia: Linking behavioral ecology with ethnoarchaeology. World Archaeology 34(1):182-208. ---. 2009. Historical ecology in Kiribati: Linking past with present. Pacific Science 63(4):567-600.

---. 2014. Shellfish gathering and conservation on low coral islands: Kiribati perspectives. The Journal of Island and Coastal Archaeology 9(2):203-218. Weisler, M.I. 1999a. The antiquity of aroid pit agriculture and signiﬁcance of buried A horizons on Paciﬁc Atolls. Geoarchaeology: An International Journal 14(7):621-654. ---. 1999b. Atolls as settlement landscapes: Ujae, Marshall Islands. Atoll Research Bulletin 460:1- 51. ---. 2000. Burial artifacts from the Marshall Islands: description, dating and evidence for extra- archipelago contacts. Micronesica 33(1/2):111-136. ---. 2001a. Life on the edge: Prehistoric settlement and economy on Utrōk Atoll, northern Marshall Islands. Archaeology in Oceania 36(3):109-133. ---. 2001b. On the Margins of Sustainability: Prehistoric Settlement of Utrōk Atoll, Northern Marshall Islands. Oxford: Archaeopress. ---. 2002. Archaeological Survey and Test Excavations on Ebon Atoll, Republic of the Marshall Islands. Majuro Historic Preservation Office, Republic of the Marshall Islands. Weisler, M.I., R. Bollt and A. Findlater. 2010. Prehistoric fishing strategies on the makatea island of Rurutu. Archaeology in Oceania 45(3):130-143. Weisler, M.I., J.K. Lum, S.L. Collins and W.S. Kimoto. 2000. Status, health, and ancestry of a late prehistoric burial from Kwajelein Atoll, Marshall Islands. Micronesica 32(2):191-220. Weisler, M.I. and D. Swindler. 2002. Rocker jaws from the Marshall Islands: Evidence for interaction between eastern Micronesia and west Polynesia. People and Culture in Oceania 18:23-33. Weisler, M.I., H. Yamano and Q. Hua. 2012. A multidisciplinary approach for dating human colonization of Pacific atolls. The Journal of Island and Coastal Archaeology 7(1):102-125. Weisler, M.I., H. Yamano, Q. Hua and M. Harris. in prep. Islet size of Pacific atolls constrains the timing of human colonisation: an example from Ebon Atoll, Marshall Islands. Whitaker, A.R. 2008. Incipient aquaculture in prehistoric California?: Long-term productivity and sustainability vs. immediate returns for the harvest of marine invertebrates. Journal of Archaeological Science 35(4):1114-1123. Wiley, A.E., P.H. Ostrom, A.J. Welch, R.C. Fleischer, H. Gandhi, J.R. Southon, T.W. Stafford, J.F. Penniman, D. Hu and F.P. Duvall. 2013. Millennial-scale isotope records from a wide- ranging predator show evidence of recent human impact to oceanic food webs. Proceedings of the National Academy of Sciences 110(22):8972-8977. Yamaguchi, T., H. Kayanne and H. Yamano. 2009. Archaeological investigation of the landscape history of an oceanic atoll: Majuro, Marshall Islands. Pacific Science 63(4):537-565.

Chapter 2: Prehistoric human impacts to marine molluscs and intertidal ecosystems in the Pacific Islands

Note: This is the final version of a peer-reviewed article published in the Journal of Island and Coastal Archaeology

Matthew Harris and Marshall I. Weisler

School of Social Science, The University of Queensland, St. Lucia, QLD 4072 Australia Corresponding author: Matthew Harris, School of Social Science, The University of Queensland, St. Lucia, QLD 4072 Australia. Email: [email protected]

Abstract From long-term stratigraphic records in Pacific Island archaeological sites researchers have documented alterations to molluskan species richness and abundance, decreases or increases in mollusk shell size and, in rare cases, human foraging may have contributed to the extirpation of mollusk taxa. Mollusks perform critical ecosystem functions in tropical intertidal environments, including improving water quality through filtration, regulating algal cover, and increasing habitat and substratum complexity through ecosystem engineering. These critical ecosystem functions can be negatively affected by human foraging, possibly contributing to decreased resilience of coral reefs to climatic alterations. We review modern ecological research on human impacts to mollusks and intertidal ecosystems that illustrates the mechanisms and effects of human foraging. We then examine centuries to millennial scale archaeological records from the Pacific Islands to understand long-term, time-averaged trends in human impacts to intertidal ecosystems.

Keywords: archaeomalacology, anthropogenic impacts, coral reefs, prehistoric foraging, zooarchaeology, marine subsistence

Introduction The study of mollusk remains from archaeological sites, known as archaeomalacology, can reveal important information on human foraging behavior (Harris and Weisler 2016; Szabó, 2009; Thomas 2009), management and exploitation of subsistence resources (Álvarez et al., 2011; Braje et al. 2007; Whitaker 2008), mobility and settlement patterns (Allen 2012), responses to climatic and environmental fluctuations (Amesbury 2007), and can inform on long-term changes to coral reefs and human interaction with the marine environment (Morrison and Addison 2008:16; Szabó and Amesbury 2011). Mollusks are the most species rich class of marine invertebrates, and a critical component of tropical intertidal ecosystems (Glynn and Enochs 2011). Active at all trophic levels, mollusks perform important ecological functions, including increasing habitat complexity (Andréfouët et al. 2005; Gutiérrez et al. 2003), improving water quality through filtration (Hutchings et al. 2007), and regulating algal cover through herbivory (Hockey and Alison 1986).

The intertidal environments that mollusks inhabit are varied and ubiquitous, and despite decades of dedicated research the ecological role of human foraging in these ecosystems is not well understood (Allen 2003; Brander et al. 2010; Castilla 1999; Castilla and Duran 1985; Catterall and Poiner 1987; de Boer and Prins 2002; Erlandson and Rick 2010; Fitzpatrick and Donaldson 2007; Poiner and Catterall 1988; Rick and Erlandson 2008; Rivadeneira et al. 2010). Measuring the impact of human foraging is difficult given the rarity of cases where single stressors can account for documented alterations to intertidal communities (McClanahan et al. 2014:459), and general lack of modern ecological data necessary to measure the impact of humans on mollusks in the tropical Pacific (Kinch 2003:5). However, studies of human interaction with marine mollusks report a range of direct and indirect effects including, alterations to community structure and composition (Hockey 1987; Hockey and Alison 1986), change in the body size of exploited species (Giovas et al. 2010; Spennemann 1987), or a lack of discernible impacts (de Boer and Longamane 1996; Morrison and Addison 2008; Weisler 2001:116). The study of mollusks from archaeological sites in the Pacific Islands has revealed that human foraging during the prehistoric period altered mollusk species richness, abundance and diversity, physical size, and may have led to the extirpation of some taxa. Data generated by archaeologists has also contributed valuable long-term historic records of human interaction with the marine environments of the Pacific Islands (Dalzell 1998; Jones 2007).

Mollusk shells from archaeological contexts provide excellent proxy records for human interaction with the intertidal zone as their shells preserve well, especially compared to other marine fauna, and 16 are ubiquitous in most coastal Pacific Island archaeological sites. Archaeologists often investigate human disturbance to mollusks by assessing alterations to species abundance, diversity and richness, and changes to body size. These data can be compared with archaeological, paleontological and modern ecological data to identify local depletion and extirpation of marine fauna (Dulvy et al. 2003; Edgar and Samson 2004:1579; Erlandson and Fitzpatrick 2006), and contribute to modern conservation efforts (e.g., Aswani and Allen 2009; Jones 2009; Lauer and Aswani 2010; Thomas 2009).

Human impacts to mollusks during Pacific Islands prehistory have been routinely documented, but analytical and methodological variation hinder a regional synthesis of these studies. Issues with the quantification of mollusk remains from archaeological sites have been discussed elsewhere (Giovas 2009; Harris et al. 2015; Mason et al. 1998), but variable or unclear reporting of quantification protocols prevents rigorous comparison of assemblages quantified variably using weight, Number of Individual Specimens (NISP), or Minimum Numbers of Individuals (MNI). As standards of archaeological excavation have changed substantially throughout the development of Pacific Island archaeology, variable screen size, sampling (not all excavated mollusks are retained for analysis), and dating protocols hinder precise comparison of archaeomalacological datasets.

Rather than focus on quantitative comparison of datasets to discern regional or temporal patterning in human foraging impacts, we discuss archaeological evidence for human impacts based on three inferred outcomes of human foraging: (1) extirpation or extinction of mollusk species, (2) changes in the size/age structure of mollusk populations, and (3) signatures of trophic alteration, most commonly present as changes to species richness, diversity and abundance. We document archaeological cases where human foraging has been reported as the primary driver for changes in archaeological mollusk assemblages and suggest future research avenues for the region. Modern ecological research on human impacts to mollusks is reviewed to elucidate the mechanisms of change that may underpin long-term, time-averaged trends in archaeological datasets from the Pacific Islands.

The Effects of Human Predation Human predation directly reduces the abundance and biomass of mollusk populations (Green and Craig 1999:210; Zhang et al. 2013:237), which can also alter the mean size of the prey population (Castilla and Bustamante 1989; Fairweather 1990:453), and in cases of intensive and sustained predation, local extirpation can occur (Dulvy et al. 2003). In addition to these direct effects,

17 predation can cause indirect effects on community structure as the targets of human predation are often keystone species that have a disproportionate influence on associated flora and fauna relative to their abundance or biomass (Paine 1995:962; Power et al. 1996:609). Predation of keystone species can effect organisms at all trophic levels and decrease the ratio of predators to prey, thus reducing interspecific competition and potentially increasing the abundance of subordinate, often r- selected species (de Boer and Prins 2002; Fairweather 1990). Conversely, human predation can increase diversity and richness by creating a mosaic of patches with different stages of succession (de Boer et al. 2000), especially through the reduction of herbivorous taxa that regulate floral cover, thus further impacting a range of ecological processes (Godoy and Moreno 1989; Hockey and Alison 1986; Sagarin et al. 2007). Human predation at a number of locations along a coastline can also cause faunal and floral communities to converge towards a common state of similar taxonomic richness, diversity and abundance (Hockey and Alison 1986).

However, equifinality and the dynamic, unpredictable effects of predation can hinder precise assignment of causation (Fairweather and Underwood 1991). As Fairweather (1990:453) surmised, predation is a form of ecological interaction that can be “maddeningly idiosyncratic”. In modern ecological studies, traditional subsistence harvesting is increasingly recognised as strongly influencing invertebrate assemblages, though the outcomes of this type of predation are difficult to measure and document due to the range of non-human stressors in the intertidal (Jimenez et al. 2012:90). The relationship between predation, recruitment, and change over time in the relative abundance of marine invertebrates in intertidal areas is complex, and even simple studies of predator-prey interaction have produced divergent results over a short period (Fairweather 1988). Fecundity and growth rates, size at maturity, escape and avoidance mechanisms such as burying or benthic mobility, the presence of adjacent populations for replenishment of stocks, and larval phases (especially taxa with pelagic larva) can complicate understanding of the impacts of human exploitation (Catterall and Poiner 1987:Table 7; de Boer et al. 2000:288). For archaeologists and ecologists alike, the lack of ecological information on commonly exploited taxa, the response of these populations to human predation in tropical intertidal settings, and limited understanding of the human role in structuring intertidal populations reduces our confidence in isolating anthropogenic influence (Castilla and Duran 1985; Catterall and Poiner 1987:119; McShane et al. 1994; Poiner and Catterall 1988).

An Historical Ecology Approach Pacific Island coral reefs are amongst the most structurally complex and taxonomically diverse ecosystems in the world (Glynn and Enochs 2011; Jackson et al. 2001:631). The human colonization and settlement of the Pacific Islands was the most expansive maritime migration in history, and local adaptations reflect the diversity of island types (continental, high volcanic, makatea and atolls) and time scales of occupation from many thousands of years to several centuries (Fitzpatrick and Callaghan 2013; Irwin 1992; Kirch 2000; Weisler et al. 2016). Pre- European human impacts to these ecosystems have been underestimated by ecologists (Erlandson and Braje 2013; Smith and Zeder 2013), and so-called ‘pristine’ environments are commonly the current representation of a deep history of anthropogenic influence (Alden Smith and Wishnie 2000:496; Myers and Worm 2003; Pinnegar and Engelhard 2008). If human occupation of Pacific Islands, and mollusk foraging in these environments has decreased the resilience of tropical coastal ecosystems to current climatic alterations (Jackson et al. 2001), the current state of coral reefs cannot be fully understood, explained, conserved or restored without an historical perspective (Hayashida 2005:45; McClanahan et al. 2014:59). Documenting human impacts across broad temporal and spatial scales requires a multidisciplinary engagement with geological, biological, ecological, archaeological and anthropological data (Baisre 2010:137; Erlandson and Fitzpatrick 2006:7). The unique long-term historical data provided by zooarchaeologists can help elucidate proximate and ultimate causes of anthropogenic ecological change (Braje et al. 2005; Carder and Crock 2012; Jackson et al. 2001; Wolverton and Lyman 2012). Zooarchaeological data alone cannot provide the answers, and integrated research utilizing the methodological frameworks of historical ecology to compare temporally disparate datasets has become increasingly common (Briggs et al. 2006; Hayashida 2005; Pinnegar and Engelhard 2008).

We follow the broad definition of historical ecology as a multidisciplinary endeavor that uses historic and prehistoric data to understand ancient and modern ecosystems, often with the goal of providing context for contemporary conservation (Rick and Lockwood 2013:46-47). Rather than treat humans and the environment as separate, historical ecologists investigate the dialogue between them, and view the landscape as an outcome of these interactions (Balée 2006:77; Braje et al. 2005:6). Analyses based in historical ecology frameworks have eliminated the bias of the shifting baselines syndrome identified by Pauly (1995), where each generation of fisheries scientists had accepted the state of marine fisheries at the beginning of their careers as a baseline for future stock assessments, an underestimation of pre-European environmental impacts (Erlandson and Braje 2013; Erlandson and Rick 2010), and a lack of time-depth in empirical records. Historical ecology

19 has facilitated deep historical analyses of current ecosystems (Baisre 2010; Bunce et al. 2008; Carder and Crock 2012; Cramer et al 2015.; Morrison and Hunt 2007; Pauly 1995; Rick and Lockwood 2013; Small and Nicholls 2003; Wake et al. 2013). Recent historic and ancient archaeological records have been compared to assess fishing down food webs (Blick 2007; Pestle 2013), regime shifts, alternative stable states, trophic cascades and trophic level analyses, extinctions (Dulvy et al. 2003; Seeto et al. 2012), and to establish baselines and reference states for restoration ecology and conservation biology (Baisre 2010; Carder and Crock 2012). Historical ecological research has revealed that even relatively minor artisanal or subsistence fisheries can have long-term effects on marine environments (Briggs et al. 2006), and comparisons with historical data have shown that current perceptions of ‘natural’ environments rarely match long- term records (Alleway and Connell 2015).

While historical ecology has many potential benefits as an integrative framework for multidisciplinary analyses, the comparison of datasets that might span days, decades, or up to tens of thousands of years can be methodologically difficult. Integrating archaeological and paleontological data with modern ecological data requires sacrificing the “apparent precision and analytical elegance” of modern ecology (Jackson et al. 2001:630). Historical ecologists must be aware of the limitations and strengths of the data, but the ability to observe time-averaged patterns of change may offset the sacrifice in analytical precision that studies of extant environments alone can provide (Rick and Lockwood 2013). Below, we review a range of modern case studies to highlight the ways that human foraging can impact mollusk populations, and present archaeological research from the Pacific Islands that has reported human impacts to mollusks.

Anthropogenic Extirpation of Mollusks in the Pacific Islands Declines in the abundance of tridacnid clams (Cardiidae: Tridacninae) have been reported in both archaeological and ecological studies in the Pacific Islands (Green and Craig 1999; Matthews et al. 2003; Morrison and Allen in press and references therein; Newman and Gomez 2002). In the late 20th century, extirpations were reported for a number of archipelagos, with Hippopus hippopus (horse’s hoof clam) likely extinct in Guam (Paulay 1996), Fiji (Lewis et al. 1988) and Tonga (McKoy 1980 in Lewis et al. 1988). Tridacna gigas (giant clam) has been extirpated from the Caroline Islands and the Fiji archipelago (Lewis et al. 1988:67), and Tridacna spp. are rare in the Gilbert Islands (Tebano and Paulay 2000). Tridacna are largely depleted from the Samoan archipelago, with an inverse relationship between human population size and tridacnid abundance

(Green and Craig 1999:210). In Palau, Matthews et al. (2003:3) reported an overall decline in tridacnids in areas where clam population density had traditionally been expected to be high.

While the causes of these declines are multi-faceted, and changes in tridacnid distribution are dynamic over time and space (Copland and Lucas 1988; Newman and Gomez 2002), human exploitation has been identified as a causal factor. Modern coastal development and alterations to settlement patterns and agricultural activity during the prehistoric period increased terrigenous runoff and water turbidity in the intertidal, negatively impacting the symbiotic relationship between photosynthetic zooxanthellae and the tridacnid host (Kirch 1983; Morrison and Cochrane 2008; Rolett 1992). Tridacnids may also be susceptible to overharvesting given their large, conspicuous size, tendency to live in dense, spatially aggregated colonies, and slow growth rate (Catterall and Poiner 1987; Morrison and Allen in press; Poiner and Catterall 1988; Whitaker 2008).

Documenting local or regional mollusk extirpation in archaeological contexts is difficult as the presence or absence of a particular species is mediated by a range of stochastic and unpredictable ecological phenomena that structure reef assemblages at any given time, many of which are archaeologically intractable (Jones 2009:637; Sale 1980; Thomas 2009:580). Establishing island, or archipelago wide archaeological evidence for the extirpation or extinction of mollusks requires temporally controlled, regional archaeomalacological datasets suitable for comparative analyses, which are rare in the Pacific Islands. Thus, anthropogenic extirpation and extinction of mollusks has rarely been reported in Pacific archaeology (Allen 2003:324; Seeto et al. 2012). However, local extirpation has been tentatively proposed for the northern Marshall Islands (Weisler 2001) and the Fiji archipelago (Seeto et al. 2012). Weisler (2001:125) documented the presence of the bullmouth helmet shell (Cypraecassis rufa) in archaeological sites at Utrōk Atoll, and the lack of modern records for this taxon in the Marshall Islands as potentially indicative of anthropogenic extirpation during the prehistoric period. Seeto et al. (2012:11-13) propose five non-mutually exclusive explanations—overexploitation, introduced predators and diseases, increased sea surface turbidity, and sea level fall—for the disappearance of Hippopus from the Boureawa and Qoqo (Fiji) archaeological sites. While sea level fall and consequent habitat loss during the Lapita occupation of the site (c. 2500 - 3050 BP) is inferred as the primary driver of the extirpation of Hippopus, a general reduction in shell size over time may indicate that exploitation pushed local Hippopus populations “beyond a threshold where the population could become re-established” (Seeto et al. 2012:11).

Changes in the Size/Age Structure of Mollusk Populations The most common model for human-induced mollusk size reduction hypothesizes that foragers will target the largest and oldest individuals, only selecting smaller and younger mollusks as larger individuals are removed. Over time, the mean size of the population will decrease, with consistent exploitation preventing growth to larger sizes, and prohibiting growth to sexual maturity (e.g., Allen 2012; Mannino and Thomas 2002; Swadling 1976). Often, physical size is used as a proxy for age, however, changes in mollusk size can be induced by a range of non-anthropogenic factors, and inferences of genuine human impact should ideally be demonstrated by both changes in size and in ontogenetic development and age (Swadling 1976; Branch and Odendaal 2003). Reduction in size can impact ecological functioning, including the regulation of subordinate prey species (Catterall and Poiner 1987; de Boer et al. 2000) and macro-algal cover (Hockey and Alison 1986; Sagarin et al. 2007). A reduction in mean body size and age can also alter sex ratios of populations and decrease average reproductive output over the lifetime of the organism (Brander et al. 2010). These changes can be critical to the long-term survival of mollusk populations, especially for sequentially hermaphroditic taxa (individuals that change sex throughout life) (Branch and Odendaal 2003; Fenberg and Kautstuv 2008).

Moreno et al. (1984) found that keyhole limpets (Fissurella spp.) were less abundant and smaller at sites where human foraging was heaviest (Moreno et al. 1984:159). Human exclusion from the intertidal over a 6000 km2 area of the Chilean coast resulted in increased densities and a larger median size of Chilean abalone (Concholepas concholepas) (Castilla and Duran 1985) and a greater size range of the owl limpet, Lottia gigantea (Sagarin et al. 2007). Human harvesting was also correlated with a decline in reproductive potential of Cymbula oculus limpets in south-east South Africa where sites protected from harvesting had 2.7 times as many adults as unprotected sites, and a 182-fold increase in female reproductive output (Branch and Odendaal 2003:259-261).

Planes et al. (1993) documented the diminution of Tridacna maxima at sites in Bora-Bora Lagoon, Society Islands, French Polynesia where tourists regularly consumed clam meat. While coastal quarrying and land reclamation negatively affected clam habitats, which likely contributed to the decline in mean shell size, T. maxima mean length was 12-16 % smaller at sites regularly exploited by tourists (Planes et al. 1993:5; Table 1). Aswani et al. (2015) found that areas of Roviana Lagoon, Solomon Islands, that were permanently closed or restricted for most of the year had larger Anadara granosa (blood clams) and Polymesoda spp. (mud clams) individuals than areas that were constantly accessed by foragers. Populations of both taxa from restricted areas had greater numbers

22 of sexually mature individuals and thus, a greater potential for clam reproduction than non- restricted coastal areas (Aswani et al. 2015:226).

Pacific Islands archaeologists have routinely inferred a link between human harvesting and reduction in mean shell size by measuring mollusks preserved whole, and/or establishing morphometric relationships between shell features to generate formulae for predicting shell size from fragmented remains (e.g., Kirch et al. 1995; Swadling 1976; Thangavelu et al. 2011). There is ongoing debate as to the interpretive validity of predictive formulae that do not account for allometric or ontogenetic growth and/or the influence of local ecological conditions on shell size (Campbell 2015; Faulkner 2010; Jerardino and Navarro 2008). Singh et al. (2015) argued that rather than using formulae for understanding mollusk ontogeny and evolutionary development, archaeologists seek to describe changes over time; current methods for generating predictive formulae based on morphometric data drawn from a wide geographic area are therefore appropriate for archaeological inquiry.

Pamela Swadling first investigated shell size change in the Pacific Islands using a combination of metric analyses and age determinations based on shell features to investigate the effects of human exploitation on mollusk populations at prehistoric sites in New Zealand, Papua New Guinea and the Solomon Islands (Swadling 1976, 1977 see also Green 1986). Grounded in molluskan biology and ecology, Swadling investigated changes in mollusk size and recorded the presence or absence of shell features indicating sexual maturity, such as the thickening of the apertural lip in the Strombidae (Swadling 1976:158-159), fluting on the lower whorls of Trochidae (Swadling 1986), and the spacing of annular growth striae in bivalves (Swadling 1976:158). Swadling sought to determine if changes observed in mollusk body size were produced by “human or environmental factors” (Swadling 1976:156, emphasis added), inferring that intensive human exploitation pressure led to a decline in the size and age of Austrovenus stutchburyi (New Zealand little neck clam, formerly Chione stuchburyi) populations at Maori pā sites in northern New Zealand. Swadling also proposed that Lapita occupation of the Reef/Santa Cruz Islands in the Solomon Islands likely led to a decline in the abundance of large top shells, Tectus niloticus (formerly Trochus niloticus), and slight reductions in the means size and age of Tectus pyramis and the ark clam, Anadara antiquata. Swadling (1986:145-146) used mollusks from Lapita period middens in the Reef/Santa Cruz sites, and compared the assemblages to natural relative abundances of Indo-Pacific taxa to determine foraging patterns. Swadling challenged Groube’s (1971) now outdated ‘Oceanic Strandlooper’ hypothesis, proposing instead that the inhabitants of the SE-RF-2 site likely relied on a mixed

23 horticultural and marine diet, rather than depleting marine resources prior to focusing on shifting cultivation.

Swadling’s work provided a foundation for other Pacific archaeologists investigating prehistoric human impacts to mollusks. Declines in shell size have been detected for Haliotis iris (Pāua), Lunella smaragda and Cellana denticulata at Pallier Bay, New Zealand (Anderson 1981), Nerita in Lakeba, Fiji (Best 1985), Gafrarium spp. at Tongatapu, Tonga (Spennemann 1987, 1989), Turbo setosus at Mangaia, Cook Islands (Kirch et al. 1995) and Turbo smaragdus at the Haratonga Beach Site, New Zealand (Allen 2012). Morphometric approaches have also been used to track decreases in shell size in north-eastern Fiji (Thomas et al. 2004), Palau (Masse et al. 2006), the western coast of New Zealand’s South Island (Jacomb et al. 2010), and Papua New Guinea (Thangavelu et al. 2011). These studies follow the model first employed in the Pacific Islands by Swadling (1976), correlating increasing human exploitation pressure with decreases in shell size over time.

Inferring human impacts based on declining shell size over time has been criticized as many archaeologists have associated size change with human predation without examining alternate explanations such as environmental and climatic influences (Giovas et al. 2010:2797; Thakar et al. in press). Assigning humans as primarily responsible for declining mollusk size in archaeological contexts often lacks detailed consideration of the complex range of factors that would have shaped these processes. While human impact can reduce mollusk shell size, the relationship between shell size and human exploitation pressure is rarely clear. Molluskan body size is a dynamic process influenced by a range of ecological processes, and causal links between human predation and body size are difficult to conclusively establish over both short and long-term records (e.g McShane et al. 1994). Recent archaeological and ecological research has highlighted that solely attributing decreasing shell size to human foraging lacks consideration of the range of non-anthropogenic phenomena that can alter intertidal community structure. Equifinality in intertidal ecosystems has traditionally received little attention in Pacific Island archaeomalacological literature (Giovas et al. 2010:2788; Jones 2009).

Influenced by historical ecological concepts and acknowledging the myriad factors influencing shell size, Pacific Islands archaeologists are increasingly considering non-human explanations for mollusk size change. Giovas et al. (2010) examined a mollusk assemblage from Chelechol ra Orrak, in Palau’s Rock Islands, documenting that the humpbacked conch, Gibberulus gibberulus (formerly Strombus gibberulus) increased 1 – 1.5 mm from c. 3000 BP to the present. Giovas et al.

(2010:2793-2796) hypothesize eight interrelated phenomena that may have contributed to this size increase at the site, including trophic alteration through foraging, changes in local and regional climatic and environmental conditions, and alterations to foraging patterns. Giovas et al. (2010:2796) conclude that “there is insufficient evidence to support any one of these eight hypotheses over all others” and stress the need for additional paleoenvironment data, and the use of high-resolution techniques for determining ontogenetic development and age of mollusk shells, rather than using size alone. Studies that consider the influence of human foraging, environment and climatic factors, and consider the ecology and biology of mollusks demonstrate that humans are only one of the complex factors that influence mollusk size through time. While alterations to physical size of mollusks can inform on human impacts, there are a range of wider impacts that can result from human foraging, such as alterations to trophic networks in tropical intertidal ecosystems.

Trophic Alteration and Impacts to Species Richness, Abundance, and Diversity Human foraging can enact a range of top-down direct and indirect alterations to intertidal trophic networks that reduce species abundance, diversity, and biomass. Intensive mollusk collecting can enact long-lasting changes to intertidal ecosystems (e.g., Hockey 1987) given the many critical roles mollusks play in trophic networks (Glynn and Enochs 2011; Odum and Odum 1955). Invertebrates, including mollusks, are ubiquitous at low trophic levels, and elevate nutrients to higher trophic levels (Klumpp and Pulfrich 1989). Large filter feeding mollusks improve water quality, with Tridacninae recorded in Tonga filtering 600 ml of water per minute (Klumpp and Lucas 1994). Aggregations of mollusk shells can also increase substratum heterogeneity and increase sedimentation rates on the reef (Gutiérrez et al. 2003). In the Fangatau Atoll lagoon, T. maxima is so hyper-abundant that it is the primary builder of lagoonal reef structures, creating large ridges and mounds of hard, topographically complex substratum (Andréfouët et al. 2005:1039). The addition of mollusk shell to reef habitats can also modulate the amount of substrate available for colonization by other organisms, or create structures which provide refuge from predation (Gutiérrez et al. 2003:81). Bivalves that bore into the substrate create refuges for fauna that live deeply embedded in these habitats (Hutchings et al. 2007; Stier and Leray 2014) and bioeroding mollusks create coral rubble and other carbonate sediments (Glynn and Enochs 2011:297). Herbivorous grazing mollusks are effective regulators of algal cover in the intertidal (Klumpp and Pulfrich 1989; Moreno et al. 1984). The removal of highly effective grazing fauna, such as limpets and herbivorous fishes, can induce a phase-shift in habitats dominated by coral or rock to algae, impacting a range of ecosystem functions and altering species composition (e.g., Hughes et al. 2007; Sagarin et al. 2007:400).

Humans are also effective ecosystem engineers (organisms that modify, create or maintain habitats), and often physically alter habitats in the intertidal by constructing architectural sites and walled fishponds along the coast (Castilla 1999:280; Jones et al. 1996; Weisler and Kirch 1985:Figure 3). Kataoka (1996) noted an increase in the bivalves Tellina palatum, Anadara antiquata, and Gafrarium pectinatum from mangrove habitats created by the construction of the Nan Madol architectural complex on Pohnpei. The basalt columns and rubble fill from artificial islet construction also created new habitats for the polished nerite, Nerita polita, which increased in abundance in archaeological assemblages around 950 -1050 BP. Although not in the Pacific, during the 19th and 20th centuries, a decline in agricultural productivity and an increasing human population on Fuerteventura Island, Canary Islands has been linked to the extinction of the Canary Island oystercatcher, Haemotopus meadewaldoi. Increased competition for food, restricting nesting sites, and a progressive reduction in invertebrate stocks (primarily the limpet Patella candei and the mussel Perna picta) due to human harvesting contributed to the extinction of H. meadewaldoi (Hockey 1987:61).

Examining the impact of human foraging on interactions between mollusks and other intertidal fauna is rare in Pacific Island archaeology. Masse et al. (2006) presents quantified predatory crab attack scars on Gibberulus gibberulus shells from archaeological assemblages to assess interspecific competition. Crab predation on mollusks can strongly structure gastropod populations (Tyler et al. 2014 and references therein) and Masse et al. (2006:124) proposed that as human foraging increased in intensity, the number of crab attack scars on the shell should decrease as both mollusks and crabs are targeted by human foragers (Masse et al. 2006:124). Masse reported a significant decline in the number of crab attack scars in assemblages from Uchularois Cave and Tmasch in Palau’s Rock Islands from the 9th century AD. Combined with the results of the analyzed fish remains (Fitzpatrick et al. 2011; Fitzpatrick and Kataoka 2005), and morphometric analysis of the G. gibberulus assemblage, significant evidence was presented for human induced resource depression in the Rock Islands (Clark and Reepmeyer 2012).

The removal of keystone predators from intertidal environments can increase subordinate taxa abundances, and reduce species diversity (Beauchamp and Gowing 1982). From modern observations, Castilla and Duran (1985) reported that human foraging of Conchohlepas conchohlepas in central Chile led to declines in mollusk species richness and an increase in the abundance of the subordinate mussel Perumytilus purpuratus. When human predation of C.

26 concholepas ceased, species richness increased as multiple taxa were able to utilize the space previously dominated by P. purpuratus. Hockey and Alison (1986:11) also reported that low-level human exploitation targeting high-trophic level predators at several sites on the south-east coast of South Africa led to increased richness, and similar taxonomic abundances and community structure at exploited sites. These studies clearly demonstrate the short-term effects of human intertidal foraging, while the archaeological record offers a unique long term perspective on the role of humans in modifying the abundance, richness and diversity of mollusk communities.

Across the Pacific Islands, archaeologists have reported declines in the relative abundance of targeted prey with a concurrent increase in the abundance of smaller taxa, taxonomic richness, or both. While alterations to relative abundance and richness are not always attributed to foraging pressure (see Best 1985), researchers in the region have routinely interpreted these patterns as the result of human exploitation. Nichol (1986) proposed that humans overexploited Cellana denticulata limpets and Amphibola at early-phase Hahei, New Zealand, driving foragers further from the site to collect Perna mussels and smaller-bodied limpets, such as Cellana radians. Walter (1998:86) proposed a similar outcome of human foraging for Turbo setosus from the algal ridge offshore from the archaic Anai‘o site, Cook Islands. Within 200 years of the occupation of Hanamiai Valley, Marquesas Islands beginning around 850 BP, the gastropod Trapezium oblongum and the chiton, Chiton marquesanus disappear from the sequence (Rolett 1992), co-occurring with increases in pig bone and anthropophilic land snails associated with Polynesian gardening. Weisler (1995) reported the rapid decline in relative abundance of Cerithium tuberculiferum at HEN-10 on Henderson Island (Pitcairn Group) around 520-670 BP as the result of human predation. Weisler (1999 2001) also noted a decline in Tridacna and concurrent increase in Cerithium in the main village site on Ujae Atoll, and documented a preference for large tridacnid clams at short-term camp sites away from the main villages on both Utrōk and Ujae Atolls. Nunn et al. (2007:119) reported a decline in the density (g/m3) of mollusk remains during the later phases of Lapita occupation of Naitabale, Moturiki Island, Fiji (c. 2950-2600 cal. BP) as indicative of early human colonizing populations depleting nearshore resources.

Other researchers have explored alterations to relative abundance using faunal abundance indices (AIs) (Broughton 1994, 1997). Studies employing abundance indices use body size as a proxy for foraging returns, positing that a decline in “large-bodied” relative to “small-bodied” taxa represents a decline in foraging efficiency (Allen 2012), or to track alterations to habitat representation in archaeological assemblages (Morrison and Cochrane 2008). AIs have been used to track changes in

27 mollusk deposits in New Zealand (Allen 2012), Fiji (Morrison and Cochrane 2008) and Hawaii (Morrison and Hunt 2007). Allen (2012) grouped mollusk taxa from the Haratonga beach site, New Zealand, based on average shell size and body weight into four classes (A-D), examining the decline in classes A (>105 g or >100 mm) and B (105-15 g or 100-60 mm) relative to classes C (15- 1.5 g or 60-25 mm) and D (1.5 to < 1 g or <25 mm). A second index examined the relative abundance of the larger-bodied trochid Cookia sulcata to the smaller-bodied nerite, Nerita atramentosa. Taxa from size classes A and B, and C. sulcata decline in abundance, while N. atramentosa abundance increases, was inferred to be the result of continuous harvesting pressures (Allen 2012:304-306). Similarly, a decline in the abundance of Turbo sandwicensis and an increase in Canarium maculatum (formerly Strombus maculatus), and stable abundances of other near-shore mollusks was reported as evidence for anthropogenic impacts to coral reef habitats at Nu‘alolo Kai, Kaua‘i, Hawaiian Islands (Morrison and Hunt 2007).

At the Natia Beach Site, Fiji, Morrison and Cochrane (2008) tracked declining relative abundances of the sand-dwelling bivalves Gafrarium tumidum and Atactodea striata from the initial occupation of the site around 2170 cal. BP, coincident with a decrease in “large-bodied” Trochus niloticus, Anadara antiquata and Turbo crassus relative to the “small-bodied” gastropods Turbo cinerus, Nerita sp., and Planaxis sulcatus. Overharvesting, the progradation of the Natia Beach terrace, increased terrigenous runoff associated with use of the uplands in Fiji around 500 BP, and increased ENSO activity post 650 BP all may have contributed to changing abundances of mollusks yet, ultimately, more analysis is needed to establish broader, regional patterns and to fully comprehend the role of humans in these changes (Morrison and Cochrane 2008:2397)

Researchers have justified the use of body size as a predictor of foraging returns based on a range of factors, including a lack of evidence to indicate that processing and handling costs differed between taxa (Morrison and Hunt 2007:339), and data from terrestrial foraging studies that show a correlation between hunting efficiency and large-bodied prey are often cited (e.g., Broughton 1994; Byers and Broughton 2004; Codding et al. 2010). Using body size to measure foraging returns has been challenged based on the predictable aggregations of some marine animals, requiring little technological investment, search time or risk for collection of large numbers of small individuals (Erlandson and Fitzpatrick 2006:11; Whitaker 2008). Tracking changes in mollusk assemblages using abundance indices can facilitate assessment of alteration to mollusk foraging when comparative analytical units are carefully considered, and have proven utility for tracking changes in habitat selection or processes of environmental or climatic alteration (Allen 2003:320; Morrison

28 and Cochrane 2008). However, researchers must be aware of the implications of defining “large” and “small” bodied taxa as a proxy measure of foraging returns. While all models rely on simplification, (Boyd and Richerson 1987 Railsback and Grimm 2011) tracking body size as a primary indicator of foraging declines risks reducing complex human foraging behaviors to “simplistic truisms such as the notion that large animals are more productive than smaller animals”, regardless of the prey type, environment (Erlandson and Fitzpatrick 2006:11), and methods of capture. In the Pacific Islands, the use of body size as an indicator of foraging returns has also received criticism for denying those aspects of marine foraging and fishing that are not fitness- maximizing, but mediated by human agency and/or cultural factors (e.g., Jones and Albarella 2009).

One method to more comprehensively model mollusk foraging is to utilize generative methods for hypothesis testing, such as agent-based modelling (ABM). ABM can be used to simulate complex social and ecological phenomena, for modelling outputs that are compared with archaeological data (Gilbert 2008). Morrison and Addison (2008) developed two simple models of the role of climate change impacts in human foraging. One model simulated mollusk resources that were replenished quickly in coral habitats unaffected by climate change, and the second modelled forager response to declining mollusks from coral reef habitats affected by climate-induced coral bleaching. When compared to the Fatu-ma-Futi site, American Samoa, the taxonomically rich and even mollusk assemblage deposited over c.1500 years correlated well with the stability model, indicating that human foraging and climatic fluctuations did not affect nearshore mollusk populations. Using more complex models, Morrison and Allen (in press) examined the influence of age at maturity, reproductive output and energetic return, and of prey aggregation on human overexploitation. The age at maturity/energetic return model is compared to several Polynesian archaeological sites that exhibit a decline in Tridacna, or other large bodied taxa over the course of occupation (Allen 1992, 2012; Kirch and Yen 1982; Morrison and Addison 2008; Morrison and Hunt 2007). Humans were considered the primary driver for the declining abundance of large-bodied taxa, but “conclusive demonstration that ancient foragers were responsible [for reported declines in mollusk taxa] will ultimately require evaluation of the contemporaneous paleoclimate conditions, and local reef histories” (Morrison and Allen in press:9). By incorporating mollusk ecology, human foraging data, and local ecological and palaeoclimatological data, ABM has the potential to test hypotheses of human impacts in a high-resolution, explicit manner across space and through time. These models also have the potential to model the impacts of the wide range of non-anthropogenic processes that can impact mollusk populations.

Non-Anthropogenic Alterations to Mollusk Assemblages There are a range of non-anthropogenic phenomena that can lead to changes in molluskan body size, extirpation, extinction, and alterations to trophic networks. Mollusks are especially vulnerable to non-anthropogenic stressors as they are generally sedentary, and unable to flee or relocate from danger. Mollusks are subject to potentially lethal desiccation and thermal stress during diurnal low tides (Kohn 1987:139; Petes et al. 2007; Yamaguchi 1975) and alterations to salinity due to rainfall, especially in the tropics (Hancock and Simpson 1962:44; Yamaguchi 1975:229). Seasonally large waves generated by storms and typhoons can lead to mortality of mollusks (Blumenstock 1958; Kohn 1980) reducing species richness, and increasing dominance of those taxa which inhabit physical refuges on intertidal platforms, often predators such as cone shells (Conidae) or Muricidae, potentially leading to changes in trophic networks (Kobluk and Lysenko 1993; Leviten and Kohn 1980).

Sea-level fluctuations can lead to mass mortality of reef organisms and wide-scale mollusk extirpations as recorded in the central Pacific during the late Cenozoic (Paulay 1990), and over short time scales including recent evidence for almost complete mortality of macroscopic organisms on the reef flat in Pago Bay, Guam following rapid sea level fall in 1972 (Yamaguchi 1975). Holocene records for the Pacific Islands show several periods of sea level regression and tectonic uplift that have been correlated with alterations to mollusk assemblages (Clark and Reepmeyer 2012; Dickinson 2003; Nunn 1990; Webb and Kench 2010; Woodroffe 2008). Amesbury (1999, 2007) proposed that sea level decline during the last 3000 – 4000 years correlated with alterations to archaeological mollusk assemblages on Guam. Pre-Latte phase middens show a decline in the mangrove adapted Anadara antiquata (c. 3500 – 1600 BP) and an increase in sand-dwelling Tellinidae in the transitional phase (c. 1600 – 1000 BP). Latte phase middens (c. 1000 BP onwards) are dominated by Strombus, ubiquitous on intertidal sand flats. Without consideration of the effects of sea level, these patterns “might have been interpreted as overharvesting by the earliest inhabitants of the Marianas” (Amesbury 2007:955). Similarly, Spennemann (1987) noted a decrease in the abundance of Anadara in middens on Tongatapu, Tonga associated with sea level fall, with a concurrent increase in Gafrarium as the lagoon became increasingly brackish.

Discussion The analysis of mollusk assemblages from Pacific Islands archaeological sites has documented that human foraging potentially impacted shell size, altered species richness and abundance, changed trophic networks and, in some cases, may have extirpated mollusk species adjacent to

30 archaeological sites, or from whole islands. However, Pacific Islands archaeologists have linked changes in mollusk assemblages with human foraging, especially changes in shell size, without detailed consideration of other explanations. Current archaeomalacological research is presenting more balanced approaches to investigating human foraging, better integrating the ecology of the intertidal zone (Giovas et al. 2010) and utilizing multi-proxy environmental data on sea level and other geological and climatic changes, as well as comparative zooarchaeological data from other faunal classes, especially finfish (Fitzpatrick and Kanai 2001). These studies draw upon the tradition of multi-disciplinary research in the Pacific Islands for understanding long-term patterns of human interaction with the oceans (Aswani and Allen 2009; Fitzpatrick and Donaldson 2007; Jones 2009; Jones and Albarella 2009; Jones and Quinn 2009; Kataoka 1996; Kittinger et al. 2011). The methodological frameworks of historical ecology have influenced the discourse on human impacts to mollusks, and allowed archaeologists to contribute to current debates on the historic role of humans and global reductions of marine fauna (Carder and Crock 2012).

There has been a bias towards analyzing large-bodied mollusks in Pacific Islands archaeology, but recent integration of 3.2 mm and finer mesh screening for recovery protocols has yielded significant data on smaller taxa in archaeological assemblages as by-catch useful for environmental reconstruction, or as evidence for fine-grained or non-selective foraging strategies (Szabó 2001, 2009). These results should encourage archaeologists to examine entire assemblages, rather than only analyzing those large-bodied taxa with the highest relative abundance. Quantification protocols are more routinely and transparently reported, and the integration of multiple measures of relative abundance (most commonly NISP, MNI, and weight) facilitates regional comparisons. The archaeological study of human impacts to mollusks in the Pacific Islands is patchy (Figure 1), especially for low coral islands (Thomas 2009; Weisler 1999, 2001). When human impacts to mollusks are identified, it is often represented by isolated sites lacking supportive data at the island or archipelago scale which deters elucidation of greater impacts.

While Pacific archaeologists have often assessed changes in individual marine resources, or assemblages of resources, there has been less attention paid to how these collective impacts might have fundamentally altered marine ecosystems through time (Allen 2003; Lambrides and Weisler 2016). However, researchers have begun to consider the broader ecological implications of human impacts in the region (Aswani and Allen 2009; Jones 2009; Morrison and Hunt 2007). Future archaeomalacology should ideally include fine-mesh screening, considering all taxa for identification, rather than a suite of large-bodied, economically important taxa. Szabó (2009:186)

31 discussed the utility of identifying taxa which may have been incidentally collected for understanding how humans interacted with the marine environment and also the risks of biasing archaeological inferences regarding gathering strategies when all taxa are not considered for identification. Quantification should be by NISP, MNI, and weight, with quantification protocols explicitly reported. Land snails still remain relatively neglected, 20 years on from Kirch and Weisler’s (1994) assessment of the lack of studies dealing with these fauna (but see Christensen and Weisler 2013).

Researchers should collaborate with natural scientists for integrating the collection of environmental data into archaeological research programs in order to generate data of wide applicability. In order to confidently identify negative impacts due to human foraging researchers must be able to evaluate the influence of climatic or environmental factors (Thakar et al. in press). The ecology and biology of taxa in the archaeological assemblage must be considered, and the complexities of an organism’s life history and local ecology must be critically evaluated before assigning human impact as the primary mechanism for change. Metric analysis should only be conducted on specimens identified to species level due to significant interspecific variation in morphology for some Indo-Pacific mollusks, for example, Cellana limpets in Hawai‘i (Bird 2011; Rogers 2015; Vermeij 1993, 2002). Schwerdtner Máñez et al. (2014:1-2) identified several key areas for a global marine historic agenda: establishing the state of the oceans before human exploitation, determining the key drivers of environmental change, the significance of marine resources for human societies over time, the circumstances that have encouraged societies to exploit or cease exploitation of the ocean, and the role of historical data in ocean governance and management (see also Lambrides and Weisler 2016 for relevance to archaeoicthyological studies). Archaeomalacology in the Pacific Islands is uniquely placed to provide historic data to engage these global research agendas in the future.

Conclusion Archaeomalacology in the Pacific Islands has provided historical data contributing to our understanding of environmental, economic, and cultural processes in prehistory. Past researchers developed a broad foundation of analytical techniques, methodological approaches and analyzed datasets, and current pursuits are expanding archaeological discourses in the region, especially those relating to historical ecology. The increasingly consistent reporting of quantification protocols, using NISP, weight, and MNI, and the incorporation of fine-mesh screening into recovery methods is facilitating new, comparative approaches providing novel data pertaining to the processes and

32 timing of human impacts to coastal marine ecosystems of the Pacific Islands, and thus enhancing our understanding of long-term trends in human interactions with the ocean.

Acknowledgements We appreciate the thoughtful comments from the two anonymous reviewers. Harris’s university studies are supported by an Australian Government Research Training Program Scholarship.

References Cited Alden Smith, E. and M. Wishnie. 2000. Conservation and subsistence in small- scale societies Annual Review of Anthropology 29(1): 493-524. Allen, M.S. 1992. Dynamic Landscapes and Human Subsistence: Archaeological Investigations on Aitutaki Islands, Southern Cook Islands. Ph.D. Dissertation. Washington: University of Washington. ---. 2003. Human Impact on Pacific Nearshore Marine Ecosystems. In C. Sand Pacific Archaeology: Assessments and Prospects. Proceedings of the International Conference for the 50th Anniversary of the First Lapita Excavation (C. Sand, ed.):317-325. Les Cahiers de L'archéologie en Nouvelle-Calédonie. Nouméa: Département Archéologie, Service des Musées et du Patrimoine. ---. 2012. Molluscan foraging efficiency and patterns of mobility amongst foraging agriculturalists: A case study from northern New Zealand. Journal of Archaeological Science 39(2):295- 307. Alleway, H.K. and S.D. Connell. 2015. Loss of an ecological baseline through the eradication of oyster reefs from coastal ecosystems and human memory. Conservation Biology 29(3): 795- 804. Álvarez, M., I. Briz Godino, A. Balbo and M. Madella. 2011. Shell middens as archives of past environments, human dispersal and specialized resource management. Quaternary International 239(1–2):1-7. Amesbury, J.R. 1999. Changes in species composition of archaeological marine shell assemblages in Guam. Micronesica 31(2):347-366. ---. 2007. Mollusk collecting and environmental change during the prehistoric period in the Mariana Islands. Coral Reefs 26(4):947-958. Anderson, A.J. 1981. A model of prehistoric collecting on the rocky shore. Journal of Archaeological Science 8(2):109-120. Andréfouët, S., A. Gilbert, L. Yan, G. Remoissenet, C. Payri and Y. Chancerelle. 2005. The remarkable population size of the endangered clam Tridacna maxima assessed in Fangatau Atoll (eastern Tuamotu, French Polynesia) using in situ and remote sensing data. ICES Journal of Marine Science: Journal du Conseil 62(6):1037-1048. Aswani, S. and M.S. Allen. 2009. A Marquesan coral reef (French Polynesia) in historical context: An integrated socio-ecological approach. Aquatic Conservation: Marine and Freshwater Ecosystems 19(6):614-625. 34

Aswani, S., C.F. Flores and B.R. Broitman. 2015. Human harvesting impacts on managed areas: ecological effects of socially-compatible shellfish reserves. Reviews in Fish Biology and Fisheries 25(1):1-14. Baisre, J.A. 2010. Setting a baseline for Caribbean fisheries. The Journal of Island and Coastal Archaeology 5(2):120-147. Balée, W. 2006. The research program of historical ecology. Annual Review of Anthropology 35:75- 98. Beauchamp, K.A. and M.M. Gowing. 1982. A quantitative assessment of human trampling effects on a rocky intertidal community. Marine Environmental Research 7(4):279-293. Best, S.B. 1985. Lakeba: The Prehistory of a Fijian Island. Ph.D. Dissertation. Auckland: University of Auckland. Bird, C.E. 2011. Morphological and behavioral evidence for adaptive diversification of sympatric Hawaiian limpets (Cellana spp.). Integrative and Comparative Biology 51(3):466-473. Bird, D.W. and R.L. Bliege Bird. 1997. Contemporary shellfish gathering strategies among the Meriam of the Torres Strait Islands, Australia: Testing predictions of a central place foraging model. Journal of Archaeological Science 24(1):39-63. ---. 2000. The ethnoarchaeology of juvenile foragers: Shellfishing strategies among Meriam children. Journal of Anthropological Archaeology 19(4):461-476. Blick, J.P. 2007. Pre-Columbian impact on terrestrial, intertidal, and marine resources, San Salvador, Bahamas (A.D. 950–1500). Journal for Nature Conservation 15(3):174-183. Blumenstock, D.I. 1958. Typhoon effects at Jaluit Atoll in the Marshall Islands. Nature 182(4645):1267-1269. Braje, T.J., J.M. Erlandson, D.J. Kennett, T.C. Rick and J.E. Peterson. 2005. Deep history: using archaeology and historical ecology to promote marine conservation. Alternate Routes 21:5- 17. Braje, T.J., D.J. Kennett, J.M. Erlandson and B.J. Culleton. 2007. Human impacts on nearshore shellfish taxa: A 7,000 year record from Santa Rosa Island, California. American Antiquity 72(4):735-756. Branch, G. and F. Odendaal. 2003. The effects of marine protected areas on the population dynamics of a South African limpet, Cymbula oculus, relative to the influence of wave action. Biological Conservation 114(2):255-269. Brander, K., L. Botsford, L. Ciannelli, M. Fogarty, M. Heath, B. Planque, L. Shannon and K. Wieland. 2010. Human Impacts on Marine Ecosystems. In M. Barange Marine Ecosystems and Global Change (M. Barange, ed.):41-71. Oxford: Oxford University Press.

