Fremantle Stuff > FHS > Fremantle Studies > 7 > Dortch

Archaeological evidence for early human presence in the western reaches of the Greater Swan Region, Western Australia

Charles Dortch and Joe Dortch

Dortch, Charles & Joe Dortch 2012, 'Archaeological evidence for early human presence in the western reaches of the Greater Swan Region, WA', Fremantle Studies, 7: 51-76.


This paper presents archaeological evidence for early human occupation in the western parts of the Greater Swan Region, Western Australia during the last ice age and immediate post-glacial period, when sea levels globally were greatly lowered. The informal label ‘Greater Swan Region’ refers to a 60-km-long and 40-km-wide, east-West transect from the Darling Escarpment through the centre of the Perth Metropolitan Region to the coast, and thence seaward to the deepening contours of the continental shelf edge, ten kilometres west of Rottnest Island (cf 100 m isobath: Fig. 1). In assessing early human presence in this study area through many millennia of lowered sea levels, one views this stretch of land from the Darling Scarp westward across the exposed (emergent) continental shelf as a single region. This is requisite, on the grounds that through hundreds of human generations this coastal sand plain was in its entirety traversed by Aboriginal hunter-gatherer groups.

The focus then is on Aboriginal presence in the western part of the Greater Swan Region, as it was before post-glacial sea level rise to present height. The study area comprises a small segment of the seaward third of the Swan Coastal Plain - as it happens, centred on the Fremantle area - and a much larger zone of adjacent, now submerged, continental shelf with its offshore islands. The onshore portion of the study area is bounded by the 40-km-long coastline between Leighton Beach and Warnbro Sound, and inland by the corridor of freshwater lakes and wetlands running from North Lake and Bibra Lake through the Beeliar Regional Park southward 15 kilometres and then shifting south-westward a further ten kilometres to the brackish Lakes Coolongup and Walyungup. The southern boundary of this onshore part of the study area runs from these two lakes westward to Warnbro Sound; the northern boundary is from Leighton Beach (Mosman Park) to the Swan River, as far upstream as Blackwall Reach (Fig 1, including boxes 2 and 3).

The archaeological information in this appraisal of early human presence in the study area consists of flaked stone artefacts made by Aboriginal tool makers; the stratigraphical contexts of these artefacts within their find sites; and these sites’ radiometric ages. Of equal importance to the early evidence examined here, though outside the scope of this paper, is Aboriginal occupation in the onshore part of the study area over the past few millennia, including the colonial/modern era. [1]

Late Quaternary context

The archaeological study is set in the Late Quaternary period, from ca 30 000 to ca 6000 years ago. This span of time covers the last phases of the last glaciation (or ‘ice age’) within the Pleistocene epoch, and the first few millennia of the Holocene epoch (or post-glacial period). The last glaciation, like the many preceding ice ages recurring during the roughly two million years of the Pleistocene epoch, was a time of massive expansion of glaciers and ice sheets, mainly in the northern hemisphere, and greatly lowered sea levels world-wide. Under full glacial conditions, the Australian continental shelves were entirely exposed, and all offshore islands, including Tasmania, were part of the mainland. During Pleistocene ice ages, the south-west of Western Australia was colder and drier than it is today, though no glaciers formed here. In the Greater Swan Region, the Swan River followed its present course through the ancient limestone canyon of Blackwall Reach and past the equally ancient limestone cliffs at Rocky Bay, before flowing across the emergent continental shelf. Its former channel lay a few kilometres to the north of the ‘pre-island’ Rottnest locality before reaching a now long submerged and long forgotten marine littoral, ten kilometres further west.

The off-shore islands

The Nyungar-speaking Aboriginal people of the Greater Swan Region and elsewhere in south-western Australia knew Rottnest Island as ‘Wadjemup’. The island was uninhabited at the beginning of the colonial period, [2] and there is neither material evidence nor apparently any Nyungar oral tradition suggesting that south-western hunter-gatherers visited Rottnest or any other offshore islands in the region after they became separated from the mainland.

A Nyungar story, recounted to the early colonist George Fletcher Moore in the 1830s, describes how the offshore islands in the Greater Swan Region were once connected to the mainland.

The natives have a tradition that Rottnest, Carnac, and Garden Island, once formed part of the mainland, and that the intervening ground was thickly covered with trees; which took fire in some unaccountable way, and burned with such intensity that the ground split asunder with a great noise, and the sea rushed in between, cutting off these islands from the mainland. [3]

This extremely old oral tradition, recorded in the first decade of the colonial era, indicates an awareness among Nyungar people that in ‘pre- island’ times all three localities had been part of the ancestral landscape.

Ten or eleven thousand years ago, in lower latitudes throughout the world, glacial climates began giving way to the warmer conditions of the Holocene or post-glacial period. With the ending of the last ice age, and the partial melting of ice sheets, sea levels rose globally, finally reaching their present height during the Early to Middle Holocene. In this encroachment of the sea over the continental shelf, Rottnest Island was cut off about 6500 years ago, [4] and Garden Island and Carnac Island probably around the same time.

Physical setting

Archaeological appraisal of past human occupation regionally requires an overview of physical settings, both past and present, in the area under study. In its westernmost onshore part, the Greater Swan Region study area covers a representative fraction of the seaward third of the Swan Coastal Plain - the lower reaches of the Swan River being the exceptional feature. This is a dune landscape, marked from West to east by a gradual lessening in topographical relief and an increase in biologically diverse terrestrial habitats, including in its eastern part the earlier noted corridor of fresh- water lakes and wetlands, rich in aquatic food resources. The study area’s offshore portion includes what in the given time frame would have been a 30-km-wide, low-lying stretch of exposed (‘emergent’) continental shelf, in places probably interspersed with wetlands and fresh-water lakes, and, as well, several slightly elevated features - today’s offshore islands.


Fig 1 The ‘Greater Swan Region’, Western Australia, showing lettered and numbered features, localities and sites.

