Revisiting Rapa Nui (Easter Island) ''Ecocide''1

Revisiting Rapa Nui (Easter Island) ‘‘Ecocide’’1

Terry L. Hunt2,4 and Carl P. Lipo3

Abstract: Easter Island (Rapa Nui) has become widely known as a case of ‘‘ecocide,’’ where the ancient Polynesians recklessly destroyed their environment and, as a consequence, suffered collapse. In recent publications, both popular and academic, scholars have promoted this perspective, drawing upon archaeological evidence and offering Rapa Nui as a parable for our current global crisis. In this paper we address recent claims and outline emerging archaeological and paleoenvironmental evidence. We consider chronology, causes and consequences of deforestation, agricultural strategies, statue transport, and the evidence for ancient population size and its demise. Although deforestation and ecological catastrophe certainly unfolded over the course of the island’s prehistory, the ensuing demographic and cultural collapse followed European contact and resulted from the devastating effects of disease and slave trading. Defores- tation and contact-induced demographic collapse were separated in time and causation. Finally, we offer alternative perspectives emerging from a variety of recent research.

In his recent best-selling book, Jared Dia- ture (e.g., Brander and Taylor 1998, Dalton mond (2005) further popularized a story of and Coats 2000). ancient Easter Islanders recklessly destroying Close examination of the Rapa Nui evi- their environment and, as a direct conse- dence from archaeology, paleoenvironmental quence, destroying their civilization. Dia- work, and biological anthropology demon- mond calls this ecological self-destruction strates that although the island suffered an ‘‘ecocide’’ and offers Rapa Nui as a parable ecological catastrophe, the story of ‘‘ecocide’’ for our current global environmental crisis. may be little more than a modern myth The notion that ancient Polynesians of Rapa (Rainbird 2002, Peiser 2005, Hunt 2006, Nui committed ‘‘ecocide’’ has been widely 2007, Hunt and Lipo 2007, 2008). After pub- shared in public statements (e.g., Young lication of Collapse, many new details have 1991, Gee 2004) and academic literature come to light with continuing field research (e.g., Redman 2001, Flenley and Bahn 2002, on the island. This new research casts doubt 2007, De Monocol et al. 2005). Indeed the on previous assertions about the sequence theme of Rapa Nui’s self-destruction with de- and timing of events in the island’s prehistory forestation induced by large-scaled invest- (see Lipo and Hunt 2005, Barnes et al. 2006, ments in megalithic statuary and architecture Hunt 2006, 2007, Hunt and Lipo 2006, 2007, pervades a multidisciplinary academic litera- 2008). New details on the island’s chronology, settlement patterns, subsistence systems, and environmental change have led research- 1 Manuscript accepted 1 February 2009. ers to question ‘‘ecocide,’’ instead raising al- 2 Department of Anthropology, University of Hawai‘i ternatives to scientific and public audiences at Ma¯noa, 2424 Maile Way, Honolulu, Hawai‘i 96822. (e.g., Gibbons 2006, Hunt 2006). 3 Department of Anthropology and Institute for Inte- grated Research in Materials, Environments, and Society, In a recent essay in Science, Jared Diamond California State University at Long Beach, 1250 Bell- (2007) reiterated the account given in his flower Boulevard, Long Beach, California 90840. book Collapse and defended his claims 4 Corresponding author (e-mail: [email protected]). against new evidence that challenges a now- conventional view of ‘‘ecocide.’’ Like Dia- Pacific Science (2009), vol. 63, no. 4:601–616 mond, Flenley and Bahn (2002, 2007) have : 2009 by University of Hawai‘i Press also continued their steadfast support for an All rights reserved ecocide story of the island’s prehistory.

601 . 602 PACIFIC SCIENCE October 2009

In this paper, we respond primarily to Di- available for fires. . . . As we try to imagine amond’s (2007) recent framing of the evi- the decline of Easter’s civilization, we ask dence in defense of his views previously ourselves, ‘Why didn’t they look around, re- published (Diamond 1995, 2005). A detailed alize what they were doing, and stop before response to Flenley and Bahn (2007) can be it was too late? What were they thinking found elsewhere (Hunt and Lipo 2007). We when they cut down the last palm tree?’’’ do not question whether prehistoric deforestation transformed Rapa Nui but instead ask evaluating the evidence if this deforestation caused population collapse before European contact. We argue that No one denies that Rapa Nui was largely de- researchers cannot uncritically accept inter- forested by the time Europeans first arrived pretations made of the archaeological and in 1722. The earliest visitors remarked on historical records without evaluating the evi- the dearth of trees or wood available on the dence and weighing alternative hypotheses. island. Evidence from pollen and charcoal In doing our best to understand and explain identifications show that more than 20 woody the past, we must open our minds to possibil- taxa once grew on the island (Orliac 2000, ities that may not find obvious analogies in 2003). Field research has shown that several our contemporary world, particularly if we million giant palm trees ( Jubaea sp. or Pascha- are to learn something new (and different) lococos disperta) once dominated the vegetation from the past. Surely this is the promise and over much of the island (Mieth and Bork strength of archaeology, and science more 2004, 2006). The palms and other woody generally. taxa appear to have declined from about A.D. 1280 and disappeared by about A.D. the conventional interpretation 1650, with some remnants surviving into his- and ‘‘ecocide’’ toric times (Hunt 2007). The fact that deforestation occurred is not in question, nor is its Visiting Rapa Nui for a single day in April association with human activities (contrary to 1786, the French explorer La Pe´rouse specu- arguments for a major role of climate in caus- lated that at some distant time the island’s in- ing deforestation as well as cultural changes, habitants had the imprudence to cut down all see Nunn et al. [2007]). Rather, the conven- their trees. La Pe´rouse lamented that loss of tional interpretations (e.g., Diamond 1995, trees exposed the soil to hot sun, dried up 2005, 2007, Flenley and Bahn 2002) for the streams, and desiccated the land, making the consequences of deforestation (i.e., demo- island almost uninhabitable. He wrote that graphic and cultural collapse) are not sup- the islanders ‘‘were indebted to the impru- ported in the archaeological and historical dence of their ancestors for their present un- records. We do not find unambiguous evi- fortunate situation’’ (1798:318–319). After dence for a huge population that crashed as decades of modern field research, contempo- the consequence of deforestation before Eu- rary writers tell a similar story (e.g., Flenley ropean contact. Although the ancient popula- and Bahn 2002, Diamond 2005) (Table 1). tion of Rapa Nui certainly contributed to the Diamond (1995:68) captured what he has island’s deforestation, they do not appear to called ‘‘ecocide’’ by arguing that ‘‘Eventually have committed ‘‘ecocide’’ as recounted in Easter’s growing population was cutting the the conventional scenario. forest more rapidly than the forest was regen- erating. The people used land for gardens Deforestation and Its Consequences and wood for fuel, canoes, and houses—and of course, for lugging statues. As forest disap- Diamond (2007) and others (e.g., Flenley peared, the islanders ran out of timber and and Bahn 2002) believe that the end of the rope to transport and erect their statues. Life forest brought huge losses for the islanders, became more uncomfortable—springs and but they assume that the palms and other streams dried up, and wood was no longer trees were integral to Polynesian subsistence. . Revisiting Rapa Nui (Easter Island) ‘‘Ecocide’’ Hunt and Lipo 603

