Lapita Colonization across the Near/Remote Oceania ...

Lapita Colonization across the Near/Remote Oceania Boundary Author(s): Peter J. Sheppard Source: Current Anthropology, Vol. 52, No. 6 (December 2011), pp. 799-840 Published by: The University of Chicago Press on behalf of Wenner-Gren Foundation for Anthropological Research Stable URL: http://www.jstor.org/stable/10.1086/662201 . Accessed: 18/06/2013 00:07 Your use of the JSTOR archive indicates your acceptance of the Terms & Conditions of Use, available at . http://www.jstor.org/page/info/about/policies/terms.jsp

. JSTOR is a not-for-profit service that helps scholars, researchers, and students discover, use, and build upon a wide range of content in a trusted digital archive. We use information technology and tools to increase productivity and facilitate new forms of scholarship. For more information about JSTOR, please contact [email protected].

.

The University of Chicago Press and Wenner-Gren Foundation for Anthropological Research are collaborating with JSTOR to digitize, preserve and extend access to Current Anthropology.

http://www.jstor.org

This content downloaded from 130.216.158.78 on Tue, 18 Jun 2013 00:07:55 AM All use subject to JSTOR Terms and Conditions

Current Anthropology Volume 52, Number 6, December 2011

799

Lapita Colonization across the Near/Remote Oceania Boundary by Peter J. Sheppard The Lapita colonization of Remote Oceania involved rapid expansion from New Guinea across onetenth of the circumference of the earth. Implicit in most discussions of this phenomenon is a standard wave-of-advance model founded on demographic growth and the economic advantage provided by food production. The Lapita movement is also routinely embedded within a much larger narrative of the expansion of Austronesian languages and peoples out of Southeast Asia into Island Melanesia and ultimately east through East Polynesia. Although this simple narrative is very attractive, as more data become available, the details of segments of the “Austronesian” expansion require revision in order to reconcile the data from archaeology, linguistics, and biology. This paper looks closely at recent data on the Lapita portion of the “Austronesian” expansion and concludes that it is best explained as a leapfrog rather than a wave-of-advance movement out of New Guinea into Remote Oceania. This has important implications for those interested in modeling linguistic and biological variation in the region and highlights the potential importance of historical accident over process in our understanding of culture history.

Chief among fascinating archaeological phenomena are sudden spreads of very distinctive material culture styles over long distances and broad areas. In the western Pacific, the spread of the Lapita “culture” (fig. 1) more than 4,000 km from the Bismarck Archipelago (Papua New Guinea) southeast into West Polynesia is such a phenomenon. This involved impressive feats of seamanship, the sudden appearance and spread in an aceramic milieu of very elaborate pottery, the long-distance transport of obsidian (up to 3,200 km), the proliferation of languages of the Austronesian family, and the colonization of previously uninhabited area. Explanatory debate has focused on the role of distinctive peoples, languages, cultures, and genes (Blust 2008; Kirch 1997; Spriggs 1997). Although the proximate motives or drivers may be unknowable (Kirch 1997:64), underlying many discussions of Lapita expansion are an implicit wave-of-advance (WOA) model (e.g., Fort 2003) and an economic driver founded on the association of food production and “Neolithic” lifeways with the appearance of Lapita (Bellwood 1978; Diamond and Bellwood 2003; Green 1979b). In Europe, evaluation of a similar Neolithic spread has matured from simple application of a WOA model (Ammerman and Cavalli-Sforza 1973) through debate over the role of “indigenous” populations to considPeter J. Sheppard is Professor in the Department of Anthropology at the University of Auckland (PB 92019, Auckland, New Zealand [[email protected]]). This paper was submitted 7 X 09 and accepted 11 XI 10.

eration of more complex models of movement and cultural transmission (Zvelebil and Lillie 2000) as well as the role of sea and river transport (Davison et al. 2006). Increasingly, models of leapfrog colonization are invoked to explain archaeological and biological evidence (Zeder 2008; Zilhaõ 2001). Development of a similar critical appraisal of the Lapita spread has been hindered by spatial and quality gaps in the archaeological record; however, in the past decade considerable new archaeological, linguistic, and biological data have been acquired. Review of these data leads to the conclusion that a leapfrog pattern of colonization, rather than the gradualist WOA model, best fits the available evidence and helps explain many of the extraordinary characteristics of Lapita. This has important implications for the historical models often used by biologists and linguists when examining the broader Austronesian spread to which Lapita is, in part, related (e.g., Friedlaender et al. 2008; Gray, Drummond, and Greenhill 2009; Hill et al. 2007; Kayser 2010; Oppenheimer 2004), and it contributes to general consideration of the role of historical accident and serendipity in our understanding of the details of colonization processes. The Lapita spread is especially interesting because it must rank among the most challenging of Neolithic colonization movements.

Pre-Lapita and the Near Oceania Setting The Lapita cultural complex or tradition (Green 1992, 2003: 109) appeared sometime shortly after 3500 (cal) BP within the

䉷 2011 by The Wenner-Gren Foundation for Anthropological Research. All rights reserved. 0011-3204/2011/5206-0002$10.00. DOI: 10.1086/662201


800


Figure 1. Western Pacific, showing the known Lapita distribution superimposed on a schematic representation of a wave of advance centered on New Britain in the Bismarck Archipelago.

Bismarck Archipelago (fig. 1), a world of mostly intervisible islands located in the center of what Irwin (1992) has called a voyaging nursery of small islands and sheltered sailing. This area extends west along the coast of New Guinea, through Island Southeast Asia to the Malay Peninsula, and north from the Philippines to Taiwan. A southern extension of this nursery reaches southeast from the vicinity of New Ireland down to Bougainville, through the Solomon Islands 1,000 km to Santa Ana off the east coast of Makira (San Cristobal). Nowhere in this region is a passage of much more than 100 km required, and generally two-way visibility between islands obtains. East of Santa Ana lies the open ocean, with a minimum passage of 350 km to the Reef/Santa Cruz Group in the Southeast Solomons. At this point, sheltered sailing and intervisibility are left behind and open-sea navigation is required. Green has labeled the region beyond this boundary “Remote Oceania” (Green 1991b), an area not settled until the Lapita period, whereas Near Oceania, to the west, has been settled, at least in part, since the Late Pleistocene. Of course, the Solomons do not form the only southern extension of this voyaging nursery, as coastal sailing along the east and south coasts of New Guinea or east from the Birds Head at the western tip of New Guinea leads to the islands of the Torres Straits, where again short hops allow passage to Australia, something accomplished in the Late Pleistocene (O’Connell and Allen 2007), when Australia and

New Guinea were joined to form the continent of Sahul during glacial low-water periods. Although New Guinea and Australia were settled before 40,000 BP, the earliest dates in the Island Melanesia sector of Near Oceania indicate occupation no later than 39,500 (uncal) BP (Leavesley and Chappell 2004) when people had managed to cross the Bismarck Archipelago (fig. 2) to New Ireland. Presumably, from this time people were capable of crossing from New Ireland to Buka, which formed, during Late Pleistocene low-water periods, the northwestern end of Greater Bougainville (or Bukida), which joined the islands of Buka, Bougainville, Choiseul, Santa Isabel, and the Florida Group, enabling people to walk 750 km to the edge of a very narrow passage (!1 km) across to Guadalcanal. Evidence that people did this is, however, slight. The oldest dates from the Solomon Islands are from Kilu cave on Buka (Wickler 2001), at the northern end of Greater Bougainville. This site dates back to circa 32,000 BP (Wickler 2001:68).1 There are no other Pleistocene-age dates from the Solomon Islands, and the nextoldest date is mid-Holocene, at the site of Vatuluma Posovi on Guadalcanal, where Roe (1993) has a date of circa 6000 1. 28,340 Ⳳ 280 BP; ANU-5990; Nerita marine shell; calibrated (OxCal 3.1) here using MARINE09 with a DR of 0, producing a 1j range of 32,650–31,550 BP.


Sheppard Lapita Colonization of Remote Oceania

801

Figure 2. Northern Island Melanesia, showing the Near/Remote Oceania divide and the proposed settlement models associated with the Lapita expansion. WOA p wave of advance.

BP2 on remnant deposits at the base of a rockshelter. Surface finds of waisted axes on Guadalcanal, similar to those of Pleistocene age from the Huon Peninsula in Papua New Guinea, have been discounted by Specht (2005) as dating, in the Solomons, from the comparatively recent past. The distribution of surface finds of obsidian stemmed tools, dated in excavated contexts by tephrachronology, and distinctive mortars argued to be of mid-Holocene age extends through the Bismarcks to Bougainville, suggesting interaction throughout this area, but as yet they do not appear to be recorded farther to the southeast in the Solomons (Torrence and Swadling 2008). Perhaps the strongest evidence for early occupation of the Solomons south of Buka is provided by linguistics. Although most of the modern languages of the Solomons and the Bismarcks are Austronesian and are generally assumed to correlate with the Lapita phenomenon, a scattering of nonAustronesian (NAN) languages are found from New Britain as far south as the tiny island of Savo off the north coast of Guadalcanal. Although these have been referred to as East Papuan, in fact they differ so much among themselves that linguists have found it very difficult to relate them to one another (Ross 2001). Recently, Dunn et al. (2005) have applied phylogenetic techniques to the study of these languages and have argued that they form a group that broke up as much 2. Sample ANU-6733, Canarium nut charcoal, gives a calibrated (OxCal 3.1) 1j range, using SHCAL04, of 6400–5910 BP from the bottom layer (3(III)). There are no artifacts associated with this layer (Roe 1993: 64). The next-oldest date (ANU-6734), also on Canarium nut charcoal, gives a calibrated 1j range of 4520–4090 BP, using SHCAL04.

as 10,000 years ago. Although there has been critique of their methods (Donohue and Musgrave 2007; Dunn et al. 2007), it remains true that there is considerable diversity among these languages, which can be easily explained only by divergence over a considerable period of time, suggesting occupation of all of Near Oceania by NAN speakers by the Early Holocene. There is no archaeological evidence of pre-Lapita occupation of Remote Oceania (Green 1988; Sand 1995; Spriggs 1997). There are also no NAN languages beyond Savo in the central Solomons. What have been described as mixed languages in the Reef/Santa Cruz Group in Remote Oceania, just across the Near/Remote Oceania boundary, were argued to have NAN admixture; however, recent close analysis of these ¨ iwoo on the Reef Islands) concludes they are languages (A Austronesian (Dunn et al. 2007; Ross and Næss 2007). Similarly, the suggestion that the high degree of variability in the Austronesian languages of Southern Melanesia (Vanuatu, New Caledonia) is a function of interaction with NAN speakers (Blust 2005, 2008) is debated (Donohue and Denham 2008; Pawley 2006). The Pleistocene inhabitants of Near Oceania were foragers living a mobile, low-density lifestyle, but there is evidence, beginning in the Pleistocene, of food-resource management and the early importance of species that came to form the subsistence base with the advent of food production. Although the marine resources of the region are considerable and were exploited, it was not until the mid-Holocene that the modern lagoons and reefs were established and stabilized, as sea levels returned to current levels (Specht 2005:251–252).


802


Many coastlines would have been inhospitable rocky shores until that time (Terrell 2002). The tropical rainforests were relatively resource poor for hunter-gatherers, and that might have led to attempts to manage and enrich the environment by transporting food plants and animals from mainland New Guinea into the Bismarck Archipelago. This included the introduction of the Pacific almond (Canarium), possibly by 23,000 BP3 at Kilu on Buka but dated confidently as a tree crop on Manus by 14,000 BP (Specht 2005:267; Yen 1996); the cuscus (Phalanger orientalis), from 23,500 BP into New Ireland; the bandicoot (Echymipera kalubu), after 13,000 BP into Manus; and the northern pademelon (Thylogale browni), about 8400 BP into New Ireland (all dates calibrated following Specht 2005 and White 2004). Although appearance of these and other animals (e.g., Rattus spp.; White, Clark, and Bedford 2000) in the archaeological record have been widely interpreted as evidence of translocation of species, Specht (2005: 256) cautions against such interpretations in the absence of good data on the indigenous fauna of the region. The patchy distribution of some of these species across the islands suggests that even if they were introduced, their presence was not important enough to ensure widespread distribution (Specht 2005:259). Evidence from starch grains and phytoliths on stone tools indicates the importance of plant foods in the diet, in particular tubers, which ultimately formed the core domesticates in the region. There is evidence for taro use as early as 28,700–20,100 BP at Kilu Cave on Buka (Colocasia esculenta, Alocasia macrorrhiza), and although there is no evidence for domesticated taro (something hard to determine from starch grains) in the Bismarck Archipelago, there is increasing evidence for cultivation of taro and other crops in Highland New Guinea by the mid-Holocene (Denham and Haberle 2008; Fullagar et al. 2006). If accepted, the introduction of these plants and animals attests to considerable potential mobility by the Late Pleistocene/Early Holocene, something that is securely demonstrated by the transport of obsidian from sources in New Britain and the Admiralty Islands. Obsidian was commonly used for stone tool production within the Bismarck Archipelago, and by 20,000 BP obsidian from the Talasea and Mopir regions of New Britain is found in some quantity in New Ireland, a straight-line distance of 320 km from the sources and a minimum 15-km water crossing via the Duke of York Islands (Summerhayes and Allen 1993). In the Admiralty Group, the small source islands of Pam Lin and Lou provided obsidian to sites on neighboring Manus (Fredericksen 1997), requiring a minimum sea passage of 25 km. However, no obsidian moved out of this group until the Lapita period (Summerhayes 2004), suggesting that the long sea passage (1350 km) to its neighbors may have limited movement or that there was no need for such long-distance interactions among these hunter-gatherer populations. Although the 3. All dates for translocated species are as calibrated and reported in Specht (2005).

movements of obsidian across the sea to New Ireland are dramatic, a more common situation may be reflected by the data from the open site of Kapona na Dari (Torrence et al. 2004), which dates to 40,000–20,000 years ago (luminescence dates). Located at the base of the Willaumez Peninsula, the site contains obsidian from all the geochemical source groups of New Britain, indicating numerous two-day trips north and south of the site by people well versed in the distribution of useful stone resources. Our understanding of obsidian movement and patterns of shared culture in the Early and midHolocene is limited by the poor quality of the archaeological record; however, the find-spot distribution of distinctive stemmed obsidian tools and styles of mortars and pestles dating to this period (Torrence and Swadling 2008) indicates broad zones of shared culture within and extending beyond the Bismarck Archipelago. Although this reflects some degree of social interaction, it may simply illustrate the slow diffusion of technologies and styles over many thousands of years, creating regional traditions. In summary, the Lapita phenomenon appeared suddenly in the central Bismarck Archipelago among populations that had adapted to that environment for more than 25,000 years. These people were foragers whom we assume to have lived at low densities, possibly focused on resource concentrations along the coast but also venturing far into the interior of the large islands, such as New Britain (Pavlides and Gosden 1994; Summerhayes 2007a:11–12). They were dependent on wild foods, but these included most of the plants, both tree and root crops, that as domesticates have been the basis of subsistence into the recent past. As foragers, these people were of necessity mobile, and their crossing of numerous water gaps indicates considerable marine-transport capability, developed over thousands of years. The distribution of obsidian from sources to distant sites demonstrates relatively routine movements across passages of up to 15–25 km as well as a widespread knowledge of stone resources. Torrence and Swadling (2008) argue that the distributions of stone tools and distinctive mortars and pestles reflect social networks that were precursors of those developed or enhanced in the Lapita period. It could also be argued that the precursors or preadaptations of the Lapita maritime sailing capability and subsistence base were also in existence. All that was needed was a trigger to, as Torrence and Swadling (2008:600) put it, “light up” the system in a flash of cultural expansion that greatly magnified all precursors.

Extraordinary Lapita Like many similar horizon spreads, Lapita is marked by a series of extraordinary features, in both material culture and behavior. The Lapita cultural complex (Green 2003:109) is marked by the sudden appearance in the Bismarck Archipelago of elaborately formed and decorated pottery in a previously aceramic region (Kirch 2001b:85). Although an apparently related ceramic technology and material culture is



803

Figure 3. Earliest dates of Lapita appearance in Near and Remote Oceania. The DR marine calibration correction for Santa Cruz is from Petchey et al. (2008).

found some 2,000 km to the west in the islands of Wallacea, the very distinctive decorative ceramic style that is the hallmark of Lapita originated in the Bismarcks (Kirch 1997:52). Lapita is also marked and was conceivably catalyzed, possibly within a few generations, by rapid expansion into Remote Oceania (i.e., Oceania east of the main Solomon Islands; Green 2003:106), breaking the limits of previous colonization of the Oceanic world. This involved feats of seamanship unprecedented in world history, including the maintenance of ongoing spatial interaction over distances of more than 2,000 km, as evidenced by transport of obsidian. Figure 3 plots the ages of the earliest Lapita occupations

across the western Pacific.4 These begin with sites in the west4. The earliest dates in Mussau are from marine shells, and as yet there is no agreement on an appropriate DR value to be used. Here I use a value of 0 for sample ANU-5088. The position of the Intertropical Convergence Zone (ITCZ) below the equator in the western Pacific during the southern summer indicates that many Lapita 14C dates from terrestrial samples could be calibrated with the INTCAL09 curve for the Northern Hemisphere, as opposed to the Southern Hemisphere calibration SHCAL04 used here. Because there is no separate marine curve for the Northern and Southern hemispheres, this does not apply to the Bismarck Archipelago dates used here, which are all from shell. The current annual movement of the ITCZ boundary, whereby it moves south as far as the eastern end of the main Solomons before moving north


804


ernmost Mussau Islands of the Bismarcks (Kirch 2001a:205), followed by slightly later ages for other island groups of that archipelago. In Remote Oceania, the earliest radiocarbon ages have come from the Reef/Santa Cruz Group; that from the SE-SZ-8 Nanggu site on Santa Cruz has been argued to be the oldest (Green, Jones, and Sheppard 2008; Jones et al. 2007), although new DR values5 for correction of marine-shell dates in the Reef/Santa Cruz Group (Petchey et al. 2008) have made them, in general, age penecontemporary with those from northern and central Vanuatu (Bedford, Spriggs, and Regenvanu 2006; Galipaud and Swete Kelly 2007a). The earliest dates from New Caledonia begin with one from the original Lapita site itself (Sand 2000), while for the Fijian group a basal date derives from the site of Bourewa on the southwestern coast of Viti Levu (table 2 in Nunn 2007). While these landfall dates from Vanuatu and New Caledonia to Fiji are statistically contemporaneous, that for Tongatapu, in the Tongan group, suggests that its settlement may have been slightly later (Burley and Dickinson 2001, 2010), a viewpoint supported by later dates for sites in the Vava’u Islands to the north (Burley and Connaughton 2007). This summation of radiocarbon dating allows no more than 250 years for the formation and dispersal of the Lapita tradition. In Remote Oceania this spread was extremely rapid, so much so that even with some slope in the rate of spread from the Reef/ Santa Cruz Group to Tonga, it permits no pauses to fully populate any of the large island groups. Within almost a radiocarbon instant, a unified ceramic tradition (fig. 4) spread over 4,000 km—one-tenth of the circumference of the earth—from New Guinea to Samoa. This represents the transport of a production technology, decorative techniques—which include not just fine dentate stamping but incision, applique´, and punctation—and an elaborate decorative system composed of sets of distinctive motifs applied to a wide variety of pot forms. As almost all of this material is locally made or, in the absence of clay, derives from communities on nearby islands (Dickinson 2006), the ceramics involve reproduction in the new lands of this complex cultural system and not simply transport of heirlooms. Why this elaborate ceramic tradition, requiring consideragain, plus the unknown longer-term history of its movement, related in part to the El Ninõ–Southern Oscillation phenomenon, makes it hard to know which curve to apply to terrestrial samples in Near Oceania. Ultimately, a combined model might be required that might also take into account the northern and southern convergence zones. I use SHCAL04 to calibrate all terrestrial samples unless otherwise stated. The offset of the Northern and Southern hemisphere curves is 56 Ⳳ 22 years (McCormac et al. 2004). 5. There is, in fact, no DR value for the high island of Santa Cruz. Jones et al. (2007) generated a value of ⫺81 Ⳳ 64 by using matched pairs of archaeological Tridacna and charcoal from the RF-2 site on the nearby low coraline Reef Islands, while most recently, Petchey et al. (2008) have provided a DR of 26 Ⳳ 11 for the Reef Islands by using museum samples of non-Tridacna species. The use of this new value here significantly reduces the age of the SZ-8 samples; however, until a DR is established for Santa Cruz, the age of SZ-8 is still imprecisely known.

able skilled labor and time investment in pot construction and decoration, was maintained and reproduced by colonizers of Remote Oceania remains a puzzle. Whatever the reason, it would appear that it materialized something fundamental to Lapita society that needed to be reproduced by small colonizing populations. Some insight into the function of the design system may be given by the emphasis on anthropomorphic motifs, including face motifs of a variety of forms (Chiu 2007; Spriggs 1990). Combined with the suggestion (Green 1979a; Kirch 1997) that the dentate stamping and incising represented tattooing of the pots, this indicates that they may have played a role in an ancestor cult. The activity of tattooing/decorating, along with more recent evidence of overpainting and covering of dentate designs (Bedford 2006), suggests that the pots were actively involved in some form of ongoing system, wherein the making of the pots and designs may have been more significant than some form of meaning directly signaled by distinctive designs. Most recently, the relationship with ancestors has been strengthened by the association of large decorated pots with burials, wherein skulls were manipulated, at the early Lapita Teouma cemetery site in Vanuatu (Bedford, Spriggs, and Regenvanu 2006). This elaborate tradition did not remain cohesive. Within a few hundred years, the fine dentate stamping and elaborate pot forms characteristic of the early sites were replaced in a process of rapid simplification of form and decorative technique (Clark and Murray 2006; Green 1979b). Ultimately, ceramic manufacture disappeared from much of the archaeological record of Near and Remote Oceania (Samoa, Tonga, Wallis/Futuna, central Vanuatu, the Southeast Solomons, the Western Solomons, and much of the Bismarck Archipelago; Spriggs 1997). It seems probable that the rate of simplification, at least in some areas (Sand 2007:217), was faster in Remote Oceania than in the Bismarck Archipelago (Summerhayes 2007b:152), so that the dentate stamp/elaborate pot form system ended in some West Polynesian sites in a few generations. This suggests a close association between this system and a decline in its utility following initial events of colonization (Kirch 1988a; Sheppard 1993). A third extraordinary aspect of Lapita is the long-distance movement of obsidian. Obsidian from local sources was used and transported around the Bismarcks before Lapita; however, the amounts and distances rose dramatically with Lapita (Specht 2002; Summerhayes 2004). In the Lapita period, Bismarck obsidian spread in small amounts west to Sabah, Malaysia, and east to Fiji and New Caledonia, covering a total range of 6,500 km. Although Lapita obsidian is often described as indicative of some long-distance trade/exchange network, there is little basis for arguing anything other than direct-access procurement and heirloom effect. The amounts of obsidian transported and the variety of sources are best seen as an index of movement and interaction. Consequently, no model of simple decline with distance applies. Sites in the Reef/Santa Cruz Group more than 2,000 km



805

Figure 4. Lapita ceramics from the SE-RF-2 site in the Reef/Santa Cruz Group, Te Motu Province, Solomon Islands.

from the source have comparatively large amounts of Bismarcks obsidian (SZ-8, 287 pieces; RF-2, 644; RF-6, 29); while sites slightly farther out, including those in Vanuatu, have little to none (!60 pieces per site in northern and central Vanuatu [Galipaud and Swete Kelly 2007a], !10 [total] in New Caledonia and Fiji, and none in Tonga). Specht (2002) has estimated that the combined amounts for 12 recorded Lapita sites in the Reef/Santa Cruz Group could be 845–1125 kg. These sites also contain a striking variety of obsidian sources. Of the 972 samples from the Reef/Santa Cruz sites sourced by Green (1987), most (97.5%) are from Talasea on the Willaumez Peninsula of New Britain, while small amounts are from the Admiralties (1.13%), in the northwest of the Bismarcks, and also from Vanua Lava (1.23%), in the Banks Islands of Vanuatu. One

piece is from West Fergusson, off the coast of New Guinea in the D’Entrecasteaux Group (Green and Bird 1989; Sheppard, Trichereau, and Milicich 2010). These are thus the most diverse Lapita sites, with four sources representing a web of connections back into the Bismarcks, south into Remote Oceania, and across the Coral Sea to New Guinea. The presence of small amounts of obsidian from the nearby Banks Islands indicates that the obsidian was not simply linked to connections back to a “homeland” in the west. Although the Banks Islands volcanic glass is of comparatively poor quality, it would be suitable for the simple cutting tasks to which obsidian was put in Lapita sites, yet it remains a minor component in Lapita obsidian assemblages of Remote Oceania, even those in Vanuatu. There is no evidence of obsidian reuse in the Reef/Santa Cruz sites (Shep-


