Content uploaded by Melanie Fillios
Author content
All content in this area was uploaded by Melanie Fillios on Aug 02, 2019
Content may be subject to copyright.
Abstract
This paper describes the results and implications of recent
excavations on the Hamersley Iron Brockman 4 tenement, near
Tom Price, Western Australia. Results concentrate on two rock
shelters with Aboriginal occupation starting at least 32,000 years
ago and extending throughout the Last Glacial period. Preliminary
observations are proposed concerning the nature of Aboriginal
foraging patterns as displayed in the flaked stone and faunal
records for the Brockman region.
There are two enduring research questions that have
dominated the archaeology of the Pilbara since research
commenced in the region some 30 years ago – antiquity and
continuity. The antiquity of the Pilbara continues to be
important in our developing understandings of the timing
and directionality of continental settlement. At the time
Steve Brown provided the first synthesis and review of
initial archaeological excavations for the Hamersley Plateau
(Brown 1987) the oldest radiocarbon date from an
archaeological site was 26,300 + 500 BP (SUA1510) for
Newman Rockshelter (P2055.2) (Brown 1987:22, citing
Troilett 1982). Despite 20 years of intensive archaeological
work (mostly consulting projects including over 50
excavations with more than 100 radiocarbon dates, see
Slack 2008), this antiquity has only recently been surpassed
with age estimates for excavations at Djadjiling at Hope
Downs in the eastern Hamersley Range, indicating possible
occupation of at least 35,159 ± 537 years BP (Morse, this
issue, Table 1).
Associated with the issue of the timing of Aboriginal
Pleistocene occupation of the Pilbara has been the question
of whether such occupation endured into times of markedly
increased aridity associated with the last glacial period (i.e.
OIS2, between about 29,000–15,000 cal. yr BP, see
Burroughs 2005:30, 93), and in particular the peak of aridity
associated with the Last Glacial Maximum (LGM) when sea
levels off northwestern Australia were at their lowest
between 22,000 and 19,000 cal. yr BP (Yokoyama et al.
2000). The nature of Aboriginal occupation spanning the
LGM has, over the last few decades, become a consistent
focus for research in the Pilbara region, with the Hamersley
and Chichester Ranges proposed as likely refuges (Hiscock
1988; Smith 1987, 1989; Veth 1989, 1993). It is only over
the same few decades that research has shown the impact of
the LGM on climate to be more severe and at an earlier time
than previously thought (and peaking at 21,200 cal. yr BP in
the Greenland ice-core isotope stratigraphy see Turney et al.
2006 and Barrows and Juggins 2005).
The nature of regional occupation patterns during the
revised and extended LGM has been summarised, based on
seven specific rockshelter sites in the Pilbara Uplands
argued to exhibit refuge occupation during the LGM (see
O’Connor and Veth 2006: 33-39) (Figure 1). Of these sites,
only Yirra (Veitch et al. 2005) and Milly’s Cave (Marwick
2002) were said to feature persuasive evidence of LGM
occupation. O’Connor and Veth (2006) concur with
Marwick (2002) that the five other sites have no
unequivocal evidence of LGM occupation. Marwick’s
analysis found that of these sites the first two, Newman
Rockshelter (Troilett 1982) and Newman Orebody XXIX
Rockshelter (Maynard 1980), have stratigraphic records and
radiocarbon chronologies that suggest, but do not confirm,
evidence of human occupation 13,000 to 17,000 years ago
[i.e. 15,000–20,000 cal. yr BP1](Marwick 2002:23; see also
Comtesse 2003). Similarly, evidence of human occupation
in this period is regarded as ambiguous. At Mesa J J24
(Hughes and Quartermaine 1992), Malea Rockshelter
(McDonald, Hales and Associate 1997) and Manganese
Gorge 8 (Veth 1995:736) uncertainties plague interpreta-
tions of artefacts and their relationship to carbon dates.
