SEARCH

SEARCH BY CITATION

Keywords:

  • Holocene;
  • prehistoric mining;
  • optical dating;
  • silcrete

Abstract

  1. Top of page
  2. Abstract
  3. Introduction
  4. The Silcrete at ODX16
  5. Background to the Site Identification and Quarry Investigation
  6. Surface Collections and Hand Excavations at ODX16
  7. The Machine Excavations
  8. Characteristics of Artefacts and Quarry Discard Materials
  9. Optical Dates from ODX16 – The Age of the Quarry
  10. Discussion and Conclusions
  11. Acknowledgements
  12. References

A site in the Olympic Dam area, recorded during archaeological surveying as a silcrete quarry, was investigated. Hand-excavated squares and subsequent machine-excavated trenches revealed an ancient “mine” rather than a simple surface quarry. Blocks of high-quality silcrete were levered from below the ground surface and many were knapped in the immediate area. Rubble in the pit backfills included large numbers of flakes. Single-grain optical dates from sediments in the backfilled pits demonstrate that the silcrete “mining” occurred during a short period in the late Holocene.


Introduction

  1. Top of page
  2. Abstract
  3. Introduction
  4. The Silcrete at ODX16
  5. Background to the Site Identification and Quarry Investigation
  6. Surface Collections and Hand Excavations at ODX16
  7. The Machine Excavations
  8. Characteristics of Artefacts and Quarry Discard Materials
  9. Optical Dates from ODX16 – The Age of the Quarry
  10. Discussion and Conclusions
  11. Acknowledgements
  12. References

During a survey at Olympic Dam in arid north-eastern South Australia (Figure 1), over 16000 archaeological sites were recorded at an average density of about 31 per km2. The main types of site recorded were stone artefact scatters (81%), artefact scatters with knapping floors (8%), single or multiple knapping floors (8%), quarries (1%) and hearths (2%). Of these, 160 sites were included in a salvage program involving detailed recording, collection of surface artefacts and excavations, as described in Hughes et al. (2011).

figure

Figure 1. The sudy area location map (prepared by BHP Billiton Base Metals, Olympic Dam).

Download figure to PowerPoint

Silcrete of various textures was the most commonly used material for artefact manufacture (71% of the artefacts recorded) and it was collected mainly as cobbles from gibber (stony) surfaces. Numerous small quarries on local outcrops of nodular silcretes and a few large quarries on extensive linear outcrops or capping layers occur in the study area. A small silcrete quarry, ODX_00016 (hereafter referred to as ODX16), was one of the sites investigated. Only at site ODX16 was there any sign of subsurface extraction, making this site unique in the region.

The Silcrete at ODX16

  1. Top of page
  2. Abstract
  3. Introduction
  4. The Silcrete at ODX16
  5. Background to the Site Identification and Quarry Investigation
  6. Surface Collections and Hand Excavations at ODX16
  7. The Machine Excavations
  8. Characteristics of Artefacts and Quarry Discard Materials
  9. Optical Dates from ODX16 – The Age of the Quarry
  10. Discussion and Conclusions
  11. Acknowledgements
  12. References

The Olympic Dam region covers ancient dissected plateau surfaces that, in common with other parts of arid Australia, underwent episodic erosion until the Mesozoic, with duricrust formation during the Tertiary (Ambrose & Flint 1981; Mabbutt 1977). The sediments forming the gibber plain at ODX16, including the silcrete (both the grey sandy and creamy-pink very fine textured varieties) and quartzite cobbles and boulders that they contain, were derived from the weathering and erosion of the capping of Cainozoic and Cretaceous sediments that occur there. These sedimentary units are as follows: Cainozoic (Tertiary) silicified sandy dune and beach-ridge deposits (Czs); and Cretaceous kaolinitic siltstones, shales, sandstones with glacially transported gravel, cobbles and boulders (Bulldog Shale) (K). Recent (Quaternary) landscape formation has involved the movement of superficial sand to form dunefields and pans.

Silcrete is indurated rock that comprises silicified clasts. The nature and origin of silcrete in eastern Australia has been covered in a number of papers in Langford-Smith (1978). Silcrete is composed largely of quartz clasts (ranging in size from silt and sand grains to boulders) cemented by micro- to cryptocrystalline silica. It contains at least 90% (and commonly more than 98%) silica.

In north-eastern South Australia, silcrete occurs mainly in the Cainozoic/Tertiary (Czs) geological unit in the form of outcrops, and cobbles and boulders on the floors of the swales and on gibber plains. In Czs at Olympic Dam, the silcrete is associated with Tertiary strandline deposits that once mantled much of the basin landscape, having formed locally through silica mobilisation and induration in the deep weathering profiles developed on those deposits (see, e.g., Ambrose & Flint 1981; Kreig & Rogers 1995; Wopfner 1978).

To a lesser extent, but significant as a source of artefact raw material, silcretes also formed in deep weathering profiles in the underlying Cretaceous Bulldog Shale (K). In several places at Olympic Dam, silcrete with numerous embedded quartzite cobbles or gravels has been recognised in the Cretaceous sediments. During the early to mid-Tertiary, the remnant Cretaceous sediments underwent prolonged deep weathering, forming porous kaolinitic siltstones in which there was extensive but patchy formation of Tertiary silcretes of different ages from Middle Eocene to Pliocene (44–6 million years ago) (Benbow et al. 1995: 201–7; Mabbutt 1977). Silcretes comprising silicified silty to clayey sediments have formed in the upper bleached zones of deep weathering profiles below the duricrust along the “breakaway” margins of Tertiary plateaux here formed on Bulldog Shale.