806


pard 1993); thus, the amount of Bismarcks obsidian implies strong ongoing connections back to the homeland and, given the dominance of New Britain material, into the Willaumez Peninsula region. Models of Lapita Expansion Models of migration proposed by Anthony (1990) and Moore (2001) are routinely cited in the Lapita literature (e.g., Green 2003; Kirch 1997); however, additional characteristics have to be taken into account to provide for variants used by biologists and linguists interested in Pacific colonization. Although there are a bewildering variety of named models or scenarios (see review by Green 2003), most are based on permutations of a small number of parameters. These include (1) demographic viability, (2) directionality, (3) speed of migration, and (4) the nature of the natural and cultural environment at the destination. The simplest model is the wave of advance (WOA), as originally proposed by Ammerman and Cavalli-Sforza (1973). This model is founded on colonization as a function of demic diffusion or demographic growth and short-distance, slow movement expanding out into suitable social and physical environments without any particular directional bias. Underlying this model is an economic driver whereby demographic growth is fueled by either improved access to resources or new food-producing technologies (e.g., Paleoindian occupation of the Americas or Neolithic expansion into Europe). Expected outcomes of this model are (1) continuous cultural distributions, (2) temporal clines, (3) nondirectional expansion into equivalent environments, and (4) considerable biological and cultural interaction with previous inhabitants, if present. In the Pacific variants of this model (e.g., Fort 2003; Kayser 2010; Oppenheimer 2004), much attention has been paid to speed of movement, giving rise to “express-train” and “slow-boat” versions. Speed is assumed to correlate with degree of interaction with previous inhabitants, which would have important implications for the genetic record. Archaeologically, this is also entangled with questions of the rise of the Lapita archaeological tradition and whether Lapita was introduced into the Bismarck Archipelago or developed in situ (Allen 1984). Green has proposed an intermediate scenario that describes Lapita as developing from external intrusion, local innovation, and ultimate integration (Green 1991a). Ultimately, in Oceania the Lapita expansion is seen as part of a WOA fueled by innovation in maritime technology and development of food production based on domestic animals and tubers. Variation in rates of spread or “pulse-pause” patterns are seen as related to the limitations of new innovations driving expansion or pulse events (e.g., Gray, Drummond, and Greenhill 2009). Most recently, a variant of the demic model, the ideal free distribution, which attempts to operationalize the model by using the principles of human behavioral ecology, has been proposed for the study of colonization in Oceania (Kennett, Anderson, and Winterhalder 2006). The major difference be-

tween this model and earlier formulations is an attempt to take into account mode of subsistence, new-habitat suitability, and population density in predicting migratory behavior. Therefore, not all habitats are of equal value, making it possible to predict leapfrog-type behavior to reach high-quality habitats. This model is still founded on demographic growth and predicts some population increase until habitat-quality differentials develop triggering dispersal. The Kennett, Anderson, and Winterhalder (2006) model for Oceania predicts episodic periods of dispersal/migration followed by population growth, with the key variable mode of subsistence determining population growth and effect on habitat quality. Faster movement is predicted by foraging-type subsistence behavior in the relatively depauperate natural environments of Remote Oceania. Speed of advance is, of course, related to transport capability. In the Island Pacific, migration involves watercraft and island hopping, although settlement of large islands under the WOA model would take some time (e.g., New Caledonia is 400 km long, with an area of 19,000 km2; Bougainville is 200 km long, with an area of 9318 km2). If, however, all islands along a front are colonized in sequence, then there is ultimately little difference from continental models, as movement requires the filling of available equivalent landscapes before expansion. Leapfrog models differ from WOA models in that they involve bypassing potential settlement areas to establish outposts in uninhabited land or enclaves in previously inhabited areas. Repeated moves of this kind might result in a string-of-pearls pattern, joining settlements in favorable environments, which Moore (2001) suggests is both a very successful strategy from a demographic perspective and possibly predicted by the ideal-free-distribution model. Such models, however, are dependent on significant founding populations or regular contact with parent or neighboring populations to buffer demographic variation. The implications of the leapfrog model depend on the size of the leaps and whether there is backfilling from colonization events. If the leaps are small, perhaps to the next high-quality habitat, then as Bar-Yosef (2002) argues, the difference between WOA and leapfrog, or saltation, models may simply be one of scale. If, however, the distance is large and crosses inhabited areas, then we would expect to see significant cultural and biological disjuncture, with implications for the genetic, linguistic, and archaeological record. If the leapfrog movement is large and into uninhabited territory, we would see a founder effect, creating cultural and biological homogeneity in the new area. If the leapfrog strategy was repeated, as under the string-of-pearls scenario, we would see cultural homogeneity over wide areas appearing in a very short period of time. Lapita Expansion Figure 1 shows the known distribution of Lapita sites and a representation of a wave-of-advance model. Contrary to what we might expect, the Lapita distribution is highly directional



(Irwin 2008). If settlement was driven by the need for new land by new food-producing populations, then given settlement success in the Bismarcks, we would expect to see the coast and coastal islands of Papua New Guinea occupied long before any move to the south and east. The only dentate ceramic material reported from mainland New Guinea consists of two sherds from the north coast (Summerhayes and Allen 2007:98). Nothing has been recovered from the New Guinea coast across the Vitiaz Strait or, with the exception of Lapita sherds at one late Lapita site on Tuam in the Siassi group (Lilley 2004; Spriggs 1997:118; Summerhayes 2007b), from the set of offshore islands that extends north along that coast. Similarly, on Manus only a handful of Lapita sherds have been recovered from three possibly late Lapita sites (Ambrose 1991; Spriggs 1997:112), leading Summerhayes (2007b: 169) to question the extent of Lapita occupation of that area. On the south coast of Papua New Guinea, no Lapita sites are reported, although there is a later Early Papuan Pottery (EPP) horizon that, as a colonization event, has many similarities with Lapita and is believed to correlate with the spread of Austronesian speakers considered to be Lapita descendants. The only other direct Lapita connection is a piece of Fergusson Island obsidian recovered from the RF-2 Site in the Reef/ Santa Cruz Group (Green and Bird 1989). In a review of the south-coast sequence, Summerhayes and Allen (2007) did not completely rule out the possibility of yet-to-be-discovered Lapita sites in that area, but they concluded that the amount of research already conducted in the Massim made it unlikely. The recent finding of “red slip” pottery in the Torres Straits, argued to date to 2400–2600 BP, may suggest earlier intrusion into this area (McNiven et al. 2006), as does the discovery of plain ceramics on Wari Island in the Massim region, which may date to the same time period (Negishi and Ono 2009). Similarly, the finding of an anomalous temper in late Lapita ceramics in Roviana Lagoon in the Western Solomons, possibly derived from granitic contexts in New Guinea or Australia, argues for contact across the Coral Sea (Felgate 2003; Felgate and Dickinson 1998). Although Clark and Bedford (2008:70) imply that these facts indicate an as-yet-unknown Lapita occupation of the east New Guinea coast, which would have the known EPP tradition at its end, Summerhayes and Allen (2007; see also Irwin and Holdaway 1996) have demonstrated that the latter has all the hallmarks of a colonization event, like that of early Lapita but substantially later. At the moment, the data suggest that any Lapita movement along the New Guinea coast may at best be contemporaneous with the late Lapita movement into the Western Solomon Islands. The clear direction of early Lapita expansion is to the southeast, along the axis of the Solomon Islands, with a later radiation into southern Island Melanesia and western Polynesia. A wave-of-advance model would predict considerable Lapita settlement of the many islands and lagoons of the main Solomons, where there is no shortage of the locations favored in the Bismarcks. There are, however, no early Lapita settlements known from the region extending from Bougainville

807

to the end of the main Solomons (Sheppard and Walter 2006). Although the possibility of a gap in settlement has been raised (Roe 1993), it has generally been balanced by the suggestion that the gap represents a knowledge gap (Kirch 1997:53; Spriggs 1997:128) and that sites would ultimately be found (Green 1979b). In fact, although members of the Southeast Solomons Culture History project (Green and Cresswell 1976) failed to find Lapita sites in the areas of Near Oceania that they surveyed in the eastern Solomon Islands (Ulawa, Makira, Santa Ana, and Ugi), it was their discovery of materials from the main Solomons, especially chert in the early Lapita sites of the Reef/Santa Cruz Group in Remote Oceania (Sheppard 1993, 1996), that strengthened the argument for an as-yetunknown main-Solomons settlement and led Green (1979b, his fig. 2.12) to propose such locations in Marau Sound in eastern Guadalcanal and in the New Georgia Group of the Western Solomons. Although such settlement made sense under a wave-of-advance model, early Lapita sites have not been found, despite what can be argued to be amounts of research at least comparable to that expended elsewhere before the discovery of Lapita deposits. This has led Sheppard and Walter (2006) to propose a revised model of Solomon Island prehistory that incorporates an early Lapita leapfrog across the main Solomons into Remote Oceania followed by a significant late Lapita appearance in the Western Solomons. Critique of this proposal has reiterated the need for more research and emphasized the need to consider the geomorphic context of the Lapita record and its preservation (Clark and Bedford 2008; Felgate 2003, 2007). Before turning to a review of recent evidence from biology, linguistics, and archaeology in an evaluation of the Solomons leapfrog proposal, it is important to note that the accepted archaeological evidence based on chronology, amounts of obsidian in the Reef/Santa Cruz Group, and the cohesiveness of the early Lapita ceramic corpus is much more supportive of leapfrogs than of wave-of-advance models (see also Kirch 1997:64). The evidence from Remote Oceania indicates that Lapita colonizers were accustomed to making long moves away from neighbors, and although the 2,000 km from the Bismarcks to the Reef/Santa Cruz Group is a very long leap, sailing is guided along the Solomons as far as Makira or Ulawa. In addition, although we have evidence of Lapita contact with the Solomons and potentially with the southern Papuan coast, the known archaeological sequence from the main Solomons is in many respects similar to the known sequence from the southern Papuan coast, with a late Lapita or Austronesian incursion creating a distinctive regional tradition well after the initial Lapita pulse out of the Bismarck Archipelago. To summarize, the currently available archaeological evidence (fig. 2) indicates (1) a leapfrog colonization of Remote Oceania out of the Bismarck Archipelago by 3200 BP, (2) late Lapita settlement of the Western Solomons circa 2600 BP, and (3) late or post-Lapita settlement of the southern Papuan coast circa 2400–2800(?) BP. Although the proposed archaeological picture may simply represent gaps in data, it


808


also has implications for linguistic and biological data as well as for models of archaeological visibility. These implications will be developed and used to evaluate that picture in the next section.

Evaluation of the Models Linguistic Evidence Linguistic data indicate that speakers of Proto-Austronesian originated in East or Southeast Asia, perhaps some 5,000– 6,000 years ago (Pawley and Ross 1993). Most recently, Gray, Drummond, and Greenhill (2009), using a phylogenetic approach, have replicated the standard historical linguistic model and placed the origin of Austronesian in Taiwan circa 5230 BP. It is argued that the spread of Austronesian, by either migration of speakers or language shift, was fueled by development of agriculture and Neolithic technology (Bellwood 1991; Diamond and Bellwood 2003), although this association has been challenged, most recently by Donohue and Denham (2010). A prehistoric association between language, material culture, and biology is very hard to determine. In Near and Remote Oceania, the association between language and material culture, as recovered by archaeology, is supported by the fact that in Remote Oceania Lapita is the founding culture and that only Austronesian languages are spoken, providing a near-perfect association (Green 1991b), at least in that region. A possible exception has been the languages of the Reef/ Santa Cruz Group, to which I return below. The linguistic expectation for a wave-of-advance model of Austronesian spread out of the Bismarck Archipelago into Remote Oceania would, following the principles outlined by Ross (1997), be the formation of an innovation-linked subgroup of languages extending from the homeland in the Bismarcks southeast into Remote Oceania. Early breakage of the linkages might occur where geographical barriers to ongoing communication existed (Ross 1997:226). In Near and Remote Oceania, significant sea gaps (lacking intervisibility) should correlate with early linkage breaking. Therefore, early breaks should occur in the Buka–New Ireland gap, the Southeast Solomons–Reef/Santa Cruz/Vanikoro/Utupua (Near-Remote Oceania) gap, the Reef/Santa Cruz/Utupua–Vanuatu gap, the Vanuatu–New Caledonia gap, and the Vanuatu-Fiji gap. Over time, linkage breakage would develop at less significant barriers to interaction. The resulting distribution of languages should show a fairly clear pattern of relationship from west to east, with major distinctions at these gaps. Subsequent history of migrations and language replacement or interaction with NAN populations might complicate this picture; however, it should be rare for all traces of earlier language substrates to be obliterated. A leapfrog model would differ from a wave of advance in that it would create much more disjuncture in linguistic distribution, potentially create founder effects, and minimize the

degree of interaction with preexisting populations. There should be no necessary close relationships between neighboring languages, and if leapfrogs occurred from a linguistically variable homeland, unexpected variety might be generated in the destination region. Given the Solomons leapfrog model, we would expect (1) no simple west-to-east pattern of language relationships, (2) high-order relationships between the Bismarck Archipelago and Remote Oceania, and (3) potentially complex relationships within Remote Oceania, all occurring as a result of distinct founder effects. Figure 5 represents some general relationships and characteristics of the languages of Island Melanesia. One characteristic shared by all of these languages is remarkable linguistic diversity, with many languages per island or island group. This contrasts remarkably with the situation in Polynesia, where there is a one-to-one relationship between island groups and languages (Pawley 1981). Whatever the drivers behind the development of this pattern, they would appear to be similar throughout the region. Apart from this general similarity, there is considerable variation across the area, with major disjuncture in language geography. The Austronesian languages of this region are all members of the Oceanic subgrouping of Austronesian. Ross (1988) argues for two high-order divisions of Oceanic: Western Oceanic, derived from a dialect linkage, and an Admiralties group. Lynch, Ross, and Crowley (2002:94) add a third division, called Central/Eastern Oceanic, that includes almost all other areas of Oceania in a poorly defined grouping. Western Oceanic includes Meso-Melanesian, which incorporates the Austronesian languages of northern New Britain east of and including the Willaumez Peninsula, the Bali-Vitu group, New Ireland, and the northern and western Solomons down to Santa Isabel (Lynch, Ross, and Crowley 2002:101; Pawley and Ross 1993: 439). The origin of this subgroup is considered to be in the Willaumez Peninsula area. The Austronesian languages of the St. Matthias Group (Mussau) appear to form a separate highorder subgrouping that has affinities to the Admiralties but is not part of that group (Lynch, Ross, and Crowley 2002:96). Western Oceanic and the Meso-Melanesian subgroup end at the eastern end of Santa Isabel in the central Solomons, where it abuts Southeast Solomonic, which includes all the remaining languages in the main Solomons to the east. This boundary is marked by the Tryon-Hackman line, which denotes a very sharp high-order linguistic break (Ross 1988; Sheppard and Walter 2006): all languages to the east of this point belong to the Central/Eastern Oceanic grouping of Lynch, Ross, and Crowley (2002). Although the geographical and archaeological boundary of Near and Remote Oceania is found at the end of the main Solomons, the equivalent linguistic divide apparently occurs, following Lynch, Ross, and Crowley (2002), at the Tryon-Hackman line. The Southeast Solomonic languages are very conservative. This distinguishes them from language groups to either side and in particular from other languages of Central/ Eastern Oceanic in Island Melanesia, which Pawley (2006) describes as highly “aberrant.” Lynch, Ross, and Crowley (2002: 112) divide the latter into a large Southern Linkage, which



809

Figure 5. General characteristics of Austronesian languages of Island Melanesia.

includes the languages of Vanuatu and New Caledonia, and the very diverse languages (six) of Utupua and Vanikoro, which they propose constitute one or two first-order divisions of Central/Eastern Oceanic. Linguists generally propose a wave-of-advance or equivalent model to explain the distribution of Austronesian languages southeast from the Bismarck Archipelago and explain deviations from expectation as a function of subsequent history. If our application of the spread zone is correct, then the CEOc (Central/Eastern Oceanic) languages represent the original eastward spread of Oceanic languages, first forming a dialect chain and then breaking up into a linkage. The WOc (Western Oceanic) linkage, on the other hand, represents a later spread that, at least in its eastern portion (New Britain, New Ireland, Bougainville, and the western Solomons) overlaid earlier CEOc languages. The eastern extent of this overlay accounts for the rather sharp boundary between WOc and CEOc in the central Solomons (Ross 1988:382–385, 1989). The fact that the CEOc languages apparently form a linkage, rather than a neatly branching tree, suggests that Oceanic speakers moved across the central and eastern Pacific rather rapidly (Pawley 1981):

there were no pauses of sufficient length to allow much linguistic innovation to take place before movement continued (Lynch, Ross, and Crowley 2002:97–98). Therefore, CEOc is proposed to have arisen in the Bismarcks and to have spread with, or at the time of, Lapita through the Solomon Islands and on through Remote Oceania. The problem raised by the Tryon-Hackman line is explained by portraying WOc and the Meso-Melanesian linkage as a later (post-Lapita?) spread through the northern and western Solomons that completely replaced all earlier Austronesian languages. The linguistic evidence for an earlier CEOc spread in this region is apparently very slight (Ross 1988:383–384), leading even Ross to question the scenario. The next aberration is the conservative Southeast Solomonic subgroup. Pawley (2006) has suggested that the relative lack of diversity in this region is the result of limited interaction with NAN populations in the central Solomons, as opposed to high levels of interaction in the western Solomons. Although there are more NAN languages surviving in the area of the Meso-Melanesian linkage, their distribution continues into the central Solomons. Savo lies 12 km off the northwest


810


coast of Guadalcanal and the Russell Group 40 km to the west. Both are occupied solely by NAN speakers and have histories of interaction with Austronesian speakers (Terrill 2003) of Southeast Solomonic. It does not seem reasonable to think that the area occupied by Southeast Solomonic was originally unoccupied by NAN speakers. In fact, Guadalcanal is the only area with archaeological remains dating to the mid-Holocene, demonstrating the presence of pre–Lapita period people, if not what language they spoke. If Southeast Solomonic represents back-movement from farther east in Remote Oceania, as has been tentatively proposed by Sheppard and Walter (2006), it is apparently not easily related to the languages of that region. They are highly variable, and it has been proposed that their variability could be caused, in part, by interaction with NAN speakers (Blust 2005); however, there is no archaeological evidence of preLapita settlement of the region (Green 1988) associated with NAN speakers. More recently, Blust (2008) has proposed that an early Austronesian colonization of Vanuatu was quickly followed by a second wave of NAN speakers who gradually adopted Austronesian from the early colonizers. This secondmigration hypothesis has some antiquity in archaeology (Spriggs 1997) as a mechanism to ultimately explain the biological and cultural differences across the Melanesia/Polynesia boundary. The archaeological basis for this has steadily weakened, leading Bedford and Spriggs (2008:112–113) to recently conclude that “these multiple traditions do not provide evidence of waves of secondary migration or high levels of sustained interaction” (see also Reepmeyer and Clark 2009). The only other area where NAN occupation has been proposed is in the Reef/Santa Cruz Group, where three languages have been interpreted as mixtures of Austronesian and NAN (Mu¨hlhaüsler et al. 1996). Recently, Ross and Næss (2007) have closely examined the languages of the Reef/Santa Cruz Group and concluded that there is no evidence for NAN ¨ iwoo and the other languages of that influence and that A region are part of an Utupua-Vanikoro group, which together they describe as the Temotu group, representing a first-order subgroup of Oceanic. Ross and Næss’s conclusion is also supported by recent phylogenetic analysis of all Solomons NAN languages (Dunn et al. 2007). Temotu cannot be related to its neighbors to the east or to the west in the Solomons and has its closest ties with the languages of the St. Matthias Group (Mussau, Tench), in the Bismarcks north of New Ireland. This is indicative of a significant leapfrog and leads Ross and Næss (2007:461) to state that “if the Temotu languages form a primary subgroup, then this suggests that the arrival of their ancestors was separate from the arrivals of either the Southeast or Northwest Solomonic Groups in the areas they currently occupy.” The most recent contribution to the study of relationships among Austronesian languages is the work of Gray, Drummond, and Greenhill (2009), who have used lexical data and phylogenetic analysis to construct a phylogeny of 400 Austronesian languages. In the region of interest, their results

support, for the most part, those of the comparative linguistic method as discussed above (supplementary online material 5 in Gray, Drummond, and Greenhill 2009). The Temotu subgroup is strongly supported and joins the phylogenetic tree at the same level as the groupings of Southeast Solomonic/ Meso-Melanesian, North and South Vanuatu/New Caledonia, and Polynesian/East Polynesian/Micronesian, implying a firstorder division created before the divergence of these groups (fig. 1 in Gray, Drummond, and Greenhill 2009). However, the St. Matthias (Mussau) subgroup is not supported, and Mussau and Vitu (65 km northwest of the Willaumez Peninsula) form distinct outliers in the large Meso-Melanesian group. Both of these conservative languages are from small, isolated island groups that have possibly experienced comparatively limited interaction with neighbors since the Austronesian expansion in the region. In summary, there is very little linguistic evidence to support a wave-of-advance model of Austronesian expansion through the Solomons into the Pacific. Instead we see considerable variation and radical disjuncture in the linguistic geography. Of course, this could be explained by a complex linguistic history, following the proposed Lapita-associated spread of CEOc (Lynch, Ross, and Crowley 2002). These arguments invoke the complete replacement of early CEOc in the northern and Western Solomons, with only limited traces remaining (Ross 1988), and an absence of Austronesian-NAN interaction in the central and eastern Solomons, where there are extant NAN languages and evidence of pre–Lapita period populations. Finally, in Remote Oceania, the explanation of “aberrant,” highly variable languages in Southern Melanesia requires reference to Austronesian interactions with preexisting or significant secondary migrations of NAN populations (Blust 2005, 2008; Donohue and Denham 2008) in a region where there is no evidence for pre-Lapita settlement or archaeological support for secondary migration. The work of Ross and Næss (2007), recently supported by Gray, Drummond, and Greenhill (2009), points directly to at least one leapfrog out into Remote Oceania, and Pawley suggests that the variability in Remote Oceania was a function of repeated direct movements from the Bismarcks: “As movements of Lapita peoples from Northwest Melanesia to Southeast Melanesia continued over the next few generations it may be that canoes carrying mixed[-language] populations came to predominate” (Pawley 2006:248). Leapfrog movement from a diverse homeland region, such as the Bismarck Archipelago, allows for the generation of diversity in the populations settling the new islands, something not easily accommodated in a wave-of-advance model. Genetic Evidence Study of the genetic structure of Pacific populations has exploded in recent years, with considerable literature on the evidence produced by mitochondrial DNA (mtDNA), nonrecombining Y-chromosome markers, and, most recently,



study of autosomal microsatellite variation (Friedlaender et al. 2008). Geneticists have tended to test their patterns against simple models (Kayser 2010) of Pacific history that seek to map the “origins” of Polynesians and consider the amount of genetic admixture contributed by indigenous populations to Austronesian colonists. Often, this can lead to the definition of “Polynesian” genetic lineages in areas thousands of kilometers to the west of Polynesia (Mona et al. 2009). In Island Melanesia, this is ultimately tied to the debate over the origins of Lapita, as either an Austronesian intrusion or a local Bismarck innovation, and the importance of reticulate versus treelike models of culture history (Terrell 1988). For biologists, low amounts of genetic admixture correlate with limited local interaction and an “express-train” model, while a “slow-boat” model postulates colonists interacting with indigenous populations, producing high amounts of admixture as they slowly pass from points of origin to Polynesia (Diamond 1988; Diamond and Bellwood 2003; Hurles et al. 2003; Oppenheimer 2004). Recently, Hunley et al. (2007) have moved to explicit biological models that drop the culture history associated with the express-train and slow-boat models. They formulate “population-fission” and “isolation-by-distance” models. The former is a model “in which populations split as they expand to occupy new territories” (Hunley et al. 2007:141), then remain relatively isolated, and as a result create treelike or evolutionary patterns of relationship. The isolation-by-distance model has people exchanging genes and language with their neighbors as a function of intermarriage, which will produce correlations with distance (i.e., neighbors should be more alike) and reticulate relationships. These models are essentially models of interaction designed to look specifically at patterns of genetic variation in Northern Melanesia. The development of such focused models and studies is an improvement over one-size-fits-all approaches that incorporate Asian origins, Austronesian expansion, Lapita development, and the origins of Polynesians in one model when these may require a variety of explanations or models for different stages (Green 2003: 103) or aspects of this complex process. A biological wave-of-advance model for Lapita expansion out of the Bismarcks into Remote Oceania would involve ongoing interaction along the expansion front with both the immediate population source and other populations, if present. This would predict considerable overall homogeneity throughout the region, with perhaps a cline of genetic similarity as foreign lineages (assuming intrusion of a Lapita “biology”) were swamped by indigenous lineages when the wave of advance moved through previously occupied areas in the Solomons. The resulting biology, at the time of colonization, should be very similar across the Lapita zone of Remote Oceania and like that of the eastern Solomons. A leapfrog model predicts close relationships between the biology of the Bismarcks and that of Remote Oceania, with little necessary similarity to that of the main Solomons. The leapfrog model also suggests repeated population movements from the Bismarcks to establish viable populations in the new region, providing

811

the potential for repeated founder effects sampling any diversity found in the Bismarcks. The modern populations of the Bismarcks exhibit considerable genetic diversity. Studies of mtDNA indicate that although some of this is driven by topography and interiorversus-coastal divisions, which correlate to some degree with NAN-versus-Austronesian language distinctions, most variation is geographic, with neighboring coastal populations showing little language-based distinction (Friedlaender et al. 2007b). Therefore, islands and sectors of islands tend to have distinctive genetic patterns (Friedlaender et al. 2007b; Kayser et al. 2008a). Most of the mtDNA haplogroups currently identified in the region are ancient. Two, B4a1a1 and E, are considered by Friedlaender to date to the Holocene and are potentially associated with the development of the Lapita phenomenon. The relatively common B4a1a1 haplogroup, which includes the “Polynesian motif,” is very common in East Polynesia and is considered a marker of Polynesian heritage. In the Bismarcks, it has a variable distribution (Kayser 2010) and is often found in low frequencies, especially in eastern New Britain (fig. 8 in Friedlaender et al. 2007b).6 Therefore, the development of Lapita throughout the Bismarcks does not correlate, on the basis of modern genetics, in any simple fashion with biology or an intrusive homogenizing population. As a source region for further expansion it is highly variable, potentially allowing identification of source populations under the leapfrog model. There are few genetic data sets or analyses that cover the Near-Remote Oceania transition in detail, and the Solomons and Vanuatu–New Caledonia are poorly studied. Friedlaender et al. (2007a, 2007b, 2008) have recently provided a mtDNA data set that covers our area of interest. Figure 6 shows a cluster analysis of data presented in table 4.2 in Friedlaender et al. (2007a), excluding the very rare F, Y, and M27 lineages, which are virtually absent (n p 3 ) from the data studied, which included all Bismarcks, Solomons, Vanuatu, New Caledonia, and Fiji data. In addition, an extra sample for Manus was added from Kayser et al. (2008a) to complement the very small sample in the original data set. The frequencies of the 23 lineages reported were converted to percentages and clustered (average linkage). The resulting plot presents a number of higher-order clusters, which show significant geographical patterning. The first grouping includes all samples from Mussau, New Ireland, western New Britain, Bougainville, the main Solomon Islands (Malaita) and Fiji. These are essentially populations with high frequencies of the “Polynesian motif,” B4a1a1. The second group includes all eastern New Britain populations and Remote Island Melanesia (Santa Cruz, Vanuatu, and New Caledonia), with the exception of Fiji. These are populations with low (!36%) to no Polynesian motif, high frequencies of the ancient Melanesian lineages Q1 and Q2, and they are distinguished by high frequencies of M28a1 6. Defined here as east of the Willaumez Peninsula and not referring solely to the province.