TABLE 1 Comparative Dimensions of Conventional and Revised Models for Rapa Nui Prehistory

Dimension Conventionala Revised Referencesb

Colonization date A.D. 400–800 @A.D. 1200c Hunt and Lipo (2006, 2008) Archaeological visibility Largely invisible record Almost immediate; visibility Hunt and Lipo (2006, 2007, of colonization and for 400–800 years depends on age 2008); cf. Anderson early settlement resolution and nature of (1995), Wilmshurst et al. the record (2008) Human population Slow (@1%?), Rapid (@3%?), logistic Birdsell (1957); see early growth rate, size 15,000 þ max. growth rate but reaching historic population about 3,000–4,000? max. estimates (McCall 1994) Deforestation and Caused by reckless use of Caused by rats, fire, and Hunt (2006, 2007), Mieth ecological impacts fire and felling (to move some felling (but palms and Bork (2006); see also moai); impacts 800–400 not likely used moving Athens et al. (2002), years after colonization, moai); impacts begin Wilmshurst et al. (2008) then sudden and massive immediately, but some forest persists 400 years þ?; historic impacts from intensive sheep ranching Resource loss, subsistence Major, loss of trees as Negligible, palms did not Orliac (2000), Louwagie et change food source and for provide major food al. (2006), Hunt (2007) canoes providing access source given rat impacts; to deep-sea fishing; soil island never had suitable depletion/erosion trees for seaworthy canoes; depletion/ erosion of already nutrient-poor soils; some historic soil erosion Competition, social Warfare (including moai ‘‘De´tente,’’ ‘‘signaling,’’ Edwards and Marchetti conflict toppling), cannibalism and smaller-scale (1996), Hunt and Lipo conflicts (most moai (2001, 2007), toppled historically, by McLaughlin (2005) an earthquake?); little physical evidence for cannibalism Cause of ‘‘collapse’’ Overinvestment in moai; Historic introduction of Rainbird (2002), Hunt overpopulation; epidemic diseases (2007), Hunt and Lipo recklessness; (2007) deforestation Chronology of ‘‘collapse’’ Precontact, A.D. 1680 Postcontact, A.D. 1722– Rainbird (2002), Hunt and ‘‘event’’ 1877 Lipo (2007)

a Interpretations and assumptions from Flenley and Bahn (2002, 2007), Diamond (2005, 2007). b Selected sources providing evidence in support of a revised model for Rapa Nui prehistory. c A date of ca. A.D. 1200 represents the best current evidence for initial colonization based on radiocarbon dating. Rapa Nui may have been colonized earlier, but currently there is no reliable evidence to support a longer chronology.

Sources of fuel from wood were lost (Orliac even in the earliest deposits (Figure 1). Poly- 2000) and other negative consequences such nesians brought several cultivated plants with as soil erosion may have followed, but impor- them, and subsistence, as elsewhere in Poly- tant losses for subsistence remain open to nesia, relied largely on cultivation, chickens, doubt. For example, archaeological faunal as- and marine resources. semblages show that wild resources, such as Some have speculated that the loss of the birds, whose habitat would be supported by forest brought an end to deep-sea fishing as forest were never a dominant part of the materials for seaworthy canoes disappeared. diet; they occur only in small frequencies However, the palm trees, at least, were not a Figure 1. Faunal class data from Anakena Sand Dune excavations: (A) from Steadman et al. (1994), (B) from Univer- sity of Hawai‘i (2005 excavations). Faunal data from Skjolsvold (1994a,b) show comparable patterns. These faunal data reveal the abundance of rat (Rattus exulans), fish, and relatively smaller numbers (by NISP [number of identified specimens]) of bird. Contrary to common claims, sea mammals are common in the deposits and even abundant in later times. The horizontal axis shows excavated levels (arbitrary) or stratigraphic layers. Layers and levels are numbered consecutively from top to bottom so smaller numbers are younger, larger numbers older. . Revisiting Rapa Nui (Easter Island) ‘‘Ecocide’’ Hunt and Lipo 605 likely source of wood for canoes given their ticularly on offshore islands of New Zealand, fibrous interior and relatively thin, brittle shows that invasive species like rats lead to bark. At least one tree on Rapa Nui (Alphito- changes in vegetation structure and the loss nia zizyphoides [see Orliac 2000]) could have of birds, including seabirds (Fukami et al. provided suitable wood but only for smaller 2006). Rats prey upon eggs, chicks, and even canoes because the largest known (mature) adult seabirds, as well as altering nesting trees reach trunk diameters of 30–65 cm. habitats leading to their extirpation (e.g., At- And using small canoes on the largely unpro- kinson 1985). This has a cascading ecosystem tected open seas off Rapa Nui would have effect because seabirds play an important role often proven treacherous. This observation in nutrient cycling, so as they disappeared— is supported by the archaeological evidence perhaps rapidly—perhaps major nutrient cy- for marine subsistence. The fish remains cling to the soils of Rapa Nui would have from Anakena and other excavations com- been lost (e.g., Fukami et al. 2006). Yet recent prise primarily inshore taxa (Martinsson- research shows that even before seabirds Wallin and Crockford 2002), with larger ma- and forest disappeared, the primeval soils of rine resources represented in sea mammals. the island may well have been nutrient-poor Despite what some have claimed, the pres- (Louwagie et al. 2006). The implication is ence of sea-mammal bones does not ne- that deforestation did not turn fertile soil cessarily indicate a deep-sea fishing strategy. into poor soil; instead relatively poor soil got In ethnographic cases, using small canoes in worse. In short, the impacts to environmental nearshore waters, or simply by swimming, carrying capacity under dryland agriculture we know that people take dolphins by striking would likely have been minimal. This obser- stones together in nearshore waters to disori- vation is relevant to arguments that Rapa ent the animals’ echo-location system and Nui once supported a large human popula- driving them into shallow waters or on shore tion that crashed as a result of human- where they are killed (e.g., Bloch et al. 1990, induced ecological impacts. It seems that the Takegawa 1996, Porcasi and Fujita 2000). basic resources would never have supported a Sea-mammal bones from Anakena (the only population as large (e.g., 15,000 or more) as location where they are reported in any quan- Diamond (2005:90) and others have specu- tity for the island) probably represent this lated. specialized capture method, because it was re- Diamond (2007:1692) has claimed that de- ported for that location in the early twentieth forestation induced changes in horticultural century. It is also noteworthy that sea mam- practices. However, he bases this supposition mals do not disappear from the faunal record. on extrapolating the findings from one area, Sea-mammal bones are present, even abun- the Poike Peninsula, to the entire island; he dant, from multiple excavations in later pre- does this despite Mieth and Bork (2006:315) historic deposits at Anakena (see Figure 1), explicitly questioning whether their results showing they were not depleted by over- could be validly extrapolated to the entire exploitation nor did people loose access to island. But Poike is a peculiar place in Rapa sea mammals as a result of deforestation. Re- Nui because it has very little surface rock. sources such as dolphins were likely taken in Elsewhere, most of the island is covered by occasional drives on to the sandy beach at bedrock outcrops and a remarkably stony Anakena—one of the very few places on the surface composed of volcanic cobbles and island where this strategy would be feasible. boulders. Poike is not the only area of early Deep-sea fishing appears never to have been cultivation, as evidence throughout the island a common subsistence strategy on Rapa Nui, reveals closely contemporaneous use of fire, with or without suitable trees for canoes. etc., beginning in the late thirteenth century As the forest disappeared on Rapa Nui (see Mann et al. 2003, 2008). Moreover, there over the course of 400 to 500 years, some is little evidence that ‘‘stone mulching’’ is ex- negative consequences would likely have en- clusively an agricultural response to defores- sued. Comparative research on islands, par- tation, or particularly to falling crop yields . 606 PACIFIC SCIENCE October 2009