Briggs, J.M., K.A. Spielmann, H. Schaafsma, K.W. Kintigh, M. Kruse, K. Morehouse and K. Schollmeyer. 2006. Why ecology needs archaeologists and archaeology needs ecologists. Frontiers in Ecology and the Environment 4(4):180-188. Broughton, J.M. 1994. Late Holocene resource intensification in the Sacramento Valley, California: the vertebrate evidence. Journal of Archaeological Science 21(4):501-514. ---. 1997. Widening diet breadth, declining foraging efficiency, and prehistoric harvest pressure: ichthyofaunal evidence from the Emeryville Shellmound, California. Antiquity 71(274):845- 862. Bunce, M., L.D. Rodwell, R. Gibb and L. Mee. 2008. Shifting baselines in fishers' perceptions of island reef fishery degradation. Ocean & Coastal Management 51(4):285-302. Byers, D.A., J.M. Broughton. 2004. Holocene environmental change, artiodactyl abundances, and human hunting strategies in the Great Basin. American Anqitquity 69(2):235-255 Campbell, G. 2015. “We want to go, where everyone knows, mussels are all the same…”: A comment on some recent zooarchaeological research on Mytilus californianus size prediction. Journal of Archaeological Science 63:156-159. Carder, N. and J.G. Crock. 2012. A Pre-Columbian fisheries baseline from the Caribbean. Journal of Archaeological Science 39(10):3115-3124. Castilla, J.C. 1999. Coastal marine communities: trends and perspectives from human-exclusion experiments. Trends in Ecology & Evolution 14(7):280-283. Castilla, J.C. and R.H. Bustamante. 1989. Human exclusion from rocky intertidal of Las Cruces, Central Chile: Effects on Durvillaea antarctica (Phaeophyta, Durvilleales). Marine Ecology Progress Series. Oldendorf 50(3):203-214. Castilla, J.C. and L. Duran. 1985. Human exclusion from the rocky intertidal zone of central Chile: The effects on Concholepas concholepas (Gastropoda). Oikos 45(3):391-399. Catterall, C.P. and I.R. Poiner. 1987. The potential impact of human gathering on shellfish populations, with reference to some NE Australian intertidal flats. Oikos 50(1):114-122. Christensen, C.C. and M.I. Weisler. 2013. Land snails from archaeological sites in the Marshall Islands, with remarks on prehistoric translocations in tropical Oceania. Pacific Science 67(1):81-104. Clark, G. and C. Reepmeyer. 2012. Last millennium climate change in the occupation and abandonment of Palau's Rock Islands. Archaeology in Oceania 47(1):29-38. Codding, B.F., D.W. Bird and R.L. Bliege Bird. 2010. Interpreting abundance indices: Some zooarchaeological implications of Martu foraging. Journal of Archaeological Science 37(12):3200-3210.

Copland, J.W. and J.S. Lucas. 1988. Giant Clams in Asia and the Pacific. Canberra: Australian Centre for International Agricultural Research. Cramer, K.L., J.S. Leonard-Pingel, F. Rodríguez and J.B.C. Jackson. 2015. Molluscan subfossil assemblages reveal the long-term deterioration of coral reef environments in Caribbean Panama. Marine Pollution Bulletin 96(1-2):176-187 Dalzell, P. 1998. The role of archaeological and cultural-historical records in long-range coastal fisheries resources management strategies and policies in the Pacific Islands. Ocean & Coastal Management 40(2):237-252. de Boer, W.F. and F.A. Longamane. 1996. The exploitation of intertidal food resources in Inhaca Bay, Mozambique, by shorebirds and humans. Biological Conservation 78(3):295-303. de Boer, W.F., T. Pereira and A. Guissamulo. 2000. Comparing recent and abandoned shell middens to detect the impact of human exploitation on the intertidal ecosystem. Aquatic Ecology 34(3):287-297. de Boer, W.F. and H.H.T. Prins. 2002. Human exploitation and benthic community structure on a tropical intertidal flat. Journal of Sea Research 48(3):225-240. Dickinson, W.R. 2003. Impact of mid-Holocene hydro-isostatic highstand in regional sea level on habitability of islands in Pacific Oceania. Journal of Coastal Research 19(3):489-502. Dulvy, N.K., Y. Sadovy and J.D. Reynolds. 2003. Extinction vulnerability in marine populations. Fish and Fisheries 4(1):25-64. Edgar, G.J. and C.R. Samson. 2004. Catastrophic decline in mollusc diversity in eastern Tasmania and its concurrence with shellfish fisheries. Conservation Biology 18(6):1579-1588. Erlandson, J.M. and T.J. Braje. 2013. Archeology and the Anthropocene. Anthropocene 4:1-7. Erlandson, J.M. and S.M. Fitzpatrick. 2006. Oceans, islands, and coasts: current perspectives on the role of the sea in human prehistory. The Journal of Island and Coastal Archaeology 1(1):5- 32. Erlandson, J.M. and T.C. Rick. 2010. Archaeology meets marine ecology: the antiquity of maritime cultures and human impacts on marine fisheries and ecosystems. Annual Review of Marine Science 2:231-251. Fairweather, P.G. 1988. Consequences of supply-side ecology: manipulating the recruitment of intertidal barnacles affects the intensity of predation upon them. The Biological Bulletin 175(3):349-354. Fairweather, P.G. 1990. Is predation capable of interacting with other community processes on rocky reefs? Australian Journal of Ecology 15(4):453-464.

Fairweather, P.G. and A.J. Underwood. 1991. Experimental removals of a rocky intertidal predator: variations within two habitats in the effects on prey. Journal of Experimental Marine Biology and Ecology 154(1):29-75. Faulkner, P. 2010. Morphometric and taphonomic analysis of granular ark (Anadara granosa) dominated shell deposits of Blue Mud Bay, northern Australia. Journal of Archaeological Science 37(8):1942-1952. Fenberg, P.B. and R. Kautstuv. 2008. Ecological and evolutionary consequences of size-selective harvesting: how much do we know? Molecular Ecology 17(1):209-220. Fitzpatrick, S.M. and R.T. Callaghan. 2013. Estimating trajectories of colonisation to the Mariana Islands, western Pacific. Antiquity 87(337):840-853. Fitzpatrick, S.M. and T.J. Donaldson. 2007. Anthropogenic impacts to coral reefs in Palau, western Micronesia during the late Holocene. Coral Reefs 26(4):915-930. Fitzpatrick, S.M., C.M. Giovas and O. Kataoka. 2011. Temporal trends in prehistoric fishing in Palau, Micronesia over the last 1500 years. Archaeology in Oceania 46(1):6-16. Fitzpatrick, S.M. and V. Kanai. 2001. An applied approach to archeology in Palau. Cultural Resource Managment 24(1):41-43. Fitzpatrick, S.M. and O. Kataoka. 2005. Prehistoric fishing in Palau, Micronesia: Evidence from the northern Rock Islands. Archaeology in Oceania 40(1):1-13. Gilbert, G.N. 2008. Agent-Based Models. Los Angeles: Sage Publications. Giovas, C.M. 2009. The shell game: analytic problems in archaeological mollusc quantification. Journal of Archaeological Science 36(7):1557-1564. Giovas, C.M., S.M. Fitzpatrick, M. Clark and M. Abed. 2010. Evidence for size increase in an exploited mollusc: humped conch (Strombus gibberulus) at Chelechol ra Orrak, Palau from ca. 3000-0 BP. Journal of Archaeological Science 37(11):2788-2798. Glynn, P.W. and I.C. Enochs. 2011. Invertebrates and Their Roles in Coral Reef Ecosystems. In Coral reefs: An Ecosystem in Transition (Z. Dubinsky and N. Stambler, eds.):273-325. Dordrecht: Springer Netherlands. Godoy, C. and C.A. Moreno. 1989. Indirect effects of human exclusion from the rocky intertidal in southern Chile: a case of cross-linkage between herbivores. Oikos 54(1):101-106. Green, A. and P. Craig. 1999. Population size and structure of giant clams at Rose Atoll, an important refuge in the Samoan archipelago. Coral Reefs 18(3):205-211. Green, R. 1986. Lapita fishing: The evidence of site SE-RF-2 from the main Reef Islands, Santa Cruz group, Solomons. In Pacific Anthropological Records, 37: Traditional Fishing in the

Pacific (A.J. Anderson, ed.):119-135. Honolulu: Department of Anthropology, Bernice Pauahi Bishop Museum. Groube, L.M. 1971. Tonga, Lapita pottery and Polynesian origins. The Journal of the Polynesian Society 80(3):278-316. Gutiérrez, J.L., C.G. Jones, D.L. Strayer and O.O. Iribarne. 2003. Mollusks as ecosystem engineers: the role of shell production in aquatic habitats. Oikos 101(1):79-90. Hancock, D.Α. and A.C. Simpson. 1962. Parameters of Marine Invertebrate Populations. In The Exploitation of Natural Animal Populations (E.D. Le Cren and Η.W. Holgate, eds.):29-50. Oxford: Blackwell. Harris, M. and M.I. Weisler. 2016. Intertidal foraging on atolls: prehistoric forager decision making at Ebon Atoll, Marshall Islands. Journal of Island & Coastal Archaeology 10.1080/15564894.2016.1167140 Harris, M., M.I. Weisler and P. Faulkner. 2015. A refined protocol for calculating MNI in archaeological molluscan shell assemblages: a Marshall Islands case study. Journal of Archaeological Science 57:168-179. Hayashida, F.M. 2005. Archaeology, ecological history, and conservation. Annual Review of Anthropology 34:43-65. Hockey, P.A.R. 1987. The influence of coastal utilisation by man on the presumed extinction of the Canarian Black Oystercatcher Haematopus meadewaldoi Bannerman. Biological Conservation 39(1):49-62. Hockey, P.A.R. and L.B. Alison. 1986. Man as an intertidal predator in Transkei: disturbance, community convergence and management of a natural food resource. Oikos 46(1):3-14. Hughes, T.P., M.J. Rodrigues, D.R. Bellwood, D. Ceccarelli, O. Hoegh-Guldberg, L. McCook, N. Moltschaniwskyj, M.S. Pratchett, R.S. Steneck and B. Willis. 2007. Phase shifts, herbivory, and the resilience of coral reefs to climate change. Current Biology 17(4):360-365. Hutchings, P., S. Ahyong, M. Byrne, R. Przeslawski, G. Wörheide and Great Barrier Reef Marine Park Authority. 2007. Chapter 11: Vulnerability of Benthic Invertebrates of the Great Barrier Reef to Climate Change. In Climate Change and the Great Barrier Reef: A Vulnerability Assessment (J.E. Johnson and P.A. Marshall, eds.):309-356. Townsville: The Great Barrier Reef Marine Park Authorty. Irwin, G. 1992. The Prehistoric Exploration and Colonisation of the Pacific Cambridge: Cambridge University Press. Jackson, J.B.C., M.X. Kirby, W.H. Berger, K.A. Bjorndal, L.W. Botsford, B.J. Bourque, R.H. Bradbury, R. Cooke, J.M. Erlandson, J.A. Estes, T.P. Hughes, S. Kidwell, C.B. Lange, H.S.

Lenihan, J.M. Pandolfi, C.H. Peterson, R.S. Steneck, M.J. Tegner and R.R. Warner. 2001. Historical overfishing and the recent collapse of coastal ecosystems. Science 293(5530):629- 638. Jacomb, C., R. Walter and E. Brooks. 2010. Living on pipi (Paphies australis): Specialised shellfish harvest in a marginal environment at Karamea, west coast, New Zealand. Journal of Pacific Archaeology 1(1):36-52. Jerardino, A. and R. Navarro. 2008. Shell morphometry of seven limpet species from coastal shell middens in southern Africa. Journal of Archaeological Science 35(4):1023-1029. Jimenez, H., P. Dumas, D. Ponton and J. Ferraris. 2012. Predicting invertebrate assemblage composition from harvesting pressure and environmental characteristics on tropical reef flats. Coral Reefs 31(1):89-100. Jones, C.G., J.H. Lawton and M. Shachak. 1996. Organisms as Ecosystem Engineers. In Ecosystem Management: Selected Readings (F.B. Samson and F.L. Knopf, eds.):130-147. New York: Springer New York. Jones, S. 2007. Human impacts on ancient marine environments of Fiji's Lau Group: current ethnoarchaeological and archaeological research. The Journal of Island and Coastal Archaeology 2(2):239-244. ---. 2009. A long-term perspective on biodiversity and marine resource exploitation in Fiji's Lau Group. Pacific Science 63(4):617-648. Jones, S. and U. Albarella. 2009. Contemporary Subsistence and Foodways in the Lau Islands Fiji: an Ethnoarchaeological Study of Non-Optimal Foraging and Irrational Economics. In Ethnozooarchaeology: The Present Past of Human-Animal Relationships (U. Albarella and A. Trentacoste, eds.):73-81. Oxford: Oxbow Books. Jones, S. and R.L. Quinn. 2009. Prehistoric Fijian diet and subsistence: integration of faunal, ethnographic, and stable isotopic evidence from the Lau Island Group. Journal of Archaeological Science 36(12):2742-2754. Kataoka, O. 1996. Prehistoric and Historic Faunal Utilization in Pohnpei: An Ecological and Ethnoarchaeological Understanding. Ph.D. Dissertation. Eugene: University of Oregon. Kinch, J. 2003. Marine mollusc use among the women of Brooker Island, Louisiage Archipelago, Papua New Guinea. SPC Women in Fisheries Information Bulletin 13:5-14. Kirch, P.V. 1983. Man's role in modifying tropical and subtropical Polynesian ecosystems. Archaeology in Oceania 18(1):26-31. ---. 2000. On the Road of the Winds : An Archaeological History of the Pacific Islands Before European Contact. Berkeley: University of California Press.

Kirch, P.V., D.W. Steadman, V.L. Butler, J. Hather and M.I. Weisler. 1995. Prehistory and human ecology in eastern Polynesia: excavations at Tangatatau Rockshelter, Mangaia, Cook Islands. Archaeology in Oceania 30(2):47-65. Kirch, P.V. and M.I. Weisler. 1994. Archaeology in the Pacific Islands: an appraisal of recent research. Journal of Archaeological Research 2(4):285-328. Kirch, P.V. and D.E. Yen. 1982. Tikopia: The prehistory and ecology of a Polynesian outlier. Bernice Pauahi Bishop Museum Bulletin 238. Kittinger, J.N., J.M. Pandolfi, J.H. Blodgett, T.L. Hunt, H. Jiang, K. Maly, L.E. McClenachan, J.K. Schultz and B.A. Wilcox. 2011. Historical reconstruction reveals recovery in Hawaiian coral reefs. PLoS One 6(10):1-14. Klumpp, D.W. and J.J. Lucas. 1994. Nutritional ecology of the giant clams, Tridacna tevoroa and T. derasa from Tonga: influence of light on filter-feeding and photosynthesis. Marine Ecology Progress Series 107:147-156. Klumpp, D.W. and A. Pulfrich. 1989. Trophic significance of herbivorous macroinvertebrates on the central Great Barrier Reef. Coral Reefs 8(3):135-144. Kobluk, D.R. and M.A. Lysenko. 1993. Hurricane effects on shallow-water cryptic reef molluscs, Fiji Islands. Journal of Paleontology 67(5):798-816. Kohn, A.J. 1980. Populations of tropical intertidal gastropods before and after a typhoon. Micronesica 16(2):215-228. ---. 1987. Intertidal Ecology of Enewetak Atoll. In The Natural History of Enewetak Atoll: Volume I the Ecosystem: Environments, Biotas, and Processes (D.M. Devaney, E.S. Reese, B.L. Burch and P. Helfrich, eds.):139-157. Oak Ridge: Office of Scientific and Technical Information, U.S. Department of Energy. Lambrides, A.B.J. and M.I. Weisler. 2016. Pacific Islands ichthyoarchaeology: implications for the development of prehistoric fishing studies and global sustainability. Journal of Archaeological Research 24(3):1-50. Lauer, M. and S. Aswani. 2010. Indigenous knowledge and long-term ecological change: detection, interpretation, and responses to changing ecological conditions in Pacific Island communities. Environmental management 45(5):985-997. Leviten, P.J. and A.J. Kohn. 1980. Microhabitat resource use, activity patterns, and episodic catastrophe: Conus on tropical intertidal reef rock benches. Ecological Monographs 50(1):55-75. Lewis, A.D., T.J.H. Adams and E. Ledua. 1988. Fiji's Giant Clam Stocks - a Review of Their Distribution, Abundance, Exploitation and Managment. In Giant clams in Asia and the

Pacific (J.W. Copland and J.S. Lucas, eds.):66-72. Canberra: Australian Centre for International Agricultural Research. Mannino, M.A. and K.D. Thomas. 2002. Depletion of a resource? The impact of prehistoric human foraging on intertidal mollusc communities and its significance for human settlement, mobility and dispersal. World Archaeology 33(3):452-474. Mason, R.D., M.L. Peterson and J.A. Tiffany. 1998. Weighing vs. counting: measurement reliability and the California school of midden analysis. American Antiquity 63(2):303-324. Masse, W.B., J. Liston, J. Carucci and J.S. Athens. 2006. Evaluating the effects of climate change on environment, resource depletion, and culture in the Palau Islands between AD 1200 and 1600. Quaternary International 151(1):106-132. Matthews, E., J. Kinch, P. Lowrey, I. Novaczek, S. Pinca, S. Owen, J. Veitayaki and I. Noaczek. 2003. Nearshore invertebrates decline as coastal development increases around Palau. SPC Women in fisheries information bulletin 13:2-5. McClanahan, T.R., N.A.J. Graham and E.S. Darling. 2014. Coral reefs in a crystal ball: predicting the future from the vulnerability of corals and reef fishes to multiple stressors. Current Opinion in Environmental Sustainability 7:59-64. McKoy, J. 1980. Biology, exploitation and management of giant clams (Tridacnidae) in the kingdom of Tonga. . Fisheries Bulletin, Tonga 1(61):541-558. McShane, P.E., D.R. Schiel, S.F. Mercer and T. Murray. 1994. Morphometric variation in Haliotis iris (Mollusca: Gastropoda): analysis of 61 populations. New Zealand Journal of Marine and Freshwater Research 28(4):357-364. Moreno, C.A., J.P. Sutherland and H.F. Jara. 1984. Man as a predator in the intertidal zone of southern Chile. Oikos 42(2):155-160. Morrison, A.E. and D.J. Addison. 2008. Assessing the role of climate change and human predation on marine resources at the Fatu-Ma-Futi site, Tutuila Island, American Samoa: an agent based model. Archaeology in Oceania 42(4):22-34. Morrison, A.E. and M.S. Allen. In Press. Agent-based modelling, molluscan population dynamics, and archaeomalacology. Quaternary International http://dx.doi.org/10.1016/j.quaint.2015.09.004. Morrison, A.E. and E.E. Cochrane. 2008. Investigating shellfish deposition and landscape history at the Natia Beach site, Fiji. Journal of Archaeological Science 35(8):2387-2399. Morrison, A.E. and T.L. Hunt. 2007. Human impacts on the nearshore environment: an archaeological case study from Kauaʻi, Hawaiian Islands. Pacific Science 61(3):325- 328,331-345.

Myers, R.A. and B. Worm. 2003. Rapid worldwide depletion of predatory fish communities. Nature 423(6937):280-283. Newman, W. and E. Gomez. 2002. On the Status of Giant Clams, Relics of Tethys (Mollusca: Bivalvia: Tridacninae). In Proceedings of the Ninth International Coral Reef Symposium, Bali, 23-27 October 2000, Vol. 2 (M.K. Moosa, S. Soemodihardjo, A. Soegiarto, K. Romimohtarto, A. Nontji, Soekarno and Suharsono, eds.):927-936. Bali: International Society for Reef Studies. Nichol, R.K. 1986. Analysis of Midden From N44/215: Hard Times at Hahei? In Pacific Anthropological Records, 37: Traditional Fishing in the Pacific (A.J. Anderson, ed.):179- 198. Honolulu: Department of Anthropology, Bernice Pauahi Bishop Museum. Nunn, P.D. 1990. Recent environmental changes on Pacific Islands. Geographical Journal 156(2):125-140. Nunn, P.D., T. Ishimura, W.R. Dickinson and K. Katayama. 2007. The Lapita occupation at Naitabale, Moturiki Island, central Fiji. Asian Perspectives 46(1):96-132. Odum, H.T. and E.P. Odum. 1955. Trophic structure and productivity of a windward coral reef community on Eniwetok Atoll. Ecological Monographs 25(3):291-320. Paine, R.T. 1995. A conversation on refining the concept of keystone species. Conservation Biology 9(4):962-964. Paulay, G. 1990. Effects of late cenozoic sea-level fluctuations on the bivalve faunas of tropical Oceanic islands. Paleobiology 16(4):415-434. Paulay, G. 1996. Dynamic clams: changes in the bivalve fauna of Pacific Islands as a result of sea level fluctuations. American Malacological Bulletin 12(1-2):45-57. Pauly, D. 1995. Anecdotes and the shifting baseline syndrome of fisheries. Trends in Ecology & Evolution 10(10):430. Pestle, W.J. 2013. Fishing down a prehistoric Caribbean marine food web: isotopic evidence from Punta Candelero, Puerto Rico. The Journal of Island and Coastal Archaeology 8(2):228- 254. Petes, L.E., B.A. Menge and G.D. Murphy. 2007. Environmental stress decreases survival, growth, and reproduction in New Zealand mussels. Journal of Experimental Marine Biology and Ecology 351(1–2):83-91. Pinnegar, J.K. and G.H. Engelhard. 2008. The ‘shifting baseline’phenomenon: a global perspective. Reviews in Fish Biology and Fisheries 18(1):1-16. Planes, S., C. Chauvet, J. Baldwin, J. Bonvallot, Y. Fontaine-Vernaudon, C. Gabrié, P. Holthus, C. Payri and R. Galzin. 1993. Impact of tourism-related fishing on Tridacna maxima

(Mollusca, Bivalvia) stocks in Bora-Bora Lagoon (French Polynesia). Atoll Research Bulletin 385:1-13. Poiner, I.R. and C.P. Catterall. 1988. The effects of traditional gathering on populations of the marine gastropod Strombus luhuanus Linne 1758, in southern Papua New Guinea. Oecologia 76(2):191-199. Power, M.E., D. Tilman, J.A. Estes, B.A. Menge, W.J. Bond, L.S. Mills, G. Daily, J.C. Castilla, J. Lubchenco and R.T. Paine. 1996. Challenges in the quest for keystones. BioScience 46(8):609-620. Railsback, S.F. and V. Grimm. 2011. Agent-based and individual-based modeling: a practical introduction. Princeton: Princeton University Press Richerson, J. and R. Boyd. 1987. Simple Models of Complex Phenomena: The Case of Cultural Evolution. In The Latest on the Best: Essays on Evolution and Optimality (Dupré, ed.):27- 52. Cambridge: The MIT Press Rick, T.C. and J.M. Erlandson. 2008. Human Impacts on Ancient Marine Ecosystems: A Global Perspective. Berkeley: University of California Press. Rick, T.C. and R. Lockwood. 2013. Integrating paleobiology, archeology, and history to inform biological conservation. Conservation Biology 27(1):45-54. Rivadeneira, M.M., C.M. Santoro and P.A. Marquet. 2010. Reconstructing the history of human impacts on coastal biodiversity in Chile: constraints and opportunities. Aquatic Conservation: Marine and Freshwater Ecosystems 20(1):74-82. Rogers, A. 2015. Assessing the Efficacy of Genus-Level Data: An Archaeological Case Study of the Hawaiian Limpet (Cellana spp.), Moloka‘i, Hawaiian islands. Honours thesis. Brisbane: University of Queensland. Rolett, B.V. 1992. Faunal extinctions and depletions linked with prehistory and environmental change in the Marquesas Islands (French Polynesia). The Journal of the Polynesian Society 101(1):86-94. Sagarin, R.D., R.F. Ambrose, B.J. Becker, J.M. Engle, J. Kido, S.F. Lee, C.M. Miner, S.N. Murray, P.T. Raimondi and D. Richards. 2007. Ecological impacts on the limpet Lottia Gigantea populations: human pressure over a broad scale on island and mainland intertidal zones. Marine Biology 150(3):399-413. Sale, P.F. 1980. Assemblages of fish on patch reefs — predictable or unpredictable? Environmental Biology of Fishes 5(3):243-249.

Schwerdtner Máñez, K., P. Holm, L. Blight, M. Coll, A. MacDiarmid, H. Ojaveer, B. Poulsen and M. Tull. 2014. The future of the oceans past: towards a global marine historical research initiative. PLoS ONE 9(7):e101466. Seeto, J., P.D. Nunn and S. Sanjana. 2012. Human-mediated prehistoric marine extinction in the Tropical Pacific? Understanding the presence of Hippopus hippopus (Linn. 1758) in ancient shell middens on the Rove Peninsula, southwest Viti Levu Island, Fiji. Geoarchaeology 27(1):2-17. Singh, G.G., I. McKechnie, T.J. Braje and B. Campbell. 2015. “All models are wrong but some are useful”: a response to Campbell's comment on estimating Mytilus californianus shell size. Journal of Archaeological Science 63:160-163. Small, C. and R.J. Nicholls. 2003. A global analysis of human settlement in coastal zones. Journal of Coastal Research 19(3):584-599. Smith, B.D. and M.A. Zeder. 2013. The onset of the Anthropocene. Anthropocene 4:8-13. Spennemann, D.H.R. 1987. Availability of shellfish resources on prehistoric Tongatapu, Tonga: effects of human predation and changing environment. Archaeology in Oceania 22(3):81- 96. ---. 1989. Effects of Human predation and changing environment on some mollusc species on Tongatapu, Tonga. In The Walking Larder : Patterns of Domestication, Pastoralism, and Predation (J. Clutton-Brock, ed.):326-335. Sydney: Unwin Hyman. Stier, A.C. and M. Leray. 2014. Predators alter community organization of coral reef cryptofauna and reduce abundance of coral mutualists. Coral Reefs 33(1):181-191. Swadling, P. 1976. Changes induced by human exploitation in prehistoric shellfish populations. Mankind 10(3):156–162. ---. 1977. The implications of shellfish exploitation for New Zealand prehistory. Mankind 11(1):11- 18. ---. 1986. Lapita shellfishing: evidence from sites in the Reef/Santa Cruz Group, southeast Solomons. In Pacific Anthropological Records, 37: Traditional Fishing in the Pacific (A.J. Anderson, ed.):137-148. Honolulu: Department of Anthropology, Bernice Pauahi Bishop Museum. Szabó, K. 2001. The reef, the beach and the rocks: an environmental analysis of mollusc remains from Natunuku, Viti Levu, Fiji. In Terra Australis 17: The Archaeology of Lapita Dispersal in Oceania (G.R. Clark, A.J. Anderson and T. Vunidilo, eds.):159-166. Canberra: ANU Press.

---. 2009. Molluscan remains from Fiji. In Terra australis 31: The Early Prehistory of Fiji (G. Clark and A.J. Anderson, eds.):183-212. Canberra: ANU E Press. Szabó, K. and J.R. Amesbury. 2011. Molluscs in a world of islands: The use of shellfish as a food resource in the Tropical Island Asia-Pacific region. Quaternary International 239(1–2):8- 18. Tebano, T. and G. Paulay. 2000. Variable recruitment and changing environments create a fluctuating resource: The biology of Anadara uropigimelana (Bivalvia: Arcidae) on Tarawa Atoll. Atoll Research Bulletin 488:1-15. Thakar, H.B., M.A. Glassow and C. Blanchette. in press. Reconsidering evidence of human impacts: implications of within-site variation of growth rates in Mytilus californianus along tidal gradients. Quaternary International http://dx.doi.org/10.1016/j.quaint.2015.10.018. Thangavelu, A., B. David, B. Barker, J.-M. Geneste, J.-J. Delannoy, L. Lamb, N. Araho and R. Skelly. 2011. Morphometric analyses of Batissa violacea shells from Emo (OAC), Gulf Province, Papua New Guinea. Archaeology in Oceania 46(2):67-75. Thomas, F.R. 2007a. The behavioral ecology of shellfish gathering in western Kiribati, Micronesia 1: prey choice. Human Ecology 35(2):179-194. ---. 2007b. The behavioral ecology of shellfish gathering in western Kiribati, Micronesia. 2: patch choice, patch sampling, and risk. Human Ecology 35(5):515-526. ---. 2009. Historical ecology in Kiribati: linking past with present. Pacific Science 63(4):567-600. Thomas, F.R., P.D. Nunn, T. Osborne, R. Kumar, F. Areki, S. Matararaba, D. Steadman and G. Hope. 2004. Recent archaeological findings at Qaranilaca Cave, Vanuabalavu Island, Fiji. Archaeology in Oceania 39(1):42-49. Tyler, C.L., E. Stafford and L. Leighton. 2014. The utility of wax replicas as a measure of crab attack frequency in the rocky intertidal. Journal of the Marine Biological Association of the United Kingdom 95(2):361-369. Vermeij, G.J. 1993. A Natural History of Shells. Princeton: Princeton University Press. ---. 2002. Characters in context: molluscan shells and the forces that mold them. Paleobiology 28(1):41-54. Wake, T.A., D.R. Doughty and M. Kay. 2013. Archaeological investigations provide late Holocene baseline ecological data for Bocas del Toro, Panama. Bulletin of Marine Science 89(4):1015-1035. Walter, R. 1998. Anai'o: The Archaeology of a Fourteenth Century Polynesian Community in the Cook Islands. Dunedin: New Zealand Archaeological Association.

Webb, A.P. and P.S. Kench. 2010. The dynamic response of reef islands to sea-level rise: evidence from multi-decadal analysis of island change in the central Pacific. Global and Planetary Change 72(3):234-246. Weisler, M.I. 1995. Henderson island prehistory: colonization and extinction on a remote Polynesian island. Biological Journal of the Linnean Society 56(1-2):377-404. ---. 1999. Atolls as settlement landscapes: Ujae, Marshall Islands. Atoll Research Bulletin 460:1-51. ---. 2001. On the Margins of Sustainability: Prehistoric Settlement of Utrōk Atoll, Northern Marshall Islands. Oxford: Archaeopress. Weisler, M.I., R. Bolhar, J. Ma, E. St Pierre, P. Sheppard, R.K. Walter, Y. Feng, J.-x. Zhao and P.V. Kirch. 2016. Cook island artifact geochemistry demonstrates spatial and temporal extent of pre-European interarchipelago voyaging in east Polynesia. Proceedings of the National Academy of Sciences 113(29):8150-8155. Weisler, M.I. and P.V. Kirch. 1985. The structure of settlement space in a Polynesian chiefdom: Kawela, Molokaʻi, Hawaiian Islands. New Zealand Journal of Archaeology 7:129-158. Whitaker, A.R. 2008. Incipient aquaculture in prehistoric California?: long-term productivity and sustainability vs. immediate returns for the harvest of marine invertebrates. Journal of Archaeological Science 35(4):1114-1123. Wolverton, S. and R.L. Lyman. 2012. Conservation Biology and Applied Zooarchaeology. Tucson: University of Arizona Press. Woodroffe, C.D. 2008. Reef-island topography and the vulnerability of atolls to sea-level rise. Global and Planetary Change 62(1–2):77-96. Yamaguchi, M. 1975. Sea level fluctuations and mass mortalities of reef animals in Guam, Mariana Islands. Micronesica 11(2):227-243. Zhang, L.-P., J.-J. Xia, P.-F. Peng, H.-P. Li, P. Luo and C.-Q. Hu. 2013. Characterization of embryogenesis and early larval development in the pacific triton, Charonia tritonis (Gastropoda: Caenogastropoda). Invertebrate Reproduction & Development 57(3):237-246

Figures and tables

Figure 1 Map of the Pacific Islands with sites mentioned in text

Chapter 3: A refined protocol for calculating MNI in archaeological molluscan shell assemblages: a Marshall Islands case study

Note: This is the final version of a peer-reviewed article published in the Journal of Archaeological Science

Matthew Harris1, Marshall I. Weisler1 and Patrick Faulkner2 1. School of Social Science, The University of Queensland, St Lucia, Queensland, 4072, Australia 2. School of Philosophical and Historical Inquiry, Faculty of Arts and Social Sciences, Department of Archaeology, University of Sydney, Sydney, New South Wales, 2006, Australia

Abstract Comprehensive and transparent protocols for calculating Minimum Number of Individuals (MNI) for archaeological faunal assemblages are critical to data quality, comparability, and replicability. MNI values for archaeological molluscan assemblages are routinely calculated by counting a select range of Non-Repetitive Elements (NREs). Most commonly, only the frequency of the spire of gastropods and the umbo or hinge of bivalves are recorded. Calculating MNI based only on the frequency of these NREs can underestimate the relative abundance of particular molluscan shell forms. Using archaeological mollusc assemblages from two sites in the Marshall Islands as a case study, we outline a new protocol (tMNI) that incorporates a wider range of NRE and calculates MNI based on the most frequently occurring NRE for each taxon. The principles that underlie the tMNI method can be modified to be regionally or assemblage specific, rather than being a universally applicable range of NRE for the calculation of MNI. For the Marshall Islands assemblages, the inclusion of additional NRE in quantification measures led to (1) a 167% increase in relative abundance of gastropods and 3% increase in bivalves (2) changes to rank order abundance, and (3) alterations to measures of taxonomic richness and evenness. Given these results for the Marshall Islands assemblages, tMNI provides more accurate taxonomic abundance measures for these and other archaeological molluscan assemblages with similar taxa. These results have implications for the quality of zooarchaeological data increasingly utilised by conservation biologists, historical ecologists and policy makers.

Keywords: Mollusc quantification; Minimum Numbers of Individuals (MNI); Shell middens; Marshall Islands; Analytical transparency

Introduction Quantification is fundamental to zooarchaeological analyses, and comprehensive analytical protocols are critical to ensure data quality, comparability, and replicability. As Wolverton (2013:381) noted, it is essential that archaeologists undertake high-quality faunal analyses as zooarchaeological data is increasingly utilised by conservation biologists, historical ecologists, and policy makers (e.g. Augustine and Dearden 2014; Gobalet 2011; Carder and Crock 2012; Erlandson and Fitzpatrick 2006; Groesbeck et al. 2014; Wake et al. 2013).

The relative merits of various faunal quantification methods and the analytical and interpretive implications for using Minimum Numbers of Individuals (MNI), Number of Identifiable Specimens (NISP) and/or weight have been widely discussed in the zooarchaeological literature, (Claassen 1998:106, 2000, Giovas 2009; Glassow 2000; Grayson 1979; Gutiérrez-Zugasti 2011; Lyman 2008:21–140, Mason et al. 1998; Reitz and Wing 2008:202–213). Post-depositional leaching of calcium carbonate from molluscan shell and differential rates of fragmentation within and between taxa can bias NISP values and weight measures of mollusc shell from archaeological deposits. As such, many analysts working on assemblages of invertebrate taxa use MNI to potentially provide a more accurate measure of taxonomic abundance (Ballbè 2005; Mannino and Thomas 2001; Nunn et al. 2007; Poteate and Fitzpatrick 2013; Robb and Nunn 2013; Seeto et al. 2012).

MNI values for molluscan remains are most commonly calculated by counting the frequency of a restricted number of Non-Repetitive Elements (NREs), such as the spire of gastropods or the umbo and hinge of bivalves (Allen 2012; Ballbè 2005; Claassen 1998:106, Chicoine and Rojas 2013; Mannino and Thomas 2001; Mason et al. 1998; Ono and Clark 2012; Poteate and Fitzpatrick 2013; Seeto et al. 2012). This method, however, has the potential to consistently underrepresent some taxa (Giovas 2009). Differential fragmentation resulting from inter-taxonomic variability in shell architecture, morphology, and human processing of mollusc shell can lead to the underrepresentation of particular shell forms; especially gastropods that lack robust, readily identifiable spires.

We propose that current MNI calculation protocols utilising only a restricted range of NRE can influence measures of taxonomic abundance, richness, evenness, and dominance, potentially affecting reconstructions of human behaviour derived from archaeological material. This hypothesis is tested using two assemblages of tropical Indo-Pacific mollusc shells from two prehistoric habitation sites on Ebon Atoll, Marshall Islands. The assemblage was quantified using current MNI

51 calculation protocols and a new, more comprehensive method for calculating molluscan MNI that utilises a wider range of molluscan NREs and calculates MNI based on the most frequently occurring NRE per taxon. Each quantification protocol is tested by comparing relative abundance and rank order abundance. Alterations to taxonomic richness, evenness, and diversity are measured using the number of taxa (NTAXA) Simpson's index of Diversity (1-D), Shannon's Evenness (E), and The Shannon–Weiner index of Diversity (H′).

Results for the Marshall Islands assemblages analysed indicate that current methods for calculating MNI routinely underrepresent gastropod abundance. Furthermore, the application of the new MNI protocol resulted in increased richness and evenness values for all taxa. Given the importance of molluscs for human subsistence and ecosystem maintenance, it is critical to consider the influence of quantification measures on inferences of human behaviour and long-term impacts to the environment. Analysts should employ quantification methods that result in the most precise and accurate measures of relative abundance. Diversity indices, biomass estimates, and mean trophic level are common measures used by zooarchaeologists and fisheries scientists for examining human impacts to marine environments (Carder and Crock 2012; Cinner 2014; Morrison and Hunt 2007; Perry et al. 2011; Wake et al. 2013), and the calculation of these indices using abundance data derived from MNI protocols which utilise only a restricted number of NREs could be misleading or erroneous. The compounding error introduced by such protocols could further bias interpretations and reconstructions of changes in prehistoric subsistence practices and impact modern management of marine resources.

Current methods of MNI calculation in mollusc shell assemblages Common protocols for calculating MNI in molluscan assemblages are described in Mason et al. (1998:308–309) and Claassen (1998:106). These protocols lack the sufficient descriptive detail required (e.g., minimum criteria for determining NRE completeness) to ensure accuracy, precision, and replicability of MNI calculations that would allow analysts to reliably compare datasets. The quantification protocol outlined by Mason et al. (1998:309–309), calculates MNI by counting only a limited range of predetermined shell features, emphasising the frequency of gastropod spires and bivalve hinge and umbones. Moreover, Mason et al. (1998:308) assert that analysts should only identify those fragments that contribute toward MNI. Claassen (1998:106) proposes similar methods for MNI calculation including the quantification of gastropod umbilici. Only the pre- selected NREs are identified and quantified for both methods, regardless of the presence of other molluscan NREs that could increase MNI and influence diversity and richness measures. While

52 mollusc shell fractures in ways that are somewhat predictable (Harris 2011; Koppell 2010; Vermeij 1979; Zuschin et al. 2003), to determine the element or feature to be counted prior to analysis of the archaeological material assumes that there is no variation in fragmentation or preservation at either the assemblage or taxon level. Differential fragmentation or variation in processing techniques could influence preservation of the pre-selected NRE across taxa, potentially under-representing gastropod taxa that lack thick shells and durable spires, and bivalve shells lacking easily identifiable, robust umbones and hinges (see also Giovas 2009; Glassow 2000).

An alternative method proposed by Giovas (2009: Fig. 3) calculates MNI using a part-scoring system based on the presence or absence of portions of mollusc shell (see also Gutiérrez-Zugasti 2011). This method produced ‘no uniform pattern that holds across all taxa and samples’ (Giovas 2009:1560) and remains a valid alternative to NRE MNI. However, NRE based MNI may be less subjective, as NRE are discrete, easily identified features compared to large zones of mollusc shell (Lyman 2008:277). Furthermore, the identification of mollusc shell is often based on diagnostic NRE, and the quantification method proposed here may expedite the quantification process compared to part-scoring methods.

The method outlined below, referred to herein as tMNI, refines and extends upon the NRE MNI method by: (1) including a wider range of NREs for counting, (2) recording element frequency by taxon, and (3) calculating MNI after the highest frequency NRE is identified for each taxon. Ideally, analysts should calculate MNI by stratigraphic layers, rather than arbitrary aggregates which inflate MNI (Grayson 1984). This method is derived from “Traditional MNI” (Giovas 2009:1558), described originally by White (1953) and elaborated by Grayson (1979, 1984), using a range of well-defined molluscan NREs that often preserve in archaeological deposits rather than only hinges, umbones, and spires. The utilisation of an expanded range of NREs reduces the inherent bias towards particular forms (i.e. overall shape) of mollusc shell that occurs when the NRE for the calculation of MNI is selected prior to analysis. While some researchers have implemented this method, or similar methods, (e.g. Allen 2012; Kataoka 1996; Szabó 2009; Rosendahl 2012), no formal protocol has been outlined in detail for the determination of MNI using a wide range of NREs.

Site Description The southernmost atoll in the Marshall Islands, Ebon Atoll (4°38′N, 168°42′E) consists of 22 islets encircling a 104 km2 lagoon. Typical of Marshall Islands atolls, prehistoric villages are marked by

53 concentrations of large marine molluscs (Tridacna spp., Conus spp., Spondylus spp.), shell artefacts, and coral gravel spread to form pavements. Villages are situated parallel to and just inland from the lagoon shore and often run most of the length of the larger islets. TP 18 and 19 (Fig. 1: a), forming a 1 × 2 m unit, were excavated 25 m from the lagoon beach and 30 m west of the council house on a small hill at site MLEb∼-1 (Rosendahl 1987:83; Weisler 1999:∼ Fig. 4, 2002:20). Continuous cultural deposits were encountered to a depth of 1.10 m. At the northeast portion of Ebon Atoll, Enekoion islet (1 km long and 300 m wide) has one major village site (MLEb-33) where TP7 (Fig. 1: b) (1 m2) was excavated 100 m from the lagoon shore just inland of a large swampy gardening area. A cultural layer, 35 cm thick, was encountered.

∼ Site deposits consist of coralline sands with a neutral pH; consequently, shell is well preserved and minimal evidence of shell degradation via dissolution was noted during routine taxonomic identification. Furthermore, in contrast to many mollusc shell assemblages from middens and mounds (Perry and Hoppa 2011 cf. Weisler 2001:Table 7.7), the assemblages analysed here are both taxonomically rich (many species present) and diverse (many taxa represented by many individuals, rather than the assemblage dominated by a single taxon [e.g.Faulkner 2009]). The richness and diversity of the samples will be discussed further in Section 5.2.

Methods A refined protocol for calculating MNI tMNI is calculated by the methods outlined below. All shell should be identified to the lowest taxonomic level and quantified by NISP and/or weight to be broadly comparable with other studies where MNI was not calculated (e.g.Amesbury 1999; Weisler 2001). Here, all specimens were quantified by NISP and weight in grams to two decimal places. The use of a range of quantification measures in concert can highlight the strengths and weaknesses of each method and are useful for addressing additional questions, such as issues relating to taphonomy (see Faulkner 2013; Lyman 2008; Grayson 1984).

To ensure accurate identification, specimens that could not be confidently assigned to species were only identified to genus or family, despite having similar morphology to dominant taxa (Szabó 2009:186). All identifications were completed to the lowest possible taxonomic level using modern Indo-Pacific focussed comparative collections held at the University of Queensland archaeology laboratory. Various reference manuals were also consulted, including: Abbott and Dance (1990), Lamprell and Healy (2006), Poppe (2008), and Röckel et al. (1995). For consistency, all 54 taxonomic names follow the World Register of Marine Species (http://www.marinespecies.org). A review of gastropod and bivalve shell features is presented in the following section, followed by a description of the NREs used to quantify archaeological mollusc remains from two prehistoric habitation sites on Ebon Atoll, Marshall Islands.

Gastropod and bivalve shell features There are five major classes of the Mollusca: the Gastropoda (e.g., whelks and winkles), Bivalvia (e.g., clams and mussels), Cephalopoda (e.g., squid, cuttlefish, Nautilidae and octopuses), Polyplacophora (chitons), and Scaphopoda (tusk shells). Only the Gastropoda and Bivalvia are discussed in detail here, as along with the Polyplacophora, they are often the main classes of molluscs recovered from archaeological sites.

Gastropod shells (Fig. 2) most commonly consist of a hollow cone coiled around a central axis, known as the columella (Stachowitsch 2002, Vermeij 1993; Pechenik 2010). Each full revolution of the shell around the columella is larger than the last as additional shell is secreted along the growth margin (Ruppert et al. 2004). This combination of shell coiling and enlargement produces the typical gastropod morphology of a coiled cone with a pointed top (posterior end) and a large opening at the anterior end. Shell architecture (e.g. micro and macrostructure, teeth, nodes, spines, and ribs) and morphology varies depending on shell growth rate, the overlap of whorls, tightness of coiling, and angle of the aperture in relation to the horizontal plane (Vermeij 1993) (Fig. 3). Additionally, environmental and ontogenetic factors can also influence shell architecture and form (Rhoads and Lutz 1980).

Bivalve shells (Fig. 4) consist of two dorsally hinged, articulating valves that enclose the animal (Gosling 2003:1). Shell shape can be classified broadly depending on the symmetry of valve pairs and individual valve symmetry. The terms equivalve and inequivalve describe the symmetry of one valve in relation to the other (Pechenik 2010). Lateral symmetry (equilateral or inequilateral) describes the symmetry of the anterior and posterior portion of individual valves (Stachowitsch and Proidl 1992). Bivalve form and architecture are determined by the way that the animal lays down shell along the ventral margin and mantle surface, influenced during the life of the animal by a range of ecological and ontogenetic factors (Gosling 2003:7, Rhoads and Lutz 1980) (Fig. 5). A range of morphological elements are defined as NRE. The location of each NRE on the shell and qualitative minimum criteria for counting fragmented NREs are described in the following sections.

If multiple fragments can be refitted to form a complete NRE they should be counted as one NRE. The location and descriptions of NREs are drawn from Stachowitsch and Proidl (1992)

Additional NRE The NREs described below are useful for calculating MNI in archaeological assemblages of tropical marine molluscs from the Marshall Islands. A wider range of NREs may be useful for quantifying other Indo-Pacific molluscan shell assemblages. We recommend that all additional NREs be reported in publication including a clear definition that includes minimum identification and quantification criteria.

Calculating gastropod MNI Gastropod NREs that were used in the quantification of Marshall Islands molluscan remains are the (1) spire (2) anterior notch or canal (3) posterior notch or canal (4) outer lip (5) aperture (6) operculum, and (7) umbilicus (Fig. 6). A quantification key is provided that illustrates examples of our tMNI method for the quantification of gastropod fragments of different shell forms (Fig. 7).

The spire The spire consists of the protoconch and all shell whorls except the body whorl. The protoconch is the shell laid down at the larval stage and can usually be differentiated from all other whorls by a difference in sculpture and smaller size. In molluscan shell from archaeological deposits—if present—the protoconch is generally eroded so that the suture lines at the intersection of whorls are muted or diminished. For this NRE to be counted, greater than 50% of the apex— consisting of the smallest whorls of the spire, and if preserved or present, the protoconch—must be present.

The anterior notch or canal The anterior notch is located on the anterior side of the aperture, at the base of the columella. Many taxa feature extended cylindrical forms of this NRE, known as canals (e.g., the Muricidae subfamily, Muricinae). Anterior canals often fragment in archaeological deposits. To account for post-depositional alteration, the minimum criterion for counting the anterior canal or notch NRE is that the base of the columella is present. Additionally, some taxa (such as the Neritidae) lack anterior notches, and additional NRE should be described and included to account for this variability between taxa.

The posterior notch or canal In taxa where posterior notches occur, the NRE is morphologically similar to the anterior notch, but occurs on the posterior margin of the aperture, rather than the 56 anterior margin. Like anterior notches, these NRE can occur as extended cylindrical protrusions, such as the posterior canals of the tropical frog shell, Bursa bufonia. More than 50% of the posterior notch or canal must be present for this NRE to be counted.

The outer lip The outside edge of the aperture is known as the outer lip. The outer lip can be ornamented with tooth-like protrusions known as denticles (e.g. some Neritidae) or thickened (e.g., Strombidae, Cassidae). The outer lip NRE can be counted if more than 50% is present. The location and morphology of denticles that ornament the outer lip is useful for this determination.

The aperture The opening of the shell at the terminal margin (anterior end) of the coiled cone is the aperture. Apertural morphology is variable, with some taxa (e.g., Trochidae, Turbinidae, Architectonidae) lacking anterior notches or canals. An aperture is counted as whole when the following elements are present: the outer lip, the inner lip (the edge of the aperture that is attached to the columella, often ornamented with folds or plaits), and depending on the specific taxon morphology, the anterior and posterior canals or notches. While this NRE duplicates counts of other NRE noted above, recording aperture frequency contributes to studies of taphonomy (e.g. Szabó 2012) and meat extraction (e.g.Sommerville-Ryan 1998).

The operculum The aperture of most marine gastropods is capped by a flexible proteinaceous or rigid calcified disk known as the operculum. In most cases proteinaceous opercula (e.g. the opercula of Terebralia spp. and Strombus spp.) will not be preserved in archaeological deposits; however, calcified opercula regularly preserve. The opercula of turban shells (Turbo spp.) are ubiquitous in Pacific island archaeological sites (e.g. Allen 1992; Morrison and Hunt 2007; Walter 1998; Szabó 2009). An operculum may be counted if the nucleus can be confidently identified.

The umbilicus If the shell is loosely coiled around the columella an umbilicus (if open, rather than closed) may be present as a small depression or cavity on the base of the columella side of the body whorl. The umbilicus may contribute to MNI only if more than 50% of the element remains. The umbilicus NRE limits the underrepresentation of taxa that lack anterior canals or notches, but may be influenced by intraspecific variation in umbilicus form. For example, some members of the genus Turbo exhibit individual variation in umbilicus form (Carpenter and Niem 1998:412–416), which may influence MNI values based on this NRE.

Gastropod NRE that are excluded from MNI There are several NRE that are excluded from contributing to MNI calculations as they are highly variable across or within taxa, are strongly influenced by shell age, or are too difficult to confidently identify in fragmented archaeological remains. These NRE are the parietal lip and wall, the columella, and exterior shell sculpture. The columella is excluded as a distinct NRE as this element is incorporated with the spire, anterior canal, and inner lip NREs. Any exterior shell sculpture such as ribs, threads, nodules, and spines are also not considered to be a robust and discrete NRE here. It is important to note that the range of NREs outlined here are provided to demonstrate a range of NRE that were useful for the Marshall Islands assemblage, and should be modified to suit the diverse assemblages, taxa, and taphonomic processes which are encountered in the analysis of molluscan remains globally.

