OFU ISLAND AND THE TO'AGA SITE: DYNAMICS OF THE NATURAL AND CULTURAL ENVIRONMENT PATRICK V. KIRCH Every Manuan thinksfirst of his own village, then of his island, then of Manua as a whole-a tiny archipelago set offfrom all the rest of Samoa. There may be rivalry between villages; there have even been wars- internecine conflicts between Fitiuta and the three villages of Tau and between OfQu and Olosega-but against the outside world Manua presents a solidfront. Mead (1930:51) THE MANU'A ISLANDS, ISOLATED from the main Samoan islands by 100 kn of often- turblent ocean, form a geographically and culturally disinctive cluster at the eastem extreme of the archipelago (fig. 2.1). Being closest to Tutuila, Manu'ans had the greatest interaction with the occupants of that island (a pattern that continues today with Manu'a and Tutuila comprising the Tern- tory of American Samoa). Although sharing in most respects the dassic characteristics of Samoan culture, Manu'ans nonetheless regard themselves as different and distinctive. Mead (1930:9) commented on this distinctiveness, for example in the lack of emphasis on "war, its paraphemalia, its ritual, and its gods." She opined that "the chief historical value of Manua lies in her easterly and isolated position, offering a valuable check upon cultural traits which are intru- sive in westem Samoa" (1930:9). Such isolation, however, was relative and should not be overempha- sized; as our archaeological investigations have revealed, the prehistory of Ofu shares much in common with the westerly islands of Samoa, and prehistoric inter-island contacts can be documented in the transport of basalt adzes (see Weisler, chapter 12), and perhaps of pottery as well (see Hunt and Erkelens, chapter 9). Thus, despite considerable isolation, Manu'a was never a "closed system." The dtee islands of fte Manu'a Group-Ta'u, Olosega, and Ofu-form an intervisible cluster and comprised a political unity in ancient times. Ta'u is by far the largest island (table 2.1) and dominated the group politically, being the seat of the Tui Manu'a paramount line of chiefs. Olosega and Ofu are practically adjacent to each other, separated only by a narrow and shallow strait, now spanned by a concrete causeway. They are similar in size, Ofu being slightly smaller (3.4 kM2), but rising to a loftier summit (at 638 m) than Olosega. Ofu and Olosega are visually exquisite, their youthful volcanic profiles dthust abruptly out of the Pacific, rising to basalt pinnacles often shrouded in misL Narrow ribbons of blindingly white coral sand fonn a no-man's-land between the rank green of tropical vegetation and the blue-green mosaic of coral reefs. Sheer cliffs towering hundreds of meters above To'aga, Sili, and other villages provide nesting sites for the white-tailed tropic birds that ride the trade winds battering the islands' mass. 10 The To'aga Site Figure 2.1 Map of the Manu'a Isands. Perhaps the most striking feature of Ofu's environment is its precipitous topography, and the scarcity of flat land. No less than ninety-one percent of the total island area consists of slopes greater than 30 degrees. Due to its geological youth, the steep sides of the volcanic cone have not yet been significantly incised by streams, the principal weathering being major landslides resulting from faulting, and creating sheer cliffs. The only level terrain is found around the margins of the island, at the critical land-sea interface, where a narrow coastal terrace has been constructed primarily from coral sand, with the addition of some talus and colluvium from the slopes inland. This coastal zone, ranging from 50 to 150 meters wide, is unevenly distributed, being found along the westem and southeastem coasts but not along the north coast, except for a small strip at the northeast- em point. The To'aga site occupies the longest span of coastal terace land on the island, along the southeastem shoreline. This coastal terrace was a critical microenvironment for the Polynesian occupants of Ofu, first as the setting for their villages, and second as a major zone of garden land. More importantly, the narrow coastal terrace is a highly dynamic environment. Understanding the morphodynanics of this land-sea interface is essential not only to reconstructing the prehistoric sequence of the To'aga site itself, but more broadly for an understanding of Manu'an prehistory. In this chapter, I outline the main characteris- tics and resources of the Ofu Island environment that were significant to the prehistoric Polynesian inhabitants with their horticultural and fishing economy. I also pay considerable attention to those aspects of the contemporary environment-both physical and biotic-that inform us about the dynamics of environmental change. Polynesian archaeologists have now abandoned an earlier perspective that viewed island ecosystems merely as static backdrops to cultural developments. Increasingly, archaeological investigations through- out Polynesia are yielding striking evidence of ecological change: forest clearance and vegeta- tional succession, extinction and extirpation of birds and other biota, alluviation of valley floors, sea level changes, and so forth (e.g., Kirch 1982, 1984; Steadman 1989; Flenley and King 1984; Olson and James 1984; McGlone 1983). Some of these changes resulted from natural processes; others reflect the impact of human populations themselves, colonizing isolated and biologically- vulnerable islands for the first time (Fosberg 1963). The prehistory encapsulated within the To'aga site is as much a history of environmental dynamics as of cultural changes in pottery Wpes or adz forms. The Natural and Cultural Environment 11 Table 2.1 Environmental Characteristics of the Manu'a Islands Characteristic Ta'u Olosega Ofu Area (kIn) 