Lapita Textpages

See discussions, stats, and author profiles for this publication at: https://www.researchgate.
net/publication/293820549
Lapita Ancestors and Descendants
Book · January 2009
CITATIONS READS
4 1,238
2 authors:
Peter J. Sheppard Tim Thomas

University of Auckland University of Otago
91 PUBLICATIONS 2,069 CITATIONS 27 PUBLICATIONS 203 CITATIONS
SEE PROFILE SEE PROFILE
All content following this page was uploaded by Peter J. Sheppard on 11 February 2016.
The user has requested enhancement of the downloaded file.

Lapita
Ancestors and Descendants
Lapita
Ancestors and Descendants
Edited by
Peter J. Sheppard
Tim Thomas
Glenn R. Summerhayes
NEW ZEALAND ARCHAEOLOGICAL ASSOCIATION MONOGRAPH 28

2009
New Zealand Archaeological Association
PO Box 56443 Dominion Rd, Auckland 1446
Monograph series editors: Dorothy Brown and Simon Holdaway

Editors this volume: Simon Holdaway
Cover photograph:
Gizo, Western Solomon Islands, November 2005 (Bilua Bifoa Project)
Photographer: Adrian Taylor
Orders to:
New Zealand Archaeological Association

PO Box 56443 Dominion Rd,
Auckland 1446
or http://nzarchaeology.org
© 2009 New Zealand Archaeological Association
ISSN 0111-5715
ISBN 978-0-9582977-1-4
Design and page layout by Hamish Macdonald

Printed by Publishing Press Ltd., Auckland.
TABLE OF CONTENTS
PREFACE vi
list of contributors vii
1. Understanding the Who, What and Why of Lapita:
What Can Archaeological Ancestors and Descendants Contribute? 1
Peter Sheppard
2. The aceramic to ceramic boundary in the Bismarck Archipelago 11

Jim Specht
3. A Review of Solomon Island Archaeology 35

Richard Walter and Peter Sheppard
4. Inter-tidal Late Lapita Sites and Geotectonics in the

Western Solomon Islands 73
Peter J. Sheppard and Richard Walter
5. Assessing the anomalous role of ceramics in late-Lapita

interaction: A View from Kolombangara, Western Solomon Islands 101
Amy F. Findlater, Glenn.R. Summerhayes, William R. Dickinson and Ian A. Scales
6. Communities of practice in the archaeological record

of New Georgia, Rendova and Tetepare 119
Tim Thomas
7. The Southeast Solomon Islands Culture History Project:

Principal Investigators’ Overview of the 1970s Project Including Recent
and Current Research 147
Roger C. Green and Douglas E Yen
8. Ethnobotany In The Southeast Solomon Islands Cultural

History Project 173
Douglas E. Yen
9. The Post-Lapita Sequence in the Reef/Santa Cruz 181

Moira Doherty
10. The Teouma Lapita site, South Efate, Vanuatu:

A Summary of Three Field Seasons (2004-2006) 215
Stuart Bedford, Matthew Spriggs, Hallie Buckley, Frédérique Valentin
and Ralph Regenvanu
11. Understanding the place properly: Palaeogeography of Selected

Lapita Sites in the Western Tropical Pacific Islands and its Implications 235
Patrick D. Nunn and Tony Ahikau Heorake
12. Return To The Entangled Bank: Deciphering the Lapita Cultural Series 255
John Edward Terrell
v
preface
This volume contains papers presented at the 7th International Lapita Conference (Lapita:
Antecedents and Successors ) which was held (July 4th - 7th 2007) in Honiara, Solomon Islands.
The conference was a considerable success, attracting over 50 delegates from around the world
(New Zealand, Australia, Vanuatu, New Caledonia, Fiji, Japan, Taiwan, USA, Canada, France)
and was attended by many Solomon Island government officials and educators. The conference
was run in association with the National Museum of the Solomon Islands through the able assis-
tance of the Director Mr Lawrence Foana‘ota and was opened by the Minister of Culture and
Tourism. Following the conference a number of additional papers were solicited to build on the
core of papers dealing with the Solomon Islands. We would like to thank those additional
contributors for helping produce a useful and unique set of papers dealing with the archaeology
of the Solomons. We hope they encourage Solomon Island students to pursue an interest in their
history and prehistory. The conference was supported by funding from the University of Auckland
and the University of Otago. Publication of this volume has been supported in part by a Marsden
Fund grant to Peter Sheppard and Richard Walter by the Royal Society of New Zealand. Thanks
also to the many colleagues who acted as referees and provided very prompt turn-around. The
editors would also like to acknowledge the very professional services of Dorothy Brown in copy
editing and Hamish Macdonald for layout and assistance with final editing.
The Editors
vi
LIST OF CONTRIBUTORS
Tony Ahikau Heorake Ian A. Scales

School of Geography Associate, State, Society and Governance in
Faculty of Islands and Oceans Melanesia Project
The University of the South Pacific, Suva, Fiji Research School of Pacific and Asian Studies
S90036810@student.usp.fj The Australian National University
Canberra, Australia
Stuart Bedford
Department of Archaeology and Natural History Peter J. Sheppard
Research School of Pacific and Asian Studies Department of Anthropology,
The Australian National University University of Auckland
Canberra, Australia Private Bag 92019, Auckland, New Zealand
stuart.bedford@anu.edu.au P.Sheppard@auckland.ac.nz
Hallie Buckley Jim Specht
Department of Anatomy & Structural Biology Australian Museum,
University of Otago 6 College Street, Sydney, NSW 2010, and
P.O. Box 913, Dunedin 9054, New Zealand School of Philosophical and Historical Inquiry,
hallie.buckley@stonebow.otago.ac.nz University of Sydney, NSW 2006
jspecht@bigpond.com
Moira Doherty
Department of Anthropology Matthew Spriggs
University of Auckland School of Archaeology and Anthropology
Auckland, New Zealand College of Arts and Social Sciences
The Australian National University
William R. Dickinson Canberra, Australia
Department of Geosciences matthew.spriggs@anu.edu.au
University of Arizona
Tuscon, Arizona 85721, USA Glenn R. Summerhayes
wrdickin@dakotacom.net Anthropology Department, University of Otago
PO Box 56, Dunedin, New Zealand
Amy F. Findlater Glenn.summerhayes@stonebow.otago.ac.nz
Anthropology Department, University of Otago
PO Box 56, Dunedin, New Zealand John E. Terrell
finam126@student.otago.ac.nz Department of Anthropology
Field Museum of Natural History
Roger C. Green 1400 S. Lake Shore Dr.
Department of Anthropology Chicago, IL 60605-2496, USA
University of Auckland terrell@fieldmuseum.org
Private Bag 92019, Auckland, New Zealand
pounamu@ihug.co.nz Tim Thomas
Anthropology Department, University of Otago
Patrick D. Nunn PO Box 56, Dunedin, New Zealand
School of Geography tim.thomas@otago.ac.nz
Faculty of Islands and Oceans
The University of the South Pacific, Suva. Fiji Frédérique Valentin
nunn_p@usp.ac.fj CNRS/Université de Paris I-Paris 10,
21 Allée de l’Université,
Ralph Regenvanu F-92023 Nanterre, France
Vanuatu National Cultural Council frederique.valentin@mae.u-paris10.fr
Vanuatu Cultural Centre
P.O. Box 184, Port Vila, Vanuatu Richard Walter
Ralph.regenvanu@vanuatuculture.org Anthropology Department, University of Otago
PO Box 56, Dunedin, New Zealand
Richard.Walter@otago.ac.nz
vii
Chapter 1
UNDERSTANDING THE WHO, WHAT AND WHY OF LAPITA:

What can Archaeological Ancestors and Descendants Contribute?
Peter J. Sheppard
INTRODUCTION
The stated theme of the 7th Lapita conference was the role of Lapita ancestors and descendants in
our investigation of Lapita. In part this was to allow colleagues who are working outside the Lapita
time period to contribute, but it also reflected the real importance of understanding pre- and post-
Lapita as a means of shedding light on the Lapita archaeological phenomenon. Most of the papers
presented at the conference did not explicitly take up this theme, however underlying many papers
about Lapita are questions of becoming and transforming – origins and endings. In this introduc-
tion I will reflect on this general theme with reference to the papers published in this volume.
As Specht notes in his paper, Lapita has played a central role in the archaeology of the Western
Pacific, possibly at the expense of research and publishing on other periods (see also Bedford and
Spriggs 2008, Walter and Sheppard 2006). Lapita is attractive as a master problem not only because
it is important, but also because it is exotic and mysterious and easily connects to a grand narrative
of travel, exploration and adventure (Terrell 1996). The mystery and exoticism of Lapita stems from
its sudden appearance, marked by a fully developed elaborate ceramic technology and decorative
system, with no obvious immediate precursors and also by the relatively sudden disappearance
of that same system. Why did it start? Why did it end? These are questions which are easily, and
often inappropriately, transformed into: Who where they? Where did they come from? Where did
they go? These mysteries are greatly heightened by the very rapid movement of people carrying
this archaeological tradition out of the Bismarck Archipelago into Remote Oceania, where they
formed the founding population and base cultural tradition in a region covering a tenth of the
earth’s circumference. The sense of adventure is enhanced by the apparent extraordinary feats of
seamanship evidenced by on-going interaction (e.g., obsidian transport) over distances of more
than 2000 km. Ultimately, however, it is the apparent contrast of Lapita with what came before or
after that defines it, and creates the mystery and exoticism. Understanding the “pre” and “post” of
Lapita is an important part of explaining most of the big Lapita questions.
Speaking of “before and after” Lapita raises, of course, the issue of definitions. Traditional
archaeological culture history has never been particularly good at systematics. There are no standard
rules for the creation of archaeological units or the assignment of meaning to those units. Within
Lapita therefore we get: Lapita, decorated Lapita, dentate-stamped Lapita, Lapita without pots,
Lapitoid, Post-Lapita, etc. Processual archaeology has addressed this issue by simply pointing
out that definitions of archaeological units are dependent on the archaeological questions asked.
There are no essential archaeological categories, only definitions relevant to specific theories or
questions. More recently, evolutionary archaeology has built on this notion to argue that our goal
as archaeologists is the establishment of historical lineages, using appropriately defined variables
and datasets (Shennan 2008). We can develop ceramic lineages or lineages of other aspects of
archaeological material culture for which we have adequate samples and suitable variation. The
notion of evolutionary lineage of course calls into question the idea and utility of bounded temporal
units. Evolution creates continuums not bounded categories, and this is perhaps more in keeping
with what anthropologists know about modern and historic cultural and ethnic variation (Moore
1
Lapita: Ancestors and Descendants
1994, Rodseth 1998) . Of course doing culture history without named space/time categories is virtually
impossible and even biologists need such categories in order to talk about their work. But are we
limited then to speaking about specific lineages of material culture categories? Archaeologists have
long recognised coherence in archaeological assemblages, different categories of material culture are
commonly found together over space and time, and most robust culture history definitions are founded
on co-occurrence of a wide variety of items as proposed in the early 20th century by Childe and others.
These become defined as archaeological cultures, although often the “archaeological” gets omitted
from the descriptor and a loose association is then implied with similar, albeit similarly poorly defined,
anthropological categories (see also papers by Terrell and Thomas this volume). Another category
long recognised by archaeological culture historians is that of “tradition” (Phillips and Willey 1953).
This is an explicitly archaeological category as it recognises long-term coherence in archaeological
assemblages often over wide areas. Recently archaeological and anthropological evolutionary
theorists have acknowledged the importance of such coherence. Material culture assemblages are not
necessarily sets of independently varying historical lineages. There is apparent cultural transmission
bias which sees at least some traits evolving and transmitted as packages, leading some evolutionary
theorists to argue for core cultural structures or bauplan which may guide or constrain long-term
variation (e.g., Boyd et al. 1997, O’Brien and Lyman 2005: 106-07, Shennan 2002) – a return to, or
recognition of, I would argue, the reality of the old concept of historical tradition.
How such core cultural structures are formed or transmitted, or the extent to which they exist
or can be observed and explained, is of considerable interest (Shennan and Collard 2005: 133-34). It
seems reasonable to argue that they exist as systems of meaning and practice (Shore 1996) expressed
both through language and the material cultural world (Kirch and Green 2001). This leads then to
considerations of the relationships between material culture and language. There is of course no
necessary relationship between language, culture and biology (see for example Hunley et al. 2008),
although neither are they necessarily independent. In fact we might argue that language and other
types of culture, and to a much lesser extent biology, are more often correlated in prehistory than
not. For archaeologists interested in Lapita, such questions are very important because language
and biology have long formed important parts of our description/explanation of Oceanic culture
history. Much debate has revolved around these issues. Rather than focus narrowly on linguistic
or biological heritage, however, I would like to suggest that we consider how much coherence, as
represented through study of the archaeological record, exists within a Lapita tradition and to what
extent it seems or needs to encompass linguistic or biological heritage. To this end it is important
to understand the evidence provided by ancestors and descendants.
ANCESTORS
What should we mean by “ancestors”? If we mean biological or cultural ancestors then we will
have already answered some of the biggest Lapita questions through our definition. Case closed!
Clearly from a purely archaeological perspective we must be looking at earlier periods within the
spatial extent of Lapita, as what we are examining is change in the archaeological record over
time. This may or may not correlate with biology or long-term cultural ancestry which is poten-
tially marked by language. Ultimately we may be able to establish linkages or hypotheses from
the record into these other areas, but the archaeological question must focus on establishing
historical lineages and examining variation over time in cultural coherence and continuity (see
also Spriggs 2003: 210).
Perhaps the biggest problem we have with this approach is the poor visibility of the
archaeological record in the mid-Holocene. As Specht (this volume) points out, the record in many
areas of the Bismarcks is often poor in this time period. In fact one of the distinguishing charac-
2
Understanding the Who, What and Why of Lapita
teristics of Lapita is the absence of an earlier record at Lapita open site locations, although Kirch
has proposed an early plainware phase for Mussau (Kirch 2001). Only in a few rockshelters do we
have any earlier record stratified below Lapita, and such records are generally impoverished. Most
of our comparative record comes from the Late Pleistocene and early Holocene, where it shows
evidence of obsidian use and transport and potential developments in food production through
translocated species and the exploitation of potential plant domesticates (Specht 2005). Only in West
New Britain do we have any systematic effort, through the work of Torrence, Specht and colleagues
(Torrence 2002), to build up a detailed Holocene record for a region where work is facilitated by
tephra chronology. However the available record is poor and the limited variation in the recovered
lithic assemblages makes it difficult to measure change from the mid to Late Holocene. Only a few
formally distinctive worked stone items of material culture are available to work with. Whether
stemmed tools cease to be made shortly prior to the appearance of Lapita is still debated, although
they appear to fall out of the record in that time range. They together with stylistically distinctive
stone mortars, which are mostly from surface contexts, are presently the best candidates for markers
of the mid-Holocene record in the area of the earliest Lapita record (Torrence and Swadling 2008).
To the southeast of the Bismarcks there is no Late Pleistocene record to work with, although the
evident mobility of Bismarck hunter-gatherers suggests it should be present as does the linguistic
record (Dunn et al. 2005). Similar problems exist in the early and mid–Holocene, and it is only in the
mid-Holocene that we see the first very limited evidence of an archaeological record on Guadalcanal
(see Walter and Sheppard this volume). In fact Vataluma Posovi rockshelter on Guadalcanal is one
of the few places where we have a record which crosses into the Lapita time period (Roe 1993),
yet the recovered material culture gives little basis for studying change or continuity in the record.
The absence of ceramics at this site is of course striking, but calling it or other aceramic sites in the
Lapita time period “Lapita without pots” must be based on positive material evidence.
Moving out into Remote Oceania requires consideration of what the appropriate “ancestral”
record should be. Study of the record indicates Lapita is the initial colonising tradition, although
the near invisibility of the mid-Holocene record in the Bismarcks provides a cautionary note. That
caution aside, however, should the ancestral record be sought in the Solomons immediately to
the west of the Reef/Santa Cruz or in the Bismarcks? Sheppard and Walter (2006) have proposed
that Remote Oceania was settled directly in a leap-frog movement from the Bismarcks (see also
Sheppard and Walter this volume), in which case comparison should be made with archaeological
records in that region. In practice such comparisons using the ceramic record have been regularly
conducted. However the probability of a leap-frog settlement raises the issue of whether there
was one source or many within the Bismarcks. Perhaps such questions will require finer-grained
ceramic analysis and sourcing studies than have been routinely conducted. Thanks to the recent
Teouma burial finds (see Bedford et al. this volume), this is one area where biological data may
inform directly on the archaeological record in the event that ancient DNA can be studied from
these remains. Given the apparent diversity and geographical specificity of modern genetic data
from the Bismarcks (Friedlaender et al. 2007), it may well be possible to identify specific source
regions for Teouma settlers if a large enough sample can be obtained. Although as Hunley et al.
(2008) demonstrate, it will say nothing directly about language relationships. The role of founder
effects may have significant importance in understanding the archaeological, linguistic and biological
records of Remote Oceania.
One area where founder effect may be seen, and a useful bridge built from the archaeological
record to other domains in Remote Oceania, is in language. In Remote Oceania there is a simple
association between the archaeological and linguistic records. There are no Non-Austronesian
(NAN) languages recorded east of Savo in the Central Solomons. Debate over the possible mixed
3
languages of the Reef/Santa Cruz appears to be resolved with Äiwoo now classified as Austronesian
(AN), with its closest relatives in the St Matthias Group (Mussau/Tench) in the Bismarcks (Ross
and Næss 2007). Careful examination of the Reef/Santa Cruz archaeological record from the Lapita
period into the historic by Doherty (2007; see also this volume) has also failed to find any evidence
of radical disjunctures which might signal introductions of new cultural traditions consistent with
post-Lapita NAN introductions. There is some debate over possible NAN admixture explaining the
“aberrant” nature of languages of Vanuatu/New Caledonia (Blust 2005, 2008, Donohue and Denham
2008, Lynch 1999, Pawley 2006). However, if we have direct settlement from the Bismarcks into
Remote Oceania, the NAN/AN admixture or sampling of both language families may have occurred
in the Bismarcks and then been transported out into Remote Oceania (see also Pawley 2006: 248).
Although Blust (2005) discounts the possibility of the Lapita tradition being shared across these
language families in Near Oceania, sharing of cultural traditions in this manner is common in Near
Oceania. A good example is the sharing of the head-hunting, shrine and shell valuable complex,
which sits at the core of cultural life in the Western Solomons, across AN and NAN speaking groups
(see Thomas; Walter and Sheppard this volume). At least in linguistic terms this appears to be a
case of strong AN influence on NAN neighbours in the context of head-hunting. Our study of the
archaeology of this region shows that it is very difficult to differentiate the AN and NAN records
in the absence of prior knowledge of linguistic differences (Sheppard and Walter 2008, Sheppard,
Walter and Roga in press).
DESCENDANTS
What can “descendants” tell us about Lapita? In general terms study of more recent periods may
provide us with the ability to investigate, in data rich contexts, hypotheses of potential interest for
explaining the Lapita phenomenon. For example, as noted above in the Western Solomons, it is
perhaps only in very recent periods that we can control for linguistic or biological heritage and
see if we can discriminate NAN from AN archaeological records. From this we may also gain an
appreciation of what is or is not a useful archaeological question.
Study of the record across the Late Holocene allows us to make a number of observations
which reflect back onto Lapita. These include the date of the disappearance of decorated ceramics
and then finally, in most areas, the complete disappearance of ceramics. The extent to which these
events are contemporaneous or the sequence of ceramic simplification may similarly shed light on
the mechanism or forces which developed and maintained the elaborate Lapita ceramic system.
Contemporaneity or close similarity in this process suggests an underlying common dynamic. Why
does the ceramic record end in some areas and not in others? Can we see any similarities in the
areas where it ends or transforms? Looking at the broader record, can we see systematic decline
in cultural coherence as measured by the complete set of material culture and archaeologically
observable behaviour? Or do we see new drivers pushing change in some aspects of the record and
not in others? Such questions inform not only on new conditions, but reflect on prior states during
the Lapita time period (see also Spriggs 2004).
Unfortunately, as with the pre-Lapita record, we generally have a poor understanding in most
areas, of the period immediately following Lapita. As Spriggs (1997: 162) noted a decade ago, the
disappearance of pottery greatly reduces archaeological visibility in many areas. In the Bismarcks
region we have a limited record (see review by Lilley 2004, Summerhayes 2007), although Wickler
(2001) and Specht’s (1969) work in the Northern Solomons has been augmented recently by work
on Tanga (Garling 2003, 2007) to the north. The incised and applied relief found in the New Ireland
region has clear links to the ceramic tradition which extends out from the late Lapita period in
the Western Solomons (Felgate 2003, Sheppard and Walter 2006) (see Sheppard and Walter this
4
volume). As with the pre-Lapita period, the work on West New Britain by Torrence, Specht and
colleagues helps to build up a continuous sequence (Specht and Torrence 2007) and the publication
of later sequences from Watom provides coverage across the Late Lapita chronological boundary
(Anson 2000).
In Remote Oceania the situation has markedly improved since Spriggs’ 1997 review. The
post-Lapita sequence from the Reef/Santa Cruz has been the subject of a thesis by Doherty (2007,
see also Doherty as well as Green and Yen this volume), which looks closely at variation in the
available record from the Lapita to historic periods. Most recently Leach and Davidson (2008) have
published a monograph on their work on Taumako just to the east of the Reef/Santa Cruz, providing
a detailed sequence for that region. In Vanuatu the work of Bedford and Spriggs has specifically
focused on improving understanding of the post-Lapita period, with early work on the relationship
of the Mangaasi ceramic style to Lapita (Bedford 2006, Bedford and Clark 2001, Bedford and
Spriggs 2008). Similarly work in Fiji improves both our knowledge of the Lapita (see Nunn and
Horeke this volume) and later records (Burley 2005, Clark 1999), although Best’s (1984) work
on Lakeba remains one of the best sequences from the region. The New Caledonia (Sand 1995,
2001a, 2002) record continues to be built up with improved understanding of both the chronology
of the end of dentate Lapita (Sand 1997) and its relationship to the following ceramic sequences,
as well as considerable focus on the comparative study of Lapita ceramic sequences (Sand 2001b,
Sand 2007). Study of the Lapita to post-Lapita record in Island Melanesia has focused primarily on
the significance of the incised and relief ceramic “traditions”, which often seem to follow Lapita
(Bedford and Clark 2001, Clark 1999, Cochrane 2004, Wahome 1997). This represents one of the
few areas of systematic study of artefact lineages across the Lapita “boundary”, although the focus is
primarily on issues of postulated post-Lapita migration (Spriggs 2004), rather than on what it might
say about Lapita. The difficulty of doing this is demonstrated by the problems Clark and Murray
(2006: 115) had while trying to create useful data to look at decay in Lapita design systems within
the region. Cochrane’s (2004) study of ceramic lineages in the Yasawa Islands of Fiji is an explicit
attempt to apply cladistic approaches to analysis and, as such, it provides a methodological guide
for further work; although once again it reinforces the need to develop primary data or to work with
standard databases. The quality of old data is a common complaint in the recent literature.
To the east of Fiji, investigation of the Lapita/post-Lapita record assumes a new focus around
the origins of Polynesians or Ancestral Polynesian culture (Kirch and Green 1987, 2001, Smith
2002). Work in Samoa (Dickinson and Green 1998, Green 2002, Rieth, Morrison, and Addison
2008) has recently targeted the evaluation of the extent of Lapita occupation and the role of
geotectonics and sea level changes on the visibility of the record. In addition Rieth and Hunt (2008)
have reviewed the radiocarbon chronology of Samoa as a basic step in evaluation of change over
time in that record. The situation in Tonga has improved dramatically over the last decade and it
has, thanks to the systematic work by Dave Burley and his colleagues (Burley and Connaughton
2007, Burley and Dickinson 2001, Dickinson and Burley 2007), perhaps the best record available
for looking at transformation from colonisation through the Polynesian Plainware period (Burley
2007, Connaughton 2007).
CONCLUSIONS
Archaeologists do not dig up “cultures” or ethnicities or peoples; we create archaeological records
from the remains we find and also from those we do not find, although that is always more
problematic (e.g. Clark and Bedford, 2007). Conventional culture history has neither rules nor
standard methodologies for creating archaeological “cultures”. For that matter, neither did
Anthropology have rules for defining “cultures”. Researchers in both areas however observed
5
constellations of features which cohered over time and space in the records available to them.
Such constellations exist. For archaeologists, temporal variation in such records required the
creation of time/space divisions loosely based on these observed constellations, if only to enable
conversations about the record. Problems arise when these defined units take on an independent
existence and inferences are casually made beyond what the record can support. In Pacific prehis-
tory much debate has occurred over the relationship between the archaeological record, language
and biology (see Terrell this volume). In practice, of course, most archaeologists have a very good
appreciation of the difficulties of inferring such relationships (Green 2003), and problems arise
most often when colleagues in other disciplines look for cultures, ethnicities or peoples in our
units. We should clearly acknowledge that we have no archaeological basis for finding such
things in our data and that in fact such goals do not provide useful archaeological questions. There
is no point in arguing whether Lapita in Near Oceania spoke only AN or NAN or both. We will
never know and linguists will never be able to tell us. On the other hand, there is no point in
ignoring data or hypotheses from other disciplines on those occasions where we can make
effective linkages. The association between Lapita and AN in Remote Oceania would appear to
be a case where a strong argument can be made, although we should not forget that NAN has
apparently disappeared without much trace over a sizeable portion of Near Oceania. Even in the
event that some early colonists did speak NAN (Donohue and Denham 2008) they appear to have
rapidly adopted an AN linguistic tradition. The question remains however of the extent to which
our knowledge of this association in Remote Oceania alters our study of the archaeological
record. What difference would it make if we didn’t know this?
As I have noted in this introduction, modern approaches to culture history are predominately
associated with evolutionary theory of various forms (Kirch and Green 1987, Shennan 2008).
Although this is a diverse theoretical approach, at least some of these theorists suggest that core
cultural structures create selection bias, which sees groups of traits preferentially transmitted over
long periods. I have suggested that this is recognition of the long-term coherence of features which
old style culture historians defined as archaeological traditions. One area where evidence of such
core structures or sets of related meanings might be found is in language. Linguistics might provide
hypotheses about core traditions of use to archaeologists in areas such as Remote Oceania, where
we can control for linguistic heritage (Kirch and Green 2001).
Evolutionary theory and methods are no panacea, although modern approaches turn our
attention from bounded categories to continuums and require systematic methodologies. Applying
cladistics to poor datasets is not going to advance research and may blind us with science. All
archaeological advances require improvements in chronology and sampling, as well as better data
analysis. Material culture studies require detailed knowledge of the relationships among variables
and the factors underlying variation (see Donohue, Wichmann and Albu 2008, Dunn et al. 2007 for
similar issues debated by linguists regarding the utility of evolutionary methods) and systematic
analysis of large, purpose built datasets. A survey of current literature suggests that such problems
are widely recognised and are being addressed. There may of course be no answer to some problems;
data in some areas may never be adequate – in which case we will need to revise our goals. However,
the creation of better data when used to investigate appropriate questions should see interesting
developments in the field of Lapita archaeology; especially if we recognise that what comes before
and after the constellation of features which we call Lapita, has an important role to play in the
archaeological investigation of the Pacific.
6
REFERENCES
Anson, D., 2000. Excavations at Vunavaung (SDI), Rakival Village, Watom Island, Papua New Guinea.
New Zealand Journal of Archaeology, 20 [1998]: 95-118.
Bedford, S., 2006. Pieces of the Vanuatu Puzzle: Archaeology of the North, South and Centre. Terra
Australis 23. Canberra: Pandanus Books, Research School of Pacific and Asian Studies, Australian
National University.
Bedford, S. and G. Clark, 2001. The rise and rise of the incised and applied relief tradition: A review
and reassessment. In G.R. Clark, A.J. Anderson and T. Vunidilo (eds), The Archaeology of Lapita
Dispersal in Oceania: Papers from the Fourth Lapita Conference, June 2000, Canberra, Australia.
Terra Australis 17. Canberra: Pandanus Books, pp. 61-74.
Bedford, S. and M. Spriggs, 2008. Northern Vanuatu as a Pacific crossroads: The archaeology of discovery,
interaction, and the emergence of the “ethnographic present”. Asian Perspectives, 47: 95-120.
Best, S., 1984. Labeka: The Prehistory of a Fijian Island. Unpublished Ph.D. thesis, University of
Auckland, Auckland.
Blust, R., 2005. Review of Lynch, Ross and Crowley 2002. Oceanic Linguistics, 44: 544-58.
Blust, R., 2008. Remote Melanesia: One history or two? An addendum to Donohue and Denham. Oceanic
Linguistics, 42: 445-59.
Boyd, R., W.H. Durham, M. Mulder and P.J. Richerson, 1997. Are cultural phylogenies possible? In P.
Weingart, P. Richerson, S. Mitchell and S. Massen (eds), Human by Nature: Between Biology and
Social Science. London: Lawrence Erlbaum Associates, Publishers, pp. 355-86.
Burley, D.V., 2005. Mid-sequence archaeology at the Sigatoka sand dunes with interpretive implications for
Fijian and Oceanic culture history. Asian Perspectives, 44: 320-48.
Burley, D.V., 2007. In search of Lapita and Polynesian plainware settlements in Vava‘u, Kingdom of
Tonga. In S. Bedford, C. Sand and S.P. Connaughton (eds), Oceanic Populations: Lapita and
Western Pacific Settlements. Terra Australis 26. Canberra: ANU E Press, pp. 187-98.
Burley, D.V. and S.P. Connaughton, 2007. First Lapita settlement and its chronology in Vava‘u, Kingdom
of Tonga. Radiocarbon, 49: 131-37.
Burley, D.V. and W.R. Dickinson, 2001. Origin and significance of a founding settlement in Polynesia.
Proceedings of the National Academy of Sciences of the United States of America, 98: 11829–31.
Clark, G.R., 1999. Post-Lapita Fiji: Cultural Transformations in the Mid-Sequence. Unpublished Ph.D.
thesis, Australian National University, Canberra.
Clark, G. and S. Bedford, 2007. Friction zones in Lapita colonisation. In G. Clark, F. Leach and S.
O’Connor (eds), Island of Inquiry: Colonisation, Seafaring and the Archaeology of Maritime
Landscapes. Terra Australis 29. Canberra: ANU E Press, pp. 59-73.
Clark, G. and T. Murray, 2006. Decay characteristics of the Eastern Lapita design system. Archaeology in
Oceania, 41: 107-17.
Cochrane, E., 2004. Explaining Cultural Diversity in Ancient Fiji: The Transmission of Ceramic Variability.
Unpublished Ph.D. dissertation, University of Hawaii, Honolulu.
Connaughton, S.P., 2007. Can we dig it? Archaeology of ancestral Polynesian society in Tonga: A first look
from Falevai. In S. Bedford, C. Sand and S.P. Connaughton (eds), Oceanic Explorations: Lapita and
Western Pacific Settlement. Terra Australis 26. Canberra: ANU E Press, pp. 199-212.
Dickinson, W.R. and D.V. Burley, 2007. Geoarchaeology of Tonga: Geotectonic and geomorphic controls.
Geoarchaeology, 22 (2): 229-59.
Dickinson, W.R. and R.C. Green. 1998. Geoarchaeological context of Holocene subsidence at the Ferry
Berth Lapita site, Mulifanua, Upolu, Samoa. Geoarchaeology, 13 (3): 239-63.
7
Doherty, M.W., 2007. Post-Lapita Developments in the Reef-Santa Cruz Islands, Southeast Solomon
Islands. Unpublished Ph.D. thesis, University of Auckland, Auckland.
Donohue, M. and T. Denham. 2008. The language of Lapita: Vanuatu and an Early Papuan presence in the
Pacific. Oceanic Linguistics, 47: 433-44.
Donohue, M., S. Wichmann and M. Albu, 2008. Typology, areality, and diffusion. Oceanic Linguistics,
47: 223-32.
Dunn, M., R. Foley, S. Levinson, G. Reesink and A. Terrill. 2007. Statistical reasoning in the evaluation of
typological diversity in Island Melanesia. Oceanic Linguistics, 46: 388-403.
Dunn, M., A. Terrill, G. Reesink, R. Foley and S. Levinson, 2005. Structural phylogenetics and the
reconstruction of ancient language history. Science, 309: 2072-75.
Felgate, M.W., 2003. Reading Lapita in Near Oceania: Intertidal Shallow-Water Pottery Scatters, Roviana
Lagoon, New Georgia, Solomon Islands. Unpublished Ph.D. thesis, University of Auckland, Auckland.
Friedlaender, J.S., F.R. Friedlaender, J. Hodgson, S. McGrath, G. Stoltz, T. Schurr and A. Merriwether,
2007. Mitochondrial DNA variation in northern Island Melanesia. In J.S. Friedlaender (ed.), Genes,
Language and Culture History in the Southwest Pacific. Oxford: Oxford University Press, pp. 61-80.
Garling, S., 2003. Tanga takes to the stage: Another model ‘transitional’ site? New evidence and a
contribution to the ‘incised and applied relief tradition’ in New Ireland. In C. Sand (ed.), Pacific
Archaeology: Assessments and Prospects. Le Cahiers de l’Archéologie en Nouvelle-Calédonie 15.
Nouméa: Service de Musées et du Patrimoine de Nouvelle-Calédonie, pp. 213-33.
Garling, S., 2007. Post-Lapita Evolutions or Revolutions? Interaction and Exchange in Island Melanesia:
The View from the Tanga Islands. Unpublished Ph.D. thesis, Australian National University, Canberra.
Green, R.C., 2002. A retrospective view of settlement pattern studies in Samoa. In T. N. Ladefoged and
M. W. Graves (eds), Pacific Landscapes: Archaeological Approaches. Los Osos: Easter Island
Foundation, pp. 125-52.
Green, R.C., 2003. The Lapita horizon and traditions - signature for one set of Oceanic migrations. In
C. Sand (ed.), Pacific Archaeology: Assessments and Prospects. Le Cahiers de l’Archéologie en
Nouvelle-Calédonie 15. Nouméa: Service des Musées et du Patrimoine de Nouvelle-Calédonie, pp.
95-120.
Hunley, K., M. Dunn, E. Lindstrom, G. Reesink, A. Terrill, M.E. Healy, G. Koki, F.R. Friedlaender and J.S.
Friedlaender, 2008. Genetic and linguistic coevolution in northern Island Melanesia. Plos Genetics,
4: e1000239.
Kirch, P.V. (ed.), 2001. Lapita and its Transformations in Near Oceania: Archaeological Investigations in
the Mussau Islands, Papua New Guinea, 1985-1988. Contributions of the University of California
Archaeological Research Facility No. 59. Berkeley: University of California.
Kirch, P.V. and R.C. Green, 1987. History, phylogeny, and evolution in Polynesia. Current Anthropology,
28: 431-56.
Kirch, P.V. and R.C. Green, 2001. Hawaiki, Ancestral Polynesia: An Essay in Historical Anthropology.
Cambridge: Cambridge University Press.
Leach, F. and J. Davidson, 2008. Archaeology on Taumako: A Polynesian Outlier in the Eastern Solomon
Islands. New Zealand Archaeological Association Special Publication. Dunedin.
Lilley, I., 2004. Trade and culture history across the Vitiaz Strait, Papua New Guinea: The emerging
Post-Lapita coastal sequence. In V.J. Attenbrow and R. Fullagar (eds), Pacific Odyssey: Archaeology
and Anthropology in the Western Pacific: Papers in Honour of Jim Specht. Records of the Australian
Museum, Supplement 29. Sydney: The Australian Museum, pp. 89-96.
Lynch, J., 1999. Language change in Southern Melanesia: Linguistic aberrancy and genetic distance. In
R. Blench and M. Spriggs (eds), Archaeology and Language IV: Language Change and Cultural
Transformation. London: Routledge, pp. 149-59.
8
Moore, J.H., 1994. Putting anthropology back together again: The ethnogenetic critique of cladistic theory.
American Anthropologist, 96: 925-48.
O’Brien, M.J. and R.L. Lyman, 2005. Cultural phylogenetic hypotheses in archaeology: Some fundamental
issues. In R. Mace, C.J. Holden and S. Shennan (eds), The Evolution of Cultural Diversity: A
Phylogenetic Approach. London: UCL Press, pp. 85-108.
Pawley, A., 2006. Explaining the aberrant Austronesian langauges of Southeast Melanesia: 150 years of
debate. Journal of the Polynesian Society, 115: 215-58.
Phillips, P. and G.R. Willey, 1953. Method and theory in American archaeology: An operational basis for
culture-historical integration. American Anthropologist, 55: 615-33.
Rieth, T.M. and T.L. Hunt, 2008. A radiocarbon chronology for Samoan prehistory. Journal of
Archaeological Science, 35: 1901-27.
Rieth, T.M., A.E. Morrison and D.J. Addison, 2008. The temporal and spatial patterning of the initial
settlement of Samoa. Journal of Island and Coastal Archaeology, 3: 214-39.
Rodseth, L., 1998. Distributive models of culture - A sapirian alternative to essentialism. American
Anthropologist, 100: 55-69.
Roe, D., 1993. Prehistory Without Pots: Prehistoric Settlement and Economy of Northwest Guadalcanal,
Solomon Islands. Unpublished Ph.D. thesis, Australian National University, Canberra.
Ross, M. and Å. Næss, 2007. An Oceanic origin for Äiwoo, the language of the Reef Islands? Oceanic
Linguistics, 46 (2): 456-98.
Sand, C., 1995. Le Temps D’Avant: La Préhistoire de la Nouvelle-Calédonie. Paris: L’Harmattan.
Sand, C., 1997. The chronology of Lapita ware in New Caledonia. Antiquity, 71: 539-47.
Sand, C., 2001a. Changes in non-ceramic artefacts during the prehistory of New Caledonia. In G.R. Clark,
A.J. Anderson and T. Vunidilo (eds), The Archaeology of Lapita Dispersal in Oceania: Papers
from the Fourth Lapita Conference, June 2000, Canberra, Australia. Terra Australis 17. Canberra:
Pandanus Books, pp. 75-92.
Sand, C., 2001b. Evolutions in the Lapita cultural complex: A view from the southern Lapita province.
Archaeology in Oceania, 36: 65-76.
Sand, C., 2002. Creations and transformations of prehistoric landscapes in New Caledonia, the
southernmost Melanesian islands. In T. Ladgefoged and M. Graves (eds), Pacific Landscapes:
Archaeological Approaches. Los Osos: Easter Island Foundation, pp. 13-34.
Sand, C., 2007. Looking at the big motifs: A typology of the central band decorations of the Lapita ceramic
tradition of New Caledonia (Southern Melanesia) and preliminary regional comparisons. In S.
Bedford, C. Sand and S.P. Connaughton (eds), Oceanic Explorations: Lapita and Western Pacific
Settlement. Terra Australis 26. Canberra: ANU E Press, pp. 265-87.
Shennan, S., 2002. Genes, Memes, and Human History: Darwinian Archaeology and Cultural Evolution.
London: Thames & Hudson.
Shennan, S., 2008. Evolution in archaeology. Annual Review of Anthropology, 37: 75-91.
Shennan, S. and M. Collard, 2005. Investigating processes of cultural evolution on the north coast of New
Guinea with multivariate and cladistic analysis. In R. Mace, C.J. Holden and S. Shennan (eds), The
Evolution of Cultural Diversity: A Phylogenetic Approach. London: UCL Press, pp. 134-64.
Sheppard, P.J. and R. Walter, 2006. A revised model of Solomon Islands culture history. Journal of the
Polynesian Society, 115: 47-76.
Sheppard, P.J. and R. Walter, 2008. The sea is not land: Comments on the archaeology of islands in the
Western Solomons. In J. Connolly and M. Campbell (eds), Comparative Island Archaeologies.
Oxford: BAR International Series, pp. 167-78.
9
Sheppard, P.J., R. Walter and K. Roga, in press. Friends, relatives, and enemies: The archaeology and
history of interaction among Austronesian and NAN speakers in the Western Solomons. In J.
Bowden and N. Himmelmann (eds), Fetschrift for Andrew Pawley on NAN and Austronesian
Languages. Canberra: Pacific Linguistics, Research School of Pacific and Asian Studies, Australian
Shore, B., 1996. Culture in Mind: Cognition, Culture and the Problem of Meaning. Oxford: Oxford
University Press.
Smith, A., 2002. An Archaeology of West Polynesian Prehistory. Terra Australis 18. Canberra: Pandanus Books.
Specht, J., 1969. Prehistoric and Modern Pottery Industries of Buka Island, T.P.N.G. Unpublished Ph.D.
Specht, J., 2005. Revisiting the Bismarcks: Some alternative views. In A. Pawley, R. Attenborough, J.
Golson and R. Hide (eds), Papuan Pasts: Cultural, Linguistic and Biological Histories of Papuan-
Speaking People. Pacific Linguistics 572. Canberra: Australian National University, pp. 235-88.
Specht, J. and R. Torrence, 2007. Lapita all over: Land-use on the Willaumez Peninsula, Papua New
Guinea. In S. Bedford, C. Sand, and S.P. Connaughton (eds), Oceanic Explorations: Lapita and
Spriggs, M., 1997. The Island Melanesians. Oxford: Blackwell.
Spriggs, M., 2003. Post-Lapita evolutions in Island Melanesia. In C. Sand (ed.), Pacific Archaeology:
Assessments and Prospects. Le Cahiers de l’Archéologie en Nouvelle-Calédonie 15. Nouméa:
Service des Musées et du Patrimoine de Nouvelle-Calédonie, pp. 205-12.
Spriggs, M., 2004. Is there life after Lapita, and do you remember the 60’s? The post-Lapita sequences of
the Western Pacific. In V.J. Attenbrow and R. Fullagar (eds), A Pacific Odyssey: Archaeology and
Anthropology in the Western Pacific. Papers in Honour of Jim Specht. Records of the Australian
Summerhayes, G.R., 2007. The rise and transformations of Lapita in the Bismarck Archipelago. In S.
Chiu and C. Sand (eds), From Southeast Asia to the Pacific: Archaeological Perspectives on the
Austronesian Expansion and the Lapita Cultural Complex. Taipei: Academia Sinica, pp. 141-69.
Terrell, J.E., 1996. Lapita as history and culture hero. In Janet Davidson, Geoffrey Irwin, Foss Leach,
Andrew Pawley and Dorothy Brown (eds), Oceanic Culture History: Essays in Honour of Roger
Green. Dunedin: New Zealand Journal of Archaeology Special Publication, pp. 51-66.
Torrence, R., 2002. Cultural landscapes on Garua island, Papua New Guinea. Antiquity, 76 [293]: 766-76.
Torrence, R. and P. Swadling, 2008. Social networks and the spread of Lapita. Antiquity, 82: 600-16.
Wahome, E.W., 1997. Continuity and change in Lapita and post-Lapita ceramics: a review of evidence from
the Admiralty Islands and New Ireland, Papua New Guinea. Archaeology in Oceania, 32: 118-23.
Walter, R. and P. Sheppard, 2006. Archaeology in Melanesia: A case study from the Western Province of
the Solomon Islands. In I. Lilley (ed.), Archaeology of Oceania: Australia and the Pacific Islands.
Oxford: Blackwell, pp. 137-59.
Wickler, S., 2001. The Prehistory of Buka: A Stepping Stone Island in the Northern Solomons. Terra
Australis 16. Canberra: Department of Archaeology and Natural History and Centre for
Archaeological Research, Australian National University.
10
Chapter 2
THE ACERAMIC TO CERAMIC BOUNDARY

IN THE BISMARCK ARCHIPELAGO
Jim Specht
INTRODUCTION
Summing up the Lapita Homeland Project conducted in the Bismarck Archipelago of Papua New
Guinea in 1985, Jim Allen expressed disappointment at “our inability to find sites which immedi-
ately predate the appearance of Lapita in the Bismarcks” (Allen 1991: 7).1 Spriggs (1997: 73-82;
cf. Allen 2000: 159-63) subsequently reviewed the archaeological evidence from the archipelago
and revealed a more encouraging picture, with aceramic cultural evidence stratified below
dentate-stamped pottery at several localities, though “Lapita sites do not generally reoccupy
previously used locations apart from where Lapita deposits occur in rock shelters” (Spriggs 1997:
88). The specific issue raised by Allen, however, related to contexts that “immediately predate”
the introduction of pottery, and the question remains whether the sites discussed by Spriggs
satisfy Allen’s condition.
This is not a trivial point, as it is central to the much-debated issue of the origins of Lapita
pottery and its posited associated cultural complex. However construed conceptually or materially,
the “Lapita phenomenon” has dominated discussion of the human past in the Western Pacific for
several decades, reflecting competing models of human history in insular Southeast Asia and Near
Oceania that contrast “revolutionary” versus “evolutionary” interpretations of the data. In the
revolutionary version, “Lapita” is firmly embedded in what Spriggs (2007: 104) terms “the master
narrative of Austronesian expansion” from insular Southeast Asia into the northern reaches of Near
Oceania, and the Bismarck Archipelago in particular. The arrival of these Austronesian-speaking
seaborn migrants in the archipelago signalled dramatic changes in regional cultural scenarios. Using
one or other version of Green’s “Triple-I” model (Green 1991, 2000, 2003), the “Lapita cultural
complex” is viewed as emerging from the amalgamation of cultural practices introduced from insular
Southeast Asia with local practices, together with various innovations (e.g., Bellwood 1995, Kirch
1995, 1996, 1997, Spriggs 1997, 2003). In this model the cultural “world” of the archipelago was
“turned upside down” (Spriggs 1997: Chapter 4).
Contrary views take an evolutionary perspective in which migration is played down in favour
of more significant roles for locally generated cultural change and innovation among the Papuan-
speaking communities of the archipelago, with some versions rejecting the need to invoke migration
as the primary causative factor in change (e.g., Allen and White 1989, Szabó and O’Connor 2004,
Terrell 1998, Terrell, Kelly and Rainbird 2001). In this scenario long-term interaction between
communities in the archipelago and those to the west and northwest provided the conduits for
mutual exchange of innovations and change, perhaps extending back to the mid-Holocene (e.g.,
Allen and Gosden 1996, Terrell 1998). The developments that eventually led to the “Lapita cultural
complex” were thus part of a long-term process.
Neither group of contestants can claim the high ground without considering in closer detail
the material evidence from the archipelago, especially the relationship between contexts containing
Lapita pottery and those that “immediately predate” them, and the apparent lack of continuity of site
use, especially at open sites. More than a decade has passed since Spriggs (1996, cf. 1997: 73-90)
presented details of the geographical and temporal distributions of various cultural elements in insular
11
Southeast Asia and Near Oceania in support of the migration model. Since then, little attention has
been paid specifically to the issue raised by Allen, and in particular to the aceramic/ceramic boundary
in the Bismarck Archipelago. In some situations the relevant assemblages are separated by several
centuries, possibly even a millennium, and this gap begs questions rather than supplies answers. Where,
if at all, can we map the transition from aceramic to ceramic material culture in this region?
Here I review the evidence for this transition in the Bismarck Archipelago, with a primary focus
on New Britain (Figure 2.1). While there is ample evidence throughout the archipelago for human
activity before the appearance of Lapita pottery, there is currently little evidence for continuity of site
use and, by extension, of populations across the aceramic/ceramic boundary. Indeed, in some cases
there is a clear hiatus, but there are situations (as noted by Spriggs) where continuity seems possible.
A commonsense explanation for the absence of a clear picture might invoke insufficient data as a
result of the lack or scarcity of fieldwork in some areas, where a focus on pottery-period sites has not
been matched by concern for pre-pottery contexts. The situation is also complicated by mid- and late
Holocene sea level fluctuations, particularly during the Lapita period when sea level was still about
1 m higher than at present (Chappell et al. 1996; cf. Kirch 1988, 2001a: 130-34), and evidence for
pre-pottery levels on old shorelines has survived in only a few cases (Gosden and Webb 1994: 34).
In the following discussions I take the starting date for dentate-stamped pottery as around
3450-3350 cal. B.P. (Kirch 2001b: 219, Specht 2007: Table 1, Spriggs 2003: 63). Radiocarbon
dates are calibrated here using the CALIB 5.0.1 suite of programs (Stuiver and Reimer 1993, as
revised), with age ranges expressed at 2σ. The default value for Delta-R of 0±0 years is used for
Figure 2.1 Map of the Bismarck Archipelago, Papua New Guinea, showing the main areas and sites
mentioned in the text (based on Gosden and Webb 1994). 1: Willaumez Peninsula and Garua
Island; 2: Watom Island; 3: Duke of York Islands; 4: Arawe Islands; 5: Kandrian; 6: Mussau; 7:
Kohin; 8: Mouk; 9: Pamwak; 10: Panakiwuk; 11: Lossu; 12: Pinikindu; 13: Lasigi, Buang
Merabak; 14: Anir; 15: Lamau; 16: Balof; 17: Pomio.
12
Table 2.1 List of radiocarbon dates cited in the text. Calibrations were carried out with the CALIB 5.0.1
programs, using the default Delta-R value of 0 ± 0 years for marine samples. Age ranges are at
2σ. For published sources, see the main text.
The Aceramic to Ceramic Boundary in The Bismarck Archipelago
continued next page

14
samples of marine origin. Several Delta-R values have been proposed for the archipelago (Kirch
2001b: 199-204, Petchey et al. 2005, Petchey, Phelan and White 2004, Summerhayes 2007: 154,
endnote 3), but these are locality-specific and do not have wide application. The default value yields
results differing slightly from those with locally determined values, but it simplifies comparisons
across the region. Within the text, age ranges are rounded to the nearest 10-year interval, with
ranges ending with -5 rounded upwards. 2
PRE-LAPITA ARCHAEOLOGICAL CONTEXTS WITHIN NEW BRITAIN
Geological Contexts
Before addressing the archaeological data I provide a brief geological introduction to the island
in general, as this is critical for understanding the evidence (cf. Torrence and Doelman 2007).
New Britain is located on a subduction zone formed by the collision of the South Bismarck and
Solomon Sea Plates (Johnson 1976, Tregoning et al. 2000). As a consequence, the island is
subject to episodic volcanism, and tectonic and seismic activity. In the late Quaternary these
factors combined with hydro-isostatic variations in sea level and erosion processes to create the
island’s present-day coasts, which can be loosely treated as two main zones separated for much
of the island’s length by the Whiteman, Nakanai and Baining mountain ranges that are formed by
older igneous and sedimentary rocks.
The north coast of New Britain has numerous areas of Quaternary volcanism forming part of
the Bismarck Volcanic Arc (Smith and Johnson 1981). There are at least seven active or dormant
volcanic centres, but detailed eruptive histories are available for only Rabaul at the eastern end of
New Britain, Witori to the south of Cape Hoskins, and Dakataua at the northern end of Willaumez
Peninsula. Since human arrival in New Britain about 40,000 years ago (Torrence et al. 2004), the
Rabaul area has witnessed at least 27 eruptions (Nairn et al. 1995: Table 2). In the Holocene there
have been 12 events at Witori and at least seven at Dakataua (Machida et al. 1996: Table 1). The
magnitude and impact of these eruptions ranged from minor events with restricted local impacts,
to major cataclysmic events with significant airfall and pyroclastic flow deposits that affected large
areas, including the Passismanua area on the south side of the Whiteman Range.3 Several eruptions
caused extensive landscape devastation and destruction, and depopulation through annihilation or
enforced abandonment of areas close to the volcanic centres (Petrie and Torrence 2008, Torrence and
Doelman 2007). Equally important would have been the loss of subsistence resources, including the
coral reef resources, even in areas beyond the most serious impacts of the eruptions (Boyd, Lentfer
and Luker 1999; cf. Pandolfi et al. 2006 for Huon Peninsula). On the positive side, tephras from
these events often enriched soils and extended the area of habitable land (Torrence and Doelman
2007), and these soils currently support some of the densest concentrations of population (25-60
persons/km2) in West New Britain (data in Bourke et al. 1996). Soil profiles in areas close to the
Witori and Dakataua volcanoes show a sequence of in situ and reworked volcanic deposits separated
by soil formation, providing archaeologists with a tephrostratigraphic sequence in which successive
periods of human activity are clearly separated.
The south coast, in contrast, is dominated by limestone formed by coral reefs uplifted during
the Pleistocene and Holocene to create coastal cliffs locally reaching 80-150 m in height, with
elevated reef flat terraces that vary in number and extent along the coast. In Jacquinot Bay six
terraces extend several kilometres inland, the youngest of them dating to early- to mid-Holocene
time (Riker-Coleman et al. 2006). No comparable studies have been carried out in the Kandrian
area, where near-vertical cliffs stand tens of metres above narrow coastal plains. A minor uplift event
15
of about 2-3 m occurred here within the last 2000 years (Boyd, Specht and Webb 1999: 286), but
apparently did not affect the Arawe Islands 50-60 kms further west (Gosden and Webb 1994: 33).
Interspersed between the stretches of uplifted reef limestone are river estuaries of varying extent,
and often associated with low, emerged reef platforms of mid-Holocene or later age, covered with
swamp forest and mangroves.
Lying beyond or at the distal end of tephra distributions from the northern volcanic centres,
the south coast generally did not suffer the severe devastation and depopulation postulated for
areas close to the source volcanoes. Indeed, parts of the south coast might have been refuge zones
for those fleeing from the worst impacted areas (Boyd, Lentfer and Luker 1999: Figs 2-6). Tephra
possibly from the W-K2 eruption has been reported at Lolmo Cave in the Arawe Islands (Gosden
et al. 1994: 106), and various Witori and Dakataua tephras are reported in the Kandrian area, along
with several currently unidentified tephras (John Webb and Peter Jackson, La Trobe University, pers.
comm.). The south coast tephras are thin and poorly preserved, and there is no tephrostratigraphic
sequence comparable to that of the north coast. The periodic deposition of tephras would locally
have improved soil quality, but with decreasing effect with increasing distance from the source
vents. Today, most of the south coast has population densities less than ~20 persons/km2, though
the Arawe Islands have 97 persons/km2 (from data in Bourke et al. 1996). This is the second highest
density in the province after the Bali-Witu islands (about 110 persons/km2), and no doubt reflects
the high bio-productivity of the area’s extensive reefs and swamps.
The Archaeological Evidence – Introduction

The island is divided administratively into two provinces, East and West New Britain. With the
exception of a brief visit to Pomio during the Lapita Homeland Project reconnaissance of 1984,
archaeological fieldwork in East New Britain has been confined to Watom Island and the Duke of
York Islands, with a specific focus on dentate-stamped pottery sites (Anson, Walter and Green 2005,
Green and Anson 2000, Specht 1968, White 2007). It is not surprising, then, that pre-Lapita period
deposits are not known in this province. In contrast, West New Britain has a substantial body of data
from the late Pleistocene to recent time (e.g., Pavlides 2004, 2006, Torrence et al. 2000, 2004).
New Britain North Coast

Landscape sampling strategies on Willaumez Peninsula and Garua Island have identified many
localities with evidence for Mid-Holocene human presence (Petrie and Torrence 2008, Torrence
2002a, Torrence and Stevenson 2000). At several localities dentate-stamped pottery occurs strati-
graphically above aceramic levels, but in each case the levels are separated by the W-K2 tephra
(Specht, Fullagar and Torrence 1991, Specht and Torrence 2007a: Table 1, 2007b: Table 2). The
interval between the emplacement of this tephra and re-colonisation of the region was about
140-150 years (Petrie and Torrence 2008: Table 7). It is not clear whether this re-colonisation was
by descendants of the former occupants of the region or by other people. They certainly intro-
duced the first pottery (Specht and Torrence 2007b: 89-90), and possibly ground stone axe and
adze blades, as none have been found in pre-W-K2 levels on New Britain.4 The W-K2 tephra also
marks the end of stemmed tools (with four possible exceptions) that occur widely in pre-W-K2
contexts (Araho, Torrence and White 2002, Rath and Torrence 2003, Torrence 2004).
The four exceptions provide only equivocal evidence for continuity of stemmed tool production,
as in three cases their presence in post-W-K2 contexts could be the result of scavenging from older
contexts. This could apply to several of the tools from site FEA on Boduna Island (Ambrose and
Gosden 1991: Fig. 7, Specht and Summerhayes 2007: 59, Fig. 4), and those in the FYS and FAAJ
16
sites on Garua Island (one tool each: Robin Torrence, pers. comm.). The post-W-K2 assemblages
are dominated by an expedient technology (Hanslip 2001, Torrence 1992), with only the unique
tiny points at FSZ on Garua Island suggesting continued production of complex tools, but these
have different forms and probably different functions to the pre-W-K2 stemmed tools (Torrence
2002b: 186-187).
The Willaumez Peninsula-Garua Island sites, then, show no continuity of site use across the
W-K2 tephra boundary, and the introduction of pottery and ground stone tools, and the loss of
stemmed tools can be read to indicate population replacement. We should bear in mind, however,
that we do not know whether stemmed tools were produced right up to the W-K2 event, or if
production ceased before it.
New Britain South Coast

On the south coast six sites are relevant to the discussion, but at none does a tephra separate
aceramic from ceramic levels.5 Four of these sites are in the Arawe Islands (Figure 2.2a).
Lolmo Cave on Kumbun Island is situated about 10 m above the sea, and was first occupied
around 6280-5870 cal. B.P. (Table 2.1: Beta-28223; Gosden et al. 1994: Table 1). Two other dates
for basal Unit 5 are stratigraphically inconsistent, but both are of mid-Holocene age (Table 2.1:
Beta-26645, Beta-26646). The overlying Units 3 and 4 contain sherds without dentate-stamping
that the excavators assign to “late Lapita pottery types”, but these sherds are not securely dated
(Gosden et al. 1994: 109-110). Sample Beta-26644 (3590-3250 cal. B.P.) for Unit 4 is younger than
Beta-26643 (4710-4210 cal. B.P.) for Unit 3; Beta-26643 is almost certainly an old shell displaced
upwards. Beta-26644 is important, as the shell could not have reached Lolmo without human
agency. As such, the result indicates human activity in the cave at or just prior to the appearance of
pottery in the Bismarck Archipelago, although this conflicts with the excavators’ comments about
the late stylistic affinities of the pottery. Lolmo is also important because of the recovery from the
mid-Holocene Unit 5 of shell artefact types that also occur in Lapita period contexts elsewhere in
the Arawes (Gosden et al. 1994: 110-111, Fig. 6, Smith 1991). Taken together, these two lines of
evidence suggest continuity of site use across the aceramic/ceramic boundary.
The coastal open sites of Apalo (FOJ) on Kumbun Island, Makekur (FOH) on Adwe Island,
and Paligmete (FNY) on Pililo Island have five dates earlier than dentate-stamped pottery. The
Paligmete date of 4790-4400 cal. B.P. (Table 2.1: Beta-27941) was on shell found with a burial at
the base of “Exc. 1” in a level that was probably below high tide at the time of the burial (Gosden
1989: 54, Gosden et al. 1989: 583, Gosden and Webb 1994: 45, Fig. 12, Summerhayes 2000a: 25,
Fig. 3.8).6 The burial lacked the cranium, though the mandible “was present under a pile of clam
shells some 50 cm from the top of the vertebrae” (Gosden and Webb 1994: 45). There do not appear
to have been any artefacts associated with the burial. Gosden and Webb (1994: 47) interpret the date
and its context to indicate a pre-Lapita presence at this site. This claimed association is problematic,
as the separation of the mandible from the rest of the body suggests post-burial disturbance. The
shells might originally have been placed with the skull, but were dislodged together with the
mandible when the cranium was removed, or “the pile of clam shells” could have been part of the
beach deposit into which the burial was placed. There is no firm evidence to associate the dated
shell with the burial and, in the absence of direct dating of the bones I exclude the date and burial
from further consideration here.7
Three samples for Apalo that pre-date dentate-stamped pottery came from squares L-T-U in the
northwest sector of the site (Summerhayes 2000a: 24, Fig. 3.7): 4080-3700 cal. B.P. (Beta-37560)
for square U1 spit 20, 4410-3990 cal. B.P. (Beta-54170) for square U4 spit 22, and 3980-3650
17
cal. B.P. (Beta-55457) for square L2 spit 17 (Table 2.1). The samples came from contexts with
large pieces of wood, “some apparently shaped as posts and planks” and tentatively identified as
remains of a structure built over the inter-tidal reef flat (Gosden and Webb 1994: 38). They were
also associated with other plant remains and obsidian flakes (Hayes 1992, Matthews and Gosden
1997, Specht and Gosden 1997: 178).
The oldest date for Makekur, 4290-3910 cal. B.P. for square F1 spit 9 (Table 2.1: Wk-8539),
was one of five samples submitted to date the pottery in squares D-E-F (Summerhayes 2001: 32,
Table 3). Four samples provided acceptable results but Summerhayes rejected Wk-8539, as it is at
least 650-750 years older than those for comparable levels in squares D3 and E2.
Do these dates from Apalo and Makekur indicate pre-Lapita activity, or should we treat them
as old wood or wood with a substantial in-built age that was burnt or otherwise used during the
Lapita period? To accept the results as indicating old wood would require that the wood survived
in usable form for 200-750 years after the trees died, while assuming that wood with considerable
in-built age was used in Lapita times would require that the trees had been growing for 200-750 years
and that heartwood was used for the proposed inter-tidal structures. None of these scenarios seems
likely to me. The dates are consistent with each other and fit neatly the mid-Holocene chronology
of Unit 5 in Lolmo Cave. We should not be surprised, therefore, that people were also active on
the shorelines of Kumbun and Adwe in the mid-Holocene.
In the Kandrian area, the open beach sites of Iangpun-Rapie (FFT) on Apugi Island and
Ngaikwo (FLX) southeast from Iumielo village have dentate-stamped sherds on the surfaces of
palaeobeaches sealed by clay (Figure 2.2b). In the Arawe Islands Gosden and Webb (1994: 41,
47) argued that such clays probably represent Lapita period erosion of soils from above the sites.
At Iangpun-Rapie and Ngaikwo, the palaeobeaches are dated several millennia older than Lapita
pottery (Table 2.1: Beta-41582, Beta-48862, Beta-48863, Beta-63617; cf. Boyd, Specht and Webb
1999: Table 1), but no pre-Lapita artefactual materials can be attributed with certainty to these
palaeobeaches. At Auraruo (FFS), also on Apugi Island, the palaeobeach has a date of 3920-3590
cal. B.P. (Table 2.1: Beta-63613; Specht and Gosden 1997: 178-79, Appendix 1). This is too old for
a Lapita pottery deposit and could relate to either pre-pottery activity or to a natural beach deposit.
The presence of sherds and obsidian flakes 30 cm below the dated sample presumably reflects
downwards displacement, and we cannot claim categorically pre-pottery usage of the site.
Two rock shelters on the Kandrian area mainland, Alanglongromo (FLF) near Iumielo village
and Alanglong (FLQ) near Analo village, have pre-pottery human presence. Initial examination of
FLQ produced a single plain body sherd, but none were recovered from the subsequent excavations.
Two dates on human bone from the base of trench I gave pre-Lapita period results (Table 2.1:
SUA-2050A and SUA-2050B), but neither is used here as they cannot be adjusted for dietary intake.
I also reject the date of 4120-3670 cal. B.P. (SUA-1523) from above the bone, as the sample came
from a level that is dated much later in trench II, and probably consisted of old shells displaced
during the burial process (Table 2.1; Specht, Lilley and Normu 1983: 94).
Trench II at FLQ has a basal shell date of 4850-4420 cal. B.P. (Table 2.1: Beta-57771), but
there is no evidence for human presence in this level other than perhaps the shells themselves.
Sample Beta-57770 from 10 cm above this date and associated with obsidian flakes was dated to
4280-3820 cal. B.P. (Table 2.1). Although the trench lacked pottery, two samples from 10-15 cm and
30-35 cm above Beta-57770 indicate deposition during the Lapita period (Table 2.1: Beta-57769,
Beta-63615). The difference of about 900-1000 years between this pair of dates and Beta-57770,
and the shallow depth of deposit separating Beta-57770 and Beta-57769 suggest a long period
when the shelter was abandoned or was used very infrequently.
18
At FLF, obsidian flakes and marine shells occur throughout the deposit and dentate-stamped
sherds are stratified above an aceramic level (Specht and Gosden 1997: 179, Summerhayes 2000a:
26, 142-143, Fig. 9.1). The top of the basal clay (Layer 7, aceramic) is dated to 4790-4410 cal.
B.P. (Beta-57768), and dentate-stamped pottery in Layer 5 spit 3 is dated to 3490-3070 cal. B.P.
(Beta-63616) (Table 2.1; Specht and Gosden 1997: Appendix 1). The gap of about a millennium
between the two dates is unlikely to reflect the lack of a date for the intervening 18-30 cm of Layer
6, as this layer contains more than half of the pottery from the shelter and, therefore, should be
of similar age to Beta-63616. If that is so, then there may be a considerable time gap between the
aceramic/ceramic levels, though further dating and analysis of the finds are needed.
Surface finds of obsidian and chert stemmed and waisted tools from several localities on the
south coast suggest widespread activity during the early- and mid-Holocene. Obsidian stemmed tools
comparable to those of Willaumez Peninsula have been reported in the Pomio, Kandrian and Arawe
areas (Specht 2005). In the Arawe Islands, a bifacially flaked chert tool, similar to mid-Holocene
Passismanua examples (e.g., Bulmer 2005: Fig. 4b, Golson 1972: Fig. 2k, Pavlides 2006: Fig. 10.2,
top right), was found at Lolmo Cave on Kumbun Island in the 1960s (not discussed in Gosden
et al. 1994). In view of the mid-Holocene dates for the basal unit of Lolmo, its presence is not
remarkable. Near Kandrian a chert biface was found just above Yimilo cave (FLB) (Figure 2.3a).
The dates for the cave place its first use during the Lapita period, though no pottery was found in
the excavation or near the cave. A waisted chert tool was found on the top of Apugi Island in an
area that is periodically gardened (Figure 2.3b). Several test pits showed thin soil coverage on the
limestone bedrock without stratification.
(a) (b)
Figure 2.2 a. The Arawe Islands, south coast of West New Britain.
b. The Kandrian area, south coast of West New Britain.
19
Figure 2.3 a. Chert biface found near Yimilo cave (FLB), near Kandrian, West New Britain.
b. Chert waisted tool found on the top of Apugi Island, near Kandrian, West New Britain.
Discussion
The Willaumez Peninsula sites show a clear discontinuity in site use of about 140-150 years
marked by the W-K2 tephra, but whether this signifies discontinuity of population remains an issue.
The transition from an aceramic to ceramic material culture is not marked only by the introduction
of pottery and the apparent loss of obsidian stemmed tools. Torrence (2002a: 771) also notes a shift
from dispersed to clustered discard of flaked obsidian on Garua Island between pre- and post-W-K2
contexts, and draws attention to an inferred “increase in the intensity of land management” (2002a:
773). Technological studies of obsidian tools and phytolith studies, however, suggest that patterns
of land use and perhaps subsistence were changing before the W-K2 tephra event (Lentfer and
Torrence 2007: 102, Torrence 1992; cf. Pavlides 2006 for the Passismanua area).
The impact of the W-K2 eruption was less severe on the south coast than on the north side
and did not cause extensive destruction and depopulation. We can reasonably expect, therefore, a
continuity of site use and population across the aceramic/ceramic boundary. Lolmo Cave arguably
provides material evidence for this in the form of shell tool types that continue across the boundary,
as well as Beta-26644 that indicates human activity around the time that pottery was introduced to
the archipelago. The open sites in the Arawes currently have a substantial gap between mid-Holocene
beach level activity and the appearance of pottery, and this might indicate discontinuity, but could
20
equally well reflect the limited number of dates for these basal levels. In the Kandrian area, further
work at the FLF shelter is clearly required to resolve whether there really is a gap of about a
millennium between the aceramic and ceramic levels. The loss of stemmed, waisted and bifacial
tools on the south coast no doubt reflects simply that they were no longer made in the source areas
and cannot be used to argue for population replacement.
PRE-LAPITA ARCHAEOLOGICAL CONTEXTS BEYOND NEW BRITAIN
New Ireland Province

Kirch (2001a: 60, cf. 1997: 39-41) observed that none of the Lapita pottery sites in the Mussau
Group overlies or continues an earlier aceramic occupation, though there are two intriguing pieces
of evidence that might indicate pre-Lapita use of the Mussau area. The first relates to three radio-
carbon dates at the ECA site, and the second to the nature of the avifaunal remains from the
Mussau sites.
Bafmatuk, Egloff and Kaiku (1981: 80) reported a date of 5040-3620 cal. B.P. (Table 2.1:
GX-5499) for the fill of a pit containing pottery at the ECA site on Eloaua Island. This result is
now regarded as having no relevance for dating pottery at ECA (Kirch et al. 1987: 125, Kirch
and Hunt 1988, Spriggs 1990: 17), and the range of 1420 years indicates only that the sample tree
died before the appearance of pottery. We could dismiss the date as a “rogue” result caused by the
sample consisting of old wood burnt during the pottery period, but this would require the wood
to have survived for several centuries before being burnt, in my view an unlikely scenario in this
tropical context. Kirch et al. (1987: 125) suggested that the sample might have incorporated other
material with in-built age, such as shell or coral, but this seems improbable as the sample consisted
of charcoal (Bafmatuk et al. 1981: 80). I suggest, instead, that we can regard the result as indicating
possible human activity prior to the appearance of pottery, and this could explain two other dates
for ECA that have ranges which mostly fall before the accepted starting date for Lapita pottery
(Table 2.1: ANU-5080, Beta-30676). Both samples came from the same palaeobeach terrace as
GX-5499. Kirch (2001b: 223) noted that ANU-5080 “is one of the oldest samples” at ECA, and
suggested some in-built age; using CALIB 5.0.1, ANU-5080 gives a range of 3700-3270 cal. B.P.
Sample Beta-30676, which Kirch (2001b: 227) describes as appearing “to be slightly too old,”
calibrates to 3780-3240 cal. B.P. (using his calculated Delta-R value for the ECA area and the
OxCal program, Kirch obtained an even older result of 4200-3550 cal. B.P.). Neither of the results
is as old as GX-5499, but allowing for the uncertainty about the Delta-R value for Beta-30676, and
accepting that both samples relate to human activity, perhaps they are telling us that this activity
occurred shortly before the introduction of pottery. There is currently no stratigraphic or material
evidence to support this, though the nature of the avifaunal assemblages from the Mussau sites
may point in this direction. In their assessment of these assemblages Steadman and Kirch (1998:
20) noted the absence of flightless elements such as rails, and raised the possibility that “earlier,
pre-Lapita peoples had lived on Mussau and already had extirpated the most vulnerable species of
birds.” Steadman (2006: 128) subsequently expressed a more cautious view, citing the likelihood
of small sample bias. The issue remains unresolved, and the best we can say is that the Mussau
evidence for pre-Lapita presence is not proven, but not beyond the bounds of possibility.
No other open sites with pottery in New Ireland Province have indications of an underlying
aceramic horizon (Pinikindu – Clay 1974; Lamau – Gorecki, Bassett and Head 1991; White 1992;
Lossu/Lesu – White and Downie 1980; Lasigi – Golson 1991, 1992; Anir – Summerhayes 2000b,
2001, 2007; Tanga – Garling 2003: 226, Fig. 5, Table 6).
21
Several cave and rock shelter sites with firm evidence for pre-Lapita use appear to have been
abandoned during the period covered by the present paper (Spriggs 1997: Table 3.1), and those
that contain pottery do not have dentate-stamped sherds. Buang Merabak had plain sherds in a
level dated to the terminal Pleistocene (Leavesley 2004: 62, Rosenfeld 1997: 218), a context that
is clearly unreliable. At Panakiwuk, three sherds from analytical unit A are dated to the last 2000
years and are irrelevant to the present discussion (Marshall and Allen 1991: 87, Table 2). The oldest
sherds from the Balof shelters are dated to 3820-2850 cal. BP (Table 2.1: ANU-4972), a range so
large that it is of no value here (White et al. 1991: 49, Table 2; cf. Downie and White 1978: Fig. 9,
White 1992: 87, Fig. 4).
Manus Province
After East New Britain, Manus is the least archaeologically studied province of the archipelago.
Only Kohin Cave on the southeast coast of Manus Island has dentate-stamped pottery stratified
above an aceramic deposit (Kennedy 1981). Sample ANU-2248, calibrated to 4150-3600 cal. B.P.
(Table 2.1), relates to the surface of Layer 10, a shell midden underlying layers with several dentate-
stamped sherds. The midden contained eight plain sherds, whereas the overlying layers 7-9
contained 883, fifteen times the sherd density of the midden; I follow Spriggs (1997: 112) in
regarding the midden sherds as intrusive. Although the layers overlying the midden reflect a marked
change in depositional activity, the excavator suggests no “chronological hiatus” between them
(Kennedy 1981: 757). In the absence of a date for Layer 9 this cannot be assessed, and all we know
is that the dentate-stamped pottery is younger than 4150-3600 cal. B.P. and older than 2720-2050
cal. B.P. for Layer 5 (Table 2.1: ANU-2212). These dates simply confirm that the dentate-stamped
sherds were deposited at some point during the duration of Lapita pottery in the archipelago.
The only other dated Manus site with dentate-stamped pottery is an open location on Mouk
Island (McEldowney and Ballard 1991), where hydration dates for obsidian found with dentate-
stamped pottery “are within the expected age range of early Lapita sites”, but no results pre-dating
pottery are reported (Ambrose and McEldowney 2000: 269, 276, Table 2). No dentate-stamped
pottery has been reported from the Pamwak site (Fredericksen 1994, Fredericksen, Spriggs and
Ambrose 1993); indeed, the site lacks deposits of that period (Spriggs 1997: Table 3.1). The coastal
open site of Father’s Water (GAC) near Papitalai and the interior Peli Louson rock shelter (GFJ)
have produced mid-Holocene dates for aceramic levels (Pavlides and Kennedy 2007: 199-200),
but the sherds from the late Holocene levels belong to the post-Lapita period on Manus (Kennedy
1983: 116).8
Bougainville Autonomous Region