Calculating bivalve MNI Bivalve NREs are the: (1) umbo and beak (2) anterior portion of the hinge (3) posterior portion of the hinge (4) anterior adductor muscle scar, and (5) posterior adductor muscle scar (Fig. 8). A quantification key is provided for the Bivalvia (Fig. 9).

All bivalve NREs must be sided in order to be quantified. There are several simple means of siding valves. When teeth are present, they interdigitate on each opposing valve so that where there is a projection on one valve, there will be a socket on the other. The valve side that contains the teeth or sockets is regular and can be used to side valves for quantification. In addition, valve hinges of inequilateral valves are often morphologically distinct on the anterior and posterior sides.

Adductor muscle scars (See Section 4.5.3) can also be used to side valves. These scars are either equal (isomyarian) or of different sizes (anisomyarian). In some taxa, such as Spondylus spp., only one muscle scar is present (monomyarian). For monomyarian taxa valves are sided according to the presence or absence of a muscle scar. In isomyarian taxa, scar morphology can indicate valve side. Anisomyarian taxa can be sided by comparing the relative size (accounting for shell growth) of muscle scars. Additionally, valves can generally be sided by the prominent direction of the umbo. For prosogyrate taxa where the umbo leans to the anterior, when the shell is placed with the dorsal (exterior) surface facing upward, if the umbo points to the right the valve is from the right side of the shell. A left-pointing umbo indicates a left valve. For opisthogyrate taxa where the umbo leans toward the posterior, the siding procedure is reversed.

The umbo and beak Like the spire of gastropods, the umbo and beak are formed during the earliest growth stages of the animal. The beak is present as a small, outwardly protruding feature just above the hinge. The strongly curved portion of the shell that is laid down after the beak is known as the umbo. The umbo can be distinguished by tightly spaced concentric growth bands that can be muted due to pre and post-depositional erosion. The distinctive morphology of the concavity inside the umbo and beak, however, can be used to confidently identify this NRE which can be counted if more than 50% of the beak is present.

The hinge The articulating surfaces which are ventral to the umbo on the interior of the shell valve are known as the valve hinges or hinge. Along with the umbo, the hinge is most commonly preserved in archaeological deposits. The hinge is divided into anterior and posterior NRE at the ventral projection of the beak. By treating the hinge as two NRE in this way, the chances of distinguishing a single individual from fragmented remains are increased and the bias towards taxa with durable umbones is reduced.

The dentition generally present on the hinge of heterodont (complex hinges with a small number of distinct teeth varying in size and shape) bivalves are classified as either cardinal or lateral teeth. Cardinal teeth are found on the centre of the hinge, ventral to the umbo. Lateral teeth are generally present in the form of ridges that lie anteriorly and posteriorly of the cardinal teeth. Cardinal and lateral teeth can be used to determine if more than 50% of the NRE is present. A hinge NRE (anterior or posterior portion) that is determined to be 50% complete can be counted. Inter- taxonomic variation in the arrangement and number of teeth present on the hinge does occur (such as the taxodont Arcidae, with many alternating teeth and sockets present on the hinge), and quantification methods must be evaluated and refined based on the morphology of taxa using similar methods to those described above.

The adductor muscle scars The parts of the live animal responsible for the opening and closing of the shell valves are known as adductor muscles and hinge ligaments. Hinge ligaments rarely preserve archaeologically, but the adductor muscles leave negative impressions or scars on the interior of the shell valve that can be identified in archaeological remains and are considered NREs. Each muscle scar (anterior and posterior) is counted as a single NRE. The minimum criteria for identifying and counting the adductor muscle scar NRE is the presence of more than 50% of the ventral margin of the scar.

NRE excluded Cardinal teeth, lateral teeth, and marginal teeth are excluded as NRE due to their inclusion in the hinge NREs. Marginal teeth are also excluded as it is problematic to determine when more than 50% of these features are present in fragmented remains. The pallial line and sinus are excluded as these features are often removed by post-depositional alteration of bivalve shell and are difficult to confidently identify.

MNI Calculation The frequency of each NRE is recorded as an integer for each taxon in the assemblage. The most frequently occurring NRE for each gastropod taxon is the MNI. For bivalves, only NREs that can be sided are used for quantification. The most frequently occurring NRE for the valve side (left or right) with the highest count is used. Only bivalve NREs that can be confidently assigned a valve side are counted toward MNI.

The quantification protocol outlined above highlights the range of distinct NRE that appear on mollusc shells. Importantly for this study, a range of NRE other than the commonly used spire are equally suitable for quantification, but not currently widely utilised in gastropod MNI calculation protocols (e.g.Claassen 1998; Mason et al. 1998). For bivalves, the adductor muscle scars of molluscs can be useful NREs, but are also not routinely used. The following case study highlights the benefits of using these methods for the quantification of molluscan shell.

To compare quantification methods MNI was calculated based on the most frequently occurring NRE for each taxon at each assemblage using the quantification method outlined above (tMNI) and routine quantification protocols using the spire of gastropods and the umbo and hinge of bivalves (NRE MNI), with the anterior and posterior portion counting as distinct NRE. All mollusc NRE counts were aggregated at the test pit level prior to the calculation of MNI. All NRE described above are used for tMNI calculations. Additionally, for the Cypraeidae, the base and labum adjacent to the aperture were treated as separate NRE (Fig. 10: a) Similarly, for the nerites (Nerita spp.), an additional NRE was utilised that counted the frequency of the distinctive whorls on the interior of the shell where the columellar deck joins the outer lip, essentially replacing the posterior and anterior canal NRE for these taxa (Fig. 10: b: I, I). While MNI is influenced by aggregation effects (Grayson 1984, 1979), site level totals for all taxa highlight the influence of each calculation method on overall abundance, richness, and diversity. In addition, by examining the values for bivalves and gastropods separately, the influence of shell morphology on MNI value can be clearly established. 60

Testing the influence of quantification protocol To assess the influence of MNI calculation protocols on relative abundance, MNI values derived using NRE MNI was compared with MNI values calculated using tMNI. The resulting rank order abundance of taxa for each quantification protocol is compared, and alterations to the top ten ranked taxa are presented. Any observed differences in relative abundance and rank order depending on quantification method highlight the potential influence of quantification measures on a range of archaeological interpretations. Rank order abundance has important implications for examining foraging practices (e.g.Szabó 2009), tracking and reconstructing environmental change (Amesbury 1999) and applying models of optimal foraging theory (Allen 2012; Bird and Bliege Bird 1997 2000; Stephens and Krebs 1986:17–24, Thomas 1999, 2002, 2007a, 2007b). To test the effect of differential fragmentation, the fragmentation ratio for each taxon (NISP:MNI) was compared for each method. This index allows the approximate calculation of the number of fragments per individual (see Faulkner 2010:1946).

Species richness is the number of species present in the analytical unit of study and evenness is the relative abundance of species (Magurran 2004:9). Species richness was calculated and compared for each quantification protocol using NTAXA. NTAXA is a count of the number of distinct taxa in each analytical unit calculated by collapsing taxa at the highest common taxonomic level. NTAXA ensures that species richness is not artificially inflated by taxa which are more easily identified to lower taxonomic levels. For example, if an analytical unit consisted of individuals from the Cypraeidae family identified to Cypraea tigris, Lycina lynx and Cypraeidae spp., then the NTAXA value would be one (1) rather than three (3), as fragments identified as Cypraeidae spp. might possibly be unidentifiable fragments of C. tigris, L. lynx, or other cypraeids.

Assemblage diversity was calculated using multiple indices, the Simpson's index of Diversity (1-D) and the Shannon–Weiner index of Diversity (H′) and Evenness (E), all common to zooarchaeological analyses. 1-D values range from 0 to 1, with higher values indicating even assemblages of many taxa with many individuals from each taxon (Magurran 2004:116). For the Shannon–Weiner index of diversity (H′), theoretical values range between 0 and 5, however, values between 1.5 and 3.5 are most common. Higher H′ values indicate greater diversity and richness. Shannon's Evenness (E) values fall between 0 and 1, with values closer to 0 reflecting assemblages dominated by a single taxon, and higher values reflecting assemblages with many taxa represented by similar numbers of individuals (Lyman 2008:195, Reitz and Wing 2008:111). Random 61 permutation tests of relative abundance data were carried out using the PAST Paleontological Statistics Package, Version 3.04 (Hammer 2001), to test for significant difference between Simpson's Index of Diversity (1-D), Shannon–Weiner Diversity (H′), and Evenness (E) values reported for both quantification methods. As no alterations to taxonomic measures of dominance and evenness were reported for bivalves, significant difference was tested for in gastropod samples only.

Results Total MNI Site level totals demonstrate a marked difference in MNI for each quantification method (Table 1). MNI calculated using tMNI for gastropods was three times larger ( 108%–207% increase in MNI) than NRE MNI calculations. However, the tMNI for bivalves was∼ at most 6% higher than NRE MNI. The most notable distinction in MNI between quantification protocols was reported for gastropods from TP18 and 19 (Table 1).

Differences in MNI depending on quantification method were also reported for the top ten gastropod and bivalve taxa ranked by abundance (Table 2). MNI for gastropods calculated using the tMNI method doubled ( 78%–115% increase) compared to NRE MNI. Bivalve MNI was similar for both quantification protocols.∼

At the site level and for the top ten ranked taxa from each test pit, tMNI resulted in a marked increase in the relative proportion of gastropods to bivalves compared with NRE MNI. Using NRE MNI for TP18 and 19, bivalves accounted for 31% of total MNI, whereas using tMNI, bivalves accounted for 13%. For TP 7, NRE MNI resulted in bivalves accounting for 38% of total MNI, but tMNI reported 23% bivalves and 77% gastropods.

Rank order abundance Each quantification method resulted in different rank order abundance for the top ten gastropod taxa at both sites (Tables 3 and 4). Rank order was distinct for all gastropod taxa for both quantification methods for TP18 and 19. In samples from TP18 and 19 (Table 3), Nerita plicata and Gutturnium muricinum are not reported in the top ten ranked taxa for NRE MNI, but are rank one and two respectively for tMNI. For bivalves, minor alterations to MNI were noted in both contexts and rank order abundance was unaltered by quantification method (Tables 5 and 6).

Species richness and evenness For gastropods, richness at the site level was higher ( 17%–90% increase) in all cases at the family, genus, and species level. NTAXA was slightly lower∼ for TP18 and 19 using the tMNI method as a result of the inclusion of NRE that could not be identified beyond the family level, such as the distinctive siphonal notch of the Muricidae (Table 7). Species richness by NTAXA increased for TP7. For the top ten ranked taxa tMNI calculations decreased richness ( 8%–54%) measured by count of family, genus, and species, and also by NTAXA (Table 7). ∼

Increased evenness and dominance measures for all gastropod taxa are also reported at both sites (Table 8). A significant difference was also reported between Simpson's Evenness and Shannon's Evenness values for the top ten ranked all gastropod taxa at TP18 and 19 (p = <0.001). For the top ten ranked taxa only, evenness was increased for tMNI assemblages at TP18 and 19, but decreased at TP7. A significant difference was also reported between Simpson's Evenness and Shannon's Evenness values for the top ten ranked gastropod taxa at TP18 and 19 (p = <0.001). The majority of MNI values for the top ten ranked taxa were derived from the apex, leading to similar or lower values when additional elements are included using tMNI. No changes to species richness or evenness due to quantification method were reported for bivalves.

Element survivorship Bivalve MNI was not substantially altered by the application of the new protocol at either of the tested sites. In only two cases were the adductor muscle scars the most frequent NRE. However adductor muscles scars were the most frequent NRE only for those taxa with the highest NISP, indicating increased fragmentation relative to other taxa. Furthermore, as many taxa were represented by one or two whole individuals, often umbo, hinge and adductor muscle scar counts were equal. This trend requires further investigation with a larger assemblage of bivalves, assessing the influence of sample size and taphonomic alteration on MNI.

The contribution of the spire to the total MNI for gastropods was low compared to other elements (Tables 9 and 10). These results reflect the differential survivorship and ease of identification of elements across taxa. The taxa where the spire was the most frequently occurring element were those with easily identifiable, high, (Cerithium nodulosum) or low dense spires (Conus spp.). Taxa with low, fragile spires, but identifiable apertures and outer lips (such as the Neritidae) are consistently underrepresented by the NRE MNI method utilising only spires. The MNI for these taxa increased as a direct result of the inclusion of a wider range of NRE, reflected in decreased 63 mean fragmentation ratio values for all gastropod taxa (TP18 & 19: NRE MNI: 10.6/tMNI: 3.4, TP7 NRE MNI: 4.2/tMNI: 2.5) and the top ten ranked gastropod taxa (TP18 & 19: NRE MNI: 9.7/tMNI: 2.3, TP7 NRE MNI: 4.2/tMNI: 2.9) when tMNI was implemented. The inclusion of taxa specific NRE, such as the base and labum of the Cypraeidae, and the additional NRE for the Neritidae increased MNI markedly, highlighting the utility of including additional NRE based on the judgment of the analyst.

Discussion The inclusion of a wider range of molluscan NRE has resulted in notable changes in abundance, diversity and evenness. Studies of the quantification of fish bone from Pacific Island archaeological sites have produced similar alterations to relative abundance and rank order when the range of elements is increased (Lambrides and Weisler 2013). These results highlight the need for analysts to carefully consider the potential loss of information which results from counting a restricted range of NREs. The application of tMNI to an assemblage of mollusc shells from the Marshall Islands has demonstrated that the morphological and architectural complexity of molluscan shell must be considered in MNI quantification protocols. The tMNI and NRE MNI methods are both reliable; if each method was tested multiple times for the same assemblage, results will be consistent (Nance 1987; Giovas 2009:1653). However, tMNI increases measurement validity and provides abundance estimates that more closely estimate actual abundance of taxa within an assemblage (Carr 1987; Mason et al. 1998:308).

Total MNI for the NRE MNI method and the tMNI method are markedly different and the relative abundance of gastropods was increased substantially. Individuals were more likely to be quantified when a wider range of NRE were utilised, indicated by decreased values for fragmentation ratio for all gastropods and the top ten ranked taxa. The rank order abundance and element survivorship for NRE MNI at MLEb-1 demonstrate a dominance of taxa with high, dense spires relative to overall size, including Cerithium columna, C. nodulosum, and Planaxis sulcatus, and gastropods with dense, low spires, such as Conus spp. and Vasum turbinellus. A more balanced range of shell forms are represented when MNI is calculated using a wider range of NRE. The MNI and rank-order abundance of globoid, weak spired taxa such as Nerita spp., and small taxa with apertural thickening, such as Monoplex nicobaricus and Monetaria moneta, were increased. Overall, the spire was rarely the most frequently occurring element when a range of NREs were quantified. Giovas (2009:1563) stated that the NRE method can produce coarse grained estimates of abundance, but

64 the results outlined here indicate that the NRE MNI method provides MNI values that are inherently biased towards particular gastropod shell forms.

Quantification results for all bivalves demonstrated that the umbo and hinge are the elements most frequently preserved in this assemblage. It is possible that the NRE MNI method is adequate for quantifying bivalves, but a larger sample of bivalves is needed for conclusive results. The counting of adductor muscle scars is not a time consuming process. Moreover, the application of tMNI allows the analyst to determine the relative contribution of each NRE to total MNI.

The application of the tMNI method resulted in variable alterations to richness, dominance and evenness. By including a wider range of elements, richness, as measured by count of family, genus, and species, was increased at both sites for all taxa. NTAXA was increased at TP7 for all taxa, as taxa lacking spires which preserve well were represented in MNI calculations. NTAXA decreased at TP18 and 19 for all taxa, as a result of family-level identifications. For example, the Muricidae were represented by individuals from six genera and eight species, but family level identification of the distinctive siphonal notches necessitated a collapsed value of 1 for NTAXA. Quantification of additional NRE is resulted in increased family, genus, and species counts. Family, genus, species and NTAXA values decreased for the top ten ranked taxa at both sites. Using the NRE MNI method, many taxa are represented by single individuals, and many taxa reported the same MNI value. The inclusion of additional NRE added more individuals of each taxon as the bias toward taxa with robust spires was ameliorated. As a result, fewer taxa reported equal MNI values.

For TP18 and 19, significant differences were noted for Simpson's Index values for all gastropod taxa, and the top ten ranked taxa only. Simpson's Index is sensitive to changes in the most abundant species in the sample, while less sensitive to species richness (Magurran 2004:115) The significant difference between quantification methods is likely due to the dominance of Cerithidae (MNI = 64, 43.2% of total MNI) in NRE MNI calculations (Table 3), However, when MNI is calculated using the tMNI method, no single taxon accounts for more than 20.6% of the total MNI. Shannon's index is focussed on species richness and relative abundance (Magurran 2004:107), and the minimal difference in NTAXA and differential distribution of individuals across taxa likely accounts for the non-significant difference between quantification methods for all taxa, and the significant difference between top-ten ranked gastropod taxa from TP18 and 19. For TP7, evenness is similar between quantification methods (Table 4), reflected in the non-significant difference between Simpson's

65 index values derived from both the NRE MNI and tMNI methods. Changes in diet breadth and evenness are critical measures for testing hypotheses relating to declines in foraging efficiency and the long-term impacts of human extraction of molluscan resources (e.g. Broughton 1997; Nagaoka 2002). Measures of taxonomic abundance and heterogeneity are also crucial to understanding forager decision making (e.g. Thomas 2007a), assisting in reconstructing environmental change (Amesbury 1999; Faulkner 2013), and informing modern conservation efforts (e.g. Carder and Crock 2012; Wake et al. 2013). While not all differences were statistically significant, alterations to measures of taxonomic composition highlight the potential influence of quantification protocol on archaeological inference.

Researchers are not able to directly compare the abundances of fish and molluscs using current protocols. The NRE MNI method tends to measure the survival of the pre-selected counting character, rather than the relative abundance of all taxa in the sample. The tMNI method allows invertebrate abundance data to be compared to vertebrate abundance data (e.g. fishbone or terrestrial animals) as the most frequently occurring NRE is used for both measures. In many coastal sites, especially in the Pacific Islands, molluscan and finfish resources are important sources of protein, often captured from the same habitats. Therefore, NRE MNI protocols may be masking changes in subsistence systems that are reflected by shifting abundances of molluscs and finfish. The preservation or lack thereof of particular NREs may also be a useful indicator of a range of human behaviours, including off-site processing or shell-working.

Conclusion The new protocol for the quantification of mollusc shell demonstrates the impact of increasing the range of NRE included in MNI calculations. This new protocol has produced MNI values that more closely match the actual abundance of taxa from an archaeological assemblage in the Marshall Islands. Using this quantification protocol, alterations to richness, diversity, and rank order abundance were reported. Existing quantification protocols for molluscan shell consistently underrepresent the relative abundance of gastropods to bivalves and underrepresent the abundance of particular shell forms. To enhance the study of prehistoric subsistence and provide data that are increasingly used by conservation biologists, historical ecologists, and policy makers, it is critical that the most accurate representation of richness, evenness, and diversity are provided. We recommend that the implications of utilising tMNI quantification protocols be routinely considered. We encourage analysts to locally adapt and refine the protocols outlined here for the analysis of molluscan assemblages.

Acknowledgements We thank the Vice Chancellor of the University of Queensland for strategic funding to the archaeology programme. Marshall Islands fieldwork, conducted under permit from the Historic Preservation Office (Republic of the Marshall Islands) was supported by a grant to Weisler from the Office of the Deputy Vice Chancellor (Research). Harris' postgraduate studies are supported by an Australian Postgraduate Award. The authors thank the three anonymous reviewers for their useful comments on the manuscript.

References Cited Abbott, R.T. and S.P. Dance. 1990. Compendium of Seashells. Melbourne, Florida: American Malacologists, Inc. Allen, M.S. 1992. Dynamic Landscapes and Human Subsistence: Archaeological Investigations on Aitutaki Islands, Southern Cook Islands. Ph.D. thesis. Washington: University of Washington. ---. 2012. Molluscan foraging efficiency and patterns of mobility amongst foraging agriculturalists: a case study from northern New Zealand. Journal of Archaeological Science 39(2):295-307. Amesbury, J.R. 1999. Changes in species composition of archaeological marine shell assemblages in Guam. Micronesica 31(2):347-366. Augustine, S. and P. Dearden. 2014. Changing paradigms in marine and coastal conservation: A case study of clam gardens in the southern Gulf Islands, Canada. The Canadian Geographer 58(3):305-314. Ballbè, E.G. 2005. Shell middens in the Caribbean coast of Nicaragua: Prehistoric patterns of molluscs collection and consumption. Oxford: Oxbow Books. Bird, D.W. and R.L. Bliege Bird. 1997. Contemporary shellfish gathering strategies among the Meriam of the Torres Strait Islands, Australia: Testing predictions of a central place foraging model. Journal of Archaeological Science 24(1):39-63. ---. 2000. The ethnoarchaeology of juvenile foragers: Shellfishing strategies among Meriam children. Journal of Anthropological Archaeology 19(4):461-476. Broughton, J.M. 1997. Widening diet breadth, declining foraging efficiency, and prehistoric harvest pressure: Ichthyofaunal evidence from the Emeryville Shellmound, California. Antiquity 71(274):845-862. Bunce, M., L.D. Rodwell, R. Gibb and L. Mee. 2008. Shifting baselines in fishers' perceptions of island reef fishery degradation. Ocean & Coastal Management 51(4):285-302. Carder, N. and J.G. Crock. 2012. A pre-Columbian fisheries baseline from the Caribbean. Journal of Archaeological Science 39(10):3115-3124. Carpenter, K.E. and V.H. Niem (eds) 1998.The Living Marine Resources of the Western Central Pacific: Volume 1. Seaweeds, Corals, Bivalves and Gastropods. Rome: Food and Agriculture Organization of the United Nations. Carr, C. 1987. Removing discordance from quantitative analysis. In Quantitative Research in Archaeology: Progress and Prospects (M. Aldenderfer, ed.):185-243. Los Angeles: Sage Publications.

Chicoine, D. and C. Rojas. 2013. Shellfish resources and maritime economy at Caylán, coastal Ancash, Peru. The Journal of Island and Coastal Archaeology 8(3):336-360. Cinner, J. 2014. Coral reef livelihoods. Current Opinion in Environmental Sustainability 7(0):65- 71. Claassen, C. 1998. Shells. Cambridge: Cambridge University Press. Erlandson, J.M. and S.M. Fitzpatrick. 2006. Oceans, islands, and coasts: Current perspectives on the role of the sea in human prehistory. The Journal of Island and Coastal Archaeology 1(1):5-32. Erlandson, J.M. and T.C. Rick. 2010. Archaeology meets marine ecology: The antiquity of maritime cultures and human impacts on marine fisheries and ecosystems. Annual Review of Marine Science 2:231-251. Faulkner, P. 2009. Focused, intense and long-term: evidence for granular ark (Anadara granosa) exploitation from late Holocene shell mounds of Blue Mud Bay, northern Australia. Journal of Archaeological Science 36(3):821-834. ---. 2010. Morphometric and taphonomic analysis of granular ark (Anadara granosa) dominated shell deposits of Blue Mud Bay, northern Australia. Journal of Archaeological Science 37(8):1942-1952. ---. 2013. Terra Australis 38: Life on the Margins: An Archaeological Investigation of Late Holocene Economic Variability, Blue Mud Bay, Northern Australia. Canberra ANU E Press Giovas, C.M. 2009. The shell game: analytic problems in archaeological mollusc quantification. Journal of Archaeological Science 36(7):1557-1564. Glassow, M.A. 2000. Weighing vs. counting shellfish remains: A comment on Mason, Peterson, and Tiffany. American Antiquity 65(2):407-414. Gobalet, K.W. 2011. A Native californian's meal of coho salmon (Oncorhynchus kisutch) has legal consequences for conservation biology. In Exploring Methods of Faunal Analysis: Insights from California Archaeology (M.A. Glassow and T.L. Joslin, eds):87-95. Los Angeles: Cotsen Institute of Archaeology Press. Gosling, E. 2003. Bivalve Molluscs. Oxford: Fishing News Books. Grayson, D.K. 1979. On the quantification of vertebrate archaeofaunas. Advances in Archaeological Method and Theory 2:199-237. Grayson, D.K. 1984. Quantitative Zooarchaeology: Topics in the Analysis of Archaeological Faunas. Orlando: Academic Press.

Groesbeck, A.S., K. Rowell, D. Lepofsky and A.K. Salomon. 2014. Ancient clam gardens increased shellfish production: Adaptive strategies from the past can inform food security today. PLoS One 9(3):1-13 Gutiérrez-Zugasti, I. 2011. Shell fragmentation as a tool for quantification and identification of taphonomic processes in archaeomalacological analysis: The case of the Cantabrian region (Northern Spain). Archaeometry 53(3):614-630. Hammer, Ø. 2001. PAST PAleontological STatistics. Oslo, Norway: Natural History Museum, University of Oslo. 3.04. Harris, M. 2011. Tools or Tucker? Developing Methods for Identifying Utilised Polymesoda (Geloina) erosa (Bivalvia:Corbiculidae) Shell Valves. Brisbane: Unviersity of Queensland. Hayek, L.-A.C. and M.A. Buzas. 2010. Surveying Natural Populations: Quantitative Tools for Assessing Biodiversity. New York: Columbia University Press. Kataoka, O. 1996. Prehistoric and historic faunal utilization in Pohnpei: An ecological and ethnoarchaeological understanding. University of Oregon. Koppel, B. 2010. Fracture and Modification of Patella shell: Distinguishing Human Working. Honours thesis. Wollongong: University of Wollongong. Lambrides, A.B.J. and M.I. Weisler. 2013. Assessing protocols for identifying Pacific Island archaeological fish remains: The contribution of vertebrae. International Journal of Osteoarchaeology 25(6):838-848. Lamprell, K. and J.M. Healy. 2006. Spiny Oysters : A Revision of the Living Spondylus Species of the World. Brisbane: Jean Lamprell. Lyman, R.L. 2008. Quantitative Paleozoology. Cambridge: Cambridge University Press. Magurran, A.E. 2004. Measuring Biological Diversity. Malden: Oxford; Blackwell. Mannino, M.A. and K.D. Thomas. 2001. Intensive Mesolithic exploitation of coastal resources? Evidence from a shell deposit on the isle of Portland (southern England) for the impact of human foraging on populations of intertidal rocky shore molluscs. Journal of Archaeological Science 28(10):1101-1114. Mason, R.D., M.L. Peterson and J.A. Tiffany. 1998. Weighing vs. counting: Measurement reliability and the California school of midden analysis. American Antiquity 63(2):303-324. Morrison, A.E. and T.L. Hunt. 2007. Human Impacts on the nearshore environment: An archaeological case study from Kaua'i, Hawaiian Islands. Pacific Science 61(3):325- 328,331-345. Nagaoka, L. 2002. The effects of resource depression on foraging efficiency, diet breadth, and patch use in southern New Zealand. Journal of Anthropological Archaeology 21(4):419-442.

Nance, J.D. 1987. Reliability, validity, and quantitative methods in Archaeology. In Quantitative Methods in Archaeology (M. Aldenderfer, ed.):244-293. Beverly Hills: Sage Publications. Nunn, P.D., T. Ishimura, W.R. Dickinson and K. Katayama. 2007. The Lapita occupation at Naitabale, Moturiki Island, central Fiji. Asian Perspectives 46(1):96-132. Ono, R. and G. Clark. 2012. A 2500-year record of marine resource use on Ulong Island, Republic of Palau. International Journal of Osteoarchaeology 22(6):637-654. Pauly, D. 1995. Anecdotes and the shifting baseline syndrome of fisheries. Trends in Ecology & Evolution 10(10):430. Pechenik, J.A. 2010. Biology of the Invertebrates. Dubuque: McGraw-Hill. Perry, C.T., P.S. Kench, S.G. Smithers, B. Riegl, H. Yamano and M.J. O'Leary. 2011. Implications of reef ecosystem change for the stability and maintenance of coral reef islands. Global Change Biology 17(12):3679-3696. Perry, J.E. and K.M. Hoppa. 2011. Subtidal shellfish exploitation on the California Channel Islands: Wavy top (Lithopoma undosum) in the middle Holocene. In Exploring Methods of Faunal Analysis: Insights from California Archaeology (M.A. Glassow and T.L. Joslin, eds): Left Coast Press. Poppe, G.T. 2008. Philippine Marine Mollusks. Hackenheim: ConchBooks. Poteate, A.S. and S.M. Fitzpatrick. 2013. Testing the efficacy and reliability of common zooarchaeological sampling strategies: A case study from the Caribbean. Journal of Archaeological Science 40(10):3693-3705. Reitz, E.J. and E.S. Wing. 2008. Zooarchaeology. Cambridge: Cambridge University Press. Rhoads, D.C. and R.A. Lutz. 1980. Skeletal Growth of Aquatic Organisms: Biological Records of Environmental Change. New York: Plenum Press. Robb, K.F. and P.D. Nunn. 2013. Changing role of nearshore-marine foods in the subsistence economy of inland upland communities during the last millennium in the tropical Pacific Islands: insights from the Bā River Valley, Northern Viti Levu Island, Fiji. Environmental Archaeology 19(1):1-11. Röckel, D., W. Korn and A.J. Kohn. 1995. Manual of the Living Conidae. Hackenheim: Verlag Christa Hemmen. Rosendahl, P.H. 1987. Report 1: Archaeology in eastern Micronesia: Reconnaissance survey in the Marshall Islands. In Marshall Islands Archaeology (T. Dye, ed.):17-166. Hawaii: Pacific Anthroplogical Records: Bernice Pauahi Bishop Museum. Ruppert, E.E., R.S. Fox and R.D. Barnes. 2004. Invertebrate zoology: A functional evolutionary approach. Belmont: Thomson-Brooks/Cole.

Seeto, J., P.D. Nunn and S. Sanjana. 2012. Human-mediated prehistoric marine extinction in the tropical Pacific? Understanding the presence of Hippopus hippopus (Linn. 1758) in ancient shell middens on the Rove Peninsula, southwest Viti Levu Island, Fiji. Geoarchaeology 27(1):2-17. Sommerville-Ryan, G. 1998. The Taphonomy of a Marshall Islands' Shell Midden. M.A. thesis. Dunedin: University of Otago. Stachowitsch, M. and S. Proidl. 1992. The Invertebrates: An Illustrated Glossary. New York: Wiley-Liss. Stephens, D.W. and J.R. Krebs. 1986. Foraging Theory. Princeton: Princeton University Press. Szabó, K. 2009. Molluscan remains from Fiji. In Terra Australis 31: The Early Prehistory of Fiji (G. Clark and A.J. Anderson, eds):183-212. Canberra: ANU E Press. ---. 2012. Terrestrial hermit crabs (Anomura: Coenobitidae) as taphonomic agents in circum- tropical coastal sites. Journal of Archaeological Science 39:931-941. Thomas, F.R. 1999. Optimal Foraging and Conservation: The Anthropology of Mollusk Gathering Strategies in the Gilbert Islands Group, Kiribati. Ph.D. thesis. Honolulu: University of Hawaii at Manoa. ---. 2002. An evaluation of central-place foraging among mollusk gatherers in western Kiribati, Micronesia: Linking behavioral ecology with ethnoarchaeology. World Archaeology 34(1):182-208. ---. 2007a. The behavioral ecology of shellfish gathering in western Kiribati, Micronesia 1: Prey choice. Human Ecology 35(2):179-194. ---. 2007b. The behavioral ecology of shellfish gathering in western Kiribati, Micronesia. 2: Patch choice, patch sampling, and risk. Human Ecology 35(5):515-526. Vermeij, G.J. 1979. Shell architecture and causes of death of Micronesian reef snails. Evolution 33(2):686-696. ---. 1993. A Natural History of Shells. Princeton: Princeton University Press. Wake, T.A., D.R. Doughty and M. Kay. 2013. Archaeological investigations provide late Holocene baseline ecological data for Bocas del Toro, Panama. Bulletin of Marine Science 89(4):1015-1035. Walter, R. 1998. Anai'o: The Archaeology of a Fourteenth Century Polynesian Community in the Cook Islands. Dunedin: New Zealand Archaeological Association. Weisler, M.I. 1999. The antiquity of aroid pit agriculture and signiﬁcance of buried A horizons on Paciﬁc Atolls. Geoarchaeology: An International Journal 14(7):621-654.

---. 2001. On the Margins of Sustainability: Prehistoric Settlement of Utrōk Atoll, Northern Marshall Islands. Oxford: Archaeopress. ---. 2002. Archaeological Survey and Test Excavations on Ebon Atoll, Republic of the Marshall Islands. Majuro Historic Preservation Office, Republic of the Marshall Islands. White, T.E. 1953. A method of calculating the dietary percentage of various food animals utilized by aboriginal peoples. American Antiquity:396-398. Wolverton, S. 2013. Data quality in zooarchaeological faunal identification. Journal of Archaeological Method and Theory 20(3):381-396. Zuschin, M., M. Stachowitsch and R.J. Stanton, Jr. 2003. Patterns and processes of shell fragmentation in modern and ancient marine environments. Earth-Science Reviews 63:33- 82.

Figures and Tables Table 1 Total MNI for all taxa, aggregated at test pit level

NRE Site tMNI MNI Gastropoda Eb-1 TP18 &19 188 577 Eb-33 TP7 125 260 Bivalvia Eb-1 TP18 &19 83 88 Eb-33 TP7 78 78 Total 474 1003

Table 2 Total MNI for the top ten ranked taxa only, aggregated at test pit level NRE Site tMNI MNI Gastropoda Eb-1 TP18 &19 170 367 Eb-33 TP7 125 223 Bivalvia Eb-1 TP18 &19 83 88 Eb-33 TP7 78 78 Total 456 756

Table 3 Unique family, genus and species counts for gastropods aggregated by test pit. For each test pit, the counts for all taxa and the top ten ranked taxa are presented. All taxa Top ten taxa

Eb-1 TP18 & 19 NRE MNI tMNI NRE MNI tMNI Families 18 21 13 6 Genera 24 32 14 8 Species 21 40 12 7 NTAXA 26 25 14 9

Eb33 TP7 NRE MNI tMNI NRE MNI tMNI Families 14 17 14 11 Genera 15 23 15 9 Species 14 21 14 11 NTAXA 15 19 15 12

Table 4 Evenness and dominance measures. D = Simpson’s Dominance, 1-D = Simpson’s Evenness, H’ = Shannon’s index, H’/lnS = Shannon’s evenness All Taxa TP18 & 19 TP7

NRE MNI tMNI NRE MNI tMNI

1-D 0.787 0.884 0.822 0.830 H’ 2.273 2.412 1.988 2.182 H’/lnS 0.698 0.749 0.734 0.741

Top ten ranked TP18 & 19 TP7

NRE MNI tMNI NRE MNI tMNI

1-D 0.762 0.858 0.841 0.825 H’ 1.856 2.069 2.084 2.057 H’/lnS 0.724 0.942 0.770 0.828

Table 5 Eb-1 T18 & 19 NRE MNI and tMNI quantification results for gastropod taxa. * = change in rank order; ** = unique to rank order abundance for that method. NRE MNI tMNI Rank Taxon MNI NISP Rank Taxon MNI NISP

1 Conus spp.* 36 190 1 Nerita plicata** 83 141

2 Cerithium columna* 32 45 2 Gutturnium muricinum** 70 132

3 Cerithidae spp.* 20 37 3 Conus spp.* 36 190

4 Vasum turbinellus** 11 73 4 Cerithium columna* 32 45

5 Cerithium nodulosum** 10 176 5 Monoplex intermedius** 31 41

6 Planaxs sulcatus** 10 16 6 Monoplex nicobaricus** 29 37

7 Drupa spp. ** 8 37 7 Nerita polita** 27 75

7 Canarium spp. ** 5 13 8 Monetaria moneta** 22 33

8 Turbo argyrostomus** 5 51 9 Cerithidae spp.* 20 37

8 Turbo spp. ** 5 85 10 Chicoreous spp.* 17 64

8 Mitra stictica** 4 18 9 Chicoreous spp.* 3 64 9 Trochus maculatus** 3 58 9 Monoplex spp. ** 3 46 9 Conus flavidus** 3 3 9 Nerita signata** 2 9 9 Melampus flavus** 2 6 9 Cerithium echinatum** 2 3 9 Pollia sp. ** 2 3 9 Trochidae spp. ** 2 8 10 Turbo setosus** 2 49

Table 6 Eb-33 TP7 NRE MNI and tMNI quantification results for gastropod taxa. * = change in rank order; ** = unique to rank order abundance for that method. NRE MNI tMNI Rank Taxon MNI NISP Rank Taxon MNI NISP

1 Melampus flavus* 35 44 1 Nerita polita* 73 122

2 Conus spp. * 17 71 2 Melampus flavus* 39 44

3 Vasum turbinellus 16 90 3 Vasum turbinellus 28 90

4 Nerita polita* 15 122 4 Conus spp.* 23 71

5 Turbo argyrostomus* 7 60 5 Bursa spp.** 13 36

6 Cerithidae spp.* 6 8 6 Nerita plicata* 9 11

7 Nerita plicata 5 11 7 Turbo argyrostomus* 7 60

7 Cerithium nodulosum 5 48 8 Cerithidae spp.* 6 8

8 Canarium spp. 2 9 8 Monoplex nicobaricus** 6 6

8 Thais armigera** 2 8 9 Cerithium nodulosum* 5 48

8 Nerita spp. ** 2 7 9 Canarium spp.* 5 9 9 Drupa ricinus** 1 4 9 Gutturnium muricinum** 5 6 9 Mitra stictica** 1 8 10 Drupa ricinus* 4 4 9 Monetaria moneta** 1 2 9 Muricidae spp.** 1 6 9 Cerithium columna** 1 1 9 Sabia conica ** 1 2 9 Pterigya sp. ** 1 1 9 Drupa rubusidaeus** 1 2 9 Mamilla sp. ** 1 1 9 Nerita albicilla** 1 1 9 Terebra spp. ** 1 2 9 Trochidae spp. ** 1 2 9 Turbo spp. ** 1 17 10 - - -

Table 7 Eb-33 TP7 NRE MNI and tMNI quantification results for bivalve taxa. NRE MNI tMNI Rank Taxon MNI NISP Rank Taxon MNI NISP

1 Fragum spp. 48 51 1 Fragum spp. 48 51

2 Asaphis violascens 10 27 2 Asaphis violascens 10 27

3 Chama spp. 10 11 3 Chama spp. 10 11

4 Corculum cardissa 4 4 4 Corculum cardissa 4 4

5 Spondylus sinensis 2 2 5 Spondylus sinensis 2 2

5 Ctena bella 2 2 5 Ctena bella 2 2

6 Arca sp. 1 1 6 Arca sp. 1 1

6 Spondylus sp. 1 1 6 Spondylus sp. 1 1

Table 8 Eb-1 TP18 and 19 NRE MNI and tMNI quantification results for bivalve taxa. NRE MNI tMNI Rank Taxon MNI NISP Rank Taxon MNI NISP

1 Asaphis violascens 25 145 1 Asaphis violascens 26 145

2 Fragum spp. 10 11 2 Fragum spp. 10 11

2 Gafraium spp. 10 23 2 Gafrarium spp. 10 23

3 Ctena bella 6 6 3 Ctena bella 6 6

4 Tridacna spp. 3 4 4 Tridacna spp. 3 4

5 Arca spp. 2 3 5 Arca spp. 2 3

5 Barbatia spp. 2 5 5 Barbatia spp. 2 5

5 Tridacna maxima 2 5 5 Tridacna maxima 2 5

5 Chama spp. 2 7 5 Chama spp. 2 7

5 Codakia divergens 2 2 5 Codakia divergens 2 2 5 Lucinidae spp. 2 2 5 Lucinidae spp. 2 2 5 Spondylus spp. 2 4 5 Spondylus spp. 2 4 6 Tridacna crocea 1 2 6 Tridacna crocea 1 2 6 Tridacna cf. gigas 1 2 6 Tridacna cf. gigas 1 2 6 Vasticardium elongatum 1 1 6 Vasticardium elongatum 1 1 6 Codakia tigerina 1 1 6 Codakia tigerina 1 1 6 Pectinidae spp. 1 2 6 Pectinidae spp. 1 2 6 Pinctada spp. 1 20 6 Pinctada spp. 1 20

Table 9 Contribution of the pre-selected bivalve counting character for all taxa. Most common element %umbo

Umbo & hinge Other & hinge Eb-1 TP18 & 19 19 2 90.5 Eb-33 TP7 8 0 100

Table 10 Contribution of the pre-selected gastropod counting character for all taxa. Most common element % spire Spire Other Eb-1 TP18 & 19 27 37 42.2 Eb-33 TP7 14 24 36.8

Figure 2 Gastropod terminology.

Figure 4 Bivalve terminology.

Figure 5 Examples of bivalve shapes a. orbicular b. alate c. auriculate d. subquadrate e. trigonal f.fan-shaped g.ensiform h. elongate-elliptical.

Chapter 4: Intertidal foraging on atolls: prehistoric forager decision making at Ebon Atoll, Marshall Islands

Note: This is the final version of a peer-reviewed article published in the Journal of Island and Coastal Archaeology

Matthew Harris and Marshall I. Weisler

School of Social Science, The University of Queensland, St. Lucia, QLD 4072 Australia

Corresponding author: Matthew Harris, School of Social Science, The University of Queensland, St. Lucia, QLD 4072 Australia. Email: [email protected]

Abstract Prehistoric molluscan assemblages provide insights into long-term patterns of human landscape use, environmental change, and human impacts to marine resources. The investigation of forager decision-making regarding the selection of certain mollusc taxa and/or the exploitation of particular habitats is fundamental to understanding human-environment interactions in the past, and is relevant for understanding trajectories of human impacts to the intertidal zone in coastal settings. We document variability in the collection of molluscs at two archaeological sites on Ebon Atoll, Republic of the Marshall Islands: one on a windward, intermittently occupied islet, and the other on a permanently inhabited leeward islet. All molluscan taxa were assigned to a range of habitats within a hierarchical classification scheme for intertidal marine environments. The relative abundance of taxa from each habitat was used as a proxy for forager decision-making. We report a generalized, non-selective, foraging strategy focused on gastropod taxa from the high intertidal and supratidal. These results indicate that rather than focusing intensively on select taxa, intertidal foragers targeted particular marine habitats, taking advantage of the predictable behaviors of the molluscs that inhabit them.

Keywords: archaeomalacology, atoll archaeology, marine subsistence, Marshall Islands, shell midden studies

Introduction Molluscs have formed a component of the human diet worldwide for at least 165,000 years (Jerardino and Marean 2010), and remain essential to many coastal communities today (Aswani et al. 2015). The analysis and interpretation of mollusc assemblages from archaeological sites have revealed the complex relationship between humans, molluscs, and marine environments (e.g Bailey and Milner 2008, Fitzpatrick et al 2009, Giovas et al. 2013, Hunt et al. 2011, Jerardino and Navarro 2008, Jiao 2007, Koike 1975, 1986, Poteate et al. 2014) which includes evidence for the formation or construction of large shell mounds in Australia (Bailey 1975, Bailey et al. 1994) Brazil (Gaspar et al. 2008) and North America (Nelson 1909, Uhle 1907), the intensive collection of particular taxa (e.g., Erlandson et al. 2011, Faulkner 2013), size-selective harvesting (O’Dea et al. 2014, Whitaker 2008), and responses to resource depression, environmental and climatic change and human impacts to ecosystems (e.g., Braje et al. 2007, Fitzpatrick and Keegan 2007, Parkington et al. 2013). Furthermore, studies of contemporary forager behaviour have shown that human foraging for molluscs in the intertidal is mediated by a range of factors including tidal movement, substrate type, socio-economic and cultural factors, and other economic considerations explored primarily through optimal foraging theory (e.g., Bird and Bliege Bird 1997, de Boer et al. 2002, Thomas 2001, 2002, 2007). While several studies have highlighted the importance of molluscan resources in prehistoric subsistence strategies in the Marshall Islands (Dye 1987; Riley 1987; Weisler 1999b, 2001b), intertidal foraging has not been investigated in detail.

Consisting primarily of unconsolidated sand and gravel atop a narrow reef platform surrounding a lagoon, low coral atolls are challenging landscapes for sustained human habitation; yet the atoll archipelago of the Marshall Islands has been occupied by humans for at least 2000 years (Kayanne et al. 2011; Weisler 1999a, 2001b; Weisler et al. 2012). The intertidal zone of atolls is a complex mosaic of benthic habitats that host a diverse range of flora and fauna exploited for subsistence by prehistoric and present day Marshallese.

Reconstruction of human foraging preferences based on the habitat preferences of marine fauna has proven utility in Pacific Islands archaeology (Allen 1992, 2012; Morrison and Addison 2008; Morrison and Hunt 2007; Spennemann 1987; Szabó 2009; Thomas 2002; Weisler et al. 2010). Changes or stability in the exploitation of marine resources from particular habitats may indicate variation in foraging strategies relating to overexploitation (Morrison and Hunt 2007), environmental change (Amesbury 2007; Spennemann 1987) or foraging preferences (Allen 2012; Szabó 2009). A more comprehensive understanding of the foraging patterns of prehistoric

Marshallese is fundamental to addressing questions of resource sustainability, human impacts to the marine environment, and overall subsistence systems on oceanic atolls.

To investigate the relationship between prehistoric settlement patterns and human foraging for molluscs, two assemblages from archaeological sites on Ebon Atoll, Marshall Islands are compared: one from a small intermittently occupied habitation on a windward islet, and another from a substantial permanent village on a leeward islet. All molluscan taxa were assigned a range of habitats using a hierarchical classification scheme adapted from modern benthic habitat maps of Majuro Atoll, Marshall Islands (Kendall et al. 2012), and the relative abundance of taxa from each habitat is used to reconstruct time-averaged patterns of forager decision making (Allen 1992; Nagaoka 2002; Szabó 2009). This method aimed to determine overall patterns of forager behaviour, highlighting the habitats from which the majority of the assemblage could have been gathered and providing new insights into prehistoric marine subsistence on Ebon Atoll.

Previous archaeology in the Marshall Islands Visiting 12 Marshall Islands atolls in 1977, Rosendahl (1987) conducted brief archaeological surveys, surface artefact collection and limited test excavations. In 1979, as part of a larger programme of survey and excavation, Riley (1987) completed transect excavations at Laura village on Majuro’s western rim to discover the depth and area of cultural deposits, establishing initial occupation of the atoll at 1970 ± 110 BP (Riley 1987:Table 2.28). A year later, Dye (1987), surveyed 132 of 133 islets on Arno Atoll, sampling 10 islets with transect excavations totalling 48.5m2. Three islets yielded prehistoric deposits indicating permanent habitation, with coral gravel spreads, midden deposits and associated aroid pit agriculture (see Weisler 1999a).

The next major phase of archaeological research began in 1993, with Weisler’s interdisciplinary project documenting variability in settlement and subsistence across the ~900 mm to 4000 mm rainfall gradient from the dry north to the wet south of the archipelago. Ebon (4°N), Maleolap (8°N), Ujae (9°N), and Utrōk (11°N) have been intensively surveyed and excavated as part of this project (Weisler 1999a, b, 2001a, b, 2002; Weisler et al. 2012). The earliest habitation dates, beginning about 2000BP, were consistently identified on the largest leeward islets associated with aroid pit agriculture (Weisler 1999a; Weisler et al. 2012) , while smaller intermittently occupied habitations were found on the windward islets.

Various other projects have been conducted, some in response to modern development activities (Shun and Athens 1990; Streck 1990), others focused on burial practices, human osteology and prehistoric interaction (Spennemann 1999; Weisler 2000; Weisler et al. 2000; Weisler and Swindler 2002), material culture (Spennemann 1993; Weisler 2000; Widdicombe 1997), prehistoric horticulture (Horrocks and Weisler 2006; Weisler 1999a), landscape history and the timing of initial colonisation (Kayanne et al. 2011; Streck 1990; Weisler et al. 2012; Yamaguchi et al. 2009) as well as the chronology for the human introduction of lizards (Pregill and Weisler 2007) and land snails (Christensen and Weisler 2013). Only one study focused on molluscs from prehistoric middens which considered primarily taphonomic issues (Sommerville-Ryan 1998).

Mollusc assemblages from Marshall Islands archaeological sites A summary of previous excavations that include the identification and quantification of molluscan remains from prehistoric-period archaeological sites in the Marshall Islands is provided (Table 1), focusing on two components of the assemblages: taxonomic composition and the habitats from which molluscs were likely collected (Dye 1987; Riley 1987; Weisler 2001b). Riley (1987) and Weisler (2001b) utilised NISP (Number of Individual Specimens) counts and weight, whereas Dye used presence/absence to characterise midden assemblages. Single quantification measures can be problematic for assessing relative abundance when used independently of other measures (i.e., estimates of Minimum Numbers of Individuals [MNI]) and hinder precise comparison between the assemblages and inferences regarding forager decision-making. The predominance of genus- or family-level identifications, and variable recovery and analytical techniques, allow only broad assessment of foraging practices for the assemblages discussed here. NISP rather than weight is used to compare relative abundance where possible; NTAXA is used to compare species richness and broad habitat assignments are based on the ecological review outlined below. NTAXA quantifies the number of distinct taxa in the assemblage by collapsing taxa at the highest common taxonomic level. For example, if an analytical unit consisted of Conomurex luhuanus, Gibberulus gibberulus, Lambis sp. and Strombidae fragments, the NTAXA value would be collapsed to the highest taxonomic level, in this case, family as all identified taxa belong to the family Strombidae. Therefore, the NTAXA value would be one rather than four. This method avoids inflating species richness when taxa are more easily identified to lower taxonomic levels, i.e., genus or species. Molluscan remains were assessed at the site level to highlight broad patterns in mollusc exploitation.

Prehistoric mollusc assemblages from three atolls (Majuro, Arno and Utrōk) Marshall Islands are generally rich and diverse. Major village sites on Arno and Utrōk atolls indicate a foraging strategy focused primarily on a wide range of taxa from the lagoonside intertidal, and secondarily from the oceanside reef flat and reef edge. Only Majuro sites indicate a focus on a single taxon, with Conomurex luhuanus accounting for at least 80% of the total NISP from each Majuro site. Village site assemblages are diverse and even, but are primarily composed of gastropods, aligning with modern ecological surveys that report a ratio of 9:1 gastropod to bivalve species in the Marshall Islands (Kay and Johnson 1987:133). Assemblages from small, intermittently occupied sites away from the main village, are similarly rich, diverse and generalised, but somewhat reflect the relative distance of the site from the oceanside or lagoonside. For example, at site MLUt-4, on Utrōk Atoll (Weisler 2001b:24-26, 51-59) where the lagoonside environment is much less expansive and occupation is adjacent to the oceanside reef flat, there is an increased emphasis on taxa occurring on the lower intertidal of the oceanside reef.

