Stone artefact scatters dominate the archaeological landscape of the inland Pilbara. While the archaeological record from rockshelter sites suggests human occupation consisting of brief. intermittent visits by small groups of people, artefact scatters which range from small, discrete single flaking events of perhaps 5-10 artefacts to widespread and varied scatters, sometimes of hundreds of thousands of stone artefacts extending over hundreds of thousands of square metres, clearly tell a different story. This paper presents an analysis of stone artefact assemblages from nine inland Pilbara surface artefact scatters and demonstrates that sites of this type have the potential to contribute much to our knowledge of human occupation of the region. We propose a testable model of Aboriginal occupation of the Pilbara during the Holocene.
Keywords: Pilbara, artefact scatters, surface archaeology
The Pilbara region of Western Australia has been subject to some of the most intensive archaeological surveying of any part of Australia. At least 10,000 sites are now listed on Western Australia's Register of Aboriginal Sites for the region and most of these have been recorded as a result of heritage surveys undertaken in response to mining and other developments. But what do we know about prehistoric Indigenous occupation of this landscape? Published site descriptions are few and far between; most are lost deep in the grey literature of consulting reports.
In a region abounding in rockshelters, archaeology in the Pilbara has been guided by questions of antiquity. Research has focused on identifying the timing of initial occupation and the response of Aboriginal hunter-gatherers to the last glacial maximum (Comtesse 2003, Edwards and Murphy 2003, Marwick 2002, Maynard 1980; Troilett 1982; Veitch et al. 2005, Veth 1993). Pleistocene layers in excavated rockshelters typically contain only small amounts of faunal material, charcoal and other organics and are characterised by small assemblages of stone tools (Maynard 1980:4-7, Troilett 1982, Brown 1987:24-33, Hughes and Quartermaine 1992). The general consensus is that occupation of the Pleistocene Pilbara consisted of brief and intermittent visits by small groups, reflecting a mode of adaptation based on a high level of residential mobility (Brown 1987:53, Marwick 2002:14).
Archaeological evidence from Holocene rockshelter sites appears to follow a similar pattern. Although organic material is more common in Holocene layers than in those of Pleistocene age, it is typically scarce (Hughes and Quartermaine 1992:102, Maynard 1980:5, Edwards and Murphy 2003, Veitch et al. 2005; Fiona Hook pets. comm. June 2006). Fragments of burnt and unburnt mammal and macropod bone dominate sparse faunal assemblages. Infrequently, transported baler shell and ochre is found. Plant remains include occasional plant fibres, resin and wooden fragments (Comtesse 2003, Marwick 2002, Ryan and Morse 2005, Veitch et al. 2005). Backed artefacts and other small tools, rock art and seed grinding technology also appear in the archaeological record at this time (Marwick 2002; Comtesse 2003). Despite this momentary excitement, the Holocene rockshelter evidence, like that of the Pleistocene, suggests an archaeological landscape of sites occupied for archaeologically brief moments by small groups of people.
But is this picture consistent with other archaeological and even ethnographic evidence of Aboriginal occupation of the inland Pilbara? Ethnographic and anecdotal accounts of Pilbara Aboriginal people (e.g. Brown 1987, Brehaut and Vitenbergs 2001) clearly show that rockshelters were not the focus of human occupation. Rockshelters and caves were used only during rain or dust storms and were sometimes used to cache wooden artefacts and other materials. Sandy creeks and their banks were the preferred location for large-scale camp sites with long term seasonal camps located where reliable sources of water could be easily accessed. In this context it is no surprise that it is open surface scatters of stone artefacts that dominate the archaeology of the Pilbara--perhaps it is not Aboriginal people who enjoy rockshelters but archaeologists (cf. Bowdler 1975, Smith 1993)!
As documented in endless heritage reports, open sites in the Pilbara typically range from small, discrete single flaking events of perhaps 5-10 artefacts to widespread and varied scatters, sometimes of hundreds of thousands of stone artefacts extending over hundreds of thousands of square metres. This is a very different archaeological signature to that provided by excavated rockshelters in the region, and is, we suggest, far more likely to include elements of material culture that are the product of everyday life of Aboriginal Australians in the past than are present in rockshelter deposits.
We argue here that the focus on rockshelters has skewed our interpretation of the Pilbara archaeological landscape. While rockshelters provide glimpses of a small component of human occupation of the region, a much wider view of the archaeological signature of hunter gatherer occupation in the inland Pilbara during Holocene times can be obtained by an analysis of surface artefact scatters. This paper presents an analysis of stone artefact assemblages from nine inland Pilbara surface artefact scatters and demonstrates that sites of this type have the potential to contribute much to our knowledge of human occupation of the region. We propose a testable model of Aboriginal occupation of the Pilbara during the Holocene.
The landscape of the inland Pilbara is very different to other parts of the Australian arid zone. Drained by a series of extensive, seasonally flooding river systems which cut through steep ironstone ranges, its topography and hydrology alone mean that patterns of life in the Pilbara have been different to those of the more arid and less diverse desert regions nearby. Dominated by the Hamersley plateau, the largest elevated area of land in Western Australia, lowland plains on its northern, western and eastern flanks are bounded by abrupt and prominent escarpments deeply incised by spectacular gorges (Beard 1975: 10-11). Geography and topography exert powerful influences over the distribution of that most important resource for hunter gathers in arid environments--water. Both the Gascoyne-Murchison area to the south and the Great Sandy Desert to the north and east lack the inland Pilbara's plateau and escarpment geography and its associated high volume of surface water (Dawe and Dunlop 1983). Climatic conditions of the Pilbara are dominated by tropical cyclones to a far greater extent than anywhere else in Western Australia (Beard 1975:10-11). Rainfall in the inland Pilbara is higher than the surrounding areas but can vary dramatically with yearly averages sometimes met in the space of a few days (Johnson and Wright 2001).
