6 RADIOCARBON CHRONOLOGY OF THE TO'AGA SITE PATRICK V. KIRCH T HE AGE OF ARCHAEOLOGIcAL deposits at To'aga may be roughly estimated by typological comparison of te excavated cermnics and basalt adzes with dated assemblages of similar artifacts from Westem Samoa (Green and Davidson 1974), and from elsewhere in Westem Polynesia (e.g., Kirch 1988; Poulsen 1987). Such comparisons suggest that prehistoric occupation at To'aga spanned much of the first millennium B.C. and on into te Christian era Typological cross-dating is always suspect, however, as it depends upon the assumption that stylistically similar artifacts frm different islands were used contemporaneously, and thus ignores the possibilities of cultral lag or of indepen- dent rates of change. For example, there is no particular reason to assume, a prioin, that the change from Early Eastem Lapita pottery (characterized by dentate stamnped designs) to Late Eastem Lapita should have occurred at the same time trughout the Samoan archipelago. hIneed, the specific aspects of ceramic change may themselves have varied spatially over the Samoan Islands. Likewise, the cessation of ceramic manufacture and use was probably not a contemporaneous event trughout Westem Polynesia, as indicated by relatively late dates for pottery from Niuatoputapu (Kirch 1988:13942). In order to obtain a more precise chronology for the To'aga site, and an independent assessment of the age of the varous arfifact assem- blages sampled from the To'aga deposits, fourteen samples of charcoal and shell were radiocaibn dated. This chapter psnts provenience and sample details for each of ftese dates, and discusses some of the broader implications of the To'aga site chmnology for Samoan prehistory. An initial assessment of the age of pottery- bearing deposits at To'aga was obained by dating a sample of Turbo shell midden obtained from the 1986 reconnaissance test pit adjacent to the landfill dump. This sample (Beta-19742) and another pottery-associated sample from Ta'u Island (Beta- 1974 1) were first reported by Hunt and Kirch (1987). The sanples were close in age, with a weighted average of 2605 ? 56.6 B.P., caibraed to 360-50 cal B.C. aftr corrcion for marine rervoir effects. Hunt and Kirch noted that these dates from Manu'a were "in close contemporaneity with corrected radiocarbn assays of wood charcoal associated with stylistically and technologically similar ceramic assemblages from tUpolu Island in Westem Sanoa" (1987:417). At the conclusion of the 1987 field season, an additional seven samples (six shell, one charcoal) were submitted to Beta Analytic for 14C assay. These dates were reported by Kirch, Hunt, and Tyler (1989), and provided "a statigphically consistent chrnologic sequence for 86 The To'aga Site human occupation of Ofu Island, spaning the period from ca. 3700-3300 B.P. up to the modem era" (1989:10). At the end of the 1989 excavation season, six more samples (two shell, four charcoal) were submitted to Beta Analytic for 14C age assess- ment These 14C ages ar reported here fortfhfirst time, along with their contexual details. The 1989 samples all proved to be irnally consistent, both saigraphically and in comparison with the 1987 14C series. In swn, a suite of fourteen radiocarbon dates are now avalable for the To'aga site, ranging in age from the second millanium B.C. up to fte end of fte first millennium A.D. All dates are stratigraphically consistent, and in aggregate provide a firm basis for establishing the chrnlogy of human occupation and of geomorphological change at To'aga. Because this series spans viually the entire "ceramnic period" of Samoan prehistory, it is also of considerable significance for the cultural history of the archi- pelago as a whole. RADIOCARBON DATING PROCEDURES: CORRECTIONS AND CAUBRATIONS hi selecting samples for radiocarbon dating, emphasis was placed on material which was in firm stratigraphic context (with no apparent distubance or inversion), and in obvious association with artifacts or featur. When possible, charcoal samples were obtained from hearth or oven features, although