Pleistocene colonisation of the Bismarck Archipelago: new

33 Pages·2016·7.49 MB·English

Southern Cross University
ePublications@SCU
School of Environment, Science and Engineering
School of Environment, Science and Engineering
Papers
2004
Pleistocene colonisation of the Bismarck
Archipelago: new evidence from West New Britain
Robin Torrence
Australian Museum
Vince Neall
Massey University
Trudy Doelman
University of Sydney
Ed Rhodes
Australian National University
Chris McKee
Geophysical Observatory, Port Moresby
See next page for additional authors
Publication details
Torrence, R, Neall, V, Doelman, T, Rhodes, E, McKee, C, Davies, H, Bonetti, R, Guglielmetti, A, Manzoni, A, Oddone, M, Parr, JF &
Wallace, C 2004, 'Pleistocene colonisation of the Bismarck Archipelago: new evidence from West New Britain',Archaeology in Oceania,
vol. 39, no. 3, pp. 101-130.
The abstract and pdf of the published article reproduced in ePublications@SCU with the permission of Archaeology in Oceania
ePublications@SCU is an electronic repository administered by Southern Cross University Library. Its goal is to capture and preserve the intellectual
output of Southern Cross University authors and researchers, and to increase visibility and impact through open access to researchers around the
world. For further information please [email protected].
Authors
Robin Torrence, Vince Neall, Trudy Doelman, Ed Rhodes, Chris McKee, Hugh Davies, Roberto Bonetti,
Alessandra Guglielmetti, Alberto Manzoni, Massimo Oddone, Jeffrey F. Parr, and Cleland Wallace
This article is available at ePublications@SCU:http://epubs.scu.edu.au/esm_pubs/783
Archaeol. Oceania 39 (2004) 101-130
Pleistocene colonisation of the Bismarck Archipelago: new evidence
from West New Britain
ROBIN TORRENCE, VINCE NEALL, TRUDY DOELMAN, ED RHODES, CHRIS McKEE,
HUGH DAVIES, ROBERTO BONETTI, ALESSANDRAGUGLIELMETTI,
ALBERTO MANZONI, MASSIMO ODDONE, JEFF PARR, CLELAND WALLACE
Abstract
The geological and archaeological signatures al the site of Kupona The results of the Lapita Homeland Project and its
na Dari on the Willaumez Peninsula, West New Britain provide successors have demonstrated that people had arrived on the
important new data about human colonisation of the Bismarck islands of Near Oceania by at least 40,000 years ago (e.g.
Archipelago. Analyses of the stratigraphy and weathering of Allen and Gosden 1991; 1996; Allen 1993; 1996; 2000;
paleosols and manuporls. when combined with fission track. 2003; Gosden 1993; 1995; Leavesley and Alleti 1998;
radiocarbon, and luminescence dating, indicate that the site was Leavesley et al. 2002; Rosenfeld 1997; Spriggs 1997: 26-
39; 2001). In New Britain they spread very rapidly from the
first occupied at about 35-45.000 years ago. During the whole
coast into the rugged interior (Pavlides and Gosden 1994;
peri(Kl of occupation, people were exposed to a series of volcanic
Pavlides 1999; 2004). On the basis of these new data,
events which varied in terms of tbeir potential impacts on the local
Gosden (1993; 1995) and Allen (1996; 2000; Allen and
environment. A PIXE-PIGME characterisation study of the
Gosden 1996) proposed a two stage scenario for
obsidian artefacts at tbe site demonstrates that from the earliest
colonisation of this region. They identified a change in
period stone resources were acquired from outcrops located across
subsistence and settlement from a pattern in which people
a relatively large region. When compared with Early-Middle
moved their camps frequently between resource patches to
Holocene assemblages from nearby localities, the Pleistocene
one in which resources from a number of different settings
stone tool technology differs in only a few minor respects. From
were transported back to base camps. The finding of
this analysis we infer tbat groups were mobile in both periods, but
Kupona na Dari (FABM) on the Willaumez Peninsula, an
slightly different strategies for tbe procurement and maintenance
open site with numerous obsidian artefacts well stratified
of tbe stone tools were required for tbe more extensive ranges
between layers of volcanic tephra (Torrence et al. 1999)
exploited during the Pleistocene. The inter-disciplinary study of
(Figure I), provides an opportunity for re-examining their
Kupona na Dari concludes that colonisation comprised a long term
model of Pleistocene colonisation in the Bismarck
process of settling into this volcanically active environment. Due
Archipelago from the perspective of a very different
to variability in the environments tbat people encountered, the
environmental setting. The active volcanoes on the north
pattern of colonisation may not have been similar across the entire coast of New Britain would have created particular
Bismarck Archipelago, challenges for the first settlers that were not present iti
limestone regions. Furthermore, the human activities at this
open site are likely to have varied from those which took
place in the rock shelters that have previously dominated
research. The purpose of this paper is to report the results of
RT: Anthropology. Tbe Australian Museum, 6 College Street,
Sydney NSW 2010, Australia; VN, CW: Institute of Natural excavations conducted at Kupona na Dari in 2001 and to
Resources. Soil and Earth Sciences. Massey University. consider what new light these shed on the processes of early
Private Bag 11 222, Paimerston North, New Zealand; TD: colonisation.