This geological history explains why the gibber plain at ODX16 consists primarily of Cretaceous and perhaps some Tertiary gravels, cobbles and boulders of silcrete and quartzite in a fine textured kaolinitic matrix. These resistant siliceous rocks have been concentrated in layers at the surface over hundreds of thousands of years, as the land surface was progressively lowered by removal of the fine textured matrix through weathering and erosion.

Because they comprise clasts or particles of available sediment, silcretes vary widely in texture, and in the Olympic Dam survey area they range from very fine chert-like materials to pebbly silcretes. All fracture through their embedded clasts to produce smooth conchoidal surfaces and all retain strong sharp edges when fractured. All forms, but especially the sandy, fine sandy and very fine or chert-like silcretes were common artefact sources. Sandy silcretes are more widespread than chert-like silcretes in the gibber cobbles, but all the silcrete outcrops observed show variation in texture over short distances, commonly from coarse sandy to very fine textured forms.

The occurrence of creamy extremely fine textured silcrete at ODX16 is very localised, indicating that there has been little lateral displacement of the cobbles and boulders of this material as erosion lowered the land surface. There are occasional cobbles and small boulders within and around the outcrop of grey sandy to pebbly silcrete, but most of the large rocks comprise extremely fine silcrete – predominantly silicified silt, with relatively few sand grains visible in hand specimens. In some blocks, the two types are admixed, indicating that the local silcretes have undergone more than one cycle of silicification:

Sandy and pebbly silcrete: Massive grey sandy silcrete with band of granules and pebbles. In hand specimen, compact pink matrix with white–cream or paler pink pebbles and coarse sand grains. At 50× magnification, very poorly sorted, mainly rounded, sandy quartz in a microcrystalline silica and clay matrix. Clasts from coarse sand to pebble size of quartzite and cream–white amorphous quartz and previously formed silcrete.

Extremely fine silcrete: Very fine chert-like pink to cream–grey silcrete. Surface textures range from completely smooth to very slightly rough. All contain very occasional visible sand grains. Slight weathering rind on cobbles. In hand specimen, all fresh samples are pinkish-cream, mottled with white, grey and paler pink, in smooth compact silcrete with very rare visible sand grains. At 50× magnification, sparse rounded to subangular silt to very fine sand-sized quartz grains in a crypto- to microcrystalline silica matrix with some fibrous (chalcedonic) silica and traces (<5%) of dispersed clay. Occasional cracks infilled with quartz grains, clay and crypto- to microcrystalline silica. Rare vugs lined with chalcedonic silica. Occasional engulfed fragments of an older sandy silcrete.

Background to the Site Identification and Quarry Investigation

  1. Top of page
  2. Abstract
  3. Introduction
  4. The Silcrete at ODX16
  5. Background to the Site Identification and Quarry Investigation
  6. Surface Collections and Hand Excavations at ODX16
  7. The Machine Excavations
  8. Characteristics of Artefacts and Quarry Discard Materials
  9. Optical Dates from ODX16 – The Age of the Quarry
  10. Discussion and Conclusions
  11. Acknowledgements
  12. References

Olympic Dam is the site of a large underground mine producing copper, gold and uranium and a major open cut expansion of the mine has been planned (Figure 1). In the context of producing (over three decades) the Environmental Impact Statements (EISs) for the establishment of the mine and proposed expansion of the existing mine, a variety of archaeological investigations has been undertaken in the lease area (see Hughes et al. 2011).

In 1980, Hughes and colleagues commenced archaeological studies (Kinhill-Stearns Roger 1982, chapter 5) for the then proposed Olympic Dam mining project. In 2009, BHP Billiton submitted an EIS for a proposed expansion of the mine (ARUP/ENSR 2009). That EIS described the Olympic Dam Agreement reached between BHP Billiton and the Barngala, Kokotha and Kuyani native title claimant groups and its Heritage Management Protocol, which specifies archaeological investigations to mitigate the impact of the mine expansion in a 600 km2 area within which most of the development will occur.

These investigations are being carried out as a combined mitigation and research program by archaeologists working with representatives of the native title claimant groups, over a 7 year period. Handheld computers running ArcGIS software were used to navigate and to record archaeological sites for the field survey stage completed during 2007, 2008 and 2009. Salvage investigations were undertaken from 2010 to 2013.

Surface Collections and Hand Excavations at ODX16

  1. Top of page
  2. Abstract
  3. Introduction
  4. The Silcrete at ODX16
  5. Background to the Site Identification and Quarry Investigation
  6. Surface Collections and Hand Excavations at ODX16
  7. The Machine Excavations
  8. Characteristics of Artefacts and Quarry Discard Materials
  9. Optical Dates from ODX16 – The Age of the Quarry
  10. Discussion and Conclusions
  11. Acknowledgements
  12. References

Site ODX16 was identified during the survey stage as a typical small quarry on nodular silcrete, as some of the surface boulders of locally distinctive, pinkish-cream, extremely fine-grained silcrete showed negative flake scars, blocks of flaked silcrete were found close to the cluster of boulders and there were numerous knapping floors surrounding them, all on the same distinctive silcrete. The site was on a stony (gibber) plain, thinly and discontinuously mantled with sand and just north of the toe of a 6 m high east–west orientated longitudinal dune. The gibber plain between this dune and the next dune to the north was 500 m wide.