in four cases finely dispersed charcoal flecks had to be dated in order to ascertain the age of features of particular interest. The charcoal samples consisted of wood charcoal (species not determined) mixed in some cases with carbonized fragments of coconut endocarp (Cocos nucifera). With the exception of two shell sanples from the stratigrahically oldest beach deposits, the shell selected for 14C dating was culurlly-modified midden material about which there was no uncer- tainty as to culturl associations. All dated shell is of marine origin, consisting of shallow-water, reef- dwelling gastrpods and bivalves. Whenever possible, we seleced the species Turbo setosus for dating, as a further control on consistency of results. This reef gastpod is one of the dominant compo- nents of the shell midden assemblages from the To'aga site (see Nagaoka, chapter 13). All 14C measurements on samples from Toaga were performed by Beta Analytic, Ic. Shell samples were pretrated by etching away the outer layers with dilute acid, to remove any adhering contaminants. The smples were then atacked with further acid to produce carbon dioxide, which was used as Xt carbon source for dating. The charcoal samples were first picked manually for rootlets, and then given a series of acid, alkali, and acid ings to remove carbonates and humic acids; they were then rinsed repeatedly to neutality. One sample (Beta-35600), consisting of very finely dispersed organics (chamoal and ofter organic material), was picked for rootlets and then dispersed in hot acid to eliminate carbonates. All benzene syntheses and countings proceeded nonnally (M. Tamers, pers. comms., 21 July 1988; 6, 13 March 1990). Four samples (three of charcoal and the organic sample) had only small quantities of carbon remaining after ptreatment, and thus were accorded extended counting times (four times the normal duration) in order to reduce statistical eror as much as practi- cable. The 13C/14C ratios were measured for all samples to establish 13C adjusted, "conventional 14C ages" following the wecommendations of Stuiver and Polach (1977). The nonadjusted "'Libby ages," conventional 13C adjusted ages, and a13C values for all samples are presented below. Calibration of conventional 14C ages for secular effect, and in the case of marine shell samples for reservoir effect as well, were made following the rece work of Stuiver, Pearson, and Braziunas (1986) for shell samples, and of Stuiver and Becker (1986) for fte tenestrial charcoal and organic samples. All calibrations and probability estimates were obtained with the use of the CALIB microcom- puter program (Rev. 2.0) developed by Stuiver and Reimer (1986). As discussed by Kirch and Hunt (1988b:23), the dating of marine shell samples from southwestem Pacific archaeological sites has posed some prob- lems of calibration due to the "reservoir effect" of older carbon present in the wodrd's oceans. Archae- ologists have recognized that marine shell samples frequently give ages somewhat older than those of charcoal from identical archaeological contexts, and several attempts have been made to develop correc- tion factors for Pacific island shell dates (Gillespie and Polach 1979; Gillespie and Swadling 1979; Jansen 1984; Atxhns 1986; Kirch 1989:139). These Radiocarbon Chronology 87 attempts suggested that southwest Pacific samples of shell from shallow reef envirornents ranged frm about 400-550 years older than the actual age. Recently, Stuiver, Pearson, and Braziunas (1986) produced a first, approximate, global model of txse reservoir effects for both an "'upper mixed layer" of the ocean (relevant to the To'aga samples) and for the deep ocean. Their model takes into account time-dependent changes in 14C activity, and allows for local geographic variations in reservoir effect (e.g., te effect of upwelling from the deep ocean along continental shelves). The local geographic variations are accounted for by a A-R correction factor. Ideally, such a A-R factor should be indepen- dently determined for Samoa by 14C assay of modem, pre-bomb marine shell. As yet, however, this has not been possible. We therefore have used as a A-R correction factor a weighted average of empirically determined A-R values from several mid-Pacific islands, specifically Eniwetok, Hawaii, and Tahiti and Mo'orea in the Society group (Stuiver, Pearson, and Braziunas 1986, table 1). This pooled A-R value of 100 ? 