Mapping prehistoric statue roads on Easter Island.

Lipo, Carl P. ; Hunt, Terry L.

Introduction

Scholars have long debated on how the colossal statues (moai) on Easter Island were transported to every corner of the island. The ancient people of Easter Island carved and moved hundreds of multi-ton statues up to 18km over rugged terrain. While early researchers tracked the course of a few ancient roads leading from the main statue quarry at Rano Raraku, new high-resolution satellite imagery reveals the remains of an extensive pattern of prehistoric roads. Here we report analyses of satellite images and the results of ground-based field surveys that show something of the newly discovered ancient network and its varied components. The distribution and structure of the roads also provide new evidence for evaluating models for how the statues were moved. The pattern of the roads suggests a hypothesis for statue movement by independent groups from across the island, rather than labour controlled by a central chiefdom. Our survey also shows that historic and modern activities have destroyed roads and where protection is urgently needed.

Since the first encounter by the Dutch explorer Jacob Roggeveen in 1722, the giant stone statues (moai) of Easter Island (Rapa Nui, Chile) have engendered speculation and debate. With sizes ranging from about 2 to 10m in height, the distribution of the completed statues meant carving and hauling an estimated 14 000 tons of stone over distances of up to 18km. Despite many experiments and much debate, just how the massive statues were moved remains a mystery, particularly given the lack of historic or ethnographic evidence. Wild claims of transport by volcanic activity (Wolff 1948) and even extra-terrestrials have been proposed (Von Daniken 1974), and Roggeveen felt that the statues were formed in situ by moulding clay or pliable earth (Sharp 1970). However, most archaeologists now agree that the stone statues were moved by some kind of rolling activity (Van Tilburg 1994), by rocking the figures in an upright fashion (Pavel 1990, 1995; Love 1990) or by levering (Lee 1998, 1999, 2000). Additionally, experimental studies have shown that a relatively small number of individuals could move the statues, rather than the large groups of organised labour once considered necessary (see Lee 1998, 1999, 2000; Love 1990, 2000; Van Tilburg 1996).

Despite the long interest and lively debate on how the statues could have been moved, most scholars have paid only passing attention to the presence of extensive road features stretching from the statue quarry at Rano Raraku in multiple directions across the island. In 1919, Routledge (1919: 194) noted that 'the level rays of the sinking sun showed up inequalities of the ground, and, looking towards the sea, along the level plain of the south coast, the old track was clearly seen; it was slightly raised over lower ground and depressed somewhat through higher, and along it every few hundred yards lay a statue'. Routledge (1919) outlined the arrangement of roads over the island and sketched a map for segments of a few of them, and while often acknowledged by others, little else was known or made of their significance. In recent years, Love (2000) has undertaken intensive research on the roads and their composite features, including some excavations in the island's southern sector.

Yet despite this recent attention, the roads of Easter Island have never been systematically documented for the island as a whole. Here, we present an extensive analysis of the island's roads as a complement to the intensive studies undertaken recently by Love. Our data have implications for evaluating models for how the statues were moved, and in particular for understanding the scale of labour organisation and investment in monumental statuary and architecture on the island. As we have outlined elsewhere, such investment in monumentality may have played a significant role in the evolution of Easter Island culture (Hunt & Lipo 2001).

Easter Island is small (171[km.sup.2]) and among the most isolated inhabited islands on earth. The island lies just outside the tropics (27[degrees]9'S) and lacks the biodiversity, abundant rainfall, permanent streams and rich reef ecosystems common elsewhere in Polynesia. Multiple lines of evidence document settlement from Eastern Polynesia no later than AD 600-700. In its extreme windward location to Polynesia, the island likely remained well isolated following colonisation (Finney 1993). Scholars estimate that over at least 500 years (c. AD 1000-1500) islanders carved more than 700 multi-ton moai, with at least 300 transported from the quarry at Rano Raraku and traversing distances up to 18km over a rugged landscape. In addition to statues, the ancient islanders constructed more than 313 monumental religious structures (ahu) (Martinsson-Wallin 1994), comprising with the statue industry large investments in cultural elaboration (Hunt & Lipo 2001). In sum, despite a remote location and limited resources, the per capita investment in monument construction on Easter Island is among the highest anywhere in the ancient world.