In mid-2011, a standard process for salvage collection and excavation commenced at ODX16, and a 1 m square grid was laid out over an area of 143 m2 that covered the cluster of silcrete boulders and numerous adjacent knapping floors (Figure 2). At that stage, it was presumed that the boulders were the surface expression of a larger outcrop of silcrete.

figure

Figure 2. A plan of the site showing the layout of the grid squares and the boundary of the artefact scatters. The silcrete outcrop is in the central-southern part of the grid.

Download figure to PowerPoint

Surface material was collected from all 143 grid squares, but – unusually when compared with other local quarries – flaked material was also found in subsurface sediment, so excavation to a depth of about 100 mm was undertaken in 56 of those squares. In six grid squares within or close to the outcrop zone, excavations were extended deeper to identify the nature of the quarried material and the quarrying process (Figure 3). These excavations extended to depths ranging from 300 mm to 1.2 m, in some cases ending on large pieces of rock that prevented further hand excavation.

figure

Figure 3. A plan of the site showing the location of the hand- and machine-excavated squares and trenches and extent of the silcrete outcrop.

Download figure to PowerPoint

In every instance, no natural stratigraphy could be identified in these six deeper excavations and the material recovered was uninformative. Very shallow gravelly sand mantled a variably looser or more compact gravelly mixture with an irregular surface. The gravel mixture comprised flaked silcrete artefacts, angular silcrete rubble and rounded to subrounded quartzite and silcrete pebbles, many encrusted with white material identified subsequently as kaolinite. The artefacts were mainly flakes (on average 95%), including about 40% broken flakes, more than half of which were longitudinally cone-split. The angular silcrete rubble, which included cobble as well as gravel-sized fragments, lacked the diagnostic features of flaked artefacts. During this first phase of salvage investigation, several tens of thousands of silcrete artefacts and pieces of angular silcrete rubble were collected at densities of up to 8000 per m2 (more than 15000 per m3) from these more deeply excavated squares.

A subsequent detailed inspection of the stratigraphy exposed in the six excavated squares demonstrated that in some there was distinct dipping of the gravelly sediments in two or three directions, and anomalous stratigraphy, with gravel- to cobble-sized pieces of angular silcrete lying both above and below sand or rounded pebbles. It was apparent that the subsurface deposits had been highly disturbed if not overturned, and it was decided to abandon further hand excavation and to use machine-excavated trenches to determine the extent of the silcrete outcrop and examine the sediments and any structures present.

The Machine Excavations

  1. Top of page
  2. Abstract
  3. Introduction
  4. The Silcrete at ODX16
  5. Background to the Site Identification and Quarry Investigation
  6. Surface Collections and Hand Excavations at ODX16
  7. The Machine Excavations
  8. Characteristics of Artefacts and Quarry Discard Materials
  9. Optical Dates from ODX16 – The Age of the Quarry
  10. Discussion and Conclusions
  11. Acknowledgements
  12. References

Stratigraphy and the extent of the silcrete outcrop

Thirteen trenches were dug using a backhoe (Figure 3) and cleaned by hand to examine the sections. The trenches were placed to define the subsurface extent of the silcrete outcrop, to allow the stratigraphy of the sediments forming the gibber plain/swale to be recorded and to locate features that might explain the degree of disturbance revealed in the hand-excavated squares. Trenches T1, T2, T12 and T13 were excavated along a north–south transect to a depth of ∼1 m. They revealed an almost continuous sequence along their length of red dune sand, decreasing in thickness northwards from 900 mm at the toe of the dune to 400 mm in the silcrete outcrop, and then to <100 mm north of the outcrop. Beneath the sand there was a sharp break to a 200–300 mm thick layer of quartzite and silcrete pebbles (Figure 4). These were derived from the gibber surface layer and were mainly subrounded, had weathered and pitted surfaces, and were encrusted with white kaolinite and varying amounts of carbonate minerals. This layer of pebbles had a white silty sand matrix and in many places it was heavily carbonate-cemented. The pebble layer graded downwards into a cobble layer consisting of cobbles and occasional small boulders of silcrete (∼80%) and quartzite (∼20%) in a matrix of white silty sandy gravel. The silcrete consisted of angular nodules of grey to cream mainly sandy textured silcrete, but much of it containing gravel-sized clasts. The quartzite consisted of both smooth, rounded, water-worn cobbles and tabular blocks of flaggy material.

figure

Figure 4. The east section of trench T1, showing the partially carbonate-cemented, mainly rounded gravels forming the gibber plain capped with red sand.

Download figure to PowerPoint

At the southern end of T1, the partially cemented gravel and cobble layers continued beneath the sand dune, the base of which dates to >50000 years (see below), indicating that these layers formed before then.

T3 to T11, each 2–3 m in length and 0.5–1 m deep, were dug primarily to determine the subsurface extent of the distinctive, creamy-pink coloured, very fine-grained silcrete (described in the following section) that was represented by boulders at the surface and was the source of most of the stone artefacts on the site surface. Only in T9 were angular cobble-sized pieces and associated artefacts of this type of silcrete found below the red sand. In T3, T4, T11 and T8, no silcrete of this kind was observed, nor in T10 further to the west or T5, T6 and T7 to the east. The results of the hand and machine excavations combined indicate that the subsurface occurrence of this creamy-coloured, very fine-grained silcrete (in contrast to the more widely occurring grey sandy silcrete in the cobble layer) is confined to an area ∼15 m east–west by ∼10 m north–south (Figure 3).