Archaeology, University of Sydney. Sydney NSW 2006,
Australia; ER: Research School of Earth Sciences and
Research School of Pacific and Asian Studies. The Australian
National University, Canberra, ACT 0200, Australia; CM: Background
Geophysical Observatory, P.O. Box 232, Port Moresby, NCD,
Papua New Guinea; HD: Department of Geology. P.O. Box
The Kupona na Dari site, located on the coastal plain about
320. University, NCD, Papua New Guinea; RB, AG, AM:
6(X) metres from the current shoreline, was identified during
Istituto di Eisica Generale Appiicata, Universita di Milano, Via
an extensive survey of Numundo Plantation (Torrence et al.
Celoria. 16 20133 Milano, Italy; MO: Dipartimento di
Chimica, Universita di Pavia. Via TaramelH, 12 271(X) Pavia, 1999) (Figures I, 2). The site consists of a series of
Italy; JP: Centre for Geoarchaeology and Pale<.>environmental interbedded paleosols and tephras preserved as a small,
Research. Southern Cross University, P.O. Box 157, Lismore, roughly oval, hill (c. 32 x 34 metres) whose top stands about
NSW, Australia 10 metres above the current high water mark. Many of the
101
Willaumez Peninsula/
\flj
New Britain V^U
Dakataua^^,^-
PNG
•> {
\
&:
In 0
Gulu, ""^ \\ Australia
r^Baki
Kutau
t
^—--^
N
/
A
10 Km
/
Witori
FABMI
Mopir•
Figure 1: The Willaumez Peninsula showing the location of Kupona na Dari (FABM)
and the obsidian sources used during the Pleistocene.
same tephra layers that make up Kupona na Dari are also eroded to form steeper sides as shown in Figure 3. Some of
present in road cuttings in the nearby foothills of the the erosion probably took place in the relatively long gap in
Numundo ridge (Torrence el al. 1999: 44). The consistent volcanic activity represented by the unconformity between
mantle bedding of the heavily weathered Pleistocene tephras the top of the Pleistocene tephras and the base of the
and the conformable nature of the sequence suggest that this Holocene tephras. This clear break is exhibited by the very
hill may have been a persistent feature at this location - large differences in weathering of the siratigraphic layers in
since at least the time represented by the lowest tephras we these two periods (see below). The early Witori tephras W-
observed, some 5 metres below the top of the hill (Figure 3). Kl and W-K2, which date to c. 5900 and 3600 cal. BP
What appears at present to be a rather isolated hill was respectively (Torrence et al. 20(X); Machida et al. 1996),
probably connected to the higher ridges to the west which were eroded off the top of the hill soon after they fell and
we suspect are old flows emanating from the remnant then piled up around its base, thereby reducing the hillside's
caldera preserved as the current Numundo ridge (Figure 2). slope angle and buttressing the base with volcanic
Since it was located on the farthest promontory or colluvium. Subsequent tephras W-K3, W-K4. and the W-H
peninsula protruding from the ridge, the hill would have series (c. 18(X)-5OO cal. BP) were draped over the hill.
been a relatively prominent place before the emplacement of There may have been additional periods of erosion during
the Holocene tephras and associated alluviation created the the Pleistocene because our excavations revealed another
current coastal plain. The site has always been relatively marked unconformity on the north side (see discussion
close to the sea. Although the shoreline might have been below). Just over half of the western portion has been
closer before recent infilling, during the Last Glacial removed by modem quarrying, but this ceased in 1999 after
Maximum it was only c. 8 kilometres further away. artifacts were discovered within the Pleistocene layers
(Torrence er a/. 1999).