Figure 2. An example of ‘‘stone mulch’’ agricultural ﬁelds at Akahanga (south coast) similar to those found over many areas of Rapa Nui.

(Figure 2). Indeed, based on current evidence enhanced crop productivity and served to stone mulching appears to develop at least minimize agricultural risk. Thus, it seems un- 300 years before the forest disappeared (Ste- likely that stone mulch is best explained as venson et al. 1999). With or without forest, evidence of a desperate effort to grow crops stone mulching was an engineering solution after failures induced by deforestation. At first that optimized cultivation in nutrient-poor impression, the stony surface of much of soils in the windswept and variable rainfall Rapa Nui is easily misinterpreted as a barren conditions of Rapa Nui. wasteland consistent with ‘‘ecocide.’’ Some As detailed research has shown, stone view rocks, from a Western agricultural per- mulch minimized transpiration, moderated spective, as something to remove so that one temperature, added essential nutrients, and can plow fields for monocropping. But on mechanically protected crops against wind Rapa Nui the stone mulch was an ingenious damage (Stevenson et al. 1999, Wozniak method that provided cultigens with nutrients 1999, Bork et al. 2004). Stone mulch likely and functioned to retain soil moisture. served to optimize stability (versus short Diamond (2007:1692, italics added) sug- term maximization) in agricultural production gested ‘‘that in the first phase, erosion was given the challenges of an unpredictable envi- negligible, and horticulture was sustainable.’’ ronment. Similar uses of ‘‘stone mulching’’ This claim overlooks the chronological evi- are documented in Hawai‘i (Allen 2004; see dence and asserts a much longer chronology also Kirch et al. [2004b, 2005] on nutrient de- not evident for Rapa Nui. Precisely the lines pletion in soils) and in the desert Southwest of evidence that Diamond (2007:1693) sug- United States (e.g., Maxwell 1995), where it gested for dating the island’s first settlement . Revisiting Rapa Nui (Easter Island) ‘‘Ecocide’’ Hunt and Lipo 607