This region is not part of the Bismarck Archipelago, but I include Nissan Island here because it has
the possibility of “Lapita without pots” preceding the appearance of dentate-stamped sherds
(Spriggs 1991: 237, 1997: 83). Spriggs (1997: 81) notes that at the Lebang Halika rock shelter there
“appears to be no hiatus in occupation” between the aceramic and ceramic levels. Notwithstanding
this claim, I discount this as evidence for continuity on two grounds. First, the Nissan pottery was
not produced locally, but was imported from New Britain or New Ireland (Dickinson 2006: Table
26, D9). This implies a prior stage of production elsewhere before the transfer of pottery to Nissan,
and this appears to be reflected in the dates for the Nissan sites. Spriggs (1997: 80) summarises the
Nissan dates as indicating the arrival of dentate-stamped pottery after 3650-3200 cal. BP, which
would place the pottery later than the oldest pottery in the Bismarck Archipelago. The suggested
phase of “Lapita without pots” could indicate earlier contacts between Nissan and pottery-produc-
ing communities to the north, but this remains to be confirmed.
22
DISCUSSION
The above review suggests continuity of site use across the aceramic/ceramic boundary at Lolmo
Cave, and possibly at the ECA open site in the Mussau area, although in the latter case this
depends on one’s reading of the radiocarbon dates and their relevance.9 The Arawe open sites
show a gap in human activity of at least several centuries, while Kohin cave on Manus and the
FLF shelter near Kandrian lack dates for critical levels. The Willaumez Peninsula-Garua Island
sites show clear discontinuity of site use, with an interval of around 140-150 years separating
pre- and post-W-K2 tephra human activity. There is also a taphonomic issue at these sites, as
organic materials do not survive in soils derived from highly acidic tephras (Specht et al. 1991:
290; cf. Spriggs 1997: 84). For these sites we have only a limited range of inorganic evidence for
comparison with the wider range of materials recovered at other sites such as Lolmo and on
Nissan. The Willaumez Peninsula evidence is important, however, as the W-K2 boundary seems
to mark the end of stemmed tools, with pottery and possibly ground stone axe and adze blades
first appearing with re-colonisation of the area.
Taphonomic issues may also come into play in relation to shorelines during the mid-Holocene
higher sea level. In most, if not all situations in the Bismarck Archipelago, today’s shorelines differ
from those of the pre-Lapita and Lapita periods (cf. Dickinson 2001, 2003, Nunn et al. 2006, Nunn
and Heorake, this volume). Prior to the stabilising of sea level after the mid-Holocene highstand,
many “precipitous” coastlines (Beaton 1995: 801) such as along parts of the north coast of New
Guinea (Terrell 2002, 2004), and along much of New Britain’s south coast, had little or no beach
space for occupation. Where reef growth kept pace with the rising sea, the reefs allowed construction
of buildings in the inter-tidal zone, as in the Arawe Islands and the Mussau Islands (Gosden and
Webb 1994: 47, Kirch 1988, 2001a: 130-34). In general the archaeological evidence for this is often
poorly preserved as unstratified finds in the inter-tidal zone (cf. Felgate 2003, 2007, Specht 1991,
Wickler 2001). In the absence of beaches and/or if the inshore sea was too deep for construction
in the inter-tidal zone, settlement at shoreline level would have been impossible. When conditions
became favourable for shoreline habitation, we could expect people to take advantage of the newly
created space, irrespective of prior patterns of behaviour or supposed ethno-linguistic affiliations.
This appears to be the case at the Apalo and Makekur sites in the Arawe Islands, and continues
today near Kandrian.10
In his list of “Lapita discontinuities”, Spriggs (1997: 88) stated that the “settlement pattern is
itself new”, citing “large villages” as an example. This statement is difficult to assess, as there is
a lack of settlement studies in the archipelago for the period preceding the appearance of pottery.
Torrence’s work on Willaumez Peninsula and Garua Island (Torrence 2002a; cf. Torrence and
Stevenson 2000) suggests a change in land use and settlement behaviour between the aceramic and
ceramic periods that reflects a shift in settlement mobility to more permanently occupied places,
though this could have begun before the W-K2 event (Torrence 1992; cf. Pavlides 2006 for the
Passismanua area).
Spriggs (1997: 88) follows the widely held view that the introduction of pottery was
accompanied by the introduction of “full-on agriculture” and “the beginnings of Pacific animal
husbandry” in the form of pigs, dogs and chickens. Despite continuing claims to the contrary (e.g.,
Allen 2000: 157-58), there is no conclusive evidence for any of the three domesticated animals in
pre-Lapita contexts, and Lapita period introductions seem most likely. Spriggs’ claim for “full-on
agriculture”, however, contrasts with Gosden and Pavlides’ (1994: 169) statement that there were
no “stable farming systems” during the Lapita period and that “only in the last century have full
scale farming systems come about in the Arawes”. An issue here is what precisely do the authors
23
mean by “full-on” and “full scale”, particularly as they agree on some form of food plant production
involving vegetation clearance as the likely cause of the erosion of clays evident at many Lapita
period shoreline sites (e.g., Gosden and Pavlides 1994: 169; Gosden and Webb 1994, Spriggs 1997:
85-86). The absence of such clays at some locations may have a geomorphological explanation,
and does not mean that erosion did not occur before the Lapita pottery period (Gosden and Webb
1994: 35). If beaches did not exist in pre-Lapita time, then conditions suitable for the capture of
sediments also did not exist.
Possible support for a Lapita period change in plant food production comes from analyses
of phytoliths from palaeosols formed on Witori tephras on Willaumez Peninsula. On Garua Island
and at Numundo Plantation there is evidence for burning of vegetation prior to the W-K2 event,
but after it the extent of burning increased and several possible cultivars appeared (Boyd, Lentfer
and Parr 2005: 388-94, Tables 2 and 6). These changes probably reflect human-induced vegetation
clearance and cultivation (Lentfer and Torrence 2007: 102), but the pre-W-K2 tephra evidence may
indicate earlier cultivation practices.
CONCLUSION
This paper has focused on the transition from an aceramic to ceramic material culture in the
Bismarck Archipelago. The main conclusion is that we know very little about this critical period,
and the reasons for this are varied. There is an important issue with the interpretation of the
situation on Willaumez Peninsula, where changes in the pre- and post-W-K2 tephra suites of
evidence are separated by about 140-150 years. In areas of New Britain covered by W-K tephras
organic materials have not survived, and this limits the range of artefactual evidence for compar-
ison between periods and sites. Some sites have gaps in their sequences and problems with the
dating of key contexts, and there is a general lack of local histories of shoreline development and
pre-Lapita settlement patterns.
It is tempting to gloss over these issues and to declare that the appearance of Lapita pottery
and other elements of a “Neolithic package” did mark a significant cultural shift. This would ignore,
however, the discontinuous and patchy nature of the pre-Lapita archaeological record, particularly
the apparent gap of several centuries between the pre-Lapita and Lapita evidence in areas unaffected
or only slightly affected by the cataclysmic W-K2 event on New Britain. We need to examine what
happened during this period. In some areas it appears to have lasted for 10-20 human generations,
during which there could have been changes in material culture, subsistence and settlement patterns
before the appearance of Lapita pottery in the Bismarck Archipelago. Moreover, we need to consider
how the competing models of revolutionary versus evolutionary change might be expressed in
time-averaged archaeological deposits that do not permit fine temporal discrimination.
These points warrant closer examination than they have hitherto received, since they have
a significant bearing on the stories that we create from the various strands of evidence relating to
the role of the Bismarck Archipelago in the “master narrative of Austronesian expansion” cited at
the start of the paper. At this stage we may do well to heed Ambrose’s (1991: 112) observation on
the Manus ceramic sequence: “whether the Lapita event was simply a trigger for local change, an
incidental marker of changes already underway, or marked the presence of sustained influential
factors from abroad, is at present not clear”.
24
NOTES
1. The term “Lapita” has been used both loosely and very precisely in different contexts, sometimes as a
noun, sometimes as an adjective. At one extreme it refers only to dentate-stamped pottery and excludes
plain pottery or pottery with other kinds of surface modifications. At the other extreme it covers a
cultural complex, with implications for language, material culture, economy and social organisation.
Except where I quote other authors, in the text “Lapita pottery” refers to the “Lapita ceramic series”, and
“ceramic period” or “Lapita period” to the time when this series existed. “Dentate-stamped pottery”
indicates only the earlier part of this series, and does not take into account the possibility of an initial
plain, red-slipped phase (Kirch 1996: 65, 2001a: 85). “Lapita sites” refers to locations with dentate-
stamped pottery, without implications about the activities conducted at the locations.
2. For Huon Peninsula, on the mainland of New Guinea opposite New Britain, Edwards et al. (1993) calcu-
lated a 14C ocean reservoir effect of 407 ± 52 years. The relevance of this value to the New Britain marine
samples cited here is unknown. There are also inconsistencies in the way dates have been published,
especially those presented as “radiocarbon years before present” rather than as “conventional radiocarbon
ages” (Stuiver and Polach 1977). Detailed discussion of this issue lies outside the present paper.
3. “Passismanua” is the census division that covers communities living inland from Kandrian (Chowning
and Goodale 1965), and includes the sites excavated by Pavlides (1999, 2004, 2006), who assigns them
to the “Yombon” area. This name was previously recorded as “Yambon” by Chowning, Goodale and
myself (Specht, Lilley and Normu 1981), and as “Iombon” on administration files in Kandrian. None of
these forms is correct. Sengseng people informed me that the name should be “Ayambon”, which is the
name of an ancestral burial ground. Use of Yombon or Yambon to indicate the extent of chert waisted,
stemmed and bifacial tools is misleading, as these tools are found over a much wider area (Chowning
and Goodale 1966: Fig. 1; Specht fieldnotes).
4. Fully or partially ground stone axe and adze blades are rare in excavations in the Bismarck Archipelago,
but may pre-date Lapita pottery. A flake “from the side of a ground stone axe” was found at Balof 2 in
a context dated about 10,000 bp (White et al. 1991: 49). A terminal Pleistocene to mid-Holocene age
has been suggested for two edge-ground stone tools at Pamwak on Manus (Fredericksen 1994: 77).
Pamwak also produced a small fully ground tool that is possibly of mid-Holocene age, though there
seems to be some uncertainty about its precise provenance (Fredericksen 1994: 54, Golson 2005: 476,
Spriggs 1997: 59). A group of ground shell adzes from this site may also be of early- to mid-Holocene
age (Fredericksen 1994: 63, Golson 2005: 477).
5. Plain sherds were found in the palaeosol on the W-K2 tephra at Auwa (FGT) in the Passismanua area
(Specht et al. 1981: 13). These sherds are not considered here as they are dated to 2850-2360 cal. B.P.
(Beta-1545: 2575±100 b.p.; Specht et al. 1981: 14), long after the appearance of Lapita pottery in the
Bismarck Archipelago. As on Willaumez Peninsula, the FGT sherds are separated from the aceramic
levels by the W-K2 tephra, which marks the end of the so-called “waisted” chert tools of this region
(Pavlides 2004, 2006).
6. Gosden (1989: 55, Gosden and Webb: 1994: 45) deducted 450 years from Beta-27941 to allow for the
ORE and cited it as 3960 ± 70 b.p. The figure used here is that reported by Beta Analytic Inc. after
adjustment for δ13C (Specht and Gosden 1997: 178, Spriggs 2003: Table 1).
7. Green (2006) includes the Paligmete burial in his list of definite and possible Lapita period burials. The
possibility of a Lapita association is based on the absence of the skull, a feature of Lapita period burials
elsewhere (e.g., Green, Anson and Specht 1989: 219-220, Valentin 2003). Removal of the cranium,
however, was not restricted to Lapita period burials, and was still practiced in New Britain in the 19th
and 20th centuries (e.g., Parkinson 1999 [1907]: 38). A post-Lapita date is improbable, as it seems
unlikely that the excavators would have missed a grave pit dug through 2 m of deposit that otherwise
retained a clear stratigraphy. Sediment disturbance is indicated, however, by aberrant radiocarbon dates
of 560±170 b.p. and “modern” (ANU-4985, ANU-4987 and ANU-4986) from the “artifact-rich sand
layer” of TP 3, ten metres from “Exc. 1” (Gosden and Webb 1994: 45, Fig. 12, where the layer is marked
as “white sand with finds”).
25
8. The scarcity of dentate-stamped pottery sites in Manus Province has been attributed to “the general lack
of exposed archaeological sites within the crucial time range” (Ambrose and McEldowney 2000: 268).
The Mouk site and the presence of worked obsidian in mid- and late Holocene contexts indicate that this
situation is not universal across the province and other reasons are likely (cf. McEldowney and Ballard
1991: 97-101). Blust (1998) suggests that this scarcity reflects the early separation of the Admiralties
sub-group from the rest of the Oceanic languages, and that development of the Lapita cultural complex
occurred after this separation. As Kennedy (2002: 22-23) points out, the archaeological evidence is too
thin to support such an interpretation.
9. There are currently no confirmed pre-Lapita sites south of Guadalcanal in the Solomon Islands.
Acceptance of the arguments used in the Mussau discussion, however, should perhaps lead to acceptance
of UCR-965 (3960-3340 cal. B.P.) from Tikopia in Remote Oceania (Kirch and Yen 1982: Table 50).
Writing when the chronology of Lapita dispersal was insecure, Kirch and Yen (1982: 312, 314) noted that
the date might indicate an exploratory visit to the island on what Irwin (1992: 61) has called “the frontier
of exploration”. In Vanuatu Bedford (2006: 259-60) found no evidence for pre-Lapita use of the islands,
exploratory or otherwise. The early Tikopia date, therefore, could reflect wood with built-in age.
10. According to the late Peter Kikin of Iumielo village near Kandrian, much of the current village beach
area was formed during the 20th century. When he was young, before the Pacific War, most people lived
on the cliff top behind Iumielo, as there was almost no coastal flat. While the cliff top residency could
have been as much a defensive posture as a reaction to lack of beach level space, Kikin pointed out a
large coral boulder today standing about 20 m from the sea that was formerly in the inter-tidal zone.
Between 1979 (my first visit to Iumielo) and 2003 the shoreline advanced at least three metres.
REFERENCES
Allen, J., 1991. Introduction. In J. Allen and C. Gosden (eds), Report of the Lapita Homeland Project.
Occasional Papers in Prehistory 20. Canberra: Department of Prehistory, Research School of Pacific
Studies, Australian National University, pp. 1-8.
Allen, J., 2000. From beach to beach: The development of maritime economies in prehistoric Melanesia.
In S. O’Connor and P. Veth (eds), East of Wallace’s Line: Studies of Past and Present Maritime
Cultures of the Indo-Pacific Region. Modern Quaternary Research in Southeast Asia 16. Rotterdam:
A.A. Balkema, pp. 139-76.
Allen, J. and C. Gosden, 1996. Spheres of interaction and integration: Modelling the culture history of the
Bismarck Archipelago. In J. Davidson, G. Irwin, F. Leach, A. Pawley and D. Brown (eds), Oceanic
Culture History: Essays in Honour of Roger Green. Dunedin: New Zealand Journal of Archaeology
Special Publication, pp. 183-97.
Allen, J. and J.P. White, 1989. The Lapita homeland: Some new data and an interpretation. Journal of the
Polynesian Society, 98 (2): 129-46.
Ambrose, W.R., 1991. Lapita or not Lapita: The case of the Manus pots. In J. Allen and C. Gosden (eds),
Report of the Lapita Homeland Project. Occasional Papers in Prehistory 20. Canberra: Department
of Prehistory, Research School of Pacific Studies, Australian National University, pp. 103-12.
Ambrose, W.R. and C. Gosden, 1991. Investigations on Boduna Island. In J. Allen and C. Gosden (eds),
Report of the Lapita Homeland Project. Occasional Papers in Prehistory 20. Canberra: Department
of Prehistory, Research School of Pacific Studies, Australian National University, pp. 182-88.
Ambrose, W.R. and H. McEldowney, 2000. An age assessment for Lapita from obsidian at the Mouk Island
site, Manus. In A. Anderson and T. Murray (eds), Australian Archaeologist. Collected Papers in
Honour of Jim Allen. Canberra: Coombs Academic Publishing, Australian National University, pp.
268-78.
Anson, D., R. Walter and R.C. Green, 2005. A Revised and Redated Event Phase Sequence for the
Reber-Rakival Lapita Site, Watom Island, East New Britain Province, Papua New Guinea.
University of Otago Studies in Prehistoric Archaeology 20. Dunedin: Department of Anthropology,
University of Otago.
26
Araho, N., R. Torrence and J.P. White, 2002. Valuable and useful: Mid-Holocene stemmed obsidian artefacts
from West New Britain, Papua New Guinea. Proceedings of the Prehistoric Society, 68: 61-81.
Bafmatuk, F., B. Egloff and R Kaiku, 1981. Islanders: Past and present. In K.R. Henderson (ed.), Hemisphere:
An Asian-Australian Annual. Canberra: Australian Government Publishing Service, pp. 77-81.
Beaton, J.M., 1995. The transition on the coastal fringe of Greater Australia. Antiquity, 69 [265]: 798-806.
Bellwood, P., 1995. Austronesian prehistory in Southeast Asia: Homeland, expansion and transformation.
In P. Bellwood, J.J. Fox and D. Tryon (eds), The Austronesians: Historical and Comparative
Perspectives. Canberra: Department of Anthropology, Research School of Pacific and Asian Studies,
Australian National University, pp. 96-111.
Blust, R., 1998. A note on higher-order subgroups in Oceanic. Oceanic Linguistics, 37: 182-88.
Bourke, R.M., R.L. Hide, B.J. Allen, D. Fritsch, R. Grau, P. Hobsbawn, M. Levett, S. Lyon, L. Nama
and T. Nen, 1996. West New Britain Province: Text Summaries, Maps, Code Lists and Village
Identification. Agricultural Systems of Papua New Guinea Working Paper 13. Canberra: Department
of Human Geography, Australian National University.
Boyd, W.E., C.J. Lentfer and G.O. Luker, 1999. Environmental impacts of major catastrophic Holocene
volcanic eruptions in New Britain, P.N.G.: A preliminary model for palaeoenvironmental change. In
J.A. Kesby, J.M. Stanley, R.F. McLean and L.J. Olive (eds), Geodiversity: Readings in Australian
Geography at the Close of the 20th Century. Special Publication Series 6. Canberra: School of
Geography and Oceanography, University College, Australian Defence Force Academy, pp. 361-72.
Boyd, W.E., C.J. Lentfer and J. Parr, 2005. Interactions between human activity, volcanic eruptions and
vegetation during the Holocene at Garua and Numundo, West New Britain. Quaternary Research,
64: 384-98.
Boyd, W.E., J. Specht, and J. Webb, 1999. Holocene shoreline change and archaeology on the Kandrian
coast of West New Britain, Papua New Guinea. In J. Hall and I.J. McNiven (eds), Australian
Coastal Archaeology. Research Papers in Archaeology and Natural History 31. Canberra: ANH
Publications, pp. 283-87.
Bulmer, S., 2005. The stones of Pasismanua revisited. In C. Gross, H.D. Lyons and D.A. Counts (eds),
A Polymath Anthropologist: Essays in Honour of Ann Chowning. Research in Anthropology and
Linguistics Monograph 6. Auckland: Department of Anthropology, University of Auckland, pp. 23-34.
Chappell, J., A. Omura, T. Esat, M. McCulloch, J. Pandolfi, Y. Ota and B. Phillips, 1996. Reconciliation
of late Quaternary sea levels derived from coral terraces at Huon Peninsula with deep sea oxygen
isotope records. Earth and Planetary Science Letters, 141: 227-36.
Chowning, A. and J. Goodale, 1965. The Passismanua Census Division, West New Britain Open Electorate.
In D.G. Bettison, C.A. Hughes and P.W. van der Veur (eds), The Papua-New Guinea Elections 1964.
Canberra: ANU Press, pp. 264-79.
Chowning, A. and J. Goodale, 1966. A flint industry from southwest New Britain, Territory of New Guinea.
Asian Perspectives, 9: 150-53.
Clay, R.B., 1974. Archaeological reconnaissance in central New Ireland. Archaeology and Physical
Anthropology in Oceania, 9 (1): 1-17.
Dickinson, W.R., 2001. Palaeoshoreline record of relative Holocene sea levels on Pacific Islands.
Earth-Science Reviews, 55 (3-4): 191-234.
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.
Dickinson, W.R., 2006. Temper Sands in Prehistoric Oceanian Pottery: Geotectonics, Sedimentology,
Petrography, Provenance. Special Paper 406. Boulder (COL): The Geological Society of America.
27
Downie, J.E. and J.P. White, 1978. Balof Shelter, New Ireland: Report on a small excavation. Records of
the Australian Museum, 31: 762-802.
Edwards, R.L., J.W. Beck, G.S. Burr, D.J. Donahue, J.M.A. Chappell, A.L. Bloom, E.R.M. Druffel, and
F.W. Taylor, 1993. A large drop in atmospheric 14C/12C and reduced melting in the Younger Dryas,
documented with 230Th ages of corals. Science, 260: 962-68.
Lagoon, New Georgia, Solomon Islands. Unpublished Ph.D. thesis, University of Auckland,
Auckland.
Felgate, M.W., 2007. Leap-frogging or limping? Recent evidence from the Lapita littoral fringe,
New Georgia, Solomon Islands. In S. Bedford, C. Sand and S.P. Connaughton (eds), Oceanic
Explorations: Lapita and Western Pacific Settlement. Terra Australis 26. Canberra: ANU E Press,
pp. 123-40.
Fredericksen, C., 1994. Patterns in Glass: Obsidian and Economic Specialisation in the Admiralty Islands.
Unpublished Ph.D. thesis, Australian National University, Canberra.
Fredericksen, C., M. Spriggs and W. Ambrose, 1993. Pamwak Rockshelter: A Pleistocene site on Manus
Island, Papua New Guinea. In M.A. Smith, M. Spriggs and B. Fankhauser (eds), Sahul in Review:
Pleistocene Archaeology in Australia and Island Melanesia. Occasional Papers in Prehistory 24.
Canberra: Department of Prehistory, Research School of Pacific Studies, Australian National
University, pp. 144-54.
Garling, S., 2003. Tanga takes the stage: Another model ‘transitional’ site? New evidence and a
Golson, J., 1972. Both sides of the Wallace Line: New Guinea, Australia, Island Melanesia and Asian
prehistory. In N. Barnard (ed.), Early Chinese Art and its Possible Influence in the Pacific Basin.
New York: Intercultural Arts Press, pp. 533-95.
Golson, J., 1991. Two sites at Lasigi, New Ireland. In J. Allen and C. Gosden (eds), Report of the Lapita
Homeland Project. Occasional Papers in Prehistory 20. Canberra: Department of Prehistory,
Research School of Pacific Studies, Australian National University, pp. 244-59.
Golson, J., 1992. The pottery from Lasigi, New Ireland. In J-C. Galipaud (ed.), Poterie Lapita et
Peuplement. Nouméa: ORSTOM, pp. 155-67.
Golson, J., 2005. The middle reaches of New Guinea history. In A. Pawley, R. Attenborough, J. Golson
and R. Hide (eds), Papuan Pasts: Cultural, Linguistic and Biological Histories of Papuan-Speaking
People. Pacific Linguistics 572. Canberra: Australian National University, pp. 451-91.
Gorecki, P., S. Bassett and J. Head, 1991. A Lapita site at Lamau, New Ireland mainland. In J. Allen and
C. Gosden (eds), Report of the Lapita Homeland Project. Occasional Papers in Prehistory 20.
Canberra: Department of Prehistory, Research School of Pacific Studies, Australian National
Gosden, C., 1989. Prehistoric social landscapes of the Arawe Islands, West New Britain Province, Papua
New Guinea. Archaeology in Oceania, 24 (2): 45-58.
Gosden, C., J. Allen, W. Ambrose, D. Anson, J. Golson, R. Green, P. Kirch, I. Lilley, J. Specht and M.
Spriggs, 1989. Lapita sites of the Bismarck Archipelago. Antiquity, 63: 561-86.
Gosden, C, and C. Pavlides, 1994. Are islands insular? Landscape vs. seascape in the case of the Arawe
Islands, Papua New Guinea. Archaeology in Oceania, 29 (3): 162-71.
Gosden, C. and J. Webb, 1994. The creation of a Papua New Guinean landscape: Archaeological and geo-
morphological evidence. Journal of Field Archaeology, 21 (1): 29-51.
Gosden, C., J. Webb, B. Marshall and G.R. Summerhayes, 1994. Lolmo Cave: A mid- to late Holocene
Site, the Arawe Islands, West New Britain Province, Papua New Guinea. Asian Perspectives, 33 (1):
97-119.
28
Green, R.C., 1991. The Lapita cultural complex: Current evidence and proposed models. In P. Bellwood
(ed.), Indo-Pacific Prehistory 1990: Proceedings of the 14th Congress of the Indo-Pacific Prehistory
Association. Canberra: Indo-Pacific Prehistory Association, pp. 295-305.
Green, R.C., 2000. Lapita and the cultural model for intrusion, integration and innovation. In A. Anderson
and T. Murray (eds), Australian Archaeologist. Collected Papers in Honour of Jim Allen. Canberra:
Coombs Academic Publishing, Australian National University, pp. 372-92.
Green, R.C., 2003. The Lapita horizon and traditions – signature for one set of Oceanic migrations. In
Nouvelle-Calédonie 15. Nouméa: Service de Musées et du Patrimoine de Nouvelle-Calédonie, pp.
95-120.
Green, R.C., 2006. Notes accompanying a set of selected Lapita-related reprints privately circulated by
R.C. Green.
Green, R.C. and D. Anson, 2000. Excavations at Kainapirina (SAC), Watom Island, Papua New Guinea.
Green, R.C., D. Anson and J. Specht, 1989. The SAC burial ground, Watom Island, Papua New Guinea.
Records of the Australian Museum, 41 (3): 215-21.
Hanslip. M.D., 2001. Expedient Technologies? Obsidian Artefacts in Island Melanesia. Unpublished Ph.D.
Hayes, L.T., 1992. Plant Macroremains from Archaeological Sites in the Arawe Islands, Papua New
Guinea. Unpublished B.A. (Hons) thesis, La Trobe University, Bundoora (Vic.).
Irwin, G.J., 1992. The Prehistoric Exploration and Colonisation of the Pacific. Cambridge: Cambridge
University Press.
Johnson, R.W., 1976. Late Cainozoic volcanism and plate tectonics at the southern margin of the Bismarck
Sea, Papua New Guinea. In R.W. Johnson (ed.), Volcanism in Australasia: A Collection of Papers in
Honour of the Late G.M. Taylor, G.C. Amsterdam: Elsevier Scientific Publishing Co., pp. 101-16.
Kennedy, J., 1981. Lapita colonization of the Admiralty Islands? Science, 213: 757-59.
Kennedy, J., 1983. On the prehistory of Western Melanesia: The significance of new data from the
Admiralty Islands. Australian Archaeology, 16: 115-22.
Kennedy, J., 2002. Manus from the beginning: An archaeological overview. In C. Kaufmann, C. Kocher
Schmid and S. Ohnemus (eds), Admiralty Islands: Art from the South Seas. Zurich: Museum
Rietberg, pp. 17-28.
Kirch, P.V., 1988. The Talepakemalai Lapita site and Oceanic prehistory. National Geographic Research, 4
(3): 328-42.
Kirch, P.V., 1995. The Lapita culture of western Melanesia in the context of Austronesian origins and
dispersal. In P. Li, C. Tsang, Y. Huang, D. Ho and C. Tseng (eds), Austronesian Studies Relating to
Taiwan. Taipei: Academia Sinica, pp. 255-94.
Kirch, P.V., 1996. Lapita and its aftermath: The Austronesian settlement of Oceania. In W.H. Goodenough
(ed.), Prehistoric Settlement of the Pacific. Transactions of the American Philosophical Society 86
(5). Philadelphia: American Philosophical Society, pp. 57-70.
Kirch, P.V., 1997. The Lapita Peoples: Ancestors of the Oceanic World. Oxford: Blackwell.
Kirch, P.V., 2001a. Three Lapita villages: Excavations at Talepakemalai (ECA), Etakosarai (ECB),
and Etapakengaroasa (EHB), Eloaua and Emananus Islands. In P.V. Kirch (ed.), Lapita and its
Transformations in Near Oceania: Archaeological Investigations in the Mussau Islands, Papua New
Guinea, 1985-1988. Contributions of the University of California Archaeological Research Facility
No. 59. Berkeley: University of California. pp. 68-145.
29
Kirch, P.V., 2001b. A radiocarbon chronology for the Mussau Islands. In P.V. Kirch (ed.), Lapita and its
Transformations in Near Oceania: Archaeological Investigations in the Mussau Islands, Papua New
Guinea, 1985-1988. Contributions of the University of California Archaeological Research Facility
No. 59. Berkeley: University of California, pp. 196-227.
Kirch, P.V., M.S. Allen, V.L. Butler and T.L. Hunt, 1987. Is there an early far western Lapita province?
Sample size effects and new evidence from Eloaua Island. Archaeology in Oceania, 22 (3): 123-27.
Kirch, P.V. and T.L. Hunt, 1988. Radiocarbon dates from the Mussau Islands and the Lapita colonization of the
Southwest Pacific. Radiocarbon, 30 (2): 161-69.
Kirch, P.V. and D.E. Yen, 1982. Tikopia: The Prehistory and Ecology of a Polynesian Outlier. Bernice P.
Bishop Museum Bulletin 238. Honolulu.
Leavesley, M.G., 2004. Trees to the Sky: Prehistoric Hunting in New Ireland, Papua New Guinea.
Lentfer, C. and R. Torrence, 2007. Holocene volcanic activity, vegetation succession, and ancient human
land use: Unraveling the interactions on Garua Island, Papua New Guinea. Review of Palaeobotany
and Palynology, 143: 83–105.
Machida, H., R. Blong, J. Specht, H. Moriwaki, R. Torrence, Y. Hayakawa, B. Talai, D. Lolok and C.F.
Pain, 1996. Holocene explosive eruptions of Witori and Dakataua caldera volcanoes in West New
Britain, Papua New Guinea. Quaternary International, 34-36: 65-78.
Marshall, B. and J. Allen, 1991. Excavations at Panakiwuk Cave, New Ireland. In J. Allen and C. Gosden (eds),
Report of the Lapita Homeland Project. Occasional Papers in Prehistory 20. Canberra: Department of
Prehistory, Research School of Pacific Studies, Australian National University, pp. 59-91.
Matthews, P.J. and C. Gosden, 1997. Plant remains from waterlogged sites in the Arawe Islands, West New
Britain Province, Papua New Guinea: Implications for the prehistory of plant use and domestication.
Economic Botany, 51(2): 121-33.
McEldowney, H. and C. Ballard, 1991. The Mouk Island site: Manus as paradox or parable in
reconstructions of the Lapita cultural complex? In J. Allen and C. Gosden (eds), Report of the
Lapita Homeland Project. Occasional Papers in Prehistory 20. Canberra: Department of Prehistory,
Nairn, I.A., C.O. McKee, B. Talai and C.P. Wood, 1995. Geologic and eruptive history of the Rabaul
Caldera area, Papua New Guinea. Journal of Volcanology and Geothermal Research, 69: 255-84.
Nunn, P.D., S. Matararaba, R. Kumar, C. Pene, L. Yuen and M.R. Pastorizo, 2006. Lapita on an island
in the mangroves? The earliest human occupation at Qoqo Island, southwest Viti Levu, Fiji.
Archaeology in New Zealand, 49: 205-12.
Pandolfi, J.M., A.W. Tudhope, G. Burr, J. Chappell, E. Edinger, M. Frey, R. Steneck, C. Sharma, A. Yeates,
M. Jennions, H. Lescinsky and A. Newton, 2006. Mass mortality following disturbance in Holocene
coral reefs from Papua New Guinea. Geology, 34 (11): 949-52.
Parkinson, R.R.H., 1999 [1907]. Thirty Years in the South Seas. Land and People, Customs and Traditions
in the Bismarck Archipelago and on the German Solomon Islands. Translated by J. Dennison.
Translation edited by J.P. White. Bathurst (NSW): Crawford House Publishing in association with
Oceania Publications, University of Sydney.
Pavlides, C., 1999. The Story of Imlo: The Organisation of Flaked Stone Technologies from the Lowland
Tropical Rainforest of West New Britain, Papua New Guinea. Unpublished Ph.D. thesis, La Trobe
University, Bundoora (Vic).
Pavlides, C., 2004. From Misisil cave to Eliva hamlet: Rediscovering the Pleistocene in interior West New
Britain. In V.J. Attenbrow and R. Fullagar (eds), Pacific Odyssey: Archaeology and Anthropology
in the Western Pacific: Papers in Honour of Jim Specht. Records of the Australian Museum,
Supplement 29. Sydney: The Australian Museum, pp. 97-108.
30
Pavlides, C., 2006. Life before Lapita: New developments in Melanesia’s long-term history. In I. Lilley
(ed.), Archaeology of Oceania: Australia and the Pacific Islands. Oxford: Blackwell, pp. 205-27.
Pavlides, C. and J. Kennedy, 2007. Archaeological Studies of the Middle and Late Holocene, Papua New
Guinea. Part V: Pre-Lapita Horizons in the Admiralty Islands: Flaked Stone Technology from GAC
and GFJ. Technical Reports of the Australian Museum, 20: 197-215. Published online: www.austral-
ianmuseum.net.au/pdf/publications/1477_complete.pdf
Petchey, F., R.C. Green, M. Jones and M. Phelan, 2005. A local marine reservoir correction value (ΔR) for
Watom Island, Papua New Guinea. New Zealand Journal of Archaeology, 26 [2004]: 29-40.
Petchey, F., M. Phelan and J.P. White, 2004. New ΔR values for the Southwest Pacific Ocean. Radiocarbon,
46 (2): 1005-14.
Petrie, C.A. and R. Torrence, 2008. Assessing the effects of volcanic disasters on human settlement in the
Willaumez Peninsula, Papua New Guinea: A Bayesian approach to radiocarbon calibration. The
Holocene, 18(5): 729-744.
Rath, P. and R. Torrence, 2003. Producing value: Stemmed tools from Garua Island, Papua New Guinea.
Australian Archaeology, 57: 119-27.
Riker-Coleman, K.E., C.D. Gallup, L.M. Wallace, J.M. Webster, H. Cheng and R.L. Edwards, 2006.
Evidence of Holocene uplift in east New Britain, Papua New Guinea. Geophysical Research Letters,
33 L18612: 1-4.
Rosenfeld, A., 1997. Excavation at Buang Merabak, central New Ireland. Bulletin of the Indo-Pacific
Prehistory Association, 16: 213-24.
Smith, A., 1991. Mollusc of the Ancient Mariner. Shell Artefacts: Typology, Technology and Pacific
Prehistory. Unpublished B.A. (Hons) thesis, La Trobe University, Bundoora (Vic).
Smith, I.E. and R.W. Johnson, 1981. Contrasting rhyolite suites in the Late Cenozoic of Papua New
Guinea. Journal of Geophysical Research, 86, B11: 10257-272.
Specht, J., 1968. Preliminary report of excavations on Watom Island. Journal of the Polynesian Society
77 (2): 117–34.
Specht, J., 1991. Kreslo: A Lapita pottery site in southwest New Britain, Papua New Guinea. In J. Allen
and C. Gosden (eds), Report of the Lapita Homeland Project. Occasional Papers in Prehistory
20. Canberra: Department of Prehistory, Research School of Pacific Studies, Australian National
Specht, J., 2005. Obsidian stemmed tools in New Britain: Aspects of their role and value in mid-Holocene
Papua New Guinea. In I. Macfarlane, M.-J. Mountain and R. Paton (eds), Many Exchanges:
Archaeology, History, Community and the Work of Isabel McBryde. Aboriginal History Monograph
11. Canberra: Aboriginal History Inc., pp. 373-92.
Specht, J., 2007. Small islands in the big picture: The formative period of Lapita in the Bismarck
Archipelago. In S. Bedford, C. Sand and S.P. Connaughton (eds), Oceanic Explorations: Lapita and
Specht, J., R. Fullagar and R. Torrence, 1991. What was the significance of Lapita pottery at Talasea? In P.
Bellwood (ed.), Indo-Pacific Prehistory 1990: Proceedings of the 14th Congress of the Indo-Pacific
Prehistory Association. Canberra: Indo-Pacific Prehistory Association, pp. 281-94.
Specht, J. and C. Gosden, 1997. Dating Lapita pottery in the Bismarck Archipelago, Papua New Guinea.
Asian Perspectives, 36 (2): 175-99.
Specht, J., I. Lilley and J. Normu, 1981. Radiocarbon dates from West New Britain, Papua New Guinea.
Australian Archaeology, 12: 13-15.
Specht, J., I. Lilley and J. Normu, 1983. More on radiocarbon dates from West New Britain, Papua New
Guinea. Australian Archaeology, 16: 92-95.
31
Specht, J. and G.R. Summerhayes, 2007. Archaeological Studies of the Middle and Late Holocene,
Papua New Guinea. Part II. The Boduna Island (FEA) Lapita Site, Papua New Guinea. Technical
Reports of the Australian Museum, 20: 51-103. Published online: www.australianmuseum.net.au/pdf/
publications/1474_complete.pdf
Specht, J. and R. Torrence, 2007a. Archaeological Studies of the Middle and Late Holocene, Papua
New Guinea. Part IV. Pottery Sites of the Talasea Area, Papua New Guinea. Technical Reports
of the Australian Museum, 20: 131-96. Published online: www.australianmuseum.net.au/pdf/
publications/1476_complete.pdf
Specht, J. and R. Torrence, 2007b. Lapita all over: Land-use on the Willaumez Peninsula, Papua New Guinea.
In S. Bedford, C. Sand and S.P. Connaughton (eds), Oceanic Explorations: Lapita and Western
Pacific Settlement. Terra Australis 26. Canberra: ANU E Press, pp. 71-96.
Spriggs, M., 1990. Dating Lapita: Another view. In M. Spriggs (ed.), Lapita Design, Form and
Composition. Occasional Papers in Prehistory 19. Canberra: Department of Prehistory, Research
School of Pacific Studies, Australian National University, pp. 6-27.
Spriggs, M., 1991. Nissan: The island in the middle. In J. Allen and C. Gosden (eds), Report of the Lapita
Homeland Project. Occasional Papers in Prehistory 20. Canberra: Department of Prehistory,
Spriggs, M., 1996. What is Southeast Asian about Lapita? In T. Azakawa and E. Szathmary (eds),
Prehistoric Mongoloid Dispersals. Oxford: Oxford University Press, pp. 324-48.
Spriggs, M., 2003. Chronology of the Neolithic transition in Island Southeast Asia and the Western Pacific:
A view from 2003. The Review of Archaeology, 24 (2): 57-80.
Spriggs, M., 2007. The Neolithic and Austronesian expansion within Island Southeast Asia and into
the Pacific. In S. Chiu and C. Sand (eds), From Southeast Asia to the Pacific: Archaeological
Perspectives on the Austronesian Expansion and the Lapita Cultural Complex. Taipei: Center for
Archaeological Studies, Academia Sinica, pp. 104-25.
Steadman, D.W., 2006. Extinction and Biogeography of Tropical Pacific Birds. Chicago: University of
Chicago Press.
Steadman, D.W. and P.V. Kirch, 1998. Biogeography and prehistoric exploitation of birds in the Mussau
Islands, Bismarck Archipelago, Papua New Guinea. Emu, 98: 13-22.
Stuiver, M. and H. Polach, 1977. Discussion: Reporting of 14C data. Radiocarbon, 19 (3): 355-63.
Stuiver, M. and P.J. Reimer, 1993. Extended 14C database and revised CALIB 3.0 radiocarbon calibration
program. Radiocarbon, 35 (1): 215-30.
Summerhayes, G.R., 2000a. Lapita Interaction. Terra Australis 15. Canberra: Department of Archaeology
and Natural History and Centre for Archaeological Research, Australian National University.
Summerhayes, G.R., 2000b. Recent archaeological investigations in the Bismarck Archipelago, Anir –
New Ireland Province, Papua New Guinea. Bulletin of the Indo-Pacific Prehistory Association, 19:
167-74.
Summerhayes, G.R., 2001. Defining the chronology of Lapita in the Bismarck Archipelago. In G.R. Clark,
Summerhayes, G.R., 2007. The rise and transformations of Lapita in the Bismarck Archipelago. In S.
Chiu and C. Sand (eds), From Southeast Asia to the Pacific: Archaeological Perspectives on the
Austronesian Expansion and the Lapita Cultural Complex. Taipei: Center for Archaeological Studies
Research, Research Center for Humanities and Social Sciences, Academic Sinica, pp. 141-69.
Szabó, K. and S. O’Connor, 2004. Migration and complexity in Holocene Island Southeast Asia. World
Archaeology, 36 (4): 621-28.
32
Terrell, J.E., 1998. 30,000 years of culture contact in the southwest Pacific. In J.G. Cusick (ed.), Studies in
Culture Contact: Interaction, Culture Change, and Archaeology. Occasional Paper 25. Carbondale:
Centre for Archaeological Investigations, Southern Illinois University, pp. 191-219.
Terrell, J.E., 2002. Tropical agroforestry, coastal lagoons, and Holocene prehistory in Greater Near
Oceania. In S. Yoshida and P.J. Matthews (eds), Vegeculture in Eastern Asia and Oceania. Osaka:
Japan Centre for Area Studies, National Museum of Ethnology, pp. 195-216.
Terrell, J.E., 2004. Island models of reticulate evolution: The ‘ancient lagoons’ hypothesis. In S.M. Fitzpatrick
(ed.), Voyages of Discovery: The Archaeology of Islands. Westport (CT): Praeger, pp. 203-22.
Terrell, J.E., K.M. Kelly and P. Rainbird, 2001. Foregone conclusions? In search of “Papuans” and
“Austronesians.” Current Anthropology, 42 (1): 97-124.
Torrence, R., 1992. What is Lapita about obsidian? A view from the Talasea sources. In J-C. Galipaud (ed.),
Poterie Lapita et Peuplement. Nouméa: ORSTOM, pp. 111-26.
Torrence, R., 2002a. Cultural landscapes on Garua Island, Papua New Guinea. Antiquity, 76 [293]: 766-76.
Torrence, R., 2002b. Thinking big about small tools. In R.G. Elston and S.L. Kuhn (eds), Thinking Small:
Global Perspectives on Microlithization. Archeological Papers of the American Anthropological
Association 12. Arlington (VA): American Anthropological Association, pp. 179-89.
Torrence, R., 2004. Pre-Lapita valuables in Island Melanesia. In V.J. Attenbrow and R. Fullagar (eds),
Pacific Odyssey: Archaeology and Anthropology in the Western Pacific: Papers in Honour of Jim
Specht. Records of the Australian Museum, Supplement 29. Sydney: The Australian Museum, pp.
163-72.
Torrence, R. and T. Doelman, 2007. Problems of scale: Evaluating the effects of volcanic disasters on
cultural change in the Willaumez Peninsula, Papua New Guinea. In J. Grattan and R. Torrence (eds),
Living under the Shadow: The Archaeological, Cultural and Environmental Impact of Volcanic
Eruptions. Walnut Creek (CAL): Left Coast Press, pp. 42-66.
Torrence, R., V. Neall, T. Doelman, E. Rhodes, C. McKee, H. Davies, R. Bonetti, A. Guglielmetti, A.
Manzoni, M. Oddone, J. Parr and C. Wallace, 2004. Pleistocene colonization of the Bismarck
Archipelago: New evidence from West New Britain. Archaeology in Oceania, 39 (3): 101-30.
Torrence, R., C. Pavlides, P. Jackson and J. Webb, 2000. Volcanic disasters and cultural discontinuities in
Holocene time, in West New Britain, Papua New Guinea. In W.G. McGuire, D.R. Griffiths, P.L.
Hancock and I.S. Stewart (eds), The Archaeology of Geological Catastrophes. Special Publication
171. London: The Geological Society of London, pp. 225-44.
Torrence, R. and C.M. Stevenson, 2000. Beyond the beach: Changing Lapita landscapes on Garua Island,
Papua New Guinea. In A. Anderson and T. Murray (eds), Australian Archaeologist. Collected
Papers in Honour of Jim Allen. Canberra: Coombs Academic Publishing, Australian National
Tregoning, P., H. McQueen, K. Lambeck, R. Jackson, R. Little, S. Saunders and R. Rosa, 2000. Present-day
crustal motion in Papua New Guinea. Earth Planets Space, 52: 727–30.
Valentin, F., 2003. Human skeletal remains from the site of Lapita at Koné (New Caledonia): Mortuary and
biological features. In C. Sand (ed.), Pacific Archaeology: Assessments and Prospects. Le Cahiers
de l’Archéologie en Nouvelle-Calédonie 15. Nouméa: Service de Musées et du Patrimoine de Nou-
velle-Calédonie, pp. 285-93.
White, J.P., 1992. New Ireland and Lapita. In J-C. Galipaud (ed.), Poterie Lapita et Peuplement. Nouméa:
ORSTOM, pp. 83-90.
White, J.P., 2007. Archaeological Studies of the Middle and Late Holocene, Papua New Guinea. Part I.
Ceramic Sites on the Duke of York Islands. Technical Reports of the Australian Museum, 20: 3-50.
Published online: www.australianmuseum.net.au/pdf/publications/1473_complete.pdf
White, J.P. and J.E. Downie 1980. Excavations at Lesu, New Ireland. Asian Perspectives, 23 (2): 193-220.
33
White, J.P., T.F. Flannery, R. O’Brien, R.V. Hancock and L. Pavlish, 1991. The Balof Shelters, New
Ireland. In J. Allen and C. Gosden (eds), Report of the Lapita Homeland Project. Occasional Papers
in Prehistory 20. Canberra: Department of Prehistory, Research School of Pacific Studies, Australian
National University, pp. 46-58.
34
Chapter 3
A REVIEW OF SOLOMON ISLAND ARCHAEOLOGY

Richard Walter and Peter Sheppard
INTRODUCTION
Archaeologists have had some difficulty synthesising the Solomon Island archaeological record
and this is not altogether surprising given the scale of the endeavour. Parts of the Solomon Islands
have been settled for 30,000 years and comprise hundreds of inhabited islands totalling 28,500
square kilometres of land scattered over 1.3 million square kilometres of ocean. With a multitude
of different ethnic and culture groups and more than 70 extant languages, it is easy to see why
archaeologists have found the task of constructing a Solomon Islands’ prehistory daunting. The
purpose of this paper is to review the history of archaeological research in the Solomon Islands
and to use the results of this work to develop an updated regional synthesis.
GEOGRAPHY AND CULTURAL HISTORY

The main Solomon Islands consist of a double island chain stretching from Bougainville in the
northwest to Makira (San Cristobal) in the southeast (Figure 3.1). The northern chain of Choiseul,
Santa Isabel and Malaita is separated from the southern chain of New Georgia, Guadalcanal and
Makira by the Central Solomons Basin or “The Slot”. Bougainville is politically part of the
Autonomous Region of Bougainville within Papua New Guinea but we include it, and adjacent
Buka, in this review since the people of Bougainville and the Western Solomons have been inter-
acting continuously for thousands of years and their prehistories are intimately connected. In fact,
until late in the Pleistocene many of the larger islands of the Solomon Group were joined with
Bougainville to form a single landmass known as Greater Bougainville or Greater Bukida
(Flannery 1995).
The main islands of Bougainville, Vella Lavella, New Georgia, Kolombangara, Choiseul,
Santa Isabel, Guadalcanal, Malaita and Makira are tropical-rainforest clad and mountainous.
They have steep central ridges cut by river valleys running on to a narrow coastal belt. Only
Guadalcanal and Bougainville have extensive areas of coastal plain. The other islands consist of
volcanic and raised reef islands, atolls, cays and islets. Large areas of inshore lagoon, which are
formed by strings of off-shore barrier islands, occur along the coasts of some of the large islands,
especially Malaita and New Georgia. The islands of the Solomon Group are mainly inter-visible
and separated by relatively sheltered waters so canoe travel between islands and island groups has
always been comparatively easy. Nineteenth century accounts suggest that canoe trips of over
200 km were fairly common (Bathgate 1985).
The people of the Solomons are diverse in character and language. The most universally
recognised cultural distinction is that made between “salt-water” and “bush” people (Roe 2000).
Bush people live inland and their life and traditions centre on the forest and gardens, while the
“salt water” people live on the coast and have a maritime based lifestyle. Today in many areas
such as the Western Solomons, where the interiors of islands are no longer inhabited, this distinc-
tion is remembered in oral tradition. The Solomon Islands has considerable linguistic diversity.
Between 70 and 80 languages are spoken today, most of which are Austronesian (AN) (Tyron and
Hackman 1983) (Figure 3.3). The remaining languages are non-Austronesian (NAN) or Papuan
languages. The internal relationships of the NAN languages of the Solomon Islands are not
35
CHANGE THROUGH TIME
36
Figure 3.1 Map of Solomon Islands.
A Review of Solomon Island Archaeology
Figure 3.2 Greater Bukida or Greater Bougainville.
entirely understood (Ross 2001), but it is generally assumed that NAN speakers first settled the
Solomon Islands and that the later expansion of AN languages commenced about the time of
Lapita colonisation. Recent research (Dunn et al., 2005) has suggested a break-up of the Solomons
NAN languages about 10,000 BP. One area where the proposed sequence has been considered to
be possibly reversed is the Reef and Santa Cruz Islands (Ross 2001: 310, Sheppard and Walter
2006a), however the most recent linguistic study suggests this is not the case and that there is no
evidence for a NAN intrusion into this region (Ross and Næss 2007). In a recent Ph.D. Doherty
(2007) came independently to the same conclusion after reviewing the archaeological and ethno-
graphic evidence (see also this volume). First settlement there was by Lapita peoples who brought
AN languages. Although the NAN followed by AN settlement sequence is the most parsimonious
explanation for the linguistic geography of the main Solomons, it should not be forgotten that
interaction and not isolation is the norm in this part of the Pacific. This emerges from the linguis-
tic data, but is equally apparent in the archaeological data and oral history (Sheppard, Walter, and
Roga in press) which show that for millennia Solomon Islanders have maintained complex
patterns of interaction. This has resulted in the formation of culture areas which are defined ethno-
graphically and archaeologically in the patterned distribution of material culture traits, language,
and religious and cult practices.
37
Figure 3.3 AN Languages of the Solomon Islands.
From an archaeological perspective the Solomon Islands occupy a vital position in Oceanic
cultural history. The Northern Solomons mark the eastern limit of known Pleistocene expansion
in the Pacific and the Eastern Solomons mark the boundary of pre-Lapita (mid-Holocene) settle-
ment. The archipelago provides an area where key questions in Pacific prehistory can be
addressed. These include understanding the nature and limits of pre-Austronesian expansion and
adaptation, the dispersal of early Lapita communities and their interaction with non-Lapita
groups, the initial colonisation of Remote Oceania and the origins of Polynesian society, and the
evolution of late period Melanesian diversity. Yet despite the obvious strategic importance of the
Solomon Islands, archaeologists have been unable to develop more than a very preliminary
account of Solomon Islands’ prehistory (Foana‘ota 1996, Green 1977, Spriggs 1997, 2005), and
several authors have lamented the paucity of published material (Kirch 1997, 2000, Spriggs 1997,
2000). Indeed, the archipelago has been described as one of the least well understood regions of
the Pacific (Kirch 2000: 131). There seems to be a view that the apparent gap in knowledge is due
to insufficient archaeological effort. But we suggest that this is misleading – in fact, a great deal
of archaeology has been done in the Solomon Islands over the last 40 years and many parts of the
record are extremely well covered (Table 3.1). There is excellent coverage of Lapita in the regions
38
where it is best represented (in the Southeast), and there is now a substantial body of data on the
ceramic record throughout the Solomons. Sequences have been established for the Shortlands,
parts of the Western Solomons, the Reef/Santa Cruz and some of the Polynesian Outliers, and
while the record elsewhere is patchy in terms of any integrated synthesis, abundant primary
material is available in the form of graduate theses and unpublished reports.
Table 3.1 Historical summary of major research initiatives in the Solomon Islands.
continued next page
39
THE ARCHAEOLOGY OF THE SOLOMON ISLANDS

The data from the various projects described in Table 3.1 can be ordered chronologically, and
arranged to fit into four major themes. These are: Pleistocene colonisation and adaptation, the
early to mid-Holocene record, the arrival of Austronesians, and late prehistory and the develop-
ment of traditional Solomon Island societies. We discuss the following archaeological record in
terms of these four themes.
40
Pleistocene Colonisation and Adaptation

The Pleistocene phase of Solomon Island prehistory is currently represented by a single site, Kilu
Cave (Site DJA), located in the raised limestone Sohano formations of southeast Buka (Figure
3.4). The Pleistocene deposit at Kilu, Layer II, dates from 29,000 to 20,000 B.P. (Wickler 2001:
33, Wickler and Spriggs 1988).
Figure 3.4 Map showing location of sites in the Northern Solomons referred to in the text.
41
Kilu was occupied late in what Leavesley and Allen (1998) refer to as the first pulse of
human activity in the region. This first pulse represents the initial colonisation of Sahul at periods
of low sea levels with possible first occupation of New Ireland just before 40,000 B.P. (Leavesley
et al. 2002). The settlement of the Solomon Islands from the Bismarcks involved a sea crossing
of up to 180 km – perhaps the longest ocean crossing in the world to that date. There is no
evidence to suggest that communication continued back to the west.
The earliest Pleistocene settlers of the Bismarcks appear to have been small, highly mobile
groups of hunter-gatherers (Allen, Gosden and White 1989). As settlers moved eastward into the
Pacific they faced a steady decline in both marine and terrestrial biodiversity, and by the time they
reached Buka many of the fauna and flora available in New Ireland would already have dropped
out of the inventory of potential food sources. Thus the Pleistocene forests of the Solomon Islands
would have been, at best, marginal for a terrestrial hunting and gathering economy. It was for this
reason that, prior to the discoveries on Buka, archaeologists had assumed that settlement of the
Solomon chain would have been delayed until the development of food production systems in the
mid-Holocene. The picture now from Kilu Cave is quite different.
The terrestrial vertebrate assemblage from Layer II at Kilu Cave documents a forest hunting
regime including the exploitation of bats, reptiles, rats, and birds from at least six families
including a megapode, a rail and several species of pigeon (Wickler 2001: 224). The marine shell
and vertebrate remains reflect an inshore fishing and foraging strategy. Kilu Cave is not particu-
larly early for shellfishing, but the range of vertebrate taxa implies a more advanced fishing
adaptation than is usually allowed for Pleistocene Oceania. Allen (1993: 144) has argued against
inferring a complex technology to account for the Pleistocene fishbone assemblages from New
Ireland sites. But a high relative frequency of benthic feeders in the families Labridae and
Serrranidae plus pelagic predators such as Carangididae spp at Kilu Cave certainly suggests an
advanced technology. These families are normally caught in Oceania using hooks or lures and,
while they could no doubt be taken in other ways, we would not expect them to have been
significant components of a fishing system based on generalised foraging methods.
The flake stone toolkit from Kilu Cave has provided the biggest upset to existing models of
Pleistocene subsistence in the Pacific. The tools themselves are quite simple and are made of a
range of coarse-grained volcanic stones, quartz, calcite, chert and limestone. Traces of plant
residues adhering to some of the flake edges contain starch grains and raphides including starches
of Colocasia esculenta and Alocasia sp from the earliest levels of Layer II (Loy, Spriggs and
Wickler 1992). Use-wear studies suggest that the tools were used for the processing of raw taro
corms (Wickler 2001: 234).
At Kilu Cave the Pleistocene levels show that by 29,000 B.P. a set of adaptations had
developed that allowed humans to expand into the true Oceanic environments of Island Melanesia
where the bands of opportunity for non-agriculturalists were even more narrow than in the
Bismarcks. The returns for plant gathering in the Pleistocene forests of Island Melanesia would
have been low, and this raises the issue of whether these early settlers were harvesting wild taro
or whether they were involved in some form of early plant management activities. Wild stands of
Colocasia would have been present in the Buka region when the first settlers arrived, but domestic
taro has larger and more edible corms than the wild varieties and these only developed as a result
of long-term human selection. Spriggs (1997: 38) suggests that the amount of residue present on
the Layer II tools is an indication that this selection process had already started by early in the
Kilu Cave occupation, which could imply that a key to early survival in the Solomon Islands was
the active management of natural stands of Colocasia and other pre-domesticates. This would
42
involve such activities as the expansion and improvement of optimal growing zones by removing
competitive plant species and increasing available light levels through bush clearance around
swamps and river edges. These types of activities have been described as a form of “cultivation
without domestication”, a possible precursor of indigenous Melanesian plant domestication
(Spriggs 1997: 32).
Figure 3.5 Map showing location of archaeological sites and core locations in the Western Solomons
referred to in the text.
43
The question remains as to how far down the Solomon chain Pleistocene settlement
advanced. Kilu Cave was occupied continuously (in archaeological terms) for 10,000 years so
there was clearly no economic barrier to further expansion. Neither was there any significant
technological barrier to overcome – indeed the settlers could have travelled as far southeast as
modern Makira without moving far from the coasts of the Greater Bukida landmass. There is, of
course, a site visibility problem – there would never have been a great number of sites and many
of those on the coast would have been submerged during late Pleistocene sea level rises. The most
likely place to look for sites of this period is along the raised reef terraces like those found on the
Huon Peninsula and New Ireland which preserve the earliest record of Pleistocene occupation in
Papua New Guinea. The authors have carried out reconnaissance surveys and small-scale excava-
tions on a number of the raised terraces of northeast coastal Rendova (Figure 3.5). Excavations
in one rockshelter (Girhu Moho) on the third terrace at a height of circa 120 m above sea level
located a water-rolled volcanic cobble at a depth of 1.87 m in a highly degraded limestone clay.
Datable materials have yet to be recovered from sites along the Rendova terrace series but further
survey work there would be desirable. Much of the New Georgia Group has experienced consid-
erable uplift (Mann et al. 1998), and our survey has included rockshelters and upland areas at a
variety of heights throughout the Group, but to date we have not found any cultural deposits of
any significant age.
The Pleistocene levels at Kilu Cave terminate around 20,000 B.P. which coincides with the
Last Glacial Maximum (22,000 – 20,000 B.P.), a period when settlement pattern change appears
to have been occurring widely in Island Melanesia. These generally involved a decline in the
movement of people between scarce resources, to one in which resources were moved to people
(Gosden 1995). In the Bismarcks this is marked by the appearance of introduced animal species
in the midden record and an increase in the transport of industrial raw materials such as New
Britain obsidian from the sources at Talasea and Mopir. Spriggs (1997: 61) also suggests that by
20,000 B.P. the economies of Island Melanesia might have shifted into a hunting-horticulture
regime foreshadowing the emergence in the early Holocene of true production systems.
Other than the abandonment of Kilu Cave after nearly 10,000 years of intermittent and low
density occupation with the implication of a shift to new patterns of resource use, there is no
direct evidence for any of these purported changes in the Solomon Island record. The New Britain
obsidians and the commensal animals do not appear until the Lapita period and the only hint of
continued contact with the west is the appearance of Canarium charcoals at 10,400 B.P. at Kilu.
This suggests an introduction of Canarium sometime in the terminal Pleistocene. Overall,
however, the evidence points to the Solomon Islands having been largely isolated from develop-
ments in the west during most of the Pleistocene and with coastal movement restricted to within
the Solomon region – and perhaps to within the Northern Solomons.
The Early to Mid-Holocene Record

From the end of the Last Glacial Maximum changes in Solomon Island coastal geomorphology
brought about new opportunities for human expansion and exploitation. The general model of
change involved sea levels rising with the retreat of the northern ice-sheets to reach their highest
stands between 5500 and 6000 B.P., followed by a period of sea level decline as a result of hydro-
static adjustment. But the specific history of coastal change varied greatly from region to region
and was the result of both sea level changes and vertical tectonic movement (Mann et al. 1998:
273 see also Sheppard and Walter, this volume). There were two major changes in coastal
geomorphology that would have impacted on human settlement. The first of these was the separa-
tion of the greater Solomon landmass (Greater Bougainville) into a dispersed island chain. This
44
loss of landmass was offset by the creation of hundreds of kilometres of new coastlines and
coastal reefs, and eventually, by extensive coastal lagoon systems. The second was the creation
of sheltered coastal flats which are now the main focus of settlement throughout most of the
Solomon Islands. Initially, as sea levels rose steadily, the coral reefs struggled to keep pace.
Narrow fringing reefs built up from the shore but dropped directly into deep water offering little
protection from the elements. However through time, coastal benches developed behind sheltered
lagoons and the modern coastal landscape of the Solomon Islands emerged.
The early to mid-Holocene is not well represented archaeologically in the Solomon Islands
but there are three areas where we can catch a sufficient glimpse to reconstruct a broad outline
of events.
Buka and the Northern Solomons
The earliest Holocene deposits are at Kilu Cave where Layer I spans the period 10,000 to 5000
B.P. From Layer I there is more midden than in the Pleistocene levels and the density of fireplac-
es increases, suggesting that the site was more intensively occupied. Flake tools were present in
the earliest Holocene levels, but from about 8000 B.P. flake production seemed to decline and
ground stone tools, shell tools and shell ornaments appear. Examples of these new artefact types
include an anvil stone, a fragment of a Tridacna shell adze, perforated shark teeth and vertebrae,
and a Terebralia shell chisel or abrader (Wickler 2001: 39). Although direct dating of shell
artifacts from the site suggest some intrusion into older levels, conus pendants which are often
thought to be Lapita associated are dated here to 7580 BP (Spriggs 2001). Changes in subsistence
are indicated by the palaeobotanical record. Charcoal remains of coconut (Cocos nucifera) and
Canarium (both C. indicum and C. solomonense) have been confidently identified from early
Holocene horizons (Wickler 2001: 234). Coconut was most probably a natural colonist but
Canarium was a human introduction from the New Guinea mainland. The evidence suggests the
introduction or expansion of tree cropping.
The tail end of the Holocene sequence from Kilu overlaps the earliest lenses at Palandraku
Cave (Site DBE) located 250 m to the north. The DBE site was first excavated by Specht in 1967,
and again, 20 years later, as part of Wickler’s research programme (Wickler 2001). The pre-ceram-
ic horizons date to about 5000 B.P. and contain a low density of cultural material interpreted as
representing a temporary and non-intensive occupation (Wickler 2001: 42). The artefact assem-
blage includes un-retouched stone flakes, similar to those found in Kilu Cave, dominated by chert,
and with examples also of quartz, calcite and fine-grained volcanic stone. The faunal remains point
to a continuation in the hunting of small forest-game including the use of endemic reptiles, bats
and rats. An earth oven was also found in the pre-ceramic levels. Phalanger orientalis, the intro-
duced rat species and the faunal domesticates do not appear until the later Lapita horizons.
New Georgia and the Western Solomons
The Western Solomons, and particularly the New Georgia Group, contain some of the most
extensive lagoon systems in the Solomon Islands, where they typically consist of long, shallow
inshore lagoons lying between a string of barrier islands and a mangrove-lined shoreline. These
rich, sheltered habitats made their appearance from the early to mid-Holocene and would have
been a major attractor for expanding coastal populations. There are no known early to mid-Holocene
archaeological sites in the Western Solomon islands, but the formation history of the lagoon
systems of the New Georgia Group holds a potential key to understanding Holocene expansion
into the Solomon Islands, and early adaptation in the region. As part of the New Georgia
Archaeological Survey programme directed from 1996 – 2000 by the authors, swamp cores were
45
taken from a number of locations in Roviana Lagoon documenting coastal change and vegetation
history in Holocene New Georgia. The work was undertaken as Ph.D. research by Sarah Grimes
(University of Western Australia, see Grimes 2003) under the direction of John Dodson.
The Roviana coring programme targeted the mangrove shores of the barrier islands, swampy
river valleys on the mainland coast, and shallow lake beds on islands along the edges of both
Roviana and Vonavona Lagoons, and on the southern side of Rendova Island (Figure 3.5). The
results show that during the Pleistocene low stand coastal plains extended out to meet a narrow
fringing reef. These plains were cut by rivers which emptied through what are now the major
passages between the barrier islands. By just before 6000 B.P. the plains were flooding to reach
a stand slightly higher than present at about 4500 – 4000 B.P. From around 4000 B.P. sea levels
started to fall, at least partly as a result of regional tectonic uplift. This led to the formation of
many of the barrier islands in Roviana and Vonavona Lagoons, and by about 2600 B.P. the current
marine landscape was formed. These results are discussed in more detail in Grimes 2003 and are
summarised in Table 3.2.
Table 3.2 Coastal changes in New Georgia based on palaeoenvironmental

research in Roviana Lagoon (Grimes 2003).
According to this model of environmental change, from around 4000 B.P. the inshore
lagoon systems started to develop behind well formed barrier islands. Although there is no
archaeological evidence of human settlement at this time, these geomorphological changes
would have made the coasts of New Georgia increasingly attractive to human colonists. This
situation from Roviana is likely to have been mirrored in the other lagoon systems of the Western
Province and perhaps also in Malaita.
Guadalcanal and the Central Solomons
The longest continuous archaeological sequence reported so far from the Solomon Islands east of
Buka has been constructed by David Roe for Guadalcanal. The sequence is based on only a small
number of sites and on limited excavation data from the northwest of the island. But the data is
important because the first phase of Roe’s sequence (the Hoana Phase) is based on an analysis of
the only known material of mid-Holocene age east of Buka. It also provides a useful, although
patchy, record of cultural and environmental interactions just prior to the arrival of Austronesians.
The Guadalcanal sequence is based on excavations of cave and rockshelter deposits in the
Poha and Vura Valleys, and on a single open settlement in the Visale area of northwest
Guadalcanal. It also draws on palaeoenvironmental data, especially the palynological work of
46
Simon Haberle (1993, 1996). The earliest cultural horizons are from the Vatuluma Posovi site
(also known in the literature as Poha Cave or Fotoruma Cave) located about two kilometres inland
in the Poha Valley (Figure 3.6).
The site was located in the 1960s and first excavated by William Davenport in 1966 (Davenport
1968), and then by amateur archaeologists Tom Russell and James Tedder (Davenport, Russell and
Tedder 1968). The Russell and Tedder excavation recovered dates from the basal deposits suggest-
ing occupation about 3000 B.P., but removed nearly all the intact material (Davenport 1972,
Davenport et al. 1968). Roe returned to the site in 1987 – 1988 and succeeded in locating a small
remnant portion of cultural deposit near the front of the cave (Roe 1993). On the basis of a study
of the remnant material, a reinterpretation of previous unpublished reports and other documents,
and an examination of the artefact collections returned to Honiara by Davenport, Roe argued that
the site contained five occupation phases with Phase 1 spanning the period 6500 to 4000 B.P.
Figure 3.6 Map showing location of sites in the Central and Eastern Solomons referred to in the text.
47
Phase 1, the mid-Holocene Phase, was defined by two dates (ANU-6733 and ANU-6744)
which bracketed a deposit containing “…small quantities of artefacts – chert flakes and chips, a
single Trochus armband fragment and Trochus fishhook blanks – and quantities of terrestrial
animals and marine fish and shell midden of predominantly freshwater species…” (Roe 1993:
66). In fact, the artefacts are more closely associated with the stratigraphically higher sample
(ANU-6734 from Layer 3 I) which calibrated at one standard deviation to about 4500 B.P. (Roe
1993: 63). As far as we can determine, no artefacts or other cultural materials were recovered
from Layer 3 III, which is dated by ANU-6733 to about 6200 B.P. (calibrated at one standard
deviation) (Roe 1993: 63). Thus a conservative interpretation of Phase 1 might place occupation
at Poha Cave closer to the upper end of the band (perhaps around 4500 B.P.) than to the lower.
From further up the Poha Valley, the Vatuluma Tavaro Cave site has produced a sequence
that might overlap part of that from Vatuluma Posovi. Phase 1, dated to around 4200 B.P.,
includes a single Trochus arm ring and an assemblage of chert chips and flakes.
Drawing on both the archaeological and palaeoenvironmental record, Roe described the settle-
ment and subsistence practices of the Hoana Phase as representing the exploitation of “relatively
undisturbed and undeveloped forest and the collection of molluscan foods from fresh and brackish
water environments” (Roe 1993: 95). The faunal assemblage includes endemic rats and reptiles and,
together with the molluscan faunas, this is indicative of a generalised forest edge and riverine
hunting and gathering system. Canarium nuts were harvested, but nut processing stones have not
yet been identified. The artefact assemblage from the Hoana Phase includes chert flakes, Trochus
shell arm rings and a few shell items which might be fishhook blanks. These occur from the earliest
levels while polished stone adzes and shell beads appear sometime around 3000 B.P.
The Arrival of Austronesians

Much of the archaeology of Island Melanesia has been concerned with the arrival and spread of
Austronesian speaking peoples, especially the Lapita peoples who had both the sailing technology
and economic strategies that allowed them to settle the remote and relatively depauperate islands
of Remote Oceania. Lapita makes its first appearance in the west, in the Bismarck Archipelago,
about 3400 B.P., so we might expect to see evidence of Lapita occupation in the Solomon Islands
some short time after that period. In Buka and the North Solomons the archaeological record
enters a hiatus about a millennium earlier and there is no further evidence of occupation until the
appearance of a late Lapita tradition at about 2700 B.P. Roe’s Hoana Phase from Guadalcanal does
indicate occupation of the Poha Valley during the period when Lapita was appearing in New
Britain, but these sites show no evidence of ceramics, and any connection they may have with the
arrival of Austronesians is likely to date to much later in the Hoana Phase (see below). The earliest
record for the appearance of Lapita ceramics in the Solomon Islands is in the Reef/Santa Cruz
Islands of the Southeast Solomons (Green, Jones and Sheppard 2008). The Lapita sites of the
Southeast Solomons were first explored during the 1970s and there were expectations that Lapita
sites of similar age or earlier would one day be found to the west, in the Northern or Western
Solomons, reflecting an inland-hopping model of Lapita migration (Green 1979). Such sites have
not been located and the model for Lapita colonisation of the Solomon Islands supported here is
based on the view that the current distribution of ceramic sites and associated radiocarbon dates
is approximately accurate. The following sections review the ceramic record and evidence for
Austronesian and Lapita settlement across the Solomon Islands.
48
The Northern Solomons

The Northern Solomons have long been known to have a long ceramic sequence and during the
historic period pottery manufacture was important in each of the islands or island groups. This
region has also benefited from considerable systematic research. Irwin (1972) described a ceramic
sequence extending back some 1100 years from survey and excavation in the Shortland Islands,
Terrell (1970, 1976, 1977, 1986, Terrell and Irwin 1972, see also Black 1977) ran a major research
programme on Bougainville which has more recently been augmented by Spriggs (1992,
1997: 171), and Specht (1969) provided the first, and still the most complete, ceramic sequence
based on his extensive survey on Buka, which has also been augmented and considerably
extended by the work of Wickler (2001). In all of these areas ceramic bearing sites were
commonly found, and in some locations (e.g., the north Bougainville coast) create near continu-
ous coastal deposits extending over many kilometres (e.g. Terrell 1970: 4, Wickler 2001: 19).
Despite the amount of research and the abundance of ceramic deposits, it was not until research
in the inter-tidal zone by Wickler on Buka that sites bearing abundant dentate ceramics were
found, although it should be noted that early excavations by Specht (1969: 194) on Buka produced
two dentate-stamped sherds.
The three inter-tidal sites (DAA, DAF, DJQ) containing dentate Lapita ware on Buka are
undated, but Wickler (2001: 241) suggests dates on stylistic grounds of between 3000 to 2800 B.P.
(DJQ) and finishing around 2300 to 2100 B.P. (DAF beach and inner reef). The DJQ site has over
56% (n=188) of decorated sherds dentate-stamped, however this falls sharply to 1.9% (n=77) at
DAF (Wickler 2001: 108) with dentate-stamping replaced by increasing proportions of unbounded
incision (Wickler 2001: 112), punctation and appliqué relief. Although dentate-stamping seems to
virtually disappear on Buka at this point, many of the characteristics of Lapita ceramic technol-
ogy and design lasted into what Specht named the Buka ceramic style. Wickler noted that the
Buka style phase appeared to mark the movement of Lapita settlement onto land where it appears
in a number of sites. Specht dated this style to the later part of the first millennium B.C. (Specht
1969: 214), and Wickler (2001: 6) suggests a span of 2500 to 2220 B.P. He labelled this phase the
“Late Lapita Phase” although he was unable to date it or the start of the Sohano ceramic style that
appears to closely follow the Buka style.
No Buka or Late Lapita style ceramic assemblages have been found in the rest of the Northern
Solomons. Survey and excavation was conducted by Terrell’s Bougainville Archaeological Survey
over multiple seasons in the Silao, Teop, Numa Numa, and Paubake (Buin) areas (Black 1977,
Terrell 1970, 1976, 1986: 222-239). More recently, Spriggs followed up Terrell’s work on
Bougainville with a study in Kieta district in south Bougainville, from which he reports a 1500 year
sequence at the Sivu Rockshelter on the Bougainville Penninsula (Spriggs 1997: 171), and in Eivo
to the north of Kieta where he recorded a parallel but shorter sequence (Spriggs 1992: 418).
The Western Solomons
Ceramic bearing sites have been reported nearly everywhere in the Western Solomons where
archaeological survey has been carried out, including most recently eastern Santa Isabel (Roe
pers. comm. 2007). During the historic period pottery was manufactured on northwest Choiseul,
which has close historical and cultural connections with the Shortland Islands and southern
Bougainville. Both these regions also practiced ceramic manufacture. Many of the ceramic sites
reported in the Western Solomons are inter-tidal sites including those few that have included a
small number of dentate-stamped ceramics (see Sheppard and Walter this volume).
49
In his early survey work in the Solomon Islands Chikamori (1967) reported ceramics on
Choiseul, and Yen later reported sherds on Kolombangara (Miller 1979: 148, Yen 1971). During
the National Site Survey programme led by Miller (1979, Foana‘ota 1979 ) ceramic find spots
were reported on Simbo, Kolombangara and Choiseul, including Vaghena Island off Choiseul’s
eastern tip. Reeve (1989) reported inter-tidal ceramic sites on Vella Lavella and New Georgia,
including in North New Georgia. The most common decorative technique reported was shallow
incising, although appliqué, punctuation, and notched rims are also mentioned. The Paniavili site
located in Roviana Lagoon in New Georgia was first visited by Spriggs and Reeve and reported
by Reeve (1989). The ceramic assemblage from Paniavili did not contain any dentate-stamped
sherds but did include incised, pinched, fingernail impressed, punctuate, and applied relief wares.
In addition, Tridacna adzes, a Cassis shell chisel, a Tridacna arm ring, and a range of Conus shell
rings and bracelets were also found in the same inter-tidal zone as well as a piece of obsidian.
None of these materials are necessarily contemporary with the ceramics but the nature of the
assemblage led Reeve (1989) and Spriggs (1997) to suggest that much of the inter-tidal deposit
dated to a late or immediately post-Lapita time period, perhaps dating to around circa 2500 B.P.
Since the work of Reeve, we have returned to Roviana Lagoon, relocated the Paniavili site and
conducted survey throughout the length of the lagoon (Sheppard et al., 1999). This has resulted
in the location of some 20 similar sites from Araroso Point in the east to Nusa Roviana at the
western end of the lagoon. We have located similar sites in the inter-tidal zone in Marovo Lagoon
(Seghe Channel, Chea, Marovo Island), off the northern end of Ghizo Island, where we also
recovered an obsidian flake and in northern Vella Lavella (Irigila region). Recently Summerhayes
and Scales (2005) have reported similar assemblages from Kolombangara.
An intensive study of the western end of the lagoon formed the basis of Ph.D. research by
Matthew Felgate (2003). Felgate’s study of sherd taphonomy and distribution demonstrated that
the sites of the early Roviana ceramic period were stilt house settlements like those reported in
Lapita contexts on Mussau (Kirch 1987), Buka (Wickler 1995) and Nissan (Spriggs 2000: 355).
Felgate (2003) also reported the first dentate-stamped ceramics from two sites (Honiavasa and
Nusa Roviana) in Roviana Lagoon and provided a radiocarbon-based chronology for the New
Georgia ceramic sequence. Charcoal from within a sherd from Panaivili was dated to 2130±90
b.p. and an exterior smoke-derived carbon deposit on a sherd from the Hoghoi site was dated to
2619±45 b.p. (Felgate 2001: 48). The sherds from these sites contain rectilinear incised, punctuate,
and pinched decoration as well as a variety of lip treatments including crenulated or horizontally
deformed lips (Felgate 2003), and confirm the late Lapita age of the earliest ceramics from New
Georgia suggested by Reeves and Spriggs (Reeve 1989, Spriggs 1997) for Paniavili. Felgate
reports a shift from complex pot forms in the earlier Honiavasa site to simpler forms and often
thinner wares in the post-Lapita sites (Felgate 2001: 53) dating to the mid- to late first millennium
B.C. This suite of ceramic styles and apparent sequence is found in inter-tidal sites throughout
much if not all of the Western Solomons.
The palaeoenvironmental record of Roviana indicates a period of vegetation and sedimen-
tary changes starting after 4000 B.P. (Grimes 2003: 237). Evidence from pollen cores sampled in
Roviana Lagoon suggests a number of periods of burning from about 3750 B.P. followed, around
3400 B.P., by increases in ferns and a decline in woody vegetation. Between 3200 and 2750 B.P.
there are records of further burning and erosion events. Although Grimes suggests that this might
be evidence for the arrival of Lapita colonists from around 3200 B.P., it is difficult to distinguish
these changes from natural events. It is only from 2600 B.P., coincident with the first appearance
of pottery, when the vegetation and sediment data can be said to clearly represent “...an anthro-
pologically-modified landscape” (Grimes 2003: 231).
50
Our interpretation of the archaeological and palaeoenvironmental records is that the estab-
lishment of hundreds of kilometres of sheltered lagoons with abundant and easily won fisheries
from 4000 B.P. was a major attractor for human settlers. It is very likely that people moved into
and through coastal New Georgia and impacted on landscapes and vegetation patterns from about
this time. But there is no evidence archaeologically for any sustained occupation until the appear-
ance of ceramics around 2600 B.P., at which time an inter-tidal settlement pattern based on stilt
houses was introduced to the region, and vegetation changes of clear anthropogenic origin are
recorded. The inter-tidal ceramic sites are the beginning point for the continuous occupation
record for Roviana Lagoon and represent a settlement pattern similar to that noted in various
locations in the Bismarck Archipelago during the Lapita Period. However, in the Western
Solomons this was associated with a Late Lapita settlement that occurred after the loss of the
dentate-stamped ceramic tradition (Sheppard and Walter 2006a).
The Central and Eastern Solomons
The most striking characteristic of the Lapita time period in the Central and Eastern Solomon
Islands is the lack of decorated Lapita pottery (dentate-stamped or incised), or for that matter
virtually no ceramics of any period. This could be a consequence of limited investigation, as
argued by Felgate (2003), but as Sheppard and Walter (2006a) point out, this argument no longer
seems plausible. The amount of survey in a number of regions within the Eastern and Central
Solomons is comparable or exceeds that in the Reef/Santa Cruz region. This includes survey and
excavation by Whitney, Green, Ward and Hendren on Ulawa (Yen 1976), survey and excavation
on the small isolated island of Uki by Green (1976a), survey and excavation on the small island
of Santa Ana by Davenport (1972), Swadling (1976), and Green (Black and Green 1975), and
survey and excavation on the Star Harbour Peninsula of Makira (Green 1976b).
Subsequent to this research the Solomon Island National Site Survey was initiated by Daniel
Miller. This involved several weeks survey and excavation in the Arosi District of Makira (Miller
1979: 87-109) and occasional subsequent survey by members of the National Museum (John Keopo
pers. com. 1999). Although archaeological research on Malaita is limited (Miller 1980), many anthro-
pologists have conducted long-term research on Malaita, and although some have paid attention to
stone artefacts (e.g., Ross 1970), none have reported finding any ceramics. If we compare this with
the situation in the Southeast, Northern and Western Solomons where virtually every survey has
produced ceramics of some sort, it is reasonable now to conclude that the main Solomons were
mostly aceramic, and almost certainly not part of the initial wave of colonisation by bearers of the
dentate-stamped Lapita tradition. The only ceramics recovered so far from the Central Solomons date
to the very end of the Lapita sequence. These are from a number of rockshelters on Santa Ana
including the Feru II rockshelter which was excavated by Davenport (1972). In the middle of a
deposit which contained cultural material to a depth of 216 cm, his team recovered a small sample of
plainware including a rim sherd with a serially incised lip (Davenport 1972: 177). This was dated by
associated charcoals (I-2878) to 1275±105 B.P. Subsequently Green revisited and reinterpreted the
stratigraphy, which he divided into an upper and lower unit divided by a sterile sand (Black and Green
1975: 30, Swadling 1976). Dates from a hearth in the base of the deposit indicated the pottery bearing
layer was created during the first millennium B.C. (SUA-114 3050±70 CRA combined Tridacna shell
sample, SUA-113 2860±250 CRA small charcoal sample (Swadling 2000). Green recovered
additional sherds from Feru II (Swadling 1976: 127) and Davenport reported a rim sherd (undeco-
rated but missing the lip) from the Rate site located in a shallow rockshelter some four kilometres
along the shore from Feru II. This sherd was recovered from a depth of 106 cm and is undated
although a charcoal date from 85 cm (I-2882, 1910±135 CRA) suggests it might also date to the first
millennium B.C. In total, about 20 small plain sherds have been recovered from these sites.
51
Although there is no archaeological support for early Lapita occupation of the Central
Solomons, Roe’s work from Guadalcanal along with the palynological work of Simon Haberle
provides some possible evidence for initial Austronesian arrival or impact in the region.
Roe’s Hoana Phase for Guadalcanal was dated at 6400 – 2200 B.P., although the evidence for
occupation at the earlier end of this bracket is slim (see above). An argument could be made that the
Hoana Phase is better described as a two-phase unit with a dividing line at about 3300 – 3000 B.P.
Within this date range shell beads and polished stone adzes make an appearance, and there is a
change in the sources of chert from predominantly Guadalcanal sources to the inclusion of more
Malaitan material. This date range also coincides with evidence for environmental change inferred
from pollen cores reported by Simon Haberle (1996). The base of two cores from the northern
alluvial plains of Guadalcanal has been interpreted, on the basis of sediment rates, to date earlier
than 3000 B.P. (although only one has actual radiocarbon dates to support a pre-3000 B.P. age). Both
cores show continuous charcoal influx and the Laukutu Swamp core shows appreciable fluctuation
after 3200 and prior to 2100 B.P. The most dramatic and sustained charcoal influx occurs just before
2100 B.P. and after the tephra dated to 2650 B.P. (Haberle 1996: 336). The northern plains of
Guadalcanal support a fire-climax vegetation community which is maintained by regular human
firing. Haberle’s sequence suggests that by 3300 B.P. at the earliest, and most likely after about 2600
B.P., the grasslands and fire were already an integral part of the environment (Roe 1993: 95).
Although there is no direct evidence in the Hoana Phase for agricultural activities (including animal
husbandry), there were “massive” human impacts on the environment in the form of increased forest
clearance and erosion between 2750 – 2150 B.P. (Roe 1993: 119). A very similar pattern indicating
a dramatic rise in charcoal influx at circa 2600 B.P. appears in the Roviana cores. In Roviana this
correlates with the oldest date available for the inter-tidal sites (the Hoghoi site, see above).
The Southeast Solomons
There are 14 known Lapita sites in the Reef/Santa Cruz Islands, which is an extraordinarily high
density of Lapita sites for such a comparatively small area surveyed. Three of the sites excavated
by Green (RF-2 and RF-6 on the Reef Islands and SZ-8 on Santa Cruz) have provided an
important chronological sequence, and an outline of ceramic change in the Southeast Solomons.
The large areal excavation (153.5 m2) at RF-2 provides the most complete excavation of a Lapita
settlement to date and contributes towards the development of a model for the organisation of
early Lapita habitation sites.
The RF-2 site was a hamlet of circa 1100 m2 which contained a small number of structures.
These included one large rectangular structure which was the focus of artefact concentrations, as
well as one or two smaller structures indicated by surface pottery distributions in the northern part
of the site. A spatial analysis of the artefacts carried out by Sheppard and Green (1991) identified
distinct patterns in the distribution of vessel forms when compared to post-hole and other sub-soil
features. These patterns appeared to reflect an association between vessel form and the function
of the structures. The site appears to be of rather short-term occupation with little evidence for the
rebuilding and reorientation of activity areas.
The earliest date for the Southeast Solomons Lapita sites is from SZ-8 at circa 3200 B.P.
(Green et al. 2008). This is closely followed, and overlapped by dates from RF-2 (Jones et al.,
2007), and then at a distance by RF-6. The SZ-8/RF-2/RF-6 sequence has been challenged by
Best (2002), but we consider that this challenge has been adequately addressed (see Green, this
volume) and must now be accepted as robust. It also provides a model of ceramic change
(Sheppard and Green 2007) that is both internally consistent, and congruent with wider regional
models of Lapita ceramic history.
52
The Reef/Santa Cruz Lapita sites have a high percentage of dentate-stamped sherds; always
at least twice that of incised sherds (Sheppard and Walter 2006a: 60). But the SZ-8, RF-2, RF-6
sequence also demonstrates a clear simplification over time in ceramic design – a fall-off in the
cultural force represented by the horizon style – especially between the two early sites and the
later Ngamanie (RF-6) site. Decorating drops out of the ceramic inventory by about 2700-2600
B.P. and is replaced by plainware, with the disappearance of ceramics altogether over the next 500
years. This is consistent with the pattern for Vanuatu where Bedford (2006) suggests the end of
the dentate-stamped tradition occurs between 2800 and 2700 B.P. The one unusual characteristic
of the Reef/Santa Cruz ceramic sequence is the fact that despite the comparatively large number
of sites located, surface collected and excavated, there have been no post-Lapita decorated
ceramics recovered. Instead of a post-Lapita incised and applied relief body decoration similar to
the Mangaasi Tradition in Vanuatu (McCoy and Cleghorn 1988: 114) or the New Caledonia
paddle impressed Podtanėan decoration style (Sand 1996), the ceramic sequence in the Reef/
Santa Cruz region ends circa 100 B.C.-100 A.D. with a plainware tradition of simple pots with at
most impressed lips. This is similar to the revised sequence now understood for much of Vanuatu
which appears to end Lapita with a predominately plainware horizon followed by considerable
regional diversification (Bedford and Spriggs 2008). This contrasts with post-Lapita develop-
ments in the Bismarck Archipelago where the Lapita dentate and incised tradition is gradually
replaced or finishes with an incised, applied relief and punctate decorative repertoire (Garling
2003, Spriggs 1997: 124-125, White and Murray-Wallace 1996).
Along with the ceramic sequence, one of the most important results from the Reef/Santa
Cruz Lapita investigations concerns the transport of exotic lithic materials. The Lapita sites of the
Reef/Santa Cruz Islands stand out in the abundance of stone imported from distant sources.
Sheppard (1993: 127) has estimated that the SZ-8, RF-2 and RF-6 sites contained more than 275
kg of obsidian originating from sources located 2000 km to the west in the Bismarck Archipelago.
A strong case can also be made that this material was transported directly and in repeat voyages
over a period of some hundreds of years (Sheppard 1993, Sheppard and Walter 2006a: 59). Most
of the material comes from the Talasea (97.5%) and Admiralties sources, but a small amount
comes from the comparatively close Banks Islands 400 km to the south, and one piece is sourced
to Fergusson Island off the coast of Papua New Guinea (Green 1987, Green and Bird 1989). By
comparison the Lapita period (Kiki Phase) on the island of Tikopia, which is also in the Southeast
Solomons and dates from 900 to 100 B.C. thus overlapping the Reef Islands Lapita period, has
only three pieces of rhyolitic obsidian sourced to Manus (Spriggs 1997: 137). Although other sites
in Vanuatu may have more obsidian than was previously anticipated (Bedford, Spriggs and
Regenvanu 2006, Galipaud and Swete Kelly 2007), the amounts are still comparatively small.
A Model for Lapita Settlement of the Solomon Islands
It has generally been assumed that the apparently patchy Lapita record for the Solomon Islands
is to a large extent a product of incomplete sampling or other taphonomic processes including
coastal change. According to this assumption, Lapita settlers would have followed an essentially
west-to-east colonisation path settling the Northern and Western Solomons before crossing to the
Central and Southeast Solomons. But in a recent review of the Solomon Islands Lapita evidence
Sheppard and Walter (2006a) put forward an alternative view that took, as its starting point, the
argument that the current picture of site distribution and chronology is approximately accurate;
that it is not fatally flawed by insufficient archaeological survey or biases in site preservation (see
also Sheppard and Walter, this volume). This revised model argued that the Lapita colonisation of
the Solomon Islands involved three basic movements as follows:
53
1. The settlement of the Reef Islands and Santa Cruz occurred during the Early Lapita period. It
was the first settlement of the Solomon Islands by Lapita peoples and involved leapfrogging
the main islands of the Solomon Group. In other words, it involved a direct settlement from
the Bismarck Archipelago. The amount of obsidian from the Bismarcks is indicative of a
special long-distance relationship and is not replicated anywhere else in Remote Oceania.
There is additional genetic support for a model of direct settlement from the Bismarcks.
Mitochondrial DNA evidence summarised in Sheppard and Walter (2006a: 61-62) shows that
the modern genetic makeup of populations in the Reef and Santa Cruz region is more similar
to those of New Britain and to the south (Vanuatu and New Caledonia) than to Solomon Island
populations to the west. Most recently (Ross and Næss 2007) re-evaluation of the Reef/Santa
Cruz Austronesian languages suggests that the closest relationships are to be found in the St
Matthias Group north of New Ireland.
2. The Northern and Western Solomon Islands were settled shortly after 2700 B.P. by an
Austronesian expansion from the west during the Late Lapita period. This established an
inter-tidal settlement pattern based on stilt houses plus the use of a ceramic suite including
incised and appliqué ceramics that spanned Near Oceania from New Ireland to New Georgia.
Evidence for this expansion is found in the ceramic record and is supported by the palaeoen-
vironmental evidence for increased forest clearance around 2600 B.P. in the Central and
Western Solomons.
3. The Central and Southeast Solomons east of the Florida Group were also settled by Austronesians
in the Late Lapita period, but this region was settled possibly not from the west, but by aceramic
populations moving to the west, out of the general Reef/Santa Cruz region. It is this hypothesised
back movement of Austronesian speakers that, we argued, may have resulted in the development
of the Tryon/Hackman Line, a major linguistic division that separates Western Oceanic from
Central and Eastern Oceanic languages (Ross 1989, Sheppard and Walter 2006a: 64). Although
this aspect of our argument has not met with much linguistic support (Pawley 2007), we find
the congruence between the archaeological data and the linguistic pattern intriguing and
indicative of a complex pattern of cultural development and origin in the region which is
potentially grossly simplified by simple wave of advance models.
Late Prehistory and the Development of Traditional Solomon Islands

Societies
The period following Lapita is something of a dark age in many parts of Near and Remote Oceania
and the Solomons Islands is no exception. Where ceramics ceased, and this apparently includes
most of the Solomon Islands, the first millennium A.D. is poorly represented archaeologically.
This is followed by a gradually increasing density of remains until about 500 years ago, from
which time there is a well preserved surface record across most of the islands. This consists of the
remains of stone platforms, shrines, wharfs, old villages, artificial islands, terracing, irrigation
works and fortifications. This record documents the emergence of “traditional” Solomon Island
societies and the formation of the recent social, ethnic and political landscape of the Solomon
Islands. As described by the first ethnographers of the 19th century, this landscape is characterised
by an amazing diversity in language and culture. But what is especially remarkable is that these
high levels of linguistic and cultural diversity have been maintained in the face of pervasive
regional interaction. The study of the emergence of “traditional” Solomon Islands societies and the
development and maintenance of diversity represents a significant challenge to archaeology. In
reviewing the late period record for the Solomon Islands below we concentrate on the key data
types which are both archaeologically accessible and relevant to such a study. These include the
54
archaeological and ethnographic evidence for settlement patterns, monuments, production systems
and exchange media. We use our own research results from the Western Province as a case study
in the emergence of a “traditional” Solomon Island society – the Roviana Chiefdom.
The Northern Solomons
The Northern Solomons contrast with the rest of the Solomon Group in that they maintain a
continuous ceramic record from the Lapita period up to, in places, the present day. Pottery
production appears to have been focused on a series of production centres scattered from Buka
down through Bougainville and across to the Shortlands and northern Choiseul (Blackwood
1935, Irwin 1972, Ogan 1970, Specht 1969), and three major regions or overlapping zones can
be distinguished. These are the Buka Straits region, central Bougainville and the Bougainville
Straits region.
The Buka sequence was first described by Jim Specht (1969) and revised by Wickler (2001)
(see Table 3.3). Wickler’s revisions involved the inclusion of a dentate-stamped Lapita period to
predate the late Lapita “Buka” style, as well as a demonstration of continuity between the Buka
and Sohano styles which Specht initially thought were unrelated (see also Terrell 1976: 230).
Table 3.3 Buka ceramic sequence with revisions based on Wickler (2001: Table 1.1).
The Sohano style differs from the Buka style in pot forms and manufacturing technique, as well
as in decoration: Sohano pottery does not include dentate-stamping and includes punctation as a
more common decorative technique (Specht 1969: 193). Through time, decoration moves towards
an increase in the use of applied relief (Hangan) then of comb incising (Malasang). Ceramics of
the Buka series are distributed throughout Buka and across Buka Passage to the Silao Penninsula
(Terrell 1976: 238). To the north Sohano pottery is reported from Nissan and the abundance of
Malasang to Recent Buka pottery on that island indicates extensive interaction after 800 A.D.
(Spriggs 1997: 196). The Buka sequence is a clear case of a process which is congruent with the
distribution of Lapita ceramics and which involves a gradual loosening of pottery design stric-
tures. This loosening of Lapita strictures may also be seen in the sudden demise in obsidian
exchange which is currently represented by only seven flakes in Sohano deposits (Wickler 2001:
243) and a few flakes in Western Solomon Island sites.
In the historic period the next pottery manufacturing centres south of Buka were in the Kieta
region of central Bougainville and in the Siwai and Buin regions of south Bougainville. In general
there appears to be a broad similarity in early ceramics stretching from Kieta in the north, through
55
Buin to the Shortlands. The early sequence at Kieta was thin and plain tempered (Sivu style)
changing to tempered, thin-incised pottery (Asio style), and is paralleled in the Early to Middle
Period transition in the Shortlands at about 1000 B.P. (Spriggs 1992). In Buin, Terrell’s sequence
from Paubake again starts with thin tempered pottery decorated by incising and appliqué (Terrell
1976: 460). The earliest deposits in this region at present only date back to 1600 B.P. but, based
on the Buka and Western Solomons situation, earlier material (similar to the Buka and Sohano
ceramics) may well be found as inter-tidal sites in lagoon and sheltered bay settings, especially in
the Bougainville Strait region. It is probable that throughout this region the sequence of ceramic
change starts from a common Late Lapita base and that the calcareous tempered plain ceramics
from Kieta, although later in time, are linked in a general sense to the plain calcareous tempered
Buka Ware and the early thin plainware in the Shortlands (Irwin 1972: 236).
The ceramic record of the Northern Solomon Islands suggests the presence of a number of
overlapping interaction spheres that have some time depth. This interpretation is supported by the
ethnographic record and oral history (e.g. Terrell and Irwin 1972). To the north the Nissan-Buka-
Northern Bougainville interaction sphere centred on the Buka Straits, while to the south the
Bougainville Straits interaction sphere extended from southern Bougainville across to the
Shortland Group (Alu, Mono, Fauro) and into northern Choiseul.
The historic settlement pattern of Bougainville and Buka is characterised by almost continu-
ous coastal settlements in the north, scattered hamlets in the mountainous interior, and interior
hamlets on the Buin Plain (Blackwood 1931: 201, 1935: 17, Terrell 1976: 95-120). This is
generally reflected also in the archaeological survey record of north Bougainville and Buka (e.g.,
Specht 1969, Terrell 1976). In the south, 19th century reports comment on the thinly populated
southern coast in contrast to the densely populated Buin Plain. Prior to WWI the fertile Buin Plain
supported small, scattered hamlets and men’s houses joined by a network of paths, and a similar
pattern seems to have prevailed in the Siwai area of southwest Bougainville (Oliver 1949, Scott
et al. 1967, Thurnwald 1934a, 1934b). Most of these hamlets would rarely have held more than
50 people, and political organisation was historically based on the competitive power politics of
Big Men whose influence could provide some cohesion above the hamlet level (Oliver 1955),
although the extent to which political organisation as seen post WWI was a colonial transforma-
tion of previous systems such as described by Thurnwald (1951) is debated (Spriggs 1997: 199).
The late prehistoric settlement pattern in the Shortlands, as observed by Irwin in 1970 (Irwin
1972, 1973), consisted of coastal villages opposite reef passages with some interior hamlets or
garden houses occupied for short periods of time. This pattern is consistent with that reported by
Guppy (1887) in the late 19th century. Using ceramics recovered from test excavations or surface
collections from about 40 sites, Irwin was able to create a stylistic seriation allowing him to place
the sites into a four stage chronological scheme. This sequence starts circa 1500-1000 B.P. and
includes six modern villages. His location and functional analysis suggests that the settlement
pattern has remained essentially unchanged, indicating no major shifts in the forces influencing
settlement. The data do however appear to reflect some change in Alu populations over time, with
a significant 19th century population decline. This is in keeping with historic records which report
the devastating impact of European disease (Guppy 1887: 176, Thurnwald 1912 [III]: 76).
Other than Irwin’s Shortlands sequence, the archaeological record of settlement pattern
change in the Northern Solomon Islands is still undeveloped. The archaeology in both the north
and in the Shortlands indicates a tendency towards patterns of long-term continuity in the distri-
bution and relative density of settlements. Ethnographic observations in the south suggest that the
major cultural factors that may have influenced change were those arising out of dynamic
56
processes in the interaction spheres. Using the historic Bougainville Straits as an example, this
may have included such events as slave raiding, headhunting and later, the impact of colonial
administrative developments and disease. Such dynamism is also well expressed in the traditions
of 19th century movement of Austronesian speaking peoples between Shortland (Alu), Mono
(Treasury Group) and southern Bougainville (Terrell and Irwin, 1972).
The greatest non-cultural influence on settlement patterns in the last 1000 years would
almost certainly have been the effects of volcanism. Bougainville is unique in the Solomon Islands
for the considerable number of large volcanoes which are found along the central spine of the
island. Volcanic activity over the last few thousand years has resulted in deep ash deposits and
pyroclastic flows which, as Spriggs has noted (1992, 1997: 171), must have strongly shaped the
cultural history of the island, creating at times large uninhabitable zones which would have repeat-
edly isolated the northern and southern regions. At the same time, volcanic ash would ultimately
enhance soil fertility and this explains the fertility and population densities of the Buin Plain.
The Central and Eastern Solomons
There is no ceramic record from the Central and Eastern Solomons, and the most common archae-
ological features of the region are monumental structures which include a range of shrines and
burial structures with the more elaborate forms occurring in the east. In Santa Ana and Star
Harbour (Makira) walled rectangular platforms contain multiple burials and in both places they
appear to be associated with settlement zones (Swadling 1976). Green (1976b) excavated one
enclosure at Star Harbour and reported a few grave goods and a range of artefacts located upon the
platform surface including shell arm ring fragments, chert flakes, pearl shell lures, pig bones and
teeth (Green 1976b: 146). Some or most of those are probably votive items. Dating such structures
is difficult but it is likely that they date mainly to the last 500 years of prehistory – as they do in
the Western Solomons (see below). Other ceremonial or religious structures on Santa Ana include
canoe houses, often with ancillary stone walling and platforms (Swadling 1976: 126). In northwest
Makira Miller recorded burial enclosures similar to those on Santa Ana and Star Harbour as well
as a range of other platforms and low mounds (Miller 1979: Fig. 5.4, 103). On Ulawa both small
shrines and U-shaped canoe houses were noted by Hendren (1976) and Ward (1976), and small
shrine features were recorded by Miller (1979: 109) in East Kwaio on Malaita.
On central and northwest Guadalcanal Roe (1993: 32-34) recorded shrines, low stone cairns,
deposits of shell valuables, stone kerbed and in-filled burial platforms containing human remains,
and deposits of exhumed bones. The latter were generally found in caves, or small stone-slab
roofed crypts covered with vegetation and soil, and were said to represent the remains of
prominent male leaders. Although the data is limited, it appears that the monumental features of
central and northwestern Guadalcanal are more similar to those of its trading partners in Malaita
and perhaps Ulawa than to those of the Eastern Solomons.
Early accounts and archaeological survey provide some insight into what late prehistoric
settlement patterns may have been like. The Spanish explorers reported dense settlement along the
Guadalcanal plains and extending into the foothills (Amherst and Thomson 1901), although by the
early 20th the density of occupation declined along the northwest Guadalcanal coast (Paravicini
1931: 60) as people fled the impact of headhunting raids from the west (Roe 1993: 23). Roe’s
surveys of the Poha and Vuru valley systems and the Themeda grasslands of the northwest cape
produced little in the way of village sites other than occasional terraces, house platforms, and chert
scatters on the ridge tops. These suggest the presence of small, dispersed hamlets.
57
Hendren (1976) and Ward (1976) report a similar pattern of coastal occupation from Ulawa,
where the archaeological evidence of settlements mainly consists of stone canoe house platforms
and small stone shrine enclosures. At the site of Haradeiwi (BS-UW-1) located on a raised inland
beach, Ward (1976: 165) excavated a hamlet-size occupation limited to a “long low mound” of
shell and chert debris and a small shrine enclosure. Dates are reported for the upper and lower
deposits suggesting occupation over some hundreds of years beginning in the 12th century A.D.
There is very little archaeological information on prehistoric settlement patterns on Malaita. A
lagoon-edge settlement pattern including the use of artificial islands (such as are now found in Langa
Langa and Lau Lagoons) was certainly part of the coastal picture. Contemporary inland settlement
patterns include small hilltop and ridge hamlets, and Miller (1979: 109) reports that in East Kwaio
there is some evidence of this settlement pattern in the late prehistoric archaeological record. But he
also notes that these types of settlements appear to have left few archaeological traces, except where
small stone shrines remain, or where stone walling was used to augment natural defences.
In the Eastern Solomons the picture is different and prehistoric coastal settlement areas are
often marked by the presence of sizeable mounds of midden debris and dense distributions of
stone features of a variety of forms. On Santa Ana Swadling (1976: Fig. 34) mapped a coastal
village at Mwaroqorafu consisting of numerous intersecting stone walls of probable horticultural
function, stone lined pathways, mounds, wells, burial enclosures, and at least one canoe house.
The site provided dates suggesting occupation between the 15th and early 18th centuries A.D.
(Swadling 1976: 128-29). At Star Harbour Green found a similar record, with numerous sites
marked by low mounds ranging from several metres to 20 m in diameter (see also Miller 1979:
90). The largest site (Na Mugha, BB-8-4) contained seven mounds and three burial enclosures
distributed over an area of about a hectare (Green 1976b: Fig. 38). This site produced a date
suggesting occupation as early as the early 17th century, but historic artefacts dating to the 18th
or early 19th century were also recovered (Green 1976b: 143). A similar settlement marked by
the presence of two large mounds was located on the small island of Uki off the northwest coast
of Makira. The site (Suena, BB-2-7) is situated in an area of shell midden scatter on a coral bench
behind the modern village of Su‘ena, and excavation revealed a total depth of deposit of nearly
two metres, representing an occupation of up to 500 years starting in the early 16th century.
The archaeological remains of horticultural practices in the Central and Eastern Solomons
pose something of a mystery. Mendana and his companions report large and intensive irrigated
gardening systems in the interior of Guadalcanal to the west of modern Honiara (Amherst and
Thomson 1901: 308-20). Yet at the time of Yen’s review of agricultural systems (1976, see also
this volume) there was no evidence of irrigation systems in the region, although he did report
enclosed field systems suggesting some degree of intensification. Roe (1993: 159) used aerial
photography to locate some pond field systems in the interior valleys of the northwest cape of
Guadalcanal and these have been tentatively dated to about 450 B.P.
On Ulawa Hendren (1976: 154) reported walled gardening systems comprising “…large
areas of adjoining stone-walled enclosures on the coastal flats of the island”, and on the higher
lands combinations of enclosures and stone faced terraces extended down valley slopes. On Santa
Ana more substantial enclosures with walls up to one metre high and varying in width from 40 to
150 cm served as walled yam gardens (Yen 1976: 64). Green’s survey at Star Harbour and on the
island of Uki did not locate any horticultural field evidence.
Historically Malaita, Ulawa, Makira, Santa Ana, Uki, and the Marua Sound area of south-
eastern Guadalcanal were bound together in an extensive trade network involving the movement
of chert from Malaita and Ulawa, metamorphic and igneous stone adzes from Marau Sound, and
58
shell beads from Makira (Green 1976c: 13). Other networks linked western Guadalcanal with the
Florida Group and Malaita in the exchange of chert, shell beads and various perishable goods
(Roe 1993: 29). The archaeology, while fragmentary, suggests that these exchange systems have
some antiquity. The key indicator of early trade in the region is the distribution of chert especial-
ly that from the extensive Ulawan and Malaitan source areas. It is actually difficult at present to
accurately distinguish cherts from the different source areas, but on hand-specimen criteria,
Davenport (1972: 171, 183) attributed the chert flakes recovered in all of his Santa Ana excava-
tions to sources on Ulawa or Makira. Furthermore, the presence of chert throughout the deposits
suggests that cherts were being moved into Santa Ana from at least the early first millennium B.C.
(Black and Green 1975: 30). At the Suena site chert flakes and adzes were ubiquitous throughout
the excavated mound indicating a time depth for chert transport of at least 500 years. From the
Star Harbour site of Na Mugha Green recovered several thousand chert flakes, adzes and drill
points, as well as a nodule of raw material which he attributed to Ulawan or Malaitan sources,
and polished adzes from surface collections of a gray-green meta-basalt he believed originated in
Marau Sound on Guadalcanal (Green 1976b: 140, 144). On Guadalcanal Roe noted a change
through time in the type of chert found in the Poha Valley sites. In the earliest levels of Poha Cave
the chert is mainly a coarse variety attributed to the Mbirao volcanics of eastern Guadalcanal.
This changes perhaps from about 2300 B.P., to the brown vitreous material characteristic of
Malaita (Roe 1993: 178). Taken together the data from the Central and Eastern Solomons
indicates considerable antiquity to the historic patterns of interaction between major islands, and
based on the Guadalcanal data it seems probable that exchange systems commenced or intensified
at about the time pollen cores on Guadalcanal indicate the start of intensive gardening during the
Hamosa Phase (from circa 2200 B.P.) (Roe 1993: 183).
The Reef/Santa Cruz Islands
The Reef/Santa Cruz Islands contain a dense record of Lapita settlement and this has produced a
rich account of early Lapita occupation. The post-Lapita record is less well represented in archaeo-
logical accounts, and far less clearly understood. A major recent contribution to the archaeology of
the region is contained in Doherty (2007) (see also this volume). This Ph.D. thesis (University of
Auckland) explores in detail the material culture, settlement pattern and linguistic evidence of post-
Lapita culture change. Doherty concludes that some core cultural schema were introduced into the
region by the Lapita settlers and persisted right through the sequence. But around these threads of
continuity Reef/Santa Cruz prehistory has been influenced by repeated cultural and linguistic
inputs, locally driven change and a constant re-sorting of inherited traditions. Doherty rejects the
idea that post-Lapita culture change can be attributed to successions of separate traditions.
Historic settlement patterns in the region include both coastal and inland occupation
(Figure 3.7). Stone structures are quite common, but the Reef/Santa Cruz Islands appear, based
on the available archaeological record, not to have shared the tradition of ancestral skull shrines
found elsewhere in the Solomons, and most especially in the Western Solomons. Archaeologically,
human remains have been found in association with houses, which is in keeping with the custom
of retaining the bones of ancestors in houses. It is not clear how far back this practice extends, but
two human burials were found in the Növlaö Rockshelter dating just earlier than 2000 B.P., and
seemingly associated with ovens and house foundations.
Settlement sites include a number of small permanent villages or hamlets and are marked by
the presence of various stone structures. Like Suena on Uki Island, these sites may have been
relatively stable, supporting at least intermittent occupation over two or three centuries. The
coastal and inland settlement sites show a great deal of similarity.
59
Figure 3.7 Map showing location of sites in the Reef/Santa Cruz region referred to in the text.
On the western peninsula of Santa Cruz Yen mapped and carried out small scale excavations
at the inland site of Naiavila (SZ-12), a historic village located close to Graciosa Bay (Yen 1976:
204). The site was about 0.6 hectares in size and contained two clusters of structures outlined by
stone foundations and joined by a stone wall described as a “pig fence” (Yen 1976: Fig. 59). The
majority of features appear to have been rectangular house foundations including large structures
identified by a former inhabitant as men’s houses. A large circular kerbed structure, ten metres in
diameter was identified as a dance circle. Structures described as god houses are reported within
villages, and at Naiavila this site type was represented by a large rectilinear foundation associated
with a historic religious cult. A pile of human skulls and bottle fragments were found in a corner
of the house identified as belonging to the chief (Yen 1976: 210). The site was abandoned in the
1920s, but radiocarbon dates suggest it might have been established at or before the time of
Spanish discovery in the 16th century A.D. A very similar coastal site was recorded by McCoy
and Cleghorn (1988) at Dai Village (SZ-11) located on the tip of the small island Tömotu Neo
which lies at the entrance to Graciosa Bay.
At Dai Village a total of 35 stone structures, including a centrally located dance circle, were
recorded in a tightly clustered area of around 0.4 hectares. Two adjacent rectilinear structures
somewhat isolated at the southwestern corner of the site were identified as god houses. Human
60
burials were present in three of the four excavation areas but absent from the dance circle
excavation. Some were articulated burials, while others were jumbles of bones from different
individuals buried in shallow pits (McCoy and Cleghorn 1988: 111). This site was abandoned
after the arrival of the mission, but a suite of radiocarbon dates provide a settlement age range
similar to that from Naiavila, with one feature dating back into the 15th century A.D. (McCoy
and Cleghorn 1988: 109).
To date there is very little archaeological evidence for horticultural practices in the Reef/Santa
Cruz Islands, although forest-swidden systems using taro and yam were likely to have always been
important and these leave little direct trace (Yen 1973: 43, 1976: 62). The other important element
of the horticultural system was an elaborate arboriculture which included breadfruit fermentation
(exceptional in this “Melanesian” context) and drying (Yen 1976). Clearly, however, this swidden-
arboricultural complex could support medium to large permanent settlements.
The Reef/Santa Cruz Islands are home to the famous red feather money exchange system
(Davenport 1962). The web of inter-island exchange that sustained the red feather system
included the Duff Islands (Taumako), the Reef Islands, and Santa Cruz, as well as Utupua and
Vanikoro to the south. But the core of the historic system involved frequent interaction within the
Reefs and Santa Cruz, where the red feather money circulated in direct support of the subsistence
and prestige economy. There are no apparent historical data suggesting regular ties to the Eastern
Solomons (Davenport 1964). This represents change from the Lapita Period when the interaction
spheres included the southern end of the Eastern Solomons and extended south to the Banks
Islands. During the Lapita Period cherts were brought in from Malaitan and Ulawan sources as
well as a small quantity from the Duff Islands. These sources do not appear to be represented in
the late period sites excavated by Yen, and McCoy and Cleghorn (1988) do not report any from
the Dai Village site. Some volcanic glass from the later period sites is provisionally sourced to the
Banks Islands, indicating interaction to the south, possibly through Vanikoro, although the people
of Vanikoro claim to have had no knowledge of the Banks Islands until the mission period
(Davenport 1964: 135).
THE WESTERN SOLOMONS – ROVIANA

In this final section we describe the post-Lapita archaeology of the Western Province from the
perspective of our project (The New Georgia Archaeological Survey) in Roviana Lagoon, New
Georgia. This synthesis includes the first archaeological based model for the emergence of a late
period Solomon Island society.
The landscape of Roviana Lagoon on the south coast of New Georgia contains an archaeo-
logical record commencing around 2600 B.P., although vegetation changes up to a millennium
earlier might signal earlier human intrusion into the area. We divide the archaeological record into
three periods: the Early Ceramic Period, the Bao Period and the Roviana Period (Figure 3.5). The
Early Ceramic Period is the late and immediately post-Lapita occupation represented by an inter-
tidal ceramic tradition (see above). The data from Roviana Lagoon indicate the presence of small
inter-tidal hamlets or villages along the lagoon margins from at least the mid-first millennium
B.C. As is the case today, these sites were mainly located near passages between barrier islands
and near fresh water sources. Our knowledge of this period is sparse and there are many gaps, but
it is likely to contain a great deal more variation than is currently apparent. The second two
periods fall within what we term the Munda Tradition, which spans the last 700 years of prehis-
tory and documents the development of the political, economic, religious and ceremonial complex
known historically as the Roviana Chiefdom.
61
Munda Tradition (700 – 100 B.P.)