At some point during the Late Pleistocene, the hill was
102
A Excavation locations
..™ Contours (80m)
Landform units
Hi coastal foothill
!•
coastal hill
coastal plain
•1 Intand mountains
Figure 2: Local setting for Kupona Na Dari showing the location of Early-Middle Holocene
assemblages used for comparison. Radius of circle is 2km.
11 m IV V
,W-K3
10 -
'- H
• G
F
y^ Es
Et
- - " D1
/ \ " Cs c
Ct
•^ W-K3
^ D2 B
/ /
B-1 B-1
C ^ - '
yy.K3 e ^ ^ " - - ' ' A A
6 —
A
Win unexcavated
masi
unexcavated
unexcavated
Figure 3: Cross section of the Kupona na Dari hill with stratigraphic units identified by McKee.
103
The site was first visited in 1998 when the quarrying had Finally, the results of stone tool analyses are used to
just begun, but no artifacts were observed until 1999. As reconstruct patterns of human mobility and resource use. A
reported in Torrence et al. (1999). obsidian flakes were comparison of human behaviour during the Late Pleistocene
found stratified under and above a distinctive tephra and Early-Middle Holocene periods within the immediate
(Unit C), which is located between 2 and 3 metres below the region helps place the site within the long term context of
current ground surface (Figure 3), At the southern end of the human colonisation.
section in the soil above this tephra, artifacts were found
directly associated with a tight cluster of fire-cracked stones
(Figure 4). In 2001 New Britain Palm Oil Ltd. pushed a pile
Site Stratigraphy
of coconut logs off the top of the hill and onto the sides.
They then lent us an excavator and operator who cleared
Following initial work in 1999 by Davies (Torrence et al.
away talus created by the collapse of the original quarry face
1999), McKee and Neall studied the stratigraphy and
and carved out four new vertical sections ranging in height
collected samples of pumice and sediments for further
from 2-6 metres (Figure 5). Excavation proceeded beneath
geochemical and pedological analyses. McKee, who visited
the level of the current ground surface for about 1,5 metres
the site in 2000 and again shortly after the sections had been
until it reached a very hard layer of volcanic tuff (Unit A).
cleared in 2001, concentrated on the tephras. Following
Running from north to south, the cuttings were labelled II-
Davies' original terminology, his basic stratigraphic units
V, and the sections within them were assigned the same
(A-H) are demarcated by the flags and drawn lines as shown
roman numeral. Section IV, located in the tallest.central part
in Figure 6 (cf. Figure 3). These were used as the strata for
of the remnant hill, provides the best preserved sequence.
luminescence dating. Neall, who arrived at the close of the
Another area labelled cutting I was excavated into the
excavations in 2001, sampled all the units in Section IV and
northern side of the hill at right angles to the quarry face and
recorded a range of pedogenic variables and anthropogenic
intersecting with the top of section III. Later, cutting VI was
features. He made a few new subdivisions within McKee's
dug out by hand. The steps created by cuttings I and VI
stratigraphy, but there were no major differences between
provided safe access to the entire profile of the hill.
the two interpretations, Neall's interpretation of sections in
Following basic surveying and a preliminary stratigraphic
the archaeological excavations, along with observations
study of the cleared sections (cf. Figure 3), excavation took
made by the archaeologists during excavation, enabled
place within Squares A, B, and C, as described below.
excavated spits to be grouped into stratigraphic units.
We begin by describing the site stratigraphy in order to Artifacts found while studying sections or during soil
reconstruct the geological history of the site during the time sampling, both in previous years and during the 2001 field
of human occupation. Following this, we apply a range of season, were assigned to a stratigraphic unit by the geologist
techniques to provide a chronological framework for these on site at the time. Since the high degree of weathering
environmental changes and associated human activities. makes recognition of the units quite difficult, a pedological
description is presented in Appendix 1 to aid future
researchers in re-locating them at this location.