DS - Darling Scarp
SR - Swan River/estuary
RI - Rottnest Island
GI - Garden Island
CI - Carnac Island
CS - Cockburn Sound
WS- Warnbro Sound
PP - Point Peron, Penguin Island
R - Rockingham
WP - Woodman Point
FFB - Five Fathom Bank
LB - Leighton Beach (Mosman Park, Victoria Street Station)
C - Cottesloe
CoF - City of Fremantle
CoP - City of Perth
BRP - Beeliar Regional Park
QD - Quindalup Dune System
SD - Spearwood Dune System
BD - Bassendean Dune System
GF - Guildford Formation
AD - alluvial deposits

box 1 - Helena River site (The Upper Swan site is 13 km north of this site.)
box 2 - Minim Cove, Rocky Bay, Blackwall Reach
box 3 - Fiona Stanley Health Precinct, North Lake, Bibra Lake

The surface and sub-surface sediments covering the onshore portion of the study area are defined by three dune systems of successive ages, as delineated for the whole of the Swan Coastal Plain. [5] The youngest of these, formed directly on the sea coast, is the Quindalup Dune System, consisting of white shelly sands with a varyingly small quartz grain percentage. In the northern half of the study area these dunes are present only in narrow strips - a larger exception being Woodman Point. However, in the study area’s southern part, the Quindalup dunes form an expansive plain composed of multiple beach ridges, which have successively accumulated (‘prograded’) from Lakes Coolongup and Walyungup westward some five to eight kilometres to form the present-day coast at Rockingham and Warnbro Sound.

The next oldest unit, the Spearwood Dune System, is a focal point within the study. This dune system is similar to other regional soil complexes developed within the Tamala Limestone, which is a major geological unit of Quaternary (mainly or largely Upper Pleistocene) age reaching over the littoral and submerged continental shelf along the Indian Ocean coast of south-western Australia. [6] Quartz dune soils of the Spearwood Dune System overlie and are intercalated (interbedded) with dune limestone formations. Weathering of dunes in the Spearwood Dune System has produced reddish-brown or orange-brown soils composed of iron-stained quartz sands. With continued weathering in the system’s older, more inland parts these soils become paler and grey-brown in colour.

The eastern edge of the study area reaches into the Bassendean Dune System, the third and oldest complex of dune soils noted here. The soils in this western margin of the Bassendean system, which covers much of the eastern two thirds of the Swan Coastal Plain, are brown to grey iron-humic podsols, ie deeply weathered quartz sands in which the limestone or carbonate substrate has been leached away. The north-south trending corridor of lakes and wetlands marking the study area’s eastern boundary is broadly aligned with the oldest, most easterly part of the Spearwood dunes and the western margins of the Bassendean sands (Fig 1).

A floral history covering the past 40 000 years in the onshore parts of the Greater Swan Region is provided by pollen records obtained from cores drilled into the peat of lake floors. [7] The radiocarbon dated pollen sequences logged in these studies indicate that much the same mosaic of woodland and wetland was present during the Late Pleistocene as exists today on the Swan Coastal Plain, though the east-west positioning of the different plant associations have shifted periodically through tens of thousands of years. These shifts in vegetative zones presumably were the result of repeated changes in global and localised climatic conditions, eg effective rainfall, from glacial times to the present day.

Stone artefacts and sites

Throughout the onshore part of the Greater Swan Region the material evidence for Aboriginal occupation from Late Pleistocene times to the arrival of European colonists consists almost entirely of stone artefacts. These scatters and concentrations of artefacts, whether surface or sub- surface, are broadly regarded as the remains of open-air campsites. Sites of all ages, including the colonial/modern era, are much more prolific in the Bassendean Dune System and other parts of the eastern parts of the region (Fig 1: Guildford Formation and alluvial deposits) than in its western parts. The region’s principal field researcher, Emeritus Professor Sylvia Hallam, reasoned that this site distribution reflects the presence of far more plentiful and reliable fresh water and bush foods in the eastern two thirds of the Swan Coastal Plain than in its Western third. [8] Four decades of at times concentrated survey and excavation in the whole of the onshore part of the Greater Swan Region have revealed no occupation sites featuring food remains, or structures connected with past subsistence such as fireplaces (hearths) or fish traps. Given this persisting lack of direct, material evidence, appraisal of past Aboriginal subsistence regionally can only be provisionally inferred from ethnographic and ethnohistoric accounts of regional Nyungar hunter-gatherer economy, and through review of site distributions relative to exploitable habitats. [9]

The flaked stone artefacts from sites across the whole of the onshore part of the Greater Swan Region are not generally recognisable as ‘tools’, as presumed from specimens having deliberately shaped (retouched) and sometimes use-damaged edges. Most flaked artefacts are merely tiny chips and larger flakes and chunks produced in the making of tools. Many of the stone artefacts from the oldest known sites in the general region are made of distinctive forms of chert; most others are made of quartz.


Figure 2 Four retouched flakes (tools) made of Eocene fossiliferous chert from the following find sites on Rottnest Island (with year of find).
a. Notched flake, Little Armstrong Bay, 1992.
b. Retouched flake, Fish Hook Bay, 1984.
c. Retouched flake, Little Armstrong Bay, 2010. This tool is shown (Fig 6) in situ in the artefact-bearing palaeosol at this site.
d. Retouched flake, Fish Hook Bay, 1992.


Figure 3 Rottnest Island with numbered Eocene chert artefact and other find sites and features as follows.
1 - Bathurst Point
2 - Little Armstrong Bay
3 - Charlotte Point
4 - City of York Bay
5 - Fish Hook Bay
6 - Barker Swamp
7 - salt lakes

One chert type for many millennia much used for artefacts throughout the Indian Ocean littoral of south-western Australia features marine microfossils of Eocene age - hence the label Eocene fossiliferous chert. [10] This particular kind of chert has yet to be identified in outcrops or other surface sources along the Indian Ocean coast, from Cape Leeuwin northwards some 700 kilometres to the region around Geraldton. In short, some of the most characteristic and prolific of the earliest identified stone artefacts throughout this long coastal zone are made of stone that has hidden local sources - perhaps buried beneath coastal dunes or, as is probable, submerged offshore. [11] Given the quantities of Eocene fossiliferous chert artefacts at sites in this zone it is improbable that this stone was traded in from elsewhere, eg from the Southern Ocean coast. [12] Geological and archaeological opinion is that throughout the south-western littoral outcrops of this stone ceased to be available around the time rising sea level reached its present position some 6000 years ago. Eocene fossiliferous chert artefacts are then practically a hallmark of Late Pleistocene and Early Holocene archaeological sites on the Indian Ocean littoral of south- western Australia, as well as along the Southern Ocean coast, at least as far to the east as the Esperance district, including the Archipelago of the Recherche. [13]

Of particular significance in the study are a half-dozen Eocene fossiliferous chert artefacts from Rottnest Island, located 19 kilometres offshore Fremantle (Figs 1-3). The presence of these distinctive chert artefacts on the island, as well as three Eocene chert flakes, a quartz flake and five calcrete (limestone) flakes found on Garden Island indicates wide-ranging human movements on this part of the emergent continental shelf during the Late Pleistocene and Early Holocene. [14]