The archaeological landscape of the Munda Tradition is dominated by sites and material culture
which are recognised by the people of Roviana as relating directly to the histories, genealogies
and traditions of their own ancestors. Ceramic use continued for most of the Munda Tradition and
is represented by a thin undecorated ware found scattered through the gardening soils and around
the habitation zones. But ceramic production centers were very few and, in Roviana at least,
pottery was probably all imported. The most distinctive site types of this archaeological landscape
are agricultural field systems, stone shrines and related architectural features. Agricultural sites
comprise stone faced terraces and channels forming pond field systems, and are reported from
throughout the Western Province including Kolombangara, Rendova (Yen 1976: 69), Roviana,
Kusaghe (North New Georgia) (Chikamori 1967, Reeve 1986, Tedder and Barrus 1976), Viru
Harbour (South New Georgia) (Miller 1979: 133), Marovo, Vangunu and Gatokae (Bayliss-
Smith, Hviding and Whitmore 2003, Hviding and Bayliss-Smith 2000. These wet field systems
range in size from very small single family sized plots, to large terraced areas of up to 100
hectares (Tedder and Barrus 1976). Hviding and Bayliss-Smith (2000: 131-37) suggest that the
larger of these systems could have supported very high population densities in the 19th century
before they were abandoned in the early 20th century.
Shrines and other architectural features made of igneous stone and coral are found predom-
inantly along the coast, but are also found in scattered inland locations. The shrines consist of low
platforms of coral or basaltic stone which often contain specialised sub-components such as small
stone skull houses, uprights and various types of stone walling. In addition, skulls (of ancestors),
animal remains and caches of shell valuables or other artefacts are often found on or adjacent to
the shrines. In the historic period, shrines were associated with ancestor cults and with the
exchange systems that networked the Western Solomons (Sheppard, Walter and Nagaoka 2000,
Walter and Sheppard 2001). In addition to the shrines, other common architectural features
include house platforms, walls, wharves, and canoe house foundations. In places, aggregates of
these sites delineate late prehistoric and early historic coastal settlements.
The cultural complex to which the Munda Tradition sites relate is known as the Roviana
Chiefdom, which was described by the earliest visitors to New Georgia in the 19th century. The
Roviana Chiefdom was a powerful polity whose leaders controlled an exchange and headhunting
network that spanned the Western Province, and which maintained links with other networks in
the Guadalcanal region to the southeast, and with Choiseul and Bougainville to the northwest
(Dureau 1994, Hocart n.d., Jackson 1978, McKinnon 1975, Miller 1978, Walter and Sheppard
2001, White 1979, Zelenietz 1979). Within this network shell valuables circulated and were used
in a variety of social, ritual and economic transactions (Aswani and Sheppard 2003). All power
in the Roviana Chiefdom was bestowed by the ancestors whose skulls were deposited on the
ancestral skull shrines (Figures 3.8 and 3.9). On these shrines ritual acts, often involving offerings
of shell valuables, invoked support of the ancestors which was expressed as mana (efficacy)
(Dureau 2000), a spiritual manifestation of power and the right to wield it in secular affairs
(Codrington 1891: 118, Walter and Sheppard 2001). At the time of European contact one of the
main ritual and political centres for the chiefdom was located on the barrier island of Nusa
Roviana, close to the settlement of Munda (Figure 3.5).
We divide the Munda Tradition into two periods based on a seriation of architectural features
and radiocarbon dating of selected shrines and other site classes (Sheppard and Walter 2006a,
Sheppard et al. 2000, Walter and Sheppard 2001).
62
Figure 3.8 Sipikelo shrine on Vella Lavella contains more than 22 skulls and many shell valuables.
Figure 3.9 The cycle linking ritual, power and political efficacy in the Roviana Chiefdom.
63
Bao Period (700 to 400 B.P.)

The Bao Period (700 to 400 B.P.) represents the earliest archaeological evidence for the local
development of the Roviana Chiefdom. The sites consist mainly of platforms interpreted as
shrines in which basalt is used to create uprights, table stones, cobble flooring and wall compo-
nents. Bao Period sites do not contain any human bone, shell valuables or ovens. These sites are
found in three zones. The earliest sites, dating to the early 13th century A.D., are located at Bao,
high in the rainforests on mainland New Georgia and consist of large basalt platforms spread
along 500 m of ridgeline. According to tradition the earliest ancestors of the Roviana people lived
at Bao, and that was the origin place for the sacred power, ritual and symbolism of Roviana
religion and politics. Elsewhere on the mainland, Bao Period sites are found as single features in
isolated positions on ridge tops up to several kilometres from the lagoon shores. With the
exception of one ceramic and midden deposit dated to the 15th century A.D. (Site 25) (Sheppard
and Walter 2006), no habitation complexes of the Bao Period have been identified, but the settle-
ment pattern inferred by the mainland site distribution is of small inland settlements, based on the
production of irrigated taro and the harvesting of nut trees located near the headwaters of larger
streams. The shrines were always located at some distance, often on the first coastal ridge, and
this implies a deliberate separation of sacred and secular space. A similar pattern of segregation
is evidenced by the Bao Period sites on the barrier islands, none of which contain any evidence
of habitation, cooking or the accumulation of midden waste. On the barrier islands, symbolic links
were maintained with the ancestral homeland at Bao by the importation of basalt from the
mainland rivers for use in shrine construction.
None of the Bao sites contain stone ovens, which in the Roviana Period were used by the
priests to prepare sacrificial offerings on the shrine. Nor do they contain skull houses or votive
offerings, both of which are abundantly represented in the later Roviana Period sites.
Roviana Period (400 – 100 B.P.)
This period is represented by a much wider range of site types and by a rich material culture.
Shrine construction continued, but the use of basalt fell off in favour of coral rubble construction.
The shrines contain human bone, shell valuables and, in the historic period, ceramics, trade axes
and muskets. Many of the shrines have skull houses of sheet coral containing skulls and votive
offerings. In addition to the shrines, other architectural features include wharves, canoe house
foundations, house platforms and walls. These site types occur in dense aggregates, and in some
areas the space between these structures is a carpet of portable artefacts, midden shell and low
stone constructions. The archaeological landscape demonstrates that by the Roviana Period major
changes had taken place in settlement pattern and the symbolic organisation of space. Within the
habitation zones shrines and house platforms are located side by side, often sharing one or more
walls. The distinction between sacred and secular space of the Bao Period had disappeared in
favour of a social landscape that appears to be entirely ritually charged (Sheppard et al. 2000).
Sites are found all along the lagoon coasts of both the mainland and the barrier islands with
several areas of particularly dense concentration. One of these places is the barrier island of Nusa
Roviana, where there is a dense cluster of shrines, wharves and stone house platforms around two
coasts surrounding a fortification constructed along the central ridge of the island. The Nusa
Roviana hill fort contains some of the most important shrines of the Roviana Chiefdom. The fort
consists of a series of transverse ditches and banks which divide the ridge into sections, each
named and associated with particular ancestors, spiritual beings and important events in the
history of the Roviana people (Thomas, Sheppard and Walter 2001). At the highest and most
64
isolated part of the hill fort complex is a small paved and terraced area supporting a small shrine
containing a statue of a dog’s head and a pile of shell valuables. The dog represents Tiola, the
most important figure in Roviana mythology. It was he who, in some versions of the tradition,
taught the people of Roviana how to make war canoes and carry out headhunting.
Elsewhere in the Western Province recent work has continued the New Georgia Archaeological
Survey focus on the archaeology of the recent past. In Vella Lavella the authors have directed a
survey and excavation programme that compares the development of exchange, ritual systems, and
socio-political changes in these non-Austronesian speaking communities with that of the Roviana
record (Sheppard and Walter 2004, 2005, 2006b). In Rendova Tim Thomas is using a similar field
survey, oral history based approach to investigate the emergence of communities of practice;
networks of shared symbols, ideologies and cultural practices (see Thomas, this volume).
CONCLUSIONS
We began this review by acknowledging the difficulty of synthesising the Solomon Island archae-
ological record given the vast areas involved, the long time depth of occupation, and the sheer
diversity of cultural and linguistic history. But we also pointed out that contrary to some views
there is sufficient archaeological data already available to begin to piece together a coherent
outline of the main threads of Solomon Island prehistory. We also argued that the Lapita threads
are mostly intact.
Archaeological work in the Solomon Islands has gone through a period of rapid growth
since the mid 1990s resulting in at least 12 Ph.D. and M.A. theses since 1998, and many publica-
tions and other outputs. There are ongoing programmes of field archaeology in the Western
Province, and Green (this volume) lists the work still in progress relating in some way to the
SSICHP he started with Yen in the early 1970s. We are also aware of new projects being planned
for the Central and Eastern Solomons. It is likely that the four main themes in Solomon Island
archaeology which we follow in this review will continue to guide new research, and of these, the
search for more Lapita sites is likely to figure most prominently. But it is our view that the archae-
ology of the recent (post-Lapita) past offers the most exciting and anthropologically interesting
potential. It is in this period when the diversity of language and cultural expression making up
traditional Solomon Island society, and which attracted early and sustained anthropological
attention, first developed.
Archaeology is the only discipline capable of creating a deeper history to support and
contextualise the rich ethnographic and historical records, and of tracing the long-term develop-
ment of the complex interaction spheres, exchange systems, ancestor cults, religious practices,
art, and material culture that structure traditional Solomon Island life. A final area of research
waiting to be opened is in the archaeology of the Pacific War. There is currently a rich landscape
and material culture record of the Solomon Island war yet to be explored, offering enormous
potential for future archaeological research.
ACKNOWLEDGMENTS
The authors would like to thank the many people of the Western Solomons who have helped us
over the years, especially our colleague Kenneth Roga. In addition we thank Lawrence Foana‘ota
and the staff of the National Museum, Solomon Islands as well as the many graduate students
from the University of Auckland and Otago who have assisted us and provided us with many fine
theses. This research was funded by major grants from the Marsden Fund of the Royal Society of
New Zealand, the National Geographic Society and the Universities of Auckland and Otago.
65
REFERENCES
Allen, J., 1993. Notions of the Pleistocene in Greater Australia. In M. Spriggs, D. Yen, W. Ambrose, R.
Jones, A. Thorne and A. Andrews (eds), A Community of Culture: The People and Prehistory of the
Pacific. Occasional Papers in Prehistory 21. Canberra: Department of Prehistory, Research School
of Pacific Studies, Australian National University, pp. 139-51.
Allen, J. and C. Gosden (eds), 1991. Report of the Lapita Homeland Project. Occasional Papers in
Prehistory 20. Canberra: Department of Prehistory, Research School of Pacific Studies, Australian
Allen, J., C. Gosden and J.P. White, 1989. Human pleistocene adaptions in the tropical island Pacific.
Antiquity, 63: 548-61.
Amherst, Lord W. and B. Thomson, 1901. The Discovery of the Solomon Islands by Alvaro de Mendana in
1568. London: Hakluyt Society.
Aswani, S. and P.J. Sheppard, 2003. The archaeology and ethnohistory of exchange in precolonial and
colonial Roviana: Gifts, commodities and inalienable possessions. Current Anthropology, 44
[supplement]: S51-S78.
Bathgate, M.A., 1985. Movement processes from precontact to contemporary times: The Ndi-Nggai, West
Guadalcanal, Solomon Islands. In M. Chapman and R. Prothero (eds), Circulation in Population
Movement: Substance and Concepts from the Melanesia Case. London: Routledge and Kegan Paul,
pp. 83-118.
Bayliss-Smith, T., E. Hviding and T.C. Whitmore, 2003. Rainforest composition and histories of human
disturbance in the Solomon Islands. Ambio, 32: 346–52.
interaction, and the emergence of the “ethnographic present”. Asian Perspectives, 47: 95-120.
Bedford, S., M. Spriggs and R. Regenvanu, 2006. The Teouma Lapita site and the early human settlement
of the Pacific Islands. Antiquity, 80: 812-28.
Best, S., 2002. Lapita: A View from the East. New Zealand Archaeological Association Monograph 24.
Auckland.
Black, S. 1977. The Excavations at Teobebe. Chicago: Field Museum of Natural History.
Black, S. and R.C. Green, 1975. Radiocarbon Dates from the British Solomon Islands to December 1973.
Working Papers in Anthropology, Archaeology, Linguistics and Maori Studies 39. Auckland:
Department of Anthropology, University of Auckland.
Blackwood, B., 1931. Report on field work in Buka and Bougainville. Oceania, II: 199-219.
Blackwood, B., 1935. Both Sides of Buka Passage: An Ethnographical Study of Social, Sexual and
Economic Questions in the North-Western Solomon Islands. Oxford: Clarendon Press.
Chikamori, M., 1967. Preliminary report on archaeological research in the Western Solomons. In S. Itoh and
M. Chikamori (eds), A Brief Report on Research into the Prehistory and Anthropology of Various
Islands in the British Solomon Islands Protectorate. Hamamatsu: Kaimeido, pp. 8-42. [in translation].
Chikamori, M. (ed.), 1975. The Early Polynesian Settlement on Rennell Island British Solomon Islands
Protectorate: Preliminary Report 1975. Volume 1. Occasional Papers of the Department of
Archaeology and Ethnology, Keio University. Tokyo: Keio University.
Chikamori, M. and H. Takasugi, 1985. Archaeology on Rennell Island. Tokyo: Department of Archaeology
and Ethnology, Keio University.
Codrington, R.H., 1891. The Melanesians: Studies in Their Culture and Anthropology. Oxford: Clarendon
Press.
66
Davenport, W., 1962. Red feather money. Scientific American, 206: 94-103.
Davenport, W., 1964. Notes on Santa Cruz voyaging. Journal of the Polynesian Society, 73: 134-42.
Davenport, W., 1968. Anthropology in the British Solomon Islands. Expedition, 11: 31-34.
Davenport, W., 1972. Preliminary excavations on Santa Ana Island, Eastern Solomon Islands. Archaeology
and Physical Anthropology in Oceania, 7: 165-83.
Davenport, W., T. Russell and J.L.O. Tedder, 1968. Excavations at Fotoruma, Poha River, Guadalcanal,
British Solomon Islands Protectorate. Manuscript, Solomon Islands National Museum, Honiara.
Dunn, M., A. Terrill, G. Reesink, R. Foley and S. Levinson, 2005. Structural phylogenetics and the
reconstruction of ancient language history. Science, 309: 2072-75.
Dureau, C., 1994. Mixed Blessings: Christianity and History in Women’s Lives on Simbo, Western
Solomon Islands. Unpublished Ph.D., Macquarie University, Sydney.
Dureau, C., 2000. Skulls, mana and causality. Journal of the Polynesian Society, 109: 71-97.
Felgate, M.W., 2001. A Roviana ceramic sequence and the prehistory of Near Oceania: Work in progress. In
G.R. Clark, A.J. Anderson and T. Vunidilo (eds), The Archaeology of Lapita Dispersal in Oceania:
Papers from the Fourth Lapita Conference, June 2000, Canberra, Australia. Terra Australis 17.
Canberra: Pandanus Books, pp. 39-60.
Lagoon, New Georgia, Solomon Islands. Unpublished Ph.D. thesis, University of Auckland,
Auckland.
Felgate, M. W. 2007. Leap-frogging or limping? Recent evidence from the Lapita littoral fringe,
Explorations: Lapita and Western Pacific Settlement. Canberra: ANU E Press. pp. 123-140.
Flannery, T., 1995. Mammals of the South-West Pacific and Moluccan Islands. Chatswood: Reed Books.
Foana‘ota, L.A., 1979. The Solomon Islands national site recording system. In J.R. McKinley and K.L.
Jones (eds), Archaeological Research Management in Australia, and Oceania. Wellington: New
Zealand Historic Places Trust, pp. 29-33.
Foana‘ota, L.A., 1996. The development of archaeological work in the Solomon Islands. In Janet
Davidson, Geoffrey Irwin, Foss Leach, Andrew Pawley and Dorothy Brown (eds), Oceanic Culture
History: Essays in Honour of Roger Green. Dunedin: New Zealand Journal of Archaeology Special
Publication, pp. 241-43.
Galipaud, J-C. and M. Swete Kelly, 2007. Makué (Aore Island, Santo, Vanuatu): A new Lapita site in the
ambit of New Britain obsidian distribution. In S. Bedford, C. Sand and S.P. Connaughton (eds),
Oceanic Explorations: Lapita and Western Pacific Settlement. Canberra: ANU E Press, pp. 151-162.
Gosden, C., 1995. Arboriculture and agriculture in coastal Papua New Guinea. Antiquity, 69 [265]: 807-17.
Green, R.C., 1976a. A late prehistoric sequence from Su‘ena Village, Uki. In R.C. Green and M.M.
Cresswell (eds), Southeast Solomon Islands Culture History: A Preliminary Survey. Bulletin 11.
Wellington: Royal Society of New Zealand, pp. 181-91.
Green, R.C., 1976b. A late prehistoric settlement in Star Harbour. In R.C. Green and M.M. Cresswell (eds),
Southeast Solomon Islands Culture History: A Preliminary Survey. Bulletin 11. Wellington: Royal
Society of New Zealand, pp. 133-47.
67
Green, R.C., 1976c. An introduction to the Southeast Solomons Culture History Programme. In R.C. Green
and M.M. Cresswell (eds), Southeast Solomon Islands Culture History: A Preliminary Survey.
Bulletin 11. Wellington: Royal Society of New Zealand, pp. 9-17.
Green, R.C., 1977. A First Culture History of the Solomon Islands. Auckland: University of Auckland Bindery.
Green, R.C., 1979. Lapita. In J.D. Jennings (ed.), The Prehistory of Polynesia. Cambridge, Mass.: Harvard
University Press, pp. 27-60.
Green, R.C., 1987. Obsidian results from the Lapita sites of the Reef/Santa Cruz Islands. In W.R. Ambrose
and J.M.J. Mummery (eds), Archaeometry: Further Studies in Australasia. Occasional Papers in
Prehistory 14. Canberra: Department of Prehistory, Research School of Pacific Studies, Australian
Green, R.C. and J.R. Bird, 1989. Fergusson Island obsidian from the D’Entrecasteaux Group in a Lapita
site of the Reef Santa Cruz Group. New Zealand Journal of Archaeology, 11: 87-99.
Green, R.C. and M.M. Cresswell (eds), 1976. Southeast Solomon Islands Cultural History: A Preliminary
Report. Bulletin 11. Wellington: Royal Society of New Zealand.
Green, R.C., M.D. Jones and P.J. Sheppard, 2008. The reconstructed environment and absolute dating of
SE-SZ-8 Lapita site on Nendö, Santa Cruz,, Solomon Islands. Archaeology in Oceania, 43: 49-61.
Grimes, S., 2003. A History of Environmental Change and Human Impact Since 3750 B.P. in the New
Georgia Group-Western Solomon Islands. Unpublished Ph.D. thesis, University of Western
Australia, Perth.
Guppy, H.B., 1887. The Solomon Islands and their Natives. London: Swan Sonnenschien, Lowrey and Co.
Haberle, S., 1993. Environmental History of the Northern Plains of Guadalcanal, Solomon Islands. In D. Roe,
Prehistory Without Pots: Prehistoric Settlement and Economy of North-west Guadalcanal, Solomon
Islands [Appendix 1, pp.1-25]. Unpublished Ph.D. thesis, Australian National University, Canberra.
Haberle, S., 1996. Explanations for paleoecological changes on the northern plains of Guadalcanal,
Solomon Islands: The last 3200 years. Holocene, 6: 333-38.
Hendren, G.H., 1976. Recent settlement pattern changes on Ulawa, Southeastern Solomon Islands. In R.C.
Green and M.M. Cresswell (eds), Southeast Solomon Islands Culture History: A Preliminary Survey.
Bulletin 11. Wellington: Royal Society of New Zealand, pp. 149-59.
Hocart, A.M., n.d. Manuscripts and Field Notes on Eddystone and New Georgia. MS-Papers-0060.
Wellington: Alexander Turnbull Library.
Hviding, E. and T. Bayliss-Smith, 2000. Islands of Rainforest: Agroforestry, Logging and Eco-Tourism in
the Solomon Islands. Aldershot: Ashgate.
Irwin, G.J., 1972. An Archaeological Survey in the Shortland Islands, B.S.I.P. Unpublished M.A. thesis,
Irwin, G.J., 1973. Man-land relationships in Melanesia. Archaeology and Physical Anthropology in
Oceania, 8: 226-52.
Jackson, K.B., 1978. Tie Hokara, Ties Vaka: Black Man White Man, A Study of the New Georgia Group to
1925. Unpublished Ph.D. thesis, Australian National University, Canberra.
Jones, M.D., F.J. Petchey, R.C. Green, P.J. Sheppard and M. Phelan, 2007. The marine ΔR for Nenumbo
(Solomon Islands): A case study in calculating reservoir offsets from paired sample data.
Radiocarbon, 49 (1): 95-102.
Keopo, J. and Y. Kawamura, 1999. Report on National Site Survey: Rereona, Kia District, Isabel Province.
Honiara: Solomon Islands National Museum.
Kirch, P.V., 1987. Lapita and Oceanic cultural origins: Excavations in the Massau Islands, Bismarck
Archipelago, 1985. Journal of Field Archaeology, 14 (2): 163-80.
Kirch, P.V., 1997. The Lapita Peoples: Ancestors of the Oceanic World. Cambridge: Blackwell.
68
Kirch, P.V., 2000. On the Road of the Winds: An Archaeological History of the Pacific Islands Before
European Contact. Berkeley and Los Angeles: University of California Press.
Lampert, R., 1966. Archaeological reconnaissance in Papua and New Guinea: 1966. Canberra, Australian
National University (mimeo).
Leavesley, M. and J. Allen, 1998. Dates, disturbance and artefact distributions: Another analysis of Buang
Merabak, a Pleistocene site on New Ireland, Papua New Guinea. Archaeology in Oceania, 33: 68–82.
Leavesley, M., M.I. Bird, L.K. Fifield, P.A. Hausladen, G.M. Santos and M.L. di Tada, 2002. Buang
Merabak: Early evidence for human occupation in the Bismarck Archipelago, Papua New Guinea.
Australian Archaeology, 54: 55–57.
Loy, T., M. Spriggs and S. Wickler, 1992. Direct evidence for human use of plants 28,000 years ago: Starch
residues on stone artefacts from the northern Solomon Islands. Antiquity, 66: 898-912.
Mann, P., F.W. Taylor, M.B. Lagoe, A. Quarles and G. Burr, 1998. Accelerating late Quaternary uplift of the
New Georgia Island Group (Solomon Island Arc) in response to subduction of the recently active
Woodlark spreading center and Coleman Seamount. Tectonophysics, 295 (3-4): 259-306.
McCoy, P.C. and P.C. Cleghorn, 1988. Archaeological excavations on Santa Cruz (Nendö), Southeastern
Solomon Islands: Summary Report. Archaeology in Oceania, 23: 104-15.
McKenzie, A., 2007. Ancestral Skull Shrines: Material Dialogues of Social Interaction in the Western
Solomon Islands. Unpublished M.A. thesis, Department of Anthropology, University of Auckland.
McKinnon, J., 1975. Tomahawks, turtles and traders: A reconstruction in the circular causation of warfare
in the New Georgia Group. Oceania, 45: 290-307.
Miller, D., 1978. An organizational approach to exchange media: An example from the Western Solomons.
Mankind, 11 (3): 288-95.
Miller, D., 1979. National Sites Survey Summary Report. Honiara: Solomon Islands National Museum.
Miller, D., 1980. Settlement and diversity in the Solomon Islands. Man, 15: 451-66.
Miller, D. and D. Roe. 1982. The Solomon Islands National Sites Survey: The first phase. IPPA Bulletin, 3:
47-51.
Nagaoka, T., 1999. Hope Pukerane: A Study of Religious Sites in Roviana, New Georgia, Solomon Islands.
Unpublished M.A. thesis, Department of Anthropology, University of Auckland.
Ogan, E., 1970. Nasioi pottery-making. Journal of the Polynesian Society, 79: 86-90.
Oliver, D.L., 1949. The Peabody Museum Expedition to Bougainville, Solomon Islands, 1938-39. Studies
in the Anthropology of Bougainville, Solomon Islands. Papers of the Peabody Museum of American
Archaeology and Ethnology, Harvard University 29 (1). Cambridge, Mass.
Oliver, D.L., 1955. A Solomon Island Society. Cambridge: Harvard University Press.
Paravicini, E., 1931. Reisen in den Britischen Salomonen. Frauenfeld und Leipzig: Verlag Huber and Co.
Pawley, A., 2007. Was There Early Lapita Settlement of the Solomon Islands? Bringing Linguistic
Evidence to an Archaeological Debate. Paper presented at 7th International Conference on Oceanic
Linguistics. Noumea, New Caledonia.
Poulsen, J., 1972. Outer archaeology: Bellona. A preliminary report on field work and radiocarbon dates.
Part 1. Archaeology. Archaeology and Physical Anthropology in Oceania, 7: 184-205.
Reeve, R., 1986. Kusage: An Archaeological Study. Gizo: Ministry of Culture Western Province, Solomon
Islands.
Reeve, R., 1989. Recent work on the prehistory of the Western Solomons, Melanesia Bulletin of the
Indo-Pacific Prehistory Association, 9: 46-67.
69
Roe, D., 2000. Maritime, coastal and inland societies in Island Melanesia: The bush-saltwater divide in
Solomon Islands and Vanuatu. In S. O‘Connor and P. Veth (eds), East of Wallace‘s Line: Studies of
Past and Present Maritime Cultures of the Indo-Pacific. Rotterdam: A.A. Balkema, pp. 197-222.
[Published as Modern Quaternary Research in SE Asia, 16].
Ross, H.M., 1970. Stone adzes from Malaita, Solomon Islands: An ethnographic contribution to Melanesian
archaeology. Journal of the Polynesian Society, 79: 411-20.
Ross, M., 1989. Early Oceanic linguistic prehistory: A reassessment. Journal of Pacific History, 24: 135-49.
Ross, M., 2001. Is there an East Papuan phylum? Evidence from pronouns. In A. Pawley, M. Ross and D.
Tryon (eds), The Boy from Bundaberg: Studies in Melanesian Linguistics in Honour of Tom Dutton.
Canberra: Pacific Linguistics, Research School of Pacific and Asian Studies, Australian National
Linguistics, 2: 456-98.
Russell, T., 2003. I Have the Honour To Be. Spennymoor: The Memoir Club.
Sand, C., 1996. Recent developments in the study of New Caledonia’s prehistory. Archaeology in Oceania,
31: 45-71.
Scott, R.M., J.R. Heyligers, J.R. McAlpine, J.C. Saunders and J.G. Speight (eds), 1967. Lands of
Bougainville and Buka Islands, Territory of Papua and New Guinea. Canberra: Commonwealth
Scientific and Industrial Research Organisation.
Sheppard, P.J., 1993. Lapita lithics: Trade/exchange and technology. A view from the Reefs/Santa Cruz.
Sheppard, P.J., M. Felgate, K. Roga, J. Keopo and R. Walter, 1999. A ceramic sequence from Roviana
Lagoon (New Georgia, Solomon Island). In J-C. Galipaud and I. Lilley (eds), The Pacific from
5000-2000 BP: Colonisation and Transformations. Paris: Editions de IRD, pp. 313-22.
Sheppard, P.J. and R.C. Green, 1991. Spatial analysis of the Nenumbo (SE-RF-2) Lapita site, Solomon
Islands. Archaeology in Oceania, 26: 89-101.
Sheppard, P.J. and R.C. Green, 2007. Sample size and the Reef/Santa Cruz Lapita sequence. In S. Bedford,
C. Sand and S.P. Connaughton (eds), Oceanic Explorations: Lapita and Western Pacific Settlement.
Canberra: ANU E Press. pp.
Sheppard, P.J. and R. Walter, 2004. Bilua Bifoa: Vella Lavella Archaeological Survey Year I. Annual Report
2005. Report to the Ministry of Education, Solomon Islands.
Sheppard, P.J. and R. Walter, 2005. Bilua Bifoa: Vella Lavella Archaeological Survey Year II. Annual
Report 2005. Report to the Ministry of Education, Solomon Islands.
Sheppard, P.J. and R. Walter, 2006a. A revised model of Solomon Islands culture history. Journal of the
Sheppard, P.J. and R. Walter, 2006b. Bilua Bifoa: Vella Lavella Archaeological Survey Year III. Annual
Report 2005. Report to the Ministry of Education, Solomon Islands.
Sheppard, P.J., R. Walter and T. Nagaoka, 2000. The archaeology of head-hunting in Roviana Lagoon, New
Georgia, Solomon Islands. Journal of the Polynesian Society. 109: 9-37.
Sheppard, P.J., R. Walter and K. Roga, in press. Friends, relatives, and enemies: The archaeology and
history of interaction among Austronesian and NAN speakers in the Western Solomons. In
J. Bowden and N. Himmelmann (eds), Fetschrift for Andrew Pawley on NAN and Austronesian
Languages. Canberra: Pacific Linguistics, Research School of Pacific and Asian Studies, Australian
Shutler, R. Jr and M. Shutler, 1964. Potsherds from Bougainville Island. Asian Perspectives, 8: 181-83.
Specht, J., 1969. Prehistoric and Modern Pottery Industries of Buka Island, T.P.N.G. Unpublished Ph.D.
70
Specht, J., 1972. The pottery industry of Buka Island, T.P.N.G. Archaeology and Physical Anthropology in
Oceania, 7: 125-44.
Spriggs, M., 1992. Archaeological and linguistic prehistory in the North Solomons. In T. Dutton, M. Ross
and D. Tryon (eds), The Language Game: Papers in Memory of Donald C. Laycock. Canberra: Pacific
Linguistics, Research School of Pacific and Asian Studies, Australian National University, pp. 417-26.
Spriggs, M., 1993. Island Melanesia: The last 10,000 years. In M. Spriggs, D. Yen, W. Ambrose, R. Jones,
A. Thorne and A. Andrews (eds), A Community of Culture: The People and Prehistory of the
Pacific. Occasional Papers in Prehistory 21. Canberra: Australian National University, pp. 187-205.
Spriggs, M., 2000. The Solomon Islands as bridge and barrier in the settlement of the Pacific. In A.
Anderson and T. Murray (eds), Australian Archaeologist. Collected Papers in Honour of Jim Allen.
Canberra: Coombs Academic Publishing, Australian National University, pp. 348-64.
Spriggs, M., 2001. How AMS dating changed my life. In M. Jones and P.J. Sheppard (eds), Australasian
Connections and New Directions: Proceedings of the 7th Australasian Archaeometry Conference.
Research Papers in Anthropology and Linguistics., Auckland: Department of Anthropology,
University of Auckland, pp. 365-74.
Spriggs, M., 2005. Bougainville’s early history: An archaeological perspective. In A. Regan and H.M.
Griffin (eds), Bougainville Before the Conflict. Canberra: Pandanus Books, pp. 1-19.
Stoddart, D.R., 1969. Geomorphology of the Solomon Islands coral reefs. Philosophical Transactions of
the Royal Society of London. Series B, Biological Sciences, 255: 355-82.
Summerhayes, G.R. and I. Scales, 2005. New Lapita pottery finds from Kolombangara, Western Solomon
Islands. Archaeology in Oceania, 40 (1): 14-20.
Swadling, P., 1976. The occupation sequence and settlement pattern on Santa Ana. In R.C. Green and M.M.
Cresswell (eds), Southeast Solomon Islands Cultural History: A Preliminary Survey. Bulletin 11.
Swadling, P., 2000. Changing marine interests and their implications for the settlement history of Santa
Ana, an island in the Southeast Solomon Islands. In A. Anderson and T. Murray (eds), Australian
Archaeologist. Collected Papers in Honour of Jim Allen. Canberra: Coombs Academic Publishing,
Tedder, M.M. and S. Barrus, 1976. Old Kusaghe. Journal of the Solomon Islands Cultural Association, 4: 41-95.
Terrell, J.E., 1970. An Archaeological Survey of Bougainville Island, T.P.N.G. Chicago: Field Museum of
Natural History.
Terrell, J.E., 1976. Perspectives on the Prehistory of Bougainville Island, Papua New Guinea: a Study in
the Human Biogeography of the Southwestern Pacific. Unpublished Ph.D. dissertation, Harvard
University, Cambridge, Mass.
Terrell, J.E., 1977. Human biogeography in the Solomon Islands. Fieldiana Anthropology, 68 (1): 1-47.
Terrell, J.E., 1986. Prehistory in the Pacific Islands: A Study of Variation in Languages, Customs, and
Human Biology. Cambridge & New York: Cambridge University Press.
Terrell, J.E. and G.J. Irwin, 1972. History and traditions in the Northern Solomons: An analytical study of the
Torau migration to Southern Bougainville in the 1860s. Journal of the Polynesian Society, 81: 317-49.
Thomas, T., 2004. Things of Roviana: Material Culture, Personhood and Agency in the Nineteenth Century
Solomon Islands. Unpublished Ph.D. thesis, University of Otago, Dunedin.
Thomas, T., P.J. Sheppard and R. Walter, 2001. Landscape, violence and social bodies: Ritualised
architecture in a Solomon Islands society. Journal of Royal Anthropological Institute, 7: 545-72.
Thurnwald, R., 1912. Forschungen auf dem Salomo-insulen und dem Bismarck-Archipel, Volume 3. Berlin:
Dietrich Reimer.
71
Thurnwald, R., 1934a. Pigs and currency in Buin; observations about primitive standards of value and
economics. Oceania, 5: 119-41.
Thurnwald, R., 1934b. Stone monuments in Buin (Bougainville, Solomon Islands). Oceania, 5: 214-17.
Thurnwald, R., 1951. Historical sequences on Bougainville. American Anthropologist, 53: 137-39.
Tyron, D. and B. Hackman, 1983. Solomon Islands Languages: An Internal Classification. Canberra: The
Australian National University.
Walter, R. and P. Sheppard, 2001. Nusa Roviana: The archaeology of a Melanesian chiefdom. Journal of
Field Archaeology, 27: 295-318.
Walter, R. and P.J. Sheppard, 2006. Archaeology in Melanesia: A case study from the Western Province of
Ward, G.K., 1976. The archaeology of settlements associated with the chert industry of Ulawa. In R.C.
Green and M.M. Cresswell (eds), Southeast Solomon Islands Cultural History: A Preliminary
Report. Bulletin 11. Wellington: Royal Society of New Zealand, pp. 161-80.
White, G., 1979. War, peace and piety in Santa Isabel, Solomon Islands. In M. Rodman and M. Cooper
(eds), The Pacification of Melanesia. Ann Arbor: University of Michigan Press, pp. 109-139.
White, J.P. and C.V. Murray-Wallace, 1996. Site ENX (Fissoa) and the incised and applied pottery tradition
in New Ireland, Papua New Guinea. Man and Culture in Oceania, 12: 31-46.
Wickler, S., 1995. Twenty-Nine Thousand Years on Buka: Long-Term Cultural Change in the Northern
Solomon Islands. Unpublished Ph.D. dissertation, University of Hawaii, Honolulu.
Wickler, S. and M. Spriggs, 1988. Pleistocene human occupation of the Solomon Islands, Melanesia.
Antiquity, 62: 703-6.
Yen, D.E., 1971. The development of agriculture in Oceania. In R.C. Green and M. Kelly (eds), Studies in
Oceanic Culture History. Volume 2. Pacific Anthropological Records 12. Honolulu: Department of
Anthropology, Bernice P. Museum, pp. 1-12.
Yen, D. E., 1973. Ethnobotany from the voyages of Mendana and Quiros in the Pacific. World Archaeology,
5: 32-43.
Yen, D.E., 1976. Agricultural systems and prehistory in the Solomon Islands. In R.C. Green and M.M.
Cresswell (eds), Southeast Solomon Islands Cultural History: A Preliminary Report. Bulletin 11.
Yen, D.E., 1982. The Southeast Solomon Islands Cultural History Programme. Indo- Pacific Prehistory
Association Bulletin, 3: 52-66.
Zelenietz, M., 1979. The end of head hunting in New Georgia. In M. Rodman and M. Cooper (eds), The
Pacification of Melanesia. Ann Arbor: University of Michigan Press, pp. 91–108.
72
Chapter 4
INTER-TIDAL LATE LAPITA SITES AND GEOTECTONICS IN

THE WESTERN SOLOMON ISLANDS
Peter Sheppard and Richard Walter
INTRODUCTION
The speed of colonisation of Remote Oceania (Green 1991) or that part of Oceania beyond the
inter-visible island chains which run down southeast from the coast of Papua New Guinea and
through the main Solomon Islands to the island of Makira, is astonishing. Although the dating of
earliest colonisation may always be debated, it is clear that within a few hundred years all of
Island Melanesia as far as New Caledonia, Fiji and Western Polynesia out to Samoa was settled
by populations making Lapita pottery. Perhaps by circa 3200-3100 B.P. in the Reef/Santa Cruz
(Green, Jones and Sheppard 2008); by 3000 B.P. in Vanuatu (Bedford 2006, Bedford, Spriggs
and Regenvanu 2006) or possibly somewhat earlier (Bedford 2007); by 3000 B.P. in Fiji (Nunn
et al. 2007), and by 2800 B.P. in Tonga (Burley and Connaughton 2007). If we initiate movement
in the region of New Britain, this colonisation represents settlement over a straight line distance
passing directly through the Solomon Islands, the Reef/Santa Cruz Groups and out to Samoa of
4200 km or approximately a tenth of the circumference of the earth. Although this is Oceanic
maritime colonisation, and consequently faster point to point type movement than in continental
style wave of advance models, there is still a significant land area to be settled, especially in the
initial leg through the Solomon Islands (27,556 sq km land area and six islands of more than
1000 sq km), if we are to assume a demographic or density based driver (Kirch 1997: 64).
Although it may be difficult to specify in any useful way motivation, or more likely diverse
motives (Parsonson 1972: 28-35), for such fast and risky behaviour (Anderson 2006, Spriggs
1997: 105) it should be feasible to distinguish between geographical patterns of advance which
can be understood as the result of systemic adaptive processes (e.g., demographic expansion) and
those which appear to be historical opportunism requiring consideration of agency (Anthony
1990). Recently we have argued (Sheppard and Walter 2006) that consideration of archaeologi-
cal, linguistic and biological data for the Solomon Islands strongly suggested that early Lapita
settlement by-passed the main Solomon islands in a leap-frog movement out to the first uninhab-
ited island group of the Reef/Santa Cruz, where they in effect created an enclave or outpost of
the Bismarck Archipelago with which they maintained effective relations over a distance of 2000
km for some hundreds of years as evidenced by the movement of significant quantities of
obsidian (Sheppard 1993). Such a rapid movement is increasingly supported by recent genetic
data (Friedlaender et al. 2008). Similar early pioneering settlements may have existed in northern
Vanuatu (Bedford 2006: 262, 2007), however the details of these sites are as yet unpublished,
leaving the sites of the Reef/Santa Cruz as comparatively unique examples of this “outpost”
behaviour. If we are correct, then the sudden explosion into Remote Oceania is not so puzzling
if seen as the result of an early discovery of unexploited islands 2000 km east of the parent
population centre. Such historical accidents of colonisation may not actually be rare in prehis-
tory. A similar pattern has been observed in the movement of agricultural populations across the
Western Mediterranean and around Iberia. There it is suggested that episodic maritime colonisa-
tion resulted in the creation of enclaves of new agricultural settlers in areas occupied by hunter-
gatherers, accounting for the speed of the Neolithic spread which cannot be accounted for by
traditional wave of advance models (Zhilão 2001).
73
The difficulty of arguing for enclave or leap-frog type behaviour is the problem of demon-
strating a negative. Any suggestion of a blank in the geography of an archaeological distribution
is likely to be met with the “absence of evidence is not evidence of absence” argument and a call
for more fieldwork. The Solomon Islands are often characterised as archaeological terra
incognita, however as we have argued (Sheppard and Walter 2006), there in fact has been
concentrated fieldwork over a considerable period of time in a number of areas, often by
researchers familiar with Lapita archaeological records and settlement patterns. In those areas
where survey has been concentrated it has, however, been suggested that the visibility of early
Lapita sites is obscured or the record destroyed by geomorphologic processes. Clearly under-
standing of the geomorphological context of Lapita sites is critical to evaluation of site distribu-
tions (Dickinson 2001), and this has been increasingly recognised (Bedford 2006, Dickinson and
Burley 2007, Dickinson and Green 1998, Gosden and Webb 1994, Green et al. 2008). In the
Western Solomon Islands Felgate (2003) has argued that geotectonic movements combined with
changing sea level may have destroyed or obscured the record of Lapita settlement. Felgate
(2003: 504) concludes that “Evidence suggesting a Lapita gap in the Near-Oceanic Solomon
Islands as avoidance or leap-frog colonisation is thus substantially eliminated”. In this paper we
will examine the archaeological and geotectonic record of the Western Solomon Islands in order
to evaluate Felgate’s hypothesis.
GEOLOGICAL CONTEXT
Examination of the map (Figure 4.1) of the Solomon Islands shows a 900 km long double chain
of islands which mark the boundary between two tectonic plates. To the east is the Pacific Plate
and to the west the Australian Plate. More specifically we find on the western margin of the Pacific
Plate the Ontong Java submarine plateau, which is the size of the contiguous United States and the
largest submarine plateau in the world. The interaction between these plates and the collision of
the Ontong Java Plateau with the Australian Plate is responsible for many of the distinctive
features of the Solomon Islands. We briefly summarise the main points here as derived from a
variety of more detailed sources (Cowley et al. 2004, Mann et al. 1998, Taylor et al. 2005).
If we look at the geological map of the Solomon Islands we will see considerable variation
between the islands in surface geology but also some similarities. Therefore both Guadalcanal and
Choiseul exhibit a very complex but similar suite of volcanic, sedimentary and metamorphic
rocks while Malaita and Santa Isabel are dominated by marine limestones, and the islands of the
Western Province show the distinctive landforms and rock types of comparatively recent
volcanism. These differences reflect both age and distinct formation processes. Cowley et al.
(2004: Fig. 20) propose the following model for the formation of the central Solomon Islands.
In the late Oligocene and Early Miocene the Pacific Plate was being subducted under the
Australian Plate. At that time island arc basement rock was exposed forming Choiseul,
Guadalcanal and the southeast edge of Santa Isabel. As the Pacific Plate was subducted this
resulted in the Ontong Java Plateau moving toward the Solomon Trench and the Australian Plate.
The arrival of this large plateau in the Pliocene resulted in a blockage of the subduction process,
and instead of being subducted the Ontong Java marine sediments were ultimately accreted onto
the Australian Plate, forming Malaita and the majority of Santa Isabel. With the subduction of the
Pacific Plate blocked, a reversal of subduction occurred with the Australian Plate beginning to
move under the Pacific Plate along the San Cristobal Trench in the late Pliocene. This reversal
then created a Plio-Pleistocene volcanic arc in the New Georgia Group, possibly leading to the
development of the New Georgia volcaniclastic sediments in the Late Pleistocene and Quaternary
74
Inter-tidal Late Lapita Sites and Geotectonics in the Western Solomon Islands
Figure 4.1 The Solomon Islands.
75
(Cowley et al. 2004: 302-4), or possibly earlier as the age of basement rock in the New Georgia
Group is poorly known. Of most interest to archaeologists is the recent geological history, in
particular the formation of the large lagoon systems, barrier islands and uplifted reefs such as
those found in New Georgia, where they form some of the largest lagoons in the world. It is these
rich maritime environments and off-shore islands which have long been considered as targeted by
Lapita settlers (Green 1979), and it is suggested that it was the formation or stabilisation of these
situations in the mid-Holocene which triggered late Holocene culture change (Allen and Gosden
1996, Terrell 2002) related to the Lapita spread.
FORMATION OF THE NEW GEORGIA LAGOON SYSTEMS

The lagoons and raised reefs of the New Georgia Group have been the subject of a considerable
body of research beginning with the work of Stoddart (1969) and more recently the extensive
survey and coring programmes led by Mann and Taylor (Mann et al. 1998, Taylor et al. 2005).
Their work indicates that the lagoons and barrier reefs of the New Georgia Group and in particu-
lar those of New Georgia were established as a result of net subsidence in the Plio-Pleistocene
period. As the islands sank, a fringing reef gradually grew upwards in classic Darwinian fashion
(Darwin 1842), ultimately establishing a barrier reef and lagoon (Mann et al. 1998: 268). This
would account for the formation of the Roviana/Vonavona, Marovo and Nggerasi lagoons which
extend from western New Georgia east to Vangunu and Gatokae Islands. Subduction of the
Pacific Plate has, however, ultimately resulted in uplift becoming the dominant process during the
Quaternary (Mann et al. 1998: 302) as the Coleman Seamount located to the southwest of New
Georgia has moved east at 97 mm per year. This has resulted in sequences of uplifted reef
throughout the New Georgia Group (Stoddart 1969), of which the Holocene portion have been
dated by Mann et al. (1998: Table 1).
Figure 4.2 is a schematic representation of rates of Holocene uplift in the Western Solomons
based on Mann et al. (1998: Fig. 8B). In general terms rates of uplift are greatest in the southwest
with the islands of Tetepare and Rannonga moving up at a very fast rate, while rates decrease to
the northeast with uplift effectively zero on a line running southeast from northern Vella Lavella
across the northern side of Kolombangara. Moving further north and across the New Georgia
Sound the drowned coastline and mangrove swamps of northwestern Santa Isabel and eastern
Choiseul indicate substantial subsidence (Mann et al. 1998: 264). Figure 4.2 is, however, a very
simplified representation of a complicated situation. Mann et al. (1998: Fig. 8A) divide the Western
Solomons into a number of tectonic blocks. A forearc zone includes the islands of Tetepare,
Rendova, Ghizo and Ranongga, while the volcanic arc blocks include the islands of Nggatokae,
New Georgia, Kohinggo, Parara, Kolombangara and Vella Lavella. Within the forearc zone
Tetepare has the highest uplift rate in the Western Solomons at 7.7 mm per year, while Rannonga
exhibits a north-south trend in uplift with rates declining from the south (5.7 mm per year) to
virtually nil in the north. Within the volcanic arc zone, Vella Lavella and Kolombangara form a
block with no or minimal uplift or tilt; the New Georgia block (New Georgia, Kohinggo, Parara
Islands) is uplifting and although Mann et al. (1998: Fig. 8a) report tilting to the northeast and
southwest, their record of Holocene uplift suggests considerable variation in rates of uplift around
the island and within a small area. Although rates in the Roviana Lagoon average somewhat higher
(1.0 to 0.6 mm per year), high rates of 0.8 mm can also be found in north New Georgia, and most
areas seem to have rates of 0.5 mm or higher. Such common overall patterns of uplift have resulted
in the lagoons and barrier reefs/islands which encircle most of New Georgia.
76
77
Figure 4.2 Schematic model of geotectonic processes after Mann et al. 1998 and locations of inter-tidal sites. Width of
arrows is indicative of rate of uplift.
In summary although uplift or stasis are predominant in the New Georgia Group, there is
considerable variation throughout the region and between fault blocks. Although Mann et al.
(1998) have dated a large number of features related to former sea levels and geotectonic
movement in the region and have a model of regional process, the number and resolution of dates
is insufficient for detailed understanding of small zones within which considerable variation in
process may apparently occur over short distances. Moreover, although the expression of uplift
as mm per year gives the impression of a continuous process, in fact the authors are careful to
point out that tectonic movements in the New Georgia Group may be discontinuous with subduc-
tion of the irregular surface of the Coleman Seamount, possibly creating periods of little
movement followed by abrupt (Mann et al. 1998: 302, Taylor et al. 2005) and potentially
catastrophic tectonic movement. Taylor et al. (2005: TC6005) cite evidence from Rendova Island
of several distinct 1–2 m uplifts over the past 2000 years, and recent events (earthquake of Easter
2007) in the Western Solomons would appear to support this model. The implications for archae-
ologists of this variation in rates and tempo are important as it indicates general regional models
of geotectonics cannot be applied in a simple fashion and the settings of individual archaeological
sites need to be carefully considered (Dickinson and Green 1998: 257).
LAPITA INTER-TIDAL SITES

The majority of reported Lapita sites are located in close proximity to modern or former beaches,
however in Near Oceania a considerable percentage are located in the inter-tidal zone. Such sites
appear to be found throughout Near Oceania, from Mussau in the west to New Georgia in the east,
and it seems probable that they are characteristic of settlement in the region and under-reported
given a generally terrestrial survey bias. We might speculate that their distribution will be found
to extend to coastal Papua New Guinea.
In a few cases deposits are found below the water table and they are interpreted as the
remains of former inter-tidal stilt villages which have been covered by prograding beach
sediments. On the island of Eloaua off the coast of Mussau Island Kirch has reported two sites
ECA (Talepakemalai) and ECB (Etakosarai), which apparently were occupied contemporane-
ously and were situated on reef flats facing one another across a small channel now covered by
sand. The site of Talepakemalai was located 30-40 m off-shore from the original island and the
waterlogged deposits capped by the prograding sand have preserved the wooden remains of what
is interpreted as a stilt house in Area B (Kirch 1987, 2001). Various lines of evidence indicate this
was a stilt house in shallow water including: (a) the lack of features cut into the deposit (hearths,
ovens, pits), (b) sand dwelling bivalves in death position, and (c) the lithology of the sediments
from the lower cultural deposit which is indicative of a marine foreshore. As Kirch (1987: 167)
notes: “In sum, the evidence indicates that Layer I11 represents a Lapita stilt-house occupation
situated over shallow, quiet water immediately adjacent to the former shoreline (indicated by a
topographic rise of about 1 m, some 30 m south of Area B)”. This zone of stilt houses extended
for over 10,000 sq m.
A similar suite of Lapita sites has been discovered in the Arawe Islands off the southwest
coast of New Britain. Here the Apalo, Makekur and Paligmete sites are all argued to have been
stilt house settlements formed over reef flats at some distance from the beach (Gosden and Webb
1994). Analysis of the sediment has suggested rapid build up of sand banks underneath stilt houses
as a result of the baffling effect of the piles promoting sand accumulation. As at Talepakemalai the
discovery of well preserved wooden posts and planks at Apalo just above the reef flat and dated
to 3680±50 b.p. and similar material at Makekur, dated at 3200±70 b.p. (on shell) (Gosden and
78
Webb 1994: 38-42), provide direct evidence of the deposits being submerged since deposition.
Dating of coral heads in the bottom of the excavation at Apalo indicates coral growing at 4430±100
years b.p. in water about 2.5 m deep at high tide assuming a 1metre higher sea level at that time.
About 3000 years ago a Lapita settlement was established in water about 2-1.5 m at high tide but
which gradually became shallower by 0.5 m as a sandbank was established. Estimates of water
depth at Makekur are 0.5. to 1 m at high tide (Gosden and Webb 1994: 42).
The Mussau and Arawe sites are the only locations where we have strong direct evidence of
stilt houses over flooded reef flats. In all other locations the sites are found in the inter-tidal zone,
although often extending beyond the tidal range into deeper water. In these situations we must
consider the possibility that the current setting is the result of subsidence and each location must
be evaluated separately. Given that most of these locations are not the primary focus of archaeo-
logical survey, they have not received such evaluations and it seems they are generally assumed
to be the remains of stilt villages. As noted above inter-tidal sites are found throughout the
Bismarck Archipelago. In addition to the Mussau and Arawe Island sites, they are found on New
Britain in the Kandrian region (Kreslo [Specht 1991]), in the Talasea region (FDK [Specht 1981]),
Garala and Boduna Islands (White et al. 2002), and on the Duke of York Islands (SDQ, SDP, SEE
sites [White and Harris 1997]), and possibly the SAD locality at Watom Island (Specht 2003) at
the east end of New Britain. The situation on New Ireland is unclear although water rolled sherds
are reported from Lossu (White and Downie 1980: 203). To the southeast of New Ireland and
across the water gap to the Northern Solomons and the island of Nissan an inter-tidal deposit is
reported at Tarmon (DES). Located on the reef flat and adjacent to the reef passage into the inner
lagoon, Spriggs (1997: 126) suggests it may be the remains of a stilt village occupation or a settle-
ment on a former sand spit. Pottery at Tarmon is sourced to Buka 60 km to the south where a
series of inter-tidal sites have been reported and studied by Wickler (2001). Reef flat sites are
found at Kessa (DJQ) on the northern tip of Buka and on the island of Sohano (DAA and DAF)
off the southern end of Buka. These sites form scatters of ceramics, oven stones and stone
artefacts over often extensive areas (30,000 sq m at DAF) of reef flats with generally sandy
bottoms and are interpreted as the remnants of stilt villages. Although none of this material has
been directly dated, Wickler (2001: 241) seriates these sites based on decorative techniques,
suggesting the oldest site to be Kessa (DJQ) with 50% of decorated sherds (n=188) dentate-
stamped and dating with similar dated assemblages to the range 3000-2800 B.P., and the latest
site with only 1.9% (n=77) dentate-stamped dating in the range 2300-2100 B.P. (DAF beach and
inner reef). Spriggs (1997: 127) suggests a starting date of circa 2700 B.P. for this sequence.
South of Buka, no Lapita or inter-tidal sites have been reported until Vella Lavella in the Western
Solomons, where they are part of a widespread phenomenon which extends at present to eastern
New Georgia. Such sites are common, with more inter-tidal sites reported here than anywhere to
the west, and most appear to date to the very late to immediately post-Lapita time period. We will
discuss these sites and their context in more detail below.
East of the New Georgia Group and throughout Remote Oceania out to Samoa there are,
with the possible exception of the Bourewa site in Fiji (see Nunn and Heorake this volume), a
locality (SE-SZ-45) on Santa Cruz (Green et al. 2008: 52) and the Mulifanua site in Samoa, no
reported inter-tidal sites. At the Ferry Berth site at Mulifanua on Upolu, Samoa, a deposit contain-
ing Lapita ceramics was discovered at 2.25 m below sea level (Dickinson and Green 1998: 240).
This was originally assumed to be the result of subsidence, however, that was at variance with
general geological models which concluded uplift was the dominant process in the region.
Therefore this site was considered either a product of localised subsidence or alternatively a stilt
79
village. Careful reconsideration of the site setting and the geology of Upolu have led Dickinson
and Green (1998: 257, Dickinson 2007a) to conclude that the site is submerged as the result of
subsidence and not deposition from a stilt village. Whatever the scenario at Mulifanua, it remains
true that no other definite inter-tidal sites have been discovered in Remote Oceania, and this
would appear to be characteristic of the region, possibly representing a change in house design or
settlement pattern in these previously uninhabited lands (Dickinson and Green 1998: 258, Green
and Pawley 1999, Kirch 1997: 184) . Elsewhere in Remote Oceania uplift and/or progradation of
shorelines appear to be significant processes, resulting in many Lapita sites being associated with
palaeobeach lines at some distance from the modern shore (Dickinson and Burley 2007,
Dickinson, Burley, and Shutler 1999, Green et al. 2008, Nunn 2005, Nunn et al. 2007). In the
following section we return to the Western Solomons to look more closely at the age and context
of the inter-tidal sites found there.
WESTERN SOLOMON ISLANDS INTER-TIDAL SITES

Figure 4.3 shows the distribution of reported inter-tidal sites in the Western Solomons. Most of
these sites have been discovered as part of our research programmes beginning in 1996 (Felgate
2001, 2003, Sheppard et al. 1999, Sheppard and Walter 2006), however the first report in Roviana
Lagoon was by Reeve (1989), who noted the presence of similar sites on north New Georgia and
Vella Lavella. This built on the work of Miller (1979), who reported finding ceramics below the
water table at the back of a beach on Choiseul (Nuatambu) during the National Museum Survey
of the late 1970s. Following our work in Roviana Summerhayes and Scales (2005) have reported
inter-tidal sites from Kolombangara discovered by Scales during his linguistic fieldwork. More
recently Felgate has surveyed along the north coast of New Georgia and reports (Felgate 2007)
Figure 4.3 Inter-tidal sites and other survey locations noted in the text.
80
discovering inter-tidal sites in the Nggerasi Lagoon. These results and our experience suggest that
inter-tidal sites are ubiquitous throughout the Western Solomon Islands. Figure 4.4 records the
location of 29 sites but does not include the site of Lolomo (Felgate 2007: 232) and other
unreported sites discovered by Felgate (pers. comm. 2004) on north New Georgia. Including
those could bring the number into the mid-30s, and our experience suggests that wherever lagoons
are located inter-tidal sites will be easily found, especially if survey is conducted in the inter-tidal
zone during the low water season of June/August when very low tides occur during mid-day
(Felgate 2003: 279).
These sites contain the earliest ceramic assemblages found in the Western Solomons.
Nowhere have similar assemblages been found above the inter-tidal, with the exception of the
Nuatambu site on Choiseul where apparently similar ceramics are found below the current water-
table (Miller 1979), and Irigila on north Vella Lavella where rolled and weathered ceramics are
found on the surface in the village behind the modern beach ridge. Reeve (1989) originally
reported Irigila, however our test excavations indicate that the deposit has been churned by land
crab activity to a depth of over 1.8 m and the original deposit may have been on the now buried
lagoon floor. A large Trochus shell which appeared to be in position just above the old beach level
was dated (Wk-16583 1372±36 b.p.) to 1050-1180 A.D. (1 sigma, Delta r=0).
Following these very late Lapita sites there is however a possibly continuous ceramic
sequence up into recent prehistory and on north Choiseul into the historic period. In every
location where we have conducted fieldwork (Roviana, Rannonga, Vella Lavella) we have found
terrestrial ceramic assemblages of a generally thin plainware. This is in keeping with the finding
of Miller (1979), who reported 13 similar sites (Simbo 4, Choiseul 4, Kolombangara 5) with
generally low density ceramic deposits. Most recently David Roe (pers. comm. 2007) has found
similar materials in the Bugotu region of eastern Santa Isabel. The extent to which these ceramics
are traded in from Choiseul or Bougainville (Kraus 1945, Ogan 1970, Specht 1972, Wheeler
1928), which had significant pottery production up into the historic period is as yet unknown,
however the low density of ceramics in late prehistory in Roviana suggests they are not locally
made and this thin ware is very similar to that in historic Choiseul pots. The point at which local
production ceases is not clear, however we have found one relatively dense deposit of thin
plainware (Site 25) in the interior of west New Georgia dated to circa 1400 A.D. Our point is not
to review these later assemblages here but to indicate that unlike the central or eastern Solomon
Islands ceramics are commonly found in the Western Solomons and form part of a tradition which
begins in the late Lapita period.
All of the inter-tidal sites we have studied share common characteristics. They are
essentially lag deposits found below the high tide line. Repeated attempts to find any in situ or
stratified deposit have been unsuccessful, most if not all of these deposits sit on relatively hard
coral gravel bottoms. This may represent the choice of suitable bottoms for establishing pile
dwellings or simply reflect the fact that the visibility of such sites will be low in soft muddy
sediments (Felgate 2003: 162). We have recently (Sheppard, December 2007) returned to these
sites as well as locations showing raised tidal notches and conducted shore profile surveys in
order to investigate the relationship between site locations and the history of lagoon formation
and possible uplift. Where possible, heights were taken above current live coral, however in many
locations no live coral was present. All results presented below were surveyed from the water line
using a laser ranger finder and then calculated, based on the time of day, as heights above 0 tide
datum (Mean Lower Low Water) based on computer model data (XTide) with a reference point
at Nusazonga opposite Munda (David Flater. XTide version 2.8.2. http://www.flaterco.com/
xtide/. 2007-12-24). This has the effect of standardising the height measurements. Where heights
81
above highest live coral could be recorded, we found close correlation with the data generated
using the computer modelled 0 datum values, with highest live coral matching the zero datum.
The sites are generally well exposed at mid-day in June to August when tides are often at
maximum low of circa -90 cm below the maximum high tide line. Many sites are located on reef/
mangrove flats (Figure 4.4) and are essentially dry throughout their extent during low water,
while others on sloping bottoms extend off into deeper water (e.g., Zangana in Roviana). The one
site which does appear to have been sub-tidal is the site of Honiavasa in Roviana Lagoon, which
was collected by Felgate on a rising tide on Sept 20th 1997 in “up to a metre of water” (Felgate
2003: 444) Tide modelling predictions indicate that low tide at the station in Roviana Lagoon
(Nusazonga) for that date was at 10:26 AM, with a reading of 0.36 m while high tide was at 5:58
PM and 0.65 m. On July 20th 1997 the low tide at 1:34 PM was -0.02m (David Flater. XTide
version 2.8.2. http://www.flaterco.com/xtide/. 2007-09-03) suggesting a maximum sub-tidal
depth of circa 50-60 cm at Honiavasa (13 km east of the Nusazonga tide reference). Today all of
the known inter-tidal sites are covered by 0.9 to 1.5 m of water at maximum high tide. All sites
are today located either directly adjacent to or a short distance out from the current beach in
protected lagoon settings suitable for the establishment of stilt house villages and offering
sheltered anchorages. Building these settlements during low water when the flats are essentially
dry would not be too difficult, however once water rose above a metre the effort would increase
substantially. Today people often build houses out over the inter-tidal zone but only very rarely
are structures built out beyond the tidal range.
The common features of these settlement locations suggest a shared cultural tradition and
this is confirmed by the range of ceramic wares, forms and decorative styles found in the assem-
blages. Of the sites recorded to date only a very few have any dentate-stamped sherds (maximum
of 6 at Honiavasa), carinated pot forms (Honiavasa, Zoraka/Nusa Roviana, Poitete) (Felgate
2001, 2003: Figs 7, 9, Summerhayes and Scales 2005), or the occasional piece of good quality
Figure 4.4 Surface collecting at the Humbi Konqu (Roviana Lagoon) inter-tidal ceramic site.
82
obsidian (Paniavile, Pusimao), indicating the earliest assemblages are of late Lapita age. The
majority of the decorated ceramic assemblages consists of various proportions of rectilinear
incised, appliqué (on surfaces and rims), punctuates and finger nail impressions (Figure 4.5). A
variety of rim treatments are found, however crenulated or “wave-deformed” (Felgate 2003: Fig.
24, 41-43, Table 42) rims are very characteristic of this ceramic tradition (Figure 4.5: bottom row
A and C). As these assemblages are all lag deposits it is difficult to assess their age or the time
span represented within a site. Two radiocarbon dates have been obtained on carbon extracted
from sherds. A charcoal inclusion in a sherd from Paniavile (AA33504 2130±90 b.p.) is dated to
390 B.C.-30 A.D. (calibrated 2 sigma range) while carbon encrustation on a sherd from Hoghoi
(NZA-1253 2619±45) is dated to 900-550 B.C. (calibrated 2 sigma range). Table 4.1 presents data
Figure 4.5 Inter-tidal ceramics from the Western Solomons (top row: a, b. Honiavasa, c. Hoghoi, Roviana;
middle row: a, b. Seghe, Marovo Lagoon, c. Matu Soroto, Vella Lavella; bottom row: a, b. Matu
Soroto, Vella Lavella, c. Totutotu Bay, Vella Lavella.
83
Table 4.1 Seriation of Roviana (Kalikoqu area) inter-tidal sites (data from Felgate 2003: Table 42).
84
on decorative attribute counts for the Roviana sites analysed by Felgate (2003) in his Table 42.
Here we have re-sorted the table using the most common attribute – lip deformation “wave”
pattern. This creates a good seriation in most of the other attributes and is in agreement with the
radiocarbon dates and the earlier dentate stamping; suggesting we have a useful chronological
order, ranging from Late Lapita through a post-Lapita period ending with Paniavile or circa 900
B.C. to 30 A.D. Therefore most, and possibly all, sites of this tradition have a minimum age of
2000 B.P. Together these sites form a chronological marker throughout the Western Solomon
Islands of potential use in studying localised geotectonic processes. In the following section we
will look more closely at the setting of inter-tidal sites in Roviana.
GEOTECTONICS AND INTER-TIDAL SITES IN ROVIANA LAGOON

Mann et al. (1998) provide considerable data on geotectonic movement in Roviana Lagoon,
making it an ideal location to integrate archaeological data with that from geology, and examine
what light can be shed on the formation of both records. Starting at the western end of the lagoon
(Figure 4.3) we will move east examining the setting of sites within each area.
The western end of Roviana Lagoon is marked by the disappearance of barrier islands which
only start again to the west of Munda at the entrance to Vonavona Lagoon just west of New
Georgia. In this area in front of Munda there is in fact extensive coral reef which is an extension
of the uplifted reefs which form islands further to the east. The position of the modern reef in front
of Munda suggests very limited uplift in this area. Moving to the east you encounter very low,
narrow barrier islands where uplift is minimal and much of the interior of the islands is swampy
and just at or above sea level. This pattern continues until the island of Nusa Roviana which
appears to represent a sudden change as it has a maximum height of 76 m, however the southern
end of the island is very low and swampy and like contiguous islands to the west and east. The
main part of Nusa Roviana appears to represent a different geological phenomenon. The central
ridge on Nusa Roviana is oriented north/south at right angles to the coast of New Georgia and to
the long axes of the other barrier islands. It is also the highest of the barrier islands and the top
appears to consist of heavily recrystallised limestone. The height and composition of this deposit
is consistent with Pre-Quaternary recrystallised limestone found at 70+ m in neighbouring east
New Georgia. It seems likely that the main body of Nusa Roviana was a small older Early
Quaternary or Pliocene offshore island formed from an eroded remanent of the last inter-glacial
reef which forms much of the New Georgia coastline (Mann et al. 1998: Fig 5a; 277). Mann et
al. (1998: Table 1) have dated a palaeoreef flat on Nusa Roviana which is 3 m above modern live
coral at (TX7596) 6220±80 b.p. This would appear to be a date on a terrace which encircles the
main ridge and sits above the low swampy flats on the south end of the island. If this dates a reef
associated with the mid-Holocene high stand which Mann et al. date in the range 5500-6000 B.P.
and 1-2 m above current sea level then it is uplifted by 1 to 2 m. Mann et al. (1998: Table 2)
calculate an uplift rate (based on 1 m high stand and 5500 B.P. age) of 0.5 mm per year for this
location. The presence of late Lapita ceramics (Zoraka/Nusa Roviana) in the inter-tidal zone off
the north side of the island indicates occupation from at least 2500 B.P. Shore profiles were not
collected at this location as it was heavily modified during WWII. Two additional inter-tidal sites
of a later period are found at the northeast corner of the island (Rango Bay, Kunduru). These
locations should have been uplifted, using the Mann et al. rate, at least 1.25-1 m if we assume a
constant rate of uplift in the time period 2500-2000 B.P. Alternatively if we simply consider
change in relative sea level of the location since the time of the 3 m reef flat formation (6670 -
6450 B.P. 68% CI), assuming a constant rate of uplift and post high stand drawdown, the uplift
from 2500 or 2000 B.P. is 1.16 or 0.89 m. Uplift however may not have been constant.
85
Stone constructions related to the Late Roviana period and associated with the large hillfort
which dates from at least 1600 A.D. are found throughout the island and down to the coast. These
include coral cobble wharfs which today project out into the inter-tidal zone and shrine and house
features which, especially on the narrow coastal flat on the eastern side of the island, extend down
to the water’s edge. Although we do not have age of construction for these features they were
clearly present in the 19th century and most probably before then. Figures 4.6 and 4.7 show the
current shore profiles based on a survey (using a Laser range finder) transect at right angles to the
modern beach (survey commenced at the water line). Both plots show terraces at circa 2 and 6
metres (above 0 tide datum). If we assume 1.0 m of uplift then both sites were at least 15 m from
the beach 2000 years ago.
Figure 4.6 Shore profile at Ranngo Bay (Nusa Roviana Island) inter-tidal site (heights above 0 tide datum).
Figure 4.7 Kunduru (Nusa Roviana Island) inter-tidal sites (height above 0 tide datum).
Additional information on uplift can be derived from pollen studies on the neighbouring
island of Ndume which is immediately east of Nusa Roviana. This is a very low island with boggy
deposits and a brackish lake in its interior. Grimes (2003) has analysed a pollen core from this
lake (Tamberamakoto, 3 km east of Nusa Roviana). The bottom of the core dates to (UWA-13)
3220±60 b.p. (3590-3330 B.P. 95% CI), at which time the immediate environment is interpreted
as a coastal (possibly estuarine) littoral forest containing large forest species including
Barringtonia which grows to over 20 m. At 3000 b.p. (3350-3000 B.P. 95% CI) the sea level is
interpreted as falling, with the production of a brackish to fresh water estuary or backswamp as
evidenced by fresh to brackish water loving gastropods (Sermyla spp) at (UWA-14) 2750±60 b.p.
(2990-2750 B.P. 95% CI) (Grimes 2003: 158, 176). The data would indicate that this very low
86
island was well established and forested by at least 3000 B.P. The change in sediments and forest
type which is interpreted as falling sea levels may of course also relate to uplift. Since 3200 B.P.
this island would have been uplifted by 1.6 m assuming the Nusa Roviana rate of 0.5 mm per year.
Figure 4.8 is a profile from the north coast south across the island to the lake. Although the island
is very low it would have been well above water 3200 years ago. The only apparent paleoshore-
line feature is indicated by an inflection at circa 3 m.
Figure 4.8 Tamberamakoto (Ndume Island) pollen coring location profile from shore (left) to lake (height
above 0 tide datum).
Figure 4.9 Miho inter-tidal site (Ndume Island) shoreline profile from top of live coral (height above 0
tide datum).
The village of Sasavele in the area of the Lagoon known as Kalikoqu is located two kilometres
east of Lake Tamberamakoto. It is situated on a low, flat, 200 m wide north-south (800 m) trending
ridge which is at least 5 m above the low lying swampy section of Ndume Island that begins
abruptly immediately west of the village. The island profile at Sasavele (Figure 4.9) consists of a
very narrow coastal strip of circa 5 metres width, a gradual slope rising up to the bottom of exposed
makatea (2.86 m elevation) which has a height of 3.16 m forming a possible terrace, followed by
another gradual rise to the level (5.3 m) area of the village. Sasavele is located on a reef passage
which separates Ndume and Honiavasa Islands. The profile on Honiavasa Island is very similar to
that at Sasavele Island, although in the Honiavasa Passage there is a low cliff on the Honiavasa side
requiring a scramble up the cliff out of a canoe of over 2 m. On the north coast (Figures 4.10 and
4.11) a low coastal flat runs inland until it hits raised makatea. The mainland New Georgia coastline
in this area typically is fringed with mangroves leading gradually up to a low coastal flat no more
than circa 1 m above the high tide and extending back a variable distance to a rise which would
87
appear to mark an old shoreline and correlate with the mapped (New Georgia Group Geology 1:
250,000 1987) contact of Pleistocene backreef and lagoonal sediments (Qs) with Holocene alluvial
deposits (Qf). Presumably this represents the mid-Holocene high stand shoreline.
In this region we have five inter-tidal sites studied by Felgate (2003) – Miho on Sasavele
Point, directly across the passage from it Honiavasa on Honiavasa Point, Hoghoi located 1.6 km
east of Honiavasa Point, while on the mainland Zangana is located 1.9 km across the lagoon east
of Sasavele, and Gharanga 5 km north of .Sasavele near the mouth of the Gharanga Stream. With
the possible exception of Honiavasa these sites are all inter-tidal. The rate of uplift for Honiavasa
could be estimated using a date (RER-A TX7593) from the western end of Rereghana Island
across the passage from Honiavasa Island. Here oysters encrusting a notch 0.42 m above living
inter-tidal oysters gave a date of 180±40 b.p. (uncalibrated or corrected for marine reservoir)
(Mann et al. 1998: Table 1). This date is anomalous, as modern marine shell routinely dates 400
years old as a result of marine fractionation, and was confirmed by our paired matching of
charcoal and shell dates on Nusa Roviana. Even if we accept the date, it is still Modern when
calibrated and can not be used to calculate a rate with any confidence. Although sudden recent
uplift (Mann et al. 1998: 285) is clearly possible in this area, as shown by recent events, this
evidence seems problematic. However if we accept the overall Holocene uplift rate for this point
of 0.7 mm per year (Mann et al. 1998: Table 2) and use it to estimate the uplift on western
Honiavasa then the Hoghoi site has been uplifted something in the range of 2.0 to 1.8 m since its
Figure 4.10 Honiavasa Point (Honiavasa Island) inter-tidal site shoreline profile from top of live coral
(heights above 0 tide datum).
Figure 4.11 Hoghoi (Honiavasa Island) inter-tidal site (heights above 0 tide datum).
88
occupation. At that time the location would have been covered at high tide by circa 3 m of water
without considering the possibility that sea level might not have yet fully receded from the
mid-Holocene high stand (assumption of a constant rate of drawdown from a 2 m high stand
would provide an additional circa 1 m of water). If we accept the seriation in Table 4.1 then the
Honiavasa site is older than Hoghoi, which given the presence of some dentate stamping seems
reasonable, therefore that location has been uplifted more than 2 m. If that site is in fact sub-tidal,
as discussed above, then the total depth of water at high tide during the occupation would have
been more than 3 m (or possibly more than 4 m given a high stand contribution). Based on the
seriation the Miho site, on the point across the channel from Honiavasa, is older than Paniavile
(390 B.C.-30 A.D.) and would have been uplifted somewhat less (<1.7 m) than Hoghoi.
If in fact such changes in elevation have occurred then all of these sites would have been, at
the time of occupation, further from the shore than they are today. On the barrier islands the
distances would have been 10 m at Miho, 15 m at Honiavasa Point and 80 m at Hoghoi. On the
mainland at Zangana and Gharanga the distance from the shore, if we assume similar rates of
uplift to Hoghoi, would have been 25 m (Zangana) and 80 m (Gharanga). The sites at Zangana
Point and Gharanga Stream (either side of the stream mouth) seem optimally located with respect
to the current shoreline and position of the current mouth of the Gharanga Stream which meanders
across the flat back of the site location.
Figure 4.12 Zangana (New Georgia) inter-tidal site (heights above 0 tide datum).
Figure 4.13 Gharanga (New Georgia) inter-tidal site (heights above 0 tide datum).
89
Examination of the profiles for paleoshoreline indicators shows: a feature marked by