This leaves Yirra and Milly’s Cave as the only sites in the
Pilbara interior occupied during the LGM. At Yirra, artefacts
said to occur between conventional radiocarbon ages of
19,270 ±140 BP (Wk-8954) (23,440–22,480 cal. yr BP) and
16,950 ± 90 (Wk-9148) BP (20,300–19,889 cal. yr BP) are
considered consistent with LGM refuge occupation (V
eitch
et al. 2005:58). However, despite acknowledged unresolved
bioturbation problems, with critical dates at the peak, and
with little additional information concerning artefact
frequencies, the site and local climatic history,it is uncertain
whether Yirra was occupied more intensively at the height
of the LGM or immediately afterwards.
Marwick’s 2002 paper claimed that the site of Milly’s
Cave provided the only clear indication of human
Archaeol. Oceania 44 Supplement (2009) 32–39
32
Aboriginal Settlement during the LGM at Brockman,
Pilbara Region, Western Australia
MICHAEL SLACK, MELANIE FILLIOS and RICHARD FULLAGAR
Keywords: excavations, refuge, LGM, Brockman, Pilbara
Scarp Archaeology, PO Box 7241, South Sydney Hub, NSW
2015; Michael.Slack@scarp.com.au
1. Radiocarbon calibrations in this article were completed using OxCal
v4.0.5, IntCal04.
occupation during the LGM period. We agree, but the recent
re-evaluation of the timing of the LGM (Yokoyama et al.
2000; Lambeck and Chappell 2001) suggests that this site
may have been occupied only sporadically until after the
LGM. The lowest radiocarbon determination and artefact
frequencies at Milly’s Cave could plausibly indicate more
intense occupation immediately after the LGM peak, and
discard rates underlying this level (between about 21,000
and 30,000 cal. yr BP) are very low (see Marwick 2002:25).
Wealso note that the lower two radiocarbon dates for the
site 14,150 ± 320 BP (18,024–16,022 cal. yr BP) and 18,750
± 460 BP (23,686–21,075 cal. yr BP) are separated by
possibly only 5 cm of deposit. As such, we suggest that the
Milly’s Cave data provide more compelling evidence of
increased occupation towards the end of the LGM, rather
than a persistent occupation throughout.
Accepting for the moment that the Hamersley Plateau
constituted a refuge for people during periods of extreme
regional aridity, then what was the nature of this occupa-
tion? Marwick, in reference to Milly’sCave, suggests that
territorial ranges reduced in area. O’Connor and Veth
consider more generally that retraction into and within the
ranges occurred, but that differences in relation to local
catchments would be evident, ranging from complete
abandonment through to increased use (O’Connor and Veth
2006:41).
A significant hindrance to understanding the utilization
of refuge areas before, during and after the LGM lies in the
dearth of evidence for subsistence. Foremost among the
evidence needed are systematically analysed organic
remains in conjunction with flaked stone. Edwards and
O’Connell (1995) discussed the shift toward broad spectrum
diets at the terminal Pleistocene, yet true understanding of
the phenomenon has yet to be achieved, primarily due to the
precious few excavated sites with evidence of occupation
that includes not just flaked stone, but faunal (and flora)
remains.
There is now solid evidence of Pilbara occupation during
and prior to the LGM, but cultural remains are limited to a
few excavated, and fewer still, published sites. Sites with
evidence of early occupation, such as Newman Rockshelter,
Newman Orebody XXIX Shelter,Malea Rockshelter and
Milly’sCave offer little in the way of faunal material.
Faunal remains were preserved at Malea (Edwards and
Murphy 2003), but thus far the published information
consists of little more than a species list accompanied by the
statement that the assemblage is highly fragmented
(Edwards and Murphy 2003:45). At Malea, faunal material
was not recovered from every excavation unit, but confined
to the upper 16 units. Edwards and Murphy argue that this
occurrence is likely due to preservational factors, not true
absence. Work at Malea has since been renewed, and we are
currently awaiting analyses which we hope will supply
much needed data to bolster our understanding of
subsistence and settlement in the area. Faunal remains were
also preserved at Marillana A, although discussion is limited
to a quantitative analysis of density per stratigraphic unit
(Marwick 2005:1363-4).