Methods

Over the past several decades, satellite imagery has become a powerful and efficient means for generating information about the structure of the earth's surface over large areas. The potential of satellite images in archaeological research, however, is only beginning to be recognised. Part of the slow adoption of satellite-based research in archaeology has been the relatively low resolution of images that was available from early satellites. Satellites such as those in the Landsat series, for example, are only able to resolve features larger than 80m in size. Thus, the earliest uses of satellite images were limited mostly to the study of landscapes or very large archaeological features (e.g., Allan & Richards 1983; Custer 1986; Ebert & Lyons 1980; Findlow & Confeld 1980; Schalk & Lyons 1976). In the past several years, this situation has radically changed with the availability of declassified military imagery and the establishment of commercial firms who have launched their own satellites. A new generation of satellites such as Corona, SPIN-2, Orbview-3/4, SPOT, EROS and Ikonos provide high-resolution images that are well below 10m and at low costs. Here we take 'high resolution' to mean imagery that represents the earth's surface at a resolution of less than 10m across (Forte 2001:132). It is possible in such images to recognise archaeological features (e.g., structures, monuments, deposits); their use has spawned a wide variety of applications in archaeological research (e.g., Failmezger 2001; Fowler, 1996, 2002; Kennedy 1998, Kouchoukos 2001; Mumford & Parcak 2002; Philip et al. 2002).

Significantly, the resolution of recent satellite images has proven to be sufficient to trace linear prehistoric features such as roads (e.g., Ur 2003; Sever & Wagner 1991) and even footpaths (McKee et al. 1994; Sheets & Sever 1991; Sheets 2003). The success of these studies suggests that it should be possible to detect roads used to move Easter Island statues (moai) which can be c. 2-5m across provided views are not obstructed by vegetation or other kinds of ground cover.

We acquired images from DigitalGlobe's QuickBird satellite (http://www.DigitalGlobe. com), a commercial satellite launched in 2001, which orbits the earth every 93.5 min and revisits its path every 1-3.5 days depending on the latitude. Remarkably, the QuickBird satellite is capable of generating panchromatic images with resolutions of 61-72cm and multispectral images with resolutions of 244-288cm. Differences in resolution depend on the degree to which the satellite is off from the nadir (overhead position) when the image is taken. Data in the DigitalGlobe images are geo-processed so that points and features can be located with an accuracy of 23m at 90 per cent circular error.

The QuickBird satellite provides images that are comparable with most aerial photographs, and is of sufficient resolution to provide a base for mapping between 1'= 200' and 1" = 400' scale (Nale 2002). One of the advantages of QuickBird is that it is able to collect data at near nadir and that the corresponding digital imagery may be considered true overheads. An additional advantage that QuickBird has over aerial photographs is that images are available in colour, panchroma and 4-band multispectra ranging from blue to near-infrared (400 nm-900 nm). This multispectral information, though at a lower resolution (c. 2.4m), is of sufficient quality to provide for a broad range of vegetation and environmental information. Natural colour imagery can provide for crop, forest and wetland information. Remarkably, all of these datasets are available for areas across most of the earth for about US $30 per square kilometre.

We used a composite of three sets of recent high-resolution (60cm) panchromatic and multispectral images in the analysis of Easter Island. These images were taken on 2 April 2003, 2 December 2002 and 3 February 2003 and at angles less than 15 degrees off of nadir. Though the images were georeferenced by DigitalGlobe, we conducted additional georectification by matching points on the images with GPS coordinates collected on the ground. This rectification allowed us to travel to points identified on the images to within the precision of our hand-held GPS units, or about 20m. The image data were processed using a nearest-neighbour resampling kernel that produced a pixel size of 70cm for the panchromatic images and 240cm for the multispectral images. While portions of the island are not visible due to cloud coverage, as a composite set the images provide a view of c. 85 per cent of the island's ground surface.

