Vanuatu in the vanguard: new frontiers in Lapita zooarchaeology

1 Introduction

Zooarchaeology, the study of faunal remains from archaeological sites (Gifford-Gonzalez, 2018), has played a foundational role in reconstructing Lapita subsistence strategies and human adaptations to tropical Pacific islands during the late Holocene (Butler, 1988; Davidson and Leach, 2001; Hawkins, 2015; Nagaoka, 1988). Since the 1950s, analysis of faunal assemblages from Lapita sites has contributed centrally to debates about economic activity and settlement patterns (Davidson and Leach, 2001; Green, 1976; Groube, 1971; Kirch, 1997, 2000, 2010), providing direct evidence for the exploitation of marine and terrestrial resources, the introduction of domestic animals, and the ecological consequences of human colonisation (Hawkins, 2015; Kirch, 2021; Steadman, 2006). By identifying and quantifying animal bones, zooarchaeologists have attempted to answer questions ranging from the nature of Lapita dispersal and the disappearance of numerous bird and reptile species which lacked adequate behavioural defences against novel predators (Hawkins and Worthy, 2019; Steadman, 2006).

Over thirty-five years ago, Nagaoka (1988) and Butler (1988) concluded that models of Lapita subsistence were undermined by insufficient zooarchaeological data, due to small and biased sample sizes, poor preservation, inconsistent recovery methods, and vague or qualified reporting, a view still echoed decades later (Anderson, 2009; Hawkins, 2015; O'Connor et al., 2011; Specht et al., 2014). Of the 293 Lapita sites catalogued by Bedford et al. (2019, 2023), most have yielded no faunal remains whatsoever (Figures 1, 2, Tables 1–3). This has severely restricted but not stopped theoretical and methodological advances. While progress in site discoveries, faunal recovery, and Pacific zooarchaeology have gradually been made, the core issues they identified are still an ongoing problem that needs to be addressed.

Table 1. Archaeological sites in New Guinea, and Solomon Islands with reported vertebrate remains from Lapita to immediately post-Lapita contexts.

Table 2. Archaeological sites, New Caledonia, Fiji, Tonga, and Samoa with reported vertebrate remains from Lapita to post-Lapita contexts.

Table 3. Archaeological sites in Vanuatu with reported vertebrate remains from lapita to post-lapita contexts .

After a quick introduction summary of our knowledge about Lapita expansion, Pacific biogeography, prehuman biodiversity, and the strategic location of Vanuatu for studying Lapita dispersal, this review will address six key aims;

• Review methodological and theoretical advances from culture history, historical ecology, Human Behavioural Ecology (HBE), Optimal Foraging Models (OFM), Costly Signalling theory, and Political Ecology and the role played by zooarchaeological methods, and archaeological science (stable isotopes, aDNA, peptide collagen fingerprinting, and archaeobotany).
• Review current knowledge about Lapita subsistence strategies including knowledge gaps by region, and the evolution of the debate between “strandlooper” (Groube, 1971) and mixed foraging/horticulture models (Green, 1991).
• Review human impacts on faunal diversity, including extinctions, resource depression, and the mechanisms driving faunal turnover.
• Review taphonomy and preservation in Lapita sites: How site formation processes, recovery methods, and post-depositional factors affect zooarchaeological interpretation and data quality.
• Position the Teouma case study as an example to follow for method and theory in Lapita zooarchaeology.
• To provide best-practice guidance for future Lapita zooarchaeology studies.

1.1 The Lapita expansion

The Lapita Cultural Complex represents one of the most remarkable colonisation events in human history. Lapita peoples were highly mobile oceanic voyagers (Irwin, 2008) who first appear in the archaeological record of the Bismarck archipelago, Island Melanesia, around 3200 BP (Kirch, 2021). Genetic, archaeological, and linguistic evidence traces their origins to East Asia and Island Southeast Asia, with affinities to Austronesian-speaking populations and distinctive dentate-stamped pottery marking their material signature (Lipson et al., 2020; Pawley, 2002, 2007; Skoglund et al., 2016; Summerhayes, 2010). The New Guinea region had been occupied by non-Austronesian speaking populations for 50,000 years (Summerhayes et al., 2016), both on mainland New Guinea and in the Bismarck archipelago, and certainly during the early Holocene right across the main Solomon Islands chain. This region is known as Near Oceania. Beyond that is Remote Oceania to the North in Micronesia where Neolithic dispersals pre-date Lapita, and to the east where Lapita populations were the first to successfully cross the 350 km water gap separating Near and Remote Oceania, reaching the Reef–Santa Cruz Islands, Vanuatu, and New Caledonia by 3100–3000 BP, and extending to Fiji, Tonga, and Samoa by 2800 BP (Bedford et al., 2023; Burley et al., 2012; Sheppard, 2011). Covering over 4,500 km, this diaspora ranks among the fastest and most extensive colonisation events in human history (Bedford et al., 2023; Sheppard, 2011). Variation in biogeography across this expanse of Oceania undoubtedly affected the availability and range of endemic terrestrial and marine fauna and influenced Lapita subsistence behaviour, with a decline of species diversity as one extends east from the Papuan source, and a stark drop off as terrestrial faunal immigration is constricted at the crossroads from Near and Remote Oceania at the end of the Solomons chain (Figure 1) (Keast and Miller, 1996). Nonetheless, the restriction in diversity was compensated by faunal abundance in a region that had yet to be impacted by the arrival of invasive humans and commensal mammals.

1.2 Vanuatu's pivotal position

Vanuatu occupies a pivotal position within this narrative (Figures 1, 2). Strategically located near the eastern boundary between Near and Remote Oceania, the archipelago was among the first regions in Remote Oceania to be settled by Lapita peoples (Bedford and Spriggs, 2008). To date, more than twenty Lapita sites have been identified across the Vanuatu chain, from the Banks Islands in the north to Aneityum in the south (Bedford, 2006; Bedford et al., 2023; Bedford and Galipaud, 2010). These sites are concentrated in northern and central Vanuatu, particularly around the larger islands of Santo and Malakula and are typically situated on uplifted beach terraces or sheltered offshore islets, locations offering access to rich marine resources and favourable settlement conditions (Bedford and Galipaud, 2010). Recent discoveries continue to expand this distribution, with new Lapita sites with abundant and well-preserved faunal assemblages identified at Pangpang on the east coast of Efate (Bedford et al., 2026), Vao and Uripiv off the northeast coast of Malakula (Bedford et al., 2011; Hawkins et al., 2016a; Ono et al., 2019) and Port Olry (Hawkins et al., 2016a) on northeast Santo in northern Vanuatu, underscoring the archipelago's centrality to early Remote Oceanic settlement.

The Teouma site on Efate Island, discovered in 2004, stands out as one of the most significant Lapita sites in the Pacific, yielding a large cemetery and associated midden deposits (Bedford et al., 2010; Bouffandeau et al., 2019; Hawkins, 2015). Excavated between 2004 and 2010, Teouma is a large Lapita site with approximately 460 square metres of midden and burial contexts exposed. Teouma is remarkable for its exceptional preservation, deeply stratified deposits, and the systematic application of fine-grained recovery methods, including wet screening through fine mesh sizes (1–2 mm) (Hawkins, 2015). Unlike most Lapita sites where faunal remains are scarce, degraded, or incidentally recovered, Teouma yielded tens of thousands of well-preserved fish, reptile, bird, and mammal bones, providing an unprecedented window into early Lapita subsistence practices in Remote Oceania (Bouffandeau, 2021; Bouffandeau et al., 2019; Hawkins, 2015). Teouma provides a rare opportunity to examine Lapita subsistence and ecological interactions at high temporal resolution across the transition from initial colonisation to immediately Post-Lapita cultural phases (c. 3000–2500 BP).