Faunal remains at Newman Orebody XXIX are limited to
one macropod molar in the top excavation unit (Maynard
1980:5), while data are absent from Newman Rockshelter
and Milly’sCave (Marwick 2003). The absence of faunal
data in the Pilbara is a significant hindrance to our
understanding of refuge areas, as well as early Aboriginal
subsistence as a whole.
Wereport here new sites in the region with the potential
to provide important subsistence data and more robust
frameworks concerning of Aboriginal settlement in the
Hamersley Plateau during the LGM.
Excavations at Brockman 4,
Hamersley Plateau
Recently Scarp Archaeology completed a series of
excavations within the Rio Tinto Brockman 4 mining
tenement, located in the Western Pilbara approximately
60 km west of the town of Tom Price. The results of
excavations at two particular sites: Juukan-1 and Juukan-2
provide further substantiation of a regional antiquity of
occupation of over 30,000 years, and compelling evidence
that occupation persisted even during the height of the LGM
(22,000–19,000 cal. yr BP). These results are interesting,
and somewhat surprising given the location of Brockman
well within the central Hamersley Plateau and over 75 km
north of the nearest substantial watercourse (albeit
ephemeral), the Ashburton River.
Both Juukan-1 and Juukan-2 are located within a small
ironstone gorge, near to a small ephemeral watercourse
known as Purlykunti Creek (Figure 2). Three other rock-
shelter sites occur within this gorge, however all feature
very recent occupation sequences. Below the gorge on an
extensive floodplain a very large open artefact scatter also
occurs. It is thought that the dominant stone raw materials
including ironstone, chert, quartz and siltstone are all
available from the creek at and near to the open scatter.
33
Figure 1. The Hamersley Range area with sites discussed.
Juukan-1
Juukan-1 is a south facing ironstone rockshelter approxi-
mately 25 m width, 8 m deep, and has a dripline about 10 m
high. The site features a higher collapsing rear chamber and
an open entrance area set at a slightly lower level with a lot
of roof fall separating the two areas. The floor of the rear
chamber consists of soft sediment that slopes down from the
rear.Flaked stone material was recorded along the front of
the site, particularly in the western end of the shelter.
During August 2008 a single 1 x 1 mtest pit was
excavated in the front chamber of the rockshelter.
Excavations reached a depth of 75 cm below the surface
where a level of solid roof fall or bedrock was encountered.
The stratigraphy of the test pit consisted of three main
layers: a topsoil of loose material overlying brown/grey
compacted sediment with many organic finds, which in turn
overlies a horizon of orange/brown. A small pink/white lens
of soft chalky material was also noted at a depth of
approximately 40 to 50 cm below the surface, as were small
lenses of charcoal (Figure 3).
A series of three radiocarbon determinations show that
discard of cultural material occurred at Juukan-1 from as
early as 32,950 ± 270 BP (conventional radiocarbon age)
(Beta-249759) at a depth of 60 cm. Until 35 cm below the
surface, which is dated to 26,640 ± 160 BP (conventional
radiocarbon age) (Beta-249758), sediment accumulation
and artefact discard were slow. From this point however, a
generally more rapid accumulation rate is proposed
extending to recent times with a near surface age
determination of 760 ± 40 BP (conventional radiocarbon
age) (Beta-249757) (740–660 cal. yr BP).
Atotal of thirty two stone artefacts were recovered from
Juukan-1. The majority of these occurred in spits 1 and 2,
with only individual artefacts recorded in lower spits. The
lowest artefact recovered from the site however was found
in spit 14 at a depth of approximately 70 cm, underlying the
date of 32,920 years BP.All of the flaked stone was
recorded in the far south eastern corner of the test square,
and it is likely that further planed extensions to the
excavations will provide a greater assemblage size.
In addition to flaked stone, a total of 67 fragments of
animal bone were recovered, of which 57 were identifiable.
