203

EFFECTS OF

C SAMPLE SELECTION IN ARCHAEOLOGY: AN EXAMPLE

FROM HAWAI‘I

Tom Dye

International Archaeological Research Institute, Inc., 2081 Young St., Honolulu, Hawai‘i 96826, USA. Email: tdye@lava.net.

ABSTRACT. A comparison of radiocarbon histograms of samples controlled and uncontrolled for sample provenance and

composition factors indicates that differences are sufficiently large to influence the discrimination of alternative hypotheses

on the age of the important, early Bellows Dune site in Hawai‘i.

INTRODUCTION

The importance of radiocarbon sample collection in archaeology to control sample provenance and

composition factors (Waterbolk 1971) is widely appreciated in the archaeological community (Tay-

lor 1987:105 ff). In the Pacific, McFadgen et al. (1994) have shown that the probable effects of

in-built age are sufficiently great to support “fallacious inferences” about the chronology of New

Zealand prehistory. They compared

C histograms of 338 marine shell and 425 wood and charcoal

samples, and of 101 short-lived and 131 long-lived woods. In both comparisons, the sample type

with greatest potential for in-built age yielded a histogram with a relatively long, heavy tail extend-

ing back in time and an earlier peak. In Hawai‘i, where the shallow time depth of prehistory,

between 1000 (Spriggs and Anderson 1993) and 2000 yr (Hunt and Holsen 1991), is comparable to

New Zealand, archaeologists have often failed to control sample provenance and composition fac-

tors in

C sample selection (Davidson 1992). This paper identifies effects of sample provenance

and composition in a suite of

C dates from two sites adjacent to the Bellows Dune Site, with impli-

cations for tests of hypotheses on the age of this important traditional Hawaiian archaeological site.

The Bellows Dune Site is often cited as an example of an early Hawaiian settlement (Kirch 1998,

1974, 1985; Shun 1992b; Miller 1991). The

C chronology of the site is a matter of dispute among

prehistorians over the timing of initial Polynesian settlement of the Hawaiian Islands, with argu-

ments for a “long” chronology with settlement prior to AD 600 (Kirch 1974, 1985; Hunt and Holsen

1991) countered by arguments from the same data for a “short” chronology with settlement around

AD 900 (Spriggs and Anderson 1993; Tuggle 1997). An attempt in 1987 to clarify the site chronol-

ogy by dating archived wood charcoal samples, the results of which are partially reported (Spriggs

and Anderson 1993; Hunt and Holsen 1991), yielded ambiguous results. The site has been mostly

destroyed since it was last excavated in 1975 and there is some question whether archaeological

contexts from the site’s earliest levels are preserved. Given this state of affairs,

C dates from

archaeological contexts immediately inland of the Bellows Dune site provide potentially important

information with which to distinguish between the “long” and “short” chronologies. In particular,

contextually secure

C dates that yield a calibrated age older than AD 600 would counter-indicate

the short chronology and provide strong support for the long chronology.

The Bellows Dune Site, designated Site 50-80-15-4852 in the State of Hawaii site numbering system,

is located north of the mouth of Stream on the seaward margin of the plain (Figure

1), a series of calcareous sand beach ridges and swales that formed as local sea level fell from a

mid-Holocene high stand of about +1.8 m (Fletcher and Jones 1996) to its present level by about 2000

yr ago (Athens and Ward 1993; Fletcher and Jones 1996). Two archaeological sites, designated Sites

50-80-15-4851 and -4853, are located inland of the Bellows Dune Site; Site 4853 is immediately

adjacent and Site 4851 is on the south side of Stream. Both sites consist of extensive, discontig-

uous, and somewhat variable subsurface deposits. The basic stratigraphic sequence at both sites can

be described as a series of three horizons. The surface Horizon I consists primarily of secondarily

204

T Dye

deposited sediments including traditional Hawaiian cultural materials, WWII-era construction fill,

and recently deposited aeolian sands. The composition, thickness, and appearance of this horizon

vary considerably over the sites. Horizon II includes intact traditional Hawaiian land surface and cul-

tural deposits. This horizon is absent in many places, and in others only the lower portion is intact;

both situations appear to be the result of sediment removal by heavy equipment. Horizon II lacks

internal stratification and generally does not preserve cultural features, which usually can be distin-

guished only when they cut into Horizon III, the basal Holocene deposit of white calcareous sand.

There are at least three sources of wood with high in-built age at or near Sites 4851 and 4853. The

heartwood of old trees growing on the plain might, at one time, have been the most ubiquitous

source. There is little information on the longevity of most native taxa, but it can be assumed that

many trees live on the order of 100 yr. Waterlogged wood is commonly found buried in and near the

banks of Stream (Athens 1988; Rolett 1992). Two pieces of wood have been dated, yielding

C dates of 770 ± 80 BP (I-16734) and 450 ± 70 BP (Beta-18237). Driftwood from the Pacific

Northwest (Strong and Skolmen 1963) that washes ashore at Beach (Hall 1839) is a

final source.

C dates on Pacific Northwest driftwood collected on the south shore of Hawai‘i

Island (Emory and Sinoto 1969) indicate the magnitude of in-built age from this source. A piece of

Douglas fir (WSU 424) returned a date of 470 ± 160 BP, and a piece of Western red cedar (WSU

428) returned a date of 1200 ± 325 BP.

Archaeological investigations at Sites 4851 and 4853 since 1986 have produced a corpus of 44

dates, nearly all of which are reported in unpublished proprietary or manuscript documents of lim-

ited circulation. Before 1997, 22 samples of wood charcoal were dated, most of which derived from

the general Horizon II matrix or alluvial deposits along the banks of stream and were not

securely associated with a particular cultural event. Samples taken from the Horizon II matrix con-

tain materials of a potentially wide range of ages, not all of which are related directly to human

activity. Sample material was identified in only one case (Beta-89698), and the material chosen for

dating was not a short-lived taxon. Thus, all of the samples dated before 1997 did not control for the

effects of in-built age. Since 1997, 22 samples of identified wood charcoal, securely associated with

burning events in small scoop hearths and earth ovens or with infilling of refuse pits, have been

dated. The identified and dated charcoals represent twigs, short-lived taxa, and nutshells selected to

minimize in-built age.

