Rapa Nui Journal: Journal of the Island Foundation Volume 16 Article 4 Issue 1 May

2002 Productive Strategies in an Uncertain Environment: Prehistoric Agriculture on Christopher M. Stevenson Virginia Department of Historic Resources

Thegn Ladefoged University of Washington, Dept. of Anthropology

Sonia Haoa Isla de Pascua,

Follow this and additional works at: https://kahualike.manoa.hawaii.edu/rnj Part of the History of the Pacific slI ands Commons, and the Pacific slI ands Languages and Societies Commons

Recommended Citation Stevenson, Christopher M.; Ladefoged, Thegn; and Haoa, Sonia (2002) "Productive Strategies in an Uncertain Environment: Prehistoric Agriculture on Easter Island," Rapa Nui Journal: Journal of the Easter Island Foundation: Vol. 16 : Iss. 1 , Article 4. Available at: https://kahualike.manoa.hawaii.edu/rnj/vol16/iss1/4

This Research Paper is brought to you for free and open access by the University of Hawai`i Press at Kahualike. It has been accepted for inclusion in Rapa Nui Journal: Journal of the Easter Island Foundation by an authorized editor of Kahualike. For more information, please contact [email protected]. - Stevenson et al.: Productive Strategies in an Uncertain Environment PftODU rI C fRATCCICS IN AN UN ftT IN E:.NVlftONMLNI: PRTHI TOftle Cftl ULTUftC N EA TCft L LA 0

Christopher M. Stevenson!, Thegn Ladejoged2 and Sonia Haoa3

'Virginia Department ofHistoric Resources, Petersburg, VA 23805; 2Department ofAnthropology, University ofWashington, Seattle, WA; 3Sonia Haoa, Isla de Pascua, Chile

I TRODUCTIO will look at how the prehi toric Rapanui utilized and changed the i land landscape. These land cape modification on Rapa he unpredictable nature of a key environmental re ource Nui and el ewhere were developed in re pon e to fundamental T (e.g., moi ture) is often viewed a a major variable that environmental constraints within which agriculture wa prac­ guide economic deci ion-making and the form of the ocial ticed. Thi was a dynamic and continuou proce . During the organization required to implement adaptive strategies that will settlement of the Pacific, human colonizer initially took ad­ en ure the urvival of the population. Several author have re­ vantage of natural landscape features such as wind-protected cently argued that environment characterized by higher uncer­ locations with plentiful upplies of fre h water. The e advanta­ tainty (e.g., ri k of crop failure) will exhibit a greater tempo of geous settings were used for early settlement and agriculture. changing organizational form and an overall higher level of The later expansion of gardens and field systems significantly organizational complexity. (Ladefoged 1995; Graves and modified the ecological structure of many i lands (Kirch 1997, Ladefoged 1995; Dunnell 1999; Ladefoged and Grave 2000; 2000; Flenley and King 1984; Flenley et al. 1991) and the ur­ Kirch 2000; Hammon 2001; Hunt and Lipo 2001). In the e face oils that support the living biota. The elimination of habi­ high-ri k environments the con truction of chiefly pon ored tat and soil ero ion have been cited as two outcomes of an ex­ monumental architecture can uppre s population number to panding agricultural y tern. Thi proces increased agricul­ below critical level . Energy that could be expended on in­ tural ri k and required innovation within the subsistence sy ­ creasing agricultural production, thereby leading to increa e tern to maintain required production level. It is this proce of in population level, is diverted into "wa teful" activities, uch innovation that we will inve tigate through an examination of as monumental construction. Thi ha the two-fold advantage the Rapa Nui land cape. of suppre sing population num- In this paper, we will ex­ ber to level u tainable during amine the Rapanui agriculture environmental perturbation Rapa Nui ystem and ask three question : and maintaining a pool of labor ~ that can be diverted into ubsi­ I) What were the risk factor tence activitie when nece sary. N La Perouse behind rainfall agriculture and In addition, monumental archi­ how did the risk vary acro the tecture can be used to bolster i land land cape? ideological beliefs that are nec­ essary for maintaining an inte­ 2) How did agricultural use of grated chiefly polity (Earle the land cape vary in response 1997). The e polities can often to different habitat conditions? Akahallga RUllga Vai provide opportunities for re- ource buffering during period 3) What i the chronology of Hallga POllkllrn of environmental perturbation prehi toric agricultural tech­ (Ladefoged 1995). The imple­ nologie and landscape u e? mentation of labor intensive o 5 forms of agricultural intensifi­ THERApA VI ENVIRONME T cation in re ponse to marginal and uncertain environmental Three volcanoe form the Rapa Figure I. Map of Easter I land (Rapa Nui) howing the Vaitea area condition require that the au­ Nui landscape. Mt. Terevaka i (center) and the Heki'i area near La Perou e. thority of the hierarchy be the principal shield volcano that trong in order that production form the central rna s of the above the hou ehold level can be obtained. i land. Strato volcanoe at the outhwe tern and southea tern In this paper, we look at how the practice of prehistoric tip, Rano Kau and Poike respectively, form the other elevated agriculture may have contributed to the development and landform . Mt. Terevaka reaches a height of approximately maintenance of a hierarchical social organization. In our ap­ 450 m. A low coa tal plain connect the three volcanoe along proach to understanding the prehi toric agricultural system, we the we tern, outhern and eastern margins of Mt. Terevaka