(i.e., first presence of Pacific rat, initial burn- stantial erosion with accumulation of collu- ing, changes in vegetation, and initial dates vium occurred in just the last 100–200 years from several sites, etc.) consistently point to (Stevenson et al. 2006). Based on dating approximately A.D. 1200 as the start of colo- wood charcoal particles in colluvium (i.e., nization (see Mann et al. 2003, 2008, Mieth from eroded soil and redeposited sediment), and Bork 2004, 2006, Hunt and Lipo 2006, even these age estimates suggest a maximum 2007, 2008). Diamond (2007) nonetheless (yet recent) age estimate for these deposition- opined that the chronology of island set- al events. It appears that much of the erosion tlement remains in question, implying that at Maunga Orito took place over the historic it may be many centuries longer. Taking this period, perhaps quite late (nineteenth cen- position allows him to speculate, in the tury), including a time when the island suf- absence of any reliable evidence, about a fered serious impacts of sheep ranching ‘‘first phase’’ of archaeologically invisible (Fischer 2005:135–198). Such late erosion land use in support of the conventional sce- raises questions about the chronology of nario. Here the tacit assumption is that peo- impacts to the landscape from deforestation ple arrived on the island centuries earlier, relative to those from historic ranching ac- left no visible impacts to the island, and re- tivities. In the absence of detailed research mained archaeologically invisible. Unfortu- distinguishing depositional events and their nately, such a scenario calls on faith given chronology, proclaiming a dramatic yet uni- the lack of evidence (see Anderson 1995; he formly high rate of soil erosion that conflates describes simplistic notions of ‘‘cryptic settle- effects from historic and prehistoric eras, as ment’’ in arguments made for long chronolo- Diamond (2007:1692) does, has little validity. gies; also see Hunt and Lipo 2007, 2008). Drawing on the detailed stratigraphic Rats and Deforestation work of Mieth and Bork (2006), Diamond (2007:1692) cited a rate of soil erosion pro- Diamond (2005, 2007, Rolett and Diamond ceeding uphill 3 m/year, implying that this 2004) is correct in suggesting that Rapa Nui’s represents a constant rate of soil erosion for deforestation can be understood in terms of the entire island. It does not and cannot rea- environmental fragility. This fragility meant sonably be extrapolated for the entire island that the island’s vegetation was vulnerable, (see Mieth and Bork 2006). This rate is esti- more than that of other islands with greater mated for the Poike Peninsula—an area of biocomplexity, to the impacts of invasive spe- fairly steep terrain with unconsolidated volca- cies such as the introduced Pacific rat (Rattus nic ash soils and almost no surface rocks. As a exulans). Concerning the role of rats, Hunt consequence, that part of the island is partic- (2007) suggested that we evaluate the relative ularly susceptible to erosion. It also remains contribution of rats, felling, and fire to the unknown how much of this erosion followed deforestation of the island. Hunt (2007) did grazing by thousands of sheep in the nine- not suggest that rats alone deforested the teenth century as well as modern efforts to island; so it is disingenuous to suggest, as farm the area with plowing. As Mieth and Diamond (2007:1693) did, that rats needed Bork (2006:289) reported for Poike, 7,000– hatchets and matches to play a role in the is- 10,000 sheep grazed an area of about 900 ha land’s ecological transformations. from 1930 to 1960. The grassland of Poike It is worth noting that many islands have was also burned annually during recent his- native forest that has persisted despite inva- toric times (Mieth and Bork 2006), and grass sion by the Pacific rat (Diamond 2007:1693). fires are still common. Intensive historic use However, a long list of islands with rats that and abuse must have taken a huge toll with were not deforested says little, if anything, soil erosion, which continues today. about Rapa Nui. Each island is unique in its Recent stratigraphic research on ancient history, biodiversity, and biogeography (e.g., land use and soil erosion at Maunga Orito ‘‘rescue’’ effects). To argue a false and simple on the west side of the island shows that sub- cause-effect that ‘‘rats mean deforestation’’ . 608 PACIFIC SCIENCE October 2009 assumes that diverse islands shared the same extinct palms of Rapa Nui ( Jubaea sp. or Pas- history, biogeography, and ecology. Even as chalococos disperta) had a life span of up to Diamond (1985:602) pointed out for island 2,000 years. So if rats were the sole agent of biota, ‘‘rats have caused catastrophic extinc- deforestation, mature trees would still exist tion waves on some islands, a few extinc- today. But a maximum life span of 2,000 tions on others, and no visible effect on still years for these extinct palms remains nothing others.’’ Otherwise, Diamond (2007) and more than speculation. In a recent review others (e.g., Flenley and Bahn 2007) should Tomlinson (2006:10) pointed out that ‘‘the explain how rats—as an invasive species in a age of a palm can only be determined accu- fragile ecosystem—could colonize success- rately from knowledge of its seed planting fully yet sustain no impact on Rapa Nui, and, date. An iconic example is the famous speci- by implication, that direct human actions men of Jubaea chilensis in the Temperate were the sole source of ecological change. House at Kew [London], planted in 1843.’’ Diamond (2007) also ignores the critical Extrapolated ages published for Jubaea palms research in Hawai‘i by Athens et al. (2002; range from only 100 to 740 years (Tomlinson see also Athens 2008) that documents the 2006:10). Finally, it is worth noting that the dramatic role of the introduced Pacific rat in Kew Garden Jubaea palm growing in the the rapid and extensive deforestation of na- equable conditions of a greenhouse (similar tive forest, including the predominant Prit- to the prime growing conditions of subtropi- chardia palms, shortly after the arrival of cal Rapa Nui?) is now 164 years old, and it is Polynesians. The paleoenvironmental evi- a large, adult tree now threatening to break dence from Hawai‘i reveals that rats had an the glass roof of Kew’s Temperate House. enormous impact on the lowland forests, even in those larger islands. Moreover, an ex- Human Population and ‘‘Ecocide’’ tensive literature documents the profound impact of rats on vegetation, and ecosystems Critical to the case for ‘‘ecocide’’ are es- generally (e.g., Towns et al. 2006). These ob- timates of changes in ancient population servations beg the question: what role did in- size and the timing of its decline. Diamond vasive rats play in Rapa Nui’s deforestation? (2007:1693) cited a recent source (Vargas et Indeed, the relative contribution of rats to al. 2006:232) as evidence of a pre-European deforestation on Pacific islands has yet to be contact ‘‘collapse.’’ Based on obsidian hy- fully explored. dration dates measured from samples taken Unfortunately, along these lines and per- from habitation sites (n ¼ 922), Diamond ar- haps in efforts to characterize the issue as gues that the occupation rises and falls, thus simplistic, Diamond (2007:1693) framed the reflecting the collapsing population. How- impact of rats as an all-or-nothing hypothe- ever, within the period when deforestation sis. For example, he (Diamond 2007:1693) progressed, about A.D. 1200 and 1700, the claimed, without documentation, that a larger population appears to have continued to rise number of palm nuts were burned rather (Figure 3). Thus, based on at least this analy- than rat gnawed. Yet no one has quantified sis (Vargas et al. 2006:232), it is apparent that the number of burned versus rat-gnawed as forest disappeared, the population grew. nuts sampled from deposits throughout the The first and only sign of sustained decline island (including all palm endocarps inspected in the population—where decline is vital to in the Englert Museum that are rat gnawed Diamond’s ‘‘ecocide’’ argument—occurs dur- [see Hunt 2007]). In any case, the observation ing an interval dated from A.D. 1750 to 1800, that some nuts were burned and others were after the arrival of the first European visitors. gnawed points to the potential of both rats Contrary to the narrative that has emerged and fire playing a role in the demise of the is- about a prehistoric collapse, a precipitous land’s forest. drop in the population with European contact Diamond (2007:1693) and others (Flenley is precisely what we would expect: as else- and Bahn 2007, Rolett 2008) assert that the where in the Pacific and the New World, . Revisiting Rapa Nui (Easter Island) ‘‘Ecocide’’ Hunt and Lipo 609

Figure 3. Habitations dated by obsidian hydration. Arrow indicates a dramatic decline in dated habitations in the historic (postcontact) period. These data do not support a hypothesis of precontact demographic collapse. Values are re- plotted from data provided by Vargas et al. (2006:232).

European germs wiped out many native pop- cated instruments are used to accurately ulations lacking immunity to Old World measure hydration rims (e.g., Stevenson et diseases (e.g., Crosby 1972, 1986, Stannard al. 1998, 2001), two kinds of problems re- 1991), as documented archaeologically (e.g., main. First, the lack of any deterministic re- see Kirch [2007] for an exemplary study lationship between the events measured by from Maui). radiocarbon and that of obsidian hydration Rapa Nui’s distribution of dated habitation means that establishing one with the other as sites lends support for a post-European con- if by ‘‘calibration’’ suffers from unacceptably tact depopulation event and provides at least low precision. Second, multiple factors can ordinal-level information about the history influence hydration of specimens and thus of prehistoric population size, its growth, their ‘‘dating,’’ so that inconsistent results and collapse, but other problems warrant cannot be adequately evaluated. As Anovitz critical evaluation. This dated distribution et al. (1999) established, obsidian hydration of structures from Vargas et al. (2006:232) cannot reliably provide interval-level chrono- derives entirely from obsidian hydration logical results. At best the method may show measurements of flaked artifacts. Several potential in large numbers as a relative mea- problems inherent in the method suggest sure of time. No amount of ‘‘recalibrating’’ that the distribution may not provide an ac- hydration dates will produce a single value curate absolute chronology. First, no valid robust enough to warrant absolute chrono- hydration rate has been independently estab- logical claims. Indeed, the establishment of lished for Rapa Nui obsidians. Attempts to do ever ‘‘better’’ rates in the field or laboratory so have proceeded by comparing experimen- to match what is assumed about chronology tal hydration results with radiocarbon dates risks a circular logic in employing obsidian from associated excavation context (Steven- hydration as an absolute dating method. son 2000). Although increasingly sophisti- Regarding the early end of this sequence . 610 PACIFIC SCIENCE October 2009