A comparison of

C histograms for the 22 samples selected to control for the effects of sample prov-

enance and composition and the 22 uncontrolled samples (Figure 2) reveals a pattern similar to the

one identified by McFadgen et al. (1994). The

C histogram for controlled samples starts at AD 1150

and peaks at AD 1600. The shape of the

C histogram for uncontrolled samples has a long tail that

starts at AD 160. The histogram peaks about AD 1490, 100 yr earlier than the histogram for controlled

samples. The long tail on the histogram of uncontrolled samples represents three

C dates, the two

oldest of which, Beta-25783 and Beta-31518, were collected from the banks of Stream. Their 2

calibrated age ranges (Stuiver and Pearson 1986) of AD 350–650 and AD 440–960, respectively, are

sufficiently old potentially to disconfirm the short chronology. However, neither date is securely asso-

ciated with a cultural event and both dates possibly incorporate materials with a high in-built age,

such as are present today in the banks of the stream. The third

C date in the tail, Beta-30891, cali-

brates to a 2

range of AD 770–1170 and would not disconfirm the short chronology.

In conclusion, suites of

C dates from two sites adjacent to the Bellows Dune site show the effects

of sample selection factors. These effects, which are likely due to in-built age and to the use for dat-

Effects of

C Sample Selection

205

ing of materials not directly associated with human activity, are sufficiently large to influence the

discrimination of alternative hypotheses on the age of the Bellows Dune site. The suite of controlled

C dates contains no evidence for traditional Hawaiian use of sites immediately inland of the Bel-

lows Dune site before AD 1150. The short chronology for the Bellows Dune site is not contradicted

by the dating evidence from adjacent sites.

Controlled Samples

Twenty-two samples of identified materials from Sites 50-80-15-4851 and -4853, selected to mini-

mize the effects of in-built age. Wood species identification by Gail M Murakami, International

Archaeological Research Institute, Inc. Wood Identification Laboratory. All samples associated with

Figure 1 Locations of sites 50-80-15-4851, -4852 (Bellows Dune), and -4853. Site locales based on

Tuggle (1997).

206

T Dye

a traditional Hawaiian cultural event from either the prehistoric period (pre AD 1778) or early his-

toric period prior to AD 1850 (Desilets and Dye 1998:97 ff). Unless otherwise noted, all samples

stratigraphically associated with Horizon II.

Beta-101868. 220 ± 40

C = -24.2‰

Sample identified as cf.

Sida fallax

, 0.09 g, Site 50-80-15-4853, Locale 5 (Figure 1), Feature 3, Unit

BT-1. Collected and submitted January 1997. Feature 3 was exposed in a backhoe trench, where it

was “V” shaped in profile, 2 m wide and 0.9 m deep. Feature fill appeared “banded” (Addison 1997:

25).

Comment

: Feature 3 interpreted by the excavator as a “refuse pit” (Addison 1997:25). Wood char-

coals identified from Feature 3 include 6 native and Polynesian-introduced taxa and

Pinus

sp., a

genus alien to the Hawaiian Islands. Feature 3 was subsequently excavated by Desilets and Dye

(1998). This excavation showed the feature to be a depression or trench >8 m long, subsequently

filled in. Food remains and combustion related artifacts indicate that cooking was performed at or

near the feature. Identification of

Pinus

sp. was interpreted as evidence that the feature “was in the

process of formation during the early historic period” (Desilets and Dye 1998:154–5).

Figure 2

C histograms of 22 controlled and 22 uncontrolled samples at Sites 50-80-15-4851 and

-4853. Samples calibrated with the s

= 100 smoothed calibration curve described by Törnqvist and

Bierkens (1994). Nonparametric 1s confidence intervals generated with an approximate bootstrap

technique (Efron and Tibshirani 1993:188). The y-axis scale, called “intensity” by Stolk et al. (1994),

is omitted as unimportant in the comparison of histogram shapes.

Effects of

C Sample Selection

207

Beta-101869. 30 ± 60

C = -12.9‰

Chamaesyce

sp., 0.07 g, Site 50-80-15-4853, Locale 5 (Figure 1), Feature 6, Unit BT-5. Collected

and submitted January 1997. Feature 6 is a “shallow depression with two distinct dark lenses” (Add-

ison 1997:29).

Comment

: Identification of wood charcoals yielded 6 native and Polynesian-introduced taxa; no his-

torically introduced or alien taxa were identified. The lensing in this feature is unusual for traditional

Hawaiian pit features in the region.

Beta-101870. 170 ± 60

C = -22.5‰

Chenopodium oahuense

, 0.19 g, Site 50-80-15-4853, Locale 5 (Figure 1), Unit TP-1, Layer II. Col-

lected and submitted January 1997. Sample taken from cultural layer near Feature 1, “a pit shaped

depression with fill slightly darker” (Addison 1997:49) than the layer from which it was cut.

Comment

: The excavator interpreted Feature 1 as an earth oven based on its size and associated

fire-altered rock. Wood charcoal identification yielded 6 native and Polynesian-introduced taxa; no

historically introduced or alien taxa were identified.

Beta-101871. 720 ± 40

C = -25.3‰

Sample identified as cf.

Osteomeles anthyllidifolia

, 0.65 g, Site 50-80-15-4853, Locale 5 (Figure 1),

Feature 9, Unit BT-23. Collected and submitted January 1997. Feature 9 is a depression with

“heavily charcoal-stained fill” (Addison 1997:45).