Sites and radiometric dates

Radiometric dating of the earliest archaeological evidence from the study area is based on analyses of organic and mineral sediments in situ within the dune soils of the different stone artefact find sites. Two different dating techniques are used. The first method - isotopic dating - relies on radioactive decay of different chemical elements such as carbon, hence the term radiocarbon dating. Isotopic dating thus measures the residual radiation remaining in the sample, following centuries or millennia of radioactive loss or decay. The isotopic or radiocarbon age assays reported below are derived from analyses of charcoal or marine mollusc samples. The second means of age assay - luminescence dating - measures the amount of radiation absorbed by minerals through time, following last exposure to heat or light (eg from sunlight, bush fires or a hearth fire) before being buried within soil or other sediment with low background radiation. Luminescence assay of sediment samples determines then the amount of absorbed radiation and uses this information to calculate the sample’s age. The luminescence dates noted below from Rottnest Island and from onshore sites were obtained through analysis of quartz sand grains present in carbonate sediment samples collected from dune limestones and calcareous palaeosols.

Fiona Stanley Health Precinct test trench

In archaeological studies covering the past 40 000 years, radiocarbon dating is the most commonly used form of radiometric age assay, and is of crucial importance in establishing regional chronologies, no less in Australia than elsewhere. For this reason it is necessary here to show how radiocarbon dates are interpreted. A dated charcoal or other organic sample collected in situ from an intact soil profile or other undisturbed deposit should provide a reliable ‘minimum age’ for artefacts or other materials found at greater depths than the position of the dated sample.

For example, three radiocarbon dates for a large archaeological test trench dug in a dune soil in the construction site for the Fiona Stanley Health Precinct in the western part of the Bassendean Dune System (Fig 1, box 3) provide a minimum age for a single stone tool - an adze flake (Fig 4) - found in situ in this trench at a depth of approximately 140cm below the original ground surface. [15] The youngest date, ca 8500 years BP (‘Before Present’ is the conventional term), comes from a charcoal sample some 40cm higher up in the deposit than the depth from which the tool had been collected. A second date, ca 13 700 years BP, is from charcoal collected from a position 20cm higher than the tool’s find position. The third date, ca 32 000 years BP, also based on a charcoal sample, is from the same depth as the tool. The tool is conceivably the same age as this oldest dated charcoal sample. However given that this solo artefact find is in a featureless dune soil in which no stratigraphical layering is visible, a safer assumption is that the two charcoal samples from shallower depths provide a minimum age for the artefact, ie it is probably older than 13 700 years and is almost certainly older than 8500 years.


Figure 4 Four views of a mylonite adze flake, Fiona Stanley Health Precinct. Top view shows front of heavily used, adze working edge. Three lower views indicate that the piece is broken (deliberately truncated by a single blow with a hammerstone).

This dated tool is a deliberately broken (truncated) and re-used woodworking adze flake (Fig 4) made of a kind of cherty stone called mylonite, as found in various outcrops in the Darling Escarpment seventeen or more kilometres east of the find site. Far older than this mylonite adze flake are the two archaeological sites yielding the oldest radiocarbon dates in the Perth Metropolitan Region, which are situated in alluvial sediments of the Helena River (Fig 1, box 1) and the Swan River. The former site is dated as old as ca 29 000 years BP; the latter - called the Upper Swan site - is radiocarbon dated as much as ca 38 000 years BP. [16]

Minim Cove

A similar example of radiocarbon dating providing a minimum absolute age for stone artefacts comes from the first dated Late or terminal Pleistocene site anywhere in the Perth Metropolitan Region. In the 1970s, Eocene fossiliferous chert artefacts were found in situ in an orange-brown dune soil of the Spearwood Dune System, as exposed in the former Mosman Park town tip located at Minim Cove on the north shore of the Swan River (Fig 1, box 2). [17] A charcoal sample collected from an archaeological test pit excavated in undisturbed soil revealed in a grader cutting within the tip yielded a radiocarbon age of ca 9930 years BP, so giving a minimum age for sixteen Eocene fossiliferous chert and quartz artefacts found in situ in the soil as much as 100cm below the dated sample. Given the artefacts’ considerable depth below the sample position, it is likely that human occupation at this site took place a good deal earlier than the ca 10 000 BP date obtained.

Eocene chert artefacts were identified in the 1970s in two other, undated, exposures of Spearwood dune soils elsewhere in Mosman Park. Both find sites were cuttings made in railway platform construction, respectively at Victoria Street (immediately inland from Leighton Beach), and the next station ‘up the line’, Mosman Park. The small assemblage at the former site comprises a dozen Eocene chert flakes, including two retouched tools, as well as four quartz flakes; the latter yielded only a few chert flakes.

Lakes Cooloongup and Walyungup

Apart from the radiocarbon dated Minim Cove artefact assemblage, the only other finds of Eocene fossiliferous chert artefacts from the mainland portion of the study area for which a minimum radiocarbon age is determinable are in situ within a Spearwood dune soil (and former marine shore line) intermittently exposed for about eight kilometres in road cuttings along on the eastern shoulder of the north-south trending Mandurah Road, immediately east of Lakes Cooloongup and Walyungup (Fig 1). A sample of marine molluscs (cockle shells: Katelysia sp.) in situ in marine sands cut into the dune soil yielded a radiocarbon date of about 8000 years BP, showing that post-glacial sea level rise had brought the marine littoral to this position at that time, thus providing the artefact- bearing dune soil with its minimum age. [18] Given that this soil pre-dates the dated Early Holocene marine deposit, a terminal to Late Pleistocene age is assumed for the Eocene fossiliferous chert artefacts buried within it.

Rottnest Island

In highly intermittent searches, beginning in 1984, Charlie Dortch identified the earlier noted Eocene fossiliferous chert artefacts from Rottnest Island (Fig 2). All are from sea-cut cliffs in the island's constituent Tamala Limestone. A single Eocene fossiliferous chert artefact has been identified in situ in a prominent fossil soil (‘palaeosol’) at each of three sites: Bathurst Point, Little Armstrong Bay and Charlotte Point (Fig 3). In each case, the artefact-bearing palaeosol exposed in the sea cliff is deeply buried beneath a cemented dune (aeolian calcarenite or dune limestone) unit (Figs 5,6. The Eocene chert tool shown in Fig 6 is also illustrated in Fig 2c).