Rapa NUl Journal 17 Vol. 16 (1) May 2002

Published by Kahualike, 2002 1 Rapa Nui Journal: Journal of the Easter Island Foundation, Vol. 16 [2002], Iss. 1, Art. 4 (Figure 1). expect that the technology of production and landscape use will Rapa Nui is located within the southern, temperate lati­ differ between the dryer Heki'i, Mahatua, Tongariki, Hanga tudes of the Pacific Ocean and a marked seasonality occurs each Poukura, Vinapu and districts and the Vaihu, Akahanga, year that is generally referred to as the rainy season and the dry Tepeu and ' districts that have sizable upland regions season. Rainfall averages over 1000 mm a year in the coastal within their boundaries. The Tahai and Vinapu district have ac­ zone near the town of but much of this rain (359 cess to moister areas near at the top of the Rano Kau volcano mm) falls in the wet season (April-June). During the warmest but the size of the productive zone is small. period of the year (December-February) an average of 240 mm of rain is reported at the Mataveri weather station. An analysis of these weather records (1972-1997) by Hunt and Lipo (2001) reports that the variation around the mean is great and reflected Rapa Nui by a high coefficient of variation (0.17-0.26). There appears to be no periodicity in these short-term rainfall records, and the occurrence of dry periods appears to be unpredictable. The high elevations on the i land surface and predominant Tepeu northwe t-southeast wind patterns create an orographic rainfall pattern. No rain gauges are present on the island other than at Akahanga the near ea-Ievel location of Mataveri (Figure 1) but a predic­ Tahai tive model by Honorato et al. (1991) proposes that up to double Ilanga Vaihu the amount of rainfall characteristic of coastal regions will fall Pokura in the upper regions of the three volcanoes beginning near the 200 meter elevation. The Poike region appears to have been o 5 used for agriculture as there is little evidence of prehistoric set­ tlement. As a result only about a sixth of the i land surface area Km benefits from the higher rainfall levels. Although the higher rainfall reduces the unpredictable nature of this upland region, Figure 2. Rapa Nui chiefdom districts circa. AD [500 (after Steven­ the arne model also predicts a decrease in temperature of 0.6­ son in press). 1.00 C for every 100 m ri e in elevation. Thi may have slowed crop maturation rates. Similarly, higher elevations may be asso­ THE RAPA NUl AGRICULTURAL SYSTEM ciated with greater wind velocities, which increase evaporation and thus reduce available moisture. Taken together with lower The prehistoric agricultural system on Rapa Nui relied exclu­ temperatures, these factors may partly offset the full benefit of sively on rainfall. Only a few years ago the location of Rapa increased rainfall amounts. Nui gardens and agricultural field systems could not be identi­ The past volcanic activity has also created the excessively fied by either archaeologists or the remaining indigenous popu­ drained soils present on the island today. The near surface soil lation. Only the frequently occurring circular stacked tone A-horizons are characterized by soil that have formed from ash structures (manavai) and planting wells (pu) were associated falls (Louwagie and Langohr 2002). As a result, it is highly po­ with agricultural production. Yet the dryland production system rou with about 50% of the volume con isting of air space. In that fed a population numbering in the thousands for nearly a some area of deeper soil, B-horizons tend to be clay-enriched millennia was assumed to have been large in scale and intensive and have a blocky pedogenic structure. Beneath this i a charac­ in form. In retrospect, this inability to find evidence of past agri­ teristically loose, regolithic soil with little pedogenic structure culture was a problem in land cape interpretation. The subtle (Wozniak 2000). This strata may occur almost directly under A­ indications of field systems had not been recognized. horizons in areas with little oil development. The clay-rich However, a series of surveys and test excavation programs soils may help to hold moisture at deeper levels but in shallow (Stevenson and Cristino 1986; Stevenson 1997; Stevenson and areas the lower C-horizon does little to retain soil moisture. In Haoa 1998; Stevenson et al. 1999; Wozniak 1998a, 1998b, general, the soils on Rapa Nui exacerbate the problem of limited 2000) has identified a variety of morphologically distinct agri­ moisture. cultural features found throughout the i land. Steven on et a1. These data show that while the entire island may be char­ (1999:803-804) note eight types of agricultural features and acterized as high risk for agriculture, there are some areas at these can be redefined as ix discrete types: higher elevations that may have been less unpredictable in terms of rainfall occurrence. The added moi ture above 200 m eleva­ 1) Mulched soils are areas with small (2 to 20 cm) rocks tion could have extended the period of wetter winter/spring intentionally mixed into the subsurface sediment to depths of 30 weather for weeks over that encountered near the coast. The to 50 cm. The source of the lithic mulch includes rock intention­ implications for agricultural production and the organization of ally broken off surrounding ba alt outcrops, and rock eroding production are significant. If the rainfall pattern is compared from the underlying bedrock. The lithic mulch would help to with the distribution of island political units (Stevenson in conserve moisture, ameliorate excess soil temperature fluctua­ press, Figure 2), we see that not all chiefdom districts had ac­ tions, and protect the garden oil from wind evapotransporation cess to the wetter interior regions. It is therefore reasonable to (Stevenson et a1. 1999:809).