Comment

: Identification of wood charcoal yielded 3 native and Polynesian-introduced taxa; no his-

torically introduced or alien taxa were identified. Feature 9 is buried by secondarily deposited mate-

rial.

Beta-101872. 680 ± 40

C = -24.7‰

Sample identified as cf.

Osteomeles anthyllidifolia

, 0.26 g, Site 50-80-15-4853, Locale 5 (Figure 1),

Feature 10, Unit BT-23. Collected and submitted January 1997. Feature 10 is a depression with

“heavily charcoal-stained fill” (Addison 1997:45).

Comment

: Identification of wood charcoal yielded 4 native taxa; no Polynesian or historic introduc-

tions were identified. Feature 10 is buried by secondarily deposited material.

Beta-101873. 320 ± 40

C = -21.8‰

Dodonea viscosa

, 0.45 g, Site 50-80-15-4853, Locale 5 (Figure 1), Feature 16, BT-13. Collected and

submitted January 1997. “Feature 16 is a shallow pit feature with more charcoal than the surrounding

sediments” (Addison 1997:40).

Comment

: The dated specimen, a native species, was the only charcoal identified from the feature.

Beta-111022. 150 ± 40

C = -27.5‰

Sida

cf.

fallax

, 0.5 g, Site 50-80-15-4853, Locale 5 (Figure 1), Feature 1, Trench 5, Layer IIC/2. Col-

lected 1997 and submitted October 1997. The Feature 1 pit, greater than 55 cm in diameter, was

filled with sand stained dark by charcoal and with numerous coarse-grained and vesicular angular to

208

T Dye

rounded pebbles and gravel, many pieces of which show evidence of alteration by heat (Dye 1998:

55 ff).

Comment

: Feature 1 is inferred to be an earth oven based on its morphology and the nature of the fill.

13 native and Polynesian-introduced taxa were identified in a charcoal assemblage collected from

the base of the feature; no historically introduced taxa were identified.

Beta-111023. 310 ± 40

C = -26.9‰

Sample identified as cf.

Rauvolfia sandwicensis

, 0.34 g, Site 50-80-15-4851, Locale 14 (Figure 1),

Feature 3, Trench 4. Collected 1997 and submitted October 1997. Feature 3 is a shallow pit cut into

the basal Holocene sand from the Horizon II cultural layer (Dye 1998:48 ff).

Comment

: The feature is interpreted as a scoop fire pit. Identification of wood charcoal yielded 5

native and Polynesian-introduced taxa; no historic introductions were identified.

Beta-111024. 140 ± 60

C = -26.8‰

Sida

cf.

fallax

, 0.36 g, Site 50-80-15-4851, Locale 14 (Figure 1), Feature 2, Trench 4. Collected 1997

and submitted October 1997. Feature 2 is a shallow fire pit cut into the basal Holocene sand (Dye

1998:46 ff).

Comment

: Feature 2 is interpreted as a fire pit. Identification of wood charcoal yielded 9 native and

Polynesian introduced taxa; no historically introduced or alien taxa were identified.

Beta-111025. 540 ± 50

C = -24.2‰

Sida

cf.

fallax

, 0.31 g, Site 50-80-15-4851, Locale 14 (Figure 1), Feature 1, Trench 4, Excavation

Unit 3, Layer III/2. Collected 1997 and submitted October 1997. Sample collected from a shallow

pit feature sealed stratigraphically by debitage from a lithic reduction event (Dye 1998:46).

Comment

: Identification of wood charcoal yielded 11 native and Polynesian-introduced taxa; no his-

torically introduced taxa were identified. Sample provides a terminus post quem for the lithic depo-

sition event.

Beta-120317. 140 ± 50

C = -21.3‰

Sida

cf.

fallax

, 0.25 g, Site 50-80-15-4853, Locale 5 (Figure 1), Feature 1, Unit IIA2, Layer II. Col-

lected 1998 and submitted July 1998. Bowl-shaped Feature 1 measures roughly 90 cm in diameter

and is 20 cm at maximum thickness (Desilets and Dye 1998:111 ff). Feature matrix consists of very

dark gray unconsolidated calcareous sand with small amounts of bone and marine invertebrates.

Comment

: Identification of wood charcoal yielded 4 native taxa and

Pinus

sp., a genus alien to the

Hawaiian Islands.

Pinus

sp. is believed by the excavator to have entered the site as lumber, rather

than driftwood, an interpretation consistent with the age of the sample.

Beta-120318. 150 ± 50

C = -26.1‰

Sida

cf.

fallax

, 0.09 g, Site 50-80-15-4853, Locale 5 (Figure 1), Feature 5, Unit IIIB6, Layer II. Col-

lected 1998 and submitted July 1998. Bowl shaped Feature 5 is circular in plan, measures 40 cm in

diameter, and is 12 cm thick at its center (Desilets and Dye 1998:161 ff). It contains small amounts

of bone and marine invertebrate as well as some thermally altered rock, coral, and charcoal.

Effects of

C Sample Selection

209

Comment

: Identification of wood charcoal yielded 8 native and Polynesian-introduced taxa and

Pinus

sp., a genus alien to the Hawaiian Islands.

Pinus

sp. is believed by the excavator to have

entered the site as lumber, rather than driftwood, an interpretation consistent with the sample’s age.

Beta-120319. 350 ± 80

C = -25.9‰

Aleurites moluccana

nutshell,

Chenopodium oahuense

Sida

cf.

fallax

, and

Osteomeles anthyllidifo-

lia

, 3.04 g, Site 50-80-15-4853, Locale 5 (Figure 1), Feature 9, Unit IIIA8, Layer II. Collected 1998

and submitted July 1998. Feature 9, round in plan and bowl-shaped in cross-section, measures 44 cm

in diameter, with a maximum thickness of 21 cm (Desilets and Dye 1998:165 ff). It contains large

amounts of coral, charcoal and thermally altered rock, with scant vertebrate and marine invertebrate

remains.