Species identified included bandicoot, kangaroo, wallaroo,
native mouse, rat and one fish fragment (Table 1). Animal
bone from small to large species was recovered from most
spits, with the exception of spit 9, with the majority of the
bone belonging to medium-large macropods. Overall faunal
density is consistently small, with the exception of spit 12,
in which nearly 50% of the recovered bone was found
(Figure 4). All bone is highly fragmented with long bone
shaft fragments accounting for 66% of the bone recovered,
followed by teeth (9%) (Figure 5).
Species Common name NISP % Total NISP
Macropus rufus Red kangaroo 13 23%
Macropus robustus Common wallaroo 40 70%
Isoodon sp. Bandicoot 1 <2%
Pseudomys sp. Native mouse 1<2%
Rattus sp. Native rat 1 <2%
Fish Fish 1 <2%
Total NISP 57
Table 1. Juukan-1 Species Frequency
Atotal of nine fragments of bone are burned, two of
which are calcined, suggesting deposition in fire for longer
34
Figure 3. Site plan and stratigraphic profile drawings
of Juukan–1.
Figure 2. Juukan gorge looking towards Purlykunti Creek
periods of time. In addition to burning, taphonomic analyses
reveal several specimens with evidence of heavy
mineralization, potentially suggestive of greater time depth.
Bone from Juukan-1 is unweathered with no obvious post-
depositional surface modifications, suggesting the faunal
assemblage is an in situ deposit with rapid burial.
The frequency of long bone shaft fragments, coupled
with the high degree of fragmentation in the faunal
assemblage is suggestive of either human or carnivore
activity. However, given the absence of animal gnaw marks
on the surviving bone, coupled with evidence of burning
from fire, humans are considered the most likely
accumulator in this assemblage. Given the small sample
size, it is difficult to draw conclusions concerning changes
in species utilization and/or frequency diachronically.
However, the extant faunal evidence suggests both species
and element frequencies remain constant throughout the
spits containing bone, showing no obvious signs of
diachronic change. Additionally, the highest frequency of
bone occurs in spit 12, associated with the oldest date (and
possibly) with the oldest artefact, providing even clearer
evidence for early occupation.
Juukan-2
Juukan-2 is a large cavernous rockshelter located 50 m to
the west of Juukan-1. The site consists of two south-facing
chambers; a large western chamber with very deep sediment
and cathedral-like roof height, and a lesser eastern chamber
which is small and largely unprotected with a bare rock
floor. The main chamber is 10 m wide, 10 m deep and has a
height at the dripline of about 8 m. There are three general
areas to the main chamber; a scoured and rocky area with
some plants in the western side where a hole in the roof has
allows rainfall to enter, a main central area where large roof
fall at the front has aided in trapping extensive sediment,
and a raised area at the eastern rear of the site where bedrock
is higher than other areas and where sediment deposition is
minimal.
Like Juukan-1 a single 1 x 1 m test pit was excavated in
2008. The excavations consisted of 21, 5 cm spits that were
concluded at a depth approximately 1.05 m below the
surface, where large pieces of roof fall stopped any further
excavation. Five stratigraphic units were recorded within
the deposit, largely related to the changes in ironstone
weathering and minerals (Figure 5). At least five hearths
were also noted. A high frequency of flaked stone and
animal bones were also recovered.
Three carbon samples from spits 2, 12 and 17 returned
dates of 470 ± 40 years BP (Beta-247330) (540–490 cal. yr
BP), 16,160 ± 80 years BP (Beta-247331) (19,490–19,080
cal. yr BP) and 20,090 ± 100 years BP (Beta-247332). The
lowest date of 20,090 years BP was obtained using AMS
techniques and is derived from a depth of 85 cm below the
surface. This date represents neither the level of the lowest
artefacts or a basal date for the site. It is likely that the
deposit of Juukan-2 might be up to 0.5 m deeper than our
excavations.