Rapa Nui Journal 18 Vol. 16 (1) May 2002

https://kahualike.manoa.hawaii.edu/rnj/vol16/iss1/4 2 Correction' paoe 18 second h, third . . 0 ~. paragrap sentence should read "Th P.. agnculture as there IS little evidence ofprehistoric settlement" e oike regIOn appears not to have been used for - Stevenson et al.: Productive Strategies in an Uncertain Environment

2) Veneer surfaces con ist of areas were a relatively uni­ Heki'i District form 20 to 30 cm layer of rocks has been spread over the A compari on of the land cape between a dry region uch ground urface. The rocks are u ually from basalt outcrop in a the Heki'i Di trict and a moi ture upland area in the Aka­ the immediate vicinity that have been intentionally broken hanga/Vaihu District reveals everal imilaritie, a well as apart. The cover of rock would aid in rainfall permeating the striking difference ,in ettlement and method of farming. The surface, would reduce evaporation providing higher moisture archaeological urvey data and geomorphological oil a e s­ levels, and inhibit weeds (Stevenson et al. 1999:809). ment (Stevenson et al. 1996, 1997,2000) how that field sys­ tem covered a vast amount of the land cape in the north coast 3) Stacked boulder concentrations are areas where a ve­ Heki'i District (Stevenson and Haoa 1998; Steven on et al. neer of rock ha been spread over the surface but additional 1999) (Figure 3). The distribution of agricultural feature com­ tacking of rocks within the garden occurs. The stacked rocks prising different garden type can be u ed to define three gen­ form low windbreaks for the cultigens. Again, the rocks usually eral zone of production. derive from ba alt outcrops in the immediate vicinity that have Coa tal: Tbi zone i below 30 m elevation and i ituated been intentionally modified. The stacked boulder and a soci­ within 1000 m of the coast. On the heavily ettled coa tal plain, ated veneer would have the advantages of veneer surfaces, with the numerou dome tic settlement con i t of hou e , the additional benefit of creating windbreaks for the cultigens. houses, earth ovens, and are associated with small hou ehold gardens and larger field systems, both of which consist of 4) Pu are steep sided depressions, ca. 50 to 60 cm in diameter, located in rocky areas. The depressions extend to the oil layer below the rock, and with addi­ tional vegetative mulch would have pro­ vided a planting medium conducive to water retention. The rocky area are the re ult of people breaking apart basalt out­ crops or are natural lava flow .