The minimal vegetation and the nature of the archaeological record on the island made features easily visible in the images. The environment of the Easter Island predominately comprises various grasses and other weedy plants. Much of the island is under grazing by cattle and horses, with some small-scale farming. Areas of the island have been planted with eucalyptus trees, but these groves are limited. As a result, archaeological features such as platforms (ahu), house foundations (hare paenga), agricultural features (manavai) and in some cases, statues (moai), were easily identified. We recognised road features as linear forms on several criteria, often found in combinations, such as vegetation differences, depressions filled with cobble scree, banks, trails between statues, erosion patterns and shadow marks. Vegetation differences appear to have been caused by compressed sediment that retains greater moisture. The same compressed, U-shaped roadbeds naturally filled with surface rocks (cobbles) as scree or eroded into troughs with surface water flow, particularly on slopes. Some roads have curbstones and other structural features, including earthworks that create shadows in angled light, or are associated with multiple statues.

We first identified alignments and linear features of stone, soil and/or vegetation anomalies measuring to approximately 5m in width (Figures 1 and 2). Roads appeared on the images as extensive, dark lines in the case of the panchromatic images or with colour and intensity differences in the case of the multispectral images. The panchromatic images were the most useful since the resolution of these images is an order of magnitude greater than the average width of the road features. For the most part, the multispectral data were too low in resolution, given the average width of the roads. However, the multispectral images often provided information about potential vegetation differences used to corroborate potential ancient roads first recognised in the panchromatic data.

[FIGURES 1-2 OMITTED]

Extensive historic sheep ranching in the late nineteenth and early twentieth centuries by Williamson-Balfour and Company and modern farming activities have led to the construction of roads, stone walls (pirca)and fences. Since the linear features identified on the satellite images included a combination of prehistoric, historic and modern roads, we initially mapped as many linear features as possible and used ground survey to eliminate roads clearly attributed to more recent activity. Survey teams travelled to all areas identified as potential roads on the satellite images. The presence of statues (moai) along roadbeds as well as constructed linear features such as curbstones with U-shaped depressions served as a primary means to confirm the prehistoric age of roads. Such features were not always continuous, but linear aggregations of features could be traced across the landscape. Use of these criteria means that our map delineates the minimum number of statue roads, but ensures that the patterns we observed are related to prehistoric activity, and not ranching or other kinds of later historic activities.

To distinguish the ancient roads from historic and modern ones as well as other linear features on the landscape, we paid particular attention to linear features crossed by historic features such as ranch roads and walls to estimate relative ordering. We also plotted the island-wide distribution of 702 statues and 87 statue topknots (pukao) of red volcanic cinder from Puna Pau Crater (Figure 3). Finally, we conducted extensive ground survey, covering approximately 70km of potential roads. The field survey consisted of walking the length of the areas identified as potential roads (ancient, historic or modern) and recording evidence of prehistoric modifications, linear patterns of erosion, and particularly the presence of statues. Indeed, investigations of the alignments plotted from the satellite images led to the discovery of several statues previously undocumented.

[FIGURE 3 OMITTED]

Results

Our analysis and field survey have identified approximately 32kin of ancient statue roads emanating like spokes from the Rano Raraku quarry across the island (Figure 3). These roads comprise a minimum of four major roads leading north-north-west, north-west, west-south-west and south-west. A fifth major road may head directly north from the quarry, but historic and modern modification and use leave this identification ambiguous without further research. Major roads are also found along the west and north side of the island and on the lower (northern) edge of Rano Kao Crater. Minor roads or remnants lead from Rano Raraku to the monumental site (ahu) of Tongariki, and along the northeast coast. Some of these roads conform to those sketched by Routledge in 1919, but overall our research shows a greater extent of roads. Field survey confirms that ancient road alignments are formed by linear depressions and/or eroded beds commonly filled with cobbles, as soil erosion (particularly on slopes), as stone constructed features (e.g., single and double alignments of curbstones lining the edge of roadbeds), as earthen/rock modifications, as clearings of surface rock and/or as vegetation differences (Figures 4 and 5). Along the southern road in Akahanga we also identified circular stone clusters (6-8m in diameter) placed roughly parallel to the road (Figure 2). These features may have played a role in statue transport; a few are found on other roads. In the field survey we also documented and distinguished historic horse trails, ranch roads and zones of modern grading or other machine work.