The analysis outlined here focuses on stone assemblages from nine open surface artefact scatters in the inland Pilbara (Table 1, Figure 1). Eight of these, the Gull to Tunkawanna Creek sites (W3-03.W3-04, W3-06, W3-07, W3-08, W3-10, W3-11 and W3-12), were recorded and salvaged by archaeological consultants Gavin Jackson Pty Lid and representatives of the Ngarluma Yindjibarndi native title group prior to construction of the then proposed Pilbara Iron railway (Jackson and Fry 2001). Following the heritage brief provided, artefacts that were within the 300 m wide disturbance corridor associated with the railway were collected from each site. This meant in most cases that only a small percentage of each site was actually salvaged (Table 2). The one exception to this was site W3-08 where the entire site was within the rail corridor and adjacent borrow pit. Sites were salvaged by imposing a grid of 10 m x 10 m squares over the area of each site and collecting all surface material in every eighth square, providing a sample of 12.5% of the area of the sites within the rail corridor. A total of 2945 artefacts collected from these sites were analysed (Ryan et al. 2005). The ninth site, DE-SAS 1-4, is located in Pilbara Iron's West Angelas project area some 150 km south east of the Gull-Tunkawanna sites (Table 1, Figure 1). Originally recorded as four discrete artefact scatters, the archaeological consultants later concluded that the small sites DE-SAS 2, 3 and 4 were most likely components of the much larger artefact scatter DE-SAS 1 (Wood and Westall 2003:17). DE-SAS 1-4 was salvaged by Pilbara Iron archaeologists and Gobawarrah Minduarra Yinhawanga native title group representatives using a series of 10 x 10 m squares placed at 100 m intervals across the site. A total of 142 artefacts, providing a sample of just less than 1% of the site, were salvaged and subsequently analysed (Ryan et al. 2006a).
[FIGURE 1 OMITTED]
It is acknowledged here that there are problems in the level of sampling undertaken at a number of these sites. In the absence of additional material, however, all systematically salvaged artefacts from each site have been included in this analysis.
All the Gull to Tunkawanna artefact scatters are located around ephemeral or seasonal creeks (W3-06. W3-07, W3-08, W3-10, W3-11 and W3-12) and gilgais (W3-03 and W3-04). Four (W3-03, W3-04, W3-06 and W3-07) are located on the plains of the Chichester Plateau and the remaining four (W3-08, W3-10, W3-11 and W3-12) are located in the undulating Fortescue Valley. DE-SAS 1-4 is located at the eastern end of the Hamersley Plateau and extends for over one kilometre on both sides of a deeply incised seasonal creek line. Summary characteristics of each of the nine sites are presented in Table 1.
Chronological control of open surface artefact scatters is inherently problematic. Furthermore. the iron-clad environment of the Pilbara does not favour the development of stratified surface sites. A metal match box, several glass artefacts, a single backed artefact and a small number of blades provide the only chronological markers at the sites discussed here and suggest that all represent late Holocene occupation of the inland Pilbara and use or presumably reuse of some of these sites during historical times.
Stone artefacts: Method and Methodology
In the following analysis, nine inland Pilbara open surface artefact scatters are characterised by comparing each stone assemblage, focusing on variation in the technology of artefact manufacture and assemblage diversity. We argue that stone tool assemblages, as a record of human behaviour, can potentially provide a wide range of information about the people that produce them (Flenniken and White 1985:131, Hiscock 1998:265, Barton 2003:32-33, Clarkson and O'Connor 2006:160). Our aim in this paper is to explore how artefact assemblages could be linked to broader patterns of human behaviour beyond simply identifying patterns within the technology of stone artefact manufacturing itself.
Modern approaches to stone tool analysis take a broad range of factors into account in explaining variation between and within lithic assemblages. They generally focus on the technology of stone tool manufacture, defined here as a tradition of knowledge made up of a series of stone working techniques conducted in a dynamic sequence. This "reduction sequence" is typically conceptualised as consisting of four main stages: procurement of raw material, flake production, secondary working (retouch) and artefact use and discard (Flenniken and White 1985:131, Holdaway and Stern 2004:206-208). Variation in this sequence can be seen in differences in the techniques employed to flake stone and in the way the sequence is organised. While lithic analysts remain uncertain about which variables provide meaningful information about the reduction sequence, the most robust claims about this sequence can be made when several variables show the same directional change (Odell 2000:291, Pelcin 1997, Shott et al. 2000, Dibble and Pelcin 1995). Data on metrics, dorsal flake scars, core rotations, cortex, platform faceting and the extent of retouch can indicate the stage or stages of reduction represented in an artefact assemblage (Amick et al. 1988, Ammerman and Andrefsky 1982, Magne and Pokotylo 1981, Odell 1989, Shott 1994). Following Holdaway and Stern (2004:107211), a standard range of metrical and technological attributes of all flaked stone was recorded. All metric measurements were recorded in millimetres to one decimal place using vernier callipers and entered on to a Microsoft Access Database.