(Figure 3) using obsidian hydration, it is also concrete-replica moai (statue) and shown in worth noting that we cannot expect measure- transport experiments on a NOVA television ment error to be normally distributed. As the documentary (Love 2000). The assumption is hydrated portion approaches zero (i.e., with that it took large numbers of people to make the youngest samples), the measured rind and especially move moai, as well as those will physically be biased more toward larger who supported their efforts with agricultural values than smaller ones (because negative surplus. Diamond (2007:1693) mistakenly re- numbers are obviously invalid). Thus the er- ferred to a 9-ton statue pulled a distance of 10 ror will be toward older age estimates, not km; he also suggested that the task would re- younger ones. A second concern is that any quire 70 people, who in turn must have been rate used to estimate age has some unspecifi- fed by 400. These speculative extrapolations able degree of error. And third, hydration are used to argue that a population of thou- rates likely vary by source and environment sands existed on at least one portion of the is- of deposition. In general, given the numbers land. This is taken as support of a huge (i.e., of dates used to generate the distribution re- well beyond carrying capacity) islandwide ported by Vargas et al. (2006) and the rela- population. But to set the record straight, tively short amount of time represented Van Tilburg and Ralston (2005:295) de- overall, it is likely that measurement and scribed two experiments pulling a 10-ton other sources of error added ages to the esti- (not 9-ton) statue with a team (46 people, mates, making chronological distributions not 70) 110 m in pulls of 40 and 70 m, and left-skewed (i.e., too early). By chance alone, in the second attempt, just 50 m (not the 10 one would expect some of the dates to be km cited). too old. This would explain the small number But these figures are likely moot. Unfor- of early obsidian dates (i.e., before ca. A.D. tunately, the experiments cannot prove the 1200, when the island appears to have been actual statue-moving methods used. More- first settled based on the radiocarbon chro- over, extrapolations from one method—never nology currently established [see Hunt and demonstrated—are used to speculate about Lipo 2006]). population size that only heaps speculation In sum, although we question the validity upon speculation. It is also important to note of details in the age distribution based on that Van Tilburg and Ralston’s (2005) exper- obsidian hydration, particularly toward the iment using an A-frame sled with sliders over early end, it may nonetheless provide some log rails, or ‘‘canoe ladders,’’ is plagued with glimpse of postcontact demographic collapse. problems, particularly in light of the coastal A postcontact demographic collapse follows locations where statues were transported and the predictable consequences of the introduc- erected (e.g., see Lee 1998, 1999). Our own tion of Old World disease. ‘‘Ecocide’’ calls field research (Lipo and Hunt 2005) and for a precontact, indeed even a ‘‘1680 AD’’ detailed observations (Sergio Rapu, pers. event (Diamond 2005; see also Ferdon 1961, comm., 2005) on more than 50 ‘‘in transit’’ Heyerdahl 1961), based on the spurious cor- statues abandoned along the ancient roadways relation of oral traditions and a radiocarbon show that the moving experiments and the date seen as the cumulative social conse- methods reconstructed from them have not quences of deforestation and overexploitation drawn directly on patterns in the archaeolog- of resources. The obsidian hydration evi- ical evidence for transport such as the roads dence, albeit problematic, does not support a and statues abandoned along them. chronology for a precontact population collapse premised on ‘‘ecocide.’’ reenvisioning rapa nui’s prehistory In support of arguments for a large prehistoric population—one that overshot the is- In contrast to a conventional view that has land’s limited carrying capacity—Diamond been promulgated for some time, we envision (2007:1693) cited experiments reported by a different outline of prehistory based on cur- Van Tilburg and Ralston (2005) using a rent evidence (see Table 1). As we consider . Revisiting Rapa Nui (Easter Island) ‘‘Ecocide’’ Hunt and Lipo 611 alternatives, evaluating them in empirically population could reach 2,000 in just 123 sufficient terms will be critical. years. Initial logistic growth rates would slow as the population reached a sustainable number. Based on Birdsell’s data, demographic Colonization Chronology models, island population densities known Research now suggests that colonization of for Polynesia, and early estimates of Rapa eastern Polynesian archipelagos such as the Nui population (see McCall 1994), we hy- Cooks, Societies, Marquesas, and Hawaiian pothesize that a sustainable number could be Islands first occurred around A.D. 900 to reached in less than two centuries, reaching 1000. Based on higher standards for radiocar- about 3,000–4,000, attaining an intermediate bon dating and its validity, these chronologies density in the range of about 20 people per for eastern Polynesia are centuries later than square kilometer. In this light, the high den- have long been assumed (e.g., see Spriggs sity of archaeological remains on Rapa Nui and Anderson 1993, Anderson and Sinoto reflects 30–40 generations living on this small 2002). Recent field research and reevaluation island and producing a dense palimpsest rec- of radiocarbon dates show that New Zealand ord often largely visible at the surface. Future (e.g., Wilmshurst et al. 2008) and the distant paleodemographic research for Rapa Nui can islands of the Austral and Gambier groups be modeled on recent successes from Hawai‘i were not colonized until after A.D. 1000– (e.g., Kirch 2007) and will necessarily address 1200 (Kirch et al. 2004a, Kennett et al. estimating a range for the island’s population 2006). In this last episode of migration, per- size. haps coincident with high frequencies of El Nin˜ o–Southern Oscillation events (Finney Deforestation 1985, 2001, Anderson et al. 2006), Polynesian colonists first reached Rapa Nui. Based on Over the first century of human settlement chronology and geographic proximity, per- on Rapa Nui the environmental impacts of haps among other lines of evidence (Green rats and the use—and runaway effects—of 1998, 2000), migrations to Rapa Nui may fire would have initiated precipitous forest have originated in the Gambier Group decline. The loss of trees, both in terms of (Mangareva-Pitcairn region). The founding mature individuals (fire and felling) and their population may have been small, numbering recruitment (rats destroying seeds), may have somewhere around 50 people or less, but this been substantial. Although forest loss is un- is only conjecture. fortunate from a biodiversity perspective, these anthropogenic changes may have had little impact on overall agricultural productiv- Population Growth and Size ity. Deforestation also meant habitat loss for The rapid colonization of scores of islands native species, including what was certainly a over the vastness of the eastern Pacific de- large population of seabirds. Seabirds would manded multiple migrations supported by a have played a role in nutrient cycling enrich- high population growth rate. As Birdsell ing the soil of the island, but those soils were (1957) pointed out, rapid colonization of is- already nutrient-poor—a problem for agri- lands requires population growth as a nec- culturalists since the island was first settled. essary foundation for migrations. From the historically documented population on Pit- Statues and Monuments cairn after its recolonization by the Bounty mutineers and their mates, Birdsell (1957) Production and transport of statues as well as showed that population growth rates of monument building began soon after the 3.4% and higher occur among human colo- island was settled, although this and other nists. If a rate approaching this (e.g., 3.0%) dimensions of the archaeological record de- occurred with the first colonization of Rapa mand better chronological control. Early stat- Nui, starting with only 50 people the island’s ues were variable in form and made from a . 612 PACIFIC SCIENCE October 2009 variety of materials found across the island. there remains much to be documented in the As the industry developed, it became rela- island’s archaeology, paleoecology, and pre- tively standardized, and production focused history. But the point we wish to reiterate more exclusively on the quarry at Rano Ra- here is that the ‘‘conventional’’ accounts pro- raku. The major investments in architecture vided by some (e.g., Diamond 1995, 2005, and statuary may have served the Rapa Nui 2007, Flenley and Bahn 2002) do not com- population well in terms of mitigating violent port well with the existing evidence. Our competition and overpopulation (Hunt and concern is that a conventional narrative dom- Lipo 2001; see Dunnell 1999, Madsen et al. inates, one that has unsubstantiated roots in 1999). The evolution of such cultural elabo- modern myths (e.g., Heyerdahl 1961), and rations can potentially be explained by evolu- still frames much of what readers believe to tionary ecology and modeled as dimensions be accurate about the island’s prehistory. of ‘‘signaling’’ (e.g., Bliege Bird and Smith This narrative continually finds its way into 2005), ‘‘advertising’’ (Neiman 1997), and popular and academic sources. We question ‘‘bet-hedging’’ (e.g., Madsen et al. 1999, the historical veracity and usefulness of these Hunt and Lipo 2001), where ecological and conventional accounts and invite open and evolutionary dimensions of explanatory mod- critical research into this island’s astonishing els are complementary. In cultural and demo- cultural and environmental story. Under- graphic terms, the human settlement of Rapa standing what happened on Rapa Nui is im- Nui can be regarded a success, more so than portant and likely to have broad implications some kind of catastrophe-induced collapse. for understanding human and environmental history. The dynamic story of this small, re- mote island has much to tell us. It is critical Demographic Collapse and Genocide that we do our best to get it right. The demographic collapse likely began as a consequence of the devastating effects of acknowledgments Old World disease (Crosby 1972, 1986, Do- byns 1993). The arrival of the Dutch in 1722 We thank David Addison, Atholl Anderson, J. may have brought the first wave of epidemics, Stephen Athens, Kelley Esh, Donald Gray- even as they remained invisible to Europeans. son, Patrick Kirch, Alex Morrison, Sergio The next contacts occurred 48 years (about Rapu, John Terrell, Francisco Torres, and Ja- two generations) later with the arrival of the net Wilmshurst for sharing sources, thoughts, Spanish in 1770 and then the English in and comments generally on our research. We 1774. It is perhaps noteworthy that from alone are responsible for the errors. Cook’s 1774 visit onward, many of the de- scriptions of the island and its people are dis- Literature Cited mal. Was the impression formed by Cook and his crew a result of grave disease impacts Allen, M. S. 2004. Bet-hedging strategies, from a Spanish visit just 4 years before? The agricultural change, and unpredictable en- genocide of the Rapa Nui population con- vironments: Historical development of tinued into the nineteenth century with slave dryland agriculture in Kona, Hawaii. J. trading, epidemics, and other devastating Anthropol. Archaeol. 23:196–224. blows from colonialism on this small popula- Anderson, A. 1995. Current approaches in tion. East Polynesian colonisation research. J. Polynesian Soc. 104:110–132. concluding thoughts Anderson, A., J. Chappell, M. Gagan, and R. Grove. 2006. Prehistoric maritime migra- The details of this narrative will certainly tion in the Pacific Islands: An hypothesis change as basic archaeological and paleo- of ENSO forcing. Holocene 16 (1): 1–6. environmental research continues on Rapa Anderson, A., and Y. Sinoto. 2002. New ra- Nui. In contrast to a widespread impression, diocarbon ages of colonization sites in . Revisiting Rapa Nui (Easter Island) ‘‘Ecocide’’ Hunt and Lipo 613