[FIGURES 4-5 OMITTED]

The course of the statue roads also reveals their independence from areas of high-density occupation. The paths do not connect areas of habitation together. Rather, the roads appear to have been primarily constructed for statue transport, and not for other primary purposes, such as travel and communication. This may distinguish, to some degree, the roads of Easter Island from general economic transportation systems elsewhere (e.g., Sheets 2003; Trombold 1991).

Some ancient statue roads had been modified and reused in historic and modern times. In many cases, ancient statue roads appear to have offered ideal paths for historic horse trails, ox carts, or more recently for motor vehicles. Recent use of the ancient roads appears to have followed a 'path of least resistance' as they are relatively free of rock or other barriers and traverse minimal topography and distances from one point to another. The termination of at least the three major west and north-west bound roads end in recently ploughed fields. Such historic and modern use makes the presence of some ancient roads difficult to confirm without further intensive study, such as the excavations conducted by Charles Love on road segments in 2001 (see Flenley & Bahn 2002). Reports from Love's excavations will provide important details on their construction, form, use and a basis for establishing an absolute chronology through dating. Tracing ancient roads across the landscape provides a vivid picture of where recent activities have destroyed archaeological evidence. The roads are a clear reminder of the urgent need for a comprehensive management plan for historic preservation on the island.

Discussion

The systematic documentation of the ancient roads of Easter Island is essential to evaluate the competing models for how the statues were moved. Data on statue size and distribution suggest that their movement required the greatest investment in energy (Figure 6), and that transport failure occurred more often with larger statues. Documenting the roads provides data on the actual minimal distances that the statues were transported, rather than hypothetical or optimal paths used in some models (Van Tilburg 1994). Direct study of the roads will also inform on the nature of the surfaces and the slopes the statues traversed in the movement over the island.

[FIGURE 6 OMITTED]

Mapping the paths of these stone giants also provides some clues for the social, economic and political organisation of the ancient islanders. The roads emanate from the quarry in a spoke-like pattern to the south, west and north coasts. Field survey confirmed the distribution of at least seven major roads/road segments, extending some 32km. Their arrangement suggests that roads were not necessarily shared, but each region (potentially related to individual social groups) had their own road for delivery of statues from Rano Raraku. The apparent independence of the roads argues against a centralised authority regulating statue production and transport. Instead, the evidence favours a model of smaller, competing groups engaged in the labour-intensive investments of statue making and transport. We suggest such a hypothesis for relatively autonomous competing groups over much of the island's prehistory should be tested against multiple lines of evidence.

The remains of the ancient roads provide a new context in which to evaluate models for the movement of the statues from the quarry at Rano Raraku to every part of the island.

If the pattern of road construction and use support a hypothesis for the movement of statues by smaller independent and competitive groups, then social and political organisation was not centralised for the island, as some have inferred (Diamond 1995). Indeed, much of the story told for Easter Island's ancient past is in need of critical re-evaluation (Hunt & Lipo 2001; Rainbird 2002). Meanwhile our survey also suggests that the ancient roads themselves are fast disappearing as ploughing and other farming activities spread across the island.

Acknowledgements

We thank Sergio Rapu, Francisco Torres Hochstetter of the Sebastian Englert Museum on Rapa Nui and Daniel O. Larson for their collaboration and support. Brett Shephardson, University of Hawaii Ph.D. student, collected data for the island-wide distribution of statues. T. Hunt and students from the 2003 University of Hawai'i field school conducted the primary survey of pukao. We also thank Ileana Bradford, Kalewa Correa, Roxanne Kennedy, Chris O'Boynick and the many students from the 2003 California State University, Long Beach and University of Hawai'i Archaeological Field Schools for their hard work in the analysis and field survey for this project. Special thanks go to Marc Kelly who provided heroic assistance in the field and in the preparation of the graphics. The satellite images were purchased with support from the College of Liberal Arts, CSULB. We are greatly indebted to Brett Shepardson who kindly provided the digitized topographic data for Rapa Nui and the results from his recent island-wide inventory of 702 statues.