Comment

: A concentration of thermally altered basalt cobbles is located immediately north of the

feature. These cobbles are likely “oven stones” used in Feature 9, which is interpreted to be an earth

oven. Identified wood charcoal includes 12 native and Polynesian-introduced taxa; no historically

introduced or alien taxa were identified.

Beta-120320. 230 ± 50

C = -25.6‰

Aleurites moluccana

nutshell, 20.63 g, Site 50-80-15-4853, Locale 5 (Figure 1), Feature 13, Unit

IVA8, Layer II. Collected 1998 and submitted July 1998. Feature 13 is round in plan, 65 cm diame-

ter, and bowl-shaped in cross-section with a 19 cm maximum thickness (Desilets and Dye 1998:133

ff). Contents include bone, charcoal, marine invertebrate, thermally altered rock, 1 basalt flake, and

1 bone tool blank.

Comment

: Feature 13 is interpreted as an open fire used for cooking or otherwise preparing a variety

of marine flesh foods and possibly pig. Identified wood charcoal includes 14 native and Polyne-

sian-introduced taxa; no historically introduced or alien taxa were identified.

Beta-120321. 110 ± 70

C = -25.0‰

Aleurites moluccana

nutshell, 5.84 g, Site 50-80-15-4853, Locale 5 (Figure 1), Feature 15, Unit

IVB8, Layer II. Collected 1998 and submitted July 1998. Feature 15 is round in plan with a 50 cm

diameter and bowl-shaped with a maximum thickness of 13 cm (Desilets and Dye 1998:136 ff).

Materials recovered from Feature 15 include bone, charcoal, coral, unburned

Aleurites moluccana

nutshell, marine invertebrates, and thermally altered rock.

Comment

: Wood charcoal identification yielded 11 native and Polynesian-introduced taxa; no his-

torically introduced or alien taxa were identified. The relatively small amount of cultural material

yielded by Feature 15 makes it difficult to infer a function for the feature.

Beta-120322. 310 ± 60

C = -16.8‰

Chamaesyce

spp., 7.85 g, Site 50-80-15-4853, Locale 5 (Figure 1), Feature 16, Unit IVA, B8.7,

Layer II. Collected 1998 and submitted July 1998. Feature 16 is round in plan with an approximate

diameter of 65 cm and bowl shaped in cross-section with a maximum thickness of 12 cm (Desilets

and Dye 1998:138 ff). Feature 16 matrix is unconsolidated black calcareous sand and includes bone,

charcoal, coral, historic debris, marine invertebrate, and thermally altered rock.

Comment

: The feature is interpreted as a firepit used to cook fish and shellfish. Wood charcoal iden-

tification yielded 12 native and Polynesian-introduced taxa; no historically introduced or alien taxa

210

T Dye

were identified. The sample has a moderate probability of association with use of the feature, which

yielded some evidence of post-depositional disturbance.

Beta-120323. 170 ± 60

C = -27.5‰

Aleurites moluccana

nutshell,

Chamaesyce

spp.,

Sida

cf.

fallax

, 5.25 g, Site 50-80-15-4853, Locale

5 (Figure 1), Feature 17, Unit IVA7, Layer II. Collected 1998 and submitted July 1998. Feature 17,

only partially revealed by excavation, appears round in plan with an estimated 70 cm diameter, and

is bowl shaped in cross-section with a maximum thickness of 17 cm (Desilets and Dye 1998:140 ff).

Matrix is a very dark grayish brown unconsolidated calcareous sand that grades gradually to the

underlying basal sand. Material recovered from Feature 17 includes bone, charcoal, coral, a small

piece of metal, unburned

Aleurites moluccana

nutshell, marine invertebrate, thermally altered rock,

and one basalt flake.

Comment

: A small piece of metal is likely to be intrusive to the feature. The feature is interpreted as

a small earth oven. Wood charcoal identification yielded 9 native and Polynesian-introduced taxa;

no historically introduced or alien taxa were identified in the assemblage.

Beta-120324. 250 ± 50

C = -25.2‰

Aleurites moluccana

nutshell, 37.27 g, Site 50-80-15-4853, Locale 5 (Figure 1), Feature 18, Unit

IVC7, Layer II. Collected 1998 and submitted July 1998. Feature 18 is oblong in plan and measures

45 cm along its major axis and 25 cm along its minor axis (Desilets and Dye 1998:142 ff). It is

bowl-shaped in cross-section with a 6 cm maximum thickness. Feature 18 contains bone, charcoal,

coral, a large amount of unburned

Aleurites moluccana

nutshell, marine invertebrate, and thermally

altered rock.

Comment

: The feature is interpreted as an earth oven in which coral was the dominant “oven stone.”

Wood charcoal identification yielded 5 native and Polynesian-introduced taxa; no historically intro-

duced taxa were identified.

Beta-120325. 270 ± 70

C = -25.2‰

Aleurites moluccana

nutshell, 5.42 g, Site 50-80-15-4853, Locale 5 (Figure 1), Feature 19, Unit

IVC7, Layer II. Collected 1998 and submitted July 1998. Feature 19 is oblong in plan and trends to

the northeast along its 67 cm major axis (Desilets and Dye 1998:144 ff). In cross-section, the feature

is bowl-shaped with a maximum thickness of 15 cm. Matrix is a very dark gray unconsolidated cal-

careous sand with bone, charcoal, coral, marine invertebrate, thermally altered rock, one basalt flake,

and a waterworn pebble.

Comment

: The feature is interpreted as an open firepit. Wood charcoal identification yielded 12

native and Polynesian-introduced taxa; no historically introduced or alien taxa were identified.