Geological history
A stratigraphic analysis of the units preserved at this
location provides the basis for a reconstruction of the
geological history of the coverbeds and, by implication, the
history of environmental change in the region during the
time that humans occupied Kupona na Dari. Before
presenting our results, we define key terms and briefly
describe our methodology.
A coverbed is an unconsolidated or weakly consolidated
surficial deposit forming a unit that mantles the landscape
V V V V VV V VV Q
and rests directly on underlying rock or on other coverbeds.
V V V v vV VV
They are usually of Quaternary age and include tephras.
loess, alluvium or aeolian sand. Stratigraphic studies at the
site have focused on Section IV with additional study of
+ obsidian artefacts
5 ctn Section VI (using sections from Squares A-C) (Figures 5-8).
The history and formation of the layers in each section have
been heavily influenced by their particular location within
the hill, as for example if one compares Section IV, which
Figure 4: Cluster of fire-cracked stones observed within
was placed in the middle of the hill, with Section VI, which
the quarry face in 1999.
is on the northern flank.
104
Conclusions about how the coverbeds were formed are
based on a visual assessment of the textural changes within
the units, the presence of the progressive accumulation of
fine-grained materials, whether soil formation was
continuous or absent, and the provenance for the sand and
silt mineralogy. An important distinction is made between
units produced by large plinian and sub-plinian volcanic
events and the process known as tephra accretion (Neall
1977). At distal localities during large magnitude eruptions
and at proximal localities in small eruptions, dustings of
volcanic ash are added to the surface of the soil without
wide scale destruction of the vegetative cover. The ash does
not fall directly onto the soil surface but is washed off leaves
and branches onto the forest floor by the next rains. Over
time, these many small additions of tephra build a soil
formed from the fine grained parent materials. Many metres
can be accumulated through tephra accretion. It is important
to recognize units built up in this way because, compared
with the deposits created over a short period in a single
eruption, units formed by tephra accretion can represent
considerable periods of time and may not necessarily disturb
Figure 5: Contour map of Kupona na Dari showing
human activities.
location of the cuttings and excavated squares.
Figure 6: Looking east toward quarry face at FABM. Section IV is in the centre. Flags and lines mark units identified
by McKee. Luminescence samples were removed from the holes. Section III which has been cut
by Squares A and B is at the far left. Square C is in the far distance and slightly to the right of A and B.
It cuts through the upper half of the layers represented in the hill. Scale is I metre.
105
East West
Disturbed
W-H
W-K4 soil
W-K4 tephra
W-K3 soil
W-K3 tephra Csoil
C tephra
B Lipper
H(G?)
0.5 m
F soil
F tephra
unexcavated
Figure 8; Section of the south side of Squares A and B
preserves the lower units at Kupona na Dari
stones
0.25 m
roots
manuports at this site, therefore, indicates human activity on
a relatively stable landscape.
Figure 7: Section of the south side of Square C preserves The eoverbeds studied in Sections IV and VI are
the upper units at Kupona na Dari. described below in order from the base up to the top. The
descriptions focus on the nature of volcanic activity that
formed them and the length of time over which they were
formed. Details about the mineralogy and additional
Manuports are rounded stones that could not have formed information about the type of eruption they represent arc
naturally within the airfall tephras and paleosols that make presented in Table 1. A summary diagram of the units based
up the stratigraphic sequence at this location. They must on the geological subdivisions identified by McKee and
therefore have been transported to this hilltop and as such pedological horizons' defined by Neall (shown in
are an important indicator of human activity. They are also parentheses below) is presented in Figure 9.
useful in interpreting the stratigraphy. Where scattered or
isolated stones are found within a coverbed, it is impossible
to determine if they are in situ or have fallen into a
subsurface pit or tree root hole. The situation is entirely Pedological horizons have been designated according to the
soil nomenclature of FAO/UNESCO 1974. The first number
different when they form a horizon at the junction of two
refers to each identified parent material. Each capital letter
units derived from independent processes (e.g. two tephras
refers to a soil master horizon; in this case all A's refer to
with different provenances). Firstly, all the stones would be
major paleosols (buried soils) and all B's refer to former
unlikely to travel down the profile at the same rate (i.e. large
subsoils or low organic matter rapidly accumulating soils.
and small stones would have moved at different speeds,
The one or two lower case letters refer to the main
thereby leading to a random spread) and secondly, it would pedological features of each horizon: p = human presence
be extremely fortuitious for the stones to become aligned on and influence; b = buried; t = horizon with clay
the same stratigraphic boundary. The presence of lines of accumulation ; w = weathered parent material in situ.