Numerous small boulders of this creamy fine-grained silcrete were scattered across the site, including beyond the inferred boundary of the outcrop. All of these were lifted and all were found to be resting on the surface or were partially buried in the surficial red sand, and none extended down into the underlying pebble and cobble layers. This demonstrated that none of them was in situ and instead they had all been carried to their present locations.

Quarry pits

At the northern end of T2, two clearly defined asymmetrical steep-sided quarry pits were identified, one 1.1 m and the other 1.4 m deep (Features 2 and 3 in Figure 5). The machine-excavated trench T2 ended at the southern edge of one of the hand-excavated squares (H1 in Figure 3). This had been excavated to a depth of 700 mm, at which point excavation was stopped because the large amounts of tightly packed rock (artefacts, other angular pieces of silcrete and pebbles) could not be penetrated using trowels and shovels. More than 2000 pieces of angular silcrete, most of them clearly artefacts, were recovered from this square. This was one of the squares in which distinct dipping of the gravelly sediments had been observed (Figure 6) and it was evident from the sections that this was yet another pit fill (Feature 4 in Figure 5).

figure

Figure 5. The east wall of square H1 and trench T2, showing Features 2, 3 and 4 and the stratigraphy of the sediments into which they were dug.

Download figure to PowerPoint

figure

Figure 6. The east wall of square H1, showing the upper fill of Feature 3 dipping down slightly to the south (right) and overlying the fill of Feature 4, which dips down more steeply to the north (left).

Download figure to PowerPoint

Only a small part of the fill of Feature 4 was exposed by the hand excavations, so its full size is not known. Features 2 and 3 had been dug through the red sand into the >50000-year-old partially cemented gravel and cobble layers. The pit fills contained some materials that dipped steeply, but in different directions. This aided in distinguishing between the different fills and helped demonstrate that Feature 3 was the youngest, and that it had been partially cut into the fills of Features 2 and 4 either side. No charcoal, heat-discoloured sediment or fire-pits were observed anywhere in the excavations or on the ground surface.

The southern pit (Feature 2) was filled with poorly sorted gravel rubble with some artefacts in a red sandy matrix. The proportions of the gravel and small cobbles increased with depth. The number of artefacts was small when compared with the fills of Features 3 and 4 to the north. In the middle of the fill was a small boulder of good-quality sandy silcrete. Halfway up the fill on its northern side was a distinct thin band of redeposited gravels, dipping into the middle of the pit.

The central pit (Feature 3) had a more complex fill:

0–150 mm: Red sand with silcrete artefacts and small amounts of gravel.

150–550 mm: Angular cobbles of silcrete and numerous artefacts in a pebbly, red sand matrix, which was faintly bedded.

550–1200 mm: Occasional silcrete artefacts in a faintly bedded sandy gravel matrix.

1200–1400 mm: Silcrete artefacts in a pale red sand matrix. A small boulder and large cobble of silcrete protruded from the section wall.

In all three pits, the angular silcrete (including flakes) and gravel pieces were not lying horizontally but dipped at various angles into the fill. The gravel was mainly subrounded white-coated pebbles derived from the gibber layer, but some of it was fractured angular pieces of silcrete with no obvious signs of having been produced by flaking.

One of the other hand-excavated squares (H2 in Figure 3) had also been dug into the fill of a steep-sided pit, which was larger and deeper than the 1 m2 excavated (Figure 7).

figure

Figure 7. Square H2 looking west, showing the partially excavated fill of a pit feature. In contrast to the natural stratigraphy (Figure 4), the gravels and cobbles here are mainly of angular silcrete, much of it flaked.

Download figure to PowerPoint

About 4 m to the west of trench T2, in the hand-excavated square H3, the stratigraphy was horizontal. In this area, flaked silcrete blocks and large broken flakes were observed in the red sand, material apparently discarded on the sandy swale surface adjacent to the quarry, and later covered by disturbed windblown red sand. The layer of worked silcrete was approximately 100–150 mm thick, overlain by about 150 mm of red slightly gravelly sand and resting on at least 300 mm of red sand (Figure 8). Samples for optical dating were collected from above and below this artefact layer.

figure

Figure 8. A layer of mainly worked silcrete in the west section of square H3. Samples for optical dating were taken immediately above and below the silcrete layer.

Download figure to PowerPoint

These pit fills and their associated layers of mainly worked silcrete were artificial, not natural, features. The pits were sunk into the open and virtually flat swale surface, and their fills were material that had slumped into voids formed by removing blocks of rock. The stratigraphy of the pit fills indicate that these pits were created by people levering out boulder-sized blocks of silcrete and breaking them up into smaller pieces. The pit upcast and blocks on the present ground surface demonstrate that some of these pieces were knapped on the spot, and some were moved and reworked at a later date to produce the numerous knapping floors recorded across the present surface of the site. Rock debris that had fallen back into the pits consisted of angular pieces of silcrete (including flakes and cores) from breaking up the boulders and then knapping the smaller blocks, along with pebbles and red sand derived from digging into the sand-mantled gibber. The dipping maintained in each fill indicates that most voids filled rapidly as the blocks were levered out and worked.

Much of the silcrete was extracted by levering out buried blocks that protruded above the surface. Quarry debris from other sites in the study area, and material discarded on adjacent dune sites, demonstrates that silcrete was extracted from outcrops by flaking or by taking detached pieces of rock from their surfaces. It is likely that at ODX16 protruding boulders were followed down and “mined” (i.e. quarried as defined by Hiscock & Mitchell 1993). Large blocks were levered out to obtain good-quality stone. Very little nodular silcrete remains at the site now, and it is likely that the quarry was exhausted after a short period of use. The nature of the silcrete, the disposition and the characteristics of the artefact assemblage and the optical dates from the site (see below) all strongly support this interpretation and provide a chronology for the episode of “mining” at this site.

