Prehistoric and early historic agriculture at Maunga Orito, Easter Island (Rapa Nui), Chile.

Stevenson, Christopher M. ; Jackson, Thomas L. ; Mieth, Andreas 等

Introduction

The prehistoric economy of Rapa Nui relied primarily on the production of the tuber crops of sweet potato (Ipomoea batatas), dryland taro (Colocasia esculenta), yam (Dioscorea spp.), and ti (Cordyline). All of these crops, with the possible exception of sweet potato, arrived with the first immigrant population to Rapa Nui around AD 800 and were the staple food for this society over the next 1000 years. It has been proposed that sweet potato was a secondary introduction to the island sometime around the thirteenth century (Wallin et al. 2005). In part this is based on 1 sigma calibrated date ranges of AD 988-1155 and AD 1409-1440 that bracket sweet potato fragments found in a cave deposit on Mangaia, central Polynesia (Hather & Kirch 1991; for a discussion see Green 2005, and Ladefoged et al. 2005). These starchy tubers were planted in small gardens, open fields and intensively cultivated plantations in both the coastal and elevated regions of Rapa Nui (Stevenson & Haoa 1998; Stevenson et al. 1999, 2002). In general, people were dispersed among the agricultural fields in single family or multiple family hamlets and produced food for household needs. In remote field systems located away from domestic sites, surplus foods were probably produced to support communal ceremonies and temple building projects. In this paper, we examine the character of agriculture on Rapa Nui, based on the age and internal structure of one garden through the analysis of a 100m long soil profile located at the base of Maunga Orito (Figure 1).

[FIGURE 1 OMITTED]

Agriculture on Rapa Nui is dependent upon suitable soil conditions (see Ladefoged et al. 2005) and annual rainfall to supply moisture. Archaeological surveys have identified water diversion alignments but there is no evidence of irrigation (McCoy 1976; Stevenson 1997). In lowland coastal areas the average annual rainfall is c. 1100mm (Mataveri Station, Figure 1). However, deviations from the mean value can be significant. Within the 25-year recording period, dispersions from the mean value can be as much as 200mm and result in periods of high humidity or reduced moisture. In the latter case, this would have resulted in significant declines in crop production in open and unprotected settings. Hunt and Lipo (2001) note that moisture fluctuations do not appear to be predictable and therefore introduce a degree of uncertainty into agricultural production.

These difficulties are exacerbated by two additional factors. The first is the presence of a nearly continuous onshore wind that removes ground moisture by evapotranspiration. In very dry years, a high rate of evapotranspiration can even result in a moisture deficit for the year (Louwagie 2003). The second factor is the high soil moisture permeability in some areas (Louwagie 2003; Louwagie & Langohr 2002). These problems may not have existed during the early part of prehistory when the island was forested with palm and understory species (Flenley 1998; Flenley et al. 1991; Orliac 2000; Orliac & Orliac 1998). Swidden fields within a forest environment would have been well protected (Mieth & Bork 2003). However, with the extensive removal of trees and surface ground cover, higher evapotranspiration, reduced soil moisture retention and soil erosion became more significant impediments to a successful harvest.

Strategies for agricultural production

The prehistoric people of Rapa Nui would have quickly recognised the impacts of deforestation and rainfall variability on the productive potential of the agricultural system. These limiting and unpredictable factors would have introduced an uncertainty about the outcome of farming from year to year. As a result, the Rapa Nui implemented a set of agricultural innovations to counteract these negative effects. These developments included the use of a surface lithic mulch to facilitate water permeability, significantly reduce evaporation, and protect the soils from erosion. In addition, rocks were placed within gardens as protection from wind, and gardens were placed at the base of slopes that benefited from surface water runoff (Stevenson et al. 1999, 2002; Wozniak 1998, 1999, 2001; Bork et al. 2004) (Figure 2). Excavations conducted beneath these surface rock distributions have documented an anthropogenic soil horizon 20-80cm deep with casts of humanly created planting pits preserved within the upper anthropogenic soil horizon and penetrating the underlying B-Horizon.

[FIGURE 2 OMITTED]

Many questions about the Rapa Nui farming system remain unanswered. In this paper, we will look at two fundamental aspects. A basic set of questions concern when were stone-based technological innovations implemented and how long did they continue to be used. It is our hypothesis that the application of lithic mulch would have appeared as deforestation severely impacted the landscape. This is estimated from the Rano Kau pollen cores to have been critical in the thirteenth century (Flenley 1998) and to have occurred on Poike peninsula between AD 1250 and AD 1450 (Mieth & Bork 2003).