1.3 Biogeographic context and pre-human biodiversity

Vanuatu's biogeographic setting further underscores its importance. The archipelago sits at a crossroads of faunal migration routes, receiving biotic influences from New Guinea via the Solomon Islands chain, and to a lesser extent from New Caledonia and Fiji (Allison, 1996; Hamilton et al., 2010). Its native terrestrial fauna is relatively depauperate compared to Near Oceanic Island groups, consisting primarily of volant mammals (fourteen bat species), birds (over 130 species), and small squamate reptiles (Alcover, 2011; Allison, 1996; Balouet, 1991; Barre et al., 2011; Bauer, 1988; Bregulla, 1992; Chambers and Esrom, 1989). Crucially, archaeological and palaeontological research has revealed that pre-human biodiversity in Remote Oceania was far richer than modern distributions suggest, with numerous endemic bird and reptile species driven to extinction following human arrival (Hawkins, 2015; Hawkins et al., 2016a; Hawkins and Worthy, 2019; Steadman, 2006; Worthy et al., 1999; Worthy and Clarke, 2009). In Vanuatu, extinct and extirpated taxa identified from Lapita contexts include a dwarf land crocodile (Mekosuchus kalpokasi), a giant tortoise (?Meiolania damelipi) of uncertain taxonomic origins, several bird species (e.g., extinctions include Eclectus infectus, an undescribed flightless rail, a large megapode, large tooth-billed pigeon, a giant Ducula sp. and extirpated shearwaters and boobies), and likely multiple fruit bat species (Hawkins, 2015; Mead et al., 2002; Steadman, 2006; White et al., 2010; Worthy et al., 1999; 2015). These discoveries highlight the profound impact of Lapita colonisation on island ecosystems and position Vanuatu as a critical case study for understanding human interactions with naïve ecosystems during initial settlement.

2 Method and theory in Lapita zooarchaeology

Lapita zooarchaeology has largely followed a processual approach that emphasizes materialist ecological explanations of the archaeological record, interpreting Lapita migration, settlement, and economy through frameworks such as culture history (descriptive accounts of material remains; see Green, 1991), historical ecology (the cumulative effects of human settlement on landscapes; e.g., Kirch and Hunt, 1997), and Human Behavioural Ecology (the use of models to generate testable predictions for socioeconomic change within ecological constraints (e.g., Cochrane, 2018; Hawkins, 2015). Post-processual approaches, which foreground human agency, symbolism, and identity in the formation of the archaeological record as alternatives to materialist ecological explanations, have been applied primarily in bioarchaeology, focusing on symbolism in mortuary practices as expressions of social identity, memory, and status (e.g., Valentin et al., 2014), and in ceramic design as a signal of group affiliation or prestige (e.g., Chiu, 2007). In zooarchaeology, post-processual perspectives have been merged with processual ones regarding the signalling of wealth and status by exploiting animals (Hawkins, 2015; Hawkins and Worthy, 2019). The following section provides a brief account of the applications of these theoretical frameworks in Lapita zooarchaeology and the methodologies employed to achieve them.

2.1 Culture history

Culture history emerged as the foundational theoretical framework in Lapita archaeology, critically defining the spatial and temporal boundaries of the Lapita Cultural Complex through the classification of material culture, most notably, distinctive dentate-stamped pottery (e.g., Green, 1991). This approach aimed to construct regional chronologies and trace migration sequences by treating stylistic pottery variation as a proxy for cultural affiliation and population movement. In Lapita subsistence studies, culture history manifested in the formulation of two competing models that continue to shape archaeological discourse. Groube's (1971) “strandlooper” hypothesis characterized Lapita peoples as highly mobile, marine-specialized foragers who rapidly colonized the Pacific through successive resource depletion and settlement abandonment. In contrast, Green's (1979, 2003) “Triple I” model (Intrusion, Innovation, Integration) portrayed Lapita colonists as transporting a “portable landscape” of domesticated plants and animals, practicing a generalized mixed economy that included flexible marine and terrestrial foraging, and which enabled successful long-term settlement across diverse island environments.

These culture-historical debates were foundational to Pacific archaeology and largely used qualitative and descriptive analyses, relying on generalized observations from limited faunal assemblages and pottery distributions to create broad dichotomies between foraging and food production. The approach employed primarily classificatory methods aimed at defining cultural units, with faunal analysis remaining opportunistic and superficial. Researchers such as Gifford and Shutler (1956) and Poulsen (1967) reported animal bone occurrences as generalized lists or anecdotal observations alongside pottery descriptions, with quantification minimal or absent. Recovery methods ranged from coarse screening (often >6 mm mesh) to no sieving at all, introducing severe biases toward large mammal bones while systematically underrepresenting smaller taxa. Taxonomic identifications were often tentative, stratigraphic provenience poorly controlled, and chronological frameworks relied on imprecise radiocarbon dates with little attention to taphonomic issues that could compromise assemblage integrity.

In essence, culture history in Lapita zooarchaeology functioned to authenticate broader narratives of migration and diffusion, without deliberate analytical or interpretive engagement with the bones themselves. Faunal remains were used to support predetermined culture-historical narratives, usually somewhere in between Groube's “strandlooper” hypothesis or Green's transported landscape model, rather than being subjected to rigorous analysis that might challenge or refine these frameworks. This approach ultimately reduced complex, adaptive subsistence strategies to static cultural packages, obscuring the regional and temporal variability in Lapita subsistence behaviour that subsequent research has revealed (Butler, 1988, 1994; Clark et al., 2015; Densmore, 2010; Hawkins, 2015; Nagaoka, 1988). The legacy of culture history was a series of broad generalizations about Lapita economy that, while stimulating debate, could not accommodate the nuanced, dynamic reality of human-environment interactions across the vast expanse of the Lapita expansion.

2.2 Historical ecology

Historical ecology emerged in Lapita studies as a framework for understanding the long-term, dynamic interactions between human colonists and island ecosystems, moving beyond culture history's focus on material culture classification to examine how human settlement transformed previously pristine environments (Kirch and Hunt, 1997, Kirch, 2021; Kirch et al., 1991). This approach integrated zooarchaeological, palaeontological, and palaeoenvironmental data to reconstruct pre-human biodiversity and document the dramatic ecological impacts following Lapita arrival, as revealed across Remote Oceania where terrestrial fauna was far more diverse than modern distributions suggest, with numerous endemic bird and reptile species driven to extinction or extirpation after human colonization.

Methodologically, scientists aligning with the historical ecology framework developed more systematic recovery and analytical techniques than its predecessor. Researchers such as Kirch (2021), Steadman (2006), Burley (1999), Worthy and Clarke (2009), Shaw et al. (2022), and Bedford (2006) employed fine-grained excavation methods including wet screening through 1–3 mm mesh, enabling recovery of small vertebrate remains—fishes, birds, bats, and rats—that had been systematically underrepresented in earlier excavations. Taxonomic identification improved substantially through the development of comparative reference collections and collaboration with palaeontologists, leading to documentation of numerous extinct and extirpated species across Remote Oceania. Methods emphasized presence/absence records, relative dating of faunal turnover, and correlation with palaeoenvironmental proxies such as pollen and charcoal to infer human impacts. Zooarchaeological data were integrated with palaeontological surveys of cave deposits to establish pre-human baselines against which archaeological assemblages could be compared, demonstrating that Lapita colonists caused rapid extinctions of naïve avifauna and reptiles across the region.

Through this approach, researchers demonstrated that Lapita peoples encountered naïve fauna lacking behavioural defences, leading to rapid resource depression and what Steadman (1995, 2006) termed a “blitzkrieg” pattern of extinction. This pattern particularly affected flightless rails, megapodes, pigeons, and giant reptiles, including tortoises, land crocodiles, and iguanas, across island groups from Tonga to Vanuatu. Historical ecology successfully established the scale and pace of human-induced ecological change, documenting what changed (species extinctions, resource depression, habitat modification) and when these changes occurred relative to human arrival. However, the approach relied heavily on circumstantial evidence and correlational reasoning, lacking explicit theoretical frameworks to explain why and how subsistence behaviours transformed over time. While historical ecology documented the consequences of human colonization in unprecedented detail, it treated environmental transformation as a generalized backdrop rather than a process that could be quantitatively tested through hypothesis-driven research, ultimately leaving causal mechanisms—whether direct overkill, habitat modification, introduced predators, or their interaction—largely inferred rather than demonstrated.

2.3 Human behavioral ecology

Human Behavioural Ecology (HBE), an evolutionary theory that assumes people make foraging decisions that maximize reproduction under natural selection (Codding and Bird, 2015), was first applied to Lapita archaeology as a framework for understanding migration and settlement patterns. This was pursued through the Ideal Free Distribution model, which integrated elements of Groube's (1971) “strandlooper” hypothesis and Green (1991) mixed economy model with push-pull dynamics. This model proposed that agricultural carrying capacity, resource depression, and elite conflict pushed populations out of established areas, while the allure of pristine, resource-rich lands pulled Lapita peoples further into Remote Oceania (Cochrane, 2018; Kennett et al., 2006). These early applications operated at a macro-scale, using meta-data from settlement patterns and chronology to generate hypotheses about the drivers of Lapita expansion, but they did not test predictions against fine-grained subsistence data.