In addition to identifying differences in stone artefact manufacture at sites in the inland Pilbara we have also focused on variation in assemblage diversity. We have explored this by comparing differences in the proportion of the following classes of artefacts: retouched flaked stone, unmodified flaked stone, cores and grinding implements. A statistical measure of diversity that can be applied to archaeological data is the Shannon-Weaver information statistic H (Shott 1989:286). This statistic expresses the probability that any randomly selected member of a population will fall into one of a group of imposed classes (Shott 1989:286). The minimum value of H, 0, occurs when all members of a population fall into one class. The maximum value. 1, occurs when all members of the population are evenly distributed, although the statistic also takes account of the number of classes present in the population. As such. it is a measure of both richness and "equitability of elements' of members of a population. This is not a measure of how significant differences in assemblage diversity are, but is an index of greater or lesser diversity. In the following analyses the H statistic has also been used to investigate raw material diversity, with classes determined by the raw materials present at each site.
Short (1986:15, 1989:283) discusses the influence of residential mobility on assemblage diversity in hunter gatherer archaeology. Shott's (1986:22-34) analysis of mobility and assemblage diversity in fourteen ethnographically observed hunter-gatherer groups suggests that there is an inverse relationship between assemblage diversity and the number of residential moves a group makes per year. It follows that the diversity of an assemblage is greatest where people have lived longest (Shott 1989:297). Given that raw material diversity appears to co-vary with assemblage diversity, we propose that the number of raw materials present within an assemblage also increase with the amount of time spent at a particular location. This is discussed further below.
The reduction sequence at all nine sites is broadly similar and is, we suggest, characteristic of stone artefact manufacture in the Pilbara during the Holocene. There is a clear pattern of largely local extraction of raw materials with assemblages from those sites on the Chichester Plain (W3-03, 04, 06 and 07) dominated by rhyolite while those located in the Fortescue Valley (W3-08, 10, 11 and 12) are dominated by chert (Table 3). Hard hammer percussion is by far the most common flake removal technique. Single platform cores are the most abundant core type at all sites (Table 4). Platform preparation is minimal with flaked and faceted flake platforms uncommon and overhang removal absent from most of the assemblages (Table 5). Flakes from the early stages of the reduction sequences, i.e. those with high proportions of dorsal cortex, are largely absent (Table 6). Most flakes derive from cores with either one or two dorsal scars that have been rotated only once, probably in order to remove uneven cortex and create relatively flat striking platforms. Flakes from the late stages of the sequence, indicated by high levels of platform faceting and high numbers of dorsal scars and core rotations, are also absent. This pattern suggests that the assemblages at these sites generally represent the middle stages of core reduction. It appears from this and other recent work (Ryan et al. 2005, Ryan et al. 2006b) that Pilbara reduction sequences rarely extended to the late stages, i.e. high levels of platform preparation and core rotation.
There are some clear patterns of variation from this general sequence of flake production at the nine sites. The most obvious differences between most of the assemblages relate to the dominant raw material type, with the structure of the assemblages strongly informed by local geology. In particular, retouched artefacts are made from the most abundant raw material at each site. Chert, however, appears to have been generally preferred as it is the only raw material from which retouched artefacts were made at seven of the eight Gull to Tunkawanna sites. Indeed, retouch is generally more common at the chert-dominated sites than at the rhyolite-dominated sites (Figure 1). There are also subtle technological differences related to the most common raw material with sites dominated by chert generally characterised by a greater level of platform preparation than those dominated by rhyolite (Table 5). Interestingly, the assemblage at DE-SAS1-4 differs from the Gull to Tunkawanna sites in having a more diverse range of raw materials than found at those sites (Table 3).
Some of the observed variation between these assemblages cannot be accounted for by raw material differences alone. The complete flakes at sites W3-06, W3-08, W3-11 and DE-SAS1-4 are distinctive in having higher levels of platform preparation than the other sites (Table 5). Assemblages at these four sites also have the highest proportion of retouched artefacts (Figure 2) and cores (Figure 3), particularly DE-SAS 1-4. All four of these sites also differ from the others in having more diverse assemblages. These differences indicate that there is a wider range of the reduction sequence present at these four sites than is found at the other sites.
[FIGURE 2 OMITTED]
[FIGURE 3 OMITTED]
Pilbara surface artefact scatters: An Interpretation
So what do the assemblages at these sites suggest about occupation of the Pilbara during the late Holocene? In particular, do they indicate use by only a small and ephemeral population, spending perhaps just a day or even hours at one place, as suggested by the rockshelter sites in the region?
The archaeological research of Veth (1987, 1989, 1993, 2005) is an important point of reference for any interpretation of archaeological sites in the arid zone of Australia and has implications for the interpretation of surface artefact scatters in the Pilbara. Veth's research concentrated on the occupation of the Great Sandy Desert in Western Australia, an area immediately to the north and east of the Pilbara. In simple terms, Veth argued that occupation of the arid interior could be described in terms of a dichotomy between large intensive camps based around permanent springs and soaks and short-term camps based around more ephemeral sources of water (Veth 1987:104-5, 1989:198-226). The stone assemblages at more permanent water sources are accordingly larger and are characterised by a wider diversity of artefact types and raw materials while those at ephemeral sources are smaller and less varied. This model has been widely applied to other arid regions of Australia including the Pilbara and is a mainstay for interpretation of site use in many consulting reports. The results of our analysis suggest, however, that more varied patterns of human behaviour can be identified from surface artefact scatters than is indicated by Veth's model. We have explored this by linking stone artefact scatters to other economic resources (cf. Barton 2003).