Environmental Context The southernmost atoll in the Marshall Islands, Ebon Atoll (4°38’N, 168°42’E) consists of 22 islets encircling a 104km2 lagoon (Figure 1). The total land area of Ebon is ~5.4km2, but the reef platform totals ~22km2 – >4:1 ratio of reef to land area. Ebon atoll is an ‘open’ atoll, with lagoon and ocean waters exchanged regularly through a pass across the reef platform. The macrobenthic fauna of ‘open’ atolls are generally highly diverse, but with low overall abundance (Paulay 2000:28). Around 1655 mollusc species occur in the Marshall Islands (Richmond et al. 2000:222), many of which can be gleaned from the intertidal marine habitats, evidenced by diverse archaeological mollusc assemblages from sites in the archipelago (Harris et al. 2015; Weisler 2001b). Finfish formed an important component of the diet in prehistory (Lambrides and Weisler 2015; Weisler 2001b:106-113) and about 860 fish taxa occur in the Marshall Islands (Richmond et al. 2000:222). Prehistoric terrestrial subsistence consisted of a mixed horticultural and arboricultural system of coconut (ni, Cocos nucifera), giant swamp taro (iraij, Cyrtosperma chamissonis), pandanus (bōb, Pandanus tectorius) arrowroot (makmōk, Tacca leontopetaloides) and breadfruit (mā, Artocarpus altilis).

The configuration of intertidal habitats on atolls is mediated by geological history, exposure to waves and currents, and the influence of the dominant north-east trade winds (Wiens 1962; Yamano et al. 2005), leading to a distinction on Ebon Atoll between the marine environments of the windward north-eastern extent and the sheltered south-eastern leeward areas (Figure 2). Further

99 distinctions exist between portions of the reef that face the open ocean (oceanside, jablik) and those which encircle the lagoon (lagoonside, jabar). Merlin et al. (1994) note that Marshallese divide the marine environment into the reefs that face the open ocean (baal), reef flat (ioon-pedped-lik), beachrocks at the subtidal fringe (pedpedilero), sandy shoreline (ioon-ippe-lik), lagoonside shoreline referred to as ioon-ippe-iaar and the lagoonside intertidal and sublittoral called lo̧ iem. The deep, interior portion of the lagoon is known as lo̧ kilmeejej. These areas correspond to the zones defined in the habitat classification scheme discussed below and broadly follow zonation and habitat conceptions of intertidal foragers on other Micronesian atolls (Thomas 2007). On Ebon, the exposed, windward north-eastern islets are generally smaller and intertidal marine environments facing the open ocean are composed of poorly sorted coral rubble washed from the subtidal reefs with coarse, gravelly sands on the lagoonside. The larger leeward islets feature high-rugosity coral reefs, inter-reefal sand flats and seagrass beds in the calm lagoonside waters and low relief, expansive reef flat pavements on the oceanside.

Moniak islet Moniak (~0.7 km2) lies on the north-eastern rim of Ebon Atoll (Figure 1). The long axis of the islet, perpendicular to the north-easterly trade winds, is approximately 350m long and 230m wide at its broadest point. The terrestrial zone of the islet reflects its exposed location, sloping upward from the lagoonside to a natural rampart built of coral cobbles and boulders deposited by waves (Wiens 1962:333; Figure 2b). The Pisonia tree dominated interior zone hosts large colonies of terns (Sternidae) and boobies (Sula spp.). Coconut crabs (Bigrus latro) and smaller crabs are found in greater numbers on Moniak than on inhabited islets. It is likely that these attractive resources drew prehistoric foragers to the islet as they continue to do today.

The oceanside reef edge of Moniak consists primarily of wave cut erosional channels known as spur and groove formations. On oceanic atolls, the ridges of the channels host communities of corals, crabs, urchins and molluscs such as Turbo, Drupa and large cypraeids (Morrison 1954; Odum and Odum 1955). In the groove between the spurs finfish including hawkfish (Cirrithidae), blennies (Blenniidae), parrotfish (Scaridae) and surgeonfish (Acanthuridae) are found, many of which can be speared from the ridges at low tide, (Harry 1953:22; Morrison 1954; Wiens 1962:Plate 42). The deeper, oceanward portions of the channels tend to be dangerous places for foraging and fishing due to the presence of sharks and strong waves (Harry 1953:22).

100

Shoreward of the spur and groove zone is a calcareous algal ridge that provides a habitat for molluscs such as the rough turban (Turbo setosus), Trochus and Drupa. Exposed, wave-swept pavements scattered with coral cobbles and reef rubble lie shoreward of the algal ridge. These pavements are often colonised by macroalgae and host populations of molluscs including herbivorous Cerithiidae, Cypraeidae and Trochidae and the predatory molluscs Conidae and Muricidae, characteristic of windward reef pavements (Leviten and Kohn 1980). The shoreline and supratidal zones of Moniak are primarily coral boulders and cobbles deposited during severe storms. These environments are easily accessed at low and high tide and provide refuge from desiccation for crabs and molluscs, such as Bursa spp, Thais spp., Vasum spp. and Nerita polita. The lagoon-facing beach is composed of coarse grained sand, and is a habitat for the bivalves Gafrarium spp. and Asaphis violascens.

Ebon islet Ebon islet is situated on the south-eastern rim of Ebon Atoll, spanning a length of roughly 12km from the north-eastern tip to the south-western Ruby Point. The land area of ~2.2km2 is over 40% of the total land area of the atoll. The spur and groove zone on the leeward oceanside is sheltered from exposure to waves, wind and storms and access to the reef edge for spearing, netting and collection of large Cypraeids, Turbo and urchins is usually safe during low tide, even at night (Personal observations, 12/29/2011). Reef flat pavements on Ebon islet are expansive and relatively free of debris compared to Moniak. Large colonies of macroalgae and shallow sandy tide pools host a wide range of mollusc species, hosting communities of Cypraeidae, Muricidae, Trochidae, Conidae and Neritidae, among others. The oceanward shoreline is composed of comparatively finer grained sand than Moniak, with beachrock outcrops demarcating previous shorelines. Gastropods such as the nerites (Neritidae), littorinids (Littorinidae), hollow shelled snails (Ellobidae) and clusterwinks (Planxinae) inhabit the shoreline fringe including the splash zone and upper intertidal.

The lagoonside is a mosaic of coral reefs, inter-reefal sand flats, turtle grass (Thalassia hemprichii) beds, coral rubble, and smooth, eroded pavements. With an increase in habitat complexity compared to Moniak islet, a general increase in the richness and diversity of mollusc communities is predicted (Gratwicke and Speight 2005; Kohn and Leviten 1976). The lagoonside sands offer habitat for the bivalves Asaphis violascens, Gafrarium spp., Vasticardium spp., Arca spp. and Atactodea striata. Giant clams (Tridacna spp. and Hippopus sp.), spider conchs (Lambis spp. and Harpago spp.) cone shells (Conus spp.), mitre shells and auger snails (Terebra spp. and members of the family Mitridae) occur in the lagoonside habitats, though current abundance and distribution is 101 not known. In addition to food, the shells were traditionally used as raw material for making a range of tools and personal ornaments (Kramer and Nevermann 1938; Weisler 2000, 2001b).

Sites and Samples MLEb-1 is located centrally within a ~2km long village system defined by seven sites along Ebon islet, lagoonward of the largest horticultural pit system on the atoll (Weisler 1999a, 2002). TP17, 18, 19 and 20 were excavated into a low mound developed by successive coral pavements (living floors), located 40m inland of the current lagoon shore and 20m northwest of the primary school. Cultural deposits consisted of dense coral pebbles and coarse sand extending to a depth of 1.75m (Weisler 1999a:Figure 4; 2002:20) including molluscan remains, fishbone and worked shell artefacts. A total of 4953 fragments (6.39kg) of molluscan shell were retained and analysed from the 6mm wet-screened material, with all but 217 fragments (4% total NISP) identifiable to family, genus or species. The assemblage is dominated by gastropods—65% of NISP and 89% of MNI.

MLEb-31 is located ~65m from the current lagoon shore on the north-eastern islet of Moniak. Cultural deposits consist of mixed coral cobbles and poorly sorted sand extending to a maximum depth of 70cm. No coral living floors were encountered, suggesting prehistoric use of Moniak as a temporary camp during sorties to the windward side of the atoll. A total of 1891 fragments (7.33kg) of mollusc shell was retained and analysed from the 6mm wet-screened material, with all but 77 fragments (3%) identifiable to family, genus or species. The assemblage is dominated by gastropod taxa in both NISP (78%) and MNI (85%).

Methods of mollusc identification and quantification All molluscan remains were identified to the lowest possible taxon using books, identification manuals (Carpenter and Niem 1998; Cernohorsky 1967; Poppe 2008) and Indo-Pacific molluscan reference collections held at the University of Queensland. Due to the richness of atoll molluscan fauna (Kay and Johnson 1987) and Indo-Pacific Mollusca generally (Bouchet et al. 2002:421), taxa were identified based only on diagnostic features present on individual fragments, rather than assuming fragments derive from dominant taxa (after Driver 2011; Szabó 2009; Wolverton 2013). Taxonomic nomenclature was verified using the World Register of Marine Species online database (WoRMS Editorial Board 2015). This approach to identification results in lower resolution foraging reconstructions due to fewer species-level identifications, yet is preferable due to errors introduced through over-identification.

102

NISP was recorded for all taxa. Weight was recorded to the nearest 0.01g. MNI is used as the primary measure of relative abundance, calculated according to the protocols outlined in Harris et al. (2015). For MLEb-1, TP17-20, MNI was aggregated by stratigraphic layer. NRE (Non- Repetitive Elements such as the spire or the siphonal notch of gastropods, or posterior and anterior adductor muscle scars of bivalves) frequency was summed for each stratigraphic layer for all test pits prior to the calculation of MNI. Due to the small sample size of molluscan assemblages from individual test pits at MLEb-31, samples from TP2-6 (essentially from the same cultural layer) were combined to increase sample size. NRE frequency was summed prior to calculation of MNI for each taxon in each layer. The utility of MNI as a primary measure for archaeomalacological analyses has been discussed in detail elsewhere (Claassen 1998; Giovas 2009; Harris et al. 2015; Mason et al. 1998; Szabó 2009) and it is critical that any inferences based on MNI are considered in light of NISP and weight data to assess the influence of post-depositional factors on measures of relative abundance. The relative abundance (MNI) of molluscan taxa in the archaeological assemblages is used here as a proxy for time-averaged patterning of forager decision making regarding exploitation of molluscs from the intertidal reef habitats of Ebon Atoll.

Methodological framework Reconstructing foraging preferences Reconstructing marine mollusc foraging preferences can inform on subsistence systems, settlement patterns and, perhaps, land tenure and broader socio-cultural relations. However, as Allen (1992:331) notes, reconstructing foraging in this way reveals probabilistic (rather than deterministic) relationships between the taxa in the assemblage and forager decision making. More explicit, deterministic inferences can be made using oxygen isotopes and sclerochronology (e.g., Andrus and Thompson 2012); however, the methodology used here relies on ecological knowledge of the preferences of particular molluscan taxa for configurations of water depth, tidal exposures, substrate type and associated biological cover. These variables are fundamental when determining the range of molluscan taxa foragers could have potentially exploited in a particular habitat. Distinction in these factors between leeward and windward islets (Figure 2) make them useful for investigating differences in forager decision making relating to local environment and settlement patterns.

Allen (1992:331) cited three potential issues with investigating foraging in this manner: (1) the specificity of faunal identifications, (2) the quality of ecological information available for each taxon and (3) the depth of information pertaining to traditional or prehistoric extraction 103 technologies and the articulation of this knowledge with the quality of information pertaining to prey behaviour. Identification and quantification of faunal material has been discussed elsewhere (Claassen 1998, 2000; Harris et al. 2015; Lyman 2008; Mason et al. 1998) and issues relating to the quality of ecological data are covered below. However, little is known about mollusc foraging technology during Marshall Islands prehistory. While there are at least 64 Marshallese words for different fishing methods (Abo et al. 1976:361-362), there are no words specifically for the collection of shellfish. However, there are at least 45 Marshallese words for particular shell taxa (Abo et al. 1976:453-454), salted clams (jiookra), the meaty party of the clam (aḷaḷ), the rod-like style from the digestive system of a Tridacna clam (lām), particular places where shells, fish, birds or clams gather, (ajañ), or sickness resulting from consumption of surf clams (kọnet, Atactodea striata) (Abo et al. 1976). Based on archaeological evidence from Pacific island archaeological sites (e.g., Allen 2012; Morrison and Addison 2008; Szabó 2009; Szabó and Amesbury 2011), modern studies of subsistence harvesting of shellfish and ethnographic data (e.g., Kinch 2003; Malm 2009; Thomas 2001; Titcomb 1978), the ease of accessing the intertidal zone of the reef on the ocean and lagoon sides of Ebon Atoll, and the authors’ observations of shellfish gathering in the Marshall Islands, we hypothesise that most shellfish gathering would likely have occurred by hand while walking or wading along the intertidal and shallow subtidal at low tide in the areas extending from the shoreline, out across the reef flat, to the reef edge.

Hierarchical classification scheme In order to assess forager decision making, a framework for assigning molluscan taxa to ecological zones must be utilised. A hierarchical classification scheme is an objective, systematic approach to benthic habitat classification and description (Mumby and Harborne 1999b:155) and is primarily utilised by researchers mapping the spatial distribution of benthic habitats for conservation and management of tropical marine environments (Kendall et al. 2012; Mumby and Harborne 1999a). Hierarchical classification schemes partition the marine environment by describing and classifying the location of discrete zones of the benthos. A habitat is defined in these schemes as a unique configuration of: (1) location in relation to the shoreline, (2) geomorphological structure and (3) benthic cover. This method of classifying the marine environment is adapted here to assign molluscan taxa to particular combinations of these three environmental categories to use in reconstructing forager decision making.

The habitat classification scheme used here was developed by Kendall et al. (2012) for mapping the marine environment of Majuro Atoll, Marshall Islands. Kendall et al. (2012) employed a simple 104 habitat classification scheme (in contrast to Mumby and Harborne 1999a) that omits detailed attributes such as the percentage of hard bottom, relative patchiness of seagrass cover and the type and/or species of coral cover. The modified classification scheme used here has only two hierarchical attributes: Zone and Geomorphological Structure (Table 2). Geomorphological structure consists of two levels of sub-attributes: Major Geomorphological Structure and Detailed Geomorphological Structure. This classification scheme incorporates aspects of marine environments critical to molluscs, including the relationship with the shoreline, tides and water depth (Zone), substrate type (Major Geomorphological Structure) and associated biological cover (Detailed Geomorphological Structure).

Data collection for molluscan zonation and ecology A range of sources on Indo-Pacific molluscan fauna were consulted with a focus on atoll environments, including field observations and surveys (e.g., Banner and Randall 1952; Bernstein 1974; Houk and Musburger 2013; Kay and Johnson 1987; Morrison 1954; Paulay 2000), reference collections of molluscs from Micronesia (Demond 1957), studies of particular species (e.g., Abbott 1960; Kohn 1980; Taylor 1983, 1984) and identification guides (principal sources include: Burgess 1985; Carpenter and Niem 1998; Röckel et al. 1995).

All taxa recorded from the Ebon Atoll archaeological mollusc assemblages were assigned a number of habitats using the hierarchical classification scheme outlined above. A total of 251 bivalve and gastropod taxa from 42 families and 111 genera were assigned a combination of Zone, Major Geomorphological Structure and Detailed Geomorphological Structure as determined by the habitat preference(s) of each taxon. Reliable information relating to habitat preferences could not be located for the following taxa: Pitar striatus, Ranularia testidunaria, Gyrineum pusillum, Chama gryphoides, Plicopura spp., Nerita exuvia and Neritopsis radula. However, these taxa are uncommon in archaeological sites on Ebon Atoll, totalling 8 NISP from both sites.

Molluscan ecology and habitat classification scheme usage MNI values for taxa associated with a particular habitat were summed. The total MNI for all taxa from each habitat was compared to total MNI for the assemblage and converted to a percentage. These values are used as a proxy for human decision making, indicating those habitats from which the taxa in the assemblage could have derived. By using site-level totals for each site, this method allows a time-averaged assessment of the areas of the marine environment from which the majority of the taxa present in the assemblage could have been collected. In the methodology adopted here 105 no assumptions are made of the intentions of prehistoric foragers. Dominant taxa are given equal analytical weight to non-dominant taxa. The analytical inclusion of these non-dominant taxa which often together account for a large portion of total MNI, highlight broad trends in foraging practices that may be masked when taxa are deemed non-economic or unimportant.

For specimens identified to family or genus, the relevant habitat information from species level entries were aggregated and, where necessary, additional data was added for genera or species not recorded in the assemblage (for example, the Conidae and Cypraeidae) based on the range of taxa recorded for the Marshall Islands (Demond 1957; Kay and Johnson 1987). These aggregated habitat designations capture the intrafamily variation in habitat preferences and the uncertainty associated with family and genus level identifications. These lower level identifications hinder high-resolution foraging reconstructions, but the sacrifice in analytical resolution minimises errors introduced through over-identification.

Care was taken to assign taxa to all appropriate habitats, however, molluscan habitat preference is a dynamic process influenced locally by a range of biological, ecological and climatic factors and can alter markedly over short periods of time (Augustin et al. 1999:Table 2; Paulay 2000). Those taxa which may have been processed off-site, or did not preserve will be invisible in foraging reconstructions (Bird and Bliege Bird 1997). In many cases taxa are assigned to multiple habitats, introducing the potential for over-representation of habitat types when numerically dominant taxa are recorded for multiple habitats. However, the foraging patterns presented in the results should be treated as probabilistic, rather than deterministic; this method highlights the habitats from which the majority of the taxa in the assemblage could have been gathered, and can provide useful data for testing hypotheses relating to the articulation of molluscan foraging with other aspects of human behaviour.

Results MLEB-1 The molluscan assemblage from MLEb-1 is both taxonomically rich and even, represented by 34 families, 62 genera and 63 species (NTAXA = 48, total MNI = 1258, Figure 3). Taxonomic measures of heterogeneity reported an even assemblage as measured by Simpson's index of diversity (1-D = 0.889), the Shannon-Weiner index of diversity (H’ = 2.668) and Shannon's evenness (E = 0.689). The majority of taxa are epifaunal gastropods, with few infaunal bivalves in the assemblage. 106

Most of the assemblage could have been gathered from the reef flat pavements (D/1/15, D/1/16), which today are most prevalent on the oceanside of Ebon islet (Figure 4a).This single habitat accounts for 58.1% of the individuals in the assemblage (Figure 5). Many of the top-ranked taxa in the assemblage are highly associated with this habitat on Micronesian atolls, including N. plicata, V. turbinellus, M. moneta, T. maculatus, M. intermedius and Drupa spp. (Demond 1957). Even when the influence of taxa that can occur in multiple habitats are excluded from foraging reconstructions, the oceanside reef flat pavement still accounts for 45.7% of MNI (Figure 4b). A large portion of the assemblage could have also been gathered from rocky areas of the reef flat. This habitat accounts for 41.8% of total MNI, represented primarily by N. polita and N. sinensis.

There is less evidence for foraging in the lagoonside coral reefs, sand flats and seagrass beds. G. muricinum and Monoplex intermedius are common in the assemblage (15% of total MNI), and taxa such as Tridacna spp., Lambis lambis and Harpago chiragra also indicate lagoonside collection, although collection of these taxa is minimal (c.1% of total MNI).

MLEB-31 Some 31 families, 53 genera and 61 species (NTAXA = 37, MNI = 650) were recovered from archaeological deposits at MLEb-31 (Figure 6). Simpson's index of diversity (1-D = 0.859), the Shannon-Weiner index of diversity (H’ = 2.528) and Shannon's evenness (E = 0.7) reported an even and diverse assemblage.

The greatest number of individuals in the assemblage could have been gathered from the reef flat cobbles and boulders, currently circumscribing the northern, western and eastern sides of Moniak islet and characteristic of windward islets (Wiens 1962; Figure 7a). Taxa characteristic of this habitat on windward islets (N. polita, T. armigera) contribute 31% of total MNI. Reef flat rocks account for 56.8% of the individuals in the assemblage. Once again, when the MNI of taxa that can inhabit multiple habitats are removed from quantification, these trends still hold (Figure 7b). Reef flat pavements are also highly ranked with V. turbinellus, C. nodulosum, T. maculatus and other taxa found commonly on pavements in Micronesia, accounting for 36.3% of total MNI. Gathering from sand flats and coral growth is also indicated primarily by the presence of A. violascens and G. muricinum, with these two taxa accounting for 13% of total MNI.

107

Discussion The molluscan assemblages from MLEb-1 and MLEb-31 indicate a generalist foraging strategy, exhibiting rich and even assemblages with the major focus on collecting small gastropods from the high intertidal and reef flat. Minor differences in taxonomic composition were noted, likely reflecting local availability of taxa on windward and leeward islets. The taxonomic composition of both assemblages suggests that foragers were not targeting particular taxa, but rather focusing foraging efforts on particular habitats, exploiting the predictable behaviours the taxa hosted there. The range of habitats, which is typically higher on leeward islets (Kendall et al. 2012), and taxa exploited at both sites seems to reflect natural variation in the habitats adjacent to the archaeological deposits. Both assemblages suggest an overall emphasis on gleaning, with the majority of the taxa living epifaunally. However, most bivalves in both assemblages live infaunally, indicating that foragers were also extracting molluscs by digging.

The predominance of Nerita at both sites suggests that these small taxa were an attractive resource for prehistoric Marshallese. N. polita, the top ranked taxa at MLEb-31, burrows into the sand between 5-10cm deep at the base of boulders and cobbles during high tides and diurnal low water. The shallow burial depth characteristic of this species requires minimal effort from foragers collecting this resource. During periods when the zone is free of water during nocturnal low tides, N. polita emerges from the sand to mate and feed on macroalgae and returns to the sand before inundation at high tide (Chelazzi 1982:453). Patterns of burrowing, foraging and emersion tend to be highly predictable, with emersion following in the wake of the retreating tide and re-burial occurring between 30 minutes and two hours prior to inundation from rising tides (Chelazzi 1982:454). N. polita also tend to re-bury close to the location of emersion. The relationship between the activity of the tides, the behaviour of N. polita and predictable burial and emersion location, would have allowed foragers easy access to this species for the entire low tide. Furthermore, large concentrations of individuals can occur on single boulders or cobbles (Chelazzi 1982), facilitating mass harvesting. Predictable patterns of behaviour are not unique to the nerites and may help to explain the focus on oceanside reef flat environments on Ebon islet.

When atoll reef flats are aerially exposed during daytime low tides, surface temperatures commonly reach as high as 38°C (Kohn 1987:139). Due to the surface temperature of the expansive reef flat at Ebon islet during daytime aerial exposure, molluscs must take refuge in tide pools to avoid desiccation (Russell and Phillips 2009:71). The vast majority of the gastropod taxa present in the MLEb-1 assemblage, including the top-ranked N. plicata, live epifaunally, either grazing on algal

108 turfs or actively preying on other molluscan taxa. Like the predictable concentrations of N. polita below and on the surface of cobbles and boulders, the predictable behaviours of molluscs sheltering from desiccation, browsing on algae, or preying on other molluscs (Kohn 1983; Taylor 1978, 1983) were likely targeted by prehistoric foragers.

While there is a strong indication of generalised gleaning from the oceanside intertidal at both MLEb-1 and MLEb-31, it is likely that foragers at MLEb-1 also exploited lagoonside habitats. The knobbly triton, G. muricinum, is the second highest ranked species at MLEb-1. While the behaviour of this species is not well understood, Govan (1995) noted that G. muricinum occurs in almost all lagoonside habitats, including sand flats, coral reefs and the inter-reefal sand flats. Furthermore, A. violascens is ranked sixth by MNI at MLEb-1 and ranked two at MLEb-31. A. violascens burrows relatively deep, with peak densities reported between 13 and 20cm (Soemodihardjo and Matsukuma 1989), preferring gravelly sands in sheltered lagoon habitats (Paulay 2000). Modern observations of mollusc collecting on Tarawa, Kiribati have noted that due to A. violascens preference for the mid- high intertidal beach slope, this species is accessible “even on the worst tides” (Paulay 2000:25). This species is often the dominant bivalve in its habitat, located easily by the tell-tale siphonal opening in the sand, and dug out (Paulay 2000:13). This combination of ecological factors likely made A. violascens a dependable and attractive resource for prehistoric foragers.

It is difficult to draw inferences from presence/absence data alone, but comparisons with mollusc assemblages from Arno Atoll (Dye 1987) seem to indicate exploitation of a similar range of taxa to assemblages from Ebon Atoll. There is some indication that foragers on Arno were focusing efforts on the lagoonside, but this assertion is tenuous given that the relative abundance of taxa cannot be adequately assessed and many identifications were only to the family or genus level. Converse to mollusc assemblages from Utrōk Atoll to the far north (Weisler 2001a, b), where lagoonside gathering seems to be the focus, Ebon Atoll assemblages tend to be gathered primarily from oceanside habitats, with a minor component of lagoonside foraging. These patterns may be attributable to differences in quantification measures, but may indicate genuine variation in foraging practices between northern and southern atolls.

A notable difference occurs between assemblages recovered from Transect 6 and 7 at Laura Village, Majuro (Riley 1987) and MLEb-1 on Ebon islet. Both sites were major villages, with permanent human populations and a well-developed zone of aroid pit horticulture. However, the molluscan assemblages from sites at Laura village are dominated by the remains of the strawberry

109 conch, C. luhuanus. This species was not reported from deposits at MLEb-1 and the molluscan assemblages from MLEb-1 or MLEb-31 do not indicate a pattern of intensive exploitation of any single taxon, or reach the densities of Majuro deposits. The dominance and density of C. luhuanus remains suggests that prehistoric foraging at Laura village was intensively focused on mass collection of a single taxon, more similar to foraging patterns recorded for modern atoll dwellers (Thomas 2014), than prehistoric assemblages from Ebon Atoll. A lack of suitable habitats or other biotic or ecological factors may have inhibited the establishment of C. luhuanus colonies on Ebon and Moniak islets. During the prehistoric period, the inhabitants of Ebon seem to have focused on generalised gleaning of molluscs primarily from the oceanside reef flat and shoreline rocks.

Conclusion An analysis of prehistoric foraging preferences of the inhabitants of two archaeological sites has shown that on Ebon Atoll, Marshall Islands, foragers practiced a generalised collection strategy focused on a rich assemblage of primarily gastropod taxa from the upper intertidal rocks, sand and cobbles, and the oceanside reef flat pavement. Foragers likely relied on predictable patterns of molluscan behaviour to harvest a wide range of taxa, rather than focusing foraging effort on any individual taxon. Minor differences in taxonomic composition were noted, likely reflecting local availability of taxa on windward and leeward islets. It is possible that this generalised foraging pattern is related to other forms of marine subsistence, such as collecting edible seaweed or the exploitation of inshore finfish. Further study will seek to understand how these patterns of molluscan exploitation relate to the exploitation of finfish (Harris et al. 2016), shell working, manufacture and curation of shell tools and atoll-wide settlement patterns. Ultimately, these data will be used to investigate potential human impacts to the marine environments of Ebon Atoll that may have been related to mollusc exploitation. The low-intensity, generalised collection strategy inferred from an analysis of molluscan taxa in the assemblages may have allowed sustained yields of molluscs from a range of environments for the entire period of human occupation, beginning ~2000 years ago.

110 university studies are supported by an Australian Postgraduate Award. Marine molluscs collected during field work have been returned to the HPO, Marshall Islands.

111

References Cited Abbott, R.T. 1960. The genus Strombus in the Indo-pacific. Indo-Pacific Mollusca 1(2):33-146. Abo, T., B.W. Bender, A. Capelle and T. DeBrum. 1976. Marshallese-English Dictionary. University of Hawai'i Press, Honolulu. Allen, M.S. 1992. Dynamic Landscapes and Human Subsistence: Archaeological investigations on Aitutaki Islands, Southern Cook Islands. Ph.D. Dissertation. Washington: University of Washington. Allen, M.S. 2012. Molluscan foraging efficiency and patterns of mobility amongst foraging agriculturalists: a case study from northern New Zealand. Journal of Archaeological Science 39(2):295-307. Amesbury, J.R. 2007. Mollusk collecting and environmental change during the prehistoric period in the Mariana Islands. Coral Reefs 26:947-958. Andrus, C.F.T. and V.D. Thompson. 2012. Determining the habitats of mollusk collection at the Sapelo Island shell ring complex, Georgia, USA using oxygen isotope sclerochronology. Journal of Archaeological Science 39(2):215-228. Aswani, S., C.F. Flores and B.R. Broitman. 2015. Human harvesting impacts on managed areas: ecological effects of socially-compatible shellfish reserves. Reviews in Fish Biology and Fisheries 25:217-230. Augustin, D., G. Richard and B. Salvat. 1999. Long-term variation in mollusc assemblages on a coral reef, Moorea, French Polynesia. Coral Reefs 18(3):293-296. Bailey, G.N. 1975. The role of molluscs in coastal economies: the results of midden analysis in Australia. Journal of Archaeological Science 2(1): 45-62. Bailey, G., J. Chappell and R. Cribb. 1994. The origin of Anadara shell mounds at Weipa, North Queensland, Australia. Archaeology in Oceania: 69-80. Bailey, G. and N. Milner. 2008. Molluscan archives from European prehistory. In Early Human Impact on Megamolluscs (A. Antczak and R. Cipriani, eds.): 111-134. Oxford: Archaeopress (BAR International Series). Banner, A.H. and J.E. Randall. 1952. Preliminary report on marine biology study of Onotoa Atoll, Gilbert Islands. Atoll Research Bulletin 13:1-62. Baron, J. 1992. Reproductive cycles of the bivalve molluscs Atactodea striata (Gmelin), Gafrarium tumidum (Roding) and Anadara scapha (L.) in New Caledonia. Australian Journal of Marine and Freshwater Resources 43:393-402.

112

Baron, J. and J. Clavier. 1992. Effects of environmental factors on the distribution of the edible bivalves Atactodea striata, Gafrarium tumidum and Anadara scapha on the coast of New Caledonia (SW Pacific). Aquatic Living Resources 5:107-114. Bernstein, A.S. 1974. Diet and Competition for Food Among the Predatory Gastropods of Limestone Benches in Hawaii and Eniwetok. Ph.D. Dissertation, Oregon: University of Oregon. Bird, D.W. and R.L. Bliege Bird. 1997. Contemporary shellfish gathering strategies among the Meriam of the Torres Strait Islands, Australia: testing predictions of a central place foraging model. Journal of Archaeological Science 24(1):39-63. Bouchet, P., P. Lozouet, P. Maestrati and V. Heros. 2002. Assessing the magnitude of species richness in tropical marine environments: exceptionally high numbers of molluscs at a New Caledonia site. Biological Journal of the Linnean Society 75(4):421-436. Braje, T.J., D.J. Kennett, M.E. Jon and B.J. Culleton. 2007. Human impacts on nearshore shellfish Taxa: a 7,000 year record from Santa Rosa Island, California. American Antiquity 72(4): 735-756. Burgess, C.M. 1985. Cowries of the World. Cape Town: Seacomber Publications. Carpenter, K.E. and V.H. Niem (eds) 1998. The Living Marine Resources of the Western Central Pacific: Volume 1. Seaweeds, Corals, Bivalves and Gastropods. FAO Species Identification Guide For Fishery Purposes. Rome: Food and Agriculture Organization of the United Nations. Cernohorsky, W.O. 1967. Marine Shells of the Pacific. Sydney: Pacific Publications. Chelazzi, G. 1982. Behavioural adaptation of the gastropod Nerita polita L. on different shores at Aldabra Atoll. Proceedings of the Royal Society of London. Series B. Biological Sciences 215(1201):451-467. Christensen, C.C. and M.I. Weisler. 2013. Land snails from archaeological sites in the Marshall Islands, with remarks on prehistoric translocations in tropical Oceania. Pacific Science 67(1):81-104. Claassen, C. 1998. Shells. Cambridge: Cambridge University Press. Claassen, C. 2000. Quantifying shell: comments on Mason, Peterson, and Tiffany. American Antiquity 65(2):415-418. de Boer, W.F., A.F. Blijdenstein and F. Longamane. 2002. Prey choice and habitat use of people exploiting intertidal resources. Environmental Conservation 29(02): 238-252. Demond, J. 1957. Micronesian reef-associated gastropods. Pacific Science 11(3):275-341. Driver, J.C. 2011. Identification, classification and zooarchaeology. Ethnobiology Letters 2:19-39.

113

Dye, T. 1987. Report 3: Archaeological survey and test excavations on Arno Atoll, Marshall Islands. In Marshall Islands Archaeology (T. Dye, ed.): 271-394. Honolulu: Pacific Anthroplogical Records 38. Bernice Pauahi Bishop Museum. Erlandson, J.M., T.J. Braje, T.C. Rick, N.P. Jew, D.J. Kennett, N. Dwyer, A.F. Ainis, R.L. Vellanoweth and J. Watts. 2011. 10,000 years of human predation and size changes in the owl limpet (Lottia gigantea) on San Miguel Island, California. Journal of Archaeological Science 38(5): 1127-1134. Faulkner, P. 2013. Life on the Margins: An Archaeological Investigation of Late Holocene Economic Variability, Blue Mud Bay, Northern Australia, Terra Australis 38. Canberra: ANU E Press. Fitzpatrick, S.M. and W.F. Keegan. 2007. Human impacts and adaptations in the Caribbean Islands: an historical ecology approach. Transactions of the Royal Society of Edinburgh-Earth Sciences 98: 29-45. Fitzpatrick, S.M., M. Kappers, Q. Kaye, C.M. Giovas, M.J. LeFebvre, M.H. Harris, S. Burnett, J.A. Pavia, K. Marsaglia and J. Feathers. 2009. Pre-Columbian Settlements on Carriacou, West Indies. Journal of Field Archaeology 34(3): 247-266. Gaspar, M., P. DeBlasis, S. Fish and P. Fish. 2008. Sambaqui (Shell Mound) Societies of Coastal Brazil. In The Handbook of South American Archaeology (H. Silverman and W. Isbell, eds.): 319-335. New York: Springer. Giovas, C.M. 2009. The shell game: analytic problems in archaeological mollusc quantification. Journal of Archaeological Science 36(7):1557-1564. Giovas, C.M., M. Clark, S.M. Fitzpatrick and J. Stone. 2013. Intensifying collection and size increase of the tessellated nerite snail (Nerita tessellata) at the Coconut Walk site, Nevis, northern Lesser Antilles, AD 890–1440. Journal of Archaeological Science 40(11): 4024- 4038. Govan, H. 1995. Cymatium Muricinum and Other Ranellid Gastropods: Major Predators of Cultured Tridacnid Clams. Penang: WorldFish. Gratwicke, B. and M. Speight. 2005. The relationship between fish species richness, abundance and habitat complexity in a range of shallow tropical marine habitats. Journal of Fish Biology 66(3):650-667. Harris, M., A.B.J. Lambrides and M.I. Weisler. 2016. Windward vs. leeward: Inter-site variation in marine resource exploitation on Ebon Atoll, Republic of the Marshall Islands. Journal of Archaeological Science: Reports 6:221-229.

114

Harris, M., M.I. Weisler and P. Faulkner. 2015. A refined protocol for calculating MNI in archaeological molluscan shell assemblages: a Marshall Islands case study. Journal of Archaeological Science 57:168-179. Harry, R.R. 1953. Ichthyological field data of Raroia Atoll, Tuamotu Archipelago. Atoll Research Bulletin 18:1-190. Horrocks, M. and M.I. Weisler. 2006. Analysis of plant microfossils in archaeological deposits from two remote archipelagos: the Marshall Islands, eastern Micronesia, and the Pitcairn Group, southeast Polynesia. Pacific Science 60(2):261-280. Houk, P. and C. Musburger. 2013. Trophic interactions and ecological stability across coral reefs in the Marshall Islands. Marine Ecology Progress Series 488:23-34. Hunt, C.O., T.G. Reynolds, H.A. El-Rishi, A. Buzaian, E. Hill and G.W. Barker. 2011. Resource pressure and environmental change on the North African littoral: Epipalaeolithic to Roman gastropods from Cyrenaica, Libya. Quaternary International 244(1): 15-26. Jerardino, A. and R. Navarro. 2008. Shell morphometry of seven limpet species from coastal shell middens in southern Africa. Journal of Archaeological Science 35(4): 1023-1029. Jerardino, A. and C.W. Marean. 2010. Shellfish gathering, marine paleoecology and modern human behavior: perspectives from cave PP13B, Pinnacle Point, South Africa. Journal of Human Evolution 59(3–4): 412-424. Jiao, T. 2007. The Neolithic of Southeast China: Cultural Transformation and Regional Interaction on the Coast. Amherst: Cambria Press. Kay, E.A. and S. Johnson. 1987. Mollusca of Enewetak Atoll. In The Natural History of Enewetak Atoll: Volume II Biogeography and Systematics (D.M. Devaney, E.S. Reese, B.L. Burch and P. Helfrich, eds.): 105-146. Tennessee: Office of Scientific and Technical Information, U.S. Department of Energy. Kayanne, H., T. Yasukochi, T. Yamaguchi, H. Yamano and M. Yoneda. 2011. Rapid settlement of Majuro Atoll, central Pacific, following its emergence at 2000 years CalBP. Geophysical Research Letters 38(20):L20405. Kendall, M.S., T.A. Battista and C. Menza 2012 Majuro Atoll, Republic of the Marshall Islands Coral Reef Ecosystems Mapping Report. Maryland: NOAA National Centers for Coastal Ocean Science. Kinch, J. 2003. Marine mollusc use among the women of Brooker Island, Louisiage Archipelago, Papua New Guinea. SPC Women in Fisheries Information Bulletin 13:5-14. Kohn, A.J. 1980. Abundance, diversity, and resource use in an assemblage of Conus species in Enewetak Lagoon. Pacific Science 34(4):359-369.

115

Kohn, A.J. 1983. Microhabitat Factors Affecting Abundance and Diversity of Conus on Coral Reefs. Oecologia 60(3):293-301. Kohn, A.J. 1987. Intertidal Ecology of Enewetak Atoll. In The Natural History of Enewetak Atoll: Volume I The ecosystem: Environments, Biotas, and Processes (D.M. Devaney, E.S. Reese, B.L. Burch and P. Helfrich, eds.): 139-157. Tennessee: Office of Scientific and Technical Information, U.S. Department of Energy. Kohn, A.J. and P.J. Leviten. 1976. Effect of habitat complexity on population density and species richness in tropical intertidal predatory gastropod assemblages. Oecologia 25(3):199-210. Koike, H. 1975. The use of daily and annual growth lines of the clam Meretrix lusoria in estimating seasons of Jomon period shell gathering. Quaternary Studies. The Royal Society of New Zealand 13:189-193. Koike, H. 1986. Prehistoric hunting pressure and paleobiomass: an environmental reconstruction and archaeozoological analysis of a Jomon shellmound area. Prehistoric Hunter-Gatherers in Japan: New Research Methods 27: 27-53. Kramer, A. and H. Nevermann. 1938. Ralik-Ratak (Marshall-Inseln), G. Thilenius, Ergebnisse der Sudsee-Expedition 1908-1910. II. Ethnographie, B: Mikronesien. Hamburg: De Gruyter and Co. Lambrides, A.B.J. and M.I. Weisler. 2015. Applications of vertebral morphometrics in Pacific Island archaeological fishing studies. Archaeology in Oceania 50(2):53-70 Leviten, P.J. and A.J. Kohn. 1980. Microhabitat resource use, activity patterns, and episodic catastrophe: Conus on tropical intertidal reef rock benches. Ecological Monographs 50(1):55-75. Lyman, R.L. 2008. Quantitative Paleozoology. Cambridge: Cambridge University Press. Malm, T. 2009. Women of the coral gardens: The significance of marine gathering in Tonga. SPC Traditional Marine Resource Management and Knowledge Information Bulletin 25:2-15. Mason, R.D., M.L. Peterson and J.A. Tiffany. 1998. Weighing vs. counting: measurement reliability and the California school of midden analysis. American Antiquity 63(2):303-324. Merlin, M., A. Capelle, T. Keene, J. Juvik and J. Maragos. 1994. Plants and Environments of the Marshall Islands. Honolulu: East West Center. Morrison, A.E. and D.J. Addison. 2008. Assessing the role of climate change and human predation on marine resources at the Fatu-ma-Futi site, Tutuila Island, American Samoa: an agent based model. Archaeology in Oceania 42(4):22-34. Morrison, A.E. and T.L. Hunt. 2007. Human impacts on the nearshore environment: an archaeological case study from Kaua'i, Hawaiian Islands. Pacific Science 61(3):325-345.

116

Morrison, J.P.E. 1954. Ecological notes on the mollusks and other animals of Raroia. Atoll Research Bulletin 34:1-18. Mumby, P.J. and A.R. Harborne 1999a Classification Scheme for Marine Habitats of Belize, Report to the UNDP/GEF Belize Coastal Zone Management Project. Washington, D.C.: Global Environment Facility Mumby, P.J. and A.R. Harborne. 1999b. Development of a systematic classification scheme of marine habitats to facilitate regional management and mapping of Caribbean coral reefs. Biological Conservation 88(2):155-163. Nagaoka, L. 2002. The effects of resource depression on foraging efficiency, diet breadth, and patch use in southern New Zealand. Journal of Anthropological Archaeology 21(4):419-442. Nelson, N.C. 1909. Shellmounds of the San Francisco Bay Region. University of California Publications in American Archaeology and Ethnology 7: 309-356. O'Dea, A., M.L. Shaffer, D.R. Doughty, T.A. Wake and F.A. Rodriguez. 2014. Evidence of size- selective evolution in the fighting conch from prehistoric subsistence harvesting. Proceedings of the Royal Society of London B: Biological Sciences 281(1782):20140159. Odum, H.T. and E.P. Odum. 1955. Trophic structure and productivity of a windward coral reef community on Eniwetok Atoll. Ecological Monographs 25(3):291-320. Parkington, J., J.W. Fisher and K. Kyriacou. 2013. Limpet gathering strategies in the later Stone Age along the Cape West Coast, South Africa. The Journal of Island and Coastal Archaeology 8(1): 91-107. Paulay, G. 2000. Benthic ecology and biota of Tarawa Atoll lagoon: influence of equatorial upwelling, circulation, and human predation. Atoll Research Bulletin 487:1-41. Poppe, G.T. 2008. Philippine Marine Mollusks. Hackenheim: ConchBooks. Poteate, A.S., S.M. Fitzpatrick, M. Clark and J.H. Stone. 2014. Intensified mollusk exploitation on Nevis (West Indies) reveals ~six centuries of sustainable exploitation. Archaeological and Anthropological Sciences 7(3): 361-374. Pregill, G.K. and M.I. Weisler. 2007. Lizards from prehistoric sites on Ebon Atoll, Marshall Islands. Micronesica 39(2):107-115. Richmond, R., R. Kelty, P. Craig, C. Emaurois, A. Green, C. Birkeland, G. Davis, A. Edward, Y. Golbuu, J. Gutierrez, P. Houk, N. Idechong, J. Maragos, G. Paulay, J. Starmer, A. Tafileichig, M. Trianni and N.V. Velde. 2000. Status of coral Reefs of American Samoa and Micronesia: US-affiliated Freely Associated Islands of the Pacific. . In Status of Coral Reefs of the World: 2000 (C. Wilkerson, ed.): 217-236. Canberra: Australian Institute of Marine Science.

117

Riley, T.J. 1987. Report 2: Archaeological survey and testing, Majuro Atoll, Marshall Islands. In Marshall Islands Archaeology (T. Dye, ed.): 169-270. Honolulu: Pacific Anthroplogical Records. Bernice Pauahi Bishop Museum. Röckel, D., W. Korn and A.J. Kohn. 1995. Manual of the Living Conidae. Hackenheim: Verlag Christa Hemmen. Rosendahl, P.H. 1987. Report 1: Archaeology in Eastern Micronesia: Reconnaissance Survey in the Marshall Islands. In Marshall Islands Archaeology (T. Dye, ed.): 17-166. Honolulu: Pacific Anthroplogical Records. Bernice Pauahi Bishop Museum. Russell, J. and N. Phillips. 2009. Species-specific vulnerability of benthic marine embryos of congeneric snails (Haminoea spp.) to ultraviolet radiation and other intertidal stressors. The Biological Bulletin 217(1):65-72. Shun, K. and J.S. Athens. 1990. Archaeological investigations on Kwajalein Atoll, Marshall Islands, Micronesia. Micronesica Supplement 2:231-240. Soemodihardjo, S. and A. Matsukuma. 1989. Ecology of sandy beach bivalves of Pari Island off the coast of Jakarta Bay, Indonesia. Bulletin Of The National Science Museum Series A (Zoology) 15:197-212. Sommerville-Ryan, G. 1998. The Taphonomy of a Marshall Islands' Shell Midden. M.A. thesis, Dunedin: University of Otago. Spennemann, D.H.R. 1987. Availability of shellfish resources on prehistoric Tongatapu, Tonga: effects of human predation and changing environment. Archaeology in Oceania 22(3):81- 96. Spennemann, D.H.R. 1993. Cowrie shell tools: fact or fiction? Archaeology in Oceania 28(1):40- 49. Spennemann, D.H.R. 1999. No room for the dead. burial practices in a constrained environment. Anthropos 94:35-56. Streck, C.S. 1990. Prehistoric settlement in eastern Micronesia: archaeology on Bikini Atoll, Republic of the Marshall Islands. Micronesica Supplement 2:247-260. Szabó, K. 2009. Molluscan remains from Fiji. In The Early Prehistory of Fiji, Terra Australis 31 (G. Clark and A.J. Anderson, eds.): 183-212. Canberra: ANU E Press. Szabó, K. and J.R. Amesbury. 2011. Molluscs in a world of islands: The use of shellfish as a food resource in the tropical island Asia-Pacific region. Quaternary International 239(1–2):8-18. Taylor, J.D. 1978. Habitats and diet of predatory gastropods at Addu Atoll, Maldives. Journal of Experimental Marine Biology and Ecology 31(1):83-103.

118

Taylor, J.D. 1983. The food of coral-reef Drupa (Gastropoda). Zoological Journal of the Linnean Society 78(4):299-316. Taylor, J.D. 1984. The diet of Indo-Pacific Vasum (Gastropoda: Vasidae). Journal of Conchology 31:375. Thomas, F.R. 2001. Mollusk habitats and fisheries in Kiribati: an assessment from the Gilbert Islands. Pacific Science 55(1):77-97. Thomas, F.R. 2002. An evaluation of central-place foraging among mollusk gatherers in Western Kiribati, Micronesia: linking behavioral ecology with ethnoarchaeology. World Archaeology 34(1):182-208. Thomas, F.R. 2007. The behavioral ecology of shellfish gathering in western Kiribati, Micronesia. 2: patch Choice, patch Sampling, and risk. Human Ecology 35:515-526. Thomas, F.R. 2014. Shellfish gathering and conservation on low coral islands: Kiribati perspectives. The Journal of Island and Coastal Archaeology 9(2):203-218. Titcomb, M. 1978. Native use of marine invertebrates in old Hawaii. Pacific Science 32(4):325-386. Uhle, M. 1907. The Emeryville Shellmound. University of California Publications in American Archaeology and Ethnology 7: 1-106. Weisler, M.I. 1999a. The antiquity of aroid pit agriculture and signiﬁcance of buried A horizons on Paciﬁc atolls. Geoarchaeology: An International Journal 14(7):621-654. Weisler, M.I. 1999b. Atolls as settlement landscapes: Ujae, Marshall Islands. Atoll Research Bulletin 460:1-51. Weisler, M.I. 2000. Burial artifacts from the Marshall Islands: description, dating and evidence for extra-archipelago contacts. Micronesica 33(1/2):111-136. Weisler, M.I. 2001a. Life on the edge: prehistoric settlement and economy on Utrōk Atoll, northern Marshall Islands. Archaeology in Oceania 36(3):109-133. Weisler, M.I. 2001b. On the Margins of Sustainability: Prehistoric Settlement of Utrōk Atoll, Northern Marshall Islands. Oxford: Archaeopress. Weisler, M.I. 2002. Archaeological Survey and Test Excavations on Ebon Atoll, Republic of the Marshall Islands. Majuro: Historic Preservation Office, Republic of the Marshall Islands. Weisler, M.I., R. Bollt and A. Findlater. 2010. Prehistoric fishing strategies on the makatea island of Rurutu. Archaeology in Oceania 45(3):130-143. Weisler, M.I., J.K. Lum, S.L. Collins and W.S. Kimoto. 2000. Status, health, and ancestry of a late prehistoric burial from Kwajelein Atoll, Marshall Islands. Micronesica 32(2):191-220.

119

Weisler, M.I. and D. Swindler. 2002. Rocker jaws from the Marshall Islands: evidence for interaction between eastern Micronesia and West Polynesia. People and Culture in Oceania 18:23-33. Weisler, M.I., H. Yamano and Q. Hua. 2012. A multidisciplinary approach for dating human colonization of Pacific atolls. The Journal of Island and Coastal Archaeology 7(1):102-125. Widdicombe, H. 1997. The Cutting Edge: A Technological Study of Adzes from Ebon, Maloelap and Ujae Atolls, Marshall Islands. M. A. thesis. Dunedin: University of Otago. Whitaker, A.R. 2008. Incipient aquaculture in prehistoric California?: Long-term productivity and sustainability vs. immediate returns for the harvest of marine invertebrates. Journal of Archaeological Science 35(4): 1114-1123. Wiens, H.J. 1962. Atoll Environment and Ecology. New Haven: Yale University Press. Willan, R. 1993. Taxonomic Revision of the Family Psammobiidae (Bivalvia: Tellinoidea) in the Australian and New Zealand region. Canberra: Australian Museum. Wolverton, S. 2013. Data quality in zooarchaeological faunal identification. Journal of Archaeological Method and Theory 20(3):381-396. WoRMS Editorial Board. 2015. World Register of Marine Species. Retrieved from . Yamaguchi, T., H. Kayanne and H. Yamano. 2009. Archaeological investigation of the landscape history of an oceanic atoll: Majuro, Marshall Islands. Pacific Science 63(4):537-565. Yamano, H., H. Kayanne and M. Chikamori. 2005. An overview of the nature and dynamics of reef islands. Global Environmental Research 9:9-20.

120

Figures and Tables Table 1 Summary of previous analyses of molluscan assemblages from archaeological sites in the Marshall Islands from Majuro (Riley 1987), Arno (Dye 1987) and Utrōk (Weisler 2001) atolls. Habitat assignments from Baron (1992), Baron and Clavier (1992), Carpenter and Niem (1998), Demond (1957), Soemodihardjo and Matsukuma (1989), Thomas (2001), and Willan (1993). Site type excavated Atoll Islet NTAXA NISP Representative taxa Predominant habitat(s) area (m2)

C. luhuanus Quidnipagus lagoon and ocean shoreline sands, lagoonside inter- Majuro Gafrarium spp. palatum Majuro Village 7 25 35062 reefal sand patches, areas of coral growth and (MiMLMj-1) Naticidae spp. A. violascens seagrass Harpa spp. Atactodea spp.