The first rigorous, site-level test of HBE in Lapita subsistence research was undertaken at the Teouma site (Hawkins, 2015). Methodologically, HBE at Teouma was built upon the foundations of robust identification protocols and fine-grained recovery practices established by historical ecology, including 1–2 mm wet screening, stratigraphic control, and comprehensive taxonomic identification, but advanced beyond descriptive documentation to hypothesis-driven, predictive modelling using Optimal Foraging Models (Prey Choice, Patch Choice, and Central Place Foraging models) that could generate predictions of subsistence behaviour and its impact on ecosystems over time which could then be tested in the Teouma archaeological record (Hawkins, 2015) (see section 4 below for a detailed account). These were tested with a suite of zooarchaeological analyses (abundance indexes of high vs. low ranked taxa/resource patches, diet breadth, skeletal element abundance, butchery abundance, demographic reconstruction) to demonstrate changing subsistence behaviours, foraging efficiency, intensification of resource use, expanding diet breadth, resource depression and mobility between resource patches local to Teouma (Hawkins, 2015). The results demonstrated that early Lapita foragers at Teouma initially focused on high-ranked, concentrated resources. These were fruit bats from roosting colonies and sea turtles from nesting beaches in proximal locations, optimizing post-encounter return rates. As these patches became depleted, foragers increasingly exploited distant giant tortoise populations, evidenced by changing skeletal element representation that reflected reduced transport costs over time. Pig management intensified, and diet breadth expanded to include lower-ranked prey.

This HBE optimal foraging approach went beyond materialist island adaptations under ecological constraints to also consider socio-political aspects of Lapita behaviour at Teouma by utilizing the theory Costly Signalling. Costly Signalling, which focuses on displays of wealth and power in the archaeological record (i.e., hunting prowess such as diving for sea turtles off the reef; Smith et al., 2003) could be seen as improving social standing and thus reproductive fitness (Codding and Jones, 2007) making it compatible within an HBE framework. Such a framework builds on previous studies that suggest complex social hierarchies in Lapita culture (Bedford et al., 2010; Chiu, 2007; Valentin et al., 2010, 2014). Costly Signalling also has strong parallels with Political Ecology, a theoretical approach that examines how power, inequality, and resource access shape human–environment interactions (Bauer et al., 2007). The rare utilization of Political Ecology approaches in Lapita research focused on how elites establish control over economic bottlenecks, such as long-distance trade, complex technologies, or highly productive lands, which enabled the emergence of social hierarchy and elite power drawing on disparate archaeological datasets (Earle and Spriggs, 2015). While the politics of Lapita can never be known, demonstrating Costly Signalling behaviour under a HBE framework at Teouma was possible by investigating pig management systems that demonstrate wealth and status (Hawkins, 2015), or the use of scarce bird feathers for chiefly headdresses and cloaks (Hawkins and Worthy, 2019).

Unlike culture history's static characterizations or historical ecology's correlational narratives, HBE at Teouma generated testable predictions that were evaluated against multiple independent datasets, demonstrating quantified links between resource depression, declining foraging efficiency, and the suite of subsistence transformations that accompanied the Lapita to Post-Lapita transition at Teouma discussed in greater detail below in Section 4.

2.4 Contributions of modern archaeological science

New science is significantly improving our ability to answer Lapita research questions of ages past. Ancient DNA studies have contributed significantly to two aspects of Lapita vertebrate studies over the last 20 years: (1) tracing the origins of domesticates to East Asia and rats to Island Southeast Asia as a proxy for migration directions and origins (Dobney et al., 2008; Horsburgh et al., 2022; Larson et al., 2007; Lum et al., 2006; Matisoo-Smith, 2007, 2009; Stanton et al., 2026; van Os et al., 2025); and (2) identifying extinct taxa and their phylogenetic relationships. aDNA studies of the giant Vanuatu tortoise have so far been inconclusive regarding origins, initially thought to derive from Australasian Gondwanan sources adjacent to New Caledonia (White et al., 2010), but sea currents and morphological considerations suggest alternative origins (Hawkins et al., 2016a; Sterli, 2015). Peptide collagen finger-printing analysis or Zoarchaeology by Mass Spectrometry (ZooMS) has contributed to taxonomic assessments globally, and more specifically in the Pacific (Oertle et al., 2025) helping refine identifications of bones that are difficult to identify morphologically. Ongoing studies are currently underway in Vanuatu on fish to determine whether species can be narrowed down to specific environments, and on giant tortoises to support aDNA studies in ascertaining the origins of the Vanuatu giant tortoise and answering questions of biogeographic distribution. These methods have an advantage of being relatively inexpensive and applicable to highly fragmented bone.

Stable isotope analysis has contributed to understanding diet using variations in isotopes (δ¹³C, δ¹⁵N, δ³⁴S) to directly reconstruct the diet of humans and pigs as a proxy for human management systems and Lapita economies (Athfield et al., 2008; Bentley et al., 2007; Fenner et al., 2021; Field et al., 2009; Kinaston et al., 2014a, 2014b, 2015; Kononenko, 2007; Leach et al., 2000; Valentin et al., 2010, 2014). Isotope analysis of human remains from a range of Lapita sites indicates a high marine dietary component for colonizing communities that becomes more terrestrially oriented in later generations (Fenner et al., 2021; Herrscher et al., 2018; Kinaston, 2010; Kinaston et al., 2014a, 2014b; Valentin et al., 2014). At Teouma, stable isotope data indicate a diet focused on marine resources and terrestrial plants, with significant inter-individual variation attributed to population movement, gender, and social status (Bentley et al., 2007; Kinaston, 2010; Kinaston et al., 2014a).

3 Current knowledge about Lapita zooarchaeology

What can current data reveal about Lapita subsistence adaptations and their variability across the vast temporal and spatial expanse of Pacific prehistory when examined through the lenses of culture history and historical ecology? The mixed economy model of horticulture and foraging in marine and island terrestrial environments (Green, 1979, 2003; Kennett et al., 2006; Kirch, 1997, 2000) appears to be supported by the archaeological evidence but requires finer-grained temporal resolution. Regional variability in the archaeological record likely stems as much from sampling and taphonomic biases as from genuine ecological differences between Near and Remote Oceania. The long-established human presence and modified landscapes of Near Oceania contrast sharply with the pristine environments and naïve fauna of Remote Oceania, making “strandlooper” and transported landscape concepts still relevant due to the variability of subsistence behaviours across the Lapita distribution (Bedford et al., 2023). Examining the evidence systematically by island group reveals clear patterns in research concentration and knowledge deficits that hinder robust interpretations of subsistence change across both space and time (Tables 1–3).

3.1 Regional patterns and knowledge gaps

Near Oceania is characterised by high potential for Lapita zooarchaeology but in many cases incomplete reporting. Papua New Guinea contributes eleven sites with meaningful assemblages, most notably Talepakemalai in the Mussau Islands, where abundant faunal remains (Over 50,000 bones) were recovered with 3–5 mm dry sieving from undisturbed stratified deposits (Kirch, 2021). At Talepakemalai, detailed analysis has been published for bird, reptiles and mammal remains (Kirch, 2021; Steadman and Kirch, 1998), and fish (Butler, 2021). The overwhelming majority of numbers of identified specimens (NISP) were fish (90%) with only very small numbers of endemic birds, mammals, and reptiles as well as domesticates (pig 73 NISP, dog 6 NISP, chicken 10 NISP) (Kirch, 2021). Gutunka in the southern New Guinea region of the Massim archipelago had 3,553 vertebrates recovered (74% fish) using 3 mm wet sieving including small numbers of translocated wild fauna (Phalanger NISP 4) and domesticates (Pig 11 NISP, dog 3 NISP) (Shaw et al., 2022). Excavations at other sites have small numbers of identified specimens that are well documented, including the early Lapita site of Emirau (1862 NISP) (Summerhayes et al., 2010), Kamgot on Anir (1,311 bones) (Summerhayes et al., 2019), and the Arawe Islands (Adwe 815 bones; Apolo 378 bones) (Henderson, 2021). Again, most of these sites contain predominantly fish remains (Eimrau 93%, Kamgot 43%, Adwe 43%, Apolo 17%), and few endemic and domesticate fauna (Kamgot pig NISP 9, dog NISP 8; Adwe pig NISP 18, dog NISP 1; Emirau pig NISP 1, dog NISP 1). The Kainapirina (SAC) site on Watom in New Britain, though disturbed to some level, provides some of the best evidence for Lapita pig husbandry (251 pig bones) in Near Oceania (Green and Anson, 2000; Smith, 2000). More promising is recent work at Caution Bay on the south coast of New Guinea, where well-preserved stratified vertebrate remains were quantified (David et al., 2022; McNiven et al., 2011, 2012). Further east in the western Solomon Islands, Late Lapita and immediately Post-Lapita sites exist but are largely located in the inter-tidal or submerged zones and have yet to yield any preserved vertebrate remains (Felgate, 2007).

