|
Article Excerpt
Abstract
We examine radiocarbon dates from Samoan archaeological sites using the fourfold division of Samoan prehistory established by Green (2002). The context of dating samples was assessed to recognize potentially "reliable" determinations in the Samoan [sup.14]C corpus. Radiocarbon dates associated with earth and stone structures were identified to one of four phases of construction/use to develop a chronology for the emergence and use of domestic and monumental architecture. The 17 radiocarbon determinations from the Pulemelei mound site were used to generate a local prehistoric sequence for the Letolo area. In general the results parallel the prehistoric sequence for Samoa, but the chronology of the Pulemelei mound demonstrates that monumental architecture in West Polynesia can have a complicated developmental history spanning several centuries.
**********
The first radiocarbon dates from the Central Pacific were obtained from Fiji by Edward Gifford (1951), and it was not until the 1960s that the first [sup.14]C determinations from Samoa, on archaeological samples collected by Golson in 1957 (Golson 1969a), showed that Polynesian pottery had an antiquity of at least 2000 years (Grant Taylor and Rafter 1963; Green and Davidson 1965). Subsequent archaeological work resulted in radiocarbon dates from significant investigations published by Green and Davidson (1969, 1974a), and Jennings and colleagues (1976, 1980).
From 2002 to 2004 excavations by the authors at the Pulemelei mound site on Savai'i provided new radiocarbon dates relating to the extensive prehistoric structures and features that had been mapped in the Letolo Plantation in the 1970s (Jennings et al. 1982). The radiocarbon results from earlier and recent archaeological projects provide the data to construct a prehistoric sequence, particularly of the last 1000 years when large mounds emerged. We have not included determinations from American Samoa as the prehistoric sequence of the small islands appears to differ from that of Samoa, particularly the absence of monumental mounds, and possibility that pottery manufacture lasted longer in American Samoa than it did in Samoa (Clark 1996; Green 2002).
In previous research Samoan prehistory has been viewed as an aperiodic cultural succession (Green and Davidson 1974a). However, the development of the settlement pattern in prehistoric Samoa, which has chronological connotations, has recently been suggested by Green (2002:134-146).
1. Initial settlement represented by distinctive Lapita ceramics;
2. Settlement patterns during the period when Polynesian plainware was produced;
3. The interval when evidence of the settlement pattern is extremely limited (the so called 'Dark Age');
4. Late prehistoric settlement patterns marked by the construction of earth and stone structures.
Green's settlement sequencing is based on evidence of the use of pottery with patterns/no patterns, use of pottery/ no pottery, the apparent absence of settlements activities, and the use of large stone and earth mounds. Tied to the discussion are various dates presented for the changes in the settlement pattern.
To date 89 [sup.14]C determinations (Table 1) have been reported for Samoa by Green and Davidson (1974b:214-5), Jennings and Holmer (1980:7-10), and from our investigations at the Pulemelei mound (Martinsson-Wallin et al. 2003, 2005; see Martinsson-Wallin et al., this publication). In this paper we outline the Samoan radiocarbon sequence. Following the discussion of radiocarbon dates for Samoa as a whole, we discuss the dated samples from the Pulemelei mound site to construct a detailed chronology for the Letolo Plantation.
Site Location
Radiocarbon assays have been obtained for archaeological sites on Upolu, Savai'i, and Manono. The excavated sites are listed (see Martinsson-Wallin this publication, Table 2). Many additional sites are known, but have yet to receive archaeological excavation (e.g. Green and Davidson 1969, 1974a). The intention of archaeological programs in the 1960s and 1970s was to investigate a range of coastal and inland sites (Green 1969a:3-11). The investigation of inland sites include prehistoric deposits and structures from the Falefa Valley on Upolu, and the Letolo area on Savai'i, in addition to prehistoric and historic settlements at Mt Olo and Luatuanu'u on Upolu. Coastal sites have been investigated at Vailele, Faleasi'u and Lotufaga on Upolu, and on Manono and Apolima (Green and Davidson 1969, 1974a; Jennings et al. 1976, 1980).
Methodology
All charcoal samples reported here have been calibrated using CALIB (Version 5.0.1, 1986-2005 based on Stuiver and Reimer 1993), with the Southern hemisphere correction data set (SHCal04, McCormac et al. 2004). Marine shell determinations were calibrated with the marine correction data of Hughen et al. (2004). The local marine reservoir value (Delta R) was set at 57 [+ or -] 23 years. This value was calculated on a marine shell collected from Upolu (Phelan 1999), and was also used by Petchey (2001) when calibrating marine shell dates from the Mulifanua Lapita site. A recent examination of selected Samoan marine shell samples by Smith (2002:93-125) employed a slightly lower Delta R value of 45 [+ or -] 30 years. Although neither Delta R value makes much difference to the calibrated results, there are also a few inconsistencies in the conventional radiocarbon ages (CRA) values from Samoa reported by Smith (2002:110), which affect the interpretation and discussion of radiocarbon results from the Jane's Camp site (see below). All calibrated dates are reported at two standard deviations.
Determinations on marine shell samples (NZ-4342B/ UGa-1485 and NZ-2726/7/8B) have been pooled when dating of the same marine shell gave similar CRAs. Some radiocarbon ages which do not appear to convincingly date prehistoric cultural activity have been excluded from Table 1, as have dates with modern calibrated ages. From the determinations reported by Jennings and Holmer (1980: 7-10), we reject UGa-1671 (14920 [+ or -] 175 BP), which is too old considering the generally accepted chronology of human entry to West Polynesia at 900 BC (Burley and Dickinson 2001), and RL-479 (3220 [+ or -] 130 BP), as it is not convincingly associated with prehistoric cultural activity, along with a modern result UGa-1486 (35 [+ or -] 70 BP). Among the dates reported by Green and Davidson (1974b) the following modern results (no CRAs reported) were excluded: GaK-1342, which was a contaminated charcoal sample, while NZ-1427 and NZ-1431 were on post wood from a recent house. NZ-854 has a reported CRA of 352 BE but no standard errors were given and it was excluded from our analysis.
A problem with some of the Samoan radiocarbon samples are that 31 samples were carried out by the Gakushuin Laboratory in Tokyo. It has been argued that dates up to c. GaK-4500 may be erroneous (Spriggs and Anderson 1993:207). However, these dates are included here, since they seem to give similar dates to other laboratories concerning what they are expected to date.
Radiocarbon dating of settlements and stone/earth structures
Prehistoric settlements consisting of structures built in earth and stone can be difficult to date accurately with radiocarbon, and [sup.14]C samples were evaluated according to whether or not they had a clearly identified archaeological context.
In situ fireplaces and earth ovens are prehistoric features that should provide relatively reliable radiocarbon determinations on wood charcoal. In both cases the burned wood derives from a localized feature that was constructed by prehistoric people, which links the sample to a specific action taking place over a short timeframe....
|