It was the finer coherent silcrete, with excellent flaking properties, that was the focus of extraction by the prehistoric miners. A “buffer” search on the ArcGIS database of Olympic Dam sites showed that the proportion of chert-like silcrete recorded during site surveys in artefact scatters was 10% within a zone of 1 km from the quarry site, but that proportion dropped to 4% in artefact scatters recorded in a zone from 1 to 2 km from the quarry. It is likely that ODX16 was the major source of this silcrete locally, but further afield similar chert-like silcrete was obtained from other outcrops and from the gibber throughout this part of the survey area. The mainly chert-like silcrete at ODX16 (Figure 9) is not different either texturally or mineralogically from that at numerous other outcrops, so it is not known why this site is unique in the region in terms of the method of extraction.

figure

Figure 9. A flake from ODX16. The right part is chert-like chalcedonic silcrete, but the left part is a mixture of fragments of chert-like silcrete engulfed in a matrix of sandy silcrete. The scale is in centimetres.

Download figure to PowerPoint

Characteristics of Artefacts and Quarry Discard Materials

  1. Top of page
  2. Abstract
  3. Introduction
  4. The Silcrete at ODX16
  5. Background to the Site Identification and Quarry Investigation
  6. Surface Collections and Hand Excavations at ODX16
  7. The Machine Excavations
  8. Characteristics of Artefacts and Quarry Discard Materials
  9. Optical Dates from ODX16 – The Age of the Quarry
  10. Discussion and Conclusions
  11. Acknowledgements
  12. References

It appeared from the excavations that the subsurface artefacts in the pits were smaller and more commonly broken than the flakes in knapping clusters on the present ground surface, which were assumed initially to have been manufactured when quarrying occurred. Eight samples of the ODX16 artefacts were analysed to compare the material discarded on the surface with the material recovered from the pit backfill areas. Figure 10 shows the locations from which the eight samples were drawn. H1–H4 are hand-excavated squares, of which H1, H2 and H4 are cut into pit backfill and H3 into an associated layer of upcast material. S1–S4 are surface collections, of which S3 and S4 were identified as discrete knapping floors at the time of collection. The surface samples range in size from 141 to 322 artefacts and the subsurface samples from 1295 to 4125 artefacts (Table 1).

figure

Figure 10. The locations of the eight samples of artefacts.

Download figure to PowerPoint

Table 1. Raw data for the catalogued artefacts
SampleSample area (m2)Number of bags cataloguedBags catalogued (%)FlakesRetouched flakesCoresHammer stonesTotal
S1881002857282322
S2282810032614696415
S3411001025331141
S42210022116252264
H11 (excavated to 500 mm)3453398413111174125
H21 (excavated to 470 mm)33692051410692170
H31 (excavated to 365 mm)15481421106311495
H41 (excavated to 380 mm)24100122885811295

Artefact assemblages

The surface and subsurface samples differ in their composition. There are much higher percentages of cores, retouched flakes (RF) and hammer stones in the surface samples. Figure 11 indicates that approximately 20% of the surface material is made up of cores and retouched flakes, compared with about 5% in the subsurface samples. The composition of the surface material suggests that formal material selection, and targeted flake creation, took place here.

figure

Figure 11. The composition of the samples, expressed as a percentage.

Download figure to PowerPoint

The proportions of cores, retouched flakes and hammer stones are significantly lower in the subsurface artefact samples, although their absolute numbers and frequency of occurrence are much higher than in the surface samples (Tables 1 and 2). Artefact densities used for comparison are expressed as the total numbers per m2, whether the artefacts were on the surface or were distributed through the sediment column at that location. The density of subsurface artefacts was estimated to be more than 140 times that of surface artefacts (Table 2). The densities of cores (35 times), retouched flakes (15 times) and hammer stones (185 times) were similarly much higher in the subsurface material. For the subsurface samples, these are minimum numbers and densities as the base of the artefact-rich material was not reached in three squares. These comparisons suggest that in addition to the large-scale smashing of silcrete blocks, which produced the extremely high quantities of subsurface material, a much greater amount of knapping also took place when silcrete was being extracted originally from the quarry pits than on the surface at a later date.

Table 2. Actual/estimated numbers and densities (number per m2) of different artefact types in the surface and subsurface samples
SampleSample area (m2)Total number of artefacts cataloguedSample catalogued (%)Actual/estimated numbers and densities (number per m2) of the different artefact types
All artefactsCoresRetouched flakesHammer stones
No.DensityNo.DensityNo.DensityNo.Density
S1 322100322 28 7 2 
S2 415100415 69 14 6 
S3 141100141 33 5 1 
S4 264100264 25 16 2 
S1–S4 combined421142 1142271554421110.26
H1 4125537783 209 25 1732
H2 2170693145 154 6 913
H3 1495483115 131 21 12
H4 12951001295 58 8 11
H1–H4 combined49085 15338383555213860152748
Density ratio subsurface/surface    142 35 15 185

The size of the flakes

Between 100 and 263 flakes from each sample were measured to determine their size. Figure 12 shows that except for S1, which will be discussed shortly, the surface flakes are consistently and significantly larger than the subsurface flakes. The box plots in Figure 12 display the strength of the variation between the samples, and the bullet graphs display the confidence with which the results are demonstrated to be real – or, rather, “are not attributable just to the vagaries of sampling” (Drennan 2009: 151). As can be seen in the bullet graphs, the means for all of the subsurface measurements lie beyond the 99% confidence level error range for the surface measurements (except for S1). This indicates that there is less than a 1% chance that three of the surface samples (S2, S3 and S4) came from the same population as the subsurface samples (after Drennan 2009). S1, on the other hand, appears to have been drawn from the same parent population as the subsurface samples.

figure

Figure 12. A comparison of the complete flake sizes by sample.