Figure 5 Sea-cut cliff face at Eocene fossiliferous chert tool find site on the east side of Little Armstrong Bay, Rottnest Island. Person at cliff base is pointing at artefact-bearing palaeosol (cf Fig 6). The palaeosol is bounded to top and bottom by massive, cross-bedded aeolian calcarenite (dune limestone) units. The mainland coast is visible on the horizon. (Mark Ward, Rottnest Island Authority, Fremantle)


Figure 6 Eocene chert tool (illustrated in Fig 2c) in situ in palaeosol horizon at Little Armstrong Bay, Rottnest Island (cf Fig 5). Note fresh graffiti carved into face of section. (Reg Yarran, Rottnest Island Authority, Fremantle)

Visible then in the sea cliffs at the Eocene chert artefact find sites and in many other cliff faces along the Rottnest littoral are palaeosols and dune limestone units dating to the time when the ‘pre-island’ locality consisted mainly of gently undulating sand hills, probably resembling those of the present-day island’s interior. These hills were originally soft calcareous dunes that through weathering processes (mainly dissolution and precipitation of calcium carbonate) over thousands of years developed into cemented palaeosols and dune limestones (aeolian calcarenite). Afterwards, during the Early to Middle Holocene, these limestone hills were sharply cut by rising seas to form the shoreline cliffs of the present-day island.


Figure 7 Eocene fossiliferous chert flaked fragment (artefact) embedded in rock- like cemented carbonate sediment (calcrete) within the prominent palaeosol at Bathurst Point, Rottnest Island.

The palaeosol horizons at these three Eocene chert find sites may be parts of a single, variably cemented, Late Pleistocene palaeosol unit exposed at these sites, and in several other cliff faces along the northern shore of the island. [19] None of the palaeosols at these sites appear to have in their sediments charcoal or other organic materials (eg marine mollusc shell) well suited for radiocarbon dating. However, an absolute age for the palaeosol at Bathurst Point and the artefact in situ within it is provided by thermoluminescence (TL) dating, which is one of two basic techniques of luminescence age assay. This dating was carried out more than a decade ago in a study independent of the archaeological research reported here. [20] Thermoluminescence analysis of quartz grains from the predominantly carbonate sediments in the limestone units below and above the artefact- bearing palaeosol at Bathurst Point respectively yielded dates of ca 67 000 years and ca 20 000 years BP. It is clear then that the Eocene chert artefact (Fig 7) in situ in the palaeosol bounded by the two dated dune limestone units is no less than 20 000 years old, and perhaps is considerably older, judging by a TL date of ca 34 000 years BP yielded by a sediment sample collected from a comparable stratigraphic succession in an eroded limestone headland at The Basin, 500 metres to the west of Bathurst Point. Provisional correlation of the Bathurst Point palaeosol with the artefact- bearing palaeosols at Little Armstrong Bay and Charlotte Point suggests that the Eocene chert artefacts found in situ at all three sites may be datable Within much the same age range.

The younger TL date from Bathurst Point is congruent with a radiocarbon date of ca18 660 years BP, based on a sample from a charcoal lens (‘thin layer’) in situ within a palaeosol section exposed in a sea cave deeply buried beneath a succession of dune limestone units at Fish Hook Bay, at the western end of Rottnest Island. Two Eocene chert tools come from the summit of the limestone cliff at Fish Hook Bay, though no stone artefacts or other cultural evidence have been identified in the vertical face of the palaeosol in which the dated charcoal lens is exposed. The radiocarbon age of this palaeosol, like the ca 20 000 year BP thermoluminescence date for the artefact-bearing palaeosol at Bathurst Point (Fig 7), is evidence for widespread soil development on the ‘pre-island’ land mass during the Late Pleistocene (see Discussion).

The two above-noted Eocene fossiliferous chert tools from Fish Hook Bay were identified in rubble derived from the erosion of a palaeosol and overlying dune limestone unit at the cliff summit. One of these (Fig 2b) has semi-lustrous (only slightly weathered) surfaces showing that it has not been subjected long to the severe weathering conditions - including copious salt spray from the sea, wind-blown sand, intense sunlight most days of the year and high ground surface temperatures - prevailing at this exposed cliff summit. The other Eocene chert tool (Fig 2d) has somewhat more weathered surfaces, though it, too, was probably in situ within a stable deposit until the past few centuries or millennia.

A third Eocene chert tool (Fig 2a), from Little Armstrong Bay, had almost certainly fallen from the vertical face of the artefact-bearing palaeosol at this site (Figs 5, 6). Since this find and the two Eocene chert tools from Fish Hook Bay (Figs 2b, 2d) were not in situ when identified, they cannot be directly dated, though they clearly belong to the period of low sea levels under review, when Eocene chert outcrops presumably were accessible on the emergent continental shelf and Rottnest Island, along with Carnac Island and Garden Island, was part of the mainland.


Site numbers in the study area

The Western Australian Department of Indigenous Affairs (DIA) register of Aboriginal sites presently lists 83 solo finds or surface scatters of stone artefacts within the onshore portion and the offshore islands of the study area. Among the onshore archaeological sites are two undated human burials and several stone artefact scatters at sites that also have mythological or ceremonial significance, suggesting they were still in use during recent centuries. Other stone artefact surface scatters include retouched and chipped pieces of glass and ceramics, showing that the find sites are at least in part datable to the colonial/ modern era.

Absolute ages for the archaeological finds in the study area mainly relate to the earliest known phases of human occupation in the region, as indicated by the Late and terminal Pleistocene ages of the four radiometrically dated stone artefact find sites - three mainland and one island - described above. Notably lacking in the study area are stone artefact assemblages featuring highly distinctive tool types called microliths or backed blades, as occur in several sites in more easterly parts of the Perth Metropolitan Region, eg at Lake Monger. [21] Microlithic tool assemblages were first prolific in south- eastern Australia as early as 4000 years BP and about a thousand years later became widely distributed in the south-west of Western Australia. [22] Thus the maximum age of the microlithic tool assemblages at Lake Monger is unlikely to be much greater than 3000 years old, the age in which these tools (backed blades, geometric microliths, backed points, etc.) come into common use regionally.

Hallam’s comprehensive, long term survey of Aboriginal occupation patterns in the Perth Metropolitan Region showed throughout the time span of occupation ranging from Late Pleistocene times to the historic period ‘. . .the heaviest concentration of sites, and by implication of resource usage,  lay not adjacent to the coast, but around the lakes and swamps of the coastal sandplain  and in the alluvial and scarpfoot zones ... [23] This zone of concentration is situated along and to the east of the corridor of fresh-water lakes and swamps running through the central part of the metropolitan region. In the onshore part of the Greater Swan Region study area outlined in this paper, this corridor of fresh-water bodies is where many of the stone artefact scatters listed in the DIA site register are located. Included is a stone artefact scatter on the north shore of North Lake, radiocarbon dated to the Late Holocene, ca 2300 years BP. (Fig 1, box 3).