The Gull to Tunkawanna sites differ slightly from the pattern predicted in Veth's model. There are no permanent soaks or springs present at the Gull to Tunkawanna sites and the sites that appear to be the largest in surface area, W3-03 and W3-04, are both located on rhyolite plains adjacent to several large gilgais. Gilgais form in clay soils that have very high water absorption. This causes them to swell when wet and shrink on drying, leading to the formation of holes and cracks - colloquially known as 'crabhole country' (Beard 1975:30). Gilgais have been described by traditional owners as 'fat country' and are known to abound with a diversity of both plant and animal resources over a period of several months following rain (P. Kendrick, Department of Conservation and Land Management, pers. comm.). As such, gilgais constitute a predictable 'resource patch' in an arid high-risk environment (cf. Barton 2003:32). It is evident from the stone assemblages at the gilgai sites described here that cores were roughly prepared from nearby sources of stone (at these sites, mainly rhyolite) and bought to these locations to provide large flakes for use during everyday tasks. Retouch is not common at these two sites, contributing only between <1% and 2.5% of the assemblages. It should be noted that the assemblages at these sites are similar to that described by Veth (1987:105) for claypan sites in the Great Sandy Desert, although gilgais differ from claypans in responding much more quickly to rain.
We propose that the extent of the gilgai sites W3-03 and W3-04 suggests that they were occupied by large groups. The structure of the assemblages, however, with their small range of raw materials and, particularly, their low level of diversity suggests that the sites were not occupied for very long periods of time. Furthermore, we propose that the large scale but short term nature of occupation at sites W3-03 and W3-04 relates to the exploitation of the short term or seasonal abundance of plant foods associated with the gilgais, in particular the bush onion Cyperus bulbosus. Plant foods like this are abundant in temporarily wet or inundated areas such as gilgais following rain (P. Kendrick, pers. comm., Bindon 1996) and would have been able to sustain a large group of people for a short period of time. Bush onion is well documented as one of the most important arid zone food sources (Bindon 1996). It is collected as small tubers which are easily harvested by hand and has a nutty flavour when roasted (Bindon 1996). Significantly it is very rich in carbohydrate, providing 62.5 g per 100 g. By way of comparison, the seeds of another arid zone staple, Acacia aneura, provide only 25.5 g of carbohydrate per 100 g (Peterson 1976:28). In this context it is interesting to note Thorley's (2001) model of site use in the Palmer River area in central Australia in which he links large scale ephemeral water sources and ceremonial activity. Furthermore it is perhaps significant that bush onion is depicted in recent artworks as a plant food strongly associated with women's ceremonies (http://www.aboriginalartistsofaustralia.com.au).
A similar pattern of exploitation has been suggested for other plant species, most notably cycads (Macrozamia spp., Beaton 1982, Smith 1982). Beaton (1982:57) refers to cycads as 'communion foods' and argues that they were used to sustain large gatherings of people for short periods of time especially during ceremonial occasions. While cycads required sophisticated processing and/or storage to detoxify them, it is the occurrence of a short term abundance of a specific plant food that is of relevance here. In this context, and in view of the patterns evident in the stone assemblage at sites located adjacent to gilgais, we suggest that occupation of these sites at a time of predictable localised resource abundance may have allowed the congregation of large groups of people for short periods of time in a way analogous to that made possible by the collection and preparation of plants foods such as cycads.
In contrast, we suggest that sites with a wider range of the reduction sequence, W3-06, W3-08, W3-11 and DE-SAS 1-4 represent longer term camps utilised by smaller groups of people, probably family or hearth groups - that is, groups of the same size as those last seen occupying some of the many Pilbara rockshelter sites. Sites W3-06, W3-08 and W3-11 provided access to sources of ephemeral water and snakewood (Acacia xiphophylla), a species with edible seeds and gum that was favoured for firewood and for the manufacture of wooden tools, including boomerangs (Kuruma elder, pers. comm.). The relatively high proportion of retouched artefacts at these sites is presumably related to tasks associated with the manufacture and maintenance of artefacts made from this wood. As such, it is possible that the assemblages reflect specialised site activities rather than, or in addition to, the duration of stay.
Although similar to W3-06, W3-08 and W3-11, the assemblage at DE-SAS 1-4 differs in featuring a very diverse range of raw materials and artefact types. This site differs from the gilgai sites (W3-03 and W3-04) discussed earlier in featuring a much wider range of the reduction sequence. The site is located on a deeply incised seasonal water course that collects drainage from the surrounding low hills and plains and is the focus of drainage in the immediate area. We suggest therefore, that while the gilgai sites were utilised specifically for short periods of time and by large groups of people exploiting a small range of plants that flourish following rains, no single resource would have been found in abundance at DE-SAS 1-4. The high diversity of both raw materials and tool types at DE-SAS 1-4 suggests perhaps that people based themselves here during the wet season and moved out to collect sources of stone and food which were then brought back to the site, somewhat similar to the 'collector' strategy proposed by Binford (1980:4-20). That is, the residential mobilio' of the group was restricted at this time of the year while the logistical mobility of the group was high.