A second set of questions concerns the internal structure of gardens. The rock gardens are highly variable with regard to the size and density of stone but they fall into four basic classes. Lithic mulch gardens consist of 2-20cm diameter rocks that were anthropogenically worked into the upper soil horizon while veneer surfaces are distributions of rock that were placed on the ground surface. The veneer surfaces may also include stacked boulder concentrations that provide a buffer against the wind. Lastly, pu are depressions within deep accumulations of stone that are frequently found at the base of hill slopes. Stevenson et al. (2002) hypothesised that different garden classes were used for different types of crops. Lithic mulch gardens were used for shallow rooted crops such as sweet potatoes while veneer surfaces and boulder gardens covered deeply rooted crops such as taro, yam and ti but may have been used for sweet potato as well. Deep rock accumulations (pu) may have also been preferred for taro and yam. This crop assignment, however, is hypothetical and requires a direct evaluation.

The Orito garden

The ancient garden reported in this paper is located at the base of the south-western slope of Maunga Orito adjacent to the road to Vinapu that parallels the airport runway (Figure 3). Maunga Orito is a steeply sided cone with a mid-section slope of approximately 30 degrees. Numerous habitation sites along with a few small shrines and statues are present on the slopes and at the base (Figure 3, Table 1) (McCoy 1976). It is also the location of one of the four obsidian outcrops that were extensively exploited in prehistory. Open pit mines reflecting this activity are located about 100m upslope. Recent soil mining at the base of the hill has created a large quarry pit with a north-west--south-east oriented soil profile that is over 100m in length (Figure 3, Figure 4). This soil removal exposed a stratigraphic profile that incorporated an anthropogenic horizon formed by prehistoric cultivation. Although we refer to this area as a domestic garden, our initial observations noted cultural features and charcoal lenses at varying depth, which suggested that multiple cultivation events had occurred over time and space.

[FIGURES 3-4 OMITTED]

The soil profile was investigated on three occasions. In early 2002, Mieth and Bork sketched a 65m section of the profile and selected radiocarbon samples from cultural features with the goal of documenting prehistoric land use activities and the resulting phenomena of soil erosion/deposition in this area. Later in the year, after further mechanical soil extraction with a front-end loader from Hanga Roa, the site was visited by Stevenson and Jackson and additional work was conducted. In this effort, a 122m section of the profile was scraped, profiled, and drawn to scale. A level reference line and 30m tapes were used to control for distance and relative elevation (Figure 5). Agricultural planting pits within an anthropogenic soil horizon and cultural hearths and living surfaces within a colluvium were identified, drawn and photographed. The rock distributions at the current surface were also recorded along the entire extent of the soil profile. A final visit was made in late 2004 by Stevenson, Ladefoged, Mieth, and Bork to re-evaluate the stratigraphic profile. It was during this visit that a buried garden horizon was encountered at the far eastern end of the profile and sampled for datable carbon and obsidian fragments. We have combined the results of the three investigations to provide a more complete picture of past agricultural activity in this area.

[FIGURE 5 OMITTED]

General stratigraphic interpretation

The profile is almost parallel to the slope contour lines and is oriented in a north-west--south-east direction. The profile has a depth of up to 3.5m. It was divided into four sections, lettered A-D (NW-SE) each about 30m long. The base of the profile (Cv-Horizon) consists of dark reddish-brown (2.5 YR 3/4), naturally reworked volcanic material. In parts of the profile two different Cv-Horizons could be distinguished (I-Cv und II-Cv).

Above the reworked volcanic stone is a B-Horizon, up to 2m in thickness and rich in clay. This soil horizon is defined by a dark greyish-brown colour (10 YR 3/2). The soil has a polyedric to subpolyedric structure. Radiocarbon dating of the humus fraction demonstrated an average age of at least 5700 BC (Table 2). The most striking feature in this soil is the remnant of root casts from the Jubaea palm (Figure 6). Numerous root channels vertically traverse the B-Horizon and protrude into the C-Horizon. The palm root casts were found along the entire extent of the profile. In some segments of the profile the root cast formation allows the differentiation of single palm trees as well as the basal trunk diameters and growing distances (Figure 7). The root casts at Orito show that dense palm vegetation once existed in this area, and are similar to the pattern of palm root casts described for the Poike peninsula (Mieth & Bork 2003).

[FIGURES 6-7 OMITTED]

In most parts of the profile the upper boundary of the B-Horizon is evidenced by irregular structures and an abrupt change in the soil texture. Without doubt this is evidence for an early gardening horizon (Aga). The irregular structures are planting pits that have diameters of 10cm to 70cm and protrude from the Aga-Horizon into the B-Horizon (Figures 8-12). Very similar structures were analysed on the south-west Poike peninsula by Mieth and Bork (2003). A clear identification of the Aga-Horizon and planting pits is possible only when the profile is freshly exposed and moist (Figure 5).

[FIGURES 8-12 OMITTED]

Our observations allow us to infer that the original forest ground surface and A-Horizon (the pre-colonisation land surface present c. AD 800) is not present in the profile. This was caused by the human removal of the palms and destabilisation of the land surface. A few centimetres to decimetres of the B-Horizon were eroded before a colluvium was deposited onto the surface from upslope locations. This process may have been augmented by gardening activity that would have penetrated the B-Horizon during the excavation of planting pits and increased the depth of the Aga-Horizon. Evidence for this is inferred from an erosional nonconformity in the upper part of palm root cones found at the north-western edge of the profile (Figure 7).