106
FABM FABM - FABM - FABM - Section \ I - Squares A & 8 FABM - Section \ i . Squares C
Section |\ (Part 1) Section l\ (Part 2) Section 1\ (Part 3)
II
4B»;
w
.B.
Sul.n
c
'.'.(• , 1
Toptft KEY
C
paleosol
Ahonzww
t iapifli
coafsoasti
1 fine ash
Figure 9: Summary sections reconstructed by Neall.
Pedological Geological > 250 }tm fraction < 250 ftm fractioD Interpreted Depositional
sample unit dominant minerals dominant minerals provenance tnechani.sm
2Bt H grey volcanic lithics opaques > brown weathered rhyolitic slow tephra accretioti
> euhedral quartz volcanic glass
3Bt G black & white volcanic lithics opaques > brown weathered dacitic slow tephra accretion
> orange pumices volcanic glass
4Bw F volcanic lilhics opaques > quartz & feldspar andesitic = dacilic slow tephra accretion
> euhedral quartz
5Apb Upper E grey & orange volcanic lithics opaques > brown weathered andesitic > dacitic soil formation and very
volcanic glass slow tephra accretion
6Bw Lower E volcanic lithics > opaques opaques > brown weathered andesitic distal plinian
(quartz absent) volcanic glass (quartz absent) eruption
7Apb D3 volcanic lilhics opaques > brown weathered dacitic > andesitic soil formation on
> euhedral quartz volcanic glass previous lephra accretion
8Bw D2 grey volcanic lithics grey volcanic lithics andesitic > dacitic distal subplinian
(crystals rare) > quartz & opaque or plinian eruption
9Bw Dl mixed volcanic lithics mixed volcanic lithics dacitic > andesitic distal plinian
> quartz & opaque > quartz & opaques eruption
lOApb Soil on C volcanic lithics brown weathered glass andesitic > dacitic soil formation on
> euhedral quartz > opaques previous tephra accretion
lOBw Middle C volcanic lithics > feldspar opaques = brown weathered andesitic > dacitic slow tephra accretion
volcanic glass
llBw C grey volcanic lithics opaques = brown weathered dacitic proximal plinian
volcanic glass eruption
l2Apb Upper B grey volcanic lithics opaques > brown weathered andesitic soil formation
> orange-brown weathered volcanic glass with very slow
lithics (quartz absent) tephra accretion
l2Bw Lower B grey volcanic lithics opaques = brown weathered dacitic slow tephra
volcanic glass accretion
l3Bw Bl brown and grey volcanic lithics brown weathered glass dacitic distal plinian eruption
> opaques during moderately fast
lephra accretion
14Bw A brown weathered glass > orange. brown weathered glass andesitic proximal plinian and
grey and green weathered lithics > opaques ?phrealoplinian eruptions
Table I: Mineralogy, provenance and deposition of the coverbeds at Kupona na Dari
Section IV (stratigraphic thickness = 3.6+ m) D3 (7Apb) is a paleosol formed within the tephra. Its
presence indicates a significantly long period of time. One
tabular stone manuport 50 mm long was found in this
Unit A (I4Bw) is a lapilli tuff that shows shower bedding
paleosoi.
indicative of an airfall origin. Due to its vesicular nature, it
may be phreatomagmatic. It was probably deposited over a
Unit Lower E (6Bw) is of primary volcanic origin
relatively short time period because it has a very coarse
texture and there are no paleosols between the beds. This comprising 15-35% orange lapilli, and is likely to bave been
unit is hard and forms an indurated base to the whole deposited over a short period of time.
section. The observed lapilli textures suggest Utiit A is the
result of a plinian eruption. During a hiatus in volcanic Unit Upper E(5Apb) is a paleosol developed within a bed
activity, represented by the break between this unit and the that accumulated by tephra accretion. A horizon containing
one above it, people brought stones to this locality scattered stone manuports up to 40mm in size occurs on the
(manuports) and deposited them on the relatively stable upper surface of tbe unit. Like the other paleosols, tbis unit
landscape. represents a considerable period of time in the succession.