Beta-120326. 330 ± 60

C = -14.0‰

Aleurites moluccana

nutshell,

Chamaesyce

spp.,

Sida

cf.

fallax

, 5.28 g, Site 50-80-15-4853, Locale

5 (Figure 1), Feature 20, Unit VA1, Layer II. Collected 1998 and submitted July 1998. Feature 20,

partially excavated, is round in plan with some irregularity in its southwest portion and with an esti-

mated diameter of 110 cm (Desilets and Dye 1998:122 ff). In cross-section, Feature 20 is irregularly

bowl-shaped with a maximum thickness of 26 cm. Feature matrix consists of dark grayish brown

(10YR 4/2) loosely consolidated calcareous sand and included large quantities of bone, charcoal,

Effects of

C Sample Selection

211

coral, lithic debitage,

Aleurites moluccana

nutshell, marine invertebrate, thermally altered rock, and

traditional artifacts.

Comment

: Feature 20 is interpreted as a “trash pit associated with... cooking” (Desilets and Dye

1998:126).

Beta-120327. 400 ± 70

C = -23.0‰

Aleurites moluccana

nutshell, 3.99 g, Site 50-80-15-4853, Locale 5 (Figure 1), Feature 24. Collected

1998 and submitted July 1998. Feature 24 was identified in the wall of a utility trench. It is bowl

shaped in cross-section with an approximate maximum thickness of 13 cm. Based on the visible pro-

file, the feature has a diameter in excess of 45 cm. Materials recovered from Feature 24 include char-

coal, a very small amount of waterworn and unidentified marine invertebrate, and a bit of thermally

altered rock.

Comment

: The feature is interpreted as a fire pit. The identified wood charcoal includes 7 native and

Polynesian-introduced taxa; no historically introduced or alien taxa were identified.

Beta-120328. 220 ± 50

C = -25.5‰

Sida

cf.

fallax

, 0.23 g, Site 50-80-15-4853, Locale 5 (Figure 1), Feature 25. Collected 1998 and sub-

mitted July 1998. Feature 25 was identified in the wall of a utility trench. It is bowl-shaped in

cross-section with an approximate thickness of 22 cm and diameter in excess of 60 cm. Materials

recovered from Feature 25 include thermally altered rock, bone fragments, and charcoal.

Comment

: The size, shape, and contents of Feature 25 suggest that it functioned as an earth oven.

The identified wood charcoal includes 7 native and Polynesian-introduced taxa; no historically

introduced or alien taxa were identified.

Uncontrolled Samples

Twenty-two wood charcoal samples on unidentified materials and/or unassociated with a cultural

event. The Layer designations in the descriptions follow the source documents, which reflect local

stratigraphic sequences. Samples are stratigraphically associated with Horizon II unless otherwise

noted.

Beta-18237. 450 ± 70

C = -27.7‰

Unidentified partially carbonized wood, 114.5 g, Site 50-80-15-4851, Locale 7 (Figure 1), Unit

ST-20, Layer III, 110-116 cm bs. Collected 1986 and submitted October 1986. Sample from a log

that was deposited on a light gray, very fine sand and covered with very dark gray clay, approxi-

mately 15 m from Stream (Athens 1988:65,97–8).

Comment

: The excavator interpreted the sample date as evidence for human use of the region. The

sample was not associated with a cultural event. The stratigraphy of Unit ST-20 reflects alluvial

deposition along Stream rather than the sequence of horizons described above. Date provides

an indication of the potential in-built age of waterlogged wood at Site 4851 (see also I-16,688).

Beta-19043A. 420 ± 60

C = -30.9‰

Unidentified wood charcoal, 23.5 g, Site 50-80-15-4851, Locale 7 (Figure 1), Unit TP-4, Layer III.

Collected 1986 and submitted December, 1986. Dispersed charcoal taken from the periphery of a

212

T Dye

pickleweed marsh adjacent to Stream (Athens 1988:58–63,65). Layer III contained preserved

wood and cultural material in a waterlogged environment.

Comment

: The excavator interpreted the sample as associated with natural “infilling of what was

formerly a marine embayment or estuary” (Athens 1988:62). It was not associated with a cultural

feature. Layer III reflects alluvial deposition along Stream rather than the sequence of horizons

described above.

Note

: The dating laboratory assigned Beta-19043 to 2 samples (see Beta-24362).

Beta-24362. 250 ± 50

C = -22.8‰

Unidentified wood charcoal, 18.8 g, Site 50-80-15-4851, Locale 3 (Figure 1), Unit TP-2, Layer IV/

2. Collected 1986 and submitted December 1987. Sample from dispersed charcoal within the cul-

tural layer.

Comment

: Not associated with a cultural event. Charcoal from the layer (not dated) was identified

and included mango, an early 19th century introduction to the Hawaiian Islands. A second sample

(Beta-19043B, see Beta-19043A) collected from “a probable hearth feature with abundant charcoal”

(Athens 1988:54) near the provenience of Beta-24362 was more active than the modern standard

(Athens 1988:56).

Beta-25783. 1540 ± 80

C = -12.8‰

Unidentified wood charcoal, Site 50-80-15-4853, Locale 2 (Figure 1). Collected 22 Feb 1988.

Comment

: Sample taken from “a probable prehistoric oven” (Streck and Watanabe 1988:5) exposed

in the bank of Stream. There are few contextual details for this early date.

Beta-25784. 360 ± 80

C = -18.2‰

Unidentified wood charcoal, Site 50-80-15-4853, Locale 2 (Figure 1). Collected 22 Feb 1988.

Comment

: Sample collected from “a postmold feature” exposed in the bank of Stream (Streck

and Watanabe 1988:5). Few contextual details.

Beta-26702. 170 ± 80

C = -27.3‰

Bulk sample of

Diospyros

sp.,

Metrosideros

sp.,

Aleurites moluccana

nutshell, and unidentified

wood charcoal, 6.7 g, Site 50-80-15-4851, Locale 3 (Figure 1), Feature 3, TP-5, Layer IV. Collected

1988 and submitted July 1988. Feature 3 described as “a circular basin shaped feature filled with

fire-cracked basalt cobbles, charcoal, and midden remains. The feature was 40 cm in diameter and

20 cm deep” (McNeill 1989:28). The feature also contained small flakes of volcanic glass.