East Polynesia. Asian Perspect. 41:242– exchange: Biological and cultural conse- 257. quences of 1492. Greenwood Press, West- Anovitz, L. M., J. M. Elam, L. R. Riciputi, port, Connecticut. and D. R. Cole. 1999. The failure of ob- ———. 1986. Ecological imperialism: The sidian hydration dating: Sources, implica- biological expansion of Europe, 900– tions, and new directions. J. Archaeol. Sci. 1900. Cambridge University Press, Cam- 26:735–752. bridge. Athens, J. S. 2008. Rattus exulans and the cat- Dalton, T. R., and R. M. Coats. 2000. Could astrophic disappearance of Hawai‘i’s native institutional reform have saved Easter Is- lowland forest. Biol. Invas. 11 (7): 1489– land? J. Evol. Econ. 10 (5): 489–506. 1501. De Monocol, P. B., E. R. Cook, D. Demeritt, Athens, J. S., H. D. Tuggle, J. V. Ward, and A. Hornborg, P. V. Kirch, R. McElreath, D. J. Welch. 2002. Avifaunal extinctions, and J. A. Tainter. 2005. Perspectives vegetation change, and Polynesian impacts on Diamond’s Collapse: How Societies in prehistoric Hawai‘i. Archaeol. Oceania Choose to Fail or Succeed. Curr. Anthro- 37:57–78. pol. 46 (Suppl.): S91–S99. Atkinson, I. A. E. 1985. The spread of com- Diamond, J. 1985. Rats as agents of extermi- mensal species of Rattus to oceanic islands nation. Nature (Lond.) 318:602–603. and their effects on island avifaunas. Pages ———. 1995. Easter’s end. Discover 9:62– 35–81 in M. J. Moors, ed. Conservation of 69. island birds: Case studies for the manage- ———. 2005. Collapse: How societies choose ment of threatened island species. ICBP to fail or succeed. Viking, New York. Technical Publications. Vol. 3. Interna- ———. 2007. Easter Island revisited. Science tional Council for Bird Preservation, Cam- (Washington, D.C.) 317:1692–1694. bridge, United Kingdom. Dobyns, H. F. 1993. Disease transfer at con- Barnes, S., L. Matisoo-Smith, and T. Hunt. tact. Annu. Rev. Anthropol. 22:273–291. 2006. Ancient DNA of the Paciﬁc rat (Rat- Dunnell, R. C. 1999. The concept of waste in tus exulans) from Rapa Nui (Easter Island). an evolutionary archaeology. J. Anthropol. J. Archaeol. Sci. 33:1536–1540. Archaeol. 18:243–250. Birdsell, J. B. 1957. Some population prob- Edwards, E., and R. Marchetti. 1996. When lems involving Pleistocene man. Cold the Earth trembled, the statues fell. Rapa Spring Harbor Symp. Quant. Biol. 22:47– Nui J. 10:1–15. 69. Ferdon, E. 1961. A summary of the excavated Bliege Bird, R., and E. A. Smith. 2005. Sig- record of Easter Island prehistory. Pages naling theory, strategic interaction, and 527–535 in T. Heyerdahl and E. N. Fer- symbolic capital. Curr. Anthropol. 46 (2): don, eds. Archaeology of Easter Island. 221–248. Monogr. Sch. Am. Res. Mus. N. Mex. 24 Bloch, D., G. Desportes, K. Hoydal, and P. (1). Jean. 1990. Pilot whaling in the Faroe Is- Finney, B. 1985. Anomalous westerlies, El lands, July 1986–July 1988. North Atl. Nino, and the colonization of Polynesia. Stud. 2 (1, 2): 36–44. Am. Anthropol. 87:9–26. Bork, H.-R., A. Mieth, and B. Tschochner. ———. 2001. Voyage to Polynesia’s land’s 2004. Nothing but stones? A review of end. Antiquity 75:172–181. the extent and technical efforts of prehis- Fischer, S. R. 2005. Island at the end of the toric stone mulching on Rapa Nui. Rapa world: The turbulent history of Easter Is- Nui J. 18:10–14. land. Reaktion Books, London. Brander, J. A., and M. S. Taylor. 1998. Flenley, J. R., and P. Bahn. 2002. The enig- The simple economics of Easter Island: mas of Easter Island. Oxford University Ricardo-Malthus model of renewable re- Press, New York. source use. Am. Econ. Rev. 88:19–38. ———. 2007. Conﬂicting views of Easter Is- Crosby, A. W., Jr. 1972. The Columbian land. Rapa Nui J. 21:11–13. . 614 PACIFIC SCIENCE October 2009