exposed makatea (2.86 m bottom, top 3.16 m) on Sasavele Point (Miho) at circa 3 m; a possibly
similar feature at Honiavasa Point is indicated by inflections at circa 2 m and 3 m; no obvious
inflection in the makatea cliff back of the coastal flat at Hoghoi; an inflection at circa 2 m at
Zangana (Figure 4.12); and an inflection at circa 2 m at Gharanga (Figure 4.13) which runs up to
the steep makatea slope whose basal elevation is 2.62 m.
The next group of inter-tidal sites are located on the mainland 5 km northeast of Hoghoi. We
have not located any inter-tidal sites on the barrier islands in this zone, although a few sherds were
recovered from the bay on the east side of Mbaraulu village at the eastern end of Honiavasa Island.
The coast at the village is very steep rising up to 40 m and forming cliffs on the passage between
Honiavasa and the next island Rereghana. A small island at the lagoon entrance to the passage
shows a cliff rising from the water with multiple wave notches above the current notch suggesting
considerable uplift in this area. Examination of a cave on this island, whose base was just above
high tide, did not produce any cultural deposit. Across from Mbaraulu village approaching the
mainland and west of the inter-tidal sites is the islet of Masighe where we conducted test excava-
tions in a rockshelter formed by a wave notch (Figure 4.14) 2.29 m (base of notch) above the 0
tide datum. Here we found shallow cultural deposits of shell midden dating from the last several
hundred years, under which was a natural beach deposit of crushed oyster shell in a muddy
sediment. Dating by the Waikato Laboratory of oyster shell from this beach deposit produced an
age of greater than 50,000 B.P. (Fiona Petchey pers. comm. 2007). Field examination of the grain
size of a cemented beach deposit (John Dodson pers. comm. 1998) in the back of the notch (Figure
4.14) indicated it was formed in a higher energy environment than exists in this sheltered lagoon
location today. A sample of oyster shell taken from the top of the cemented deposit in the back of
the notch (2.79 m above 0 tide datum) and identical to the beach deposit under the excavated
cultural deposit was collected in December 2007. It has also returned an age (Beta – 240200) of >
50,000 B.P. (AMS acid etched, 13C/12C +2.1 o/oo), indicating the formation of this notch greatly
precedes the Holocene highstand and possibly relates to the 5a or 5c inter-glacial (Linsley 1996).
The inter-tidal sites of Omaru and Pikoro are clustered together along a short stretch of the
mainland. Omaru is opposite a fresh water spring which seeps out on shore above the high tide
line and Pikoro along the banks and at the mouth of the Pikoro Stream. Like many of the streams
and rivers along the mainland coast, the Pikoro Stream, like its neighbour (Kopo Stream) to the
west, meanders through the very low coastal flat, which in this case is a sago palm swamp
extending for over 100 m from the Pikoro Stream mouth to where it hits the contact with the ridge
Figure 4.14 Masighe Rockshelter profile (Masighe Island) (heights above 0 tide datum).
90
which then rises steeply into the interior. Sherds are common along the edges of the stream and
at its mouth, apparently deposited on the surface by crab activity and clearance of silt from the
stream. People have clearly been using it as a source of fresh water for some time. The exact time
depth represented by this activity is unclear and detailed analysis of the ceramics has not been
completed, however sherds with rectilinear incised motifs and crenulated rims, like those found
in the early inter-tidal sites form part of this assemblage and that at Omaru. Local landowners
reported finding a large decorated (appliqué lizard on rim) ceramic sherd while digging a drain in
the sago swamp, circa 40 m inland from the coast. Test pitting in the area failed to find any more
definite cultural deposit, beyond some fire cracked rock and a few small sherd fragments,
although excavation was carried out to below the water table (-20 cm dbs) in a number of
locations. In both the Pikoro and Omaru locations only very small increases in sea level or depres-
sion of the coast would destroy the relationships between the sites and the fresh water sources to
which they are clearly oriented. Unfortunately, since our earlier studies the area has been heavily
modified by the construction of logging roads and coastal loading areas.
Seven kilometres west of Pikoro is the next group of inter-tidal sites, all located on the north
coast of the barrier island of Ndora. The first of these is Paniavile, the site originally reported by
Reeve (1989). The site is located in the shallow lagoon just around the point from the deep
passage between Ndora and Rereghana Islands. At low water there are extensive reef flats
exposed in front of the modern village located above the site. The site is found just off shore and
extends for a considerable distance parallel to the modern shore. The shore profile (Figure 4.15)
is similar to that found to the west with a narrow (10-15 m) coastal flat, which then quickly rises
up some metres to a wide (circa 30 m) terrace, above which is a high steep makatea cliff which
has at its base a wave cut notch of considerable depth. As you progress south through the village
and along the passage, the cliff approaches the shore at the point where the modern village ends,
forming coastal cliffs estimated to be over 20 m in height.
Excavation in the rockshelter formed by the wave notch revealed a shallow aceramic
cultural deposit from which we have obtained a date (WK-4586 2510±50) on mangrove clam
shell of 2310-2040 B.P. (95% CI delta r=0). Contemporaneous with this is a charcoal inclusion in
a sherd from Paniavile (AA33504 2130±90 b.p.) dated to 2340-1990 B.P. (calibrated 2 sigma
range). If the rate of uplift is similar to that estimated for the western end of Rereghana Islands
(0.7), then 2000 years would produce uplift of 1.4 m and the site would have been located 10+ m
off-shore from a steep coastline leading up to the base (3 m) of a terrace (circa 5+ m) on which
most of the modern village is located.
Figure 4.15 Paniavile (Ndora Island) inter-tidal site (heights above 0 tide datum).
91
Two more sites are located to the east (7.3 km) of this site in Humbi Kongu Bay. Punala
(Figure 4.16), located in mangrove flats on the west side of the bay and Humbi Konqu (Figure
4.17) on the eastern point. The site at Punala extends 100 m across a mangrove flat and up to the
mouth of a small fresh water stream which has sherds scattered along its banks. Behind a very
narrow beach strip is a low backswamp flat which extends back 50 m distance into the coastal
scrub where it climbs to an inflection at 2 m and another at 3 m. A similar coastal profile is found
back of the Humbi Konqu site with an inflection at circa 2 m which runs up to the base (2.3 m)
of a makatea cliff which has a height of 3.88 m. In both cases a 1 m increase in water level would
leave both sites a very considerable distance from shore, and remove the association of the Punala
site with the fresh water stream which would have been a significant resource on these generally
waterless barrier islands. The age of these sites is unknown, however they include the standard
suite of rectilinear incised decoration and crenulated lip treatment, indicating they are more than
2000 years old based on the seriation (Table 4.1) from the western end of the lagoon.
In the far eastern end of the lagoon beyond Humbi Konqu we have only discovered two
inter-tidal ceramic sites, both located on the mainland. The site at Kazu is a small deposit of
sherds just off a beach and wide low coastal flat, while only a few sherds were recovered from a
similar situation at Mare Point where construction of a logging wharf and booming area had
apparently disturbed a site. On the barrier islands we have not found any inter-tidal sites, however
Figure 4.16 Punala (Ndora Island) inter-tidal site (heights above 0 tide datum).
Figure 4.17 Humbi Kongu (Ndora Island) inter-tidal site (heights above 0 tide datum).
92
we located remains of a fortification (post-1800 A.D.) at Saikile at the eastern end of Ndora Island
where extensive stonework is located just above the high tide mark. To the east on the eastern end
of Petani Island at Araroso Point, we found a skull shrine and midden under the shelter of a deep
wave cut notch (Table 4.2) cut into a 16.5 m high cliff located 20 m back of the modern beach at
a height of 2.3 m (base of notch) above mean low tide. This location is along the passage to the
right of the location shown by Mann et al. (1998) in their Figure 10b and the site is located within
the lowest very deep notch seen in the photo. This notch extends along the passage and beyond
and is the lowest of at least 4 distinct notches, with the higher notches most probably pre-Holocene
in age (Mann et al. 1998: 285). Excavation of the deposit found under the notch, which is up to
5 m wide, showed 30 cm of aceramic midden rich in shellfish and fishbone, suggesting the site
was used as a fishing shelter up into the recent past (Walter and Sheppard 1997). The shelter was
occupied when visited in December 2007. This lowest notch is clearly not active today.
Table 4.2 Poroporo Shelter (Araroso Passage).
The notch above this lowest one has been dated, further along the passage, by Mann et al.
(1998: Fig. 10) at 6430±70 b.p. (TX7595) (6920 -68% CI- 6710 B.P.), using coral from the base
of the notch at 3.7 m above living coral. This site provided Mann et al. an annual Holocene uplift
rate of 0.7 mm per year, although the cliff heights at this end of the lagoon are over 20 m, indicat-
ing over the longer term higher rates of uplift (or distinct episodes) than in the western end of
the lagoon. If we ignore the problematic estimate of the height of the mid-Holocene high stand
(Mann et al. 1998 have used 1 m) and directly calculate the amount of relative sea level change
since the formation of the coral, the change is slightly more than 1 m in 2000 years and 1.3 m
in 2500 years.
In summary a number of general observations can be made about the record of uplift and the
distribution and location of inter-tidal sites in Roviana.
1. There is a general trend of increasing height of barrier islands from west to east (0 to 20+ m),
suggesting limited long-term effect of uplift in front of Munda and significant effect at the
eastern end of the lagoon.
2. Multiple distinct notches are found in the eastern end of the lagoon, as seen at Araroso
Passage (4 notches), but also found as far west as eastern Honiavasa Island.
3. The presence of distinct notches would suggest stasis at a series of heights indicating
sporadic uplift events rather than gradual uplift (Taylor et al. in press).
93
4. The study of Mann et al. (1998) date a notch (2nd notch) at 3.7 m above live coral at Araroso
Passage and a reef flat on Nusa Roviana at 3 m as older than 6000 B.P. A beach deposit on the
floor of a notch incised at 2.29 m (above 0 tide datum) on Masighe Island in the central lagoon
dates older than 50,000 B.P., and an associated cemented deposit at the back of the notch (2.79
m) also returns a greater than 50,000 B.P. age. The grain size of these sediments suggests
deposition in a radically different energy environment than the currently protected lagoon.
5. The lowest notch (base 2.3 m, top 3.48) as seen at Araroso is very deep, suggesting stasis at
that level for some considerable period of time since the mid-Holocene. A notch with a
similar elevation is found at Masighe, and related features can be seen as inflections on
profiles at circa 2 and 3 m above mean low water throughout the lagoon, where in the
absence of exposed cliffs it can form a steep bank at the back of the coastal flat. If this is
correct, then a significant uplift rate or event after 6000 B.P., uplifted notch 2 followed by a
long period of stasis.
6. The height of this lowest notch would suggest it correlates, at least in part, with the
mid-Holocene high stand. Mitrovica and Peltier (1991) infer a most probable high stand of
2.0 m for Fiji and 2.5 m for New Caledonia. Dickinson (2001) reports 2.2 m for Tuvalu and
2.4 m for Majuro in the Marshall Islands. It seems very probable that the value for the
Solomon Islands should be in the 2-3 m range and not the 1 m used by Mann et al. (1998).
7. Inter-tidal sites with similar ceramic assemblages, dating to the same time range (2600-2000
B.P.), are found in similar locations both on barrier islands and the mainland along the length
of the lagoon.
8. The association of inter-tidal sites with current stream mouths and springs suggests little
change since the period of occupation.
9. Sites located on the mainland are off-shore from a low swampy coastline and often near the
mouths of meandering streams. It seems probable that the current shore is prograding with
sediment deposition from the interior since the re-establishment of the lagoon in the
mid-Holocene. There is little evidence of recent uplift on this coast, with the current shoreline
dating to pre-2000 B.P.
DISCUSSION
If the inter-tidal archaeological record of the Western Solomons is biased through destruction of
the early Lapita part of the record, then it requires a widespread systematic process which
destroys the early record and leaves the late Lapita sites. If we were to get uplift which raised the
early record above the tide line then weathering and swash activity would quickly destroy salt
impregnated sherds. This has likely happened in some areas with very high rates of uplift,
although these coastlines, such as the rough steep coastline of Rannonga or much of Rendova,
would never have been conducive to construction of stilt house settlements. In these areas there
is evidence to indicate uplift has often occurred episodically, either from geological study in the
case of Rendova (Taylor et al. 2005: T C6005), or most recently on eastern Rannonga with an
uplift of over 2 m after the earthquake of April 2007 (Taylor et al. in press). This most recent event
has also demonstrated the spatial variation across the Western Solomons with significant differ-
ences in uplift over short distances. It has also demonstrated significant subsidence in areas where
there is net overall Holocene uplift such as Roviana. Taylor et al. (in press) measured considerable
subsidence in the western end of Roviana and Vonavono Lagoons as part of a general survey
shortly after the April 2007 earthquake. This formed a west-east trend from a maximum of -0.66
94
m west of Munda, through -0.49 m at Sasavele and -0.36 m at Ndora Island in central Roviana.
At an uplift rate of 0.5 mm a year (Nusa Roviana), this represents 1320 to 720 years of uplift.
Taylor et al. (in press) indicate that each major earthquake or mega-thrust event in this region will
result in similar subsidence with some unknown time before net uplift is resumed.
One effect of this behaviour would be creation of wide solution notches such as the lowest
notch at Araroso Passage. It is difficult to factor unknown but expected periods of subsidence into
our understanding of the Roviana inter-tidal record. Subsidence is something which is hard to
incorporate into models of Holocene geotectonic movement in the absence of coring programmes
(Dickinson 2007b). A half metre of subsidence would tend to preserve archaeological sites,
however it may also mask earlier periods of faster uplift which may have destroyed sites in
Roviana. Mann et al. (1998) have documented considerable geotectonic variation across the
region, and although there seems to be a northwest-southeast trend with uplift effectively zero on
a line from north Kolombangara across through north Vella Lavella, there seems to be consider-
able variation over short distances even within tectonic blocks such as New Georgia. The very
recent record of significant subsidence (Taylor et al. in press) adds to the picture of variability
throughout the region, again indicating the difficulty of applying simple regional models to
processes impacting the archaeological record.
Although archaeological survey cannot be argued to be systematic across the New Georgia
Group of the Western Solomons, it has been conducted throughout the region for more than 30
years and inter-tidal sites have been known for more than 20 years. Our own research beginning
in 1996 and lasting until 2007 has involved considerable survey over a period of years with
particular attention to inter-tidal sites, and has been conducted in widely separated areas (from
north Vella Lavella to Marovo Lagoon) on different tectonic blocks where uplift has ranged from
zero to the very high rates of Rannonga. This work has been supplemented by that of others on
Kolombangara and north New Georgia. In all these areas a similar suite of late Lapita ceramics
has been recovered. The probability of not finding an early Lapita inter-tidal assemblage rich in
dentate-stamped ceramics, such as are commonly found in the Bismarck Archipelago and
northern Solomons, given the survey coverage, would appear to be low. Clearly a uniform geotec-
tonic process cannot be invoked to explain the absence.
In Roviana we have located a large number of inter-tidal sites in a comparatively well
studied geotectonic and geomorphological setting (Mann et al. 1998). If we assume a Holocene
uplift rate of 0.7 mm per year over the length of the lagoon, then most of the inter-tidal sites would
be uplifted a minimum of 1.0 m, resulting in them originally being covered by more than 2 m of
water at high tide without considering potentially higher water levels in the run-down from the
mid-Holocene high stand. Assuming a linear draw down from a 2 m high stand at 5500 B.P., it
would contribute circa 0.75 m at 2000 B.P. This would have the sites constructed in waist high or
considerably deeper water. Similarly if we lower many of these sites and their surrounds by 1.0
m, they lose their association with current shorelines, streams and springs; in many cases leaving
them a considerable distance from the shore. It seems most probable then that the situation of
these sites, including relationship to sea level, has not altered much if at all since the time of
occupation. However this does not rule out the possibility of an earlier episodic or higher rate of
uplift in Roviana some time after 6000 B.P. and before 2500-2600 B.P. producing the 3+m
palaeosea level marker recorded by Mann et al. (1998), which they attribute to the mid-Holocene
high stand. If a major uplift event occurred in Roviana Lagoon after 3200 B.P. and before 2600
B.P. then it could have destroyed the early Lapita record, however such an event should be
marked by uplifted reef flats and tidal notches below the 3+ m feature. This might correlate with
95
the 2750±60 b.p. (2990-2750 B.P. 95% CI) change from a marine to fresh water environment in
the pollen core from the Tamberamakoto swamp, although that is most likely associated with
falling mid-Holocene sea levels. Similarly if a significant uplift was followed at some later date
by 0.5 m of subsidence then the event could be masked in the lowest solution notch such as the
wide basal notch seen at Araroso
CONCLUSION
There is a considerable variety of coastline types in the Western Solomon Islands, from steep
exposed coasts with villages often perched up on coastal cliffs (e.g., Rannonga, Rendova) to
lagoons fringed by mangroves, coastal swamps and makatea barrier islands (e.g., Roviana,
Marovo, north Vella Lavella). As we have discussed, this variation reflects considerable variation
in geotectonic processes and rates throughout the region. The earthquake of Easter 2007 demon-
strated that both sudden and significant uplift and subsidence could occur within the region over
comparatively short distances. This suggests even more variability across the region than indicated
by previous research and demonstrates that models based on gradual geotectonic movement may
be too simplistic. Our study in Roviana supports the argument that there has been limited uplift in
that lagoon over at least the last 2000-2500 years and that shoreline profiles are consistent with a
mid-Holocene high stand forming a palaeoshoreline 2-3 m above Mean Lower Low Water (tidal
range of 1 m), followed by declining sea levels and the creation of the modern lagoon form.
Inter-tidal sites are found throughout the Western Solomons (north Vella Lavella,
Kolombangara, Ghizo, Roviana, north New Georgia and Marovo) in considerable numbers. One
of these sites has 6 dentate-stamped sherds, the majority none, and a few sites 2-3 sherds, certainly
none are like the earlier inter-tidal sites from Buka which have high percentages of dentate sherds
(50% of 188 decorated sherds at DJQ). These sites share a common incised and appliqué decora-
tive tradition which is very closely related to late/post-Lapita assemblages from the Bismarck
Archipelago, in particular the assemblages reported by Garling (2003, 2007) from Tanga, and the
New Ireland region (White and Murray-Wallace 1996). Current evidence suggests movement of
people making this pottery tradition and carrying small amounts of obsidian out of the New
Ireland region in the very late Lapita period into the Western Solomons (Sheppard and Walter
2006). There is no evidence of a region wide geotectonic process which could have destroyed an
earlier Lapita site record.
ACKNOWLEDGMENTS
We thank the Marsden Fund of the Royal Society of New Zealand which has funded much of our
Solomon Island research and that of our students over the last 11 years. We also thank our many
colleagues in the Solomons: Lawrence Foana‘ota (National Museum), Kenneth Roga (Western
Province), Gaudrey Kama (Roviana) and the many people of Roviana, Rannonga, Ghizo and
Vella Lavella who have assisted in our research, often taking considerable interest in the location
of inter-tidal sites. Unfortunately many of these same people have suffered considerable loss and
on-going trauma as a result of the recent earthquake and tsunami. For these people the history of
geotectonic movement has taken on new meaning, and we hope we can make some small contri-
bution to that study.
96
REFERENCES
Allen, J. and C. Gosden, 1996. Spheres of interaction and integration: Modelling the culture history of
the Bismarck Archipelago. In Janet Davidson, Geoffrey Irwin, Foss Leach, Andrew Pawley and
Dorothy Brown (eds), Oceanic Culture History: Essays in Honour of Roger Green. Dunedin: New
Zealand Journal of Archaeology Special Publication, pp. 183-97.
Anderson, A.J., 2006. Islands of exile: Ideological motivation in maritime migration. The Journal of Island
and Coastal Archaeology, 1: 33-47.
Anthony, D.W., 1990. Migration in archaeology: The baby and the bathwater. American Anthropologist, 92:
895-914.
Bedford, S., 2007. Crucial first steps into Remote Oceania. In S. Chiu and C. Sand (eds), From Southeast
Asia to the Pacific: Archaeological Perspectives on the Austronesian Expansion and the Lapita
Cultural Complex. Taipei: Centre for Archaeological Research, Academica Sinica, pp. 185-205.
Burley, D.V. and S.P. Connaughton, 2007. First Lapita settlement and its chronology in Vava‘u, Kingdom
of Tonga. Radiocarbon, 49: 131-37.
Cowley, S., P. Mann, M.F. Coffin and T.H. Shipley, 2004. Oligocene to recent tectonic history of the
Central Solomon intra-arc basin as determined from marine seismic reflection data and compilation
of onland geology. Tectonophysics, 389 (3-4): 267-307.
Darwin, C., 1842. The Structure and Distribution of Coral Reefs. London: Smith Elder.
Dickinson, W.R., 2001. Paleoshoreline record of relative Holocene sea levels on Pacific Islands.
Dickinson, W.R., 2007a. Upolu (Samoa): Perspective on island subsidence from volcano loading. Journal
of Island and Coastal Archaeology, 2: 236-38.
Dickinson, W.R., 2007b. Holocene paleoshoreline and paleoenvironmental history of Palau: Implications
for human settlement. Journal of Island and Coastal Archaeology, 2: 175-96.
Dickinson, W.R. and D.V. Burley, 2007. Geoarchaeology of Tonga: Geotectonic and geomorphic controls.
Geoarchaeology, 22 (2): 229-59.
Dickinson, W.R., D.V. Burley and R. Shutler Jr, 1999. Holocene paleoshoreline record in Tonga:
Geomorphic features and archaeological implications. Journal of Coastal Research, 15 (3): 682-700.
Dickinson, W.R. and R.C. Green, 1998. Geoarchaeological context of Holocene subsidence at the Ferry
Berth Lapita Site, Mulifanua, Upolu, Samoa. Geoarchaeology, 13 (3): 239-63.
Felgate, M.W., 2007. Leap-frogging or limping? Recent evidence from the Lapita littoral fringe,
Explorations: Lapita and Western Pacific Settlement. Terra Australis 26. Canberra: ANU E Press,
pp. 123-40.
97
Friedlaender, J.S., F.R. Friedlaender, F.A. Reed, K.K. Kidd, J.R. Kidd, G.K. Chambers, R.A. Lea, J-H. Loo,
G. Koki, J.A. Hodgson, D. A. Merriwether and J.L. Weber, 2008. The genetic structure of Pacific
Islanders. PLoS Genetics, 4 (1): e19. doi: 10.1371/journal.pgen.0040019.
Garling, S., 2007. Post-Lapita Evolutions or Revolutions? Interaction and Exchange in Island Melanesia: The
View from the Tanga Islands. Unpublished Ph.D. thesis, Australian National University, Canberra.
Gosden, C. and J. Webb, 1994. The creation of a Papua New Guinean landscape: Archaeological and
geomorphological evidence. Journal of Field Archaeology, 21: 29-51.
Green, R.C., 1979. Lapita. In J.D. Jennings (ed.), The Prehistory of Polynesia. Cambridge, Mass.: Harvard
Green, R.C., 1991. Near and Remote Oceania: Disestablishing “Melanesia” in culture history. In A. Pawley
(ed.), Man and a Half: Essays in Pacific Anthropology and Ethnobiology in Honour of Ralph
Bulmer. Auckland: The Polynesian Society, pp. 491-502.
Green, R.C., M.D. Jones and P.J. Sheppard, 2008. The reconstructed environments and absolute dating of
SE-SZ-8 Lapita site on Nendö, Santa Cruz, Solomon Islands. Archaeology in Oceania. 43: 9-61.
Green, R.C. and A. Pawley, 1999. Early Oceanic architectural forms and settlement patterns: Linguistic,
archaeological and ethnological perspectives. In R. Blench and M. Spriggs (eds), Archaeology and
Language III: Artefacts, Languages and Texts. London: Routledge, pp. 31-89.
Grimes, S., 2003. A History of Environmental Change and Human Impact Since 3750 B.P. in the New
Australia, Perth.
Kirch, P.V., 1987. Lapita and Oceanic cultural origins: Excavations in the Mussau Islands, Bismarck
Kirch, P.V. (ed.), 2001. Lapita and its Transformations in Near Oceania: Archaeological Investigations in
the Mussau Islands, Papua New Guinea, 1985-1988. Contributions of the University of California
Archaeological Research Facility, No. 59. Berkeley: University of California.
Kraus, B. S., 1945. Preliminary report on the discovery of surface sherds on Mono Island, Treasury Group,
Solomon Islands. American Antiquity, 11: 102-4.
Linsley, B.K., 1996. Oxygen-isotope record of sea level and climate variations in the Sulu Sea over the past
150,000 years. Nature, 380 [6571]: 234-37.
Mann, P., F.W. Taylor, M.B. Lagoe, A. Quarles and G. Burr, 1998. Accelerating late Quaternary uplift of the
New Georgia Island Group (Solomon Island Arc) in response to subduction of the recently active
Woodlark spreading center and Coleman Seamount. Tectonophysics, 295 (3-4): 259-306.
Mitrovica, J.X. and W.R. Peltier, 1991. On postglacial geoid subsidence over the equatorial oceans. Journal
of Geophysical Research-Solid Earth, 96 (B12): 20053-71.
Nunn, P.D., 2005. Reconstructing tropical paleoshorelines using archaeological data: Examples from the
Fiji Archipelago, southwest Pacific. Journal of Coastal Research, 21 [Special Issue 42]: 15-25.
Nunn, P.D., T. Ishimura, W.R. Dickinson, K. Katayama, F. Thomas, R. Kumar, S. Matararaba, J. Davidson
and T.H. Worthy, 2007. The Lapita occupation at Naitabale, Moturiki Island, Central Fiji. Asian
Perspectives, 46: 96-132.
Ogan, E., 1970. Nasioi pottery-making. Journal of the Polynesian Society, 79: 86-90.
98
Parsonson, G.S., 1972. The settlement of Oceania: An examination of Accidental Voyage Theory. In
J. Golson (ed.), Polynesian Navigation: A Symposium on Andrew Sharp’s Theory of Accidental
Voyages. Wellington: A.H. & A.W. Reed, pp. 11-63.
Reeve, R., 1989. Recent work on the prehistory of the Western Solomons, Melanesia. Bulletin of the
Indo-Pacific Prehistory Association, 9: 46-67.
Archaeology in Oceania, 28 (3): 121-37.
Sheppard, P.J., M. Felgate, K. Roga, J. Keopo and R. Walter, 1999. A ceramic sequence from Roviana
Lagoon (New Georgia, Solomon Island). In J-C. Galipaud and I. Lilley (eds), The Pacific from
5000-2000 BP: Colonisation and Transformations. Paris: Editions de IRD, pp. 313-22.
Specht, J., 1972. The pottery industry of Buka Island, T.P.N.G. Archaeology and Physical Anthropology in
Oceania, 7 (3): 125-44.
Specht, J., 1981. Obsidian sources at Talasea, West New Britain, Papua New Guinea. Journal of the
Specht, J., 1991. Kreslo: A Lapita pottery site in southwest New Britain, Papua New Guinea. In J. Allen
and C. Gosden (eds), Report of the Lapita Homeland Project. Occasional Papers in Prehistory
20. Canberra: Department of Prehistory, Research School of Pacific Studies, Australian National
Specht, J., 2003. Watom Island and Lapita: Observations on the Reber-Rakival localities. In C. Sand (ed.),
Pacific Archaeology: Assessments and Prospects. Le Cahiers de l’Archéologie en Nouvelle-Calédo-
nie 15. Nouméa: Service de Musées et du Patrimoine de Nouvelle-Calédonie, pp. 121-34.
Stoddart, D.R., 1969. Geomorphology of the Marovo Elevated Barrier Reef, New Georgia. Philosophical
Transactions of the Royal Society of London. Series B, Biological Sciences, 255, (800): 383-402.
Taylor, F.W., R.W. Briggs, C. Frohlich, A. Brown, M. Hornbauch, A.K. Papabatu, A.J. Meltzner and D.
Billy, in press. Rupture across arc segmen and plate boundaries in the 1 April 2007 Solomons
earthquake. Nature Geosciences.
Taylor, F.W., P. Mann, M.G. Bevis, R.L. Edwards, H. Cheng, K.B. Cutler, S.C. Gray, G.S. Burr, J.W. Beck,
D.A. Phillips, G. Cabioch and J. Recy, 2005. Rapid forearc uplift and subsidence caused by impinging
bathymetric features: Examples from the New Hebrides and Solomon arcs. Tectonics, 24 (6): TC6005.
Terrell, J.E., 2002. Tropical agroforestry, coastal lagoons, and Holocene prehistory in Greater Near
Oceania. In S. Yoshida and P.J. Matthews (eds), Vegeculture in Eastern Asia and Oceania. Osaka:
Japan Centre for Area Studies, National Museum of Ethnology, pp. 191-219.
Walter, R. and P.J. Sheppard, 1997. Excavation of Poroporo Cave, New Georgia, Solomon Islands.
Archaeology in New Zealand, 40 (1): 50-62.
Wheeler, G.C., 1928. On some pottery from Alu, Bougainville Strait Solomon Islands. Man, 28: 38-40.
White, J.P., C. Coroneos, V.E. Neall, W. Boyd and R. Torrence, 2002. FEA site, Boduna Island: Further
investigations. In S. Bedford, C. Sand and D. Burley (eds), Fifty Years in the Field: Essays
in Honour and Celebration of Richard Shutler Jr’s Archaeological Career. New Zealand
Archaeological Association Monograph 25. Auckland, pp. 101-7.
White, J.P. and J.E. Downie, 1980. Excavations at Lesu, New Ireland. Asian Perspectives, 23: 193-220.
White, J.P. and M. Harris, 1997. Changing sources: Early Lapita period obsidian in the Bismarck
Archipelago. Archaeology in Oceania, 32: 97-107.
99
White, J.P. and C.V. Murray-Wallace, 1996. Site ENX (Fissoa) and the incised and applied pottery tradition
in New Ireland, Papua New Guinea. Man and Culture in Oceania, 12: 31-46.
Zhilão, J., 2001. Radiocarbon evidence for maritime pioneer colonization at the origins of farming in west
Mediterranean Europe. Proceedings of the National Academy of Sciences of the United States of
America, 98 (24):14180-85.
100
Chapter 5
ASSESSING THE ANOMALOUS ROLE OF CERAMICS IN LATE

LAPITA INTERACTION: A View from Kolombangara, Western
Solomon Islands
A.M. Findlater, G.R. Summerhayes, W.R. Dickinson and I.A. Scales
INTRODUCTION
Kolombangara, located in the New Georgia Group, Western Solomon Islands, is important for
understanding models for prehistoric interaction between Near and Remote Oceania. The area of
the Western Solomons is unique. With an absence of early Lapita and presence of only late Lapita
and post-Lapita pottery (Felgate 2001, Summerhayes and Scales 2005), this region is significant
in constructing models of Lapita settlement and occupation for the Western Pacific. It has been
argued that Lapita peoples spread throughout the Western Pacific from the Bismarck Archipelago
to Samoa, bypassing the major islands of the Solomon Islands. During this early movement of
peoples, models focusing on high mobility and or interactions with a homeland community have
been used (Anderson 2001, Kirch 1988, Summerhayes 2000a). Recently, an intensive analysis of
the production and distribution of pottery has been undertaken on long Lapita sequences from the
Bismarck Archipelago, to demonstrate the changing modes of mobility and interaction during the
Lapita occupation in the Bismarck Archipelago and the Western Pacific (Summerhayes 2000b,
2003, 2004, Summerhayes and Allen 2007). It was argued that later Lapita communities were
more sedentary with pot assemblages having a more standardised mode of production. Such
studies are, however, limited and restricted to a few sites in the Bismarck Archipelago.
Analyses on pottery production are especially important as they are one of the few material
culture items that do survive archaeologically and have the potential to be identified to a source
(Weisler 1997: 10). Archaeologists have the opportunity for determining (1) whether production
was limited to a few centres or produced at a large number of locations, (2) if production was
indigenous or exotic to geologically distinct island groups, and (3) the geographical and temporal
scale of ceramic transfer operating. The demonstration of these aspects of ceramic transfer can be
used to infer the temporal and spatial nature of Lapita interaction and colonisation and more
specifically commercial trade, ceremonial exchange and mobility.
The present study is a unique opportunity to apply these techniques to assemblages from
Kolombangara with the aim of understanding the nature of Lapita occupation in the Western
Solomons. Previous archaeological research on Kolombangara consisted mainly of site surveys
which recognised mostly late prehistoric sites and early contact sites (refer to Keopo 1981, Kirch
2000, Miller 1979, Miller and Roe 1982, Yen 1976). However none of these sites recorded were
Lapita or inter-tidal pottery-bearing sites. Scales later identified nine localities with ceramic sites
on Kolombangara, of which two ceramic sites have affinities with Lapita ware (see Summerhayes
and Scales 2005: 14). The present study is on late and post-Lapita pottery found and collected by
Scales at Poitete and Tanahuka, from Kolombangara (Figure 5.1) (Summerhayes and Scales
2005). In the absence of early Lapita, the pottery from these two sites provides a chance to closely
assess the nature of late Lapita interaction in the region by examining pot production and transfer
through techniques of chemical characterisation.
The primary aim of this paper is to firstly examine the nature of ceramic production and
movement in detail by chemical characterisation using the electron microprobe and petrographic
101
Figure 5.1 Map of the Solomon Islands showing the location of places mentioned in text.
analysis. The results are used to assess the nature of production and exchange and how they fit
into broader regional production patterns. Thus the data from Kolombangara is contextualised
within the greater regional social system (after Earle 1982: 11), to allow the assessment of regional
models of mobility and interaction. This is contrasted with late Lapita and the Bismarck
Archipelago to the northwest (Kirch 1990, 1991, Summerhayes and Allen 2007: 2), and the Reef
and Santa Cruz sequences to the southeast (Sheppard 1993) of the same period from which conser-
vative ceramic production, regionalisation and the maintenance of short-distance exchange has
been inferred. Therefore the ceramics from Kolombangara have the potential to reassess the nature
of Lapita occupation, settlement and interaction within the greater Solomon Island region.
The results from this study provide useful comparisons with work undertaken in nearby
Roviana Lagoon where petrographic analysis of late Lapita and post-Lapita pottery by Felgate
and Dickinson (2001: 119) raised a number of issues that need addressing. On the basis of the
presence of a quartz-calcite temper, they suggested trans-Solomon sea interaction during the late
Lapita period to explain this temper at Roviana Lagoon. Testing the spatio-temporal distribution
and frequency of this anomalous temper through further geomorphological and archaeological
survey in the Western Solomons in order to understand origins of this temper was recommended
(Dickinson 2000, Felgate 2001, Felgate and Dickinson 2001).
METHODS
Chemical characterisation using the electron microprobe and petrographic analysis are both
highly useful analytical techniques used to identify production patterns from ceramics. For the
application of these techniques, representative samples of pottery from Poitete and Tanahuka
were chosen using the initial form and fabric analysis, as described in detail by Summerhayes
and Scales (2005).
102
Assessing the Anomalous Role of Ceramics in Late Lapita Interaction
Chemical Analysis (Clay Matrix and Mineral Inclusions)

An electron microprobe with an energy dispersive spectrometer (EDS) was used to chemically
characterise the ceramics sampled. The advantages of using this analytical technique has been
documented extensively, especially in regards to understanding production centres through
separate analysis of the ceramic clay matrix and mineral inclusions (Summerhayes 1997, 2000a:
37). Mineral inclusions need to be accounted for separately because their manual addition to clay
as temper can affect the analysis of clays. Kolombangara ceramic samples from Tanahuka (n=10)
and Poitete (n=10) were prepared in briquettes and impregnated with epoxy (araldite) resin under
vacuum. Analysis was undertaken on a JEOL JXA-8600 electron microprobe using an energy
dispersive spectrometer (EDS) attachment at the Geology Department, University of Otago.
Acceleration voltage and beam current were 15 KeV and 20nA respectively. We measured Na,
Mg, Al, Si, P, S, Cl, K, Ca, Ti, V, Cr, Mn, Fe and Ni as oxides. Backscattered-electron images
were taken of the pottery samples at 40 times magnification and the spots of analysis were
recorded on these images. Separate data was therefore obtained for the clay matrix and mineral
inclusions which could be discriminated.
In order to make archaeological sense of the clay matrix data and to understand chemical
paste compositional reference units (CPCRUs) (after Bishop, Rands and Holley 1982,
Summerhayes 2000a: 38-41), a series of multivariate techniques were performed on this data once
culled, normalised and averaged according to each sample. Principal Component Analysis (PCA)
was used to account for the maximum amount of variation in the data while still reducing it to
dimensions manageable for plotting and visualisation. The object scores from this were then used
in hierarchical clustering analysis (HCA) to produce dendrograms, which were compared with the
PCA plots. Several clustering techniques (Wards, Average Linkage between groups and Average
Linkage within groups) were used to check for common clusters to insure that the outcome was
valid (after Summerhayes 2000a). If clusters are present, hierarchical clustering is most useful for
identifying discrete groupings within larger continuous clusters, such as those observed in PCA
plots (for details on methodology refer to Summerhayes 2000a: 39-41).
The analysis of mineral inclusions via electron microprobe was intended to be qualitative
representing the range of inclusions present. The data obtained for the mineral inclusions via
electron microprobe were compared and synthesised with the data on temper types obtained via
petrographic analysis resulting in the identification of main fabric types.
Petrographic Analysis (Mineral Inclusions)

In addition to electron microprobe analysis, samples were petrographically analysed using thin
sections. Counts were made of individual grains or mineral types which were used to define main
temper types. This type of analysis has been recognised as the most effective way to characterise
ceramic temper types (Dickinson 2006, Dickinson and Shutler 2000). Dickinson characterised the
temper of six Kolombangara sherds via standard petrographic analysis techniques (see Dickinson
2007a). This provided both supplementary and complementary data to that of the microprobe for
the mineral inclusions. The data derived from petrographic analysis was also particularly useful
for comparative purposes with the Roviana Lagoon petrographic information.
103
RESULTS
Mineral Inclusions
Table 5.1 presents the results of characterisation of mineral inclusions using the electron micro-
probe. In these sherds there was a noticeable difference in the size between the largest particles
in the clay matrix and the finest mineral inclusions, thus signalling that the mineral inclusions
were manually added temper. Four main fabric types were identified through the combination of
results from petrographic analysis and the electron microprobe (see Table 5.2). Some of the initial
interpretations made by Summerhayes and Scales (2005) can now be revised. Only Poitete was
thought to have hornblende rich sherds but this can now be dismissed as both sites contain
examples. Three fabric groups were identified for the ceramics from each site, with the quartz-
calcite group present in Tanahuka ceramics only, and the feldspathic one in Poitete ceramics only.
The relationship of fabric with vessel form is unclear from this data, with jars and pots composed
of a range of different types.
Pyroxenic: Table 5.2 shows that half of the ceramic samples examined contain a pyroxene rich
placer sand of well sorted and sub-rounded to sub-angular grains (see Figure 5.2a). Microprobe
analysis identified orthopyroxene in four samples (with two sherds having both orthopyroxene
and clinopyroxene). There is little or no olivine or hornblende observed in these ceramics.
Petrographic analysis noted that the tempers were not identical and probably derived from
different coastal collecting sites and from basaltic bedrock. This hypothesis may be tested by
looking at the number of production centres as derived from the results of the analysis of the
ceramic matrix.
Hornblendic: Samples from both Poitete and Tanahuka have a hornblende rich volcanic placer
sand of well sorted and sub-rounded to sub-angular grains (see Figure 5.2b). This group contrasts
noticeably with the pyroxene rich group above as apart from hornblende dominating the ferro-
magnesium minerals, it also contains significantly less pyroxene and plagioclase grains.
Petrographic analysis noted that sparse volcanic rock fragments were dominantly felsic, indicat-
ing derivation from andesitic as opposed to basaltic bedrock.
Feldspathic hybrid: The feldspathic rich hybrid group is represented by three ceramic samples
from Poitete only and they contain unplacered volcanic terrigenous sand and calcareous grains of
reef detritus (see Figure 5.2c). Petrographic analysis showed that of the non-calcareous grains
69% were plagioclase, with some hornblende, clinopyroxene and iron oxide grains.
Quartzose hybrid: Only ceramic samples from Tanahuka (40%) contain the moderately sorted
quartz-calcite fabric (see Figure 5.2d). From petrographic analysis this sand is thought to be
hybrid as there is a bimodal distribution of grain sizes. Calcareous grains from reef detritus are
rounded and the quartz grains are sub-rounded reflecting their greater resistance to abrasion.
Counts of terrigenous grains in a stained thin section were made by Dickinson to check the ratios
of grain types for comparison with the Roviana Lagoon quartz-calcite temper later (see Dickinson
2007b). These sherds contain both plagioclase and alkali feldspars (the alkali feldspar being
identified as a microcline microperthite by the electron microprobe).
104
Table 5.1 Electron microprobe results characterising mineral inclusions for Poitete and Tanahuka (Sample
numbers correspond to those in Figure 3, T = Tanahuka, P = Poitete).
105
continued next page
106
Table 5.2 Electron microprobe results for mineral inclusions summarised to show the
percentage of samples from Poitete and Tanahuka for each fabric type identified.
Figure 5.2 a-d. Backscatter images to visualise each of the four main mineral inclusion types.
a. Pyroxenic. b. Hornblendic. c. Feldspathic Hybrid. d. Quartzose Hybrid.
107
(a)
(b)
Figure 5.3 a-b. Score plots of PCA and CPCRU comparison. a. Dendrogram from HCA showing CPCRUs.
b. PCA plot components 1 and 2 showing the relationship between CPCRUs, fabrics and sites.
108
Clay Matrix
The results of chemical analysis from the electron microprobe on the clay matrix, for each site are
presented together here. Six CPCRUs were defined from PCA and HCA. Figure 5.3a-b shows the
PCA plots and dendrogram with clusters. Applying the three different clustering methods
produced the same groupings and CPCRUs, reinforcing the archaeological significance of these
groups, and so only the dendrogram resulting from Wards method is presented here. CPCRUs can
be seen to separate out on the first, second, and third components where Mg loads heavily on the
first component and Ca shows the highest correlation to the second and third components. Tables
3, 4, and 5 present the relationship of CPCRUs to site, mineral inclusions and vessel form respec-
tively and illustrate the results presented here.
Table 5.3 Relationship between CPCRUs and sites.
Table 5.4 Relationship between CPCRUs and fabric types.
Table 5.5 Relationship between CPCRUs and vessel form.
109
CPCRU 1: This is the only chemical grouping that contains ceramic samples from both Poitete and
Tanahuka. Most are from Tanahuka, except for one sample from Poitete. This latter sherd has a
hornblendic fabric and is one of only two ceramics identified confidently to vessel form as a pot.
The Tanahuka ceramics have no exclusive fabric or form, with pyroxenic, hornblendic and quartzose
examples present, and are either from jars or pots or are non-attributable to a specific vessel form.
CPCRU 2: This CPCRU is represented by four samples, all from Poitete. All three of the samples
found to have a feldspathic hybrid fabric are found in this chemical grouping. These vessels are
all jars. The other sample is non-attributable to vessel form and is pyroxenic based. Upon closer
examination this group is not as coherent as the others, and PCA plots of components 2 and 3 and
1 and 3 (not shown here) indicate that the pyroxenic sample appears to separate out and lie closely
with CPCRU 3 which also contains pyroxenic ceramics.
CPCRU 3: Again this CPCRU is represented only by samples from Poitete. It contains an almost
restricted range of fabrics, with the majority of pyroxenic tempers bar one, found in the samples
from this site. This CPCRU has comparatively a wide distribution which may reflect the origins
of the pyroxenic temper from different coastal collecting sites. One sample is the only hornblen-
dic ceramic to be identified in the Poitete collection (except for the one from CPCRU 1 mentioned
above). Jars and pots are both represented by this CPCRU.
CPCRU 4: CPCRU 4 is represented by four samples from Tanahuka only. It is similar to CPCRU
3 as it is made up from mostly pyroxenic ceramics. The other fabric represented here, also by one
sample, is a quartzose sample. The relationship to vessel form is unclear, with samples either being
from jars or pots or simply non-attributable. This CPCRU contains the one sherd identified as
being from an incurving bowl, which is thought to be a vessel form rare in Lapita assemblages and
more reminiscent of later Melanesian ceramic assemblages (Summerhayes and Scales 2005: 17).
CPCRU 5: This CPCRU represents one sample from Tanahuka that is an outlier from all of the
rest. This sample is pyroxenic and is non-attributable to a particular vessel form.
CPCRU 6: This CPCRU has most in common with CPCRU 5, but only in terms of remoteness.
It represents another single sample from Tanahuka that is an outlier from all of the rest. This
sample has a quartzose fabric and is either from a jar or a pot.
Summary
Approximately six CPCRUs were identified from the electron microprobe results. Ceramic samples
from the two sites generally appear to separate out, except for CPCRU 1 which contains samples
from both sites. Even with the small sample size, there appears to be a clear division between the
pyroxenic samples from each site. Some CPCRUs have more than one fabric and the fabrics are
found in different CPCRUs, except for the Feldspathic Hybrid which appears to be only present in
CPCRU 2 found only in Poitete. The quartz-calcite fabric is found in three of the CPCRUs with
Tanahuka only sherds. Tanahuka and Poitete both share CPCRU 1 with the hornblende rich temper
(see Tables 3-5). The relationship of chemical groupings with vessel form is unclear.
DISCUSSION
Discussion will focus on three areas. Firstly, the results are interpreted in terms of understanding
the geological origins of the ceramic raw materials and thus the geographical scale of interaction.
Next, the nature of production centres is inferred from the CPCRUs, followed lastly by a compar-
ison of these results primarily with the Roviana Lagoon ceramics.
110
Geological Origins
Because islands or island groups have a restricted geology, sands, especially terrigenous ones, are
important for understanding the origins of ceramic raw materials. Ceramic fabrics, whether as
manually added or natural temper in sherds, can be interpreted as either indigenous or exotic to
an island or island group depending on their compatibility with an island’s geological make-up
(Dickinson 2006: 1-4). The ceramics from Kolombangara exhibit fabrics that reflect both indig-
enous and exotic origins (see Table 5.6).
The pyroxenic fabric most likely originates from local Kolombangara beach sands as it is
consistent with local geology. The island of Kolombangara itself is geologically part of the New
Georgia Group and almost entirely quaternary volcanic, with stratified ash deposits, basaltic
flows and pyroclastic sediments with one small upthrusted block of sedimentary bedrock in the
southwest. The most common rock on the island is olivine basalt with basaltic andesites, with rare
exposures of hornblende andesite and clinopyroxene as a common phenocryst resulting from the
basaltic bedrock (Dunkley 1986: 16). The absence and presence in small quantities of olivine in
the Kolombangara ceramics also points to Kolombangara origins. On Kolombangara the
non-picritic basalt is composed of sparse olivine phenocrysts which are easily susceptible to
tropical weathering. Basalts rich in olivine are found elsewhere in the New Georgia Group, for
example in southeast New Georgia (Dunkley 1986: 13), and so the presence of olivine is thought
to signal wider New Georgia origins. Orthopyroxene rich sands are thought to be restricted to
Simbo and Nggatokai where hypersthene andesites are mainly restricted (Dickinson 2000: 2,
Dunkley 1986: 14). Although orthopyroxene was identified by the electron microprobe in four or
the pyroxenic samples, it is assumed at this stage that the quantity is low as petrographic analysis
did not detect any. However even though the possibility that these samples came from further a
field in the New Georgia Group cannot be ruled out, Kolombangara origins are the most probable.
The prevalence of the pyroxenic type is explained by its local origin.
The dominance of hornblende in the hornblendic fabric points to geological origins
somewhere in the wider New Georgia Group. Because exposures of hornblende andesite are
common on Vella Lavella and rare on Kolombangara, this points to Vella Lavella as the likely
origin (Dunkley 1986: 13). However Kolombangara cannot be ruled out entirely. Hornblende-
bearing rocks are by volume much less important in the New Georgia Group than olivine basalt,
basaltic andesites and picritic basalts (Dunkley 1986: 13), and so this probably accounts for the
frequency of this type.
Table 5.6 Fabric as indicative of the geographical scale of interaction.
111
The feldspathic hybrid samples are more difficult to pinpoint to an origin. These ceramics
may be from the wider New Georgia Group, if not the wider Solomon Island region, due to the
plagioclase rich nature of the sand which may be derived from plagioclase rich basalts in these
wider regions (Dickinson 2007a: 3). Of note is that feldsparphyric basalts are found to post-date
the volcanic cones on the islands of New Georgia and Vangunu (Dunkley 1986: 22).
Like the Roviana Lagoon ceramics, Tanahuka displays evidence of a quartzose hybrid temper
which is thought to be exotic. It has an unknown origin (Dickinson 2006: 115, Felgate and
Dickinson 2001), but the presence of both plagioclase feldspar and K-feldspar grains suggests
possible origins from granite rock and the epidote could be from alteration of a plutonic complex.
Because the plutonic complexes in the Solomon Islands, the Bismarcks and Vanuatu are considered
to be mafic in character, they are rejected as possible origins for this type due to the idea that these
complexes would not provide enough K-feldspar or quartz consistent with what is present in the
samples. Ceramic transfer on a larger geographical scale is argued for this type, which is definitely
not from the New Georgia Group, perhaps signalling continental origins (Dickinson 2006: 115).
Nature of Production Centres

The data presented above demonstrates that a single potting industry or centre does not make all
of the pots from Tanahuka and Poitete, or even from one of the sites. Production is complex (see
Figure 5.4), with six sources of clay of unknown proximity identified here and many combina-
tions of clays and fabrics. Although the exact origins are uncertain, interpretations can be made
based on mineral identifications. The results suggest that finished pots or raw materials are
moving between communities at different geographical scales, both indigenous and exotic.
If particular tempers are thought to derive from local sources, such as the pyroxene-rich
tempers, it could be argued that production was local. But if additional exotic tempers are found
within the same clay then either the pots would have been made beyond the sites or the raw
materials brought in. It can be suggested that except for one sample most clays are local as they
are pyroxenic based, but there is the possibility of these samples originating from elsewhere in
the island group. It is highly probable that at least two different potting industries were operating
on Kolombangara: one at Tanahuka and one at Poitete. These centres made use of local pyroxenic
rich sand as temper, signalling local production using one clay and temper. Because of the
variation (as noted by petrographic analysis and PCA plots), it is also likely that a number of
collecting points or centres may have been in operation along the Kolombangara coast. The clay
forming CPCRU 5 with the local pyroxenic fabric is likely to be such an example of a separate
collecting point or alternatively ceramic transfer into Tanahuka. As all of the feldspathic ceramics
are found in one CPCRU, this signals evidence of a production centre or collecting point making
use of a single clay associated with one temper.
It is suggested that Tanahuka and Poitete also shared a production centre as signalled by
CPCRU 1, perhaps with closer ties to Tanahuka as indicated by the fabrics of the remainder of
ceramics in CPCRU 1 and the fact that no quartzose hybrid based samples were found in any of
the Poitete ceramics. This indicates possible ceramic transfer from Tanahuka to Poitete.
Fabrics possibly indicate that interaction was also occurring further afield in the region of the
New Georgia Group and perhaps the wider Solomon Islands. Mostly it appears that there was
ceramic transfer into Tanahuka and Poitete. However, the outlying sample forming CPCRU 6
suggests something else. This sample clearly indicates a separate clay source from all of the other
samples and additionally has exotic temper. This is suggestive of the movement of finished ceramic
vessels associated with migration and the movement of people, as opposed to transfer or exchange.
112
Figure 5.4 Diagram illustrating the nature of ceramic production centres, transfer and movement.
To summarise, most clays used at the two sites, Tanahuka and Poitete, are local or at least
regional, signalling a number of production centres. There is also evidence of ceramic transfer
from Tanahuka to Poitete and movement of raw materials between other localities on the island.
There are examples of conservative practice using only one clay and one temper type. There are
also combinations of clays with several temper types and exotic tempers which suggest a high
degree of mobility and experimentation. One ceramic fabric provides evidence of a foreign or
intrusive addition of finished vessels associated with the movement of people.
Comparison and Contextualisation

A broader picture of late Lapita interaction is available through the comparison and contextualisa-
tion of the Kolombangara ceramics, especially in terms of their relationship with the ceramics
from Roviana Lagoon. Dickinson (2007a, 2007b) has made a number of observations through
petrographic analysis that deserve summarising here:
113
1. No orthopyroxene has yet been detected in the Roviana Lagoon samples, unlike in the
Kolombangara samples. Although its presence in low quantities may mean this is insignificant
but worth noting for further studies.
2. The pyribole index of one Kolombangara hornblendic sample is comparable to the index of
feldspathic tempers in Roviana Lagoon sherds also thought to derive from Vella Lavella,
strengthening the interpretation that this is the most likely source for this Kolombangara sherd.
3. Dickinson looked at one feldspathic hybrid sherd from Kolombangara. This temper was
observed to not resemble three sherds recovered previously on Kolombangara (of non-Lapita
origin) thought to derive from Choiseul as looked at by Dickinson previously.
4. Most significantly Dickinson looked at one quartzose hybrid sherd and reported that this was
indistinguishable compositionally and texturally from comparable quartz-calcite tempers in
exotic sherds previously found in Roviana Lagoon (Felgate and Dickinson 2001).
Comparisons with other assemblages have improved the chances of establishing source
areas and have begun to address the question of the spatial and temporal distribution of the
anomalous quartz-calcite temper. As Felgate and Dickinson (2001: 119) note “appearance of the
temper for a brief period in low frequency over a modest area of the Solomon Islands, with no
pattern of local centre of origin where the temper occurs in high frequency, would suggest exotic
origin”. This temper occurs at a higher rate in the Tanahuka samples (four out of ten samples),
compared with the lower frequency of the temper from Roviana Lagoon as reported in the figures
for the Hoghoi site (60 out of 861) (Felgate and Dickinson 2001: 112). Even though the percent-
age is higher from Tanahuka there is a marked difference in sample sizes. Chemical characterisa-
tion of the clay matrix of quartz-calcite tempered ceramics from Roviana Lagoon may prove an
important avenue of enquiry to understand if production centres might be comparable between
sites in the Western Solomons.
The results suggest that the Kolombangara ceramics generally fit within the local regional
pattern, as observed in Roviana Lagoon. The majority of ceramics appear to be produced locally
as argued for other late Lapita sites, such as those in the Bismarck Archipelago. Here late produc-
tion is thought to be standardised and conservative with simple combinations of clays with one
fabric type. Yet, the results do not completely conform to a pattern of late Lapita production as
defined in the Bismarck Archipelago. The results from Kolombangara lend themselves to models
of mobility for the makers of this pottery, as seen in the complexity in the combinations of clays
and fabrics. The quartz-calcite temper provides further evidence for a high degree of mobility and
the possibility of trans-Solomon sea interaction during late Lapita times. It must also be remem-
bered that evidence of long distance transfer is not new to the Solomon Islands, where the long
distance movement of obsidian in during the early Lapita period has been documented for some
time (Green and Bird 1989). Thus this kind of pattern with multiple production centres is thought
to be reminiscent of early Lapita and may perhaps represent a later Lapita colonisation signature
(Summerhayes and Allen 2007).
The Solomon Islands has a substantial but anomalous archaeological record of ceramics from
which to explore Lapita interaction (q.v. Sheppard and Walter 2006: 48, 56-60). Early Lapita terres-
trial sites are present on Buka and on the Reef and Santa Cruz Islands (Green 1978, Green and
Creswell 1976, Wickler 2001). In the Western Province of the Solomons, there is an absence of
early Lapita sites, and a presence of inter-tidal late Lapita and post-Lapita ceramics from a number
of localities. This is unlike the central Solomons where these later inter-tidal sites are absent
(Sheppard and Walter 2006: 57). A number of explanations, including both cultural and natural
114
processes and the nature of archaeological practice, have been offered to explain the absence of
Lapita sites in parts of the Solomons (see Felgate 2001, Sheppard and Walter 2006, Summerhayes
and Scales 2005). Clearly there are combinations of factors affecting what is known as the archae-
ological record of Lapita in this region which can only be addressed through further archaeological
and geomorphological research. However, on the basis of the ceramic evidence presented above, a
model of Lapita “leap frogging” the Solomons (Sheppard and Walter 2006) during its early phase,
and a late Lapita colonisation of this area for the first time, is favoured here.
CONCLUSION
Summerhayes and Scales (2005: 18) originally addressed late Lapita ceramics from Kolombangara
in terms of their importance for filling the anomalous gaps in Solomon Island prehistory. The
evidence from chemical characterisation has not only greatly extended our understanding of late
Lapita interaction in the region, but it has increased our knowledge of the distribution of the
anomalous quartz-calcite temper as originally reported by Felgate and Dickinson (2001). This
temper is not restricted to Roviana Lagoon, and is now known from at least one other site in the
Western Province – Tanahuka. It can now be hypothesised with more certainty that the temper’s
distribution may extend into the wider Western Province and that it is indeed exotic due to its
low frequency.
The results presented here offer new insights into the nature of late Lapita settlement and
interaction in the Solomon Island region. The sites on Kolombangara show evidence of local
ceramic production, but also exhibit evidence of exotic ceramic transfer. Interaction may have
been occurring at many scales between Tanahuka and Poitete and within the wider island
landscape. Production appears complex, with no single production centre for the ceramics on the
island. The results appear to contrast with current models of late Lapita social processes based on
ceramic data and may represent a colonisation signature. In Kolombangara the relationship
between production and vessel form is still unclear. Further archaeological research, such as
chemical characterisation, will be able to refine these observations and eventually lead to an
explanation of the anomalous nature of ceramics in the region.
ACKNOWLEDGMENTS
The authors would like to thank the 2007 Lapita Antecedents and Successors Conference organis-
ers, and Peter Sheppard the session chair. We acknowledge the support of the University of Otago
Humanities Division Postgraduate Conference Funding, University of Otago Students Association
Clubs and Societies Grant, and the University of Otago Anthropology Society. Thanks for the
assistance of Lorraine Patterson, Microprobe Officer, Geology Department, and Les O’Neill,
Illustrations Unit, Department of Anthropology, University of Otago. The support of the people
of Kolombangara and Lawrence Foanaota, Solomon Islands National Museum is gratefully
appreciated. Thanks also to an anonymous reviewer for comments on this paper.
REFERENCES
Anderson, A.J., 2001. Mobility models of Lapita migration. In G.R. Clark, A.J. Anderson and T. Vunidilo
(eds), The Archaeology of Lapita Dispersal in Oceania: Papers from the Fourth Lapita Conference,
June 2000, Canberra, Australia. Terra Australis 17. Canberra: Pandanus Books, pp. 15-23.
Bishop, R.L, R. Rands and G. Holley, 1982. Ceramic compositional analysis in archaeological perspective.
Advances in Archaeological Method and Theory 5: 275-320.
115
Dickinson, W.R., 2000. Petrographic Report WRD 196. Petrography of Sand Tempers in Prehistoric Sherds
from Roviana Lagoon, New Georgia, Solomon Islands. Unpublished report in possession of author.
Dickinson, W.R., 2006. Temper Sands in Prehistoric Oceanian Pottery: Geotectonics, Sedimentology,
Petrography, Provenance. Special Paper 406. Boulder: The Geological Society of America.
Dickinson, W.R., 2007a. Petrographic Report WRD 258. Petrography of Temper Sands in Prehistoric
Sherds from Kolombangara, New Georgia Group, Solomon Islands. Unpublished report in
possession of author.
Dickinson, W.R., 2007b. Petrographic Report WRD 260. Quartz-Calcite Tempers in Exotic Sherds from the
New Georgia Group. Unpublished report in possession of author.
Dickinson, W.R. and R. Shutler Jr, 2000. Implications of petrographic temper analysis for Oceanian
prehistory. Journal of World Prehistory, 14 (3): 203-66.
Dunkley, P.N., 1986. Geology of the New Georgia Group, Western Solomon Islands. British Geological
Survey Overseas Directorate Report MP/86/6.
Earle, T.K., 1982. Prehistoric economics and the archaeology of exchange. In J.E. Ericson and T.K. Earle
(eds), Contexts for Prehistoric Exchange. New York: Academic Press, pp. 1-12.
Felgate, M.W. and W.R. Dickinson, 2001. Late-Lapita and post Lapita pottery transfers: Evidence from
intertidal-zone sites of Roviana Lagoon, Western Province, Solomon Islands. In M. Jones and P.
Sheppard (eds), Australasian Connections and New Directions. Proceedings of the 7th Australasian
Archaeometry Conference. Research in Anthropology and Linguistics No.5. Auckland: Department
of Anthropology, University of Auckland, pp. 105-20.
Green, R.C., 1978. New Sites with Lapita Pottery and their Implications for an Understanding of the
Settlement of the Western Pacific. Working Papers in Anthropology, Archaeology, Linguistics and
Maori Studies 51. Auckland: Department of Anthropology, University of Auckland.
Green, R.C. and J.R. Bird, 1989. Fergusson Island obsidian from the D’Entrecasteaux Group in a Lapita
site of the Reef Santa Cruz Group. New Zealand Journal of Archaeology, 11: 87-99.
Survey. Royal Society of New Zealand Bulletin 11. Wellington.
Keopo, J., 1981. Report of the Kolombangara and North New Georgia Archaeological Site Survey.
Solomon Islands National Museum, Honiara. Manuscript.
Kirch, P.V., 1988. Long-distance exchange and island colonization: The Lapita case. Norwegian
Archaeological Review, 21: 103-17.
Kirch, P.V., 1990. Specialization and exchange in the Lapita complex of Oceania (1600-500 B.C.). Asian
Perspectives, 29 (2): 117-33.
Kirch, P.V., 1991. Prehistoric exchange in Western Melanesia. Annual Review of Anthropology, 20: 141-65.
European Contact. Berkeley and Los Angeles, California: University of California Press.
Miller, D., 1979. Report of the National Sites Survey: Summary Report. Honiara: Solomon Islands National
Museum.
Miller, D. and D. Roe, 1982. The Solomon Islands National Sites Survey: The First Phase. Indo-Pacific
Prehistory Association Bulletin, 3: 46-51.
116
Summerhayes, G.R., 1997. Losing your temper: The effect of mineral inclusions on pottery analyses.
Summerhayes, G.R., 2000a. Lapita Interaction. Terra Australis 15. Canberra: Department of Archaeology
and Natural History and Centre for Archaeological Research, Australian National University.
Summerhayes, G.R., 2000b. Far Western, Western and Eastern Lapita – A re-evaluation. Asian
Summerhayes, G.R., 2003. Modelling differences between Lapita obsidian and pottery distribution patterns
in the Bismarck Archipelago. In C. Sand (ed.), Pacific Archaeology: Assessments and Prospects. Le
Cahiers de l’Archéologie en Nouvelle-Calédonie 15. Nouméa: Service des Musées et du Patrimoine,
pp. 135-45.
Summerhayes, G.R., 2004. The nature of prehistoric obsidian importation to Anir and the development of a
3,000 year old regional picture of obsidian exchange within the Bismarck Archipelago, Papua New
Guinea. In V.J. Attenbrow and R. Fullagar (eds), Pacific Odyssey: Archaeology and Anthropology
in the Western Pacific: Papers in Honour of Jim Specht. Records of the Australian Museum,
Supplement 29. Sydney: The Australian Museum, pp. 245-56.
Summerhayes, G.R. and J. Allen, 2007. Lapita writ small? Revisiting the Austronesian colonisation of the
Papuan south coast. In S. Bedford, C. Sand and S.P. Connaughton (eds), Oceanic Explorations:
Lapita and Western Pacific Settlement. Terra Australis 26. Canberra: ANU E Press, pp. 97-122.
Weisler, M.I., 1997. Introduction. In M.I. Weisler (ed.), Prehistoric Long-Distance Interaction in
Oceania: An Interdisciplinary Approach. New Zealand Archaeological Association Monograph 21.
[Dunedin], pp. 7-18.
Yen, D.E., 1976. Agricultural systems and prehistory in the Solomon Islands. In R.C. Green and M.M.
Cresswell (eds), Southeast Solomon Islands Culture History: A Preliminary Survey. Bulletin 11.
117
Chapter 6
COMMUNITIES OF PRACTICE IN THE ARCHAEOLOGICAL

RECORD OF NEW GEORGIA, RENDOVA AND TETEPARE
Tim Thomas
INTRODUCTION
Since the mid-1990s, a series of archaeological projects in the Western Province of the Solomon
Islands has produced one of the most thorough post-Lapita cultural sequences we have for Near
Oceania (Felgate 2003, Nagaoka 1999, Sheppard, Walter and Nagaoka 2000, Thomas 2004,
Walter and Sheppard 2000, 2006). Perhaps this says as much about the patchiness of research into
this time period as it does about the success of the projects in question. But it is indicative of the
potential richness of this part of the record, particularly over the last 1000 years, that research in
the Western Province has been so productive. I would also argue that the results of the research
signal how important this period is for our understanding of wider themes in Melanesian prehis-
tory. In this chapter I present some initial results from my own ongoing work in New Georgia,
Rendova and Tetepare, articulating this with some of these wider themes.
It has long been recognised that the key to understanding the patterns of culture evident in
the Island Melanesian ethnographic record lies in the post-Lapita period (notwithstanding an
occupation history that stretches into the Pleistocene). This is often conceptualised as involving a
process of increasing diversification and regionalisation, from the comparative unity of early
Lapita to the enormous diversity of 19th century Melanesia. Drawing largely on data from coastal
Papua New Guinea, it has been hypothesised that diversification involved the breakdown of
region-wide exchange networks into smaller circuits of increasing specialisation and complexity
(Allen 1984), accompanied by localised cultural and socio-political transformations, and
fragmentation of dialect chains into the myriad languages of the “ethnographic present” (Kirch
2000: 163, Spriggs 1997: 154-60, 185-86). This model is one akin to speciation by network decay,
like a pond evaporating in the sun leaving behind puddles teeming with rapidly mutating life.
Given the patchiness of the post-Lapita record however, the general applicability of this model
still needs to be assessed region by region. Data like that produced for the Western Solomons is
useful in this regard, and can be profitably extended through more fine-grained tracing of cultural
change regionally, as well as through more work on the precise nature of local exchange networks,
their extent and temporality. This has been the primary aim of my research on the islands of
Rendova and Tetepare, and similar projects on the island of Vella Lavella conducted by Sheppard,
Walter and students.
As well as filling gaps in our knowledge, such studies may also enable assessment of the
network decay model’s central assumptions – e.g., that contact and exchange engenders similar-
ity, holding at bay “natural” divergence with isolation. The recent end of the Solomon Islands
record for example, indicates that some exchange networks embraced, perhaps even fostered,
diversity, being based on social forms that ethnographers have referred to as “relations which
separate” (Strathern 1988: 191). That is, exchange consisted of a shared set of understandings
(and objects) for eliciting difference (Thomas 2004), amongst multiple tribal and linguistic groups
(see below). The degree to which such fine-grained patterns played out over the long term is
currently unknown, but there is potential here for building more theoretically sophisticated
models of network formation and decay.
119
Figure 6.1 The New Georgia Group in the Western Solomon Islands.
If such work can lead us to question our assumptions, it also brings up the issue of how
applicable well-documented later forms of exchange are to earlier periods, or more generally, how
much we should allow the ethnographic record to colour our interpretations of the archaeological
record. Certainly key aspects of debates about the origins and nature of Lapita centre on the
applicability of ethnographically known characteristics of Melanesian lifeways to those of the
past (Bellwood 1996: 883, 2005: 10, Terrell, Hunt and Gosden 1997). Opinions differ about the
extent to which the apparent homogeneity of Lapita as a “culture” is without ethnographic parallel
or can be challenged by ethnographic analogy. This problem is partly due to differences in data
resolution – micro-level classification of social formations characterises interpretations of high
resolution ethnographic data, while only broad classifications are encouraged by the lower resolu-
tion archaeological data. Thus our tendency to see Lapita as relatively homogenous in comparison
to the ethnographic record is to some (unknown) extent biased on both sides by resolution. This
will always be an issue if our only reference points are two widely separated periods in time with
only partial accounts of their connection. It can only be resolved by developing linked-up culture
histories, where the sequence of change from the past to the present is better known. This reiter-
ates the importance of studies into the recent past: somewhat paradoxically, one key to under-
standing Lapita diversity lies in post-Lapita culture histories.
To a greater or lesser extent the development of full sequences for Island Melanesia may
lessen the need for analogical reasoning in the form of drawing direct parallels between the
archaeological and ethnographic records – something that is inherently problematic (Gosden
1992: 22). But it should not limit our drawing theoretical or conceptual lessons from one context
120
Communities of Practice in the Archaeological Record of New Georgia, Rendova and Tetepare
and using them to think about problems in another. In the remainder of this chapter I do just that
by interrogating the concept “community of culture” through the lens of the recent archaeological
record of the Western Solomons. My goal is not to say anything in particular about contexts other
than the one in question, but rather to draw attention to and question the ways we tend to explain
certain kinds of phenomena.
COMMUNITIES OF CULTURE
To my knowledge the phrase “community of culture” was first introduced into Pacific archaeol-
ogy by Jack Golson (1961: 176) in his well known recognition that the distribution of Lapita
pottery required a rethink of the ethnographic Polynesia / Melanesia boundary as it was then
understood. Since then the phrase has cropped up when there has been a need to specify other
situations where suites of cultural traits co-occur over large regions. The debated “incised and
applied relief tradition” of post-Lapita ceramics, for example, has sometimes been referred to as
a watered down community of culture, although opinions continue to differ on the matter
(Bedford 2006: 175, Lilley 2006: 18, Spriggs 2004, Wahome 1999). The phrase has also been
deployed to characterise the ethnographic pattern of artefact usage amongst the varied tribal and
linguistic groups of the North Coast of New Guinea (Terrell et al. 1997, Welsch, Terrell and
Nadolski 1992). In each of these cases there is similarity in meaning – the phrase implies shared
cultural participation as indexed by the presence of identical or similar objects amongst predefined
units (archaeological sites, language groups). Consequently each case employs a trait based,
classificatory approach to the definition of culture.
The New Guinea coast case, however, has been controversial in that the phrase “community
of culture” was mobilised to draw a purposeful parallel with Lapita, claiming in the process that
the distribution of Lapita pottery need not necessarily imply homogeneity in language or biology
as postulated in some conventional accounts (Roberts, Moore and Romney 1995, Terrell et al.
1997, Welsch et al. 1992: 591-92). As noted above, some of the room for disagreement here is
created by differences in data resolution, but this is compounded, I think, by a problem with
ambiguities in the “community of culture” concept, and the underlying trait based approach.
Culture is a polysemic term (Terrell, this volume) but at root it entails a relational, complex
whole deriving from shared involvement and history (Kroeber and Kluckhohn 1952). Some have
argued that “culture” implies homogeneity and boundedness however negotiated, and encourages
us to think in hypostatised terms (i.e., what begins as an analytical classification becomes thought
of as a real entity) and is thus of dubious utility (see Brumann 1999 and responses for a good
summary of the various “for” and “against” arguments). Indeed this is the basis for Green’s
(1992:12) objection to Lapita being termed “a culture”, preferring instead to specify a polythetic
“cultural complex” on the grounds of assemblage heterogeneity. But a further problem is that
when a set of immobilised variables (shared traits) is used to define a community of culture,
exactly how or why those traits are shared can be elided, such that the explanation may be the
same as the observation: groups share a culture. The underlying ambiguity is that when the term
culture is used it is not immediately obvious whether the usage is merely classificatory (i.e.,
shared object use defines a group and who belongs in it), or whether it is held to have explana-
tory power (i.e., objects are shared because people have a common culture). If both are intended
then we risk tautology.
This latter problem is not new and was addressed in 1950s and 1960s by a shift from talking
about cultural behaviours and traits to cultural meanings and shared understandings, since these
were felt to have more explanatory weight (e.g., Geertz 1973). Another very early attempt at a
121
solution is evident in Radcliffe-Brown’s (1952) turn towards an anthropological theory based on