and M28a2. Within this group, Mamusi, Ata, Mali, and Kaket populations can be distinguished by an almost complete absence of the Polynesian motif and the presence of a variety of older “Melanesian” lineages. The final cluster is distinguished by very low to no “Polynesian” motif and very high frequencies of Q1. This would appear to reflect older populations either isolated in the interior of New Britain and Bougainville or within the remote Admiralties Group. These data show a clear, close relationship between Remote Oceania and the Bismarcks, in particular eastern New Britain, and a lack of the relationship with the main Solomons and Bougainville that would be expected from a wave-of-advance model. Surprisingly, the Reef/Santa Cruz group clusters most closely with the coastal language group located next to the Talasea Peninsula and the major Lapita obsidian sources. The exception to this pattern is Fiji, which has a very high (60%) percentage of the “Polynesian motif” (Friedlaender et al. 2007a), making it more like its Polynesian neighbors to the east and perhaps reflecting Polynesian interaction into Fiji. Recently, additional genetic data (Kayser 2010) have become available from the nonrecombining portion of the Y chromosome (NRY) for the Bismarck region, providing a picture of male lineages in the region (Scheinfeldt et al. 2006). These, like the mtDNA data, have shown considerable diversity within the region and an unexpectedly very low Asian signature. The geographic pattern of variation is, however, similar to that seen in the mtDNA, with a distinctive difference between New Britain and New Ireland–Bougainville (fig. 2 in Scheinfeldt et al. 2006). Unfortunately, comparable data are not available for the Solomons, Santa Cruz, Vanuatu, and New Caledonia, although the limited Vanuatu data are comparable to those from the Bismarcks. Most recently, Friedlaender et al. (2008) have reported the analysis of 687 microsatellite and 203 insertion/deletion polymorphisms in 952 individuals from 41 Pacific populations; unfortunately, as with the NRY data, crucial areas south of Bougainville in Island Melanesia are not represented. However, within the Bismarcks and Bougainville, the earlier patterns from mtDNA and NRY are maintained and refined, with new lower-level discrimination of the north, south, east, and west coasts of New Britain and of New Ireland from Bougainville (fig. 8 in Friedlaender et al. 2008). Application of these methods to the rest of Island Melanesia should be able to clearly identify source populations in the Bismarck Archipelago. If, as the leapfrog model suggests, Remote Oceania was settled by repeated direct colonization events from the Bismarcks, fine-grained genetic data from Remote Oceania should be able to test the details of this hypothesis. Although many more genetic data are re-

813

quired from the Solomon Islands, Vanuatu, and New Caledonia, authors of the most recent work on mtDNA from a small western Solomon Islands sample (Ricaut et al. 2010) support the leapfrog proposal. Archaeological Evidence It is rather difficult to predict what an archaeological record of early Lapita occupation of the main Solomon Islands would look like. Although there may be some Lapita preference for small islands, that may simply reflect relative ease of archaeological survey (Bedford and Spriggs 2008:112; Specht 2007). The Solomons have an abundance of rich lagoons and large islands fringed by barrier reef/island complexes, which are easily comparable to areas settled in the Bismarcks and Remote Oceania. There would appear, then, to be no environmental basis for avoiding the region, yet there is no evidence for early Lapita settlement south of Buka. For virtually all of the central and eastern Solomons, there is no record at all of any ceramics from any period. Only on the small island of Santa Ana, on the edge of Remote Oceania, has a handful of very poor quality plain ceramics been recovered, after extensive excavations by William Davenport (1972),7 followed up by work by Roger Green and Pamela Swadling (Swadling 2000). My new fieldwork there in 2009 failed to find additional ceramic evidence. In the western Solomons and Bougainville, there is a relatively abundant ceramic record extending up to the historic period but not extending back beyond the late Lapita period. The standard explanation for this apparent gap is limited archaeological reconnaissance; however, as Sheppard and Walter (2006) have argued, there has been considerable Solomons research over the past 40 years that is certainly comparable to the amount of research conducted before the first finding of early Lapita sites in the rest of the Lapita range. Perhaps most telling is the work of the Southeast Solomons Culture History project, which found numerous Lapita sites in the Reef/Santa Cruz Group and Tikopia in Remote Oceania but, employing the same approach, found none in Ulawa, Ughi, Santa Ana, and Makira in the Eastern Solomons. Similarly, 7. The results of the work by Davenport and later by Green on Santa Ana have not been well reported; however, Davenport screened very large volumes of deposit in the rockshelters and caves he excavated on this very small island. The Feru excavation is still open, and I estimate, on the basis of observations in 2009, that at least 52 m3 of deposit were excavated along the face of that rockshelter. Although such places may not preserve good Lapita-period records, there is generally, as on Santa Cruz, some evidence of Lapita occupation in rockshelters where that occurs.

Figure 6. Average-linkage dendrogram of mtDNA haplogroup variation in Island Melanesia (data adapted from Friedlaender et al. 2007a and Kayser et al. 2008). Underlined languages are non-Austronesian. NI p New Ireland; Boug p Bougainville; Sol p Solomons; WNB p western New Britain; SENB p southeastern New Britain; NENB p northeastern New Britain; ENB p eastern New Britain.


814


the excavation of large volumes of cultural deposit, including deposits of Lapita age, from Vatuluma Posovi (Poha) Cave at the back of a coastal flat on Guadalcanal revealed no ceramics from any period. This, plus extensive fieldwork on western Guadalcanal, led Roe (1993) to conclude that the island had been bypassed by Lapita settlement. Beginning in 1996, I initiated a program of research in the Western Solomons, one goal of which was to search for the presumed missing Lapita link between the Bismarcks and the Southeast Solomons. I conducted, along with my colleague Richard Walter (University of Otago) and numerous PhD and MA students, an initial archaeological survey over a period of years (1996–2001) in the Roviana Lagoon region of New Georgia, followed by another program of research 70 km to the west on the island of Vella Lavella (2001–2007). During this time, we recorded 24 late Lapita or Lapita-derived intertidal sites, and research by others added four sites to the total of 28 presented in figure 7. This research indicates that such sites are ubiquitous in the Western Solomons, extending eastward a tradition of Lapita intertidal “stilt-village” occupation commonly found in the Bismarck Archipelago (Gosden and Webb 1994; Kirch 1987). Although the ceramic tradition (incised and applique´ designs) represented in these sites is clearly late Lapita

or Lapita derived, only a very small sample of dentate stamped sherds have been recovered, with the Honiavasa site in Roviana Lagoon having the most (six) and only a few sites having some of the complex carinated pot forms (Honiavasa, Zoraka/Nusa Roviana, and Poitete; Felgate 2001, 2003 [figs. 7, 9]; Summerhayes and Scales 2005) of classic Lapita. Direct dating of carbon inclusions on or in sherds provides late Lapita ages for some of these sites (charcoal inclusion in a sherd from Paniavile [AA33504; 2130 Ⳳ 90 BP; calibrated with SHCAL04, 2j range 2350–1800 BP]; carbon encrustation on a sherd from Hoghoi [NZA-1253; 2619 Ⳳ 45 BP; calibrated with SHCAL04, 2j range 2770–2490 BP]). Given the abundance of late Lapita evidence in the Western Solomons recovered over the past 20 years by a variety of researchers and research programs, the question remains of how well the known record reflects the actual occupation history. Perhaps earlier sites are yet to be found or the early record has been destroyed. Proving the absence of evidence is rather difficult; however, assessing specific hypotheses of systematic destruction of the record is feasible. Felgate (2003, 2007) has proposed that geotectonic processes have significantly altered the record of intertidal sites in the Western Solomons. Using the work of Mann et al. (1998), Felgate has

Figure 7. Distribution of late Lapita sites in the Western Solomons and general patterns of geotectonic movement. The width of arrows is proportional to rates of uplift proposed by Mann et al. (1998).



argued that significant rates of uplift in areas such as Roviana Lagoon have moved earlier intertidal sites up into the swash zone and ultimately into dry areas, where the salt-impregnated sherds would rapidly deteriorate, based on our observations of the fate of untreated sherds removed from intertidal sites. Felgate’s hypothesis depends on the accuracy of the Mann et al. (1998) proposed model of regional uplift and the careful assessment of the geotectonic setting of the individual intertidal sites. The strong earthquake and associated tsunami of Easter 2007 has renewed interest in the geotectonics of the Western Solomons, and in December 2007 I returned to Roviana to assess the setting of the intertidal sites and the potential effect of such events (Sheppard and Walter 2009). Figure 7 summarizes the Mann et al. (1998) model of geotectonic movement in the Western Solomons and the distribution of reported intertidal sites. This figure illustrates that there is considerable variability in the rates of geotectonic movement in the region and that late Lapita or Lapita-derived sites are found throughout the area, including in places where no uplift has been measured (e.g., north Kolombangara and north Vella Lavella). Therefore, we cannot propose a uniform geotectonic process to explain the known distribution of sites throughout the Western Solomons. Roviana Lagoon, however, is one region where such an explanation is feasible, because Mann et al. (1998) propose an annual rate of uplift of 0.6– 0.9 mm for the lagoon. Detailed consideration of geotectonics and the archaeological record in Roviana are presented in Sheppard and Walter (2009), where each site context is reviewed. That study involved recording height profiles across the intertidal sites and the adjacent shoreline and also recording raised beach terraces and old wave notches. The results are summarized as follows. 1. There is evidence of a raised terrace, which can be found at circa 2–3 m above mean low tide around the lagoon and is consistent with estimates of the mid-Holocene high stand elsewhere in the western Pacific (Dickinson 2001; Mitrovica and Peltier 1991). 2. Intertidal ceramic sites are found at the mouths of rivers and adjacent to springs on the current shore of the lagoon mainland, indicating that the wide, low, swampy shore through which streams meander dates at least to 2000 BP and is probably the result of progradation that started when the lagoon was established after the mid-Holocene high stand. 3. Dated intertidal sites (e.g., Hoghoi) on the inner shore of the barrier islands should be raised up to 2 m, given a constant uplift rate of 0.7 mm per year. This would require an original depth of more than 3 m, without considering potentially higher water levels associated with the run down from the mid-Holocene high stand. Construction of stilt houses in an original depth of water of 3⫹ m seems improbable. The Mann et al. (1998) model for uplift in Roviana assumed constant uplift in the absence of other data. Whether geotectonic events in the region are a function of slow and constant or episodic interactions between the Pacific and Aus-

815

tralian plates has been debated (Taylor et al. 2005). Study of the archaeological site distribution on the mainland coast of Roviana indicates limited uplift over at least the past 2,000 years. Most recently, the earthquake of 2007 has resulted in dramatic, instantaneous uplift in parts of the New Georgia Group; on Rannonga this was more than 2 m, while in Roviana it has resulted in sudden subsidence of up to 65 cm at the western end of the lagoon and of 39 cm in the central section of Roviana (Taylor et al. 2008). This indicates that Roviana is subject to episodic subsidence after major tectonic activity, potentially followed by slow rates of uplift. The archaeological evidence suggests that the sum effect for at least the past 2,000 years has been minimal. It is possible that a sudden uplift before circa 2600 BP destroyed an earlier Lapita record, but at present there is no evidence of such an event. What is clear is the fact that the Roviana geotectonic record is complex through both time and space and cannot be modeled with a simple constant rate of uplift. As a consequence, no uniform model can be employed to explain the distribution of intertidal sites in the region and the lack of an early Lapita record. Summary of the Linguistic, Biological, and Archaeological Data A simple wave-of-advance model predicts smooth clines of spatial variability with neighbors looking most like one another, while leapfrog models predict disjuncture and founder effects that potentially create considerable spatial heterogeneity. The linguistic evidence demonstrates that the languages spoken by the descendants of the Lapita settlers who first crossed into Remote Oceania and settled in the region of the Reef/Santa Cruz Group and neighboring islands (Reef/Santa Cruz–Utupua-Vanikoro) are most closely related to languages in the eastern Bismarck Archipelago, possibly to languages in the vicinity of Mussau. They are not closely related to those of their nearest neighbors to the west in the main Solomon Islands. The genetic data create a highly variable spatial picture throughout Island Melanesia, and even within the Bismarck Archipelago there is no simple “Lapita” genetic signature or association between genetics and the distribution of Lapita pottery. Examination of mitochondrial DNA shows samples from Santa Cruz and Vanuatu to be most similar to samples from New Britain east of the Willaumez Peninsula, not to samples from elsewhere in the Bismarcks or the Northern Solomons. This eastern New Britain–Santa Cruz–Vanuatu grouping is characterized by very low levels of the “Polynesian motif” and high frequencies of ancient Melanesian lineages. The Santa Cruz sample is most like the sample from the Nakanai language group, which is currently located just to the east of the Talasea Peninsula, where most of the obsidian in the Reef/Santa Cruz sites was obtained. Although much more genetic study is required southeast of the Bismarcks, available data show considerable disjuncture in genetic spatial distributions, supportive of a leapfrog model. Recent genetic


816


research in the Western Solomons has concluded by supporting that model (Ricaut et al. 2010). On the basis of archaeological research to date, there is a clear gap in the distribution of Lapita sites, with only late Lapita ceramics found in the Western Solomons and no Lapita in the main Solomons east of New Georgia. But is this gap real, or does it simply reflect inadequate research and/or systematic destruction and poor archaeological visibility? It is perhaps significant to note that in every location in Island Melanesia where there is a Lapita record, there is also a visible post-Lapita ceramic record, although in some instances ceramics production eventually ceases. Even in the Western Solomons, limited amounts of plainware are present in the late prehistoric record, and these have been regularly reported by all archaeologists working in the region. In contrast, the only ceramics ever reported from the central and eastern Solomon Islands are the handful of early plainware sherds from Santa Ana; beyond that there is no ceramic tradition, suggesting the lack of a Lapita foundation. Whether the record is obscured or destroyed is hard to determine. Variation in geomorphological processes and rates of sedimentation and weathering exist across the Lapita range. Deep volcanic ash on northern Vanuatu has both buried and preserved Lapita deposits, creating in most areas a landscape devoid of Lapita-age surface deposits, while tectonic uplift has altered the location of shorelines (Bedford 2007). A thin layer of ash may have helped preserve numerous Lapita deposits in the Reef/Santa Cruz Group (Green, Jones, and Sheppard 2008); however, they are being gardened onto the surface today. In the main Solomons south of Bougainville, there are no volcanic ash deposits, and Lapita sites should often be close to the surface. In the Western Solomons, tectonic movement may have destroyed part of the record of the ubiquitous intertidal sites, but as we have discussed, this process is not uniform and cannot be invoked to explain the destruction of an early Lapita record throughout the studied region. Elsewhere to the north, on Buka and in the Bismarck Archipelago, early Lapita intertidal sites are relatively common (Kirch 1997), and given the large number of intertidal sites reported in the Western Solomons, it seems highly unlikely they would have been missed. Taken together, the present linguistic, biological, and archaeological data are very consistent with one another and present a disjunct spatial pattern indicative of a leapfrog model. This strongly supports the leapfrog model of settlement of Remote Oceania directly from the Bismarck Archipelago by people making Lapita pottery and the avoidance of the main Solomon Islands until the late Lapita period.

Discussion: Why a Leapfrog? Push and Pull Factors The rapid movement of Lapita out into Remote Oceania suggests the presence of a trigger event or a sudden change of conditions. Two such push factors have been proposed: major volcanic activity on New Britain and the development of fa-

vorable eastward sailing conditions during a period of ENSO (El Ninõ–Southern Oscillation) activity. The massive WK-2 (Machida et al. 1996) plinian eruption of the Witori volcano covered much of West New Britain east of the Willaumez Peninsula with thick layers of volcanic ash, with the 20-cm isopach covering the breadth of the island for a distance of more than 100 km and creating much of the modern landscape. Earlier dating of this event placed it in the 3830–3350-BP range (SUA-2814; 3370 Ⳳ 100; Machida et al. 1996), which locates it outside most of the Lapita time frame. New dating and Bayesian analysis of the event place it well within the Lapita period (3480–3150 BP, modal age 3315 BP; table 5 of Petrie and Torrence 2008) and potentially just before the Lapita expansion into Remote Oceania. Petrie and Torrence (2008:740) have drawn attention to the association of this event with the appearance of Lapita and have suggested that the widespread extinction or disruption of cultural groups in West New Britain may have facilitated the introduction of new immigrants or the local adoption of pottery. Although no Lapita pottery has been found under the WK-2 ash, there is no doubt that local non-pottery-producing residents were displaced, altering regional population dynamics and creating an unpopulated zone that was eventually reoccupied, after a few hundred years, by people making Lapita pottery (Torrence, Neall, and Boyd 2009a). It should, however, also be noted that obsidian from the Talasea region, within the affected area, is found at all the earliest Lapita sites, indicating continuity of contact with the Willaumez region throughout the Lapita period. Whatever the exact effect, it is reasonable to propose that the eruption was a dramatic event visible throughout the Bismarcks, resulting in major population disruption that had knock-on effects throughout the New Britain/New Ireland region. Irwin (1992, 2008) has demonstrated the feasibility and utility of purposeful sailing upwind against the prevailing southeasterly trade winds in the settlement of Near and Remote Oceania, taking advantage of periodic westerlies for outward downwind legs. Although travel within the Bismarck and Solomon archipelagos would have provided comparatively easy coastal voyaging, in a region where historic movement of war canoes often involved paddle voyages of more than 200 km (Bathgate 1985), movement out into Remote Oceania would have been greatly facilitated by westerly winds. Recently, Anderson et al. (2006) have noted that the ENSO record applicable to the western Pacific indicates a sudden increase in El Ninõ events starting at 3200 BP and continuing until 2200 BP. Such events reduced the strength of the southeasterly trade winds and facilitated downwind sailing to the east. Although, as Irwin (2008) argues, the Anderson et al. (2006) model minimizes the role of sailing innovation, the theories are mutually reinforcing, and together they suggest that the period starting at 3200 BP was an optimum time for colonizing Remote Oceania, when both innovation in sailing capability and favorable weather conditions potentially converged for the first time.



Once the Lapita colonizers of Remote Oceania moved across the 350-km water gap from the main Solomon Islands, they would have encountered something neither they nor their ancestors had seen before: islands without people and without any of the commensal or symbiotic plants, animals, insects, or microbes that had shaped their lives for generations. Although movement out into the Pacific was movement into a terrestrially depauperate environment, where no land mammals other than bats could be found, this was more than made up for by the untouched marine, bird, and reptilian fauna, including now extinct megapodes, crocodiles, and tortoise. In the first generations, the return from exploitation of these resources would have been outstanding, and it is perhaps significant that none of the earlier translocations of indigenous New Guinea animals into the Bismarcks, with the exception of the large spiny rat (Rattus praetor), which was transported to the Reef/Santa Cruz Group, Tikopia, Vanuatu, and Fiji (White, Clark, and Bedford 2000), occurred into Remote Oceania. Ultimately, chickens, pigs, dogs, and the Polynesian rat made it out into Remote Oceania, but is notable that the distribution of these animals is patchy and not at all strongly associated with colonization (Anderson 2003; Matisoo-Smith 2007). It has been argued that initial Lapita colonizers were “strandloopers” (Anderson 2003; Davidson and Leach 2001; Groube 1971) living off of the abundant resources of untouched islands. It is clear that Lapita peoples had knowledge of and used domesticates (Green 1979b), including plants (Horrocks and Bedford 2005; Horrocks, Bedford, and Spriggs 2009; Horrocks and Nunn 2007), which they transported into Remote Oceania. It is also clear that early Lapita settlements show very heavy exploitation of rich marine and bird resources, such as the very abundant turtle remains found in early Tongan and Tikopian sites or the large megapodes of New Caledonia, which became extinct shortly after Lapita settlement (Burley 1999; Kirch 1997:199–226). Perhaps the most striking evidence of the early importance of mixed terrestrial and marine animal protein in the diet in Remote Oceania is the isotopic data on teeth and bone from the early Lapita Teouma cemetery in Vanuatu (Bentley et al. 2007; Valentin et al. 2010). This shows that the majority of the population appears to have spent their childhood locally and had a significant marine signature, while four individuals interpreted to have been immigrants had a significantly reduced marine signature. At least one of these individuals is interpreted as a man of some importance, and the inference is that these immigrants were from the original colonizing movement out of Near Oceania. Isotopic evidence from late Lapita–age burials in New Britain indicates a diet dominated by terrestrial protein and plants (Beavan Athfield et al. 2008). The authors of the most recent isotopic study at Teouma conclude by favoring “the hypothesis of a mixed economy including maritime exploitation, terrestrial foraging and lowlevel food production” (Valentin et al. 2010:1827). Although the abundant untouched resources of the new islands would have been immediately apparent, it would have

817

been only a short time before another powerful advantage became apparent: there was no malaria (Buckley 2005; Kirch 1997:110). The current mortality rate among children and pregnant women is very high in Near Oceania, and at any one time a very significant percentage of the population suffers malaria attacks. The modern incidence of malaria in the Solomon Islands is among the highest in the world (152 per 1000 in 2005), and the associated mortality (7.54 per 100,000 in 2005) is exceeded, within the Pacific, only by Papua New Guinea. Although the incidence of malaria in Vanuatu was relatively high in 2005 (44 per 1000), there was zero mortality (table 7 in Breman and Holloway 2007). Today, the Buxton line marks the limits of Anopheles mosquitoes and malaria, which extend only as far south as Vanuatu, with the north being more malarious than the south. The rest of Remote Oceania is malaria free. When people first crossed into Remote Oceania in the Reef/Santa Cruz Group, there would have been no malaria, given the absence of the human host. It also appears likely that there was no vector for transmission, because in Remote Oceania there is only one species (Anopheles farauti s.s.) of the common mosquito vector (the Anopheles punctulatus group), while in the Solomons there are 12 (Beebe et al. 2000; Lum et al. 2007). Although there is evidence that A. farauti can feed on nonhuman hosts, Lum et al. (2007) argue that this pattern of species diversity, which correlates with the timing and length of human colonization of the Pacific, and the fact that A. farauti is salt tolerant, indicate that the Vanuatu mosquito came ultimately at some unknown point after human colonization as a hitchhiker on boats (Lum et al. 2007). When this occurred is as yet unknown, although Buckley (2005) attributes comparatively high levels of anemia, as evidenced by skeletal remains from Taumako just east of the Reef/Santa Cruz Group and dating after 470 BP (Leach and Davidson 2008), to malaria.

Conclusion The prehistory of the settlement of Near and Remote Oceania involves debate over a series of puzzles. When was Near Oceania settled? What is the “origin” of the Lapita cultural complex? Why did that elaborate cultural complex suddenly move out into Remote Oceania and almost as quickly disappear? How do we account for the linguistic, cultural, and biological diversity of Remote Oceania, especially the Melanesia/Polynesia division, when it appears to have a common Lapita ancestor? The wave-of-advance model makes most of this problematic, because it promotes the idea of homogeneity and, at most, clines of variation. Explaining the observed diversity and disjuncture in linguistic, biological, and archaeological patterns under such a scenario calls for subsequent diversity in localized but powerful historical forces capable of radically altering and, in some cases, obliterating earlier patterns. Therefore, explanation of the Tryon-Hackman line in the Solomons requires a second wave of Austronesian speakers into


818


the Western Solomons; explanation of the conservative nature of the Southeast Solomonic languages requires significant interaction with NAN speakers in the Western Solomons but not in the central Solomons; explanation of linguistic diversity in the Island Melanesia sector of Remote Oceania calls for a second wave of “NAN”-speaking peoples into that region after the Lapita settlement, of such magnitude as to significantly alter cultural and linguistic patterns before being subsumed into the earlier Austronesian cultural pattern; and finally, the explanation of biological differences between populations called Melanesian and Polynesian in the Western Pacific requires genetic bottlenecks in the Lapita period, at a time of unsurpassed regional interaction, and/or subsequent migration of biologically Melanesian populations into Vanuatu, New Caledonia, and Fiji. The leapfrog model of Lapita settlement into Remote Oceania greatly simplifies many of these scenarios by allowing for rapid expansion and radical spatial disjuncture in linguistic, cultural, and biological patterns and by generating the potential for significant founder effects. Linguistic, biological, and archaeological evidence all support the leapfrog hypothesis. Why the leapfrog occurred is perhaps unknowable, but serendipity and historical accident, rather than some systematic process, must underlie this event. Specific historical factors, such as a massively destructive eruption on New Britain at the start of the Lapita period (Petrie and Torrence 2008) and the increase in El Ninõ events facilitating sailing to the east (Anderson et al. 2006) by people with suitable boats, must have played a role. Once found, pull factors must have included the lure of untouched marine and sea bird resources (Kirch 1997:199–226), which isotopic analysis shows formed a significant part of the diet in Remote Oceania. Most significantly, the absence of malaria (Kirch 1997:110), which has an extremely high incidence in the Solomons, seems very likely to have been another pull factor. The implications for archaeological, biological, and linguistic models relate predominately to the importance of founder effects in understanding Pacific prehistory. A leapfrog will create unexpected geographic breaks in data and conceivably result in repeated sampling of the homeland variation. Biological and linguistic similarities are to be expected between the Bismarcks and remote Island Melanesia and reflect variation in the Bismarcks. More biological research is required in the Solomons, Vanuatu, New Caledonia, and Fiji, and the first focus must be on the origins of these Lapita descendants. For these are the direct descendants of Lapita settlers of Remote Oceania, and there is little need to invoke some subsequent process of “Melanization.” Once Lapita had reached Remote Oceania, much of the subsequent Lapita colonization may represent direct movements, or leapfrogs, as opposed to a wave of advance. The recent report of Lapita pottery with distinctive temper from Island Melanesia in the oldest Tongan Lapita site is argued by the excavators to be the result of direct settlement from that region (Burley and Dickinson 2001, 2010). Ultimately, Polynesian origins are related to Lapita settlement but form a separate and potentially more complicated issue, as suggested by

the Polynesian/Micronesian linkages in Gray, Drummond, and Greenhill’s (2009, their fig. 1) phylogenetic study of Austronesian languages and as recently discussed by Addison and Matisoo-Smith (2010).