Sea floor appraisal

Given the concentration of sites in the present-day hinterland of the Swan Coastal Plain, it would seem reasonable to think that the environs of the present-day sea coast may also have been a focal point of occupation during the Late Pleistocene when sea level was low, and the emergent continental shelf reached some 30 kilometres further westward. As earlier noted, fresh-water bodies - perhaps similar to North Lake, Bibra Lake and the swamps and lakes in Beeliar Regional Park - are likely to have been present in the off-shore zone of the continental shelf, eg the deepest part of Cockburn Sound and in the trough of deeper water between the western shore of Garden Island and the submerged ridge of Tamala Limestone (designated ‘Five Fathom Bank’ on nautical charts) running 20 kilometres northward toward Rottnest Island (Fig 1). In the 1990s, several dive searches for former terrestrial sites on some of these parts of submerged shelf hardly surprisingly failed to identify any ancient cultural evidence. [24] However, calcrete root or solution pipes (‘pinnacles’) and other Tamala Limestone landscape features are occasionally identifiable at varying depths (Fig 8). Systematic recording of former landscape features on the sea floor has the potential to reveal much about past terrestrial settings, including ones likely to have been frequented by hunter-gatherer groups.


Fig 8 Small calcrete root or solution pipe (‘pinnacle’) in outcrop of Tamala Limestone on the sea floor near Little Island, Marmion Marine Park, 40km north of Fremantle. (Photograph by Clay Bryce, WA Museum, Perth)

An indirect approach to sea floor study was attempted in the 1990s through searches of carbonate shell and sand tailings dredged from the floor of Cockburn Sound in industrial operations.  Present among the tailings are occasional large lumps of reddish cemented soil and pieces of preserved wood indicative of former terrestrial settings buried beneath marine shell and sand accumulated since post-glacial sea level rise created the inshore marine body of Cockburn Sound. These lumps of cemented soil clearly resembling soils formed in the Spearwood Dune System, as well as the pieces of tree wood recovered from the dredge tailings, suggest that stone artefacts or other cultural evidence could also be buried perhaps only one or two metres below the marine floor sediments. [25]

Archaeological potential of Rottnest Island

As noted, the six Eocene fossiliferous chert artefacts [26] identified on Rottnest Island (Figs 2,6,7) were in situ within or almost certainly derive from fossil soil (palaeosol) horizons (Figs 5, 6) exposed in wave-cut cliffs of dune limestone (aeolian calcarenite). These palaeosols developed during pauses in dune building when physical conditions were perhaps less arid and less windy, and thus more suitable for the formation of vegetative habitats attractive to animal and human populations. The palaeosol horizons exposed in these cliffs, eg the artefact-bearing palaeosol unit at Bathurst Point, TL dated to the Late Pleistocene, often incorporate rock-like calcrete (indurated limestone) horizons, considered to have formed under humid/subhumid conditions regionally. The presence of massive calcrete horizons associated with these palaeosols is then evidence that during Late Pleistocene low sea levels the ‘pre-island’ locality was not always deficient in fresh water. [27]

The Eocene fossiliferous chert artefact in Fig 7 remains embedded in a slab of rock-like calcrete developed in the prominent palaeosol at Bathurst Point, from which the slab was cut with an angle grinder in 2003. This dated find implies human occupation in the locality before or during a Late Pleistocene humid phase.

Definitive fossil evidence for the presence of an ancient fresh-water body at Rottnest dates to the Early to Middle Holocene, shortly before the island’s separation from the mainland. The biotic remains identified from a core drilled at Barker Swamp include fresh-water mollusc shells and crustaceans, as well as pollen indicating the presence of Tuart and Jarrah eucalypt woodland near this fresh-water lake or swamp (Fig 3). This evidence is from the ‘Lower Peat’, which is the lowermost sediment layer recorded in this core, and radiocarbon dated ca 7220-6850 years BP. [28] These findings from Barker Swamp suggest that other of the island’s eight existing swamps were also fresh-water bodies situated in congenial terrestrial settings. If so, these features could have supplied hunter-gatherer groups with aquatic and woodland foods and fresh water perhaps not available at other, now submerged localities this far out on the then emergent shelf in the Early Holocene. The only other topographical features on Rottnest likely to have been sources of abundant or at least appreciable supplies of fresh water during times of low sea level would seem to be the limestone caverns and depressions now filled by the present-day island's salt lakes. [29]

Test excavations around the margins of one or other of the island’s formerly fresh-water swamps and salt lakes may well yield stone artefacts or other evidence for Aboriginal occupation on the shores of the fresh- water bodies once present there. Rottnest Island thus has an archaeological potential additional to that of further possible rare finds of stone artefacts in Tamala Limestone palaeosols exposed in sea cliffs.

The four Rottnest Island cliff-face and cliff-top Eocene chert artefact find spots (eg Fig 5) discussed in this paper are striking examples of archaeological sites whose present topographical and environmental settings are radically different from those existing at the time of ancient human occupation. These differences are of course only to be expected, given the rising seas that created the island, and in doing so cut the cliff faces of its littoral, which has been further sculpted and reduced through a subsequent six millennia of wind and water erosion. These climatic and geomorphological events have laid bare much of the ‘pre-island’ Tamala Limestone stratigraphy, so revealing the presence of the ancient Eocene chert artefacts (Figs 2, 6, 7) that make a compelling case for human activities on this outer part of the emergent continental shelf during the Late Pleistocene.

Older evidence?

Radiocarbon dating of Eocene chert and other stone artefacts from the two mainland sites at Fiona Stanley Health Precinct and Minim Cove, and luminescence dating at the Bathurst Point Eocene chert artefact find site on Rottnest Island indicate human occupation in the western parts of the Greater Swan Region from more than 10 000 years ago to at least 20 000 and perhaps as much as 30 000 years ago. [30] This age range is supported by the radiocarbon-dated marine shoreline record implying a Late to terminal Pleistocene age for the artefact-bearing Spearwood dune soil immediately east of Lakes Cooloongup and Walyungup.