Figures 4 and 5 illustrate the relationship between both assemblage diversity and raw material diversity at all sites on the one hand and the proportion of cores and retouched artefacts on the other. We suggest, following Shott (1989), that there is clearly a continuum of variation in assemblage and raw material diversity in the sites discussed here, with W3-04 and DE-SAS1-4 representing the extreme ends of a spectrum of mobility. The trendline shown in Figure 4 represents this continuum of mobility: sites on the left hand side of the diagram represent those sites at which residential mobility was highest while those on the right hand side represent those sites at which residential mobility was lowest. Sites between these groups of sites are presumably transitional. Movement away from the trendlines in both figures is interpreted here to represent sites at which specific activities have taken place. In other words, we propose that sites furthest from the trendline on the right side of Figure 5 represent sites at which restricted mobility is associated with a particular task. For example, the particularly high proportion of cores, but comparatively low proportion of retouched artefacts at W3-06 might represent a site associated with core reduction rather than processing resources other than stone. On the other hand, W3-08, which has a high proportion of retouched artefacts, but a lower proportion of cores than W3-06 and DE-SAS1-4, might represent a site associated with a task associated with retouch, in this context, the manufacture and maintenance of wooden artefacts, presumably made of snakewood.
[FIGURE 4 OMITTED]
[FIGURE 5 OMITTED]
DE-SAS1-4 conforms closely to the proposed mobility trendline. The site features a higher proportion of cores than W3-06 and less retouched artefacts than W3-08. The fact that this site has a much higher range of raw materials than W3-06 and W3-08 suggests that occupants have introduced tool stone from a variety of contexts, rather than processing locally available raw materials, a pattern that is characteristic of the Gull to Tunkawanna sites.
Rockshelters and open artefact scatters clearly differ in the insights they may provide into the life of Aboriginal people in the Pilbara. Pilbara rockshelters, as we have seen, appear to skew our interpretation of the archaeology of this landscape as one characterised primarily by intermittently occupied sites with sparse assemblages of archaeological material suggesting a small, highly mobile population. We know, both from ethnohistoric and ethnographic accounts that, at least in the late Holocene, this is not the case (Brown 1987, Brehaut and Vitenbergs 2001:28).
The model proposed here identifies a series of stone assemblage characteristics that, in tandem with other relevant environmental and social information, has been used to suggest how some surface sites may have been used. We suggest that this approach to comparing stone artefact assemblages might act as a useful heuristic device in exploring variation between lithic assemblages in the inland Pilbara. Regional studies in Australia too often pigeonhole settlement systems into either residential or logistical types and the causal mechanisms at work are regularly simplified (cf. Guilfoyle 2005). Recent analyses of other Pilbara surface assemblages (Ryan et al. 2006a, b and c, Ryan et al. 2008, Ryan and Carson 2006) appear to support the fundamental basis of our model, while indicating the presence of a range of other site types, for example, blade production quarries, specialist wood working sites and sites where groups appear to have gathered to escape cyclonic flooding.
It is clear that potentially much more can be learned about Aboriginal people and their occupation of the inland Pilbara in the past by systematic recording and comprehensive analysis of surface artefact scatters. There is great potential and need for consulting archaeologists to contribute here. It is essential that the information collected by consultants is sufficient in terms of sample size and artefact and site detail to provide archaeologically meaningful data that can be used to test models such as the one presented here. Pilbara archaeology needs to improve the resolution of our understanding of surface sites and continue to explore variation in artefact scatters as a reflection of Aboriginal settlement patterns.
We thank Jane Balme, Sandra Bowdler and Annie Carson for reading earlier drafts of this paper, and Peter White for his comments on an earlier draft. We acknowledge the work of Ben Marwick in the initial analysis of the Gull to Tunkawanna materials. We also acknowledge Gavin Jackson and members of the Ngarluma Yindjibarndi native title consulting group, staff of Pilbara Irons's Aboriginal Training and Liaison unit and representatives of the Gobawarrah Minduarra Yinhawanga native title consulting group for enabling us to undertake the analysis of archaeological material discussed here. Figure 1 was drawn by Annie Carson, Ryan Coughlan and Stafford Smith.
Amick, D.S., Maulden, R.P. and S.A. Tomka 1988 An evaluation of debitage produced by experimental biracial core reduction of a Georgetown chert nodule. Lithic Technology 17:26-36.
Ammerman, A.J. and W. Andrefsky 1982 Reduction sequences and the exchange of obsidian in Neolithic Calabria, in J. Ericson and T. Earle (eds) Contexts for Prehistoric Exchange. Academic Press: New York, pp. 149-72.
Barton, H. 2003 The thin film of human action: interpretations of arid zone archaeology. Australian Archaeology 57:32-41.
Beard J. 1975 Pilbara. Explanatory notes to Sheet S, 1:1 000 000 Series, Vegetation Survey of Western Australia. University of Western Australia Press: Perth.
Beaton, J.M. 1982 Fire and water: Aspects of Australian management of Cycads. Archaeology in Oceania 17:51-58
Binford, L. 1980 Willow Smoke and dogs' tails: hunter-gatherer settlement systems and archaeological site formation. American Antiquity 45:4-20.
Bindon, P. 1996 Useful bush plants. Western Australian Museum: Perth.
Bowdler, S. 1975 Caves and Aboriginal Man. Australian Natural History 18(6):216-219.
Brehaut, L. and A. Vitenbergs (eds) 2002 The Guruma Story. IAD Press: Alice Springs.
Brown, S. 1987 Toward a Prehistory of the Hamersley Plateau, Northwest Australia. Occasional Papers in Prehistory 6, Department of Prehistory, Research School of Pacific Studies, Australian National University, Canberra.