Acknowledgments I would like to thank the people of the Western Solomons and the National Museum of the Solomon Islands, who have assisted research for more than a decade. An early draft of this paper was improved by comments from Andrew Pawley, Roger Green, Geoff Irwin, Thegn Ladefoged, and John Allen. I also thank the reviewers, who very carefully read and improved this paper. All errors of fact or interpretation remain, of course, my own. I thank Briar Sefton and Tim Mackerell for work on the figures. This work was supported in part by grants from the Royal Society of New Zealand Marsden Fund and the University of Auckland. I dedicate this paper to my mentor and friend Roger Green, who provided this North African archaeologist with the opportunity to study Pacific prehistory when he sent me to the Solomons more than 20 years ago.

Comments Stuart Bedford Department of Archaeology and Natural History, College of Asia and the Pacific, Australian National University, Canberra, Australian Capital Territory 0200, Australia ([email protected]). 4 VI 11 From the mid-1990s Peter Sheppard and colleagues began one of the most comprehensive archaeological programs undertaken in the Solomon Islands. Focused in the Western Province, it has produced spectacular results, particularly for the Lapita period and the past 700 years, published in a series of articles and chapters in books. It involved collaboration with colleagues from a range of disciplines and has seen the completion of a number of successful PhDs. It entailed literally years of accumulated survey and excavation in the field. These researchers did not find any evidence of pre-Lapita or Early Lapita occupation but instead found many late Lapita and immediately post-Lapita ceramic sites. I do not address the linguistic or genetic arguments that are advanced by Sheppard to support his leapfrog model of Lapita colonization but rather focus on the archaeological ones, which, unlike the other two lines of evidence, cannot significantly shift location over millennia. The question of whether gaps in Lapita across its distribution are real or imagined has been around for a long time, across much of the Pacific, but over the past 15 years either the gaps have been steadily filled in or there have been sound explanations as to why they might be extremely difficult to find, that is, Vanuatu



and Samoa (see Clark and Bedford 2008 for a review). The real concern with the leapfrog argument is the comprehensiveness of our knowledge of the archaeological record in areas east of the zone studied by Sheppard and colleagues in the Western Province. Sheppard here and elsewhere claims that the Solomons, through efforts of various researchers over a 40-year period, now has good archaeological coverage across the whole archipelago (Sheppard and Walter 2006). Certainly, most of the islands have been visited at some time by archaeologists, but to suggest that the research coverage is comprehensive is a significant exaggeration. The only surveys that have been carried out on many of the large islands were initial reconnaissance surveys undertaken during the National Site Survey Program of the National Museum in the 1970s, which primarily targeted customary sites. Pottery was frequently noted in the Western Province, but no Lapita sites were then identified. Even after two field seasons of the Sheppard project focusing on the question of Lapita, it was reported that Classic Lapita dentate-stamped ceramics had not been found, and it was hypothesized that there was no Lapita occupation in the Western Solomons (Sheppard et al. 1999). Clearly, finding these sites in the Solomons is a challenge, the earliest of them even more so, and it requires the sort of longterm research projects mounted by Sheppard and colleagues. Sheppard, of course, cites the results of the earlier Southeast Solomons Culture History project. Many Lapita sites were abundant in islands located in Remote Oceania, but none were found farther east on the islands of Ulawa, Ughi, Santa Ana, or Makira. In fact, apart from Santa Ana, most of the sites found dated to not much earlier than the past 1,000 years. This record is clearly far from complete. Similarly, on Guadalcanal, no Lapita or other pottery sites have been recorded. At least there are two sites there, as opposed to most of the rest of the Solomons, that do encompass the pre-Lapita, Lapita, and post-Lapita periods, the inland cave sites excavated by Roe (1993). But does the excavation of three inland cave sites on the north coast and a surface survey of agricultural features and other sites in the west really constitute comprehensive coverage of an island of more than 5,000 km2 that has a history of human occupation at least 6,000 years old? Accumulated evidence now indicates that cave sites are generally poor indicators of Lapita presence, and humaninduced erosion brought about by increased agricultural activity is likely to bury coastal sites (Lapita) that are backed by high hillsides, as is the situation in many of the islands of the Solomons. Surface surveys are often not enough when looking for sites that date to ca. 3000 BP. None of this comment is to knock the abilities or strategy followed by any of the archaeologists, amateur or professional, who have worked in the Solomons. Rather, it highlights the challenging nature of archaeological research in the Solomons, with its many islands, both large and small, and their very complex geomorphology. What are the drivers for a migratory population who have arrived from farther west, always traveling through regions

819

that are already populated, to decide that once they have spent a short time in the Bismarcks, they should spectacularly leapfrog the entire Solomons and land safely in the Reef/Santa Cruz Islands? One of the major pulls for rapid migration east must have been the discovery of uninhabited islands, but presumably in Sheppard’s model there were Lapita scouts who found the Reef/Santa Cruz Islands then roared back to spread the word. Certainly, I agree that the wave-of-advance model in unadulterated form does not explain the Lapita phenomenon, but neither, I suggest, do 2,000-km leaps bypassing whole archipelagos.

Peter Bellwood School of Archaeology and Anthropology, Australian National University, Canberra, Australian Capital Territory 0200, Australia ([email protected]). 17 VI 11 This is an interesting paper that had to be written, given the current expansion of knowledge about Lapita issues in general. The author makes much of the apparent differences between the wave-of-advance and leapfrog models of population expansion. But I find it difficult to see any clear separation between them. Both are purely propulsive concepts; waves flow (sometimes quickly), and frogs leap. Sheppard defines a wave of advance as reflecting demographic growth and a presence of food production and as being gradualist. Leapfrogging is not really defined at all beyond its saltatory connotations, but it is linked to cultural and genetic disjunction, founder effects, minimal interaction with preexisting populations, and “serendipity” or historical accident in causation. Why these concepts cannot apply to a wave of advance as much as to a leapfrog model is not made clear. Both models are two aspects of the same historical phenomenon: human migration fed by population growth, for whatever reason, and a desire to expand into new territory. Sheppard opens with the statement that the Lapita migration started “out of New Guinea.” He then amends the starting point to the Bismarck Archipelago, politically a part of modern Papua New Guinea but certainly distinct from New Guinea island proper. As he notes, New Guinea itself had no connection with early Lapita. I agree fully with Sheppard that Lapita people avoided New Guinea and leapfrogged across the Solomons from the Bismarcks to the Reefs/Santa Cruz Group. But I do not agree that Lapita expansion was serendipitous. Had it been so, it could have happened any time within the past 50,000 years. It did not. In fact, it happened exactly when major population movements and advances in Neolithic and maritime technology were occurring all over southern China and Southeast Asia, both island and mainland. A quite different view of Lapita origins will have been published by the time this paper appears (Bellwood 2011; Bellwood et al. 2011; Hung et al. 2011). It runs as follows. About 3500 BP, a group speaking a language created by the migratory


820


breakup of Proto-Malayo-Polynesian settled the Mariana Islands from the northern Philippines. This 2,300-km opensea crossing is our first evidence of a major ability to cross wide ocean spaces in human history. From the Mariana Islands, a daughter group crossed a similar expanse of ocean to reach the Bismarcks by 3350 BP, where they introduced a red-slipped pottery tradition decorated with alternating zones of stamped circles and parallel incised lines infilled with punctate stamping. These migrants apparently did not carry domestic animals. Those who stayed behind in the Marianas came to speak a Malayo-Polynesian language outside the Oceanic subgroup of Austronesian, one whose modern descendant we know as Chamorro. The Bismarck Archipelago, before 4000 BP, contained populations of potential arboriculturalists and vegeculturalists who lacked domesticated animals. Data are scarce, but we have a better-known manifestation of indigenous (pre-Austronesian) food production in the New Guinea Highlands, so a lowland counterpart in the Bismarcks can also be considered likely. The shock of the cultural, linguistic, and biological confrontation with long-existent others in the Bismarcks led somehow to the development of Lapita as an archaeological identity, together with the linguistic creation and breakup, through continuing migration from the Bismarcks, of the Proto-Oceanic linguistic node. By close to 3300 BP, some Lapita migrants had leapfrogged from the Bismarcks over the previously inhabited Solomons to reach the first of the empty islands of Remote Oceania. At the same time, New Britain obsidian was carried to Borneo, probably reflecting a meeting with cultural close cousins in equatorial Island Southeast Asia who lacked pottery ancestral to Lapita at this time but who also spoke languages very recently derived from Proto-Malayo-Polynesian. Lapita was thus only one end of a vast and multidirectional movement between 3500 and 3000 BP that encompassed much of Island Southeast Asia, but not New Guinea itself until much later. Lapita stands out for archaeologists only because of its striking pottery decoration. Otherwise, it contained little that was unique, beyond its geographical distribution. I also part company with Sheppard when he suggests that existing ethnolinguistic boundaries in western Oceania were mainly laid down by leapfrogging during the period of Lapita dispersal. The early Oceanic daughter languages of Lapita times were too homogeneous to have created such enduring patterns. The past 2,500 years must have seen massive population shifts within Melanesia. I am also puzzled as to why a “voyaging nursery” is defined to include Melanesia and Island Southeast Asia, and in this case even Taiwan, but not the many small and very strategic islands off the coastline of southern China, from Zhejiang to Guangdong. The archaeological record in Zhejiang has produced several canoes and paddles from Early to Middle Holocene waterlogged sites, discoveries so far unparalleled in Oceania. My overall opinion of this paper is positive, but I see the Lapita colonization as part of a much larger picture than just

the islands of Melanesia. The early Malayo-Polynesian world at 3000 BP was much greater in extent.

David V. Burley Department of Archaeology, Simon Fraser University, Burnaby, British Columbia V5A 1S6, Canada ([email protected]). 20 VI 11 For the past decade-and-a-half, Peter Sheppard and colleagues have searched unsuccessfully for an early Lapita settlement landscape between Bougainville in the west and the Reef/Santa Cruz Group to the east, a distance of almost 1,500 km. In a logical sense, archaeologists assumed (some still do) an orthodox wave-of-advance model for Oceanic settlement, where progressive expansion through the central and eastern Solomons eventually extended into Remote Oceania, reaching its ultimate limits in Fiji/western Polynesia. The failure of Sheppard and others to find early Lapita sites did not dissuade logical assumption. Rather it led to further assumptions, ones rationalizing the absence of evidence. Since stilt-house villages on the reef are a documented Lapita feature of the Bismarck Archipelago, why should we expect an archaeological signature to have survived, especially given potentially volatile geomorphic contexts across the region? Sheppard and Walter (2006) proposed a radical alternative in 2006—an early Lapita leapfrog model where the area was avoided altogether. Sheppard’s testing and validation of this model in the current paper expands this interpretation and presents a broad range of supplementary data for its support. Whether all of the data presented here have veracity remains to be determined. That Sheppard is forcing archaeologists to recognize the complexity of Oceanic settlement and alternative models in lieu of wave of advance is without question. Many archaeologists, including Sheppard, appreciate and attempt to reconcile their data with the vast array of studies documenting language and genetic relationships in Oceania. His review of these data and their lack of apparent support for a wave-of-advance model are presented as central to the argument at hand. The review also hints at the all-too-often contradictory, confusing, or short-lived conclusions that many of these studies generate (e.g., cf. Kayser et al. 2008b, Richards, Oppenheimer, and Sykes 1998, and Soares et al. 2011). Sheppard’s case seemingly may be strengthened by this review, but it truly must rest on an archaeological foundation. In this respect, his case is a difficult one, for as he notes, an absence of evidence is at the heart of the argument. His counterarguments to claims of early Lapita site destruction, poor archaeological visibility, or inadequate research are well made but potentially debatable. What is more difficult to debate is the simple but astute observation that a Lapita ceramic record in Oceania inevitably gives way to an abundant post-Lapita ceramic record. The complete absence of ceramics in the cen-



tral and eastern Solomon Islands, with one minor exception, becomes substantively telling. The implications of Sheppard’s paper go well beyond his arguments for a leapfrog versus a wave-of-advance model. Volumes of Talasea obsidian in Reef/Santa Cruz Lapita sites indicate continued interaction with a homeland 2,000 km distant. Without intervening settlements from a wave-ofadvance progression, some degree of sophistication in the maritime capacity of Lapita seafaring must be anticipated. This topic not only is contentious but also is debated in vitriolic fashion (Anderson 2008). A leapfrog movement through Near Oceania raises the further probabilities that Lapita exploration was strategic and that leapfrogs might have occurred elsewhere in the Lapita settlement of Remote Oceania. Indeed, Dickinson and I (Burley and Dickinson 2010) propose this type of event for the Polynesian founder settlement in Tonga on the basis of petrographic analysis of ceramic temper sands, as Sheppard notes in his conclusions. The leapfrog, in this case, begins on a dacitic high island in centralisland Melanesia or, perhaps, even farther to the west. It involves a distance of no less than 2,000 km, and it bypasses or skips through an already settled Fiji. Founder effects, isolation, and regional integration in Tonga present an obvious recipe for the Melanesian/Polynesian divide. Sheppard’s caution that Gray, Drummond, and Greenhill’s (2009) phylogenetic study of Austronesian languages make this “a separate and potentially more complicated issue” is difficult to fathom. What Gray, Drummond, and Greenhill (2009:482) propose is the possibility of a “common underlying factor [read homeland] between the subsequent pulses into Polynesia and Micronesia.” With the understanding that these pulses occurred at very different times (there is no Lapita in Micronesia), that may well have been the case.

Scarlett Chiu Research Center for Humanities and Social Sciences, Academia Sinica #128, Section 2, Academia Road, Nankang, Taipei 11529, Taiwan ([email protected]). 12 VII 11

Relatedness of Lapita Peoples After examining the complex and sometimes contradictory linguistic, genetic, and archaeological records related to the Solomons, the author proposes a leapfrog pattern of colonization, instead of the original wave-of-advance model, to explain the absence of Lapita sites between Bougainville and the end of the main Solomon chain and the loose associations among neighboring languages and genetics. The strongest evidence for the leap concentrates on similar linguistic and genetic characteristics found at both the Reef/Santa Cruz Islands and St. Matthias/Mussau Island (Friedlaender et al. 2005; Ross and Næss 2007) and the unusually high amount of Talasea obsidian from

821

the St. Matthias/Mussau region found at the Reef/Santa Cruz Lapita sites (Green and Kirch 1997; Sheppard and Walter 2006: 59). Talepakemalai, in the Mussau Islands, and the Reef/Santa Cruz Lapita sites both show consistent preference for Talasea obsidian during the early period, even while having access to various nearby sources. It was only during the post-Lapita period that Talasea obsidian was replaced, indicating a decrease in the long-distance exchange system over time in both regions (e.g., Doherty 2007; Green 1987, 2010; Green and Kirch 1997; Kirch 1988b; Sheppard 1993). Green (1987:246) proposed that these Lapita communities “wished to maintain ‘ties’ with their relatives . . . by importing a luxury and status-maintaining item with social and ideological significance.” Thus, direct contact between inhabitants at a distance of more than 2,000 km, is assumed to have lasted for hundreds of years and to have resulted in such strong linguistic and genetic similarities as to separate these two island groups from their surrounding neighbors. However, since the Reef/Santa Cruz case is highly unusual in the study of Lapita cultural complex, whether the leapfrog model may be applied to other areas remains questionable at this stage. There are several issues that deserve more attention. First of all, if obtaining the luxury and status-maintaining Talasea obsidian and marriage partners was the main reason for the Reef/ Santa Cruz islanders to sail up to 2,000 km, what did they have to offer in exchange? Could it have been other forms of ritual or luxury items made in the Reef/Santa Cruz Islands? Since there is no report of a vast amount of Reef/Santa Cruz pottery showing up in the archaeological records of the Bismarcks and there is also no indication of any shell-ornament workshop, such as the one found at Talepakemalai, the possibility of identifying such items must be further investigated, especially they were perishable in nature, such as feather money or root crops. Although this may never be able to be proved archaeologically, it is possible that these “valuables” also included expected friendships in terms of water and food supplies, with a resting spot needed after passage of the 360-km open sea. This may well also have included sailing knowledge about the location of islands, the zenith star path, and seasonal wind direction from the main Solomons to the Santa Cruz group and beyond (Irwin 2006:73), very useful knowledge after the destructive WK-2 volcanic explosions. As we cannot rule out the possibility that the Reef/Santa Cruz islanders might have also accumulated Talasea obsidian through exchange with their local partners, emphasis on direct contact between these two regions may be a bit risky. Second, in contrast to the similarities observed from linguistics, genetics, and obsidian procurement, such relatedness does not appear in the Lapita pottery assemblages of these two groups. Judging from vessel forms, motif preferences, the use of lime infills or carving, and the maximum thickness of pottery vessel walls, these two areas do not share the same stylistic and technical preferences, as has been pointed out by several authors (e.g., Anson 1986; Green 1978; Kirch 1997). As I have argued in the case of Lapita Site 13A and the Reef/


822


Santa Cruz sites, special emphasis on different face motifs of these two regions may reflect the local efforts of constructing social identities, in the sense of competing with one another while still using the same limited inventory of materialized symbols (Chiu 2005, 2007). Whether a similar scenario had also occurred between the Reef/Santa Cruz and St. Matthias/ Mussau groups remains to be investigated. At this stage, a database for the study of Lapita pottery has been developed, and collaborations with various scholars have been conducted through the past six years; these will no doubt provide us with more insights into this issue (Chiu 2011). Only through more detailed analysis of the contexts and environmental constrains in which Lapita pottery was used may we one day be able to illustrate with greater accuracy what Lapita may have meant to the migrating Austronesians and non-Austronesians, as they strived to make history in the Pacific some 3,000 years ago.

Geoff Irwin Department of Anthropology, University of Auckland, Private Bag 92019, Auckland, New Zealand ([email protected]). 10 VII 11 The proposition that Lapita colonists leapfrogged past the Solomon Islands is a very reasonable one that could be falsified by the discovery of one early Lapita site there. Furthermore, if Lapita colonists did not settle in the main Solomons chain, they appear to have treated mainland New Guinea in the same way and probably for the same reasons. Coastal regions of both mainland New Guinea and the Solomons must have been traversed, and certain resources were discovered, but there is still no evidence for early settlement. As Sheppard notes, both the Lapita sites in the Western Solomons and those newly reported for the south Papuan coast near Port Moresby (McNiven et al. 2011) appear to be later than the early colonizing period. Sheppard argues that the expectations of a leapfrog model would be disjunctions in the evidence and that these are to be found perhaps especially in linguistic and modern genetic data, which are outcomes of prehistory rather than a close record of events. By contrast, it could be argued that the expectations of a wave-of-advance model fit continuities in the archaeological evidence, insofar as early Lapita sites share an elaborate design system, most of their pottery is locally made and, beyond the Reef and Santa Cruz Islands, there is little imported obsidian, and that could have traveled in colonizer mode, as Sheppard explains. However, this dispersal was not driven by demographic pressure arising from food production. By its very speed Lapita simply outran the ability to diversify, and the broad material similarities among sites camouflage the process by which they spread. The leapfrog model draws attention to interesting questions, such as How many groups bypassed the Solomons,

From which sources in the Bismarcks did they come, and Were they independent of one another? My own early assumption was that there was no continuing and substantial supply of recruits from the Bismarcks to the frontier of settlement once it had advanced deeply into Island Melanesia: “The spread of Lapita communities in Oceania, while not necessarily a simple passage eastward, occurred rapidly. Insofar as these communities had to reproduce biologically and materially, it was a gradual process that took some time. Yet insofar as archaeological sampling and dating are imprecise, the Lapita expansion appears practically instantaneous” (Irwin 1981:483). The kind of exciting work now being done on human remains from the Lapita Teouma cemetery promises to throw light on such questions (e.g., Bentley et al. 2007; Valentin et al. 2010). It would be easy to exaggerate the influence of ENSO as an explanation for the timing of Lapita expansion. El Ninõ westerlies were available from around 1400 BC (Anderson et al. 2006; Veth, Hiscock, and Williams 2011) but were not used then. The northwesterly monsoon was always available. El Ninõ was just another source of westerly winds, but these sometimes brought their own problems. They varied in intensity and extent, and they could create unstable sailing conditions, with reduced visibility, and on occasion they could be difficult to return against. Sailing canoes were fundamental to Lapita culture and able to sustain a rapid and extensive maritime migration. A likely Lapita type was a single-outrigger canoe with a simple twospar spritsail able to sail by reaching with and across the direction of the wind. That level of performance allowed the strategic use of known seasonal wind patterns and was sufficient to explore an ocean (Irwin 2008). However, to sail beyond the Solomons represented a quantum leap in navigational competence, from coastal navigation within a corridor of intervisible islands to wayfinding in open ocean. Lapita sites were widespread in the Bismarck Archipelago by 1300 BC, but there was evidently a short interval before they sailed offshore after 1200 BC. During this time, Lapita sailors could have learned how to search at sea and survive, perhaps as they explored the coast of the main Solomons. If sailors knew the zenith stars of their own islands— as they surely did—then they had control of a conceptual equivalent of north and south. The seasonal alternation of monsoonal and trade winds allowed them to sail east or west. These basic elements of navigation permitted search and return. For steering a course at sea they had the direction of the sun, stars, and ocean swells. For estimating their position they had dead reckoning based on memorized information about the route sailed. To make landfalls they had high islands, clouds, and birds feeding at sea. We will never know exactly how Lapita sailors navigated, but archaeological outcomes show that their methods were equal to the task. The push and pull factors for the dispersal of Lapita are conjectural and often expressed as metaphor; for example, a pulse from some pump, the leap of a frog, the trigger of a



gun, or a sudden gust of wind. Peter Sheppard invokes serendipity and historical accident, to which I would add choice and contingency.

Patrick V. Kirch Departments of Anthropology and Integrative Biology, University of California, Berkeley, California 94720, U.S.A. ([email protected]). 10 VII 11 Peter Sheppard proposes that a wave-of-advance model of Lapita expansion out of the Bismarck Archipelago through the Solomon Islands and beyond into Remote Oceania be replaced with a leapfrog model in which the main Solomons group was initially bypassed. In this model, colonization proceeded directly from the Bismarcks homeland region to the Reef–Santa Cruz Group of the far-eastern Solomons, where Roger Green and others documented extensive early Lapita occupations that included imports such as Talasea obsidian. Sheppard argues that such a leapfrog model correlates better with our current understanding of patterns of linguistic and human biological variation across this region than does a wave-of-advance model. He may be right, and I certainly agree that the use of alternative working models or hypotheses is important in any science. Thus, Sheppard’s article is a valuable and welcome contribution. However, acceptance of the leapfrog model depends entirely on the assumption that archaeological survey of the main Solomon Islands has been sufficiently thorough or comprehensive as to demonstrate that early Lapita sites are, in fact, absent in this region. I hate to be a spoiler, but I am not yet convinced that this assumption is warranted. With regard to the main Solomon Islands region from Bougainville in the west to San Cristobal in the east, Sheppard writes that “early Lapita sites have not been found, despite what can be argued to be amounts of research at least comparable to that expended elsewhere before the discovery of Lapita deposits.” As a participant in both phases the Southeast Solomons Culture History Programme directed by Roger Green and Doug Yen in the 1970s, I question whether the work at Ughi, Makira, Ulawa, and Santa Ana was, in fact, in any way sufficient to answer the question of whether Lapita is present in those islands. Much of the 1970s work was specifically targeted at later-period sites, in order to help define a culture-historical sequence extending from prehistory into the European-contact period. Moreover, in the 1970s none of us had any idea that early Lapita settlement patterns favored stilt-house occupations over shallow lagoons, resulting in depositional contexts (such as the one I excavated in the 1980s in Mussau) that often are elusive to standard surface survey. The Solomon Islands constitute a very large archipelago, one that has had the least archaeological investigation of any comparable island group in Melanesia. Large islands such as Malaita and Choiseul have had effectively no archaeological

823

survey at all. Coastlines are often dominated by mangrove swamps, transport and access (usually by dugout canoe) is challenging, and as Sheppard himself notes, there are major issues of geomorphic changes within the past 3,000 years, occasioned both by eustatic sea level and by local tectonic and anthropogenic land-use changes. All of these factors greatly confound and complicate survey and site discovery. As an example of how intensive fieldwork can suddenly change our understanding of the archaeological record, one may cite the case of the southern Papuan coast. Sheppard states that “there are no Lapita sites reported, although there is a later Early Papuan Pottery (EPP) horizon that, as a colonization event, has many similarities with Lapita and is believed to correlate with the spread of Austronesian speakers considered to be Lapita descendants.” Despite extensive survey work by Jim Allen, Geoff Irwin, and others along the southern Papuan coastline, nothing predating the EPP horizon had been discovered. However, the situation changed radically in 2010, when intensive subsurface sampling near Port Moresby (occasioned by impending construction of a major natural gas terminal) turned up ceramics (including some with dentate stamping) that predate the EPP period and appear to be from a late Lapita phase (Bruno David, personal communication, May 2010). Thus, it now seems that the EPP period was not the initial colonization phase along the southern Papuan coast, radically changing our model of colonization of this region. Time—and a great deal more archaeological effort—may prove Sheppard’s leapfrog model to be correct. Or, when intensive archaeological work eventually does take place on islands such as Malaita, Choiseul, and San Cristobal, it may prove to be the case that there are Lapita sites distributed through the main Solomons, as Roger Green originally predicted on the basis of the presence of chert and other materials in his Reef/Santa Cruz sites (materials that arguably have a provenance in the main Solomons). For the time being, it is very useful to have Sheppard’s alternative model as a different way of thinking about Lapita expansion and colonization. But the hard work of developing a rich archaeological record for the main Solomon Islands still lies before us.