A total of 272 flaked stone artefacts were recovered from
the test pit excavation. Artefacts were noted in all but spit 16
of the excavation, with the lowest recorded in spit 18 (at a
depth of 90 cm), below the lowest age determination of
20,090 years BP. The assemblage is dominated by
unmodified flakes (95.2%), with few retouched flakes
(4.4%) and even fewer cores (0.4%). Retouched artefacts
are chiefly comprised of chert (n = 8), and ironstone (n = 4).
The rates of discard for the flaked stone assemblage are
low, but remain steady throughout the period of occupation,
until spit 4 at approximately 5000 years BP where discard
increases fourfold (Figure 6). Although limited by the
sample size, there does not appear to be evidence of a hiatus
in occupation or sedimentation at the site before, during and
after the LGM peak.
Analysis of raw material richness and diversity of
artefacts shows that the assemblage is comprised of five
types of stone; ironstone, chert, quartz, chalcedony and
siltstone. Chert and quartz dominate the assemblage (55.9%
and 29%) with lesser amounts of ironstone (13.6%),
chalcedony (1.1%) and siltstone (0.4%). Interestingly
ironstone is as dominant a raw material in the lower spits as
chert and quartz until the aforementioned massive increase
in discard from 5000 years BP onwards. The first retouch in
35
Figure 5. Juukan-1 Skeletal Element Frequency (NISP).
Figure 4. Juukan-1 Faunal density by excavation unit
(NISP).
the assemblage occurs in spit 14 at about 19,000 years BP,
with a sharp peak occurring in spit 5 at about 7000 years BP
where the first evidence of backing occurs.
The rate of fragmentation of flakes shows dominance of
complete flakes (81.9% n = 222), with far lesser quantities
of broken flakes – distal account for 7.7% (n = 21), proximal
for 5.2% (n = 14) and medial for 4.1% (n = 11). The ratio of
broken to complete flakes is very low until spit 3 at about
4000 years BP where complete flakes account for just
65.2% (n = 43) and broken flakes total 34.8% (n = 23). This
is most likely the result of treadage, with more intensive use
of the shelter proposed at this time during the Mid-Holocene
El Nino arid phase experienced in Northern Australia.
Up until the mid Holocene, the assemblage is dominated
by ironstone flakes. These are generally heavier and have a
greater size range especially during the periods between
about 15,000 and 5000 BP.After 5000 BP, chert is the
dominant raw material and the average weight of flakes is
much less than 1 g. Retouched flakes are generally heavier
than unmodified flakes, and in the case of ironstone,
significantly so.
The relationships between the size of complete flakes and
the extent of their reduction is further supported by analysis
of the amount of cortical surface still present on the dorsal
surface of the flake. This analysis shows that ironstone
flakes are far more likely than chert to have more cortex,
indicating that they have been less reduced. Additionally
both chert and quartz flakes are far more likely than
ironstone flakes to have smaller and more reduced
platforms, as evidenced by single and multiple flake scar
platforms.
In addition to the flaked stone, 857 fragments of animal
bone were recovered from Juukan-2. A wide variety of
species is present, with small species (native rats/mice,
lizard, snake) comprising the majority of the recovered
specimens (61%, NISP = 523). Medium-large macropods
(kangaroo, wallaby) comprise 30% (NISP = 255), and the
assemblage is rounded out by fish and bird fragments.
Species identified include: Red kangaroo, common
wallaroo, bandicoot, possum, pygmy possum, echidna,
bettong, native mouse, rat, gekko, skink, small bird and fish
(Table 2), with bone recovered from nearly every spit
(Figure 8).
Species Common name NISP % Total NISP
Macropus rufus Red kangaroo 68 18%
Macropus robustus Common wallaroo 143 38%
Isoodon sp./ Bandicoot 9 2%
Perameles sp.
Trichosurus vulpecula/. Possum 5 1%
Pseudocheirus sp
Cercartetus sp. Pygmy possum 2 <1%
Tachyglossus aculeatus Echidna 9 2%
Bettongia lesueur Bettong (boodie) 3 <1%
Antechinus sp./ Native mouse 37 10%
Pseudomys sp.