Remote Oceania presents an increasingly uneven record. In the southeast Solomon Islands, the Reef Santa Cruz sites like SE-RF-2 have produced some analysed fishbone (882 bones) and shellfish assemblages (Green, 1986; Swadling, 1986), but most sites suffer from disturbance and poor preservation. Tikopia stands out as an exception, with well-stratified deposits yielding one of the largest and most detailed faunal studies (35,000+ bones) in the Pacific (Kirch and Yen, 1982; Steadman et al., 1990). However, there is no extinct fauna reported and along with the pottery assemblage recovered the bulk of the material appears to be immediately post-Lapita (Kirch and Yen, 1982).

Vanuatu shows a trajectory of increasing research activity but delayed outputs. Before 1995, zooarchaeology was virtually non-existent, with only qualified reports from disturbed sites like those from Malo Island (Hedrick, n.d., 1971). Through the 1990s and 2000s, excavations at sites such as Arapus/Mangassi on Efate, Ponamla and Ifo on Erromango, Yalo on Malakula (Bedford, 2000, 2006), and Atanoasao on Malo (Galipaud, 1998) produced quantified but small assemblages. More recent excavations on the Malakula offshore islets of Vao, Uripiv, Atchin, and Wala recovered relatively large vertebrate assemblages (Bedford, 2007), but final analysis and publication are ongoing (Table 3). The fish and giant tortoise have been published in detail from Vao and Uripiv indicating dense vertebrate accumulations (Hawkins et al., 2016a; Ono et al., 2019). Even well-stratified early sites like Makué on Aore Island await full analysis (Galipaud and Swete-Kelly, 2007). Recent discoveries of larger well-preserved vertebrate samples at Pangpang on Efate (Bedford et al., 2026), and Port Olry (Hawkins et al., 2016a) on Santo in northern Vanuatu add to the inventory but analysis is on-going and results remain unpublished, leaving the transition from earliest Lapita through Post-Lapita periods as generally a substantial gap in understanding. The Teouma site, further detailed below, is a significant exception providing a rich assemblage (Hawkins, 2015) associated with the colonizing phase of Lapita through to several hundred years later.

New Caledonia and Fiji present contrasting pictures. New Caledonia is among the best-studied regions for Lapita marine exploitation, with quantified fish assemblages from sites including Lapita, St Maurice-Vatcha, and Goro (Davidson et al., 2002; Sand, 1999; Sand et al., 2001). However, terrestrial faunal interactions remain largely unstudied, although the Pindai Cave site has yielded the bones of numerous extinct reptile and bird species with uncertain human associations (Anderson et al., 2010). In Fiji, numerous Lapita sites have been excavated, but most are shallow, disturbed deposits yielding degraded and small assemblages (Irwin et al., 2011; Worthy and Clarke, 2009). The Qaranipuqa rock-shelter on Lakeba stands out as one of the few well-stratified, dated, and quantified Lapita faunal collections (Best, 1984). Bourewa, one of the most extensively excavated Lapita sites in Fiji at over 200m², has in situ midden deposits spanning 2900-2500 BP, but the vertebrate analysis remains unpublished (Nunn, 2007, 2009; Nunn and Petchey, 2013).

Earlier works in Tonga on Niuatoputapu (NT-90) paved the way with smaller assemblages (Kirch, 1988, Kirch and Dye, 1979). Tonga currently offers one of the most productive regions for understanding Lapita ecological interactions, due to systematic fine-grained recovery methods over several decades (Burley, 1999, Burley, 2023; Burley et al., 2001; Clark et al., 2015). Excavations in the Ha'apai group at multiple sites including Touma'a Well, Tongoleleka, and Holopeka, conducted with fine mesh sieving, yielded over 100,000 vertebrate remains spanning Lapita and Polynesian plainware periods (Burley, 1999; Pregill and Steadman, 2004; Steadman et al., 2002a). Bird and reptile extinctions following Lapita arrival have been documented in exceptional detail (Steadman, 2006). Subsequent work in the Vava'u Islands has added further stratified deposits with fauna under analysis (Burley and Connaughton, 2007; Densmore, 2010). The contrast with Samoa, which has only one confirmed Lapita site (Mulifanua, submerged with no fauna) and the immediately Post-Lapita site of To'aga on Ofu Island with quantified fish and bird remains (Kirch and Hunt, 1993; Nagaoka, 1993; Steadman, 1993a), underscores the regional patchiness of the record.

3.2 Lapita marine resource exploitation

Across the Lapita distribution, marine resources formed the foundation of subsistence economies, comprised a broad spectrum of highly adaptive planned behaviours with opportunism over time and space, adjusting to varying social and marine environmental situations (Bouffandeau, 2021; Butler, 1994; Clark et al., 2015; Ono, 2003; Ono et al., 2019; Walter, 1989). Lapita communities exploited every niche of rich marine systems, from deep-sea and pelagic fish to lagoons, inner shores, reefs, and beaches with abundant fish, shellfish, and crustacea. Evidence from fishhook assemblages and pumice floats indicates diversity of fishing techniques and technologies including drop-line fishing, trolling with nets, spears, traps, and poisoning (Bouffandeau, 2021; Butler, 2021; Ono et al., 2019). While the wider marine environment was exploited, the overwhelming component of fish species recovered from sites are lagoon and reef fish, those most easily accessible, requiring more limited resources and skills (Bouffandeau, 2021; Butler, 2021; Cannon et al., 2019; Ono et al., 2019).

At the Mussau sites, 49,848 fish bones were recovered, with 4,687 identified to forty-four taxonomic categories. While deep-sea fish such as shark and tuna were present, the vast majority were inshore species (Butler, 2021), a pattern that has been documented throughout the Lapita distribution for some time now (Butler, 1988; Butler, 1994). At Gutunka (Brooker Island), 2,627 finfish and shark/ray vertebrates of near shore species were recovered (Shaw et al., 2022). In Vanuatu, the Teouma assemblage contains 39,310 fish bones, with 13,047 identified to thirty-five families (Bouffandeau, 2021). Reef species dominated, but the large nearby river also proved a well-exploited niche, Teouma is one of the few Lapita sites where freshwater fish species are prominent, representing 16% of the sample (Bouffandeau, 2021; Bouffandeau et al., 2019). In Tonga, a similar pattern of reef species dominance occurs at all sites, with only a very small component of deep-sea species (Burley, 2023; Cannon et al., 2019). A few studies of shellfish and fish from Remote Oceania indicate some level of marine resource depression in Lapita and Post-Lapita sites (Butler, 2001; Densmore, 2009; Gyngell, 1997; Heorake, 2009; Spennemann, 1987).

Sea turtles were identified across all regions as taxa regularly targeted during the initial Lapita period (Allen, 2007; Burley, 2023; Frazier, 2003; Hawkins, 2015; Kirch, 2021; Shaw et al., 2022). At St Maurice-Vatcha in New Caledonia, abundant “turtle” bones were quantified (NISP = 1,612; Sand, 1999), these could possibly include terrestrial meiolaniid specimens. At Gutunka and Teouma, sea turtles were targeted initially in great abundance, resulting in resource depression over time (Hawkins, 2015; Shaw et al., 2022).