Maunga Puko Puhi interior Valley 5) Manavai are 2 to 6 m long, 70 to N ...... 150 cm high, tacked stone garden enclo­ ure . The e would have protected fragile Figure 3. The dislribution of settlement and agricultural fields in the Heki 'i Dislrict plants such a from wind damage and reduced evaporation. manavai, veneer urfaces and tacked boulder concentrations. Dome tic ettlement tend to be located at the margin of their 6) Planting circles are formed by ring of mall tone 1.0 garden while the larger field y tern have few adjacent ettle­ to 1.5 m in diameter and define the perimeter of a planting pit. ment. With the addition of vegetative mulch, the pit would have pro­ vided a localized soil medium conducive for plant growth. Lowland inland: At a distance of approximately 1000 m from the coast at an elevation of ca. 50 to 80 m above sea level, The precise cultigen grown in these agricultural feature a large hilltop plantation i encountered on top of Maunga Puko is unknown, but we hypothesize that sweet potatoes were pum. Thi flat hilltop i heavily mulched with small basalt grown in the mulched soils; dryland , yams, and possibly stones that cover a shallow (50 cm deep) agricultural soil many weet potatoes were grown in the veneer surfaces and tacked hectare in ize. Settlement con ist primarily of low-density boulder concentration ; dryland taro and yam were grown in lithic tool catter that repre ent temporary dwellings. The the pu; banana and perhap young propagation plants were houses of elite per onnel and a small shrine (ahu) have also grown in the manavai; and dryland taro, yams, and sweet po­ been identified in thi zone. They are hypothesized to repre ent tato were grown in the planting circle . a managerial pre ence. These ix type of agricultural features are differentially di tributed throughout the island. In combination, these feature Lowland interior: Further inland at a distance of about 2 form the gardens and field systems that would have fed the sub­ km from the coa t and located on the northern edge of Maunga stantial prehistoric population. The once thought haphazard Puko Puhi lies an interior valley. Permanent domestic settle­ distribution of outcrops and boulder can now be een a well ment with a sociated veneer surface and stacked boulder con­ articulated and strategically placed growing area on the ter­ centration garden resume although the overall density of et­ rain. The investigation and interpretation of rock garden varie­ tlement is lower than the near coa tal region. The garden in­ tie permit a reconstruction of past agricultural technologie clude both mall hou ehold garden and larger field y tems. and land cape use. We compare the dry lowland Heki'i Di trict and the moi ter Akahanga/Vaibu (Vaitea) upland region to ex­ Akahanga/Vaihu District amine difference in ettlement and agricultural technology. An inten ive archaeological ettlement urvey ha been

Rapa Nui Journal 19 Vol. 16 (1) May 2002

Published by Kahualike, 2002 3 Rapa Nui Journal: Journal of the Easter Island Foundation, Vol. 16 [2002], Iss. 1, Art. 4 conducted in the Akahanga/Vaihu districts and has recorded -... .. many hundreds of archaeological features (Cristino et al. 1981; ,. ::. McCoy 1976). We have recently re-surveyed a portion of the ..., .. . o .. upland Vaitea area of the e districts (Stevenson et al. 2002). I,..•..... The result of the e urveys indicate three general zones of land .,,, • , u e. t, ...... Coastal: Thi zone con ists of the lowland coastal area, ••

extend inland ca. 1.75 kIn, and includes the lope up to an ele­ ••Oll vation of ca. 50 m above sea level. The settlement in this area is , . imilar to coastal zone in the Heki'i District and consists of per­ 0.' ~ '" .... c· ..' ... manent architectural features such as houses, habitation caves, • 1.8 ~" ... ., earth ovens, and chicken house (Cri tino et al. 1981, Stevenson / 1984, 1997). A number of maller household garden and larger ·'00 field y tern consi ting of manavai, veneer surface and ". .." .. tacked boulder concentration are known to exist but have not , been urveyed...... ~' Lowland inland: Thi zone extends from ca. 1.75 kIn to 4 ." kIn inland between elevation of 50 to 150 m. Settlement is , \. imilar to the coa tal zone, with the exception that the den ity of \ .' re idential and agricultural feature is considerably Ie .