Unit F (4Bw) represents a long period of fme tepbra
Unit BI (J3Bw) represents a moderately weathered tephra
accretion. Rare stone manuports up to 70 mm long occur on
which contains distinctive orange lapilli and may be
top of tbe unit. Tbey reflect a hiatus in soil formation before
phreatomagmatic. This unit was formed by tephra accretion
the next unit was emplaced.
over a relatively long time period. Phytoliths are common in
the <32/^m fraction. A prominent stone line at the top of the
Unit G i3Bt) on first glance appears similar to Unit F but is
unit indicates another hiatus in volcanic activity.
subtly differentiated by the presence of waxy coatings on
the soil aggregates (argillans) and tbe visual appearance of
Unit Lower B (I2Bw) lacks orange lapilli in Section IV,
mafic and creamy weatbered coarse ash grains. The whole
although they were seen within this unit in Sections III and
bed contains <5% of stone manuports between 20 and 50
VI. The unit seems to represent another period of tephra
mm diameter throughout. Our interpretation is tbat this unit
accretion with a similar provenance to Unit B1.
represents a long period of tephra accretion during wbich
there was considerable anthropogenic mixing. There may be
Unit Upper B (l2Aph) represents a long period of relative
a small phreatomagmatic event preserved within it.
volcanic quiescence when a soil was able to be formed.
Unit H (2Bi) differs from unit G in colour and structure, but
Unit C (JJBw) is one of the key tephra marker beds in the
also displays a number of similarities. It also bas waxy,
section. It has a loamy sand texture, wbich contrasts with
chocolate-coloured coatings on tbe soil aggregates
the high amount of clay in the encompassing units. It would
(argillans) and contains rare stone manuports throughout,
have been deposited as a coarse ash in a single volcanic
suggestive of human presence during its formation. Like
episode lasting over a relatively short time period. The unit
unit G, it probably also represents a long period of tephra
has the potential to be a valuable chronological marker at
accretion contemporary with buman occupation.
other Pleistocene sites in the region.
Soil on H (2Apb) is a thin (now exhumed) paleosol which
Middle C (lOBw) has a much finer texture than llBw. It developed within tbe uppermost part of unit H. When tbe
represents another long period of tephra accretion. site was visited in 1999 and 2000, it was overlain by the
W-K 3 pumice lapilli, but this has since been removed by
Soil on C (JOAph) again represents a long period of volcanic recent human activities. The paleosol on H at this site
quiescence during which soil formation extended down into therefore represents a longer time interval than the paleoso!
theparent material of Unite. Between 5 and 15% of the soil found beneath W-K 1 elsewhere in the region. The absence
is comprised of strongly weathered stone manuports, one of of W-K I and W-K 2 prior to the deposition of W-K 3
which measured 60mm across. Many phytoliths occur in the implies removal in the time interval 1700 to 5600 yr BP
m fraction. (Machida et al. 1996). Although erosion by heavy rainfall
cannot be discounted, the unusual removal of the Holocene
DI {9Bw) is clearly a short term airfall eruptive with a few tepbras at tbis location may imply human interference.
lapilli. Like Unit C, it bas potential value as a regional
chronostratigraphic marker at late Pleistocene sites in the
region. Section VI (straiigraphic thickness - 4.1 + m)
D2 (8Bw) is another tephra of coarser ash texture whicb As in Section IV, the absence of the W-K I and W-K2
forms discontinuous lithified lumps in tbe soil (known as tepbras forms an obvious unconformity between the W-K 3
'creamcakes'). It may be phreatomagmatic. It would have tephra and tbe Soil on H (2Apb) (Figure 7). The underlying
been deposited over a relatively short period of time. units, H, G, and F, show markedly different properties to
108