Above the Aga/B-Horizon interface is a very dark grey (10 YR 3/1) colluvium with a maximum thickness of about 80 centimetres. The sharp discontinuity with the B-Horizon, the loose structure, and finally, the lack of palm root casts in this layer indicate that the colluvium represents a deposition from A- and B-Horizons that were eroded from upslope locations. These soils have been reworked over time by agricultural activity. Embedded in the colluvium are numerous cultural features such as obsidian flakes, charcoal fragments, and the remains of hearths represented by charcoal and ash lenses. Radiocarbon dating of the organic features and obsidian hydration dating allows a chronological assessment of the erosion and deposition processes at the base of Maunga Orito. The vertical distribution of the carbon lenses and hearths demonstrates the chronological sequence of various land use phases (Figures 8-12).

After forest clearance, the area was covered by sediments derived from upslope locations and the present surface represents late prehistoric/early historic land use in this area. This phase saw the placement of basalt stones on the surface of the garden. Veneer surfaces with basalt stones of 10cm to 20cm in diameter, and boulder concentrations with stones of 30cm to 40cm in diameter cover large segments of the current surface on top of a younger colluvial Aga-Horizon. Almost no stones are embedded in the colluvium.

The lack of stones within the profile suggested that a long period of weathering and down slope movement had not occurred and that the surface rocks had been intentionally placed. This is also supported by our inspection of the upslope basalt outcrop and surrounding stones and boulders where numerous linear alignments and stacked stone windbreak features are present. These arrangements reflect the purposeful creation of a garden surface and support our contention that the adjacent veneer surface and boulder distribution at lower locations was intentionally created for agricultural purposes.

Chronological interpretation

Radiocarbon dating

Five radiocarbon samples were recovered from the profile in Section A. These samples were used principally to provide age estimates for the near-surface colluvium. Four samples were submitted to the Leibniz-Laboratory at the Christian-Albrechts-University of Kiel, Germany and one to Beta Analytic, Miami, Florida (Table 2). Samples KIA 17117 and 17118 were single charcoal pieces in the colluvium located within the Aga-Horizon. Sample 17116 was from a hearth in the colluvium 40cm below the present surface. Also, sample Beta-178860 was from a hearth (F101A). By this stratigraphic sequence the samples postdate the planting pits located below. These features contain carbonised wood fragments from open fires used over a limited time range. Sample KIA 17120 was taken from undisturbed soil (A/B-Horizon) below the colluvium and represents the average humus age for this soil. Two small radiocarbon samples were selected from within a planting pit feature within the buried Aga-Horizon in Section D (Figure 13). This context precedes the later colluvial build up dated in Section A. The fragments were suitable for AMS dating and were submitted to Beta Analytic (Table 2). Another charcoal sample was taken from a planting pit in the buried Aga-Horizon 90 centimetres below the present surface. This sample was dated in the Leibniz-Laboratory, Kid, Germany (Table 2).

[FIGURE 13 OMITTED]

All of the radiocarbon assays on cultural carbon samples from Section A had multiple intercepts on the calibration curve. However, the probabilities associated with each of the intercepts show the greatest likelihood of the true age in three of the samples (KIA 17116, 17117, Beta-178860) to occur in the middle to late AD 1700s. The fourth date (KIA 17118) suggests later activity in the nineteenth century. Two AMS samples from Section D (Beta 196925, 196926) had only single intercepts on the calibration curve and both samples dated to the middle of the fifteenth century with 2-sigma calibrations of AD 1420-1490 and AD 1410-1480. The third AMS sample (KIA 25975) from Section D had two intercepts on the calibration curve and is of younger age. The 2-sigma calibrated age is AD 1477-1531 and 1545-1635.

Obsidian hydration dating was used to assess the ages of colluvial deposits, agricultural planting pits, and the buried Aga-Horizon. Dates from the cultural features, other than planting pits, tend to be after AD 1500 and extend into the middle 1600s (Table 3). There is one eleventh century date that likely represents redeposition from another context. The dates from the planting pits and buried soil horizon range from AD 904 through AD 1677 but the vast majority (18 of 21) occur before AD 1500. The later dates from the non-agricultural cultural features are generally associated with the late radiocarbon dated habitation in Section A while most of the earlier dates come from the planting pits in Sections B, C, and D.