Clarkson, C. and S. O'Connor 2006 An Introduction to Stone Artefact Analysis. In Peterson, A. and J. Balme (eds) Archaeology in Practice." A Student Guide to Archaeological Analyses. Blackwell Publishing Ltd: Boston, pp. 159-206.
Comtesse, S. 2003 Mt. Newman Sites Re-analysed: Newman Orebody XXIX, Newman Rockshelter and P0959. Unpublished Bachelor of Science (Honours) thesis in Archaeology, University of Western Australia.
Dawe, C. and J. Dunlop. 1983 Introduction to Hamersley Range National Park in B. Muir (ed) A fauna survey of the Hamersley Range National Park, Western Australia: Bulletin No. 1. National Parks Authority of Western Australia: Perth, pp. 3-6.
Dibble, H.L. and A. Pelcin 1995 The effect of hammer mass and velocity on flake mass. Journal of Archaeological Science 22: 429-39.
Edwards, K. and A. Murphy 2003 A preliminary report on archaeological investigations at Malea Rockshelter, Pilbara region, Western Australia. Australian Archaeology 56:44-46.
Flenniken, J.J. and J.P. White 1985 Australian flaked stone tools: a technological perspective. Records of the Australian Museum 36:131-151
Guilfoyle, D.R. 2005 Socializing stone artefact assemblages: Regionalization and raw material availability. Australian Archaeology 60:34-40.
Hiscock, R 1998 Revitalising artefact analysis. In T. Murray (ed) Archaeology of Aboriginal Australia. Unwin and Allen: Sydney, pp. 257-265.
Holdaway, S. and N. Stern 2004 Written in Stone: Methods and Theories for Analysing Australian Flaked Stone Artefacts. Museum Victoria: Melbourne.
Hughes, P.J. and G. Quartermaine 1992 Investigations of Aboriginal Sites in the Mesa J Development Area, Pannawonica. Unpublished report held by the Western Australian Department of Indigenous Affairs, Perth.
Jackson, G. and R. Fry 2001 A report of an Aboriginal archaeological assessment of the proposed Tom Price to Yandi Mine 220 kV power line route, Pilbara, Western Australia (stage 1 : Tom Price to Turee Creek and West Angelas spur sections). Unpublished Report held by the Western Australian Department of Indigenous Affairs, Perth.
Johnson, S.L. and A.H. Wright 2001 Central Pilbara Groundwater Study. Water and Rivers Commission, Hydrogeological Record Series HG 8.
Magne M.P. and D. Pokotylo 1981 A pilot study in bifacial lithic reduction sequences. Lithic Technology 10:34-47.
Marwick, B. 2002 Inland Pilbara archaeology: A study of variation in Aboriginal occupation over time and space on the Hamersley Plateau. Unpublished Master of Arts thesis, University of Western Australia, Perth.
Maynard, L. 1980 A Pleistocene date from an occupation deposit in the Pilbara region, Western Australia. Australian Archaeology 10:3-8.
Odell, G.H. 1989 Experiments in lithic reduction. In D.S. Amick and R.P. Maudlin (eds) Experiments in Lithic Technology. BAR International Series 528: Oxford, pp. 163-98.
Odell, G.H. 2000 Stone tool research at the end of the millennium: Procurement and technology. Journal of Archaeological Research 8:269-331.
Pelcin, A.W. 1997 The formation of flakes: The role of platform thickness and exterior platform angle in the production of flake initiations and terminations. Journal of Archaeological Science 24:1107-113.
Ryan, I. and A. Carson 2006 Report on an analysis of stone artefacts salvaged from two sites located along the proposed Robe Rail Duplication, Tunkawanna to Cockatoo Sections, inland Pilbara, Western Australia, for Pilbara Iron. Unpublished Report prepared by Eureka Archaeological Research and Consulting, University of Western Australia.
Ryan, I., Carson, A. and Morse, K. 2006a An analysis of salvaged artefacts from five artefact scatters at Deposit E, West Angelas, Pilbara, Western Australia. Unpublished report prepared for Pilbara Iron by Eureka Archaeological Research and Consulting, University of Western Australia.
Ryan, I., Carson, A. and K. Morse 2006b Report on the Results of an Archaeological Survey of Nominated Areas within Mining and Exploration Tenements M52/776 and E52/1455, 5 Mile Creek, Ashburton Region, Western Australia for Abra Mining Limited. Unpublished Report prepared by Eureka Archaeological Research and Consulting, University of Western Australia.
Ryan, I., Carson, A. and K. Morse 2006c Cane River AS05-01: Report on an Aboriginal Archaeological Site on Mt Smart Station in the inland Pilbara, Western Australia for Pilbara Native Title Service. Unpublished Report prepared by Eureka Archaeological Research and Consulting, University of Western Australia.
Ryan, I., Carson, A., Brown, V. and K. Morse 2008 An Analysis of Stone Artefacts Salvaged from three stone artefact scatters located on Mesa A, near Pannawonica, inland Pilbara, Western Australia, for Pilbara Iron. Unpublished Report prepared by Eureka Archaeological Research and Consulting, University of Western Australia.
Ryan, I. and K. Morse 2005 In Transit: An Analysis of Archaeological Material Excavated From Rockshelter TP00-04, Tom Price, Western Australia, for Pilbara Iron. Unpublished Report prepared by Eureka Archaeological Research and Consulting, University of Western Australia.