Ian Lilley Aboriginal and Torres Strait Islander Studies Unit, University of Queensland, Brisbane, Queensland 4072, Australia ([email protected]). 22 VI 11 At an abstract methodological level, the test for this paper is whether Sheppard can convincingly make his case not only archaeologically but also linguistically and biologically, without any one data set overdetermining interpretation of the others. The latter has long been a failing in Pacific archaeology, where linguistics and biology have very often been used to decide how archaeological patterns are described and explained. In the pre-


824


sent case, however, the boot could have been on the other foot, had Sheppard used the archaeological case against wave-ofadvance colonization as a hammer to fit linguistic or biological square pegs into the archaeological round hole of leapfrog migration. Much to his credit, he succeeds in avoiding this trap. He explains the archaeology in archaeological terms, and (re)interprets the linguistics and biology largely in their own terms, ending up with three sets of analytically separate but mutually reinforcing interpretations. Whether the whole exercise is empirically warranted by the archaeology is another matter. One can never say “never” in our game, as we are regularly reminded by finds that overturn generations of orthodoxy. The Papuan south coast is an apt example in this context. Pacific archaeologists (including Sheppard, as is plain in this article) have all been holding their collective breath while Ian McNiven, Bruno David, and colleagues (2011) confirmed new Lapita finds dating from around 2900 BP on the Papuan coast near Port Moresby. We have known for decades that the hitherto earliest ceramics in the region were Lapita derived. Yet until now—and despite decades of searching—no Lapita material other than a couple of sherds from the Sepik coast has been found anywhere on the New Guinea mainland. The researchers who found the Port Moresby Lapita material had previously suggested that it might be present on the strength of an earlier unanticipated discovery in Torres Strait of stylistically undiagnostic ceramics of New Guinean origin dating older than the oldest Papuan pottery. That they found it, however, is owed not to standard archaeological research in the places we would normally look. Rather, it was the requirements of a large-scale resource-development project, which forced archaeologists to look in places dictated by nonarchaeological criteria, that led to the discovery. The Papuan Lapita is not where we would expect it to be on the basis of our collective experience elsewhere in the Pacific. How, in this context, should we assess Sheppard’s claims concerning the absence of Lapita in the Solomons? Have he and his colleagues looked in all the right—or “wrong”—places? It would seem that the answer is “yes.” Of course, it remains possible that early, “classic” Lapita will turn up somewhere in the main Solomons chain. However, Sheppard and his coworkers appear not only to have looked very thoroughly in all the places Lapita would normally be expected but also to have gone to considerable lengths to identify unusual locations in which the ceramics might be found owing to geotectonic activity. In “going the extra mile” in this fashion, the researchers have addressed as comprehensively as currently possible the question of whether the absence of classicLapita evidence is really evidence for the absence of classic Lapita in the Solomons. On that basis, the leapfrog model Sheppard proposes thus seems the best hypothesis we have at the moment to describe and explain the initial Lapita colonization of Remote Oceania.

Lisa Matisoo-Smith Department of Anatomy and Structural Biology and Allan Wilson Centre for Molecular Ecology and Evolution, University of Otago, P.O. Box 913, Dunedin 9054, New Zealand ([email protected]). 1 VII 11 It has become increasingly obvious that simplistic models such as the wave-of-advance (WOA) model do not fit with the archaeological, linguistic, or biological data, so it is good to be explicitly discussing other scenarios for understanding the settlement history of the Pacific. The leapfrog model, initially proposed by Sheppard and Walter (2006) for the Solomon Islands and now expanded on here to address the Pacific in general, is one possible alternative. This model does better explain the current archaeological data, particularly with regard to the speed and directionality of migration and the demographic implications. Unfortunately, however, as Sheppard points out, the direct connection between archaeology, language, and biology is very difficult to determine. Sheppard discusses the current genetic data from Pacific populations and argues that the disjunction observed in the patterns of mtDNA, Y chromosome, and microsatellite data are all inconsistent with a WOA model. I certainly agree with him. However, I would have to say that the current human genetic data for the Pacific are not really adequate for testing any alternatives, which reflect the likely much more complex history of Pacific population origins and interactions. To date, the majority of the fine-grained genetic data for the Pacific come from the collections of Friedlaender and colleagues, which focus on northern Island Melanesia. These studies have opened our eyes to the complexities of population histories in this region, as only well-collected and focused population sampling and analyses of a range of genetic markers can do. However, if we are to ever be able to test models of the colonization of Remote Oceania, we will require much better sampling of key populations there—for example, in New Caledonia, Vanuatu, and Fiji. Micronesia is a region that is consistently overlooked or underrepresented in most genetic studies, as is East Polynesia. Recently published, well focused, and specific modern population sampling, combined with ancient DNA analyses in the Gambiers (French Polynesia), identified several individuals (both ancient and modern) carrying haplotypes that do not belong to what was believed to be the near-ubiquitous B4a1a1a “Polynesian” mtDNA lineages (Deguilloux et al. 2011). Both population sampling and subsequent analyses of genetic data really must take into consideration the demographic effects of documented historical events as well as those associated with prehistoric patterns of trade, exchange, and political interaction identified through the archaeological record. In addition to testing simplistic models for Pacific settlement such as the WOA, we also need to challenge some other dogma that seems to have developed when discussing the biology and prehistory of the Pacific. The assumption that the identification



of Polynesian origins will fully explain the origin of the Lapita culture and peoples is illogical. Polynesians became Polynesian after the disappearance of Lapita, and while they may be Lapita derived, like most other Remote Oceanic populations, the assumption that they alone are somehow the representation of what Lapita people were, biologically, both when they left their point of origin and when they arrived in Remote Oceania, is clearly absurd; yet this is regularly implied in the various genetic studies of Polynesian/Lapita origins (Soares et al. 2011). In addition, a “Lapita-only” model, in which the settlement of Remote Oceania is the result of a single population migration, even a genetically and phenotypically variable population with necessary bottleneck events, is becoming an increasingly difficult position to argue—and one that is virtually untestable by analyses of modern population data alone (Addison and Matisoo-Smith 2010). In reality, the only way that we are ever going to be able to reconstruct the biological origins and history of Pacific populations is through analyses of ancient remains, from a combination of ancient DNA and morphological studies. Unfortunately, until relatively recently, skeletal remains from early Remote Oceanic sites were rare and generally in poor condition. Recent excavations in New Guinea, Vanuatu, and elsewhere, however, are providing opportunities to assess the real genetic signatures and other biological characteristics of Lapita and post-Lapita populations across the Pacific. Recovery of these new samples, combined with recently developed techniques in next-generation DNA sequencing and the increasing frequency of community consent and even requests for the study of these ancestral populations, makes this a particularly exciting time for those interested in Pacific prehistory. Should DNA be preserved in these samples (which, unfortunately, given the hot and wet conditions of most Pacific sites, may be too much to ask) and if studies of genetic change through time are possible, I suspect that we will see a much more complex history of migrations and interactions than most models of Pacific settlement currently acknowledge. It is only through focused and integrated multidisciplinary approaches that we might start providing some reliable data for both generating and testing new hypotheses for prehistoric population origins, interactions, and adaptations and the effects of these on the biological, cultural, and linguistic variation we see in the Pacific today. Perhaps the model proposed by Sheppard will help in this pursuit.

Andrew Pawley Department of Linguistics, College of Asia and the Pacific, Australian National University, Canberra, Australian Capital Territory 0200, Australia ([email protected]). 5 VII 11 I welcome Sheppard’s assessment of factors that may have led to or shaped the rapid Lapita colonization of Remote Oceania.

825

However, I do not think that his focus on the relative merits of a wave-of-advance model, positing gradual (slow, islandby-island) expansion driven by population increase, and a leapfrog model, positing sudden long-range leaps, accurately represents recent debate on the nature of the Lapita expansion. It is hard to think of any mainstream archaeologist who over the past 30 years has supported a gradualist model of the settlement of Remote Oceania. It has been widely accepted since the 1980s (e.g., Irwin 1981) that the initial Lapita dispersal (1) was from northwest Melanesia into Remote Oceania, (2) quickly reached the Reefs/Santa Cruz Group, (3) within a few more centuries spread as far east as Tonga and Samoa, and (4) was associated with the initial expansion of Oceanic languages into a Remote Oceania that probably had no pre-Lapita inhabitants. Of course, this scenario raises many questions about the process. A central question is How did Lapita migrants manage to generate or recruit enough people to colonize in rapid succession several extensive island groups so far from the homeland? Here a more pertinent contrast is between two kinds of rapid-dispersal models. One (call it A) posits initial maintenance of long-distance links between Remote Oceanic colonies and the Lapita homeland in the Bismarcks, allowing the recruitment of new waves of colonists from the homeland to the frontier, along with valuable resources such as obsidian and domesticated plants and animals. Another (call it B) posits almost immediate loss of contact with the homeland, so that in order to extend the frontier, colonists were dependent on natural population increase or recruitment from places nearer to them than the Bismarcks. Sheppard makes a strong case for model A, as far as the settlement of the Reefs/Santa Cruz Group and perhaps Vanuatu is concerned. However, it seems that model B may fit better the history of Lapita settlement in New Caledonia, Fiji, and western Polynesia. Sheppard’s review of the historical-linguistics literature is largely confined to subgrouping, and here, understandably, he struggles to separate the wheat from the chaff. Let it be said that the linguistic evidence does not support a gradualist model of the initial spread of Oceanic languages into and across Remote Oceania. If we leave out the western Solomons, whose Oceanic languages represent a later expansion out of New Ireland (Ross 1988), the subgrouping evidence points to a swift initial spread of an almost undifferentiated Oceanic language across Island Melanesia, reaching western Polynesia without any long pauses. Such pauses would be indicated by the existence of a hierarchy of well-defined high-order subgroups in Remote Oceania in which some branches span more than a single major island group. No such hierarchy exists, except in Polynesia, which stands at the end of the line. Putative subgroups that rest on flimsy evidence do not imply a significant pause. Thus, Sheppard refers to the Central/Eastern Oceanic subgroup, tentatively posited by some linguists to comprise most of the languages of Remote Oceania, as a case where linguists have assumed a gradual spread, punctuated by pauses. However, this putative group never had much


826


evidence going for it and is now rejected, at least by Malcolm Ross and myself. The Central Pacific group, extending over Fiji and Polynesia, is also very weakly defined. Incidentally, the limitations of archaeological evidence as a means of reconstructing prehistoric cultures are starkly revealed by a question asked about Lapita in the concluding section: “Why did that elaborate cultural complex suddenly move out into Remote Oceania and almost as quickly disappear?” The cultural complex carried by Lapita people did not quickly disappear. Lexical agreements, supported by comparative ethnography, allow us to reconstruct many elements of the material and nonmaterial culture of Proto-Oceanic speakers, the bearers of early Lapita culture, that have been largely or completely lost in the archaeological record. These elements include extensive terminologies for canoe and sailing technology, fishing techniques, horticultural practices, architecture, kinship, and so on (Pawley 2007; Ross, Pawley, and Osmond 1998–2011). Lapita-style dentate-stamped pottery may have disappeared, but in certain Oceanic-speaking societies there has been a high degree of continuity in these other terminologies through to modern times. On this point, see also Kirch and Green (2001).

Malcolm Ross Department of Linguistics, School of Culture, History and Language, College of Asia and the Pacific, Australian National University, Canberra, Australian Capital Territory 0200, Australia ([email protected]). 29 VI 11 This comment evaluates the linguistic evidence that Sheppard draws on in support of his leapfrog model of Solomons settlement. He refers to my 1988 claim (Ross 1988), restated by Pawley (2009), that the boundary separating Northwest Solomonic (NWS) from Southeast Solomonic (SES) languages reflects two separate settlements by Oceanic speakers, with SES languages reflecting initial settlement by Oceanic speakers and NWS languages a later spread from the southern New Ireland area by speakers of a Meso-Melanesian language. My 1988 scenario also assumed that NWS languages had replaced older languages left by an initial Oceanic settlement. Sheppard replaces this wave-of-advance model with a leapfrog model that has implications for linguistic history. Under his scenario, Lapita-culture Oceanic speakers first leapfrogged from the Bismarcks to the Reefs and Santa Cruz (RSC) Islands, then late Lapita speakers of Proto-NWS moved into the northwest Solomons, and then, at some post-Lapita date, SES speakers spread back from the southeast into their present range (Sheppard and Walter 2006). I discuss the three stages separately. Crucial to an evaluation of the linguistic implications of Sheppard’s hypothesis is Ross, Pawley, and Osmond’s (2008: 8–15) reassessment of the evidence for higher Oceanic subgrouping, which gives nine first-order subgroups: Yapese (one language), Admiralty, Mussau-Tench, Western Oceanic (New

Guinea, Meso-Melanesian), SES, Temotu, Southern Oceanic (Vanuatu, New Caledonia), Micronesian, and Central Pacific (Fijian, Polynesian). The Central/Eastern Oceanic grouping is no longer supported. As Sheppard mentions, Ross and Næss (2007) showed that the languages of the Temotu archipelago, which includes RSC, form a first-order subgroup of Oceanic and could therefore reflect an initial leap out of the Bismarcks. This means, however, that Sheppard’s suggestion that the closest relatives of RSC can be identified is incorrect, as RSC is simply one of several first-order groups. In this connection, Sheppard also refers briefly to Gray, Drummond, and Greenhill’s (2009) Oceanic phylogeny, but the latter’s database is still under development, and differences between their results and those obtained by the comparative method remain unexplained. The rethinking of Oceanic subgrouping also bears on the suggestion that SES languages reflect a back-migration from somewhere farther east. With no Central/Eastern Oceanic grouping, SES is not especially associated with more easterly languages (but Sheppard makes no strong claim in this regard). The latter, incidentally, are diverse but by no means all aberrant, as Sheppard’s figure 5 implies. In 1988, I suggested that the SES languages have remained more or less in situ since the original Oceanic eastward settlement out of the Bismarcks. Pawley (2009) also argues, on the basis of innovations in various SES subgroups and languages, that the SES languages must have begun to spread through their present range at least 2,500 years ago. Under my 1988 scenario, the NWS languages spread later, replacing languages resulting from the original Oceanic spread. The later spread accords well with Sheppard’s claim that NWS reflects late Lapita settlement (Pawley 2009), but he infers that there were no earlier Oceanic languages to be replaced because the earlier settlers had bypassed the Solomons. Ross (2010) shows that NWS languages have a small number words of Oceanic origin that have unexpected forms and appear to have been borrowed from earlier Oceanic languages. They are (albeit weak) evidence for the replacement scenario. Pawley (2009) and Ross (2010) both highlight how radical the differences between NWS and SES histories are. NWS and SES each form an innovation-defined subgroup, but NWS languages display far more interisland and local diversity and far more lexical innovations than SES. The obvious explanation of this is that NWS speakers have had quite intense contact (through bilingualism) with speakers of diverse preexisting Papuan languages, whereas SES speakers have not. In sum, the RSC evidence is consistent with Sheppard’s leapfrog hypothesis, but both the NWS and SES evidence speak for a modified version of it. On linguistic evidence, the earliest settlement (either before or after RSC settlement) probably left Oceanic-speaking villages on the small islands of the Solomons and tiny enclaves on the coasts of the larger ones (this accords with Lapita settlement patterns in the Bismarcks). In the northwest, Oceanic speakers intermarried with prior Papuan populations and, given the continuing Papuan prevalence on Bougainville, were in some cases probably



absorbed by them. The later settlement by speakers of NWS languages established a continuing Oceanic presence, but one that would be significantly affected by contact with Papuan languages, even though these have largely disappeared outside Bougainville. The SES languages almost certainly reflect a continuing presence since the earliest Oceanic settlement, but the Proto-SES-speaking population was probably small, later expanding to occupy the present SES range, where there has apparently been little or no contact with Papuan speakers.

Christophe Sand Institute of Archaeology of New Caledonia and the Pacific, Rue Teyssandier de Laubare`de, Montravel, BP: 11423, 98802 Noumeá, New Caledonia (christophe.sand@iancp .nc). 10 VII 11 This paper hypothesizes a leapfrog process in the settlement of the uninhabited Reef/Santa Cruz Islands in Remote Oceania directly from the Bismarck Archipelago. Although wishing to be convincing, the demonstration appears to have a number of weaknesses. After having rapidly dealt with the linguistic and genetic arguments, I will concentrate on the archaeological context of the model. The linguistic demonstration proposed appears too simple, mainly because of the very limited control that linguists have on the chronology of language development in Melanesia. Language replacements are documented ethnographically in different Near Oceania contexts and in Remote Oceania. Sheppard is right to point out the possibility of different Lapita groups introducing a diversity of Oceanic languages in Remote Oceania. Yet to envision that the groups living in the Willaumez Peninsula today, in an area subject to regular eruptions, are the same as those who lived there 3,200 years ago is misleading, and to try to use the poorly dated modern languages in Talasea and the Reef/Santa Cruz Islands to infer a first-settlement hypothesis is unconvincing. Likewise, the few genetic data at hand for the region are not robust enough to be used in a definitive way. Sheppard says rightly that “there is no simple ‘Lapita’ genetic signature,” especially when postdiscovery founder effects are taken into account. So to advocate that more than 3,000 years after the spread of Lapita—after 130 generations of complex Melanesian history and a number of catastrophic volcanic events, not to mention the population crash of the post–European contact era—we can directly and indisputably see genetic links related to first Austronesian settlement between Bismarck and Reef/Santa Cruz populations is premature. The archaeological part of Sheppard’s demonstration is also open to other interpretations. The large islands of the Solomons have a complex geomorphology. This implies that, even if the different projects undertaken in this region have failed to find early Lapita occupations, to infer that those sites just do not exist is too simple. A similar model of Lapita “bypassing” was

827

proposed for Vanuatu more than 20 years ago, before being dismissed by more detailed research by Bedford. The absence of Lapita sherds in the Central Solomons echoes the near absence of discovered pre-Lapita sites and calls above all for renewed studies. The recent discovery of a series of Lapita sites in the Port Moresby region of southern Papua is testimony that sometimes things can be missed. This does not mean that there was necessarily widespread establishment of the ceramic tradition: Lapita sites might be few in the Central Solomons. If this is the case, it would simply imply that the technology was never picked up by the main inhabitants of the large islands, although they knew about the technique from neighboring pioneer Lapita communities. Several examples across Remote Oceania (southern Vanuatu, the Loyalty Islands) show that the present archaeological pattern can also be a result of rapid abandonment of pottery making in this region after Lapita, leaving only a faint archaeological signature. Likewise, to advocate that the apparent absence of reuse of obsidian in the Reef/Santa Cruz Islands is testimony of a direct access to the Talasea source is not conclusive, as recent excavations in the Teouma cemetery of Vanuatu show the presence of a significant amount of Talasea obsidian, which cannot be explained by direct importation from New Britain. Other explanations, such as a “down-the-line” exchange network of a Kula type, might as well explain the amount of accessible raw material. In conclusion, the arguments proposed by Sheppard to advocate a leapfrog Lapita settlement of the Reef/Santa Cruz Islands needs further archaeological testing and must be seen mainly as a working model on available data. His stronghold is probably the “push” factor owing to the combined absence of malaria and presence of empty islands, which certainly attracted a whole set of different groups with different sublanguages. But to envision that the large islands between the Bismarck Archipelago and the entry point to Remote Oceania, a distance of around 1,500 km, were not used at least as stepping stones is questionable. Consequently, the paper is too reductive in the definition of the “wave-of-advance” model and its biological consequences: there can be a rapid movement across a region already occupied to reach better places farther afield, with just some settlements along the route. This implies more a “string-of-pearls” model of settlement than an exclusive leapfrog process. One way of testing this might be a detailed analysis of motif/form variation in Lapita ceramics, an avenue completely neglected in the present paper.

Jim Specht Australian Museum, 6 College Street, Sydney, New South Wales 2010, Australia ([email protected]). 2 VI 11 Sheppard offers a welcome contribution to discussion about the settlement of the western Pacific islands. It is an impressive review of three fields of archaeology, historical linguistics, and


828


human genetics relating to the dispersal of people(s) making Lapita pottery into the hitherto uninhabited islands of southeastern Remote Oceania. All three lines of evidence, however, have major gaps in geographical coverage and sampling depth, and there is some way to go before we can be confident that they support the “leapfrog” model over the “wave-ofadvance” one, rather than simply appearing to do so, on account of imperfect evidence. Notwithstanding this caveat, the paper presents sound overviews of the two models, their implications, and the available data. Sheppard’s interpretation reduces the need for multiple migrations of different ethnolinguistic groups to account for “aberrant” Austronesian (AN) in Remote Oceania that some linguists have invoked (e.g., Lynch, Ross, and Crowley 2002; Blust 2008) and the apparent difficulty in deriving a “Polynesian” phenotype from Melanesia (e.g., Spriggs 2003: 207–208). Rather than focus on narratives of migration, we should pay more attention to the evidence for repeated contacts between the Bismarck Archipelago and Remote Oceania, viewed as a “lifeline” back to the homeland communities to minimize risks to newly established Lapita settlements (Kirch 1988a:113; Summerhayes 2000, 2010; cf. Soares et al. 2011 for Island Southeast Asia and the Bismarck Archipelago). There is no reason to assume that all journeys associated with that lifeline were leapfrog movements; some intermediate, short-term “way stations” no doubt developed, although these need not have left a significant archaeological signature. Sheppard nevertheless presents a reasonably convincing case for the leapfrog model in terms of differential tectonic movement in the western Solomon Islands. On Buka to the north, decorated Lapita sherds found in intertidal contexts on the fringing reef are well preserved (Wickler 2001), and I see no reason why this would not apply to the western Solomon Islands pottery sites if Early and Middle Lapita pottery were present there. At the same time, few coastal areas of the Solomon Islands have been studied in as much detail as these islands, and given the scale of geomorphological changes that occur over both the short and long terms, it may be premature to dismiss completely the possibility of Early and Middle Lapita sites in other intertidal locations. The distribution of obsidian from the Bismarck Archipelago has been used as a possible indicator of source populations for the colonization of Remote Oceania and of subsequent interactions between islands (e.g., Green 1987; Specht 2002; Galipaud and Swete Kelly 2007b; Reepmeyer and Clark 2009; Sheppard, Trichereau, and Milicich 2010; Spriggs, Bird, and Ambrose 2010). While Willaumez Peninsula obsidian dominates in the Reef/Santa Cruz Group, this need not mean that the people who brought it came directly from that source area. Communities living outside the natural range of obsidian could have also contributed to the colonization of Remote Oceania. This might be picked up through detailed stylistic analyses (cf. Chiu 2005, 2007; Noury 2005) and perhaps through linguistics, genetics, and stable-isotope analyses (cf. Bentley et al. 2007 for Vanuatu). In this context, the presence

of Fergusson Island obsidian in the RF-2 site in the Reef Islands (Green and Bird 1989) reminds us to look widely in our comparisons. The sharing of the M28a2 haplogroup between Santa Cruz and New Britain groups speaking both AN and non-Austronesian (NAN) languages (table 4.2 in Friedlaender et al. 2007a) adds a new twist to the picture: the groups that colonized Remote Oceania were not only speakers of Oceanic languages but probably included speakers of NAN languages (cf. Pawley 2006:248). The equation of Lapita pottery makers with speakers of Oceanic languages has a long history (e.g., Pawley and Green 1973) but now needs a serious rethinking (Donohue and Denham 2008). This has implications for narratives of the origin of the “Polynesians,” for which Addison and Matisoo-Smith (2010) have proposed a new model that questions the “Lapita cultural complex” as the sole foundational culture of the various Polynesian groups of today (e.g., Groube 1971; Kirch 1997; Kirch and Green 2001). Two major points emerge from Sheppard’s paper. The first is that unilinear narratives of the settlement of the Pacific islands obscure deficiencies and gaps in data and theory and miss the much more complex and nuanced pictures that are now emerging. These pictures are founded firmly on data, particularly data derived from new analytical techniques, but further development of theory is needed. The second point is that the time has come for archaeologists, linguists, and geneticists to abandon treating the “Lapita cultural complex” as an ethnolinguistic entity and to stop assigning such tags to archaeological materials; there are many more interesting archaeological questions relevant to the realm of archaeology that can be addressed.