Comment

: Identified wood charcoal contains native and Polynesian-introduced taxa; no historically

introduced or alien taxa were identified. Feature 3 is a rare example of a cultural feature preserved

within Horizon II.

Beta-30888. 240 ± 80

C = -23.3‰

Unidentified wood charcoal, 12 g, Site 50-80-15-4851, Locale 11 (Figure 1), Trench 15 B, Layer III.

Collected 1989 and submitted April 1989. Layer III described as “buried A-horizon containing plen-

tiful midden and artifacts” (Hammatt and Shideler 1989:32).

Comment

: Sample not securely associated with a cultural event.

Effects of

C Sample Selection

213

Beta-30889. 440 ± 60

C = -26.0‰

Unidentified wood charcoal, 3 g, Site 50-80-15-4851, Locale 11 (Figure 1), Trench 15 B, Layer III.

Collected 1989 and submitted April 1989. Layer III described as “buried A-horizon containing plen-

tiful midden and artifacts” (Hammatt and Shideler 1989:32).

Comment

: Sample described as coming from an undisturbed cultural deposit (Hammatt and Shideler

1989:58). Sample not securely associated with a cultural event.

Beta-30890. 730 ± 120

C = -24.3‰

Unidentified wood charcoal, 8 g, Site 50-80-15-4851, Locale 10 (Figure 1), Trench 18B, Layer III.

Collected 1989 and submitted April 1989. Layer III described as a “gray sandy loam with well sorted

very fine sand and silt containing historic artifacts in upper portion. Contains fire-cracked rocks,

basalt flakes, charcoal and midden” (Hammatt and Shideler 1989:33).

Comment

: The excavator interpreted Layer III as an undisturbed cultural deposit (Hammatt and

Shideler 1989:58). Sample depth (135–145 cm) is below the reported lower boundary of Layer III

(115 cm) (Hammatt and Shideler 1989:33), but a profile of Trench 18B (Hammatt and Shideler

1989:38) shows a pit feature at the northern end that extends from the base of the cultural layer into

the sterile Holocene sand layer to a depth consonant with the sample depth. The apparent pit feature

is not described. Sample not securely associated with a cultural event.

Beta-30891. 1060 ± 90

C = -24.5‰

Unidentified wood charcoal, 5 g, Site 50-80-15-4851, Locale 10 (Figure 1), Trench 18C, Layer III.

Collected 1989 and submitted April 1989. Layer III described as a “gray well sorted fine to very fine

sand, top 10 cm compacted massive structure. Contains basalt flakes, shell, volcanic glass, charcoal,

midden” (Hammatt and Shideler 1989:34).

Comment

: Sample described as coming from an undisturbed cultural deposit (Hammatt and Shideler

1989:58). Sample not securely associated with a cultural event.

Beta-30892. 750 ± 90

C = -23.2‰

Unidentified wood charcoal, 5 g, Site 50-80-15-4851, Locale 10 (Figure 1), Trench 18C, Layer III.

Collected 1989 and submitted April 1989. Layer III described as a “gray well sorted fine to very fine

sand, top 10 cm compacted massive structure. Contains basalt flakes, shell, volcanic glass, charcoal,

midden” (Hammatt and Shideler 1989:34).

Comment

: Sample described as coming from an undisturbed cultural deposit (Hammatt and Shideler

1989:58). Sample not securely associated with a cultural event.

Beta-30893. 640 ± 90

C = -25.9‰

Unidentified wood charcoal, 5 g, Site 50-80-15-4851, Locale 10 (Figure 1), Trench 18E, Layer III.

Collected 1989 and submitted April 1989. Layer III is described as a white sand (Hammatt and

Shideler 1989:35).

Comment

: Layer description appears to be incorrect. Sample described as coming from an undis-

turbed cultural deposit (Hammatt and Shideler 1989:58). Sample not securely associated with a cul-

tural event.

214

T Dye

Beta-31518. 1340 ± 130

C = -25.4‰

Unidentified wood charcoal, Site 50-80-15-4853, Locale 2 (Figure 1), NW Quad., Layer VI. Col-

lected 1989. Sample collected from a charcoal concentration exposed in bank of Stream.

Comment

: Layer VI interpreted by the excavator as a secondary “alluvial river-bank deposit” (Shun

1993:5). Sample not stratigraphically associated with Horizon II. Sample interpreted by the excava-

tor as dating cultural activity upstream from the collection location, but is not securely associated

with a cultural event.

Beta-31519. 380 ± 60

C = -22.1‰

Unidentified wood charcoal, Site 50-80-15-4853, Locale 2 (Figure 1), Unit TU-1, Layer IV. Col-

lected 1989. Layer IV described as a light yellowish brown very fine sand with traditional Hawaiian

cultural material.

Comment

: Sample “retrieved from a charcoal lens” (Shun 1993:47) located just above the water

table. Also present in Layer IV were copper wire and small pieces of rubber, which the excavator

believed “could easily have fallen in from the upper strata” (Shun 1993:34). Sample not securely

associated with a cultural event.

Beta-32226. 10 ± 40

C = -21.5‰

Unidentified wood charcoal, Site 50-80-15-4851, Locale 3 (Figure 1), Layer III, dark gray sediment

above Feature 5. Collected in late 1988 or early 1989 and submitted 1989.

Comment

: The excavator interpreted this sample, along with Beta-32227, as dating “the terminal

points of occupation” (Shun 1992b:37) associated with Horizon II, and supporting the idea that

occupation of the site extended to the late prehistoric or early historic period (post AD 1778). Sam-

ple not securely associated with a cultural event.

Beta-32227. 690 ± 110

C = -26.7‰

Unidentified wood charcoal, Site 50-80-15-4851, Locale 3 (Figure 1), Fire-hearth feature 1A, Layer

III base. Collected in late 1988 or early 1989 and submitted 1989.