24 represents a reasonable working value for tropical, central Pacific island samples of shallow-water marine organisms. We have also used this A-R value of 100 ? 24 in calibrating radiocarbon dates from a wide range of Lapita sites (Kirch and Hunt 1988a, b). It is signifi- cant that following calibration of the marine shell samples according to the above procedure, these are entirely consistent with the calibrated charcoal dates. This is shown for example by samples Beta-35603 and -35604, both from the same excavation level of Unit 28 in Transect 9, which overlap at one standard deviation. (hat the charcoal sample appears to be slightly older may reflect the use of wood from a tree that had been growing for a century or more prior to cutting and buming.) THE RADIOCARBON CORPUS FROM TO'AGA The fourteen 14C age determinations from To'aga are listed below, by laboratory number, with all relevant information on provenience, associa- tions, samnple details, and calibrations. The uncorrected "Libby age" is given first, followed by the a13C correCted "conventional 14C age" in bold- face type, along with the D13C value in paits per thousand. Calibrations for each date are given at one standard deviation; values within parenteses indicate intercepts on the calibration curve. Beta-19742. 1890 ? 50 B.P.; 2350 ? 50 B.P.; a3C= + 2.9 0/o0. Marine shell (Turbo setosus) from the initial test pit excavation adjacent to the Ofu landfill dump (Layer D, Level 10). The Turbo shell consisted of midden refuse in an organicaly-enriched midden deposit, with calcareous sand matrix. The sample was direcly associated with small quantities of thick, coarse-tempered pottery. Cal 28 B.C. (A.D. 45) A.D. 108 at 1 a; cal B.P. 1977 (1905) 1842 at 1 a. Beta-25033. 2190 ? 80 B.P.; 2640 ? 80 B.P.; al3C = + 2.3 %/o. Marine shell (Turbo setosus, 71 g) from 1987 main excavation trench, Unit 6, Layer IIA-1. The sample consisted of cultrally deposited shell midden in direct association with an earth oven feature, and with small quantities of Polynesian Plainware ceramics. Cal B.C. 362 (244) 145 at 1 a; cal B.P. 2311 (2193) 2094 at 1 a. Beta-25034. 2120 ? 80 B.P.; 2570 ? 80 B.P.; al3C = + 2.5 %oo. Marine shell (Turbo setosus, 70 g) from 1987 main excavation trench, Unit 6, Layer IIB. The sample consisted of culurally-deposited shell midden from the pnincipal occupation deposit in Layer II, in association with Polynesian Plainware ceramics and with Turbo-shell fishhooks. Cal B.C. 295 (161) 58 at 1 a; cal B.P. 2244 (2110)2007 at 1 a. Beta-25035. 3370 ? 70 B.P.; 3820 ? 70 B.P.; a13C = + 2.4 %/o. Marine shell (2 specimens: Asaphis violascens and Lunella cincerea, total weight 48 g) from 1987 main excavation trench, Unit 6, Layer V, 314 cm below surface. Both specimens retained their surface coloration and were not water-rolled, thus indicating deposition in Layer V soon after death. These shells consist of naturally deposited marine 88 The To'aga Site shell (rather than cultural food refuse) in a calcareous beach deposit also containing isolated tn-fine ware ceramic sherds. Cal B.C. 1765 (1682) 1600 at 1 a; cal B.P. 3714 (3631) 3549 at 1 a. Beta-25673. 3170 ? 80 B.P.; 3620 ? 80 B.P.; al3C = + 2.2 0/oo. Marine shell (Phalium sp.,45 g) from 1987 main excavation trench, Unit 1, Layer V, 290 cm below surface. The dated specimen was a single whole gastpod, not waterwom and retaining original surface coloration, thus indicating rapid deposition in the Layer V beach deposit soon after death. Layer V contained isolated thin, fine-tem- pered ceramic sherds. Cal B.C. 1526 (1441) 1377 at 1 ai; cal B.P. 3475 (3390) 3326 at 1 a. Beta-26463. 1460 ? 