28.5 4.5 3.4 Highest Point (m) 965 494 638 Area<30% slope(%) 41 10 9 Coastline (kan) 32.5 13.3 10.4 Population (1980) 1,146 340 254 GEOLOGY AND GEOMORPHOLOGY OF OFU ISLAND Initial studies of Ofu geology by Friedlander (1910), Daly (1924), and Stearns (1944) were based on short reconnaissance surveys. Although Steams spent only one or two days in Manu'a, he was able to produce "a remarkably accurate geologic sketch map" (Stice and McCoy 1968:429). The most intensive study to date was carried out by Stice and McCoy (1968), whose report and geologic map of Ofu we have drawn upon for the following summary. Subsequent work by Natland (1980) and McDougall (1985) has corrected certain errors made by Stice and McCoy in dating the island's geological forma- tions. Ofu and Olosega comprise an integral geologi- cal complex "of volcanic cones that have been buried by lava flows from two coalescing shields" (Stice and McCoy 1968:443). At least five cones make up the Asaga Formation; these include a "composite cone exposed in the cliff behind To'aga," as well as "an explosion breccia cone with an associated intrusive plug at Fatuaga Point" (1968:443-44). The latter plug is exposed as the visually impressive spire at the eastern point of Ofu (fig. 2.2). Stice and McCoy believed that the Asaga Formation was of lower Pliocene age, although this has now been shown to be incorect (see discussion below). All of these Asaga Fonnation cones were subsequently buried by lava of the Tuafanua Formation, from the A'ofa and Sili coalescing shields. "After summit collapse of the shields, volcanism decreased so that a sea cliff about 300 feet high was cut around the islands" (1968:456). In the To'aga area, the high cliff seems to have resulted from a combination of faulting and sea erosion. Finally, renewed volcan- ism resulted in the construction of the tuff cone of Nu'utele Islet, as well as in several hawaiite and olivine basalt flows that filled in deeply eroded stream valleys on westem Ofu. Whereas Stice and McCoy (1968) regarded the geological history of Ofu to have extended over a fairly long period, from the lower Pliocene through to the late Pleistocene, recent work by Natland (1980) and McDougall (1985) has con- firmed that the island is very recent in age. McDougall reports that "these volcanoes are quite youthful, confirmed by unpublished K-Ar ages from this laboratory [Austalian National Univer- sity] averaging 0.3 Ma for Ofu/Olosega and less than 0.1 Ma for Ta'u" (1985:318). As Natland argues, "there can be little doubt that both Tau and Ofu-Olosega are substantially younger than the shield volcanoes of Tutuila, which were extinct and extensively eroded before drowning of the Pleistocene reefs" (1980:721). Indeed, the Sa- moan chain illustrates a typical instance of linear, "hot-spot" progressive volcanism, with the islands increasing in age from east to west. (Recent, renewed volcanism on Savai'i is evidently due to the proximity of the westem end of the archi- pelago to the Pacific Plate margin.) The youthful age of Ofu and Olosega is of considerable impor- tance to the geomorphology and geoarchaeology of the To'aga site, as we shall argue in greater detail in chapter 4. This is because the islands are 12 The To'aga Site still tectonically unstable, due to point-loading on the oceanic crust, and thus are undergoing a phase of subsidence (Menard 1986). The rocks of the island consist primarily of olivine basalts, hawaiites, and ankaramites which were extruded as both pahoehoe and aa lava flows, interbedded with various pyroclastic tuffs and breccias. Intrusive rocks consist primarily of dikes, and of the plug at Fatuaga Point, a "hypabyssal intrusion of ankaramite" (1968:449). A major swarm of near vertical dikes rns through the central spine of Ofu, and is clearly visible in the cliffs behind the To'aga site. Stice and McCoy report that "the razorback ridge of eastem Ofu is the topographic expression of a dike complex about 400 feet wide. The dikes are nearly vertical ... Most are dense basalt, although olivine basalt, ankaramite, and feldspar-phyric basalt also are present" (1968:449). We were particularly interested in this dike complex as a possible source of dense, fine-grained basalt that could have been exploited by the prehistoric occupants of Ofu for manufacture of adzes or other flake tools. The recent road cutting at Fa'ala'aga provided an opporunity to examine this dike cluster at close hand, and to collect samples for XRF composi- tional analysis (see Weisler, chapter 12). As shown in figure 2.3, the dikes are closely spaced, cutting trugh older, weathered basalts of the Asaga Formation. On the edges of several of these dikes, where they came into contact with the older basalts, we observed glassy "chills" of low-silica volcanic glass ("obsidian"), up to 2 cm thick. Although of poor quality, such volcanic glass could have been exploited for the production of small flakes. The appearance of local volcanic glass flakes in archaeological contexts in the Manu'a Group is thus a distinct possibility. We did not observe other possible sources of fine-grained basalts suitable for adz manufacture, except for a rather weathered dike exposure on the edge of Mako Ridge at about 350-400 m elevation (also sampled for XRF analysis). It is entirely possible, however, that prehistoric Manu'ans exploited small exposures of suitable basalt or hawaiite for adz manufacture. We are not aware of any adz quaries in the Manu'a Group, but this negative evidence certainly does not preclude local adz manufacture. There has been relatively little stream erosion on Ofu, and the radial drainage pattem is poorly developed. A number of shallow valleys drain to the north and to the west from the slopes of Tumutumu Mountain. All of these are