Hydromys sp./Rattus sp. Native rat 89 24%
Gekkonidae sp./ Gekko 8 2%
Scincidae sp. Skink
Aves sp. Bird 2 <1%
Fish Fish 1 <1%
Total NISP 376
Table 2. Juukan-2 species frequency.
All major skeletal elements are represented in the
assemblage, with long bone shaft fragments contributing the
highest number of fragments (NISP =120). Differences in
element frequency are present, however, between smaller
and larger species. Elements belonging to smaller species
are dominated by limb bones and are largely unfragmented,
with a large proportion of complete skeletal elements. The
most frequently occurring elements are teeth and vertebra,
respectively. In contrast, bone belonging to medium-large
individuals is highly fragmented and heavily weighted
toward long bone shaft fragments (Figure 9). There is
however, a relative paucity of macropod lower limb
elements which can likely be accounted for by the long bone
shaft fragments. Teeth from medium-large macropods show
amixture of tooth wear stages ranging from unworn to
extremely worn, suggesting a mixture of young and old
individuals.
Seven percent (NISP = 61) of the assemblage shows
evidence of burning, with a third (NISP =18) of these being
calcined, suggesting deposition in fire for longer periods of
time in a defleshed state. Burned bone belongs to a range of
36
Figure 6. Plan and section drawing of Juukan-2.
37
Figure 7. Graph of flaked
stone artefact discard at
Juukan-2.
Figure 9. Juukan-2 Skeletal
Element Representation
(NISP).
Figure 8. Juukan-2 Faunal
density by excavation unit
(spit).
species, and is not confined to any one class of individual.
In addition to burning, five specimens show evidence of
tooth marks, and five fragments from spit 15 have possible
cut marks, including a kangaroo sacrum, pelvis and long
bone shaft fragment. Taphonomic analysis suggests the
faunal assemblage represents an in situ deposit with rapid
burial, as the bone is unweathered and there is no physical
evidence on the bone surfaces for either aeolian or fluvial
transport.
At Juukan-2 analysis of the fauna may provide important
diachronic information on subsistence strategies in the
region. While the majority of species and skeletal element
distributions appear to remain consistent through time,
medium macropods (wallaroos) were more common at the
beginning of occupation, while the frequency of large
macropods (kangaroos) increases toward the end of the
occupation (Figure 10). It is likely that the macropod
presence is a cultural rather than natural accumulation, as
burned and calcined bone, taken in conjunction with
evidence for hearths, suggests people were likely
responsible for some of the faunal accumulation. Heavy
fragmentation of macropod lower limb bones also supports
aprimary human role in the accumulation, as do possible cut
marks. The presence of bone from less common species
such as echidna and fish lends further weight to the notion
of people as bone accumulators. Preserved faunal remains
are rare in rockshelters, and further faunal analysis at sites in
which bone is present has the potential to contribute
significant information regarding species exploitation in the
Pilbara for which there is a severe dearth of published data.
Discussion
The results of our excavations at Brockman provide
important new information concerning the prehistory of the
Pilbara. First, our data provide further support for early
occupation of beyond 35,000 BP. Second, the cultural
sequence at Juukan-2 indicates a continual, albeit
infrequent, occupation of the Brockman region during OIS
2, and even at the height of the LGM.