50 B.P.; 1910 50 B.P.; al3C = + 2.5 %1oo. Marine shell (Turbo setosus, 72 g) from Unit 3 of the 1987 excavation tranet; in Layer 11,40-70 cm below surface. The dated specimen was a single large Turbo shell, with the apenural margin display- ing chipping due to cultual removal of the opercu- lum (presumably to extract the edible soft pails). The specimen was stradgraphically situated at the basal contact of the cultural, aceramic midden deposit and the underlying sterile calcareous beach sand. The specimen's context thus indicates a time period after the commencement of rapid progradation of the To'aga coastal terrace. Cal A.D. 561 (620) 663 at 1 a; cal B.P. 1389 (1330) 1287 at 1 a. Beta-26464. 2660 ? 140 B.P.; 2620 ? 140 B.P.; a13C --27.8 %1oo. Charcoal flecks from 1987 excavation Unit 10, Layer HB, at 70-80 cm below surface. The charcoal was in association with Polynesian Plainware in the Layer IIB occupation deposit. The field sample weighed ca. 1 g, yielding 0.2 g of carbon after pretreatment in the laboratory; the sample was accorded extended counting time. Cal B.C. 967 (801) 454 at 1 a; cal B.P. 2916 (2750) 2403 at 1 a. Beta-2646S. 1160 ? 70 B.P.; 1600 ? 70 B.P.; a13C = + 2.0 %1oo. Marine shell (Turbo setosus, 66.4 g) from 1987 excavation Unit 13, Layer III, 35-45 cm below surface. The dated sample consisted of 1 nearly complete shell and 2 small fragmentary shells, all displaying cultually-induced fractures and chipping (possibly from ardfact manufactre). The Turbo shells were in direct association with an aceramic cultural midden near the base of a pebble-paved house platfoim, in which Unit 13 was excavated. Cal A.D. 828 (914) 1000 at 1 a; cal B.P. 1122 (1036) 950 at 1 a. Beta-35600. 1200 ? 70 B.P.; 1190 ? 70 B.P.; a13C = - 26.1 %/o. Finely dispersed charcoal and ash ("organics') from Trnsect 5, Unit 17,53 cm below surface in the southeast comer of the unit. The field sample weighed 510 g. The charcoal and ash were dis- persed in the interstices of a gravel ('ili'iIi) pavement layer (see chapter 5), presumably a dwelling house floor. Cal A.D. 694 (781, 789, 805, 821, 829, 839, 862) 943 at 1 a; cal B.P. 1256 (1169, 1161,1145, 1129, 1121, 1111, 1088) 1007 at 1 a;. Beta-35601. 2950 ? 110 B.P.; 2900 ? 110 B.P.; al3C = - 27.8 %1o. Charcoal flecks from Transect 5, Unit 28, base of Layer 11,290-300 cm below surface. The field sample weighed 12 g, and produced 0.6 g of carbon after laboratory pretreatment; it was accorded extended counting time. The charcoal was from a non-concentrated midden deposit containing several sherds of a red-slipped, thin, fine-tempered ceramic ware. Cal B.C. 1308 (1188, 1184, 1127, 1126, 1107, 1105, 1083, 1059, 1054) 930 at 1 a; cal B.P. 3257 (3137, 3133,3076, 3075, 3056,3054, 3032, 3008, 3003) 2879 at 1 a. Beta-35602. 2660 ?100 B.P.; 2630 ? 100 B.P.; a13C = - 26.9 0/oo. Charcoal in an ash and sediment matrix (470 g) fm Transect 9, Unit 23. The charcoal was con- tained within an earth oven feaure cut from the upper pan of Layer HIA into Layer IlIB, which contained predominantly thickware pottery, with a small quantity of thinware. Cal B.C. 896 (803) 663 at 1 a; cal B.P. 2845 (2752) 2612 at 1 a. Radiocarbon Chronology 89 Beta-35603. 2660 ? 170 B.P.; 2600 ? 170 B.P.; a13C = - 28.4 0/oo. Charcoal (70g) from Tnsect 9, Unit 23, from the base of Layer IUB, 190-206 cm below surface. The field sample of 70 g yielded 0.15 g of carbon after labortory patment, and was accorded extended counting time. The charcoal flecks were collected from a culturl occupation deposit contain- ing thin, fine-tempered pottery, including some rnms with notched decortion. This sample was from the same depositional context as shell sample Beta- 35604. Cal B.C. 968 (797) 433 at 1 a; cal B.P. 2917 (2746) 2382 at 1 a. Beta-35604. 2330 ? 80 B.P.; 2770 ? 80 B.P.; D13C = + 1.7 0/0o. Marine shell (Tridacna maxima, 455 g) from Transect 9, Unit 23, Layer IIB, 198 cm below surface. The dated specimen consisted of a single valve of T. maxina, showing no signs of water abrasion or rounding; the shell was presumably collected live for food and deposited directly in the cultual matrix. It was in direct association with ceramics and other portable artifacts, in the sane stratigraphic position as charcoal sample Beta- 35603. Cal B.C. 495 (385) 340 at 1 a;; cal B.P. 2444 (2334) 2289 at 1 a. Beta-35924. 