Comment

: Interpreted by the excavator as establishing the date of initial site occupation around AD

1300, and with Beta-32226 as indicating an occupational duration of “at least 500 years” (Shun

1992b:37 ff).

Beta-36065. 440 ± 50

C = -20.6‰

Unidentified wood charcoal, Site 50-80-15-4851, Locale 11 (Figure 1), Unit TU 18, Layer III/3. Col-

lected 1989. Sample collected near a feature interpreted by the excavator as a “probable fire hearth”

in a portion of Layer III “with pockets of charcoal concentrations and fire-cracked rocks” (Shun

1992a:40).

Comment

: Although the sample was not securely associated with a cultural event the excavator

interpreted it as “material put aside from fire-hearth cleaning” (Shun 1992a:46).

Effects of

C Sample Selection

215

Beta-89698. 450 ± 60

C = -27.5‰

Metrosideros

sp., 0.53 g, Site 50-80-15-4851, Locale 11 (Figure 1), Unit TP-1, Layer IIIc/8. Col-

lected 1995 and submitted January 1996. Collected from general matrix at base of cultural deposit.

Comment

: Identified wood charcoals from the cultural deposit (not dated) include 12 endemic,

indigenous, and Polynesian introduced taxa, none of which are historically introduced or alien to

Hawai‘i. The excavator interpreted the date as evidence for “pre-contact Hawaiian habitation”

(Erkelens 1996:19). The genus

Metrosideros

contains a wide variety of forms, from low shrubs to

trees. Trees in the genus are long-lived. Sample not securely associated with a cultural event.

I-16,652. 270 ± 80

C = -26.0‰

Unidentified wood charcoal, Site 50-80-15-4853, Locale 1 (Figure 1), Feature Hearth, Unit I14,

Zone I. Collected 1990 and submitted June 1991.

Comment

: Sample collected from a feature interpreted as a hearth (Rolett 1992:Table 1). Few con-

textual details.

I-16,654. 240 ± 110

C = -18.7‰

Unidentified wood charcoal, Site 50-80-15-4853, Locale 1 (Figure 1), Feature Imu, Unit I9-H10,

Zone I. Collected 1990 and submitted June 1991.

Comment

: Sample collected from a feature intepreted as an earth oven (Rolett 1992:Table 1). Few

contextual details.

I-16,688. 230 ± 80

C = -29.0‰

Unidentified wood charcoal, Site 50-80-15-4853, Locale 1 (Figure 1), Feature Hearth, Unit H8,

Zone I. Collected 1990 and submitted July 1991.

Comment

: Sample collected from a feature interpreted as a hearth (Rolett 1992:Table 1). Few con-

textual details. The conventional

C age of this sample was misreported by the laboratory as 290 ±

80 BP, the value reported by Rolett (1992). A report of the dating process provided by Teledyne

Brown Engineering, indicates the correct age of the sample, which is consistent with the reported

value D

C =

28 ± 10.

I-16,734. 770 ± 80

C = -28.4‰

Unidentified waterlogged wood and charcoal, Site 50-80-15-4853, Locale 1 (Figure 1), Unit H9,

Zone I. Collected 1990 and submitted September 1991. The “materials analyzed... are not from a

secure cultural context, and may thus represent natural materials unassociated with the archaeolog-

ical deposits” (Rolett 1992:3).

Comment

: Date provides an indication of the potential in-built age of waterlogged wood at Site 4853

and is not stratigraphically associated with Horizon II (see also Beta-18237).

ACKNOWLEDGMENTS

The author thanks J S Athens, R Cordy, M Desilets, T Higham, P V Kirch, B McFadgen, B V Rolett,

D Shideler, K Shun, H D Tuggle, and A C Ziegler for access to data and for helpful comments on

early versions of this paper. A Latham and M A Keenan, Teledyne Brown Engineering, provided

216

T Dye

copies of dating reports prepared by Teledyne Isotopes in 1991. Recent field research at Bellows Air

Force Station has benefitted from the support given by C S Streck and K Shun, US Army Corps of

Engineers, Pacific Ocean Division, and by J Simons, US Navy Pacific Division, Facilities Engineer-

ing Command. International Archaeological Research Institute, Inc. graciously supported the

research reported here. The author is responsible for any errors.

REFERENCES

Addison DJ. 1997. Archaeological subsurface testing for

the removal of pipeline utilities under the installation

restoration program at Bellows Air Force Station,

, Ko‘olaupoko District, O‘ahu, Hawai‘i.

Draft report prepared for US Army Corps of Engi-

neers, Pacific Ocean Division. Honolulu, Interna-

tional Archaeological Research Institute, Inc.

Athens JS. 1988. Archaeological survey and testing for

airfield perimeter fence project, Bellows Air Force

Station, Oahu, Hawaii. Report prepared for US Army

Engineer District, Pacific Ocean Division. Honolulu,

International Archaeological Research Institute, Inc.

Athens JS, Ward JV. 1993. Environmental change and

prehistoric Polynesian settlement in Hawai‘i. Asian

Perspectives 32(2):205–23.

Davidson J. 1992. Editorial. New Zealand Journal of Ar-

chaeology 14:5–7.

Desilets M, Dye TS. 1998. Archaeological monitoring

and sampling during bellows OU7 UST removal

project interim remedial action, phase I, Bellows Air

Force Station, , Ko‘olaupoko, O‘ahu.

Draft report prepared for US Army Corps of Engi-

neers, Pacific Ocean Division. Honolulu, Interna-

tional Archaeological Research Institute, Inc.

Dye TS. 1998. Archaeological services in support of the

final environmental impact statement for proposed ex-

pansion of military training and the construction of

improvements to existing recreational resources at

Bellows Air Force Station, , Hawai‘i. Re-

port prepared for Department of the Navy, Pacific Di-

vision, Naval Facilities Engineering Command. Ho-

nolulu, International Archaeological Research

Institute, Inc.