Fukami, T., D. A. Wardle, P. J. Bellingham, ———. 2008. Evidence for a shorter chro- C. P. H. Mulder, D. P. Towns, G. W. nology on Rapa Nui (Easter Island). J. Isl. Yeates, K. I. Bonner, M. S. Durrett, M. Coastal Archaeol. 3:140–148. N. Grant-Hoffman, and W. M. William- Kennett, D., A. Anderson, M. Prebble, E. son. 2006. Above- and below-ground im- Conte, and J. Southon. 2006. Prehistoric pacts of introduced predators in seabird- human impacts on Rapa, French Polyne- dominated island ecosystems. Ecol. Lett. sia. Antiquity 80:1–15. 9:1299–1307. Kirch, P. V. 2007. Palaeodemography in Ka- Gee, H. 2004. Environmental geography: hikinui, Maui: An archaeological approach. Treeless at Easter. Nature (Lond.) 431:411. Pages 90–107 in P. V. Kirch and J.-L. Gibbons, A. 2006. Dates revise Easter Island Rallu, eds. The growth and collapse of Pa- history. Science (Washington, D.C.) cific island societies: Archaeological and 311:1360. demographic perspectives. University of Green, R. C. 1998. Rapanui origins prior to Hawai‘i Press, Honolulu. European contact: The view from eastern Kirch, P. V., J. Coil, A. S. Hartshorn, M. Polynesia. Pages 87–110 in P. Vargas Ca- Jeraj, P. M. Vitousek, and O. A. Chadwick. sanova, ed., Easter Island and East Polyne- 2005. Intensive dryland farming on the sian prehistory. Instituto de Estudios Isla leeward slopes of Haleakala, Maui, Hawai- de Pascua, Facultad de Arquitectura y Ur- ian Islands: Archaeological, archaeobotani- banismo, Universidad de Chile, Santiago. cal, and geochemical perspectives. World ———. 2000. Origins for the Rapanui of Archaeol. 37:240–258. Easter Island before European contact: Kirch, P. V., J. Coil, M. I. Weisler, E. Conte, Solutions from holistic anthropology to and A. J. Anderson. 2004a. Radiocarbon an issue no longer much of a mystery. dating and site chronology. Pages 94–105 Rapa Nui J. 14:71–76. in E. Conte and P. V. Kirch, eds. Archaeo- Heyerdahl, T. 1961. An introduction to Eas- logical investigations in the Mangareva ter Island. Pages 21–90 in T. Heyerdahl Islands (Gambier Archipelago), French and E. N. Ferdon, eds. Archaeology of Polynesia. Contribution No. 62, Archaeo- Easter Island. Monogr. Sch. Am. Res. logical Research Facility, University of Mus. N. M. 24 (1). California, Berkeley. Hunt, T. L. 2006. Rethinking the fall of Kirch, P. V., A. S. Hartshorn, O. A. Chad- Easter Island: New evidence points to an wick, P. M. Vitousek, D. R. Sherrod, J. alternative explanation for a civilization’s Coil, L. Holm, and W. D. Sharp. 2004b. collapse. Am. Sci. 94:412–419. Environment, agriculture, and settlement ———. 2007. Rethinking Easter Island’s patterns in a marginal Polynesian land- ecological catastrophe. J. Archaeol. Sci. scape. Proc. Natl. Acad. Sci. U.S.A. 34:485–502. 101:9936–9941. Hunt, T. L., and C. P. Lipo. 2001. Cultural La Pe´rouse, J. F. G. de. 1798. A voyage round elaboration and environmental uncertainty the world performed in the years 1785, in Polynesia. Pages 103–115 in C. M. 1786, and 1788. J. Johson, London. Stevenson, G. Lee, and F. J. Morin, eds. Lee, V. R. 1998. Rapa Nui rocks: Impressions Pacific 2000. Proceedings of the 5th Inter- from a brief visit. Rapa Nui J. 12:69–72. national Conference on Easter Island and ———. 1999. Rapa Nui rocks update. Rapa the Pacific. Easter Island Foundation, Los Nui J. 13:16–17. Osos, California. Lipo, C. P., and T. L. Hunt. 2005. Mapping ———. 2006. Late colonization of Easter prehistoric statue roads on Easter Island. Island. Science (Washington, D.C.) Antiquity 79:158–168. 311:1603–1606. Louwagie, G., C. M. Stevenson, and R. Lan- ———. 2007. Chronology, deforestation, gohr. 2006. The impact of moderate to and ‘‘collapse’’: Evidence vs. faith in Rapa marginal land suitability on prehistoric Nui prehistory. Rapa Nui J. 21:85–97. agricultural production and models of . Revisiting Rapa Nui (Easter Island) ‘‘Ecocide’’ Hunt and Lipo 615