Ryan, I., Marwick, B., Balme, J., Morse, K. and A. Carson 2005 Gull --Tunkawanna: a report on the analysis of stone artefact scatters salvaged during the proposed West Angelas rail project, Pilbara, Western Australia. Unpublished Report prepared for Gavin Jackson Pty Ltd by Eureka Archaeological Research and Consulting, University of Western Australia.
Shott, M.J. 1986 Technological organization and settlement mobility, an ethnographic examination. Journal of Anthropological Research 42:15-51.
Shott, M.J. 1989 Diversity, organization, and behaviour in the material record. Ethnographic and archaeological examples. Current Anthropology 30:283-315.
Shott, M.J., Bradbury, A.P., Carr, P.J. and G.H. Odell 2000 Flake size from platform attributes: Predictive and empirical approaches. Journal of Archaeological Science 27:877-94.
Smith, M. 1982 Late Pleistocene Zamia exploitation in southwestern Western Australia. Archaeology in Oceania 17:117-121.
Smith, M.A. 1993 Biogeography, human ecology and prehistory in the sand ridge deserts, Australian Archaeology 37:35-50.
Thorley, P. 2001 Uncertain supplies: water availability and regional archaeological structure in the Palmer River catchment, central Australia. Archaeology in Oceania 36:1-14.
Troilett, G. 1982 Report on Ethel Gorge salvage project. Unpublished Report held by Western Australian Department of Indigenous Affairs.
Veitch, B., Hook, E and E. Bradshaw 2005 A Note on Radiocarbon Dates from the Paraburdoo, Mount Brockman and Yandicoogina areas of the Hamersley Plateau, Pilbara, Western Australia. Australian Archaeology 60:58-61.
Veth, P. 1987 Martujarra Prehistory: variation in arid zone adaptations. Australian Archaeology 25:102-11.
Veth, P. 1989 The prehistory of the sandy deserts: spatial and temporal variation in settlement and subsistence behaviour within the arid zone of Australia. Unpublished PhD thesis, University of Western Australia, Perth.
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. 2005 Cycles of Aridity and Human Mobility: Risk Minimization among Late Pleistocene Foragers of the Western Desert, Australia. In Veth, P., M. Smith and P. Hiscock (eds), Desert Peoples: Archaeological Perspectives. Blackwell Publishing: Carlton, pp. 100-115.
Wood, V. and C. Westall 2003 An archaeological survey of the West Angelas Project Deposits E and F, Hamersley Ranges, Western Australia. Unpublished report prepared for Hamersley Iron and Robe River Mining, by Vivienne Wood Heritage Consultant Pty Ltd. April 2003.
Eureka Archaeological Research and Consulting, University of Western Australia, Crawley, Perth, WA 6009. firstname.lastname@example.org
Table 1. Summary characteristics Gull-Tunkawanna sites (after Jackson
and Fry 2001) and site DE-SAS 1-4.
W3-03 On flat featureless rhyolite plain, spinifex,
gasses, east of existing HI rail line, artefact
scatter focused around small gilgais
W3-04 On flat featureless rhyolite plain, spinifex.
Brasses, east of existing HI rail line, artefact
scatter focussed around small gilgais and one
very large gilgai
W3-06 South side of broad but shallow creek in
snakewood (Acacia sp.) grove
W3-07 South side of broad but shallow creek in
snakewood (Acacia sp.) grove, adjacent to low hill
W3-08 On alluvial plain on south side of creek.
Dense Spinifex and acacia
W3-10 South side of broad, well defined creek on
alluvial plain with eucalypts, acacia scrub.
spinifex and snakewood (Acacia sp.) grove
W3-11 On alluvial plain which forms the high northern
bank of substantial creek, low eucalypts, Acacia
scrub and spinifex
W3-12 Gravel covered alluvial 'island' between branches
of creek, eucalypts, acacia and spinifex
DE-SASI-4 In Acacia woodland on both sides of large
seasonal creek in shallow valley
Site Max dimensions Artefact density: Approx no.
N-S x E-W (m); min, max, mean of artefacts
estimated area ([m.sup.2])
W3-03 800 x 480; 310,000 0.04, 0.92, 0.24 >73,000
W3-04 450 x 230; 60.000 0.04, 0.6, 0.21 >13,000
W3-06 130 x 650; 39,000 0.08, 2.28, 0.37 >14,000
W3-07 300 x 280: 65,000 0.04, 0.6, 0.18 11,700
W3-08 180 x 300: 43,000 0.04, 0.24, 0.10 4,000
W3-10 180 x 200; 40.000 0.04, 11.16, 0.22 8.800
W3-11 210 x 70: 9,500 0.04, 1.24, 0.08 760
W3-12 250 x 290; 45,000 0.08, 0.68, 0.27 12,000
DE-SASI-4 1500 x 310: 464,750 0.01, 0.17, 0.14 17,700
Table 2. Summary details of salvage undertaken at the Gull
to Tunkawanna and DE-SASl-4 sites.
Site Estimated % of No. of 10 x 10 Total no.
total area total site [m.sup.2] of artefacts
([m.sup.1]) salvaged squares salvaged
W3-03 307,200 3.6 113 1037
W3-04 61,200 10.1 62 577
W3-06 39,000 4.6 18 237
W3-07 65,000 4.1 27 93
W3-08 43,000 13 56 115
W3-10 40,000 8 32 448
W3-11 9,500 17.8 17 94
W3-12 45,000 4.3 19 354
DE-SAS 1-4 450,000 < 1 45 142
Table 3. Assemblage of stone artefacts from the Gull-Tunkawanna sites
and DESAS 1-4: raw materials (%). BIF: Banded ironstone formation.