the concepts of process, structure and function. Expressly moving away from his earlier usage of
the term “culture” (1952: 14), Radcliffe-Brown felt the processes and interactions of social life to
be the appropriate units of anthropological analysis. Since these are naturally always in a state of
flux or change, explanation was to focus on instances of observed temporal stability – i.e., struc-
tured practices of recurring duration. Thus, those classificatory and somewhat passive “cultural
traits” were now problematised as “social practices”, a focal point for explanations in the form of
determining their role (“function”) in societal continuity. As is well known, this shift was at the root
of the split between British social anthropology and American cultural anthropology (although they
were later to converge, particularly with Radcliffe-Brown’s move to Chicago). I mention Radcliffe-
Brown here not because I think we should return to early 20th century social theory, but because a
similar shift in emphasis might be useful for clarifying the kinds of phenomena we seek to under-
stand in the Island Melanesian archaeological record – implying a “social archaeology” perhaps.
In the archaeology of Lapita there are a few notable examples of such an approach, and these
might clarify what I mean. Gosden (1992) for example, attempts to explain the temporal and
spatial stability and coordinated change evident in Lapita material culture by deploying the
concept “dynamic traditionalism”. Rather than approach the problem from the direction of
inherited culture, he approaches it from the direction of ongoing practice, interpreting the material
evidence as indicative of a practical strategy for coping with diversity and change as mobile
populations colonised new land. The emphasis is on spatio-temporal process, practical solutions
and their material expression. Lilley (1999) employs a similar strategy, arguing that the stability
of Lapita pottery worked as a materialised ideology, an attempt to unify people at a time of
diaspora and creolisation. Both examples can be thought of as social archaeology in the sense that
they attempt to comprehend stable phenomena with reference to their “structural function”, (if I
can put it in such old-fashioned terms), rather than as data for externally defining the unity or
disunity of groups. Insofar as identity is an issue it is revealed to be the product of a process of
practical action, rather than a static “thing”. In fact the general impression is of a social field, which,
in particular times and places, coalesces around shared involvement in certain projects: communi-
ties of practice rather than culture. I want to now expand on this theme by looking at the late period
archaeological record of the Western Province, where, in common with much of Island Melanesia,
shared items of material culture are found amongst disparate tribal and linguistic groupings.
NEW GEORGIA
Exactly when humans first colonised New Georgia, and the central Solomons in general, is unknown
(see Walter and Sheppard, this volume). On New Georgia itself, however, the archaeological
sequence begins at about 2800 B.P. with concentrated scatters of pottery sherds deposited on the
inter-tidal sands of lagoon fringes, probably marking the presence of stilt house villages. About 20
such sites have been recorded in Roviana Lagoon (Felgate 2003), and are also present in north New
Georgia (Felgate 2007), on Kolobangara (Summerhayes and Scales 2005), Vella Lavella and Ghizo.
Stylistically, the pottery from these sites conforms to regional trends in the late- and post-Lapita
periods. There is something of a lacuna in the record between 2200 B.P. and 1000 B.P. possibly
caused by a decline in the intensity of coastal settlement or usage of ceramics. Palaeoenvironmental
evidence indicates that human impact on forest ecosystems intensified greatly from 2500-1000 B.P.
(Grimes 2003), suggesting that this period saw a generalised movement inland.
Walter and Sheppard (2006) have usefully ordered the more recent archaeological record of
New Georgia into periods within an encompassing “Munda Tradition” dating from 700-100 B.P.
By the early limit of this time range populations had established extensive inland settlements,
122
with irrigated taro pondfields in the fresh water river basins lining valley floors on the larger
islands (Bayliss-Smith, Hviding and Whitmore 2003, Tedder 1976). Reflecting the intensity of
this agricultural activity, massive depositions of clay, gravel and silt began to accumulate on
several mainland river floodplains from 1100-500 B.P. (Grimes 2003: Table 5.1). Associated with
these inland settlements a tradition of megalithic shrine construction has been documented
(Chikamori 1967, Miller 1979, Nagaoka 1999) and is related to practices that persisted until the
early 20th century, albeit with various significant transformations. It is this portion of the record
that I wish to focus on.
The Munda Tradition is divided into two periods by Walter and Sheppard (2006: 148-52):
the Bao Period (700-400 B.P.) and the Roviana Period (400-100 B.P.). The nomenclature here
recognises that most of the available data for this part of the sequence comes from the southern
coast of New Georgia – centred on the modern town of Munda and Roviana Lagoon. Much of
this data was produced in the years 1996-2000 by the New Georgia Archaeological Survey project
led by Sheppard and Walter, and has been published in numerous articles (Sheppard et al. 2000,
Walter and Sheppard 2000). Here I give a brief account of the sequence as it is currently under-
stood, but focus particularly on the implications of the record for understanding social practice.
This will serve as a baseline for my discussion of more recent, ongoing work on Rendova and
Tetepare in which I will outline the argument for the development of a regional community of
practice at this time.
The Bao Period is named for a site in the interior of New Georgia held to be a key origin
place for several currently defined tribal groupings (butubutu) in Roviana. At Bao a collection of
archaeological features has produced the earliest radiocarbon dates for the Munda Tradition. The
site is located in the highest part of Kazukuru tribal territory behind Munda and consists of a
series of large platforms descending a ridge. These structures are recognised locally as shrines
(hope), once associated with land fertility and cultivation. Each platform is made from earth and
rubble fill, faced with basalt slabs that sometimes exceed a metre in height. Most are elongate
rectangular structures and some are stepped, having small indented “cists” at the highest point.
The largest sit on long paved areas and are associated with an altar-like “table stone” oriented
down-slope, consisting of a massive slab suspended on top of smaller rounded boulders. Some
table stones have upper surfaces that are pitted from having been used as nut anvils or are concave
from grinding, and were perhaps used for making food-offerings. There is, however, an absence
of midden debris or ovens, and the sites are isolated from habitation structures. In addition, no
artefacts or human remains have been recovered (Sheppard et al. 2000).
These attributes are shared with other shrines located in interior New Georgia as well as a
few rare sites on the barrier islands of the lagoon. Together they form a typological class usefully
termed “slab-constructed shrines”. According to radiocarbon dates this class has a temporal extent
of 700-400 B.P. (Figure 6.2) thus defining the Bao Period. The date range compares favourably
with regional oral traditions. The shrines at Bao are held to have been made by ancestors of the
Kazukuru butubutu during particular mythical events, beginning with the birth of an apical
ancestor, Vakorige, out of the mountain (Hall 1964: 133), and culminating in migration to the
coast generations later. This latter development is said to have begun with intermarriage and
alliances forged with the neighbouring Tagosage and Roviana butubutu and a consequent mutual
extension of territory incorporating the coast, forming a Kazukuru-Roviana polity. Genealogically
the process is crystallised in the person of Ididubaŋara, a chief (baŋara) who is said to have
abandoned the last shrines of Bao to take up residence on the barrier reef island of Nusa Roviana
some 14 generations ago (Aswani 2000: 46-47). Oral histories of this island associate a shrine
there with the arrival of Ididubaŋara – the site (NGAS 79, Kogu) is a series of coral slab platforms
123
incorporating basalt columns and a “table stone” all imported from mainland New Georgia. The
site has been radiocarbon dated to the mid 14th century, at the terminal end of the “slab shrine”
period (Sheppard et al. 2000: 32-33). Several other coral slab shrines were constructed on other
barrier reef islands at this time, and possibly relate to the extension of Kazukuru-Roviana territory
throughout the lagoon – rendered in oral tradition as stories of Ididubaŋara and his successors
warring against resident coastal tribes, particularly the Koloi and Vuragare butubutu (Aswani
2000: 48). These coastal slab shrines were, however, the last of their kind (see Walter and
Sheppard 2006: 149-50).
From about 400 B.P. a different but related form of shrine construction developed in consort
with a suite of cultural shifts that persisted until the early 20th century, thus defining the Roviana
Period. The genealogy of shrines at Bao is extended on Nusa Roviana with another series of sacred
origin places proceeding down the only ridge on that island (Thomas, Sheppard and Walter 2001).
These sites are associated with the immediate descendants of Ididubaŋara, embedding Kazukuru-
Roviana in a new locale and consolidating authority on the coast. The dislocation is mirrored in
genealogy and associated oral history with a series of events that created a new focal point of
origin. Nine of Ididubaŋara’s descendants are said to have died while living near the summit of
Nusa Roviana, before magically transforming into a class of spirits called mateana. These were
violently potent beings that appeared riding rainbows and lightning bringing sickness and death,
and were associated with headhunting and its symbolic ally, bonito fishing (Hocart 1922: 268-69).
The bodies of the nine dead sank into the earth at the summit of Nusa Roviana leaving their
mateana spirits to haunt the skies. Their descendants went on to form the current tribes of the
lagoon in a series of movements, fissions and amalgamations with neighbouring coastal and inland
groups, establishing enclaves at Munda, Kalikoqu and Saikile. Today the chiefly lineages of these
areas claim an authority and potency derived from the nine mateana and focus on the place of their
deaths as a self-evident embodiment of authority (e.g., Parker 1994, Schneider 1998).
Each sacred locale, as at Bao, is marked by a shrine, but in contrast these consist of small
mounds of coral cobbles, often grouped together or within walled compounds. Earth ovens lined
with stone are sited in front of many shrines, traditionally used for burning food-offerings.
Importantly, the shrines frequently house collections of human skulls, exchange media such as
Tridacna shell rings and other artefacts, as well as pig mandibles and food waste. Furthermore,
the sites are situated within a fortified settlement complex, surrounded by defensive walls and
ditches, as well as house platforms and terraces stretching along much of the length of the ridge
(Thomas et al. 2001). Some of these features were in use by 1600 A.D. (Figure 6.2), and in form
are typical of the organisation of settlements established throughout the lagoon after this date,
with loose cobble shrines, domestic areas and defensive features in close association – creating
an often densely packed landscape of contiguous hamlet clusters (Walter and Sheppard 2000).
The archaeological landscape of Nusa Roviana is the densest in the region, marking a real
intensification of settlement. Along the coastal fringes a continuous blanket of midden material,
domestic terraces and house platforms, wharves, canoe houses, shrines and artefact caches covers
the land (Figure 6.3). This more recent archaeological record bears attributes recognisable from
the photographs and ethnographic descriptions made by early European visitors to New Georgia
(Hocart 1922, Somerville 1897), and indeed contains items of European origin – axes, guns,
ceramics, and other trade objects. These sites track the emergence of the Roviana chiefdom, and
their density and wealth are indicative of its success. The shrines particularly, relate to a period
that saw ancestor veneration and headhunting become increasingly vital to social reproduction
and the structure of chiefly authority (Aswani 2000, Sheppard et al. 2000).
124
NZA10855 830±60
NZA10856 789±70
WK7917 610±50
NZA6235 468±62
WK6155* 1060±45
WK7914* 1010±50
NZA9457 556±57
WK6760* 810±50
NZA10854 325±60
WK6156 300±45
WK6157* 720±50
WK7916* 710±50
WK6756* 680±50
WK6758 250±50
Figure 6.2 Radiocarbon dates documenting the “Munda Tradition” of New Georgia.
[Data from Sheppard et al. 2000].
The persistence of social agency at this time was dependent on a careful maintenance of
relations with the potent spirits (tomate) of dead ancestors, which could be induced to accompany
the living in important endeavours, making them mana or efficacious. Shrines, housing ancestral
skulls and shell valuables, were the focal point of offerings and communication with tomate,
standing as elements of a complex funerary practice that served to ensure the safe transition of
the soul of the dead to the afterlife, while assembling the potent remains, an embodied spirit,
embedded in the landscape (Walter, Thomas and Sheppard 2004). This in itself may have been
little different from previous times – certainly the shrines at Bao can be viewed as places of
ancestor worship or propitiation – but in terms of context, the change was great.
The Kazukuru shrines at Bao and elsewhere were apparently restricted in their use to rituals
of land clearance and garden fertility (Hall 1964). This is reflected in the position of many inland
shrines on high land above river systems associated with irrigated taro agriculture. In contrast, the
later shrines of the coast reveal an efflorescence of ritual practice, with each loosely cobbled
feature locally classified by gender and realm of efficacy. Different shrines were used in rituals
of weather control, fishing (for bonito, or turtle, or with a net), purification, childbirth, disease
(either causation or cure), gardening, pig hunting, and crucially, headhunting (Nagaoka 1999).
The rising importance of this latter category is articulated by the hilltop origin shrines of Nusa
Roviana, which, surrounded by walls and modified topography, were heavily defended and
125
126
Figure 6.3 Archaeological landscape of Nusa Roviana as documented by the New Georgia Archaeological Survey 1996-2000.
explicitly affiliated with war. The nine dead ancestors had, after all, taken the form of death-
dealing mateana spirits, and so their shrines were places where ancestral agency was called upon
prior to raids to provide invincibility for local warriors or to bring misfortune to enemies (Thomas
et al. 2001: 557, 569).
By the 19th century headhunting was ubiquitous throughout the New Georgia Group, with
Roviana something of a stronghold. Butubutu there came to coalesce around chiefly lineages
headed by men who had achieved fame for their ability to manage and conduct successful raids
against neighbouring islands. Success at taking heads was one of the ingredients of mana, the
state of being that promised perpetual efficacy for leader and butubutu, and part of a project of
constructing local utopias where “living well” (Dureau 2000: 86) meant ancestral spirits joined
their descendants in all endeavours: gardens would be bountiful, fish would be caught, enemies
would be vanquished, and the butubutu would prosper. This was not a simple thing, for success
necessitated careful management of a field of relationships with the living as well as the dead,
involving the elicitation of work and its compensation. To stage a single raid a leader might need
to sponsor the construction of war canoes by experts, secure the service of warriors, form
alliances, and so on – each of which in turn required participation in a reciprocal economy,
whereby feasts would be staged (and thus gardens prepared, and foodstuffs secured), shell ring
exchange media would be accumulated and distributed, and ancestral spirits propitiated at every
step. Headhunting was a practical net within which much of the butubutu was caught up and
necessarily implicated in its success or failure (Thomas 2004).
Accordingly headhunting was an integral part of refocusing power on the coast, in that it
played upon the affordances of that landscape in opposition to interior New Georgia. Hviding
(1996: 96-97) has noted how, in nearby Marovo, a distinction is maintained between butubutu that
are historically “of the coast” and those “of the bush”, with attendant practical differences in
lifestyle and social emphasis. Groups living on lagoon shorelines tend to stress cumulative patri-
filial ties to place as embodied in chiefly lineages – a “men-leadership-territory” complex, where
leadership centres on the predominantly male activities of inter-island exchange, fishing, long-
distance raiding and associated rituals. Inland groups, on the other hand, stress cumulative matri-
filiation and blood ties – a “women-blood-territory” complex, centred on female cultivation of
people (birth) and garden land (1996: 147-49). Hviding emphasises that these are not exclusive
systems, but differential biases have clearly emerged from engagements in dissimilar but
connected social landscapes. In Roviana the coast/bush opposition is not currently a matter of
identity because histories of movement, amalgamation and intermarriage allow members of most
butubutu to associate with both spheres. But, the archaeological and oral evidence indicates there
was a gradual accumulation of population and social gravity on the coast over the last 400 years,
with links between headhunting and leadership ensuring that the political economy of coastal
raiding took a dominant role. This is also evident in the extinction of the Kazukuru language and
its replacement with Roviana soon after the abandonment of the interior. The new context of the
Nusa Roviana shrines and their material assemblages suggests this process facilitated an increas-
ing prioritisation of travel, trading and raiding such that headhunting shifted a prior social
interface between inland and coastal groups to one between Roviana and other islands.
We can think of headhunting as an axis around which regional relationships were structured,
having at least two dimensions: allies and enemies. In terms of the first, the same fields of reliance
and reciprocation that pertained internally within the butubutu had a tendency to expand outwards,
encompassing others from neighbouring islands. Alliances were fostered resulting in intermar-
riage and familiarity as well as material co-dependence and participation. This was particularly
127
important in a region where the production capacity of tribal estates was unevenly distributed, and
particular groups specialised in certain products necessitating trade expeditions (Hocart [n.d.a]: 7).
Such trips were conducted along exchange pathways that redistributed specialist products
throughout a series of inter-butubutu partnerships between persons classified as baere (‘friends’).
So, for example, there were trips from Roviana to Vella Lavella to procure Canarium nuts in
exchange for locally made shell rings (Hocart [n.d.b]: 7), or to the Kusage region of north New
Georgia for wicker war-shields. Down-the-line exchange occurred with certain parties acting as
mediators: Roviana people gave Kusage shields and shell rings to Simbo baere in exchange for
pigs, or packaged nuts acquired from Vella Lavella, and so on (Hocart [n.d.a]: 7). Archaeologically
we see interactions through the presence of non-local items such as chert (from Isabel or Malaita),
stone adzes and plain pottery (possibly from Choiseul) in sites of the Roviana period.
Baere partnerships were established and framed by reciprocal exchanges of shell rings, and
entailed that members of each party were treated according to the moral obligations of siblingship
(Hocart 1931: 305, Thomas 2004: 286). This implied mutual respect and care as well as safe
harbour, to the extent that chiefs held privileges in the territories of their baere, with the ability
to stage feasts, or make requests for aid in war and other endeavours (Hocart [n.d.c]: 9). The flip
side of this was that any transgression of the moral bond (e.g., murder, adultery) resulted in the
annihilation of relations, and led to all-out conflict.
This brings us to the second dimension. If performing raids (and indeed other forms of
action) entailed alliance with those who were not necessarily kin, then it also entailed there being
enemy others entirely outside the bounds of sociality. Or to put it another way, the existence of
“friendships” was predicated on the fact that there were ‘enemies’ (kana) who were afforded little
or no social status at all. By the late 19th century most Roviana headhunting expeditions were
apparently conducted against groups in Choiseul and Santa Isabel, with whom peaceful interac-
tion was increasingly negligible, fighting having escalated beyond local conflicts. Aswani (2000:
62) has argued that people from these regions were systematically denied personhood, compared
metaphorically to pigs that were hunted, or fish that were caught. More importantly, headhunting
enacted an absence of enemy efficacy in that through defeat other parties lost the means by which
communication with life-giving ancestors was effected: the victors absconded with the heads of
the dead, preventing their enshrinement (Dureau 2000: 91). Haunted by the angry ghosts of their
slain kin, surviving victims were considered to be severely limited in the extent to which they
could participate as effective agents in regional social networks.
Like baere friendships however, this boundary of sociality was not total – just as friends
could become enemies via transgression, enemies could become like kin through contribution and
care. Cessation of conflict was frequently negotiated and evinced by ceremonial exchanges of
shell rings, resulting again in “friendly” relations, if not alliance. Perhaps more interesting though,
is the fact that captive enemies were routinely incorporated into butubutu. Known as pinausu,
captives were, according to Hocart’s informants, “really supposed to be dead” and were counted
amongst the heads taken during a raid (1931: 306). The implication is that such persons were
severed from obligatory ties to their homelands and rendered free to be incorporated anew in a
number of possible guises: although some were killed in sacrificial rites of inauguration, others
reportedly became ritual attendants at shrines or servants, and some married their captors.
Pinausu is the nominal form of pausu ‘to adopt’, and indeed captives were included in the termi-
nology of kinship as classificatory children and siblings. Like kin, captives engaged in work on
the land, producing and sharing food with their captors, and were thus engaged in familial
reciprocity and care. In recognition, their skulls might be placed at death in a lineage shrine.
128
Though their incorporation was certainly ambiguous, pinausu were seen to be crucial to social
reproduction, to the extent that in certain circumstances they were recruited specifically to
“replace the butubutu” of their captors, ensuring continuity in practical ties to land (see McDougall
2000). Today descendants of captives live on gifted land throughout New Georgia, albeit in
occasionally tenuous positions, testifying to the duration and extent of such practices.
DISCUSSION
The archaeological and ethnographic record of the Munda Tradition in New Georgia traces the
material emergence of a set of social practices that facilitated management and control of group
membership, attachment to place, and successful action in the world. To summarise: genealogies
describe a situation whereby an inland group (Kazukuru) allied and amalgamated with neighbour-
ing inland (Tagosaghe) and coastal (Roviana) groups to form a social network and territory that
excluded others (Vuragare, Koloi). Archaeological evidence suggests that by 400 B.P. the coastal
realm of this network had garnered more social weight, with population levels increasing on the
barrier islands of Roviana Lagoon. The architecture and material culture of this time emphasises
that political legitimacy was partially fostered through the dual “exchange” practices of trade and
headhunting, with group efficacy and membership increasingly defined with reference to coastal
travel and engagement with distant “others” rather than a more local coast/bush division. As noted
by Walter and Sheppard (2000: 315, 2006: 152) the structural contrast between defended ancestral
skull shrines linked to inland origins, and war-canoe houses where enemy heads captured from
overseas were displayed, epitomises the development of this new dynamic.
The account I have presented stresses the “structural function” (i.e., societal role) of
material practice from a Roviana perspective. But by the time of European arrival, we know that
most islands within the New Georgia Group were involved in similar practices. Ultimately the
result was an efflorescence of coastal activity, movement, trade, alliance, raiding, and mutual
participation in social networks regionally – all of which is reasonably well documented in
historical accounts of the period 1880-1910 (cf. Bennett 1987). A number of questions spring
from this. The first is whether this historic period pattern is the late outcome of the expansion of
the Roviana chiefdom, or whether the sequence documented in Roviana is the local expression
of a much wider regionally coordinated process. A secondary question relates to localised
variability, or the extent to which different groups shared a common understanding of the
meaning and outcome of the practices they engaged in. Certainly there is much linguistic
diversity today in the New Georgia archipelago with about 12 different languages spoken. Ten
of these are very closely related Austronesian languages and the two remaining are tentatively
classified as East Papuan (Dunn, Reesink and Terrill 2002). In addition, several languages may
have suffered extinction relatively recently – Kazukuru, of course, is one example (Dunn and
Ross 2007, Waterhouse and Ray 1931) but rumours of others persist in local traditions of
“original tribes” wiped out during periods of conflict or disease. Thus, the region may well have
been more linguistically diverse in the recent past than at present. In order for this to have
developed there must have been either a degree of isolation at some point, or a process of active
boundary definition through language markers.
The only way to address these issues is to conduct more work on neighbouring islands, devel-
oping parallel sequences to the Munda Tradition of southern New Georgia. Since the end of my
involvement in the New Georgia Archaeological Survey, I have been doing exactly this. For the
remainder of this chapter I review my initial findings on two islands located off the coast of Roviana:
Rendova and Tetepare. This work is ongoing, and so the results presented here are preliminary.
129
RENDOVA
Rendova is a volcanic island located only 4 km south of Roviana Lagoon’s barrier reef.
Geologically it consists of two parts – a large symmetrical volcanic cone to the north, and a second
eroded volcano to the south with a long uplifted peninsula attached. Like New Georgia, much of
the island is without permanent inland settlement. The northern coastal fringe, facing Roviana, is
occupied by the Ughele people whose language is very closely related to Roviana. There has been
no archaeological work carried in Ughele, but the people here are traditionally linked to the
Vuragare people of Roviana, and it is possible that they moved to Rendova during the period of
Kazukuru-Roviana expansion. The greater part of southern Rendova is home to scattered coastal
settlements of the speakers of Touo, a non-Austronesian language possibly belonging to an East
Papuan familial group (Dunn et al. 2002). It is here that my surveys have concentrated, particu-
larly within the Lokuru and Baniata regions (Figure 6.4) during 2004 and 2005.
Non-Austronesian languages are, of course, generally thought to descend from languages
spoken by Pleistocene era settlers of Island Melanesia, due to their comparative diversity and
uncertain familial relationships. But their presence in any given locale need not imply continuous
in situ occupation since that time. On Rendova the entire early portion of the record is currently
unknown. A single test excavation within a solution cave on an uplifted Pleistocene reef terrace
Figure 6.4 Rendova late period archaeological sites (RDV prefix removed for space).
130
inland from Busana Bay was conducted during July 1999 by members of the NGAS, but this
recovered only a single volcanic oven stone beneath two metres of degraded limestone clay, with
no datable material. Further surveys along the face of this terrace (the remains of a fringing reef
around the now uplifted volcanic cone) may yet turn up some evidence of very early occupation.
Much of the Holocene record too has yet to be found. The late- and post-Lapita era inter-
tidal sites of Roviana have no parallel in southern Rendova, at least within the survey area. This
is probably due to environmental conditions – this region is characterised by a long black-sand
coastline, which drops off steeply into deep water, and on the western weather coast there are surf
beaches. There is very little reef formation, with coral growth inhibited by fresh water runoff,
swift currents and water depth. In addition this portion of the island has been effected by very
high rates of episodic uplift during the Holocene (Mann et al. 1998) exposing a substrate of old
marine sediments, primarily water rolled volcanic rocks deposited by ocean currents and debris
flows from the island chain to the north and northeast, that are being rapidly eroded by water
runoff. Winding rivers cut down from the mountain range between Lokuru and Baniata, dissect-
ing the loose sediment, and this has created a series of very steep ridgelines and gully systems.
The land between the Fatoani and Rano Rivers on the Lokuru side, and the entire Baniata coast,
is especially steep and ridged. Consequently this landscape may not have been particularly attrac-
tive to settlement by people more accustomed to lagoon ecosystems.
In late prehistory however, this ridged landscape offered the distinct advantage of being
easily defensible. In fact the current range and ancestral lands of the Touo speakers are those of
classic refugia, focused almost exclusively on the steepest most difficult terrain. There are two
major villages on the eastern coast, Vanikuva/Bangopingo and Rano, with scattered settlement to
the north and south of these relating to very recent logging company activity. The entire southern
peninsula, known as Rava, has been devoid of people within living memory, as has the more
gently sloping land to the north. On the western coast, the villages of Ou/Au and Hopongo occupy
thin strips of flat land on the edge of the Baniata Mountains.
According to oral tradition, people moved to the coast at the behest of missionaries in the
1930s, abandoning their previous settlements behind the beach in the current garden lands. As in
New Georgia there is a tradition of ancestral movement from the interior of the island – in Lokuru
and Baniata people claim descent from tribes (rana) who lived at “Loulo” (allegedly the original
pronunciation of Lokuru), a now forgotten place somewhere high on the mountain. Genealogical
memory here, however, is comparatively short with most traditions going back only 6-7 genera-
tions. Interestingly this lack of time depth is mirrored by the archaeological record.
My surveys of late period archaeological sites concentrated on the Lokuru region, with a
single trip to Baniata documenting only a handful of coastal sites. Most features cluster around
the knife-edge ridges behind Rano and Vanikuva/Bangopingo (Figure 6.5), comprising defended
settlement complexes that probably date to no earlier than about 1700 A.D. on the basis of radio-
carbon samples (Figure 6.6 and Table 6.1). All tend to conform to a similar organisational pattern.
Below each ridgeline, on the lower sloping land near the rivers, the remains of old gardens sit
amongst newly planted crops. Occasional platforms of basalt cobbles are positioned at the edge
of river cut bluffs, comprising recent house foundations (e.g., RDV 14) and isolated places of first
interment (e.g., RDV 30, 34, 36), where the recently dead were left to decompose before removal
of the skull for enshrinement in a skull house.
Above these areas the ground slopes steeply upwards, and a first series of shrines is often
found at the head of each ridge overlooking the surrounding terrain, typically facing out to sea.
Examples of these include the shrine complexes at Borongozo (RDV 5, 6, 7), Taravilo (RDV 3
131
Figure 6.5 Late period archaeological sites, Lokuru, Rendova.
132
Figure 6.6 Radiocarbon dates from late period Rendova sites.
Table 6.1 Radiocarbon dates from Rendova.
– 3.8), and Utuovo (RDV 24 – 27). Each consists of a group of large skull shrines, sitting on ground
that has been levelled and terraced. Shrine form varies from simple piles of river rolled stones, to
earth platforms and small constructed cists. The building materials are basalt rocks brought up the
slopes from the river beds below, and slabs of coral. The shrines sometimes contain the remains of
human skulls, pig jawbones, deposits of shell rings and other artefacts, and most have a small earth
oven located nearby for the burning of offerings. Old nut cracking anvils are frequently found on
shrines, but it is not clear whether these were used opportunistically as building materials or
whether they were deposited for ritual purposes. Whatever the case the landscape of Lokuru is
scattered with these pitted stones and it is not uncommon to find them alongside paths, old gardens
and overgrown living spaces. Carved stone figures were found in several shrine locations (e.g.,
RDV 5, 4) and these were clearly the focus of offerings, with piles of artefacts and pig mandibles
commonly deposited at their base. An unusual artefact form was noted on many shrines – large
basalt cobbles pecked or ground into a U-shape. It is thought that these are the remains of worn out
stone bowls, perhaps used as anvils for pounding puddings or in the preparation of other foods.
133
Some shrines have pieces of branch coral or cobbles made of coral limestone deposited amongst
the basalt rocks. These are likely to have been brought up from the coast, or from areas of coral
reef uplift to the north and south of the survey area. Low rectangular paved platforms are commonly
located in the vicinity of the shrines, marking the position of house foundations.
The construction and contents of these sites are in most respects similar to those found in
Roviana period sites of New Georgia. The major difference is that the material used is invariably
basalt – obviously due to the different geological affordances of the landscape. Interestingly
however, the occasional presence of imported coral rocks and reef fragments is an inversion of the
pattern seen in Roviana – in each case non-local materials are incorporated into ritual structures.
The shell valuables present in the shrines and the stone figures are identical in style to those found
in Roviana during the same period (compare Figure 6.7 with Sheppard et al. 2000: Fig. 8). Tridacna
shell rings are traditionally a specialisation of Roviana Lagoon where abundant uplifted reef
provides access to sub-fossil clam shell, but people in Lokuru were clearly involved in this too,
since several sites contain manufacture debris and stone pump-drill counterweights. All of which is
to say that during this period Lokuru people were involved in similar practices of ritual and trade.
Figure 6.7 Carved stone figure (RDV5). Pig mandible “offerings” are deposited in the foreground.
134
Violent activity is also written into the landscape. The shrines located at the heads of ridges
typically mark the beginning of defended settlements which may also have served as places of
refuge and ritual importance. On the flat land beyond the first shrines, patches of wilding crops (e.g.,
banana and nut trees) and scattered house platforms are typically found. In places where the ridge
rises again or narrows, terracing occurs (e.g., RDV 4.3) making the way virtually impassable, and
ditches are sometimes cut across the width of the ridge (RDV 7.3, 4.1, 4.2), creating easily defended
barriers. These areas make the most of an extremely steep and eroded landscape, with features
placed to impede access to the maximum extent. Finally, at the highest point of each ridge and the
end of the modified areas, another group of shrines usually occurs (RDV 8, 9, 4, 19-21). These sites
may have had particular ritual importance since they are the least accessible, most defended areas.
Perhaps the most spectacular and effective defensive site is located at a fork in the river
below a ridge leading up to the shrine complex and settlement extending from Taravilo (RDV 3)
to Beoroko (RDV 4). Known as Hasiri (RDV 15), this site consists of a 40m long ditch and bank
surmounted by a 1 m high basalt cobble wall, situated above an access path from the river bed.
The slope here approaches 90 degrees, and the wall would serve as ample defence against anyone
coming up from the deep river gorge below.
These ridgeline sites find their closest parallel in the Nusa Roviana hillfort with its similar
arrangement of defended ancestral shrines and settlement areas (Thomas et al. 2001). Although
the oral traditions associated with the Lokuru sites are nowhere near as detailed as those of Nusa
Roviana, it is clear from local interpretations that they served much the same social function,
remembered as the haunts of powerful warriors, witness to great deeds of heroism, and remaining
useful markers of ancestral attachment to land. However, on the basis of current dating, these sites
appear in the record more recently than those of Roviana. There are a number of possible explana-
tions for this, including the possibility that (a) earlier sites are to be found in as yet unsurveyed
regions, (b) date samples relate to last use rather than first construction, or (c) practices diffused
from a Roviana centre and were adopted later here. The explanation I consider most likely
however, is that the sites relate to the recent arrival of Lokuru people in this part of the landscape
– probably in response to increased headhunting activity taking place at the time. The shallowness
of genealogical history, and the depopulation of less defensible land to the north and south
(including all of Tetepare, see below), is indicative of this. In addition there is tentative evidence
of earlier occupation in the northern interior.
Three sites (RDV40, 47, 49) were recorded to the north of Lokuru in the high land of
Haforai, and all are of a typologically similar construction, related to but different from the sites
mentioned above. Each consists of an elongate earthen mound faced with flat basalt stones, and
supporting one or more stone platforms on top. The mounds vary between 5-10 m wide and 12-70
m long. Each has a very large basalt upright slab positioned at the coastward or down slope end
of the mound. Although associated with stone paving, small mounds and peripheral platforms, all
of these sites are somewhat more isolated than those mentioned above. A test excavation at
RDV49 recovered datable burnt Canarium nutshell at the foot of the upright, associated with
large angular chunks of basalt and a large chert flake (9 x 11 x 4 cm) exhibiting use wear and edge
damage. On geological grounds the latter must have been imported, probably originating from the
vicinity of Isabel or Malaita. Radiocarbon dating returned an age of 431±30 B.P. placing site use
at around 1450 A.D. (Figure 6.6) – near the end of the Bao Period of the Munda Tradition.
Given their similarity of construction and location we can consider all three sites as evidence
of an earlier period of occupation in southern Rendova. The sites are remarkable for their similar-
ity to the Bao Period shrines of New Georgia – they are similarly faced with basalt slabs, have
135
stepped platforms, are isolated from other settlements, lack shell rings and other recent artefact
forms, and have a general elongate, rectangular shape. There are some obvious differences – the
Rendova sites lack “table stones”, and the New Georgia sites tend to have several frontal standing
stones rather than just one. But the parallels are enough to indicate some similarity of practice and
age. Interestingly oral traditions relate that Haforai people migrated to coastal Ozireqo tribal land
in Lokuru, after forging an alliance by exchanging the only wealth they had – a charm called
torozoqi for ensuring “food bearing mana”. The similarity to the Kazukuru narrative is striking,
both implying a shift in emphasis from garden land and its fertility to coastal activities of trade
and war. In contrast to Roviana however, Lokuru people remain strongly matrilineal today,
suggesting that the role of male entrepreneurial and warrior based leadership may not have
achieved the salience it did in Roviana.
In any case, site typology coupled with the presence of the imported chert flake at site RDV
49, indicates that regional interaction networks have an antiquity extending beyond the recent
record in southern Rendova. Thus it is possible that the changes we see happening in the New
Georgia sequence are part of a regional phenomenon. Roviana, particularly, may have achieved
dominance in these networks, and the depopulated landscape surrounding that centre is indicative
of the role of headhunting in its success. This brings us to Tetepare.
TETEPARE
Widely regarded as the largest unoccupied island in the South Pacific, Tetepare is currently
managed by a local charitable organisation called the Tetepare Descendants Association as an
eco-stay, with various active conservation programmes supported by funding from the WWF and
others. Descendents trace ancestry to the island through links to individuals (typically women)
who are recorded in genealogies as having left through marriage or captivity probably in the late
1700s or early 1800s. Captives are held to have married men in Nusa Roviana, Kolobangara,
Ranonga and locations in Marovo, again testifying to the extent of regional movement in later
periods. Local hypotheses for the total depopulation of the island refer to the impact of headhunt-
ing and sickness induced by magic. In one such story a Roviana chief called Odikana (of Saikile,
a key butubutu that split from the Nusa Roviana based Kazukuru-Roviana polity) attacked a settle-
ment at the eastern end of Tetepare but later died as a result of injuries suffered there. In response,
the Saikile people burnt a spell called loduvuvusu and released it over Tetepare, killing everyone.
My archaeological surveys began in 2007, and have thus far recorded numerous late period
village settlements and shrine complexes along the northern coast of the island, supplemented by
a few sites located on the eastern and western tips (Figure 6.8). Like Rendova much of the
weather coast to the south is extremely steep and rugged, with rough seas, steep beaches and high
rainfall. Coral development is limited to fringing reef, but there is a small lagoon and good reef
passage at the western end of the island. The north coast is sheltered, with deep bays and
numerous river mouth sand bars, but no lagoon or reef, although there are some very high coral
bluffs towards the eastern end.
As on Rendova most of the village settlements occupy positions on ridges between river
valleys, or on top of coral bluffs. However, the landscape is less steep and the ridgelines less
precipitous. Many sites are close to the sea and are not positioned in defensible locations, although
several have low walls surrounding them, and it is possible these served some defensive purpose.
Oral traditions suggest that the western end of the island has greater antiquity of settlement, with
people moving progressively eastward in response to raids from Roviana. The veracity of this
narrative is yet to be established, but the most easterly sites do have more wall structures and tend
136
Figure 6.8 Tetepare late period archaeological sites (TET prefix removed).
137
to be located in less accessible places (e.g., TET 57-63, TET 52-56). The most easterly site of all,
on Kupa point (TET 72-76), consists of a walled enclosure sectioning off a high coral bluff,
surrounded by big seas with no canoe landings nearby. It is held to be the last place permanently
occupied on Tetepare. Numerous closely spaced platforms inside the walls are of unique construc-
tion and form, being faced with stacked slabs of flat coral (Figure 6.9), and the preservation
suggests that they may indeed be relatively recent. Although investigation and mapping of this
site is still underway, the general impression is of a densely packed, defended habitation,
positioned as far away from Roviana as possible.
Site preservation generally on Tetepare is very good in comparison to Lokuru where logging
activity and looting have damaged the record. Many features are structurally intact with clear
definition from the surrounding forest floor. Crisp rectangular platforms constructed of coral or
basalt cobbles predominate, and except in a few cases, distinctions between shrines and houses
are quite difficult to make on architectural grounds alone (i.e., without artefactual evidence). This
is mostly due to a relative scarcity of constructed skull houses, which in Lokuru and Roviana
Figure 6.9 Stone walling at “Hindi” (TET-72) Kupa Point, Tetepare.
138
clearly mark ancestral shrines. Ritual practice in Tetepare appears to have varied slightly, with
most sites which contain human skeletal material being elongate mounds of cobbles with indented
cists or depressions. A particularly common type consists of a basal platform with dual elongate
mounds placed in parallel on top – a form not documented elsewhere in the region. Human bones
are deposited within the cobble matrix of such sites and both cranial and other bones are present
– this is not typical in New Georgia or Rendova where only the skull is kept in shrines, the rest
of the body being left to decay in isolated locations. Several ritual sites are rather massive, either
by virtue of the sheer amount of rock used in construction (e.g., TET 12), or the size of the
boulders themselves (e.g., TET 40). Labour investment is comparatively high – shrines placed on
60 m high vertical coral bluffs have platforms constructed of massive water rolled boulders trans-
ported from rivers at sea level below (e.g., TET 57).
In terms of artefact contents and activity areas, the Tetepare sites are very similar to those
of Lokuru and Roviana, containing complete and broken shell rings of identical types, nut anvils
and hammer stones. Large boulders with numerous facets produced during the grinding and
polishing of shell valuables or stone tools are very common finds, occasionally associated with
cut Tridacna clam shell and other manufacturing waste. Some faceted stones are of flared cylin-
drical form, suggested by informants to be of a type used in the preparation of bark cloth fabric.
A single stone axe blade of lenticular cross section was recovered from site TET 4 and this is
likely to have been of local manufacture on geological grounds. A single fragment of shell ring
found at TET 29 is likely to have been imported from Choiseul on stylistic grounds. No items of
European origin have yet been recovered from these sites, and this is consistent with the
estimated date of island abandonment.
Radiocarbon dates for six sites span the period from 1500-1800 A.D., with the majority falling
within the 1700s (Figure 6.10 and Table 6.2). TET 44 and TET 8 are the earliest sites dated so far,
and are part of a settlement area surrounding the Hokata River. TET 44 consists of a collection of
four platforms aligned in parallel, with earth ovens at their centre. Behind these platforms, two
rectangular alignments of stone and scattered grinding stones and nut anvils suggest that the site
Figure 6.10 Radiocarbon dates from late period Tetepare sites.
139
Table 6.2 Radiocarbon dates from Tetepare.
represents a collection of domestic structures and house floors. TET 8 consists of two high basalt
cobble mounds marking the entrance to a rectangular area enclosed by a coral cobble wall. Human
skeletal fragments and shell rings are deposited within the mounds, and there are grinding stones
and anvils scattered about. A central earth oven is positioned within the wall. This site is similar to
ritual compounds recorded in Roviana and elsewhere, with a clear demarcation of space restricting
access (see Thomas et al. 2001). Both sites are part of a wider landscape containing shrine platforms
and domestic structures and these probably date to a similar period. They are contemporaneous with
the construction of the hillfort on Nusa Roviana, but unlike that site, each feature is within easy
reach of the coast, and they are mostly undefended. The only effort towards defence is provided by
two platforms (TET 48) situated either side of a gap in a low coral bluff below the settlement area,
forming a gateway or lookout. Skirting this however, would not be much of a problem.
Amongst the sites dated to a more recent period is TET 65 which is located on the ridgeline
above TET 8. It marks the beginning of a collection of house platforms and shrine structures that
proceed up to the top of this ridge (TET 66-70). The dating may be evidence that people moved
further inland to more defensible, or at least less easily detected, locations with time – perhaps in
response to increased raiding and conflict. Other sites dated to a similar period are also positioned
up narrow ridges (TET 12) or sit atop high uplifted coral bluffs looking directly out to sea (TET
60, TET 54). The latter have walled areas sectioning off an area of flat land at the edge of the cliff,
within which house platforms and shrines are placed. Again the position of these settlement
complexes seems primarily to be determined by defensive considerations.
Research has only just started on Tetepare but already we have some indication of the
position people living here occupied within regional networks. It is clear in terms of trade and
ritual activity, and in the general organisation of settlement space, that social structures were like
those apparently common in the New Georgia group during the last 400 years – people were
engaged in the same realms of practice. Subtle differences in site construction, funerary ritual, and
settlement location appear to simply represent local variations of shared social structures.
Nevertheless it is equally clear that groups living on Tetepare were not able to sustain a balance
in power through alliance or force, eventually dropping out of the network and off the map. It is
not clear yet why this should have happened, although geography, choice of settlement location,
and perhaps some degree of relative social isolation were probably factors. The wider lesson is
that networks of interaction in the region must have had a complex topology, with fluctuations in
the direction and intensity of linkages through time – networks formed, decayed, expanded and
contracted in probably quite localised ways. Fluctuations in group size may have had an important
140
impact on the generation of social and linguistic diversity, with localised abandonment and migra-
tions possibly acting as internal founder effects. Much work remains to be done however, partic-
ularly in establishing practice and variation during earlier parts of the record so that we may
assess how these localised fluctuations played out over the long term.
CONCLUSION
I regard the evidence presented above for the late period sequences of Rendova and Tetepare to
be indicative of involvement in a shared “community of practice” that also encompassed New
Georgia. I prefer this terminology over “community of culture” because I think it has higher
fidelity to what was actually going on at the time, and because it has more explanatory power
encouraging us to attend to the role those practices played in defining “community”. The practical
complex of headhunting, ritual and trade that is embodied in the materiality of settlement compo-
nents is but one facet of regional social life, albeit a particularly important one. It was a complex
that increasingly focused on coastal interaction in opposition to inland-oriented practices symbol-
ically allied to group origins, tenure and fertility. The former played out at a greater regional scale
than the latter, which may well have found expression in much more locally variable forms,
propagating diversity. The gendering of this division in local understanding is interesting, creating
gradients of interaction within social groups and their settlements – coastal dwelling men clearly
had the highest degree of access to neighbouring communities, and inland dwelling women
probably had the least. This implies that the influence of isolation on diversity was graded and
partial, through time, space and society (e.g., with women and children perhaps fuelling linguistic
change, even as men circulated cultural practices and artefacts throughout wider regions).
The Touo language for example has a definite non-Austronesian structure and much of the
vocabulary is unique. But with respect to practices, objects and roles associated with coastal
oriented life (involving the trade of shell rings, canoe travel, and headhunting) the vocabulary is
shared with neighbouring Austronesian groups (Table 6.3). We could regard this as evidence of
“borrowing” if we wanted, but I think this makes less sense than regarding it as evidence of active
engagement in particular realms of shared practice, a common framework relating to social inter-
action (raiding, exchange and associated ritual) in the region. Touo people appear to have been
active participants in the development of these practices throughout their duration, and were
invested in their ontological salience. Borrowing has little explanatory weight in such situations
because complex aggregate practices do not have pin-point origins, developing instead within
practical networks through time/space. Equally, classifying Touo people as part of a regional
culture would obscure our understanding of the processes giving rise to diversity, ignoring the
particular nature of the practices that were shared.
As noted, we can regard the opposition between headhunting and trade as a mirror around
which relationships were structured. These practices set up collections of structured relations
which separated parties into defined wholes – creating identities. My argument is that this enabled
understandings of difference to be worked out at a regional scale, representing an escalation or
amplification of related practices that once turned on a more local coast/inland division. The
archaeological evidence suggests that this developed as a regional phenomenon, rather than as
something that began in one centre and was emulated by others. What we are probably seeing
over the long term is the evolution of a network, variously contracting and reconnecting, with
different nodes joining in or dropping out in the process, but ultimately guided by a common
understanding of practical ways for negotiating relations. Thus, by the historic period we have a
collection of language groups and tribal identities linked in a community of practice that helped
people make sense of their differences.
141
Table 6.3 Touo – English, [Roviana] word list.
ACKNOWLEDGEMENTS
Fieldwork on Rendova was conducted during a postdoctoral appointment at the Leverhulme
Centre for Human Evolutionary Studies, Cambridge, funded by the AHRC through the European
Science Foundation EUROCORES Origins of Man, Language, and Languages (OMLL) program
“Pioneers of Island Melanesia”. I would like to thank Robert Foley and Marta Lahr for their
support during that appointment, and Kenneth Roga and Matthew Suka for their field assistance.
Subsequent work on Rendova was funded by the University of Otago. Research on Tetepare is
funded by the Royal Society of New Zealand Marsden Fund. I thank the Tetepare Descendants
Association for permission to work on Tetepare, and for logistic assistance on the island. Allan
Tippet Bero, Mike and Jeanine D’Antonio, and especially my long-term guide Hanakolo Suka
deserve thanks for their help in enabling this research.
142
REFERENCES
Allen, J., 1984. Pots and poor princes: A multidimensional approach to the role of pottery trading in coastal
Papua. In S.E. van der Leew and A.C. Pritchard (eds), The Many Dimensions of Pottery: Ceramics
in Archaeology and Anthropology. Amsterdam: Inst. Prae-Protohist, pp. 409-63.
Aswani, S., 2000. Changing identities: The ethnohistory of Roviana predatory head-hunting. Journal of the
Bayliss-Smith, T., E. Hviding and T. Whitmore. 2003. Rainforest composition and histories of human
disturbance in Solomon Islands. Ambio, 32: 346-52.
Bellwood, P., 1996. Phylogeny vs, reticulation in prehistory. Antiquity, 70: 881-90.
Bellwood, P., 2005. First Farmers, the Origins of Agricultural Societies. Oxford: Blackwell.
Bennett, J.A., 1987. Wealth of the Solomons. A History of a Pacific Archipelago, 1800-1978. Pacific Islands
Monograph Series 3. Honolulu: University of Hawai‘i Press.
Bronk Ramsey, C., 2001, Development of the radiocarbon calibration program OxCal, Radiocarbon,
43 (2A) 355-363.
Bronk Ramsey, C., 2007, Deposition Models for Chronological Records. Quaternary Science Reviews,
(INTIMATE special issue) in press.
Brumann, C., 1999. Writing for culture: Why a successful concept should not be discarded. Current
Anthropology, 40: s1-s27.
Chikamori, M., 1967. Preliminary report on archaeological research in the Western Solomons. In S. Itoh
and M. Chickamori (eds), A Brief Report on Research into the Prehistory and Anthropology of
Various Islands in the British Solomon Islands Protectorate. Hamamatsu: Kaimeido, pp. 8-42. [in
translation].
Dunn, M., G. Reesink and A. Terrill, 2002. The East Papuan languages: A preliminary typological
appraisal. Oceanic Linguistics, 41: 28-62.
Dunn, M. and H.M. Ross, 2007. Is Kazukuru really Non-Austronesian? Oceanic Linguistics, 46: 210-31.
Dureau, C., 2000. Skulls, mana and causality. Journal of the Polynesian Society, 109: 71-97.
Felgate, M.W., 2007. Leap-frogging or limping? Recent evidence from the Lapita littoral fringe, New
Georgia, Solomon Islands. In S. Bedford, C. Sand and S.P. Connaughton (eds), Oceanic Explorations:
Lapita and Western Pacific Settlement. Terra Australis 26. Canberra: ANU E Press, pp. 123-40.
Geertz, C., 1973. The Interpretation of Cultures. New York: Basic Books.
Golson, J., 1961. Report on New Zealand, Western Polynesia, New Caledonia and Fiji. Asian Perspectives,
5: 166-80.
Gosden, C., 1992. Dynamic traditionalism: Lapita as a long term social structure. In J-C. Galipaud (ed.),
Poterie Lapita et Peuplement. Nouméa: ORTSOM, pp. 21-26.
Green, R.C., 1992. Definitions of the Lapita cultural complex and its non-ceramic component. In J-C.
Galipaud (ed.), Poterie Lapita et Peuplement. Nouméa: ORTSOM, pp. 7-20.
Grimes, S., 2003. A History of Environmental Change and Human Impact Since 3750 BP in the New
Australia, Perth.
143
Hall, A.M., 1964. Customs and culture from Kazukuru: Folklore obtained after the discovery of the shrine
at Bao. Oceania, 35: 129-35.
Hocart, A.M., [n.d.a.]. Chieftainship. Wellington: Alexander Turnbull Library.
Hocart, A.M., [n.d.b]. Roviana: Topography & Chiefs. Wellington: Alexander Turnbull Library.
Hocart, A.M., [n.d.c]. Trade & Money. Wellington: Alexander Turnbull Library.
Hocart, A.M., 1922. The cult of the dead in Eddystone of the Solomons. Journal of the Royal Anthropologi-
cal Institute of Great Britain and Ireland, 52: 71-117, 259-305.
Hocart, A.M., 1931. Warfare in Eddystone of the Solomon Islands. Journal of the Royal Anthropological
Institute of Great Britain and Ireland, 61: 301-24.
Hviding, E., 1996. Guardians of Marovo Lagoon: Practice, Place and Politics in Maritime Melanesia.
Pacific Islands Monograph Series 14. Honolulu: University of Hawai‘i Press.
Kroeber, A.L. and C.K.M. Kluckhohn, 1952. Culture: A Critical Review of Concepts and Definitions. New
York: Vingate Books.
Lilley, I., 1999. Lapita as politics. In J-C. Galipaud and I. Lilley (eds), The Pacific from 5000-2000 BP:
Colonisation and Transformations. Paris: Editions de IRD, pp. 21-29.
Lilley, I., 2006. Archaeology of Oceania: Themes and issues. In I. Lilley (ed.), Archaeology of Oceania:
Australia and the Pacific Islands. Oxford: Blackwell, pp. 1-30.
Mann, P., F.W. Taylor, M.B. Lagoe, A. Quarles and G. Burr, 1998. Accelerating late Quaternary uplift of
the New Georgia Island Group (Solomon island arc) in response to subduction of the recently active
Woodlark spreading center and Coleman seamount. Tectonophysics, 295 (3-4): 259-306.
McDougall, D., 2000. Paths of Pinauzu: Captivity and social reproductions in Ranongga. Journal of the
Nagaoka, T., 1999. Hope Pukerane: A Study of Religious Sites in Roviana, New Georgia, Solomon Islands.
Parker, R., 1994. Maekera: The Life Story of Hereditary Chief Nathan Kera and the Saikile Community of
Solomon Islands. [Honiara]: Solomonesia Productions.
Radcliffe-Brown, A., 1952. Structure and Function in Primitive Societies. London: Cohen and West.
Reimer, P.J., M.G.L. Baillie, E. Bard, A. Bayliss, J.W. Beck, C.J.H. Bertrand, P.G. Blackwell, C.E. Buck,
G.S. Burr, K.B. Cutler, P.E. Damon, R.L. Edwards, R.G. Fairbanks, M. Friedrich, T.P. Guilderson,
A.G. Hogg, K.A. Hughen, B. Kromer, G. McCormac, S. Manning, C. Bronk Ramsey,
R.W. Reimer, S. Remmele, J.R. Southon, M. Stuiver, S. Talamo, F.W. Taylor, J. van der Plicht
and C.E. Weyhenmeyer, 2004. IntCa184 terrestrial radiocarbon age calibration, 8-26 cal kyr BP,
Radiocarbon, 46 (3) 1829-1858
Roberts, J., C. Moore and A. Romney, 1995. Predicting similarity in material culture among New Guinea
villages from propinquity and language: Alog linear approach. Current Anthropology, 36: 769-99.
Schneider, G., 1998. Reinventing identities: Redefining cultural concepts in the struggle between
villagers in Munda, Roviana Lagoon, New Georgia Island, Solomon Islands, for the control of
land. In J. Wassmann (ed.), Pacific Answers to Western Hegemony: Cultural Practices of Identity
Construction. Oxford: Berg, pp. 191-211.
Sheppard, P.J., R. Walter and T. Nagaoka, 2000. The archaeology of head-hunting in Roviana Lagoon, New
Georgia. Journal of the Polynesian Society, 109 (1): 9-37.
Somerville, B.T., 1897. Ethnographical notes in New Georgia, Solomon Islands. Journal of the Royal
Anthropological Institute, 26: 357-413.
144

Spriggs, M., 2004. Is there life after Lapita and do you remember the 60s? The post-Lapita sequences of
the Western Pacific. In V.J. Attenbrow and R. Fullagar (eds), Pacific Odyssey: Archaeology and
Anthropology in the Western Pacific: Papers in Honour of Jim Specht. Records of the Australian
Strathern, M., 1988. The Gender of the Gift. Berkeley and Los Angeles: University of California Press.
Stuiver, M., P.J. Reimer, E. Bard, J.W. Beck, G.S. Burr, K.A. Hughen, B. Kromer, G. McCormac, J. Van der
Plicht and M. Spurk, 1998. INTCAL98 radiocarbon age calibration, 24,888-8 cal BP, Radiocarbon,
48 (3) 1841-1883.
Tedder, M.M., 1976. Old Kusaghe: With additional field notes by Susan Barrus. Journal of the Cultural
Association of the Solomon Islands, 4: 41-95.
Terrell, J.E., T. Hunt and C. Gosden, 1997. The dimensions of social life in the Pacific: Human diversity
and the myth of the primitive isolate. Current Anthropology, 38: 155-96.
Thomas, T., 2004. Things of Roviana: Material Culture, Personhood and Agency in the Nineteenth Century
Solomon Islands. Unpublished Ph.D. thesis, University of Otago, Dunedin.
Thomas, T., P.J. Sheppard and R. Walter, 2001. Landscape, violence and social bodies: Ritualized
architecture in a Solomon Islands society. Journal of the Royal Anthropological Institute, 7: 545-72.
Wahome, E., 1999. Ceramics and Prehistoric Exchange in the Admiralty Islands, Papua New Guinea.
Walter, R. and P.J. Sheppard, 2000. Nusa Roviana: the archaeology of a Melanesian chiefdom. Journal of
Field Archaeology, 27: 295-318.
Walter, R. and P.J. Sheppard, 2006. Archaeology in Melanesia: A case study from the Western Province of
Walter, R., T. Thomas and P.J. Sheppard, 2004. Cult assemblages and ritual practice in Roviana Lagoon,
Solomon Islands. World Archaeology, 36: 142-57.
Waterhouse, J.H.L. and S.H. Ray, 1931. The Kazukuru language of New Georgia. Man, 31: 123-26.
Welsch, R., J. Terrell and J. Nadolski, 1992. Language and culture on the north coast of New Guinea.
American Anthropologist, 94: 568-600.
145
Chapter 7
THE SOUTHEAST SOLOMON ISLANDS CULTURE HISTORY

PROJECT: Principal Investigators’ Overview of the 1970s
Project Including Recent and Current Research
Roger C. Green and Douglas E. Yen
INTRODUCTION
The overall objective of this paper is to outline the current research and the documentation efforts
that have been or are being undertaken for the whole body of data that was initially gathered by
the participants in the Southeast Solomon Islands Culture History Project (SESP). This essay
should also serve as a guide to the majority of publications that have appeared since 1996, or are
now “in press” or under submission review, and to those that are in preparation or planned in
relation to the ongoing research on the databases and collections. In short, the publication of this
essay serves to inform other researchers of the existence of various papers and work in progress
or recently completed. It is a condensed and more coherent summation of earlier informal
documents that are in circulation.
Numerous articles and monographs stemming from the Southeast Solomon Islands Culture
History Project (SESP) appeared during the 1970s and in the decade or so thereafter. Most of
them are readily traced from the bibliographies in Kirch (1997, 2000) and Spriggs (1997), or can
be ascertained from the literature cited by these authors. Thus, 1996 has been taken as a general
rather than a specific cut-off point for the citations appearing in the various sections of this
document which outline the current research projects and efforts toward their publication in a
number of formats. Some relate to unpublished or “grey” literature texts bearing on the body of
data that was gathered in the field between 1973 and 1979 by the original SESP participants. The
rest are, or will in due course, be reasonably accessible. The extent of these programmes is fully
outlined by Green (1973) and by Yen (1982), and has been assessed by Foana‘ota (1996) in a
more general survey. Hence those aspects do not require repetition, and an overview can be seen
through a glance at Figure 7.1.
For the convenience of those with particular interests, and for coherence in presentation, we
employ an arbitrary set of categories in which to frame our discussion. The boundaries adopted
are fuzzy and overlapping so there is an occasional repetition in citation from one section to
another. Enquiries about all ongoing research projects, or intended publications, should in the first
instance be addressed to the people concerned. More general questions about access to the collec-
tions, or proposals for new research should be addressed to Peter Sheppard or Geoffrey Irwin, of
the Department of Anthropology, University of Auckland. From June 2009 they will commence
the supervision, curation and storage of these collections and their related documentation until
such time as they may safely be returned to the Solomon Islands National Museum.
Researchers using this document should be aware that while it attempts to be fully indicative
of the current situation, it makes no claim to be exhaustive in its content. Things invariably
change, and we have been selective in the citations from the published literature since 1995,
providing only those that seem most pertinent to each section. We have endeavoured to record all
those items known to be “in press”, in preparation, or intended for publication.
147
CHANGE THROUGH TIME
148
Figure 7.1 Archaeological coverage in the Solomon Islands (Sheppard and Walter 2006: Figure 1).
The Southeast Solomon Islands Culture History Project
THE TOPIC SECTIONS: I TO VI
Section I: Oceanic Prehistory Records

This series was originally published in microfiche during the period 1976 to 1989, under the
auspices of the University of Auckland Archaeology Society based in the Department of
Anthropology. Roger Green, who was one of the early editors, has been engaged in facilitating
the re-issue of microfiche papers in a compact digital form that could be accessed through
libraries and the data therein made readily available. The objective is to have the full series
available online as an e-series of the new Research in Anthropology and Linguistics [RAL-e]
series, as a means of reaching a much wider audience. An electronic version, which owes much
to Dorothy Brown who scanned all the microfiche copies onto disk, should enable this to happen
either this year or the next.
In the Oceanic Prehistory Records [OPR] Microfiche Series, seven items related to SESP
furnish basic information of use to those researchers working with these collections or the
databases related to them. To highlight their existence, they are listed here:
OPR 4: Joint Issue on data from the Solomon Islands Cultural History Project. 1977.
(a) Black, Stephen J. and Roger C. Green. Radiocarbon Dates from the Solomon Islands to
1975.
(b) Burnett, William C. and Charles D. Fein. Geological Report on Samples from Lapita Sites
SE-RF-2 Solomon Islands Cultural History Project.
OPR 5: Green, Roger C. and A.Y. Dessaint. Measurements of Polynesian Adzes. Some
Comparative Results from Samoa and New Zealand. 1978.
OPR 7: Studies in the Sourcing of Materials from the Southeast Solomons. 1978.
(a) Dickinson, William R. Sand Tempers in Southeast Solomon Islands Sherds from Feru Rock
shelter on Santa Ana and Lapita Sites in the Main Reef and Santa Cruz Group (Petrographic
Report WRD-68).
(b) Moore, P.R. [Appendix by B.D. Hackman]. Petrography of Adzes from the Southeast
Solomon Islands.
(c) Green, Roger C. Notes on Adze Flakes, Oven Stones, Pumice, Muscovite-Garnet-Schist and
Metamorphosed Sandstone Specimens from the Main Reef/Santa Cruz Lapita Sites, Southeast
Solomons.
OPR 8: Sayes, Shelley. The Ethnohistory of Arosi. 1989.
Section II: Protohistoric [16th and 17th Century Spanish Text Aided] and
Historic [18th to 20th Century Text Aided] Archaeology
Martin Gibbs, a historic archaeologist in the Department of Archaeology, School of Archaeology
and Historical Inquiry at the University of Sydney, has decided – in association with David Roe,
James Cook University – to further investigate the whole issue of Spanish contact with the central
and eastern Solomon Islands in the 16th and 17th century A.D. In late 2008 Gibbs assessed the
Pamua site [the Spanish hilltop camp site and that on the flat below which was an occupied indig-
enous settlement when the ship arrived] as well as the Spanish encampment at the head of
Graciosa Bay, to determine their potential for further research.
149
Besides extensive unpublished data in the primary field records and a few laboratory reports
(copies of which have been made available to Gibbs), the collections from the Pamua sites, up to
now held by the Department of Anthropology at the University of Hawaii, Manoa, have been
transferred to the University of Auckland are now included with other collections held in the
storage facilities of the Department of Anthropology (refer Appendix 1 for details).
Nine items in print or in preparation serve to update the published literature on these sites.
References
Bedford, Stuart, William Dickinson, Roger C. Green, Graeme K. Ward, 2009. “Detritus of
Empire: Seventeenth Century Spanish Pottery from Taumako, Southeast Solomon Islands,
and Mota, Northern Vanuatu”. Journal of the Polynesian Society, 118 (1): 69-89.
Doherty, Moira W., 2007. Post-Lapita Developments in the Reef-Santa Cruz Islands, Southeast
Solomons. Unpublished Ph.D. thesis, University of Auckland, Auckland. [For this topic,
see especially Appendix 3.1 entitled “The Historic Period AD 1595 – AD 1935”, as well
as archaeologically excavated indigenous assemblages recovered from both Nendö and
the Main Reef Islands belonging to this temporal interval].
Green, Roger C. and M.W. Kaschko, [ms. in revision]. This is based on M.W. Kaschko
[n.d.]. Field Report: 1975 Excavations on Site BB-2-15, Makira, S.E. Solomons. [An
unpublished progress report by the author for the Department of Anthropology, University
of Hawaii at Manoa, Honolulu].
Leach, F., and J. Davidson. 2008. Archaeology on Taumako: A Polynesian Outlier in the Eastern
Solomon Islands. Dunedin: New Zealand Archaeological Association Special Publication.
OPR 8: Sayes, Shelley. The Ethnohistory of Arosi. 1989.
Spriggs, Matthew, 1997. Pages 235-49 of Chapter 8: Ships from the West, in The Island
Melanesians. Oxford: Blackwell.
Walter, Richard and Roger C. Green [in final stages of preparation]. Monograph about the
Su‘ena site on the island of Ugi [Uki], and its 500-year sequence starting in the 15th
century and lasting into the 20th century A.D. Note that the longer sequence of this site
overlaps in part that of the Pamua flatland indigenous settlement site (1500-1700 A.D.).
White, Moira, 2002. The Spanish sherds from San Cristobal. Journal of the Polynesian Society,
111 (3): 249-54.
White, Moira, 2007. The material culture of Makira. In A. Anderson, K. Green and F. Leach
(eds), Vastly Ingenious: The Archaeology of Pacific Material Culture in Honour of Janet
M. Davidson. Dunedin: Otago University Press, pp. 243-62.
Section III: Projects Based on Ugi, Ulawa and Santa Ana in the Near
Oceania Area of the Southeast Solomon Islands
1. The excavated Su‘ena assemblages from the island of Ugi [or Uki] between San Cristobal and
Ulawa [or Ulaua] are at present held in the Department of Anthropology, University of Otago, in
the custody of Richard Walter. This collection has been the subject of several M.A. theses and
Honours Research dissertations, summations of which are to be incorporated in the monograph
referenced above. In due course, after this monograph is published, the collections will be trans-
ferred to Auckland for storage (Appendix 1).
150
This invaluable record of assemblages, from c. 1500 years B.P. into the 20th century A.D., is
intended for publication in a monograph series of either the New Zealand Archaeological Association
or the University of Otago Studies in Prehistoric Anthropology, or as a museum bulletin.
2. The Graeme Ward collections, largely composed of chert artefacts [both surface and excava-
tion], are from sites on the island of Ulawa, the southernmost island in the Malaita chain. The
Ward Collection, held in the research facilities of the Department of Anthropology, University of
Auckland, is available to students interested in its study under the supervision of Peter Sheppard.
It forms a part of Peter’s long-standing research programme involving items made from chert.
Graeme Ward currently holds the documentary records for these collections and is willing to
make copies available to any researcher undertaking their further investigation.
3. Field research, carried out in large part by Pamela Swadling [1976], which was focused on the
small part-raised atoll island of Santa Ana, off the southern end of San Cristobal [Makira] Island.
In addition, Roger Green conducted the re-excavation of one rock shelter site [Feru II]: further
potsherds were recovered from its basal layers and dated to 2900-3000 B.P. Douglas Yen investi-
gated the ethnohistory of Santa Ana’s present day and past agricultural practices. Both of the
initial reports from Swadling and Yen have been published. However, more recent publications
should also be consulted, as they serve to further enhance the interpretation of all sites on the
island. Additional articles include:
References
Swadling, P., 1988. Radiocarbon dates from a mound in Nafinuatogo village, Santa Ana. Bulletin
of the Indo-Pacific Prehistory Association, 9: 101-3.
Swadling, P., 1996. The distribution of Lapita sites and coral reef resources in the Solomon
Islands. In Janet Davidson, Geoffrey Irwin, Foss Leach, Andrew Pawley and Dorothy
Brown (eds), Oceanic Culture History: Essays in Honour of Roger Green. Dunedin: New
Zealand Journal of Archaeology Special Publication, pp. 237-39.
Swadling, P., 2000. Changing marine interests and their implications for the settlement history
of Santa Ana, an island in the Southeast Solomon Islands. In Atholl Anderson and Tim
Murray (eds), Australian Archaeologist. Collected Papers in Honour of Jim Allen.
Canberra: Coombs Academic Publishing, Australian National University, pp. 365-71.
The second of the articles above includes a brief discussion of the Feru II rock shelter
excavations by Green. Although containing an overly reduced illustration of the stratigraphy, it
briefly discusses and provides the context for the dating of its lowermost deposits. The dates are
described in greater detail in OPR 4, while the few ceramic sherds recovered are discussed within
a comparative context by Kirch and Rosendahl (1976: 235-36) in their article on the early ceramic
period of settlement of Anuta (included in the Green and Cresswell [1976] edited volume). Still
further commentary on the dating and interpretation of these pottery sherds, the only ones
recovered from this part of the Southeast Solomons so far, appears in Sheppard and Walter (2006:
57-58). Peter Sheppard intends to carry out new investigations in Santa Ana in 2009 and thereaf-
ter, and will any include new insights on the Feru II data in his publications.
Section IV: Overview of Ethnobotany Projects

No general summation of the SESP investigations, largely conducted by Yen, has been published.
Here, we provide information on the location of various ethnobotanical collections made by Yen
and references to his recent papers (since 1995) that have furthered their interpretation after his
subsequent broader-scale comparative work in New Guinea and the Bismarck Archipelago. A
151
separate paper by Yen, appearing in this volume, identifies the principal outcomes from the wide-
ranging endeavours he started during the SESP. One of Yen’s concerns is irrigation systems and
their discontinuous distribution. Another is the current status of research on the seven species of
Canarium that have now been identified, including the two among them of most importance to
this project. One issue is the dominance of C. indicum in the forests of San Cristobal and the main
Solomon Islands, and its virtual absence (except by modern introduction) in Santa Cruz and the
Outer Eastern Islands, where the cultivated C. harveyi – with larger edible nuts – dominates. The
incidence of both in the archaeological materials from the Banks Islands, excavated later by
Graham Ward, gave a jolt to any neat, simple interpretation of the Canarium finds in our project.
References
Yen, D. E., 1996. Melanesian arboriculture: historical perspectives with emphasis on the genus
Canarium. In M.L. Stevens, R.M. Bourke and B.R. Evans (eds), South Pacific Indigenous
Nuts. Canberra: ACIAR, pp. 36-44.
Ethnobotanical Collections: Depositions
Plant remains from archaeological excavations in Tikopia – B.P. Bishop Museum (Department of
Anthropology), Honolulu, Hawaii.
Canarium harveyi seed (nut) collection representative of the Vanikoro and Santa Cruz tree
populations – B.P. Bishop Museum (Department of Botany, herbarium), Honolulu, Hawaii.
Canarium indicum seed collection by David Roe, from tree populations of Northwest Guadalcanal
– Department of Prehistory, Australian National University, Canberra.
Canarium remains from San Cristobal archaeology (excavator: M. Kaschko) – These were
in the University of Hawaii (Department of Anthropology), and following a recent re-examination
by Yen, have been returned to the University for formal deposition in the B.P. Bishop Museum
(Department of Botany, herbarium) for safekeeping and future study.
Section V: The Outer Eastern Islands of the Solomons, the Western-most

Region Among the Island Groups of Southeastern Island Melanesia
1. The Outer Eastern Islands of the Solomons lie in Remote Oceania and were first settled around
3150 B.P. The sequences for these islands and island groups have been described in summary
form by Green (1997) and by Kirch (2000: 93-98, 135-47). The content of these articles can now
be updated. Yet, it is unnecessary to expand on them here since research during the past decade
has refined those summaries. Much of the current research effort relates to establishing a regional
Lapita tradition within the Outer Eastern Islands of the Solomons, covering the period from
c.3150 B.P. to c.1900 B.P. and documenting its changing content.
Doherty’s (2007) analysis of information from her Group III [the sub-phase plainware]
ceramic assemblages within the Lapita tradition for Nendö and the Main Reef Islands has filled
a gap in that sequence not adequately covered by previous literature. Doherty also compares the
Lapita sequences of all the other island groups within the Outer Eastern Islands to that which she
constructed for the Reef/Santa Cruz Group. However, an archaeologically undocumented interval
occurs in the Reef/Santa Cruz sequence following the regional Lapita tradition: it constitutes a
cultural blank, during which a major volcanic event takes place. Subsequently, entirely new infor-
mation – derived from the investigations of post-Lapita sites dating to the last 500-600 years –
attests to much later developments. That data includes a few Early Aceramic Group II site assem-
blages: there are also a number of assemblages assigned to the Doherty Group I (i.e., belonging
to the Late Aceramic/ Protohistoric/ Historic period).
152
2. A Taumako sequence has been developed from SESP excavation assemblages recovered by
Janet Davidson and Foss Leach, although its content has never been presented in any detail. A
monograph under their authorship has now been published. The assemblages recovered span
some 3000 years, starting with ceramic phases that are part of the Outer Eastern Islands regional
Lapita tradition. The sequence ends with the Protohistoric period of Spanish contact in 1606 A.D.
and continues into the historic period well known from ethnographic accounts.
The Namu burial ground, belonging to the mid-15th through the 16th century A.D., is included
in the sequence: excavated materials have been the subject of intensive studies in paleo-anthropol-
ogy. The skeletal collection, currently held in the Department of Anatomy and Structural Biology
at the University of Otago Medical School is overseen by Hallie Buckley of that department.
The portable artefact and ecofact part of that SESP collection remains analysed by Janet
Davidson and Foss Leach is now held in storage at Te Papa Tongarewa Museum in Wellington.
More information on these two collections and their future repositories appears in Appendix 1.
The following published accounts document the research completed on skeletal material
recovered from the Namu burial ground. To this date, its anthropometry has been described by
quite a number of physical anthropologists – although only some of the other results from these
repeated studies are readily available.
References
Buckley, H.R., 2000a. A possible fatal wounding in the prehistoric Pacific Islands. International
Journal of Osteoarchaeology, 10: 135-41.
Buckley, H.R., 2000b. Health and Disease in the Prehistoric Pacific Islands. Part I. Unpublished
Ph.D. thesis, University of Otago.
Buckley, H.R. and N. Tayles, 2003a. Skeletal pathology in a prehistoric Pacific Island sample:
Issues in lesion recording, quantification, and interpretation. American Journal of Physical
Anthropology, 122: 303-24.
Buckley. H.R. and N. Tayles, 2003b. The functional cost of yaws (Treponema pertenue) in a
prehistoric Pacific Island skeletal sample. Journal of Archaeological Science, 30: 1301-14.
Houghton, Philip, 1996. People of the Great Ocean. Aspects of Human Biology of the Early
Pacific. Cambridge: Cambridge University Press [especially pp. 43, 55, 185, 192-93, 206,
214-15, 232-33].
Leach, F., and J. Davidson. 2008. Archaeology on Taumako: A Polynesian outlier in the Eastern
Solomon Islands. Dunedin: New Zealand Archaeological Association Special Publication.
Section VI: An Outline of Current Research on the Reef/Santa Cruz

Decorated and Plainware Lapita Assemblages (Periods I and II)
All of the research projects in this section are either reanalyses or new ventures which add to the
large information base provided by earlier literature that is currently in use by most scholars.
These ongoing investigations are presented here under a number of sub-headings, along with the
articles that relate to each of the sub-headings.
1. Dating the Reef/Santa Cruz Regional Tradition of Lapita Sites – Decorated and Plainware
Because a large number of reservations have been expressed about the adequacy of the published
14C dates for these sites, the whole issue has been subjected to review. The articles involve the
addition of many new dates on samples recovered in the initial investigations, which at that time
153
could not provide satisfactory 14C determinations that new protocols in the radiocarbon dating
field now permit. One of these is AMS dating. Another is the determination of appropriate Delta
R marine offsets to apply to animals with varying marine intakes in their diet. A third is the
pooling and calibration of these dates through Bayesian analysis. For cogent reasons, attempts at
dating these sites by obsidian hydration never proved satisfactory as a useful method for deter-
mining their age (Ambrose 1996).
The separate treatment of Lapita sites from deep cave deposits with multiple layers, spanning
long periods of time, need to be treated as special cases. This is because decorated or rim sherds
are frequently few in number or nil throughout the time they were used and their inhabitants were
using ceramics. For example the dated deposits at the very base of Növlaö Cave (SE-SZ-47) are
as early as those of earliest sites associated with the highly decorated Lapita potsherds of SZ- and
RF-2 (Anson, Walter and Green 2005: 20).
Among the three decorated Lapita sites there are seven new 14C dates. Two (on marine shell)
related to the SE-RF-2 site of Nenumbo, have recently been published. The refinement in dating that
they provide includes the calculation of an appropriate regional ΔR value for a period around 3000
years ago. An unpublished 14C determination on a chicken bone from RF-2 presents a very interest-
ing case study in respect to the need for determining an animal’s diet when calculating its age. This
is to be the subject of a separate article by Alice Storey, Nancy Beavan-Athfield and Green.
The published and “in preparation” articles on the age determinations available for RF-2 are:
References
Jones, M., F. Petchey, R.C. Green, P. Sheppard and M. Phelan, 2007. The marine ΔR for
Nenumbo: A case study in calculating reservoir offsets from paired sample data.
Radiocarbon, 49 (1): 95-102.
Storey, Alice, Nancy Beavan-Athfield and R.C. Green, MS. Early chicken from the Lapita site
of Nenumbo, Outer Eastern Islands of the Solomons. [Journal to which this article will be
submitted is yet to be decided].
A third article evaluates the time differential between two of the decorated phase Lapita sites
from the Outer Eastern Islands of the Solomons. It rests on three charcoal 14C determinations
along with site plans and stratigraphy for RF-6, compared to the six for RF-2. The Bayesian statis-
tical analysis shows that RF-6 represents a significantly younger site. The article also covers the
geological evidence for a major Tinakula ashfall covering the Main Reef Islands, and discusses its
relationship to the sites of the Lapita tradition, all of which are covered by this ashfall.
Jones, M. and R.C. Green, 2008. The absolute age of SE-RF-6 (Ngamanie) and its relation to
SE-RF-2 (Nenumbo): Two decorated Lapita sites in the Southeast Solomon Islands. New
Zealand Journal of Archaeology, 29 [2007]: 5-18.
The fourth article presents the information on four shell dates for the Nendö SZ-8 site within
a three-part text. The first part provides an environmental reconstruction for the Lapita site of
Nendö, and relates SZ-8 to other Lapita tradition sites and to the following Tinakula ashfall and
soils derived from it. A second part sets out plans, maps, and stratigraphic cross-sections of SZ-8,
and locates the contexts of the four dates. This establishes those priors adopted during the Bayesian
analysis. The outcome strongly suggests SZ-8 is probably slightly earlier in age than RF-2.
Green, R.C., M. Jones and P. Sheppard, 2008. The reconstructed environment and absolute
dating of SE-SZ-8 Lapita site on Nendö, Santa Cruz, Solomon Islands. Archaeology in
Oceania, 43 (2): 49-61.
154
Finally, in one chapter of Doherty’s 2007 Ph.D. thesis, the dating of the Group II plainware
site assemblages is shown (after calibration) to encompass a 500-600 year interval that includes
the already published dates of McCoy and Cleghorn as well as the additional dates from the Reef
Island plainware Lapita site of SE-RF-19. Overall, the regional Lapita tradition within the Outer
Eastern Islands of the Solomons – including also Tikopia, Anuta and Taumako – now stands as
one of the best dated of any of the regional sequences from Remote Oceania. The number of sites
belonging to this 1000-year duration regional Lapita tradition is also impressive.
2. Responses to Negative Commentary by Best and by Felgate Questioning the Dating
and Adequacy of Site Samples from SE-RF-6, RF-2 and SZ-8
Negative commentary is the life-blood of science, and when offered by informed commentators,
usually requires a well-argued rejoinder whenever the content and deductions appear unsound. A
short response to Best (2002) and Felgate (2003), and others who agree with their position, has
been published; a longer comprehensive assessment, examining in detail the motif analyses and
the issues involved, has been completed for appearance in e-series form.
In addition it is necessary to reiterate that Anson et al. (2005: 36) have yet again empha-
sised that previous models for the Lapita phenomena developed from the empirical evidence
recovered either in sites of the Eastern Lapita region or those of other regions of Remote Oceania
cannot and should not be applied throughout all those widely distributed locations with Lapita
ceramics. In short such models simply do not conform to situations farther to the west of the
easternmost region, and such models certainly do not fit with the empirical evidence from
various regions in Near Oceania. No single model has ever proven feasible or persuasive for the
Lapita phenomena as a whole.
References
Green, Roger C., MS. An Evaluation of Sample Adequacy for the Lapita-Style Ceramic
Assemblages from Three Sites in the Reef/Santa Cruz Group, Outer Eastern Islands of the
Solomons. Submitted to Research in Anthropology and Linguistics [RAL-e], Department
of Anthropology, University of Auckland.
Sheppard, P.J. and R.C. Green, 2007. Sample size and the Reef/Santa Cruz Lapita sequence. In
Stuart Bedford, Christophe Sand and Sean Connaughton (eds), Oceanic Explorations: Lapita
and Western Pacific Settlement. Terra Australis 26. Canberra: ANU E Press, pp. 141-50.
3. Further Details About Obsidian and Chert Sourcing Studies Prior to 1987
These SESP collections, well studied by Sheppard (1993, 1996), currently require little further
examination. However, one unpublished document may assist those wishing to investigate this
subject further. In 1985, while at ANU, Green wrote a manuscript that was subsequently
mentioned in a publication that appeared in Australasian Archaeometry 1987, edited by Wal
Ambrose (who had assisted in data compilation on which Green drew). Hard copies of this
manuscript are held by four people with an interest in the additional details – Peter Sheppard,
Glenn Summerhayes, Jim Specht, and Robin Torrence. The original tables were reformatted and
re-edited by Dorothy Brown, who prepared the text from a surviving photocopy of the handwrit-
ten manuscript (now lost). She thoroughly checked all tables and assembled a bibliography
covering items cited in that text at that time. The manuscript therefore represents thoughts up to
1985 on how best to determine the source of an item in obsidian. It also furnishes specific data
relevant to each item (not available elsewhere), its context and the analytical treatments used to
determine its source.
155
References
Doherty (2007), reports in detail on a small sample of obsidian items from the now 14C dated
Group II plainware ceramic site of RF-19 that have been assigned with a high probability
to a known source. They were run by Glenn Summerhayes using PIXE-PIGME at the
Lucas Heights facility in Australia. In her discussion of obsidian items from other sites
around Graciosa Bay on Nendö, Doherty also reassesses [with much lower probabilities]
the sources of the obsidian items recovered by the McCoy and Cleghorn (based on
photocopies of the original data sheets for each piece as supplied by McCoy). Green
independently confirmed her assessments as better than the previous attributions, although
the samples should still be analysed by elemental methods to confirm their probable
sources more precisely.
Green, R.C., 1985. MS. An interpretation of source allocations for obsidian from three Lapita
sites in the Reef/Santa Cruz Island group, Southeast Solomons. Researchers should note
that a random sample of 50 flake items from the density separations, argued in 1985
to derive from the Talasea sources, have all been shown through further PIXE-PIGME
analyses conducted by Summerhayes et al. (1998: Table 6.3, pers. com. and a printout)
to be from the Kutau/Bao subsource precisely as was anticipated. This confirms to a high
degree of probability that all those items assigned to the “Talasea” source in the 1987
article on the basis of density are from the Kutau/Bao subsource on the Talasea peninsula,
and not just from some generalised source in that vicinity, nor from the Mopir source on
the Hoskins Peninsula.
Sheppard, P.J., B. Trichereau and C. Milicich. MS. Pacific Obsidian Sourcing by Portable XRF.
(Submitted for review to a Pacific archaeological journal). This re-analysis duplicates,
using new technology, the previously published and/or manuscript allocations, but greatly
refines the sub-source attributions.
4. Shellfish
An analysis of the midden content of the three sites, [SE-RF-6, RF-2 and SZ-8], was first
published by Swadling in 1986. Szabó has taken up the subject from this baseline and expanded
the analysis further by:
(a) Examining the assemblage from the point of view that inhabitants of the hamlet of RF-2 were
also involved in the manufacture of shell tools and valuables. Interestingly, she identifies the
limited presence of industrial yellow lip pearl shell as probable manufacturing debris, some
of it possibly waste while crafting simple fishhooks or similar items not recovered from the
excavated portion of the site.
(b) Re-studying species content of the midden content, thus providing much fuller outcomes for
any comparative exercises done in the future. In this respect she was unable to lend further
support to the effects of predation on the larger individuals of certain shellfish species studied
by Swadling.
(c) Discussing all the classes of shell artefacts recovered from the RF-2 site, relating these to
those found more widely in assemblages of the Lapita Cultural Complex. A contrast with the
common shell objects found in Lapita sites to those from Island Southeast Asia is an
important theme developed in Szabó and O’Connor (2004).
(d) Observing that some of the more robust shells, especially those of Tridacna sp. had been
used as packing around former wooden posts that once stood in the postholes recovered by
156
the archaeologists when they were first set into the coral beach sand substrate to firm up and
stabilise these uprights (Anson et al. 2005: 20).
References
Szabó, Katherine, 2004. Technique and Practice: Shell-working in the Western Pacific and Island
Southeast Asia. Three volumes. Unpublished Ph.D. thesis, Department of Archaeology and
Natural History, Research School of Pacific Studies, Australian National University, Canberra
[see Chapter 5, section 5.3 entitled “Nenumbo (SE-RF-2): Shell assemblage and artefacts”].
Szabó, Katherine, 2007. An assessment of shell fishhooks of the Lapita cultural complex. In
A. Anderson, K. Green and F. Leach (eds), Vastly Ingenious: The Archaeology of Pacific
Material Culture in Honour of Janet M. Davidson. Dunedin: Otago University Press, pp.
227-41.
Szabó, Katherine and S. O’Connor, 2004. Migration and complexity in Holocene Island
Southeast Asia. World Archaeology, 36 (4): 621-28.
Doherty (2007) expands on this topic in new and important ways to include not only the
plainware Lapita sites, but also those from later periods (III, IV, and V of her temporal schema).
Those with an interest in worked shell objects, or the results of additional shell midden analyses,
especially for the later periods following major environmental changes after the Tinakula ashfall
and sediment runoff into the lagoon system will profit from the enhanced discussions of this topic
within this region.
The initial article investigating land snails from sites in the Outer Eastern Islands of the
Solomons was published by Carl C. Christensen and P.V. Kirch in Pacific Science in 1981. There
was however further reportage of results from other sites in this region, which their paper should
have engendered, but they failed to reach published states. In an appendix appearing in the Leach
and Davidson (2008) monograph on the archaeology of Taumako, results from that island group
and from the Nenumbo (SE-RF-2) Lapita site are presented. Moreover, additional analyses on
sub-samples taken from bulk soil samples related to RF-2 and RF-6, examined by Alison
Crowther for her Ph.D., have also been wet sieved by Marshall Weisler to recover shell specimens
of the micro-sized snails from them. The results will be analysed and interpreted in due course in
a publication by Carl Christensen [carlcc@hawaii.edu]. One chapter in Szabó’s Ph.D. thesis also
contains informative data on the macro-sized fossil land snails from that site. Doherty’s Ph.D.
covers the macro and micro-sized land snails found in sites RF-19 and RF-3. There is some old
unpublished data on the land snails from the Feru site excavations on the island of Santa Ana, that
is held by Christensen.
5. Fishing and Fishbone
While the Green (1986) study of fishbone and fishing gear has served as a baseline for the RF-2
site, Yolanda Vogel recently completed a contract re-study of this assemblage. A forthcoming joint
publication by Yolanda and Richard Walter will update the importance of the assemblage for
Lapita studies focusing on the marine component of the subsistence diet at this early stage when
people first moved into Remote Oceania 3000 years ago. It presents outcomes in both MNIs and
NISPs derived from the analyses of all the boney fragments, including those previously assigned
to a “not able to be identified” category, and not just MNI numbers as in Green’s original article.
Doherty (2007) discusses an analysis she conducted of the fishbones from the later Lapita
plainware sites and those from the Aceramic assemblages of the late to historic period, and their
implications. These allow one to address the question of those differences, if any, that may be
157
suggested to have occurred in a preliminary fashion among the various fishing strategies the
inhabitants of the various sites employed between 3000 years ago and 1900 years ago, and then
again during the last 500 years.
Reference
Vogel, Yolanda and Richard Walter, MS. Department of Anthropology, University of Otago.
6. Evidence for Plant Remains
Alison Crowther of the University of Queensland, School of Social Research Archaeology
Program has just completed her Ph.D. on plant residues on decorated assemblages of Lapita
pottery from selected Lapita sites. One of her analyses focuses on sherds from RF-2, with selected
others from RF-6 and SZ-8. She is also examining starch residues and phytoliths in the soil
samples from RF-2 and RF-6, together with Dr Carol Lentfer, a Research Fellow at the Queensland
School of Social Science Archaeology Program.
This phytolith research follows that done by Lentfer and Green (2004) on phytoliths from
the Reber-Rakival Lapita site on Watom Island, and includes one of the dietary reconstructions
completed for Lapita inhabitants and indicative of their subsistence practices: such phytolith
studies demonstrate the presence of otherwise invisible plant remains, including various cultivars
such as banana. From the analyses of the sediment samples completed by Lentfer and Crowther’s
Ph.D. thesis results, several manuscripts are now in preparation for publication.
Researchers with an interest in this topic should also consult Doherty (2007: 408-15) for
discussion of macro-flora remains from a few of the later Lapita plainware sites, along with those
belonging to the Aceramic periods that follow.
References
Crowther, Alison, Investigating Lapita Subsistence and Pottery Use Through Microscopic
Residues on Ceramics: Methodological Issues, Feasibility and Potential. Ph.D. thesis now
under examination. Archaeology Programme, University of Queensland.
7. Faunal Remains
Various attempts, by Lisa Matisoo-Smith and her team of analysts, to extract mtDNA from the rat
bones in the three sites [RF-2, RF-6, and SZ-8] have not proved successful: nor have those carried
out on the pig teeth associated with these assemblages. Currently, Alice Storey has submitted her
Ph.D. thesis on the Asian jungle fowl in the Pacific; this includes chicken bones from the RF-2
site. Although a 14C determination from a chicken bone and its isotopic signatures for delta 13C,
I5N and S34, have been successfully retrieved (see Sub-section 2), the taphonomic processes in
3000 year old open sites such as RF-2 appears to have removed the mtDNA signal from all six
chicken bones. The same phenomenon also applies to the rat and pig bones from this site, which
were also tested for mtDNA.
Although a full listing of the faunal remains from the three sites compiled by the late Alan
Ziegler, with expert identifications, has been available to any interested party, only the overall
content of these investigations was presented in summary form in the published reports of the
1970s. Copies of all this information have now been placed with Melinda Allen, who will curate
the collections of all bones from these assemblages while they are undergoing renewed study in
the laboratories of the Department of Anthropology, University of Auckland. Initially, Melinda is
looking at the extensive assemblage of turtle bone from these collections, particularly those
constituting a substantial collection from RF-2.
158
Despite pig bone, a worked pig tusk, or pig teeth – from one or more of these three sites –
having been reported in the literature, researchers writing on the subject of pig bone in Lapita sites
of Remote Oceania (e.g., A. Anderson 2003) choose to ignore or devalue these sites as possessing
early evidence for the pig in Lapita assemblages from Remote Oceania. As a consequence, when
discussing the significance of an apparent absence of pig bone in the early Lapita sites of New
Caledonia, Fiji and Western Polynesia, some investigators using an absence of pig bones as
indicative of an absence of horticulture employ a far too simplistic formulation. Certainly, in
many regions of Remote Oceania, the presence of pig bone with other evidence for horticulture
implies the early practice of arboriculture and horticulture. However, an absence of early pig
bones in some Lapita sites further to the east in Fiji and Tonga is insufficient grounds on which
to justify a claim that the earliest inhabitants of that region lacked both arboriculture and horti-
culture from the beginning. On the current evidence commencing 3000 B.P., the archaeological
record shows that New Caledonia lacked pigs, yet during the last 2000 years an ever increasing
and extensive practise of irrigation horticulture is undeniable. The ethnographic and linguistic
record of recent centuries supports the associated archaeological evidence.
An important aspect of these bone assemblages from the Reef/Santa Cruz sites, in addition
to good evidence of pig in each, is the noticeable absence within them of extinct or extirpated
birds, and the presence of Rattus exulans and Rattus praetor and the Asian jungle fowl.
References
Matisoo-Smith, E. and J. Robins, 2004. Origins and dispersals of Pacific peoples: Evidence from
mtDNA phylogenies of the Pacific Rat. Proceedings of the National Academy of Science,
1001 (24): 9167-72.
Storey, Alice A., 2008. Migrations Most Fowl: Archaeology and ancient mitochondrial DNA
signatures of Pacific chickens [Ph.D. thesis now under examination, University of
Auckland, Auckland].
Storey, Alice A., Thegn Ladefoged and Elizabeth A. Matisoo-Smith, 2008. Counting your
chickens: Density and distribution of chicken remains in archaeological sites of Oceania.
International Journal of Osteoarchaeology, 18: 240-61.
Doherty (2007) provides identification of and offers discussion about the additional terrestrial
faunal remains from the sites of the post-decorated Lapita period, and especially for the rat bones
from all those temporal periods of Groups IV to I employed in her analyses.
7. Portable Artefacts
An illustrated set of drawings for the adzes and other formal portable artefacts recovered from
these three decorated Lapita sites was prepared in the 1970s. Joan Lawrence has redrawn the
adzes, and is reworking the other drawings. They have also been relabelled and rearranged into
their respective artefacts classes by Green.
Doherty (2007) describes and illustrates the rather different set of portable artefacts found
in the Lapita plainware period assemblages, as well as the very different set from the last millen-
nium A.D. associated with the late assemblages totally lacking in ceramics. Comments by some
researchers that few of the artefacts from the three decorated Lapita sites of the Reef/Santa Cruz
have been illustrated previously indicate oversight on their part (see R.C. Green ”Lapita” (Figs
2.4 and 2.5), in J.D. Jennings (ed.), 1979, The Prehistory of Polynesia, where the majority of the
main artefact forms recovered from the decorated Lapita assemblages are pictured and identified
by site and artefact category).
159
We have formed an impression that a degree of disinterest in comparative portable Lapita

artefacts is countered by overwhelming focus on one set of dentate-decorated ceramic pot forms,
to the exclusion of all else (ceramic and non-ceramic) by some researchers, to deny any status for
Lapita as comprising a wide-ranging cultural complex. It is a contemporary bias that only time
will overcome. Summerhayes appropriately labels as “dentate-centric” the viewpoint of Lapita
archaeologists who dwell on a single decorative technique to define Lapita, a perception with
which we wholeheartedly agree (for a fuller discussion of this issue see Green 2003: 102-12).
Having made this point, some recent developments in the methods and protocols for the
comprehensive study of Lapita decoration have opened up anew the potential for more detailed
analyses of the Reef/Santa Cruz decorated Lapita assemblages. To this end, a re-study of these by
Scarlett Chiu, following her pioneering research on the Lapita site 13A assemblages from the
Koné Lapita site in New Caledonia, provides a new basis for comparative work incorporating the
decorated assemblages recovered from all the Lapita sites of Remote Oceania. Moreover, it is
indicative of how things might precede in future studies of the Lapita ceramics of Near Oceania.
References
Chiu, S., 2005. Meanings of a Lapita face: Materialized social memory in ancient house
societies. Taiwan Journal of Anthropology, 3: 1-47.
Chiu, S., 2007. The ever-changing Lapita face motifs: Case studies from Reef/Santa Cruz sites
and New Caledonia Site 13A. In Stuart Bedford, Christophe Sand and Sean Connaughton
(eds), Oceanic Explorations: Lapita and Western Pacific Settlement. Terra Australis 26.
Canberra: ANU E-Press, pp. 241-64.
Section VII. Integration of Historical Linguistic Investigations with

Outcomes Derived from Archaeology
One never fully realised objective set forth for the two phases of the Southeast Solomon Island
Cultural History Project was to integrate the complex linguistic situation within that region
around the time of European contact with the varied archaeological outcomes the proposed
research was expected to recover. While achieving this aim would not always prove possible, it
was to be explored wherever useful insights into the region’s history might result. In the portion
of the region covered by the languages of the Cristobal-Malaitan branch of the Southeast
Solomonic subgroup, a failure to find coincident archaeological sequences for the full 3000+
years [hinted at by evidence from the rock shelter of Feru, on the islet of Santa Ana (see Section
III) with plainware sherds of Lapita pottery] rendered moot the realisation of this objective.
The most recent historical linguistic enquiries continue to support a date in the 3000 year
B.P. range for the initial establishment of an Oceanic language immediately ancestral to Proto
Southeast Solomonic (PSS) in the eastern portion of the main Solomon Island chain (Pawley
2007a, 2007b, 2007c). This finding indicates a reasonably early movement into the Southeast
Solomons of one migrant community of people from the Bismarck Archipelago bearing a version
of the Lapita Cultural Complex and speaking the language that became PSS. At present this
inference constitutes a linguistics-based proposition that awaits demonstration through the
extensive recovery of appropriately dated Lapita archaeological assemblages in that region. In
addition, the task of finding one or more of possible pre-Lapita assemblages in this region, which
lack pottery and obsidian signatures and are therefore rather ephemeral, remains a truly formi-
dable and very likely an elusive one.
160
161
Figure 7.2 Language sub-groups of the Eastern Solomon Islands.

In contrast, the SESP aim of rectifying a rather confusing linguistic situation in the Outer
Eastern Islands has garnered some degree of success. Here the first steps have been taken toward
the overall objective, which required clarifying the historical relationships among an uncertain
number of languages and dialects found on the large island of Nendö and at least one on the Main
Reef Islands that together comprise the Santa Cruz Island Group. It also involved sorting out just
how many languages existed on the smaller volcanic island groups of Utupua and Vanikoro
whose circumstances and character were then very poorly known. Another part of the same
objective was to place in time the later arrivals from somewhere within the Triangular Polynesia
region far to the east. These migrants spoke some of the better known Outlier Polynesia languages
of Pileni, Taumako, Anuta and Tikopia, each distributed among rather remote and tiny atolls and
volcanic islands lying at a distance from the core set of much older languages found in the central
and western parts of this region.
An introductory article, which formed part of a volume that followed completion of Stage I
of the SESP (Green 1976), outlined the linguistic situation throughout the entire project region,
in light of the preliminary linguistic enquiries undertaken by various participants in the project. It
is well established that the languages spoken in the San Cristobal-Ugi-Ulawa and Santa Ana
region belong to the Cristobal-Malaitan branch of Southeast Solomonic (Lichtenberk 1988).
Moreover, the status of Southeast Solomonic as a high-order subgroup of Oceanic, and one whose
conservatism makes it of great utility in linguistic reconstruction, has seen this group grow in
significance (Ross, Pawley and Osmond 2003: 6-11). In contrast, the situation in the Outer
Eastern Islands of the Solomons has remained something of a puzzle. Various of the Outlier
languages of the region, most have come to agree, should be assigned to a Nuclear Polynesian
subgroup of Samoic derived languages. Yet, in respect to Tikopia and Anuta, debate continues to
ensue over possible Tongan inputs either directly or indirectly by means of demonstrable Tongan
influence on the language of East ‘Uvea (Feinberg 1989, Kirch and Yen 1982: 341-43).
Just before the SESP project began, Wurm (1969, 1970) put forth the evidence he had
gathered to support a claim that the dialects and languages of Nendö and the Main Reefs were
definitely not Austronesian in origin, though they did exhibit signs of extensive borrowing from
various Oceanic Austronesian languages. This included a late period of Polynesian borrowing, a
regional reflection of incoming Outlier neighbours as permanent residents. It was Wurm’s view
that the Reef/Santa Cruz languages instead displayed inherited affinities with certain of the
Papuan languages of Near Oceania that occupied particular areas in western Island Melanesia.
However, the evidence put forward for last these proposals failed to convince many others in the
field of Pacific historical linguistics.
Two SESP members, Peter Lincoln and Christine Cashmore, who recorded information
about the core set of non-Polynesian languages and dialects in the region, on the basis of enquiries
among their speakers, were left unconvinced by Wurm’s interpretations. Lincoln (1978) concluded
that the three main languages of the Reef/Santa Cruz group constitute rather aberrant forms of
Oceanic Austronesian, whose uncommon features may in a few cases have been due to various
degrees of borrowing from Papuan languages in the past. The three languages of Utupua and
another three on Vanikoro, however, were judged to be just as Davenport (1968) had reported,
somewhat unfamiliar kinds of Oceanic Austronesian that were very difficult at that time to place
with confidence within any particular Oceanic subgroup. And there these matters tended to rest
for several decades, without further resolution.
Recently, other scholars with a continuing interest in the history of the Oceanic subgroup,
and its dissolution into a stable set of well defined primary subgroups, have turned their attention
162
to unravelling the likely subgrouping situation among the disparate languages of the Outer
Eastern Island region. Thus, Ross and Næss (2007) have produced a very convincing reassess-
ment of the Main Reef Island language of Äiwoo and its historical relationship to the languages
of Nendö, Utupua, and Vanikoro. This research, considered with that of others, has confirmed the
proposition that all these languages are in fact of Oceanic Austronesian origin, bearing little or no
trace of significant influence from past contact with Papuan languages further to the west.
Those outcomes throw entirely new light on refinements to the linguistic subgroups that can
be postulated as occurring within the Outer Eastern Island region and indications of their relation-
ships one to another. Ross and Næss demonstrate, on the basis of uniquely shared innovations
found only among the languages of Äiwoo and the dialects and two languages of Nendö that these
languages form a single subgroup, descended from Proto Reef/Santa Cruz. Similarly the
languages of Utupua and Vanikoro probably constitute another related subgroup marked by a
different set of innovations. These two subgroups prove to share sufficient phonological innova-
tions to enable Ross and Næss to assign both to an entirely new high-order subgroup, which they
call Temotu. Moreover, they (like Pawley 2007a) regard this new subgroup as one of at least six
primary subgroups of Oceanic. Under this schema a second primary subgroup of Oceanic, though
present only in part within the project region, is Southeast Solomonic, whose languages are
distributed across the small Nggela Island group, the large high island of Guadalcanal, the islands
forming the Malaitan chain, and into the San Cristobal region of Makira Province.
During the last decade, Næss and Hovdhaugen (2007) have engaged in further documenta-
tion and analyses of the various ‘Pileni’ dialects of Outlier Polynesian. One cluster of widely
distributed dialects, spread across a number of tiny atolls and one raised coral island of the Outer
Reef Islands group, has now been convincingly joined to the dialect of Taumako spoken in the
high volcanic islands of the Duff group. The resulting Outlier language has been designated
Vaeakau-Taumako rather than ‘Pileni’. Anuta and Tikopia continue to remain another set of
closely related Outlier languages, though they themselves seem to have had somewhat separate
origins within the Western Polynesia region, as well as an origin different to a still uncertain one
for Vaeakau-Taumako somewhere within Western Polynesia.
In addition to these new perspectives on the Reefs/Santa Cruz group, Sheppard and Walter
(2006) have also proposed, on archaeological grounds, that these languages may well have had
their immediate origin in the Near Oceania region of northern and western New Britain and its
offshore islands, and in particular among those languages of Willaumez Peninsular area from
whence the majority of obsidian items in the very early decorated Lapita sites of Nendö and the
Main Reef Islands have their source (see Section VI, under the subsection on obsidian and chert).
If so, those migrants would appear to have established few, if any, enduring way stations on their
early long distance voyages to the Outer Eastern Islands of the Solomons. Moreover, on the
evidence of the obsidian and its sources in later Lapita sites of the Outer Eastern Islands, those
who initially settled that region continued to maintain their voyaging contacts between the two
well-separated localities over many centuries thereafter, despite distances of up to 2000 km.
During the recent period of all this renewed research effort, Green (2003: 106-7) too
returned to an abiding interest in the languages of the Outer Eastern Islands and their probable
archaeological correlates. He equated the ancestors of speakers of these languages with the
bearers of what – over a 1000 years – became a distinctive regional Lapita tradition: it was
initiated with the first Lapita occupations found on nearly all the major islands and their sequences
documenting the prehistory of the Outer Eastern Islands region (see Green 2003: the upper row
regional sequence of Figure 4). Green then employed a research assistant to assemble in electron-
163
ic format, vocabulary lists for 356 lexical items and their meanings that had been documented for
each of the six languages belonging to the putative Utupua/Vanikoro subgroup. He has continued
to use this for his own research focused on refining the subgrouping, borrowing among the
languages of the region, and back migration to an earlier home island involving speakers of two
of the languages on Utupua and two on Vanikoro. In the first instance, this material was also
provided to colleagues in the Psychology Department of the University of Auckland for entry into
their widely supported, co-operative and inter-institutional compilation of an Austronesian Basic
Vocabulary database of massive size: currently, this encompasses up to 500 languages from the
AN language family – (http://language.psy.auckland.ac.nz ).
This Austronesian vocabulary project, under the direction of Simon Greenhill, Robert Blust
and Russell Gray, made it possible for Greenhill to oversee the entry of 210 lexical items and
languages from certain of lists from Vanikoro and Utupua into this ever growing database. The
entry of the new languages was done with a view to placing them within an Austronesian family
tree under development: it is based on statistical algorithms devised by Gray and Greenhill to
independently determine the time depths involved among the sets of major splits and clusters
recognised within that widely spread Austronesian language family and assess their compatibility
with previous efforts at subgrouping identified using the traditional methods employed by
comparative linguistics.
Some preliminary analyses employing this database were presented at the COOL7 confer-
ence in Noumea (Gray 2007, Greenhill 2007) that unfortunately overlapped the period when a
Lapita Conference was taking place in Honiara. For the Outer Eastern Islands one striking
outcome of the preliminary results was that while each of the Utupua and Vanikoro languages
grouped together as expected, they fell in a cluster with seven languages representative of the
Admiralty Island subgroup, one of the long recognised primary subgroups of Oceanic (Gray and
Greenhill 2007). That outcome suggested they had a great antiquity, of an order similar to that of
the Reef/Santa Cruz subgroup languages – all of which are seen by Ross and Næss (2007) as
constituting a primary Temotu subgroup. A far more searching analysis in much greater depth has
been completed by Greenhill (2008), some of the core results have already been published (Grey,
Drummond and Greenhill 2009).
Green (in prep.) is writing an exploratory essay, which exploits these new propositions in
order to establish more precisely a probable linguistic history for the whole region of the Outer
Eastern Islands of the Solomons. It attempts to relate the new linguistic perspectives to evidence
emerging from recent projects informing the complex history of an island or island groups within
this region. These results feature in other sections of this document as items that have appeared
during the last decade in published form, are in press, or are currently in preparation. Not
unexpectedly, given all this new information for two sub-disciplines and the altered perspectives
it furnishes, the attempts at synthesis have run into many complex and difficult problems.
However, a substantive paper on this topic is currently at an advanced stage of preparation.
References:
Davenport, W.H., 1968. Social organisation notes on the Southern Santa Cruz Islands: Utupua
and Vanikolo. Baessler-Archiv, 16: 207-75.
Feinberg, Richard, 1989. Possible prehistoric contacts between Tonga and Anuta. Journal of the
164
Gray, Russell, 2007. Tangled Trees: What do Phylogenetic Networks Reveal About Linguistic
History? COOL7: Paper presented at the Seventh International Conference on Oceanic
Linguistics. Noumea, Nouvelle-Caledonie, 2-6 Juillet 2007.
Gray, R.D., A.J. Drummond and S.J. Greenhill, 2009. Language phylogenies reveal expansion
pulses and pauses in Pacific settlement. Science, 323: 479-83.
Gray, R.D. and S.J. Greenhill, 2007. A Preliminary Cladistic Analysis Employing the
Austronesian Basic Vocabulary Database. Circulated at the Seventh International
Conference on Oceanic Linguistics. Noumea, Nouvelle-Caledonie, 2-6 Juillet 2007.
Green, R.C., 1976. Languages of the Southeast Solomons and their historical relationships. In
R.C. Green and M.M. Cresswell (eds), Southeast Solomon Islands Cultural History. A
Preliminary Survey. Bulletin 11. Wellington: Royal Society of New Zealand, pp. 47-60.
Green, R.C., [in preparation]. The Outer Eastern Islands of the Solomons: A Puzzle for the Holistic
Approach to Historical Reconstruction. [Incomplete manuscript in possession of author].
Greenhill, S.J., 2007. Language Trees and the Austronesian Basic Vocabulary Database.
COOL7: Paper presented at the Seventh International Conference on Oceanic Linguistics.
Noumea, Nouvelle-Caledonie, 2-6 Juillet 2007.
Greenhill, S.J., 2008. Language as the Archives of History. Ph.D. thesis. University of Auckland,
Auckland.
Kirch, Patrick V. and D.E. Yen, 1982. Tikopia: The Prehistory and Ecology of a Polynesian
Outlier. Bernice P. Bishop Museum Bulletin 238. Honolulu.
Lichtenberk, Frantisek, 1988. The Cristobal-Malaitan subgroup of Southeast Solomonic.
Oceanic Linguistics, 27 (1): 24-62.
Lincoln, P.C., 1978. Reef-Santa Cruz as Austronesian. In S.A. Wurm and L. Carrington (eds),
Second International Conference on Austronesian Linguistics: Proceedings. Pacific
Linguistics C-61. Canberra: Department of Pacific Linguistics, Australian National
Næss, Åshild and Even Hovdhaugen, 2007. The history of Polynesian settlement in the Reef and
Duff Islands: The linguistic evidence. Journal of the Polynesian Society, 116 (4): 433-49.
Pawley, Andrew, 2007a. Linguistic Subgrouping and the Settlement History of the Solomon
Islands. COOL7: Paper presented at the Seventh International Conference on Oceanic
Linguistics. Noumea, Nouvelle-Caledonie, 2-6 Juillet 2007.
Pawley, Andrew, 2007b. Was There Early Lapita Settlement of the Solomon Islands? Bringing
Linguistic Evidence to an Archaeological Debate. [Manuscript intended for publication
provided by its author to Green for use in his current research].
Pawley, Andrew, 2007c. Why Do Polynesian Island Groups Have One Language and Melanesian
Island Groups Have Many? Patterns of Interaction and Diversification in the Austronesian
Colonization of Remote Oceania. Paper presented at the Workshop on Migration, Ile de
Porquerolles, France, 5-7 September 2007.
Ross, M.D. and Åshild Næss, 2007. An Oceanic origin for Äiwoo, the language of the Reef
Islands? Oceanic Linguistics, 46 (2): 456-98.
165
Ross, M. D., Andrew Pawley and Meredith Osmond, 2003. Introduction. In Malcolm Ross,
Andrew Pawley and Meredith Osmond (eds), The Lexicon of Proto Oceanic. The Culture
and Environment of Ancestral Oceanic Society: II. The Physical Environment. Pacific
Linguistics 545. Canberra: Department of Pacific Linguistics. Australian National
Sheppard, Peter and Richard Walter, 2006. A revised model of Solomon Islands cultural history.
Journal of the Polynesian Society, 115 (1): 47-76.
Wurm, S.A., 1969. The linguistic situation in the Reef and Santa Cruz Islands. In A. Capell, A.
Chowning and S.A. Wurm, Papers in Linguistics of Melanesia, No.2. Pacific Linguistics
Series A – Occasional Paper No 21. Canberra: Department of Linguistics, Australian
Wurm, S.A., 1970. Austronesian and the vocabulary of languages of the Reef and Santa
Cruz Islands – A preliminary approach. In S.A. Wurm and D.C. Laycock (eds), Pacific
Linguistic Studies in Honour of Arthur Capell. Pacific Linguistics C–13. Canberra,
Section VIII: Catalogue of Collections and Associated Databases

Numerous handwritten or non-electronic catalogues of all SESP collections have existed since the
1970s. In the past few years, efforts have been made to produce a full electronic catalogue for the
entire corpus of collections, covering all categories and sites. It has proven a slow process, partly
because our colleagues, Peter Sheppard and Geoffrey Irwin, decided it was an appropriate time
to digitally photograph the principal decorated sherds from the three early Lapita phase sites for
which they will now assume curatorial responsibility. Many images of those Lapita sherds are
already posted on the Department of Anthropology, University of Auckland website, and more are
destined to appear there as the cataloguing progresses to completion
[http://magic.lbr.auckland.ac.nz/anthpd]. A great number of the SESP collections are now housed
in a suitable storeroom and their contents are in the process of being re-catalogued in a digital
format (see Appendix 1).
Because the curation of the larger portion of SESP collections presents an ongoing problem,
the Director of the Solomon Islands National Museum prefers that they are not yet returned to
storage in Honiara for understandable reasons of security and insufficient suitable space in which
to house them. The political situation has improved in Honiara: a formal request has been made
to secure an updated statement of the official policy of those in the Solomon Islands Government,
its Department of Culture and Tourism and its National Museum on how they wish the matters of
the collections in temporary storage overseas to be handled in the immediate future. This deter-
mination will probably come from a newly appointed National Museum director who is scheduled
to take up that position in 2009.
Meanwhile, Chiu and Green have undertaken a separate project to make as much as possible
of the unpublished information available through a series of DVDs concerned with data from
three decorated Lapita sites on Outer Eastern Islands. These also should be available by the end
of 2008, or at least the first four in the series should be issued by then. Subsequently, others
scheduled in this series will be added as they are completed.
A protocol safeguarding this latter set of raw data compilations in DVD format will be set
out in an accompanying Note of Confidential Information when they are issued to scholars for
their own research purposes by –
166
Dr Scarlett Chiu Professor Roger C. Green