intennit- tent, flowing only after heavy rains. "'The stream valleys are all youthful and nowhere exceed 50 feet in depth" (Stice and McCoy 1968:455). The main erosional forces at work on Ofu since the cessation of volcanism have been marine wave attack and mass wasting, especially land- slides. As noted above, an extensive sea cliff more than 100 m high "was carved into the island by the sea" (Stice and McCoy 1968:455) during the late Pleistocene to early Holocene.. These high cliffs, such as the one behind the To'aga site, "originated by faulting and/or foundering," but were certainly extended and modified by marine erosion. Landslides continue to be active ero- sional forces, contributing to the talus screes that border the inland edge of the coastal terrace. "Individual blocks also work loose from the cliff face and fall, fonning talus slopes that extend almost continuously around the islands at the base of these cliffs" (1968:455). One such massive block fell between our 1987 and 1989 field seasons: measuring at least 4 m in diameter, this giant boulder crashed through the banana and breadfruit orchards near the western edge of To'aga to land by the roadside, a daily reminder of the geological dynamism of Ofu. Similar large blocks dot the surface of the To'aga site, and one of these was modified in late prehistory for use as an adz grinding stone (see Hunt, chapter 3). For the prehistoric occupants of the To'aga site, such falling boulders posed a continual hazard. As Stice and McCoy relate, a local legend tells of a young girl who "was killed by a large block that rolled across the reef at Sili, where she was fishing" (1968:456). We have no reason to doubt the veracity of this story. RESOURCES OF THE ISLAND For aso b taditionally upon inten- sive root bphoriculture, and on arboricubw-. s more than an abstract The Natural and Cultural Environment 13 Figure 2.2 View of the exposed volcanic plugs at Fa'ala'aga, from the beach at To'aga. The high mountain on the right is the summit of Olosega Island. 0 1 2m Figure 2.3 Sketch of basalt dikes exposed in the road cut across the Le[olo Ridge at Fa'ala'aga. concept. The Samoan archipelago lies within the humid tropics. Buxton, who spent much time studying the natural history of Samoa, summed up the climate thusly: "As the temperature is nearly constant, as rain is abundant and well distributed, and as the islands are surrounded by the ocean, it follows that the atmosphere is moist at all times and seasons" (1930:17). Indeed, relative humidity usually ranges between 80 and 86 percent, with the temperature varying only between 25.7-26.20C. Buxton also observed the effects of such humidity on humans: "the perspiring entomologist soon learns to recognize that his comfort depends entirely on the wind, for the warm damp air does not cool him unless it is in motion" (1930:18). Green (1969:4) echoed these remarks from the viewpoint of the field archaeologist, quoting Curry (1962) to the effect that the Samoan climate is at "the upper limit of thermal comfort for half-naked men at rest." There is, however, a distinct and significant seasonality to the Samoan climate, expressed most clearly in patterns of rainfall and of wind direc- tion. We have no rain gauge records for Ofu itself, 14 The To'aga Site but the records for Tutuila Airport (Nakamura 1984, table 1) are probably similar, indicating a total annual average of 124 inches [3100 mm] (see also Coulter 1941:10-11). (Buxton gives 2,738 mm [107.8 in] for Apia [1930:16].) There is a distinctly drier season from about June through September, in which the monthly rainfall averages about 6 or 7 inches [175 mm], and a wet season from October through May, with montly aver- ages of from 11-14 inches [350 mm]. Even during the "dry season," however, torrential downpours may occur, as beset us all too frequently during the 1989 fieldwork at To'aga. As elsewhere in West- ern Polynesia, this seasonality played an important role in scheduling the agricultural calendar, particularly the clearing and planting of swidden cultivations (Kirch 1978, forthcoming). In some years, the dry season may be particularly acute or lengthy, and the resulting drought can significantly reduce agricultural yields. During the wet season, the opposite may occur, with torrential rains and serious flooding. "Some floods are associated with hurricanes and tropical storms, but flooding can occur at other times as well" (Nakamura 1984:3). The dry season also corresponds ap- proximately to the period of prevailing trade winds (from about April through September). During October to March the winds are more variable, and westerly reversals occur. This latter period was important to early Polynesian voyagers, as it allowed exploratory voyages from west to east (Finney 1985; Irwin 1981, 1992). In addition to the annual pattem of seasonal- ity, there are stochastically recurring environmen- tal hazards that seriously affect Samoan life. We have already mentioned the problem of periodic drought, which can inhibit agricultural production. Even more severe are the hurricanes or tropical cyclones that periodically lash the islands. Coulter notes that "Samoa suffers hurricanes at irregular intervals during the hotter season" (1941:12). Visher (1925:27, table 6) indicates an average annual frequency of two to three hurricanes in the Samoan area, although not all of these are equally intense, nor do their paths always cross the Manu'a Group. When a hurricane bears directly down on the islands, however, the effects are often devas- tating-to crops and orchards, to houses, and to human life itself. The most recent hurricane to lash Manu'a occurred early in 1987, between our first and second field seasons. Ta'u Island was particularly hard hit, with virtually every house