«MARA ELENA ROSENTHAL IN HIS CLASSIC TEXT on Oceanic canoes, Piroques Oceaniennes (1976), Father Jean N eyret argued that the canoes of the Fiji ...»
The Archaeological Excavation of
an Outrigger Canoe at the Nasilai Site,
Rewa Delta, Viti Levu, Fiji
MARA ELENA ROSENTHAL
IN HIS CLASSIC TEXT on Oceanic canoes, Piroques Oceaniennes (1976), Father Jean
N eyret argued that the canoes of the Fiji Islands were among the "masterpieces of
Oceanic naval construction." According to Neyret, this was due to Fiji's location
at the "crossroads" of the Pacific, which permitted the successful integration of diverse Polynesian, Melanesian, and Micronesian maritime technologies to pro- duce vessels that were uniquely rapid, solid, and seaworthy (Neyret 1976: 47).
This paper reports on the 1987 archaeological excavation in Fiji of portions of a late prehistoric to early historic canoe at the site of N asilai. The evidence collected indicates a canoe that unites diverse Oceanic maritime technologies as described by N eyret. Further, the vessel revealed by the excavations provides the first direct example of a riverine outrigger canoe of a type previously known only from historical reference and secondary description. Referring to this "Type Ancien," Neyret writes: "This model is very interesting, as it reveals characteristics of Fijian canoes of 100 years ago, before European influence was evidenced in matters of naval construction" (Neyret 1976: 59; my translation).
This paper presents a description and analysis of the Nasilai remains, excep- tional in the quality of their preservation and their significance to the growing understanding of Oceanic maritime technology and the sociopolitical develop- ments it made possible.
Although Pacific island communities have been classically portrayed as ideal "laboratories" for the study of bounded social groups developing in pelagic isolation (Kirch 1984: 2-4, 20), recent research suggests that a "sea of islands" (Friedman 1982; Hau' ofa 1994; Kirch 1984: 82) perspective-a view of waters and landmasses as a total integrated Pacific setting-may represent a more accurate portrayal of the Pacific context, at least with regard to late prehistoric sociopoliti- cal developments in western Polynesia and Fiji (Kaeppler 1971, 1978; Rosenthal 1991, 1993). Late prehistoric developments in naval technology like those evidenced at Nasilai indicate that the facilitation of maritime exploitation and Mara Elena Rosenthal is a research associate at the Smithsonian Institution and a visiting fellow at the East-West Center, University of Hawai'i at Manoa, Honolulu, Hawai'i.
© 1995 by University of Hawai'i Press.
Asian Perspectives, Vol. 34, No.1, 34(1). 1995
ASIAN PERSPECTIVES. SPRING92 LOWER
increased ease of interisland interaction may have been integral to the rise of the complex, late prehistoric sociopolitical forms of western Polynesia and Fiji.
The site of Nasilai is located at 18°06'S longitude and 178°34'E latitude in the southeastern segment of the main Fijian island of Viti Levu. It lies on the outside of a broad meander of the Bucona River, one of the tributaries of the Rewa River in the lower part of the delta (Fig. 1). The area is a prograding delta subject to extreme flooding. Aerial photo analysis and historical evidence indicate that there has been some shifting of river courses in this area within the past several hundred years (Parry 1977).
The area is presently settled by inhabitants living in small nucleated villages of 50-200 people as well as in a few larger village clusters. The area is intensively cultivated primarily in cassava and taro, including the giant swamp taro, or via kana (Cyrtosperma chamissonis), for which the area is famed.
The 1987 archaeological excavations were situated on the outskirts of the village of Nasilai adjacent to the Bucona River. Nasilai is one of several villages among a dense cluster of villages or residential sections along the Bucona tributary each of which performs specific traditional duties and directly serves the Rewan chiefly capital of Lomanikoro. Nasilai village is inhabited by fishermen who serve the Sacred King of Rewa, the Roko Tui Dreketi. Representing one
ROSENTHAL. FIJIAN OUTRIGGER CANOE 93of the traditional roles performed by specific groups within a Fijian kingdom, the Nasilai villagers are kai wai, literally, "water peoples," serving as fishermen and traditional navigator clans for their chiefs.