Download figure to PowerPoint

As all of the batches displayed skewing towards larger flakes, it was necessary to use a logarithm transformation to normalise the shape of the batches before statistical significance testing could be conducted (Drennan 2009: 52, 55–60). The bullet graphs (Figure 12) displaying confidence therefore have a logarithmic scale on the Y-axis that represents an increase in the order of magnitude, rather than a linear increase.

S1

The flakes from S1 appear to have been drawn predominantly from the same parent population as the subsurface material. The S1 sample encompasses square H1 and the northern end of trench T2, which contains Features 2, 3 and 4 (Figures 5, 6 and 10). The ground surface in the S1 area is slightly higher than surrounding parts of the site and the sterile red sand that lies immediately below the surface artefact scatter over much of the rest of the site (see, e.g., H3 in Figure 8) is discontinuous to absent (Figures 5 and 6). Either sand never accumulated on this slightly raised area or it has been stripped from it by wind or wash erosion. As a result, the surface artefactual material is not separated from the underlying quarry pit fill by a distinct sterile sand layer. Rather, the subsurface material was exposed at the time of collection, so the surface sample S1 includes subsurface/pit backfill material. The inclusion of a large number of subsurface artefacts in the S1 sample is also suggested from Figure 11, which shows S1 as having the highest percentage of flakes of the surface samples.

H3

Another aberration is seen in the H3 material, where there are more tertiary and more broken flakes than in the other subsurface samples. Within the H3 sample, breakage has also occurred more frequently subsequent to manufacture. The H3 sample had the highest breakage rate of 48%, compared with other subsurface samples, which had breakage rates between 27% and 39%. Within the broken flakes, H3 had the lowest rate of longitudinally cone-split flakes, at 41%, compared with 57–65% in the other subsurface samples, indicating that fewer of the H3 flakes were broken during manufacture and instead were snapped sometime afterwards. In addition, H3 had more tertiary flakes than the other samples at 73.5%, compared with 53–65%, indicating that less of this material was produced during the preliminary stages of reduction and more during later stages.

The lower percentage of primary flakes and higher percentage of post-manufacture breakages suggest that the H3 flakes form part of a secondary deposit, comprising material discarded from quarrying activity nearby. The stratigraphy of the pit fills supports this interpretation. In Figure 8, it can be seen that this excavated square contains a distinct horizontal artefact and rubble layer. In contrast, the pits lack any coherent stratigraphy and are instead made up of tightly packed artefacts and stone rubble (see Figure 6), dipping at various angles and in several directions.

Optical Dates from ODX16 – The Age of the Quarry

  1. Top of page
  2. Abstract
  3. Introduction
  4. The Silcrete at ODX16
  5. Background to the Site Identification and Quarry Investigation
  6. Surface Collections and Hand Excavations at ODX16
  7. The Machine Excavations
  8. Characteristics of Artefacts and Quarry Discard Materials
  9. Optical Dates from ODX16 – The Age of the Quarry
  10. Discussion and Conclusions
  11. Acknowledgements
  12. References

The matrix of the backfills that have slumped into the spaces left by levering out blocks of rock is predominantly local red quartz sand, mixed with some pebbles from the underlying pebble layer. Throughout the Olympic Dam sand desert area, this sand has been demonstrated to have optically stimulated luminescence (OSL) characteristics, making it highly suitable for optical dating (see, e.g., Sullivan et al. 2012).

Five samples were collected for optical dating: one from near the base of the sand dune on the margin of the quarry, one each from the bases of the rubbly fills of two of the mining pits revealed in the machine-excavated trench T2 (Features 2 and 3) and two bracketing flaked silcrete in square H2 (Figure 4). These were collected during on-site discussions with Professor Nigel A. Spooner and dated under his supervision at the Environmental Luminescence Laboratory at the University of Adelaide. An outline of the dating procedure follows; full details will be provided elsewhere.

In the field, sediment samples for optical dating were collected in lightproof stainless steel coring tubes, along with approximately 500 g of sediment for radioisotope assay for uranium (U), thorium (Th) and potassium (K). A portable NaI gamma-ray spectrometer was then used to measure the in situ gamma-ray spectrum.

The cores were subsequently opened in the laboratory under darkroom conditions, and quartz grains in the 180–250 μm size range were extracted by a process including hydrochloric acid (HCl) digestion, density separation using heavy liquid (lithium heteropolytungstate) and hydrofluoric acid (HF) etching. Quartz grains were then positioned in arrays of 100 grains each on custom-produced sample discs for optically stimulated luminescence (OSL) measurement of the equivalent dose of each grain, using the single aliquot regeneration (SAR) protocol (Jacobs et al. 2006; Murray and Wintle 2000, Olley et al. 2004; Wintle & Murray 2006). Luminescence measurements were conducted using a Risø single-grain module interfaced to a Risø DA-20 OSL/TL reader; OSL was stimulated by a 532 nm laser and detected with an EMI 9235QB photomultiplier tube optically filtered by a 7 mm thick Hoya UV-bandpass U340 filter, and laboratory irradiations were administered using a calibrated 1.5 GBq 90Sr/90Y beta particle source onboard the Risø reader.