1640 ? 70 B.P.; 2100 ? 70 B.P.; a13C = + 2.7 /co. Marine shell (Turbo setosus, 400 g) from Transect 5, Unit 15, Layer II, 60-70 cm below surface. The dated sample consisted of numerous cultually modified (chipped and broken) pieces of T. setosus, representing food refuse and probably also detritus from the manufacture of Turbo shell fishhooks. The shell was dispersed in a dark grey, 'greasy' (carbon-nich) deposit with much fire-cracked rock, y a cookhouse culural deposit. The deposit also contained thick, coarse-tempered ceranics. This is the youngest 14C age from To'aga in direct association with pottery. Cal A.D. 319 (410) 473 at 1 a; cal B.P. 1631 (1540) 1477 at 1 a. The To'aga Sequence The fourteen samples described above provide a stratigraphically consistent series of dates s the period from ca. 3700 to 1000 cal B.P. In figure 6.1 these dates are plotted as calibraed age ranges at one standard deviation, in chronological order. The cultual associations of the samples are indicated on the right-hand side of the diagram. The eau?iest two dates (Beta-25035 and -25673) are frm the Layer V beach deposit of te 1987 main trench excavation and are associated with thin, fine-tempered ceramics. Sanple Beta-35601, also associated with this early type of pottery, is from an in situ cultrl deposit. A suite of four dates (Beta-35602, -26464, -35603, and -35604) is associated with assemblages having significant frequencies of thin, fine-tempered pottery, fishhooks, and other artifacts as well as coarse-tempered thickware. Another suite of four dates (Beta-25033, -25034, -19742, and -35924) is in association with predominantly thick, coarse- tempered pottery, of the sort known in Westem Samoa as "Polynesian Plainware" (Green 1974). Some thinware, however, continues to be repre- sented in these later deposits. These ages indicate a time span of ca. 2400- 1500 cal B.P. for the phase of dominant use of thick pottery. The final three dates (Beta-26463, -35600, and -26465) are from aceramic cultural contexts, and all are younger than 1500 cal B.P. Although we have no 14C dates from To'aga younger than about 1000 Cal B.P., it is lkely that the site continued to be occupied at least until European contact, based on historic artifacts and varous surface archaeological indications (house mounds, grindstones, masi pits, etc.). DISCUSSION The radiocarbon sequence from the To'aga site is the largest suite of dates from any single ardhaeo- logical site in the Samoan archipelago, and provides an imporant baseline not only for establishing the cultural chronology of the Manu'a Islands, but for comparisons with sequences from Tutuila and Westem Samoa. An initial radiocarbon and strati- graphic sequence for the Samoan archipelago was established by Green and Davidson (1974) on the basis of forty-five dates from a number of sites on 'Upolu and from two sites on Savai'i. The University of Utah Samnoan Archaeological Program (Jennings and Holmer 1980) obtained another thirty-four 14C dates from dteir excavations of sites on 'Upolu and -4 B-26465 * 4 B-35600 o-. B-26483 -4 B-35924 '-4 B-19742 i 4 B-25034 i- 4 B-25033 i B-35604 ;--- i B-35603 i I B-26464 i- B-35602 I--4 B-35*01 '- B-25673 '4 |B-25035 I I I 3 2 1 KYR CAL B.P. Figwre 6.1 Plot of radiocarbon age determinations from the To'aga site (cal B.P. at 1 sigma). Manono Islet. These seventy-nine dates from Westem Samoa provide a firm basis for an 'abso- lute' chronology. In American Samoa, the situation is less well developed. Frost (1978) obtained five radiocarbon dates from tee sites (AS-21-1, AS-21-2, AS-25-1) excavated by her. One sample, from Tulauta Village (AS-21-1), yielded a very early date of 810 ? 140 B.C. along with a significandy later date of A.D. 1320 ? 