+886-2-26523129 (office) +886-2-26523175 (fax) +64-9-8177608 (home) +64-9-8172015
chius@gate.sinica.edu.tw pounamu@ihug.co.nz
Academia Sinica PO Box 60-054 Titirangi, Waitakere
Research Center for Humanities and Social Sciences Auckland 0642
#128, Section 2, Academia Road NEW ZEALAND
Nanking, Taipei, 11529
TAIWAN
Interim Note on the Confidential Nature of the contents on DVD disks

The information on the DVD disks will not be available as a formal publication for unrestricted
circulation: the format is intended to provide a means of sharing the largely unpublished informa-
tion within a community of scholars, each of whom will have received an identifiable DVD from
the providers, Scarlett Chiu and Roger C. Green.
Insofar as the DVD contents are concerned:
(a) the intellectual property rights remain with the providers
(b) the authorised recipients are provided with access to this data for the sole purpose of their
personal scholarly research: the use of the material is legally privileged and copying,
dissemination, or distribution of the contents is prohibited
(c) the authorised recipients will need written permission from one of the providers to cite or use
content in publications, reports, other publicly accessible documents, websites or
internet-based communications.
We expect serious scholars with good will to employ these DVD databases appropriately.
Polished publications they are not. But informative they most certainly are, for those researching
a topic related to Lapita. The intended contents of these DVDs are massive in file size and include
texts, images, and partial as-yet unpublished analyses. To further supplement this collection of
data and difficult to access “grey” literature, other documents of a similar nature are also destined
to appear on the web site of the University of Auckland Library [http://researchspace.auckland.
ac.nz] as they become available and with the agreement of their authors.
CONCLUSION
Public money, and the efforts of many Solomon Islands inhabitants, much of it voluntary (except
for the work crews paid by each researcher when doing fieldwork), in addition to the sustained
activities of a large number of research investigators, produced these SESP collections and their
accompanying textual and illustrative records and images. The only ethical position is that they
should be shared, although shared with due respect to a concept that they also constitute, often
in large part, the intellectual property of those who originally created them. They are, after all,
part of individual’s contributions to the archaeological research for this portion of the Pacific
region. Perhaps when this essay is published and the research data DVDs are released later this
year, they will help to overcome the “until death do us part” mentality not uncommon among
Pacific colleagues. Once a reasonable period of time has passed for the material to be published
in some form by the originators, useful scholarly access to field data and many kinds of unpub-
lished analyses carried out on the SESP collections, appears to us to be the most responsible
option available from the repertoire of choices. It is our hope that others among the colleagues
once involved in the Southeast Solomon Islands Culture History Project will also come to a
167
similar conclusion and allow their field data and unpublished analytical results to be accessible
to present day scholars who could use them in their own research on the SESP physical collec-
tions that will remain under curation. Meanwhile, we believe this overview article will serve as
a guide for interested parties in the field of Pacific cultural history who may require information
of the kind presented, compiled in an ordered way within a single source. Perhaps, from time to
time, other researchers will also find there is a need for similar project overviews, in addition to
articles focused on particular results and a conclusion to a piece of research undertaken by an
author or a set of authors.
APPENDIX 1
Current Location of Collections from the Southeast Solomon Islands
Cultural History Project
Department of Anthropology, University of Auckland
Within the central store space of the Archaeology Section, Department of Anthropology, one
room has been allocated to the storage of all the ecofacts and artefacts from the SESP other than
those collections listed below. They will remain there under the temporary custody for an
undefined period. A separate electronic catalogue is being developed to itemise the total collection
held in accessible storage. This will not only list the various separate storage boxes or similar bulk
containers, but also provide a general guide to their contents.
It has become necessary to revise the conditions of custodial care of the SESP excavation
collections, informally transferring the responsibility for managing and overseeing curation of
these collections from the authors to Peter Sheppard and Geoffrey Irwin of the Department of
Anthropology of the University of Auckland. This has been done with the agreement of the
Solomon Islands National Museum and its long time Director, Lawrence Foana‘ota. An official
request has been made through this museum to the Solomon Islands Government and its
Department of Culture and Tourism to confirm or implement any new conditions.
The initial arrangements were made through reciprocal internal emails and set out in a
memorandum of July-August 2006. Next, in July 2007, a grey literature document was circulated
to those currently carrying out research on these collections. Most of the collections will now be
held in safekeeping at the University of Auckland, under the supervision of Peter and Geoff, until
it proves feasible to return this government property to secure conditions in the Solomons, when it
can be expected that the Solomon Islands National Museum will undertake their ongoing curation.
The exceptions to those two locations are annotated in this document and in Appendix 1.
University of Otago, Dunedin, New Zealand
(a) School of Medicine, Department of Anatomy and Structural Biology.
This department currently holds the human skeletal remains from the Namu burial ground site on
the island of Taumako, in the Duff Group. The collection, formally lodged in this department with
the approval of the University of Otago Registrar in July 1977, was initially supervised and
curated by Dr Philip Houghton. On his retirement, the role and its responsibilities were trans-
ferred to Dr Hallie R. Buckley of that department.
In early 2008, the Solomon Islands National Museum made direct formal contact with the
Anatomy and Structural Biology Department, Otago Medical School, to request the return of this
collection of human skeletal remains from the Namu burial ground site in the Duff Group to the
168
National Museum in Honiara, as was agreed when permission for its removal and study was
originally granted. A plan to facilitate this transfer, under culturally appropriate protocols with
minimum disruption and strict security, is currently under negotiation between the Head of the
Anatomy and Structural Biology Department and the Director of the Solomon Islands Museum.
(b) Department of Anthropology
Portable artefacts from the Su‘ena site on the island of Ugi are retained and supervised by
Dr Richard Walter in the Department of Anthropology laboratories until the monograph (in prepa-
ration) involving them has been published. At that time, these artefacts will be returned to the
central store, Department of Anthropology, University of Auckland, where the catalogued shell
midden from this site is in storage. Thus, these two specified site collections will again be associ-
ated and form one part of the comprehensive Southeast Solomon Island Cultural History Project
collection now held in the Department of Anthropology store: they will remain in this location
until their return to secure storage facilities within the National Museum in the Solomons is
deemed appropriate and/or requested.
Museum of New Zealand/Te Papa Tongarewa
All collections (except the Namu burial ground human skeletal remains identified above)
recovered during the two periods of research carried out in the Duff Islands have for some time
been stored in the Museum of New Zealand for the convenience of Leach and Davidson until they
published their monograph on Taumako. These collections, which consist of both ecofacts and
portable artefacts, will now remain there. In due course, the National Museum of the Solomon
Islands will determine the time of their return to the museum in Honiara.
Bernice P. Bishop Museum, Honolulu, Hawaii
(a) Ethnobotanical Collections.
Information concerning the Ethnobotany collections, in large part the outcome of field research
by Dr Douglas E. Yen, is provided within Section IV of this document.
(b) Faunal remains from excavations
In general, the faunal remains from project excavations conducted by several former staff
members of the B. P. Bishop Museum have been retained at the museum. The following storage
containers, with their relevant faunal remains identified to a particular island or island group, are
known to form a part of these faunal assemblages:
T-Box 1352 = (Anuta); T-Box 1355 = (Santa Cruz); T-Box 1356 = (Santa Cruz); T-Box 3577 =
(Tikopia).
Dr Lisa Matisoo-Smith holds a more detailed record of the contents of each of these
numbered boxes: she compiled this record in April 2007 when she had brief access to the museum
storage facilities to search for faunal items for her DNA research.
In addition, the faunal remains from the excavations by McCoy and Cleghorn on the island
of Nendö are stored in the Bishop Museum collections.
A formal request for an inter-institutional loan of a selection of specimens from these faunal
collections was made by Matisoo-Smith and Melinda Allen to assist them in their research on the
comparable faunal collections already held and currently under study in the Department of
Anthropology, University of Auckland. This has now been implemented.
169
(c) Archaeologically recovered artefactual collections

The entire set of assemblages of portable artefact items recovered during excavations by Patrick
McCoy and Paul Cleghorn – in and around the east coast of Graciosa Bay, as well as along coastal
areas of Western Nendö and Te Motu Neo – are held in various storage facilities in the Museum’s
main collection of portable artefacts from the Pacific. They are available for study by other
scholars upon appropriate application to the Museum’s curator of collections.
Other such items from excavations by D.E. Yen, P.V. Kirch and others – from Tikopia and
Anuta – were returned to the Solomon Islands National Museum in Honiara following the publi-
cation of that research. One exception is a group of chert and obsidian artefacts (flakes and cores)
that were retained at the B.P. Bishop Museum at that time to allow McCoy to analyse them in
detail: this study was part of a comparative exercise focused on similar materials he had recovered
from the sites he and Cleghorn investigated on the island of Nendö. The other consists of the
fishbones from Tikopia, Anuta and Nendö that will be studied.
(d) Ethnological Collections
A set of recently manufactured traditional objects of material culture, which were collected and
fully documented by Roger Rose as part of the Southeast Solomon Islands Cultural History
Project, were shortly thereafter deposited in the Pacific portion of the Ethnological Collections of
the Bishop Museum. These items originally came from Nendö, Utupua and Vanikoro and remain
available for future research or comparative studies of a kind that is badly needed to set them
within a wider context.
Department of Anthropology, University of Hawaii – Manoa
Although most of the collections from the Spanish hilltop encampment at Pamua (San Cristobal),
and the site in Graciosa Bay (Nendö), have long since been returned to the Solomon Islands
National Museum, those from the flat land indigenous settlement on the Pamua Peninsula had
been kept in University of Hawaii off-campus storage supervised by the Department of
Anthropology. Recently, Jo Lynn Gunness, the Archaeological Laboratory Manager for that
department, reconfirmed their location and summarised their content. Plans are well advanced to
return this collection to the central store in the Department of Anthropology, University of
Auckland, where the other related collections from this region are held. This transfer will facili-
tate their long delayed study by several New Zealand and Australian researchers.
GENERAL REFERENCES CITED

Ambrose, W.R., 1996. Obsidian hydration dating of the Reef/Santa Cruz Lapita obsidians. In Janet
Anderson, A.J., 2003. Initial human dispersal in Remote Oceania: Pattern and explanation. In C. Sand (ed.),
Anson, D., R. Walter and R.C. Green, 2005. A Revised and Redated Event Phase Sequence for the
Reber-Rakival Lapita Site, Watom Island, East New Britain Province, Papua New Guinea.
University of Otago Studies in Prehistoric Anthropology No. 20. Dunedin: Department of
Anthropology, University of Otago.
170
Auckland.
Felgate, M.W., 2003. Reading Lapita in Near Oceania: Intertidal Shallow-Water
Pottery Scatters, Roviana Lagoon, New Georgia, Solomon Islands. Unpublished Ph.D. thesis,
University of Auckland, Auckland.
Foana‘ota, L.A., 1996. The development of archaeological work in the Solomon Islands. In Janet
Green, R.C., 1973. Southeast Solomon Islands Culture History Program. Royal Ontario Museum
Ethnography Monograph, 1: 1-6.
Green, R.C., 1986. Report 8 ‑ Lapita fishing: The evidence of the Site SE‑RF‑2 from the Main Reef
Islands, Santa Cruz Group, Solomons. In A. Anderson (ed.), Traditional Fishing in the Pacific.
Pacific Anthropological Records 37. Honolulu: Bernice P. Bishop Museum, pp. 119-35.
Green, R.C., 1997. Linguistic, biological and cultural origins of the initial inhabitants of Remote Oceania.
Green, R.C., 2003. The Lapita horizon and traditions – signature for one set of Oceanic migrations. In C.
Sand (ed.), Pacific Archaeology: Assessments and Prospects. Le Cahiers de l’Archéologie en Nou-
velle-Calédonie 15. Nouméa: Service de Musées et du Patrimoine de Nouvelle-Calédonie, pp. 95-120.
Survey. Royal Society of New Zealand Bulletin 11. Wellington.
Kirch, P.V. and P. Rosendahl, 1976. Early Anutan settlement and the position of Anuta in the prehistory of
the Southwest Pacific. In R.C. Green and M.M. Cresswell (eds), Southeast Solomon Islands Cultural
History: A Preliminary Survey. Bulletin 11. Wellington: Royal Society of New Zealand, pp. 225-44.
Lentfer, C. and R.C. Green, 2004. Phytoliths and evidence for banana cultivation at the Lapita Reber-Rakival
site on Watom Island, Papua New Guinea. In V.J. Attenbrow and R. Fullagar (eds), Pacific Odyssey:
Archaeology and Anthropology in the Western Pacific: Papers in Honour of Jim Specht. Records of
the Australian Museum, Supplement 29. Sydney: The Australian Museum, pp. 75-88.
Archaeology in Oceania, 28 (3): 121-37.
Sheppard, P.J., 1996. Hard rock: Archaeological implications of chert sourcing in Near and Remote
Oceania. In Janet Davidson, Geoffrey Irwin, Foss Leach, Andrew Pawley and Dorothy Brown (eds),
Oceanic Culture History: Essays in Honour of Roger Green. Dunedin: New Zealand Journal of
Archaeology Special Publication, pp. 99-115.
Summerhayes, G., J.R. Bird, R. Fullagar, C. Gosden, J. Specht and R. Torrence, 1998. Application of
PIXE-PIGME to archaeological analysis of changing patterns of obsidian use in West New Britain,
Papua New Guinea. In M.S. Shackley (ed.), Archaeological Obsidian Studies: Method and Theory.
New York: Plenum Press, pp. 129-58.
171
Swadling, P., 1976. The occupation sequence and settlement pattern on Santa Ana. In R.C. Green and M.M.
Swadling, P., 1986. Lapita shellfishing: Evidence from sites in the Reef Santa Cruz group, Southeast
Solomon Islands. In A. Anderson (ed.), Traditional Fishing in the Pacific. Pacific Anthropological
Records 37. Honolulu: Bernice P. Bishop Museum, pp. 137-48.
Yen, D.E., 1982. The Southeast Solomon Islands Cultural History Programme. Indo-Pacific Prehistory
Association Bulletin, 3: 52-66.
172
Chapter 8
Ethnobotany In The Southeast Solomon Islands

Cultural History Project
Douglas E. Yen
INTRODUCTION
Ethnobotany in the research programmes of the Southeast Solomon Island Culture History Project
(SESP) had two aspects. The first was descriptions of agricultural endpoints which indicate the
directions of past human-guided development and environmental differences among the island
landscapes – broadly defined as high volcanic, raised coral and coral atolls – that may correspond
with contrasting cultures. This aspect characterises the diversity of the terrestrial subsistence
systems found in the eastern Solomons, now and in the past. The second aspect is the more strictly
botanical study of the cultivated species that again may indicate regional differentials of adapta-
tion – with implications for prehistory of human migrants as carriers of the species themselves or
even as domesticators of new-found species. The material evidence for their history, of course,
was the establishment of standards of identification of plant remains that could be anticipated
from the archaeological excavations on the islands.
It was the customary delays in the islands’ shipping schedules to the eastern islands of the
Solomons that gave us the opportunity for brief explorations of the western islands as well,
particularly Kolombangara. This allowed us to draw some contrasts with the islands of the eastern
region, the major focus of the overall project, and of most of the projects involving archaeological
investigations. In the following discussion, some subsequent research programmes undertaken
further to the west (concerned with Island Melanesia and its prehistory) are also invoked,
especially those of the Lapita Homeland Project organised by Jim Allen (1984). In these compar-
isons the ethnobotanical connections of the New Guinea region with the research in the Solomons
reflects, at least in some measure, the importance of Lapita in its Oceanic distribution from the
Bismarck Archipelago eastward into Remote Oceania, and especially its downward leg into
southeastern Island Melanesia.
THE AGRICULTURAL ENVIRONMENTS

The comprehensive finds of the remains of earlier riverine irrigation systems on Kolombangara
in 1970 (together with informants’ claims of similar on Rendova) – evocative of the Polynesian
systems of Futuna, Nukuhiva and Hawaii – confirmed the earlier, unsubstantiated reports. They
also supplemented knowledge of similar archaeological features in the hinterland of neighbouring
northern New Georgia Island. Hviding and Bayliss-Smith (2000) later recorded more similarly
abandoned systems, not only on the southeast area of main New Georgia but also on an offshore
island. The antiquity of this extinct form of taro cultivation in the region is unknown. One carbon
date (I-6373 230 ± 90, 1 sigma 1630-1820 A.D. [50.3%], 1830-1890 A.D. [11.2%] 1920-1960
A.D. [6.7%] 2 sigma 1500-1590 A.D. [11.0%], 1610-1960 A.D. [84.4%]) (Black and Green 1975:
19), from a terrace wall of a tributary subsystem of a western river on Kolombangara, was some
indication of perhaps earlier age of the major systems constructed on the natural river terraces.
More recent investigations on the island have revealed evidence of antiquity in the form of
dentate-decorated sherds signifying a late phase in the Lapita Cultural Complex (Summerhayes
and Scales 2005).
173
There would be sufficient evidence from the New Georgia group to suggest a discrete
western region of prehistoric irrigation agriculture were it not for the Spanish chronicles of their
1568 A.D. discovery of the Solomons. That description of the practice inland on Guadalcanal is
the first known recording of water-control agriculture for the Pacific islands. David Roe (1993)
has confirmed archaeologically this early discovery in northwestern Guadalcanal, but to date, no
other example has been reported either from Guadalcanal or any of the other central Solomons,
for example the ethnographically well-studied Malaita. It transpired that none of our surveys on
Makira (San Cristobal) could produce evidence of irrigation practice – archaeological, traditional
or contemporaneous. Similarly in the southeast, Santa Cruz, with tracts of swamp, river and stream
flats seemingly adaptable for water-controlled cultivation produced the same negative result.
The discontinuities in the expression of agricultural intensification are palpable in the
eastern Solomons, with some obvious environmental correlates such as the high volcanic islands,
raised coral islands and low sandy atolls. A levelling effect is the periodically active Tinakula
Island volcano whose major ashfall deposition of the first millennium A.D. and chemical interac-
tion on the calcareous substrate of the neighbouring Reef Islands produced cultivable land as a
basis for a significant expansion of human settlement following that enriching tephra event. In
due course it allowed that group’s more recent inhabitants to grow the full range of crops that are
also produced by the inhabitants of the volcanic islands in the Santa Cruz Island Group.
Direct hydrological control is expressed as drainage of swampy areas on the high islands of
Utupua and Vanikoro, but even this form of modification to promote horticulture was not found on
the largest island, Nendö (Santa Cruz). Colocasia taro is grown now and, as claimed by cultivators,
in the past on selected low-lying tracts, often without firing regimes, with yams occupying appro-
priate dry land sites. Excavations on one inland complex of abandoned villages on the western
plateau of Nendö in the Santa Cruz Group (with the earliest date of only some 400 years B.P.)
suggested rather ephemeral occupation (Yen 1976), along the pattern of Southeast Asian shifting
cultivators. Furthermore, the building of a huge pig fence connecting two villages and bisecting an
earlier settlement indicated a change in pig husbandry from a less restrictive practice.
On the Buma Islet of Vanikoro, test excavations of a low-lying tract attested claims for former
taro cultivation some two generations ago; water channels tended to follow crooked, natural courses
(and thus non-rectangular planting plots), one of which yielded two buried fragments of wooden
digging sticks (promptly appropriated as “sacred” by the headman of Buma village). Contemporary
cultivation of Cyrtosperma taro on the swampy areas below the hilly forested areas was a feature
of Utupua Island subsistence, with Colocasia taro plots near the villages on the coastal low-lying
plain, with channels dug to drain runoff from the naturally watered Cyrtosperma plantings.
The digging of pits (to exploit the Ghyben-Herzberg fresh water lens) or the use of small
swampy areas for growing Cyrtosperma taro were found in Anuta and Tikopia, with indications
of such practice on the coral atolls, Nupani and Nukapu, in the past. In no case however, were we
to see the elaborate, formal pits (with woven planting “pots” and mulching with coconut leaves)
of the Gilbert Islands (Kiribati) or the rectangular raised drainage plots (again mulched) of
swampy land on East ‘Uvea or some of the islands in the Fijian archipelago.
Further intensification that is today exclusively Polynesian in this region is the pit-ensiling
of breadfruit in the manufacture of ma, or in this region, masi; and here, the technique extending
the season for such food is also applied not only to the root crops, but also to other tree fruit,
notably Burckella obovata. Yen (1973: 127-30) also lists bananas, and the root crops of Colocasia
taro, recently introduced manioc, Cyrtosperma and Alocasia aroids, and even the inner pith of the
sago tree among the other cultivars treated in this fashion. The only other traditional storage
174
Ethnobotany In The Southeast Solomon Islands Cultural History Project
technique was that of the Reef Islanders, where the limited production of fresh breadfruit was
extended by drying of sectioned cooked breadfruit (see also Kirch and Yen [1982: 43-45, 124,
333] for ethnographic description and archaeological evidence of pit ensilage on Tikopia
appearing circa 1200 A.D. in the Tuakamali Phase, and description of ethnographic practice on
Anuta [Yen 1973: 127-31]).
The arduous mulching of dry land plots, while still recognised by most subsistence farmers,
is practiced on Anuta (Yen 1973) and Tikopia (Kirch and Yen 1982). On these small volcanic
islands, fallow periods in crop rotation were shortened by this practice to the point where many
plots could be classed as permanent. Comprehensive excavations were conducted on a major
agricultural site on Tikopia – a valley and contiguous talus slope devoted to intensive cultivation
of manioc and taro. A series of pre-European 14C-datings indicated there had been two major
changes in agricultural development (Kirch and Yen 1982: Fig. 128), the early cessation of the
use of fire in cultivation practice and the expansion of the valley fields to cover the narrow marine
beach inlet evidenced by the sandy substrate with evidence of natural in situ sea grass remains.
The latter was accounted for by an erosion effect of continuous cultivation of the upper reaches
of the valley fields as well as from the surmounting fields on the old talus. A third subsistence
modification, of cultural significance in its decision, was the abandonment of pig husbandry,
evidenced in the excavations undertaken by Kirch (Kirch and Yen 1982).
THE PLANT SPECIES AND ARBORICULTURE

The taro/yam root crop dominance of garden plot agriculture in Oceania was reflected in the
eastern Solomons, as elsewhere, fading in favour of sweet potato when that crop was introduced
following European contact (Allen 2005), and shortly thereafter manioc also. Such was the case
even in Tikopia and Anuta, where, notwithstanding the importance of taro as the major tradi-
tional crop, manioc (easier to cultivate) was adopted as raw material for masi, probably in the 19th
century. The most conspicuous feature of village and garden landscapes in the project region for
the ethnobotanist, however, was the profusion of tree species beyond the customary coconut,
breadfruit, Pandanus, Terminalia and other less important food sources shared with all of
Oceania. These nut and fruit bearing species certainly extended the production season and dietary
variety and value, and had spatial as well as rotational implications of land use. Whether such
development should be called “intensification” may be debated, but the outcomes are easily
described and functionally productive in increasing the food supply.
For most of the SESP investigators questions of origin arose in respect to various cultivars
and their probable sources further west. The domestication of the fruit trees of the genera
Burckella and Pometia were to be assigned to the floristically rich region of northern New
Guinea, but in terms of variation and ethnographic enquiry, the conspicuous nut-producing
species of Barringtonia and Canarium were native to the Santa Cruz Group. The showy
Barringtonia found as specimen trees in villages of Kolombangara and north coastal Malaita were
said by those local informants to have “come from the east.”
Of the two species there is little doubt about the endemic status of Canarium harveyi, with
its external distribution limited to northern Vanuatu, the nearby Banks Islands with an easterly
extension to Fiji and as a rarely cultivated tree in Western Polynesia. This species, whose
variation (especially in seed characteristics) seems greater on the island of Vanikoro, has only
been grown as specimen trees in the central Solomons, with a single example in Honiara town
claimed to have been introduced by a Reef Islander. In this project, the 15th to 17th century A.D.
archaeological excavations of Kaschko at Pamua on Makira agreed with Yen’s observations of
175
absence of this species in the contemporary flora. Hence, the far more widely distributed C.
indicum to the west of Makira could be identified as dominant plant macro-fossils among the
remains of the flatland Pamua indigenous inhabitants during 15th and 16th century assemblages
before the 1595 A.D. Spanish pottery appears in the uppermost layers of each mound, as well as
on the surface of the site.
It should be noted here that these finds preceded the more detailed accounts of the
Melanesian tree crops (as plant descriptions by Walter and Sam 2002; in “agroforestry” produc-
tion systems by Clarke and Thaman 1993, and in Bayliss-Smith, Hviding and Whitmore 2003).
Thus at the time, arboriculture was not judged to be significant in indigenous subsistence systems
of Melanesia. Yet all had been described taxonomically, some even assigned a status of “culti-
vated” (i.e. domestication). Consequently, it was the study of C. harveyi in the SESP that was the
forerunner to the establishment of a series of seven distinctive domesticated species of Canarium.
With some overlaps (probably due to human intervention after local domestications), these have
their Oceanic centres in the Philippines, Indonesia, New Guinea, and the Solomons (Yen 1996).
An observed characteristic in all of the major cultivated trees was their variability in form, but
especially noticeable in the fruits and nuts – size, colour, shape. In the account that follows,
therefore, this genus assumes a perhaps disproportionate importance in terms of ethnobotany,
because it was, in subsequent investigations of the remains from no longer occupied habitation
localities in western Oceania, (especially those of the Lapita Homeland Project), to yield fossil
remains of Canarium nuts from the excavations of these sites.
The collection of live specimens of Canarium nuts from living trees allowed us to establish
criteria for the identification of often carbonised fragments of shell (pericarp) in archaeological
deposits. The three domesticated species from the project region, C. harveyi, from Santa Cruz and
Vanikoro, and C. salomonense and C. indicum from San Cristobal, all display distinguishing
features that could survive in macrofossil remains. The extension of this coverage to the Lapita
Homeland Project and thence to the New Guinea representatives of the species through other
investigations allowed us to see the dominance of C. indicum. This was especially the case in the
northern and western islands of the Bismarck Archipelago, with the aforementioned extension to
the Solomons and Vanuatu.
This dominance of C. indicum is reflected by all the archaeological materials excavated in
the Melanesian region up to 1990 (Yen 1991: 86), except for the Outer Eastern Islands of the
Solomons. There are also some indications of more recent but prehistoric exchange of Canarium
species that include C. harveyi during and after the period of Lapita settlement. One such late
occurrence is in a transitional stage following late Lapita assemblages dating to as late as 800
A.D. It is perhaps indicative of a still earlier probable transfer from or to the Reef/Santa Cruz
region of tentatively identified C. harveyi to or from the Reber-Rakival Lapita site on Watom
Island (Anson 2000: 104). Indeed, the possible pre-Lapita finds of this species on Nissan Island
(Spriggs 1991) indicates the probability of earlier, more than one-way contact between the eastern
islands of New Guinea and the Outer Eastern Islands of the Solomons. In the future, further
analyses of charred macro-fossil evidence, that could include Canarium and similar nuts, from
decorated and plainware phase Lapita sites on the Reef and Santa Cruz Island Group may
elucidate the status of endemism in the region, including Vanikoro, that I have deduced from the
detailed study of C. harveyi.
Graeme Ward’s excavations on the Banks Islands yielded both remains of C. harveyi and C.
indicum, though these specimens are not yet associated with archaeological contexts or dating.
Certainly they show prehistoric correspondence with present day cultivation of both species on
176
the nearby islands of Vanuatu. As elsewhere these locality specific present day ethnobotanical
observations, when joined with those of archaeological specimens from sites in same region,
confirm the prehistoric dispersal of both species. The antiquity of C. indicum as a food source is
attested by the archaeological finds of nut shell remains on the mainland of New Guinea
(according to this writer, the origin centre for this and other tree and crop species), especially in
sites on the Ramu River (Swadling, Araho and Ivuyo 1991) dated to c. 6000 years B.P. and the
Sepik (Gorecki, pers. comm.) in the late Pleistocene. Thus an appearance of the C. indicum at this
time depth (Green 2000: 376, Yen 1995: 843) in human food waste deposits of sites in the New
Guinea lowland region possess importance in support for claims of its early spread to the east, as
well as its establishment as far to the west as Indonesia.
A SUMMATION
This account of the ethnobotany programme that formed a part of the Southeast Solomon Island
Culture History Project and especially of the significance of the Canarium remains recovered by
archaeology within the Western Pacific may be viewed as another step in the enrichment of an
aspect of a generation of scholars researching the prehistory of Oceania in the mid-20th century.
Its genesis is firmly embodied in Jacques Barrau’s symposium (1963) title, “Plants and the
Migration of Pacific Peoples.” Still in that decade and the next there were hardly any plant
remains on which to build hypotheses – though build we did! Consequently this account does not
plumb the potential of finding macrofossils of many of the other tree crops that careful excavation
like that of Kirch (1989) on the Mussau Islands, and those on Watom (Lentfer and Green 2004)
can uncover. Nor does it engage with the study of phytoliths and starch and raphide residues now
under development as yet a further stage adding to such enquiry.
However, to indulge in a bit of “traditional” conjecture; assuming that the Lapita peoples
(and bearers of crop plants) were colonisers from several source islands to the west, the absence
of C. indicum in the Outer Eastern Islands of the Solomon Group (except for recent introduction
to Nendö Island), and its successful transfer to Vanuatu might simply have been a random
phenomenon. On the other hand, if the Lapita movement to the east was multilineal, we might
recognise the Canarium distribution as representing at least two of such routes, the one due east
to the Outer Eastern Islands of the Solomons and then veering southeast to Vanuatu, another
down the Solomon Islands chain more directly to the eastern Solomons. The latter may represent
failure of transport or adaptation, but it was the extreme variability of C. harveyi in the region
that favoured a hypothetical origin and domestication of the species there. We were able to
recognise three wild, small-fruited species of the genus (two being widely distributed in the
Solomons, the other endemic to Vanikoro) along with the cultigen in the high islands of the
group. Apart from the further speculation that the parentage of C. harveyi might have involved
an earlier dispersal of C. indicum, there was little we could add to the question of origin. The
co-existence of the latter two species in northern Vanuatu may indicate quite separate Lapita
transfers (see for example, Kirch 1989: 234, and the evidence from Tikopia [Kirch and Yen 1982:
306]). As important is the demonstration that C. harveyi, on the basis of modern distribution, was
dispersed to Fiji, Tonga and Samoa by the Lapita people from the region of the Outer Eastern
Islands which still awaits future archaeological attestation. Whatever proves to be the case, this
“plant migration” was not to be continued into Eastern Polynesia (until modern times). Instead it
fits pretty well with its Lapita-ware correlate, and botanically with St. John’s 1953 observation
(unpublished conference paper quoted by Barrau 1963: 2) of loss of cultigen species west to east
with a successive series of serial human migrations. In this sketch of “migration” scenarios based
on plant transfers, two unexplainable factors stand out – the complete absence of C. harveyi plus
177
only a hint so far of Lapita wares of the right date on Santa Ana Island off the eastern end of
Makira and a total non-association – so far – of the “famed” Lapita of New Caledonia with any
kind of Canarium whatsoever.
If the formulation of the plants-human migration in this region is problematic, that for
agricultural technology and its diffusion remains more so. At what we assume to be an ultimate
development in cultivation practices, namely irrigation terracing, Vanuatu and New Caledonia
both again provide unanswered comparative questions in their possession of complex systems of
water control in taro cultivation, forming a “southern Melanesian” grouping with the Fijian
Islands. One question for agricultural prehistory is whether the absence of irrigation is correlated
with the general dominance of Dioscorea yam, a practice that has not been maintained among
depleted indigenous populations, due to the adoption of the American sweet potato as the dry land
crop of choice.
Finally, the information assembled to date leads to the consideration of the Outer Eastern
Islands of the Solomons as a region for domestication of species of Canarium harveyi. Based on
herbarium specimens, Leenhouts (1959) proposed systematising C. harveyi into three subspecies,
but, as Walter and Sam (2002) have indicated, such sub-division may not be warranted. Field
observations during our project of whole cultivated populations indicated that these distinctions
are largely quantitative, just as the cultivated species Barringtonia edulis and B. procera were
inseparable. When similar variation in the fruit trees Pometia and Burckella are considered, it
must be concluded that the selection methods pursued by indigenous cultivators preserve variabil-
ity that is expressed in formal genetic terms, as continuous.
One cultural distinction consistent with such variability is that there is no varietal naming,
cf. the contrast with varieties of root crops; rather, each tree is attributed to an owner, or may be
described for its exceptional qualities. Arguments will continue (depending on the region in which
an investigator is working) on whether the suite of cultivated trees originated in New Guinea,
Vanuatu or the Santa Cruz Islands, as field studies indicate wide variation throughout the whole
region of Melanesia. New solutions may well lie in biomolecular analysis exemplified by the
survey of Oceanic field crops initiated by Lebot (1999).
ACKNOWLEDGEMENTS
I owe much to Roger and Valerie Green and her sister Beverley Lang for their valued enhance-
ment of this essay. Further, I would like to use this opportunity to acknowledge with gratitude
Roger’s part in the forming of my ethnobotanical career, of which the Southeast Solomons
Culture History Project has been a major component; his stimulating friendship is a privilege that
I share with many.
REFERENCES
Allen, J., 1984. In search of the Lapita homeland: Reconstructing the prehistory of the Bismarck
Archipelago. Journal of Pacific History, 19 (4): 186-201.
Allen, M.G., 2005. The evidence for sweet potato in Island Melanesia. In C. Ballard, P. Brown, R.M.
Bourke and T. Harwood (eds), The Sweet Potato in Oceania: A Reappraisal. Oceania Monograph
56. Sydney: University of Sydney, pp. 99-108.
Anson, D., 2000. Excavations at Vunavaung (SDI), Rakival Village, Watom Island, Papua New Guinea.
Barrau, J. (ed.), 1963. Plants and the Migrations of Pacific Peoples. Honolulu: Bishop Museum Press.
178
Bayliss-Smith, T., E. Hviding and T. Whitmore, 2003. Rainforest composition and histories of human
disturbance in the Solomon Islands. Ambio, 32: 346-52.
Black, S. and R.C. Green, 1975. Radiocarbon Dates from the British Solomon Islands to December 1973.
Working Papers in Anthropology, Archaeology, Linguistics and Maori Studies 39. Auckland:
Clarke, W.C. and R.R. Thaman (eds), 1993. Agroforestry in the Pacific Islands. Tokyo: United Nations
University Press.
Green, R.C., 2000. Lapita and the cultural model for intrusion, integration and innovation. In A. Anderson
and T. Murray (eds), Australian Archaeologist. Collected Papers in Honour of Jim Allen. Canberra:
Coombs Academic Publishing, Australian National University, pp. 372-92.
Hviding, E. and T. Bayliss-Smith, 2000. Islands of Rainforest: Agroforestry, Logging and Eco-Tourism in
the Solomon Islands. Aldershot: Ashgate.
Kirch, P.V., 1989. Second millennium BC arboriculture in Melanesia: Archaeological evidence from the
Mussau Islands. Economic Botany, 43 (2): 225-40.
Kirch, P.V. and D.E. Yen, 1982. Tikopia: The Prehistory and Ecology of a Polynesian Outlier. Bernice P.
Lebot, V., 1999. Biomolecular evidence for plant domestication in Sahul. Genetic Resources and Crop
Evolution, 46: 619-28.
Leenhouts, P.W., 1959. A Monograph of the Genus Canarium. Leiden: Ijdo N.V.
Lentfer, C. and R.C. Green, 2004. Phytoliths and evidence for banana cultivation at the Lapita Reber-Rakival
site on Watom Island, Papua New Guinea. In V.J. Attenbrow and R. Fullagar (eds), Pacific Odyssey:
Archaeology and Anthropology in the Western Pacific: Papers in Honour of Jim Specht. Records of the
Australian Museum, Supplement 29. Sydney: The Australian Museum, pp. 75-88.
Spriggs. M., 1991. Lapita origins, distribution, contemporaries and successors revisited. In P. Bellwood
(ed.), Indo-Pacific Prehistory 1990: Proceedings of the 14th Congress of the Indo-Pacific Prehistory
Association. Canberra: Indo-Pacific Prehistory Association, pp. 306-12.
Swadling, P., N. Araho and B. Ivuyo, 1991. Settlements associated with the Sepik-Ramu Sea. In P.
Bellwood (ed.), Indo-Pacific Prehistory Association 1990: Proceedings of the 14th Congress of the
Indo-Pacific Prehistory Association. Canberra: Indo-Pacific Prehistory Association, pp. 92-112.
Walter, A. and C. Sam, 2002. Fruits of Oceania. Canberra: ACIAR.
Yen, D.E., 1973. Agriculture in Anutan subsistence. In D.E. Yen and J. Gordon (eds), Anuta: A Polynesian
Outlier in the Solomon Islands. Pacific Anthropology Records 21. Honolulu: Bernice P. Bishop
Museum, pp. 113-48.
Yen, D.E., 1976. Inland settlement on Santa Cruz Island (Nendö). In R.C. Green and M.M. Cresswell (eds),
Southeast Solomon Islands Culture History: A Preliminary Survey. Bulletin 11. Wellington: Royal
Yen, D.E., 1991. Polynesian cultigens and cultivars. In P.A. Cox and S.A. Banack (eds), Islands, Plants and
Polynesians. Portland: Dioscorides Press, pp. 67-95.
Yen, D.E., 1995. The development of Sahul agriculture with Australia as a bystander. Antiquity, 69 [265]:
831-47.
Yen, D.E., 1996. Melanesian arboriculture: Historical perspectives with emphasis on the genus Canarium.
In M.L. Stevens, R.M. Bourke and B.R. Evans (eds), South Pacific Indigenous Nuts. Canberra:
ACIAR, pp. 36-44.
179
Chapter 9
POST-LAPITA DEVELOPMENTS IN THE REEF-SANTA CRUZ

ISLANDS, SOUTHEAST SOLOMON ISLANDS
Moira Doherty
The Reef-Santa Cruz Islands lie c. 350 km southeast of the main Solomon Island chain (see
Figure 9.1). While the earliest Lapita occupation of these islands has been studied and published
extensively, comparatively little work has focused on later periods. These islands provide a useful
case study for evaluating the hypothesis that multiple migrations have influenced culture history
in Island Melanesia as linguists have debated whether the languages spoken on Santa Cruz and
the Main Reef Islands should be classified as Austronesian (AN) or non-Austronesian (NAN)
languages (Dunn, Reesink and Terrill 2002, Lincoln 1978, Tryon and Hackman 1983, Wurm
1969, 1970, 1975, 1976, 1978, 1982, 1986, 1987).
METHOD AND DATASETS

The archaeological datasets examined were excavated in 1979 by Green, from two sites on the
Main Reef Islands – SE-RF-19 (Ngatoponu) and SE-RF-3 (Sie). Ceramics from Mdailu (SE-SZ-
33) and Növlaö (SE-SZ-47) provided by McCoy (McCoy and Cleghorn 1988) were also
examined. McCoy also provided much unpublished information related to the fieldwork
summarised in McCoy and Cleghorn (1988). Ethnographies of the islands, particularly the
numerous publications of William Davenport (1962a, 1962b, 1964a, 1964b, 1965, 1968a, 1968b,
1969, 1972, 1975a, 1975b, 1990, 2005), and historical narratives of contact between indigenes
and visiting Europeans provided valuable information relating to the historic period. These
European encounters began with the failed attempt to colonise Santa Cruz by the Spanish in 1595
A.D. (Markham 1967 [1904]).
Using the direct historic approach, Doherty (2007) evaluated the narrative and archaeologi-
cal evidence for coherence or disjuncture in multiple traditions (the built environment, faunal
remains, portable artefacts and ceramics), and compared the archaeological evidence with the
prevailing interpretation of history derived from linguistics and human biology. That work is
summarised below.
THE TWO NEW SITES: SE-RF-19 and SE-RF-3

The two sites that are the focus of the Ph.D. study – SE-RF-3 and SE-RF-19 – are both located
on the eastern end of Lomlom (Ngambelipa), which is the largest of the uplifted coral reef islands,
at c. 30 km2 (see Figure 9.2).
SE-RF-19 Ngatoponu (“Some Turtle”) Mound Site

Site details are derived from Green’s field notebook (1979 ms: 17-41), and site record form of
16.12.77. Ngatoponu is an open mound site, located approximately 1 km inland from the coast.
The mound is c. 44 m long, 28 m wide, and up to 1.65 m high. A parallel, lower mound, c. 25 m
to the northwest (Mound B), also contained pottery (see Doherty 2007: Figs 2.3-2.7 for maps, and
site drawings).
181
182
Figure 9.1 Location of Reef-Santa Cruz Islands, Outer Eastern Islands District, Solomon Islands.
Post-Lapita Developments in the Reef-Santa Cruz Islands, Southeast Solomon Islands
Figure 9.2 Main and Outer Reef Islands.
SE-RF-19 Excavation, Stratigraphy and Features

The whole mound was gridded in 3 m squares, and a surface collection made of pottery from this
site, through the gully leading to Mound B, and the opposite side of Mound B. On Ngatoponu
mound, three areas 2 m x 2 m were excavated. Each stratigraphic layer was excavated in 10 cm
units, and a maximum depth was reached of just over 1 m in square VV-WW 48-49. The mound
was built up of rubbish, largely ovenstones, in which pottery, bone and some shellfish had been
incorporated. The stratigraphy was simple, consisting of two cultural layers, the upper a darker
brown soil, and the second a lighter brown matrix. The base of the site consisted of an almost
sterile light reddish-brown soil of the type formed on coral limestone.
Only in square SS-TT 48-49 was the top of the third layer excavated because the basal layer
here appeared slightly less red-brown than the base in other squares. A further 10 cm (top of Layer
3), was excavated to test for features. None was found, and the deposit was described as almost
sterile (Green 1979 ms: 36-37). Bone preservation at the site was considerably better than shell,
which was severely affected by weathering. All categories of calcareous materials – shell temper
in pottery, shell midden, and shell artefacts – had deteriorated.
Ngatoponu was not a habitation mound, and contained only three features – a large oven pit
filled with stones, a pit c. 40 cm deep, and a possible post-hole.
183
SE-RF-19 Dating
The site contained no charcoal-bearing features suitable for radiocarbon dating, and although it
was intended that three radiocarbon dates on marine shell would be obtained, because of the
extent of remineralisation that had taken place, only two radiocarbon dates could be processed,
one from each of Layers 1 and 2, in square VV-WW 48-49. Wk-7852 is an AMS date. The dated
shells were not associated with a particular feature. A marine correction factor (Delta R) of
-81±64, calculated previously for SE-RF-2 (Jones et al. 2007) was applied in the calibration. Date
ranges for the two levels overlap, and suggest use of the site sometime in the interval c. 900 B.C.−
600 B.C. (68% CI) (see Table 9.1).
Table 9.1 Radiocarbon determinations for SE-RF-19.
Soil Samples from SE-RF-19

Soil samples were taken from square PP-QQ 48-49, top of Layer 1 and base of Layer 3, and from
square WW-49 Layer 1 at 15-25 cm deep, and Layer 2 at 60-70 cm deep, and were examined by
P. Shane (Geology Department, University of Auckland). The soils were derived from weathered
tephra which “appeared to be of basic to intermediate in composition, i.e., andesite or basalt”
(Shane pers. comm. 2000).
SE-RF-3 Sie Village Site

All details are derived from Green’s field notebook (1979 ms: 42-49) and site record form of 9.7.70.
The site was one of numerous mounds within the old Sie village. The village was located approx.
200 m beyond Gnimoa, and was surrounded by a coral block wall, c. 2.2 m high, and 60-70 cm wide
at the base. The mounds inside the wall were 1-2 m high and 10-15 m in size, with midden visible
on the surface. Coral blocks, ovenstones, shell and charcoal debris were scattered all over the area.
Excavation, Stratigraphy and Features
The mound was excavated by Green and a team of local workmen over a period of 6 days, from
15-20 January, 1979. The surface of the site was cleared, and an area of 9 m2 (3 m x 3 m grid) was
excavated in the centre of the mound (see Doherty 2007: 93, Fig. 2.8). Green’s excavation notes
(1979 ms.) recorded that the mound was “minutely stratified with no major layers”, but incorporated
many local ash lenses, all oriented horizontally (see Doherty 2007: Fig. 2.9 and Plates 2.5 – 2.6). A
large white ash lens, 10 cm thick, extended through squares O-25 and O-26, and half of squares O
and P-27. A subtle change of colour was observed in the NE corner of N-27 (between Levels 4 and
8), which suggested “a major change and earlier mound in the corner” (Green 1979 ms: 45). Given
the difficult excavation conditions, pressure of time, and inexperience of workmen, the mound was
excavated in 10 cm. levels, “paralleling strata in large degree, but not stratigraphically” (Green 1979
184
ms: 45). This strategy was sufficient to acquire the specific data sought to describe the late sequence,
namely shellfish, change in cultural content over time, and use-span of site. Because of the copious
quantity of shell in the mound, samples were collected from all nine squares at Levels 1, 4, 7, and
10. A reference collection of six seagrass and mangrove shells was made, and these shells were
counted for each sampled level and then discarded in the field. There was abundant charcoal in the
mound, and samples were taken from squares at Level 12 (110-120 cm depth).
Excavation proceeded to a depth of 1.20 m, but did not continue to the base of the mound
in all squares. Test excavations were made in the NE corner (square N-27) to a depth of 2.6 m
below datum, and in the SW corner (P-25) to a depth of 2.65 m below datum, to reach the brown
layer base, and soil samples were taken at these locations. This dark coloured midden occupation
extended for a further 25-30 cm. The site contained only one feature, a possible coral pavement
in Level 10, square N-25, and parts of O-25 and N-26 (see Doherty 2007: 95, Fig. 2.10).
Table 9.2 Radiocarbon determinations for SE-RF-3.
185
Soil Samples from SE-RF-3

Soil samples were taken from the brown base layer in two locations – in the southwest corner,
square P-25 at 2.65 m below datum, and in the northeast corner, square N-27 at 2.6 m below
datum. These were examined by Dr. P. Shane. The report for these soils is the same as that for
SE-RF-19 samples. All exhibit a similar degree of clay weathering, and are derived from
weathered tephra, probably of andesitic or basaltic lithology (Shane pers. comm. 2000). The
ashfall has therefore been incorporated into SE-RF-19, and forms the base of SE-RF-3.
Dating
Six radiocarbon dates have been estimated for this site, three on marine shell and three on wood
charcoal (see Table 9.2). The samples were taken from within the excavated layers, so do not date
the beginning of mound accumulation. A marine reservoir correction factor of -110+/-85 was calcu-
lated by M.D. Jones (2002 ms). A Bayesian analysis of the five radiocarbon dates from layers 10 and
12 indicates an activity period spanning up to 510 years (0.938%), starting some time in the interval
A.D.1310-1650 (Jones 2002 ms). The possibility that these lower levels pre-date European contact
is reinforced by the absence of European artefacts, which appear only in the uppermost levels.
SEQUENCE SUMMARY
The following schemata for discussing sites are derived from published and unpublished dates and
excavation reports, and are used throughout the thesis (Doherty 2007), and this summary report
(see Table 9.3).
The data sources are as follows: Yen 1976a (Naiavila), McCoy and Cleghorn 1988, and
unpublished information provided by Pat McCoy (Dai, Mateone, Növlaö, Mdailu); Kirch n.d.
(Napo); Green (1976, 1979 ms., 1986, 1987) for the three Lapita sites (Nenumbo, Nanngu, and
Ngamanie), with reworking of radiocarbon dating by Jones (Jones and Green [in prep.], Jones et
al. [2007], Green, Jones and Sheppard 2008).
Table 9.3 Reef-Santa Cruz Sites and Ceramic Sequence.
186
THE BUILT ENVIRONMENT

Davenport described four main structures in traditional villages on the Main and Outer Reef
Islands, and Santa Cruz: the men’s house(s), family dwellings, the cult house and a circular dance
ground (Davenport 1964a: 66, 1968b: 149, 1969: 170, 1972: 75). Archaeological and narrative
evidence for these structures is reviewed in Doherty 2007: 132-98. Changes over time in the
shape of dwellings, location of burials and ovens, and the increasingly monumental character of
construction are discussed below.
Shape of Dwellings
Dwelling shapes on Santa Cruz change over time. In the historic period both round and rectan-
gular houses coexisted on parts of Santa Cruz, with a more tenuous possibility that oval-ended
houses were built inland. Round houses were observed by the Spanish at Graciosa Bay, and were
described by other observers (e.g., O’Ferrall 1908, and Beattie ms. 1906 for Tömotu Neo, plus
photographs). Members of the Melanesian Mission regarded the form as confined to the Te Motu
area (i.e., residents of Tömotu Neo and their neighbours in Graciosa Bay). The round form disap-
peared quite quickly. Speiser resided briefly in Graciosa Bay in 1912, and reported that the
houses were square, with only a few circular ones at that time (1913: 284). While rectangular
house forms predominate in the Group I sites, three round houses were identified in Dai village
(McCoy and Cleghorn 1988).
For Santa Cruz, we have no information about structures for the period from c. A.D. 100 –
A.D. 1400. Coral foundations interpreted as dwelling houses were recovered from the Mdailu site
(associated with plainware levels c. 2600 B.P.) and Növlaö (in the late ceramic period c. 2000 B.P.)
(McCoy and Cleghorn 1988). The earlier interpretation that these were the foundations of round
houses has been re-evaluated (McCoy pers comm. to Green 2002). In my view, too little of the
foundations have been exposed at Mdailu to reconstruct the complete structures with any accuracy,
but the alignments at Növlaö appear to be more elliptical or oval-shaped than round.
On the Main Reef Islands the traditional dwelling was a rectangular structure, the interior of
which was divided into male and female sides, with the female side containing the earth oven and
cooking equipment (Davenport 1969: 170). Here, the information gap is even greater and there is
no data about structures which bridges the time between the rectangular Nenumbo (SE-RF-2)
residential unit of c. 3000 B.P. and the rectangular dwellings of the historic period. The main
structure at Nenumbo (SE-RF-2) was interpreted as a rectangular dwelling occupied by a single
household (Green and Pawley 1999). For descriptions of structural elements of the site and
artefactual associations, see Sheppard and Green (1991). The Nenumbo house is much larger than
any later dwellings, which may reflect a reduction in the size of the family unit occupying houses,
a change in activities within the dwelling, or a reduction in the number of people sleeping there.
There are as yet no archaeological examples from the Main Reef Islands of elliptical or oval-
shaped houses, and it is unclear from the historical narratives whether round houses were
constructed there. The possibility arises therefore, that the two sequences may not have changed
in similar directions, although a linking feature is that earth ovens are also located inside the late
houses on the Reefs (see below).
Technological considerations do not provide an explanation for the change to round house
forms, as they are not easier to build than rectangular houses, and the same methods for increasing
the roof span (using central posts, trusses or space frames) apply to both shapes (Rappoport 1969:
116-22). Round houses are easier to roof, and rectangular houses are easier to orient (to signifi-
cant compass points or landmarks) (Rappoport 1969: 25, 41). If the Cruzian innovation to a round
187
form is symbolically motivated, then its builders diverged in this respect for the wider shared
symbol system of duka posts, ghost houses and dance circles that linked Santa Cruz, the Main and
Outer Reefs. Whatever the inspiration for the round form, the organisational principles governing
the relationship between dwellings and other buildings and the objective of separating the genders
was also shared with the Main Reef Islands. The interiors of homes seem also to have been
similar, containing an altar and storage platform.
Location of Earth Ovens

A feature of Cruzian and Main Reefs houses in Davenport’s descriptions of traditional villages is
a cooking oven/hearth inside the house, although Davenport provides no details about the size or
form these took. On the evidence of Naiavila, and some of the narratives, communal ovens
located outside dwellings were also used at this time. However, there is no archaeological
evidence from ceramic sites for earth ovens located within the dwelling. At both Mdailu and
Növlaö, clusters of coral cobble-lined ovens were located separately from houses, and were inter-
preted as evidence of communal cooking (McCoy and Cleghorn 1988: 107). At the Lapita site of
Nenumbo (SE-RF-2), a number of small pits located to the area south of the main structure were
interpreted as ovens (Sheppard and Green 1991: 95).
Burials
Although the archaeological data relating to burials in the Reef-Santa Cruz case is sparse and highly
variable, there does appear to be a change over time in the location of burials. At Növlaö, burials
were found in both the mid-late and late ceramic periods, but in both cases were separate from
houses (McCoy and Cleghorn 1988). The traditional repertoire of mortuary practices on both Santa
Cruz and the Reef Islands included burying a deceased person (sometimes described as a deceased
man) within the dwelling (Codrington 1972 [1891]: 262, Coombe 1911: 190, Girieud and
Herrenschmidt 1898: 107, O’Ferrall 1908: 13). Subsequently the bones were disinterred, the skull
was retained, and some of the bones were made into arrowheads (O’Ferrall 1908: 13). Archaeological
confirmation of burial within the house comes from examples in Dai village, associated with both
rectangular and round houses, although this was not the sole burial location, and the treatment of
the body and burial positions were also variable (McCoy and Cleghorn 1988: 111-12).
Increasingly Monumental Construction

A particularly noticeable change is the increasingly monumental character of construction over
time. In the Group I sites, coral cobble terraces and house platforms, pavements and massive
walls are highly visible features of villages. Coral pavements and house foundations are not
features of the Reef-Santa Cruz Lapita sites of the early dentate decorated phase. Coral house
foundations/outlines first appear at Mdailu and Növlaö in plainware contexts. The Spanish
reported walls around houses (Markham 1967 [1904]: 52), but later descriptions of walls around
sections of villages or the entire village, and bordering paths, suggest such constructions became
more extensive in the historic period. Archaeological and historical sources suggest that the
monumental form of the dance circle is a very recent innovation, perhaps from the reed-enclosed
courtyard noted by the Spanish (see Doherty [2007: 161-63] for details).
Coherence in Architectural Traditions

Doherty (2007) proposed that symbolic coherence in architectural traditions might be assessed in
relation to the concept of “House societies”, which some scholars have found useful for describ-
ing ethnographic Austronesian-speaking communities of Island Southeast Asia (e.g., contributors
188
to Carsten and Hugh-Jones (eds) 1995, Fox 1993, Waterson 1990). Green has argued that the
House society may represent a very enduring social structure, its beginnings perhaps dating back
4000-4500 years ago to Malayo-Polynesian speakers (Green 2002: 25). Green and Kirch have
developed the proposition that the House society provides a framework for understanding the
social organisation of Lapita societies (Green 2002, Green and Pawley 1999, Kirch 1997: 143-44,
188-191), Ancestral Polynesian society (Kirch and Green 2001: 201-18), and their Polynesian
descendants (Green 1998, Kirch 1996, 2000).
The House society is not an unproblematic concept, and in my view the archaeological case
for Lapita as a kind of House society is considerably weaker than arguments proceeding from
historical linguistics or comparative ethnology. However, if we accept that the Lapita colonists in
the Reef-Santa Cruz Islands organised themselves as House societies of some kind, then the
question arises as to whether this social structure persisted over the long term. Davenport’s
fieldwork was conducted before House societies became such a topic of interest, so that his work
is not framed in those terms, but he provided sufficient information to answer the question – the
communities on Santa Cruz and the Main Reefs do not have this kind of social organisation
(Davenport 1964a, 1969).
Summary of Evidence Relating to Architectural Traditions

At no point in the sequence was it possible to identify two co-existing but unrelated architectural
traditions. At wider spatial and temporal scales, locally idiosyncratic features such as the change
to oval-shaped or round houses appear less incongruous and do not require recourse to external
explanations. In the regional context, there are cross-cutting features that are shared differentially.
If the original colonisers were organised as an early form of House society, it was not possible to
identify when in the sequence this form of organisation changed.
SUBSISTENCE SYSTEMS
Faunal remains from SE-RF-3 and SE-RF-19 comprise the datasets. External comparisons were supple-
mented by unpublished notes on bone identification supplied by McCoy for Mdailu and Növlaö.
Rats
Both SE-RF-3 and SE-RF-19 contained small and medium rat bones, but large bones were few
(N=9) and were recovered only from SE-RF-3. The rat bone assemblages were examined by Peter
White (University of Sydney) and Ken Aplin (Curator of Mammals, Western Australian Museum),
who reported on 26 mandibles and maxillae which they measured and compared to Australian
Museum material. In their assessment, both sites contain R. exulans and probably R. praetor.
Most of the medium bones were likely to be R. praetor, and the large bones from SE-RF-3 also
fitted comparative material for R. praetor, the size difference perhaps being accounted for by age.
No European rat was identified in either assemblage. One medium mandible from SE-RF-19 had
much narrower incisors, and was not securely identified to species.
Specimens from both sites were submitted for mtDNA analysis, but sequences could not be
established for bones from SE-RF-19. However, the mtDNA sequence from one small SE-RF-3 rat
proved to be R. exulans Haplogroup II, which is the easternmost limit of distribution for this haplo-
group. At the time of thesis submission, the geographic distributions of Haplogroups II and III were
suggestive of different dispersal histories, and it was postulated that Haplogroup III rats were “an
intrusive component of Lapita” (Matisoo-Smith and Robins 2004: 9171). Finding Haplogroup II
rats on the Reef Islands is therefore unexpected. The principal of competitive exclusion apparently
189
precludes a successful second introduction when one population is already established, so while it
would be tempting to see the Haplogroup II lineage as indicative of a subsequent movement of
people from somewhere amongst the large islands of Near Oceania, there is presently no convinc-
ing evidence of an event sufficiently catastrophic to entirely eliminate an earlier rat population in
the Reef Islands. Further work on R. exulans distributions may clarify this anomaly. Attempts to
extract mtDNA from R. exulans specimens from decorated Lapita and plainware sites have been
unsuccessful, so we do not know the genealogical relationships between earlier and later rats.
One of the “large” bones from SE-RF-3 was also subjected to mtDNA analysis. At the time
it was not possible, with the reference samples available, to distinguish between R. praetor and
R. mordax using mtDNA (Robins pers. comm. 2006). The sample submitted fell within the R.
praetor/R. mordax clade. R. praetor is present in Main Reefs Lapita sites, but insufficient work
has been done on the lineages of this species to make any inferences.
Pig
Pig bones were excavated only from SE-RF-3, where they were well distributed throughout the
site. The bones were identified by Alan Ziegler (zoological consultant, Hawaii), and examined by
Ian Smith (University of Otago). Smith (pers. comm. 2001) judged that the total assemblages
“appears to reflect deliberate pig rearing for food production, and may also provide evidence of
ritual behaviour associated with this species”. A minimum number of 9 pigs were represented, of
which only one was older than two years, and one was aged between 1-3 months.
Although absent from SE-RF-19, small quantities of pig bone are attested in Lapita sites on
the Main Reefs (SE-RF-6 and SE-RF-2) (Green 1976: 255). Pig bone is also represented in Group
I and ceramic sites (III and IV) on Santa Cruz. In summary, although no genetic data can be brought
to bear on the question, the retention of POC terms for pig (see Lynch 1991), and the archaeologi-
cal demonstration of pigs in decorated and plainware sites, and in sites of the late period, all suggest
that pigs are a continuous feature of foodways on Santa Cruz and the Main Reefs.
Dogs
No dog bone was recovered from SE-RF-3, but a variety of elements (N=30) were recovered from
SE-RF-19, most from Layer I. In addition, another 33 bones in Layer I were “very likely” Canis
familiaris (Ziegler pers. comm. 2001). Attempts to extract mtDNA from several of the dog teeth
were unsuccessful.
Dog bones have not so far been excavated from decorated Lapita contexts on either the Main
Reefs or Santa Cruz, but two bones have been identified from a second plainware context, in the
Mdailu site (McCoy pers. comm. 2003). The absence from sites with decorated Lapita but
presence in plainware contexts raises the possibility that dog was not successfully introduced with
early Lapita colonists, but may represent a later introduction during the plainware stage.
Bats
Large Pteropodid bones were common in SE-RF-19 (N=93) and SE-RF-3 (N=128), and are found
in all levels in both sites. From published and unpublished data it is apparent that bats are a
continuous feature of sites on the Main Reef Islands and Santa Cruz.
Other Vertebrates
See Doherty (2007: 443 and 444) for tabulation of miscellaneous vertebrates.
190
Fish Bone
In the assemblage from SE-RF-19, a total of 1199 cranial and special bones were identified to 15
families (with Sparidae included in Lethrinidae) and one class. (At least 3-4 species of Lethrinidae
were represented, including Monotaxis; 3 species of Scaridae, and at least 2 Labridae). In
addition, another 20 cranial elements could not be identified. Complete listings and weights are
given in Doherty (2007: 430-32 and Tables 8.4-8.7). This site has all but four of the families
represented in SE-RF-2 (Green 1986) – not represented are Scorpaenidae, Belonidae, Sphyraenidae
and Pempheridae, which were only a minor component of the assemblage from Nenumbo,
together comprising less than 2.5% of the total. SE-RF-19 includes one taxa not found at SE-RF-2,
Fistulariidae, represented by bone fragments identified by Ziegler.
The top ranked taxa using the paired cranial bones only are Scaridae (10), Lethrinidae (2),
Serranidae (3), Diodontidae (4), and Carangidae (5). These rankings are broadly consistent with
the rank orders for SE-RF-2, the major differences being in the greater abundance of Carangidae
in SE-RF-19, and the relative unimportance in this site of Labridae.
The fishbone assemblage from SE-RF-3 was numerically smaller, with 362 identified bones.
Four families represented in SE-RF-19 are absent from SE-RF-3, namely Diondontidae,
Fistulariidae, Mullidae and Scombridae. The last three are minimally present in SE-RF-19, so
their absence from a smaller assemblage is unsurprising (Grayson 1984). The complete absence
of Diodontidae is noteworthy, and is unlikely to be an artefact of recovery method or preservation.
This phenomenon, in which Diodon hystrix was represented archaeologically, but not in recent
sites, was also observed on Tikopia (Kirch and Yen 1982: 292). The authors speculated that on
Tikopia the tapu on Diodon may have followed deaths from poisoning.
The SE-RF-3 assemblage also lacks the four families present in SE-RF-2 but absent from
SE-RF-19. One family – Kyphosidae – is tentatively identified in this site but is not represented
in either of the others. The rank order for cranial elements is Scaridae (1), Lethrinidae (2),
Serranidae (3), and Lutjanidae (4), with the rest contributing small percentages only.
In summary, the rank orders in the three sites clearly indicate the predominance over time
of Scaridae, Lethrinidae and Serranidae, and although there is some variation between sites below
the top ranked taxa, overall the fish families exploited are remarkably consistent, except for the
absence of Diodontidae in SE-RF-3. The pattern of inshore exploitation described for SE-RF-2 is
maintained over time, while a minor component of pelagic fishing, present in SE-RF-2, is again
documented, by Scombridae in SE-RF-19 and Elasmobranchii in SE-RF-19 and SE-RF-3.
The two fishing techniques that Anell (1955) regarded as late introductions to the area – kite
fishing for Belonidae, and shark noosing – are not reflected by increases in target taxa in the late
assemblage. A low archaeological visibility of some culturally important practices is no surprise,
whether explained by dietary practices, disposal methods, or taphonomy (see Kirch and Yen
1982: 292, Leach and Davidson 1988).
Molluscan Remains
All shell excavated from SE-RF-19 was retained for analysis, and was quantified by weight. Rank
order of species was compared with published (Swadling 1986) and unpublished (Szabó pers.
comm. 2004) analyses of shellfish remains from SE-RF-2. In SE-RF-19, Trochus and Turbo
species occupy the top two ranks, as they do in SE-RF-2, and Tridacnidae also rank amongst the
top taxa. Notably lacking from SE-RF-19 are Neritidae and various bivalves which rank highly
in SE-RF-2. The Neritas are not under-represented because of low shell weight, there are very few
191
in the assemblage. Taphonomic factors cannot be discounted so easily, however, and it is not
possible to determine whether the rather low bivalve component indicates a change in collecting
strategy or results from post-depositional weathering. A similar problem was investigated by
Robins and Stock (1990), who sought explanations for the under-representation in shell midden
on Moreton Island of two edible bivalves available in the local environment. SE-RF-19 also lacks
the large intact Trochus and Conus sp. observed in the SE-RF-2 assemblage.
Allowing for taphonomic factors, and internal variation within midden, SE-RF-19 generally
shows a similar exploitation strategy to Swadling’s earlier Main Reefs decorated Lapita site, in
that a diverse range of species was collected, without targeting a single (or small number of)
species. Overall, the assemblage indicates continuation of broad spectrum gathering predomi-
nantly from coral reef taxa, with a smaller component of soft shore taxa.
In SE-RF-3, different field collection methods were used. Shell was so abundant and the
component taxa so consistent, that samples were taken from all squares at Levels 1, 4, 7 and 10.
The 6 most abundant taxa were counted and discarded in the field, with examples being retained
for identification purposes. Although the range of taxa is very diverse, collecting strategies from
SE-RF-3 differ in two significant ways from the earlier sites. Firstly there is targeting of a few taxa,
in particular the small sand-dwelling stromb S. mutabilis, which dominates at all levels, and was
probably collected locally from along the rocks at Gnimoa. In each level, between 5-8 taxa account
for 87% or more of the total. Three is insufficient evidence to determine the cause of this change,
whether over-predation, environmental changes affecting mollusca or simply a change in collecting
location. Secondly, there is a marked mangrove component – Geloina sp. and Terebralia palustris
– which was brought to the site from other parts of the island, rather than collected on the coast
immediately adjacent. The increase in mangrove species has been attributed to silting resulting from
run off of volcanic ash deposited over Santa Cruz and the Reef Islands (Green 1986: 124-25).
Summary of Faunal Remains

Coherence in subsistence systems was assessed in three ways. The first was continuity in the pattern-
ing of resource, encapsulated in the notion of cuisines. Agricultural practices on Santa Cruz and the
Main Reefs develop from a common Oceanic plant roster, but with an emphasis on arboricultural
practices which differentiates the region from techniques practiced to the west (Yen 1973, 1974,
1976b). The faunal remains show basic continuity in mammalian taxa, with the possible exception
of dog, which was not found in Green’s three Lapita sites, or in SE-RF-3. Dugong and medium
Pteropodid are the only identified vertebrates recovered from SE-RF-2 not also found in the
SE-RF-19 and SE-RF-3. Allowing for differences explicable by sample size effects, there is continu-
ity in fish remains, except for the absence of Diodontid in SE-RF-3. A change in collecting strategy
is evident in the late molluscan remains, in which a smaller range of species was collected, and there
is a marked mangrove component. This trend may result from the combined effects over time of
collecting pressure in a local habitat, and accumulation of volcanic ash in lagoonal waterways.
The second criterion was the presence of introduced taxa requiring human intervention to
disperse. The significance of the Haplogroup II rats from SE-RF-3 has yet to be determined, as the
absence of Haplogroup III in Near Oceania is the subject of further investigation. Dogs may also
have a complicated history in these islands. The absence of dog bones from the earliest Lapita
sites, and their presence in SE-RF-19 may indicate an introduction during the plainware phase.
The third criterion was the introduction of new prey capture techniques of sufficient
complexity that they were more likely to have been learned than developed locally. The fish
assemblages on the Main Reefs do not provide robust evidence of the fishing techniques – kite
192
fishing and shark snaring – that have supposedly been acquired recently through cultural contacts
with other communities (Anell 1955: 41, 45). Such discrepancies between faunal remains,
archaeological evidence of fishing equipment, and historical narratives are unsurprising. As
shellfish gathering requires only simple techniques and tools, it is relatively uninformative about
minor changes to traditional gathering techniques.
Overall, a major anomaly in the faunal remains is the Haplogroup II R. exulans from
SE-RF-3. Otherwise developments that are observable can be adequately explained by local
innovation and response to environmental change.
CERAMICS
In the Main Reef Islands and Santa Cruz, the Lapita decorative phase ends about 650 – 750 B.C.,
and is followed by a plainware phase lasting until c. 100 B.C. – A.D.100 (McCoy and Cleghorn
1988), when pottery manufacture ceases and is never resumed. There is no evidence for either the
incised and applied relief traditions that develop from the Western Solomons to New Ireland, or
for the Mangaasi tradition of Vanuatu.
Several hundred sherds from Green’s three Lapita sites – representing both decorated and
undecorated vessel forms – were examined (i.e., Group IV sites, including the plainware component
of Lapita assemblages). These provided the baseline typology for rim and vessel forms.
Plainwares were examined from SE-RF-19 and SE-SZ-47 (Növlaö) (i.e., Group III sites),
and both decorated and plainwares from SE-SZ-33 (Mdailu) (i.e., Group III and IV sites).
Although proveniences were not available for all sherds from SE-SZ-33, McCoy and Cleghorn
(1988: 110) indicated that Layer V represents “the effective terminus of the Lapita ceramic
tradition at this site. The few incised sherds found above layer V can be readily attributed to
cultural and/or natural site formation processes”. Unpublished information provided by McCoy
indicated that 1613 undecorated sherds and 67 decorated sherds were recovered from SE-SZ-33.
The material sent to me included 13 dentate sherds, 35 incised sherds, 2 with possible finger-
pinching and incising, and another 4 that were possibly decorated. The lower number of decorated
sherds partly results from my discounting as making marks some lines apparently counted as
incising (discussed below). There are a small number of unusual rims in Mdailu (some from
plainware contexts) without precursors in Green’s Lapita sites, which were also not represented
in literature searches (Doherty 2007: 391, Fig.7.19). The decorated component of the Mdailu site
is described in Doherty (2007: 350-51). Figures 9.3 and 9.4 illustrate the larger decorated sherds
from Mdailu (SE-SZ-33). Figures 9.5 to 9.10 illustrate diagnostic rim sherds from this site.
On the basis of similarity in form, fabric and technology, the sherds excavated and surface
collected from SE-RF-19, plus sherds surface collected from adjacent Mound B, were treated as
one assemblage. The SE-RF-19 assemblage is in very poor condition, but contains the most
diverse range of lip and rim forms – see Figure 9.11. It differs from the Mdailu and Növlaö assem-
blages in the poor quality of some of the sherds, and the presence of robust sherds not seen in the
other Group III assemblages. Few large sherds survived in this site, making reconstruction of
vessel forms and comparison with other sites difficult. Of 671 body sherds measured that were
greater than 1 cm2 in area, only 9% are greater than 5 cm2. The percentage is even less if the
fragments are included in the sherd count. The largest excavated sherd was 23 cm2 in area, and
another five were 12 cm2 or larger.
Doherty (2007: 341-74) includes sherd numbers, provenience, rim categories and vessel
types. The following sections summarise the argument for coherence in morphological and
technological criteria.
193
Figure 9.3 SE-SZ-33 – Dentate and incised rim sherds and carinations. Levels not known. a) V9-26 incised;
b) M8-9 incised; c) M8-5 dentate; d) M9-16 dentate; e) N7-23 incised; f) 09-4 incised; g) N7-33
?finger pinch with incised line above.
194
Figure 9.4 SE-SZ-33 – Dentate and incised shaped sherds. Levels not known. a) 18 23A+23B shaped
sherd, dentate; b) N7-14 neck, incised; c) H13-11 ? neck, incised; d) N9-11 shaped sherd,
dentate; e) M9-14 ? neck, incised; f) 18-24 shaped sherd, incised; g) Q9-13 shaped sherd,
incised; h) N7 18A+18B shaped sherd, incised; i) Q9-15 shaped sherd, dentate.
Figure 9.5 SE-SZ-33 – Everted flared rims – Converging rim course. a) 19-27 layer III; b) M7-16 layer IV.
195
Morphological Attributes
In my view, on morphological grounds, the Group III plain assemblages developed out of the
Lapita plainwares. There are no new distinctive vessel forms, or parts (e.g., handles or spouts) nor
any form of body decoration (e.g., perforations, nubbins, applied bands, blocks of fingernail
impression, incised motifs) that suggest external influence or would allow the source of any
influence to be traced.
Continuity between the Group IV and Group III plainwares is evidenced by continuity in
some vessel forms, specifically Parker (1981) Form 6, globular/round necked vessel (the predom-
inant form in SE-RF-19 and SE-SZ 47); and Parker Form 3 carinated vessel (well represented in
plainware layers in SE-SZ-33, present in low numbers in SE-RF-19, and absent from SE-SZ-47).
Although bowls are also present, there are few of them and they are deeper than Parker’s Form 2
bowls, and have quite variable rim forms. The vertical, narrow necked vessel from SE-SZ-8
(Form 7) is not represented in any of the plainware assemblages, which apparently also lack the
shallow dish (Form 1), and carinated Form 4 vessel.
There is also continuity in some rim forms between the undecorated Lapita rims (Group IV)
and rims from plainware assemblages (Group III). The horizontal variant of the rarer complex/
faceted rims is present in Mdailu and SE-RF-19 plainwares (see Figures 9.9 and 9.11). However
the typical plainware rim (Group III sites) is a simple everted, converging or parallel sided rim,
sometimes with simple notching across the lip. These rims are typically shorter, everted more
acutely, and converge to finer lips than Group IV everted rims (see Figures 9.7 and 9.8 for
SE-SZ-33, and Figure 9.11 for SE-RF-19). By default – the absence or low incidence of shaped
shoulder sherds in Növlaö and SE-RF-19 – this rim probably derives from a globular vessel,
although it is difficult to estimate the extent of neck constriction. This form is not common in the
plain component of assemblages from Green’s three Lapita sites, nor amongst the decorated
Lapita levels in Mdailu. There are rare examples in the Lapita assemblages, and Scarlett Chiu
(pers. comm. 2006) advised me she has seen them in other Lapita assemblages. They can therefore
be regarded as evidence of continuity, but exhibit a change in frequency.
Continuity is also exhibited in simple lip decoration techniques. Diamond notching, incision,
and “stick” notching are present in all ceramic phases, although the locus of decoration shifts
from a preference for the inner lip margin in Lapita (Group IV) plainwares, to across the lip in
the Group III plainwares. Two sherds in Mdailu and SE-RF-19 exhibit very similar deeply-
excised notches across the lip.
Figure 9.6 SE-SZ-33 – Everted flared rims – Parallel rim course. MI7-17+ 1 layer VI.
196
Figure 9.7 SE-SZ-33 – Everted, curved rim – Converging rim course. Orientation angle uncertain.
a) H14-10 layer IV; b) M8-4 layer IV; c) E9-7 layer II/III contact; d) 09-3 layer I; e) 09-5 layer I.
Figure 9.8 SE-SZ-3 – Everted straight rim – Converging rim course. Orientation angle uncertain.
a) H13-46+47 layer?; b) U9-10 layer II.
There are, however, a number of differences between the Lapita plainwares (Group IV) and
Group III plainwares, in addition to the loss of some vessel forms noted above. The elongated
soft-necked everted rim so common in the Group IV plainwares is rare in the Group III plainwares
(although present in SE-RF-19). As noted above, a more acutely outcurving rim with converging
rim course predominates. Bilateral asymmetrical incising of the lip is also absent from Group III
plainwares. External making marks, consisting of deep indentations usually slightly below, and
running parallel with, the vessel neck, were common in Group III assemblages – see Figure 9.5,
Figure 9.7 (a and c) and Figure 9.11 (e). Such pronounced marks were not observed on plain
vessels from Green’s three Lapita sites, which had either smooth inflection points, or very light
marks. This may indicate a different process for forming the rim i.e., attaching a separate band of
clay rather than extruding the rim from the upper shoulder of the vessel.
197
Figure 9.9 SE-SZ-33 – Rims with complex/faceted forms. a) N8-10a+l0b layer IV; b) N8-9 layer VI.
a)a)
b)b)
Figure 9.10 SE-SZ-33 – Bowls. a) 18-22 layer V; b) N8-2a-2h layer I.
198
Figure 9.11 SE-RF-19 – Assorted rim sherds. a) 202/1; b) P54; c) P67; d) 208/6; e) P180; f) P93; g) 208/9;
h) 209/3. Orientation angle uncertain for a-d.
199
In my view before external influence is invoked to explain the variation observable between
plainware assemblages (Group III sites), local factors such as temporal differences, and the possi-
bility that assemblages were manufactured in different locations must first be disproved.
Manufacturing Technology
Temper composition, volume and grain size were compared for the decorated Lapita and
plainware assemblages.
Temper Composition
Dickinson has previously described a Reef-Santa Cruz temper suite, based on examinations of
sherds from Green’s three Lapita sites, and sherds submitted by McCoy from Lapita and plainware
sites on Santa Cruz and Tömotu Neo, including four sherds from SE-SZ-47 (Növlaö) and four
from SE-SZ-33 (one from Level VI and three from plainware levels) (Dickinson 1978, 1979,
1995, 1998). These analyses have established that sherds from both Santa Cruz and Reef Islands
sites contain manually added, variably placered temper sands of beach origin, which are petro-
graphically indistinguishable and are indigenous to Santa Cruz. The grain types include varying
proportions of (i) ferromagnesian grains (principally clinopyroxene but also opaque iron oxides,
minor olivine and rare hornblende; (ii) plagioclastic feldspar; and (iii) volcanic rock fragments
derived from Santa Cruz bedrock assemblages. The amount of calcareous grains varied. The
temper relationship established that either pots or raw materials had been transported from Santa
Cruz to the Reef Islands, on a huge scale.
Twenty-six sherds from SE-RF-19 were submitted to Dickinson for petrographic analysis
(Dickinson 2000). All but one (which was unsuitable for analysis because of its low mineral
content) contained manually added volcanic sand tempers from the Santa Cruz temper suite.
Various grain proportionality indices indicate that the SE-RF-19 sherds are indistinguishable from
Nendö-Reef Lapita tempers analysed previously. SE-RF-19 did not contain the lithic variant of
the suite, an observation also applying to sherds from Graciosa Bay previously examined by
Dickinson. Peculiar to this site was a “super pyroxenic” variant (>80% FM grains) not seen in
any of the sites analysed previously. Dickinson (2000: 4) concluded “ ... the Nendö-Reef Lapita
and Reef post-Lapita tempers are qualitatively and quantitatively comparable, suggesting that
prehistoric potters used the same general kind of sand as temper throughout the ceramic period in
the Nendö-Reef portion of the Santa Cruz group”.
Tempers in the Mdailu and Növlaö assemblages were examined microscopically, and all
were found to be of local origin.
Assessment of temper composition and volume were impeded by post-depositional leaching
of calcareous grains, which was observed in all assemblages to varying degrees. It is difficult to
accurately assess the extent of calcareous leaching, and until this effect is controlled for, statistical
counts of calcareous vs non-calcareous tempers and estimates of temper volumes suggest a level
of accuracy that is in fact spurious. My observations of the tempers in all assemblages (including
decorated Lapita) are: (i) the paste compositions are generally medium-dense (Snow and Shutler
1985), with open textures being rare; (ii) grain size is generally medium, with a smaller propor-
tion of very fine and fine sizes also present, but coarse grains are rare. Both calcareous sand
tempers and mineral-only tempers were selected for decorated Lapita, plainwares in decorated
assemblages, and plainwares. The calcareous component is quite variable, from incidental grains,
to medium amounts of shell, to high volumes of shell. Overall, temper selection appears very
expedient. This accords well with Clough’s (1992) assessment of the heterogeneity of Reef-Santa
Cruz pastes, and the versatility of the open-firing technology in use in Oceania.
200
Manufacturing Techniques
It is difficult to assess surface treatments for all assemblages given the condition of sherds from
SE-RF-19. It was not apparent that slips (other than self-slips) were used on any assemblage. The
surface of sherds from Növlaö generally exhibited less weathering, and some retained a shiny
surface perhaps indicative of an applied surface treatment. This has yet to be investigated.
Certain attributes are common to all assemblages (including plain sherds from both
decorated Lapita and plainware levels at Mdailu). These are: (i) the presence of unoxided cores;
(ii) anvil marks/impressions on sherd interiors, indicative of paddle and anvil forming/finishing
techniques; (iii) laminating edge fractures, indicative of beating (Rye 1981), and (iv) exterior
blackening from heating the vessel.
The most obvious finishing technique is the making/tool marks running horizontally along
the exterior neck. These constitute a small point of difference in the plainwares, as such marks are
present but rare and not so pronounced in the plain sherds from Green’s three Lapita sites. (One
example from these sites is SE-SZ-8 11/2, a large convex upper shoulder/neck sherd).
The assemblages are very similar overall, with approx. 70-80% of measured bodysherds
between 4-8 mm, and occasional finer and thicker sherds in all assemblages. No changes over
time in bodysherd thickness are discernible.
Summary of Ceramics
Morphologically, the plainwares lack indications of external influence that can be traced, and the
argument was made that they are derived from the Lapita ceramic tradition, exhibiting continuity
in temper, many aspects of manufacturing technology, and a number of vessel and rim forms. The
variation that exists between plainwares can be explained by local factors such as temporal differ-
ences, and the likelihood that assemblages were manufactured in different locations.
WORKED SHELL
All shell objects from SE-RF-19 were very badly weathered, hindering identification of shell
species, and reconstruction of artefact forms.
Tridacna Breast Plates

Three pieces of worked Tridacna sp. from SE-RF-3 (one excavated, two surface collected) are
likely to be pieces of completed Tridacna breast ornaments or kapkap. The form of the tortoise-
shell overlay served to make the Reef-Santa Cruz version of this widely distributed ornament
distinctive. These ornaments are better represented in late sites, but at present are not documented
in the Lapita ornament repertoire. The earliest archaeological examples come from Taumako,
where Davidson and Leach (1991) excavated disks in early (2600 B.P.) and late contexts. Watom
may provide evidence of another early disk, from SAD Trench IV, spit 11, one of the trenches that
had been reworked and were “largely post-Lapita formations (Specht 2003: 130, 132). Despite
the problem of mixing, the object was tabulated in the category “[d]efinitely or probably of
pottery age” (Specht 2003: 130 Table 6).
Other Tridacna Ornaments

Another very weathered ornament was excavated from SE-RF-19. It is difficult to identify the
material, although Green thought it was Tridacna. This appears to have been a disk rather than a
ring, with a perforation about 1 cm from the edge. It is not readily identifiable as a Lapita form.
201
Trochus Rings
Thirty-four fragments of Trochus rings were excavated from SE-RF-3. Some are definitely T.
niloticus. The internal diameter varies from 4-9 cms, with most size classes (at ½ cm intervals)
represented between these extremes. On Santa Cruz, Trochus rings represent a point of continuity
between late ceramic and early aceramic sites (McCoy and Cleghorn 1988), and were also
recovered from a number of Group I sites.
Tridacna Adzes
Seven reconstructible Tridacna adzes were excavated from SE-RF-19, another one was acces-
sioned as a “micro adze” (Green 1979 ms), and two others were too damaged to determine their
original forms. Only one dorsal form was excavated, made across the lip or body of the shell and
retaining evidence of the external architecture. Three hinge adzes were excavated from this site,
two of which are Kirch and Yen (1982) Type 6 adzes, while the third is somewhat like a Type 6.
Particularly intriguing in this site are three adzes made along the rib or fold of the Tridacna,
one of which, a robust example, is probably made from Hippopus (Szabó pers. comm. 2001). The
adzes are illustrated in Figure 9.12. Rib/fold adzes are not documented formally in the Lapita adze
suite, although Szabó has identified one piece of worked Hippopus hippopus within the SE-RF-2
assemblage, comprised of two major ribs, cut down one side, which may be an unfinished adze
(Szabó pers. comm. 2004). It is also possible from the descriptions and illustrations of Tikopian
and Anutan adzes that a small number of rib/fold adzes may have been subsumed into hinge or
dorsal categories (e.g., Kirch and Rosendahl (1973: 75 Fig. 23d) looks very like SE-RF-19 A-12,
with a concave bevel and thin cutting edge made from the edge of the shell. The rib adzes from
SE-RF-19 are quite diverse in size and finished form, unlike the more standardised types recog-
nised as rib adzes elsewhere), for example, Takuu (Moir 1989), Ryuku (Asato 1990: 34), or
Pamwak and Golo (Szabó 2004). Rib adzes do not become more common over time in Santa Cruz
or Reef sites, as only one possible example was recovered from Mateone, which had 65 shell
adzes in total (McCoy pers. comm. 2004). On information from Takuu (Moir 1989), no addition-
al technical skills are needed to produce them.
Also excavated from SE-RF-19 was a Tridacna sp. chopper which fits snugly into the palm
and retains probable use wear scars on three edges. It may be a failed preform, and utilises the
hinge section and two adjacent folds of the shell, giving it an asymmetrical shape in plan and
cross section. The field identification of the micro adze should in my view be discarded as it quite
unlike any micro adzes I have seen illustrated.
Five of six adzes recovered from SE-RF-3 were surface collected. The single excavated adze
from SE-RF-3 is a dorsal type, while the forms of the other two could not be reconstructed because
of the degree of weathering and damage. The adze forms recovered do not reflect the diversity of
adze forms present in the historic period, judging by illustrations in Koch (1971) and unpublished
accession records from McCoy. Both hinge and dorsal region adzes are present in Group I contexts.
Mitra/Terebra Adzes
Twenty-seven adzes of Mitra or Terebra shell were recovered from SE-RF-3: 3 excavated, 23
surface collected in 1976 and 1979, plus one ethnographic example collected by Green. Terebra/
Mitra adzes have a particular significance as they are “the most visible archaeological evidence
for late contact between Micronesia and Melanesia” (Intoh 1999: 413). They are a distinctive
horizon marker, well documented over a wide area in Micronesia and Melanesia by c. 1000-1200
202
Figure 9.12 Tridacna adzes – Rib/fold forms – SE-RF-19. a) A-9, SS-TT 48-49 Layer 1 Level 3; b) A-I2,
VV-49 Layer 2 Level 8; c) A-I3, VV-WW48-49 Layer 2 Level 8.
203
A.D. No examples of shells modified at both ends were found in SE-RF-3. Two types were identi-
fied based on the thickness of the cutting edge and the treatment of the internal face.
Lambis Adze?
One possible Lambis adze was surface collected from SE-RF-3. It was very weathered and the
bevel end is completely eroded. At present it is an anomaly, as Lambis adzes are not reported from
any of the sites investigated by McCoy and Cleghorn; nor from Green’s Lapita sites, nor from
Tikopia and Anuta (Kirch and Rosendahl 1973, Kirch and Yen 1982). They are known from
Vanuatu, however, where they are confined to the later end of the sequence, probably 600 B.P.
and after (Bedford 2000: 193).
Worked Pinctada sp.