References

ALLAN, J.A. & T.S. RICHARDS. 1983. Use of satellite imagery in archaeological surveys. Annual Report for Libyan Studies 14: 4-8.

CUSTER, J.F. 1986. Application of LANDSAT data and synoptic remote sensing to predictive models for prehistoric archaeological sites: an example from the Delaware coastal plain. American Antiquity 51: 572-88.

DIAMOND, J. 1995. Easter Island's end. Discovery magazine 16: 63-9.

EBERT, J.I. & T.R. LYONS. 1980. Prehistoric irrigation canals identified from Skylab III and Landsat imagery in Phoenix, Arizona, in T.R. Lyons & F.J. Mathien (ed.). Cultural resources remote sensing 209-28. Washington: Cultural Resources Management Division, National Park Service.

FAILMEZGER, V. 2001. High resolution aerial color IR, multi-spectral, hyper-spectral, and SAR imagery over the Oatlands Plantation archaeological site near Leesburg--a pilot study of the new high resolution satellite imagery, in S. Campana & M. Forte (ed.). Remote sensing in archaeology: 143-48. Florence: All'Insegna Del Giglio.

FINDLOW, F.J. & L. CONFELD. 1980. LANDSAT imagery and the analysis of archaeological catchment territories: a test of the method. Anthropology 10: 31-52.

FINNEY, B. 1993. Voyaging and Papa Nui prehistory. Rapa Nui Journal 7: 1-6.

FLENLEY, J. & P. BAHN. 2002. The enigmas of Easter Island. Oxford: Oxford University Press.

FORTE, M. 2001. Telerilvamento e paesaggi archeologici tridimensionali, in S. Campana & M. Forte (ed.). Remote sensing in archaeology: 95-141. Florence: All'Insegna Del Giglio.

FOWLER, M.J.F. 1996. High-resolution satellite imagery in archaeological application: a Russian satellite photograph of the Stonehenge region. Antiquity 70: 667-71.

--2002. Satellite remote sensing and archaeology: a comparative study of satellite imagery of the environs of Figsbury Ring, Wiltshire. Archaeological Prospection 9: 55-69.

HUNT, T.L. & C.P. LIPO. 2001. Cultural elaboration and environmental uncertainty in Polynesia. The proceedings of the fifth international conference on Easter Island and the Pacific. Los Osos: The Easter Island Foundation.

KENNEDY, D. 1998. Declassified satellite photographs and archaeology in the Middle East: case studies from Turkey. Antiquity 72: 553-61.

KOUCHOUKOS, N. 2001. Satellite images and Near Eastern landscapes. Near Eastern Archaeology 64: 80-91.

LEE, V.R. 1998. Rapa Nui rocks: Impressions from a brief visit. Rapa Nui Journal 12: 69-72.

--1999. Rapa Nui rocks update. Rapa Nui Journal 13: 16-17.

--2000. Vince Lee's uphill battle (http://www.pbs. org/wgbh/nova/lostempires/easter/vince.html).

LOVE, C. 1990. How to make and move an Easter Island statue, in H.M. Esen-Bauer (ed.). State and perspectives of scientific research in Easter Island culture. 139-40. Frankfurt: Courier Forschungsinstitut Senckenburg.

--2000. More on moving Easter Island statues, with comments on the NOVA program. Rapa Nui Journal 14: 115-18.

MARTINSSON-WALLIN, H. 1994. Ahu--the ceremonial stone structures of Easter Island. Analyses of variation and interpretation of meanings. Uppsala: Aun 19.