Upland interior: The archaeological features in the upland " Vaitea urvey area extend from ca. 4 to 6 kIn inland, above an .:. 11 .."... "'. e:l » •• ,: •••l, .. -. elevation of 150 m, and consist of rectangular house founda­ '. ;4 :_:eA ~.1> .A, ,. " tion , rectangular platform of crude field stone, u-shaped truc­ "..' e. .~A ." ture , cleared walkways with stone borders, circular enclosure -,"eA with low perimeter wall and open hearths. Noticeably absent are features uch as larger hou e patios, chicken house, Figure 4. The distribution of settlement manavai and earth oven. In the Vaitea area, these urface fea­ (after eri tino et aJ. 1981). ture are a ociated with an undulating weathered lava flow con i ting of numerou ridges and swales defined by projecting ba alt outcrop . Prehi toric garden of varying dimen ions con- abandoned and agricultural production was conducted at lower i ting of veneer urface and tacked boulder concentrations elevation. are located within the walled enclosures and in the protected The u e of the land cape in the Heki'i District show a area at the ba e of slope and between the ba alt outcroppings. irnilar dynamic pattern of landscape u e. The earliest ettle­ The rectangular hou e are typically associated with the e agri­ ment wa al 0 along the coa t. Excavated ceremonial structures cultural fields. (ahu) on the north coa t in the adjacent 'Anakena District date to AD 900-1000 (SkjSJ}lsvold 1994). Domestic sites in the THE CHRONOLOGY OF LANDSCAPE USE coa tal plain were long-term and stable residence locations with occupational pan of over 500 years (AD 1300-1800) The body of chronological data in the form of radiocarbon and (Steven on et al. 1996). In contrast, the heavily mulched hilltop obsidian hydration date ha grown significantly in the last 15 field located about a kilometer inland were only used for ev­ year (Martino on-Wallin 1994; Stevenson 1984,2000; Steven- eral centune (ca. AD 1400-1700) and represent an effort to ig­ on et al. 2000). The context of these dates are not only from nificantly increa e production near the apex of chiefdom com­ ahu and domestic site but al 0 from agricultural features. Col­ plexity by the introduction of elite managed plantations lectively, these data allow us to see where the island landscape (Steven on et al. 1999). The e fields are in production at the wa occupied in the pa t and how it was modified. same time as the moister upland region and are a direct expre - A ettlement pattern model based on earlier work in the ion of an inten ified agricultural ystem. Further inland within Akahanga District and the upland interior Vaitea (Maunga Tari) an interior valley, initial ettlement wa early and began in the area propo ed an early (ca. AD 800) settlement for coa tal re­ AD 1100-1200 and continued until it too was abandoned in the gion and a gradual expan ion from core areas to other near early AD 1700 . From this time on, virtually all of the i land shore region. The building of new dwellings gradually ex­ population wa Living on the i land coastal plain. panded inland and the first forays into upland regions occurred The e data sugge t that the coa tal zone wa occupied con­ around AD 1250. By AD 1400-1500 the uplands were routinely tinuou Iy from initial ettlement of the island at ca. AD 800. utilized (Stevenson 1997) and continued to be farmed until the The relatively low rainfall in thi zone would reduce agricul­ early AD 1700 . In the early AD 1700s the upland region were tural productivity, but the zone had the advantage of being