In terms of hunter gatherer landscape use, our analysis of
the data is limited by sample size, however a number of
observations and hypotheses are suggested. That these two
rockshelters have evidence that people had been in this area
of the Hamersley during the LGM indicates that a local
population may have actually been more residentially
mobile during enhanced aridity than what we might expect,
given the dominant refuge models and their previous
application to the Pilbara (but see Veth 2005:101). It is clear
that people were not just retreating into gorges on the
margins of the ranges near to the main river courses, but that
amore complex use of the landscape, perhaps following
local weather patterns and allowing access to the less
drained areas occurred. Further to this we consider that it is
likely that with greater rainfall after the LGM residential
mobility decreased. This would explain the greater levels of
discard at Brockman, the increase in the density of faunal
remains at Juukan-2 in the later phases of occupation, and
also those trends observed in the flaked stone by Marwick at
Milly’sCave (2002:29). However at Brockman, during
periods of greater rainfall, we think it possible that while
residential mobility decreased, logistical mobility increased
(at least on a local level). This is suggested by the greater
range of raw materials and larger sizes of flaked stone
during the last few thousand years of the Pleistocene and up
to the middle of the Holocene. Also apparent is that the
intensity of reduction and frequency of artefact discard
increased slightly in the mid and late Holocene. Behavioural
implications suggested by the faunal remains provide
additional support for the increase in occupation during later
periods (however,the fauna also show that both Juukan-1
and Juukan-2 were used continually in all periods). Given
38
Figure 10. Juukan-2 Frequency
of large vs. medium
macropods by excavation unit
(NISP).
the results of other excavations within the region that all
date to this period, this trend is likely to be related to
increased population levels – as suggested by Marwick
(2002).
Conclusion
The results of this ongoing project further emphasise that
the archaeology of the Pilbara region will continue to play
an important role in developing our understandings of the
timing of arid settlement, and the nature of hunter gatherer
subsistence during periods of uncertainty like the LGM.
Further excavations at Brockman will focus on these issues
and with hopefully greater sample sizes obtained, on the
nature of technological innovation and its relationship to
mobility in such a marginal landscape over such a long
period of time.
Acknowledgements
This project would not have been possible without the
tireless work of the following people from the Puuntu Kunti
Kurruma and Pinnikura and Scarp Archaeology who were
involved in the excavation and analysis of the Juukan
sites; Harold Ashburton, Corbett Ashburton, Charleston
Ashburton, Jimmy Ashburton, John Ashburton, Lenny
Ashburton, Kate Connell, Burchell Hayes, Arness James,
Robert James McKay, and Sarah Robertson. We would like
to thank Rio Tinto Iron Ore, particularly Ed Clarke, Merv
Lockyear, Jason Masters and Amy Stevens. We also
acknowledge the support of Pilbara Native Title Services.
References
Barrows, T.T. and Juggins, S. 2005. Sea-surface temperatures around
the Australian margin and Indian Ocean during the Last Glacial
Maximum, Quaternary Science Reviews 24:1017-1047.
Brown, S. 1987. Toward a Prehistory of the Hamersley Plateau,
Northwest Australia.Occasional Papers in Prehistory 6, Depart-
ment of Prehistory,Research School of Pacific Studies, Australian
National University.
Burroughs, W.J. 2005. Climate Change in Prehistory.Cambridge: CUP.
Comtesse, S. 2003. Mt. Newman Sites Re-analysed: Newman Orebody
XXIX, Newman Rockshelter and PO959, Unpublished BA (hons)
thesis, University of Western Australia.
Edwards, D.A. and O’Connell, J.F. 1995. Broad spectrum diets in arid
Australia, Antiquity 69:769-83.
Edwards, K. and Murphy, A. 2003. A preliminary report on
archaeological investigations at Malea rockshelter, Pilbara region,
Western Australia, Australian Archaeology 56:44-46.
Hiscock, P.1988. Prehistoric Settlement Patterns and Artefact
Manufacture at Lawn Hill, Northwest Queensland, Unpublished
PhD thesis, University of Queensland: Brisbane.
Hughes, P.J. and Quartermaine, G. 1992. Investigations of Aboriginal
archaeological sites in the Mesa J development area, Pannawonica.
Unpublished report to the Western Australian Department of
Indigenous Affairs, Perth.
Lambeck, K. and Chappell, J. 2001. Sea level change through the last
glacial cycle. Science 292:679-86.
McDonald, Hales and Associates. 1997. Hope Downs Aboriginal
Heritage Investigations. Unpublished report to the Western
Australian Department of Indigenous Affairs, Perth.