destroyed; the area was declared a Federal disaster area. Flying in to Ta'u in June of 1987, one was struck by the devestation of the forest cover of the central volcanic cone which had not yet recovered. Aside from their obvious significance to the prehistoric Samoan population, hurrcanes have considerable archaeological importance as agents of landscape transformation, and of site formation as well as destruction. The torrential rains un- leashed during these events can cause severe flooding and result in major landslides and in the deposition of colluvium and alluvium. Storm surges and high energy waves are capable of moving large quantities of sand and larger clastics (up to boulder size) in the coastal zone. During our 1986 reconnaissance survey on Ta'u Island, we observed a massive rampart of coral cobbles and shingles at Saua which had been thrown up during a previous storm surge. We shall give further consideration to these high-energy processes below, and elsewhere in this monograph (see especially chapters 4 and 6), as they obviously played a significant role in the geomorphological history of the To'aga site. Ofu soils are young and undeveloped, a reflection of the island's geological youth. The soils have been mapped and described by Nakamura (1984). Most of the steep interior is covered with "Ofu silty clay," a deep, well-drained soil formed in volcanic materials. In the steeper areas (slopes 40 to 70 percent) these silty clays are covered in forest. Where slopes range between 15 and 40 percent, the Ofu silty clays provide the main gardening soils. These gardening areas are confined to two zones on the western slopes, and to one area on the northern slope of Tumutumu Mountain (fig. 2.4). The very steep slopes and talus regions lying inland of To'aga, and along the northem coast, are described as "Fagasa family- Lithic Hapludolls-Rock outcrop association, very steep" (Nakamura 1984:1 1). At the inland edge of the coastal terrace are found strips of "Aua very stony silty clay loam" (1984:10), described as "very deep, well drained soil ... on talus slopes . . . formed in colluvium and alluvium derived domi- nantly from basic igneous rock." These areas The Natural and Cultural Environment 15 z r44 0 0D E cJ- 0! 0d C. Yim-. o0 0 4' CY 0 ~14 0 a jo 3 a *: o8 I'Q *E4 z 16 The To'aga Site are used for subsistence gardening, primarily of tree crops such as breadfruit and bananas. Finally, the coastal terrace itself, as at To'aga, consists from the pedological viewpoint of "Ngedebus mucky sand" (Nakamura 1984:15). This is a "somewhat excessively drained soil ... derived from coral and sea shells." The natural vegetation of Manu'a has been significantly modified in the lower elevations by three millennia of human land use. The coastal terrace, as well as the less precipitous mountain slopes, comprise a mosaic of coconut stands, breadfruit and banana orchards, and aroid gardens interspersed with second growth (further discus- sion of coastal terrace vegetation below). On the higher and steeper slopes, however, the original ain forest vegetation persists. Yuncker, who studied the flora of Manu'a, lists of total of 421 species for the three islands, including mosses and pteridophytes as well as flowering plants (1945:4). The Samoan root-tuber and tree crop complex, of course, is wholly adventive to the archipelago, having been introduced and established by the early colonists. The terrestrial fauna of the Manu'a Group is very restricted in vertebrates, somewhat richer in invertebrates (especially land molluscs and insects). The only indigenous mammal is the fruit bat, Pteropus samoensis. These are commonly seen soaring over the forest canopy high above the To'aga site, especially at dusk. The diminutive Pacific rat, Ratus exulans, was introduced by early Polynesian settlers, and its bones are com- mon in the archaeological deposits at To'aga (see Nagaoka, chapter 13). Also purposively intro- duced by the Polynesians were the domestic pig (Sus scrofa) and dog (Canisfamliaris). The richest diversity of vertebrates is among the birds, both native land birds and nesting seabirds (Waading 1982). In the higher elevation forests on Ofu are found the lupe or Pacific Pigeon (Ducula pac4fica) and the Crimson-Crowned Fruit Dove or manutagi (Ptilinopus porphyraceus); both are occasionaUy taken for food. White-Collared Kingfishers, ti'otala (Halcyon chloris manuae), of which there is a distinct Manu'a subspecies, are frequently perched high on poles or telephone wire along the coast. Common in the coastal bush at To'aga and elsewhere on the island is the ve'a or Banded Rail (Gallirallus philippensis), often seen making its characteristic headlong dash across the dirt road from the security of one patch of under- growth to another. The iao or Wattled Honeyeater (Foulehaio carunculata) is abundant in the banana groves at To'aga, where it feeds on banana flower nectar. This bird is sometimes hunted by young Samoan boys, who then pluck and roast the tiny carcasses over a fire in the bush. (Having sampled this delicacy myself, I can attest that while there is little flesh, it is sweet and a delicious complement to roasted bananas.) Also present on Ofu is the Polynesian Starling, mitivao (Aplonis tabuensis manuae). A number of seabirds, and some migratory species, also nest on Ofu. The White-Tailed Tropic Bird, tava'e (Phaethon lepturus), nests in the high cliffs towering over the To'aga site. These elegant birds, soaring high ovetiead, were always a wonderful visual diversion from the perspective of a grimy test pit. A seasonal mi- grant, the Golden Plover or tuli (Pluvialis dominicafulva) is also seen along the coast at To'aga. The contemporary avifauna of Manu'a is only an impoverished remnant of the pre-human bird life. Archaeological