Discussion

The radiocarbon and hydration dating data demonstrate the chronological sequence of different land use phases during the active period of erosion and deposition in this area. The first phase of agricultural activity after forest clearance took place in the twelfth century. Farming continued up to the seventeenth century with the most evidence for agricultural activity in the area during the fourteenth and fifteenth centuries. It is toward the end of this period that the early planting structures were buried by colluvium that preserved the agricultural field sections from later disturbance. Some areas have evidence of habitation features (e.g. hearths, a storage pit, carbon lenses) within the colluvium. These features mark a post-agricultural period that was indicated by deposition in the downslope area caused by erosion in the midslope and upslope area. Quarrying activity for obsidian extraction might have triggered this erosion. About 150-200 years ago, a veneer pavement and boulder garden were added to the ground surface. The process of erosion and deposition appears to have slowed since the surface stones are not covered with sediment. This likely coincides with the decline in obsidian mining.

The prehistoric settlement pattern in the vicinity of Orito is now fragmentary as a result of modern farming and the construction of roads and an airport. However, the remnant archaeological sites show that habitation sites are numerous within the vicinity of the obsidian quarry (Table 1). Obsidian reduction seems to have been largely confined to the quarry proper since few outlier workshops are present. The mechanically excavated soil profile exposed a portion of a habitation site likely to be Site 4-17 (Figure 3). Numerous carbon lenses and hearths, a storage pit and planting pits reflect human occupation in this area. The occurrence of these features in the profile reveals the chronological development of land use. Radiocarbon samples from a portion of these features suggest that this dynamic lasted until the middle to late AD 1700s, and may have extended into the nineteenth century. The planting pits below the colluvium document the oldest period of land use in this area that date between the twelfth through fifteenth centuries. However, their stratigraphic location below the habitation site and within an intact and earlier Aga-Horizon show that a period of garden activity preceded other types of land use.

Our inspection of the soil profile documents at least three garden forms. They include:

1) Extensive garden activity that occurred after forest clearance. This period is characterised by well-differentiated planting pits at the base of the early Aga-Horizon without any forms of stone cover;

2) Garden activity at the end of the erosion/deposition period, represented by a surface veneer pavement, and a strongly homogenised colluvial Aga-Horizon;

3) Garden activity represented by surface boulder gardens on top of the colluvial AgaHorizon.

Veneer pavements and boulder gardens have been previously documented (Stevenson et al. 1999, 2002). At Maunga Orito, the planting pits within the Aga-Horizon have no relationship to the younger stone structures at the current surface. Obsidian hydration dating indicates that all of the pits dated to before AD 1700 and are not contemporary with the eighteenth century rock garden. Lithic mulch consisting of high densities of small stone worked onto the upper parts of the cultivated soil horizon was not present in any section of the profile.

The shapes and dimensions of the planting pits within the Aga-Horizon suggest the cultivation of a variety of cultigens. Shallow rooted crops such as sweet potato, sugar cane and gourds were likely planted in shallow pits near the surface. Areas with deep planting pits would have likely been planted with taro, yam or ti. Two forms of deep planting pits are present and consist of conical and sub-rounded forms. These are not rigid pit types as there is significant variation between each form. The conical forms often have asymmetrical sides, which indicate that a plant was extracted by excavation from one side. We hypothesise that these were the location of taro plants. The pits with a sub-rounded base tend to be wider across the midsection. The tops of these pits are also asymmetrical in form that was likely created during extraction. The larger form and sub-rounded base, however, would have been created during preparation of the soil for planting. This larger planting pit type may have held yams, which require more moisture than taro (Louwagie 2003). All of these interpretations require evaluation through additional analyses that could include phytolith, starchy and pollen studies.

The location and frequency of deep planting pit structures suggest that the first gardens were installed after the clearance of the palm forest and may have been cultivated only extensively. The casts of the single planting events are well preserved within the B-Horizon and penetrate into the palm root zone. It could be interpreted that extended use of the garden for many centuries was unlikely because there is rarely any overlap or intersection of deep planting pits with one another. In intensively cultivated fields, the spacing of deep pits can be much closer and the upper portion of the B-Horizon has been reworked except for small pockets or islands of subsoil (Stevenson & Haoa 1998). However, the obsidian hydration dates show a 300-year period of garden cultivation. It is possible that planting pits were an exception to the general rule of planting in surface mounds or simply within the upper 30-40cm of the soil. Alternately, pits may have been generally shallow and the profile has preserved the less frequent and deeper pits that may have been created in drier years. The chronological assessments indicate that the garden was abandoned in the fifteenth century. This is clearly the case in Section D where the soil horizon was not impacted by later agriculture. Why was this the case? The deep soils and partially protected setting of the garden at the base of Maunga Orito would have helped deter moisture loss promoted by the on-shore winds. However, we hypothesise that open gardens became less attractive as aridity increased from widespread landscape deforestation and that farming shifted to lithic-mulched field systems. In the area of Orito large supplies of stone are not present and may have limited the implementation of lithic mulch technology. Approximately 300 years later gardening resumed in the area and a veneer pavement and low-density boulder garden were installed. In this post-collapse period of the island chiefdom hierarchy (after AD 1680) high yields may not have been required and gardening at Maunga Orito reflects the household level of production.