Five pieces of worked Pinctada sp. were excavated from SE-RF-3, two of which are perforated. It
is not possible to be certain of the function. Ethnographically, this shell was used for pendants, and
for spoons which were often perforated (Speiser 1916: 59-61 and Edge-Partington 1969: 159 Figs
5, 6, and 7). Worked pearlshell is also present Group III and IV sites (Mdailu and SE-RF-2).
Miscellaneous Shell
A Cypraea tigris cap, with no obvious signs of working, was excavated from SE-RF-3. This
appears to conform to Spennemann’s Type 210 breakage pattern (Spennemann 1993: 43 Fig. 5X)
which resulted from dropping the shell on its side. However the use of tiger cowrie caps as
octopus lures is attested ethonographically from Santa Cruz (Finsch 1914: 135, cited in
Spennemann 1993: 42), even if this particular example is not certain. A piece of Isognomen sp.
with cut/worked edges was also excavated from this site. Its function is unknown.
Worked Pencil Sea Urchin Spines

Three examples of worked sea urchin spines were excavated from SE-RF-3, two of which have
bevelled ends. Although not documented from Green’s three Lapita sites, this form is well
documented from early contexts on Tikopia (Kirch and Yen 1982) and Anuta (Kirch and
Rosendahl 1973: 82-83).
Worked Shell Summary

The effects of post-depositional weathering were significant for shell objects recovered from
SE-RF-19. There are few components of the Lapita shell ornament suite that survive into Group
I sites – a fact completely congruent with the direction of change observed in neighbouring
sequences. Nor were there any additions to the ornament suite that permit identification of
external derivation. The rib form Tridacna adzes in SE-RF-19 are unusual but require no addition-
al technological skill to produce. Given the lack of standardisation of form, no external source
can be identified for them, and in my view they represent a local innovation rather than external
influence. The preference for shell adzes developed over time on the Main Reefs can be explained
by environmental constraints. The late adze suite contains some Tridacna adzes that resemble
both Lapita dorsal and hinge forms, with the addition of Terebra/Mitra adzes which form a
regional horizon.
204
LITHICS
Obsidian and Volcanic Glass
No references to obsidian have been found in the historic literature, and no obsidians or volcanic
glasses were recovered from SE-RF-3. Fourteen pieces of obsidian were recovered from SE-RF-19
− six pieces were excavated, six were collected from the surface of the excavated mound, and another
two surface collected from adjacent Mound B. PIXE-PIGME analysis carried out by Summerhayes
on 13 pieces determined that 12 derived from a Banks Islands source (probably Vanua Lava), and
one was sourced to Umrei in the Admiralty group (Summerhayes pers. comm. 2001). The small
assemblage from SE-RF-19, consisting of 14 pieces of broken flakes and shatter, nine of which weigh
less than 2.0 gms, contributes nothing additional to the technological analyses conducted by Sheppard
(1992, 1993) of the obsidians in Green’s three Lapita sites (SE-SZ-8, SE-RF-2, and SE-RF-6).
The single piece of Umrei obsidian in SE-RF-13, and three pieces of Admiralties obsidian in
the Kiki Phase on Tikopia (Spriggs 1997: 137), in both cases in conjunction with more numerous
Banks Island glasses, suggest some ongoing contacts with the west. By the plainware period,
however, the preferred source had switched from Talasea to the lower quality but closer Banks
Islands source(s), which was only a minor component in Green’s Lapita sites. Banks source(s) also
predominate in the six pieces argued to be associated with ceramics on Taumako (Leach 1985).
Obsidian is rare in the late (Group 1) sites – certainly it was not found in SE-RF-3 or
Naiavila, or Kirch’s midden mound (SE-SZ-4), but was reported by McCoy and Cleghorn (1988)
from Mateone, where it is probable all samples derive from a Banks Islands source. The authors
referred to obsidian in later sites (i.e., more than one) but provide no other information about
which sites specifically (McCoy and Cleghorn 1988: 114). Obsidian has not been reported from
late contexts in Taumako, nor on Vanikoro or Anuta, but four pieces were found in the Sinapupu
Phase on Tikopia, with a greater abundance from the Tuakamali Phase, all of which derives from
a Banks source (Kirch and Yen 1982: 260).
Because no obsidian was reported from the early aceramic layers at Növlaö, it is not possible
to argue that the obsidian in Mateone represents the maintenance of a continuous supply from the
Banks Islands, rather than the establishment of a new supply stream late in prehistory, at about
the same period when Banks obsidian reappears in quantity in Tikopian sites. Furthermore, given
the patchy distribution in late (Group 1) sites, the possibility that supply in the late period was
restricted to certain villages cannot be excluded.
Chert
Only two pieces of volcanic tuff/silicified mudstone were excavated from SE-RF-3, while 28
pieces of chert and chert-like materials were excavated SE-RF-19. The 12 true chert samples
excavated from SE-RF-19 included nine of probable Ulawan/Malaitan origin c. 400 km to the
northwest; two pieces possibly from the Duff Islands about 150 km east of Nenumbo; and a single
zebraic chalcedony sample that was somewhat similar to a sample from Niuatoputapu (# NT-90-
SA-5). The Niuatoputapu sample, in turn, has been reported as being similar to chert examined
from Tikopia (Sheppard 1996: 109, 111). There was also a single piece of zebraic chalcedony in
SE-SZ-8. The source of this material is not known, but may be located somewhere in Northern
Vanuatu, or north of Tikopia (Sheppard 1996: 109). There is continuity in the sources of chert
from decorated Lapita to the SE-RF-19 plainware site. The small sample from SE-RF-19, which
included only two complete flakes, is insufficient for generalisations about reduction technology.
Points, which were made almost exclusively of chert in Green’s three Lapita sites, are absent.
205
Slingstones
Although no pointed-end slingstones were excavated from either SE-RF-3 or SE-RF-19, continu-
ity in the use of shaped slingstones is the most parsimonious explanation for their presence in the
SE-RF-2 Lapita site, and in multiple sites at the end of the Reef-Santa Cruz sequence. However,
the case is not unimpeachable, as there are no shaped slingstones in Reef-Santa Cruz sites
bridging the time gap, and although the narratives record stones as missiles, they do not record
slingshots or shaped missiles.
Other Lithics
A number of artefacts from SE-RF-3 and SE-RF-19 have been interpreted as manufacturing tools
of various sorts. In most cases, there are only a few examples of each tool, and some are repre-
sented by a single specimen. While some of the large objects are definitely artefacts, the status of
others is less secure. The patchy distributions do not make a robust base from which to make
arguments about continuity and change.
Adzes
Although no stone adzes were recovered from SE-RF-19, continuity into plainware contexts of
Lapita types and the fondness for green stone is suggested by unpublished information supplied
by McCoy. Over time, however, and probably commencing in the plainware phase, stone adzes
become relatively unimportant relative to shell adzes, a situation which persisted to the end of the
sequence. Codrington included the Cruzians among those users of shell rather than stone tools
(1972 [1891]: 16, 313), and there is scant mention of stone adzes in the historic narratives. Koch’s
(1971) ethnological study illustrates a number of stone adzes manufactured from Cruzian stone
which are quite variable in form, and include some that could be accommodated in a Lapita
typology, the “Melanesian” form, and others with no obvious stylistic affiliations. At the late end
of the sequence, this area is a participant in the process which sees adzes of West Polynesian form
and Samoan material distributed over a wide area, extending as far as San Cristobal, and including
Tikopia, Anuta and Taumako (Best et al. 1992).
An adze of metavolcanic green schist was surface collected from SE-RF-3, and Dickinson’s
report indicates possible sources including the metamorphic belt of New Caledonia or the Pacific
fringe of the Solomon chain – Malaita/Ulawa and Santa Isabel (Dickinson 2004). The adze could
be accommodated within the Lapita adze suite, and may be a Lapita remnant. Two other objects
from SE-RF-3, one an intermediate volcanic rock and a slightly altered piece of andesite, were
accessioned as possible adze fragments.
Stone Knife
A flaked stone knife of intermediate volcanic rock with hornblendes was surface collected from
SE-RF-3. It is bifacially flaked, reduced for hafting, and shows edge damage with slight retouch
only. No mention of such knives has been found in the ethnographic literature. This is a unique
object with no comparable examples in earlier or contemporaneous sites.
Hammerstones, Sharpening-Stone/Abraders and Pestle-Type Pounders
Small pebbles of intermediate volcanic rock were excavated from both SE-RF-3 and SE-RF-19.
All retain evidence of hammering. An oval-shaped cobble of intermediate volcanic rock, which
lacks obvious use wear, but has a piece avulsed from the back, was recovered from SE-RF-3. An
elongated object of intermediate volcanic rock, showing flattened ends due to hammering, was
excavated from SE-RF-19. A similar but somewhat larger object, used at both ends, was recovered
from SE-RF-2, and was interpreted as a nut cracking pestle.
206
Worked Stone Summary

Overall, the lithic evidence supports continuity in lithic forms between decorated Lapita and
plainware sites but contraction in obsidian networks by the plainware period. Despite some conti-
nuity in Mdailu, the absence of exotic adzes from SE-RF-19, and local derivation of the assorted
lithic materials in that site suggests the change is already underway during the plainware period,
and there may be differential access to exotic materials. Probably commencing during the
plainware period, stone adzes apparently become less important relative to shell ones. There are
no late parallels for the small meta-volcanic chisel excavated from Ngamanie (SE-RF-6) or the
stone bowl or pounding cup excavated from Nenumbo (SE-RF-2).
Raw materials in the late sites are predominantly intermediate volcanics and miscellaneous
volcanic rocks of local origin, either from Santa Cruz or Tinakula, representing a reduced range over
which materials are acquired. The schist adze surface collected from SE-RF-3 seems likely to be a
Lapita Period remnant when viewed in the context of the other lithic sources in use at the time.
WORKED BONE
The bone component of the Lapita Cultural Complex lacks the coherence demonstrable in
ceramics and shell artefacts, perhaps for taphonomic reasons relating to bone preservation. Green
(1979: 40) included bone needles, awls, tattooing chisels, and spear points in the Lapita inventory,
but of these only awls and needles are well represented and can be reliably documented in Lapita
and post-Lapita deposits elsewhere.
No bone artefacts were recovered from SE-RF-19. The ten excavated from SE-RF-3, included
one awl in mammal or fish bone and one probably of Pteropodid [bat]; a needle in Pteropodid, two
arrow shafts and one arrow point of mammal or fish bone, and three other unidentified worked
bone fragments. In addition, one piece of possible human bone, the R. proximal fibula of a gracile
individual (J. Littleton pers. comm. 2002), may be part of a compound arrow point, which would
accord with ethnographic descriptions (e.g., Coote 1883: 105, Speiser 1913: 282).
There is little to link early and late period bone assemblages. Continuity in a generic sense is
evidenced by needles and awls. Early use of Pteropus bone is not documented for Green’s three
Lapita sites, but is reported for other early sites, and is also found in late (Group I) Reef-Santa
Cruz sites. Bone arrow points lack earlier precursors, but the use of bows and arrows for combat,
of composite arrows, tips of human bone, and poisoned arrows have parallels in neighbouring
sequences. Green (2003: 16 and Table 5) has suggested the Lapita tattooing kit included three
kinds of tools − Pteropus wing bone used as both needle and tattoing needle; a toothed comb-like
chisel in shell or bone; and obsidian graver points. None of these tattoing implements can be
documented for Santa Cruz or the Reef Islands in the late period, but diversification in needle
material is not unusual in a wider geographical context, at which scale no patterning is apparent.
CONCLUSIONS
On Santa Cruz and the Main Reefs, the material complex of late sites is an accumulation or
palimpsest of components of different longevity, origins and variable geographic distribution.
Retentions and deletions from the founding Lapita Cultural Complex occur over time, but not at
a rate or scale that is incongruous in wider context. Some changes in material culture were influ-
enced by local materials and conditions, while others choices are less constrained. It is not
possible on data presently available to indicate a point of disjuncture in the sequence likely to
indicate the arrival of a different cultural group.
207
Santa Cruz and the Main Reefs Islands are geographically situated to receive castaways as
well as purposeful visitors from east and west. At different times, different influences washed
through here (e.g., from Micronesia and Polynesia) − probably more than we know of from the
archaeology − but not at a rate or scale that is incongruous in the wider regional context. In the
late sites, Terebra/Mitra adzes appear, which are durable evidence of a regional horizon encom-
passing sites in Melanesia and Micronesia. The backstrap loom is another non-durable link. A
very few adzes of Polynesian affiliation also appear in late sites, again part of a wider process
extending to San Cristobal and Vanuatu. Networks of interaction of varying geographical scales
can be recognised at different time periods. To this extent, history in the islands is reticulate.
From the historical sources, some aspects of material culture that are innovations post-dating
the Spanish visit could be identified − i.e., the chronology could be refined beyond what was
possible from the late archaeological record. Red feather money seems to be in this category, and
possibly the turtle shell overlay on the breast ornament. On present evidence, the monumental form
of the dance circle was a very recent development, adopted from Santa Cruz, to the Main and Outer
Reefs, but not beyond. The round houses of the late period may be confined to Santa Cruz, but
otherwise the interior layout of houses and their uses are similar on the Main Reefs, as are the main
components of settlements. For some elements of late material culture there is no evidence concern-
ing their ultimate age. The dried breadfruit nabo, for example, has not so far been documented
archaeologically, but has an antiquity of at least 400 years based on the Spanish reports.
Overall, taking a wider geographic perspective, it was not apparent from the rate of change
or the types of change in material culture that there was a different history of process occurring
in these islands. The material culture of the late period is of diverse origins. Given these facts, the
archaeological evidence overall suggests continuity, with additional inputs over time from diverse
sources, not cultural replacement. The evidence from archaeology is congruent with the revised
assessment of the status of the languages of Santa Cruz and the Main Reefs, as Oceanic languages
(Ross and Næss 2007).
Acknowledgements
Special thanks are due my two supervisors Roger Green and Peter Sheppard, for practical assis-
tance and encouragement through the long years of Ph.D. study. I am particularly indebted to Pat
McCoy for providing access to comparative material and unpublished records, but also offer my
thanks to many others who helped with analysis, or by providing access to unpublished material.
PIXE-PIGME analysis of the obsidian was carried out by Glen Summerhayes under AINSE
Research Grant 01/134. The thesis was funded by grants from the University of Auckland
Graduate Research Fund and the Green Foundation for Polynesian Research. The maps and
drawings were produced by Joan Lawrence and Seline McNamee.
REFERENCES
Anell, B., 1955. Contribution to the History of Fishing in the Southern Seas. Uppsala: Almquist and
Wiksells Boktryckere AB.
Asato, S., 1990. The Urasoko Site: A Sketch of the Excavation in Photographs. Okinawa, Japan: The
Gusukube Town Board of Education.
Beattie, J., n.d. Voyage in the “Southern Cross” August-November 1906. Unpublished Diary: Auckland
War Memorial Museum Library M.S. 1045.
Bedford, S., 2000. Pieces of the Vanuatu Puzzle: Archaeology of the North, South and Centre. Unpublished
Ph.D. thesis, Australian National University, Canberra.
208
Best, S., P.J. Sheppard, R.C. Green and R. Parker, 1992. Necromancing the stone: Archaeologists and adzes
in Samoa. Journal of the Polynesian Society, 101: 45-85.
Carsten, J. and S. Hugh-Jones (eds), 1995. About the House: Levi-Strauss and Beyond. Cambridge and
New York: Cambridge University Press.
Clough, R., 1992. Firing temperatures and the analysis of Oceanic ceramics: A study of the Lapita ceramics
from Reef/Santa Cruz, Solomon Islands. In J-C. Galipaud (ed.), Poterie Lapita et Peuplement.
Nouméa: ORSTOM, pp. 177-92.
Codrington, R., 1972 [1891]. The Melanesians: Studies in Their Anthropology and Folk-lore; With
Illustrations. New York: Dover Publications.
Coombe, F., 1911. Islands of Enchantment: Many-Sided Melanesia Seen Through Many Eyes and Recorded
by Florence Coombe. London: Macmillan.
Coote, W., 1883. The Western Pacific: Being a Description of the Groups of Islands to the North and East
of the Australian Continent. London: Sampson Low, Marston, Searle and Rivington.
Davenport, W., 1962a. Red-feather money. Scientific American, 206: 94-103.
Davenport, W., 1962b. Comments on A. Capell, “Oceanic linguistics today”. Current Anthropology, 3: 400-2.
Davenport, W., 1964a Social structure of Santa Cruz Island. In W. Goodenough (ed.), Explorations in
Cultural Anthropology. New York: McGraw-Hill, pp. 57-93.
Davenport, W., 1964b. Notes on Santa Cruz voyaging. Journal of the Polynesian Society, 73: 134-42.
Davenport, W., 1965. Sexual patterns and their regulation in a society of the Southwest Pacific. In F.
Ambrose (ed.), Sex and Behaviour. National Research Council (U.S.), Committee for Research in
Problems of Sex, University of California, Berkeley. New York: Wiley, pp. 164-207.
Davenport, W., 1968a. Social oganisation notes on the Southern Santa Cruz Islands: Utupua and Vanikoro.
Baessler-Archiv, 16: 207-75.
Davenport, W., 1968b. Social organisation notes on the Northern Santa Cruz Islands: The Duff Islands
(Taumako). Baessler-Archiv, 16: 137-205.
Davenport, W., 1969. Social organisation notes on the Northern Santa Cruz Islands: The Main Reef Islands.
Baessler-Archiv, 17: 151-243.
Davenport, W., 1972. Social organization notes on the Northern Santa Cruz Islands: The Outer Reef
islands. Baessler-Archiv, 20: 11-95.
Davenport, W., 1975a. The population of the Outer Reef Islands, British Solomon Islands Protectorate. In
V. Carroll (ed.), Pacific Atoll Populations. Association for Anthropology in Oceania Monograph No.
3. Honolulu: University of Hawaii Press, pp. 64-116.
Davenport, W., 1975b. Lyric verse and ritual in the Santa Cruz Islands. Expedition, 8 (1): 39-47.
Davenport, W., 1990. The figurative sculpture of Santa Cruz Island. In F. Hanson and L. Hanson (eds), Art
and Identity in Oceania. Honolulu: University of Hawaii Press, pp. 98-110.
Davenport, W., 2005. Santa Cruz Island Figurative Sculpture and its Social and Ritual Contexts.
Philadelphia: University of Pennsylvania Museum of Archaeology and Anthropology.
Davidson, J. and F. Leach, 1991. Bird-man amulets and Tridacna shell discs from Taumako, Solomon
Islands. In A. Pawley (ed.), Man and a Half: Essays in Pacific Anthropology and Ethnobiology in
Honour of Ralph Bulmer. Auckland: The Polynesian Society, pp. 478-83.
Dickinson, W.R., 1978. Sand Tempers in Southeast Solomon Sherds from Feru Rockshelter on Santa Ana
and Lapita Sites in the Main Reef and Santa Cruz Group. Oceanic Prehistory Records 7: Studies in
the Sourcing of Materials from the Southeast Solomons. (Petrographic Report WRD-68). Auckland:
University of Auckland Archaeology Society. [Microfiche].
Dickinson, W.R., 1979. Sand Tempers in Nendo Sherds. Petrographic Report WRD-81.
209
Dickinson,W.R., 1995. Temper Types in Prehistoric Vanuatu Potsherds Indigenous to Efate, Santo,
Malakula, and Other Islands of the New Hebrides Island Arc. Petrographic Report WRD-117.
Dickinson,W.R., 1998 [revised]. Petrographic Comparison of Placer Sand Tempers in Erromango (Vanuatu)
and Nendo-Reef (Santa Cruz Group) Lapita Sherds. Petrographic Report WRD-170.
Dickinson,W.R., 2000. Petrography of Sand Tempers in Post-Lapita Sherds from the Reef Islands, Santa
Cruz Group, Eastern Solomon Islands Outliers. Petrographic Report WRD-197.
Dickinson,W.R., 2004. Petrography of a Cobble and Adze from the Reef Islands. Petrographic Report
WRD-242.
Dunn, M., G. Reesink and A. Terrill, 2002. The East Papuan languages: A preliminary typological
appraisal. Oceanic Linguistics, 41: 28-62.
Edge-Partington, J., 1969. An Album of the Weapons, Tools, Ornaments, Articles of Dress of the Natives of
the Pacific Islands: A Facsimile of the Original Work. London: Holland Press.
Finsch, O., 1914. Südseearbeiten. Gewerbe und Kunstfleiss, Tauschmittel und ‘Geld’ der Eingeborenen auf
Grundlage der Rohstoffe und der geographischen Verbreitung. Friederichsen, Hamburg. Cited in
Spenneman 1993.
Fox, J.J., 1993. Comparative perspectives on Austronesian houses: An introductory essay. In J.J. Fox (ed.),
Inside Austronesian Houses: Perspectives on Domestic Designs for Living. Canberra: Australian
Girieud, J. and H. Herrenschmidt, 1898. La Nouvelle-Calédonie, Suivi d’un Aperçu sur les Nouvelles-
Hébrides, les Iles Santa-Cruz et les Swallow. Rouen: J.Girieud.
Grayson, D., 1984. Quantitative Zooarchaeology: Topics in the Analysis of Archaeological Faunas.
Orlando: Academic Press.
Green, R.C., 1976. Lapita sites in the Santa Cruz Group. In R.C. Green and M.M. Cresswell (eds),
Southeast Solomon Islands Cultural History: A Preliminary Survey. Bulletin 11. Wellington: Royal
Green, R.C., 1979. Lapita. In J. Jennings (ed.), The Prehistory of Polynesia. Cambridge, Mass.: Harvard
Green, R.C., 1979, ms. Field Notebook III of Stage 2, Southeast Solomon Islands Cultural History
Programme. [author’s copy]
Green, R.C., 1986. Lapita fishing: The evidence of Site SE-RF-2 from the Main Reef Islands, Santa Cruz
Group, Solomons. In A. Anderson (ed.), Traditional Fishing in the Pacific. Pacific Anthropological
Records 37. Honolulu: Department of Anthropology, Bernice P. Bishop Museum, pp. 119-35.
Green, R.C., 1987. Obsidian results from the Lapita sites of the Reef/Santa Cruz Islands. In W.R. Ambrose
and J.M.J. Mummery (eds), Archaeometry: Further Studies in Australasia. Canberra: Department of
Prehistory, Research School of Pacific Studies, Australian National University, pp. 239-49.
Green, R.C., 1998. From Proto-Oceanic *Rumaq to Proto-Polynesian *Fale: A significant reorganisation in
Austronesian housing. Archaeology in New Zealand, 41 (4): 253-72.
Green, R.C., 2002. Rediscovering the social aspects of Ancestral Oceanic societies through archaeology,
linguistics and ethnology. In S. Bedford, C. Sand and D. Burley (eds), Fifty Years in the Field:
Essays in Honour and Celebration of Richard Shutler Jr’s Archaeological Career. New Zealand
Archaeological Association Monograph 25. Auckland, pp. 21-35.
Green, R.C., 2003. The Lapita horizon and traditions – signature for one set of Oceanic migrations. In C.
Sand (ed.), Pacific Archaeology: Assessments and Prospects. Le Cahiers de l’Archéologie en Nouv-
elle-Calédonie 15. Nouméa: Service des Musées et du Patrimoine de Nouvelle-Calédonie, pp. 95-120.
210
Green, R.C. and A.K. Pawley, 1999. Early Oceanic architectural forms and settlement patterns: Linguistic,
archaeological and ethnological perspectives. In R. Blench and M. Spriggs (eds), Archaeology and
Language III: Artefacts, Languages and Texts. London and New York: Routledge, pp. 31-89.
Intoh, M, 1999. Cultural contacts between Micronesia and Melanesia. In J-C. Galipaud and I. Lilley (eds),
The Pacific from 5000 to 2000 BP: Colonisation and Transformation. Paris: Éditions de IRD, pp.
407-22.
Jones, M.D., 2002. Calibration of SE-RF-3 14C Data. Unpublished manuscript University of Auckland.
Jones, M.D. and R.C. Green, in prep. The absolute dating of SE-RF-6 (Ngamanie) and its relative dating
with respect to RF-2 (Nenumbo): Two decorated phase Lapita sites from the Outer Eastern Island
(sic) of the Solomons.
Jones, M.D., F. Petchey, R.C. Green, P. Sheppard and M. Phelan, 2007. The marine ΔR for Nenumbo: A
case study in calculating reservoir offsets from paired sample data. Radiocarbon, 49 (1): 95-102.
Green, R.C., M. Jones and P. Sheppard, 2008. The reconstructed environment and absolute dating of
SE-SZ-8 Lapita site on Nendö, Santa Cruz, Solomon Islands. Archaeology in Oceania, 43: 49-61.
Kirch, P.V., n.d. Test Excavation of a Coastal Midden (SE-SZ-4), Graciosa Bay, Santa Cruz, Eastern
Solomon Islands. Honolulu: Bernice P. Bishop Museum. [Unpublished excavation report].
Kirch, P.V., 1996. Tikopia social space revisited. In Janet Davidson, Geoffrey Irwin, Foss Leach, Andrew
Pawley and Dorothy Brown (eds), Oceanic Culture History: Essays in Honour of Roger Green.
Dunedin: New Zealand Journal of Archaeology Special Publication, pp. 257-74.
Kirch, P.V., 2000. Temples as “Holy Houses”: The transformation of ritual architecture in traditional
Polynesian societies. In R. Joyce and S. Gillespie (eds), Beyond Kinship: Social and Material
Reproduction in House Societies. Philadelphia: University of Pennsylvania Press, pp. 103-14.
Kirch, P.V. and R.C. Green, 2001. Hawaiki, Ancestral Polynesia: An Essay in Historical Anthropology.
Cambridge: Cambridge University Press.
Kirch, P.V. and P. Rosendahl, 1973. Archaeological investigation of Anuta. In D. Yen and J. Gordon (eds),
Anuta: A Polynesian Outlier in the Solomon Islands. Pacific Anthropological Records 21. Honolulu:
Department of Anthropology, Bernice P. Bishop Museum, pp. 25-108.
Kirch, P.V. and D.E. Yen, 1982. Tikopia: The Prehistory and Ecology of a Polynesian Outlier. Bernice. P.
Koch, G., 1971. Materielle Kultur der Santa Cruz-Inseln: Unter Besondere Berücksichtigung der
Riff-Inseln. Neue Folge 21. Berlin: Museum für Völkerkunde.
Leach, B.F., 1985. The source of prehistoric obsidian artefacts from the Polynesian Outlier of Taumako in
the Solomon Islands. New Zealand Journal of Archaeology, 7: 117-23.
Leach, B.F. and J. Davidson, 1988. The quest for the Rainbow Runner: Prehistoric fishing on
Kapingamarangi and Nukuoro Atolls, Micronesia. Micronesica, 21: 1-22.
Lincoln, P., 1978. Reef-Santa Cruz as Austronesian. In S.A. Wurm and Lois Carrington (eds), Second
International Conference on Austronesian Linguistics: Proceedings. Pacific Linguistics C-61.
Canberra: Australian National University, pp. 929-67.
Lynch, J., 1991. Pigs and dogs in Island Melanesia. In A. Pawley (ed.), Man and a Half: Essays in Pacific
Anthropology and Ethnobiology in Honour of Ralph Bulmer. Auckland: The Polynesian Society, pp.
421-32.
Markham, C., 1967. The Voyages of Pedro Fernandez de Quiros, 1595 to 1606 Reprint of the 1904 edition.
Works issued by the Hakluyt Society. Second Series, No. 14-15. Nendeln: Kraus Reprint.
Matisoo-Smith, E. and J. Robins, 2004. Origins and dispersals of Pacific peoples: Evidence from mtDNA
phylogenies of the Pacific rat. Proceedings of the National Academy of Science, 101 (24): 9167-72.
211
McCoy, P.C. and P.C. Cleghorn, 1988. Archaeological investigations on Santa Cruz (Nendö)), southeast
Solomon Islands: Summary report. Archaeology in Oceania, 23: 104-15.
Moir, M., 1989. Mariculture and Material Culture on Takuu Atoll: Indigenous Cultivation of Tridacna Gigas
(Mollusca: Bivalvia) and its Implications for Pre-European Technology, Resource Management, and
Social Relations on a Polynesian Outlier. Ann Arbor: University Microfilms International.
O’Ferrall, W., 1908. Santa Cruz and the Reef Islands. London: Melanesian Mission.
Parker, M., 1981. Vessel Forms of the Reef Island SE-RF-2 Site and Their Relationships to Other
Vessel Forms in Other Western Lapita Sites of the Reef/Santa Cruz and Island Melanesian Area.
Rappoport, A., 1969. House Form and Culture. Englewood Cliffs, N.J.: Prentice-Hall.
Robins, R. and E. Stock, 1990. The burning question: A study of the molluscan remains from a midden on
Moreton Island. In S. Solomon, I. Davidson and D. Watson (eds), Problem Solving in Taphonomy:
Archaeological and Palaeontological Studies from Europe, Africa and Oceania. Tempus No. 2.
Brisbane: Anthropology Museum, University of Queensland, pp. 80-100.
Linguistics, 46 (2): 456-98.
Rye, O., 1981. Pottery Technology: Principles and Reconstruction. Washington: Taraxacum.
Sheppard, P.J., 1992. A report on the flaked lithic assemblage from three Southeast Solomons Lapita sites
(SE-SZ-8, SE-RF-2 and SE-RF-6). In J-C. Galipaud (ed.), Poterie Lapita et Peuplement. Nouméa:
ORTSOM, pp. 145-53.
Sheppard, P.J., 1996. Hard rock: Archaeological implications of chert sourcing in Near and Remote
Oceania. In Janet Davidson, Geoffrey Irwin, Foss Leach, Andrew Pawley and Dorothy Brown (eds),
Oceanic Culture History: Essays in Honour of Roger Green. Dunedin: New Zealand Journal of
Archaeology Special Publication, pp. 99-115.
Sheppard, P.J. and R.C. Green, 1991. Spatial analysis of the Nenumbo (SE-RF-2) Lapita site, Solomon
Islands. Archaeology in Oceania, 26: 89-101.
Snow, B. and R. Shutler Jr, 1985. The Archaeology of Fuga Moro Island. Cebu City: San Carlos
Publications.
Specht, J., 2003. Watom Island and Lapita: Observations on the Reber-Rakival localities. In C. Sand (ed.),
Speiser, F., 1913. Two Years With the Natives in the Western Pacific. London: Mills and Boon Ltd.
Speiser, F., 1916. Völkerkundeliches von den Santa-Cruz Inseln. Ethnologica, II: 153-214. Translated by
Frieda Schütze. [eHRAF Collection of Ethnography Doc. No. 2., 2003, ms. pp.1-131.
http://80-ets.umdl.umich.edu.ezproxy.auckland.ac.nz/cgi/e/ehraf/ehraf-idx?c=ehrafe;view:…]
Spennemann, D., 1993. Cowrie shell tools: Fact or fiction? Archaeology in Oceania, 28: 40-49.
Swadling, P., 1986. Lapita shellfishing: Evidence from sites in the Reef/Santa Cruz Group, Southeast
Solomons. In A. Anderson (ed.), Traditional Fishing in the Pacific. Pacific Anthropological Records
37. Honolulu: Department of Anthropology, Bernice P. Bishop Museum, pp. 137-48.
Szabó, K., 2004. Technique and Practice: Shell-Working in the Western Pacific and Island Southeast Asia.
Three volumes. Unpublished Ph.D. thesis, Australian National University, Canberra.
Tryon, D. and B. Hackman, 1983. Solomon Islands Languages: An Internal Classification. Canberra:
Australian National University.
212
Valentin, F., 2003. Human skeletal remains from the site of Lapita at Koné (New Caledonia): Mortuary and
biological features. In C. Sand (ed.), Pacific Archaeology: Assessments and Prospects. Le Cahiers
de l’Archéologie en Nouvelle-Calédonie 15. Nouméa: Service de Musées et du Patrimoine de Nou-
velle-Calédonie, pp. 285-93.
Waterson, R., 1990. The Living House: An Anthropology of Architecture in South-East Asia. Oxford:
Oxford University Press.
Wurm, S.A., 1969. The linguistic situation in the Reef and Santa Cruz Islands. In A Capell, A. Chowning
and S.A. Wurn, Papers in the Linguistics of Melanesia 2. Pacific Linguistics A-21. Canberra:
Wurm, S.A., 1970. Austronesian and the vocabulary of languages of the Reef and Santa Cruz Islands – A
preliminary approach. In S.A. Wurm and D. Laycock (eds), Pacific Linguistic Studies in Honour of
Arthur Capell. Pacific Linguistics C-13. Canberra; Research School of Pacific Studies, Australian
Wurm, S.A., 1975. The East Papuan phylum in general. In S.A. Wurm (ed.), New Guinea Area Languages
and Language Study, Vol. 1: Papuan Languages and the New Guinea Linguistic Scene. Pacific
Linguistics C-38. Canberra: Research School of Pacific Studies, Australian National University, pp.
783-804.
Wurm, S.A., 1976. The Reef Islands-Santa Cruz family. In S.A. Wurm (ed.), New Guinea Area Languages
and Language Study, Vol. 2: Austronesian Languages. Pacific Linguistics C-39. Canberra: Department
of Linguistics, Research School of Pacific Studies, Australian National University, pp. 637-74.
Wurm, S.A., 1978. Reefs-Santa Cruz: Austronesian, but....! In S.A. Wurm and L. Carrington (eds), Second
International Conference on Austronesian Linguistics: Proceedings. Pacific Linguistics C-61.
Canberra: Australian National University, pp. 969-1010.
Wurm, S.A., 1982. Papuan Languages of Oceania. Tübingen: Gunter, Narr, Verlag.
Wurm, S.A., 1986. Remarks on some language problems in the Santa Cruz archipelago, Solomon Islands.
In B. Elson (ed.), Language in Global Perspective. Papers in Honour of the 50th Anniversary of the
Summer Institute of Linguistics. Dallas, TX: Summer Institute of Linguistics, pp. 507-23.
Wurm, S.A., 1987. Semantics and world view in languages of the Santa Cruz Archipelago, Solomon
Islands. In R. Steele and T. Treadgold (eds), Language Topics: Essays in Honour of Michael
Halliday. Amsterdam and Philadelphia: John Benjamins, pp. 439-51.
Yen, D.E., 1973. Ethnobotany from the voyages of Mendaña and Quiros in the Pacific. World Archaeology,
5: 32-43.
Yen, D.E., 1974. Arboriculture in the subsistence of Santa Cruz, Solomon Islands. Economic Botany, 28: 247-84.
Yen, D.E., 1976a. Inland settlement on Santa Cruz Island (Nendö). In R.C. Green and M.M. Cresswell
(eds), Southeast Solomon Islands Cultural History: A Preliminary Survey. Bulletin 11. Wellington,
New Zealand: Royal Society of New Zealand, pp. 203-24.
Yen, D.E., 1976b. Agricultural systems and prehistory in the Solomon Islands. In R.C. Green and M.M.
Wellington, New Zealand: Royal Society of New Zealand, pp. 61-74.
213
Chapter 10
THE TEOUMA LAPITA SITE, SOUTH EFATE, VANUATU:

A Summary of Three Field Seasons (2004-2006)
Stuart Bedford, Matthew Spriggs, Hallie Buckley,
Frédérique Valentin and Ralph Regenvanu.
INTRODUCTION
The Teouma Lapita site, located on the south coast of Efate Island, Central Vanuatu (Figure 10.1a
and b) was found in January 2004, following a series of serendipitous circumstances (Bedford et
al. 2004, Bedford, Spriggs and Regenvanu 2006). Tectonic uplift, volcanic ashfall and alluvial
deposits have repositioned this once immediately-coastal colonisation period site, situated on an
uplifted reef terrace, to a current location some 800 m from the sea and buried beneath up to 80
cm of black tephra-rich sediment. Earthmoving activity in 2003 removed 1000 m2 of the overlying
black tephra-rich layer and in the process exposed the Lapita site. An initial assessment prior to
excavation recognised it as an important well-preserved early Lapita site, a site type that had been
missing from the Central Vanuatu sequence. Excavations in 2004, however, quickly revealed that
the first use of the site was as a cemetery, and it was therefore not simply important at a local
Vanuatu level, but provided the first real opportunity in more than 50 years of research to describe
a group of Lapita people themselves, and at the same time glean insights into their ritual and
mortuary practices. Three field seasons have been undertaken at the site (2004-2006). Those of
2004 and 2005 have been previously been outlined (Bedford et al. 2006) and will only be further
referred to here to provide background and comparison with results from excavations undertaken
in 2006 (Figure 10.2a-c). Some aspects of the on-going analysis have been reported on in detail
elsewhere, and will be into the future, in collaboration with various specialists, but some general
observations and summary conclusions are outlined in this paper.
ARCHAEOLOGICAL EXCAVATIONS: SITE EXTENT, STRATIGRAPHY AND

CHRONOLOGY
Excavations 2004-2005
A total of 57 m2 were excavated during initial investigations undertaken in July 2004 (Figure
10.2b). The excavations were designed to assess the extent of the damage caused to the site by the
mechanical earthmoving activity and define the boundaries of the site, along with its stratigraphy,
chronology and composition. A primary focus, where areal excavation of c. 24 m2 was undertaken,
was in a part of the site (Area 3) where Lapita sherds and other midden were seen eroding out of
a more deeply machine-mined zone. A further nine test-pits or test-trenches totalling 33 m2 were
excavated across the wider site to try and determine its stratigraphy and extent. At the close of
excavations in 2004 it had been established that the first use of the site had been as a Lapita period
cemetery of considerable extent, with a concentration of 10 burials recovered from Area 3, and a
further three burials identified in Trenches 3 and 3a and Test-pit 4.3.
In 2005 the focus of excavations was a 100 m2 area adjacent and east of Area 3 (Area 2) (Figure
10.2c) where the greatest concentration of burials had been located in 2004 (Bedford et al. 2006).
The northwestern half of the 100 m2 had been scraped by machine almost down to the underlying
reef, facilitating access to the Lapita layers, but also potentially affecting conditions of preservation.
The southeastern part of the area had not been modified. Thirteen burials were uncovered, the
215
Figure 10.1 a. Southwest Pacific and Vanuatu; b. Vanuatu and south Efate (inset).
216
The Teouma Lapita Site, South Efate, Vanuatu
Figure 10.2 a. Excavated areas 2004-2006; b. Excavated areas 2004; c. Excavated area 2005.
cemetery extending only really into the southwestern corner of the 100 m2 area. Two outliers were
found in the rest of the excavated area, an inhumation burial in a solution hole in the far northeast
corner and a possibly later cremation burial in the southeast corner (Valentin et al. in prep. a).
Excavations 2006
The aims of the excavations in 2006 were to complete the investigation of areas that had been
damaged but not destroyed by mechanical scraping, focusing in an area that was considered to be
an extension of the focal burial area for the site and which was likely to contain multiple inter-
ments. Further definition of the area of the cemetery and any associated activity areas, particu-
larly any habitation remains, were also a priority. A laser theodolite was employed during the
course of the excavations in 2006 to facilitate detailed recording of the excavated areas, including
burial features and artefacts, and to complete a topographic map of the wider site.
217
The targeted area of the cemetery, bounded on the west by the north-south aligned Trench 3
and on the east by part of Area 3, both of which had been excavated in 2004, eventually comprised
a contiguous area of some 53 m2 (Figure 10.2a). A further 22 interments were found concentrated
in this area, and as was the case in both 2004 and 2005, they displayed wide variation in burial
position and mortuary practice. Excavations confirmed that this part of the site was more towards
the centre of the cemetery, which probably extends at least as far as Area 4.3. The other nine test-
trenches comprising a total of 65 m2 were both adjacent to the cemetery area and spread across
the terrace (Figure 10.2a). They all contributed to a greater understanding of the overall site, but
at the same time have highlighted the challenges of thoroughly investigating a site that covers a
potential area of more than 10,000 m2.
Area 6B (9 x 1 m) was a northern extension of Area 6, a largely sterile 3 by 2 m test-pit that
had been excavated on the flat part of the terrace in 2004. Area 6B was laid out in a north-south
alignment running down the slope at the edge of the terrace to determine if midden dumping, as
had been identified further west in 2005, was present. Similarly the 8 by 2 m eastern extension of
Area 2 was aligned along the edge of the terrace, to both determine the extent of midden deposits
and if there was any further sign of burials. The combined excavated 26 m2 revealed concen-
trated Lapita and post-Lapita midden deposits but no sign of any further burials, confirming that
the eastern edge of the cemetery had been established in 2005. The Lapita midden deposits
uncovered in this area suggest that this may have been the immediate area of initial habitation at
the site and may even be contemporary with cemetery activity. The boundary of the cemetery was
certainly respected in terms of midden dumping during this period. The concentrated, overlying
post-Lapita midden deposits were contiguous with and paralleled those that had been identified
right across the cemetery area in 2004 and 2005, and relate to later habitation at the site which
spilled over the cemetery area and across the former beach towards the current stream to the
north. Area 6A, a trench measuring 5 by 1 m, just to the south of Area 6, on the edge of the flat
terrace, revealed the familiar stratigraphy of a black tephra-rich soil, overlying a yellow brown
tephra on top of the uplifted reef, but was archaeologically fairly sterile. Sparse, weathered
dentate-stamped and plain sherds and occasional shells were recovered.
To the west of the site, Area 4B, a north-south aligned trench measuring 15 by 1 m, was laid
out in order to further determine the location of the edge of the reef terrace, the extent of the
cemetery and later midden deposits. This trench revealed the edge of the reef terrace but no
further skeletal remains or the later concentrated post-Lapita midden deposits were present. The
stratigraphy paralleled that which had been identified elsewhere but only weathered dentate-
stamped, plain and Erueti-style sherds were recovered. Area 4A (5 by 1 m) to the southeast and
on the flat area of the terrace was similar to Area 6A, showing increasingly sterile deposits,
although occasional weathered dentate-stamped sherds were still present.
Three further trenches were excavated to the southwest of the site, two on the edge of the
terrace in areas that would have once faced across the bay (Area 7A [5 by 1 m] and Area 7B [4
by 1 m]) and another, Area 7C (3 by 1 m), more than 75 m south of the cemetery area, on top of
a low mound feature. These were seen as potential locations for habitation. Areas 7A and B
demonstrated the standard stratigraphy identified across the site but were largely culturally sterile
apart from, again, the presence of weathered dentate-stamped and Erueti-style sherds. Area 7C,
however, revealed evidence of at least two mortuary features apparently associated with scattered
Lapita and early Erueti-style sherds. These burials were left in place for later investigation and
are not included in the burial inventory.
218
A final trench, Area 8A (3 by 1 m), was excavated in the far northeastern edge of the terrace
where sherds had been exposed during fencing activity. This area again proved to have evidence
of Lapita presence albeit only in the form of occasional weathered sherds.
Site Extent and Stratigraphy
The excavations undertaken in 2006 have confirmed some aspects of the site that had been estab-
lished earlier, but have also led to some reinterpretation and a broader understanding of its devel-
opment over time. In all some 275 m2 were excavated at the Teouma site over the 2004-2006 field
seasons. The areal excavations, undertaken in Areas 2 and 3 (177 m2) have provided fine detail
of the Lapita cemetery and the subsequent history of the site in the northern area of the terrace,
while the other trenches and test-pits have provided evidence of site and activity boundaries and
site formation. We are dealing with a site that potentially encompasses an area of more than
10,000 m2, situated on a promontory that during the initial Lapita occupation was bounded to the
north by a small river, by the sea on its western side, and uplifted limestone cliffs extending
several hundred metres to the east and southeast.
Excavations have shown that there is archaeological evidence dating to the Lapita and Early
Erueti Phases, albeit in some cases fairly ephemeral, in all the trenches and test-pits excavated to
date. Thus far we have concentrated investigations on or near the edge of the large terrace where
we would expect habitational evidence. The centre of the terrace appears not to have similar
evidence of occupation and may have been used for gardening activities.
When people first arrived at Teouma Bay they would have been greeted by a wide sheltered
bay. Located on the eastern side was a large flat uplifted karstic reef terrace, with a coral-rubble
beach at its edge and a permanent water source in the form of a stream draining from a swamp in
the nearby eastern interior. Around 3300-3200 B.P., just prior to human arrival, a thick and still
largely unweathered orange/yellow tephra, which has been identified in other areas of Efate
(Bedford 2006, Bedford and Spriggs 2000, Spriggs and Bedford 2001), was deposited across the
uplifted terrace. This created a level surface across the once jagged uplifted reef (Figures 10.3a-c).
The initial use of the area was as a cemetery with burials placed in shallow graves dug into the
tephra in gaps in the uplifted reef and in the upper beach zone. The Lapita cemetery is concen-
trated along a north-east to south-west trending zone, 10-15 metres wide adjacent and parallel to
the former beach. There is potentially a further 200 m2 of the site where burials may be located,
across from Area 3 at least as far as Area 4.3 but presumably no further than Area 4B. Evidence
of initial habitation, possibly contemporary with the life of the cemetery, was identified in the
form of concentrated midden dumping, adjacent and east of the cemetery, in Area 6B and the
east-west aligned trench extension of Area 2. Evidence of Lapita period activity, although sparse,
has been identified in all other excavated areas, most of which were located on or near the edge
of the old reef terrace. Further burials were revealed in the southern-most trench (Area 7C) on a
low mound feature, but not fully investigated due to time constraints. It is possible that they mark
a cemetery associated with the later habitation areas of the site.
Some 50 cm of midden was subsequently deposited across the site; presumably after its
significance or location as a cemetery had been forgotten (Figure 10.3a). It is associated with
habitation activities dating to the Early Erueti phases of Central Vanuatu at c. 2800-2500 B.P.
(Bedford 2006, Bedford et al. 2004, 2006). These later deposits provide insights into changing
settlement pattern and site use over time. They overlie the earlier Lapita deposits but also cover
a significantly larger area of the site. The southern extent of the Erueti deposits parallels that of
the Lapita deposits on the edge of the old reef but extends at least a further 20 m in a northerly
219
Figure 10.3 a. Stratigraphic section showing the black tephra-rich layer (Layer 1), overlying the midden
layer (Layer 2), which has been deposited across the cemetery and uplifted reef, that is seen in
the trench at the rear. Willy Damelip, facing the camera, is seated on top of the former upper
beach; b. Excavation of a Lapita burial in the top of Layer 3, Area 3 (2006). Layers 1 and 2 have
been completely removed across the site but can be seen in the section at the rear; c. Completed
excavation of Area 2, 2005. The uplifted reef is exposed across the site. All photos have been
taken from the northern end of the excavations, looking south.
direction as evidenced in Areas 1, 4.1, 4.2, but also in an easterly and westerly direction as seen
in Area 4.3 and Area 6B. This pattern of deposition, representing realigned and expanding settle-
ment has been recorded at the Arapus/Mangaasi site on the northwest coast of Efate (Bedford
2006, Bedford and Spriggs 2000, Spriggs and Bedford 2001). During occupation of the site there
was on-going tectonic activity which created a progressively prograding shoreline. The preferred
location for habitation remained adjacent and parallel to the contemporary beach, requiring
episodic realignment of the settlement. The continued uplift and alluvial infilling of the bay led
to the abandonment of the settlement as access to the sea became increasingly problematic.
Once abandoned, the site was subsequently capped by a weathered tephra-rich black layer
probably associated with volcanic activity on Nguna Island to the north around 2300 B.P., and
which has again been identified across much of west coast of Efate (Figure 10.3a). Further tephra-
rich soil completes the stratigraphy of the site, and this may relate to further activity on Nguna or
220
indeed to the massive 1452 A.D. Kuwae eruption further to the north (Robin, Monzier and Eissen
1994). The site shows no signs of use or occupation after 2500 B.P. until the development of the
coconut plantation on the site about 100 years ago. There is no sign at Teouma of the late Erueti
or Mangaasi-style pottery (c. 2500-1200 B.P.) that is common in many other archaeological sites
on Efate and the Shepherd Group (Bedford 2006).
Chronology
Multiple radiocarbon dates, utilised in combination with the already established, dated tephrostratig-
raphy, well-defined ceramic chronology from Efate, and regional Lapita ceramic sequences are
providing a detailed and broadly reliable chronology for Teouma. A single conventional date1
(4227±38 B.P. [Wk-17766] 4410-4160 B.P.) on natural shell embedded in growth position in the
crevices of the old reef gives some idea of the period when the reef was initially uplifted above sea
level. People arrived after 3300 B.P., the earliest possible date for the underlying tephra-rich layer
that overlay the reef. Occupation would have begun between 3200 B.P., the date of initial occupa-
tion suggested for the Reefs/Santa Cruz Group to the north (Green, Jones and Sheppard 2008) and
the Makue site on Aore, northern Vanuatu (Galipaud and Swete Kelly 2007), and 3000 B.P. A range
of decorative features and vessel forms associated with the dentate-stamped ceramics from the site,
which are discussed below, also support this date for arrival, as does the recovery of obsidian from
the Bismarck Archipelago to the northwest. As noted earlier, pottery recovered from the midden
dumped across the cemetery belonged to the Early Erueti phases of Central Vanuatu dated to c.
2800-2500 B.P. The site was abandoned no later than 2500 B.P., after which time the late Erueti-
style pottery, not present at Teouma, appears across Efate (Bedford 2006).
Radiocarbon dates thus far have concentrated on attempting to define the chronology of the
cemetery. Four previously published AMS dates (Bedford et al. 2006) included those assayed on
Conus sp. shell rings associated with two burials (3139±36 B.P. [Wk-16831] 2980-2755 B.P.
Burial 4, and 3162±34 B.P. [Wk-15729] 3016-2774 B.P. Burial 11), and two charcoal samples
unidentified as to species. One charcoal sample came from just above the lowest levels of the
midden deposit in Trench 3A (2848±35 B.P. [Wk-15728] 3070-2867 B.P.), and another sample
was retrieved from the more recent Erueti deposits of Trench 3A (2961±36 B.P. [Wk-16830]
3254-3001 B.P.). The last date is not consistent with the associated ceramics that would be
expected to date in the range of 2800-2500 B.P. These results highlight the difficulties of defini-
tively dating the period of use of the cemetery. There are a number of factors that hinder such
precision: the flatness of the calibration curve in this time period (Pearson 1993: Fig. 1b),
variability in Delta R values for marine shells of the region, and the potential problem of in-built
age in unidentified charcoal samples.
However, more recent radiocarbon results from the direct dating of human bone from 20
individuals do suggest that the site was used as a burial place for a time interval of perhaps a
century around 3000 B.P. (Spriggs et al. in prep). The similarity of mortuary practice across the
site and the very limited cases of inter-cutting or disturbance by subsequent burials confirm this
relatively short period of use. They also indicate that people had a level of awareness of the
location of particular burials, perhaps through grave markers, a situation that has some parallels
with the colonisation phase cemetery associated with Wairau Bar in New Zealand (Higham,
Anderson and Jacomb 1999).
221
THE POTTERY ASSEMBLAGE

The discovery of dentate-decorated vessels in direct association with skeletal remains has
confirmed a long-proffered hypothesis that decorated pots had a significant role in Lapita
ceremonial activity. Confirmation of the association of dentate-decorated vessels with the mortuary
rituals at Teouma was established from excavations in 2004. In 2005 this was further detailed when
three whole pots were recovered in amongst and enclosing skeletal remains. One was a complete
flat-bottomed dish that had been up-turned and used as a lid. It lay on top of a dentate-stamped
carinated vessel into which a human skull had been placed (B19). Another inverted carinated
vessel lay nearby alongside another burial (B18: Bedford et al. 2007: 223-40). Evidence of
secondary deposition of skeletal remains in a pot was also identified (Bedford and Spriggs 2007).
During the excavations of 2006 the association of decorated Lapita pots and mortuary ritual
was further confirmed. Another whole carinated vessel and a plain stand were recovered (Bedford
et al. 2007: 223-40) amongst burials. There was also another instance of a pot base containing
assorted human bone being recorded in situ, and associated upper parts of the vessel were scattered
nearby. In many cases discrete concentrations of sherds from other vessels, broken and disturbed
during the use of the site, were recovered over several square metres amongst the burials.
The range of Lapita vessel forms recovered from the cemetery area includes flat-bottomed
dishes, a wide assortment of carinated vessels and cylinder stands. The flat-bottomed dishes, some
15 of which have been identified thus far, range in diameter from 28 to 45 cm and have an
assortment of dentate-stamped designs, including a number with varied anthropomorphic motifs
(Figure 10.4ai, 10.4aii, 10.4bi and 10.4bii). Several of the lips of these dishes are “stepped” and are
excised (Figure 10.4biii and 10.4ciii), a feature also found near the base on some of the dishes. In
addition there are occasional paired pierced holes near the lip (Figure 10.4ci). The dish used as a
lid (10.4ai) was decorated with a double face motif, and had a circular groove in the centre of its
basal exterior, indicating that it was originally made to sit on top of a cylinder stand. Also recovered
from the burial area was a single example of a flat-bottomed dish with attached pedestal foot.
Three separate cylinder stands have thus far been identified at the site, although a further two
to three single sherds may represent others. The difficulty of being able to identify these vessel
forms from small sherds appears to have contributed to the relatively rare positive identification
of such forms from Lapita sites (Best 2002: 82, Kirch 1997: 137), although this is a situation that
is beginning to change (Specht 2007: 60). Their current known geographic range stretches from
the far west of the Lapita distribution in the Bismarck Archipelago to Vanuatu and New Caledonia.
The Teouma examples are decorated with zigzag motifs and a range of anthropomorphic figures
(Figure 10.5i and 10.5ii) including different stylistic forms on the same stand. These tubular vessel
types, which are approximately 30 cm tall, have no decoration on the upper lip, suggesting that
they were indeed used as stands on to which other vessels, such as flat-dishes, were placed.
Carinated vessels dominate the cemetery pottery assemblage and comprise a wide array in
terms of form, size and dentate-stamped designs. Vessel shapes include carinated vessels with
both concave and convex upper body forms associated with outcurving rims (Figure 10.6a and b).
These dominate the assemblage, but there are also a number of carinated vessels with incurving
rims. The double or composite rim form (Sand 2000: Fig 2) is also present. Just to add to the
variation, two of the carinated vessels have flat bases (Bedford et al. 2007: Fig. 7) and one has
modelled birds perched on the inner rim (Bedford and Spriggs 2007). Diameters of these pots
range from the small at c. 15 cm to the very large at c. 55 cm. In some cases the rim diameter is
larger than the carination and in others the carination forms the widest point. The variation in
222
Figure 10.4 Three Teouma flat dishes. ai-bii. Anthropomorphic designs; biii. Excised and dentate-stamped
lip; ci-iii. Flat dish with excised lip, base and pierced holes.
(Design motifs drawn by Richard Shing).
vessel form and size is paralleled in the design motifs, in their construction, application and
positioning. Both fine and coarse dentate decoration is present amongst the recovered vessels.
There may be a chronological aspect to this, but the complete vessels found in association indicate
that complex finely executed dentate designs were being deposited contemporary to vessels
decorated with much coarser designs. This evidence reinforces conclusions reached by others in
terms of gauging chronological variation in Lapita ceramics: that vessel form, decorative finesse,
and design structure and content are not always a definitive marker of chronological divergence
(Chiu 2005: 27, Sand et al. 1998: 41).
223
Figure 10.5 i Cylinder stand (Fidel Yoringmal); ii full motif design (Richard Shing).
Figure 10.6 a. Carinated vessel with double central frieze motif; b. Vessel with convex upper carination; c.
Common motif which is also cut in half and used on the upper part of the carination (d); e.
incised vessel (a, b and e drawn by Fidel Yoringmal; c and d by Richard Shing).
224
Dentate-stamped designs on the carinated vessels include geometric, curvilinear, anthropo-

morphic, labyrinth and “house” motifs along with applied vertical lugs (Bedford et al. 2007:
223-40) and a whole range of other combinations. One of the more common motifs used on all
vessel forms is a variation of Donovan’s M13 (Figure 10.6c; Donovan 1973). There are a number
of cases of central friezes being cut in half and used on the upper part of the carination of the same
vessel (Figure 10.6d). There is also certainly one case and possibly two of the central frieze
comprising two different motifs (Figure 10.6a). There are two cases where decoration continues
just below the carination. Incision is very rare and found only on carinated vessels. A full compi-
lation of vessel forms and design motifs is on-going.
Sherds recovered from the Lapita midden dumping area, east of the cemetery, are as one
would expect from such deposits, much more fragmentary and mixed. There are both fine and
coarse dentate-stamped sherds on the same range of vessel forms as found in the cemetery.
Incised and plain pots are also present in this area. The pottery at the site changes after a hundred
years or so, leading to a dramatic restriction of vessel form and decorative technique. Later vessel
forms of the Early Erueti Phase are dominated by globular out-curving rim vessels with wide flat
lips and very little evidence for decoration, apart from notching on the lip. The only other late
vessel form identified, in much fewer numbers, is a carinated vessel that invariably displays
incised decoration above the carination (Figure 10.6e). These vessel forms and decoration
comprise a restricted sample of the typical Early Erueti Phase ceramics originally identified at the
Erueti site some nine kilometres further east along the coast from Teouma (Garanger 1972: Figs
17-24), and subsequently on the northwest coast of Efate (Bedford 2006).
Limited fabric analysis has been undertaken so far but distinctive black glassy inclusions, a
characteristic of Efate temper, can be seen in many of the sherds (Bedford 2006, Garanger 1972:
110-12). Intriguingly, of the 11 sherds studied thus far, two are exotic to Efate. One is sourced to
New Caledonia and is from a flat-bottomed dish. The other is potentially from the Solomon-
Bismarck region far to the northwest (Dickinson 2006a), and is from the very large carinated
vessel placed upside down next to Burial 18 (Bedford et al. 2007: Fig. 10).
Wider Connections and Implications
Preliminary analysis of the Teouma Lapita ceramics places most of them in the category of
Western or Middle Lapita style, although there are hints of Early/Far Western and Late/Eastern
style ceramics present as well (Anson 1983, Summerhayes 2000). Distinctive traits of the Early/
Far Western Lapita style that are not present at Teouma include cutouts on pedestal bases (Kirch
1997: 121, Summerhayes 2000), open bowls with a curvilinear base (Summerhayes 2000: 34),
and spherical vessels and flasks (Summerhayes 2001: 58). While there are certainly missing
aspects of the Early/Far Western style at Teouma, there is a range of very similar vessel forms and
designs that have links with this earliest of Lapita phases. Very distinctive excised decorations
found both on the “stepped” lip and at the base of flat-bottomed dishes (Figure 10.4b, c) have
affinities with early sites in the Far West including Mussau (Kirch 1987: Fig 4b) and the Arawes
(Summerhayes 2000: 87), but also the SZ-8 site in the Reefs/Santa Cruz Group (Donovan 1973).
In New Caledonia, despite the investigation of some 40 Lapita sites, decoration of this type is
extremely rare with only one or two examples having been found (Sand 2000, Siorat 1990: 81).
It has not to date been found anywhere further east.
The greatest similarities are with ceramics recovered from the Reefs/Santa Cruz Group and
New Caledonian Lapita sites, but comparison has also highlighted regional specificities. None of
the designs on the complete vessels from Teouma, for example, have direct parallels with any of the
whole pots from New Caledonia. Many elements of the full designs, however, are indeed present,
225
and virtually all the motifs present at Teouma can be found in the assemblages from the Reefs/Santa
Cruz Group, which points to some level of chronological correspondence, and a high level of inter-
action during the early phase of expansion into Remote Oceania, something that is further confirmed
by the presence of exotic vessels at the Teouma cemetery. More simplified dentate-stamped designs
and the later incised designs (Figure 10.6e), along with evidence of shell impression, hint at some
level of continued interaction in the region for several hundred years or so.
TEOUMA LAPITA BURIALS

2004-2005
All seasons at Teouma have produced the unexpected, and this is particularly the case in relation
to the burials and associated mortuary practices. In the 2004 and 2005 seasons it was established
that there was a wide range of burial positions, and that mortuary ritual had been a multi-faceted
and on-going lengthy process, rather than a one-off event. The burials were generally placed in
shallow graves dug into the underlying tephra deposits amongst gaps in the uplifted reef and coral
boulders on the upper part of the beach. There is evidence that suggests the manipulation of the
corpse prior to burial or at least during the early stages of decomposition, and that the graves were
repeatedly visited. All the burials had had their skulls, and often many other bones, removed
during the extended mortuary ritual. The majority of the burials were primary interments, although
there are also examples of secondary deposition. There is a wide array of burial positions but most
are laid out in a supine position, often with legs in a flexed or crossed position, possibly due both
to particular burial practices and in some cases the restricted space available in reef cavities. One
of the interments uncovered in 2005 and containing burnt human remains, is consistent with a
cremation (Fuller 2006), adding to the diverse array of burial practices at Teouma.
At the conclusion of the 2004-2005 seasons a total of 26 burials had been excavated but only
four adult crania had been found. All were associated with practices post-dating initial interment. In
one case three skulls were found across the chest of a headless adult skeleton laid in a supine position,
and in another a single skull was found inside a carinated vessel topped with an upturned flat-
bottomed dish (Bedford et al. 2006: Fig. 9, 2007: Figs 3-5). The total lack of in situ skulls with any
of the adult burials, the evidence for removal of skulls after decomposition of the soft tissue, the
group placement of skulls and the identification of a sequence of ritual events occurring over time,
suggests their curation for some period after initial retrieval. There was a total absence of children in
the cemetery, but disturbed infant remains, none older than a few weeks at death, were recovered.
As noted previously, dentate-decorated pottery has been found in association with the
burials, both whole pots adjacent to interments and assorted sherds both near to and within the
grave fills. Some originally whole pots associated with earlier burials were disturbed and spread
amongst the fill of later ones during subsequent burial activities. There are indications that many
of the pots were not completely buried and may have been used to mark individual graves at the
surface. Other associated grave goods are few, but include Conus sp. shell rings near where the
skull would have been, and valves of large mangrove shellfish (Geloina coaxans) on the pelvic
region or limbs. In one case sections of tortoise or turtle carapace were placed over parts of the
body. Finer detail on the burial practices relating to the 2004-2005 excavations at Teouma, set
within a Pacific-wide context, will be found in a forthcoming paper (Valentin et al. in prep. a).
Preliminary analysis of the skeletal remains from the 2004-2005 field seasons has already
returned significant new information relating to the health, morphology, diet and migratory
patterns of this colonising population (Bentley et al. 2007, Buckley 2007, Buckley et al. 2008).
226
The 36 individuals studied to date from Teouma comprise generally robust, large males and
gracile females with high frequencies of dental and degenerative joint disease and pathology
associated with physical activity (Buckley et al. 2008). A relatively high incidence (n = 7/20) of
erosive arthritis was also noted, which has been identified as potentially evidence of gout, indicat-
ing that this significant modern health problem amongst Pacific populations may have a lengthy
history (Buckley 2007). All these data suggest that, although these populations were able to cope
with a significant disease burden, the biological costs of colonisation were not insignificant.
Isotopic analysis of the tooth enamel of 17 individuals has shown that the majority were
local residents who had at least spent their childhood near Teouma and were ultimately buried
there. There were, however, four individuals who were probably immigrants from another coastal
location. A number of aspects suggest that they were also treated differently in death. One of the
individuals was the adult male with three skulls on his chest, and he and two other migrants were
buried in a supine position with their heads towards the south, unlike the locally-raised population
among whom much greater variation in burial orientation was noted (Bentley et al. 2007).
2006
The excavations in 2006 uncovered almost the same number of burials that had been recorded in
2004-2005 field seasons, 22 as opposed to 26. There was an equally surprising array of burial
positions and arrangements of skeletal elements (see Table 10.1 for a full summary). Consistent
with the 2004-2005 findings was the lack of in situ skulls, the removal of a range of other bones,
supine positioning, flexed limbs, secondary deposition in pots, an absence of children or subadults
and a similar range of grave goods, including pottery. The most unusual aspects, unlike anything
that had been found previously, were two examples of arrangements of skeletal elements. In one
case, a cluster of forearm bones, one of the frequently missing elements from earlier graves, was
found sitting on top of arranged scapulae, clavicles and sternum (Figure 10.7a). Three mandibles
sitting adjacent and aligned in a row presented another case (B29) of the appearance of previ-
ously missing elements, underneath a cluster of postcranial bones encompassing the spectrum of
the postcranial skeleton of at least 3 adults (Figure 10.7b). These were all placed on top of a
further burial (B34).
Four infants were added to the range of individuals present at the site, including an individual
that was around 6 months in utero at death and most likely stillborn. The infants were generally
buried individually, but in one case on the shoulder of a female. Limbs were in a flexed position
and unlike the adult burials, the skulls were in situ. A further three adult skulls were found in
2006, associated with another single burial. Two skulls were positioned between the legs and one
to the side of the leg of an individual lying supine with legs outstretched (B44: Figure 10.7ci,
10.7cii). The undisturbed remains of children have still not been found at the site although there
are a few bones of a child aged around 6-10 years of age at death which were included in the pile
of postcranial remains associated with the three crania buried by the legs of Burial 44. Also
included in this pile of postcranial remains was one of the infants.
Another complete decorated pot was found deposited in association with a burial (B49), but
had been disturbed by later activity at the site in the form of a large post-hole originating from
the later midden layer. Structural failure had occurred in the carination area, with the lower part
of the neck shifting inwards while the upper part of the pot collapsed outward (Bedford et al.
2007: Fig. 6), suggesting that the upper part of the pot may have been left projecting above the
ground surface at its deposition. A concentration of small unmodified Cypraea sp. shells was
found in the base of this pot.
227
Table 10.1 Age and sex of individuals from 2004-2006 Teouma field seasons.
228
Figure 10.7 a. Cluster of forearm bones resting on two scapula (Frédérique Valentin excavating); b. three
adjacent mandible sitting beneath a cluster of leg bones and above another burial; ci and cii.
Burial 44 with three skulls placed between (2) and to the side of the legs of one individual.
229
Findings from analyses of the individuals found in 2006 are consistent with those from the
first two field seasons (Buckley et al. 2008), with high levels of degenerative joint disease, poor
dental health, and further evidence of an erosive arthropathy consistent with gout. A study looking
at muscle insertion sites as a way of assessing physical activity in the Teouma population has
found possible evidence of pottery manufacture in the hands of the adults by comparing the
sample with a non-pottery making population (Foster 2007). Publication of these findings is
currently in preparation. Isotopic analyses of the adults suggest that the Teouma population had
consumed a mixed marine and terrestrial diet (Valentin et al in prep.b). Further isotopic analyses
of the infants and females from Teouma suggest that maternal and infant mortality may have been
influenced by maternal ill-health, adding another layer of understanding to the health of these
colonising people (Kinaston et al. in prep).
DISCUSSION AND CONCLUSION

Excavations undertaken in 2006, building on those of 2004-2005, greatly increased our knowledge
of the layout of the cemetery, the complex mortuary rituals, and the population sample at the
Teouma site. A broader understanding of the development of the site over time has also been
achieved. Some 48 burial features representing at least 71 individuals have been excavated,
providing a statistically robust sample of Lapita individuals for the first time. The cemetery at
Teouma is providing insight into the ritual life of the earliest Lapita settlers of Remote Oceania.
The evidence thus far demonstrates that the mortuary ceremony and associated ritual were a
multi-faceted and lengthy process. In many respects these rituals were not dissimilar to contact
burial ritual in Vanuatu and other areas of the Pacific (Speiser 1996: 272-81). More pertinent is
the fact that certain aspects of the burial practices at Teouma, such as the placement of skulls and
other bones in pots and the use of pots in burial ceremony, have close parallels with burial
practices in Neolithic Island Southeast Asia, including Taiwan, during contemporary and slightly
earlier periods (Bellwood 1997).
The site provides the first real detail for Vanuatu of the various vessel forms and elaborate
designs associated with early Lapita settlement of the archipelago. More significantly, the remains
also provide insight into wider issues such as the association of vessel form and use, and the
chronology and pattern of Lapita settlement across the Southwestern Pacific. We suggest that to
have a range of very similar vessel forms and designs at Teouma as some of those found in several
Early/Far Western sites implies rapid movement of people out of the Bismarcks “homeland”
region into western Remote Oceania. The long-postulated “pause” in the Bismarcks region needs
to be reassessed and may need to be further shortened (Bedford and Sand 2007, Specht 2007).
The decorative techniques, motifs and petrography of the pottery indicate some level of regional
interaction throughout the period of the use of the cemetery. There is striking variation in vessel
form and design structure found amongst the ceramics recovered from the cemetery area,
suggestive of a dispersed household mode of production, as opposed to any significant level of
craft specialisation (Clark 2007).
An intriguing question is whether these vessels associated with ritual activity were being
produced in a settlement or settlements nearby, or whether a significant number of them were
transported from afar? Preliminary fabric analysis indicates that the Teouma pottery was mostly
made locally, incorporating sand from the Teouma River. Two sherds analysed by Dickinson
(2006a) were exotic, amplifying results he had previously reported from other Lapita sites (conve-
niently summarised in Dickinson 2006b). Dickinson had earlier found that some Lapita pottery
in Tonga is made from clay and temper from the same source as some of the pottery in the Reefs/
Santa Cruz Islands of the Southeast Solomons. The source is not yet identified but may well be
230
somewhere in Northern Vanuatu on current evidence (Dickinson 2006b: 63,119). One of the
Teouma cemetery pots comes from the same source as some further pottery from the Reef Islands.
The most likely source is thought to be either in the main Solomon Islands or in the Bismarck
Archipelago (Dickinson 2006a, cf. 2006b: 63). Another Teouma pot derived from New Caledonia
and fits a previously-identified pattern (Dickinson 2006b: 115) of the import of occasional pots
from there both to the south of Vanuatu (Erromango Island) and to the north (Malo and Santo
Islands). Further petrographic analysis of all individual vessels is planned.
Prior to the excavation of Teouma (and of Makue on Aore Island in Northern Vanuatu
[Galipaud and Swete Kelly 2007]), it was thought that there was a major Lapita threshold in the
Reefs/Santa Cruz Islands of the Southeast Solomons, the last sites to show considerable exchange
with the Lapita “homeland” in the Bismarcks in terms of obsidian from New Britain sources and
close parallels in Lapita vessel form and design (Green 1976, Green and Kirch 1997, Sheppard
and Walter 2006: 59). The Teouma Project has now extended this interaction sphere down to
Central Vanuatu (Bedford and Spriggs 2008). Only very small quantities of obsidian have been
found in Lapita sites beyond Teouma to the east in Fiji and south in New Caledonia, rather than
the tens to hundreds of pieces now known in northern and central Vanuatu. The idea, (most
recently in Anderson 2006), that Vanuatu was somehow leapfrogged in the settlement of the
further Pacific – in particular Fiji and Tonga – can now be well and truly laid to rest. The Teouma
pottery shows very close parallels to the Reefs/Santa Cruz pottery, and is more complex than all
Fijian and Tongan and all but a very restricted set of New Caledonian Lapita pottery.
The Teouma cemetery represents one of the very rare examples of archaeological remains
of a true pioneering population of the first one or two generations of settlement of an island group
anywhere in the world. Many questions of course remain, and one of the more challenging is
determining the existence or not of a settlement contemporary with the use of the cemetery. This
will be a focus of excavations to be undertaken in 2008 and 2009 funded by the Australian
Research Council (DP0880789). The Teouma site is providing a whole new impetus for Lapita
research, leading to new interpretations and layers of understanding of Lapita settlement and
society both in Vanuatu and the wider region.
ACKNOWLEDGMENTS
We dedicate this paper to the late Willy Nemen Damelip who passed away in December 2007. He
was a key member of innumerable excavations carried out in Vanuatu by Bedford and Spriggs
since 1996, including the three seasons at Teouma. Colleague, friend, guide and raconteur
extraordinaire, you are sadly missed. The Teouma 2004-2006 Archaeological Project was a joint
initiative of the Vanuatu National Museum and The Australian National University (ANU),
directed by Matthew Spriggs and Stuart Bedford of the ANU and Ralph Regenvanu, Director of
the Vanuatu Cultural Centre until the end of 2006. Funding of the project during 2004-2005 was
provided by the Pacific Biological Foundation, the Department of Archaeology and Natural
History, and School of Archaeology and Anthropology at the ANU, the Snowy Mountains
Engineering Corporation Foundation, and Mr Brian Powell. Excavations in 2006 were funded
primarily through a National Geographic Scientific Research Grant (8038-06), with further
support from an ARC Discovery Grant (DP0556874). Thanks must also be extended to our
collaborators in the field, Jacques Bolé, Rachel Fuller, Stuart Hawkins, Andrew Hoffmann, Iarawoi
Philip, Richard Shing, Nancy Tayles, Martha Yamsiu and Fidel Yoringmal. Students who partici-
pated in 2006 from the University of the South Pacific, Fiji, were Tony Heorake, Bronwyn Oloni,
Michael Leodoro, Tristelle Kavae and Anne Tosiro. The laboratory research and travel for Hallie
Buckley were funded by The Royal Society of New Zealand Marsden Fund and a University of
231
Otago Research Grant. Laboratory research and travel costs for Frédérique Valentin were
primarily provided by CNRS. Christophe Sand provided valuable comment on the Teouma
pottery and comments on this paper. The support of the leaseholder M. Robert Monvoisin and
family is acknowledged, as is the support and assistance of the VCC fieldworker for the area,
Silas Alben, and the traditional landowners and community of Eratap village. William Dickinson
analysed the petrography of the Teouma sherds discussed in this paper, Fidel Yoringmal and
Richard Shing produced the pottery illustrations and Mikael Michelland Figure 10.7ci.
ENDNOTES.
The radiocarbon dates are presented in the following format: date, sample number and calibrated
age at two standard deviations using the Calib. Program REV 5.0.1 (Reimer et al. 2004, Stuiver,
Reimer and Reimer 2005). A Delta R value of 45±19 was applied to shell samples (Petchey,
Phelan and White 2004).
REFERENCES
Anderson, A.J., 2006. Crossing the Luzon Strait: Archaeological chronology in the Batanes Islands,
Philippines and the regional sequence of Neolithic dispersal. Journal of Austronesian Studies, 1 (2):
25-44.
Anson, D., 1983. Lapita Pottery of the Bismarck Archipelago and its Affinities. Unpublished Ph.D. thesis,
University of Sydney, Sydney.
Bedford, S., A. Hoffmann, M. Kaltal, R. Regenvanu and R. Shing, 2004. Dentate-stamped Lapita reappears
on Efate, central Vanuatu: A four decade-long drought is broken. Archaeology in New Zealand, 47
(1): 39-49.
Bedford, S. and C. Sand, 2007. Lapita and Western Pacific settlement: Progress, prospects and persistent
problems. In S. Bedford, C. Sand and S.P. Connaughton (eds), Oceanic Explorations: Lapita and
Bedford, S. and M. Spriggs, 2000. Crossing the Pwanmwou: Preliminary report on recent excavations
adjacent to and southwest of Mangaasi, Efate, Vanuatu. Archaeology in Oceania, 35: 120-26.
Bedford, S. and M. Spriggs, 2007. Birds on the rim: A unique Lapita carinated vessel in its wider context.
interaction and the emergence of the ‘ethnographic present’. Asian Perspectives, 47: 95-120.
Bedford, S., M. Spriggs, R. Regenvanu, C. Macgregor, T. Kuautonga and M. Sietz, 2007. The excavation,
conservation and reconstruction of Lapita burial pots from the Teouma site, Efate, Central Vanuatu.
In S. Bedford, C. Sand and S.P. Connaughton (eds), Oceanic Explorations: Lapita and Western
Pacific Settlement. Terra Australis 26. Canberra: ANU E Press, pp. 223-40.
Bellwood, P., 1997. Prehistory of the Indo-Malaysian Archipelago. Honolulu: University of Hawaii Press.
Bentley, A., H. Buckley, M. Spriggs, S. Bedford, C. Ottley, G. Nowell, C. Macpherson and D. Pearson,
2007. Lapita migrants in the Pacific’s oldest cemetery: Isotopic analysis at Teouma, Vanuatu.
American Antiquity, 72 (4): 645-56.
Auckland.
232
Buckley, H.R., 2007. Possible gouty arthritis in Lapita – Associated skeletons from Teouma, Efate Island,
Central Vanuatu. Current Anthropology, 48: 741-49.
Buckley, H.R., N. Tayles, M. Spriggs and S. Bedford, 2008. A preliminary report on health and disease
in early Lapita skeletons: Possible biological costs of colonization. Journal of Island and Coastal
Archaeology, 3 (1): 87-114.
Chiu, S., 2005. Meanings of a Lapita face: Materialized social memory in ancient house societies. Taiwan
Journal of Anthropology, 3: 1-47.
Clark, G.R., 2007. Specialisation, standardisation and Lapita ceramics. In S. Bedford, C. Sand and S.P.
Connaughton (eds), Oceanic Explorations: Lapita and Western Pacific Settlement. Terra Australis
26. Canberra: ANU E Press, pp. 266-89.
Dickinson, W.R., 2006a. Petrography of Sand Tempers in Prehistoric Potsherds from Selected Islands in
Central and Northern Vanuatu, New Hebrides Island Arc. Unpublished Petrographic Report WRD-255.
Dickinson, W.R., 2006b. Temper Sands in Prehistoric Oceanian Pottery: Geotectonics, Sedimentology,
Petrography, Provenance. Special Paper 406. Boulder: The Geological Society of America.
Donovan, L.J., 1973. Inventory of Design Elements and Motifs in Lapita Reef-Santa Cruz Island Pottery.
Unpublished appendices to M.A. research essay, University of Auckland, Auckland.
Foster, A., 2007. Pottery Production and Markers of Occupational Stress: Musculoskeletal Stress Markers
and Osteoarthritis in Two Pacific Island Skeletal Samples. Unpublished B.A. Honours dissertation,
University of Otago, Dunedin.
Fuller, R., 2006. A Study of the Non-Burial Human Bone from the Lapita Site of Teouma, Vanuatu.
Dissertation submitted as partial fulfillment for a Post-Graduate Diploma of Science, University of
Otago.
Galipaud, J-C. and M.C. Swete Kelly, 2007. Makué (Aore Island, Santo, Vanuatu): A new Lapita site in
the ambit of New Britain obsidian distribution. In S. Bedford, C. Sand and S.P. Connaughton (eds),
Oceanic Explorations: Lapita and Western Pacific Settlement. Terra Australis 26. Canberra: ANU E
Press, pp. 151-62
Garanger, J., 1972. Archéologie des Nouvelles-Hébrides: Contribution à la Connaissance des îles du
Centre. Publications de la Société des Océanistes, No. 30. Paris: ORSTOM.
Green, R.C., 1976. Lapita sites in the Santa Cruz Group. In R.C. Green and M.M. Cresswell (eds),
Southeast Solomon Islands Cultural History: A Preliminary Survey. Bulletin 11. Wellington: Royal
Society of New Zealand pp. 245-65.
Green, R.C., M. Jones, and P.J. Sheppard. 2008. The reconstructed environments and absolute dating of
SE-SZ-8 Lapita site on Nendo, Santa Cruz, Solomon Islnds. Archaeology in Oceania, 43:49-61.
Green, R.C. and P.V. Kirch, 1997. Lapita exchange systems and their Polynesian transformations: Seeking
explanatory models. In M. Weisler (ed.), Prehistoric Long-Distance Interaction in Oceania: An
Interdisciplinary Approach. New Zealand Archaeological Association Monograph 21. [Dunedin],
pp. 19-37.
Higham, T., A. Anderson and C. Jacomb, 1999. Dating the first New Zealanders: The chronology of Wairau
Bar. Antiquity, 73: 420-27.
Kinaston, R., H.R. Buckley, S.E. Halcrow, M. Spriggs and S. Bedford, in prep. Investigating Mortality and
Infant Feeding in Prehistoric Skeletal Samples: A case Study from a 3000 Year Old Pacific Island
Cemetery Site.
Kirch, P.V., 1987. Lapita and Oceanic cultural origins: Excavations in the Mussau Islands, Bismarck
Pearson, G.W., 1993. High precision bidecadal calibration of the radiocarbon time scale 500-2500 BC.
Radiocarbon, 35 (1): 25-33.
233
Petchey, F., M. Phelan and J.P White, 2004. New ΔR values for the Southwest Pacific Ocean. Radiocarbon,
46 (2): 1005-14.
Reimer, P.J., M.G.L Baillie, E. Bard, A. Bayliss, J.W. Beck, C. Bertrand, P.G. Blackwell, C.E. Buck, G.
Burr, K.B. Cutler, P.E. Damon, R.L. Edwards, R.G. Fairbanks, M. Friedrich, T.P. Guilderson, K.A.
Hughen, B. Kromer, F.G. McCormac, S. Manning, C. Bronk Ramsey, R.W. Reimer, S. Remmele,
J.R. Southon, M. Stuiver, S. Talamo, F.W. Taylor, J. van der Plicht and C.E. Weyhenmeyer, 2004.
Intcal04 Terrestrial Radiocarbon Age Calibration, 0–26 Cal Kyr BP. Radiocarbon, 46: 1029-58.
Robin, C., M. Monzier and J-P. Eissen, 1994. Formation of the mid-fifteenth century Kuwae caldera (Vanuatu)
by an initial hydroclastic and subsequent ignimbritic eruption. Bulletin of Volcanology, 56: 170-83.
Sand, C., 2000. The specificities of the ‘Southern Lapita Province’: The New Caledonian case.
Sand, C., K. Coote, J. Bole and A. Ouetcho, 1998. A pottery pit at locality WKO013A, Lapita (New
Caledonia). Archaeology in Oceania, 33: 37-43.
Siorat, J-P., 1990. A technological analysis of Lapita pottery decoration. In M. Spriggs (ed.), Lapita Design,
Form and Composition. Occasional Papers in Prehistory 19. Canberra: Department of Prehistory,
Specht, J., 2007. Small islands in the big picture: The formative period of Lapita in the Bismarck
Archipelago. In S. Bedford, C. Sand and S.P. Connaughton (eds), Oceanic Explorations: Lapita and
Speiser, F., 1996. Ethnology of Vanuatu. An Early Twentieth Century Study. Bathurst: Crawford House Press.
Spriggs, M. and S. Bedford, 2001. Arapus: A Lapita site at Mangaasi in Central Vanuatu. In G.R. Clark,
Spriggs, M., . F. Petchey, F. Leach, S. Bedford, H. Buckley and F. Valentin. in prep. Direct Dating of Lapita
Skeletons from the Teouma Cemetery Site.
Stuiver, M., P.J. Reimer and R.W. Reimer, 2005. Calib. REV 5.0.1 manual, http://calib.qub.ac.uk/calib/manual/.
Summerhayes, G.R., 2000. Lapita Interaction. Terra Australis 15. Canberra: Department of Archaeology
and Natural History and the Centre for Archaeological Research, Australian National University.
Summerhayes, G.R., 2001. Lapita in the Far West: Recent developments. Archaeology in Oceania, 36: 53-64.
Valentin, F., S. Bedford, H. Buckley and M. Spriggs. in prep. a. Inhumations, Exhumations and Bone
Treatment in a Lapita Community as Reflected at the Teouma Burial Ground, Vanuatu.
Valentin, F., H. Buckley, E. Herrscher, R. Kinaston, S. Bedford, M. Spriggs, S. Hawkins And K. Neal
in prep. b. Lapita Subsistence Strategies And Food Consumption In The Community Of Teouma
(Efate, Vanuatu): Insights From Stable Isotope Analyses.
234
Chapter 11
Understanding the place properly:

Palaeogeography of Selected Lapita Sites in the Western
Tropical Pacific Islands and its Implications
Patrick D. Nunn and Tony Ahikau Heorake
INTRODUCTION
The island groups of the western tropical Pacific include both large high islands with often conti-
nental affinities, such as are found widely in Papua New Guinea, New Caledonia and Solomon
Islands, and smaller lower islands of wholly oceanic origin, such as most in Fiji, Samoa and
Tonga. Within the Holocene (last 10,000 years), these islands experienced the effects of sea level
changes of tens of metres, changes that have not proved easy to isolate because of both the
tectonic instability of many parts of the region and the spatially-variable pattern of Holocene sea
level change (Dickinson 2001, Nunn 1994). In particular, there was once considerable debate
(reviewed by Nunn 1995) as to whether or not this region experienced a higher-than-present sea
level during the mid-Holocene. It is now accepted that it did so, with sea level in the western
low-latitude Pacific having attained a maximum of perhaps 2.1 m around 4200 cal yr B.P.
(Grossman, Fletcher and Richmond 1998, Nunn and Peltier 2001).
The times at which Lapita people colonised islands of the western tropical Pacific (approxi-
mately 3400-2900 cal B.P.) coincided with a time of slowly falling sea level, one effect of which
was to transform Pacific Island coastlines, both their terrestrial and offshore components
(Dickinson 2001, Nunn 1994). It has recently been suggested that the fall of sea level below a
particular threshold transformed coastal environments in this region so profoundly that it led to
the end of the Lapita culture at approximately the same time in every part of the Lapita realm
(Carson 2008, Nunn forthcoming).
Comparatively little attention has been paid to the reconstruction of Lapita-era environments
in this region. Noteworthy exceptions include Frimigacci (1980) and Gosden and Webb (1994)
and the synthesis of Lepofsky (1988). Many accounts of Lapita settlements and settlement
patterns have apparently assumed that Lapita palaeoenvironments were similar to modern
environments, and that the reasons why particular sites were selected for settlement by the Lapita
people can be explained by the configuration of modern environments, particularly their coast-
lines, river systems, anchorages, and coral reefs. In line with more recent, more enlightened work
by archaeologists in the tropical Pacific (such as Sheppard and Walter in press), we argue that
post-settlement environmental changes, brought about by a range of interacting human and
non-human driven processes, have profoundly altered many coasts, disguising their character at
the time of colonisation and potentially misleading many investigators trying to understand the
reasons behind Lapita site selection.
Following a brief discussion of environmental influences on Lapita-era environments, this
paper looks at several case study areas, presenting new palaeoenvironmental reconstructions for
many, and then a new analysis of the factors that Lapita people considered to be paramount in site
selection. We have tried to use examples where local environmental factors (like uplift and subsid-
ence) have been clearly subordinate to regional factors (principally sea level change and reef
upgrowth) so that these examples could be considered as representative rather than singular.
235
CLIMATE AND SEA LEVEL IN THE WESTERN TROPICAL PACIFIC

DURING THE LAPITA ERA
Mean sea level was higher by as much as 2.1 m during the mid-Holocene, about 4200 B.P., and
began falling shortly afterwards until about 1200 B.P. when it reached its present level (Grossman
et al. 1998, Nunn 1995, Nunn and Peltier 2001). Shore platforms, both erosional and reefal,
developed at low tide level along tropical Pacific coasts around the time of the sea level maximum
and, when sea level fell subsequently, these platforms emerged to provide foundations for the
coastal plains favoured by many colonists (Nunn 1994). The rates of late Holocene sea level fall
varied within the tropical Pacific, meaning that coastal plains in some island groups became
available for occupation before those in others, a factor that may explain, in part at least, the
episodic nature of island colonisation across the entire Pacific (Dickinson 2003, Nunn 2009).
The association of coral reef growth with Holocene sea level changes is a critical issue to
consider when trying to understand Lapita colonisation chronologies and settlement location
choices. Prior to the mid-Holocene sea level highstand, sea level had been generally rising for
10,000 years or more, as land-grounded ice melted in response to rising temperatures and entered
the ocean. The rate of this postglacial sea level rise varied, with shorter faster rises (the catastroph-
ic rise events [CREs] of Blanchon and Shaw 1995) occasionally punctuating the overall slower
rise. During the late Quaternary in much of the western tropical Pacific, coral reefs began flour-
ishing again only when ocean-surface waters had warmed sufficiently: in general, by around
13000 B.P. Yet there was also considerable geographical variation. For example, in New
Caledonia postglacial coral reef growth began only about 7200 B.P. while in neighbouring
Vanuatu it was apparently continuous from 24000 to 6000 B.P. (Cabioch 2003).
Not all reefs that existed 13,000 years ago, when sea level was about 70 m lower, exist today.
Some were drowned, simply unable to grow upwards at the same rate as sea level rose, particu-
larly during the rapid bursts that punctuated the slower, more usual rise. The upgrowth of most
coral reefs in the low-latitude Pacific lagged behind rising sea level, only to “catch up” with it at
a later stage, typically after it began to fall during the later Holocene (4200-1200 B.P.). Only a
few reefs in optimal oceanographic locations were able to “keep up” with rising sea level during
the early Holocene (Neumann and Macintyre 1985, Nunn 1994, 1999).
By the time the Lapita people reached the islands of the western tropical Pacific east of the
Bismarcks, there were evidently some coral reefs at the ocean surface (low tide level) but probably
far fewer than there are today. Furthermore, those that did exist may have been mostly compara-
tively ephemeral and lacking the biodiversity of their modern counterparts (Kan et al. 1995,
Montaggioni, 2005). In much of the Fiji archipelago for example, there is evidence that most coral
reefs did not “keep up” with rising Holocene sea levels but only “caught up” during the later
Holocene, commonly during post-settlement times (Nunn 1990). Thus, when the Lapita people first
entered the Fiji archipelago, they encountered far fewer reefs, yet, desiring reef foods, plausibly
chose to settle where they found the broadest ocean-surface (rather than submerged) reefs.
At the time of arrival of the Lapita people in Remote Oceania, about 3000 B.P., sea level
was still about 1.5 m above its present level (from Nunn and Peltier 2001) and most ocean-surface
coral reefs in large archipelagoes like most in the western tropical Pacific would have been found
along those windward (generally east- or southeast-facing) coasts with no obstruction to the
southeast to prevent long-fetch waves from reaching the shore.
Reef-fringed coasts may also have been favoured sites for initial Lapita settlement in the
western tropical Pacific because of both the associated morphology of nearshore areas and the
236
Palaeogeography of selected Lapita sites in the western tropical Pacific Islands and its implication
sediment that they produced. At low tide today, the shallow reef-enclosed lagoon is a generally
benign environment for people collecting marine foods, one that owes both its existence and its
biological productivity to the presence of ocean-surface reefs offshore. Reef-produced sediment
is critical to the growth and maintenance of many sandy shorelines in the Pacific Islands, and it
is likely that Lapita colonists favoured coastal locations that were not eroding, which, in other
words, were sustained by reef-derived sediment.
The following discussion identifies several places occupied by the Lapita people for which
sufficient information is available to enable the palaeoenvironment of the site about the time of
its initial occupation to be reconstructed. Several examples of smaller-island sites are discussed.
Larger-island sites appear to be comparatively few but the possibility that post-Lapita landscape
processes – more active on larger than smaller islands – may have effectively obscured and/or
removed the evidence for Lapita occupation in such locations cannot be everywhere discounted
(Spriggs 1984). We then distinguish two generic types of site environment that appear to have
been especially common – sand spit sites and tombolo sites. Both types of site are vulnerable by
modern yardsticks governing the choice of coastal settlement, but perhaps these were not those
that prevailed in Lapita times. These issues are revisited in the Implications section.
SMALLER-ISLAND COASTAL SITES

Most known Lapita sites occur on smaller islands. Some of the single smaller islands considered
here occur just offshore a larger island, others are more isolated. The latter may represent an
opportunist colonisation, the former perhaps the start of a sub-regional colonisation – a gateway
site (Clark 1999).
There are many reasons why a smaller island off a larger island might have been selected
for initial settlement by a group of people unfamiliar with a particular region. They may have
considered the offshore island safer, not knowing who or what inhabited the larger. It has been
suggested that an association of larger wetter islands with malaria may have been a reason why
Lapita people generally avoided settling them (Kirch 1997, Serjeantson and Gao 1995). In more
pragmatic terms, the reef resources of the smaller island, being in generally less turbid, more
saline and better aerated ocean water, may have been greater per unit area and more favoured by
shore or stilt house dwellers. Smaller islands may also have been favoured by colonising peoples
carrying food animals such chickens, rats and dogs, so that feral populations established subse-
quently would be easier to capture than on larger islands.
Talepakemalai, Mussau Islands, Papua New Guinea