The 1987 excavations were precipitated by an episode of flooding that caused severe erosion of a segment of the bank of the river adjacent to the village of Nasilai, Rewa. The erosion exposed a number of worked wooden objects, which were brought to the attention of the Fiji Museum. IdentifYing the artifacts as a bokola (cannibal victim) serving dish and two probable canoe planks, the Fiji Museum authorized test excavations at Nasilai under my direction.
In early 1987, three and a half months of archaeological field excavations were conducted at the site. The goal of the work was to determine the nature of the occupation levels and cultural remains. 1 At the outset of the fieldwork, exploratory cores were sunk into the site area at stratified unaligned intervals of two m in order to test for the presence or absence of cultural deposits across the area within the confines of the riverbank and the presently inhabited area of the village. Organic deposits indicative of a possible cultural level were detected across the entire site area. Excavation pits were thus located as indicated in Figure 2 in order to provide a cross-sectional view of the central site area in both north/ south and east/west directions. A total of five 1.5 m 2 pits with one 50 cm extension were sunk during the 1987 field season. Placement of the pits turned out to be fortunate indeed, as the boundaries of a significant structure-a post-lined ditch-within which the canoe fragments were deposited, was identified.
Waterlogged conditions at the site created reducing or anaerobic conditions that allowed for the exceptional preservation of organic remains at the site, including worked wood, woven and plaited reeds, magimagi or sennit cord, and leaf, nut, and seed material. Pumps were run continuously in order to clear the pits for excavation. All soil removed from the pits was screened through 0.5 cm mesh. Spot heights were recorded for all artifacts before they were lifted. Field descriptions were recorded, measured plans drawn, and photographs taken every 10-15 cm as the excavations progressed. Detailed section drawings were made once excavation of the pits was completed. All objects removed from the Nasilai site were brought to the Fiji Museum for conservation by Gladys Fulman, acting director of the Fiji Museum.
THE NASILAI STRATIGRAPHY AND DEPOSITIONAL HISTORY
Nasilai site geology and depositional history were researched by the author in conjunction with Patrick Nunn, senior lecturer in geography at the University of the South Pacific. Figure 3 indicates the stratigraphy of the site as evidenced in the northernmost section. According to Nunn, no true soils were observed at the site. Several indications of stratification are present, though these are generally poorly formed and discontinuous from one exposure to another. Of particular note, Nunn identifies one clear stratigraphic discontinuity between the gray sediments associated with the canoe deposits of Level IV and the orange-brown silts (Levels I, II, and III) that overlie them (Nunn, personal communication 1987).
34(1). 1995 94 ASIAN PERSPECTIVES • SPRING,,.,,
Fig. 2. Nasilai Level IV site plan: 1, pilaster; 2, basal portion of mast; 3, club head; 4, outrigger float portion (fragmentary); 5, cabin buttress; 6, stanchions. * = posts.
The upper three levels at Nasilai are distinguished on the basis of gradations in sediment color, texture, and density. Level I is comprised of loosely packed, structureless or thinly laminated silt, dark brown in color (7.SYR 3/4). Level II is hard-packed, structureless or thinly laminated silt, dark yellowish brown in color (10YR 3/2), and Level III is hard-packed, moist structureless silt, yellowish red in color (SYR 3/6). It is the type of sediment found in Level III that is utilized by the villagers today as clay for pottery production.
ROSENTHAL • FIJIAN OUTRIGGER CANOE 95
Cultural features within the upper three levels included three former cultivation pits in Levell. In addition, two pits in 0N0W, one beginning in Level I and the other in Level II, were cut down into the lowest-lying blue-gray clays. Historic artifacts, ceramics, and shell found in these pits suggest that they were refuse dumps. There is no evidence of other cultural features in Levels I, II, and III.