adaptive strategies for Easter Island (Rapa V. Winiwarter, eds. Soils and societies: Nui, Chile). J. Anthropol. Archaeol. Perspectives from environmental history. 25:290–317. White Horse Press, Isle of Harris, United Love, C. 2000. More on moving Easter Island Kingdom. statues, with comments on the NOVA Neiman, F. D. 1997. Conspicuous consump- program. Rapa Nui J. 14:115–118. tion as wasteful advertising: A Darwinian Madsen, M., C. Lipo, and M. Cannon. 1999. perspective on spatial patterns in Classic Fitness and reproductive trade-offs in Maya terminal monument dates. Pages uncertain environments: Explaining the 267–290 in C. M. Burton, G. Clark, and evolution of cultural elaboration. J. D. Bamforth, eds. Rediscovering Darwin: Anthropol. Archaeol. 18:251–281. Evolutionary theory and archaeological ex- Mann, D., J. Chase, J. Edwards, R. Beck, R. planation. Archaeol. Pap. Am. Anthropol. Reanier, and M. Mass. 2003. Prehistoric Assoc. 7. destruction of the primeval soils and vege- Nunn, P. D., R. Hunter-Anderson, M. T. tation of Rapa Nui (Isla de Pascua, Easter Carson, F. Thomas, S. Ulm, and M. J. Island). Pages 133–153 in J. Loret and Rowland. 2007. Times of plenty, times of J. T. Tanacredi, eds. Easter Island: Scien- less: Last-millennium societal disruption in tific exploration in the world’s environ- the Pacific Basin. Hum. Ecol. 35:385–401. mental problems in microcosm. Kluwer Orliac, C. 2000. The woody vegetation of Academic/Plenum, New York. Easter Island between the early 14th and Mann, D., J. Edwards, J. Chase, W. Beck, R. the mid-17th centuries A.D. Pages 211– Reanier, M. Mass, B. Finney, and J. Loret. 220 in C. Stevenson and W. Ayres, eds. 2008. Drought, vegetation change, and hu- Easter Island archaeology: Research on man history on Rapa Nui (Isla de Pascua, early Rapanui culture. Easter Island Foun- Easter Island). Quat. Res. 69:16–28. dation, Los Osos, California. Martinsson-Wallin, H., and S. J. Crockford. ———. 2003. Ligneux et palmiers de l’ile de 2002. Early settlement of Rapa Nui (Easter Paques du XIeme au XVIIeme siecle de Island). Asian Perspect. 40 (2): 244–278. notre ere. Pages 184–199 in C. Orliac, ed. Maxwell, T. D. 1995. The use of comparative Archeologie en Oceanie insulaire: Peuple- and engineering analyses in the study of ment, societes et paysages. Editions Art- prehistoric agriculture. Pages 113–128 in com, Paris. P. A. Teltser, ed. Evolutionary archaeol- Peiser, B. 2005. From genocide to ecocide: ogy: Methodological issues. University of The rape of Rapa Nui. Energy Environ. Arizona Press, Tucson. 16:513–539. McCall, G. 1994. Rapanui: Tradition and Porcasi, J. F., and H. Fujita. 2000. The dol- survival on Easter Island. University of phin hunters: A specialized prehistoric ma- Hawai‘i Press, Honolulu. rine adaptation in the southern California McLaughlin, S. 2005. Cannibalism and Channel Islands and Baja California. Am. Easter Island: Evaluation, discussion of Antiq. 65:543–566. probabilities, and survey of the literature Rainbird, P. 2002. A message for our future? on the subject. Rapa Nui J. 19:30–50. The Rapa Nui (Easter Island) ecodisaster Mieth, A., and H.-R. Bork. 2004. Easter and Pacific island environments. World Island—Rapa Nui: Scientific pathways to Archaeol. 33:436–451. secrets of the past. Man and Environment Redman, C. L. 2001. Human impacts on an- 1. Department of Ecotechnology and Eco- cient environments. University of Arizona system Development, Ecology Center, Press, Tucson. Christian-Albrechts-Universitat zu Kiel, Rolett, B. 2008. Avoiding collapse: Pre- Kiel, Germany. European sustainability on Pacific islands. ———. 2006. The dynamics of soil, land- Quat. Int. 184:4–10. scape and culture on Easter Island (Chile). Rolett, B., and J. Diamond. 2004. Environ- Pages 273–321 in J. R. McNeill and mental predictors of pre-European defor- . 616 PACIFIC SCIENCE October 2009

estation on Pacific islands. Nature (Lond.) Stevenson, C. M., B. Scheetz, and J. J. Mazer. 431:443–446. 1998. Laboratory obsidian hydration rates: Skjolsvold, A. 1994a. Archaeological investi- Theory, method, and application. Pages gations at Anakena, Easter Island. Pages 181–204 in S. Shackley, ed. Archaeological 5–120 in A. Skjolsvold, ed. Archaeological obsidian studies: Methods and theory. Ple- investigations at Anakena, Easter Island. num Press, New York. Vol. 3. The Kon Tiki Museum Occasional Stevenson, C., J. Wozniak, and S. Haoa. Papers, Oslo. 1999. Prehistoric agricultural production ———. 1994b. Appendix: Osteological analy- on Easter Island (Rapa Nui), Chile. Antiq- sis. Pages 203–211 in A. Skjolsvold, ed. Ar- uity 73:801–812. chaeological investigations at Anakena, Takegawa, D. 1996. The method of dolphin Easter Island. Vol. 3. The Kon Tiki Mu- hunting and the distribution of teeth and seum Occasional Papers, Oslo. meat: Dolphin hunting in the Solomon Is- Spriggs, M., and A. Anderson. 1993. Late col- lands 2. Pages 67–80 in T. Akimichi, ed. onisation of East Polynesia. Antiquity Coastal foragers in transition. Senri Eth- 67:200–217. nological Studies No. 42. National Mu- Stannard, D. E. 1991. The consequences seum of Ethnology, Osaka. of contact: Toward an interdisciplinary Tomlinson, P. B. 2006. The uniqueness of theory of native responses to biological palms. Bot. J. Linn. Soc. 151:5–14. and cultural invasion. Pages 519–539 in Towns, D. R., I. A. E. Atkinson, and C. H. D. H. Thomas, ed. Columbian conse- Daugherty. 2006. Have the harmful effects quences. Vol. 3. The Spanish borderlands of introduced rats on islands been exagger- in Pan-American perspective. Smithsonian ated? Biol. Invas. 8:863–891. Institution Press, Washington, D.C. Van Tilburg, J. A., and T. Ralston. 2005. Steadman, D. W., P. Vargas, and C. Cristino. Megaliths and mariners: Experimental ar- 1994. Stratigraphy, chronology, and cul- chaeology on Easter Island. Pages 279– tural context of an early faunal assemblage 303 in K. Johnson, ed. Onward and up- from Easter Island. Asian Perspect. 33 (1): ward! Papers in honor of Clement W. 79–96. Meigan. Stansbury Publications, Chico, Stevenson, C. M. 2000. Estimating Easter Is- California. land obsidian hydration rates from glass Vargas, P., C. Cristino, and R. Izaurieta. composition. Pages 205–210 in C. M. Ste- 2006. 1000 An˜ os en Rapa Nui. Arqueolo- venson and W. S. Ayres, eds. Easter Island gıá del asentamiento. Editorial Universita- archaeology: Research on early Rapanui ria, Santiago de Chile. culture. Easter Island Foundation, Los Wilmshurst, J. M., A. J. Anderson, T. F. G. Osos, California. Higham, and T. H. Worthy. 2008. Dating Stevenson, C. M., I. M. Abdelrehim, and the late prehistoric dispersal of Polyne- S. W. Novak. 2001. Infra-red photoacous- sians to New Zealand using the com- tic and secondary ion mass spectrometry mensal Pacific rat. Proc. Natl. Acad. Sci. measurements of obsidian hydration rims. U.S.A. 105:7676–7680. J. Archaeol. Sci. 28:109–115. Wozniak, J. 1999. Prehistoric horticultural Stevenson, C. M., T. L. Jackson, A. Mieth, practices on Easter Island: Lithic mulched H.-R. Bork, and T. N. Ladefoged. 2006. gardens and field systems. Rapa Nui J. Prehistoric and early historic agriculture 13:95–99. at Maunga Orito, Easter Island (Rapa Young, L. B. 1991. Easter Island: Scary para- Nui), Chile. Antiquity 80:919–936. ble. World Monitor August: 40–45.