3.3 Lapita terrestrial resource exploitation

Moving from a pioneering phase to settlement occurred rapidly as gardens were established and naïve fauna driven to extinction (Bedford et al., 2023; Burley, 2023). The focus of subsistence became increasingly garden-based over time. Isotope analysis of human remains from a range of Lapita sites indicates a high marine component for colonizing communities that becomes more terrestrially oriented in later generations (Herrscher et al., 2018; Kinaston et al., 2014b; Valentin et al., 2014).

The pattern of terrestrial vertebrate exploitation shows a very different composition from the largely homogenous marine pattern, reflecting variation between previously inhabited landscapes in Near Oceania vs. pristine landscapes and regional biodiversity differences from west to east already discussed above.

Some sites in Near Oceania are proportionately far more dominant in fish remains compared to Remote Oceania (Bedford et al., 2023). At Mussau, 6,186 non-fish vertebrates (10%) were recovered (Kirch, 2021). After removing turtle and porpoise, 4,337 bones remain, the vast majority identified only as medium mammal (333) or vertebrate (3,325), with pig (229), birds (103), and fruit bats (30) being the largest identified specimens. At Gutunka on Brooker Island in the Massim region only 26% of vertebrates were terrestrial vertebrates including Cuscus (Phalanger) which suggests that translocation of wild species from the New Guinea mainland was a strategy for island colonisation (Shaw et al., 2022). Few Near Oceanic Lapita vertebrates belong to extinct species, which is not surprising since endemic species had already coevolved with humans and invasive mammals for several millennia in Island Melanesia (Flannery et al., 1988; Steadman, 2006).

At Teouma, 49,212 non-fish bones were identified; removing marine species reduces the total to 46,737 bones, more than 50% of the recovered faunal assemblage. Species dominating the collection include fruit bat (21,766), pig (5,335), bird (3,038), rat (2,234), and tortoise (1,915) (Hawkins, 2015). The abundance and morphological diversity of fruit bat vertebrates recovered from Teouma and now observed at other Vanuatu sites including Vao, Uripiv, and Pangpang, far exceeds that reported from elsewhere in the Lapita distribution (Hawkins, 2015, Hawkins et al., 2016b). This disparity likely reflects either genuine biogeographical variation in bat populations between regions, cultural practices and/or, alternatively, methodological artefacts arising from differences in sampling intensity and recovery techniques.

3.4 Lapita impacts on island terrestrial fauna

Identifying novel terrestrial taxa has been critical for understanding human impacts and faunal turnover during the Lapita dispersal. However, establishing the true scale of human impacts on fauna in Oceania is compounded by the limited palaeontological record, which is essential for reconstructing pre-human ecologies within the region of Lapita expansion. While New Caledonia, Fiji, and Tonga possess reasonable fossil records (Anderson et al., 2010; Balouet, 1991; Balouet and Buffetaut, 1987; Balouet and Olson, 1989; Koopman and Steadman 1995; Molnar et al., 2002; Worthy, 2000, 2001, 2004; Worthy et al., 1999), Vanuatu has none (Alcover, 2011). Systematic palaeontological analysis, relying on comparisons with known type specimens, has been spectacularly successful in documenting the disappearance of several bird and reptile species following Lapita arrival (Hawkins et al., 2016a; Steadman, 2006; Worthy and Clarke, 2009).

Far fewer extinctions have been documented in Near Oceanic Lapita sites, a consequence of several millennia of coevolution of fauna with mammals and the refuge capacity of inter-visible island groups close to Papua (Steadman, 2006). Large endemic reptiles were recorded, such as the saltwater crocodile, but no extinct reptiles have been identified in Near Oceania (Kirch, 2021). In total, fifty bird species are recorded as having become extinct or extirpated shortly after Lapita arrival across Near and Remote Oceania (Hawkins and Worthy, 2019). Only four of these are identified in Lapita sites in the Bismarck's, including a petrel (Pterodroma sp.), brown booby (Sula leucogaster), cockatoo (cf. Cacatua sp.), and owl (Tyto sp.) (Kirch, 2021; Steadman and Kirch, 1998). For Remote Oceania, the story is far different. In isolated island contexts, naïve endemic bird and reptile species experienced rapid biodiversity loss soon after Lapita arrival. Of the fifty extinct bird species throughout the Lapita dispersal, 46 are from Remote Oceania (Hawkins and Worthy, 2019). These sites also revealed a suite of large-bodied endemic reptile extinctions (Hawkins et al., 2016a).

In New Caledonia, the ancient Gondwana-derived fossil fauna, mostly from the Pindai Caves, reveals at least 20 extinct birds, a few bat species, as well as large-bodied reptiles including giant horned meiolaniid turtles, the extinct land crocodilian Mekosuchus inexpectatus and a monitor lizard (Varanus sp.) (Anderson et al., 2010; Balouet and Buffetaut, 1987; Balouet and Olson, 1989; Hand and Grant-Mackie, 2012). Birds include two giant moa-like galliforms: Sylviornis neocaledoniae, a massive flightless bird (∼30 kg) considered one of the most distinctive extinct birds of the Pacific region, and Megapodius molistructor, a slightly smaller giant megapode that used vegetation mounds to incubate its eggs. These extinct species have not yet been found in association with Lapita sites in New Caledonia due to previously identified sampling and methodological issues.

In Vanuatu, extinct species have yet to be identified in pre-human fossil-bearing deposits (Alcover, 2011). However, Lapita sites in Vanuatu have revealed the extinction of at least 15 endemic species of birds, mostly ground-dwelling, forest, and wetland species (Hawkins and Worthy, 2019; Worthy et al., 2015). Additionally, the extinct dwarf land crocodile (Mekosuchus kalpokasi) has been identified in Lapita sites in Vanuatu (Hawkins, 2015; Mead et al., 2002), along with the giant tortoise (?Meiolania damelipi) (Hawkins, 2015; Hawkins et al., 2016a; White et al., 2010), and possibly several extinct bat species (Hawkins, 2015; Hawkins et al., 2016b). The extinct Vanuatu giant tortoise has been documented at five sites in Vanuatu (Port Olry, Vao, Uripiv, Arapus, Teouma) (Hawkins et al., 2016a). Since that study, the list has grown with the recent discovery of the Pangpang site on eastern Efate (Bedford et al., 2026), which has yielded over 400 tortoise bones to date in exploratory test pits. Specimens from all localities exhibit consistent morphological traits that differ markedly from the horned meiolaniid tortoises known from New Caledonia (Hawkins et al., 2016a; Sterli, 2015).

In Fiji, at least 14 extinct or extirpated bird species have been identified from Lapita sites and fossil deposits at Wainibuku Cave, Viti Levu (Hawkins and Worthy, 2019; Irwin et al., 2011; Worthy, 2000, 2004; Worthy et al., 1999). Large bodied extinct novel reptiles have also been identified from Lapita sites and Wainibuku Cave including the giant tortoise?M. damelipi at Naigani and Yanuca Lapita sites (Hawkins et al., 2016a), and an extinct giant iguana (Lapitiguana impensa), likely of Neotropical origins (Irwin et al., 2011; Worthy et al., 1999). The dwarf land crocodile Volia athollandersoni was identified in Wainibuku Cave, Viti Levu (Molnar et al., 2002; Worthy et al., 1999), and likely also at the Naigani Lapita site (Irwin et al., 2011). Additionally, a giant ground frog (Platymantis megabotoniviti) identified at Wainibuku Cave, was a large, extinct amphibian that once inhabited Viti Levu, and disappeared 3000 years ago (Worthy, 2001).

As far east as Tonga, faunal diversity was much lower, with giant tortoises and land crocodiles absent from fossil and Lapita sites. However, extensive documentation exists of 13 bird extinctions following Lapita arrival (Hawkins and Worthy, 2019; Steadman, 2006) and an extinct species of giant iguana (Brachylophus gibbonsi), likely of Neotropical biogeographical origins to the east (Pregill and Dye, 1989; Pregill and Steadman, 2004; Steadman et al., 2002b).