Conclusion

The agricultural system of Rapa Nui is remarkably varied and complex with many subtle changes in technology and use of the landscape over time. The garden at Maunga Orito is but one small piece of this sequence. The events here begin with a period of deforestation and erosion that is followed by a period of open-field gardening prior to development of the rock garden techniques. Plantings appear to have been spatially extensive. In the early phases of island occupation, when population numbers were lower, demands on the agricultural system were likely modest and may not have exceeded a household level of production.

As population growth and island deforestation continued, we hypothesise that the conditions of production changed significantly by the fourteenth century. On an exposed landscape, the added evapotranspiration and wind velocities reduced available moisture and overall levels of productivity. Farming was intensified by the creation of lithic-mulched fields. The open and rock-flee field at Maunga Orito, and similar areas, were abandoned in favour of more protected environments. The extensive rock gardens in all parts of the island testify to the adoption of this strategy (Stevenson et al. 1999; Wozniak 1999). The demands for production above the household level also began by the thirteenth century as larger scale ceremonial centres were constructed in many of the island territorial segments (Martinsson-Wallin & Wallin 2000). These added demands on production within a higher risk environment probably promoted a more rigid hierarchy that was involved in the direct monitoring of farming activities. This is inferred from the presence of chiefly structures and small temples located amidst remote field systems at higher elevations (Stevenson et al. 2005).

This degree of control is not seen at the Maunga Orito garden in that it appears to have been little used in the sixteenth to seventeenth centuries. Only after the period of hierarchical collapse and termination of megalithic temple construction does farming resume at this location. However, concerns about successful production were still on the minds of Rapa Nui farmers who continued to construct veneer pavements and boulder gardens to protect their plants.

Acknowledgments

We thank Sonia Haoa for her help and support. We are also grateful to St. Enrique Pakarati, Governor, Rapa Nui, the Consejo de Desarollo, Rapa Nui, and to Angel Cabeza, Consejo de Monumentos Nacionales de Chile, Santiago, for permission to conduct research on Easter Island. We extend our appreciation to the Earthwatch Institute and the many Earthwatch volunteers whose participation funded this research and resulted in its successful completion. Patrick McCoy kindly provided the site identifications for the survey map from his original field notes. Finally, we are grateful to the teams of the Leibniz-Laboratory, Kiel, Germany and Beta Analytic, Miami, Florida for the radiocarbon dating.

Received: 23 May 2005; Accepted: 20 October 2005; Revised: 12 December 2005

References

BORK, H.-R., A. MIETH & B. TSCHNOCHNER. 2004. Nothing but stones? A review of the extent and technical efforts of prehistoric stone mulching on Rapa Nui. Papa Nui Journal 18(1): 10-14.

FLENLEY, J.R. 1998. New data and new thoughts about Rapa Nui, in C.M. Stevenson, G. Lee & F.J. Morin (ed.) Easter Island in Pacific context, south seas symposium. Proceedings of the fourth international conference on Easter Island and east Polynesia: 125-8. Los Osos (CA): Easter Island Foundation, Bearsville & Cloud Mountain Presses.

FLENLEY, J.R., S.M. KING, J. TELLER, M. PRENTICE, J. JACKSON & C. CHEW. 1991. The late Quaternary vegetational and climatic history of Easter Island. Journal of Quaternary Science 6:85-115.

GREEN, R.C. 2005. Sweet potato transfers in Polynesian prehistory, in C. Ballard, E Brown, R.M. Bourke & T. Harwood (ed.) The Sweet Potato in Oceania: A Reappraisal: 43-62. Sydney and Pittsburgh: Oceania Monograph.

HATHER, J.G. & P.V. KIRCH. 1991. Prehistoric sweet potato (Ipamoea batatas) from Mangaia Island, central Polynesia. Antiquity 65: 887-93

HUNT, T.L. & C.P. LIPO. 2001. Cultural elaboration and environmental uncertainty, in C.M. Stevenson, G. Lee & EJ. Morin (ed.) Pacific 2000, proceedings of the fifth international conference on Easter Island and the Pacific. 103-16. Los Osos (CA): Easter Island Foundation, Bearsville & Cloud Mountain Presses.

LADEFOGED, T.N., C.M STEVENSON, P. VITOUSEK O. CHADWICK. 2005. Soil nutrient depletion and the collapse of Rapa Nui society. Rapa Nui Journal 19(2): 100-105.

LADEFOGED, T.N., M.W. GRAVES & J. COIL. 2005. The introduction of sweet potato in Polynesia: Early sweet potato remains in Hawai'i. Journal of the Polynesian Society 114(4): 359-73.

LOUWAGIE, G. 2003. Palaeo-environment reconstruction and evaluation based on land characteristics on archaeological sites. Case Study I: Verrebroek "Dok" and Doel "Duerganckdok", Case Study II: Easter Island. Ph.D. Dissertation, University of Ghent, Belgium.