The above dating results are a good deal younger than radiometric dates for archaeological sites in several regions in Australia indicating that Aboriginal populations were dispersed across the continent more than 40 000 years ago. This older evidence is well represented in Western Australia, with dated 35-45 000 year old occupation at sites in the Kimberley, the Pilbara, and the South-west, including the Perth Metropolitan Region, where mention has already been made of the Upper Swan and Helena River sites (Fig 1, box 1). [31] One of the most closely studied, most comprehensively dated and among the oldest, reliably dated sites anywhere in Australia is the limestone cave occupation site of Devil’s Lair in the Leeuwin-Naturaliste Region, where the stratigraphically lowermost of several fireplaces (hearths) identified in the sandy cave floor deposit is dated by radiocarbon and luminescence assay to 45 500 years BP, with stone artefacts found deeper in the floor deposit dated by these and other means of age assay to ca 47-48 000 years BP. [32]

The study area defined in this paper has no limestone caves with floor deposits like those in the Leeuwin-Naturaliste Region, or as found in Yanchep National Park, 60 kilometres north of Fremantle. A good question then is what kinds of topographical features in the western parts of the Greater Swan Region could yield discoveries of sites as old as those mentioned above? For the present, the best chances for finding old sites in the study area would seem to be in the Tamala Limestone cliffs and headlands on the offshore islands and along the mainland coast in several places between Cottesloe and Warnbro Sound. Dating of quartz samples from some of these dune limestone features by means of thermoluminescence assay (TL) was carried out over a decade ago by the team of three dating specialists from the University of Wollongong that also determined the Late Pleistocene age of Bathurst Point on Rottnest Island. This team obtained even greater ages through TL assay of sediment samples from cliff faces at Cottesloe, Fremantle, Point Peron and Penguin Island. [33]

Searches for palaeosols or other possible artefact-bearing deposits are much needed at these dated coastal sites. A beginning has already been made at Point Peron, where three Eocene fossiliferous chert artefacts have been identified in situ within orange-brown soils of Spearwood type and a fourth Eocene chert artefact still lies embedded in a cemented carbonate unit in the heavily eroded Tamala Limestone landscape at this site. [34] Further radiometric dating assay at Point Peron, combined with renewed searches for Eocene chert and other stone artefacts in palaeosols exposed in Tamala Limestone cliffs and outcrops on the mainland and adjacent offshore islands may yet provide an additional dated record for early human occupation in the study area. This potential is not limited to the Greater Swan Region study area , but holds equally true for all the Tamala Limestone sea cliffs and outcrops on the coastlines and offshore islands along the Indian Ocean coast of south-western Australia, and in similar Late Pleistocene-age coastal limestone features along the region’s Southern Ocean coast, eastward to the Archipelago of the Recherche and beyond. [35]

A final comment

All that is significant in archaeology is by no means ‘old’. A coherent and relatively complete archaeological record of the Aboriginal hunter- gatherer past in the Greater Swan Region can only emerge through further investigation of sites ranging in age from the earliest times until the colonial/modern era.


The co-authors wish to acknowledge with sincere thanks the help given by those persons who made it possible to produce this paper. Nyungar elders generously gave their permission and support in 2011 for Charlie Dortch to carry out searches for and to collect ancient stone tools on Rottnest Island. Harriet Wyatt, Hannah Eames, Reg Yarran, Wendy Bradshaw and Patsy Vizents of the Cultural Heritage Unit, Rottnest Island Authority provided essential help and cooperation in enabling Charlie to carry out several one-day field searches on the island since late 2010. Patsy and Mark Ward, Design and Production Coordinator, Rottnest Island Authority drew the map in Fig 3. Mark also produced the stunning coastal shot in Fig 5. Reg took the diverting photograph in Fig 6.

Several Nyungar families of the Perth region supported and helped Joe Dortch in the archaeological investigations he directed at the Fiona Stanley Health Precinct building site. Moya Smith and Ross Chadwick of Anthropology and Archaeology Department, Western Australian Museum provided much help and encouragement, and Alice Beale of that department photographed stone artefacts. Francois Mazieres of Eureka Archaeological Research and Consulting, University of Western Australia drew the line drawings of most of the artefacts in figures 2 and 4. Phil Playford kindly reviewed a late draft of the paper.

Fremantle Studies Day, 2010


1 The many published records of these later periods of Aboriginal occupation in the onshore portion of the Greater Swan Region include the following four works covering, respectively, Nyungar history, ethnohistoric, ethnography and language, as well as hunter-gatherer land usage and economy: L Collard, ‘Kura, Yeye, boorda Walwalinyup: from the past, today and the future Fremantle’ , Fremantle Studies, v 5, 2007, pp 9-21; N Tindale, Aboriginal Tribes of Australia, University of California, Berkeley, 1974; GF Moore, A descriptive vocabulary of the language in common use amongst the aborigines of Western Australia, (facsimile ed), University of Western Australia, Perth, 1978 [1884]; SJ Hallam, Fire and hearth, a study of Aboriginal usage and European usurpation in South-western Australia, Australian Institute of Aboriginal Studies, Canberra, 1975.

2 PP King, Narrative of a survey of the intertropical and western coasts of Australia, performed between the years 1818 and 1822 ... John Murray, London, v 1 , 1827, p 166; FA Peron and L Freycinet, Voyage de découvertes aux Terres Australes ... Paris, Imprimerie Impériale, 1807, p 119.

3 Moore, A descriptive vocabulary of the language in common use amongst the Aborigines of Western Australia, p 8.

4 PE Playford, ‘Geology and hydrogeology of Rottnest Island’ in Geology and hydrogeology of carbonate islands, Developments in sedimentology, v 54, HL Vacher and T Quinn (eds), 1997, p 804.

5 WM McArthur and E Bettenay, Development and distribution of soils of the Swan Coastal Plain, Western Australia, CSIRO Australia, Soil Publication No. 16, 1960 (second edition 1974). WM McArthur, Reference soils of south-western Australia, Department of Agriculture, Western Australia, 1991, pp 27-29.

6 PE Playford, A E Cockbain and G H Low, Geology of the Perth Basin Western Australia. Geological Survey of Western Australia, Bulletin 124, 1976; PE Playford ‘Geological research on Rottnest Island’, journal of the Royal Society of Western Australia, v 66, 1983, pp 10-15; PE Playford, Guidebook to the geology of Rottnest Island, Geological Society of Australia, WA Division, WA Geological Survey of WA, Perth, 1988.

7 J Newsome and EJ Pickett, Palynology and palaeoclimatic implications of two Holocene sequences from south-western Australia, Palaeogeography, Palaeoclimatology, Palaeoecology, v 101, pp 245-261; EJ Pickett, The late Pleistocene and Holocene vegetation history of three lacustrine sequences from the Swan Coastal Plain, Southwestern Australia. Unpublished Ph D thesis, Geography Department, The University of Western Australia, 1997.