The mechanisms of faunal extinction during the Lapita expansion remain debated. Overkill/blitzkrieg models dominate archaeological interpretations (Bedford, 2006; Martin and Steadman, 1999; Steadman et al., 2002b), proposing that concentrated human predation drove rapid extinctions of naïve fauna. Few analyses have quantified changes in terrestrial endemic prey populations due to sustained Lapita hunting pressure that led to reductions in prey age or size (see Hawkins, 2015). Recent palaeoecological models indicate a more nuanced process involving multiple interacting factors, including introduced mammalian predators (rats, pigs, dogs) and forest clearance from swidden horticulture (Hawkins, 2015; Prebble and Wilmshurst, 2009; Steadman, 2006).

3.5 Lapita horticulture and animal husbandry

Archaeobotanical evidence from Near and Remote Oceania Lapita sites (Clark et al., 2015; Crowther, 2005, 2009; Horrocks et al., 2009; Horrocks and Bedford, 2005, 2010; Horrocks and Nunn, 2007; Lentfer and Green, 2004; Mathews and Gosden, 1997; Tromp et al., 2020) confirms the presence of introduced cultigens across the Lapita distribution. Microfossil remains include taro, yam, and banana (Burley et al., 2020; Horrocks et al., 2009; Ussher, 2015), and many fruit and nut trees were likely distributed by Lapita colonists. At the Mussau site, a full suite of macrobotanical remains provides direct evidence of plant exploitation (Kirch, 2021). At Teouma, dental calculus studies reveal the importance of forest plant species and arboriculture (Tromp et al., 2020). Although Remote Oceania had a reduced selection and number of edible species compared to New Guinea, there was no shortage of such species, many already familiar to Lapita colonists (Lebot and Chanel, 2019). Isotopic analyses of human teeth from Lapita skeletal remains indicate that forest species were a significant dietary contributor (Roberts et al., 2023).

Pigs are critical to the socio-political economies of Pacific Island communities today (Hide, 2003; Rappaport, 1968), with Vanuatu having highly elaborate feasting rituals for chiefly status and the establishment of circular pig tusks (Bedford, 2018; Bonnemaison et al., 1996). Thus, pig husbandry serves as a key proxy for social complexity and agricultural intensification in Lapita economies (Hawkins, 2015). Regional variation is pronounced across the Lapita distribution from west to east. Pigs have been identified at some of the earliest Lapita sites in the Bismarck Archipelago (Kirch, 2021). At the disturbed Kainapirina (SAC) site on Watom Island, pig age profile data demonstrate that pigs were managed for slaughter during their second year rather than hunted (Smith, 2000). The earliest reliable date for pig in the New Guinea region is a single tooth from Watom dated to 2760–2547 cal BP (Athfield et al., 2008), while at Caution Bay the earliest pig dates to 2500–2350 cal BP (McNiven et al., 2012). At Gutunka, the presence of a single pig bone immediately predating Lapita pottery was used as evidence for Lapita interactions with Indigenous peoples of the Massim region in southern New Guinea (Shaw et al., 2022). In Remote Oceania, pig bones are reported in small quantities in the Reef-Santa Cruz Islands and Tikopia (Green, 1976; Kirch and Yen, 1982), and yet to be reported in New Caledonia Lapita sites (Hawkins, 2015). It is in Vanuatu that pig bones have been found in Lapita sites in great abundance, particularly at Teouma where over 5,000 pig bones were identified (Hawkins, 2015), as well as at Vao, Uripiv, and Pangpang, with detailed analyses currently being compiled for publication. At Teouma, demographic, butchery, and skeletal element abundance data indicate a complex pig husbandry management system that included raising some pigs for feasts while most were free-range, with some clearly ritualized practices occurring in association with the cemetery (Hawkins, 2015). Farther east, however, evidence for pigs is absent from early Lapita sites in Tonga (Anderson, 2009; Burley, 1999) and Fiji (Clark and Anderson, 2009).

Dog remains are considerably rarer and more contentious. Present at Talepakemalai (NISP = 6), Kamgot (NISP=8), Gutunka (NISP=3) and Adwe (NISP=1) in Lapita contexts, and in very small numbers (Henderson, 2021; Kirch, 2021; Summerhayes et al., 2019) the earliest reliably dated dog bones come from Caution Bay (2500–2350 cal BP; McNiven et al., 2012) and Post-Lapita deposits on Buka Island (1900 cal BP; Flannery et al., 1988). Dogs do not appear in Remote Oceania until much later than Lapita (Anderson, 2009; Hawkins, 2015). Chickens had a widespread distribution in Remote Oceanic Lapita sites but were largely absent from early Near Oceanic levels (Storey et al., 2008), though small amounts reported at Mussau and Watom suggest that sampling and preservation biases may obscure their true prevalence. The Pacific rat (Rattus exulans) dispersed with Lapita communities throughout the entire Lapita distribution (Matisoo-Smith, 2007, 2009). Previously, the larger New Guinea spiny rat (R. praetor) was documented only in immediately Post-Lapita or late Lapita deposits in Vanuatu and Fiji (White et al., 2000). However, analysis at Teouma has demonstrated that the large New Guinea spiny rat was introduced by Lapita colonists in substantial numbers, declining relative to the much smaller Pacific rat (R. exulans) over time, a pattern that illuminates the complex dynamics of commensal populations under changing human settlement and land-use regimes (Hawkins, 2015).

3.6 Persistent taphonomy and methodological problems

Across all regions, a cluster of persistent problems has limited zooarchaeological interpretation. Taphonomic issues are pervasive. Lapita sites are often shallow, mixed, or submerged intertidal deposits subjected to a range of post-depositional processes including, agricultural disturbance, bioturbation, acidic water saturation, and coastal erosion (Anderson, 2009; Bedford, 2006; Green, 1979). Many sites yield little or no bone, and recovered specimens are frequently in poor condition (Bedford et al., 2023). These taphonomic factors are rarely addressed systematically in site reports, despite their fundamental importance for assessing vertebrate assemblage integrity.

On exceptional study which addressed taphonomy was Steadman et al. (2002b) who analyzed ∼500 bird bones from three cultural sites in Tonga (3000 cal BP–AD 1800), comparing taphonomic attributes (element representation, breakage, burning, cut marks, rodent gnawing, age) between domestic chickens and native birds. Cut marks occurred on only 2%–3% of specimens and burning on ∼8%; chickens and native birds exhibited nearly identical preservation patterns. Given the unequivocally cultural context and lack of non-human depositional agents, the authors argue that native bird bones, like chicken bones, entered the sites as human food refuse. They conclude that bird bones should not be held to a higher standard of direct anthropogenic modification (e.g., cut marks) than other vertebrate remains; rather, contextual and comparative taphonomic evidence suffices to establish cultural origin. At Teouma, detailed taphonomic studies of pig skeletal element representation considered the possibility that post-depositional bone destruction, rather than cultural practices, could explain observed patterns (Hawkins, 2015), given that denser bones tend to survive preferentially (Lyman, 1994). To evaluate this, percentage survivorship of each skeletal element was calculated by comparing observed Minimum Number of Element values against expected counts derived from Minimum Numbers of Individuals, and these values were plotted against artiodactyl bone density indices (Lyman, 1982). Statistical testing revealed no significant correlation between skeletal element survivorship and bone volume density in any temporal unit, confirming that taphonomic destruction did not substantially bias the pig assemblage and supporting the validity of zooarchaeological inferences about cultural behaviour at Teouma (Hawkins, 2015). These studies are rare and highlights what could be achieved with well-preserved deposits.

Some studies have demonstrated the effects of sieve size on the recovery of vertebrates from Pacific sites (Densmore, 2009; Hawkins, 2015; Nagaoka, 1994; 2005a). Recovery methods have initially been coarse, with large mesh sizes or limited sieving during earlier phases of research (e.g., Garanger, 1972; Gifford, 1951; Gifford and Shutler, 1956; Poulsen, 1967, 1987), and many publications fail to explicitly describe recovery techniques, compromising assessment of sample integrity. The adoption of fine-grained wet screening in recent decades, at Kulu Bay (Clark and Szabó, 2009), Caution Bay (McNiven et al., 2011, 2012), Massim region (Shaw et al., 2022), New Caledonia (Davidson et al., 2002), as well as Vanuatu (Bedford, 2006; Bedford et al., 2010, Hawkins, 2015) and Tongan sites (Burley, 1999), has dramatically increased sample sizes and species diversity, demonstrating the method's critical importance.