excavations throughout Polynesia have revealed a significant pattem of bird extinctions and extirpations due to human- induced forest clearance and habitat destruction, and to direct predation (Olson and James 1984; Steadman 1989; Steadman, Pahlavan, and Kirch 1990; Steadman and Kirch 1990). As reported by Steadman in chapter 14, the To'aga excavations added further evidence of this widespread pattem of avifaunal extinction and extirpation. These faunal depletions are just one aspect of the dy- namic nature of Polynesian ecosystems within the span of human occupation. The only other vertebrates indigenous to the island are a number of lizard species in the fami- lies Geckonidae and Scincidae. There is a rich endemic and indigenous insect fauna, but this is of little archaeological relevance. More important from the viewpoint of the prehis- torian are the land snails, which include a number of endemic and indigenous taxa, as well as several species which have been introduced by humans, prehistorically as well as after European contact. The Natural and Cultural Environmen 17 Among the important families represented in Samoa are Partulidae, Assimineidae, Toma- tellinidae, Helicinidae, and Microcystinae. Land snails frequently preserve well in archaeological deposits, including those at To'aga, and are excellent indicators of microenvironmental change. Several synanthropic 'garden-snail' species, including Lamellaxis gracillis, were inadvertently transported around the Pacific basin by prehistoric people, presumably with crop plants and adhering soil. Their appearances in archaeo- logical contexts are therefore important signals of habitat modification, and indirect evidence for horticulture. In chapter 8, I present an analysis of land snails recovered from the To'aga excavations, and their implications for environmental change at the site. Finally, we cannot ignore the marine environ- ment, so crucial to the indigenous Samoan economy. Ofu is surrounded by a fringing reef, widest and most sheltered on the western side (opposite Ofu Village). The reef is a complex mosaic of micro-habitats and a source of shellfish and fish for the human population. A diverse array of molluscs occupy the reef flat and algal crest, including various bivalves such as Periglypta reticulata, Tridacna maxima, Hippopus hippopus, and Asaphis violascens, and gastropods such as Trochus maculatus, Turbo setosus, Nerita spp., Cypraea spp., Drupa spp., Thais armigera, and Conus spp. Many of these species were heavily exploited by the occupants of the To'aga site and occur in dense concentrations in the midden deposits (see Nagaoka, chapter 13). Also present on the reef are spiny lobsters (Panulirus sp.), sea slugs (holothurians), sea urchins (echino- derms), octopus, and various edible seaweeds. Approximately 800 species of inshore fishes occur around Ofu (Jordan and Seale 1906) and are still taken by the Samoans with a variety of traditional fishing strategies, using spears, nets, hooks, and other gear. Farrell's description of fishing activi- ties in Westem Samoa is equally appropriate for Ofu: In the lagoon itself there is some activity at almost any time, day or night. Hundreds of yards from shore, near the edge of the reef, women hunt for sea foods between the breaking of the larger waves. Fishermen in small canoes (paopao) equipped with goggles and spear search the placid waters inshore, while in shallower waters nets are tossed to enshroud passing shoals of small fish. Further out, midway between the shore and the reef, fish traps are built of hard-riven, close-spaced stakes and wire netting, or of mounds of coral rock. Into these trps fish are driven by 'beaters'; later the catch is shared (1962:179). Among the fish commonly taken are jacks (Caranx spp.), parrot fish (Scarus spp.), wrasses (Labridae), and acanthruids. Bones of these fishes also occur in great frequency in the To'aga ar- chaeological deposits. The open sea beyond the reef is less heavily exploited but is the zone of the prized tunas (Scombridae) and flying fish. Marine turtles (Chelonia mydas and Eretmochelys imbricata) are rarely sighted in the waters off Ofu today but must have nested on the island's sand beaches in substantial numbers prior to early Polynesian settlement. The bones of these turtles are one of the most commonly occurring taxa in the To'aga faunal assemblages. THE CULTURAL AND SOCIAL LANDSCAPE As in all Polynesian islands, the landscape of Ofu is culturally and socially ordered. This includes the system of land use, the pattem of land tenure, and the village settlement pattem and its intemal structure, all of which have evolved over several millennia. One goal of our archaeological investigations in Manu'a has been to contribute to an understanding of how this distinctive cultural and social landscape has developed over the course of prehistory. A brief description of some of the key aspects of the contemporary landscape is therefore apposite, as an ethnographic reference or 'endpoint.' Farrell eloquently evokes the essence of the Samoan settlement pattem of "villages ... strung like beads unevenly along the thread of the coastline" (1962:177). Mead elaborates: "The cliff behind, the sea before it, defines the ground plan of a Manuan village, which may spread out in either direction as far as the land permits" (1930:45). On Ofu today, there is only one such village (with two named sectors, Ofu and Alaufau), 18 The To'aga Site strung along the westem coastal terrace. TIhis village lies conveniently between the widest and most sheltered expanse of reef flat and lagoon (protected by Nu'utele and Nuusilaelae Islets), and te largest expanse of arable mountain slope inland. Previ- ously, the zone of coastal villages extended around most of the southem and