Rapa Nui Journal 20 Vol. 16 (1) May 2002 https://kahualike.manoa.hawaii.edu/rnj/vol16/iss1/4 4 - Stevenson et al.: Productive Strategies in an Uncertain Environment clo er to marine resources. The lowland inland zone wa not ized a weathered and undulating lava flow for protection against inten ively u ed until much later, beginning around AD 1400 the wind. Millions of tones were intentionally brought together and extending to ca. AD 1700. This zone receive less rainfall to form a protective rock di tribution in veneer urface and than the upland and would have required inten ive amount of stacked boulder concentration . Our ob ervation note that the labor devoted to lithic mulching to make it agriculturally pro­ outcrop were al 0 di mantled to provide additional stone. ductive. The lowland interior zone apparently had greater agri­ These gardens were removed from the coa tal areas by at least 4 cultural potential using extensive techniques than had the low­ kilometers and would have involved ignificant travel time to land inland zone, and was utilized relatively early from ca. AD tend the fields and significant transport co t to move the weet 1100 to 1700. The upland interior zone receives the highest potatoes and taro to lowland hou ehold . rainfall, but the greater distance to the coast, the increa e in The hilltop plantations in the Heki'i Di trict were located wind velocity (thereby increasing evaporation and reducing the only about 1-2 Ian from the coast thereby reducing travel and amount of available moisture), and the lower temperatures made transport co ts. However, the level of effort inve ted in rock it Ie s desirable for agriculture and it wa not initially u ed until garden con truction wa ignifIcantly greater. Here, the rock AD 1250, and not extensively used until ca. AD 1400. cover is a deep lithic mulch of billion of mall tone over a shallow oil. Individual planting pit penetrated into regolithic DISCUSSIO C-horizon in an attempt to create a uitable growing environ­ ment with higher moisture content. The archaeological remains from the AkahangaIVaibu (Vaitea) It seem clear now that the development of labor-intensive area were initially used to develop the first regional settlement field systems correlates well with the occurrence of larger cere­ model for Rapa Nui (Stevenson 1997). It wa ba ed upon the monial complexes. In the early part of Rapa Nui prehistory we concept of the Hawaiian ahupua 'a where each ub-district hypothesize that extensive forms of agricultural production were population egment had acce to a strip of land from the coast practiced in the coa tal region (from AD 800), in the lowland to the interior and the full range of resource contained within interior (from AD 1100) and in the upland interior (from AD it. Implied in this model is that the resources available to each 1250) and wa probably ba ed on dryland taro cultivation. population segment were more or less equal. On Rapa Nui the Yield from this form of production would not have re ulted in variation in elevation and moi ture availability suggests that large urplu es. We propo e that intensive forms of agriculture orne chiefdom districts had more favorable habitats within in each of the e arne zone occurred after AD 1400 with the their boundarie . later introduction of the sweet potato. This tuber reached east It has been proposed that prior to chiefdom collapse in the Polyne ia around AD 1000 (Kirch et al. 1995) and was later 17 th century, the political center of each district was located near taken to outlier islands where its cultivation produced yields the largest ceremonial center (ahu). A cluster of chiefly hou es well above that required for ubsistence need . wa located immediately inland of the ahu. Habitation sites of The architectural trajectory on Rapa Nui appears to sup­ non-elite personnel were located inland of the coa tal sacred port our general cherne. Although large formal ahu are present zone up to an elevation of approximately 150 m. Settlement be­ by AD 1000-1100 (Skjl2ll void 1994) many of the largest truc­ came les frequent with increa ing elevation and eventually tures were built in the AD 1400 to AD 1600s (Martins on­ changed to the settlement characteri tic of the upland interior Wallin 1994). The presence of landform in two part of the Vaitea area. We now know that the rectangular houses repre- island devoted to intensive agricultural production beginning ented temporarily occupied dwellings, some of which may around AD 1400 and ending around AD 1700 i evidence to have been the residence of elite managers overseeing agricul­ support the proposition that corporate work effort required a tural production (Stevenson 1997; Stevenson et al. 1999). larger amount of food than could be produced at the household Archaeological urvey in the Heki'i District has increased level. However, this elevated production had to occur in an un­ our understanding of how the land cape wa utilized and re­ certain environment and required extraordinary modifications of quire that the initial model be revised to account for districts the land cape to be realized. We propo e that the return on labor that do not have the moi ter upland regions. The similarity in investment was very low and required an ideologically trong the coastal ettlement pattern between the Heki'i and Akahanga and strictly ranked hierarchy, one that emerged in only a few Districts suggests that the coastal zone in both areas included centurie after ettlement. hou ehold garden and lowland field systems. In contrast, fur­ ther inland on the edge of the coastal plain of the Heki'i Di trict REFERE CES were large and labor-intensive plantation on low hilltop that were managed by elite personnel and where non-elite occupa­ Cri tino, c., P. Vargas and R. lzuarieta. 1981. Atlas Arqueologi­ tions did not require permanent surface architecture. Further co de Isla de Pascua. Facultad de Arquitectura y Urbanis­ inland, low-density ettlement with permanent architecture and mo, Centro de Estudios, I la de Pa cua. Santiago, Chile. a ociated field sy tems resume. These latter two areas served Dunnell, R. 1999.The Concept ofWa te in Evolutionary Ar­ a the main production region in Heki'i District where surplu chaeology. Journal ofAnthropological Archaeology food upplies were generated. 18:243-250 Both the upland region and the hilltop plantations reflect Earle, T. 1997. How Chiefs Come to Power: The Political Econ­ intensive forms of agricultural production. The upland interior omy in Prehistory. University of Stanford Pre ,Stanford. Vaitea area consi t of rock gardens or field sy tems that util- Henley, J. and S. King. 1984. Late Quaternary Pollen Records