Even when adequate samples are recovered, analysis often remains superficial, limited to taxonomic lists or focused on select taxa (e.g., extinct birds and reptiles), without investigating skeletal element frequencies, taphonomy, or demographic changes that could illuminate human-ecological interactions. Earlier studies were based on limited collections often utilising poorly preserved remains with insecure contexts from where it was difficult to make definitive conclusions. This was the case for example with the debate over the presence or absence of pigs being associated with Lapita. Very few secure pig bones were identified, and scholars justifiably questioned whether Lapita communities did travel with pigs (O'Connor et al., 2011; Specht et al., 2014). It was not until larger secure quantities of vertebrates, including pig bones were recovered from well-preserved sites both in Near and Remote Oceania, such as Talepakemalai and Teouma (Hawkins, 2015; Kirch, 2021) that the question of Lapita animal domestication began to be clarified.

Chronological control has been another major challenge in Lapita studies over many decades. A whole range of factors such as in-built age of samples, the marine reservoir effect associated with shell and the suitability of different species have all been refined and the once early and lengthy period proposed for Lapita has been greatly reduced. However, the radiocarbon calibration plateau (c. 2650–2350 BP) that covers the Late Lapita to immediately post-Lapita period, greatly distorts estimated occupation periods, obscuring the precise timing and nature of subsistence shifts during this period of transition (Bedford et al., 2023; McNiven et al., 2011, 2012; Petchey et al., 2014). This presents a unique challenge for demonstrating fine-grained changes in subsistence behaviour and extinction events across this critical period (see Lyman, 2003 for effects on the application of foraging theory).

Although many bird, bat, and reptile extinctions likely occurred within decades after human arrival (Bedford et al., 2023), precise dating remains difficult due to limited direct dating of extinct fauna.

4 The Teouma case study

The Teouma case study demonstrates that when well-preserved sites are excavated with appropriate methods and analysed using robust quantitative techniques, Lapita zooarchaeology can move beyond taxonomic listing to address questions of socio-political economy, ritual practice, and human impacts on island ecosystems (Hawkins, 2015). By overcoming the taphonomic and methodological problems—small sample sizes, recovery biases, and poor preservation—that have long hindered previous research, the Teouma vertebrate assemblage enabled detailed quantitative analyses of taxonomic composition, diet breadth, skeletal element frequencies, demographic profiles, and butchery practices. Applying a Human Behavioural Ecology (HBE) framework allied with Optimal Foraging Models (Prey Choice, Patch Choice, Central Place Foraging), Hawkins (2015) generated testable predictions for Lapita subsistence activity at Teouma, yielding insights into subsistence change, socioeconomic developments, and ecological impact across the Lapita to Post-Lapita transition. As such, Teouma resets expectations for what is achievable in the region and provides a model for future research design.

The Prey Choice Model predicts that high-ranked taxa, those offering the highest net returns in protein/fat and energy after accounting for search, capture, and processing costs, will be targeted first (Nagaoka, 2002). Intensive predation leads to declines in prey abundance and foraging efficiency, subsequently increasing the targeting of lower-ranked species and broadening the diet (Nagaoka, 2002). At Teouma, this model was tested using diet breadth and diversity indices, along with quantitative abundance indices that track changes in the ratio of high to low-ranked taxa (e.g., tortoise to fruit bat, sea turtle to sea bird) across seven well-dated spatio-temporal units to monitor foraging efficiency in different resource patches (Hawkins, 2015). Bone morphometrics (tortoise humeri measurements) directly linked human predation on giant tortoises at Teouma to resource depression, demonstrating declining prey ages under sustained hunting pressure, moving beyond the circumstantial correlations of earlier approaches (Hawkins, 2015). Butchery intensity was quantified using multiple metrics, including NISP/MNE (Number of Identified Specimens/Minimum Number of Elements) ratios, cut mark frequencies, and cultural fracture proportions, measuring the intensification of carcass processing in response to declining foraging efficiency.

The Patch Choice Model predicts that taxa cluster in specific resource patches and that foragers will concentrate on patches offering the highest net returns, shifting to new patches once existing ones are depleted (Nagaoka, 2002). This approach required significant theoretical adaptation, as traditional foraging models were developed for hunter-gatherer economies (Codding and Bird, 2015) and had to be modified to accommodate Lapita's mixed economy, which combined marine and terrestrial foraging with horticulture and animal husbandry (Hawkins, 2015). By dividing vertebrate taxa into three discrete resource patches, terrestrial (fruit bats, birds, tortoises), coastal (sea turtles, fish, shellfish), and domestic (pigs, rats, chickens), the framework generated specific, testable predictions about how Lapita peoples would respond to resource depression and declining foraging efficiency over time (Hawkins, 2015).

Central Place Foraging models predict that once a proximal resource patch is depleted, foragers will travel farther from a central base and will make transport and butchery decisions to reduce distance-related costs by bringing only high ranked skeletal elements back to base (Nagaoka, 2005b). Skeletal element representation analysis was used to test these predictions at Teouma as distance-related costs changed over time. This required the development of ordinal meat utility rankings for turtles, a taxon for which no standardized utility indices existed, and the application of established Meat Utility Indices (MUI) for pigs (Rowley-Conwy et al., 2002).

To test Costly Signalling at Teouma, two lines of evidence were used (Hawkins, 2015): pig management strategies using pig epiphyseal fusion and tooth eruption/wear analysis (slaughtering pigs beyond optimal ages signals feasting for status and wealth; Hide, 2003), and the presence of worked turtle shell and raptor wing bones. These are species and skeletal elements unlikely to have been consumed as food. Large flight feathers from these raptor wing skeletal elements would have been ideal for ceremonial headdresses and cloaks (Hawkins and Worthy, 2019), serving as displays of chiefly status, while the turtle shell artefacts may have served some ceremonial purpose (Hawkins, 2015).

This approach enabled the identification of three broad temporal exploitation phases at Teouma, documenting major shifts in subsistence behaviour across the Lapita to Post-Lapita transition. These phases are defined by changes in foraging efficiency, resource patch selection, and the socioeconomic role of animals, tracked through multiple zooarchaeological datasets including prey diversity, relative abundance, demographic profiles, skeletal element representation, and butchery intensity. Three Lapita exploitation phases were quantified using the Cochran-Armitage test of linear trend (Hawkins, 2015; Zar, 2010) as follows;

Exploitation Phase 1 (EP1) was characterized by uneven specialization on concentrated, proximal resource patches by early Lapita. Initial foraging focused intensively on fruit bat roosts and sea turtle nesting sites, with fruit bats dominating the terrestrial patch despite their small body size, mass harvesting of densely aggregated colonies yielded high net returns. Large-bodied sea turtles and giant tortoises featured prominently but declined rapidly by the end of EP1, while large-bodied crocodiles, never abundant, disappeared very quickly. Settlement intensity increased markedly, evidenced by rapid faunal deposition, yet foraging efficiency began declining due to resource depression and agriculturally induced landscape change. Forest bird abundance decreased relative to wetland birds, signalling habitat destruction, and several bird and fruit bat species disappeared from the archaeological record entirely by the end of EP1. Giant tortoises initially contributed less to the diet relative to sea turtles due to distance-related transport costs, with only high-quality meat portions returned to the site. Within the domestic patch, pig abundance increased rapidly alongside high infant pig mortality, attributable to nutritional deficiency, selective culling to reduce labour costs, or ritual use. Evidence for costly signalling appears through pig feasting and the use of bird feathers and turtle shells for status display.

Exploitation Phase 2 (EP2), this later Lapita phase witnessed a dramatic intensification of settlement and population growth. Tortoise exploitation reached its zenith despite declining mean prey age and increasing distance-related costs, as foragers compensated for the depletion of proximal fruit bat and sea turtle resources. This intensification is reflected in increasing proportions of intermediate-ranked tortoise skeletal elements being returned to the site, indicating that transport cost considerations diminished as high-ranked local patches became exhausted. Sea turtle populations continued to decline, while forest birds further decreased relative to wetland birds, with several additional species, including the Tongan tooth-billed pigeon, disappearing from the record. Foraging efficiency continued to decline as subsistence broadened to include a more even range of taxa. Within the domestic patch, pig relative abundance peaked, with animals serving multifunctional socioeconomic roles spanning household consumption and displays of wealth. Infant pig mortality improved, likely reflecting either more established gardens providing better nutrition or less intensive management strategies. Pigs were butchered more intensively to maximize protein returns. Rat abundance increased, resulting in declining rat ages and a decrease in the abundance of the larger Spiny New Guinea rat relative to the Pacific rat. This may have been the result of inter-species competition or human predation, although this is difficult to ascertain in taphonomic studies (Colman, 1998).