southeastem parts of th island, including the To'aga area (see Hunt, chapter 3). There is also limited evidence that some occupa- tion may have extended inland into the intensive gardening zone in late prehistorc times. Although in the past decade or two most houses on Ofu have been rebuilt using westem materials- largely concrete with corrugated roofs-the ground plan of the village remains essentially traditional. The central focus is the malae or "village green" where important ceremonies and feasts may be held, the church, and the guest house of the high chief (which functions also as a council house for the fono). The individual households, strips of land that extend in principle from the beach to the mountain slope, spread out on either side of fte sandy road- way. Traditionally, a household had thee main stuctures: a guest house, the main dwelling, and fte cookhouse (Handy and Handy 1924; Buck 1930:8- 97). Mead aptly descrbed the Manu'an guest houses "which stand by the sea, are round and high-ircles of posts about four feet high topped by a twenty-foot thathed circular cone. They stand upon a founda- tion of small stones which rises in slightly higher, narrowing concentric circles, each terrace about five or six inches high, and edged by larger stones" (1930:46). The main dwelling or sleeping house was taditionally the long house (fale o'o), with rounded ends. The cookhouse or shed is thefale wnu, literally house which shelters the wnu or earth oven (Buck 1930:13). These are situated farthest from the guest houses: "a small shack supported on four pillars and roofed, not with sugar cane thatch but with mats woven of palm leaves" (Mead 1930:48). The layout of house types within the household complex reflects a distinctive social and symbolic structure to the village organization. The main axis of orientation is perpendicular to the beach, the interface of land and sea, and thus extends in two directions: i tai, toward the sea, and i uta, inland. The seaward direction is higher ranked, associated with chiefs and persons of status; irnland is lower ranked, associated with prduction and with the economic basis of society. Mead describes the seaward-irdand dichotomy thusly: The term i tai (towards the sea) stands for the optimum position; the village on the seashore, the house on the sea side of the village, the place of honor in the front of the house. And as the trails lead back from the village over narrow stils, through stony piaces, swamp pices, into deep gulches, and up slippery inclines; so the channel marks the way out to sea (1930:50). This strctural dichotomy between land and sea, and the way in which it organizes the spatial structure of the village, is a widespemad-and therefore probably ancient-pattern within Westem Polynesia, includ- ing Fiji. Salins (1976:3745, fig. 6), for example, has described this structure for Fiji, and in another paper (1981) has outlined some of the cultural associations between the sea and high-ranking chiefs. The Polynesian Outlier of Tikopia (Firth 1936; Kirch and Yen 1982) likewise has a character- istic spatial organization very much like that de- scribed above for Samoa, with household units differentiated along a seaward-landward axis, canoe houses toward the beach, cookhouses inland, with the main dwelling mediating between. (The Tikopian dwelling itself is divided into seaward:male and landward:female divisions; see Firth [1936].) Similarly, on Niuatoputapu Lsland in nofferm Tonga, the prehistoric coastal villages appear to follow this pattem (Kirch 1988). And in Fuuna Island, Burrows (1936) describes a village structure very reminiscent of the Manu'an situation. In sum, this kind of cultural and symbolic ordering of space along a seaward-landward axis has a wide distribu- tion in Westem Polynesia, and arguably has a deep prehistory in the region as a "structure of the long run"'(Braudel 1980). Our tansect excavations at the To'aga site provided some evidence of a coastal village settement pattem ta may also have been organized in this characteristic Westem Polynesian model some 2,500 years ago. TO'AGA: LAND USE AND VEGETATION PAT1-ERNS While the fonmation of the coastal terrace at To'aga was primarily a geomorphological phenom- enon controlled by such extemal factors as sea level change, teonics, and their effect on sediment budget (see Kirch, chapter 4), humans have also The Natural and Cultal Environment 19 played an active role in the evolution of this land- fonm. Inchar5, for example, we s argue on the basis of our sratigraphic data that the input of tIrestrial sediments onto the coastl terace in- creased after human colonization due to forest clearance and agnicultral activities on the talus slope and mounain. PeNaps the most obvious effect of humans, however, results frm the pate s of land use and vegetaion on the coastal terrace, for the whole zone constitutes an an ogenic, managed environment from the phytogeographic perspetive. The coastal terrace of Ofu provides the only flat land on the island, the setting for both villages and for certain Inds of intsive horticul- tura and aiboncultual production. The develop- ment of this pattem of intensive land use is one prbem that we have attempted to addrss in our archaeological investigations at To'aga As back- gound to this study, it is necessary to charactenize te pattens of land use and vegetaton found on the Toaga coastal terrace today. Vegetion Transects Yuncker's (1945) study of the Manu'an flora enumerated the plant sources of the islands, but neglected patterns of plant distribution. In 1987 and 1989 we reoorded the horizontal distribution of dominant species along three of our trasc across the To'aga coastal terce that had been cleared for archaeological subsurface sampling (see Kirch and Hunt, chapter