Marwick, B. 2002. Milly’s Cave: Evidence for human occupation of
the inland Pilbara during the Last Glacial Maximum, Tempus 7:
21-33.
Marwick, B. 2005. Element concentrations and magnetic susceptibility
of anthrosols: indicators of prehistoric human occupation in the
inland Pilbara, Western Australia, Journal of Archaeological
Science 32:1357-68.
Maynard, L. 1980. A Pleistocene date from an occupation deposit in
the Pilbara Region, Western Australia, Australian Archaeology
10:3-8.
O’Connor, S. and Veth, P. 2006. Revisiting the Past: Changing
Interpretations of Pleistocene Settlement Subsistence and
Demography in Northern Australia, in Lilley, I. (ed.) Archaeology
of Oceania,Blackwell Publishing: Victoria, pp 31-48.
Slack, M.J. 2007. Between the Desert and the Gulf: evolutionary
anthropology and Aboriginal Prehistory in the Riversleigh/Lawn
Hill region, Northern Australia. Unpublished PhD thesis,
University of Sydney.
Slack, M.J. 2008. Report of Archaeological Survey and Excavation at
the Proposed Brockman 4 Syncline Project, Pilbara Region,
Western Australia. Unpublished Report for Pilbara Iron Pty Ltd.
Smith, M.A. 1987. Pleistocene Occupation in arid central Australia,
Nature 328:710-11.
Smith, M.A. 1989. The case for a resident human population in the
central Australian ranges during full glacial aridity, Archaeology in
Oceania, 24:93-105.
Thomas, M.F. 2008. Understanding the impacts of Late Quaternary
climate change in tropical and sub-tropical regions,
Geomorphology 101:146-158.
Troilett, G. 1982. Report on Ethel Gorge Salvage Project, Unpublished
report to Mt Newman Mining Co. Perth, Department of Aboriginal
Studies, Western Australian Museum.
Turney,C.S.M., Haberle, S., Fink, D., Kershaw, A.P., Barbetti, M.,
Barrows, T.T., Black, M., Cohen, T.J., Corre’ge, T., Hesse, P.P.,
Hua, Q., Johnston, R., Morgan, V., Moss, P., Nanson, G., Van
Ommen, T., Rule, S., Williams, N.J., Zhao, J.X., D’Costa, D., Feng,
Y.X., Gagan, M., Mooney, S., and Xia, Q. 2006. Integration of ice-
core, marine and terrestrial records for the Australian Last Glacial
Maximum and Termination: a contribution from the OZ
INTIMATE group, Journal of Quaternary Science, 21:751-761.
Veth, P. 1989. Islands in the Interior: A Model for the Colonisation of
Australia’s Arid Zone. Archaeology in Oceania, 24:81-92.
Veth, P. 1993. Islands in the Interior: The Dynamics of Prehistoric
Adaptations Within the Arid Zone of Australia.International
Monographs in Prehistory, Archaeological Series 3. Michigan: Ann
Arbor.
Veth, P. 1995. Marginal returns and fringe benefits: characterizing the
prehistory of the lowland deserts of Australia, Australian
Archaeology 40:32-38.
Veth, P. 2005. Cycles of Aridity and Human Mobility: Risk
Minimisation Among Late Pleistocene Foragers of the Western
Desert, Australia. in P.Veth, M.A. Smith, P. Hiscock (eds) Desert
Peoples: archaeological perspectives,Melbourne: Blackwell,
pp. 100-115.
Veitch, B. and Hook, F.2005. Anote on radiocarbon dates from the
Paraburdoo, Mount Brockman and Yandicoogina areas of the
Hamersley Plateau, Pilbara, Western Australia, Australian
Archaeology, 60:58-61.
Yokoyama, Y
., K. Lambeck, P.De Decker, P. Johnston and L.K.
Fifield. 2000. Timing of the Last Glacial Maximum from observed
sea-level minima, Nature 406:713-6.
39