Exploitation Phase 3 (EP3), the early Post-Lapita period saw a dramatic fall in foraging efficiency coinciding with the cultural transition from Lapita to Post-Lapita. Faunal deposits became sparse despite continued intensive settlement, signalling that localized ecologies had been severely degraded. Sea turtle exploitation became opportunistic and greatly reduced, targeting only individuals foraging on the reef or within the harbour rather than breeding colonies. The giant tortoise population was devastated, either already extinct or on the brink of extinction, with cascading effects on co-evolved and co-dependent ecologies. Pig production declined markedly, though butchery of remaining animals intensified. Rat populations continued to be affected by human activity. The settlement appears to have been winding down, with subsistence increasingly focused on vegetable consumption. A notable increase in the proportion of high-utility pig meat portions may indicate either a switch to feral pig hunting, with associated transport cost decisions, or the importation of high-ranked meat portions for ritual feasting. By the end of EP3, the Teouma settlement was abandoned, likely as inhabitants relocated to more fertile agricultural land or other coastal zones to maintain access to marine resources.

5 Best future practices for Lapita zooarchaeological analyses

Recent discoveries of new sites through systematic island field surveys continue to expand the Lapita site inventory and offer potential for addressing current knowledge gaps. In Vanuatu, new Lapita sites at Vao, Uripiv, Pangpang, and Port Olry in northern Vanuatu add to the distribution but await detailed analysis (Bedford et al., 2026; Hawkins et al., 2016a; Ono et al., 2019). The recent discovery of a Lapita site on Aneityum (Bedford et al., 2016), Vanuatu's southernmost inhabited island, extends the known range and offers potential for examining latitudinal variation in subsistence adaptations. In the Massim region of Papua New Guinea, recent research has identified Lapita sites that promise to illuminate interactions with resident Papuan populations (Shaw et al., 2022). At Caution Bay, ongoing analysis of well-preserved vertebrate assemblages continues to provide insights into Lapita subsistence on the New Guinea mainland (David et al., 2022; McNiven et al., 2011, 2012). New Caledonia is in serious need of reanalysis of Lapita vertebrate collections, particularly the St Maurice-Vatcha site which has 1,612 turtle bones identified and may possibly be able to link New Caledonia Lapita peoples with the extinction of megafauna for the first time. These new sites and methods when analysed in combination with advanced theories discussed above have the potential to transform regional understanding. HBE, Optimal Foraging Models, and Costly Signalling provide powerful predictive frameworks for understanding subsistence change, as demonstrated at Teouma (Hawkins, 2015). The integration of multiple lines of evidence including zooarchaeology, stable isotopes, archaeobotany, ancient DNA, and peptide analysis, offers the most promising path forward. Each method has strengths and limitations, but combined they provide a robust and nuanced understanding of Lapita subsistence and its transformations.

Based on the above review, we outline in detail below the following future directions for best practices in Lapita zooarchaeology that will increase the pace of scientific advances;

• Renew island field surveys to identify more well-preserved Lapita sites, especially in regions with identified research gaps such as New Caledonia, Fiji, Samoa, Solomons, and Papua New Guinea.
• Develop standardized recovery (wet screening through 1–2 mm mesh) and recording methods to ensure comparability between sites and regions, including explicit reporting of mesh sizes, sampling strategies, and quantification units.
• Expand fine-grained zooarchaeological analysis at a wider range of new and old key sites, applying comprehensive quantification methods to enable robust inter-site comparisons.
• Validate cultural vertebrate patterns through systematic taphonomic analysis of bone modification and site formation.
• Improve chronological frameworks using U/Th dating on branch coral files and Bayesian modelling of radiocarbon dates to refine the timing of subsistence transitions, particularly across the 2650–2350 BP radiocarbon plateau.
• Integrate genetic studies of extinct fauna (aDNA) to resolve taxonomic uncertainties and biogeographic origins, particularly for giant tortoises and other enigmatic extinct taxa.
• Apply peptide analysis (ZooMS) to refine taxonomic identifications of fragmented bones and to address questions of biogeographic origins and species distributions.
• Adapt HBE frameworks to mixed economies combining foraging, horticulture, and animal husbandry, a challenge this review has only begun to address.
• Expand on the latest theoretical frameworks outlined above that allow for the investigation of the role of fauna in ritualized contexts by documenting bone modification and depositional patterning that signals the use of animals in the performance and negotiation of wealth, status, and political authority. A verification of such activity could be done by examining large bird wing elements for cut marks associated with feather removal (Hawkins and Worthy, 2019).
• Combine zooarchaeology with palaeoecological studies to demonstrate more nuanced models of ecological disruption due to invasive species and Lapita agricultural-mediated habitat change.
• Combine zooarchaeology with material culture studies to link subsistence change with social transformations, including changes in ceramic design, settlement patterns, and mortuary practices.
• Conduct targeted palaeontological surveys to locate pre-human fossil deposits in Vanuatu, and elsewhere, to establish baselines for pre-human biodiversity against which archaeological impacts can be measured.
• Develop actualistic studies to improve estimates of prey return rates, skeletal element utility indices for turtles and other taxa, and bone density values to strengthen interpretations of transport decisions and taphonomic processes.

6 Conclusions

Lapita subsistence appears to have been dynamic, flexible, and regionally variable, adapting rapidly to new island environments and changing socio-political contexts across the Southwest Pacific. The evidence synthesized in this review supports the mixed economy model first proposed by Green (1979) and aspects of Groube's (1971) “strandlooper” hypothesis, combining broad-spectrum marine and terrestrial foraging with horticulture and animal husbandry. However, these models require significant refinement to accommodate the temporal and spatial variability documented at sites like Teouma and across the Lapita distribution.

Human arrival in Remote Oceania caused significant and rapid faunal extinctions. The mechanisms that include direct hunting, habitat modification, and introduced predators, remain debated. The “blitzkrieg” model finds support with direct evidence for human hunting pressure on giant tortoises at Teouma and in the pattern and pace of extinctions across multiple island contexts. However, more nuanced ecological models recognizing multiple interacting factors are likely closer to reality and need to be addressed in future. The Teouma sequence demonstrates that declining forest habitats, resource depression and declining foraging efficiency were directly linked to socioeconomic transformation, including shifts in patch choice, and intensification of animal management, and eventual settlement abandonment. The Teouma data also indicate a level of socioeconomic complexity previously unreachable, demonstrated through complex pig management patterns and the potential use of feathers as wearable art to display wealth and status.

Methodological challenges continue to constrain understanding. Variable recovery methods, poor preservation, small sample sizes, and chronological imprecision limit the questions that can be addressed with most Lapita faunal assemblages. The Teouma site stands as a transformative exception, demonstrating what is possible when these challenges are systematically overcome.

The trajectory of Lapita subsistence research over the past half-century has moved from culture-historical characterization through historical ecological documentation to theoretically rigorous, hypothesis-testing frameworks grounded in evolutionary ecology. The Teouma site has played a pivotal role in this evolution, demonstrating the power of large, well-preserved assemblages analysed within robust theoretical frameworks. As new sites are discovered and existing collections analysed with increasingly sophisticated methods, our understanding of Lapita subsistence, associated with one of the most remarkable colonisation events in human history, will continue to deepen.

Generative AI statement

The author(s) declared that generative AI was not used in the creation of this manuscript.

Any alternative text (alt text) provided alongside figures in this article has been generated by Frontiers with the support of artificial intelligence and reasonable efforts have been made to ensure accuracy, including review by the authors wherever possible. If you identify any issues, please contact us.

Publisher's note

All claims expressed in this article are solely those of the authors and do not necessarily represent those of their affiliated organizations, or those of the publisher, the editors and the reviewers. Any product that may be evaluated in this article, or claim that may be made by its manufacturer, is not guaranteed or endorsed by the publisher.

Footnotes

Edited by: Loukas Koungoulos, The University of Sydney, Australia

Reviewed by: Weronika Tomczyk, Dartmouth College, United States

Simon-Pierre Gilson, Federal University of Santa Catarina, Brazil