Majuro C. luhuanus Atactodea spp. Majuro Village 7 38 9460 ibid. (MiMLMj-10) Gafrarium spp. A. violascens

Arno Nerita polita Tridacna sp. lagoonside shoreline, sand patches, coral growth Arno (Ar-5-1, Ar-5-0, Village 23 14 N/A Turbo spp. Quidnipagus sp. seagrass, oceanside shoreline, reef flat pavement/ Islet 116) Cypraeidae Lambis spp. reef edge

N. polita Bikareij Tridacna spp. Turbo spp. Arno (Islet 17) Village 6 10 N/A Cerithium spp. ibid. Cypraeidae spp. Trochus spp. Quidnipagus spp.

Tridacna spp. Lambis spp. Jebu Arno Village 1 12 N/A Cerithium spp. Cypraeidae ibid. (112) Nerita polita Turbo spp.

A. striata Vasum spp. Lagoonside and oceanside reef flat, sands, shoreline, Utrōk Utrōk (MLUt-1) Village 26 35 8487 Tridacna spp. Codakia spp. lagoonside coral, oceanside reef edge

Tridacna spp. Conus spp. Utrōk Aon (MLUt-5) Village 12 21 524 Asaphis sp. Trochidae ibid. Turbo sp. C. nodulosum 121

Temporary Turbo spp. Vasum spp. Utrōk Aon (MLUt-4) 12 27 1299 Oceanside shoreline rocks, reef flat, and reef edge camp Tridacna spp. Cypraea spp.

Temporary C. nodulosum N. polita Lagoonside and oceanside reef flat, sands, shoreline, Utrōk Allok (MLUt-2) 2 21 237 camp Tridacna spp. Fragum spp. lagoonside coral

C. nodulosum Temporary Vasum spp. Utrōk Bikrak (MLUt-3) 10 32 1243 Tridacna spp. ibid. camp N. polita Codakia spp.

122

Table 2 List of zones, major geomorphological structures and detailed geomorphological structures used in the Ebon archaeological project hierarchical classification scheme (after Kendall et al. 2012:8-12). Zone J, dredged/excavated and Detailed Geomorphological structure 13, aggregated patch reefs was not used for the analysis presented here, as these classes relate to methods for mapping modern day atoll benthic habitats. Detailed Geomorphological structure 19, Algal Ridge, was added by the authors due to the distinctive range of molluscan taxa associated with this habitat (Morrison 1954). Zones cod e name description A Land Terrestrial features at or above the high tide line. B Shoreline Intertidal Area between the spring high tide line and lowest spring tide level C Lagoon Area of water inside the atoll, surrounded by the Back Reef D Reef Flat Shallow, low relief area exposed at low tide between the Shoreline Intertidal and Fore Reef or Back Reef E Back Reef Area on the lagoonside of an atoll sloping inward from the Shoreline Intertidal or Reef Flat down to the seaward edge of the Lagoon floor. F Fore Reef Area along the seaward (oceanside) edge of the reef flat that slopes into deeper water to the landward edge of the Bank/Shelf Escarpment G Bank/Shelf Deeper water extending offshore from the seaward edge of the Fore Reef to the beginning of the escarpment where the insular shelf drops off into deep, oceanic water H Bank/Shelf Escarpment Begins on the seaward edge of the Fore Reef, where depth increases rapidly into deep, oceanic water. I Channel Naturally occurring channels in the seafloor that often cut across several other zones. K Pinnacle High-relief features occurring in the Lagoon that are separated from the Back Reef by the deeper waters of the Lagoon. L Unknown Habitat proclivities could not be assessed.

123

habitats. 14 Spur and Groove Alternating sand and coral formations that are oriented perpendicular to the Shoreline intertidal or Fore Reef. The coral formations (spurs) of this feature typically have a high vertical relief and are separated by 1 to 5m of sand or hard bottom (grooves). Occurs only in the Fore reef or Bank/Shelf Escarpment zone. 15 Pavement Flat, low-relief, solid rock in broad areas, often with partial coverage of sand, algae, hard coral, Alcyonacea (sea whips or fans), zoozanthids or other sessile invertebrates. 16 Pavement with Sand Areas of pavement with alternating sand/surge channel formations that are oriented perpendicular to Channels the Shoreline Intertidal or Bank/Shelf escarpment. 17 Reef Rubble Dead, unstable coral rubble often colonised with turf, filamentous, calcareous or encrusting macroalgae. Often occurs due to storm waves piling up dead coral. 18 Rock/Boulder Large, irregularly shaped carbonate blocks often extending from the island bedrock, indicating higher sea-levels, or aggregations of loose coral cobbles and boulders that have been detached and transported from their native beds. Individual cobbles and boulders often range in diameter from 0.25-3m 19 Algal Ridge Area of consolidated coral pavement colonised by calcareous algae occurring shoreward of the Bank/Shelf Escarpment or Fore Reef and demarcates the seaward margin of the Reef Flat. Often slightly higher elevation that the seaward and shoreward areas of the reef. 2 Unconsolidated Substrate Areas of the seafloor consisting of small, unattached or uncemeneted particles with less than 10% cover of large stable substrate. 21 Sand Areas of the seafloor consisting of small, unattached or unncemented particles. 22 Sand with Scattered Coral Primarily sand bottom with scattered rocks or small, isolated coral heads. and Rock 23 Seagrass Primarily sand bottom colonised by seagrass. 3 Other Delineations Any other type of structure not classified as Coral Reef and Hard bottom or Unconsolidated Substrate. 31 Land Terrestrial features beyond the Shoreline Intertidal. 4 Unknown Habitat proclivities could not be assessed. 41 Unknown Habitat proclivities could not be assessed.

124

Figure 1 Map of the Marshall Islands and Ebon Atoll, showing archaeological sites on leeward (Ebon) and windward (Moniak) islets.

125

126

Figure 3 Mollusc species from TP17-20, MLEb-1, represented by 15 or more individuals.

127

Figure 4 Habitats accounting for more than 20% of MNI MLEb-1, TP17-20; a. all taxa b. Conus spp., Monetaria moneta and Cypraeidae spp. removed. See Table 1 for classification scheme key.

128

Figure 5 Mollusc species from TP17-20, MLEb-1 assigned to D/1/15 or D/1/16, represented by 15 or more individuals.

129

Figure 6 Mollusc species from TP2-6, MLEb-31, represented by 15 or more individuals.

130

Figure 7 Habitats accounting for more than 20% of MNI, TP2-6, MLEb-31 a. all taxa; b. Taxa assigned to D/1/18, represented by 15 or more individuals. See Table 1 for classification scheme key.

131

Chapter 5: Windward vs. leeward: Inter-site variation in marine resource exploitation on Ebon Atoll, Republic of the Marshall Islands

Note: This is the final version of a peer-reviewed article published in the Journal of Archaeological Science

Matthew Harris, Ariana B.J. Lambrides and Marshall I. Weisler School of Social Science, The University of Queensland, St Lucia, Queensland, 4072, Australia

Corresponding author: Matthew Harris, School of Social Science, The University of Queensland, St. Lucia, QLD 4072 Australia. Email: [email protected]

132

Abstract The variation in windward and leeward marine environments has been linked to distinctions in marine subsistence on large, high volcanic Pacific Islands, but these patterns have not been explored on low coral atolls. We document windward vs. leeward islet site variation in the taxonomic composition of fish bone and mollusc shell assemblages from three archaeological sites at Ebon Atoll, Republic of the Marshall Islands, to elucidate the relationship between local environment, archaeological site type and the taxonomic composition of marine archaeofaunal assemblages. While the representation of taxa at each site was broadly similar in terms of measures of taxonomic heterogeneity (richness, evenness and dominance), chord distance and correspondence analysis reported variation in taxonomic composition at each site. For mollusc shell assemblages, variation in taxonomic abundance indicates the influence of the marine environments adjacent to each site and the relative exposure of these coastlines to heavy surf, wind, waves and extreme weather events. Fish bone assemblages recovered from 6.4 mm screens had less inter-site variation in richness, evenness and rank order, but differences were noted in the rank order of fish taxa recovered from selective 3.2 mm screening of archaeological deposits when compared between sites. In contrast to patterns for molluscs, variation in the taxonomic composition of fish bone assemblages likely relates to site function, rather than the marine environments adjacent to each site. These trends highlight for the first time the complex range of factors that influenced the prehistoric acquisition of marine resources between leeward and windward islets, and document variation in prehistoric marine subsistence within one atoll.

Keywords: Atoll archaeology; marine subsistence; Marshall Islands; Micronesia; Pacific fishing; shell midden studies; zooarchaeology; archaeomalacology; ichthyoarchaeology

133

Introduction Marine resources were a critical component of prehistoric subsistence systems across the Pacific Islands (Allen 2012; Fitzpatrick et al. 2011; Jones 2009; Leach & Davidson 1988; Ono & Clark 2012; Szabó & Amesbury 2011; Thomas 1999; Weisler et al. 2010). Finfish and molluscan remains are ubiquitous in Marshall Islands archaeological sites (Beardsley & Athens 1994; Dye 1987; Riley 1987; Rosendahl 1987; Shun & Athens 1990; Weisler 1999b, 2001b), and ethnographic and linguistic evidence highlights the varied and complex ways that Marshallese interact with the marine landscape (Abo et al. 1976; Erdland 1914; Kramer & Nevermann 1938; Spoehr et al. 1949; Tobin 2002). The intertidal reef platforms of the Marshall Islands host at least 1000 species of mollusc (Vander Velde & Vander Velde 2008), and over 800 fish taxa occur within 60 m ocean depth (Myers 1999). These expansive reefs—often greater in area than the terrestrial zone of atolls—provided predictable and possibly sustainable yields of marine subsistence resources throughout prehistory (e.g., Ono & Addison 2013; Thomas 2014); indeed, on Utrōk Atoll situated at the northern limit of permanently-inhabited atolls in the Marshall Islands, an 1800-year sequence points to human populations, albeit in low numbers, with no indications of marine resource depression (Weisler 2001b:128).

The degree of wave exposure has been recognised as a key factor influencing a range of important ecological processes which are critical for structuring faunal composition in tropical intertidal environments (Drumm 2005) and has been related to variation in human behaviour in archaeological contexts. However, determining whether taxonomic composition of archaeofaunal assemblages is driven by ecological conditions and/or human agency is undoubtedly a complex endeavour. Patterned variation in site use and diet as it relates to site location (windward vs. leeward) has been documented on large, high volcanic Pacific Islands (e.g., Bayman and Dye 2013; Kirch & Dye 1979:58; Palmer et al. 2009; Weisler and Kirch 1985). In the Hawaiian archipelago particular agricultural practices are more commonly associated with leeward or windward locations, with rain-fed agriculture and sweet potato cultivation associated with dry, leeward landscapes, and irrigated taro cultivation better suited to wetter windward regions (Palmer et al. 2009: 1444, Earle 1977:224, Weisler and Kirch 1985). The differences in windward and leeward environments have also been recognised as influencing the exploitation of marine fauna. Ethno-archaeological research into Niuan fishing strategies (Niuatoputapu, Polynesia) recognised the relationship between fishing practices and distinctions in reef structure, tidal activities and faunal communities that were related to windward or leeward location (Kirch and Dye 1979). Similarly, Kirch (1982) noted variation in exploitation of fishes from three Hawaiian archaeological sites that was inferred to be driven by

134 local environment. Also in Hawaiʻi, fishponds are more common on the leeward coasts than on the exposed windward zone (Weisler and Kirch 1985). Distinctions in taxonomic composition between two assemblages, Tangarutu (leeward) and Akatanui 3 (windward), from Rapa Island were attributed to their windward/leeward location, and while similar species were identified at each site, their rank-ordering was variable, with pomacentrids more common at the windward site of Akatanui 3 (Vogel and Anderson 2012).

Mollusc assemblages from large, high volcanic Pacific Islands also demonstrate a link between site location and taxonomic composition. Assemblages from windward sites in Hawaiʻi are dominated by limpets (Cellana spp.) and turban shells (Turbo spp.), characteristic of rocky shores, but are rare in leeward sites (Kirch 1982, Morrison and Hunt 2007). Similarly, mollusc assemblages from Vaunautu (Bedford 2007), Fiji (Szabó 2009) and Rapa (Szabó and Anderson 2012) reflect local environmental conditions across the windward/leeward divide and have been interpreted as the result of non-selective foraging strategies operating in varying environments. On atolls, the configuration of marine environments and the distribution and relative abundance of fauna is determined by geological history, exposure to wind, waves and currents as well as myriad stochastic, local ecological, biological and abiotic factors and relates to the windward and leeward exposure of each islet (Wiens 1962). However, atoll settlement patterns and subsistence practices reflected in the variation between windward and leeward environments have not been investigated. We explore inter-islet and inter-site variation in the taxonomic composition of fish bone and mollusc shell assemblages from archaeological deposits on three islets on Ebon Atoll, Marshall Islands to elucidate the relationship between human foraging behaviour, site function, site location (windward v. leeward) and local environment. A range of statistical techniques are employed to investigate whether differences in the taxonomic composition of the assemblages is a reflection of ecological variability and site location (windward vs. leeward marine habitats) or site function (village vs. camp site). Future research avenues for exploring spatial variation in atoll settlement patterns and subsistence are then suggested.

Sites and Samples Ebon Atoll is the southernmost atoll in the Marshall Islands. Consisting of 22 islets encircling a 104 km2 lagoon, the total land area is approximately 5.4 km2 (Figure 1). The reef platform totals over 22 km2, roughly a 4:1 ratio of reef to land area. Two field seasons (1995/1996 and 2011/2012) of survey and excavation were conducted on Ebon Atoll as part of a larger project directed by Weisler to investigate regional variation in Marshall Islands archaeology as it relates to the 700+ km north-

135 south rainfall gradient, as well as documenting intra-atoll differences in settlement patterns and subsistence (Weisler 1999a, b, 2000, 2001a, b, 2002; Weisler & Swindler 2002; Weisler et al. 2012). We report results from the analyses of the fish bone and mollusc shell remains retained in 6.4 mm and 3.2 mm screens at three sites excavated during the 2011/2012 Ebon Atoll field season. A total of 68 m2 was excavated across seven archaeological sites on three islets, one situated on the leeward rim (Ebon Islet), one with windward exposure (Moniak Islet) and Enekoion Islet located between the two extremes. The archaeological sites chosen for analysis here include two lagoonside villages (MLEb-1 and MLEb-33) (Weisler 2001b) and a much smaller, shorter-term occupation site (MLEb-31) on Moniak (Weisler 2002). Lagoonside deposits were explicitly selected at all three sites to minimise diachronic effects on these analyses as they all represent a later phase of Marshallese prehistory in which habitation sites occur adjacent to the lagoon, in contrast to earlier phases where deposits are in the interior and associated with horticultural pits (Weisler 2001a:129). The following description of mollusc and finfish habitats on oceanic atolls derives from: Carpenter and Niem (1998), Demond (1957), Hiatt and Strasberg (1960), Kohn (1987) and Weins (1962). The largest islets and widest reef platforms are on the leeward south-eastern, western and north-western rim of Ebon Atoll. MLEb-1 and MLEb-33 lie in this zone, relatively sheltered from waves, winds and currents and feature high-rugosity coral reefs, and fine grained inter-reefal sand flats and seagrass beds on the lagoonside, and expansive, low relief pavements on the oceanside. Habitat complexity is highest on leeward islets, with a corresponding increase in faunal diversity predicted (Gratwicke and Speight 2005; Kohn and Leviten 1976). While mollusc taxa are generally sessile, and strongly associated with particular benthic habitats, finfish taxa are more difficult to associate with windward or leeward environments. Fish often track across multiple habitats with day/night cycles, tides, and during feeding. However, some taxa are strongly associated with certain substrate types, which vary in predominance between leeward and windward reef habitats as described below.

Leeward oceanside mollusc communities are highly diverse, with colonies of macroalgae and shallow tide pools hosting large numbers of cowries (Cypraeidae) drupes and other murex shells (Muricidae) top shells (Trochidae), cone shells (Conidae) and nerites (Neritidae). These reef platforms are also associated with large schools of parrotfish (scarids), surgeonfish (acanthurids), wrasse (labrids), goatfish (mullids), and small bodied sharks (carcharhinids), and algal turfs provide grazing for rabbitfish (siganids), sea chubs (kyphosids), butterfly fish (chaetodontids), acanthurids, damselfish (pomacentrids) and triggerfish (balistids).

136

The lagoonside reefs, seagrass beds and sand flats host communities of giant clams (Tridacna spp. and Hippopus spp.), spider conchs (Lambis spp. and Harpago spp.) Conidae, mitre shells and auger snails (Terebra spp. and members of the family Mitridae). The upper intertidal sand flats provide habitat for the easily accessible sand dwelling bivalves including the violet asaphis (Asaphis violascens), venus clams (Gafrarium spp.), cockles (Vasticardium spp.), ark clams (Arca spp.) and surf clams (Atactodea striata). Areas of coral growth on the lagoonside are associated with diverse herbivorous, carnivorous, and omnivorous fish communities including scarids, balistids, chaetodontids, pomacentrids, moray eel (muraenids), squirrelfish and soldierfish (holocentrids), grouper (serranids), snapper (lutjanids), labrids, filefish (monacanthids), and pufferfish (tetraodontids).

MLEb-31 is located on a windward islet, which is smaller and generally more exposed to winds, waves and currents. The intertidal zone is primarily composed of poorly sorted coral rubble washed from the ocean facing subtidal reefs, wave cut erosional channels, and coarse, gravelly sands on the lagoonside. Habitat complexity is generally lower on windward islets, with a decrease in richness and diversity of mollusc and fish communities predicted (Gratwicke and Speight 2005; Kohn and Leviten 1976). Mollusc communities characterised by large and robust Turbo, drupes (Drupa spp.), Cypraeidae, Conidae, vase shells (Vasum spp.), polished nerites (Nertia polita) and frog shells (Bursa spp.). Finfish communities on windward islets are varied and complex, but taxa characteristic of exposed surge channels include muraenids, carcharhinids, hawkfish (cirrhitids), serranids, lutjanids, acanthurids, pomacentrids, labrids, scarids, combtooth blenny (blenniids), and balistids.

The following analyses are based on all fish bone and mollusc shells retained in the 6.4 mm screens from lagoonside deposits on Ebon Atoll; Test Pit (TP) 17 to 20 at site MLEb-1, TP 2 to 6 at site MLEb-31 and TP 2 and 8 at site MLEb-33 (Figure 1). A single unit from each site—MLEb-1 (TP 17), MLEb-31 (TP 2) and MLEb-33 (TP 8)—was sieved with 3.2 mm screens during the 2011/12 fieldwork. All excavated sediments were wet-sieved. Hereafter, the single-unit sub-samples of fish bones recovered from nested 3.2 mm and 6.4 mm screens are described as the 3.2 mm samples.

MLEb-1 MLEb-1 is located at the centre of a ~2 km long village system on Ebon Islet (Weisler 2002). Ebon Islet is the largest islet of the atoll, featuring high-rugosity coral reefs and sand flats in the lagoon intertidal and expansive, relatively calm intertidal reef flats on the oceanside (Figure 1a). Cultural 137 material, including molluscan remains, fish bone, charcoal, oven (um) stones and worked shell artefacts were recovered from a 2 x 2 m unit (TP 17, 18, 19 and 20) excavated into a low mound, built by the accumulation of successive coral pavements, located 40 m inland of the current lagoon shore and 20 m northwest of the Primary School. Cultural deposits extend to a depth of 1.75 m. Cultural material retained in the 6.4 mm screens yielded 3464 fragments of mollusc shell (MNI [Minimum Number of Individuals] = 1258), and 4188 fish bones (MNI = 509), with 94.1% and 39.3% of fragments, respectively, identified to family, genus or species. The 3.2 mm fish bone samples (TP 17) yielded a NISP (Number of Identified Specimens) of 2655 (MNI = 378), with 14.4% of fragments identified to family, genus or species.

MLEb-33 Situated on Enekoion Islet, MLEb-33 is a sparse to dense midden deposit surrounding a large aroid pit situated from 100 m to 25m from the lagoon shore. A 1 x 2 m trench was excavated on the lagoonward side of the aroid pit where cultural deposits extended to 40 cm below surface. The oceanside reef flat is generally wider than at Ebon or Moniak, but is mainly composed of coral rubble, boulders and eroded beachrock slabs, more similar to the exposed reef flat of Moniak than Ebon Islet. The lagoonside environment features expansive seagrass meadows (Thalassia spp., Figure 1b) and sand flats, and some coral growth in the intertidal, similar to the lagoon environments at Ebon Islet. Cultural material retained in the 6.4 mm screens yielded 617 fragments of mollusc shell (MNI = 230), and 144 fish bones (total MNI = 67), with 96.8% and 60.5% of fragments, respectively, identified to family, genus or species. The 3.2 mm fish bone samples (TP 8) yielded a NISP of 98 (MNI = 34), with 20.1% of fragments identified to family, genus or species.

MLEb-31 The small midden site MLEb-31 is located ~75 m from the lagoon shore on the windward islet of Moniak (Weisler 2002). Cultural deposits extend to a maximum depth of 70 cm. The oceanside intertidal is characterised by coral boulder ramparts and cobbles deposited by extreme weather events (Figure 1c), in contrast to the relatively protected oceanside of Ebon and Enekoion Islets. The lagoonside sands of Moniak are coarse and the shore declines steeply to the lagoon floor. Cultural material retained in the 6.4 mm screens yielded 1740 fragments of mollusc shell (MNI = 650), and 1084 fish bones (MNI = 326), with 95% and 53.5% of fragments, respectively, identified to family, genus or species. The 3.2 mm fish bone samples (TP 6) yielded a NISP of 648 (MNI = 192), with 20.4% of fragments identified to family, genus or species.

138

Methods Identification and quantification protocols Fish remains were identified by Lambrides and mollusc remains by Harris; all identifications were completed to the lowest taxonomic level using Indo-Pacific comparative reference collections housed at The University of Queensland archaeology laboratory (see Lambrides & Weisler 2015: 5; Weisler 2001b: appendix 3, for a description of the fish reference collection). Reference manuals were also used for molluscan identification, including: Abbott and Dance (1990), Poppe (2008), Röckel et al. (1995) and Burgess (1985). Due to the richness of Indo-Pacific marine fauna, all fish bone (see Lambrides & Weisler 2013) and mollusc shell fragments were attempted for identification, but lower order taxonomic identifications (e.g. genus and species) were assigned with caution to avoid over-identification (Driver 1992; Wolverton 2013). Taxonomic abundance of archaeological fish bone and mollusc shell were quantified by NISP and MNI. For fish remains, MNI values were calculated following standard zooarchaeological protocols for vertebrate fauna (Grayson 1984; Lyman 2008; Reitz & Wing 2008) and for molluscs following Harris et al. (2015). The quantification methods used here allow comparison of fish and mollusc taxonomic abundance as MNI values were consistently determined using the most frequently occurring Non-Repetitive Element (NRE).

Statistical analyses Both mollusc shell and fish bone samples were aggregated at the site level. The NRE frequency for each taxon was summed by cultural layer for calculating MNI. Relative taxonomic abundance is used here to examine differences in the taxonomic composition of fish bone and mollusc shell assemblages from the three sites to explore the interaction between windward vs. leeward islets, local environment and the extraction of marine fauna. A range of statistical tests were utilised including taxonomic richness and diversity as measured by NTAXA, the Shannon-Weiner index of diversity (H’), Shannon’s evenness (E), Simpson’s dominance (1-D), and Fisher’s α. Similarity and difference in faunal composition was analysed using chord distance analysis, and correspondence analysis (CA). These statistics have proven utility for examining similarities and differences in taxonomic composition for archaeological assemblages, including faunal (Faith 2013) and archaeobotanical samples (Wright et al. 2015). All statistical analyses reported below were carried out using MNI values for comparability with other Pacific Island assemblages, but it should be noted that statistical analyses of NISP values were tested and revealed similar trends (Grayson 1984; Lyman 2008). All statistical analyses were completed using PAST Paleontological Statistics Package, version 3.06 (Hammer et al. 2001). 139

Species richness (the number of species in an analytical unit) was assessed using NTAXA. Evenness, being the relative abundance of species in the assemblage, was measured using the Shannon-Weiner index of diversity (H’) and Shannon’s evenness (E). H’ values range between 0 and a theoretical maximum of 5, but values between 1.5 and 3.5 are most common. Higher H’ values indicate greater species diversity and richness. E values range between 0 and 1, with 0 indicating assemblages dominated by a single taxon, and values closer to 1 indicating rich, even assemblages (Lyman 2008: 195; Reitz & Wing 2008: 111). The dominance of few species in the assemblage was assessed using Simpson’s index of diversity (1-D). 1-D values range from 0 to 1, with lower values indicating assemblages dominated by a single taxon (Magurran 2004: 116). Fisher’s α, a measure of diversity, was also utilised as Shannon’s indices and NTAXA can be influenced by sample size (Faith 2013). Fisher’s α values are considered to be relatively independent of sample size (Hayek & Buzas 2010: 295-296). Fisher’s α tracks the occurrence of taxa represented by single individuals as a measure of overall diversity (Karlson et al. 2004). Significant difference between diversity indices calculated for each sample were also carried out using random permutation tests of relative abundance data.

Chord distance and exploratory CA analyses were conducted using non-aggregated (i.e. not collapsed by NTAXA) relative abundance data. NTAXA quantifies richness by collapsing taxa at the highest common taxonomic level for each assemblage. NTAXA values are generally correlated with sample size and can be influenced by identification protocols, but do ensure richness values are not inflated by species that are more easily identified to lower taxonomic levels. Chord distance analysis is a scaled measure of Euclidean distance for examining the dissimilarity between samples in relative abundance of species, such as sites or cultural units (Faith et al. 1987; Legendre & Gallagher 2001). Chord distance values range from 0, indicating samples with no difference in relative abundance, up to indicating no species in common between samples. Chord distance values are useful measures of dissimilarity for the assemblages examined here, as species represented by single individuals are not highly weighted.

This suite of statistical tests allows an exploration of the role of site location (windward vs. leeward marine habitats) and site function (village vs. camp site) on the taxonomic composition of the assemblages. Taxonomic measures of heterogeneity and chord distance are complementary analyses which can be used to assess human collection strategies (i.e. non-selective or selective behaviours)

140 that are linked to local faunal community structure (i.e. number of species, dominance of particular taxa, etc.) and site function. Correspondence analysis is used to further explore the relationship between particular taxa, ecological variables, and site function, and provides useful data for comparison with the results of the other statistical analyses used here.

Results Figure 2 presents the relative abundance of mollusc and fish taxa identified from the 6.4 mm and 3.2 mm screened assemblages at sites MLEb-1, MLEb-31 and MLEb-33. To highlight broad trends in taxonomic composition at each site, quantification data are aggregated at the family level. Characteristic of Indo-Pacific marine archaeofaunas, many species are represented (e.g., Morrison & Addison 2009; Ono & Intoh 2011; Riley 1987; Szabó 2009; Weisler 2001b). The molluscan assemblage is dominated by gastropods (x̄ = 84.2% MNI / 81.7% NISP), with bivalves contributing minimally to MNI and NISP. Both 6.4 mm and 3.2 mm fish bone assemblages are dominated by piscivores and omnivores/benthic carnivores, which account for 75.2 % and 84.4 % of total MNI, respectively.

Taxonomic measures of heterogeneity for all samples report high richness and evenness and low dominance overall (Table 2). Random permutation tests for significant difference between index values at each site reported significant values only for 1-D values for molluscs between MLEb-1 and MLEb-31 (p = 0.0002) and E values for 3.2 mm fish bone samples from MLEb-1 and MLEb-31 (p = 0.0264). 1-D values are sensitive to differences in the relative abundance of the top ranked taxa, explaining the significant result for mollusc samples. E values are sensitive to alterations in the relative abundance of all taxa, once again explaining the significant difference reported for 3.2 mm fish samples from MLEb-1 and MLEb-31.

Chord distance was used to measure the dissimilarity between the relative abundance of taxa in each assemblage. The greatest faunal dissimilarity as measured by chord distance was reported for molluscan assemblages from site pairs MLEb-1/MLEb-31 and MLEb-1/MLEb-33. Minimal dissimilarity was reported for the mollusc assemblage from MLEb-31/MLEb-33, 6.4 mm fish bone assemblages from all site pairs and 3.2 mm fish bone assemblage from site pair MLEb-1/MLEb-31. Moderate dissimilarity was noted for 3.2 mm fishbone assemblages from site pairs MLEb-1/MLEb- 33 and MLEb-31/MLEb-33. Chord distance analysis indicates that molluscan assemblages tend to be more taxonomically dissimilar than fish bone assemblages across all site pairs, except for MLEb-

141

31/MLEb-33. Both taxonomic measures of heterogeneity and chord distance analysis indicate that all assemblages were relatively similar in terms of richness, evenness, and relative abundance of taxa, with the most pronounced differences generally between mollusc assemblages from MLEb-1 and MLEb-31.

CA of taxonomic abundance (6.4 mm data) was used to investigate whether differences in the taxonomic composition of the assemblages as initially determined by the results of the diversity measures, and chord distance, was better explained by local ecological variability (windward vs. leeward marine habitats) or site function (village vs. camp site). Figure 3a-c plots CA axis 1 and 2 for all samples, which account for 77.8% and 22.2% of the variance in taxonomic abundance, respectively. Axis 1 discriminates between windward and leeward islets, with MLEb-1, on the most leeward islet, reporting the lowest axis 1 score and MLEb-31, on the most windward islet, reporting the highest axis 1 score. MLEb-33, which is moderately exposed to windward waves, currents and wind, reports an intermediate axis 1 score. The negative axis 1 score that characterises MLEb-1 is associated with 43 mollusc taxa and six fish taxa (Carcharhinus spp., Decapterus spp., Elagatis bipinnulata, Ostraciidae, Sphyraena spp., and Thunnus spp.) that occur only at that site, and account for 11% and 4.3% of total site MNI, respectively. MLEb-31 is characterised by positive axis 1 scores, and is associated with 34 mollusc taxa and a single fish taxon (Zebrasoma sp.) that occur only at that site, and account for 7.7% and 0.3% of total site MNI, respectively. MLEb-33 is characterised by negative axis 2 scores, and is associated with five molluscan taxa (Conus leopardus, C. lividus, Corculum cardissa, Harpa spp., Periglypta spp.) that occur only at that site (2.5% of total site MNI), but no distinct fish species. Axis 1 scores are negatively loaded by reef flat pavement dwelling gastropod taxa, the most common habitat exploited for mollusc gathering at MLEb-1 (Harris and Weisler 2016). Conversely, axis 2 scores are negatively loaded by sand- dwelling gastropod and bivalve taxa. Interestingly, the extant lagoon environment adjacent to MLEb-33 is predominately turtle grass (Thalassia spp.) beds and sand flats. Habitat proclivities are more difficult to assess for the non-sessile fish but, generally, the representation of feeding behaviours (piscivores, omnivores/benthic carnivores and herbivores) for non-distinct taxa were broadly similar at each site. In contrast, the CA of 3.2 mm fish bone taxonomic abundance data (Figure 3d) report similar levels of variance for both axis 1 and axis 2, 58.2 % and 41.8 %, respectively. This suggests that there is less taxonomic similarity between sites than represented by the 6.4 mm data, which is also reflected by the chord distance scores (Table 1). Similar to the 6.4 mm fish bone CA, distinct taxa from all sites only accounted for a small percentage of the

142 assemblage, specifically 5.6 % of total MNI. CA results reflect the substantially different rank ordering of taxa at each site as represented by the 3.2 mm data.

Discussion Spatial variation in marine subsistence as it relates to windward and leeward settlement patterns on a single atoll has not previously been assessed in Pacific Island archaeology. A range of statistical analyses were implemented using mollusc shell and fish bone relative taxonomic abundance data reported from three habitation deposits on Ebon Atoll. All sites were located adjacent to the extant lagoon shore, with one a temporary habitation site (MLEb-31), and the other two large villages (MLEb-1 and MLEb-33).

The taxonomic composition of archaeological mollusc and fish bone assemblages from each site evidenced a similar suite of families with assemblages characterised by highly rich and even measures of taxonomic diversity, and no strongly dominant taxa, despite probable differences in habitat complexity at each site. All mollusc assemblages are dominated by gastropods, with bivalves contributing minimally to MNI. The dominance of gastropods is characteristic of macrobenthic mollusc communities recorded for other atolls in the Marshall Islands (Kay & Johnson 1987), potentially indicating a generalised molluscan foraging strategy at all sites (Harris & Weisler 2016, Szabó 2009, Kirch 1982). Richness was generally greatest at MLEb-1, which is unsurprising as this is both the largest sample, and from the most complex habitat (Gratwicke and Speight 2005).

Mollusc assemblages at each site have relatively high ranks for nerites, muricids (Drupa spp. and Thais spp.), Conus spp., Cerithidae, Cypraeidae, Turbo spp. and Asaphis violascens. Fish bone assemblages recovered from the 6.4 mm screens are generally dominated by piscivorous species (e.g., serranids, lutjanids and carangids) followed in rank order by omnivorous/benthic carnivorous species (e.g., holocentrids, lethrinids and balistids) and herbivorous species (e.g. acanthurids and scarids). Fish bone assemblages recovered from the 3.2 mm screens showed increased dominance of omnivores/benthic carnivores (e.g., holocentrids, exocoetids and mullids) which may be related to average fish bone size for these taxa, bone density and taphonomy. For fish bone assemblages from the 6.4 mm screens, scarids and serranids are rank 1 and 2, respectively, at MLEb-1 and MLEb-31, and rank 2 and 3 at MLEb-33. However, the 3.2 mm samples, while similar in taxonomic composition, showed greater variation in the rank ordering of these taxa when compared

143 to the 6.4 mm samples, similar trends were noted from Rapa Islands archaeological sites, where 2 mm screens were utilised (Vogel and Anderson 2012).

Mollusc assemblages reported generally greater inter-site variation in taxonomic abundance than fish bone assemblages. The statistical analyses utilised here indicate variation in relative taxonomic abundance of mollusc assemblages is due to differences in the local environment at each site. At MLEb-1, the majority of the molluscan assemblage could have been gathered from the oceanside reef flat and coral reefs (Harris & Weisler 2016). Conversely, CA shows that those molluscan taxa that prefer sandy lagoon substrates are most strongly associated with MLEb-33, a marine environment today which is characterised by large lagoonal sand flats and turtle grass beds. Furthermore, assemblages at MLEb-31 consist principally of those taxa which either inhabit the boulder ramparts, typical of windward islets (i.e., Nerita polita), or those taxa which are suitably adapted to constant exposure to wind, waves and currents on the reef edge (i.e., Mauritia mauritiana, Vasum turbinellus, and Thais armigera). The variability in the relative abundance of Nerita plicata, Nerita polita and the ranellids (Monoplex intermedius, Monoplex nicobaricus and Gutturnium muricinum) likely explains the significant difference in measures of dominance between molluscan assemblages from MLEb-1 and MLEb-31. This result also potentially indicates the influence of local environment, as Ranellidae are most common in areas of coral growth (Govan 1995), which are characteristic of Ebon Islet (MLEb-1), but are rare at Moniak (MLEb-31). The correlation between local environment and the molluscan taxa in the assemblage, in addition to the even, rich and diverse taxonomic composition indicates that a non-selective foraging strategy, mediated by local environment, operated at each site. These patterns are broadly similar to mollusc assemblages from other oceanic islands where richness and evenness are high, and taxonomic composition varies predominately with changes in site location and local environment (e.g. Szabó 2009).

Fish bone assemblages, however, generally have lower values of dissimilarity between sites and are harder to link to marine environments adjacent to each site because fish track across different habitats while foraging, unlike molluscs that are generally sessile. For example, Katsuwonus pelamis, a pelagic-oceanic dwelling species, is dominant at the largest village site, MLEb-1 but rare at the campsite, MLEb-31 indicating that variation in fish bone assemblages may be related to site function rather than the marine environments adjacent to the sites. This trend requires further analysis (e.g. assessment of temporal variation and inclusion of additional sites), but could relate to a number of variables, including settlement patterns and fish capture strategies operating at each

144 site. Variability in inter-site taxonomic composition for 3.2 mm samples was generally greater than 6.4 mm fish bone assemblages. Mesh size has been linked to alterations in species richness and diversity (e.g., Nagaoka 2005; Ono & Clark 2012) and inferences made regarding capture methods, morphometric reconstructions of fish size, ontogenetic growth and associated live fish behaviour can be useful for predicting fishing technology (Bertrando & McKenzie 2011).

Conclusion Exploratory data analyses were implemented to determine whether there are differences and/or similarities in the taxonomic composition of mollusc shell and fish bone assemblages from three archaeological sites situated on windward and leeward islets at Ebon Atoll, Marshall Islands. Results indicate broad similarity in assemblage composition, reflected in similar richness, evenness and dominance scores at each site, regardless if the sites were small intermittently occupied habitations on windward islets or large villages on leeward islets. Where variation in taxonomic composition occurs, the configuration of marine environments at each site may account for much of the differences in the mollusc assemblages, while fishing technology (capture techniques) and site function (i.e. village sites v. campsites) could account, in part, for the variation in fish bone assemblages at each site. Intra-islet variation in taxonomic abundance, metric analysis of fish bone to assess body size over time and possible effects of human impacts to marine fisheries, studies of associated material culture and temporal analysis of alterations to foraging strategies will provide additional datasets for testing the influence of local marine environments and human settlement patterns on relative taxonomic abundance of mollusc shell and fish bone assemblages from Ebon Atoll.

This study has shown that even within a single atoll, human foraging patterns can differ over small spatial scales. The observed patterns follow documented evidence from other Pacific Islands where molluscan assemblages broadly reflect local environmental conditions. In contrast, fishbone assemblages possibly reflect capture methods and site function. The results presented here highlight the importance of atolls for examining the dialogue between human behaviour and local marine environments when investigating long-term trajectories of human-environment interaction. Assessing variation in the composition of archaeofaunal assemblages from windward and leeward islets can yield useful information for understanding variation in settlement patterns and intra-atoll subsistence practices—the latter previously not recognised. These intra-atoll analyses are critical for assessing variability in marine subsistence practices and are applicable to other island types across the Pacific.

145

Acknowledgements Permission to conduct archaeological research in the Republic of the Marshall Islands was granted to Weisler by the Historic Preservation Office, Ministry of Internal Affairs and on Ebon Atoll, former mayor Lajan Kabua. Marshall Islands fieldwork was supported by a grant to Weisler from the Office of the Deputy Vice Chancellor (Research), University of Queensland. Harris’ and Lambrides’ postgraduate studies are supported by an Australian Postgraduate Award.

146

References Cited Abbott, R.T. and S.P. Dance. 1990. Compendium of Seashells. Melbourne, Florida: American Malacologists, Inc. Abo, T., B.W. Bender, A. Capelle and T. DeBrum. 1976. Marshallese-English Dictionary. Honolulu: University of Hawai'i Press. Allen, M.S. 2012. Molluscan foraging efficiency and patterns of mobility amongst foraging agriculturalists: a case study from northern New Zealand. Journal of Archaeological Science 39(2):295-307. Anderson, A. and D.J. Kennett. 2012. Terra Australis 37: Taking the High Ground: The Archaeology of Rapa, a Fortified Island in Remote East Polynesia. Canberra: ANU E Press. Bayman, J.M. and T.S. Dye. 2013. Hawaii's Past in a World of Pacific Islands. Washington D.C.: The SAA Press. Beardsley, F.R., J.S. Athens, G.M. Murakami, J.V. Ward and M. Pietrusewsky 1994 Archaeological Investigations on Kwajalein Atoll, Marshall Islands, International Archaeological Research Institute, Inc., Honolulu. Bedford, S. 2007. Terra Australis 23: Pieces of the Vanuatu Puzzle: Archaeology of the North, South and Centre. Canberra: ANU E Press. Bertrando, E.B. and D.K. McKenzie. 2011. Identifying fishing techniques from the skeletal remains of fish. In Exploring Methods of Faunal Analysis: Insights from California Archaeology (M.A. Glassow and T.L. Joslin, eds):169-186. Los Angeles: Cotsen Institute of Archaeology Press. Burgess, C.M. 1985. Cowries of the World. Cape Town: Seacomber Publications. Carpenter, K.E. and V.H. Niem (eds) 1998.The Living Marine Resources of the Western Central Pacific: Volume 1. Seaweeds, Corals, Bivalves and Gastropods. Rome: Food and Agriculture Organization of the United Nations. Demond, J. 1957. Micronesian reef-associated gastropods. Pacific Science 11(3):275-341. Driver, J.C. 1992. Identification, classification, and zooarchaeology. Circaea 9:35-47. Drumm, D.J. 2005. Habitats and Macroinvertebrate Fauna of the Reef-top of Rarotonga, Cook Islands: Implications for Fisheries and Conservation Management. Ph.D. thesis. Dunedin: University of Otago. Dye, T. 1987a. Marshall Islands Archaeology. Honolulu: Bernice Pauahi Bishop Museum. ---. 1987b. Report 3: Archaeological survey and test excavations on Arno Atoll, Marshall Islands. In Marshall Islands Archaeology (T. Dye, ed.):271-394. Honolulu: Pacific Anthroplogical Records: Bernice Pauahi Bishop Museum.

147

Erdland, A. 1914. Die Marshall Insulaner. Leben und Sitte, Sinn und Religion eines Südsee-volkes. Anthropos Bibliothek. Internationale Sammlung Ethnologischer Monographien 2(1). Faith, D., P. Minchin and L. Belbin. 1987. Compositional dissimilarity as a robust measure of ecological distance. Vegetatio 69(1-3):57-68. Faith, J.T. 2013. Taphonomic and paleoecological change in the large mammal sequence from Boomplaas Cave, western Cape, South Africa. Journal of Human Evolution 65(6):715-730. Fitzpatrick, S.M., W.R. Dickinson and G. Clark. 2003. Ceramic petrography and cultural interaction in Palau, Micronesia. Journal of Archaeological Science 30(9):1175-1184. Fitzpatrick, S.M., C.M. Giovas and O. Kataoka. 2011. Temporal trends in prehistoric fishing in Palau, Micronesia over the last 1500 years. Archaeology in Oceania 46(1):6-16. Govan, H. 1995. Cymatium muricinum and Other Ranellid Gastropods: Major Predators of Cultured Tridacnid Clams. Penang: WorldFish. Gratwicke, B. and M. Speight. 2005. The relationship between fish species richness, abundance and habitat complexity in a range of shallow tropical marine habitats. Journal of Fish Biology 66(3):650-667. Grayson, D.K. 1984. Quantitative Zooarchaeology: Topics in the Analysis of Archaeological Faunas. Orlando: Academic Press. Hammer, Ø. 2001. PAST PAleontological STatistics. Oslo, Norway: Natural History Museum, University of Oslo. 3.04. Harris, M. and M.I. Weisler. 2016. Intertidal foraging on atolls: Prehistoric forager decision making at Ebon Atoll, Marshall Islands. The Journal of Island and Coastal Archaeology 12(2):200- 223. Harris, M., M.I. Weisler and P. Faulkner. 2015. A refined protocol for calculating MNI in archaeological molluscan shell assemblages: A Marshall Islands case study. Journal of Archaeological Science 57:168-179. Hayek, L.-A.C. and M.A. Buzas. 2010. Surveying Natural Populations: Quantitative Tools for Assessing Biodiversity. New York: Columbia University Press. Hiatt, R.W. and D.W. Strasburg. 1960. Ecological relationships of the fish fauna on coral reefs of the Marshall Islands. Ecological Monographs 30(1):65-127. Jones, S. 2009. A Long-term perspective on biodiversity and marine resource exploitation in Fiji's Lau Group. Pacific Science 63(4):617-648. Karlson, R.H., H.V. Cornell and T.P. Hughes. 2004. Coral communities are regionally enriched along an oceanic biodiversity gradient. Nature 429(6994):867-870.

148

Kay, E.A. and S. Johnson. 1987. Mollusca of Enewetak Atoll. In The Natural History of Enewetak Atoll: Volume II Biogeography and Systematics (D.M. Devaney, E.S. Reese, B.L. Burch and P. Helfrich, eds):105-146. Tennessee: Office of Scientific and Technical Information, U.S. Department of Energy. Kirch, P. 1982. The ecology of marine exploitation in prehistoric Hawaii. Human Ecology 10(4):455-476. Kirch, P.V. and T.S. Dye. 1979. Ethno-archaeology and the development of Polynesian fishing strategies. The Journal of the Polynesian Society 88(1):53-76. Kohn, A.J. 1987. Intertidal ecology of Enewetak Atoll. In The Natural History of Enewetak Atoll: Volume I The Ecosystem: Environments, Biotas, and Processes (D.M. Devaney, E.S. Reese, B.L. Burch and P. Helfrich, eds):139-157. Tennessee: Office of Scientific and Technical Information, U.S. Department of Energy. Kohn, A.J. and P.J. Leviten. 1976. Effect of habitat complexity on population density and species richness in tropical intertidal predatory gastropod assemblages. Oecologia 25(3):199-210. Kramer, A. and H. Nevermann. 1938. Ralik-Ratak (Marshall-Inseln), G. Thilenius, Ergebnisse der Sudsee-Expedition 1908-1910. II. Ethnographie, B: Mikronesien. Hamburg: De Gruyter & Co. Lambrides, A.B.J. and M.I. Weisler. 2013. Assessing protocols for identifying Pacific Island archaeological fish remains: The contribution of vertebrae. International Journal of Osteoarchaeology 25(6):838-848. Lambrides, A.B.J. and M.I. Weisler. 2015. Applications of vertebral morphometrics in Pacific Island archaeological fishing studies. Archaeology in Oceania 50(2):53-70. Leach, B.F. and J.M. Davidson. 1988. The quest for the rainbow runner: Prehistoric fishing on Kapingamarangi and Nukuoro Atolls, Micronesia. Micronesica 21:1-22. Legendre, P. and E. Gallagher. 2001. Ecologically meaningful transformations for ordination of species data. Oecologia 129(2):271-280. Lyman, R.L. 2008. Quantitative Paleozoology. Cambridge: Cambridge University Press. Magurran, A.E. 2004. Measuring Biological Diversity. Malden: Oxford; Blackwell. Morrison, A.E. and D.J. Addison. 2009. Examining causes and trends in marine trophic level change: 1500 years of fish exploitation at Fatu-ma-Futi, Tutuila Island, American Sāmoa. The Journal of Island and Coastal Archaeology 4(2):177-194. Morrison, A.E. and T.L. Hunt. 2007. Human Impacts on the nearshore environment: An archaeological case study from Kaua'i, Hawaiian Islands. Pacific Science 61(3):325- 328,331-345.

149

Myers, R.F. 1999. Micronesian Reef Fishes: A Comprehensive Guide to the Coral Reef Fishes of Micronesia. Barrigada: Coral Graphics. Nagaoka, L. 1994. Differential recovery of Pacific Island fish remains: Evidence from the Moturakau Rockshelter, Aitutaki, Cook Islands. Asian Perspectives 33(1):1-17. Ono, R. and D. Addison. 2013. Historical ecology and 600 years of fish use on Atafu Atoll, Tokelau. In Terra Australis 39: Prehistoric Marine Resource Use in the Indo-Pacific Regions (R. Ono, A. Morrison and D. Addison, eds):59-83. Canberra: ANU E Press Ono, R. and G. Clark. 2012. A 2500-year record of marine resource use on Ulong Island, Republic of Palau. International Journal of Osteoarchaeology 22(6):637-654. Ono, R. and M. Intoh. 2011. Island of pelagic fishermen: Temporal changes in prehistoric fishing on Fais, Micronesia. The Journal of Island and Coastal Archaeology 6:255–286. Palmer, M.A., M. Graves, T.N. Ladefoged, O.A. Chadwick, T. Ka'eo Duarte, S. Porder and P.M. Vitousek. 2009. Sources of nutrients to windward agricultural systems in pre-contact Hawai'i. Ecological Applications 19(6):1444-1453. Poppe, G.T. 2008. Philippine Marine Mollusks. Hackenheim: ConchBooks. Reitz, E.J. and E.S. Wing. 2008. Zooarchaeology. Cambridge: Cambridge University Press. Riley, T.J. 1987. Report 2: Archaeological survey and testing, Majuro Atoll, Marshall Islands. In Marshall Islands Archaeology (T. Dye, ed.):169-270. Honolulu: Pacific Anthroplogical Records: Bernice Pauahi Bishop Museum. Röckel, D., W. Korn and A.J. Kohn. 1995. Manual of the Living Conidae. Hackenheim: Verlag Christa Hemmen. Rosendahl, P.H. 1987. Report 1: Archaeology in eastern Micronesia: Reconnaissance survey in the Marshall Islands. In Marshall Islands Archaeology (T. Dye, ed.):17-166. Hawaii: Pacific Anthroplogical Records: Bernice Pauahi Bishop Museum. Shun, K. and J.S. Athens. 1990. Archaeological investigations on Kwajalein Atoll, Marshall Islands, Micronesia. Micronesica Supplement 2:231-240. Spoehr, A., P.S. Martin and L.A. Ross. 1949. Majuro, a village in the Marshall Islands. Fieldiana Anthropology 39:1-262. Szabó, K. 2009. Molluscan remains from Fiji. In Terra Australis 31: The Early Prehistory of Fiji (G. Clark and A.J. Anderson, eds):183-212. Canberra: ANU E Press. Szabó, K. and J.R. Amesbury. 2011. Molluscs in a world of islands: The use of shellfish as a food resource in the tropical island Asia-Pacific region. Quaternary International 239(1–2):8-18.