Raw Material 03 04 06 07 08 10 11
BIF 0 0 0 0 0 0 0
Chalcedony <1 0 1.2 11.7 3.4 3.3 0
Chert <1 1.2 6.3 5.9 87.9 90.4 60.6
Dolerite 0 0 <1 0 0 <1 0
Glass 2.9 0 6.3 0 0 <1 0
Granite 12.2 7.2 <1 0 1.7 <1 0
Ironstone 0 0 0 0 0 0 0
Metasediment 0 0 <1 0 0 0 0
Quartz 0 0 1.6 0 0 0 0
Quartzite 0 0 6.7 0 0 0 0
Sandstone 0 0 0 0 0 0 0
Sedimentary 1.8 0 2.1 0 5.2 0 37.2
Silcrete <1 0 <1 1 <1 0 0
Rhyolite 81.7 91.5 73.6 87.2 <1 5.1 2
Raw Material 12 DE-SAS 1-4
BIF 0 38
Chalcedony <1 4.2
Chert 74.3 32.4
Dolerite 0 0
Glass 0 0
Granite 0 0
Ironstone 0 14.8
Metasediment 0 0
Quartz 0 8.4
Quartzite 0 0
Sandstone 0 1.4
Sedimentary 22.6 0
Silcrete 1.9 0.7
Rhyolite <1 0
Table 4. Assemblage makeup of sites.
Artefact Type W3-03 W3-04 W3-06
n % n % n %
All Retouched 28 2.7 3 0.5 14 5.9
Complete 494 48.1 313 54.2 105 44.3
Fragments 98 9.5 6 1.04 24 10.1
Longitudinally 60 5.8 55 9.5 10 4.2
Transversely 170 16.5 106 18.3 25 10.5
Both Breaks 0 0 0 0 0 0
Single Platform 13 1.2 3 0.5 6 2.5
Multi Platform 3 0.2 0 0 5 2.1
Core Other 0 0 0 0 4 l.6
Angular 157 15.2 91 15.7 39 16.4
Grinding Stones 2 0.1 0 0 0 0
Pebble Tools 2 0.1 0 0 4 1.6
Hammerstones 0 0 0 0 1 0.4
TOTAL 1027 100 577 100 237 100
Artefact Type W3-07 W3-O8 W3-10
n % n % n %
All Retouched 4 4.3 17 14.7 19 4.2
Complete 61 65.5 42 36.5 196 43.7
Fragments 2 2.1 14 12.1 8 1.7
Longitudinally 0 0 11 9.5 8 1.7
Transversely 7 7.5 11 9.5 29 6.4
Both Breaks 0 0 0 0 0 0
Single Platform 2 2.1 3 2.6 3 0.6
Multi Platform 0 0 1 0.8 1 0.2
Core Other 0 0 0 0 0 0
Angular 16 17.2 14 12.1 184 41
Grinding Stones 1 1 2 1.7 0 0
Pebble Tools 0 0 0 0 0 0
Hammerstones 0 0 0 0 0 0
TOTAL 93 100 115 100 448 100
Artefact Type W3-11 W3-12 DE-SAS
n % n % n %
All Retouched 7 7.4 14 3.9 16 11.2
Complete 19 20.2 116 32.7 72 50.7
Fragments 5 5.3 21 5.9 18 12.6
Longitudinally 1 1 19 5.3 6 4.2
Transversely 9 9.5 49 13.8 7 4.9
Both Breaks 0 0 0 0 3 2.1
Single Platform 2 2.l 4 1.1 1 0.7
Multi Platform 1 I 0 0 4 2.8
Core Other 0 0 1 0.2 2 1.4
Angular 48 51 129 36.4 10 7
Grinding Stones 2 2.1 1 0.2 3 2.1
Pebble Tools 0 0 0 0 0 0
Hammerstones 0 0 0 0 0 0
TOTAL 94 100 354 100 142 100
Table 5. Gull to Tunkawanna sites: platform preparation of
Site Number of Platform Attributes
artefacts flaked or faceted overhang removal
n % n %
03 485 31 6.3 6 1.2
04 313 0 0 1 <1
06 101 13 12.8 7 6.6
07 61 0 0 0 0
08 42 13 33.3 15 35.7
10 193 1 <1 7 3.5
11 19 11 61.1 2 10.5
12 111 5 4.5 1 <1
1-4 72 11 15.2 7 9.7
Table 6. Gull to Tunkawanna sites: dorsal attributes of
Site Number Dorsal Attributes
of without >2 dorsal >1 rotations
artefacts cortex scars
n % n % n %
W3-03 485 133 28.7 95 19.2 72 14.5
W3-04 313 172 55 56 17.9 73 23.3
W3-06 1Ol 27 29.5 34 32.3 36 34.2
W3-07 6l 22 36 19 31.6 22 36
W3-08 42 17 40.5 30 71.4 19 42.8
W3-10 193 124 64.6 56 28.1 46 23.5
W3-II 19 2 10.5 5 26.3 6 31.5
W3-12 111 64 59.2 21 18.2 28 24.3
DE-SAS 1-4 72 42 58.3 28 38.8 15 20.8