Rapa Nui Journal 21 Vol. 16 (1) May 2002

Published by Kahualike, 2002 5 Rapa Nui Journal: Journal of the Easter Island Foundation, Vol. 16 [2002], Iss. 1, Art. 4 from Ea ter Island. Nature 307:47-50 Stevenson, C. M. 1997. Archaeological Investigations on Ea ter Flenley, J., S. King, J. Jackson, C. Chew, J. T. Teller and M. E. Island, Maunga Tari: An Upland Agricultural Complex. Prentice. 1991. TheLate Quaternary Vegetational and Cli­ The Ea ter 1 land Foundation, Los Osos. matic Hi tory of Easter I land. Journal ofQuaternary Sci­ Stevenson, C. M. 2000. E timating Easter Island Ob idian Hy­ ence 6:85-115. dration Rates from Glass Composition, Easter Island Ar­ Grave, M. and T. N. Ladefoged. 1995. The Evolutionary Sig­ chaeology: Research on Early Rapa Nui Culture. C. M. nificance of Ceremonial Architecture in Polyne ia. Evolu­ Stevenson and W. S. Ayre , eds.:205-21O. The Ea ter I­ tionary Archaeology: Methodological Issues. P. A. Teltser, land Foundation, Lo 0 os. ed.:149-174. Univer ity of Arizona, Tuc on. Steven on, C. M. 2002. Easter Island Territorial Divi ion in Hommon, R. J. 200 I. The Emergence of Large-scale Society in the 16th Century. Pacific Landscapes. T. . Ladefoged Hawai'i. Pacific 2000. Proceedings ofthe Fifth Interna­ and M. Grave, eds. The Easter Island Foundation, Los tional Conference on Easter Island and the Pacific. C. M. o 0 . (in press) Steven on, G. Lee and F. J. Morin, ed .:141-149. The Stevenson, C. and C. Cri tino. 1986. The residential settlement Ea ter I land Foundation, Lo 0 0 . history of Easter Island's outhern coastal plain. Journal Honorato, R., A. Norero, J. Rodriguez and D. Suarez. 1991. ofNew World Archaeology 7:29-38 Ambiente Agrofisico y Fertilidad de Suelos. E tudio de la Stevenson, C. and S. Haoa. 1998. Prehistoric Gardening Sys­ Productividad Silvoagropecuaria del Fundo Vaitea, Isla de tems and Agricultural Intensification in the La Perouse Pascua. Humberto Benedetti R., ed.:25-43. Pontificia Uni­ Area of Easter Island. Easter Island in Pacific Context, ver idad Catolica de Chile, Facultad de Agronomia. South Seas Symposium. Proceedings ofthe Fourth Interna­ Hunt, T. L. and C. P. Lipo. 2001. Cultural Elaboration and En­ tional Conference on Easter Island and East . C. vironmental Uncertainty in Polynesia. Pacific 2000. Pro­ M. Stevenson, G. Lee and F. J. Morin, eds.:205-213. ceedings ofthe Fifth International Conference on Easter Easter Island Foundation, Lo Osos. Island and the Pacific. C. M. Stevenson, G. Lee and F. J. Stevenson, C., T. Ladefoged and S. Haoa. 2002. La Pro pecci6n Morin, eds.: 103-115. The Ea ter I land Foundation, Lo Arqueol6gica de Vaitea: Informe de Terreno 2001. Report 00. Submitted to the Con ejo de Monumentos Nacionale de Kirch, P. 1997. The Lapita Peoples. Blackwell, Oxford. Chile, Santiago. Kirch, P. 2000. On the Road ofthe Winds: An Archaeological Stevenson, c., J. Wozniak and S. Haoa. 1999. Prehi toric agri­ History ofthe Pacific