Efron B, Tibshirani RJ. 1993. An introduction to the

bootstrap. Monographs on statistics and applied prob-

ability. Nr 57. New York: Chapman & Hall.

Emory KP, Sinoto YH. 1969. Age of the sites in the South

Point area, Ka‘u, Hawaii. Pacific Anthropological

Records. Nr 8. Honolulu: Department of Anthropol-

ogy, BP Bishop Museum.

Erkelens C. 1996. Archaeological monitoring and data

recovery, underground storage tanks removal, Bel-

lows Air Force Station, Waimanalo, O‘ahu, Hawai‘i.

Draft report prepared for US Army Corps of Engi-

neers, Pacific Ocean Division. Honolulu, Interna-

tional Archaeological Research Institute, Inc.

Fletcher III CH, Jones AT. 1996. Sea-level highstands re-

corded in Holocene shoreline deposits on Oahu, Ha-

waii. Journal of Sedimentary Research 66:632–41.

Hall EO. 1839. Notes of a tour around Oahu. Hawaiian

Spectator 2(1):94–112.

Hammatt HH, Shideler DW. 1989. Archaeological sur-

vey and testing at Bellows Air Force Station for new

antennas and trench lines, Waimanalo, Ko‘olaupoko,

O‘ahu. Report prepared for US Army Corps of Engi-

neers. Kailua, Hawaii, Cultural Surveys Hawaii.

Hunt TL, Holsen RM. 1991. An early radiocarbon chro-

nology for the Hawaiian Islands: a preliminary analy-

sis. Asian Perspectives 30(1):147–61.

Kirch PV. 1974. The chronology of early Hawaiian set-

tlement. Archaeology and Physical Anthropology in

Oceania IX(2):110–9.

Kirch PV. 1985. Feathered gods and fishhooks. Hono-

lulu: University of Hawai‘i Press. 349 p.

Kirch PV. 1998. Archaeology. In: SP Juvik, JO Juvik, ed-

itors. Atlas of Hawai‘i. Honolulu: University of Ha-

wai‘i Press. 3rd edition. p 161–8.

McFadgen BG, Knox FB, Cole TRL. 1994. Radiocarbon

calibration curve variations and their implications for

the interpretation of New Zealand prehistory. Radio-

carbon 36(2):221–36.

McNeill JR. 1989. Intensive archaeological survey and

data recovery at site 50-80-15-3709, Bellows Air

Force Station, O‘ahu, Hawaii. Report prepared for US

Army Engineer District, Pacific Ocean Division. Ho-

nolulu, International Archaeological Research Insti-

tute, Inc.

Miller LO. 1991. Archaeological monitoring of the

Tinker Road bridge repair (replacement) project, Bel-

lows Air Force Station, Waimanalo, O‘ahu Island,

Hawai‘i. Prepared for T. Iida Contracting, Ltd. Hono-

lulu, Public Archaeology Section, Applied Research

Group, BP Bishop Museum.

Rolett BV. 1992. 1990 archaeological excavations at site

50-80-15-3300 (Bellows Air Force Station, Oahu)

conducted by the University of Hawaii Archaeological

Field School. Typescript in State of Hawaii Historic

Preservation Division Library.

Shun K. 1992a. Archaeological reconnaissance survey

and subsurface testing for proposed electric tie cir-

cuit, Bellows Air Force Station, Waimanalo, Kool-

aupoko District, Island of Oahu, Hawaii. Report pre-

pared for US Army Engineer District, Honolulu.

Kaneohe, Hawaii, Archaeological Associates Oce-

ania.

Shun K. 1992b. Archaeological sampling and monitor-

ing during installation of perimeter security fencing,

Bellows Air Force Station, Waimanalo, Koolaupoko

Effects of

C Sample Selection

217

District, Island of Oahu, Hawaii. Report prepared for

US Army Engineer District, Honolulu. Kaneohe, Ha-

waii, Archaeological Associates Oceania.

Shun K. 1993. Archaeological monitoring and sampling

during emergency flood repair construction, Waim-

analo and Inoaole streams, Bellows Air Force Station,

Waimanalo, Koolaupoko District, Island of Oahu, Ha-

waii. Report prepared for US Army Engineer District,

Honolulu. Kailua, Hawaii, Archaeological Associates

Oceania.

Spriggs M, Anderson A. 1993. Late colonization of East

Polynesia. Antiquity 67:200–17.

Stolk A, Törnqvist TE, Hekhuis KPV, Berendsen HJ, Van

der Plicht J. 1994. Calibration of

C histograms: a

comparison of methods. Radiocarbon 36(1):1–10.

Streck CF Jr, Watanabe FK. 1988. Memorandum for the

record. CEPOD-ED-PV. Typescript in State of Ha-

waii, Historic Preservation Division Library.

Strong CC, Skolmen RG. 1963. Origin of drift-logs on

the beaches of Hawaii. Nature 197:890.

Stuiver M, Pearson G. 1986. High-precision calibration

of the radiocarbon time scale, AD 1950–500 BC. Ra-

diocarbon 28(2B):805–38.

Taylor RE. 1987. Radiocarbon dating: an archaeologi-

cal perspective. New York: Academic Press. 212 zp.

Törnqvist TE, Bierkens MFP. 1994. How smooth should

curves be for calibrating radiocarbon ages? Radiocar-

bon 36(1):11–26.

Tuggle HD. 1997. Archaeological research of areas pro-

posed for development of military family housing and

expansion of military training at Bellows Air Force

Station, O‘ahu. Report prepared for Belt Collins Ha-

waii. Honolulu, International Archaeological Re-

search Institute, Inc.

Waterbolk HT. 1971. Working with radiocarbon dates.

Proceedings of the Prehistoric Society 37:15–33.