70. The other sites all dated to the last thousand years. Best et al. (1989) reported seven radiocarbon dates from the Tataga-matau adz quarry on Tutuila, indicating a sequence of use spaning the past 1000 years. Clark (in press) reviews an additional fifteen radiocarbon dates, largely from Aoa and Leone valleys on Tutuila Island. These include a date of 3389-2749 cal B.P. for initial occupation of the Aoa Valley, as well as very late dates for the use of ceramnics. The late dates associated with ceramics have caused Clark (in press) to argue that "pottery was not abandoned uniformly and wholesale trughout the archipelago." Dating Initial Hwnan Colonization On archaeological (ceramic) critera, t oldest known site in te Samoan archipelago is the sub- merged Early Eastem Lapita occupation at Mulifanu'a, Upolu (Green and Davidson 1974; Leach and Green 1989). A sample of marne shell 90 The To'aga Site w w __) 4 Radiocabon Chronology 91 (NZ-1958) from this site was originally reportd as having an age of 2980 ? 80 B.P. (Green and Richards 1975:317). More recently, Leach and Green have indicated that the correct convendonal 14C age for this sample should be 3251 ? 155 B.P. (1989:319). They suggest that the "most probable age" for this sample is 3116 cal B.P. (1 166 cal B.C.), after taking into accou te marine resevoir effect (Using the 100 ? 24 A-R value applied to the To'aga samples, the Mulifanua date would be calibrated to 1280-800 B.C.) This is consistent with other dates on Early Eastem Lapita assemblages from Fiji, Tongatapu, and Niuatoputapu (Kirch 1988, table 48), suggesting initial Lapita colonization of the Fiji-Westem Polynesian region by about cal 1200 B.C. (see also Kirch and Hunt 1988a, b). The earliest two 14C detemninations from the To'aga site, from the Layer V beach deposit in the 1987 main trench excavation, range between about 3700-3300 B.P. Although small numbers of thin, fine-ware sherds are present in this beach deposit, we must be cautious in interpreting these dates, because the deposidonal context is not specifically cultural. As argued in chapter 5, it is likely that the sherds derive from an in situ occupation situated on a beach ridge somewhat inland of the Layer V beach, and now buried under several meters of talus and colluvium. That the shell samples used for dating showed no signs of water-rolling or prolonged weathering is noteworthy, indicating that they were deposited in Layer V soon after death. It is conceiv- able, however, that the sherds became incorporated into the beach deposit somewhat later in time. Although these Layer V dates are not inconsistent with the earliest radiocarbon ages for Lapita col- onization in the region, they do fall at the early end. Sample Beta-35601, from a deep test unit directly against the talus slope, is unquestionably in primary, cultural context and is associated with thin, fine-tempered ware. It dates to 3257-2870 cal B.P. (1308-930 cal B.C.). Thus, we can be certain that Ofu Island, and the To'aga site, were settled by the end of the second millennium B.C. as part of te proess of discovery and colonization of the Fiji- Westem Polynesian region by Lapita populations (Kirch and Green 1987). Calibrating the Morphodynamic Model A second objective of some importance in our work has been to calibrate th sequence of shoreline change and progradation esponsible for the creation of the To'aga coastal ten-ace. Based on the model developed in chapter 4 (Kirch), it was suggested that the coastal tenace would not have begun to prograde significantly until sometime after about 2000 cal B.P., when te Holocene sea-level maximum began to drop to its modem level. This hypothesis is substantially confinned by the series of 14C dates from To'aga which reveals at the coastal terrace was very narrow and confined to a zone adjacent to the high cliffs until sometime early in the second millennium A.D. Sample Beta-26463, at 1389-1287 cal B.P., comes from a straigraphic context ta post- dates fte onset of significant prograation of fte coastal terrace. The Samoan Ceramic Sequence Fmally, the To'aga radiocarbon suite provides a finn chronology for the changes in the ceramic sequence revealed by our excavations and analyses (see chapter 5). Although classic dentate-stamped Early Eastem Lapita pottery was not recovered from To'aga, the red- and orange-slipped