LOUWAGIE, G. & R. LANGOHR. 2002. Testing land evaluation methods for crop growth on two soils of the La Perouse area (Easter Island, Chile). Papa Nui Journal 16: 23-8.

MARTINSSON-WALLIN, H. & P. WALLIN. 2000. Ahu and settlement: Archaeological excavations at 'Anakena and La Perouse, in C.M. Stevenson & W.S. Ayres (ed.) Easter Island Archaeology: Research on Early Rapanui Culture: 27-44. Los Osos (CA): Easter Island Foundation, Bearsville Press.

McCoy, P.C. 1976. Easter Island settlement patterns in the Late Prehistoric and Protohistoric periods. New York (NY): Bulletin 5, International Fund for Monuments Inc., Easter Island Committee.

MIETH, A. & H.-R. BORK. 2003. Diminution and degradation of environmental resources by prehistoric land use on Poike Peninsula, Easter Island (Rapa Nui). Papa Nui Journal 17(1): 34-41.

ORLIAC, C. 2000. The woody vegetation of Easter Island between the early fourteenth and mid-seventeenth centuries AID, in C.M. Stevenson & W.S. Ayres (ed.) Easter Island Archaeology: Research on Early Rapanui Culture: 211-20. Los Osos (CA): Easter Island Foundation, Bearsville Press.

ORLIAC, C. & M. ORLIAC. 1998. The disappearance of Easter Island's forest: over exploitation or climatic catastrophe, in C.M. Stevenson, G. Lee & F.J. Morin (ed.) Easter Island in Pacific context, south seas symposium. Proceedings of the fourth international conference on Easter Island and east Polynesia: 129-34. Los Osos (CA): Easter Island Foundation, Bearsville & Cloud Mountain Presses.

STEVENSON, C.M. 1997. Archaeological investigations on Easter Island. Maunga Tari: an upland agricultural complex. Los Osos (CA): Easter Island Foundation, Bearsville & Cloud Mountain Presses.

STEVENSON, C.M. & S. Haoa. 1998. Prehistoric gardening systems and agricultural intensification in the La Perouse area of Easter Island, in C.M. Stevenson, G. Lee & F.J. Morin (ed.) Easter Island in Pacific context, south seas symposium. Proceedings of the fourth international conference on Easter Island and east Polynesia: 205-13. Los Osos (CA): Easter Island Foundation. Bearsville & Cloud Mountain Presses.

STEVENSON, C.M., J. WOZNIAK & S. HAOA. 1999. Prehistoric agricultural production on Papa Nui. Antiquity 73: 801-12.

STEVENSON, C.M., T. LADEFOGED & S. HAOA. 2002. Productive strategies in an uncertain environment: prehistoric agriculture on Easter Island. Rapa Nui Journal 16(1): 17-22.

STEVENSON, C.M., T.N. LADEFOGED, S. HAOA & A.T. GUERRA. 2005. Managed agricultural production in the Vaitea region of Rapa Nui, Chile, in C.M. Stevenson, J.M. Ramirez, F. Morin & N. Barbacci (ed.) VI International Conference on Easter Island and the Pacific. 125-36. Los Osos (CA): Easter Island Foundation, Bearsville Press.

WALLIN, P., T. LADEFOGED & C.M. STEVENSON. 2005. Sweet potato production on Rapa Nui, in C. Ballard, P. Brown, R.M. Bourke & T. Harwood (ed.) The Sweet Potato in Oceania: A Reappraisal: 85-8. Sydney and Pittsburgh: Oceania Monograph.

WOZNIAK, J.A. 1998. Settlement patterns and subsistence on the northwest coast of Rapa Nui, in C.M. Stevenson, G. Lee & F.J. Morin (ed.) Easter Island in Pacific context, south seas symposium. Proceedings of the fourth international conference on Easter Island and east Polynesia: 185-92. Los Osos (CA): Easter Island Foundation, Bearsville & Cloud Mountain Presses.

--1999. Prehistoric horticultural practices on Easter Island-lithic mulched gardens and field systems. Rapa Nui Journal 13: 95-9.

--2001. Landscapes of food production on Easter Island: successful subsistence strategies, in C.M. Stevenson, G. Lee & F.J. Morin (ed.) Pacific 2000, proceedings of the fifth international conference on Easter Island and the Pacific: 91-102. Los Osos (CA): Easter Island Foundation, Bearsville & Cloud Mountain Presses.