S for futer details of hese transets). One of xse, Trnsect 7, is graphically depicted in figure 25. Although there are minor differences in the distribution of species, the same overll pattem is evident in all trsec . Whisler (1980) described and ilustrated most of the species recorded in these tran s. Three main vegetation zones can be discerned, from seaward to ladward: 1. Strand Vegetation. Beginning at the high- water mark and extending to the seaward edge of the coastal road is a zone dominated by halophytic, liUttoral species. Ovehganing the beach are Scaevola taccada and Messerschmidia argentea, with the vines Canalia maritima and Jpomoea pes-caprae trailing over the sands. Larger trees sunnounting the beach ridge inlude Barringtonia asiatica (traditon- ally used for fish poisoning), Hernandia nyphae4folia, Cocos nucifera, and Terminala samoensis. Pandaiws tectorius and Hibiscus tiliaceous shrubs dominate the iner edge of this zone, lning the sandy radway. 2. Arborlulural-Horticultural Zone. Commencing on the inland side of the road, and extending acrmss the width of the coastal terrace to the base of the talus slope is the main zone of economic plants. Coconut palms dot the area, especially toward the seaward half of the zone, where the soil is sandier. The badfruit tree, Artocarpus altilis, commences not far frm the roadway and is the main upper story dominant across the terace. Under and between the beadfruit and coconuts are planted a number of fruit and root cropS, the most important being Ewnusa banas and the large arid, Alocasia macrrorhiza. Some taro, Colocasia esculent and the historically inrduced American amid Xanthosoma saggitolia, are also found, although in lower frequency. The Alocasia aroids are often densely planted in clearings which, after cropping, ae secondarily planted in bananas. In sone areas, the undestory beneath the coconut palms and badfruit trees is a tangle of second growth shubs, dominated by Hibiscus tiliaceous and Macaranga st4ulosa. Othr useful trees occuring less frequently though the zone include mosooi (Cananga odorata), the flowers of which are used for scening coconut oil, andfisoa (Colubrina asiatica), which has medicinal value and can be used as a soap substitute (Whistler 1980:41). This zone of tree, fruit, and root crops also exhibits a high frequency of feral or natalized species which are commonly cultivated in Oceanic agricultural systems, and which unquestionably are present in the Toaga area as survivals frm an earlier phase of (presumably) more intensive cultivation. These include the t plant, Cordylinefruicosum, the arrowrot, Tacca leontopetaloides, the bitter yam, Dioscorea budbfera (which twines in gat abun- dance over the trees and shrubs of this zone), and in lesser quantities, the a'a or Pueraria Lobata. All of these plants are mcognized by the Ofu people as having edible subtemnean roots or tubers and are regarded as potential famine resources. They are part of the complex of plants Barrau (1965) has termed "witnsse of the past," indicators of eatlier cultivation praices in the Pacific islands. 3. Talus Slope Vegetation. The main zone of economic plants terminates abrupdy at the base of e talus slope, strewn with large volcanic boulders. 20 The To'aga Site Alocasia macrorrhiza Colubrina asiatica Morinda citrifolia 4I i 4-- 4~~~~~~~~~~~- Cordy/ine fruticosum Artocarpus a/tilis Macaranga sp. Tacca leontopetal/odes Dioscorea bulbifera Musa hybrids -4 Hibiscus tiliaceous i Cocos nucifera Pandanus tectorius - . Barringtonia asiatica -' @Hemandia nymphaeifo/ia I4 e MassarsMhmida arnantna 10 *\ 0 8 * Scaevola tacc4 D . ARBORICULTURAL ZONE fi - _ _ _ _ _ _ _ ada Figure 2.5 Distribution of major floral dominants along Transect 7 at To'aga A few breadfruit trees and the occasional banana plant may be found extending a few meters up the slope, but the main dominant here is Hibicus tiliaceous, which fonns a dense tangle over the boulders, and larger forest trees such as Erythrina varigata and others. Land Use and Site Formation Although no longer a locus of pennanent village habitation, the To'aga coastal terrace remains an important zone of intensive horticul- ture and arboriculture for the Ofu Island popula- tion, as indicated by the analysis of vegetation patterns. Indeed, on an island where flat land comprises less than nine percent of the total area, the economic importance of this coastal terrace to the human population cannot be overemphasized. It is certain that the pattem of intensive cultivation of this area extended back at least into late prehis- tory, but it remains an archaeological problem to deternine just when this pattem first developed. The pattem of land use in prehistory must also have had consequences for archaeological site formation processes. For example, the clearance of indigenous forest cover on the talus slope above the flat would have exposed unstable soil and rock, and thus accelerated erosion and deposition of colluvial sediment onto the coastal terrace. Cultivation on the flat itself would have resulted in a continual reworking of the upper soil layer (through the actions of digging sticks as well as 4- -4 4 -- -4 v . # i - The Natural and Cultural Environmen 21 thrugh floral-turbation by plant roots and tubers). The mixing of terruginous and calcareous sedi- ments thugh cultivation would have created a well-drained, highly fertile edaphic medium which itself was probably more suitable for root crop cultivation than either the heavy colluvial clays, or the calcareous sands themselves (see Kirch and Yen 1982). In the chapters to follow, we pay particular attention to several lines of evidence that point to the gradual, historical development of the inten- sive land use pattem at To'aga. These lines include evidence for