MCKEE B.R., T.L. SEVER & P. SHEETS. 1994. Prehistoric footpaths in Costa Rica: remote sensing and field verification, in P. Sheets & B. McKee (ed.). Archaeology, volcanism and remote sensing in the Arenal region, Costa Rica: 142-57. Austin: University of Texas.

MUMFORD, G. & S. PARCAK. 2002. Satellite image analysis and archaeological fieldwork in El-Markha plain (south Sinai). Antiquity 76: 353-4.

NALE, D.K. 2002. QuickBird aerial product comparison. Manuscript prepared by Enap International for Digital Globe, Inc. (http://www. digitalglobe.com/downloads/QuickBird%20%20Aerial%20Photography%20Comparison% 20Reporcpdf).

PAVEL, P. 1990. Reconstruction of the transport of moai, in H.M. Esen-Bauer (ed.). State and perspectives of scientific research in Easter Island culture: 141-44. Frankfurt: Courier Forschungsinstitut Senckenburg.

--1995. Reconstruction of the transportation of the moai statues and pukao hats. Rapa Nui Journal 9: 69-72.

PHILIP, G., D. DONOGHUE, A. BECK & N. GALIATSATOS. 2002. CORONA satellite photography: an archaeological application from the Middle East. Antiquity. 76: 108-18.

RAINBIRD, P. 2002. A message for our future? The Rapa Nui (Easter Island) ecodisaster and Pacific Island environments. World Archaeology 33: 436-51.

ROUTLEDGE, K. 1919. The mystery of Easter Island. Illinois: Adventures Unlimited Press.

SCHALK, R.F. & T.R. LYONS. 1976. Ecological applications of Landsat imagery in archeology: an example from the San Juan basin, New Mexico, in T.R. Lyons (ed.). Remote sensing experiments in cultural resource studies: non-destructive methods of archaeological exploration, survey, and analysis: 173-86. Albuquerque: Reports of the Chaco Center, No. 1. Chaco Center, National Park Service and University of New Mexico.

SEVER T.L. & D.W. WAGNER. 1991. Analysis of prehistoric roadways in Chaco Canyon using remotely sensed digital data, in C. Trombold (ed.). Ancient road networks and settlement hierarchies in the New World: 42-52. Cambridge: Cambridge University Press.

SHARP, A. 1970. Journal of Jacob Roggeveen. Oxford: Oxford University Press.

SHEETS, P. 2003. Prehistoric human footpaths in Costa Rica indicate intimate ties with villages, cemeteries. Science Daily. 8 October 2003 (http://www. sciencedaily.com/releases/2003/10/ 031008064657.htm).

SHEETS, P.D. & T.L. SEVER 1991. Prehistoric footpaths in Costa Rica: transportation and communication in a tropical rainforest, in C. Trombold (ed.). Ancient road networks and settlement hierarchies in the New World: 53-65. Cambridge: Cambridge University Press.

TROMBOLD, C.D. (ed.). 1991. Ancient road networks and settlement hierarchies in the New World. Cambridge: Cambridge University Press.

UR, J. 2003. CORONA satellite photography and ancient road networks: a northern Mesopotamian case study. Antiquity 77:96-115.

VON DANIKEN, E. 1974. In search of ancient gods: my pictorial evidence for the impossible. London: Souvenir Press.

VAN TILBURG, J.A. 1994. Easter Island archaeology, ecology and culture. London: British Museum Press.

VAN TILBURG, J.A. 1996. Mechanics, logistics and economics of transporting Easter Island (Rapa Nui) statues. Rapa Nui Journal 10:110-15.

WOLFF, W. 1948. Island of death. New York: Hacker Art Books.

Received: 27 October 2003; Accepted: 9 February 2004; Revised: 28 January 2004

Carl P. Lipo (1) & Terry L. Hunt (2)

(1) Department of Anthropology, California State University Long Beach, 1250 Bellflower Blvd., Long Beach, CA 90840, USA (Email: clipo@csulb.edu)

(2) Department of Anthropology, University of Hawaii-Manoa, 2424 Maile Way Honolulu, HI 96822, USA