Matthew Spriggs School of Archaeology and Anthropology, Australian National University, AD Hope Building 14, Canberra, Australian Capital Territory 0200, Australia (matthew.spriggs@ anu.edu.au). 12 VII 11 Is the sample of known sites in the Solomon Islands sufficient to state with confidence that the absence of evidence of Lapita sites in much of the main Solomons chain represents a real evidence of absence? Sheppard believes this to be the case, but I remain to be convinced. I agree with Sheppard that a wave-of-advance model is not a suitable one in conceptualizing Lapita colonization and that models incorporating historical accident and serendipity are nearer to the reality. Settlers of any one Lapita site could have arrived from almost any direction, given the rapidity of settlement across the Western Pacific. But I find it hard to believe that in this early phase of exploration and colonization the main Solomons chain would be the only archipelago in the entire region to be leapfrogged over almost entirely and not



subject to rapid backfill settlement in an archaeologically instantaneous time span. It cannot be because the island chain was already settled, although it doubtless was; the Bismarck Archipelago was, similarly, previously inhabited. As Sheppard notes, there does not seem to be any particular environmental distinction making the Solomons less suitable for settlement than anywhere else. Malaria may be a relevant factor, but surely it was for the Bismarcks too. A theory of post-Lapita environmental change was long ago invoked to explain the apparent lack of Lapita settlement, with sites postulated as being buried by alluvial and colluvial deposition and therefore difficult to discover without a major project of subsurface testing (Spriggs 1984; cf. Spriggs 2010). This hypothesis is hard to dismiss in the Solomons context, where large volcanic and easily erodible islands most certainly create difficult conditions for finding any sites older than a few hundred years. I remain skeptical, for the same reason, of Dickinson and Green’s (1998) explanation of the absence of Lapita sites in Samoa, apart from the easternmost Lapita outpost at Mulifanua, as being solely due to subsidence. Felgate (2007) has more recently added other complicating factors relating to local geotectonics in the Western Solomons. It is notable that in the Bismarcks, many Lapita sites have been found on smaller offshore islands. This was previously explained by a Lapita penchant for such locations, but it is now generally seen as reflecting relative site visibility. Sheppard claims that “early Lapita sites have not been found despite what can be argued to be amounts of research at least comparable to that expended elsewhere before the discovery of Lapita deposits.” But when is the amount of research sufficient to be able to state the negative with certainty? An exactly parallel argument was used until recently in relation to the apparent lack of Lapita along the southern Papuan coast in Papua New Guinea: numerous surveys since the 1960s by several teams of researchers had failed to find any sites of this age. The recent location of multiple Lapita sites in a small area at Caution Bay, 30 km to the west of Port Moresby, makes the earlier claimed certainties there seem hollow (McNiven et al. 2011). Indeed, the negative argument can be realistically sustained only when one can produce coastal sites in Near Oceania of the same age as Lapita that are not Lapita associated. The only island in the main Solomons where this may be the case is Guadalcanal and the rockshelter site there of Vatuluma Posovi (Roe 1993); but rockshelters are not generally favored Lapita locales. The vast majority of Lapita sites are opensettlement sites, but no such sites of the requisite time period, Lapita or otherwise, have been found on Guadalcanal. There are no sites of any type in Sheppard’s major research area of the Western Solomons that have produced dates either earlier than or contemporary with early Lapita sites elsewhere (ca. 3200–3000 BP). As Sheppard would be the first to admit, this absence of evidence is most unlikely to represent evidence of absence of human groups in the area. But he does by impli-

829

cation take it to represent evidence of absence of specifically Lapita groups. I am quite happy to entertain the idea of a special relationship between Lapita sites of the Bismarcks and the Reefs/Santa Cruz Group of the Southeast Solomons. But in postulating such a relationship Sheppard does not need to assert a lack of settlement of the main Solomons during early Lapita times. The islands are large, the geotectonics are complex, and little or no systematic test-pitting survey has been carried out anywhere in the main islands. Lapita sites should be more visible on some of the smaller, low-lying islands, but these have not generally been subject to survey of any kind outside of the Western Solomons. It is therefore not surprising that the only pottery found in the more easterly islands comes from tiny Santa Ana, at the very eastern end of the archipelago.

Glenn R. Summerhayes Department of Anthropology and Archaeology, University of Otago, P.O. Box 56, Dunedin, New Zealand ([email protected]). 4 VII 11 The discussion of “leapfrog” versus “wave-of-advance” makes for interesting reading of where Lapita research is at in 2011. Although the two models are presented here as opposing positions, I never saw them as either one or the other, with the “leapfrog” more of a description rather than an explanation of what was happening and the “wave-of-advance” model (as presented here) an anachronism that very few took seriously in modeling the Early Lapita colonization of the western Pacific. Sheppard’s paper supports the “leapfrog” camp, which basically arose out of trying to provide an explanation for the absence of Early Lapita from the Solomon Island chain, where both he and Richard Walter had undertaken extensive research. That is, Early Lapita colonizers leapfrogged over the Solomons. I agree with the description of “leapfrogging” within a wider context of highly mobile groups of people. I have just a few comments/questions. First, the absence of Early Lapita settlements in some parts of Near Oceania, such as the Solomon Island chain, is not surprising, given that this region had been populated for about 45,000 years. Early Lapita settlements within Near Oceania were located on small offshore islands or reefs away from existing populations. Their absence in the Solomons could simply reflect their inability to locate in someone else’s turf. Second, what exactly is a “wave-of-advance model”? Does it really equate with filling in all the dots in the landscape before moving on because of population pressure? As noted by Sheppard and others, the narrow time period for Lapita’s spread from the Bismarck Archipelago to Fiji would negate any simple population growth model to explain the rapid movements of people in this region. Sheppard is right: models using demic diffusion or demographic growth just do not apply to the Pacific. Given what we know of mobility within


830


the Early Lapita populations of the Bismarck Archipelago, a better model to use would center around highly mobile groups of people interacting across the Pacific for a number of reasons, including finding marriage partners, for instance (Summerhayes 2010). Kirch pointed out years ago that continued interaction through long-distance exchange between Lapita communities was needed until areas reached demographic levels needed for survival (Kirch 1990:128). It could be argued that the distribution of Lapita sites represent both “leapfrogging” and a “wave of advance.” As has been noted, “Most archaeologists do see a progression from west to east which was not simple. That is, the earliest occupation is in the Bismarcks, and the latest colonization events are in Tonga, with Fiji earlier than Tonga and later than the Bismarck. Yet, I do not remember reading any article that said all the dots had to filled before a new area was colonized or visited” (Summerhayes 2010:28). In such a scenario, although populations leapfrogged over the Solomons, they were part of a wave of highly mobile and interactive groups of people eventually progressing west to east. The problem here is one of scale, an issue that Sheppard does raise in this article. Finally, there is an issue in integrating linguistics into archaeological modeling. For instance, Sheppard notes similarities between the languages of the St. Matthias Group (to the north of New Ireland) and Temotu (Solomons) as an indication for a “significant leapfrog.” This is based on the assumption that the St. Matthias island group has been relatively isolated over the past 3,000 years. The problem is that we know from local oral histories and material-cultural studies that the St. Matthias Group was not isolated and that it had interactions not only with Micronesia to the north but also with areas to the east, such as the Polynesian Outliers, let alone New Hanover to the south. This island group has not been isolated in a vacuum for the past 3,000 years (see Nevermann 1933 and Parkinson 1907). Similar arguments apply to studies of DNA using modern populations. People move around. Sheppard has conducted many years of successful archaeological research in the Solomon Islands and has used the term “leapfrog” to appropriately model the absence of Early Lapita occupation in this region. “Leapfrogging” is an interesting term for a highly mobile group of people who moved from the Bismarck Archipelago to Vanuatu and then reached Fiji and Tonga within a generation or two, an archaeological instant. This article is a valuable one for Lapita studies.

John Edward Terrell Field Museum of Natural History, 1400 South Lake Shore Drive, Chicago, Illinois 60605, U.S.A. ([email protected]). 6 VI 11 At the Honiara Lapita Conference in early July 2007 organized by Peter Sheppard and others, I noted that for the past halfcentury or so, it had been largely taken for granted by scholars

and others that Lapita was an orderly series of ancient events all linked with one another by some definite, although now obscure, logic of relatedness (Terrell 2009). I argued that this was, however, unlikely to have been the case, for several reasons. I suggested that we have been mostly fooling ourselves into believing that the history of this ornate pottery and its makers adds up to a systematic and cascading series of migrations that ultimately culminated in the settlement of Polynesia. A number of other scholars have similarly concluded that the time has come to abandon what Sheppard talks about here as the once-prevailing “wave-of-advance” model of Austronesian settlement in the Pacific (e.g., Addison and MatisooSmith 2010; Burley, Sheppard, and Simonin 2011). As Matisoo-Smith and Robins (2009:1526) remarked, “the history of human settlement of the Pacific is much more complex than some early models suggest. Additional studies will most likely only add to our appreciation of how complex that history was.” Sheppard contributes a useful benchmark summary of current thinking on this issue and presents several lines of evidence supporting his proposition that “leapfrog movement from a diverse homeland region, such as the Bismarck Archipelago, allows for the generation of diversity in the populations settling the new islands [Remote Oceania], something not easily accommodated in a wave-of-advance model.” I agree, and I also concur with him that in all likelihood there were numerous population movements from the Bismarcks into Remote Oceania, resulting in founder events variously sampling from the human diversity present in the Bismarcks 3,000 or so years ago, something that has long made good sense to at least some of us working in the Pacific (Terrell 1986). I have a few minor points. I did not argue in 2002, contrary to what Sheppard suggests, that many coastlines in the Pacific would have been inhospitable rocky shores until the midHolocene. My observations referred specifically to the possibly unusual case of the north coast of New Guinea (Terrell 2006). Sheppard repeats the popular idea that the Lapita design system emphasized anthropomorphic motifs. Esther Schechter and I have argued to the contrary, that the iconography on decorated Lapita pots and dishes is probably referencing sea turtles and their associations with human beings (Terrell and Schechter 2007, 2009). Recent genetic studies (e.g., Soares et al. 2011; Terrell 2010) have further challenged how directly biological evidence derived from contemporary populations can be used to reconstruct Pacific prehistory. Case in point: it is now evident that the so-called “Polynesian motif” is poorly named—yet another example of the bias toward all things “Polynesian” that has long plagued scholarship on prehistory in the Pacific Islands. If this trait must be given a geographic tag, it would be much more informative and accurate to call it the “Melanesian motif,” or possibly even the “Bismarcks motif” (provided we include the North Solomons in the designation).



Tim Thomas Department of Anthropology and Archaeology, University of Otago, P.O. Box 56, Dunedin, New Zealand ([email protected]). 7 VII 11 This article presents the strongest and most coherent case yet for a real gap in the general distribution of early Lapita sites in the western Pacific. It is the proposed scale of this gap, requiring avoidance of the entire Solomons archipelago, that will cause most debate. The idea that Lapita migration was characterized by leapfrog patterns is not itself new (e.g., Spriggs 1997:104–105). Many Pacific archaeologists accept that early Lapita settlements are patchily distributed, particularly in Near Oceania, strongly suggesting leaps and gaps along coastlines. The wave-of-advance model is clearly inappropriate at this scale. It is also worth noting that orthodox accounts of the origins of Lapita also postulate some degree of long-distance leapfrog migration, by Austronesian speakers arriving in the Bismarck Archipelago from somewhere to the west of New Guinea, apparently skirting the northern coast of that landmass. But did Lapita settlers really avoid all of the Solomon Islands during their subsequent expansion into Remote Oceania? Sheppard’s argument, as well as the form it takes, stems mostly from the absence of archaeological evidence for early Lapita sites in the Solomons: he asks us to evaluate how well this absence matches (and explains) irregular patterns in the distribution of languages and genetic markers. As such, the argument is based on plausibility and, to some extent, elegance. If Lapita had its origins in landscapes depopulated by the WK-2 eruption and migrating colonists subsequently favored similarly unpopulated places, then the absence of sites in the already-populated Solomon Islands and nearby coastal areas of New Guinea might be expected. This could also explain why the languages of the Reefs/Santa Cruz Group are more closely related to those of the Bismarcks than to those of the western Solomons and why genetic diversity patterns appear to follow suit. Such an account appears elegant enough, but it does require subsequent complexities: a backmigration of people from the Reefs/Santa Cruz Group to the southeast Solomons to account for the Tryon-Hackman line and a late Lapita migration from the Bismarcks into the western Solomons to account for the presence of ceramic sites and Meso-Melanesian languages. The problem with arguments that rely on plausibility, however, is that other scenarios can be similarly plausible. Ross and Næss (2007:459), for example, find that a back-migration of people from the Reefs/Santa Cruz Group is unsupported by the linguistics, while a leapfrog event from the Bismarck Archipelago to the Reefs/Santa Cruz Group is likely (p. 461; also Næss and Boerger 2008:210), thus implying something even more complex than what Sheppard envisages. Felgate (2007) has proposed that early Lapita expansion was demographically weak in the Solomons, “limping” because of extant

831

populations and the effects of malaria and other natural predators—explaining why early sites are far rarer there than in Remote Oceania, where expansion was explosive. Similarly, if Lapita people in the Bismarck Archipelago occupied landscapes depopulated by the WK-2 eruption, then we might expect their descendents to group more closely with Lapita descendants in Remote Oceania than with Lapita descendants who mixed with previously resident populations in the Solomons. More pessimistically, modern genetic and linguistic patterns may not track cultural patterns operating more than 3,000 years ago very well at all. My own view is that the linguistic, biological, and archaeological data for the Solomon Islands are still too sparse and poorly understood to draw firm conclusions from. I have argued in support of the leapfrog model on genetic grounds, particularly the distribution of the M27 haplogroup (Ricaut et al. 2010), but I do not think that this is conclusive. I take Sheppard’s point that numerous archaeological surveys over a 40-year period should have discovered Lapita sites if they were in the Solomons, but I also know that much of the landscape is forested and difficult to access and has had little development impact, with consequently less accidental site discovery. So we clearly need to agree on what constitutes archaeological evidence for absence, criteria allowing us to accept that we have looked hard enough and found nothing. Clark and Bedford (2008) have suggested that a demonstration that imported pottery in early Reefs/Santa Cruz Lapita assemblages, like the obsidian, was brought only from locations west of the Solomon Islands would be sufficient proof. But at the risk of being annoying, I think that more systematic archaeological surveys targeting likely islands are still warranted. Overall, however, the stance that Sheppard has taken is immensely productive, in that it proposes a genuinely falsifiable model and embodies an epistemologically conservative position. A key underlying theme here is the way we should approach data in describing prehistory: Should we create broad narratives from sparse data points, filling in the gaps with our expectations, or should we simply speak from the data as they currently stand, being always prepared to be corrected by new evidence? If the latter, then leapfrogs are a good fit, for now at least.

Robin Torrence Australian Museum, 6 College Street, Sydney, New South Wales 2010, Australia ([email protected]). 11 VII 11 Despite the inherent difficulties of relying on the absence rather than the presence of archaeological evidence, Sheppard’s argument about Lapita colonization is convincing. The geotectonic, linguistic, and genetic data he presents support a leapfrog pattern rather than a wave of advance. The rapid speed of Lapita potters moving from the Bismarck Archipel-


832


ago across the Solomon Islands chain does not conform to models requiring populations to reach relatively high levels before groups bud off and move onward. In contrast, why such a huge leap and why at this particular time are important questions not satisfactorily addressed. Sheppard retreats into “serendipity and historical accident, rather than some systematic process.” Although I agree that a stochastic event played a crucial role, it does not follow that larger processes did not shape why and how the Solomon Islands leapfrog occurred. Competition for status and subsequent changes in ideology are potential key factors in the colonization of the Remote Pacific (Torrence 2009; cf. Gosden 1992) but are missing in the current models discussed by Sheppard. I agree with Sheppard that the WK-2 volcanic eruption in New Britain was a major trigger in stimulating migration, but a further process is needed to explain the leapfrog. This massive natural disaster destroyed a vast region and forced a large population to abandon their homes and seek refuge elsewhere (Torrence, Neall, and Boyd 2009a; Torrence and Doelman 2007). But why did some refugees head out to Remote Oceania and skip over the Solomons? As Torrence and Grattan (2002) have argued, it is important not to confuse correlation with causation. WK-2 was not a unique event. The WK-1 event, ca. 2,600 years earlier, had similar environmental effects, but cultural continuity was not significantly disrupted (Torrence and Doelman 2007). The key difference between the two disasters was the particular cultural setting in which WK-2 occurred. Until recently, understanding the origin and spread of Lapita pottery has been prohibited by lack of information about the cultural context in which it was invented. Recent research shows that before WK-2, a social system in which some individuals maintained a high status, partly through the competitive exchange of valuables such as obsidian stemmed tools, operated in New Britain and possibly over a much larger area (Torrence 2011; Torrence and Swadling 2008; Torrence et al. 2009b). The destruction of gardens and other resources, such as obsidian sources, by WK-2 curtailed the ability of elites to support craft specialists and other activities, such as feasting. I have proposed that either they scrambled to hang on to their status or, more likely, others who had previously been disenfranchised or were in competition with them took advantage of the disruption to create a new ideological system, in which Lapita pottery played a part. This new cult, developed in competition with the status quo, incorporated concepts promoting movement out of the settled areas, for example, searching for a new “promised land” (Torrence 2009). It is notable that Lapita pottery occurs only in areas that were not currently occupied, whether in the Bismarck Archipelago or Remote Oceania. The Lapita cult followers were not welcomed in areas practicing a different means of status enhancement and control. Since the Solomons were already occupied at this time—as were New Ireland and the Admiralty Islands, where Lapita is found only in new, empty spaces such as small offshore islands—people practicing the new Lapita

pottery cult were forced to move on until they found pristine lands. Although the Witori eruption provided an opportunity for migration, it was the transformation of the existing ideological system that actually provided the rationale for movement. Once underway, a process—competition between the new cult and existing societies—shaped the leapfrog pattern of colonization within both the Bismarck Archipelago and the Solomon Islands. To better understand Lapita pottery, we need to accept that what differentiates this period is not primarily subsistence or settlement but ideology and social practices (cf. Spriggs 2011). Sheppard uses starch granules with identical properties to argue that before Lapita pottery people had familiarity only with plants but that during Lapita they “used domesticates,” although he notes that domestication is “hard to determine from starch grains.” This represents a case of finding what you were looking for and not being equally critical of the same evidence in different contexts. Many taro and yam starch granules have been recovered from Early Holocene contexts (e.g., Torrence 2006:19), but lacking secure criteria to differentiate wild progenitors from domesticates, claims for gardening have not been made, unlike the Lapita cases. Sheppard’s assumption of low population density and mobile settlement before Lapita also must be reexamined, given recent evidence from extensive excavation and use-wear studies that in New Britain there is very little difference in population levels and settlement patterns before and after the Lapita pottery period (Torrence 2011; Kononenko 2007, 2011).

Reply I would like to thank all my colleagues who have commented on this paper; it is a real testament to the vitality of the Lapita research community. Venturing outside your own discipline to pull together data from a number of different disciplines is always difficult and inevitably results in some simplification, which can irritate the specialist. In this case, my colleagues outside archaeology have been very tolerant, and it is instructive to compare the views from linguistics and biology with those from archaeology. Given the large number (17) of comments, I cannot respond to them individually but will focus on a number of recurring themes. Bellwood notes that the Lapita expansion is ultimately part of population movements out of Southeast Asia facilitated by advances in Neolithic and maritime technology. I would certainly agree that at the large scale this is true. Without the boats and, ultimately, without the domesticates, occupation of Remote Oceania would not have been possible, although settlement and interisland movement were obviously not determined by these factors in Near Oceania and Island Southeast Asia. However, the main point of my paper is that if one wishes to explain the details of the archaeological record and



also the linguistic and biological record, one needs to consider regional specifics. Fast-train models from Taiwan to Easter Island or simple linguistic correlations are convenient but will ultimately leave much unexplained variation. At the end of the day, it is the regional variation that is of most interest to the culture historian working in the field. I have focused on the Near/Remote Oceania boundary because I think that there is much unexplained variation across this boundary. I leave Island Southeast Asia to Bellwood and other colleagues working in the area, and in addition, I have stayed away from questions of the “origins” of Lapita, at least as regards linkages to language and biology that seem irresolvable. Bellwood, along with a number of others, comments on the nature of the wave-of-advance (WOA) and leapfrog models. They are correct in stating that WOA is an explanatory model, as it has demographic drivers, while leapfrog has no driver and is thus a description. Bellwood and Summerhayes argue that there is no necessary difference between the two. However, unless one wishes to remove the theoretical basis from WOA, it requires demographic growth and filling of new lands. A better model closer to the leapfrog is the ideal free distribution, which is a WOA with an evolutionary optimizing basis that can create leapfrogs or the string of pearls noted by Sand. My leapfrog description has no theoretical basis other than to acknowledge that much of known history at the middle to small scale involves accidents and individual decisions that can have far-reaching effects. Of course, in linguistics and biology these might be theorized as founder effects; however, archaeology typically finds such events hard to “explain.” The linguists and many archaeologists agree with the notion of an early leapfrog across the main Solomons. For them, nothing else makes sense. Pawley goes so far as to state that “mainstream” archaeologists have known this to be true for a long time. Roger Green certainly believed that Lapita would be found in the main Solomons, marking on his 1979 map of Lapita sites New Georgia and Marau Sound in Guadalcanal as likely locations. The finding of chert from Malaita/Ulawa in the Reef/ Santa Cruz Lapita sites seemed to confirm the probability. Of course, the finding of Lapita anywhere in the Solomons from any period does not preclude a leapfrog, and it is likely that this is the scenario that Pawley is referring to, as is Summerhayes. In my opinion, finding early Lapita in the Eastern Solomons would only make the situation even more anomalous. Of course, it is always possible that Lapita may be found throughout the Solomons and at all time periods. As Irwin and Thomas note, the leapfrog hypothesis is at least falsifiable; the alternative can never be falsified, just as the suggestion that pre-Lapita occupation may exist in Vanuatu or New Caledonia, which might make some linguists happy, can never be falsified. Many commentators note the cautionary tale provided by the surprise recent finding of Lapita sites near Port Moresby after extensive survey and testing (McNiven et al. 2011) as part of a cultural resource management survey. There was, of course, a well-known strong ceramic signal in the region that had been linked to Lapita before the finding of these recent sites. Such

833

a signal is as yet unknown from most of the Solomons. The pottery reported from the Bogi 1 site and presented at the recent Samoa Lapita conference is clearly late Lapita, and the open dentate designs are like those of the late Lapita from the Western Solomons. The earliest Lapita component is described as contemporaneous with early Lapita expansion by the excavators and is dated to 2900 BP. My calibration of their early date (WK27077), using the SHCAL04 curve as in my paper, gives a 1j range of 2860–2775 BP, which makes it slightly earlier than the earliest late Lapita date from the Western Solomons. I would still see this as part of a late Lapita expansion into the Solomons and along the coast of New Guinea. I urge my colleagues in New Guinea to look for intertidal sites. Recent research on the exotic quartz-calcite temper found in the Western Solomons shows that it is found in more sites, indicating considerable interaction across the Coral Sea into areas where coastal sands derived from granite are found. Recent as-yet-unpublished work by Dickinson (personal communication, 2011) appears to finally resolve the source of the exotic temper, and I await its publication with interest. Has there been enough research in the Solomons to make a strong argument for absence? Clearly, for many there has not been, and I for one know very well how hard it is to survey and test in these extreme conditions and how easy it is to walk by even stone features covered in thick jungle. The general rule is this: if the undergrowth looks impenetrable, then it probably hides something of interest, just as perfect crocodile habitat probably contains an intertidal site. Kirch reports that the Southeast Solomons Culture History project did not expend as much effort in its survey of Near Oceania as it did in Temotu; however, I would suggest that the tiny island of Santa Ana, located on the eastern margin of Near Oceania, which was extensively excavated by Davenport, then by Green, and finally by Swadling working in open sites near modern villages, might have experienced more effort per area than anywhere in Temotu. I have recently dated small charcoal samples collected by Green from Feru II on Santa Ana and associated with the small samples of very-poor-quality plain ceramic. The oldest returned a calibrated 1j range of 2752– 2726 BP, putting this material at the end of the Lapita period and providing another site that can be added to the list of Solomons sites extending into the Lapita period that Spriggs requires for a suitable test of absence. Having conducted fieldwork on Santa Ana and most recently visited Santa Cruz, I am still struck with the differences in density of sites. Numerous new ceramic sites were reported to me when I was circumnavigating Santa Cruz and visiting SE-SZ-8 in December 2010. It should not be forgotten that the first Lapita ceramics found by Roger Green on Santa Cruz were shown to him by schoolchildren and were not the result of intensive survey and test-pitting. It may be that the thin post-Lapita volcanic ash has preserved sites in the Reef/Santa Cruz Group. Commentators suggest that sites are buried in the main Solomons and can be found only by extensive test-pitting. In many areas where we have worked in the Solomons, coastal


834


soils are very thin, and it may be that ceramics are not preserved as a result of lack of burial in this extreme weathering environment. I have argued that simple geotectonic models cannot explain the age and distribution of intertidal sites; however, more work is needed on the geotectonic history of the region. Ongoing work by Fred Taylor (personal communication, 2011) and his team from the University of Texas will provide much more information on that topic in the near future, as will analysis of pollen cores collected by John Dodson from Vella Lavella and Ghizo as part of a project run by Richard Walter and myself in that region. Similarly, pollen cores collected by Matthew Prebble (Australian National University) during my 2009 fieldwork on Santa Ana should inform us on the age of garden activity on that island. A number of commentators have been critical of my attempts, using linguistics and biology, to find a source region for Reef/Santa Cruz Lapita populations in the Bismarcks. As we would all agree, linking biology, linguistics, and archaeology in Near Oceania is very problematic. The simple linkage of Lapita with Austronesian or some Lapita genetic signal in that region is an assumption; however, we have much better luck making those linkages in Remote Oceania and then attempting to match back into the Bismarcks. Such matching does not have to assume that Lapita in the Bismarcks is a monothetic cultural, linguistic, and biological entity. Not all Lapita people in the Bismarcks need to have spoken Austronesian, and not all Austronesian speakers need to have made Lapita pottery, including possibly the earliest Austronesian occupants of the main Solomons. Although Ross and Næss (2007:471) note many linguistic similarities between Temotu and Mussau and describe Mussau as the closest to the Temotu languages, they do not include them in the same subgroup. I take Ross’s point that because they are first-order divisions of Oceanic, Mussau (St. Matthias Group) cannot necessarily be the origin of both groups. I have not repeated nor meant to argue, in this paper, the tentative suggestion made in Sheppard and Walter (2006) that the Southeast Solomonic languages might originate from back-movement out of Temotu. As a first-order division of Oceanic, the Temotu languages cannot be closely related to those of the Southeast Solomons, a point that supports the leapfrog hypothesis. Of course, we are still left with explaining the puzzling position and origin of the conservative Southeast Solomonic languages. Biological data might suggest a New Britain source for Santa Cruz populations, a proposal that is supported by the dominance of Talasea obsidian in the Reef/Santa Cruz Group. The early sites in Mussau have very high percentages of Admiralty Islands obsidian. Sand and Matisoo-Smith critique the use of modern linguistic and biological data. Sand states that there can be no simple relationships between modern language distributions and early prehistory because people move around. As I noted in the paper, such is most probably the case in the Bismarcks, given volcanic activity, and it may be very hard to make detailed connections to patterns existing 3,200 years ago. However, for both the linguistic and biological data there

are strong geographical patterns at the middle to large scale that demand explanation, not a randomized me´lange of language and biology. Some of these, such as the pattern formed by the movement out into Remote Oceania, clearly reflect events occurring 3,200 years ago. Certainly, more and better biological data are required, and in the best of all possible worlds it would be ancient samples, not modern. However, modern data will always be much more common, and it, too, is deserving of study and explanation, although the details of history may be harder to tease out. Finally, and as suggested by Sand, any direct archaeological light on the subject of Reef/ Santa Cruz Lapita origins will have to come from detailed comparative analysis of ceramics, as this is ultimately an archaeological issue. We never dig up language, and it is very unlikely that we will ever have sufficient ancient DNA or isotopic data to provide more than a general picture. —Peter J. Sheppard