With regard to the environmental dose rate: contributions from cosmic radiation were calculated following Prescott and Hutton (1994); the soil radioisotope component was calculated using the concentrations of U, Th and K measured both by the field gamma-ray spectrometry and assayed by Genalysis Laboratories, Perth, using inductively coupled plasma – mass spectroscopy (ICP–MS) for U and Th, and inductively coupled plasma – atomic emission spectrometry (ICP–AES) for K. The water content of the samples was measured as the percentage of dry weight by weighing, drying overnight at 125°C and reweighing. These data were combined with the individual grain equivalent doses using the AGE program of Grün (2009) and the subsequent sets of grain ages for each sample were analysed to determine the sample ages: the results obtained are shown in Table 3.

Table 3. Optical dates from ODX16
SampleLaboratory codeMaterial datedOptical age (ka)
OD16 sample 1 Feature 2 Trench 2, 700 mm depthAd12010Red sand in base of fill of Feature 22.4 ± 0.2
OD16 sample 2 Feature 3 Trench 2, 1040 mm depthAd12011Red sand in base of fill of Feature 32.3 ± 0.2
OD16 sample 3 square H2 (PAD402), 140 mm depthAd12012In red sand above buried layer of artefacts and rock rubble1.0 ± 0.1
OD16 sample 4 square H2 (PAD402), 280 mm depthAd12013In red sand below buried layer of artefacts and rock rubble3.1 ± 0.2
OD16 Trench 1, sample 5, 900 mm depthAd12014K surface extending just below north side of dune (mixed ages on predominantly older grains from dune core and younger grains from overlying loose sand)51.9 ± 2.1 and 7.4 ± 0.9

The age of the core of the dune immediately to the south of the outcrop indicates that a linear sand ridge has been present in that location for at least 50000 years and that the gibber plain surface that this overlies is even older. The use of the silcrete quarry occurred over a relatively short time period, at about 2300–2400 years ago. The fill of the quarry pits gives a firm age for the process of block removal. The artefact/rock rubble layer in square H3 adjacent to the quarry pits was deposited sometime between 3100 and 1000 years ago, and almost certainly corresponds to the short time period time when quarrying occurred. Knapping floors found on the present ground surface in the vicinity of these dated features must be younger than 1000 years. This suggests that there was reuse of blocks of suitable silcrete like those on the surface today, which occurred well after they were quarried.

Discussion and Conclusions

  1. Top of page
  2. Abstract
  3. Introduction
  4. The Silcrete at ODX16
  5. Background to the Site Identification and Quarry Investigation
  6. Surface Collections and Hand Excavations at ODX16
  7. The Machine Excavations
  8. Characteristics of Artefacts and Quarry Discard Materials
  9. Optical Dates from ODX16 – The Age of the Quarry
  10. Discussion and Conclusions
  11. Acknowledgements
  12. References

The artefacts in the subsurface material are different from those on the present ground surface, indicating that two different kinds of knapping took place at ODX16. The first knapping event occurred at the time when large blocks were levered from the outcrop. It is characterised by an exceptionally high density of artefacts, including thousands of small flakes. This was a period of intense quarrying, during which the silcrete blocks were levered from the outcrop and were literally being smashed as they were lifted up. Cores, retouched pieces and hammer stones are also present, indicating that in addition to the preliminary breaking up of the silcrete blocks, knapping also took place immediately above and adjacent to the voids, producing some of the pit backfill material.

Dating of the site indicates that the pits were worked between 2300 and 2400 years ago, with some upcast material being thrown onto the adjacent ground surface as the pits were dug. The pits were then abandoned, with a period of time passing in which a discontinuous layer of sand accumulated over most of the pit fills and associated upcast material. The upper sandy layer returned an optical date of 1.0 ± 0.1 ka. People returned to the site within the past 1000 years and collected and knapped silcrete blocks left on the ground surface from the previous quarrying activities. This produced the knapping floors and scatters of artefacts now visible on the ground surface.

The surface knapping that took place at least a thousand years later was unrelated to the initial quarrying/knapping event. The surface material is characterised by significantly larger flakes, a higher percentage of cores and retouched pieces, and a very much lower density of artefacts. This suggests that knapping took place on the surface, with larger remnant blocks of silcrete lying on the surface being selected for working.

Site ODX16 is an unusual procurement site in the Olympic Dam area. Silcrete was not extracted through splitting or flaking of the outcrop surface, but through a “mining” process. People apparently dug and levered silcrete blocks from the outcrop, and followed the edges of the blocks below the ground surface to do this. There is no evidence of the tools used to achieve this, but mulga (Acacia aneura) is abundant in the area, and its roots and branches are hard and strong enough to have been used as levers to extract large blocks of rock.

The trench sections demonstrated that the silcrete had formed on and partially incorporated weathered Cretaceous sediments, including cobbles, and that the outcrop was of a limited size. Importantly, it was clear from the sections that there had been selective removal by quarrying of fine, good-quality silcrete (silicified silts and very fine sands), with coarser commonly pebbly textured silcrete left within the rubble.

The rubble in the backfills was intimately mixed with flakes and other discarded artefactual material. It is estimated that at least 20000 artefacts remain to be analysed, but these preliminary analyses indicate that around 40% of the discarded flakes were broken, probably during manufacturing.