150

Szabó, K. and A. Anderson. 2012. The Tangarutu invertebrate fauna. In Terra Australis 37: Taking the High Ground: The Archaeology of Rapa, a Fortified Island in Remote East Polynesia (A. Anderson and D.J. Kennett, eds):135-144. Canberra: ANU E Press. Thomas, F.R. 1999. Optimal Foraging and Conservation: The Anthropology of Mollusk Gathering Strategies in the Gilbert Islands Group, Kiribati. Ph.D. thesis. Honolulu: University of Hawaii at Manoa. ---. 2014. Shellfish gathering and conservation on low coral islands: Kiribati perspectives. The Journal of Island and Coastal Archaeology 9(2):203-218. Tobin, J., A. 2002. Stories from the Marshall Islands. Honolulu: University of Hawai'i Press. Vander Velde, N. and B. Vander Velde. 2008. Seashells and Other Mollusks of the Marshall Islands: Chitons and Snails and Relatives. Majuro: Republic of Marshall Islands Historic Preservation Office. Vogel, Y. and A. Anderson. 2012. Prehistoric fishing on Rapa Island. In Terra Australis 37: Taking the High Ground: The Archaeology of Rapa, a Fortified Island in Remote East Polynesia (A.J. Anderson and D. Kennett, eds):115-133. Canberra: ANU E Press. Weisler, M.I. 1999a. The antiquity of aroid pit agriculture and signiﬁcance of buried A horizons on Paciﬁc Atolls. Geoarchaeology: An International Journal 14(7):621-654. ---. 1999b. Atolls as settlement landscapes: Ujae, Marshall Islands. Atoll Research Bulletin 460:1- 51. ---. 2000. Burial artifacts from the Marshall Islands: description, dating and evidence for extra- archipelago contacts. Micronesica 33(1/2):111-136. ---. 2001a. Life on the edge: Prehistoric settlement and economy on Utrōk Atoll, northern Marshall Islands. Archaeology in Oceania 36(3):109-133. ---. 2001b. On the Margins of Sustainability: Prehistoric Settlement of Utrōk Atoll, Northern Marshall Islands. Oxford: Archaeopress. ---. 2002. Archaeological Survey and Test Excavations on Ebon Atoll, Republic of the Marshall Islands. Majuro Historic Preservation Office, Republic of the Marshall Islands. Weisler, M.I., R. Bollt and A. Findlater. 2010. Prehistoric fishing strategies on the makatea island of Rurutu. Archaeology in Oceania 45(3):130-143. Weisler, M.I. and P.V. Kirch. 1985. The structure of settlement space in a Polynesian chiefdom: Kawela, Molokai, Hawaiian Islands. New Zealand Journal of Archaeology 7:129-158. Weisler, M.I. and D. Swindler. 2002. Rocker jaws from the Marshall Islands: Evidence for interaction between eastern Micronesia and west Polynesia. People and Culture in Oceania 18:23-33.

151

Weisler, M.I., H. Yamano and Q. Hua. 2012. A multidisciplinary approach for dating human colonization of Pacific atolls. The Journal of Island and Coastal Archaeology 7(1):102-125. Wiens, H.J. 1962. Atoll Environment and Ecology. New Haven: Yale University Press. Wolverton, S. 2013. Data quality in zooarchaeological faunal identification. Journal of Archaeological Method and Theory 20(3):381-396. Wright, N.J., A.S. Fairbairn, J.T. Faith and K. Matsumura. 2015. Woodland modification in Bronze and Iron Age central Anatolia: an anthracological signature for the Hittite state? Journal of Archaeological Science 55(0):219-230.

152

Figures and Tables Table 1 Chord distance values for mollusc shell and fish bone assemblages retained in the 6.4 mm and 3.2 mm sieves for each site pair. 6.4 mm samples 3.2 mm samples

Site Pair molluscs fish bone fish bone

MLEb-1/MLEb-31 1.085 0.463 0.481 MLEb-1/MLEb-33 1.045 0.538 0.732 MLEb-31/MLEb-33 0.373 0.566 0.741

153

Table 2 Measures of taxonomic heterogeneity: NTAXA, Shannon's index of diversity (H') and evenness (E), Simpson's dominance (1-D) and Fisher’s α, as calculated for mollusc shell and fish bone assemblages retained in the 6.4 mm and 3.2 mm sieves for all sites (MLEb-1, MLEb-31 and MLEb-33). 6.4 mm samples 3.2 mm samples Index molluscs fish bone fish bone MLEb-1 MLEb-31 MLEb-33 MLEb-1 MLEb-31 MLEb-33 MLEb-1 MLEb-31 MLEb-33 NTAXA 47 37 26 27 25 18 29 25 18 1-D 0.887 0.859 0.869 0.921 0.926 0.909 0.936 0.940 0.919 H’ 2.648 2.528 2.512 2.780 2.822 2.605 2.968 2.972 2.705 E 0.687 0.700 0.711 0.844 0.877 0.901 0.881 0.923 0.936 Fisher’s α 9.631 8.508 7.534 6.803 6.306 8.071 7.316 7.670 15.51

154

155

156

157

Chapter 6: Two millennia of mollusc foraging on Ebon Atoll, Marshall Islands: sustained marine resource use on a Pacific atoll.

Note: This is the final version of a peer-reviewed article published in Archaeology in Oceania

Matthew Harris and Marshall I. Weisler

School of Social Science, The University of Queensland, St. Lucia, QLD 4072 Australia

Corresponding author: Matthew Harris, School of Social Science, The University of Queensland, St. Lucia, QLD 4072 Australia. Email: [email protected]

158

Abstract Small, remote islands, such as low coral atolls, with nutrient-poor, biogenic soils for food crops and vulnerability to extreme weather, have long been considered marginal environments for human habitation. Yet, four decades of archaeological research in the atoll archipelago of the Marshall Islands, eastern Micronesia have demonstrated sustained human occupation there for over two millennia. Here, we present a fine-grained analysis of mollusc remains from four recently excavated archaeological sites (4476 total MNI/ 14843 total NISP) combined with mapping and analysis of extant benthic habitats, on Ebon Atoll, Marshall Islands. We examine spatial and temporal variability in mollusc foraging practices from prehistoric village sites and ephemeral campsites across the windward-leeward exposure gradient. Our analysis demonstrates that foragers targeted a rich assemblage of taxa from different habitats, reflecting a foraging strategy that was adapted to local environmental conditions. Human foraging over 2000 years documented no observable human impacts to molluscs or nearshore intertidal marine ecosystems, challenging previous notions of atolls as marginal, exceptionally difficult settings for human habitation.

Keywords: archaeomalacology, shell midden, human impacts, Pacific Islands archaeology, intertidal ecosystems.

159

Introduction Coral reefs provided critical resources to the inhabitants of the Marshall Islands during the prehistoric period, but the long-term patterns of human behaviour and potential impacts to these environments as a result of mollusc foraging has not been well studied. The coral reefs of the Republic of the Marshall Islands (RMI) were considered to be in good condition about 10 years ago relative to global patterns of reef decline (Beger et al. 2008:412; Hughes et al. 2003;), but a range of anthropogenic stressors threaten these ecosystems (Houk and Musburger 2013). Rapid growth in human populations along coastlines during the last century has resulted in substantial impacts to intertidal marine environments, especially coral reefs (Brander et al. 2010). These stresses reduce biodiversity and cause substantial negative impacts to the communities that depend on coastal environments for subsistence, shoreline protection, water quality, lifestyle, and personal and cultural identity (Brander et al. 2010:65; Cinner 2014). The reefs of the Marshall Islands host at least 1655 mollusc species, and ~860 fish taxa (Richmond et al. 2000) that provide critical economic resources. Archaeological research has demonstrated that Marshall Islanders exploited marine resources for subsistence and raw materials since colonisation about 2000 years ago (Dye 1987a; Kayanne et al. 2011; Weisler 1999a, 2000), with little negative impact (Lambrides and Weisler in press; Weisler 2001b). However, the long-term patterns of mollusc foraging in these reef habitats is not well understood, especially as it relates to potential human impacts to coral reef ecosystems. Hayashida (2005:45) and other researchers have stressed the importance of historical data for understanding the current state of ecosystems due to the time-lag between disturbance and ecosystem change (Alleway and Connell 2015; Rick and Lockwood 2013).

Nearly four decades of archaeological research in the Marshall Islands has revealed a sequence of occupation that begins soon after atoll emergence, likely some time just before 2000 BP (Dye 1987a; Riley 1987: Table 2.28, Kayanne et al. 2011; Weisler 1999a, 2001a, 2001b; Weisler et al. 2012). In 1993, Weisler commenced an interdisciplinary project documenting the variability in colonisation dates, marine and terrestrial subsistence practices, and human impacts across the north- south rainfall gradient in the archipelago. As part of this project, Weisler (2001b) synthesised data on mollusc foraging in Marshallese palaeoeconomies, and broad foraging preferences and the differences in faunal assemblage composition at windward and leeward sites have also been explored (Harris et al. 2016; Harris and Weisler 2016). For the first time, we present a fine-grained spatio-temporal analysis of mollusc remains from recent archaeological excavations on Ebon Atoll, Marshall Islands. We examine mollusc assemblages from four archaeological sites representative of the major environmental and settlement types on atolls (windward/leeward and

160 permanent/ephemeral occupation). Our diachronic analysis, spanning two millennia of human occupation, examines long term human interaction with the marine environment and also investigates potential human impacts to the intertidal zone.

Traditional Marshall Islands Economy Atolls consist primarily of unconsolidated sediments forming atop a narrow reef platform that surrounds a lagoon formed through the development of fringing reefs around a subsiding volcanic island. These landscapes have been considered marginal and difficult for sustained human habitation due to a lack of standing fresh water and only nutrient-poor biogenic soils for the cultivation of food crops (Fosberg 1954). However, the subterranean Ghyben-Herzberg freshwater lens has been utilised for pit-cultivation of giant swamp taro (iraij, Cyrtosperma chamissonis) since colonisation (Weisler 1999a). Arboriculture including varieties of pandanus (bō b, Pandanus tectorius), as well as coconut (ni, Cocos nucifera), breadfruit (mā , Artocarpus altilis), and the low- lying arrowroot (makmō k, Tacca leontopetaloides) were also cultivated during the prehistoric period and continues to the present.

Marshall Islanders are accomplished seafarers, with frequent prehistoric interisland contact suggested by a single language spoken across the entire archipelago encompassing 2 million km2, with only minor, mutually intelligible dialect differences in the Ratak (eastern) and Ralik (western) island chains (Bender 1969:xii-xiii). There are more than 100 different Marshallese words describing fishing techniques, shell taxa, and the intertidal zone (Merlin et al. 1994) attesting to a deep and intimate knowledge of marine ecosystems (Abo et al. 1976:361-362, 453-454). The iroj (chief) sometimes restricted access to the reef, known locally as mo, which may have been invoked to conserve resources (Sudo 1984:208). Although birds and dogs contributed to the diet, the major source of protein was from the sea, with the intertidal zone utilised for , fishing and mollusc foraging. Archaeological deposits in the RMI indicate continuous exploitation of molluscs and finfish over the course of human occupation on Utrō k, Ujae, Arno, Majuro and Ebon Atolls (Dye 1987b; Harris and Weisler 2016; Riley 1987; Weisler 1999b, 2001b), with most assemblages consisting of a rich, even, and diverse range of taxa. Weisler (2001b:116) documented no indications of human impacts to molluscs on Utrō k Atoll, inferred to be the result of low human populations and expansive reef flats. There is also no evidence for adverse prehistoric human impacts to finfish resources (Lambrides and Weisler in press; Weisler 2001b).

161

Sites and Samples We focus on excavations from four archaeological sites on Ebon Atoll: prehistoric village sites MLEb-1 and MLEb-5 on Ebon Islet, MLEb-33 on Enekoion Islet, and an ephemeral campsite, MLEb-31, on Moniak Islet (Figure 1). All deposits were excavated in arbitrary 10 cm spits within stratigraphic layers. Although sediments of selected units were wet-sieved through 3.2 mm mesh, mollusc remains analysed here were wet screened through 6.4 mm mesh. Water worn shell, samples with evidence of clionid sponge adherence to the interior of the shell, or samples exhibiting other forms of damage indicating secondary deposition were not retained for analysis. The few gastropod specimens (< 1%) showing evidence of hermitting were excluded from analysis. All samples were cleaned of sediment prior to identification and quantification. Smaller taxa such as Melampus flavus, Planaxis sulcatus, and Ctena bella are included in the following analyses as they may represent areas of the littoral environment that were accessed by foragers while collecting other taxa, and are useful for characterising the assemblage.

Archaeological deposits on Ebon Atoll villages generally consist of gravelly sand layers, the product of reworked coral gravel pavements, deposited atop sterile sands representing the original ground surface prior to human habitation. Ephemeral or short-term occupation campsites lack coral gravel pavements, while agricultural trenches have sparse midden and occasional combustion features. Agricultural trenches sometimes revealed a buried A horizon that was covered with spoil during the initial excavation of aroid pits (Weisler 1999a). Cultural material at all sites consisted primarily of fish bones, mollusc shells, shell artefacts, and combustion features. We present five AMS radiocarbon age determinations on coconut endocarp and Pandanus drupe that were calibrated to two standard deviations using the IntCal09 curve and the OxCal program (Reimer et al. 2009). The details of all chronometric dates from Ebon Atoll will be presented elsewhere (Weisler et al. in prep.).

Ebon Islet MLEb-1 is located centrally within the ~2 km long prehistoric village near the lagoonside of Ebon Islet. This village system consists of several archaeological sites and associated aroid pits, and is currently the major population centre of the atoll. A single test pit (TP 6) and three larger excavations (TP 17 – 20, TP 8 – 12, TP 13 – 16) totalling 14 m2 were conducted along a transect from the lagoon shore to the interior aroid pits. A 2 m × 2 m area excavation (TP 17 – 20) was positioned 20 m north-west of the primary school and 40 m inland of the lagoon shore on a low mound primarily built from successive coral gravel pavements. Cultural deposits extended to 1.75 162 m. Radiocarbon age determinations from Layer IIIA on coconut endocarp from the base of the lowest cultural deposits in TP17 and TP 19 indicate initial occupation here sometime after 925-790 cal BP ( 940 ± 25 BP OZP-927). TP 6 is located 150 m inland of the lagoon, within an area previously identified as an early occupation site (Weisler et al. in prep.). Cultural deposits extended to 1 m and included mollusc remains, charcoal, um or oven stones, and prehistoric artefacts. A single radiocarbon date from TP 6, Layer IIIA revealed a basal age of 1895 – 1730 cal BP (1890 ± 30 BP, OZP925). Additional excavations of two trenches excavated across the spoil heap of aroid pits, with a 1 × 5 m trench (TP 8 – 12) located 185 m inland of the lagoon and TP13 – 16 located 250 m from the lagoon. Cultural deposits were relatively shallow, with material located to 0.23 m deep in TP 8 – 12 and 0.30 m at TP13 – 16. A radiocarbon date from the base of TP 15 yielded a date of 460 – 295 cal BP (300 ± 25 BP, OZP926).

MLEb-5 is located ~650 m southeast of MLEb-1. A 3 m × 5 m unit (TP 1, 13, and 15-27) was excavated into a low mound 120 m from the lagoon shore. Cultural deposits extended to a depth of 0.5 m, totalling 8.25 m3. The earliest dates for the atoll were encountered at MLEb-5, with initial occupation of the mound at 2295 - 1995 cal BP (2115 ± 30 BP, OZP932). Cultural material included coral gravel spreads, charcoal, um stones, primarily fish and molluscs, but also human remains, and shell artefacts.

Enekoion Islet A prehistoric village site (MLEb-33) was located on Enekoion Islet. We analysed mollusc assemblages from six test pits excavated around a large (~236 m2) aroid pit between 25 m and 100 m from the lagoon shore. Here, TP 1, 6, and 7 are aggregated due to proximity and to maximise sample size. Units 2 and 8 form a contiguous 1 m × 2 m trench, but TP 3 is analysed individually. Cultural material from MLEb-33 consisted of mollusc remains, fish bone, worked shell artefacts, coral gravel spreads, and um stones. Cultural deposits were relatively shallow, extending to a maximum depth of 0.4 m in TP 2 and 8. A single radiocarbon age determination from an oven, 27 cmbs (centimetres below surface) in TP1 yielded a date of 505 - 315 cal BP (370 ± 30BP, OZP318 BP).

Moniak Islet Previous excavations by Weisler had located cultural deposits at MLEb-31 indicating a relatively small site occupied for short periods of time for acquiring seabirds, intertidal foraging for molluscs, and as a place for staging fishing sorties (Weisler 2002). During the 2011-2012 field season, four 163 test pits (TP 2 – 6) were excavated over an area between 55 m and 125 m from the current lagoon shore. Cultural deposits extended to 0.70 m, interrupted between 20 – 45 cmbs by a culturally sterile storm deposit. Large storm events are historically recorded for the atoll, with a large typhoon that occurred during the 1850s causing widespread food shortages and damage to horticulture (Mason 1950).

Habitat Mapping Methods and Description of Marine Habitats Atoll environments are dynamic and the configuration of intertidal habitats can vary over short time scales, but modern marine habitats adjacent to archaeological sites on Ebon (MLEb-1 and 5), Enekoion (MLEb-33), and Moniak (MLEb-31) islets provide useful comparative data for understanding prehistoric subsistence practices. Benthic habitat maps were generated using protocols outlined in Kendall et al. (2012). The minimum mapping unit (MMU) is the minimum size of a single feature (e.g. area of sandy bottom, patch of coral) for delineation. The MMU size can influence landscape metrics as well as indices of heterogeneity and diversity (Kendall and Miller 2008;). Broadly, a smaller MMU will more effectively capture diminutive and rarer habitats, where larger MMU will merge these smaller habitats into broader categories, sacrificing some analytical resolution (Kendall & Miller 2008). We adopted a minimum mapping unit (MMU) of 1000 m2 to align with existing maps of Majuro Atoll (Kendall 2012, Kendall & Miller 2008) and to account for the lack of in situ calibration data. High-resolution (e.g. 100 m2 MMU) differences in habitat structure are not necessary when broad characterisation of extant marine environments is the aim for archaeological hypothesis testing. All maps were digitised at 1:2500 scale, providing a balance between map detail and expediency (Kendall et al. 2001:38). Orthorectification for all satellite imagery was fixed as imagery was generated by the Arc2Earth plugin for ArcGIS. Village structures and paths present in the 1995-1996 and 2011-2012 field seasons were used as ground control points for site and test pit locations. The relative areas of each habitat mapped within a 2 km radius of each study site are presented in Appendix A. Below we summarise the major differences between islets, noting common mollusc taxa and report the Shannon-Wiener index of diversity (H’) for each mapped area.

Habitats on each islet broadly reflect the degree of exposure to winds, waves, and currents (Figure 2). Habitats adjacent to MLEb-1 and MLEb-5 on the largest islet (~2.2 km2) of Ebon were the most diverse (MLEb-1, H = 2.663; MLEb-5, H = 2.557) of the mapped sites. Diversity was marginally lower at MLEb-33 on Enekoion Islet (H = 2.165) and least diverse at MLEb-31 on Moniak Islet (H = 0.693). Ebon islet features relatively calm, expansive reef flat pavements that host a range of

164 molluscs on the oceanside, characteristic of large, leeward islets (Kendall et al. 2012; Wiens 1962). These pavements are colonised by macroalgae and feature shallow tide pools and sand channels that commonly host Cypraeidae, Muricidae, Trochidae, Conidae, and Neritidae (Figure 2C). Enekoion (MLEb-33), to the north, is moderately more exposed than Ebon, reflected in a large area of reef flat rubble on the oceanside (Figure 2E). Moniak also features an expansive reef flat pavement, but these habitats are more exposed to high-energy waves and are consequently scattered with coral cobble and reef rubble. These habitats characteristically host predatory molluscs such as Conidae and Muricidae, and Cerithiidae, Cypraeidae, and Trochidae. The spur and groove zone that occurs oceanward of the reef flat habitats is expansive on Ebon, Moniak and Enekoion islets. These habitats characteristically host Turbinidae, Drupa spp., and large Cypraeidae. An algal ridge is also present on all three islets, just shoreward of the spur and groove topography, hosting Trochidae, Turbinidae, and Drupa spp.

The lagoonside habitats of Ebon and Enekoion islets are a complex mosaic of large areas of live coral, sand flats, and seagrass beds (Figure 2A-E). Patch reefs and aggregate reef account for a large portion of mapped area for these islets, but were not present in mapped areas of Moniak Islet (Supplementary Appendices A, B). The less complex habitats on Moniak are likely correlated with a decrease in richness and diversity of mollusc communities compared with Ebon and Enekoion (Gratwicke and Speight 2005; Kohn and Leviten 1976). Expansive seagrass (cf. Thalassia hemprichii) beds are present in the upper- and mid-intertidal of Ebon and Enekoion lagoonside, but are absent at Moniak. These habitats provide contemporary inhabitants of Enekoion Islet with strombs, indicated by the frequency of discarded Conomurex luhuanus shells around the modern village. Modern studies of I-Kiribati atoll dwellers have highlighted the importance of seagrass beds for modern foraging activity and archaeological deposits from Palau and Majuro demonstrate that this taxon was exploited prehistorically (Giovas et al. 2010; Riley 1987). The shoreline and high- intertidal of Ebon and Enekoion is generally sand, where Moniak is primarily rocks and boulders that slope upwards from the reef flat to the islet surface (Figure 2G). The sandy habitats of Ebon and Enekoion host Asaphis violascens and Gafrarium spp., where the rocky shoreline of Moniak offers habitat for Cerithiidae, Cypraeidae, Trochidae, Turbinidae, and Neritidae.

Methods of Mollusc Analysis Molluscs were identified to the lowest possible taxonomic level using Indo-Pacific molluscan reference collections housed at the University of Queensland Pacific Archaeology Lab, and reference books and manuals (Carpenter and Niem 1998; Cernohorsky 1967; Poppe 2008).

165

Identifications were based only on diagnostic features on individual fragments, rather than assuming fragments derive from common taxa (Driver 2011; Szabó 2009; Woo et al. 2015). All taxonomic nomenclature was verified following the World Register of Marine Species online database (WoRMS Editorial Board 2015). The Number of Identified Specimens (NISP), weight (recorded to the nearest 0.01g), and Minimum Numbers of Individuals (MNI) were recorded for all taxa. MNI is used as the primary measure of relative abundance, with NISP and weight providing comparative data for assessing taphonomic factors such as fragmentation and post-depositional leaching of carbonate (Faulkner 2011; Giovas 2009; Szabó 2009). Metric analysis of major taxa and meat-weight reconstructions (e.g. Allen 2012) were not undertaken as MNI, NISP and weights were more appropriate measures for our analytical purposes. MNI was calculated according to the protocols outlined in Harris et al. (2015). Non-repetitive element (NRE) frequency for each taxon was summed from 10 cm spits aggregated to cultural layer prior to the calculation of MNI. As stratigraphy was broadly continuous across MLEb-31 deposits, TP 2-6 were treated as a single cultural deposit to increase sample size. All other test pits were analysed individually or as aggregated deposits where units were contiguous.

Assemblage richness (the number of species in an analytical unit) was measured by collapsing taxonomic categories to the highest common level (NTAXA) to avoid duplicate counts of categories and to provide conservative estimates of richness. Taxonomic diversity and evenness was measured using the Shannon-Wiener index of diversity (H’), Shannon’s evenness (E), Simpson’s dominance (1-D), and Fisher’s α. H’ and E provide a measure of the number of individuals per species in an analytical unit (taxonomic evenness). H’ values generally fall below a theoretical maximum of 5, between 1.5 and 3.5, with higher values indicating greater species richness and diversity. E values range between 0 and 1, with values closer to 1 indicating assemblages with a rich and even composition (Lyman 2008:195). Dominance was measured using 1-D, where values range between 0 and 1, with low values indicating dominance of a single taxon (Magurran 2004:116). Due to the sensitivity of H’ and E to sample size, Fisher’s α was used as values are relatively sample size independent, tracking the number of taxa represented by only single individuals to measure overall diversity. Statistical significance of alterations to these measures of taxonomic heterogeneity was assessed using random permutation tests of relative abundance data. Chord distance, a scaled measure of Euclidean distance, was used to analyse similarity and difference in taxonomic composition between cultural layers (Legendre and Gallagher 2001). Chord values range between 0 and , with higher values indicating increasing distance, or dissimilarity

166 between samples. All statistical analyses were conducted using PAST Palaentological Statistics Package, Version 3.14 (Hammer 2001).

Habitat preferences were explored using the methods outlined in Harris and Weisler (2016). This is a probabilistic method for reconstructing forager behaviour that assigns mollusc taxa to habitat categories using a hierarchical classification scheme for benthic habitats (Appendix B). We also included an analysis of potential alterations to trophic networks and other indirect effects of human foraging by assigning mollusc taxa feeding-habit and behavioural characteristics (Appendix C, Tables 1- 5). All ecological data were drawn originally from the Neogene Marine Biota of Tropical America Molluscan Life Habits Database (Todd 2001) and verified with appropriate literature if necessary. Gastropod taxa are assigned to one of seven categories (Appendix C, Table 1) and as either epifaunal or infaunal based on dominant mode of life. Bivalve taxa were assigned an organism/substrate relationship, feeding type, mobility, and shell fixation method (Appendix C, Table 2 - 5). Analysis of alterations to the relative abundance of taxa assigned to these categories allows assessment of potential alteration to trophic networks due to human exploitation or other factors. Only genus-level identifications were used for these analyses. Family-level identifications were rare (3 % of total MNI), but introduce considerable uncertainty in foraging reconstructions that potentially mask variation.

These statistical tests are used to assess the relationship between site location (e.g. windward v. leeward, near lagoon v. interior), inferred site function (e.g. village or campsite), and the taxonomic composition of mollusc assemblages from our four study sites on Ebon Atoll, Marshall Islands. For the first time, these methods will also be used to assess temporal changes in human foraging of molluscs in the southern Marshall Islands to assess long term patterning in human interaction with the intertidal zone. Taxonomic measures of heterogeneity (H’, E, 1-D, Fisher’s α), chord distance, and habitat representation are complementary analyses that allow a comprehensive characterisation of assemblage variation (Faith 2013; Harris et al. 2016) and potential human impacts to mollusc fauna and intertidal environments on Ebon Atoll over the last 2000 years.

Results The Ebon Atoll mollusc assemblage (Table 1) is well preserved. Minimal chemical dissolution and burning (< 1 %) of shell were noted for the sites analysed. Fragmentation measured by MNI/NISP was moderate at 0.31, with many diagnostic fragments or complete mollusc shells recovered. A total MNI of 4476 (14843 NISP) was reported from 199 taxonomic categories, with 119 to species,

167

60 to genus, and 20 to family or subfamily (Appendix D). Gastropod taxa dominate, accounting for 81.8 % of total MNI (MNI 3822). Gastropod taxa are predominately epifaunal (80.1 % MNI) predatory carnivores (33 % MNI) or herbivores on rock and rubble substrates (40.5 %). Bivalve taxa are largely unattached (16 % MNI) infaunal siphonate taxa (13.7 % MNI). Archaeological molluscs (Figure 3) are mostly taxa that inhabit the reef flat rocks and boulders (e.g. Nerita spp., Vasum turbinellus, Thais armigera), pavements (e.g. Nerita spp., Cerithium nodulosum, Conus spp., Drupa spp., Monetaria moneta), and sand flats (e.g. Guturnium muricinum, A. violascens) (Supplementary 4), but previous analyses have demonstrated that locally available habitats mediate human selection on Ebon Atoll and that there is evidence of collection from a wide range of habitats (Harris et al. 2016; Harris and Weisler 2016). Measures of taxonomic heterogeneity reported a diverse and even assemblage overall (NTAXA = 49, H’ = 2.660, E = 0.683, 1-D = 8.914). Richness is significantly correlated with sample size (rs = 0.67, p = <0.001). We present analytical summaries of trends in measures of taxonomic heterogeneity, relative abundance of feeding types and mode of life, and the habitats from which most of the taxa in the assemblage could have been collected.

MLEb-5 3m × 5m area excavation A large sample of 4401 NISP yielded an MNI of 988 from the only cultural layer, Layer I. The assemblage was rich (NTAXA = 45), but reported low evenness for Ebon Atoll deposits (Table 1) with low density overall (136.75 MNI/m3). The assemblage was dominated by gastropods (72.1 % MNI), primarily Nerita plicata and N. polita (33 % of total MNI). Major taxa aside from the nerites include A. violascens (MNI 108), Gafrarium spp. (MNI 65), Conus spp. (MNI 62), and Cypraeidae (MNI 29). The range of taxa present indicate an emphasis on collection of epifaunal, primarily herbivorous gastropod taxa from mid-intertidal and upper intertidal habitats, including reef flat and shoreline rocks and boulders, reef flat pavements, sand and rubble. The largest sample of Tridacna (primarily T. maxima) was recovered from this deposit (MNI 42, NISP 698), accounting for 39 % of total MNI for Tridacna across all sites. This sample was the most heavily fragmented of any Tridacna assemblage on the atoll (MNI/NISP = 0.04). The only other deposit with comparably high fragmentation was from the other early period deposit on Ebon Islet, MLEb-1, TP 6 (MNI/NISP = 0.11).

MLEb-1 TP 6 Molluscan samples from early deposits (Layers IIB – IIA) are small, with a total MNI of 25 and taxonomically even with low richness (Figure 4). In contrast with most Ebon assemblages that are dominated by gastropods, most individuals from Layers IIB and IIA are infaunal bivalves 168 derived from reef flat sands, with A. violascens and Gafrarium spp. accounting for 64 % of MNI. The assemblage from Layer IB is similar, with a continued dominance (53.7 % MNI) of molluscs from sand habitats, Gafrarium spp. and A. violascens. NTAXA and evenness increase with sample size. Chord distance reported high dissimilarity (chord = 1.235) between Layer IB and the uppermost cultural layer, IA. This is correlated with an increase in the relative abundance of Conus spp., Nerita polita and V. turbinellus in Layer IA, with an overall increased abundance of taxa derived from the reef flat pavement (52.3 % of MNI). Concurrently, there is an increase in the abundance of carnivorous gastropods, accounting for 53.6 % of MNI compared with 20-36 % in Layers IIB-IB.

TP 17-20 The largest sample of molluscs on Ebon Islet was recovered from TP 17-20 (Figure 5), with a total MNI of 1273 (NISP 3597). Lower deposits (Layers IIB-IIA) are typically diverse, but somewhat less even than most Ebon assemblages (Table 1). Three taxa, N. plicata, N. polita and A. violascens account for 43.7 % of MNI in IIB and 40.7 % of MNI in IIA. Taxa from these relatively early layers could have been gathered from the reef flat pavements (IIB: 57 % of MNI, IIA: 54.5 %), rocks and boulders on the reef flat (IIB: 53.5 %, IIA: 49.1 %) and shoreline (IIB: 41.9 %, IIA: 40.8 %), and reef flat rubble (IIB: 44.2 %, IIA: 49.9 %).

While reef flat pavements are highly represented in all layers, and N. plicata, N. polita, and A. violascens are also highly ranked throughout, Layer IB to IA have an increased abundance of taxa from coral habitats. In Layers IIB and IIA coral habitats account for 2.7 – 17. 4 % of MNI. In Layer IB, reef flat coral habitats account for 32.8 % of MNI and peak in abundance in Layer IA, accounting for 44.3 %. Across the four cultural layers, molluscs from shoreline rocks and rubble decline from 41.9 % in IIB to 14.7 % in IA and taxa from seagrass beds increase from 11.6 % in Layer IIB to 30.5 % in Layer IA, potentially indicating a development or expansion of these habitats on the lagoonside, or a shift in foraging behaviour. A concurrent increase in predatory carnivores driven by increasing abundance of Monoplex spp. and G. muricinum, marks layers IA and IB as taxonomically quite different to earlier layers. Chord distance reported high dissimilarity between Layer IIA and IB, and paired diversity permutation tests reported significant difference for 1-D, H and E between IIA/IB (1-D: p = < 0.01; H: p = < 0.01; E: p = 0.02) and IB/IA (1-D: p = < 0.01; H: p = 0.02; E: p = < 0.01). A chi-square test indicated a significant association between layer and feeding type X2 (9, n = 1415) = 93.234, p = < 0.001, Monte-Carlo p = 0.0001, V = 0.14781. Cramer’s V showed that the effect size was small to medium (Cohen 1988). Examination of adjusted residuals showed that there were significantly greater numbers of predatory carnivores

169

(CP) in IA and significantly more herbivorous (HR/HM) individuals in Layers IIA and IIB. These trends also hold when herbivores and carnivores are collapsed to a binary carnivore or herbivore, rather than the more detailed categories of CP, HO, HR, and HM.

Agricultural trenches (TP 8-16) Mollusc assemblages from the two agricultural trenches on Ebon Islet had relatively small samples (TP 8-12: 47 MNI/141 NISP, TP13-16: 130 MNI/257 NISP) that were rich, even, and diverse. Fragum fragum, N. polita, and A. violascens were top ranked in the single cultural layer in TP 8-12, with 61.4 % of total MNI derived from reef flat sands. TP 13-16, oceanward of TP 8-12, had increased abundances of taxa that could have been collected from reef flat pavements in Layers A to I, with M. moneta, Conus spp., N. plicata, and Engina mendicaria accounting for 17 % of MNI in Layer I and Bursa spp., V. turbinellus, Tridacna maxima, and C. nodulosum accounting for 20 % of total MNI Layer A. The differences in taxonomic composition are reflected in high dissimilarity as measured by chord distance (TP 8-12/A – TP 13-16/I = 1.122, TP 8-12/A – TP 13-16/A = 0.910). Both trenches reported high abundances of epifaunal gastropods, with bivalves more abundant in TP 8-12 than in TP 13-16, possibly due to the increased proximity to the lagoonside sands.

MLEb-33 TP 1,6-7 The mollusc assemblage from the single cultural layer (Layer I) within TP 1, 6, and 7 was dense (447.2 MNI/m3), rich, diverse, and even (Table 1, Figure 6). A total NISP of 1392 yielded 559 MNI. The five highest ranked taxa are N. polita (MNI 121), Fragum sp. (MNI 67), V. turbinellus (MNI 45), N. plicata (MNI 43), and M. flavus (MNI 32). Most taxa in the assemblage likely derive from rocks and boulders on the reef flat (57.4 % MNI) and shoreline (38.9 % MNI), and reef flat pavements (38.2 % MNI). The majority of individuals are herbivorous gastropods (50.3 % MNI) that live epifaunally. Predatory carnivores account for 25.8 % of MNI, primarily V. turbinellus and Conus spp., characteristic of the rocky, exposed shorelines of the oceanside reef flat of Enekoion Islet.

TP 2 and 8 N. polita, V. turbinellus, Conus spp., Turbo argyrostomus, and C. nodulosum constitute 62.3 % of MNI from TP 2 and 8 cultural deposits. The assemblage was generally rich and even (Figure 6) with most taxa derived from reef flat and shoreline rocks and boulders, and reef flat pavements (34.2 % – 44.2 % of MNI). Epifaunal herbivorous (48.5 % of MNI) and carnivorous (35.9 % of MNI) gastropod taxa were typically most common.

170

TP 3 A total of 331 MNI was recovered from only 0.2 m3, making this assemblage the densest recovered from Ebon Atoll (1655 MNI/m3). Four gastropod taxa (C. nodulosum, Conus spp., V. turbinellus, and N. polita) comprise 56.5 % of total MNI, reflected in relatively low evenness values compared with other Ebon Atoll mollusc assemblages (H’ =2.4330, E = 0.756). Reef flat pavements (67.1 % MNI) and shoreline rocks and boulders (49.2 % MNI) are the two highest-ranked habitats (Figure 6).

MLEb-31 TP 2-6 The cultural layers of MLEb-31 (Figure 7) are divided by a storm deposit (Layers IC and SD) that is delineated by large coral cobbles and coarse biogenic gravel dividing Layers IA and IB from IIA and IIB. Assemblage composition was relatively static, with all cultural layers primarily composed of epifaunal gastropods, with a generally even contribution of both herbivorous and carnivorous taxa. Layers IIB and IIA report high relative abundance of N. polita, V. turbinellus, Conus spp., A. violascens, and C. nodulosum, with these taxa accounting for 65 % of total MNI in Layer IIA. These taxa are highly abundant in the layers proceeding the storm event, with the same taxa accounting for 51.5 % of total MNI in IA, and 45.2 % in IB. Moniak assemblages reported some of the lowest evenness scores for Ebon Atoll, but still represent a non-selective foraging strategy as reflected in Fisher’s α and overall richness. The majority of taxa across all cultural layers could have been gathered from the shoreline rocks and boulders and the reef flat pavement, consistent with the results of previous analysis of this assemblage (Harris et al. 2016; Harris and Weisler 2016).

Tests for human impact Diversity permutation tests reported non-significant differences between cultural layers in all contexts except for MLEb-1, TP 17-20, and MLEb-31, TP 2-6. Significant differences were reported for MLEb-31 between layers with and without evidence of storm deposited molluscs. Fisher’s α reported significant difference from cultural layers both above (Layer IB) and below (IIA) the storm deposit layers (SD/IC). These differences in Fisher’s α reflect that almost all taxa in the storm deposit layers are represented by only a single individual. Results from MLEb-1 indicate significant differences between cultural layers for 1-D, H and E between Layers IA-IB (1-D: p = < 0.01; H: p = 0.02; E: p = < 0.01) and Layers IB-IIA (1-D: p = < 0.01; H: p = < 0.01; E: p = 0.02).

Overall, chord distance reported moderate dissimilarity between cultural layers (x̄ = 0.924). The greatest dissimilarity between cultural layers were reported for MLEb-1, TP 17-20 between Layers 171

IIA and IB (Chord = 1.187), MLEb-1 TP 6 IIA-IB / IB – IA (Chord = 1.236 / 1.235). The high chord distance values for TP 17-20 correlate with the general shift in species abundances and assemblage composition between upper and lower strata. At MLEb-31, high chord distance values were reported for layers above and below the storm deposits, but this is likely due to sample size changes between layers, rather than reflecting assemblage composition overall. In addition, a chi- square test indicated a significant association between layer and feeding type at both MLEb-1 (X2 (9, n = 1415) = 93.234, p = < 0.001, Monte-Carlo p = 0.0001, V = 0.14781) and MLEb-31 (X2 (15, n = 829) = 32.86, p = 0.005, Monte-Carlo p = 0.007, V = 0.11495). Cramer’s V showed that the effect size was small to medium in both cases (Cohen 1988). For MLEb-1, examination of adjusted residuals showed that there were significantly greater numbers of predatory carnivores in IA and significantly less in Layers IIA and IIB. Additionally, in Layers IIA and IIB there were significantly more herbivores on rock, rubble or coral substrates, and more herbivores on fine-grained substrates. At MLEb-31, examination of residuals showed that there were significantly greater numbers of herbivorous omnivores in IB, just after the storm deposits and significant numbers of greater predatory carnivores in Layer IIB, the earliest cultural layer prior to the storm event.

Discussion Temporal and spatial trends in mollusc foraging Previous analyses (Harris et al. 2016; Harris and Weisler 2016) have demonstrated that site-level taxonomic composition of mollusc assemblages on Ebon Atoll is broadly reflective of a fine grained, non-selective foraging strategy mediated by the availability of taxa in marine habitats adjacent to archaeological sites and differences in local habitat according to the windward-leeward exposure gradient. The analysis presented here supports these results, demonstrating that these foraging practices can be inferred for most of the sequence. Ebon assemblages from early period through late deposits are rich, even, and diverse. The taxa broadly represent local habitats, though additional palaeoenvironmental and geomorphological data are necessary to reconstruct the configuration of marine habitats at the time of initial colonisation. Here, analysis of the extant marine environment with benthic habitat maps has demonstrated broad similarities between intertidal configuration on each islet and the molluscs represented in archaeological assemblages.

While there are minor alterations to relative abundance through time, five taxa (Nerita polita, N. plicata, Conus spp., Vasum turbinellus, and A. violascens) are highly ranked at all sites, with N. plicata most common in leeward Ebon deposits and N. polita and V. turbinellus more common from assemblages on windward islets, likely due to local environment. Small-bodied gastropod taxa seem 172 to have been the major focus, with bivalves only a minor component of most assemblages. Gastropods are mostly epifaunal, herbivorous taxa that could have been collected by gleaning. Where bivalve taxa were present, infaunal siphon-feeding taxa were most common, suggesting, unsurprisingly, digging for bivalves from the intertidal and shoreline sands. However, the majority of prehistoric collecting activity seems to have occurred on the reef flat pavement and reef flat boulders, with these habitats consistently ranked first at all sites.

Variation in mollusc assemblages relates to the windward-leeward exposure gradient (Harris et al. 2016; Harris and Weisler 2016). There is some evidence for intra-islet variation in assemblage composition that may be related to site function (i.e. village sites v. midden deposits associated with aroid pits) and distance to the ocean or lagoon shore. Minor differences in taxonomic composition were noted between lagoonside and interior deposits at MLEb-1 and MLEb-33. Lagoonside deposits tend to be richer and evenness values for interior deposits was generally higher. Lagoonside deposits at MLEb-1 and MLEb-5 reported some of the lowest evenness scores for the atoll, with the majority of MNI contributed by few taxa, notably nerites, ranellids, and psammobids. Interior deposits also have a higher relative abundance of taxa from the oceanside reef edge, including V. turbinellus, T. argyrostomus, and T. maculatus. In other Marshall Islands assemblages, distance to the oceanside is correlated with higher abundances of these and other lower intertidal/reef edge gastropod taxa (Weisler 2001b). This is in line with expectations, as taxonomic composition at other sites on Ebon Atoll tend to reflect adjacent local habitats.

There were minor differences from these non-selective foraging patterns, both from early deposits on Ebon Islet. Early deposits at MLEb-1, TP 6 reported the highest relative abundance of bivalves of any cultural deposit, suggesting the possibility of an early focus on A. violascens and Gafrarium spp. soon after Ebon was colonised. However, sample sizes were small and these conclusions remain tentative as other, larger early deposits at MLEb-5 are dominated by gastropods. At MLEb- 5, however, the largest assemblage of Tridacna (primarily T. maxima) was recovered. Several worked Tridacna fragments appear to be tool-making debitage. This site has also previously been identified as the location of the largest assemblage of skipjack tuna (Katsuwonis pelamis) on Ebon Atoll (Lambrides and Weisler in press) and the early date in addition to the substantial K. pelamis and Tridacna spp. assemblage point towards a site of some importance in the early period occupation on Ebon Atoll.

173

Assemblages from late period village sites on Ebon and Enekoion islets have similar taxonomic composition and are consistent with mollusc assemblages from village sites on other Marshall Islands atolls, being rich, diverse, and even (Dye 1987b; Riley 1987; Weisler 2001b). While there was some variation in taxonomic composition between deposits at MLEb-1 and MLEb-33, with rocky, oceanside taxa such as V. turbinellus more common at Enekoion, and coral dwelling taxa more common on Ebon, both sites reflect a similar foraging strategy focused on the intertidal reef flat and gleaning. Riley (1987) reports assemblages from sites at Laura Village, Majuro Atoll, composed of as much as 80 % C. luhuanus, but no such focus on a single taxon was reported for any Ebon site. The seagrass beds that currently dominate the lagoonside habitats of Enekoion Islet were either not accessed by foragers in prehistory, or more likely, given the exploitation of these habitats in the region throughout prehistory (e.g. Giovas et al. 2010), were not present on Enekoion until recently, perhaps only historically. While late-period village sites reported relatively low evenness scores for Ebon Atoll assemblages, no taxa accounted for greater than 25 % of MNI. Even early village site assemblages at MLEb-5, where evenness was relatively low, no single taxon accounted for greater than 22 % of MNI. Ephemeral village sites showed a similarly wide range of taxa exploited, with taxonomic composition reflective of local environment, but still no taxa accounts for greater than 32 % of MNI for any cultural layer. The non-selective foraging strategy inferred here was mediated strongly by the richness and diversity of local marine environments. Thus, the low scores for evenness at Moniak may also relate to wave exposure, local habitat complexity, and substrate rugosity as taxonomic richness and diversity are generally lower in less complex marine environments, such as pavements on windward islets, compared with, for example, areas of coral growth on leeward islets (Gratwicke and Speight 2005; Kohn and Leviten 1976). However, a focus on a wide range of taxa, spread relatively equally across functional groups and feeding guilds, likely mitigated long-term human impacts on mollusc populations on Ebon Atoll (Giovas 2016).

A range of archaeological and neo-ecological studies have demonstrated that human foraging for molluscs can lead to alterations in species richness, abundance, and diversity contributing to wider scale ecosystem impacts (Allen 2003; de Boer and Prins 2002; Erlandson et al. 2011; Giovas et al. 2013; Giovas et al. 2010; Harris and Weisler in press;). However, the overall picture from Ebon Atoll is one of molluscs as a minor and sustainable contributor to diet during the prehistoric period, similar to Utrō k Atoll to the far north (Weisler 2001b) and other small islands occupied during the Holocene both within and beyond the Pacific Ocean (Giovas 2016; Thomas 2014). There is scant evidence for major ecological alteration resulting from human foraging of molluscs, with little

174 evidence for changes in habitat selection, richness, evenness or diversity, and little indication of impacts to particular trophic groups based on feeding type or mode of life analysis. Archaeologists exploring small islands and historical ecology have posited that conservation was not necessary in these contexts as small population size and abundant marine resources limited human impacts (Thomas 2014; Weisler 2001b). Additionally, foragers on small islands practicing generalised collection strategies or ‘reef-sweeping’ may have mitigated impacts by taking a range of taxa from similar feeding guilds or functional groups (Giovas 2016; Szabó 2009, cf. Szabó and Anderson 2012). A similar foraging strategy on Ebon Atoll, focused on relatively equal proportions of taxa from a range of functional groups from the reef flat pavements and reef flat boulders, may have limited human impacts to a degree that they are not visible in the archaeological record.

While significant differences were noted between measures of heterogeneity, chord distance reported high dissimilarity and a chi-square test showed association between layer (midden versus storm deposit) and feeding type at MLEb-31. The cultural deposits at MLEb-31 are interrupted by storm-deposited coral rubble and the alterations in mollusc composition are likely related to this event rather than human impacts. Kohn (1980) surveyed populations of predatory intertidal gastropods before and after a typhoon on Eniwetok Atoll in the northern Marshall Islands, reporting that taxa that inhabit refuges such as cracks, crevices, and depressions in the reef flat will be adequately protected during storm events. Shell architecture was also an important determinant in mortality rates, with Muricidae, especially Drupa spp., suffering lower mortality rates due to a shell and foot that are well adapted to resist damage and dislodgement from wave action. However, taxa that inhabit algal turfs were significantly impacted and suffered high mortality. Trends in the assemblage generally correlate well with these findings, with the storm event coinciding with a decline in Conidae relative abundance and an increase in Muricidae, especially Drupa and Thais. In Layer IB, the increase in herbivorous taxa is driven primarily by an increase in N. plicata that inhabit refuges in the upper intertidal and reproduce and disperse via planktonic larvae (Chelazzi and Vannini 1980; Crandall et al. 2008; Demond 1957). It is possible that N. plicata resisted high storm mortality rates through a combination of reproductive strategies and being adapted to high- energy environments led to increased abundances. Therefore, high chord values and significant difference between measures of heterogeneity in this layer likely result from these factors, rather than an anthropogenic cause.

Only a single context had any alterations to assemblage composition that could represent human impact: TP 17-20 at MLEb-1. Significant differences were noted for 1-D, H, E, rank order, and

175 feeding categories between cultural layers, concurrent with a shift in sub-dominant habitats and an increase in richness and density. From Layer IIB-IA, the proportion of carnivorous gastropods increased, peaking to 52.5 % of assemblage MNI in IA. At the same time, evenness declines sharply due to the increase in the abundance of G. muricinum, from 0.4 % of MNI in IIB to 16.3 % of MNI in IA. This shift is concomitant with a decrease in evidence for foraging among shoreline rocks and boulders and an increase in evidence for both seagrass and coral habitats on the reef flat. The predominance of reef flat pavements remains steady, accounting for around 50 % of MNI for all layers. Despite these alterations, it is unlikely that these changes reflect human impacts to the marine environments adjacent to MLEb-1. G. muricinum is dominant in Layer IA, but there is no evidence to suggest a decline in predominant taxa from earlier layers. N. plicata, Conus spp., ranellids, and A. violascens continue to be highly ranked in Layer IA. There is little evidence from an analysis of habitats that these changes reflect degradation of the intertidal zone due to trophic cascade or other alterations to trophic networks. While an increase in macroalgal cover may signal loss of hardbottom habitats that have been associated with human foraging in other studies (Castilla and Duran 1985; Godoy and Moreno 1989;), at MLEb-1, these changes are concurrent with both an increase in predatory carnivores and proportional representation of coral habitats, opposing evidence for a phase shift. The increase in predatory carnivores, coral habitats, and seagrass habitats is concurrent with an increase in shell density (MNI/m3) and species richness. These changes may be more adequately explained by an increase in foraging activity at the site, possibly due to human population increase and the establishment of the large village site at MLEb-1. An increase in population may have correlated with an increase in foraging activity and the number of different habitats encountered or accessed by foragers practising a non-selective foraging strategy. While more data are needed to confirm this hypothesis for TP 17-20, at this stage there is little to no indication that human foraging for molluscs on Ebon Atoll resulted in negative impacts to the intertidal ecosystem that are visible in the archaeological record.

Weisler (2001b) and Thomas (2014) argued that atolls, with a high ratio reef to land area and relatively low population sizes, are unlikely to present evidence for long-term impacts from human foraging for molluscs. On Ebon Atoll, a total of ~22 km2 of reef area surrounds ~5.4 km2 of land area. The reefs of the Marshall Islands are considered to be ‘in excellent condition’, especially on atolls distant from the urban centres of Majuro and Kwajalein (Beger et al. 2008:388). Though there are many modern threats to RMI reefs, the fine-grained, broad, and non-selective prehistoric mollusc foraging likely practiced on Ebon Atoll does not appear to have generated long-term negative impacts. Giovas (2016) has recently argued that fine-grained, non-selective foraging for

176 molluscs on small islands is unlikely to result in human impacts to molluscs or the wider marine environment, primarily through a lack of intense disturbance to any particular functional group. Indeed, many cases where both modern and archaeological studies have demonstrated that human foraging for molluscs has resulted in negative impacts have been due to an intensive focus on a single taxon, or members of the same functional group (often, large predatory gastropods or efficient herbivorous regulators of algae) (Braje and Erlandson 2013; Castilla and Duran 1985; Erlandson et al. 2008; Jerardino and Navarro 2008; O'Dea et al. 2014). We presented archaeological evidence from Ebon Atoll to support these assertions in parallel with other studies from Grand Bay and Sabazan in the West Indies (Giovas 2016, Poteate et al. 2014), Fiji (Szabó 2009), and Atafu Atoll, Tokelau (Ono and Addison 2013). At Utrō k atoll in the north (Weisler 2001b), foragers targeted a wide array of taxa, with major differences in assemblage composition related to flexible foraging strategies that reflect the configuration of the local environment adjacent to archaeological sites, with no discernible human impacts. It seems likely that foraging for molluscs on Ebon Atoll provided a predictable resource to supplement the products of extensive horticultural and arboricultural activity, fishing, and capturing seabirds and crustaceans.