changes in the rate of deposi- tion of colluvial sediments, stratigraphic evidence of buried soil surfaces, evidence of reworking of soils, and the presence of several species of synantrpic land snails that are markers of Oceanic horticultural activities. REFERENCES CITED Barrau, J. 1965. Histoire et prehistoire horticoles de l'Ocdanie tropical. Journal de la Societe des Ocdanistes 21:55-78. Braudel, F. 1980. On History. Chicago: Univer- sity of Chicago Press. Buck, P. 1930. Samoan Material Culture. Hono- lulu: Bernice P. Bishop Museum Bulletin 75. Burrows, E. 1936. Ethnology of Futuna. Honolulu: Bernice P. Bishop Museum Bulletin 138. Buxton, P. A. 1930. Description of the environ- ment. Insects of Samoa, Part IX, Fasc. 1. London: The British Museum. Coulter, J. W. 1941. Land Utilization in American Samoa. Bemice P. Bishop Museum Bulletin 170. Honolulu. Curry, L. 1962. Weather and climate. IN J. W. Fox and K. B. Cumberland, eds., Western Samoa: Land, Life, and Agriculture in Tropical Polynesia, pp. 48-62. Christchurch: Whit- combe and Tombs. Daly, R. A. 1924. The geology of American Samoa. Carnegie Institution Publication 340:93-152. Farrell, B. H. 1962. The village and its agrculture. IN J. W. Fox and K. B. Cumberland, eds., Western Samoa: Land, Life, and Agriculture in Tropical Polynesia, pp. 177-238. Christchurch: Whitcombe and Tombs. Finney, B. 1985. Anomalous westerlies, El Niflo, and the colonization of Polynesia. American Anthropologist 87:9-26. Firth, R. 1936. We The Tikopia. London: George Allen and Unwin. Flenley, J., and M. King 1984. Late quatemary pollen records from Easter Island. Nature 307:47-50. Fosberg, R. 1963. Disturbance in island ecosys- tems. IN J. L. Gressitt, ed., Pacific Basin Biogeography, pp. 557-61. Honolulu: Bishop Museum Press. Friedlander, I. 1910. Beitrage zur der Samoa Inseln. K. Bayer. Akad. Wiss., Math-Phys. El., Munich, Band 27:358-69. Green, R. C. 1969. Archaeological investigation of Westem Samoan prehistory. IN R. C. Green and J. Davidson, eds., Archaeology in Western Samoa, Vol. 1, pp. 3-11. Bulletin of the Auckland Institute and Museum, Auckland. Handy, E. S. C., and W. C. Handy 1924. Samoan House Building, Cooking, and Tattooing. Bemice P. Bishop Museum Bulletin 15. Honolulu. Hiroa, Te Rangi (P. H. Buck) 1930. Samoan Material Culture. Honolulu: Bemice P. Bishop Museum Bulletin 75. Irwin, G. 1981. How Lapita lost its pots. Journal of the Polynesian Society 90:481-94. . 1992. The Prehistoric Exploration and colonisation of the Pacific. Cambridge: Cambridge University Press. Jordan, D. S., and A. Seale 1906. The Fishes of Samoa. Bureau of Fisheres Bulletin 25, pp. 173455. Washington D.C.: Government Printing Office. Kirch, P. V. 1978. Indigenous agriculture on 'Uvea (Westem Polynesia). Economic Botany 32: 157-81. 1982. The impact of the prehistoric Polynesians on the Hawaiian ecosystem. Pacifc Science 36:1-14. 1983. Man's role in modifying tropical and subtropical Polynesian ecosystems. Archaeol- ogy in Oceania 18:26-3 1. . 1984. The Evolution of the Polynesian Chiefdoms. Cambridge: Cambridge Univer- sity Press. -. 1988. Niuatoputapu: The Prehistory of a 22 The To'aga Site Polynesian Chiefdom. Thomas Burke Memo- rial Washington State Museum Monograph No. 5. Seattle: Burke Museum. . forthcoming. The Wet and the Dry: Irrigation and Agricultural Intensification in Polynesia. Chicago: University of Chicago Press. Kirch, P. V., and D. E. Yen 1982. Tikopia: The Prehistory and Ecology of a Polynesian Outlier. Honolulu: Bemice P. Bishop Mu- seum Bulletin 238. McDougall, I. 1985. Age and evolution of the volcanoes of Tutuila, American Samoa. Pacific Science 39:311-20. McGlone, M. 1983. The Polynesian deforestation of New Zealand: A preliminary synthesis. Archaeology in Oceania 18:11-25. Mead, M. 1930. Social Organization of Manua. Honolulu: Bernice P. Bishop Museum Bulletin 76. Menard, H. W. 1986. Islands. New York: Scien- tific American Library. Nakamura, S. 1984. Soil Survey of American Samoa. Soil Conservation Service, U.S. Departnent of Agriculture. Washington D.C.: Govermnent Printing Office. Natland, J. H. 1980. The progression of volcanism in the Samoan linear volcanic chain. Ameri- can Journal of Science 280-A:709-35. Olson, S., and H. James 1984. The role of Polynesians in the extinction of the avifauna of the Hawaiian Islands. IN P. S. Martin and R. G. Kline, eds., Quaternary Extinctions, pp. 768-80. Tucson: University of Arizona Press. Sahlins, M. D. 1976. Culture and Practical Reason. Chicago: University of Chicago Press. -. 1981. The stranger-king, or Dumezil among the Fijians. Journal of Pacific History 16:107-32. Steadman, D. 1989. Extinction of birds in Eastem Polynesia: A review of the record, and comparisons with other island groups. Journal of Archaeological Science 16:177- 205. Steadman, D. W., D. S. Pahlavan, and P. V. Kirch 1990. Extinction, biogeography, and human exploitation of birds on Tikopia and Anuta, Polynesian outliers in the Solomon Islands. Bishop Museum Occasional Papers 30:118- 53. Steadman, D. W., and P. V. Kirch 1990. Prehis- toric extinction of birds on Mangaia, Cook Islands, Polynesia. Proceedings of the National Academy of Sciences U.S.A. 87:9605-9609. Steams, H. T. 1944. Geology of the Samoan Islands. Bulletin of the Geological Society of America 55:1279-1332. Stice, G. D., and F. W. McCoy, Jr. 1968. The geology of the Manu'a Islands, Samoa. Pacific Science 22:427-57. Visher, S. S. 1925. Tropical Cyclones of the Pacific. Bernice P. Bishop Museum Bulletin 20. Honolulu. Wading, D. 1982. Birds of Fiji, Tonga, and Samoa. Wellington: Millwood Press. Whistler, W. A. 1980. Coastal Flowers of the Tropical Pacific. Honolulu: Oriental Publish- ing Company. Yuncker, T. G. 1945. Plants of the Manua Islands. Honolulu: Bernice P. Bishop Museum Bulletin 184.