Christopher M. Stevenson (1), Thomas L. Jackson (2), Andreas Mieth (3), Hans-Rudolf Bork (4) & Thegn N. Ladefoged (5)

(1) Virginia Department of Historic Resources, 2801 Kensington Avenue, Richmond, VA 23221, USA (Email: christopher.stevenson@dhr.virginia.gov)

(2) Pacific Legacy, 1525 Seabright Ave., Santa Cruz, CA 95062, USA (Email: legacytj@yahoo.com)

(3) University of Kiel, Ecology Centre, Olshausenstrasse 40, D 24098 Kiel, Germany (Email: amieth@ecology.unikiel.de)

(4) University of Kiel, Ecology Centre, Olshausenstrasse 40, D 24098 Kid, Germany (Email: hrbork@ecology.unikiel.de)

(5) Department of Anthropology, University of Auckland, Private Bag 92019, Auckland, New Zealand (Email: t.ladefoged@auckland.ac.nz)

Table 1. Archaeological sites at Maunga Orito

 Site classification Site numbers (Quadrangle 4)

Alignment 66, 73, 75
Shrine (ahu) 74, 80
Chicken house (hare moa) 14, 24
Circular house 33
Earthovcn (umu pae) 21, 36
Habitation 2, 3 (not shown), 4, 6-19, 21-33, 34-35,
 37-39, 44-48, 50-53, 55-64, 66-70, 76-79
Obsidian quarry 5, 32, 29
Petroglyphs 1, 49
Statue 71-72, 77
Stone cairn (pipi horeko) 41-43, 54, 65
Terrace 40
Stone tower (tupa) 20

* Site 4-3 was destroyed by the soil barrow pit.

* Repeated number entries represent sites with multiple features.

Table 2. Radiocarbon dates from the south-
western downslope area of Maunga Orito

 Lab No. Provenience Context

KIA 17116 HK 7 IPC M. Orito 1 Charcoal from hearth,
 0.40m below present
 surface within colluvium

KIA 17117 HK 6 IPC M. Orito 1 Charcoal within
 colluvium, 0.63m
 below present surface

KIA 17118 HK 8 IPC M. Orito 1 Charcoal within colluvium,
 0.13m below present
 surface

KIA 17120 M1 IPC M. Orito 1 Soil substrate from
 the Aga/B-Horizon taken
 1.2m below the present
 surface

KIA 25975 Section D-Buried Aga- Charcoal from planting pit
 Horizon (Figure 13, #3)

Beta-178860 Feature 101A Charcoal from hearth

Beta 196925 Section D-Buried Aga- Charcoal from planting pit
 Horizon (Figure 13, #1)

Beta 196926 Section D-Buried Aga- Charcoal from planting pit
 Horizon (Figure 13, #2)

 Lab No. Radiocarbon age Calibrated ages

KIA 17116 BP 177 [+ or -] 18 Cal AD 1674, 1777,
 1800, 1941, 1954

KIA 17117 BP 148 [+ or -] 18 Cal AD 1682, 1734,
 1807, 1930, 1954

KIA 17118 BP 114 [+ or -] 21 Cal AD 1696, 1725, 1814,
 1834, 1877, 1916, 1950

KIA 17120 BP 6607 [+ or -] 107 Cal BC 5600, 5593,
 5556, 5552, 5533

KIA 25975 BP 347 [+ or -] 21 Cal AD 1516, 1598, 1618

Beta-178860 BP 200 [+ or -] 50 Cal AD 1668, 1782, 1795

Beta 196925 BP 450 [+ or -] 40 Cal AD 1440

Beta 196926 BP 460 [+ or -] 40 Cal AD 1440

 Calibrated ages 2-sigma
 Lab No. range (Probability 95.4%)

KIA 17116 Cal AD 1665-1683 (15.3%)
 Cal AD 1733-1784 (47.7%)
 Cal AD 1789-1808 (13.4%)
 Cal AD 1928-1955 (19.1%)

KIA 17117 Cal AD 1672-1699 (16.2%)
 Cal AD 1723-1779 (38.2%)
 Cal AD 1798-1814 (9.5%)
 Cal AD 1832-1879 (12.4%)
 Cal AD 1915-1944 (19.1%)

KIA 17118 Cal AD 1682-1734 (28.6%)
 Cal AD 1807-1903 (55.3%)
 Cal AD 1905-1930 (11.4%)

KIA 17120 Cal BC 5761-5367 (95.4%)

KIA 25975 Cal AD 1477-1531 (35.3%)
 Cal AD 1545-1635 (60.1%)

Beta-178860 Cal AD 1533-1539 (4.0%)
 Cal AD 1636-1714 (26.8%)
 Cal AD 1715-1888 (58.3%)
 Cal AD 1910-1950 (14.5%)

Beta 196925 Cal AD 1430-1490 (100%)

Beta 196926 Cal AD 1410-1480 (100%)