The Talepakemalai (ECA) and Etakosarai (ECB) Lapita sites in the Mussau Islands of Papua New
Guinea involved two stilt-island and onshore settlements that were established in association with
two islands; now they are all part of Eloaua Island (11 km2). Their palaeogeography was described
by Kirch (2001a) and is shown in Figure 11.1.
The earliest occupation at Talepakemalai about 3500 B.P. involved both stilt houses and
onshore houses. The later Lapita occupation here about 2800 B.P. was transformed through
coastal progradation, attributable to both sedimentation associated with sea level fall and human
activities, and involved only stilt houses: land areas appear to have been utilised solely for horti-
culture. Both stilt villages here were centrally located on the contemporary reef flat, suggesting
that reefal foods played a critical role in the subsistence economy of the Lapita people here. It is
possible that, as with Lapita stilt house occupations elsewhere, these ones did not spread out
across a reef flat into which it may have been difficult to sink posts, but were confined to places
237
Figure 11.1 Settlements at Etakosarai and Talepakemalai on Eloaua Island, Mussau Islands, Papua New
Guinea during the period of Lapita occupation 3500-2800 B.P. (after Kirch 2001a).
where this was covered by a layer of sand thick enough for this purpose. Many other stilt house
occupations – at Boliu (Papua New Guinea) and Bourewa (Fiji), for example, – appear to have
been built along submerged sand spits or tombolos (Kirch 2001b, Nunn 2007a).
Lapita Sites SZ-10 and SZ-50, Tömotu Noi Island, Temotu Group,
Solomon Islands
The Mbanoimba site (SZ-10) and another Lapita site (SZ-50) on Tömotu Noi Island (18 km2) are
only a few kilometres from the large Nanggu site on larger Nendö Island for which post-Lapita
environmental changes are described below.
Tömotu Noi is an emerged limestone island. Much has been revealed about its Lapita
environments by the coring of Lake Luedambu (Powell 1976) on the shore of which SZ-50 is
located. Powell’s work shows that during Lapita times, Lake Luedambu was part of a lagoon that
was connected to the sea; the lake is now fresh-brackish and separated from the rest of the lagoon
by a 250 m-long “coral bridge” (1976: 89), probably an emerged reef. This identification is
critical because it suggests, within the limits of the available data, that the Lapita occupants of
SZ-50 had ready access to both coral reef resources and to lagoonal resources. The association of
SZ-50 with these aspects of the Tömotu Noi palaeoenvironment was made by Green, Jones and
Sheppard (2008).
Naitabale, Moturiki Island, Fiji

In Fiji, a Lapita settlement was established on the offshore island Moturiki (10.9 km2 in area)
perhaps around 2950-2850 cal yr B.P. on the coastal flat named Naitabale (Nunn et al. 2007).
Owing to subsequent sea level fall and shoreline progradation, the Lapita site is now nearly 300
m inland from the shoreline. Yet at the time of its establishment, it was located at the mouth of
the Mataloaloa River on the inland side of a shore-parallel beach ridge that extended east into a
sand spit (Figure 11.2).
238
Figure 11.2 a. The landscape of the Naitabale area, Moturiki Island, central Fiji, about 900 B.C. when it was
occupied by the Lapita people. The Lapita beach ridge was actively forming and ran eastwards into
a sand spit. The Lapita settlement, its likely extent shown by the broken line, was in the lee of the
beach ridge at the head of a short estuary. b. The modern landscape of the Naitabale area showing
the location of the pits (sizes exaggerated) excavated on and around the ancient beach ridge (the
“Lapita beach ridge”). The entire area is now around 300 m inland of the present shoreline.
The Naitabale site is on a southeast-facing, long-fetch wave-affected coast, and it is likely

that 3000 years ago coral reef had developed at the low tide level here earlier than along many
other coasts in Fiji. The reef flat, then around 1.2 km broad, that adjoined the site is likely to have
been the main reason for site selection. This is reflected by the huge numbers of shellfish
evidently consumed by the Lapita occupants of Naitabale, although they may also have been
attracted by the environments at the head of the then-estuary of the Mataloaloa, particularly if
there was a mangrove forest there as seems possible, given its sheltered location and estuarine
environment (Nunn et al. 2007).
Lolokoka, Niuatoputapu Island, Tonga

The initial Lapita settlement at Lolokoka on Niuatoputapu Island (4 km2) was located on the north
coast of the island. It was backed by cliffs of emerged Pleistocene reef and adjoined on its
northwest side a 1.8-km broad sub-tidal reef flat that evidently played an important role in
sustaining the island’s earliest human inhabitants. Lolokoka is thought to have been a beach site,
but it also abutted a lagoon on its northeast side (Kirch 1988). This is evidence that the Lapita
people sought a range of marine-subsistence options.
While it cannot be certainly ascertained, it seems likely that a reef had developed around
much of the Niuatoputapu coast at the time of its earliest occupation more than 3000 yrs B.P. In
this regard its is worth mentioning that the oceanographic criteria described earlier that are likely
to control reef growth in archipelagoes are subordinate on smaller isolated islands like
Niuatoputapu to factors like the existence of a suitable submarine shelf for reef upgrowth. As a
result of sea level fall since the island’s initial occupation, most of the reef surrounding the
239
Pleistocene core of Niuatoputapu has emerged, increasing island area by 312% yet reducing reef
area by some 50% (Dickinson, Burley and Shutler 1999, Kirch 1988).
LARGER-ISLAND COASTAL SITES

Throughout the Lapita region, most of the earliest sites located and described appear to have been
founded on smaller islands (<50-200 km2 in area depending on context). Whether this represents
a real preference for smaller-island environments or whether it reflects reduced site visibility
arising from the generally greater amount of post-Lapita landscape change along larger-island
coasts (Spriggs 1984) remains a key issue, raised in the Implications section below. Yet critical to
the understanding of this question is chronology, specifically the establishment ages of smaller-
island and larger-island Lapita settlements. It might be, for instance, that the earliest settlements
were established on smaller islands where reef resources were generally more abundant but that,
after a short time, linked (daughter) settlements were established along neighbouring larger-island
coasts where a broader subsistence economy could be developed. A lack of knowledge about all
the Lapita settlements in a particular archipelago and the lack of reliable dates for their establish-
ment are two issues that hinder the complete evaluation of such ideas.
Palaeogeographical reconstructions of the selected larger-island Lapita sites described in
this section show that they were all established on larger-island coasts. Some others, like those on
the Foué Peninsula in New Caledonia and in the Bourewa area of southwest Viti Levu Island in
Fiji, are today part of larger islands yet at the time of their establishment by Lapita peoples appear
to have been located on smaller offshore islands. These two examples are therefore described
elsewhere in this paper.
Nanggu, Nendö Island, Temotu Group, Solomon Islands

The Nanggu (SZ-8) Lapita site lies on the southeast coast of Nendö [Santa Cruz] Island (500 km2).
It now lies nearly 500 m inland and 5-6 m above high tide level, something that is likely to be a
result of both post-occupation sea level fall and uplift; the island of Nendö experienced consider-
able differential block movement during the late Holocene (Hughes, Craig and Daniels 1981).
Coring of southeast Nendö, where the Nanggu site is located, suggests that it is underlain by
swamp facies, suggesting that it was a stilt house occupation along the fringe of an ocean-linked
lagoon or coastal embayment that was not fringed, as it is today, by coral reefs (Green et al. 2008).
More than that, it seems clear from the palaeogeographic reconstruction suggested by these
authors that Nanggu was located on a narrow sand spit separating a brackish-water lagoon from
the ocean, a location identical to many other Lapita settlements (see below).
Natunuku, Viti Levu Island, Fiji

The Lapita site at Natunuku, described by Davidson et al. (1990), is located on the north coast of
Viti Levu Island (10,388 km2) in Fiji, the largest in the group. The site lies on the northeast-facing
coast of the 10 m high neck of a bedrock promontory. Today it looks out across a broad fringing
reef with large areas of mangrove within one kilometre of the site (Figure 11.3).
Around the time of initial Lapita settlement here, perhaps 2800 B.P., the neck of the promon-
tory was much narrower and it was surrounded by fringing reef. Most of the subsequent lowland
growth and spread of mangrove forest has been a result of post-Lapita sea level changes and
human impact on island landscapes, particularly those of the interior within the last millennium
(Nunn 2005). Although probably not one of the earliest Lapita settlements in the Fiji archipelago,
240
Figure 11.3 a. Palaeogeography of the Natunuku area of northern Viti Levu Island, Fiji, around the time of
initial settlement 2800 B.P, reconstructed from field survey and assuming that mean sea level
was 1.25 m higher than present (Nunn 2005, Nunn and Peltier 2001). b. Modern geography of
the Natunuku area showing the modern villages of Natunuku and Vatutavui and the Lapita
settlement (after Nunn 2005).
the Natunuku site is one of the earliest known to have been established on the largest island, Viti
Levu, and as such may mark a developmental stage involving the diversification of the Lapita diet
into larger-island wild foods and horticulture.
Fanga ‘Uta Lagoon, Tongatapu Island, Tonga

The likely founder settlement in the Tonga Islands was established 2900-2800 cal yr B.P. (950-850
B.C.) at Nukuleka (discussed separately under “Sand Spit Sites” below) on the eastern shore of
the Fanga ‘Uta Lagoon that dominates the north coast of Tongatapu Island (260 km2 in area:
Burley and Dickinson 2001). Lapita settlements were eventually established at numerous points
around the shores of this lagoon (Figure 11.4). It seems clear that these settlements were estab-
lished with optimal access to food resources in mind.
When we look at the reconstruction of the Lapita landscape in Figure 11.4a, it is clear that
about half the Lapita settlements within the Fanga ‘Uta Lagoon were in locations marginal
between sea and land: either on sand spits or islands. Thus we could infer that the Lapita occupa-
tion of Tongatapu Island was not a genuine occupation of a large island but one that was filtered
through perhaps several transitional (sea to land) phases of settlement: perhaps stilt houses above
lagoon waters, offshore islands, mangrove islands and sand spits.
It is inferred from the lack of Lapita sites around the contemporary Folaha Peninsula
(Figure 11.4b) that this area was then fringed by comparatively unproductive sand (not sand-
coated reef) flats.
Fourteen Lapita settlements have been discovered along the coast of the Lapita-era lagoon
or on smaller islands within it (Burley et al. 2001). Today the lagoon is almost wholly enclosed,
both access and water circulation being reduced by this and the screen of offshore islands (see
241
Figure 11.4 a. Palaeogeography of the Fanga ‘Uta Lagoon, Tongatapu Island, Tonga (after Burley et al. 2001)
around 2850 cal yr B.P. when sea level was perhaps 2 m higher with the location of all known
Lapita settlements and a suggested reconstruction of the nature of offshore areas. It is likely that
waves had caused sediment deposition around the Folaha Peninsula reducing the amounts of
marine foods found here. Conversely, healthy reef extended close to the shore in the west and
east at this time which explains why Lapita coastal settlements are concentrated in these areas.
b. Modern geography of the Fanga ‘Uta Lagoon. Note the closed nature of much of the lagoon,
the coasts of which are generally moderately populated.
Figure 11.4b). Around the time of Lapita colonisation, the situation was quite different (see Figure
11.4a). The “lagoon” was actually a pair of separate coastal inlets, both readily accessible, in
which water circulation was undoubtedly conducive to reef health that attracted the earliest
people to establish settlements at the closest points on the shore. Lapita diets in Tonga, at least for
the first 100-200 years, are likely to have been dominated by shellfish and other marine foods
obtainable from fringing reefs (Spennemann 1987). The numbers of Lapita-era shell middens
around the Fanga ‘Uta Lagoon and their degree of compaction are remarkable, suggesting that
shellfish collection was the principal reason for settlement establishment here (Burley et al.
2001). The main shellfish genus found here, Anadara, depends on detrital foods produced in
mangrove ecosystems (Amesbury 2007), suggesting that these were present during Lapita times
in the Fanga ‘Uta Lagoon.
SAND SPIT SITES

Irrespective of the size of the island on which they were located, several Lapita sites were built
on sand spits (or tombolos – discussed in the following section), mostly submerged, at least at
high tide. This is an observation of considerable significance, allowing insights into the Lapita
world view that have not always been apparent. For instance, most modern humans might rate the
land a superior option for settlement than offshore areas or transitional locales such as (submerged)
sand spits and tombolos.
It is likely that some Lapita sites, for which palaeogeography cannot be effectively recon-
structed at present (given the published information available), were also founded on sand spits.
These include the Mbanoimba site (SZ-10) on Tömotu Noi Island, and SZ-23 and SZ-45 on
Tömotu Neo Island, both in the Temotu Group of Solomon Islands (Lepofsky 1988), possibly
several of the uplifted sites on Efate and Malo Islands in Vanuatu (reviewed by Bedford 2003),
and the Wakea site on Lakeba Island, Fiji (Best 1984).
242
Figure 11.5 Photo of part of the Lapita site at Bourewa, Viti Levu Island, Fiji, following its excavation in
February 2008. The earliest Lapita occupation was in stilt houses along an underwater sand spit
on the top of which a shell midden accumulated. The higher sea level meant that there was a tidal
inlet (or brackish-water lagoon) on the landward side of this sand spit. One wonders why the
adjacent hills were apparently eschewed by the Lapita people for settlement.
Sand spits appear to have been an important landscape of choice for Lapita settlements, which
raises the question why Lapita people chose such places – vulnerable to multiple environmental
stresses – when more secure sites were available close by, typically just inland (Figure 11.5).
Bourewa, Viti Levu Island, Fiji

The likely founder settlement in Fiji, dating from perhaps about 3050 cal B.P., is located along
the southwest coast of Viti Levu Island at the place called Bourewa (Nunn 2007a). Bourewa today
is one of the longest windward beaches along this part of the coast and adjoins an uncommonly
broad (up to 2.7 km) reef flat. It is part of an area of limestone bedrock (the Rove Peninsula)
delimited by the ocean and two distributaries of the Tuva River, which today is joined to the main
part of Viti Levu only by the beach barrier at Natadola (Figure 11.6).
Geomorphological mapping of the area allows reconstruction of the form of the island
around 3050 cal yr B.P. when the sea level was approximately 1.3 m higher (Nunn and Peltier
2001). The available radiocarbon dates (in Nunn 2007a) suggest that Bourewa was the founder
site and that subsequently Lapita settlement spread to main-island locations to the southeast, and
to the smaller offshore islands, Qoqo and Rove (Figure 11.6b).
Around 3000 years ago, the offshore island, which today forms the Rove Peninsula, was
fringed along its windward (southeast-facing) side by a broad coral reef that was probably an
attractive target for Lapita voyagers. Coupled with the fact that this location is where the barrier
243
Figure 11.6 a. Palaeogeography of the Bourewa area, southwest Viti Levu Island, Fiji (after Nunn 2005)
approximately 2700 cal yr B.P. when sea level was about 1 m higher (calculated from Nunn and
Peltier 2001). The largest offshore island (now joined to the Viti Levu mainland – see map (b)
had at least four Lapita settlements including the large one at Bourewa regarded as the likely
founder site in Fiji, dating from 3210-2920 cal yr B.P. (1260-970 B.C.). Lapita settlements were
also established on the smaller islands Rove and Qoqo, probably surrounded by reef at the time,
but not on Likuri and Navo islands, inferred to have been isolated from ocean-surface reef at
the time. Most data from unpublished field survey. b. Modern geography of the Bourewa area.
Note the broad fringing reef and the extensive mangrove swamps in north of the area. The land
on which Vusama Village is located is joined to the Viti Levu (island) mainland only by the
barrier beach at Natadola.
reef enclosing the smaller (drier) islands of southwest Fiji (Mamanuca and Yasawa Groups) first
meets land, there is strong inferential evidence for the earliest human settlements in Fiji being
along the Rove Peninsula, something supported by the cluster of early radiocarbon ages
3210-2850 cal yr B.P. (Nunn 2007a).
Navutulevu, Viti Levu Island, Fiji

The coastal village of Navutulevu on the south coast of Viti Levu Island, Fiji, is located on a 3-4 m
high beach ridge that is joined at its eastern end to a coastal plain situated between two high
sedimentary bedrock cliffs. Along its western margin, a thriving mangrove community exists where
several streams drain from an inland marsh out into a small estuary. Extending southwards is an
offshore coral-fringing reef about 2 km broad bisected by a large sea passage – Navutulevu Pass.
The Navutulevu site represents the easternmost evidence yet obtained of a Lapita presence
on Viti Levu Island (Kumar, Nunn and Dickinson 2004). Preliminary observations there suggest
that during the mid-Holocene (4-5000 years ago), a sand spit was created at the entrance to a
shoreline embayment. When the Lapita colonists of Navutulevu arrived at perhaps 2850 years
B.P., sea level had dropped from its Holocene maximum to about 1.15 m above present level
(Nunn 2005, Nunn and Peltier 2001), and they established their settlement on the sand-spit. This
location gave them ready access to both the nearshore reefs and the fresh-brackish lagoon.
244
Subsequent entrapment and accumulation of terrigenous sediments continued to extend the

coastal plain and fill the palaeobay. The coral reefs that lagged behind during the mid-Holocene
(e.g., Naquaquavarua, Tabati and Cakaunimona reefs) also “caught up” with the ocean surface about
the time of Lapita occupation, leading to progradation of the coastal lowlands. Terrigenous materials
washed into the lagoons were trapped by the fringing reefs and the gradually emerging sea floor.
Nukuleka, Tongatapu Island, Tonga

The earliest-known settlement in the Kingdom of Tonga is at Nukuleka at the entrance to the
Fanga ‘Uta Lagoon on Tongatapu Island. Both the radiocarbon ages and the ceramic associations
suggest that Nukuleka was the founder settlement in the Tonga archipelago from which popula-
tions subsequently dispersed (Burley and Dickinson 2001).
The Lapita settlement at Nukuleka was located on a sand spit at the entrance to the Fanga ‘Uta
Lagoon adjoining, according to the reconstruction in Figure 11.4a, a broad coral reef as well as an
area of non-reefal coastal inlet. In this sense it matches what is emerging as a clear site preference of
Lapita settlers for sites that allowed optimal access to both reef and low-saline lagoon food sources.
The modern village of Nukuleka occupies a coastal promontory that represents the emerged
and enlarged sand spit on which the original settlement was founded. While the site may conceivably
have emerged because of uplift, the degree of emergence can be satisfactorily explained by the net
sea level fall of around 1 m that took place after Lapita arrival at Nukuleka about 2850 cal yr B.P.
Faleloa, Foa Island, Ha‘apai Islands, Tonga

At the northern end of Foa Island in the Ha‘apai Group of Tonga there is a Lapita site where the
modern village Faleloa now lies (Dickinson, Burley and Shutler 1994). Although the modern
village is on a broad tombolo, the Lapita village was probably on a sand spit that had grown out
from the southwest coast of the northern island and bordered an open-water passage between
what were then two islands (Figure 11.7). It may be that this spit was chosen for Lapita settlement
because the passage was a place where fish, driven in one or both directions with the tide, were
relatively easy to catch. Yet it also seems likely that fringing reef on either side of the island either
continued through this passage (as shown in Figure 11.7a), and thus made the site suitable for safe
sheltered gathering of reef foods at low tide, or that only the northwest end of the passage was
fringed by a broad reef, explaining why this was the favoured end.
Since Lapita times, emergence attributable largely to sea level fall of some 1.0-1.8 m have
caused the sand spits to extend across the passage and eventually to fill it completely, forming a
tombolo (Figure 11.7b). A younger progradational beach complex marks this emergence along the
northwest end of the former passage, which explains why the Faleloa Lapita site is now some 200
m inland (Dickinson et al. 1994).
TOMBOLO SITES
A tombolo is a sand spit that links two larger, higher pieces of land. From the standpoint of the
settlement choices of the Lapita people, a tombolo would have been little different from a sand
spit (see previous section), given its similar composition and environmental context. Where it
does differ is in doubling the proximal opportunities for true terrestrial settlement which, in every
case known to the authors, appear to have been rejected.
For some Lapita sites, it is difficult to know for certain whether they were founded on sand
spits (promontories) or tombolos. The distinction may not have unduly influenced potential
245
Figure 11.7 a. Palaeogeography of part of northern Foa Island, Ha‘apai Group, Tonga during its early Lapita
occupation (after Dickinson et al. 1994) showing the distribution of fringing reef and the
nascent sand spits that later developed into a tombolo here. b. Modern geography and geomor-
phology of the area. The modern village named Faleloa is believed to have been built on the
approximate site of its earliest counterpart.
Lapita settlers, as both sand spits and tombolos allow settlements to be located on dry land (or at
least covered by only shallow slow-moving water) with optimal access to reef/nearshore resources
on both their ocean and land-facing sides.
Boliu Island, Mussau Islands, Papua New Guinea

Modern Boliu Island, which comprises the northern and southern parts shown in Figure 11.8
together with the emerged sand spit, has a total surface area of less than 1 km2. The two Lapita
settlements established there were both comprised of stilt houses (Kirch 2001b).
The one in the north (EKF) lies today between the mangrove fringe and the limestone escarp-
ment marking the island’s north coast, but at the time of Lapita settlement it is inferred to have
occupied a gap in the mangrove fringe through which there was ready access to the reef flat (Figure
11.8). Since that time, sea level fall and associated changes in nearshore sediment dynamics have
seen the mangrove fringe extend seawards beyond the remains of the EKF stilt village.
246
Figure 11.8 Stilt house settlements off Boliu Island (north), Mussau Islands, Papua New Guinea, dating
from its period of Lapita occupation beginning around 3350-3150 cal yr B.P. (interpreted
from Kirch 2001b).
The stilt village (EKE) in the south of Boliu Island during late (and post-) Lapita times
extended along the axis of a submerged sand spit (tombolo?) that emerged subsequently as a
result of sea level fall. In Lapita times, the spit would have provided a more easily workable
foundation for house posts than the sand-free reef flat, but also the location of EKE may simply
have been the optimum one for access to the most productive reef flats, or a place where seine
nets could be lain easily across the reef flat (Kirch 2001b).
Foué, La Grande Terre, New Caledonia

Several Lapita sites, including the eponymous one of Lapita (site 13), are found along the sandy
coasts of the Foué Peninsula on the west coast of La Grande Terre (New Caledonia). Figure 11.9
shows the modern situation and the palaeogeographic reconstruction which represents three of the
sites as having been located along a low sand tombolo connecting the main island to what was
then an offshore island (Foué Island).
Today the three sites along the tombolo are being eroded out of a sandy strip of land 2-3
metres high which connects the Foué Peninsula with the rest of La Grande Terre. On the south
side there is a lagoon with a reef offshore, on the north a sand-fringed mangrove swamp. The
placement of the three sites in the palaeogeographic reconstruction suggests that their location on
the tombolo gave their inhabitants optimal access to the resources of an estuarine inlet to the north
and an open ocean coast to the south. It is possible sand has buried a reef that was closer to the
south shore of the tombolo in Lapita times. These changes could be explained by the late
Holocene sea level fall of 1.0-1.5 m together with the associated infilling of the estuary to the
north of the tombolo and the burial of any productive nearshore ecosystem to its south.
247
Figure 11.9 a. Palaeogeography of part of western New Caledonia around during the period of early Lapita
settlement. b. Modern geography of the Foué Peninsula and adjacent parts of La Grande Terre,
New Caledonia (after Sand 1997).
Matanamuani, Naigani Island, Fiji

The long-known Lapita site at Matanamuani (VL 21/5) on Naigani Island (2 km2) is probably of the
same vintage as Naitabale (see above), just 25 km distant, although possibly slightly younger (Best
1981, 2002). The site was established on a sand tombolo connecting two islands along the eastern
(windward) side of which a fringing coral reef had recently formed (Figure 11.10a). The precise
location of the site allowed optimal access to the windward reef and possibly a leeward mangrove
swamp, as well as utilising the shelter afforded by the proximity of the two high islands.
No uplift is necessary to explain post-Lapita landscape changes in Naigani (Figure 11.10b),
only the drop in sea level of around 1.2 m and an increase in lowland sedimentation resulting from
a period of intensive occupation of the island’s interior during the last millennium (see Figure 6.9
in Nunn 2007b).
Qoqo Island, Viti Levu Island, Fiji

Located in a large mangrove swamp at the mouth of the Tuva River, Qoqo Island (40,000 m2)
today comprises two rounded hills reaching more than 20 metres above sea level connected by a
20-50 metres broad coastal flat, interpreted as a tombolo (Figure 11.11). Reconstruction of the
form of the tombolo suggests that it was perhaps 15 metres wide in Lapita times, submerged at
high tide, and that a Lapita stilt house occupation straddled the tombolo proper, extending onto
the adjacent reef flat (Nunn et al. 2006).
Marine (reef-flat) shellfish appear to have been important components of Lapita diets at
Qoqo and point to the fact, confirmed by other work in the vicinity, that at the time of this Lapita
occupation, the island was surrounded by coral reef which was converted to mangrove as much
as 2000 years later, following sharply increased sediment loads of the Tuva River. The recovery
of around 5% of freshwater shellfish (principally Batissa violacea) from the midden at this site
implies that its Lapita occupants occasionally utilised the river for subsistence purposes, although
the reef was the generally preferred target.
248
Figure 11.10 a. Reconstruction of the environment of Naigani Island, central Fiji, during its initial Lapita
occupation about 900 B.C. (after Nunn 2007b). The settlement at Matanamuani is located in
a tombolo. b. Modern environment of Naigani Island.
Figure 11.11 The island of Qoqo in the estuary of the Tuva River, southwest Viti Levu Island, Fiji (after
Nunn et al. 2006). The form of the tombolo that was occupied during Lapita times is shown.
Note that the mangrove swamp that surrounds the island today was a recent (post-Lapita)
environmental change; in Lapita times, the island was fringed with coral reef.
249
It appears clear that in the case of Lapita settlement on tombolos, not even a doubling of the
immediate possibilities of true terrestrial settlement could tempt the Lapita people away from
such vulnerable locations. This became clear to the authors’ team when they were excavating the
Lapita site on Qoqo and found no trace of any settlement on the two higher areas (former islands
and their coasts) in contrast to the abundant evidence for Lapita settlement on the 15-metre wide
tombolo that connected them in Lapita times.
This raises an important point, namely that because there are numerous examples of tombolo
occupations by the Lapita people, we can reasonably infer that they had good reasons for
eschewing true terrestrial settlement – at least during their earlier settlement period – in favour of
what appears to us today to be a unnecessarily exposed location. This is discussed further in the
following section.
IMPLICATIONS
A complete discussion of the factors influencing Lapita site selection is premature, given that
palaeoenvironmental information is not available (and cannot be readily obtained) for many
Lapita sites. But the information obtained to date, some of which is presented above, does allow
some general points to be made about the nature of Lapita settlement choices throughout the
Lapita realm.
• Smaller islands, whether alone or “tethered” to a larger-island site, were favoured settlement
locations.
• Most (early Lapita-era) sites are located along those windward coasts where coral reef is likely
to have reached low-tide level and would therefore provide an important source of food.
• Most sites were located so as to allow optimal access to marine foods on reefs and in
low-saline lagoons.
• Most (early Lapita-era) sites were stilt-house occupations built on sand spits or tombolos that
were awash at high tide.
• No (early Lapita-era) sites are found inland of the contemporary coastline.
It appears clear that most early period Lapita settlements were located on smaller islands,
something that has only become manifest in some places since palaeogeographic reconstruction
has shown that areas that are now parts of larger islands were formerly smaller offshore islands.
We would argue that the choice of smaller over larger islands for settlement was largely a
pragmatic one that became a cultural trait and thus a tradition. The pragmatic drivers of this
choice probably related largely to food resources, for example:
• the superior extent, condition, and bioproductivity of coral reefs fringing smaller rather than
larger islands,
• the superior boundedness of smaller islands, meaning that resources (both naturally occurring
and introduced) could be more readily controlled than they could be on larger islands, and
• the generally healthier situation of smaller islands, particularly those in windward locations,
rather than larger islands where diseases associated with slow-moving or stagnant water,
heavy rainfall, and disease-carrying insects are more common.
250
But palaeogeographic reconstruction also allows other insights. For while in some places, it
seems probable that Lapita people from a smaller island subsequently colonised a larger island
nearby, there are also examples of situations where one smaller island after another was colonised
and a nearby larger island avoided. This also emphasises the importance of the above factors, and
probably also signals their continuing importance throughout much of Lapita-era history.
Also we draw attention to the well-known (but under-discussed) fact of Lapita preferences
for settlement in places that are transitional between ocean and land, such as (submerged) sand
spits and tombolos. We need to answer the question as to why the Lapita people chose – through-
out the Lapita realm – to build complex high-maintenance structures (stilt house settlements) when
a short distance away they could have lived in far simpler dwellings on land (Nunn 2007c).
It is possible that the past tendency to recognise that Lapita palaeoenvironments were often
not very different from their modern counterparts may have led to errors regarding the nature of
Lapita subsistence. For example, most of the palaeoenvironmental analyses reported in this paper
show that Lapita settlements were closer to the sea than the areas in which they are found are
today. Consequently it seems reasonable to posit that the degree to which Lapita people depended
on marine foods may have been underestimated in some studies and that, conversely, their depen-
dence on horticulture and other land-based subsistence overstressed. One could follow this point
with the idea that, because most Lapita settlements are found on the land today, it has been widely
assumed – in fact taken as indisputable – that the Lapita people sought to reside on the land rather
than on the water or over it. In support of the possibility that the latter is correct might be ranged
the presence of over-water stilt house occupations throughout the Lapita realm, the likelihood that
(at least in Remote Oceania) there is no evidence for horticulture associated with the earliest
occupations, and various models that imply it was sea nomads rather than inter-island voyagers
who were the Lapita colonisers (Di Piazza and Pearthree 1999, Nunn 2007a, 2007c, 2009,
Sheppard and Walter, this volume).
CONCLUSIONS
There are manifest dividends to be paid to those who study past cultures in the Pacific Islands and
elsewhere by considering palaeoenvironments. We contend that it will be possible to gain consider-
able insights into Lapita lifeways once we have a better and more complete knowledge of the
palaeoenvironments in which their sites were located. We would encourage Lapita archaeologists
to routinely gather the data that are needed to allow reconstruction of Lapita palaeoenvironments.
Within this general call, we would also highlight the need for better chronologies of settle-
ment spread in some areas. Much emphasis is naturally placed by Lapita archaeologists on estab-
lishing the time when people first settled a particular key location, but less on determining when
they moved to adjacent locations. Such settlement chronologies would allow better insights into
Lapita lifeways than we have at present because it would tell us when movement took place,
allowing us to speculate about why it did.
ACKNOWLEDGEMENTS
We thank Christophe Sand for help with the Lapita-era reconstruction of the Foué Peninsula,
David Burley and William R. Dickinson for information about sites in Tonga, and Peter Sheppard
for a preprint of his paper (with Roger Green and Martin Jones) on Nanggu. Comments from two
anonymous referees improved the manuscript greatly.
251
REFERENCES
Amesbury, J.R., 2007. Mollusk collecting and environmental change during the prehistoric period in the
Mariana Islands. Coral Reefs, 26: 947-58.
Bedford, S., 2003. The timing and nature of Lapita colonisation in Vanuatu: The haze begins to clear. In
Nouvelle-Calédonie 15. Nouméa: Service de Musées et du Patrimoine de Nouvelle-Calédonie,
pp. 147-58.
Best, S., 1981. Excavations at Site VL 21/5 Naigani Island, Fiji: A Preliminary Report. Auckland:
Best, S., 1984. Lakeba: The Prehistory of a Fijian Island. Ph.D. thesis, University of Auckland. Ann Arbor:
University Microfilms International.
Auckland.
Blanchon, P. and J. Shaw, 1995. Reef drowning during the last deglaciation: Evidence for catastrophic
sea-level rise and ice-sheet collapse. Geology, 23: 4-8.
Burley, D.V., 1998. Tongan archaeology and the Tongan past, 2850-150 B.P. Journal of World Prehistory,
12: 337-92.
Burley, D.V. and W.R. Dickinson, 2001. Origin and significance of a founding settlement in Polynesia.
National Academy of Sciences, Proceedings, 98: 11829-831.
Burley, D.V., W.R. Dickinson, A. Barton and R. Shutler Jr, 2001. Lapita on the periphery: New data on old
problems in the Kingdom of Tonga. Archaeology in Oceania, 36: 89-104.
Cabioch, G., 2003. Postglacial reef development in the south-west Pacific: Case studies from New
Caledonia and Vanuatu. Sedimentary Geology, 159: 43-59.
Carson, M.T., 2008. Correlation of environmental and cultural chronology in New Caledonia.
Geoarchaeology, 23: 695-714.
Clark, G.R., 1999. Post-Lapita Fiji: Cultural Transformation in the Mid-Sequence. Unpublished Ph.D.
Davidson, J., E. Hinds, S. Holdaway and F. Leach, 1990. The Lapita site of Natunuku, Fiji. New Zealand
Journal of Archaeology, 12: 121-55.
Di Piazza, A. and E. Pearthree, 1999. The spread of the ‘Lapita people’: A demographic simulation. Journal
of Artificial Societies and Social Simulation, 2 (3) – online at jasss.soc.surrey.ac.uk/2/3/4.html
Dickinson, W.R., 2001. Paleoshoreline record of relative Holocene sea levels on Pacific Islands.
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: 489-502.
Dickinson, W.R., D.V. Burley and R. Shutler Jr, 1994. Impact of hydro-isostatic Holocene sea-level change
on the geologic context of island archaeological sites, Northern Ha‘apai Group, Kingdom of Tonga.
Geoarchaeology, 9: 85-111.
Dickinson, W.R., D.V. Burley and R. Shutler Jr, 1999. Holocene paleoshoreline record in Tonga:
Geomorphic features and archaeological implications. Journal of Coastal Research, 15 (3): 682-700.
Frimigacci, D., 1980. Localisation éco-géographique et utilisation de l’espace de quelques sites Lapita de
Nouvelle-Calédonie. Journal de la Société des Océanistes, 36 (66-67): 5-11.
Gosden, C. and J. Webb, 1994. The creation of a Papua New Guinean landscape: Archaeological and geo-
morphological evidence. Journal of Field Archaeology, 21: 29-51.
Green, R.C., M.D. Jones and P.J. Sheppard, 2008. The reconstructed environmental and absolute dating of
SE-SZ-8 Lapita site on Nendö, Santa Cruz, Solomon Islands. Archaeology in Oceania, 43: 49-61.
252
Grossman, E.E., C.H. Fletcher and B.M. Richmond, 1998. The Holocene sea level highstand in the
equatorial Pacific: Analysis of the insular palaeosea-level database. Coral Reefs, 17: 309-27.
Hughes, G.W., P.M. Craig and R.A. Dennis (eds), 1981. Geology of the Eastern Outer Islands. Solomon
Islands Geological Survey Bulletin 4: 1-33. Honiara.
Kan, H., N. Hori, Y. Nakashima and K. Ichikawa, 1995. The evolution of narrow reef flats at high-latitude
in the Ryukyu Islands. Coral Reefs, 14: 123-30.
Kirch, P.V., 1988. Niuatoputapu: The Prehistory of a Polynesian Chiefdom. Thomas Burke Memorial
Washington State Museum Monograph 5. Seattle: The Burke Museum.
Kirch, P.V., 2001a. Three Lapita villages: Excavations at Talepakemalai (ECA), Etakosarai (ECB), and
Etapakengaroasa (EHB), Eloaua and Emananus Islands. In P.V. Kirch (ed.), Lapita and its Transfor-
mations in Near Oceania: Archaeological Investigations in the Mussau Islands, Papua New Guinea,
1985-88. Contributions of the University of California Archaeological Research Facility, No. 59.
Berkeley: University of California pp. 68-145.
Kirch, P.V., 2001b. The Boliu site (site EKE). In P.V. Kirch (ed.), Lapita and its Transformations in Near
Oceania: Archaeological Investigations in the Mussau Islands, Papua New Guinea, 1985-88.
Contributions of the University of California Archaeological Research Facility, No. 59. Berkeley:
University of California, pp. 162-74.
Kumar, R., P.D. Nunn and W.R. Dickinson, 2004. The emerging pattern of earliest human settlement in Fiji:
Four new Lapita sites on Viti Levu Island. Archaeology in New Zealand, 47: 108–17.
Lepofsky, D., 1988. The environmental context of Lapita settlement locations. In P.V. Kirch and T.L. Hunt
(eds), Archaeology of the Lapita Cultural Complex: A Critical Review. Seattle: Burke Museum, pp.
33-47.
Montaggioni, L.F., 2005. History of Indo-Pacific coral reef systems since the last glaciation: Development
patterns and controlling factors. Earth Science Reviews, 71: 1–75.
Neumann, A.C. and I. MacIntyre, 1985. Reef response to sea-level rise: Keep-up, catch-up or give-up.
Proceedings of the 5th International Coral Reef Congress, 3: 105-10.
Nunn, P.D., 1990. Coastal geomorphology of Beqa and Yanuca islands, South Pacific Ocean, and its
significance for the tectonic history of the Vatulele-Beqa Ridge. Pacific Science, 44: 348-65.
Nunn, P.D., 1994. Oceanic Islands. Oxford: Blackwell.
Nunn, P.D., 1995. Holocene sea-level changes in the South and West Pacific. Journal of Coastal Research
[Special Issue], 17: 311-19.
Nunn, P.D., 1999. Environmental Change in the Pacific Basin: Chronologies, Causes, Consequences.
London: Wiley.
Nunn, P.D., 2005. Reconstructing tropical paleoshorelines using archaeological data: Examples from the
Fiji Archipelago, southwest Pacific. Journal of Coastal Research, 21 [Special Issue] 42: 15-25.
Nunn, P.D., 2007a. Echoes from a distance: Progress report on research into the Lapita occupation of the
Rove Peninsula, southwest Viti Levu Island, Fiji. In S. Bedford, C. Sand and S.P. Connaughton
(eds), Oceanic Explorations: Lapita and Western Pacific Settlement. Terra Australis 26. Canberra:
ANU E Press, pp. 163-76.
Nunn, P.D., 2007b. Climate, Environment and Society in the Pacific During the Last Millennium.
Amsterdam: Elsevier.
Nunn, P.D., 2007c. Space and place in an ocean of islands: Thoughts on the attitudes of the Lapita people
towards islands and their colonization. South Pacific Studies, 27: 24-35.
Nunn, P.D., 2009. Vanished Islands and Hidden Continents of the Pacific. Honolulu: University of Hawai‘i
Press.
253
Nunn, P.D., forthcoming. Geographical influences on settlement-location choices by initial colonizers: a

case study of the Fiji Islands. Geographical Research.
Nunn, P.D., T. Ishimura, W.R. Dickinson, K. Katayama, F. Thomas, R. Kumar, S. Matararaba, J. Davidson
and T.H. Worthy, 2007. The Lapita occupation at Naitabale, Moturiki Island, Central Fiji. Asian
Nunn, P.D., S. Matararaba, R. Kumar, C. Pene, L.Yuen and M.R. Pastorizo,, 2006. Lapita on an island
in the mangroves? The earliest human occupation at Qoqo Island, southwest Viti Levu, Fiji.
Archaeology in New Zealand, 49: 205-12.
Nunn, P.D. and W.R. Peltier, 2001. Far-field test of the ICE-4G (VM2) model of global isostatic response
to deglaciation: Empirical and theoretical Holocene sea-level reconstructions for the Fiji Islands,
southwest Pacific. Quaternary Research, 55: 203-14.
Powell, J.M., 1976. Pollen, plant communities and prehistory in the Solomon Islands. In R.C. Green and
M.M. Cresswell (eds), Southeast Solomon Islands Cultural History: A Preliminary Survey. Bulletin
11. Wellington: Royal Society of New Zealand, pp. 75-105.
Sand, C., 1997. The chronology of Lapita ware in New Caledonia. Antiquity, 71: 539-47.
Serjeantson, S.W. and X. Gao, 1995. Homo sapiens is an evolving species: Origins of the Austronesians.
In P. Bellwood, J.J. Fox and D. Tryon (eds), The Austronesians: Historical and Comparative
Perspectives. Canberra: Department of Anthropology, Research School of Pacific and Asian Studies,
Sheppard, P.J. and R. Walter, in press. Inter-tidal late Lapita sites and geotectonics in the western Solomon
Islands. This volume.
Spennemann, D.H.R., 1987. Availability of shellfish resources on prehistoric Tongatapu, Tonga: Effects of
human predation and changing environment. Archaeology in Oceania, 22: 81–96.
Spriggs, M., 1984. The Lapita cultural complex: Origins, distribution, contemporaries and successors.
Journal of Pacific History, 19: 202-23.
254
Chapter 12
RETURN TO THE ENTANGLED BANK:

Deciphering the Lapita Cultural Series
John Edward Terrell
INTRODUCTION
In his acclaimed novel The Oxford Murders, the Argentinean writer and mathematician Guillermo
Martínez engagingly shows how easy it is to hide the truth from others by getting them to think
that a series of similar events—in this instance, a series of murders—is happening because, when
taken in sequence, they appear to add up to a coded message that we are being taunted to decipher.
Judging by appearances, each murder apparently symbolises one of the logical steps in a predict-
able sequence, just as most of us would probably agree that the next logical number in the familiar
series 2, 4, 8, and 16 must be the multiple 32. Perhaps, but as the philosopher Ludwig Wittgenstein
famously observed, any finite sequence of numbers can be continued in a variety of different
ways, not just in the one way that may seem reasonable (Biletzki and Matar 2006). For example,
the narrator of The Oxford Murders, whose name we are never told, is asked early in the novel if
he can figure out what is the next symbol in the odd series reproduced here as Figure 12.1a.
Figure 12.1 Alternative possible solutions to Martínez’s cryptic symbol series.
Although Martínez never shows us the solution he has in mind (the narrator merely tells us
later on that the answer is the number series 1, 2, 3, 4), we suspect those who find riddles like this
one appealing are likely to say the solution shown in Figure 12.1b is the right resolve: an answer
simply derived from the rules of symmetry (Figure 12.1c). Yet in keeping with Martínez’s
revealing observations about both logic and magic set here and there in this story, what if the
proper solution is not so playful? For example, what if the three symbols already revealed follow
instead the alternative rule that one stroke equals 1? If this were so, then the missing fourth
symbol in this cryptic series would not be an “M” with a bar drawn horizontally through it (in
keeping with our different rule, this strange symbol could stand instead for the number 5), but
disconcertingly could be drawn either as a single stroke (Figure 12.1d), or possibly as an inscribed
circle, the letter “O,” or a zero (Figure 12.1e).
255
Doubt as to the proper resolve of Martínez’s series of symbols illustrates Wittgenstein’s

(1953: §201) cryptic and oft-quoted remark: “This was our paradox: no course of action could be
determined by a rule, because every course of action can be made out to accord with the rule. The
answer was: if everything can be made out to accord with the rule, then it can also be made out
to conflict with it. And so there would be neither accord nor conflict”.
I am not a philosopher, nor a novelist. It seems to me, however, that Martinez’s tale and
Wittgenstein’s remark both tell us something about ourselves, about how we are given to looking
for similarities among things and events proving that what we are seeing makes sense not by
chance but necessity. It might even be argued that human beings are strongly predisposed to
equate similarity with necessity. This is why we need statisticians, however much statistics may
sometimes seem only a cultivated way of lying for effect. They keep us from foolishly jumping
to the conclusion that similarities in appearance or similarities in effect are necessarily similarities
of cause. And in this regard, we need to remember that when statisticians say that something
should be attributed to “chance,” they do not mean “without cause.” Far from it: the point they
are making is that the cause (or causes) is not necessarily the one we think it is.
LAPITA
In the annals of modern archaeology, the word “Lapita” minimally refers to a kind of ancient
pottery that has been found at more than 200 archaeological sites across a broad arc of islands in
the southwestern Pacific, starting in the west at the small town of Aitape on New Guinea’s Sepik
coast and stretching far eastward to the small-island archipelagoes of Fiji, Tonga, and Samoa. As
Christophe Sand has remarked, Lapita’s similarities from one end of its geographic distribution
to another in the Pacific are conspicuous, so much so that the commonalities observed certainly
appear to go beyond “some simple broad similarity demonstrable through tables and statistics”
(Sand 2003: 1). In short, there is something about Lapita that seems to be begging us, perhaps
even taunting us, to decipher.
On other occasions (Terrell 1990, 1996) I have written about how Pacific scholars have
answered the question What is Lapita? My goal previously was to document some of the strengths
and weaknesses of writing stories—or to give them a more scholarly turn, historical narratives—
about “Lapita people” (or “peoples”) to account for Lapita’s distribution and shared characteris-
tics. Here I would like to consider a related issue. For the last half century or so, it has been taken
more or less for granted that Lapita adds up to an orderly series of events all linked with one
another by some definite, although now obscure, logic of relatedness. But what if we are being
fooled—or are fooling ourselves—by appearances?
Similarities of Appearance
Since the 1950s, Lapita pottery and the places where it has been found have been the focus of
much research and academic debate. Yet as Christophe Sand (2003: 1) has written: “every archae-
ologist who has been given the opportunity to compare the dentate-stamped potsherds, shell
ornaments and polished adzes from Lapita sites of the Bismarcks, the Reef/Santa Cruz Islands,
New Caledonia, Fiji and Tonga [and also Vanuatu], is faced with the clear evidence that a link
unites these archaeological items from sites scattered over thousands of kilometres.”
I know nobody who doubts the basic truth of Sand’s remark, however cryptic it may be.
Most who are familiar with Lapita pottery, in other words, would agree with what Jack Golson
(1961) wrote decades ago: this pottery is evidence for some kind of “early community of culture”
that once spanned an astonishingly wide swath of the southwestern Pacific over three thousand
256
years ago. In doing so, this culture straddled the ethnological boundary line conventionally drawn
by 19th century commentators between the darker-skinned peoples of Melanesia and their lighter-
skinned counterparts in Polynesia.
A Community of Culture
In the middle of the last century, Kroeber and Kluckhohn extensively documented that what the
word culture means in the social sciences is ambiguous. Its usage is polysemantic. Their published
analysis lists 160 definitions of the word in the English language alone. They concluded: “In
anthropology at present we have plenty of definitions but too little theory” (Kroeber and
Kluckhohn 1963: 357).
In spite of this embarrassment of semantic riches, some nowadays see this concept as more
or less equivalent to the phrase “normative behaviour,” an interpretation that readily invokes
research questions about contingencies of reinforcement, the prescriptive character of social
relationships, the causes of deviancy, and the like. In seeming contrast, many anthropologists
evidently prefer to think about culture instead as basically a mental or cognitive phenomenon
having an inherent structure and logical organisation—in short, as social rules, symbolic repre-
sentations, and learned constraints capable of motivating (or at least invoking) human actions and
responses (e.g., Geertz 1973).
Both ways of thinking about culture are compatible, for both are alike ways of trying to
understand the simple truth that much of what we do as individuals is predictable not just because
of our genes, personalities, and life histories, but also because we are influenced by what others
around us are doing, and by what others tell or show us is right or wrong, good or bad, wise or
foolish. It would seem to follow, therefore, from these two alternative ways of thinking about
culture as a phenomenon or as a state of mind that there must also be similarly contrastive ways
of thinking about Lapita as a community of culture.
As a Cultural Complex
Archaeologists remain uncertain exactly when and where people in the Pacific first started
making Lapita pottery: sometime in the middle of the 2nd millennium B.C., and somewhere in
the Bismarck Archipelago are the best guesses at the moment. Nor do archaeologists agree on
when and where the last potter made the last Lapita pot. Ironically, in spite of the widespread
popularity of this pottery in its halcyon days, this time may have been sometime early in the
Current Era and somewhere again in the Bismarck Archipelago. Much depends on what you think
a Lapita pot should look like.
Setting aside these uncertainties as to time and place, it is probably safe to say that nobody
disputes that what some have called the “Lapita phenomenon” (Spriggs and Chippendale 1989)
lasted for at least several centuries once it got started, and that in its heyday, Lapita pottery
achieved remarkably wide popularity. Furthermore, most experts would insist that Lapita was
much more than simply a pottery style. It is now conventional to say that Lapita was a whole
“cultural complex” of many diverse traits—material, technological, economic, social, and
spiritual, too—all shared by a people or peoples popularly identified by archaeologists and
journalists alike as the “Lapita people.”
As a Series of Migrations
According to Roger C. Green and others, Lapita was also the “signature” of “one of the significant
sets of step-by-step pulses” in an inter-related series of episodes, cycles, or “pulse-like spreads”
of human colonisation “leading to the complete settlement of the Pacific” (Green 2003: 95, 97-98,
257
Fig. 2). That is, what archaeologists have been finding at Lapita sites may be seen collectively as
one step in a “continuing series of migrations” marking “the episodic pattern of colonising
mobility within Oceania.” Moreover, these “serial migrations of related peoples” were made
possible by “a series of adaptive steps”—a sequence of indispensable accomplishments charac-
terising “the increasingly specialised series of cultural innovations and adaptations required to
make such migration possible and successful” (Green 2003: 95, 97, 98).
Judging by what others, too, have written about Lapita as an historical phenomenon, many
experts on Pacific prehistory similarly have in their mind’s eye something akin to what Green sees
as a series of events or episodes. There is apparently wide support for the idea that after Lapita
got started somewhere in the Bismarck Archipelago sometime in the middle of the 2nd millen-
nium B.C., the people or peoples making this pottery initiated a cascading series of migrations
that eventually culminated in the colonisation of Polynesia. Moreover, Green is not alone in
heralding this unfolding series of migrations as one of humankind’s last great adventures before
the development of technologies advanced enough to let us break free of the Earth’s gravita-
tional field and begin exploring outer space (Green 2003: Table 1; Irwin 1992).
Nobody, however, need appeal to Wittgenstein to observe that just as a finite sequence of
numbers can be continued in a variety of ways, so too are there many possible ways to account
for the human settlement of the Pacific, however free-flowing or episodic this chapter in our
human story may have been. Needless to say, too, bringing into play a rule that people couldn’t
have done what they needed to do to write this chapter until they were able to do what they had
to do isn’t an explanation. Furthermore, just as parents are given to saying to children, just
because one can does not mean one should. Even when people in the Pacific had mastered (if they
needed to master) the sorts of “adaptive steps” Green and others say were sine qua non for the
settlement of Oceania, they did not have to put these new “adaptations” to work in ways that
would or should expand their settlement horizons. In sum, as Anderson (2001: 15) has also noted,
many explanations for Lapita’s dispersal in Oceania have been proposed or are plausible, and in
spite of claims sometimes to the contrary, it has not been possible to rule out, or definitively
endorse, any of the contending explanations.
Even Anderson (2001), however, would grant that Lapita’s wide geographic distribution in
the Pacific is evidence pointing to a “Lapita migration” on the grounds that at least some people
making Lapita pottery were obviously mobile and were quite open to setting up shop in new
places sometimes quite far from home. I would note, nonetheless, that Anderson has repeatedly
also pointed out that the broad resemblances in material culture exhibited by Lapita sites through-
out their distributions in the Pacific may be misleading us into believing that other sides of the
Lapita story—their subsistence practices, social institutions, and so forth—must have been as
similar and widely shared in common.
As an Extraordinary Series of Migrations
Nowadays it is not unusual to label what Green and others (e.g., Anderson 2001, Lilley 2006) see
as the serial and episodic patterning of settlement in Oceania as part of a great “Austronesian
expansion” out into the Pacific (Terrell 2004). However, unless you are prepared to argue that
speaking an Austronesian language was somehow needed to getting the job of Pacific colonisa-
tion done, this labelling would seem misleading. After all, as statisticians tell us, a correlation is
not a cause. Even if it is likely that the first Lapita settlers of Remote Oceania were Austronesian
speakers of one Oceanic sort or another, it would be a far stretch to insist that being an
Austronesian speaker was in itself instrumental in colonising Remote Oceania.
258
Figure 12.2 Episodic pattern of relative colonising mobility in Oceania (Anderson 2001: Fig. 4).
Nonetheless, it is not only the scale, or scope, of Lapita’s geographic distribution that leads
some experts today to talk about Lapita pottery and all that was once associated with it as a
remarkable human phenomenon. Again Roger Green is not alone in seeing Lapita as merely one
step, phase, or episode in a much longer and bigger series of historical events or developments
(e.g., Figure 12.2), a lengthy story of ever onward and expanding human dispersal popularly
called the “Austronesian migration.” As Peter Bellwood (2006: 62) has phrased the idea:
“Austronesian spread was a demographically powerful but chronologically long-term historical
process that moved people, languages and cultures over vast distances”.
How Bellwood would explain this apparently bigger and older story of island discovery and
settlement is familiar to many by now, and I will not repeat what he has so often and so eloquent-
ly written on this popular theme (e.g., Bellwood 2006). It seems worth pointing out, however, that
his historical arguments are rather less nuanced than Green’s various accounts of step-by-step
settlement “pulses” out into the farthest reaches of the Pacific. Specifically, Bellwood sees the
invention of agriculture elsewhere in the world as the decisive adaptive step in our species’
cultural evolution: “I infer that increasing dependence on food production led to upward trends in
population density, ultimately resulting in population dispersal” (Bellwood 2007a: 88).
However true this may be as a broad historical generalisation about our species (or any species,
for that matter), what is not certain is how important agriculture was to Lapita’s own expansion.
Bellwood (2007b: 104) acknowledges, for instance, that the people making this pottery were
evidently not so dependent on crop growing that this dependency kept them on the move. Indeed
there seems to be archaeological evidence to the contrary (Anderson 2004, Burley et al. 2001,
Leavesley 2006). Green (2003: 103) has contested whether we yet know enough about the subsis-
tence practices of Lapita people to be able to tell for sure how they went about making a living, or,
by extrapolation, how driven they may have been by economic or demographic necessity.
The Hidden Logic of Lapita
Given what I have summarised about the various ways in which experts today are thinking about
Lapita as a community of culture, it seems that many people currently take it to be more or less
self-evident that anything as widely similar and as wide-ranging as Lapita both as a ceramic style
and also as a discernible cultural assemblage, or complex, must be decipherable as a coherent
series of events that were governed by a logic that we ought be able to make sense of if we could
somehow “loosen the conceptual logjam” (Anderson 2001).
259
This may well be so, but I would caution that we should not forget that just as any finite
sequence of numbers can be continued in a variety of different ways—and not just in the one way
that may look reasonable to us—so too, any series of events that can be arranged in a sequence
can be told as a story in a variety of ways (Terrell 1990). There’s the rub. If the particular story
we elect to tell about a sequence of events we have pieced together is told well enough—as
Martínez illustrates so effectively—not only others but also we ourselves can be all too easily
enticed by the flow and logic of our tale into believing that merely placing one event after another
explains the sequence thus strung together—a logical flaw that philosophers refer to as post hoc,
ergo propter hoc: the fallacy of arguing from temporal sequence to causal relations.
If, like Peter Bellwood (2007b: 104), you are not persuaded by me invoking rules of logic,
let me hasten to add that there is now growing factual evidence supporting the thought that
caution should be observed when it comes to “telling about” Lapita. Let me explain.
LAPITA PEOPLE
Recognition that sites yielding Lapita pottery are not only widely distributed in the Pacific but are
also to be found in both Melanesia and Polynesia has led to much debate over the last half century
about what the makers of this pottery must have been like, biologically speaking. Initially it seems
to have been taken for granted that the people behind this pottery must have been as distinctive
and readily identifiable as their pottery. Additionally, since the oldest archaeological sites in
Polynesia have consistently yielded Lapita potsherds, it has also been easy for decades to
conclude that “Lapita people” were not only the first settlers of Polynesia, but were also none
other than the long-sought heroic ancestors of the Polynesians: “Vikings of the Sunrise,” as Sir
Peter Buck dubbed them (Buck 1938).
If all this seems too obvious to need saying, allow me to observe that the sequences of events
commonly pieced together nowadays as the story of Lapita is a sequence that can be read differ-
ently depending on the direction you read it in. When read from left to right—that is, from west to
east—the Lapita story usually comes across as a tale about an ancient folk called “the Austronesians”,
who are strongly given to migrating. But when read from right to left—east to west—the story is
conventionally read as one about today’s Polynesians and their prehistoric forebears.
While journalists rarely seem to have any problem with this alternative left-to-right or right-
to-left rendering of Lapita’s history (perhaps because writing about “Polynesians” has a lot more
headline cachet than writing about “Austronesians”), as a rule archaeologists are more cautious
than journalists need to be. Unlike the average journalist, archaeologists know that while
Polynesians are Austronesian speakers, not all Austronesian speakers are Polynesian. Journalists
may not care one way or the other, but it actually does make a difference which way you decide
to write about the people behind Lapita’s pots.
A Genetic Strain that was Stable and Bred True
As early as 1960, the anthropologist Robert Suggs was linking Lapita—although not yet labelled
as such—with the forebears of the Polynesians in his book The Island Civilizations of Polynesia:
Archaeological evidence shows that the basic population pool from which the Polynesians
diverged developed on the coast of South Asia…. [Their subsequent movement from there] can
be traced across the Pacific in the archaeological records of the Philippines, Melanesia, Papua,
and Indonesia, using certain characteristic artifact types as cultural indices…. The evidence at
hand indicates that by approximately 1000 B.C., at least, the fringe areas of Polynesia (i.e.,
260
Melanesia) were already occupied by Malayo-Polynesian speakers bearing the antecedents of

Polynesian culture.” (Suggs 1960: 72)
Or in fewer words: “After mating [on the Asian coast] had produced a genetic strain that was
stable and bred true, the group filtered into the Pacific . . . and ultimately arrived in the Polynesian
triangle” (Suggs 1960: 37).
Like Lapita, there may be more than one way to read these words written by Robert Suggs,
but I think we can take them to mean that he assumed human biology would be able to support
archaeology in showing us that Lapita marks a coherent series of events linking Polynesians directly
with Asians. It would probably seem gratuitous for me to add that equating Polynesians with Asians
is an assumption, or foundational premise, that continued to be a favoured idea in Pacific academic
circles for decades after Suggs’ book appeared (e.g., Bellwood 1978: 47-48). Even today this idea
has its advocates (e.g., Diamond 1997). In its most extreme journalistic form, the first Austronesians
and the first Polynesians are seen as basically one and the same (Diamond 1988).
How Pure is Pure?

Diamond and others aside, experts have been disputing for years how stable Suggs’ primal human
“genetic strain” continued to be after it left Asia and launched itself out on the ocean blue. The
grounds for both caution and confusion are not only logical and rhetorical, but also biological.
The linguist Andrew Pawley (2002: 268) has written that anthropologists have long assumed
that the Polynesians are biologically of mixed origin. But in the 1970s the renowned Harvard
somatologist W.W. Howells felt able to argue that Polynesians and Melanesians are “too different
physically” to be related to one another (Howells 1973: 234). For Howells (1973: 234; see also
1973: 253, 261), the voyagers who finally reached Tonga and Samoa were not only Lapita potters
by trade but were already physically Polynesian in appearance.
Other experts at the time were more open to seeing a certain amount of cross-breeding
between Pre- or Proto-Polynesian Lapita migrants and the darker-skinned (and presumably long-
resident) Papuan-speaking people of Melanesia—provided room was left for saying that those
people making Lapita pots who got all the way to Polynesia had themselves received “relatively
little influence from Non-Austronesian populations” en route (Bellwood 1975: 17).
People or Peoples?
By the 1990s, however, some experts began to question whether we should be thinking of the
Lapita people as a singular (and genetically pure?) biological population at all. In the middle of
the decade, Matthew Spriggs (1995: 124) wrote: “there may have been a moment in the Bismarcks
when there was a single people using Lapita pottery, genetically, linguistically and culturally
distinct from their neighbours. But this unity and distinctiveness would have been short-lived.”
During the same decade, Patrick Kirch added his own perspective to the continuing discussion:
… we must not simply assume at the outset that the makers and users of Lapita pottery conformed
to linguistic and biological entities; rather, this is a hypothesis to be carefully tested. Indeed there
are strong reasons for arguing that the Lapita peoples were a dynamic group who recruited a
diversity of populations (biological entities) and speech communities into their sphere over the
millennium or so that Lapita existed as an archaeologically-identifiable entity… [I]t is for this
reason that we must speak of Lapita peoples, not Lapita people, and of Lapita societies, not
society. (Kirch 1997: 16, 18; see also Kirch 2000: 92-101).
261
I am not sure all experts still see wisdom in talking about Lapita “peoples” rather than Lapita
“people” (e.g., Bedford, Spriggs and Regenvanu 2006), and some, such as Green (2003: 96-97),
appear to argue for both: “For a brief period of a few centuries they appear to have . . . possessed
a unified base in language and kinship, before they began to diverge into a multiplicity of groups,
many of these ancestral to the larger scale Pacific ethnic groupings recognised today.
The Advent of Molecular Genetics

During this first decade of the new millennium, human molecular genetics has radically changed
how human biological diversity in Island Southeast Asia and the Pacific is being mapped and
interpreted. What seemed controversial at the beginning of the decade (Terrell, Kelly and Rainbird
2001) has now almost become the new orthodoxy. Far from today’s Polynesians simply being the
descendants of what Suggs saw as a genetic strain that was stable and bred true, current molecular
evidence seems to be saying that the biological characteristics of Polynesians firmly link them
genetically to both Southeast Asia and Melanesia.
According to Kayser and his colleagues, for example, roughly two-thirds of the typifying
characteristics of Polynesian Y-chromosomes can be traced back to Melanesia; nearly 94% of
their mitochondrial DNA variation, in contrast, looks Asian in origin (Kayser et al. 2006). In a
more recent study, these same experts assert: “Our data indicate that on average about 79% of the
Polynesian autosomal gene pool is of East Asian origin and 21% is of Melanesian origin” (Kayser
et al. 2008: 194)
Not everyone agrees with such a radical reassessment of Polynesian genetic history. Stephen
Oppenheimer and Martin Richards, for example, favor a reading of the new molecular evidence
that more or less fits the sort of scenario that Howells and others once favored. Recall that
Howells insisted that “the Pre-Polynesians who arrived in Fiji and Tonga were physically
Polynesian, distinct from anything which can be called Melanesian” (Howells 1973: 234).
Oppenheimer and Richards would not phrase things as strongly since they do not draw the same
sharp divide between Polynesians and Melanesians that Howells called for (Richards 2007: 100).
But this may be only a quibble, because like Howells, they would also derive most of the genetic
characteristics of people in Polynesia today more or less directly from a place of origin in the
west. While they disagree strongly with Peter Bellwood, for instance, on precisely where the
genetic homeland of the Polynesians was located—Bellwood favors Taiwan, but they say “coastal
Melanesia, Wallacea and Island Southeast Asia are all potential offshore homelands for the
ancestors of today’s Polynesian peoples” (Oppenheimer and Richards 2002: 290)—like Howells,
Bellwood, and many others, these authors are similarly focused on testing “homeland hypotheses
attempting to explain Austronesian and/or Polynesian origins and dispersals” (2002: 287).
Please do not get me wrong. The impressive track record being developed by molecular
genetics investigations in the Pacific proves that Richards is right when he says that using biolog-
ical data may “finally allow us to get a more direct purchase on past human movements and help
us to distinguish alternative hypotheses (at least within an archaeological framework)” (Richards
2007: 98). I think there is a nagging concern here, nonetheless.
Finding More in the Evidence Available

When Suggs and Howells were writing about the prehistoric Pacific decades ago, few seemed to
question whether it was worthwhile to trace the ancient movements of recognisable “races”,
“populations”, or “genetic strains.” Nowadays, molecular genetics is repeatedly showing how
simplistic this old ambition was (e.g., Friedlaender et al. 2007). If so, what else beyond getting a
purchase on past human movements can be done with genetics evidence?
262
One obvious resolve is to use this new kind of biological evidence to see if it is possible to
find out what people were doing in the Pacific over time regardless of where their ancestors came
from. For example, two unique α-globin gene deletions resulting in forms of α-thalassaemia—
labeled by geneticists as the α3.7III type and the α4.2 type—can be found throughout coastal and
lowland Melanesia. Both types are local mutations; they are not found in Southeast Asia. One
occurs in Polynesia; the other does not (Oppenheimer and Richards 2002: 292). Oppenheimer and
Richards see such local mutations as an indication that the ancestors of the Polynesians must have
stopped somewhere in northern Melanesia long enough to intermarry locally:
If the Polynesians’ ancestors ever stopped on the north coast of New Guinea long enough to pick
up the α3.7III-deletion, however, they failed to pick up the α4.2-deletion although it is present in
over 80 per cent of the people now living there. The alternative is that they travelled out to the
Pacific 3500 years ago via offshore islands such as the Bismarck Archipelago, where the α3.7III-
deletion is the dominant variant today. Such a bypass interpretation certainly fits the aspect of
the common archaeological model that identifies the Proto-Polynesians with Lapita pottery
because, with one exception, there are no Lapita pottery sites anywhere on the New Guinea
mainland. (Oppenheimer and Richards 2002: 292)
Whether Oppenheimer and Richards have read these α-globin gene deletions correctly is
beside the point I want to make. Such an interpretation of what can be done with genetic evidence
illustrates how local events in the Pacific—such as something as unpredictable as gene mutation,
and something else as contingent as marriage—would have modified and reshaped the biology of
people shifting their homes from wherever it was that their forebears had been living to wherever
they themselves decided to set down roots. Similarly, there is growing molecular evidence
nowadays supporting the thought that by the very act of shifting their homes from place to
place—something people often do even when scholars do not say they are “migrating”—may
have in itself unintentionally helped transform the biology of people in the Pacific.
Here’s some of that evidence.
Pattern and Process

A number of different biological and historical processes, singly or in combination, can affect the
genetic character and variability of any population, human or otherwise. These include natural
selection, inbreeding, intermarriage (or in more generic terms, “intermixture”), genetic drift,
population bottlenecks, and founder events (such as lineal fission). Furthermore, it has long been
recognised in population genetics that these different processes can lead to similar biological
outcomes—that is, similar patterns of genetic variability and relatedness. I won’t call upon
Martínez and Wittgenstein again out of fear that you may be finding my allusions annoying, but
please imagine these gentlemen lurking in the background nearby, nonetheless.
An apt example can be found in a recently published research paper by the biological anthro-
pologist Jonathan S. Friedlaender and his colleagues (2008). Most biologists today accept that the
ancestors of today’s Polynesians went through numerous genetic founder events or episodes
(Kayser et al. 2006, Oppenheimer and Richards 2002). Many years ago I phrased the potential
consequences of such events this way:
[the] distinctive features of the modern Polynesian physique (and gene pool) might be due to the
combined effects of (a) lineal fissioning and (possibly) immigrant selection during the unknown
events that led to the initial colonization of Fiji-Tonga-Samoa; and (b) later in situ change and
diversification fostered by local geography, the fissioning and fusion of communities consequent
with population growth, and increasing isolation-by-distance. Specifically, it can be argued that a
263
particular pathway, or sequence, of events would explain how the “Polynesians” of the present day
emerged out of an earlier “Melanesian” population matrix not only in language and culture, as
many now accept, but also in their fundamental and often distinctive biology. (Terrell 1986: 192)
The expression “lineal fissioning” refers to the establishment of new social groups by small
numbers of biologically related individuals—the genetic effect of which “is to reduce the number
of independent genomes in splinter groups, thereby augmenting the nonrepresentativeness of the
genetic sample” (Fix 1979: 215).
Judging by their recent publications (Friedlaender et al. 2007, 2008), Friedlaender and his
co-authors may be more comfortable invoking biological processes such as drift, inbreeding, and
intermarriage to explain geographic patterns of genetic variation in the Pacific than social and
demographic processes such as founder events and lineal fissioning, but they acknowledge that
such events have happened in the genetic past of our species as a whole (Friedlaender et al.
2008: 176).
Now as I said earlier, different processes can lead to similar biological outcomes. Hence it
is somewhat surprising to find Friedlaender and his co-authors (2008: 173) confidently saying—
even while acknowledging the genetic distinctiveness of Polynesians vis-à-vis other Pacific
Islanders—that “the ancestors of Polynesians moved through Melanesia relatively rapidly and
only intermixed to a very modest degree with the indigenous populations there”. This claim is a
surprise to me because, as Kevin Kelly, Paul Rainbird, and I remarked some years ago, founder
events—particularly lineal fissioning—by definition and by fact lead to loss of genetic variation.
Since there may be no easy way to tell what has been lost during such events given only the
evidence of what has survived, founder events can lead to loss of genetic history.
Therefore, it seems unwise to conclude much about the genetic history of Polynesians based
on the absence of certain genetic traits among today’s Polynesians, since there is evidence that
their ancestors went through founder events (Kayser et al. 2006). As Kayser and his colleagues
observed recently: “the estimation of the East Asian and Melanesian components in the current
Polynesian autosomal gene pool is potentially complicated by the severe bottlenecks that accom-
panied the migrations to and through the Pacific (Remote Oceania)” (Kayser et al. 2008: 194).
“Bottlenecks” are another way—and in this instance, probably an inappropriate way—of talking
about founder events.
Allow me to appeal to logic once more. One of the most notorious errors in science and in
logic is mistaking a lack of evidence as evidence for or against some idea. In the case of
Polynesians, far from their current genetic characteristics confidently telling us (a) where their
ancestors came from, (b) how much (or how little) their ancestors may have resembled other
people also living in the southwest Pacific during the heyday of Lapita pottery, and (c) how
rapidly they may have journeyed on their way to what is now Polynesia, it is sensible to take
the thesis advanced by Friedlaender and his colleagues with a grain of salt. Like Martínez’s
cryptic series of murders, the biological distinctiveness of modern Polynesians (and Fijians)
may be hiding more than it is showing about their origins and history (Terrell et al. 2001: 103).
While here is not the place to discuss the issue further, reanalysis of their co-ancestry data
(Friedlaender et al. 2008: supplementary Table 7) using network rather than cladistic algorithms
supports the inference that genetic relationships among Polynesians, Melanesians, and Asians
have been distorted by the (largely unknown) founder events that ultimately led to the human
colonisation of Polynesia.
264
DISCUSSION
I think most of us would agree that there is something about Lapita and its pottery that appears to
be begging us, perhaps even taunting us, to decipher. It makes no difference what we call it—a
culture, cultural complex, archaeological horizon, people, peoples, migration, episode, pulse,
spread, cycle, or signature. Most of us see Lapita, however variable it may once have been as an
historical phenomenon, as something singular, coherent, and rule-bound. Moreover, it seems just
as certain that many of us also see Lapita as but one installment in a much bigger, much older
story featuring journeys by people starting on or near the Asian mainland and continuing ever
onward and outward, however intermittently, until the distant islands of Hawai‘i and Rapa Nui
had been brought into our human fold.
Said more simply, Lapita presents itself to us—so we often seem to be telling ourselves—as
a series or sequence of intimately related historical events. Robert Suggs was not the first, nor the
last, to be convinced that this sequence of events must have been driven by logic, necessity and
rules rather than chance or blind human opportunism. Proof that this was so would surely be ours
if we could only decipher what Lapita means, if we could only figure out what was driving all
those ancient Polynesians, Austronesians, or at any rate, Pacific Islanders to do what they did.
Toward this end, molecular genetics is now not only influencing how many of us are trying
to decipher Lapita’s impact on Pacific prehistory, but also—and more to my point—how many of
us see this past in our mind’s eye. The good news here is that apparently because of the fondness
for hypothesis testing in the biological sciences, there is now also greater openness in academic
circles to construct and evaluate alternative hypotheses (e.g., Green 2003) and weigh statistical
probabilities. But there is also a downside to the advent of molecular genetics into the field of
island studies.
Over a decade ago, I reported on the words being using to write about Lapita as ceramics
and as history (Terrell 1996). We were often extravagant back then in attributing remarkable
qualities and motivations to the makers of Lapita pottery. In gloomy contrast, the vocabulary now
commonly encountered in scholarly publications on the Pacific could be described as bloodless
and impersonal. Formerly there was much talk about voyaging, colonisation, settlement, conquest,
and the like. Now the talk is more often about transmission, movement, spread, flow, expansion,
pulsed expansion, phylogenetic spread and diversification, dispersal, dispersion, entry, entry
time, expansion time, replacement, admixture, drift, haplotype sharing, waves of advance, major
migration events, impulses of influence, and so on. Where are the people (or peoples) in all this
talk of biological spreads and pulses?
The concern here is not that the words we are now using are politically incorrect or craven.
Peter Bellwood has been forthright in saying that archaeology is for him a useless discipline if
it cannot illuminate large-scale processes in human history because, in his own words, “big
pictures are far more significant than minutiae” (Bellwood 2006: 63). In science, big pictures and
minutiae are complementary, but we can hear the passion in his declaration. Nonetheless, leaving
out the people and talking instead only about big-picture dispersals, pulses, or waves of advance
strikes me as objectionable. I also cannot help but think that many Pacific Islanders, too, may be
far more interested in what their forebears did to shape their lives in the islands than in our
strange scholarly passion for finding where their ancestors may or may not have come from in
the first instance.
265
Nevertheless, it seems plain that most of us who write about Lapita, however we do it, do
so out of passion as well as perhaps Lapita’s career enhancing academic possibilities. We really
do want to decipher the riddle of Lapita. We want to decode the symbols in this series, so to speak.
But The Oxford Murders should give us pause:
1. It is second nature for archaeologists to order things and places into sequences, but we need
to worry about the fallacy of arguing from temporal sequences to causal relations. How
historically connected were the places and things we label as Lapita?
2. Given our hindsight on events long past, it may be true that the settlement of the Pacific can
be divided into a series of geographic moves. The timing and direction of these moves may
be debatable, but mapping this series of events does not appear to be just an exercise in
fantasy and imagination. But what does it mean to say that these moves were historically
connected? What does it mean to suggest, for example, that Lapita people were caught up in
a “colonising pulse” that grew weaker as the distribution of Lapita expanded eastward
(Anderson 2001: 18)?
3. As Anderson has also remarked, we need to be open to the probability that Lapita, however
we characterise it, was more variable than we may suspect: “The evidence of a generic
coherence in material culture across the Lapita range may be lulling us, then, into a false sense
of similarity about other aspects of Lapita prehistory, such as subsistence patterns, ecological
relationships, and social structures” (Anderson 2001: 18-20). Even if there is logic to the
series of events we label as Lapita, the logic and rules involved surely must have changed to
fit new circumstances, new opportunities, and new personalities.
CONCLUSION
At the start I said something in passing that I would like to return to. I suggested that human
beings are predisposed to look for similarities among things and events in hopes of finding that
the world today—or in the past—makes sense not by chance, but by necessity. It might even be
said, I noted, that people are predisposed to equate similarity with necessity. This is why we need
statisticians. They keep us from jumping too quickly to the conclusion that similarities in appear-
ance or similarities in effect are necessarily similarities of cause. Then I added this thought in
passing: when statisticians say something should be attributed to chance, they do not mean
without cause. Instead, the point they are making is that due caution is advised. The cause or
causes we have in mind may not necessarily be the causes responsible.
Fundamental to the concept of culture is the freedom of association between words and
things. This, of course, is why ideas are able to take on a life of their own. A leopard may not be
able to change its spots, but ideas are so malleable that, for example, a political party calling itself
“liberal” can be as conservative as voters will allow, and adopting an idea as your own cannot be
controlled by even the most repressive of political regimes. The lesson here is that even in the
most extreme interpretations of Lapita as a unitary historical phenomenon in which material
culture, ethnicity, language, and human biology were all strongly correlated (in the old jargon of
20th century anthropology: race = language = culture), cause and causation cannot solely lie
within these purported correlations, but also beyond them.
If so, then understanding Lapita calls for looking beyond Lapita as a cultural phenomenon
regardless how unitary or variable it was. Lapita was more than ideas; Lapita was also people.
People are not leopards. They can change their minds, reinterpret even the most solemn tradi-
tional ideas, and take advantage of new opportunities to get along in life. This being so, I think
266
we must entertain the likelihood that Lapita’s history in the Pacific was more Markovian in its
expansions, transformations, and demise than sometimes thought (in sequences of events
described as Markovian, it is enough to know the present state of a system to predict what its
future state may be like).
Said less grandly, there is indeterminacy to any series of events however much in hindsight
they may appear to us to have been premeditated migrations, pulses, dispersals, or what have you.
Just because things followed one another, one after the other, down through time and out across
space, this does not have to mean the sequence of events (or the flow of people) reconstructed
was rule-like and predictable over the long run. Along with Martínez, we need to think twice if
we are inclined to believe that a series of similar events—in this case, a series of Lapita events—
happened because, when set in sequence, they look like a coded message we ought to be able to
decipher, if only we knew the rules—they would be Austronesian rules, wouldn’t they?—that
made them all happen.
REFERENCES
Anderson, A.J., 2001. Mobility models of Lapita migration. In G. Clark, A. Anderson and T. Vunidilo (eds),
The Archaeology of Lapita Dispersal in Oceania: Papers from the Fourth Lapita Conference, June
2000, Canberra, Australia. Terra Australis 17. Canberra: Pandanus Books, pp. 15-23.
Anderson, A.J., 2004. Islands of ambivalence. In Scott M. Fitzpatrick (ed.), Voyages of Discovery: The
Archaeology of Islands. Westport, Conn.: Praeger, pp. 251-73.
Bedford, S., M. Spriggs and R. Regenvanu, 2006, The Teouma Lapita site and the early human settlement
Bellwood, P., 1975. The prehistory of Oceania. Current Anthropology, 16: 9-28.
Bellwood, P., 1978. Man’s Conquest of the Pacific. Auckland: Collins.
Biletzki, A., and A. Matar, 2008. Ludwig Wittgenstein. In Edward N, Zalta (ed.), The Stanford
Encyclopedia of Philosophy (Winter 2008 Edition), forthcoming URL = http://plato.stanford.edu/
archives/win2008/entries/wittgenstein/.
Bellwood, P., 2006. The early movements of Austronesian-speaking peoples in the Indonesian region. In
T. Simanjuntak, I.H.E. Pojoh and M. Hisyam (eds), Austronesian Diaspora and the Ethnogenesis of
People in Indonesian Archipelago: Proceedings of the International Symposium. Jakarta: LIPI Press,
pp. 61-82.
Bellwood, P., 2007a. Review Feature. First Farmers: The Origins of Agricultural Societies by Peter
Bellwood. Overview. Cambridge Archaeological Journal, 17: 88-91.
Bellwood, P., 2007b. Review Feature. First Farmers: The Origins of Agricultural Societies by Peter
Bellwood. Reply. Cambridge Archaeological Journal, 17: 102-09.
Buck, P.H., 1938. Vikings of the Sunrise. New York: Frederick A. Stokes.
Burley, D.V., W.R. Dickinson, A. Barton and R. Shutler Jr, 2001. Lapita on the periphery. New data on old
problems in the Kingdom of Tonga. Archaeology in Oceania, 36: 89-104.
Diamond, J.M., 1988. Express train to Polynesia. Nature, 336: 307-8.
Diamond, J.M., 1997. Guns, Germs, and Steel: The Fates of Human Societies. New York: W.W. Norton.
Fix, A.G., 1979. Anthropological genetics of small populations. Annuals Review of Anthropology, 8: 207-30.
Friedlaender, J.S., F.R. Friedlaender, J.A. Hodgson, M. Stoltz, G. Koki, G. Horvat, S. Zhadanov, T.G.
Schurr and D.A. Merriwether, 2007. Melanesian mtDNA complexity. PLoS ONE 2 (2): e248.
doi:10.1371/journal.pone.0000248.
267
Friedlaender, J.S., F.R. Friedlaender, F.A. Reed, K.K. Kidd, J.R. Kidd, G.K. Chambers, R.A. Lea, J-H. Loo,
G. Koki, J.A. Hodgson, D.A. Merriwether and J.L. Weber, 2008. The genetic structure of Pacific
Islanders. PLoS Genetics, 4 (1): 173-190 e19. doi: 10.1371/journal.pgen.0040019.
Geertz, C., 1973. The Interpretation of Culture. New York: Basic Books.
Golson, J., 1961. Report on New Zealand, Western Polynesia, New Caledonia, and Fiji. Asian Perspectives,
5: 166-80.
Green, R.C., 2003. The Lapita horizon and traditions – signature for one set of oceanic migrations. In C.
Sand (ed.), Pacific Archaeology: Assessments and Prospects. Le Cahiers de l’Archéologie en Nou-
velle-Calédonie 15. Nouméa: Service de Musées et du Patrimoine de Nouvelle-Calédonie, pp. 95-120.
Howells, W.W., 1973. The Pacific Islanders. New York: Charles Scribner’s Sons.
Irwin, G.J., 1992. The Prehistoric Exploration and Colonisation of the Pacific. Cambridge: Cambridge
University Press.
Kayser, M., S. Brauer, R. Cordaux, A. Casto, O. Lao, L.A. Zhivotovsky, C. Moyse-Faurie, R.B. Rutledge,
W. Schiefenhoevel, D. Gil, A.A. Lin, P.A. Underhill, P.J. Oefner, R.J. Trent and M. Stoneking, 2006.
Melanesian and Asian origins of Polynesians: mtDNA and Y-chromosome gradients across the
Pacific. Molecular Biology and Evolution, 23: 2234-44.
Kayser, M., O. Lao, K. Saar, S. Brauer, X. Wang, P. Nürnberg, R.J. Trent and M. Stoneking. 2008.
Genome-wide analysis indicates more Asian than Melanesian ancestry of Polynesians. American
Journal of Human Genetics, 82: 194–98.
Kroeber, A.L. and C.K.M. Kluckhohn. 1963. Culture: A Critical Review of Concepts and Definitions.
[Originally published in 1952 by the Peabody Museum, Harvard University. New York: Vintage
Books].
Leavesley, M., 2006. Late Pleistocene complexities in the Bismarck Archipelago. In Ian Lilley (ed.),
Archaeology of Oceania: Australia and the Pacific Island. Oxford: Blackwell, pp. 189-204.
Lilley, I., 2006. Archaeology of Oceania: Themes and issues. In Ian Lilley (ed.), Archaeology of Oceania:
Australia and the Pacific Islands. Oxford: Blackwell, pp. 1-30.
Martínez, G., 2005 The Oxford Murders. New York: Penguin.
Oppenheimer, S.J. and M. Richards, 2002. Polynesians: Devolved Taiwanese rice farmers or Wallacean
maritime traders with fishing, foraging and horticultural skills. In P. Bellwood and C. Renfrew
(eds), Examining the Farming/Language Dispersal Hypothesis. Cambridge: McDonald Institute for
Archaeological Research, pp. 287-97.
Pawley, A., 2002. The Austronesian dispersal: Languages, technologies and people. In P. Bellwood and C.
Renfrew (eds), Examining the Farming/Language Dispersal Hypothesis. Cambridge: McDonald
Institute for Archaeological Research, pp. 251-73.
Richards, Martin, 2007. Is genetics coming between archaeology and linguistics? Cambridge
Archaeological Journal, 17: 98-100.
Sand, C., 2003. Introduction to the conference: Commemorating the first excavation at Lapita. In C. Sand
(ed.), Pacific Archaeology: Assessments and Prospects. Le Cahiers de l’Archéologie en Nouvelle-
Calédonie 15. Nouméa: Service de Musées et du Patrimoine de Nouvelle-Calédonie, pp. 1-10.
Spriggs, M., 1995. The Lapita culture and Austronesian prehistory in Oceania. In P. Bellwood, J.J. Fox, and
D. Tryon (eds), The Austronesians: Historical and Comparative Perspectives. Canberra: Department
of Anthropology, Research School of Pacific Studies, Australian National University, pp. 112-33.
Spriggs, M. and C. Chippindale, 1989. Introduction to special section: “Early Settlement of Island
Southeast Asia and the Western Pacific.” Antiquity, 63: 547.
268
Suggs, R.C., 1960. The Island Civilizations of Polynesia. New York: New American Library.
Terrell, J.E., 1986. Causal pathways and causal processes: Studying the evolutionary prehistory of human
diversity in language, customs, and biology. Journal of Anthropological Archaeology, 5: 187-98.
Terrell, J.E., 1990. Storytelling and prehistory. In M.B. Schiffer (ed.), Archaeological Method and Theory,
vol. 2. Tucson: University of Arizona Press, pp. 1-29.
Terrell, J.E., 1996. Lapita as history and culture hero. In Janet Davidson, Geoffrey Irwin, Foss Leach,
Andrew Pawley and Dorothy Brown (eds), Oceanic Culture History: Essays in Honour of Roger
Green. Dunedin: New Zealand Journal of Archaeology Special Publication, pp. 51-66.
Terrell, J. E., 2004. The ‘sleeping giant’ hypothesis and New Guinea’s place in the prehistory of Greater
Near Oceania. World Archaeology, 36: 601-609.
Terrell, J.E., K.M. Kelly and P. Rainbird, 2001. Foregone conclusions? In Search of “Austronesians” and
“Papuans.” Current Anthropology, 42: 97-124.
Wittgenstein, L., 1953. Philosophical Investigations. Oxford: Blackwell Publishers.
269
View publication stats

Lapita Textpages

Uploaded by

Document Information

Copyright

Available Formats

Share this document

Share or Embed Document

Sharing Options

Did you find this document useful?

Is this content inappropriate?

Copyright:

Available Formats

Lapita Textpages

Uploaded by

Copyright:

Available Formats

See discussions, stats, and author profiles for this publication at: https://www.researchgate.

Lapita Ancestors and Descendants

Book · January 2009

Peter J. Sheppard Tim Thomas

SEE PROFILE SEE PROFILE

The user has requested enhancement of the downloaded file.

NEW ZEALAND ARCHAEOLOGICAL ASSOCIATION MONOGRAPH 28

Monograph series editors: Dorothy Brown and Simon Holdaway

New Zealand Archaeological Association

© 2009 New Zealand Archaeological Association

Design and page layout by Hamish Macdonald

2. The aceramic to ceramic boundary in the Bismarck Archipelago 11

3. A Review of Solomon Island Archaeology 35

4. Inter-tidal Late Lapita Sites and Geotectonics in the

5. Assessing the anomalous role of ceramics in late-Lapita

6. Communities of practice in the archaeological record

7. The Southeast Solomon Islands Culture History Project:

8. Ethnobotany In The Southeast Solomon Islands Cultural

9. The Post-Lapita Sequence in the Reef/Santa Cruz 181

10. The Teouma Lapita site, South Efate, Vanuatu:

11. Understanding the place properly: Palaeogeography of Selected

Tony Ahikau Heorake Ian A. Scales

UNDERSTANDING THE WHO, WHAT AND WHY OF LAPITA:

THE ACERAMIC TO CERAMIC BOUNDARY

continued next page

PRE-LAPITA ARCHAEOLOGICAL CONTEXTS WITHIN NEW BRITAIN

The Archaeological Evidence – Introduction

New Britain North Coast

New Britain South Coast

PRE-LAPITA ARCHAEOLOGICAL CONTEXTS BEYOND NEW BRITAIN

New Ireland Province

Bougainville Autonomous Region

A REVIEW OF SOLOMON ISLAND ARCHAEOLOGY

GEOGRAPHY AND CULTURAL HISTORY

Figure 3.2 Greater Bukida or Greater Bougainville.

Figure 3.3 AN Languages of the Solomon Islands.

continued next page

THE ARCHAEOLOGY OF THE SOLOMON ISLANDS

Pleistocene Colonisation and Adaptation

The Early to Mid-Holocene Record

Table 3.2 Coastal changes in New Georgia based on palaeoenvironmental

The Arrival of Austronesians

The Northern Solomons

Late Prehistory and the Development of Traditional Solomon Islands

THE WESTERN SOLOMONS – ROVIANA

Munda Tradition (700 – 100 B.P.)

Bao Period (700 to 400 B.P.)

INTER-TIDAL LATE LAPITA SITES AND GEOTECTONICS IN

Figure 4.1 The Solomon Islands.

FORMATION OF THE NEW GEORGIA LAGOON SYSTEMS

LAPITA INTER-TIDAL SITES

WESTERN SOLOMON ISLANDS INTER-TIDAL SITES

GEOTECTONICS AND INTER-TIDAL SITES IN ROVIANA LAGOON

Examination of the profiles for paleoshoreline indicators shows: a feature marked by

Table 4.2 Poroporo Shelter (Araroso Passage).

ASSESSING THE ANOMALOUS ROLE OF CERAMICS IN LATE

Chemical Analysis (Clay Matrix and Mineral Inclusions)

Petrographic Analysis (Mineral Inclusions)

Table 5.3 Relationship between CPCRUs and sites.

Table 5.4 Relationship between CPCRUs and fabric types.

Table 5.5 Relationship between CPCRUs and vessel form.

Table 5.6 Fabric as indicative of the geographical scale of interaction.

Nature of Production Centres

Comparison and Contextualisation