thin, fine- tempered ware recovered from the deepest units would appear to be contemporary with the Mulifanu'a Lapita assemblage from rUpolu. The manufacture and use of thin, fine-tempered ware, some decorated with rim notching, spwanned the period from about 2800-2400 cal B.P. This was followed by a phase in which the manufacture and use of thicker, coarse-tenpered pottery became dominant between about 2400-1500 cal B.P. The use of pottery appears to have been discontinued entirely by about 1500 cal B.P., thus matching closely the chronology for 'Upolu derived by Green and Davidson (1974). Furtr details of this ceramic chronology are presented in chapter 9 by Hunt and Erkelens. REFERENCES CITED Athens, S. 1986. Archaeological investigations at the Tarague Beach site, Guam. Report prepared for the Department of the Air Force. San Fran- cisco, Califomia. Best, S., H. Leach, and D. Witter 1989. Report on the second phase of fieldwork at the Tataga- matau site, American Samoa, July-August 92 The To'aga Site 1988. Department of Anthropology, University of Otago, New Zealand. Clarlc, J. T. 1989. The Eastem Tutila archaeological project: 1988 final report Report prepared for the Govenmnent of American Samoa. in press. Radiocarbon dates from American Samoa. Radiocarbon. Frst, J. 1978. Archaeological investigations on Tutuila Island, American Samoa. Unpublished Ph.D. dissetation, University of Oregon, Eugene. Gillespie, R., and H. A. Polach 1979. The suitability of marine shells for radiocarbon dating of Australian prehistory. IN R. Berger and H. Suess, eds., Proceedings of the Ninth Interna- tional Coniference on Radiocarbon Dating, pp. 404-421. Berkeley: University of Califomia Press. Gillespie, R., and P. Swadling 1979. Marine shells give reliable ages for middens. Search 10:92- 93. Green, R. C., and J. M. Davidson 1974. A radiocar- bon and stratigraphic sequence for Samoa. IN R. C. Green and J. M. Davidson, eds., Archae- ology in Western Samoa, Vol. 11, pp. 212-24. Auckland stitute and Museum Bulletin 7. Hunt, T. L., and P. V. Kirch 1987. Radiocarbon dates from two coastal sites in the Manu'a Group, American Samoa. Radiocarbon 29:417- 19. Jansen, H. S. 1984. Inshtie of Nuclear Sciences INS-R-328: Radiocarbon Datng for Contribu- tors. Lower Hut, New Zealand: New Zealand Institute of Nuclear Sciences. Jennings, J. D., and R. N. Holmer 1980. Archaeo- logical Excavations in Western Samoa. Pacific Antrpological Records 32. Honolulu: Bemice P. Bishop Museum. Kirch, P. V. 1988. Niuatoputapu: The Prehistory of a Polynesian Chiefdom. Thomas Burke Memorial Washington State Museum Mono- graph No. 5. Seattle. Kirch, P. V., and R. C. Green 1987. History, phylog- eny, and evolution in Polynesia. Current Anthropology 28:431-56. Kirch, P. V., and T. L. Hunt 1988a. Radiocarbon dates from the Mussau Islands and the Lapita colonization of the southwestem Pacific. Radiocarbon 30:161-69. 1988b. The spatial and temporal boundaries of Lapita. IN P. V. Kirch and T. L. Hun, eds., Archaeology of the Lapita Cultural Complex: A Critical Review, pp. 9-32. Burke Memorial Museum Research Report No. 5. Seattle. Kirch, P. V., T. L. Hunt, and J. Tyler 1989. A radiocarbon sequence from the To'aga site, Ofu Island, American Samoa. Radiocarbon 31:7- 13. Leach, H., and R. C. Green 1989. New information for the Ferry Berth Site, Mulifanua, Westem Samoa. Journal of the Polynesian Society 98:319-29. Poulsen, J. 1987. Early Tongan Prehistory: The Lapita Period on Tongatapu and its Relon- ships. 2 vols. Terra Australis 12. Deparunent of Prehistory, Australian National University. Canberra. Stuiver, M., and B. Becker 1986. High-precision decadal calibraion of the radiocarbon time scale, A.D. 1950-2500 B.C., Radiocarbon 28(2B):863-910. Stuiver, M., G. W. Pearson, and T. Braziunas 1986. Radiocarbon age calibration of marine samples back to 9000 cal B.P. Radiocarbon 28(2B):980- 1021. Stuiver, M., and H. A. Polach 1977. Discussion: Reporting of 14C data. Radiocarbon 19:355-33. Stuiver, M., and P. J. Reimer 1986. A computer program for radiocarbon age calibration. Radiocarbon 28(2B): 1022-1030.