Acknowledgements

  1. Top of page
  2. Abstract
  3. Introduction
  4. The Silcrete at ODX16
  5. Background to the Site Identification and Quarry Investigation
  6. Surface Collections and Hand Excavations at ODX16
  7. The Machine Excavations
  8. Characteristics of Artefacts and Quarry Discard Materials
  9. Optical Dates from ODX16 – The Age of the Quarry
  10. Discussion and Conclusions
  11. Acknowledgements
  12. References

BHP Billiton funded the Olympic Dam archaeological salvage and research program, and drew the location map. The hand excavations were supervised by Dr Oliver Macgregor and the field team included Mark Agostini, Andrew Ball, Dawn Buerdlmayer, Harold Dare (Barngala representative), Ruki Hartman-Thomas, Heather Leasor, Mychal Ludwig, Michael Mckenzie (Kuyani representative), Robin McAlpine, Bonnie Mayo, Harrison Pitts, Paulina Przystupa, Sarah Robertson, Jo Threadgold and Glen Wingfield (Kokotha representative). Peter Hiscock participated in field discussions with the authors. Katarina Sporcic drew the site plan from field survey data. The field gamma-ray spectrometry, sample collection and optical dating were undertaken by Nigel Spooner, Frances Williams, Daniele Questiaux and Rochelle Dumaua from the Environmental Luminescence Laboratory at the University of Adelaide. We thank them for helpful discussions about this unusual dating situation. Amy Way undertook the artefact analyses as a student intern in the ANU Masters in Archaeological Science program.

References

  1. Top of page
  2. Abstract
  3. Introduction
  4. The Silcrete at ODX16
  5. Background to the Site Identification and Quarry Investigation
  6. Surface Collections and Hand Excavations at ODX16
  7. The Machine Excavations
  8. Characteristics of Artefacts and Quarry Discard Materials
  9. Optical Dates from ODX16 – The Age of the Quarry
  10. Discussion and Conclusions
  11. Acknowledgements
  12. References
  • Ambrose, G.J. and Flint, R.B. 1981. A regressive Miocene lake system and silicified strandlines in northern South Australia: Implications for regional stratigraphy and silcrete genesis. Journal of Geological Society of South Australia 28: 8194.
  • ARUP/ENSR. 2009. Olympic Dam expansion draft environmental impact statement. BHP Billiton, South Australia.
  • Benbow, M.C., Callen, R.A., Bourman, R.P. and Alley, N.F. 1995. Deep weathering, ferricrete and silcrete. In J.F. Drexel , W.V. Preiss and A.J. Parker (eds), The Geology of South Australia. Bulletin 54, Vol. 2, pp. 201207. Mines and Energy, Adelaide, South Australia.
  • Drennan, R.D. 2009. Statistics for Archaeologists: A Common Sense Approach, 2nd edn. Springer, New York.
  • Grün, R. 2009. The “AGE” program for the calculation of luminescence age estimates. Ancient TL 27: 4546.
  • Hiscock, P. and Mitchell, S. 1993. Stone artifact quarries and reduction sites in Australia: Towards a type profile. Australian Heritage Commission Technical Publication Series No. 4. AGPS, Canberra.
  • Hughes, P.J., Hiscock, P., Sullivan, M.E. and Marwick, B. 2011. Archaeological investigations at Olympic Dam in arid northeast South Australia. Journal of the Anthropological Society of South Australia Special Edition: Issues in South Australian Aboriginal Archaeology 34: 2137.
  • Jacobs, Z., Duller, G.A.T. and Wintle, A.G. 2006. Interpretation of single grain De distributions and calculation of De. Radiation Measurements 41: 264277.
  • Kinhill-Stearns Roger. 1982. Olympic Dam Project Draft Environmental Impact Statement. Roxby Management Services Pty Ltd, Adelaide, South Australia.
  • Kreig, G.W. and Rogers, P.A. 1995. Stratigraphy – marine succession. In J.F. Drexel , W.V. Preiss and A.J. Parker (eds), The Geology of South Australia. Bulletin 54, Vol. 2, pp. 112123. Mines and Energy, Adelaide, South Australia.
  • Langford-Smith, T. (ed.). 1978. Silcrete in Australia. Department of Geography, University of New England, Armidale.
  • Mabbutt, J. 1977. Desert Landforms. Australian National University Press, Canberra.
  • Murray, A.S. and Wintle, A.G. 2000. Luminescence dating of quartz using an improved single-aliquot regenerative-dose protocol. Radiation Measurements 32: 5773.
  • Olley, J.M., Pietsch, T. and Roberts, R.G. 2004. Optical dating of Holocene sediments from a variety of geomorphic settings using single grains of quartz. Geomorphology 60: 337358.
  • Prescott, J.R. and Hutton, J.T. 1994. Cosmic ray contribution to dose rates for luminescence and ESR dating: Large depths and long-term variations. Radiation Measurements 23: 497500.
  • Sullivan, M., Field, T.L., Hughes, P., Marwick, P., Przystupa, P. and Feathers, J.K. 2012. OSL ages that inform late phases of dune formation and human occupation near Olympic Dam in northeastern South Australia. Quaternary Australasia 29 (1): 411.
  • Wintle, A.G. and Murray, A.S. 2006. A review of quartz optically stimulated luminescence characteristics and their relevance in single-aliquot regeneration dating protocols. Radiation Measurements 41: 369391.
  • Wopfner, H. 1978. Silcretes of northern South Australia and adjacent regions. In T. Langford-Smith (ed.), Silcrete in Australia, pp. 93142. Department of Geography, University of New England, Armidale.