References Cited Addison, David J., and Elizabeth Matisoo-Smith. 2010. Rethinking Polynesians origins: a West-Polynesia Triple-I model. Archaeology in Oceania 45(1):1–12. Allen, Jim. 1984. In search of the Lapita homeland: reconstructing the prehistory of the Bismarck Archipelago. Journal of Pacific History 19(4):186–201. Ambrose, Wallace R. 1991. Lapita or not Lapita: the case of the Manus pots. In Report of the Lapita Homeland Project. Jim Allen and Chris Gosden, eds. Pp. 103–112. Occasional Papers in Prehistory 20. Canberra: Department of Prehistory, Research School of Pacific Studies, Australian National University. Ammerman, Albert J., and Luca Cavalli-Sforza. 1973. A population model for the diffusion of early farming in Europe. In The explanation of cultural change. Colin Renfrew, ed. Pp. 343–357. London: Duckworth. Anderson, Atholl. 2003. Initial human dispersal in Remote Oceania: pattern and explanation. In Pacific archaeology: assessments and prospects. Christophe Sand, ed. Pp. 71–84. Les Cahiers de l’archeólogie en Nouvelle-Cale´donie 15. Noumeá, New Caledonia: De´partement Archeólogie, Service des Museés et du Patrimoine. ———. 2008. Traditionalism, interaction, and long-distance seafaring in Polynesia. Journal of Island and Coastal Archaeology 3(2): 240–250. [DVB] Anderson, Atholl, John Chappell, Michael Gagan, and Richard Grove. 2006. Prehistoric maritime migration in the Pacific islands: an hypothesis of ENSO forcing. The Holocene 16(1):1–6. Anson, Dimitri. 1986. Lapita pottery of the Bismarck Archipelago and its affinities. Archaeology in Oceania 21(3):157–165. [SC] Anthony, David W. 1990. Migration in archaeology: the baby and the bathwater. American Anthropologist 92(4):895–914. Bar-Yosef, Ofer. 2002. The Natufian culture and the early Neolithic: social and economic trends in southeastern Asia. In Examining the farming/language dispersal hypothesis. Peter Bellwood and Colin Renfrew, eds. Pp. 113–126. Cambridge: McDonald Institute for Archaeological Research. Bathgate, Murray A. 1985. Movement processes from precontact to contemporary times: the Ndi-Nggai, west Guadalcanal, Solomon Islands. In Circulation in population movement: substance and concepts from the Melanesia case. Murray Chapman and R. Mansell Prothero, eds. Pp. 83–118. London: Routledge & Kegan Paul.


Sheppard Lapita Colonization of Remote Oceania Beavan Athfield, Nancy, Roger C. Green, Jacqueline Craig, Bruce McFadgen, and Simon Bickler. 2008. Influence of marine sources on 14C ages: isotopic data from Watom Island, Papua New Guinea inhumations and pig teeth in light of new dietary standards. Journal of the Royal Society of New Zealand 38(1):1–23. Bedford, Stuart. 2006. The Pacific’s earliest painted pottery: an added layer of intrigue to the Lapita debate and beyond. Antiquity 80(309):544–557. ———. 2007. Crucial first steps into Remote Oceania. In From Southeast Asia to the Pacific: archaeological perspectives on the Austronesian expansion and the Lapita cultural complex. Scarlett Chiu and Christophe Sand, eds. Pp. 185–201. Taipei: Centre for Archaeological Research, Academica Sinica. Bedford, Stuart, and Matthew Spriggs. 2008. Northern Vanuatu as a Pacific crossroads: the archaeology of discovery, interaction, and the emergence of the “ethnographic present.” Asian Perspectives 47(1):95–120. Bedford, Stuart, Matthew Spriggs, and Ralph Regenvanu. 2006. The Teouma Lapita site and the early human settlement of the Pacific islands. Antiquity 80(310):812–828. Beebe, Nigel W., Robert D. Cooper, Desmond H. Foley, and John T. Ellis. 2000. Populations of the south-west Pacific malaria vector Anopheles farauti s.s. revealed by ribosomal DNA transcribed spacer polymorphisms. Heredity 84(2):244–253. Bellwood, Peter. 1978. Man’s conquest of the Pacific: the prehistory of Southeast Asia and Oceania. Auckland: Collins. ———. 1991. The Austronesian dispersal and the origin of languages. Scientific American 265(1):88–93. ———. 2011. Holocene population history in the Pacific region as a model for world-wide food producer dispersals. Current Anthropology 52(suppl. 4):S363–S378. [PB] Bellwood, Peter, Geoffrey Chambers, Malcolm Ross, and Hsiao-chun Hung. 2011. Are “cultures” inherited? multidisciplinary perspectives on the origins and migrations of Austronesian-speaking peoples prior to 1000 BC. In Investigating archaeological cultures: material culture, variability, and transmission. Benjamin Roberts and Marc Vander Linden, eds. Pp. 321–354. Dordrecht: Springer. [PB] Bentley, R. Alexander, Hallie R. Buckley, Matthew Spriggs, Stuart Bedford, Chris J. Ottley, Geoff M. Nowell, Colin G. Macpherson, and D. Graham Pearson. 2007. Lapita migrants in the Pacific’s oldest cemetery: isotopic analysis at Teouma, Vanuatu. American Antiquity 72(4):645–656. Blust, Robert. 2005. Untitled review of The Oceanic Languages. Oceanic Linguistics 44(2):544–558. ———. 2008. Remote Melanesia: one history or two? an addendum to Donohue and Denham. Oceanic Linguistics 47(2):445–459. Breman, Joel G., and Cherice N. Holloway. 2007. Malaria surveillance counts. American Journal of Tropical Medicine and Hygiene 77(6 suppl):36–47. Buckley, Hallie R. 2005. “The predators within”: investigating the relationship between malaria and health in the prehistoric Pacific islands. In Bioarchaeology of Southeast Asia. Marc Oxenham and Nancy Tales, eds. Pp. 309–332. Cambridge: Cambridge University Press. Burley, David V. 1999. Lapita settlement to the east: new data and changing perspectives from Ha’apai (Tonga) prehistory. In The Pacific from 5000 to 2000 BP: colonisation and transformations. Ian Lilley and Jean-Christophe Galipaud, eds. Pp. 189–200. Paris: Institut de Recherche pour le De´veloppement. Burley, David V., and Sean P. Connaughton. 2007. First Lapita settlement and its chronology in Vava’u, Kingdom of Tonga. Radiocarbon 49(1):131–137. Burley, David V., and William R. Dickinson. 2001. Origin and significance of a founding settlement in Polynesia. Proceedings of the National Academy of Sciences of the USA 98(20):11829–11831. ———. 2010. Among Polynesia’s first pots. Journal of Archaeological Science 37(5):1020–1026.

835 Burley, David V., Peter J. Sheppard, and Maia Simonin. 2011. Tongan and Samoan volcanic glass: XRF analysis and implications for constructs of ancestral Polynesian society. Journal of Archaeological Science 38(10):2625–2632. [JET] Chiu, Scarlett. 2005. Meanings of a Lapita face: materialized social memory in ancient house societies. Taiwan Journal of Anthropology 3(1):1–47. [SC, JS] ———. 2007. Detailed analysis of Lapita face motifs: case studies from Reef/Santa Cruz Lapita sites and New Caledonia Lapita Site 13A. In Oceanic explorations: Lapita and western Pacific settlement. Stuart Bedford, Christophe Sand, and Sean P. Connaughton, eds. Pp. 241–264. Terra Australis 26. Canberra: ANU E-Press. ———. 2011. Lapita-scape: research possibilities using the digital database of Lapita pottery. People and Culture in Oceania 27:39– 63. [SC] Clark, Geoffrey, and Stuart Bedford. 2008. Friction zones in Lapita colonisation. In Island of inquiry: colonisation, seafaring and the archaeology of maritime landscapes. Geoffrey Clark, Foss Leach, and Sue O’Connor, eds. Pp. 59–73. Terra Australis 29. Canberra: ANU E-Press. Clark, Geoffrey, and Tim Murray. 2006. Decay characteristics of the eastern Lapita design system. Archaeology in Oceania 41(3):107–117. Davenport, William. 1972. Preliminary excavations on Santa Ana island, eastern Solomon Islands. Archaeology and Physical Anthropology in Oceania 7(3):165–183. Davidson, Janet, and Foss Leach. 2001. The strandlooper concept and economic naivety. In The archaeology of Lapita dispersal in Oceania: papers from the Fourth Lapita Conference, June 2000, Canberra, Australia. Geoffrey R. Clark, Atholl J. Anderson, and Tarisi Vunidilo, eds. Pp. 115–123. Terra Australis 17. Canberra: Pandanus. Davison, Kate, Pavel Dolukhanov, Graeme R. Sarson, and Anvar Shukurov. 2006. The role of waterways in the spread of the Neolithic. Journal of Archaeological Science 33(5):641–652. Deguilloux, Marie-France, Marie-He´le`ne Pemonge, Vincent Dubut, Sandrine Hughes, Catherine Ha¨nni, Lionel Chollet, Eric Conte, and Pascal Murail. 2011. Human ancient and extant mtDNA from the Gambier Islands (French Polynesia): evidence for an early Melanesian maternal contribution and new perspectives into the settlement of easternmost Polynesia. American Journal of Physical Anthropology 144(2):248–257. [LM-S] Denham, Tim, and Simon Haberle. 2008. Agricultural emergence and transformation in the Upper Wahgi Valley, Papua New Guinea, during the Holocene: theory, method and practice. The Holocene 18(3):481–496. Diamond, Jared, and Peter Bellwood. 2003. Farmers and their languages: the first expansions. Science 300(5619):597–603. Diamond, Jared M. 1988. Express train to Polynesia. Nature 336(6197):307–308. Dickinson, William R. 2001. Paleoshoreline record of relative Holocene sea levels on Pacific islands. Earth-Science Reviews 55(3–4): 191–234. ———. 2006. Temper sands in prehistoric Oceanian pottery: geotectonics, sedimentology, petrography, provenance. Special Paper 406. Boulder, CO: Geological Society of America. Dickinson, William R., and Roger C. Green. 1998. Geoarchaeological context of Holocene subsidence at the Ferry Berth Lapita site, Mulifanua, Upolu, Samoa. Geoarchaeology 13(3):239–263. [MS] Doherty, Moira W. 2007. Post-Lapita developments in the Reef–Santa Cruz Islands, southeast Solomon Islands. PhD thesis, University of Auckland. [SC] Donohue, Mark, and Tim Denham. 2008. The language of Lapita: Vanuatu and an early Papuan presence in the Pacific. Oceanic Linguistics 47(2):433–444. ———. 2010. Farming and language in Island Southeast Asia: reframing Austronesian history. Current Anthropology 51(2):223–256. Donohue, Mark, and Simon Musgrave. 2007. Typology and the lin-


836


guistic macrohistory of Island Melanesia. Oceanic Linguistics 46(2): 348–387. Dunn, Michael, Robert Foley, Stephen Levinson, Ger P. Reesink, and Angela Terrill. 2007. Statistical reasoning in the evaluation of typological diversity in Island Melanesia. Oceanic Linguistics 46(2): 388–403. Dunn, Michael, Angela Terrill, Ger Reesink, Robert A. Foley, and Stephen Levinson. 2005. Structural phylogenetics and the reconstruction of ancient language history. Science 309(5743):2072– 2075. Felgate, Matthew W. 2001. A Roviana ceramic sequence and the prehistory of Near Oceania: work in progress. In The archaeology of Lapita dispersal in Oceania: papers from the Fourth Lapita Conference, June 2000, Canberra, Australia. Geoffrey R. Clark, Atholl J. Anderson, and Tarisi Vunidilo, eds. Pp. 39–60. Terra Australis 17. Canberra: Pandanus. ———. 2003. Reading Lapita in Near Oceania: intertidal shallowwater pottery scatters, Roviana Lagoon, New Georgia, Solomon Islands. PhD thesis, University of Auckland. ———. 2007. Leap-frogging or limping? recent evidence from the Lapita littoral fringe, New Georgia, Solomon Islands. In Oceanic explorations: Lapita and western Pacific settlement. Stuart Bedford, Christophe Sand, and Sean P. Connaughton, eds. Pp. 123–140. Terra Australis 26. Canberra: ANU E-Press. Felgate, Matthew W., and William R. Dickinson. 1998. Late-Lapita and post-Lapita pottery transfers: evidence from intertidal-zone sites of Roviana Lagoon, Western Province, Solomon Islands. In Australasian connections and new directions: proceedings of the Seventh Australasian Archaeometry Conference. Martin Jones and Peter J. Sheppard, eds. Pp. 103–122. Research in Anthropology and Linguistics 5. Auckland: Department of Anthropology, University of Auckland. Fort, Joaquim. 2003. Population expansion in the western Pacific (Austronesia): a wave of advance model. Antiquity 77(297):520– 530. Fredericksen, Clayton. 1997. The maritime distribution of Bismarck Archipelago obsidian and Island Melanesian prehistory. Journal of the Polynesian Society 106(4):375–393. Friedlaender, Jonathan S., Françoise R. Friedlaender, Jason A. Hodgson, Stacy McGrath, Matthew Stoltz, George Koki, Theodore G. Schurr, and D. Andrew Merriwether. 2007a. Mitochondrial DNA variation in northern Island Melanesia. In Genes, language and culture history in the Southwest Pacific. Jonathan Scott Friedlaender, ed. Pp. 61–80. Oxford: Oxford University Press. Friedlaender, Jonathan S., Françoise R. Friedlaender, Jason A. Hodgson, Matthew Stoltz, George Koki, Gisele Horvat, Sergey Zhadanov, Theodore G. Schurr, and D. Andrew Merriwether. 2007b. Melanesian mtDNA complexity. PLoS ONE 2:e248. Friedlaender, Jonathan S., Françoise R. Friedlaender, Floyd A. Reed, Kenneth K. Kidd, Judith R. Kidd, Geoffrey K. Chambers, Rodney A. Lea, et al. 2008. The genetic structure of Pacific Islanders. PLoS Genetics 4:e19. Friedlaender, Jonathan S., Fred Gentz, Françoise R. Friedlaender, Frederika Kaestle, Theodore G. Schurr, George Koki, Charles S. Mgone, et al. 2005. Mitchondrial genetic diversity and its determinants in Island Melanesia. In Papuan pasts: cultural, linguistic and biological histories of Papuan-speaking peoples. Andrew Pawley, Robert Attenborough, Jack Golson, and Robin Hide, eds. Pp. 693– 716. Pacific Linguistics 572. Canberra: Australian National University. [SC] Fullagar, Richard, Judith Field, Tim Denham, and Carol Lentfer. 2006. Early and mid Holocene tool-use and processing of taro (Colocasia esculenta), yam (Dioscorea sp.) and other plants at Kuk Swamp in the highlands of Papua New Guinea. Journal of Archaeological Science 33(5):595–614. Galipaud, Jean-Christophe, and Mary Clare Swete Kelly. 2007a.

Makue´ (Aore Island, Santo, Vanuatu): a new Lapita site in the ambit of New Britain obsidian distribution. In Oceanic explorations: Lapita and western Pacific settlement. Stuart Bedford, Christophe Sand, and Sean P. Connaughton, eds. Pp. 151–162. Terra Australis 26. Canberra: ANU E-Press. ———. 2007b. New evidence relating to the transport of obsidian from New Britain to Vanuatu. Antiquity 81(312). http:// antiquity.ac.uk/projgall/swete/index.html. Accessed June 2, 2011. [JS] Gosden, Chris. 1992. Production systems and the colonisation of the Western Pacific. World Archaeology 24(1):55–69. [RT] Gosden, Chris, and John Webb. 1994. The creation of a Papua New Guinean landscape: archaeological and geomorphological evidence. Journal of Field Archaeology 21(1):29–51. Gray, Russell D., Alexei J. Drummond, and Simon J. Greenhill. 2009. Language phylogenies reveal expansion pulses and pauses in Pacific settlement. Science 323(5913):479–483. Green, Roger C. 1978. New sites with Lapita pottery and their implications for understanding the settlement of the western Pacific. Working Papers in Anthropology, Archaeology, Linguistics, and Maori Studies 51. Auckland: University of Auckland. [SC] ———. 1979a. Early Lapita art from Polynesia and Island Melanesia: continuities in ceramic, barkcloth and tattoo decorations. In Exploring the visual art of Oceania: Australia, Melanesia, Micronesia, and Polynesia. Sidney M. Mead, ed. Pp. 13–31. Honolulu: University of Hawaii Press. ———. 1979b. Lapita. In Prehistory of Polynesia. Jesse D. Jennings, ed. Pp. 27–60. Cambridge, MA: Harvard University Press. ———. 1987. Obsidian results from the Lapita sites of the Reef/ Santa Cruz Islands. In Archaeometry: further Australasian studies. W. R. Ambrose and J. M. J. Mummery, eds. Pp. 239–249. Canberra: Australian National University. ———. 1988. Those mysterious mounds are for the birds. Archaeology in New Zealand 31:153–158. ———. 1991a. Lapita cultural complex: current evidence and proposed models. Bulletin of the Indo-Pacific Prehistory Association 11: 295–305. ———. 1991b. Near and Remote Oceania: disestablishing “Melanesia” in culture history. In Man and a half: essays in Pacific anthropology and ethnobiology in honour of Ralph Bulmer. Andrew Pawley, ed. Pp. 491–502. Auckland: Polynesian Society. ———. 1992. Definitions of the Lapita cultural complex and its nonceramic component. In Poterie Lapita and peuplement: actes du Colloque Lapita, Noumeá, Nouvelle Cale´donie, Janvier 1992. JeanChristophe Galipaud, ed. Pp. 7–20. Noumeá: ORSTROM. ———. 2003. The Lapita horizon and traditions: signature for one set of oceanic migrations. In Pacific archaeology: assessments and prospects. Christophe Sand, ed. Pp. 95–120. Les Cahiers de l’archeólogie en Nouvelle-Cale´donie 15. Noumeá, New Caledonia: De´partement Archeólogie, Service des Museés et du Patrimoine. ———. 2010. The outer eastern islands of the Solomons, a puzzle for the holistic approach to the anthropology of history. In A journey through Austronesian and Papuan linguistic and cultural space: papers in honour of Andrew Pawley. John Bowden, Nikolaus P. Himmelmann, and Malcolm Ross, eds. Pp. 207–223. Pacific Linguistics 615. Canberra: Pacific Linguistics. [SC] Green, Roger 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, Roger C., and M. M. Cresswell, eds. 1976. Southeast Solomon Islands cultural history: a preliminary report. Bulletin 11. Wellington: Royal Society of New Zealand. Green, Roger C., and Patrick V. Kirch. 1997. Lapita exchange systems and their Polynesian transformations: seeking explanatory models. In Prehistoric long-distance interaction in Oceania: an interdisciplinary approach. Marshall I. Weisler, ed. Pp. 19–37. New Zealand


Sheppard Lapita Colonization of Remote Oceania Archaeological Association Monograph A21. Auckland: New Zealand Archaeological Association. [SC] Green, Roger C., Martin Jones, and Peter Sheppard. 2008. The reconstructed environments and absolute dating of SE-SZ-8 Lapita site on Nendo, Santa Cruz, Solomon Islands. Archaeology in Oceania 43(2):49–61. Groube, L. M. 1971. Tonga, Lapita pottery and Polynesian origins. Journal of the Polynesian Society 80(3):278–314. Hill, Catherine, Pedro Soares, Maru Mormina, Vincent Macaulay, Dougie Clarke, Petya B. Blumbach, Matthieu Vizuete-Forster, et al. 2007. A mitochondrial stratigraphy for Island Southeast Asia. American Journal of Human Genetics 80(1):29–43. Horrocks, Mark, and Stuart Bedford. 2005. Microfossil analysis of Lapita deposits in Vanuatu reveals introduced Araceae (aroids). Archaeology in Oceania 40(2):67–74. Horrocks, Mark, Stuart Bedford, and Matthew Spriggs. 2009. A short note on banana (Musa) phytoliths in Lapita, immediately postLapita and modern period archaeological deposits from Vanuatu. Journal of Archaeological Science 36(9):2048–2054. Horrocks, Mark, and Patrick D. Nunn. 2007. Evidence for introduced taro (Colocasia esculenta) and lesser yam (Dioscorea esculenta) in Lapita-era (c. 3050–2500 cal. yr BP) deposits from Bourewa, southwest Viti Levu Island, Fiji. Journal of Archaeological Science 34(5): 739–748. Hung, Hsiao-chun, Mike T. Carson, Peter Bellwood, Fredeliza Z. Campos, Philip J. Piper, Eusebio Dizon, Mary Jane Louise A. Bolunia, Marc Oxenham, and Zhang Chi. 2011. The first settlement of Remote Oceania: Luzon to the Marianas. Antiquity 85(329): 909–926. [PB] Hunley, Keith, Michael Dunn, Eva Lindstro¨m, Ger Reesink, Angela Terrill, Heather Norton, Laura Scheinfeldt, et al. 2007. Inferring prehistory from genetic, linguistic, and geographic variation. In Genes, language, and culture history in the Southwest Pacific. Jonathan Scott Friedlaender, ed. Pp. 141–154. Oxford: Oxford University Press. Hurles, Matthew E., Elizabeth Matisoo-Smith, Russell D. Gray, and David Penny. 2003. Untangling Oceanic settlement: the edge of the knowable. Trends in Ecology & Evolution 18(10):531–540. Irwin, Geoffrey. 1981. How Lapita lost its pots: the question of continuity in the colonization of Polynesia. Journal of the Polynesian Society 90(4):481–494. [GI, AP] ———. 1992. The prehistoric exploration and colonisation of the Pacific. Cambridge: Cambridge University Press. ———. 2006. Voyaging and settlement. In Vaka Moana: voyages of the ancestors. Kerry R. Howe, ed. Pp. 54–91. Auckland: David Bateman. [SC] ———. 2008. Pacific seascapes, canoe performance, and a review of Lapita voyaging with regard to theories of migration. Asian Perspectives 47(1):12–27. Irwin, Geoffrey, and Simon Holdaway. 1996. Colonisation, trade and exchange: from Papua to Lapita. In Oceanic culture history: essays in honour of Roger Green. Janet M. Davidson, Geoffrey Irwin, Foss Leach, Andrew Pawley, and Dorothy Brown, eds. Pp. 225–235. Dunedin: New Zealand Journal of Archaeology. Jones, Martin, Fiona Petchey, Roger Green, Peter Sheppard, and Matthew Phelan. 2007. The marine DR for Nenumbo (Solomon Islands): a case study in calculating reservoir offsets from paired sample data. Radiocarbon 49(1):95–102. Kayser, Manfred. 2010. The human genetic history of Oceania: near and remote views of dispersal. Current Biology 20(4):R194–R201. Kayser, Manfred, Ying Choi, Mannis van Oven, Stefano Mona, Silke Brauer, Ronald J. Trent, Dagwin Suarkia, Wulf Schiefenho¨vel, and Mark Stoneking. 2008a. The impact of the Austronesian expansion: evidence from mtDNA and Y chromosome diversity in the Admiralty Islands of Melanesia. Molecular Biology and Evolution 25(7):1362–1374.