8 SJ Hallam, ‘Coastal does not equal littoral’,Australian Archaeology,v 25, 1987, pp 10-29.

9 SJ Meagher, ‘The food resources of the Aborigines of the south-west of Western Australia’, Records of the Western Australian Museum, v 3, 1973, pp 14-65; Hallam, Fire and hearth; Hallam, ‘Coastal does not equal littoral’.

10 JE Glover, ‘The geological sources of stone for artefacts in the Perth Basin and nearby areas’, Australian Aboriginal Studies, v 1, 1984, pp 17-25.

11 Geochemical differences detected in Eocene chert fossiliferous artefacts sampled from 78 mainland archaeological sites broadly dispersed between Geraldton and Cape Leeuwin accord with the hypothesis that now concealed Eocene chert outcrops or other sources quarried by Aboriginal tool makers are widely distributed, and lie offshore the Indian Ocean littoral. JE Glover and R Lee, ‘Geochemistry and provenance of Eocene chert artifacts, south-western Australia’, Archaeology in Oceania, v 19, 1984, pp 16-20.

12 Only two miniscule outcrops of this stone are recorded on the Southern Ocean coast west of King George Sound, and neither features artefacts or any evidence of quarrying or other human activities. CE Dortch and G Gardner, ‘Archaeological sites in the Northcliffe district, Western Australia’, Records of the Western Australian Museum, v 4, 1976, p 268.

13 Glover, ‘The geological sources of stone for artefacts in the Perth Basin and nearby areas’; Glover and Lee, ‘Geochemistry and provenance of Eocene chert artifacts, south-western Australia’; CE Dortch, ‘Rottnest and Garden Island prehistory and the archaeological potential of the adjacent continental shelf, Western Australia’, Australian Archaeology, v 33, 1991, pp 40-41; MV Smith, Recherche a L’Esperance.' a prehistory of the Esperance region, South- western Australia, Unpublished Ph D thesis, University of Western Australia, 1993, p 131. Present in the Esperance district are two other varieties of Eocene fossiliferous chert, each distinguishably different from the Eocene fossiliferous chert as found both along the Indian Ocean coast, and the Southern Ocean coasts. Ibid. pp 128-129. Because one of these other variants is still found locally in outcrops onshore, the case for artefacts made from Eocene fossiliferous chert deriving from sources submerged by post-glacial sea level rise is a little more complex in Southern Ocean coast districts east of King George Sound than it is thought to be on this coast to the west of the Sound, and particularly as considered for the Indian Ocean coast of South- western Australia.

14 CE Dortch, ‘Rottnest and Garden Island prehistory and the archaeological potential of the adjacent continental shelf, Western Australia’, pp 40-41; CE Dortch and PA Hesp, ‘Rottnest Island artifacts and palaeosols in the context of Greater Swan Region prehistory , journal of the Royal Society of I/Western Australia, v 77, 1994, p 30. In the 1990s Charlie Dortch carried out a brief search for stone artefacts on Carnac Island (Fig 1), but found no stone artefacts or other evidence of Aboriginal occupation.

15 J Dortch, Report on the Aboriginal heritage survey cf the proposed Fiona Stanley Health Precinct, Murdoch WA. (unpublished consultancy report). The mylonite adze flake in Fig 4 has been deliberately truncated (broken) by a single blow with a hammer stone against its bulbar or inner face. Its lower right-hand corner (as depicted in the lowermost part of the lower central view in Fig 4), including the negative flake scar produced when the piece was truncated, shows edge-wear or damage consistent with the piece having been used (and presumably re-hafted) as an adze stone after having been truncated. This reliably dated specimen contributes to the continental record of considerable antiquity for the hafting of stone tools (adze flakes, hatchets, spear barbs, etc) to wooden handles and shafts. DJ Mulvaney and J Kamminga, Prehistory of Australia, Allen and Unwin, St Leonards, 1999, pp 43-44, 231, 245-254.

16 A preliminary report for the Helena River site is: M Schwede, Supertrench - phase two: a report on excavation results.’, In M Smith (ed), Archaeology at ANZAAS 1983, Western Australian Museum, Perth, 1983, pp 52-62. The Upper Swan site report is: RH Pearce and M Barbetti, ‘A 38,000 year old archaeological site at Upper Swan, Western Australia’, Archaeology in Oceania, v 16, 1981, pp 173-178. The following comment refers to both sites. ‘The association of charcoal [sampled in the radiocarbon dating] and artefacts, and the stratigraphic integrity of these [two] sites, must be conceded to be open to argument.’ S Bowdler, L Strawbridge and M Schwede, ‘Archaeological mitigation in the Perth Metropolitan Region’, Australian Archaeology, v 32, 1991, p 24. Review of the evidence from both sites is clearly warranted.

17 J Clarke and CE Dortch, ‘A 10 000 year BP radiocarbon date for archaeological finds within a soil of the Spearwood Dune System, Mosman Park, WA’, Search, v 8,1977, pp 36-38.

18 Both lakes Cooloongup and Walyungup are relict (former) marine bodies; a half-dozen samples of marine mollusc shells collected from these lakes shores have yielded radiocarbon dates in the range 6000 to 7000 years BP, which is shortly before both were cut off from the sea by dune ridge progradation that continued westward for some six thousand years before stabilising along the present-day coast between Rockingham and Warnbro Sound. DJ Searle, V Semeniuk and PJ Woods, ‘Geomorphology, stratigraphy and Holocene history of the Rockingham-Becher Plain, South-western Australia’, journal of the Royal Society of Western Australia, v 70, 1988, pp 89-109. Although outside the scope of this paper, concerned as it is with the earliest evidence for human occupation in the Greater Swan Region, these prograded Holocene beach sands (multiple beach ridges attributable to the Quindalup Dune System) running westward from the Mandurah Road to the present coast seem as lacking in archaeological evidence for Aboriginal occupation as the remnant patches of Quindalup dunes in the northern half of the onshore part of the study area in this paper, and as reported by Hallam for the areas of Quindalup Dunes in her 15 km wide study transect running east-west across the width of the Swan Coastal Plain, immediately to the north of the study area here. Hallam, ‘Coastal does not equal littoral’, Map 2, Tables 1, 2.

19 ‘The presence of what may be a single palaeosol unit extending for several km along Rottnest Island’s northern shore offers perhaps the best opportunities for further prehistoric investigations on the island.’ CE Dortch and PA Hesp, ‘Rottnest Island artifacts and palaeosols in the context of Greater Swan Region prehistory’, p 31.