The depositional history of the orange-brown silts of Levels I, II, and III is
described by Nunn:
On account of sorting and granular characteristics, (these levels are thought)... to have been accumulated on a point-bar surface on the inside of a meander. Most of the deposits were probably laid down during high-water stages... likely to have been linked to tides and moderate, regular floods rather than extreme, uncommon events. The discontinuities which can be observed in this deposit probably indicate localized erosion... suggesting that this part of the record is... incomplete in both space and time. The presence of sand, perhaps as much as 15-20 percent of this deposit in places, may indicate aeolian deposition and date from a time when the site was open to the sea.
The final environmental change is from the point-bar aggradation environment indicated by the orange-brown silts to the present day erosional situation. This can be explained most easily by shifts in the main river channel and changes in its regime. (Nunn, personal communication 1987) Under the above-mentioned sediments lies Level IV, the canoe-bearing stratum. It is comprised of moist silty clay that is dark-gray in color (SY 4/1).
Deposits are dense in organic material including seeds, leaves, plaited and woven reed, and fibers, as well as numerous worked wood artifacts. The frequency of ceramics is high throughout this level. A sharp discrepancy is found in the frequency of shell material within these gray silty clays. Upper and lower horizons were distinguished in level IV on the basis of the density of shell and organic material. The upper horizon is almost barren of shell. However, underlying the densest patches of woven and plaited reed and fibrous material, the lower gray 34(1). 1995
ASIAN PERSPECTIVES • SPRINGsediment horizon evidences extremely dense shell deposits, including two distinct shell lenses running across 6N3W. With the exception of one piece of porcelain from 4N1 W, no European historic objects were recovered from Level IV. Level IV is deposited within the confines of a post-lined ditch structure dug
into Level V. According to Nunn:
the sediments within this main occupation layer are well sorted, their coloring suggestive of reducing or anaerobic conditions. Sediments appear to have been deposited in the context of slow-moving water conditions. Within and in some places just above [this] main occupation layer is found the prominent disconforrnity which is probably an erosional unconformity. This implies that accumulation of the [sediments] continued but then, by the time the character of the deposition was starting to change into the overlying orange-brown silts, a major erosional event, undoubtedly flood related, took place and perhaps as much as 60 cm of material were removed from the contemporary surface. Clearly, this could have included human artifacts. The figure of 60 cm is a maximum based on the assumed character of the "missing" transition zone.
The former presence of a transition zone between the two contrasting strata is supported by considerations of paleoenvironments. The most likely reason for the filling of the backwater/swamp is that movements of the main river channel had brought its outlet into a position on the inside of a broad meander bend. Deposition in such an environment is indicated by the orange-brown silts. Further, when the post-lined ditch structure was constructed, probably on the side of a creek full of standing water at high tide (the inhabitants)... undoubtedly found that they had to excavate a large area to allow (continuing) access to the main river. Such excavations probably caused a comparatively abrupt transformation of the environment from a backswamp to point-bar condition, as suggested by the sediments, although there where undoubtedly several major scour events. (Nunn, personal communication 1987) The level below the canoe-bearing stratum, Level V, is hard-packed, moist fine-grained silt, yellowish red in color (5YR 3/6) with some gray mottling. No shell and very little pottery is found in this layer. Some seed material is present.
Parallel rows of worked wooden posts were set into Level V. Further, as indicated above, a ditch was intentionally dug into Level V. Level IV, above, was deposited within the confines of posts lining the ditch excavated into Level V.
The underlying Level VI is loosely packed, moist clay with some patches of sand. It is very dark-gray in color (N 4/1.5, 7.5R 3/0) as a result of reducing or anaerobic conditions. No shell, very few pottery fragments, and a small quantity of seed are found in this level. Decayed unworked wood is evident. No indication of cultural features and no historic artifacts were found at this level. According to Nunn: "It seems likely that [these sediments] accumulated in a backwater or swamp area. Were more of these clays exposed, they would probably exhibit a change in colour from bottom to top which would show the intensification of reducing conditions interpreted, from the limited observations made, as gradual infilling of the backwater or swamp" (Nunn, personal communication 1987).
In sum, the stratigraphic evidence indicates that the canoe fragments were deposited within Level IV and are associated with an intentionally excavated ditch that was lined with parallel rows of posts and dug into the underlying Level V.