Islands before European Contact. cultural production on Rapa Nui. Antiquity 73:801-812 Univer ity of California Pre , Berkeley. Steven on, c., J. Wozniak and S. Haoa. 2000. A Geomor­ Kirch, P., Steadman, D., Butler, V., Hather, J. and M. Wei . phological and Archaeological Study of Prehi toric Gar­ 1995. Prehistory and Human Ecology in Eastern Polyne ia: den in the La Perou e Area of Ea ter Island. Report ub­ Excavation at Tangatatau Rockshelter, Mangaia, Cook mitted to the Con ejo de Monumento Nacionales de Chile, I land. Archaeology in 30:47-65 Santiago. Ladefoged, T. N. 1995. The Evolutionary Ecology of Rotuman Steven on, c., S. Haoa, P. Mills, J.M. Ramirez and C. Gomez. Political Integration. Journal ofAnthropological Archae­ 1996. The La Perouse Archaeological Project, Report of the ology 14:341-358. 1995 Field Investigation. Report submitted to the Con ejo Ladefoged T. N. and M. W. Grave. 2000. Evolutionary Theory de Monumentos Nacionales de Chile, Santiago. and the Hi torical Development of Dry Land Agriculture in Steven on, c., S. Haoa , P. Wallin, H. Wallin-Martins on and J. North Kohala, Hawai 'i. American Antiquity 65:423-448. M. Ramirez. 1997. The La Perouse Archaeological Project: Louwagie, G. and R. Langohr. 2002. Testing Land Evaluation Report of the 1996 Field Investigations. Unpublished Re­ Method for Crop Growth on Two Soils of the La Perouse port. Area (Easter Island, Chile). Rapa Nui Journal (This i sue) Wozniak, J. A. 1998a. Settlement Patterns and Sub istence on Martin on-Wallin, H. 1994. Ahu - The Ceremonial Stone the Northwest Coast of Ea ter I land. Easter Island and Structures ofEaster Island. Analysis ofVariation and In­ East Polyne ian Prehi tory. P. Vargas c., ed.:171-178. terpretation ofMeanings. Aun 19. Upsalla, Sweden. Univer idad de Chile, Santiago. McCoy, P. C. 1976. Easter Island Settlement Patterns in the Wozniak, J. A. 1998b. Settlement Pattern and Sub i tence on Late Prehistoric and Protohistoric Periods. Bulletin No.5, the Northwest Coa t of Rapa Nui. Easter Island in Pacific International Fund for Monument, Inc., New York. Context, South Seas Symposium. Proceedings ofthe Fourth Skj~lsvold, A. 1994. Archaeological Inve tigation at International Conference on Easter Island and East Poly­ 'Anakena, Ea er I land. The Kon-Tiki Museum Occasional nesia. C. M. Steven on, G. Lee and F. J. Morin, ed .:185­ Skj~lsvold, Papers, Vol. 3. A. ed. The Kon-Tiki Mu eum 192. Easter Island Foundation, Lo 0 0 . Institute for Pacific Archaeology and Cultural Hi tory, Wozniak, J. A. 2000. Prehi toric Horticultural Practice on Oslo. Ea ter Island - Lithic Mulched Gardens and Field Sy terns. Steven on, C. M. 1984. Corporate De cent Group Structure in Rapa Nui Journal 13:95-99 Ea ter I land Prehistory. Ph.D. Di ertation, Department of Anthropology, The Penn ylvania State Univer ity, State College.

Rapa Nui Journal 22 Vol. 16 (1) May 2002 https://kahualike.manoa.hawaii.edu/rnj/vol16/iss1/4 6