Table 3. Obsidian hydration dates for the Orito Garden

Lab No. Provenience ABS Rim (um) Density EHT

DL-2005-1 Pit IOTA, Hearth 0.1325 1.22 2.3900 21.5
DL-2005-3 Pit IOTA, Hearth 0.1396 1.28 2.3900 21.5
DL-2005-2 Pit IOTA, Hearth 0.1560 1.43 2.3900 21.5
DL-2005-4 Pit 137 (Storage) 0.1607 1.48 2.3900 21.5
DL-2005-5 Fea. 2 (Hearth) 0.2287 2.10 2.3900 21.5
DL-2005-6 Pit 57 0.1265 1.16 2.3900 21.5
DL-2005-7 Pit 51 0.1396 1.28 2.3900 21.5
DL-2005-8 Pit 74 0.1570 1.44 2.3900 21.5
DL-2005-9 Pit 22 0.1630 1.50 2.3900 21.5
DL-2005-10 Pit 66 0.1730 1.59 2.3900 21.5
DL-2005-11 Pit 74 0.1870 1.72 2.3900 21.5
DL-2005-12 Pit 94 0.1911 1.76 2.3900 21.5
DL-2005-13 Pit 68 0.1925 1.77 2.3900 21.5
DL-2005-14 Pit 65 0.1959 1.80 2.3900 21.5
DL-2005-15 Pit 30 0.1982 1.82 2.3900 21.5
DL-2005-16 Pit 59 0.1982 1.82 2.3900 21.5
DL-2005-17 Pit 70 0.2099 1.93 2.3900 21.5
DL-2005-18 Pit 1 0.2163 1.99 2.3900 21.5
DL-2005-19 Pit 12 0.2474 2.27 2.3900 21.5
DL-2005-20 Pit 15 0.2732 2.51 2.3900 21.5
DL-2005-21 Buried Aga 0.1825 1.68 2.3900 21.5
DL-2005-22 Buried Aga 0.1830 1.68 2.3900 21.5
DL-2005-23 Buried Aga 0.1949 1.79 2.3900 21.5
DL-2005-24 Buried Aga 0.1998 1.84 2.3900 21.5
DL-2005-25 Buried Aga 0.2007 1.84 2.3900 21.5
DL-2005-26 Buried Aga 0.2054 1.89 2.3900 21.5

Lab No. %RH/100 OH- A E

DL-2005-1 0.98 0.097 1.03 85699
DL-2005-3 0.98 0.097 1.03 85699
DL-2005-2 0.98 0.097 1.03 85699
DL-2005-4 0.98 0.097 1.03 85699
DL-2005-5 0.98 0.097 1.03 85699
DL-2005-6 0.98 0.097 1.03 85699
DL-2005-7 0.98 0.097 1.03 85699
DL-2005-8 0.98 0.097 1.03 85699
DL-2005-9 0.98 0.097 1.03 85699
DL-2005-10 0.98 0.097 1.03 85699
DL-2005-11 0.98 0.097 1.03 85699
DL-2005-12 0.98 0.097 1.03 85699
DL-2005-13 0.98 0.097 1.03 85699
DL-2005-14 0.98 0.097 1.03 85699
DL-2005-15 0.98 0.097 1.03 85699
DL-2005-16 0.98 0.097 1.03 85699
DL-2005-17 0.98 0.097 1.03 85699
DL-2005-18 0.98 0.097 1.03 85699
DL-2005-19 0.98 0.097 1.03 85699
DL-2005-20 0.98 0.097 1.03 85699
DL-2005-21 0.98 0.097 1.03 85699
DL-2005-22 0.98 0.097 1.03 85699
DL-2005-23 0.98 0.097 1.03 85699
DL-2005-24 0.98 0.097 1.03 85699
DL-2005-25 0.98 0.097 1.03 85699
DL-2005-26 0.98 0.097 1.03 85699

Lab No. RATE Date BP AD S.D.

DL-2005-1 4.94 300 1650 51
DL-2005-3 4.94 333 1617 54
DL-2005-2 4.94 416 1534 60
DL-2005-4 4.94 441 1509 62
DL-2005-5 4.94 894 1056 87
DL-2005-6 4.94 273 1677 49
DL-2005-7 4.94 333 1617 54
DL-2005-8 4.94 421 1529 60
DL-2005-9 4.94 454 1496 63
DL-2005-10 4.94 511 1439 66
DL-2005-11 4.94 597 1353 72
DL-2005-12 4.94 624 1326 73
DL-2005-13 4.94 633 1317 74
DL-2005-14 4.94 656 1294 75
DL-2005-15 4.94 671 1279 76
DL-2005-16 4.94 671 1279 76
DL-2005-17 4.94 753 1197 80
DL-2005-18 4.94 799 1151 82
DL-2005-19 4.94 1046 904 94
DL-2005-20 4.94 1275 675 104
DL-2005-21 4.94 569 1381 70
DL-2005-22 4.94 572 1378 70
DL-2005-23 4.94 649 1301 74
DL-2005-24 4.94 682 1268 76
DL-2005-25 4.94 688 1262 77
DL-2005-26 4.94 721 1229 78

* The early AD 675 date is before accepted island
settlement and is not considered in the discussion.