837 Kayser, Manfred, Oscar Lao, Kathrin Saar, Silke Brauer, Xingyu Wang, Peter Nu¨rnberg, Ronald J. Trent, and Mark Stoneking. 2008b. Genome-wide analysis indicates more Asian than Melanesian ancestry of Polynesians. American Journal of Human Genetics 82(1): 194–198. [DVB] Kennett, Douglas J., Atholl Anderson, and Bruce Winterhalder. 2006. The ideal free distribution, food production, and the colonization of Oceania. In Behavioral ecology and the transition to agriculture. Douglas J. Kennett and Bruce Winterhalder, eds. Pp. 265–288. Berkeley: University of California Press. Kirch, Patrick V. 1987. Lapita and Oceanic cultural origins: excavations in the Mussau Islands, Bismarck Archipelago, 1985. Journal of Field Archaeology 14(2):163–180. ———. 1988a. Long-distance exchange and island colonization: the Lapita case. Norwegian Archaeological Review 21(2):103–117. ———. 1988b. The Talepakemalai Lapita site and Oceanic prehistory. National Geographic Research 4(3):328–342. [SC] ———. 1990. Specialization and exchange in the Lapita complex of Oceania. Asian Perspectives 29(2):117–133. [GRS] ———. 1997. The Lapita peoples: ancestors of the Oceanic world. Cambridge: Blackwell. ———. 2001a. A radiocarbon chronology for the Mussau Islands. In Lapita and its transformations in Near Oceania: archaeological investigations in the Mussau Islands, Papua New Guinea, 1985– 1988. Volume I. Introduction, stratigraphy, chronology. Patrick V. Kirch, ed. Pp. 196–222. Contributions of the University of California Archaeological Research Facility 59. Berkeley: University of California. ———. 2001b. Three Lapita villages: excavations at Talepakemalai (ECA), Etakosarai (ECB), and Etapakakengaroasa (EHB), Eloaua and Emananus Islands. In Lapita and its transformations in Near Oceania: archaeological investigations in the Mussau Islands, Papua New Guinea, 1985–1988. Volume I. Introduction, stratigraphy, chronology. Patrick V. Kirch, ed. Pp. 68–145. Contributions of the University of California Archaeological Research Facility 59. Berkeley: University of California. Kirch, Patrick V., and Roger C. Green. 2001. Hawaiki: ancestral Polynesia: an essay in historical anthropology. Cambridge: Cambridge University Press. [AP, JS] Kononenko, Nina. 2007. The contribution of use-wear/residue studies of obsidian artefacts for understanding changes in settlement and subsistence patterns in West New Britain, Papua New Guinea. Bulletin of the Indo-Pacific Prehistory Association 27:135–143. [RT] ———. 2011. Experimental and archaeological studies of use-wear and residues on obsidian artefacts from Papua New Guinea. Technical Reports of the Australian Museum Online 21. Sydney: Australian Museum. http://australianmuseum.net.au/journal/Kononenko2010-Tech-Rep-Aust-Mus-Online-21-1244. [RT] Leach, Foss, and Janet Davidson. 2008. Archaeology on Taumako: a Polynesian outlier in the eastern Solomon Islands. Dunedin: New Zealand Archaeological Association Special Publication. Leavesley, Matthew G., and John Chappell. 2004. Buang Merabak: additional early radiocarbon evidence of the colonisation of the Bismarck Archipelago, Papua New Guinea. Antiquity 78(301). http: //antiquity.ac.uk/projgall/leavesley/index.html. Lilley, Ian. 2004. Trade and culture history across the Vitiaz Strait, Papua New Guinea: the emerging post-Lapita coastal sequence. Records of the Australian Museum Supplement 29:89–96. Lum, J. Koji, Akira Kaneko, George Taleo, Morris Amos, and Dana M. Reiff. 2007. Genetic diversity and gene flow of humans, Plasmodium falciparum, and Anopheles farauti s.s. of Vanuatu: inferred malaria dispersal and implications for malaria control. Acta Tropica 103(2):102–107. Lynch, John, Malcolm Ross, and Terry Crowley. 2002. The Oceanic languages. Richmond: Curzon. Machida, Hiroshi, Russell J. Blong, Jim Specht, Hiroshi Moriwaki,


838


Robin Torrence, Yukio Hayakawa, Benjamin Talai, David Lolok, and Colin 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. Mann, Paul, Frederick W. Taylor, Martin B. Lagoe, Andrew Quarles, and George 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. Matisoo-Smith, Elizabeth. 2007. Animal translocations, genetic variation, and the human settlement of the Pacific. In Genes, language, and culture history in the Southwest Pacific. Jonathan Scott Friedlaender, ed. Pp. 157–170. Oxford: Oxford University Press. Matisoo-Smith, Elizabeth, and Judith Robins. 2009. Mitochondrial DNA evidence for the spread of Pacific rats through Oceania. Biological Invasions 11(7):1521–1527. [JET] McCormac, F. G., Alan G. Hogg, Paul G. Blackwell, Caitlin E. Buck, Thomas F. G. Higham, and Paula J. Reimer. 2004. SHCAL04 Southern Hemisphere calibration, 0–11.0 cal kyr BP. Radiocarbon 46(3):1087–1092. McNiven, Ian J., William R. Dickinson, Bruno David, Marshall Weisler, Friedrich Von Gnielinski, Melissa Carter, and Ugo Zoppi. 2006. Mask Cave: red-slipped pottery and the Australian-Papuan settlement of Zenadh Kes (Torres Strait). Archaeology in Oceania 41(2): 49–81. McNiven, Ian J., Bruno David, Thomas Richards, Ken Aplin, Brit Asmussen, Jerome Mialanes, Matthew Leavesley, Patrick Faulkner, and Sean Ulm. 2011. New direction in human colonisation of the Pacific: Lapita settlement of south coast New Guinea. Australian Archaeology 72:1–6. Mitrovica, Jerry X., and W. R. Peltier. 1991. On postglacial geoid subsidence over the equatorial oceans. Journal of Geophysical Research–Solid Earth 96(B12):20053–20071. Mona, Stefano, Katharina E. Grunz, Silke Brauer, Brigitte Pakendorf, Loredana Castrı`, Herawati Sudoyo, Sangkot Marzuki, et al. 2009. Genetic admixture history of eastern Indonesia as revealed by Ychromosome and mitochondrial DNA analysis. Molecular Biology and Evolution 26(8):1865–1877. Moore, John H. 2001. Evaluating five models of human colonization. American Anthropologist 103(2):395–408. Mu¨hlhaüsler, Peter, Tom Dutton, Even Hovdhaugen, Jeff Williams, and Stephen A. Wurm. 1996. Precolonial patterns of intercultural communication in the Pacific Islands. In Atlas of languages of intercultural communication in the Pacific, Asia and the Americas, vol. II.1. Stephen A. Wurm, Peter Mu¨hlhaüsler, and Darrell T. Tryon, eds. Pp. 401–437. Trends in Linguistics: Documentation 13. Berlin: Mouton de Gruyter. Næss, A˚shild, and Brenda H. Boerger. 2008. Reefs–Santa Cruz as Oceanic: evidence from the verb complex. Oceanic Linguistics 47(1):185–212. [TT] Negishi, Yo, and Rintaro Ono. 2009. Kasasinabwana shell midden: the prehistoric ceramic sequence of Wari Island in the Massim, eastern Papua New Guinea. People and Culture in Oceania 25:23–52. Nevermann, Hans. 1933. St Matthias-Gruppe: Ergebnisse der Su¨dseeExpedition 1908–1910. Hamburg: Friederichsen, de Gruyter. [GRS] Noury, Arnaud. 2005. Le reflet de l’aˆme Lapita. Versailles: privately published. [JS] Nunn, Patrick. 2007. Echoes from a distance: research into the Lapita occupation of the Rove Peninsula, southwest Viti Levu, Fiji. In Oceanic explorations: Lapita and western Pacific settlement. Stuart Bedford, Christophe Sand, and Sean P. Connaughton, eds. Pp. 163– 176. Terra Australis 26. Canberra: ANU E-Press. O’Connell, James F., and Jim Allen. 2007. Pre-LGM Sahul (Pleistocene Australia-New Guinea) and the archaeology of early modern humans. In Rethinking the human revolution: new behavioural and biological perspectives on the origin and dispersal of modern

humans. Paul Mellars, Katie Boyle, Ofer Bar-Yosef, and Chris Stringer, eds. Pp. 395–410. McDonald Institute Monographs. Oxford: McDonald Institute for Archaeological Research Oppenheimer, Stephen. 2004. The “express train from Taiwan to Polynesia”: on the congruence of proxy lines of evidence. World Archaeology 36(4):591–600. Parkinson, Richard. 1907. Dreißig Jahre in der Su¨dsee. Stuttgart: Strecker & Schro¨der. [GRS] Pavlides, Christina, and Chris Gosden. 1994. 35,000-year old sites in the rainforests of West New Britain, Papua New Guinea. Antiquity 68(260):604–610. Pawley, Andrew. 1981. Melanesian diversity and Polynesian homogeneity: a unified explanation for language. In Studies in Pacific languages and cultures in honour of Bruce Biggs. Jim Hollyman and Andrew Pawley, eds. Pp. 269–309. Auckland: Linguistic Society of New Zealand. ———. 2006. Explaining the aberrant Austronesian languages of southeast Melanesia: 150 years of debate. Journal of the Polynesian Society 115(3):215–258. ———. 2007. The origins of early Lapita culture: the testimony of historical linguistics. In Oceanic explorations: Lapita and western Pacific settlement. Stuart Bedford, Christophe Sand, and Sean P. Connaughton, eds. Pp. 17–49. Terra Australis 26. Canberra: ANU E-Press. [AP] ———. 2009. The role of the Solomon Islands in the first settlement of Remote Oceania: bringing linguistic evidence to an archaeological debate. In Austronesian historical linguistics and culture history: a festschrift for Robert Blust. Alexander Adelaar and Andrew Pawley, eds. Pp. 515–540. Pacific Linguistics 601. Canberra: Pacific Linguistics. [MR] Pawley, Andrew, and Roger C. Green. 1973. Dating the dispersal of the Oceanic languages. Oceanic Linguistics 12:1–67. [JS] Pawley, Andrew, and Malcolm Ross. 1993. Austronesian historical linguistics and culture history. Annual Review of Anthropology 22: 425–459. Petchey, Fiona, Atholl Anderson, Albert Zondervan, Sean Ulm, and Alan Hogg. 2008. New marine DR values for the South Pacific subtropical Gyre region. Radiocarbon 50(3):373–397. Petrie, Cameron A., and Robin 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. Reepmeyer, Christian, and Geoffrey Clark. 2009. Post-colonization interaction between Vanuatu and Fiji reconsidered: the re-analysis of obsidian from Lakeba Island, Fiji. Archaeometry 52(1):1–18. Ricaut, François-Xavier, Timothy Thomas, Maru Mormina, Murray P. Cox, Maggie Bellatti, Robert A. Foley, and Marta Mirazon-Lahr. 2010. Ancient Solomon Islands mtDNA: assessing Holocene settlement and the impact of European contact. Journal of Archaeological Science 37(6):1161–1170. Richards, Martin, Stephen Oppenheimer, and Bryan Sykes. 1998. mtDNA suggests Polynesian origins in eastern Indonesia. American Journal of Human Genetics 63(4):1234–1236. [DVB] Roe, David. 1993. Prehistory without pots: prehistoric settlement and economy of north-west Guadalcanal, Solomon Islands. PhD thesis, Australian National University. Ross, Malcolm. 1988. Proto Oceanic and the Austronesian languages of western Melanesia. Pacific Linguistics C-98. Canberra: Australian National University. ———. 1989. Early Oceanic linguistic prehistory: a reassessment. Journal of Pacific History 24(2):135–149. ———. 1997. Social networks and kinds of speech-community event. In Archaeology and language I: theoretical and methodological orientations. Roger Blench and Matthew Spriggs, eds. Pp. 209–261. London: Routledge. ———. 2001. Is there an East Papuan phylum? evidence from pro-


Sheppard Lapita Colonization of Remote Oceania nouns. In The boy from Bundaberg: studies in Melanesian linguistics in honour of Tom Dutton. Andrew Pawley, Malcolm Ross, and Darrell Tryon, eds. Pp. 301–321. Pacific Linguistics 514. Canberra: Pacific Linguistics. ———. 2010. Lexical history in the Northwest Solomonic languages: evidence for two waves of settlement in Bougainville and the northwestern Solomons. In A journey through Austronesian and Papuan linguistic and cultural space: papers in honour of Andrew Pawley. John Bowden, Nikolaus P. Himmelmann, and Malcolm Ross, eds. Pp. 245–270. Pacific Linguistics 615. Canberra: Pacific Linguistics. [MR] ¨ iwoo, Ross, Malcolm, and A˚shild Næss. 2007. An Oceanic origin for A the language of the Reef Islands? Oceanic Linguistics 46(2):456– 498. Ross, Malcolm, Andrew Pawley, and Meredith Osmond. 2008. Introduction. In The lexicon of Proto Oceanic: the culture and environment of ancestral Oceanic society. Vol. 3, Plants. Malcolm Ross, Andrew Pawley, and Meredith Osmond, eds, Pp. 1–24. Pacific Linguistics 599. Canberra: Pacific Linguistics. [MR] ———, eds. 1998–2011. The lexicon of Proto Oceanic: the culture and environment of ancestral Oceanic society. Vol. 1, Material culture (1998); vol. 2, The physical environment (2003); vol. 3 Plants (2008); vol. 4, Animals (2011). Pacific Linguistics 599. Canberra: Pacific Linguistics. [AP] Sand, Christophe. 1995. Le temps d’avant: la pre´histoire de la NouvelleCale´donie. Paris: L’Harmattan. ———. 2000. La datation du premier peuplement de Wallis et Futuna: contribution a` la de´finition de la chronologie Lapita en Polyne´sie occidentale. Journal de la Socie´te´ des Oceánistes 111:165–172. ———. 2007. The eastern frontier: Lapita ceramics in the Fiji–West Polynesia region. In From Southeast Asia to the Pacific: archaeological perspectives on the Austronesian expansion and the Lapita cultural complex. Scarlett Chiu and Christophe Sand, eds. Pp. 214– 234. Taipei: Centre for Archaeological Research, Academia Sinica. Scheinfeldt, Laura, Françoise Friedlaender, Jonathan Friedlaender, Krista Latham, George Koki, Tatyana Karafet, Michael Hammer, and Joseph Lorenz. 2006. Unexpected NRY chromosome variation in northern Island Melanesia. Molecular Biology and Evolution 23(8):1628–1641. Sheppard, Peter, Barbara Trichereau, and Christany Milicich. 2010. Pacific obsidian sourcing by portable XRF. Archaeology in Oceania 45(1):21–30. Sheppard, Peter J. 1993. Lapita lithics: trade/exchange and technology. A view from the Reefs/Santa Cruz. Archaeology in Oceania 28(3):121–137. ———. 1996. Hard rock: archaeological implications of chert sourcing in Near and Remote Oceania. In Oceanic culture history: essays in honour of Roger Green. Janet M. Davidson, Geoffrey Irwin, Foss Leach, Andrew Pawley, and Dorothy Brown, eds. Pp. 99–115. Dunedin: New Zealand Journal of Archaeology. Sheppard, Peter J., Matthew Felgate, Kenneth Roga, John Keopo, and Richard Walter. 1999. A ceramic sequence from Roviana Lagoon (New Georgia, Solomon Island). In The Pacific from 5000 to 2000 BP: colonisation and transformations. Ian Lilley and Jean-Christophe Galipaud, eds. Pp. 313–322. Paris: Institut de Recherche pour le De´veloppement. [SB] Sheppard, Peter J., and Richard Walter. 2006. A revised model of Solomon Islands culture history. Journal of the Polynesian Society 115(1):47–76. ———. 2009. Inter-tidal late Lapita sites and geotectonics in the western Solomon Islands. In Lapita: ancestors and descendants. Peter J. Sheppard, Tim Thomas, and Glenn R. Summerhayes, eds. Pp. 73–100. Monograph 28. Auckland: New Zealand Archaeological Association. Soares, Pedro, Teresa Rito, Jean Trejaut, Maru Mormina, Catherine Hill, Emma Tinkler-Hundal, Michelle Braid, et al. 2011. Ancient

839 voyaging and Polynesian origins. American Journal of Human Genetics 88(2):239–247. [DVB, LM-S, JS, JET] Specht, Jim. 2002. Obsidian, colonising and exchange. In Fifty years in the field: essays in honour and celebration of Richard Shutler Jr.’s archaeological career. Edited by Stuart Bedford, Christophe Sand, and David Burley, pp. 37–49. Monograph 25. Auckland: New Zealand Archaeological Association. ———. 2005. Revisiting the Bismarcks: some alternative views. In Papuan pasts: cultural, linguistic and biological histories of Papuanspeaking peoples. Andrew Pawley, Robert Attenborough, Jack Golson, and Robin Hide, eds. Pp. 235–288. Pacific Linguistics 572. Canberra: Pacific Linguistics. ———. 2007. Small islands in the big picture: the formative period of Lapita in the Bismarck Archipelago. In Oceanic explorations: Lapita and western Pacific settlement. Stuart Bedford, Christophe Sand, and Sean P. Connaughton, eds. Pp. 51–70. Terra Australis 26. Canberra: ANU E-Press. Spriggs, Matthew. 1984. The Lapita cultural complex: origins, distribution, contemporaries and successors. Journal of Pacific History 19(4):202–223. [MS] ———. 1990. The changing face of Lapita: transformation of a design. In Lapita design, form and composition: proceedings of the Lapita Design Workshop, Canberra, Australia, December 1988. Matthew Spriggs, ed. Pp. 83–122. Occasional Papers in Prehistory 19. Canberra: Dept. of Prehistory, Research School of Pacific Studies, Australian National University. ———. 1997. The Island Melanesians. Oxford: Blackwell. ———. 2003. Post-Lapita evolutions in Melanesia. In Pacific archaeology: assessments and prospects. Christophe Sand, ed. Pp. 205– 212. Les Cahiers de l’archeólogie en Nouvelle-Cale´donie 15. Noumeá, New Caledonia: De´partement Archeólogie, Service des Museés et du Patrimoine. [JS] ———. 2010. Geomorphic and archaeological consequences of human arrival and agricultural expansion on Pacific Islands: a reconsideration after 30 years of debate. In Altered ecologies: fire, climate and human influence on terrestrial landscapes. Simon Haberle, Janelle Stevenson, and Matthew Prebble, eds. Pp. 239–252. Terra Australis 32. Canberra: ANU E-Press. [MS] ———. 2011. Archaeology and the Austronesian expansion: where are we now? Antiquity 85(328):510–528. [RT] Spriggs, Matthew, Roger Bird, and Wallace Ambrose. 2010. A reanalysis of the Tikopia obsidians. Archaeology in Oceania 45(1): 31–38. [JS] Summerhayes, Glenn. 2000. Lapita interaction. Terra Australis 15. Canberra: Department of Archaeology and Natural History and Centre for Archaeology, Australian National University. [JS] ———. 2004. The rocky road: the selection and transport of Admiralties obsidian to Lapita communities. Australian Archaeology 57:135–142. ———. 2007a. Island Melanesian pasts: a view from archaeology. In Genes, language, and culture history in the Southwest Pacific. Jonathan Scott Friedlaender, ed. Pp. 10–35. Oxford: Oxford University Press. ———. 2007b. The rise and transformations of Lapita in the Bismarck Archipelago. In From Southeast Asia to the Pacific: archaeological perspectives on the Austronesian expansion and the Lapita cultural complex. Scarlett Chiu and Christophe Sand, eds. Pp. 141– 169. Taipei: Centre for Archaeological Research, Academia Sinica. ———. 2010. Lapita interaction: an update. In 2009 International Symposium on Austronesian Studies. Masegseg Z. Gadu and Hsiuman Lin, eds. Pp. 11–40. Taitong, Taiwan: National Museum of Prehistory. [JS, GRS] Summerhayes, Glenn R., and Jim Allen. 1993. The transport of Mopir obsidian to Late Pleistocene New Ireland. Archaeology in Oceania 28(3):144–148. ———. 2007. Lapita writ small? revisiting the Austronesian colo-


840


nisation of the Papuan south coast. In Oceanic explorations: Lapita and western Pacific settlement. Stuart Bedford, Christophe Sand, and Sean P. Connaughton, eds. Pp. 97–122. Terra Australis 26. Canberra: ANU E-Press. Summerhayes, Glenn R., and Ian Scales. 2005. New Lapita pottery finds from Kolombangara, western Solomon Islands. Archaeology in Oceania 40(1):14–20. Swadling, Pamela. 2000. Changing marine interests and their implications for the settlement history of Santa Ana, an island in the southeast Solomon Islands. In Australian archaeologist: collected papers in honour of Jim Allen. Atholl Anderson and Tim Murray, eds. Pp. 365–371. Canberra: Coombs. Taylor, Frederick W., Richard W. Briggs, Cliff Frohlich, Abel Brown, Matt Hornbach, Alison K. Papabatu, Aron J. Meltzner, and Douglas Billy. 2008. Rupture across arc segment and plate boundaries in the 1 April 2007 Solomons earthquake. Nature Geoscience 1(4): 253–257. Taylor, Frederick. W., Paul Mann, Michael G. Bevis, R. Lawrence Edwards, Hai Cheng, Kirsten B. Cutler, Sarah C. Gray, et al. 2005. Rapid forearc uplift and subsidence caused by impinging bathymetric features: examples from the New Hebrides and Solomon arcs. Tectonics 24:TC6005. Terrell, John. 1986. Causal pathways and causal processes: Studying the evolutionary prehistory of human diversity in language, customs, and biology. Journal of Anthropological Archaeology 5(2):187– 198. [JET] ———. 1988. History as a family tree, history as an entangled bank: Constructing images and interpretations of prehistory in the South Pacific. Antiquity 62(237):642–657. ———. 2002. Tropical agroforestry, coastal lagoons, and Holocene prehistory in Greater Near Oceania. In Vegeculture in eastern Asia and Oceania. Shuji Yoshida and Peter J. Matthews, eds. Pp. 195– 216. Symposium Series 16. Osaka: Japan Centre for Area Studies, National Museum of Ethnology. ———. 2006. Human biogeography: evidence of our place in nature. Journal of Biogeography 33(12):2088–2098. [JET] ———. 2009. Return to the entangled bank: deciphering the Lapita cultural series, In Lapita: ancestors and descendants. Peter J. Sheppard, Tim Thomas, and Glenn R. Summerhayes, eds. Pp. 255– 269. Monograph 28. Auckland: New Zealand Archaeological Association. [JET] ———. 2010. Social network analysis of the genetic structure of Pacific Islanders. Annals of Human Genetics 74(3):211–232. [JET] Terrell, John E., and Esther M. Schechter. 2007. Deciphering the Lapita code: the Aitape ceramic sequence and late survival of the “Lapita face.” Cambridge Archaeological Journal 17(1):59–85. [JET] ———. 2009. The meaning and importance of the Lapita face motif. Archaeology in Oceania 44(2):45–55. [JET] Terrill, Angela. 2003. Linguistic stratigraphy in the central Solomon Islands: lexical evidence of early Papuan/Austronesian interaction. Journal of the Polynesian Society 112(4):369–401. Torrence, Robin. 2006. Starch and archaeology. In Ancient starch research. Robin Torrence and Huw Barton, eds. Pp. 17–34. Walnut Creek, CA: Left Coast. [RT] ———. 2009. The role of volcanism in the origin and spread of Lapita pottery. Paper presented at the 19th Congress of the IndoPacific Prehistory Association, Hanoi, Vietnam, November 29– December 5. [RT] ———. 2011. Finding the right question: learning from stone tools

on the Willaumez Peninsula, Papua New Guinea. Archaeology in Oceania 46(2):29–41. [RT] Torrence, Robin, and Trudy Doelman. 2007. Problems of scale: evaluating the effects of volcanic disasters on cultural change in the Willaumez Peninsula, Papua New Guinea. In Living under the shadow: the cultural impacts of volcanic eruptions. John Grattan and Robin Torrence, eds. Pp. 42–66. Walnut Creek, CA: Left Coast. [RT] Torrence, Robin, and John Grattan. 2002. The archaeology of disasters: past and future trends. In Natural disasters and culture change. Robin Torrence and John Grattan, eds. pp. 1–18. London: Routledge. [RT] Torrence, Robin, Vince Neall, and William E. Boyd. 2009a. Volcanism and historical ecology on the Willaumez Peninsula, Papua New Guinea. Pacific Science 63(4):507–535. Torrence, Robin, Vince Neall, Trudy Doelman, Ed Rhodes, Chris McKee, Hugh Davies, Roberto Bonetti, et al. 2004. Pleistocene colonisation of the Bismarck Archipelago: new evidence from West New Britain. Archaeology in Oceania 39(3):101–130. Torrence, Robin, and Pamela Swadling. 2008. Social networks and the spread of Lapita. Antiquity 82(317):600–616. Torrence, Robin, Pamela Swadling, Nina Kononenko, Wallace Ambrose, Pip Rath, and Michael D. Glascock. 2009b. Obsidian stemmed tools and mid-Holocene interaction. Asian Perspectives 48(1):118–147. [RT] Valentin, Fre´de´rique, Hallie R. Buckley, Estelle Herrscher, Rebecca Kinaston, Stuart Bedford, Matthew Spriggs, Stuart Hawkins, and Ken Neal. 2010. Lapita subsistence strategies and food consumption patterns in the community of Teouma (Efate, Vanuatu). Journal of Archaeological Science 37(8):1820–1829. Veth, Peter, Peter Hiscock, and Alan Williams. 2011. Are tulas and ENSO linked in Australia? Australian Archaeology 72:7–14. [GI] White, J. Peter. 2004. Where the wild things are: prehistoric animal translocation in the circum New Guinea Archipelago. In Voyages of discovery: the archaeology of islands. Scott M. Fitzpatrick, ed. Pp. 147–164. London: Praeger. White, J. Peter, Geoffrey Clark, and Stuart Bedford. 2000. Distribution, present and past, of Rattus praetor in the Pacific and its implications. Pacific Science 54(2):105–117. Wickler, Stephen. 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. Yen, Douglas E. 1996. Melanesian arboriculture: historical perspectives with emphasis on the genus Canarium. In South Pacific indigenous nuts. M. L. Stevens, R. Michael Bourke, and Barry R. Evans, eds. Pp. 36–44. Canberra: Australian Centre for International Agricultural Research. Zeder, Melinda A. 2008. Domestication and early agriculture in the Mediterranean Basin: origins, diffusion, and impact. Proceedings of the National Academy of Sciences of the USA 105(33):11597– 11604. Zilhaõ, Joaõ. 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 USA 98(24): 14180–14185. Zvelebil, Marek, and Malcolm Lillie. 2000. Transition to agriculture in eastern Europe. In Europe’s first farmers. T. Douglas Price, ed. Pp. 57–92. Cambridge: Cambridge University Press.