20 DM Price, BP Brooke and CD Woodroffe, ‘Thermoluminescence dating of aeolianites from Lord Howe Island and south-west Western Australia’, Quaternary Science Reviews, v 20, 2001, pp 841-846. The other luminescence dating technique in current use is optically stimulated luminescence (Q SL).

21 These are large assemblages of finely made microlithic tools (‘geometric’ microliths, backed points, tiny adze flakes), mainly flaked from mylonite (probably multiple sources: Darling Scarp) unsystematically collected around Lake Monger during velodrome and other lake-shore excavations during the 1960s and never described in publication. These assemblages are held in the archaeological collections of the Western Australian Museum.

22 Mulvaney and Kamminga, Prehistory of Australia, p 235. Line drawings on this page depict two geometric microliths and two backed points. For radiocarbon ages for the oldest microlithic assemblages in the South-west cf CE Dortch and G Gardner, ‘Archaeological investigations in the Northcliffe district, Western Australia’, pp 269-274; RH Pearce, ‘Changes in artefact assemblages during the last 8000 years at Walyunga, Western Australia’, journal of The Royal Society of Western Australia V 61, 1978, pp 1-10.

23 Hallam, ‘Coastal does not equal littoral’, p 10.

24 CE Dortch, ‘The erosion factor in sea floor investigations of prehistoric occupation deposits. ‘Australian archaeology, v 47, 1998, pp 30-32 This paper emphasises the destructive forces of erosion due to surge, particularly that taking place at the time a near shore archaeological site is first inundated by rising seas. It makes the point that perhaps the best chance for finding prehistoric occupation sites on the sea floor would lie in the search for former soils or cave floor deposits that are indurated (cemented) or capped by calcium carbonate (calcrete) or other cemented sediment. The diver searches at localities offshore the western shore of Garden Island and elsewhere in the study area were encouraged by the successful underwater prehistoric studies at Lake Jasper, 50 km east of Cape Leeuwin. CE Dortch, ‘Prehistory Down Under: archaeological investigations of submerged Aboriginal sites at Lake Jasper, Western Australia’, Antiquity, v 71, 1997, pp 116-123.

25 CE Dortch, ‘Rottnest and Garden Island prehistory and the archaeological potential of the adjacent continental shelf, Western Australia’, pp 41-42.

26 Two other ancient stone artefacts from the island are both made of calcrete (limestone). One flake was recovered from a calcareous sand dune blow-out near the summit of a limestone ridge 300 m east of Fish Hook Bay. The other probably derives from the artefact-bearing palaeosol at Little Armstrong Bay. These two heavily weathered flakes are the only ancient artefacts yet identified on Rottnest Island not made of Eocene fossiliferous chert. Ibid. p 38, Fig 2.

27 Calcrete is a crypto-crystalline calcium carbonate deposit typically forming in dune soils, both in the aerated, upper zones and in water-saturated lower zones (water tables). V Semeniuk, ‘Holocene climate history of coastal southwestern Australia using calcrete as an indicator’, Palaeogeography, palaeoclimatology, palaeoecology, v 53, 1986, p 296. Field studies show that ‘... modern calcrete occurs mainly in humid and subhumid climatic settings in Holocene coastal sands ....’ Ibid.

28 J Backhouse, ‘Holocene vegetation and climate record from Barker Swamp, Rottnest Island, Western Australia’, journal of The Royal Society of Western Australia, v 76, 1993, pp 53-61. Pollen from the core taken at Barker Swamp imply that the original fresh water body was ‘... surrounded by sedges and Callitris [native pine] low woodland, with a restricted jarrah/ tuart woodland present nearby.’ Ibid. p 53. Today, six of the island’s eight, formerly brackish or fresh-water swamps are hyper-saline, mainly due to their being mined for road-building material (eg marl) in the early 1970s. Ibid.

29 Playford, ‘Geology and hydrogeology of Rottnest Island’, p 786.

30 Some of the Rottnest Island Eocene fossiliferous chert artefact find sites have been dated by a third means of absolute age assay called aspartic acid racemisation technique (AAR), though the results, suggesting ages two to even three times greater than the luminescence age reported here for the Bathurst Point chert artefact find site, still seem provisional. PA Hesp, CV Murray-Wallace and CE Dortch, ‘Aboriginal occupation on Rottnest Island, Western Australia, provisionally dated by Aspartic Acid Racemisation assay of land snails to great than 50 ka’, Australian Archaeology, v 49, 1999, pp 7-12.

31 R Gillespie, ‘Dating the first Australians’, Radiocarbon, v 44, 2002, pp 455 -472; Mulvaney and Kamminga, Prehistory of Australia, pp 135-138, 143-144, 176-178, 204. The two latter authors point out that the Upper Swan site has yet to be adequately reported. Ibid. p 138. Cf Pearce and Barbetti, ‘A 38,000 year old archaeological site at Upper Swan, Western Australia’.

32 CSM Turney, MI Bird, LK Fifield, RG Roberts, M Smith, CE Dortch, R Grün, E Lawson, LK Ayliffe, GH Miller, J Dortch and RG Creswell, ‘Early human occupation at Devil’s Lair, southwestern Australia 50,000 years ago’, Quaternary Research, v 55, 2001, pp 3-13. For current interpretive analyses and inventories of the Devil’s Lair stone and bone artefact assemblages and terrestrial faunal remains see J Dortch, Palaeo-environmental change and the persistence of human occupation in South-western Australian forests, British Archaeological Reports International Series 1288, 2004.

33 Price, Brooke and Woodroffe, ‘Thermoluminescence dating of aeolianites from Lord Howe Island and south-west Western Australia’, p 843, Table 2.

34 It is uncertain whether stratigraphic association can be established between the Eocene chert artefact identified in situ in a cemented carbonate unit at Point Peron and a sediment sample collected from this site that through thermoluminescence (TL) assay Price et al. obtained an age of ca 78 000 year BP. Ibid. In any case, an age this great for early human presence in Australia would be as controversial, though one would expect not to be as groundless, as a 70 000 year age estimate for human occupation on Rottnest Island, published by a newspaper journalist. C Amalfi, ‘Island's fossil secret’, The West Australian, 19 July 2003.

35 CE Dortch and K Morse ‘Prehistoric stone artefacts on some offshore islands in Western Australia’, Australian Archaeology, v 19, 1984, pp 31-47; M V Smith, Recherche a L’Esperance.' a prehistory of the Esperance region, South-western Australia.

Garry Gillard | New: 1 May, 2018 | Now: 16 December, 2018