Conclusion A temporal and spatial analysis of the mollusc assemblage from three major village sites and one ephemeral campsite has revealed little evidence for human impacts to molluscs during prehistory on Ebon Atoll. While small islands, and especially atolls, have long been considered marginal, fragile environments difficult for sustained human occupation, we presented evidence for long-term continuity in foraging practices spanning two millennia from colonisation through to the historic period. We also highlight that in contrast to other island types in the Pacific region, the differences in mollusc assemblage composition between windward and leeward locations on atolls are subtle and that a gradient, rather than a dichotomy, exists between mollusc assemblages from windward and leeward habitats (Kirch 1982; Morrison and Hunt 2007; Wiens 1962). However, all mollusc assemblages from Ebon Atoll were consistently rich, even, and diverse, with the major source of variation between archaeological sites related to the configuration of adjacent marine habitats. Other authors working in Micronesia have noted the remarkable productivity of the nearshore habitats on atolls, positing that low population density and abundant marine resources mitigated long-term impacts to the reef (Thomas 2014; Weisler 2001b). These results are also consistent with current discourses on the role of small islands in the human story, with Fitzpatrick et al. (2016) among a number of researchers challenging traditional conceptions of small islands as remote, marginal, and isolated (Giovas 2016), a position developed early on by Weisler (1995, 1996, 1997)

177 for the geographically isolated but socially connected Pitcairn Group. Linguistic evidence clearly points to consistent and continuous contact between atolls across the Marshall Islands throughout prehistory (Rehg 1995). Ethnographic evidence and storage pits from archaeological sites demonstrate that Marshall Islanders developed effective methods for preserving terrestrial foods in preparation for times of scarcity (Horrocks and Weisler 2006; Pollock 1984; Spoehr et al. 1949; Weisler 2001b). Studies of marine resource extraction in the archipelago, including Weisler (2001b) and Lambrides and Weisler (in press), have not identified evidence for human impacts to finfish resources in archaeological contexts, despite rich and dense fishbone assemblages. This analysis demonstrates that foraging on Ebon Atoll focussed on a broad range of taxa from different habitats, with most foraging occurring on reef flat pavements on leeward islets, with windward islets reporting higher abundances of taxa from rocks and boulders. These patterns reflect a foraging strategy that was adapted to local environmental conditions, with molluscs providing sustained returns throughout prehistory. Our evidence challenges previous conceptions of atolls as environmentally marginal for human habitation and document that human populations had no discernible negative impacts on molluscs for at least two millennia on Ebon Atoll.

Acknowledgements Permission to conduct archaeological research in the Republic of the Marshall Islands was granted to Weisler by the Historic Preservation Office (HPO), Ministry of Internal Affairs, and on Ebon Atoll, former mayor Lajan Kabua. Marshall Islands fieldwork was supported by a grant to Weisler from the Office of the Deputy Vice Chancellor (Research), University of Queensland. An Australian Institute of Nuclear Science and Engineering grant (ALNGRA 12004) to Weisler funded the radiocarbon age determinations. Harris’s university studies are supported by an Australian Research Training Scheme award. Molluscs collected during fieldwork have been returned to the HPO, Marshall Islands. We thank the two anonymous reviewers for their comments on the manuscript.

178

References Cited Abo, T., B.W. Bender, A. Capelle and T. DeBrum. 1976. Marshallese-English Dictionary. Honolulu: University of Hawai'i Press. Allen, M.S. 2003. Human impact on Pacific nearshore marine ecosystems. In Pacific Archaeology: Assessments and Prospects. Proceedings of the International Conference for the 50th anniversary of the first Lapita excavation (C. Sand, ed.):317-325. Nouméa: Les cahiers de l'archéologie en Nouvelle-Calédonie: Département Archéologie, Service des Musées et du Patrimoine. ---. 2012. Molluscan foraging efficiency and patterns of mobility amongst foraging agriculturalists: a case study from northern New Zealand. Journal of Archaeological Science 39(2):295-307. Alleway, H.K. and S.D. Connell. 2015. Loss of an ecological baseline through the eradication of oyster reefs from coastal ecosystems and human memory. Conservation Biology 29(3):795- 804. Beger, M., D. Jacobson, S. Pinca, Z. Richards, D. Hess, F. Harriss, C. Page, E. Peterson and N. Baker. 2008. The state of coral reef ecosystems of the Republic of the Marshall Islands. In The State of Coral Reef Ecosystems of the US and Pacific Freely Associated States (W. JE and A. Clarke, eds):387-417. Silver Spring: NOAA. Bender, B.W. 1969. Spoken Marshallese: An Intensive Language Course with Grammatical Notes and Glossary. Honolulu: University of Hawaii Press. Braje, T.J., J.M. Erlandson, D.J. Kennett, T.C. Rick and J.E. Peterson. 2005. Deep history: Using archaeology and historical ecology to promote marine conservation. Alternate Routes 21:5- 17. Brander, K., L. Botsford, L. Ciannelli, M. Fogarty, M. Heath, B. Planque, L. Shannon and K. Wieland. 2010. Human impacts on marine ecosystems. In Marine Ecosystems and Global Change (M. Barange, ed.):41-71. Oxford: Oxford University Press. Carpenter, K.E. and V.H. Niem (eds) 1998.The Living Marine Resources of the Western Central Pacific: Volume 1. Seaweeds, Corals, Bivalves and Gastropods. Rome: Food and Agriculture Organization of the United Nations. Castilla, J.C. and L. Duran. 1985. Human exclusion from the rocky intertidal zone of central Chile: the effects on Concholepas concholepas (Gastropoda). Oikos 45(3):391-399. Cernohorsky, W.O. 1967. Marine Shells of the Pacific. Sydney: Pacific Publications. Chelazzi, G. and M. Vannini. 1980. Zonal orientation based on local visual cues in Nerita plicata L. (Mollusca: Gastropoda) at Aldabra Atoll. Journal of Experimental Marine Biology and Ecology 46(2):147-156.

179

Cinner, J. 2014. Coral reef livelihoods. Current Opinion in Environmental Sustainability 7(0):65- 71. Crandall, E.D., M.A. Frey, R.K. Grosberg and P.H. Barber. 2008. Contrasting demographic history and phylogeographical patterns in two Indo-Pacific gastropods. Molecular Ecology 17(2):611-626. de Boer, W.F. and H.H.T. Prins. 2002. Human exploitation and benthic community structure on a tropical intertidal flat. Journal of Sea Research 48(3):225-240. Demond, J. 1957. Micronesian reef-associated gastropods. Pacific Science 11(3):275-341. Driver, J.C. 2011. Identification, classification and zooarchaeology. Ethnobiology Letters 2:19-39. Dye, T. 1987a. Marshall Islands Archaeology. Honolulu: Bernice Pauahi Bishop Museum. ---. 1987b. Report 3: Archaeological survey and test excavations on Arno Atoll, Marshall Islands. In Marshall Islands Archaeology (T. Dye, ed.):271-394. Honolulu: Pacific Anthroplogical Records: Bernice Pauahi Bishop Museum. Erlandson, J.M., T.J. Braje, T.C. Rick, N.P. Jew, D.J. Kennett, N. Dwyer, A.F. Ainis, R.L. Vellanoweth and J. Watts. 2011. 10,000 years of human predation and size changes in the owl limpet (Lottia gigantea) on San Miguel Island, California. Journal of Archaeological Science 38(5):1127-1134. Erlandson, J.M., T.C. Rick, T.J. Braje, A. Steinberg and R.L. Vellanoweth. 2008. Human impacts on ancient shellfish: a 10,000 year record from San Miguel Island, California. Journal of Archaeological Science 35(8):2144-2152. Faith, J.T. 2013. Taphonomic and paleoecological change in the large mammal sequence from Boomplaas Cave, western Cape, South Africa. Journal of Human Evolution 65(6):715-730. Faulkner, P. 2011. Quantifying shell weight loss in archaeological deposits. Archaeology in Oceania 46(3):118-129. Fitzpatrick, S.M., V.D. Thompson, A.S. Poteate, M.F. Napolitano and J.M. Erlandson. 2016. Marginalization of the margins: The importance of smaller islands in human prehistory. The Journal of Island and Coastal Archaeology 11(2):155-170. Fosberg, F.R. 1954. Soils of the northern Marshall atolls, with special reference to the Jemo series. Soil Science 78(2):99-108. Giovas, C.M. 2009. The shell game: analytic problems in archaeological mollusc quantification. Journal of Archaeological Science 36(7):1557-1564. ---. 2016. Though she be but little: Resource resilience, Amerindian foraging, and long-term adaptive strategies in the Grenadines, West Indies. The Journal of Island and Coastal Archaeology 11(2):238-263.

180

Giovas, C.M., M. Clark, S.M. Fitzpatrick and J. Stone. 2013. Intensifying collection and size increase of the tessellated nerite snail (Nerita tessellata) at the Coconut Walk site, Nevis, northern Lesser Antilles, AD 890–1440. Journal of Archaeological Science 40(11):4024- 4038. Giovas, C.M., S.M. Fitzpatrick, M. Clark and M. Abed. 2010. Evidence for size increase in an exploited mollusc: humped conch (Strombus gibberulus) at Chelechol ra Orrak, Palau from ca. 3000-0 BP. Journal of Archaeological Science 37(11):2788-2798. Godoy, C. and C.A. Moreno. 1989. Indirect effects of human exclusion from the rocky intertidal in Southern Chile: A case of cross-linkage between herbivores. Oikos 54(1):101-106. Gratwicke, B. and M. Speight. 2005. The relationship between fish species richness, abundance and habitat complexity in a range of shallow tropical marine habitats. Journal of Fish Biology 66(3):650-667. Hammer, Ø. 2001. PAST PAleontological STatistics. Oslo, Norway: Natural History Museum, University of Oslo. 3.04. Harris, M., A.B.J. Lambrides and M.I. Weisler. 2016. Windward vs. leeward: Inter-site variation in marine resource exploitation on Ebon Atoll, Republic of the Marshall Islands. Journal of Archaeological Science: Reports 6:221-229. Harris, M. and M.I. Weisler. 2016. Intertidal foraging on atolls: Prehistoric forager decision making at Ebon Atoll, Marshall Islands. The Journal of Island and Coastal Archaeology 12(2):200- 223. Harris, M. and M.I. Weisler. in press. Prehistoric human impacts to marine molluscs and intertidal ecosystems in the Pacific Islands. The Journal of Island and Coastal Archaeology. Harris, M., M.I. Weisler and P. Faulkner. 2015. A refined protocol for calculating MNI in archaeological molluscan shell assemblages: A Marshall Islands case study. Journal of Archaeological Science 57:168-179. Hayashida, F.M. 2005. Archaeology, ecological history, and conservation. Annual Review of Anthropology 34:43-65. Horrocks, M. and M.I. Weisler. 2006. Analysis of plant microfossils in archaeological deposits from two remote archipelagos: The Marshall Islands, eastern Micronesia, and the Pitcairn Group, southeast Polynesia. Pacific Science 60(2):261-280. Houk, P. and C. Musburger. 2013. Trophic interactions and ecological stability across coral reefs in the Marshall Islands. Marine Ecology Progress Series 488:23-34. Hughes, T.P., A.H. Baird, D.R. Bellwood, M. Card, S.R. Connolly, C. Folke, R. Grosberg, O. Hoegh-Guldberg, J.B.C. Jackson, J. Kleypas, J.M. Lough, P. Marshall, M. Nyström, S.R.

181

Palumbi, J.M. Pandolfi, B. Rosen and J. Roughgarden. 2003. Climate change, human impacts, and the resilience of coral reefs. Science 301(5635):929-933. Kayanne, H., T. Yasukochi, T. Yamaguchi, H. Yamano and M. Yoneda. 2011. Rapid settlement of Majuro Atoll, central Pacific, following its emergence at 2000 years CalBP. Geophysical Research Letters 38(20):L20405. Kendall, M.S., T.A. Battista and C. Menza 2012 Majuro Atoll, Republic of the Marshall Islands Coral Reef Ecosystems Mapping Report, NOAA National Centers for Coastal Ocean Science, Center for Coastal Monitoring, Biogeography Branch, Silver Spring, Maryland. Kendall, M.S. and T. Miller. 2008. The influence of thematic and spatial resolution on maps of a coral reef ecosystem. Marine Geodesy 31(2):75-102. Kendall, M.S., M.E. Monaco, K.R. Buja, J.D. Christensen, C.R. Kruer, M. Finkbeiner and R.A. Warner. 2001. Methods Used to Map the Benthic Habitats of Puerto Rico and the U.S. Virgin Islands, NOAA Technical Memorandum Maryland: NOAA National Centers for Coastal Ocean Science, Center for Coastal Monitoring, Biogeography Branch. Kirch, P. 1982. The ecology of marine exploitation in prehistoric Hawaii. Human Ecology 10(4):455-476. Kohn, A.J. 1980. Populations of tropical intertidal gastropods before and after a typhoon. Micronesica 16(2):215-228. Kohn, A.J. and P.J. Leviten. 1976. Effect of habitat complexity on population density and species richness in tropical intertidal predatory gastropod assemblages. Oecologia 25(3):199-210. Lambrides, A.B.J. and M.I. Weisler. in press. Alterations in the late Holocene Marshall Islands archaeological tuna fishery provide proxy evidence for western and central Pacific Ocean ENSO variability. The Journal of Island and Coastal Archaeology. Legendre, P. and E. Gallagher. 2001. Ecologically meaningful transformations for ordination of species data. Oecologia 129(2):271-280. Lyman, R.L. 2008. Quantitative Paleozoology. Cambridge: Cambridge University Press. Magurran, A.E. 2004. Measuring Biological Diversity. Malden: Oxford; Blackwell. Mason, L. 1950. Notes on Marshallese Culture in Ebon Atoll, Unpublished Interview with Dwight Heine. . Honolulu. Merlin, M., A. Capelle, T. Keene, J. Juvik and J. Maragos. 1994. Plants and Environments of the Marshall Islands. Honolulu: East West Center. Morrison, A.E. and T.L. Hunt. 2007. Human Impacts on the nearshore environment: An archaeological case study from Kaua'i, Hawaiian Islands. Pacific Science 61(3):325- 328,331-345.

182

O'Dea, A., M.L. Shaffer, D.R. Doughty, T.A. Wake and F.A. Rodriguez. 2014. Evidence of size- selective evolution in the fighting conch from prehistoric subsistence harvesting. Proceedings of the Royal Society of London B: Biological Sciences 281(1782):1-9. Ono, R. and D. Addison. 2013. Historical ecology and 600 years of fish use on Atafu Atoll, Tokelau. In Terra Australis 39: Prehistoric Marine Resource Use in the Indo-Pacific Regions (R. Ono, A. Morrison and D. Addison, eds):59-83. Canberra: ANU E Press Pollock, N.J. 1984. Breadfruit fermentation practices in oceania. Journal de la Société des Océanistes 40(79):151-164. Poppe, G.T. 2008. Philippine Marine Mollusks. Hackenheim: ConchBooks. Poteate, A.S., S.M. Fitzpatrick, M. Clark and J.H. Stone. 2015. Intensified mollusk exploitation on Nevis (West Indies) reveals~ six centuries of sustainable exploitation. Archaeological and Anthropological Sciences 7(3):361-374. Presley, T.K. 2005. Effects of the 1998 Drought on the Freshwater Lens in the Laura Area, Majuro Atoll, Republic of the Marshall Islands. Denver: U.S. Geological Survey. Rehg, K.L. 1995. The significance of linguistic interaction spheres in reconstructing Micronesian prehistory. Oceanic Linguistics 34(2):305-326. Reimer, P.J., M.G. Baillie, E.Bard, A. Bayliss, J.W. Beck, P.G. Blackwell, C.B. Ramsey, C.E. Buck, G.S. Burr and R.L. Edwards. 2009. IntCal09 and Marine09 radiocarbon age calibration curves, 0–50,000 years cal BP. Radiocarbon 51(4):1111-1150. Richmond, R., R. Kelty, P. Craig, C. Emaurois, A. Green, C. Birkeland, G. Davis, A. Edward, Y. Golbuu, J. Gutierrez, P. Houk, N. Idechong, J. Maragos, G. Paulay, J. Starmer, A. Tafileichig, M. Trianni and N.V. Velde. 2000. Status of coral reefs of American Samoa and Micronesia: US-affiliated freely associated islands of the Pacific. In Status of Coral Reefs of the World: 2000 (C. Wilkerson, ed.):217-236. Canberra: Australian Institute of Marine Science. Rick, T.C. and R. Lockwood. 2013. Integrating paleobiology, archeology, and history to inform biological conservation. Conservation Biology 27(1):45-54. Riley, T.J. 1987. Report 2: Archaeological survey and testing, Majuro Atoll, Marshall Islands. In Marshall Islands Archaeology (T. Dye, ed.):169-270. Honolulu: Pacific Anthroplogical Records: Bernice Pauahi Bishop Museum. Spoehr, A., P.S. Martin and L.A. Ross. 1949. Majuro, a village in the Marshall Islands. Fieldiana Anthropology 39:1-262. Sudo, K. 1984. Social organization and types of sea tenure in Micronesia. In Maritime Institutions in the Western Pacific (K. Ruddle and T. Akimichi, eds):203-230.

183

Szabó, K. and A. Anderson. 2012. The Tangarutu invertebrate fauna. In Terra Australis 37: Taking the High Ground: The Archaeology of Rapa, a Fortified Island in Remote East Polynesia (A. Anderson and D.J. Kennett, eds):135-144. Canberra: ANU E Press. Thomas, F.R. 2014. Shellfish gathering and conservation on low coral islands: Kiribati perspectives. The Journal of Island and Coastal Archaeology 9(2):203-218. Todd, J.A. 2001. Neogene Marine Biota of Tropical America Molluscan Life Habits Database London: Smithsonian Tropical Research Institute Panama, Natural History Museum Basel Switzerland,. Weisler, M.I. 1995. Henderson Island prehistory: colonization and extinction on a remote Polynesian island. Biological Journal of the Linnean Society 56(1-2):377-404. ---. 1996. An archaeological survey of Mangareva: Implications for regional settlement models and interaction studies. Man and Culture in Oceania 12:61-85. ---. 1997. Prehistoric long-distance interaction at the margins of Oceania. In Prehistoric Long- Distance Interaction in Oceania: An Interdisciplinary Approach (M.I. Weisler, ed.):149- 172. Dunedin: New Zealand Archaeological Association. ---. 1999a. The antiquity of aroid pit agriculture and signiﬁcance of buried A horizons on Paciﬁc Atolls. Geoarchaeology: An International Journal 14(7):621-654. ---. 1999b. Atolls as settlement landscapes: Ujae, Marshall Islands. Atoll Research Bulletin 460:1- 51. ---. 2001a. Life on the edge: Prehistoric settlement and economy on Utrōk Atoll, northern Marshall Islands. Archaeology in Oceania 36(3):109-133. ---. 2001b. On the Margins of Sustainability: Prehistoric Settlement of Utrōk Atoll, Northern Marshall Islands. Oxford: Archaeopress. ---. 2002. Archaeological Survey and Test Excavations on Ebon Atoll, Republic of the Marshall Islands. Majuro Historic Preservation Office, Republic of the Marshall Islands. Weisler, M.I., H. Yamano and Q. Hua. 2012. A multidisciplinary approach for dating human colonization of Pacific atolls. The Journal of Island and Coastal Archaeology 7(1):102-125. Wiens, H.J. 1962. Atoll Environment and Ecology. New Haven: Yale University Press. Woo, K., P. Faulkner and A. Ross. 2015. The effects of sampling on the analysis of archaeological molluscan remains: A quantitative approach. Journal of Archaeological Science: Reports 7:730-740. WoRMS Editorial Board 2015 World Register of Marine Species. Retrieved from .

184

Figures and Tables Table 1 Summary data for mollusc assemblages from MLEb-1, MLEb-5, MLEb-31, and MLEb-33; Wt. = weight MLEb-1 and MLEb-5, Ebon Islet Unit Layer MNI NISP Wt. (g) MNI/m3 NTAXA H′ E 1-D Fisher's α 6 IA 115 347 831.5 287.50 32 2.998 0.865 0.929 14.69 IB 59 215 207.78 306.49 27 3.036 0.921 0.939 19.26 IIA 41 115 126.32 230.99 7 1.671 0.859 0.766 3.677

17-20 IA 804 2330 3923.83 525.49 43 2.609 0.694 0.873 9.71 IB 247 735 1014.62 251.40 32 2.877 0.83 0.923 9.797 IIA 118 277 315.54 232.51 24 2.408 0.758 0.836 9.104 IIB 87 223 442.68 122.97 27 2.528 0.767 0.832 13.41

8-12 A 47 141 305.75 61.84 23 2.921 0.932 0.934 17.8

13-16 A 85 174 497.22 244.60 32 3.187 0.920 0.948 18.66 13-16 I 45 83 233.74 230.77 21 2.767 0.909 0.92 15.33

3 x 5 I 988 4401 19469.7 119.83 45 2.509 0.659 0.843 9.716

MLEb-31, Moniak Islet Unit Layer MNI NISP Wt. (g) MNI/m3 NTAXA H′ E 1-D Fisher's α 2-6 IA 171 494 2973.21 2988.12 23 2.38 0.759 0.864 7.154 IB 261 742 1535.38 1553.41 35 2.558 0.719 0.819 10.87 SD 32 66 281.35 281.87 13 2.157 0.841 0.83 8.155 IC 18 29 63.78 63.78 13 2.447 0.954 0.901 21 IIA 176 449 1666.62 1673.26 26 2.449 0.752 0.841 8.425 IIB 42 130 404.27 407.4 20 2.676 0.893 0.905 14.96

MLEb-33, Enekoion Islet Unit Layer MNI NISP Wt. (g) MNI/m3 NTAXA H′ E 1-D Fisher's α 1, 6, 7 I 559 1392 2685.71 2699.21 35 2.589 0.728 0.869 8.28

2 IA 96 243 1007.96 1012.01 17 2.272 0.802 0.854 6 IB 39 81 393.74 398.52 16 2.287 0.825 0.847 10.14

8 I 96 315 1076.91 1093.93 27 2.793 0.847 0.905 12.49

3 I 331 1807 4388.48 4398.68 25 2.433 0.756 0.876 6.274

185

Figure 1 Map of the Republic of the Marshall Islands, with Ebon Atoll and the location of sites MLEb-1, MLEb-31 and MLEb-33

186

Figure 2 Representative mollusc taxa from Ebon Atoll archaeological deposits

187

188

Figure 4 Summary of analysis for MLEb- 1 TP6 showing a. evenness (E), dominance (1-D) and diversity (H’) b. relative abundance of top three ranked taxa overall c. predominant habitat categories, and d. feeding types.

189

190

191

192

Chapter 7: Conclusion

Introduction This chapter presents a discussion of the main research findings of this thesis. Each research question (RQ1 – RQ5) is addressed individually. Future research objectives, and concluding remarks are also provided.

Overview of thesis results RQ1: Has human foraging for molluscs impacted mollusc populations and intertidal ecosystems in the Pacific Islands during the prehistoric period?

A number of neo-ecological and archaeological studies have demonstrated the potential for human foraging for molluscs to have strong direct (reductions in abundance and biomass of targeted species, reduction in body-size over time) and indirect (alterations to trophic networks, interspecific competition, community structure) effects. These impacts are routinely investigated by archaeologists working in the Pacific Islands, and outcomes ranging from changes in shell size, trophic alteration and possibly extirpation of particular taxa have been linked to human foraging.

This thesis presented three key findings from a literature review of archaeomalacological studies in the Pacific Islands.

First, previous studies commonly linked changes in mollusc assemblages to human foraging, primarily via a decrease in mean shell size or a decline in the abundance of large-bodied taxa, without fully exploring other explanations. Human foraging can result in a decline in shell size, but there is rarely a clear link between the two, even in controlled studies (e.g. McShane et al. 1994). These interpretations are less frequently presented in current literature, with researchers routinely considering the ecology of the intertidal zone and incorporating multi-proxy datasets including climate, sea-level, geological and other faunal data into analyses. Historical ecological analyses of archaeological data sets have demonstrated the complex link between human foraging and shell size or the abundance of particular taxa in the intertidal (Faulkner 2009; Giovas 2016; Giovas et al. 2013; Giovas et al. 2010; Thomas 2014) within and outside the Pacific Islands.

Second, large-bodied gastropod taxa have received disproportionate representation in archaeomalacological analyses due to a lack of fine-mesh screening or exclusion of small taxa from

193 identification or analysis. However, recovery methods that include < 3.2 mm screening or attempt identification on the total assemblage have demonstrated, for example, the importance of small- bodied taxa for environmental reconstruction or as resulting from non-selective foraging strategies. In addition, researchers in the Pacific Islands are more commonly presenting clear, detailed identification and quantification protocols. The continued reporting of these methods will facilitate comparison of archaeological datasets across the region to understand long-term human interaction with the marine environment, and possible impacts to intertidal ecosystems.

Third, in concert with other researchers (Allen 2003; Aswani and Allen 2009; Jones 2009; Jones and Quinn 2009), this thesis stressed the importance of analysing impacts to individual faunal classes as well as wider ecological impacts of marine subsistence on intertidal ecosystems in the Pacific Islands. Analyses aimed at generating these data will allow broad assessment of the role of long-term human interaction with the marine environment in structuring intertidal ecosystems in the region, and the resilience of coral reefs to future impacts including climate change. These studies would ideally include collaborative research with other disciplines including ecology, marine biology, and geology.

RQ2: Does the inclusion of an increased number of non-repetitive shell elements (NREs) in quantification protocols influence measures of relative abundance and taxonomic heterogeneity?

Ebon Atoll assemblages were used to test a new quantification protocol for Indo-Pacific mollusc taxa that incorporates an increased number of non-repetitive shell elements (NREs). Commonly utilised quantification protocols in archaeomalacology count only a select number of NREs. For example, for gastropods, the spire is most common, and the umbo and/or hinge for bivalves (e.g. Allen 2012; Claassen 1998; Mannino and Thomas 2001). This thesis presented the hypothesis that commonly utilised quantification protocols may systematically over-represent taxa with well- preserved spires, umbones or hinges. Measures of taxonomic heterogeneity and assemblage characteristics (richness, evenness, and diversity) derived from these quantification protocols would also be influenced by these biases, and as such these quantification methods may be altering inferences of past human behaviour relating to the collection of molluscs in the Pacific.

In order to test this hypothesis, two mollusc assemblages from Ebon Atoll, Marshall Islands were quantified using common quantification methods (referred to as NRE MNI), and a new quantification protocol, referred to as tMNI. The new MNI method, tMNI, uses an expanded 194 number of NREs for the quantification of mollusc shells, records element frequency per taxon, and MNI is calculated based on the most frequently occurring NRE for each taxon per analytical unit, rather than the total number of a single or restricted range of NRE. A formal protocol for this quantification method was outlined, to assure transparency, clarity and replicability of results (Driver 2011; Wolverton 2013; Woo et al. 2015). Each quantification protocol was assessed by comparing MNI values calculated using the NRE MNI protocol with the new, tMNI protocol. The influence on rank-order abundance, and derived measures of taxonomic heterogeneity including richness (NTAXA), and diversity and evenness (Simpson’s index of diversity, and the Shannon- Weiner indices of diversity and evenness). The influence of differential fragmentation was assessed by comparing a ratio of NISP to MNI for each protocol.

This comparison yielded several key findings related to shell form and the influence of quantification protocols. Principally, common methods of MNI quantification (NRE MNI) using a restricted range of elements systematically underrepresent gastropod abundance for Indo-Pacific taxa. As much as a two-fold increase in gastropod MNI was noted for the same assemblage quantified using the tMNI protocol. Turrifom taxa with high, dense spires such as Cerithium nodulosum were over-represented using NRE MNI, while globoid-form taxa with weak spires, such as Nerita spp. were consistently under-represented. The tMNI method did not produce lowered MNI values for these turriform taxa, but instead more accurately represented the relative abundance of these taxa to the total assemblage. Conversely, tMNI protocols resulted in only modest increases in bivalve MNI, however, the inclusion of additional NRE did increase MNI and count individuals that would not have been represented by NRE MNI protocols.

Divergent results were also reported for measures of taxonomic richness and heterogeneity for each quantification method. In all cases, expanding the number of NRE included for quantification resulted in an increase in richness as measured by NTAXA. As with the results for abundance, richness was increased due to the inclusion of taxa with spires that do not preserve well in archaeological deposits. Measures of evenness and diversity were also influenced by quantification protocol. Assemblage evenness was also impacted, with significant difference for Simpson’s index of diversity results for TP18 and 19 at MLEb-1 due to the over-representation of the turriform Cerithiidae when the NRE MNI protocol was used. While no other significant differences were reported, the variation in results for derived measures of assemblage composition highlight the potential influence of quantification protocol on inferences of human foraging behaviour.

195

Finally, the tMNI method allows a more comprehensive comparison of archaeomalacological data with other faunal datasets. Because both tMNI and common methods for the analysis of vertebrate fauna use the most commonly occurring NRE as the basis for calculating MNI, raw abundance data can be compared. For many coastal sites, mollusc shells and fishbones are ubiquitous components of the assemblage, and the major classes of subsistence remains.

This comparative study demonstrated that the inclusion of an increased number of non-repetitive shell elements (NREs) in quantification protocols has a marked impact on measures of relative abundance, and species richness, diversity, and evenness. Quantification protocols for mollusc shells that incorporate a wide range of NREs were demonstrated to provide a more valid measurement of relative abundance, avoiding biasing any type of shell architecture over another. Furthermore, these methods can be adapted for use in a wide range of regions and archaeomalacological contexts, allow comparison with other faunal data, and establish a clear, replicable quantification protocol for use in future research.

RQ3: What methods can be generated to explore and understand forager decision-making and habitat selection in atoll environments? Patterns of human foraging for molluscs, and the decisions made by foragers to collect from particular areas of the intertidal zone, are mediated by a range of physical (e.g., tidal movements, reef structure, ease of access, substrate type) and sociocultural (e.g., individual preferences, cultural taboos, gender roles, age) factors. The influence of these factors on the accumulation of mollusc assemblages in archaeological sites has traditionally been explored in archaeology through optimal foraging theory, or by reconstructing foraging preferences by assigning mollusc taxa to single, relatively broad habitat types (Bedford 2007; Morrison and Addison 2008; Morrison and Hunt 2007), or a range of detailed classes (Allen 2012; Clark et al. 2001; Szabó 2009). These methods of reconstructing habitat selection are referred to by Allen (1992:331) as defining probabilistic relationships between foraging patterns and the taxa in the assemblage. These methods rely on linking knowledge of the ecology of the intertidal zone and the molluscs that are hosted there with mollusc taxa recovered from archaeological sites. However, by assigning each taxon to only a single habitat, the complex relationship between molluscs and the complex variation in water depth, tidal exposures, substrate type, and associated biological cover in the intertidal zone is highly simplified.

196

Hierarchical classification schemes are a robust framework for assigning mollusc taxa to habitat categories to assess forager decision making. This method for classifying the marine environment is commonly used to map the distribution of benthic habitats and assign a combination of location in relation to the shoreline, geomorphological structure, and benthic cover to define a habitat class. However, this scheme can be adapted to assign mollusc taxa to areas of the intertidal zone where they would most likely be encountered by foragers. Using a classification scheme developed by Kendall et al. (2012) for Majuro Atoll, all mollusc taxa in the assemblage were assigned to one or more habitats based on a literature review of Indo-Pacific mollusc ecology. With each taxon assigned to an area of the intertidal zone, forager decision making as it relates to both site location (windward v. leeward) and site function (village v. campsites) was investigated for two archaeological sites on Ebon Atoll.

An analysis of the range of habitats represented at each site demonstrated that a broad-based, generalist foraging strategy was employed during the prehistoric period. All assemblages were rich, and even, with many habitat classes represented. Gastropods from the high intertidal and reef flat were the most common component of both windward and leeward assemblages, with differences in the taxa exploited likely reflecting local habitat variation. Though there were some taxa that were more common than others, foraging seems to have targeted habitats, and the predictable aggregations of taxa that inhabit them, rather than any specific focus on a single species. Only minor differences were noted between assemblages from the major village site on Ebon Islet and the ephemeral campsite on Moniak Islet. The source of this variation likely relates to be the configuration of intertidal habitats adjacent to each site, as the more exposed Moniak Islet featured greater numbers of taxa adapted to strong waves and currents (e.g. Vasum turbinellus), while Ebon assemblages featured greater numbers of taxa that inhabit the high-rugosity reefs of the leeward lagoon. These findings relate to RQ4, and are discussed in more detail below, however, this thesis has presented a new method for detecting these differences in mollusc assemblages from Indo- Pacific archaeological sites. This thesis also includes a clear and transparent description of the method that can be adapted to investigate other archaeological sites within and beyond the Indo- Pacific for comparative studies. The foraging preferences of human foragers for molluscs can yield data regarding long-term interactions of humans with coastal environments, and how these may have changed through time. These findings can be related to other archaeological datasets, and compared with ancillary datasets including climate records, geological datasets, and modern studies of forager behaviour.

197

RQ4: What is the influence of spatial (settlement patterns and local habitat) and temporal factors on the richness (number of species present), abundance (number of individuals of each species present), and diversity (richness and relative distribution of individuals of each species within a population) of mollusc assemblages on Ebon Atoll?

As stated above, a study of mollusc assemblages from the leeward Ebon Islet and windward Moniak Islet demonstrated that variation in assemblage composition most likely relates to the configuration of intertidal habitats adjacent to each site. Variation in the patterning of archaeological data depending on windward or leeward coastal exposure have been noted for other Pacific Islands. Subsistence practices differed between windward and leeward exposed sites in Hawaii (Bayman and Dye 2013 ; Earle 1977; Kirch and Dye 1979; Palmer et al. 2009; Weisler and Kirch 1985), and different reef configurations on other high islands (Kirch and Dye 1979; Morrison and Hunt 2007; Szabó 2009; Szabó and Anderson 2012). The configuration of intertidal habitats and mollusc- bearing environments on atolls are heavily influenced by the relative exposure to winds, waves, and currents, geological history and other stochastic factors.

In addition to these local ecological factors, archaeological and geographical investigations indicate that human settlement patterns of Marshall Islands atolls tend to be patterned in a predictable manner. Permanent villages that are the population centres of atolls tend to be located on islets with large land area and a well-developed Ghyben-Herzberg freshwater lens that are sheltered from waves, winds and currents. Islets that are exposed to strong waves, winds and currents on the windward side of atolls are generally smaller, with insufficient freshwater lenses to support permanent habitation. These windward islets are typically associated with ephemeral occupation sites, and were likely used during the prehistoric period for staging fishing sorties and collecting seabirds and molluscs. This thesis presented both a way to investigate these differences by presenting a new method for analysing human foraging preferences, and for the first time on atolls, an analysis of the role of local ecological variation in mollusc foraging. Additionally, this thesis examined the role of human settlement patterns on mollusc foraging, which has been explored primarily for northern atolls in the Marshall Islands, but not in a dedicated study.

An analysis of mollusc assemblages from four archaeological sites on Ebon Atoll including leeward, village sites (MLEb-1, MLEb5), a windward ephemeral occupation site (MLEb-31), and a moderately exposed village site (MLEb-33) has demonstrated three key findings (1) regardless of site location or site type, mollusc assemblages are rich, even and diverse (2) foragers seem to have

198 practiced a fine grained, non-selective foraging strategy; and (3) differences in assemblage composition seems to have been most strongly mediated by the configuration of intertidal habitats adjacent to archaeological sites, rather than site function. However, for leeward islets there was a tendency for increased proximity to the lagoon (and therefore to permanent village sites) to be associated with relatively low evenness scores, and higher abundance of nerites, ranellids and psammobids. Increased proximity to the oceanside was associated with an increased evenness and abundance of taxa that prefer the reef edge.

A comparative analysis of mollusc and fish bone assemblages from MLEb-1 and MLEb-31 demonstrated that unlike molluscs, taxonomic composition of fishbone assemblages tends to be more strongly structured by site type, rather than location. Dissimilarity scores for fishbone assemblages were generally lower than molluscs. Though requiring further analysis, this thesis proposed that while mollusc assemblages are strongly mediated by local environment, variation in fish bone likely relates to the range of fish capture strategies practiced at different site types.

Atolls have traditionally been considered especially difficult settings for long-term human habitation. A lack of standing fresh water, poor soils, and vulnerability to extreme weather have been cited as challenges to survival. However, archaeological research in the Marshall Islands has demonstrated that human populations have persisted since initial colonisation soon after atoll emergence. Marine resources are key to this survival, and the literature review presented in Chapter 2 demonstrated the potential for long-term exploitation of marine fauna to alter intertidal ecosystems. This thesis investigated 2000 years of foraging for molluscs on Ebon Atoll for the first time to both understand long-term human interaction with the marine environment, and also to investigate potential impacts. Two new methods, one for calculating MNI (Chapter 3), and one for tracking foraging preferences (Chapters 4 and 5) for investigating human foraging preferences, and modern benthic habitat maps were generated and incorporated into a temporo-spatial analysis of mollusc assemblages from four archaeological sites on Ebon Atoll.

The variability in mollusc assemblages relating to the windward-leeward exposure gradient on atolls has been discussed in this chapter, and extensively in Chapters 4 and 5. Temporal analysis of 199 mollusc assemblages demonstrated three key findings: (1) a suite of five taxa, (Nerita polita, N. plicata, Conus spp., Vasum turbinellus and A. violascens) are highly ranked at all sites through time, but all assemblages are rich, even and diverse; (2) molluscs formed a minor, stable component of the human diet, likely from initial colonisation to the historic period; and (3) there is little indication of negative impacts to molluscs or intertidal ecosystems as a result of human foraging for molluscs based on these analyses.

The five most common taxa in Ebon assemblages are typical Indo-Pacific fauna that are abundant in intertidal atoll environments, and reflect the non-selective foraging strategy employed throughout the sequence. Minor alterations to rank-order abundance was noted, but there was no indication of over-exploitation of any particular taxon at any analysed site. The only evidence for focused targeting of a single taxon was reported from early deposits at the village site at MLEb-5, Ebon Islet. This site had the highest relative and absolute abundance of Tridacna and may be associated with the production of shell tools. While other archaeological sites in the Marshall Islands have yielded evidence for focussed exploitation of a single taxon, this was not the case for any Ebon assemblage. In addition, densities of mollusc remains from Ebon Atoll assemblages were generally low, especially compared to other Micronesian assemblages (Giovas et al. 2010), indicating that molluscs were a minor and predictable component of the diet.

The intertidal zone of atolls is remarkably productive, and expansive compared to the small land area available for habitation. This marine productivity and generally low human populations have been suggested by other researchers to be key to the lack of observable long-term impacts to molluscs in these ecosystems (Thomas 2014; Weisler 2001). The broad-based, generalised foraging strategy practiced during the prehistoric period on Ebon Atoll also likely spread direct and indirect impacts across multiple trophic levels and functional groups. In archaeological and ecological studies that have demonstrated human impacts, foraging is usually centred on intensively harvesting a single, or narrow range of taxa. This thesis presents results to support those of other small-island researchers both within and outside the Pacific Islands (Giovas 2016; Thomas 2014; Weisler 2001) that assert that a lack of disturbance to any particular functional group, high marine productivity and low human population density did not result in observable impacts to molluscs. These results also support current discourses on small islands in the human story, challenging traditional notions of these landscapes as harsh, remote, marginal and isolated settings for human habitation. The foraging strategy inferred here provided sustained yields of molluscs for at least two millennia on Ebon Atoll, Marshall Islands.

200

Future research objectives This thesis has demonstrated the utility of employing a suite of transparent and robust quantification protocols to accurately and reliably measure the relative abundance of taxa in archaeomalacological assemblages. The advantages of utilising a wide range of NRE for quantification have been demonstrated, especially for gastropod taxa. Further studies on a larger, more diverse assemblage of bivalve taxa would elucidate the role of NRE selection on this class of molluscs more clearly. However, gastropod taxa were consistently underrepresented, and future research should endeavour to consider NRE selection, and the influence these analytical decisions can have on reconstructions of the past. Researchers should clearly and transparently report quantification protocols, and seek to build comparative datasets, especially for low coral atolls, where data is patchy.

Furthermore, the development of novel methods for understanding foraging decision-making presented here have highlighted the utility of probabilistic reconstructions of habitat selection. With this new method, this thesis presented for the first time a detailed analysis of foraging practices during the prehistoric period on Ebon Atoll. While no indications of human impact were observed here, the generation of multi-proxy data to assess the condition of atoll reefs throughout the sequence of human occupation would greatly enhance understanding of long-term human interaction with these ecosystems. At present, there is limited geomorphological data on atoll islet development for the atolls of the Marshall Islands, and other high-resolution proxy data for climate in the region is lacking. These datasets would be useful for a range of archaeological and historical ecological studies, and should be incorporated into future research objectives.

Concluding Remarks This thesis presented new quantification and analytical methods for understanding human foraging for molluscs in tropical Indo-Pacific settings, and reviewed archaeological evidence for human impacts to molluscs in the Pacific Islands. A new quantification protocol was presented that considered a wide range of NRE for the calculation of MNI, more accurately representing the relative abundance of taxa in mollusc assemblages from Ebon Atoll, Marshall Islands. The new MNI method presented here can be adapted for use with any assemblage, not only those from the Indo-Pacific region. This method allows a more reliable comparison of mollusc relative abundance with other faunal classes as both methods use the most frequently occurring NRE to calculate MNI. Researchers in other regions have highlighted the utility of this method for both archaeomalacological analyses, and broader analyses of multiple faunal classes (Jerardino et al.

201

2016; Morin et al. in press; Popejoy et al. in press; Popejoy et al. 2016). The method for reconstructing forager-descision presented in this thesis facilitates understanding of long-term patterning in foraging behaviour. These data are critical to further elucidation of the role of humans in structuring marine environments through time, and the impacts of foraging on intertidal ecosystems worldwide.

This thesis also undertook a high-resolution study of the archaeomalacological record of Ebon Atoll, demonstrating that molluscs had been a stable component of the diet for two millennia. Braje et al. (in press) recently emphasised the critical importance of the archaeology of islands for both understanding the role of humans in altering or adapting to these environments, but also as critical to enhancing future global sustainability. This thesis presents new methods and high-resolution data to contribute to these research aims. Mollusc assemblages from Ebon were rich, even and diverse, incorporating a broad range of taxa from different habitats. Variation in assemblage composition is likely related to the configuration of intertidal habitats on windward and leeward exposed islets, rather than site function. No discernible human impacts were noted, indicating that this generalised foraging strategy, low human populations and a productive marine environment produced sustained yields for molluscs by spreading impact across trophic levels and functional groups. These data contest traditional perceptions of atolls as marginal, and are in line with current discourses that challenge notions of small islands, and especially atolls as remote, isolated and marginal settings for human habitation.

202

References Cited Allen, M.S. 1992. Dynamic Landscapes and Human Subsistence: Archaeological Investigations on Aitutaki Islands, Southern Cook Islands. Ph.D. thesis. Washington: University of Washington. ---. 2003. Human impact on Pacific nearshore marine ecosystems. In Pacific Archaeology: Assessments and Prospects. Proceedings of the International Conference for the 50th anniversary of the first Lapita excavation (C. Sand, ed.):317-325. Nouméa: Les cahiers de l'archéologie en Nouvelle-Calédonie: Département Archéologie, Service des Musées et du Patrimoine. ---. 2012. Molluscan foraging efficiency and patterns of mobility amongst foraging agriculturalists: a case study from northern New Zealand. Journal of Archaeological Science 39(2):295-307. Aswani, S. and M.S. Allen. 2009. A Marquesan coral reef (French Polynesia) in historical context: an integrated socio-ecological approach. Aquatic Conservation: Marine and Freshwater Ecosystems 19(6):614-625. Bayman, J.M. and T.S. Dye. 2013. Hawaii's Past in a World of Pacific Islands. Washington D.C.: The SAA Press. Bedford, S. 2007. Terra Australis 23: Pieces of the Vanuatu Puzzle: Archaeology of the North, South and Centre. Canberra: ANU E Press. Braje, T.J., T.P. Leppard, S.M. Fitzpatrick and J.M. Erlandson. in press. Archaeology, historical ecology and anthropogenic island ecosystems. Environmental Conservation:1-12. Claassen, C. 1998. Shells. Cambridge: Cambridge University Press. Clark, G., A. Anderson and S. Matararaba. 2001. The Lapita Site at Votua, Northern Lau Islands, Fiji. Archaeology in Oceania 36(3):134-145. Driver, J.C. 2011. Identification, classification and zooarchaeology. Ethnobiology Letters 2:19-39. Earle, T.K. 1977. A reappraisal of redistribution: Complex Hawaiian chiefdoms. In Exchange systems in prehistory (T.K. Earle and J.E. Ericson, eds):29. New York: Academic Press. Faulkner, P. 2009. Focused, intense and long-term: evidence for granular ark (Anadara granosa) exploitation from late Holocene shell mounds of Blue Mud Bay, northern Australia. Journal of Archaeological Science 36(3):821-834. Giovas, C.M. 2016. Though she be but little: Resource resilience, Amerindian foraging, and long- term adaptive strategies in the Grenadines, West Indies. The Journal of Island and Coastal Archaeology 11(2):238-263. Giovas, C.M., M. Clark, S.M. Fitzpatrick and J. Stone. 2013. Intensifying collection and size increase of the tessellated nerite snail (Nerita tessellata) at the Coconut Walk site, Nevis,

203

northern Lesser Antilles, AD 890–1440. Journal of Archaeological Science 40(11):4024- 4038. Giovas, C.M., S.M. Fitzpatrick, M. Clark and M. Abed. 2010. Evidence for size increase in an exploited mollusc: humped conch (Strombus gibberulus) at Chelechol ra Orrak, Palau from ca. 3000-0 BP. Journal of Archaeological Science 37(11):2788-2798. Jerardino, A., P. Faulkner and C. Flores. 2016. Current methodological issues in archaeomalacological studies. Volume 427(Part A):1-4. Jones, S. 2009. A Long-term perspective on biodiversity and marine resource exploitation in Fiji's Lau Group. Pacific Science 63(4):617-648. Jones, S. and R.L. Quinn. 2009. Prehistoric Fijian diet and subsistence: Integration of faunal, ethnographic, and stable isotopic evidence from the Lau Island Group. Journal of Archaeological Science 36(12):2742-2754. Kendall, M.S., T.A. Battista and C. Menza 2012 Majuro Atoll, Republic of the Marshall Islands Coral Reef Ecosystems Mapping Report, NOAA National Centers for Coastal Ocean Science, Center for Coastal Monitoring, Biogeography Branch, Silver Spring, Maryland. Kirch, P.V. and T.S. Dye. 1979. Ethno-archaeology and the development of Polynesian fishing strategies. The Journal of the Polynesian Society 88(1):53-76. Mannino, M.A. and K.D. Thomas. 2001. Intensive Mesolithic exploitation of coastal resources? Evidence from a shell deposit on the isle of Portland (southern England) for the impact of human foraging on populations of intertidal rocky shore molluscs. Journal of Archaeological Science 28(10):1101-1114. McShane, P.E., D.R. Schiel, S.F. Mercer and T. Murray. 1994. Morphometric variation in Haliotis iris (Mollusca: Gastropoda): Analysis of 61 populations. New Zealand Journal of Marine and Freshwater Research 28(4):357-364. Morin, E., E. Ready, A. Boileau, C. Beauval and M.-P. Coumont. in press. Problems of identification and quantification in archaeozoological analysis, part II: Presentation of an alternative counting method. Journal of Archaeological Method and Theory. Morrison, A.E. and D.J. Addison. 2008. Assessing the role of climate change and human predation on marine resources at the Fatu-ma-Futi site, Tutuila Island, American Samoa: An agent based model. Archaeology in Oceania 42(4):22-34. Morrison, A.E. and T.L. Hunt. 2007. Human Impacts on the nearshore environment: An archaeological case study from Kaua'i, Hawaiian Islands. Pacific Science 61(3):325- 328,331-345.

204

Palmer, M.A., M. Graves, T.N. Ladefoged, O.A. Chadwick, T. Ka'eo Duarte, S. Porder and P.M. Vitousek. 2009. Sources of nutrients to windward agricultural systems in pre-contact Hawai'i. Ecological Applications 19(6):1444-1453. Popejoy, T., C.R. Randklev, S. Wolverton and L. Nagaoka. in press. Conservation implications of late Holocene freshwater mussel remains of the Leon River in central Texas. Hydrobiologia. Popejoy, T., S. Wolverton, L. Nagaoka and C.R. Randklev. 2016. An interpretive framework for assessing freshwater mussel taxonomic abundances in zooarchaeological faunas. Quaternary International Volume 427(Part A):36-46. Szabó, K. 2009. Molluscan remains from Fiji. In Terra Australis 31: The Early Prehistory of Fiji (G. Clark and A.J. Anderson, eds):183-212. Canberra: ANU E Press. Szabó, K. and A. Anderson. 2012. The Tangarutu invertebrate fauna. In Terra Australis 37: Taking the High Ground: The Archaeology of Rapa, a Fortified Island in Remote East Polynesia (A. Anderson and D.J. Kennett, eds):135-144. Canberra: ANU E Press. Thomas, F.R. 2014. Shellfish gathering and conservation on low coral islands: Kiribati perspectives. The Journal of Island and Coastal Archaeology 9(2):203-218. Weisler, M.I. 2001. On the Margins of Sustainability: Prehistoric Settlement of Utrōk Atoll, Northern Marshall Islands. Oxford: Archaeopress. Weisler, M.I. and P.V. Kirch. 1985. The structure of settlement space in a Polynesian chiefdom: Kawela, Molokai, Hawaiian Islands. New Zealand Journal of Archaeology 7:129-158. Wolverton, S. 2013. Data quality in zooarchaeological faunal identification. Journal of Archaeological Method and Theory 20(3):381-396. Woo, K., P. Faulkner and A. Ross. 2015. The effects of sampling on the analysis of archaeological molluscan remains: A quantitative approach. Journal of Archaeological Science: Reports 7:730-740.

205

Appendix A

Summary of extant marine habitats mapped within a 2km radius of archaeological sites on Ebon Islet (MLEb-1, MLEb-5) Enekoion Islet (MLEb-33) and Moniak Islet (MLEb-31)

206

2 Habitats km % MLEb MLEb MlEb MlEb MLEb MLEb MLEb MLEb

Code Name - - - - 33 33 31 31 - - - - 5 1 5 1

B Shoreline Intertidal B/1 Coral Reef and Hardbottom B/1/18 Rock/Boulder 0.016 0.006 0.023 0.052 0.80 0.35 0.79 2.41 B/2 Unconsolidated Substrate B/2/21 Sand 0.081 0.070 0.037 0.010 4.16 4.18 1.28 0.47 D Reef Flat D/1 Coral Reef and Hardbottom D/1/11 Aggregate Reef 0.106 0.133 0.336 5.48 7.90 11.68 0.00 D/1/12 Aggregated Patch Reef 0.092 0.125 0.122 4.75 7.41 4.25 0.00 D/1/13 Individual Patch Reef 0.007 0.007 0.34 0.39 2.23 0.00 D/1/15 Pavement 0.294 0.144 0.974 1.214 15.12 8.56 33.83 55.90 Pavement with Sand D/1/16 0.283 0.354 14.60 21.01 0.00 0.00 Channels D/1/17 Reef Rubble 0.122 0.034 0.511 6.30 2.00 17.74 0.00 D/1/18 Rock/Boulder 0.012 0.012 0.109 0.070 0.60 0.72 3.78 3.21 D/2 Unconsolidated Substrate D/2/21 Sand 0.078 0.030 0.011 0.007 4.04 1.78 0.38 0.31 Sand with Scattered Coral D/2/22 0.047 0.047 0.065 0.000 2.43 2.80 2.27 0.00 and Rock D/2/23 Seagrass 0.186 0.114 0.125 0.000 9.57 6.79 4.34 0.00 E Back Reef B/1 Coral Reef and Hardbottom E/1/11 Aggregate Reef 0.057 0.052 0.022 2.92 3.12 0.75 0.00 E/1/12 Aggregated Patch Reef 0.093 0.121 0.060 4.79 7.18 2.10 0.00 E/1/17 Reef Rubble 0.030 0.138 1.55 0.00 0.00 6.35 E/2 Unconsolidated Substrate E/2/21 Sand 0.000 0.000 0.000 0.167 0.00 0.00 0.00 7.70 Sand with Scattered Coral E/2/22 0.144 0.146 0.115 0.092 7.42 8.69 3.98 4.25 and Rock H Bank/Shelf Escarpment H/1 Coral Reef and Hardbottom H/1/14 Spur and Groove 0.211 0.192 0.144 0.238 10.85 11.38 5.01 10.95 H/1/19 Algal Ridge 0.083 0.097 0.159 0.184 4.28 5.74 5.51 8.45 L Unknown L/1 Coral Reef and Hardbottom L/1/18 Rock/Boulder 0.000 0.000 0.002 0.000 0.00 0.00 0.07 0.00

207

Appendix B

List of zones, major geomorphological structures and detailed geomorphological structures used in the Ebon archaeological project hierarchical classification scheme (after Kendall et al. 2012:8-12). Zone J, dredged/excavated and Detailed Geomorphological structure 13, aggregated patch reefs was not used for the analysis presented here, as these classes relate to methods for mapping modern day atoll benthic habitats. Detailed Geomorphological structure 19, Algal Ridge, was added by the authors due to the distinctive range of molluscan taxa associated with this habitat (Morrison 1954).

208

Zones code name description A Land Terrestrial features at or above the high tide line. B Shoreline Intertidal Area between the spring high tide line and lowest spring tide level C Lagoon Area of water inside the atoll, surrounded by the Back Reef D Reef Flat Shallow, low relief area exposed at low tide between the Shoreline Intertidal and Fore Reef or Back Reef E Back Reef Area on the lagoonside of an atoll sloping inward from the Shoreline Intertidal or Reef Flat down to the seaward edge of the Lagoon floor. F Fore Reef Area along the seaward (oceanside) edge of the reef flat that slopes into deeper water to the landward edge of the Bank/Shelf Escarpment G Bank/Shelf Deeper water extending offshore from the seaward edge of the Fore Reef to the beginning of the escarpment where the insular shelf drops off into deep, oceanic water H Bank/Shelf Escarpment Begins on the seaward edge of the Fore Reef, where depth increases rapidly into deep, oceanic water. I Channel Naturally occurring channels in the seafloor that often cut across several other zones. K Pinnacle High-relief features occurring in the Lagoon that are separated from the Back Reef by the deeper waters of the Lagoon. L Unknown Habitat proclivities could not be assessed.

Major geomorphological structures and detailed geomorphological structures 1 Coral Reef and Hard bottom Solid substrates, including bedrock, boulders and reef building organisms. A thin veneer of sediment may be present. 11 Aggregate Reef Continuous, high-relief coral formation of variable shapes, lacking sand channels of Spur and Groove formations. 13 Individual patch reef Coral formations that are isolated from other coral reef formations by bare sand, seagrass or other habitats. 14 Spur and Groove Alternating sand and coral formations that are oriented perpendicular to the Shoreline intertidal or Fore Reef. The coral formations (spurs) of this feature typically have a high vertical relief and are separated by 1 to 5m of sand or hard bottom (grooves). Occurs only in the Fore reef or Bank/Shelf Escarpment zone. 15 Pavement Flat, low-relief, solid rock in broad areas, often with partial coverage of sand, algae, hard coral, Alcyonacea (sea whips or fans), zoozanthids or other sessile invertebrates. 16 Pavement with Sand Areas of pavement with alternating sand/surge channel formations that are oriented perpendicular to the Channels Shoreline Intertidal or Bank/Shelf escarpment. 209

17 Reef Rubble Dead, unstable coral rubble often colonised with turf, filamentous, calcareous or encrusting macroalgae. Often occurs due to storm waves piling up dead coral. 18 Rock/Boulder Large, irregularly shaped carbonate blocks often extending from the island bedrock, indicating higher sea- levels, or aggregations of loose coral cobbles and boulders that have been detached and transported from their native beds. Individual cobbles and boulders often range in diameter from 0.25-3m 19 Algal Ridge Area of consolidated coral pavement colonised by calcareous algae occurring shoreward of the Bank/Shelf Escarpment or Fore Reef and demarcates the seaward margin of the Reef Flat. Often slightly higher elevation that the seaward and shoreward areas of the reef. 2 Unconsolidated Substrate Areas of the seafloor consisting of small, unattached or uncemeneted particles with less than 10% cover of large stable substrate. 21 Sand Areas of the seafloor consisting of small, unattached or unncemented particles. 22 Sand with Scattered Coral Primarily sand bottom with scattered rocks or small, isolated coral heads. and Rock 23 Seagrass Primarily sand bottom colonised by seagrass. 3 Other Delineations Any other type of structure not classified as Coral Reef and Hard bottom or Unconsolidated Substrate. 31 Land Terrestrial features beyond the Shoreline Intertidal. 4 Unknown Habitat proclivities could not be assessed. 41 Unknown Habitat proclivities could not be assessed.

210

Appendix C

Neogene Marine Biota of Tropical America Molluscan Life Habits Database (NmiTA) categories (after Todd 2001)

211

Table A Gastropod feeding type categories Code Feeding Category Description

Predators feeding on and killing whole sedentary and mobile macro-organisms and also selective ingesters of foraminifera (foraminiferivores). Included here are CP Predatory carnivores scavengers, which with just a few known exceptions, are also predators, shifting facultatively when carrion is present (Britton and Morton 1994)

Predators which feed on sedentary, and typically clonal, animals (e.g. corals and other cnidarians, sponges, ascidians) without killing CB Browsing carnivores them. This also includes those ‘parasites’, which are ectoparasitic upon mostly relatively larger sedentary or mobile prey.

Browsing macroherbivores with unselective HO Herbivorous omnivores omnivory, typically of epifauna attached to macroalgae

Microalgivores, detritivores, microphages and unselective deposit feeder. Also included here Herbivores on fine-grained HM is a miscellany of herbivorous Non HR and Substrates HP categories, including those living on wood or mangrove substrates.

Herbivores on rock, rubble or coral Microalgivores HR substrates

Micro and macroalgivores and detritivores on HP Herbivores on plant or algal substrates macroalgal and seagrass substrates.

Includes taxa feeding solely or dominantly SU Suspension feeders upon suspended particles, including mucociliary feeders.

212

Table B Bivalve organism/substrate relationship categories Code Organism/Substrate Relationship Notes

ER Epifaunal recliner

On a range of substrates; including sediment; consolidated substrates EP Epifaunal including biogenic substrates (e.g. coral), and macroalgal and seagrass substrates

SI Semi-infaunal

IS Infaunal siphonate

IA Infaunal asiphonate

WN Nestler on or within hard substrates Excluding active borers

Includes taxa feeding solely or dominantly WB Borer, nestling in hard substrate upon suspended particles, including mucociliary feeders.

Nestler within burrow of another organism WU in unconsolidated substrate

213

Table C Bivalve feeding type categories Code Feeding Type Description

SU Suspension feeder

DU Subsurface deposit feeder

Surface and subsurface deposit feeders food sources and strategies have been compared and contrasted by Jumars et al. (1990). Suspension feeders may ingest deposited material and surface deposit feeders may suck in material from the water column (Kamermans 1994). Despite this, the two groups reflect distinct feeding strategies with often very different food sources. There is growing evidence that some tellinid species, among surface deposit feeders, may DS Surface deposit feeder facultatively suspension feed. This swop between suspension and deposit feeding may occur as a response to food quality and quantity, hydrodynamics and predation pressure. Nevertheless, this ability may vary between congeners (Levinton 1991). To help resolve ecological patterns, for the present I have simply coded all tellinoids as surface deposit feeders except those taxa which have been examined and are only known to suspension feed.

DC Chemosymbiotic deposit feeder

CAR microcarnivore

214

Table D Bivalve mobility categories Code Feeding Category Notes

Includes cemented, boring, nestling and reclining taxa with no means of repositioning, apart from that which IM Immobile may result from growth

Sluggish forms which have at least some capacity to SE Sedentary reposition in response to disturbance

MA Actively mobile Including active crawlers and burrowers.

Those which have the ability to swim and which are SW Swimming believed to do so not solely as an escape response.

215

Table E Bivalve attachment type categories Code Attachment type

UN Unattached

BA Bysally attached

CE Cemented

216

References Cited Britton, J.C. and B. Morton. 1994. Marine carrion and scavengers. Oceanography and Marine Biology: an annual review 32. Jumars, P.A., L.M. Mayer, J.W. Deming, J.A. Baross and R.A. Wheatcroft. 1990. Deep-sea deposit- feeding strategies suggested by environmental and feeding constraints. Philosophical Transactions of the Royal Society of London A: Mathematical, Physical and Engineering Sciences 331(1616):85-101. Kamermans, P. 1994. Similarity in food source and timing of feeding in deposit-and suspension- feeding bivalves. Marine Ecology-Progress Series 104:63-63. Levinton, J. 1991. Variable feeding behavior in three species ofMacoma (Bivalvia: Tellinacea) as a response to water flow and sediment transport. Marine Biology 110(3):375-383. Todd, J.A. 2001. Neogene Marine Biota of Tropical America Molluscan Life Habits Database London: Smithsonian Tropical Research Institute Panama, Natural History Museum Basel Switzerland.

217

Appendix D

Relative abundance data for molluscs from Ebon Archaeological sites

218

MLEb-1, Ebon Islet

Table A Relative abundance of all taxonomic categories for MLEb-1, TP 17-20 by MNI, NISP and Weight in grams. IA IB IIA IIB NISP Weight NISP Weight NISP Weight NISP Weight MNI. MNI. MNI. MNI. P. (g) P. (g) P. (g) P. (g) Bivalvia Arcidae Arca spp. 2 3 2.96 1 2 1.06 Barbatia spp. 2 5 2.64 1 1 1.15 Cardiidae Acrosterigma spp. 1 1 0.21 1 2 2.79 Fragum fragum 3 7 1.18 6 11 3.02 1 2 0.69 Fragum spp. 1 1 0.12 1 1 0.33 Hippopus spp. 1 1 2.46 Tridacna crocea 1 7 27.18 1 1 2.45 Tridacna gigas 1 1 5.24 1 1 3.59 Tridacna maxima 1 10 223.89 1 2 17.35 Tridacna spp. 3 29 119.25 1 3 5.97 1 2 5.29 Tridacna squamosa 1 3 27.83 Vasticardium elongatum 1 1 4.66 Vasticardium spp. 1 1 5 1 2 13.48 Chamidae Chama pacifica 3 11 18.43 Chama spp. 6 23 49.34 2 5 42.99 Lucinidae Lucinidae 1 1 0.22 Ctena bella 2 4 1.3 2 5 2.58 2 3 0.81 2 2 0.36 Mesodesmatidae Atactodea striata 1 1 0.18 Psammobiidae Asaphis violascens 27 219 173.63 10 97 59.41 10 67 37.06 7 50 53.61 Gari spp. 1 3 6.42 Pteriidae Isognomon spp. 1 1 0.49 Pinctada spp. 1 3 1.48 Spondylidae Spondylus spp. 1 4 2.36 1 2 31.3 1 1 0.64 Tellinidae Tellina palatum 1 1 0.4 219

Tellina scobinata 1 1 0.16 Veneridae Gafrarium spp. 6 18 11.79 8 29 19.11 5 14 11.65 2 4 2.74 Periglypta puerpera 1 1 2.1 Gastropoda Buccinidae Pollia spp. 1 1 2.65 Pollia undosa 3 9 7.94 1 1 0.32

Bursidae Bursa spp. 13 74 130.85 4 8 25.29 1 1 0.53 3 4 26.46 Cassidae Cassis cornuta 1 1 9.66 Cerithidae Cerithium columna 33 54 25.8 5 11 4.3 3 3 1.44 2 2 1.6 Cerithium echinatum 3 4 3.86 2 2 0.79

Cerithium nodulosum 8 211 580.32 4 57 114.69 4 16 60.31 2 14 53.38

Cerithium spp. 17 29 9.6 3 6 1.45 4 7 2.79 1 1 0.73

Clypeomorus spp. 1 1 0.93

Conidae Conus araneosus 1 1 0.93

Conus ebraeus 2 3 3.33

Conus flavidus 1 1 9.11 1 1 20.93

Conus marmoreus 1 1 4.07

Conus miliaris 1 1 2.11

Conus spp. 57 254 405.57 13 58 94.83 1 5 9.54 2 5 11.56

Cypraeidae Cypraeidae 21 52 60.96 5 26 19.26 2 4 3.02 Cypraea spp. 7 6 13.74 2 5 7.84 Cypraea tigris 1 1 0.72 1 1 2.62 1 1 5.28

Erosaria erosa 1 1 3.33

Luria spp. 1 1 0.48 1 1 2.12

Lyncina spp. 2 3 12.52

Mauritia maculifera 1 2 3.57

Mauritia spp. 14 15 55.2 2 4 11.77 1 1 2.43 1 1 3.11

Moneta annulus 2 2 0.98

Monetaria caputserpentis 1 1 1.61

Monetaria moneta 25 34 27.29 12 21 15.65 2 2 1.4 1 1 0.71

220

Monetaria spp. 2 2 4.64

Ellobiidae Melampus flavus 8 12 4.23 2 2 0.8 Melampus spp. 1 5 0.25 1 1 0.15

Fasciolariidae Fasciolariidae 1 1 0.88 Hipponicidae Sabia conica 2 3 1.18 Mitridae Mitra mitra 1 1 0.52 Mitra stictica 4 16 19.31 1 3 2.47

Muricidae Muricidae 3 8 26.5 1 3 17.73 1 1 1.85 Chicoreus spp. 15 57 266.56 5 28 81.08 2 6 20.78 1 2 15.26 Drupa aperta 1 1 0.52

Drupa clathrata 1 2 15.95 1 1 1.12

Drupa grossularia 6 6 10.07

Drupa morum 17 22 36.12 2 3 10.53

Drupa ricinus 22 25 29.85 4 4 7.46

Drupa spp. 8 43 44.38 6 15 16.7 2 5 5.71 1 1 1.93

Mancinella alouina 1 2 9.31

Morula granulata 3 7 5.85 1 1 0.52

Morula spp. 2 3 2.34 4 6 4.35 1 1 0.75 2 2 1.06

Morula uva 4 4 3.73 1 1 0.99

Plicopurpura spp. 1 1 0.33

Thais armigera 5 17 200.88 1 3 15.04 1 2 12.74

Thais spp. 1 1 13.18

Naticidae Mammilla spp. 1 1 0.56

Polinices spp. 2 3 2.18 1 3 2.65

Neritidae Nerita exuvia 2 2 1.16 Nerita plicata 77 97 52.62 32 49 24.06 31 43 23.09 24 52 28.72

Nerita polita 13 34 18.22 15 42 24.03 7 20 11.32001 7 17 11.35

Nerita signata 4 4 1.43 2 2 0.53 3 6 1.85 1 1 0.14

Nerita spp. 1 2 0.48 1 1 0.58

Neritopsidae Neritopsis radula 1 1 0.56 2 2 1.72 1 1 0.38 221

Planaxinae Planaxis sulcatus 11 14 4.23 1 3 0.7 2 3 1.16 1 1 0.23 Ranellidae Ranellidae 3 5 2.79 1 1 0.25 1 3 3.09 Cymatium spp. 1 1 0.55 Guturnium muricinum 131 195 87.92 5 6 2.21 2 3 1.45 1 1 0.45

Monoplex intermedius 38 46 40.45 15 18 13.4 6 6 4.21 2 2 1.84

Monoplex nicobaricus 40 50 47.42 14 16 12.66 3 3 1.06 1 1 0.22

Monoplex spp. 15 56 24.03 1 9 5.52 1 2 1.22 1 2 0.83

Strombidae Strombidae 1 1 0.72 1 3 1.84 1 1 6.39 Canarium spp. 5 8 5.19 2 4 1.37 1 1 1.98 Canarium microurceus 2 2 1.67 Harpago chiragra 1 1 0.67

Lambis lambis 1 9 17.28 1 2 2.02

Lambis spp. 1 5 7.42 1 8 12.89 1 2 1.46 1 10 14.71

Strombus maculatus 2 2 0.74 1 1 0.92

Strombus spp. 1 1 0.3

Terebridae Terebridae 1 1 0.38 1 1 1.07 Myurella spp. 3 10 7.57 Oxymeris areolata 1 1 2.12

Oxymeris crenulata 2 12 14.37 1 2 11.8

Oxymeris spp. 1 1 0.64

Terebra spp. 2 6 2.07 1 1 0.47

Tonnidae Tonnidae 1 1 0.42 Tonna spp. 1 2 0.75 Triviidae Trivirostra spp. 2 2 1.19 1 1 0.6 Trochidae Trochus maculatus 19 56 95.09 9 26 47.01 3 9 12.08 1 3 14 Trochus spp. 2 6 5.96 1 5 2.36 1 1 0.5

Turbinellidae Vasum turbinellus 14 83 218.09 5 23 56.74 2 5 12.91 2 4 78.21 Vasum spp. 2 4 13.05 1 1 2.86

Turbinidae Turbo argyrostomus 9 72 272.43 1 12 36.35 1 1 3.56 1 5 46.4 Turbo setosus 2 55 183.47 1 10 19.48 1 4 31.19 1 3 17.08

222

Turbo spp. 7 83 60.27 2 14 15.92 1 7 7.15 1 3 5.67

Vermetidae Vermetidae 1 1 2.43 1 1 0.65 Unidentified 68 27.12 23 11.02 4 0.68 8 1.17

223

Table B Relative abundance of all taxonomic categories for MLEb-1, TP 6 by MNI, NISP and Weight in grams. IA IB IIA IIB

Weight Weight Weight Weight MNI. NISPP. MNI. NISPP. MNI. NISPP. MNI. NISPP. (g) (g) (g) (g) Bivalvia Arcidae Barbatia spp. 1 1 0.31 Cardiidae Cardiidae 1 2 1.84 Acrosterigma spp. 1 1 6.82 Fragum fragum 2 4 0.77 2 4 0.99 1 1 0.49 Fragum spp. 1 2 0.27

Hippopus hippopus 1 1 19.97

Hippopus spp. 1 1 0.39

Tridacna crocea 1 6 43.92 1 2 20.81

Tridacna maxima 1 9 115.56 1 1 11.92

Tridacna spp. 1 22 55.61 1 18 56.35 1 5 17.72 1 11 19.37

Vasticardium 1 1 1.36 1 1 14.62 elongatum Vasticardium spp. 1 2 3.84

Chamidae Chama spp. 1 1 1.33 Lucinidae Codakia spp. 1 1 0.79 Ctena bella 2 5 6.31 1 2 0.64 Mesodesmatidae Atactodea striata 2 2 0.42 1 10 3.1 Psammobiidae Asaphis violascens 5 38 29.35 3 31 12.83 8 33 41.1 2 10 1.95 Pteriidae Pteriidae 1 2 0.53 Pinctada spp. 1 3 3.82 1 2 0.52 Spondylidae Spondylus spp. 1 3 3.82 Tellinidae Tellinidae 1 1 0.83 Tellina palatum 2 15 12.63 2 14 7.41 1 2 0.6 1 2 0.81 Tellina spp. 1 1 0.48 1 5 2.67 Veneridae Veneridae 1 2 1.21 224

Antigona spp. 1 1 24.89 Gafrarium pectinatum 1 2 1.17 Gafrarium spp. 2 9 3.24 5 31 22.13 5 13 11.37 1 3 2.39 Periglypta puerpera 1 1 1.06 Periglypta spp. 1 3 1.98 Gastropoda Bursidae Bursa rhodostoma 1 1 8.03 Bursa spp. 1 1 0.43 Cassidae Cassis cornuta 1 1 53.45

Cerithidae Cerithium columna 2 4 1.21 1 1 0.77

Cerithium nodulosum 2 38 110.43 1 4 4.95

Cerithium spp. 3 4 1.22 1 1 0.31

Conidae Conus distans 1 1 1.81

Conus marmoreus 1 1 5

Conus spp. 14 25 37.41 4 14 8 3 4 2.65

Cypraeidae Cypraeidae 1 1 0.68 1 1 1.97 Cypraea spp. 1 3 1.25 Cypraea tigris 1 1 16.22 1 1 1.48

Erosaria helvola 1 2 0.82

Mauritia spp. 1 1 3.85

Monetaria moneta 2 3 1.3 1 3 1.42 1 1 0.5

Monetaria obvelata 1 1 0.35

Ellobiidae Melampus flavus 1 1 0.11

Haliotidae Haliotidae 1 1 0.16

Littorinidae Littoraria coccinea 1 1 0.4 Mitridae Mitra spp. 1 1 1.07 1 1 0.16

Mitra stictica 1 3 5.96 1 1 0.66

Muricidae Drupa morum 3 4 5.47

Drupa ricinus 2 2 1.49

Drupa spp. 1 1 0.47 1 1 0.77

225

Morula granulata 1 2 0.4 1 1 0.57

Thais armigera 2 2 29.22

Naticidae Naticidae 2 3 2.82

Neritidae Nerita plicata 5 10 3.99 1 3 0.87 3 5 1.1 3 4 0.82 Nerita polita 11 11 5.97 8 25 9.78 3 7 3.26 1 2 0.52

Nerita spp. 1 8 2.51

Nerita undata 1 1 0.31

Planaxinae Planaxis sulcatus 1 1 0.26 1 1 0.55 Ranellidae Ranellidae 1 2 9.36 Cymatium spp. 1 1 1.15 Guturnium muricinum 1 2 0.71 1 1 0.52

Monoplex intermedius 3 3 4.15 1 1 0.67

Monoplex nicobaricus 2 2 0.71 1 1 0.2

Monoplex spp. 1 1 0.31 1 1 0.21

Strombidae Harpago chiragra 1 1 7.28

Lambis lambis 1 1 0.72

Lambis spp. 1 10 27.76 1 2 1.7 1 1 0.95

Terebridae Terebra spp. 1 1 0.38

Trochidae Trochus maculatus 1 2 3.57

Trochus spp. 1 1 0.46 1 1 0.57

Turbinellidae Vasum turbinellus 10 28 89.82 1 2 12.58

Turbinidae Turbo spp. 1 2 1.03 1 5 1.25 1 2 1.61

Turbo argyrostomus 1 12 50.63 1 2 10.6 Turbo setosus 1 1 0.81 1 1 1.76

Turbo stenogyrus 1 3 5.11

Unidentified 21 8.8 21 8.56 13 5.54 7 0.38

226

Table C Relative abundance of all taxonomic categories for MLEb-1, TP 8-12 by MNI, NISP and Weight in grams A

Weight MNI. NISPP. (g)

Bivalvia Arcidae Barbatia (Savignyarca) spp. 1 1 0.09 Cardiidae Acrosterigma spp. 1 2 8.88 Fragum fragum 6 21 6.51 Tridacna crocea 1 1 18.5 Tridacna maxima 1 4 50.61 Tridacna spp. 1 10 62.13 Mesodesmatidae Atactodea striata 1 39 11.25 Fragum fragum 6 21 6.51 Psammobiidae Asaphis violascens 5 8 4.11 Veneridae Gafrarium spp. 2 5 3.09 Gastropoda Buccinidae Engina mendicaria 1 1 1 Cerithidae Cerithium columna 1 1 0.28 Cerithium nodulosum 1 3 4.64

Conidae Conus spp. 3 3 19.14 Cypraeidae Monetaria moneta 2 2 0.95 Naticidae Naticidae 3 3 0.77 Neritidae Nerita plicata 2 2 0.44 Nerita polita 5 8 9.07

Ranellidae Monoplex spp. 1 1 0.47 Strombidae Harpago chiragra 1 1 0.73 Lambis lambis 1 1 40.87

Lambis spp. 1 2 4.6

Terebridae Oxymeris crenulata 1 2 0.78 Trochidae Trochus maculatus 1 1 0.3 Trochus spp. 1 3 3.89

Turbinellidae Vasum turbinellus 1 1 2.06 Turbinidae Turbo argyrostomus 1 3 41.69 Turbo stenogyrus 1 1 1.68

Unidentified - 11 7.22

227

Table D Relative abundance of all taxonomic categories for MLEb-1, TP 13-16 by MNI, NISP and Weight in grams A I

Weight Weight MNI. NISPP. MNI. NISPP. (g) (g) Bivalvia

Arcidae Barbatia spp. 2 2 0.34 Cardiidae Fragum fragum 8 16 4.55 1 1 0.44 Trachycardium spp. 1 1 0.18 Tridacna crocea 1 2 10.53 Tridacna maxima 1 3 42.47 1 1 30.31 Tridacna spp. 1 1 0.44 1 2 5.31 Cardiidae Fragum fragum 8 16 4.55 1 1 0.44 Trachycardium spp. 1 1 0.18 Chamidae Chama spp. 1 1 1.09 Lucinidae Ctena bella 2 2 0.45 Mesodesmatidae Atactodea striata 1 3 0.68 1 16 3.97 Psammobiidae Asaphis violascens 2 10 5.83 2 3 7.34 Veneridae Antigona spp. 1 2 6.05 Gafrarium spp. 1 1 0.4 1 1 1.21 Gastropoda Bursidae Bursa spp. 5 5 11.2 2 2 4.38 Cerithidae Cerithium columna 1 1 0.72 1 1 0.27 Cerithium nodulosum 3 11 114.56 1 4 9.19

Cerithium spp. 1 3 0.87

Conidae Conus ebraeus 1 1 4.83

Conus spp. 1 4 3.92 3 4 6.62

Cypraeidae Cypraeidae 3 8 9.18 2 2 2.13 Cypraea tigris 1 1 1.95 Lyncina spp. 1 1 4.56

Mauritia mauritiana 1 1 4.69 Monetaria 1 1 4.86 caputserpentis Monetaria moneta 3 3 1.65 Janthinidae Janthinidae 1 1 0.36 Muricidae Muricidae 1 1 2 Chicoreus spp. 2 3 15.7

Drupa morum 1 1 1.41

Drupa ricinus 2 2 1.98

Drupa spp. 1 3 2.51 1 2 1.49

Morula granulata 1 1 0.72

Morula uva 1 1 0.71

Neritidae Nerita exuvia 1 1 2.99

Nerita plicata 2 2 1.56

Nerita polita 10 20 13.46 1 2 0.43

228

Planaxinae Planaxis sulcatus 3 3 0.88 1 1 0.24 Ranellidae Guturnium muricinum 1 1 0.4

Monoplex intermedius 2 2 1.95

Monoplex nicobaricus 3 3 4.5 2 3 3.11

Monoplex spp. 1 1 0.61

Strombidae Harpago chiragra 1 2 40.09

Lambis lambis 1 1 34.18

Lambis spp. 1 1 0.63

Strombus spp. 1 2 1.28

Terebridae Oxymeris crenulata 1 1 1.61 1 2 21.38 Tonnidae Malea spp. 1 1 2.18

Trochidae Trochus maculatus 4 8 18.79 2 4 11.74 Trochus spp. 1 1 0.19

Turbinellidae Vasum turbinellus 5 9 31.13 1 4 6.26 Vasum ceramicum 1 2 5.11

Turbinidae Turbo argyrostomus 3 6 101.04 2 3 34.55 Turbo setosus 1 1 4.5

Turbo spp. 1 10 13.73 2 6 13.27

Turbo stenogyrus 1 1 13.72

Vermetidae Vermetidae 1 3 2.9

Unidentified 12 7.79 2 0.71

229

MLEb-5, Ebon Islet

Table E Relative abundance of all taxonomic categories for MLEb-1, 3 x 5m area excavation by MNI, NISP and Weight in grams A

Weight MNI. NISPP. (g) Bivalvia Arcidae Arca spp. 1 1 0.96 Arca ventricosa 1 1 0.6 Barbatia (Savignyarca) spp. 5 11 24.07

Cardiidae Acrosterigma spp. 2 10 13.41 Fragum fragum 4 21 10.77 Fragum spp. 5 13 5.59 Hippopus hippopus 2 15 225.24 Trachycardium spp. 1 1 0.51 Tridacna crocea 2 5 312.27 Tridacna maxima 39 211 6953.26 Tridacna spp. 1 752 3976.55 Tridacna squamosa 3 4 30.98 Vasticardium elongatum 6 43 230.35 Vasticardium spp. 1 11 116.89 Vasticardium angulatum 1 1 0.77 Chamidae Chama pacifica 1 1 4.05 Chama spp. 2 20 50.45 Lucinidae Codakia punctata 2 2 24.07 Ctena bella 6 16 5.92 Mesodesmatidae Atactodea striata 1 1 0.14 Spondylidae Spondylus spp. 1 5 11.9 Spondylus squamosus 1 1 35.52 Pectinidae 1 1 0.1 Psammobiidae Asaphis violascens 108 786 810.19 Tellinidae Tellina palatum 8 56 59.57 Tellina scobinata 1 5 7.53 Tellina spp. 1 3 6.13 Veneridae Gafrarium spp. 65 269 221.11 Periglypta puerpera 1 21 54.43 Gastropoda Buccinidae Pollia undosa 2 3 4.36 Bursidae Bursa bufonia 1 1 11.32 Bursa spp. 4 17 73.62

Cassidae Cypraecassis rufa 1 1 0.65 Cerithidae Cerithium columna 5 6 2.79 Cerithium echinatum 2 2 1.24

Cerithium nodulosum 13 64 295.59

Cerithium spp. 2 9 4.78

230

Conidae Conus chaldeus 1 2 5.85 Conus distans 1 9 90.15

Conus ebraeus 2 2 10.46 Conus marmoreus 2 4 35.43 Conus miliaris 1 1 18.52 Conus spp. 62 159 553.04 Cypraeidae Cypraeidae 29 130 189.04 Erosaria helvola 1 1 0.42 Erosaria spp. 2 2 0.95 Lyncina lynx 1 1 7.68 Lyncina spp. 8 26 53.91 Mauritia spp. 10 13 26.91 Moneta annulus 4 4 2.74 Monetaria caputserpentis 7 7 41.07 Monetaria moneta 15 21 22.63 Talparia talpa 1 1 1.83 Ellobiidae Melampus flavus 4 8 3.36 Fasciolariidae Fasciolariidae 1 1 1.07 Harpidae Harpa amouretta 1 1 1.54 Harpa spp. 1 3 1.44 Hipponicidae Sabia conica 2 2 1.02 Mitridae Mitridae 1 1 0.43 Mitra spp. 1 1 2.81 Mitra stictica 2 6 6.48 Muricidae Muricidae 5 44 28.56 Chicoreus spp. 1 2 3.56 Drupa grossularia 1 1 1.16 Drupa morum 7 14 21.16 Drupa ricinus 7 9 6.57 Drupa spp. 3 14 9.12 Morula granulata 1 1 0.57 Morula spp. 1 1 0.49 Morula uva 5 5 3.01 Plicopurpura spp. 1 1 0.37 Thais armigera 3 10 110.45 Thais spp. 1 2 1.17 Thalessa virgata 4 9 13.73 Naticidae Naticidae 1 1 0.3 Mammilla spp. 1 2 1.2 Notocochlis spp. 26 33 31.55 Neritidae Nerita exuvia 1 1 0.64 Nerita plicata 212 320 247.11 Nerita polita 112 219 149.45 Nerita signata 6 8 2.78 Nerita spp. 13 45 8.96 Nerita undata 1 1 0.26 Patellidae Scutellastra flexuosa 3 3 0.8 231

Planaxinae Planaxis sulcatus 4 6 1.08 Pteriidae Isognomon spp. 1 7 27.23 Pinctada spp. 2 46 67.56 Ranellidae Ranellidae 1 2 0.31 Guturnium muricinum 6 8 3.97 Monoplex intermedius 7 9 5.19 Monoplex nicobaricus 3 4 1.68 Monoplex spp. 3 8 3.1 Strombidae Strombidae 1 1 0.27 Canarium maculatum 1 1 1.98 Canarium spp. 5 7 10.69 Canarium wilsonorum 1 1 3.05 Harpago chiragra 16 117 2399.84 Lambis lambis 1 10 65.68 Lambis spp. 1 25 86.88 Strombus spp. 1 1 0.21 Terebridae Terebridae 1 3 2.15 Oxymeris crenulata 1 1 0.32 Oxymeris spp. 1 1 6.92 Triviidae Trivirostra spp. 2 2 0.83 Trochidae Trochus maculatus 6 55 63.55 Trochus spp. 1 8 5.48 Turbinellidae Vasum turbinellus 6 30 134.2 Turbinidae Turbo argyrostomus 9 84 241.49 Turbo setosus 22 43 714.79 Turbo spp. 5 77 107.88 Vermetidae Vermetidae 1 3 1.17 Unidentified 299 202.84

232

MLEb-31, Moniak Islet

Table F Relative abundance of all taxonomic categories for MLEb-31, TP 2 - 6 by MNI, NISP and Weight in grams IA IB IIA IIB IC SD

Wt. Wt. Wt. Wt. Wt. Wt. MNI NISP MNI NISP MNI NISP MNI NISP MNI NISP MNI NISP (g) (g) (g) (g) (g) (g) Bivalvia

Arcidae Barbatia spp. 1 1 0.31

Cardiidae Cardiidae 1 6 0.66

Fragum spp. 1 1 0.02 4 5 0.34 1 1 0.69 Hippopus 1 1 87.42 hippopus Tridacna 2 2 136.68 crocea Tridacna 4 7 988.12 1 3 64.8 1 5 127.4 1 6 102.28 maxima Tridacna spp. 1 15 90.5 1 6 16.06 1 3 37.55 1 1 0.09 Tridacna 1 1 7.71 1 1 1.73 squamosa Vasticardium 1 1 1.56 1 1 2.33 spp. Beguina Carditidae 1 1 0.51 semiorbiculata Chamidae Chama spp. 1 1 4.14 3 4 12.47

Lucinidae Ctena bella 1 1 0.1

Psammo- Asaphis 32 139 150.5 15 105 103.11 10 41 61.15 2 11 13.49 3 10 9.52 2 6 7.1 biidae violascens 233

Spondylus Spondylidae 1 2 49.41 1 2 34.37 sinensis Spondylus 1 2 9.09 1 1 0.28 spp. Periglypta Veneridae 1 3 3.54 1 1 1.92 puerpera Pitar striatus 1 1 2.55

Gastropoda

Buccinidae Pollia undosa 2 3 3.95

Bursidae Bursa spp. 2 3 9.01 3 9 7.03 2 10 11.19 1 4 3.6

Cerithium Cerithidae 1 1 0.43 columna Cerithium 6 23 128.39 2 28 68.57 10 74 318.21 1 6 15.3 1 2 3.27 1 4 10.19 nodulosum Cerithium 1 1 0.25 spp. Conidae Conus catus 1 2 1.45 1 1 0.96 1 2 3.74

Conus miles 2 2 45.21 1 2 6.9 Conus 1 1 32.21 1 1 2.66 miliaris Conus spp. 5 19 103.04 10 38 119.96 12 49 137.27 1 9 75.09 1 3 8.76 2 3 30.2

Cypraeidae Cypraeidae 4 9 7.87 2 6 4.24 1 2 1.32 1 2 1.66 1 1 2.44

Cypraea spp. 1 1 2.1 1 2 4.75

Erosaria 1 1 0.47 helvola

234

Luria isabella 1 1 14.38

Luria spp. 1 1 1.56

Mauritia 1 2 5.7 1 1 0.55 eglantina Mauritia spp. 2 2 6.46 5 10 16.23 3 3 5.33 1 1 1.68

Monetaria 2 2 9.26 1 1 6.45 caputserpentis Monetaria 4 4 3.97 1 1 0.66 moneta Monetaria 1 1 2.16 spp. Lyncina leviathan Melampus Ellobiidae 2 2 1.14 flavus

Fascio- Latirus 1 1 6.02 lariidae maculatus Harpa Harpidae 1 1 0.95 amouretta Hippon- Sabia conica 2 3 0.62 icidae Mitra Mitridae 1 1 1.6 imperialis

Muricidae Muricidae 1 1 1.27 Chicoreus 1 1 2.71 brunneus Chicoreus 1 2 4.28 2 3 18.23

235

spp.

Drupa aperta 1 1 0.73 3 3 7.29

Drupa 1 1 10.45 clathrata Drupa 1 1 1.45 1 1 2.09 grossularia Drupa morum 3 3 9.98 1 2 1.7

Drupa ricinus 2 2 3.24 1 1 0.56 1 2 3.42

Drupa spp. 5 11 7.87

Mancinella 1 1 3.11 alouina Morula 1 1 1.55 granulata Morula spp. 1 2 0.49 1 1 0.47 1 1 1.02

Naquetia 1 1 1.79 cumingii Thais 24 88 568.99 4 35 142.4 2 10 99.78 1 6 21.19 1 1 12.13 1 1 3.61 armigera Thalessa 1 2 2.44 1 2 15.81 virgata Nassarius Nassariidae 1 1 1.55 1 1 2.12 papillosus Nassarius 1 1 0.08 spp. Naticidae Polinices spp. 1 1 0.85

Mammilla

236

simiae

Nerita Neritidae 1 1 0.87 albicilla Nerita plicata 2 2 1.79 8 8 3.48 2 2 0.88 1 1 0.6

Nerita polita 36 48 51.79 82 130 103.87 57 75 78.86 1 26 26.67 1 1 1 11 16 20.2

Nerita signata 4 6 4.77 1 1 1.58

Nerita spp. 1 2 0.96 2 14 5.76 1 4 1.73 1 1 0.72

Nerita undata 1 2 1.1 9 12 10.96 1 0 0

Oliva Olividae 1 1 3 annulata Oliva minicea 1 1 3.23 1 1 0.74

Scutellastra Patellidae 1 1 0.37 flexuosa Isognomon Pteriidae 1 1 0.53 spp. Charonia Ranellidae 1 3 5.51 1 1 3.24 tritonis Guturnium 5 5 1.71 14 17 7.4 7 6 3.74 1 1 0.33 muricinum Gyrineum 1 1 0.79 pusillum Monoplex 2 2 1.86 4 5 4.4 1 2 1.57 2 2 2.21 intermedius Monoplex 6 6 5.15 2 3 6.15 1 1 0.88 nicobaricus Monoplex 1 1 0.8 1 3 0.82 1 1 0.31 1 1 0

237

spp.

Ranularia 1 1 2.17 testudinaria

Strombidae Strombidae 1 2 2.7 1 7 2.16 1 1 0.97 Canarium 5 10 4.44 1 1 3.28 spp. Harpago 1 2 31.99 chiragra Lambis 1 1 11.34 1 2 14.07 lambis Lambis spp. 1 1 0.37

Lentigo 1 1 3.23 lentiginosus Strombus 2 2 5.01 luhuanus Strombus spp. 1 6 2.31 1 2 1.57

Terebridae Terebridae 1 3 0.77 Hastula 1 1 0.63 solida Oxymeris 1 5 9.82 1 1 3.19 1 1 0.5 crenulata Oxymeris spp. 1 1 1.59

Tonnidae Malea pomum 1 2 1.71 1 3 3.87 1 1 1.25

Tonna perdix

Trochus Trochidae 3 10 495.2 7 31 42.81 4 5 21.34 1 3 6.49 maculatus

238

Vasum Turbinellidae 9 19 113.95 9 49 118.64 19 66 389.91 9 22 118.36 3 12 34.05 turbinellus Turbo Turbinidae 2 26 99.19 4 22 98.05 4 15 74.66 5 10 51.81 3 4 20.71 2 8 55.09 argyrostomus Turbo setosus 3 6 31.53 3 13 125.17 1 4 35.09 1 1 1.62 2 2 9.57

Turbo spp. 1 19 32.04 3 39 71.1 1 11 38.76 1 6 10.44 1 1 3.59

Turritellidae Turritella spp. 1 1 0.87

Vermetidae Vermetidae 1 2 1.02

Turridae Turridae 1 1 1.72

Unidentified 15 14.91 37 18.03 20 6.64 3 3.13 1 0.52

239

MLEb-33, Enekoion Islet

Table G Relative abundance of all taxonomic categories for MLEb-3, TP 1, 6, and 7 by MNI, NISP and Weight in grams I

MNI. NISPP. Weight (g) Bivalvia Arcidae Barbatia spp. 1 1 0.24 Cardiidae Acrosterigma spp. 1 1 0.4 Corculum cardissa 5 9 2.32 Fragum spp. 67 162 54.09 Tridacna maxima 2 3 263.23 Tridacna spp. 1 2 9.91 Chamidae Chama spp. 11 28 263.34 Lucinidae Ctena bella 2 3 0.87 Psammobiidae Asaphis violascens 19 125 106.15 Tellinidae Tellina palatum 1 1 0.05 Veneridae Veneridae 1 1 0.15 Gafrarium spp. 1 1 0.45 Gastropoda Buccinidae Pollia undosa 2 2 1.54 Bursidae Bursa spp. 15 42 78.71 Cerithidae Cerithidae 1 4 0.87 Cerithium columna 1 1 0.28 Cerithium nodulosum 7 85 253.8

Cerithium spp. 23 26 7.66

Conidae Conus marmoreus 1 1 6.81 Conus spp. 29 89 320.82 Cypraeidae Cypraeidae 6 8 14.19 Erosaria helvola 1 2 0.82 Mauritia spp. 1 1 7.8 Monetaria moneta 2 5 1.78 Ellobiidae Melampus flavus 32 41 15.91 Melampus spp. 1 1 0.08 Harpidae Harpa amouretta 1 1 0.34 Harpa spp. 1 1 0.4 Hipponicidae Sabia conica 1 2 0.71 Mitridae 1 3 1.02 Mitridae Mitra spp. 1 1 0.93 Mitra stictica 3 8 10.03

Pterygia spp. 1 1 1.27 Muricidae Muricidae 3 8 12.16 Chicoreus spp. 1 1 1.59 Drupa aperta 1 1 2.8 Drupa morum 2 3 8.14 Drupa ricinus 5 5 4.36

240

Drupa rubusidaeus 1 5 25.42 Drupa spp. 1 1 0.4 Morula uva 2 2 1.88 Thais armigera 3 11 49.44 Thais spp. 1 1 1.1 Thalessa virgata 1 2 1.99 Nassariidae Nassarius papillosus 1 1 0.26 Naticidae Mammilla spp. 2 2 0.33 Neritidae Nerita albicilla 2 2 2.45 Nerita plicata 43 65 42.7 Nerita polita 121 225 160.41 Nerita signata 1 1 0.6 Nerita spp. 3 9 1.83 Patellidae Scutellastra flexuosa 1 1 0.06 Planaxinae Planaxis sulcatus 8 9 2.42 Ranellidae Ranellidae 1 1 0.4 Guturnium muricinum 11 12 4.51 Monoplex intermedius 3 3 4.15 Monoplex nicobaricus 8 8 4 Monoplex spp. 1 7 2.7 Spondylidae 1 2 1.72 Spondylidae Spondylus sinensis 2 3 60.7 Spondylus spp. 1 1 0.63 Strombidae Canarium spp. 5 11 4.94 Lambis spp. 1 6 4.55 Terebridae Terebridae 1 2 1.16 Tonnidae Tonnidae 1 12 0.49 Malea spp. 1 1 2.34 Tonna pennata 1 2 0.7 Trochidae Trochus maculatus 3 8 12.66 Trochus spp. 1 1 0.28 Turbinellidae Vasum turbinellus 45 137 438.75 Turbinidae Turbo argyrostomus 16 80 290.59 Turbo setosus 6 14 57.4 Turbo spp. 3 34 45.33 Vermetidae Vermetidae 1 2 0.40 Unidentified 28 13.5

241

Table H Relative abundance of all taxonomic categories for MLEb-33, TP 2 by MNI, NISP and Weight in grams IA IB

Weight Weight MNI. NISPP. MNI. NISPP. (g) (g) Bivalvia Cardiidae Fragum spp. 5 5 11.2 2 2 4.38 Tridacna maxima 8 16 4.55 1 1 0.44 Tridacna spp. 1 1 0.18 Chamidae Chama spp. 1 1 0.44 1 2 5.31 Psammobiidae Asaphis violascens 1 1 4.86 Spondylidae Spondylus sinensis 1 3 0.68 1 16 3.97 Tellinidae Tellina palatum 1 3 2.51 1 2 1.49 Gastropoda Bursidae Bursa spp. 2 2 0.34 Vasticardium elongatum 1 2 10.53

Cerithidae Cerithium nodulosum 1 3 42.47 1 1 30.31

Conidae Conus spp. 1 1 0.72 1 1 0.27 Cypraeidae Cypraeidae 3 11 114.56 1 4 9.19 Muricidae Chicoreus spp. 1 1 1.09

Drupa spp. 1 1 4.83

Naquetia cumingii 1 4 3.92 3 4 6.62

Neritidae Nerita plicata 1 1 1.95

Nerita polita 3 8 9.18 2 2 2.13

Nerita signata 1 1 4.56

Nerita spp. 1 1 4.69

Ranellidae Guturnium muricinum 3 3 1.65

Monoplex intermedius 1 1 0.36

Monoplex spp. 2 2 0.45

Strombidae Strombidae 2 2 1.98 Harpago chiragra 12 7.79 2 0.71 Lambis lambis 2 3 15.7

Lambis spp. 1 1 1.41

Trochidae Trochus maculatus 1 1 0.72

Turbinellidae Vasum turbinellus 1 1 0.71

Turbinidae Turbo argyrostomus 1 1 2

Turbo setosus 1 1 2.99

Turbo spp. 2 2 1.56

Unidentified 1 3 0.87

242

Table I Relative abundance of all taxonomic categories for MLEb-33, TP 8 by MNI, NISP and Weight in grams

MNI. NISPP. Weight (g) Bivalvia Cardiidae Fragum spp. 2 2 1.54 Tridacna maxima 15 42 78.71 Tridacna spp. 1 1 0.4 Vasticardium elongatum 5 9 2.32 Chamidae Chama spp. 2 3 263.23 Psammobiidae Asaphis violascens 1 1 0.93 Pteriidae Pinctada spp. 3 8 10.03 Spondylidae Spondylus spp. 1 1 1.59 Veneridae Gafrarium spp. 1 2 1.99 Periglypta spp. 1 1 0.26 Gastropoda Bursidae Bursa spp. 1 1 0.24 Cerithidae Cerithium nodulosum 67 162 54.09 Conidae Conus distans 1 2 9.91 Conus leopardus 1 1 0.28

Conus lividus 7 85 253.8

Conus spp. 23 26 7.66

Cypraeidae Cypraeidae 1 4 0.87 Monetaria moneta 11 28 263.34

Fasciolariidae Fasciolariidae 1 1 6.81 Harpidae Harpa spp. 29 89 320.82 Mitridae Mitra mitra 6 8 14.19 Muricidae Muricidae 1 1 0.08 Chicoreus spp. 1 1 7.8 Drupa ricinus 2 5 1.78

Drupa spp. 32 41 15.91

Thais armigera 1 1 0.34

Neritidae Nerita plicata 1 1 0.4 Nerita polita 1 2 0.71

Nerita signata 2 3 0.87

Ranellidae Guturnium muricinum 1 1 1.27 Monoplex nicobaricus 1 3 1.02 Monoplex spp. 28 13.5 Strombidae Strombidae 2 3 8.14 Lambis spp. 1 1 2.8 Terebridae Terebridae 1 5 25.42 Oxymeris crenulata 5 5 4.36 Trochidae Trochus maculatus 1 1 0.4 Turbinellidae Vasum turbinellus 2 2 1.88

243

Turbinidae Turbo argyrostomus 3 8 12.16 Turbo setosus 3 11 49.44 Turbo spp. 1 1 1.1 Unidentified 1 2 0.82

244

Table J Relative abundance of all taxonomic categories for MLEb-33, TP 3 by MNI, NISP and Weight in grams I

MNI. NISPP. Weight (g) Bivalvia Arcidae Arca spp. 1 1 1.18 Barbatia spp. 1 1 0.31 Cardiidae Fragum spp. 7 20 0.58 Tridacna spp. 1 2 11.16 Chamidae Chama spp. 1 1 1.56 Psammobiidae Asaphis violascens 6 52 27.03 Veneridae Antigona spp. 1 1 1.2 Gastropoda Bursidae Bursa spp. 8 54 100.25 Cerithidae Cerithium nodulosum 67 769 1763.65 Cerithium spp. 5 5 1.14

Conidae Conus distans 2 6 70.9 Conus leopardus 1 2 4.74

Conus marmoreus 2 7 33.21

Conus spp. 55 216 675.57

Cypraeidae Cypraeidae 2 17 12.64 Erosaria helvola 1 2 2.58

Monetaria moneta 3 3 2.58

Harpidae Harpa spp. 1 1 0.17 Hipponicidae Sabia conica 1 1 0.52 Mitridae Mitra stictica 1 3 4.07 Muricidae Muricidae 1 1 0.47 Chicoreus spp. 2 15 72.11 Drupa morum 1 1 0.78

Drupa ricinus 1 1 0.64

Drupa rubusidaeus 2 9 42.88

Morula spp. 1 1 0.64

Thais armigera 2 10 124.43 Naticidae Naticidae 1 1 0.35 Neritidae Nerita plicata 14 20 11.48 Nerita polita 30 44 35.16 Nerita signata 2 3 1.36 Nerita spp. 1 14 1.14 Planaxinae Planaxis sulcatus 5 5 1.82 Ranellidae Guturnium muricinum 23 24 14.7 Monoplex intermedius 1 1 1.98 Monoplex nicobaricus 11 12 9.09 Monoplex spp. 1 10 3.99 Strombidae Strombidae 1 5 5.01 Harpago chiragra 2 4 4.06

245

Lambis lambis 3 3 78.2 Lambis spp. 1 1 2.45 Terebridae Terebridae 1 3 3.49 Trochidae Trochus maculatus 7 44 45.46 Turbinellidae Vasum turbinellus 35 322 1046.46 Turbinidae Turbo argyrostomus 11 33 141.48 Turbo setosus 3 6 12.32 Turbo spp. 1 37 11.49 Unidentified 13 10.2

246